Surface plasmon resonance as a tool in the functional analysis of an immunodominant site in foot

Ph. D. Thesis

Ph. D. Program: Fundamental Chemistry,Organic Chemistry, 1997-1999.

Surface plasmon resonance as a tool in the

functional analysis of an immunodominant site

in foot-and-mouth disease virus

Paula Alexandra de Carvalho Gomes

Department of Organic Chemistry, Faculty of Chemistry.Division of Experimental and Mathematical Sciences.

University of Barcelona, 2000

Department of Organic Chemistry, Faculty of Chemistry.Division of Experimental and Mathematical Sciences.

University of Barcelona, 2000

Dissertation presented by

Paula Alexandra de Carvalho Gomes

to apply for the degree of Doctor in Chemical Sciences

at the University of Barcelona

and revised by

Dr. David Andreu

En primer lugar, me gustaría agradecer al Dr. David Andreu, supervisor de la presente tesis, y al Dr.Ernest Giralt por la calurosa acogida en su laboratorio, dándome la gran oportunidad de formar partede su grupo de investigación. También quisiera agradecerles las fructuosas discusiones y el constanteapoyo siempre que me sentía perdida entre “biacores” y péptidos...

A Jaume y Chary, de los SCT, porque siempre han traído un rayo de Sol a mis días de autismo cuandotocaba hacer biacore (al final, la “unidad” paula se puede desacoplar del BIACORE 1000!)

Estoy igualmente en deuda con Wendy Fernández, Dr. Núria Verdaguer y Dr. Ignacio Fita porpresentarme el guapísimo mundo de la cristalografía de proteínas y por toda la paciencia que tuvieronguiando mis pasitos infantiles en este campo de investigación (eh, Wendy? A que tendrás undescansito, ahora que no estaré todo el rato preguntándote “y ahora qué se hace?”...muchas gracias,de verdad! Eres un sol!).

Bueno, y ahora...cómo agradecer a todos vosotros que sois el alma del inolvidable (y enorme!) “grup10”? Vuestra alegría y buena voluntad inagotables, las palabras de consuelo siempre que la ciencia seresistía a colaborar (lo que suele ser la regla, no la excepción...), las tardes de cine (lo siento, sigopreferiendo las V.O.S.!!! ...bueno, creo que solo tú, Mac, me comprendes en este tema...), las noches de“vicio” por Poble Nou (no diré nombres para no daros mala fama, pero sabéis quienes sois), lasexcursiones “torpedillas” por la montaña (demasiado “sanas” para mí, pero siempre podía gritar“Javiiii, no me dejeeees! Me da miedo pasar por aquíiiii!!!!”), las clases de RMN y de ornitología delgrande (en altura y como persona) Doctor Millet, las carcajadas siempre que Maria José (la “femmeformidable”) decía alguna palabrota en gallego... La incansable Super-Molina y la mítica CristinaCarreño (la mejor guía práctica de laboratorio que he conocido nunca)... el equipo “carmanyola”, quetambién tenía compañeros “del otro pasillo”, y que animaba la hora de comer con temas de actualidadtan diversos como la receta de codornices de Cris Chiva o la película de anoche, todo estoacompañado por unas negociatas sospechosas relacionadas con las mafias de los puntos Danone...DelFresni, qué tal si nos montamos un restaurante? (tu pones el arroz negro y yo el pulpo, vale?)

Tampoco puedo olvidar el super buen rollo del “lab 3”...empezando por Eva, mi compañera “aftosa” ysiguiendo por Alberto, con quien disfruté buenos momentos musicales “after nine”, cuando ya noquedaba nadie (ní tu, Paul...) y podíamos poner la radio a “toda leche” (aunque los jóvenes no nosentendieran cuando nos animábamos con algun “greatest hit” de los 80... debe ser el “generationgap”...), y, como no, mi primera jefecilla, Mary-Light (ahora Señora Dueña), toda ella un encanto depersona... también las nuevas generaciones han sido clave para trabajar en el lab 3 con placer:Judit[h], aunque te estés haciendo mayorcita (cómo va Peter Pan?), seguirás siendo “mimonstruito”...Ferrán, todo él un personaje, tan carismático como su “ForFi”...y el último fichage,Miquel (buen chaval), con carita de ángel y comentarios de diablete...

Luego la inolvidable “Granada Connection”: la sonriente y altruísta Melena (buena contadora deleyendas nazaríes), Miriam Royo, mi compañera en foto-adicción (podremos vender nuestras fotos a laNational Geographic, no?), y la dulce Lorena, de calidades humanas inigualables (aunque con ciertosproblemillas relacionados con calcetines blancos...).

Para cerrar el “sector en castellano”, debo expresar mi gratitud a Txell (alias, Cielitos, alias Mary-Heaven) por su amistad y su valor, sea peleándose con retro-inversos ó con 400 millones de péptidosen forma de ocho, sea aprendiendo el portugués (creo que Lia está de acuerdo en darte “matrícula”),sea pateando todo Porto (con sus infinitas subidas y bajadas) arrastrada (literalmente) por mí...

Antes de mais devo agradecer à Prof. Doutora Maria Joaquina Amaral Trigo, não só por me apoiar eacreditar em mim profissionalmente, mas também pela sua valiosa amizade.

Agradeço à Fundação Calouste Gulbenkian (Lisboa, Portugal) a bolsa de doutoramento que me foiconcedida, o apoio financeiro para assistir a congressos internacionais e a gentileza com que semprefui tratada.

À Lia, catalã por motivos geográfico-sentimentais, portuguesa pelo idioma (si beim qui teim um cerrtosôtaqui dá cidádji dji São Paulo, né?) e com a bondade e a doçura de brasileira que é. És daquelasjóias raras com quem se pode sempre contar.

Aos meus amigos de sempre, Helena, Carla e Alfredo, por continuarmos a ser uma espécie deD’Artagnan e os, ou melhor, as Três Mosqueteiras...”voltar à terrinha” não seria o mesmo sem vocês enunca me cansarei de agradecer a vossa amizade!

Aos meus pais, Tina e Quim, e ao Richard, por

tudo o que são para mim, dedico a presente Tese.

i

General index

Abstract iiiResumen ivAbbreviations vAmino acids viiAmino acid protecting groups ixResins, handles and coupling reagents x

0. Introduction 3Surface plasmon resonance biosensors 5

0.1 Surface plasmon resonance 70.2 Real-time biospecific interaction analysis 110.3 Measuring kinetics of biospecific interactions 15References 22

Foot-and-mouth disease virus 250.4 Foot-and-mouth disease 270.5 Foot-and-mouth disease virus 290.6 The development of anti-FMDV vaccines 33References 37

Objectives 45

1. SPR screening of synthetic peptides from the GH loop of FMDV 491.0 Introduction 511.1 Optimisation of the experimental set-up 521.2 Application to the systematic screening of FMDV peptides 561.3 Use of other site A – directed monoclonal antibodies 601.4 Probing subtle differences in peptide and mAb behaviour by SPR 641.5 Validity of the experimental kinetic constants 641.6 Relevance of the SPR data for FMDV studies 67References 68

2. Antigenic determinants in the GH loop of FMDV C1-Barcelona (or C-S30) 692.0 Introduction 712.1 Peptides mimicking the GH loop of FMDV C1-Barcelona and the corresponding partial mutants 71

2.2 SPR study of the C-S30 peptides 752.3 Competition ELISA analysis of the C-S30 pentadecapeptides 832.4 Size effects in the antigenicity of C-S30 peptides 852.5 Input from parallel X-ray diffraction studies 892.6 Effect of conformation in the antigenicity of C-S30 peptides 902.7 Antigenic evaluation of C-S30 peptides through solution affinity SPR analysis 952.8 Two-dimensional proton nuclear magnetic resonance studies of C-S30 peptides 1012.9 Recapitulation 107References 109

3. Antigenic peptides with non-natural replacements within the GH loop of FMDV 1113.0 Introduction 1133.1 Peptides that combine antigenicity-enhancing replacements in the GH loop 1133.2 Direct kinetic SPR analysis 1163.3 Indirect SPR kinetic analysis using a high molecular weight competitor antigen 1193.4 Solution affinity SPR analysis of the peptide antigens 1273.5 Two-dimensional 1H-NMR analysis of peptide A15(FPS) 1313.6 X-ray diffraction crystallography analysis of a peptide-antibody complex 1343.7 Recapitulation 144References 145

ii

Conclusions 147

4. Materials & Methods 1514.1 General procedures 153

4.1.1 Solvents and chemicals 1534.1.2 Instrumentation 1564.1.3 Analytical methods 1574.1.4 Chromatographic methods 157References 158

4.2 Solid-phase peptide synthesis 1594.2.1 Solid-phase peptide synthesis protocols 1594.2.2 Synthesis of peptides from the GH loop of FMDV 162References 164

4.3 Antigenic evaluation of the FMDV peptides 1654.3.1 SPR analysis of peptide-antibody interactions 1654.3.2 Enzyme-linked immunosorbent assays 174References 175

4.4 Structural studies of the FMDV peptides 1764.4.1 Two-dimensional proton nuclear magnetic resonance 1764.4.2 Protein X-ray diffraction crystallography 177References 178

iii

Abstract

A fast and direct surface plasmon resonance (SPR) method for the kinetic analysis of the interactions between

peptide antigens and immobilised monoclonal antibodies (mAb) has been established. Protocols have been

developed to overcome the problems posed by the small size of the analytes (< 1600 Da). The interactions

were well described by a simple 1:1 bimolecular interaction and the rate constants were self-consistent and

reproducible. The key features for the accuracy of the kinetic constants measured were high buffer flow rates,

medium antibody surface densities and high peptide concentrations. The method was applied to an extensive

analysis of over 40 peptide analogues towards two distinct anti-FMDV antibodies, providing data in total

agreement with previous competition ELISA experiments.

Eleven linear 15-residue synthetic peptides, reproducing all possible combinations of the four replacements

found in foot-and-mouth disease virus (FMDV) field isolate C-S30, were evaluated. The direct kinetic SPR

analysis of the interactions between these peptides and three anti-site A mAbs suggested additivity in all

combinations of the four relevant mutations, which was confirmed by parallel ELISA analysis. The four-point

mutant peptide (A15S30) reproducing site A from the C-S30 strain was the least antigenic of the set, in

disagreement with previously reported studies with the virus isolate. Increasing peptide size from 15 to 21

residues did not significantly improve antigenicity. Overnight incubation of A15S30 with mAb 4C4 in solution

showed a marked increase in peptide antigenicity not observed for other peptide analogues, suggesting that

conformational rearrangement could lead to a stable peptide-antibody complex. In fact, peptide cyclization

clearly improved antigenicity, confirming an antigenic reversion in a multiply substituted peptide. Solution NMR

studies of both linear and cyclic versions of the antigenic loop of FMDV C-S30 showed that structural features

previously correlated with antigenicity were more pronounced in the cyclic peptide.

Twenty-six synthetic peptides, corresponding to all possible combinations of five single-point antigenicity-

enhancing replacements in the GH loop of FMDV C-S8c1, were also studied. SPR kinetic screening of these

peptides was not possible due to problems mainly related to the high mAb affinities displayed by these synthetic

antigens. Solution affinity SPR analysis was employed and affinities displayed were generally comparable to or

even higher than those corresponding to the C-S8c1 reference peptide A15. The NMR characterisation of one

of these multiple mutants in solution showed that it had a conformational behaviour quite similar to that of the

native sequence A15 and the X-ray diffraction crystallographic analysis of the peptide – mAb 4C4 complex

showed paratope – epitope interactions identical to all FMDV peptide – mAb complexes studied so far. Key

residues for these interactions are those directly involved in epitope – paratope contacts (141Arg, 143Asp, 146His)

as well as residues able to stabilise a particular peptide global folding. A quasi-cyclic conformation is held up by

a hydrophobic cavity defined by residues 138, 144 and 147 and by other key intrapeptide hydrogen bonds,

delineating an open turn at positions 141, 142 and 143 (corresponding to the Arg-Gly-Asp motif).

iv

Resumen

Se diseñó un método rápido y sencillo para el análisis cinético por resonancia de plasmón superficial (RPS) de

las interacciones entre antígenos peptídicos de bajo peso molecular (< 1600 Da) y anticuerpos monoclonales

(AM) inmovilizados en la superficie de un chip sensor. Dichas interacciones se ajustaron a un modelo de

interacción bimolecular 1:1 y las constantes cinéticas obtenidas resultaron fiables y reproducibles. Los

parámetros clave para la calidad de las constantes cinéticas medidas fueron un flujo de tampón elevado, una

densidad superficial de AM intermedia y una elevada concentración de péptido. El método se extendió a más

de 40 análogos peptídicos frente a dos AM contra el virus de la fiebre aftosa (VFA), obteniéndose total

correlación con datos anteriores de ELISA competitivo.

Se sintetizaron once pentadecapéptidos con todas las combinaciones posibles de las cuatro mutaciones que

caracterizan el bucle GH del aislado C-S30 del VFA respecto a la secuencia de referencia C-S8c1. Los

resultados del análisis cinético directo, por RPS, de la antigenicidad de estos péptidos frente a tres AM

sugirieron que dichas combinaciones eran aditivas, observación que fué confirmada por ELISA competitivo.

Así, el tetramutante (A15S30) que mimetiza el bucle GH de C-S30 resultó ser el peor antígeno de la serie, en

contraste con resultados anteriores con este aislado. Aumentando el tamaño del tetramutante de 15 a 21

aminoácidos no afectó significativamente su antigenicidad. En cambio, una incubación prolongada con el AM

llevó a un aumento de reactividad no observado para otros análogos. Posiblemente una reordenación

conformacional del péptido pudo conllevar a la formación de un complejo estable con el anticuerpo.

Experimentos de RPS con un análogo cíclico del péptido A15S30 confirmaron una reversión en la

antigenicidad del tetramutante inducible a través de restricciones conformacionales. Estudios de ambos

péptidos, lineal y cíclico, por resonancia magnética nuclear (RMN) mostraron que características estructurales

anteriormente correlacionadas con la antigenicidad eran más pronunciadas en el análogo cíclico.

Se prepararon veintiseis péptidos con todas las posibles combinaciones de cinco sustituciones específicas en el

bucle GH del VFA C-S8c1. Dichas sustituciones individuales habían sido objeto de estudios anteriores,

obteniéndose una elevada antigenicidad para los correspondientes péptidos mutantes frente a AM anti-VFA.

No se pudo sistematizar el análisis cinético por RPS de los nuevos mutantes multiples, debido a problemas

tanto en la determinación de las constantes cinéticas de disociación, como en la regeneración de las superficies

de AM. Se utilizó así la RPS para la determinación de la afinidad péptido – AM en solución, obteniéndose

antigenicidades comparables o incluso superiores a las del péptido nativo A15 (VFA C-S8c1). Se estudió uno

de los mutantes multiples (A15FPS) por RMN, observándose una conformación identica a la del péptido

nativo. El estudio del complejo cristalino entre el péptido A15FPS y el AM 4C4 por difracción de RX mostró

que las interacciones parátopo – epítopo eran similares a las observadas con el péptido nativo. Se concluyió

que los residuos clave para el reconocimiento son tanto aquellos involucrados en contactos directos (141Arg,143Asp, 146His) como aquellos que estabilizan el plegamiento adecuado del péptido. Así, una conformación casi

cíclica es soportada por una cavidad hidrofóbica definida por los residuos 138, 144 y 147 y por puentes de

hidrógeno intra-peptídicos clave, diseñándose un bucle abierto centrado en las posiciones 141, 142 and 143

(triplete Arg-Gly-Asp).

v

Abbreviations

AA Amino acid

AAA Amino acid analysis

AcOH Acetic acid

AM 2-[4-aminomethyl-(2,4-dimethoxyphenyl)phenoxy]acetic acid

APS Ammonium persulphate

ATR Attenuated total reflection

BSA Bovine serum albumin

CDR Complementarity determining region

Da Dalton

DCM Dichloromethane

DIEA Diisopropylethylamine

DIP Diisopropylcarbodiimide

DMF dimethylformamide

EDC N-ethyl-N’-(dimethylaminopropyl)carbodiimide

EDTA Ethylenediaminotetraacetic acid

ELISA Enzyme-linked immunosorbent assay

eq equivalent

ESI Electro-spray ionisation

Fab Fragment, antigen-binding

Fc Fragment, crystallisable

FMD Foot-and-mouth disease

FMDV Foot-and-mouth disease virus

FT-IR Fourier-transform infrared spectroscopy

HBcAg Hepatitis B core antigen

HCA Human carbonic anhydrase

HEPES 4-(2-hydroxyethyl)piperazine-1-ethanesulphonic acid

HOBt 1-Hydroxybenzotriazole

HPLC High performance liquid chromatography

HRV Human rhino virus

HS Heparan sulphate

IC50 Antigen concentration giving 50% inhibition

IFC Integrated fluidic cartridge

Ig Immunoglobulin

ka Association rate constant / M-1s-1

KA Affinity constant (association) / M-1

kd Dissociation rate constant / s-1

KD Affinity constant (dissociation) / M

KLH Keyhole limpet hemocyanin

ks Apparent/global rate constant / M-1s-1

LED Light-emitting diode

vi

mAb Monoclonal antibody

MALDI-TOF Matrix-assisted laser desorption ionisation – time-of-flight

MAP Multiple antigenic peptide

MBHA p-methylbenzhydrylamine resin

MBS m-maleimidobenzoyl-N-hydroxysuccinimide

MeCN acetonitrile

MeOH methanol

MPLC Medium-pressure liquid chromatography

MS Mass spectrometry

MW Molecular weight

NHS N-hydroxysuccinimide

NMM N-methylmorpholine

NMP N-methylpyrrolidone

NMR Nuclear magnetic resonance

NOE Nuclear Overhauser effect

NOESY Nuclear Overhauser effect spectroscopy

OD Optical density

PBS Phosphate buffer saline

PEG Polyethylene glycol

PS Polystyrene

PVC Polyvinyl chloride

R Response

Req Response at equilibrium

RI Refractive index

Rmax Maximal response

RNA Ribonucleic acid

Rtot Total response

RU Resonance unit

SD Standard deviation

SDS Sodium dodecylsulphate

SDS-PAGE Sodium dodecylsulphate – polyacrylamide gel electrophoresis

SPPS Solid-phase peptide synthesis

SPR Surface plasmon resonance

SPW Surface plasmon wave

VP Viral protein

TBTU N-[(1H-benzotriazol-1-yl)dimethylaminomethylene]-N-methylmethaneaminium N-oxidetetrafluoroborate

TEMED N,N,N’,N’-tetramethylethylenediamine

TFA Trifluoroacetic acid

TFE 2,2,2-trifluoroethanol

TIR Total internal reflection

TOCSY Total correlation spectroscopy

UV - Vis Ultraviolet – visible spectroscopy

vii

NH CH CO

CH3

NH CH CO

H2C (CH2)2 NH C NH2

NH

NH CH CO

H2C CONH2

NH CH CO

H2C COOH

NH CH CO

H2C SH

NH CH CO

H2C CH2 CONH2

NH CH CO

H2C CH2 COOH

NH COCH2

NH CH CO

CH(CH3)CH2CH3

NH CH CO

CH2CH(CH3)2

NH CH CO

CH2(CH2)3NH2

NH CH CO

CH2CH2SCH3

HN CH CO

H2C NH

N

HN CH CO

H2C

Amino acids

Three-

letter

code

One-letter

code

Name Formula

Ala A Alanine

Arg R Arginine

Asn N Asparagine

Asp D Aspartic acid

Cys C Cysteine

Gln Q Glutamine

Glu E Glutamic acid

Gly G Glycine

His H Histidine

Ile I Isoleucine

Leu L Leucine

Lys K Lysine

Met M Methionine

Phe F Phenylalanine

viii

NH CH CO

CH(OH)CH3

NH CO(CH2)5

HN CH CO

H2C NH

HN CH CO

H2C

CH3

NH CH CO

CH(CH3)2

NH CH CO

H2C OH

N CH CO

Three-

letter

code

One-letter

code

Name Formula

Pro P Proline

Ser S Serine

Thr T Threonine

Trp W Triptophan

Tyr Y Tyrosine

Val V Valine

Ahx * 6-aminohexanoicacid

Table I Abbreviations used for amino acid residues according to theBiochemistry Nomenclature Committee of the IUPAC-IUB [specified in Eur. J.Biochem. 138, 9-37 (1984) and J. Biol. Chem. 264, 633-673 (1989)]. α carbonside chains are presented in the non-ionic form for the twenty coded amino acids;All amino-acid residues employed corresponded to the natural L-configuration.

* Ahx is a non-coded amino acid residue used in this work.

ix

H3C C

CH3

CH3

O

O

O

O

S

NNO2

H3C C

CH3

CH3

C

CH3H3C

S

CH3H3C

H3C

OO

Amino acid protecting groups

Abbreviation Name Stability Formula

Boc t-butyloxycarbonyl Stable to bases,labile to TFA

Fmoc 9-fluorenylmethyloxycarbonyl Stable to acids andlabile to bases

Npys 3-nitro-2-pyridylsulphenylStable to acids and

bases, labile tonucleophiles

Pmc2,2,5,7,8-

pentamethylchromane-6-sulphonyl

Stable to bases,labile to TFA

tBu t-butyl Stable to bases,labile to TFA

Trt Triphenylmethyl (trityl) Stable to bases,labile to 1% TFA

Table II Amino acid protecting groups employed in this work.

x

NH2

OOH

O

OCH3

H3CO

H3CCH

H3CN C

CH3CH

CH3

N

NN

N

NN

CH3

CH3

CH3

H3C

+BF4-

O-+

H2N

Polystyrene

H2NHN

NH

ONH

HN

O

O O

O

NH

O

Polystyrene

( )n

Resins, handles and coupling reagents

Abbreviation Structure

AM

MBHA

PEG-PS

DIP

TBTU

Table III Resins, handles and coupling reagents used in this work.

“En el campo hubo de todo: sequía, caracol, fiebre aftosa.”Isabel Allende

in La casa de los espíritus

0. Introduction

Surface plasmon resonance biosensors

SPR as a tool in the functional analysis of an immunodominant site in FMDV

6


7

0.1 Surface Plasmon Resonance

0.1.1 The physical phenomenon1-6

When a beam of light propagating through a first medium of higher refractive index, n1 (e.g. a glass

or quartz prism) meets an interface with a second medium of lower refractive index, n2 (e.g. an

aqueous solution), then it will be totally internally reflected for all incident angles greater than a

critical angle θc:

θc=sin-1(n2/n1) (0.1)

where θ is the angle between the incident

beam and the axis normal to the plane of

the interface. This phenomenon is known

as total internal reflection (TIR)4. Despite

being totally reflected, the incident beam

establishes an electromagnetic field that

penetrates a small distance into the

second medium, where it propagates

parallel to the plane of the interface (Fig.

0.1). This electromagnetic field is called the evanescent wave. The intensity of the evanescent

electric field, I(z), decays exponentially with perpendicular distance z from the interface:

I(z)=Ioe-z/d (0.2)

where d is the penetration depth for angles of incidence θ<θc and light of wavelength λo:

d=(λo/4π)(n12sin2θ-n2

2) (0.3)

d is independent of incident light polarisation but depends on its wavelength. The evanescent

electric field intensity at z=0, Io, depends both on θ and the incident beam polarisation. When the

beam is polarised parallel to the plane of the interface, Io is given by Io//:

Io//=I//[4cos2θ(2sin2θ-n2)]/(n4cosθ+sin2θ-n2) (0.4)

When the beam is polarised perpendicular to the plane of the interface, the field intensity at z=0 is

equal to:

Figure 0. 1 Schematic view of the total internal reflection(TIR) phenomenon3.


Io⊥ =I⊥ (4cos2θ)/(1-n2) (0.5)

I// and I⊥ are the intensities of the incident light beam polarised parallel or perpendicular to the

interface, respectively, and n=(n2/n1)<1. Therefore, two major characteristics of the evanescent

wave are worth to notice:

i) the depth of the evanescent wave is typically less than a wavelength, thus extending a few

hundred nm into the dielectric (liquid phase of refractive index n2);

ii) the evanescent field intensity, Io, for angles a few degrees above the critical angle θc is

several times the incident intensity, I.

When a thin metal film is inserted at the prism/dielectric interface, a new phenomenon called surface

plasmon resonance (SPR) can occur5. Surface plasmons are waves of oscillating surface charge

density (conducting electrons) propagating along the metal surface, at a metal (e.g. silver or

gold)/dielectric (e.g. aqueous solution) interface. The field amplitude of the surface plasmon is

maximal at the interface and decays evanescently, i.e., perpendicularly to it, with a penetration into

the dielectric of about 100-200 nm. Due to high loss in the metal, the surface plasmon wave

propagates with high attenuation in the visible and the near-infrared spectral regions. The

distribution of the electromagnetic field of a surface plasmon wave is highly asymmetric and the

majority of the field is concentrated in the dielectric (Table 0.1).

Table 0.1 Major characteristics of surface plasmon waves (SPW) at the metal-water interface6.

Metal layer supporting SPW Silver GoldWavelength (nm) 630 850 630 850Propagation length(µm) 19 57 3 24Penetration depth into metal (nm) 24 23 29 25Penetration depth into dielectric (nm) 219 443 162 400Concentration of field in dielectric (%) 90 95 85 94

Surface plasmons cannot be directly excited

(resonate) by light, since the frequency and

wave vector requirements cannot be

simultaneously matched. Nevertheless, indirect

excitation can be achieved by an evanescent

wave created by the internal reflection of a p-

polarised incident beam at the metal-coated

surface of the prism.
Figure 0. 2 Schematic view of the surface plasmonresonance (SPR) phenomenon3.
8

This excitation, or resonance, occurs only at a well-defined angle of incidence, θsp, given by:

θsp =sin-1(ksp/ngko) (0.6)


where ng is the refractive index of the prism, λ is the wavelength of the incident light in the vacuum,

ksp is the wave vector of the surface plasmon and ko is the wave vector of the light in the vacuum

ko=2π/λ. When the resonance condition is fulfilled, energy from the incident light is transferred to

the non-radiative surface plasmon and converted to heat. This energy loss is recognised by a sharp

minimum (≤1o at half-width) in the angle-dependent reflectance (Fig. 0.2).

0.1.2 SPR optical sensors and their applications1-31

The phase matching, i.e., the resonance angle θsp, is very sensitive to changes in wavelength, metal

thickness and refractive indices of the prism (n1) and of the dielectric (n2). However, if the first three

factors are all kept constant, θsp will depend only on the refractive index (i.e., on the dielectric

constant) of the dielectric (n2). A change in the dielectric refractive index very close to the metal

surface will originate a change in the resonance angle θsp, which is the principle of SPR sensing.

Considering the unique characteristics of SPR detection, it is possible to design SPR sensors that,

under the proper geometries, allow the sensing of physical, chemical or biochemical phenomena

which give rise to changes in the optical properties of the solution (dielectric) very close to the metal

surface.

Generally, an SPR optical sensor is composed of an optical system, a transducing medium inter-

relating the optical and the reagent domains, and an electronic system controlling the optoelectronic

devices and allowing data processing. The transducing unit transforms changes in the refractive

index, determined by continuously monitoring the SPR angle, into changes in the quantity of

interest. Measurement of the angular dependence of reflectance from an SPR sensor surface requires

a monochromatic light source, such as a

small gas laser (HeNe at 633 nm) or a solid

state diode in the far red. The optical

pathway can include elements for

polarisation in the incident plane (p-

polarisation), attenuation, spatial filtering

and shaping the beam to a convergent

wedge focused at the SPR surface of a

hemicylindrical glass prism (Fig. 0.3).

SPR sensing has a wide variety of applica

displacement and angular position using SPR s

phenomena occurring in optical transducing

sensors, such as humidity detectors using hum

layers and polymers9. The thermooptic effect in

to create an SPR-based temperature sensor10.

Figure 0. 3 Optical apparatus for the measurement ofthe angular dependence of resonance1.

9

tions. Measurement of physical properties such as

ensors has been described7,8. Exploitation of physical

materials allowed the development of specific SPR

idity-induced refractive index changes in porous thin

hydrogenated amorphous silicon has also been used


Chemical SPR sensors are based on the measurement of SPR variations due to adsorption or

chemical reaction of an analyte within a transducing medium which causes changes in its optical

properties. Examples of chemical applications of SPR due to analyte adsorption include: monitoring

of hydrocarbons, aldehydes and alcohols adsorbing on polyethyleneglycol films11; monitoring of

chlorinated hydrocarbons adsorbing on polyfluoroalkylsiloxane12; detection of aromatic

hydrocarbons adsorbing on Teflon film13. SPR devices using palladium are effective in the detection

of molecular hydrogen14; also, chemisorption of NO2 on gold has been used for NO2 detection15.

Copper and nickel phtalocyanine films have been used for SPR detection of toluene16, while bromo-

cresol purple films have been employed for the detection of NH3 vapours17. SPR detection of Cu

and Pb ions was also made possible by combination with anodic stripping voltammetry18.

Affinity SPR biosensors are the most

widely employed, where SPR, as a

surface-oriented method, allows real-time

analysis of biospecific interactions

without the use of labelled biomolecules.

The SPR biosensor technology has been

commercialised and has become a

central tool for characterising and
Figure 0. 4 SPR detection caused by biospecific binding ofligand in solution to an immobilised receptor22.
10

quantifying biomolecular interactions.

Since the first demonstration, in 1983, of the viability of SPR biosensing19, SPR detection of

biospecific interactions was developed until the appearance, in 1994, of the first analysis methods

for surveying biomolecular interactions in real-time3. These methods have been improved for the

study of kinetic and thermodynamic constants of those interactions. Generally, SPR biosensing relies

on the immobilisation of the biological receptor at the chemically modified gold surface, which is in

contact with a buffer solution20. Upon addition of a specific ligand to the solution, binding occurs

close to the gold surface, allowing for SPR detection due to mass increase, and consequent change

in the refractive index in this region (Fig. 0.4)21. The shift in the resonance angle acts as a mass

detector and the continuous angular interrogation in the SPR-sensing device allows for the real-time

monitoring of binding, providing kinetic data on the biospecific interaction. Prism-based SPR

biosensors using angular interrogation have been employed in studies of antigen-antibody23-26,

protein-protein27,28, protein-DNA interactions29and epitope mapping30,31. Many other biomolecular

studies are presently among the applications of SPR biosensing, which has become a part of modern

analytical methods.


0.2 Real-Time Biospecific Interaction Analysis

The use of optical biosensors for interaction analysis has made it possible to obtain affinity and

kinetic data for a large number of protein-protein, protein-peptide and protein-DNA systems.

Biosensor AB (Uppsala, Sweden) is undoubtedly the biosensor market leader, since it launched, in

late 1990, the first commercial SPR-based instrument, BIAcore32-36.

0.2.1 The BIAcore technology3,20-22,37,38

BIAcore uses SPR to investigate biospecific

interactions at the surface of a sensor chip. One

of the components in the interaction is

immobilised on the sensor chip surface and the

other flows over the surface free in solution. As

the interaction proceeds, the concentration

(mass) of analyte in the surface layer changes,

giving an SPR response which can be followed

in real-time in the form of a sensorgram (Fig.

0.5). The instrument consists of a processing

unit, reagents for ligand immobilisation,

exchangeable sensor chips and a personal

computer for control and evaluation. The

processing unit contains the SPR detector and

an integrated microfluidic cartridge that,

together with an autosampler, controls the

delivery of sample plugs into a transport buffer

which flows continuously over the sensor chip

surface (Fig. 0.6). All the injection and

detection systems are thermostatically

controlled so that BIAcore measurements are carr

in BIAcore is a near-infrared light-emitting diode

wedge-shaped beam giving a fixed range of incid

fixed array of light-sensitive diodes covering the

and samples are delivered to the sensor chip surfa

three main parts: the pumps, the sample injector

the pumps is used to maintain the continuous fl

injection of samples and reagents via the auto

programmed to take defined volumes of liquid

sample positions or to the IFC injection port.

Figure 0. 5 Sensorgram: monitoring the SPRresponse in terms of binding to receptor22.

Figure 0. 6 Scheme of the BIAcore instrument21.

11

ied out at constant temperature. The light source

(LED) and light is focused on the gold film as a

ent angles. The SPR response is monitored by a

whole wedge of reflected light. Reagents, buffers

ce through a liquid handling system composed by

and the integrated fluidic cartridge (IFC). One of

ow over the surface, while the other is used for

sampler (sample injector). This autosampler is

from specified sample positions to either other


0.2.2 Ligand immobilisation3,20-22,37,39,40

Sensor chip architecture

The sensor chip is a glass slide onto which a 50-nm thick gold

film has been deposited. Immobilisation by physical adsorption

on gold has disadvantages, namely ligand denaturation, non-

specific binding and steric hindrance. Therefore, the gold surface

has been chemically modified to allow ligand covalent

immobilisation39 in order to obtain a stable ligand surface, with

the possibility of repeated analyses and maximum exposure of

the ligand to the solution containing the biospecific partner (Fig.

0.7).

Immobilisation chemistry

Proteins are, by far, the most widely employed ligands in biospecific interaction analysis. Therefore,

the development of chemistries for ligand immobilisation was based on protein chemistry, namely

on the reaction between protein primary amino groups and the carboxyl groups from the

carboxymethyldextran matrix to form amide bonds. Immobilisation starts by activation of the matrix

COOH groups as N-hydroxysuccinimide active esters, upon reaction with N-hydroxysuccinimide

(NHS) in the presence of N-ethyl-N’-(dimethylaminopropyl)carbodiimide (EDC), in water. Next, a

protein solution at low ionic strength and pH below the isoelectric point is passed over the surface

and protein-matrix amide bonds are formed (Fig. 0.8). The efficiency of the immobilisation step

relies simultaneously on two factors:

i) Electrostatic pre-concentration of positively charged protein in the negatively charged

carboxymethyldextran matrix, and

ii) Reaction between the protein primary amines and the matrix active esters.

1 2 34

Figure 0. 7 Sensor chip CM521,22.

Figure 0. 8 Steps in the standard ligand immobilisation on CM5 sensor chips: 1. COOH activation withEDC/NHS; 2. Ligand coupling; 3. Blocking of remaining reactive NHS-ester groups with ethanolamine; 4.Final ligand surface20.

12


Remaining active esters after protein immobilisation are

converted into inactive amides via reaction with

ethanolamine. The SPR detector continuously monitors

the immobilisation steps (Fig. 0.8) and the amount of

immobilised protein can be controlled either by protein

concentration, reaction time or other factors such as ionic

strength or pH20. NHS-esters are also reactive with other

nucleophilic groups from the ligand, such as thiol or

aldehyde groups (Fig. 0.9). Other chemical modifications

based on NHS-active esters were proposed40:

i) Formation of an amine derivative by reaction of the

ii) Similar preparation of a hydrazide derivative upo

active esters;

iii) Obtention of a maleimide derivative

hydroxysuccinimide ester (sulfo-MBS) to the amine

Tailor-made sensor chips

The high versatility of the carboxymethyldextran matrix in

very high binding capacity and low non-specific binding

analyses, particularly those involving kinetic studies o

concentration analysis. However, the size and charge dens

specific studies, such as those involving high-molecular weig

Presently, a set of sensor chips in which the dextran m

experimental studies is available, ranging from chips with

polymers to chips with reduced charge density (B1) in th

chips are also available, such as a plain gold surface (J1) su

a streptavidin (SA) surface to capture biotinylated ligands,

capture via nickel chelation and a flat hydrophobic (HPA) su

0.2.3 General methodology3,21-31,37,41-54

Binding strategies

Methods for real-time biospecific interaction analysis inclu

sensor chip surface and direct or indirect measurement of a

one component to the immobilised ligand is measured,

binding of two or more components is monitored. When th

modified sensor surface is monitored, the method is direct

with increasing amount of analyte. Indirect methods rely up

either:

Figure 0. 9 Ligand immobilisationbased on NHS-ester activation22.

13

NHS-esters with ethylenediamine;

n reaction of hydrazine with the NHS-

adding sulfo-m-maleimidobenzoyl-N-

surface prepared according to i).

sensor chip CM5 is accompanied by a

suitable for the majority of biospecific

n low-molecular weight analytes or

ity of the matrix can be detrimental for

ht molecules or complex culture media.

atrix has been tailored to suit various

absent (C1) or shortened (F1) dextran

e dextran matrix. Other specific sensor

itable to create new surface chemistries,

a nitrilotriacetic acid (NTA) surface with

rface for membrane biochemistry.

de single- or multi-step binding to the

nalyte. In single-step methods binding of

while in multi-step methods sequential

e interaction of the analyte itself with the

and in such case the response increases

on measurement of a component which


i) interacts with analyte in solution and the remaining free concentration in solution is

measured (solution affinity), or

ii) competes with the analyte of interest for the same ligand binding site (surface competition).

In indirect methods, the response is inversely related to the amount of analyte.

Direct single-step methods (Fig. 0.10) are the simplest

way to study biospecific interactions and are commonly

used for kinetic studies and for concentration

measurement of macromolecules at relatively high

quantities (medium to large analytes above ca. 1

µg/ml)23,27,28,41-45.
Figure 0. 10 Direct single-step detection ofanalyte on the sensor surface22.
14

Direct multi-step methods are assays in which each

stage in the series of binding steps is recorded in the

sensorgram. A common use of these methods consists

in the immobilisation of a capturing molecule (e.g.

streptavidin, anti-immunoglobulin) that specifically

binds the ligand (biotinylated molecule,

immunoglobulin), which is the receptor of the target

analyte (Fig. 0.11)29,46-48. This affinity capture allows

the use of non-pure samples of ligand (e.g. from cell culture media) and also the oriented non-

covalent immobilisation of ligand. These methods are often employed in binding site analysis such

as epitope mapping30,31,49. Another application of multi-step methods is the use of a secondary

molecule to enhance analyte response in sandwich assays where an analyte binds an immobilised

ligand and a second macromolecule is then injected to bind the bound analyte.

Indirect methods are most widely employed for small

analytes (molecular weight<1000 Da) in solution. Direct

detection of such analytes is often difficult and these

usually lack multiple independent binding sites necessary

for response enhancement with sandwich techniques. In

solution affinity experiments the analyte and a specific

receptor interact in solution and, once equilibrium is

reached, the remaining free receptor is determined by

SPR using a sensor chip where another ligand (e.g. the

analyte itself) is immobilised (Fig. 0.12)24,25,50,51. Thus,

the solution affinity between analyte and receptor can be

determined.

Figure 0. 11 Direct multi-step detection ofanalyte on the sensor surface22.

Figure 0. 12 Solution affinity studies –the target analyte is pre-equilibrated withits biospecific receptor in solution andremaining free receptor is measured on thesurface.

Sensor surface

Target analyte

Analyte receptor(measured molecule)


15

In surface competition assays a high molecular

weight analyte is usually employed to compete with

the low molecular weight target analyte for the

same ligand binding site. Since response due to

small analyte binding is unappreciable, only the

response from the large analyte is monitored.

Therefore, the effects on the kinetics of

macromolecule binding due to additions of small

competing analyte can be measured and the

kinetics of small analyte binding can be indirectly

determined (Fig. 0.13)52.

Surface regeneration

Regeneration of the ligand surface allows for the re-utilisation of the same biospecific surface for

series of measurements, obviating the need to replicate identical surfaces. The most general

regeneration methods rely on pH reduction below 2.5 using strong inorganic acids such as 10-100

mM HCl or H3PO4 or weaker acids such as glycine buffers. Nevertheless, ligand tolerance to acids is

variable and, on the other hand, many ligand-analyte complexes may not be disrupted under acidic

conditions. Therefore, regeneration procedures must be optimised and many regenerating agents

other than acids (bases such as 10 mM NaOH, high ionic strength solutions such as 1M NaCl, etc.)

may be found to be more effective. A systematic regeneration optimisation protocol has been

recently described and successfully applied to antibody surfaces53,54. Stock solutions are mixtures of

similar components (e.g. all acids, all bases, all salts, all detergents, etc.) and regeneration cocktails

are different combinations of such stock solutions. Fine-tuning of regeneration cocktails may provide

the answer to problems not overcome with standard regeneration methods and allow for the use of

molecules that, otherwise, would not be suitable as easy-to-regenerate ligands.

0.3 Measuring kinetics of biospecific interactions

Characterisation of the affinities and rates of biospecific interactions is fundamental in many areas of

biochemical research. Methods that measure changes in optical parameters, such as fluorescence or

absorbance, can be employed for direct kinetic analysis. However, these methods require that one

of the reactants is often labelled with a radioactive or fluorescent probe and thus no longer in native

form. SPR detection is more general than these methods, since it is sensitive to changes in mass and

no labelling is required. When analyte is injected across a ligand surface, the resulting sensorgram

displays three essential phases, namely, association of analyte with ligand during sample injection,

equilibrium (if reached) during sample injection, where the rate of analyte binding is balanced with

complex dissociation, and dissociation of analyte-ligand complex due to buffer flow immediately

Figure 0. 13 Surface competition between thesmall target analyte and the SPR-detectedmacromolecule.

Sensor surface

Target analyte(small competitor)

Large analyte(measured molecule)


16

after the end of analyte injection. With suitable analysis of binding data, reliable affinity and kinetic

data can be obtained from SPR experiments. However, for the majority of experimental purposes,

semi-quantitative ranking of rates and/or affinities is sufficient.

0.3.1 Basic theory3,31,37,55,56

A 1:1 interaction between the analyte (A) continuously flowing in solution over the ligand (B)

surface may be described by:

Considering that, in the association phase, the sensor surface is continuously replenished with free

analyte solution and the amount of bound analyte is negligible with respect to the total analyte

concentration (C), pseudo-first order kinetics can be assumed. Thus, the rate of complex formation

is given by the equation:

d[AB]/dt=ka[A][B]-kd[AB] (0.7)

which, in terms of SPR response, can be expressed as:

dR/dt=kaC(Rmax-R)-kdR= kaCRmax-(kaC+kd)R (0.8)

where:

! R is the SPR response (in resonance units, RU) at time t;

! Rmax is the maximum analyte binding capacity (in RU), which reflects the number of ligand

binding sites, i.e., total ligand concentration;

! ka is the association rate constant;

! kd is the dissociation rate constant.

Therefore, a plot of dR/dt against R will be a straight line of slope -(kaC+kd) or -ks, where ks is the

apparent binding rate, and a plot of ks against analyte concentration will give a straight line with

slope ka and intercept on the ordinate kd.

When equilibrium is reached, the total binding rate (dR/dt) is zero and, from equation 0.8:

kaC(Rmax-Req)=kdReq (0.9)

Where Req is the total response at equilibrium. Considering that the affinity association constant (KA)

is given by ka/kd, binding affinity can be determined from equilibrium measurements, as it can be

inferred from substituting and rearranging equation 0.9:

A+Bka

kd

AB


17

Req/C=KARmax-KAReq (0.10)

Thus, by plotting Req/C against Req, a straight line is obtained and KA and Rmax can be calculated

from the slope and the intercept on the ordinate, respectively.

During the dissociation phase, analyte solution is replaced by a continuous flow of running buffer

solution and analyte concentration drops to zero. For the pseudo-first order kinetics model, complex

dissociation can be described by:

dR/dt=-kdR (0.11)

which, in the logarithmic form, can be given by:

ln(R0/Rt)=kd(t-t0) (0.12)

where R0 is the response at an arbitrary start dissociation time t0. Consequently, a plot of ln(Rt/R0)

will give a straight line with slope -kd.

This basic theoretical model only applies when the interaction is homogeneous and when the

pseudo-first order kinetics is actually observed.

0.3.2 Fitting and evaluating biosensor data55-58

Curve fitting methods

In early kinetic studies based on SPR biospecific interaction analysis, data evaluation relied upon

linearisation of the binding data, according to the equations described in the previous section.

Nevertheless, linear transformations also transform the parameter-associated errors, which decreases

the quality of primary data.

On the other hand, it requires data from many analyte

concentrations. Therefore, non-linear least squares analysis

has been introduced for fitting and evaluating biosensor

data57. Non-linear least squares methods optimise

parameter values by minimising the sum of the squared

residuals (S), being the latter the difference between the

fitted (rf) and the experimental (rx) curves at each point

(residuals are squared in order to equal the weight of

deviations above and below the experimental curve, Fig.

0.14, Eq. 0.13).

( )2

1∑ −=

n

xf rrS (0.13)

Figure 0. 14 Schematicrepresentation of non-linear leastsquares fitting by minimising squaredresiduals55.


18

Non-linear least squares analysis has been applied to curve fitting based on the integrated rate

equations (Table 0.2, page 21). This analytical integration is the simplest tool for systems with rate

equations that can be readily integrated. However, many interactions studied on biosensors do not

fit simple kinetic models, which can be seen by curved plots when linearisation is applied or by poor

fits when using the integrated rate equations. The software currently employed for biosensor data

evaluation includes several kinetic fitting models (Table 0.2, page 21). Those models corresponding

to binding that can be described by well-known rate equations use analytical integration while more

complex models, such as interactions with mass transfer limitations or conformational changes, rely

upon curve fitting with numerical integration58. Numerical integration is more computationally-

intensive but allows evaluation when the rate equations cannot be integrated analytically. In

numerical integration methods, each species is assigned an initial concentration and the reaction is

stepped through in discrete time intervals. At the end of each interval the concentration of each

species is calculated considering its rate of formation or disappearance according to the rate

equations. Numerical integration can be used to model any kinetic mechanism and also to analyse

biosensor data by curve fitting as is done with analytical integration. However, with numerical

methods data is usually analysed globally by fitting both the association and dissociation phases for

several concentrations simultaneously. Global curve fitting is advantageous, because it minimises the

possibility of having a good fit with a wrong kinetic model and it lowers the variance in the estimates

of the rate constants.

Evaluating fitted data

The fitting algorithms are purely mathematical tools without any biochemical “knowledge”.

Therefore, it is always important to examine the results of fitted data to check for “reasonableness”

of the parameters found. This must be kept in mind at the time of choosing the “best fit”. This best

fit depends on the ability of the fitting algorithm to converge for the true minimum in the sum of

squared residuals and on the number of parameters that can be varied in the model, i.e., the

complexity of the model. Increasing model complexity also increases the probability to fall in local

minima and obtain misleading fits. Wrong fits are usually evident from markedly poor curve fits or

unreasonable results and are often due to bad data quality or inadequate choice of the fitting model.

Increasing the complexity of a model will also increase the ability of fitting the experimental curves

to the equation, since there is a wider range for varying parameters in order to obtain a closer fit.

Therefore, it is important to accept the simplest model that fits the sensorgrams when evaluating

kinetic data and judge whether a slightly better fit with a more complex model is experimentally

significant. The quality of the fit is described by the chi squared statistical parameter, defined as:

χ2=S/(n-p) (0.14)

where n is the number of data points, p is the number of fitted parameters and S is the sum of the

residuals (Eq. 0.13). Since n>>p, χ2 reflects the average squared residual per data point and, when


the model fully fits the experimental data, χ2 represents the mean square of the signal noise. In

practice it is useful to check for the shape of the residual plot, since non-random distribution of

residuals is often a symptom of an incorrect fit.

0.3.3 Deviations from the langmuirian behaviour2,3,37,51,55,58-72

a) Mass-transport limitations2,3,51,55,59-66

Transport of mobile analyte to the sensor surface (Fig.

0.15) may be a serious problem when the interaction is

fast. Insufficient transport rate will not allow to obtain

meaningful kinetics (the rate-limiting step will be the

diffusion into the dextran matrix and not the interaction

itself) and the assumption that analyte bulk

concentration is constant and equal to the injected

concentration is no longer valid. Consequently, the rate

equations corresponding to pseudo-first-order kinetics

are not applicable to systems under diffusion-controlled ki

using high flow rates (> 30 µl/min), low density ligand s

Nevertheless, systems with very high interaction rates w

implies an upper limit to the range of association rate cons

effect related to mass-transport limitations is analyte reb

analyte depletion from the surface is not fast enough, analy

response no longer follows a single exponential decay.

b) Ligand heterogen

Random immobilisa

lead to heterogeinet

longer equivalent ne

is more pronounced

with decreasing num

affinity” ligand sites

dextran layer, lo

immobilisation leve

heterogeneity effects

c) Analyte heterogeneity3,55,59

Although biospecific analysis allows for the utilisation

measurements of bioactive molecules in biological sample

that samples for kinetic analysis do not contain molecules

with the ligand. Otherwise, the SPR response will reflect

cannot be described by simple kinetics.

Figure 0. 15 Scheme of the differentfactors influencing transport of mobileanalyte to the sensor surface withimmobilised ligand2.

netics. Diffusion effects can be minimised

urfaces and high analyte concentrations.

ill be always diffusion-controlled, which

tants amenable to study by SPR. Another

inding during the dissociation phase. If

te molecules will rebind to the ligand and

eity2,3,55,58-61,67,68

tion chemistries and high surface density

y of ligand sites, which therefore are no

ither independent (Fig. 0.16). This effect

with high analyte concentrations, i.e.,

ber of free “readily accessible⇔higher

. Oriented attachment of ligand to the

w analyte concentrations and low

ls are the best measures to avoid

Figure 0. 16 Illustration ofheterogeneous binding of analyte toligand molecules immobilised inexposed and buried sites59.

19

.

of non-purified samples (concentration

s, ligand fishing, etc.), it must be ensured

, other than the analyte, that can interact

the sum of different binding events and


d) Steric hindrance2,59,60

The formation of a complex between a large analyte and

immobilised ligand can mask additional binding sites. Although it

could be argued that such steric hindrance would not affect

binding kinetics (it would only decrease Rmax), this is not strictly

true, since the flexibility/fluidity of the dextran matrix allows

temporarily masked sites to become accessible during analyte

injection, adding complexity to the kinetics of the interaction.

This problem, also named the parking problem, assumes greater

proportions for large macromolecular analytes, higher analyte

Figure 0. 17 Illustration of theparking problem: masking ofligand binding sites byattachment of large analytemolecules59.

20

concentrations and high density ligand surfaces (Fig. 0.17).

e) Analyte multivalency and ligand co-operativity51,55,59,69

Poor fits with pseudo-first-order kinetics should be expected whenever analyte is multivalent (e.g.

antibody bivalency), since 1:1 stoichiometry is no longer observed. Also, it is difficult to ensure that

both analyte binding sites (in the case of bivalency) are equivalent and independent, as well as to

know to which extent has the interaction 1:1 or 1:2 stoichiometry. Another situation where binding

sites may not be independent occurs when there is ligand co-operativity. Although equivalent, the

ligand binding sites may not interact independently from each other and negative or positive co-

operative interactions will prevent the system from following simple pseudo-first-order kinetics.

f) Conformational changes58-60

It has been suggested that non-conformity of sensorgrams with the langmuirian model could be due

to additional steps involving isomerisation of the AB complex. Such two-state-reactions, where there

is conformational change upon binding, are not well described by pseudo-first-order kinetics and

other models must be employed to fit the data.

0.3.4 Experimental design in kinetic SPR analysis2,3,37,51,55,61,68-72

Deviations to pseudo-first-order kinetics predicted for 1:1 interactions could be interpreted as due to

more complex interaction mechanisms describing the interactions. However, they are often

produced by artefacts which can be minimised by careful experimental design. Low ligand

immobilisation levels are advisable to avoid mass-transport limitations, ligand heterogeneity or steric

hindrance. Analyte concentration must be high enough to avoid diffusion-controlled kinetics and low

enough not to saturate the surface (between 0.1KD and 10 KD). Buffer flow must be kept at high rate

to minimise diffusion-controlled binding and, whenever possible, soluble ligand must be added to

buffer in the dissociation phase to avoid rebinding effects. Oriented immobilisation chemistries

should be used when random amine coupling is seen to be a significative source of surface

heterogeneities. Instrumental drifts or non-specific binding to the carboxymethyldextran matrix are

often eliminated by subtraction of a blank run, using either an inactive analyte or a suitable

reference cell with inactive ligand. If problems with non-specific binding are persistent, the choice of

another kind of surface (other model of sensor chip) may be the solution.


21

Table 0.2 Rate equations used in the pre-defined fitting models included in the BIAevaluation 3.0 software55.

Simultaneous ka/kd fitDifferential equations Total response Reaction scheme

(a) 1:1 langmuirian

binding

d[B]/dt=-(ka[A][B]-kd[AB])

d[AB]/dt=ka[A][B]-kd[AB] [AB]+RIA+B⇔AB

[A]=C, [B]0=Rmax,

[AB]0=0

(b) 1:1 binding with

drifting baselinethe same as in (a) [AB]+drift(t-

ton)+RI

the same as in (a)

(c) 1:1 binding with

mass transfer

the same as in (a) plus

d[A]/dt=kt(C-[A])- (ka[A][B]-kd[AB])[AB]+RI

Abulk⇔A+B⇔AB

[A] bulk=C, [B]0=Rmax,

[AB]0=0

(d) Heterogeneous

ligand (2 different

binding sites)

d[B1]/dt=-(ka1[A][B1]-kd1[AB1])

d[AB1]/dt=ka1[A][B1]-kd1[AB1]

d[B2]/dt=-(ka2[A][B2]-kd2[AB2])

d[AB2]/dt=ka2[A][B2]-kd2[AB2]

[AB1]+[AB2]+RI

A+B1⇔AB1

A+B2⇔AB2

[A]=C, [B1]0=Rmax1,

[B2]0=Rmax2

[AB1]0=[AB2]0=0

(e) Heterogeneous

analyte (competition

between two different

analytes)

d[B]/dt=-(ka1[A1]mw1[B]-

kd1[A1B])/mw1n1-(ka2[A2]mw2[B]-

kd2[A2B])/mw2n2

d[A1B]/dt=ka1[A1]mw1[B]-kd1[A1B]

d[A2B]/dt=ka2[A2]mw2[B]-kd2[A2B]

[A1B]+[A2B]+RI

A1+B⇔A1B

A2+B⇔A2B

[A1]=C1, [A2]=C2,

[B]0=Rmax/mw1,

[A1B]0=[A2B]0=0

(f) Bivalent analyte d[B]/dt=-(ka1[A][B]-kd1[AB])-

(ka2[AB][B]-kd2[AB2])

d[AB]/dt=(ka1[A][B]-kd1[AB])-

(ka2[AB][B]-kd2[AB2])]

d[AB2]/dt=ka2[AB][B]-kd2[AB2]

[AB]+[AB2]+RI

A+B⇔AB

AB+B⇔AB2

[A]=C, [B]0=Rmax,

[AB]0=[AB2]0=0

(g) Conformational

change (two-state

reaction)

d[B]/dt=-(ka1[A][B]-kd1[AB])

d[AB]/dt=(ka1[A][B]-kd1[AB])-

(ka2[AB]-kd2[AB*])

d[AB*]/dt=ka2[AB]-kd2[AB*]

[AB]+[AB*]+RIA+B⇔AB⇔AB*

[A]=C, [B]0=Rmax,

[AB]0=[AB*]0=0

Separate ka/kd fitIntegrated rate equations

(h) 1:1 langmuirianbinding

( )( )[ ]RI

kCk

eRCkR

da

tkCk

maxada

++

−=

+−1

R = R0e-kdt + offset

[AB]+RI

[AB]+offset

the same as in (a)


22

References

1 Garland, P. B. (1996) Optical evanescent wave methods for the study of biomolecular interactions,Quart. Rev. Biophys. 2, 91-117.

2 Schuck, P. (1997) Use of surface plasmon resonance to probe the equilibrium and dynamic aspects ofinteractions between biological macromolecules, Annu. Rev. Biophys. Biomol. Struct. 26, 541-566.

3 “BIAcore Instrument Handbook”, (Pharmacia Biosensor AB, 1994) Uppsala, Sweden.4 Hirshcfield, T. (1967), Physics of Thin Films 9, 145.5 Kretschmann, E. and Raether, H. (1968) Radiative decay of non-radiative surface plasmons excited

by light, Z. Natuforsch. 23A, 2135-2136.6 Homola, J., Yee, S. S. and Gauglitz, G. (1999) Surface plasmon resonance sensors: review, Sens. &

Actuat. B 54, 3-15.7 Margheri, G., Mannoni, A. and Quercioli F. (1996) A new high-resolution displacement sensor based

on surface plasmon resonance, Proc. SPIE 2783, 211-220.8 Schaller, J. K., Czepluch, R. and Stojanoff, C. G. (1997) Plasmon spectroscopy for high resolution

angular measurements, Proc. SPIE 3098, 476-486.9 Weiss, M. N., Srivastava, R. and Groger, H. (1996) Experimental investigation of a surface plasmon-

based integrated-optic humidity sensor, Electron. Lett. 32, 842-843.10 Chadwick, B. and Gal, M. (1993) An optical temperature sensor using surface plasmons, Japn. J.

Appl. Phys. 32, 2716-2717.11 Miwa, S. and Arakawa, T. (1996) Selective gas detection by means of surface plasmon resonance

sensors, Thin Solid Films 281-282, 466-468.12 Abdelghani, A., Chovelon, J. M., Jaffrezic-Renault, N., Ronot-Trioli C., Veillas, C. and Gagnaire, H.

(1997) Surface plasmon resonance fiber-optic sensor for gas detection, Sens. & Actuat. B 38-39, 407-410.

13 Podgorsek, R. P., Sterdenburgh, T. Wolters, J. Ehrenreich, T., Nischwitz, S. and Franke, H. (1997)Optical gas sensing by evaluating ATR leaky mode spectra, Sens. & Actuat. B 39, 349-352.

14 Chadwick, B. and Gal, M. (1994) A hydrogen sensor based on the optical generation of surfaceplasmons in a palladium alloy, Sens. & Actuat. B 17, 215-220.

15 Ashwell, G. J. and Roberts, M. P. S. (1996) Highly selective surface plasmon resonance sensor forNO2, Electron. Lett. 32, 2089-2091.

16 Granito, C., Wilde, J. N., Petty, M. C., Houston, S. and Iradale, P.J. (1996) Toluene vapor sensingusing copper and nickel phthalocyanine Langmuir-Blodgett films, Thin Solid Films 284-285, 98-101.

17 Van Gent, J., Lambeck, P. V., Bakker, R. J., Popma, T. J., Sudholter, E.J.R. and Reinhoudt, D.N.(1991) Design and realization of a surface plasmon resonance-based chemo-optical sensor, Sens. &Actuat. A 26, 449-452.

18 Chinowsky, T. M., Saban, S. B. and Yee, S. S. (1996) Experimental data from a trace metal sensorcombining surface plasmon resonance with anodic stripping voltammetry, Sens. & Actuat. B 35, 37-43.

19 Liedberg, B., Nylander, C. and Lundström, K. (1983) Surface plasmons resonance for gas detectionand biosensing, Sens. & Actuat. 4, 299-304.

20 Johnsson, B., Löfås, S. and Lindquist, G. (1991) Immobilization of proteins to acarboxymethyldextran-modified gold surface for biospecific interaction analysis in surface plasmonresonance sensors, Anal. Biochem. 198, 268-277.

21 Fägerstam, L. G., Frostel-Karlsson, Å., Karlsson, R., Persson, B. and Rönnberg, I. (1992) Biospecificinteraction analysis using surface plasmon resonance detection applied to kinetic, binding site andconcentration analysis, J. Chrom. 597, 397-410.

22 http://www.biacore.com/23 Altschuh, D., Dubs, M. C., Weiss, E., Zeder-Lutz, G. and Van Regenmortel, M. H. V. (1992)

Determination of kinetic constants for the interaction between monoclonal antibody and peptidesusing surface plasmon resonance, Biochemistry 31, 6298-6304.

24 Lasonder, E., Bloemhoff, W. and Welling, G. W. (1994) Interaction of lysozyme with synthetic anti-lysozyme D1.3 antibody fragments studied by affinity chromatography and surface plasmonresonance, J. Chrom. A 676, 91-98.

25 Lasonder, E., Schellekens, G. A., Koedijk, D. G. A. M., Damhof, R. A., Welling-Wester S., Feijlbrrief,M., Scheffer, A. J. and Welling, G. W. (1996) Kinetic analysis of synthetic analogues of linear-epitopepeptides of glycoprotein D of herpes simplex virus type I by surface plasmon resonance, Eur. J.Biochem. 240, 209-214.

26 Houshmand, H., Fröman, G. and Magnusson, G. (1999) Use of bacteriophage T7 displayed peptidesfor determination of monoclonal antibody specificity and biosensor analysis of the binding reaction,Anal. Biochem. 268, 363-370.

27 Wu, Z., Johnson, K., Choi, Y. and Ciardelli, T. L. (1995) Ligand binding analysis of soluble interleukin2-receptor complexes by surface plasmon resonance, J. Biol. Chem. 270, 16045-16051.


23

28 Lessard, I. A. D., Fuller, C. and Perham, R. N. (1996) Competitive interaction of component enzymeswith the peripheral subunit-binding domain of the pyruvate-dehydrogenase multienzyme complex ofBaccilus stearothermophilus: kinetic analysis using surface plasmon resonance detection, Biochemistry35, 16863-16870.

29 Cheskis, B. and Freedman, L. P. (1996) Modulation of nuclear receptor interactions by ligands: kineticanalysis using surface plasmon resonance, Biochemistry 35, 3309-3318.

30 Dubs, M. C., Altschuh, D. and Van Regenmortel, M. H. V. (1992) Mapping of viral epitopes withconformationally specific monoclonal antibodies using biosensor technology, J. Chrom. 597, 391-396.

31 Saunal, H. and Van Regenmortel, M. H. V. (1995) Mapping of viral conformational epitopes usingbiosensor measurements, J. Immunol. Meth. 183, 33-41.

32 Hodgson, J. (1994) Light, Angles, Action: Instruments for label-free, real-time monitoring ofintermolecular interactions, Biotechnology 12, 31-35.

33 Malmqvist, M. and Karlsson, R. (1997) Biomolecular interaction analysis: affinity biosensortechnologies for functional analysis of proteins, Curr. Op. Chem. Biol. 1, 378-383.

34 Pathak, S. and Savelkoul, H. F. J. (1997) Biosensors in immunology: the story so far, Immunol.Today 18, 464-467.

35 Fivash, M., Towler, E. M. and Fisher, R. J. (1998) BIAcore for macromolecular interaction, Curr. Op.Biotechnol. 9, 97-101.

36 Lakey, J. H. and Raggett, E. M. (1998) Measuring protein-protein interactions, Curr. Op. Struct. Biol.8, 119-123.

37 “BIAapplications Handbook”, (Pharmacia Biosensor AB, 1994) Uppsala, Sweden.38 Sjölander, S. and Urbaniczky, C. (1991) Integrated fluid handling system for biomolecular interaction

analysis, Anal. Chem. 63, 2338-2345.39 Löfås, S. and Johnsson, B. (1990) A novel hydrogel matrix on gold surfaces in surface plasmon

resonance sensors for fast and efficient covalent immobilization of ligands, J. Chem. Soc., Chem.Commun., 1526-1528.

40 O’Shannessy, D. J., Brigham-Burke, M. and Peck, K. (1992) Immobilization chemistries suitable foruse in the BIAcore surface plasmon resonance detector, Anal. Biochem. 205, 132-136.

41 Brigham-Burke, M., Edwards, J. R. and O’Shannessy, D. J. (1992) Detection of receptor-ligandinteractions using surface plasmon resonance: model studies employing the HIV-1 gp120/CD4interaction, Anal. Biochem. 205, 125-131.

42 Lemmon. M. A., Ladbury, J. E., Mandiyan, V., Zhou, M. and Schlessinger, J. (1994) Independentbinding of peptide ligands to the SH2 and SH3 domains of Grb2, J. Biol. Chem. 269, 31653-31658.

43 Tamamura, H., Otaka, A., Murakami, T., Ishihara, T., Ibuka, T., Waki, M., Matsumoto, A.,Yamamoto, N. and Fujii, N. (1996) Interaction of an anti-HIV peptide, T22, with gp120 and CD4,Biochem. Biophys. Res. Comm. 219, 555-559.

44 Chao, H. Houston, M. E., Grothe, S., Kay, C. M., O’Connor-McCourt, M., Irvin, R. T. and Hodges, R.S. (1996) Kinetic study on the formation of a de novo designed heterodimeric coiled-coil: use ofsurface plasmon resonance to monitor the association and dissociation of polypeptide chains,Biochemistry 35, 12175-12185.

45 England, P., Brégére, F. and Bedouelle, H. (1997) Energetic and kinetic contributions of contactresidues of antibody D1.3 in the interaction with lysozyme, Biochemistry 36, 164-172.

46 Huyer, G., Li, Z. M., Adam, M., Huckle, W. R. and Ramachandran, C. (1995) Direct determination ofthe sequence recognition requirements of the SH2 domains of SH-PTP2, Biochemistry 34, 1040-1049.

47 Shen, B. J., Hage, T. and Sebald, W. (1996) Global and local determinants for the kinetics ofinterleukin-4/interleukin 4 receptor α chain interaction: a biosensor study employing recombinantinterleukin-4 binding protein, Eur. J. Biochem. 240, 252-261.

48 Lookene, A., Chevreuil, O., Østergaard, P. and Olivecrona, G. (1996) Interaction of lipoprotein lipasewith heparin fragments and with heparan sulfate: stoichiometry, stabilization and kinetics,Biochemistry 35, 12155-12163.

49 Van Regenmortel, M. H. V., Altschuh, D., Pellequer, J. L., Richalet-Sécordel, P., Saunal, H., Wiley, J.A. and Zeder-Lutz, G. (1994) Analysis of viral antigens using biosensor technology, Methods: AComp. Meth. Enzymol. 6, 177-197.

50 Zeder-Lutz, G., Rauffer, N., Altschuh, D. and Van Regenmortel, M. H. V. (1995) Analysis ofcyclosporin interactions with antibodies and cyclophilin using BIAcore, J. Immunol. Meth. 183, 131-140.

51 Nieba, L., Krebber, A. and Plükthun, A. (1996) Competition BIAcore for measuring true affinities:large differences from values determined from binding kinetics, Anal. Biochem. 234, 155-165.

52 Karlsson, R. (1994) Real-time competitive kinetic analysis of interactions between low-molecular-weight ligands in solution and surface-immobilized receptors, Anal. Biochem. 221, 142-151.

53 Andersson, K., Hamalainen, M. and Malmqvist, M. (1999) Identification and optimization ofregeneration conditions for affinity-based biosensor assays. A multivariate cocktail approach, Anal.Chem. 71, 2475-2481.


24

54 Andersson, K., Areskoug, D. and Hardenborg, E. (1999) Exploring buffer space for molecularinteractions, J. Molec. Recogn. 12,

55 “BIAevaluation Software Handbook: version 3.0”, (Biosensor AB, 1997) Uppsala, Sweden.56 Atkins P. W., “Physical Chemistry, 5th ed., Oxford University Press, Oxford, U. K. 1994.57 O’Shannessy, D. J., Brigham-Burke, M., Soneson, K. K., Hensley, P. and Brooks, I. (1993)

Determination of rate and equilibrium binding constants for macromolecular interactions using surfaceplasmon resonance: use of nonlinear least squares analysis methods, Anal. Biochem. 212, 457-468.

58 Morton, T. A., Myszka, D. and Chaiken, I. (1995) Interpreting complex binding kinetics from opticalbiosensors: a comparison of analysis by linearization, the integrated rate equation and numericalintegration, Anal. Biochem. 227, 176-185.

59 O’Shannessy, D. J. and Winzor, D. J. (1996) Interpretation of deviations from pseudo-first-orderkinetic behavior in the characterization of ligand binding by biosensor technology, Anal. Biochem.236, 275-283.

60 Bowles, M. R., Hall, D. R., Pond, S. M. and Winzor, D. J. (1997) Studies of protein interactions bybiosensor technology: an alternative approach to the analysis of sensorgrams deviating from pseudo-first-order kinetic behavior, Anal. Biochem. 244, 133-143.

61 Schuck, P. (1997) Reliable determination of binding affinity and kinetics using surface plasmonresonance biosensors, Curr. Op. Biotech. 8, 498-502.

62 Glaser, R. W. (1993) Antigen-antibody binding and mass transport by convection and diffusion to asurface: a two-dimensional computer model of binding and dissociation kinetics, Anal. Biochem. 213,152-161.

63 Myszka, D., Arulanatham, P. R., Sana, T., Wu, Z., Morton, T. A. and Ciardelli, T. L. (1996) Kineticanalysis of ligand binding to interleukin-2 receptor complexes created on an optical biosensor surface,Prot. Sci. 5, 2468-2478.

64 Hall, D. R., Cann, J. R. and Winzor, D. J. (1996) Demonstration of an upper limit to the range ofassociation rate constants amenable to study by biosensor technology based on surface plasmonresonance, Anal. Biochem. 235, 175-184.

65 Myszka, D., Morton, T. A., Doyle, M. L. and Chaiken, I. M. (1997) Kinetic analysis of a proteinantigen-antibody interaction limited by mass transport on an optical biosensor, Biophys. Chem. 64,127-137.

66 Witz, J. (1999) Kinetic analysis of analyte binding by optical biosensors: hydrodynamic penetration ofthe analyte flow into the polymer matrix reduces the influence of mass transport, Anal. Biochem. 270,201-206.

67 Oddie, G. W., Gruen, L. C., Odgers, G. A., King, L. G. and Kortt, A. A. (1997) Identification andminimization of nonideal binding effects in BIAcore analysis: ferritin/anti-ferritin Fab’ interaction as amodel system, Anal. Biochem. 244, 301-311.

68 Kortt, A. A., Oddie, G. W., Iliades, P., Gruen, L. C. and Hudson, P. J. (1997) Nonspecific amineimmobilization of ligand can be a potential source of error in BIAcore binding experiments and mayreduce binding affinities, Anal. Biochem. 253, 103-111.

69 Kalinin, N. L., Ward, L. D. and Winzor, D. J. (1995) Effects of solute multivalence on the evaluationof binding constants by biosensor technology: studies with concanavalin A and interleukin-6 aspartitioning proteins, Anal. Biochem. 228, 238-244.

70 Catimel, B., Nerrie, M., Lee, F.T., Scott, A. M., Ritter, G., Welt, S., Old, L. J., Burgess, A. W. andNice, E. C. (1997) Kinetic analysis of the interaction between the monoclonal antibody A33 and itscolonic epithelial antigen by the use of an optical biosensor: a comparison of immobilisation strategies,J. Chrom. A 776, 15-30.

71 Karlsson, R. and Fält, A. (1997) Experimental design for kinetic analysis of protein-protein interactionswith surface plasmon resonance biosensors, J. Immunol. Meth. 200, 121-133.

72 Ober, R. J. and Ward, E. S. (1999) The choice of the reference cell in the analysis of kinetic data usingBIAcore, Anal. Biochem. 271, 70-80.

Foot-and-mouth disease virus


26


27

0.4 Foot-and-mouth disease

0.4.1 General features1-5

Foot-and-mouth disease (FMD) is an acute systemic infection affecting even-toed ungulates, both

domesticated and wild, including cattle, swine, sheep and goats. FMD generally involves mortality

rates below 5%, but even so it is considered the most important disease of farm animals since it

causes important decreases in livestock productivity and trade. The main route of infection of

ruminants is the inhalation of airborne virus, but infection via the alimentary tract or skin lesions is

also possible, although requiring higher doses of virus. After primary replication in the pharynx, the

virus enters the bloodstream and, following a 3 to 5 days period of febrile viræmia, it spreads

throughout the organs and tissues where new sites for secondary infection are established. Some

clinical symptoms of FMD are fever, anorexia, weight loss, lameness, salivation and vesicular lesions

(mouth and skin). Although FMD only rarely causes death in adult animals, the virus can cause

severe lesions in the myocardium of young animals, leading in this case to high mortality rates1-3.

An asymptomatic persistent infection can be established in ruminants for periods of a few weeks to

several years as a consequence either of the acute infection or of vaccination with live-attenuated

virus. Animals affected by this long-term persistent infection are known as carrier animals and are an

important reservoir of the FMD virus in nature. Also, it has been suggested that carrier cattle are a

possible source of FMD outbreaks by virus transmission to susceptible animals. The impossibility to

cure carrier animals by vaccination, together with the extraordinary genetic and antigenic complexity

of the FMD virus, are major drawbacks for the control of the disease1-5.

0.4.2 Natural distribution of FMD

The earliest reports on FMD were descriptions of outbreaks in Northern Italy in 1514 and in

Southern Africa in 1780, written by Fracastorri6 and Le Vailant7, respectively. Seven

immunologically different serotypes of the FMD virus are known, namely A, O, C, Asia-1, South-

African Territories (SAT) -1, -2 and -3, which comprise more than 65 subtypes. The global

distribution of the disease in 1997 was as represented in Fig. 0.18, with no significant changes over

the last 30 years. FMD is endemic in South America, sub-Saharan Africa, India and Middle/Far

East8. Countries such as Chile, French Guyana, Guyana and Surinam have been FMD free for the

last decade, while the members of the Mercasur (Argentina, Uruguay, Paraguay and Brazil) have

greatly improved the control of the disease through vaccination programmes8. In sub-Saharan

Africa, the control of FMD has been motivated by the exportation of beef to Europe. However, there

is occasional spread of the disease from the African buffalo, which is usually restricted to game

parks. On the other hand, poor surveillance and diagnostic facilities as well as deterioration of some

control programmes are causative of FMD spread to domestic cattle, including North-African

countries like Tunisia, Morocco and Algeria8.


Figure 0. 18 Estimated world distribution of FMD in 1997. Dark zones represent regions where the disease isendemic, striped zones regions where the disease is controlled and under vaccination programmes and lightzones are FMD free (adapted from reference 8).

The control of the disease in India and other countries in the Far East is very difficult due to the

extremely large number of sheep, goats and cattle and to the poverty of many of the farmers. New

outbreaks occurred in Asian countries where the disease had been controlled (Malaysia, Philippines,

Japan) and FMD was recently introduced in Taiwan8. The uncontrollable movement of livestock

between countries of the Middle East has made it impossible to effectively control the disease in this

region of the world. Partial control of FMD was achieved only in Israel, upon immunisation with

vaccines produced in Europe. Of concern to Europe has been the situation in Turkey, since this is

the traditional route by which FMD enters the Balkans8. Sanitary measures such as movement

restrictions and quarantine, total slaughter of affected and in-contact animals (“stamping out”) and

extensive vaccination employing inactivated whole-virus have been successful in the control and

eradication of the disease in Europe (Fig. 0.19), which led to the decision of the European Union to

cease vaccination in 1991. This decision was followed, for the sake of trading agreements, by the

remaining European countries, and control of imports, quarantine and “stamping out” replaced

vaccination as the measures to exclude the disease. Nevertheless, several outbreaks have been

reported in Europe since 1991, namely in Bulgaria

(1991, 1993, 1996), Italy (1993), Greece (1994, 1996),

Russia (1995), Albania (1996), Macedonia (1996),

Kosovo (1996) and the Turkish Thrace (1995, 1996)8.

In the particular context of the Iberian Peninsula, the last

recorded outbreak occurred in Spain during 1986.

Although the peninsula has been FMD-free since then,

the large (and insufficiently controlled) flow in persons

and goods from and into Northern African territories is a

cause of grave concern for the animal health authorities,

even if not explicitly acknowledged.

0

5000

10000

15000

20000

25000

30000

35000

1960

1965

1970

1975

1980

Year

Outbreaks

Figure 0. 19 Estimated FMD outbreaksin Europe from 1960 to 1982 (reproducedfrom reference 1).

28


29

0.5 Foot-and-mouth disease virus

0.5.1 The virus particle

Foot-and-mouth disease virus (FMDV) was the first recognized viral pathogen (by Loeffler and

Frosch in 18989) and is the sole member of the genus Aphthovirus belonging to the Picornaviridæ

family. The viral particle, or virion, contains a single-stranded RNA of positive polarity,

approximately 8500 nucleotides long. The RNA is covalently linked to a small protein, VPg, at its 5’

terminus and translation of the RNA yields a single polypeptide (L-P1-P2-P3) which is then cleaved

into the structural (from the P1 region) and non-structural (such as the viral-specific protease 3C and

the viral-specific RNA polymerase 3D) proteins. The virus capsid is non-enveloped and has

icosahedral symmetry with a diameter of approximately 300 Å, consisting of 60 copies of each of

the structural proteins VP1, VP2, VP3 and VP4 (Fig. 0.20). While the first three structural proteins

(MW≈24 kDa) have surface components, the fourth (MW≈8.5 kDa) is internal. The virion is also

usually composed by one or two units of VP0, the precursor of VP2 and VP410,11.

Figure 0. 20 Illustration of the structure of three picornaviruses (FMDV, Mengo andHRV14) and their capsid proteins VP1-4 (reproduced from reference 11).


0.5.2 Architecture of the FMD virion

The structure of the FMDV particle was first resolved

for serotype O1 BFS 1860 by X-ray diffraction

analysis11. Since then, other serotypes of FMDV

have been crystallised and analysed, allowing some

of the phenotypic (e.g., high buoyant density in

CsCl, acid lability) and serological (immunological

and antigenic) properties of the virus to be

explained from a structural point of view12-15. The

overall shape of the outer virus surface is

approximately spherical and relatively smooth. The

virus has icosahedral symmetry (Fig. 0.21); each

ntamer11-15 formed upon assembly of five copies of the
asymmetric unit (1/12 of the particle) is a pe
Figure 0. 21 Structure of the viral capsid forFMDV C-S8c1 (only α C are represented); VP1– 4 are represented in blue, green, red andyellow, respectively (reproduced from reference

30

biological protomer of FMDV16. The arrangement of VP1, VP2 and VP3 in the biological protomer

is as represented in Fig. 0.22, where the internal VP4 is not displayed.

The viral proteins VP1 – 3 of FMDV are quite similar in size, position, orientation and tertiary

structure to those of other picornaviruses, with VP1 showing the most significant rearrangements.

VP4 is the most variable protein among picornaviruses, the one belonging to FMDV being the larger

(Fig. 0.20).

0.5.3 Antigenic structure of FMDV

It has long been known that the main cell

attachment site and the immunodominant region

of FMDV are both located on a solvent exposed

region at the surface of the virion, namely in

trypsin-sensitive areas of VP117,18. Earlier

serological studies showed that different serotypes

of FMDV shared a highly variable region of VP1,

comprising residues 135 to 155 (Fig. 0.23)19, as

one of the major antigenic sites of the virus.

Several overlapping B-cell epitopes are located

within this region and are able to induce both

neutralising and non-neutralising antibody

responses19-23. The high sequence variability

found in this region accounts for the low cross-

reactivity observed among different serotypes21-23.

Figure 0. 22 Ribbon protein diagram of theFMDV C-S8c1 protomer composed of proteinsVP1 – blue, VP2 – green and VP3 – yellow(reproduced from reference 16).


This immunodominant region was seen to correspond to the

loop which connects β-sheets G and H of the VP1 β-barrel,

named the GH loop11-15. Since the first evidences pointing to

the relevance of the GH loop in both the infectivity and

immune response in FMD, an enormous volume of research

has been focused on this region14,19-55. Unfortunately, the first

crystal structure of FMDV (strain O1 BFS) showed this region

to have very low electron density11, indicating high mobility

and thus lack of a defined structure. Based on the

assumption that such disordered conformation was

dependent on a native disulphide bond linking Cys134 of

VP1, at the base of the loop, and Cys130 of VP2, the crystal

structure of FMDV O1 BFS was analysed under reducing

conditions and the conformation of the loop was thus

resolved12 (shown in yellow in Fig. 0.23). Other important

antigenic and immunogenic sites have been identified in

several FMDV serotypes; for instance, the C-terminal stretch

of VP1 (which, together with the GH loop, defines the main

antigenic site 1 in serotype O), or sites involving different

loops from the three accessible viral proteins (e.g. sites 2, 3

Figure 0. 23 Localisation of the GHloop within the FMDV protomer(above) and detailed illustration of theconformation of this loop (below) –yellow for isolate O1 BFS; magenta forisolate C-S8c1. The RGD motif isshown in detail (reproduced fromreference 51).

31

and 4 of FMDV O)28,56-59. The absence of cross-reactivity between the different types of FMDV,

together with the lack of steric hindrance between serotype-specific mAbs in competition

experiments, clearly show that antigenic sites in these serotypes are topologically independent from

each other. Resolution of the crystal structures of other FMDV variants, such as FMDV C-S8c115, or

peptide/virus – antibody complexes33,36,42,51,52, provided further evidence of such topological

differences, as shown in Fig. 0.23.

0.5.4 FMDV cell attachment sites: the Arg-Gly-Asp motif

Studies on surface topology, sequence conservation and inhibition of cell attachment of different

picornaviruses have shown that the majority of these pathogens share a common strategy for hiding

their cell attachment sites from the immune system. Such sites are usually placed inside canyons or

pits, out of reach from antibody footprints60. The absence of any such canyons or pits in the smooth

FMDV surface11-15, as well as the existence of a highly conserved Arg-Gly-Asp (RGD) motif within

the hypervariable GH loop of VP111-59, led to suspect that this motif could have a key role in

infectivity, since RGD is known to promote cell attachment in several different systems61.

Immunochemical and structural studies have shown that the RGD motif is, in fact, critically involved

in FMDV infectivity, upon cell attachment via the integrin αvβ3, the vitronectin receptor18,27,36,41,42,62-70.

Being placed in a highly exposed region of FMDV, the RGD motif has been surprisingly conserved


32

among the different serotypes, in spite of the high immune pressure exerted on this region. The

strategy of FMDV to elude antibody recognition is based on surrounding RGD with hypervariable

residues within a disordered loop. Thus, a mechanism for escape from antibody neutralisation

would involve subtle structural modifications which preserve the integrin-recognisable open-turn

conformation of the RGD triplet (Fig. 0.23)41,42,44,62-69.

Despite its obvious relevance, the RGD motif is not the only possible route for FMDV to be

internalised by the host cells70-77. Increasing evidence that FMDV clones lacking the RGD triplet can

infect host cells has made the essentiality of this motif questionable. In fact, it is now known that

there are at least three different mechanisms for cell recognition by FMDV. Apart from the RGD-

integrin mechanism, there are isolates of FMDV which use heparan sulphate (HS) as the

predominant cell surface ligand72-75 (e.g., certain strains of FMDV O1, cell culture-adapted FMD

viruses) and even others which can establish RGD- and HS-independent infections76.

It has also been reported that FMDV can cause infection via the antibody-dependent enhancement

pathway, in which FMDV bound to virus-specific antibodies could enter cells via the Fc receptor,

thus bypassing the RGD mechanism70,71.

0.5.5 Antigenic and genetic variability of FMDV

RNA viruses are characterised by an error-prone RNA replication, which gives them great potential

for variation1,10. In FMDV genomes, the sequence homology between different serotypes can be as

low as 25-40% while homologies between subtypes of a same serotype are usually above 60-70%23.

Natural populations of FMDV from a single disease outbreak have been shown to be heterogeneous

and, moreover, “individual” isolates have been reported to include two different nucleotide

sequences. The high variability of FMDV (Table 0.3) led to the proposal that FMDV natural

populations are quasispecies, i. e., pools of variant genomes statistically defined but individually

indeterminate1,10,78. High mutation rates during replication allow FMD viruses to continuously evolve

and adapt to new environments. Although most mutations will be detrimental and eliminated by

natural selection (negative selection), others can be of value under the particular conditions where

the virus is replicating and are therefore selected (positive selection)1,10,75,78-87. Despite the high

heterogeneity of FMDV populations, there is a potential for long-term conservation of sequences

due to the continuous selection of a same consensus sequence in a situation of equilibrium79.

Whenever this equilibrium is ruptured, rapid evolution and selection of new master sequences take

place1,10,79.

One of the most troubling consequences of genetic variability is antigenic diversity. Immunochemical

studies have shown that isolates of the same geographical and chronological origin as well as viral

clones derived from single isolates may be antigenically distinct10,12,30,57,70,79-95. Antigenic variants

have been isolated under variable conditions, such as in partially immune animals, persistently

infected cattle4 and in cell culture96-98, in the latter case both in the presence or the absence of

immune pressure99,100. Therefore, antigenic variants result from the high mutation rates during RNA


33

replication and from the negative selection of most of the mutant phenotypes. This would mean that

substitutions at antigenic sites such as the FMDV GH loop are very likely to occur, since these are

disordered, flexible, and therefore, permissive sites, not subject to intensive negative selection.

This antigenic diversity has serious implications in vaccine design since synthetic vaccines should

include multiple independent epitopes in order to decrease the probability of selection of FMD

viruses resistant to the immune response.

Table 0.3 Variability of FMDV (reproduced from reference 12 and based on data from references therein).

Genetic heterogeneity Substitutions/genome

During a disease outbreakAmong consensus sequences of different isolatesAmong consensus sequences of contemporary isolatesAmong individual genomes of one isolate

Of clonal populations in cell cultureAmong consensus sequences of independently passagedplaque-purified virusesAmong individual genomes of clonal, passaged population

Frequency of mAb-resistant mutantsIn viruses from lesions of infected animalsIn viruses from cell culture fluid

60 – 702 – 200.6 – 2

14 – 572 – 8

2××××10-6 – 2××××10-5

4××××10-5

Evolution Substitutions/nucleotide/year

Rate of fixation of mutationsAcute diseasePersistent infection

<4××××10-4 – 4.5××××10-2

9××××10-3 – 7.4××××10-2

0.6 The development of anti-FMDV vaccines

0.6.1 Conventional vaccines

Vaccination has been one of the most powerful tools for efficient control of infectious diseases such

as poliomyelitis, measles, yellow fever and smallpox, the latter having been totally eradicated world-

wide. Conventional vaccines are whole-virus vaccines where attenuated variants or inactivated

viruses are employed. The possibility to control viral RNA quasispecies with classical vaccines relies

on two important factors: i. attempts of the virus to escape immune response upon mutation lead to

non-viable phenotypes which cannot adapt to the environment, and ii. the constant actualisation of

vaccine strains to include field variants from new outbreaks provides a broad coverage of the genetic

and antigenic heterogeneity found in the field. Nevertheless, high mutational rates are still an

obstacle to the efficacy of RNA viral vaccines. Also, on a more practical level, not all viruses are

easily grown in cell culture, a fact that can often prevent the production of classical vaccines, such as

hepatitis A101.

Most vaccines against FMDV are prepared by growing the virus in surviving bovine tongue epithelial

fragments, pig or calf kidney cell monolayers, or in baby hamster kidney cell culture and subsequent

inactivation with ethyleneimine (aziridine). The inactivated virus is then adsorbed onto aluminium


34

hydroxide and mixed with saponin prior to inoculation; vaccine delivery can also rely on emulsions

with an oil adjuvant. These classical anti-FMDV vaccines, given as a single dose, have been effective

in the control of the disease in Western Europe3,8,101-105.

The need of cold chains to keep viral vaccines at low temperatures in order to preserve their

immunogenicity is one of the reasons for the unsuccessful vaccination programmes in countries with

difficult terrain and climate conditions (e.g., tropical countries). Other disadvantages of whole-virus

vaccines arise from occasional deficient inactivation of the infective particle and consequent escape

to the field, causing new FMD outbreaks and eventually establishing persistent infections in cattle,

which act as important reservoirs and factories of new variants. But, clearly, the most important

problem of anti-FMDV classical vaccines comes from the high antigenic diversity of this virus, since

convalescent animals recovering from infection with a particular serotype are not protected against

other serotypes. Furthermore, each serotype consists of a wide spectrum of variant isolates and often

the virus strain used to prepare vaccines against a certain serotype does not offer the same degree of

protection against other isolates of the same serotype. Moreover, adaptation of an outbreak virus to

growth in cell culture can lead to the selection of variants that are antigenically different from their

parental virus1,2,10,12,30,70,79-98.

0.6.2 Synthetic vaccines

In view of the difficulties posed by conventional vaccines, the development of synthetic, molecularly

engineered vaccines has become a priority for the control of viral diseases. In particular, the use of

peptide-based synthetic vaccines offers significant advantages over classical procedures in terms of

stability, availability, safety, purity and cost101,106. These benefits are not easily achieved, however.

Thus, in order to design effective candidate vaccines, the antigenic and immunogenic determinants

of the pathogen must be adequately understood.

Intensive research has been focused on FMDV B-cell epitopes with the hope that they could be

mimicked by short linear peptides capable of eliciting protective virus-specific immune

responses20,25,26,39,43,45,107-115. As mentioned before, early studies allowed the recognition of major

antigenic sites located within protein VP1, namely the GH loop (antigenic site A) and the C-terminal

region (antigenic site C). Peptide vaccines based on site A or on constructs including both A and C

sites (the so-called “DiMarchi” peptides, Fig. 0.24) induced significant levels of anti-FMDV

neutralising antibodies and protected either guinea-pigs or natural hosts (pigs and cattle)116-118.

CysCys ProCysGlyVP1 C-terminus(residues 200-213)

VP1 G-H loop(residues 141-158)

ProProSer

Figure 0. 24 The “DiMarchi” peptide antigen: the VP1 C-terminal and GH loop regions from FMDV O1

Kaufbeuren were brought together in a linear construct; a ProProSer spacer was used to induce a turn andCys residues were placed at each terminus to allow oligomerisation and bypass the use of a carrier protein.


The immunogenicity of synthetic FMDV antigens was shown to be

generally lower than that of classical vaccines in natural hosts. Also

differences between anti-virus and anti-peptide immune responses were

detected, namely, the good correlation between neutralising activity of anti-

virus sera and host protection was not always well established in peptide-

immunised animals119-123. Attempts to increase the immunogenicity of small

FMDV peptides include the design of constructs containing tandem repeats

of the linear peptide, attachment of peptide to carrier proteins such as

bovine serum albumin (BSA) and keyhole limpet hæmocyanin (KLH)107-115

(Fig. 0.25) or insertion in scaffolds such as multiple antigenic peptide

(MAP) systems124 (Fig. 0.26), recombinant proteins (e.g., β-galactosidase

from Escherichia coli35,40,125,126) and hepatitis B virus core (HBc) protein

which self-assembles into a spherical virus-like particle, the hepatitis B core

antigen (HBcAg)67,101 (Fig. 0.27). Recombinant technology is also

important in protection against FMDV, as shown by recent results using

recombinant viruses or transgenic plants where VP1 or the precursor

polypeptide P1 of FMDV capsid proteins have been inserted127-130.

+

Carrier protein(BSA or KLH)

Carrier-peptideconjugate

Figure 0. 25Representation of apeptide-carrier proteinconjugate (peptiderepresented as a black“loop”).

35

NH2

NH2

NH2

NH2NH2

NH2

NH2

NH2

+

NH

NH

NH

NHNH

NH

NH

NH

Figure 0. 26 Multiple antigenic peptide (MAP) system: a poly-lysine scaffoldis used to present in a single chimera several copies (8, in the present case) ofthe synthetic antigen, represented as a black “loop” (adapted from reference101).

In terms of antigenicity and immunogenicity, comparison between different peptide vaccine

candidates clearly shows that peptide presentation and orientation are important25,35,40,101,125,126,131.

Thus, insertion of a peptide reproducing site A of FMDV C-S8c1 on different solvent-exposed loops

of the homotetrameric enzyme β-galactosidase yielded different antigenicity levels of the resulting

chimeras, some of them more antigenic than the corresponding KLH conjugate35,40,125,126. These

results prove the sensitivity of anti-FMDV antibody responses to peptide conformation, regardless of

the localisation of antigenic site A on a linear and flexible loop. Another evidence of such

dependence on orientation was reported by Schaaper et al., who observed anti-peptide immune

responses dependent on the peptide-carrier coupling method131.


36

+

HBc protein

HBc protein-peptide

HBcAg-peptide assembly

Figure 0. 27 Antigenic peptide (black “loop”) insertion into hepatitis B viruscore protein and self-assembly of the latter into hepatitis B core antigen(HbcAg) particle (adapted from reference 101).

Despite these differences, short linear and cyclic peptides have been shown to reproduce rather

faithfully the features of antigenic site A from different isolates, including C-S strains 29,31,34,36-

38,41,44,49,50. This opened the possibility of analysing in detail the effects of amino acid replacements

found in natural isolates and the repercussions of antigenic variation in the field132,133. Moreover, it

allowed an extensive screening of the effects of single-point replacements of amino acids spanning

the entire GH loop41 (Fig. 0.28) which, together with the recently resolved structures of some

peptide-antibody complexes36,42,44, provided further insight into the mechanisms of interaction

between the GH loop and anti-FMDV neutralising antibodies. The knowledge of such mechanisms

at the molecular level can provide the basis for the design of FMDV peptides with strong antigenic

character, suitable to be inserted in constructs including T-cell epitopes and other immunogenic

components to produce efficient synthetic anti-FMDV vaccines. Recent advances with retro-inverso

FMDV peptides135-138 (increasing peptide resistance to host proteases) and with synthetic models of

important discontinuous antigenic sites (site D from FMDV C-S8c1 isolate)59 are also encouraging

regarding the future of fully synthetic anti-FMDV vaccines.

Y T A S A . R G D L A H L T T T

SD6

4C4

6D11

7JD1

7CA11

5A2

7FC12

infectivity

Figure 0. 28 Sequence of the GH loop from C-S8c1 clone of FMDV (adapted from reference 134).

Above the sequence: ( ) variable residues found in 50 field isolates of serotype C FMDV; () residues found replaced in 97laboratory FMDV mutants (89 of them derived from C-S8c1) selected by antibodies; () replaced residues found after 25independent passages of FMDV in cell culture, in the absence of immune pressure. Below the sequence: Average effects ofsingle-point replacements within site A on antigenicity towards 7 anti-FMDV monoclonal antibodies, where a black boxstands for IC50>100, a vertically striped box for 30<IC50<100 and a white box to IC50<5. The last row represents theaverage effects of single-point mutations within the GH loop on inhibition of infectivity (FMDV C-S8c1), where a black boxstands for IC50>300, a vertically striped box for 30<IC50<300 and a horizontally striped box for 5<IC50<30.


37

References

1 Domingo, E., Mateu, M. G., Martínez, M. A., Dopazo, J., Moya, A. and Sobrino, F. Genetic variabilityand antigenic diversity of foot-and-mouth disease virus, in: Kurstak, E., Marusk, R. G., Murphy, S.A .and Van Regenmortel, M. H. V., eds. “Applied Virology Research” – vol. 2, Plenum Publishing Co.,New York: 1990; pp. 233-266.

2 Woodbury, R. L. (1995) A review of the possible mechanisms for the persistence of foot-and-mouthdisease virus, Epidemiol. & Infect. 114, 1-13.

3 Doel, T. R. (1996) Natural and vaccine-induced immunity to foot-and-mouth disease: the prospectsfor improved vaccines, Ver. Sci. Tech. Off. Int. Epiz. 15, 883-911.

4 Gebauer, F., de la Torre, J. C., Gomes, I., Mateu, M. G., Barahona, H., Tiraboschi, B., Bergmann, I.,Augé de Mello, P. and Domingo, E. (1988) Rapid selection of genetic and antigenic variants of foot-and-mouth disease virus during persistence in cattle, J. Virol. 62, 2041-2049.

5 Hernández, A. M. M., Carrilo, E. C., Sevilla, N. and Domingo, E. (1994) Rapid cell variation candetermine the establishment of a persistent viral infection, Proc. Natl. Acad. Sci. USA 91, 3705-3709.

6 Fracastorri, H. (translation) in: Wright, W. C., eds. “Contagion, Contagious diseases and theirtreatment.”, London, Putnam: 1545

7 Le Vailant, “Travels into the interior parts of Africa 1780-1785” – vol. 2: 1795.8 Kitching, R. P. (1998) A recent history of foot-and-mouth disease, J. Comp. Path. 118, 89-108.9 Lœffler, F., Frosch, P. (1898) Berichte der Kommission zur Erforschung der Maul- und Klauenseuche

bei dem Institut für Infectrionskrankheiten in Berlin, Zbl. Bakter Abt. I. Orig. 23, 371-391.10 Domingo, E., Escarmís, C., Martínez, M. A., Martínez-Salas, E. and Mateu, M. G. (1992) Foot-and-

mouth disease virus populations are quasispecies, Curr. Topics Microbiol. Immunol. 176, 33-47.11 Acharya, R., Fry, E., Stuart, D., Fox, G., Rowlands, D. and Brown, F. (1989) The three-dimensional

structure of foot-and-mouth disease virus at 2.9 Å resolution, Nature 337, 709-716.12 Parry, N., Fox, G., Rowlands, D., Brown, F., Fry, E., Acharya, R., Logan, D. and Stuart, D. (1990)

Structural and serological evidence for a novel mechanism of antigenic variation in foot-and-mouthdisease virus, Nature 347, 569-572.

13 Curry, S., Abu-Ghazaleh, R., Blakemore, W., Fry, E., Jackson, T., King, A., Lea, S., Logan, D.,Newman, J. and Stuart, D. (1992) Crystallization and preliminary X-ray analysis of three serotypes offoot-and-mouth disease virus, J. Mol. Biol. 228, 1263-1268.

14 Logan, D., Abu-Ghazaleh, R., Blakemore, W., Curry, S., Jackson, T., King, A., Lea, S., Lewis, R.,Newman, J., Parry, N., Rowlands, D. J., Stuart, D. and Fry, E. (1993) Structure of a majorimmunogenic site on foot-and-mouth disease virus, Nature 362, 566-568.

15 Lea, S., Hernández, J., Blakemore, W., Brocchi, E., Curry, S., Domingo, E., Fry, E., Abu-Ghazaleh,R., King, A., Newman, J., Stuart, D. and Mateu, M. G. (1994) The structure and antigenicity of a typeC foot-and-mouth disease virus, Structure 2, 123-139.

16 http://www.rscb.org/pdb/ (structure code 1FMD)17 Laporte, J. and Lenoir, G. (1973) Structural proteins of foot-and-mouth disease virus, J. Gen. Virol.

20, 161-168.18 Cavanagh, D., Sangar, D. V., Rowlands, D. J. and Brown, F. (1977) Immunogenic and cell

attachment sites of foot-and-mouth disease virus: further evidence for their location in a single capsidpolypeptide, J. Gen. Virol. 35, 149-158.

19 Strohmaier, K., Franze, R. and Adam, K. H. (1982) Location and characterization of the antigenicportion of the foot-and-mouth disease virus immunizing protein, J. Gen. Virol. 59, 295-306.

20 Oulridge, E. J., Barnett, P. V., Parry, N. R., Syred, A., Head, M. and Rweyemamu, M. M. (1984)Demonstration of neutralizing and non-neutralizing epitopes on the trypsin-sensitive site of foot-and-mouth disease virus, J. Gen. Virol. 65, 203-207.

21 Cheung, A., DeLamarter, J., Weiss, S. and Küpper, H. (1983) Comparison of the major antigenicdeterminants of different serotypes of foot-and-mouth disease virus, J. Virol. 48, 451-459.

22 Beck, E., Feil, G. and Strohmaier, K. (1983) The molecular basis of the antigenic variation of foot-and-mouth disease virus, EMBO J. 2, 555-559.

23 Grubman, M. J., Zellner, M. and Wagner, J. (1987) Antigenic comparison of the polypeptides of foot-and-mouth disease virus serotypes and other picornaviruses, Virology 158, 133-140.

24 Robertson, B. H., Moore, D. M., Grubman, M. J. and Kleid, D. (1983) Identification of an exposedregion of the immunogenic capsid polypeptide VP1 on foot-and-mouth disease virus, J. Virol. 46,311-316.

25 Parry, N. R., Syred, A., Rowlands, D. J. and Brown, F. (1988) A high proportion of anti-peptideantibodies recognize foot-and-mouth disease virus particles, Immunology 64, 567-572.

26 Pfaff, E., Thiel, H. J., Strohmaier, K. and Scaller, H. (1988) Analysis of neutralizing epitopes on foot-and-mouth disease virus, J. Virol. 62, 2033-2040.


38

27 Fox, G., Parry, N. R., Barnett, P. V., McGinn, B., Rowlands, D. J. and Brown, F. (1989) The cellattachment site on foot-and-mouth disease virus includes the amino acid sequence RGD (Arginine-Glycine-Aspartic Acid), J. Gen. Virol. 70, 625-637.

28 McCahon, D., Crowther, J. R., Belsham, G. J., Kitson, J. D. A., Duchesne, M., Have, P., Meloen, R.H., Morgan, D. O. and de Simone, F. (1989) Evidence for at least four antigenic sites on type O foot-and-mouth disease virus involved in neutralization; Identification by single and multiple-sitemonoclonal antibody-resistant mutants, J. Gen. Virol. 70, 638-644.

29 Camarero, J. A., Andreu, D., Cairó, J. J., Mateu, M. G., Domingo, E. and Giralt, E. (1993) Cyclicdisulfide model of the major antigenic site of serotype C foot-and-mouth disease virus (synthetical,conformational and immunochemical studies), FEBS Lett. 328, 159-164.

30 Krebs, O., Ahl, R., Strauß, O., C. and Marquardt, O. (1993) Amino acid changes outside the G-Hloop of capsid protein VP1 of type O foot-and-mouth disease virus confer resistance to neutralizationby anti-peptide G-H serum, Vaccine 11, 359-362.

31 Roig, X., Novella, I. S., Giralt, E. and Andreu, D. (1994) Examining the relationship betweensecondary structure and antibody recognition in immunopeptides from foot-and-mouth disease virus,Lett. Pept. Sci. 1, 39-49.

32 France, L. L., Piatti, P. G., Newman, J. F. E., Toth, I., Gibbons, W. A. and Brown, F. (1994) Circulardichroism, molecular modeling and serology indicate that the structural basis of antigenic variation infoot-and-mouth disease virus is α-helix formation, Proc. Natl. Acad. Sci. USA 91, 8442-8446.

33 Verdaguer, N., Mateu, M. G., Bravo, J., Tormo, J., Giralt, E., Andreu, D., Domingo, E. and Fita, I.(1994) Crystallization and preliminary X-ray diffraction studies of a monoclonal Fab fragment againstfoot-and-mouth disease virus and of its complex with the main antigenic site peptide, Proteins 18,201-203.

34 Valero, M. L., Camarero, J. A., Adeva, A., Verdaguer, N., Fita, I., Mateu, M. G., Domingo, E., Giralt,E. and Andreu, D. (1995) Cyclic peptides as conformationally restricted models of viral antigens:application to foot-and-mouth disease virus, Biomed. Peptides, Prot. & Nucl. Acids 1, 133-140.

35 Benito, A., Mateu, M. G. and Villaverde, A. (1995) Improved mimicry of a foot-and-mouth diseasevirus antigenic site by a viral peptide displayed on β-galactosidase surface, Biotechnology 13, 801-804.

36 Verdaguer, N., Mateu, M. G., Andreu, D., Giralt, E., Domingo, E. and Fita, I (1995) Structure of themajor antigenic loop of foot-and-mouth disease virus complexed with a neutralizing antibody: directinvolvment of the Arg-Gly-Asp motif in the interaction, EMBO J. 14, 1690-1696.

37 Mateu, M. G., Camarero, J. A., Giralt, E., Andreu, D. and Domingo, E. (1995) Direct evaluation ofthe immunodominance of a major antigenic site of foot-and-mouth disease virus in a natural host,Virology 206, 298-306.

38 Mateu, M. G., Andreu, D. and Domingo, E. (1995) Antibodies raised in a natural host andmonoclonal antibodies recognize similar antigenic features of foot-and-mouth disease virus, Virology210, 120-127.

39 Zamorano, P., Wigdorovitz, A., Pérez-Filguera, M., Carrillo, C., Escribano, J. M., Sadir, A. M. andBorca, M. V. (1995) A 10-amino acid linear sequence of VP1 of foot-and-mouth disease viruscontaining B- and T-cell epitopes induces protection in mice, Virology 212, 614-621.

40 Carbonell, X., Benito, A. and Villaverde, A. (1996) Converging antigenic structure of a recombinantviral peptide displayed on different frameworks of carrier proteins, FEBS Lett. 397, 169-172.

41 Mateu, M. G., Valero, M. L., Andreu, D. and Domingo, E. (1996) Systematic replacement of aminoacid residues within an Arg-Gly-Asp containing loop of foot-and-mouth disease virus and effect on cellrecognition, J. Biol. Chem. 271, 12814-12819.

42 Verdaguer, N., Mateu, M. G., Bravo, J., Domingo, E. and Fita, I. (1996) Induced pocket toaccomodate the cell attachment site Arg-Gly-Asp motif in a neutralizing antibody against foot-and-mouth disease virus, J. Mol. Biol. 256, 364-376.

43 Pegna, M., Molinari, H., Zetta, L., Melacini, G., Gibbons, W. A., Brown, F., Rowlands, D., Chan, E.and Mascagni, R. (1996) the solution conformational features of two highly homologous antigenicpeptides of foot-and-mouth disease virus serotype A, variants A and USA, correlate with theirserological properties, J. Pep. Sci. 2, 91-105.

44 Verdaguer, N., Sevilla, N., Valero, M. L., Stuart, D., Brocchi, E., Andreu, D., Giralt, E., Domingo, E.,Mateu, M. G. and Fita, I. (1998) A similar pattern of interaction for different antibodies with a majorantigenic site of foot-and-mouth disease virus: implications for intratypic antigenic variation, J. Virol.72, 739-748.

45 Volpina, O. M., Yarov, A. V., Zhmak, M. N., Kuprianova, M. A., Chepurkin, A. V., Toloknov, A. S.and Ivanov, V. T. (1996) Synthetic vaccine against foot-and-mouth disease based on a palmitoylderivative of the VP1 protein 135-159 fragment of the A22 virus strain, Vaccine 14, 1375-1380.

46 De Prat-Gay, G. (1997) Conformational preferences of a peptide corresponding to the major antigenicdeterminant of foot-and-mouth disease virus: implications for peptide-vaccine approaches, Arch.Biochem. Biophys. 341, 360-369.


39

47 Zamorano, P., Wigdorovitz, A., Pérez-Filguera, M., Escribano, J. M., Sadir, A. M. and Borca, M. V.(1998) Induction of anti-foot-and-mouth disease virus T and B cell responses in cattle immunized witha peptide representing ten amino acids of VP1, Vaccine 16, 558-563.

48 Valero, M. L., Giralt, E. and Andreu, D. (1999) A comparative study of cyclization strategies applied tothe synthesis of head-to-tail cyclic analogs of a viral epitope, J. Peptide Res. 53, 56-67.

49 Haack, T., Camarero, J. A., Roig, X., Mateu, M. G., Domingo, E., Andreu, D. and Giralt, E. (1997) Acyclic disulfide peptide reproduces in solution the main structural features of a native antigenic site offoot-and-mouth disease virus, Int. J. Biol. Macromol. 20, 209-219.

50 Valero, M. L., Camarero, J. A., Haack, T., Mateu, M. G., Domingo, E., Giralt, E. and Andreu, D.(2000) Native-like cyclic peptide models of a viral antigenic site: finding a balance between rigidityand flexibility, J. Mol. Recognit. 13, 5-13.

51 Hewat, E. A., Verdaguer, N., Fita, I., Blakemore, W., Brookes, S., King, A., Newman, J., Domingo, E.,Mateu, M. G. and Stuart, D. I. (1997) Structure of the complex of an Fab fragment of a neutralizingantibody with foot-and-mouth disease virus: positioning of a highly mobile antigenic loop, EMBO J.16, 1492-1500.

52 Verdaguer, N., Schoehn, G., Ochoa, W. F., Fita, I., Brookes, S., King, A., Domingo, E., Mateu, M. G.,Stuart, D. and Hewat, E. A. (1999) Flexibility of the major antigenic loop of foot-and-mouth diseasevirus bound to a Fab fragment of a neutralizing antibody: structure and neutralization, Virology 255,260-268.

53 Domingo, E., Verdaguer, N., Ochoa, W. F., Ruiz-Jarabo, C. M., Sevilla, N., Baranowski, E., Mateu,M. G. and Fita, I. (1999) Biochemical and structural studies with neutralizing antibodies raised againstfoot-and-mouth disease virus, Virus Res. 62, 169-175.

54 Feliu, J. X., Benito, A., Oliva, B., Avilés, F. X. and Villaverde, A. (1998) Conformational flexibility in ahighly mobile protein loop of foot-and-mouth disease virus: distinct structural requirements for integrinand antibody binding, J. Mol. Biol. 283, 331-338.

55 Benito, A. and Van Regenmortel, M. H. V. (1998) Biosensor characterization of antigenic site A offoot-and-mouth disease virus presented in different vector systems, FEMS Immunol. & Med.Microbiol. 21, 101-115.

56 Barnett, P. V., Samuel, A. R., Pullen, L., Ansell, D., Butcher, R. N. and Parkhouse, R. M. E. (1998)Monoclonal antibodies, against O1 serotype foot-and-mouth disease virus, from a natural bovine host,recognize similar antigenic features to those defined by the mouse, J. Gen. Virol. 79, 1687-1697.

57 Mateu, M. G., Hernández, J., Martínez, M. A., Feigelstock, D., Lea, S., Pérez, J. J., Giralt, E., Stuart,D., Palma, E. L. and Domingo, E. (1994) Antigenic heterogeneity of a foot-and-mouth disease virusserotype in the field is mediated by very limited sequence variation at several antigenic sites, J. Virol.68, 1407-1417.

58 Mateu, M. G., Escarmís, C. and Domingo, E. (1998) Mutational analysis of discontinuous epitopes offoot-and-mouth disease virus using an unprocessed capsid protomer precursor, Virus Res. 53, 27-37.

59 Borràs, E., Giralt, E. and Andreu, D. (1999) A rationally designed synthetic peptide mimic of adiscontinuous viral antigenic site elicits neutralizing antibodies, J. Am. Chem. Soc. 121, 11932-11933.

60 Chapman, M. S. and Rossmann, M. G. (1993) Comparison of surface properties of picornaviruses:strategies for hiding the receptor site from immune surveillance, Virology 195, 745-756.

61 Hynes, R. O. (1992) Integrins: versatility, modulation and signaling in cell adhesion, Cell 69, 11-25.62 Novella, I. S., Borrego, B., Mateu, M. G., Domingo, E., Giralt, E. and Andreu, D. (1993) Use of

substituted and tandem-repeated peptides to probe the relevance of the highly conserved RGDtripeptide in the immune response against foot-and-mouth disease virus, FEBS Lett. 330, 253-259.

63 Berinstein, A., Roivainen, M., Hovi, T., Mason, P. W. and Baxt, B. (1995) Antibodies to thevitronectin receptor (integrin αvβ3) inhibit binding and infection of foot-and-mouth disease virus tocultured cells, J. Virol. 69, 2664-2666.

64 Villaverde, A., Feliu, J. X., Harbottle, R. P., Benito, A. and Coutelle, C. (1996) A recombinant,arginine-glycine-aspartic acid (RGD) motif from foot-and-mouth disease virus binds mammalian cellsthrough vitronectin and, to a lower extent, fibronectin receptors, Gene 180, 101-106.

65 J. Hernández, Valero, M. L., Andreu, D., Domingo, E. and Mateu, M. G. (1996) Antibody and hostcell recognition of foot-and-mouth disease virus (serotype C) cleaved at the Arg-Gly-Asp (RGD) motif:a structural interpretation, J. Gen. Virol. 77, 257-264.

66 Jackson, T., Sharma, A., Abu-Ghazaleh, R., Blakemore, W. E., Ellard, F. M., Simmons, D. L.,Newman, J. W. I., Stuart, D. I. and King, A. M. Q. (1997) Arginine-glycine-aspartic acid-specificbinding by foot-and-mouth disease viruses to the purified integrin αvβ3 in vitro, J. Virol. 71, 8357-8361.

67 Sharma, A., Rao, Z., Fry, E., Booth, T., Jones, E. Y., Rowlands, D. J., Simmons, D. L. and Stuart, D.I. (1997) Specific interactions between human integrin αvβ3 and chimeric hepatitis B virus coreparticles bearing the receptor-binding epitope of foot-and-mouth disease virus, Virology 239, 150-157.


40

68 Villaverde, A., Feliu, J. X., Arís, A., Harbottle, R.. P., Benito, A. and Coutelle, C. (1998) A celladhesion peptide from foot-and-mouth disease virus can direct cell targeted delivery of a functionalenzyme, Biotechnol. Bioeng. 59, 295-301.

69 Neff, S., Sá-Carvalho, D., Rieder, E., Mason, P. W., Blystone, S. D., Brown, E. J. and Baxt, B. (1998)Foot-and-mouth disease virus virulent for cattle utilizes the integrin αvβ3 as its receptor, J. Virol. 72,3587-3594.

70 Ruiz-Jarabo, C. M., Sevilla, N., Dávila, M., Gómez-Mariano, G., Baranowski, E. and Domingo, E.(1999) Antigenic properties and population stability of a foot-and-mouth disease virus with an alteredArg-Gly-Asp receptor recognition motif, J. Gen. Virol. 80, 1899-1909.

71 Mason, P. W., Baxt, B., Brown, F., Harber, J., Murdin, A. and Wimmer, E. (1993) Antibody-complexed foot-and-mouth disease virus, but not poliovirus, can infect normally insusceptible cells viathe Fc receptor, Virology 192, 568-577.

72 Mason, P. W., Rieder, E. and Baxt, B. (1994) RGD sequence of foot-and-mouth disease virus isessential for infecting cells via the natural receptor but can be bypassed by an antibody-dependentenhancement pathway, Proc. Natl. Acad. Sci. USA 91, 1932-1936.

73 Jackson, T., Ellard, F. M, Abu-Ghazaleh, R., Brookes, S. M., Blakemore, W. E., Corteyn, A. H.,Stuart, D. I., Newman, J. W. I. and King, A. M. Q. (1996) Efficient infection of cells in culture by typeO foot-and-mouth disease virus requires binding to cell surface heparan sulfate, J. Virol. 70, 5282-5287.

74 Sá-Carvalho, D., Rieder, E., Baxt, B., Rodarte, R., Tanuri, A. and Mason, P. W. (1997) Tissue cultureadaptation of foot-and-mouth disease virus selects viruses that bind to heparin and are attenuated incattle, J. Virol. 71, 5115-5123.

75 Fry, E. E., Lea, S. M., Jackson, T., Newman, J. I., Ellard, F. M., Blakemore, W. E., Abu-Ghazaleh, R.,Samuel, A., King, A. M. Q. and Stuart, D. I. (1999) The structure and function of a foot-and-mouthdisease virus-oligosaccharide receptor complex, EMBO J. 18, 543-554.

76 Baranowski, E., Ruiz-Jarabo, C., Sevilla, N., Andreu, D., Beck, E. and Domingo, E. (2000) Cellrecognition by foot-and-mouth disease virus that lacks the RGD integrin-binding motif: flexibility inAphthovirus receptor usage, J. Virol. 74, 1641-1647.

77 Brown, F., Benkirane, N., Limal, D., Halimi, H., Newman, J. F. I., Van Regenmortel, M. H. V.,Briand, J. P. and Müller, S. (2000) Delineation of a neutralizing subregion within theimmunodominant epitope (GH loop) of foot-and-mouth disease virus VP1 which does not contain theRGD motif, Vaccine 18, 50-56.

78 Novella, I. S., Domingo, E. and Holland, J. J. (1996) Rapid viral quasispecies evolution: implicationsfor vaccine and drug strategies, Molec. Med. Today 53, 248-253.

79 Piccone, M. E., Kaplan, G., Giavedoni, L., Domingo, E. and Palma, E. L. (1988) VP1 of serotype Cfoot-and-mouth disease virus: long-term conservation of sequences, J. Virol. 62, 1469-1473.

80 Sáiz, J. C. and Domingo, E. (1996) Virulence as a positive trait in viral persistence, J. Virol. 70, 6410-6413.

81 Sevilla, N. and Domingo, E. (1996) Evolution of a persistent Aphthovirus in cytolytic infections: partialreversion of phenotypic traits accompanied by genetic diversification, J. Virol. 70, 6617-6624.

82 Domingo, E., Mateu, M. G., Escarmís, C., Martínez-Salas, E., Andreu, D., Giralt, E., Verdaguer, N.and Fita, I. (1996) Molecular evolution of aphthoviruses, Virus Genes 11, 197-207.

83 Charpentier, N., Dávila, M., Domingo, E. and Escarmís, C. (1996) Long-term, large-populationpassage of Aphthovirus can generate and amplify defective noninterfering particles deleted in theleader protease gene, Virology 223, 10-18.

84 Martínez, M. A., Verdaguer, N., Mateu, M. G. and Domingo, E. (1997) Evolution subvertingessentiality: dispensability of the cell attachment Arg-Gly-Asp motif in multiply passaged foot-and-mouth disease virus, Proc. Natl. Acad. Sci. 94, 6798-6802.

85 Carrillo, C., Borca, M., Moore, D. M., Morgan, D. O. and Sobrino, F. (1998) In vivo analysis of thestability and fitness of variants recovered from foot-and-mouth disease virus quasispecies, J. Gen.Virol. 79, 1699-1706.

86 Sevilla, N., Ruiz-Jarabo, C. M., Gómez-Mariano, G., Baranowski, E. and Domingo, E. (1998) An RNAvirus can adapt to the multiplicity of infection, J. Gen. Virol. 79, 2971-2980.

87 Baranowski, E., Sevilla, N., Verdaguer, N., Ruiz-Jarabo, C. M., Beck, E. and Domingo, E. (1998)Multiple virulence determinants of foot-and-mouth disease virus in cell culture, J. Virol. 72, 6362-6372.

88 Mateu, M. G., Rocha, E., Vicente, O., Vayreda, F., Navalpotro, C., Andreu, D., Pedroso, E., Giralt, E.,Enjuanes, L. and Domingo, E. (1987) Reactivity with monoclonal antibodies of viruses from anepisode of foot-and-mouth disease, Virus Res. 8, 261-274.

89 Mateu, M. G., da Silva, J. L., Rocha, E., de Brum, D. L., Alonso, A., Enjuanes, L., Domingo, E. andBarahona, H. (1988) Extensive antigenic heterogeneity of foot-and-mouth disease virus serotype C,Virology 167, 113-124.


41

90 Mateu, M. G., Martínez, M. A., Andreu, D., Parejo, J., Giralt, E., Sobrino, F. and Domingo, E. (1989)Implications of a quasispecies genome structure: effect of frequent, naturally occurring, amino acidsubstitutions on the antigenicity of foot-and-mouth disease virus, Proc. Natl. Acad. Sci. USA 86,5883-5887.

91 Mateu, M. G., Martínez, M. A., Cappucci, L., Andreu, D., Giralt, E., Sobrino, F., Brocchi, E. andDomingo, E. (1990) A single amino acid substitution affects multiple overlapping epitopes in themajor antigenic site of foot-and-mouth disease virus of serotype C, J. Gen. Virol. 71, 629-637.

92 Martínez, M. A., Hernández, J., Piccone, M. E., Palma, E. L., Domingo, E., Knowles, N. and Mateu,M. G. (1991) Two mechanisms of antigenic diversification of foot-and-mouth disease virus, Virology184, 695-706.

93 Feigelstock, D., Mateu, M. G., Piccone, M. E., de Simone, F., Brocchi, E., Domingo, E. and Palma, E.L. (1992) Extensive antigenic diversification of foot-and-mouth disease virus by amino acidsubstitutions outside the major antigenic site, J. Gen. Virol. 73, 3307-3311.

94 Feigelstock, D. A., Mateu, M. G., Valero, M. L., Andreu, D., Domingo, E. and Palma, E. L. (1996)Emerging foot-and-mouth disease virus variants with antigenically critical amino acid substitutionspredicted by model studies using reference viruses, Vaccine 14, 97-102.

95 Sevilla, N., Verdaguer, N. and Domingo, E. (1996) Antigenically profound amino acid substitutionsoccur during large population passages of foot-and-mouth disease virus, Virology 225, 400-405.

96 Díez, J., Dávila, M., Escarmís, C., Mateu, M. G., Domínguez, J., Pérez, J. J., Giralt, E., Melero, J. A.and Domingo, E. (1990) Unique amino acid substitutions in the capsid protein of foot-and-mouthdisease virus from a persistent infection in cell culture, J. Virol. 64, 5519-5528.

97 Meyer, R. F. and Brown, F. (1995) Sequence identification of antigenic variants in plaque isolates offoot-and-mouth disease virus, J. Virol. Meth. 55, 281-283.

98 Piatti, P., Hassard, S., Newman, J. F. E. and Brown, F. (1995) Antigenic variants in plaque-isolate offoot-and-mouth disease virus: implications for vaccine production, Vaccine 13, 781-784.

99 Borrego, B., Novella, I. S., Giralt, E., Andreu, D. and Domingo, E. (1993) Distinct repertoire ofantigenic variants of foot-and-mouth disease virus in the presence or absence of immune selection, J.Virol. 67, 6071-6079.

100 Holguín, A., Hernández, J., Martínez, M. A., Mateu, M. G. and Domingo, E. (1997) Differentialrestrictions on antigenic variation among antigenic sites of foot-and-mouth disease virus in theabsence of antibody selection, J. Gen. Virol. 78, 601-609.

101 Brown, F. (1990) Synthetic peptides as potential vaccines against foot-and-mouth disease, Endeavour14, 87-94.

102 Twomey, T., Newman, J., Burrage, T., Piatti, P., Lubroth, J. and Brown, F. (1995) Structure andimmunogeneicity of experimental foot-and-mouth disease and poliomielytis vaccines, Vaccine 13,1603-1610.

103 Brown, F. (1999) Foot-and-mouth disease and beyond: vaccine design, past, present and future,Arch. Virol. [Suppl.] 15, 179-188.

104 Mason, P. W., Piccone, M. E., McKenna, T. St.-C., Chinsangaram, J. and Grubman, M. J. (1997)Evaluation of a live-attenuated foot-and-mouth disease virus as a vaccine candidate, Virology 227,96-102.

105 Doel, T. R. (1999) Optimization of the immune response to foot-and-mouth disease vaccines, Vaccine17, 1767-1771.

106 Sela, M. and Arnon, R. (1984) Synthetic antigens and vaccines, Interdisc. Sci. Rev. 9, 271-282.107 Pfaff, E., Mussgay, M., Böhm, H. O., Schulz, G. E. and Schaller, H. (1982) Antibodies against a pre-

selected peptide recognize and neutralize foot-and-mouth disease virus, EMBO J. 1, 869-874.108 Geysen, H. M., Meloen, R. H. and Barteling, S. J. (1984) Use of peptide synthesis to probe viral

antigens for epitopes to a resolution of a single amino acid, Proc. Natl. Acad. Sci. USA 81, 3998-4002.

109 Francis, M. J., Fry, C. M., Rowlands, D. J., Brown, F., Bittle, J. L., Houghten, R. A. and Lerner, R. A.(1985) Immunological priming with synthetic peptides of foot-and-mouth disease virus, J. Gen. Virol.66, 2347-2354.

110 Geysen, H. M., Meloen, R. H. and Barteling, S. J. (1985) Small peptides induce antibodies with asequence and structural requirement for binding antigen comparable to antibodies raised against thenative protein, Proc. Natl. Acad. Sci. USA 82, 178-182.

111 Geysen, H. M., Rodda, S. J. and Mason, T. J. (1986) A priori delineation of a peptide which mimics adiscontinuous antigenic determinant, Molec. Immunol. 23, 709-715.

112 Meloen, R. H. and Barteling, S. J. (1986) Epitope mapping of the outer structural protein VP1 of threedifferent serotypes of foot-and-mouth disease virus, Virology 149, 55-63.

113 Meloen, R. H., Puyk, W. C., Meijer, D. J. A., Lankhof, H., Posthumus, W. P. A. and Shaaper, W. M.M. (1987) Antigenicity and immunogenicity of synthetic peptides of foot-and-mouth disease virus, J.Gen. Virol. 68, 305-312.

114 Francis, M. J., Hastings, G. Z., Clarke, B. E., Brown, A. L., Bedell, C. R., Rowlands, D. J. and Brown,F. (1990) Neutralizing antibodies to all seven serotypes of foot-and-mouth disease virus elicited bysynthetic peptides, Immunology 69, 171-176.


42

115 Siligardi, G., Drake, A. F., Mascagni, P., Rowlands, D. J., Brown, F. and Gibbons, W. A. (1991) A CDstrategy for the study of polypeptide folding/unfolding: a synthetic foot-and-mouth disease virusimmunogenic peptide, Int. J. Peptide Protein Res. 38, 519-527.

116 DiMarchi, R., Brooke, G., Gale, C., Cracknell, V., Doel, T. and Mowat, N. (1986) Protection of cattleagainst foot-and-mouth disease by a synthetic peptide, Science 232, 639-641.

117 Doel, T. R., Gale, C., Brooke, G. and DiMarchi, R. (1988) Immunization against foot-and-mouthdisease with synthetic peptides representing the C-terminal region of VP1, J. Gen. Virol. 69, 2403-2406.

118 Steward, M. W., Stanley, C. M., DiMarchi, R., Mulcahy, G. and Doel, T. R. (1991) High-affinityantibody induced by immunization with a synthetic peptide is associated with protection of cattleagainst foot-and-mouth disease, Immunology 72, 99-103.

119 Murdin, A. D. and Doel, T. R. (1987) Synthetic peptide vaccines against foot-and-mouth disease. I.Duration of the immune response and priming in guinea pigs, rabbits and mice, J. Biol. Standardiz.15, 39-51.

120 Murdin, A. D. and Doel, T. R. (1987) Synthetic peptide vaccines against foot-and-mouth disease. II.Comparison of the response of guinea pigs, rabbits and mice to various formulations, J. Biol.Standardiz. 15, 58-65.

121 Francis, M. J., Fry, C. M., Rowlands, D. J. and Brown, F. (1988) Qualitative and quantitativedifferences in the immune response to foot-and-mouth disease virus antigens and synthetic peptides,J. Gen. Virol. 69, 2483-2491.

122 Mulcahy, G., Gale, C., Robertson, P., Iysan, S. DiMarchi, R. and Doel, T. R. (1990) Isotype responsesof infected, virus-vaccinated and peptide-vaccinated cattle to foot-and-mouth disease virus, Vaccine 8,249-256.

123 Taboga, O., Tami, C., Carrillo, E., Núñez, J. I., Rodríguez, A., Sáiz, J. C., Blanco, E., Valero, M. L.,Roig, X., Camarero, J. A., Andreu, D., Mateu, M. G., Giralt, E., Domingo, E., Sobrino, F. and Palma,E. L. (1997) A large-scale evaluation of peptide vaccines against foot-and-mouth disease: lack of solidprotection in cattle and isolation of escape mutants.

124 Francis, M. J., Hastings, G. Z., Brown, F., McDermed, J., Lu, Y. A. and Tam, J. P. (1991)Immunological evaluation of the multiple antigen peptide (MAP) system using the major immunogenicsite of foot-and-mouth disease virus, Immunology 73, 249-254.

125 Feliu, J. X. and Villaverde, A. (1998) Engineering of solvent-exposed loops in Escherichia coli β-galactosidase, FEBS Lett. 434, 23-27.

126 Carbonell, X., Feliu, J. X., Benito, A. and Villaverde, A. (1998) Display-induced antigenic variation inrecombinant peptides, Biochem. Biophys. Res. Comm. 248, 773-777.

127 Sanz-Parra, A., Vázquez, B., Sobrino, F., Cox, S. J., Ley, V. and Salt, J. S. (1999) Evidence of partialprotection against foot-and-mouth disease in cattle immunized with a recombinant adenovirus vectorexpressing the percursor polypeptide (P1) of foot-and-mouth disease virus capsid proteins, J. Gen.Virol. 80, 671-679.

128 Wigdorovitz, A., Carrillo, C., dos Santos, M. J., Trono, K., Peralta, A., Gómez, M. C., Ríos, R. D.,Franzone, P. M., Sadir, A. M., Escribano, J. M. and Borca, M. V. (1999) Induction of a protectiveantibody response to foot-and-mouth disease virus in mice following oral or parenteral immunizationwith alfalfa transgenic plants expressing the viral structural protein VP1, Virology 255, 347-353.

129 Sanz-Parra, A., Jímenez-Clavero, M. A., García-Briones, M. M., Blanco, E., Sobrino, F. and Ley, V.(1999) Recombinant viruses expressing the foot-and-mouth disease virus capsid precursor polypeptide(P1) induce cellular but not humoral antiviral immunity and partial protection in pigs, Virology 259,129-134.

130 Wigdorovitz, A., Pérez-Filguera, D. M., Robertson, N., Carrillo, C., Sadir, A. M., Morris, T. J. andBorca, M. V. (1999) Protection of mice against challenge with foot-and-mouth disease virus (FMDV)by immunization with foliar extracts from plants infected with recombinant tobacco mosaic virusexpressing the FMDV structural protein VP1, Virology 264, 85-91.

131 Schaaper, W. M. M., Lankhof, H., Puijk, W. C. and Meloen, R. H. (1989) Manipulation of anti-peptide immune response by varying the coupling of the peptide with the carrier protein, Molec.Immunol. 26, 81-85.

132 Mateu, M. G., Andreu, D., Carreño, C., Roig, X., Cairó, J. J., Camarero, J. A., Giralt, E. andDomingo, E. (1992) Non-additive effects of multiple amino acid substitutions on antigen-antibodyrecognition, Eur. J. Immunol. 22, 1385-1389.

133 Carreño, C., Roig, X., Camarero, J. A., Cairó, J. J., Mateu, M. G., Domingo, E., Giralt, E. andAndreu, D. (1992) Studies on antigenic variability of C-strains of foot-and-mouth disease virus bymeans of synthetic peptides and monoclonal antibodies, Int. J. Peptide Protein Res. 39, 41-47.

134 Valero, M. L. “Mimetización estructural e inmunogénica del sitio antigénico principal del virus de lafiebre aftosa” (Ph. D. Thesis), Department of Organic Chemistry – University of Barcelona: 1997.

135 Briand, J. P., Benkirane, N., Guichard, G., Newman, J. F. E., Van Regenmortel, M. H. V., Brown, F.and Müller, S. (1997) A retro-inverso peptide corresponding to the GH loop of foot-and-mouthdisease virus elicits high levels of long-lasting protective neutralizing antibodies, Proc. Natl. Acad. Sci.USA 94, 12545-12550.


43

136 Müller, S., Benkirane, N., Guichard, G., Van Regenmortel, M. H. V. and Brown, F. (1998) Thepotential of retro-inverso peptides as synthetic vaccines, Exp. Opin. Invest. Drugs 7, 1429-1438.

137 Petit. M. C., Benkirane, N., Guichard, G., Du, A. P. C., Marraud, M., Cung, M. T., Briand, J. P. andMüller, S. (1999) Solution structure of a retro-inverso peptide analogue mimicking the foot-and-mouthdisease virus major antigenic site, J. Biol. Chem. 274, 3686-3692.

138 Nargi, F., Kramer, E., Mezencio, J., Zamparo, J., Whetstone, C., Van Regenmortel, M. H. V., Briand,J. P., Müller, S. and Brown, F. (1999) Protection of swine from foot-and-mouth disease with one doseof an all-D retro peptide, Vaccine 17, 2888-2893.


44

1. SPR screening of synthetic

peptides from the GH loop of FMDV


50

SPR screening of synthetic peptides from the GH loop of FMDV

1.0 Introduction

The first objective of the present work was the study of the applicability of SPR biosensors1 to

kinetically characterise the interactions between peptides related to viral antigenic sites and relevant

monoclonal antibodies2. In particular, the research was focused on the interactions between anti-

FMDV mAbs and synthetic peptides reproducing an immunodominant region of FMDV (antigenic

site A, residues 136-150 of envelope protein VP1, isolate C-S8c1)3-5 to examine the main structural

features involved in the recognition of this site by neutralising antibodies. Synthetic peptides

reproducing different mutations at this site are particularly useful in identifying residues involved in

recognition or escape events6,7. Given the large number of such peptides and the relatively small

number of relevant mAbs, the most productive approach would seem to be mAb immobilisation and

analysis of the peptides as soluble analytes. However, a limitation of the SPR technique is that

interactions between low molecular weight (<5 kDaA) analytes and their immobilised binding

partners cannot, in principle, be studied directly since the increase in mass on the sensor chip is too

small to provide reliable data8. Not only small responses are a problem, but also bulk refractive

index effects together with non-specific

binding and mass-transport limitations can

affect true binding kinetics, particularly in

the direct detection of small analytes. A

possible way to circumvent detection

problems associated with small analytes is

to use a competitive kinetic analysis with a

high molecular weight analyte for the same

ligand binding site9. However, this

approach was not initially feasible, since a

high molecular weight representative of

antigenic site A, e. g., capsid protein VP1 of

FMDV, was unavailable. In view of this, it was

the difficulties associated with the small size (≈1

A 1:1 bimolecular interaction kinetics is to b

antigen-binding fragments are considered inde

A The experimental work described in the present thesis hwork was completed, improved versions of the BIAcore inhave been commercialised.

Fab fragment

Fab fragment

antigen antigen

kakd kd

ka

Fc fragment

Figure 1. 1 Representation of an antigen – antibodyinteraction; Fab stands for antigen-binding fragment andFc for crystallisable fragment.

51

decided to work with immobilised mAb and address

.6 kDa) of the peptide analytes.

e expected for peptide-antibody interaction, if both

pendent and equivalent (Fig. 1.1)10.

as been carried out using a BIACORE 1000 instrument. Since thisstrumentation with higher sensitivity (BIACORE 2000, 3000, X...)


52

1.1 Optimisation of the experimental set-up

Antigenic site A of FMDV C-S8c1 contains several distinct, overlapping, B-cell epitopes and is

located in the GH loop (residues 136 to 150) of the envelope protein VP12-7,11. It can be reproduced

by peptide A15B, corresponding to the sequence:

136YTASARGDLAHLTT150T

A few sets of experiments, using A15 as analyte, were run on mAb SD6C surfaces with different

densities (8, 1.7 and 0.8 ng/mm2), using peptide concentrations between 1 and 2440 nM and two

different buffer flow rates (5 and 60 µl/min). mAb immobilisation and peptide injection procedures

are described in section 4.3.1.1.

1.1.1 High mAb density

In a first approach, a very high mAb surface density (8 ng/mm2) and high A15 concentrations were

employed in an attempt to overcome the low responses that were to be expected from the small size

of the analyte. Peptide injections were carried out at 5 µl/min, using the kinject mode to avoid

sample dispersion at injection plugs, and association and dissociation times were of 7 and 6

minutes, respectively. The surface was regenerated, at the end of each cycle, by a 3-min pulse of

100 mM HCl. The sensorgrams generated under these conditions (Fig. 1.2 A) could not be fitted to

the expected 1:1 bimolecular interaction kinetics, as inferred from the high and non-random

residuals observed in the dissociation phase (Fig. 1.2 B), and from the concentration-dependence of

the association rate constant, ka (Fig. 1.2 C).

B Peptides used for optimisation and validation of the SPR experimental set-up were kindly given by Dr. Mari-Luz Valero(Dept. Q.O. - U. B., Barcelona).C Monoclonal antibody SD6 is a site-A directed neutralising mAb raised against FMDV C-S8c1; it was kindly supplied by Dr.Nuria Verdaguer and Wendy F. Ochoa (IBMB/CSIC – Barcelona).


53

C

0.00E+00

2.00E+05

4.00E+05

6.00E+05

8.00E+05

1.00E+06

1.20E+06

1.40E+06

1.60E+06

0 200 400 600 800 1000 1200

Peptide concentration/nM

k a/M

-1s-1

s

Figure 1. 2 First approach to the SPR kinetic analysis of the interaction between immobilised mAb SD6 andpeptide analyte A15: A. Experimental sensorgrams; B. Distribution of residual data points for the dissociationphase corresponding to [A15]=310 nM (detailed view of the experimental and modelled curves); C. Variationof ka with peptide concentration.

The shape of the dissociation curves suggested that some analyte rebinding to the surface was

occurring, affecting true kinetics. At the same time, the apparent association rate constant decreased

with increasing analyte concentration, i. e., the more peptide was injected, the more difficult became

its binding to mAb molecules. So, it appeared that antibody molecules were heterogeneously

distributed in the dextran matrix, with different accessibility levels12. Upon analyte injection, the first

peptide molecules would occupy the most accessible mAb receptors and the following ones would

have increasing difficulty in reaching free mAb binding sites, such effect becoming larger with higher

peptide concentrations. This seemed to be confirmed by the better fit obtained when a

heterogeneous ligand kinetic model was employed to fit the experimental data (data not shown).

This model, however, considers only two different types of ligand, which is most probably far from

reality. Since lowering peptide concentration (1 to 50 nM) did not improve the results (data not

shown), new conditions were searched in order to obtain experimental data consistent with a

langmuirian kinetic behaviour (1:1 bimolecular interaction).


54

1.1.2 Medium mAb density

The poor results obtained in the previous section were symptomatic of significant heterogeneity in

ligand accessibility and orientation. Also, diffusion-controlled delivery of analyte to the most

hindered SD6 molecules would be a further cause for the observed deviations. Therefore, a second

SD6 surface was prepared with much lower density (1.7 ng/mm2) and another set of injections was

run at the same flow rate, spanning peptide concentrations from 1 to 2440 nM. In this case, peptide

concentrations below 70 nM were too low for a clear response to be observed, since sensorgrams

were hardly distinguished from mere bulk refractive index effects. Higher peptide concentrations led

to results better than those described in section 1.1.1, but still presenting some degree of data

inconsistency (not shown).

1.1.3 Low mAb density

Further lowering of mAb surface density (to 0.8 ng/mm2), in an attempt to eliminate the non-ideal

effects observed so far, did not work either. In fact, this density was seen to be too low for the

detection of the FMDV peptides injected, even at analyte concentrations as high as 2.44 µM (not

shown).

1.1.4 High buffer flow rate

Since the previous results, all of them obtained at 5 µl/min flow, persistently deviated from the

expected behaviour at different mAb surface densities and peptide concentrations, the flow rate

seemed an important parameter to manipulate in order to optimise the SPR analysis. Low buffer

flow rate could be favouring diffusion-controlled kinetics, affecting true binding constants12-14.

Therefore, a fourth set of SPR experiments was run, this time using the medium density SD6 surface

(1.7 ng/mm2) and high A15 concentrations (152 to 2440 nM), and raising the buffer flow rate to 60

µl/min. Both association and dissociation times were decreased (90 and 240 seconds, respectively)

to diminish sample and buffer consumption. Under these experimental conditions, consistent and

apparently reliable data were obtained. Experimental and modelled curves were virtually

superimposable (Fig. 1.3 A) with a random distribution of residuals within an interval of ca. ±0.4 RU

(Fig. 1.3 B).

Linearity of ks versus peptide concentration over the 32-fold concentration range was observed, as

required for a concentration-independent ka (Fig.1.3 C). The chi-squared (χ2) value was smaller than

0.1 and data self-consistency15 was further confirmed by the total agreement between the values for

the equilibrium association constant, KA, obtained from either the ka/kd ratio or from the plot of Req

versus peptide concentration (Fig. 1.3 D)D.

D The theoretical basis for SPR kinetic analysis is exposed in section 0.3.


55

Data analysis produced good quality fits, reproducing the same rate and affinity constants

independently from fitting curves globally, locally or with separate association/dissociation phases to

the langmuirian kinetics model (Table 1.1).

A

-10

-5

0

5

10

15

20

25

30

35

40

-10 40 90 140 190Time / s

Res

po

nse

/ R

U

152 nM 305 nM

610 nM 1220 nM

2440 nM 152 nM, calc

305 nM, calc 610 nM, calc

1220 nM, calc 2440 nM, calc

C

r2 = 0.9992

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.00E+00 5.00E-07 1.00E-06 1.50E-06 2.00E-06 2.50E-06

Peptide concentration / M

ks /

s-1

ks (1/s) vs. Conc of analyte

Linear (ks (1/s) vs. Conc ofanalyte)

D

14

15

16

17

18

19

20

0.00E+00 5.00E-07 1.00E-06 1.50E-06 2.00E-06 2.50E-06

Peptide concentration / M

Res

po

nse

at

equ

ilib

riu

m (

Req

) / R

U

Req (RU) vs. Conc ofanalyte

B

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

0 50 100 150 200 250

Res

idu

als

/ RU

152 nM305 nM

610 nM1220 nM

2440 nM

Figure 1. 3 Results obtained in the SPR kinetic analysis of the interaction between immobilised mAbSD6 and soluble peptide A15: A. Sensorgrams (experimental and modelled); B. Distribution of residualdata points; C. Plot of locally fitted ks (apparent rate constant, see section 0.3) versus peptideconcentration; D. Plot of Req (response at equilibrium) vs. peptide concentration (see section 0.3).


56

Table 1.1 Quantitative data on the 1:1 langmuirian interaction between immobilised mAb SD6 andsoluble peptide A15.

Curve fitting [peptide]/ nM ka/M-1s-1 kd/s-1 KA/M-1

Global - 6.2×104 2.6×10-3 2.3×107

152 6.0×104 2.4×10-3 2.5×107

305 5.8×104 2.6×10-3 2.3×107

Local, simultaneous ka/kd 610 5.9×104 2.6×10-3 2.3×107

1220 6.1×104 2.7×10-3 2.3×107

2440 6.2×104 2.9×10-3 2.1×107

152 6.7×104 2.4×10-3 2.8×107

305 6.2×104 2.6×10-3 2.4×107

Local, separate ka/kd 610 5.5×104 2.6×10-3 2.1×107

1220 5.8×104 2.8×10-3 2.1×107

2440 5.9×104 2.7×10-3 2.2×107

1.2 Application to the systematic screening of FMDV peptides

1.2.1 Screening of 44 FMDV peptides as antigens towards mAb SD6

Having found suitable experimental conditions for the kinetic analysis of the A15/SD6 interaction, a

similar protocol was applied to the systematic screening of 43 other A15 analogues. The antigenicity

of these peptides had been previously characterised by competition ELISA6, which made them

excellent models to evaluate the reliability of our SPR optimised analysis conditions.

An additional peptide, A15scr, with the same constitutive amino acids as A15 but randomly ordered

(RAGTATTLADLHYST), was used as a negative control. The scrambled sequence A15scr had no

apparent specific binding, but gave rise to a substantial bulk refraction index response (Fig. 1.4 A),

as observed for all other peptides analysed. Therefore, the curves for each site A peptide were

corrected by subtraction of the corresponding A15scr sensorgrams (Fig. 1.4 B and C). The

consistency and accuracy of the fitted kinetic data for the whole set of A15 analogues were in every

aspect similar to those described for A15 under the same conditions (Fig. 1.4 D, E and F). The

stability of the SD6 surface to the repeated strong acid regeneration cycles allowed the screening of

the entire set over the same surface without any detectable loss in mAb activity, thus providing

reliable comparison among the different peptides. This is a clear advantage of the present SPR

configuration, since in the alternative immobilisation of peptides one cannot control the similarity of

the different peptide surfaces. The constants obtained for the interaction between mAb SD6 and the

44 peptides screened are shown in Table 1.2.


57

A

-10

0

10

20

30

40

50

60

-10 40 90 140 190 240Time/s

Res

po

nse

/RU

157 nM

314 nM627 nM1254 nM2509 nM

B

-10

0

10

20

30

40

50

60

70

80

-10 40 90 140 190 240Time/s

Res

po

nse

/RU

163 nM326 nM652 nM1305 nM

2610 nM

C

-20

-15

-10

-5

0

5

10

15

20

25

-10 40 90 140 190 240Time/s

Res

po

nse

/RU

163 nM326 nM652 nM

1305 nM2610 nM

E

r2 = 0.9991

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

0.00E+00 5.00E-07 1.00E-06 1.50E-06 2.00E-06 2.50E-06

Peptide concentration/M

ks/

s-1

F

15

15.5

16

16.5

17

17.5

18

18.5

19

19.5

0 0.0000005 0.000001 0.0000015 0.000002 0.0000025


Req

/RU

D

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

0 50 100 150 200 250

Time/s

Res

idu

als/

RU

Figure 1. 4 A. Sensorgrams for the A15scr/SD6 interaction; B. Sensorgrams for the interaction betweenSD6 and an FMDV peptide: A15 (140P); C. Sensorgrams of the same interaction as in B, after correctionupon subtraction of sensorgrams shown in A; D. Residual distribution after fitting sensorgrams C to the 1:1bimolecular interaction kinetics model; E. Linear plot of locally fitted ks versus peptide concentration; F. Plotof locally fitted Req versus peptide concentration (for comparison between the association equilibriumconstants as obtained from this plot or from the ka/kd ratio).


58

Table 1.2 Kinetic dataa of the interactions between mAb SD6 and 44 site A peptidesb.PEPTIDE ka/M-1s-1 kd/s-1 KA/M-1 ELISA PEPTIDE ka/M-1s-1 kd/s-1 KA/M-1 ELISA

A15 7.3×104 1.4×10-3 5.4×107 A15(148S) 9.4×104 6.7×10-3 1.4×107

A15(137P) 5.8×104 1.9×10-3 3.1×107 A15(138D) ni ni niA15(138P) ni ni ni A15(138E) ni ni niA15(139P) ni ni ni A15(138F) 8.6×104 3.9×10-3 2.2×107

A15(140P) 7.4×104 2.1×10-3 3.5×107 A15(138K) ni ni niA15(141P) ni ni ni A15(138R) ni ni niA15(142P) ni ni ni

A15(138V)1.3×105 7.1×10-3 1.8×107

A15(143P) ni ni ni A15(138Y) 2.3×105 8.8×10-3 2.6×107

A15(144P) ni ni ni A15(145D) ni ni niA15(145P) ni ni ni A15(145E) 5.3×104 4.1×10-3 1.3×107

A15(146P) ni ni ni A15(145F) ni ni niA15(147P) ni ni ni A15(145I) ni ni niA15(148P) 6.6×104 6.2×10-2 1.1×107 A15(145K) 4.5×104 4.1×10-3 1.1×107

A15(137S) 1.2×105 3.5×10-3 3.5×107 A15(145R) 1.4×104 9.7×10-3 1.4×106

A15(138S) 1.1×105 1.2×10-2 8.8×106 A15(147A) 9.2×104 2.1×10-3 4.4×107

A15(140S) 1.5×105 8.6×10-4 1.8×108 A15(147D) ni ni niA15(141S) 4.5×104 2.6×10-3 1.8×107 A15(147E) 6.7×104 3.1×10-3 2.2×107

A15(142S) 6.1×104 5.6×10-3 1.1×107 A15(147G) 3.2×105 6.1×10-3 3.7×107

A15(143S) ni ni ni A15(147K) 6.6×104 3.4×10-3 2.0×107

A15(144S) ni ni ni A15(147N) 4.6×104 5.0×10-3 9.2×106

A15(145S) 4.0×104 5.1×10-3 7.8×106 A15(147R) 7.9×104 3.5×10-3 2.3×107

A15(147S) 1.3×104 1.1×10-2 1.2×106 A15(147V) 9.2×104 6.6×10-3 1.4×107

a corrected for non-specific binding; b qualitative relative antigenicities from ELISA competition assays are represented, with a black box corresponding to IC50>100, a dark grey box toIC50 = 30 to 100, a light grey box to IC50 = 5 to 30 and a white box to IC50<5. “ni” - no measurable interaction.


59

Table 1.2 also includes previous data from enzyme-linked immunosorbent assays (ELISA)6. These

data had been expressed as IC50 values (competitor peptide concentration giving a 50 % drop in

maximal absorbance), normalised to the IC50 of peptide A15 (see section 4.3.2). A general

agreement between both SPR and ELISA techniques was observed, thus supporting the reliability of

the functional characterisation of small antigenic FMDV peptides using SPR.

1.2.2 Reproducibility in the SPR constants measured on the SD6 surface

Although systematic repetition of assays for every analyte was not possible, given the large number

of peptides, the reproducibility of our SPR analysis was nevertheless assessed by repeated injection

of a representative sub-set of peptides. Six A15 analogues were independently analysed six times

under similar conditions, with the results shown in Table 1.3. Standard deviations of the measured

kinetic parameters oscillate between 2 and 11% of the mean value, which is quite good considering

the small size of the analytes. A sole exception was seen with kd for the SD6/A15(137I) complex

(SD=20%), which is not surprising given the very low dissociation rate observed for this complex,

making it more prone to be affected by experimental error.

Table 1.3 Reproducibility in kinetic SPR analyses of SD6/peptide interactions (six assays per peptide).

Peptide ka/M-1s-1 kd/s

-1 Peptide ka/M-1s-1 kd/s

-1

1.0×105 6.5×10-4 2.88×104 (*) 3.74×10-3 (*)

9.7×104 4.7×10-4 4.69×104 4.13×10-3

A15 (137I) 9.1×104 4.3×10-4 A15 (145E) 5.76×104 3.63×10-3

9.6×104 5.1×10-4 4.82×104 4.23×10-3

8.7×104 (*) 8.0×10-5 (*) 6.23×104 4.28×10-3

9.3×105 (*) 2.7×10-4 (*) 5.21×104 4.15×10-3

mean±SD (9.6±±±±0.4)××××104 (5±±±±1)××××10-4 mean±SD (5.3±±±±0.6)××××104 (4.1±±±±0.3)××××10-3

n.i. n.i. 2.87×104 (*) 4.67×10-3 (*)

n.i. n.i. 4.15×104 4.55×10-3

A15 (138K) n.i. n.i. A15 (145K) 4.86×104 4.00×10-3

n.i. n.i. 4.40×104 3.90×10-3

n.i. n.i. 4.52×104 4.07×10-3

n.i. n.i. 2.45×104 3.99×10-3

mean±SD - - mean±SD (4.5±±±±0.3)××××104 (4.1±±±±0.3)××××10-3

1.22×105 1.90×10-3 2.09×104 (*) 9.65×10-3 (*)

9.33×104 (*) 1.98×10-3 (*) 3.05×104 (*) 5.40×10-3 (*)

A15(148I) 1.19×105 2.43×10-3 A15(145R) 1.44×104 9.01×10-3

7.24×104 2.44×10-3 1.38×104 9.46×10-3

1.00×105 1.88×10-3 1.37×104 1.03×10-3

1.13×105 2.17×10-3 1.37×104 1.01×10-3

mean±SD (1.1±±±±0.1)××××105 (2.1±±±±0.2)××××10-3 mean±SD (1.39±±±±0.03)××××104 (9.7±±±±0.6)××××10-3

(*) data was not considered for calculating mean and standard deviation values.


60

1.3 Use of other site A-directed monoclonal antibodies

A desirable general applicability of our direct kinetic SPR antigenic analysis of small site A peptides

would obviously require not only the ability to distinguish between different analytes (antigens) but

also between different receptors (antibodies). Therefore, it was decided to adapt the procedure to a

new mAb, 4C4E, as the immobilised receptor.

1.3.1 Adaptation of the experimental set-up to a new mAb

In order to obtain good quality data with mAb 4C4, slight changes had to be introduced in the

protocols of SPR analysis previously described for mAb SD6. MAb 4C4 coupled more efficiently to

the dextran matrix under the same conditions employed for the immobilisation of SD6:

immobilisation levels had to be therefore adjusted by dilution of the mAb solution, to achieve a final

surface densities of ca. 1600 RU (1.6 ng/mm2).

Also, it was observed that 4C4 surfaces were not suitably regenerated with hydrochloric acid. A clear

symptom for this problem was that sensorgrams from the same 4C4 surface showed an increase in

baseline response and a concomitant decrease in signal for identical A15 concentrations over

repetitive cycles (not shown). Alternative regeneration procedures, using other acids (phosphoric or

formic) or bases (10 mM glycine, pH 12 or 10 mM sodium hydroxide) were tested and sodium

hydroxide was found to be the most efficient regenerating agent.

Further, while for SD6 the optimal analyte concentration range was generally between ca. 75 and

1250 nM, for 4C4 saturation was already reached at concentrations above 600 nM (not shown).

This observation suggested that mAb 4C4 possessed higher affinity than SD6 towards the site A

peptides, which was later confirmed upon peptide analysis on 4C4 surfaces (see following section).

1.3.2 Screening of 44 FMDV peptides as antigens towards mAb 4C4

A systematic screening similar to that described in section 1.2.1 was performed on a 4C4 surface

(Fig. 1.5). Surface mAb density and injection parameters were quite the same, with the only

difference being the peptide concentrations used (from 35 to 1250 nM). The kinetic data were fitted

as before, generally displaying identical accuracy and consistency levels. Once more, the global

agreement between SPR-derived affinities and previous ELISA data was remarkable and further

validated the experimental SPR set-up (Table 1.4).

E mAb 4C4 is a site-A directed, anti-FMDV neutralising antibody, raised against strain C1-Brescia; it was kindly supplied byDr. Nuria Verdaguer and Ms. Wendy F. Ochoa (IBMB/CSIC – Barcelona, Spain).


61

B

-5

15

35

55

75

95

115

-10 40 90 140 190 240Time/s

Res

po

nse

/RU

35 nM

70 nM

140 nM

280 nM

560 nM

1120 nM

C

-5

5

15

25

35

45

55

65

-10 40 90 140 190 240Time/s

Res

po

nse

/RU

35 nM70 nM140 nM280 nM560 nM1120 nM

D

-5

5

15

25

35

45

55

65

-10 40 90 140 190 240

Time/sR

es

po

ns

e/R

U

35 nM (sim) 70 nM (sim)

140 nM (sim) 280 nM (sim)

560 nM (sim) 1120 nM (sim)

35 nM 70 nM

140 nM 280 nM

560 nM 1120 nM

A

-5

5

15

25

35

45

55

65

-10 40 90 140 190 240

Time/s

Res

po

nse

/RU

39 nM78 nM157 nM314 nM627 nM1254 nM

Figure 1. 5 A. Binding curves for non-specific peptide A15scr on a mAb 4C4 surface. The remaining plots are sensorgrams forbinding of peptide A15(142S) to mAb 4C4: B. Raw data; C. After correction for non-specific binding; D. Overlay plot ofexperimental (corrected) and simulated (sim) sensorgrams. [Note: higher total (Rtot) responses correspond to higher peptide concentrations,except for 280 nM peptide injection (second smaller response) which presented a lower bulk RI jump].


62

Table 1.4 Kinetic dataa of the interactions between mAb 4C4 and 44 site A peptidesb.PEPTIDE ka/M-1s-1 kd/s-1 KA/M-1 ELISA PEPTIDE ka/M-1s-1 kd/s-1 KA/M-1 ELISA

A15 3.8×105 1.9×10-3 1.9×108 A15(148S) 3.0×105 7.9×10-3 3.8×107

A15(137P) 1.2×105 6.1×10-4 2.0×108 A15(138D) n.i. n.i. n.i.A15(138P) 4.1×105 5.1×10-2 8.0×106 A15(138E) n.i. n.i. n.i.A15(139P) n.i. n.i. n.i. A15(138F) 5.5×105 5.7×10-3 9.8×107

A15(140P) 1.9×105 1.9×10-3 1.0×108 A15(138K) 3.5×105 2.3×10-2 1.5×107

A15(141P) n.i. n.i. n.i. A15(138R) 1.4×105 1.9×10-2 7.4×106

A15(142P) n.i. n.i. n.i. A15(138V) 2.7×105 1.4×10-3 2.0×108

A15(143P) n.i. n.i. n.i. A15(138Y) 4.0×105 1.3×10-3 3.0×108

A15(144P) n.i. n.i. n.i. A15(145D) n.i. n.i. n.i.A15(145P) n.i. n.i. n.i. A15(145E) 1.5×105 2.2×10-3 6.9×107

A15(146P) n.i. n.i. n.i. A15(145F) 1.4×105 5.1×10-3 2.7×107

A15(147P) n.i. n.i. n.i. A15(145I) 1.7×105 6.4×10-3 2.7×107

A15(148P) 1.6×105 1.4×10-2 1.2×107 A15(145K) 2.5×105 2.3×10-3 1.1×108

A15(137S) 1.7×105 3.0×10-3 5.6×107 A15(145R) 6.6×104 5.9×10-3 1.1×107

A15(138S) 2.5×105 2.2×10-3 1.1×108 A15(147A) n.i. n.i. n.i.A15(140S) 2.5×105 2.4×10-3 1.1×108 A15(147D) n.i. n.i. n.i.A15(141S) 1.2×105 3.5×10-3 3.2×107 A15(147E) n.i. n.i. n.i.A15(142S) 6.3×104 3.8×10-3 1.6×107 A15(147G) n.i. n.i. n.i.A15(143S) n.i. n.i. n.i. A15(147K) n.i. n.i. n.i.A15(144S) n.i. n.i. n.i. A15(147N) n.i. n.i. n.i.A15(145S) 3.8×105 7.6×10-3 5.1×107 A15(147R) n.i. n.i. n.i.A15(147S) n.i. n.i. n.i. A15(147V) 1.4×105 5.6×10-2 2.5×106

a corrected for non-specific binding; b qualitative relative antigenicities from ELISA competition assays are represented, with a black box corresponding to IC50>100, a dark grey box to IC50

= 30 to 100, a light grey box to IC50 = 5 to 30 and a white box to IC50<5. “ni” - no measurable interaction.


63

1.3.3 Reproducibility in the constants measured on the 4C4 surface

As already described in section 1.2.1, reproducibility of the SPR-measured constants was evaluated

through repetitive analyses of a small set of site A peptides. Again, six peptides were independently

analysed six times on a mAb 4C4 surface. Results for mAb 4C4 are presented in Table 1.5 and

show very good reproducibility, with standard deviations less than 9% of the mean values.

Table 1.5 Reproducibility in kinetic SPR analysis of 4C4/peptide interactions (six assays per peptide).


-1 ka/M-1s-1 kd/s

-1

n.i. n.i. 1.06×105 (*) 2.16×10-2 (*)n.i. n.i. 1.46×105 1.94×10-2

A15(138D) n.i. n.i. A15(138R) 1.45×105 1.88×10-2

n.i. n.i. 1.51×105 2.08×10-2

n.i. n.i. 1.42×105 1.94×10-2

n.i. n.i. 1.39×105 1.88×10-2

mean±SD _ _ mean±SD (1.45±±±±0.05)××××105 (1.94±±±±0.07)××××10-2

5.54×105 5.06×10-3 2.56×105 6.59×10-4

6.00×105 6.23×10-3 2.90×105 1.34×10-3

A15(138F) 4.70×105 5.44×10-3 A15(138V) 2.49×105 1.15×10-3

5.68×105 5.55×10-3 2.67×105 1.34×10-3

5.70×105 6.04×10-3 2.72×105 1.60×10-3

4.90×105 (*) 1.16×10-3 (*) 2.83×105 1.62×10-3

mean±SD (5.5±±±±0.5)××××105 (5.7±±±±0.5)××××10-3 mean±SD (2.7±±±±0.2) ×105 (1.4±±±±0.2)××××10-3

3.05×105 2.35×10-2 4.46×105 (*) 1.41×10-3 (*)

3.62×105 2.33×10-2 3.99×105 1.32×10-3

A15(138K) 3.74×105 2.33×10-2 A15(138Y) 4.05×105 1.12×10-3

3.78×105 2.40×10-2 3.82×105 1.41×10-3

3.38×105 2.25×10-2 3.80×105 1.48×10-3

4.65×105 (*) 2.25×10-2 (*) 3.82×105 1.34×10-3

mean±SD (3.5±±±±0.3)××××105 (2.32±±±±0.06)××××10-2 mean±SD (4.0±±±±0.3)××××105 (1.3±±±±0.1)××××10-3

(*) data was not considered for calculating mean and standard deviation values.


64

1.4 Probing subtle differences in peptide and mAb behaviour by SPR

Comparison of the results in sections 1.2 and 1.3 leads to the immediate conclusion that not only

the different features of the peptides analysed can be distinguished through SPR, but also distinct

mAb “personalities” can be appreciated.

Peptides screened on the same mAb surface are mainly distinguished by their dissociation rate

constants. A closer look into Tables 1.2 or 1.4 shows that ka varies over a 10-fold range, while kd

varies over a 100-fold range. This observation has already been reported16,17 and it has been

proposed that the biologically relevant SPR-derived parameter is, in fact, kd, since it is a measure of

the life-time of the ligand-receptor complex. Correlations between dissociation rate constants and

neutralisation have also been found18.

On the other hand, comparison of data from the two mAb surfaces seems to suggest that each

antibody has its own “avidity range”, i. e., its own range of association rate constants, which provide

a measure of the accessibility of a particular paratope towards similar antigens. Thus, while average

ka for SD6 is 7.9×104 M-1s-1, for mAb 4C4 it is 2.4×105 M-1s-1, a three-fold increase.

The validity of the SPR approach for the screening of antigenic site A peptides is better illustrated in

Fig. 1.6, which shows the good correlation between antigenicity data measured with SPR and

previous results from competition ELISA. Also, the distinct recognition requirements imposed by

different antibodies is clearly demonstrated upon comparison of Figs. 1.6 A and 1.6 B, particularly

in what concerns recognition of A15 analogues displaying mutations at position 147 (corresponding

to a leucine in the native sequence). This provides further proof of the suitability of SPR to the

functional study of antigenic determinants in viral epitopes using synthetic peptides.

1.5 Validity of the experimental kinetic constants

Evaluation of mass-transport influence on kinetic data is often advisable. Tests should include

analysis over a concentration range from 0.1 to 10 KD, variation of the buffer flow rate and also

variation of the binding capacity using different surface densities13-15. In this work, consistent results

were observed for peptides analysed over a 30-fold concentration range, from as high as 10 KD

down to ca. 40 nM. The 0.1 KD condition was possible only for peptides with KD values at or above

the µM level, since response could not be accurately measured at lower peptide concentrations.

Mass-transport effects were evaluated on peptide A15 at two different buffer flow rates (2 and 60

µl/min) and three different surface capacities (ca. 0.5, 1.6 and 2.5 ng/mm2) as shown in Table 1.6.

Binding was not measurable at the lowest density surfaces, as expected from both previous results

(section 1.1) and small size of the analytes.


65

A

0

20

40

60

80

100

120

137P

139P

141P

143P

145P

147P

137S

140S

142S

144S

147S

138D

138F

138R

138Y

145E 14

5I

145R

147D

147G

147N

147V

ELISA

SPR

B

0

20

40

60

80

100

120

137P

139P

141P

143P

145P

147P

137S

140S

142S

144S

147S

138D

138F

138R

138Y

145E 14

5I

145R

147D

147G

147N

147V

ELISA

SPR

Figure 1. 6 Comparison of SPR [relative KD =KD (peptide X)/KD (peptide A15)] and ELISA[relative IC50=IC50 (peptide X)/IC50 (peptide A15)] affinity data for the 43 variants of A15towards: A. mAb SD6; B. mAb 4C4. Peptides displaying IC50 or KD values too high to beaccurately measured are represented by bars truncated at 100. In the horizontal axis arerepresented the A15 analogues screened, with the number corresponding to the A15 positionreplaced and the letter corresponding to the capital case code of the amino acid residueintroduced at that position (only half of the peptide labels are shown for simplicity).


66

Higher densities did not show important differences in kinetic rate constants, with all data sets giving

best fits to the 1:1 langmuirian interaction model. Despite deviations observed when the smaller

buffer flow rate was employed, these hardly affected the magnitude of the kinetic rate constants or

the quality of the fitted data.

Table 1.6 Kinetic data for the mAb SD6/peptide A15 and mAb 4C4/peptide A15 binding interactions underdifferent buffer flow rate and surface density conditions.

Buffer

flow rate

SD6/ng.mm-2 4C4/ ng.mm-2

(µµµµL/min) 0.5 1.6 2.5 0.4 1.7 2.7

2 *

ka=5.9×104M-1s-1

kd=1.3×10-3s-1

χ2=0.3

ka=9.0×104M-1s-1

kd=1.2×10-3s-1

χ2=1.1

*

ka=2.1×105M-1s-1

kd=1.6×10-3s-1

χ2=2

ka=2.6×105M-1s-1

kd=8.4×10-4s-1

χ2=0.4

60 *

ka=7.3×104M-1s-1

kd=1.4×10-3s-1

χ2=0.2

ka=1.1×105M-1s-1

kd=1.8×10-3s-1

χ2=2.2

*

ka=3.8×105M-1s-1

kd=1.9×10-3s-1

χ2=1.0

ka=5.0×105M-1s-1

kd=2.0×10-3s-1

χ2=0.5

* no reliable measurements at this surface density.

Mass-transport limitations are not usually dramatic for small analytes and can be avoided with

careful experimental set-ups, where high buffer flow rates and low surface densities are key features.

However, even when careful experimental design is applied and apparently consistent data is

obtained, one cannot rule out the possibility of diffusion-controlled kinetics. Hence, the SPR-derived

kinetic rate constants cannot be considered as absolutely “true” values. Further, one cannot fully

compare the events taking place at the biosensor surface, where the biological receptor is

immobilised, with those occurring in solution or in physiologic media. Although agreement with

ELISA experiments provides a valuable check for the reliability of biosensor data, one cannot write

off the possibility that mass-transport affects actual ka and kd values by a similar factor, thus

providing thermodynamic constants apparently consistent with equilibrium experiments.

Nevertheless, the real usefulness of the SPR technology lies in the comparative analysis of the kinetic

behaviour of analogous analytes screened under the same experimental conditions and this was the

purpose of the present work.


67

1.6 Relevance of the SPR data for FMDV studies

Antigenic site A is a key component of the immune response against FMDV, and some of its

constituent amino acid residues play a decisive role in the mechanisms of FMDV escape under

immune pressure19. The involvement of the RGD tripeptide motif in both antibody and host cell

recognition, as well as the importance of key adjacent residues such as Leu 144 and Leu 147 were

well-established in previous studies, where site A variant peptides proved very useful in probing the

antigenic structure of this site5,6,11,20. Since only equilibrium data had been reported so far, the

dynamic aspects of site A peptide-antibody interactions remained unexplored and real-time

biospecific SPR analysis seemed the right tool to perform such an exploration. Forty-three analogues

of the site A reference peptide A15 (from the C-Sc8c1 FMDV clone) were chosen to show how

structural variation within site A can be correlated with and adequately explained by kinetic SPR

data.

The choice of peptides focused on several structural features of antigenic site A. A proline scan was

first performed from residues 137 to 148 of the GH loop. The well-known structure-disrupting effect

of Pro was reflected in a complete absence of measurable binding when Pro was replacing residues

within the RGD triplet or the following short helical stretch at Asp 142 – Leu 1475,6. Replacement at

the N-terminal region did not affect binding in positions 137 and 140, but produced a slight

decrease and a significant loss in antigenicity for positions 138 and 139, respectively. This agrees

with reported observations that Ser 139 participates in important polar interactions6 in which Pro is

unable to engage. Next, a serine scan was performed, given the striking preservation of antigenicity

in site A variants having Ser at critical positions6. Ser replacements at the 137 – 142 region are in

general well-tolerated, including positions 141 and 142, corresponding to Arg and Gly of the RGD

motif. On the other hand, changes at either Asp 143, Leu 144 and Leu 147 were clearly detrimental

to recognition, a result which can be explained by (i) the role of the Asp residue of the RGD motif in

antibody recognition and (ii) the fact that both Leu 144 and Leu 147 are involved in hydrophobic

interactions in all available three-dimensional structures of peptide A15 – antibody complexes5,6.

A third group of variant peptides included replacements at positions 138, 145 and 147 to illustrate

the subtle effects that structural variation can bring about in antibody recognition. For instance,

replacements with charged basic (Arg, Lys) residues at Ala 138 are better tolerated by mAb 4C4

than by SD6. This is in agreement with the higher percentage of residue contact observed for the

latter mAb in the crystal structure of A15 – antibody complexes6. Non-polar aliphatic or aromatic

amino acids seem to be acceptable by both mAbs at this position. Changes at position 145 (Ala) are

similarly interesting. While SD6 does not recognise peptides with non-polar replacements, 4C4

easily binds the same mutated peptides. Even more striking is the reactivity of both mAbs with the

Glu-replaced peptide, whereas the Asp mutation is not recognised. Finally, position 147 provides

the more critical differentiation between both mAbs assayed: while SD6 is quite tolerant to

mutations (except Asp), 4C4 is extremely sensitive to changes at this position.


68

References

1 Fägerstam, L. G., Frostell-Karlsson, A., Karlsson, R., Persson, B. and Rönnberg, I. (1992) Biospecificinteraction analysis using surface plasmon resonance detection applied to kinetic, binding site andconcentration analysis, J. Chromatogr. 597, 397.

2 Van Regenmortel, M. H. V., Altschuh, D., Pellequer, J. L., Richalet-Sécordel, P., Saunal, H., Wiley, J.A., Zeder-Lutz, G. (1994) Analysis of viral antigens using biosensor technology. Methods: A Comp.Meth. Enzymol. 6, 177.

3 Carreño, C., Roig, X., Camarero, J., Mateu, M. G., Domingo, E., Giralt, E., Andreu, D. (1992) Studieson antigenic variability of C strains of foot-and-mouth disease virus by means of synthetic peptidesand monoclonal antibodies. Int. J. Peptide Protein Res. 39, 41.

4 Feigelstock, D. A., Mateu, M. G., Valero, M. L., Andreu, D., Domingo, E., Palma, E. L. (1996)Emerging foot-and-mouth disease virus variants with antigenically critical amino acid substitutionspredicted by model studies using reference viruses. Vaccine 14, 97.

5 Verdaguer, N., Mateu, M. G., Andreu, D., Giralt, E., Domingo, E., Fita, I. (1995) Structure of themajor antigenic loop of foot-and-mouth disease virus complexed with a neutralizing antibody: directinvolvement of the Arg-Gly-Asp motif in the interaction. EMBO J. 14, 1690.

6 Verdaguer, N., Sevilla, N., Valero, M. L., Stuart, D., Brocchi, E., Andreu, D., Giralt, E., Domingo, E.,Mateu, M. G., Fita, I. (1998) A similar pattern of interaction for different antibodies with a majorantigenic site of foot-and-mouth disease virus: implications for intratypic antigenic variation. J. Virol.72, 739.

7 Mateu, M.G., Andreu, D., Carreño, C., Roig, X., Cairó, J. J., Camarero, J. A., Giralt, E. andDomingo, E. (1992) Non-additive effects of multiple amino acid substitutions on antigen-antibodyrecognition. Eur. J. Immunol. 22, 1385.

8 Karlsson, R., Ståhlberg R. (1995) Surface plasmon resonance detection and multispot sensing fordirect monitoring of interactions involving low-molecular-weight analytes and for determination of lowaffinities. Anal. Biochem. 228, 274.

9 Karlsson, R. (1994) Real-time competitive kinetic analysis of interactions between low-molecular-weight ligands in solution and surface-immobilized receptors. Anal. Biochem. 221, 142.

10 Abbas, A. K., Lichtman, A. H. and Pober, J. S. “Cellular and molecular immunology”, 3rd edition; W.B. Saunders Co., New York (1997).

11 Mateu, M. G., Valero, M. L., Andreu, D. and Domingo, E. (1996) Systematic replacement of aminoacid residues within an Arg-Gly-Asp-containing loop of foot-and-mouth disease virus and effect on cellrecognition. J. Biol. Chem. 271, 12814.

12 O’Shannessy, D. J., Winzor, D. J. (1996) Interpretation of deviations to pseudo-first-order kineticbehavior in the characterization of ligand binding by biosensor technology. Anal. Biochem. 236, 275.

13 Schuck, P. (1997) Use of surface plasmon resonance to probe the equilibrium and dynamic aspects ofinteractions between biological macromolecules. Ann. Rev. Biophys. Biomol. Struct. 26, 541.

14 Schuck, P. (1997) Reliable determination of binding affinity and kinetics using surface plasmonresonance biosensors. Curr. Op. Biotechnol. 8, 498.

15 Shuck, P. and Minton, A. P. (1996) Kinetic analysis of biosensor data: elementary tests for auto-consistency. Trends Biochem. Sci. 21, 458-460.

16 Altschuh, D., Dubs, M. C., Weiss, E., Zeder-Lutz, G., Van Regenmortel, M. H. V. (1992)Determination of kinetic constants for the interactions between a monoclonal antibody and peptidesusing surface plasmon resonance. Biochemistry 31, 6298.

17 England, P., Brégére, F. and Bedouelle, H. (1997) Energetic and kinetic contributions of contactresidues of antibody D1.3 in the interaction with lysozyme. Biochemistry 36, 164.

18 VanCott, T. C., Bethke, F. R., Polonis, V. R., Gorny, M. K., Zolla-Pazner, S., Redfield, R. R. and Birx,D. L. (1994) Dissociation rate of antibody-gp120 binding interactions is predictive of V3-mediatedneutralization of HIV-1. J. Immunol. 153, 449.

19 Martínez, M. A., Hernández, J., Piccone, M. E., Palma, E. L., Domingo, E., Knowles, E. and Mateu,M. G. (1991) Two mechanisms of antigenic diversification of foot-and-mouth disease virus. Virology184, 695.

20 Hernández, J., Valero, M. L., Andreu, D., Domingo, E. and Mateu, M. G. (1996) Antibody and hostcell recognition of foot-and-mouth disease virus (serotype C) cleaved at the Arg-Gly-Asp (RGD) motif:a structural interpretation. J. Gen. Virol. 77, 257.

2. Antigenic determinants

in the GH loop of FMDV C1-Barcelona


70

Antigenic determinants in the GH loop of FMDV C1-Barcelona

71

2.0 Introduction

On previous genomic studies of FMDV field isolates, a natural variant, C1-Barcelona (or C-S30), was

characterised as containing four mutations in the GH loop (Ala138→Thr, Ala140→Thr,

Leu147→Val and Thr149→Ala) relative to the reference strain C-S8c1 (Table 2.1). Analyses of this

four-point mutant by immuno-enzymatic assays showed that it was fully recognised by site

A-directed mAbs such as 4C4. Further, this behaviour was confirmed in antigenicity studies of

peptide – keyhole limpet hemocyanin (KLH) conjugates reproducing the four relevant mutations1-7.

The fact that one of the mutations, Leu147→Val, was found to be detrimental for antibody and cell

recognition of site A peptides, makes the GH loop of FMDV C-S30 an interesting example to learn

more about antigen-antibody recognition mechanisms in FMDV.

The second objective of the present work was, therefore, the synthesis and analysis of peptides

mimicking not only the GH loop of the C-S30 strain but also all possible partial mutants of this

natural isolate.

2.1 Peptides mimicking the GH loop of FMDV C1-Barcelona and the corresponding

partial mutants

A set of fifteen pentadecapeptides was synthesised, corresponding to all possible combinations of

the four mutations found in antigenic site A of C1-Barcelona (C-S30), taking peptide A15 (GH loop

of FMDV C-S8c1) as the reference sequence (Table 2.1).

These peptides were synthesised by machine-assisted parallel solid-phase peptide synthesis, using

standard Fmoc/tBu protocols 8-10 as shown in Fig. 2.1 and described in more detail in section 4.2

(Materials & Methods). Crude peptides were obtained following cleavage from the resin and

submitted to further purification (Fig. 2.2) by medium-pressure liquid chromatography (MPLC).

Purified products were all satisfactorily identified (MALDI-TOF MS, AAA) as the target peptides, with

global yields of ca. 50% (Table 2.2). Peptides were lyophilised and stored at – 20 oC prior to their

utilisation in subsequent studies.


72

Table 2.1 Pentadecapeptides reproducing all possible combinations of the mutations found in the GHloop of FMDV C1-Barcelona (C-S30).

Name Sequence Mutants

A15 YTASARGDLAHLTTT GH loop of FMDV C-S8c1

A15(138T) --T------------A15(140T) ----T---------- One-pointA15(147V) -----------V---A15(149A) -------------A-

A15(138T,140T) --T-T----------A15(138T,147V) --T--------V--- Two-pointA15(138T,149A) --T----------A-A15(140T,147V) ----T------V---A15(140T,149A)a ----T--------A- GH loop of FMDV C1-Brescia

A15(147V,149A) -----------V-A- Two-point

A15(138T,140T,147V) --T-T------V---A15(138T,140T,149A) --T-T--------A- Three-pointA15(138T,147V,149A) --T--------V-A-A15(140T,147V,149A) ----T------V-A-

A15(138T,140T,147V,149A)b --T-T------V-A- GH loop of FMDV C-S30a termed A15Brescia further on.b termed A15S30 further on.

0 30 min

5% B 95% B

0 30 min

5% B 35% B

MPLC

Figure 2. 2 Typical HPLC profiles obtained in the synthesis (left) andpurification (right) of the FMDV C-S30 peptides.


73

Figure 2. 1 Schematic representation of the general protocol in Fmoc/tBu solid-phase peptide synthesis (described under Materials & Methods, section 4.2)8-10.

Boc

TFA

Fmoc

+ Fmoc-linker-OH+ coupling agent

+ Fmoc- -OH+ coupling mixture

PG PG

Fmoc

+ Fmoc- -OH+ coupling mixture

PG

PG

PG

Fmoc

PG

PG

PG

PG

PG

PG

PG

TFA (+ scavengers)

resin bead (MBHA, PEG-PS, ...)

amino acid residue

PG protecting groups (tBu, OtBu, Trt, Pmc, ...)

bi-functional spacer (handle)

bond scission

piperidine


74

Table 2.2 Yield, purity (HPLC) and characterisation (MALDI-TOF MS and AAA) of the C-S30 series.

Peptide Global yield(%)

Purity(% HPLC)

MW found

MWexpected

Amino acid analysis(AAA)

A15(138T) 52 90 1607.2 1607 Asp, 1.12 (1); Ser, 0.94 (1); Gly, 0.98 (1); Ala 2.03 (2); His, 1.08 (1); Arg, 0.85 (1)A15(140T) 49 81 1607.1 1607 Asp, 0.97 (1); Ser, 0.85 (1); Gly, 1.05 (1); Ala 2.06 (2); His, 1.09 (1); Arg, 0.88 (1)

A15(147V) 43 87 1562.9 1563 Thr, 4.06 (4); Ser, 0.97 (1); Gly, 0.98 (1); Ala 3.06 (3); His, 0.86 (1); Arg, 0.89 (1)

A15(149A) 51 80 1547.1 1547 Asp, 0.92 (1); Ser, 0.80 (1); Gly, 1.03 (1); Ala 4.10 (4); His, 1.06 (1); Arg, 0.99 (1)

A15(138T,140T) 46 85 1636.9 1637 Asp, 1.10 (1); Ser, 0.88 (1); Gly, 1.07 (1); Ala 1.06 (1); Leu, 2.01 (2); His, 0.87 (1)

A15(138T,147V)* 10 76 1592.8 1593 Asp, 0.96 (1); Ser, 1.13 (1); Val, 0.95 (1); Leu, 1.15 (1); His, 0.95 (1); Arg 1.01 (1)

A15(138T,149A) 50 96 1576.9 1577 Asp, 0.94 (1); Ser, 0.91 (1); Gly, 1.06 (1); Leu, 2.04 (2); His, 0.97 (1); Arg, 0.91 (1)

A15(140T,147V) 41 94 1592.7 1593 Asp, 0.90 (1); Ser, 0.87 (1); Gly, 1.07 (1); Ala 2.09 (2); Leu, 1.10 (1); Arg, 0.84 (1)

A15Brescia 36 91 1576.9 1577 Asp, 0.95 (1); Ser, 0.87 (1); Gly, 1.06 (1); Ala 3.06 (3); Leu, 1.99 (2); His, 0.94 (1)

A15(147V,149A)* 21 79 1533.1 1533 Asp, 0.71 (1); Gly, 1.07 (1); Ala 4.04 (4); Val, 0.68 (1); Leu, 1.03 (1); Arg, 0.86 (1)

A15(138T,140T,147V) 48 93 1623.6 1623 Asp, 1.03 (1); Ser, 0.88 (1); Gly, 1.09 (1); Ala 1.08 (1); Leu, 1.08 (1); Arg, 0.84 (1)

A15(138T,140T,149A) 41 91 1607.6 1607 Asp, 0.92 (1); Ser, 0.81 (1); Gly, 1.06 (1); Ala 2.14 (2); Leu, 1.96 (2); His, 0.92 (1)

A15(138T,147V,149A) 52 98 1562.9 1563 Asp, 0.98 (1); Ser, 0.90 (1); Gly, 1.01 (1); Ala 3.06 (3); Leu, 1.04 (1); Arg, 1.01 (1)

A15(140T,147V,149A) 98 1562.9 1563 Asp, 1.02 (1); Ser, 0.86 (1); Gly, 1.03 (1); Ala 3.04 (1); Leu, 1.05 (1); Arg, 0.82 (1)

A15S30 59 99 1592.7 1593 Asp, 1.10 (1); Ser, 0.88 (1); Gly, 1.07 (1); Ala 1.06 (1); Leu, 2.01 (2); His, 0.87 (1)

* These syntheses were carried out under sub-optimal conditions due to instrumental malfunction.Relative amino acid ratios found by AAA are followed by the expected value in parenthesis.


75

2.2 SPR study of the C-S30 peptides

Having found suitable experimental conditions for the SPR kinetic study of interactions between

FMDV peptides in solution and immobilised anti-FMDV antibodies, as described in chapter 1, SPR

was again chosen for the characterisation of the C-S30 peptides. This would allow the detailed study

of the effects caused by the stepwise introduction of the four mutations found in the GH loop of

FMDV C-S30 and provide a possible explanation for the peculiar behaviour of this virus isolate.

The C-S30 pentadecapeptides were, therefore, screened by SPR against three anti-site A

monoclonal antibodies, SD6, 4C4 and 3E5A.

The three mAbs were immobilised on CM5 sensor chips following standard protocols, with final

immobilisation densities of about 1600 RU. Sensorgrams were obtained and analysed as previously

described (chapter 1) and all measurable interactions fitted to the 1:1 langmuirian interaction kinetic

model (often considering baseline drift). Whenever interactions could not be reliably measured,

sensorgrams had a square-wave like shape, either due to bulk refractive index response or to

extremely fast on/off rates. Although the monitored on/off rates were generally high, and

consequently liable to be under mass-transport effects, the quantitative data obtained appeared to

be self-consistent and were thus considered reliable. Discussion of the results is presented in the

following sections.

2.2.1 One-point mutants

As observed in previous studies by competition ELISA, SPR analysis showed that substitutions

Ala140→Thr and Thr149→Ala were well tolerated by the three mAbs. This was to be expected,

since both replacements are present in field isolate C1-Brescia, previously shown to be recognised by

these mAbs4.

Mutations Ala138→Thr and Leu147→Val affected antibody recognition to different extents: mAb

SD6 was more sensitive to Ala138→Thr than to Leu147→Val, quite the opposite to mAb 4C4,

which tolerated the first replacement much better than the second one. Similar effects were observed

for mAb 3E5. These results are consistent with recent crystallographic studies, where it was found

that Ala138 has a higher percentage of residue contact with mAb SD6 than with mAb 4C414.

Peptide affinities to each mAb are mainly reflected in the different dissociation rate constants

observed for the corresponding peptide-mAb complexes (Table 2.3), as further illustrated in Fig. 2.3.

A this anti-site A mAb was raised against FMDV strain C1-Brescia. Ascitic fluid of mAb 3E5 was kindly supplied by Dr.Emiliana Brocchi (IZSLE – Brescia, Italy).


76

Table 2.3 Kinetic SPR analysis of the interactions between FMDV C-S30 peptides and mAbs SD6, 4C4 and 3E5.mAb SD6 4C4 3E5


-1 KA/M-1 ka/M-1s-1 kd/s

-1 KA/M-1 ka/M-1s-1 kd/s

-1 KA/M-1

A15 7.3×104 1.4×10-3 5.4××××107 3.8×105 1.9×10-3 1.9××××108 1.6×105 1.6×10-3 9.4××××107

A15(138T) 1.0×105 1.5×10-2 6.5××××106 2.5×105 5.9×10-3 4.2××××107 1.1×105 8.5×10-3 1.3××××107

A15(140T) 1.4×105 3.0×10-3 4.7××××107 6.0×105 2.6×10-3 2.3××××108 2.6×105 1.5×10-3 1.8××××108

A15(147V) 1.1×105 1.0×10-2 1.0××××107 9.5×104 4.4×10-2 2.2××××106 3.3×105 5.0×10-2 6.6××××106

A15(149A) 1.2×105 2.2×10-3 5.5××××107 6.4×105 3.1×10-3 2.1××××108 4.8×105 1.5×10-3 3.2××××108

A15(138T,140T) 1.6×105 1.5×10-2 1.1××××107 2.6×105 8.2×10-3 3.1××××107 2.4×105 8.4×10-3 2.9××××107

A15(138T,147V) 3.8×104 4.2×10-2 9.1××××105 2.4×105 4.2×10-2 5.7××××106 2.6×105 4.2×10-2 6.3××××106

A15(138T,149A) 1.2×105 1.7×10-2 7.0××××106 2.5×105 3.2×10-3 7.6××××107 3.2×105 8.3×10-3 3.9××××107

A15(140T,147V) 7.8×104 1.3×10-2 6.1××××106 2.4×105 6.5×10-2 3.7××××106 3.8×105 4.8×10-2 8.0××××106

A15Brescia 9.0×104 8.0×10-3 1.2××××107 2.6×105 1.6×10-3 1.6××××108 4.4×105 4.2×10-3 1.0××××108

A15(147V,149A) 1.0×105 1.7×10-2 6.0××××106 2.1×105 4.4×10-2 4.7××××106 4.8×105 3.7×10-2 1.3××××107

A15(138T,140T,147V) ni 3.4×105 1.6×10-1 2.1××××106 ni

A15(138T,140T,149A) 2.2×105 2.2×10-2 9.7××××106 2.1×105 9.6×10-3 2.2××××107 3.4×105 1.4×10-2 2.5××××107

A15(138T,147V,149A) ni 3.1×105 5.2×10-2* 6.0××××106 3.6×105 4.1×10-2 8.9××××106

A15(140T,147V,149A) 1.7×105 1.8×10-2 9.1××××106 4.4×105 7.2×10-2* 6.1××××106 5.9×105 5.2×10-2* 1.1××××107

A15S30 3.8×104 8.8×10-2* 4.3××××105 2.2×105 1.2×10-1* 2.0××××106 3.2×105 7.2×10-2* 4.5××××106

* Although resulting from apparently reliable data fits, kd values equal or higher than 5×10-2 should be regarded with some caution, since this value is considered the limit ofreliable SPR measurement of rate constants; “ni” denotes interactions which could not be reliably measured.


77

A

-10

0

10

20

30

40

50

60

70

80

90

-10 40 90 140 190 240 290Time/s

Re

sp

on

se

/RU

A15138T 39 nMA15138T 78 nMA15138T 156 nMA15138T 312 nMA15138T 625 nMA15138T 1250 nM

B

-5

15

35

55

75

95

115

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

A15140T 39 nMA15140T 78 nMA15140T 156 nMA15140T 312 nMA15140T 625 nMA15140T 1250 nM

C

-45

-35

-25

-15

-5

5

15

25

35

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

A15147V 39 nMA15147V 78 nMA15147V 156 nMA15147V 312 nMA15147V 625 nMA15147V 1250 nM

D

-5

15

35

55

75

95

115

135

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

A15149A 39 nMA15149A 78 nMA15149A 156 nMA15149A 312 nMA15149A 625 nMA15149A 1250 nM

Figure 2. 3 Sensorgrams obtained in the SPR kinetic analysis of the interactions between immobilised mAb 4C4 and: A. peptide A15(138T); B. peptideA15(140T); C. peptide A15(147V) and D. peptide A15(149A). This figure illustrates the differences observed in the dissociation rate constants for eachmAb-peptide interaction.


78

2.2.2 Two- and three-point mutants

Analysis of data in Table 2.3 immediately suggests that two- and three-point combinations of the

amino acid replacements are additive. Indeed, antigenicities of the two- and three-point mutant

peptides towards the three mAbs employed reflect the combined effects of the single-point mutations

present in each particular sequence. This effect is further confirmed when comparing the

experimental relative affinities [KArel=KA(peptide)/KA(A15)] with the calculated relative affinities

assuming additive effects in the combination of single-point mutations [expected KArel =KArel(single-

point mutant 1) × KArel(single-point mutant 2) × ... × KA(peptide A15)], as illustrated in Fig. 2.4.

Peptide A15Brescia (with replacements A140→T and T149→A) was the most antigenic within the

group of multiple-point mutants (Table 2.4). Interestingly, for this peptide the correlation between

experimental and calculated relative affinities was poorer than for all the other mutants, suggesting a

compensatory effect (i.e., lack of additivity) between both replacements. On the other hand,

peptides containing the L147→V substitution were the poorest antigens, particularly when the

A138→T replacement was also present. Further, affinities were again almost exclusively determined

by the dissociation rates of peptide – mAb complexes (compare data within each mAb set in Table

2.3).

2.2.3 Four-point mutant: peptide A15S30

The most striking result obtained in this SPR screening of FMDV peptides was the low affinity

observed for peptide A15S30 (reproducing the GH loop of FMDV C-S30) towards all three mAbs

assayed. Although such low antigenicity was totally in agreement with the additive effects observed

when combining the different individual substitutions in the partial mutants (Fig. 2.4), previous

studies with FMDV C-S30 had shown that this natural isolate was neutralised by mAb 4C415 and,

further, that the KLH conjugate of a 21-mer peptide reproducing the C-S30 loop had been fully

recognised by the same mAb in immuno-enzymatic assays6,7.

Despite the fact that SPR data was apparently self-consistent and reliable, we decided to perform a

qualitative comparison between the SPR affinities of peptides A15, A15Brescia and A15S30 using a

reverse SPR configuration: peptide immobilisation and mAb as soluble analyte, as described under

Materials & Methods (section 4.3.1). Even though this configuration was not optimised (to avoid

artefacts such as diffusion controlled kinetics or ligand heterogeneity11), the performance of all assays

under identical conditions and the high molecular weight of the mAb analytes would provide both

high SPR responses and a reliable qualitative comparison between the three peptide antigens. As

shown in Table 2.4 and Fig. 2.5, this assay further confirmed the above SPR data: although rate

constants for peptide – mAb interaction depended on the analysis format (consistently lower in the

format with mAb as analyte), thermodynamic affinities were virtually the same and the antigenicity

ranking was identical to the one derived from the first SPR analysis of these peptides. Since these


79

SPR results did not agree with the neutralisation and immuno-enzymatic data discussed above, a

competition ELISA screening of all C-S30 peptides synthesised towards the three mAbs was

performed, as described in section 2.3.

Figure 2. 4 Experimental and calculated affinities of A15 mutantpeptides towards mAbs SD6 (A), 4C4 (B) and 3E5 (C). Calculated valueshave been determined assuming additive effects of the amino acidreplacements (see text).

A

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

A1

5(1

38T

,14

0T)

A1

5(1

38T

,14

7V)

A1

5(1

38T

,14

9A)

A1

5(1

40T

,14

7V)

A1

5(1

40T

,14

9A)

A1

5(1

47V

,149

A)

A1

5(1

38T

,14

0T,1

49A

)

A1

5(1

40T

,14

7V,1

49A

)

A1

5S30

Peptide

Re

lati

ve

aff

init

y

experimental

calculated

B

-0.05

0.15

0.35

0.55

0.75

0.95

1.15

1.35

A1

5(1

38T

,14

0T)

A1

5(1

38T

,14

7V)

A1

5(1

38T

,14

9A)

A1

5(1

40T

,14

7V)

A1

5(1

40T

,14

9A)

A1

5(1

47V

,149

A)

A1

5(1

38T

,14

0T,1

47V

)

A1

5(1

38T

,14

0T,1

49A

)

A1

5(1

38T

,14

7V,1

49A

)

A1

5(1

40T

,14

7V,1

49A

)

A1

5S30

Peptide

Re

lati

ve

aff

init

y

experimental

calculated

C

-0.05

0.95

1.95

2.95

3.95

4.95

5.95

A1

5(1

38T

,14

0T)

A1

5(1

38T

,14

7V)

A1

5(1

38T

,14

9A)

A1

5(1

40T

,14

7V)

A1

5(1

40T

,14

9A)

A1

5(1

47V

,149

A)

A1

5(1

38T

,14

0T,1

49A

)

A1

5(1

38T

,14

7V,1

49A

)

A1

5(1

40T

,14

7V,1

49A

)

A1

5S30

Peptide

Re

lati

ve

aff

init

y

experimental

calculated


80

Peptidebound slow dissociation

A

-5

0

5

10

15

20

25

30

35

40

45

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

A15, 38 nMA15, 76 nMA15, 152 nMA15, 305 nMA15, 610 nM

B

-5

15

35

55

75

95

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

A15S30, 39 nMA15S30, 78 nM

A15S30, 156 nMA15S30, 312 nMA15S30, 625 nM

Peptidebound

fast dissociation

C

-5

15

35

55

75

95

115

135

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

4C4, 16 nM4C4, 31 nM

4C4, 62 nM4C4, 125 nM4C4, 250 nM4C4, 500 nM

D

-5

15

35

55

75

95

115

135

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

4C4, 31 nM4C4, 62 nM

4C4, 125 nM4C4, 250 nM4C4, 500 nM

Figure 2. 5 Sensorgrams from the SPR direct analysis of the interactions between peptides A15 and A15S30 with mAb 4C4: A. immobilised mAb vs.A15; B. immobilised mAb vs. A15S30; C. immobilised A15 vs. 4C4; D. immobilised A15S30 vs. 4C4.


81

Table 2.4 Kinetic analyses using peptides immobilised on the chip.

mAb Peptide ka/M-1s-1 kd/s

-1 KA/M-1

A15 1.2×104 2.5×10-4 5.0×107

SD6 A15Brescia* 1.8×104 - -A15S30 2.9×103 1.8×10-2 1.6×105

A15 2.3×104 2.2×10-4 1.1×108

4C4 A15Brescia* 1.8×104 - -A15S30 2.5×104 1.1×10-2 2.3×106

A15 1.2×105 7.3×10-4 1.7×108

3E5 A15Brescia* 1.7×104 - -A15S30 3.5×104 1.1×10-2 3.2×106

* kd too small to be reliably measured; ka determined from the lineardependence of ks (global rate constant=ka×C+kd) on analyte concentration.

2.2.4 A possible significance for kinetic rate constants in antigen-antibody interactions

As already mentioned, a consistent observation in the present study was that peptide – antibody

affinities seem mainly determined by the dissociation rate constant, kd (Table 2.3). Hence, for a

given family of peptide analogues, a higher or lower antigenicity towards a given mAb would

exclusively depend on the half-life (t1/2) of the complex. Thus, all analogues would be similarly

capable of approaching the mAb paratope but, once there, their different ability to establish

complex-stabilising interactions would dictate the life-time of the complex and, consequently,

peptide – antibody affinity (KD). If this interpretation was true, the association rate constant (ka)

would be sequence-independent and only determined by the global fitness of the antibody paratope

(KD ∝ kd ⇒ ka = constant) to a series of analogue antigens. Indeed, a closer look at the kinetic data

of the SPR screening of 15-mer peptides (Table 2. 3) suggests that this might be the case. Plotting

affinities (KD) against dissociation rate constants (kd) gives for each mAb a set of points for which a

correlation line with a satisfactory r2 coefficient (>0.9) can be derived (Fig. 2.6). The slopes of these

three lines correlate well with the reciprocal of the average association rates of the interactions

among each mAb and the entire set of peptide antigens. Thus, slope of SD6 correlation line is

1×10-5 vs. 1/average ka = 9.0×10-6. Similarly, for 4C4, slope = 4×10-6 vs. 1/average ka = 3.1×10-6;

and for 3E5, slope = 3×10-6 vs. 1/average ka = 3.4×10-6.

Of course, these correlations would only apply to relatively similar sets of analogue antigens; more

drastic changes in antigen size, folding or amino acid composition would certainly be expected to

lead to substantial changes in both dissociation and association rate constants.


82

slope = 4 x 10-6

r2 = 0.99

slope = 3 x 10-6

r2 = 0.92

slope = 1 x 10-5

r2 = 0.94

0.00E+00

1.00E-07

2.00E-07

3.00E-07

4.00E-07

5.00E-07

6.00E-07

7.00E-07

0.00E+00 2.00E-02 4.00E-02 6.00E-02 8.00E-02 1.00E-01 1.20E-01

kd/s-1

KD/M

mAb SD6mAb 4C4mAb 3E5

Figure 2. 6 Correlation between thermodynamic affinity (KD) and kinetic dissociation constant (kd) for theSPR-measured interactions between C-S30 peptides and three anti-FMDV mAbs SD6, 4C4 and 3E5.


2.3 Competition ELISA analysis of the C-S30 pentadecapeptides

As already mentioned, in view of the surprisingly low affinity shown by the A15S30-4C4 complex

on SPR, a competition ELISA (Fig. 2.7) screening of all the C-S30 pentadecapeptides was

performed16,17. For this purpose, 96-well plates were coated with an FMDV C-S8c1 reference

antigen, the KLH conjugate of peptide A21 (corresponding to a 6-residue extension at the C-

terminus of peptide A15):

YTASARGDLAHLTTTHARHLP A21

Equilibrium mixtures of the competitor at different concentrations with constant concentrations of

the mAb were incubated and the amount of mAb which bound to antigen was measured. Plotting

the variation of plate-bound mAb with competitor peptide concentration gave inhibition curves such

as those in Fig. 2.8, thus providing an evaluation of competitor peptide antigenicity. This

antigenicity was expressed in terms of relative IC50 (concentration of competitor producing 50%

inhibition), where relative means normalised with respect to the IC50 of peptide A15. As can be seen

from Figs. 2.8 and 2.9, data derived from competition ELISA were in good agreement with previous

SPR results, in the sense that the poorest antigens in SPR were also the worse competitors in ELISA.

These results demonstrate that the low A15S30-4C4 affinity previously obtained did not come from

eventual artefacts in SPR analysis.

Plate-bound antigen (competed)

Competitor antigen in solution

Specific monoclonal antibody

Anti-Fc antibody (conjugated to a carrierthat provides a means for detection)

Figure 2. 7 Scheme of the central stepsin competition ELISA analysis: pre-equilibrated peptide competitor – mAbmixtures are incubated with plate-boundantigen; the amount of mAb thatpreferably binds the immobilised antigenis detected after incubation of a tagged(e.g. peroxidase) anti-Fc antibody.

83


84

0

20

40

60

80

100

120

0.1 1 10 100 1000

peptide concentration (pmol/100 µµµµl)

% a

bso

rban

ceA15S30/SD6

A15/SD6

A15S30/4C4

A15/4C4A15S30/3E5

A15/3E5

Figure 2. 8 Competition between plate-bound A21 – KLH and pentadecapeptides A15and A15S30 for anti-site A mAbs.

% Percentage of the maximal absorbance measured for mAb incubated with plate-bound antigen inthe absence of peptide competitor; all absorbances were corrected by subtraction of mean absorbanceobtained for negative controls.

0

10

20

30

40

50

60

A1

5(1

38

T)

A1

5(1

40

T)

A1

5(1

47

V)

A1

5(1

49

A)

A1

5(1

38

T,1

40

T)

A1

5(1

38

T,1

47

V)

A1

5(1

38

T,1

49

A)

A1

5(1

40

T,1

47

V)

A1

5B

resc

ia

A1

5(1

47

V,1

49

A)

A1

5(1

38

T,1

40

T,1

47

V)

A1

5(1

38

T,1

40

T,1

49

A)

A1

5(1

38

T,1

47

V,1

49

A)

A1

5(1

40

T,1

47

V,1

49

A)

A1

5S

30

Peptide

rela

tive

IC50

mAb SD6mAb 4C4

mAb 3E5

Figure 2. 9 Screening of the C-S30 pentadecapeptides by competition ELISA.

IC50 values were normalised to IC50 of peptide A15; IC50 higher than the maximumcompetitor concentration are truncated at 60.


85

2.4 Size effects in the antigenicity of C-S30 peptides

The observation of a low antigenicity for peptide A15S30 in solution, and the confirmation that it

was not artefactual (e.g. peptide instability or aggregation) prompted us to investigate if the

discrepancies between our measurements and previous immunoenzymatic results1-7 could be due to

the small size of our peptides. Therefore, two additional 21-mer versions of antigenic site A in FMDV

strains C-S8c1 (peptide A21) and C-S30 (peptide A21S30) were synthesised and analysed by SPR

and competition ELISA.

YTASARGDLAHLTTTHARHLP A21

YTTSTRGDLAHVTATHARHLP A21S30

2.4.1 Synthesis of 21-mer peptides reproducing antigenic site A from FMDV strains C-S8c1 and

C-S30

Synthetic procedures and results of the synthesis and purification of these 21-mer peptides were

identical to those described in section 2.1, employing either manual or machine-assisted Fmoc/tBu

solid-phase peptide synthesis (see Materials & Methods, section 4.2). The results are summarised in

Table 2.5.

Table 2.5 Synthesis data of A21 and A21S30.

Peptide Globalyield (%)

Purity(% HPLC)

MWfound

MWexpected

AAA*

A21 37 94 2302.6 2303 Asp, 0.97 (1); Gly, 1.06 (1); Ala, 3.89 (4);Leu, 3.19 (3); His, 2.87 (3); Pro, 1.12 (1)

A21S30 31 98 2287.8 2288 Asp, 0.93 (1); Gly, 0.91 (1); Ala, 3.12 (3);Leu, 2.17 (2); His, 3.02 (3); Pro, 1.08 (1)

* Relative amino acid ratios given by AAA are followed by the theoretical values in parenthesis.


86

2.4.2 Antigenic analyses of A21 and A21S30 by SPR and competition ELISA

Peptides A21 and A21S30 were studied by SPR using peptide as either analyte or immobilised

ligand. As already explained in section 2.2.3, SPR analyses using immobilised peptide were not fully

optimised for artefacts such as ligand heterogeneity or mass-transport limitations. This is reflected in

the lower kinetic constants measured when using mAb as analyte. Nevertheless, one of our major

goals in the present work has been the application of SPR analysis to screen small antigenic peptides

as analytes and all experiments in the reverse format should be regarded as merely comparative.

SPR data of the interactions between peptide A21S30 and the anti-GH loop mAbs are displayed in

Table 2.6. Parallel analyses of peptide A21 were also performed but could not be accurately

quantitated due to either insufficient surface regeneration (using mAb immobilised on the chip) or

extremely small off-rates (using peptide immobilised on the chip). Only the interaction between

immobilised SD6 and A21 could be measured: ka=1.3×105 M-1s-1, kd=6.1×10-3 s-1, KA=2.1×107 M-1.

Although this was a serious drawback for an accurate evaluation of the antigenicity of A21, it

provided further evidence of the high affinity of the C-S8c1 peptides towards anti-GH loop mAbs. In

turn, peptide A21S30 displayed high dissociation rate constants (kd) as previously observed with the

shorter 15-mer peptide A15S30 (Fig. 2.10). A slight increase in affinity could be observed upon

addition of further 6 amino amino acid residues to the sequence of the C-S30 GH loop, but such

increase was also qualitatively observed for the C-S8c1 sequence, thus maintaining the antigenicity

ranking already observed and discussed in previous sections.

Table 2.6 Interactions of anti-site A mAbs with A21S30.

PeptideA21S30

mAbas analyte immobilised

on the chipka/M

-1s-1 8.6×104 5.4×103

SD6 kd/s-1 2.9×10-2 4.3×10-3

KA/M-1 3.0××××106 1.2×106

ka/M-1s-1 2.4×105 2.5×104

4C4 kd/s-1 4.3×10-2 3.7×10-3

KA/M-1 5.6××××106 6.8××××106

ka/M-1s-1 2.8×105 5.0×104

3E5 kd/s-1 5.0×10-2 4.8×10-3

KA/M-1 5.6××××106 1.0××××107


87

A

12690

12700

12710

12720

12730

12740

12750

12760

12770

12780

-10 40 90 140 190 240 290Time/s

Re

sp

on

se

/RU

A21, 40 nMA21, 80 nMA21, 159 nMA21, 318 nMA21, 636 nMA21, 1278 nM

B

-5

5

15

25

35

45

55

65

75

85

95

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

A21S30, 39 nMA21S30, 78 nMA21S30, 155 nMA21S30, 310 nMA21S30, 621 nMA21S30, 1242 nM

C

-5

495

995

1495

1995

2495

2995

3495

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

4C4, 15 nM4C4, 31 nM4C4, 62 nM4C4, 125 nM4C4, 250 nM4C4, 500 nM

D

-10

0

10

20

30

40

50

60

70

80

90

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU


Figure 2. 10 Interactions between mAb 4C4 and peptides A21 and A21S30: A. peptide A21 vs. immobilised mAb (note the increasing baseline responsedue to incomplete surface regeneration); B. peptide A21S30 vs. immobilised mAb (note the high dissociation rate); C. mAb 4C4 vs. immobilised A21(note the extremely low dissociation rate); D. mAb 4C4 vs. immobilised A21S30 (high dissociation rate).


88

Competition ELISA analysis of the two 21-mer peptides was also performed as described in section

2.3. Once again, peptide A21S30 was seen to be less antigenic than peptide A21 towards the three

mAbs assayed (Fig. 2.11).

0

20

40

60

80

100

120

0.1 1 10 100 1000

Peptide concentration (pmol/100 µµµµl)

% a

bso

rban

ce

A21S30_SD6A21S30_4C4A21S30_3E5A21_SD6A21_4C4A21_3E5

Figure 2. 11 Screening of peptides A21 and A21S30 by competition ELISA.

In summary, results from the antigenic characterisation of these 21-mer peptides showed that

peptide size did not account for the low antigenicity observed for peptide A15S30 in solution. So the

question of why the C-S30 peptide sequences were less antigenic than expected from previous

immunological studies1-7 remained open.

The total agreement between SPR and ELISA experiments proved that the discrepancies between

our data and previous immunoenzymatic results could not come from technical or experimental

artefacts.

At this point, we could only envisage a last-resource explanation, namely that in this particular case

peptide conformation was responsible for the different behaviour of peptide A15S30 (or A21S30)

when analysed by either ELISA/SPR or by immunodot. This remote possibility would run contrary

to the general observation that the continuous antigenic site A is perfectly mimicked by linear

peptides under all circumstances.


89

2.5 Input from parallel X-ray diffraction studies

At this point of our work we had access to X-ray diffraction studies performed by Wendy F. Ochoa

and co-workers (IBMB, CSIC - Barcelona) showing that peptide A15S30 could form a complex with

the Fab fragment of mAb 4C418. The peptide adopted a nearly cyclic conformation in the complex,

almost identical to the one observed for a similar complex with peptide A15. In the A15S30-4C4

complex, the more critical positions, residues 138 and 147, showed very few direct contacts with the

antibody. Therefore, these residues influence antigenicity by altering the conformation of the peptide

as a whole, rather than by local interactions. In the A15S30-4C4 complex, an additional water

molecule (which could not occupy the same position in the A15-4C4 complex due to steric

hindrance) was seen to bridge the side chain hydroxyl group of 138Thr and the main chain oxygen

atoms of 144Leu and 147Val (Fig. 2.12). Thus, this water molecule seemed to be a key feature in

holding the compact fold of the peptide, which has been further confirmed by molecular dynamics

simulations performed by the same researchers18.

Figure 2. 12 Conformation of peptide A15S30 complexed with theFab fragment of mAb 4C4; interactions between peptide (dark blue)and antibody (light blue) or water molecules (red spheres) are markedwith dashed lines; the RGD motif is located in an open turnconformation, closely related to the one previously observed for theviral loop as part of the virion; the water molecule that bridgesresidues 138 (Thr side chain hydroxyl), 144 and 147 (main chainoxygen atoms) is the one further on the right side of the image.(Figure was provided by Ms. Wendy F. Ochoa).


90

2.6 Effect of conformation in the antigenicity of C-S30 peptides

When earlier FMDV C-S30 studies, recent X-ray diffraction data and our own results on antigenicity

of C-S30 peptides are all brought together, it seems plausible that the apparent discrepancies

between the former and the latter data are due to factors such as distinct conformations of the

antigenic site under the different assay systems employed. Therefore, we decided to analyse the

effect of introducing conformational constraints in the C-S30 peptides. For that purpose, two

additional cyclic peptides (cyc16S30 and cyc16147Val) were synthesised.

2.6.1 Synthesis of peptides cyc16S30 and cyc16147Val

The design and synthesis of these cyclic peptides was based on previous research in our group,

involving the same type of constructs based on the GH loop sequence belonging to the C-S8c1 viral

strain. Cyclic versions of the viral antigenic site were formed by intra-molecular disulphide formation

between N- and C-terminal cysteine residues added to the sequences. An extra 6-aminohexanoic

acid (Ahx) residue was also included to give some flexibility to the constructs, so that the known

disordered structure of the viral loop could be better mimicked19. Production of these cyclic peptides

relied on the synthesis and purification of the corresponding linear bis-thiol precursors (A15S30 and

A16147Val) using methods similar to those previously described in section 2.1, followed by air

oxidation of the thiol groups at pH 8 and high peptide dilution (Fig. 2.13)20. Cyclization was

monitored by HPLC and the qualitative Ellman21 assay (see Materials & Methods), reaching

completion within one hour (Fig. 2.14). Cyclic peptides were then lyophilised without further

purification. General data on the synthesis and cyclization of these peptides are presented in Table

2.7.

CysThrXaaSerXaaArgGlyAspLeuAlaHisValThrYaaAhxCys

[O], pH 8

SH SH

CysThrXaaSerXaaArgGlyAspLeuAlaHisValThrYaaAhxCys

Figure 2. 13 Oxidation of bis-thiol precursors to cyclic peptidescyc16S30 (Xaa=Thr, Yaa=Ala) and cyc16147Val (Xaa=Ala,Yaa=Thr); air oxidation was performed at pH 8 and high dilution(50 µM) to favour intra-molecular cyclization.


91

Figure 2. 14 HPLC profiles showing both linear and cyclic forms of the different steps in the synthesis ofcyc16S30; A. crude A16S30 (bis-thiol form), B. MPLC-purified A16S30, C. air oxidation at 30 min, D.oxidation at 1 hour.

Table 2.7 Synthesis of the C-S30 cyclic peptides and their bis-thiol precursors.

PeptideGlobalyield(%)

Purity

(%HPLC)

MWfound

MWexpected AAA*

A16S30 37 97 1647.2 1647 Asp, 1.02 (1); Ser, 1.02 (1); Gly, 0.99 (1);Ala, 1.99 (2); Leu, 0.98 (1)

cyc16S30 82 95 1645.4 1645 Asp, 0.99 (1); Ser, 1.07 (1); Gly, 0.97 (1);Ala, 1.99 (2); Leu, 0.99 (1)

A16147Val 34 94 1617.9 1618 Asp, 1.05 (1); Ser, 1.01 (1); Gly, 0.99 (1);Ala, 2.76 (3); Leu, 0.97 (1)

cyc16147Val 85 93 1615.9 1616 Asp, 1.07 (1); Ser, 1.04 (1); Gly, 0.93 (1);Ala, 2.78 (3); Leu, 0.96 (1)

*Relative amino acid proportions given by AAA are followed by the expected value in parenthesis.

0 30 min

5%B 95%B

A16S30

cyc16S30

0 30 min

5%B 95%B0 30 min

5%B 95%B0 30 min

5%B 95%B

A B C D


92

2.6.2 Antigenic evaluation of cyclic C-S30 peptides using SPR

The affinities of the cyclic peptides towards the three anti-site A mAbs were measured by SPR

analysis, both with peptide as analyte and as immobilised ligand (Table 2.8 and Fig. 2.15).

Table 2.9 Interactions between peptides cyc16S30 and cyc16147Val and anti-site A mAbs.

as analyte immobilised on the chipPeptide

mAb cyc16S30 cyc16147Val cyc16S30 cyc16147Val

ka/M-1s-1 7.2x104 1.8x105 2.0x104 1.5x104

SD6 kd/s-1 1.8x10-2 4.5x10-3 4.1x10-3 1.1x10-3

KA/M-1 4.0x106 4.1x107 4.8x106 1.3x107

ka/M-1s-1 5.0x105 1.7x105 9.3x103 1.5x104

4C4 kd/s-1 3.5x10-3 5.0x10-3 8.9x10-5 2.6x10-4

KA/M-1 1.4x108 3.3x107 1.1x108 5.0x107

ka/M-1s-1 4.5x105 1.6x105 1.0x104 1.0x104

3E5 kd/s-1 5.4x10-3 2.9x10-3 2.0x10-4 4.1x10-4

KA/M-1 8.4x107 5.5x107 5.0x107 2.5x107

Before discussing the SPR data of the cyclic peptides, it must be re-emphasised that analyses using

mAb as analyte were not subject to optimisation and, therefore, data from such analyses should be

regarded as purely comparative. In fact, kinetic constants in Table 2.9 show that mass-transport

limitations are probably occurring when mAb is used as analyte, since rate constants measured

under this analysis configuration are lower than when the small peptides are the analytes11-13.

However, since both rate constants seem to be affected by mass-transport artefacts to a similar

extent, the affinities displayed are within the same order of magnitude as those measured in the

reverse configuration and, furthermore, the antigenicity ranking of the peptides is maintained in

both analysis formats.


93

A

-40

-30

-20

-10

0

10

20

30

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

cyc16S30, 39 nMcyc16S30, 78 nMcyc16S30, 156 nMcyc16S30, 312 nMcyc16S30, 625 nMcyc16S30, 1250 nM

B

-30

-25

-20

-15

-10

-5

0

5

10

15

20

25

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

cyc16Val, 39 nMcyc16Val, 156 nMcyc16Val, 312 nMcyc16Val, 625 nMcyc16Val, 1250 nM

C

-5

95

195

295

395

495

595

695

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

4C4, 15 nM4C4, 31 nM4C4, 62 nM

4C4, 125 nM4C4, 250 nM4C4, 500 nM

D

-5

15

35

55

75

95

115

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU


Figure 2. 15 Sensorgrams of the SPR kinetic analysis of the interactions between mAb 4C4 and: peptide cyc16S30 as analyte (A) and as immobilisedligand (C); peptide cyc16147Val as analyte (B) and as immobilised ligand (D).


94

Data in Table 2.9 show that cyclic versions of peptides A15S30 (i.e., peptide cyc16S30) and

A15(147Val) (i.e., peptide cyc16147Val) are clearly more antigenic than their linear counterparts

against all three mAbs. Further, the increase in affinity is reflected in both association and,

especially, dissociation rate constants, indicating that cyclic peptides bind more readily to the mAbs

and that the resulting complexes are stabilised to a greater extent. A similar result had been

observed by M. L. Valero and co-workers in the analysis of the interactions between mAb SD6 and

both peptide A15 and its corresponding cyclic disulphide analogue (AhxA16SS)19. In this previous

study, an increase of about one order of magnitude in mAb affinity was observed upon peptide

cyclization (KA from 1.9×107 to 1.2×108 M-1), but almost exclusively due to an increase in association

rate constant (ka from 3.7×103 to 2.6×104 M-1s-1). This indicated that, despite the easier entry of the

cyclic peptide into the mAb paratope, the final complex had the same half-life as the one formed

with the linear analogue (kd was 2.2×10-4 and 2.0×10-4 s-1 for the cyclic and linear peptides,

respectively).

The most striking observation made with our cyclic peptides corresponding to FMDV C-S30 and to

a hypothetical Leu147→Val mutant (a field isolate with such single-point replacement has not yet

been isolated) was the fact that, while the C-S30 sequence was less antigenic than the 147Val mutant

towards mAb SD6, this ranking was inverted when the other two mAbs were considered (the results

with mAb 4C4 being the most significant). This was precisely what had been observed in earlier

immuno-enzymatic studies of field isolate C-S30 as well as of KLH conjugates of site A peptides

displaying either the corresponding four replacements or the single Leu147→Val substitution.

Further, this was the first evidence, using small peptides, of the reversion observed in 4C4 – antigen

affinity when both Ala138→Thr and Leu147→Val replacements were brought together.

Additionally, these results confirmed what had been postulated by W. F. Ochoa and co-workers,

concerning the fact that residues 138Thr and 147Val in the A15S30-4C4 complex have only minor

contacts with the antibody paratope and, thus, differences in binding affinities observed for peptides

with replacements at these positions would be due to reduced stability of such peptides in the “mAb-

recognisable” conformation18.

A question still remained, however: peptide A15S30 can be crystallised in complex with antibody

4C4, so both molecules can undergo a considerable induced fit to form a stable complex. Therefore,

would prolonged (e.g., overnight) incubation of peptide with antibody, both species in solution,

result in higher affinities than those measured in kinetic SPR or competition ELISA assays?


2.7 Antigenic evaluation of C-S30 peptides through solution affinity SPR analysis

2.7.1 Basic concepts22,23

Measurement of affinity in solution with SPR biosensors is based on the determination of the free

concentration of one of the interactants in equilibrium mixtures. Measurements are made on known

concentrations of the free interactant for a standard curve to be built and also on the equilibrium

mixtures for determination of affinity. If binding in solution is written as:

the experiment is designed so that a constant concentration of B is incubated with a series of known

concentrations of A and, then, SPR is used to measure the remaining free concentration of B in

solution (Fig. 2.16). Such measurement is performed on a sensor chip surface where a specific

ligand for B (A’) had been previously immobilised and on which a calibration curve using known

concentrations of B had been built. The affinity constant can then be calculated with the

BIAEvaluation software24, where the variation of free B with the concentration of A is fitted to the

equation:

[ ] [ ] [ ] [ ]( ) [ ] [ ]BAKBAKAB DD ×−

+++

−−42

2

(2.1)

In the particular case of antigen –

antibody interactions (taking, for

instance, B for antibody and A for

antigen), one should ensure that

reactions take place at a 1:1

stoichiometry. Thus, antibody Fab

fragments instead of the whole

immunoglobulin must be

employed. Fab fragments of the

relevant antibodies were produced

by digestion with papain, as

described under Materials &

MethodsB (section 4.3.1.3).

B Purified Fab of SD6 was kindly supplied by Dr. Esteban Domingo and Ms. Mercedes D4C4 was a kind gift of Dr. Nuria Verdaguer and Ms. Wendy F. Ochoa (IBMB – CSIC, Bprepared by digestion of mAb isolated from ascitic fluid kindly supplied by Dr. Emiliana B

A+B AB

Chip-bound antigen

Antigen in solution

Fab fragment of specific mAb

Figure 2. 16 Main steps insolution affinity SPR analysis:pre-incubation of bothinteractants in solution(above) and injection of theequilibrium mixtures over asensor chip surface with pre-immobilised specific antigen(below).

95

ávila (CBMSO – UAM, Madrid); Fabarcelona); Fab from mAb 3E5 was

rocchi (IZSLE, Brescia – Italy).


96

The main steps in antibody digestion and Fab purification were monitored by SDS-PAGE (Fig.

2.17).

Solution affinity SPR analysis relies on concentration measurements. In the SPR biosensor,

concentration measurements are based either on binding level (avoiding bulk refractive index

contributions) or on binding rate determinations. Under conditions of limiting mass transfer of

analyte to the surface, the initial binding rate is independent of ligand density and interaction

kinetics, being exclusively determined by analyte concentration and diffusion characteristics. In the

present study, we were able to see that SPR analyses with immobilised peptide and antibody as

analyte were influenced by mass-transport artefacts even at flow rates as high as 60 µl/min (sections

2.2.3, 2.4.2 and 2.6.2). Thus, free Fab concentrations in the present study had to be derived from

initial binding rate measurements at 5 µl/min on sensor chips with peptide A15 (0.3 ng/mm2) pre-

immobilised by standard procedures (see Materials & Methods, section 4.3.1.3). Binding rates were

taken from curve slopes at a given injection time, chosen as the earliest possible where the influence

of bulk refractive index or other artefacts at injection plugs would be negligible.

It could be argued that there was no evidence that the Fab – A15 interactions were 100% diffusion-

controlled. Nevertheless, all measurements were performed on the same A15 surface and the

analyte (Fab) was always the same for each data set, so antigenicity ranking of the peptide

analogues screened under these conditions is totally meaningful.

The dependence of the measured free Fab concentration on antigen concentration gives an

inhibition curve that can be fitted to Eq. 2.1 (BIAEvaluation general fit→solution affinity model) so

that affinity is calculated (as either KA or its reciprocal KD). However, the influence of immobilised

antigen (A15) on free Fab concentration measurements is not taken into account when fitting data

as described; in fact, the immobilised peptide is competing with the soluble analogue for the same

Fab molecules as in competition ELISA. The Cheng and Prusoff’s formula (Eq. 2.2)25 can be used to

obviate this problem:

[ ]50

'1

IC

BKK A

i += (2.2)

Figure 2. 17 SDS-PAGE monitoring (at 12%acrylamide) of the digestion and purification of theFab fragment of mAb 3E5; lanes A and B: purifiedmAb 4C4 and its corresponding Fab fragment,used instead of MW protein standards; lane C:papain digestion of mAb 3E5 at 3 h of reaction;starting mAb, plus Fab2 and Fab products can bedistinguished; lane D: Fab fragment of mAb 3E5after a two-step (affinity and gel filtrationchromatography) purification.

A B C D


97

where Ki is the “real” affinity for the interaction of A and B in solution, K’A is the affinity for the

interaction of B with immobilised A’ and IC50 is the concentration of A causing a 50% drop in the

total concentration of B.

2.7.2 Results

Injection of known Fab (SD6, 4C4 and 3E5) standards on the A15 surface allowed the building of

calibration curves (Fig. 2.18), which were subsequently employed in the determination of Fab

molecules that remained free after overnight incubation with peptide antigens in solution.

0

0.5

1

1.5

2

2.5

3

3.5

0.00E+00 5.00E-08 1.00E-07 1.50E-07 2.00E-07 2.50E-07 3.00E-07 3.50E-07

Fab concentration/M

Slo

pe/

RU

.s-1

data points (SD6)data points (4C4)data points (3E5)fitted curve (SD6)fitted curve (4C4)fitted curve (3E5)

Figure 2. 18 Plots of initial binding rate/RU.s-1 vs. Fab concentration (M) for the three antibodiesemployed in the present study. Measurements were made at a 5 µl/min flow rate on a 0.3 ng/mm2 A15surface; data points were fitted to a four-parameter equation using BIAEvaluation software in order tobuild the corresponding calibration curves.

The quantification of remaining free Fab in solution for each incubated mixture (where Fab total

concentration is constant and peptide antigen concentrations varied) allowed to build inhibition

curves (Fig. 2.19) from which the peptide – antibody solution affinities were calculated, either

through direct fitting to the BIAEvaluation solution affinity model (Eq. 2.1) or using the Cheng &

Prusoff’s formula (Eq. 2.2). Results are summarised in Table 2.9.


98

0.00E+00

1.00E-08

2.00E-08

3.00E-08

4.00E-08

5.00E-08

6.00E-08

7.00E-08

8.00E-08

9.00E-08

1.00E-07

0.00E+00 1.00E-07 2.00E-07 3.00E-07 4.00E-07 5.00E-07 6.00E-07 7.00E-07


Fab

co

nce

ntr

atio

n/M

A15

A15Brescia

A15(147V)

A15S30

A21S30

cyc16val

cyc16S30

A15Scr

Figure 2. 19 Inhibition curves obtained in the SPR analysis of the interactions between Fab 4C4 andpeptides A15, A15Brescia, A15(147Val), A15S30, A21S30, cyc16S30, cyc16147Val and A15Scr in solution;a constant total Fab concentration was used and competitor peptide concentrations were varied from 0 to625 nM.

Table 2.9 Affinity data of interactions between peptides and mAbs in solution.

Fab Peptide KA (solutionaffinity fit)a/M-1

Ki (Cheng & Prusoff’s)b/M-1

KA (kinetic analysis)c/M-1

A15 4.3×107 6.3××××107 5.4×107

A15(147V) 3.6×106 4.5××××106 1.0×107

A15Brescia 4.3×107 6.5××××107 1.2×107

SD6 A15S30 6.0×105 ND 4.3×105

A21S30 1.6×106 ND 3.0×106

cyc16S30 2.8×106 7.5××××106 4.0×106

cyc16147Val 5.5×107 5.0××××107 4.1×107

A15 8.2×107 2.0××××108 1.9×108

A15(147V) 3.8×106 2.8××××107 2.2×106

A15Brescia 5.5×107 1.5××××108 1.6×108

4C4 A15S30 2.3×106 1.1××××107 2.0×106

A21S30 5.2×106 3.8××××107 5.6×106

cyc16S30 1.4×108 1.8××××108 1.4×108

cyc16147Val 2.7×107 8.2××××107 3.3×107

A15 1.6×108 2.0××××108 9.4×107

A15(147V) 4.5×106 2.8××××107 6.6×106

A15Brescia 6.5×107 1.1××××108 1.0×108

3E5 A15S30 2.1×106 8.2××××106 4.5×106

A21S30 2.9×106 1.1××××107 5.6×106

cyc16S30 4.4×107 1.1××××108 8.4×107

cyc16147Val 4.5×107 8.5××××107 5.5×107

a Direct curve fitting with the BIAEvaluation software.b Application of the Cheng & Prusoff’s formula.c Data from previous SPR kinetic assays.ND, not determined.


99

2.7.3 Discussion

SPR measurement of affinities of C-S8c1, C1-Brescia and C-S30 peptides for anti-site A mAbs SD6,

4C4 and 3E5 provided a conclusive confirmation of the data discussed in the present chapter.

Indeed, as can be seen in Fig. 2.20 and Table 2.9, antigenicity rankings observed for these

interactions were in good agreement with those previously obtained by competition ELISA and

kinetic SPR analysis. Peptide A15S30 was less antigenic than its counterparts from strains C-S8c1

(peptide A15) or C1-Brescia (A15Brescia). At the same time, important differences were observed

and are discussed in the following paragraphs.

i) Solution affinities obtained by curve fitting or by the Cheng & Prusoff’s formula

Comparing the affinities in Table 2.9 as obtained by solution affinity fit or by the Cheng & Prusoff’s

formula, the influence of immobilised antigen A15 becomes quite evident. Thus, Eq. 2.1 describes

phenomena such as those occurring in competition ELISA. In this case, peptides with lower

antigenicity will be the most affected by the immobilised antigen competitor and fitted affinities will

be lower than affinities determined in solution. In contrast, the Cheng & Prusoff’s formula (Eq. 2.2)

allows to obtain data independent from the immobilised antigen and major differences between

both methods of affinity calculation can be observed for the least antigenic peptides A15(147V),

A15S30 and A21S30 towards mAbs 4C4 and 3E5. Differences of about one order of magnitude can

be found in these cases, relative to affinities calculated by direct fit of the inhibition curves.

ii) Solution affinity versus kinetic SPR data

Comparing affinity data calculated from the inhibition curves using Eq. 2.2 with previous data

obtained by kinetic SPR analysis, an excellent agreement is observed with three important

exceptions: peptides A15(147V), A15S30 and A21S30 (towards mAbs 4C4 and 3E5). Even though

relative ranking of all antigens is maintained, an increase in affinity of about one order of magnitude

is measured in solution equilibrium experiments involving these peptides. The fact that such

observation is more pronounced for these three particular peptide mutants appears to be quite

significant. Indeed, it seems that when antigen and antibody are allowed to interact overnight both

free in solution, they can rearrange so that more stabilised complexes are formed. Thus, the lower

affinities measured in kinetic SPR analysis would possibly be due to interaction times (1.5 min) too

small for such conformational changes to be detected. This ability of the C-S30 GH loop to

rearrange into a mAb-recognisable structure, leading to stable antibody-antigen complexes, could be

the basis for the recognition and neutralisation of FMDV C-S30 by mAb 4C4.


100

iii) A role for peptide conformation

The above observations are further supported by the modulation of peptide antigenicity upon

cyclization. Both four-point and one-point (147Leu→Val) replaced sequences produced important

increases in mAb affinities when presented as cyclic peptides. Particularly in the case of mAb 4C4, it

was observed that the cyclic C-S30 sequence was about one order of magnitude more antigenic

than the cyclic one-point mutant 147Leu→Val, which confirms previous studies suggesting a positive

reversion in the antigenicity of the four-point mutant1-7. These observations have important

implications vis à vis the simplistic view of continuous antigenic sites as conformation-independent.

If this was in fact the case, peptide cyclization would have only minor effects on antigenicity. Further,

the loss of antigenicity due to amino acid replacements in positions that are not in direct contact with

the antibody paratope18 (e.g., 138Ala→Thr and 147Leu→Val in linear peptides) cannot be easily

explained by factors other than peptide conformation.


101

2.8 Two-dimensional proton nuclear magnetic resonance studies of C-S30 peptides26

Most linear and many cyclic peptides possess a high conformational flexibility, meaning that they

can adopt a fairly large set of interconvertible conformations in solution. In a few favourable cases,

the population of one conformer is large enough to be distinctly detected by spectroscopy. Among

the techniques commonly employed for the detection and identification of peptide conformation

such as circular dichroism, Raman spectroscopy, FT-IR and NMR, the latter has a clear advantage in

that it allows not only the global detection of the preferred structure, but also the characterisation of

the individual amino acid residues defining such structure. Thus, it provides a picture of peptide

folding in aqueous solution, which is desirable when searching for structure – activity relationships.

Due to their natural abundance, gyromagnetic constant and localisation in peptides, protons are the

best probes for peptide conformational NMR studies. Although a simple one-dimensional NMR

spectrum of the peptide should always be obtained in order to check for peptide purity,

concentration or aggregation, the level of complexity is usually too high for a complete assignment.

So, two-dimensional NMR experiments are needed for a full and unequivocal proton assignment in

peptide studies.

Peptides A15S30 and cyc16S30 were both studied by 2D – 1H NMR in solution, in an attempt to

define secondary structure elements that could explain the antigenic features of both peptides. NMR

experiments (TOCSY27, NOESY28 and ROESY29) were carried out both in aqueous solution and in

the presence of the structuring agent trifluoroethanol, as described under Materials & Methods

(section 4.4.1).

2.8.1 Basic concepts

2D – 1H NMR spectra of peptides are interpreted according to the sequential assignment method

developed by Wüthrich for proteins27. The first step in the procedure relies on the total correlation

scalar experiment (TOCSY), in which peaks are detected for protons that can correlate with each

other by means of a magnetisation transfer sequence involving, at each step, H – H couplings. This

allows the identification of each amino acid residue independently from all the others in the

sequence, since magnetisation cannot be transferred through the amide bond from one residue to

the following one. Thus, each amide proton (HN) will be correlated with all other protons from the

same spin system; the number and chemical shift of such protons provide an identification of the

amino acid residue in question. If the peptide includes an amino acid residue that is unique in the

whole sequence, the assignment is immediate and unequivocal. However, if a certain amino acid

residue is repeatedly present along the sequence, it will be necessary to carry out another kind of

NMR experiment where protons from different amino acid residues are correlated. Such correlation

is based on nuclear Overhauser effect (NOE) experiments (NOESY28, ROESY29). The NOE arises

from the dipolar relaxation that occurs between two nuclei that are spatially close to each other,


102

regardless of their belonging or not to the same spin system. An NOE between a pair of protons is,

therefore, observed when there is a population of peptide structures where both nuclei are within

4.5 Å from each other. Thus, when all residues are attributed a set of signals in the TOCSY

spectrum, the second step will be the use of NOE experiments to establish connectivities among

them. This will be possible if sequential distances such as dαNi,i+1, dNni,i+1, dβNi,i+1, etc., can be

observed, i. e., correlations between proton Hα of residue i with HN of residue i+1, or HN of residue i

with HN of residue i+1, or Hα of residue i with HN of residue i+1, respectively. Since the α proton of

a given residue is usually close in space to the HN of the following residue, sequential dαNi,i+1 NOEs

are useful to assign the amino acid sequence.

Given the slow time-scale of NMR spectroscopy relative to optical spectroscopy, one must keep in

mind that NMR spectral parameters are all averaged. Thus, all conformational information provided

by NMR corresponds to the average of all structures adopted by the peptide in solution. Of the

several parameters that can be used in NMR peptide structural studies, only conformational chemical

shifts and NOEs have been considered in the present work.

Chemical shifts are the most easily measured NMR parameters and are quite susceptible to subtle

changes in the chemical environment of the proton. The large number of protein structures assigned

by NMR made possible a statistical study correlating differences in chemical shifts with peptide

secondary structure. The most useful chemical shift differences have been found to be those

between the Hα of folded and random coil structures, the latter ones derived from model

oligopeptides. Thus, Hα conformational chemical shift differences (defined as ∆δHα=δexp – δ random coil)

are found to be negative for helices (average, - 0.39 ppm) and β turns, and positive for β sheets

(average, + 0.37 ppm). NOEs usually provide the most unambiguous information about peptide

structure. The sole observation of a NOE between two protons implies that there are conformational

populations in which these two protons are spatially close (d ≤ 4.5 Å), independently of their being

or not close in the primary sequence. Energy studies on the conformational space of proteins allow

to establish NOE patterns that correlate with peptide secondary structure. A general guide for

structural interpretation of NOEs is presented in Table 2.10.

Table 2.10 Useful NOEs for the identification of peptide secondary structure elements.

StructureNOE Helix ββββ sheet ββββ turn

αααα 310 I IIdααααNi,i+1

dNNi,i+1 (2) (1)

dNNi,i+2

dααααNi,i+2

dααααNi,i+3

dααααNi,i+4

dαβαβαβαβi,i+3

- expected NOE (the number of ticks corresponds to the expected intensities); - not observed.


103

2.8.2 Results

Sequence assignment based on TOCSY and NOESY/ROESY spectra is shown in Fig. 2.20 for both

A15S30 and cyc16S30 peptides in water. Chemical shifts observed for both peptides, in water as

well as in 30% TFE, are presented in Table 2.11.

145A

150T

147V 144L

137T

149A

141R

139S

143D

146H142G

140T

148T

138T

139S151C

142G

141R

143D

146H

150Ahx

147V145A

149A

144L

140T

148T138T 137T

Figure 2. 20 Expansion of the TOCSY experiments (70 ms) of peptides A15S30 (above) andcyc16S30 (below) performed at 25 oC in water; the different spin systems are indicated and wereassigned upon analysis of TOCSY, NOESY and ROESY spectra.


104

Table 2.11 Chemical shifts (ppm) measured in the 2D – 1H NMR study of peptides A15S30 and cyc16S30 in water and in 30% TFE at 25 oC.

A15S30 cyc16S30Residue H2O 30% TFE H2O 30% TFE

HN Hα Hβ Other HN Hα Hβ Other HN Hα Hβ Other HN Hα Hβ Other136Tyr/Cys nd 4.32 3.15

3.19nd nd 4.32 3.13

3.20nd nd 4.48 3.34

3.23ne nd 4.48 3.40

3.26ne

137Thr 8.49 4.47 4.15 1.20 8.40 4.52 4.20 1.24 8.34 4.42 4.25 1.20 8.03 4.22 4.18 1.28138Thr 8.36 4.32 4.25 1.24 8.25 4.41 4.32 1.28 8.29 4.38 4.28 1.20 8.25 4.45 4.31 1.25139Ser 8.43 4.55 3.86

3.92ne 8.33 4.60 3.91

3.98ne 8.43 4.57 3.88 ne 8.33 4.60 3.96

3.90ne

140Thr 8.27 4.37 4.27 1.20 8.18 4.40 4.31 1.25 8.36 4.40 4.24 1.20 8..18 4.43 4.33 1.24141Arg 8.35 4.32 1.89

1.773.301.66

8.28 4.30 1.901.80

3.201.70

8.36 4.34 1.891.76

3.221.64

8.28 4.36 1.931.80

3.221.70

142Gly 8.40 3.93 ne ne 8.32 3.95 ne ne 8.44 3.94 ne ne 8.34 3.97 ne ne143Asp 8.32 4.66 2.85

2.80ne 8.21 4.65 2.86 ne 8.25 4.61 2.73 ne 8.15 4.67 2.83 ne

144Leu 8.19 4.29 1.61 0.920.85

8.08 4.26 1.71 0.950.90

8.16 4.24 1.66 0.920.84

8.14 4.24 1.69 0.950.90

145Ala 8.12 4.23 1.32 ne 8.02 4.22 1.38 ne 8.10 4.18 1.28 ne 8.03 4.20 1.34 ne146His 8.36 4.71 3.15

3.26nd 8.09 4.67 3.37

3.23nd 8.16 4.70 3.30

3.14nd 8.05 4.69 3.43

3.20nd

147Val 8.19 4.18 2.07 0.92 8.05 4.12 2.20 1.00 8.09 4.17 2.10 0.94 8.04 4.18 2.19 1.00148Thr 8.33 4.33 4.19 1.21 8.07 4.37 4.29 1.25 8.26 4.32 4.21 1.20 8.07 4.36 4.29 1.24149Ala 8.44 4.41 1.41 8.12 4.40 1.47 8.18 4.25 1.36 7.91 4.32 1.41

150Thr/Ahx 8.11 4.29 4.24 1.20 7.87 4.33 nd 1.23 7.84 3.15 1.63 1.301.502.33

7.66 3.22 1.65 1.351.542.35

151----/Cys - - - - - - - - 8.44 4.68 3.303.00

8.25 4.76 3.323.04

ne, non existing; nd, not determined.


105

Conformational chemical shifts

The first structural information derived from the NMR experiments was based on the conformational

chemical shifts plotted in Fig. 2.21. As can be seen, neither peptide shows any marked tendency to

adopt a predominant canonical (helix or β-sheet) conformation in solution. The small absolute

values of the conformational chemical shifts suggest that both peptides exist mainly in aperiodic

(random coil) form in solution. This lack of structuration could not be modified by environmental

changes, as shown by the similarity between conformational chemical shifts observed in water and

in 30% TFE.

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

X1 T T S T R G D L A H V T A X2 X3

residue

∆δ ∆δ∆δ∆δH

α ααα/p

pm

cyc16S30, 30% TFEcyc16S30, waterA15S30, 30% TFE

A15S30, water

Figure 2. 21 Conformational chemical shifts (∆δHα) observed at 25 oC for peptides A15S30 (Xaa=Tyr,Yaa=Thr, Zaa= - ) and cyc16S30 (Xaa=Cys, Yaa=Ahx, Zaa=Cys) both in water and in 30% TFE.

Nevertheless, a slight tendency to structuration in the central region of the peptides can be

distinguished. As previously observed by T. Haack and co-workers for the linear (A15) and cyclic

disulphide (AhxA16SS) versions of FMDV C-S8c119,30, the conformational chemical shifts in the

region that includes the RGD tripeptide are compatible with a tendency for an open turn

conformation. Further, the cluster of negative conformational chemical shifts from 143Asp to 146His

could be suggestive of an incipient short helix in this region, as previously observed for the C-S8c1

peptides by the same authors. A significant difference between the peptides of both strains is,

however, the fact that this short helix extends, in the C-S8c1 peptides, to the 147Leu residue. In the

case of the C-S30 peptides, the replacement of leucine by valine at this position seems to shorten

this pre-helical stretch, and this could be related to the lower antigenicities observed in peptides

including this replacement. This seems further supported by the fact that absolute values of

conformational chemical shifts in this region are higher for cyclic peptide cyc16S30, which is the

best of the two C-S30 antigens under study.


106

Structural information from NOEs

Data from NOE experiments were not particularly conclusive. A continuous series of αNi,i+1 and

βNi,i+1 NOEs was observed for both peptides either in water or in 30% TFE (Fig. 2.22). These NOEs

are compatible with practically all conformations and therefore not very informative. Some weak

NNi,i+1 NOEs – compatible with α helix – were observed for peptide A15S30 in 30% TFE, in the

same region where a tentative assignment of incipient helix had been made (see previous section).

In turn, peptide cyc16S30 displayed weak NNi,i+1 NOEs both in water and 30% TFE (Fig. 2.23).

Peak assignment was often ambiguous due to identical chemical shifts for different HN and peak

overlap possibly prevented the detection of other informative HN – HN connectivities.

A15S30 Y T T S T R G D L A H V T A T

NN(i,i+1)ΗΗΗΗ2222ΟΟΟΟ αN(i,i+1)

βN(i,i+1)

NN(i,i+1)30% TFE αN(i,i+1)

βN(i,i+1)

cyc16S30 C T T S T R G D L A H V T A X C

NN(i,i+1)αN(i,i+1)

ΗΗΗΗ2222ΟΟΟΟ NN(i,i+2)βN(i,i+1)

NN(i,i+1)30% TFE αN(i,i+1)

NN(i,i+2)βN(i,i+1)

Figure 2. 22 Summary of the NOEs observed for peptides A15S30 and cyc16S30 in water and 30% TFE;relative NOE intensities are represented by bar thickness; dotted lines stand for overlapping or ambiguousNOEs.

149A - 150T

143D - 144L

145A - 146H

142G - 143D

140T - 141R

142G - 143D

145A - 146H

149A - 150Ahx

138T - 139S ?

147V - 148T ?

151C - 137T ?

Figure 2. 23 Expansion of the ROESY experiments (200 ms) performed at 25 oC: peptide A15S30 in 30%TFE (left) and peptide cyc16S30 in water (right). Diagonal peaks (along the dashed lines) were omitted forsimplicity.


107

2.9 Recapitulation

The work described in the present chapter involved a major goal:

Finding out why FMDV C-S30 was recognised and neutralised by anti-site A antibodies such as 4C4,

even though it possesses, within this site, mutations known to be detrimental for mAb recognition

(e.g., 147L→V).

To accomplish such purpose, a total of sixteen peptides were synthesised and studied by SPR.

ELISA and NMR analyses were also performed on some of these peptides, to further complement

the study of the GH loop from FMDV C-S30. Extensive research previously reported on 15-mer

peptide mimics of FMDV antigenic site A provided a solid basis for the adequacy of such peptides as

models of this antigenic site.

The first set of eleven peptides corresponded to linear pentadecapeptides reproducing all possible

combinations of the four mutations (138A→T, 140A→T, 147L→V, 149T→A) present in FMDV C-S30

antigenic site A (taking as reference sequence the antigenic site A of FMDV C- S8c1).

A direct kinetic SPR analysis of the mAb – peptide interactions was performed according to a

protocol previously optimised and validated with similar peptide FMDV antigens (chapter 1). Results

pointed to additive effects in all combinations of the four relevant mutations. For each mAb,

association rate constants were virtually equal, while dissociation rate constants varied in a relatively

broad range, increasing with decreasing peptide antigenicity.

The four-point mutant linear 15-mer peptide from the C-S30 GH loop (A15S30) was shown to be

the least antigenic of the set.

The surprisingly low antigenicity of peptide A15S30 led to a competition ELISA screening of all 15-

mer peptides, in order to confirm the SPR data.

Peptide A15S30 was again shown to be a poor competitor in this format of analysis, thus

confirming its low antigenicity relative to C-S8c1 or C1-Brescia peptides.

Having confirmed that the unexpected results obtained for A15S30 were not due to any technical

artefact from the SPR biosensor, our attention was then focused on the peptide itself: was the

peptide too short?

Two 21-mer linear peptide models of the C-S8c1 and C-S30 GH loops were then studied by kinetic

SPR analysis and competition ELISA.

Peptide C-S30 was, once again, clearly less antigenic than peptide C-S8c1.


108

At this point of the investigation, parallel studies performed by Wendy F. Ochoa and co-workers

showed that the 15-mer peptide A15S30 could be crystallised in complex with the Fab fragment of

mAb 4C4. The peptide adopted a nearly cyclic conformation similar to the one previously described

for the C-S8c1 peptide A15 in a similar complex. Further, it was observed that the two more critical

mutations (138A→T and 147L→V), although not in direct interaction with the antibody, were both

involved in keeping the peptide in a pseudo-cyclic conformation, through hydrogen bonding

involving the Thr side chain hydroxyl, one water molecule and the main chain oxygen atoms of144Leu and 147Val.

So, a new question was immediately raised: Were the linear C-S30 peptides too flexible in solution?

Two cyclic disulphide peptides were thus studied by SPR, one of them reproducing the C-S30

sequence (cyc16S30) and the other containing the one-point mutation 147L→V (cyc16147Val) in

order to analyse the effects of conformation modulation on peptide antigenicity.

An increase in peptide affinity was observed upon cyclization. While the single-point was still

more antigenic than the four-point mutant towards mAb SD6, a reversion in this ranking was

observed with mAb 4C4.

The higher antigenicity of C-S30 cyclic peptides and the results obtained by Wendy F. Ochoa with

the linear C-S30 peptide led to another question: Would the flexible peptide A15S30 be able to

rearrange in order to form a stable complex with anti-site A mAb 4C4 in solution, after prolonged

incubation?

To answer this question, a solution affinity SPR experiment was performed using peptides

A15(147V), A15S30, cyc16S30 and cyc16147Val as target analytes and peptides A15 and

A15Brescia as reference analytes. Despite confirming that linear C-S30 peptides were again less

antigenic than their C-S8c1 or C1-Brescia counterparts after overnight incubation with mAb in

solution, a significant increase (about one order of magnitude) in 4C4 affinity towards peptide

A15S30 was observed. This suggests that, indeed, incubation of 4C4 with A15S30 can lead the

peptide to rearrange and form a stable complex with antibody.

Solution conformation NMR studies of both A15S30 and cyc16S30 peptides, though not totally

conclusive, were quite suggestive. Both peptides were seen to be very flexible in solution, even in

the presence of conformation-inducing solvents. Nevertheless, both displayed a tendency for an

open turn in the RGD region, followed by an incipient short helical path, as previously observed

for C-S8c1 linear and cyclic peptides. Remarkably, while in the C-S8c1 peptides this helical

stretch extends up to the 147Leu, in the C-S30 peptides it stops at the 146His, suggesting that a147Leu→→→→Val helix-disruptive mutation could be the basis for the lower antigenicites observed in

peptides including this mutation. Further, this short helix is more pronounced in the cyclic model

of the C-S30 GH loop, which can be a reason for the higher antigenicities observed for this

peptide.


109

References

1 Mateu, M. G., Rocha, E., Vicente, O., Vayreda, F., Navalpotro, C., Andreu, D., Pedroso, E., Giralt, E.,Enjuanes, L. and Domingo, E. (1987) Reactivity with monoclonal antibodies of viruses from anepisode of foot-and-mouth disease, Virus Res. 8, 261-274.

2 Mateu, M. G., da Silva, J. L., Rocha, E., de Brum, D. L., Alonso, A., Enjuanes, L., Domingo, E. andBarahona, H. (1988) Extensive antigenic heterogeneity of foot-and-mouth disease virus serotype C,Virology 167, 113-124.

3 Mateu, M. G., Martínez, M. A., Andreu, D., Parejo, J., Giralt, E., Sobrino, F. and Domingo, E. (1989)Implications of a quasispecies genome structure: effect of frequent, naturally occurring, amino acidsubstitutions on the antigenicity of foot-and-mouth disease virus, Proc. Natl. Acad. Sci. USA 86,5883-5887.

4 Mateu, M. G., Martínez, M. A., Cappucci, L., Andreu, D., Giralt, E., Sobrino, F., Brocchi, E. andDomingo, E. (1990) A single amino acid substitution affects multiple overlapping epitopes in themajor antigenic site of foot-and-mouth disease virus of serotype C, J. Gen. Virol. 71, 629-637.

5 Martínez, M. A., Hernández, J., Piccone, M. E., Palma, E. L., Domingo, E., Knowles, N. and Mateu,M. G. (1991) Two mechanisms of antigenic diversification of foot-and-mouth disease virus, Virology184, 695-706.

6 Mateu, M. G., Andreu, D., Carreño, C., Roig, X., Cairó, J. J., Camarero, J. A., Giralt, E. andDomingo, E. (1992) Non-additive effects of multiple amino acid substitutions on antigen-antibodyrecognition, Eur. J. Immunol. 22, 1385-1389.

7 Carreño, C., Roig, X., Camarero, J. A., Cairó, J. J., Mateu, M. G., Domingo, E., Giralt, E. andAndreu, D. (1992) Studies on antigenic variability of C-strains of foot-and-mouth disease virus bymeans of synthetic peptides and monoclonal antibodies, Int. J. Peptide Protein Res. 39, 41-47.

8 Fields, G. B. and Noble, R. L. (1990) Solid-phase peptide synthesis utilizing 9-fluorenylmethoxycarbonyl amino acids, Int. J. Peptide Protein Res. 53, 161-214.

9 Knorr, R., Trzeciak, A., Bannwarth, W. and Gillesen, D. (1989) New coupling reagents in peptidechemistry, Tetrahedron Lett. 30, 1927-1930.

10 Bernatowicz, M. C., Daniels, S. B. and Köster, H. (1989) A comparison of acid labile linkage agentsfor the synthesis of peptide C-terminal amides, Tetrahedron Lett. 30, 4645-4648.





15 Mateu, M. G., personal communication16 Abbas, A. K., Lichtman, A. H. and Pober, J. S. “Cellular and molecular immunology”, 3rd ed., W. B.

Saunders Co., United States of America (1997).17 Mateu, M. G., Andreu, D. and Domingo, E. (1995) Antibodies raised in a natural host and

monoclonal antibodies recognize similar antigenic features of foot-and-mouth disease virus, Virology210, 120-127.

18 Ochoa, W. F., Kalko, S., Mateu, M., Gomes, P., Andreu, D., Domingo, E., Fita, I. and Verdaguer, N.(2000) A multiply substituted GH loop from foot-and-mouth disease virus in complex with aneutralizing antibody: a role for water molecules, J. Gen. Virol. 81, 1495-1505.


20 Andreu, D., Albericio, F., Solé, N. A., Munson, M. C., Ferrer, M. and Barany, G. Formation ofdisulfide bonds in synthetic peptides and proteins in “Methods in molecular biology, vol. 35: Peptidesynthesis protocols”, Pennington, M. W. and Dunn, B. M. (Eds.), Humana Press Inc., Totowa, NewJersey (1994), pp 91-169.

21 Ellman, G. L. (1958) A colorimetric method for determining low concentrations of mercaptans, Arch.Biochem. Biophys. 74, 443-450.

22 “BIAapplications Handbook”, (Pharmacia Biosensor AB, 1994) Uppsala, Sweden.23 Nieba, L., Krebber, A. and Plükthun, A. (1996) Competition BIAcore for measuring true affinities:

large differences from values determined from binding kinetics, Anal. Biochem. 234, 155-165.


110

24 “BIAevaluation Software Handbook: version 3.0”, (Biosensor AB, 1997) Uppsala, Sweden.25 Lazareno, S. and Birdsall, N. J. (1993) Estimation of competitive antagonist affinity from functional

inhibition curves using the Gaddum, Schild and Cheng-Prusoff equations, British J. Pharmacol. 109,1110-1119.

26 Wüthrich, K. “NMR of proteins and nucleic acids”, Wiley, New York (1986).27 Braunschweiler, L. and Ernst, R. R. (1983), J. Magn. Reson. 53, 521.28 Kumar, A., Ernst, R. R. and Wüthrich, K. (1980), Biochem. Biophys. Chem. Comm. 95, 1.29 Bothner-By, A. A., Stephens, R. L., Lee, J., Warren, C. D. and Jeanloz, R. W. (1984) Structure

determination of a tetrasaccharide: transient nuclear overhauser effects in the rotating frame, J. Am.Chem. Soc. 106, 811-813.


3. Antigenic peptides with non-natural

replacements within the GH loop of FMDV


112

Antigenic peptides with non-natural replacements within the GH loop of FMDV

3.0 Introduction

In an attempt to analyse the contribution of each amino acid residue to the antigenicity of site A of

FMDV C-S8c1, M. L. Valero and co-workers evaluated a set of 250 peptides corresponding to the

systematic replacement of all residues within the sequence of peptide A151-3. Peptide antigenicity

was quantitated by competition ELISA, using a panel of seven anti-site A mAbs: SD6, 4C4, 5A2,

6D11, 7DJ1, 7FC12 and 7CA11. In this systematic screening, five singly replaced peptides were

found to be antigenic for at least three mAbs, being comparable to or even better than the native

A15 sequence. These peptides corresponded to the substitutions Thr137→Ile, Ala138→Phe,

Ala140→Pro, Gly142→Ser and Thr148→Ile. Although the first two and the last residue

replacements correspond to hyper-variable regions of the GH loop, the contrast in size between the

Phe and the Ala residues, the structural “personality” of Pro and, even more, the Gly→Ser mutation

within the highly conserved RGD motif, altogether raise many questions about what would be the

contribution of these residues in antibody recognition.

In the present chapter, further studies on the above mentioned amino acid replacements are

presented. Both the one-point mutants (Table 3.1), and a set of peptides reproducing all possible

combinations of the five mutations were synthesised and antigenically characterised by SPR. Some

NMR and X-ray diffraction studies were also performed, as a structural complement to the functional

SPR characterisation of the peptide antigens.

3.1 Peptides that combine antigenicity-enhancing replacements in the GH loop

Thirty-one peptides, corresponding to the combinations of the five mutations (Table 3.1), were

synthesised (Fig. 3.1, Table 3.2) by methods similar to those mentioned in chapter 2 and further

described in section 4.2 (Materials & Methods).

MPLC

0 30 min 0 30 min

Figure 3. 1 HPLC of crude(left) and purified (right)peptide A15(138F,140P,142S),representative of the thirty-one15-mers of this chapter.

113


114

Table 3.1 Pentadecapeptide library reproducing all possible combinations of the substitutions Thr137→Ile,Ala138→Phe, Ala140→Pro, Gly142→Ser and Thr148→Ile.

Name Sequence Mutants

A15 YTASARGDLAHLTTT GH loop of FMDV C-S8c1

A15(137I) -I-------------A15(138F) --F------------A15(140P) ----P---------- One-pointA15(142S) ------S--------A15(148I) ------------I--A15(137I,138F) -IF------------A15(137I,140P) -I--P----------A15(137I,142S) -I----S--------A15(137I,148I) -I----------I--A15(138F,140P) --F-P---------- TwoA15(138F,142S) --F---S-------- PointA15(138F,148I) --F---------I--A15(140P,142S) ----P-S--------A15(140P,148I) ----P-------I--A15(142S,148I) ------S-----I--A15(137I,138F,140P) -IF-P----------A15(137I,138F,142S) -IF---S--------A15(137I,138F,148I) -IF---------I--A15(137I,140P,142S) -I--P-S--------A15(137I,140P,148I) -I--P-------I-- ThreeA15(137I,142S,148I) -I----S-----I-- PointA15(138F,140P,142S) --F-P-S--------A15(138F,140P,148I) --F-P-------I--A15(138F,142S,148I) --F---S-----I--A15(140P,142S,148I) ----P-S-----I--A15(137I,138F,140P,142S) -IF-P-S--------A15(137I,138F,140P,148I) -IF-P-------I--A15(137I,138F,142S,148I) -IF---S-----I-- Four-pointA15(137I,140P,142S,148I) -I--P-S-----I-- “A15(138F,140P,142S,148I) --F-P-S-----I-- “

A15(137I,138F,140P,142S,148I) -IF-P-S-----I-- Five-point


115

Table 3.2 General data (yield and product characterisation by HPLC, ES or MALDI-TOF MS and AAA) of the syntheses of the 15-mer peptides.Peptide Global

yield (%)Purity

(% HPLC)MW

foundMW

expectedAAA

A15(137I) 89 99 1589.2 1589 Asp, 1.07 (1); Ser, 1.04 (1); Gly, 1.08 (1); Ala, 3.10 (3); Leu, 1.96 (2); His, 0.92 (1)A15(138F) 65 97 1653.3 1653 Asp, 0.96 (1); Ser, 0.96 (1); Gly, 1.03 (1); Ala, 2.05 (2); Leu, 2.05 (2); His, 0.99 (1)A15(140P) 84 98 1603.1 1603 Asp, 1.04 (1); Ser, 0.97 (1); Gly, 1.05 (1); Ala, 2.07 (2); Leu, 1.96 (2); Arg, 1.01 (1)A15(142S) 79 99 1607.2 1607 Asp, 1.05 (1); Ser, 1.99 (2); Ala, 3.08 (3); Leu, 2.00 (2); His, 0.92 (1); Arg, 0.95 (1)A15(148I) 87 98 1589.1 1589 Asp, 1.05 (1); Ser, 1.01 (1); Gly, 1.04 (1); Ala, 3.03 (3); Leu, 1.82 (2); Arg, 0.94 (1)A15(IF)* 15 93 1665.1 1665 Asp, 1.00 (1); Ser, 0.98 (1); Gly, 1.03 (1); Ala, 1.98 (2); Leu, 1.98 (2); His, 1.02 (1)A15(IP) 55 91 1615.6 1615 Asp, 1.06 (1); Ser, 1.00 (1); Gly, 1.10 (1); Ala, 1.93 (2); Pro, 0.99 (1); Arg, 1.02 (1)A15(IS) 79 90 1619.2 1619 Asp, 0.99 (1); Ser, 2.07 (2); Ala, 3.08 (3); Leu, 2.03 (2); His, 0.88 (1); Arg, 0.95 (1)A15(II) 78 97 1601.3 1601 Asp, 1.00 (1); Ser, 0.99 (1); Gly, 1.01 (1); Ala, 3.04 (3); His, 0.92 (1); Arg, 0.99 (1)A15(FP) 72 95 1679.6 1679 Asp, 1.06 (1); Pro, 1.01 (1); Gly, 1.09 (1); Ala, 1.05 (1); Leu, 1.89 (2); Arg, 1.07 (1)A15(FS) 86 89 1684.1 1684 Asp, 1.05 (1); Ser, 2.10 (2); Ala, 2.05 (2); Leu, 1.88 (2); His, 0.92 (1); Arg, 1.00 (1)A15(FI) 86 91 1665.4 1665 Asp, 1.03 (1); Ser, 0.95 (1); Gly, 1.03 (1); Ala, 2.00 (2); His, 0.90 (1); Arg, 1.09 (1)A15(PS) 81 87 1632.4 1633 Asp, 1.02 (1); Pro, 1.03 (1); Ala, 2.04 (2); Leu, 1.93 (2); His, 0.92 (1); Arg, 1.11 (1)A15(PI) 79 89 1615.6 1615 Asp, 1.03 (1); Ser, 0.95 (1); Gly, 1.00 (1); Ala, 2.01 (2); His, 0.95 (1); Arg, 1.07 (1)A15(SI) 79 86 1619.6 1619 Asp, 1.05 (1); Ser, 1.96 (2); Ala, 3.10 (3); Tyr, 0.87 (1); His, 0.92 (1); Arg, 1.14 (1)A15(IFP) 78 94 1691.6 1691 Asp, 1.03 (1); Ser, 0.96 (1); Gly, 1.05 (1); Ala, 1.01 (1); Pro, 0.98 (1); Arg, 1.05 (1)A15(IFS) 95 86 1694.5 1695 Asp, 1.10 (1); Ser, 1.90 (2); Tyr, 0.91 (1); Phe, 1.01 (1); His, 0.97 (1); Arg, 1.08 (1)A15(IFI) 83 95 1677.4 1677 Asp, 1.01 (1); Ser, 0.91 (1); Gly, 1.01 (1); Ala, 1.97 (2); His, 1.07 (1); Arg, 1.04 (1)A15(IPS) 82 94 1644.3 1645 Asp, 1.03 (1); Pro, 1.04 (1); Ala, 2.01 (2); Leu, 1.90 (2); His, 0.91 (1); Arg, 1.12 (1)A15(IPI) 71 92 1626.6 1627 Asp, 1.05 (1); Ser, 0.99 (1); Pro, 1.00 (1); Gly, 1.08 (1); Ala, 1.93 (2); Arg, 1.03 (1)A15(ISI) 80 92 1631.1 1631 Asp, 1.05 (1); Ser, 2.07 (2); Ala, 2.97 (3); Leu, 1.84 (2); His, 0.96 (1); Arg, 0.95 (1)A15(FPS) 73 97 1708.3 1709 Asp, 0.97 (1); Ser, 1.97 (2); Ala, 1.04 (1); Leu, 2.08 (2); Phe, 1.03 (1); Arg, 1.02 (1)A15(FPI) 87 90 1690.1 1691 Asp, 1.02 (1); Pro, 1.00 (1); Gly, 1.06 (1); Ala, 1.02 (1); Phe, 0.86 (1); His, 0.90 (1)A15(FSI) 88 91 1695.0 1695 Asp, 1.07 (1); Ala, 2.09 (2); Leu, 1.85 (2); Tyr, 0.87 (1); Phe, 0.91 (1); His, 1.07 (1)A15(PSI) 89 92 1645.0 1645 Asp, 1.00 (1); Pro, 1.02 (1); Ala, 2.02 (2); Tyr, 0.93 (1); His, 0.91 (1); Arg, 1.11 (1)A15(IFPS) 77 95 1720.9 1721 Asp, 0.98 (1); Ser, 1.91 (2); Pro, 0.98 (1); Ala, 0.99 (1); His, 1.07 (1); Arg, 1.07 (1)A15(IFPI) 74 94 1703.1 1703 Asp, 0.99 (1); Ser, 0.94 (1); Pro, 0.96 (1); Gly, 1.03 (1); Ala, 0.96 (1); Arg, 1.01 (1)A15(IFSI) 84 98 1706.7 1707 Asp, 1.07 (1); Ser, 2.06 (2); Ala, 2.04 (2); Leu, 1.91 (2); His, 0.93 (1); Arg, 0.89 (1)A15(IPSI) 80 91 1656.0 1657 Asp, 1.02 (1); Ser, 1.91 (2); Pro, 1.02 (1); Ala, 1.98 (2); His, 0.96 (1); Arg, 1.11 (1)A15(FPSI) 56 92 1721.2 1721 Asp, 1.04 (1); Ser, 2.01 (2); Pro, 1.04 (1); Ala, 1.01 (1); Phe, 0.91 (1); Arg, 1.08 (1)A15(IFPSI) 80 94 1732.1 1733 Asp, 1.09 (1); Ser, 2.03 (2); Pro, 1.02 (1); Ala, 1.05 (1); Leu, 1.99 (2); Arg, 0.95 (1)

Note: relative amino acid proportions given by AAA are followed by the expected value in parenthesis (for simplicity, mutant peptides are designed only with the capital case lettercode for the replaced amino acids without the corresponding position number).* synthesis performed under sub-optimal conditions due to instrumental malfunction.


116

3.2 Analysis of the mutated peptides by direct kinetic SPR

Kinetic SPR screening of the thirty-one peptide antigens was performed as previously described in

chapters 1 and 2. In this particular case, it was observed that most interactions could not be

quantitated, either due to high association rates, extremely slow dissociation rates or even

insufficient surface regeneration (Fig. 3.2). This was particularly frequent with mAbs 4C4 and 3E5.

Also, surface saturation was observed for peptide concentrations higher than ca. 600 nM.

Interaction data that could be reasonably fitted as a 1:1 bimolecular interaction (Table 3.3)

presented, in most cases, rate constants in the limit of reliable kinetic information4-6. Non-ideal

effects, such as ligand rebinding in the dissociation phase, seemed to be affecting binding kinetics

(Fig. 3.3).

A

-5

-3

-1

1

3

5

7

9

11

13

15

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

A15(138F,148I), 41 nM

B

-5

0

5

10

15

20

25

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

A15(137I,138F)

C

11510

11520

11530

11540

11550

11560

11570

11580

11590

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

A15(IFI), 36 nMA15(IFI), 73 nMA15(IFI), 146 nMA15(IFI), 292 nMA15(IFI), 584 nMA15(IFI), 1167 nM

Steep ascentNegligible

descent

Increasing baseline level

Decreasing response (for increasing concentrations)

Figure 3. 2 Sensorgrams illustrating problems often observed in the kinetic SPR analysis of theinteractions between anti-site A mAbs and peptides combining mutations Thr137→Ile, Ala138→Phe,Ala140→Pro, Gly142→Ser and Thr148→Ile; A. extremely slow dissociation (as can be observed, slopeis, in fact, positive, possibly due to the sum of a negligible dissociation slope and a positive slope frominstrumental drift); B. extremely fast association (a steep ascent can be observed, giving the sensorgrama square wave-like shape in the association phase); C. Insufficient surface regeneration: the baselinelevel increases and the response level decreases from cycle to cycle.


117

SD6 4C4 3E5mAbPeptide ka/M-1s-1 kd/s

-1 KA/M-1 ka/M-1s-1 kd/s-1 KA/M-1 ka/M-1s-1 kd/s

-1 KA/M-1

A15 7.3×104 1.4×10-3 5.4××××107 3.8×105 1.9×10-3 1.9××××108 1.6×105 1.6×10-3 9.4××××107

A15(137I) 9.6×104 5.0×10-4 1.9××××108 2.9×105 2.0×10-3 1.4××××108 3.5×105 1.0×10-3 3.4××××108

A15(138F) 8.6×104 3.9×10-3 2.2××××107 5.5×105 5.7×10-3 9.8××××107 6.0×105 1.4×10-3 4.0××××108

A15(140P) 7.4×104 2.1×10-3 3.5××××107 1.9×105 1.9×10-3 1.0××××108 2.0×105 1.5×10-3 1.4××××108

A15(142S) 6.1×104 5.6×10-3 1.1××××107 6.3×104 3.8×10-3 1.6××××107 1.8×105 7.0×10-3 2.5××××107

A15(148I) 1.1×105 2.1×10-3 5.3××××107 7.0×105 2.3×10-3 3.0××××108 5.9×105 1.1×10-3 6.0××××108

A15(IF) 2.9×105 5.1×10-3 5.6××××107 6.4×105 1.7×10-3 3.8××××108 A15(IP) 1.6×105 2.6×10-3 5.9××××107 A15(IS) 7.3×104 4.1×10-3 1.8××××107 2.8×105 1.0×10-3 2.8××××108 A15(II) 2.3×105 3.0×10-3 7.6××××107 A15(FP) 1.1×105 8.5×10-3 1.3××××107 A15(FS) 1.1×105 1.1×10-2 1.0××××107 3.6×105 5.9×10-4 6.0××××108 A15(FI) 2.6×105 8.4×10-4 3.0××××108 A15(PS) 2.0×105 7.9×10-4 2.5××××108 A15(PI) 1.5×105 3.5×10-3 4.0××××107 A15(SI) 3.7×104 6.0×10-3 6.1××××106 2.0×105 7.0×10-4 2.9××××108 A15(IFP) A15(IFS) 3.5×104 3.5×10-3 1.0××××107 7.6×104 1.8×10-3 4.3××××107 A15(IFI) 1.2×105 4.2×10-3 2.9××××107 A15(IPS) 9.3×104 5.4×10-4 1.7××××108 A15(IPI) A15(ISI) 4.5×104 1.5×10-3 3.1××××107 A15(FPS) 8.5×104 8.8×10-3 9.7××××106 2.0×105 4.2×10-4 4.8××××108 A15(FPI) 1.4×105 4.3×10-3 3.3××××107 A15(FSI) 5.0×104 9.4×10-3 5.4××××106 4.6×105 1.6×10-3 1.6××××108 A15(PSI) 5.9×104 7.1×10-4 8.4××××107 4.0×105 2.9×10-4 1.4××××109 A15(IFPS) 1.5×105 4.2×10-3 3.6××××107 A15(IFPI) 2.5×105 5.5×10-4 4.5××××108 A15(IFSI) 1.3×105 1.7×10-3 7.7××××107 4.7×105 8.5×10-4 5.5××××108 A15(IPSI) 1.7×105 2.5×10-3 6.7××××107 A15(FPSI) 8.6×104 6.1×10-3 1.4××××107 A15(IFPSI) 2.1×105 2.2×10-4 9.8××××107

Table 3.3 Kinetic SPRanalysis of the peptidestowards mAbs SD6, 4C4 and3E5.

Notes:Although resulting fromapparently reliable data fits,kinetic parameters should beregarded with some caution,since most ka values are in thelimit for reliable kineticmeasurements not affected bymass-transport limitations. non-measurable interactions.


118

A

-5

0

5

10

15

20

-10 40 90 140 190 240 290Time/s

Re

sp

on

se

/RU

A15(140P,142S), 20 nMA15(140P,142S), 39 nMA15(140P,142S), 78 nMA15(140P,142S), 156 nMA15(140P,142S) 312 nMA15(140P,142S), 625 nM

B

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.00E+00 1.00E-07 2.00E-07 3.00E-07 4.00E-07 5.00E-07 6.00E-07 7.00E-07


k s/s

-1

r2=0.984

r2=0.999

Figure 3. 3 Sensorgrams of the interaction between peptide A15(140P,142S) and mAb 4C4 (A); despite thegood quality of data when using global curve fitting, slight deviations were observed in the dissociation rateconstants (increasing with peptide concentration) when local curve fitting was employed; this fact together withthe slight curvature observed when plotting ks against peptide concentration (B) indicate that non-ideal effectswere affecting binding kinetics.


119

3.3 Indirect SPR kinetic analysis using a high molecular weight competitor antigen7

As explained in section 0.2.3, there are indirect SPR methods for the kinetic characterisation of

biospecific interactions between an immobilised receptor and small ligands in solution. One of them

is a kind of surface competition assay, where a high molecular weight ligand, specific to the

immobilised receptor, is injected and detected on the sensor chip surface. The kinetics of the

macromolecular ligand – receptor interaction is characterised and, then, mixtures of this

macromolecule with varying concentrations of the small analyte of interest are injected (Fig. 3.4).

The effects observed on the macromolecule – receptor interaction upon addition of the small

competitor analyte provide an indirect means to measure the kinetics of the small analyte – receptor

interaction.

Figure 3. 4 Main steps in the surface competition SPR kinetic analysis: left – the kinetics of the interactionbetween surface – immobilised receptor (brown) and the macromolecular analyte (yellow) is measured; right –effects due to addition of a small competitor (target analyte, green) on the macromolecule – receptor interactionare evaluated and the kinetics of the small analyte – receptor interaction is thus determined.

Due to problems often observed in the direct kinetic SPR described in the previous section, it was

thought that perhaps an indirect approach would be more appropriate. Indeed, peptide – antibody

interactions seemed to be affected by diffusion-controlled kinetics, due to the apparently high

association rates involved. Further, symptoms of both analyte rebinding and surface saturation at

higher analyte concentrations were equally observed. Thus, it would seem advisable to decrease

mAb surface density and analyte concentrations as much as possible, in order to minimise such non-

ideal effects4-6. However, this would imply severe losses in response levels, given the small size of the

target analytes. Therefore, an indirect approach based on the detection of a macromolecular analyte

would be a possible solution for these problems7.

In the absence of a natural macromolecular FMDV antigen available for such surface competition

SPR assays, alternative high molecular weight antigens were chosen, as described in the following

sections.


120

3.3.1 Peptide – protein (1:1) conjugate: the A15 – human carbonic anhydrase I (HCA I) construct

To ensure unambiguous mechanisms for peptide – macromolecule competitive binding to the

antibody paratope, a macromolecular antigen with only one specific epitope per molecule was

needed. This implied the conjugation of a site A representative peptide such as A15 to a protein

carrier in a 1:1 stoichiometry. A safe way to achieve such stoichiometry would be to link through a

heterodisulphide bond the A15 sequence (with an additional Cys residue) to a carrier protein

bearing a single Cys.

Search in the Protein Data Bank for a protein suitable for such purposes (i.e., with a single cysteine

residue, weighing over ca. 10 kDa, commercially available and affordable) led to human carbonic

anhydrase I (HCA I – E. C. 4.2.1.1).

3.3.1.1 Synthesis of peptide (Npys)Cys – A15 and heterodimerisation with protein HCA I

The A15 sequence was assembled as described in previous sections, then Boc–Cys(Npys)–OH was

coupled as N-terminal residue by similar protocols and cleavage/side chain deprotection with TFA

was then performed as usual (see Materials & Methods, section 4.3), leading to peptide (Npys)Cys–

A15 (Table 3.4). The 3-nitro-2-pyridylsulphenyl (Npys) group is stable to TFA and therefore not

removed in the cleavage. This thiol protecting group was chosen for its well-known applicability to

direct peptide – protein conjugation through cysteine residues. Such property of the Npys group is

due both to the fact that it is stable to the standard acidolytic cleavage conditions (even to hydrogen

fluoride acidolysis in Boc/Bzl chemistry) and also to its thiol-activating character, which allows

regiospecific peptide – protein coupling through cysteine residues (Fig. 3.5)8.

Table 3.4 General data concerning the synthesis of peptide (Npys)Cys – A15.

Globalyield (%)

Purity(% HPLC)

MWfound

MWexpected

AAA

67 81 1833.6 1833 Asp, 1.04 (1); Ser, 0.93 (1); Gly, 1.07 (1);Ala, 3.10 (3); Leu, 2.00 (2); His, 0.97 (1)

S S

S

S

NNO2

S

HNNO2SH pH 4.2

+

Figure 3. 5 Regiospecific formation of a disulphide heterodimer via the Npys thiol activation; Npys not onlyserves as thiol-protection during peptide synthesis, but also activates the Cys sulphur atom toward othernucleophiles, such as other Cys thiol groups8.


121

The heterodimerisation reaction was carried out overnight under acidic, denaturating conditions (6

M guanidine hydrochloride in water, pH 4.2) using excess peptide, and was monitored by HPLC

(Fig. 3.6). The final reaction mixture was dialysed (MW cut-off = 15 to 20 kDa) against decreasing

concentrations of guanidine hydrochloride in water for 48 hours. The major product in the final

dialysed solution was characterised by MALDI-TOF MS (Fig. 3.7) as the target HCA I – CysA15 1:1

conjugate.

Figure 3. 6 HPLC monitoring of theheterodimerisation reaction: A. crude(Npys)Cys – A15 peptide; B. HCA I;C, D and E progress of the reaction at2, 8 and 24 h, respectively.

Figure 3. 7 MALDI-TOF MS spectrum of the finalheterodimerisation product, which contains unreacted HCA Iprotein (the MALDI-TOF MS spectrum of the commercial HCA Iis shown in the upper left corner).

A B

C D E

1 2

1

2

3

1

2

31

2

3

15 - 65%B, 30 min 15 - 65%B, 30 min

15 - 65%B, 30 min 15 - 65%B, 30 min 15 - 65%B, 30 min


122

3.3.1.2 Antigenic evaluation of the HCA I – CysA15 conjugate

The adequacy of the HCA I – CysA15 construct as an FMDV antigen was evaluated by direct SPR

detection on SD6, 4C4 and 3E5 surfaces (with mAb densities of approximately 600 RU, i. e, 0.6

ng/mm2). Unfortunately, absence of specific response was repeatedly observed, showing that the

construct was not antigenic towards these mAbs (Fig. 3.8). Some assays at pH values (5.8 and 8)

different from the usually employed (7.3) did not improve the results (only non-specific binding

between protonated protein – negatively-charged dextran layer was observed at the lowest pH).

These results were confirmed by competition ELISA experiments using A21 – KLH conjugate as the

plate-bound antigen (Fig. 3.9). Inadequate conformational presentation or inaccessibility of peptide

A15 were considered as probable causes for such lack of antigenicity.

A

-5

5

15

25

35

45

55

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

31 nM

62 nM125 nM

250 nM

500 nM

B

-200

4800

9800

14800

19800

24800

29800

34800

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

31 nM

62 nM

125 nM

250 nM

500 nM

C

-200

4800

9800

14800

19800

24800

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

31 nM

62 nM

125 nM

250 nM

500 nM

Figure 3. 8 SPR analysis of conjugate HCA I – CysA15: A. injection of conjugate samples over a mAb 4C4surface (600 RU), employing the conditions used for FMDV peptides; B. and C. injection of protein samples atpH 5.8 over the same 4C4 surface and over an non-derivatised sensor chip surface, respectively.

Figure 3. 9 Inhibition curves fromcompetition ELISA analysis of theHCA I – CysA15 conjugate (curvesfor reference peptide A15 wereomitted).

0

20

40

60

80

100

120

140

0.1 1 10 100 1000

conjugate concentration (pmol/100µµµµL)

% a

bso

rban

ce

HCA-A15 vs. 3E5HCA-A15 vs. SD6

HCA-A15 vs. 4C4


123

3.3.2 Recombinant engineered proteins bearing the FMDV GH loop peptide: protein JX249A

3.3.2.1 Preliminary assays

The extensive research work on recombinant proteins bearing the FMDV C-S8c1 GH loop carried

out by A. Villaverde and co-workers9-12 opened the possibility to use one of these engineered

proteins as such macromolecular antigen.

A recombinant β-galactosidase from Escherichia coli, protein JX249A, has solvent exposed-loops

which have been engineered for the insertion of a peptide from the GH loop of FMDV C-S8c1

(TT136YTASARGDLAHLTT150THARHLP). JX249A is a homotetramer with four GH loops per

protein, having a total molecular weight of 472 kDa. Given its high antigenicity, JX249A was tested

in preliminary SPR assaysA, where each protein molecule would be regarded as four independent

antigenic monomers to simplify data processing. The first assays confirmed the antigenicity of

JX249A towards mAbs SD6, 4C4 and 3E5, and preliminary analyses also indicated that peptide

A15 competed with JX249A in binding to surface-immobilised mAb (surface densities of about 600

RU), while non-specific peptide A15Scr did not (Fig. 3.10). Nevertheless, important problems due to

insufficient surface regeneration and consequent protein accumulation on the surface led to short

surface life-times and prevented a systematic screening of the peptide antigens.

A

-10

10

30

50

70

90

110

-10 40 90 140 190 240 290Time/s

Res

po

nse

/RU

JX249A, 10 nM

JX249A, 20 nM

JX249A, 40 nMJX249A, 80 nM

JX249A, 160 nM

B

-10

0

10

20

30

40

50

60

70

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

A15, 0 nM

A15, 19 nMA15, 38 nM

A15, 76 nM

A15, 152 nM

decreasing slopewith increasing

peptideconcentration

faster protein dissociation and lower bound proteinlevel with increasing peptide concentration

C

-10

0

10

20

30

40

50

60

70

80

90

-10 40 90 140 190 240 290 340Time/s

Res

po

nse

/RU

A15Scr, 0 nM

A15Scr, 20 nM

A15Scr, 40 nM

A15Scr, 80 nMA15, 160 nM

slope independentof peptide

concentration

dissociation rate and amount of protein boundindependent of peptide concentration

A Protein engineering, production and antigenic evaluation by ELISA were performed by Dr. A. Villaverde and co-workers (U.A. B., Bellaterra – Barcelona), who kindly offered protein samples to the author.

Figure 3. 10 SPR assays using JX249A forindirect SPR kinetic analysis of peptide-antibodyinteractions: A. JX249A – mAb 4C4 interaction;B. competition between JX249A with peptideA15; C. competition between JX249A and peptideA15Scr; JX249A was used at a constant 80 nMconcentration in the competition assays.


124

3.3.2.2 Screening of alternative regeneration conditions

Insufficient surface regeneration was, once again, preventing the study of the kinetics of peptide –

mAb interactions by SPR. Therefore, a study of regeneration conditions, as described by Andersson

and co-workers13, was carried out. This study was based on the screening of several regeneration

cocktails, consisting of mixtures of the stock solutions presented in Table 3.5. This multi-cocktail

approach is based on the principle that what one kind of chemical property (acidic, basic, saline,

organic, denaturating) cannot disrupt, perhaps another one can or, even better, combination of

several distinct chemical properties will act synergistically and solve the regeneration problem.

Table 3.5 Stock solutions13used for the multi-cocktail surface regeneration approach in SPR analysis.

Cocktail Main chemical properties Composition

A AcidEqual volumes of 0.15 M phosphoric, formic andmalonic acids, adjusted to pH 5 with 4 M NaOH

B BasicEqual volumes of 0.20 M ethanolamine, sodiumphosphate, piperazine and glycine, adjusted to pH 9with 2 M HCl

I Ionic/denaturatingPotassium thiocianate (0.46 M), magnesium chloride(1.83 M), urea (0.92 M) and guanidine hydrochloride(1.83 M)

U Non-polar/organicEqual volumes of dimethylsulfoxide, formamide,acetonitrile, ethanol and 1-butanol

D Detergent0.3% (w/w) CHAPS, 0.3% (w/w) zwittergent 3-12,0.3% (v/v) Tween 80, 0.3% (v/v) Tween 20 and0.3% (v/v) Triton X-100

C Chelating 20 mM EDTA aqueous solution

The general protocol consists of a screening and an optimisation step. In the first step, diluted

solutions or simple binary combinations of the above described cocktails are tested (Table 3.6). The

evaluation of the screening cocktails is carried out in the biosensor and, afterwards, an optimisation

step is performed upon combinatorial mixing of the stock cocktails rated as the best.

Table 3.6 Cocktails used in the screening step of the multi-cocktail regeneration approach13.

Composition of the screening cocktails

Bww Iww Dww Uww Cww BDw BCw Aiw Adw Auw Acw Idw Icw Duw DCw Ucw ABw

* equivalent amounts (v/v) of each cocktail represented by the corresponding letter (see Table 3.5); w stands for water.


125

Having an immunoglobulin as the surface-immobilised receptor, the use of cocktails I or D, as

defined in Table 3.5, could be harmful. Therefore, cocktails A, B, C, U and a modified ionic cocktail

I’ (differing from I in that denaturating chemicals such as urea or guanidine hydrochloride were not

added) were screened as described. The screening was performed on high density mAb surfaces (ca.

3 ng/mm2) to ensure high responses for an unequivocal evaluation of the cocktail regeneration

efficacy. In neither case was a significant improvement observed (Fig. 3.11).

650 RU 640 RU 638 RU 630 RU 627 RU 1152 RU

Cocktail B

Cocktail C

Cocktail A

Cocktail I’

Cocktail U

-20

480

980

1480

1980

-20 180 380 580 780 980 1180 1380 1580 1780

Time/s

Re

sp

on

se

/RU

Figure 3. 11 Successive injections of stock regeneration solutions Bww, Cww, Aww, I’ww and Uww (see Table3.6) on a mAb 4C4 surface (3 ng/mm2), after binding of protein JX249A (300 nM).

Results in Fig. 3.11 show the inefficacy of the stock regeneration solutions tested for recovering the

initial baseline level. Further tests, either with triplicate injections of each regeneration cocktail or

with binary combinations of the stock solutions (AI’ and CI’ at different proportions), did not

improve the results. The increase in response level observed when injecting cocktail Uww was

attributed to compression of the hydrophilic dextran matrix due to the organic solvents present in

this cocktail.

The cocktail approach was also tested on some problematic peptide – mAb interactions (mentioned

in 3.2), but results were none the better.


126

3.3.2.3 Capping of JX249A free cysteine thiol groups

We hypothesised that the free thiol groups from the JX249A cysteine residues might be the source

for the irreversible binding of the protein to the mAb surfaces, upon disulphide bridge cross-linking.

Usually, experiments with bacterial β-galactosidases at room temperature require the addition of β-

mercaptoethanol, in order to maintain the cysteine thiol groups in their native free form and to

avoid protein aggregation. In the above SPR experiments, β-mercaptoethanol was not added, since

protein solutions would be put in contact with an antibody surface and native folding of the

immunoglobulin had to be preserved. Thus, capping of the JX249A cysteine thiol groups seemed to

be a wise precaution to avoid both protein aggregation and, possibly, disulphide cross-linking to the

mAb surfaces.

Capping of the free thiol groups was carried out by nucleophilic substitution with iodoacetic acid

(Fig. 3.12) as described under Materials & Methods (section 4.3); this converted cysteine residues

into carboxymethylcysteine14. Protein integrity after carboxymethylation was checked by SDS-PAGE

(Fig. 3.13) and capping yield (75%) was assessed by AAA, using carboxymethylcysteine standards.

SH

S

OH , ∆HI

COOH

+COOHI

Figure 3. 12 Nucleophilic attack of a thiol sulphur atom on the methylene group of iodoacetic acid14.

Figure 3. 13 SDS-PAGE (10% acrylamide) analysis ofprotein JX249A before (lane A) and after (lane C)carboxymethylation of the cysteine side-chain thiolgroups; the following protein standards (lane B) wereemployed: from higher to lower MW – myosin, β-galactosidase, phosphorylase B, bovine albumin,ovalbumin and carbonic anhydrase.

Use of the capped JX249A in SPR assays as those described in the previous section did not,

unfortunately, solve the regeneration problems already observed (not shown). This indicated that

such problems were due either to intrinsic features of the protein – surface interactions or to the

presence of some free JX249A mixed with the carboxymethylated protein.

JX249A after carboxymethylationJX249A

205 kDa

116 kDa97 kDa

66 kDa

45 kDa

29 kDa

A B C


127

3.4 Solution affinity SPR analysis of the peptide antigens

Several features of anti-FMDV mAb interactions with either the peptide antigens or protein JX249A

prevented their dynamic characterisation by means of SPR kinetic analysis. Therefore, the

characterisation of the synthetic peptides under study was alternatively carried out by solution

affinity SPR analysis15,16, as previously described for the FMDV C-S30 peptides (previous chapter,

section 2.7). The same experimental set-up was employed under similar conditions (described in

Materials & Methods, section 4.3). A confirmative competition ELISA screening of the peptides was

carried out in parallel, as described in section 2.3.

Inhibition curves such as those exemplified in Fig. 3.14 were observed and affinity constants were

determined by the Cheng & Prusoff’s formula17, as previously exposed in section 2.7. These

constants are presented in Table 3.7, where the corresponding results from kinetic SPR analysis and

competition ELISA are included, for comparison purposes.

0.0E+00

1.0E-08

2.0E-08

3.0E-08

4.0E-08

5.0E-08

6.0E-08

7.0E-08

8.0E-08

9.0E-08

0.0E+00 1.0E-07 2.0E-07 3.0E-07 4.0E-07 5.0E-07 6.0E-07

Peptide Concentration/M

Fre

e F

ab

Co

nc

en

tra

tio

n/M

A15scr

A15(138F)

A15(140P)

A15(142S)

A15(FPS)

A15

Figure 3. 14 Inhibition curves in the SPR analysis of the interactions between Fab 4C4 andpeptide antigens in solution; a constant 80 nM total concentration of Fab was employed; peptideA15Scr was included as a negative control.

Analysis of quantitative data in Table 3.7 shows that all the peptides are highly antigenic, taking the

native A15 antigen as reference. Differences observed between SPR data in kinetic or in solution

equilibrium experiments were not generally significant. This is a further evidence of the reliability of

the kinetic SPR methodology employed along the present study of FMDV peptides. More

pronounced differences were due either to mass transport-influenced kinetic data or to the fact that

immobilised mAb – free peptide and free mAb – free peptide interactions are intrinsically different

and therefore only peptide ranking should be compared.


128

Table 3.7 Affinity constants for the interactions between mAbs (Fab) SD6, 4C4 and 3E5 with the peptide antigens bearing combinations of the replacements Thr137→Ile,Ala138→Phe, Ala140→Pro, Gly142→Ser and Thr148→Ile (columns labelled KA-sol.aff.).

SD6 4C4 3E5FabPeptide KA-sol.aff. /M

-1 KA-kin /M-1 E KA-sol.aff. /M

-1 KA-kin /M-1 E KA-sol.aff. /M

-1 KA-kin /M-1 E

A15 6.3××××107 5.4×107 ++ 2.0××××108 1.9×108 ++ 2.0××××108 9.4×107 ++A15(137I) 8.5××××107 1.9×108 ++ 2.0××××108 1.4×108 ++ 1.4××××108 3.4×108 ++A15(138F) 3.6××××107 2.2×107 ++ 2.1××××108 9.8×107 ++ 2.1××××108 4.0×108 ++A15(140P) 7.1××××107 3.5×107 ++ 1.8××××108 1.0×108 ++ 1.6××××108 1.4×108 ++A15(142S) 2.7××××107 1.1×107 ++ 7.3××××107 1.6×107 ++ 6.2××××107 2.5×107 ++A15(148I) 6.7××××107 5.3×107 ++ 2.0××××108 3.0×108 ++ 2.0××××108 6.0×108 ++A15(IF) 5.1××××107 5.6×107 ++ 2.2××××108 3.8×108 ++ 1.6××××108 ++A15(IP) 6.1××××107 5.9×107 ++ 2.1××××108 ++ 1.3××××108 ++A15(IS) 5.2××××107 1.8×107 ++ 1.9××××108 2.8×108 ++ 1.3××××108 ++A15(II) 9.1××××107 7.6×107 ++ 2.3××××108 ++ 2.0××××108 ++A15(FP) 2.8××××107 1.3×107 ++ 2.2××××108 ++ 2.0××××108 ++A15(FS) 7.1××××106 1.0×107 ++ 2.0××××108 6.0×108 ++ 1.6××××108 ++A15(FI) 2.1××××107 3.0×108 ++ 2.1××××108 ++ 1.9××××108 ++A15(PS) 3.6××××107 ++ 2.0××××108 2.5×108 ++ 6.2××××107 ++A15(PI) 6.0××××107 4.0×107 ++ 1.8××××108 ++ 1.9××××108 ++A15(SI) 1.6××××107 6.1×106 ++ 1.9××××108 2.9×108 ++ 1.2××××108 ++A15(IFP) 7.4××××107 ++ 2.2××××108 ++ 1.9××××108 ++A15(IFS) 2.0××××107 1.0×107 ++ 8.5××××107 4.3×107 ++ 5.9××××107 ++A15(IFI) 5.6××××107 2.9×107 ++ 2.4××××108 ++ 1.8××××108 ++A15(IPS) 7.4××××107 1.7×108 ++ 2.2××××108 ++ 1.4××××108 ++A15(IPI) 7.9××××107 ++ 2.2××××108 ++ 2.0××××108 ++A15(ISI) 4.5××××107 3.1×107 ++ 2.2××××108 ++ 1.7××××108 ++A15(FPS) 6.5××××106 9.7×106 ++ 2.1××××108 4.8×108 ++ 1.3××××108 ++A15(FPI) 5.0××××107 3.3×107 ++ 1.3××××108 ++ 1.8××××108 ++A15(FSI) 1.3××××107 5.4×106 ++ 2.1××××108 1.6×108 ++ 1.5××××108 ++A15(PSI) 2.5××××107 8.4×107 ++ 2.1××××108 1.4×109 ++ 1.5××××108 ++A15(IFPS) 4.7××××107 3.6×107 ++ 2.2××××108 ++ 1.3××××108 ++A15(IFPI) 7.1××××107 4.5×108 ++ 2.0××××108 ++ 1.8××××108 ++A15(IFSI) 1.9××××107 7.7×107 ++ 2.1××××108 5.5×108 ++ 1.2××××108 ++A15(IPSI) 5.0××××107 6.7×107 ++ 1.5××××108 ++ 1.3××××108 ++A15(FPSI) 1.7××××107 1.4×107 ++ 2.2××××108 ++ 1.3××××108 ++A15(IFPSI) 3.8××××107 9.8×107 ++ 1.9××××108 ++ 1.7××××108 ++

Note: values measured by kinetic SPR are included in columns labelled KA-kin; total Fab concentrations were kept constant for each peptide dilution series and peptide concentrations variedbetween 0 and 625nM; results from a competition ELISA screening of the peptides are also included (column E), where signs – , + and ++ stand for low (IC50 rel>30), medium (30<IC50rel<10)and high antigenicity (IC50 rel<10), respectively.


129

The one-point mutants are closely equivalent regarding antigenicity. In spite of this, mAb SD6

slightly “resents” mutations A138→F and G142→S, while the other two mAbs only disfavour the

mutation within the RGD motif. The higher involvement of residue 138 in peptide-SD6 complexes

and the important role of the RGD triplet are both in agreement with these observations1,18-21. All

mAbs are “indifferent” to mutations T137→I and T148→I, which is consistent with the almost

absent participation of these residues in the mAb-peptide interactions21.

Generally, the multiply substituted peptides displayed similar affinities, close to those expected from

additive effects in the combination of the one-point mutations (Fig. 3.15). However, for mAbs 4C4

and 3E5, affinities of multiple mutants containing the G142→S replacement were systematically

superior to those expected from the “additivity rule”. Unless there was an undetected error in the

affinities of all peptides containing this mutation towards both 4C4 and 3E5, these differences

suggest a small positive synergistic effect in these multiple mutants. So it seems that mutations

outside the RGD triplet compensate the slight decrease in affinity provoked by replacing glycine by

serine in this important motif. Possibly, such compensation comes from peptide conformational

features, which are not as favourable in the A15(142S) single-point mutant as when the other

replacements are combined with the RSD motif.

Interestingly, peptides including multiple mutations within the GH loop of C-S8c1 FMDV still display

antigenicities as high as those of the native sequence. This is even more relevant when one of these

mutations is located in the RGD triplet and involves the substitution of a glycine by a serine residue.

For these reasons, the multiply substituted peptide A15(FPS) was submitted to further structural

studies, both by two-dimensional 1H-NMR of free peptide in solution and by X-ray diffraction

crystallography of its complex with antibody 4C4. Peptide A15(FPS) was chosen since it combines

the three more relevant mutations. Also, W. F. Ochoa and co-workers have observed very low

electron densities for residues placed at both ends of FMDV pentadecapeptides (residues≤137 and

≥148) preventing the unequivocal location of such residues in the structure of peptide-4C4

complexes. This has been interpreted as due to the lack of strong interactions between the terminal

residues and the antibody paratope21.


130

Figure 3. 15 Comparison betweenmeasured and expected (for additivecombination of partial mutations) solutionaffinity constants.

A

0.0E+00

1.0E+07

2.0E+07

3.0E+07

4.0E+07

5.0E+07

6.0E+07

7.0E+07

8.0E+07

9.0E+07

1.0E+08

IF IP IS II

FP

FS FI

PS PI

SI

IFP

IFS IFI

IPS

IPI

ISI

FP

S

FP

I

FS

I

PS

I

IFP

S

IFP

I

IFS

I

IPS

I

FP

SI

IFP

SI

peptide

Aff

init

y co

nst

ant/

M-1

measured

expected

B

0.0E+00

5.0E+07

1.0E+08

1.5E+08

2.0E+08

2.5E+08

IF IP IS II

FP

FS FI

PS PI

SI

IFP

IFS IFI

IPS IPI

ISI

FP

S

FP

I

FS

I

PS

I

IFP

S

IFP

I

IFS

I

IPS

I

FP

SI

IFP

SI

peptide

Aff

init

y c

on

sta

nt/

M-1

measured

expected

C

0.0E+00

5.0E+07

1.0E+08

1.5E+08

2.0E+08

IF IP IS II

FP

FS FI

PS PI

SI

IFP

IFS IFI

IPS IPI

ISI

FP

S

FP

I

FS

I

PS

I

IFP

S

IFP

I

IFS

I

IPS

I

FP

SI

IFP

SI

peptide

Aff

init

y c

on

sta

nt/

M-1

measured

expected


131

3.5 Two-dimensional 1H-NMR analysis of peptide A15(FPS)

The structural features of peptide A15(FPS) in solution were analysed by 2D – 1H NMR22, under

conditions identical to those described in section 2.8 for C-S30 peptides. In the present case, peptide

A15(FPS) was studied both in water and in 30% TFE, through TOCSY and NOESY/ROESY

experiments (Fig. 3.16)23-25. The chemical shifts measured are presented in Table 3.8.

143D

142S

146H

139S

138F141R

137T148T

149T147L

150T

145A

144L

142S -144L

149T -150T

145A -146H138F -139S

142S -143D

Figure 3. 16 Expansions of the TOCSY (above) and ROESY at 200 ms (below) spectra ofpeptide A15(FPS) in water at 25 oC.


132

Table 3.8 Chemical shifts measured in the 2D 1H-NMR analysis of of peptide A15(FPS) at 25 oC.

A15(FPS)H2O 30% TFEResidue

HN Hα Hβ Other HN Hα Hβ Other136Tyr 4.23 3.03 7.14

6.804.23 3.05 7.04

6.84137Thr 8.34 4.28 4.22 1.12 8.25 4.34 4.03 1.16138Phe 8.39 4.58 3.07 7.29 8.24 4.64 3.09 7.30139Ser 8.20 4.68 3.75 8.04 4.76 3.84

3.78140Pro 4.36 2.30

1.951.903.673.60

4.38 2.301.99

3.733.57

141Arg 8.35 4.30 1.861.75

1.653.207.31

8.03 4.28 1.901.77

1.683.227.36

142Ser 8.29 4.40 3.903.85

8.06 4.37 3.973.89

143Asp 8.49 4.68 2.852.80

8.34 4.66 2.85

144Leu 8.07 4.26 1.60 0.900.85

8.01 4.23 1.701.62

0.930.88

145Ala 8.08 4.25 1.30 7.96 4.17 1.38146His 8.29 4.69 3.27

3.157.607.18

8.04 4.63 3.383.24

8.067.33

147Leu 8.25 4.42 1.60 0.900.85

8.07 4.36 1.781.66

0.930.90

148Thr 8.28 4.45 4.25 1.26 7.98 4.42 4.34 1.26149Thr 8.23 4.46 4.30 1.20 7.93 4.47 4.28 1.25150Thr 8.12 4.33 4.28 1.20 ambiguous due to peak overlap

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

Y T F S P R S D L A H L T T T

residue

∆δ ∆δ∆δ∆δH

α ααα/p

pm

Water30% TFE

Figure 3. 17 Conformational chemical shifts observed for peptide A15(FPS) in water and30% TFE, at 25 oC.


133

Conformational chemical shifts

Fig. 3.17 shows absolute conformational chemical shifts slightly higher for peptide A15(FPS) than

those previously observed for the C-S30 peptides (chapter 2). The global shape of the plots

resembles those observed for peptides C-S30 or C-S8c1 under identical conditions. However, the

region containing the open turn at the Arg-Gly-Asp triplet and the following short helix usually

observed in FMDV peptides26,27 is longer for peptide A15(FPS). This region extends in a continuous

manner from position 140 (in which an alanine is replaced by a proline) to the 146His-147Leu

positions, including the altered Arg-Ser-Asp motif. The almost identical plots obtained either in

water or in 30% TFE suggest that peptide A15(FPS) is not particularly sensitive to structure-inducing

solvents, and thus conformationally stable.

NOEs observed

The above observations are further supported by the NOEs observed for this peptide (Fig. 3.16):

despite its globally disordered structure, peptide A15(FPS) presented some interesting NOEs

corresponding to NNi,i+1 and NNi,i+2 connectivities, located precisely in the short helical path

mentioned above (Fig. 3.18). These NOEs reinforce that helical character is indeed present in the142S – 146H stretch.

Compared to what was described in chapter 2 for C-S30 peptides, peptide A15(FPS) is more prone

to adopt a defined structure in solution. In fact, linear peptide A15S30 did not exhibit any clearly

observable NOEs in water. Only in the presence of a structuring solvent (TFE) did the C-S30

peptide exhibit NOEs similar to those observed in the case of A15(FPS). As for the more antigenic

version cyc16S30, NOEs coincident to those described for A15(FPS) were observed both in water

and 30% TFE. These observations further support a significant relationship between a stable peptide

conformation in solution and antigenicity.

Solvent NOE Y T F S P R S D L A H L T T TααααNi,i+1

H2O NNi,i+1

NNi,i+2

ββββNi,i+1

ααααNi,i+1

30% TFE NNi,i+1

NNi,i+2

ββββNi,i+1

Figure 3. 18 Distribution of NOEs observed for peptide A15(FPS) in water and 30% TFE; NOErelative intensity is represented by the thickness of the bars and dotted lines correspond to possiblyoverlapping NOEs.


134

3.6 X-ray diffraction crystallography analysis of a peptide – antibody complex

3.6.0 Introduction

X-ray diffraction crystallography was employed in the structural study of the complex formed by

peptide A15(138F,140P,142S) and the Fab fragment of mAb 4C4. This study was performed in

collaboration with W. F. Ochoa, Dr. N. Verdaguer and Dr. I. Fita, who had been working on similar

peptide – antibody complexes1,19-21, including those between Fab 4C4 and FMDV peptides A15,

A15(138F), A15(140P) and A15(142S). The present structural study had the purpose of analysing

in detail eventual structural changes in the paratope – epitope interaction region, caused by the

simultaneous introduction of three important mutations in the FMDV GH loop. Also, we were

interested in observing how both antibody and peptide molecules managed to fit together forming a

stable complex, in the presence of such amino acid replacements.

3.6.1 Determination of protein structure by X-ray crystallography28

Solving the three-dimensional structure of proteins by X-ray crystallography requires well-ordered

and strongly X-ray diffracting crystals. However, well-ordered protein crystals are difficult to grow,

due to the large size and irregular surface of protein molecules. These pack in a crystal forming large

channels that occupy more than half the volume of the crystal and are filled with solvent molecules.

Different protein molecules within a crystal seldom are in direct contact with each other and

interactions are indirect, through several solvent layers. This feature is the reason why structures of

proteins determined by X-ray crystallography are considered as the same as those for the

biologically active proteins in solution. The high solvent content makes protein crystals much less

resistant than their inorganic counterparts and protein crystallisation is difficult to achieve, being

critically dependent on factors such as pH, temperature, protein concentration and purity, solvent,

ionic strength and precipitant. Crystals

are formed upon slow precipitation from

supersaturated solutions. The most

widely employed technique for protein

crystallisation is the hanging-drop vapour

diffusion method (Fig. 3.19), in which a

droplet of protein solution (plus adequate

additives) is placed on a glass cover,

facing down a larger reservoir of a similar

solution (with higher precipitant

concentration and without protein). The

droplet looses water gradually by vapour

Figure 3. 19 Scheme of the hanging-drop method used inprotein crystallisation (adapted from reference 30).

seal

glasscover

precipitant solution

proteinsolution


diffusion to the reservoir and precipitation occurs.

Crystals are submitted to X-ray diffraction analysis. X-rays are short wavelength electromagnetic

radiation, resulting from electronic transitions from excited to low energy levels. Conventional X-ray

sources are high-voltage tubes with a metal plate (anode) that is bombarded with accelerating

electrons, X-rays of specific wavelength being emitted. Rotating anode X-ray generators are the most

commonly found in X-ray crystallography laboratories. Much more powerful X-ray generators are

synchrotron storage rings, in which electrons or positrons travel close to the speed of light. Strong

radiation is emitted at all wavelengths, covering the X-ray spectrum. After passing through a

collimator, monochromatic X-ray radiation is produced with an intensity several orders of

magnitude higher than that produced by conventional X-ray sources. This allows very short

exposure times in diffraction experiments and useful data can be collected from small and more

sensitive crystals.

The primary beam must strike the crystal from several different directions so that all possible

diffraction spots are produced. The crystal is therefore rotated during the experiment and the

diffraction spots are recorded either on film or by an electronic detector. Electronic area detectors,

such as the imaging plate detector, are a kind of electronic film where a plate covered with a

photosensitive material is used to store the diffraction spots. The image thus produced is then

digitised into a computer.

When a crystal is put in the path of an X-ray primary beam, some of the X-rays interact with the

electrons on each crystal atom, causing them to oscillate. The oscillating electrons, in turn, emit new

X-rays in all directions, a phenomenon known as scattering. Due to the regular three-dimensional

arrangement of atoms in a crystal, the radiations emitted by the different electrons interfere with

each other and, in most cases, cancel each other out. Some of them interfere positively, giving

beams that are recorded as diffraction spots (Fig. 3.20).

crystalX-raysource

primary beam

detector

diffracted beams

A

B

Figure 3. 20 Representationof a diffraction experiment: A.diffracted X-ray beams afterthe primary beam hits thecrystal; B. a diffraction patternfrom a protein crystal (adaptedfrom reference 30).

135


136

Thus, each spot is originated by interference of all X-rays emerging from all crystal atoms with

identical diffraction angle. According to Bragg’s law, diffraction is regarded as reflection of the

primary beam by a set of parallel planes through the unit cells of the crystal. X-rays reflected from

adjacent planes travel different distances and diffraction only occurs when this difference equals the

wavelength of the beam. The position of the diffraction data on the detector film relates each spot to

a specific set of planes through the crystal, from which the size of the unit cell can be determined.

Each recorded spot is related to a diffracted beam characterised by its amplitude, wavelength and

phase. These three parameters are needed to determine the spatial arrangement of the atoms.

However, the phase is lost in X-ray diffraction experiments and this is the major problem in X-ray

crystallography. The classical way to circumvent this problem in protein X-ray crystallography is

based on the preparation of heavy atom protein derivatives. X-ray diffraction data from the protein

alone, from the heavy atom alone and from the heavy atom protein derivative are then used to

attribute initial phases to the protein atoms.

A simpler method to determine initial phases to work with is the molecular replacement method,

which requires a known protein structure similar to the molecule under study. The phases belonging

to the search model are assigned as the initial phases of the new protein structure and an electron

density map is calculated.

Then, with the aid of computer graphics, a trial-and-error process is started in order to build up a

model: the polypeptide main chain and side chains are matched with the electron densities and

computer-aided crystallographic refinement of the model is performed. In this refinement, the model

is slightly changed to minimise the differences between the experimental diffraction data and the

calculated model. This difference can be given in terms of the R factor, a residual disagreement that

is zero for total agreement and around 0.59 for total disagreement. The R factor lies between 0.15

and 0.20 for high quality data.

Non-zero R factors are seldom due to errors in the protein model. Rather, they derive from

imperfections in the experimental data, such as variations in protein conformation, inaccurate

solvent corrections or orientation of the micro-crystals. Therefore, the final model is an average of

molecules that differ slightly both in conformation and orientation, not corresponding exactly to the

real crystal.


137

3.6.2 General structure of immunoglobulins28

The basic structure of all immunoglobulins (Ig) involves two identical heavy chains and two identical

light chains, linked through disulphide bonds (Fig. 3.21). The major type of immunoglobulin in

human serum is the class G Ig (IgG), which is a monomer of the basic structural unit. The IgG

polypeptide chain is divided into domains of 110 amino acid residues each; the light chains contain

two of such domains and the heavy chains contain four. A light chain is composed of a variable

amino-terminal domain (VL) and a constant carboxy-terminal (CL), whereas a heavy chain is built up

from an amino-terminal variable domain (VH) followed by three constant domains (CH1, CH2 and

CH3). The variable VL and VH domains coincide with the antigen binding sites of the IgG and are not

uniformly variable along their lengths: three sub-domains, called hypervariable or complementarity

determining regions (CDR1, CDR2 and

CDR3), show much higher variability,

both in sequence and size. The CDRs

are the regions that determine the

specificity of the antigen – antibody

interactions. Complete IgG molecules

are difficult to crystallise, but their

enzymatic digestion with papain or

pepsin cleaves the Ig by the hinge

region, with one Fc and two identical

Fab fragments being obtained (Fig.

3.21). High resolution X-ray structural

information on these fragments has

shown all domains to have a similar

structure, either in light or heavy chains, either in variable or constant regions. This structure is the

so-called immunoglobulin fold, where a constant domain is formed by seven anti-parallel strands,

four of which form one β sheet and the remaining three form another.

Both β sheets are closely packed together in a barrel-like arrangement (Fig. 3.22 A). The loops

connecting the strands are short and thus the majority of the framework invariant residues are in the

β sheets. These structural features are similar for both heavy and light chains. Variable domains are

structurally similar to constant domains, but contain nine instead of seven β strands. The two

additional strands are placed in the important loop region that contains the hypervariable CDR2

(Fig. 3.22 B). These extra strands provide the scaffold that renders CDR2 closer to the other two

hypervariable loops CDR1 and CDR3. CDR2 and CDR3 are hairpin loops linking different strands

in the five-strand β sheet, while CDR1 is a cross-over between one strand from the five-strand sheet

and another from the four-strand sheet.

Figure 3. 21 Basic structure of an IgG and its fragments,produced by enzymatic digestion with papain (partiallyadapted from reference 30).

papain

antigen binding sites

heavychain

lightchain

CH1

CH2

CH3

VH

VL

CL

+ +

Fab Fab Fc

hingeregion CDR1

CDR2CDR3


138

Figure 3. 22 Theimmunoglobulin fold: A.general structure of aconstant IgG domain; B.general structure of avariable IgG domain(adapted from reference30).

The Fab fragment is the “arm” of the IgG molecule

that contains an intact antigen binding site. In this

fragment, the heavy and light chains are tightly and

extensively associated, in such a way that CL

associates with CH1 and VL with VH. Thus, a Fab

fragment consists of two globular regions, one with

the two constant domains and the other with the two

variable domains (Fig. 3.23). While the constant

domains associate by close interactions between the

almost perpendicular four-strand sheets from CH and

CL, the variable domains associate in a very different

manner. In this case, the interaction area is formed

by the five-strand β sheets, almost parallel to each

other, and defining a barrel structure of eight (four

from each five-strand sheet) antiparallel β strands.

This allows the CDR loops from both variable

domains to be located at the same end of the barrel,

forming the complete antigen binding site.

antigen

VL VH

CL CH1

CDRs

Figure 3. 23 Schematic representation ofa Fab fragment: VL/VH and CL/CH1 domainsassociate in such a way that CDRs can“grab” the antigen.

N

C

N

C

CDR1

CDR2

CDR3

A B


139

3.6.3 Molecular structure of the A15(FPS) – 4C4 complex in the crystal state

The protocols previously described for the crystallisation of similar FMDV peptide – Fab 4C4

complexes20,21 proved to be readily applicable to the present case (see Materials & Methods, section

4.4.2) and crystals as those shown in Fig. 3.24 were formed. Despite the difficulties found in growing

clean and perfect crystals, good diffraction data were acquired using synchrotron radiation at the

European Synchrotron Radiation Facility at Grenoble, France.

Figure 3. 24 Crystals of the A15(FPS) – 4C4complex.

After evaluating and internally scaling the diffraction data29 (Table 3.9), initial phases were assigned

taking the known structure of Fab 4C4 as the search model20,30. The model was initially subjected to

rigid body refinement and then treating each constant and variable domains as independent

structural units. The computed electron density maps clearly showed extra density at the antigen

binding site, corresponding to the peptide ligand. The peptide was then added to the model

structure, which was improved by cycles of manual rebuilding with program O31 and refinement with

the CNS package32. The final model had an R factor of 0.22 at a 2.3 Å resolution (Table 3.9) and is

represented in Fig. 3.25. The epitope – paratope contacts through hydrogen bonds are listed in

Table 3.10 and matched those previously observed with the native peptide A15 in complex with the

same mAb1.

A better illustration of the present structural study requires a global appreciation of the complex, as

well as of those formed between the same mAb and the three relevant single-point mutants,

previously solved by W. F. Ochoa and co-workers21. All four peptides were shown to interact with

Fab 4C4. As can be seen in Fig. 3.26, the three single-point mutants and the triple mutant all adopt

a similar quasi-cyclic conformation, also shared by the native sequence A151,21. This conformation

seems therefore to be a key feature in the antibody – FMDV peptide recognition process. A

stereoview of the A15(FPS) peptide fold in complex with Fab 4C4 is shown in Fig. 3.27.


140

Table 3.9 Crystallisation and diffraction data ofthe A15(FPS) – 4C4 complex.

Crystalisation and data collection

Space group P212121

Cell parameters (Å) 48.417

68.792

145.404

Resolution (Å) 25 – 2.2

Overall Completeness (%) 99.7

Rsymm (%) 7.2

Average I/σ 10.2

Total # of residues

Fab 429

Peptide 13

Total # of solvent

molecules 269

Volume solvent (%) 48.81

Diffraction agreement

Resolution (Å) 15 – 2.3

# of reflections 22153

Rfree 0.266

Rfactor 0.225

rms deviations from ideal distance

Bond length (Å) 0.0179

Bond angle (º) 2.2193

Average thermal factor (Å)

Fab 24.8

Peptide 24.3

Stereochemistry of main chain

Omega angle std. dev. 2.0

Bad contacts/100 res. 0.8

Zeta angle std. dev. 2.0

Stereochemistry of side chain

Chi-1 pooled std. dev. 11.5

Figure 3. 25 Structure of the A15(FPS) – 4C4complex (only Fab variable domains areshown); peptide side chains and some peptide– antibody hydrogen bonds are shown in moredetail (structures built with program SETOR33).

Figure 3. 26 Superposition of the structuresadopted by peptides A15(138F) – red,A15(140P) – dark blue; A15(142S) – orangeand A15(FPS) – light blue, when complexedwith Fab 4C4 (structures built using SETOR33).


141

Table 3.10 Hydrogen bonds between antigenic peptideA15(FPS) and the Fab fragment of mAb 4C4.

Bond interaction site in

Peptide FabLocation Distance

(Å)Tyr136 O Asn34 Nδ2 L1 2.6

Thr137 O Asp104 N H3 2.8

Thr137 Oγ1 Ser103 Oγ H3 3.0

Ser139 Oγ Asn96 Oδ1 L3 2.5

Arg141 O Asp98 N L3 2.9

Arg141 Nη1 Asn96 O L3 3.3

Arg141 Nη2 Asn96 Nδ2 L3 3.1

Asp143 O Tyr59 Oη H2 3.8

Asp143 Oδ1 Thr50 Oγ1 H2 2.7

Asp143 Oδ1 Arg99 Nε H3 2.7

Asp143 Oδ2 Arg99 Nη2 H3 2.7

His146 Nδ1 Tyr59 Oη H2 2.9

His146 Nδ2 Thr33 Oγ1 H1 3.0

Note: letters L and H in the location column stand for light andheavy chain, respectively, and are followed by numbers indicatingthe CDR where the antibody residue is located.

Figure 3. 27 Stereoview of the Fo-Fc omit map of peptide A15(FPS) at 2.3 Å resolution; the peptide finalmodel, including water molecules, was also shown for clarity; residues 149 and 150 were not considered in themodel (image generated with program SETOR33).


142

The similarity found in both peptide folding and epitope – paratope interactions for all FMDV

peptide – antibody complexes studied so far is remarkable. In view of this, we can devise the

following requirements for the recognition of FMDV peptides by anti-site A mAbs:

Fab-peptide interactions

Important interactions between peptide residues and the Fab cannot be bypassed, which is

confirmed by the absence of Fab-recognition of peptide mutants where key residues had been

replaced1,34,35. These interactions mainly involve peptide residues 141, 143 and 146 (Table 3.10).

They are similar to several other resolved structures1,20,21 and strong hydrogen bonds are seen to

stabilise the complex scaffold. Other observed interactions, particular to each complex, can explain

affinity differences between the different complexes, but do not seem to be absolutely necessary.

Peptide conformation

The hydrogen bonds between peptide residues and Fab do not seem to be sufficient to ensure a

strong interaction, since a precise peptide conformation seems to be an important requirement for

peptide-Fab union. Such conformation involves two important features:

Hydrophobic cavity

In all peptides studied a hydrophobic cavity was observed1,18-21, mainly formed by

residues 138, 144 and 147. These residues engage in strong hydrophobic interactions

through their side chains, stabilising peptide conformation. Mutations at these

positions would imply the loss of such cavity and, therefore, a decrease in the stability

of the peptide-Fab complex. However, that does not occur with the 138Ala→Phe

replacement; in this case, the Phe side chain is oriented into the cavity, and not only

does not disrupt hydrophobic interactions, but in fact stabilises the cavity itself (Fig.

3.26, 3.27).

Intrapeptide interactions

There is a set of interactions between the different peptide residues which contribute to

peptide folding, such as hydrogen bonds between the different nitrogen and oxygen

atoms of the main chain and also the presence of several water molecules bridge-

bonding peptide residues. These are key interactions that restrict mutations to those

amino acids able to preserve this type of structural arrangement. A remarkable

example is the 142Gly→Ser mutation, in which the side chain hydroxyl group of Ser

replaces a water molecule present in the structures where 142Gly is conserved (Fig.

3.28), maintaining the turn characteristic of the RXD motif (X=Gly or Ser). Pro at

position 140 also helps in the stabilisation of such a turn.


143

Figure 3. 28 Detailed view of the RXD turn, with X=Gly and X=Ser for the upper[peptide A15(138F)] and lower [peptide A15(142S)] structures, respectively; thecorresponding Fo-Fc omit maps are also shown. As it can be seen, the RGD turn isheld up by hydrogen bridging between Ser139 O – H2O – Leu144 N (above), whilethe RSD turn is stabilised by similar hydrogen bridging between Ser139 O – Ser142Oγ – Leu144 N (below). Structures were built with SETOR33.


144

3.7 Recapitulation

This chapter was focused on the study of 15-residue peptides from the GH loop of FMDV. The

peptide sequences were based on the reference FMDV strain C-S8c1, bearing combinations of the

replacements 137T→I, 138A→F, 140A→P, 142G→S and 148T→I. These amino acid replacements were

interesting in the sense that the corresponding single-point mutants had been previously found to be

significantly antigenic towards several anti-GH loop neutralising mAbs1. Therefore, the question was

raised whether multiple combination of these replacements could lead to positive synergistic effects,

yielding promising peptide antigens. Also, three such replacements deserved further attention:

position 138 was known to play a role in intrapeptide interactions1,20 and the much larger size of the

Phe side chain seemed unlikely to be “ignored” by antibodies; the introduction of a Pro in the loop,

with its peculiar structural behaviour, deserved to be analysed; and the effect of replacing Gly by Ser

in the highly conserved, key RGD motif, also captured our attention.

These peptides were fully characterised as FMDV antigens by means of SPR studies, towards three

anti-site A mAbs. It was immediately observed that this peptide family was quite different from the

one discussed in chapter 2. In fact, all these peptides displayed high antigenicities, which prevented

the kinetic study of their interactions with the mAbs due to mass-transport limitations. High

association rate constants, very slow dissociations and/or incomplete surface regeneration were

persistently observed and alternative solution affinity SPR analyses were performed. These

confirmed the high peptide – antibody affinities expected, generally comparable to or even higher

than those displayed by the native sequence represented by peptide A15.

The fact that these multiple mutants were fully recognised by anti-site A neutralising mAbs was quite

interesting and led to other questions, such as how did the mutations affect peptide conformation

and how did the antibody paratope adapt to these mutations. We approached the first question

through a two-dimensional 1H-NMR study of peptide A15(FPS) in solution and the second one by

means of an X-ray diffraction crystallography study of the complex formed between mAb 4C4 and

the same peptide. The NMR characterisation of peptide A15(FPS) in solution showed that this

peptide had a conformational behaviour quite similar to that previously observed for the native

sequence A1526,27. Data suggested an open turn in the RSD region followed by an incipient short α-

helix up to residue 147, features which had been previously recognised in peptide A15 and

regarded as antigenically relevant26,27,34.


145

The diffraction study of the A15(FPS) – 4C4 complex showed that the pattern of antibody – antigen

interactions is identical for all FMDV peptides studied so far1,18-21. Thus, a stable mAb – peptide

complex can be formed as long as some key requisites are fulfilled. These involve specific residues

committed in direct epitope – paratope contacts (141Arg, 143Asp, 146His) and residues able to stabilise

a particular peptide conformation. This conformation corresponds to a quasi-cyclic folding around a

hydrophobic cavity defined by residues 138, 144 and 147 and to other important intrapeptide

hydrogen bonds defining the central open turn involving positions 141, 142 and 143. Amino acid

replacements that not only do not disrupt, but even help to promote these essential requirements for

mAb recognition can yield peptides with significant reactivity towards anti-FMDV neutralising mAbs

and thus useful as FMDV antigens.

References


2 Valero, M. L. “Mimetización estructural e inmunogénica del sitio antigénico principal del virus de lafiebre aftosa” (Ph. D. Thesis), Department of Organic Chemistry – University of Barcelona: 1997.

3 Mateu, M. G., Valero, M. L., Andreu, D. and Domingo, E. (1996) Systematic replacement of aminoacid residues within an Arg-Gly-Asp containing loop of foot-and-mouth disease virus and effect on cellrecognition, J. Biol. Chem. 271, 12814-12819.





8 Albericio, F., Andreu, D., Giralt, E., Navalpotro, C., Pedroso, E., Ponsati, B. and Ruiz-Gayo, M.(1989) Use of the Npys thiol protection in solid phase peptide synthesis, Int. J. Peptide Protein Res.34, 124-128.

9 Benito, A., Mateu, M. G. and Villaverde, A. (1995) Improved mimicry of a foot-and-mouth diseasevirus antigenic site by a viral peptide displayed on β-galactosidase surface, Biotechnology 13, 801-804.

10 Carbonell, X., Benito, A. and Villaverde, A. (1996) Converging antigenic structure of a recombinantviral peptide displayed on different frameworks of carrier proteins, FEBS Lett. 397, 169-172.


146

11 Feliu, J. X., Benito, A., Oliva, B., Avilés, F. X. and Villaverde, A. (1998) Conformational flexibility in ahighly mobile protein loop of foot-and-mouth disease virus: distinct structural requirements for integrinand antibody binding, J. Mol. Biol. 283, 331-338.

12 Feliu, J. X. and Villaverde, A. (1998) Engineering of solvent-exposed loops in Escherichia coli β-galactosidase, FEBS Lett. 434, 23-27.


14 Lapko, V. N., Jiang, X. Y. and Smith, D. L. (1998) Surface topography of phytocrome A deducedfrom specific chemical modification with iodoacetamide, Biochemistry 37, 12526-12535.

15 Nieba, L., Krebber, A. and Plükthun, A. (1996) Competition BIAcore for measuring true affinities:large differences from values determined from binding kinetics, Anal. Biochem. 234, 155-165.

16 “BIAapplications Handbook”, (Pharmacia Biosensor AB, 1994) Uppsala, Sweden.17 Lazareno, S. and Birdsall, N. J. (1993) Estimation of competitive antagonist affinity from functional


18 Verdaguer, N., Mateu, M. G., Andreu, D., Giralt, E., Domingo, E. and Fita, I. (1995) Structure of themajor antigenic loop of foot-and-mouth disease virus complexed with a neutralizing antibody: directinvolvment of the Arg-Gly-Asp motif in the interaction, EMBO J. 14, 1690-1696.

19 Verdaguer, N., Mateu, M. G., Bravo, J., Domingo, E. and Fita, I. (1996) Induced pocket toaccomodate the cell attachment site Arg-Gly-Asp motif in a neutralizing antibody against foot-and-mouth disease virus, J. Mol. Biol. 256, 364-376.

20 Ochoa, W. F., Kalko, S., Mateu, M., Gomes, P., Andreu, D., Domingo, E., Fita, I. and Verdaguer, N.(2000) A multiply substituted GH loop from foot-and-mouth disease virus in complex with aneutralizing antibody: a role for water molecules, J. Gen. Virol. 81, 1495-1505.

21 Ochoa, W. F. et al., manuscript in preparation.22 Wüthrich, K. “NMR of proteins and nucleic acids”, Wiley, New York (1986).23 Braunschweiler, L. and Ernst, R. R. (1983), J. Magn. Reson. 53, 521.24 Kumar, A., Ernst, R. R. and Wüthrich, K. (1980), Biochem. Biophys. Chem. Comm. 95, 1.25 Bothner-By, A. A., Stephens, R. L., Lee, J., Warren, C. D. and Jeanloz, R. W. (1984) Structure




28 Branden, C. and Tooze, J., “Introduction to protein structure”, Garland Publishing Inc., New York(1991).

29 Otwinowsky, Z. and Minor, W. (1997) Processing of X-ray diffraction data collected in oscillationmode, Methods Enzymol. 276, 307-326.

30 Rossman, M. G. (Ed.) “The molecular replacement method”, Gordon & Breach, New York (1972).31 Jones, T. A., Zou, J. Y., Cowan, S. W. and Kjeldgaard, M. (1991) Improved methods for building

protein models in electron density maps and the location of errors in these models, Acta Crystallogr. A47, 110-119.

32 Brünger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang,J.-S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T. and Warren, G.L. (1998) Acta Crystallogr. D 54, 905-921.

33 Evans, S. V. (1993) SETOR: hardware-lighted three-dimensional solid model representations ofmacromolecules, J. Molec. Graphics 11, 134-138.

34 Mateu, M. G. (1995) Antibody recognition of picornaviruses and escape from neutralisation: astructural view, Virus Res. 38, 1-24.

35 Domingo, E., Verdaguer, N., Ochoa, W. F., Ruiz-Jarabo, C. M., Sevilla, N., Baranowski, E., Mateu,M. G. and Fita, I. (1999) Biochemical and structural studies with neutralising antibodies raised againstfoot-and-mouth disease virus, Virus Res. 62, 169-175.

Conclusions


148

Conclusions

149

1 A reliable SPR method for the kinetic analysis of binding between peptide antigens and

immobilised antibodies has been established, despite the small size of the analytes. The

interactions were well described by a simple 1:1 bimolecular interaction model and data

were self-consistent, reproducible and in total agreement with previous competition ELISA

screenings.

SPR kinetic analysis was, therefore, proven to be adequate for the functional

characterisation of small FMDV peptide antigens.

2 Different combinations, reproduced by linear 15-residue peptides, of the four amino acid

replacements found in the GH loop of FMDV isolate C-S30 were seen to be additive in

ELISA and kinetic SPR assays. Whereas increasing the size of the C-S30 peptide did not

cause any marked effect, overnight incubation with mAb in solution led to an antigenic

reversion of peptide A15S30 towards mAbs 4C4 and 3E5, but not SD6. A similar effect was

observed upon peptide cyclization. Solution NMR studies of both linear and cyclic C-S30

peptides showed that structural features formerly associated with peptide antigenicity were

more pronounced in the cyclic peptide.

Although the FMDV GH loop is a continuous (i.e., linear) antigenic region usually well

mimicked by linear peptides, conformation seems to have subtle, but important, effects

in the reproduction of recognition events involving peptides derived from field isolate

C-S30.

3 Antigenic FMDV peptides, comparable to or even better antigens than the wild type

sequence, can be obtained by combination of adequate amino acid replacements. The

peptide mutants display conformational and antibody – binding behaviour similar to those

characterising the native peptide.

A stable mAb – peptide complex can be formed as long as key requisites are fulfilled,

involving both residues committed in direct mAb – peptide contacts (141Arg, 143Asp,146His), and residues able to promote/stabilise a quasi-cyclic folding held up by a

hydrophobic cavity (defined by positions 138, 144 and 147) and by intra-peptide

hydrogen bonds delineating an open turn at the central region (positions 141, 142 and

143).


150

4. Materials & Methods


152

Materials & Methods

153

4.1 General procedures

4.1.1 Solvents and chemicals

Amino acids and resins for SPPS

Supplier

Advanced Chemtech, Propeptide,Bachem Feinchemikalien AG andCalbiochem-Novabiochem

Coupling reagents for SPPS

DIP, HOBtTBTU

FlukaPropeptide/Neosystem

Solvents and other reagents for SPPS

DCM, Normasolv p. a.DMF, peptide synthesisNMP, peptide synthesisNMM, peptide synthesisDIEA, p. s.Piperidine, p. a.TFA, p. s.Tert-butylmethylether* >99%Water

MeOH, MeCN, hplc gradeGlacial AcOH, p. a.Anisole, p. s.1,2-ethanedithiol, 90%+Thioanisole, >99%Triethylsilane, > 99%Hydrochloric acid 37%, p. a.Phenol, p. s.Ninhydrin, p. a.

*Stabilised with BTH and stored with sodium

ScharlauScharlau/PanreacApplied BiosystemsMerckMerck, AcrosAldrichKalieChemieFlukaDe-ionised and filtered with a MilliQ Plussystem (Millipore) to a resistivity superiorto 18 MΩ cm-1

Merck, Panreac, ScharlauMerckMerckAldrichFlukaAldrichMerckAldrichMerck


154

Materials and reagents for ELISA

PBS, tablets for buffer preparationBSA, fraction VGoat-anti-mouse IgG antibody - peroxidaseconjugate, blotting gradeorto-phenylenediamineHydrogen peroxide 35%, stabilised p. a.Tween 20, PlusonePVC 96-well plates, highly activated

SigmaBoehringer Manheim

Bio RadSigmaAcrosPharmacia BiotechTitertek, 77-172-05

Materials and reagents for SPR analysis

EDC (amine coupling kit), BIAcertifiedNHS (amine coupling kit), BIAcertifiedEthanolamine hydrochloride (amine couplingkit), BIAcertifiedHBS buffer, BIAcertifiedSensor chip CM5, BIAcertified

Oxalic acid, p. s.orto-phosphoric acid, p. s.Formic acid, p. a.Malonic acid, p. a.Sodium phosphate, p. a.Ethanolamine, p. a.Piperazine, p. a.Glycine, p. a.Potassium thiocianate, p. s.Magnesium chloride, p. a.

Biosensor AB

MerckScharlauMerckMerckMerckAldrichAldrichFlukaMerckMerck

Materials & Methods

155

General reagents and materials forbiochemistry

KLH, 18 mg/ml in 65% aqueousammonium sulphateGlutaraldehyde, p. a.HCA I (E. C. 4.2.1.1)Guanidine hydrochloride, purum >98%Citric acid, p. a.Sodium citrate, p. a.E.D.T.A. (Titriplex III), p. a.Sodium chloride, p. a.Tris, p. a.Urea, p. a.Iodoacetic acid, p. a.Cysteine, for molecular biologyβ-mercaptoethanol, purum >99%Ammonium sulphate, for biochemistrySodium azide, purum p. a. >99%Papain (E. C. 3.4.22.2)Dialysis membrane (MW cut-off=15-20 kDa),∅ =16 mmCentriprep-3 concentrators

CalbiochemSigmaSigma, C-4396FlukaMerckMerckMerckMerckMerckMerckMerckSigmaFlukaMerckFlukaSigma, P-4762

Servapor, 4415Amicon

Reagents for SDS-PAGE

Acrylamide/Bis 37.5:1, ultrapure gradeAPS, p. a.TEMED, p. a.Glycine, for molecular biologySDSGlycerol, p. a.Bromophenol blueCoomassie brilliant blue R-250

AmrescoServaMerckSigmaBoehringerSigmaBio RadBio Rad

Solvents and materials for NMRspectroscopy

Deuterium oxide, 99.95% Uvasol2,2,2-Trifluoroethanol - d3

1,4 – dioxane, for spectroscopy UvasolNMR quartz tubes, 5 mm OD (highmagnetic field)

MerckSDSMerckSDSSDS


156

4.1.2 Instrumentation

Amino acid analysis*

Mass spectrometry**- ES-MS- MALDI-TOF (matrices: ACH, SA)

UV-Vis spectrometry

pH-meter

Centrifuge

Lyophiliser

ELISA spectrometer

SPR instrument

NMR spectrometry***

X-ray diffraction****

SDS-PAGE

Beckman System 6300- elution with sodium salts- 250 × 4 mm column containing a

polysulphonate resin for cationic exchange- post-column detection by the ninhydrin

reaction

Fisons Instruments VG QuatroFinnigan MAT Lasermat 2000, Bruker II Biflex

Perkin-Elmer Lambda 5

Crison MicropH 2002

Beckman GS-15R

Virtis Freezemobile 12EL

Labsystem Multiskan MS

BIAcore 1000, IFC4 with recovery

Varian VXR500

MarResearch image plate detector (180 × 0.10mm, 1800 pixels); Rigaku RU-200B rotatinganode

Bio Rad Mini-PROTEAN II electrophoresis cell;Bio Rad gel dryer, model 583.

* amino acid analyses were performed by Dr. C. Carreño, Dr. M. L.Valero and Ms. M. E. Méndez, at the Servei deSintesi de Pèptids de la Universitat de Barcelona.** mass spectra were acquired by Drs. I. Fernández, M. Vilaseca, M. L. Valero and E. de Oliveira, at the Servei deEspectrometria de Masses de la Divisió de Ciències de la UB.*** NMR spectra were recorded by Dr. M. A. Molins at the Servei de RMN, SCT, Universitat de Barcelona.**** X-ray diffraction data were acquired by Ms. W. F. Ochoa at the European Synchrotron Facility in Grenoble.

Materials & Methods

157

4.1.3 Analytical methods

4.1.3.1. Qualitative ninhydrin assay

This assay serves to detect free amine groups on polymeric supports (resins) for SPPS and is

performed as described by Kaiser et al.1

4.1.3.2 Qualitative Ellman assay

This assay allows the detection of free thiol groups either in solution or on polymeric supports

compatible with aqueous media, according to Ellman et al.2

4.1.3.3. Amino acid analysis

The content and proportion of amino acid residues present in a free or resin-bound peptide are

determined by amino acid analysis (AAA), following a previous hydrolysis step. To hydrolyse a

peptide-resin3, 1 – 10 mg of dried resin are placed into a Pyrex glass tube and 250 µl of a 1:1 (v/v)

mixture of 12 M hydrochloric and propionic acids are added. The tube is sealed and hydrolysis is

carried out at 155 oC for 90 minutes. The procedure for a free peptide4 is similar, hydrolysing with 6

M HCl for 45 minutes. The hydrolysed mixture is then evaporated to dryness and the residue

dissolved in a known volume of a 0.06 M citrate buffer, pH 2.0. After filtration through a nylon filter

(∅ pore=0.45 µm), the sample is ready for AAA.

4.1.4 Chromatographic methods

4.1.4.1 High performance liquid chromatography

Analytical HPLC is performed in either of the following systems:

Waters – composed by a controller and a quaternary pump 600E with a low pressure

mixer, an automatic injection system Waters 712, a variable wavelength UV-Vis detector

490E and a integrator/recorder either D-2000 (Merck-Hitachi) or Chromatopac C-R5A.

Shimadzu – composed by two LC-6A pumps with a high pressure mixer, an SCL-6B

controller, an SIL-6B auto-injection system, a variable wavelength UV-Vis detector SPD-6A

and a integrator/recorder Chromatopac C-R6A.

The HPLC column is a 250 × 4 mm Nucleosil C18, with a reverse solid phase of octadecylsyloxane

(∅ beads=5 µm; ∅ pore=120 Å). The mobile phases are gradients of H2O (0.045% v/v TFA) and MeCN

(0.036% v/v TFA) at a 1 ml/min flow.


158

4.1.4.2 Medium pressure preparative liquid chromatography

Peptides are purified by MPLC in systems composed by LCD/Milton or Duramat ProMinent piston

pumps, variable wavelength Applied Biosystems or single wavelength Uvicord 2158 SD (LKB)

detectors at 220 nm, Ultrorac 2070II (LKB) or Gilson FC205 fraction collectors and Servoscribe

(Phillips) or Pharmacia-LKB REC101 recorders. Glass columns (!=200-300 mm, ∅ internal=25 mm)

with Vydac C18 reverse phase (∅ beads=15-20µm, ∅ pore=300 Å) are used and the mobile phase

consisted on a binary linear gradient of H2O and MeCN with 0.05% TFA at a constant flow of 120-

150 ml/h.

References

1 Kaiser, E., Colescott, R. L., Bossinger, C. D. and Cook, P. I. (1970) Color test for detection of freeterminal amino groups in solid-phase synthesis of peptides, Anal. Biochem. 34, 595-598.

2 Ellman, G. L. (1958) A colorimetric method for determining low concentrations of mercaptans, Arch.Biochem. Biophys. 74, 443-450.

3 Scotchler, J., Lozier, R. and Robinson, A. B. (1970) Cleavage of single amino acid residues fromMerrifield resin with hydrogen chloride and hydrogen fluoride. J. Org. Chem. 35, 3151-3152.

4 Steward, J. M. and Young, J. D. in “Solid Phase Peptide Synthesis” , 2nd ed., Pierce Chemical Co.,Rockford, Illinois (1984).

Materials & Methods

159

4.2 Solid-phase peptide synthesis

4.2.1 Solid-phase peptide synthesis protocols

4.2.1.1 Preparation of resins for peptide synthesis

The dry resin (MBHA1 or PEG-PS2) is placed in a polypropylene syringe containing a polyethylene

filter. The resin is allowed to swell in 40% TFA in DCM (1 × 1 min + 1 × 20 min), filtered and

neutralised with 5% DIEA in DCM (3 × 1 min), then washed with DCM (5 × 30 s) and DMF (3 × 1

min).

A mixture of the chosen two-functional spacer (or handle) and HOBt (3 eq each) is dissolved in a

minimum volume of DMF and added to the resin in the syringe. The coupling agent DIP (3 eq) is

then added and coupling is allowed to proceed overnight or until a negative ninhydrin test is

obtained. If the ninhydrin test remains positive after 18 hours of reaction, the resin is washed and a

re-coupling step (1 eq of all reactants in similar conditions) is performed. When coupling is

complete, the resin is filtered, thoroughly rinsed with DMF and dried.

4.2.1.2 Fmoc/ tBu chemistry

4.2.1.2.1 Manual synthesis3

Manual syntheses are performed in polypropylene syringes with a polyethylene porous filter. The

volumes of solvents and reagent solutions added should cover the entire resin to allow optimal

solvating and swelling of the beads. Stirring is done with a teflon rod and, at the end of each cycle,

excess reagents, by-products and solvents are eliminated by filtration and washing with DMF and

DCM.

Peptide chain elongation is performed according to the following steps:


160

Step Reagenta Treatment Time/min1 DMF Wash 3 × 12 20 % piperidine in DMF Pre-equilibrate 13 20 % piperidine in DMF Deblock (Fmoc removal) 104 DMF Wash 3 × 15 Fmoc-AA-OH/coupling agent/DIEAb Coupling 45 – 606 DMF Wash 3 × 17 DMF Wash 3 × 18 Ac2O/DIEA 1:1 in DMF Acetylation (block non-reacted

amino groups)15

9 DCM Wash 3 × 1a volumetric reagent/solvent proportionsb coupling agents used: DIP, TBTU; Base (DIEA) is required with TBTU only (2 eq DIEA/ 1 eq TBTU)

The coupling agents employed in this work were DIP and TBTU4, the latter requiring the addition of

base (DIEA) in a 2:1 molar proportion between base and reagent. The Fmoc-AA-OH were dissolved

in the minimum volume of DMF and, once coupling time was over and washing steps performed

(step 6), a ninhydrin assay was done. When the assay was negative, chain elongation proceeded to

the incorporation of the following amino acid residue, starting with the removal of the Fmoc group

(step 2 and the following). When the ninhydrin test was positive, a recoupling cycle was performed

(step 5 and the following). If the addition of recoupling steps could not improve coupling efficiency,

then acetylation (steps 7 – 9) could be used to block the non-reacted amino groups.

4.2.1.2.2 Machine-assisted synthesis

Peptides can also be synthesised in a MilliGen 9050 Plus PepSynthesiser, which dissolves amino

acids and coupling reagents and works with a continuous flow Fmoc/tBu chemistry (instead of

filtration steps after each cycle). The inlet/outlet detectors allow a constant monitoring of the

synthesis at each step and it is possible to choose synthesis scale and coupling reagents.

The general protocol consists on Fmoc-AA-OH/coupling agent dissolution in DMF, followed by

addition of 0.6 M DIEA in DMF. The coupling mixture is activated through a 5 min bubbling step

with nitrogen and subsequent transfer of the solution to the column reactor, which contains the

previously de-blocked resin. Chain elongation proceeds as follows:

Step Reagenta Treatment Flow /ml.min-1 Time1 DMF Wash 3 15 s2 20 % piperidine in DMF Pre-equilibrate 3 1 min3 20 % piperidine in DMF Deblock (Fmoc removal) 3 5 min4 DMF Wash 3 7 min5 Activated Fmoc-AA-OHb Coupling 3 5 s6 Activated Fmoc-AA-OH Coupling 3 60 min7 DMF Wash 3 4 mina volumetric reagent/solvent proportionsb coupling agents used: TBTU

Once the synthesis is completed, the peptide-resin is transferred to a polypropylene syringe to be

washed and dried as described in 4.2.1.2.1.

Materials & Methods

161

4.2.1.2.3 Machine-assisted parallel synthesis

Multiple peptide synthesis can be performed on an Abimed MAS 422 synthesiser, which allows the

simultaneous synthesis of up to 48 peptide-resins, using Fmoc/tBu chemistry with in situ activation.

The synthesis programmes are quite flexible in what concerns synthesis scale, number of coupling

steps per amino acid residue and duration of each step. This synthesiser operates as follows: the

Fmoc-AA-OH (0.6 M in DMF, except for Fmoc-His(Trt)-OH and Fmoc-Phe-OH, which are

dissolved in NMP), the coupling reagent (0.5 M TBTU) and the base (4 M NMM), which are

previously dissolved and placed in appropriate racks and sealed with septa, are added to 2 ml

syringes containing previously deblocked and washed resin, and fitted to a 48-port manifold system.

The addition is done in a pre-defined sequence, according to the reactivity of each amino acid

residue which is added to each syringe. Chain elongation proceeds according to steps 1-6:

Step Reagenta Treatment Number ofrepeats

Time/min

1 20 % piperidine in DMF Deblock 2 × 1 ml 5b

2 DMF Wash 2 × 1 ml 0.53 DMF Wash 3 × 0.3 ml 0.54 Fmoc-AA-OH/TBTU/NMM Coupling 1 30c

5 DMF Wash 12 × 1 ml 0.56 DMF Wash 2 × 0.3 ml 0.57d DMF Wash 2 × 1 ml 0.58 DMF Wash 3 × 0.3 ml 0.59 20 % piperidine in DMF Deblock 2 × 1 ml 5b

10 DMF Wash 2 × 1 ml 0.511 DMF Wash 3 × 0.3 ml 0.512 DCM Wash 3 × 0.3 ml 0.5a volumetric reagent/solvent proportions;b reaction time is increased along chain elongation;c 100 µl of DCM are added at 80% of the coupling total time ;d steps 7-12 correspond to the final cycles for resin deblocking, washing and drying.

4.2.1.2.4 Peptide cleavage from the resin and removal of side-chain protecting groups

Up to 500 mg of dry resin (Nα - Fmoc previously removed) are placed in a Falcon centrifuge tube.

The cleavage reagent (cocktail R5) is prepared: 90% TFA, 2% anisole, 5% thioanisole and 3% 1,2 –

ethanedithiol, and added to the resin at the proportion of 1 ml cocktail : 100 mg resin. The reaction

is carried out at room temperature for 2 hours with constant shaking. Anhydrous tert-

butylmethylether (40 ml) is then added and the mixture cooled at -78 oC for peptide precipitation.

The suspension is stirred, then centrifuged at 4 oC and 4000 r. p. m. for 15 minutes, after which the

supernatant is decanted. The procedure is repeated 5 times from the ether addition step. The final

peptide precipitate is dried with nitrogen, resuspended in AcOH 10% and filtered through a

polypropylene syringe containing a polyethylene filter. The peptide solution is then lyophilised.


162

4.2.2 Synthesis of peptides from the GH loop of FMDV

4.2.2.1 Peptides for SPR and ELISA

Peptide sequences and their characterisation are compiled in chapters 2 and 3.

4.2.2.1.1 Linear 15-mer peptides from the FMDV strain C1-Barcelona (C-S30)

These peptides were prepared by machine-assisted parallel synthesis at a 25 µmol scale (section

4.2.1.2.3) on an Fmoc-AM-MBHA resin (0.51 mmol/g), where the handle AM6 was incorporated as

described in 4.2.1.1. The usual side-chain protecting groups in Fmoc/tBu synthesis were employed:

Asp(OtBu), Arg(Pmc), His(Trt), Ser(tBu), Thr(tBu) and Tyr (tBu).

The peptides, which were obtained as C-terminal carboxamides, were cleaved and deprotected as

described in 4.2.1.2.4, with some modifications: a polystyrene pipette tip was adapted to each resin-

containing syringe and the latter was introduced in a 10 ml Sarsted polypropylene tube, previously

containing the cleavage cocktail (1 ml). The cocktail was then sucked up into the syringe and, after

air bubbles were carefully expelled, the reaction was carried out as previously described. Once the

first 2-hour period was over, the peptide crudes were expelled from the syringes to the Sarsted tubes

and additional 0.5 ml of fresh cleavage cocktail were sucked up into the syringes and reaction

proceeded for further 30 minutes. The second filtrates were mixed with the first ones and processed

as described in chapter 4.2.1.2.4.

Crude products were analysed by HPLC (5→95% B and 10→45% B), AAA and ES MS(+) or

MALDI-TOF MS (section 4.1.4.1). Peptides with more than 15% of byproducts were purified by

reverse phase MPLC (5→25% B, section 4.1.4.2).

4.2.2.1.2 Larger versions of the FMDV GH loop: peptides A21 and A21S30

The 21-residue peptides A21 and A21S30 were synthesised either by manual Fmoc/ tBu chemistry

(section 4.2.1.2.1) or by machine-assisted synthesis on a MilliGen 9050 Plus PepSynthesiser (section

4.2.1.2.2) at a 50 µmol scale. In either case, an Fmoc-AM-PEG-PS resin (0.20 mmol/g) was

employed and procedures were as already described in previous chapters. Peptide cleavage from

the deblocked resin was done as described in 4.2.1.2.4. Both peptides were purified by reverse

phase MPLC (10→30% B) and characterised as usual.

4.2.2.1.3 Cyclic versions of the FMDV GH loop: peptides cyc16S30 and cyc16147Val

The cyclic peptides, cyc16S30 and cyc16147Val, were synthesised by intra-molecular disulphide

bridge formation7 of the corresponding linear bis-thiol precursors (see chapter 2). The linear bis-thiol

Materials & Methods

163

peptides were synthesised by similar methods as those described in 4.2.2.1.2, using Fmoc-Cys(Trt)-

OH and Fmoc-Ahx-OH in addition to the other protected amino acids usually employed (section

4.2.2.1.1). Peptide cleavage, characterisation and purification by MPLC (10→25% B) were

performed as already described, having the extra care that peptides were always kept under acidic

conditionsA to avoid intermolecular disulphide bridge cross-linking. Once the linear bis-thiol

precursors were purified and lyophilised, cyclization proceeded by air oxidation at high peptide

dilution and pH 8. The peptide was added stepwise to 100 mM ammonium bicarbonate buffer, pH

8, to a final concentration of 50 µM, under vigorous stirring, and left to react at open air.

The extent of cyclization was monitored by HPLC (10→45% B) and by the Ellman qualitative assay,

and usually reached completion within 1 hour. The reaction was stopped upon dropwise addition of

glacial acetic acid until pH 3. Cyclic peptides were characterised by HPLC, AAA and MALDI-TOF

MS and repeatedly lyophilised from water to eliminate the ammonium salts.

4.2.2.1.4 Linear 15-mer peptides bearing the mutations 137T→I, 138A→F, 140A→P, 142G→S

and 148T→I

These peptides were prepared by machine-assisted parallel synthesis at a 25 µmol scale on an

Fmoc-AM-MBHA resin (0.51 mmol/g), by similar methods as those described in section 4.2.2.1.1.

Peptide cleavage was performed according to the same section and, again, crude peptides having

more than 15% of impurities were purified by reverse phase MPLC (15→45% B). Peptide

characterisation and quantification was as previously described.

4.2.2.2 Single syntheses of peptides for structural studies

The set of FMDV peptides for NMR and X-ray diffraction studies were prepared in individual

syntheses and exhaustively purified to meet the purity requirements of both techniques. Syntheses

were performed at a 100 µmol scale on an Fmoc-AM-MBHA resin (0.20 mmol/g) in the MilliGen

9050 Plus PepSynthesiser by methods similar to those described. Peptide purification by MPLC was

carried out as usual, regarding that at least 5 mg of 99% pure peptide should be obtained.

A MPLC peptide purification fractions were collected on tubes containing 100 µl of 0.1 M AcOH.


164

References

1 Barany, G. and Merrifield, R. B. in “The peptides”, vol. 2, 1st ed., Gross, E. and Meinhofer, R. B.(Eds.), Academic Press, New York (1980).

2 Barany, G. and Albericio, F., Mild orthogonal solid-phase peptide synthesis, in “Peptides 1990:Proceedings of the 21st European Peptide Symposium”, Giralt, E. and Andreu, D. (Eds.), ESCOM,Leiden (1991), pp 139.

3 Fields, G. B. and Noble, R. L. (1990) Solid-phase peptide synthesis utilizing 9-fluorenylmethoxycarbonyl amino acids, Int. J. Peptide Protein Res. 53, 161-214.

4 Knorr, R., Trzeciak, A., Bannwarth, W. and Gillesen, D. (1989) New coupling reagents in peptidechemistry, Tetrahedron Lett. 30, 1927-1930.

5 Albericio, F., Kneib-Cordonier, N., Biancalana, S., Gera, L., Masada, R. I., Hudson, D. and Barany,G. (1990) Preparation and application of the 5-(4-(9-fluorenylmethyloxycarbonyl)-aminomethyl-3,5-dimethoxyphenoxy)valeric acid (PAL) handle for the solid-phase peptide synthesis of C-terminalpeptide amides under mild conditions, J. Org. Chem. 55, 3730-3743.

6 Bernatowicz, M. C., Daniels, S. B. and Köster, H. (1989) A comparison of acid labile linkage agentsfor the synthesis of peptide C-terminal amides, Tetrahedron Lett. 30, 4645-4648.

7 Andreu, D., Albericio, F., Solé, N. A., Munson, M. C., Ferrer, M. and Barany, G. Formation ofdisulfide bonds in synthetic peptides and proteins in “Methods in molecular biology, vol. 35: Peptidesynthesis protocols”, Pennington, M. W. and Dunn, B. M. (Eds.), Humana Press Inc., Totowa, NewJersey (1994), pp 91-169.

Materials & Methods

165

4.3 Antigenic evaluation of the FMDV peptides

4.3.1 SPR analysis of peptide-antibody interactions

The technical and scientific bases for real-time surface plasmon resonance biospecific interaction

analysis are exposed in sections 0.1 to 0.3 of the present work. A BIAcore SPR biosensor was used,

and standard amide immobilisation chemistry on a CM5 sensor chip were employed. Both

immobilisation chemistry and sensor chip features are described in section 0.2. Standard

procedures, following manufacturer’ s instructions, were employed as far as possible. Equipments

and reagents for biosensor analysis are specified in section 4.1.

4.3.1.1 Peptide and mAb solutions for biosensor analysis

Peptide stock solutions ca. 2.5 mM in 0.1 M acetic acid were prepared and quantitated by AAA.

Solutions for BIAcore analysis were obtained by 1000-fold and subsequent serial dilutions in HBS.

Stock solutions of mAbs SD6 and 4C4 (in PBS with 0.02% sodium azide, pH 7.3) were desalted

and buffer-exchanged on an NAP-5 Sephadex G-25 column (Pharmacia Biotech) and final mAb

concentrations were determined by measurement of optical density at 280 nm, considering that 1

OD280 ≈ 0.75 mg (protein)/ml.

4.3.1.2 Optimisation of the direct kinetic analysis of immobilised mAb – peptide interactions

SD6 solutions (100 and 50 µg/ml, in either 10 mM sodium acetate, pH 5.5, or 5 mM sodium

maleate, pH 6.5) were separately injected (30 µl) at 5 µl/min over a non-activated sensor surface, to

determine which gave the most efficient mAb pre-concentration into the dextran matrix.

Three SD6 surfaces were prepared using the standard amine coupling procedure as described by the

manufacturer1: each carboxymethyl surface was activated with a 35 µl injection (at 5 µl/min) of a

solution containing 0.2 M EDC and 0.05 M NHS, and SD6 was then coupled at three different

densities by injecting over each surface 35 µl of 50, 5 and 3 µg/ml SD6 in 10 mM sodium acetate


166

buffer, pH 5.5, respectively. Non-reacted activated groups were then blocked by a 30 µl injection of

ethanolamine hydrochloride and remaining non-covalently bound material was washed off in a

regeneration step with a 3-min pulse of 100 mM HCl. Surface densities obtained were of 8000,

1700 and 800 RU, respectively, where 1 RU (resonance unit) corresponds to 1 ng (protein)/mm2

(surface).

A few sets of experiments, using A15 as analyte, were run on the three SD6 surfaces at different

peptide concentrations (ranging from 1 to 2500 nM) and flow rates (5 and 60 µl/min). All

experiments were done with HBS as running buffer at 25 oC, using the kinjection mode.

Sensorgrams were generated with 7-min peptide injections in the HBS flow, followed by 6-min

dissociation in running buffer and then by a 2-min regeneration step with 100 mM HCl.

Biosensor data were prepared, modelled and fitted by means of the BIAEvaluation 3.0.2 software2

(Biosensor AB, 1994-97, run on Windows ’ 95). The quality of the fits was assessed by visual

comparison between experimental and modelled sensorgrams, as well as by statistical parameters

such as χ2 and standard errors associated to the calculated constants, or by further inspection of

residual distribution.

4.3.1.3 Systematic screening of FMDV peptide antigens: validation of the SPR methodology

Immobilisation of mAbs SD6 and 4C4 was performed as described in the previous section.

Biospecific surfaces were obtained by injecting 35 and 16 µl of the 5 µg/ml SD6 and 4C4 solutions

in 10 mM acetate buffer pH 5.5, respectively. Following the capping step with ethanolamine

hydrochloride, remaining non-covalently bound molecules were washed off with a 3-min pulse of

100 mM HCl or 10 mM NaOH for SD6 or 4C4 surfaces, respectively. The final immobilisation

responses were of about 1600 RU.

All kinetic SPR analyses were run at a 60 µl/min HBS flow and each peptide was analysed at 6

different concentrations, ranging from ca. 80 to 2500 nM for SD6 and ca. 40 to 1250 nM for 4C4.

Sensorgrams were generated by kinjections of peptide solutions with 90 s association steps followed

by 240 s dissociation in running buffer. A 90 s pulse of 100 mM HCl or 10 mM NaOH (SD6 and

4C4 surfaces, respectively) was applied to regenerate the surfaces at the end of each cycle and wash

steps (needle, IFC, system flush) were added to avoid carry-over. The pentadecapeptide A15scr,

containing the constituent amino acids of A15 in scrambled form, was injected under the same

conditions as a control for non-specific binding to the sensor chip surfaces.

After subtracting the response of peptide A15scr to the responses of the relevant peptides, data were

prepared, modelled and fitted by means of BIAevaluation software as already described.

Materials & Methods

167

4.3.1.4 Antigenic evaluation of FMDV C-S30 peptides by direct SPR kinetic analysis

Peptides from the FMDV C-S8c1 and C-S30 GH loops (syntheses described in section 4.2.2) were

screened by SPR as described in section 4.3.1.3. This screening included an additional anti-FMDV

monoclonal antibody, mAb 3E5. This mAb was purified from ascitic fluid as follows.

4.3.1.5 Purification of mAb 3E5 from ascitic fluid

Ascitic fluid was unfrozen and divided into 1 ml aliquots, which were then centrifuged for 5 min at

10000 r.p.m. (4 oC). The supernatants were pooled and an equivalent volume of buffer A was

added (buffer A: 112.4 g/l glycine, 175.4.g/l NaCl, pH 8.9). This mixture was again divided into 1

ml aliquots and centrifuged. The aqueous fraction was separated from lipids and pellets.

A HiTrap protein A – Sepharose affinity column (Pharmacia Biotech), coupled to a 2132

Microperpex (LKB) peristaltic pump, was prepared by extensive rinsing, with 100 mM sodium citrate

buffer, pH 3, then with buffer A, at a constant flow of 20 ml/h.

The sample was applied to the column and eluted with buffer A until OD280≤0.01; elution then

proceeded with 100 mM sodium citrate buffer, pH 5.0, and fractions were collected on glass tubes

containing 100 µl of 100 mM Tris-HCl buffer, pH 8.5. Fractions were monitored at 280 nm and,

once an absorbance peak was observed, the elution buffer was changed to 100 mM sodium citrate,

pH 3, until a second smaller peak was observed. Fractions collected at pH 5 with OD280≥0.5 were

pooled and dialysed against PBS overnight (PBS, phosphate buffered saline: 137 mM NaCl, 2.7

mM KCl, 8 mM Na2HPO4, 1.5 mM KH2PO4, pH 7.3; 3 × 1 l) and final concentration was

determined by optical density at 280 nm. For the present purpose, further steps for mAb

concentration (upon precipitation with ammonium sulphate and subsequent dialysis against PBS

and gel filtration on Sephadex G-100) were not required.

4.3.1.6 SPR interaction analysis of free FMDV C-S30 peptides with immobilised mAbs SD6, 4C4

and 3E5

Immobilisation of mAbs SD6 and 4C4 were as described in section 4.3.1.3. MAb 3E5 was

immobilised similarly to the described for mAb 4C4. Kinetic analyses were run as described in

section 4.3.1.3 and peptide concentrations injected over the 3E5 surface ranged from ca. 35 to ca.

625 nM. Data evaluation procedures were as already described, using the 1:1 langmuirian binding

(either with or without baseline drift) kinetic model2.


168

4.3.1.7 SPR interaction analysis of immobilised FMDV C-S30 peptides with free mAbs SD6, 4C4

and 3E5

Peptides A15, A15S30, A21, A21S30, A15Brescia, cyc16S30, cyc16147Val and A15scr were

immobilised by methods identical to those described for mAb immobilisation. Following a surface-

activation step similar to that previously described, 25 µl of the peptide solutions (200 µg/ml in 10

mM acetate buffer, pH 5.0) were injected over the surface and final immobilisation responses of ca.

170 RU were obtained. Surfaces were regenerated by 2-min pulses of 100 mM HCl.

mAb solutions, 25 to 800 nM in HBS, were injected as previously described for peptides and

sensorgrams were run, modelled and fitted as before.

4.3.1.8 Antigenic evaluation of multiply substituted FMDV peptides by direct SPR kinetic analysis

Direct SPR kinetic analysis of peptides reproducing combinations of the substitutions 137T→I,138A→F, 140A→P, 142G→S and 148T→I could not be accomplished due to experimental problems

related either to extremely slow complex dissociation or to insufficient surface regeneration (see

chapter 3).

4.3.1.9 Alternative strategies for surface regeneration

A screening of alternative regeneration conditions, based on the multi-cocktail approach of

Andersson and co-workers5, was performed (see chapter 3, Tables 3.5 and 3.6).

The evaluation of the screening cocktails is performed by a 10-min analyte injection at a flow rate of

2 µl/min, followed by 30 s injections (at 20 µl/min) of a given cocktail until analyte level decreases to

30% or less of the original value. At this point, a new analyte injection is applied, followed by

injections of another regeneration cocktail. After a regeneration effectiveness is attributed to each

cocktail, a more extensive optimisation can be performed, relying on combinatorial mixing of the

different stock cocktails.

Cocktails including stock solutions A, B, C, U or I’ (see chapter 3) were screened as described, but

none led to satisfactory results.

4.3.1.10 Indirect SPR kinetic analysis by competition assays using an FMDV peptide – carrier

protein conjugate

The preparation of a 1:1 peptide – protein conjugate made use of the chemistry of the Npys thiol

protecting group7. An Fmoc-peptide A15-AM-MBHA resin was prepared by solid phase peptide

Fmoc/tBu chemistry (scale 50 µmol) on an Fmoc-AM-MBHA resin (0.20 mmol/g) in the MilliGen

9050 Plus PepSynthesiser. After removal of the Nα - Fmoc protecting group and subsequent

Materials & Methods

169

washing cycles, Boc-Cys(Npys)-OH was manually coupled (using 4 eq. of DIP in DCM) to the

peptide-resin. Once coupling was completed and resin conveniently washed and dried, peptide

cleavage and side-chain deprotection (except for the Npys group, stable to strong acids) was

performed with a 90% TFA, 5% H2O and 5% Et3Si cleavage mixture The Cys(Npys)A15 peptide

was characterised by AAA, HPLC and MALDI-TOF MS and lyophilised prior to heterodimerisation

with protein HCA I (human carboxy anhydrase I, a 28 kDa monomer with a single Cys residue)8.

The reaction of disulphide heterodimerisation between Cys(Npys)A15 and HCA I was performed

under denaturating acidic conditions. HCA I (5 mg, 0.17 µmol) was dissolved in the minimum

amount of 6 M guanidine hydrochloride in water (pH 4.2) and left under magnetic stirring for 30

min, until a positive Ellman test was obtained. Peptide Cys(Npys)A15 (3.13 mg, 1.7 µmol) was

added to the protein solution and reaction was allowed to proceed overnight, monitored by HPLC

(15→65% B) and by the qualitative Ellman assay. When reaction reached equilibrium, the mixture

was diluted to 2 ml with 6 M guanidine hydrochloride and dialysed during 48 h against decreasing

concentrations of guanidine hydrochloride (4 M, 1 × 1 l → 2 M, 1 × 1 l → 1 M, 1 × 1 l → water, 3

× 1 l). The dialysed solution was characterised by HPLC, MALDI-TOF MS and AAA, and the total

protein content was determined by optical density at 280 nm. After adding sodium azide (0.03%),

the protein solution was divided into 500 µl aliquots that were stored at –20 oC prior to SPR assays.

The protein heterodimer HCA I – CysA15 was injected on sensor chip flow cells where mAbs SD6,

4C4 and 3E5 had previously been immobilised by standard procedures (ca. 600 RU of each mAb

were immobilised). A total protein concentration of 300 nM was injected over the three mAb

surfaces and also over a fourth mock surface (EDC/NHS activation plus ethanolamine hydrochloride

capping, without protein injection) for non-specific binding evaluation. Injections were performed as

already described for peptide injection. Protein response was studied at three different pH values,

upon protein dilution in either HBS (pH 7.3), 10 mM Tris – HCl (pH 8.5) or 10 mM sodium acetate

(pH 5.5) buffers. In neither case was a specific response observed. A non-specific response was

observed at pH 5.5, due to the electrostatic attraction between protonated protein – pI ≈ 6 – and the

negatively charged carboxymethyl dextran matrix.

4.3.1.11 Indirect SPR kinetic analysis by competition assays using an engineered recombinant

protein expressing the GH loop of FMDV C-S8c1

Protein JX249A is a recombinant β-galactosidase from Escherichia coli9, with solvent exposed loops

where a 24-residue peptide from the GH loop of FMDV C-S8c1

(TT136YTASARGDLAHLTT150THARHLP)10 has been inserted. The protein is a 472 kDa

homotetramer with one GH loop per monomer and is highly antigenic towards a panel of anti-GH

loop antibodies.

Protein samples (305 µg/ml in buffer Z: 0.06 M sodium hydrogen phosphate, 0.04 M sodium di-

hydrogen phosphate, 0.01 M potassium chloride and 1 mM magnesium sulphate) were diluted in

HBS to a total FMDV peptide concentration of 624 nM (and subsequent serial dilutions) and tested


170

for SPR analysis. Protein was injected over SD6, 4C4 and 3E5 surfaces (ca. 600 RU of mAb

immobilisation level), as previously described. Insufficient mAb surface regeneration was observed,

which could not be overcome by alternative regeneration procedures based on the multi-cocktail

approach described in section 4.3.1.9. The regeneration strategy that yielded better results consisted

on three successive 1-min injections of 40 mM NaOH, but even so surface life-time was significantly

reduced due to the inefficient removal of bound protein and consequent decreasing availability of

mAb binding sites.

In spite of the surface regeneration problems observed, a set of preliminary tests for SPR surface

competition analysis was performed using peptides A15 and A15scr as competitors. A constant

amount of protein JX249A (total final concentration in FMDV peptide = 160 nM) was added to six

peptide solutions with concentrations ranging from 0 to 300 nM. The peptide – protein mixtures

were then injected as previously described for peptide injections and three 1-min pulses of 40 mM

NaOH for partial surface regeneration were added at the end of each injection. Data was processed

with the BIAEvaluation 3.0.2 software, using the heterogeneous analyte kinetic model2 (competition

between two different analytes, section 0.3).

4.3.1.12 Indirect SPR kinetic analysis by competition assays using cysteine-capped protein JX249A

Since one of the possible causes for JX249A irreversible binding to mAb surfaces could be the fact

that all cysteine thiol groups in native bacterial β-galactosidases are reduced, capping of the thiol

groups was performed using iodoacetic acid11. A denaturating solution (2 ml, 7.5 M urea, 4 mM

EDTA, 0.25 Tris-HCl, pH 8.5) was added to a 170 µg/ml JX240A solution in buffer Z (2 ml),

corresponding to 46 nmol of total cysteine. β-mercaptoethanol was added (2 µl, 26 µmol) to cleave

any disulphide bonds in the protein and the mixture was left to stand at 60 oC for 1 h in the dark,

followed by another hour at room temperature. Then, iodoacetic acid (5 µl, 39 mM in 0.1 M NaOH)

was added to the mixture and reaction was allowed to proceed for further 30 min in the dark at

room temperature. Reaction was quenched by excess β-mercaptoethanol (100 µl) and the mixture

was then dialysed for 48 hours against decreasing concentrations of urea (5 M, 1 × 1 l → 2 M, 1 ×

1 l → 1 M, 1 × 1 l → water, 3 × 1 l).

The dialysed solution was analysed by AAA to determine the degree of cysteine carboxymethylation

(sample and two carboxymethylcysteine standards were submitted to the same AAA protocol) and

by SDS-PAGE to check for protein integrity.

SDS-PAGE analysis12 of the protein JX249A before and after cysteine carboxymethylation was

performed on a BIO RAD Mini-PROTEAN II electrophoretic cell. An 8% acrylamide gel (7 × 8 cm)

was prepared by mixing a 40% acrylamide solution (2 ml of an acrylamide/bis-acrylamide mixture,

37.5:1) with H2O (5.5 ml), “ lower” buffer (2.5 ml, 1.5 M Tris-HCl and 0.4% SDS, pH < 8.7), 15 %

APS (40 µl) and TEMED (5 µl); the mixture was poured in the aligned clamp assembly, covered

with a water layer and left to polimerise at room temperature for 40 min. The upper gel layer for

sample loading was prepared by mixing the 40% acrylamide solution (150 µl) with H2O (1.3 ml),

Materials & Methods

171

“upper” buffer (500 µl, 0.5 M Tris-HCl and 0.4% SDS, pH < 8.7), 15% APS (20 µl) and TEMED (2

µl); this mixture was poured over the lower solidified gel and left to polimerise at room temperature

for 50 min (a teflon comb was used to mould the sample loading wells). Protein samples (JX249A

and carboxymethylated JX249A, 200 µg/ml) were diluted in “sample” buffer (1:1 v/v dilution in

20% glycerol, 4% SDS, 0.125 M Tris-HCl, 0.04% bromophenol blue, pH 6.8) and, after adding β-

mercaptoethanol (2 µl), were heated at 110 oC for 2 min. A mixture of protein molecular weight

standards including carbonic anhydrase (28 kDa), ovalbumin (45 kDa), bovine albumin (66 kDa),

phosphorylase B (97 kDa), β-galactosidase (116 kDa) and myosin (205 kDa) was prepared by

similar methods. The gel assembly was introduced in the inner cooling core and completely covered

with “ running” buffer (500 ml, 1.92 M glycine, 0.25 M Tris-HCl, 1 % SDS, pH<8.7). Both samples

and standards were loaded (20 µl) in the corresponding wells. The gel was then run at a constant

voltage of 150 V for approximately 1 hour. After cutting off the upper layer, the gel was submerged

into a Coomassie blue staining bath (0.1% Coomassie blue R-250 in fixative medium: 40%

MeOH/10% AcOH) and left under mechanic shaking for 30 min. Destaining of background colour

was done with several changes of 40% MeOH/10% AcOH (3 changes, overnight). Colour-

developed gel was dried under vacuum and heat (2 h) on a BIO RAD 583 gel dryer, using a slowly

increasing temperature gradient, followed by constant heating at 80 oC and a final fast cooling step.

The carboxymethylated JX249A fraction was analysed by SPR under conditions identical to those

described for the original protein. Similar results were obtained.

4.3.1.13 Indirect SPR affinity analysis by solution competition experiments

A solution competition SPR approach was employed for the determination of peptide – antibody

affinities13 (section 0.2). In this approach, similar to a competition ELISA experiment, a known

constant Fab concentration is incubated with known increasing competitor antigen (peptide)

concentrations. When equilibrium is reached, the peptide – antibody mixtures are put in contact

with a surface covered with specific antigen (for instance, the C-S8c1 GH loop peptide A15) and the

relationship between free Fab in competitor concentration provides a measure for competitor

peptide – antibody affinity.

Fab fragments of both SD6 and 4C4 monoclonal antibodies were kindly supplied by Ms. Wendy F.

Ochoa and Dr. Nuria Verdaguer (IIQAB – CSIC, Barcelona). Isolation and purification of Fab 3E5

were performed at 4 oC (except where mentioned otherwise) as follows: mAb was purified from

ascitic fluid as described in section 4.3.1.5 and then concentrated by precipitation with 45%

ammonium sulphate. The suspension was centrifuged (10000 r.p.m.) for 20 min and pellet was

resuspended in the minimum volume of PBS buffer. This suspension was then dialysed against PBS

overnight (3 × 1 l). To a Falcon centrifuge tube containing the antibody solution (3 mg in 2 ml of

PBS) were added the following reagents: 24 µl of 0.1 M EDTA, 126 µl of 100 mM cysteine and 30

µg of papain. The volume was completed to 3 ml with PBS buffer and the tube was sealed and left


172

at 37 oC for 5 hours. The digestion was then quenched by addition of iodoacetamide (180 µl). The

digestion mixture was analysed by SDS-PAGE on a 12% acrylamide gel as described in section

4.3.1.12, except for pre-treatment of samples, which were not submitted to heating neither to β-

mercaptoethanol addition prior to loading in the gel. mAb and Fab 4C4 samples were used as

standards. Proteins in the digestion mixture were precipitated with 85% ammonium sulphate and

the suspension was centrifuged (10000 r.p.m.) for 20 min. Pellet was resuspended in the minimum

volume of 1:1 PBS/buffer A and the suspension dialysed overnight against buffer A (3 × 1 l). After

centrifuging at 12000 r.p.m. to remove remaining solid particles, the protein solution was eluted in a

protein A – Sepharose column as previously described for mAb purification. Fractions with

OD280≥0.5 (first elution peak) were pooled and concentrated to a final volume of 2 ml, using a

Centriprep-3 concentrator* at 2000 r. p. m.

Fab was purified by gel filtration on a Sephadex G-100 support previously conditioned and

equilibrated (overnight) at a constant PBS flow of 20 ml/h. Sample elution was performed at the

same buffer flow and monitored at 280 nm. Three peaks were collected and their composition was

analysed by SDS-PAGE as previously described in this section. Fab-containing fractions were

pooled, concentrated with a Centriprep-3 concentrator and quantitated by optical density at 280

nm.

A biospecific surface was prepared upon peptide A15 immobilisation on a CM5 sensor chip as

previously described (section 4.3.1.7), injecting 50 µl of the peptide solution (200 µg/ml in 10 mM

acetate buffer, pH 5.5) in order to obtain high surface peptide density (ca. 300 RU) and therefore

favour mass transport limitations (see chapter 3).

Fab SD6, 4C4 and 3E5 stock solutions were diluted in HBS to a final concentration of 320 nM (and

subsequent serial dilutions). Series of 7 different Fab concentrations (ranging from 0 to 320 nM)

were injected over the sensor chip surface with immobilised A15: 5-min injections at 5 µl/min were

applied, and 1-min pulses of 100 mM HCl were used for regeneration. Under mass transport

limitations, initial binding rate is related to analyte concentration14, therefore a calibration curve for

initial rate = ƒ (Fab concentration) could be built from the dependence of curve initial slope

(measured at the 100th second of injection time with a 10-second time window) on Fab

concentration (see chapter 3).

* this system is used for concentration and desalting of 5 – 15 ml samples, having a 3 kDa molecular weight cut-off. Sample isplaced in a container where a filtrate collector is immersed and twist-locked. Immersion creates a slight hydrostatic pressuredifferential which is increased upon centrifugation of the assembly. Therefore, solvent and materials below the molecularweight cut-off are forced through the membrane into the filtrate collector until equilibrium is reached (hydrostatic pressuredifferential = 0). Upon removal of filtrate solution, the differential is restablished and successive concentration cycles can becarried out.

Materials & Methods

173

Peptide solutions (5 to 1250 nM in HBS) were incubated with Fab (80 nM) overnight at 4 oC. The

solutions were then allowed to stand at room temperature for 1 hour for re-equilibration prior to

injection in the SPR system. Each peptide – Fab mixture was then injected over the sensor chip

surface (5-min injections at 5 µl/min) and 100 mM HCl 1-min pulses were used to regenerate the

peptide surface after each injection. Remaining free Fab in each injected mixture was measured

from curve initial slope and subsequent intrapolation in the corresponding calibration curve. The

dependence of remaining free Fab on competitor peptide concentration was plotted and processed

by the following two methods:

Data points from the titration series where free Fab concentration was measured from the binding

rate were fitted to the equation (Fab total concentration is constant, peptide concentration is the

independent variable and KD is the fitted parameter):

[ ] [ ] [ ] [ ]( ) [ ] [ ]FabpeptideKFabpeptideKpeptideFab DD ×−

+++

−−42

2

that is included in the BIAEvaluation solution affinity model2 (chapter 3). This evaluation of KD

(KD=1/KA) does not take into account the effects of the immobilised peptide antigen.

Another method, that takes into account the influence of the immobilised peptide, is based on the

Cheng and Prusoff’ s formula15 (chapter 3):

[ ]50

1IC

FabKK A

i +=

where KA is the immobilised peptide – Fab affinity (determined independently, for instance, by SPR

kinetic analysis), [Fab] is the Fab total concentration and IC50 is the 50% inhibitory concentration for

the competitor peptide in solution (determined from the free Fab = ƒ (peptide concentration) plot).


174

4.3.2 Enzyme-linked immunosorbent assays – ELISA

The antigenicity of the FMDV synthetic peptides towards mAbs SD6, 4C4 and 3E5 was also

determined by immuno-enzymatic assays16, namely, competition ELISA17. Procedures were as

follows:

Peptide A21 conjugated to KLH# (5 pmol of peptide in 100 µl PBS per well) was incubated

overnight at 4 oC as coating antigen in micro-titer ELISA 96-well plates. The latter were saturated for

3 h with 5% BSA in PBS and then liquid was removed upon suction under reduced pressure with a

Pasteur pipette. This and all subsequent steps were carried out at room temperature. Then, 100 µl of

a solution containing a non-saturating, constant amount of mAb – pre-incubated for 1.5 h with

different concentrations of the competitor peptide antigens (serial dilutions from 243 to 0.1

pmol/100 µl) in 1% BSA in PBS – was added to the wells and further incubated for 1 h. After

washing with 0.1% BSA, 0.1% Tween 20 in PBS, 100 µl of peroxidase-conjugated goat anti-mouse

IgG (1:3000 dilution in PBS) were added to each well. Incubation was for 1 h, followed by thorough

rinsing with 0.1% BSA, 0.1% Tween 20 in PBS. Bound antibody was detected using H2O2 and orto-

phenylenediamine as substrate. Colour was allowed to develop in the dark for 10 minutes and

reaction was quenched upon addition of 2 M H2SO4 (100 µl/well). The absorbance at 492 nm was

immediately read.

The assay included a series of positive and negative controls: a positive control A21-KLH + mAb

(without competitor peptide) in triplicate and five negative controls, respectively, A21-KLH + PBS

(× 2), PBS + mAb (× 2) and PBS + PBS (× 1). Absorbances were corrected upon subtraction of

the average absorbance measured for negative controls and expressed as percentages of the

maximum absorbance (average of positive controls). Competitor peptide antigenicity was expressed

as IC50, that is, 50% of inhibitory concentration (competitor concentration leading to a 50%

decrease in maximum absorbance) and normalised to the IC50 obtained for the standard peptide

A15 (IC50 rel = IC50 competitor/IC50 A15, sections 2 and 3).

# this conjugate had been already prepared by Dr. M. L. Valero and M. E. Méndez.

Materials & Methods

175

References

1 Johnsson, B., Löfås, S. and Lindquist, G. (1991) Immobilization of proteins to acarboxymethyldextran-modified gold surface for biospecific interaction analysis in surface plasmonresonance sensors, Anal. Biochem. 198, 268-277.

2 “BIAevaluation Software Handbook: version 3.0”, (Biosensor AB, 1997) Uppsala, Sweden.3 Morton, T. A., Myszka, D. and Chaiken, I. (1995) Interpreting complex binding kinetics from optical

biosensors: a comparison of analysis by linearization, the integrated rate equation and numericalintegration, Anal. Biochem. 227, 176-185.

4 O’ Shannessy, D. J., Brigham-Burke, M., Soneson, K. K., Hensley, P. and Brooks, I. (1993)Determination of rate and equilibrium binding constants for macromolecular interactions using surfaceplasmon resonance: use of nonlinear least squares analysis methods, Anal. Biochem. 212, 457-468.



7 Albericio, F., Andreu, D., Giralt, E., Navalpotro, C., Pedroso, E., Ponsati, B. and Ruiz-Gayo, M.(1989) Use of the Npys thiol protection in solid phase peptide synthesis, Int. J. Peptide Protein Res.34, 124-128.

8 http://www.rscb.org/pdb/ (structure code 1ca1)9 Jacobson, R. H., Zhang, X. J., DuBose, R. F. and Matthews, B. W. (1994) Three-dimensional

structure of β-galactosidase from E. coli, Nature 369, 761-766.10 Carbonell, X., Feliu, J. X., Benito, A. and Villaverde, A. (1998) Display-induced antigenic variation in

recombinant peptides, Biochem. Biophys. Res. Comm. 248, 773-777.11 Lapko, V. N., Jiang, X. Y. and Smith, D. L. (1998) Surface topography of phytocrome A deduced

from specific chemical modification with iodoacetamide, Biochemistry 37, 12526-12535.12 Hames, B. D. and Rickwood, D. “Gel electrophoresis of proteins – a practical approach”, 2nd ed., IRL

Press, Oxford (1990).13 Nieba, L., Krebber, A. and Plükthun, A. (1996) Competition BIAcore for measuring true affinities:

large differences from values determined from binding kinetics, Anal. Biochem. 234, 155-165.14 “BIAapplications Handbook”, (Pharmacia Biosensor AB, 1994) Uppsala, Sweden.15 Lazareno, S. and Birdsall, N. J. (1993) Estimation of competitive antagonist affinity from functional


16 Abbas, A. K., Lichtman, A. H. and Pober, J. S. “Cellular and molecular immunology”, 3rd ed., W. B.Saunders Co., United States of America (1997).

17 Mateu, M. G., Andreu, D. and Domingo, E. (1995) Antibodies raised in a natural host andmonoclonal antibodies recognize similar antigenic features of foot-and-mouth disease virus, Virology210, 120-127.


176

4.4 Structural studies of the FMDV peptides

4.4.1 Two-dimensional proton nuclear magnetic resonance1

NMR spectra were acquired both in aqueous solution (85% H2O + 15% D2O) and in the presence

of the structuring agent TFE (30% TFE + 60% H2O + 10% D2O) at a peptide concentration of 2

mM. Experiments were run at 25 oC and 1,4-dioxane was added to all samples as an internal

reference standard.

All experiments were carried out with a Varian VXR-500 S NMR spectrometer and further processed

with the VNMR3 software programs. The 2D 1H-NMR experiments performed were:

! TOCSY3, with 70 millisecond mixing time;

! NOESY4, with mixing time of either 200 or 400 milliseconds;

! ROESY5, with 200 millisecond mixing time.

Water signal elimination was carried out either upon pre-saturation or using the WATERGATE6

method. Prior to the Fourier transform, both FIDs and interferograms were multiplied by an

exponential function.

Materials & Methods

177

4.4.2 Protein X-ray diffraction crystallography7,8

4.4.2.1 Protein crystallisation

Crystals of the complex between the Fab of 4C4 and peptide A15(138F,140P,142S) were obtained

by the hanging drop vapour diffusion9 technique and subsequent micro and macro-seeding steps.

Fab (40 µl, 18 mg/ml in PBS) and peptide (8 µl, 10 mg/ml in H2O) were incubated at 4 oC for 2

hours. A simple search for crystallisation conditions was performed in the vicinity of the conditions

found by W. F. Ochoa for the crystallisation of similar Fab 4C4 – FMDV peptide complexes at 20oC: 1 µl droplets of the peptide-Fab mixture were mixed with equivalent volumes of the precipitating

agents; these agents were based on different dilutions of PEG 4K in water, 100 mM Tris-HCl buffer

at variable pH and 400 mM LiCl. Each precipitating solution (1 ml) was poured on a well of 24-well

cell culture plates, which acted as solution reservoirs. Protein droplets were put on pre-treated& glass

covers that were then inverted and stuck, using silicone grease, to the top of the corresponding

solution reservoir.

Small twined needles were formed at 18% polyethyleneglycol (PEG) 4K, pH 8.5 and then used for

micro-seeding: a cat whisker was soaked in a needle-containing drop and then in a fresh protein

drop that was equilibrated against a solution reservoir as previously described. This micro-seeding

produced larger needles at 16% PEG 4K, which were harvested (upon suction with a capillary

quartz tube, ∅ =0.2 mm) and washed in crystallising solution. These needles were used for macro-

seeding in 2 µl droplets containing 7 mg/ml of Fab, 1.8 mg/ml of peptide, 6.5% PEG 4K, 0.2 M LiCl

with 50 mM Tris HCl (pH=8.5), equilibrated against a reservoir containing 13% PEG 4K equally

buffered at room temperature. Small needle-shaped crystals (0.6 × 0.05 × 0.03 mm) were

reproducibly formed under these conditions and, occasionally, unstable hexagonal crystals were also

observed.

4.4.2.2 Data collection

Crystals for cryogenic data collection were soaked in harvesting solutions with 20% of glycerol and

flash-frozen under a stream of boiled-off nitrogen at 100 K. X-ray data sets were collected by W. F.

Ochoa on a MarResearch image plate detector (180 × 0.10 mm, 1800 pixels) system using a Rigaku

RU-200B rotating anode, on the European Synchrotron Radiation Facility at Grenoble. A 2.2 Å

resolution data set was collected with 1o rotations (a total of 91 rotations) at a crystal-detector

distance of 180 mm. Crystals were orthorhombic, space group P212121, and unit cell parameters as

presented in chapter 3, containing one molecule of the complex per asymmetric unit. Diffraction

data were auto-indexed and integrated and merged using programs DENZO and SCALEPACK10.

& glass covers were treated as follows: 30 min in a dichloromethylsilane bath (hood) → 30 min in a water bath → 30 min in afresh water bath → 30 min in an ethanol bath; the covers were then allowed to dry prior to their utilisation.


178

4.4.2.3 Structure solution and refinement

Crystals of the complex seemed related to crystals formed with the same Fab and the wild-type

peptide A15, whose structure had been previously solved. However, the unit cell parameters

differed and the structure was newly determined by molecular replacement11 using the AmoRe

package12, employing the 4C4 Fab coordinates as searching model. The initial solutions were then

optimised by allowing to move as four separated rigid bodies the variable heavy, variable light,

constant heavy and constant light domains. Examination of the electron density maps, calculated at

this stage, clearly showed extra densities corresponding to peptide occupying the antigen binding

site. The final model for the structure of the complex was obtained by iterative cycles of manual

modelling of water molecules and rebuilding of protein/peptide chains using the program O13,

alternating with positional refinement using standard protocols in the CNS package14. Bulk solvent

correction was applied, allowing the use of all reflections in the resolution shell 15.0 – 2.3 Å. The

refined models converged to satisfactory crystallographic agreement factors, as presented in chapter

3. Structural refinement analysis was done with PROCHECK15 and graphic representation of the

structure was processed with program SETOR16.

References

1 Wüthrich, K. “NMR of proteins and nucleic acids”, Wiley, New York (1986).2 Braunschweiler, L. and Ernst, R. R. (1983), J. Magn. Reson. 53, 521.3 http://www.varianinc.com/nmr/4 Kumar, A., Ernst, R. R. and Wüthrich, K. (1980), Biochem. Biophys. Chem. Comm. 95, 1.5 Bothner-By, A. A., Stephens, R. L., Lee, J., Warren, C. D. and Jeanloz, R. W. (1984) Structure


6 Piotto, M., Saudek, V. and Sklenár, V. (1992), J. Biomol. NMR 2, 661-665.7 Drenth, J., “Principles of protein X-ray crystallography”, Cantor, C. R. (Ed.), Springer-Verlag, New

York (1987).8 Ducruix, A. and Giegé, R., “Crystallization of nucleic acids and proteins – a practical approach.”,

Oxford University Press, Oxford (1992).9 McPherson, A., “Preparation and analysis of protein crystals”, Wiley, New York (1982).

10 Otwinowsky, Z. and Minor, W. (1997) Processing of X-ray diffraction data collected in oscillationmode. Methods Enzymol. 276, 307-326.

11 Rossman, M. G. (Ed.) “The molecular replacement method”, Gordon & Breach, New York (1972).12 Navaza, J. (1994) AmoRe: an automated package for molecular replacement, Acta Crystallogr. A 50,

157-163.13 Jones, T. A., Zou, J. Y., Cowan, S. W. and Kjeldgaard, M. (1991) Improved methods for building

protein models in electron density maps and the location of errors in these models, Acta Crystallogr. A47, 110-119.

14 Brünger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang,J.-S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T. and Warren, G.L. (1998) Acta Crystallogr. D 54, 905-921.

15 Laskowski, R. A., MacArthur, M. W., Smith, D. K., Jones, D. T., Hutchinson, E. G., Morris, A. L.,Naylor, B., Moss, D. and Thornton, J. M. “PROCHECK Manual, version 3.0”, Oxford Molecular Ltd.,Oxford, UK (1994).

16 Evans, S. V. (1993) SETOR: hardware-lighted three-dimensional solid model representations ofmacromolecules, J. Molec. Graphics 11, 134-138.

Surface plasmon resonance as a tool in the functional analysis of an immunodominant site in foot

Documents

Transcript of Surface plasmon resonance as a tool in the functional analysis of an immunodominant site in foot