Assessing the Generation and Bioactivity of neo-...

AS

SE

SS

I NG

TH

E G

EN

ER

AT

I ON

AN

D B

I OA

CT

I VI T

Y O

F N

EO

-FO

RM

ED

CO

MP

OU

ND

S I

N T

HE

RM

AL

LY

TR

EA

TE

D F

OO

DS

Assessing theGeneration andBioactivity of neo-formed Compoundsin Thermally treatedFoods

Cristina Delgado-AndradeJosé Ángel Rufián-Henares(Editors)

Editorial Atrio

ISBN: 978-84-96101-76-0

a t r i oe d i t o r i a l

port. cost 2009 12/3/09, 13:551

ASSESSING THE GENERATION AND BIOACTIVITYOF NEO-FORMED COMPOUNDS

IN THERMALLY TREATED FOODS

Editors:CRISTINA DELGADO-ANDRADE

JOSÉ ÁNGEL RUFIÁN-HENARES

GRANADA , 2009

ASSESSING THE GENERATIONAND BIOACTIVITY OF NEO-FORMED

COMPOUNDS IN THERMALLYTREATED FOODS

© Los Autores

© Editorial Atrio, S.L.

EDITORIAL ATRIO, S.L.

C./ Dr. Martín Lagos, núm. 2 - 1.º C

18005 Granada

Tlf./Fax: 958 26 42 54e-Mail:atrioeditorial@ telefonica.net

ISBN: 978-84-96101-76-0

Depósito Legal: Gr.-000/2009

Summary

1. Non-enzymatic browning: The case of the Maillard reactionJOSÉ ÁNGEL RUFIÁN-HENARES; CRISTINA DELGADO ANDRADE; FRANCIS-CO J. MORALES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2. Colour and fluorescence measurement as unspecific markers for the MaillardreactionCRISTINA DELGADO-ANDRADE; JOSÉ ÁNGEL RUFIÁN-HENARES; FRANCIS-CO J. MORALES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

3. Gastrointestinal digestion as first conditioning of nutrient bioavailabilityCRISTINA DELGADO-ANDRADE; ANA HARO; ROSA CASTELLANO; JOSÉ ÁN-GEL RUFIÁN-HENARES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

4. Antimicrobial activity of Maillard reaction productsJOSÉ ÁNGEL RUFIÁN-HENARES; CRISTINA DELGADO-ANDRADE . . . . . . . . 63

Non-enzymatic browning:The case of the Maillard reaction

JOSÉ ÁNGEL RUFIÁN-HENARESDepartment of Nutrition and Food Science, Faculty of Pharmacy,

University of Granada, Spain.

CRISTINA DELGADO-ANDRADEUnit of Animal Nutriton, Estación Experimental del Zaidín,

CSIC, Granada, Spain

FRANCISCO J. MORALESInstituto del Frío, CSIC, Madrid, Spain

Non-enzymatic browning (NEB) is a set of complex reactions produced inthermally treated foods giving rise to the formation of brown colours (Cheftel andCheftel, 1980). NEB produces undesirable effects during the processing and storageof different liquid foods such as milk or fruit juices whereas for other solid foods thechanges are favourable (in the case of bread, breakfast cereals, candies, coffee, cho-colate, etc.).

NOB can be divided in three different reactions called ascorbic acid degradation,caramelisation (degradation of sugars) and the Maillard reaction (sugar-amino acidreaction). The conditions where such reactions take place are reviewed in the nextdiagram.

1. ASCORBIC ACID DEGRADATION

L-ascorbic acid, or vitamin C, is a highly water-soluble and strongly reducingsubstance with acidic properties. This chemical behaviour is related to its enodiolstructure conjugated with a carbonyl group (a lactone) which makes this molecule

Mechanism Oxygen Amino groups Optimum ph Heat Aw

Ascorbic acid degradation

yes/no no slightly acid mild medium/high

Caramelisation no no basic/acid strong low

Maillard reaction no yes basic mild low/medium

José Ángel Rufián-Henares; Cristina Delgado-Andrade; Francisco J. Morales10

very sensitive to different ways of degradation (Finholt y col., 1965). The degradationof ascorbic acid is accomplished in the absence of amino groups, slightly acidic pH,water activity medium/high and moderate temperature. There are two differentpathways, one oxidative and another non-oxidative, although one of the maindifferential characteristics among them is the higher production of furfural in the non-oxidative pathway. The selection of the oxidative or the non-oxidative pathwaysdepends on the presence of metallic catalysts. In an acidic medium, the non-oxidativedegradation of ascorbic acid starts with the hydrolysis of the lactone group, followedby decarboxilation and dehydration to diketogulonic acid and then, 3-deoxypentosoneand furfural. During the oxidative degradation, the ascorbic acid molecule istransformed to dehydroascorbic acid, which is the main precursor of the degradationproducts. Then, the dehydroascorbic acid is converted to diketogulonic acid, whichis the same molecule obtained in the non-oxidative pathway. The scheme of thereaction is shown in figure 1.

H2A: reduced ascorbic acid HA: ascorbic acid monoanion. A: dehidroascorbic acid. A+: ascorbateanion. DKG: diketogulonic acid. Mn+: metalic catalyst. HO.

2: hydroperoxyl radical. DP: 3-deoxypentose.X: xylose. F: furfural. FA: furancarboxílic acid.

Figure 1Ascorbic acid degradation pathway

Líneas con trazo más oscuro: principales formas con actividad vitamínica. H2A: ácido

córbico reducido. HA: monoanión del ácido ascórbico. A: ácido dehidroascórbico. A+: radical

ión ascorbato. DKG: ácido dicetogulónico. Mn+: catalizador metálico. HO.

2: radical

droperoxilo. DP: 3-desoxipentosa. X: xilosa. F: furfural. FA: ácido furancarboxílico.

b) Degradación enzimática: Es producida gracias a la acción de la enzima ascórbico

idasa y se realiza mediante la transformación del ácido ascórbico en ácido dehidroascórbico,

sde donde se producen los correspondientes pigmentos. La enzima necesita cobre como

factor y es producida principalmente en productos cítricos y sus derivados. Los tratamientos

rmicos inhiben la enzima y la falta de oxigeno hace que la acción se produzca a mucha menos

locidad, al ser, también, oxígeno-dependiente.

ketonisation

Non-oxydative

pathway

s roft educing

agents

O2

slow

Oxydative

pathway

(not catalysed)

Oxydative

pathway

(catalysed)

Non-o

xydat

ive

pathw

ay

± amino acids

Melanoidins

reductones

ketonisation

Non-oxydative

pathway

s roft educing

agents

O2 Oxydative

pathway

(catalysed)

Oxydative

pathway

(not catalysed)

slow

Non-o

xydat

ive

pathw

ay

± amino acids

reductones

Melanoidins

Non-enzymatic browning: The case of the Maillard reaction 11

1.2. Caramelisation

Caramelisation is another kind of NEB obtained when sugars are heated over theirfusion temperature, giving rise to an enol intermediate and final dehydration products(Krow, 1994). This pathway and other relevant reactions can be seen in figure 2. Ifthe reactive sugar is a disaccharide, such as sucrose, first a previous hydrolysis stepmust be performed in order to release two monosaccharides. Pentoses give rise to theformation of furfural as the main degradation product whereas hexoses produce 5-hydroxymethylfurfural (HMF). As an example the formation of HMF can bee seen infigure 3.

Figure 2Selected sugar degradation reactions


1.3. Maillard reaction

1.3.1. Chemistry of the Maillard reaction

The Maillard reaction (MR) is a set of chain chemical reactions that give rise tothe formation of brown pigments with modifications in colour, odor and taste ofdifferent thermally treated foods (Hodge, 1958). It usually is produced at low-intermediate water activity and basic pH. The general scheme of the reaction can beseen in figure 4.

A) The first step of the MR consists on the condensation of a carbonyl groupwith an amino one and, after dehydration an unstable Schiff base is formed, which istransformed rapidly in a N-substituted-glycosylamine. This reaction is reversiblebecause of in a strong acidic medium the sugar and the aminoacid can be re-generated.

Figure 31,2 and 2,3-enolisation of D-glucose and formation of HMF, HDF and HAF via

3- and 4-deoxyhexosulose

Figure 4. Maillard reaction

C = O + N - R C C = N - R + H2O(H )H

H

+OH

NH - R

aldosa,cetosa

am inoácidoproteina

glicosilam ina om ás generalm entecarbonilam ina

base de Schiff,inestable

Aldose orketose

Amino group

Glycosylamine Schiff base

C = O + N - R C C = N - R + H2O(H )H

H

+OH

NH - R

aldosa,cetosa

am inoácidoproteina

glicosilam ina om ás generalm entecarbonilam ina

base de Schiff,inestable

Aldose orketose

Amino group

Glycosylamine Schiff base


The amino group can be a free aminoacid, the side chain of an aminoacid (likelysine) incorporated in a protein or the amino group of the last aminoacid in eachprotein. In the case of the carbonyl groups, they are usually reducing sugars, althoughthey can be also carbonyl compounds from the intermediate stages of the MR andlipid oxidation.

+NH2

+NH2

+NH2

+NH2

+2H

Aldose+

RNH2

N-substituted glycosylamine -H2O

1-amino-1-deoxi-2-ketose

Amadori rearrangement

HMF/Furfural ReductonesFission products

(carboniles, dicarboniles)

-2H2O-2H2O

Dehydroreductone

-2H

aldehydes

Aldols

Streckerdegradation

+NH2

-CO2

+NH2

MELANOIDINS(Brown nitrogenous polymers and co-polymers)

Figure 4Maillard reaction


B) The next step consists on the irreversible rearrangement of the N-substitutedglycosylamine: when the molecule is an N-substituted-aldosylamine, by means of theAmadori rearrangement it is formed the 1-amine-1-deoxy-2-ketose. However, whenthe starting product is an N-substituted-ketosylamine it is formed a 2-amine-2-deoxy-2-ketose by means of the Heyns rearrangement. The Amadori and Heyns productsare decomposed, depending on the pH and temperature of the medium, giving rise tothe formation of different intermediate compounds: these are the intermediate stepsof the MR.

C) At low pH a 1,2-enolysation occurred, giving rise to the formation ofdicarbonyl compounds (powerful precursors of brown compounds) and finally HMF(figure 5). Contrary, at basic pH a 2,3-enolysation takes place (figure 6), beingreductones the final compounds. Such reductones can be dehydrated to formdehydroreductones, which form polymers by reacting with amino groups at advancesstages of the MR.

D) The Amadori compounds can be splitted to different fission products(dicarbonyl compounds) such as acetal or acetaldehyde.

Figure 51,2-enolysation

H - C - N - R

H

C = O

H

(H - C - OH) n

CH2 - OH

H - C - N - R

H

C - OH

H2O

H - C - OH

CH2 - OH

(H - C - OH) n - 1

H - C = N - R

H

C - OH

H - C

CH2 - OH

(H - C - OH) n - 1

+

H2OH

H - C = OC = O

H - C

CH2 - OH

(H - C - OH) n - 1

H - C = O

C - OH

CH2

CH2 - OH

(H - C - OH) n - 1

C = O

CH2 - OH

H - C - OH

H - C = O

CH

CH

RNH3+

H2O

1-amino-1-desoxi-2-cetosa

forma 1,2 enólica (muy inestable)

regeneracióndel catalizador aminado

pH óptimo 5,5

+

ompuesto alfa- (3-desoxi hexosona 3-desoxi hexosulosa)

medio ácido

compuesto dicarboniloinsaturado (3,4-didesoxi 3-eno hexosona = hexosulosa

insaturada)

O C

O

H

HOCH2

5-Hidroximetilfulfural

1-amino-1-deoxy-2-ketose

Unstable 1,2 enol

optimum pH

5.5

3-deoxyoone

acidic medium

5-HMF

H - C - N - R

H

C = O

H

(H - C - OH) n

CH2 - OH

H - C - N - R

H

C - OH

H2O

H - C - OH

CH2 - OH

(H - C - OH) n - 1

H - C = N - R

H

C - OH

H - C

CH2 - OH

(H - C - OH) n - 1

+

H2OH

H - C = OC = O

H - C

CH2 - OH

(H - C - OH) n - 1

H - C = O

C - OH

CH2

CH2 - OH

(H - C - OH) n - 1

C = O

CH2 - OH

H - C - OH

H - C = O

CH

CH

RNH3+

H2O

1-amino-1-desoxi-2-cetosa

forma 1,2 enólica (muy inestable)

regeneracióndel catalizador aminado

pH óptimo 5,5

+

ompuesto alfa- (3-desoxi hexosona 3-desoxi hexosulosa)

medio ácido

compuesto dicarboniloinsaturado (3,4-didesoxi 3-eno hexosona = hexosulosa

insaturada)

O C

O

H

HOCH2

5-Hidroximetilfulfural


Unstable 1,2 enol

optimum pH

5.5

3-deoxyoone

acidic medium

5-HMF


E) The interaction of amino acids with dicarbonyl compounds (dehydroreductonesfission products) is known as the Strecker degradation and implies the loss of aminoacids in foods. As a result of this degradation pathway, new aldehydes with one carbonatom less (lost as CO2) are formed (figure 7).

The next steps (F and G) are known as the advanced steps of the MR, where twodifferent classes of compounds are formed: melanoidins and volatile aromaticcompounds.

F) The volatile aromatic compounds (low molecular weight) are formed directlyfrom the Amadori compounds and don’t need the mediation of free amino groups.However, a 1% of the total volatile compounds produced are formed by the reactionof 2-deoxyglucose with aminoacids.

G) Melanoidins are brown polimeric compounds produced by jeans of thecondensation of aminated porducts of the intermediate stages of the MR Duch as N-substituted pyrrols, 2-formilpyrrols N-substituted, 2-furaldehyde, etc. Melanoidinshave a wide molecular weight and different absorbance spectra with maximums inthe ultraviolet (280 nm) and visible (420 nm) ranges.

Figure 62,3-enolysation

CH2 - N - R

H

C = O

H - C - OH

CH2 - OH

H - C - OH

CH2 - N - R

H

C = OH

H - C - OH

CH2 - OH

H - C - OH

C - OH

CH3

C = O

H - C - OH

CH2 - OH

H - C - OH

C = O

I

1-amino-2-

desoxi-2-cetosa

II

forma 2,3 enólica III

compuesto metil-alfa

dicarbonilo

(1-desoxi 2,3-diulosa)

(inestable)

CH3

C - OH

CH2 - OH

H - C - OH

C = O

C - OH

CH3

H - C - OH

CH2 - OH

C - OH

C = O

C - OH

IV

reductona V

enolización 3,4

enolización 2,3 catálisis básica

R - NH2

H - C

H - C

HO - C

H3C - C

C - OH

C - C

O

O

CH3

C - O

CH

OCH3

isomaltol furanona

escisiones hidrolíticas

aldehidos y cetonas con olor

H - C - OH

2,3-enolysation basic pH

1-amino-2-

deoxi-2-ketose

2,3-enol

Unstable

compound

hydrolytic rupture

3,4-enolysation

reductone furanoneisomaltol

CH2 - N - R

H

C = O

H - C - OH

CH2 - OH

H - C - OH

CH2 - N - R

H

C = OH

H - C - OH

CH2 - OH

H - C - OH

C - OH

CH3

C = O

H - C - OH

CH2 - OH

H - C - OH

C = O

I

1-amino-2-

desoxi-2-cetosa

II

forma 2,3 enólica III

compuesto metil-alfa

dicarbonilo

(1-desoxi 2,3-diulosa)

(inestable)

CH3

C - OH

CH2 - OH

H - C - OH

C = O

C - OH

CH3

H - C - OH

CH2 - OH

C - OH

C = O

C - OH

IV

reductona V

enolización 3,4

enolización 2,3 catálisis básica

R - NH2

H - C

H - C

HO - C

H3C - C

C - OH

C - C

O

O

CH3

C - O

CH

OCH3

isomaltol furanona

escisiones hidrolíticas

aldehidos y cetonas con olor

H - C - OH

2,3-enolysation basic pH

1-amino-2-

deoxi-2-ketose

2,3-enol

Unstable

compound

hydrolytic rupture

3,4-enolysation

reductone furanoneisomaltol


1.3.2. Variables in the Maillard reaction

— Substrates: The main substrates involved in the MR are the carbonyl groups(mainly from reducing sugar) and the amino groups (mainly from free amino acidsand those amino acids with side chain amino groups). The first point to take intoaccount is that although the MR is an isomolecular reaction between a sugar and aminoacid, the loss of the first one is higher than of the second one, mainly due to otherchemical reactions that taking place at the same time (like caramelisation).

Low molecular weight reducing sugars are the most reactive species due to theirlow steric hindrance. The degree of browning produced by different sugars followsthe following range: 1st Pentoses: ribose > xilose > arabinose; 2nd Hexoses: galactose> glucose > fructose; 3rd Disaccharides: maltose lactose. In the case of the aminogroups, the e-amino group of lysine is the main responsible of the development ofthe MR in proteical foodstuffs. In this sense, the reactivity of whey proteins is higherthan that of caseins, although they are more reactive than soy proteins and these, morereactive than cereal ones (mainly glutenin and glyadin).

— pH: The initial pH of foods and their buffering capacity plays an importantrole in the type and intensity of the MR. At pH<3 the rate of browning is low and itincreases as the pH raise up to a maximum of 10. However, the advance of the MRsupposes a decrease of the pH due to the formation of short chain fatty acids and thedisappearance of basic amino acids.

— Water activity: The intensity of browning depends also in the mediumhydration. The maximum activity is reached at water content between 10-15%, which

Figure 7Strecker degradation

- C = O

- C = O+

HOOC

H2N - CH - R- C = O

- C = N - CH - R

C = O

HO

H2O

- C - OH

- C - N = CH - R

- C = O

- CH - OH

CO2

NH3

R - C +O

H

- C - OH

- C - NH2

nuevoscom puestos carbonilo

α - dicarbonilo α - am inoácidoα-dicarbonyl α-amino acid

New carbonylcompounds

- C = O

- C = O+

HOOC

H2N - CH - R- C = O

- C = N - CH - R

C = O

HO

H2O

- C - OH

- C - N = CH - R

- C = O

- CH - OH

CO2

NH3

R - C +O

H

- C - OH

- C - NH2

nuevoscom puestos carbonilo

α - dicarbonilo α - am inoácidoα-dicarbonyl α-amino acid

New carbonylcompounds


means a aw of 0.67-0.70. At high moisture there is a decrease in the rate of the MRdue to a dilution factor of the reactants. Contrary, if there is low water content thediffusion of reactants is inhibited.

— Temperature: The MR is produced both at room temperature (food storage)and high temperatures (sterilisation procedures). Higher temperatures give rise to adevelopment of the MR more intense that at low temperatures but, the key factor isnot only the temperature employed but also the time applied; then, the same degreeof browning is obtained if a product is heated at a high temperature for a short timeperiod that if the same product is heated at a lower temperature for a longer time. Inthis sense, in must be taken into account that the best variable that predicts thedevelopment of the MR is the thermal load: amount of calories applied to the product.It can be obtained (for the same product processed in the same heater) by multiplyingthe time of heating (minutes) by the temperature (degrees centigrades).

1.3.3. Chemical indicators of the Maillard reaction

The extent of the MR in foods can be monitored with different chemical indexes.The main objectives of such indicators are to define the nutritional status, organolepticcharacteristics or even the possible toxicity of the foodstuff after the thermal processand storage, then obtaining food products of good quality and high nutritional value.Depending on the method assayed it will be obtained information of the different stepsof the MR (figure 8):

Figure 8Indicators of the MR.

Aldose + Protein


Intermediate dicarbonyles

Melanoidins

Furosine

HMF

Fluorescent compounds

Streckerdegradation

Volatile aldehydes

Loss of availableamino groups

Chemical assays

Bioassays

Colour measurement

Pirraline

Carboxymethyllysine

Aldose + Protein



Melanoidins

Aldose + Protein



Melanoidins

FurosineFurosine

HMFHMF

Fluorescent compoundsFluorescent compounds

Streckerdegradation

Volatile aldehydesStreckerdegradation

Volatile aldehydes


Chemical assays

Bioassays


Chemical assays

Bioassays

Colour measurementColour measurement

Pirraline

Carboxymethyllysine

Pirraline

Carboxymethyllysine


The continuous development of the analytical techniques makes possible theisolation and characterisation of more specific compounds of each step of the MR.

A) Measurement of amino acids loss: Amino acids can be measured after acidichydrolysis with concentrated hydrochloric acid (6 N) at 110°C for 10-24 hours. Thishydrolysis not release the biologically available lysine, which is the most studiedamino acid. The free e-amino group can react with different specific chemical reagentssuch as fluorodinitrobenzene (Carpenter, 1960), guanidine (Mauron and Bujard, 1963),trinitrobenzene-sulphonic acid (Tomarelli et al., 1985), borohydride reduction (Hurrelland Carpenter, 1974) or fluorescent compounds (Vigo et al., 1992).

The 2,4 dinitrofluorobenzene (FDNB) is the classic reactive used to react withthe amino group. Carperter and Ellinger (Carpenter and Ellinger, 1955) were the firstresearchers who applied such reactive to measure the e-amino group of lysine bycolorimetry (435 nm) of the derivative formed (dinitrophenillysine, DNP-lysine) al-ter acidic hydrolysis. The separation of DNP-lysine from the other DNP-amino acidswith amberlite columns improves the specificity of the technique. A later improvementwas the measurement of DNP-lysine by means of HPLC.

O-phtaldyaldehyde (OPA) is a fluorogenic reactive used in the determination ofamino groups, mainly lysine (Vigo et al., 1992). The method is highly sensitive, needslow amount of simple and don’t need protein hydrolysis (Goodno y col., 1981). Theloss of fluorescence occurred after thermal treatment has been used for controllingthe industrial processing of milk (Morales et al., 1996a) infant formula (Ferrer et al.,2000) and enteral formulas (Rufián-Henares et al., 2002a, b).

B) Amadori compounds: The measurement of Amadori compounds is useful forcontrolling the thermal damage of processed foods because of such compounds arethe first molecules relatively stable formed in the Maillard reaction. In milk productsthe measured Amadori compound is εεεεε–lactulosyl-lysine (Henle et al., 1991) whereasin the case of dehydrated orange juice the Amadori compound evaluated is fructosyl-γ-aminobutiric acid (Del Castillo et al., 1998). However, the measurement of Amadoricompounds is not analytically easy because of their derivatisation is quite complicateand the derivative tend to decompose. Then, the main approach to measure the amountof Amadori compounds is to analyse derivative formed after acidic hydrolysis. Thee-fructosyl-lysine produce two new aminoacids named furosine (20%) and pyridosine(10%) and release a 50% of blocked lysine (Finot y Mauron, 1972). In the sameconditions, the e-lactulosyl-lysine produces a 32% of furosine and a 40% of lysine(Bujard y Finot, 1978). Of special importance is to take into account the amount offurosine formed could vary depending on the hydrolysis conditions (Guerra and Cor-zo, 1996).

Erbersdobler and Zucker (1966) were the first ones who observed the formationof a new compound that eluted close to arginine in a chromatogram obtained by ionicexchange from the acidic hydrolysate of over-processed milk. This compoundincreased as function of the thermal treatment applied. Heyns et al. (1968) and Finotet al. (1968) synthesised the molecule calling it furosine (ε-N-furoylmethyl-L-lysine).The first methods used for measuring furosine were the ionic exchange employed for


analyse amino acids (Möller et al., 1977; Finot et al., 1981; Erbersdobler et al., 1987).This method is expensive and produce an overestimation of the thermal damageoccurred in lysine (Henle y col., 1991). Gas chromatography (Ruttkat andErbersdobler, 1994) and capillary electrophoresis (Delgado-Andrade et al, 2005) hasbeen also used but HPLC is, at the moment, the most used analytical technique.Because of furosine is one of the most efficient parameters for evaluating the loss oflysine availability in foodstuffs, it is measured by HPLC in many foods like potatoes,rice, and carrots (Resmini and Pellegrino, 1991), jams and fruit-based infant foods(Rada-Mendoza et al, 2004), dehydrated vegetables (Rufian-Henares et al., 2008) pasta(Resmini et al, 1991), commercial baby cereals (Guerra-Hernández et al, 1999) bread(Ramírez-Jiménez et al, 2000), breakfast cereals (Delgado Andrade et al, 2007), milk(López-Fandiño et al, 1993), enteral formulas (Rufián-Henares et al, 2002a, b) or eggs(Hidalgo y col., 1995).

In addition to furosine, other furoyl-methyl derivatives such as 2-furoyl-methyl-γ-aminobutiric acid and 2-furoyl-methyl-arginine has been proposed as usefulindicators for controlling the dehydration of fruits (Sanz et al., 2001) and differentvegetables (Cardelle-Cobas et al., 2005). However, the application of these productsis difficult because of they are not commercially available at the present moment.

C) 5-Hidroxymetylfurfural (HMF): HMF is naturally formed as an intermediatein the MR and from dehydration of hexoses under mild acidic conditions (Kroh, 1994)during thermal treatment applied to foods. In the same way, furfural is another Maillardintermediate product which is also formed from the degradation of pentoses in acidsolution (Kroh, 1994). Formation of HMF is directly related to the heat load appliedto many foods. HMF is not present in fresh, untreated foods, but rapidly accumulatesduring the heat treatment and storage of carbohydrate-rich products, sometimesexceeding 1 g/kg in certain dried fruits and caramel products (Akkan et al., 2001;Ibarz et al., 2000; Rada-Mendoza et al., 2004). The content of HMF can vary largelyin the various food products (Bachmann et al., 1997). Another source of HMF in foodproducts is related to ingredients used in the formulation such as the addition ofcaramel solutions, honey or severely heated ingredients (Rufián-Henares et al., 2006).

HMF is considered a heat-induced marker for a wide range of carbohydrate-containing foods such as processed fruits (Ibarz et al., 2000; Rada-Mendoza et al.,2004), dehydrated vegetables (Rufián-Henares et al., 2008), coffee (Dauberte et al.,1990), honey (Fallico et al., 2004; Tosi et al., 2002) and milk (Morales and Jimenez-Perez, 2001). HMF is also used for monitoring the heating process applied to cerealproducts such as bread baking (Ramirez-Jimenez et al., 2000) as well as baby cereals(Fernandez–Artigas et al., 1999) and breakfast cereals (Rufián-Henares et al., 2006).HMF is also considered for alcoholic beverages such as brandies (Frischkorn et al.,1982), beer (lo Coco et al., 1995) or wine (Laszlavik et al., 1995). But, significanceof the measurement of HMF is different depending of the type of food being mainlyused as marker of either adulteration (i.e. honey) or over-processing (i.e. milk orjuices), and marker of quality of ageing (i.e vinegar). In other foods, analysis of HMFis taken together with other furanic compounds such as furfural.


Literature describes wide variety of methods which mainly could be classified ascolorimetric, spectrophotometric and chromatographic. Initially the classical methodsfor the qualitative identification of carbonyl compounds were based on colorimetricprocedures, and were later adapted for quantitative measurement of furaniccompounds, and are often the official method for the determination of HMF in certainfoods. Colorimetric methods are based on reactions to give coloured derivates thatcan be measured at the visible range. Resorcinol, diphenylamine, aniline, p-toluidineand thiobarbituric acid has widely been used for HMF determination in an acidic me-dia. The so-called Winkler method for assessing quality in honey involves the use ofthe toxic compounds p-toluidine and barbituric acid (Winkler, 1955). This reactionand the reaction with 2-thiobarbituric acid (TBA) have been widely used to quantifyHMF in honey (Wood, 1993), wine (Malik et al., 1985; Navara et al., 1986), brandy(Queseda-Granados et al., 1992; Duran Meras et al., 1995), grape syrup, and must(Malik et al., 1981). In coffee HMF is overestimated, nearly 9%, by furfural contentbecause both substances form red colours with p-toluidine/barbituric acid (Kanjahnet al. 1996).

The TBA method has been widely applied in dairies since Keeney and Bassette(1959) developed a simple spectrophotometric method for determining HMF in dairyproducts using a TBA reaction product (after hydrolysis with oxalic acid). Howeverthe main drawbacks of the method are the lack of specificity of TBA for HMF, sinceother aldehydes may take part in the reaction, and the limited stability of the colouredcomplex (Morales et al., 1996). These authors compared the chromatographic andcolorimetric methods for HMF determination and found that 71.6% of the amount ofHMF measured by the colorimetric method was due to interferences. This probablyexplains the overestimation of HMF reported in some papers (Ferrer at al., 2002; vanBoekel and Rehman, 1987). Hence, the TBA method is not a reliable measurementof HMF content, although it is still used as a quick, cost-effective measurement ofthe heat load of milk products. Summarising, both colorimetric and spectrophotometricmethod present several drawbacks, since are time consuming, interference in stronglycoloured foods, make use of toxic or anyhow hazardous chemicals such as aromaticamines, require a strict control of both reaction time and temperature since theinstability of the reaction product derivative may lead to low recoveries and widestatistical variations in the results, reaction is dependent with the age of the reagents,limited sensibility and sometimes a lack of specificity.

Nowadays chromatographic techniques are preferably used for accurate andreliable measurement of furanic compounds in several food products. These techniquescan determine HMF and furfural specifically, and the formation of a colouredderivative is not required because of the strong UV absorption of furfurals atapproximately 280 nm. For many analysis, a detection wavelength of 280 nm is chosenbecause it is located between the maxima of HMF (284 nm) and furfural (277 nm).HMF shows a band at 284 nm (18,000 molar absorptivity) and a less intense band at230 nm. Then, reversed-phase HPLC (rp-HPLC) methods has been widely used todetermine the contents of 5-HMF and furfural in many food items, such as apple juices


and concentrates (Blanco-Gomis et al., 1991), commercial brandies and caramels(Villalón-Mir et al., 1992), and coffee (Chambel et al., 1997), milk (Morales et al.,1992), infant formulas (Albala-Hurtado et al., 1997), breakfast cereals (Rufián-Henareset a., 2006), baby cereals (Fernandez-Artigas, 1999), tomato products (Porretta andSandey, 1991), jams (Rada-Mendoza et al., 2002) and enteral formulas (Rufián-Henares, 2001). As compared with other approaches. HMF is often eluted isocraticallywith mobile phases containing 5-10% acetonitrile or methanol in water, acidified water(0.1% acetic acid) or Na-acetate buffers (pH 3.6), although gradient elution is reportedfor coffee and whiskies (Kanjahn et al., 1996; Jaganahathan et al., 1999).

REFERENCES

AKKAN, A. A., ÖZDEMIR, Y., EKIZ, H. L. (2001). Derivative spectrophotometric determinationof 5(hydroxymethyl)-2-furaldehyde HMF and furfural in Locust an extract. Nahrung/Food,45(1), 43–46

ALBALA-HURTADO, S.; VECIANA-NOGUÉS, M. T.; IZQUIERDO-PULIDO, M.; VIDAL-CAROU, M.C. (1997) Determination of free and total furfural compounds in infant milk formulas byhigh-performance liquid chromatography. Journal of Agricultural and Food Chemistry, 45,2128-2133.

BACHMANN S., MEIER M. and KA NZIG A. (1997) 5-Hydroxymethyl-2-furfural (HMF) inLebensmitteln. Lebensmittelchemie 51, 49-50.

BLANCO-GOMIS, D.; GUTIÉRREZ-ALVAREZ, M. D.; SOPENA-NAREDO, L.; MANGAS-ALONSO,J. J. (1991) High-performance liquid chromatographic determination of furfural andhydroxymethylfurfural in apple juices and concentrates. Chromatographia, 32, 45-48.

BUJARD, E. and FINOT, P.A. Measure de la disponibilite et du blocage de la lysine dans leslaits industriels. Ann. Nutr. Aliment., 1978, 32,:291-305.

CARDELLE-COBAS, A.; MORENO, F.J.; CORZO, N.; OLANO, A.; VILLAMIEL, M. Assessment ofinitial stages of Maillard reaction in dehydrated onion and garlic samples. J. Agric. FoodChem. 2005, 53, 9078-9082.

CARPENTER, K. J. and ELLINGER, G. M. Protein Quality and available lysine in animal products.Poultry Sci., 1955, 34: 1451-1452.

CARPENTER, K.J. The estimation of the available lysine in animal-protein foods. Biochem. J.,1960, 77: 604-610.

CHAMBEL, P.; OLIVEIRA, M. B.; ANDRADE, P. B.; SEABRA, R. M.; FERREIRA, M. A. (1997)Development of an HPLC/diode-array detector method for simultaneous determination of5-HMF, furfural, 5-O-caffeoylquinic acid and caffeine in coffee. J. Liq. Chromatogr. Relat.Technol. 20, 2949-2957.

CHEFTEL, J.C. and CHEFTEL, H. Introducción a la bioquímica de los alimentos. Ed. Acribia,Zaragoza, 1980: 306-308.

DAUBERTE, B.; ESTIENNE. J.; GUERRE, M.; GUERRA, N. (1990) Contribution a‘ l’etude de laformation d’hydromethylfurfural dans les boissons a base da jus de fruits et dans les cafesTorrefacto. Ann. Falsif. Expert. Chim. 1990, 889, 231-253.

DEL CASTILLO, D., CORZO, N., POLO, M. C., PUEYO, E. and OLANO, A. Changes in the aminoacid composition of dehydrated orange juice during accelerated nonenzymatic browning.J. Agric. Food. Chem., 1998, 46: 277-280.


DELGADO-ANDRADE, C.; RUFIÁN-HENARES, J.A. and MORALES, F.J. Fast method to determi-ne furosine in breakfast cereals by capillary zone electrophoresis. Eur. Food Res. Tech.2005, 221: 707-711

DELGADO-ANDRADE, C.; RUFIÁN-HENARES, J.A. and MORALES, F.J. Lysine availability is diminishedin commercial fibre-enriched breakfast cereals. Food Chem. 2007, 100(2): 725-731.

DURAN-MERAS, I.; ESPINOSA-MANSILLA, A.; SALINAS, F. (1995) Simultaneous kineticspectrophotometric determination of 2-furfuraldehyde and 5-hydroxymethyl-2-fulfuraldehyde by application of a modified Winkler’s method and partial least squarescalibration. Analyst, 120, 2567-2571.

ERBERSDOBLER, H. and ZUCKER, H. Untersuchungen zum gealt and lysin und verfügbaremlysin in trockenmagermilch. Milchwis., 1966, 21: 564-568.

ERBERSDOBLER, H.F.; DEHN-MÜLLER, B.; NANGPAL, A. and REUTER, H. Determination offurosine in heated milk as a mesure of heated intensity during processing. Dairy Res., 1987,54: 147-151.

FALLICO, B., ZAPPALÀ, M., ARENA, E., VERZERA, A. (2004). Effects of conditioning on HMFcontent in unifloral honeys. Food Chemistry, 85 (2), 305–313.

FERNÁNDEZ–ARTIGAS, P., GUERRA-HERNÁNDEZ, E., GARCÍA-VILLANOVA, B. (1999). Browningin model systems and baby cereals. Journal of Agricultural and Food Chemistry, 47, 2872–2878

FERRER, E., ALEGRÍA, A., FARRÉ, R., ABELLÁN, P., ROMERO, F. (2002). High-performanceliquid chromatographic determination of furfural compounds in infant formulas, changesduring heat treatment and storage. Journal of Chromatography A, 947(1), 85–95.

FERRER, E.; ALEGRÍA, A.; FARRÉ, R.; ABELLÁN, P. and ROMERO, F. Effects of thermalprocessing and storage on available lysine and furfural compounds contents of infent for-mulas. J. Agric. Food Chem., 2000, 48:1817-1892.

FINHOLT, P.; ASLOS, L. and HIGUCHI, T. Rate studies on the anaerobic degradation of ascorbicacid III. Rate of formation of furfural. J. Pharm. Sci. 1965, 54:181-186.

FINOT, P.A. and MAURON, J. Le blocage de la lysine par la reaction de Maillard. II. Propriétéchimiques des dérivés N-(désoxy-1-D-frutosyl-l) et N-(désoxy-l-D-lactulosyl-l) de la lysine.Helv. Chim. Acta, 1972, 55:1153-1164.

FINOT, P.A.; BRICOUT, J.; VIANI, R. and MAURON, J., Identification of a new lysine derivativeobtained upon acid hydrolysis of heated milk. Experientia, 1968, 24: 1097-1099.

FINOT, P.A.; DEUTSH, R. and BUJARD, E., The extent of the Maillard reaction during theprocessing of milk. Prog. Food Nutr. Sci., 1981, 5: 345-355.

FRISCHKORN, H. E.; WANDERLY CASADO, M.; FRISCHKORN, C. G. (1982) Schnelle Bestimmungvon Furfural und Hydromethylfurfural in alkoholischen Getränken mit Hilfe derUmkehrphasen-Chromatographie. Zeitung Lebensmittlel Untersuchung und Forschung, 174,117-121

GOODNO, C.C.; SWAISGOOD, H.E. and CATIGNANI, G.L. A fluorimetric assay for available lysinein proteins. Anal. Biochem., 1981, 115: 203-211.

GUERRA-HERNÁNDEZ, E. and CORZO, N. Furosine determination in baby cereal by Ion-PairReversed Phase Liquid Chromatography. Cereal Chem., 1996,73: 729-731.

GUERRA-HERNÁNDEZ, E., CORZO, N. and GARCÍA-VILLANOVA, B. Maillard reaction by furosinedetermination during infant cereal processing. J. Cereal Sci., 1999, 29: 171-176.

GUERRA-HERNÁNDEZ, E., CORZO, N., & GARCÍA-VILLANOVA, B. (1999). Maillard reactionevaluation by furosine determination during infant cereal processing. Journal of CerealScence, 29, 171-176.


HENLE, T.; WALTER, H. and KLOSTERMEYER, H. Evaluation of the extent of the early Maillardreaction in milk products by direct measurement of the Amadori product lactuloselysine.Z. Lebensm. Unters Forsch, 1991, 193: 119-122.

HEYNS, K.; HENKESHOVEN, J. and BROSE, K.H. Degradation of fructose amino acids to N-(2-Furomethyl) amino acids. Intermediates in browning reactions. Ang. Chem. Internat., 1968,7: 628-629.

HIDALGO, A.; ROSSI, M. and POMPEI, C. Furosine as a freshnes parameter of shell egg. J.Agric.Food Chem.,1995, 43: 1673-1677.

HODGE, J.E., Dehydrated foods. Chemistry of browning reactions in model systems. J. Agric.Food Chem., 1953, 1: 928-943.

HURREL, R.F. and CARPENTER, K.J. Mechanisms of heat damage in proteins. IV. The reactivelysine content of heat-damaged material as mesured in different ways. Brit. J. Nutr., 1974,32: 589-604.

IBARZ, A., PAGÁN, J., GARZA, S. (2000). Kinetic models of nonenzymatic browning in applepuree. Journal of Science and Food Agriculture, 80(8), 1162–1168

JAGANATHAN, J., DUGAR, S. (1999) Authentication of straight whiskey by determination ofthe ratio of furfural to 5-hydroxymethylfurfural. Journal of the AOAC International, 82,4, 997-1001

KANJAHN, D., JARMS, U., MAIER, H.G. (1996) Hydroxymethylfurfural and furfural in coffeeand related beverages. Deutsche Lebensmittel-Rundschau, 92, 328-331

KEENEY, M., & BASSETTE, R. (1959). Detection of intermediate compounds in the early stagesof browning reaction in milk products. Journal of Dairy Science, 42(6), 945–960.

KROH, L. W. (1994) Caramelisation in food and beverages. Food Chemistry, 51, 373-379LASZLAVIK, M.; GAL, L.; MISIK, S.; ERDEI, L. (1995) Phenolic compounds in two Hungarian

red wines matured in Quercus robur and Quercus petrea barrels: HPLC analysis and diodearray detection. American Journal of Enology and Viticulture. 46, 67-74.

LO-COCO, F. L.; VALENTINI, C.; NOVELLI, V.; CECCON, L. (1995) Liquid chromatographicdetermination of 2-furaldehyde in beer. Analitica Chimica Acta. 306, 57-64.

LÓPEZ-FANDIÑO, R.; CORZO, N.; VILLAMIEL, M.; DELGADO, T.; OLANO, A. and RAMOS, M.Assessment of quality of comercial UHT milks by chromatographic and electrophoreticmethods. J. Food Prot., 1993, 56: 263-269.

MALIK, F.; DRAK, M.; CRHOVA, K. (1981) 5-Hydromethylfurfural in Produkten derWeinerzeugung. Wein-Wiss. 36, 360-365.

MALIK, F.; NAVARA, A.; MINARIK E. (1985) 5-Hydroxymethylfurfural-Gehalte inTraubenweinen. Mitt. Klosterneuburg, 35, 45-47.

MAURON, J. and BUJARD, E. Guanidation, an alternative approach to the determination ofavailable lysine in foods. En: Nutrition. Proc. 6th Int. Congr., Edimbourgh, 1963, Mills,C.F. y Passmore, R., (Eds.), E. y S. Livingstone Ltd., 1964: 489-498.

MÖLLER, A.B.; ANDREWS, A.T. and CHEESMAN, G.C. Chemical changes in ultra-heat-treatedmilk during storage. II. Lactuloselysine and fructoselysine formation by the Maillardreaction. J. Dairy Res., 1977a, 44: 267-275.

MORALES, F. J., JIMÉNEZ-PÉREZ, S. (2001). Hydroxymethylfurfural determination in infantmilk-based formulas by micellar electrokinetic capillary chromatography. Food Chemistry,72(4), 525–531.

MORALES, F. J.; ROMERO, C.; JIMÉNEZ-PÉREZ, S. 1995. New methodologies for kinetic studyof 5-(hydroxymethyl)furfural formation and reactive lysine blockage in heat-treated milkand model systems. Journal of Food Protection, 58, 310-315.


NAVARA, A.; MALIK, F.; MINARIK, E. (1986) 5-Hydroxymerhylfurfural in Weinen aus demostslowakischen und Tokayer Weinbaugebiet der CSSR. Mitt. Klosterneuburg, 36, 28-33.

PORRETTA, S., SANDEI, L. (1991). Determination of 5-(hydroxymethyl)-2-furfural (HMF) intomato products: proposal of a rapid HPLC method and its comparison with the colorimetricmethod. Food Chemistry, 39, 51–57.

QUESADA-GRANADOS, J.; VILLALON, M.; LOPEZ-SERRANA, H.; LOPEZ-MARTINEZ, M. C. (1992)Comparison of specrophotometric and chromatographic methods of determination of furanicaldehydes in wine distillates. Food Chemistry, 52, 203-208.

RADA-MENDOZA, M., OLANO, A., VILLAMIEL, M. (2002). Determination ofhydroxymethylfurfural in commercial jams and in fruit-based infant foods. Food Chemistry,79, 513–516

RADA-MENDOZA, M., SANZ, M.L., OLANO, A., & VILLAMIEL, M. (2004). Formation ofhydroxymethylfurfural and furosine during the storage of jams and fruit-based infant foods.Food Chemistry, 85, 605-609.

RAMÍREZ-JIMÉNEZ, A., GUERRA-HERNÁNDEZ, E., & GARCÍA-VILLANOVA, B. (2000). Browningindicators in bread. Journal of Agricultural and Food Chemistry, 48, 4176–4181.

RAMÍREZ-JIMÉNEZ, A., GUERRA-HERNÁNDEZ, E., & GARCÍA-VILLANOVA, B. (2000). Browningindicators in bread. Journal of Agricultural and Food Chemistry, 48, 4176-4181.

RAMÍREZ-JIMÉNEZ, A.; GUERRA-HERNÁNDEZ, E. and GARCÍA-VILLANOVA, B. Browningindicators in bread. J. Agric. Food Chem. 2000a, 48:4176-4181.

RESMINI, P. and PELLEGRINO, L. Analysis of food heat damage by direct HPLC of furosine.Int. Chrom. Lab., 1991, 6: 7-11.

RESMINI, P., & PELLEGRINO, L. (1991). Analysis of food heat damage by direct HPLC offurosine. Inernational Chromatography Laboratory, 6, 7-11.

RESMINI, P., PAGANI, A.M., & PELLEGRINO, L. (1991). Evaluation of the thermal damage topasta by determination of epsilon-furoylmethyl-lysine (furosine) by HPLC. TecnicaMolitoria, 41, 821-826.

RUFIAN-HENARES, J.A, DELGADO-ANDRADE C., MORALES, F.J. (2006) Application of a reversephase high-performance liquid chromatography method for simultaneous determination offuranic compounds and glucosylisomaltol in breakfast cereals. Journal of AOACInternational, 89, 1-5.

RUFIÁN-HENARES, J.A., GARCÍA-VILLANOVA, B. and GUERRA-HERNÁNDEZ, E. Occurrence offurosine and hydroxymethylfurfural as markers of thermal damage in dehydrated vegetables.Eur. Food Res. Technol. 2008, 228: 249-256.

RUFIÁN-HENARES, J.A.; GARCÍA-VILLANOVA, B. and GUERRA-HERNÁNDEZ, E. Loss of o-phthaldialdeyde reactivity, fluorescence and colour in stored enteral formula. Int. J. DairyTechnol., 2002a, 54 (4).

RUFIÁN-HENARES, J.A.; GARCÍA-VILLANOVA, B.; GUERRA-HERNÁNDEZ, E. Maillard reactionin enteral formula processing: furosine, loss of o-phthaldialdehyde reactivity andfluorescence. Food Res. Int., 2002b, 35 (6).

RUFIÁN-HENARES, J.A., GARCIA-VILLANOVA, B. and EDUARDO GUERRA-HERNANDEZ. 2001.Determination of furfural compounds in enteral formula. Journal of Liquid Chromatographyand Related Technologies, 24, 3049–3061.

RUTTKAT, A. and ERBERSDOBLER, H.F. Degradation of furosine during heptafluorobutyricanhydride-derivatization for gas-chromatographic determination. J. Chrom. A, 1994, 673:103-110.

SANZ, H.L.; CASTILLO, M.D.; CORZO, N. and OLANO, A. Formativo of Amadori compoundsin dehydrated fruits. J. Agric. Food Chem. 2001, 49(11):5228-5231.


SANZ, M.L.; DEL CASTILLO, M.D.; CORZO, N.; OLANO, A. Formation of Amadori compoundsin dehydrated fruits. J. Agric. Food Chem. 2001, 49, 5228-5231.

TOMARELLI, R.M.; YUHAS, R.J., FISHER, A. and WEABER, J.R. An HPLC method for thedetermination of reactive (available) lysine in milk and infant formulas. J. Agric. FoodChem., 1985, 33: 316-318.

TOSI, E., CIAPPINI, M., RÉ, E., LUCERO, H. (2002). Honey thermal treatment effects onhydroxymethylfurfural content. Food Chemistry, 77(1), 71–74.

VAN BOEKEL, M. A. J. S., REHMAN, Z. (1987). Determination of hydroxymethylfurfural inheated milk by high – performance liquid chromatography. Netherlands Milk and DairyJournal, 41, 297–306.

VIGO, M.S.; MALEC, L.S.; GÓMEZ, R.G. and LLOSA, R.A. Spectrophotometric assay using o-phtalaldehyde for determination of reactive lysine in dairy products. Food Chem., 1992,44: 363-365.

VILLALON MIR, M.; QUESADA GRANADOS, J.; LOPEZ; G.; DE LA SERRANA, H.; LOPEZ

MARTINEZ, M. C. (1992) High performance liquid chromatography determination of furaniccompounds in commercial brandies and caramels. Journal of Liquid Chromatography, 15,513-524.

WINKLER, O. (1955). Beitrag zum Nachwals und zur Bestimmung von Oxymethylfurfural inHonig und Kunsthonig. Zeitschrift fur Lebensmittel Untersuchung und Forshung, 102(3),161–167.

WOOD, R. M. (1993) A HMF validated method for the analysis of foodstuffs. J. Assoc. Publ.Anal. 28, 195-199.


Practical assay 1Measurement of HMF and Furfural

CHEMICALS

Carrez I (potassium ferrocyanide, 15% w/v) and Carrez II (zinc acetate 30% w/v) as precipitating reagents; HMF and furfural standards; acetonitrile for liquidchromatography. In addition, all the chemicals necessary to perform the in vitrogastrointestinal digestion as described in the respective chapter.

SAMPLES

Wheat and wheat bran breads (W and BW, respectively) commercialised as pre-baked breads to be finally baked at home. They were kept in an air oven for 12, 20 or26 minutes at 200°C, and, after aerated, they were lyophilised. One piece of eachkind of bread was also lyophilised without heat treatment as a control of the initialstage in the pre-baked breads. All of them were milled in a grinder. Different powderedsamples obtained were named as follows:

— Wheat breads: W0, W12, W20 and W26.— Wheat bran breads: WB0, WB12, WB20 and WB26.

ANALYTICAL PROCEDURES

Procedure: The present method is based on Rufián-Henares et al (2006). Briefly,0.5 g of sample are suspended in 5 ml of deionized water and the tube is shakenvigorously for 1 min and clarified with 0.25 ml each of Carrez I and Carrez IIsolutions. The resulting mixture is centrifuged, the supernatant is collected in a 10ml volumetric flask and two further extractions are performed using 2 mL of deionizedwater. The supernatants are mixed and the volume is made up to 10 ml with deionizedwater. Solutions are filtered and stored in vial after HPLC analysis.

The LC mobile phase is made up of a mixture of acetonitrile in water (5% v/v)delivered at the flow rate of 1 ml/min under isocratic conditions through the analytical


column (C18 5 μm, 25 x 0.4 cm) termostatized at 32°C. The UV detector is set at 280nm and 20 μL are injected. HMF and furfural are quantified by the external standardmethod within the range 2-100 μM and 1-20 μM respectively .

RESULTS

Data from HMF and furfural in pre-baked breads

Sample Heating time HMF Furfural0

12

Wheat bread20

26

0

12

Wheat20bran bread

26


Data from HMF and furfural measurement before and after the in vitrogastrointestinal digestion

Before in vitro digestion After in vitro digestionHeatingSample

Time HMF Furfural HMF Furfural0

12Wheat bread

20

26

0

12Wheat bran

bread 20

26


Practical assay 2Measurement of furosine

CHEMICALS

Hydrochloric acid as hydrolysis reagent; furosine standard; acetonitrile andmethanol for liquid chromatography; formic acid and sodium heptane sulphonate forthe HPLC mobile phase. In addition, all the chemicals necessary to perform the invitro gastrointestinal digestion as described in the respective chapter.

SAMPLES




Procedure: The present method is based on Delgado-Andrade et al. (2007).Briefly, 30 mg of the sample are hydrolysed with 4 ml of 7.95 M HCl at 110°C for23 h in a Pyrex screw-capped vial with PTFE-faced septa. Hydrolysis tubes must besealed under nitrogen. The hydrolysates are aerated and cooled at room temperatureand subsequently centrifuged at 14,000 × g for 10 min. A 0.5-ml portion of thesupernatant is applied to a Sep-pak C18 cartridge (Millipore) pre-wetted with 5 ml ofmethanol and 10 ml of deionized water, and was then eluted with 3 ml of 3 M HCl.The dried sample was dissolved in 1 ml of a water, acetonitrile and formic acid (95 :5 : 0.2) mixture. An analytical column (C18 25 × 0.40 cm, 5-μm) is used at 32°C,isocratically eluted at a 1.2 ml/min flow rate with degassed mobile phase prepared


with 5 mM sodium heptane sulphonate containing 20 % of acetonitrile and 0.2% offormic acid. The injection volume was 50 μL and detection was performed at 280nm. Furosine is quantified by the external standard method.

RESULTS

Data from furosine in pre-baked breads

Sample Heating time Furosine0

12

Wheat bread20

26

0

12

Wheat20bran bread

26


Data from furosine measurement before and after the in vitro gastrointestinal digestion

After in vitro

digestion

Heatingtime

Before in vitro digestion

Sample

0

12Wheat bread

20

26

0

12Wheat bran

bread 20

26

Colour and fluorescence measurementas unspecific markers for the Maillard reaction

CRISTINA DELGADO-ANDRADEUnit of Animal Nutriton, Estación Experimental del Zaidín, CSIC,

Granada, Spain

JOSÉ ÁNGEL RUFIÁN-HENARESDepartment of Nutrition and Bromatology, Faculty of Pharmacy,

University of Granada, Spain

FRANCISCO J. MORALESInstituto del Frío, CSIC, Madrid, Spain

As it is well-known, one of the main modifications induced by heating and longstorage conditions of foods is the Maillard Reaction, which involves amino acids andreducing carbohydrates and can produce a loss in nutritive value (Henle et al., 1991;Lowry & Baker, 1989).

The development of colour is an evident and extremely important indicator ofthe extent of the advanced Maillard reaction (Nursten, 1986). The colours producedrange from pale yellow to very dark brown, depending on the type of food and extentof the reaction (Morales & Van Boeckel, 1999). From an organoleptic point of view,browning is desirable in some types of food (bakery products and coffee), whereas itis undesirable in sterilized or dried milk products (powder milk, powder whey). Milkand milk-based systems darken with heating intensity (temperature and time) andstorage.

The quantitative measurement of the browning rate can be considered to indicatethe severity of heat treatment (Andrews & Morant, 1987; Rhim, Jones, & Swartzel,1988) or to assess the efficiency of technological industrial processes in milk(Giangiacommo & Messina, 1988) or any other foods, since the colour is the resultof coloured natural products associated with raw material and/or coloured compoundsgenerated as a result of processing (Rizzi, 1997). In this sense, it could be useful toexpress browning as the changes in the visual colour of the foods.

There are various techniques for determining browning, including opticalmeasurements without coupled separation techniques (Morales & Van Boeckel, 1999)or analytical determinations such as thin layer chromatography (Rizzi, 1997), RP-HPLC (Ames, Apriyantono, & Arnoldi, 1993) or IE-HPLC (Ingles & Gallimore, 1985).Some of them are detailed in this chapter.

Cristina Delgado-Andrade; José Ángel Rufián-Henares; Francisco J. Morales34

1. TRISTIMULUS COLOUR OR CIELAB SYSTEM

Optical analysis has been performed by the spectrophotometric measurement oftristimulus colour values both in liquid (Bertelli, et al., 1996; Kessler & Fink, 1986;Pagliarini, et al., 1990; Rampilli & Andreini, 1992) and solid (Fernández-Artigas, etal., 1999; Ramírez-Jiménez, et al., 2001) foods, what is known as CIELab system.

The system provides the values of three colour components; L* (black-whitecomponent, luminosity), and the chromaticness coordinates, a* (+red to –greencomponent) and b* (+yellow to –blue component) (Hunter, 1942). Samples are placedinto a glass dish (with different diameter depending on the concrete apparatus) andilluminated with D65-artificial daylight (10° standard angle) according to conditionsprovided by manufacturer.

Some interesting equations can be calculated from these parameters. The E index(E) is obtained from the following expression:

And Chroma value (C) according to the next one:

Solid colours are named according to Kelly and Judd (1976).The E index is mainly influenced by the colour lightness and describes the

proximity or farness between two samples in the colour space (Morales & van Boekel,1998).

The Chroma value is a more specific formula indicating the degree of saturation,purity or intensity of visual colour (Rhim et al., 1988).

Francis and Clydesdale (1975) also proposed the use of yellowing index (YI) as acolour measurement related to browning intensity:

All the mentioned parameters are usually applied by many authors to evaluatecolour development associated to Maillard reaction development. Thus, Pagliarini etal. (1990) used the yellow index to evaluate milk colour changes after 60min of heattreatment between 90 and 130°C; Lowry and Baker (1989) used the L* parameter toobserve colour changes during the storage of enteral formulas; Rufián-Henares et al.

E = (L*2 + a*2 + b*2)1/2

C = (a*2 + b*2)1/2

YI = 142.86b*/L*

Colour and fluorescence measurement as unspecific markers for de Maillard reaction 35

(2002) employed the colour measurement (L*, a*, b*, C and YI) for controlling themanufacture and storage of enteral formulas.

As an example of the employment of these indexes, let us analyse the resultsobtained in our research group working with an equilibrate diet designed for a healthygroup of adolescents which was processed by two different ways (Delgado-Andradeet al., 2007):

On one hand, the diet was cooked using culinary techniques to avoid as far as possiblethe development of the Maillard reaction (B-diet), and on the other, another culinarytechniques promoters of the Maillard reaction were used (A-diet). Lunch and dinner mealsfor the 7-d menu were prepared in a local catering, always under the strict control of theinvestigators (Table 1). Breakfast and afternoon snack were prepared in the laboratory asfollows: A-diet, whole milk with cocoa powder and breakfast cereals in the morning andmilk shake and sandwich (pâté, boiled ham or mortadella) or buns in the afternoon; B-diet,whole milk and butter and bread in the morning and whole milk and sandwich without crust(pâté, mortadella or boiled jam) in the afternoon. Dishes of each menu were prepared twiceand mixed.

On each diet, the edible portion (the part of the food that one can eat: in some foods,there is an inedible portion that has to be removed, eg, chicken or fish bones or fruit skin)of lunch and dinner (LD) for each day (prepared in duplicate) were removed, weighed andhomogenised with a hand blender. Aliquots of the 7 days from both diets separately werewell mixed and lyophilised; these samples were names as LD-A or LD-B, respectively for Aor B-diet. In the same manner, breakfast and afternoon snack (BA) from both diets (also induplicate) were mixed and lyophilised; samples were named BA-A and BA-B. Table 2 showsL*, a*, b*, E and C values.

The more positive values of a* in the A meals compared with the corresponding B onesindicate the progress of the colour to the red zone, in accordance with the higher presenceof MRP, although lipid oxidation could also be partly responsible of this observation. Inview of the decreased b*-value for the BA-A sample, this parameter might be also affectedby oxidation, since data suggested the predominance of blue, when a higher portion of yellowcomponent was expected in this more processed meal. The luminosity (L*), is the black-white component of the sample. Luminosity significantly decreased in the more severelycooked meals (LD-A and BA-A vs. LD-B and BA-B, respectively). Even LD-B luminositywas lower than BA-B value, showing that B lunch and dinner was darker, and so moreprocessed, than B breakfast and afternoon snack. Concerning the E index, is mainly influencedby the colour lightness and describes the proximity or farness between two samples in thecolour space [35]. Comparing E-values from LD-B vs. LD-A and BA-B vs. BA-A a decreasedin the index was manifested, confirming the loss of lightness likely due to the more severecooking treatment of meals and the formation of brown pigments in the advance of theMaillard reaction. C value indicates the degree of saturation, purity or intensity of visualcolour (Rhim et al., 1988). The decrease in the Chroma value denotes the introduction ofthe blue colour in the LD-A and BA-A meals, as it has been described above for the b*values.


DIE

T #

1st2n

d3r

d 4t

h5t

h6t

h7t

h

L U N C H

Em

pana

dill

as†

wit

h sa

lad

Leg

umes

(le

ntil

s)

Cho

cola

te y

oghu

rtB

read

Gra

tin

mac

aron

i w

ith

béch

amel

sau

ce

Ban

anas

Bre

ad

Span

ish

omel

ette

wit

h ha

mM

eatb

alls

(ve

al)

wit

h ve

geta

bles

Cho

cola

te c

usta

rd

Bre

ad

Leg

umes

(c

hick

peas

) C

hoco

late

yog

hurt

Bre

ad

Sal

ad (

lett

uce,

to

mat

oes…

)P

aell

aA

pple

Bre

ad

Frie

d ch

icke

n an

d fr

ied

pota

toes

Sal

ad (

lett

uce,

to

mat

oes…

)Pe

ars

Bre

ad

Tro

pica

l sal

ad

Spa

nish

om

elet

te

Ric

e w

ith

mil

k an

d ci

nnam

onB

read

A

D I N N E R

Con

som

mé

wit

h n

oodl

es

Bre

aded

fis

h Pe

ars

Bre

ad

Sau

té v

eget

able

s G

ridd

le lo

in o

f po

rk a

nd

frie

d po

tato

es

Tor

rija

*B

read

Piz

za

Ban

anas

Bre

ad

Soup

of

past

a an

d ch

icke

nH

ambu

rger

wit

h f

ried

pot

atoe

s B

anan

asB

read

Veg

etab

les

crea

m

with

cro

uton

s B

read

ed h

ake

fish

-fi

nger

sC

hoco

late

cus

tard

Pa

n

Fish

cro

quet

tes

Car

amel

cus

tard

B

read

Bre

aded

fis

h an

d ri

ce

Ora

nges

Bre

ad

L U N C H

Rus

sian

sal

ad w

ith

tuna

L

egum

es (

lent

ils)

St

raw

berr

y yo

ghur

tB

read

with

out c

rust

Spa

ghet

ti w

ith

tom

ato

sauc

e, c

hees

e an

d ha

m

Ban

anas

Bre

ad w

ithou

t cru

st

Boi

led

pota

toes

, boi

led

eggs

and

ham

B

aked

mea

t (ve

al)

wit

h ve

geta

bles

Cus

tard

Bre

ad w

ithou

t cru

st

Soup

of

past

a an

dch

icke

nSa

usag

es w

ith

mas

hed

pota

toes

B

anan

asB

read

with

out c

rust

Sal

ad (

lett

uce,

to

mat

oes…

)S

tew

ed r

ice

App

leB

read

with

out c

rust

Boi

led

chic

ken

and

pota

toes

S

alad

(le

ttuc

e,

tom

atoe

s…)

Pear

s B

read

wit

hout

c

rust

Tro

pica

l sal

ad

Tun

a-fi

lled

egg

s R

ice

wit

h m

ilk

Bre

ad w

ithou

t cru

st

BD I N N E R

Con

som

mé

wit

h n

oodl

es

Bak

ed f

ish

with

boi

led

pota

toes

Pear

s B

read

with

out c

rust

Veg

etab

le s

tew

B

aked

loin

of

pork

and

bo

iled

pot

atoe

s Sy

rup

peac

h B

read

with

out c

rust

Past

a w

ith

tom

atoe

s an

d ch

eese

Ban

anas

Bre

ad w

ithou

t cru

st

Leg

umes

(ch

ickp

eas)

St

raw

berr

y yo

ghur

tB

read

with

out c

rust

Veg

etab

les

crea

m

Fish

pud

ding

C

usta

rdB

read

with

out c

rust

Bak

ed f

ish

Cus

tard

Bre

ad w

ithou

t cru

st

Fish

wit

h cr

eam

and

ri

ceO

rang

esB

read

with

out c

rust

† Sm

all t

una-

fill

ed b

read

ed p

asti

es

* Fr

ied

brea

d w

ith m

ilk, s

ugar

and

cin

nam

on.

# Die

ts: A

, die

t sim

ilar

to th

ose

usua

lly c

onsu

med

by

adol

esce

nt p

opul

atio

n; B

, die

t con

tain

ing

the

sam

e fo

ods

than

in th

e A

-die

t, bu

t with

sof

ter

proc

essi

ng w

hene

ver

poss

ible

.

Tabl

e 1

Lun

ch a

nd d

inne

r 7-

d m

enu

of t

he d

iffe

rent

die

ts


2. BROWNING MEASUREMENT

The development of colour is an extremely important and obvious feature of theextent of the advanced Maillard reaction (Nursten, 1986). This stage is characterisedby the formation of unsaturated, brown nitrogenous polymers and copolymers,although nitrogen-free polymers are also formed from condensation reactions fromfurfurals or dehydroreductones (Ames, 1992; Hodge, 1953). The colours producedrange from pale yellow to very dark brown, depending on the type of food and/or theextent of the reactions. Many of the studies are refereed to low molecular weightcompounds, which can contain conjugated double bounds, responsible of yellow andorange-like colours (Ames et al., 1993). Heavier compounds use to give rise to thered-purpure colours (Hayashi y col., 1985), and their appearance is linked to a higherpH level in the food system (Reineccius, 1990)

The aldehydes formed by the Strecker degradation are also a source of browningsince they can condense with themselves, with sugar fragments, with furfurals, orwith other dehydration products to form brown pigments. However, Strecker´s pathwayis not the major colour-producing reaction and is better known as the source of off-flavours associated with Maillard browning (Namiki, 1988); moreover, the Streckerdegradation is impossible in the absence of amino acids, i.e. is not relevant withprotein.

Ames and Nursten (1989) grouped the coloured compounds into two generalclasses: low molecular weight compounds, which typically possess two-to-four linkedrings containing extended double-bond conjugation (Ledl & Sverin, 1978, 1982; Ames,1992; Ames et al., 1993), and melanoidins, which are brown polymers and possessmolecular weights of several thousand daltons and discrete chromophore groups (Ames& Nursten, 1989; O’Brien & Morrisey, 1989).

Quantitative measurement of browning rate (brown compounds or colour) maybe considered an indicator of the severity of the heat treatment (Bosset & Balnc, 1978;Mauron, 1981; Horak & Kessler, 1981; Andrews & Morant, 1987; Rhim et al., 1988)

Table 2Colorimetrics parameters of mealsA

MealsB L* a* b* EC CD

LD-A 71.71 ± 0.11 a -1.46 ±0.12 ac 23.46 ± 0.20 a 75.46 ± 0.05 a 23.51 ± 0.19 a

LD-B 75.29 ± 0.02 b -1.70 ± 0.03 a 24.55 ± 0.12 b 79.21 ± 0.05 b 24.60 ± 0.11 b

BA-A 65.17 ± 0.22 c 4.74 ± 0.09 b 19.02 ± 0.09 c 68.05 ± 0.20 c 19.60 ± 0.04 c

BA-B 77.57 ± 0.04 d -1.38 ± 0.05c 23.18 ± 0.14 a 80.97 ± 0.08 d 23.22 ± 0.14 a

A Values represent mean ± S.D. (n = 4). Different lower-case letters in the same column indicate significantdifferences (One way Anova and Duncan Test, P < 0.05) for each studied parameter.

B LD-A, one-week pool of lunch and dinner from the A-diet; LD-B, one-week pool of lunch and dinnerfrom the B-diet; BA-A, one-week pool of breakfast and afternoon snack from the A-diet; BA-B, one-week pool of breakfast and afternoon snack from the B-diet.

C E = (L*2 + a*2 + b*2)1/2.DC = (a*2 + b*2)1/2.


or for evaluating the efficiency of technological industrial processes (Giangiacomo& Messina, 1988). Methods which have been employed to determine browningreactions include both chemical analyses and optical measurements, as well as visualexamination (Maillard, 1912; Webb & Holm, 1930; Ellis, 1959; Reyes et al., 1982;Petriella et al., 1985). Traditionally, the advanced stage of the Maillard reaction withproteins has been monitored by measuring pigment production measured after trysindigestion and aqueous extraction (Choi et al., 1949). This method was modified byLabuza and Saltmarch (1981) using a solution of three proteolytic enzymes (trypsin,chymotrypsin and peptidase). Later, Palombo et al. (1984), studying the browning ofdairy powders, proposed to liberate brown pigments from protein molecules by meansof a non-specific proteolytic enzyme, pronase, followed by determination of thebrowning index by measuring the absorbance at 420 nm.

Currently, browning measurement, with or without a previous enzymatic digestionstep, at both 280 and 420 nm, is considered as a good unspecific indicator of theextent of the Maillard reaction. In the first stage of the Maillard reaction reducingsugars react with amino acids giving rise to non-colour compounds which do notabsorb in the visible spectra (400-700 nm) (Renn & Sathe, 1997). So that absorbancemeasurement between 280 and 294 nm is related to the early stages of the Maillardreaction and to low Maillard reaction compounds (Jing & Kitts, 2004). Thisdetermination have been used by different authors to confirm the initial step of thereaction and the appearance of those initial compounds in foods, such as in roastedcoffee (del Castillo et al., 2002), as well as in sugar-protein model systems (Benjakulet al., 2005). The progress of the reaction supposes the production of high molecularweight brown products, so-called melanoidins, whose chromophore groups aredifferent depending on the reactants and the reaction conditions. These compoundshave a characteristic absorbance maximum at 420 nm (Morales & Jiménez-Pérez,2004). In some cases, it has been described an initial increase in the absorbance valueat this wavelength followed by a decrease when increasing the heating time of thestudied model system or the heat processing of foods (Delgado-Andrade et al., 2004;Fogliano et al., 1999). This fact has been attributed to the existence of insolublepigments which are not included in the soluble fraction submitted to thespectrophotometric analysis.

Due to the complexity of the chemistry of Maillard reaction, most studies describedin the literature have been confined to model systems. For that reason, let us analysethe absorbance values obtained for glucose-lysine and glucose-methionine modelsystems at 420 nm:

Maillard reaction compounds were generated in both model systems by heatingequimolar mixtures of glucose-lysine (GL0) or glucose-methionine (GM0), 40% moistureand in unbuffered systems at 150°C for 30 min (GL30 and GM30), 60 min (GL60 and GM60)or 90 min (GL90 and GM90), using open recipients (Delgado-Andrade et al., 2004).

The development of brown colour at 420 nm was measured suspending 100 mg of eachsample (GL, GL30, GL60, GL90, GM, GM30, GM60 or GM90) in a final volume of 10mlof demineralized and neutralized water. Samples were then stirred for 1h in a water bath at


room temperature and centrifuged at 5000 rpm for 45min. Absorbance was measured in thesolutions or in the corresponding supernatants when samples were partially insoluble in water(Table 3).

The time-course of colour formation in the glucose-lysine model system was moreaccelerated than in the glucose-methionine one, and the browning degree reached after eachheating time was also higher in samples containing lysine. Similarly, after the heat treatmentof various amino acid-sugar combinations, Ashoor and Zent (1984) showed that the glucose-lysine system produced a higher browning rate at 420 nm.

According to Bates et al. (1998), the dissociation constant of the two amino groups oflysine favour the initial step of MR, as well as the more chromogenic routes of the reaction.In the present glucose-lysine system, the absorbance value at 420 nm was maximum afteran hour of heating time (Table 3), probably due to the high presence of soluble brownpigments. In the last 30 min. of treatment insoluble melanoidins were formed, evidenced bythe appearance of a dark pellet. Thus, the spectrophotometric measure of absorbance,performed on the soluble fraction, decreased in the GL90 sample. In this sense, Fogliano etal. (1999) in a gluten-glucose system heated to 150°C also observed an important increaseof browning until 45 min. heating, followed by a decrease at longer heating times. Monti etal. (1998), using different amino acid-sugar mixtures refluxed for up to 120 min., alsoreported that in the glucose-lysine system the highest absorbance values appeared after onehour of heating, to subsequently reach a plateau.

Water activity (aw) is an important factor for MR, which is favoured within an intervalof moisture content (Reineccius, 1990). Thus, the MR activation energies increase withdecreasing water proportion (Cremer & Eichner, 2000). In the present experiment the moisturecontent of the assayed samples drastically decreased after 60 min of heating, and thus, thelower aw could be another limiting factor in the browning rate. Fogliano et al. (1999) pointedout that the samples heated under wet conditions produce more colour at all temperaturesstudied, compared to those heated under dry conditions.

3. FLUORESCENCE ASSOCIATED TO MAILLARD REACTION DEVELOPMENT

In the last decades, the determination of fluorescence has been proposed as aneffective procedure to assess the extent of the MR, being suitable to monitor the earlystages as well as the advanced steps of the reaction. Amadori rearrangement productundergoes dehydration and fission and yield colourless reductones as well asfluorescent substances are formed (Baisier & Labuza, 1992).

Different letters in the same column mean significant differences (p < 0.05, ANOVA one-way and Duncantest). Values are mean ± SD of six determinations.

SampleGlucose-Lysine Glucose-Methionine

Heatingtime

Precipitate Absorbance Precipitate Absorbance0 - 0.001±0.000a - 0.007±0.001a30 - 0.573±0.001b - 0.011±0.001a60 - 0.989±0.001c + 0.126±0.002b90 + 0.348±0.001d ++ 0.219±0.002c

Table 3Absorbance at 420 nm of water sample solutions or corresponding supernatants


Many authors showed that the fluorophores are precursors of the brown pigmentsbut are not identical to them. In view of this non-identity and of the fact that inhibitionof pigment formation does not affect the formation of fluorescent products (Burtonet al., 1962a; Burton et al., 1962b, Burton et al., 1962c) and resultant deteriorationof nutritional quality, they proposed the measurement of fluorescence as a more usefulvariable than browning development for assessing nutritional quality.

Each fluorescent molecule has a characteristic excitation and emission spectrum,which can be used to separate and identify molecules as well as to differentiatebetween substitutions and conformations of the same molecule.

Traditionally, fluorescence has been applied to clear solutions with knownfluorophores. The measurements are carried out using an angle of 90° between thesample and the excitation light. In such situations, and when the concentration is belowa certain level, the measured intensity is proportional to the concentration and followsLambert–Beer’s law. Scatter, quenching, and inner filter effects destroy thisrelationship when the concentration is high or when the sample is turbid or solid.Instead, front-face fluorescence spectroscopy can be applied. It measures fluorescenceemitted only from the sample surface, which reduces the influence of nonfluorescencedisturbances. In front-face fluorescence, the angle between the sample and the lightbeam is changed to, for instance, 60° (Andersen & Mortensen, 2008).

Advanced Maillard reaction products (such as pyrrole and imidazole derivatives)are fluorophores and are formed as intermediary products during heating of manyfoods. Thus, fluorescence, measured by using excitations at 340, 350, or 360 nm andemissions at 415 or 440 nm has been used as a measure of the level of AGE-modifiedproteins in vitro (Morales & van Boekel, 1997; Tessier et al., 2002). From literature,fluorescence measurement is frequently used in MR studies at physiological conditionswith regards to glycosylation of proteins in the human body, the AGEs formation,and AGEs relates pathologies as well (Sell et al., 1989). The measurement offluorescence related to Maillard compounds (FIC) in milk and milk-resemblingsystems has been studied in detail by Morales et al. (1996). Later, determination offluorescence and front-face fluorescence has been also used as an indicator ofnutritional damage in heat-treated milk, breakfast cereals, cooked salmon and roastedsoy (Birlouez-Aragón et al., 2001).

It must be remarked within the food matrix the total pool of fluorescent Maillardcompounds is formed by those that are free in the media and the linked-to-proteinsfraction. Both free and total FIC index can be determined by different analyticalprocedure, and then the linked-to-protein fraction can be estimated by subtraction (Mo-rales & van Boekel, 1997). To release the linked-to-protein FIC some procedureincluding enzymatic digestion must be applied, such as an in vitro gastrointestinaldigestion or the employment of a non-specific proteolytic enzyme, i.e. pronase. Asobvious, after this enzymatic treatment the FIC values will be much higher than incase of the determination of free FIC.


Our research group carried out a study of total and free fluorescence associatedto Maillard compounds in breakfast cereals (Delgado-Andrade et al., 2006). Here arepresented some of the most interesting results:

Free FIC determination: samples were suspended in 5 ml of deionized water and thetube was shaken vigorously for 1 min and clarified with 0.25 ml each of Carrez I (potassiumferrocyanide, 15% w/v) and Carrez II (zinc acetate 30% w/v) solutions. The resulting mixturewas centrifuged, the supernatant was collected in a 10 ml volumetric flask and two furtherextractions were performed using 2 mL of deionized water. The supernatants were mixedand the volume was made up to 10 ml with deionized water. Solutions were filtered,adequately diluted to prevent quenching effects, and measured at an excitation wavelengthof 347 nm and emission wavelength of 415 nm. The linearity of fluorescence response waschecked with a quinine sulphate solution of 1 μg/mL dissolved in 0.1 mol/L H2SO4. Thissolution was assigned 100% of relative fluorescence intensity (FI) and the results wereexpressed as percentage of relative fluorescence respect to the quinine sulphate solution asreference. FIC values are expressed as percentage of fluorescence intensity (FI) per mg ofsample or per mg of protein.

Total FIC (free + linked-to-protein backbone) determination: an enzymatic hydrolysisusing pronase E was performed. Briefly, samples were digested with a 0.375 mg/mL pronaseE solution (1500 U/mL in 1M sodium-borate solution, pH 8.2) in a stoppered test tube for36 h at 40°C in a water bath under shaking. After cooling, the solution was centrifuged at4,500 g for 10 min at 4°C. The supernatant was filtered and adequately diluted to preventquenching effects. Afterwards, the measurement of the fluorescence was done setting thesame conditions as free FIC. Results are presented in Table 4 and Figure 1.

As expected, values of free FIC were significantly lower than the corresponding totalFIC values, since total FIC comprises free fluorescence compounds in the food matrix aswell as that linked-to-protein backbone (Table 4). Results agree with stated previously byMorales and van Boekel (1997) for milk-resembling systems. Fluorescent Maillard reactioncompounds in breakfast cereals behave as described for milk-based systems, whereas protein-linked fluorescent structures are quantitatively more relevant in the assessment of the overall

Factor Free FIC Total FIC

FI /mg sample FI /mg protein FI /mg sample FI /mg protein

TotalAverageMedian

Minimum Maximum 1st quartile 3rd quartile

114 ± 13

86 30 697 61 126

1439 ± 130 1279 332 6043 789 1857

1005 ± 69 892 234 2408 560 1355

12745 ± 29 12980 4148 30921 7907 16444

Table 4Average free and total FIC present in a group of 60 breakfast cereals

marketed in Spain


fluorescence. Ratio between total to free FIC was 10.4 ± 4.12 (mean ± sd), but ratio wassignificantly higher in the group of rice-based breakfast cereals (Figure 1). Number of rice-based breakfast cereals analysed were limited (n = 3) therefore result could not berepresentative enough.

REFERENCES

AMES, J. M. and NURSTEN, H. E. (1989). Recent advances in the chemistry of colouredcompounds formed during the Maillard reaction. In Trends in Food Science, eds W. S.Lien and C. W. Foo. Singapore Institute of Food Science and Technology,

AMES, J.M., APRIYANTONO, A. and ARNOLDI, A. (1993). Low molecular weight colouredcompounds formed in xylose–lysine model systems. Food Chem., 46, 121–127.

ANDERSEN, C.M. and MORTENSEN, G. (2008). Fluorescence spectroscopy: A rapid tool foranalyzing dairy products. J. Agric. Food Sci., 56, 720-729.

ANDREWS, G.R. and MORANT, S.V. (1987). The Maillard browning of common amino acidsand sugars. J. Food Sci,. 49, 1206–1207.

ASHOOR, S.H. and ZENT, J.B. (1984). Maillard Browning of common amino acids and sugars.J. Food Sci., 49, 1206-1207.

BAISIER, W.M., LABUZA, T.P. (1992). Maillard browning kinetics in a liquid model system. J.Agric. Food Chem., 40, 707-713.

BATES, L., AMES, J.M., MACDOUGALL, D. and TAYLOR, P.C. (1998). Laboratory reaction cellto model Maillard color development in a starch-glucose-lysine system. J. of Food Sci.,63, 991-996.

0

300

600

900

1200

1500

1 2 3 4

FIC

/ mg s

ampl

e

corn-based wheat-based rice-based mixture

Free FIC Total FIC

10.3 ± 0.6a 10.1 ± 0.8a 15.9 ± 3.8b 9.9 ± 1.1a

Free FIC / Total FIC ratio

0

300

600

900

1200

1500

1 2 3 4

FIC

/ mg s

ampl

e

corn-based wheat-based rice-based mixture

Free FIC Total FICFree FICFree FIC Total FICTotal FIC

10.3 ± 0.6a 10.1 ± 0.8a 15.9 ± 3.8b 9.9 ± 1.1a

Free FIC / Total FIC ratio

Figure 1Free and total FIC (FI /mg sample) content in different groups of breakfast cereals

depending on the type of grain. Free FIC / Total FIC ratio for each group is alsodisplayed. Different letters indicates significant differences (One-way Anova and

Duncan Test P < 0.05)


BENJAKUL, S., LERTITTIKUL, W., BAUER, F. (2005). Antioxidant activity of Maillard reactionproducts from a porcine plasma protein-sugar model system. Food Chem., 93, 189-196.

BERTELLI, L., TORREGGIANI, D. and BERTOLO, G. (1996). Nonenzymatic browning inhydrolysed concentrated cheese whey permeate. Food Chem., 55, 353–358.

BIRLOUEZ-ARAGÓN, I., LECLERE, J., QUEDRAOGO, C.L., BIRLOUEZ, E., GRONGNET, J.F. (2001).The FAST method, a rapid approach of the nutritional quality of heat-treated foods.Nahrung, 45, 201-205.

BOSSET, J. O. and BLANC, B. (1978) Determination of the colour of milk and milk productsusing a tristimulus refractance photometer (Hunter system). Twentieth International DairyCongress, Paris, pp. E427-E428.

BURTON, H. S, MC WEENY, D. J. and BILTCLIFFE, D. O. (1962b). Sulphites and aldose–aminoreactions. Chemistry & Industry, 219.

BURTON, H. S, MC WEENY, D. J., & BILTCLIFFE, D. O. (1962c). The influence of pH on aldoseamino reactions. Chemistry and Industry, 1962, 1682.

BURTON, H. S, MC WEENY, D. J., PANDHI, P. N. and BILTCLIFFE, D. O. (1962a). Fluorescentcompounds and nonenzymatic browning. Nature, 98, 948–950.

CHOI, P., KONCUS, F., OMALLEY, M. and FAIRBANKS, W. (1949).A proposed method for thedetermination of colour of dry products of milk. J. Dairy Sci., 32, 580-586.

CREMER, D.R. and EICHNER, K. (2000). The reaction kinetics for the formation of streckeraldehydes in low moisture model systems and in plant powders. Food Chem., 71, 37-43.

DEL CASTILLO, M.D., AMES, J.M., GORDON, M.H. (2002). Effect of roasting on the antioxidantactivity of coffee brews. J. Agric. Food Chem., 50, 3698-3703.

DELGADO-ANDRADE, C., RUFIÁN-HENARES, J.A. and MORALES, F.J. (2006). Study onfluorescence of Maillard reaction compounds in breakfast cereals. Mol. Nutr. Food Res.,50, 799-804.

DELGADO-ANDRADE, C., SEIQUER, I., NAVARRO, M.P. and MORALES, F.J. (2007). Maillardreaction indicators in diets usually consumed by adolescent population. Mol. Nutr. FoodRes., 51, 341-351.

DELGADO-ANDRADE, C., SEIQUER, I., NIETO, R. and NAVARRO, M.P. (2004). Effects of heatedglucose-lysine and glucose-methionine model systems on mineral solubility. Food Chem.,87, 329-337.

ELLIS, G. P. (1959) The Maillard reaction. Advances Carbohydrate Chem., 14, 63-134.FERNÁNDEZ-ARTIGAS, P., GUERRA-HERNÁNDEZ, E. and. GARCÍA-VILLANOVA, B. (1999).

Browning indicators in model system and baby cereals. J. Agric. Food Chem., 47, 2872–2878.

FOGLIANO, V., MONTI, S.M., MUSELLA, T., RANDAZZO, G. and RITIENI A. (1999). Formationof coloured Maillard reaction products in a gluten-glucose model system. Food Chem.,66, 293-299.

FRANCIS, F.J., and CLYDESDALE, F.H. (1975). Food colorimetry: Theory and applications, AviPublishing Co, Westport, Connecticut.

GIANGIACOMMO, R. and MESSINA, G. (1988). Determinazione oggetiva del colore del lattealimentare mediante colorimetria tristimolo. Scienza e Tecnica Lattiero-Caesaria 39, 20–39.

HAYASHI, T., TERAO, A., UEDA, S. and NAMIKI, M. (1985). Red pigment formation by thereaction of oxidized ascorbic and protein in a food model system of low moisture content.Agric. Biol. Chem., 49, 3139-3144.

HENLE, T., WALTER, H. and KLOSTERMEYER, H. (1991). Evaluation of the extent of the earlyMaillard reaction in milk products by direct measurement of the Amadori productlactuloselysine, Zeitschrift für Lebensmittel-Untersuchung und-Forschung, 193, 119–122.


HODGE, J. E. (1953) Chemistry of browning reactions in model systems. J. Agric. Food Chem.,1, 928-934.

HORAK, F. P. and KESSLER, H. G. (1981). Colour measurement as an indicator of heat treatmentof foods, as exempli ed by milk products]. Zeitschrift fur Lebensmittel-Technologie undderfahrenstechnik 32, 180-184.

HUNTER, R.S., Photoelectric tristimulus colorimetry with three filters. NBS Circular C429.Department of Commerce, Washington D.C., 1942.

INGLES, D.L. and GALLIMORE, D., (1985). A new method for the isolation of melanoidins fromthe Maillard reaction of glucose and glycine. Chemistry and Industry, London, p. 194.

JING, H. and KITTS, D.D. (2004). Antioxidant activity of sugar-lysine Maillard reaction prductsin cell free and cellculture systems. Arch. Biochem. Biophys., 429, 154-163.

KELLY, K., JUDD, D., Colour: Universal language and dictionary of names. U.S. National Bureauof Standards, NBS Special publication 440. Department of Commerce, Washington D.C.,1976.

KESSLER, H.G. and FINK, R. (1986). Changes in heated and stored milk an interpretation byreaction kinetics. J. Food Sci. 51, 1105–1111.

LABUZA, T. P. and SALTMARCH, M. (1981). Kinetics of browning and protein quality loss inwhey powders during steady state and nonsteady state storage conditions. J. Food Sci.,47, 92-96.

LEDL, F. and SEVERIN, T. (1978). Braunungsreaktionen von pentosen mit aminen. Zeitschriftfur Lebensmittel Intersuchung und Forschung 167, 410-413.

LEDL, L. and SEVERIN, T. (1982) Investigations of the Maillard Reactions. XVI (1). Formationof coloured compounds from hexoses. Zeitschrift fur Lebensmittel Intersuchung undForschung, 175, 262-265.

LOWRY, K.R. and BAKER, D.H. (1989). Effect of storage, carbohydrate composition and heatprocessing on protein quality and amino acid bioavailability of a commercial enteralproduct. J. Food Sci. 54, 1024–1030.

MAILLARD, L. C. (1912) Action des acides amine«s sur les sucres; formation des melanoidinespar voi methodique. Comptes Rendues d´Academie de Science 154, 66-68.

MAURON, J. (1981) The Maillard reaction in food, critical review from the nutritionalstandpoint. In Progress in Food and Nutritional Science 5, 5-35.

MONTI, S.M., BAILEY, R.G. and AMES, J.M. (1998). The influence of pH on the non-volatilereaction products of aqueous Maillard model systems by HPLC with diode array detection.Food Chem., 62, 369-375.

MORALES, F. J. and VAN BOEKEL, M. A. J. S. (1997). A study on advanced Maillard reactionin heated casein/sugar solutions: fluorescence accumulation. Int. Dairy J., 7, 675–683.

MORALES, F.J. and JIMÉNEZ-PÉREZ, S. (2004). Peroxyl radical scavenging activity ofmelanoidins in aqueous systems. Eur. Food Res. Tech.,218, 515-520.

MORALES, F.J. and VAN BOECKEL, M.A.J.S. (1999). A study on advanced Maillard reactionin heated casein/sugar solutions: Colour formation. Int. Dairy J., 8, 907–915.

MORALES, F.J., ROMERO, C., JIMÉNEZ-PÉREZ, S., Fluorescence associated with Maillardreaction in milk and milk-resembling systems. Food Chem., 57, 423-428.

MORALES, F.J., VAN BOEKEL, M.A.J.S., A study on advanced Maillard reaction in heated casein/sugar solutions: colour formation. Int. Dairy J. 1998, 8, 907-915.

NAMIKI, M. (1988) Chemistry of Maillard reactions: recent studies on the browning reactionmechanism and the development of antioxidants and mutagens. Advances Food Res., 32,115-184.


NURSTEN, H.E. (1986). Maillard browning reactions in dried foods, Concentration and Dryingof Foods, Elsevier Applied Science Publishers, London, pp. 53–68.

O´BRIEN, J. and MORRISSEY, P. A. (1989). Nutritional and toxicological aspects of Maillardbrowning reaction in foods. Crit. Rev. Food Sci. and Nutr.K,. 28, 211-248.

PAGLIARINI, E., VERNILE, M. and PERI, C. (1990). Kinetic study on colour changes in milkdue to heat. J. Food Sci., 55, 1766–1767.

PALOMBO, R., GERTLER, A. and SAUGY, I., (1984). A simpli ed method for determination ofbrowning in dairy powders. J. Food Sci., 49, 1609-1613.

PETRIELLA, C., RESNIK, S. L., LOZANO, R. D. and CHIRIFE, J. (1985). Kinetics of deteriorativereaction in model food systems of high water activity: colour changes due to non-enzymaticbrowning. J. Food Sci., 50, 622-626.

RAMÍREZ-JIMÉNEZ, A., GARCÍA-VILLANOVA, B. and GUERRA-HERNÁNDEZ, E. (2001). Effectof toasting time on the browning of sliced bread. J. Sci. Food Agric., 81, 513–518.

RAMPILLI, M. and ANDREINI, R. (1992). Evaluation of colour components in sterilized milk.Italian J. Food Sci., 4, 285–291.

REINECCIUS, G.A. (1990). The influence of Maillard Reaction on the sensory properties offoods. In P.A. Finot, H.U. Aeschbacher, R.F. Hurrell, & R. Liardon, The Maillard Reactionin Food Processing, Human Nutrition and Physiology. Birkhaüser, Basel, pp.: 157-170.

RENN, P.T. and SATHE, S.K. (1997). Effects of pH, temperature, and reactant molar ration onL-leicin and D-glucose Maillard browning reaction on an aqueous system. J. Agric. FoodChem., 45, 3782-3787.

REYES F. G. R., POOCHAROEN B. and WORLSTAD, R. E. (1982). Maillard browning reactionof sugar glycine model systems: changes in sugar concentration, colour and appearance.J. Food Sci., 47, 1376-1377.

RHIM, J.W., JONES, V.A. and SWARTZEL, K.R. (1988). Kinetic studies in the colour changes ofskim milk. Lebensmittelwissenschaft und Technologie, 21, 334–338.

RHIM, J.W., JONES, V.A., SWARTZEL, K.R., Kinetics studies in the colour changes of skim milk.Lebensm-wiss Technol. 1988, 21, 334-338.

RHIM, J.W., JONES, V.A., SWARTZEL, K.R. Kinetics studies in the colour changes of skim milk.Lebensm-wiss Technol. 1988, 21, 334-338.

RIZZI, G.P. (1997). Chemical structure of coloured Maillard reaction products. Food Rev. Int.,13, 1–28.

RIZZI, G.P., Chemical structure of coloured Maillard reaction products. Food Rev. Int. 1997,13, 1-28.

RUFIÁN-HENARES, J.A., GARCÍA-VILLANOVA, B. and GUERRA-HERNÁNDEZ, E. (2002). Lossof o-phthaldialdeyde reactivity fluorescence, and colour in stored enteral formula, Int. J.Dairy Tech., 35, 34–39.

SELL, D.R., MONNIER, V.M. (1989). Structure elucidation of a senescence crosslink from humanextra-cellular matrix. J. Biol. Chem., 264, 597-602.

TESSIER, F., MONNIER, V. M. and KORNFIELD, J. A. (2002). Characterization of novelchromophores, fluorophores and cross-links from glyceraldehydes, lysine and arginine.International Congress Series, 1245, 303–311.

WEBB, B. H. and HOLM, G. E. (1930). Color of evaporated milks. J. Dairy Sci., 13, 25-39.


Practical assays 3Colour and browning measurement

CHEMICALS

Carrez I (potassium ferrocyanide, 15% w/v) and Carrez II (zinc acetate 30% w/v).All the chemical necessary to perform the in vitro gastrointestinal digestion as

described in the respective chapter.

SAMPLES

Wheat and wheat bran breads (W and WB, respectively) commercialised as pre-baked breads to be finally baked at home. They were kept in an air oven for 12, 20 or26 minutes at 200°C, and, after aerated, they were lyophilised. One piece of eachkind of bread was also lyophilised without heat treatment as a control of the initialstage in the pre-baked breads. All of them were milled in a grinder. Different powderedsamples obtained were named as follows:



1. Tristimulus colour or CIELab system

After setting up the instrument and perform the autocero with the blanks providedby the manufacturer, samples are placed into the glass dish (variable diameterdepending on the concrete apparatus). The sample is illuminated with D65-artificialdaylight (10° standard angle) according to conditions provided by manufacturer andthe parameter L*, a* and b* are shown in the display of the instrument. Measurementmust be performed at least in triplicate with different portions of the samples to ensurethe reproducibility.


2. Browning measurement

Free browning compounds

There are many different procedures in the bibliography to perform thismeasurement and all of them offer a little different results, but all of them are suitableto compare the sample each other within the study. The advantage of the next one isthat in the sample filtered solution obtained not only absorbance can be measured,but also fluorescence as well as HMF and furfural.

Procedure: 0.5 g of sample are suspended in 5 ml of deionized water and the tubeis shaken vigorously for 1 min and clarified with 0.25 ml each of Carrez I and CarrezII solutions. The resulting mixture is centrifuged, the supernatant is collected in a 10ml volumetric flask and two further extractions are performed using 2 mL of deionizedwater. The supernatants are mixed and the volume is made up to 10 ml with deionizedwater. Solutions are filtered, adequately diluted if necessary, and measured at 280(early and low molecular weight Maillard compounds) and 420 (final and highmolecular weight Maillard compounds) nm.

Total browning compounds

The measurement of total browning compounds (free + linked-to-proteinbackbone) needs a previous step including an enzymatic hydrolysis. Proceduresincluding a proteolytic digestion with pronase E are efficiency, but in this case wehave selected a complete in vitro gastrointestinal digestion to better simulate thephysiological conditions. The inclusion of α-amylase in the digestion protocol of thesesamples is important since they are cereal-derived products.

Absorbance is measured at both 280 and 420 nm in the supernatants of the invitro gastrointestinal digestion of the samples.

3. Fluorescence associated to Maillard compounds

Currently, due to technical reasons, fluorescence measurement is not possible inour facilities. In any case, as previously mentioned, free FIC could be determined inthe same filtered solution obtained for the free absorbance measurement and total FICin the supernatants of the in vitro gastrointestinal digestion following the proceduredescribe in Delgado-Andrade et al. (2006).


RESULTS

Data from CIE-Lab colour

ParametersSample

HeatingTime L* a* b*

0

12

20

26

Wheat bread

0

12

20

26

Wheat bran bread


Data from absorbance measurement before and after thein vitro gastrointestinal digestion

Before in vitro digestion After in vitro digestion Sample

HeatingTime 280 nm 420 nm 280 nm 420 nm

0

12

20

26

Wheat bread

0

12

20

26

Wheat bran bread

Gastrointestinal digestion as first conditioningof nutrient bioavailability

CRISTINA DELGADO-ANDRADE1, ANA HARO1, ROSA CASTELLANOUnit of Animal Nutrition, Estación Experimental del Zaidín, CSIC,

Granada, Spain

JOSÉ ÁNGEL RUFIÁN-HENARESDepartment of Nutrition and Bromatology, Faculty of Pharmacy,

University of Granada, Spain

The nutritional status of man is largely influenced by the level and bioavailabilityof nutrients in food. Bioavailability can be defined as the proportion of a nutrientthat can be utilized for normal physiological functions. The main component ofbioavailability refers to the digestion and absorption of nutrients in the gut, which isthe main rate-limiting factor. Studying this socalled intestinal bioavailability(Ekmekcioglu, 2000) in humans is very costly and not always easy due to ethicalreasons and large interindividual variations. Therefore, experimental models are usedto overcome these problems. A variety of in vitro methods were presented in the lasthalf of the twentieth century yielding more or less reliable results. These techniquesare mainly based on:

1. In vitro digestion of homogenized foods in a closed system and determinationof the soluble nutrient fraction (Narasinga Rao & Prabhavathi, 1978).

2. In vitro digestion and dialyzability of soluble nutrients across a semipermeablemembrane with or without removement of the dialyzed nutrients (Wolterset al., 1993).

3. Usage of Caco-2cells and either measurement of:a. Cellular uptake of nutrients after transmembranous dialysis during intes-

tinal digestion of foods (Glahn, Wien, Van Campen, & Miller, 1996)b. Transepithelial transport of solubilized nutrients from digested foods

(Ekmekcioglu, Pomazal, Steffan, Schweiger, & Marktl, 1999). The maindisadvantages of these models are the lack of the complex mucosal barrierwith all of its regulatory processes and also the static transport conditionsnot allowing an accurate calculation of whole fractional transport and fluxrates (Ekmekcioglu, 2002).

Therefore, although in vivo methods cannot reproduce in vivo studies, some ofthe in vitro methods demonstrate similar results with human studies. They are basedon the observation of the physiologic conditions of the gastrointestinal digestion(Cummings, 1991) to reproduce them in vitro:

Cristina Delgado-Andrade; Ana Haro; Rosa Castellano; José Ángel Rufián-Henares52

— The pH in the oral cavity is about 6.5 and the residence time is from secondsto minutes

— The compound (plus matrix) is swallowed through the esophagus into thestomach.

— The half-emptying time for the stomach is 8 to 15 min for fasting conditions,and 0.5 to 3 h for fed conditions. The gastric pH is between 1 and 2, andbetween 2 and 5 for fasting and fed conditions, respectively.

— The small intestine consists of three sections: duodenum, jejunum, and ileum.The residence time in the first one is 0.5 to 0.75 h, at a pH between 4 and5.5. The residence time in the jejunum is 1.5 to 2 h, at a pH between 5.5and 7.0. The residence time in the ileum is 5 to 7 h, at a pH between 7.0and 7.5.

— The small intestine goes over into the colon, with a residence time between15 and 60 h, at a pH generally between 6.0 and 7.5.

Besides all these pH conditions, diverse enzymes, saliva and different salts arepresent in the gastrointestinal tract to enable the absorption of the nutrients containedin the foods. Of course, the gastrointestinal tract is also the entry door for non-nutrients, if they are able to pass the gut barrier. This could be the case of MRP ortheir degradation products.

In this sense, significant quantities of dietary MRPs and melanoidins enter thehuman intestines every day, but very little is known about their metabolism therein.Neither the effects of digestive enzymes in the small intestines nor those of the bac-teria in the large bowel on melanoidin degradation have been described yet. In a recentstudy, a roasted mixture of glucose-glycine was demonstrated to change the dynamicsof large bowel bacteria (Ames et al., 1999). After a maximum time of 24 h ofincubation, the total number of anaerobes, bacteroides, clostridia, and bifidobacteriaincreased in a batch culture fermenter containing human fecal bacteria. Although nodegradation product was identified in this study, a bacterial growth clearly indicatedthat gut bacteria are able to utilize melanoidins prepared from a roasted glucose-glycine mixture as energy substrates.

Concerning the absorption rate of MRP, many studies have shown that at leastlow molecular weight components of melanoidins or their intestinal degradationproducts are absorbed in quantities up to about 30% (Somoza, 2005). Results fromour research group in a first approach to test hydroxymethylfurfural (HMF) absorptionin Caco-2 cells have pointed out that the compound is absorbed in different rates fromdigested breakfast cereals, depending on the composition of the ready-to-eat cereal(Delgado-Andrade et al., 2008). However, it seems that high molecular fractions ofMRP are mainly excreted in faeces and the absorption rate is scarce. The role ofunabsorbed melanoidins and their degradation products on human health also remainsan open question (Somoza, 2005).

Gastrointestinal digestion as first conditioning of nutrient bioavailability 53

Rufián-Henares and Delgado-Andrade have applied the in vitro gastrointestinaldigestion to a group of breakfast cereals to establish its effect on the bioaccessibilityof certain MRP as the Amadori compounds (measured as furosine) and HMF (Ru-fián-Henares & Delgado-Andrade, 2009). Tables 1 and 2 show average contain ofHMF and furosine on the studied breakfast cereals and the distribution between solu-ble and insoluble fraction after the in vitro digestion.

Table 2Mean furosine content in the groups of the studied breakfast cereals andfurosine distribution in the soluble and insoluble fractions after in vitro

digestion of samples.

Table 1Mean HMF content in the groups of the studied breakfast cereals and HMF

distribution in the soluble and insoluble fractions after in vitro digestionof samples.

º

GroupsHMF

(mg/Kg)

SolubleHMF(%)

InsolubleHMF(%)

Standard 133 ± 10a 89 ± 23a 14 ± 4a

Fibre enriched

63 ± 4b 92 ± 26a 23 ± 10b

Total mean 98 ± 5 91 ± 22 19 ± 8

a - b Different letters within the same column indicate statistical differences (One-way ANOVA andDuncan test, p<0.05).

º

GroupsFurosine(mg/Kg)

SolubleFurosine

(%)

InsolubleFurosine

(%)

Standard 14 ± 0.3a 49 ± 14a 50 ± 13a

Fibre enriched

103 ± 0.7b 50 ± 8a 50 ± 8a

Total mean 59 ± 1.0 49 ± 10 51 ± 10

a – b Different letters within the same column indicate statistical differences (One-way ANOVA andDuncan test, p<0.05).


After the digestion process two samples (one of each group) showed an amountof HMF in the soluble fraction higher than that found in the corresponding raw ce-real (112.3 and 121.7%). The high HMF content of the soluble fraction could comefrom the release of protein-bound Amadori products and its conversion to HMF dueto the digestive process. That could be the reason why the mean recovery is higherthan 100% after the digestion process (90.6% HMF in the soluble fraction and 18.5%still attached to the insoluble fraction). No statistically significant differences (p<0.05)were found in the soluble fraction of the two groups (Table 1) whereas a statisticallyhigher HMF concentration (p<0.05) was present in the insoluble matter of fibre addedcereals. This behaviour could be explained if HMF were linked to the fibre matrix.

Regarding to furosine, contrary to HMF, fibre enriched cereals showed astatistically significant (p<0.05) ten-time higher furosine content than standard corncereals (103.2 and 14.2 mg/Kg respectively). This could be explained taking intoaccount that the addition of fibre supposes an increase in the protein content and adecrease on the amount of reducing sugars (data not shown); therefore, fibre increasesthe rate of Maillard reaction in cereals (which is responsible of furosine formation)whereas at the same time decreases the caramelisation rate (which is thought to bethe main responsible of HMF generation in breakfast cereals) due to a diluting effectover reducing sugars. In the case of digested samples almost a 50% was found inboth soluble and insoluble fractions (Table 2), obtaining no significant differencesamong fibre added/non-added groups. This means that fibre does not play an importantrole over the Amadori products solubility but probably other cereal components couldinterfere with furosine solubility.

Protein-linked fructoselysine (FL, the Amadori compound of lysine) has beendemonstrated to be hardly accessible to digestive enzymes of the gastrointestinal tract(Erbersdobler, 1977). Whereas the urinary excretion of dietary free FL can reach upto 60% in rats (Finot & Magnenat, 1981), only 10-30% of the orally ingested FL isreleased out of the food protein by gastrointestinal digestion (Erbersdobler & Faist,2001). A recent study conducted by Somoza et al (Somoza et al., 2006) have shownin rats that only a 3-5% of the ingested FL derived from heat-treated casein modelswas excreted in urine. These data are in line with those found in our study, whereonly a 50% of the furosine is located in the soluble fraction after digestion, whichcould give rise to a reduced absorption and then a low urinary excretion.

On the other hand, in some cases it is also interesting to analyses not only theeffects of the gastrointestinal digestion on the bioaccessibility of MRP, but also howthe bioaccessibility of other nutrients/compounds are affected by the presence of MRP.This is the case of mineral nutrients.

Diverse studies have suggested that MRPs are capable of behaving as anionicpolymers and of complexing metal ions, producing soluble and insoluble complexes.According to several authors (Rendleman, 1987; Wijewickeme & Kitts, 1998) calciumand copper are bound by soluble and insoluble melanoidins derived from differentamino acid–sugar model systems. MRPs, prepared from monosodium glutamate andglucose, are capable of fixing Ca, Zn, Cu and Mg (O´Brien & Morrissey, 1997). Other


Maillard browned compounds, from proline, leucine, glycine, glutamic acid or lysine,autoclaved with glucose, bind zinc and decrease its apparent availability, and lysineproducts present the greatest binding capacity (Whitelaw & Weaver, 1988). After theaddition of a heated casein–glucose–fructose sample to a calcium solution, a higherpercentage of metal appears insoluble in comparison with the raw sample, while therest remains soluble but, partially, in non-ionic form (Aspe et al., 1993). Homma andhis coworkers (Homma & Murata, 2001; Homma, et al., 1986) also demonstrated theexistence of metal chelating compounds in browned pigments of co?ee with di?erenta?nities for iron, copper and zinc.

As described, once again, studies in model systems help to monitor the chelatingbehaviour of these complex compounds. In this sense, some trials carried out by ourresearch group have shown the different capability of MRP from glucose-lysine orglucose-methionine model systems, obtained at 150° for 30, 60 or 90 min, to modifythe mineral solubility at intestinal conditions of pH, ionic strength and tempetature(Delgado-Andrade et al., 2004) (Figures 1 and 2).

0

20

40

60

80

100

GL0 GL30 GL60 GL90

Ca

Mg

Fe

Cu

Zn

Figure 1Effects of the heat treatment of glucose-lysine samples on mineral solubility

(%). GL0, unheated glucose-lysine mixture; GL30, GL60 and GL90, glucose-lysine mixtures heated at 150˚C for 30, 60 or 90 min, respectively.

The findings showed that heating mixtures of glucose–lysine and glucose–methionine has little or no effect on Ca and Mg solubility. With respect the traceelements, early compounds formed in both model systems seem to favour mineralsolubility, maximum after 30 min of heating in the glucose–lysine system, and after60 min in the glucose–methionine one supporting the slowing down of the MR whenmethionine is one of the reactants, what was also reflected in the lower browningdevelopment. It was therefore concluded that the effects of browning products


generated during food processing should be taken into account, especially regardingtrace element availability, as in vitro solubility may be considered as a prior indicationof in vivo mineral utilization.

One of there most interesting point of the MRP chelating activity is the possibleeffect on mineral bioavailability. Regarding to this, our own data have highlightedthe negative effects of MRP consumption on Ca and Mg concentrations in bone (Del-gado-Andrade et al., 2005; Delgado-Andrade et al., 2008), or their effects in Cu andFe metabolism (Delgado-Andrade et al., 2002; Delgado-Andrade et al., 2004; Mesíaset al., 2009).

In summary, the in vitro gastrointestinal digestion results an interesting tool topredict nutrient availability as well as the bioaccessible fraction to be absorbed in afood. The enzymatic treatment is extremely important when the target compoundsare linked to proteins, otherwise the present of the mentioned compounds/nutrientand their biological actions could be underestimated. Besides the simulatedgastrointestinal digestion, alternative treatments including non-specific enzymes canbe applied (i.e. pronase E treatment), but none of them simulate the physiologicalconditions as the in vitro gastrointestinal digestion.

Figure 2Effects of the heat treatment of glucose-methionine samples on mineral

solubility (%). GM0, unheated glucose-methionine mixture; GM30, GM60 andGM90, glucose-methionine mixtures heated at 150˚C for 30, 60 or 90 min,

respectively.

0

20

40

60

80

100

GM0 GM30 GM60 GM90

Ca

Mg

Fe

Cu

Zn


REFERENCES

AMES, J. M., WYNNE, A., HOFMANN, A., PLOS, S. and GIBSON, G.R. (1999). The effect of amodel melanoidin mixture on faecal bacterial populations in vitro. Br. J. Nutr., 82, 489–495.

ASPE, T., VAQUERO, M.P. & NAVARRO, M.P. (1993). Changes in the speciation and utilizationof calcium in rats by the presence of untreated or heated casein, separately or in combinationwith glucose and fructose. In: Bioavailability’93-nutritional, chemical and food processingimplications of nutrient bioavailability. Nutrient-Bioavailability Symposium (Part II),University of Antwerp, Belgium. (pp. 68–72).

CUMMINGS, J.H., MACFARLANE, G.T. (1991). The control and consequences of bacterialfermentation in the human colon. J. Appl. Bacteriol., 70, 443-459.

DELGADO-ANDRADE, C., SEIQUER, I., and NAVARRO, M.P. (2002). Copper metabolism in ratsfed diets containing Maillard reaction products. J. Food Sci., 67, 855-860.

DELGADO-ANDRADE, C., SEIQUER, I., NAVARRO, M.P. (2004). Bioavailability of iron from aheat treated glucose-lysine model food system: assays in rats and in Caco-2 cells. J. Sci.Food Agric., 84, 1507-1513.

DELGADO-ANDRADE, C., SEIQUER, I., NAVARRO, M.P. (2005). Comparative effects of glucose-lysine versus glucose-methionine Maillard reaction products consumption: in vitro and invivo calcium availability. Mol. Nutr. Fod Res., 49, 679-684.

DELGADO-ANDRADE, C., SEIQUER, I., NAVARRO, M.P. (2008). Maillard reaction productsconsumption: magnesium bioavailability and bone mineralization in rats. Food Chem., 107,631-639.

DELGADO-ANDRADE, C., SEIQUER, I., NAVARRO, M.P., Morales, F.J. (2008). Estimation ofhydroxymethylfurfural availability in breakfast cereals. Studies in Caco-2 cells line. FoodChem. Toxicol., 46, 1600-1607.

DELGADO-ANDRADE, C., SEIQUER, I., NIETO, R. and NAVARRO, M.P. (2004). Effects of heatedglucose-lysine and glucose-methionine model systems on mineral solubility. Food Chem.,87, 329-337.

EKMEKCIOGLU C. (2002) A physiological approach for preparing and conducting intestinalbioavailability studies using experimental systems Food Chem., 76, 225–230.

EKMEKCIOGLU, C. (2000). Intestinal bioavailability of minerals and trace elements from milkand beverages in humans. Nahrung, 44, 390–397.

EKMEKCIOGLU, C., POMAZAL, K., STEFFAN, I., SCHWEIGER, B., and MARKTL, W. (1999).Calcium transport from mineral waters across Caco-2 cells. J. Agric. Food Chem., 47, 2594–2599.

ERBERSDOBLER, H.F. (1977). Adv. Exp. Med. Biol. In: Friedman, M. (Ed.), Protein Crosslinking– Nutritional and Medical Consequences, (Vol 86B) (pp. 367–378). New York: PlenumPress.

ERBERSDOBLER, H.F. and FAIST, V. (2001). Metabolic transit of Amadori products. Nahrung,45, 177-181.

FINOT, P.A. and MAGNENAT, E. (1981). Metabolic transit of early and advanced Maillardproducts. Prog. Food Nutr. Sci., 5, 193-207.

GLAHN, R. P., WIEN, E. M., VAN CAMPEN, D. R., and MILLER, D. D. (1996). Caco-2cell ironuptake from meat and casein digests parallels in vivo studies: use of a novel in vitro methodfor rapid estimation of iron bioavailability. J. Nutr., 126, 332–339.

HAZELL, T. and JOHNSON, I.T. (1987). In vitro estimation of iron availability from a range ofplants foods: influence of phytate, ascorbate and citrate. Br. J. Nutr., 57, 223-233.


HOMMA, S., and MURATA, M. (2001). Metal chelating compounds in instant coffee. Ann. Nutr.Metab., 45(S), 394.

HOMMA, S., AIDA, K., and FUJIMAKI, M. (1986). Chelation of metal with brown pigments incoffee. In M. Fujimaki, M. Nakmiki, & H. Kato (Eds.). In:Amino-carbonyl reactions infood and biology systems. Amsterdam: Elsevier Science. (pp. 165–172).

LUTEN, J., CREWS, H., FLYNN, A., VAN DAEL, P., KASTENMAYER, P., HURRELL, R., DEELSTRA,H., SHEN, L.H., Fairweather-Tait, S., Hickson, K., Farré, R., Schlemmer, U. and Frøhlich,W. (1996). Interlaboratory trial on the determination of the in vitro iron dialysability fromfood. J. Sci. Food Agric., 72, 415-424.

MESÍAS, M., SEIQUER, I., DELGADO-ANDRADE, C., GALDÓ, G., NAVARRO, M.P. (2009). Intakeof Maillard reaction products reduces iron bioavailability in male adolescents. Mol. Nutr.Food Res, in press.

MILLER, D.D., SCHRICKER, B.R., RASMUSSEN, R.R. and VAN CAMPER, D. (1981). An in vitromethod for estimation of iron availability from meals. Am. J. Clin. Nutr., 34, 2248-2256.

MILLS, D.J.S., TUOHY, K.M., BOOTH, J., BUCK, M., CRABBE, M.J.C., GIBSON, G.R. and AMES,J.M. (2008). Dietary glycated protein modulates the colonic microbiota towards a moredetrimental composition in ulcerative colitis patients and non-ulcerative colitis subjects.J. App. Microbiol., 105, 706-714.

NARASINGA RAO, B. S., and PRABHAVATHI, T. (1978). An in vitro method for predicting thebioavailability of iron from foods. Am. J. Clin. Nutr., 31, 169–175.

NAVARRO, P., ASPE, T. and SEIQUER, I. (2000). Zinc transport in Caco-2 cells and Zinc in rats:influence of the heat treatment of a caseinglucose-fructose mixture..J Agric. Food Chem.,48, 3589–3596.

O´BRIEN, J. and MORRISSEY, P.A. (1997). Metal ion complexation by products of the Maillardreaction. Food Chem., 58, 17-27.

RENDLEMAN, J. A. (1987). Complexation of calcium by the melanoidin and its role indetermining bioavailability. J. Food Sci., 6, 1699–1705.

RUFIÁN-HENARES, J.A. and DELGADO-ANDRADE, C. (2009). Effect of digestive process onMaillard reaction indexes and antioxidant properties of breakfast cereals. Food Res. Int.,in press.

SEIQUER, I., DELGADO-ANDRADE, C. and NAVARRO, P. (2000). Effects of the heat treatmentof casein on Zn transport and uptake by Caco-2 cells. Pro.c Nutr. Soc., 59,134A.

SOMOZA, V. (2005). Five years of research on health risks and benefits of Maillard reactionproducts: An update. Mol. Nutr. Food Res., 49, 663-67.

SOMOZA, V., WENZEL, E., WEISS, C., CLAWIN-RDECKER, I, GRUBEL, N. and ERBERSDOBLER,H.F. (2006). Dose-dependent utilisation of casein-linked lysinoalanine, N(epsilon)-fructoselysine and N(epsilon)carboxymethyllysine in rats. Mol. Nutr. Fod Res.,, 50, 833–841.

WHITELAW, M. L., and WEAVER, M. L. (1988). Maillard browning effects on in vivo availabilityof zinc. J. Food Sci., 53, 1508– 1510.

WIJEWICKREME, A.N. and KITTS, D.D. (1998). Metal chelating and antioxidant activity ofmodel Maillard reaction product. In: Process-induced chemical changes in food. Shahidiand others (Ed.) p 245-254. New York: Plenum Press.

WOLTERS, M. G., SCHREUDER, H. A., VAN DEN HEUVEL, G., VAN LONKHUIJSEN, H. J.,HERMUS, R. J., and VORAGEN, A. G. (1993). A continuous in vitro method for estimationof the bioavailability of minerals and trace elements in foods: application to breads varyingin phytic acid content. Br. J. Nutr., 69, 849–861.


Practical assays 4In vitro gastrointestinal digestion

CHEMICALS

Solutions— 1mM Cl2Ca solution pH 7.0— 6N HCl solution— 0.1N HCl solution— 1 and 0.1 M NaHCO3 solution

Enzymes— α-amylase— pepsin— pancreatin— bile salts

SAMPLES

Wheat and wheat bran breads (W and WB, respectively) commercialised as pre-baked breads to be finally baked at home. They were kept in an air oven for 12, 20 or26 minutes at 200°C, and, after aerated, they were lyophilised. One piece of eachkind of bread was also lyophilised without heat treatment as a control of the initialstage in the pre-baked breads. All of them were milled in a grinder. Different powderedsamples obtained were named as follows:


IN VITRO GASTROINTESTINAL DIGESTION

The technique of Miller et al. (Miller et al., 1981), modified to our requirements,was followed. The inclusion of α-amylase is done according to Mills et al. (2008).


It comprised three stages:

1. Oral digestion:— 1 g of each lyophilised sample is brought to a final volume of 10 ml of

milli-Q water and shaken vigorously in a vortex.— Shortly before use, prepare an a-amylase (32.5 mg) solution in 25 ml 1mM

ClCa pH 7. Add 250 μL to the sample solution.— Incubate at 37°C for 30 min.

2. Gastric digestion:— After incubating, pH of sample must be adjustment to 2 with HCl 6N.— Shortly before use, 0.4 g of pepsin was dissolved in 2.5 ml of 0.1 M HCl

and added at a proportion of 0.05 g of pepsin/g of sample (313 μL of thesolution in our particular case).

— Check pH and readjust to 2 if necessary.— Incubate at 37°C in a shaking water bath at 110 oscillations/min for 2 h.— Take out the samples and keep in crushed ice for 5 min and afterwards 5

min at room temperature.— Adjust pH at 6 with 1M NaHCO3 solution dropwise.

3. Intestinal digestion:— For intestinal digestion, 0.1 g of pancreatin and 0.625 g of bile salts were

dissolved in 25 ml of 0.1 M NaHCO3.— Adjust the pH of the digest pH 6 with 1M NaHCO3 dropwise.— Add 2.50 ml of pancreatin + bile salts mixture. The pH is then adjusted to

pH 7.5 with 1M NaHCO3, and samples are incubated at 37°C at 110oscillations/min for 2 h.

After oral and gastrointestinal digestion, the digestive enzymes must be inactivatedby heat treatment for 4 min at 100°C in a polyethylenglycol bath. The samples werethen cooled by immersion in a crushes ice bath and centrifuged at 3200g for 60 minat 4°C to separate soluble and non-soluble fractions. It is useful to measure the finalvolume of supernatant obtained aimed to perform calculations of concentration ofanalysed nutrients/compounds.

MEASUREMENT OF TRACE MINERAL SOLUBILITY AFTER IN VITRO DIGESTION AS

INDICATOR OF MINERAL AVAILABILITY

As previously mention, diverse studies have suggested that MRPs are capable ofbehaving as anionic polymers and of complexing metal ions, producing soluble andinsoluble complexes. According to several authors (Rendleman, 1987; Wijewickeme& Kitts, 1997) calcium and copper are bound by soluble and insoluble melanoidinsderived from different amino acid–sugar model systems. MRPs, prepared from


monosodium glutamate and glucose, are capable of fixing Ca, Zn, Cu and Mg (O‘Brien& Morrissey, 1997). Rendleman (1987) demonstrated the importance of MRP in thebinding of metal ions in foods, noting that toast made from bread containing milkhad a stronger affinity for Ca and Cu than toast prepared from milk-free bread.Consequently, the chelating mineral capacity of MRP and the conjoint presence ofboth in the intestinal lumen suggest the possibility that mineral availability may bealtered (Seiquer et al., 2000; Navarro et al., 2000).

Therefore, since mineral solubility is a first conditioning of its bioavailability,the determination of the soluble minerals after the in vitro gastrointestinal digestion,as the bioaccessible fraction of the food, is suitable to establish the portion availableto be absorbed. Despite the utility of this procedure, since it simulates the physiologicconditions, one never must forget that in vivo the situation can always be a littledifferent because the insoluble fraction could be available in some extent due to theactions of the gut microflora on the digested food matrix. Moreover, the non-solublefraction could also have some kind of local action in the gut (i.e. antioxidant,antimicrobial, etc.).Therefore, the activity (i.e. antioxidant activity) or the nutrient(i.e. mineral or some MRP) measured in the soluble fraction after the in vitro digestioncould underestimated the global values in vivo, as a consequence of the presence ofboth fractions in the gut.

After performing the in vitro gastrointestinal digestion of all the bread samples,the effects of different heating time (and then different degree of Maillard reaction)on mineral solubility can be determined by measuring different elements in thesupernatants and pellet separated. In particular, we will measure Fe, Cu and Zn, sinceMRP seem to have a very high affinity by them, giving rise to stables chelates thatcould modify mineral availability.

Aliquots of the powdered samples as well as the soluble and insoluble fractionsobtained after the in vitro digestion are completely digested by the addition ofconcentrated HNO3 (5ml), HNO3:HClO4 (1:4) (5ml) and by heating at hightemperatures (180-220°C) in a sand beaker. Once finished the mineralisation step, allsamples were diluted with milli-Q water to an appropriate volume for measurement.The mineral determinations are carried out with flame AAS in a Perkin-Elmer Analyst700 Spectrophotometer building a calibration curve for each mineral analysed.


RESULTS

Data from mineral content in the samples and mineral solubilityafter in vitro digestion

SamplesFe

(mg/g)Cu

(mg/g)Zn

(mg/g)0

Total content % Soluble

% Insoluble 12


% Insoluble 20


% Insoluble

Wheat bread

26Total content

% Soluble % Insoluble

0Total content


12Total content


20Total content


Wheat bran bread

26Total content


Antimicrobial activity of Maillardreaction products

JOSÉ ÁNGEL RUFIÁN-HENARESDepartment of Nutrition and Food Science, Faculty of Pharmacy,

University of Granada, Spain.

CRISTINA DELGADO-ANDRADEUnit of Animal Nutriton, Estación Experimental del Zaidín, CSIC,

Granada, Spain

The Maillard reaction (MR) produces a large amount of compounds with differentchemical structures and molecular weights. The high molecular weight compoundsformed in the last stage of the Maillard reaction are called melanoidins. Melanoidinsare widely distributed in foods and could exert different in-vitro functional propertiessuch as antioxidant (Delgado-Andrade et al., 2005), antihypertensive (Rufián-Henares& Morales, 2007), metal-binding activity (Morales et al., 2005), antimicrobial (Ru-fián-Henares & Morales, 2006) and prebiotic (Borrelli et al., 2004). In recent years,several studies have been mainly focused in the effect of melanoidins on the humandiet and their possible nutritional, biological and health implications (Ames et al.,1999; Faist & Erbersdobler, 2001).

Some investigations highlight the role of melanoidins in vivo since melanoidinsescape digestion and pass through the upper gastrointestinal tract (Faist &Erbersdobler, 2001) and then can interact with the different microbial species presentin the hindgut (Finot & Magnenat, 1981). Recently Borrelli and Fogliano (2005) haveshown that bread crust melanoidins can be metabolised/fermented by the humanhindgut microflora and that these melanoidins selectively enhance the growth ofbifidobacteria, which are desirable bacteria in the gut due to their health-promotingproperties.

Most investigations focussed on the effect of MRP on microorganisms have beendone in specific microbial growth media, which show that MRP can stimulatemicrobial growth (Jemmali. 1969) or inhibit it (Einarsson. 1987; Stecchini et al. 1993).Antimicrobial activity of Maillard reaction products (MRPs) has been previouslystudied in model systems (Einarsson et al., 1983; Einarsson, 1987; Stecchini et al.,1991; Lanciotti et al., 1999; Del Castillo et al., 2007) and coffee (Daglia et al, 1994a,b;Daglia et al., 1998) but specific studies on antimicrobial activity of melanoidins arescarce (Rufian-Henares and Morales, 2006).

Einarsson and co-workers (1983) measured the antimicrobial activity of arginine-xilose and histidine-glucose Maillard reaction mixtures. They fractionated the model

José Ángel Rufián-Henares; Cristina Delgado-Andrade64

systems by dialysis (molecular cut-off at 1000 Da) and found that the high molecularweight fraction exerted a higher bacteriostatic antimicrobial activity than the lowmolecular weight one, although the culture media, pH and temperature play animportant role. This authors stated that the high molecular weight fraction (melanoidinsprincipally) could develop its antimicrobial action by lowering iron solubility (whichis essential for growth and survival of pathogenic bacteria) so that there is a decreaseon glucose, serine and oxygen uptake. In addition, other authors referred theantimicrobial activity of Maillard reaction compounds to inhibit the sugar catabolisingenzymes of microorganisms (Lanciotti et al., 1999) or their potential antioxidantactivity (Mattila and Sandholm, 1989). However, the exact mechanism by whichmelanoidins or MRP affect the bacteria growth is not known.

Chemical structure of melanoidins has not been completely elucidated, but, inthe last few years, more data became available. In particular, it was shown that theycan have a different structure according to the different starting material and behaveas anionic material: in some cases they are mainly formed by a carbohydrate skeletonwith few unsaturated rings and a small nitrogen component, in other cases they canhave a protein structure linked to small chromophores or phenolic residues (Nunes &Coimbra, 2007; Bekedam et al., 2007; Cämmerer & Kroh, 1995; Morales, 2002). Otherauthors have found that melanoidins can form stable complexes with metal cations(Gomyo and Horikoshi, 1976; Migo et al., 1997). Hashiba (1986) reported that theketone or hydroxyl groups of pyranone or pyridone residues can act as donor groupsin melanoidins and participate in the chelation with metals. However, Morales et al.(2005) found that the chromophoric groups were not the main co-ordination sites foriron complexation in the melanoidin structure. The results in the literature seem toindicate that the antimicrobial activity of melanoidins could be related to their anioniccharge and the ability of chelating some cations like Fe, Zn and Cu (Homma andMurata, 1995), which are essential for growth and survival of pathogenic bacteria.

1. METHODS FOR MEASURING THE ANTIMICROBIAL ACTIVITY OF MELANOIDINS

There are many biological assay techniques that have been developed to monitorand measure the antimicrobial activity of natural compounds (Casey et al., 2004).The standard disk diffusion and broth dilution methods are the wider extended methodsused to study the bioactivity of chemical compounds (Wanandy et al., 2004). Both ofthem need the use of standardised microbial techniques and are time consuming. Inaddition, the disk diffusion method proves unreliable in certain applications (Swensonet al., 1989) and can lead to interpretational problems such as growth in the inhibitionzone (Piliouras et al., 2002) and subjectivity associated with visual assessments(Deighton and Balkau, 1990).

Microtiter plate-based assays have been developed for a wide range of applicationsincluding monitoring of bioactive compound production in fermentation samples(Cassey et al., 2004), determination of antimicrobial susceptibility patterns ofmicroorganisms (Jones and Dudley., 1997), etc. The major problem of this kind of

Antimicrobial activity of Maillard reaction products 65

assays is that they use turbidity or absorbance to monitor biomass growth and deathrates (Nayak et al., 2002; Lopez-García et al., 2003). In this way such assays aretime consuming because of the need of reach an enough growth to take the readings.

The development of a microtiter plate-based assay method that could measurefast and accurately the antimicrobial activity of natural compounds like melanoidinswould be of significant benefit. In this sense, a fast, economic and easy test has beendeveloped. It is based on a commercial test, can be used in laboratories withoutmicrobiological facilities and address the demand for rapid antimicrobial screeningof natural products from plants, spices, food, etc.

1.1. Commercial test

Antimicrobial assay is performed with a ready-to-use test (Eclipse®-100) suppliedby Zeu-inmunotec (Spain). Melanoidins are dissolved in distilled water at differentconcentrations. Sample (100 μl) is added to every well of a 96-well plate. They areallowed to diffuse for 1 hour and subsequently the wells were washed 4 times withdistilled water. For assay I (static method) the plate is incubated for 2.5 hours at 65± 0.1°C and finally the absorbance of each well is read at 600 nm on a microplatereader. For assay II (dynamic method) the plate is incubated in the same way but theabsorbance of each well is read every 15 minutes. In every assay it is used as positivecontrol oxytetracyclin at 100 μg/l (ppb) and distilled water as negative one.

The test used in this work, Eclipse®-100, is a commercial test developed for thequalitative detection of microbial growth inhibitors and/or antibiotics in raw andthermally treated milk samples. This test is based on a gram-positive thermophilicbacterium, Geobacillus stearothermophylus var. calidolactis that grows at 65°C andis very sensitive to substances which exert an antimicrobial effect. In this way, thetest is suitable to detect low concentrations of antibiotics as low as 0.003 mg/l forpenicillin.

The culture medium has an appropriate concentration of G. Stearothermophylusspore and an acid-basic indicator. When the culture medium is incubated at 65°C thespore germinate and start to multiply producing lactic acid as a by-product of itsmetabolism; thus, the indicator turns from blue to yellow, which can be followed bymeasuring at 600 nm, the maximum absorption point for blue. In other way, when anantimicrobial agent or a substance with antimicrobial action is placed in the culturemedium, the microbial multiplication is inhibited, consequently lactic acid is avoidedand the pH indicator remains blue. The test are useful over a pH range from 5.2 to8.8 although it is necessary a high dilution for coffee and wine melanoidins becauseof its initial low pH. In addition, melanoidin colour did not interfere with thespectrophotometric reading.

— Static method: The static method is a useful and fast way to detect the presenceof substances with inhibitory bacterial growing activity. The use of a referenceantimicrobial substance as positive control allows the comparison between its activityand the sample activity and so, to obtain another way to compare the different


antimicrobial power of different substances. In this sense, the use of an antibiotic ispractical to standardise the evolution of the antimicrobial activity of differentsubstances.

— Dynamic method: The assay with two different antibiotics (Oxytetracyclin-bacteriostatic; penicillin-bactericide) during different periods of time allowed to findout that oxytetracyclin colour changed from blue to yellow at prolonged times whereaspenicillin colour remained blue. Thus, the change of colour during the assaydevelopment could differentiate between a bacteriostatic and a bactericide activity.Figure 1 shows the profile of penicillin, oxytetracyclin and water during the assay.Penicillin response shows a low decrease during the assay whereas oxytetracyclinabsorbance decrease very fast at about 400 minutes and stabilised about 100 minuteslater. This behaviour means that at 400 minutes the antimicrobial activity started todecrease and consequently microorganisms’ growth increased, being observed a colourchange that stabilised after another 100 minutes.

Figure 1Time course of antimicrobial activity of standards

(oxytetracyclin and penicillin at 100 µg/l and 3 µg/l respectively) and water (blank)

0,000

0,500

1,000

1,500

2,000

0 200 400 600 800

Time

(minutes)

Abs

orba

nce

0 200 400 600 800

H2O

Oxytetracyclin

0.0

0.5

1.0

1.5

2.0 Penicillin

0,000

0,500

1,000

1,500

2,000

0 200 400 600 800

Time

(minutes)

Abs

orba

nce

0 200 400 600 800

H2O

Oxytetracyclin

0.0

0.5

1.0

1.5

2.0 Penicillin

Time

(minutes)

Abs

orba

nce

0 200 400 600 800

H2O

Oxytetracyclin

0.0

0.5

1.0

1.5

2.0 Penicillin

Time

(minutes)

Abs

orba

nce

0 200 400 600 800

H2O

Oxytetracyclin

0.0

0.5

1.0

1.5

2.0 Penicillin

Time

(minutes)

Abs

orba

nce

0 200 400 600 800

H2O

Oxytetracyclin

0.0

0.5

1.0

1.5

2.0 Penicillin


In figure 2 it is shown the absorbance/time plot obtained for a 25 μg/loxytetracyclin solution with the dynamic assay. There is an initial period where theabsorbance remained constant and a turning point at about 300 minutes whereabsorbance started to decrease fast until about 450 minutes where it stabilised. Thefirst turning point is called T0 and is defined as the time where the antimicrobialactivity starts to decline, the point T50 is defined as the time where the antimicrobialactivity has decreased up to the 50% and finally, the second turning point is calledT100 and is the time where there is no remainder antimicrobial activity. This lack ofresidual antimicrobial activity is determined by the blank assay, when the samplesreach a final absorbance value of 0.350 ± 0.050 absorbance units. T0, T50 and T100 canbe calculated mathematically. T0 is calculated as the intersection point between theline obtained for the initial straight absorbance period and the line obtained for theabsorbance decreasing period. In the same way, T100 is calculated as the intersectionpoint between the line obtained for the absorbance decreasing period and the lineobtained for the final straight absorbance period. Finally, T50 is calculated as the meanvalue of T0 and T100.

Figure 2Determination of T0, T50 and T100 for oxytetracyclin solution (25 µg/l)

Time(minutes)

Abs

orba

nce

0 200 400 600 800

0.0

0.4

0.8

1.2

1.6

(1) T0

(3) T100

2.0

(2) T50

(1) (2) (3)

There is a better lineal correlation between T100 and oxytetracyclin concentration(r2 = 0.9999) than that obtained with T0 (r2 = 0.9897) probably due to the more irre-gular behaviour of T0 (starting time of antimicrobial activity decrease). The correlationobtained between the oxytetracyclin concentration and T100 was used as standard curveto evaluate the antimicrobial activity of the samples. In this way, substituting the T100

of a sample it could be obtained the value of the oxytetracyclin concentration (μg/l)that produces an equivalent antimicrobial activity than the sample does. This value


was called OTEV (OxyTetracyclin Equivalent Value). Figure 3 reports the results ofthe application of the dynamic assay to melanoidins. As can be seen there is asubstantially difference between the T100 values of the assayed melanoidins. In thissense, the higher the T100 value, the stronger the antimicrobial activity (wine > lowroasted coffee > light beer). In addition, the form of the curve obtained for melanoidinsis the same to that observed for oxytetracyclin, which confirms their bacteriostaticaction, previously reported by Einarsson et al. (1983).

Figure 3Time course of antimicrobial activity of light beer (LB), coffee (CTn110)

and wine (W)

As stated with the static method, the higher antimicrobial activity was for themelanoidin obtained from coffee CTn60 –the coffee with the higher thermaltreatment— which exerts an antimicrobial activity equivalent to an oxytetracyclinsolution of 1221 μg/l (table 2). In addition, it can be observed an increase inantimicrobial activity with the roasting degree of coffees, from an OTEV of 260 forlight roasting to 4060 for dark roasting. A similar behaviour is observed for beermelanoidins, showing the black beer a higher activity (OTEV of 820) than thatobtained for lager beer (OTEV of 20). The antimicrobial activity of wine, toastedand black beer melanoidins ares higher than that for CTn85 coffee melanoidins butmuch lower than CTn60 melanoidin.

Abs

orba

nce

Time(minutes)

0.0

0.4

0.8

1.2

1.6

2.0

0 200 400 600 800

LB

W

CTn110


1.2. Non-commercial test

The methods used for the screening of food antimicrobials may be divided intoendpoint and descriptive methods (Skyttä & MattilaSandholm, 1991). The most widelyused endpoint method is the agar diffusion assay although its limitations are generallyrecognized. Since the size of the inhibition zone is dependent upon the rate of diffusion(Davidson & Parish, 1989) misleading results may be obtained if the agent to be testedis hydrophobic and will not diffuse into the agar. Furthermore, the end point methodsscarcely provide information of the effects of the compounds on the growth kineticsof the microorganisms. In descriptive tests sampling or automated recording ofmicrobial growth is carried out at timed intervals. Consequently, one can evaluatethe effects of an antimicrobial over a longer period of time owing to the kineticrecording of results. Various growth curve parameters can be used to describe theinhibitory effects. The extension of the lag phase is probably the most widely usedparameter. It has been shown that in a food system, even a slight delay of a lag phasemay have an important influence on its shelf life. Other parameters, such as end-absorbance (Mattila, 1987), slope, which is a gradient of the exponential growth phase(Adams & Hall, 1988), and area under the growth curve have also been used (Borrelli& Fogliano, 2005; MattilaSandholm, 1989). The latter was considered to be the bestdescriptor of overall inhibitory effect since it covers the entire growth period insteadof referring to single points of time. Continuous monitoring offers an even moreimportant advantage over the agar diffusion test in that the inhibition of growth canbe detected as soon as the measured parameter deviates significantly from uninhibitedcontrol.

In a previous study (Rufián-Henares et al., 2008a), a first approximation to assessthe antimicrobial activity of melanoidins was evaluated by a generic gram-positivethermophilic bacterium (Geobacillus stearothermophilus var. calidolactis) as a veryefficient discriminate test (Rufian-Henares & Morales, 2006). But former approachhad not a direct application to strains of interest in Food Science and Technologyalthough is useful for a rapid evaluation of the potential antimicrobial activity. Then,a new rrapid plate-based method to study the effect of model and food melanoidinson the growth of some selected pathogenic organism (Escherichia coli andStaphylococcus aureus) was developed. Antimicrobial activity is evaluated as areaunder the growth curve as compared to a control. In addition, results can be alsoexpressed as Oxytetracyclin equivalent concentration.

Overnight suspensions of E. coli or S. aureus are growth at 37°C in brain heartinfusion medium (BHI) until a concentration of 109 colony forming units (cfu)/mlare reached. Culture of bacteria are diluted 1/1000 in fresh broth used to culture themprior to each experiment to approximately 106 cfu/ml. Based on their routine growthcharacteristics, bacterial strains are diluted to optical densities (600 nm) of 1.480 ±0.020 and 1.150 ± 0.015 for S. aureus and E. coli, respectively. Then 250 μl of bacterialcells suspensions are inoculated into a sterile 96-well microplate. Subsequently 50 μlof sample, sterile distilled water (blank assay) or 100 μg/l solution of oxytetracyclin


(positive control of bacteriostatic activity) are. Microbial growth kinetics are recordedon a microplate reader incubated at 37°C ± 1°C for 24 hours. Turbidity is measuredas absorbance at 600 nm, being lectures taken every 5 minutes. The area under thecurve (AUC) is calculated according to the equation

AUC = Abs0/Abs0 + Abs0/Abs5 + Abs0/Abs10 + … + Abs0/Absn

being Abs0 = Absorbance at time 0, Abs5 = Absorbance at time 5 minutes, and soforth. However, in order to eliminate the differences in the AUC corresponding tothe different absorbidities of the melanoidins, the initial absorbance of each sampleis subtracted to their following readings, then obtaining the same initial lecture (equalto the absorbance of the BHI medium at 600 nm). Finally, the Net AUC is calculatedby subtracting the AUC of the blank sample to the AUC of each melanoidin or positivecontrol. The addition of increasing concentrations of a coffee melanoidin to the BHImedium shows a classical dose-dependent inhibition profile (figure 4).

Figure 4

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

0 250 500 750 1000 1250 1500

Incubation time (minutes)

OD

blank

10 µg/l

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

0 250 500 750 1000 1250 1500


OD

blank

10 µg/l

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

0 250 500 750 1000 1250 1500


OD

blank

10 µg/l

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

0 250 500 750 1000 1250 1500


OD

blank

10 µg/l

Net AUC vs. oxytetracyclin concentration is plotted for both strains (figure 5)and a linear response in the working area is obtained. Where higher oxytetracyclinconcentration in the media, later the microbial growth start. Working area, expressedas Net-AUC, is estimated up to 130 and 350 for E. coli and S. aureus respectively.Experimental results shows that working area could be much extended foroxytetracyclin but it is possible some side-effects at high concentrations of theunknown substance for testing. Inhibitory activity of target substances which givesNet-AUC higher than the assigned maximum values must be re-evaluated at a morediluted concentration.


2. ANTIMICROBIAL ACTIVITY OF MELANOIDINS

2.1. Bacteriostatic and bactericide activity

The results obtained with the above stated methods have been published indifferent papers (Rufián-Henares & Morales, 2006, 2007, 2008a and 2008b). In thecase of the bacteriostatic activity the results obtained are shown in table 1. Thepercentage of sample activity at a concentration of 1 mg/ml against oxytetracyclin100 ppb gives information of the activity of every melanoidin at a standardisedconcentration compared with our reference antibiotic at a concrete concentration. Then,the higher and the lower % activity against oxytetracyclin were for dark coffee andpilsner beer respectively. Melanoidins isolated from more severely treated samplesexerted higher inhibitory bacterial growing activity, such as dark coffee and dark beer.

As stated with the static method, with the dynamic one the higher antimicrobialactivity was for the melanoidin obtained from dark coffee, which exerts anantimicrobial activity equivalent to an oxytetracyclin solution of 1221 μg/l (table 2).In addition, it could be observed an increase in antimicrobial activity with the roastingdegree of coffees, from an OTEV of 260 for light roasting to 4060 for dark roasting.A similar behaviour was observed for beer melanoidins, showing the black beer ahigher activity (OTEV of 820) than that obtained for lager beer (OTEV of 20). Theantimicrobial activity of wine, toasted and black beer melanoidins was higher thanthat for medium roasting coffee melanoidins but much lower than dark coffeemelanoidin.

Figure 5

y = 0,1134x + 0,026

R2 = 0,999y = 0,0437x + 0,1746

R2 = 0,9991

0

4

8

12

16

0 100 200 300 400

Net-AUC

Ox

yte

tra

cc

yc

lin

(µ

g/l)


Because of the successfully application of the commercial method to food-derivedmelanoidins, they were assayed with the non-commercial methodology. The resultsobtained are shown in table 3 as OTEV, which gives information of the activity ofevery melanoidin at a standardised concentration compared with a referencebacteriostatic antibiotic. OTEV value is useful for interlaboratory comparison for thesame methodology and if different approaches are being used. In the case of S. aureusthe OTEV values range from 7.9 to 152.3 μg/l, with a median value of 17.3. In the

Table 1Antimicrobial activity of melanoidins (1 mg/ml) measured with static assay

Melanoidin % Activity against Oxytetracyclin 100 g/l

Pilsner beer 105

Toasted beer 1347

Black beer 1754

Sweet wine 1675

CTn60 (dark roasting) 7423

CTn85 (medium roasting) 1901

CTn110 (light roasting) 534

Table 2Antimicrobial activity of melanoidins (1 mg/ml) measured with dynamic assay

Melanoidin % Activity against Oxytetracyclin 100 g/l

Pilsner beer 23

Toasted beer 230

Black beer 260

Sweet wine 257

CTn60 (dark roasting) 1221

CTn85 (medium roasting) 198

CTn110 (light roasting) 94


case of E. coli the values were lower than those obtained for S. aureus, probablybecause of this is a Gram-positive bacteria whereas E. coli is a Gram-negative one,which are known to be more resistant to the action of antibiotics in a general way.There are statistically significant differences (p < 0.01) between the differentmelanoidins of the same type of food in the case of both S. aureus and E. coli. Inaddition, as reported above melanoidins isolated from more severely treated samplesexerted higher inhibitory bacterial growing activity for both S aureus and E. coli. Inour samples, there was a very important increase in the antimicrobial activity withthe roasting degree of coffees, from an OTEV of 9.1 for light roasting to 152.3 formedium roasting, in the case of S. aureus whereas in the case of E. coli that increasewas lower, although statistically significant. A similar behaviour was observed forbeer melanoidins, showing the black beer a higher activity (OTEV of 19.7) than thatobtained for lager beer (OTEV of 7.9).

Table 3Antimicrobial activity of melanoidins (2 mg/ml) obtained from food

OTEVMelanoidin S. aureus E. coli

Pilsner beer 7.9 5.8

Toasted beer 9.1 7.5

Black beer 19.7 8.9

Sweet wine 15.5 7.3

CTn85 (medium roasting) 152.3 8.5

CTn110 (light roasting) 9.1 6.7

In other different work (Rufián-Henares and Morales, 2008b) the antimicrobialactivity of coffee and biscuits melanoidins was studied. Three fractions with differentmolecular weight (> 10 kDa, between 10-3 kDa and < 3 kDa) were obtained fromroasted coffee and a classical breakfast biscuit. Antimicrobial activities were evaluatedusing the minimum inhibitory concentration (MIC) values against a Gram-negativebacterium, Escherichia coli. Among the fractions, the most intense antimicrobialactivity was obtained for the 10 kDa fractions (10.0 and 7.5 μg/ml for coffee andbiscuits respectively) whereas up to 5 times lower activity was found for the LMWfraction; in the case of the 10-3 kDa fractions their antimicrobial activity wasintermediate to that of the above reported fractions (table 4).


On the other hand, the high or intermediate molecular weight (IMW) compoundsof biscuits exerted a slightly higher antimicrobial activity compared to the onesobtained for the coffee ones. On the contrary, the LMW fraction of coffee showed ahigher activity when compared to that of biscuits. This could be a consequence ofthe most intense thermal treatment applied to coffee, which gives rise to the generationof most LMW compounds with potential antimicrobial activity (such as heterocyclicamines, furans, etc) (Daglia et al., 1994a).

Some authors reported previously (Einarsson et al., 1983; Lanciotti et al., 1997)that melanoidins and other compounds from the Maillard reaction exerted abacteriostatic activity over different bacterial strains increasing the lag phase. Ourown investigation (Rufian-Henares & Morales, 2006) corroborated this hypothesis inthe case of isolated melanoidins from other sources at concentrations not higher than2 mg/l. In order to evaluate if the antibacterial activity obtained for every fractionwas bacteriostatic or bactericide, mixtures of sample - culture medium were incubatedduring 3 h. After this period, cells were harvested by centrifugation (in order toeliminate the active compounds present in the culture medium), incubated for 24h inBHI agar plates and finally the number of CFU was determined. It was found thatthe HMW fractions exerted a bacteriostatic activity at low concentrations (2.5 mg/ml) whereas this bacteriostatic activity was observed at higher concentrations forthe IMW and LMW fractions (up to 30 - 35 mg/ml for LMW). In addition, a bactericideactivity was obtained when the concentration raised some limits (5.0 mg/ml for HMW,10 - 15 mg/ml for intermediate compounds and 35-40 for LMW). At the MICconcentrations E. coli cells could not grown in the culture media, indicating that allthe cells were killed. These results support the hypothesis that melanoidins damage

Table 4MIC values of coffee and biscuits melanoidins fractions against E. coli.

FractionMIC value

(mg/ml)

RC10 10.0

RC3 25.0Coffee

FC3 45.0

RB10 7.5

RB3 20.0Biscuit

FB3 50.0

RC10 sample: Retentate of Coffee at 10 KDa; RC3: Retentate of Coffee at 3 KDa; FC3; Filtrate of Coffeeat 3 KDa. RB10 sample: Retentate of Biscuit at 10 KDa; RB3: Retentate of Biscuit at 3 KDa; FB3;Filtrate of Biscuit at 3 KDa.


cells in an irreversible way that inhibits their growth as well as if cells are exposedto melanoidins for a short period of time and then removed from the culture media.

Melanoidins, which are the main constituents of the HMW fraction, are brownanionic polymeric compounds (Bekedam et al., 2007; Nunes & Coimbra, 2007) withdifferent charge/mass ratio that depends on the degree of saturation of the reactivegroups in the core structure (Morales, 2002). In addition, melanoidins can bindessential metals like iron (Morales et al., 2005). In this sense some substances withanionic behavior and chelating properties like EDTA (Vaara, 1992) have antimicrobialactivity because of the profound effect on the OM permeability barrier of Gram-negative enteric bacteria. It removes, by chelation, stabilizing divalent cations fromtheir binding sites in LPS resulting under certain conditions in the rupture of the OM.

2.2. Mechanisms of antimicrobial activity

2.2.1. Effect of melanoidins over microbial membrane

Cell Integrity. The cytoplasmic cell membrane is the target of many antimicrobialagents, which interact with bacterial membranes causing changes in bacterialmembrane (Je & Kim, 2006). When bacterial membranes become compromised lowmolecular mass substances followed by macromolecules are leaving out. Theseintracellular components, such as DNA or RNA, are easily detected by UV at 260nm as an indication of membrane damage (Chen & Cooper, 2002). The release ofintracellular components from E. coli treated with coffee or biscuit fractions is shownin figure 6. The absorbance at 260 nm was increased in a time dependent manner dueto the addition of the different fractions. The release of intracellular components withthe HMW fraction (RB10 and RC10) was higher than that of IMW fractions (RB3and RC3) whereas slight changes were obtained for the LMW extracts. There was adramatic increase up to 40 min; thereafter, the absorbance was almost unchanged.These results indicate that the cell membrane disruption is faster for melanoidins,which correlates with their low MIC value.

OM Permeabilization of E. coli. Fluorescence of the NPN probe increases whenincorporated into the hydrophobic core of a Gram-negative bacterial cell membrane (afterpermeation) compared with the fluorescence of a non-permeating bacterial cell. Basedon this principle, the OM permeabilization of live E. coli by the different fractions wasmonitored using the hydrophobic NPN fluorescent probe. As shown in Figure 7 (bothA and B), the addition of high or intermediate molecular weight fractions E. colisuspensions in the presence of NPN caused an increase in fluorescence, which indicatesthat E. coli cells membranes were damaged by these fractions. In a general way theOM permeabilization exerted by the HMW fraction was faster and almost doubled theactivity of the IMW fractions. This result also strengthens the fact, previously statedfor the release of intracellular substances, that melanoidins quickly disrupts cellmembranes of bacteria more than the intermediate fractions. Moreover, it was observedthat the activity of RB10 was one third higher than the one of RC10.


Figure 6Effect of different coffee and biscuits fractions (10 mg/ml) on the cell integrity

measured as release of intracellular components at 260 nm.

RC10

RB10

RC3

RB3

FC3

FB3

Blank0.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 15 30 45

Ab

so

rba

nc

e(2

60

nm

)

Time (min)

RC10

RB10

RC3

RB3

FC3

FB3

Blank

RC10

RB10

RC3

RB3

FC3

FB3

Blank0.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 15 30 45

Ab

so

rba

nc

e(2

60

nm

)

Time (min)

0.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 15 30 450.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 15 30 450.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 15 30 45

Ab

so

rba

nc

e(2

60

nm

)

Time (min)

The OM of Gram-negative bacteria consists of lipopolysaccharides (LPSs) andproteins, and these are maintained together by electrostatic interactions with divalentcations required to stabilize the OM. Polycationic molecules such as chitosan, couldbind to the negatively charged O-specific oligosaccharide units of E. coli LPSs (Je &Kim, 2006). However, anionic substances such as EDTA, remove by chelationstabilizing divalent cations from their binding sites in LPS (Chen & Cooper, 2002)and thus disrupting the integrity of the OM and resulting in loss of the barrier functionor blocking of the nutrient flow with concomitant bacterial death due to depletion ofthe nutrients. It could be hypothesized that melanoidins, because of their negativesurface charge (Morales, 2002; Nunes and Coimbra, 2007) could bind to the OM byelectrostatic interactions and chelate Mg2+, which is a divalent cation that plays aspecific structural role stabilizing the prokaryotic membranes as metal ion bridgesbetween phosphate groups of phospholipids (Ibraim et al., 2000). In order to checkthis hypothesis, the effect of MgCl2 addition over the uptake of NPN by E. coli treatedwith the different fractions at a concentration of 10 mg/ml was assayed. The additionof MgCl2 produced a strong decrease on the fluorescence production, indicating theinteraction of melanoidins and the compounds isolated in the intermediate fractionwith Mg2+ as shown in table 5, which corroborates the above mentioned hypothesis.


Figure 7Effect of different coffee (A) and biscuits (B) fractions (10 mg/ml) on the outer

membrane permeabilization of E.coli as measured by uptake of NPN.

A( )

RC10

FC3

RC3

Time (min)

Re

lati

ve

Flu

ore

sce

nc

e

0

10

20

30

40

0 2 4 106 8

RB10

RB3

FB3

B( )

Time (min)

Re

lati

ve

Flu

ore

sce

nc

e

0

10

20

30

40

0 2 4 106 8

50

60


IM Permeabilization of E. coli. Destabilization of the OM is necessary to gainaccess to the IM. Because of OM permeation was demonstrated, the HMW and theIMW fractions could interact with the IM. In order to study this possible effect, theIM permeation was evaluated as a function of cytoplasmic μ-galactosidase release,with bacteria grown in lactose containing medium. When cells were treated with a10 mg/ml concentration of each fraction (figure 8A), a lag time of about 35 min wasfollowed by a progressive release of the cytoplasmicμμ-galactosidase in the case ofFG3 and FC3. However, in the case of the HMW fractions the release of μμ -galactosidase started immediately, reaching a steady state at 70 minutes. Thisbehaviour could be explained taking into account that the HMW fractions were presentin the culture media at bactericide concentrations, whereas IMW compounds werenot. In addition, the highestμμ-galactosidase release by RB10 corroborates its mostintense activity against E coli. In order to get more insight into the permeation activityof HMW fractions, they were used to follow the study because of their strong effectover IM permeation.

CCCP is an uncoupler of oxidative phosphorilation in mitochondria, being capableof depolarizing both plasmatic and mitochondria membranes and then it can inhibitthe IM permeation of those species that based its activity in the transmembranepotential. As seen in figure 8B, pretreatment of E coli with 100 μM CCCP resultedin a 2-fold increase of fractions mediated IM permeation, suggesting that membranedisruption is independent of the bacterial transmembrane potential. In order to seewhether the effect of the HMW fractions on IM increased the release of thecytoplasmic enzyme or membrane lesions allowing ONPG uptake into cells, thespecific activity of μ -galactosidase released in the cell-free medium was measured(Figure 8C). HMW fractions caused considerable release of the enzyme into the

Table 5Effect of MgCl2 addition over the uptake of NPN by E. coli treated with 10 mg/

ml of coffee or biscuits melanoidins fractions. Results are expressed as % offluorescence intensity ± standard deviation.

MgCl2Fraction

- +RC10 38.2 ± 3.1 8.9 ± 1.0

RC3 22.4 ± 1.3 3.8 ± 0.5Coffee

FC3 1.1 ± 0.2 0.9 ± 0.3

RB10 57.2 ± 2.4 10.8 ± 0.9

RB3 32.1 ± 1.4 4.6 ± 0.6Biscuit

FB3 0.8 ± 0.4 0.6 ± 0.2


medium within 60 min of incubation with cells, being double for biscuits HMWfraction. This results demonstrate that melanoidins disintegrates the bacterialmembranes and might allow for interference with biosynthetic processes occurringat the membrane, which are important in bacterial viability, such as inhibition of thetransport of nutrient and macromolecular precursors.

Figure 8Effect of different coffee and biscuit fractions on the inner membrane cell

integrity as measured by the release of cytoplasmic µ-galactosidase activity of E.coli. (A) Time dependence of permeation of cells treated with 10 mg/ml of coffeeor biscuits fractions. (B) Bacteria treated with 10 mg/ml of RC10 or RB10 for 30

minutes in the absence or presence of uncoupler CCCP. (C) Bacteria (after 60minutes incubation) were removed by centrifugation and enzyme release was

assayed in the cell-free supernatant. Ctrl, cells treated with saline alone.

( ) (B)A

0.04

0.08

0.12

00 30 60 90

RB10

RC10

FB3FC3

Time (min)

Ab

so

rba

nc

e(4

20

nm

)

0.04

0.08

0.12

0

(+)(-)

RC10Ctrl

RB10

CCCP 100 μM

(C)

Ctrl RB10RC100

0.2

0.4

0.6

β-G

ala

cto

sid

ase

–R

ele

ase

(μM

ON

P /

min

/ m

L)


2.2.2. Effect of coffee melanoidins over metal chelation: relationship with theirantimicrobial activity.

2.2.2.1. Antimicrobial activity of coffee melanoidins over different bacterial strains

The activity of coffee melanoidins was measured against eight different bacterialstrains, 2 Gram-positive and 6 Gram-negative. Figure 9 shows a classical dose-dependent inhibition profile of coffee melanoidins over E. coli ATCC 35150 (nosiderophore production). A decrease of the final absorbance —related with the numberof bacteria in the culture medium— was found when increasing the amount ofmelanoidin added, giving rise to a total inhibition of bacterial growth at the highestmelanoidin concentration assayed. This concentration was selected as the MIC forthat specific bacterial strain. Bacterial cells were harvested by centrifugation and re-incubated again in fresh culture medium without melanoidin, obtaining no bacterialgrowth after a 24h incubation period. This confirmed that the antimicrobial activityof coffee melanoidins at such high concentration is bactericidal. A higher melanoidinconcentration was necessary to inhibit the microbial growth in the case of the E. colistrain which is able to produce siderophores (ATCC 33475). This indicates that suchbacterial strain was more resistant to the melanoidin activity.

Figure 9

0 250 500 750 1000 1250 1500

Time (minutes)

0 (Blank)

0.5mg/ml1.0mg/ml1.5mg/ml2.5mg/ml5.0

0

0.2

0.4

0.6

1.2

1.0

0.8

Ab

sorb

an

ce

1.4

1.6

5.0 mg melanoidin/ml

0 (blank)





The results for all the strains studied, expressed as MIC values (table 6), showedthat the minimum concentration of melanoidin (with no rise to microbial growth) iscomprised between 2.0 and 8.0 mg/ml. Gram-positive microorganisms are moresensitive to the antimicrobial activity of melanoidins whereas Gram-negative bacte-


ria had higher MICs. This could be related to the absence in Gram-positivemicroorganisms of a low-permeability barrier such as the outer membrane, which madethem more susceptible to antimicrobial substances (Nikaido, 1998). Lastly, isnoteworthy to underline that those microorganisms with siderophore productioncapability (E. coli ATCC 33475 and P. aeruginosa 15692) showed the highest MICs,even higher than the same microorganism but without capability to producesiderophores (E. coli ATCC 35150 and P. aeruginosa 27853). Siderophores —considered as virulence factors— are molecules designed to form tight and stablecomplexes with ferric iron, allowing the microorganisms colonisating of adverse mediasuch as the human body (Miethke & Marahiel, 2007). Therefore, attending to thechelating properties of siderophores against iron, the higher resistance of those bac-teria capable of siderophore production could indicate that part of the antimicrobialactivity of coffee melanoidins is linked to iron chelation.

Table 6MIC values of coffee melanoidins

Bacteria Strain Shape/Gram Siderophore production

MIC value (mg/mL)

Bacillus cereus ATCC 11778 Rod/G+ - 2.5

Staphylocuccus aureus ATCC 25923 Coccus/G+ - 2.0

Proteus mirabilis ATCC 7002 Rod /G- - 4.5

Pseudomonas aeruginosa ATCC 27853 Rod /G- - 4.0

Salmonella typhimurium ATCC 13311 Rod /G- - 4.5

Escherichia coli ATCC 35150 Rod /G- - 5.0

Pseudomonas aeruginosa ATCC 15692 Rod /G- + 7.0

Escherichia coli ATCC 33475 Rod /G- + 8.0

3.2. Effect of coffee melanoidins over microbial membrane

Melanoidins cause E. coli membrane disruption by chelating Mg2+ ions from theouter membrane, giving rise to a destabilization of the inner membrane. In order tocheck this hypothesis in other different bacterial strains, the release of low molecularmass substances was recorded for each kind of bacterium incubated with an amountof melanoidin equal to its MIC. As depicted in figure 10, all the microorganismsreached the same final absorbance, which means that the final degree of cell damagewas similar. However, the speed at which every type of bacterium reached this finaldamage was different, being faster for Gram-positive and slower for Gram-negativesones, who had the highest MICs. Gram-negative bacteria have a strong outermembrane where melanoidins bind first and permeabilize (sublethal action). Thispermeation allows melanoidins to enter the cytoplasmic membrane where they cause


the leakage of cytoplasmic components (the lethal action). However, in the case ofGram-positive bacteria, they are surrounded by a thick peptidoglycan cell wall toocoarse, which can be rearranged easily by melanoidins. This explains the lower MICof Gram-positive bacteria and the faster release of cellular compounds.

Figure 10Release of intracellular components at 260 nm of different bacterial strains treated

with coffee melanoidins at their minimum inhibitory concentration (MIC)

E. coli ATCC 35150

P. aeruginosa ATCC 27853

S. typhimurium ATCC 13311

P. mirabilis ATCC 7002

E. coli ATCC 33475


B. cereus ATCC 11778

S. aureus ATCC 25923

0

0.3

0.6

0.9

1.2

1.5

1.8

Ab

sorb

an

ce

0 5 10 15 20 25 30

Time (minutes)

35 40 45

E. coli ATCC 35150


S. typhimurium ATCC 13311

P. mirabilis ATCC 7002

E. coli ATCC 33475


B. cereus ATCC 11778

S. aureus ATCC 25923

0

0.3

0.6

0.9

1.2

1.5

1.8

Ab

sorb

an

ce

0 5 10 15 20 25 30

Time (minutes)

35 40 45

0

0.3

0.6

0.9

1.2

1.5

1.8

Ab

sorb

an

ce

0

0.3

0.6

0.9

1.2

1.5

1.8

Ab

sorb

an

ce

0 5 10 15 20 25 30

Time (minutes)

35 40 450 5 10 15 20 25 30

Time (minutes)

35 40 45

In order to check this possibility, the experiment was repeated again by adding ahigh concentration of Mg2+ before adding melanoidin. No release of intracellularcompounds in Gram-negative bacteria was observed whereas the behaviour of Gram-positive ones was unaffected and the same amount of intracellular molecules wasmeasured. This corroborates the idea that the sublethal action of melanoidins is linkedto the removal of Mg2+ from the outer membrane of Gram-negatives, while the lethalaction in both

3.2.1. Chelating activity of coffee melanoidins against iron

In order to study the iron chelating properties of coffee melanoidins, differentmelanoidin concentrations were incubated with the culture medium —which showeda mean iron concentration of 10 μg/ml— and the final free iron content was measured(Figure 11). The addition of growing amounts of melanoidins produced a decrease offree iron and an increase of chelated iron, which remained linked to the melanoidin


core and was not available for bacteria. The 50% of iron was chelated at a melanoidinconcentration of 2.5 mg/ml, where the bacteriostatic activity was observed. However,the remaining free iron ranged between a 20 and a 5% for the different strains at theMICs found (4 and 5 mg/ml).

The effect of iron chelation over the E. coli ATCC 35150 growth (in those withno siderophore production) is depicted in figure 12. As shown in the figure, there is alineal relationship between the net area under the curve (Net AUC, calculated bysubtracting the AUC of the blank sample to the AUC of each melanoidin) and thechelated iron, indicating that the decrease in bacterial growth is mediated by ironstarvation. The chelation of iron has little effect up to a melanoidin concentration of1.5 mg/ml, where a 31% of the total iron was chelated. From this concentration, theincrease in the slope of the curve indicates that iron chelation has a deep effect overbacterial growth, suggesting the beginning of the bacteriostatic activity. In this sense,the inflexion point between both curves allows the calculation of the amount ofchelated iron (or melanoidin concentration) where the bacteriostatic activity starts,corresponding in the case of E. coli ATCC 35150 to 3.55 μg of iron chelated or 1.75mg of coffee melanoidin per millilitre of culture medium.

The acquisition of iron is possibly the major determinant as to whether amicroorganism within an animal is able to maintain itself therein (Migo et al., 1997).Without this ability, it would be unable to grow and will effectively be eliminated bydirect attack from the host defence mechanisms or would die of nutrient starvation.Therefore, the acquisition of iron is recognized as one of the key steps in thedevelopment of any pathogen in its host. Siderophores, which are specific Fe3+-binding

Figure 11Effect of coffee melanoidins on the amount of free and chelated iron (total

amount 10 µg/mL) in the culture media (BHI broth)

0

2

4

6

8

10

12

0 0.5 1.0 1.5 2.0 2.5 4.0 5.0

Melanoidin concentration (mg/ml)

Chelated Fe

Free Fe

Fe c

on

cen

tra

tion

(μg/m

L)



agents, are produced by many but not all microorganisms in response to a deficiencyof iron (Miethke and Marahiel, 2007). In the case of a siderophore producer bacteria(E. coli ATCC 33475) a lineal relationship between Net AUC and chelated iron wasfound. At low melanoidin concentrations, iron chelation has little effect (low curveslope) over bacterial growth up to a melanoidin concentration of 2 mg/ml (42% ofiron chelated). This could be related to the strong binding power of siderophores foriron —dissociation constant values ranging from 1022 to 1050 (Drechsel & Winkelmann,1997)— which is regarded as sufficiently strong for the siderophore to remove ironattached to molecules such as ferritin, probably able to remove iron from melanoidins.A second curve with a more pronounced slope was also found, indicating a deep effectof iron starvation over microbial growth. In this case, since no bacteriostatic activitywas observed in the dose-response plot, other different effect involving siderophoresmust be present, which are the final responsible molecules which provide iron to thesemicroorganisms.

A second set of experiments were performed in order to study the effect of coffeemelanoidins over the concentration of siderophores in the culture medium (figure 13).Siderophore production was initiated when cells were grown with a deficiency of ironof a 23%, corresponding to a melanoidin concentration of 1.0 mg/ml. At a

Figure 12Relationship between chelated iron concentration (µg/ml) and the Net-AUC of

E. coli ATCC 35150 (siderophore production)

r2 = 0.9976

r2 = 0.9950

0

50

100

150

200

250

0 2 4 6 8 10

Chelated Fe (μg/ml)

Net

AU

C1

Interval where Fe concentration has little effect over bacterial growth

Interval where Fe concentration has a deep effect over bacterial growth


concentration of 1.5 mg/ml, there was a decrease (although not statistically significant)of the % of free siderophores, while such decrease was statistically significant (P <0.05) from a concentration of 2.0 mg/ml. In addition, with the aid of a high ionicstrength siderophores were removed from the melanoidin core, indicating that theyare chelated by melanoidins at high concentrations by non-covalent interactions. Themaximum chelating activity was obtained at a melanoidin concentration of 5.0 mg/ml where a 97.2% of the siderophores was chelated. These results are in line withthose previously published by other authors where coffee melanoidins showedremarkable binding properties toward low molecular weight substances in the foodmatrix such as furfural or pyrrole derivatives (Hofmann, 1998), chlorogenic acid(Delgado-Andrade & Morales, 2005) or other phenol-like compounds (Bekedam etal., 2007).

Figure 13Effect of coffee melanoidins over siderophores concentration

in the culture medium


0

20

40

60

80

100

0 0.5 1.0 1.5 2.0 2.5 4.0 5.0

% S

idero

ph

ore

s

Chelated siderophores

Free siderophores

ND1 ND1

Melanoidins comprise a substantial proportion, up to a 35%, of several foods suchas baked cereals or roasted coffee and are widely consumed dietary components. Theantimicrobial activity of coffee melanoidins against different bacterial strains has beenstudied, finding that it is mediated by the metal chelating properties of melanoidins.Three different mechanisms have been observed and are reviewed in Figure 14. Atlow concentrations, melanoidins exert a bacteriostatic activity mediated by thechelation of iron from the culture medium, although the sequestering of other essentialcations cannot be ruled out. In addition, for those bacterial strains able to producesiderophores for iron acquisition, the chelation of the siderophore-Fe3+ complex bymelanoidins has been demonstrated, which could decrease the virulence of suchpathogenic bacteria. Finally, coffee melanoidins also exerted a bactericidal activity


at high concentrations by removing Mg2+ cations from the outer membrane, promotingthe disruption of the cell membrane and allowing the release of intracellular molecules.These results reinforce the idea that water-soluble melanoidins may be good candidatesas naturally formed antimicrobial agents in thermally processed foods.

REFERENCES

ADAMS, M.R. & HALL, C.J. (1988) Growth Inhibition of food-borne pathogens by lactic andacetic acids and their mixtures, Int. J. Food Microbiol. 23, 287-292.

AMES, J.M., WYNNE, A., HOFMANN, A., PLOS, S. & GIBSON, G.R. (1999) The effect of a modelmelanoidin mixture on faecal bacterial populations in vitro, British J. Nutr. 82, 489-495.

BEKEDAM, E.K., DE LAAT, M.P.F.C., SCHOLS, H.A., VAN BOEKEL, M.A.J.S. & SMIT, G. (2007)Arabinogalactan proteins are incorporated in negatively charged coffee brew melanoidins.J. Agric. Food Chem. 55, 761-768

BORRELLI, R.C. & FOGLIANO, V. (2005) Bread crust melanoidins as potential prebioticingredients, Mol. Nutr. Food Res. 49, 673 – 678

Figure 14Proposed mechanisms for the antimicrobial activity of coffee melanoidins

INSIDE

( Microbial cell )

OUTSIDE

Melanoidin Mg2+

Mg2+

Mg2+Melanoidin Mg2+

Melanoidin

MembraneDisruption

SiderophoreFe3+

Siderophore Fe3+

Fe3+

Melanoidin Siderophore Fe3+

SiderophoreChelation

IronStarvation

Melanoidin

Melanoidin Fe3+

Fe3+

Fe3+

Fe3+

Fe3+

INSIDE

( Microbial cell )

OUTSIDE

Melanoidin Mg2+

Mg2+

Mg2+Melanoidin Mg2+

Melanoidin

MembraneDisruption

SiderophoreFe3+

Siderophore Fe3+

Fe3+



INSIDE

( Microbial cell )

OUTSIDE

Melanoidin Mg2+

Mg2+

Mg2+Melanoidin Mg2+

Melanoidin

MembraneDisruption

SiderophoreFe3+

Siderophore Fe3+

Fe3+



INSIDE

( Microbial cell )

OUTSIDE

Melanoidin Mg2+Mg2+

Mg2+Mg2+

Mg2+Mg2+Melanoidin Mg2+Mg2+

Melanoidin

MembraneDisruption

SiderophoreFe3+Fe3+

Siderophore Fe3+Siderophore Fe3+Fe3+

Fe3+Fe3+

Melanoidin Siderophore Fe3+Melanoidin Siderophore Fe3+Siderophore Fe3+Fe3+


IronStarvation

Melanoidin

Melanoidin Fe3+

Fe3+

Fe3+

Fe3+

Fe3+IronStarvation

Melanoidin

Melanoidin Fe3+

Fe3+IronStarvation

Melanoidin

Melanoidin Fe3+Fe3+

Fe3+Fe3+

Fe3+Fe3+

Fe3+Fe3+

Fe3+Fe3+


BORRELLI, R.C., ESPOSITO, F., NAPOLITANO, A., RITIENI, A., FOGLIANO, V. (2004)Characterization of a new potential functional ingredient: coffee silverskin, J. Agric. FoodChem. 52, 1338–1343.

CÄMMERER, B. & KROH, L.W. (1995) Investigation of the influence of the reaction conditionson the elementary composition of melanoidins. Food Chemistry, 53, 55–59.

CASEY, J.T., O’CLEIRIGH, C., WALSH, P.K. & O’SHEA, D.G. (2004). Development of a robustmicrotiter plate-based assay method for assessment of bioactivity. J. Microbiol. Methods58, 327-334.

DAGLIA, M.; CUZZONI, M.T.; DACARRO, C. Antibacterial activity of coffee. J. Agric. FoodChem. 1994a, 42, 2270-2272.

DAGILA, M.; CUZZONI, M.T.; DACARRO, C. Antibacterial activity of coffee: Relationshipbetween biological activity and chemisal markers. J. Agric. Food Chem. 1994b, 42, 2273-2277.

DAGLIA, M.; PAPETTI, A.; DACARRO, C.; GARRAN, G. Isolation of an antibacterial componentfrom roasted coffee. J. Pharm. Biomed. Anal. 1998, 18, 219-225.

DAVIDSON, P.M., PARISH, M.E. 1989, Methods for testing the efficacy of food antimicrobials,Food Technol. 43, 148-155.

DEIGHTON, M.A. & BALKAU, B. (1990). Adherence measured by microtiter assay as a virulencemarker for Staphylococcus epidermidis infections. J. Clin. Microbiol. 28, 2442-2447.

DEL CASTILLO, M.D.; FERRIGNO, A.; ACAMPA, I.; BORRELLI, R.C.; OLANO, A.; MARTÍNEZ-RODRÍGUEZ, A.; FOGLIANO, V. In vitro release of angiotensin-converting enzyme inhibitors,peroxyl-radical scavengers and antibacterial compounds by enzymatic hydrolysis ofglycated gluten, J. Cer. Sci., 2007, 45, 327-334.

DELGADO-ANDRADE, C., RUFIÁN-HENARES, J.A. & MORALES, F.J. (2005) Assessing theantioxidant activity of melanoidins from coffee brews by different antioxidant methods,J. Agric. Food Chem. 53, 7832-7836.

DELGADO-ANDRADE, C.; MORALES, F. J. Unraveling the contribution of melanoidins to theantioxidant activity of coffee brews. J. Agric. Food Chem. 2005, 53, 1403-1407.

DRECHSEL, H.; WINKELMANN, G. Iron chelation and siderophores. In: Winkelmann, G.; Carrano,C.J. eds., Transition Metals in Microbial Metabolism (Amsterdam), 1997, pp. 1–9.

EINARSSON, H. The effect pH and temperature on the antibacterial effect of Maillard reactionproducts, Lebens. Wissen. Technol. 1987, 20, 51-55

EINARSSON, H.; SNIGG, B.G.; ERIKSSON, C. Inhibition of bacterial growth by Maillard reactionproducts, J Agric Food Chem. 1983, 31, 1043-1047.

FAIST, V. & ERBERSDOBLER, H.F. (2001) Metabolic transit and in vivo effects of melanoidinsand precursors compounds deriving from the Maillard reaction, Ann. Nutr. Metab. 45, 1-12

FINOT, P.A. & MAGNENAT, E. (1981) Metabolitic transit of early and advanced Maillardproducts. In: Eriksson, C. (Ed.), Maillard Reactions in Food. (Pergamon Press, Oxford),pp. 193-207.

GOMYO T.; HORIKOSHI M. On the interaction of melanoidin with metallic ions, Agric.Biol.Chem. 1976, 40, 33–40.

HASHIBA, H. Oxidative browning of Amadori compounds. Color formation by iron withMaillard reaction products, In: Fujimaki, M.; Kato, H.; Namiki, M. eds., Amino-carbonylreactions in food and biological systems (Amsterdan), 1986, pp. 155–164.

HOFMANN T. Studies on the relationship between molecular weight and the colour potency offractions obtained by thermal treatment of glucose/amino acid and glucose/protein solutionsby using ultracentrifugation and colour dilution techniques. J. Agric. Food Chem. 1998,46, 3891–3895.


HOMMA, S.; MURATA, M. Characterization of metal-chelating compounds in instant coffee. In16th International Scientific Colloquium on Coffee (Kyoto); ASIC, Paris, France, 1995;pp. 183-191.

IBRAIM, H.R., SUGIMOTO, Y.; AOKI, T. Ovotransferrin antimicrobial peptide (OTAP-92) killsbacteria through a membrane damage mechanism. Biochim. Biophys. Acta. 2000, 1523,196-205.

JE, J.Y., KIM, S.K. Chitosan derivatives killed bacteria by disrupting the outer and innermembrane. J. Agric. Food Chem. 2006, 54, 6629-6633.

JEMMALI M.J. Influence of Maillard reaction products on some bacteria of intestinal flora. J.Appl. Bacteriol. 1969, 32, 151-154.

JONES, R.N. & DUDLEY, M.N. (1997). Microbiologic and pharmacodynamic principles appliedto the antimicrobial susceptibility testing of Ampicillin/Sublactam: análisis of thecorrelation between in vitro test results and clinical response. Diagn. Microbiol. Infect.Dis. 28, 5-18.

LANCIOTTI, R.; ANESE, M.; SINIGAGLIA, M.; SEVERINI, C.: MASSINI, R. Effects of heatedglucose-fructose-glutamic acid solutions on the growth of Bacillus stearothermophilus,Lebensm. Wiss. U. Technol. 1999, 32, 223-230.

LOPEZ-GARCÍA, B., VEYNAT, A., PEREZ-PAYA, E., GONZALEZ-CANDELAS, L. & MARCOS, J.F.(2003). Comparison of the activity and antifungal hexapeptides and the fungicidesthiabendazole and imazalil against postharvest fungal pathogens. Int. J. Food Microbiol.89, 163-170.

MATTILA, T. (1987) A modified Kelsey-Sykes method for testing disinfectants with 2,3,5-triphenyltetrazolium chloride reduction at an indicator of bacterial growth, J. Appl.Bacteriol., 62, 551-554.

MATTILA, T.; SANDHOLM, M. Antibacterial effect of the glucose oxidase-glucose system onfood-poisoning organisms, Int. J. Food Microbiol. 1989, 8, 165-174.

MATTILASANDHOLM, M. (1989) Antibacterial effect of the glucose oxidase-glucose system onfood-poisoning organisms, Int. J. Food Microbiol., 8, 165-174.

MIETHKE, M; MARAHIEL, M.A. Siderophore-Based Iron Acquisition and Pathogen Control.Microbiol. Mol. Biol. Rev. 2007, 71, 413–451.

MIGO, V.P.; DEL ROSARIO, E.J.; MATSUMURA, M. Flocculation of melanoidins induced byinorganic ions. J. Ferment. Bioeng. 1997, 83, 287–291.

MORALES F.J.; FERNANDEZ-FRAGUAS C.; JIMÉNEZ-PÉREZ S. Iron binding ability of melanoidinsfrom food and model systems, Food Chem. 2005, 90, 821-827.

MORALES, F.J. Application of capillary zone electrophoresis to the study of food and food-model melanoidins. Food Chemistry, 2002, 76, 363–369

NAYAK, R., KHAN, S.A., WATSON, R.H. & CERNIGILA, C.E. (2002). Influence of growth me-dia on Vancomycin resistence Enterococcus isolates and correlation with resistance genedeterminants. FEMS Microbiol. Lett. 214, 159-173

NIKAIDO, H. Prevention of drug access to bacterial targets: Permeability barrier and activeefflux. Science, 1994, 364, 282-388.

NUNES, F.M. & COIMBRA, M.A. (2007) Melanoidins from coffee infusions. Fraction, chemicalcharacterization, and effect of the degree of roast. J. Agric. Food Chem., 55, 3967-3977.

PILIOURAS, P., ULETT, G.C., ASHHURST-SMITH, C., HIRST, R.G. & NORTON, R.E. (2002). Acomparison of antibiotic susceptibility testing methods for cotrimoxazole with Burkholderiapseudomullei. Int. J. Antimicrobiol. Methods, 1794, 1-8.

RUFIÁN-HENARES J.A.; MORALES, F.J. Microtiter plate-based assay for screening the specificantimicrobial activity of melanoidins. Food Chem. 2008a, 111, 1069-1074.


RUFIÁN-HENARES J.A.; MORALES, F.J. Antimicrobial activity of melanoidins againstEscherichia coli is mediated by a membrane damage mechanism, J. Agric. Food Chem.2008b, 56, 2357-2362.

RUFIÁN-HENARES, J.A. & MORALES, F.J. (2007) Angiotensin-I converting enzyme inhibitoryactivity of coffee melanoidins, Journal of Agricultural and Food Chemistry, 55, 1480-1485.

RUFIAN-HENARES, J.A. & MORALES, F.J. (2007) Antimicrobial activity of melanoidins. J. FoodQuality., 30, 160-168.

RUFIÁN-HENARES, J.A.; MORALES, F.J. A new application of a commercial microtiter plate-based assay for assessing the antimicrobial activity of Maillard reaction products, FoodRes. Int. 2006, 39, 33–39.

SKYTTÄ, E. & MATTILASANDHOLM, T. (1991) A quantitative method for assessing bacteriocinsand other food antimicrobials by automated turbidometry, J. Microbiol. Meth. 14, 77— 88.

STECCHINI, M.L.; GIAVEDONI, P.; SARAIS, I.; LERICI., C.R. Effect of Maillard reaction productson the growth of selected food-poisoning micro-organism. Lett. Appl. Microl. 1991, 13,93-96.

SWENSON, J.A., HILL, B.C. & THRONSBERRY, C. (1989). Problems with the disk diffusion assayfor the detection of Vancomycin resistance in Enterococci. J. Clin. Microbiol. 27, 2140-2142.

VAARA, M. Agents that increase the permeability of the outer membrane. Microbiol. Rev. 1992,56, 395-411.

WANANDY, S., BROUWER, N., LIU, Q., MAHON, A., CORK, S., KARUSO, P., VEMULPAD, S. &JAIME., J. (2004). Optimisation of the fluorescein diacetate antibacterial assay. J. Microbiol.Methods, in press.


Practical assay 5Measurement of antioxidant activity

(general procedure)

CHEMICALS

2,2-diphenyl-1-picrylhydrazyl (DPPH), 2,2´-azobis-(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS), 2,4,6-tri(2-pyridyl)-s-triazine (TPTZ), potassium persulphate,methanol and FeCl3

.H20. In addition, all the chemicals necessary to perform the invitro gastrointestinal digestion as described in the respective chapter.

SAMPLES




Three different approaches are tested: antiradical activity in aqueous (ABTS) andmethanolic medium (DPPH) and reducing capability (FRAP). The samples measuredare the extracts obtained for HMF and furfural determination and the extracts obtainedafter the enzymatic digestion.

(1) DPPH assay. The antiradical activity of different samples in methanolicmedium is estimated according to the procedure reported by Morales and Jiménez-Pérez (2001). A 300 μl aliquot of sample is added to 3 ml of DPPH· 74 mg/l inmethanol. If the sample produces some precipitate, it is centrifuged beforese the


absorbance measurement. A daily-prepared solution of DPPH· gave a final absorptionat 520 nm of 1.8 AU. The mixture was stored in the dark for 1h, and then absorptionis measured at 520 nm. Temperature in the measurement chamber is set at 30°C.Aqueous solutions of trolox at various concentrations are used for calibration (0.15 -1.15 mM). The results are expressed μmol equivalents of trolox per g of sample.

(2) ABTS+· assay. The antioxidant capacity is estimated in terms of radicalscavenging activity in an aqueous medium, following the procedure described byJiménez-Escrig et al. (2003). Briefly, ABTS+· is produced by reacting 7 mM ABTSstock solution with 2.45 mM potassium persulphate and allowing the mixture to standin the dark at room temperature during 12-16 h before use. The ABTS+·solution (stablefor two days) is diluted with 5 mM phosphate buffered saline (pH 7.4) to an absorbanceof 0.70 ± 0.02 at 730 nm. After addition of 25 μl of sample (previously diluted 1/10)to 1 ml of diluted ABTS+·solution, absorbance reading is taken at 20 min. Calibrationis performed, as described previously, with trolox stock solution. Results are expressedas μmol equivalents of trolox per g of sample.

(3) FRAP assay. The ferric reducing ability of each sample is estimated accordingto the procedure described by Benzie and Strain (1996). Briefly, 1 ml of FRAP reagent,prepared freshly and warmed at 37°C, is mixed with 150 μl of sample or water asappropriate reagent blank. The FRAP reagent contains 2.5 ml of a 10 mM TPTZsolution in 40 mM HCl plus 2.5 ml of 20 mM FeCl3

.H20 and 25 ml of 0.3 M acetatebuffer, pH 3.6. Readings at the absorption maximum (595 nm) are taken after 30minutes of incubation at 37°C in the dark. Trolox stock solutions are used to performthe calibration curve. Results are also expressed as μmol equivalents of trolox per gof sample.


RESULTS

Data from antioxidant activity in pre-baked breads

Sample Heating time DPPH ABTS FRAP0

12

20

26

Wheat bread

0

12

20

26

Wheatbran bread


Data from antioxidant activity in the soluble fraction after the in vitrogastrointestinal digestion

SampleHeating

timeDPPH ABTS FRAP

0

12

20

26

Wheat bread

0

12

20

26

Wheat brand bread

LITERATURE

MORALES, F.J.; JIMÉNEZ-PÉREZ, S. Free radical scavenging capacity of Maillard reactionproducts as related to color and fluorescence. Food Chem. 2001, 72, 119-125.

JIMÉNEZ-ESCRIG, A.; DRAGSTED, L.O.; DANESHVAR, B.; PULIDO, R.; SAURA-CALIXTO, F. Invitro antioxidant activities of edible artichoke and effect of intake on biomarkers ofantioxidant status in rats. J. Agric. Food Chem. 2003, 51, 540-545.

BENZIE, I.F.F.; STRAIN, J.J. The ferric reducing ability of plasma (FRAP) as a measure of“antioxidant power”: The FRAP assay. Anal. Biochem. 1996, 239, 70-76.


Practical assay 6Measurement of the antioxidant activity

of insoluble material

CHEMICALS

2,2´-azobis-(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS), potassiumpersulphate and microcrystalline cellulose. In addition, all the chemicals necessaryto perform the in vitro gastrointestinal digestion as described in the respective chapter.

SAMPLES




The procedure assayed is based on the method of Serpen et al. (2008) and allowthe measurement of the antioxidant activity of insoluble and soluble substances.

ABTS Method. Ten mg of ground sample are transferred to a centrifuge tube. Thereaction is started by adding 6 ml of ABTS reagent previously prepared by reacting a7 mmol/l aqueous solution of ABTS with 2.45 mmol/l potassium persulfate and furtherdissolved in a mixture of ethanol:water (50:50, v/v). The tube is placed in an orbitalshaker and the mixture is rigorously shaken until centrifugation to facilitate a surfacereaction between the solid particles and the ABTS reagent. After centrifugation at9200g for 2 min, optically clear supernatant is separated and absorbance measurements


are performed at 734 nm after exactly 30 min. The antioxidant activity is expressedas millimole of Trolox equivalent antioxidant capacity (TEAC) per kilogram sampleby means of a dose–response curve for Trolox. The solid-state samples can be dilutedprior to measurement if measured absorbance values are outside the linear responserange of the radical discoloration solution. For insoluble fractions, dilution isperformed with cellulose powder, which is inert toward both ABTS and DPPHreagents. Dilution with cellulose allows weighing 10 mg per sample, thus ensuringgood reproducibility also for high antioxidant materials. For soluble fractions, dilutionis performed with distilled water.

RESULTS

Data from antioxidant activity in pre-baked breads

Sample Heating time ABTS0

12

20

26

Wheat bread

0

12

20

26

Wheatbran bread


Data from antioxidant activity in the insoluble fraction after the in vitrogastrointestinal digestion

LITERATURE

SERPEN, A., GÖKMEN, V., PELLEGRINI, N. FOGLIANO, V. (2008). Direct measurement of thetotal antioxidant capacity of cereal products. Journal of Cereal Science, 48, 816-820.

SampleHeating

timeABTS

0

12

20

26

Wheat bread

0

12

20

26

Wheat brand bread

AS

SE

SS

I NG

TH

E G

EN

ER

AT

I ON

AN

D B

I OA

CT

I VI T

Y O

F N

EO

-FO

RM

ED

CO

MP

OU

ND

S I

N T

HE

RM

AL

LY

TR

EA

TE

D F

OO

DS

a t r i oe d i t o r i a l

port. cost 2009 12/3/09, 13:551

Assessing the Generation and Bioactivity of neo-...

Documents

Transcript of Assessing the Generation and Bioactivity of neo-...