Non-standard amino acids in peptide design and protein engineering Padmanabhan Balaram

Indian Inst i tute of Science, Bangalore, India

The introduction of non-coded amino acids with well defined stereochemi- cal and functional properties will greatly enhance the scope of protein design and engineering. The present state of methodologies for incorpora- tion of non-coded residues into proteins is examined. The prospects for conformationally constrained amino acid residues are evaluated in the light of peptide structural studies. Templates for secondary-structure nucleation and recent experiences in the incorporation of novel residues into proteins

are considered.

Current Opinion in Structural Biology 1992, 2:845-851

Introduction

Protein engineering using the standard 20 genetically c(xled amino acids is necessarily restricted by the limited repertoire of residues. Although remarkable structural and functional diversity" can be generated by variations in polypeptide primary ,~quences, the incorporation of non-standard amino acids with well defmed stereochemi- c'al and functional properties would greatly enhance the ,scope of protein engineering. The rich literature on the use of an enonnously diverse range of unusual residues in designing analogs of biologically active peptides is testimony to the potential usefulness of incorporating side-chain and backtxmc m(xlifications as a means of modulating the stability, and activi W of peptide sequences [1,2].

In the arc~a of protein engineering, two immediate goals appear desirable. First, unusual amino acids with de- fined contormational properties could be used to imlx~se local restrictions on polypeptide chain stereochemistry, thereby conferring stability, on specific regions of sec- ondary structure. In d e n o v o design approaches, novel residues and templates may be employed to nucleate and stabilize specific secondary structures as a prelude to controlled tertiary organization [3]. Second, residues with versatile chemical functions in their side chains can be used to impart novel binding and catalytic properties. Two requirements critical to the further development of protein engineering are: the development of techniques to incorlx)rate unusual residues into proteins by genetic meth(xts [4,5oo], total chemic~d sTnthesis [6.,7 oo] or by semisynthesis using enz2,,'me mediated ligation [8°,9]; and the development of a palette of 'designer residues' which will permit definitive structural control in the process of

assembling sequences with predictable folding proper- ties.

Methodologies for unusual residue incorporation

Total chemical sTnthesis using highly developed solid- phase approaches is undoubtedly the most direct way of introducing novel residues into proteins. However, this method is restricted to relatively short pol~q3eptide chains of ~ 100 residues or less in length. The present level of technology is exemplified by the recent sTnthesis of the ~)-residue sequence of human immun(x.teficiencT ~rus (HIV)-I protc~,~se, which incorporates all amino acids of thc unnatural I) configuration 16"1. A particularly dra matic example of the power of chemical approaches is the intrcxtuction of a thioester bond into 1 llV-1 protease b.v chemical ligation of the 1-51 fragment (i.e. residues 1-51 ), which incorporates a carboxy-terminal thiol nucle- ophilc, with the 52-99 fragment, which features an alkyl bromide at its amino temainus [7"']. Solid-phase methcxt- olo~" is, however, less attractive because of low coupling yields in the g'nthesis of segments containing stereo- chemic'ally hindered residues, such as a-aminoisobutyric acid. In such situations, enzymatic methcxts hoM greater promise, as illustrated by recent applications to short- peptide antibiotics of the alamethicin family [8"]. Site- specific incorporation of non-coded residues ma,v also be achieved by semi-sTnthesis, where the chemical s~aathesis of a short segment is followed by protease-mediated lig- ation to a longer polypeptide in organic-aqueous solvent mixtures [8"]. This approach is most suitable for modifi- cations at amino- or carbox3'-terminal segments. The reg-

Abbreviations Acnc--l-amino-cycloa]kane-l-carboxylic acid; Aib~Qc-aminoisobutyric acid; HIV

Pip~pipecolic acid. human immunodeficiency virus; Lac.--lactic acid;

c (~) Current Biology Ltd ISSN 0959-440X 845

846 Proteins

ular and routine application of non-standard residues in protein engineering will, however, become a reality ord~ hen genetic methods of incorporation become readily accessible. In early work, non-coded amino acids which are largely isosteric with one of the coded residues have been incorporated into proteins following the misacy lation of tRNA by the appropriate aminoacy-tRNA syn- thetase [lo].

The substrate specificity of the tRNA synthetases is strict, however, and only structurally very similar residues can be introduced in this way. Site-specific incorporation is also generally not possible, with the exception of cases where the replaced residue occurs at a lone position in the sequence. For example, L-2aminohexanoic acid (or norleucine) has been substituted for methionine at posi- tion 21 in human epidermal growth factor [ 111. Another intriguing report has suggested that the novel residue fu- ranomycin (1) [Fig. 11, which is structurally quite distinct from isoleucine, is charged onto tRNA1le by the isoleucy- tRNA synthetase and, indeed, incorporated into a j!-lac- tamase precursor in an in vitro biosynthetic experiment [=I. Recent approaches which appear to point towards gen- erally applicable procedures involve the use of semi- synthetic aminoacylated suppressor tRNAs [ 139•,14*,15*] and the incorporation of an appropriate translatable am- ber codon at the site chosen for mutagenesis [4,5**,14*]. Interestingly, a biochemical precedent for such a strategy does exist. The biosynthetic incorporation of selenocys- teine into E. coli formate dehydrogenase is directed by a UGA codon and utilizes selenocysteinyl-tRNA and a unique translation factor [16]. Efficient procedures for the chemical acylation of suppressor tRNAs by any amino acid analog are a prime requirement for the success of this method.

Elaborate protection and deprotection strategies and low yields are major limitations in the use of tRN4.s acylated with non-coded residues. Procedures developed recently for chemical aminoacylation of tRNAs using a photola- bile protecting group (nitroveratryl) have allowed prepa- ration of the aminoacyl tRNAs in high yields [13-l. Chemical strategies may, however, be intrinsically lim- ited in scope because of synthetic difficulties. The high specificity of aminoacyltRNA synthetases precludes their use at present, although an engineered, non-discriminat- ing tRNA synthetase presents an attractive possibility for the future.

Chemoenzymatic procedures that permit acylation of the terminal 2’,3’-diol of the suppressor tRNA with an amino acid ester using a non-selective lipase constitute another potential strategy [ 171. Initial reports have suggested that the suppressor tRNA route can be used to incorporate a wide variety of amino acids into T4 lysozyrne (see below) [5-l.

Conformationally constrained residues

Stereochemically constrained non-coded residues may have great potential in protein engineering studies once

YN\C<COOH

c

'\ n2N>fooH 0 COOH

“H#!

HJC ‘R N H

4 5 6

YN,/OO+' HO, ,CCOH

AH

YN-CCH$,,-COOH ,A,

7 8 9

10 11

"PHx>,+ "zNHxC""

W 12 b

13

Fig. 1. Structures of some representative non-coded amino acids and templates discussed in the text. (1) Furanomycin; (2) a- aminoisobutyric acid; (3) qwdi-n-alkylglycines (n = 1, diethyl- glycine; n = 2, dipropylglycine; (4) l-aminocycloalkane-l-car- boxylic acids; (5) cr-methyl-a-amino acid CR groups can be identi- cal to those found in the coded amino acids); (6) pipecolic acid; (7) a#-dehydroamino acids CR = H, dehydroalanine; R = phenyl, Z-a,P-dehydrophenylalanine); (8) w-amino acids (n = 2, p-ala- nine; n = 3, y-aminobutyric acid; n = 5 &-aminocaproic acid); (9) u-hydroxy acid; (10) helical template; (11) template for P-sheets; (12) 2-amino-4methylhexanoic acid; (13) amino-3-cyclopentyl- propanoic acid.

the problem of incorporation has been overcome. The a+dialkylated glycines constitute the most widely stud- ied class of amino acids that introduce backbone rigid- ity. a-Aminoisobutyric acid (2) [Aib; also known as 2- methylalanine] (Fig. 1) is the most extensively investi- gated residue of this group because of its widespread occurrence in voltage-gated membrane-channel-forming polypeptides of fungal origin [l&19]. The presence of an additional methyl group at the C” atom in Aib greatly restricts the sterically allowed region of confor- mational space because of van der Waals clashes. Ener- getically favourable minima are limited to helical regions (~z:660~20”and$~ f 30 f 20”) [20]. Experimen- tally, the Aib residue has been found to be strongly helicogenic and to stabilize 310- and a-helical confor- mations in a very large number of synthetic peptides, a feature which has been established conclusively by crystallographic studies (Fig. 2) [21,22*,23].

Non-standard amino acids in peptide design and protein engineering Baiaram 847

-~80 °

x

x

.~80 °

V'

t o

--~x ÷~80 °

- i e 0 o

, . . J

Fig. 2. Crystallographically observed ~,~1/values for at-aminoiso- butyric acid (Aib) residues in peptide structures. 305 Aib residues from 108 independent crystal structures are represented [24]. In the case of achiral peptides crystallizing in centrosymmetric space groups, the choice ol ~ the sign of the dihedral angles is ar- bitrary. The sterically allowed regions of the Ramachandran map for an t-alanine residue are marked (these are also valid for all coded residues except glycine and proline). Note the extremely intense clustering of residues in right- and left-handed helical (3m/0t) conformations. Residues lying in non-helical regions are almost always observed at the carboxyl terminus of very short peptides or in cyclic systems.

The use of this residue in a de n o v o approach to the design of helical super-secondary structural motifs has been explored [24,25"]. In this 'Meccano (Leg()) Set' ap- proach, prefabricated conformationally rigid helical frag- ments containing strategically positioned Aib residues are covalently connected by flexible linking segments. The crystallographic characterization of a linked helix structure, employing c-aminocaproic acid as a connect- ing element between two helical heptapeptide modules, each of which is stabilized by a single centrally placed Mb residue, is a good illustration of this strategy [26"]. The achiral Aib residue can adopt l-x)th right- (atR) and left- handed (atl) helical conformations, depending on the context of the sequence. In short oligopeptides, Aib residues at the penultimate position (position n - 1 in an n-residue sequence) adopt % conformations which result in the formation of a n-turn involving a 6--+1 hydrogen bond, leading to helix terminati(m (IL Karle el a/., unpublished data). This feature is reminiscent of helix-tem~ination signals in proteins, which generally involve glycine or, less often, asparagine residues in ot L conformations [27]. A pronounced feanare of the large number of studies on model Aib peptides is the high de- gree of local stereochemical rigidity that accompanies the incorporation of this residue into peptkle chains. Ks a consequence, relatively long oligopeptides (up to 16-20 residues in length) crystallize readily and there is generally excellent agreement between the solid-state

and solution conformations. In contrast, oligopeptides comprising the 20 coded amino acids only are highly conformationally flexible. Most biologically active pep- tides (such as bradykinin, angiotensin, endorphins and adrenocorticotropic hormone) have resisted efforts to crystallize them over many years. Definitive conforma- tional information can be deduced only in specific cases where complexes with macromolecules are crystallizable, as illustrated by a recent structure of an angiotensin com- plex with an anti-idiotypic antibody [28].

The single reported attempt at introducing Aib into pro- teins was the replacement of Ala82 in T4 lysozyme by Schultz and colleagues [5"] . In the wild-tytye enzyme, Ma82 has qb,~ values of - 6 7 ° and - 2 4 °. The Ma82Mb mutant was shown to be marginally more stable ( ~ 1 °C) in thermal denaturation experiments. Mthough the re- suhs are undramatic, this study" clearly emphasizes the feasibility of intr(x:lucing unusual residues into proteins (the terms 'unusual' or 'non-coded' are preferable to the term 'unnatural'). In retrospect, it appears that the choice of Ma82, which is a surface residue and not part of a regular helix, may have contributed to the small ob- served stabiliTation. Residues in surface loops may be less important in determining the temperature at which ther- mal unfolding commences. In general, the introduction of Aib at sites on protein helices might stabilize local secondary structure, with a consequent effect on over- all melting temperature. It is also conceivable that refold- ing processes might be facilitated by the more etl~cient nucleation of secondary structures.

Other residues that might impose conformational re- straints and which have been well characterized in pep- tides arc shown in Fig. 1. The achiral Aib homologs, at,0t-dialkylated glycines (3) , have been shown to favour flflly extended (~ ~ ~ 180 °) confonnations in homo- oligopeptides, for the cases of diethyl and dipropyl- glyvines [29"]. Nevertheless, this stereochemical prc'f- erence is not absolute. Energy' minima have been de- fined theoretically in the region of helical conformations (qb ~ :1:50 °, ~/~, 4-50°). A recent crystal structure of a fully p r()tected h()mt)-t ri txeptide ( )f at, ¢t-di-n- p r()pylglycine did indeed reveal a gl>e III [3-turn conformation [30].

The 1-amino-cycloalkane-l-carbo~,lic acids (4) [Acnc, where n is the number of carbon atoms in the W- cloalkane ring] are limited almost exclusively to heli- cal conformations in peptides [31]. ACnC residues x~4th n = 3-6 have been incorporated at position 82 of "1"4 lysozyme, but no details of mutant enzyme stabilities were reported [5"]. The high yield preparation of chiral at-methyl-:t-amino acids (5) by chen~oenzymatic proce- dures [32] promises to make available conformationally restricted analogs of many of the cox.ted amino acids. This, in turn, should provide great scope for the im- position of backbone conformational restraints without major side chain alterations. The introduction of such residues would only Jinx)lye replacement of the C ~ hydro- gen with a methyl group. Such mutations should, in gen- eral, stabilize helical conformations at the site of replace- ment.

Pipecolic acid (6) [Pip], a proline homolog, has also been shown to be a useful local conformational determi

848 Proteins

nant in small peptides. This residue is conformationally distinct from proline because the carboxylic acid occu- pies an axial I-x)sition in the cyclohexane ring. Wher~.~as Lproline residues prefer conformations in the helical (qb ,~ - 6 0 °, ~ ~ - 3 0 °) and semi-extended, pol~l")roline- like (qb ~ - 6 0 °, ~ ~ 120 °) regions of q~,~/ space, the(}- retical studies have indicated that low positive values of t~ may be preferred for the Pip residue [33]. Indeed, the incorporation of a Pip residue at position 82 of T4 ly- sozyme resulted in a lowering of the melting temper- ature by ~2°C, whereas the Ala82Pro mutant is more stable by 2°C [5"]. Although Schultz and coworkers attributed this to a larger q5 angle in Pip, peptide co's- tal s!oructures [34] suggest that ~l,ip is indeed close to - 6 0 , a value similar to that for AlaS2 in the wild-t~t)e enzyme. The observed destabilization may indeed be a consc~tuence of changes in the ~ value at the Pip residue.

0q~-Dehydro residues ('7) also offer a means of restrict- ing backbone coiR'ormations, l)ehydroalanine residues have indeed been found in enzymes, albeit very infre- quently, and are formed by the post translational dehy- dration of serine residues [35]. Recent studies of model peptides suggest that ~-turns are often stabilized in short scNuences, whereas helix formation may be promoted in longer peptides, t towever, the available experimental data are limited, with most of the studies that have been reportt.x.I featuring relatively short lxeptides containing Z- dehydrophenylalanine [36], whereas veo' few fc~ltured investigations of dehydroalanine [37]. Interestingly, the incorporation of dehydroalanine into proteins may be possible by chemical modification of charged tRNA set or tR~NA(:.vs.

ever, the hydroxyl residue was at a t-x)sition uninvolved in intramolecular hydrogen bonding. The intr(x.tuction of ester bonds into proteins may eventually provide an ex- cellent chemical handle for controlled site-specific cleav- age, a feature which may be useful in the processing of expressed fusion proteins to x4eld desired products.

t)-Residues provide an attractive means of stabilizing con- fomlations with tx}sitive q~ values, which are generally adopted in proteins by glycine and, to a lesser extent, by asparagine [27]. Such conformations (al) occur in [3-turns and are critical in detennining folding patterns. Although D-residues are used extensively in the design of analogs of biologically active peptide, particularly with a ~iew to enhancing proteolytic stability, few g,stematic investigations of their influence on regular secondary structures have been reported. A study of a model 29- residue peptide with a centrally placed [>alanine residue revealed a helix destabilization of 0.95 kcal mol - 1, com- pared with the parent L-alanine peptide I40]. Although two recent reports have described the chemical g,nthe- sis of all D analogs of HIV 1 prote,ase [6"J and rubredoxin [41"], the stereochemical conseqt,ences of the site-spe- citic introduction of D-residues into proteins have not been explored. Attempts to incorporate D-alanine into [3- lacmmase [4], "1'4 lysozy~lle [ 5"] and model 16-residue sc~quences [14"] by the suppressor tRNA route proved unsuccessful. This failure m W be a consequence of dis- crimination at the lm'el of the formation of the complex be~'een aminoacT1 tRNA and elongation factor EF-Tu [5"] , illustrating the nature of the difficulties that may be encountered in the biosynthetic intr(x]uction of non- coded residues.

Flexible residues and conformational variability

m-kallino acids [glycine homologs] (8) , which incorpo- rate additional methylene groups between the amino and carboxyl groups (e.g. ]3-alanine, y-aminobutyric acid and ~;-aminocaproic acid), might be expected to enhance confonnational flexibility. Relatively little information is available on their conformational influences in pep- tides 138]. It is noteworthy that an attempt to intro- duce [3-alanine at l'x)sition 82 of "I'4 lysozy'nle by the suppressor tRNA route was not particularly successful, with less than 5% suppression being observed [5"J. a-Hydroxy acids (9) , e.g. lactic acid (lac) and phenyl- lactic acid, have been incortx~rated at position 82 in "I'4 lysozy,me [5"]. The Ala82Iac mutant is thermally desta- bilized by 3.7°C, a rather surprising result `as the AlaS2 NH group is only hydrogen bonded to water in the wild- type enzyme. Hydros ' acid residues may be expected to enhance local conformational freedom because they lack the NIt group fbr intramolecular hydrogen bonding. Fur- ther, differences in the barriers to rotation about the ester and amide groups may also contribute. The incorpora tion of a-hydrox~3sobug,ric acid into proline-containing peptides has been shown to lead to structures an~ogous to the ¢t-aminoisobutyryl peptide [39]. In this case, how-

Templates

A survey of non-standard residues and protein structure would be incomplete without mentioning at least some of the many ingenious templates that have been used to nucleate specific secondaw structures in oligopeptides and to assemble super-secondao' structural motifs in d e

n o v o protein-design approadles. Template 10, which is tbmlally related to a Pro-Pro dit~ptide, has two bonds tixed in pyrrolidine rings at qb values necessary for he- lix fonnation while a -CII2-S- bridge contains a third rotatable bond. This template nucleates or-helical con- tbmlations in oligoq, alanine sequences [42",43"]. In an- other study, diac3'laminoepindoleidiones (11) have been shown to act ,as templates for ~-sheet formation [44]. In these cases, chimeric molect, les were constructed by attaching linear peptide sequences to non-peptidc functions bearing appropriately oriented hydrogen-bond donors and acceptors. The studies ret'x.~rted, however, are conlined to relatively short sequences.

A fi'uitful use of templates is the orientation of heli- cal peptides to generate helix bundles mimicking those found in proteins. The covalent attachment of multiple helices to side-chain functions on a rigid cTclic pep tide base has been examined in an attempt to prepare template-assembled synthetic proteins [45"]. In addition,

Non-standard amino acids in peptide design and protein engineering Balaram 849

metal-ion-promoted oligomerization of helical peptides bearing a terminal pyri~/l or bipyridyl metal-coordinat- ing site has been used to demonstrate self assembly of designed three- and four-helix bundles [40-48"]. The insertion of templates into native protein sequences appcmrs to be technically forbidding at present, yet tem- plates may be central to d e tlot,o design strategies using total chemic'al sTnthesis.

Prospects for proteins incorporating non-coded residues

Data on the purposeful intrcKtuction of non coded residues into proteins in order to address issues of structure and stability., are extremely limited. In the ini- tial study of Schultz's group [4], Phe66 in a [3 lactmnase could be replaced by closely related analog resktues such as pnitrophenylalanine, p-flu()rophenylalanine and homophenylalanine, which differ from phcnylalanine in electr(mk: properties and to some extent in steric bulk. Interestingly, the pnitro and homo-phenylalanine mu- tams appeared to be less stable and were degraded proteolytically during purificatkm [4]. Attempts to modify the backb(me by the introduction of an ester linkage, an addition;d methylene group or a D-phenylalanine residue were unsuccessful [4]. A later stud.,,' of T4 lysozyme demonstrated the successful introduction of an ester link- age (hydroxyacid) but yew little or no incorporation of [3 alanine and D-residues was observed [5"']. This study also established the pt)ssibilit T of incorporating 0t,at-dialk- ylated residues and Pip at i~.)siti()n 82 ()f T4 lysozyme [49"'], Ks discussed alxwe.

A recent report has explored the use of non-¢xMed residues to replace Phe133, a buried residue, in T4 ly- sozTme [.t9"']. Norvaline, ethylglycine, O-methylserine, S,S-2-amino-,t-methyl hexanoic acid (12) and S-2-mnino- 3-cTclopenwl propanoic acid (13) were used to examine the cost of stepwise removal of mcthvl grot, ps, side-chain solvation, the effect of packing densi W and side-chain confimnational entropy on protein stability'. These ex- perinaents were supported by m(Melling and molecular dynamics simulatkms and ~.ielded two interesting 'non- ctmlcd' ca~ity-filling mutants which are stabilized by 1.9 °C and 4.3 *C compared with the v~ild-t3Tm end,me, whk'h has a mehing temperature of 43.5 °C under identical con- ditions [49"']. Clearly, the availability of a structurally diverse set of anaino acids can enhance the scope of protein engineering for stabilization. In particular, amino acids that intrtxtuce confomtational constraints may be usefully employed in stabilizing specific secondaw struc- tural features. The introduction of such residues should also reduce conformational entropy of the unfolded state, so indirectly stabilizing the folded form. At present, Aib and related residues offer a me, ms of reinlbrcing helical and [3-tum backbones. Obligatoo' [3-sheet or extended structure stabilizers are, as yet, less well characterized. Achiral or D-residues may also help in promoting a L conformations, which are generally attained in proteins only by glycine or asparagine residues.

Rational site-directed mutagenesis, however, requires a clear understanding of the structural role of the residue being replaced and an equally clc~ar definition of the con- formational properties of the non-coded replacement. The identification of t-x)tential weak sites (Achilles' heels) in protein structures is a pre-requisite for rationally en- gineering stability. The s3aathetic development of novel anaino acids with a strong preference for well defined re- gions of backbone (qS,~) conformational space would incrc~ase the tools available to the protein engineer. Novel residues that permit covalent disulfide crosslinks over longer distances than cTstine bridges have also l>een develt)ped [50"]. Although enhancing the reper- toire of available residues is a desirable goal, the sub- tie stereochemical requirements of ribosomal sTnthesis may lead to unexpected problems in incorporation. For relatively small proteins ( < 100 residues), total chemical sTnthesis ,nay provide an alternative approach, although solid-phase methodolog5" may falter with increasing stere- ochemical constraints [8"].

Although this review has focussed largely on the effect of non-c(Med residues on backbone ct)nfornmtion, modili- cation of functional side chains m W provide interesting insights into en~'me mechanisms. At present, site-specific changes are restricted to special situations, as illustrated in a recent study of conversion of an active-site arginine in aspartate aminotransferase into homoarginine [51"]. Novel resktues and templates are particularly attractive elements in d e not ,o design, as the}," offer a definitive mc~ans of controlling the fold of the polypeptide back- b( }ne.

Acknowledgements

Work in the area of pcptide design in this laboratory has been supporttu.l by grants from the l.X'pam'nent of Science and Technolog, y. India. 1 thank R Gumnath, S Ag,trwalla and R Gokh,de for helpfl_tl suggestk ms.

References and recommended reading

Paf~ers of particular interest, published within the annual fmriod of re- ~ie',',, have been highlighted as: , of s[xecial interest , , of outstanding interest

1. IIIRS(:ItMA'.;X R: Medicinal Chemistry in the Golden Age of Biology: l.essons from Steroid and Peptide Re.~arch. Angeu' (.7~em 1991. 30:12~8 1301.

2. liRt'm" \ 'J, NIKIFOROV1CH GV: The Ramachandran Plot and Beyond: Conformational and Tolx)graphical Considerations in the Design of Peptidcs and Proteins. In Molecuktr C'on. .fl~rmations and Biological Interactions. F.dited b.v Balaram P, Rama.seshan S B:mgalore: Indian Academy of Sciences; 1 c~) 1 :-+29-44=,.

3. DEGRAIx) ~rF, RAIJ':IGI! l)P. IIANDEI. T: De novo Ih'otein De- sign: What Arc We l.earning? Curr Opin Struct Biol 1991, 1:984- ~)3.

• 1. N()RFN cJ, ANq'IIONW-(~.AIIIU. ~J, GRIFFfl'tt Me, SCltl'LTZ PG: A General Method for Incorlx~ration of t;nnatural Amino Acids into Proteins. Science 1989. 244:182 188.

850 Proteins

5. EILMAN JA, MENDEL D, SCHUL'IZ PG: Site Specific Incorpora- ,)o tion of Novel Backbone Structures into Proteins. Science

19(-)2, 255:197-200. This is the first rel'x)rt of the successful intr(xluction of conformationally constrained, non.c{xted amino acids into a protein using the suppres.'~)r tRNA strategy. Various replacements at position 82 of T4 l.vsoDTne are relx)rted. Replacements include :(.:x-dialkytated residues, N-alkyl amino acids and lactic acid.

6. MII.TON RCDEI. MII:rON SCF, KE>,'T SBtt: Total Chemical Syn. • thesks of a D-Enzyme: the Enantiomers of HIV-I Protease

Show I)emonstrat ion of Reciprocal Chiral Substrate Speci- ficity. ,Science 1992, 256:1445--1448.

The preparation of an all-D analog of a of) residue polylmptide illustrates the pr(_,~nt state-of-the art technology for assembling peptide chains on solid supports, and demonstrates the fea,sibilit T of incorporating non- cox.led residues by chemical sy'nthesis.

7. ,C')CHNOIZER M, KEN'I' SBH: Construct ing Proteins by IM)vetail- . . hag Unprotected Synthetic Peptides: Backbone-engineered

HIV Protease. Science 1992, 256:221-225. The introduction o fa thioester Ixrod between residues GI~%l and Gly52 of the 99-residue protein IIIV.1 prote~a.se is achieved by a chemical strategy invoMng the ligation of two large tYagmenLs prepared by to- tal sy'nthesis. The 1-51 peptide has a carl'x)xT-terminal sulfur nucle. ophile, whereas the amino terminus of the 52-99 fragment bears an alkyl bromide function. Control of stereochemistl 3' is facilitated by the choice of achiral glycine residues at the site of backbone engineering. This approach offers the possibili W of intrcMucing "fixc~d elements of 31)-structure" and the "synthesis of hybrid protein- nonprotein macro molecules".

8. SI.O.MCZYNSKA l_~, BEUSI..'N 1)I), ZABROCKI J, KOCIOLEK K, • REDIJNSKI A, REUSSER F. HU'I'I't)N WC, I£PIAWY MT, MAR.~,HAIJ.

GR: Emerimicins Ill and IV and their Ethylalanine Epimers. Facilitated Chemical-enzymatic Synthesis and a Qualitative Evaluation of their .Solution Structures. J A m C~,rn Soc 1992, 114:4095-4106.

A description of chemoenz}~natic strategnes for the s3althesis of l~p - tides containing a,a-dialM3ated amino acid residues.

9. Roy RP, KIIANDKE KM, MANJI:IA BN, ACHARYA AS: Hclix For- mation in Enzymatically Ligated Peptidcs as a Driving Force for the Synthetic Reaction: Example of cx-Globin .¢,emisTn- thetic Reaction. Bicx'hemL~tr), lof)2, 31:7249--7255.

10. WII.~3N MJ, HATFIEII ~) DI.: Incorlx)ration of Modified Amino Acids into Proteins in vivo. Biochim Biopt~s Acla 1984, 781:205-215.

11. KOII)E II, YOKOYAMA S, KAWM (), 11,~ JM, OKA "I', KAWA/ S, MIYAKA T, I:UX~'A T. Mn'AZAWA "1": Biosynthesis of a Protein Containing a Nonprotein Amino Acid by Escherichda coli: t.-2-Aminohexanoic Acid at Position 21 in Human Epidermal Growth Factor. Proc Natl Acad Sci L'SA 1988, 85:623"45241.

12. KOItNO T, KOHDA D, HARUKI M, YOKOYA.~,b'~ S. MIYAZA~A T: Nonprotein Amino Acid Furanomycin, Unlike Isoleucine in Chemical Structure, Is Charged to isoleucine tRNA by Isoleucyl-tRNA Synthetasc and Incorporated into Pro- te in . . l Biol C&,m l°ft.)(), 265:6931-6935.

13. ROBI.:R'I,'SON SA. ELLMAN JA, SClItnL'IZ PG: A General and ERi- • • cient Route for Chemical Aminoacylation of Transfer RNA.s.

J Am Chem Soc 1991, 113:2722 2729. The authors describe protocols which greatb' simplify the preparation of tRN&s acTlatcxi with unusual amino acid residues. The aminoacyla. tion of unprotected dinucleotide with an Nblocked anfino acid residue, efiqcient ligation to tRNA and the sub~xtuent photcx:hemical unmasking of the amino group arc important ~atures of these methods.

14. BMN JD, ~"ACKER DA, Kt'o EE, CHAMBERHN AR: Site-specific • Incorporation of Non-natural Residues into Peptides: Ef-

fect of Residue Structure on Suppression and Translational Efficiencies. 7k'trahedron 1991, 47:23b~)-2400.

Relative translation efficiencies for a series of 12 semi-synthetic a(Tlated suppres,,,or tRNA.s are reported using a rapid assay procedure for tie remaining suppression and translational el~ciencies.

15. BAIN JD, '~X,'ACKER DA, KUO EE, LYTrI.E MH, CtlAMBERLIN AR: • Preparation of Chemically Misacylated Semisynthetic N o n -

s e n s e Suppressor tRNA.s Employed in Biosynthetic Incor- poration of Non-natural Residues into Proteins. J Org Cbem 1991, 56:4615-4625.

Several .,synthetic routes for acylated dLNAs are described. Direct chemi- cal ..synthesis is presented as an alternative to run-off transcription from a direct DNA template for tRNA synthesis. An t-3-iodotyrosine-charged suppres.~)r tRNA is prepar~.xt by four pr(x't.xlures and an in vitro trans. lation reaction is used to demonstrate similar suppression efficiencies in -,ill cases.

16. BOCK /L FORCIIltAMM£R K, IIEIDER J, BARON C: Selenopro- tein Synthesis: an Expansion of the Genetic Code. Trends Biochem Sci lof)l, 16:463--467.

17. AN"I'HONY-CAHIIJ. SJ, GRIFFITH MC. NOREN CJ, St'ICH DJ, ,(~Clltll'lX PG: Site-specific Mutagenesis with Unnatural Amino Acids. Trend¢ 13iocbem Sci 1989, 14:400-403.

18. AGARWAI.IA S, MEI£OR IR, SKNSOM MSP, KARI.E IL, FLIPPEN- ANDER.~ON JL, IZMA K, KRISHNA K, SUKUMAR M, BAIARAM P: Zervamicins, a Structurally Characterised Peptide Model for Membrane Ion Channels. Biochem Bioply).~ Res Commun lof)2, 186:8 15.

19. KARLF. IL, FLIPPEN-A.NDERY,()N JL. AGARWAIJA S, BAIARAM P: Crys- tal Structure of [Leu I] Ze~'amicin, a Membrane ion-chan- nel Peptide: Implications for Gating Mechanisms. I'roc Nail Acad .~ci USA 1991, 88:5307-5311.

20. PRASAD BVV, 19d.AK,Wt P: "llae Stereochemistry of :(- Aminoisobutyric Acid Containing Peptides. CRC Oi t Rev Biochem 1984. 16:307-347.

21. KARLE IL BAIARAM P: Structural Characteristics of :(-Helical Pepttde Molecules Containing Aib Residues. Bi(x~hern~t~, 10fX), 29:67474756.

22. KARt.E IL Folding, Aggregation and Molecular Recognition • in Peptides. Acta 03,stallogr [B/ 1992, 48:341 356. An ove~iew of structural investigations of txeptides, which contains a summaB" of cI3'stallographic anal,,:ses of ,~weral Aib-containing peptide helices and discusses helix packing in ci3,stals.

23. DIBt.&slo 13. PAVONE V SAVIANO M, IL)MBARI)I A, NaS"l'Va F, PI'I)ONE (.:, IaJENE[)E'I°II 1~., CRISMA M. ANZOIJN M, TONIOLO C: Structural Charactcrization of the [LBend Spiral: ('rys- tallographic Analysis of Two Long (I.-Pro-Aib) n Sequential Pcptides. J Ant Cbem 50:. 1992. 114:6273-6278,

24. KARIa ¢ 11., I:UPPI'.'N-ANI)ERSON .II, IrMA K, BAt,\RAM P: Helix Construct ion Using a-Aminoisobutyryl Residues in a Mod- ular Approach to Synthetic Protein Design. Conformational Properties of an Apolar Dccapcptide in Two Different CtTs- tal Forms and in ,'~)lution. (,'urr .~'ci lof)O, 59:875 885.

2S. Bat.t,R.~t P: The Design and Construct ion o f Synthetic Pro- ]'his tein Mimics. Pure Appl CI)em lof)2, (')4:1061- 1066.

paper Immdes a summary of a design strate~, that employs Aib residues and de.,~:ribes the sTnthesis of linked helix sTstems. The strong helix nucleation prol:x:rties of Aib are demonstrated by examples of the ctTscallographic characteriTation of several long peptides.

26 KARI.E IL, FLIPI'EN-ANI)ER."~)N JL. SUKI MAR M, UMA K, BAIARAM • P: Modular Design of Synthetic Protein Mimics. Crys-

tal Structure of Two Seven-residue Itelical Peptide .Seg- ments Linked by ~.-Aminocaproic Acid. J Am ('./~em .S'(~: 1991, 113:3952-3956.

The authors present an unambiguous conformation;d characteriTation of m'o peptkte helkz's conn(x'tc~ by a t]exible chain linking seg-ment. A single centrally kx'ated Mb residue suffices for helix nucleation in the heptapeptide fragments.

27. RI(:ItARI),"K)N JS, RICHARI)%()N IX~: Principles and Patterns of Protein Confbrmation. In t'rediction c f Protem ConJormation. Edited by Fasman GI). New York: Plenum; 1989:1 98.

28. GARCIA KC, RONCO PM, VERROt'ST PJ, BRINGER AT. AMZEI. LM: Three Dimensional Structure of an Angiotensin lI-Fab Corn-

Non-standard amino acids in peptide design and protein engineering Balaram 851

plex at 3A:" Hormone Recognit ion by an Anti-idiotypic An- tibody. Science 1992, 257:502-507.

29. TON1OLO C, BENEDEq-lq D: The Fully Extended Polypeptide • Conformation. In Molecular Confi~rmation a n d Biological In.

teractions Edited by Balaram P, Ramaseshan S. Bangalore: Indian Academy of Sciences; 1992:511 521.

A brief review of the use of a,~t.dialkylated amino acids in stabiliz. ing fully extended (d~ ~ ~. Ig0 °) conformations in synthetic homo. oligopeptides. The crystal structures of short peptides adopting Cs conformations are illustrated

30. DIBLASIO I3, PAVONE V, ISERNIA C, PI:DONI" C, BENEI)ETTI E, TONIOt.O C, IIARDY PM, IJNGH&'a IN: A Helical IX'~ Homo- peptide. J Cbem Soc Perkin TrarLs 2 1%)2, 523 526.

31. VALI.E (}, CRISMA M, TONIOLO C, St'DHANANI), RAO RB, SIIKUMAR M, BAIARAM P: Stereochemistry of Peptides Containing I- Amimn.Tcloheptane-l-carlx)xylic Acid (At-re). Crystal Struc- tures of Model Peptides. Int J Pept Protein Re• 1991, 38:511 -518.

32. KAMPHUIS J, I:IOES'fEN WIIJ, BROXTERMAN QB, HERMES HFM, VAN BALKEN JAM, MEIJER EM, SCHOE,'aAKI..'~ HE: New Developments in the Chememnzymatic Production of Amino AcidS. Adz, Biex:hern Eng Biotechnol 1990, 42:133 186.

33. TONIOI.O C, P, ARDI R, PIAT_7.ESl AIM, ('RISMA M, P, AtARAM P, StJKUMAR M, PAUL PKC: Preferred Conformations of Pipecolic Acid Containing Peptides. In PtsOtides 1 9 ~ , Proceedings" of the 20th European Peptide .S~ymf~sium. Editcx.l by Jung G, Bayer F.. Berlin.'Ne~' York: WMter d r Gm.,,¢er; 1989:477.-479.

34. BARDI R, PIAZZESI AiM, CRISMA M, TONIOI.O C, SI'KUMAR M, BAIAR~t P: Structure of N-tert-Butyloxycarbonyl-a- aminoisohutyryl-I)L-pipecolyl-,x-aminoi~)butyric Acid Methyl Ester. Acta C~.,stallogr /C/ 1~)2, in press.

35. STA.~aMER CH: Dehydroamino Acids and Peptides. In GSem. L~t~ attd Bi(xchernt~t O, ¢ f Amino Acids, Peptides a n d Pr(> teins. Edited by Weinstein B. New York, q~,asel: Marcel-l)ekker; 1982:33-74.

36. CIAJOLO MR, "I'uzl A, PRA'h'SEI CR, FISSl A, PIERONI O: Crystal and Molecular Structure of the Doubly Unsaturated De- hydropeptide Ac-APhe-AIa-APhe-NHMe. Biol~lymers 19<)2, 32:717-724.

37. PALMER DE, PA'FrARONI C, NUNAMI K, (.~IIAI)HA RK, (R)ODbtAN M, WAKAMIYA T, I:I;KASE K. tlORIMO'IO S, KITATAWA M, EUJITA II, m" At.: Effects of Dehydroalanine on Peptide Conformations. .l Am (Yoem Soc 1992, 114:5634-5642.

38. ToP~Io J, PUK;GAIJ J, VIx~s J, EI'ITL. I. IJOVERAS J, BFIJA J, AYMAMI J, SI;BIRAN^ JA: Crystal Structure of a itelical Oligopcptide Model of Polyglycine II and of Other Polyamides: Acetyl- (glycyl-[~-alanyl)2-NH propyL Bi¢qx)lymers 1~)2, 32:64.~-648.

39. V^HJ-: G, BARI)I R, PL~ZI'.'St AM, CRtSMA M, TONIOI.O "C, ('AV1CCII1ONI (.}, UMA K, BAI.ARAM P: Crystal State Confor- mation of Three Model Monomer Units for the 13-Bend Ribbon Structure. Bi¢~-~)l)vners 1991, 31 : 1669- 1676.

40. FAIRMAN R, ANTHONY-CAHIIJ. SJ, DEGRAt~ WF: The Helix Forming Propensity of D-Alanine in a Right-handed a-ltelix. J Am Ox, m Soc 1992, 114:5458-5459.

41. ZAWADZKI { I.E, BERG JM: A Racemic Protein. . l Am C&,m Soc 19<)2, 114:4002--4003.

] 'he preparation of all.D rubrcMoxin "*'as motivated by a desire to grcr,~' centrosymmetric crystals of a racemic protein, ttowever, the success ful .synthesis suggests that site-specitic incoq'x)ration of D-residues by chemical methcxts should t'x)se no problems in small proteins.

42. KEMP DS, BOYD JG, MUENDEL CC: The Helical S Constant for • Alanine in Water Derived from Template-nucleated Helices.

Nature 1991, 352:451-454. ,See [43"]

43. KEMP I)S, CtTRRAN TP, l~)Y~) JG, ALLEN TJ: Studies • o f N-terminal Templates for a-Helix Formation. Syn-

thesis and Conformational Analysis of Peptide Conju- gates of (2S,5S,8S,11S)-l-Acetyl-l,4-diaza-3-keto-5-carboxy- lO-thiatricyclo[2.8.1.O4,8]-tridecane (Ac.Hell-OH). J Org (',hem 19<)1, 56:6(xq3-6697.

These two papers [42o,43 • ] describe a novel sTnthetic template for nu- cleating peptide helices, which is based on a mcxtification of a Pro-Pro dipeptide.

44. KEMP DS, BOWEN BR, MUENI)EL CC: Synthesis and Conforma- tional Analysis of EpindoUdione Derived Peptides. Models for ~-Sheet Formation. J Org Chem 1990, 55:4650-4657.

45. Mt:'I'I'ER M, TUCIISCIIERER C-G, .MILLER C, AI.TMANN K-H, CAREY • 16, WYSS 1)1:, LABHARIYI AM, RIVIF.R JM: Template-assembled

Synthetic Proteins with Eour Helix Bundle Topol¢~y. Total Chemical Synthesis and Conformational Studies. J Am Cbem Soc V.D2, 114:1463 1470.

A cTclic carrier molecule (.decapeptide) bearing four lysine residues is used as a template for the toy'.dent attachment of four putative heli- cal (16-residue) peptides to the ly.wl t:.amino Aroups. Limited circular- dichroism and NMR data demonstrate the helix.promoting effect of the template bt, t do not unambiguously establish a h)ur.helix bundle tolx)l- ogy.

,t6. GnADtm MR, .~)ARES C, CHOI C: Design of an Artificial • I:our-helix Bundle Metalloprotein via a Novel Ruthenium

(I!) Assisted Self-assembly Process . . I Am Chem Soc 1992, 1 1 4 : 4 0 0 0 ~ i 0 0 2

Sc~" [48"1.

47. IIEBERMAN M, SA.SAKI "r: Iron (il) Organises a Synthetic • Peptide into Three-helix Bundles. J Am Cbem Soc 19<)1,

113:1470-1471. Sc'e 148"1.

,ibm. GHAI)IRI MR, .'~)ARES C, ClIOl C: A Convergent Approach to • Protein Design. Metal Ion-assisted Spontaneous ,Self-assem-

bly of a Polypeptide into a Triple-helix Bundle Protein. J Am Chem Sot: 1992. 114:825-831.

This and the two preceding papers 146o,47 .] address the common theme of using a metal-ccx)rdination site at the terminal ends of helical lxrptides to promote self assembly into c)ligomeric bundles. Detailed NMR or ct2,'stallographic characterizatums of these synthetic 'metallo- proteins' remain to t~e accomplished.

49. MENDEL I), Et.LMA.X JA, CIIANG Z, VEENSTRA 1)I. KOLLMAN P, • • SCut VV..: PG: Probing Protein Stability with Unnatural Amino

Acids. Science 1992, 256:1798 1802. This paper is a fc)mmnner of things to come in the area of protein engineering with non-coded residues. "1'4 b'sozyme is the model and position 133 is probed with a variety of replacements to examine pack ing, ,~flvation and side-chain conformational entropy effects on stability. Two 'non-oxkxl mutants' with marginalb,' higher thermal stabilities are obtained.

50. JACKSON l)g, KIN(; US, CIIMIELEWSKI J, SINt.}H S, ,'¢~C.IIIII.TZ PG: • General Approach to the Synthesis of Short ,,x-Helical Pep-

tides..1 Am Chem S(,,c 1991, 113:9391-9392. A higher homolog of t3'steine featuring the insenion of three additional methylene groups into the side chain, which is derived by chemical syn- thesis from lysine, disulfide bridges the i and i + 7 residues in tx)tentially helical peptides (where i represents residue number from the amino terminus). Disulfide eros•linking enhances peptide helicit3,.

51. WllrI'E PW, KIPs•eli JF: Sequential Site-directed Mutagenesis • and Chemical Modification to Convert the Active Site Argi-

nine 292 of Aspartate Aminotransfera.se to Ilomoarginine. J Am C&,m Sc;c 1~)2, 114:3567-3568.

This homologation of Arg292 is ba.%xt on an initial site-specific muta- tion to lysine followed by chemical guanidination with O-methyli~)urea. Although modification oecurs at several lysine residues, guanidinated en~Tnes often ix)ssess the same activity and are stable.

P Balaram, Molecular Biophysics Unit, Indian Institute of Science, Ban- galore 560 012, India.