100 years of peptide
‘100years of peptide
synthesis’:ligation methods for peptide and protein
synthesis with applications to b -peptide assemblies *
T.Kimmerlin D.Seebach
Authors'affiliations:
T.Kimmerlin and D.Seebach ,Department of Chemistry and Applied Biosciences,Laboratory of Organic Chemistry,Swiss Federal Institute of Technology,ETH-Ho ¨nggerberg,Zu ¨rich,Switzerland.
Correspondence to:Dieter Seebach
Department of Chemistry and Applied Biosciences
Laboratory of Organic Chemistry Swiss Federal Institute of Technology ETH-Ho ¨nggerberg Wolfgang-Pauli-Str.10CH-8093Zu ¨rich Switzerland
E-mail:seebach@org.chem.ethz.ch
Key words:peptide chemistry;chemical ligation;thioligation;b -amino acids;b -peptide
Abstract:A brief survey of the history of peptide chemistry from Theodore Curtius to Emil Fischer to Bruce Merri?eld is ?rst presented.The discovery and development of peptide ligation,i.e.of actual chemical synthesis of proteins are described.In the main chapter,‘Synthesis of Proteins by Chemical Ligation’a detailed discussion of the principles,reactivities and mechanisms involved in the various coupling strategies now applied (ligation,chemical ligation,native chemical ligation)is given.These include coupling sites with cysteine and methionine (as well as the seleno analogs),histidine,glycine and pseudo-prolines,?unrestricted ?amino-acid residues (using the Staudinger reaction),as well as solid-phase segment coupling by thioligation of unprotected peptides.In another section,?Synthesis of b -peptides by Thioligation ?,couplings involving b 2-and b 3-peptides are described (with experimental details).
Dates:
Received 4October 2004Revised 25October 2004Accepted 20November 2004
*Dedicated to Murray Goodman,a pioneer in peptide chemistry.To cite this article:
Kimmerlin,T.&Seebach,D.‘100years of peptide synthesis’:ligation methods for peptide and protein synthesis with applications to b -peptide assemblies *.J.Peptide Res .,2005,65,229–260.DOI 10.1111/j.1399-3011.2005.00214.x Copyright Blackwell Munksgaard,2005
Introduction
Peptides and proteins play a central role in numerous bio-logical and physiological processes in living organisms;they are involved as hormones and neurotransmitters in
intercellular communication,act as antibodies in the im-mune system to protect organisms against foreign invaders,and are also involved in the transport of various substances through biological membranes.Understanding the control,at the molecular level,of the mechanisms and principles governing structural and functional properties of bioactive proteins is an important objective in biological and medical research.The ?rst requirement for the study of proteins is to assess their ease of availability in terms of purity and quantity.There are three main routes to consider:(i)native
229
protein isolation,(ii)recombinant techniques for the expression of proteins in microorganisms,and(iii)chemical synthesis.Each of these methods has its advantages and disadvantages but only chemical peptide synthesis permits the incorporation of unnatural amino acids and the pro-duction of large quantities of pure peptides.Since the?rst synthesis of a dipeptide by Emil Fischer in1901,peptide science has made tremendous progress and with recent innovations it is currently possible to?routinely?synthesize proteins,well over200amino acids in length.The follow-ing sections focus on the development of peptide synthesis. First,a brief overview of the history of synthetic peptides will be given,followed by a presentation of the most recent methodology developed for the synthesis of large peptides: the so-called?chemical ligation?.Applications of this method in the synthesis of b-peptides are described.
A Brief History of Peptide Chemistry
The publication in1901,by E.Fischer(with E.Fourneau) (1),of the preparation of the dipeptide glycylglycine by hydrolysis of the diketopiperazine of glycine,is consid-ered to be the beginning of peptide chemistry.However, T.Curtius had synthesized and characterized a related peptide20years earlier,during his PhD studies with H.Kolbe,preparing the?rst N-protected dipeptide,ben-zoylglycylglycine,by treatment of the silver salt of glycine with benzoylchloride(2)(Fig.1).
The intensive work of T.Curtius with diazo compounds led to the development of the?rst practical method for peptide synthesis:the azide-coupling method,which was successfully employed in the synthesis of benzoylglycine peptides of a de?ned length(Fig.2)(3).E.Fischer,on the contrary,developed a method of peptide coupling based on the use of acylchlorides,prepared from the corresponding free amino acid using PCl5and acetyl chloride as solvent (4).
The major problems,and therefore limitations,of both methods stemmed from dif?culties at that time in obtaining enantiomerically pure l-amino acids and from the absence of an easily removable amino-protecting group. The introduction in1931,by M.Bergmann(a former stu-dent of E.Fischer)and L.Zervas,of the carbobenzoxy(Cbz) group(Fig.3left)for the temporary protection of the amino function solved part of the problem and led to a new era in peptide synthesis.From this point,numerous small pep-tides such as glutathione(5),and carnosine(6)were syn-thesized,culminating some20years later in the synthesis of an active hormone,the octapeptide oxytocin,by V.du Vigneaud et al.(7)(see Fig.4).V.Du Vigneaud was awarded the Nobel Prize2years later.It should be mentioned here that M.Bergmann and L.Zervas showed,during their synthesis of glutathione,that the use of the Cbz-protecting group prevented racemization during the formation of the acyl chloride,whilst with N-acyl-or N-benzoyl-protected amino acids,the reaction led to an almost complete racemization.It was later shown that this
con?gurational
Figure1.The?rst synthesis of a dipeptide by
T.Curtius and later by E.Fischer. Kimmerlin&Seebach.100years of peptide synthesis
230J.Peptide Res.65,2005/229–260
stability is a general property of urethane-protecting groups.
The introduction,in1957,by L.A.Carpino(8)and F.C. McKay and N.F.Albertson(9)of a new,acid-labile pro-tecting group,the tert-butyloxycarbonyl group(Boc)(Fig.3 right),which is stable toward hydrogenation,Birch reduc-tion,strong alkali and therefore totally orthogonal to the Cbz(or modi?ed Cbz)group and also to benzyl esters and ethers,greatly enhanced the arsenal of protecting groups available to the peptide chemist at the time.The combina-tion of Boc-and Cbz-protecting groups was then used for the synthesis of several peptides,the most spectacular example, for this period,being the synthesis of b-corticotropin Adre-nocorticotrophic Hormone(ACTH),a39-residue porcine hormone,in1963,by R.Schwyzer and P.Sieber(10)(Fig.4). The development of new protecting groups was accom-panied by intensive research toward discovering new coupling methods.The most important innovation in this regard was probably the introduction of carbodiimides in 1955by J.C.Sheehan and G.P.Hess(11)and H.G.Khorana (12)(Fig.5).The carbodiimide activation method has, however,a high propensity for racemization because of
the 231
high reactivity of the O-acyl-isourea,which can lead, through intramolecular cyclization,to the formation of an oxazolone:this cyclic intermediate can easily racemize via an aromatic intermediate as shown in Fig.5.The racemi-zation mechanism through the formation of the oxazolone has been extensively investigated.In their pioneering work, Goodman and McGahren(13)established all the factors affecting the optical purity of the oxazolone intermediate during the peptide bond formation,such as the nucleophi-licity/basicity ratio of the incoming amino group and sol-vent effects.
Activation of the carboxylic group can always lead to racemization via the oxazolone intermediate;however,the introduction of?additives?such as1-hydroxybenzotriazole (HOBt)to the reaction mixture was then shown to min-imize this by the formation of a less-reactive HOBt ester. Over the years,numerous other coupling reagents have been developed,and a selection of the most widely used ones is shown in Fig.6.
The next major breakthrough in peptide chemistry came in1963when B.Merri?eld published a historic paper des-cribing the principles and the applications of his
invention: J.Peptide Res.65,2005/229–260
solid-phase peptide synthesis (SPPS)(14).In contrast to the solution-phase methodology,where after each reaction the product has to be isolated and puri?ed before the next step,the growing peptide in the solid-phase approach is linked to an insoluble support and therefore,after each reaction step,the byproducts are simply removed by ?ltration and washing (Fig.7).Furthermore,because of the repetitive nature of peptide synthesis (deprotection,washing,coup-ling,washing,deprotection,…),the use of an insoluble support in a single reaction vessel allows for automatiza-tion of the processes.
In the beginning,the solid support used was a styrene–divinylbenzene co-polymer,functionalized by chlorome-thylation.The resulting benzyl chloride derivative (the Merri?eld resin)was used for anchoring the ?rst amino acid to the resin via an ester linkage.The peptide was then assembled using a carbodiimide as coupling reagent,with a combination of Boc as the protecting group for the N-te-rminus and benzyl for the side-chain functionalities.Upon completion of the last coupling,the peptide could be cleaved from the resin,with concurrent deprotection of the side-chain-protecting groups,using liquid hydrogen ?uoride (HF).The use of liquid HF can,however,lead to numerous side reactions,including catalytic Friedel–Crafts reactions between the aromatic groups of the resin and the side
chains of the peptide,and/or promotion of an N ?O acyl shift involving the side-chain groups of serine and threon-ine.These side reactions were minimized by the introduc-tion by Tam et al.(15)of a two-stage,low–high HF concentration cleavage protocol.
In 1970,L.A.Carpino and G.Y.Han (16)introduced a totally different protecting-group strategy.This was based
on the use of the base-labile 9-?uorenylmethyloxycarbonyl (Fmoc)group for the protection of the a -amino group,thereby allowing the orthogonal protection of side-chain groups through use of acid-labile protection (Fig.8).The mechanization of the SPPS process permits,in a fully automatic manner,the incorporation of more than 10amino acid residues per day and since the ?rst introduction of this methodology,has accounted for the synthesis of thousands of peptides.In 1984,B.Merri?eld was awarded the Nobel Prize in chemistry for his invention.In recent years,numerous new types of resins have been developed,allowing for the preparation,through the Boc/benzyl (Bzl)or the Fmoc/t -Butyl (Bu)strategy,of peptides bearing different functionalities at the C-terminus such as peptide-acid,-amide,-thioester,or -alcohol (a selection of derivatized resins for Fmoc and Boc SPPS is given in Tables 1and 2).There is an important aspect of peptide chemistry that has not been mentioned so far,but which is an essential consideration:the puri?cation and analysis of peptides.Parallel to the intensive development of synthetic meth-odology for the production of peptides,considerable pro-gress has been made to address these factors.Today,high-pressure (high-performance)liquid chromatography (HPLC)is the most widely used method and indeed a highly effective method for puri?cation,allowing the routine separation of complex product mixtures (17).Mass spectr-ometry,on the contrary,is the most powerful tool
for
Figure 7.Schematic presentation of the principles of solid-phase peptide
synthesis.
Figure 8.Fmoc strategy in solid-phase peptide synthesis.The Fmoc group is cleaved under basic conditions with piperidine,while the side-chain-protecting groups and the linker are cleaved under acidic condi-tions using TFA.
Kimmerlin &Seebach.100years of peptide synthesis
J.Peptide Res.65,2005/229–260
233
product analysis:determining the exact mass of an analyt-ical sample is the best proof that a synthesis was,at least in part,successful or not.The development of new technol-ogies in mass spectrometry permits the sequencing of pep-tides or proteins and therefore represents a direct analysis of their primary structures (18).
Synthesis of Proteins by Chemical Ligation
Since its introduction,the solid-phase methodology has advanced signi?cantly and now allows for the routine synthesis of peptides and small proteins with sizes up to 50amino acid residues.There are only a few examples describing the synthesis of longer chains,such as ribonuc-lease A (124residues)(19)and human immunode?ciency virus (HIV)-1TaT (86residues)(20,21)or the green ?uor-escent protein,a 238-residue peptide chain synthesized by fragment coupling on solid phase (22).These examples are,however,rather exceptional as the solid-phase methodology is usually limited because of the accumulation of side products arising from incomplete deprotection or coupling reactions.Unfortunately,the average length of a protein is
approximately 250amino acids and consist of two func-tional domains of 15kDa in size (23,24).In order to cir-cumvent this limitation of solid-phase methodology for the preparation of longer proteins,new approaches have been developed.The most useful and important one of these is chemical ligation,which allows for the coupling of unpro-tected peptide fragments in aqueous solution.
Table 1.Resins for Fmoc/t Bu solid-phase peptide synthesis (the solid supports are generally either cross-linked polystyrene or polyethylene glycol-based polymers)Peptide acid
Resin
Cleavage
Wang resin 95%TFA
Protected peptide
acid
2-Chlorotrityl resin 1%TFA in DCM
Peptide
carboxamide
Rink amide resin 95%TFA
Peptide ester and N
-alkylamide
4-Fmoc-hydrazinobenzoyl resin Cu(OAc)2,RNH 2or Cu(OAc)2,ROH,pyridine
Peptide
thioester
4-Sulfamylbutyryl resin (1)TMSCH 2N 2(2)RSH
Peptide
alcohol
HMBA resin
(4-hydroxymethylbenzoic acid)
NaBH 4/EtOH
Kimmerlin &Seebach.100years of peptide synthesis
234
J.Peptide Res.65,2005/229–260
Discovery and ?rst approach:prior ligation followed by intramolecular acyl shift
In 1953,Wieland et al.(25)reported the possibility of gen-erating an amide bond in aqueous solution through an intramolecular acyl shift (Fig.9).
It is exactly this principle that was used 25years later by Kemp et al.(26)and Kemp and Kerkman (27)in the devel-opment of the ?rst ligation methodology for the coupling of two peptide fragments:the ?prior thiol capture strategy ?.The principle is described in Fig.10.
Peptide fragment A is derivatized at the C-terminus in the form of an ester group,using a hydroxy-group that serves as a template to bring the acyl group and the amino group in close proximity.The thiol functionality present in the hydroxy-template moiety is referred to as the ?capture site ?and serves to attach the second peptide fragment to the template.In the ?rst step,this thiol function is deprotected and allowed to react with the thiol function of the N-terminal cysteine residue present in peptide frag-ment B .A subsequent intramolecular acyl transfer leads to the formation of a new amide bond between peptides A
Table 2.Resins for Boc solid-phase peptide synthesis (the solid supports are generally either cross-linked polystyrene or polyethylene glycol-based polymers)Peptide acid
Resin
Cleavage
Merri?eld HF,
TFMSA
PAM (4-hydroxymethyl-phenylacetamidomethyl)HF,TFMA
Protected peptide
acid
Oxime resin NaOH/dioxane
Peptide
carboxamide
MBHA resin (4-methylbenzhydrylamine)HF,TFMSA
Peptide
ester
Oxime resin MeOH/DMF/TEA
Peptide
hydrazide
Brominated PPOA resin
[4-(2-bromopropionyl)phenoxy]-acetic acid
NH 2NH 2/DMF
Kimmerlin &Seebach.100years of peptide synthesis
J.Peptide Res.65,2005/229–260
235
and B .Finally,the resulting peptide is released from the template.Extensive research into the design of the best template led to a dibenzofuran derivative shown in yellow in Fig.11(28).The kinetics of the acyl transfer was studied on a model reaction (Fig.11)and it was shown that peptide bond formation is extremely sensitive to steric interactions (29).Indeed,the reaction took 51h with R ?Val,which has a branched side chain,and only 3h with R ?Ala (Table 3).
This methodology was then successfully used for the synthesis of the 29-residue C-terminal segment of the pro-tein basic pancreatic trypsin inhibitor (BPTI)(30).As shown in Fig.12,a linear strategy was applied:four fragments were coupled sequentially through a prior thiol-capture reaction,from the C-to the N-terminus.Fragment 1was prepared in solution while the protected fragments 2,3,and 4were prepared by solid-phase methodology on an amino-methyl polystyrene resin,derivatized with the dibenzofuran tem-plate.The reaction sequence for each fragment coupling is as follows:
?Step 1:The ?rst step is the thiol-capture reaction,leading to the formation of a disul?de bond between the thiol functionality of the N-terminal cysteine present in one fragment and the thiol group of the template attached at the C-terminus of the second fragment.
?Step 2:The second step is the acyl transfer which leads to the formation of the amide bond between the two frag-
ments.
Figure 10.The principle of the ?prior thiol-capture strategy ?of
Kemp.
Figure 11.Model reaction for the determination of the kinetics of the acyl transfer.The part of the molecule in the yellow box is the template developed by
Kemp.
Table 3.Half-lifes for the intramolecular acyl-shift reactions shown in Fig.11,as a function of the acyl
fragment
J.Peptide Res.65,2005/229–260
?Step 3:The disul?de-linked template is then cleaved using PEt 3and the free thiol group of the cysteine residue (present at the ligation site)is protected with 2,4-dini-trophenyl (DNP)to prevent interference during the next fragment coupling.
?Step 4:All the side-chain-protecting groups except those of the cysteine residues are removed using TFA.?Step 5:Finally,the acetamidomethyl (Acm)group of the N-terminal cysteine is replaced by a sulfenylcarbometh-oxy (Scm)group (the Scm group allows for clean and rapid formation of the disul?de linkage in the capture steps:R–S–Scm +R ￠–SH ?R–S–S–R ￠).
The native chemical ligation
General principle of thioligation and synthesis of ?unnatural ?proteins
In 1992,Schno ¨lzer and Kent (31)introduced a novel strategy for the coupling of unprotected peptide fragments in aque-ous solution.The basis for this new approach,called ?chemical ligation ?,is the presence in each peptide fragment of a unique,mutually reactive functionality which enables a ?chemoselective ?reaction between the two components.The chemistry initially used by Schno ¨lzer and Kent for this purpose is a nucleophilic substitution reaction between an SH group of a thioacid attached to the C-terminus of one peptide,and an alkyl bromide attached to the N-terminus of the second fragment,leading to the formation of a thioester at the ligation site.This reaction can be performed in
aqueous solution:the selectivity of the reaction allows the use of unprotected peptide fragments.The characteristics of the chemical ligation methodology overcomes all the limitations of the traditional convergent approach for the synthesis of large peptides or proteins (i.e.poor solubility and dif?culty in purifying the fully protected peptide frag-ments),and provides access to new synthetic systems.One of the ?rst total syntheses of a protein by chemical ligation using unprotected peptides in aqueous solution was that of the human immunode?ciency virus-1protease (HIV-1PR)by Schno ¨lzer and Kent (31).The HIV-1PR protein is a homodimer of a 99-amino acid polypeptide chain with a molecular weight of 22.5kDa.The monomer was prepared by chemical ligation of a 51-residue peptide bearing a C-terminal thioacid and a 48-residue peptide having an N-terminal alkyl bromide (Fig.13).The reaction was per-formed in a phosphate buffer at pH 4.5,and was followed by analytical HPLC.After 3h,the reaction was almost com-plete and the product was puri?ed and characterized.The homodimer of this synthetic HIV-1protease analog (Fig.14)had the same biological activity as the native enzyme.
Native thioligation through N-terminal cysteine residues
The major disadvantage of the initial chemical ligation approach was that the reaction leads to an ?unnatural structure ?at the ligation site.A second generation of liga-tion chemistry,referred to as ?the native chemical ligation (NCL)?,was introduced in 1994by Dawson et al.(33).Similar to the previous strategy,this new
methodology
Figure 12.Synthesis of a portion of the basic pancreatic trypsin inhibitor using the prior thiol-capture strategy.
Kimmerlin &Seebach.100years of peptide synthesis
J.Peptide Res.65,2005/229–260
237
allows the coupling of two unprotected peptide fragments in aqueous solution,but now the ligation reaction leads to the formation of a ?native ?amide bond at the ligation site.The principle of this strategy is outlined in Fig.15.The ?rst step is a chemoselective transthioesteri?cation reaction involving the thiol group of the N-terminal cysteine present in unprotected peptide 6,and the C-terminal thioester moiety present in peptide 5.The thioester-linked interme-diate (7)then undergoes a rapid intramolecular S ?N acyl shift,forming the amide bond at the ligation site (8).The thioester intermediate is not isolated,but evidence for its formation has been obtained by using an N -acetyl-cysteine for the ligation reaction (33).In this case,the acyl transfer cannot proceed and the thioester intermediate could be isolated (Fig.16).
The nature of the thioester-leaving group greatly in?u-ences the rate of the reaction.Initially,benzyl thioesters were used and it has been shown that the addition of an excess of thiophenol to the reaction leads to an increase in the coupling rate (34).The thiophenol probably replaces the benzyl thiol before the ligation and therefore provides a better leaving group (Fig.17).Furthermore,the presence of an excess of thiol suppresses the risk of oxidation of the sulfhydryl group of the N-terminal cysteine residue.This is important because a ?cystin moiety ?cannot react in the ligation reaction.
An investigation of the in?uence of pH on the rate of the ligation reaction has shown that at pH 7the coupling pro-ceeds rapidly and that the reaction is almost complete
after
Figure 13.Total synthesis through chemical ligation of the HIV-1PR
analogue.
Figure 14.Structure of the synthetic HIV-1protease complex with a
substrate-based inhibitor at 2.3A
?resolution (PDB entry:4HVP)(32
).Figure 15.Synthesis of peptides by native chemical ligation of unpro-tected peptide fragments.
Kimmerlin &Seebach.100years of peptide synthesis
238
J.Peptide Res.65,2005/229–260
5min,while below pH 6the same reaction was only 50%complete after 10min.This suggests that the ionized form of the thiol moiety of the cysteine is directly involved in the reaction (33).
The impact on the rate of the coupling reaction in the NCL of the C-terminal amino acid present in the peptide fragment bearing the C-terminal thioester has also been investigated (35).This was performed in a model reaction where each of the 20possible thioesters was examined (Fig.18).The results are shown in Table 4:the ligation reaction proceeding in all the cases.Similar to the results obtained by Kemp (29)(see Table 3),the coupling reaction is much slower with Pro or when b -branched amino acids like Val,Ile,Thr are used.Curiously,the rate of the reaction with Cys and His is similar to that observed with Gly and
faster than that with Ala.This indicates that the side-chain groups facilitate the ligation reaction.
Finally,to complete these mechanistic studies,the extent of epimerization during the ligation process was also investigated.A model peptide was synthesized using the native chemical ligation methodology and for HPLC com-parison,the peptide containing the epimer of the amino acid located next to the ligation site was prepared by standard SPPS.Analysis by HPLC revealed that no epime-rization took place during the ligation reaction (36).
SPPS with a C-terminal thioester group
Peptides bearing a C-terminal thioester group are key intermediates in the synthesis of proteins by chemical ligation and consequently,several methods have been developed for their preparation on the solid phase.In the Boc strategy,the benzylic thiol 10(Fig.19)was used as a linker (36–38).This linker is prepared from chloride 9by reaction with thiourea and subsequent hydrolysis of the resulting thiouronium salt using aqueous base.The ?rst amino acid is then derivatized with this linker,and the resulting thioester 11is ?nally attached to an aminomethyl polystyrene resin.Standard Boc-SPPS is then applied for preparation of the desired peptide sequence.Cleavage of the peptide from the resin with HF leads to the formation of the peptidic thioacid 12,which is subsequently treated with benzyl bromide to yield the target peptide thioester 13
.
Figure 18.Model reaction for the determin-ation of the effect of the amino acid residue adjacent to the ligation site (yellow box)on the kinetics of the chemical ligation reaction.
Kimmerlin &Seebach.100years of peptide synthesis
J.Peptide Res.65,2005/229–260
239
A second protocol for SPPS containing a C-terminal thioester by the Boc strategy was developed by Tam and co-workers (39–41).The method is based on the use of a 4-methylbenzhydrylamine (MBHA)resin (Fig.20),which is ?rst loaded with S -trityl mercaptopropionic acid.After removal of the trityl-protecting group,the desired poly-peptide chain is assembled using standard Boc strategy.The thioester is obtained after cleavage with HF.
Despite the fact that thioesters are sensitive to basic conditions,Fmoc-compatible strategies for the preparation of peptide thioesters have also been developed.Ingenito et al.?s (42)methodology is based on the use of an acylsul-fonamide ?safety-catch linker ?14(Fig.21).This linker,?rst
introduced by Kenner et al.(43)and later modi?ed by Bac-kes and Ellmann (44),is stable to both strongly basic and acidic conditions.The peptide is assembled using the standard Fmoc protocol to afford solid-phase bound peptides
of type 15.After the ?nal peptide coupling,the resin is activated for cleavage by treatment with diazomethane to give an N ,N -methylacylsulfonamide 16,or by treatment with iodoacetonitrile to give an N ,N -cyanomethylacylsul-fonamide 17.The peptide is then released from the activa-ted resin by nucleophilic displacement involving a thiol group to yield 18,and ?nally the side chains are deprotected in solution to give the desired peptide thioester 19.Sewing and Hilvert (45)reported a second useful strategy (46).The target peptide is assembled by standard Fmoc chemistry on a 4-hydroxymethyl-phenylacetamidomethyl
Table 4.Rate of the coupling reaction in the chemical ligation step,depending on the C-terminal amino
acid
Coupling time
(h)
Coupling time
(h)
Coupling time
(h)
Coupling time (h)
Gly ￡4Ala ￡9Arg ￡24Ile ?48Cys ￡4Met ￡9Asn ￡24Leu ?48His
￡4
Phe ￡9Asp ￡24Pro ?48Trp ￡9Gln ￡24Thr ?48Tyr
￡9
Glu ￡24Val
?48
Lys ￡24Ser
￡
24
Figure 19.Kent methodology for the synthesis of a peptide thioester on solid
phase.
Figure 20.Boc-solid-phase synthesis of a peptide possessing a C-terminal thioester according to the method developed by Tam.
Kimmerlin &Seebach.100years of peptide synthesis
240
J.Peptide Res.65,2005/229–260
(PAM)or 4-hydroxymethylbenzoic acid (HMBA)resin (see Tables 1and 2),and the cleavage performed by activation of the ester linkage with AlMe 2Cl in the presence of a large excess of a nucleophilic thiol (Fig.22).The side-chain-pro-tecting groups are subsequently removed in solution by treatment with TFA.
Besides the synthetic methodology described here,recombinant strategies based on protein splicing have also been developed for the preparation of peptides bearing a C-terminal thioester [for an extensive review see Ref.(47)].
A protein and a glycopeptide synthesis by native thioligation
The ?rst example is the synthesis of a membrane protein:the 136-residue mechanosensitive ion channel from Myco-bacterium tuberculosis (Tb-MscL)(Fig.23)by Clayton et al.(48).
Membrane proteins are dif?cult to synthesize because of their large hydrophobic segment.They tend to couple inef?ciently,form aggregates,and are generally dif?cult to purify and solubilize.In order to overcome these problems,
Kochendoerfer applied the native chemical ligation meth-odology to the successive coupling of three fragments as shown in Fig.24(48).The wild-type protein does not con-tain any cysteine residues,and these were introduced by replacing two amino acids of the original sequence that are not critical for channel function.The two thioester frag-ments 20and 21were prepared on a thioester-generating resin similar to the procedure described above (Fig.20),and the third fragment 22was prepared on a PAM resin (Table 2)using standard Boc chemistry.As a result of the limited solubility of the peptide fragments,the ligation reactions were performed at 40°C in a phosphate buffer (pH 7.5)containing a high concentration of denaturant (8m urea).After the last ligation reaction,the ?non-native ?cysteine residues were selectively masked by treatment with bromoacetamide.The synthetic Tb-MscL protein prepared in this fashion displayed almost the same ion-channel activity as the native protein.
The second example illustrates the usefulness of the native chemical ligation methodology for the synthesis of large and highly functionalized biomolecules.As shown in Fig.25,Warren et al.(50)applied the native chemical liga-tion approach to the development of a convergent strategy for the synthesis of complex glycopeptides.The unpro-tected glycosylamines 23and 24are ?rst attached to pep-tides 25and 26,respectively,via the side chain of the Asp residue using O -(7-azabenzotriazol-1-yl)-N ,N ,N ￠,N ￠-tetramethyluronium hexa?uorophosphate (HATU)as a coupling reagent [Landsbury aspartylation (51)].After reductive cleavage of the disul?de bond in 27,and Fmoc deprotection of 28,the glycopeptides are coupled together via the native chemical ligation.It appears that the cleavage of the disul?de bond in 27with sodium 2-mercapto-ethanesulfonate (MES-Na)leads almost instantaneously to the formation of the MES-Na-derived thioester which can then react with the sulfhydryl functionality of the N-ter-minal cysteine present in fragment 29,giving after an intramolecular acyl shift,the desired amide bond (30).This methodology was successfully used for the synthesis of normal and transformed fragments of prostate-speci?c antigen glycopeptides (52
).
Figure 22.Formation of a C-terminal peptide thioester by dimethylaluminum thiolate-mediated
cleavage.
Figure 21.Fmoc-solid-phase synthesis of a peptide thioester on a sul-fonamide ?safety-catch ?resin.
Kimmerlin &Seebach.100years of peptide synthesis
J.Peptide Res.65,2005/229–260
241
Ligations without cysteine at the coupling site
The native chemical ligation methodology introduced by the Kent group (33,53)has proven very useful in the syn-thesis of large peptides and proteins.This strategy has,however,some limitations,principally because of the fact that a cysteine is required at the ligation site,and that naturally occurring proteins do not always contain a cys-teine residue in the right position of their sequences.Therefore,several modi?cations of the initial method have been introduced to overcome this drawback.
Methionine,histidine,selenocysteine and homoselenocysteine ligation
Tam and Yu (54)have shown that the chemical ligation reaction can also be performed with a peptide bearing an N-terminal homocysteine residue instead of a cysteine.In
this case,the intramolecular acyl shift proceeds through a six-membered ring (vs.a ?ve-membered ring for cysteine).The product of this ligation contains a homocysteine at the ligation site,which is subsequently methylated with methyl p -nitrobenzenesulfonate to produce a methionine side chain (Fig.26)(55).
Similar to cysteine,it has been shown by Hilvert and co-workers (56,57)and van der Donk and co-workers (58,59)that the native chemical ligation can also be mediated by selenocysteine and selenohomocysteine (60)amino acids instead of cysteine.This is of particular interest in biology as the speci?c physical properties of selenium (SeH is more acidic and is a better nucleophile than SH)(61)are useful for the mechanistic investigation of bioactive peptides and proteins containing a Cys in the active site.Furthermore,selenium-containing peptides can be used as a probe in
nuclear magnetic resonance (NMR)spectroscopy (62).Finally,selenocysteine can also be reduced to an alanine or converted through oxidative elimination to a dehydroala-nine residue (57)(Fig.27).
Zhang and Tam have shown that the imidazole side chain of histidine,if located at the N-terminus of a peptide,can also be used as a nucleophile in the ligation reaction.In this case,the C-terminus of the second peptide fragment is activated as a disul?de,generated in situ by treatment of the thioacid 31with 5,5￠-dithiobis(2-nitrobenzoic acid)(Fig.28).The reaction between the two fragments,peptide 33having an N-terminal histidine and peptide 32bearing an activated thioacid,?rst generates an amide 34which then undergoes a N im ?N a acyl transfer,via a six-member d-ring intermediate,leading to the ligated product 35(63).
Thioligation with a removable auxiliary –an application of the thiol-capture strategy
Another way to overcome the N-terminal cysteine residue requirement in the native chemical ligation methodology
is
Figure 23.Structure of the homopentamer of the MscL homologue from Mycobacterium tuberculosis (PDB entry:1MSL)(49
).
J.Peptide Res.65,2005/229–260
to mimic the characteristic of this Cys by the use of a removable auxiliary.For this purpose,the Canne et al.replaced the N-terminal Cys by an oxyethanethiol-substi-tuted N-terminus (Fig.29)(64).The ?rst step in the ligation reaction involve a thioester exchange between peptide 36,bearing a C-terminal thioester,and the thiol function present at the N-terminus of peptide 37.The ligation product
38then undergoes an S ?N acyl shift with formation of an amide bond (39)having an N -(oxyalkyl)substituent on nitrogen.The N–O bond can then be cleaved under reducing conditions,e.g.with zinc dust,to give peptide 40.
Several groups have shown that the N-terminal cysteine can also be replaced by a 2-mercaptobenzylamine linker (Fig.30)(65–67).In this case,the intramolecular acyl shift proceeds through a six-membered ring with the linker subsequently removed by treatment with an acid such as tri?uoromethanesulfonic acid (TFMSA).
A third type of auxiliary for the native chemical ligation of unprotected peptide fragments was introduced by Kawakami and Aimoto (69)in Japan and shortly after by Marinzi et al .(68).The auxiliary attached at the N-termi-nus of one peptide fragment is based on an o -nitrobenzyl scaffold (Fig.31),with which the thioester present in the second peptide fragment can react.The reaction is per-formed under standard conditions and provides a substi-tuted amide at the ligation site as shown in Fig.31.The
photolytic properties of the o -nitrobenzyl group permits the use of mild conditions for the cleavage of the auxiliary from the ligated product (68,69).
Other ligation strategies
Besides the native chemical ligation methodology des-cribed previously,and characterized by the reaction of
a
Figure 26.Native chemical ligation with homocysteine.
Kimmerlin &Seebach.100years of peptide synthesis
J.Peptide Res.65,2005/229–260
243
N-terminal Cys with a C-terminal thioester,numerous other approaches for the coupling of unprotected peptide fragments have been developed.Each of these methods can be catagorized by the types of reactions used to bring the peptide fragments together(the capture step)before the acyl transfer
reaction.Figure27.Synthesis of peptides by seleno-cysteine-mediated chemical
ligation.
Figure28.Histidine-mediated native chem-
ical ligation of unprotected peptide frag-
ments.
Figure29.Chemical ligation through an
N a
-(oxyethanethiol).
J.Peptide Res.65,2005/229–260
Imine ligation with incorporation of pseudo-prolines The imine ligation developed by Tam and co-workers (70–73)involves a peptide segment having a C-terminal glycoaldehyde (41)and a peptide bearing an N-terminal Cys (42),Thr (43)or Ser (44)residue (Fig.32).The ?rst step in this ligation process is the formation of the imines 45,46,or 47by addition of the N-terminal amino group of peptides 42,43,or 44to the C-terminal aldehyde function of peptide 41.The presence of a nucleophilic SH or OH functionality in the side chain of the N-terminal amino acid of the segment leads,through a ring-chain tautomerism,to a thiazolidine 48or an oxazolidine 49,50,respectively.This heterocyclic intermediate then undergoes an intramolecular O ?N acyl shift via a favorable bicyclic ?ve-membered ring interme-diate to form an amide bond and a hydroxymethyl-substit-ued pseudo-proline (W pro )moiety at the ligation site (51,52,53).
The peptide bearing the C-terminal aldehyde function-ality was prepared on a benzaldehyde polystyrene resin (74).Treatment of this resin with glycerol in the presence of a catalytic amount of p -toluenesulfonic acid leads to the for-mation of a benzylic acetal.The peptide sequence can then be assembled using standard Fmoc solid-phase strategy.
After cleavage with TFA,the diol of the glyceric ester is oxidatively cleaved,using NaIO 4,to give the desired pep-tide having an aldehyde function at its C-terminus (Fig.33).The formation of the thiazolidine ring during the imine ligation is 104times faster than that of the oxazolidine ring (75)and the reaction can be performed in aqueous solu-tion,whilst for the oxazolidine anhydrous conditions are required (for oxazolidine ligation the best solvent was found to be a 1:1mixture of pyridine and acetic acid)(72).An important aspect in the imine ligation process is the fact that during the formation of the heterocycle,a new asym-metric carbon is created at position C(2)of the pseudo-proline ring (Fig.34).Furthermore,the new amide bond formed in the acyl-transfer step can adopt a cis or trans conformation (Fig.34).The con?guration of the new center and the cis –trans conformational ratio was determined in a model reaction as shown in Fig.34.In the case of the thi-azolidine ester (55+58),both diastereoisomers were formed and are stable enough to be separated by HPLC.However,after the acyl transfer only one diastereoisomer was detec-ted.The oxazolidine ester intermediates (54+56,55+56,54+57,and 55+57)were not stable enough to be detected by HPLC;however,as with the thiazolidine a
single
Figure 33.Solid-phase synthesis of a peptide bearing a C-terminal aldehyde function.
Kimmerlin &Seebach.100years of peptide synthesis
J.Peptide Res.65,2005/229–260
245
diastereoisomer was isolated after the acyl transfer.The con?guration of the new centre was assigned in both cases to be (R )by 2D NMR experiments (Table 5)(72).The dia-stereoselectivity of the ligation reaction can be explained as follows:the formation of the thiazolidine or oxazolidine ring is not stereoselective and therefore two diastereoiso-mers are formed.However,in the acyl transfer reaction,the C(2)-(R )-diastereoisomer reacts faster than the (S )-epimer,and gives the more stable cis -substituted product and because of ring-chain tautomerization the (S )-epimer undergoes a re-equilibration to the (R /S )mixture.The cis –trans amide rotamer ratio was determined by NMR spectroscopy and it was shown that this ratio varied,depending on the amino acids located in the vicinity of the pseudo-proline ring (72).
The imine-ligation methodology was then successfully applied by Miao and Tam (78)for the synthesis of ?ve analogs of a proline-rich helical antimicrobial peptide,the 59-residue bactenecin 7(Bac 7)(Fig.35).They used a three-segment ligation strategy,applying the thiaproline and oxaproline ligation methodology simultaneously.Two analogs 65and 66,were prepared by ligation in a C to N direction based on the fact that the ligation with thiaproline is much faster than that with oxaproline,and therefore the N-terminal Ser (resp.Thr)present in 60(resp.61)does not need to be protected (Fig.36).The C-terminal glycoalde-hyde ester in 60(resp.61)then reacts selectively with the N-terminal Cys present in 59.This thiaproline ligation was
performed under two-stage aqueous conditions;?rst at pH 5.2for 10h which led to formation of the two diastereo-meric thiazolidine-esters (see Fig.32),and then at pH 6.6for 20h to give,after O ?N acyl transfer,a single product 62(or 63)as determined by RP-HPLC.The third fragment 64was attached through an oxaproline ligation.
The three other analogs 73,74,and 75were prepared through an N to C three-segment sequential ligation as shown in Fig.37.Fragments 64and 67were ?rst coupled by a thiaproline ligation (68),followed by oxidative cleavage
using NaIO 4of the C-terminal glycerol ester of the result-ing peptide,to give glycoaldehyde ester 69.Subsequent ox-aproline or thiaproline ligation between 69and the fragment bearing an N-terminal Ser 70,Thr 71or Cys 72yielded the three Bac analogs 73,74,and 75,respectively.Circular dichroism (CD)investigations revealed that the ?ve analogs formed a stable polyproline helical structure in aqueous solution.However,the replacement of the Pro residues present in Bac 7by OPro and SPro in the analogs 65,66,73,74and 75resulted in a minor population of polypro-line type I structure,while Bac 7adopts a polyproline type II helical structure.The antimicrobial activities of the Bac 7analogs were similar to that of the natural product.
Oxime and hydrazone ligations
Similar to the imine ligation,hydrazide-or aminoxy-deri-vatized N-terminal peptides can be ligated,through hydra-zone or oxime linkages respectively,to peptides having a C-terminal aldehyde function.Hydrazides and oximes are good nucleophiles and highly reactive toward aldehydes under acidic condition where the basic side chains are protonated and therefore excluded from the reaction.The reaction is extremely selective and is compatible with unprotected side-chain functionality with the exception of Cys (thiazolidine formation).This type of ligation has been used to prepare cell-permeable lipopeptides (79),a dimeric transcription-factor-related protein containing 172amino acids (38),and peptide dendrimers (80,81),for
examples.
Table 5.Result of the pseudo -proline-ligation reaction Segments
X
R
1
R
2
C(2)
cis /trans amide conformation
54+56O Me H R 68:3255+56O iPr H R 56:4454+57O Me Me R 54:4655+57O iPr Me R 43:5755+58
S
iPr
H
R
40:
60
Figure 35.Primary sequence of the 59-resi-due antimicrobial peptide bactenecin 7(Bac 7).
Kimmerlin &Seebach.100years of peptide synthesis
246J.Peptide Res.65,2005/229–260
An interesting application of the use of a hydroxyl amino group is the synthesis of the glycopeptide analog 78(82)which was prepared through an oxime ligation of four disaccharide 77(T-antigen:b -d -Gal(1?3)-a -d -GalNac)units to a cyclic decapeptide 76derivatized with four-alde-hyde functionalities (Fig.38).
The Staudinger ligation
An elegant way to overcome the limitation of the original native chemical ligation was developed independently by two research groups:Raines (83–86)and Bertozzi (87,88)and is based on the Staudinger reaction.The Staudinger reac-tion,which was ?rst reported in 1919(89),is the reaction of
a phosphane with an azide to produce an iminophospho-rane.This iminophosphorane intermediate can then be trapped by different electrophiles (Fig.39)[for an extensive review on the Staudinger reaction see Ref.(90)].
This reaction can be applied to the coupling of two peptide fragments,one bearing a C-terminal phosphino-thioester group and the second an N-terminal azido group [for a review on the application of the Staudinger reaction for peptide ligation see Ref.(91)].In the ?rst step of this Staudinger ligation,the phosphinothioester reacts with the azide to give an iminophosphorane,which then undergoes an intramolecular S ?N acyl shift leading to an amido-phosphonium salt.The amidophosphonium salt is
then
Figure 36.Three-segment tandem ligation in a C to N direction affording two analogs of Bac 7
.
Figure 37.Synthesis of three Bac 7analogs through an N to C three segment sequential ligation.
Kimmerlin &Seebach.100years of peptide synthesis
J.Peptide Res.65,2005/229–260
247
hydrolyzed to produce the amide product and a phosphine oxide.The high reactivity of the aza-ylide does not,how-ever,permit the presence of unprotected side-chain func-tionalities and therefore the Staudinger ligation is limited to the coupling of fully protected peptide fragments (Fig.40).
Solid-phase thioligation of unprotected peptide segments
Recently,Kent (92)and Dawson (35,93)have developed a solid phase approach for the coupling of unprotected peptide fragments using the native chemical ligation methodology.There are two ways to assemble the desired
polypeptide
Figure 38.Synthesis of a glycopeptide using oxime
ligation.
J.Peptide Res.65,2005/229–260