Download PDF
Synlett 2019; 30(11): 1289-1302
DOI: 10.1055/s-0037-1612417
account
© Georg Thieme Verlag Stuttgart · New York

Peptide Modifications: Versatile Tools in Peptide and Natural Product Syntheses

Phil Servatius
,
Lukas Junk
,
Uli Kazmaier  *
Department of Organic Chemistry, Saarland University, D-66123 Saarbrücken, Germany   Email: u.kazmaier@mx.uni-saarland.de
› Author Affiliations
Further Information

Publication History

Received: 14 January 2019

Accepted after revision: 07 February 2019

Publication Date:
02 April 2019 (online)

 
 PDF Download Permissions and Reprints All articles of this category


Abstract

Peptide modifications via C–C bond formation have emerged as valuable tools for the preparation and alteration of non-proteinogenic amino acids and the corresponding peptides. Modification of glycine subunits in peptides allows for the incorporation of unusual side chains, often in a highly stereoselective manner, orchestrated by the chiral peptide backbone. Moreover, modifications of peptides are not limited to the peptidic backbone. Many side-chain modifications, not only by variation of existing functional groups, but also by C–H functionalization, have been developed over the past decade. This account highlights the synthetic contributions made by our group and others to the field of peptide modifications and their application in natural product syntheses.

1 Introduction

2 Peptide Backbone Modifications via Peptide Enolates

2.1 Chelate Enolate Claisen Rearrangements

2.2 Allylic Alkylations

2.3 Miscellaneous Modifications

3 Side-Chain Modifications

3.1 C–H Activation

3.1.1 Functionalization via Csp3–H Bond Activation

3.2.2 Functionalization via Csp2–H Bond Activation

3.2 On Peptide Tryptophan Syntheses

4 Conclusion


#

Key words

C–H activation - natural products - peptide enolates - peptide modifications - total synthesis

Biographical Sketches

Zoom Image

Phil Servatius was born in Saarbrücken, Germany, in 1989 and graduated from Saarland University in 2014 with a master’s degree in chemistry. After an internship at Bayer Healthcare in Wuppertal, he performed his Ph.D. studies in Uli Kazmaier’s lab at Saarland University. He obtained his Ph.D. in 2018 and has a longstanding interest in the synthesis of biologically active natural products and derived structures. Furthermore, he is engaged in the development of new transition-metal-catalyzed reactions.

Zoom Image

Lukas Junk was born in Trier, Germany, in 1993 and studied chemistry at Saarland University in Saarbrücken from 2009 to 2014. After an internship at Bayer Healthcare in 2015, he returned to Uli Kazmaier’s group in 2015, where he earned his Ph.D. in January 2019. He is currently working as a postdoctoral researcher in Uli Kazmaier’s group, continuing his research on natural product synthesis and the development of new methodologies.

Zoom Image

Uli Kazmaier was born in 1960 and studied chemistry at the University of Stuttgart where he obtained his diploma in 1985 and his Ph.D. in 1989. Afterwards, he undertook postdoctoral research in the groups of M. T. Reetz (Marburg) and B. M. Trost (Stanford). In 1992 he moved to Heidelberg, starting his own scientific work as a “habilitand” at the Institute of Organic Chemistry. In 2000 he received a Novartis Chemistry Lectureship and an offer of a full professorship at the University of Bayreuth. In 2001 he obtained an offer of a full professorship at Saarland University, which he accepted. His current research interest extends to new organometallic reagents and reactions, especially amino acid and peptide synthesis. Besides the development of new synthetic protocols, the application of these new reactions towards the synthesis of natural products and other pharmaceutically relevant structures plays a central role.

1

Introduction

Zoom Image
Scheme 1 Regio- and diastereoselective modification of cyclosporin A according to Seebach et al.

Amino acids, along with oligomeric peptides and proteins, play key roles in biological processes and are involved in many biochemical pathways including metabolism, signal transduction and cell differentiation. Peptidic natural products, often produced by microorganisms and plants as secondary metabolites, are in the spotlight of many recent research projects, based on their interesting biological properties.[1] These small peptides differ significantly from larger proteins ubiquitously found in all living organisms. While proteins mainly consist of the 20 proteinogenic amino acids, along with some post-translational modifications including disulfide bridging or phosphorylation, small peptidic secondary metabolites often contain a variety of complex non-canonical amino acids.[2] While often the natural products, especially the highly active and therefore interesting ones, are isolated in only tiny amounts, total synthesis may be the only way not only to confirm the proposed structure, but also to produce sufficient material for biological studies. In addition, total synthesis also allows for the modification of the natural products for structure–activity relationship (SAR) studies or to convert them into tools for chemical biology.[3]

Therefore, the direct introduction or manipulation of a side chain onto a given peptide represents a powerful alternative to classical peptide synthesis, since a variety of modified analogues can be prepared from only one parent peptide precursor.[4] Therefore, it is not surprising that numerous different approaches towards the modification of small- to medium-sized peptides have been developed over recent years, by either backbone modifications (introduction of side chains) or side-chain manipulations.

Fundamental work on backbone modifications has been reported by Seebach et al., who investigated the introduction of side chains into small peptides via enolate chemistry.[5] Probably the most spectacular application of this protocol is the site-selective alkylation of cyclosporin A at the sarcosine subunit (Scheme [1]).[6] They observed that deprotonation of amide N–H bonds ‘protects’ adjacent amino acid residues from deprotonation and consequently, epimerization. Therefore, deprotonation with LDA afforded a hexalithiated compound, which readily underwent nucleophilic substitution with electrophiles. The modified cyclosporins were obtained in good yields (up to 90%) and diastereoselectivity (up to dr 5:1). The stereochemical outcome of the alkylation step was determined by the conformation of the cyclic peptidic backbone. It should be mentioned that the selectivities observed with linear peptides are generally lower because of the flexibility of the peptide chain.

The concept of side-chain introduction is not limited to peptide enolates but can also be performed with other reactive intermediates, such as peptide cations[7] or radicals.[8] But also here, the stereochemical outcome in the modification step could generally not be controlled, except if cyclic and therefore rigid structures were modified.[9]

To face this issue of stereocontrol, we decided to investigate reactions of chelated amino acid ester enolates.[10] These enolates can easily be obtained from protected amino acid esters (Scheme [2]) (PG = protecting group) via deprotonation with LDA (or LHMDS) and subsequent transmetalation by addition of metal salts (MXn). They show the usual metal enolate reactivity towards electrophiles (E), undergoing typical enolate reactions such as aldol additions,[11] epoxide openings[12] or Michael additions,[13] including Michael-induced ring-closing reactions.[14]

Zoom Image
Scheme 2 Generation and reactions of chelated amino acid ester enolates

These chelated enolates have several advantages in comparison to their non-chelated analogs:

1. As a result of the fixation of the enolate geometry by chelation, many reactions of these enolates proceed in a highly stereoselective fashion.

2. The chelated enolates are significantly more stable than the corresponding lithium enolates. In contrast to them, the chelate enolates can be warmed to room temperature or even refluxed in THF without decomposition. This allows for the expansion of the field of enolate chemistry to reactions which in general cannot be carried out with non-stabilized enolates.

3. Presumably the coordination sphere of the metal ion (M) is not saturated in the bidentate enolate complex, and therefore the coordination of additional chiral ligands on the chelated metal is possible,[15] or as we hoped, a small peptide chain might coordinate multifold towards the chelating metal. Under optimized conditions, the side chains of the chiral amino acids incorporated into the peptide chain should shield one face of the enolate, directing the incoming electrophile to the opposite face.


# 2

Peptide Backbone Modifications via Peptide Enolates

Based on the pioneering work of Seebach et al.[5] and our previous work on amino acid synthesis[10], we tried to apply the different protocols also to peptide modifications, with the goal of using the chiral information of a given peptide chain to control the stereochemical outcome of the modification step.

2.1

Chelate Enolate Claisen Rearrangements

Actually, we began our investigation of chelate enolate reactions with Claisen rearrangements. Because of a preferred chair-like transition state, the simple diastereoselectivity of the rearrangement product provides important information about the enolate geometry.[16] The selectivities obtained with the chelated enolates are excellent, a clear indication for the intermediate chelate complex formation.[17] The use of chiral allylic alcohols allows for the synthesis of γ,δ-unsaturated amino acids, not only in a highly diastereoselective manner, but also in a highly enantioselective fashion (Scheme [3]).[18] Meanwhile, this approach has found widespread application in the synthesis of unusual amino acids[19] as well as some natural products such as the microcystines,[20] cylindrospermopsin,[21] lucentamycin A,[22] chlamydocin,[23] Cyl-1[24] and the cyclomarins.[25]

Zoom Image
Scheme 3 [3,3]-Sigmatropic rearrangement of chelated ester enolates

Alternatively, we also investigated the rearrangement of peptide allylic esters. The stereochemical outcome of the reaction strongly depends on the N-protecting group and the chelating metal salt. Tosyl-protected peptides are the substrates of choice if the Claisen rearrangement should be controlled by the peptide chain. In general, the induced diastereomeric ratios (dr) are in the range of 90:10 to 96:4 with yields up to 90%.[26] While different metal salts give good selectivities, the best results are often obtained with Ni salts. In all cases, an (R)-amino acid is formed if an (S)-amino acid is placed in the peptide chain, and vice versa. For the reactions with NiCl2, the stereochemical outcome can be rationalized by the formation of a square planar chelate complex in which one face of the enolate is shielded by the side chain of the adjacent (S)-amino acid. The rearrangement occurs on the sterically less hindered ‘opposite’ (unlike) face of the enolate, giving rise to the (R)-amino acid. Since N-tosyl and related protecting groups are generally difficult to cleave, we also investigated rearrangements of peptides bearing ‘classical’ N-protecting groups. Relatively good results were also obtained with N-Boc-protected peptides in combination with SnCl2 as a chelating metal salt (Scheme [4]).[26]

Zoom Image
Scheme 4 Stereoselective peptide Claisen rearrangements
Zoom Image
Scheme 5 Peptide Claisen rearrangements of chiral allyl esters
Zoom Image
Scheme 6 Cyl-1 derivatives via peptide enolate Claisen rearrangements

Excellent yields and simple diastereoselectivities were also obtained if manganese salts were utilized, but no significant induced diastereoselectivity could be observed in this case.[27] Because the influence of the adjacent amino acid on the rearrangement can be neglected, this allows for the stereoselective synthesis of peptides if esters of chiral allylic alcohols are used (Scheme [5]). The corresponding dipeptide could be obtained not only in good to excellent yield, but also in a highly diastereoselective fashion. Depending on the chiral alcohol used, both configurations can be obtained on request.

A similar situation is found in peptides containing N-methyl amino acids or prolines and pipecolic acids, where the multidentate coordination of the deprotonated peptide towards the chelating metal salt is suppressed. Very recently, we applied such a peptide Claisen rearrangement in the synthesis of Cyl-1 derivatives (Scheme [6]).[28] By using three equivalents of ZnCl2, the N- and the C-terminus of the tetrapeptide can form independent chelate complexes, and the results obtained were comparable to those of glycine ester enolate rearrangements.


# 2.2

Allylic Alkylations

During our investigation of peptide Claisen rearrangements, we observed that the reaction could be ‘catalyzed’ by palladium complexes, however, interestingly not by the generally used Pd(II) complexes[29] but by Pd(0).[30] In this case, the peptide allylation no longer occurred via a Claisen rearrangement, but via an allylic alkylation.[31] Chelated enolates were found to be excellent nucleophiles in allylic alkylation.[32] Based on their high nucleophilicity as metal enolates, they react readily at temperatures as low as –78 °C, which leads to suppression of π–σ–π-isomerizations,[33] which are typical for Pd-allyl complexes. Besides Pd catalysts, also Rh-[34] or Ru-complexes[35] can be used, showing different allylation behaviors.

Because the intermolecular allylic alkylation was observed as a ‘side reaction’ in the ‘Pd-catalyzed’ Claisen rearrangement, it is not really surprising that peptide enolates can also be applied as nucleophiles in allylic alkylations.[36] Also here, the best results are obtained by deprotonation in the presence of ZnCl2. The N-TFA-protecting group in many cases proved to be superior to other standard protecting groups with respect to yield and selectivity, while t-butyl esters are the C-terminal protecting groups of choice. Excellent diastereoselectivities can be obtained for various allylic substrates and highly functionalized side chains can also be incorporated with excellent yields and selectivities (Scheme [7]). The stereochemical outcome of the reaction is exclusively controlled by the peptide chain as long as terminal π-allyl-palladium complexes are involved. Probably, in the case of dipeptides, the deprotonated peptide backbone coordinates threefold to the chelating zinc ion. Again, one face of the peptide enolate is thought to be shielded by the side chain of the adjacent amino acid, directing the electrophilic attack to the opposite face. This rationalizes why an (S)-amino acid preferably generates an (R)-amino acid (and vice versa) as in the case of the Claisen rearrangements. The excellent diastereoselectivities can be explained by steric interactions between the peptide chelate complex and the sterically highly demanding π-allyl-Pd complex. Excellent yields are also obtained with N-terminal-alkylated dipeptides (no NH), but without any selectivity since no tricoordinated chelate complex can be formed.

Zoom Image
Scheme 7 Stereoselective Pd-catalyzed allylic alkylations of dipeptides. Reaction conditions: (1) 3.5 equiv LHMDS, 1.2 equiv ZnCl2, THF, –78 °C; (2) 0.7 equiv allylic substrate, 1.4 mol% [allylPdCl]2, 6.3 mol% PPh3, THF, –78 °C to rt.

Recently, we employed this approach in the construction of the peptidic backbone of the natural HDAC inhibitor trapoxin A.[37] We envisioned the combination of (R)-configured pipecolic acid and the adjacent (S)-configured non-proteinogenic Aoe to be almost ideal for a Pd-catalyzed allylic alkylation. But the pipecolic acid as a secondary amino acid represents the same situation as N-methylamino acids at the N-terminus, and therefore, the pipecolic acid had to be generated after the allylation step. Consequently, we investigated the allylation of a dipeptide with a functionalized N-terminal amino acid obtained via the previously discussed chelate Claisen rearrangement. The chiral induction caused by this not very sterically demanding linear amino acid was slightly worse than in the previous cases, but still in the synthetically useful range (Scheme [8]). Hydrogenation of both the double bond and benzyl ether functionalities, followed by Mitsunobu cyclization, afforded the pipecolic acid containing dipeptide in excellent yield. Subsequent coupling steps, macrocyclization and installation of the α-epoxy ketone moiety provided trapoxin A as a single diastereomer.

Zoom Image
Scheme 8 Pd-catalyzed allylic alkylation in the synthesis of trapoxin A

Although this concept allows for the introduction of a wide range of side chains into a given peptide, for each side chain a corresponding allylic substrate has to be synthesized first. To make the protocol even more flexible and suitable for library synthesis, it is desirable to introduce a functionalized side chain which can later on be further modified, e.g., via cross-coupling chemistry. Ideal candidates for this approach are peptides with a vinylstannane side chain, which can be subjected to Stille couplings under completely neutral conditions,[38] which is mandatory to avoid epimerization of the peptide chain.[39] The required stannylated allyl acetates and carbonates can easily be obtained by regioselective molybdenum-catalyzed hydrostannation of the corresponding propargyl alcohol derivatives.[40] Introduction of the bulky tributylstannyl substituent increases the size of the π-allyl complex even further, and therefore, the peptide allylation products are formed with almost perfect diastereoselectivity (Scheme [9]).[41] Subsequent couplings with allyl or benzyl halides provided γ-substituted amino acids with a terminal double bond, which can be further modified, e.g., via a thiol ene click reaction.[42] On the other hand, couplings with acyl halides give access to amino acids with a vinyl ketone in the side chain, which are excellent Michael acceptors.[43] Tin–iodine exchange results in an ‘umpolung’ of the side chain which is now a suitable candidate for cross-coupling reactions with a wide range of organometallics. Stille coupling with vinyl stannane generates a 1,3-diene side chain, which can be subjected to Diels–Alder reactions.[44] If carbonylation reactions are performed in alcohols as solvents, the corresponding α,β-unsaturated esters are formed, while amines, amino acids and peptides give access to the corresponding amides and cross-linked peptides. All these products can be obtained from one intermediate, the configuration of which was determined in the allylation step.

Zoom Image
Scheme 9 Synthesis and reactions of stannylated peptides

This protocol is not limited to C-terminal peptide esters, but can also be applied to secondary amides, which are able to form amide enolates. This is especially interesting because it allows the allylation of proline- or N-alkylated amino acid containing peptides in the middle of a peptide chain (Scheme [10]).[45] The ruthenium-catalyzed allylic alkylation is an interesting alternative to the palladium-catalyzed process, since it can provide products which are not accessible under Pd catalysis.[46] Chiral terminal allylic substrates could be reacted with perfect stereo- and regioretention (rr), and (Z)-configured allylic substrates could be converted without isomerization. The configuration at the α-position of the newly generated α-amino acid can be controlled by the peptide, and at the β-position by using chiral allylic substrates.

Zoom Image
Scheme 10 Allylic alkylations of peptide amides

# 2.3

Miscellaneous Modifications

In principle, chelated peptide enolates can also be reacted with other classical electrophiles such as alkyl halides, Michael acceptors or aldehydes, although the diastereoselectivities observed are significantly lower due to the lower steric demand of the electrophiles. Recently, we used such an approach in the synthesis of the miuraenamides, marine natural products with high cytotoxicity towards a wide range of cancer cell lines (Scheme [11]).[47] Deprotonation of a cyclic peptide intermediate afforded a trilithiated enolate which was treated with benzaldehyde in an aldol addition. In this case, no chelating metal was used, since the stereochemical outcome of the reaction was not relevant. The mixture of the four diastereomers formed was oxidized using Dess–Martin periodinane (DMP) to give the corresponding β-ketoester which is configurationally labile. However, only one stereoisomer was observed by NMR. Deprotonation and O-alkylation provided a mixture of the two isomeric enol ethers, which were O-deallylated to obtain the natural products miuraenamides A, D and E. Miuraenamide E was formed during flash chromatography from the (Z)-configured miuraenamide D, which is obviously less stable than the (E)-isomer. This late-stage modification not only allowed the synthesis of the natural products, but also of more than 50 derivatives by using different substituted benzaldehydes,[48] or via modification of the cyclic enolate by alkylation, allylic alkylation or Michael addition.[49]

Zoom Image
Scheme 11 Synthesis of miuraenamides via peptide modification

#
# 3

Side-Chain Modifications

The side chains of functionalized amino acids pose the most obvious handles for modifications on a peptide. Therefore, many different modifications of peptide side chains have been described in recent years.[50] Most of these methods aim to conjugate, for example, fluorescent probes or other functionalities to native peptide chains by using the functionalities of, for instance, Cys or Lys. These examples will not be discussed in this Account, but we will rather focus on modifications of small peptides, which might be useful for the diversification of natural products.

3.1

C–H Activation

In the past decade, the direct functionalization of C–H bonds via transition metal-catalyzed activation has proven its potential in many different transformations and these developments have been covered in a number of recent reviews.[51] Consequently, several methods have also been described for the direct functionalization of amino acids and peptides.[51k] [52] This protocol is advantageous compared to other functionalized side-chain modifications based on functional group interconversions, since non-activated C–H bonds can be functionalized. The only issue is the selective functionalization of a distinct C–H bond. This can be solved by employing suitable directing groups[53] or by tackling predetermined positions in the side chain. In principle, C–H activations can be used for backbone modifications,[54] as discussed for the peptide enolates, but the stereochemical outcome is again hard to control. More attractive are side-chain modifications, where the configuration of a given amino acid can be incorporated into the product.

3.1.1

Functionalization via Csp3–H Bond Activation

Therefore, a wide range of methods has been developed to introduce aryl, alkyl, alkenyl or alkynyl substituents into the β-position of N-phthaloylated amino acids (Scheme [12]), but most protocols are somehow limited to this protecting group.[52] [55] There are only very few examples with other protecting groups,[56] which cannot be used with amino acid derivatives containing an N–H functionality, probably because primary amides can act as a directing group themselves.[57] Recently, Qin et al. reported the β-arylation of N-benzylated Boc-protected non cyclic amino acids.[58]

Zoom Image
Scheme 12 β-Functionalization of N-phthaloylated amino acids

By far the best investigated examples are highly stereoselective β-functionalizations of N-protected cyclic amino acids such as proline or pipecolic acid.[59] During our work on peptide modifications we were able to show that the palladium-catalyzed stereoselective β-arylation of phenylalanine, proline and pipecolic acid containing peptides was especially suitable for this purpose.[60] The reactions proceed without epimerization of the stereogenic centers in the peptide chain, and the syn-configured substitution products are formed exclusively (Scheme [13]). If suitable functionalized aryl iodides are introduced, subsequent cross-coupling reactions can be conducted for further modifications. The 8-aminoquinoline (AQ) directing group[61] can easily be removed afterwards, allowing the extension of the peptide chain at the C-terminus.

Zoom Image
Scheme 13 Stereoselective β-functionalization of proline and pipecolic acid peptides

Very recently, we demonstrated that the synthetic potential of this protocol could be increased significantly by carrying out these modifications with N-methylated amino acids and peptides, which are structural motifs widely found in natural products (Scheme [14]).[62] As an application, we synthesized the two cyclopeptide alkaloids abyssenine A and mucronine E.

Zoom Image
Scheme 14 β-Functionalization of N-methylated amino acids and peptides

# 3.1.2

Functionalization via Csp2–H Bond Activation

In contrast to the described Csp3–H bond activation, which requires a directing group either at the N- or the C-terminus of a peptide to direct the functionalization to a certain position, amino acids with aromatic or heteroaromatic side chains are privileged structures for Csp2–H bond activation. Therefore, the proteinogenic amino acids phenylalanine, tyrosine, tryptophan and histidine have been mostly used for side-chain functionalizations, but in principle, other heterocyclic side chains can also be modified. Besides the often used palladium complexes,[63] rhodium-[64] and iridium catalysts[65] also provide good results.

An interesting application of this protocol was reported by Elhammer et al. using an Ir-catalyzed Csp2–H borylation as a key step in the synthesis of a library of aureobasidin A derivatives (Scheme [15]).[66] Interestingly, in the reaction with [Ir(cod)Cl2]2 and B2pin2, only one of the two phenylalanine residues was borylated selectively, probably due to a better accessibility. A mixture of meta- and para-borylated derivatives was obtained, which could then be subjected to iodination followed by Suzuki–Miyaura coupling. In this manner, a range of derivatives could be synthesized, which were tested for their biological activity.

Zoom Image
Scheme 15 C–H functionalization of aureobasidin A followed by Suzuki–Miyaura coupling

By far the most reported examples so far are on the modification of tryptophans, which can be functionalized selectively at the 2-position.[67] Albericio and Lavalla et al. reported the palladium-catalyzed direct 2-arylation of tryptophans using microwave irradiation. The same protocol could also be applied for the racemization-free modification of peptides under rather mild conditions (Scheme [16]). Except for methionine, all other amino acids were tolerated in the peptide chain.[68]

This approach can also be applied in the synthesis of cyclic[69] and bicyclic peptides (Scheme [16]).[68b] Instead of aryl iodides, aryl diazonium salts[70] or arylboronic acids in combination with Cu(OAc)2 [71] as a cocatalyst can be used.

Zoom Image
Scheme 16 Pd-catalyzed modifications of tryptophans

Ackermann et al. investigated the application of alternative catalysts for tryptophan modifications. Good results were obtained with Ru(II) complexes, which activate only N-2-pyridyl (2-py) substituted tryptophans, while free tryptophans are not affected (Scheme [17]). This allows for the epimerization-free regio- and chemoselective functionalization of peptides, while a wide range of other functionalities are tolerated.[72] Besides the commonly used noble metals, cheaper and especially less toxic transition metals, such as manganese, perform the reaction smoothly.[73] They were found to be particularly suitable for the introduction of alkyne substituents, and the results obtained are comparable to those of gold-catalyzed reactions.[74] The manganese-catalyzed alkynylation was also used in cyclization reactions and the best results were achieved with a 2-pyrimidyl (2-pym) directing group on the tryptophan nitrogen (Scheme [17]).

Zoom Image
Scheme 17 Ru- and Mn-catalyzed modifications of tryptophans

#
# 3.2

On Peptide Tryptophan Syntheses

Although the indole ring system is a privileged motif for modifications, all these reactions occur at the more electron-rich 5-membered ‘pyrrole’-ring, while variations at the 6-membered ‘phenyl’-ring are a non-trivial issue. To gain access to natural products with substituents on the larger ring, a de novo synthesis of the indole ring system is required. Several tryptophan syntheses have already been described so far.[75] However, most of them require harsh conditions (thermal, basic or acidic), limiting their applicability to the late-stage modification of peptides. Chen et al. described an indole synthesis from aldehydes and o-iodoanilines under Pd catalysis, which proceeds via enamine formation and subsequent intramolecular Heck cyclization.[76] This indole formation has already been used by several groups for the construction of indoles from δ-oxoamino acids.[77] However, the first tryptophan incorporation into peptides using this process was described by Suh et al. during their synthesis of ohmyungsamycins A and B (Scheme [18]).[78] They subjected a N-methylated homoallylglycine tripeptide to a Lemieux–Johnson oxidation to obtain an aldehyde, which was then reacted with 2-iodo-3-methoxyaniline under Pd catalysis to yield the corresponding tryptophan derivative, which was then transformed in a convergent synthesis into the ohmyungsamycins, leading to the structural revision of ohmyungsamycin B.

Zoom Image
Scheme 18 On peptide synthesis of the substituted tryptophan of ohmyungsamycin A

Based on our work on Pd-catalyzed allylic alkylations of peptide enolates with stannylated allylic substrates, we aimed to develop synthetic routes towards a series of tryptophan-containing natural products. Keramamides A[79] and L,[80] isolated from marine sponges, were chosen as synthetic targets since these molecules only differ in the substitution pattern of the l-tryptophan moiety and little was known about their biological properties thus far. During the synthesis of these natural products we could, however, not apply the allylic alkylation strategy for the installation of the stannylated residue. An (S)-configured tryptophan is present in the natural products, but the adjacent (S)-amino acids would guide the alkylation towards the (R)-configuration, as outlined above. Thus, we envisioned a strategy based on late-stage hydrostannation of a propargylic residue.[81] The propargylglycine-containing cyclic peptide was prepared by standard methods and was subjected to Ru-catalyzed hydrostannation[82] to access the requisite stannylated residue (Scheme [19]). This reaction proceeded smoothly and provided the stannylated peptide regioselectively.

For the Stille coupling with the corresponding 2-iodoanilines we had to choose milder conditions than before in order to avoid undesired cyclization in the urea side chain. Using Fürstner’s conditions[82] (CuTC and Ph2PO2NBu4), we obtained the coupling product in a satisfactory yield. Azidation, followed by photochemical nitrene insertion, delivered the desired tryptophan moiety and subsequent cleavage of the methyl ester led to the natural product. Employing this method, we obtained both keramamides A and L from a common precursor. Having established this synthetic route, we were able to show that the configuration of the lysine unit in both keramamides A and L were originally assigned incorrectly. Furthermore, our sequence for the late-stage installation of indoles allowed for the introduction of different substituents on the tryptophan scaffold and the synthesis of derivatives for SAR studies. Very recently, we applied the same protocol to the synthesis of the structurally related mozamide A, and again the originally proposed structure had to be revised.[83]

Zoom Image
Scheme 19 Synthesis of keramamide L via late-stage modification

#
# 4

Conclusion

The highlighted examples demonstrate the synthetic potential of peptide modifications for the construction of non-proteinogenic amino acids via peptide backbone modifications. The introduced functionalities often allow for further modifications by manifold side-chain derivatizations. In particular, chelated ester enolates have emerged as useful intermediates for the introduction of numerous substituents at the α-position. The diastereoselectivity of the utilized reaction can generally be controlled by the chiral peptidic backbone. The methods highlighted in the section on side-chain modifications clearly demonstrate the potential of site-selective derivatization methods to introduce new substituents or functionalities into a given peptidic natural product.


#
#

Acknowledgment

We thank Saarland University and Deutsche Forschungsgemeinschaft (DFG) for continuous support of our work, the Fonds der Chemischen Industrie for a Ph.D. fellowship for P. Servatius, and all the highly engaged coworkers for their outstanding contributions.



   Permissions and Reprints All articles of this category
Zoom Image
Zoom Image
Zoom Image
Zoom Image
Scheme 1 Regio- and diastereoselective modification of cyclosporin A according to Seebach et al.
Zoom Image
Scheme 2 Generation and reactions of chelated amino acid ester enolates
Zoom Image
Scheme 3 [3,3]-Sigmatropic rearrangement of chelated ester enolates
Zoom Image
Scheme 4 Stereoselective peptide Claisen rearrangements
Zoom Image
Scheme 5 Peptide Claisen rearrangements of chiral allyl esters
Zoom Image
Scheme 6 Cyl-1 derivatives via peptide enolate Claisen rearrangements
Zoom Image
Scheme 7 Stereoselective Pd-catalyzed allylic alkylations of dipeptides. Reaction conditions: (1) 3.5 equiv LHMDS, 1.2 equiv ZnCl2, THF, –78 °C; (2) 0.7 equiv allylic substrate, 1.4 mol% [allylPdCl]2, 6.3 mol% PPh3, THF, –78 °C to rt.
Zoom Image
Scheme 8 Pd-catalyzed allylic alkylation in the synthesis of trapoxin A
Zoom Image
Scheme 9 Synthesis and reactions of stannylated peptides
Zoom Image
Scheme 10 Allylic alkylations of peptide amides
Zoom Image
Scheme 11 Synthesis of miuraenamides via peptide modification
Zoom Image
Scheme 12 β-Functionalization of N-phthaloylated amino acids
Zoom Image
Scheme 13 Stereoselective β-functionalization of proline and pipecolic acid peptides
Zoom Image
Scheme 14 β-Functionalization of N-methylated amino acids and peptides
Zoom Image
Scheme 15 C–H functionalization of aureobasidin A followed by Suzuki–Miyaura coupling
Zoom Image
Scheme 16 Pd-catalyzed modifications of tryptophans
Zoom Image
Scheme 17 Ru- and Mn-catalyzed modifications of tryptophans
Zoom Image
Scheme 18 On peptide synthesis of the substituted tryptophan of ohmyungsamycin A
Zoom Image
Scheme 19 Synthesis of keramamide L via late-stage modification