Chemoenzymatic and template-directed synthesis of bioactive macrocyclic peptides

Abstract

Non-ribosomally synthesized peptides have compelling biological activities ranging from antimicrobial to immunosuppressive and from cytostatic to antitumor. The broad spectrum of applications in modern medicine is reflected in the great structural diversity of these natural products. They contain unique building blocks, such as d-amino acids, fatty acids, sugar moieties, and heterocyclic elements, as well as halogenated, methylated, and formylated residues. In the past decades, significant progress has been made toward the understanding of the biosynthesis of these secondary metabolites by nonribosomal peptide synthetases (NRPSs) and their associated tailoring enzymes. Guided by this knowledge, researchers genetically redesigned the NRPS template to synthesize new peptide products. Moreover, chemoenzymatic strategies were developed to rationally engineer nonribosomal peptides products in order to increase or alter their bioactivities. Specifically, chemical synthesis combined with peptide cyclization mediated by nonribosomal thioesterase domains enabled the synthesis of glycosylated cyclopeptides, inhibitors of integrin receptors, peptide/polyketide hybrids, lipopeptide antibiotics, and streptogramin B antibiotics. In addition to the synthetic potential of these cyclization catalysts, which is the main focus of this review, different enzymes for tailoring of peptide scaffolds as well as the manipulation of carrier proteins with reporter-labeled coenzyme A analogs are discussed.

FIG. 1.

A selection of non-ribosomally synthesized peptides. Characteristic structural features that confer rigidity to the peptide backbone are highlighted.

FIG. 2.

Diversity of acidic lipopeptide antibiotics. At least 27 compounds have been characterized so far. A54145 is produced by Streptomyces fradiae, daptomycin is produced by Streptomyces roseosporus, CDA is produced by Streptomyces coelicolor, and friulimicins and amphomycins are derived from Actinoplanes friuliensis. Conserved acidic residues are indicated in red, and d-configured/achiral residues at equivalent positions are highlighted in blue. The cyclization site is indicated by shading.

FIG. 3.

Comparison of enzymatic subunits of the daptomycin (DptA, DptBC and DptD), A54145 (LptA, LptB, LptC, and LptD), and CDA (CDAI, CDAII and CDAIII) NRPSs that are responsible for the synthesis of the respective peptide cores. Parts of the peptide cores that are synthesized by their dedicated enzymatic subunits are surrounded by red dotted lines. The modules indicated in red and white are subdivided into catalytically independent domains responsible for substrate recognition/activation. FA, fatty acid; hAsn, 3-hydroxyasparagine; mGlu, 3-methylglutamate; Sar, sarcosine; omAsp, 3-methoxyaspartate; Orn, ornithine; Kyn, kynurenine.

FIG. 4.

Chemical principles of nonribosomal peptide synthesis. Domains in action are indicated in red and the respective crystal structures are shown above. First, the A-domain specifically recognizes a dedicated amino acid and catalyzes formation of the aminoacyl adenylate under consumption of ATP. Second, the activated aminoacyl adenylate is tethered to the free thiol group of the PCP-bound phosphopantetheine (ppan) cofactor. Third, the C-domain catalyzes peptide elongation. Here, the nucleophilic amine of the acceptor substrate nucleophilically attacks the electrophilic thioester of the donor substrate (a, acceptor site; d, donor site). The crystal structure of the A-domain is derived from the Phe-activating A-domain (PheA) of the first module of gramicidin S synthetase of B. brevis (22). The NMR-structure of the PCP is derived from the third module of the B. brevis tyrocidine synthetase (141), and the C-domain is derived from the crystal structure of VibH, a stand alone C-domain of the V. cholerea vibriobactin synthetase (60).

FIG. 5.

Proposed mechanism of the lipidation of daptomycin. Decanoic acid is activated as decanoyl-adenylate under the consumption of ATP (1). This step is catalyzed by DptE. The fatty acid is transferred on to the ppan cofactor of the putative acyl-carrier protein DptF (2). DptF interacts with the starter C-domain of DptA, which catalyzes the subsequent acylation of Trp₁ (3). Finally, DptF is released (4).

FIG. 6.

Proposed mechanisms underlying amino acid epimerization. (A) The E-domain converts the PCP-tethered aminoacyl substrate into a d/l equilibrium. The stereoselective donor site (d) of the C-domain of the downstream module uses only the d-configured amino acid for subsequent peptide elongation. (B) In some cases, an external racemase (Rac) catalyzes the racemization of a freely diffusible amino acid. Here, a stereoselective A-domain is the determinant that activates solely the corresponding d-enantiomer. (C) d-Amino acid incorporation into arthrofactin, syringomycin, and syringopeptin is catalyzed by a new type of condensation domain (C/E-domain). Epimerization does not take place unless the PCP downstream of this C/E-domain is loaded with the dedicated amino acid. It is not yet known whether the epimerization reaction is reversible or not. After epimerization of the upstream aminoacyl/peptidyl thioester, the C/E-domain mediates the elongation of the peptidyl chain with ^DC_L chirality.

FIG. 7.

Macrocyclization strategies. (A) The tridecapeptidyl chain of daptomycin tethered to the ppan cofactor of the most downstream PCP is transferred to an active site serine of the TE-domain forming the acyl-O-TE intermediate. Subsequent product release is carried out by the attack of an internal nucleophile (l-Thr₄) on the oxoester bond to give the cyclic branched macrolactone. (B) Head-to-tail macrolactamization of the undecapeptide cyclosporine is catalyzed by the most downstream C-domain. Mechanistic details are still unknown. (C) Macrocyclic imine formation of nostocyclopeptide. First, the C-terminal residue of the ppan-tethered peptide is reduced by the action of an NAD(P)H-dependent R-domain to give an aldehyde, which is intramolecularly captured by the N terminus to give a macrocyclic imine. Future research will show if the R-domain also mediates this final macrocyclization step.

FIG. 8.

The experimental design for the study of excised cyclases exemplified for Tyc-TE. First, the NRPS multienzyme machinery for tyrocidine synthesis is replaced by solid-phase peptide synthesis. Second, the TE-domain is used as an excised enzyme for in vitro peptide cyclization. Third, recognition of the artificial substrate by the excised cyclase is ensured by the phosphopantetheine cofactor mimic SNAC (highlighted by shading).

FIG. 9.

Dynamic kinetic resolution of a streptogramin B SNAC substrate with Phg at the C terminus, which is prone to in situ substrate racemization. The resulting two diastereomers are able to acylate the active Ser residue of SnbDE TE resulting in two different peptidyl-O-TE intermediates. However, only the peptidyl-O-TE intermediate with the l-Phg-configuration is able to undergo cyclization to the natural product with l-Phg configuration.

FIG. 10.

Chemoenzymatic synthesis of novel bioactive compounds by excised TE-domains. Substrates can be presented to the cyclase either bound to an artificial solid support (PEGA resin) or by soluble thioester leaving group.

FIG. 11.

Chemoenzymatic approaches to glycopeptide antibiotics. (A) A chemically synthesized propargylglycine-substituted tyrocidine SNAC substrate is cyclized by Tyc TE. The resulting alkyne-containing macrolactam is then conjugated to an azido sugar to produce a glycosylated cyclic peptide by using copper(I)-catalyzed [3 + 2] cycloaddition. (B) A glycosylated amino acid is incorporated into a linear tyrocidine SNAC thioester via SPPS. Tyc TE-catalyzed head-to-tail-cyclization then produces a glycosyl-tyrocidine analog.

FIG. 12.

Posttranslational modification of PCPs. (A) Ppan transferases, such as Sfp from B. subtilis, transfer a 4′-phosphopantetheine residue from CoA to a strictly conserved serine residue of apo-PCP, creating the holo-PCP and 3′,5′-ADP. (B) CoA-reporter analogs are created through a Michael addition of the free thiol group in CoA across the double bond of reporter-linked maleimide. Subsequent Sfp-catalyzed reporter labeling of PCP fusion proteins may be used to label the target protein with small molecules, i.e., fluorophores and affinity tags.

FIG. 13.

In vivo tagging of carrier protein fusion VibB within E. coli. To allow cellular uptake, a cell-permeable fluorophore-labeled pantetheine analog was used. This reporter-labeled pantetheine is converted to reporter-labeled CoA by CoAA, CoAD, and CoAE in vivo. This process is followed by reaction of coexpressed Sfp to yield fluorophore-labeled VibB. CP, carrier protein; ICL, isochorismate lyase.

FIG. 14.

C-methylation of nonribosomal peptides. (A) Proposed mechanism of C-methylation of l-glutamate in CDA. l-Glu is converted into 2-oxoglutarate in the presence of a transaminase, which uses pyridoxal phosphate (PLP) as a cofactor. After deprotonation by an as yet unknown base (B), GlmT catalyzes the stereoselective methylation of the resulting enolate by using SAM. The methylated product 2-oxo-3-methylglutarate is finally converted to l-threo-3-methylglutamate through consumption of pyridoxamine phosphate (PMP). (B) β-Methylaspartate formation in friulimicin. The conversion of l-Glu into l-threo-3-methylaspartate is mediated by a GlmA/GlmB mutase complex, which uses B₁₂ as a cofactor.

FIG. 15.

Oxidative cross-linking of phenol side chains occurs only on NRPS-bound peptide chains. (A) Chemical structure of vancomycin. The order of oxidative cross-linking mediated by OxyB, OxyA, and OxyC is indicated. (B) A chemically synthesized balhimycin-like hexapeptide CoA thioester is loaded onto a PCP using Sfp. Subsequent incubation with OxyB leads to cross-linking between rings C and D.

FIG. 16.

Proposed mechanism for Ebony-catalyzed binding of primary amines to β-alanine. After posttranslational modification of the PCP by a dedicated ppan transferase, the A-domain of holo-Ebony specifically activates β-alanine as β-alanyl-AMP and subsequently transfers it onto the ppan cofactor. To generate the peptidoamine, the AS-domain presumably catalyzes the attack of the nucleophilic primary amine onto the electrophilic thioester of the PCP-bound β-alanine. Biochemical studies revealed that various primary amines are suitable for peptidoamine formation.

FIG. 17.

COM domain swapping (origin of the proposed COM domains are in parentheses). Native in trans communication between NRPS modules of tyrocidine synthetase is indicated by shading. Productive interaction between these partner modules can also be enforced by the presence of matching COM domains derived from different modules [TycA(B3)/(C1)TycB1 and TycB3(A)/(B1)TycC1]. Likewise, cross talk can be induced between nonpartner NRPSs [TycB3/(C1)TycB1, TycA/(B1)TycC1, TycB3(A)/TycB1, and TycA(B3)/TycC1].