Cyclic peptides can be prepared either chemically or biosynthetically. The former is considered a more traditional method to produce these molecules. Chemical methods provide greater control over the architecture of the backbone, side-chains and non-natural modifications. They can be a major challenge, particularly for the larger cyclic peptides, with issues of yield, epimerisation and scale-up being common. They also often require more complex protecting group strategies. Biosynthesis, whether through enzymatic systems or engineered cell factories, can avoid many of these issues as it uses enzymes that were selected for their regio- and stereospecificity and often work in aqueous conditions. The resulting cyclic peptides can be more difficult to produce chemically, but are less common. Biosynthetic methods also include those using non-ribosomal peptide synthetase (NRPS) and ribosomally synthesized cyclic peptides. However, these processes can lack substrate promiscuity and only have a limited ability to accept non-natural substrates. The pathway engineering of megasynthases or ribosomal machinery is non-trivial. Current efforts have seen a blurring of the lines, with many chemically inspired efforts in the biomimetic space, and many enzymatic efforts in the chemoenzymatic space. The most successful approaches are likely to lie in the middle.
Chemical cyclization processes, both non-catalytic and catalytic, can be divided into subtypes, such as direct (traditional) amide bond formation using phosphonium or aminium activated reagents, chemoselective ligation chemistries, metal-catalyzed couplings, and more recently photoredox-mediated cyclizations. Biological cyclizations use either NRPS-module type logic to form the cyclic product or protein translation to form the linear precursor with a cyclase or an appropriate modifying enzyme, like lanthipeptide dehydratase or macrocyclase, to cyclize with high precision. In this regard, chemical processes tend to allow a broader range of modifications and derivatives to be made but are limited by dilution-controlled kinetics, competing oligomerization, and epimerization at the activated site, especially in solution. Biosynthetic processes result in much higher atom efficiency, turnover, and selectivity (e.g. enantio- or diastereoselectivity) but remain dependent on fermentation conditions and the downstream stability of the enzymes involved, as well as being much more challenging to incorporate nonnative substrates without a dramatic loss in efficiency. A growing field of chemoenzymatic approaches has emerged that can be used to chemically synthesize precursors that are enzymatically cyclized in a single reaction.
Fig. 1 Cyclodipeptide synthesis by the NRPS GliP: the synthesis of cyclo(L-Phe-l-Ser), an intermediate chemical in gliotoxin biosynthesis.1,2
Chemical methods comprise direct backbone cyclization, native chemical ligation, aldehyde-based ligations, bioorthogonal reactions and disulfide formation. Chemical methods allow for exquisite control of cyclization and can be performed in vitro. Chemical methods are precise and flexible but less efficient.
The most widely practiced method for macrocyclization of peptides is the head-to-tail macrolactamization approach which essentially involves an intramolecular nucleophilic attack of the free N-terminus on an activated C-terminal carboxylate. The activation step for the carboxylic acid group is normally achieved by using a modern coupling reagent like HATU or PyBOP which act as carboxyl-activating agents to produce a more reactive O-acyluronium or phosphonium intermediate. The resulting linear peptide precursor, either in resin bound or soluble form, is subjected to activation of the C-terminus in the presence of a tertiary amine base and this activation step is slowly neutralized to generate the free amine for intramolecular capture. Mechanistically, this method would first lead to the formation of an activated ester which can be protonated to exist as its equilibrium protonated form. Under diluted conditions, the nearby amine attacks the activated carbonyl leading to the displacement of the leaving group and the closure of the macrocycle. In reality, this deceptively simple approach can suffer from numerous side reactions that can lead to a very low overall yield. Epimerization of the C-terminal residue is a common problem due to formation of an oxazolone intermediate during activation, where the oxazolone is capable of abstracting a proton from the chiral carbon atom resulting in an epimerization before cyclization can occur. This side reaction is more common when a sterically bulky residue like valine or isoleucine is present at either terminus, as this can make the cyclization slower, giving more time for epimerization to occur. Low overall yields are also common because of not only epimerization but also intermolecular oligomerization, particularly when synthesizing smaller rings, where the linear precursor is more likely to dimerize or polymerize rather than cyclize under conditions of high effective molarity to drive the intramolecular reaction. Additionally, poor solubility of protected peptide segments in the organic solvents needed for coupling can lead to incomplete activation and slow cyclization, and if deprotection is not complete, then other nucleophilic side-chains may be available for acylation. The high dilution necessary to minimize oligomerization also works against the slow reaction rates, making this method less suitable for large scale production despite its wide-spread use in academic labs.
Side-chain-to-side-chain cyclization can provide an alternative approach to head-to-tail macrolactamization, affording topological control through the orthogonal manipulation of side-chain functional groups. It can also sometimes avoid some of the inherent limitations of termini-directed cyclization. In general, this approach is achieved through the use of orthogonal protecting groups with mutually orthogonal deprotection chemistries so that the side-chain functional groups of interest can be deprotected while the rest of the peptide remains protected. For example, a lactam bridge may be formed in a manner analogous to a structural constraint present in several peptide hormones and toxins. In this example, a lysine side-chain amine and aspartate or glutamate side-chain carboxylate are to be connected by an amide bond. Orthogonal protecting groups like allyl for the carboxylate and ivDde or another base-labile protecting group for the amine, can then be deprotected in a stepwise fashion and the resulting free amine and acid intramolecularly coupled without affecting the other protected residues. Disulfide bonds are another popular side-chain cyclization motif and rely on the unique redox reactivity of cysteine thiols. This is especially useful for defining the tertiary structure of many natural product peptides such as conotoxins and antimicrobial defensins, and is generally performed by oxidative folding. The partially or fully deprotected peptide is then subjected to oxidizing conditions (usually air, dimethyl sulfoxide, or a glutathione redox buffer) to selectively form the S–S bond. The advantage of this methodology is that the oxidation of thiols can be performed in the presence of unprotected amines, acids, and other nucleophiles as long as pH and redox conditions are controlled. More advanced versions of this chemistry also make use of thiol-ene or thiol-yne click chemistry to form a C–S bond rather than a disulfide. Side-chain linkers have also been used to cyclize the peptide backbone. Here, the peptide chain is attached to resin through a side-chain linker which, upon cleavage, leaves both termini free to be cyclized while the linker to the resin mimics dilution to minimize oligomerization. This method can be especially useful for the construction of bicyclic or polycyclic systems since multiple orthogonal handles are often present and can be sequentially "clicked" into rings through different chemical reactivity.
NCL is a form of chemoselective, traceless coupling that has had a major impact in the field of peptide macrocyclization. NCL reactions can be performed under mild, aqueous conditions with unprotected side chains and therefore avoid many of the issues with traditional amide coupling methods. The strategy involves two steps: thiol-thioester exchange followed by acyl transfer. In the first step, a peptide thioester reacts with another peptide which contains an N-terminal cysteine, resulting in an acyl transfer from the alkylthiol to the cysteine. In the second step, the S-to-N acyl shift results in the formation of a native amide bond and the regeneration of the cysteine thiol. Thioester exchange is reversible and typically catalyzed with an external thiol to drive equilibration. The S-to-N shift is irreversible due to the thermodynamic stability of the amide bond and is considered the signature step in NCL. The cyclization version of NCL is generated using a single peptide containing both a C-terminal thioester and an N-terminal cysteine, making the reaction intramolecular. The development of this method has allowed the synthesis of many complex macrocycles, including cyclotides, a family of over 30 residue, cystine-knot peptides from plants that are known for their stability and diverse biological activities. NCL is performed under slightly acidic to neutral conditions and the reaction rate can be increased with aryl thiols, which can catalyze thioester exchange. In the cyclotide synthesis, the macrocyclization framework helps preorganize the termini for more efficient ring closure. In addition to cyclotides, NCL has been used to synthesize many conotoxin analogues, which allows control over disulfide pairing and to engineer antimicrobial peptides, in which backbone cyclization can make the molecule more proteolytically stable, but still membrane-permeabilizing. NCL has also been adapted for on-resin strategies, in which peptide thioesters are prepared on solid support using safety-catch linkers, followed by cyclization, global deprotection and purification. This strategy allows for NCL-mediated cyclization of large and densely functionalized macrocycles, which resemble natural products and provide handles for the incorporation of noncanonical amino acids or post-translational modifications through convergent fragment assembly.
Biosynthetic and enzymatic cyclization is an alternative strategy to chemical synthesis for the production of cyclic peptides. Ribosomal and nonribosomal peptide synthetic pathways, or individual enzymes, may be used in a variety of host systems, in vivo or in vitro, for the production of cyclic peptides. These methods offer advantages in terms of cyclization efficiency, and stereoselectivity. The use of biosynthetic or enzymatic methods requires consideration of a number of factors, such as host organism, peptide complexity, yields, and cyclization chemistry.
The nonribosomal nature of NRPSs means that the strategies for macrocyclization differ from those used in ribosomal peptide pathways. NRPSs are megasynthases made up of modules. Each module consists of at least three domains: adenylation, thiolation, and condensation. Each of these domains selects and activates its specific aminoacyl building block, which is covalently linked to a phosphopantetheinyl arm and transferred to the nascent peptide chain. Head-to-tail macrolactamization is carried out by thioesterase (TE) domains through a nucleophilic attack of the N-terminus onto the thioester at the C-terminus, while side-chain-to-backbone cyclization is accomplished by cyclization domains through capture of nucleophilic side-chains in intramolecular reactions. In TE-mediated cyclization, an acyl-O-to-alkyl-S cleavage is catalyzed by a catalytic triad, which releases the product. The system is not particularly discriminating in terms of peptide length or sequence, although substrate range narrows in the presence of unnatural residues that disrupt the normal binding pocket. Siderophore biosynthesis by Cy domains installs an oxazoline or thiazoline heterocycle by cyclodehydration of a serine/threonine or cysteine residue, respectively. Non-collinear assembly and domain skipping can be combined with iterative use of modules to create polyketide-peptide hybrids. In RiPP pathways, a precursor peptide is first translated on ribosomes. It usually has an N-terminal leader sequence and a C-terminal core, which becomes the macrocycle. The leader sequence provides a handle for the post-translational modification machinery to selectively modify the core. Cyclization can be achieved by proteolytic removal of the leader, followed by formation of a peptide bond between the N- and C-termini (cyclotides are modified by asparaginyl endopeptidases in this way), or by enzymatic crosslinking of side-chains to create a backbone-to-side-chain macrocycle (lanthipeptide synthetases form thioethers by Michael addition of cysteine to dehydrated residues).
The biocatalysts to have been co-opted for peptide macrocyclization (including through conjugation to a protein) include sortase A, butelase 1 and subtiligase. These represent different mechanistic classes and have differing and generally complementary preferences and kinetic characteristics:
(1) Sortase A is a transpeptidase enzyme found in Gram-positive bacteria that recognizes a specific pentapeptide motif, LPXTG (found in surface proteins), and, upon cleavage of the thioester bond between threonine and glycine residues, forms an acyl-enzyme thioester intermediate. The sortase A-bound thioester is resolved upon nucleophilic attack by the amine group on an oligoglycine nucleophile. The same chemistry has been adapted for cyclization purposes by fusing the recognition sequence to the C-terminus and the triglycine recognition sequence to the N-terminus of a linear peptide. Upon addition of sortase A, intramolecular ligation occurs to generate the cyclic product. Sortase A requires Ca2+ for coordination of the active site and thus assembly, and has relatively low turnover numbers. It has been possible to evolve sortase A variants to recognize APXTG and a variety of other sequences, but these have lost some of their catalytic activity.
(2) Butelase 1 is a ligase-type asparaginyl endopeptidase found in plants. It cyclizes peptides in a reverse-proteolysis reaction in which the C-terminal Asn-His-Val (NHV) recognition motif is recognized and cleaved after the asparagine residue to form a thioester intermediate that is cyclized by nucleophilic attack from the N-terminal amine. As a result, butelase 1 can rapidly cyclize peptides with the N-terminal amine to C-terminal NHV tag with virtually no epimerization. It works over a very wide pH range, and has one of the highest turnover numbers of the currently known enzyme cyclization reactions. The requirement for the NHV recognition tag is a potential limitation as it requires the attachment of this tripeptide to the target sequence, which may be a burden in the product, but it is possible to use an engineered version of OaAEP1 (OaAEP1-C247A) to avoid the acid-activation requirement for the wild-type zymogen. This engineered enzyme can be used directly in neutral buffers and is also useful for protein ligation and site-specific labeling.
(3) Subtiligase is a serine protease in the subtilisin family that has been rationally engineered to avoid the kinetic trap of acyl-enzyme intermediate stability. This has been accomplished by destabilizing the transition state to hydrolysis relative to aminolysis through active-site remodeling. As a result, the enzyme will favor nucleophilic attack from the N-terminal amine over hydrolysis, and thus peptide ligation can be directed to occur in solution via the intermediate thioester. In contrast to sortase A and butelase 1, subtiligase is much more tolerant of substrates and will accept, in principle, any N-terminus (albeit with variable selectivity), and has sufficient selectivity for the more accessible C-terminal thioesters. This makes it especially useful for cyclizing compound libraries with minimal redesign of the biocatalyst. Its catalytic efficiency is lower than some optimized natural ligases. In particular, hydrolysis is a competing side reaction with cyclization and the reaction conditions need to be well-controlled.
Engineering host organisms for in vivo cyclization of peptides involves manipulating the cellular machinery of microorganisms or cells to produce cyclic peptides. This can be achieved by introducing genes encoding enzymes or peptide precursors, as well as optimizing the intracellular environment for efficient cyclization. Host organisms such as bacteria, yeast, and mammalian cells can be engineered for this purpose. Factors such as peptide complexity, required post-translational modifications, and scalability of production may influence the choice of the host system. By leveraging the biosynthetic capabilities of these engineered hosts, it is possible to produce cyclic peptides in a scalable and efficient manner, providing a valuable approach for the development of new therapeutic agents and bioactive molecules.
Comparison of chemical and biosynthetic cyclization strategies in several technical aspects: Chemical strategies enable the use of a wider variety of building blocks such as D- and β-amino acids and N-methylated residues. At small scale, synthetic methodologies can be more rapid and amenable to iterative chemical optimization, but are dependent on the use of expensive reagents and chromatographic methods for purification. These factors, combined with the use of organic solvents, make chemical synthesis less scalable and less sustainable. In contrast, biosynthetic methods use aqueous conditions, renewable carbon sources and operate under enzyme catalysis which, in general, provides higher atom economy and less waste. These methods, however, require production and purification of large protein catalysts, are limited by precursor availability, and generally have less tolerance for non-native substrates, which can limit substrate scope or require significant engineering efforts. Kinetics of these reactions also often differ, with chemical reactions often being slower, less selective, and suffering from oligomerization at higher concentrations while enzyme-catalyzed reactions are often fast, specific, and do not require dilute conditions but are prone to product inhibition and proteolysis. Regulation of these products can also differ, with chemical synthesis methods being more familiar to the highly controlled, well-defined purification methods for therapeutic agents while biologically produced products require fermentation-specific considerations such as metabolite profiling and immunogenicity concerns. Finally, chemical synthesis is often preferred for small quantities of material in discovery settings, while biosynthesis is often preferred at very large scales due to the advantages of fermentation-based production.
| Parameter | Chemical | Biosynthetic |
|---|---|---|
| Precision | Moderate | High |
| Scalability | High | Moderate |
| Complexity | High | Low |
| Application | Synthetic chemistry | Biotech & pharma |
An attractive alternative to the direct enzymatic synthesis of macrocycles is the combination of chemical and enzymatic methods, combining the advantages of the two approaches and avoiding their respective pitfalls. Linear precursors are synthesized on solid- or solution support using purely chemical procedures, and then treated with purified cyclization enzymes or crude lysates that perform the final cyclization step with high regio- and stereo-specificity. The motivation for this approach is clear: chemical synthesis provides ready access to non-canonical amino acids, post-translational mimics, and backbone modifications which are not accessible via ribosomal or non-ribosomal synthetic systems, while the subsequent enzymatic step guarantees that cyclization occurs without epimerization and oligomerization which are common problems for macrolactamization reactions. Thioester-substrates which resemble the natural phosphopantetheinyl carrier moiety, such as the simple N-acetylcysteamine (SNAC) conjugates can be used as substrates for the thioesterase domains of NRPSs. These T-domains, when removed from the rest of the NRPS, remain catalytically competent, and can accept chemically synthesized peptide substrates with the appropriate native recognition motifs and perform head-to-tail cyclization with high efficiency and specificity, similar to the native megasynthase. This has allowed for the reprogramming of native non-ribosomal peptide platforms, where the chemical step introduces variations, and the enzymatic step provides the necessary topological constraints to induce bioactivity. This approach can be applied to glycosylated and lipopeptide macrocycles as well. In these cases, the carbohydrate moieties can be chemically elaborated on the linear precursor before enzymatic cyclization, generating a hybrid peptide/oligosaccharide. The ability of some cyclases to accept unnatural modifications in length, stereochemistry, and residues is a testament to the potential of this approach, but in each case requires an appropriate match of the chemical linker to the enzyme's active site.
Automation and AI Integration in Peptide Cyclization: In the realm of peptide cyclization, the future holds significant potential in the development of automated chemoenzymatic workflows and the integration of AI-driven peptide design. Automated systems can facilitate the cyclization process, minimizing human intervention, reducing errors, and increasing throughput. The incorporation of artificial intelligence technologies can revolutionize peptide design by predicting optimal cyclization conditions and designing novel peptide sequences with desired properties. By integrating automated peptide synthesis platforms with machine learning algorithms, researchers can streamline the discovery and optimization of cyclic peptides. This synergy between automation and AI integration promises to enhance the efficiency, precision, and speed of peptide engineering, opening new avenues for the development of novel therapeutic agents and bioactive molecules.
Chemical and biosynthetic cyclization each offer unique advantages—and leveraging them effectively requires deep technical expertise. Our platform is built to support researchers and development teams seeking high-performance cyclic peptides, regardless of complexity or scale.
We provide:
With a multidisciplinary team spanning synthetic chemistry, peptide engineering, and computational design, we help partners shorten development timelines and achieve superior macrocycle performance. If you are exploring new cyclization strategies or require reliable, high-quality cyclic peptide synthesis, our team is ready to collaborate.
References
