The increasing demand for new cyclic peptide-based molecules and nanostructures, and specifically for biomedical applications, has driven a strong interest in the development and improvement of their synthetic methods. In this sense, three different methodologies have been developed: the classical solution peptide synthesis (CSPS), the SPPS and the liquid-phase peptide synthesis (LPPS). Even though the first methodology was the first one applied, the long time and energy necessary to produce even short sequences is the major disadvantage of this synthesis method. However, CSPS has many advantages, including the possibility of isolating and characterizing intermediates at each step, and avoiding side reactions. This methodology is sometimes considered old-fashioned, but is still used for large-scale peptide synthesis. More specifically, CSPS is still an excellent alternative for the production of pharmaceutical peptides up to 22 amino acids.
Fig. 1 Current representative mainstream strategies for peptidomimetic PPI inhibitors design and optimization.1,2
Head-to-tail macrolactamization is the most obvious approach to construct a cyclic peptide but it is often a frustrating process in practice. The linear precursor needs to fold into an entropically unfavourable conformation to juxtapose N- and C-termini; furthermore, the activated C-terminal carboxylate is susceptible to epimerization, giving rise to diastereomeric by-products that are hard to separate. Intermolecular oligomerization is usually suppressed using high-dilution conditions (≤1 mM), but these conditions also severely decelerate the desired intramolecular process and cyclization often fails to reach full conversion, yielding isolated products in poor yields (< 30 %). Protected peptides tend to have poor solubility in aprotic coupling solvents, leading to heterogeneous reactions and further decreased efficiency. Diketopiperazine formation (especially with Pro or Gly at the C-terminus) and aspartimide rearrangement are two other side reactions that can deplete the starting material and worsen purification. These problems are tackled by a three-pronged approach: (i) pre-organization of the linear peptide by turn-inducing D-Pro or N-methylated residues to minimise the entropic penalty; (ii) application of low-epimerization coupling reagents (COMU, PyOxim) in combination with non-nucleophilic bases; and (iii) use of microfluidic reactors that create pseudo-dilution at gram scale routinely leads to cyclization yields >70 % with racemization < 1 %.
Peptide cyclization can be a formidable challenge with increasing peptide length (~15 residues) or in the presence of multiple disulfide bridges. At longer sequence length, polypeptides have an increased tendency to aggregate on resin or in solution via β-sheet or α-helix bundle formation that often results in a steric clash with the loop segment and prevents ring closure. A larger ring size also compounds the conformational entropy penalty, making the cyclization transition state less favorable. Furthermore, the increasing steric bulk of side chain protecting groups often negatively impacts the accessibility of reactive termini leading to slow coupling and high levels of truncated or oligomeric impurities. To address these issues we take a convergent fragment-condensation approach: the macrocycle is conceptually divided into two or three shorter fragments that are separately cyclized and then ligated via native chemical ligation or chemoselective click reaction under aqueous conditions. Oxidative folding with multiple disulfide bridges is carried out in a defined redox buffer with glutathione redox pairs to promote correct disulfide connectivity while avoiding intermolecular cross-linking. Use of microwave-assisted SPPS with pseudoproline dipeptides limits aggregation during chain elongation and in-line UPLC-MS monitoring allows real-time optimization of coupling equivalents to keep stepwise yields >99 %. Combined, these strategies have allowed the successful cyclization of 57-mer peptides and bicyclic scaffolds in overall isolated yields of >50 %.
The use of non-natural amino acids (ncAAs) is key to optimizing the potency, stability and membrane permeability of cyclic peptides, but their use comes with its own unique set of bottlenecks. Firstly, many ncAAs are not commercially available in Fmoc-protected form and need to be synthesized from scratch, with multi-step, multi-week custom syntheses slowing down timelines. Bulky side chains of ncAAs like cyclohexyl-alanine or fluorinated phenylalanines steric hinders coupling during SPPS resulting in incomplete acylation and deletion sequences. Sensitivity of some ncAAs to standard cleavage cocktails (azido- or alkyne-functionalized residues, for instance, decompose under prolonged exposure to TFA) necessitates orthogonal protection strategies or two-stage cleavage protocols. Enzymatic cyclization methods (sortase, butelase) require recognition motifs for which ncAA substitutions are not always tolerated, narrowing available route options. These challenges can be mitigated by investing in a dedicated ncAA pipeline: (i) an in-house ncAA catalogue of >250 Fmoc-ncAAs available on demand with scalable syntheses validated to GMP standards, (ii) coupling protocols at elevated temperatures (50 °C) using super-active reagents (HATU/HOAt) to push difficult acylations to completion, (iii) mild cleavage cocktails (TFA/TIS/H2O 95:2.5:2.5 v/v, 30 min) supplemented with scavengers to protect sensitive functional groups and (iv) enzymatic cyclization variants which have been engineered to accept ncAA-containing recognition sequences. With these in place, we are able to introduce multiple ncAAs into the same cyclic peptide without loss of synthetic efficiency or stereochemical integrity, enabling rapid SAR exploration across multiple chemical spaces.
Cyclic peptides are of high interest, because many of them can have increased cell permeability, improved metabolic stability, and a better pharmacokinetic profile. This means that there is a great demand in the field, since they could help with many problems that conventional peptides have, such as low stability, low bioavailability, off-target binding, and others. Hence, by being able to form stable, high-affinity, highly specific complexes with their targets, cyclic peptides are very suitable to increase the efficacy and decrease the side effects of drugs. This is why they are being considered in a very broad scope of potential applications: from small molecule inhibitors to various biologics. Additionally, cyclic peptides give the possibility to include non-natural amino acids, as well as other modifications, which further expands their functional space. One of the potential problems with using cyclic peptides in biomedical nanotechnologies is the robust strategies of their synthesis and design.
The selection of macrocyclization chemistry relies on both the peptide's sequence length and its ultimate biological application. Head-to-tail lactamization is conceptually straightforward (an amide bond linking the N- and C-termini) but becomes increasingly inefficient for > ~12 residues due to the entropically disfavoured "closed" conformation required of the peptide. To achieve high efficiency we often introduce turn-inducing elements (D-Pro, N-methylated residues) or switch to head-to-side-chain or side-chain-to-tail cyclization, where one terminus is connected to the side chain of an Asp, Glu, or Lys residue (which shortens the effective ring size, leading to higher yields). Side-chain-to-side-chain methods (e.g. disulfide, thioether, or triazole bridges) are used when a flexible N- or C-terminus is required for receptor binding. Disulfide bridges are rapidly formed under oxidative conditions but are unstable in cytosol, so we often supplement them with redox-stable, bio-isosteric amide RCM or CuAAC triazole linkages. A decision tree (ring strain calculation, solvent polarity, functional-group tolerance) helps in the selection process; for example, CuAAC is used for peptides bearing azide/alkyne ncAAs because the reaction is quantitative on-resin at 60 °C under microwave irradiation in 10 min and gives >90 % cyclization yield with minimal epimerization.
Fully automated modern synthesizers turn cyclic-peptide synthesis into a reproducible process, instead of an artisanal craft. CEM's Liberty PRO microwave peptide synthesizers perform all Fmoc-SPPS cycles at 90 °C which reduces coupling times down to 4 min with<0.1 % racemization. In addition, each deprotection and coupling step is monitored in real time by LC-MS feedback: if conversion falls below 99.5 %, the synthesis automatically re-couples with fresh reagent, removing the need for scouting runs. For aggregation-prone sequences, low-loading ChemMatrix or TentaGel resins are used in combination with pseudoproline dipeptides, which mitigate β-sheet formation, and increase stepwise yields by 10–20 %. The cyclization step itself takes place on-resin under pseudo-dilution conditions (i.e., the peptide is covalently anchored which prevents intermolecular oligomerization), and excess of coupling reagents can be rapidly washed away. After ring closure, cleavage cocktails of TFA/TIS/H2O are delivered from a high-pressure module in a closed loop to minimize exposure and oxidation. Data generated during each run (temperature, pH, coupling efficiency, impurity profile, etc.) are recorded in a digital batch record to feed machine-learning models that predict optimal synthesis parameters for high yield. The result is ready-to-use 50–100 mg pilot lots of complex macrocycles, within five working days with process parameters already locked for scale-up.
Orthogonal protection is the key to achieving regio- and chemoselective cyclization of multifunctional peptides. Head-to-tail lactamization requires a minimum of three orthogonal protecting groups: (i) an acid-labile resin linker (Rink amide or Wang) for the C-terminus; (ii) an Fmoc group for the N-terminus; and (iii) side-chain protecting groups (t-Bu for Asp/Glu, Mtt for Lys, Trt for Cys/His) which are stable to both Nα-deprotection and cyclization but are removed in the final TFA cleavage step. Side-chain-to-side-chain cyclization is based on allyl/alloc or Dde/Ddz orthogonality: the peptide is synthesized on a CTC (2-chlorotrityl) resin, the allyl ester of an Asp side chain is selectively deprotected with Pd(PPh3)4, and spontaneous lactamization takes place after neutralization, with all other side-chain protecting groups remaining intact. For multiple disulfide bridges, Acm, Trt and Mob cysteine protections are introduced in a stepwise manner; oxidation with iodine forms the first disulfide, and TFA cleavage and air oxidation of the second disulfide ensures regioselectivity. For chemoselective CuAAC or RCM stapling, azide- or alkene-bearing amino acids are incorporated with Mmt or ivDde side-chain masks that are removed by mild base treatment (2 % hydrazine) without disturbing the backbone. Traceless linkers (e.g. safety-catch sulfonamide resin) allow simultaneous cyclization and cleavage, reducing the number of synthetic steps by 20 % and eliminating work-up after cyclization. This modular protection scheme ensures that even peptides containing six different reactive handles can be cyclized as single topological isomers with >90 % yield and<1 % racemization.
Fig. 2 Copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC)-mediated oligomerization, cyclodimerization, and macrocyclization.3.4
Accurate quantification of cyclization yield is the gatekeeper between bench-scale discovery and robust, reproducible production. Because the macrocyclization step often co-produces epimerized, oligomerized, or truncated species, we deploy a tiered analytical suite that combines real-time monitoring with orthogonal verification. First, reversed-phase UPLC-MS is run on crude reaction mixtures using sub-2-µm C18 columns and shallow acetonitrile gradients buffered with 0.1 % formic acid. The chromatogram immediately reveals the disappearance of the linear precursor (m/z = [M+H]+linear) and the emergence of the cyclic product (m/z = [M+H]+cyclic), allowing us to calculate crude cyclization yield directly from UV peak areas after calibration with an internal standard. To rule out false positives caused by in-source fragmentation, we collect HR-ESI-MS/MS spectra and look for diagnostic fragments that retain the intact macrocycle (e.g., the loss of water from a lactam bridge is absent in linear controls). For CuAAC or thioether cyclizations—where precursor and product share identical molecular weight—we complement mass spectrometry with FT-IR to monitor the disappearance of azide (2100 cm-1) or thiol (2550 cm-1) stretches. When the peptide exceeds 25 residues or contains multiple disulfide bonds, size-exclusion chromatography (SEC) is added to quantify soluble aggregates that would otherwise bias the UV-based yield. Finally, we benchmark every method using quantitative NMR with an internal standard (e.g., DSS) to establish an absolute yield traceable to SI units. This multi-layered approach routinely delivers cyclization yields with a relative standard deviation (RSD)<3 % across five independent runs, ensuring that SAR decisions are made on statistically robust data.
Ensuring scalability and robustness of the synthetic process to meet regulatory and translational requirements are critical for making the right molecule (batch purity ≥95 %) on both pilot and production scales. The Quality-by-Design (QbD) approach is anchored on the predefined quality attributes: single impurities ≤0.5 %; overall impurities ≤2 %; epimeric impurities ≤0.1 %. The release of each batch is based on a three-step protocol: analytical RP-HPLC on two orthogonal columns (C18 and phenyl-hexyl) and at two wavelengths (214 nm and 280 nm) to identify the peaks and quantify impurities; high-resolution MS to confirm the exact mass and isotope pattern with the theoretical values; and chiral HPLC analysis after Marfey derivatization to monitor possible racemization at the site of cyclization. To control the process and ensure batch-to-batch reproducibility, critical process parameters (CPPs) such as temperature, coupling equivalents, base strength, and reaction time were "locked" based on the PAT instrumentation: in-line Raman spectroscopy to monitor the deprotection completeness; and UPLC-MS to divert the off-spec material on-line. Statistical process control (SPC) charts were used to monitor any CPP drift over 30+ batches, and corrective measures were taken if any drift beyond 2σ. Stability studies were performed at 40 °C/75 % RH for 6 months to ensure that the final product has >98 % purity.
The primary rate-limiting step during lead optimization of peptides is macrocyclization since the resulting ring strain along with competing oligomerization and epimerization keep yields below 20 %. We have systemically de-risked these variables through a decade-long program on "impossible" cyclizations. We operate a bifurcated platform consisting of: a chemistry track where head-to-tail, side-chain-to-side-chain, and chemoselective click or metathesis cyclizations are evaluated on the basis of calculated conformational entropy and steric maps; and an enzymology track where engineered sortase A and butelase 1 variants are used to perform aqueous cyclizations at pH 7.4 to entirely eliminate protecting-group manipulations. For peptides >25 residues in length or containing multiple disulfide bridges, we divide the sequence into smaller fragments, cyclize each fragment under microwave-enhanced SPPS conditions, and ligate them via native chemical ligation, which routinely results in overall isolated yields >50 %. Critical process parameters (temperature, base strength, and activating reagent stoichiometry) are locked by Design-of-Experiments (DoE) studies so that the same route used to deliver 50 mg in discovery can be transplanted to our 200 L GMP suite with no re-optimization required. The result has been the delivery of >100,000 cyclic peptides to date, including bicyclic scaffolds with defined topologies that display picomolar affinities for previously "undruggable" PPI targets.
The standard data package for each batch of cyclic peptides is far beyond regulatory requirements. Stage 1: UPLC-MS using sub-2-µm C18 columns to separate epimers and oxidation products, with exact-mass confirmation on a qTOF instrument with < 1 ppm mass accuracy. Stage 2: 2D 1H-13C HSQC NMR to confirm regioselective cyclization and detect linear precursor if >0.1 % is present. Stage 3: Quantitation of aggregate levels by size-exclusion chromatography, and confirmation of stereochemical purity by Marfey's derivatization/chiral LC-MS/MS. All raw data are uploaded to a secure electronic lab notebook within 24 h of synthesis completion and interactive dashboards enable clients to drill down on purity trends for each batch.
Synthesizing either 5 mg of material for a first SPR screen or 5 kg of API for GLP tox studies, our vertically integrated workflow removes the hand-offs that often result in lost time during development. Discovery tier: automated microwave synthesizers assemble and cyclize a 20-mer peptide in 48 h, delivering milligram-to-gram quantities with >95 % purity. Pilot tier: 20 L solid-phase reactors and flow-cyclization skids scale up hundreds of grams of material under GMP-like conditions with in-line PAT (Raman, UPLC-MS) that guarantees batch-to-batch CV<3 %. Most importantly, the same analytical methods and release specifications are used at all scales to ensure the material used in early pharmacology studies is chemically identical to the batches used in IND-enabling tox work, thereby removing scale-up risk and compressing overall development timelines.
Peptide Synthesis Services at Creative Peptides
References
