High-throughput screening (HTS) has emerged as a critical technology for the discovery of cyclic peptides. HTS enables the rapid and systematic screening of large peptide libraries, often with the use of automated experimental methods and advanced computational analysis. This approach allows for the identification of high-affinity cyclic peptides in a fraction of the time and with less resource expenditure than traditional methods. HTS is particularly useful for investigating complex biological interactions and for identifying novel therapeutic candidates.
HTS is an established platform in the discovery of peptide therapeutics, in particular of cyclic peptides, due to their proteolytic stability and topology which can potentially interact with even the so-called undruggable protein-protein interaction targets for which small molecule drugs are challenging to find. To this end, combinatorial library design and synthesis in parallel is commonly used and this has proven to be complementary to HTS approaches to enable the discovery of macrocyclic ligands in a relatively short time period. A major difference of the cyclic peptides to their linear counterparts is their more rigid structure, which in some cases even preorganizes their pharmacophoric features. When used together, the technologies for combinatorial library generation and automated screening allow for the targeted generation and systematic testing of millions of unique cyclic peptide derivatives against a target or disease phenotype in a short period of time. Screening such a large number of entities is a prerequisite since due to the combinatorial complexity of a peptide macrocycle regarding its ring size, amino acid sequence and stereochemistry the number of unique variants is exponentially large and cannot be tested in a serial manner. In addition, modern HTS platforms use advanced cheminformatics and machine learning to analyze and deconvolute screening results in order to directly distill biological data into knowledge, i.e. structure-activity relationships for follow-up lead optimization campaigns. The workflow therefore not only addresses hit finding but encompasses an entire pipeline from library design, library display, automated HTS platform execution and computational hit triage to produce validated lead candidates at an early stage with reduced loss rates during development.
Fig.1 General classification of high-throughput screening (HTS) assays.1,5
HTS workflows for cyclic peptides are defined as complete pipelines of connected stages that tightly integrate library generation, screening, and hit deconvolution into a continuous operational process. This process is initiated by the translation of in silico or rationally-designed and diversified libraries into a testable format, where the display technology (biological encoding systems or synthetic beads) impacts the automation needs and data processing strategies. Seamless connectivity between preparation and screening is critical, as the workflow has to be capable of processing thousands to millions of distinct peptide variants to a target protein under identical assay conditions. Automated scheduling of compound dispensing, incubation periods, and detection across multiple experimental batches is key to reproducibility and cycle time reduction. Primary screening data are filtered in real-time to remove artifacts, with selected candidates being subjected to orthogonal confirmation assays to confirm binding specificity and functional activity. This is then followed by hit triage, which typically involves structural characterization of the selected cyclic peptides, resynthesis, and dose-response determination, thus closing the loop with the lead optimization process. Quality control measures for compound integrity, liquid handling accuracy, and assay performance are integrated throughout, providing feedback for library optimization and screening parameter tuning. In this way, HTS becomes a dynamic knowledge generating activity, where each screening campaign is used to improve predictive models, increase chemical diversity sampling, and expedite the translation from molecular hits to therapeutic leads.
The format of cyclic peptide libraries has a great impact on screening results. Three general approaches for peptide library design that have been adapted to HTS pipelines include phage display, mRNA display, and combinatorial chemical synthesis. Phage display uses bacteriophages to present cyclic peptides on protein scaffolds that are exposed on the phage surface, physically linking genotype to phenotype in a manner that allows for easy sequence recovery upon amplification after selection. This format can be used for screening with immobilized targets on solid phase, and can be easily eluted, amplified, and re-incubated to allow for multiple rounds of selection. However, this format cannot easily include unnatural amino acids and N-methylated amino acids, which are important for properties such as cell permeability and metabolic stability, thus limiting the potential chemical diversity of the library. mRNA display can solve this problem by chemically linking the peptide to its mRNA template via a puromycin linkage to the ribosome during translation. This reaction can be performed in cell-free translation, which can have an expanded genetic code to allow for unnatural amino acids. The mRNA-peptide product can then be reverse transcribed to cDNA and amplified by PCR and sequenced using next-generation sequencing to deconvolute hits. This format can be used with libraries with complexities of over 1012 and allows for screening in solution, but the tags required for mRNA display can be sensitive to nucleases, and the tags can sterically hinder binding events in affinity selections. A combinatorial library made by split-and-pool solid-phase peptide synthesis is another fully synthetic method for generating a cyclic peptide library, allowing for the greatest chemical diversity in inclusion of unnatural amino acids, D-amino acids, and backbone modifications. A one-bead-one-compound approach can be used to spatially separate individual library members on resin, which cyclizes on-bead, before screening. A limitation to this method is that incomplete cyclization of the peptides leads to heterogenous mixtures, and the display of multiple copies of the same peptide per bead can increase apparent affinity through avidity and may require further validation.
Robotic platforms for HTS of cyclic peptides have been developed that perform robotic liquid handling of very small, reproducible volumes in microplates to enable high-throughput screening. Integrated robotic systems include workcells with compound dispensing, reagent dispensing, incubation, detection, etc. to reduce the need for manual handling of microplates and typically report a coefficient of variation (CV) of less than 5% in 384-well or 1536-well microplates. For compound dispensing, acoustic droplet ejection technology has been used for a contact-free method to dispense small volumes (typically in the nanoliter range) of compound with reduced risk of cross-contamination and increased accuracy in dispensing that avoids tip-washing steps that can deplete the available amounts of cyclic peptides. Screening in 384- or 1536-well microplates remains the most common HTS platform used in drug discovery as it has a well-established format and is compatible with many different biochemical or cell-based assay technologies. For assays of cyclic peptides, homogeneous assay formats are preferred for cyclic peptides so that no wash steps are required which might disturb the equilibrium binding state. The binding state can be monitored by fluorescent detection or directly with label-free biosensors. With label-free detection such as biolayer interferometry, immobilized biosensors in a microplate format can be interrogated directly to obtain real-time binding kinetics information. Reaction volumes can be reduced to less than a microliter which not only conserves reagents, but also allows screening of million compound libraries in a cost-effective manner. Quality control (QC) of the automated processes can be provided by integrated liquid handlers with built in gravimetric and fluorescent dye-based volume checking to confirm the accuracy of the dispensing. Compound QC can also be performed by mass spectrometry to identify degradation or precipitation from source plates before screening to avoid false negatives. Barcode information is typically collected from the microplates automatically, with sample positions added to a data system. Hit calling is performed using statistics and cut-offs are applied such as those recommended for industry wide use to filter out nonspecific aggregators and fluorescent contaminants.
Fig. 2 Screening of biologically active cyclic peptides.2,5
The detection strategy used is one of the most important aspects for a successful implementation of HTS for cyclic peptides. The chosen detection method needs to be sensitive enough to measure the molecular recognition event and also needs to consider the physicochemical properties that macrocyclic compounds display. In addition, for HTS the detection method must yield decision quality information to avoid false positives caused by the often complex nature of the screening matrix. At the same time, it is important to note that there is an intrinsic trade-off between information content, sensitivity and throughput. Consequently, detection methods used for screening purposes are often combined and/or used at different stages of the screening process (i.e. primary screening vs hit validation vs lead optimization). In general, both label-based and label-free technologies are used in screening pipelines for cyclic peptides, each with its own strengths and weaknesses. Label-based technologies offer very high sensitivity and often provide homogeneous (solution phase), single-step (in contrast to off-rate kinetic methods) readouts, which can be very important for capturing the native dynamics of these conformationally flexible molecules. On the other hand, label-based approaches have the disadvantage of requiring the covalent attachment of the reporter moiety to the cyclic peptide. This reporter group could potentially interfere with the binding epitope of the molecule or could in some other way perturb binding. Label-free binding assays, on the other hand, measure binding events directly, usually by monitoring a change in mass, refractive index or charge-to-mass ratio of the target-bound reagent. The most commonly used label-free screening methods are SPR, bio-layer interferometry, and mass spectrometry-based detection. The selection process for detection methods of cyclic peptide binding in screening requires a balance between collecting comprehensive binding data and obtaining a high volume of hits for subsequent analysis. In practice, this often means that multiple orthogonal techniques are used and/or that different readouts are used at different stages of a screening pipeline.
Binding and affinity assays are commonly used in cyclic peptide detection and analysis during HTS. Surface plasmon resonance (SPR) is a popular label-free method that detects changes in refractive index at a sensor surface. It can be used to monitor molecular interactions in real-time and thus offers kinetic information about binding affinity, specificity and association/dissociation rates. SPR also has high-throughput applications and has been used in fragment-based screening to find compounds with some binding affinity for a desired target. Another frequently used assay is fluorescence polarization (FP). FP measures the rotational motion of fluorescently labeled molecules. Upon binding of a labeled molecule to a macromolecule, rotational diffusion is slowed which increases the fluorescence polarization. FP can be used to screen libraries of small molecules and peptides, and is fast and sensitive.
MS-based screening is a type of screening method that utilizes mass spectrometry to analyze the molecular weight and structure of cyclic peptides, which can be used to identify and quantify the target peptides. This method can be very sensitive and is often used for HTS of large libraries of cyclic peptides. Label-free screening is a screening method that does not require any additional labeling or modification of the cyclic peptides. In this method, techniques such as bio-layer interferometry (BLI) or SPR can be used to directly measure the binding affinity and kinetics of the cyclic peptides to their target of interest. These methods can be highly sensitive and accurate, and can provide real-time data. Overall, MS-based and label-free screening methods are highly sensitive, accurate, and have high-throughput capabilities, making them ideal for HTS of cyclic peptides.
Next-generation sequencing (NGS) can be used to deconvolute hits from HTS in cyclic peptides. After performing the initial screening campaign and obtaining a large dataset of hits, NGS can be used to systematically analyze and identify true positives that require further investigation. This technology allows for the rapid and accurate determination of the sequences of cyclic peptides that have demonstrated promising binding affinities. The DNA or RNA linked to the peptide libraries can also be sequenced and then correlated back to the sequences of the identified hits. This can be particularly advantageous for phage display and mRNA display libraries since each phage or mRNA molecule is encoding the sequence of the displayed peptide. NGS provides an easy, high-throughput and cost-effective means for deconvolution, as well as the ability to look for multiple hit candidates in parallel. In addition, the NGS data can be easily fed into other computational tools for further analysis of the sequence space and prioritization of the best candidates to be resynthesized and validated. Confirmatory resynthesis and biological testing is used to confirm hits identified by HTS. Hit cyclic peptides are resynthesized to confirm their identity and activity. The resynthesized putative hits are then subjected to biological testing to confirm their activity and specificity. This can include binding assays using techniques such as SPR or fluorescence polarization (FP), as well as functional assays to evaluate the biological effects of the peptides in cells or biochemical systems. Confirmatory testing is essential to rule out false positives and to ensure that the putative hits are indeed active against the target of interest. Orthogonal testing methods such as nuclear magnetic resonance (NMR) or isothermal titration calorimetry (ITC) can also be used to provide complementary information on the interaction between the cyclic peptide and its target. This multi-step process is critical to ensure that only the most promising candidates are advanced to further stages of drug discovery and development.
Lead optimization workflows involve the systematic modification of hit cyclic peptides to improve their potency, selectivity, and drug-like properties. After hit identification and initial validation, medicinal chemists assess the chemical feasibility of scaling up the synthesis of the hit compounds. This involves evaluating the scalability of the synthetic route and optimizing reaction conditions to improve yield and purity. Computational tools are used to predict the physicochemical properties of the hit compounds, including solubility, membrane permeability, and metabolic stability. These predictions can be used to design analogs with improved pharmacokinetic properties. ADME-tox (Absorption, Distribution, Metabolism, Excretion-Toxicity) assessments are performed early in the development process to evaluate the absorption, distribution, metabolism, and excretion properties of the hit compounds. Toxicity profiles are also determined to identify any potential safety issues. This information is used to prioritize the most promising candidates for further optimization. Intellectual property assessments are conducted to determine the patentability of the hit compounds and identify any potential legal or competitive challenges. HTS has been applied in screening for enzyme inhibitors. Cyclic peptides are screened for their ability to inhibit the activity of specific enzymes. Enzyme inhibition assays can be performed using biochemical activity assays or fluorescence-based displacement assays. Receptor ligands are identified by screening cyclic peptides for their ability to bind to cell surface receptors. This can be accomplished using cell-based assays or surface plasmon resonance. Antimicrobial peptides are screened for their ability to inhibit the growth of bacteria, fungi, or other pathogens. This can be done by measuring the minimum inhibitory concentration (MIC) of the peptide or by assessing its ability to disrupt microbial membranes. These examples illustrate the versatility of HTS in identifying cyclic peptides with a range of biological activities, making it a valuable tool in drug discovery and therapeutic development.
There is always a tradeoff between throughput and data accuracy in HTS. While the goal of high-throughput screening methods is to process as many samples as possible in the shortest amount of time, this can sometimes come at the expense of data quality. Assay variability, signal-to-noise ratios, and false-positive rates are just a few of the factors that can impact data accuracy. Efforts to improve data accuracy in HTS for cyclic peptides are focusing on developing more robust assay formats and implementing more advanced data analysis techniques. Automation and microfluidics integration are also being used to increase precision and reproducibility in HTS workflows. One trend that is currently gaining traction in the field of HTS for cyclic peptides is the integration of microfluidics and artificial intelligence (AI)-driven data curation. Microfluidic platforms allow for the precise manipulation of small volumes of samples and reagents, resulting in lower reagent consumption and higher assay throughput. These platforms can be coupled with automated liquid handling systems for even greater efficiency. AI-driven data curation tools are also being developed to help researchers analyze and interpret the massive amounts of data generated during HTS. These tools can identify patterns and trends that might not be apparent to the human eye, helping to prioritize the most promising hit candidates for further validation. AI algorithms can also be used to predict the binding affinities and functional activities of cyclic peptides based on their sequences and structural features. This can help to guide the design of optimized libraries and minimize the number of experimental iterations required.
The methodologies used to identify bioactive peptides have evolved over time. Initially, single peptides were synthesized and screened manually, a process with limited throughput and diversity, as the number of peptides synthesized and tested over a period of months could be in the dozens. Only the compounds that induced a desired readout were considered and no objective way to identify SARs existed. With the rise of combinatorial chemistry, large and diverse peptide libraries could be synthesized and screened in parallel in a single experiment. The large number of compounds generated from combinatorial methods created new challenges for deconvoluting hits from complex libraries and managing the increased data load. These problems were among the first for which solutions were sought through the development of informatics tools. Automation in peptide discovery took a step further with the development of high-throughput screening methods, which included automated liquid handling and detection. However, these systems were not tightly integrated with informatics analysis tools, often requiring data to be manually transferred from the screening instrument to separate statistical analysis software. Hit validation was generally performed in separate, iterative rounds of screening with little automation or integration. The modern high-throughput screening workflow tightly integrates synthesis of large and diverse peptide libraries with next-generation sequencing, label-free biosensors, and AI, and deconvolution to occur within a self-contained system. In this manner, the library synthesis, screening, hit identification, and deconvolution steps are fully integrated to form a continuous screening process. The process is highly automated and uses robotics for screening which enables highly reproducible assays. Advances in microfluidics also allow for higher throughput screening while using small sample sizes. In addition, data analysis is often performed using artificial intelligence, which rapidly triages hits and may also predict hit liabilities and suggest analogs for further design and testing. Screening with this integrated process is thus an iterative process of knowledge generation, where an initial round of screening is used to build a predictive model, and the model is then used to guide the design of new libraries to further optimize chemical diversity to improve hit and lead rates.
High-throughput screening (HTS) enables rapid identification of active cyclic peptides across vast libraries and diverse scaffolds. Our automated HTS workflows are designed to support rapid, data-rich discovery programs for pharmaceutical and biotechnology partners.
We support:
From screening to validation and optimization, we provide the throughput, expertise, and technical depth your discovery program needs.
References
