Peptide design by optimization on a data-parameterized protein interaction landscape

PISAD: De novo peptide design for target protein with iterative stochastic searching algorithm and docking assessment

Rapid identification of peptides that bind specifically to a target protein is essential for disease diagnostics and drug development. However, de novo design of peptides without any prior structural knowledge remains a long-standing challenge. Herein, we present Peptide Iterative design with Stochastic Algorithm and Docking (PISAD), a de novo peptide design method, which combines an iterative stochastic searching algorithm with docking assessment. The searching algorithm simulates the evolution of peptide sequences by introducing mutations and crossovers iteratively. After every round of evolution, the peptide sequences undergo docking assessments with the target protein based on the structural prediction of AlphaFold2. We demonstrated PISAD's efficacy by designing peptides targeting four proteins, namely ARF6, ARF1, TGF-β1, and IL-6. For each target, PISAD managed to output peptides with ideal binding affinity within only four iterations of evolutions, and no more than 1250 sequences were assessed. Particularly, the best-performing peptide achieved a KD value of 3.4 nM with ARF6, which has been further experimentally validated. These results demonstrate the efficiency and accuracy of PISAD, which may serve as a universal tool for rapid de novo design of peptides targeting specific proteins.

Molecular Dynamics of Peptide Sequencing through MoS2 Solid-State Nanopores for Binary Encoding Applications

Biological peptides have emerged as promising candidates for data storage applications due to their versatility and programmability. Recent advances in peptide synthesis and sequencing technologies have enabled the development of peptide-based data storage systems for realizing novel information storage technologies with enhanced capacity, durability, and data access speeds. In this study, we performed coarse-grained peptide sequencing of 12 distinct sequences through single-layer MoS2 solid-state nanopores (SSNs) using molecular dynamics (MD). Peptide sequences were composed of 1 positively charged, 1 negatively charged, and 4 neutral amino acids, with the position of amino acids in the sequence being shuffled to generate all possible configurations. From MD, the goal was to evaluate the efficiency of these peptide sequences to encode binary information based on ionic current traces monitored during their passage through the SSNs. Classification approaches using LightGBM were trained and tested to analyze different sequence factors such as the position of amino acids or the spacing between charged amino acids in the sequences. Our findings reveal the presence of two distinct groups of sequences determined by the relative position of the positively charged amino acid compared to the negatively charged amino acid. Furthermore, we observe a strong correlation between discrimination accuracy and the separation in the sequence between charged amino acids, depending on the number of adjacent neutral amino acids between them. Finally, MD allowed us to establish the nonlinear relationship between amino acid positions inside the pore (called sequence motifs) and fluctuations in ionic current traces to discriminate false positives and to enable effective training of machine learning classification algorithms. These very promising results emphasized the best approaches to design peptide sequences as building blocks for molecular data storage. Finally, this study highlights the potential of the proposed approach for designing peptide sequence combinations that could help the development of efficient, scalable, and reliable molecular data storage solutions, with future research focused on encoding longer binary chains to enhance storage capacity and support the goal of stable, energy-free biological systems.

In silico screening of protein-binding peptides with an application to developing peptide inhibitors against antibiotic resistance

The field of therapeutic peptides is experiencing a surge, fueled by their advantageous features. These include predictable metabolism, enhanced safety profile, high selectivity, and reduced off-target effects compared with small-molecule drugs. Despite progress in addressing limitations associated with peptide drugs, a significant bottleneck remains: the absence of a large-scale in silico screening method for a given protein target structure. Such methods have proven invaluable in accelerating small-molecule drug discovery. The high flexibility of peptide structures and the large diversity of peptide sequences greatly hinder the development of urgently needed computational methods. Here, we report a method called MDockPeP2_VS to address these challenges. It integrates molecular docking with structural conservation between protein folding and protein-peptide binding. Briefly, we discovered that when the interfacial residues are conserved, a sequence fragment derived from a monomeric protein exhibits a high propensity to bind a target protein with a similar conformation. This valuable insight significantly reduces the search space for peptide conformations, resulting in a substantial reduction in computational time and making in silico peptide screening practical. We applied MDockPeP2_VS to develop peptide inhibitors targeting the TEM-1 β-lactamase of Escherichia coli, a key mechanism behind antibiotic resistance in gram-negative bacteria. Among the top 10 peptides selected from in silico screening, TF7 (KTYLAQAAATG) showed significant inhibition of β-lactamase activity with a K i value of 1.37 ± 0.37 µM. This fully automated, large-scale structure-based in silico peptide screening software is available for free download at https://zougrouptoolkit.missouri.edu/mdockpep2_vs/download.html.

Deep neural networks for predicting the affinity landscape of protein-protein interactions

Studies determining protein-protein interactions (PPIs) by deep mutational scanning have focused mainly on a narrow range of affinities within complexes and thus include only partial coverage of the mutation space of given proteins. By inserting an affinity-reducing N-terminal alanine in the N-terminal domain of the tissue inhibitor of metalloproteinases-2 (N-TIMP2), we overcame the limitation of its narrow affinity range for matrix metalloproteinase 9 (MMP9CAT). We trained deep neural networks (DNNs) to quantitatively predict the binding affinity of unobserved wild-type variants and variants carrying an N-terminal alanine. Good correlation was obtained between predicted and observed log2 enrichment ratio (ER) values, which also correlated with the affinity of N-TIMP2 variants to MMP9CAT. Our ability to predict affinities of unobserved N-TIMP2 variants was confirmed on an independent dataset of experimentally validated N-TIMP2 proteins. This ability is of significant importance in the field of PPI prediction and for developing therapies targeting these interactions.

Cold Spot SCANNER: Colab Notebook for predicting cold spots in protein-protein interfaces

BACKGROUND

Protein-protein interactions (PPIs) are conveyed through binding interfaces or surface patches on proteins that become buried upon binding. Structural and biophysical analysis of many protein-protein interfaces revealed certain unique features of these surfaces that determine the energetics of interactions and play a critical role in protein evolution. One of the significant aspects of binding interfaces is the presence of binding hot spots, where mutations are highly deleterious for binding. Conversely, binding cold spots are positions occupied by suboptimal amino acids and several mutations in such positions could lead to affinity enhancement. While there are many software programs for identification of hot spot positions, there is currently a lack of software for cold spot detection.

RESULTS

In this paper, we present Cold Spot SCANNER, a Colab Notebook, which scans a PPI binding interface and identifies cold spots resulting from cavities, unfavorable charge-charge, and unfavorable charge-hydrophobic interactions. The software offers a Py3DMOL-based interface that allows users to visualize cold spots in the context of the protein structure and generates a zip file containing the results for easy download.

CONCLUSIONS

Cold spot identification is of great importance to protein engineering studies and provides a useful insight into protein evolution. Cold Spot SCANNER is open to all users without login requirements and can be accessible at: https://colab.

RESEARCH

google.com/github/sagagugit/Cold-Spot-Scanner/blob/main/Cold_Spot_Scanner.ipynb .

Peptide-based allosteric inhibitor targets TNFR1 conformationally active region and disables receptor-ligand signaling complex

Tumor necrosis factor (TNF) receptor 1 (TNFR1) plays a pivotal role in mediating TNF induced downstream signaling and regulating inflammatory response. Recent studies have suggested that TNFR1 activation involves conformational rearrangements of preligand assembled receptor dimers and targeting receptor conformational dynamics is a viable strategy to modulate TNFR1 signaling. Here, we used a combination of biophysical, biochemical, and cellular assays, as well as molecular dynamics simulation to show that an anti-inflammatory peptide (FKCRRWQWRMKK), which we termed FKC, inhibits TNFR1 activation allosterically by altering the conformational states of the receptor dimer without blocking receptor-ligand interaction or disrupting receptor dimerization. We also demonstrated the efficacy of FKC by showing that the peptide inhibits TNFR1 signaling in HEK293 cells and attenuates inflammation in mice with intraperitoneal TNF injection. Mechanistically, we found that FKC binds to TNFR1 cysteine-rich domains (CRD2/3) and perturbs the conformational dynamics required for receptor activation. Importantly, FKC increases the frequency in the opening of both CRD2/3 and CRD4 in the receptor dimer, as well as induces a conformational opening in the cytosolic regions of the receptor. This results in an inhibitory conformational state that impedes the recruitment of downstream signaling molecules. Together, these data provide evidence on the feasibility of targeting TNFR1 conformationally active region and open new avenues for receptor-specific inhibition of TNFR1 signaling.

Machine learning to predict continuous protein properties from binary cell sorting data and map unseen sequence space

Proteins are a diverse class of biomolecules responsible for wide-ranging cellular functions, from catalyzing reactions to recognizing pathogens. The ability to evolve proteins rapidly and inexpensively toward improved properties is a common objective for protein engineers. Powerful high-throughput methods like fluorescent activated cell sorting and next-generation sequencing have dramatically improved directed evolution experiments. However, it is unclear how to best leverage these data to characterize protein fitness landscapes more completely and identify lead candidates. In this work, we develop a simple yet powerful framework to improve protein optimization by predicting continuous protein properties from simple directed evolution experiments using interpretable, linear machine learning models. Importantly, we find that these models, which use data from simple but imprecise experimental estimates of protein fitness, have predictive capabilities that approach more precise but expensive data. Evaluated across five diverse protein engineering tasks, continuous properties are consistently predicted from readily available deep sequencing data, demonstrating that protein fitness space can be reasonably well modeled by linear relationships among sequence mutations. To prospectively test the utility of this approach, we generated a library of stapled peptides and applied the framework to predict affinity and specificity from simple cell sorting data. We then coupled integer linear programming, a method to optimize protein fitness from linear weights, with mutation scores from machine learning to identify variants in unseen sequence space that have improved and co-optimal properties. This approach represents a versatile tool for improved analysis and identification of protein variants across many domains of protein engineering.

Design of target specific peptide inhibitors using generative deep learning and molecular dynamics simulations

We introduce a computational approach for the design of target-specific peptides. Our method integrates a Gated Recurrent Unit-based Variational Autoencoder with Rosetta FlexPepDock for peptide sequence generation and binding affinity assessment. Subsequently, molecular dynamics simulations are employed to narrow down the selection of peptides for experimental assays. We apply this computational strategy to design peptide inhibitors that specifically target β-catenin and NF-κB essential modulator. Among the twelve β-catenin inhibitors, six exhibit improved binding affinity compared to the parent peptide. Notably, the best C-terminal peptide binds β-catenin with an IC50 of 0.010 ± 0.06 μM, which is 15-fold better than the parent peptide. For NF-κB essential modulator, two of the four tested peptides display substantially enhanced binding compared to the parent peptide. Collectively, this study underscores the successful integration of deep learning and structure-based modeling and simulation for target specific peptide design.

Position-Specific Enrichment Ratio Matrix scores predict antibody variant properties from deep sequencing data

MOTIVATION

Deep sequencing of antibody and related protein libraries after phage or yeast-surface display sorting is widely used to identify variants with increased affinity, specificity, and/or improvements in key biophysical properties. Conventional approaches for identifying optimal variants typically use the frequencies of observation in enriched libraries or the corresponding enrichment ratios. However, these approaches disregard the vast majority of deep sequencing data and often fail to identify the best variants in the libraries.

RESULTS

Here, we present a method, Position-Specific Enrichment Ratio Matrix (PSERM) scoring, that uses entire deep sequencing datasets from pre- and post-selections to score each observed protein variant. The PSERM scores are the sum of the site-specific enrichment ratios observed at each mutated position. We find that PSERM scores are much more reproducible and correlate more strongly with experimentally measured properties than frequencies or enrichment ratios, including for multiple antibody properties (affinity and non-specific binding) for a clinical-stage antibody (emibetuzumab). We expect that this method will be broadly applicable to diverse protein engineering campaigns.

AVAILABILITY AND IMPLEMENTATION

All deep sequencing datasets and code to perform the analyses presented within are available via https://github.com/Tessier-Lab-UMich/PSERM_paper.

Protein engineering via sequence-performance mapping

Discovery and evolution of new and improved proteins has empowered molecular therapeutics, diagnostics, and industrial biotechnology. Discovery and evolution both require efficient screens and effective libraries, although they differ in their challenges because of the absence or presence, respectively, of an initial protein variant with the desired function. A host of high-throughput technologies-experimental and computational-enable efficient screens to identify performant protein variants. In partnership, an informed search of sequence space is needed to overcome the immensity, sparsity, and complexity of the sequence-performance landscape. Early in the historical trajectory of protein engineering, these elements aligned with distinct approaches to identify the most performant sequence: selection from large, randomized combinatorial libraries versus rational computational design. Substantial advances have now emerged from the synergy of these perspectives. Rational design of combinatorial libraries aids the experimental search of sequence space, and high-throughput, high-integrity experimental data inform computational design. At the core of the collaborative interface, efficient protein characterization (rather than mere selection of optimal variants) maps sequence-performance landscapes. Such quantitative maps elucidate the complex relationships between protein sequence and performance-e.g., binding, catalytic efficiency, biological activity, and developability-thereby advancing fundamental protein science and facilitating protein discovery and evolution.

Position-Specific Enrichment Ratio Matrix scores predict antibody variant properties from deep sequencing data

Motivation

Deep sequencing of antibody and related protein libraries after phage or yeast-surface display sorting is widely used to identify variants with increased affinity, specificity and/or improvements in key biophysical properties. Conventional approaches for identifying optimal variants typically use the frequencies of observation in enriched libraries or the corresponding enrichment ratios. However, these approaches disregard the vast majority of deep sequencing data and often fail to identify the best variants in the libraries.

Results

Here, we present a method, Position-Specific Enrichment Ratio Matrix (PSERM) scoring, that uses entire deep sequencing datasets from pre- and post-selections to score each observed protein variant. The PSERM scores are the sum of the site-specific enrichment ratios observed at each mutated position. We find that PSERM scores are much more reproducible and correlate more strongly with experimentally measured properties than frequencies or enrichment ratios, including for multiple antibody properties (affinity and non-specific binding) for a clinical-stage antibody (emibetuzumab). We expect that this method will be broadly applicable to diverse protein engineering campaigns.

Availability

All deep sequencing datasets and code to do the analyses presented within are available via GitHub.

Contact

Peter Tessier, ptessier@umich.edu.

Supplementary information

Supplementary data are available at Bioinformatics online.

BinderSpace: A package for sequence space analyses for datasets of affinity-selected oligonucleotides and peptide-based molecules

Discovery of target-binding molecules, such as aptamers and peptides, is usually performed with the use of high-throughput experimental screening methods. These methods typically generate large datasets of sequences of target-binding molecules, which can be enriched with high affinity binders. However, the identification of the highest affinity binders from these large datasets often requires additional low-throughput experiments or other approaches. Bioinformatics-based analyses could be helpful to better understand these large datasets and identify the parts of the sequence space enriched with high affinity binders. BinderSpace is an open-source Python package that performs motif analysis, sequence space visualization, clustering analyses, and sequence extraction from clusters of interest. The motif analysis, resulting in text-based and visual output of motifs, can also provide heat maps of previously measured user-defined functional properties for all the motif-containing molecules. Users can also run principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE) analyses on whole datasets and on motif-related subsets of the data. Functionally important sequences can also be highlighted in the resulting PCA and t-SNE maps. If points (sequences) in two-dimensional maps in PCA or t-SNE space form clusters, users can perform clustering analyses on their data, and extract sequences from clusters of interest. We demonstrate the use of BinderSpace on a dataset of oligonucleotides binding to single-wall carbon nanotubes in the presence and absence of a bioanalyte, and on a dataset of cyclic peptidomimetics binding to bovine carbonic anhydrase protein. BinderSpace is openly accessible to the public via the GitHub website: https://github.com/vukoviclab/BinderSpace.

Rapid Evaluation of Staple Placement in Stabilized α Helices Using Bacterial Surface Display

There are a wealth of proteins involved in disease that cannot be targeted by current therapeutics because they are inside cells, inaccessible to most macromolecules, and lack small-molecule binding pockets. Stapled peptides, where two amino acids are covalently linked, form a class of macrocycles that have the potential to penetrate cell membranes and disrupt intracellular protein-protein interactions. However, their discovery relies on solid-phase synthesis, greatly limiting queries into their complex design space involving amino acid sequence, staple location, and staple chemistry. Here, we use stabilized peptide engineering by Escherichia coli display (SPEED), which utilizes noncanonical amino acids and click chemistry for stabilization, to rapidly screen staple location and linker structure to accelerate peptide design. After using SPEED to confirm hotspots in the mdm2-p53 interaction, we evaluated different staple locations and staple chemistry to identify several novel nanomolar and sub-nanomolar antagonists. Next, we evaluated SPEED in the B cell lymphoma 2 (Bcl-2) protein family, which is responsible for regulating apoptosis. We report that novel staple locations modified in the context of BIM, a high affinity but nonspecific naturally occurring peptide, improve its specificity against the highly homologous proteins in the Bcl-2 family. These compounds demonstrate the importance of screening linker location and chemistry in identifying high affinity and specific peptide antagonists. Therefore, SPEED can be used as a versatile platform to evaluate multiple design criteria for stabilized peptide engineering.

Mutational scan inferred binding energetics and structure in intrinsically disordered protein CcdA

Unlike globular proteins, mutational effects on the function of Intrinsically Disordered Proteins (IDPs) are not well-studied. Deep Mutational Scanning of a yeast surface displayed mutant library yields insights into sequence-function relationships in the CcdA IDP. The approach enables facile prediction of interface residues and local structural signatures of the bound conformation. In contrast to previous titration-based approaches which use a number of ligand concentrations, we show that use of a single rationally chosen ligand concentration can provide quantitative estimates of relative binding constants for large numbers of protein variants. This is because the extended interface of IDP ensures that energetic effects of point mutations are spread over a much smaller range than for globular proteins. Our data also provides insights into the much-debated role of helicity and disorder in partner binding of IDPs. Based on this exhaustive mutational sensitivity dataset, a rudimentary model was developed in an attempt to predict mutational effects on binding affinity of IDPs that form alpha-helical structures upon binding.

Titrating Avidity of Yeast-Displayed Proteins Using a Transcriptional Regulator

Yeast surface display is a valuable tool for protein engineering and directed evolution; however, significant variability in the copy number (i.e., avidity) of displayed variants on the yeast cell wall complicates screening and selection campaigns. Here, we report an engineered titratable display platform that modulates the avidity of Aga2-fusion proteins on the yeast cell wall dependent on the concentration of the anhydrotetracycline (aTc) inducer. Our design is based on a genomic Aga1 gene copy and an episomal Aga2-fusion construct both under the control of an aTc-dependent transcriptional regulator that enables stoichiometric and titratable expression, secretion, and display of Aga2-fusion proteins. We demonstrate tunable display levels over 2-3 orders of magnitude for various model proteins, including glucose oxidase enzyme variants, mechanostable dockerin-binding domains, and anti-PDL1 affibody domains. By regulating the copy number of displayed proteins, we demonstrate the effects of titratable avidity levels on several specific phenotypic activities, including enzyme activity and cell adhesion to surfaces under shear flow. Finally, we show that titrating down the display level allows yeast-based binding affinity measurements to be performed in a regime that avoids ligand depletion effects while maintaining small sample volumes, avoiding a well-known artifact in yeast-based binding assays. The ability to titrate the multivalency of proteins on the yeast cell wall through simple inducer control will benefit protein engineering and directed evolution methodology relying on yeast display for broad classes of therapeutic and diagnostic proteins of interest.

Neural network-derived Potts models for structure-based protein design using backbone atomic coordinates and tertiary motifs

Designing novel proteins to perform desired functions, such as binding or catalysis, is a major goal in synthetic biology. A variety of computational approaches can aid in this task. An energy-based framework rooted in the sequence-structure statistics of tertiary motifs (TERMs) can be used for sequence design on predefined backbones. Neural network models that use backbone coordinate-derived features provide another way to design new proteins. In this work, we combine the two methods to make neural structure-based models more suitable for protein design. Specifically, we supplement backbone-coordinate features with TERM-derived data, as inputs, and we generate energy functions as outputs. We present two architectures that generate Potts models over the sequence space: TERMinator, which uses both TERM-based and coordinate-based information, and COORDinator, which uses only coordinate-based information. Using these two models, we demonstrate that TERMs can be utilized to improve native sequence recovery performance of neural models. Furthermore, we demonstrate that sequences designed by TERMinator are predicted to fold to their target structures by AlphaFold. Finally, we show that both TERMinator and COORDinator learn notions of energetics, and these methods can be fine-tuned on experimental data to improve predictions. Our results suggest that using TERM-based and coordinate-based features together may be beneficial for protein design and that structure-based neural models that produce Potts energy tables have utility for flexible applications in protein science.

Tertiary motifs as building blocks for the design of protein-binding peptides

Despite advances in protein engineering, the de novo design of small proteins or peptides that bind to a desired target remains a difficult task. Most computational methods search for binder structures in a library of candidate scaffolds, which can lead to designs with poor target complementarity and low success rates. Instead of choosing from pre-defined scaffolds, we propose that custom peptide structures can be constructed to complement a target surface. Our method mines tertiary motifs (TERMs) from known structures to identify surface-complementing fragments or "seeds." We combine seeds that satisfy geometric overlap criteria to generate peptide backbones and score the backbones to identify the most likely binding structures. We found that TERM-based seeds can describe known binding structures with high resolution: the vast majority of peptide binders from 486 peptide-protein complexes can be covered by seeds generated from single-chain structures. Furthermore, we demonstrate that known peptide structures can be reconstructed with high accuracy from peptide-covering seeds. As a proof of concept, we used our method to design 100 peptide binders of TRAF6, seven of which were predicted by Rosetta to form higher-quality interfaces than a native binder. The designed peptides interact with distinct sites on TRAF6, including the native peptide-binding site. These results demonstrate that known peptide-binding structures can be constructed from TERMs in single-chain structures and suggest that TERM information can be applied to efficiently design novel target-complementing binders.

Design of Protein Segments and Peptides for Binding to Protein Targets

Recent years have witnessed a rise in methods for accurate prediction of structure and design of novel functional proteins. Design of functional protein fragments and peptides occupy a small, albeit unique, space within the general field of protein design. While the smaller size of these peptides allows for more exhaustive computational methods, flexibility in their structure and sparsity of data compared to proteins, as well as presence of noncanonical building blocks, add additional challenges to their design. This review summarizes the current advances in the design of protein fragments and peptides for binding to targets and discusses the challenges in the field, with an eye toward future directions.

Discovery of DNA-Carbon Nanotube Sensors for Serotonin with Machine Learning and Near-infrared Fluorescence Spectroscopy

DNA-wrapped single walled carbon nanotube (SWNT) conjugates have distinct optical properties leading to their use in biosensing and imaging applications. A critical limitation in the development of DNA-SWNT sensors is the current inability to predict unique DNA sequences that confer a strong analyte-specific optical response to these sensors. Here, near-infrared (nIR) fluorescence response data sets for ∼100 DNA-SWNT conjugates, narrowed down by a selective evolution protocol starting from a pool of ∼10¹⁰ unique DNA-SWNT candidates, are used to train machine learning (ML) models to predict DNA sequences with strong optical response to neurotransmitter serotonin. First, classifier models based on convolutional neural networks (CNN) are trained on sequence features to classify DNA ligands as either high response or low response to serotonin. Second, support vector machine (SVM) regression models are trained to predict relative optical response values for DNA sequences. Finally, we demonstrate with validation experiments that integrating the predictions of ensembles of the highest quality neural network classifiers (convolutional or artificial) and SVM regression models leads to the best predictions of both high and low response sequences. With our ML approaches, we discovered five DNA-SWNT sensors with higher fluorescence intensity response to serotonin than obtained previously. Overall, the explored ML approaches, shown to predict useful DNA sequences, can be used for discovery of DNA-based sensors and nanobiotechnologies.

Multiplexed Small-Molecule-Ligand Binding Assays by Affinity Labeling and DNA Sequence Analysis

Small-molecule binding assays to target proteins are a core component of drug discovery and development. While a number of assay formats are available, significant drawbacks still remain in cost, sensitivity, and throughput. To improve assays by capitalizing on the power of DNA sequence analysis, we have developed an assay method that combines DNA encoding with split-and-pool sample handling. The approach involves affinity labeling of DNA-linked ligands to a protein target. Critically, the labeling event assesses ligand binding and enables subsequent pooling of several samples. Application of a purifying selection on the pool for protein-labeled DNAs allows detection of ligand binding by quantification of DNA barcodes. We demonstrate the approach in both ligand displacement and direct binding formats and demonstrate its utility in determination of relative ligand affinity, profiling ligand specificity, and high-throughput small-molecule screening.

Lessons in Organic Fluorescent Probe Discovery

Fluorescent probes have gained profound use in biotechnology, drug discovery, medical diagnostics, molecular and cell biology. The development of methods for the translation of fluorophores into fluorescent probes continues to be a robust field for medicinal chemists and chemical biologists, alike. Access to new experimental designs has enabled molecular diversification and led to the identification of new approaches to probe discovery. This review provides a synopsis of the recent lessons in modern fluorescent probe discovery.

Towards optimizing peptide-based inhibitors of protein-protein interactions: predictive saturation variation scanning (PreSaVS)

A simple-to-implement and experimentally validated computational workflow for sequence modification of peptide inhibitors of protein–protein interactions (PPIs) is described.

An experimentally validated approach for in silico modification of peptide based protein–protein interaction inhibitors is described.

CpACpP: In Silico Cell-Penetrating Anticancer Peptide Prediction Using a Novel Bioinformatics Framework

Cell-penetrating anticancer peptides (Cp-ACPs) are considered promising candidates in solid tumor and hematologic cancer therapies. Current approaches for the design and discovery of Cp-ACPs trust the expensive high-throughput screenings that often give rise to multiple obstacles, including instrumentation adaptation and experimental handling. The application of machine learning (ML) tools developed for peptide activity prediction is importantly of growing interest. In this study, we applied the random forest (RF)-, support vector machine (SVM)-, and eXtreme gradient boosting (XGBoost)-based algorithms to predict the active Cp-ACPs using an experimentally validated data set. The model, CpACpP, was developed on the basis of two independent cell-penetrating peptide (CPP) and anticancer peptide (ACP) subpredictors. Various compositional and physiochemical-based features were combined or selected using the multilayered recursive feature elimination (RFE) method for both data sets. Our results showed that the ACP subclassifiers obtain a mean performance accuracy (ACC) of 0.98 with an area under curve (AUC) ≈ 0.98 vis-à-vis the CPP predictors displaying relevant values of ∼0.94 and ∼0.95 via the hybrid-based features and independent data sets, respectively. Also, the predicting evaluation of Cp-ACPs gave accuracies of ∼0.79 and 0.89 on a series of independent sequences by applying our CPP and ACP classifiers, respectively, which leaves the performance of our predictors better than the earlier reported ACPred, mACPpred, MLCPP, and CPPred-RF. The described consensus-based fusion method additionally reached an AUC of 0.94 for the prediction of Cp-ACP (http://cbb1.ut.ac.ir/CpACpP/Index).

A new class of sensing elements for sensors: Clamp peptides for Zika virus

The design of a new class of selective and high affinity antibody mimetics termed clamp peptide (CP) that incorporate three short peptides structurally and mechanically mimicking a clamp is proposed as sensing elements for a reliable detection sensor platform. The CPs consist of two short peptides functioning as arms that recognize two different epitopes in the target protein and are connected by a third short peptide that acts as a hinge between the peptide arms. For the construction of CPs, we employed a rational design combined with computational methods. To illustrate our approach, we designed a CP that binds selectively to the envelope protein of the Zika virus (ZIKV). The virtual docking cycles were run maximizing the discrimination between ZIKV and Dengue virus (DENV) envelope proteins. DENV was chosen among the flavivirus family because it has high structural similarity with ZIKV. When employed in a colorimetric binding assay or in label-free electrochemical impedance sensor format, the CP was selective for ZIKV vs DENV particles showing detection limit under 10⁴ copies/mL, comparable to anti-ZIKV antibodies. Apparent dissociation binding constants (Kd) confirmed a better performance of CPs than mono-arm peptides (Kd of best CP = 162 nM ± 23 nM; Kd of best mono-arm peptide = 11.15 ± 2.76 μM). The performance of the assays based on CPs was also verified in serum and urine (diluted 1:10 and 1:1 respectively). The detection limits of CPs decreased about one order of magnitude for ZIKV detection in serum or urine, with a distinct analytical signal starting from 10⁵ copies/mL of ZIKV.

Data-driven computational protein design

Computational protein design can generate proteins not found in nature that adopt desired structures and perform novel functions. Although proteins could, in theory, be designed with ab initio methods, practical success has come from using large amounts of data that describe the sequences, structures, and functions of existing proteins and their variants. We present recent creative uses of multiple-sequence alignments, protein structures, and high-throughput functional assays in computational protein design. Approaches range from enhancing structure-based design with experimental data to building regression models to training deep neural nets that generate novel sequences. Looking ahead, deep learning will be increasingly important for maximizing the value of data for protein design.

Breakthroughs in computational design methods open up new frontiers for de novo protein engineering

Proteins catalyze the majority of chemical reactions in organisms, and harnessing this power has long been the focus of the protein engineering field. Computational protein design aims to create new proteins and functions in silico, and in doing so, accelerate the process, reduce costs and enable more sophisticated engineering goals to be accomplished. Challenges that very recently seemed impossible are now within reach thanks to several landmark advances in computational protein design methods. Here, we summarize these new methods, with a particular emphasis on de novo protein design advancements occurring within the past 5 years.

PepFun: Open Source Protocols for Peptide-Related Computational Analysis

Peptide research has increased during the last years due to their applications as biomarkers, therapeutic alternatives or as antigenic sub-units in vaccines. The implementation of computational resources have facilitated the identification of novel sequences, the prediction of properties, and the modelling of structures. However, there is still a lack of open source protocols that enable their straightforward analysis. Here, we present PepFun, a compilation of bioinformatics and cheminformatics functionalities that are easy to implement and customize for studying peptides at different levels: sequence, structure and their interactions with proteins. PepFun enables calculating multiple characteristics for massive sets of peptide sequences, and obtaining different structural observables derived from protein-peptide complexes. In addition, random or guided library design of peptide sequences can be customized for screening campaigns. The package has been created under the python language based on built-in functions and methods available in the open source projects BioPython and RDKit. We present two tutorials where we tested peptide binders of the MHC class II and the Granzyme B protease.

Computational design and experimental optimization of protein binders with prospects for biomedical applications

Protein-based binders have become increasingly more attractive candidates for drug and imaging agent development. Such binders could be evolved from a number of different scaffolds, including antibodies, natural protein effectors and unrelated small protein domains of different geometries. While both computational and experimental approaches could be utilized for protein binder engineering, in this review we focus on various computational approaches for protein binder design and demonstrate how experimental selection could be applied to subsequently optimize computationally-designed molecules. Recent studies report a number of designed protein binders with pM affinities and high specificities for their targets. These binders usually characterized with high stability, solubility, and low production cost. Such attractive molecules are bound to become more common in various biotechnological and biomedical applications in the near future.

Selective Affimers Recognise the BCL-2 Family Proteins BCL-xL and MCL-1 through Noncanonical Structural Motifs*

The BCL-2 family is a challenging group of proteins to target selectively due to sequence and structural homologies across the family. Selective ligands for the BCL-2 family regulators of apoptosis are useful as probes to understand cell biology and apoptotic signalling pathways, and as starting points for inhibitor design. We have used phage display to isolate Affimer reagents (non-antibody-binding proteins based on a conserved scaffold) to identify ligands for MCL-1, BCL-xL , BCL-2, BAK and BAX, then used multiple biophysical characterisation methods to probe the interactions. We established that purified Affimers elicit selective recognition of their target BCL-2 protein. For anti-apoptotic targets BCL-xL and MCL-1, competitive inhibition of their canonical protein-protein interactions is demonstrated. Co-crystal structures reveal an unprecedented mode of molecular recognition; where a BH3 helix is normally bound, flexible loops from the Affimer dock into the BH3 binding cleft. Moreover, the Affimers induce a change in the target proteins towards a desirable drug-bound-like conformation. These proof-of-concept studies indicate that Affimers could be used as alternative templates to inspire the design of selective BCL-2 family modulators and more generally other protein-protein interaction inhibitors.

UMI-linked consensus sequencing enables phylogenetic analysis of directed evolution

The success of protein evolution campaigns is strongly dependent on the sequence context in which mutations are introduced, stemming from pervasive non-additive interactions between a protein's amino acids ('intra-gene epistasis'). Our limited understanding of such epistasis hinders the correct prediction of the functional contributions and adaptive potential of mutations. Here we present a straightforward unique molecular identifier (UMI)-linked consensus sequencing workflow (UMIC-seq) that simplifies mapping of evolutionary trajectories based on full-length sequences. Attaching UMIs to gene variants allows accurate consensus generation for closely related genes with nanopore sequencing. We exemplify the utility of this approach by reconstructing the artificial phylogeny emerging in three rounds of directed evolution of an amine dehydrogenase biocatalyst via ultrahigh throughput droplet screening. Uniquely, we are able to identify lineages and their founding variant, as well as non-additive interactions between mutations within a full gene showing sign epistasis. Access to deep and accurate long reads will facilitate prediction of key beneficial mutations and adaptive potential based on in silico analysis of large sequence datasets.

Computational Design of BH3-Mimetic Peptide Inhibitors That Can Bind Specifically to Mcl-1 or Bcl-XL: Role of Non-Hot Spot Residues

Interactions between pro- and anti-apoptotic Bcl-2 proteins decide the fate of the cell. The BH3 domain of pro-apoptotic Bcl-2 proteins interacts with the exposed hydrophobic groove of their anti-apoptotic counterparts. Through their design and development, BH3 mimetics that target the hydrophobic groove of specific anti-apoptotic Bcl-2 proteins have the potential to become anticancer drugs. We have developed a novel computational method for designing sequences with BH3 domain features that can bind specifically to anti-apoptotic Mcl-1 or Bcl-X_L. In this method, we retained the four highly conserved hydrophobic and aspartic residues of wild-type BH3 sequences and randomly substituted all other positions to generate a large number of BH3-like sequences. We modeled 20000 complex structures with Mcl-1 or Bcl-X_L using the BH3-like sequences derived from five wild-type pro-apoptotic BH3 peptides. Peptide-protein interaction energies calculated from these models for each set of BH3-like sequences resulted in negatively skewed extreme value distributions. The selected BH3-like sequences from the extreme negative tail regions have highly favorable interaction energies with Mcl-1 or Bcl-X_L. They are enriched in acidic and basic residues when they bind to Mcl-1 and Bcl-X_L, respectively. With the charged residues often away from the binding interface, the overall electric field generated by the charged residues results in strong long-range electrostatic interaction energies between the peptide and the protein giving rise to high specificity. Cell viability studies of representative BH3-like peptides further validated the predicted specificity. This study has revealed the importance of non-hot spot residues in BH3-mimetic peptides in providing specificity to a particular anti-apoptotic protein.

Uncovering the basis of protein-protein interaction specificity with a combinatorially complete library

Protein-protein interaction specificity is often encoded at the primary sequence level. However, the contributions of individual residues to specificity are usually poorly understood and often obscured by mutational robustness, sequence degeneracy, and epistasis. Using bacterial toxin-antitoxin systems as a model, we screened a combinatorially complete library of antitoxin variants at three key positions against two toxins. This library enabled us to measure the effect of individual substitutions on specificity in hundreds of genetic backgrounds. These distributions allow inferences about the general nature of interface residues in promoting specificity. We find that positive and negative contributions to specificity are neither inherently coupled nor mutually exclusive. Further, a wild-type antitoxin appears optimized for specificity as no substitutions improve discrimination between cognate and non-cognate partners. By comparing crystal structures of paralogous complexes, we provide a rationale for our observations. Collectively, this work provides a generalizable approach to understanding the logic of molecular recognition.

Comprehensive Prediction of Molecular Recognition in a Combinatorial Chemical Space Using Machine Learning

In combinatorial chemical approaches, optimizing the composition and arrangement of building blocks toward a particular function has been done using a number of methods, including high throughput molecular screening, molecular evolution, and computational prescreening. Here, a different approach is considered that uses sparse measurements of library molecules as the input to a machine learning algorithm which generates a comprehensive, quantitative relationship between covalent molecular structure and function that can then be used to predict the function of any molecule in the possible combinatorial space. To test the feasibility of the approach, a defined combinatorial chemical space consisting of ∼10¹² possible linear combinations of 16 different amino acids was used. The binding of a very sparse, but nearly random, sampling of this amino acid sequence space to 9 different protein targets is measured and used to generate a general relationship between peptide sequence and binding for each target. Surprisingly, measuring as little as a few hundred to a few thousand of the ∼10¹² possible molecules provides sufficient training to be highly predictive of the binding of the remaining molecules in the combinatorial space. Furthermore, measuring only amino acid sequences that bind weakly to a target allows the accurate prediction of which sequences will bind 10-100 times more strongly. Thus, the molecular recognition information contained in a tiny fraction of molecules in this combinatorial space is sufficient to characterize any set of molecules randomly selected from the entire space, a fact that potentially has significant implications for the design of new chemical function using combinatorial chemical libraries.

Quantitative mapping of binding specificity landscapes for homologous targets by using a high-throughput method

To facilitate investigations of protein-protein interactions (PPIs), we developed a novel platform for quantitative mapping of protein binding specificity landscapes, which combines the multi-target screening of a mutagenesis library into high- and low-affinity populations with sophisticated next-generation sequencing analysis. Importantly, this method generates accurate models to predict affinity and specificity values for any mutation within a protein complex, and requires only a few experimental binding affinity measurements using purified proteins for calibration. We demonstrated the utility of the approach by mapping quantitative landscapes for interactions between the N-terminal domain of the tissue inhibitor of metalloproteinase 2 (N-TIMP2) and three matrix metalloproteinases (MMPs) having homologous structures but different affinities (MMP-1, MMP-3, and MMP-14). The binding landscapes for N-TIMP2/MMP-1 and N-TIMP2/MMP-3 showed the PPIs to be almost fully optimized, with most single mutations giving a loss of affinity. In contrast, the non-optimized PPI for N-TIMP2/MMP-14 was reflected in a wide range of binding affinities, where single mutations exhibited a far more attenuated effect on the PPI. Our new platform reliably and comprehensively identified not only hot- and cold-spot residues, but also specificity-switch mutations that shape target affinity and specificity. Thus, our approach provides a methodology giving an unprecedentedly rich quantitative analysis of the binding specificity landscape, which will broaden the understanding of the mechanisms and evolutionary origins of specific PPIs and facilitate the rational design of specific inhibitors for structurally similar target proteins.

Predicting gene expression using morphological cell responses to nanotopography

Cells respond in complex ways to their environment, making it challenging to predict a direct relationship between the two. A key problem is the lack of informative representations of parameters that translate directly into biological function. Here we present a platform to relate the effects of cell morphology to gene expression induced by nanotopography. This platform utilizes the ‘morphome’, a multivariate dataset of cell morphology parameters. We create a Bayesian linear regression model that uses the morphome to robustly predict changes in bone, cartilage, muscle and fibrous gene expression induced by nanotopography. Furthermore, through this model we effectively predict nanotopography-induced gene expression from a complex co-culture microenvironment. The information from the morphome uncovers previously unknown effects of nanotopography on altering cell–cell interaction and osteogenic gene expression at the single cell level. The predictive relationship between morphology and gene expression arising from cell-material interaction shows promise for exploration of new topographies.

The surface nanotopography of biomaterials direct cell behavior, but screening for desired effects is inefficient. Here, the authors introduce a platform that enables prediction of nanotopography-induced gene expression changes from changes in cell morphology, including in co-culture environments.

Optimization of Ligands Using Focused DNA-Encoded Libraries To Develop a Selective, Cell-Permeable CBX8 Chromodomain Inhibitor

Polycomb repressive complex 1 (PRC1) is critical for mediating gene expression during development. Five chromobox (CBX) homolog proteins, CBX2, CBX4, CBX6, CBX7, and CBX8, are incorporated into PRC1 complexes, where they mediate targeting to trimethylated lysine 27 of histone H3 (H3K27me3) via the N-terminal chromodomain (ChD). Individual CBX paralogs have been implicated as drug targets in cancer; however, high similarities in sequence and structure among the CBX ChDs provide a major obstacle in developing selective CBX ChD inhibitors. Here we report the selection of small, focused, DNA-encoded libraries (DELs) against multiple homologous ChDs to identify modifications to a parental ligand that confer both selectivity and potency for the ChD of CBX8. This on-DNA, medicinal chemistry approach enabled the development of SW2_110A, a selective, cell-permeable inhibitor of the CBX8 ChD. SW2_110A binds CBX8 ChD with a K_d of 800 nM, with minimal 5-fold selectivity for CBX8 ChD over all other CBX paralogs in vitro. SW2_110A specifically inhibits the association of CBX8 with chromatin in cells and inhibits the proliferation of THP1 leukemia cells driven by the MLL-AF9 translocation. In THP1 cells, SW2_110A treatment results in a significant decrease in the expression of MLL-AF9 target genes, including HOXA9, validating the previously established role for CBX8 in MLL-AF9 transcriptional activation, and defining the ChD as necessary for this function. The success of SW2_110A provides great promise for the development of highly selective and cell-permeable probes for the full CBX family. In addition, the approach taken provides a proof-of-principle demonstration of how DELs can be used iteratively for optimization of both ligand potency and selectivity.

Generating quantitative binding landscapes through fractional binding selections combined with deep sequencing and data normalization

Quantifying the effects of various mutations on binding free energy is crucial for understanding the evolution of protein-protein interactions and would greatly facilitate protein engineering studies. Yet, measuring changes in binding free energy (ΔΔG_bind) remains a tedious task that requires expression of each mutant, its purification, and affinity measurements. We developed an attractive approach that allows us to quantify ΔΔG_bind for thousands of protein mutants in one experiment. Our protocol combines protein randomization, Yeast Surface Display technology, deep sequencing, and a few experimental ΔΔG_bind data points on purified proteins to generate ΔΔG_bind values for the remaining numerous mutants of the same protein complex. Using this methodology, we comprehensively map the single-mutant binding landscape of one of the highest-affinity interaction between BPTI and Bovine Trypsin (BT). We show that ΔΔG_bind for this interaction could be quantified with high accuracy over the range of 12 kcal mol⁻¹ displayed by various BPTI single mutants.

Quantifying the effect of mutations on binding free energy is important to understand protein-protein interaction (PPI). Here the authors develop a method based on yeast display and next-generation sequencing to generate quantitative binding landscapes for any PPI regardless of their Kd value.

Molecular evolution of peptides by yeast surface display technology

Genetically encoded peptides possess unique properties, such as a small molecular weight and ease of synthesis and modification, that make them suitable to a large variety of applications. However, despite these favorable qualities, naturally occurring peptides are often limited by intrinsic weak binding affinities, poor selectivity and low stability that ultimately restrain their final use. To overcome these limitations, a large variety of in vitro display methodologies have been developed over the past few decades to evolve genetically encoded peptide molecules with superior properties. Phage display, mRNA display, ribosome display, bacteria display, and yeast display are among the most commonly used methods to engineer peptides. While most of these in vitro methodologies have already been described in detail elsewhere, this review describes solely the yeast surface display technology and its valuable use for the evolution of a wide range of peptide formats.

Tertiary Structural Motif Sequence Statistics Enable Facile Prediction and Design of Peptides that Bind Anti-apoptotic Bfl-1 and Mcl-1

Understanding the relationship between protein sequence and structure well enough to design new proteins with desired functions is a longstanding goal in protein science. Here, we show that recurring tertiary structural motifs (TERMs) in the PDB provide rich information for protein-peptide interaction prediction and design. TERM statistics can be used to predict peptide binding energies for Bcl-2 family proteins as accurately as widely used structure-based tools. Furthermore, design using TERM energies (dTERMen) rapidly and reliably generates high-affinity peptide binders of anti-apoptotic proteins Bfl-1 and Mcl-1 with just 15%-38% sequence identity to any known native Bcl-2 family protein ligand. High-resolution structures of four designed peptides bound to their targets provide opportunities to analyze the strengths and limitations of the computational design method. Our results support dTERMen as a powerful approach that can complement existing tools for protein engineering.