Machine learning-enabled discovery and design of membrane-active peptides

An active machine learning discovery platform for membrane-disrupting and pore-forming peptides

Membrane-disrupting and pore-forming peptides (PFPs) play a substantial role in bionanotechnology and can determine the life and death of cells. The control of chemical and ion transport through cell membranes is essential to maintaining concentration gradients. Likewise, the delivery of drugs and intracellular proteins aided by pore-forming agents is of interest in treating malfunctioning cells. Known PFPs tend to be up to 50 residues in length, which is commensurate with the thickness of a lipid bilayer. Accordingly, few short PFPs are known. Here we show that the discovery of PFPs can be accelerated via an active machine learning approach. The approach identified 71 potential PFPs from the 25.6 billion octapeptide sequence space; 13 sequences were tested experimentally, and all were found to have the predicted membrane-disrupting ability, with 1 forming highly stable pores. Experimental verification of the predicted pore-forming ability demonstrated that a range of short peptides can form pores in membranes, while the positioning and characteristics of residues that favour pore-forming behaviour were identified. This approach identified more ultrashort (8-residues, unmodified, non-cyclic) PFPs than previously known. We anticipate our findings and methodology will be useful in discovering new pore-forming and membrane-disrupting peptides for a range of applications from nanoreactors to therapeutics.

Hollow CoFe Nanozymes Integrated with Oncolytic Peptides Designed via Machine-Learning for Tumor Therapy

Developing novel substances to synergize with nanozymes is a challenging yet indispensable task to enable the nanozyme-based therapeutics to tackle individual variations in tumor physicochemical properties. The advancement of machine learning (ML) has provided a useful tool to enhance the accuracy and efficiency in developing synergistic substances. In this study, ML models to mine low-cytotoxicity oncolytic peptides are applied. The filtering Pipeline is constructed using a traversal design and the Autogluon framework. Through the Pipeline, 37 novel peptides with high oncolytic activity against cancer cells and low cytotoxicity to normal cells are identified from a library of 25,740 sequences. Combining dataset testing with cytotoxicity experiments, an 80% accuracy rate is achieved, verifying the reliability of ML predictions. Peptide C2 is proven to possess membranolytic functions specifically for tumor cells as targeted by Pipeline. Then Peptide C2 with CoFe hollow hydroxide nanozyme (H-CF) to form the peptide/H-CF composite is integrated. The new composite exhibited acid-triggered membranolytic function and potent peroxidase-like (POD-like) activity, which induce ferroptosis to tumor cells and inhibits tumor growth. The study suggests that this novel ML-assisted design approach can offer an accurate and efficient paradigm for developing both oncolytic peptides and synergistic peptides for catalytic materials.

Machine learning designs new GCGR/GLP-1R dual agonists with enhanced biological potency

Several peptide dual agonists of the human glucagon receptor (GCGR) and the glucagon-like peptide-1 receptor (GLP-1R) are in development for the treatment of type 2 diabetes, obesity and their associated complications. Candidates must have high potency at both receptors, but it is unclear whether the limited experimental data available can be used to train models that accurately predict the activity at both receptors of new peptide variants. Here we use peptide sequence data labelled with in vitro potency at human GCGR and GLP-1R to train several models, including a deep multi-task neural-network model using multiple loss optimization. Model-guided sequence optimization was used to design three groups of peptide variants, with distinct ranges of predicted dual activity. We found that three of the model-designed sequences are potent dual agonists with superior biological activity. With our designs we were able to achieve up to sevenfold potency improvement at both receptors simultaneously compared to the best dual-agonist in the training set.

Identification of Potent siRNA Delivery Peptides Using Computer Modeling

Peptides with suitable aggregation behavior and electrical properties are potential siRNA delivery vectors. However, identifying suitable peptides with ideal delivery and safety features is difficult owing to the variations in amino acid sequences. Here, a holistic program based on computer modeling and single-cell RNA sequencing (scRNA-seq) is used to identify ideal siRNA delivery peptides. Stage one of this program consists of a sequential screening process for candidates with ideal assembly and delivery ability; stage two is a cell subtype-level analysis program that screens for high in vivo tissue safety. The leading candidate peptide selected from a library containing 12 amino acids showed strong lung-targeted siRNA delivery capacity after hydrophobic modification. Systemic administration of these compounds caused the least damage to liver and lung tissues and has little impact on macrophage and neutrophil numbers. By loading STAT3 siRNA, strong anticancer effects are achieved in multiple models, including patient-derived xenografts (PDX). This screening procedure may facilitate the development of peptide-based RNA interference (RNAi) therapeutics.

Protein-based nanoparticles for therapeutic nucleic acid delivery

To realize the full potential of emerging nucleic acid therapies, there is a need for effective delivery agents to transport cargo to cells of interest. Protein materials exhibit several unique properties, including biodegradability, biocompatibility, ease of functionalization via recombinant and chemical modifications, among other features, which establish a promising basis for therapeutic nucleic acid delivery systems. In this review, we highlight progress made in the use of non-viral protein-based nanoparticles for nucleic acid delivery in vitro and in vivo, while elaborating on key physicochemical properties that have enabled the use of these materials for nanoparticle formulation and drug delivery. To conclude, we comment on the prospects and unresolved challenges associated with the translation of protein-based nucleic acid delivery systems for therapeutic applications.

Machine learning assisted rational design of antimicrobial peptides based on human endogenous proteins and their applications for cosmetic preservative system optimization

Preservatives are essential components in cosmetic products, but their safety issues have attracted widespread attention. There is an urgent need for safe and effective alternatives. Antimicrobial peptides (AMPs) are part of the innate immune system and have potent antimicrobial properties. Using machine learning-assisted rational design, we obtained a novel antibacterial peptide, IK-16-1, with significant antibacterial activity and maintaining safety based on β-defensins. IK-16-1 has broad-spectrum antimicrobial properties against Escherichia coli, Staphylococcus aureus, Pseudomonas aeruginosa, and Candida albicans, and has no haemolytic activity. The use of IK-16-1 holds promise in the cosmetics industry, since it can serve as a preservative synergist to reduce the amount of other preservatives in cosmetics. This study verified the feasibility of combining computational design with artificial intelligence prediction to design AMPs, achieving rapid screening and reducing development costs.

In-Cell Penetration Selection-Mass Spectrometry Produces Noncanonical Peptides for Antisense Delivery

Peptide-mediated delivery of macromolecules in cells has significant potential therapeutic benefits, but no therapy employing cell-penetrating peptides (CPPs) has reached the market after 30 years of investigation due to challenges in the discovery of new, more efficient sequences. Here, we demonstrate a method for in-cell penetration selection-mass spectrometry (in-cell PS-MS) to discover peptides from a synthetic library capable of delivering macromolecule cargo to the cytosol. This method was inspired by recent in vivo selection approaches for cell-surface screening, with an added spatial dimension resulting from subcellular fractionation. A representative peptide discovered in the cytosolic extract, Cyto1a, is nearly 100-fold more active toward antisense phosphorodiamidate morpholino oligomer (PMO) delivery compared to a sequence identified from a whole cell extract, which includes endosomes. Cyto1a is composed of d-residues and two non-α-amino acids, is more stable than its all-l isoform, and is less toxic than known CPPs with comparable activity. Pulse-chase and microscopy experiments revealed that while the PMO-Cyto1a conjugate is likely taken up by endosomes, it can escape to localize to the nucleus without nonspecifically releasing other endosomal components. In-cell PS-MS introduces a means to empirically discover unnatural synthetic peptides for subcellular delivery of therapeutically relevant cargo.

Bacteria-Specific Feature Selection for Enhanced Antimicrobial Peptide Activity Predictions Using Machine-Learning Methods

There are several classes of short peptide molecules, known as antimicrobial peptides (AMPs), which are produced during the immune responses of living organisms against various infections. In recent years, substantial progress has been achieved in applying machine-learning methods to predict the activities of AMPs against bacteria. In most investigated cases, however, the outcome is not bacterium-specific since the specific features of bacteria, such as chemical composition and structure of membranes, are not considered. To overcome this problem, we developed a new computational approach that allowed us to train several supervised machine-learning models using a specific set of data associated with peptides targeting E. coli bacteria. LASSO regression and Support Vector Machine techniques have been utilized to select, among more than 1500 physicochemical descriptors, the most important features that can be used to classify a peptide as antimicrobial or ineffective against E. coli. We then performed the classification of active versus inactive AMPs using the Support Vector classifiers, Logistic Regression, and Random Forest methods. This computational study allows us to make recommendations of how to design more efficient antibacterial drug therapies.

Targeted nanomedicine: Lessons learned and future directions

Designing a therapeutic modality that will reach a certain organ, tissue, or cell type is crucial for both the therapeutic efficiency and to limit off-target adverse effects. Nanoparticles carrying various drugs, such as nucleic acids, small molecules and proteins, are promoting modalities to this end. Beyond the need to identify a target for a specific indication, an adequate design has to address the multiple biological barriers, such as systemic barriers, dilution and unspecific distribution, tissue penetration and intracellular trafficking. The field of targeted delivery has developed rapidly in recent years, with tremendous progress made in understating the biological barriers, and new technologies to functionalize nanoparticles with targeting moieties for an accurate, specific and highly selective delivery. Implementing new approaches like multi-functionalized nanocarriers and machine learning models will advance the field for designing safe, cell -specific nanoparticle delivery systems. Here, we will critically review the current progress in the field and suggest novel strategies to improve cell specific delivery of therapeutic payloads.

Rational design of stapled antimicrobial peptides

The global increase in antimicrobial drug resistance has dramatically reduced the effectiveness of traditional antibiotics. Structurally diverse antibiotics are urgently needed to combat multiple-resistant bacterial infections. As part of innate immunity, antimicrobial peptides have been recognized as the most promising candidates because they comprise diverse sequences and mechanisms of action and have a relatively low induction rate of resistance. However, because of their low chemical stability, susceptibility to proteases, and high hemolytic effect, their usage is subject to many restrictions. Chemical modifications such as D-amino acid substitution, cyclization, and unnatural amino acid modification have been used to improve the stability of antimicrobial peptides for decades. Among them, a side-chain covalent bridge modification, the so-called stapled peptide, has attracted much attention. The stapled side-chain bridge stabilizes the secondary structure, induces protease resistance, and increases cell penetration and biological activity. Recent progress in computer-aided drug design and artificial intelligence methods has also been used in the design of stapled antimicrobial peptides and has led to the successful discovery of many prospective peptides. This article reviews the possible structure-activity relationships of stapled antimicrobial peptides, the physicochemical properties that influence their activity (such as net charge, hydrophobicity, helicity, and dipole moment), and computer-aided methods of stapled peptide design. Antimicrobial peptides under clinical trial: Pexiganan (NCT01594762, 2012-05-07). Omiganan (NCT02576847, 2015-10-13).

Incorporation of Non-Canonical Amino Acids into Antimicrobial Peptides: Advances, Challenges, and Perspectives

The emergence of antimicrobial resistance is a global health concern and calls for the development of novel antibiotic agents. Antimicrobial peptides seem to be promising candidates due to their diverse sources, mechanisms of action, and physicochemical characteristics, as well as the relatively low emergence of resistance. The incorporation of noncanonical amino acids into antimicrobial peptides could effectively improve their physicochemical and pharmacological diversity. Recently, various antimicrobial peptides variants with improved or novel properties have been produced by the incorporation of single and multiple distinct noncanonical amino acids. In this review, we summarize strategies for the incorporation of noncanonical amino acids into antimicrobial peptides, as well as their features and suitabilities. Recent applications of noncanonical amino acid incorporation into antimicrobial peptides are also presented. Finally, we discuss the related challenges and prospects.

Improvement of the Antibacterial Activity of Phage Lysin-Derived Peptide P87 through Maximization of Physicochemical Properties and Assessment of Its Therapeutic Potential

Phage lysins are a promising alternative to common antibiotic chemotherapy. However, they have been regarded as less effective against Gram-negative pathogens unless engineered, e.g., by fusing them to antimicrobial peptides (AMPs). AMPs themselves pose an alternative to antibiotics. In this work, AMP P87, previously derived from a phage lysin (Pae87) with a presumed nonenzymatic mode-of-action, was investigated to improve its antibacterial activity. Five modifications were designed to maximize the hydrophobic moment and net charge, producing the modified peptide P88, which was evaluated in terms of bactericidal activity, cytotoxicity, MICs or synergy with antibiotics. P88 had a better bactericidal performance than P87 (an average of 6.0 vs. 1.5 log-killing activity on Pseudomonas aeruginosa strains treated with 10 µM). This did not correlate with a dramatic increase in cytotoxicity as assayed on A549 cell cultures. P88 was active against a range of P. aeruginosa isolates, with no intrinsic resistance factors identified. Synergy with some antibiotics was observed in vitro, in complex media, and in a respiratory infection mouse model. Therefore, P88 can be a new addition to the therapeutic toolbox of alternative antimicrobials against Gram-negative pathogens as a sole therapeutic, a complement to antibiotics, or a part to engineer proteinaceous antimicrobials.

Machine Learning in Antibacterial Drug Design

Advances in computer hardware and the availability of high-performance supercomputing platforms and parallel computing, along with artificial intelligence methods are successfully complementing traditional approaches in medicinal chemistry. In particular, machine learning is gaining importance with the growth of the available data collections. One of the critical areas where this methodology can be successfully applied is in the development of new antibacterial agents. The latter is essential because of the high attrition rates in new drug discovery, both in industry and in academic research programs. Scientific involvement in this area is even more urgent as antibacterial drug resistance becomes a public health concern worldwide and pushes us increasingly into the post-antibiotic era. In this review, we focus on the latest machine learning approaches used in the discovery of new antibacterial agents and targets, covering both small molecules and antibacterial peptides. For the benefit of the reader, we summarize all applied machine learning approaches and available databases useful for the design of new antibacterial agents and address the current shortcomings.

Research progress in the screening and evaluation of umami peptides

Umami is an important element affecting food taste, and the development of umami peptides is a topic of interest in food-flavoring research. The existing technology used for traditional screening of umami peptides is time-consuming and labor-intensive, making it difficult to meet the requirements of high-throughput screening, which limits the rapid development of umami peptides. The difficulty in performing a standard measurement of umami intensity is another problem that restricts the development of umami peptides. The existing methods are not sensitive and specific, making it difficult to achieve a standard evaluation of umami taste. This review summarizes the umami receptors and umami peptides, focusing on the problems restricting the development of umami peptides, high-throughput screening, and establishment of evaluation standards. The rapid screening of umami peptides was realized based on molecular docking technology and a machine learning method, and the standard evaluation of umami could be realized with a bionic taste sensor. The progress of rapid screening and evaluation methods significantly promotes the study of umami peptides and increases its application in the seasoning industry.

Machine Learning-Assisted Design of Material Properties

Designing functional materials requires a deep search through multidimensional spaces for system parameters that yield desirable material properties. For cases where conventional parameter sweeps or trial-and-error sampling are impractical, inverse methods that frame design as a constrained optimization problem present an attractive alternative. However, even efficient algorithms require time- and resource-intensive characterization of material properties many times during optimization, imposing a design bottleneck. Approaches that incorporate machine learning can help address this limitation and accelerate the discovery of materials with targeted properties. In this article, we review how to leverage machine learning to reduce dimensionality in order to effectively explore design space, accelerate property evaluation, and generate unconventional material structures with optimal properties. We also discuss promising future directions, including integration of machine learning into multiple stages of a design algorithm and interpretation of machine learning models to understand how design parameters relate to material properties. Expected final online publication date for the Annual Review of Chemical and Biomolecular Engineering, Volume 13 is October 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

PTML modeling for peptide discovery: in silico design of non-hemolytic peptides with antihypertensive activity

Hypertension is a medical condition that affects millions of people worldwide. Despite the high efficacy of the current antihypertensive drugs, they are associated with serious side effects. Peptides constitute attractive options for chemical therapy against hypertension, and computational models can accelerate the design of antihypertensive peptides. Yet, to the best of our knowledge, all the in silico models predict only the antihypertensive activity of peptides while neglecting their inherent toxic potential to red blood cells. In this work, we report the first sequence-based model that combines perturbation theory and machine learning through multilayer perceptron networks (SB-PTML-MLP) to enable the simultaneous screening of antihypertensive activity and hemotoxicity of peptides. We have interpreted the molecular descriptors present in the model from a physicochemical and structural point of view. By strictly following such interpretations as guidelines, we performed two tasks. First, we selected amino acids with favorable contributions to both the increase of the antihypertensive activity and the diminution of hemotoxicity. Then, we assembled those suitable amino acids, virtually designing peptides that were predicted by the SB-PTML-MLP model as antihypertensive agents exhibiting low hemotoxicity. The potentiality of the SB-PTML-MLP model as a tool for designing potent and safe antihypertensive peptides was confirmed by predictions performed by online computational tools reported in the scientific literature. The methodology presented here can be extended to other pharmacological applications of peptides.

Deep learning to design nuclear-targeting abiotic miniproteins

There are more amino acid permutations within a 40-residue sequence than atoms on Earth. This vast chemical search space hinders the use of human learning to design functional polymers. Here we show how machine learning enables the de novo design of abiotic nuclear-targeting miniproteins to traffic antisense oligomers to the nucleus of cells. We combined high-throughput experimentation with a directed evolution-inspired deep-learning approach in which the molecular structures of natural and unnatural residues are represented as topological fingerprints. The model is able to predict activities beyond the training dataset, and simultaneously deciphers and visualizes sequence-activity predictions. The predicted miniproteins, termed 'Mach', reach an average mass of 10 kDa, are more effective than any previously known variant in cells and can also deliver proteins into the cytosol. The Mach miniproteins are non-toxic and efficiently deliver antisense cargo in mice. These results demonstrate that deep learning can decipher design principles to generate highly active biomolecules that are unlikely to be discovered by empirical approaches.

Taxonomic bias in AMP prediction of invertebrate peptides

Invertebrate antimicrobial peptides (AMPs) are at the forefront in the search for agents of therapeutic utility against multi-resistant microbial pathogens, and in recent years substantial advances took place in the in silico prediction of antimicrobial function of amino acid sequences. A yet neglected aspect is taxonomic bias in the performance of these tools. Owing to differences in the prediction algorithms and used training data sets between tools, and phylogenetic differences in sequence diversity, physicochemical properties and evolved biological functions of AMPs between taxa, notable discrepancies may exist in performance between the currently available prediction tools. Here we tested if there is taxonomic bias in the prediction power in 10 tools with a total of 20 prediction algorithms in 19 invertebrate taxa, using a data set containing 1525 AMP and 3050 non-AMP sequences. We found that most of the tools exhibited considerable variation in performance between tested invertebrate groups. Based on the per-taxa performances and on the variation in performances across taxa we provide guidance in choosing the best-performing prediction tool for all assessed taxa, by listing the highest scoring tool for each of them.

PyPEF-An Integrated Framework for Data-Driven Protein Engineering

Data-driven strategies are gaining increased attention in protein engineering due to recent advances in access to large experimental databanks of proteins, next-generation sequencing (NGS), high-throughput screening (HTS) methods, and the development of artificial intelligence algorithms. However, the reliable prediction of beneficial amino acid substitutions, their combination, and the effect on functional properties remain the most significant challenges in protein engineering, which is applied to develop proteins and enzymes for biocatalysis, biomedicine, and life sciences. Here, we present a general-purpose framework (PyPEF: pythonic protein engineering framework) for performing data-driven protein engineering using machine learning methods combined with techniques from signal processing and statistical physics. PyPEF guides the identification and selection of beneficial proteins of a defined sequence space by systematically or randomly exploring the fitness of variants and by sampling random evolution pathways. The performance of PyPEF was evaluated concerning its predictive accuracy and throughput on four public protein and enzyme data sets using common regression models. It was proved that the program could efficiently predict the fitness of protein sequences for different target properties (predictive models with coefficient of determination values ranging from 0.58 to 0.92). By combining machine learning and protein evolution, PyPEF enabled the screening of proteins with various functions, reaching a screening capacity of more than 500,000 protein sequence variants in the timeframe of only a few minutes on a personal computer. PyPEF displayed significant accuracies on four public data sets (different proteins and properties) and underlined the potential of integrating data-driven technologies for covering different philosophies by either predicting the fitness of the variants to the highest accuracy accounting for epistatic effects or capturing the general trend of introduced mutations on the fitness in directed protein evolution campaigns. In essence, PyPEF can provide a powerful solution to current sequence exploration and combinatorial problems faced in protein engineering through exhaustive in silico screening of the sequence space.

Antimicrobial/anticancer peptides: bioactive molecules and therapeutic agents

Antimicrobial peptides (AMPs) have been known as host-defense peptides. These cationic and amphipathic peptides are relatively short (∼5-50 L-amino acids) with molecular weight less than 10 kDa. AMPs have various roles including immunomodulatory, angiogenic and antitumor activities. Anticancer peptides (ACPs) are a main subset of AMPs as a novel therapeutic approach against tumor cells. The physicochemical properties of the ACPs influence their cell penetration, stability and efficiency of targeting. Up to now, several databases and web servers for in silico prediction of AMPs/ACPs have been established prior to the lab analysis. The present review focuses on the recent advancement about AMPs/ACPs activities including their in silico prediction by computational tools and their potential applications as therapeutic agents especially in cancer.

Recent advances in design of antimicrobial peptides and polypeptides toward clinical translation

The recent outbreaks of infectious diseases caused by multidrug-resistant pathogens have sounded a piercing alarm for the need of new effective antimicrobial agents to guard public health. Among different types of candidates, antimicrobial peptides (AMPs) and the synthetic mimics of AMPs (SMAMPs) have attracted significant enthusiasm in the past thirty years, due to their unique membrane-active antimicrobial mechanism and broad-spectrum antimicrobial activity. The extensive research has brought many drug candidates into clinical and pre-clinical development. Despite tremendous progresses have been made, several major challenges inherent to current design strategies have slowed down the clinical translational development of AMPs and SMAMPs. However, these challenges also triggered many efforts to redesign and repurpose AMPs. In this review, we will first give an overview on AMPs and their synthetic mimics, and then discuss the current status of their clinical translation. Finally, the recent advances in redesign and repurposing AMPs and SMAMPs are highlighted.

Molecular engineering of antimicrobial peptides: microbial targets, peptide motifs and translation opportunities

The global public health threat of antimicrobial resistance has led the scientific community to highly engage into research on alternative strategies to the traditional small molecule therapeutics. Here, we review one of the most popular alternatives amongst basic and applied research scientists, synthetic antimicrobial peptides. The ease of peptide chemical synthesis combined with emerging engineering principles and potent broad-spectrum activity, including against multidrug-resistant strains, has motivated intense scientific focus on these compounds for the past decade. This global effort has resulted in significant advances in our understanding of peptide antimicrobial activity at the molecular scale. Recent evidence of molecular targets other than the microbial lipid membrane, and efforts towards consensus antimicrobial peptide motifs, have supported the rise of molecular engineering approaches and design tools, including machine learning. Beyond molecular concepts, supramolecular chemistry has been lately added to the debate; and helped unravel the impact of peptide self-assembly on activity, including on biofilms and secondary targets, while providing new directions in pharmaceutical formulation through taking advantage of peptide self-assembled nanostructures. We argue that these basic research advances constitute a solid basis for promising industry translation of rationally designed synthetic peptide antimicrobials, not only as novel drugs against multidrug-resistant strains but also as components of emerging antimicrobial biomaterials. This perspective is supported by recent developments of innovative peptide-based and peptide-carrier nanobiomaterials that we also review.

PACAP is a pathogen-inducible resident antimicrobial neuropeptide affording rapid and contextual molecular host defense of the brain

Defense of the central nervous system (CNS) against infection must be accomplished without generation of potentially injurious immune cell-mediated or off-target inflammation which could impair key functions. As the CNS is an immune-privileged compartment, inducible innate defense mechanisms endogenous to the CNS likely play an essential role in this regard. Pituitary adenylate cyclase-activating polypeptide (PACAP) is a neuropeptide known to regulate neurodevelopment, emotion, and certain stress responses. While PACAP is known to interact with the immune system, its significance in direct defense of brain or other tissues is not established. Here, we show that our machine-learning classifier can screen for immune activity in neuropeptides, and correctly identified PACAP as an antimicrobial neuropeptide in agreement with previous experimental work. Furthermore, synchrotron X-ray scattering, antimicrobial assays, and mechanistic fingerprinting provided precise insights into how PACAP exerts antimicrobial activities vs. pathogens via multiple and synergistic mechanisms, including dysregulation of membrane integrity and energetics and activation of cell death pathways. Importantly, resident PACAP is selectively induced up to 50-fold in the brain in mouse models of Staphylococcus aureus or Candida albicans infection in vivo, without inducing immune cell infiltration. We show differential PACAP induction even in various tissues outside the CNS, and how these observed patterns of induction are consistent with the antimicrobial efficacy of PACAP measured in conditions simulating specific physiologic contexts of those tissues. Phylogenetic analysis of PACAP revealed close conservation of predicted antimicrobial properties spanning primitive invertebrates to modern mammals. Together, these findings substantiate our hypothesis that PACAP is an ancient neuro-endocrine-immune effector that defends the CNS against infection while minimizing potentially injurious neuroinflammation.

In search of a novel chassis material for synthetic cells: emergence of synthetic peptide compartment

Giant lipid vesicles have been used extensively as a synthetic cell model to recapitulate various life-like processes, including in vitro protein synthesis, DNA replication, and cytoskeleton organization. Cell-sized lipid vesicles are mechanically fragile in nature and prone to rupture due to osmotic stress, which limits their usability. Recently, peptide vesicles have been introduced as a synthetic cell model that would potentially overcome the aforementioned limitations. Peptide vesicles are robust, reasonably more stable than lipid vesicles and can withstand harsh conditions including pH, thermal, and osmotic variations. This mini-review summarizes the current state-of-the-art in the design, engineering, and realization of peptide-based chassis materials, including both experimental and computational work. We present an outlook for simulation-aided and data-driven design and experimental realization of engineered and multifunctional synthetic cells.

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.

Discovery and mechanistic characterization of a structurally-unique membrane active peptide

Membrane active peptides (MAPs) have gained wide interest due to their far reaching applications in drug discovery and drug delivery. The search for new MAPs, however, has been largely skewed with bias selecting for physicochemical parameters believed to be important for membrane activity, such as alpha helicity, cationicity and hydrophobicity. Here we carry out a search-and-find strategy to screen a 100,000-membered one-bead-one-compound (OBOC) combinatorial peptide library for lead compounds, agnostic of those physicochemical constraints. Such a synthetic strategy also permits expansion of our peptide repertoire to include unnatural amino acids. Using this approach, we discovered a structurally unique lead peptide LBF14, a linear 14-mer peptide, that induces gross morphological disruption of membranes, irrespective of membrane composition. Further, we demonstrate that the unique insertion mechanism of the peptide, visualized by spinning disc confocal microscopy and further analyzed by electron paramagnetic resonance measurements, may be the cause of this large scale membrane deformation. We also demonstrate the robustness, reproducibility, and potential application of this technique to discover and characterize new membrane active peptides that display activity by local insertion and subsequent allosteric effects leading to global membrane disruption.

Discovery of Novel Type II Bacteriocins Using a New High-Dimensional Bioinformatic Algorithm

Antimicrobial compounds first arose in prokaryotes by necessity for competitive self-defense. In this light, prokaryotes invented the first host defense peptides. Among the most well-characterized of these peptides are class II bacteriocins, ribosomally-synthesized polypeptides produced chiefly by Gram-positive bacteria. In the current study, a tensor search protocol-the BACIIα algorithm-was created to identify and classify bacteriocin sequences with high fidelity. The BACIIα algorithm integrates a consensus signature sequence, physicochemical and genomic pattern elements within a high-dimensional query tool to select for bacteriocin-like peptides. It accurately retrieved and distinguished virtually all families of known class II bacteriocins, with an 86% specificity. Further, the algorithm retrieved a large set of unforeseen, putative bacteriocin peptide sequences. A recently-developed machine-learning classifier predicted the vast majority of retrieved sequences to induce negative Gaussian curvature in target membranes, a hallmark of antimicrobial activity. Prototypic bacteriocin candidate sequences were synthesized and demonstrated potent antimicrobial efficacy in vitro against a broad spectrum of human pathogens. Therefore, the BACIIα algorithm expands the scope of prokaryotic host defense bacteriocins and enables an innovative bioinformatics discovery strategy. Understanding how prokaryotes have protected themselves against microbial threats over eons of time holds promise to discover novel anti-infective strategies to meet the challenge of modern antibiotic resistance.

Functional Reciprocity of Amyloids and Antimicrobial Peptides: Rethinking the Role of Supramolecular Assembly in Host Defense, Immune Activation, and Inflammation

Pathological self-assembly is a concept that is classically associated with amyloids, such as amyloid-β (Aβ) in Alzheimer's disease and α-synuclein in Parkinson's disease. In prokaryotic organisms, amyloids are assembled extracellularly in a similar fashion to human amyloids. Pathogenicity of amyloids is attributed to their ability to transform into several distinct structural states that reflect their downstream biological consequences. While the oligomeric forms of amyloids are thought to be responsible for their cytotoxicity via membrane permeation, their fibrillar conformations are known to interact with the innate immune system to induce inflammation. Furthermore, both eukaryotic and prokaryotic amyloids can self-assemble into molecular chaperones to bind nucleic acids, enabling amplification of Toll-like receptor (TLR) signaling. Recent work has shown that antimicrobial peptides (AMPs) follow a strikingly similar paradigm. Previously, AMPs were thought of as peptides with the primary function of permeating microbial membranes. Consistent with this, many AMPs are facially amphiphilic and can facilitate membrane remodeling processes such as pore formation and fusion. We show that various AMPs and chemokines can also chaperone and organize immune ligands into amyloid-like ordered supramolecular structures that are geometrically optimized for binding to TLRs, thereby amplifying immune signaling. The ability of amphiphilic AMPs to self-assemble cooperatively into superhelical protofibrils that form structural scaffolds for the ordered presentation of immune ligands like DNA and dsRNA is central to inflammation. It is interesting to explore the notion that the assembly of AMP protofibrils may be analogous to that of amyloid aggregates. Coming full circle, recent work has suggested that Aβ and other amyloids also have AMP-like antimicrobial functions. The emerging perspective is one in which assembly affords a more finely calibrated system of recognition and response: the detection of single immune ligands, immune ligands bound to AMPs, and immune ligands spatially organized to varying degrees by AMPs, result in different immunologic outcomes. In this framework, not all ordered structures generated during multi-stepped AMP (or amyloid) assembly are pathological in origin. Supramolecular structures formed during this process serve as signatures to the innate immune system to orchestrate immune amplification in a proportional, situation-dependent manner.

The value of antimicrobial peptides in the age of resistance

Accelerating growth and global expansion of antimicrobial resistance has deepened the need for discovery of novel antimicrobial agents. Antimicrobial peptides have clear advantages over conventional antibiotics which include slower emergence of resistance, broad-spectrum antibiofilm activity, and the ability to favourably modulate the host immune response. Broad bacterial susceptibility to antimicrobial peptides offers an additional tool to expand knowledge about the evolution of antimicrobial resistance. Structural and functional limitations, combined with a stricter regulatory environment, have hampered the clinical translation of antimicrobial peptides as potential therapeutic agents. Existing computational and experimental tools attempt to ease the preclinical and clinical development of antimicrobial peptides as novel therapeutics. This Review identifies the benefits, challenges, and opportunities of using antimicrobial peptides against multidrug-resistant pathogens, highlights advances in the deployment of novel promising antimicrobial peptides, and underlines the needs and priorities in designing focused development strategies taking into account the most advanced tools available.

Antimicrobial Susceptibility Testing of Antimicrobial Peptides to Better Predict Efficacy

During the development of antimicrobial peptides (AMP) as potential therapeutics, antimicrobial susceptibility testing (AST) stands as an essential part of the process in identification and optimisation of candidate AMP. Standard methods for AST, developed almost 60 years ago for testing conventional antibiotics, are not necessarily fit for purpose when it comes to determining the susceptibility of microorganisms to AMP. Without careful consideration of the parameters comprising AST there is a risk of failing to identify novel antimicrobials at a time when antimicrobial resistance (AMR) is leading the planet toward a post-antibiotic era. More physiologically/clinically relevant AST will allow better determination of the preclinical activity of drug candidates and allow the identification of lead compounds. An important consideration is the efficacy of AMP in biological matrices replicating sites of infection, e.g., blood/plasma/serum, lung bronchiolar lavage fluid/sputum, urine, biofilms, etc., as this will likely be more predictive of clinical efficacy. Additionally, specific AST for different target microorganisms may help to better predict efficacy of AMP in specific infections. In this manuscript, we describe what we believe are the key considerations for AST of AMP and hope that this information can better guide the preclinical development of AMP toward becoming a new generation of urgently needed antimicrobials.

Drug Conjugation Induced Modulation of Structural and Membrane Interaction Features of Cationic Cell-Permeable Peptides

Cell-penetrating peptides might have great potential for enhancing the therapeutic effect of drug molecules against such dangerous pathogens as Mycobacterium tuberculosis (Mtb), which causes a major health problem worldwide. A set of cationic cell-penetration peptides with various hydrophobicity were selected and synthesized as drug carrier of isoniazid (INH), a first-line antibacterial agent against tuberculosis. Molecular interactions between the peptides and their INH-conjugates with cell-membrane-forming lipid layers composed of DPPC and mycolic acid (a characteristic component of Mtb cell wall) were evaluated, using the Langmuir balance technique. Secondary structure of the INH conjugates was analyzed and compared to that of the native peptides by circular dichroism spectroscopic experiments performed in aqueous and membrane mimetic environment. A correlation was found between the conjugation induced conformational and membrane affinity changes of the INH-peptide conjugates. The degree and mode of interaction were also characterized by AFM imaging of penetrated lipid layers. In vitro biological evaluation was performed with Penetratin and Transportan conjugates. Results showed similar internalization rate into EBC-1 human squamous cell carcinoma, but markedly different subcellular localization and activity on intracellular Mtb.

In silico design and optimization of selective membranolytic anticancer peptides

Membranolytic anticancer peptides represent a potential strategy in the fight against cancer. However, our understanding of the underlying structure-activity relationships and the mechanisms driving their cell selectivity is still limited. We developed a computational approach as a step towards the rational design of potent and selective anticancer peptides. This machine learning model distinguishes between peptides with and without anticancer activity. This classifier was experimentally validated by synthesizing and testing a selection of 12 computationally generated peptides. In total, 83% of these predictions were correct. We then utilized an evolutionary molecular design algorithm to improve the peptide selectivity for cancer cells. This simulated molecular evolution process led to a five-fold selectivity increase with regard to human dermal microvascular endothelial cells and more than ten-fold improvement towards human erythrocytes. The results of the present study advocate for the applicability of machine learning models and evolutionary algorithms to design and optimize novel synthetic anticancer peptides with reduced hemolytic liability and increased cell-type selectivity.

Toward computer-made artificial antibiotics

Merging concepts from synthetic biology and computational biology may yield antibiotics that are less likely to elicit resistance than existing drugs and that yet can fight drug-resistant infections. Indeed, computer-guided strategies coupled with massively parallel high-throughput experimental methods represent a new paradigm for antibiotic discovery. Infections caused by multidrug-resistant microorganisms are increasingly deadly. In the current post-antibiotic era, many of these infections cannot be treated with our existing antimicrobial arsenal. Furthermore, we may have already exhausted the category of large molecules produced in nature having antimicrobial activity: the antibiotic scaffolds we have discovered so far may represent the majority of those that exist. The rise in drug-resistant bacteria and lack of new antibiotic classes clearly call for out-of-the-box strategies. Recent advances in computational synthetic biology have enabled the development of antimicrobials. New molecular descriptors and genetic and pattern recognition algorithms are powerful tools that bring us a step closer to developing efficient antibiotics. We review several computational tools for drug design and a number of recently generated antibiotic candidates, with an emphasis on peptide-based molecules. Design strategies can generate a diversity of synthetic antimicrobial peptides, which may help to mitigate the spread of resistance and combat multidrug-resistant microorganisms.

How Melittin Inserts into Cell Membrane: Conformational Changes, Inter-Peptide Cooperation, and Disturbance on the Membrane

Antimicrobial peptides (AMPs), as a key component of the immune defense systems of organisms, are a promising solution to the serious threat of drug-resistant bacteria to public health. As one of the most representative and extensively studied AMPs, melittin has exceptional broad-spectrum activities against microorganisms, including both Gram-positive and Gram-negative bacteria. Unfortunately, the action mechanism of melittin with bacterial membranes, especially the underlying physics of peptide-induced membrane poration behaviors, is still poorly understood, which hampers efforts to develop melittin-based drugs or agents for clinical applications. In this mini-review, we focus on recent advances with respect to the membrane insertion behavior of melittin mostly from a computational aspect. Membrane insertion is a prerequisite and key step for forming transmembrane pores and bacterial killing by melittin, whose occurrence is based on overcoming a high free-energy barrier during the transition of melittin molecules from a membrane surface-binding state to a transmembrane-inserting state. Here, intriguing simulation results on such transition are highlighted from both kinetic and thermodynamic aspects. The conformational changes and inter-peptide cooperation of melittin molecules, as well as melittin-induced disturbances to membrane structure, such as deformation and lipid extraction, are regarded as key factors influencing the insertion of peptides into membranes. The associated intermediate states in peptide conformations, lipid arrangements, membrane structure, and mechanical properties during this process are specifically discussed. Finally, potential strategies for enhancing the poration ability and improving the antimicrobial performance of AMPs are included as well.

Externalized histone H4 orchestrates chronic inflammation by inducing lytic cell death

The perpetuation of inflammation is an important pathophysiological contributor to the global medical burden. Chronic inflammation is promoted by non-programmed cell death1,2; however, how inflammation is instigated, its cellular and molecular mediators, and its therapeutic value are poorly defined. Here we use mouse models of atherosclerosis-a major underlying cause of mortality worldwide-to demonstrate that extracellular histone H4-mediated membrane lysis of smooth muscle cells (SMCs) triggers arterial tissue damage and inflammation. We show that activated lesional SMCs attract neutrophils, triggering the ejection of neutrophil extracellular traps that contain nuclear proteins. Among them, histone H4 binds to and lyses SMCs, leading to the destabilization of plaques; conversely, the neutralization of histone H4 prevents cell death of SMCs and stabilizes atherosclerotic lesions. Our data identify a form of cell death found at the core of chronic vascular disease that is instigated by leukocytes and can be targeted therapeutically.

Incorporation of non-standard amino acids into proteins: challenges, recent achievements, and emerging applications

The natural genetic code only allows for 20 standard amino acids in protein translation, but genetic code reprogramming enables the incorporation of non-standard amino acids (NSAAs). Proteins containing NSAAs provide enhanced or novel properties and open diverse applications. With increased attention to the recent advancements in synthetic biology, various improved and novel methods have been developed to incorporate single and multiple distinct NSAAs into proteins. However, various challenges remain in regard to NSAA incorporation, such as low yield and misincorporation. In this review, we summarize the recent efforts to improve NSAA incorporation by utilizing orthogonal translational system optimization, cell-free protein synthesis, genomically recoded organisms, artificial codon boxes, quadruplet codons, and orthogonal ribosomes, before closing with a discussion of the emerging applications of NSAA incorporation.

Unifying structural signature of eukaryotic α-helical host defense peptides

Diversity of α-helical host defense peptides (αHDPs) contributes to immunity against a broad spectrum of pathogens via multiple functions. Thus, resolving common structure-function relationships among αHDPs is inherently difficult, even for artificial-intelligence-based methods that seek multifactorial trends rather than foundational principles. Here, bioinformatic and pattern recognition methods were applied to identify a unifying signature of eukaryotic αHDPs derived from amino acid sequence, biochemical, and three-dimensional properties of known αHDPs. The signature formula contains a helical domain of 12 residues with a mean hydrophobic moment of 0.50 and favoring aliphatic over aromatic hydrophobes in 18-aa windows of peptides or proteins matching its semantic definition. The holistic α-core signature subsumes existing physicochemical properties of αHDPs, and converged strongly with predictions of an independent machine-learning-based classifier recognizing sequences inducing negative Gaussian curvature in target membranes. Queries using the α-core formula identified 93% of all annotated αHDPs in proteomic databases and retrieved all major αHDP families. Synthesis and antimicrobial assays confirmed efficacies of predicted sequences having no previously known antimicrobial activity. The unifying α-core signature establishes a foundational framework for discovering and understanding αHDPs encompassing diverse structural and mechanistic variations, and affords possibilities for deterministic design of antiinfectives.

Helical antimicrobial peptides assemble into protofibril scaffolds that present ordered dsDNA to TLR9

Amphiphilicity in ɑ-helical antimicrobial peptides (AMPs) is recognized as a signature of potential membrane activity. Some AMPs are also strongly immunomodulatory: LL37-DNA complexes potently amplify Toll-like receptor 9 (TLR9) activation in immune cells and exacerbate autoimmune diseases. The rules governing this proinflammatory activity of AMPs are unknown. Here we examine the supramolecular structures formed between DNA and three prototypical AMPs using small angle X-ray scattering and molecular modeling. We correlate these structures to their ability to activate TLR9 and show that a key criterion is the AMP's ability to assemble into superhelical protofibril scaffolds. These structures enforce spatially-periodic DNA organization in nanocrystalline immunocomplexes that trigger strong recognition by TLR9, which is conventionally known to bind single DNA ligands. We demonstrate that we can "knock in" this ability for TLR9 amplification in membrane-active AMP mutants, which suggests the existence of tradeoffs between membrane permeating activity and immunomodulatory activity in AMP sequences.

Snails In Silico: A Review of Computational Studies on the Conopeptides

Marine cone snails are carnivorous gastropods that use peptide toxins called conopeptides both as a defense mechanism and as a means to immobilize and kill their prey. These peptide toxins exhibit a large chemical diversity that enables exquisite specificity and potency for target receptor proteins. This diversity arises in terms of variations both in amino acid sequence and length, and in posttranslational modifications, particularly the formation of multiple disulfide linkages. Most of the functionally characterized conopeptides target ion channels of animal nervous systems, which has led to research on their therapeutic applications. Many facets of the underlying molecular mechanisms responsible for the specificity and virulence of conopeptides, however, remain poorly understood. In this review, we will explore the chemical diversity of conopeptides from a computational perspective. First, we discuss current approaches used for classifying conopeptides. Next, we review different computational strategies that have been applied to understanding and predicting their structure and function, from machine learning techniques for predictive classification to docking studies and molecular dynamics simulations for molecular-level understanding. We then review recent novel computational approaches for rapid high-throughput screening and chemical design of conopeptides for particular applications. We close with an assessment of the state of the field, emphasizing important questions for future lines of inquiry.

Antimicrobial and Cell-Penetrating Peptides: How to Understand Two Distinct Functions Despite Similar Physicochemical Properties

Antimicrobial and cell-penetrating peptides are both classes of membrane-active peptides sharing similar physicochemical properties. Both kinds of peptides have attracted much attention owing to their specific features. AMPs disrupt cell membranes of bacteria and display urgently needed antibiotic substances with alternative modes of action. Since the multidrug resistance of bacterial pathogens is a more and more raising concern, AMPs have gained much interest during the past years. On the other side, CPPs enter eukaryotic cells without substantially affecting the plasma membrane. They can be used as drug delivery platforms and have proven their usefulness in various applications. However, although both groups of peptides are quite similar, their intrinsic activity is often different, and responsible factors are still in discussion. The aim of this chapter is to summarize and shed light on recent findings and concepts dealing with differences and similarities of AMPs and CPPs and to understand these different functions.

Discovering de novo peptide substrates for enzymes using machine learning

The discovery of peptide substrates for enzymes with exclusive, selective activities is a central goal in chemical biology. In this paper, we develop a hybrid computational and biochemical method to rapidly optimize peptides for specific, orthogonal biochemical functions. The method is an iterative machine learning process by which experimental data is deposited into a mathematical algorithm that selects potential peptide substrates to be tested experimentally. Once tested, the algorithm uses the experimental data to refine future selections. This process is repeated until a suitable set of de novo peptide substrates are discovered. We employed this technology to discover orthogonal peptide substrates for 4’-phosphopantetheinyl transferase, an enzyme class that covalently modifies proteins. In this manner, we have demonstrated that machine learning can be leveraged to guide peptide optimization for specific biochemical functions not immediately accessible by biological screening techniques, such as phage display and random mutagenesis.

The discovery of peptide substrates for enzymes with selective activities is a central goal in chemical biology. Here, the authors develop a hybrid method combining machine learning and experimental testing for fast optimization of peptides for specific, orthogononal functions.

Machine learning antimicrobial peptide sequences: Some surprising variations on the theme of amphiphilic assembly

Antimicrobial peptides (AMPs) collectively constitute a key component of the host innate immune system. They span a diverse space of sequences and can be α-helical, β-sheet, or unfolded in structure. Despite a wealth of knowledge about them from decades of experiments, it remains difficult to articulate general principles governing such peptides. How are they different from other molecules that are also cationic and amphiphilic? What other functions, in immunity and otherwise, are enabled by these simple sequences? In this short review, we present some recent work that engages these questions using methods not usually applied to AMP studies, such as machine learning. We find that not only do AMP-like sequences confer membrane remodeling activity to an unexpectedly broad range of protein classes, their cationic and amphiphilic signature also allows them to act as meta-antigens and self-assemble with immune ligands into nanocrystalline complexes for multivalent presentation to Toll-like receptors.

Unraveling the bioactivity of anticancer peptides as deduced from machine learning

Cancer imposes a global health burden as it represents one of the leading causes of morbidity and mortality while also giving rise to significant economic burden owing to the associated expenditures for its monitoring and treatment. In spite of advancements in cancer therapy, the low success rate and recurrence of tumor has necessitated the ongoing search for new therapeutic agents. Aside from drugs based on small molecules and protein-based biopharmaceuticals, there has been an intense effort geared towards the development of peptide-based therapeutics owing to its favorable and intrinsic properties of being relatively small, highly selective, potent, safe and low in production costs. In spite of these advantages, there are several inherent weaknesses that are in need of attention in the design and development of therapeutic peptides. An abundance of data on bioactive and therapeutic peptides have been accumulated over the years and the burgeoning area of artificial intelligence has set the stage for the lucrative utilization of machine learning to make sense of these large and high-dimensional data. This review summarizes the current state-of-the-art on the application of machine learning for studying the bioactivity of anticancer peptides along with future outlook of the field. Data and R codes used in the analysis herein are available on GitHub at https://github.com/Shoombuatong2527/anticancer-peptides-review.

What Can Pleiotropic Proteins in Innate Immunity Teach Us about Bioconjugation and Molecular Design?

A common bioengineering strategy to add function to a given molecule is by conjugation of a new moiety onto that molecule. Adding multiple functions in this way becomes increasingly challenging and leads to composite molecules with larger molecular weights. In this review, we attempt to gain a new perspective by looking at this problem in reverse, by examining nature's strategies of multiplexing different functions into the same pleiotropic molecule using emerging analysis techniques such as machine learning. We concentrate on examples from the innate immune system, which employs a finite repertoire of molecules for a broad range of tasks. An improved understanding of how diverse functions are multiplexed into a single molecule can inspire new approaches for the deterministic design of multifunctional molecules.