Aggrescan3D (A3D) 2.0: prediction and engineering of protein solubility

Accelerating therapeutic protein design with computational approaches toward the clinical stage

Therapeutic protein, represented by antibodies, is of increasing interest in human medicine. However, clinical translation of therapeutic protein is still largely hindered by different aspects of developability, including affinity and selectivity, stability and aggregation prevention, solubility and viscosity reduction, and deimmunization. Conventional optimization of the developability with widely used methods, like display technologies and library screening approaches, is a time and cost-intensive endeavor, and the efficiency in finding suitable solutions is still not enough to meet clinical needs. In recent years, the accelerated advancement of computational methodologies has ushered in a transformative era in the field of therapeutic protein design. Owing to their remarkable capabilities in feature extraction and modeling, the integration of cutting-edge computational strategies with conventional techniques presents a promising avenue to accelerate the progression of therapeutic protein design and optimization toward clinical implementation. Here, we compared the differences between therapeutic protein and small molecules in developability and provided an overview of the computational approaches applicable to the design or optimization of therapeutic protein in several developability issues.

Liquid-Liquid Phase Separation Primes Spider Silk Proteins for Fiber Formation via a Conditional Sticker Domain

Many protein condensates can convert to fibrillar aggregates, but the underlying mechanisms are unclear. Liquid-liquid phase separation (LLPS) of spider silk proteins, spidroins, suggests a regulatory switch between both states. Here, we combine microscopy and native mass spectrometry to investigate the influence of protein sequence, ions, and regulatory domains on spidroin LLPS. We find that salting out-effects drive LLPS via low-affinity stickers in the repeat domains. Interestingly, conditions that enable LLPS simultaneously cause dissociation of the dimeric C-terminal domain (CTD), priming it for aggregation. Since the CTD enhances LLPS of spidroins but is also required for their conversion into amyloid-like fibers, we expand the stickers and spacers-model of phase separation with the concept of folded domains as conditional stickers that represent regulatory units.

Sequence-based prediction of pH-dependent protein solubility using CamSol

Solubility is a property of central importance for the use of proteins in research in molecular and cell biology and in applications in biotechnology and medicine. Since experimental methods for measuring protein solubility are material intensive and time consuming, computational methods have recently emerged to enable the rapid and inexpensive screening of solubility for large libraries of proteins, as it is routinely required in development pipelines. Here, we describe the development of one such method to include in the predictions the effect of the pH on solubility. We illustrate the resulting pH-dependent predictions on a variety of antibodies and other proteins to demonstrate that these predictions achieve an accuracy comparable with that of experimental methods. We make this method publicly available at https://www-cohsoftware.ch.cam.ac.uk/index.php/camsolph, as the version 3.0 of CamSol.

Structural Analysis of Genomic and Proteomic Signatures Reveal Dynamic Expression of Intrinsically Disordered Regions in Breast Cancer and Tissue

Structural features of proteins capture underlying information about protein evolution and function, which enhances the analysis of proteomic and transcriptomic data. Here we develop Structural Analysis of Gene and protein Expression Signatures (SAGES), a method that describes expression data using features calculated from sequence-based prediction methods and 3D structural models. We used SAGES, along with machine learning, to characterize tissues from healthy individuals and those with breast cancer. We analyzed gene expression data from 23 breast cancer patients and genetic mutation data from the COSMIC database as well as 17 breast tumor protein expression profiles. We identified prominent expression of intrinsically disordered regions in breast cancer proteins as well as relationships between drug perturbation signatures and breast cancer disease signatures. Our results suggest that SAGES is generally applicable to describe diverse biological phenomena including disease states and drug effects.

Autooxidation of curcumin in physiological buffer causes an enhanced synergistic anti-amyloid effect

Curcumin is an important food additive that shows multiple medical-benefits including anticarcinogenic, anti-inflammatory, antibiotic and antiamyloid properties. However, understanding the mechanism of curcumin-mediated effects becomes rather complicated since it has low bio-viability and it undergoes autooxidation, influenced by temperature, pH and buffer. We find that curcumin's antiamyloid-potential is not primarily due to curcumin alone, rather due to a synergistic action of curcumin and its autooxidized-products generated during inhibition reactions. In physiological buffer curcumin undergoes thermally induced autooxidation and yields stable compounds which can synergistically work for both inhibition of amyloid aggregation and promotion of amyloid-disaggregation into soluble protein species. Curcumin also showed substantial inhibition effect against coaggregation of different food proteins. Curcumin's strong affinity for the hydrophobic moieties of the aggregation-prone partially-folded insulin structures seems crucial for the inhibition mechanism. Further, autooxidized curcumin products were found to protect UV-induced protein damage. The results provide conceptual foundations highlighting the link between chemistry and antiamyloid-activity of curcumin and may inspire curcumin-based therapeutics against amyloidogenesis.

A guide to studying protein aggregation

Disrupted protein folding or decreased protein stability can lead to the accumulation of (partially) un- or misfolded proteins, which ultimately cause the formation of protein aggregates. Much of the interest in protein aggregation is associated with its involvement in a wide range of human diseases and the challenges it poses for large-scale biopharmaceutical manufacturing and formulation of therapeutic proteins and peptides. On the other hand, protein aggregates can also be functional, as observed in nature, which triggered its use in the development of biomaterials or therapeutics as well as for the improvement of food characteristics. Thus, unmasking the various steps involved in protein aggregation is critical to obtain a better understanding of the underlying mechanism of amyloid formation. This knowledge will allow a more tailored development of diagnostic methods and treatments for amyloid-associated diseases, as well as applications in the fields of new (bio)materials, food technology and therapeutics. However, the complex and dynamic nature of the aggregation process makes the study of protein aggregation challenging. To provide guidance on how to analyse protein aggregation, in this review we summarize the most commonly investigated aspects of protein aggregation with some popular corresponding methods.

Rational design of a Nivolumab-based ANTI-PD-1 single chain variable fragment that blocks the interaction between PD-1 expressed on T-CELLS and PD-L1 ON CHO cells

Antibodies that block the interaction between PD-1 expressing T-cells and cancer cells expressing PD-L1 play a central role in contemporary immunotherapy regimes [1-3]. We previously reported the isolation of a single chain variable fragment (scFv) of the monoclonal anti-PD-1 antibody Nivolumab, that binds to purified PD-1 and blocked its interaction with PD-L1 [4]. This anti-PD-1 scFv did not, however, function in a cell-based assay designed to detect the disruption of the PD-1/PD-L1 interaction, a result likely due to its poor solubility in tissue culture media. Herein we report that following a series of structure-based rational design analyses, including Aggreescan3D, we have isolated a variant of the anti-PD-1 scFv having significantly improved solubility in tissue culture medium. Moreover, this soluble anti-PD-1 scFv variant disrupted the interaction between PD-1 expressed on Jurkat Cells and PD-L1 expressed on CHO cells. These findings are discussed in terms of the related observation that the residues mutated to form the anti-PD-1 variant are conserved in many other scFvs; thus, the properties of a range of scFvs will likely be enhanced by similar mutations of the conserved residues.

Design, Production, and Characterization of Catalytically Active Inclusion Bodies

Catalytically active inclusion bodies (CatIBs) are promising biologically produced enzyme/protein immobilizates for application in biocatalysis, synthetic chemistry, and biomedicine. CatIB formation is commonly induced by fusion of suitable aggregation-inducing tags to a given target protein. Heterologous production of the fusion protein in turn yields CatIBs. This chapter presents the methodology needed to design, produce, and characterize CatIBs.

High concentration formulation developability approaches and considerations

The growing need for biologics to be administered subcutaneously and ocularly, coupled with certain indications requiring high doses, has resulted in an increase in drug substance (DS) and drug product (DP) protein concentrations. With this increase, more emphasis must be placed on identifying critical physico-chemical liabilities during drug development, including protein aggregation, precipitation, opalescence, particle formation, and high viscosity. Depending on the molecule, liabilities, and administration route, different formulation strategies can be used to overcome these challenges. However, due to the high material requirements, identifying optimal conditions can be slow, costly, and often prevent therapeutics from moving rapidly into the clinic/market. In order to accelerate and derisk development, new experimental and in-silico methods have emerged that can predict high concentration liabilities. Here, we review the challenges in developing high concentration formulations, the advances that have been made in establishing low mass and high-throughput predictive analytics, and advances in in-silico tools and algorithms aimed at identifying risks and understanding high concentration protein behavior.

Bioinformatics tools and data resources for assay development of fluid protein biomarkers

Fluid protein biomarkers are important tools in clinical research and health care to support diagnosis and to monitor patients. Especially within the field of dementia, novel biomarkers could address the current challenges of providing an early diagnosis and of selecting trial participants. While the great potential of fluid biomarkers is recognized, their implementation in routine clinical use has been slow. One major obstacle is the often unsuccessful translation of biomarker candidates from explorative high-throughput techniques to sensitive antibody-based immunoassays. In this review, we propose the incorporation of bioinformatics into the workflow of novel immunoassay development to overcome this bottleneck and thus facilitate the development of novel biomarkers towards clinical laboratory practice. Due to the rapid progress within the field of bioinformatics many freely available and easy-to-use tools and data resources exist which can aid the researcher at various stages. Current prediction methods and databases can support the selection of suitable biomarker candidates, as well as the choice of appropriate commercial affinity reagents. Additionally, we examine methods that can determine or predict the epitope - an antibody’s binding region on its antigen - and can help to make an informed choice on the immunogenic peptide used for novel antibody production. Selected use cases for biomarker candidates help illustrate the application and interpretation of the introduced tools.

Identification of potential candidate vaccines against Mycobacterium ulcerans based on the major facilitator superfamily transporter protein

Buruli ulcer is a neglected tropical disease that is characterized by non-fatal lesion development. The causative agent is Mycobacterium ulcerans (M. ulcerans). There are no known vectors or transmission methods, preventing the development of control methods. There are effective diagnostic techniques and treatment routines; however, several socioeconomic factors may limit patients' abilities to receive these treatments. The Bacillus Calmette-Guérin vaccine developed against tuberculosis has shown limited efficacy, and no conventionally designed vaccines have passed clinical trials. This study aimed to generate a multi-epitope vaccine against M. ulcerans from the major facilitator superfamily transporter protein using an immunoinformatics approach. Twelve M. ulcerans genome assemblies were analyzed, resulting in the identification of 11 CD8+ and 7 CD4+ T-cell epitopes and 2 B-cell epitopes. These conserved epitopes were computationally predicted to be antigenic, immunogenic, non-allergenic, and non-toxic. The CD4+ T-cell epitopes were capable of inducing interferon-gamma and interleukin-4. They successfully bound to their respective human leukocyte antigens alleles in in silico docking studies. The expected global population coverage of the T-cell epitopes and their restricted human leukocyte antigens alleles was 99.90%. The population coverage of endemic regions ranged from 99.99% (Papua New Guinea) to 21.81% (Liberia). Two vaccine constructs were generated using the Toll-like receptors 2 and 4 agonists, LprG and RpfE, respectively. Both constructs were antigenic, non-allergenic, non-toxic, thermostable, basic, and hydrophilic. The DNA sequences of the vaccine constructs underwent optimization and were successfully in-silico cloned with the pET-28a(+) plasmid. The vaccine constructs were successfully docked to their respective toll-like receptors. Molecular dynamics simulations were carried out to analyze the binding interactions within the complex. The generated binding energies indicate the stability of both complexes. The constructs generated in this study display severable favorable properties, with construct one displaying a greater range of favorable properties. However, further analysis and laboratory validation are required.

Salt-Induced Dissolution of Protein Aggregates

Protein aggregation is mediated by a complex interplay of noncovalent interactions and is associated with a broad range of aspects from debilitating human diseases to the food industry and therapeutic biotechnology. Deciphering the intricate roles of noncovalent interactions is of paramount importance for the design of effective inhibitory and disaggregation strategies, which remains a formidable challenge. By using a combination of spectroscopic and microscopic tools, here we show that the surfactant-mediated protein aggregation can be modulated by an intriguing interplay of hydrophobic and electrostatic effects. Additionally, our results illuminate the unique role of salt as a potent disaggregation inducer that alters the protein-surfactant electrostatic interactions and triggers the dissolution of preformed protein aggregates resulting in restoring the native protein structure. This unusual salt-induced dissolution and refolding offers a unique approach to regulating the balance between protein self-assembly and disassembly and will offer a potent strategy to design electrostatically targeted inhibitors.

SARS-CoV-2 spike protein aggregation is triggered by bacterial lipopolysaccharide

SARS-CoV-2 spike (S) protein is crucial for virus invasion in COVID-19. Here, we showed that lipopolysaccharide (LPS) can trigger S protein aggregation at high doses of LPS and S protein. We demonstrated the formation of S protein aggregates by microscopy analyses, aggregation and gel shift assays. LPS at high levels boosts the formation of S protein aggregates as detected by amytracker and thioflavin T dyes that specifically bind to aggregating proteins. We validated the role of LPS by blocking the formation of aggregates by the endotoxin-scavenging thrombin-derived peptide TCP-25. Aggregation-prone sequences in S protein are predicted to be nearby LPS binding sites, while molecular simulations showed stable formation of S protein-LPS higher-order oligomers. Collectively, our results provide evidence of LPS-induced S protein aggregation.

Design, construction and in vivo functional assessment of a hinge truncated sFLT01

Gene therapy for the treatment of ocular neovascularization has reached clinical trial phases. The AAV2-sFLT01 construct was already evaluated in a phase 1 open-label trial administered intravitreally to patients with advanced neovascular age-related macular degeneration. SFLT01 protein functions by binding to VEGF and PlGF molecules and inhibiting their activities simultaneously. It consists of human VEGFR1/Flt-1 (hVEGFR1), a polyglycine linker, and the Fc region of human IgG1. The IgG1 upper hinge region of the sFLT01 molecule makes it vulnerable to radical attacks and prone to causing immune reactions. This study pursued two goals: (i) minimizing the immunogenicity and vulnerability of the molecule by designing a truncated molecule called htsFLT01 (hinge truncated sFLT01) that lacked the IgG1 upper hinge and lacked 2 amino acids from the core hinge region; and (ii) investigating the structural and functional properties of the aforesaid chimeric molecule at different levels (in silico, in vitro, and in vivo). Molecular dynamics simulations and molecular mechanics energies combined with Poisson-Boltzmann and surface area continuum solvation calculations revealed comparable free energy of binding and binding affinity for sFLT01 and htsFLT01 to their cognate ligands. Conditioned media from human retinal pigment epithelial (hRPE) cells that expressed htsFLT01 significantly reduced tube formation in HUVECs. The AAV2-htsFLT01 virus suppressed vascular development in the eyes of newborn mice. The htsFLT01 gene construct is a novel anti-angiogenic tool with promising improvements compared to existing treatments.

Enhancement of solubility of recombinant alcohol dehydrogenase from Rhodococcus ruber using predictive tool

Solubility is one of key factors influencing the heterologous production of recombinant proteins in biotechnology. Among many aggregation-prone proteins, alcohol dehydrogenase (ADH-A) from Rhodococcus ruber (in this work abbreviated RrADH) shows a great potential in processes involved in the biotransformation of natural compounds. As ADH-A is a potentially high value asset in industrial biotransformation processes, improvement of its solubility would be of major commercial benefit. Predictive tools and in silico analysis provide a fast means for improving protein properties, for selecting appropriate changes, and ultimately for saving costs. We have therefore focused on enhancement of the solubility of RrADH using an online accesible predictive tool Aggrescan 3D 2.0. Selected mutations were introduced into the protein amino acid sequence by using site-directed PCR. This led to a 17% increase in the protein solubility of RrADHmut1 and a 98% increase for RrADHmut2. Moreover, the basic kinetics of the enzyme reaction were positively affected, further optimizing the overall performance of the production process.

Tetramerization of the S100B Chaperone Spawns a Ca2+ Independent Regulatory Surface that Enhances Anti-aggregation Activity and Client Specificity

Alzheimer's disease (AD) hallmarks include the aggregation of amyloid-β (Aβ), tau and neuroinflammation promoted by several alarmins. Among these is S100B, a small astrocytic homodimeric protein, upregulated in AD, whose multiple biological activities depend on localization, concentration, and assembly state. S100B was reported to inhibit the aggregation and toxicity of Aβ42 and tau similarly to a holdase-type chaperone. This activity is dependent of Ca²⁺-binding, which triggers the exposure of a regulatory binding cleft at the S100B dimer interface with which amyloidogenic clients dynamically interact. Although the dimer prevails, a significant portion of secreted S100B in the human brain occurs as higher order multimers, whose protective functions remain uncharacterized and which we here investigate. Resorting to ThT-monitored aggregation kinetics, we determined that unlike the dimer, tetrameric S100B inhibits Aβ42 aggregation at sub/equimolar ratios, an effect that persists in the absence of Ca²⁺ binding. Structural analysis revealed that S100B tetramerization spawns a novel extended cleft accommodating an aggregation-prone surface that mediates interactions with monomeric Aβ client via hydrophobic interactions, as corroborated by Bis-ANS fluorescence and docking analysis. Correspondingly, at high ionic strength that reduces solvation and favours hydrophobic contacts, the inhibition of Aβ42 aggregation by tetrameric S100B is 3-fold increased. Interestingly, this extended Ca²⁺-independent surface favours Aβ42 as substrate, as tau K18 aggregation is not inhibited by the apo tetramer. Overall, results illustrate a mechanism through which oligomerization of the S100B chaperone fine-tunes anti-aggregation activity and client specificity, highlighting the potential functional relevance of S100B multimers in the regulation of AD proteotoxicity.

Structure-Activity Relationship of the Dimeric and Oligomeric Forms of a Cytotoxic Biotherapeutic Based on Diphtheria Toxin

Protein aggregation is a well-recognized problem in industrial preparation, including biotherapeutics. These low-energy states constantly compete with a native-like conformation, which is more pronounced in the case of macromolecules of low stability in the solution. A better understanding of the structure and function of such aggregates is generally required for the more rational development of therapeutic proteins, including single-chain fusion cytotoxins to target specific receptors on cancer cells. Here, we identified and purified such particles as side products of the renaturation process of the single-chain fusion cytotoxin, composed of two diphtheria toxin (DT) domains and interleukin 13 (IL-13), and applied various experimental techniques to comprehensively understand their molecular architecture and function. Importantly, we distinguished soluble purified dimeric and fractionated oligomeric particles from aggregates. The oligomers are polydisperse and multimodal, with a distribution favoring lower and even stoichiometries, suggesting they are composed of dimeric building units. Importantly, all these oligomeric particles and the monomer are cystine-dependent as their innate disulfide bonds have structural and functional roles. Their reduction triggers aggregation. Presumably the dimer and lower oligomers represent the metastable state, retaining the native disulfide bond. Although significantly reduced in contrast to the monomer, they preserve some fraction of bioactivity, manifested by their IL-13RA2 receptor affinity and selective cytotoxic potency towards the U-251 glioblastoma cell line. These molecular assemblies probably preserve structural integrity and native-like fold, at least to some extent. As our study demonstrated, the dimeric and oligomeric cytotoxin may be an exciting model protein, introducing a new understanding of its monomeric counterpart’s molecular characteristics.

Association of NOTCH3 Variant Position With Stroke Onset and Other Clinical Features Among Patients With CADASIL

BACKGROUND AND OBJECTIVES

Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is caused by a cysteine-altering variant in 1 of the 34 epidermal growth factor-like repeat (EGFR) domains of the NOTCH3 protein. CADASIL has a variable phenotypic presentation, and NOTCH3 variants in EGFRs 1-6 have been found correlated with greater disease severity. We examined clinical and radiologic features and performed bioinformatic annotation of variants in a large CADASIL cohort to further understand these associations.

METHODS

We examined the association of NOTCH3 variant position on stroke onset and other clinical features among patients with CADASIL from the United Kingdom. We also explored how in silico predicted protein aggregation differed by variant position and the extent to which this affected stroke risk.

RESULTS

We identified 76 different cysteine-altering NOTCH3 variants in our cohort of 485 patients (mean age: 50.1 years; % male: 57.5). After controlling for cardiovascular risk factors, variants in EGFRs 1-6 were associated with earlier onset of stroke (hazard ratio [HR]: 2.05, 95% CI: 1.43-2.94) and encephalopathy (HR: 2.70, 95% CI: 1.15-6.37), than variants in EGFRs 7-34. Although the risk of stroke was higher in the patients with predicted protein aggregation (HR: 1.50, 95% CI: 1.05-2.14), this association was no longer significant after controlling for variant site. Further analysis suggested that lower stroke risk was observed for variants in EGFRs 10-17 compared with variants in the other EGFR domains.

DISCUSSION

NOTCH3 variant position is a predictor of stroke and encephalopathy in CADASIL independent of cardiovascular risk factors. Lower stroke risk was found for variants in EGFRs 10-17. Molecular factors that influence CADASIL disease severity remain to be determined.

The effect of mutation on an aggregation-prone protein: An in vivo, in vitro, and in silico analysis

Aggregation of initially stably structured proteins is involved in more than 20 human amyloid diseases. Despite intense research, however, how this class of proteins assembles into amyloid fibrils remains poorly understood, principally because of the complex effects of amino acid substitutions on protein stability, solubility, and aggregation propensity. We address this question using β2-microglobulin (β2m) as a model system, focusing on D76N-β2m that is involved in hereditary amyloidosis. This amino acid substitution causes the aggregation-resilient wild-type protein to become highly aggregation prone in vitro, although the mechanism by which this occurs remained elusive. Here, we identify the residues key to protecting β2m from aggregation by coupling aggregation with antibiotic resistance in E. coli using a tripartite β-lactamase assay (TPBLA). By performing saturation mutagenesis at three different sites (D53X-, D76X-, and D98X-β2m) we show that residue 76 has a unique ability to drive β2m aggregation in vivo and in vitro. Using a randomly mutated D76N-β2m variant library, we show that all of the mutations found to improve protein behavior involve residues in a single aggregation-prone region (APR) (residues 60 to 66). Surprisingly, no correlation was found between protein stability and protein aggregation rate or yield, with several mutations in the APR decreasing aggregation without affecting stability. Together, the results demonstrate the power of the TPBLA to develop proteins that are resilient to aggregation and suggest a model for D76N-β2m aggregation involving the formation of long-range couplings between the APR and Asn76 in a nonnative state.

An In Silico Methodology That Facilitates Decision Making in the Engineering of Nanoscale Protein Materials

Under the need for new functional and biocompatible materials for biomedical applications, protein engineering allows the design of assemblable polypeptides, which, as convenient building blocks of supramolecular complexes, can be produced in recombinant cells by simple and scalable methodologies. However, the stability of such materials is often overlooked or disregarded, becoming a potential bottleneck in the development and viability of novel products. In this context, we propose a design strategy based on in silico tools to detect instability areas in protein materials and to facilitate the decision making in the rational mutagenesis aimed to increase their stability and solubility. As a case study, we demonstrate the potential of this methodology to improve the stability of a humanized scaffold protein (a domain of the human nidogen), with the ability to oligomerize into regular nanoparticles usable to deliver payload drugs to tumor cells. Several nidogen mutants suggested by the method showed important and measurable improvements in their structural stability while retaining the functionalities and production yields of the original protein. Then, we propose the procedure developed here as a cost-effective routine tool in the design and optimization of multimeric protein materials prior to any experimental testing.

A3D database: structure-based predictions of protein aggregation for the human proteome

SUMMARY

Protein aggregation is associated with many human disorders and constitutes a major bottleneck for producing therapeutic proteins. Our knowledge of the human protein structures repertoire has dramatically increased with the recent development of the AlphaFold (AF) deep-learning method. This structural information can be used to understand better protein aggregation properties and the rational design of protein solubility. This article uses the Aggrescan3D (A3D) tool to compute the structure-based aggregation predictions for the human proteome and make the predictions available in a database form. In the A3D database, we analyze the AF-predicted human protein structures (for over 20.5 thousand unique Uniprot IDs) in terms of their aggregation properties using the A3D tool. Each entry of the A3D database provides a detailed analysis of the structure-based aggregation propensity computed with A3D. The A3D database implements simple but useful graphical tools for visualizing and interpreting protein structure datasets. It also enables testing the influence of user-selected mutations on protein solubility and stability, all integrated into a user-friendly interface.

AVAILABILITY AND IMPLEMENTATION

A3D database is freely available at: http://biocomp.chem.uw.edu.pl/A3D2/hproteome. The data underlying this article are available in the article and in its online supplementary material.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Tools for computational design and high-throughput screening of therapeutic enzymes

Therapeutic enzymes are valuable biopharmaceuticals in various biomedical applications. They have been successfully applied for fibrinolysis, cancer treatment, enzyme replacement therapies, and the treatment of rare diseases. Still, there is a permanent demand to find new or better therapeutic enzymes, which would be sufficiently soluble, stable, and active to meet specific medical needs. Here, we highlight the benefits of coupling computational approaches with high-throughput experimental technologies, which significantly accelerate the identification and engineering of catalytic therapeutic agents. New enzymes can be identified in genomic and metagenomic databases, which grow thanks to next-generation sequencing technologies exponentially. Computational design and machine learning methods are being developed to improve catalytically potent enzymes and predict their properties to guide the selection of target enzymes. High-throughput experimental pipelines, increasingly relying on microfluidics, ensure functional screening and biochemical characterization of target enzymes to reach efficient therapeutic enzymes.

Computational models for studying physical instabilities in high concentration biotherapeutic formulations

Computational prediction of the behavior of concentrated protein solutions is particularly advantageous in early development stages of biotherapeutics when material availability is limited and a large set of formulation conditions needs to be explored. This review provides an overview of the different computational paradigms that have been successfully used in modeling undesirable physical behaviors of protein solutions with a particular emphasis on high-concentration drug formulations. This includes models ranging from all-atom simulations, coarse-grained representations to macro-scale mathematical descriptions used to study physical instability phenomena of protein solutions such as aggregation, elevated viscosity, and phase separation. These models are compared and summarized in the context of the physical processes and their underlying assumptions and limitations. A detailed analysis is also given for identifying protein interaction processes that are explicitly or implicitly considered in the different modeling approaches and particularly their relations to various formulation parameters. Lastly, many of the shortcomings of existing computational models are discussed, providing perspectives and possible directions toward an efficient computational framework for designing effective protein formulations.

Computational methods to predict protein aggregation

In most cases, protein aggregation stems from the establishment of non-native intermolecular contacts. The formation of insoluble protein aggregates is associated with many human diseases and is a major bottleneck for the industrial production of protein-based therapeutics. Strikingly, fibrillar aggregates are naturally exploited for structural scaffolding or to generate molecular switches and can be artificially engineered to build up multi-functional nanomaterials. Thus, there is a high interest in rationalizing and forecasting protein aggregation. Here, we review the available computational toolbox to predict protein aggregation propensities, identify sequential or structural aggregation-prone regions, evaluate the impact of mutations on aggregation or recognize prion-like domains. We discuss the strengths and limitations of these algorithms and how they can evolve in the next future.

Bioinformatics Methods in Predicting Amyloid Propensity of Peptides and Proteins

Several computational methods have been developed to predict amyloid propensity of a protein or peptide. These bioinformatics tools are time- and cost-saving alternatives to expensive and laborious experimental methods which are used to confirm self-aggregation of a protein. Computational approaches not only allow preselection of reliable candidates for amyloids but, most importantly, are capable of a thorough and informative analysis of a protein, indicating the sequence determinants of protein aggregation, identifying the potential causal mutations and likely mechanisms. Bioinformatics modeling applies several different approaches, which most typically include physicochemical or structure-based modeling, machine learning, or statistics based modeling. Bioinformatics methods typically use the amino acid sequence of a protein as an input, some also include additional information, for example, an available structure. This chapter describes the methods currently used to computationally predict amyloid propensity of a protein or peptide. Since the accuracy of bioinformatics methods may be highly dependent on reference data used to develop and evaluate the predictors, we also briefly present the main databases of amyloids used by the authors of bioinformatics tools.

Machine learning prediction of antibody aggregation and viscosity for high concentration formulation development of protein therapeutics

Machine learning has been recently used to predict therapeutic antibody aggregation rates and viscosity at high concentrations (150 mg/ml). These works focused on commercially available antibodies, which may have been optimized for stability. In this study, we measured accelerated aggregation rates at 45°C and viscosity at 150 mg/ml for 20 preclinical and clinical-stage antibodies. Features obtained from molecular dynamics simulations of the full-length antibody and sequences were used for machine learning model construction. We found a k-nearest neighbors regression model with two features, spatial positive charge map on the CDRH2 and solvent-accessible surface area of hydrophobic residues on the variable fragment, gives the best performance for predicting antibody aggregation rates (r = 0.89). For the viscosity classification model, the model with the highest accuracy is a logistic regression model with two features, spatial negative charge map on the heavy chain variable region and spatial negative charge map on the light chain variable region. The accuracy and the area under precision recall curve of the classification model from validation tests are 0.86 and 0.70, respectively. In addition, we combined data from another 27 commercial mAbs to develop a viscosity predictive model. The best model is a logistic regression model with two features, number of hydrophobic residues on the light chain variable region and net charges on the light chain variable region. The accuracy and the area under precision recall curve of the classification model are 0.85 and 0.6, respectively. The aggregation rates and viscosity models can be used to predict antibody stability to facilitate pharmaceutical development.

Assessment of Therapeutic Antibody Developability by Combinations of In Vitro and In Silico Methods

Although antibodies have become the fastest-growing class of therapeutics on the market, it is still challenging to develop them for therapeutic applications, which often require these molecules to withstand stresses that are not present in vivo. We define developability as the likelihood of an antibody candidate with suitable functionality to be developed into a manufacturable, stable, safe, and effective drug that can be formulated to high concentrations while retaining a long shelf life. The implementation of reliable developability assessments from the early stages of antibody discovery enables flagging and deselection of potentially problematic candidates, while focussing available resources on the development of the most promising ones. Currently, however, thorough developability assessment requires multiple in vitro assays, which makes it labor intensive and time consuming to implement at early stages. Furthermore, accurate in vitro analysis at the early stage is compromised by the high number of potential candidates that are often prepared at low quantities and purity. Recent improvements in the performance of computational predictors of developability potential are beginning to change this scenario. Many computational methods only require the knowledge of the amino acid sequences and can be used to identify possible developability issues or to rank available candidates according to a range of biophysical properties. Here, we describe how the implementation of in silico tools into antibody discovery pipelines is increasingly offering time- and cost-effective alternatives to in vitro experimental screening, thus streamlining the drug development process. We discuss in particular the biophysical and biochemical properties that underpin developability potential and their trade-offs, review various in vitro assays to measure such properties or parameters that are predictive of developability, and give an overview of the growing number of in silico tools available to predict properties important for antibody development, including the CamSol method developed in our laboratory.

SOD1 mutations associated with amyotrophic lateral sclerosis analysis of variant severity

Mutations in superoxide dismutase 1 gene (SOD1) are linked to amyotrophic lateral sclerosis (ALS), a neurodegenerative disorder predominantly affecting upper and lower motor neurons. The clinical phenotype of ALS shows inter- and intrafamilial heterogeneity. The aim of the study was to analyze the relations between individual SOD1 mutations and the clinical presentation using in silico methods to assess the SOD1 mutations severity. We identified SOD1 causative variants in a group of 915 prospectively tested consecutive Polish ALS patients from a neuromuscular clinical center, performed molecular modeling of mutated SOD1 proteins and in silico analysis of mutation impact on clinical phenotype and survival analysis of associations between mutations and hazard of clinical end-points. Fifteen SOD1 mutations were identified in 21.1% familial and 2.3% sporadic ALS cases. Their effects on SOD1 protein structure and functioning inferred from molecular modeling and in silico analyses correlate well with the clinical data. Molecular modeling results support the hypothesis that folding intermediates rather than mature SOD1 protein give rise to the source of cytotoxic conformations in ALS. Significant associations between type of mutation and clinical end-points were found.

Progress and challenges for the machine learning-based design of fit-for-purpose monoclonal antibodies

Although the therapeutic efficacy and commercial success of monoclonal antibodies (mAbs) are tremendous, the design and discovery of new candidates remain a time and cost-intensive endeavor. In this regard, progress in the generation of data describing antigen binding and developability, computational methodology, and artificial intelligence may pave the way for a new era of in silico on-demand immunotherapeutics design and discovery. Here, we argue that the main necessary machine learning (ML) components for an in silico mAb sequence generator are: understanding of the rules of mAb-antigen binding, capacity to modularly combine mAb design parameters, and algorithms for unconstrained parameter-driven in silico mAb sequence synthesis. We review the current progress toward the realization of these necessary components and discuss the challenges that must be overcome to allow the on-demand ML-based discovery and design of fit-for-purpose mAb therapeutic candidates.

Protocols for Rational Design of Protein Solubility and Aggregation Properties Using Aggrescan3D Standalone

Protein aggregation is a major hurdle in the development and manufacturing of protein-based therapeutics. Development of aggregation-resistant and stable protein variants can be guided by rational redesign using computational tools. Here, we describe the architecture and functionalities of the Aggrescan3D (A3D) standalone package for the rational design of protein solubility and aggregation properties based on three-dimensional protein structures. We present the case studies of the three therapeutic proteins, including antibodies, exploring the practical use of the A3D standalone tool. The case studies demonstrate that protein solubility can be easily improved by the A3D prediction of non-destabilizing amino acid mutations at the protein surfaces.

A3D 2.0 Update for the Prediction and Optimization of Protein Solubility

Protein aggregation propensity is a property imprinted in protein sequences and structures, being associated with the onset of human diseases and limiting the implementation of protein-based biotherapies. Computational approaches stand as cost-effective alternatives for reducing protein aggregation and increasing protein solubility. AGGRESCAN 3D (A3D) is a structure-based predictor of aggregation that takes into account the conformational context of a protein, aiming to identify aggregation-prone regions exposed in protein surfaces. Here we inspect the updated 2.0 version of the algorithm, which extends the application of A3D to previously inaccessible proteins and incorporates new modules to assist protein redesign. Among these features, the new server includes stability calculations and the possibility to optimize protein solubility using an experimentally validated computational pipeline. Finally, we employ defined examples to navigate the A3D RESTful service, a routine to handle extensive protein collections. Altogether, this chapter is conceived to train and assist A3D non-experts in the study of aggregation-prone regions and protein solubility redesign.

Isotype-specific binding patterns of serum antibodies to multiple conformational epitopes of Bet v 1

BACKGROUND

Birch pollen is an important elicitor of respiratory allergy. The major allergen, Bet v 1, binds IgE exclusively via conformational epitopes.

OBJECTIVE

To identify Bet v 1-specific epitope repertoires of IgE and IgG from birch pollen-allergic and non-allergic subjects.

METHODS

Chimeric proteins were created by grafting individual epitope-sized, contiguous surface patches of Bet v 1 onto a non-allergenic structural homologue and expressed in Escherichia coli. Binding of IgE, IgG1 and IgG4 from sera of 30 birch pollen-allergic and 11 non-allergic subjects to Bet v 1, 13 chimeric proteins and four bacterial Bet v 1 homologues were measured by ELISA. The proportion of epitope-specific in total Bet v 1-specific IgE and the cross-reactivity of Bet v 1-specific IgE with bacterial homologues were determined by competitive ELISA.

RESULTS

Thirteen soluble, correctly folded chimeric proteins were produced. IgE from 27/30 birch pollen-allergic patients bound to 1-12 chimeric proteins (median 4.0) with patient-specific patterns. Three chimeras binding IgE from the majority of sera were identified, whose pgrafted patches overlapped with previously published epitopes. Patterns of IgG1 and IgG4 binding to the chimeric proteins did not correspond to the binding patterns of IgE. Sera of 19/30 birch pollen-allergic patients contained low amounts of IgE to bacterial homologues. Bacterial proteins were able to partially inhibit IgE binding to Bet v 1.

CONCLUSION

Epitopes recognized by Bet v 1-specific antibodies from birch pollen-allergic patients are specific to each patient and differ between IgE, IgG1 and IgG4.

Evidence of Anti-amyloid Characteristics of Plumbagin via Inhibition of Protein Aggregation and Disassembly of Protein Fibrils

The biological consequences associated with the conversion of soluble proteins into insoluble toxic amyloids are not only limited to the onset of neurodegenerative diseases but also to the potential health risks associated with supplements of protein therapeutic agents as well. Hence, finding inhibitors against amyloid formation is important, and natural product-based anti-amyloid compounds have gained much interest because of their higher efficacy and biocompatibility. Plumbagin has been identified as a potential natural product with multiple medical benefits; however, it remains largely unclear whether plumbagin can act against amyloid formation of proteins. Here, we show that plumbagin can effectively inhibit the temperature-induced amyloid aggregation of important proteins (insulin and serum albumin). Both experimental and computational data revealed that the presence of plumbagin in protein solutions, under aggregating conditions, promotes a direct protein-plumbagin interaction, which is predominantly stabilized by stronger H-bonds and hydrophobic interactions. Plumbagin-mediated retention of the native structures of proteins appears to play a crucial role in preventing their conversion into insoluble β-sheet-rich amyloid aggregates. More importantly, the addition of plumbagin into a suspension of protein fibrils triggered their spontaneous disassembly, promoting the release of soluble proteins. The results highlight that a possible synergistic effect via both the stabilization of protein structures and the restriction of the monomer recruitment at the fibril growth sites could be important for the mechanism of plumbagin's anti-aggregation effect. These findings may inspire the development of plumbagin-based formulations to benefit both the prevention and treatment of amyloid-related health complications.

Dimerization underlies the aggregation propensity of intrinsically disordered coiled-coil domain-containing 124

Coiled-coil domain-containing 124 (CCDC124) is a recently discovered ribosome-binding protein conserved in eukaryotes. CCDC124 has regulatory functions on the mediation of reversible ribosomal hibernation and translational recovery by direct attachment to large subunit ribosomal protein uL5, 25S rRNA backbone, and tRNA-binding P/A-site major groove. Moreover, it independently mediates cell division and cellular stress response by facilitating cytokinetic abscission and disulfide stress-dependent transcriptional regulation, respectively. However, the structural characterization and intracellular physiological status of CCDC124 remain unknown. In this study, we employed advanced in silico protein modeling and characterization tools to generate a native-like tertiary structure of CCDC124 and examine the disorder, low sequence complexity, and aggregation propensities, as well as high-order dimeric/oligomeric states. Subsequently, dimerization of CCDC124 was investigated with co-immunoprecipitation (CO-IP) analysis, immunostaining, and a recent live-cell protein-protein interaction method, bimolecular fluorescence complementation (BiFC). Results revealed CCDC124 as a highly disordered protein consisting of low complexity regions at the N-terminus and an aggregation sequence (151-IAVLSV-156) located in the middle region. Molecular docking and post-docking binding free energy analyses highlighted a potential involvement of V153 residue on the generation of high-order dimeric/oligomeric structures. Co-IP, immunostaining, and BiFC analyses were used to further confirm the dimeric state of CCDC124 predominantly localized at the cytoplasm. In conclusion, our findings revealed in silico structural characterization and in vivo subcellular physiological state of CCDC124, suggesting low-complexity regions located at the N-terminus of disordered CCDC124 may regulate the formation of aggregates or high-order dimeric/oligomeric states.

The intrinsic amyloidogenic propensity of cofilin-1 is aggravated by Cys-80 oxidation: A possible link with neurodegenerative diseases

Cofilin-1, an actin dynamizing protein, forms actin-cofilin rods, which is one of the major events that exacerbates the pathophysiology of amyloidogenic diseases. Cysteine oxidation in cofilin-1 under oxidative stress plays a crucial role in the formation of these rods. Others and we have reported that cofilin-1 possesses a self-oligomerization property in vitro and in vivo under physiological conditions. However, it remains elusive if cofilin-1 itself forms amyloid-like structures. We, therefore, hypothesized that cofilin-1 might form amyloid-like assemblies, with a potential to intensify the pathophysiology of amyloid-linked diseases. We used various in silico and in vitro techniques and examined the amyloid-forming propensity of cofilin-1. The study confirms that cofilin-1 possesses an intrinsic tendency of aggregation and forms amyloid-like structures in vitro. Further, we studied the effect of cysteine oxidation on the stability and structural features of cofilin-1. Our data show that oxidation at Cys-80 renders cofilin-1 unstable, leading to a partial loss of protein structure. The results substantiate our hypothesis and establish a strong possibility that cofilin-1 aggregation might play a role in cofilin-mediated pathology and the progression of several amyloid-linked diseases.

In silico analysis of the aggregation propensity of the SARS-CoV-2 proteome: Insight into possible cellular pathologies

The SARS-CoV-2 virus causes the coronavirus disease 19 emerged in 2020. The pandemic triggered a turmoil in public health and is having a tremendous social and economic impact around the globe. Upon entry into host cells, the SARS-CoV-2 virus hijacks cellular machineries to produce and maintain its own proteins, spreading the infection. Although the disease is known for prominent respiratory symptoms, accumulating evidence is also demonstrating the involvement of the central nervous system, with possible mid- and long-term neurological consequences. In this study, we conducted a detailed bioinformatic analysis of the SARS-CoV-2 proteome aggregation propensity by using several complementary computational tools. Our study identified 10 aggregation prone proteins in the reference SARS-CoV-2 strain: the non-structural proteins Nsp4, Nsp6 and Nsp7 as well as ORF3a, ORF6, ORF7a, ORF7b, ORF10, CovE and CovM. By searching for the available mutants of each protein, we have found that most proteins are conserved, while ORF3a and ORF7b are variable and characterized by the occurrence of a large number of mutants with increased aggregation propensity. The geographical distribution of the mutants revealed interesting differences in the localization of aggregation-prone mutants of each protein. Aggregation-prone mutants of ORF7b were found in 7 European countries, whereas those of ORF3a in only 2. Aggregation-prone sequences of ORF7b, but not of ORF3a, were identified in Australia, India, Nepal, China, and Thailand. Our results are important for future analysis of a possible correlation between higher transmissibility and infection, as well as the presence of neurological symptoms with aggregation propensity of SARS-CoV-2 proteins.

Polyphenol-solubility alters amyloid fibril formation of α-synuclein

Amyloid fibril formation is associated with various amyloidoses, including neurodegenerative diseases such as Alzheimer's and Parkinson's diseases. Amyloid fibrils form above the solubility of amyloidogenic proteins or peptides upon breaking supersaturation, followed by a nucleation and elongation mechanism, which is similar to the crystallization of solutes. Many additives, including salts, detergents, and natural compounds, promote or inhibit amyloid formation. However, the underlying mechanisms of the opposing effects are unclear. We examined the effects of two polyphenols, that is, epigallocatechin gallate (EGCG) and kaempferol-7─O─glycoside (KG), with high and low solubilities, respectively, on the amyloid formation of α-synuclein (αSN). EGCG and KG inhibited and promoted amyloid formation of αSN, respectively, when monitored by thioflavin T (ThT) fluorescence or transmission electron microscopy (TEM). Nuclear magnetic resonance (NMR) analysis revealed that, although interactions of αSN with soluble EGCG increased the solubility of αSN, thus inhibiting amyloid formation, interactions of αSN with insoluble KG reduced the solubility of αSN, thereby promoting amyloid formation. Our study suggests that opposing effects of polyphenols on amyloid formation of proteins and peptides can be interpreted based on the solubility of polyphenols.

AlphaFold and the amyloid landscape

Protein aggregation is a widespread phenomenon with important implications in many scientific areas. Although amyloid formation is typically considered as detrimental, functional amyloids that perform physiological roles have been identified in all kingdoms of life. Despite their functional and pathological relevance, the structural details of the majority of molecular species involved in the amyloidogenic process remains elusive. Here, we explore the application of AlphaFold, a highly accurate protein structure predictor, in the field of protein aggregation. While we envision a straightforward application of AlphaFold in assisting the design of globular proteins with improved solubility for biomedical and industrial purposes, the use of this algorithm for predicting the structure of aggregated species seems far from trivial. First, in amyloid diseases, the presence of multiple amyloid polymorphs and the heterogeneity of aggregation intermediates challenges the "one sequence, one structure" paradigm, inherent to sequence-based predictions. Second, aberrant aggregation is not the subject of positive selective pressure, precluding the use of evolutionary-based approaches, which are the core of the AlphaFold pipeline. Instead, amyloid polymorphism seems to be constrained by the need for a defined structure-activity relationship in functional amyloids. They may thus provide a starting point for the application of AlphaFold in the amyloid landscape.

Protocols for Fast Simulations of Protein Structure Flexibility Using CABS-Flex and SURPASS

Conformational flexibility of protein structures can play an important role in protein function. The flexibility is often studied using computational methods since experimental characterization can be difficult. Depending on protein system size, computational tools may require large computational resources or significant simplifications in the modeled systems to speed up calculations. In this work, we present the protocols for efficient simulations of flexibility of folded protein structures that use coarse-grained simulation tools of different resolutions: medium, represented by CABS-flex, and low, represented by SUPRASS. We test the protocols using a set of 140 globular proteins and compare the results with structure fluctuations observed in MD simulations, ENM modeling, and NMR ensembles. As demonstrated, CABS-flex predictions show high correlation to experimental and MD simulation data, while SURPASS is less accurate but promising in terms of future developments.

Insight into thermally induced structural changes of lupin seed γ-conglutin

A multidimensional analysis aimed to determine the thermal impact on γ-conglutin at the two oligomeric states was carried out. A wide range of biophysical and bioinformatic methods allowed to get insight into a thermal unfolding mechanism. The determined midpoint transition temperature (T_m) values were remarkably different, being 56.5 °C and 71.1 °C for γ-conglutin monomer and hexamer, respectively. The unfolding pattern for hexamer molecules included aggregation/precipitation, while monomers tended to form soluble aggregates after heat exposure. Interestingly, differences in the aromatic amino acid residues movements indicate that during thermal treatment of γ-conglutin hexamer red-shift occurred contrary to the monomer in the case of which blue-shift was noted. The obtained results provide an essential contribution to expand our knowledge about the molecular characterization of this intriguing lupin seed protein.

Solubility-Weighted Index: fast and accurate prediction of protein solubility

Motivation

Recombinant protein production is a widely used technique in the biotechnology and biomedical industries, yet only a quarter of target proteins are soluble and can therefore be purified.

Results

We have discovered that global structural flexibility, which can be modeled by normalized B-factors, accurately predicts the solubility of 12 216 recombinant proteins expressed in Escherichia coli. We have optimized these B-factors, and derived a new set of values for solubility scoring that further improves prediction accuracy. We call this new predictor the ‘Solubility-Weighted Index’ (SWI). Importantly, SWI outperforms many existing protein solubility prediction tools. Furthermore, we have developed ‘SoDoPE’ (Soluble Domain for Protein Expression), a web interface that allows users to choose a protein region of interest for predicting and maximizing both protein expression and solubility.

Availability and implementation

The SoDoPE web server and source code are freely available at https://tisigner.com/sodope and https://github.com/Gardner-BinfLab/TISIGNER-ReactJS, respectively. The code and data for reproducing our analysis can be found at https://github.com/Gardner-BinfLab/SoDoPE_paper_2020.

Supplementary information

Supplementary data are available at Bioinformatics online.

Web-based tools for computational enzyme design

Enzymes are in high demand for very diverse biotechnological applications. However, natural biocatalysts often need to be engineered for fine-tuning their properties towards the end applications, such as the activity, selectivity, stability to temperature or co-solvents, and solubility. Computational methods are increasingly used in this task, providing predictions that narrow down the space of possible mutations significantly and can enormously reduce the experimental burden. Many computational tools are available as web-based platforms, making them accessible to non-expert users. These platforms are typically user-friendly, contain walk-throughs, and do not require deep expertise and installations. Here we describe some of the most recent outstanding web-tools for enzyme engineering and formulate future perspectives in this field.

Dialysis-related amyloidosis associated with a novel β2-microglobulin variant

Till date, there had been no reported case of dialysis-related amyloidosis (DRA) associated with a β₂-microglobulin variant. We report here a 41-year-old haemodialysis patient with systemic amyloidosis, exhibiting macroglossia and swelling salivary glands, uncommon clinical manifestations for DRA. Molecular analysis showed that the patient had a new variant of β₂-microglobulin (V27M). Extracted amyloid protein was predominantly composed of variant β₂-microglobulin. In vitro analysis revealed that this variant β₂-microglobulin had a strong amyloidogenic propensity, probably owing to the decreased stability caused by a bulky methionine residue. Our data clearly show that V27M variant is amyloidogenic and this mutation results in unusual clinical manifestations. To date, only one amyloidogenic β₂-microglobulin variant (D76N) has been reported in non-dialysis patients. It is noteworthy that the V27M and D76N variants show substantial differences in both clinical phenotypes and pathomechanical features. This is the first case of DRA associated with a naturally occurring β₂-microglobulin variant.

Title: insoluble proteins catch heterologous soluble proteins into inclusion bodies by intermolecular interaction of aggregating peptides

Background

Protein aggregation is a biological event observed in expression systems in which the recombinant protein is produced under stressful conditions surpassing the homeostasis of the protein quality control system. In addition, protein aggregation is also related to conformational diseases in animals as transmissible prion diseases or non-transmissible neurodegenerative diseases including Alzheimer, Parkinson’s disease, amyloidosis and multiple system atrophy among others. At the molecular level, the presence of aggregation-prone domains in protein molecules act as seeding igniters to induce the accumulation of protein molecules in protease-resistant clusters by intermolecular interactions.

Results

In this work we have studied the aggregating-prone performance of a small peptide (L6K2) with additional antimicrobial activity and we have elucidated the relevance of the accompanying scaffold protein to enhance the aggregating profile of the fusion protein. Furthermore, we demonstrated that the fusion of L6K2 to highly soluble recombinant proteins directs the protein to inclusion bodies (IBs) in E. coli through stereospecific interactions in the presence of an insoluble protein displaying the same aggregating-prone peptide (APP).

Conclusions

These data suggest that the molecular bases of protein aggregation are related to the net balance of protein aggregation potential and not only to the presence of APPs. This is then presented as a generic platform to generate hybrid protein aggregates in microbial cell factories for biopharmaceutical and biotechnological applications.

Protein aggregation: in silico algorithms and applications

Protein aggregation is a topic of immense interest to the scientific community due to its role in several neurodegenerative diseases/disorders and industrial importance. Several in silico techniques, tools, and algorithms have been developed to predict aggregation in proteins and understand the aggregation mechanisms. This review attempts to provide an essence of the vast developments in in silico approaches, resources available, and future perspectives. It reviews aggregation-related databases, mechanistic models (aggregation-prone region and aggregation propensity prediction), kinetic models (aggregation rate prediction), and molecular dynamics studies related to aggregation. With a multitude of prediction models related to aggregation already available to the scientific community, the field of protein aggregation is rapidly maturing to tackle new applications.

Machine Learning Feature Selection for Predicting High Concentration Therapeutic Antibody Aggregation

Protein aggregation can hinder the development, safety and efficacy of therapeutic antibody-based drugs. Developing a predictive model that evaluates aggregation behaviors during early stage development is therefore desirable. Machine learning is a widely used tool to train models that predict data with different attributes. However, most machine learning techniques require more data than is typically available in antibody development. In this work, we describe a rational feature selection framework to develop accurate models with a small number of features. We applied this framework to predict aggregation behaviors of 21 approved monospecific monoclonal antibodies at high concentration (150 mg/mL), yielding a correlation coefficient of 0.71 on validation tests with only two features using a linear model. The nearest neighbors and support vector regression models further improved the performance, which have correlation coefficients of 0.86 and 0.80, respectively. This framework can be extended to train other models that predict different physical properties.

Predicted aggregation-prone region (APR) in βB1-crystallin forms the amyloid-like structure and induces aggregation of soluble proteins isolated from human cataractous eye lens

The aggregation of β-crystallins in the human eye lens constitutes a critical step during the development of cataract. We anticipated that the presence of Aggregation-Prone Regions (APRs) in their primary structure, which might be responsible for conformational change required for the self-assembly. To examine the presence of APRs, we systematically analyzed the primary structures of β-crystallins. Out of seven subtypes, the βB1-crystallin found to possess the highest aggregation score with 9 APRs in its primary structure. To confirm the amyloidogenic nature of these newly identified APRs, we further studied the aggregation behavior of one of the APRs spanning from 174 to 180 residues (¹⁷⁴LWVYGFS¹⁸⁰) of βB1-crystallin, which is referred as βB1_(174-180). Under in vitro conditions, the synthetic analogue of βB1_(174-180) peptide formed visible aggregates and displayed high Congo red (CR) bathochromic shift, Thioflavin T (ThT) binding and fibrilar morphology under transmission electron microscopy, which are the typical characteristics of amyloids. Further, the aggregated βB1_(174-180) was found to induce aggregation of the soluble fraction of proteins isolated from the human cataractous lens. This observation suggests that the presence of APRs in βB1-crystallin might be serving as one of the intrinsic supplementary factors responsible for constitutive aggregation behavior of βB1-crystallin and development of cataract.

SOLart: a structure-based method to predict protein solubility and aggregation

MOTIVATION

The solubility of a protein is often decisive for its proper functioning. Lack of solubility is a major bottleneck in high-throughput structural genomic studies and in high-concentration protein production, and the formation of protein aggregates causes a wide variety of diseases. Since solubility measurements are time-consuming and expensive, there is a strong need for solubility prediction tools.

RESULTS

We have recently introduced solubility-dependent distance potentials that are able to unravel the role of residue-residue interactions in promoting or decreasing protein solubility. Here, we extended their construction by defining solubility-dependent potentials based on backbone torsion angles and solvent accessibility, and integrated them, together with other structure- and sequence-based features, into a random forest model trained on a set of Escherichia coli proteins with experimental structures and solubility values. We thus obtained the SOLart protein solubility predictor, whose most informative features turned out to be folding free energy differences computed from our solubility-dependent statistical potentials. SOLart performances are very good, with a Pearson correlation coefficient between experimental and predicted solubility values of almost 0.7 both in cross-validation on the training dataset and in an independent set of Saccharomyces cerevisiae proteins. On test sets of modeled structures, only a limited drop in performance is observed. SOLart can thus be used with both high-resolution and low-resolution structures, and clearly outperforms state-of-art solubility predictors. It is available through a user-friendly webserver, which is easy to use by non-expert scientists.

AVAILABILITY AND IMPLEMENTATION

The SOLart webserver is freely available at http://babylone.ulb.ac.be/SOLART/.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Computational prediction of protein aggregation: Advances in proteomics, conformation-specific algorithms and biotechnological applications

Protein aggregation is a widespread phenomenon that stems from the establishment of non-native intermolecular contacts resulting in protein precipitation. Despite its deleterious impact on fitness, protein aggregation is a generic property of polypeptide chains, indissociable from protein structure and function. Protein aggregation is behind the onset of neurodegenerative disorders and one of the serious obstacles in the production of protein-based therapeutics. The development of computational tools opened a new avenue to rationalize this phenomenon, enabling prediction of the aggregation propensity of individual proteins as well as proteome-wide analysis. These studies spotted aggregation as a major force driving protein evolution. Actual algorithms work on both protein sequences and structures, some of them accounting also for conformational fluctuations around the native state and the protein microenvironment. This toolbox allows to delineate conformation-specific routines to assist in the identification of aggregation-prone regions and to guide the optimization of more soluble and stable biotherapeutics. Here we review how the advent of predictive tools has change the way we think and address protein aggregation.

Coiled-coil inspired functional inclusion bodies

BACKGROUND

Recombinant protein expression in bacteria often leads to the formation of intracellular insoluble protein deposits, a major bottleneck for the production of soluble and active products. However, in recent years, these bacterial protein aggregates, commonly known as inclusion bodies (IBs), have been shown to be a source of stable and active protein for biotechnological and biomedical applications. The formation of these functional IBs is usually facilitated by the fusion of aggregation-prone peptides or proteins to the protein of interest, leading to the formation of amyloid-like nanostructures, where the functional protein is embedded.

RESULTS

In order to offer an alternative to the classical amyloid-like IBs, here we develop functional IBs exploiting the coiled-coil fold. An in silico analysis of coiled-coil and aggregation propensities, net charge, and hydropathicity of different potential tags identified the natural homo-dimeric and anti-parallel coiled-coil ZapB bacterial protein as an optimal candidate to form assemblies in which the native state of the fused protein is preserved. The protein itself forms supramolecular fibrillar networks exhibiting only α-helix secondary structure. This non-amyloid self-assembly propensity allows generating innocuous IBs in which the recombinant protein of interest remains folded and functional, as demonstrated using two different fluorescent proteins.

CONCLUSIONS

Here, we present a proof of concept for the use of a natural coiled-coil domain as a versatile tool for the production of functional IBs in bacteria. This α-helix-based strategy excludes any potential toxicity drawback that might arise from the amyloid nature of β-sheet-based IBs and renders highly active and homogeneous submicrometric particles.