Biology subjects

Ward, A.

Publications and source records attributed to Ward, A..

At least 37 records · Page 2Linked to original sources

bioRxiv2025

Replaying germinal center evolution on a quantified affinity landscape

Darwinian evolution of immunoglobulin genes within germinal centers (GCs) underlies the progressive increase in antibody affinity following antigen exposure. Whereas the cellular mechanics of how competition between B cells produces increases in affinity are well established, the evolutionary dynamics of this process are less clear. We developed an experimental evolution model where we "replay" over one hundred monoclonal GC reactions, assigning affinities to each cell using deep mutational scanning. Our data reveal how GCs achieve predictable outcomes by means of noisy but persistent selection on an affinity landscape whose exploration is heavily constrained by somatic hypermutation biases. We infer a fitness landscape that quantitatively recapitulates the affinity maturation trajectory of our clone and find that apparent features of GC selection such as permissiveness to low-affinity lineages and rapid plateauing of affinity are likely artifacts of survivorship biases that distort our view of how B cell affinity progresses over time.

immunology↗

bioRxiv2025

Programmed inhibition of an innate immune receptor via de novo designed transmembrane proteins

Transmembrane domains of immune complexes transmit precise signals across lipid bilayers. Probing their interactions has the potential to yield mechanistic insights relevant to therapeutic design. However, our capability to generate molecules directed to bind lipid-embedded sites is limited. Here, we demonstrate a computational strategy to design polypeptides targeting Toll-like receptor 4 (TLR4), a mediator of inflammatory signaling, directly within membranes. TLR4 poses a formidable molecular recognition challenge, as its transmembrane domain is largely apolar, lacks a defined sequence motif, and exhibits an underdetermined structure-function relationship. One synthetic protein binds TLR4s transmembrane domain and antagonizes NF{kappa}B signaling in human cells, proving that precise transmembrane domain interactions are essential for cross-membrane conformational coupling. This work refines design principles for encoding stable interactions in cellular membranes and expands the range of lipid-embedded mechanisms accessible to probe with computationally derived molecules.

biochemistry↗

bioRxiv2025

An O-acetylated derivative of piericidin A1 produced by Kitasatospora sp. A2-31 has potent activity against the cacao mirid bug, Helopeltis bakeri Poppius

There is an increasing demand for novel biopesticides to protect agricultural products and improve yields. To address this need, extracts from a library of Actinomycetes collected in the Philippines were evaluated against the cacao mirid bug, Helopeltis bakeri Poppius. Analysis of an active ethyl acetate extract from Kitasatospora sp. A2-31 led to the identification of a novel metabolite acetylpiericidin A1 (3) along with the known natural products piericidin A1 (1), piericidin A5 (2), chromomycin A2 (4), chromomycin A3 (5), and olivomycin A (6). Acetylpiericidin A1 demonstrated 100 % mortality against H. bakeri, while other metabolites exhibited either weak or no activity. Whole genome sequencing followed by bioinformatics analysis identified a gene downstream of the piericidin biosynthetic gene cluster that encodes a putative acetyltransferase proposed to catalyze acetylation of the C10 hydroxyl group of piericidin A1. The involvement of this gene in acetylpiericidin A1 biosynthesis was confirmed by (i) introducing an additional copy under the control of the ermE* promoter into Kitasatospora sp. A2-31, resulting in elevated production levels and (ii) through an in vitro enzymatic assay of the corresponding purified recombinant enzyme.

microbiology↗

bioRxiv2025

Tissue resident memory T cells populate the human uveal tract

The current concept is that the eye is an immune privileged site endowed with innate immune regulatory networks to maintain organ function. We now have evidence that resident T cells occupy intraocular tissues. In immune-mediated inflammatory diseases, such as psoriasis and rheumatoid arthritis, tissue resident T cells trigger disease flares in the skin and joints. This suggests resident T cells in the uvea may have similar functions in non-infectious immune-mediated uveitis, a collective term for autoinflammatory and autoimmune diseases of the uveal tract causing intraocular inflammation. Here, we demonstrate by spectral cytometry and immunofluorescence imaging that non-inflamed uveal tissue contains multiple T cell subtypes including CD8+ CD103+ tissue resident memory T (TRM) cells. Using single cell RNA & T cell receptor (TCR) sequencing to profile aqueous humour cells from donors with acute, active uveitis, we identify clonally expanded T cells which are enriched for TRM -associated genes. We further show that in donors with active uveitis, CD8+ CD103+ T cells persist within tissue in the uveal tract. Using bulk RNA sequencing and weighted gene co-expression network analysis (WGCNA) we show that quiescent iris tissue from donors with a history of uveitis are enriched for genes associated with T cell activation and antigen presentation. Finally, we demonstrate that TRM cells persist in the anterior uvea in mice following resolution of experimental autoimmune uveoretinitis (EAU). Our results show that the human eye contains T cells both in health and during active inflammation. Our findings challenge the dogma that the eye is devoid of lymphocytes and supports the concept of resident T cell involvement in the pathogenesis of non-infectious immune-mediated uveitis and as promising targets for therapy. One Sentence SummaryT cells infiltrate aqueous humour during intraocular inflammation and have capacity to migrate into uveal tissue where they remain long-lived.

immunology↗

bioRxiv2025

Characterization of the glycoproteins of novel fish influenza B-like viruses

Novel influenza-like virus sequences previously identified in fish and amphibians were found to cluster as a sister clade of influenza B viruses, but have thus far remained uncharacterized. We demonstrate that salamander influenza-like virus (SILV) HA is functionally divergent from influenza B virus HA and does not bind to 2,3- and 2,6-linked sialic acids. However, the HAs of Siamese algae-eater influenza-like virus (SAEILV) and chum salmon influenza-like virus (CSILV) bind to 2,3 linked sialic acid. Furthermore, SAEILV HA binds to sialyated Lewis X, is activated by human airway enzymes and is fusogenic at a wide range of pH conditions. SAEILV NA has a highly conserved active site and a similar structure to other known NAs. We also determined the cryo-electron microscopy structure of the HA of a previously described virus from the same sister clade, the Wuhan spiny eel influenza virus (WSEIV). Importantly, no cross-reactive antibodies against these HAs or NAs were found in the human serum, suggesting that humans are immunologically naive to these viruses. One sentence summaryNovel influenza like-viruses, displayed different target receptor specificity and limited antigenic conservation of the HA and NA relative to influenza B Viruses.

microbiology↗

bioRxiv2025

Light-guided molecular patterning for programmable multiplexed single-molecule manipulation

Single-molecule force spectroscopy enables the detailed probing of molecular interactions, providing new insights into molecular mechanisms--yet studying biology "one molecule at a time" can lead to throughput challenges that limit applications. While multiplexed single-molecule assays can address these issues, suitable functionalization of surfaces is required, which remains a technical challenge--many commonly used approaches are constrained by random and sparse biomolecules arrangements, limiting programmability and throughput. An ideal anchoring method would enable (i) high surface densities to maximize throughput, (ii) precise control of spatial position and molecular identity for maximum control, (iii) covalent linking for high force application, and (iv) efficient patterning without the need for expensive facilities to maximize accessibility. To achieve these aims, we have developed a light-guided surface patterning method that can covalently organize oligonucleotides (oligos) without the need for lithographic equipment. Oligos with 3-Cyanovinylcarbazole (CNVK) nucleoside are crosslinked by UV patterns reflected through a digital micromirror device (DMD), with beads arranged accordingly. To demonstrate compatibility with established single-molecule methods, we performed single-molecule force spectroscopy experiments on patterned coverslips, using both magnetic tweezers and hydrodynamic-based systems. Our light-guided approach provides a scalable and accessible solution for biomolecular patterning that allows precise control over molecular identity and spatial positioning, enabling high-throughput measurements in single-molecule research.

biophysics↗

bioRxiv2025

Microbiota-derived secondary bile acids promote STING activation and antitumor activity

Microbiota metabolism generates diverse bile acids that are associated with health and disease, but the molecular targets and mechanisms of action for these metabolites have not been fully elucidated. Using deoxycholic acid (DCA) photoaffinity probes, with cholic acid (CA) probe as a comparator, we found many protein targets of microbiota-derived DCA in mammalian cells. Of note, we discovered that DCA binds the transmembrane domain of stimulator of interferon genes (STING), and enhances agonist-induced accumulation of higher-order STING species, type I interferon signaling and antitumor immunity. Moreover, DCA-producing microbiota species enhanced STING agonist-induced antitumor immunity primarily through the activation of natural killer (NK) cells. This reverse chemical microbiology approach revealed an unpredicted mechanism of action for DCA on STING activation and suggests that specific secondary bile acids and their associated microbiota species may promote NK activation and impact the efficacy of STING-targeted therapeutics.

immunology↗

bioRxiv2025

Bayesian analysis and efficient algorithms for single-molecule fluorescence data and step counting

With the growing adoption of single-molecule fluorescence experiments, there is an increasing demand for efficient statistical methodologies and accurate analysis of the acquired measurements. Existing analysis frameworks, such as those that use kinetic models, often rely on strong assumptions on the dynamics of the molecules and fluorophores under study that render them inappropriate for general purpose step counting applications, especially when the systems of study exhibit uncharacterized dynamics. Here, we propose a novel Bayesian nonparametric framework to analyze singlemolecule fluorescence data that is kinetic model independent. For the evaluation of our methods, we develop four MCMC samplers, ranging from elemental to highly sophisticated, and demonstrate that the added complexity is essential for accurate data analysis. We apply our methods to experimental data obtained from TIRF photobleaching assays of the EphA2 receptor tagged with GFP. In addition, we validate our approach with synthetic data mimicking realistic conditions and demonstrate its ability to recover ground truth under high- and low-signal-to-noise data, establishing it as a versatile tool for fluorescence data analysis. Significance statementProtein complexes are critical for cell function. Advances in fluorescence experiments have facilitated their direct study, achieving single-molecule resolution in the determination of their stoichiometry. However, the analysis of raw fluorescence data remains a challenge. Traditional visual inspection methods are often time-consuming and susceptible to user bias, while conventional statistical approaches are limited by strong assumptions regarding molecular and probe dynamics or exceedingly high computational requirements. In this study, we introduce a novel statistical methodology for robust and efficient analysis of fluorescence data. Our innovative approach not only enhances the accuracy and reliability of data analysis following a fluorescence experiment but also allows high-throughput applications. Our methods present a significant advancement that may extend the scope of current fluorescence techniques.

biochemistry↗

bioRxiv2025

Human monoclonal antibodies that target clade 2.3.4.4b H5N1 hemagglutinin

The highly pathogenic avian influenza H5N1 virus clade 2.3.4.4b has been spreading globally since 2022, causing mortality and morbidity in domestic and wild birds and mammals, including infection in humans, raising concerns about its pandemic potential. We aimed to generate a panel of anti-hemagglutinin (HA) human monoclonal antibodies (mAbs) against the H5 protein of clade 2.3.4.4b. H2L2 Harbour Mice(R), which express human immunoglobulin germline genes, were immunized with H5 and N1 recombinant proteins from A/mallard/New York/22-008760-007-original/2022 H5N1 virus, enabling the generation of human chimeric antibodies. Through hybridoma technology, sixteen full human mAbs were generated, most of which showed cross-reactivity against H5 proteins from different virus variants. The functionality of the sixteen mAbs was assessed in vitro using hemagglutination inhibition and microneutralization assays with viruses containing a clade 2.3.4.4b HA. Fourteen out of the sixteen mAbs neutralized the virus in vitro. The mAbs with the strongest hemagglutination inhibition activity also demonstrated greater neutralizing capacity and showed increased protective effects in vivo when administered prophylactically or therapeutically in a murine H5N1 challenge model. Using cryo-electron microscopy, we identified a cross-clonotype conserved motif that bound a hydrophobic groove on the head domain of H5 HA. Akin to mAbs against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) during the coronavirus 2019 (COVID-19) pandemic, these mAbs could serve as important treatments in case of a widespread H5N1 epidemic or pandemic.

microbiology↗

bioRxiv2025

Low level contamination confounds population genomic analysis

Genome sequence contamination has a variety of causes and can originate from within or between species. Previous research focused primarily on cross-species contamination or on prokaryotes. This paper visualizes B-allele frequency to test for intra-species contamination, and measures its effects on phylogenetic and admixture analysis in two fungal species. Using a standard base calling pipeline, we found that contaminated genomes superficially appeared to produce good quality genome data. Yet as little as 5-10% genome contamination was enough to change phylogenetic tree topologies and make contaminated strains appear as hybrids between lineages (genetically admixed). We recommend the use of B-allele frequency plots to screen genome resequencing data for intra-species contamination.

genomics↗

bioRxiv2024

Metabolic Switch in Endocrine Resistant Estrogen Receptor Positive Breast Cancer

PurposeThe development of endocrine resistance remains a significant challenge in the clinical management of estrogen receptor-positive (ER+) breast cancer. Metabolic reprogramming is a prominent component of endocrine resistance and a potential therapeutic intervention point. However, a limited understanding of which metabolic changes are conserved across the heterogeneous landscape of ER+ breast cancer or how metabolic changes factor into ER DNA binding patterns hinder our ability to target metabolic adaptation as a treatment strategy. This study uses dimethyl fumarate (DMF) to restore tamoxifen (Tam) and fulvestrant (Fulv) sensitivity in endocrine-resistant cell lines and investigates how metabolic changes influence ER DNA-binding patterns. Experimental DesignTo address the challenge of metabolic adaptation in anti-endocrine resistance, we generated Tam and Fulv resistance in six ER+ breast cancer (BC) cell lines, representing ductal (MCF7, T47D, ZR75-1, and UCD12), lobular (MDA-MB-134--VI), and HER2 amplified (BT474) BC molecular phenotypes. Metabolomic profiling, RNA sequencing, proteomics, and CUT&RUN assays were completed to characterize metabolic shifts, transcriptional and protein changes, and ER DNA-binding patterns in resistant cells. Dimethyl fumarate was assessed for its ability to reverse Tam and Fulv resistance, restore tricarboxylic acid cycle (TCA) cycle function, and restore parental cell (endocrine sensitive) ER DNA binding patterns. ResultsTamoxifen-resistant (TamR) and fulvestrant-resistant (FulvR) cells exhibited disrupted TCA cycle activity, reduced glutathione levels, and altered nucleotide and amino acid metabolism. DMF treatment replenished TCA cycle intermediates and reversed resistance in both TamR and FulvR cells. DMF also increased mevalonate pathway enzyme expression in both TamR and FulvR cells, with TamR cells upregulating enzymes in the cholesterol synthesis phase and FulvR enhancing enzymes in the early part of the pathway. DMF restored ER DNA-binding patterns in TamR cells to resemble parental cells, re-sensitizing them to Tam. In FulvR cells, DMF reversed resistance by modulating ER-cofactor interactions but did not restore parental ER DNA-binding signatures. ConclusionsOur findings provide new insights into how metabolic reprogramming affects ER DNA-binding activity in endocrine-resistant breast cancer. We demonstrate how altering metabolism can reprogram ER signaling and influence resistance mechanisms by targeting metabolic vulnerabilities, such as TCA cycle disruptions. Additionally, our data provide a comprehensive metabolomic, RNA-seq, and CUT&RUN data set relevant to tumor metabolic adaptation leading to acquired endocrine resistance in highly utilized ER+ breast cancer cell lines. This study improves our understanding of how metabolic states alter ER function in endocrine-resistant breast cancer.

cancer biology↗

bioRxiv2024

Cell-specific delivery of GJB2 restores auditory function in mouse models of DFNB1 deafness and mediates appropriate expression in NHP cochlea

Mutations in the GJB2 gene cause the most common form of human hereditary hearing loss, known as DFNB1. GJB2 is expressed in two cell groups of the cochlea--epithelial cells of the organ of Corti and fibrocytes of the inner sulcus and lateral wall--but not by sensory hair cells or neurons. Attempts to treat mouse models of DFNB1 with AAV vectors mediating nonspecific Gjb2 expression have not substantially restored function, perhaps because inappropriate expression in hair cells and neurons could compromise their electrical activity. Here, we used genomic chromatin accessibility profiling to identify candidate gene regulatory elements (GREs) that could drive cell-type-specific expression of Gjb2 in the cochlea. HA-tagged GJB2, delivered to a conditional knockout model in an AAV vector with GRE control of expression, was localized to the appropriate cell types, prevented the cochlear degeneration observed in untreated knockout mice, and partially rescued hearing sensitivity. In a Gjb2 partial knockdown mouse model, such exogenous GJB2 prevented degeneration and completely restored hearing sensitivity. We tested control of expression by these GREs in nonhuman primate cochleas and found that vector-delivered human GJB2.HA was located in the appropriate cell types and caused little or no reduction in hearing sensitivity. Together, these findings suggest that GRE-mediated expression of GJB2 could prevent hearing loss in DFNB1 patients.

neuroscience↗

bioRxiv2024

Somatic mtDNA Mutation Burden Shapes Metabolic Plasticity in Leukemogenesis

The role of somatic mitochondrial DNA (mtDNA) mutations in leukemogenesis remains poorly characterized. To determine the impact of somatic mtDNA mutations on the process, we assessed the leukemogenic potential of hematopoietic progenitor cells (HPCs) from mtDNA mutator mice (Polg D257A) with or without NMyc overexpression. We observed a higher incidence of spontaneous leukemogenesis in recipients transplanted with heterozygous Polg HPCs and a lower incidence of NMyc-driven leukemia in those with homozygous Polg HPCs compared to controls. Although mtDNA mutations in heterozygous and homozygous HPCs caused similar baseline impairments in mitochondrial function, only heterozygous HPCs responded to and supported altered metabolic demands associated with NMyc overexpression. Homozygous HPCs showed altered glucose utilization with pyruvate dehydrogenase inhibition due to increased phosphorylation, exacerbated by NMyc overexpression. The impaired growth of NMyc-expressing homozygous HPCs was partially rescued by inhibiting pyruvate dehydrogenase kinase, highlighting a relationship between mtDNA mutation burden and metabolic plasticity in leukemogenesis. TEASERSomatic mtDNA mutations as drivers of metabolic change in the development of leukemia.

cancer biology↗

bioRxiv2024

Potent and selective repression of SCN9A by engineered zinc finger repressors for the treatment of neuropathic pain

Peripheral neuropathies are estimated to affect several million patients in the US with no long-lasting therapy currently available. In humans, the Nav1.7 sodium channel, encoded by the SCN9A gene, is involved in a spectrum of inherited neuropathies, and has emerged as a promising target for analgesic drug development. The development of a selective Nav1.7 inhibitor has been challenging, in part due to structural similarities among other Nav channels. Here we present preclinical studies for the first genomic medicine approach using engineered zinc finger repressors (ZFRs) specifically targeting human/nonhuman primate (NHP) SCN9A. AAV-mediated delivery of ZFRs in human iPSC-derived neurons resulted in 90% reduction of SCN9A with no detectable off-target activity. To establish proof-of-concept, a ZFR targeting the mouse Scn9a was assessed in the SNI neuropathic pain mouse model, which resulted in up to 70% repression of Scn9a in mouse DRGs and was associated with reduction in pain hypersensitivity as measured by increased mechanical- and cold-induced pain thresholds. AAV-mediated intrathecal delivery of ZFR in NHPs demonstrated up to 60% repression of SCN9A in bulk DRG tissue and on single-cell levels in the nociceptors. The treatment was well tolerated in NHPs, and no dose-limiting findings were observed four weeks after a single intrathecal injection. Taken together, our results demonstrate that AAV-delivered ZFR targeting the SCN9A gene is promising and supports further development as a potential therapy for peripheral neuropathies.

neuroscience↗

bioRxiv2024

Cholesterol inhibits oncogenic EphA2 activation

The receptor tyrosine kinase EphA2 drives cancer malignancy by facilitating metastasis. EphA2 can be found in different self-assembly states: as a monomer, dimer, and oligomer. However, our understanding remains limited regarding which EphA2 state is responsible for driving pro-metastatic signaling. To address this limitation, we have developed SiMPull-POP, a single-molecule method for accurate quantification of membrane protein self-assembly. Our experiments revealed that a reduction of plasma membrane cholesterol strongly promoted EphA2 self-assembly. Indeed, low cholesterol caused a similar effect to the EphA2 ligand ephrinA1-Fc. These results indicate that cholesterol inhibits EphA2 assembly. Phosphorylation studies in different cell lines revealed that low cholesterol increased phospho-serine levels, the signature of oncogenic signaling. Investigation of the mechanism that cholesterol uses to inhibit the assembly and activity of EphA2 indicate an in-trans effect, where EphA2 is phosphorylated by protein kinase A downstream of beta-adrenergic receptor activity, which cholesterol also inhibits. Our study not only provides new mechanistic insights on EphA2 oncogenic function, but also suggests that cholesterol acts as a molecular safeguard mechanism that prevents uncontrolled self-assembly and activation of EphA2.

biochemistry↗

bioRxiv2024

IL-23 drives uveitis by acting on a novel population of tissue-resident entheseal T cells

Recurrent acute anterior uveitis is a frequent extra-articular manifestation of the axial spondyloarthropathies (AxSpA); chronic inflammatory diseases affecting the spine, enthesis, peripheral joints, skin, and gastrointestinal tract. Pathology in AxSpA has been associated with local tissue-resident populations of interleukin (IL)-23 responsive lymphoid cells. Here we reveal a novel population of ocular T cells defined by CD3+CD4-CD8-{gamma}{delta}TCR+IL-23R+ that reside within the anterior uvea as an ocular entheseal analogue of the mouse eye. Localised cytokine expression demonstrates that uveal IL-23R+ IL-17A-producing cells are both necessary and sufficient to drive uveitis in response to IL-23. This T cell population is also present in humans, occupying extravascular tissues of the anterior uveal compartment. Consistent with the concept of IL-23 as a unifying mediator in AxSpA, we present evidence that IL-23 can also act locally on tissue resident T cells in the anterior compartment of the eye at sites analogous to the enthesis to drive ocular inflammation.

immunology↗

bioRxiv2024

Repeat modules and N-linked glycans define structure and antigenicity of a critical enterotoxigenic E. coli adhesin

Enterotoxigenic Escherichia coli (ETEC) cause hundreds of millions of cases of infectious diarrhea annually, predominantly in children from low-middle income regions. Notably, in children, as well as human volunteers challenged with ETEC, diarrheal severity is significantly increased severity in blood group A (bgA) individuals. EtpA, is a secreted glycoprotein adhesin that functions as a blood group A lectin to promote critical interactions between ETEC and blood group A glycans on intestinal epithelia for effective bacterial adhesion and toxin delivery. EtpA is highly immunogenic resulting in robust antibody responses following natural infection and experimental challenge of human volunteers with ETEC. To understand how EtpA directs ETEC-blood group A interactions and stimulates adaptive immunity, we mutated EtpA, mapped its glycosylation by mass-spectrometry (MS), isolated polyclonal (pAbs) and monoclonal antibodies (mAbs) from vaccinated mice and ETEC-infected human volunteers, and determined structures of antibody-EtpA complexes by cryo-electron microscopy. Both bgA and mAbs that inhibited EtpA-bgA interactions and ETEC adhesion, bound to the C-terminal repeat domain highlighting this region as crucial for ETEC pathogen-host interaction. MS analysis uncovered extensive and heterogeneous N-linked glycosylation of EtpA and cryo-EM structures revealed that mAbs directly engage these unique glycan containing epitopes. Finally, electron microscopy-based polyclonal epitope mapping revealed antibodies targeting numerous distinct epitopes on N and C-terminal domains, suggesting that EtpA vaccination generates responses against neutralizing and decoy regions of the molecule. Collectively, we anticipate that these data will inform our general understanding of pathogen-host glycan interactions and adaptive immunity relevant to rational vaccine subunit design. Author summaryEnterotoxigenic E. coli (ETEC), a leading cause of diarrhea disproportionately affecting young children in low-income regions, are a priority for vaccine development. Individuals possessing A blood-type are more susceptible to severe cholera-like disease. EtpA, a secreted, immunogenic, blood group A binding protein, is a current vaccine target antigen. Here, we determined the atomic structure of EtpA in complex with protective as well as non-protective monoclonal antibodies targeting two different domains of the protein, allowing us to pinpoint key regions involved in blood-group A antigen recognition and uncover the mechanism of antibody-based protection. In addition, we show through mass-spectrometry that EtpA is extensively and heterogeneously glycosylated at surface-exposed asparagine residues by a promiscuous and low-fidelity glycosyltransferase, EtpC, and that this unique form of bacterial glycosylation is critical for to development of protective immune responses. Lastly, polyclonal antibodies from vaccinated mice as well as monoclonal antibodies obtained from ETEC-infected human volunteers revealed that the highly antigenic surface of EtpA exhibits both protective and non-protective epitopes. These results greatly expand our understanding of ETEC pathogenesis, and the immune responses elicited by these common infections, providing valuable information to aid in the rational design and testing of subunit vaccines.

microbiology↗

bioRxiv2024

Assessing AF2's ability to predict structural ensembles of proteins

Recent breakthroughs in protein structure prediction have enhanced the precision and speed at which protein configurations can be determined, setting new benchmarks for accuracy and efficiency in the field. However, the fundamental mechanisms of biological processes at a molecular level are often connected to conformational changes of proteins. Molecular dynamics (MD) simulations serve as a crucial tool for capturing the conformational space of proteins, providing valuable insights into their structural fluctuations. However, the scope of MD simulations is often limited by the accessible timescales and the computational resources available, posing challenges to comprehensively exploring protein behaviors. Recently emerging approaches have focused on expanding the capability of AlphaFold2 (AF2) to predict conformational substates of protein structures by manipulating the input multiple sequence alignment (MSA). These approaches operate under the assumption that the MSA also contains information about the heterogeneity of protein structures. Here, we benchmark the performance of various workflows that have adapted AF2 for ensemble prediction focusing on the subsampling of the MSA as implemented in ColabFold and compare the obtained structures with ensembles obtained from MD simulations and NMR. As test cases, we chose four proteins namely the bovine pancreatic inhibitor protein (BPTI), thrombin and two antigen binding fragments (antibody Fv and nanobody), for which reliable experimentally validated structural information (X-ray and/or NMR) was available. Thus, we provide an overview of the levels of performance and accessible timescales that can currently be achieved with machine learning (ML) based ensemble generation. In three out of the four test cases, we find structural variations fall within the predicted ensembles. Nevertheless, significant minima of the free energy surfaces remain undetected. This study highlights the possibilities and pitfalls when generating ensembles with AF2 and thus may guide the development of future tools while informing upon the results of currently available applications.

biophysics↗