Biology subjects

Matsen, F. A.

Publications and source records attributed to Matsen, F. A..

At least 19 recordsLinked to original sources

bioRxiv2026

Phylogenetically estimated neutral rates and fitness effects of mutations to influenza proteins

Influenza viruss rapid evolution is shaped by both neutral mutation and selection. Phylogenetics can be used to study these processes, but this approach has typically only been applied to a few thousand influenza genome sequences at once. Here, we built phylogenetic trees with >100,000 influenza sequences, and then used these trees to estimate neutral rates of mutations to the viruss genome. Neutral rates varied by up to ~100-fold among the 12 nucleotide mutation types (A[->]C,A[->]G, etc.). These rates were highly correlated among influenza, SARS-CoV-2, and HIV, though more nuanced context-dependent patterns showed marked differences between influenza and SARS-CoV-2. We also estimated fitness effects of mutations by comparing the number of times a mutation was observed to occur along the branches of a tree to the number of times we expect it to have occurred under neutrality. We estimated effects for ~33,000 nonsynonymous and ~8,000 synonymous mutations spanning all influenza proteins. This compendium of estimated effects helps map the relationship between sequence and fitness in a natural setting, including regions where synonymous mutations are under functional constraint, and for proteins with limited experimentally measured effects. We built interactive heatmaps of the estimated fitness effects to help readers explore these data (see https://matsen.group/flu-mut-rates). Altogether, this work places influenzas mutation rates in a broader cross-viral context and deepens our understanding of how mutation and selection shape influenza evolution in nature at a site-specific level.

bioinformatics↗

bioRxiv2026

Entrenchment of germline amino-acid differences in antibody affinity maturation

Entrenchment -- epistasis that locks in amino acid differences between homologous proteins, so each disfavors substitutions toward the others state -- has been demonstrated along individual protein lineages over deep evolutionary time. Antibodies offer a unique system for studying entrenchment: multiple homologous germline V gene paralogs provide diverse starting points, and the rapid somatic evolution of affinity maturation generates dense phylogenies from which selection on germline-encoded residues can be inferred. Using DASM, a deep-learning model that separates selection from mutation in antibody repertoire data, we test for entrenchment across immunoglobulin heavy chain variable (IGHV) genes. We detect entrenchment at two levels of germline divergence, driven by different sources of epistasis. Within V gene families (up to [~]20% amino acid divergence), entrenched sites cluster at the borders of the complementarity-determining regions (CDRs, the antigen-binding loops) and show high germline diversity. These sites contact antigen, light chain, and the heavy chain CDR3 loop, all of which are encoded independently of the IGHV germline. This pattern is consistent with epistasis from genetically uncoupled partners. Between V gene families, at deeper levels of divergence (25-40%), entrenchment additionally includes positions in the framework scaffold distant from binding interfaces, suggesting a larger contribution from intra-heavy-chain structural constraints. Observed mutation frequencies in human repertoires corroborate these predictions where data are sufficient. Together, these results demonstrate that the rapid somatic evolution of antibodies can serve as a lens for revealing epistatic constraints acting on germline-encoded residues, including constraints imposed by genetically uncoupled partners assembled during B cell development.

evolutionary biology↗

bioRxiv2026

Unifying phylogenetic traversal and deep learning to guide tree exploration

Deep learning offers hope for more efficient phylogenetic inference methods. However, it has yet to have the transformative effect on phylogenetics that it has had in other fields. Here we present a novel approach that combines deep learning with concepts behind current successful phylogenetic algorithms. Specifically, we give the deep learning algorithm access to the output of a phylogenetic dynamic program on the sequence alignment, rather than the raw sequence alignment. The algorithm then learns features based on these phylogenetically processed versions of the sequence data, providing information to guide local tree search. For this paper, our goal is simple: predict for each edge in a tree whether it is in a maximum parsimony tree or not. Our model consists of a recurrent neural network that learns features while traversing the input tree, which are used to classify the edge. The model makes high-quality predictions for this NP-complete problem on simulated and empirical datasets for trees of various sizes. We believe it is a stepping stone towards efficient phylogenetic inference using deep learning.

bioinformatics↗

bioRxiv2025

larch: mapping the parsimony-optimal landscape of trees for directed exploration

Phylogenetic inference algorithms for large data sets typically return a single tree. However, there are often many optimal trees, especially when sequence data is closely related. We develop a compact representation of large collections of maximally parsimonious histories--trees with mutations mapped onto tree edges. Our C++ implementation, larch, leverages this representation for a highly parallel search algorithm. The storage component uses our history DAG structure to compactly represent large families of optimal trees. The search algorithm integrates this storage with matOptimize for rapid tree optimization; the DAG structure allows us to accept thousands of conflicting tree rearrangements in parallel. The integration enables a new type of tree search: one that systematically maps out the collection of good trees, enabling moves that are directed away from the current set of optimal trees to cross valleys and increase the diversity of the set of optimal trees. It is able to identify more parsimonious trees than are found by other methods. We find diverse optimality landscapes for viral datasets, including many distinct plateaux. We also find that our implementation produces similar results whether using a variety of single starting trees or an ensemble of starting trees, indicating effective global optimization.

evolutionary biology↗

bioRxiv2025

Separating selection from mutation in antibody language models

Antibodies are encoded by nucleotide sequences that are generated by V(D)J recombination and evolve according to mutation and selection processes. Existing antibody language models, however, focus exclusively on antibodies as strings of amino acids and are fitted using standard language modeling objectives such as masked or autoregressive prediction. In this paper, we first show that fitting models using this objective implicitly incorporates nucleotide-level mutation processes as part of the protein language model, which degrades performance when predicting effects of mutations on functional properties of antibodies. To address this limitation, we devise a new framework: a Deep Amino acid Selection Model (DASM) that learns the selection effects of amino-acid mutations while explicitly factoring out the nucleotide-level mutation process. By fitting selection as a separate term from the mutation process, the DASM exclusively quantifies functional effects: effects that change some aspect of the function of the antibody. This factorization leads to substantially improved performance on standard functional benchmarks. Moreover, our model is an order of magnitude smaller and multiple orders of magnitude faster to evaluate than existing approaches, as well as being readily interpretable.

evolutionary biology↗

bioRxiv2025

Inference of germinal center evolutionary dynamics via simulation-based deep learning

B cells and the antibodies they produce are vital to health and survival, motivating research on the details of the mutational and evolutionary processes in the germinal centers (GCs) from which mature B cells arise. It is known that B cells with higher affinity for their cognate antigen (Ag) will, on average, tend to have more offspring. However, the exact form of this relationship between affinity and fecundity, which we call the "affinity-fitness response function", is not known. Here we use deep learning and simulation-based inference to learn this function from a unique experiment that replays a particular combination of GC conditions many times. All code is freely available at https://github.com/matsengrp/gcdyn, while datasets and inference results can be found at https://doi.org/10.5281/zenodo.15022130.

evolutionary biology↗

bioRxiv2025

Nucleotide context models outperform protein language models for predicting antibody affinity maturation

Antibodies play a crucial role in adaptive immunity. They develop as B cell receptors (BCRs): membrane-bound forms of antibodies that are expressed on the surfaces of B cells. BCRs are refined through affinity maturation, a process of somatic hypermutation (SHM) and natural selection, to improve binding to an antigen. Computational models of affinity maturation have developed from two main perspectives: molecular evolution and language modeling. The molecular evolution perspective focuses on nucleotide sequence context to describe mutation and selection; the language modeling perspective involves learning patterns from large data sets of protein sequences. In this paper, we compared models from both perspectives on their ability to predict the course of antibody affinity maturation along phylogenetic trees of BCR sequences. This included models of SHM, models of SHM combined with an estimate of selection, and protein language models. We evaluated these models for large human BCR repertoire data sets, as well as an antigen-specific mouse experiment with a pre-rearranged cognate naive antibody. We demonstrated that precise modeling of SHM, which requires the nucleotide context, provides a substantial amount of predictive power for predicting the course of affinity maturation. Notably, a simple nucleotide-based convolutional neural network modeling SHM outperformed state-of-the-art protein language models, including one trained exclusively on antibody sequences. Furthermore, incorporating estimates of selection based on a custom deep mutational scanning experiment brought only modest improvement in predictive power. To support further research, we introduce EPAM (Evaluating Predictions of Affinity Maturation), a benchmarking framework to integrate evolutionary principles with advances in language modeling, offering a road map for understanding antibody evolution and improving predictive models.

immunology↗

bioRxiv2025

Biological causes and impacts of rugged tree landscapes in phylodynamic inference

Phylodynamic analysis has been instrumental in elucidating epidemiological and evolutionary dynamics of pathogens. Bayesian phylodynamics integrates out phylogenetic uncertainty, which is typically substantial in phylodynamic datasets due to limited genetic diversity. Phylodynamic inference does not, however, scale with modern datasets, partly due to difficulties in traversing tree space. Here, we characterize tree space and landscape in phylodynamic inference and assess its impacts on analysis difficulty and key biological estimates. By running extensive Bayesian analyses of 15 classic large phylodynamic datasets and carefully analyzing the posterior samples, we find that the posterior tree landscape is diffuse yet rugged, leading to widespread tree sampling problems that usually stem from sequences in a small part of the tree. We develop clade-specific diagnostics to show that a few sequences--including putative recombinants and recurrent mutants--frequently drive the ruggedness and sampling problems, although existing data-quality tests show limited power to detect them. The sampling problems can significantly impact phylodynamic inferences or distort major biological conclusions; the impact is usually stronger on "local" estimates (e.g., introduction history) associated with particular clades than on "global" parameters (e.g., demographic trajectory) governed by general tree shape. We evaluate existing and newly-developed MCMC diagnostics, and offer strategies for optimizing phylodynamic analysis settings and mitigating sampling problem impacts. Our findings highlight the need and directions to develop efficient traversal over rugged tree landscapes, ultimately advancing scalable and reliable phylodynamics. Significance StatementBayesian phylodynamics is central to epidemiological studies, but exploring the vast and complex tree space is computationally challenging. Phylodynamic datasets comprise many highly similar sequences, sampled through time, creating a uniquely structured landscape of optimal trees. Here, we show that phylodynamic tree landscapes are often highly rugged, with multiple peaks separated by difficult-to-cross valleys. These features lead to widespread sampling problems which are often driven by a few sequences. These problems can significantly impact phylodynamic estimates, especially those associated with particular clades, distorting biological conclusions. We develop diagnostics to identify problematic sequences and provide solutions to mitigate their impacts. We offer strategies to optimize phylodynamic analysis workflows and to develop algorithms for navigating rugged landscapes, thereby advancing infectious disease investigation.

evolutionary biology↗

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

Increased immunogen valency improves the maturation of vaccine-elicited HIV-1 VRC01-class antibodies

Antibodies belonging to the VRC01-class display broad and potent neutralizing activities and have been isolated from several people living with HIV (PLWH). A member of that class, monoclonal antibody VRC01, was shown to reduce HIV-acquisition in two phase 2b efficacy trials. VRC01-class antibodies are therefore expected to be a key component of an effective HIV-1 vaccine. In contrast to the VRC01-class antibodies that are highly mutated, their unmutated forms do not engage HIV-1 envelope (Env) and do not display neutralizing activities. Hence, specifically modified Env-derived proteins have been designed to engage the unmutated forms of VRC01-class antibodies, and to activate the corresponding naive B cells. Selected heterologous Env must then be used as boost immunogens to guide the proper maturation of these elicited VRC01-class antibodies. Here we examined whether and how the valency of the prime and boost immunogens influences VRC01-class antibody-maturation. Our findings indicate that, indeed the valency of the immunogen affects the maturation of elicited antibody responses by preferentially selecting VRC01-class antibodies that have accumulated somatic mutations present in broadly neutralizing VRC01-class antibodies isolated from PLWH. As a result, antibodies isolated from animals immunized with the higher valency immunogens display broader Env cross-binding properties and improved neutralizing potentials than those isolated from animals immunized with the lower valency immunogens. Our results are relevant to current and upcoming phase 1 clinical trials that evaluate the ability of novel immunogens aiming to elicit cross-reactive VRC01-class antibody responses. AUTHOR SUMMARYPA performed ELISA, B cell sorting, BCR sequencing analysis, oversaw the analysis of all immunochemical assays, interpreted the results and wrote the manuscript; AK processed tissues from immunized animals, performed B cell staining, and generated mAbs; KRS performed BLI and BCR sequencing; AM performed BCR sequencing; DR and FAM performed phylogenetic analysis and contributed to writing the manuscript; AR performed ELISA and processed tissues from immunized animals, ABS expressed and purified recombinant Envs; LKS expressed and purified mVRC01 mAb; XS and DM performed neutralization assays; LS conceived and oversaw the study and wrote the manuscript.

immunology↗

bioRxiv2025

Clonal interference and changing selective pressures shape the escape of SARS-CoV-2 from hundreds of antibodies

SARS-CoV-2 rapidly evolves to evade human immunity. While the viruss overall resistance to human polyclonal antibody responses has steadily increased over time, the dynamics by which it escaped individual monoclonal antibodies within these responses have not been thoroughly explored. Recently, a series of studies by Cao et al. [1, 2, 3] used deep mutational scanning (DMS) to identify which mutations allow the Wuhan-Hu-1 receptor-binding domain to escape binding by individual antibodies, doing so for thousands of antibodies. Here, we sought to use these data to retrospectively examine the evolutionary dynamics of escape from a set of 1,603 antibodies. For each antibody, we used the DMS data to predict an antibody-escape score for each of thousands of globally circulating viral sequences from the first 3.5 years of the pandemic, and then computed an escape trajectory that quantifies how the populations average escape score changed over time. We use pseudovirus neutralization data from Cao et al. and Wang et al. [4] to validate common patterns in escape trajectories. While some trajectories increase monotonically over time, others show large fluctuations as a result of clade-displacement events that reduce the frequency of antibody-escape mutations in the viral population. Fitness effects of mutations estimated from natural sequences suggest that the mutations are displaced due to clonal interference. Further, these estimates suggest that the order in which escape mutations arose is shaped by changing selective pressures. Overall, this work helps describe how SARS-CoV-2 evaded the individual components of a polyclonal immune response in nature, and suggests that evasion occurred via complex evolutionary dynamics.

microbiology↗

bioRxiv2025

The mutation rate of SARS-CoV-2 is highly variable between sites and is influenced by sequence context, genomic region, and RNA structure

RNA viruses like SARS-CoV-2 have a high mutation rate, which contributes to their rapid evolution. The rate of mutations depends on the mutation type (e.g., A[->]C, A[->]G, etc.) and can vary between sites in the viral genome. Understanding this variation can shed light on the mutational processes at play, and is crucial for quantitative modeling of viral evolution. Using the millions of available SARS-CoV-2 full-genome sequences, we estimate rates of synonymous mutations for all 12 possible nucleotide mutation types and examine how much these rates vary between sites. We find a surprisingly high level of variability and several striking patterns: the rates of four mutation types suddenly increase at one of two gene boundaries; the rates of most mutation types strongly depend on a sites local sequence context, with up to 56-fold differences between contexts; consistent with a previous study, the rates of some mutation types are lower at sites engaged in RNA secondary structure. A simple log-linear model of these features explains [~]15-60% of the fold-variation of mutation rates between sites, depending on mutation type; more complex models only modestly improve predictive power out of sample. We estimate the fitness effect of each mutation based on the number of times it actually occurs versus the number of times it is expected to occur based on the model. We identify several small regions of the genome where synonymous or noncoding mutations occur much less often than expected, indicative of strong purifying selection on the RNA sequence that is independent of protein sequence. Overall, this work expands our basic understanding of SARS-CoV-2s evolution by characterizing the viruss mutation process at the level of individual sites and uncovering several striking mutational patterns that arise from unknown mechanisms.

bioinformatics↗

bioRxiv2024

Thrifty wide-context models of B cell receptor somatic hypermutation

Somatic hypermutation (SHM) is the diversity-generating process in antibody affinity maturation. Probabilistic models of SHM are needed for analyzing rare mutations, for understanding the selective forces guiding affinity maturation, and for understanding the underlying biochemical process. High throughput data offers the potential to develop and fit models of SHM on relevant data sets. In this paper we model SHM using modern frameworks. We are motivated by recent work suggesting the importance of a wider context for SHM, however, assigning an independent rate to each k-mer leads to an exponential proliferation of parameters. Thus, using convolutions on 3-mer embeddings, we develop "thrifty" models of SHM of various sizes; these can have fewer free parameters than a 5-mer model and yet have a significantly wider context. These offer a slight performance improvement over a 5-mer model, and other modern model elaborations worsen performance. We also find that a per-site effect is not necessary to explain SHM patterns given nucleotide context. Also, the two current methods for fitting an SHM model -- on out-of-frame sequence data and on synonymous mutations -- produce significantly different results, and augmenting out-of-frame data with synonymous mutations does not aid out-of-sample performance.

immunology↗

bioRxiv2024

Learning antibody sequence constraints from allelic inclusion

Antibodies and B-cell receptors (BCRs) are produced by B cells, and are built of a heavy chain and a light chain. Although each B cell could express two different heavy chains and four different light chains, usually only a unique pair of heavy chain and light chain is expressed--a phenomenon known as allelic exclusion. However, a small fraction of naive-B cells violate allelic exclusion by expressing two productive light chains, one of which has impaired function; this has been called allelic inclusion. We demonstrate that these B cells can be used to learn constraints on antibody sequence. Using large-scale single-cell sequencing data from humans, we find examples of light chain allelic inclusion in thousands of naive-B cells, which is an order of magnitude larger than existing datasets. We train machine learning models to identify the abnormal sequences in these cells. The resulting models correlate with antibody properties that they were not trained on, including polyreactivity, surface expression, and mutation usage in affinity maturation. These correlations are larger than what is achieved by existing antibody modeling approaches, indicating that allelic inclusion data contains useful new information. We also investigate the impact of similar selection forces on the heavy chain in mouse, and observe that pairing with the surrogate light chain significantly restricts heavy chain diversity.

immunology↗

bioRxiv2024

Statistical analysis of repertoire data demonstrates the influence of microhomology in V(D)J recombination

V(D)J recombination generates the diverse B and T cell receptors essential for recognizing a wide array of antigens. This diversity arises from the combinatorial assembly of V(D)J genes and the junctional deletion and insertion of nucleotides. While previous in vitro studies have shown that microhomology---short stretches of sequence homology between gene ends---can bias the recombination process, the extent of microhomologys impact in vivo, particularly in humans, remains unknown. In this paper, we assess how germline-encoded microhomology influences trimming and ligation during V(D)J recombination using statistical inference on previously-published high-throughput TCR repertoire sequencing data. We find that microhomology increases both trimming and ligation probabilities, making it an important predictor of recombination outcomes. These effects are consistent across different receptor loci and sequence types. Further, we demonstrate that accounting for microhomology effects significantly alters sequence annotation probabilities and rankings, highlighting its practical importance for accurately inferring the V(D)J recombination events that generated an observed sequence. Together, these results enhance our understanding of how microhomologous nucleotides shape the human V(D)J recombination process. Significance StatementHumans rely on diverse adaptive immune receptor repertoires to effectively defend against infections. The receptor generation process, known as V(D)J recombination, is designed to create this diversity by stochastically joining V(D)J gene segments and modifying their junctions through nucleotide deletions and insertions. Previous studies, conducted in vitro, have suggested that short stretches of homologous nucleotides between gene segments can bias these recombination steps. In this study, we explore the extent to which these homologous nucleotides influence V(D)J recombination in humans using statistical inference on large-scale receptor repertoire sequencing data. Our findings reveal that microhomology significantly biases several recombination steps, with important practical implications for the analysis, processing, and interpretation of receptor sequences.

immunology↗

bioRxiv2024

Optics-free reconstruction of 2D images via DNA barcode proximity graphs

Spatial genomics technologies include imaging- and sequencing-based methods. Sequencing-based spatial methods typically require surfaces coated with coordinate-associated DNA barcodes, but the physical registration of these barcodes to spatial coordinates is challenging, necessitating either high density printing of oligonucleotides or in situ sequencing/probing of randomly deposited, DNA-barcode-bearing beads. As a consequence, the surface areas available to sequencing-based spatial genomic methods are constrained by the time, labor, cost and instrumentation required to either print or decode a coordinate-tagged surface. To address this challenge, we developed SCOPE (Spatial reConstruction via Oligonucleotide Proximity Encoding), an optics-free, DNA microscopy-inspired method. With SCOPE, the relative positions of DNA-barcoded beads within a 2D shape, 2D image or 3D volume are inferred from the ex situ sequencing of chimeric molecules formed from diffusing "sender" and tethered "receiver" oligonucleotides. To demonstrate the potential of this approach, we applied SCOPE to reconstruct 2D shapes, 2D images or 3D volumes defined by 104-106 x 20-100 {micro}m DNA barcoded beads, including an asymmetric "swoosh" resembling the Nike logo (44 mm2), a "color" Snellen eye chart (704 mm2) and the surface topology of 3D molds of a teddy bear, star, butterfly or block letter (75-100 mm3). Each of the resulting "DNA barcode proximity graphs" was computationally reconstructed in an automated fashion, across fields of view and at resolutions that were determined by sequencing depth, bead size and diffusion kinetics, rather than by microarray or microscope instrument time. Because the ground truth shapes are known, these datasets may be particularly useful for the further development of computational algorithms by this nascent field.

genomics↗

bioRxiv2024

Delineating the functional activity of antibodies with cross-reactivity to SARS-CoV-2, SARS-CoV-1 and related sarbecoviruses

The recurring spillover of pathogenic coronaviruses and demonstrated capacity of sarbecoviruses, such SARS-CoV-2, to rapidly evolve in humans underscores the need to better understand immune responses to this virus family. For this purpose, we characterized the functional breadth and potency of antibodies targeting the receptor binding domain (RBD) of the spike glycoprotein that exhibited cross-reactivity against SARS-CoV-2 variants, SARS-CoV-1 and sarbecoviruses from diverse clades and animal origins with spillover potential. One neutralizing antibody, C68.61, showed remarkable neutralization breadth against both SARS-CoV-2 variants and viruses from different sarbecovirus clades. C68.61, which targets a conserved RBD class 5 epitope, did not select for escape variants of SARS-CoV-2 or SARS-CoV-1 in culture nor have predicted escape variants among circulating SARS-CoV-2 strains, suggesting this epitope is functionally constrained. We identified 11 additional SARS-CoV-2/SARS-CoV-1 cross-reactive antibodies that target the more sequence conserved class 4 and class 5 epitopes within RBD that show activity against a subset of diverse sarbecoviruses with one antibody binding every single sarbecovirus RBD tested. A subset of these antibodies exhibited Fc-mediated effector functions as potent as antibodies that impact infection outcome in animal models. Thus, our study identified antibodies targeting conserved regions across SARS-CoV-2 variants and sarbecoviruses that may serve as therapeutics for pandemic preparedness as well as blueprints for the design of immunogens capable of eliciting cross-neutralizing responses. AUTHOR SUMMARYThere is a large collection of sarbecoviruses related to SARS-CoV-2 circulating in animal reservoirs with the potential to spillover into humans. Neutralizing antibodies have the potential to protect against infection, although viral escape is common. In this study, we isolated several monoclonal antibodies that show broad activity against different sarbecoviruses. The antibodies target epitopes in the core of the receptor binding domain that are highly conserved in sequence across sarbecoviruses and emerging SARS-CoV-2 variants. One antibody showed remarkable breadth against both SARS-CoV-1 variants as well as diverse sarbecoviruses. The results of deep mutational scanning suggest that mutations at these predicted sites of escape may functionally constrain viral fitness. Our functional profiling of cross-reactive antibodies highlights vulnerable sites of sarbecoviruses, with some antibodies poised as broadly neutralizing candidates for therapeutic use against future sarbecovirus emergence.

immunology↗

bioRxiv2024

The S2 subunit of spike encodes diverse targets for functional antibody responses to SARS-CoV-2

The SARS-CoV-2 virus responsible for the COVID-19 global pandemic has exhibited a striking capacity for viral evolution that drives continued evasion from vaccine and infection-induced immune responses. Mutations in the receptor binding domain of the S1 subunit of the spike glycoprotein have led to considerable escape from antibody responses, reducing the efficacy of vaccines and monoclonal antibody (mAb) therapies. Therefore, there is a need to interrogate more constrained regions of Spike, such as the S2 subdomain. Here, we describe a collection of S2 mAbs from two SARS-CoV-2 convalescent individuals that target multiple regions in the S2 subdomain and can be grouped into at least five epitope classes. Most did not neutralize SARS-CoV-2 with the exception of C20.119, which bound to a highly conserved epitope in the fusion peptide and showed broad binding and neutralization activity across SARS-CoV-2, SARS-CoV-1, and closely related zoonotic sarbecoviruses. Several of the S2 mAbs tested mediated antibody-dependent cellular cytotoxicity (ADCC) at levels similar to the S1 mAb S309 that was previously authorized for treatment of SARS-CoV-2 infections. Three of the mAbs with ADCC function also bound to spike trimers from HCoVs, such as MERS-CoV and HCoV-HKU1. Our findings suggest there are diverse epitopes in S2, including functional S2 mAbs with HCoV and sarbecovirus breadth that likely target functionally constrained regions of spike. These mAbs could be developed for potential future pandemics, while also providing insight into ideal epitopes for eliciting a broad HCoV response. AUTHOR SUMMARYThe early successes of vaccines and antibody therapies against SARS-CoV-2, the virus responsible for the COVID-19 global pandemic, leveraged the considerable antibody response to the viral entry protein, spike, after vaccination or infection. These initial interventions were highly effective at protecting from infection and reducing severe disease or death. However, SARS-CoV-2 has shown no sign of abating, with the continued rise of new variants that have escaped some of the antibody defense due to distinct alterations most significantly in regions of the spike protein that elicit most of the anti-viral, functional antibody response. These findings suggest a critical need to identify vaccine approaches and therapies that provide the broadest possible antibody responses, focused on regions of spike critical for SARS-CoV-2 infection and, therefore, do not undergo changes that could lead to immune evasion. Our study describes a panel of functional antibodies, from individuals after SARS-CoV-2 infection, that recognize the S2 spike subdomain that is responsible for carrying out viral fusion with host cells. These regions in S2 are generally well conserved across SARS-CoV-2 variants and other closely related viruses and thus, could guide more effective vaccine design in the face of continued viral evolution.

immunology↗