Predicting glycan structure from tandem mass spectrometry via deep learning.

Glycans constitute the most complicated post-translational modification, modulating protein activity in health and disease. However, structural annotation from tandem mass spectrometry (MS/MS) data is a bottleneck in glycomics, preventing high-throughput endeavors and relegating glycomics to a few experts. Trained on a newly curated set of 500,000 annotated MS/MS spectra, here we present CandyCrunch, a dilated residual neural network predicting glycan structure from raw liquid chromatography-MS/MS data in seconds (top-1 accuracy: 90.3%). We developed an open-access Python-based workflow of raw data conversion and prediction, followed by automated curation and fragment annotation, with predictions recapitulating and extending expert annotation. We demonstrate that this can be used for de novo annotation, diagnostic fragment identification and high-throughput glycomics. For maximum impact, this entire pipeline is tightly interlaced with our glycowork platform and can be easily tested at https://colab.research.google.com/github/BojarLab/CandyCrunch/blob/main/CandyCrunch.ipynb . We envision CandyCrunch to democratize structural glycomics and the elucidation of biological roles of glycans.

A hydrophobic funnel governs monovalent cation selectivity in the ion channel TRPM5.

A key capability of ion channels is the facilitation of selective permeation of certain ionic species across cellular membranes at high rates. Due to their physiological significance, ion channels are of great pharmaceutical interest as drug targets. The polymodal signal-detecting transient receptor potential (TRP) superfamily of ion channels forms a particularly promising group of drug targets. While most members of this family permeate a broad range of cations including Ca2+, TRPM4 and TRPM5 are unique due to their strong monovalent selectivity and impermeability for divalent cations. Here, we investigated the mechanistic basis for their unique monovalent selectivity by in silico electrophysiology simulations of TRPM5. Our simulations reveal an unusual mechanism of cation selectivity, which is underpinned by the function of the central channel cavity alongside the selectivity filter. Our results suggest that a subtle hydrophobic barrier at the cavity entrance ("hydrophobic funnel") enables monovalent but not divalent cations to pass and occupy the cavity at physiologically relevant membrane voltages. Monovalent cations then permeate efficiently by a cooperative, distant knock-on mechanism between two binding regions in the extracellular pore vestibule and the central cavity. By contrast, divalent cations do not enter or interact favorably with the channel cavity due to its raised hydrophobicity. Hydrophilic mutations in the transition zone between the selectivity filter and the central channel cavity abolish the barrier for divalent cations, enabling both monovalent and divalent cations to traverse TRPM5.

Classification of likely functional class for ligand binding sites identified from fragment screening.

Fragment screening is used to identify binding sites and leads in drug discovery, but it is often unclear which binding sites are functionally important. Here, data from 37 experiments, and 1309 protein structures binding to 1601 ligands were analysed. A method to group ligands by binding sites is introduced and sites clustered according to profiles of relative solvent accessibility. This identified 293 unique ligand binding sites, grouped into four clusters (C1-4). C1 includes larger, buried, conserved, and population missense-depleted sites, enriched in known functional sites. C4 comprises smaller, accessible, divergent, missense-enriched sites, depleted in functional sites. A site in C1 is 28 times more likely to be functional than one in C4. Seventeen sites, which to the best of our knowledge are novel, in 13 proteins are identified as likely to be functionally important with examples from human tenascin and 5-aminolevulinate synthase highlighted. A multi-layer perceptron, and K-nearest neighbours model are presented to predict cluster labels for ligand binding sites with an accuracy of 96% and 100%, respectively, so allowing functional classification of sites for proteins not in this set. Our findings will be of interest to those studying protein-ligand interactions and developing new drugs or function modulators.

Role of N343 glycosylation on the SARS-CoV-2 S RBD structure and co-receptor binding across variants of concern.

Glycosylation of the SARS-CoV-2 spike (S) protein represents a key target for viral evolution because it affects both viral evasion and fitness. Successful variations in the glycan shield are difficult to achieve though, as protein glycosylation is also critical to folding and structural stability. Within this framework, the identification of glycosylation sites that are structurally dispensable can provide insight into the evolutionary mechanisms of the shield and inform immune surveillance. In this work, we show through over 45 µs of cumulative sampling from conventional and enhanced molecular dynamics (MD) simulations, how the structure of the immunodominant S receptor binding domain (RBD) is regulated by N-glycosylation at N343 and how this glycan's structural role changes from WHu-1, alpha (B.1.1.7), and beta (B.1.351), to the delta (B.1.617.2), and omicron (BA.1 and BA.2.86) variants. More specifically, we find that the amphipathic nature of the N-glycan is instrumental to preserve the structural integrity of the RBD hydrophobic core and that loss of glycosylation at N343 triggers a specific and consistent conformational change. We show how this change allosterically regulates the conformation of the receptor binding motif (RBM) in the WHu-1, alpha, and beta RBDs, but not in the delta and omicron variants, due to mutations that reinforce the RBD architecture. In support of these findings, we show that the binding of the RBD to monosialylated ganglioside co-receptors is highly dependent on N343 glycosylation in the WHu-1, but not in the delta RBD, and that affinity changes significantly across VoCs. Ultimately, the molecular and functional insight we provide in this work reinforces our understanding of the role of glycosylation in protein structure and function and it also allows us to identify the structural constraints within which the glycosylation site at N343 can become a hotspot for mutations in the SARS-CoV-2 S glycan shield.

Functional implications of glycans and their curation: insights from the workshop held at the 16th Annual International Biocuration Conference in Padua, Italy.

Dynamic changes in protein glycosylation impact human health and disease progression. However, current resources that capture disease and phenotype information focus primarily on the macromolecules within the central dogma of molecular biology (DNA, RNA, proteins). To gain a better understanding of organisms, there is a need to capture the functional impact of glycans and glycosylation on biological processes. A workshop titled "Functional impact of glycans and their curation" was held in conjunction with the 16th Annual International Biocuration Conference to discuss ongoing worldwide activities related to glycan function curation. This workshop brought together subject matter experts, tool developers, and biocurators from over 20 projects and bioinformatics resources. Participants discussed four key topics for each of their resources: (i) how they curate glycan function-related data from publications and other sources, (ii) what type of data they would like to acquire, (iii) what data they currently have, and (iv) what standards they use. Their answers contributed input that provided a comprehensive overview of state-of-the-art glycan function curation and annotations. This report summarizes the outcome of discussions, including potential solutions and areas where curators, data wranglers, and text mining experts can collaborate to address current gaps in glycan and glycosylation annotations, leveraging each other's work to improve their respective resources and encourage impactful data sharing among resources. Database URL: https://wiki.glygen.org/Glycan_Function_Workshop_2023.

A cooperative knock-on mechanism underpins Ca2+-selective cation permeation in TRPV channels.

The selective exchange of ions across cellular membranes is a vital biological process. Ca2+-mediated signaling is implicated in a broad array of physiological processes in cells, while elevated intracellular concentrations of Ca2+ are cytotoxic. Due to the significance of this cation, strict Ca2+ concentration gradients are maintained across the plasma and organelle membranes. Therefore, Ca2+ signaling relies on permeation through selective ion channels that control the flux of Ca2+ ions. A key family of Ca2+-permeable membrane channels is the polymodal signal-detecting transient receptor potential (TRP) ion channels. TRP channels are activated by a wide variety of cues including temperature, small molecules, transmembrane voltage, and mechanical stimuli. While most members of this family permeate a broad range of cations non-selectively, TRPV5 and TRPV6 are unique due to their strong Ca2+ selectivity. Here, we address the question of how some members of the TRPV subfamily show a high degree of Ca2+ selectivity while others conduct a wider spectrum of cations. We present results from all-atom molecular dynamics simulations of ion permeation through two Ca2+-selective and two non-selective TRPV channels. Using a new method to quantify permeation cooperativity based on mutual information, we show that Ca2+-selective TRPV channel permeation occurs by a three-binding site knock-on mechanism, whereas a two-binding site knock-on mechanism is observed in non-selective TRPV channels. Each of the ion binding sites involved displayed greater affinity for Ca2+ over Na+. As such, our results suggest that coupling to an extra binding site in the Ca2+-selective TRPV channels underpins their increased selectivity for Ca2+ over Na+ ions. Furthermore, analysis of all available TRPV channel structures shows that the selectivity filter entrance region is wider for the non-selective TRPV channels, slightly destabilizing ion binding at this site, which is likely to underlie mechanistic decoupling.

Undecaprenol kinase: Function, mechanism and substrate specificity of a potential antibiotic target.

The bifunctional undecaprenol kinase/phosphatase (UdpK) is a small, prokaryotic, integral membrane kinase, homologous with Escherichia coli diacylglycerol kinase and expressed by the dgkA gene. In Gram-positive bacteria, UdpK is involved in the homeostasis of the bacterial undecaprenoid pool, where it converts undecaprenol to undecaprenyl phosphate (C55P) and also catalyses the reverse process. C55P is the universal lipid carrier and critical to numerous glycopolymer and glycoprotein biosynthetic pathways in bacteria. DgkA gene expression has been linked to facilitating bacterial growth and survival in response to environmental stressors, as well being implicated as a resistance mechanism to the topical antibiotic bacitracin, by providing an additional route to C55P. Therefore, identification of UdpK inhibitors could lead to novel antibiotic treatments. A combination of homology modelling and mutagenesis experiments on UdpK have been used to identify residues that may be involved in kinase/phosphatase activity. In this review, we will summarise recent work on the mechanism and substrate specificity of UdpK.

An antibiotic-resistance conferring mutation in a neisserial porin: Structure, ion flux, and ampicillin binding.

Gram-negative bacteria cause the majority of highly drug-resistant bacterial infections. To cross the outer membrane of the complex Gram-negative cell envelope, antibiotics permeate through porins, trimeric channel proteins that enable the exchange of small polar molecules. Mutations in porins contribute to the development of drug-resistant phenotypes. In this work, we show that a single point mutation in the porin PorB from Neisseria meningitidis, the causative agent of bacterial meningitis, can strongly affect the binding and permeation of beta-lactam antibiotics. Using X-ray crystallography, high-resolution electrophysiology, atomistic biomolecular simulation, and liposome swelling experiments, we demonstrate differences in drug binding affinity, ion selectivity and drug permeability of PorB. Our work further reveals distinct interactions between the transversal electric field in the porin eyelet and the zwitterionic drugs, which manifest themselves under applied electric fields in electrophysiology and are altered by the mutation. These observations may apply more broadly to drug-porin interactions in other channels. Our results improve the molecular understanding of porin-based drug-resistance in Gram-negative bacteria.

A two-lane mechanism for selective biological ammonium transport.

The transport of charged molecules across biological membranes faces the dual problem of accommodating charges in a highly hydrophobic environment while maintaining selective substrate translocation. This has been the subject of a particular controversy for the exchange of ammonium across cellular membranes, an essential process in all domains of life. Ammonium transport is mediated by the ubiquitous Amt/Mep/Rh transporters that includes the human Rhesus factors. Here, using a combination of electrophysiology, yeast functional complementation and extended molecular dynamics simulations, we reveal a unique two-lane pathway for electrogenic NH4+ transport in two archetypal members of the family, the transporters AmtB from Escherichia coli and Rh50 from Nitrosomonas europaea. The pathway underpins a mechanism by which charged H+ and neutral NH3 are carried separately across the membrane after NH4+ deprotonation. This mechanism defines a new principle of achieving transport selectivity against competing ions in a biological transport process.