Decoding glycosylation potential from protein structure across human glycoproteins with a multi-view recurrent neural network.

Glycosylation is described as a non-templated biosynthesis. Yet, the template-free premise is antithetical to the observation that different N-glycans are consistently placed at specific sites. It has been proposed that glycosite-proximal protein structures could constrain glycosylation and explain the observed microheterogeneity. Using site-specific glycosylation data, we trained a hybrid neural network to parse glycosites (recurrent neural network) and match them to feasible N-glycosylation events (graph neural network). From glycosite-flanking sequences, the algorithm predicts most human N-glycosylation events documented in the GlyConnect database and proposed structures corresponding to observed monosaccharide composition of the glycans at these sites. The algorithm also recapitulated glycosylation in Enhanced Aromatic Sequons, SARS-CoV-2 spike, and IgG3 variants, thus demonstrating the ability of the algorithm to predict both glycan structure and abundance. Thus, protein structure constrains glycosylation, and the neural network enables predictive in silico glycosylation of uncharacterized or novel protein sequences and genetic variants.

Simulated digestions of free oligosaccharides and mucin-type O-glycans reveal a potential role for Clostridium perfringens.

The development of a stable human gut microbiota occurs within the first year of life. Many open questions remain about how microfloral species are influenced by the composition of milk, in particular its content of human milk oligosaccharides (HMOs). The objective is to investigate the effect of the human HMO glycome on bacterial symbiosis and competition, based on the glycoside hydrolase (GH) enzyme activities known to be present in microbial species. We extracted from UniProt a list of all bacterial species catalysing glycoside hydrolase activities (EC 3.2.1.-), cross-referencing with the BRENDA database, and obtained a set of taxonomic lineages and CAZy family data. A set of 13 documented enzyme activities was selected and modelled within an enzyme simulator according to a method described previously in the context of biosynthesis. A diverse population of experimentally observed HMOs was fed to the simulator, and the enzymes matching specific bacterial species were recorded, based on their appearance of individual enzymes in the UniProt dataset. Pairs of bacterial species were identified that possessed complementary enzyme profiles enabling the digestion of the HMO glycome, from which potential symbioses could be inferred. Conversely, bacterial species having similar GH enzyme profiles were considered likely to be in competition for the same set of dietary HMOs within the gut of the newborn. We generated a set of putative biodegradative networks from the simulator output, which provides a visualisation of the ability of organisms to digest HMO and mucin-type O-glycans. B. bifidum, B. longum and C. perfringens species were predicted to have the most diverse GH activity and therefore to excel in their ability to digest these substrates. The expected cooperative role of Bifidobacteriales contrasts with the surprising capacities of the pathogen. These findings indicate that potential pathogens may associate in human gut based on their shared glycoside hydrolase digestive apparatus, and which, in the event of colonisation, might result in dysbiosis. The methods described can readily be adapted to other enzyme categories and species as well as being easily fine-tuneable if new degrading enzymes are identified and require inclusion in the model.

HumanLectome, an update of UniLectin for the annotation and prediction of human lectins.

The UniLectin portal (https://unilectin.unige.ch/) was designed in 2019 with the goal of centralising curated and predicted data on carbohydrate-binding proteins known as lectins. UniLectin is also intended as a support for the study of lectomes (full lectin set) of organisms or tissues. The present update describes the inclusion of several new modules and details the latest (https://unilectin.unige.ch/humanLectome/), covering our knowledge of the human lectome and comprising 215 unevenly characterised lectins, particularly in terms of structural information. Each HumanLectome entry is protein-centric and compiles evidence of carbohydrate recognition domain(s), specificity, 3D-structure, tissue-based expression and related genomic data. Other recent improvements regarding interoperability and accessibility are outlined.

Protein glycosylation and glycoinformatics for novel biomarker discovery in neurodegenerative diseases.

Glycosylation is a common post-translational modification of brain proteins including cell surface adhesion molecules, synaptic proteins, receptors and channels, as well as intracellular proteins, with implications in brain development and functions. Using advanced state-of-the-art glycomics and glycoproteomics technologies in conjunction with glycoinformatics resources, characteristic glycosylation profiles in brain tissues are increasingly reported in the literature and growing evidence shows deregulation of glycosylation in central nervous system disorders, including aging associated neurodegenerative diseases. Glycan signatures characteristic of brain tissue are also frequently described in cerebrospinal fluid due to its enrichment in brain-derived molecules. A detailed structural analysis of brain and cerebrospinal fluid glycans collected in publications in healthy and neurodegenerative conditions was undertaken and data was compiled to create a browsable dedicated set in the GlyConnect database of glycoproteins (https://glyconnect.expasy.org/brain). The shared molecular composition of cerebrospinal fluid with brain enhances the likelihood of novel glycobiomarker discovery for neurodegeneration, which may aid in unveiling disease mechanisms, therefore, providing with novel therapeutic targets as well as diagnostic and progression monitoring tools.

Glycosaminoglycans: What Remains To Be Deciphered?

Glycosaminoglycans (GAGs) are complex polysaccharides exhibiting a vast structural diversity and fulfilling various functions mediated by thousands of interactions in the extracellular matrix, at the cell surface, and within the cells where they have been detected in the nucleus. It is known that the chemical groups attached to GAGs and GAG conformations comprise "glycocodes" that are not yet fully deciphered. The molecular context also matters for GAG structures and functions, and the influence of the structure and functions of the proteoglycan core proteins on sulfated GAGs and vice versa warrants further investigation. The lack of dedicated bioinformatic tools for mining GAG data sets contributes to a partial characterization of the structural and functional landscape and interactions of GAGs. These pending issues will benefit from the development of new approaches reviewed here, namely (i) the synthesis of GAG oligosaccharides to build large and diverse GAG libraries, (ii) GAG analysis and sequencing by mass spectrometry (e.g., ion mobility-mass spectrometry), gas-phase infrared spectroscopy, recognition tunnelling nanopores, and molecular modeling to identify bioactive GAG sequences, biophysical methods to investigate binding interfaces, and to expand our knowledge and understanding of glycocodes governing GAG molecular recognition, and (iii) artificial intelligence for in-depth investigation of GAGomic data sets and their integration with proteomics.

BiotechLec: an interactive guide of commercial lectins for glycobiology and biomedical research applications.

For decades, lectins have been used as probes in glycobiology and this usage has gradually spread to other domains of Life Science. Nowadays, researchers investigate glycan recognition with lectins in diverse biotechnology and clinical applications, addressing key questions regarding binding specificity. The latter is documented in scattered and heterogeneous sources, and this situation calls for a centralized and easy-access reference. To address this need, an on-line solution called BiotechLec (https://www.unilectin.eu/biotechlec) is proposed in a new section of UniLectin, a platform dedicated to lectin molecular knowledge.

The Glycan Structure Dictionary-a dictionary describing commonly used glycan structure terms.

Recent technological advances in glycobiology have resulted in a large influx of data and the publication of many papers describing discoveries in glycoscience. However, the terms used in describing glycan structural features are not standardized, making it difficult to harmonize data across biomolecular databases, hampering the harvesting of information across studies and hindering text mining and curation efforts. To address this shortcoming, the Glycan Structure Dictionary has been developed as a reference dictionary to provide a standardized list of widely used glycan terms that can help in the curation and mapping of glycan structures described in publications. Currently, the dictionary has 190 glycan structure terms with 297 synonyms linked to 3,332 publications. For a term to be included in the dictionary, it must be present in at least 2 peer-reviewed publications. Synonyms, annotations, and cross-references to GlyTouCan, GlycoMotif, and other relevant databases and resources are also provided when available. The purpose of this effort is to facilitate biocuration, assist in the development of text mining tools, improve the harmonization of search, and browse capabilities in glycoinformatics resources and help to map glycan structures to function and disease. It is also expected that authors will use these terms to describe glycan structures in their manuscripts over time. A mechanism is also provided for researchers to submit terms for potential incorporation. The dictionary is available at https://wiki.glygen.org/Glycan_structure_dictionary.

Worldwide Glycoscience Informatics Infrastructure: The GlySpace Alliance.

The GlySpace Alliance was formed in 2018 among the principal investigators of three major glycoscience portals: Glyco@Expasy, GlyCosmos, and GlyGen, representing Europe, Asia, and the United States, respectively. While each of these portals has its unique user interface, the aim is to provide the same basic data set of glycan-related omics data. These portals will be introduced with the aim to enable users to find their target information in the most efficient manner, in particular, in terms of the chemical structures of glycans and their functions.

The choanoflagellate pore-forming lectin SaroL-1 punches holes in cancer cells by targeting the tumor-related glycosphingolipid Gb3.

Choanoflagellates are primitive protozoa used as models for animal evolution. They express a large variety of multi-domain proteins contributing to adhesion and cell communication, thereby providing a rich repertoire of molecules for biotechnology. Adhesion often involves proteins adopting a ß-trefoil fold with carbohydrate-binding properties therefore classified as lectins. Sequence database screening with a dedicated method resulted in TrefLec, a database of 44714 ß-trefoil candidate lectins across 4497 species. TrefLec was searched for original domain combinations, which led to single out SaroL-1 in the choanoflagellate Salpingoeca rosetta, that contains both ß-trefoil and aerolysin-like pore-forming domains. Recombinant SaroL-1 is shown to bind galactose and derivatives, with a stronger affinity for cancer-related α-galactosylated epitopes such as the glycosphingolipid Gb3, when embedded in giant unilamellar vesicles or cell membranes. Crystal structures of complexes with Gb3 trisaccharide and GalNAc provided the basis for building a model of the oligomeric pore. Finally, recognition of the αGal epitope on glycolipids required for hemolysis of rabbit erythrocytes suggests that toxicity on cancer cells is achieved through carbohydrate-dependent pore-formation.

This is GlycoQL.

MOTIVATION: We have previously designed and implemented a tree-based ontology to represent glycan structures with the aim of searching these structures with a glyco-driven syntax. This resulted in creating the GlySTreeM knowledge-base as a linchpin of the structural matching procedure and we now introduce a query language, called GlycoQL, for the actual implementation of a glycan structure search. RESULTS: The methodology is described and illustrated with a use-case focused on Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) spike protein glycosylation. We show how to enhance site annotation with federated queries involving UniProt and GlyConnect, our glycoprotein database. AVAILABILITY AND IMPLEMENTATION: https://glyconnect.expasy.org/glycoql/.

Glycoinformatics in the Artificial Intelligence Era.

Artificial intelligence (AI) methods have been and are now being increasingly integrated in prediction software implemented in bioinformatics and its glycoscience branch known as glycoinformatics. AI techniques have evolved in the past decades, and their applications in glycoscience are not yet widespread. This limited use is partly explained by the peculiarities of glyco-data that are notoriously hard to produce and analyze. Nonetheless, as time goes, the accumulation of glycomics, glycoproteomics, and glycan-binding data has reached a point where even the most recent deep learning methods can provide predictors with good performance. We discuss the historical development of the application of various AI methods in the broader field of glycoinformatics. A particular focus is placed on shining a light on challenges in glyco-data handling, contextualized by lessons learnt from related disciplines. Ending on the discussion of state-of-the-art deep learning approaches in glycoinformatics, we also envision the future of glycoinformatics, including development that need to occur in order to truly unleash the capabilities of glycoscience in the systems biology era.

In silico analysis of the human milk oligosaccharide glycome reveals key enzymes of their biosynthesis.

Human milk oligosaccharides (HMOs) form the third most abundant component of human milk and are known to convey several benefits to the neonate, including protection from viral and bacterial pathogens, training of the immune system, and influencing the gut microbiome. As HMO production during lactation is driven by enzymes that are common to other glycosylation processes, we adapted a model of mucin-type GalNAc-linked glycosylation enzymes to act on free lactose. We identified a subset of 11 enzyme activities that can account for 206 of 226 distinct HMOs isolated from human milk and constructed a biosynthetic reaction network that identifies 5 new core HMO structures. A comparison of monosaccharide compositions demonstrated that the model was able to discriminate between two possible groups of intermediates between major subnetworks, and to assign possible structures to several previously uncharacterised HMOs. The effect of enzyme knockouts is presented, identifying ß-1,4-galactosyltransferase and ß-1,3-N-acetylglucosaminyltransferase as key enzyme activities involved in the generation of the observed HMO glycosylation patterns. The model also provides a synthesis chassis for the most common HMOs found in lactating mothers.

Bioinformatics Resources for the Study of Glycan-Mediated Protein Interactions.

The Glyco@Expasy initiative was launched as a collection of interdependent databases and tools spanning several aspects of knowledge in glycobiology. In particular, it aims at highlighting interactions between glycoproteins (such as cell surface receptors) and carbohydrate-binding proteins mediated by glycans. Here, major resources of the collection are introduced through two illustrative examples centered on the N-glycome of the human Prostate Specific Antigen (PSA) and the O-glycome of human serum proteins. Through different database queries and with the help of visualization tools, this article shows how to explore and compare content in a continuum to gather and correlate otherwise scattered pieces of information. Collected data are destined to feed more elaborate scenarios of glycan function. Glycoinformatics introduced here is, therefore, proposed as a means to either strengthen, shape, or refute assumptions on the specificity of a protein glycome in a given context.

LectinOracle: A Generalizable Deep Learning Model for Lectin-Glycan Binding Prediction.

Ranging from bacterial cell adhesion over viral cell entry to human innate immunity, glycan-binding proteins or lectins are abound in nature. Widely used as staining and characterization reagents in cell biology and crucial for understanding the interactions in biological systems, lectins are a focal point of study in glycobiology. Yet the sheer breadth and depth of specificity for diverse oligosaccharide motifs has made studying lectins a largely piecemeal approach, with few options to generalize. Here, LectinOracle, a model combining transformer-based representations for proteins and graph convolutional neural networks for glycans to predict their interaction, is presented. Using a curated data set of 564,647 unique protein-glycan interactions, it is shown that LectinOracle predictions agree with literature-annotated specificities for a wide range of lectins. Using a range of specialized glycan arrays, it is shown that LectinOracle predictions generalize to new glycans and lectins, with qualitative and quantitative agreement with experimental data. It is further demonstrated that LectinOracle can be used to improve lectin classification, accelerate lectin directed evolution, predict epidemiological outcomes in the context of influenza virus, and analyze whole lectomes in host-microbe interactions. It is envisioned that the herein presented platform will advance both the study of lectins and their role in (glyco)biology.

An Interactive View of Glycosylation.

The present chapter focuses on the interactive and explorative aspects of bioinformatics resources that have been recently released in glycobiology. The comparative analysis of data in a field where knowledge is scattered, incomplete, and disconnected from main biology requires efficient visualization, integration, and interactive tools that are currently only partially implemented. This overview highlights converging efforts toward building a consistent picture of protein glycosylation.

UniLectin, A One-Stop-Shop to Explore and Study Carbohydrate-Binding Proteins.

All eukaryotic cells are covered with a dense layer of glycoconjugates, and the cell walls of bacteria are made of various polysaccharides, putting glycans in key locations for mediating protein-protein interactions at cell interfaces. Glycan function is therefore mainly defined as binding to other molecules, and lectins are proteins that specifically recognize and interact non-covalently with glycans. UniLectin was designed based on insight into the knowledge of lectins, their classification, and their biological role. This modular platform provides a curated and periodically updated classification of lectins along with a set of comparative and visualization tools, as well as structured results of screening comprehensive sequence datasets. UniLectin can be used to explore lectins, find precise information on glycan-protein interactions, and mine the results of predictive tools based on HMM profiles. This usage is illustrated here with two protocols. The first one highlights the fine-tuned role of the O blood group antigen in distinctive pathogen recognition, while the second compares the various bacterial lectin arsenals that clearly depend on living conditions of species even in the same genus. © 2021 The Authors. Current Protocols published by Wiley Periodicals LLC. Basic Protocol 1: Searching for the structural details of lectins binding the O blood group antigen Basic Protocol 2: Comparing the lectomes of related organisms in different environments.

Bioinformatics in Immunoglobulin Glycosylation Analysis.

Analytical methods developed for studying immunoglobulin glycosylation rely heavily on software tailored for this purpose. Many of these tools are now used in high-throughput settings, especially for the glycomic characterization of IgG. A collection of these tools, and the databases they rely on, are presented in this chapter. Specific applications are detailed in examples of immunoglobulin glycomics and glycoproteomics data processing workflows. The results obtained in the glycoproteomics workflow are emphasized with the use of dedicated visualizing tools. These tools enable the user to highlight glycan properties and their differential expression.

Glycan Compositions with GlyConnect Compozitor to Enhance Glycopeptide Identification.

Glycoproteomics is unquestionably on the rise and its current development benefits from past experience in proteomics, in particular when attending to bioinformatics needs. An extensive range of software solutions is available, but the reproducibility of mass spectrometry data processing remains challenging. One of the key issues in running automated glycopeptide identification software is the selection of a reference glycan composition file. The default choices are often too broad, and a fastidious literature search to properly target this selection can be avoided. This chapter suggests the use of GlyConnect Compozitor to collect relevant information on glycosylation in a given tissue or cell line and shape an appropriate glycan composition set that can be input in the majority of search engines accommodating user-defined compositions.