Structure-based neural network protein-carbohydrate interaction predictions at the residue level

Carbohydrates dynamically and transiently interact with proteins for cell-cell recognition, cellular differentiation, immune response, and many other cellular processes. Despite the molecular importance of these interactions, there are currently few reliable computational tools to predict potential carbohydrate-binding sites on any given protein. Here, we present two deep learning (DL) models named CArbohydrate-Protein interaction Site IdentiFier (CAPSIF) that predicts non-covalent carbohydrate-binding sites on proteins: (1) a 3D-UNet voxel-based neural network model (CAPSIF:V) and (2) an equivariant graph neural network model (CAPSIF:G). While both models outperform previous surrogate methods used for carbohydrate-binding site prediction, CAPSIF:V performs better than CAPSIF:G, achieving test Dice scores of 0.597 and 0.543 and test set Matthews correlation coefficients (MCCs) of 0.599 and 0.538, respectively. We further tested CAPSIF:V on AlphaFold2-predicted protein structures. CAPSIF:V performed equivalently on both experimentally determined structures and AlphaFold2-predicted structures. Finally, we demonstrate how CAPSIF models can be used in conjunction with local glycan-docking protocols, such as GlycanDock, to predict bound protein-carbohydrate structures.

Natural variations in AAVHSC16 significantly reduce liver tropism and maintain broad distribution to periphery and CNS

Adeno-associated viruses derived from human hematopoietic stem cells (AAVHSCs) are naturally occurring AAVs. Fifteen AAVHSCs have demonstrated broad biodistribution while displaying differences in transduction. We examine the structure-function relationships of these natural amino acid variations on cellular binding. We demonstrate that AAVHSC16 is the only AAVHSC that does not preferentially bind to terminal galactose. AAVHSC16 contains two unique amino acids, 501I and 706C, compared with other AAVHSCs. Through mutagenesis, we determined that residue 501 contributes to the lack of galactose binding. Structural analysis revealed that residue 501 is in proximity to the galactose binding pocket, hence confirming its functional role in galactose binding. Biodistribution analysis of AAVHSC16 indicated significantly less liver tropism in mice and non-human primates compared with other clade F members, likely associated with overall binding differences observed in vitro. AAVHSC16 maintained robust tropism to other key tissues in the peripheral and central nervous systems after intravenous injection, including to the brain, heart, and gastrocnemius. Importantly, AAVHSC16 did not induce elevated liver enzyme levels in non-human primates after intravenous injection at high doses. The unique glycan binding and tropism of AAVHSC16 makes this naturally occurring capsid an attractive candidate for therapies requiring less liver tropism while maintaining broad biodistribution.

Identification of the Glycan Binding Profile of Human and Rodent Plasmodium Sporozoites

The transmission of Plasmodium spp. sporozoites to the mammalian host is the first step in the initiation of the mosquito-borne disease known as malaria. The exact route of transmission from the bloodstream to the liver is still not clearly elucidated, and identification of the host glycan structures bound by the sporozoites may inform as to which host cells are involved. Here, we provide a comprehensive analysis of the glycan structures that sporozoites from the human pathogen, P. falciparum, and the rodent pathogen, P. yoelii, recognize and bind. Glycan array analysis was used to profile the glycans bound by the sporozoites, and the binding affinities of these sporozoite-glycan interactions were then determined by surface plasmon resonance. Data showed that the different Plasmodium spp. bind different classes of glycans. P. falciparum was observed to bind to glycans with terminal N-acetylgalactosamine (GalNAc) or Galactose (Gal) linked to a GalNAc, and the highest-affinity observed was with the GalNAc monosaccharide (12.5 nM). P. yoelii bound glycosaminoglycans, mannosyl glycans, Gal linked to N-acetylglucosamine structures, and the αGal epitope. The highest-affinity interaction for P. yoelii was with the αGal epitope (31.4 nM). This is the first study to identify the key host glycan structures recognized by human and rodent Plasmodium spp. sporozoites. An understanding of how Plasmodium sporozoites interact with the specific glycan structures identified here may provide further insight into this infectious disease that could help direct the design of an effective therapeutic.

Mechanism of Ganglioside Receptor Recognition by Botulinum Neurotoxin Serotype E

The botulinum neurotoxins are potent molecules that are not only responsible for the lethal paralytic disease botulism, but have also been harnessed for therapeutic uses in the treatment of an increasing number of chronic neurological and neuromuscular disorders, in addition to cosmetic applications. The toxins act at the cholinergic nerve terminals thanks to an efficient and specific mechanism of cell recognition which is based on a dual receptor system that involves gangliosides and protein receptors. Binding to surface-anchored gangliosides is the first essential step in this process. Here, we determined the X-ray crystal structure of the binding domain of BoNT/E, a toxin of clinical interest, in complex with its GD1a oligosaccharide receptor. Beyond confirmation of the conserved ganglioside binding site, we identified key interacting residues that are unique to BoNT/E and a significant rearrangement of loop 1228–1237 upon carbohydrate binding. These observations were also supported by thermodynamic measurements of the binding reaction and assessment of ganglioside selectivity by immobilised-receptor binding assays. These results provide a structural basis to understand the specificity of BoNT/E for complex gangliosides.

Tools for generating and analyzing glycan microarray data

Glycans are one of the major biological polymers found in the mammalian body. They play a vital role in a number of physiologic and pathologic conditions. Glycan microarrays allow a plethora of information to be obtained on protein–glycan binding interactions. In this review, we describe the intricacies of the generation of glycan microarray data and the experimental methods for studying binding. We highlight the importance of this knowledge before moving on to the data analysis. We then highlight a number of tools for the analysis of glycan microarray data such as data repositories, data visualization and manual analysis tools, automated analysis tools and structural informatics tools.

Structural and Functional Studies of a Klebsiella Phage Capsule Depolymerase Tailspike: Mechanistic Insights into Capsular Degradation

Capsule polysaccharide is a major virulence factor of Klebsiella pneumoniae, a nosocomial pathogen associated with a wide range of infections. It protects bacteria from harsh environmental conditions, immune system response, and phage infection. To access cell wall-located receptors, some phages possess tailspike depolymerases that degrade the capsular polysaccharide. Here, we present the crystal structure of a tailspike against Klebsiella, KP32gp38, whose primary sequence shares no similarity to other proteins of known structure. In the trimeric structure of KP32gp38, each chain contains a flexible N-terminal domain, a right-handed parallel β helix domain and two β sandwiches with carbohydrate binding features. The crystal structure and activity assays allowed us to locate the catalytic site. Also, our data provide experimental evidence of a branching architecture of depolymerases in KP32 Klebsiella viruses, as KP32gp38 displays nanomolar affinity to another depolymerase from the same phage, KP32gp37. Results provide a structural framework for enzyme engineering to produce serotype-broad-active enzyme complexes against K. pneumoniae.

The Interplay between the Host Receptor and Influenza Virus Hemagglutinin and Neuraminidase

The hemagglutinin (HA) and neuraminidase (NA) glycoproteins of influenza A virus are responsible for the surface interactions of the virion with the host. Entry of the virus is mediated by functions of the HA: binding to cellular receptors and facilitating fusion of the virion membrane with the endosomal membrane. The HA structure contains receptor binding sites in the globular membrane distal head domains of the trimer, and the fusion machinery resides in the stem region. These sites have specific characteristics associated with subtype and host, and the differences often define species barriers. For example, avian viruses preferentially recognize α2,3-Sialic acid terminating glycans as receptors and mammalian viruses recognize α2,6-Sialic acid. The neuraminidase, or the receptor-destroying protein, cleaves the sialic acid from cellular membrane constituents and viral glycoproteins allowing for egress of nascent virions. A functional balance of activity has been demonstrated between the two glycoproteins, resulting in an optimum level of HA affinity and NA enzymatic cleavage to allow for productive infection. As more is understood about both HA and NA, the relevance for functional balance between HA and NA continues to expand, with potential implications for interspecies transmission, host adaptation, and pathogenicity.

The cholesterol-dependent cytolysins pneumolysin and streptolysin O require binding to red blood cell glycans for hemolytic activity

The cholesterol-dependent cytolysin (CDC) pneumolysin (Ply) is a key virulence factor of Streptococcus pneumoniae. Membrane cholesterol is required for the cytolytic activity of this toxin, but it is not clear whether cholesterol is the only cellular receptor. Analysis of Ply binding to a glycan microarray revealed that Ply has lectin activity and binds glycans, including the Lewis histo-blood group antigens. Surface plasmon resonance analysis showed that Ply has the highest affinity for the sialyl LewisX (sLeX) structure, with a K(d) of 1.88 × 10(-5) M. Ply hemolytic activity against human RBCs showed dose-dependent inhibition by sLeX. Flow cytometric analysis and Western blots showed that blocking binding of Ply to the sLeX glycolipid on RBCs prevents deposition of the toxin in the membrane. The lectin domain responsible for sLeX binding is in domain 4 of Ply, which contains candidate carbohydrate-binding sites. Mutagenesis of these predicted carbohydrate-binding residues of Ply resulted in a decrease in hemolytic activity and a reduced affinity for sLeX. This study reveals that this archetypal CDC requires interaction with the sLeX glycolipid cellular receptor as an essential step before membrane insertion. A similar analysis conducted on streptolysin O from Streptococcus pyogenes revealed that this CDC also has glycan-binding properties and that hemolytic activity against RBCs can be blocked with the glycan lacto-N-neotetraose by inhibiting binding to the cell surface. Together, these data support the emerging paradigm shift that pore-forming toxins, including CDCs, have cellular receptors other than cholesterol that define target cell tropism.

Carbohydrate-binding protein identification by coupling structural similarity searching with binding affinity prediction

Carbohydrate-binding proteins (CBPs) are potential biomarkers and drug targets. However, the interactions between carbohydrates and proteins are challenging to study experimentally and computationally because of their low binding affinity, high flexibility, and the lack of a linear sequence in carbohydrates as exists in RNA, DNA, and proteins. Here, we describe a structure-based function-prediction technique called SPOT-Struc that identifies carbohydrate-recognizing proteins and their binding amino acid residues by structural alignment program SPalign and binding affinity scoring according to a knowledge-based statistical potential based on the distance-scaled finite-ideal gas reference state (DFIRE). The leave-one-out cross-validation of the method on 113 carbohydrate-binding domains and 3442 noncarbohydrate binding proteins yields a Matthews correlation coefficient of 0.56 for SPalign alone and 0.63 for SPOT-Struc (SPalign + binding affinity scoring) for CBP prediction. SPOT-Struc is a technique with high positive predictive value (79% correct predictions in all positive CBP predictions) with a reasonable sensitivity (52% positive predictions in all CBPs). The sensitivity of the method was changed slightly when applied to 31 APO (unbound) structures found in the protein databank (14/31 for APO versus 15/31 for HOLO). The result of SPOT-Struc will not change significantly if highly homologous templates were used. SPOT-Struc predicted 19 out of 2076 structural genome targets as CBPs. In particular, one uncharacterized protein in Bacillus subtilis (1oq1A) was matched to galectin-9 from Mus musculus. Thus, SPOT-Struc is useful for uncovering novel carbohydrate-binding proteins. SPOT-Struc is available at http://sparks-lab.org.

Rapid assays for lectin toxicity and binding changes that reflect altered glycosylation in mammalian cells

Glycosylation engineering is used to generate glycoproteins, glycolipids, or proteoglycans with a more defined complement of glycans on their glycoconjugates. For example, a mammalian cell glycosylation mutant lacking a specific glycosyltransferase generates glycoproteins, and/or glycolipids, and/or proteoglycans with truncated glycans missing the sugar transferred by that glycosyltransferase, as well as those sugars that would be added subsequently. In some cases, an alternative glycosyltransferase may then use the truncated glycans as acceptors, thereby generating a new or different glycan subset in the mutant cell. Another type of glycosylation mutant arises from gain-of-function mutations that, for example, activate a silent glycosyltransferase gene. In this case, glycoconjugates will have glycans with additional sugar(s) that are more elaborate than the glycans of wild type cells. Mutations in other genes that affect glycosylation, such as nucleotide sugar synthases or transporters, will alter the glycan complement in more general ways that usually affect several types of glycoconjugates. There are now many strategies for generating a precise mutation in a glycosylation gene in a mammalian cell. Large-volume cultures of mammalian cells may also generate spontaneous mutants in glycosylation pathways. This article will focus on how to rapidly characterize mammalian cells with an altered glycosylation activity. The key reagents for the protocols described are plant lectins that bind mammalian glycans with varying avidities, depending on the specific structure of those glycans. Cells with altered glycosylation generally become resistant or hypersensitive to lectin toxicity, and have reduced or increased lectin or antibody binding. Here we describe rapid assays to compare the cytotoxicity of lectins in a lectin resistance test, and the binding of lectins or antibodies by flow cytometry in a glycan-binding assay. Based on these tests, glycosylation changes expressed by a cell can be revealed, and glycosylation mutants classified into phenotypic groups that may reflect a loss-of-function or gain-of-function mutation in a specific gene involved in glycan synthesis.

Characterizing carbohydrate-protein interactions by nuclear magnetic resonance spectroscopy

Interactions between proteins and soluble carbohydrates and/or surface displayed glycans are central to countless recognition, attachment and signaling events in biology. The physical chemical features associated with these binding events vary considerably, depending on the biological system of interest. For example, carbohydrate-protein interactions can be stoichiometric or multivalent, the protein receptors can be monomeric or oligomeric, and the specificity of recognition can be highly stringent or rather promiscuous. Equilibrium dissociation constants for carbohydrate binding are known to vary from micromolar to millimolar, with weak interactions being far more prevalent; and individual carbohydrate-binding sites can be truly symmetrical or merely homologous, and hence, the affinities of individual sites within a single protein can vary, as can the order of binding. Several factors, including the weak affinities with which glycans bind their protein receptors, the dynamic nature of the glycans themselves, and the nonequivalent interactions among oligomeric carbohydrate receptors, have made nuclear magnetic resonance (NMR) an especially powerful tool for studying and defining carbohydrate-protein interactions. Here, we describe those NMR approaches that have proven to be the most robust in characterizing these systems, and explain what type of information can (or cannot) be obtained from each. Our goal is to provide the reader the information necessary for selecting the correct experiment or sets of experiments to characterize their carbohydrate-protein interaction of interest.

Molecular basis of the structure and function of H1 hemagglutinin of influenza virus

Influenza virus hemagglutinin (HA) contains antigenic sites recognized by the host immune system, cleavage sites cleaved by host proteases, receptor binding sites attaching to sialyl receptors on the target cell, and fusion peptides mediating membrane fusion. Change in an amino acid(s) in these sites may affect the potential of virus infection and spread within and between hosts. Influenza viruses with H1 HA infect birds, pigs and humans and have caused two of the four pandemics in the past 100 years: 1918 pandemic that killed 21-50 million people and 2009 pandemic that caused more than 18,000 deaths. Understanding the relationship between antigenic structure and immune specificity, the receptor binding specificity in virus transmission, how the cleavage site controls pathogenicity, and how the fusion peptide causes membrane fusion for the entry of influenza virus into the host cell should provide information to find more effective ways to prevent and control influenza.