Deciphering gp120 sequence variation and structural dynamics in HIV neutralization phenotype by molecular dynamics simulations and graph machine learning

Mining channel-regulated peptides from animal venom by integrating sequence semantics and structural information

Channel-regulated peptides (CRPs) derived from animal venom hold great promise as potential drug candidates for numerous diseases associated with channel proteins. However, discovering and identifying CRPs using traditional bio-experimental methods is a time-consuming and laborious process. While there were a few computational studies on CRPs, they were limited to specific channel proteins, relied heavily on complex feature engineering, and lacked the incorporation of multi-source information. To address these problems, we proposed a novel deep learning model, called DeepCRPs, based on graph neural networks for systematically mining CRPs from animal venom. By combining the sequence semantic and structural information, the classification performance of four CRPs was significantly enhanced, reaching an accuracy of 0.92. This performance surpassed baseline models with accuracies ranging from 0.77 to 0.89. Furthermore, we employed advanced interpretable techniques to explore sequence and structural determinants relevant to the classification of CRPs, yielding potentially valuable bio-function interpretations. Comprehensive experimental results demonstrated the precision and interpretive capability of DeepCRPs, making it an accurate and bio-explainable suit for the identification and categorization of CRPs. Our research will contribute to the discovery and development of toxin peptides targeting channel proteins. The source data and code are freely available at https://github.com/liyigerry/DeepCRPs.

Molecular Dynamics Simulations Reveal the Conformational Transition of GH33 Sialidases

Sialidases are increasingly used in the production of sialyloligosaccharides, a significant component of human milk oligosaccharides. Elucidating the catalytic mechanism of sialidases is critical for the rational design of better biocatalysts, thereby facilitating the industrial production of sialyloligosaccharides. Through comparative all-atom molecular dynamics simulations, we investigated the structural dynamics of sialidases in Glycoside Hydrolase family 33 (GH33). Interestingly, several sialidases displayed significant conformational transition and formed a new cleft in the simulations. The new cleft was adjacent to the innate active site of the enzyme, which serves to accommodate the glycosyl acceptor. Furthermore, the residues involved in the specific interactions with the substrate were evolutionarily conserved in the whole GH33 family, highlighting their key roles in the catalysis of GH33 sialidases. Our results enriched the catalytic mechanism of GH33 sialidases, with potential implications in the rational design of sialidases.

Subtle Longitudinal Alterations in Env Sequence Potentiate Differences in Sensitivity to Broadly Neutralizing Antibodies following Acute HIV-1 Subtype C Infection

Broadly neutralizing antibodies (bNAbs) for HIV-1 prevention or cure strategies must inhibit transmitted/founder and reservoir viruses. Establishing sensitivity of circulating viruses to bNAbs and genetic patterns affecting neutralization variability may guide rational bNAbs selection for clinical development. We analyzed 326 single env genomes from nine individuals followed longitudinally following acute HIV-1 infection, with samples collected at ~1 week after the first detection of plasma viremia; 300 to 1,709 days postinfection but prior to initiating antiretroviral therapy (ART) (median = 724 days); and ~1 year post ART initiation. Sequences were assessed for phylogenetic relatedness, potential N- and O-linked glycosylation, and variable loop lengths (V1 to V5). A total of 43 env amplicons (median = 3 per patient per time point) were cloned into an expression vector and the TZM-bl assay was used to assess the neutralization profiles of 15 bNAbs targeting the CD4 binding site, V1/V2 region, V3 supersite, MPER, gp120/gp41 interface, and fusion peptide. At 1 μg/mL, the neutralization breadths were as follows: VRC07-LS and N6.LS (100%), VRC01 (86%), PGT151 (81%), 10-1074 and PGT121 (80%), and less than 70% for 10E8, 3BNC117, CAP256.VRC26, 4E10, PGDM1400, and N123-VRC34.01. Features associated with low sensitivity to V1/V2 and V3 bNAbs were higher potential glycosylation sites and/or relatively longer V1 and V4 domains, including known "signature" mutations. The study shows significant variability in the breadth and potency of bNAbs against circulating HIV-1 subtype C envelopes. VRC07-LS, N6.LS, VRC01, PGT151, 10-1074, and PGT121 display broad activity against subtype C variants, and major determinants of sensitivity to most bNAbs were within the V1/V4 domains. IMPORTANCE Broadly neutralizing antibodies (bNAbs) have potential clinical utility in HIV-1 prevention and cure strategies. However, bNAbs target diverse epitopes on the HIV-1 envelope and the virus may evolve to evade immune responses. It is therefore important to identify antibodies with broad activity in high prevalence settings, as well as the genetic patterns that may lead to neutralization escape. We investigated 15 bNAbs with diverse biophysical properties that target six epitopes of the HIV-1 Env glycoprotein for their ability to inhibit viruses that initiated infection, viruses circulating in plasma at chronic infection before antiretroviral treatment (ART), or viruses that were archived in the reservoir during ART in subtype C infected individuals in South Africa, a high burden country. We identify the antibodies most likely to be effective for clinical use in this setting and describe mutational patterns associated with neutralization escape from these antibodies.

The applications of machine learning in HIV neutralizing antibodies research-A systematic review

Machine learning algorithms play an essential role in bioinformatics and allow exploring the vast and noisy biological data in unrivaled ways. This paper is a systematic review of the applications of machine learning in the study of HIV neutralizing antibodies. This significant and vast research domain can pave the way to novel treatments and to a vaccine. We selected the relevant papers by investigating the available literature from the Web of Science and PubMed databases in the last decade. The computational methods are applied in neutralization potency prediction, neutralization span prediction against multiple viral strains, antibody-virus binding sites detection, enhanced antibodies design, and the study of the antibody-induced immune response. These methods are viewed from multiple angles spanning data processing, model description, feature selection, evaluation, and sometimes paper comparisons. The algorithms are diverse and include supervised, unsupervised, and generative types. Both classical machine learning and modern deep learning were taken into account. The review ends with our ideas regarding future research directions and challenges.