Deep-learning-based design of synthetic orthologs of SH3 signaling domains.

Evolution-based deep generative models represent an exciting direction in understanding and designing proteins. An open question is whether such models can learn specialized functional constraints that control fitness in specific biological contexts. Here, we examine the ability of generative models to produce synthetic versions of Src-homology 3 (SH3) domains that mediate signaling in the Sho1 osmotic stress response pathway of yeast. We show that a variational autoencoder (VAE) model produces artificial sequences that experimentally recapitulate the function of natural SH3 domains. More generally, the model organizes all fungal SH3 domains such that locality in the model latent space (but not simply locality in sequence space) enriches the design of synthetic orthologs and exposes non-obvious amino acid constraints distributed near and far from the SH3 ligand-binding site. The ability of generative models to design ortholog-like functions in vivo opens new avenues for engineering protein function in specific cellular contexts and environments.

Vanin-2 is expressed in peripheral blood T cells and upregulated in SLE patients.

Members of the vanin gene family include VNN1, VNN2 and VNN3 in humans. Although the functions of vanins have been widely examined in myeloid cells, their expression and functions have not been clarified in T lymphocytes. This study aimed to elucidate the significance of Vanin-2 (VNN2) on human peripheral blood T lymphocytes and study its expression in systemic lupus erythematosus (SLE). The differential expression of Vanins was analysed by bioinformatics. VNN2 expressions in peripheral blood T cell subsets were analysed by single-cell RNA sequencing data and flow cytometry. Changes of VNN2 expression before and after T cell activation were further clarified by western blot. The function of VNN2+ cells was studied by granzyme B and perforin detection. Changes in VNN2+ proportions in T cell subsets of SLE patients were further analysed. In the present study, only VNN2 among vanins showed distinguishable expression in T cells. VNN2+ percentages were higher in CD8+ T cells than in CD4+ T cells. VNN2+ T cells were with a higher memory T cell composition. VNN2 expression was significantly increased after T cell stimulation. VNN2+ T cells had higher levels of granzyme B and perforin secretion than VNN2- T cells. Clinically, VNN2+ percentages in T cells of SLE patients were upregulated. Together, these data suggested that VNN2 is expressed in peripheral blood T cells characterized more granzyme B and perforin secretion, and increased VNN2+ T cells in SLE patients could reflect altered T cell functions in vivo.

VNN2-expressing circulating monocytes exhibit unique functional characteristics and are decreased in patients with primary Sjögren's syndrome.

OBJECTIVE: This study aims to elucidate the significance of VNN2 expression in peripheral blood monocytes and its clinical relevance in primary Sjögren's syndrome (pSS). METHODS: We investigated VNN2 expression by analyzing single-cell RNA sequencing (scRNA-seq) data from peripheral blood mononuclear cells. Flow cytometry was used to detect and compare VNN2 expression in total monocytes, classical monocytes (cMo), intermediate monocytes (iMo) and non-classical monocytes (ncMo). Additionally, we examined the expression of HLA, ICAM1, CD62L, ITGAM, S100A8, S100A9, CCR2, CCR6, CX3CR1 and CXCR3 in VNN2+ and VNN2- cells. We analyzed the correlation between VNN2 expression and clinical indicators and assessed the clinical utility of VNN2+ monocytes in pSS diagnosis using receiver operating characteristic curves. RESULTS: We observed high VNN2 expression in monocytes, with significantly higher levels in CD14++ monocytes compared to ncMo. VNN2+ monocytes exhibited decreased expression of HLA and CD62L and increased expression of ICAM1, ITGAM, S100A8, S100A9, CCR2, CCR6, CX3CR1 and CXCR3 compared to VNN2- monocytes. Although scRNA-seq data showed that VNN2 mRNA was upregulated, cell surface expression of VNN2 was decreased in monocytes from pSS patients compared to healthy controls. The reduced levels of VNN2+ monocyte subpopulations in pSS patients were negatively correlated with anti-ribosome antibody levels and positively correlated with complement 4 levels. Detection of VNN2 expression in monocytes can aid in the auxiliary diagnosis of pSS. CONCLUSION: Monocytes expressing cell surface VNN2 are significantly reduced in pSS patients. This suggests a potential role for VNN2 in pSS development and its potential use as a diagnostic marker for pSS.

ProtWave-VAE: Integrating Autoregressive Sampling with Latent-Based Inference for Data-Driven Protein Design.

Deep generative models (DGMs) have shown great success in the understanding and data-driven design of proteins. Variational autoencoders (VAEs) are a popular DGM approach that can learn the correlated patterns of amino acid mutations within a multiple sequence alignment (MSA) of protein sequences and distill this information into a low-dimensional latent space to expose phylogenetic and functional relationships and guide generative protein design. Autoregressive (AR) models are another popular DGM approach that typically lacks a low-dimensional latent embedding but does not require training sequences to be aligned into an MSA and enable the design of variable length proteins. In this work, we propose ProtWave-VAE as a novel and lightweight DGM, employing an information maximizing VAE with a dilated convolution encoder and an autoregressive WaveNet decoder. This architecture blends the strengths of the VAE and AR paradigms in enabling training over unaligned sequence data and the conditional generative design of variable length sequences from an interpretable, low-dimensional learned latent space. We evaluated the model's ability to infer patterns and design rules within alignment-free homologous protein family sequences and to design novel synthetic proteins in four diverse protein families. We show that our model can infer meaningful functional and phylogenetic embeddings within latent spaces and make highly accurate predictions within semisupervised downstream fitness prediction tasks. In an application to the C-terminal SH3 domain in the Sho1 transmembrane osmosensing receptor in baker's yeast, we subject ProtWave-VAE-designed sequences to experimental gene synthesis and select-seq assays for the osmosensing function to show that the model enables synthetic protein design, conditional C-terminus diversification, and engineering of the osmosensing function into SH3 paralogues.