In-silico prioritization of pathogenic Interleukin-37b variants and a fusion platform for high-yield soluble production.

Abstract

Human interleukin-37 isoform 1 (IL-37b) is a key anti-inflammatory cytokine with significant therapeutic potential for inflammatory and immune-mediated disorders. However, its clinical translation is limited by poor understanding of disease-associated genetic variants and lack of an expression system for soluble production. While addressing both challenges, this study presents (a) a prioritized catalog of high-confidence, pathogenic IL-37b variants, and (b) a fusion-based expression platform for its soluble production, providing essential resources for future functional validations. Screening of over 3000 IL-37b variants using various computational tools and multi-algorithm consensus approach identified 25 potentially pathogenic non-synonymous single nucleotide variants (nsSNVs). Amongst these, 16 variants (e.g., D64 V/N, L72R, L111Q, V113F, C122R, F154S, I155 N, Y157C, E168G, G174A, I111T) were predicted to significantly destabilize IL-37b's structure and impair its interaction with the IL-18 receptor. Further, guided by complementary in-silico predictions, an aspartate-rich lunasin peptide yielded the highest soluble expression, constituting ∼40 % of total E. coli cellular proteins. The fusion expression system achieved ∼80 % solubility (compared to <10 % for wild-type IL-37b) and a yield of 167 mg/L following Ni²⁺-affinity purification under optimized conditions (25 °C, lactose autoinduction). The findings underscore the significance of complementary computational workflows in establishing an end-to-end pipeline for variant-to-solution analysis of IL-37b - a dual foundation linking in silico discovery to therapeutic development.