Skip to main navigation Skip to search Skip to main content

Structural and immunogenic evaluation of silk proteins from Bombyx mori using advanced bioinformatics and deep learning for biomaterials applications

Research output: Contribution to scientific journalArticle in an indexed scientific journalpeer-review

Abstract

Silk from Bombyx mori is a premier biomaterial. Yet, comprehensive structural and immunoinformatic characterization of its protein components—fibroin subunits (FibH, FibL, P25) and five sericin isoforms—remains incomplete, hindering rational design of biocompatible medical devices. We integrated AlphaFold3 structural prediction with multi-algorithm immunoinformatic profiling (VaxiJen, NetMHCpan, BepiPred, AllerTOP, ToxinPred) to establish structure-immunogenicity relationships across the silk proteome. Structural modeling revealed that FibL and P25 adopt well-defined architectures (pTM = 0.85), whereas FibH exhibited low confidence (ipTM/pTM = 0.28), reflecting its intrinsically disordered pre-assembly state. Sericins displayed predominantly disordered conformations that undergo partial ordering upon complexing with FibL-P25-Cu2+, supporting a disorder-to-order templating mechanism for fiber assembly. Immunoinformatic analysis revealed striking antigenic heterogeneity: P25 emerged as uniquely hypoimmunogenic, with subthreshold antigenicity (VaxiJen: 0.395), a single strong MHC-I binder for a single HLA allele, and minimal MHC-II reactivity. Conversely, FibL and FibH showed the highest potential for inducing CD8+ and CD4+ cell responses among the fibroin subunits, respectively. Ser-4 and Ser-1 exhibited a broad MHC coverage, presenting ≥ 10 strong binders in 81% and 30% of MHC-I alleles, respectively, as well as ≥ 4 high-priority peptides across all 27 and 6 tested MHC-II alleles, respectively. Allergenicity prediction classified FibH, FibL, Ser-2, and Ser-5 as probable allergens. All proteins were non-toxic. These findings challenge the paradigm that sericins exclusively drive silk immunogenicity, revealing instead an HLA-dependent risk profile dominated by FibL, FibH, Ser-4, and Ser-1. This computational framework provides a rational foundation for engineering hypoimmunogenic silk variants through P25 enrichment, epitope deletion, or HLA-matched biomaterial selection.

Original languageEnglish
Article number100692
JournalJournal of Genetic Engineering and Biotechnology
Volume24
Issue number2
DOIs
StatePublished - Jun 2026

Bibliographical note

Publisher Copyright:
© 2026 The Author(s). Published by Elsevier Inc. on behalf of Academy of Scientific Research and Technology. This is an open access article under the CC BY-NC-ND license. http://creativecommons.org/licenses/by-nc-nd/4.0/

Keywords

  • Epitope mapping
  • Fibroin
  • Immunoinformatics profile
  • Protein structure prediction
  • Sericin
  • Silkworm
  • Supervised machine learning

Types Minciencias

  • Artículos de investigación con calidad A2 / Q2

Fingerprint

Dive into the research topics of 'Structural and immunogenic evaluation of silk proteins from Bombyx mori using advanced bioinformatics and deep learning for biomaterials applications'. Together they form a unique fingerprint.

Cite this