A Computational Framework for Synthetic Minimal Bacteriophage Gene Set Design: Application to Streptococcus mutans and Codon-Optimized Expression in Escherichia coli

Minimal gene set design, reducing a system to its smallest functional part set, is a core synthetic biology strategy. Bacteriophages are attractive but underused chassis, yet wild-type phages are difficult to engineer. We present a comparative genomics-guided computational framework for synthetic minimal phage gene set design, demonstrated in silico on Streptococcus mutans, the primary agent of dental caries. Three S. mutans phages (M102AD, φAPCM01, SMHBZ8) were compared by synteny analysis, and their 123 proteins were annotated with InterProScan, EggNOG-mapper, and HHpred. Conserved essential genes were selected by cross-phage conservation and annotation evidence, assembled as standardized transcriptional units, codon-optimized for E. coli K-12 with a modern folding-aware optimizer, and cloned in silico into the broad host range vector pBBR1MCS-2. Seventeen genes were assembled into a 16,403 bp synthetic insert; the final 21,559 bp construct raised the mean codon adaptation index (CAI) from 0.58 to 0.96, brought GC content into the optimal range, and eliminated rare codons, while removing all internal cloning sites. A 5′ mRNA-folding screen confirmed accessible ribosome binding sites for all 17 genes; a sensitivity analysis confirmed that the gene set was robust to the selection thresholds; and AlphaFold-Multimer modeling of representative inter-module interfaces is reported. Multi-host CAI analysis indicates strong codon usage compatibility with Enterobacteriaceae hosts. We emphasize that the construct is a computationally designed multi-gene expression plasmid, not a self-replicating or packageable phage genome, and that all results are in silico predictions requiring experimental validation. Gibson Assembly, not restriction cloning, is recommended. The design logic is transferable to other phage–host systems.