Designing Novel Aspartic Protease (NAP) Using Ancestral Sequence Reconstruction (ASR) for Cheesemaking.

Rhizomucor miehei protease (RMP), a chymosin-like fungal enzyme, is among the most widely used non-animal milk coagulant enzymes (MCEs) in cheese production. However, like other current MCEs, RMP exhibits non-specific proteolytic activity and high thermostability, leading to excessive casein degradation and negatively impacting cheese yield, texture, and flavor. To address these limitations, we employed computational enzyme engineering to design a novel aspartic protease (NAP) with high κ-casein affinity and reduced thermal stability, using ancestral sequence reconstruction (ASR). ASR-aided modifications targeted amino acid residues were introduced adjacent to the catalytic and flap regions, while conserving the DTGS and DTGT catalytic motifs. Specifically, region 1 (LLFDTGSSDTWV) was modified to VLFDTGSSNLWV; region 2 (FTIDTGTNFFIM) to AIVDTGTSLLYG; and the flap sequence (YGTGGAN) to YGTGSAT. These changes enhanced substrate accommodation, structural flexibility, and intermolecular interactions. Molecular docking, molecular dynamics (MD) simulations, and structural analyses were performed to compare NAP with the wild-type RMP. The NAP-κ-casein complex showed improved substrate interaction, reflected in a more favorable HADDOCK score (- 188.7 ± 7.4) versus the wild type (- 126 ± 7.5). NAP also formed new hydrogen bonds and salt bridges, particularly involving Asp38, Asp237, and Gln81, not present in the wild type. MM-GBSA analysis indicated a stronger binding affinity for NAP at 40 °C (- 88.87 kcal/mol) compared to RMP at 45 °C (- 33.9 kcal/mol), suggesting higher catalytic efficiency under milder conditions. Structural metrics (SASA, Rg, RMSD, RMSF), along with free energy landscape (FEL) and principal component analysis (PCA), confirmed reduced thermostability and enhanced flexibility in NAP. These findings highlight NAP as a promising milk coagulant with improved specificity and reduced off-target proteolysis. Moreover, they underscore the potential of ASR-driven computational design in developing next-generation coagulants tailored for industrial dairy applications, pending experimental validation.