Using protein hidden Markov models to probe gut metagenome-assembled genomes, 42 mucinases were identified as significantly enriched in colorectal cancer patients compared to healthy subjects, with two novel mucinases (WL42 and LLN1) experimentally validated to cleave glycosylated MUC1 and MUC2 substrates.
Key Findings
Results
A total of 1,869 mucinases were detected across gut metagenome-assembled genomes from CRC patients and healthy subjects, widely distributed across both cohorts.
Protein hidden Markov models were constructed based on 11 known mucinases and used to systematically identify mucinase sequences.
Metagenome-assembled genomes were derived from 80 colorectal cancer (CRC) patients and 86 healthy subjects.
The majority of detected mucinases originated from Bacteroides, Phocaeicola, and Akkermansia species.
Results
42 mucinases differed significantly in abundance between CRC patients and healthy subjects, and all 42 were enriched in CRC patients.
Further analysis of the 1,869 detected mucinases identified 42 that showed statistically significant differences in abundance between the two groups.
All 42 differentially abundant mucinases were enriched in CRC patients rather than healthy subjects.
These findings suggest potential involvement of these mucinases in CRC pathogenesis.
Results
The taxonomic origins of enriched mucinases differed between CRC patients and healthy subjects, with notable differences in Akkermansia contribution.
In CRC patients, the 42 enriched mucinases were primarily derived from Bacteroides (36.0%), Phocaeicola (30.6%), Akkermansia (8.8%), Alistipes (8.6%), and Escherichia (6.4%).
In healthy subjects, mucinases mainly originated from Bacteroides (26.1%), Akkermansia (22.7%), and Phocaeicola (20.3%).
Akkermansia contributed a notably higher proportion of mucinases in healthy subjects (22.7%) compared to CRC patients (8.8%).
Results
Two mucinases, WL42 and LLN1, exhibited significantly higher abundance levels compared to the other 40 enriched mucinases in CRC patients.
WL42 and LLN1 were identified as the most abundant among the 42 differentially enriched mucinases.
Phylogenetic and predicted structural analyses indicated that WL42 belongs to the M60 metalloprotease family.
LLN1 was predicted to belong to the M98 family based on phylogenetic and structural analyses.
Results
Functional validation demonstrated that both WL42 and LLN1 can cleave glycosylated MUC1 and MUC2 substrates but not the corresponding non-glycosylated proteins.
Co-incubation experiments were used to functionally validate the mucinase activity of WL42 and LLN1.
Both enzymes cleaved glycosylated forms of MUC1 and MUC2, confirming mucinase activity.
Neither enzyme cleaved the non-glycosylated versions of MUC1 or MUC2, indicating glycan-dependent substrate recognition.
These results confirm the feasibility of discovering novel mucinases directly from gut metagenomic data.
Methods
A bioinformatics pipeline using protein hidden Markov models was established as a feasible method for systematically identifying microbial mucinases within the gut microbiome.
Hidden Markov models were built based on 11 known mucinase sequences and applied to gut metagenome-assembled genomes.
The pipeline enabled large-scale mining of functionally active mucin-degrading enzymes from metagenomic data.
The authors describe this as providing 'a methodological foundation for large-scale mining of functionally active mucin-degrading enzymes.'
What This Means
This research suggests that bacteria living in the gut of colorectal cancer patients produce higher levels of mucin-degrading enzymes (mucinases) compared to healthy individuals. The researchers developed a computational approach using mathematical models of known mucinase proteins to scan large collections of genetic data from gut bacteria. By analyzing gut microbiome data from 80 colorectal cancer patients and 86 healthy people, they identified 1,869 mucinase sequences in total, and found that 42 of these enzymes were significantly more abundant in cancer patients — with all 42 enriched in the cancer group rather than the healthy group.
Two of these enzymes, named WL42 and LLN1, stood out as particularly abundant in colorectal cancer patients. Laboratory experiments confirmed that both enzymes can break down MUC1 and MUC2, two important proteins that make up the protective mucus layer lining the gut — but only when these proteins carry their normal sugar attachments, not in their bare protein form. This suggests these enzymes are specifically tuned to degrade real mucus as it naturally exists in the body. The research also found that while bacteria from the genera Bacteroides, Phocaeicola, and Akkermansia are the main producers of mucinases in both groups, the cancer patients had a notably lower proportion coming from Akkermansia.
This research matters because the mucus layer lining the colon acts as a critical barrier protecting the gut wall from harmful bacteria and other threats. This study suggests that in colorectal cancer patients, an imbalance in gut bacteria leads to elevated production of mucus-degrading enzymes, which could weaken this protective barrier and potentially contribute to cancer development or progression. Maintaining the proper balance of these mucinase-producing bacteria might be important for gut health, and the newly validated enzymes WL42 and LLN1 could serve as targets for future research into colorectal cancer prevention or treatment.
Li Y, Zhang H, Xiang B, Zhang Y, Zhang M. (2026). Enhanced microbiota-derived mucinases in colorectal cancer patients revealed by gut metagenome probing coupled with functional validation.. Applied and environmental microbiology. https://doi.org/10.1128/aem.01903-25