A hidden structure may explain how infant gut bacteria gets nutrients

By Dr. Liji Thomas, MDReviewed by Lauren HardakerNov 3 2026

A newly identified β-sheet structure helps an enzyme from infant-associated gut bacteria recognize and process sugar-bearing protein fragments, revealing a specialized mechanism that could shed some light on how microbes access nutrients in the developing infant gut.

Study: Structural basis for N-glycopeptide preference in an infant-derived Bifidobacterium bifidum PNGaseA. Image credit: Corona Borealis Studio/Shutterstock.com

A study published in Nature Communications has revealed how an enzyme from an infant-associated gut bacterium recognizes and removes sugars from short protein fragments. The findings identify a previously unobserved structural feature that helps explain the enzyme's preference for glycopeptides and may contribute to understanding how gut bacteria process host-derived glycans.

Understanding how gut bacteria access N-glycans

N-glycans are complex sugar structures attached to proteins, including proteins found in breast milk. These glycans represent potential substrates for bacteria in the infant gut, where specialized enzymes help release them from proteins.

Among these enzymes are peptide (PNGases), which cleave N-glycans from their protein or peptide attachments. However, much of the existing structural understanding of these enzymes comes from PNGaseF, leaving questions about how other PNGase families recognize glycosylated peptides.

The researchers investigated a PNGaseA-family enzyme, termed PNGaseABif, identified in two strains of Bifidobacterium bifidum, including the infant-associated strain LMG13195. The enzyme is predicted to be anchored to the bacterial cell surface, suggesting a possible role in processing extracellular glycopeptides.

A bacterial enzyme shows a strong preference for glycopeptides

The researchers tested PNGaseABif against glycoproteins containing different types of N-glycans, comparing native proteins, denatured proteins and smaller glycopeptides generated through proteolytic digestion.

PNGaseABif strongly preferred glycopeptides, showed limited activity against denatured glycoproteins and had no detectable activity against the native forms of the three principal glycoproteins under the conditions tested. By comparison, the control enzyme PNGaseL was active against all three substrate preparations.

The enzyme released several types of N-glycans, including high-mannose and complex structures. It also released core α1,3-fucosylated glycans from horseradish peroxidase but not from honeybee phospholipase A2, suggesting that the surrounding protein environment influences substrate accessibility and enzyme activity.

These findings indicated that PNGaseABif's substrate preference depends not only on glycan composition but also on how the glycan is presented on the attached protein or peptide.

A previously unseen β-sheet cradle explains peptide recognition

To investigate this preference, the researchers determined three crystal structures of PNGaseABif, including a catalytically impaired double mutant bound to a glycopeptide. This provided the first reported structure of a PNGase-superfamily enzyme in complex with a glycopeptide.

The structures revealed an additional ten-stranded β-sheet region, termed a "cradle", positioned alongside the enzyme's catalytic core. Together, these structures form a C-shaped pocket that can accommodate a short glycopeptide.

Unlike the more open catalytic cleft of PNGaseF, this pocket appears suited to positioning and partially enclosing short peptide substrates, potentially explaining PNGaseABif's preference for glycopeptides over larger glycoproteins.

The enzyme recognizes the N-glycan primarily through its two core N-acetylglucosamine residues and central mannose, while the attached peptide interacts with both the catalytic core and the cradle.

Particularly important is the threonine residue located two positions after the glycosylated asparagine in the NXT recognition sequence. Structural analysis identified extensive interactions involving this residue, suggesting that it helps position the glycopeptide for cleavage. However, the researchers examined only one glycopeptide sequence in the bound structure, leaving recognition of serine-containing NXS sequences unresolved.

Comparisons between the unbound and glycopeptide-bound structures suggested that the cradle moves toward the catalytic core when the substrate binds. Molecular dynamics simulations supported this proposed closure mechanism, although experimental validation of individual cradle contacts is still needed.

Catalytic mechanism and implications for the infant gut

The structures also provided clues about how PNGaseABif removes N-glycans. Two catalytic residues, Asp107 and Glu236, are positioned near the bond linking the glycan to the peptide.

The researchers proposed that Asp107 most likely activates a water molecule to facilitate cleavage, although direct nucleophilic attack by Asp107 remains an alternative possibility. The precise catalytic mechanism has not yet been established.

Computational analyses further suggested that similar β-sheet cradles may occur across the PNGaseA superfamily, potentially indicating a broader specialization for glycopeptide processing. However, the substrate preferences of these predicted homologs require experimental confirmation.

The researchers also proposed a model in which cell-surface PNGaseABif acts on glycopeptides generated by host or microbial proteases. Because the studied B. bifidum strain lacks the α-mannosidases needed to fully degrade the released N-glycan cores, these structures could potentially become available to other gut microbes through cross-feeding.

This proposed ecological role remains hypothetical, as the study did not directly demonstrate glycan cross-feeding or investigate the enzyme's contribution to bacterial growth, infant gut colonization or health outcomes.

Study limitations and conclusions

The study was primarily based on biochemical experiments, crystal structures, and computational modeling. Its structural evidence for peptide recognition came from a single NXT glycopeptide, while the proposed cradle movement and catalytic mechanism require further experimental validation. Some biochemical analyses also had limited replication.

Nevertheless, the findings establish a structural explanation for PNGaseABif's preference for short N-glycopeptides. By identifying how the β-sheet cradle works with the catalytic core to recognize and enclose these substrates, the research advances understanding of bacterial N-glycan processing and provides a basis for investigating the enzyme's biological role in the infant gut.