Integrative metagenomics and structural bioinformatics identify explainable gut microbial variants associated with Crohn’s disease

Nadeem Khan; Muhammad Muneeb Nasir; Ubair Aziz; Haseeb Manzoor; Muhammad Faheem Raziq; Zamir Hussain; Ishrat Jabeen; Masood Ur Rehman Kayani

doi:10.1371/journal.pone.0340748

Abstract

Metagenomics has revealed disease-associated shifts in microbial taxa and functions in inflammatory bowel disease (IBD) patients. However, the role of genomic variation in gut commensals remains poorly understood. Here, we integrated metagenomic profiling, variant calling, and structural bioinformatics to identify disease-associated variants in the gut microbes. Crohn’s disease (CD) and ulcerative colitis (UC) showed significant negative associations with Bacteroides uniformis, Bacteroides vulgatus, and Eubacterium rectale. These bacteria exhibited 190,712 single-nucleotide polymorphisms, including 479 CD-specific and 235 UC-specific variants. Variant prioritization identified a CD-specific Val170Leu substitution in the conserved starch-binding domain of the Starch Utilization System D (SusD) protein in B. uniformis. Structural modeling and cyclodextrin docking indicated reduced binding affinity in the mutant, while 200-ns molecular dynamics simulations showed stable ligand retention only in the wild type. These findings suggest that impaired starch metabolism driven by SusD variation may contribute to B. uniformis depletion in CD and demonstrate the value of integrating metagenomics with structural analyses to identify functionally relevant microbial variants.

Citation: Khan N, Nasir MM, Aziz U, Manzoor H, Raziq MF, Hussain Z, et al. (2026) Integrative metagenomics and structural bioinformatics identify explainable gut microbial variants associated with Crohn’s disease. PLoS One 21(7): e0340748. https://doi.org/10.1371/journal.pone.0340748

Editor: Andrea D. Tyler, Public Health Agency of Canada, CANADA

Received: December 26, 2025; Accepted: June 10, 2026; Published: July 10, 2026

Copyright: © 2026 Khan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting information files.

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

Background

Inflammatory bowel diseases (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC), represent a growing global public health challenge. In 2017, the worldwide burden of IBD reached approximately 6.8 million cases, with prevalence rising from 79.5 to 84.3 per 100,000 population between 1990 and 2017 [1]. IBD is characterized by chronic, uncontrolled inflammation of the gastrointestinal tract, leading to debilitating symptoms such as abdominal pain, severe diarrhea, rectal bleeding, and intestinal obstruction. While the exact etiology remains incompletely understood, both CD and UC arise from a complex interplay of genetic susceptibility and environmental triggers [2–6].

A critical concern in IBD management is the heightened risk of intestinal malignancies. Due to overlapping histopathological features, misdiagnosis or delayed diagnosis can accelerate progression to colorectal cancer (CRC), anal cancer, intestinal lymphoma, and small bowel adenocarcinoma [7–9]. A global meta-analysis of 116 studies revealed that UC patients exhibit a 3.7% higher prevalence of CRC compared to the general population [8]. Similarly, a population-based study in Manitoba, Canada, reported that individuals with UC and CD face a 2–3-fold increased risk of colon cancer [9]. These findings underscore the urgent need for novel molecular insights to improve early detection and precision management of IBD.

Emerging evidence highlights the gut microbiome as a key player in IBD pathogenesis. Dysbiosis marked by shifts in microbial composition and function has been implicated in the distinct pathophysiology of CD and UC [10,11]. For instance, Raed M. Alsulaiman et al. (2023) identified phenotype-specific differences in genera such as Phascolarctobacterium, Maegamonas, Harryflintia, and Akkermansia between UC and CD patients [12]. CD patients also exhibit enriched Escherichia coli compared to healthy controls, a trend not observed in UC [10]. Functional studies further reveal disease-associated microbial pathways, including amino acid metabolism, nucleotide biosynthesis, and carbohydrate degradation [11]. Notably, Bacteroides vulgatus and its protease proteins were overabundant in UC, suggesting a potential role in mucosal inflammation [13].

Despite these advances, the contribution of microbial genomic variation to IBD pathogenesis remains poorly characterized. Single-nucleotide polymorphisms (SNPs) in microbial genomes may significantly alter gene function and host-microbe interactions, yet comprehensive analyses of these variations in IBD are lacking [14,15]. We hypothesized that CD and UC are associated with distinct profiles of microbial genomic variants that functionally contribute to disease-specific pathophysiology. To test this hypothesis, we implemented an analytical pipeline (Fig 1); to systematically identify and characterize gut microbial genomic variants in CD and UC patients. By combining high-resolution metagenomic analysis with microbial variant calling, we identified phenotype-specific variations and employed structural bioinformatics approaches, including molecular docking and molecular dynamics (MD) simulations, to investigate their potential functional implications. We believe our findings provide new insights into the distinct molecular features of CD and UC, which could inform the development of more precise diagnostic and therapeutic strategies in the future.

Download:

Fig 1. Overview of the Research workflow.

The figure shows the complete workflow for identifying gut microbial SNPs specific to CD and UC and associating their effect using structural bioinformatics approaches.

https://doi.org/10.1371/journal.pone.0340748.g001

Materials and methods

Data collection and preprocessing

We analyzed a total of 150 gut metagenomic samples, comprising CD and UC patients, and healthy controls, obtained from the NCBI SRA database using accession number PRJNA945504 (see S1 Table for detailed sample characteristics) [14]. Read quality was assessed through FastQC (v0.12.1) [15], whereas preprocessing was performed using fastp (v0.23.4) [16]. Preprocessing parameters including: -f 15 (for trimming the first 15 bases of each read), -t 20 (for trimming the last 20 bases of each read), -D (for removing duplicated reads) and -l 70 (for setting the minimum allowed read length to 70) were used. Furthermore, the removal of host reads was performed using BBduk by aligning the reads to the human reference genome (GRCh38) [17] to obtain high-quality reads. The average raw read count was ~21.24 ± 4.08 Mbps per sample, whereas the post-processed read count was ~20.46 ± 4.40 Mbps per sample.

Taxonomic profiling and association analysis

The post-processed reads were taxonomically classified using MetaPhlAn3 (v3.0.14) [18] with the default parameters, except “--ignore_eukaryotes” and “--tax_lev s”. The taxonomic profiling identified 557 bacterial species from ~30 genera from all the samples (S2 Table). To identify taxa associated with the three phenotypes, we employed Microbiome Multivariable Associations with Linear Models (MaAsLin2) [19] using parameters; --reference Healthy (setting the healthy group as the baseline for comparisons), --analysis_method LM (linear model framework), --min_prevalence 0.1 (including taxa present in at least 10% of samples), --min_abundance 0.0 and --min_variance 0.0 (retaining all features without abundance or variance filtering), --correction BH (Benjamini–Hochberg procedure for multiple testing correction), and --max_significance 0.05 (FDR threshold of ≤0.05 for statistical significance). Furthermore, significant associations common to CD and UC with prevalence >80% across samples were selected for downstream variant calling.

Variant calling, filtration, and annotation

To characterize genomic variations in the significantly associated microbial species (S3 Table), we performed variant calling using Snippy (v4.6.0) [20]. Quality-filtered metagenomic reads were directly aligned against reference genomes of B. uniformis, B. vulgatus, and E. rectale independently, without metagenomic assembly or genome binning. Although Snippy was originally developed for isolate genome analysis, its use here was justified by species-specific reference alignment and retention of a single high-confidence nonsynonymous variant from B. uniformis (depth: 562×). The analysis was not intended for strain reconstruction or population-level strain resolution, but rather for the identification of high-confidence nonsynonymous variants with potential functional relevance within dominant disease-associated taxa. This approach identified multiple variant classes, including SNPs, multiple nucleotide polymorphisms (MNPs), insertions, deletions, and indels, which were further classified as synonymous and nonsynonymous variants based on their predicted functional effects. For downstream analysis, we focused exclusively on nonsynonymous SNPs, as these variants alter amino acid sequences and are more likely to impact protein function.

To identify disease-specific microbial signatures, we implemented a multi-step filtering approach. First, we excluded SNPs commonly present in both healthy controls and diseased (CD/UC) samples, as these likely represent benign polymorphisms of the core microbial genome. This filtering step was crucial to eliminate variants unrelated to disease pathophysiology. Next, we treated SNPs shared between CD and UC cases as baseline microbial genomic variation (background SNPs), as these might reflect general IBD-associated dysbiosis rather than phenotype-specific characteristics. By focusing exclusively on variants unique to either CD or UC, we aimed to SNP identify phenotype-specific microbial genomic variations potentially contributing to each disease’s distinct pathogenesis. Finally, we applied stringent quality thresholds, prioritizing nonsynonymous SNPs with ≥100 × sequencing coverage. This depth filter served two critical purposes: (1) ensuring high-confidence variant calls by exceeding typical requirements for microbial variant detection (typically 30–50×), and (2) minimizing false positives that could arise from low-abundance strains or sequencing errors. The resulting high-quality, phenotype-specific SNP list formed the basis for our downstream functional analyses.

Functional annotation of SNPs

To functionally characterize the prioritized microbial SNPs in CD and UC, we performed comprehensive annotation of their corresponding proteins using multiple bioinformatics resources. Protein domains and families were analyzed through InterPro, while KEGG facilitated pathway mapping and functional classification. We examined protein-protein interaction networks using STRING and gathered standardized protein annotations from UniProtKB [21–24]. Additionally, we conducted systematic literature searches in PubMed to identify established associations between these proteins and IBD pathogenesis, as well as to evaluate potential functional impacts of the observed mutations.

Structure modeling and analysis

Through functional annotation and literature curation, we identified a clinically relevant SNP in the Bacteroides uniformis SusD gene for detailed analysis, given its established role in polysaccharide utilization and potential to modulate host-microbe interactions. The selected variant encodes a valine-to-leucine substitution at position 170 (V170L) in the SusD protein. We designated the Val170 form as wild-type SusD and the Leu170 variant as mutant SusD. Three-dimensional structures for both protein versions were generated using MODELLER v10.6 through homology-based modeling [22]. The optimal structural template was identified by BLASTp searches against the Protein Data Bank (PDB), which revealed Bacteroides thetaiotaomicron SusD (PDB ID: 3CK7; 2.10 Å resolution) as the closest homolog, with ~25% sequence identity and ~90% query coverage. This template was subsequently employed to generate high-confidence structural models for downstream comparative analyses [25,26]. Ten models each for the wild-type and mutant SusD proteins were generated to evaluate SNP-induced conformational changes. To improve structural representation in regions potentially affected by reduced sequence identity, additional loop refinement was performed during model optimization. The best models were further refined through energy minimization and MD simulations using Maestro v12.8.117 (Schrödinger Suite), employing the OPLS-2005 force field [27].

Molecular docking

Following protein modeling and structural refinement, we performed molecular docking to investigate the biological impact of SNP-induced variations on substrate binding. The binding site of the query proteins was selected that includes 10 Å of area surrounding the SNP residue, to precisely evaluate the SNP-induced structural alterations upon ligand interactions. Subsequently, docking simulations were performed using MOE 2015 with default parameters for placement and refinement. The GBVI/WSA dG (Generalized Born Volume Integral/ Weighted Surface Area) scoring function was used for both pose generation and refinement, generating 100 ligand conformations for each wild and mutant protein [27]. The mentioned scoring function estimates the free energy of binding (ΔG_bind) through the following empirical equation (1):

(1)

where E_(inter) is the interaction energy (electrostatic and van der Waals), ΔG_(solv) is the desolvation energy from the GBVI model, ΔG_(hydro) represents the hydrophobic contribution estimated via weighted surface area (WSA), and ΔG_(entropy) is the entropy penalty upon ligand binding. This function integrates interaction, solvation, and entropic effects to provide a refined estimate of binding affinity.

MD simulations

MD simulations were conducted for 200 ns using Maestro v12.8.117 (Schrödinger suite) to assess the stability and conformational changes of the docked complexes over time [26]. The OPLS-2005 force field was employed to model molecular interactions, as it provides a well-validated balance between accuracy and computational efficiency in simulating biological systems [25]. The total potential energy computed for the complexes followed equation (2):

(2)

Where E_bond, E_angle and E_torsion represent bonded interactions including bond stretching, angle bending, and torsional rotations, respectively. E_non-bonded accounts for van der Waals and electrostatic interactions between atoms not directly bonded. This comprehensive formulation enables realistic modeling of molecular geometry and energetics in biological systems. Furthermore, the simulation system was solvated using a TIP3P water model within an orthorhombic box, maintaining a 10 Å buffer around the complex [28]. The system was further neutralized with 0.15 M NaCl, and standard simulation conditions were applied, including a temperature of 300K and pressure of 1 atm. Equilibration was performed using the NVT and NPT ensembles for 100 ps before running the production phase of 200 ns [29]. Parameters such as Root Mean Square Deviation (RMSD), Root Mean Square Fluctuation (RMSF), and hydrogen bonding were calculated to assess the structural stability, flexibility, and interaction patterns of wild type and mutant complexes.

Results

Identification of disease-associated microbial signatures and functional variants

To identify microbial taxa, whose abundance patterns distinguish IBD phenotypes from healthy states, we applied MaAsLin2 using linear model (ML) framework. This approach was specifically chosen to quantify the strength of phenotype-specific associations and identify potential protective species showing consistent depletion across both CD and UC. This revealed 190 significantly associated species, with 68 showing CD-specific and 122 showing UC-specific associations (S1 Fig). Among these, three commensal species, B. uniformis (CD FDR = 3.01 × 10⁻⁶, coefficient = 6.18; UC FDR = 2.02 × 10⁻⁵, coefficient = 5.1), E. rectale (CD FDR = 1.99 × 10⁻⁷, coefficient = 6.14; UC FDR = 2.32 × 10⁻⁵, coefficient = 4.37), and B. vulgatus (CD FDR = 0.00163, coefficient = 3.12; UC FDR = 0.0421, coefficient = 1.97) demonstrated strong positive associations with health (Fig 2). Since healthy status was used as the reference, these results indicate significant depletion of these species in both CD and UC patients, suggesting their protective roles in gut homeostasis.

Download:

Fig 2. Association analysis of bacterial species with CD and UC.

The box plots represent the abundance of bacterial species across the phenotypes, along with significant association coefficients (FDR ≤ 0.05). Positive coefficient values, with healthy as the reference, indicate that the abundance of the mentioned bacterial species is inversely proportional to CD (A) and UC (B), and vice versa.

https://doi.org/10.1371/journal.pone.0340748.g002

We hypothesized that strain-level genomic variations in these negatively associated commensals might contribute to their depletion patterns in CD and UC. To investigate this possibility, we performed comprehensive variant analysis of these three species, identifying >1.4M total genomic variations across all samples, including 269,844 non-synonymous mutations (S4 Table and S2 Fig). The SNP distribution patterns revealed intriguing phenotype-specific differences: while B. uniformis showed progressive SNP reduction from healthy (22,290) to CD (18,009) to UC (11,837), B. vulgatus exhibited an opposite trend with CD-associated increase (healthy: 34,912; CD: 40,361; UC: 22,718). From these, we focused on high-confidence variants unique to each phenotype (Table 1), ultimately prioritizing 714 SNPs with sequencing depth >300× (CD:479; UC:235) (S3 Fig and S5 Table) for downstream functional analysis. This stringent selection was based on the premise that these strain-specific variations might mechanistically explain the observed depletion patterns of beneficial microbes in IBD pathogenesis.

Download:

Table 1. Comparative summary of synonymous and nonsynonymous variants across bacterial species associated with CD, UC, and healthy phenotypes.

https://doi.org/10.1371/journal.pone.0340748.t001

SNP in SusD implicates disruption of the starch utilization pathway of B. uniformis in CD patients

To identify functionally relevant variants, we implemented a stringent filtering approach that selected for nonsynonymous SNPs within evolutionarily conserved protein domains. This analysis yielded 244 high-priority SNPs (164 CD-associated and 80 UC-associated variants). The biological relevance of these variants was further supported by existing literature demonstrating that impairment of bacterial polysaccharide utilization systems particularly those involved in SCFA production can disrupt mucosal homeostasis and promote inflammatory responses. To functionally characterize the 244 prioritized variants, we performed comprehensive annotation of the affected proteins and filtered for those involved in carbohydrate metabolism using relevant keywords (polysaccharide, glucan, starch, binding, and metabolism). This analysis identified 41 proteins functionally linked to polysaccharide utilization pathways, including 10 in B. uniformis, 15 in B. vulgatus, and 16 in E. rectale (S6 Table).

Notably, several proteins identified in this subset, including phosphopyruvate hydratase (eno), glucose-1-phosphate adenylyltransferase subunit GlgD (glgD), L-arabinose isomerase (araA), glycerol kinase (glpK), and 2,3-bisphosphoglycerate-independent phosphoglycerate mutase (gpmI), play established roles in carbohydrate metabolism. Among these, a nonsynonymous SNP in the SusD protein, a key component of the starch utilization system (Sus) emerged as particularly relevant. This SNP, with a high sequencing depth (562X), involved a valine-to-leucine substitution at the 170th amino acid position (Val170Leu), specifically in B. uniformis from CD patients. Furthermore, the SNP was located within the conserved domain of the SusD protein (residues 89–230), which is essential for its starch-binding function. Network analysis further underscored the functional importance of SusD, revealing direct associations with other core proteins of the starch utilization system (SusB, SusC, SusE, and SusF), with positive co-expression scores of 0.067, 0.050, 0.073, and 0.103, respectively, each predicted with a confidence level exceeding 70% (S4 Fig).

Molecular modeling revealed local structural perturbations in mutant SusD

We hypothesized that the Val170Leu mutation in Bacteroides uniformis SusD would induce local structural alterations that affect its starch-binding capacity, while maintaining overall protein stability. To test this, we modeled structures of wild-type (Val170) and mutant (Leu170) SusD followed by MD simulations. Comprehensive sequence alignment revealed substantial regions of discontinuity between template and target sequences (S5 Fig). Most critically, the mutation locus and its flanking residues (Ile169, Thr171, Ser172, Tyr173, Glu174, Gln175, Pro176, Gln177, Asp178, Glu179, and Val180) resided within an extensive unstructured loop region completely absent from the template structure. This structural gap encompassed eleven consecutive residues proximal to the polymorphic site, including several residues predicted to participate in substrate coordination based on homologous systems.

Targeted loop refinement was performed using MODELLER v10.6 with the ‘very_slow’ refinement protocol (incorporating slow, iterative optimization steps) to specifically address the unresolved region. This high-stringency approach successfully resolved the backbone conformation for four consecutive residues (Ile169, Val170, Thr171, Ser172) in the wild-type structure and three residues (Ile169, Val170, Thr171) in the mutant variant, thereby enabling comparative structural analysis of this functionally critical region.

Simulation studies revealed critical differences in protein behavior between wild-type and mutant SusD variants. Global stability analysis showed comparable RMSD profiles for both structures, maintaining an average deviation of ~6 Å after 80 ns with similar fluctuation patterns throughout the 200 ns simulation (S6A Fig). RMSF analysis demonstrated that while the overall structure remained stable, the mutant SusD exhibited increased flexibility at specific residues, including Ser58, Glu59, Asp87, Asn209, Met211, Leu346, Ala347, and Arg466. These residues showed substantially greater fluctuations in the mutant variant (average ~5 Å) compared to the wild type (~2 Å) (S6B Fig). The mutation-induced flexibility changes were particularly pronounced in regions surrounding the substrate-binding site, suggesting potential functional consequences for starch binding and processing.

We hypothesized that the Val170Leu substitution would induce localized conformational changes in SusD’s starch-binding domain while preserving global protein architecture. Supporting this prediction, comparative analysis of the 200 ns simulation endpoints revealed significant local reorganization around the mutation site (Thr171-Asp178; Fig 3), despite maintained global structural integrity. Quantitative superposition with the wild-type structure identified three distinct perturbation clusters: (1) a severely displaced region centered on Glu174 (21.0 Å deviation), (2) intermediate displacements at Pro176 (13.1 Å) and Gln177 (10.3 Å), and (3) moderate but functionally relevant shifts at the mutation site (Leu170: 6.0 Å) and adjacent residues (average regional deviation: 12.0 Å). These findings demonstrate that the Val170Leu mutation creates a cascade of structural perturbations radiating through the binding domain, with magnitude inversely proportional to distance from the mutation site, a pattern consistent with allosteric modulation of substrate recognition geometry.

Download:

Fig 3. Structural superimposition of wild and mutant SusD.

(A) Superimposed structures of wild and mutant SusD, where the region, encompassing the variation (Val170Leu) along with its neighboring residues, showed a major structural deviation. (B) Conformational deviations of each residue present along the variation site, demonstrating the extent of deviation between each residue of wild and mutant SusD.

https://doi.org/10.1371/journal.pone.0340748.g003

Docking analysis indicates weakened cyclodextrin binding in mutant SusD

To evaluate whether the observed structural perturbations affected ligand binding, we performed molecular docking of cyclodextrin, an oligosaccharide mimicking starch into the binding sites of both wild-type and mutant SusD. Cyclodextrin (S7 Fig), originally co-crystallized with the template protein (PDB ID: 3CK7), was docked into the binding pocket of each version of protein, generating 100 conformations per complex. Comparative analysis revealed significant differences in binding energies and interaction networks between the wild-type and mutant complexes. The wild-type SusD exhibited stronger cyclodextrin binding, with an average binding energy of −6.274 kcal/mol across all conformations compared to −5.496 kcal/mol for the mutant. This trend was particularly evident in the top-scoring poses, where the wild-type complex achieved a binding energy of −9.125 kcal/mol versus −6.092 kcal/mol for the mutant (S8 Fig).

Detailed examination of binding interactions showed that the wild-type SusD maintained a more extensive network of stabilizing contacts with cyclodextrin. Specifically, five residues (Ser123, Ser126, Asp131, Ser172, and Asp178) participated in hydrogen bonding with the ligand, while only three residues (Asn124, Glu174, and Gln175) were involved in the mutant complex (Fig 4A and 4C). Similarly, non-bonded interactions were more numerous in the wild-type complex, involving ten residues (Leu122, Leu127, Ile130, Ser134, Arg153, Val170, Thr171, Pro176, Gln177, and Val180) compared to seven residues (Leu116, Val120, Tyr173, Pro176, Lys523, Asp526, and Leu527) in the mutant (S9A and S9C Fig).

Download:

Fig 4. Residues of wild and mutant SusD making hydrogen bonds with cyclodextrin before and after simulation.

(A) Residues of wild SusD involved in hydrogen bond formation with cyclodextrin after docking, while (B) Interacting residues following molecular dynamics simulation. (C) Residues of mutant SusD contributing to hydrogen bonding with cyclodextrin after initial docking. (D) Corresponding interactions after simulation.

https://doi.org/10.1371/journal.pone.0340748.g004

Comparative MD simulations highlight impaired ligand retention in mutant SusD

Building upon our previous findings that the Val170Leu mutation induces local structural perturbations in SusD (average deviation of ~12 Å around residues 170–178), we hypothesized that these conformational changes would specifically disrupt cyclodextrin binding stability. Molecular dynamics simulations of the docked complexes confirmed this hypothesis, revealing marked differences in ligand interaction patterns between wild-type and mutant forms. The wild-type complex maintained structural integrity throughout the 200 ns simulation, with cyclodextrin preserving both its binding pose and key interactions observed in docking studies. Hydrogen bonds with Ser123, Ser126, Leu127, Asp131, and Asp178 remained stable (Fig 4B), while non-bonded contacts with ten binding site residues persisted (S9B Fig). This stability contrasted sharply with the mutant complex, where cyclodextrin progressively dissociated from its original position – particularly from the mutation-affected region (residues 170–178) that showed the greatest structural deviations in our earlier analysis. The displaced ligand formed alternative interactions with Val444, Glu448, and Gly450 (Fig 4D), while losing contact with original binding site residues. Non-bonded interactions similarly shifted to Lys443, Arg445, Asn446, Gly447, Ser451, Leu452, and Glu453 (S9D Fig).

Comparative RMSD analysis revealed distinct structural stabilization patterns between the wild-type and mutant SusD-cyclodextrin complexes during the 200 ns simulation (Fig 5A). The wild-type complex underwent rapid conformational adjustments during the initial 20 ns, as evidenced by an RMSD increase to ~5.5–6 Å, after which it maintained stable structural equilibration with minimal fluctuations. In contrast, the mutant complex displayed slower structural reorganization, gradually increasing to ~5 Å RMSD over 150 ns before achieving relative stability. RMSD analysis of the docked cyclodextrin revealed distinct behaviors: within the mutant complex, it showed initial fluctuations before stabilizing at ~80 Å displacement, while the wild type reached ~18 Å RMSD within 45 ns and remained stable (S10 Fig). RMSF analysis (Fig 5B) demonstrated greater flexibility (~5.2 Å) in the mutant’s Thr171-Asp178 region and alternative binding site Val444-Glu453 (~5.6 Å) compared to the wild-type, consistent with the observed binding site displacement and local structural perturbations.

Download:

Fig 5. Molecular dynamics simulation of docked complexes.

(A) Comparative RMSD of docked wild and mutant SusD along a 200 ns simulation. Both the docked proteins initially deviate from their initial conformation at the beginning of the simulation and stabilize (minimal conformational changes) after the 150th time point of the simulation. (B) Comparative conformational fluctuations of each residue of both docked complexes along the simulation run, where multiple residues of the mutant complex showed notable conformational changes compared to the wild complex, demonstrating the variable interaction of cyclodextrin to multiple residues of the mutant complex at variable locations.

https://doi.org/10.1371/journal.pone.0340748.g005

Lastly, we hypothesized that the Val170Leu mutation would disrupt the stable hydrogen bonding network required for productive cyclodextrin binding in SusD. Supporting this prediction, our hydrogen bond analysis revealed markedly different interaction profiles between wild-type and mutant complexes. The wild-type SusD maintained persistent hydrogen bonds with Thr171, Ser172, Tyr173, and Glu174 throughout the simulation (S11A Fig), preserving the initial binding mode. In striking contrast, the mutant complex failed to sustain these native interactions, instead forming new hydrogen bonds with Val444-Glu453 residues (S11B Fig), a clear shift from the original binding site. Between 50–150 ns, the mutant complex established consistent but alternative interactions with Glu448, Asp449, Gly450, Ser451, Leu452, and Glu453. Time-resolved conformational analysis (Fig 6) visually confirmed these differences, with cyclodextrin maintaining stable positioning in the wild-type complex while progressively dissociating in the mutant. These results demonstrate that the Val170Leu mutation not only alters local protein structure but also critically disrupts the hydrogen bonding network essential for proper substrate recognition and binding.

Download:

Fig 6. Comparative binding conformations of cyclodextrin at different time points of the MD simulation.

Binding conformations of cyclodextrin with wild and mutant SusD at ten evenly spaced time points over the total simulation period (200 ns) were selected. (A) Cyclodextrin exhibited overall stable binding within the chosen docked site (yellow) of wild-type SusD (green). (B) In the mutant SusD (pink), cyclodextrin displayed unstable binding, characterized by notable positional shifts and interactions with other regions (detailed in the main text), highlighting the effect of mutations compared to the wild-type structure.

https://doi.org/10.1371/journal.pone.0340748.g006

Discussion

The chronic inflammation characterizing IBD arises from complex host-microbe interactions that remain incompletely understood. While previous studies have identified shifts in microbial abundance in CD and UC patients, our work provides unprecedented mechanistic insight by characterizing strain-level genomic variations and their structural-functional consequences in key commensal species. Specifically, we confirmed the (1) depletion of B. uniformis, B. vulgatus, and E. rectale in IBD phenotypes, (2) identified phenotype-specific genomic variants within these species, and (3) structurally characterized a disease-associated SNP in the B. uniformis SusD protein that impaired starch-binding interactions, suggesting a potential mechanistic link between microbial genomic variation and CD pathogenesis.

Our taxonomic profiling and subsequent association analysis are consistent with previous reports of Bacteroides depletion in IBD [30,31]. Furthermore, previous studies have also demonstrated that B. uniformis, B. vulgatus, and E. rectale exhibit a significant positive association with healthy controls compared to CD and UC phenotypes [32], thus indicating their beneficial role in gut homeostasis. The protective role of B. uniformis has been experimentally validated in murine models, where administration of this commensal bacterium maintained intestinal homeostasis without inducing adverse histological effects. [33,34]. Similarly, the abundant nature of E. rectale and B. vulgatus in the normal human gut compared to CD and UC, and their role in alleviating colitis, has also been reported previously [35,36].

While the differential abundance patterns of these three bacteria in CD and UC patients provide important ecological insights, these observations alone cannot explain the mechanistic basis of disease pathogenesis. To bridge this knowledge gap, we investigated whether genomic variations within these commensals might contribute to their depletion and functional impairment in CD and UC. In metagenomes, genomic variations can be influenced by strain heterogeneity. Our analysis was intentionally restricted to highly prevalent bacterial taxa with strong disease associations, and only high-confidence nonsynonymous variants supported by stringent filtering criteria were retained. Specifically, variants shared between healthy and disease groups were excluded, and only SNPs supported by high sequencing depth (>100×) were prioritized to minimize potential false-positive calls arising from low-abundance strains or sequencing noise.

Our analysis revealed substantial genetic divergence, with E. rectale exhibiting the highest number of variations, followed by B. uniformis and B. vulgatus. Among these, a particularly noteworthy finding was a high-confidence, non-synonymous SNP (Val170Leu) located within the functional starch-binding domain of B. uniformis SusD, a key component of the Bacteroides starch utilization system [30,37]. This mutation, exclusively identified in CD patients with high sequencing depth (562×), represents a strong candidate for functional impairment, as domain-localized nonsynonymous SNPs are more likely to disrupt protein-substrate interactions critical for commensal fitness and host benefit.

Starch serves as a critical nutritional substrate for Bacteroides, functioning as both their primary energy source and a fundamental requirement for maintaining metabolic homeostasis and ecological fitness in the gut environment [36]. Bacteroides-mediated starch metabolism plays a central role in gut microbial cross-feeding, generating metabolic byproducts that serve as key nutrients for other commensal bacteria. This metabolic cooperation underpins critical symbiotic relationships within the gut ecosystem [38]. The starch utilization system (Sus) in Bacteroides species comprises a highly conserved seven-protein complex (SusA-G) that orchestrates the efficient breakdown of starch into metabolically useful oligosaccharides and monosaccharides [39–41]. The SusD, SusE, and SusF proteins localized on the outer membrane of Bacteroides facilitate starch binding and capture, while SusG mediates its initial hydrolysis into oligosaccharides. These oligosaccharides are then transported across the outer membrane via SusC into the periplasm, where SusA and SusB further degrade them into glucose monomers for cellular uptake and metabolism [26,37,42–44]. The complete degradation of starch by gut microbes yields short-chain fatty acids (SCFAs) as primary metabolic end-products, which serve as crucial signaling molecules and energy substrates for both microbial communities and the host. These SCFAs particularly acetate, propionate, and butyrate orchestrate key physiological processes including immune modulation, epithelial barrier maintenance, and energy homeostasis through specific receptor-mediated pathways [43]. SCFAs play a pivotal role in immunomodulation, with Bacteroides-derived fermentation products serving dual functions as both the primary energy source for colonocytes and key regulators of inflammatory responses. Through inhibition of histone deacetylases (HDACs) and suppression of NF-κB signaling pathways, these microbial metabolites downregulate proinflammatory cytokine production, thereby maintaining intestinal immune homeostasis and preventing aberrant inflammation [36].

SusD serves as the main starch-binding protein in the Bacteroides starch utilization system, playing an essential role in substrate capture within the gut environment. This critical function was experimentally validated by Shipman et al. (2000), who demonstrated that domain-specific mutations in SusD of B. thetaiotaomicron completely abolished starch binding capacity, while analogous mutations in other starch-binding proteins maintained partial activity [40]. Our identification of a functional domain mutation (Val170Leu) in the SusD protein of B. uniformis from CD patients provides a plausible mechanistic link between microbial genetic variation and intestinal inflammation. We hypothesize this mutation impairs starch binding and disrupts the starch utilization system in B. uniformis, reducing glucose and SCFA production. This may explain B. uniformis depletion in CD patients, while SCFA deficiency could contribute to disease through: (1) diminished colonocytes energy, (2) dysregulated immune responses, and (3) compromised epithelial integrity [45–48].

Importantly, an alternative but complementary interpretation should also be considered. The inflammatory microenvironment characteristic of CD and UC, including immune cell infiltration, altered intestinal pH, oxidative stress, and elevated inflammatory mediators, may itself drive substantial shifts in microbial composition and metabolic activity [47,48]. Under such conditions, the observed depletion of B. uniformis may represent a secondary ecological consequence of host-driven inflammation rather than a primary event initiated by microbial genomic variation. Within this context, microbial mutations and inflammatory selection pressures are not mutually exclusive and may instead act synergistically in shaping disease-associated dysbiosis.

The structural perturbations induced by the Val170Leu mutation provide compelling mechanistic support for our hypothesis of impaired starch utilization in CD-associated B. uniformis. The persistent conformational changes observed in the binding domain (residues Thr171-Asp178), evidenced by both static structural comparisons and dynamic simulations (~5.2 Å RMSF increase), suggest a substantial reorganization of the starch-binding interface. While global protein folding remained stable (as indicated by comparable overall RMSD), the localized destabilization of this critical functional region would likely compromise several aspects of substrate interaction: (1) initial starch recognition by SusD, (2) productive positioning for SusG-mediated hydrolysis, and (3) efficient transfer to downstream Sus components. These observations align with established structure-function relationships in starch-binding proteins, where even minor perturbations in binding regions can dramatically reduce substrate affinity [49]. The mutation’s location within SusD’s conserved domain (residues 89–230) further underscores its likely functional significance, as this domain has been evolutionarily optimized for polysaccharide interactions [24]. Notably, the affected residues (Thr171-Asp178) constitute a critical segment that, in homologous systems, directly participates in oligosaccharide coordination through both hydrogen bonding and hydrophobic contacts. This structural evidence, combined with our docking results showing reduced cyclodextrin affinity, strongly suggests that the Val170Leu variant could be a loss-of-function mutation with potential consequences for B. uniformis ecology and host-microbe interactions in CD.

Similarly, site-specific docking supports our hypothesis, demonstrating that fewer residues participate in cyclodextrin binding in mutant SusD compared to wild type, resulting in reduced binding affinity. The comparative binding energies, lacking absolute reference values, prompted validation through 200 ns simulation studies [50]. MD simulations revealed distinct dynamic behavior of cyclodextrin between the mutant and wild-type SusD complexes. While cyclodextrin maintained stable binding throughout the simulation in the wild-type complex, the mutant complex exhibited early conformational fluctuations followed by stabilization at a significantly displaced position. These observations demonstrate how structural perturbations affect ligand binding and strongly support our hypothesis of mutation-induced disruption in substrate recognition.

The hydrogen bonding patterns provide critical molecular-level insights into the mutation’s functional consequences. The wild-type complex maintained stable interactions with residues 170–174 throughout simulations, explaining its consistent binding pose. In striking contrast, the mutant’s inability to sustain these interactions, instead forming transient bonds before relocating to residues 443–490 – reveals fundamental defects in substrate recognition. While these alternative interaction sites reside in α-helical regions (Koropatkin et al., 2008), their apparent inability to support productive starch binding suggests the Val170Leu mutation forces compensatory, but functionally inferior, binding modes. This observation aligns with our structural data showing local domain destabilization, collectively demonstrating how single-residue changes can propagate to disrupt entire functional networks in microbial nutrient acquisition systems [24].

While our findings provide mechanistic insights into how microbial genomic variations may contribute to CD pathogenesis, several limitations warrant consideration. First, although we identified functionally disruptive variants in B. uniformis SusD, we did not comprehensively investigate parallel variations in the other two protective taxa (B. vulgatus and E. rectale), which may harbor equally or more significant disease associations. Second, our computational predictions require validation in independent IBD cohorts to confirm the diagnostic specificity of the Val170Leu mutation. Third, while our structural models strongly suggest impaired starch binding, experimental validation through in vitro affinity measurements and in vivo colonization assays remains essential. Finally, the limited template similarity (24.9% identity) for homology modeling, particularly in unresolved substrate-binding loops, introduces uncertainty in local conformational predictions. Addressing these limitations through multi-cohort validation, functional assays will strengthen the translational potential of these findings.

Conclusions

Our integrated analysis identifies the Val170Leu mutation in B. uniformis SusD as a potential CD-specific biomarker. By combining high-resolution metagenomic analysis, microbial variant calling, and structural bioinformatics, we demonstrate how this variant might disrupt starch-binding functionality, a defect that may contribute to inflammation through impaired SCFA production. The consistent computational evidence (reduced binding affinity, ligand displacement, and altered interaction networks) implicates this mutation in CD pathogenesis. While requiring experimental validation, these findings establish a hypothetical framework for targeting microbial genomic variants in precision IBD management.

Supporting information

S1 Fig. Bacterial association across phenotypes.

The Venn diagram illustrates the total number of significantly associated species with CD and UC, along with the number of species uniquely and commonly associated with each phenotype.

https://doi.org/10.1371/journal.pone.0340748.s001

(PDF)

S2 Fig. Variant counts summary.

The overall comparative variation number in highly associated bacterial species is shown. SNP variations are high in number in all the species, accordingly with respect to the other variations.

https://doi.org/10.1371/journal.pone.0340748.s002

(PDF)

S3 Fig. Catalogue of microbial genes with reported SNPs.

The bacterial species exhibit a variable number of reported SNPs across their genomes, affecting many essential genes that encode specific proteins. These SNPs vary in depth, as shown, and have a significant impact on the corresponding proteins. Notably, a higher number of affected proteins at both lower and higher depths is predominantly observed in CD samples compared to UC samples.

https://doi.org/10.1371/journal.pone.0340748.s003

(PDF)

S4 Fig. Network analysis of starch utilization system proteins.

The network showed the close interactions of SusD with other proteins (nodes). The thickness of the interacting lines indicated the strength of the association between proteins and SusD. Furthermore, the green nodes, representing the neighboring proteins (see text), show close interactions with SusD, clarifying the intricate role of all the proteins in starch metabolism. Yellow nodes represent hypothetical proteins that might have some contributing associations with SusD accordingly.

https://doi.org/10.1371/journal.pone.0340748.s004

(PDF)

S5 Fig. Sequence alignment of wild-type and mutant SusD with template sequence.

The alignment of wild and mutant SusD with the template sequence (3CK7) is shown in (A) and (B). The residues identical to the template residues are represented in black color, further demonstrated by dark green bands. Residues showing similarity to the template sequence are colored grey, while non-matching residues are in white. Gaps are shown to dotted lines. Furthermore, the site where the mutation occurs and its neighboring residues are enclosed in red box, and the alignment results show that most of the residues show no similarity with the template sequence. Holistically, there are multiple residues in the query sequences that show similarity to the template sequence.

https://doi.org/10.1371/journal.pone.0340748.s005

(PDF)

S6 Fig. Simulation of wild and mutant SusD.

(A) illustrate the comparative RMSD of wild and mutant SusD along a 200 ns simulation run. Both proteins initially deviate from their initial conformation at the beginning of the simulation and stabilize (minimal conformational changes) after the 100^th time point of simulation. (B) Comparative conformational fluctuations of each residue of wild and mutant SusD are shown along the simulation run, where multiple residues of mutant SusD showed notable conformational changes compared to wild SusD.

https://doi.org/10.1371/journal.pone.0340748.s006

(PDF)

S7 Fig. Structure of cyclodextrin.

A cyclic chain of uniformly distributed hydrogen, oxygen and hydrogen atoms.

https://doi.org/10.1371/journal.pone.0340748.s007

(PDF)

S8 Fig. Comparative docking scores.

The plot showed the binding scores of cyclodextrin with wild and mutant SusD across 100 conformations. Each docked conformation of cyclodextrin with wild-type SusD showed a strong binding affinity (more negative energy value) as compared to mutant SusD.

https://doi.org/10.1371/journal.pone.0340748.s008

(PDF)

S9 Fig. Non-bonded interaction patterns before and after simulation in wild-type and mutant complexes.

(A) shows the residues of wild SusD involved in non-bonded interactions with cyclodextrin after docking, while (B) presents the interacting residues following molecular dynamics simulation. Similarly, (C) illustrates the residues of mutant SusD contributing to non-bonded interactions with cyclodextrin after initial docking, whereas (D) depicts the corresponding interactions after simulation.

https://doi.org/10.1371/journal.pone.0340748.s009

(PDF)

S10 Fig. RMSF of docked complexes.

Comparative conformational fluctuations of each residue of wild and mutant docked complexes are shown along the simulation run, where multiple residues of the mutant complex showed notable conformational changes compared to the wild-type SusD complex, demonstrating the variable interaction of cyclodextrin with multiple residues of the mutant SusD complex at variable locations.

https://doi.org/10.1371/journal.pone.0340748.s010

(PDF)

S11 Fig. Hydrogen bonding patterns of wild and mutant SusD complex.

(A) Hydrogen bond-forming residues of wild SusD with cyclodextrin over the 200 ns simulation. Residues 172^nd to 174^th showed consistent interactions throughout, while residues 130^th contributed up to 130 ns with minor fluctuation. Additional transient interactions were observed with residues 136^th, 176^th, and 178^th. (B) Mutant SusD residues forming hydrogen bonds with cyclodextrin. Residues 448^th and 453^rd maintained notable interactions throughout, while residue 174^th interacted stably up to 20 ns. Residues 444^th and 450^th to 452^nd exhibited intermittent bonding.

https://doi.org/10.1371/journal.pone.0340748.s011

(PDF)

S1 Table. Information of metagenomics samples used in this study.

https://doi.org/10.1371/journal.pone.0340748.s012

(XLSX)

S2 Table. Relative abundances of bacterial species across CD, UC and healthy phenotypes.

https://doi.org/10.1371/journal.pone.0340748.s013

(XLSX)

S3 Table. Significant associations of bacterial species across CD and UC with reference to healthy controls.

https://doi.org/10.1371/journal.pone.0340748.s014

(XLSX)

S4 Table. Catalogue of all types of microbial variations present in CD, UC and healthy phenotypes.

https://doi.org/10.1371/journal.pone.0340748.s015

(XLSX)

S5 Table. Catalogue of prioritized SNPs specific to CD and UC phenotypes.

https://doi.org/10.1371/journal.pone.0340748.s016

(XLSX)

S6 Table. List of SNPs functionally associated with polysaccharide metabolism.

https://doi.org/10.1371/journal.pone.0340748.s017

(XLSX)

Acknowledgments

We thank the members of the Metagenomics Discovery Lab for their continuous feedback and support in this research. We also thank the SINES Supercomputing Lab for providing the computational resources.

References

1. Alatab S, et al. The global, regional, and national burden of inflammatory bowel disease in 195 countries and territories, 1990--2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Gastroenterol Hepatol. 2020;5(1):17–30.
- View Article
- Google Scholar
2. El Hadad J, Schreiner P, Vavricka SR, Greuter T. The genetics of inflammatory bowel disease. Mol Diagn Ther. 2024;28(1):27–35. pmid:37847439
- View Article
- PubMed/NCBI
- Google Scholar
3. Hisamatsu T, Suzuki M, Reinecker H-C, Nadeau WJ, McCormick BA, Podolsky DK. CARD15/NOD2 functions as an antibacterial factor in human intestinal epithelial cells. Gastroenterology. 2003;124(4):993–1000. pmid:12671896
- View Article
- PubMed/NCBI
- Google Scholar
4. Fiocchi C. Omics and multi-omics in IBD: no integration, no breakthroughs. Int J Mol Sci. 2023;24(19):14912. pmid:37834360
- View Article
- PubMed/NCBI
- Google Scholar
5. Schofield JB, Haboubi N. Histopathological mimics of inflammatory bowel disease. Inflamm Bowel Dis. 2019;26(7):994–1009.
- View Article
- Google Scholar
6. Sawyerr AM, Ferguson A. Ulcerative colitis and Crohn’s disease: similarities and distinctions. J R Coll Physicians Lond. 1994;28(3):197–9. pmid:7932315
- View Article
- PubMed/NCBI
- Google Scholar
7. Xie J, Itzkowitz SH. Cancer in inflammatory bowel disease. World J Gastroenterol. 2008;14(3):378–89. pmid:18200660
- View Article
- PubMed/NCBI
- Google Scholar
8. Nadeem MS, Kumar V, Al-Abbasi FA, Kamal MA, Anwar F. Risk of colorectal cancer in inflammatory bowel diseases. Semin Cancer Biol. 2020. p. 51–60.
- View Article
- Google Scholar
9. Bernstein CN, Blanchard JF, Kliewer E, Wajda A. Cancer risk in patients with inflammatory bowel disease: a population-based study. Cancer. 2001;91(4):854–62.
- View Article
- Google Scholar
10. Dahal RH, Choi Y-J, Kim S, Kim J. Differentiation of Escherichia fergusonii and Escherichia coli isolated from patients with inflammatory bowel disease/ischemic colitis and their antimicrobial susceptibility patterns. Antibiotics (Basel). 2023;12(1):154. pmid:36671355
- View Article
- PubMed/NCBI
- Google Scholar
11. Liang G. Altered gut bacterial and metabolic signatures and their interaction in inflammatory bowel disease. Synth Syst Biotechnol. 2021;6(4):377–83. pmid:34820530
- View Article
- PubMed/NCBI
- Google Scholar
12. Alsulaiman RM, Al-Quorain AA, Al-Muhanna FA, Piotrowski S, Kurdi EA, Vatte C, et al. Gut microbiota analyses of inflammatory bowel diseases from a representative Saudi population. BMC Gastroenterol. 2023;23(1):258. pmid:37507685
- View Article
- PubMed/NCBI
- Google Scholar
13. Mills RH, Dulai PS, Vázquez-Baeza Y, Sauceda C, Daniel N, Gerner RR, et al. Multi-omics analyses of the ulcerative colitis gut microbiome link Bacteroides vulgatus proteases with disease severity. Nat Microbiol. 2022;7(2):262–76. pmid:35087228
- View Article
- PubMed/NCBI
- Google Scholar
14. Kang D-Y, Park J-L, Yeo M-K, Kang S-B, Kim J-M, Kim JS, et al. Diagnosis of Crohn’s disease and ulcerative colitis using the microbiome. BMC Microbiol. 2023;23(1):336. pmid:37951857
- View Article
- PubMed/NCBI
- Google Scholar
15. Andrews S, et al. FastQC: a quality control tool for high throughput sequence data. Cambridge, United Kingdom. 2010.
- View Article
- Google Scholar
16. Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884–90. pmid:30423086
- View Article
- PubMed/NCBI
- Google Scholar
17. Bushnell B. BBMap: a fast, accurate, splice-aware aligner. 2014.
18. Beghini F, McIver LJ, Blanco-Míguez A, Dubois L, Asnicar F, Maharjan S, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife. 2021;10:e65088. pmid:33944776
- View Article
- PubMed/NCBI
- Google Scholar
19. Mallick H, Rahnavard A, McIver LJ, Ma S, Zhang Y, Nguyen LH, et al. Multivariable association discovery in population-scale meta-omics studies. PLoS Comput Biol. 2021;17(11):e1009442. pmid:34784344
- View Article
- PubMed/NCBI
- Google Scholar
20. Seemann T. Snippy: rapid haploid variant calling and core SNP phylogeny. https://github.com/tseemann/snippy. 2015.
- View Article
- Google Scholar
21. Szklarczyk D, Kirsch R, Koutrouli M, Nastou K, Mehryary F, Hachilif R, et al. The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res. 2023;51(D1):D638–46. pmid:36370105
- View Article
- PubMed/NCBI
- Google Scholar
22. Webb B, Sali A. Protein structure modeling with MODELLER. Springer. 2021.
23. Burley SK, Bhikadiya C, Bi C, Bittrich S, Chao H, Chen L, et al. RCSB Protein Data Bank (RCSB.org): delivery of experimentally-determined PDB structures alongside one million computed structure models of proteins from artificial intelligence/machine learning. Nucleic Acids Res. 2023;51(D1):D488–508. pmid:36420884
- View Article
- PubMed/NCBI
- Google Scholar
24. Koropatkin NM, Martens EC, Gordon JI, Smith TJ. Starch catabolism by a prominent human gut symbiont is directed by the recognition of amylose helices. Structure. 2008;16(7):1105–15. pmid:18611383
- View Article
- PubMed/NCBI
- Google Scholar
25. Jorgensen WL, Maxwell DS, Tirado-Rives J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J Am Chem Soc. 1996;118(45):11225–36.
- View Article
- Google Scholar
26. Schrödinger L, et al. Protein preparation wizard. Epik version 2011;2.
- View Article
- Google Scholar
27. Inc CCG. Molecular operating environment (MOE). Chem Comput Gr. 2016;1010.
28. Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML. Comparison of simple potential functions for simulating liquid water. J Chem Phys. 1983;79(2):926–35.
- View Article
- Google Scholar
29. Amini M, Forsyth M, Gruenhut S, Meakin P, MacFarlane DR. Equilibration strategies in molecular dynamics simulation of polyethers. Electrochim Acta. 1998;43(10–11):1381–6.
- View Article
- Google Scholar
30. Wang C, Guo H, Bai J, Yu L, Tian F, Zhao J, et al. The roles of different Bacteroides uniformis strains in alleviating DSS-induced ulcerative colitis and related functional genes. Food Funct. 2024;15(7):3327–39. pmid:38465411
- View Article
- PubMed/NCBI
- Google Scholar
31. Knoll RL, Forslund K, Kultima JR, Meyer CU, Kullmer U, Sunagawa S, et al. Gut microbiota differs between children with Inflammatory Bowel Disease and healthy siblings in taxonomic and functional composition: a metagenomic analysis. Am J Physiol Gastrointest Liver Physiol. 2017;312(4):G327–39. pmid:28039159
- View Article
- PubMed/NCBI
- Google Scholar
32. Saarela MH. Safety aspects of next generation probiotics. Current Opinion in Food Science. 2019;30:8–13.
- View Article
- Google Scholar
33. Tuson HH, Foley MH, Koropatkin NM, Biteen JS. The starch utilization system assembles around stationary starch-binding proteins. Biophys J. 2018;115(2):242–50. pmid:29338841
- View Article
- PubMed/NCBI
- Google Scholar
34. Reeves AR, Wang GR, Salyers AA. Characterization of four outer membrane proteins that play a role in utilization of starch by Bacteroides thetaiotaomicron. J Bacteriol. 1997;179(3):643–9. pmid:9006015
- View Article
- PubMed/NCBI
- Google Scholar
35. Liu L, Xu M, Lan R, Hu D, Li X, Qiao L, et al. Bacteroides vulgatus attenuates experimental mice colitis through modulating gut microbiota and immune responses. Front Immunol. 2022;13:1036196. pmid:36531989
- View Article
- PubMed/NCBI
- Google Scholar
36. Cheng J, Hu J, Geng F, Nie S. Bacteroides utilization for dietary polysaccharides and their beneficial effects on gut health. Food Science and Human Wellness. 2022;11(5):1101–10.
- View Article
- Google Scholar
37. D’Elia JN, Salyers AA. Contribution of a neopullulanase, a pullulanase, and an alpha-glucosidase to growth of Bacteroides thetaiotaomicron on starch. J Bacteriol. 1996;178(24):7173–9. pmid:8955399
- View Article
- PubMed/NCBI
- Google Scholar
38. Rakoff-Nahoum S, Foster KR, Comstock LE. The evolution of cooperation within the gut microbiota. Nature. 2016;533(7602):255–9. pmid:27111508
- View Article
- PubMed/NCBI
- Google Scholar
39. Cameron EA. The structure and function of the starch utilization system in Bacteroides thetaiotaomicron. 2014.
40. Shipman JA, Berleman JE, Salyers AA. Characterization of four outer membrane proteins involved in binding starch to the cell surface of Bacteroides thetaiotaomicron. J Bacteriol. 2000;182(19):5365–72.
- View Article
- Google Scholar
41. Koropatkin NM, Smith TJ. SusG: a unique cell-membrane-associated alpha-amylase from a prominent human gut symbiont targets complex starch molecules. Structure. 2010;18(2):200–15. pmid:20159465
- View Article
- PubMed/NCBI
- Google Scholar
42. Kitamura M, Okuyama M, Tanzawa F, Mori H, Kitago Y, Watanabe N, et al. Structural and functional analysis of a glycoside hydrolase family 97 enzyme from Bacteroides thetaiotaomicron. J Biol Chem. 2008;283(52):36328–37. pmid:18981178
- View Article
- PubMed/NCBI
- Google Scholar
43. Vinolo MAR, Rodrigues HG, Nachbar RT, Curi R. Regulation of inflammation by short chain fatty acids. Nutrients. 2011;3(10):858–76. pmid:22254083
- View Article
- PubMed/NCBI
- Google Scholar
44. Clausen MR, Mortensen PB. Kinetic studies on the metabolism of short-chain fatty acids and glucose by isolated rat colonocytes. Gastroenterology. 1994;106(2):423–32. pmid:8299908
- View Article
- PubMed/NCBI
- Google Scholar
45. Smith PM, Howitt MR, Panikov N, Michaud M, Gallini CA, Bohlooly-Y M, et al. The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science. 2013;341(6145):569–73. pmid:23828891
- View Article
- PubMed/NCBI
- Google Scholar
46. Mortensen PB, Clausen MR. Short-chain fatty acids in the human colon: relation to gastrointestinal health and disease. Scand J Gastroenterol. 1996;31(sup216):132–48.
- View Article
- Google Scholar
47. Sahoo DK, Heilmann RM, Paital B, Patel A, Yadav VK, Wong D, et al. Oxidative stress, hormones, and effects of natural antioxidants on intestinal inflammation in inflammatory bowel disease. Front Endocrinol (Lausanne). 2023;14:1217165. pmid:37701897
- View Article
- PubMed/NCBI
- Google Scholar
48. Kunst C, Schmid S, Michalski M, Tümen D, Buttenschön J, Müller M, et al. The influence of gut microbiota on oxidative stress and the immune system. Biomedicines. 2023;11(5):1388. pmid:37239059
- View Article
- PubMed/NCBI
- Google Scholar
49. Huang C-H, Chen Y-W, Huang T-T, Kao Y-T. Effects of distal mutations on ligand-binding affinity in E. coli dihydrofolate reductase. ACS Omega. 2021;6(40):26065–76. pmid:34660967
- View Article
- PubMed/NCBI
- Google Scholar
50. Camargo PG, et al. In silico evaluation of N-aryl-1, 10-phenanthroline-2-amines as potential inhibitors of T. cruzi GP63 zinc-metalloprotease by docking and molecular dynamics simulations. Sci Rep. 2025;15(1):6036.
- View Article
- Google Scholar

[ref1] 1. Alatab S, et al. The global, regional, and national burden of inflammatory bowel disease in 195 countries and territories, 1990--2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Gastroenterol Hepatol. 2020;5(1):17–30.
View Article
Google Scholar

[2] View Article

[3] Google Scholar

[ref2] 2. El Hadad J, Schreiner P, Vavricka SR, Greuter T. The genetics of inflammatory bowel disease. Mol Diagn Ther. 2024;28(1):27–35. pmid:37847439
View Article
PubMed/NCBI
Google Scholar

[5] View Article

[6] PubMed/NCBI

[7] Google Scholar

[ref3] 3. Hisamatsu T, Suzuki M, Reinecker H-C, Nadeau WJ, McCormick BA, Podolsky DK. CARD15/NOD2 functions as an antibacterial factor in human intestinal epithelial cells. Gastroenterology. 2003;124(4):993–1000. pmid:12671896
View Article
PubMed/NCBI
Google Scholar

[9] View Article

[10] PubMed/NCBI

[11] Google Scholar

[ref4] 4. Fiocchi C. Omics and multi-omics in IBD: no integration, no breakthroughs. Int J Mol Sci. 2023;24(19):14912. pmid:37834360
View Article
PubMed/NCBI
Google Scholar

[13] View Article

[14] PubMed/NCBI

[15] Google Scholar

[ref5] 5. Schofield JB, Haboubi N. Histopathological mimics of inflammatory bowel disease. Inflamm Bowel Dis. 2019;26(7):994–1009.
View Article
Google Scholar

[17] View Article

[18] Google Scholar

[ref6] 6. Sawyerr AM, Ferguson A. Ulcerative colitis and Crohn’s disease: similarities and distinctions. J R Coll Physicians Lond. 1994;28(3):197–9. pmid:7932315
View Article
PubMed/NCBI
Google Scholar

[20] View Article

[21] PubMed/NCBI

[22] Google Scholar

[ref7] 7. Xie J, Itzkowitz SH. Cancer in inflammatory bowel disease. World J Gastroenterol. 2008;14(3):378–89. pmid:18200660
View Article
PubMed/NCBI
Google Scholar

[24] View Article

[25] PubMed/NCBI

[26] Google Scholar

[ref8] 8. Nadeem MS, Kumar V, Al-Abbasi FA, Kamal MA, Anwar F. Risk of colorectal cancer in inflammatory bowel diseases. Semin Cancer Biol. 2020. p. 51–60.
View Article
Google Scholar

[28] View Article

[29] Google Scholar

[ref9] 9. Bernstein CN, Blanchard JF, Kliewer E, Wajda A. Cancer risk in patients with inflammatory bowel disease: a population-based study. Cancer. 2001;91(4):854–62.
View Article
Google Scholar

[31] View Article

[32] Google Scholar

[ref10] 10. Dahal RH, Choi Y-J, Kim S, Kim J. Differentiation of Escherichia fergusonii and Escherichia coli isolated from patients with inflammatory bowel disease/ischemic colitis and their antimicrobial susceptibility patterns. Antibiotics (Basel). 2023;12(1):154. pmid:36671355
View Article
PubMed/NCBI
Google Scholar

[34] View Article

[35] PubMed/NCBI

[36] Google Scholar

[ref11] 11. Liang G. Altered gut bacterial and metabolic signatures and their interaction in inflammatory bowel disease. Synth Syst Biotechnol. 2021;6(4):377–83. pmid:34820530
View Article
PubMed/NCBI
Google Scholar

[38] View Article

[39] PubMed/NCBI

[40] Google Scholar

[ref12] 12. Alsulaiman RM, Al-Quorain AA, Al-Muhanna FA, Piotrowski S, Kurdi EA, Vatte C, et al. Gut microbiota analyses of inflammatory bowel diseases from a representative Saudi population. BMC Gastroenterol. 2023;23(1):258. pmid:37507685
View Article
PubMed/NCBI
Google Scholar

[42] View Article

[43] PubMed/NCBI

[44] Google Scholar

[ref13] 13. Mills RH, Dulai PS, Vázquez-Baeza Y, Sauceda C, Daniel N, Gerner RR, et al. Multi-omics analyses of the ulcerative colitis gut microbiome link Bacteroides vulgatus proteases with disease severity. Nat Microbiol. 2022;7(2):262–76. pmid:35087228
View Article
PubMed/NCBI
Google Scholar

[46] View Article

[47] PubMed/NCBI

[48] Google Scholar

[ref14] 14. Kang D-Y, Park J-L, Yeo M-K, Kang S-B, Kim J-M, Kim JS, et al. Diagnosis of Crohn’s disease and ulcerative colitis using the microbiome. BMC Microbiol. 2023;23(1):336. pmid:37951857
View Article
PubMed/NCBI
Google Scholar

[50] View Article

[51] PubMed/NCBI

[52] Google Scholar

[ref15] 15. Andrews S, et al. FastQC: a quality control tool for high throughput sequence data. Cambridge, United Kingdom. 2010.
View Article
Google Scholar

[54] View Article

[55] Google Scholar

[ref16] 16. Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884–90. pmid:30423086
View Article
PubMed/NCBI
Google Scholar

[57] View Article

[58] PubMed/NCBI

[59] Google Scholar

[ref17] 17. Bushnell B. BBMap: a fast, accurate, splice-aware aligner. 2014.

[ref18] 18. Beghini F, McIver LJ, Blanco-Míguez A, Dubois L, Asnicar F, Maharjan S, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife. 2021;10:e65088. pmid:33944776
View Article
PubMed/NCBI
Google Scholar

[62] View Article

[63] PubMed/NCBI

[64] Google Scholar

[ref19] 19. Mallick H, Rahnavard A, McIver LJ, Ma S, Zhang Y, Nguyen LH, et al. Multivariable association discovery in population-scale meta-omics studies. PLoS Comput Biol. 2021;17(11):e1009442. pmid:34784344
View Article
PubMed/NCBI
Google Scholar

[66] View Article

[67] PubMed/NCBI

[68] Google Scholar

[ref20] 20. Seemann T. Snippy: rapid haploid variant calling and core SNP phylogeny. https://github.com/tseemann/snippy. 2015.
View Article
Google Scholar

[70] View Article

[71] Google Scholar

[ref21] 21. Szklarczyk D, Kirsch R, Koutrouli M, Nastou K, Mehryary F, Hachilif R, et al. The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res. 2023;51(D1):D638–46. pmid:36370105
View Article
PubMed/NCBI
Google Scholar

[73] View Article

[74] PubMed/NCBI

[75] Google Scholar

[ref22] 22. Webb B, Sali A. Protein structure modeling with MODELLER. Springer. 2021.

[ref23] 23. Burley SK, Bhikadiya C, Bi C, Bittrich S, Chao H, Chen L, et al. RCSB Protein Data Bank (RCSB.org): delivery of experimentally-determined PDB structures alongside one million computed structure models of proteins from artificial intelligence/machine learning. Nucleic Acids Res. 2023;51(D1):D488–508. pmid:36420884
View Article
PubMed/NCBI
Google Scholar

[78] View Article

[79] PubMed/NCBI

[80] Google Scholar

[ref24] 24. Koropatkin NM, Martens EC, Gordon JI, Smith TJ. Starch catabolism by a prominent human gut symbiont is directed by the recognition of amylose helices. Structure. 2008;16(7):1105–15. pmid:18611383
View Article
PubMed/NCBI
Google Scholar

[82] View Article

[83] PubMed/NCBI

[84] Google Scholar

[ref25] 25. Jorgensen WL, Maxwell DS, Tirado-Rives J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J Am Chem Soc. 1996;118(45):11225–36.
View Article
Google Scholar

[86] View Article

[87] Google Scholar

[ref26] 26. Schrödinger L, et al. Protein preparation wizard. Epik version 2011;2.
View Article
Google Scholar

[89] View Article

[90] Google Scholar

[ref27] 27. Inc CCG. Molecular operating environment (MOE). Chem Comput Gr. 2016;1010.

[ref28] 28. Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML. Comparison of simple potential functions for simulating liquid water. J Chem Phys. 1983;79(2):926–35.
View Article
Google Scholar

[93] View Article

[94] Google Scholar

[ref29] 29. Amini M, Forsyth M, Gruenhut S, Meakin P, MacFarlane DR. Equilibration strategies in molecular dynamics simulation of polyethers. Electrochim Acta. 1998;43(10–11):1381–6.
View Article
Google Scholar

[96] View Article

[97] Google Scholar

[ref30] 30. Wang C, Guo H, Bai J, Yu L, Tian F, Zhao J, et al. The roles of different Bacteroides uniformis strains in alleviating DSS-induced ulcerative colitis and related functional genes. Food Funct. 2024;15(7):3327–39. pmid:38465411
View Article
PubMed/NCBI
Google Scholar

[99] View Article

[100] PubMed/NCBI

[101] Google Scholar

[ref31] 31. Knoll RL, Forslund K, Kultima JR, Meyer CU, Kullmer U, Sunagawa S, et al. Gut microbiota differs between children with Inflammatory Bowel Disease and healthy siblings in taxonomic and functional composition: a metagenomic analysis. Am J Physiol Gastrointest Liver Physiol. 2017;312(4):G327–39. pmid:28039159
View Article
PubMed/NCBI
Google Scholar

[103] View Article

[104] PubMed/NCBI

[105] Google Scholar

[ref32] 32. Saarela MH. Safety aspects of next generation probiotics. Current Opinion in Food Science. 2019;30:8–13.
View Article
Google Scholar

[107] View Article

[108] Google Scholar

[ref33] 33. Tuson HH, Foley MH, Koropatkin NM, Biteen JS. The starch utilization system assembles around stationary starch-binding proteins. Biophys J. 2018;115(2):242–50. pmid:29338841
View Article
PubMed/NCBI
Google Scholar

[110] View Article

[111] PubMed/NCBI

[112] Google Scholar

[ref34] 34. Reeves AR, Wang GR, Salyers AA. Characterization of four outer membrane proteins that play a role in utilization of starch by Bacteroides thetaiotaomicron. J Bacteriol. 1997;179(3):643–9. pmid:9006015
View Article
PubMed/NCBI
Google Scholar

[114] View Article

[115] PubMed/NCBI

[116] Google Scholar

[ref35] 35. Liu L, Xu M, Lan R, Hu D, Li X, Qiao L, et al. Bacteroides vulgatus attenuates experimental mice colitis through modulating gut microbiota and immune responses. Front Immunol. 2022;13:1036196. pmid:36531989
View Article
PubMed/NCBI
Google Scholar

[118] View Article

[119] PubMed/NCBI

[120] Google Scholar

[ref36] 36. Cheng J, Hu J, Geng F, Nie S. Bacteroides utilization for dietary polysaccharides and their beneficial effects on gut health. Food Science and Human Wellness. 2022;11(5):1101–10.
View Article
Google Scholar

[122] View Article

[123] Google Scholar

[ref37] 37. D’Elia JN, Salyers AA. Contribution of a neopullulanase, a pullulanase, and an alpha-glucosidase to growth of Bacteroides thetaiotaomicron on starch. J Bacteriol. 1996;178(24):7173–9. pmid:8955399
View Article
PubMed/NCBI
Google Scholar

[125] View Article

[126] PubMed/NCBI

[127] Google Scholar

[ref38] 38. Rakoff-Nahoum S, Foster KR, Comstock LE. The evolution of cooperation within the gut microbiota. Nature. 2016;533(7602):255–9. pmid:27111508
View Article
PubMed/NCBI
Google Scholar

[129] View Article

[130] PubMed/NCBI

[131] Google Scholar

[ref39] 39. Cameron EA. The structure and function of the starch utilization system in Bacteroides thetaiotaomicron. 2014.

[ref40] 40. Shipman JA, Berleman JE, Salyers AA. Characterization of four outer membrane proteins involved in binding starch to the cell surface of Bacteroides thetaiotaomicron. J Bacteriol. 2000;182(19):5365–72.
View Article
Google Scholar

[134] View Article

[135] Google Scholar

[ref41] 41. Koropatkin NM, Smith TJ. SusG: a unique cell-membrane-associated alpha-amylase from a prominent human gut symbiont targets complex starch molecules. Structure. 2010;18(2):200–15. pmid:20159465
View Article
PubMed/NCBI
Google Scholar

[137] View Article

[138] PubMed/NCBI

[139] Google Scholar

[ref42] 42. Kitamura M, Okuyama M, Tanzawa F, Mori H, Kitago Y, Watanabe N, et al. Structural and functional analysis of a glycoside hydrolase family 97 enzyme from Bacteroides thetaiotaomicron. J Biol Chem. 2008;283(52):36328–37. pmid:18981178
View Article
PubMed/NCBI
Google Scholar

[141] View Article

[142] PubMed/NCBI

[143] Google Scholar

[ref43] 43. Vinolo MAR, Rodrigues HG, Nachbar RT, Curi R. Regulation of inflammation by short chain fatty acids. Nutrients. 2011;3(10):858–76. pmid:22254083
View Article
PubMed/NCBI
Google Scholar

[145] View Article

[146] PubMed/NCBI

[147] Google Scholar

[ref44] 44. Clausen MR, Mortensen PB. Kinetic studies on the metabolism of short-chain fatty acids and glucose by isolated rat colonocytes. Gastroenterology. 1994;106(2):423–32. pmid:8299908
View Article
PubMed/NCBI
Google Scholar

[149] View Article

[150] PubMed/NCBI

[151] Google Scholar

[ref45] 45. Smith PM, Howitt MR, Panikov N, Michaud M, Gallini CA, Bohlooly-Y M, et al. The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science. 2013;341(6145):569–73. pmid:23828891
View Article
PubMed/NCBI
Google Scholar

[153] View Article

[154] PubMed/NCBI

[155] Google Scholar

[ref46] 46. Mortensen PB, Clausen MR. Short-chain fatty acids in the human colon: relation to gastrointestinal health and disease. Scand J Gastroenterol. 1996;31(sup216):132–48.
View Article
Google Scholar

[157] View Article

[158] Google Scholar

[ref47] 47. Sahoo DK, Heilmann RM, Paital B, Patel A, Yadav VK, Wong D, et al. Oxidative stress, hormones, and effects of natural antioxidants on intestinal inflammation in inflammatory bowel disease. Front Endocrinol (Lausanne). 2023;14:1217165. pmid:37701897
View Article
PubMed/NCBI
Google Scholar

[160] View Article

[161] PubMed/NCBI

[162] Google Scholar

[ref48] 48. Kunst C, Schmid S, Michalski M, Tümen D, Buttenschön J, Müller M, et al. The influence of gut microbiota on oxidative stress and the immune system. Biomedicines. 2023;11(5):1388. pmid:37239059
View Article
PubMed/NCBI
Google Scholar

[164] View Article

[165] PubMed/NCBI

[166] Google Scholar

[ref49] 49. Huang C-H, Chen Y-W, Huang T-T, Kao Y-T. Effects of distal mutations on ligand-binding affinity in E. coli dihydrofolate reductase. ACS Omega. 2021;6(40):26065–76. pmid:34660967
View Article
PubMed/NCBI
Google Scholar

[168] View Article

[169] PubMed/NCBI

[170] Google Scholar

[ref50] 50. Camargo PG, et al. In silico evaluation of N-aryl-1, 10-phenanthroline-2-amines as potential inhibitors of T. cruzi GP63 zinc-metalloprotease by docking and molecular dynamics simulations. Sci Rep. 2025;15(1):6036.
View Article
Google Scholar

[172] View Article

[173] Google Scholar

Figures

Abstract

Background

Materials and methods

Data collection and preprocessing

Taxonomic profiling and association analysis

Variant calling, filtration, and annotation

Functional annotation of SNPs

Structure modeling and analysis

Molecular docking

MD simulations

Results

Identification of disease-associated microbial signatures and functional variants

SNP in SusD implicates disruption of the starch utilization pathway of B. uniformis in CD patients

Molecular modeling revealed local structural perturbations in mutant SusD

Docking analysis indicates weakened cyclodextrin binding in mutant SusD

Comparative MD simulations highlight impaired ligand retention in mutant SusD

Discussion

Conclusions

Supporting information

S1 Fig. Bacterial association across phenotypes.

S2 Fig. Variant counts summary.

S3 Fig. Catalogue of microbial genes with reported SNPs.

S4 Fig. Network analysis of starch utilization system proteins.

S5 Fig. Sequence alignment of wild-type and mutant SusD with template sequence.

S6 Fig. Simulation of wild and mutant SusD.

S7 Fig. Structure of cyclodextrin.

S8 Fig. Comparative docking scores.

S9 Fig. Non-bonded interaction patterns before and after simulation in wild-type and mutant complexes.

S10 Fig. RMSF of docked complexes.

S11 Fig. Hydrogen bonding patterns of wild and mutant SusD complex.

S1 Table. Information of metagenomics samples used in this study.

S2 Table. Relative abundances of bacterial species across CD, UC and healthy phenotypes.

S3 Table. Significant associations of bacterial species across CD and UC with reference to healthy controls.

S4 Table. Catalogue of all types of microbial variations present in CD, UC and healthy phenotypes.

S5 Table. Catalogue of prioritized SNPs specific to CD and UC phenotypes.

S6 Table. List of SNPs functionally associated with polysaccharide metabolism.

Acknowledgments

References