DNA-utilization loci enable exogenous DNA metabolism in gut Bacteroidales

Significance

The gut microbiome plays a crucial role in nutrient metabolism, yet the fate of exogenous DNA within this ecosystem remains poorly understood. This study identifies Bacteroidales species that actively metabolize exogenous DNA, revealing a conserved pathway that converts DNA-derived nucleotides into deaminated nucleobases. We show that Bacteroides thetaiotaomicron utilizes a specialized genetic locus, ddbABCDEF, to facilitate this process, contributing to nucleobase production in a defined colonization model. Comparative genomic analyses suggest that ddbABCDEF is evolutionarily linked to bacterial natural transformation systems but has diverged into distinct metabolic subtypes. These findings establish DNA as a metabolic substrate in the gut microbiome, with potential implications for microbial ecology, host–microbe interactions, and gut health.

Abstract

The human gut microbiome plays a central role in nutrient metabolism, yet the fate of exogenous nucleic acids within this ecosystem remains poorly understood. Here, we show that multiple Bacteroidales species efficiently metabolize exogenous DNA, with Bacteroides thetaiotaomicron converting it into the deaminated nucleobases uracil and xanthine. Using genetic and biochemical approaches, we identify ddbABCDEF, a six-gene locus encoding secreted nucleases and an outer membrane transporter, essential for exogenous DNA metabolism in B. thetaiotaomicron. Colonization of gnotobiotic mice with ddbABCDEF mutants reveals that this pathway significantly alters nucleobase pools in a gnotobiotic mouse model. Comparative genomic analyses demonstrate that ddbABCDEF is evolutionarily related to a natural transformation system present in Bacteroidota and has diversified into four distinct subtypes, each linked to unique DNA-processing activities in closely related gut Bacteroidales strains. These findings thus expand our understanding of DNA metabolism in the gut microbiome and reveal a distinctive pathway for nucleobase production with implications for host–microbe interactions.

The human gut microbiome comprises a complex community of microorganisms that play pivotal roles in host physiology. A key mechanism by which the gut microbiome impacts host health is through the metabolism of dietary- and host-derived compounds, products of which include bioactive metabolites that influence systemic functions (1). While microbial metabolism of carbohydrates and proteins has been extensively studied, the processing of dietary nucleic acids by the gut microbiome remains underexplored.

Nucleic acids are abundant in plant- and animal-derived foods and undergo partial enzymatic digestion in the host gastrointestinal tract, yielding nucleotides, nucleosides, and nucleobases (2). While these molecules are partially absorbed in the small intestine, some fraction reaches the lower gastrointestinal tract, where the microbial density is highest (3, 4). Additionally, host-derived nucleic acids, including DNA from sloughed epithelial cells, infiltrating immune cells, and lysed microbial cells, further contribute to the pool of exogenous nucleic acids available for microbial metabolism (5, 6). The presence of dietary- and host-derived DNA in fecal matter confirms that nucleic acids remain a persistent resource available for microbial breakdown in the gut (3, 4, 7, 8).

Recent studies have implicated nucleic acids as an important resource for the gut microbiome. The RNA backbone sugar ribose serves as a growth substrate for Bacteroides thetaiotaomicron, an abundant gut symbiont, in the gastrointestinal environment (9). Furthermore, microbial nucleic acid metabolism generates bioactive compounds with significant systemic effects. Microbiome-derived nucleosides inosine and guanosine modulate immune function, enhancing cancer immunotherapy outcomes and promoting tolerance against food allergens, respectively (10, 11). Similarly, nucleobases such as xanthine and hypoxanthine have been shown to promote intestinal wound healing and reinforce mucosal barrier function through anti-inflammatory mechanisms that may be relevant for irritable bowel syndrome (12–14). Gut nucleobases also contribute to pathogenesis of Clostridioides difficile, a major enteric pathogen (15). Collectively, these findings indicate that microbial nucleic acid metabolism plays multifaceted roles in shaping microbial community dynamics and influencing host physiology within the intestinal ecosystem.

In this study, we investigate mechanisms of exogenous DNA metabolism by gut bacteria. Through genetic, biochemical, and metabolomic approaches, we demonstrate that B. thetaiotaomicron utilizes a locus, designated ddbABCDEF, to mediate exogenous DNA degradation and the conversion of DNA nucleotides into deaminated nucleobases, such as xanthine and uracil. Comparative genomic analyses reveal that the ddbABCDEF locus is part of a larger family of functionally diverse nucleic acid processing systems widespread among gut Bacteroidales, with distinct subtypes linked to unique DNA-depleting activities. These findings define a class of DNA metabolism systems in the gut microbiome and offer new insights into the ecological and metabolic roles of microbial DNA processing in the mammalian gut.

Results

Human Fecal Microbial Communities and Gut Bacteroides Deplete Exogenous DNA.

To determine whether human gut microbial communities metabolize exogenous DNA, we cultured fecal samples from five different donors under anaerobic conditions with herring sperm DNA or synthetic oligonucleotides and monitored DNA depletion over time. DNA levels declined progressively in all fecal cultures tested, but the rate of depletion varied considerably between donors (Fig. 1 A and B and SI Appendix, Fig. S1). These results demonstrate that the gut microbiome can effectively degrade exogenous nucleic acids, with potential donor-dependent distinctions in the nature of this activity.

Fig. 1.

Human fecal microbial communities and gut Bacteroidales deplete exogenous DNA. (A) Representative agarose gel illustrating progressive degradation of herring sperm DNA over time in a human fecal slurry (donor 1), demonstrating community-level DNA degradation activity. (B) Quantification of DNA depletion kinetics in fecal slurries from five distinct donors. (C) DNA depletion activity of monocultures from taxonomically diverse gut bacterial strains.

As 16S rRNA amplicon sequencing of the fecal microbial communities failed to reveal an obvious link between community composition and DNA-depleting activity (SI Appendix, Fig. S1C), we pursued a targeted approach to identify specific microbial taxa responsible for the observed phenotype. We screened a panel of 35 taxonomically diverse gut bacterial strains for DNA-depleting activity and identified four species (Bacteroides cellulosilyticus, B. thetaiotaomicron, Bacteroides uniformis, and Sutterella wadsworthensis) with substantial DNA-degrading capabilities (Fig. 1C and SI Appendix, Fig. S2). These results establish that exogenous DNA undergoes depletion by gut microbiomes and highlight a role for Bacteroidales in this metabolic activity.

B. thetaiotaomicron ddbABCDEF Is Essential for Exogenous DNA Depletion.

As abundant members of the gut microbiome, Bacteroidales with DNA-depleting capacity could play a significant role in determining the fate of exogenous nucleic acids within this ecosystem. Bacteroidales cultures incubated with herring sperm DNA revealed a gradual decrease in fragment size over time, consistent with extracellular nuclease activity (SI Appendix, Fig. S1). Given that secreted nucleases are commonly involved in DNA metabolism, we tested whether B. thetaiotaomicron DFI.1.47 produced such enzymes. Consistent with secreted factors contributing to DNA degradation, we found that culture supernatants from this strain exhibited robust nuclease activity (Fig. 2A).

Fig. 2.

The ddbABCDEF locus encodes secreted nucleases and is essential for DNA depletion in B. thetaiotaomicron. (A) DNA depletion activity of B. thetaiotaomicron cultures, comparing whole cells and secreted factors. (B) The ddbABCDEF locus. DUF stands for domain of unknown function. (C) DNA degradation activity of purified recombinant proteins encoded by the ddbABCDEF locus. (D) DNA depletion activity of wild-type B. thetaiotaomicron compared to uncomplemented and complemented ΔddbABCDEF mutants. Statistical significance was assessed using one-way ANOVA followed by Dunnett’s post hoc test.

To identify the genetic basis for this activity, we searched the B. thetaiotaomicron DFI.1.47 genome for putative nucleases with predicted extracellular or periplasmic localization. The sole candidate nucleases identified in our analysis localized to a six-gene locus, which we designated ddbABCDEF (for DNA Depletion in Bacteroidales). The locus encodes three predicted nucleases (DdbA, DdbB, and DdbF)—including a homolog of a previously identified variant (ddbF) proposed to encode an extracellular nuclease required for bile-dependent biofilm formation (16)—alongside an outer membrane transporter (DdbC), a hypothetical protein (DdbD), and an additional predicted extracellular protein (DdbE) (Fig. 2B). Consistent with the annotated functions, heterologous expression and in vitro nuclease assays confirmed that DdbA, DdbB, and DdbF exhibit DNA-degrading activity (Fig. 2C).

In gut Bacteroidales, distinct genetic loci encode proteins with degradative enzymes and outer membrane transporters that facilitate uptake of specific complex exogenous polymers, including polypeptides and polysaccharides (17–19). Considering these parallels, we reasoned the ddbABCDEF locus might be relevant for DNA-utilization B. thetaiotaomicron DFI.1.47. To test this hypothesis, we generated and assayed a scarless deletion mutant (ΔddbABCDEF) and a complemented (ΔddbABCDEF + pNBU2-ddbABCDEF) strain. We found the ΔddbABCDEF strain was completely deficient in DNA depletion activity, but that this defect was restored by complementation (Fig. 2D). These results establish that ddbABCDEF is required for exogenous DNA degradation in B. thetaiotaomicron DFI.1.47.

ddbABCDEF Is Required for Deaminated Nucleobase Production from DNA.

We next sought to investigate the physiological significance of B. thetaiotaomicron DNA metabolism. As preliminary experiments revealed that exogenous DNA provided a poor sole carbon, nitrogen, or phosphorus source, we asked whether B. thetaiotaomicron could metabolize DNA into biologically relevant secondary products. We performed targeted metabolomics analyses of B. thetaiotaomicron culture supernatants grown in the presence or absence of exogenous DNA and observed significantly elevated levels of the deaminated nucleobases xanthine and uracil in cultures supplemented with DNA (Fig. 3A).

Fig. 3.

Exogenous DNA fuels production of deaminated nucleobases by B. thetaiotaomicron. (A) Quantification of uracil and xanthine levels in culture supernatants of B. thetaiotaomicron grown with or without exogenous herring sperm DNA. (B) Growth of B. thetaiotaomicron in chemically defined media supplemented with different nitrogen sources. (C) Proposed pathway linking DNA degradation to nucleobase production. (D) Uracil and xanthine levels in the cecal contents of germfree mice gavaged with indicated B. thetaiotaomicron strains or a vehicle control. Statistical significance was assessed using one-way ANOVA followed by Dunnett’s post hoc test.

To clarify the pathway responsible for this conversion, we systematically tested the ability of B. thetaiotaomicron to metabolize DNA-derived nucleotides, nucleosides, and nucleobases. We found that cytosine and guanine nucleotides could serve as sole nitrogen sources, albeit suboptimally, and were readily converted into uracil and xanthine, respectively (Fig. 3 A and B and SI Appendix, Fig. S3). Notably, deletion of ddbABCDEF abolished uracil and xanthine production from exogenous DNA but not from deoxycytidine and deoxyguanosine nucleosides (Fig. 3A and SI Appendix, Fig. S4). This finding indicates that ddbABCDEF is specifically required for exogenous DNA uptake and that other presently unidentified downstream enzymes are required for subsequent reactions to generated deaminated nucleobases (Fig. 3C).

ddbABCDEF Promotes Deaminated Nucleobase Production in the Mouse Gut.

To test whether ddbABCDEF-mediated DNA metabolism contributes to nucleobase production in vivo, we colonized germ-free mice with wild-type or ΔddbABCDEF mutant B. thetaiotaomicron strains and quantified uracil and xanthine levels in the cecum. Monocolonization with wild-type B. thetaiotaomicron resulted in a significant increase in cecal uracil and xanthine concentrations relative to germ-free controls (Fig. 3D). Compared to wild-type B. thetaiotaomicron, ΔddbABCDEF monocolonization resulted in a significantly lower cecal concentration of these metabolites (Fig. 3D). These results provide evidence that B. thetaiotaomicron generates deaminated nucleobases in the gut of monocolonized mice and that ddbABCDEF is essential for maximal flux through this pathway.

ddbABCDEF Is Related to the Riemerella anatipestifer Natural Transformation System.

Having identified a role for the ddbABCDEF locus in exogenous DNA metabolism, we next sought to investigate the functions of nonnuclease proteins encoded within the locus. Natural transformation systems enable bacteria to import and integrate foreign genetic material, thereby serving as a mechanism for horizontal gene transfer. We observed that DdbC and DdbD share homology with proteins RA0C_RS04915 and RA0C_RS04920, which were recently identified as essential components of the natural transformation machinery in the bird pathogen Riemerella anatipestifer and were proposed to facilitate DNA transport across the outer membrane in this Bacteroidota species (20). Further examination revealed that these R. anatipestifer genes neighbored a gene with sequence homology to the ddbB nuclease, thus presenting a locus strikingly similar to ddbBCD in B. thetaiotaomicron (Fig. 4A).

Fig. 4.

The ddbABCDEF locus resembles both natural transformation and polysaccharide-utilization loci. (A) Schematic comparison of the ddbABCDEF locus with a homologous gene cluster in Riemerella anatipestifer involved in natural transformation. Percentages indicate amino acid sequence identity between homologous proteins. Names refer to locus tags RA0C_RS049XX. (B) AlphaFold-predicted structural model of the DdbC–DdbD complex. The gray box indicates the approximate position of the outer membrane (OM). (C) Structural alignment of DdbD with the DNA-binding domain of murine p202 complex with double-stranded DNA (PDB: 4JBM). (D) Comparison of the ddbABCDEF locus with the B. thetaiotaomicron starch-utilization system (susABCDEFG), the archetypal polysaccharide-utilization loci. Signal peptide classifications (SPI vs. SPII) indicate the predicted localization of encoded proteins.

The conserved synteny of R. anatipestifer natural transformation genes with DdbC and DdbD led us to interrogate the relationship between these proteins. DdbC is annotated as a transmembrane subunit of a TonB-dependent outer membrane transporter, while DdbD is annotated as a domain of unknown function (DUF5689). Bacteroidales TonB-dependent transporters often include an extracellular substrate-binding subunit encoded by a gene directly adjacent to the transmembrane subunit (19, 21). Based on this organization, we hypothesized that DdbD may act as the substrate-binding partner for DdbC.

AlphaFold protein structural modeling was recently shown to enable detection of interactions between Bacteroides TonB–dependent transporter transmembrane and substrate-binding subunits (22). Consistent with DdbC and DdbD forming a heterodimeric complex, AlphaFold3 predicted a high confidence DdbC/DdbD interface that reasonably positioned DdbD for a potential substrate-binding role (Fig. 4B). Despite possessing low levels of sequence identity, we further observed that DdbD AlphaFold structure exhibited high structural similarity to previously characterized DNA-binding proteins, including the DNA-binding domain of p202, a murine immune protein that nonspecifically binds double-stranded DNA (23) (Fig. 4C). Collectively, these observations are consistent with DdbCD being part of a structurally distinctive family of outer membrane transporters associated with natural transformation and DNA metabolism in bacteria.

ddbABCDEF Resembles Gut Bacteroidales Polysaccharide-Utilization Loci.

Numerous polysaccharide-utilization loci (PULs) that encode proteins responsible for metabolism of a specific polysaccharides have been identified in gut Bacteroidales (reviewed in ref. 17). PULs typically encode a) extracellular polysaccharide-binding proteins that position the polysaccharide for hydrolysis, b) extracellular glycoside hydrolases that hydrolyze polysaccharides to oligosaccharides, c) a TonB-dependent outer membrane transporter that transports oligosaccharides into the periplasm, and d) periplasmic glycoside hydrolases that hydrolyze oligosaccharides into mono- and disaccharides.

Within PULs, the distinct cellular localization of proteins is often reflected in the signal peptides that govern their secretion. Periplasmic proteins generally contain an SPI signal peptide that results in retention in the periplasm, whereas extracellular proteins typically contain an SPII signal peptide that results in lipid-mediated anchoring to the outer membrane (17). Lipoprotein export signals further mark a subset of extracellular proteins for packaging into secreted outer membrane vesicles (24–26).

As illustrated by a comparison to the emblematic starch PUL susABCDEFG, the ddbABCDEF loci is notable for its PUL-like organization, but differing in the nature of the polymer-hydrolyzing enzymes (Fig. 4D). DdbA and DdbF have SPII signal peptides and hydrolyze nucleic acids, analogous to the extracellular polysaccharide hydrolase SusG. DdbCD is a TonB-dependent transporter that likely transports products generated from extracellular enzymes into the periplasm, analogous to SusCD. DdbB has a SPI signal peptide and hydrolyzes nucleic acids, analogous to periplasmic oligosaccharide hydrolases SusA and SusB. Finally, DdbE has a SPII and encodes a domain of unknown function that could engage in substrate-binding properties, analogous to SusEF. Collectively, these analyses thus demonstrate parallels between ddbABCDEF and both natural transformation and polymer-degrading Bacteroidales loci.

Gut Bacteroidales Encode Multiple ddb Subtypes.

To assess the distribution of ddbABCDEF-like loci, we performed a comparative genomic analysis across diverse members of the phylum Bacteroidota. Consistent with the DNA-depleting activity observed for multiple Bacteroidales strains in our initial screen (Fig. 1C), bioinformatics analyses identified numerous loci across multiple species anchored by the presence of a DdbC transporter homolog (Fig. 5A). These loci consistently encoded annotated secreted and/or cytosolic nucleases, suggesting a conserved relevance for exogenous nucleic acids (Fig. 5A).

Fig. 5.

Distinct ddbABCDEF-like subtypes are present in gut Bacteroidales and confer variable DNA-depleting activities. (A) Representative ddbABCDEF-like subtype gene clusters from B. thetaiotaomicron DFI.1.47 (Bt1), B. thetaiotaomicron MSK.14.51 (Bt2), and P. vulgatus DFI.1.83 (Pv1) and P. vulgatus DFI.4.110 (Pv2). Percentages indicate that amino acid sequence identity between homologous proteins. (B) Heatmap illustrating amino acid sequence identity among DdbC homologs from Bt1, Bt2, Pv1, and Pv2 ddbABCDEF-like gene clusters. (C) Distribution of ddbABCDEF-like subtypes across gut Bacteroidales genomes. (D) Representative agarose gel demonstrating activity of B. thetaiotaomicron and P. vulgatus strains with indicated ddbABCDEF subtypes incubated with herring sperm DNA.

Further analyses revealed that ddbABCDEF-like loci could be subdivided into two types based on DdbC protein sequence homology. DdbC homologs from B. thetaiotaomicron and Phocaeicola vulgatus exemplify these two types and share approximately 25% sequence identity. B. thetaiotaomicron and P. vulgatus types could be further parsed into subtypes that share less than 50% sequence identity, which we designate as Bt1 and Bt2 (for B. thetaiotaomicron 1 and 2) and Pv1 and Pv2 (for P. vulgatus 1 and 2) subtypes (Fig. 5B).

The distribution of ddb subtypes did not generally track with taxonomy and exhibited a striking pattern of mutual exclusivity across genomes. While some genomes encoded both Bt- and Pv-type loci, different Bt- and Pv-subtypes (such as Bt1 and Bt2 or Pv1 and Pv2) were rarely found in the same genome (Dataset S1). This exclusivity was especially apparent when comparing strains of a single species. Among 525 B. thetaiotaomicron genomes analyzed, 208 encoded Bt1 and 317 encoded Bt2, but none encoded both (Fig. 5C and Dataset S1). Similarly, among 495 P. vulgatus genomes analyzed, 203 encoded Pv1 and 292 encoded Pv2, but none encoded both (Fig. 5C and Dataset S1). Notably, this exclusivity could reflect whole loci substitution events, as the different subtypes consistently localized to the same genomic region within each species (SI Appendix, Fig. S5 and Datasets S2 and S3). We further identified five additional species with comparable patterns of variable strain-level mutual exclusivity across multiple subtypes (Fig. 5C and Dataset S1).

ddb Subtypes Are Associated with Distinct DNA-Depleting Activities in Closely Related Gut Bacteroidales.

The presence of multiple ddb subtypes across gut Bacteroidales raises the possibility that these variants encode functionally distinct DNA-processing activities. To assess this, we evaluated the DNA-depleting capacity of B. thetaiotaomicron and P. vulgatus strains carrying different ddb subtypes. Across tested strains, those encoding the same subtype exhibited consistent DNA degradation phenotypes, but striking differences emerged between subtypes. Strains harboring the Bt1 subtype fully degraded exogenous DNA, while Bt2 strains exhibited only partial degradation (Fig. 5D). Similarly, Pv1 strains partially depleted DNA, whereas Pv2 strains—which lack a secreted nuclease within their ddbABCDEF-like locus—exhibited minimal DNA-depleting activity (Fig. 5 A and D). These results suggest that ddb subtypes have diverged functionally, with some variants more effective at degrading exogenous DNA than others.

Discussion

This study highlights the role of B. thetaiotaomicron and related gut Bacteroidales in the metabolism of exogenous DNA. Our findings demonstrate that the ddbABCDEF locus encodes multiple nucleases and a predicted TonB-dependent transporter and is essential for DNA degradation and deaminated nucleobase production in the mouse gut. These components presumably operate in concert to mediate exogenous DNA cleavage and transport, enabling B. thetaiotaomicron to metabolize DNA to deaminated nucleobases.

A particularly intriguing feature of the ddbABCDEF locus is its evolutionary and functional resemblance to the natural transformation machinery of R. anatipestifer, a member of the Flavobacteriales within the Bacteroidota phylum. Unlike Flavobacteriales, gut Bacteroidales are not known to be naturally competent. Consistent with this, our attempts to detect DNA uptake in a diverse set of gut Bacteroidales yielded inconsistent results. This suggests that while ddbABCDEF may share a common ancestral origin with natural transformation systems, it has since diverged functionally.

Notably, a comparable functional versatility exists in Pseudomonadota, where a pilus-based apparatus facilitates DNA transport across the outer membrane (27). In Pseudomonadota, this machinery is used for either natural transformation or DNA-based nutrient acquisition, depending on species-specific adaptations and environmental context (27–29). The characterization of ddbABCDEF thus contributes to a growing body of evidence indicating that phylogenetically diverse bacteria have evolved distinct yet functionally analogous mechanisms to acquire and metabolize exogenous DNA.

Existence of multiple ddbABCDEF subtypes further underscores the functional diversity of DNA processing systems in gut Bacteriodota. The observed mutual exclusivity of subtypes across strains and the association of this with distinct DNA-depleting activities strongly suggests that unknown factors functionally differentiate ddbABCDEF subtypes. These findings thus provide evidence that significant gaps remain in our understanding of the full functional potential of exogenous DNA uptake and processing by gut Bacteriodota, which could be an interesting topic for future study.

Finally, our findings also raise questions about the ecological and host-relevant impacts of microbial DNA metabolism in the gut. While future studies incorporating more complex microbial communities will be needed to fully elucidate the relative contribution of Bacteroidales DNA metabolism to the overall nucleobase landscape of the gut, the production of deaminated nucleobases, such as xanthine and uracil, within the mouse gastrointestinal tract may have implications for downstream effects on host physiology, including immune regulation and epithelial cell regeneration (12–14). While these molecules could be generated via multiple routes, our findings establish the microbial metabolism of exogenous DNA as a plausible source of bioactive nucleobase-derived metabolites within the gut.

Materials and Methods

Bacterial Strains and Growth Conditions.

Bacterial strains used in this study were acquired from either the American Type Culture Collection (ATCC, Manassas, Virginia) or the Duchossois Family Institute (DFI) Biobank strain collection (University of Chicago, Illinois) and are listed in SI Appendix, Table S1 (30, 31). Strains were routinely cultured under anaerobic conditions in a Coy vinyl anaerobic chamber (2 to 5%H₂, 2 to 5% CO₂, N₂ balance) at 37 °C. Growth media included Brain Heart Infusion Medium with Hemin (BHIM), basal, and chemically defined media. The chemically defined medium contained 1 g/L NH₄Cl, 6 g/L Na₂HPO₄, 3 g/L KH₂PO₄, 0.5 g/L NaCl, 14.7 mg/L CaCl₂, 246 mg/L MgSO₄·7H₂O, 0.5% glucose, 0.05% L-cysteine, 2.5 mg/L vitamin K3, 2 mg/L FeSO₄, 5 mg/L hemin, 5 ng/mL vitamin B12. Basal medium was formulated as previously described and contained 19.2 g/L Na₂HPO₄·7H₂O, 4.5 g/L KH₂PO₄, 0.75 g/L NaCl, 1.5 g/L NH₄Cl, 2 g/L sodium acetate, 1.5 g/L sodium formate, 0.1 g/L tryptone, 0.1 g/L Bacto yeast extract, 1 g/L MgSO₄, and an 1x stock of minerals and vitamins and was adjusted to pH 6.5 (32). The 1× minerals and vitamins stock solution contained 0.1 mg/L FeCl₂·4H₂O, 0.846 mg/L MnSO₄·H₂O, 0.028 mg/L ZnSO₄·7H₂O, 0.148 mg/L CaCl₂·2H₂O, 0.002 mg/L CuSO₄·5H₂O, 0.002 mg/L CoCl₂·7H₂O, 0.002 mg/L H₃BO₃, 0.002 mg/L Na₂MoO₄·2H₂O, 50 mg/L NaCl, 1.2 mg/L tri-sodium citrate, 0.05 mg/L biotin, 0.1 mg/L D-pantothenic acid, 0.05 mg/L lipoic acid, 0.1 mg/L niacinamide, 0.1 mg/L para-aminobenzoic acid, 0.1 mg/L pyridoxal HCl, 0.05 mg/L riboflavin, 0.1 mg/L thiamine HCl, and 0.01 mg/L vitamin B12. Anaerobic growth curves were obtained by shaking cells at 200 rpm and measuring the optical density at 600 nm (OD₆₀₀) every 20 min for 72 h, using a LogPhase 600 (BioTek) plate reader.

Human Fecal Microbial Community DNA Depletion Assays.

Fecal samples from five healthy, antibiotic-free individuals were collected under an IRB-approved protocol (University of Chicago Protocol IRB20-1384). All samples were deidentified prior to use in this study in accordance with the approved protocol. Samples were frozen in anaerobic 20% glycerol and stored at −80 °C until use. Thawed samples were washed in phosphate-buffered saline (PBS) and resuspended in basal medium supplemented with 0.5 mg/mL herring sperm DNA (Sigma D3159) or random sequence 60-bp synthetic oligonucleotides (Integrated DNA Technologies). Cultures were incubated anaerobically at 37 °C for 96 h, with samples collected at indicated timepoints. Cultures were centrifuged at 5,000×g for 10 min at 4 °C, and 10 μL supernatant was mixed with 6x TriTrack DNA loading dye (Sigma R1161) before loading onto a 1% agarose gel. Electrophoresis was conducted at 120 V for 45 min using a Bio-Rad Sub-Cell GT system. DNA bands were visualized using an iBright 1,500 imaging system, and band intensities were analyzed with ImageJ software (33).

16S Amplicon Sequencing.

The V4–V5 region within the 16S ribosomal RNA (rRNA) gene was amplified using universal bacterial primers—563F (5′-nnnnnnnn-NNNNNNNNNNNN-AYTGGGYDTAAA-GNG-3′) and 926R (5′-nnnnnnnn-NNNNNNNNNNNN-CCGTCAATTYHT-TTRAGT-3′), where “N” represents the barcodes and “n” is an additional nucleotide added to offset primer sequencing. Amplicons were then purified using magnetic beads, then quantified and pooled at equimolar concentrations. Illumina sequencing-compatible combinatorial dual index adapters were ligated onto pooled amplicons using the QIAseq one-step amplicon library kit (Qiagen). Library quality control was performed using Qubit and Tapestation and sequenced on an Illumina MiSeq platform to generate 2 × 250 bp reads, generating 5,000 to 10,000 reads per sample. Raw V4–V5 16S rRNA gene sequence data are demultiplexed and processed through the dada2 pipeline into amplicon sequence variants (ASVs) with minor modifications in R (v4.0.3). Specifically, reads were first trimmed at 210 bp for forward reads and 150 for reverse reads to remove low-quality nucleotides. Chimeras were detected and removed using the default consensus method in the dada2 pipeline. Then, ASVs with length between 300 bp and 360 bp were kept and deemed as high-quality ASVs. The taxonomy of the resultant ASVs was assigned to the genus level using the RDP Classifier (v2.13) with a minimum bootstrap confidence score of 80.

Bacterial Strain DNA Depletion Assays.

Strains grown overnight anaerobically in BHIM were inoculated into basal media supplemented with 0.5 mg/mL DNA and incubated anaerobically at 37 °C for 96 h. DNA depletion was visualized via agarose gel electrophoresis, with band intensities measured in ImageJ software.

Nitrogen Utilization Assays.

For experiments testing sole nitrogen sources, B. thetaiotaomicron DFI.1.47 overnight anaerobic cultures were washed three times in PBS before inoculation into chemically defined medium in which NH₄Cl was replaced with 2 mM nucleic acids, nucleotides, nucleosides, or nucleobases (from Sigma or Fisher Scientific).

B. thetaiotaomicron Genetic Manipulation and Mutant Generation.

A scarless deletion mutant (ΔddbABCDEF) was generated in B. thetaiotaomicron DFI.1.47 using counter-selection with the Bacteroides suicide vector pLGB13 (34). Flanking 1000-bp regions of ddbABCDEF were PCR amplified and cloned into pLGB13. The construct was introduced into B. thetaiotaomicron via conjugation with E. coli S17-λpir. Cointegrants were selected on BHIM plates containing erythromycin (50 μg/mL) and gentamycin (200 μg/mL). Deletion mutants were isolated by counter-selection on anhydrotetracycline (50 ng/mL) and were confirmed by PCR. For complementation studies, the entire ddbABCDEF operon from B. thetaiotaomicron DFI.1.47 was cloned into the expression vector pNBU2 and introduced into the ΔddbABCDEF mutant strain by conjugal transfer (35).

Heterologous Protein Expression and Purification.

His-tagged versions of DdbA, DdbB, DdbD, DdbE, and DdbF were created by cloning each of these genes into pET28a and transforming chemically competent E. coli BL21 (DE3) for expression as previously described (36, 37). Briefly, cultures were grown to OD₆₀₀ 0.7 to 1.0, shaking (RPM 200) at 37° C, before induction with 1 mM isopropyl β-D-1-thiogalactopyranoside and overnight incubation at 20 °C. After which, cells were resuspended in lysis buffer containing 300 mM NaCl, 1 mM dithiothreitol, 10 mM imidazole, 50 mM Tris-HCl (pH 7.5), and lysozyme (50 mg/mL). Resuspended cells were placed in an ice water bath, subjected to eight 30-s sonication pulses, and then centrifugated at 40,000×g for 30 min at 4 °C. His-tagged proteins were purified from the resulting lysate using Profinity IMAC resin, washed, and eluted in 500 mM imidazole buffer. Purity was assessed by SDS-PAGE.

Nuclease Assays.

Nuclease activity was assessed in bacterial supernatants, cell lysates, or his-tagged purified proteins. B. thetaiotaomicron cultures were grown anaerobically at 37 °C for 72 h, centrifuged at 5,000×g, and passed through a 0.22 μm membrane filter. DNA substrates (500 ng/mL) were incubated with 100 μL supernatant, lysate, or 10 μg of purified protein for 8 h at 37 °C. Reactions were analyzed by agarose gel electrophoresis, and band intensity was quantified using ImageJ software.

Metabolite Analysis.

Culture supernatants from B. thetaiotaomicron strains incubated in chemically defined medium with or without 0.5 mg/mL herring sperm DNA or 0.15 mg/mL deoxycytidine and deoxyguanosine were collected after 72 h. Uracil and xanthine concentrations were quantified using gas chromatography-mass spectrometry (GC−MS) and liquid chromatography-mass spectrometry (LC−MS), respectively. Deaminated nucleobases recovered in DNA-supplemented culture supernatants represented approximately 51% of the total nitrogenous bases present in the supplemented DNA (based on the estimated nucleotide composition of herring sperm DNA).

Gnotobiotic Mouse Experiments.

All mouse experiments were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Chicago and were conducted in accordance with ethical regulations for animal research. Germ-free C57BL/6 J mice (5 to 10 wk old) were housed in gnotobiotic isolators at the Gnotobiotic Research Animal Facility (GRAF). Female adult mice were orally gavaged with 200 µL of log-phase cultures of either wild-type B. thetaiotaomicron DFI.1.47 or the ΔddbABCDEF strain. Cecal contents were harvested on day 7 and subjected to metabolomic analysis to quantify uracil and xanthine concentrations, as described above.

Protein Structural Modeling.

To predict the structure of the DdbC/DdbD complex, AlphaFold3 was used with default parameters (38). Input sequences for DdbC and DdbD were retrieved from B. thetaiotaomicron DFI.1.47 and submitted for cofolding analysis. Model confidence was evaluated using the predicted local distance difference test (pLDDT) and predicted aligned error (PAE) scores. Structural alignments were performed using PyMOL (Schrödinger, LLC) and homology comparisons to known DNA-binding proteins were assessed using the DALI server (39).

Comparative Genomics Analyses.

NCBI accessions of DdbABCDEF proteins are as follows: WP_103878007.1 (DdbA), WP_008767132.1 (DdbB), WP_008762511.1 (DdbC), WP_048698952.1 (DdbD), WP_146072611.1 (DdbE), and WP_103878006.1 (DdbF). Putative signal peptides in protein amino acid sequences were identified using SignalP 5.0 (40). Lipoprotein export signals were identified based on the presence of a previously described C(x/xx)(D/E)(D/E) amino acid sequence (with x signifying “any amino acid” and the first cysteine corresponding to the predicted lipidation site) (24–26).

To investigate the evolutionary distribution and diversity of ddbABCDEF-like loci, we conducted comparative genomic analyses across gut Bacteroidales using publicly available genomes from NCBI RefSeq. Distinct subtypes of ddbABCDEF-like loci were identified in B. thetaiotaomicron and P. vulgatus, designated as Bt1 (B. thetaiotaomicron DFI.1.47, NZ_JAJCPU010000061:25232..34119) and Bt2 [B. thetaiotaomicron MSK.14.51, NZ_JAHOLZ010000003:complement(83009..91766)], as well as Pv1 (P. vulgatus MSK.23.57, NZ_JAHOBM010000001:10853..115503) and Pv2 (P. vulgatus MSK.9.20, NZ_JAHPXN010000009:50891..56657). Additionally, we identified a homologous locus in Riemerella anatipestifer ATCC 11845 (NC_017045:1014432..1020133).

To determine the taxonomic distribution of ddb subtypes, protein sequences of DdbC homologs from B. thetaiotaomicron (Bt1: WP_008762511.1, Bt2: WP_055220680.1) and P. vulgatus (Pv1: WP_005843284.1, Pv2: WP_117829583.1) were used as queries (Dataset S4) for BLASTp searches against a database comprising the proteomes of 9,910 Bacteroidota genomes with unambiguous genus and species designations downloaded from NCBI (Dataset S5). Blastp hits with an e-value ≤ 1e-15, with >50% sequence identity and >75% sequence coverage to the query sequences were considered significant.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Dataset S02 (XLSX)

Dataset S03 (XLSX)

Dataset S04 (XLSX)

Dataset S05 (XLSX)

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.