Dual membrane-spanning anti-sigma 2 controls OMV biogenesis and colonization fitness in Bacteroides thetaiotaomicron

ABSTRACT

Bacteroides spp. are gram-negative, gut commensals that shape the enteric landscape by producing outer membrane vesicles (OMVs) that degrade dietary fibers and traffic immunomodulatory biomolecules. Understanding the mechanism behind OMV biogenesis in Bacteroides spp. is necessary to determine their role in the gut. Recent studies showed that mutation of dual membrane-spanning anti-sigma factor 1 (Dma1) increased OMV production in Bacteroides thetaiotaomicron (Bt) by modulating the expression of its downstream regulon. Additional members of the Dma family have been identified, but very little is known regarding their roles in Bt. Here, we investigate the role of Dma2 in controlling OMV biogenesis in Bt. We employ biochemical and proteomic analyses to show that mutation of dma2 increases OMV production. This induction is dependent on the expression of its cognate sigma factor, das2, but the precise mechanism by which dma2 increases OMV biogenesis remains elusive. Transcriptome analyses revealed that Δdma2 displays decreased expression of select polysaccharide utilization loci (PULs) that primarily target host-associated glycans. Follow-up comparative proteomics showed that the PUL repertoire was most impacted in the OMV fraction. In vitro growth assessments confirmed that Δdma2 exhibits delayed growth in the presence of select host-associated glycans. In vivo co-colonization studies in mice revealed that Δdma2 is outcompeted by the wild-type in the gut, which indicates that Dma2 is a key determinant of colonization fitness in Bt. Altogether, these findings expand our knowledge of the Dma family’s role in OMV biogenesis and demonstrate their importance in Bacteroides physiology.

IMPORTANCE

Dual membrane-spanning anti-sigma factors (Dma) are a novel class of regulatory proteins found solely among Bacteroidota. Previous studies demonstrated the importance of Dma1 in vesiculation, but the overall role of the Dma family in Bacteroides physiology remains poorly understood. Here, we show that Dma2 modulates vesiculation and the expression of select polysaccharide utilization loci (PULs) that target host-associated glycans in vitro. Mouse studies revealed that Dma2 is an important fitness determinant in vivo when competing against kin bacteria. This work begins characterizing the multifaceted involvement of Dma2 in OMV biogenesis, PUL regulation, and colonization fitness.

INTRODUCTION

The gut microbiota is the consortium of trillions (~10¹²) of microbes that inhabit the human gastrointestinal tract (1–3). This collection of microbes promotes the proper development of the gut epithelium and immune system through various mechanisms (2). Bacteroides spp. are one of the most abundant genera, making up ~40% of the bacterial species in the human gut. These microbes help maintain intestinal homeostasis by outcompeting select pathogens, breaking down indigestible dietary fibers, producing short-chain fatty acids, and modulating intestinal immunity to reduce inflammation (3–6).

Bacteroides spp. can stably colonize and successfully compete within the gut due to their ability to utilize a diverse array of dietary polysaccharides and host-associated glycans to promote their growth (3, 4). This process is mediated by numerous encoded polysaccharide utilization loci (PULs) that can account for ~20% of the genome in Bacteroides spp. PULs are complex nutrient acquisition systems that sense, degrade, and import polysaccharides and other nutrients to be utilized by Bacteroides spp. (4, 6–8). Each PUL targets a particular class of polysaccharide and is characterized by the presence of SusC and SusD orthologs. SusD-like proteins are surface-exposed lipoproteins that bind to polysaccharides, which enables them to be broken down further by surface glycosyl hydrolases and imported into the periplasm via a SusC-like TonB-dependent outer membrane (OM) porin (6, 9)

Previous mass spectrometry (MS) analyses revealed that outer membrane vesicles (OMVs) from Bacteroides thetaiotaomicron (Bt) and Bacteroides fragilis (Bf) are preferentially enriched with surface-exposed glycosyl hydrolases, SusD-like proteins, and other proteins typically encoded in PULs (10–12). OMVs are small, spherical, membranous compartments derived from the active blebbing of the OM of Gram-negative bacteria to traffic cellular contents (13). Due to their glycolytic activity, Bacteroides OMVs are viewed as “public goods” because they can degrade various intestinal fibers at a distance, and the resulting breakdown products are readily accessible to kin bacteria and other commensal microbes (11, 14, 15). Producing OMVs as “public goods” is energetically costly; hence, Bacteroides spp. must tightly control the co-expression of PULs along with OMV biogenesis (11, 13). We recently reported evidence for this model, demonstrating that Bt alters their OMV PUL repertoire to adapt to the extracellular glycan landscape (11). This phenomenon provides further support for the idea that OMVs produced by Bacteroides spp. are important for these microbes to effectively compete in the gut. Despite their importance, very little is known regarding how OMV biogenesis and regulation occurs. Determining how Bacteroides OMVs are produced and regulated is key to understanding the physiology of these microbes and how they function in the gut.

Our recent studies have gained insight into how OMV biogenesis is regulated in Bt (11, 16). Briefly, we expressed fluorescent OMV reporters in live cells and employed fluorescence microscopy to visualize OMVs actively blebbing from the OM of live Bt cells (11). By adapting this visualization system, we developed an OMV reporter screen to allow the identification of genes involved in OMV biogenesis and regulation in vitro in a high-throughput manner (16). We found that mutation of Dual Membrane-spanning Anti-sigma factor 1 (Δdma1) induces OMV production in Bt by relieving the repression on its cognate ECF21 family sigma factor, das1 (16). Dma1 is the first representative of a new class of structurally novel anti-sigma factors that have domains spanning from the OM into the cytosol (16). Additional Dma family members, termed Dma2 and Dma3, were identified in Bt (16). In this work, we investigate the role of Dma2 in Bt. Our findings demonstrate that Dma2 plays dual roles in modulating OMV biogenesis and is an important determinant of colonization fitness by regulating host-glycan targeting PULs.

RESULTS

Mutation of Dma2 (BT_1558) increases OMV production

Our preliminary experiments suggested that Dma2 controls OMV biogenesis in a similar manner to Dma1 in Bt (16). To begin our analysis, we isolated total membranes (inner and outer membranes; TM) and OMVs from the wild-type (WT), Δdma2, and its corresponding complemented strain (Δdma2_Comp) and analyzed their protein profiles via SDS-PAGE, followed by Coomassie staining. We found that Δdma2 exhibited a distorted electrophoretic profile when compared to the WT and Δdma2_Comp (Fig. 1A). The OMV fractions obtained from hypervesiculating strains display irregular SDS-PAGE profiles, due to the increased abundance of lipopolysaccharide (LPS), a key OMV structural component, in these samples (16). Furthermore, we performed LPS Silver Stains to measure the relative amounts of LPS and quantified the total protein present in the TM and OMV fractions from the WT, Δdma2, and Δdma2_Comp (Fig. 1B and C). Although no differences were observed in the content of LPS and proteins in the TM fractions, the OMV fraction from Δdma2 contained significantly more LPS and protein when compared to the WT and Δdma2_Comp (Fig. 1B and C). Together, these findings all suggest that mutation of dma2 causes Bt to hypervesiculate. To directly quantify OMV production, we isolated OMVs from the WT, Δdma2, and Δdma2_Comp strains, imaged them by transmission electron microscopy (TEM), and then quantified the number of vesicles present per image. Quantification of OMVs by TEM confirmed that Δdma2 increases OMV production by ~50% when compared to the WT (Fig. 1D). The complemented strain was found to be significantly different from the WT, but this is likely due to partial complementation. Our findings indicate that the observed phenotypes are due to increased vesiculation and not compositional changes in the contents of the OMV fraction in Δdma2 (Fig. 1D).

Fig 1.

Mutation of Dma2 leads to increased OMV biogenesis in Bt. (A) Coomassie Blue stain comparing electrophoretic profiles between TM and OMV fractions from Bt WT, Δdma2, and Δdma2_Comp. Samples were normalized by OD₆₀₀ prior to being run on 10% SDS-PAGE gels. This suggests that deletion of dma2 induces vesiculation in Bt. (B) LPS Silver Stains and (C) Bio-Rad DC protein assays comparing TM and OMV fractions from Bt WT, Δdma2, and Δdma2_Comp. These show that Δdma2 contains more LPS and proteins in its OMV fraction, which is consistent with increased OMV production. Data represents the mean and standard error of three biological replicates performed in triplicate. (D) TEM confirms that Δdma2 produces significantly more OMVs than the WT. Left: Results of the quantification of 90 TEM images of OMVs from the OMV fraction from each strain (FoV: Field of view). Right: Representative TEM images of OMVs from each strain. Three biological replicates of aliquots from the OMV fraction of each strain were fixed onto grids in triplicate (in Materials and Methods). Ten random images were taken from each grid (n = 90 per strain), and OMVs were counted manually. (E) Principal component analysis (PCA) of WC, TM, and OMV proteomic data from Bt WT and Δdma2. This demonstrates that the overall composition of each cellular fraction is similar between the two strains. For panels C and D, two-tailed unpaired t-tests were performed to determine statistical significance. Significance threshold corresponds to: (*) P-value ≤ 0.05, (**) P-value ≤ 0.01, (***) P-value ≤ 0.001, and (****) P-value ≤ 0.0001.

Previous studies have demonstrated that Bacteroides OMVs contain select protein cargo that consists primarily of surface-exposed lipoproteins derived from PULs (10–12). To ensure that the increase in vesiculation in Δdma2 is not caused by cell lysis, we performed comparative proteomic analyses of whole cells (WC), TM, and OMV from the WT and Δdma2. The resulting PCA shows that each fraction from the WT and Δdma2 contains similar protein composition (Fig. 1E). Vesicles generated by lysis usually carry ribosomal proteins and other cytoplasmic components. However, since the proper OMV cargo selection is maintained in Δdma2, we can rule out that the increased OMV production observed in Δdma2 is due to membrane instability and cell lysis.

Next, we performed cryo-electron tomography (cryoET) on the WT and Δdma2 to assess the morphology of Bt OMVs. The size and electron density of OMVs were not impacted in Δdma2 (Fig. 2A and B; Fig. S1). However, we found that Δdma2 exhibited a higher proportion of coccoid cells (53.6%) when compared to the WT (24.2%) (Fig. 2C through H). This suggests a potential link between OMV production and cell shape maintenance.

Fig 2.

Mutation of Dma2 causes morphological changes in Bt cells. Representative slices through cryo-electron tomograms of OMVs from (A) Bt WT and (B) Δdma2. Slices through cryo-electron tomograms of rod-shaped and coccoid cells, and their corresponding ratios from (C-E) Bt WT and (F-H) Δdma2. This shows that the mutation of dma2 increases the number of coccoid cells present in Bt.

Das2 (BT_1559) is required to induce OMV production in Δdma2

Dma2, like other Dma family members, exhibits a unique domain organization, consisting of (i) an N-terminal anti-sigma binding domain, (ii) a transmembrane helix, (iii) a long, intrinsically disordered tether-like region, and (iv) a C-terminal β-barrel domain (Fig. 3A). We showed previously that Dma1 modulates OMV biogenesis by directly controlling the activity of its cognate sigma factor, Das1 (16). Dma2 is encoded in a three-gene operon with BT_1557, a protein of unknown function, and BT_1559 (das2), a putative ECF21 family sigma factor (Fig. 3B). ECF21 family sigma factors are found solely amongst Bacteroidota and are encoded adjacent to Dma family members (16–18). This strongly suggests that Dma2 and Das2 form a sigma/anti-sigma pair (Fig. 3C). We hypothesized that the hypervesiculation observed in Δdma2 is due to the liberation and subsequent activation of Das2. To confirm whether Das2 is required to induce OMV biogenesis in Bt, we generated clean das2 deletion mutants in the WT (Δdas2) and Δdma2 background (Δdma2-das2). Growth curves were performed with these strains and revealed that Δdma2 grew slightly slower than the WT and other tested strains in BHI media (Fig. S2). Next, we isolated OMVs from the WT, Δdma2, Δdas2, Δdma2-das2, and Δdma2_Comp and compared the electrophoretic profiles, as a proxy for OMV biogenesis, by SDS-PAGE, followed by Coomassie staining. While we observed no phenotype in Δdas2, we found that Δdma2-das2 displayed a WT electrophoretic profile (Fig. 3D). This confirms that the increased OMV production observed in Δdma2 requires the activity of its sigma factor, das2.

Fig 3.

ECF21 family sigma factor, das2, is required to induce OMV biogenesis in Bt. (A) AlphaFold structural predictions of Dma2 (19). (B) Schematic and putative functions of each gene present in the dma2 operon (Created with BioRender.com). (C) Proposed model of how Dma2 induces OMV production in Bt by controlling the activity of Das2 (Created with BioRender.com). (D) Coomassie Blue stain comparing electrophoretic profiles of OMV fractions from Bt WT, Δdma2, Δdas2, Δdas2-dma2, and Δdma2_Comp. Samples were normalized by OD₆₀₀ values and run on 10% SDS-PAGE gel. This confirms that deletion of das2 in the Δdma2 background restores WT levels of vesiculation.

Dma2 controls OMV biogenesis in a manner that is distinct from that of Dma1

To investigate how Dma2 controls vesiculation, we compared the transcriptome of Δdma2 to that of the WT. Our analysis revealed that the most differentially regulated genes belonged to two main categories: (i) genes encoded in and adjacent to the dma2 operon and (ii) genes that are components of PULs (Fig. 4A; Data sets S1 and S2).

Fig 4.

Dma2 primarily regulates its own operon and PULs targeting host-associated glycans in Bt. (A) Volcano plot representations of transcriptome data comparing Bt WT and Δdma2 (Data sets S1 and S2). (B) Coomassie Blue stain of OMV fractions isolated from Bt WT and strains containing deletions in the downstream genes of the dma2 operon. Samples were normalized by OD₆₀₀ and run on 10% SDS-PAGE. This experiment shows that the dma2 operon downstream genes are not responsible for the induction of OMV production observed in Δdma2. In the gel, Δbt_1555/dma2 represents the bt_1555 and dma2 double mutant, Δbt_1556/dma2 represents the bt_1556 and dma2 double mutant, Δbt_1557/dma2 represents the bt_1557 and dma2 double mutant, Δbt_1555-dma2 is lacking bt_1555, bt_1556, bt_1557, and dma2, while Δbt_1555-das2 is missing bt_1555, bt_1556, bt_1557, dma2, and das2.

In Δdma2, bt_1555 (Log₂FC: 2.48), bt_1556 (Log₂FC: 5.78), bt_1557 (Log₂FC: 6.00), and das2 (bt_1559) (Log2FC: 4.54) were among the most upregulated genes (Fig. 3B; Fig. 4A; Table S2; Data set S1). No studies have attributed functions to these genes; however, Foldseek predicts that BT_1555 is structurally similar to enoyl-acyl carrier protein (ACP) reductases and nitronate monooxygenases, while BT_1556 and BT_1557 are annotated as DUF4858- and DUF4943 domain-containing proteins, respectively (20). To test whether these genes impact OMV production, we generated mutant strains in each of these genes in the Δdma2 background. OMV analysis revealed that deletion of the upregulated genes encoded near Dma2 and Das2 does not have an impact on OMV biogenesis (Fig. 4B).

We previously showed that Δdma1 induces the expression of nigD1 (bt_4005) along with other select genes in Bt (16). NigD1 belongs to a class of proteins called NigD-like proteins that are found solely amongst Bacteroidota, and it is required to increase OMV production in Δdma1 (16). Since Dma1 and Dma2 both belong to the same family and increase OMV biogenesis, we hypothesized that mutation of dma2 could also induce OMV production by increasing the expression of nigD1. However, our RNA-seq analysis revealed that the expression of nigD1 (Log₂FC: −0.79) was slightly downregulated in Δdma2, which counters this idea (Table S2; Data set S1).

Finally, to determine whether there are shared genes regulated by both dma1 and dma2 that could provide additional insight regarding how they modulate OMV biogenesis, we compared our previous transcriptome data from Δdma1 (16) to that collected from Δdma2. We found that genes encoding a putative type V pilus (bt_2655-2660) (21), an orphan ECF type sigma factor (bt_2569), the dma3 locus (bt_2778-2779) (16), and many hypothetical proteins are upregulated, while components from PUL36 (α-mannan/host N-glycans), PUL52 (unknown), PUL67 (mucin-O-glycans), PUL68 (α-mannan/host N-glycans) (22, 23), S-layer proteins (bt_1926-1927) (24), and glycine betaine/L-proline transport system permeases (bt_1750-1751) (25) are downregulated in both Δdma1 and Δdma2 (Table S3; Data sets S3 and S4). Many of the shared genes identified here are not likely to be implicated in OMV biogenesis, but the dma3 locus is of significant interest. Since the dma3 locus is upregulated in both strains, this suggests that there is potentially crosstalk occurring between members of the Dma family. To determine whether dma3 plays a role in inducing OMV biogenesis, we generated clean deletion mutants in dma3 in the WT, Δdma1, and Δdma2 backgrounds prior to isolating the OMV fraction and visualizing the electrophoretic profile by SDS-PAGE followed by Coomassie staining. However, mutation of dma3 does not revert the distortion in the electrophoretic profiles observed in Δdma1 and Δdma2. This indicates that dma3 activity is not required for these strains to hypervesiculate (Fig. S3). Additional studies are required to determine the role of Dma3 in Bt. Altogether, our findings support the conclusion that Δdma1 and Δdma2 induce OMV production through distinct regulatory cascades.

Absence of Dma2 results in altered OMV cargo selection

PULs are complex, nutrient acquisition systems encoded by Bacteroidota that enable them to utilize a wide array of dietary-, microbial-, and host-derived glycans (4, 7, 8, 23). The transcriptomic data from Δdma2 revealed that genes in select PULs are differentially regulated. We found that PUL36 (α-mannan/host N-glycans), PUL52 (unknown substrate), PUL68 (α-mannan/host N-glycans), and PUL72 (mucin-O-glycans/high mannose mammalian N-glycan) are the most repressed PULs when compared to the WT. On the other hand, PUL38 (mucin-O-glycans) and PUL56 (1,6-β-glucan) were induced in Δdma2 when compared to the WT (Fig. 4A; Table 1; Data sets S1 and S2).

TABLE 1.

List of differentially expressed PUL genes from Δdma2 vs WT RNA sequencing

Downregulated PULs
PUL	Gene (new locus tag)	Gene (old locus tag)	Log2FC	Padj	Substrate
PUL68	BT_RS19095	BT_3787	−2.8421101	6.585E-11	α-Mannan/host N-glycans
	BT_RS19100	BT_3788	−2.4206965	8.656E-11	α-Mannan/host N-glycans
	BT_RS19060	BT_3779	−2.0816286	8.501E-14	α-Mannan/host N-glycans
	BT_RS19105	BT_3789	−1.972928	3.105E-06	α-Mannan/host N-glycans
	BT_RS19085	BT_3784	−1.905731	3.898E-56	α-Mannan/host N-glycans
	BT_RS19110	BT_3790	−1.5110252	0.0091832	α-Mannan/host N-glycans
	BT_RS19065	BT_3s780	−1.4576049	0.0204024	α-Mannan/host N-glycans
	BT_RS19120	BT_3792	−1.4516525	1.902E-08	α-Mannan/host N-glycans
	BT_RS19115	BT_3791	−1.4283173	0.0010753	α-Mannan/host N-glycans
	BT_RS19090	BT_3786	−1.3291173	1.298E-05	α-Mannan/host N-glycans
	BT_RS19045	BT_3776	−1.3034711	0.0621142	α-Mannan/host N-glycans
	BT_RS19050	BT_3777	−1.2124131	0.0627374	α-Mannan/host N-glycans
	BT_RS19070	BT_3781	−1.1770001	3.468E-10	α-Mannan/host N-glycans
PUL52	BT_RS16400	BT_3239	−1.8079871	1.037E-14	Unknown
	BT_RS16395	BT_3238	−1.6402074	2.635E-06	Unknown
	BT_RS16415	BT_3240	−1.5664591	1.853E-20	Unknown
	BT_RS16390	BT_3237	−1.4894948	1.561E-08	Unknown
	BT_RS16385	BT_3236	−1.4012068	1.418E-12	Unknown
	BT_RS16380	BT_3235	−1.3593774	1.576E-20	Unknown
	BT_RS16425	BT_3242	−1.2688034	4.602E-14	Unknown
	BT_RS16420	BT_3241	−1.252067	1.2E-06	Unknown
	BT_RS16430	BT_3243	−1.2089576	5.266E-09	Unknown
	BT_RS16435	BT_3244	−1.0403248	1.594E-10	Unknown
PUL72	BT_RS20105	BT_3984	−3.3194878	2.338E-41	Mucin-O-glycans/high mannose mammalian N-glycan
	BT_RS20100	BT_3983	−2.9338803	2.737E-36	Mucin-O-glycans/high mannose mammalian N-glycan
	BT_RS20110	BT_3985	−2.8586692	8.722E-17	Mucin-O-glycans/high mannose mammalian N-glycan
	BT_RS20115	BT_3986	−2.4411336	3.071E-14	Mucin-O-glycans/high mannose mammalian N-glycan
	BT_RS20145	BT_3992	−2.2819004	2.178E-44	Mucin-O-glycans/high mannose mammalian N-glycan
	BT_RS20135	BT_3990	−2.2630706	3.503E-28	Mucin-O-glycans/high mannose mammalian N-glycan
	BT_RS20140	BT_3991	−2.1626853	9.057E-67	Mucin-O-glycans/high mannose mammalian N-glycan
	BT_RS20120	BT_3987	−2.1192517	1.675E-12	Mucin-O-glycans/high mannose mammalian N-glycan
	BT_RS20125	BT_3988	−1.5356378	5.419E-06	Mucin-O-glycans/high mannose mammalian N-glycan
PUL36	BT_RS13295	BT_2629	−2.1856439	3.453E-68	α-Mannan/host N-glycans
	BT_RS13290	BT_2628	−1.6324335	5.416E-14	α-Mannan/host N-glycans
	BT_RS13280	BT_2626	−1.1388348	2.477E-06	α-Mannan/host N-glycans
	BT_RS13260	BT_2622	−1.126785	2.413E-09	α-Mannan/host N-glycans
	BT_RS13270	BT_2624	−1.0544695	1.269E-06	α-Mannan/host N-glycans
PUL77	BT_RS21005	BT_4163	−1.2380494	1.52E-06	Unknown
	BT_RS21010	BT_4164	−1.138667	0.000873	Unknown
	BT_RS20950	BT_4152	−1.1385313	0.0020674	Unknown
	BT_RS21015	BT_4165	−1.037398	0.0047151	Unknown
PUL74	BT_RS20620	BT_4085	−1.3791482	4.887E-05	Host glycans (unknown type, PMG phase 2)
	BT_RS20585	BT_4078	−1.0984951	0.129151	Host glycans (unknown type, PMG phase 2)
PUL67	BT_RS18915	BT_3750	−1.1872617	2.098E-11	Mucin-O-glycans
	BT_RS18910	BT_3749	−1.082215	6.278E-06	Mucin-O-glycans
PUL5	BT_RS01345	BT_0273	−1.1012221	0.029263	Unknown
	BT_RS01330	BT_0270	−1.0807573		Unknown
PUL80	BT_RS21705	BT_4299	−1.3966409	1.584E-06	Host glycans (unknown type, likely mucin O-glycans)
PUL7	BT_RS01770	BT_0363	−1.129689		Host/residual dietary glycans (unknown type)
PUL83	BT_RS22555	BT_4472	−1.0642391	0.0007893	Unknown
PUL73	BT_RS20385	BT_4039	−1.0457806	0.000107	Mucin-O-glycans
PUL75	BT_RS20765	BT_4114	−1.0406832	0.0011298	Host/residual dietary glycans (unknown type)
PUL71	BT_RS19970	BT_3958	−1.0170786	7.251E-06	Mucin O-glycans (core 1 disaccharide)

Since Dma2 is an anti-sigma factor that functions by controlling the activity of Das2, we propose that the selective dysregulation of PULs observed in Δdma2 is caused by Das2 activity. Since the WT bacteria express basal levels of Das2, to gain more insight regarding how Dma2 and Das2 modulate the expression of PULs in Bt, we conducted additional transcriptomic analyses that compared Δdma2, where Das2 is constitutively active, to Δdma2-das2, where the system is completely inactive. Like the previous RNA-seq, the dma2 operon and its neighboring genes were the most induced, while the expression of various PULs was also changed (Table S4; Data set S5). In this new data set, additional genes related to capsule biosynthesis were also differentially regulated, which we postulate is a consequence of phase variation, a common phenomenon in Bacteroides (Data set S5) (26, 27). Remarkably, PUL1 (unknown), PUL31 (unknown), PUL38 (mucin-O-glycans), PUL56 (1,6-β-glucan), and PUL69 (α-mannan/host N-glycans) were significantly induced, while PUL22 (levan/fructooligosaccharides), PUL36 (α-mannan/host N-glycans), PUL37 (ribose/ribonucleosides), PUL68 (α-mannan/host N-glycans), PUL72 (mucin-O-glycans), and PUL84 (mucin-O-glycans) were significantly repressed in Δdma2 (Fig. 5A; Table S5; Data set S5). Overall, a majority of the differentially expressed PUL genes identified (44/68; ~65%) are predicted to target host glycans (23).

Fig 5.

Select PULs targeting host-associated glycans are downregulated in Δdma2. Volcano plot representations of (A) RNA-seq, (B) OMV, and (C) TM proteomic data comparing Bt Δdma2 and Δdma2-das2 (Data sets S5 through S7). These analyses reveal that Δdma2 primarily alters PULs that target host-associated glycans. These changes are most prevalent in the OMV fraction. (D) Heatmap comparing the differential expression (Average log2FC) of each PUL present in Bt from Bt Δdma2 and Δdma2-das2 transcriptome and proteome data sets. Average log2FC was calculated by filtering out each gene found to be encoded within a particular PUL and averaging the significant log2FC values (log2FC ≥ |±1| and P-value < 0.05). Non-significant log2FC values were included when calculating the Average log2FC for each PUL, but these were assigned the arbitrary value of 0. To ascribe significance, we used the cutoff |Average Log2FC| > 0.8. These analyses enabled us to distinguish whether entire PULs are altered in Δdma2 or if only specific components of certain PULs are changed. PULs that are significantly altered are indicated by the following symbols: “◊” for RNA sequencing, “■” for TM proteomics, and “★” for OMV proteomics.

To evaluate whether the modulation of PUL expression observed in Δdma2 affects OMV cargo, we performed comparative proteomic analyses on OMVs isolated from Δdma2 and Δdma2-das2. In our analysis, we also included total membranes (inner and outer; TM). In Δdma2 OMVs, PULs comprised ~50% of the total proteins found to be significantly altered. On the contrary, the TM fraction displayed fewer changes to PULs (Fig. S4; Data sets S6 and S7). Our analyses revealed that PULs are primarily repressed at the protein level in these subcellular fractions of Δdma2, while very few are induced (Fig. 5B and C; Table S6; Data sets S6 through S8). We found that of the PULs shown to be differentially expressed in the RNA sequencing, only PUL1, PUL31, and PUL68 were also altered in the OMV fraction (Fig. 5D; Data sets S5 through S8). In addition, PUL17 (host glycans), PUL45 (host glycans), PUL67 (mucin O-glycans), PUL80 (host glycans), and PUL88 (unknown) are altered in the OMV fraction, while PUL10 (unknown), PUL18 (unknown), PUL25 (galactooligosaccharides), PUL30 (mucin), PUL52 (unknown), and PUL80 (host glycans) are changed in the TM fractions, although these PULs were not differentially expressed at the transcriptional level (Fig. 5D; Data set S5 through S8). Previous studies have shown that Bt selectively tailors its OMV cargo based on the extracellular nutrient landscape (11). However, our findings are the first to implicate the Dma family in shaping OMV protein cargo through the modulation of PULs in Bt.

Δdma2 exhibits delayed growth on select host-associated glycans in vitro

Since Δdma2 represses the activity of many PULs that primarily target host and microbial glycans (Fig. 5D; Tables S5 and S6; Data set S5 through S8), we hypothesized that Δdma2 would exhibit stunted growth compared to the WT when grown in minimal media supplemented with these types of glycans as a sole carbon source. PUL68 is important for growth in the presence of yeast α-mannan and was the only PUL shown to be significantly repressed in Δdma2 for each of our comparative analyses (Fig. 5D). Interestingly, we found that Δdma2 exhibited slower exponential phase growth compared to the WT and complemented strains when grown in the presence of yeast α-mannan, while the lag phase was unaffected (Fig. 6A; Fig. S5A). Together, these findings suggest that when yeast α-mannan is present, Δdma2 can induce the expression of PUL68 but likely not as efficiently or to the same extent as the WT. Since yeast α-mannans represent a diverse group of polysaccharides, it is possible that growth kinetics would differ in the presence of other α-mannan types (22).

Fig 6.

Δdma2 exhibits delayed growth on select host-derived and microbially derived glycans. Growth curves showing the growth of Bt WT, Δdma2, and Δdma2_Comp in the presence of minimal media containing (A) Saccharomyces cerevisiae α-mannan, (B) heparin, (C) 1,6-β-glucan, (D) hyaluronan, and (E) amylopectin. Growth curves were generated from the results of at least three independent experiments, each including four technical replicates from each strain. (Supported by Fig. S5). Time points and error bars on the graph represent the mean and standard error of the mean.

To determine whether the dysregulation of PULs in Δdma2 impacts their growth in the presence of other carbon sources in vitro, we tested a panel of monosaccharides (glucose, fructose, galactose, arabinose, mannose, rhamnose, and xylose), plant polysaccharides (amylopectin, gum arabic, levan, pectin, and rhamnogalacturonan), and host-derived and microbially derived glycans (hyaluronan, heparin, 1,6-β-glucan, and porcine mucin type II/III). When monosaccharides and most plant polysaccharides were the sole carbon source, Δdma2 grew comparable to the WT (Fig. S6). However, we found that Δdma2 exhibited delayed growth in the presence of the host-associated glycans, such as heparin, which is present in the intestinal mucosa and secreted by mast cells in the gut; 1,6-β-glucan, which is primarily found in yeast and fungal cell walls; and hyaluronan, which is a component of the extracellular matrix in different mammalian cell types (Fig. 6B through D; Fig. S5B through D) (28–30). On the other hand, amylopectin was the only plant polysaccharide where Δdma2 exhibited altered growth kinetics, with a longer lag phase but no difference in exponential phase growth rate (Fig. 6E; Fig. S5E). Our findings confirm that Δdma2 is less able to utilize select carbon sources, primarily host-associated glycans, in vitro.

Dma2 is an important determinant of in vivo fitness in Bt

PULs that target host-associated glycans are important for Bt to stably colonize the human gut (23). Since Δdma2 displayed delayed in vitro growth primarily in the presence of host-associated glycans (Fig. 6A through D), we hypothesized that Δdma2 may be defective in colonizing in vivo. To test this, we treated C57/BL6 mice with an antibiotic cocktail for 7 days prior to colonizing with either Bt WT or Δdma2 by oral gavage and measuring CFUs in feces in intervals for 14 days (Fig. 7A). Mono-colonization experiments revealed that Δdma2 colonized to the same degree as the WT in our antibiotic-treated mouse model (Fig. 7B and C). On the other hand, co-colonization with the WT and Δdma2 strains revealed that Δdma2 initially colonizes at higher levels than the WT, but after 5 days post oral gavage, Δdma2 rapidly declines in abundance, while the WT remains stable in the population (Fig. 7D and E). To ensure that the fitness defect in Δdma2 is not due to the overexpression of the genes surrounding dma2, we also co-colonized with the WT and Δbt_1555-dma2, which possesses das2 but lacks dma2 and the rest of the neighboring genes shown previously to be upregulated in Δdma2. We confirm that Δbt_1555-dma2 is still outcompeted by the WT, but unlike Δdma2, which initially colonizes better than the WT, Δbt_1555-dma2 starts to decline in abundance immediately before appearing to stabilize (Fig. S7A and B). Altogether, these findings demonstrate that Dma2 is an important determinant of in vivo fitness in Bt.

Fig 7.

Dma2 is an important colonization factor in Bt. (A) Schematic outlining in vivo colonization of antibiotic-treated mice experiments (Created with BioRender.com). (B) Initial inocula for both WT and Δdma2 show that mice were colonized with equal amounts of each strain from the start. (C) Monocolonization studies comparing Bt WT and Δdma2. This experiment shows that Bt WT and Δdma2 colonize to comparable levels in vivo. (D, E) Co-colonization experiment comparing Bt WT and Δdma2. This shows that Δdma2 is unable to maintain stable colonization when the WT is present. Points on the graph represent the mean and standard deviation of data collected from three independent experiments containing four mice each per condition. Two-tailed unpaired t-tests were performed to determine statistical significance. Significance threshold corresponds to: (*) P-value ≤ 0.05, (**) P-value ≤ 0.01, (***) P-value ≤ 0.001, and (****) P-value ≤ 0.0001.

DISCUSSION

The Dma family is a novel class of anti-sigma factors found solely amongst Bacteroidota (16, 17, 31). To date, members of the Dma family are known to play a role in OMV biogenesis, but we still lack a complete understanding of their role in these microbes. In this study, we showed that deletion of dma2 in Bt results in a significant increase in vesiculation. Dma2-mediated hypervesiculation requires the activity of its cognate sigma factor, Das2. In addition, transcriptome and proteomic analyses show that Dma2 and Das2 are required for proper OMV cargo selection. Finally, we demonstrate through in vivo studies that Dma2 is a key determinant of colonization fitness.

Dma1 and Das1 induce OMV production by increasing the expression of nigD1 (16). On the contrary, we demonstrated that Dma2 and Das2 increase OMV production but do not increase the expression of NigD1 (Table S2; Data set S1). This suggests that Dma1 and Dma2 regulate OMV biogenesis through distinct regulatory cascades. Even so, the role of NigD1 and other NigD-like proteins in Bt is currently unknown. Future studies are required to understand the precise mechanism by which Dma2 controls OMV production.

In Δdma2, OMV size and structure are not affected (Fig. 2A and B; Fig. S1), but we observed a higher abundance of coccoid cells (Fig. 2C through H). To produce OMVs, cells must properly coordinate the production and trafficking of LPS, other membrane lipids, and OM proteins. This process is likely energetically costly to the bacteria because the secreted cellular contents must be replenished. Our findings could indicate that the induction of OMV biogenesis in Δdma2 impacts the structure of the cell envelope due to an imbalance between the production and export of OM contents.

We previously showed that when Bt is grown in the presence of different polysaccharides, they package their OMVs with enzymes required to degrade them (11). This established a direct link between PUL induction and OMV cargo selection. Interestingly, mucin is a special case where Bt induces the requisite PULs, but these components are retained at the cell surface, instead of localizing to OMVs (11). Here, we demonstrate that Δdma2 primarily alters the expression of various PULs, but many of these transcriptional changes are not necessarily conserved at the protein level (Fig. 5A through D). These findings make Dma2 the first gene shown to impact OMV cargo selection. Most significantly, some of these changes in OMV cargo occur independent of transcriptional regulation, which indicates that additional regulatory features are at play that impact OMV cargo selection in this context. It remains a possibility that Dma2 functions to preclude specific components from host glycan targeting PULs from reaching OMVs. This is physiologically important because it has been postulated that Bt aims to avoid degrading host glycans in an uncontrolled fashion because this has been shown to cause inflammation in certain dietary contexts (11, 32).

Comparisons of our transcriptome analyses between Δdma1 and Δdma2 revealed that dma3 is significantly induced in both cases. Very little is known about Dma3, except that it is the most structurally unique member of the Dma family because it is not encoded adjacent to a cognate ECF21 family sigma factor; instead, Dma3 encodes a domain that functions as its cognate sigma factor at the N-terminus of the protein (16). We showed that Dma3 does not impact OMV production in Δdma1 and Δdma2 (Fig. S3). However, by comparing transcriptome data sets, we found that Δdma1 and Δdma2 both impact the expression of select PULs (Table S3; Data sets S3 and S4); hence, it remains a possibility that Dma3 could play a role in this process. Overall, understanding the role of the Dma family in regulating the PUL repertoire is of significant interest.

In Bacteroides spp., PULs are primarily controlled by SusR-like regulators (9, 33), hybrid two-component systems (HTCS) (34–37), and ECF-type sigma/anti-sigma factor systems (23). These systems enable Bt to maintain a strict glycan hierarchy where monosaccharides and plant polysaccharides are preferred over host mucosal glycans (23, 38, 39). PULs that target mucin and other host glycans are known to disproportionately be associated with ECF-type sigma/anti-sigma factors. However, to the best of our knowledge, this class of regulators has yet to be shown to be important for maintaining the glycan hierarchy in Bacteroides spp. (23). Our data demonstrate that Dma2 modulates the activity of PULs; however, it is tempting to speculate that it could play a role in fine-tuning the glycan hierarchy by downregulating select PULs that target host-associated glycans (Fig. 5A, Table S2, S4 and S5; Data sets S1, S2, and S5). Because of this, we hypothesize that Dma2 can sense the presence of a currently unknown glycan at the OM surface and then release Das2 to modulate the activity of PULs from lower priority host glycan targeting PULs.

Our experiments provide preliminary support for this model. In Fig. 6, we demonstrate that Δdma2 exhibits delayed growth when various host-associated glycans are the sole carbon source. In Bt, heparin has been considered a high-priority glycan because PUL85, which targets heparin, is not significantly repressed when monosaccharides are present (39). Our proteomic analyses showed that BT_4659 (SusD-like) and BT_4660 (SusC-like) from PUL85 are repressed in Δdma2 (Table S6; Data sets S6 and S7), which suggests that Dma2 could be involved in sensing a higher-priority polysaccharide. In tandem, Das2 could function by modulating the activity of lower-priority PULs by inhibiting the activity of their native regulatory systems (23). Future studies should aim to elucidate the substrates that are sensed by Dma2 and decouple whether Das2 functions by directly or indirectly repressing the activity of host glycan-targeting PULs in this context.

Tightly controlling the expression of PULs is key for Bacteroides spp. to adapt to and thrive in the human gut (23). Our experiments revealed that the lack of dma2 drastically impacts the ability of Bt to maintain stable colonization when there is competition from other kin bacteria (Fig. 7D and E). We hypothesize that the dysregulation of PULs and OMV cargo selection in Δdma2 causes the observed fitness defect, but many questions remain. The genes encoded near dma2 (bt_1555-bt_1557) are among the most upregulated genes in Δdma2. Our findings demonstrate that these genes are not involved in OMV biogenesis but could be important for the initial colonization fitness of Δdma2 and potentially the WT (Fig. 4B; Fig. S7). The function of these genes has not been experimentally confirmed; therefore, elucidating the exact function of these genes is required to understand their role in vivo.

The Dma2 operon is only encoded in Bt and closely related Bacteroides spp. (16). This suggests that there is likely a specific context where expressing the dma2 operon gives these microbes a competitive advantage in the gut. Since many factors can impact colonization fitness, and we only tested mice consuming a standard chow diet, we cannot rule out that the ability of Δdma2 to colonize in vivo may vary depending on the host diet and metabolic state. The gut microbiota also consists of many different bacterial species; so, the lack of colonization fitness observed in Δdma2 relative to the Bt WT begs the question of how this strain would compete in the presence of other Bacteroides spp. and commensal microbes (Fig. 7D and E). Future studies will involve determining the exact role of Dma2 and other members of the Dma family within the host.

MATERIALS AND METHODS

Bacterial strains and growth conditions

Strains, oligonucleotides, and plasmids are described in Table S1 in the supplemental material. Escherichia coli was grown aerobically at 37°C in Luria-Bertani (LB) medium. Bacteroides strains were grown in an anaerobic chamber (Coy Laboratories) at 37°C containing an atmosphere of 10% H₂, 5% CO₂, 85% N₂. Bt was cultured in Brain Heart Infusion (BHI) medium (Fisher Scientific) supplemented with 5 µg/mL Hemin and 1 µg/mL vitamin K3. When applicable, antibiotics were used as follows: 100 µg/mL ampicillin, 200 µg/mL gentamicin, 25 µg/mL erythromycin, and 10 μg/mL tetracycline. When required, Bacteroides was grown in minimal medium (MM) containing 100 mM KH₂PO₄ (pH 7.2), 15 mM NaCl, 8.5 mM (NH4)2SO4, 4 mM L-cysteine, 1.9 mM hematin/200 mM L-histidine (prepared together as a 1,000× solution), 100 mM MgCl2, 1.4 mM FeSO4.7H2O, 50 mM CaCl2, 1 µg/mL vitamin K3, and 5 ng/mL vitamin B12. Carbohydrates used to supplement MM include glucose, fructose, galactose, arabinose, mannose, rhamnose, xylose, amylopectin, gum arabic from Acacia Tree, levan, pectin from citrus peel, rhamnogalacturonan, Saccharomyces cerevisiae α-mannan, 1,6-β-glucan, heparin, hyaluronan, porcine mucin type II, porcine mucin type II, porcine mucin type III.

Genetic manipulation of Bt

We employed the pSIE1 vector described in Bencivenga-Barry et al. 2020 to develop constructs for generating clean deletion mutants in Bt (40). Briefly, ~750 base pair regions flanking our genes of interest were cloned into pSIE1. Vectors containing flanking regions of target genes were then transformed into E. coli s17λ-pir by electroporation. Transformants were identified by selection on LB agar plates containing ampicillin, followed by colony PCR to confirm the presence of the plasmid. Logarithmic to early stationary phase cultures of E. coli transformants and Bt were mixed (2:1 ratio) to facilitate conjugation of the vector into Bt. Transconjugants, containing the vector integrated into the Bt genome, were identified by selection on BHI plates containing gentamicin and erythromycin. To delete the gene of interest, Bt transconjugants were cultured overnight; to perform counterselection, 5 μL of overnight culture was diluted in 95 μL of BHI media prior to the entire volume being spread on BHI plates containing 125 ng/mL anhydrotetracycline (aTc). Mutants were identified by PCR prior to whole-genome sequencing.

Complementation of Δdma2 was achieved by cloning dma2 into the pWW3867 vector backbone under the control of the constitutive RpoD (BT_1311) promoter. This vector was originally designed in the study by Whitaker et al. 2017 (41).

OMV isolation

OMVs were purified by ultracentrifugation from cell-free culture supernatants according to our previously published methods (10–12, 16). Briefly, 50 mL of Bt cultures grown to late stationary phase was centrifuged twice at 6,500 rpm at 4°C for 10 min. Supernatants were then filtered using a 0.22-µm-pore membrane (Millipore) to remove residual cells. The filtrate was subjected to ultracentrifugation at 200,000 × g for 2 h (Optima L-100 XP ultracentrifuge; Beckman Coulter). Resulting supernatants were discarded, and the pellets, which contain OMVs, were resuspended in phosphate-buffered saline (PBS). For OMVs, the amount of PBS used for resuspension was based on the measured OD₆₀₀. For example, if the original OD₆₀₀ = 1, then the OMV pellet was resuspended in 100 μL. When performing MS analysis, purified OMV preparations were lyophilized.

Subcellular fractionation

TM preparations were isolated by cell lysis and ultracentrifugation. Briefly, late stationary phase cultures were harvested by centrifugation at 6,500 rpm at 4°C for 10 min. The pellets were gently resuspended in a mixture of PBS containing complete EDTA-free protease inhibitor mixture (Roche Applied Science). Cells were then lysed using two passes through a cell disruptor at 35 kPa. Next, centrifugation at 8,500 rpm at 4°C for 8 min was performed to remove unbroken cells. TMs were collected by ultracentrifugation at 200,000 × g for 1 h at 4°C. Supernatants were discarded, and pellets were resuspended in PBS. For TMs, the amount of PBS used for resuspension was based on the measured OD₆₀₀. If the original OD₆₀₀ = 1, then the TM pellet was resuspended in 1 mL. TM fractions were lyophilized for MS analysis.

SDS-PAGE analyses

To compare protein profiles from Bt WT and Δdma2 strains, samples, either TM or vesicle fractions, were normalized by OD₆₀₀, and equivalent volumes were loaded onto a 10% Tris-glycine SDS-PAGE gel, followed by Coomassie Blue staining to analyze protein profiles.

Abundance of LPS was measured by adapting the methods from the study of Tsai CM and Frasch CE (42). Briefly, samples were standardized by OD₆₀₀, then diluted in PBS, 1:3 (sample: total volume) for OMVs and 1:5 for TMs, and treated with proteinase K for 4 h at 37°C. Next, 2 μL of proteinase K-digested sample was added to 13 μL of 1× Laemmli buffer and boiled for 3 min prior to loading equal amounts (5–10 μL) onto a 15% SDS-PAGE gel. After running, the gels were fixed overnight in 200 mL of 40% ethanol in 5% acetic acid. Next, the gels were oxidized for 5 min in 100 mL of 0.7% fresh periodic acid in 40% ethanol and 5% acetic acid. Upon completion, gels underwent three washes (15 min each) in milliQ H₂O. The gels were then stained for 10 min in the dark with 28 mL 0.1M NaOH, 2mL NH₄OH, 5mL 20% AgNO₃, and 115 milliQ H₂O. Gels underwent three additional washes prior to developing in 200 mL H₂O with 10 mg citric acid and 100 μL formaldehyde.

Protein quantification

To quantify protein content, we utilized the Bio-Rad DC Proteins Assay, which is a colorimetric assay that is like the Lowry assay. Follow the manufacturer’s instructions when performing the assay. At least three biological replicates were performed for each sample in triplicate. The data presented represent the mean and standard error for each sample tested. Two-tailed unpaired t-tests were performed to determine significance.

Growth assay with Bt strains

Bt WT and Δdma2 strains were cultured overnight at 37 °C under anaerobic conditions in BHI broth. Overnight cultures were centrifuged at 6,500 rpm for 10 min before being washed with PBS. The cells were then resuspended in MM supplemented with the indicated carbon sources (0.5% wt/vol final concentration) and normalized to OD₆₀₀ =0.1. Growth assays were conducted in sterile, round-bottom 96-well polystyrene microplates. Cultures were incubated anaerobically at 37 °C under static conditions. OD₆₀₀ readings were recorded every 30 min using a Smart Reader 96-T (Accuris Instruments), following 10 s of orbital shaking to ensure homogenization. Each growth condition was tested in technical quadruplicate and independently repeated for at least three biological replicates. Growth curves were determined to be significantly different by measuring (i) the duration of the lag phase, period before bacteria are actively dividing, and (ii) the growth rate during exponential phase (ΔOD₆₀₀/ΔTime (h)), which was calculated by using the OD₆₀₀ at the start of exponential phase growth and the maximum OD₆₀₀ prior to the plateau of bacterial growth when they enter stationary phase. Lag phase duration and exponential phase growth rate were determined for all technical and biological replicates. These data were compiled, and two-tailed unpaired t-tests were employed to determine significance. Statistics were provided for growth curves that were found to be significantly altered.

Cryo-ET sample preparation and imaging

Bt cells were grown on BHI-agar plates as described previously. Subsequently, the cells were collected from the plate using a loop and resuspended in 1 mL of 1× PBS to a final OD₆₀₀ of 3.0 for cryo-ET experiments. R2/2 carbon‐coated 200 mesh copper Quantifoil grids (Quantifoil Micro Tools) were glow‐discharged for 30 s. Then, the cells were mixed with a solution of 10‐nm gold beads treated with bovine serum albumin; 4 μL of this mixture was applied to the grids in a Vitrobot chamber (FEI). Subsequently, the extra fluid was blotted off using a Whatman filter paper in the Vitrobot chamber with 100% humidity, and the grids were plunge‐frozen in a cryogen (liquid ethane). Sample imaging was performed at the Advanced Electron Microscopy at the University of Chicago. Cells were imaged using a Titan Krios transmission electron microscope operating at 300 kV and equipped with a BioQuantum K3 imaging filter (Gatan). Data were collected using Tomography 5 software with each tilt series ranging from −60° to 60° in 3° increments with a pixel size of 3.35 Å, an underfocus of 7 μm, and a total dose of 130 e^-/Å (2). Subsequently, three‐dimensional reconstructions of tilt series and further visualization were performed using the IMOD software package (43).

To determine cell morphology, we measured two perpendicular axes through the center of the cell. Cells where the ratio between these two axes was ~1–1.3 (to allow for inaccuracy in measuring the axes) were classified as coccoid, while those where the ratio was >1.4 were classified as bacilli.

RNA sequencing sample collection, library preparation, and analysis

RNA was isolated from Bt cells according to our methods outlined in the study of Pardue et al. (16). Briefly, WT and Δdma2 were grown overnight in BHI media before being diluted to the equivalent of OD 0.1 in 10 mL and grown anaerobically for 4 h at 37°C. Four individual overnight and 10 mL culture biological replicates were prepared. Cultures were normalized, and an amount of culture equivalent to an OD₆₀₀ of 4.0 was pelleted for 90s at 8,000 rpm. Pellets were resuspended on ice in 1 mL TRIzol (Invitrogen) with 10 μL of 5 mg/mL glycogen. Samples were flash frozen and stored at −80°C until extraction. Prior to extraction, samples were thawed on ice, then pelleted, and supernatants were treated with chloroform. RNA was extracted from the aqueous phase using the RNeasy minikit (Qiagen, Inc.), and RNA quality was checked by agarose gel electrophoresis and A₂₆₀/A₂₈₀ measurements. RNA was stored at −80°C with SUPERase-IN RNase inhibitor (Life Technologies) until library preparation.

RNA sequencing prep (RNA-Seq) was performed as previously described (44). Briefly, 400 ng of total RNA from each sample was used for generating cDNA libraries following our RNAtag-Seq protocol. PCR amplified cDNA libraries were sequenced on an Illumina NextSeq500, obtaining a high-sequencing depth (over 7 million reads per sample). RNA-seq data were analyzed using our in-house developed analysis pipeline, Aerobio. Raw reads are demultiplexed by 5’ and 3’ indices, trimmed to 59 base pairs, and quality filtered (96% sequence quality>Q14). Filtered reads are mapped to the corresponding reference genomes using bowtie2 with the --very-sensitive option (-D 20 –R 3 –N 0 –L 20 –i S, 1, 0.50). Mapped reads are aggregated by feature Count, and differential expression is calculated with DESeq2 (44). In each pair-wise differential expression comparison, significant differential expression is filtered based on two criteria: |log2foldchange| > 1 and adjusted P-value (padj) <0.05. All differential expression (DE) comparisons are made between the WT and Δdma2 mutants under the conditions mentioned above. The reproducibility of the transcriptomic data was confirmed by an overall high Spearman correlation across biological replicates (R > 0.95). BioProject: PRJNA1298834.

Sample preparation for proteomic analysis

WT, Δdma2, and Δdma2-das2 were grown overnight anaerobically in 3 mL of BHI media prior to being diluted into 50 mL and grown for 20 h. Whole cells, total membranes, and vesicles were collected from each strain. Four individual biological replicates of each fraction were performed for each strain. Samples were lyophilized in preparation for MS analysis.

Proteomic analysis

Acetone-precipitated protein biological replicates/fractions were solubilized in 4% SDS, 100 mM HEPES by boiling for 10 min at 95 °C, then protein concentrations were assessed using bicinchoninic acid protein assays (Thermo Fisher Scientific). 200 μg of each biological replicate/fraction was prepared for digestion using S-trap mini columns (Protifi, USA) according to the manufacturer’s instructions. Briefly, samples were reduced with 10 mM dithiothreitol for 10 min at 95 °C and then alkylated with 40 mM Iodoacetamide in the dark for 1 h. Samples were acidified to 1.2% phosphoric acid and diluted with seven volumes of S-trap wash buffer (90% methanol, 100 mM Tetraethylammonium bromide, pH 7.1) before being loaded onto S-traps and washed 3 times with 400 μL of S-trap wash buffer. Samples were then digested with 4 μg of Trypsin (a 1:50 protease/protein ratio) in 100 mM Tetraethylammonium bromide overnight at 37 °C before being collected by centrifugation with washes of 100 mM Tetraethylammonium bromide, followed by 0.2% formic acid, and then 0.2% formic acid/50% acetonitrile. Samples were dried down and further cleaned up using C18 Stage (1, 2) tips to ensure the removal of any particulate matter.

C18 cleaned up peptide samples were re-suspended in Buffer A* (2% acetonitrile, 0.1% trifluoroacetic acid in Milli-Q water) and separated using a two-column chromatography set-up on a Dionex Ultimate 3000 UPLC composed of a PepMap100 C18 20 mm × 75 μm trap and a PepMap C18 500 mm × 75 μm analytical column (Thermo Fisher Scientific) coupled to a Orbitrap Fusion™ Lumos Tribrid Mass Spectrometer (Thermo Fisher Scientific) with a FAIMS Pro interface (Thermo Fisher Scientific); 145-minute gradients were run for each sample, with samples loaded onto the trap column with 98% Buffer A (2% acetonitrile, 0.1% formic acid in Milli-Q water) and 2% Buffer B (80% acetonitrile, 0.1% formic acid) with peptides separated by altering the buffer composition from 2% Buffer B to 28% B over 126 min, then from 28% B to 40% B over 9 min, then from 40% B to 80% B over 3 min, the composition was held at 80% B for 2 min, then dropped to 2% B over 2 min, and held at 2% B for another 3 min. A data-dependent stepped FAIMS approach was utilized with two different FAIMS CVs of −45 and −65, as previously described (3). For each FAIMS CV, a single Orbitrap MS scan (500–2,000 m/z, maximal injection time of 50 ms, an AGC of maximum of 4*10⁵ ions and a resolution of 60k) was acquired every 2 s, followed by Orbitrap MS/MS HCD scans of precursors (NCE 30%, maximal injection time of 80 ms, an AGC set to a maximum of 1.25*10⁵ ions and a resolution of 30k).

Proteomic data analysis

Prior to identification and LFQ analysis, files were separated into individual FAIMS fractions using the FAIMS MzXML Generator (4). Separated FAIMS fractions were searched against the Bt VPI-5482 proteome (Uniprot: UP000001414) using MaxQuant (v1.6.17.0) (5), allowing carbamidomethylation of cysteine set as a fixed modification and oxidation of methionine as a variable modification. Searches were performed with Trypsin cleavage specificity, allowing two miscleavage events with a maximum false discovery rate (FDR) of 1.0% set for protein and peptide identifications. The LFQ and “Match Between Run” options were enabled to allow comparison between samples. The resulting data files were processed using Perseus (v1.4.0.6) (6) with missing values imputed based on the total observed protein intensities with a range of 0.3 σ and a downshift of 1.8 σ. Statistical analysis was undertaken in Perseus using two-tailed unpaired t-tests. Individual proteins were deemed significantly altered if the “Student’s t-test Difference” (equivalent to Log₂FoldChange) was greater than |±1| and the “-Log Student’s t-test P-value” was greater than 1.3 (equivalent to P-value = 0.05).

Negative staining and analysis by TEM

For quantitative analyses at the ultrastructural level, 200 mesh formvar/carbon-coated copper grids (Ted Pella Inc., Redding, CA) were coated with 50µg/mL poly-L-lysine (Sigma, St Louis, MO) for 10 min at 37^○C. Excess fluid was removed, and grids were allowed to air dry. Poly-L-lysine coating allowed for even distribution of material across the grid; 5 μL spots of normalized OMV fractions from our strains were fixed with 1% glutaraldehyde (Ted Pella Inc.) and allowed to absorb onto freshly glow-discharged poly-L-lysine-coated grids for 10 min. Grids were then washed in dH₂O and stained with 1% aqueous uranyl acetate (Ted Pella Inc.) for 1 min. Excess liquid was gently wicked off, and grids were allowed to air dry. Samples were viewed on a JEOL 1200EX transmission electron microscope (JEOL USA, Peabody, MA) equipped with an AMT 8-megapixel digital camera (Advanced Microscopy Techniques, Woburn, MA). Three biological replicates were prepared for each strain, and each biological replicate was processed in triplicate (three grids per biological replicate for a total of nine grids for each of the three strains tested here). Ten random images were taken at a magnification of 25,000× from various areas of each grid for a total of 90 images per strain. Finally, the total number of OMVs on each grid was manually counted. The gathered data were used to construct violin plots that show the median, interquartile range, and overall data distribution. Two-tailed unpaired t-tests were performed to determine statistical significance.

Competitive colonization of antibiotic-treated mice

All animal experiments were approved by the Washington University Animal Care and Use Committee, and we have complied with all relevant ethical regulations. All mice used were from the inbred C57/BL6 line. Six-week-old animals were used for colonization experiments. Mice were administered an antibiotic cocktail consisting of ampicillin (333.3 mg/mL; 15 μL), neomycin (333.3 mg/mL; 15 μL), metronidazole (10 mg/mL; 100 μL), and vancomycin (166.7 mg/mL; 30 μL), each mouse receiving 160 μL by oral gavage, every 24 h for 7 consecutive days to deplete the normal intestinal flora. Next, mice were given an inoculum of a single Bt strain, for monocolonization experiments, or two Bt strains, for co-colonization experiments (~10¹⁰ CFUs/oral gavage total; an aliquot was taken from the input inoculum and plated on BHI agar to count CFUs) for 2 consecutive days. To differentiate our strains, the WT expresses an erythromycin resistance cassette, while the mutants express tetracycline resistance cassettes from the pNBU2 backbone. Fresh fecal pellets were collected 2, 3, 5, 7, 10, and 14 days post-oral gavage and used to quantify CFU/mL to track colonization throughout the duration of the experiment. Four mice were utilized per condition, and each experiment was conducted in triplicate for a total of 12 mice per condition. Competitive index represents the ratio of mutant CFU/g of feces to that of the WT. Two-tailed unpaired t-tests were performed to determine statistical significance.

DATA AVAILABILITY

The mass spectrometry proteomics data has been deposited in the Proteome Xchange Consortium via the PRIDE partner repository (https://www.ebi.ac.uk/pride/) and is accessible with the data set identifier: PXD066605.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jb.00442-25.

Data Set S1. jb.00442-25-s0001.xlsx.

RNA sequencing data for Δdma2 vs. WT.

Data Set S2. jb.00442-25-s0002.xlsx.

Polysaccharide utilization loci in Bacteroides thetaiotaomicron.

Data Set S3. jb.00442-25-s0003.xlsx.

Genes upregulated in Δdma2 and Δdma1.

Data Set S4. jb.00442-25-s0004.xlsx.

Genes downregulated in Δdma2 and Δdma1.

Data Set S5. jb.00442-25-s0005.xlsx.

RNA sequencing data for Δdma2 vs Δdma2-das2.

Data Set S6. jb.00442-25-s0006.xlsx.

OMV proteomics data for Δdma2 vs Δdma2-das2.

Data Set S7. jb.00442-25-s0007.xlsx.

TM proteomics data for Δdma2 vs. Δdma2-das2.

Data Set S8. jb.00442-25-s0008.xlsx.

List of polysaccharide utilization loci from Δdma2 vs. Δdma2-das2 omics.

Supplemental figures. jb.00442-25-s0009.pdf.

Fig. S1 to S7.

Supplemental tables. jb.00442-25-s0010.xlsx.

Tables S1 to S6.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.