Fig 1.
RbmB plays a role in V. cholerae biofilm dispersal.
(A) Schematic of a working model suggesting cellular escape is helped by RbmB-mediated cleavage of VPS. Created in BioRender. Olson, R. (2022) BioRender.com/m42l648 and Olson, R. (2022) BioRender.com/n45c416. (B) Schematic version of the tetrasaccharide structure of VPS based on [21]. The asterisk represents the modified gulose moiety. (C) AlphaFold2 [24] model for RbmB structure. (D) RbmB was expressed as a thrombin-cleavable fusion protein with DsbA and a periplasmic secretion signal peptide at the N-terminus. (E) Size exclusion chromatography trace showing purified RbmB fusion before and after cleavage of the DsbA fusion partner. (F) SDS-PAGE gel showing purified DsbA-RbmB fusion and cleaved and purified RbmB. The expected molecular weight for cleaved RbmB is ~42.7 kDa.
Fig 2.
(A) Fluorescent-imaged native-PAGE gel using Bap1-GFPUV to label VPS polymer. Addition of Bap1 to uncut VPS shows a shift to a large complex that is unable to enter the gel. Increasing amounts of RbmB show a decrease in the aggregates and increase in free Bap1, suggesting cleavage of VPS into small fragments. Lanes left to right: Bap1 alone, Bap1+VPS, Bap1+VPS pre-treated with 0.05 μg RbmB, Bap1+VPS pre-treated with 1 μg RbmB, Bap1+VPS pre-treated with 2.5 μg RbmB, Bap1+VPS pre-treated with 5 μg RbmB, Bap1+VPS pre-treated with 10 μg RbmB, and Bap1+VPS pre-treated with 15 μg RbmB. (B) Negative-stained EM images of Bap1/VPS complexes. Bap1 particles form dense aggregates of size >0.5 μm when added to VPS. Digestion with RbmB disrupts aggregates back to smaller Bap1/VPS particles of size <0.1 μm.
Fig 3.
VPS cleavage activity of purified RbmB.
(A) Quantification of reducing ends as a function of time formed during RbmB-mediated cleavage of VPS. A standard curve using galactose as a substrate is shown in the inset and S2 Fig. The effects of pH (B), temperature (C), salt concentration (D), and divalent cations (E) on the activity of RbmB against VPS are also shown. All data are represented as the mean ± SD (n = 3). 100% activity in each graph is calculated based on the condition with the highest activity. Statistical significance was determined using an unpaired, two-tailed t-test with Welch’s correction. ns = not significant, *p<0.05, **p<0.01. Solid lines indicate which samples were compared in the statistical tests. All experiments were conducted with 25 μg of VPS and 20 μg of RbmB.
Fig 4.
Mass spectrometry analysis of VPS fragments produced by RbmB cleavage.
(A) Liquid chromatographic analysis of an overnight digestion of VPS by RbmB indicates a separable mixture of tetra- and octameric oligosaccharides of VPS. (B,C) Mass spectrometry analysis of liquid chromatography peaks indicates a primary species consistent with the published tetrameric form of VPS, including formation of a C4-C5 double bond in the gulose moiety due to a lyase mechanism. (D) Full negative mode mass spectrometry spectrum showing modified species and the chemical structure of the primary species. The spectra also indicate species consistent with O-acetylation and N-acetylation of the tetrasaccharide, presumably on the gulose and glucose saccharides as indicated in red. Annotation of major peaks in the spectrum is shown in S1 Table. (E) MS2 analysis is consistent with cleavage between the galactose and gulose moieties. Four possible cleavage sites would produce four different tetrameric species (shown in inset boxes). To identify the exact cleavage site, VPS was borate-reduced following RbmB cleavage and subjected to MS2 mass spectrometry analysis. The fragmentation pattern is consistent with a single cleavage site between the galactose and gulose, designating RbmB as an α-1,4 lyase (red box indicates resulting tetrasaccharide). (F) Integration of peak intensity for the tetra- and octasaccharide over time (ON = overnight). The tetramer:octamer peak intensity ratio (red) remains relatively constant throughout the digestion. All data are represented as the mean ± SD (n = 3).
Fig 5.
NMR analysis of uncleaved and cleaved VPS.
(A) Superposition of 1D 13C NMR spectra collected on uncleaved VPS (using CPMAS solid-state NMR spinning at 10,000 Hz, blue) and RbmB-cleaved VPS (solution NMR, red) shows the presence of new peaks in the cleaved sample (carbons numbered as shown on proposed modified gulose structure). The chemical shifts of these two new peaks are consistent with formation of alkene carbons predicted to result from lyase activity on the gulose moiety. The inset table shows chemical shift values (in ppm) for uncleaved VPS as reported in [21] and cleaved VPS from the solution NMR data. (B) Solution 2D [13C, 13C] COSY experiment run on RbmB-cleaved VPS shows assignments for carbons around the gulose ring. Green dotted lines illustrate the connectivity of carbon atoms around the gulose ring as labeled in (A).
Fig 6.
RbmB disrupts V. cholerae biofilms.
(A) After biofilm growth in M9 medium with 0.5% glucose and 0.5% casamino acids as nutrient source, we induced nutrient limitation by removing the carbon sources in the medium. Biomasses before and after dispersal for 5 hours at 30°C were quantified using a confocal microscope. The ΔrbmB mutant retains more surface biofilm mass following incubation, suggesting a weaker dispersal phenotype. Induction of rbmB from a plasmid under arabinose control shows partial recovery of the dispersal phenotype. Statistical analyses were performed using an unpaired, two-tailed t-test with Welch’s correction. ns = not significant, *p < 0.05, ***p < 0.001, ****p < 0.0001. (B) Recombinantly purified RbmB (50 μg/mL) from V. cholerae (Vc) or V. coralliilyticus (Vcor) is sufficient to nearly completely disrupt V. cholerae biofilms within 1 hour, but heat-inactivated RbmB has no effect. All data are depicted as mean ± SD. Statistical analyses were performed using an unpaired, two-tailed t-test with Welch’s correction. ns = not significant, ****p < 0.0001. (C) Cross-sectional images of V. cholerae biofilms at the bottom cell layer, before and after 1 hour of nutrient limitation. 50 μg/mL RbmB from V. cholerae (Vc) or V. coralliilyticus (Vcor) were present during the nutrient limitation step, in the middle and bottom rows, respectively. All data in this figure are depicted as mean ± SD (n = 3 biologically independent samples).
Table 1.
Bacterial strains used in this study.