Unintended Laboratory-Driven Evolution Reveals Genetic Requirements for Biofilm Formation by Desulfovibrio vulgaris Hildenborough

Abstract

Biofilms of sulfate-reducing bacteria (SRB) are of particular interest as members of this group are culprits in corrosion of industrial metal and concrete pipelines as well as being key players in subsurface metal cycling. Yet the mechanism of biofilm formation by these bacteria has not been determined. Here we show that two supposedly identical wild-type cultures of the SRB Desulfovibrio vulgaris Hildenborough maintained in different laboratories have diverged in biofilm formation. From genome resequencing and subsequent mutant analyses, we discovered that a single nucleotide change within DVU1017, the ABC transporter of a type I secretion system (T1SS), was sufficient to eliminate biofilm formation in D. vulgaris Hildenborough. Two T1SS cargo proteins were identified as likely biofilm structural proteins, and the presence of at least one (with either being sufficient) was shown to be required for biofilm formation. Antibodies specific to these biofilm structural proteins confirmed that DVU1017, and thus the T1SS, is essential for localization of these adhesion proteins on the cell surface. We propose that DVU1017 is a member of the lapB category of microbial surface proteins because of its phenotypic similarity to the adhesin export system described for biofilm formation in the environmental pseudomonads. These findings have led to the identification of two functions required for biofilm formation in D. vulgaris Hildenborough and focus attention on the importance of monitoring laboratory-driven evolution, as phenotypes as fundamental as biofilm formation can be altered.IMPORTANCE The growth of bacteria attached to a surface (i.e., biofilm), specifically biofilms of sulfate-reducing bacteria, has a profound impact on the economy of developed nations due to steel and concrete corrosion in industrial pipelines and processing facilities. Furthermore, the presence of sulfate-reducing bacteria in oil wells causes oil souring from sulfide production, resulting in product loss, a health hazard to workers, and ultimately abandonment of wells. Identification of the required genes is a critical step for determining the mechanism of biofilm formation by sulfate reducers. Here, the transporter by which putative biofilm structural proteins are exported from sulfate-reducing Desulfovibrio vulgaris Hildenborough cells was discovered, and a single nucleotide change within the gene coding for this transporter was found to be sufficient to completely stop formation of biofilm.

Keywords: Desulfovibrio vulgaris; biofilms; genetic polymorphisms; secretion systems; sulfate reduction.

FIG 1

Growth and batch biofilm comparison of DvH-MO and DvH-MT. (a) Growth of DvH-MO (diamonds) and DvH-MT (squares) in LS4D medium at 30°C. Error bars denote standard deviations from the average from triplicate samples. (b) To compare biofilms, anaerobic cultures in Balch tubes containing LS4D with a halved glass microscope slide were inoculated and incubated for 48 h. The slides were removed and imaged. (c) Biofilm was measured as protein, scraped from the glass slides. Error bars denote the standard deviations from two biological and two technical replicates.

FIG 2

Biofilm reactors of DvH-MO and DvH-MT. Biofilms of DvH-MO (solid lines) and DvH-MT (dashed line) from reactors were scraped from halved glass microscope slides sampled at intervals and measured by protein. The limit of detection was 1.3 μg protein/cm². Time 0 denotes the time at which the cultures were transitioned from batch to continuous growth in order to control the growth rate by substrate availability. Two representatives of four DvH-MO reactor runs are displayed. Error bars denote the standard deviations from three technical replicates; most are within the symbols. The red triangle (168 h) indicates the biofilm sample collected for DvH-MO genome resequencing. Two biofilm samples were collected at 24 h for DvH-MT and are nearly overlapping.

FIG 3

DvH-MO SNP mutation in a conserved α-helix of the ABC transporter DVU1017 of a putative type I secretion system (T1SS). (a) Cartoon representation of a T1SS complex comprised of three subunits: the outer membrane protein, the membrane fusion protein (MFP), and the ABC transporter (16). The red X marks the subunit with the SNP found in planktonic DvH-MO. (b) Sequence and amino acid changes found at the 635th residue of the DVU1017 protein, the predicted ABC transporter. Shown is a cartoon representation of the possible proline kink in this predicted α-helix and the change in direction that the adjacent amino acids are facing (depicted by shading). (c) Secondary structure prediction of the 622- to 650-amino-acid region of DVU1017 that encompasses residue 635. Prediction was with PSIPRED. Pink cylinders denote a helical structure, H, that is flanked by coils, C. The PSIPRED confidence of the prediction is shown along the top. (d) A sequence logo generated from an alignment of the DVU1017 amino acid sequence to the top 475 BLAST hits, excluding other Desulfovibrio spp., shows the conservation of Ala at 635. The frequency of amino acid is depicted by the size of the letter in bits. The red box highlights Ala₆₃₅ in both panels c and d.

FIG 4

D. vulgaris Hildenborough biofilm formation in bioreactors of strains with altered DVU1017 or mutants deleted for predicted cargo proteins. Representative reactors are presented; replicated reactors are shown in Fig. S1. Time 0 denotes the time at which the cultures were transitioned from batch to continuous growth in order to control the growth rate by substrate availability. Biofilm of the parental strain JWT700 (DvH-MT Δupp) is shown as black triangles. (a) Biofilm formation by DvH-MT strains lacking the entire DVU1017 gene (orange squares), a mutant lacking the DVU1017 ATP-binding domain region (bp 1543 to 2331 [green diamonds]), or the latter strain complemented with DVU1017 (blue circles). (b) SNP mutations were introduced into JWT700, resulting in DVU1017 amino acid changes of A635P or A635L (dark green circles and light green X, respectively). (c) Biofilm formation by the DvH-MO DVU1017 Pro₆₃₅ parental strain (JW710 Δupp [dark green squares]), a DVU1017(P635A) revertant (blue circles). (d) Single and double mutants of the putative T1SS-secreted, biofilm structural proteins DVU1012 and DVU1545 (red squares and gold diamonds, respectively, and a purple X for the double mutant) were created in the DvH-MT background and tested for altered biofilm formation. Error bars denote the standard deviations from triplicate samples in the protein assay; most are within the symbols.

FIG 5

Electron microscopy images of DVU1012 (a) and DVU1545 (b) antibody-treated DvH-MT cells. Antibody binding of representative cells of the DvH-MT parental strain JWT700, the putative biofilm structural protein Δ(DVU1012 DVU1545) double mutant, and the ΔDVU1017 ABC transporter mutant. Binding is depicted by black spots of 10-nm-diameter gold particles, and examples are highlighted by red arrows. Images are enlargements of the white boxed region from the inset. Scale bars in the insets represent 200 nm.

FIG 6

DvH ortholog arrangement (a) compared to the lap operon (b) of Pseudomonas putida KT2440. Arrows represent the relative size and position of the predicted coding regions, and the arrowhead represents the direction of transcription. The T1SS and DVU1012 region of DvH (DVU1012 to DVU1020) was compared to the lap operon (PP0168 to PP0164) of P. putida (43). Genes of the same color are presumed orthologs. For select genes, the amino acid length for the encoded protein is displayed below the encoding D. vulgaris Hildenborough gene.