The LacI-Family Transcription Factor, RbsR, Is a Pleiotropic Regulator of Motility, Virulence, Siderophore and Antibiotic Production, Gas Vesicle Morphogenesis and Flotation in Serratia

Abstract

Gas vesicles (GVs) are proteinaceous, gas-filled organelles used by some bacteria to enable upward movement into favorable air/liquid interfaces in aquatic environments. Serratia sp. ATCC39006 (S39006) was the first enterobacterium discovered to produce GVs naturally. The regulation of GV assembly in this host is complex and part of a wider regulatory network affecting various phenotypes, including antibiotic biosynthesis. To identify new regulators of GVs, a comprehensive mutant library containing 71,000 insertion mutants was generated by random transposon mutagenesis and 311 putative GV-defective mutants identified. Three of these mutants were found to have a transposon inserted in a LacI family transcription regulator gene (rbsR) of the putative ribose operon. Each of these rbsR mutants was GV-defective; no GVs were visible by phase contrast microscopy (PCM) or transmission electron microscopy (TEM). GV deficiency was caused by the reduction of gvpA1 and gvrA transcription (the first genes of the two contiguous operons in the GV gene locus). Our results also showed that a mutation in rbsR was highly pleiotropic; the production of two secondary metabolites (carbapenem and prodigiosin antibiotics) was abolished. Interestingly, the intrinsic resistance to the carbapenem antibiotic was not affected by the rbsR mutation. In addition, the production of a siderophore, cellulase and plant virulence was reduced in the mutant, whereas it exhibited increased swimming and swarming motility. The RbsR protein was predicted to bind to regions upstream of at least 18 genes in S39006 including rbsD (the first gene of the ribose operon) and gvrA. Electrophoretic mobility shift assays (EMSA) confirmed that RbsR bound to DNA sequences upstream of rbsD, but not gvrA. The results of this study indicate that RbsR is a global regulator that affects the modulation of GV biogenesis, but also with complex pleiotropic physiological impacts in S39006.

Keywords: Serratia; antibiotics; gas vesicles; gene regulation; motility; ribose operon; virulence.

Figure 1

The rbsR mutation affects the GV phenotype in S39006. (A) The genetic organization of the gas vesicle genetic cluster in S39006. The first operon begins with gvpA1 and the second with gvrA. Regulatory genes are indicated with an R. (B) The genetic organization of the ribose operon in S39006. The arrows indicate the positions of transposon insertions identified in this study. (C) Predicted protein domains in the rbsR amino acid sequences. (D) Comparison of the wild type and the rbsR mutant. Flotation assays of the wild type and the rbsR mutant 48 h after inoculation (top image). Bacterial culture spot test on plates with normalized cell number (middle image). PCM images from bacterial colonies on agar plates; the scale bar represents 1 μm (bottom image). (E) Representative TEM images showing the rsbR mutant with no GVs. The scale bar at the right bottom represents 500 nm. (F) The effect of ectopic expression of rbsR from pBAD-rbsR in the rbsR mutant at the indicated concentrations of arabinose. The top image indicates flotation assays of the wild type, the rbsR mutant carrying the empty plasmid or the rbsR mutant carrying pBAD-rbsR 48 h after inoculation. The middle image shows bacterial patches on plates with normalized cell number. The bottom image shows PCM images from bacterial colonies on an agar plates; the scale bar represents 1 μm.

Figure 2

The expression of gvpA1 and gvrA in wild type and rbsR backgrounds under aerobic and microaerophilic conditions. β-glu activity from a chromosomal gvpA1::uidA fusion strain was assayed in wild type (blue) and the rbsR mutant (red) under (A) aerobic or (C) microaerophilic conditions. β-glu activity was measured from a chromosomal gvrA::uidA fusion in the wild type (blue) and rbsR background (red) under (B) aerobic or (D) microaerophilic conditions. Solid lines represent β-glu assays, dashed lines represent the optical density (OD₆₀₀) of wild type and the rbsR mutant. Data shown are mean values ± SD (n = 3). ANOVA two-factor analysis comparing β-glu activity of the indicated fusion in wild type to the rbsR mutant throughout growth in (A) found F = 1344.01 > F_crit = 4.20, p = 3.28^*10⁻²⁵; in (B) F = 1027.51 > F_crit = 4.20, p = 1.29^*10⁻²³; in (C) F = 990.65 > F_crit = 4.20, p = 2.13^*10⁻²³; and in (D) F = 464.83 > F_crit = 4.20, p = 5.61^*10⁻¹⁹.

Figure 3

The disruption of rbsR has diverse effects on S39006 physiology. The production of (A) the carbapenem antibiotic and (B) prodigiosin in a gvpA1::uidA fusion strain (blue) and the rbsR mutant (red) throughout growth in LB. Solid lines represent carbapenem or prodigiosin production and dashed lines represent the optical density (OD₆₀₀) of the bacterial culture. Data shown are means ± SD (n = 3). (C) Carbapenem resistance assay. The carbapenem producing strain Erwinia carotovora subsp. carotovora (Ecc) ATTN10 (left spot on all plates in C) and carbapenem resistant mutant, Ecc ATTn10 SM10 (right spot on all plates in C) were spotted on the top agar lawn seeded with the rbsR mutant (top plate), S39006 (middle plate, positive control), or E. coli ESS (bottom plate, negative control) and incubated at 30°C for 48 h. (D) Carbapenem production assay of wild type and the rbsR mutant; S39006 ΔcarA was used as a negative control. Test strains with normalized bacterial cell number were spotted on a lawn of E. coli ESS and grown at 30°C for 48 h. The production of (E) cellulase and (F) siderophore, plus (G) swarming and (H) swimming motility of the wild type and the rbsR mutant is shown. Overnight cultures of wild type and the rbsR mutant with normalized cell number were spotted on appropriate agars or indicator plates.

Figure 4

The predicted RbsR-binding motif of S39006. (A) Predicted binding sequence of RbsR generated by MEME. (B) Sequence alignment of representative binding sequences of RbsR in the S39006 genome.

Figure 5

RbsR bound to Cy5-LUEGO-based rbsD fragment but not gvrA fragment in an EMSA assay. EMSA titrations of the indicated concentration of RbsR protein with (A) rbsD or (B) gvrA fragment. The competition experiment was carried out as a control with excess (20X) unlabeled specific (S) probe and nonspecific (NS) probe.

Figure 6

The rbsR mutation reduced plant virulence compared to wild type. (A) Comparison of rotted tissues produced by wild type and the rbsR mutant after 5 days of incubation with initial inoculation of 10⁶ cfu of bacterial cells. Values are the average of three biological replicates and error bars indicate ± SD, asterisk indicates significant differences (t-test, p < 0.05). (B) Representative potato with rotted tissue areas (stained by iodine to enhance clarity) injected with either wild type or the rbsR mutant.