Vibrio fischeri σ⁵⁴ Controls Motility, Biofilm Formation, Luminescence, and Colonization

In a search for flagellar mutants of V. fischeri, we identified a transposon insertion mutant, KV618 (Table 1), that failed to migrate in tryptone-based soft agar plates. We complemented the defect in this strain by introducing a plasmid library of BglII-digested chromosomal DNA and screening for motility in soft agar plates (38). From a motile clone, we isolated plasmid pES2-2 (Table 1), which carried a 2.5-kb chromosomal insert resembling the rpoN locus from a number of other organisms, including E. coli, V. harveyi, and V. cholerae (11, 13, 14). It contained putative genes for an ABC-type transporter (yhbG, open reading frame 1 [ORF1]), σ⁵⁴ (rpoN), a σ⁵⁴ regulatory gene (yhbH, ORF 95), and a nitrogen regulatory phosphotransferase component (ptsN) (Table 2). We cloned the DNA flanking the transposon insertion and determined that the transposon had inserted into codon 10 of the putative rpoN gene.

To characterize the rpoN gene further, we first constructed a null mutation in the symbiosis-competent wild-type V. fischeri strain ES114. We introduced plasmid pLMS71 (Table 1) into ES114 by conjugation and isolated a nonmotile recombinant, designated KV1513 (Table 1). We confirmed the identity of the mutant with Southern analysis. The mutation introduces a KpnI site; therefore, we probed KpnI-digested chromosomal DNA with pES2-2 and found, as expected, that the mutant contained two smaller bands in place of the single larger band of the wild-type strain.

Examination by transmission electron microscopy revealed that KV1513 lacked flagella (data not shown). Introduction of pES2-2 restored motility to KV1513 (data not shown), although migration of the complemented strain on soft agar plates was delayed relative to that of the plasmid-bearing wild-type counterpart. These data suggest that complementation with rpoN on a low-copy-number plasmid is not optimal for the motility of V. fischeri. Prolonged incubation of the rpoN null mutant KV1513 on soft agar plates failed to yield motile revertants (data not shown), an observation similar to that made with null mutants of V. cholerae rpoN (13) and V. fischeri flrA (22). These results suggest that V. fischeri motility absolutely requires σ⁵⁴, which likely operates in conjunction with FlrA.

In V. cholerae, σ⁵⁴ controls flagellar gene transcription on at least two levels (13, 27). In conjunction with the transcriptional activator FlrA, σ⁵⁴ induces transcription of the class II genes flrB and flrC, which encode a two-component signaling pathway. With FlrC, σ⁵⁴ then activates class III gene transcription, including the major flagellin subunit, flaA. Since V. fischeri motility also requires flrA and flaA homologs (22; D.S. Millikan and E.G. Ruby, unpublished data), we determined whether rpoN controls transcription of the V. fischeri flaA gene. Into the rpoN mutant KV1513 and its wild-type parent we introduced the plasmid pDM104-6 (Table 1), which carries a lacZ reporter gene fused downstream of the flaA promoter. We grew the resultant transconjugants in SWT medium (2) and measured β-galactosidase activity when the cells reached the mid-exponential phase of growth (A₆₀₀ = 1.1 to 1.3). Transcription of the reporter gene was reduced by 10-fold in the rpoN mutant, a result similar to that achieved by the flrA mutant (Table 3). These data confirm that σ⁵⁴ plays a role in motility by activating transcription of at least one flagellar gene.

In some organisms, biofilm formation is enhanced by the presence of flagella (25, 26, 33, 37). Therefore, we asked whether the ability of V. fischeri to form biofilms in culture was enhanced by the presence of the rpoN gene. To assay biofilm formation, we pregrew cells in SWT at 28°C with shaking and then transferred the cultures to glass test tubes. The cells were incubated without shaking for 10 h at 28°C and then exposed to a solution of 1% crystal violet to visualize cells that had formed a biofilm on the test tube (25). After further incubation for 15 min, the tubes were rinsed with distilled H₂O. Biofilms formed at the air-liquid interface were stained purple. Under these conditions (Fig. 1A) and others (data not shown), the rpoN mutant formed a biofilm that differed from that of the wild type: it was consistently broader and stained less intensely. Since levels of growth of the rpoN mutant and its wild-type parent were similar (Fig. 1B), these data support a role for rpoN in biofilm formation. To control for the role of flagella in biofilm formation, we examined the biofilm-forming ability of another mutant, KV661 (Table 1), that cannot form flagella because of a disruption in the key biosynthetic flgB operon. Since this mutant produced biofilms similar to those of its wild-type parent (data not shown), the role of rpoN in biofilm formation must be independent of its role in flagellar gene expression.

In a number of organisms, a functional copy of the rpoN gene is required for the assimilation of nitrogen from nonammonia sources, e.g., glutamine and serine (19, 28). Thus, we examined growth of the V. fischeri rpoN mutant in the presence of a variety of nitrogen sources. The rpoN mutant was not defective for growth in rich media, such as SWT, at a variety of temperatures (data not shown), nor was it defective for growth in a glucose minimal medium (MM-G) (29) containing ammonium chloride and Casamino Acids (Fig. 2A), ammonium chloride (Fig. 2B), or glutamine (data not shown). In contrast to its parent, the rpoN mutant grew poorly in MM-G containing serine as the sole nitrogen source (Fig. 2C). Thus, the V. fischeri rpoN gene controls assimilation of nitrogen from serine.

In V. fischeri, siderophore production and thus iron uptake depend upon GlnD (6), a protein that modulates the E. coli and Salmonella enterica response to nitrogen availability (reviewed in reference 16). The pathway by which V. fischeri GlnD regulates siderophores remains unknown. However, in E. coli, GlnD (the uridylyltransferase/uridylyl-removing enzyme) modulates PII, a protein whose activity ultimately affects activation of the σ⁵⁴-dependent transcriptional activator NtrC. We anticipated, therefore, that a defect in σ⁵⁴ might affect iron uptake in a manner similar to that of the V. fischeri glnD mutant. Instead, the rpoN mutant retained a significant ability to sequester iron from the environment, in contrast to the isogenic glnD mutant (data not shown). Although the rpoN mutant grew more slowly than did its wild-type parent on chrome azurol S siderophore assay plates, its growth was not significantly different from that of wild-type cells in the presence of the iron chelator ethylenediamine-di(o-hydroxyphenyl-acetic acid) (data not shown). Thus, GlnD likely controls siderophore production by a pathway that does not involve rpoN.

Luminescence regulation in V. fischeri requires the prototypical quorum-sensing system, which consists of the transcriptional activator LuxR (a member of the TetR family) and the autoinducer synthase LuxI (4). Regulation of V. fischeri luminescence also involves homologs of the V. harveyi luxO and luxR (distinct from the V. fischeri luxR) genes (3, 15, 23). Inactivation of the V. fischeri luxR homolog, litR, delayed and reduced light emission in culture (3), while inactivation of the luxO homolog in V. fischeri strains MJ1 and ES114 resulted in increased luminescence (15, 23). The latter phenotype is identical to that of a V. harveyi luxO mutant (1).

Because the V. harveyi LuxO protein functions as a σ⁵⁴-dependent transcriptional activator (14), we asked whether regulation of bioluminescence in V. fischeri involves the rpoN gene. We monitored the light emission of the rpoN mutant and its wild-type parent during growth of the cells in SWT medium. Under a variety of aeration, medium, and temperature conditions, the rpoN mutant consistently achieved higher bioluminescence levels, such as the three- to fourfold difference at the peak of luminescence shown in Fig. 3. This result is similar to that seen with a V. fischeri luxO mutant (15). This observation supports a role for rpoN in repressing luminescence in V. fischeri, likely in conjunction with the luxO homolog. The target for these regulators remains to be determined.

Because colonization of E. scolopes requires flagella and σ⁵⁴ controls flagellar biogenesis, we determined whether colonization requires σ⁵⁴ by exposing juvenile squid to the rpoN mutant KV1513 or to its wild-type parent. Consistent with previous reports of nonmotile strains (5, 22), the rpoN mutant did not initiate symbiotic colonization, as measured by symbiotic bioluminescence and viable counts of squid-associated bacteria in a representative experiment (seven squid). In contrast to wild-type-exposed squid, which contained on average 3.46 × 10⁵ CFU/squid, animals inoculated with KV1513 either were uncolonized (four of seven animals) or exhibited colonization levels of <100 CFU/squid (three of seven animals). The apparent low level of colonization in the latter squid group likely reflects bacteria aggregated on the surface of the light organ rather than bacteria present inside the light organ, as previously reported (22, 24).

Complementation with a functional copy of rpoN (pLD6) restored the ability to initiate symbiotic colonization (84 to 90% of animals inoculated with the rpoN mutant versus 90 to 96% of animals inoculated with the wild-type parent). This rate of initiation was greater than that reported (49%) for the complemented flrA mutant (22) and likely reflects differences in the ways the assays were performed. The colonization levels achieved by squid inoculated with the complemented rpoN mutant varied, but some animals achieved wild-type levels of colonization. Thus, while these data clearly demonstrate a role for rpoN in symbiotic initiation, a role for this regulator in subsequent stages cannot be assessed with the tools currently available; further investigation of its role in symbiotic colonization will require construction of strains with a tightly regulated promoter upstream of the rpoN gene.

This work shows that σ⁵⁴ of V. fischeri plays multiple roles. It controls flagellar biogenesis and, thus, motility. It contributes to biofilm development, nitrogen assimilation, and the regulation of bioluminescence. Finally, σ⁵⁴ plays an essential role in the establishment of symbiotic colonization, most likely due to its requirement for motility. Whether other σ⁵⁴-dependent traits contribute to symbiotic initiation or to later stages of colonization remains to be determined. This work provides a foundation for a deeper understanding of the contributions to symbiotic colonization of σ⁵⁴-dependent determinants such as motility, bioluminescence, and biofilm development.

Nucleotide sequence accession number. The 2.5-kb chromosomal insert from plasmid pES2-2 has been assigned GenBank accession number AY082659 .

We thank Cindy R. DeLoney for examining the rpoN mutant by transmission electron microscopy, Erika Simel for her assistance in cloning and sequencing the rpoN gene, Therese M. Bartley for expert technical assistance, and members of our labs for critical reading of the manuscript.

This work was supported by NIH grant GM59690 awarded to K.L.V.