![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Microbiology and Immunology, Loyola University Chicago, Maywood, Illinois 60153,1 Department of Medical Microbiology and Immunology, University of Wisconsin—Madison, Madison, Wisconsin 537062
Received 13 February 2007/ Accepted 8 June 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
One large-scale approach to understanding the biology of an organism is to mutagenize regulatory genes (such as two-component response regulators [RRs]), as each mutation potentially impacts the expression and/or function of a number of proteins. Both gram-negative and gram-positive bacteria utilize two-component signal transduction systems to sense and respond to different environmental stimuli, such as nutrient availability, pH, osmolarity, and host factors (5, 10). These signaling pathways are known to influence numerous processes in bacteria, including motility, bioluminescence, biofilm formation, and pathogenesis (31). Simple two-component signal transduction systems utilize two main protein components: (i) the sensor histidine kinase protein (SK), which functions as a dimer to autophosphorylate at a conserved histidine residue (using ATP as a phosphodonor) in response to a specific signaling event, and (ii) the RR, which catalyzes transfer of phosphate from the cognate SK to a conserved aspartate residue on its receiver or REC domain. More-complex "hybrid" systems consist of multiple phosphoryl-accepting histidine and aspartate residues within the same pathway. Regardless of the complexity of the upstream signal transduction pathway, the RR ultimately changes its activity upon phosphorylation to elicit a cellular response to the initial signal.
Typically, RRs contain two functional domains, a REC domain that participates in signal transduction and an effector domain necessary for the cellular response. Residues in the activation domain that are highly conserved among RRs provide a basis for identifying putative RRs. Studies of a number of such regulators, such as the well-characterized CheY, provide a basis for understanding the function of conserved residues in the activation domain. In Escherichia coli CheY, six amino acid residues play key roles in activation. Phosphorylation occurs on Asp57 (9), while two additional Asp residues, at positions 12 and 13, chelate a Mg2+ ion necessary for transfer of the activating phosphoryl group from the donor His (81). Three other amino acids, Thr87, Tyr106, and Lys109, convey the phosphorylation-associated conformational change that facilitates effector function (6, 12, 37, 38, 70).
In addition to their conserved activation domains, RRs contain effector domains responsible for eliciting a cellular response to stimulation. Non-DNA-binding RRs (such as CheY) may directly impact protein function by affecting, for example, folding or stability; as a result they may also indirectly impact transcription of downstream target genes. DNA-binding RRs typically fall into three distinct subclasses, exemplified by NarL, OmpR, and NtrC (74). Members of each subclass maintain unique DNA sequence specificities and mechanisms of action. Members of the NarL subclass contain a classical helix-turn-helix (HTH) DNA-binding motif (4). Those of the OmpR subclass share a winged helix-turn-helix (wHTH) motif. In this latter motif, the traditional recognition helix used in major groove interactions with DNA is flanked by extended looped regions, which may interact with the minor groove of the target DNA sequence (46). The third, NtrC-like subclass consists of a structurally distinct group of RR proteins that serve as coactivators for RNA polymerase holoenzymes associated with the alternative sigma factor 
54, and members of this subclass are known as 54-dependent RRs. These proteins contain two distinct effector domains. One domain binds DNA, while the other, approximately 240 amino acid residues in length, is responsible for oligomerization and ATP hydrolysis. Together with RNA polymerase containing 54, the two RR domains promote open complex formation during transcriptional initiation (54, 55, 59, 68, 80).
The importance of two-component regulators is underscored by their abundance in many sequenced bacterial genomes. For example, E. coli contains 32 two-component regulators (53). In V. fischeri, only five RRs have been described thus far: GacA, LuxO, ArcA, SypG, and SypE (8, 52, 83, 84, 90); of these, only the first three have been characterized by genetic disruption. Mutants that lack GacA exhibit pleiotropic effects in culture, fail to initiate symbiotic colonization (84), and fail to induce host development (83). LuxO and ArcA both serve as negative regulators of luminescence genes (8, 40-42, 52); however, of these two, only LuxO appears necessary during symbiotic initiation, suggesting that it controls symbiosis genes in addition to those governing luminescence (40-42, 52). Given that so few of these regulators have been characterized, we chose to identify, mutagenize, and characterize specific phenotypes mediated by V. fischeri RRs. This work demonstrates the feasibility of performing global screens for gene function in V. fischeri and represents an initial characterization of many two-component signal transduction pathways in this organism.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Strains and media.
V. fischeri strain ES114 was the parent strain used in this study (7). V. fischeri strain KV1421 (58), which contains a chromosomal erythromycin resistance cassette (Tn7::erm) (48), was used as a control for motility and bioluminescence experiments. E. coli strains DH5
pir (17), CC118pir (29), and TAM1pir (Active Motif, Carlsbad, CA) were used for cloning. Triparental matings to introduce plasmid DNA into V. fischeri utilized E. coli strain CC118pir carrying the conjugation helper plasmid pEVS104 (73).
V. fischeri strains were grown in the following media. To test growth in a minimal medium, HEPES-MM was used (64). For routine culturing, either LBS (72) or SWT medium was used; the latter contains 0.5% tryptone, 0.3% yeast extract, and seawater salts (90). For motility experiments, cells were grown in TB-SW, a tryptone-based medium containing seawater salts, to maintain consistency with previously reported assays of V. fischeri motility (15). Previous work within the field indicates that increased medium osmolarity correlates with increased luminescence in culture (71). Thus, additional NaCl was added to SWT to a final concentration of 510 mM to assay for bioluminescence; we designated this medium as SWTS. E. coli strains were grown in LB (14), brain heart infusion (Difco), or SOC (66) medium. Where appropriate, antibiotics were added to the following final concentrations: chloramphenicol at 25 µg ml–1 for E. coli and 2.5 µg ml–1 for V. fischeri, kanamycin at 50 µg ml–1, and erythromycin (ERY) at 5 µg ml–1 for V. fischeri and 150 µg ml–1 for E. coli. Agar was added to a final concentration of 1.5% for solid media.
Molecular and genetic techniques. Standard molecular biology techniques were used for all plasmid constructions. Restriction and modifying enzymes were purchased from New England Biolabs (Beverly, MA) or Promega (Madison, WI). To generate vector integration mutations (11) in each RR, DNA oligonucleotides used for amplifying internal fragments of RR genes (Table 1) were obtained from MWG Biotech (High Point, NC). Amplified PCR fragments were cloned into the oriR6K-based suicide vector pEVS122 (17), which was digested with the restriction enzyme SmaI. In some cases, internal fragments of RRs first were cloned using the TOPO TA cloning kit per the manufacturer's instructions (Invitrogen, Carlsbad, CA) and then subsequently subcloned into either pEVS122 or its derivative, pESY20 (58). Strains of E. coli carrying internal fragments of each RR in pEVS122 or pESY20 were used in triparental conjugations to introduce the DNA into wild-type V. fischeri strain ES114. ERY-resistant colonies that arose were presumptive RR mutants. To confirm each mutant, Southern blot analysis was performed as previously described (79) using digoxigenin-labeled pEVS122 as a probe to detect complementary sequences (Roche Molecular Biochemicals, Indianapolis, IN). Hybridization and detection were carried out as described previously (79). RR mutants resulting from these manipulations are shown in Table 2.
|
|
Luminescence assays. To assay for potential bioluminescence defects, the RR mutants and ES114 were grown overnight in SWT- and subcultured into SWTS to a starting optical density at 600 nm (OD600) of approximately of 0.03. At various times after inoculation, 1-ml samples were taken for luminescence and OD measurements. A TD-20/20 luminometer (Turner Biosystems, Sunnyvale, CA) set at factory settings was used to determine the relative light units integrated over a 6-second count.
Motility assays.
To assay motility, KV1421 (58) and individual RR mutants were grown to the mid-exponential phase of growth (OD, 
0.3) in TB-SW, and 10-µl aliquots were inoculated onto TB-SW plates containing 0.225% agar (15, 85). Each mutant was spotted onto an individual plate along with a wild-type control. Over the course of 4 to 5 hours of incubation at 28°C, the diameter of migrating rings was measured.
Competitive symbiosis and growth assays.
To assess the colonization capability of mutant V. fischeri strains, juvenile squid were inoculated into artificial seawater (Instant Ocean; Aquarium Systems, Mentor, OH) containing between 1,000 and 6,000 cells per ml of seawater. For competitive colonization assays, animals were inoculated with an approximately 1:1 ratio of mutant and wild-type cells. In the individual inoculation experiment, animals were inoculated with ES114 or KV1787 (
sypG) alone. The presence of bacterial strains within the light organs of juvenile squid was assessed through luminescence and homogenization/plating assays as described previously (63). The limit of detection based on these methods is 14 V. fischeri cells per squid. In competitive assays, colonization is reported as the log-transformed relative competitive index (log RCI). This number is generated by dividing the ratio of mutant to wild-type cells within the homogenate by the ratio present in the inoculum and subsequently determining the log10 value of that number.
To assess competitive growth in culture, V. fischeri RR mutant and marked wild-type strains were coinoculated into SWT medium at an approximately 1:1 ratio and the inoculum was plated. After at least 20 generations of growth in SWT, cells were diluted and plated again. The log RCI was calculated as described above.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
The RRs in the second group contain, in addition to REC domains, HTH DNA-binding motifs found in RRs such as NarL (26) (Table 4). This class of six proteins includes GacA, as well as proteins with sequence similarity to known E. coli RRs. These include potential homologs of NarP, a regulator of genes involved in anaerobic respiration during nitrate limitation (61); UhpA, which regulates hexose phosphate transport (23, 32, 82); and YehT, an RR involved in multidrug resistance (30).
|
|
54-dependent RRs with sequence similarity to the nitrogen regulator NtrC (Table 6). In addition to REC and DNA-binding domains, proteins in this class contain a nucleotide binding domain required for 54-dependent transcriptional activation (59, 80). In addition to an NtrC homolog, likely encoded by VF0095, this group includes the luminescence regulator LuxO and VF1854, a putative homolog of the flagellar synthesis regulator FlrC, based on similarity to the Vibrio cholerae protein (34). SypG is another 54-dependent activator, which regulates expression of the syp cluster of genes (90). VF0454 exhibits similarity to the V. cholerae regulator of polysaccharide synthesis, VpsR. VF1401 has no previously characterized putative homolog.
The genes of cognate RR and SK pairs are often physically linked on the chromosome. Therefore, we extended our analysis of RR genes to examine the functions of nearby genes. Using BLAST analysis, genes encoding the conserved SK domain involved in dimerization and phosphorylation, termed HisKA, were considered putative SKs. Interestingly, each of the RRs within the OmpR-like family is linked to a putative SK (Table 5). The RRs of the HTH and 54-dependent families each contain five (of six) seemingly paired members (Tables 4 and 6). In contrast, only 7 of the 15 RRs listed in Table 3 (lacking a DNA-binding motif) appear linked to SK genes.
Finally, we analyzed the distribution of RRs across each of V. fischeri's two chromosomes. The larger chromosome, chromosome 1 (2.9 Mb, designated VF), encodes 22, or about half, of the RRs, 16 of which have known or tentatively assigned identities. The remaining 18 RRs are encoded on the smaller chromosome, chromosome 2 (1.3 Mb, designated VFA), and only five of them have assigned identities. Thus, in addition to the imbalance in the number of RRs with tentative assignments, the proportion of RRs on each chromosome is not representative of the size of the respective chromosomes. With two exceptions, we have not confirmed the putative identities of these genes; however, to aid the reader, we will include the putative assignments of the RR genes in parentheses following their gene numbers throughout this work.
Construction of RR mutants. To begin to characterize the two-component signal transduction systems of V. fischeri, we constructed mutant strains defective for individual V. fischeri RR genes. We chose to eliminate from our study the RR gacA gene (VF1627), as it has been characterized extensively, but included luxO (VF0937) and arcA (VF2120) as controls.
To mutate the RR genes with relative ease, we cloned an internal portion of each RR gene into a suicide vector (pEVS122 [17] or pESY20 [58]), marked with ERY resistance (Eryr). Recombination of these constructs into the V. fischeri genome generates two truncated copies of the gene, which generally results in a null mutation (11). Although such mutations have the potential for polar effects on downstream genes (a possibility for numerous RR genes disrupted in this study), we will refer to our mutants simply as RR mutants; future work such as complementation and genetic analyses of the operon structure will be necessary to confirm the specific roles of RRs identified in this survey. Potential mutants were isolated by selection for Eryr and subsequently confirmed using Southern blot analysis (see Materials and Methods). Resistant colonies arose at a frequency generally proportional to internal fragment length (i.e., larger fragments typically resulted in more-numerous resistant exconjugants). Although rarely, Southern analysis indicated that some isolates appeared to contain two copies of the suicide vector integrated in the RR gene. These strains were not used for further study. We constructed mutations in 35 of the 39 targeted RR genes in this manner. The remaining four RR genes (VF0923, VF1830 [cheB], VFA0436, and VFA0608) were recalcitrant to disruption despite multiple attempts; further experiments are required to determine whether these represent essential genes.
Since this method of making mutations has been used to a limited extent in V. fischeri and certainly has not been utilized on such a large scale, some additional characterization of this technique was warranted. First, we probed the lower size limit of homologous DNA that could be recombined successfully into the V. fischeri chromosome. We generated derivatives of our arcA mutant construct that truncated the original 314-bp gene fragment to 101 bp and 85 bp and introduced them into V. fischeri. ERY-resistant colonies were rare but arose with greater frequency with prolonged periods of conjugation. Southern analysis verified appropriate chromosomal insertion of the smaller arcA constructs in KV1703 and KV1704. In addition, we found that these strains (as well as our original arcA mutant, KV1548) exhibited increased luminescence (Fig. 2A), consistent with the function of ArcA as a negative regulator of bioluminescence (8). These results demonstrated that an 85-bp fragment is sufficient for homologous recombination in V. fischeri; thus, this method can be readily utilized for constructing mutations in small genes.
|
Impact of RR mutations on growth and luminescence. One of the hallmarks of V. fischeri biology is the ability to grow to high cell density and produce light. To address the involvement of two-component pathways in this process, we analyzed the ability of the 35 RR mutants to grow and luminesce. Because we have previously shown that the presence of the Eryr marker decreases the luminescence and motility of V. fischeri, even in the absence of selection (58), we compared the growth and luminescence of the Eryr RR mutants to those of KV1421, a derivative of the wild-type strain also marked with Eryr. Upon plating of each mutant on solid medium composed of either the complex medium SWT or MM supplemented with a variety of carbon sources, we observed that only the arcA mutant generated consistently smaller colonies than did the control strain, KV1421. We further examined each of these strains in liquid SWT medium and found that most grew at a rate indistinguishable from that of the control. However, consistent with the observations of growth on solid medium, the arcA strain exhibited a significant, yet minor, deficit in growth, as measured by OD in SWT broth compared to that of the control (Fig. 1A). This result is consistent with previous observations (8). Conversely, a strain carrying a mutation in the putative homolog of the flagellar regulator flrC exhibited consistently higher OD measurements than did the control, indicating a possible growth advantage (Fig. 1B).
|
Consistent with previous observations, mutations in arcA and luxO resulted in increased bioluminescence at high cell density (Fig. 2A) (42). Three different mutants defective for the RR ArcA displayed high levels of luminescence; all reached a level above the limit of detection at cell densities near or above an OD600 of 2. Additionally, the luxO mutant consistently achieved specific luminescence levels about fourfold above that of KV1421 (Fig. 2B). With few exceptions, most other RR mutant strains induced light production at rates similar to those of the wild type and achieved the same maximal levels (data not shown). For example, a mutant defective for the RR SypG, which exhibits substantial sequence similarity to the luminescence regulator LuxO, was indistinguishable from KV1421 (Fig. 2C). However, we did identify three RRs that appeared to enhance bioluminescence, as mutations that disrupted VFA0698 (cheV), VF1854 (flrC), or VF1148 (yehT) caused a small but reproducible decrease in bioluminescence levels compared to the control (Fig. 2D, E, and F, respectively). These data thus reveal additional minor players (either direct or indirect) in bioluminescence control.
Role of RRs in motility. In many bacteria, numerous environmental factors influence bacterial motility, making it another phenotype likely controlled by two-component regulation in V. fischeri. Therefore, we examined the motility of the 35 RR mutants by inoculating TB-SW soft agar plates each with an RR mutant and a control and measuring the diameter of the outer chemotaxis ring over time. Again, many RR mutants, such as the sypG mutant (Fig. 3A), exhibited no significant defect in motility compared to the control. However, we identified three classes of motility phenotype mutants: (i) those that appeared nonmotile, (ii) those with decreased migration, and (iii) those with increased migration relative to the control.
|
The second class of mutants included five strains with decreases in migration (Fig. 3C). Two of these were strains defective for arcA and luxO. The arcA mutant (Fig. 3C, closed diamonds) exhibited a slight yet significant defect in migration rate. This is a phenotype not previously reported for this regulator; however, given the growth defect of this strain, verification of this result will require a growth-independent means of validation not part of this study. The luxO mutant (Fig. 3C, open triangles) exhibited the most severe motility defect of this class, with a greater-than-twofold decrease in the rate of motility on TB-SW soft agar plates relative to the control. This phenotype is consistent with that previously reported for a luxO mutant (41).
The second class of mutants also included those defective for VF0454 (vpsR) (Fig. 3C, closed squares), VF1909 (narP) (Fig. 3C, closed circles), and VF2343 (cpxR) (Fig. 3C, closed triangles). Disruption of VF2343 resulted in a slightly decreased rate of motility compared to the control. The two remaining strains were delayed in migration; each exhibited a strongly decreased rate of migration compared to that of the control until about 3 h postinoculation.
The third class of mutants contained a single strain, defective for VFA0698, which exhibited a small but reproducible increase in motility relative to the control (Fig. 3D). Intriguingly, VFA0698 is one of three RRs that exhibit sequence similarity to cheV (Table 3). In summary, these experiments supported the identification of LuxO as a motility regulator and revealed up to seven additional regulators that impact motility; additional experiments will be required to determine the mechanisms by which these RRs affect motility.
Multiple RRs play a role in colonization. Two-component regulators often mediate a bacterium's ability to sense and respond to host-specific signals during colonization. In fact, at least three such regulators, the RRs GacA and LuxO and the SK RscS, are known to affect the ability of V. fischeri to initiate colonization of the squid host (41, 79, 83, 84). To characterize our RR mutants with respect to symbiosis, we assayed the ability of the mutants to compete with the wild-type strain for colonization. In these experiments, we inoculated squid with approximately 1:1 mixtures of each RR mutant and a wild-type control. We evaluated successful colonization by measuring bioluminescence, which occurs when the bacteria have achieved a high cell density within the symbiotic light organ. We then isolated bacterial symbionts from colonized squid and determined the relative amounts of each of the two strains. To facilitate differentiation of the two strains, we marked one of the two strains using a stable plasmid that constitutively expresses lacZ, which does not affect competitive colonization (18), and plated the strains on medium containing X-Gal (5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside). Finally, we calculated the RCI by normalizing the ratio of output bacteria to the ratio of the input bacteria, which we report as a log10-transformed value (log RCI) for purposes of comparison (see Materials and Methods). A log RCI value below zero indicates a competitive disadvantage for an RR mutant; we report here only those mutants with a mean log RCI significantly less than zero as evaluated by Student's t test.
Of the 35 mutants that we tested, 23 exhibited no competitive defect in colonization. For example, squid inoculated with a strain defective for VF1396 consistently exhibited a log RCI of zero or greater (Fig. 4A). In contrast, 12 of the mutants tested exhibited a significant defect in competitive colonization in at least two independent experiments (Fig. 4B to D and data not shown). Subsequently, these 12 strains were also tested for competitive growth in culture. Most of the RR mutants that were competitively defective for colonization did not exhibit growth disadvantages in coculture (data not shown). Those that did exhibit statistically significant coculture differences are reported below.
|
We also found five mutants that exhibited no defects in growth, luminescence, or motility and yet failed to compete effectively for symbiotic colonization (Fig. 4C). They were disrupted for the following RR genes: VF0095 (ntrC), VF1689 (expM), VF1988 (phoB), VFA0179, and VFA0181. Of these, only the ntrC mutant exhibited a slight competitive growth disadvantage in coculture experiments (average log RCI of –0.553, P = 0.0018). Further studies will be required to identify the pathways governed by each of these RRs and to determine whether they represent previously uncharacterized colonization determinants.
Finally, we also found that a strain that lacked the RR SypG was severely defective in competition. SypG has been shown to regulate transcription of a cluster of genes, syp, which are required for colonization (90); however, no mutations in sypG have previously been characterized. As shown in Fig. 4D, of nine animals inoculated with labeled ES114 and the sypG mutant, all were exclusively colonized with ES114.
Because SypG was the only RR that was completely excluded from the light organ and exhibited no motility or luminescence defect, we characterized sypG further by constructing an in-frame mutation to eliminate potential polar effects on the downstream gene sypH. We tested the sypG mutant in single animal inoculations and found that under these conditions, the sypG strain was severely defective in initiating symbiosis compared to the wild-type strain (a representative experiment is depicted in Fig. 5; P = 0.01). In fact, 5 of the 10 animals inoculated with the sypG strain yielded no bacteria upon homogenization (below the limit of detection, 14 CFU/squid). These data thus supported the prediction that SypG would be an important symbiotic regulator (90) and also revealed the utility of performing competition experiments to identify genes that might be necessary for symbiotic colonization even in the absence of competition.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To identify RRs, we relied on bioinformatics approaches, searching the ERGO Light database (http://www.ergo-light.com/ERGO/) and then analyzing the candidates using rps-BLAST (44). We subsequently compared our results with data recently made available through the Center for Biological Sequence Analysis (CBS; http://www.cbs.dtu.dk/), which lists 47 putative RRs. The RRs that we identified through our analysis matched those listed by CBS, with the following exceptions. First, CBS listed 13 ORFs that we eliminated from our study, because further analysis indicated that those 13 ORFs contain both REC domains and sensory domains (i.e., the H box), and thus, they likely function as hybrid sensor kinase proteins rather than RRs. In addition, we identified five genes not listed by CBS. Four of these are CheY-like regulators that lack a DNA-binding domain: VF0923, VF1054, VFA0216, and VFA1017. Each of these contains a REC domain and most of the highly conserved residues therein (Table 3). We also included in our analysis VF0454 (vpsR), although it is unclear whether this regulator will indeed function as a RR. Alignments of this protein with LuxO (VF0937) revealed an overall conservation within the REC domain of the latter protein, supporting this assignment. However, like its V. cholerae homolog (88), VF0454 (vpsR) lacks several of the highly conserved residues in the REC domain (Table 6). Because our identifications were based on bioinformatics rather than biochemical analyses, further studies will be necessary to verify the function of these putative RRs as phosphate-accepting two-component proteins.
Our vector integration approach to interrupt the RR genes successfully yielded disruptions in 35 of the 39 RRs that we targeted for mutation with relative ease. As controls, we disrupted two previously characterized RRs, luxO and arcA. These mutations each resulted in phenotypes consistent with previous reports (8, 40-42). We can conclude, therefore, that our mutagenesis approach is capable of generating null mutations. Because many of the predicted RRs disrupted in this work appear to be encoded within operons (data not shown), the specific roles of these regulators must be addressed by complementation experiments in future work. Regardless of this caveat, this study reveals novel loci affecting downstream phenotypes such as colonization. It remains unclear why we were unable to obtain mutations in the remaining four genes. These may represent essential genes; however, this seems unlikely to be the case for VF1830 (cheB), which is predicted to play a role in chemotaxis.
Finally, we assessed the roles of the 35 RRs in two primary physiological traits in culture, bioluminescence and motility, as well as in the ability to compete with a wild-type strain for symbiotic colonization. We identified eight RR mutants with motility phenotypes, five with bioluminescence phenotypes, and 12 with symbiosis phenotypes. Of the symbiosis-defective strains, six exhibited alterations in either motility or bioluminescence or both, while six were wild type for these phenotypes. RRs required for symbiosis were found on each of the two chromosomes, with proportionally more of them mapping to the larger chromosome. They were also found in each of the four RR classes (regulators with no predicted DNA-binding sequences, with HTH or wHTH, or with 54 activation domains). Intriguingly, of the six RRs that comprise the subgroup of 54-dependent RRs, five played roles in symbiosis, indicating the importance of this regulatory subclass in symbiotic colonization.
Of the six mutants defective for both symbiosis and other traits, the nonmotile flrC (VF1854) mutant is perhaps easiest to interpret: motility is essential for the initiation of symbiotic colonization (25, 49, 50), as demonstrated by the lack of flrC mutants found in the light organ in our experiments (Fig. 4B). A cheY mutant was similarly excluded from the light organ. This mutant also failed to migrate in soft agar plates; however, consistent with the predicted role of CheY as a chemotaxis regulator, the cheY mutant was actually motile but exhibited predominantly smooth swimming in liquid medium. These experiments indicate that chemotaxis may be important for symbiotic colonization. In support of this, we have similarly found that a mutant defective for cheR, another chemotaxis regulator whose loss results in smooth swimming, was also defective for colonization (14a). However, since both cheY and cheR mutants fail to migrate through soft agar, it remains formally possible that these strains fail to colonize not due to a lack of chemotaxis but due to an inability to penetrate the squid-secreted mucus layer. It will be necessary to construct specific chemoreceptor mutants or identify a squid-secreted chemoattractant to fully address this question.
Of the five additional mutants that exhibited decreases in motility, three also displayed symbiosis phenotypes (VF0454 [vpsR], VF1909 [narP], and luxO). Thus, it is tempting to speculate that reduced motility reduces colonization proficiency. This appears not to be true, however, in the case of the VF2343 and arcA mutants, which exhibited decreased motility but remained proficient for colonization. In addition, an interpretation for the luxO mutant, which was among the most severely compromised for motility, is complicated by the role which this regulator plays in the control of bioluminescence and other genes (40-42, 52).
Motility and colonization are also linked in the case of the mutant disrupted for VFA0698 (cheV). In this case, however, the mutant displayed an increase in motility. It also exhibited a decrease in bioluminescence. VFA0698 is one of three RRs predicted to encode a CheV-like protein, with both REC and CheW domains. In Bacillus subtilis, CheV functions to couple methyl-accepting chemotaxis proteins to the SK CheA, allowing chemotactic signaling to occur (33, 62). Further experiments are required to determine whether any of the three V. fischeri CheV-like proteins performs such a function. The relative contribution to the symbiotic defect, if any, of the bioluminescence and motility phenotypes of the VFA0698 mutant remains unclear.
Perhaps most exciting are the six RRs whose loss caused a symbiosis defect but no alterations in bioluminescence or motility: VF0095 (ntrC), VF1689 (expM), VF1988 (phoB), VFA0179, VFA1081, and sypG. Of these, the most severe was the sypG mutant, which was excluded from the light organ. Because the insertional mutation in this strain could cause polar defects on a downstream gene, we further investigated the role of sypG by constructing an in-frame deletion and evaluating its ability to colonize juvenile squid in single-inoculation experiments. The results confirmed that this regulator plays an important role in symbiotic initiation (Fig. 5). The sypG gene resides within a large cluster of genes that are necessary for symbiotic colonization and function in polysaccharide biosynthesis (90). Our recent work demonstrates that sypG overexpression induces transcription of the syp genes, making it likely that this regulatory role accounts for the symbiotic requirement for SypG (90). Whether SypG controls additional colonization factors remains to be determined.
Of the remaining five RRs important for colonization, roles for VF0095 (ntrC), VF1689 (expM), and VF1988 (phoB) can be hypothesized based on bioinformatics. For example, VF1988, a putative PhoB homolog, could play a role in controlling genes based on phosphate availability. Similarly, the possible NtrC homolog VF0095 may respond to changing nitrogen levels. VF1689 may be a homolog of ExpM in Erwinia carotovora, which does not bind DNA but rather impacts the stability of the stationary-phase sigma factor S (2). The remaining regulators (VFA0179 and VFA0181) contain predicted DNA-binding domains, which will likely facilitate investigations into their functions. Intriguingly, these two regulators are encoded near one another on the chromosome. These genes might thus represent a novel symbiosis locus.
Many of the RR mutants exhibited no defects or relatively minor defects in our competitive colonization experiments; however, this does not rule out the possibility that other RRs function in symbiosis. Whereas the competitive colonization experiments described herein were terminated 24 h after inoculation, other similar assays described in the literature are often extended to 48 h to account for the ability of a given bacterial strain to persist within the light organ after initial colonization. Thus, those RRs that are essential for persistence may not be defective (or may be only slightly defective) during the initial 24 h of symbiosis. Indeed, although we found no significant competitive defect for an arcA mutant at 24 h, recent work indicates that an arcA mutant exhibits a defect within 48 h (8). Additionally, it is possible that mixed inoculation with wild-type strains may complement some RR mutant phenotypes required for colonization. For example, an RR mutant unable to produce an unknown extracellular symbiosis determinant may be able to obtain that determinant from its wild-type neighbors during inoculation, thus masking a colonization phenotype. Thus, future experiments may identify additional RRs that function in symbiosis.
In summary, our results highlight the utility of a global approach for investigating gene function in V. fischeri. Together these experiments have provided a wealth of information regarding the roles of V. fischeri RRs in a variety of phenotypes, including symbiotic colonization. These results will serve as a starting point for understanding the connections between these global regulators and the physiological traits which they impact, particularly as microarray technology is applied to the mutant strains generated here.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by NIH grants RR12294, awarded to E.G.R., and GM59690, awarded to K.L.V.
|
FOOTNOTES |
|---|
Published ahead of print on 22 June 2007.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
D-dependent gene encodes a novel protein with both CheW and CheY homologous domains. J. Bacteriol. 176:2727-2735.54-dependent transcriptional activator in Vibrio fischeri, is required for motility and symbiotic light-organ colonization. J. Bacteriol. 185:3547-3557.54 bacterial enhancer-binding protein family: mechanism of action and phylogenetic relationship of their functional domains. J. Bacteriol. 175:6067-6074.
