Inhibition of SypG-Induced Biofilms and Host Colonization by the Negative Regulator SypE in Vibrio fischeri

Vibrio fischeri produces a specific biofilm to promote colonization of its eukaryotic host, the squid Euprymna scolopes. Formation of this biofilm is induced by the sensor kinase RscS, which functions upstream of the response regulator SypG to regulate transcription of the symbiosis polysaccharide (syp) locus. Biofilm formation is also controlled by SypE, a multi-domain response regulator that consists of a central regulatory receiver (REC) domain flanked by an N-terminal serine kinase domain and a C-terminal serine phosphatase domain. SypE permits biofilm formation under rscS overexpression conditions, but inhibits biofilms induced by overexpression of sypG. We previously investigated the function of SypE in controlling biofilm formation induced by RscS. Here, we examined the molecular mechanism by which SypE naturally inhibits SypG-induced biofilms. We found that SypE’s N-terminal kinase domain was both required and sufficient to inhibit SypG-induced biofilms. This effect did not occur at the level of syp transcription. Instead, under sypG-overexpressing conditions, SypE inhibited biofilms by promoting the phosphorylation of another syp regulator, SypA, a putative anti-sigma factor antagonist. Inhibition by SypE of SypG-induced biofilm formation could be overcome by the expression of a non-phosphorylatable SypA mutant, indicating that SypE functions primarily if not exclusively to control SypA activity via phosphorylation. Finally, the presence of SypE was detrimental to colonization under sypG-overexpressing conditions, as cells deleted for sypE outcompeted wild-type cells for colonization when both strains overexpressed sypG. These results provide further evidence that biofilm formation is critical to symbiotic colonization, and support a model in which SypE naturally functions to restrict biofilm formation, and thus host colonization, to the appropriate environmental conditions.


Introduction
Bacterial biofilms, or surface-associated communities of cells encapsulated in an extracellular matrix, are ubiquitous in the environment and likely represent the preferred lifestyle mode for many bacterial species [1]. The formation of a biofilm can impact multiple aspects of a bacterium's physiology, including sensitivity to antibiotics and other antimicrobials [2], cellular metabolism [3], and gene expression [4]. Biofilms have also been demonstrated to play critical roles in mediating interactions between bacteria and their eukaryotic hosts [5,6].
The symbiotic association between the marine bacterium Vibrio fischeri and its host the Hawaiian squid Euprymna scolopes provides a natural model to study the bacterial processes necessary to promote host colonization [7,8,9]. Importantly, one of the earliest stages of the colonization process involves the formation of a specific biofilm or bacterial aggregate on the surface of the squid's symbiotic light organ [10]. Formation of this biofilm requires the symbiosis polysaccharide (syp) locus, which is regulated at the transcriptional level by the sensor kinase (SK) RscS and the syp-encoded response regulator (RR) SypG [5,11,12]. As a result, both rscS and sypG are required for initiation of host colonization [12,13]. The current model predicts that, upon detection of an as-yet unidentified signal(s), RscS autophosphorylates and serves as a phosphodonor to ultimately activate the RR SypG [12] (Fig. 1). Phosphorylated SypG is thought to directly promote transcription of the individual syp operons, which encode the structural genes necessary for polysaccharide production and thus biofilm formation [5,14].
The natural signal(s) that activate RscS remain unknown, and thus conditions under which wild-type cells form syp-dependent biofilms, other than during symbiosis, have not been established. However, overexpression of either rscS or sypG can promote the production of biofilms in laboratory culture, and provides a powerful tool to dissect the regulatory components involved in this signaling pathway. Overexpression constructs have been similarly employed in a variety of bacterial systems in which the natural signals remain unidentified, such as Vibrio cholerae and Salmonella enterica serovar Typhimurium [15,16,17]. The overexpression of rscS in wild-type cells results in the activation of syp transcription and syp-dependent biofilm formation, in particular the formation of wrinkled colonies on solid agar media and pellicles in static liquid cultures [11,12]. The formation of these biofilms depends on an intact, functional SypG [12]. rscS overexpression also results in increased symbiotic aggregation, which correlates with a colonization advantage over wild-type cells [11].
Intriguingly, overexpression of sypG in wild-type cells similarly induces activation of the syp locus, but fails to promote biofilm phenotypes as observed for RscS [12]. In a previous study, Hussa et al. (2008) demonstrated that the inability of sypG overexpression to induce syp biofilm formation was due to the inhibitory activity of a second syp-encoded RR, SypE; overexpression of sypG in a DsypE mutant resulted in dramatic biofilm formation similar to that observed with rscS-overexpressing wild-type cells [12]. Complementation with a wild-type allele of sypE fully restored the wild When sypG is overexpressed, transcription of the syp locus is activated, resulting in the production of Syp structural proteins necessary for polysaccharide production and biofilm formation, as well as the production of regulatory proteins SypA and SypE. Our data here show that, under these conditions, SypE functions as a serine kinase and phosphorylates the downstream target protein, the putative anti-sigma factor antagonist protein SypA. Phosphorylated SypA is inactive to promote biofilm formation and host colonization. In contrast, when rscS is overexpressed, the kinase activity of SypE is inactivated (presumably through phosphorylation of SypE's REC domain); instead, it functions as a serine phosphatase to dephosphorylate SypA, which promotes biofilm formation and colonization through an unknown mechanism. Co-overexpression of sypG and sypA (not shown) leads to biofilm formation, presumably because high levels of SypA permit some SypA to escape phosphorylation and inactivation by SypE. doi:10.1371/journal.pone.0060076.g001 type, non-biofilm-forming phenotypes [12]. Two main conclusions can be drawn from these studies: 1) RscS and SypG both function to induce biofilm formation, and 2) RscS may have an additional function in controlling SypE activity that SypG lacks.
SypE is a multi-domain RR consisting of a central RR receiver (REC) domain flanked by an N-terminal, serine kinase domain and a C-terminal, serine phosphatase domain (Fig. 1A) [18,19,20]. The mechanism by which SypE naturally inhibits biofilms under SypG-inducing conditions has not yet been investigated. However, we recently characterized the role of SypE in biofilms formed when rscS is overexpressed. Under these conditions, deletion of sypE slightly, but reproducibly delayed the formation of wrinkled colonies. This phenotype could be complemented by expression of the full-length SypE protein from an exogenous location in the chromosome. In addition, it could be complemented by expression of SypE's C-terminal, serine phosphatase domain alone. Surprisingly, however, expression of SypE's N-terminal kinase domain alone not only failed to complement, but in fact fully inhibited biofilms induced by rscS overexpression [19]. These data support the idea that RscS must modulate the activity of SypE such that it functions primarily to promote biofilm formation.
These studies also predicted that SypE regulates biofilms by controlling the phosphorylation state of a downstream regulatory protein. That downstream regulator has since been identified as SypA, a putative anti-sigma factor antagonist ( Fig. 1B) [20]. Under rscS-overexpressing conditions, SypA is largely unphosphorylated and promotes biofilm formation [20]. Furthermore, SypA is largely phosphorylated, even when rscS is overexpressed, when the cells express a mutant SypE protein (SypE D192A ) that constitutively inhibits biofilm formation. SypE thus mediates control over biofilms induced by rscS overexpression by controlling the phosphorylation state of SypA [20].
These previous studies investigated the regulation of biofilms induced by overexpression of rscS, a condition in which SypE functions as a positive regulator. Our goal here was to further our understanding of SypE function by determining whether its ability to naturally inhibit biofilm formation induced by sypG overexpression occurs through mechanisms similar to that observed with RscS overexpression. Indeed, we obtained compelling evidence that under sypG overexpressing conditions, SypE functions as a kinase to inhibit biofilm formation: (1) SypE's Nterminal kinase domain was both required and sufficient to mediate inhibition of SypG-induced biofilms and (2) SypE inhibits biofilms in sypG-overexpressing cells by promoting phosphorylation, and thus inactivation, of the regulatory protein SypA. In addition, we further extend our understanding of SypE function by determining that the impact of SypE occurs at a level downstream of syp transcription. Finally, we find that the ability of SypE to inhibit biofilm formation in laboratory culture is mirrored in its ability to inhibit symbiotic colonization of the squid host. Together, these data demonstrate that the negative regulation of SypG-induced biofilms by SypE is a critical mechanism by which V. fischeri restricts host colonization.

Bacterial Strains, Plasmids, and Media
Wild-type (WT) V. fischeri ES114 [21] was used as the parent strain for these studies. Strains derived from ES114 and plasmids utilized in this study are listed in Tables 1 and 2, respectively. E. coli strains TAM1 (Active Motif, Carlsbad, CA), TOP10 F (Invitrogen, Carlsbad, CA), and GT115 (InvivoGen, San Diego, CA) were used for cloning. To generate Tn7 insertions in V. fischeri, tetraparental matings were carried out as previously described [22]. To generate the sypG overexpression plasmid pCLD56, restriction digest was utilized to remove the Cm R cassette from plasmid pEAH73 [12]. To generate the various sypE and sypA alleles, we performed PCR using the primers listed in Table 3. The resulting PCR products were cloned into the cloning vector pJET1.2 (Fermentas, Glen Burnie, MD) and then subcloned into mobilizable plasmid pVSV105 [23], pKV282 [19], or pEVS107 [22]. V. fischeri strains were grown in LBS [24] or sea water tryptone (SWT) [5]. Escherichia coli strains were grown in LB [25] or brain heart infusion medium (Difco, Detroit, MI). The following antibiotics were added to V. fischeri media, where necessary, at the indicated concentrations: chloramphenicol (Cm) 2.5 mg mL 21 , erythromycin (Em) at 5 mg mL 21 , and tetracycline (Tc) at 5 mg mL 21 in LBS and 30 mg mL 21 in SWT. The following antibiotics were added to E. coli media, where necessary, at the indicated concentrations: Cm at 25 mg mL 21 , kanamycin (Kan) at 50 mg mL 21 , Tc at 15 mg mL 21 , or ampicillin (Ap) at 100 mg mL 21 . For solid media, agar was added to a final concentration of 1.5%.

Wrinkled Colony Assays
To observe wrinkled colony formation, the indicated V. fischeri strains were streaked onto LBS agar plates containing Tc and Cm overnight. Single colonies were subsequently cultured with shaking in LBS broth with Tc at 28uC overnight and then sub-cultured to an optical density at 600 nm (OD 600 ) of 0.2 in 5 mL of fresh medium. Cells were spun down, washed twice in 70% artificial seawater (ASW), and re-suspended in 70% ASW and diluted to an OD of 0.2. 10 mL of re-suspended cultures were spotted onto LBS agar plates and grown overnight at 28uC [26]. Images of spotted cultures were acquired at the indicated time points using the Zeiss Stemi 2000-C dissecting microscope as previously described [26].

Static Pellicle Assays
Strains were grown with shaking in LBS containing Tc and Cm at 28uC overnight and then subcultured to an OD of 0.1 in 1.5 mL of fresh medium in 24-well microtiter dishes. Cultures were then incubated at 28uC for 48 h. The strength of each pellicle was evaluated by disrupting the air-liquid interface with a sterile pipette tip after 48 h of incubation. Cultures with little/no pellicle were scored as (-); cultures with a strong pellicle that remained intact following disruption were scored as (+). Images of pellicles were acquired at the indicated time points using a Zeiss Stemi 2000-C dissecting microscope.

SypE FLAG Protein Expression
FLAG epitope fusions were generated to the C-terminus of the sypE DNTD and CTD mutants using the reverse primer 921 and the forward primers 868 and 910, respectively ( Table 3). The FLAG-sypE alleles were cloned into the expression vector pVSV105 and subsequently conjugated into DsypE cells containing the sypG expressing plasmid (pCLD56). Strains were cultured in LBS containing Tc and Cm for 18 h at 28uC. 1 mL of overnight culture was pelleted by centrifugation, lysed in 500 mL 2X SDS  Table 3. Primers used in this study.

Analysis of SypA Phosphorylation Using Phos-tag TM Acrylamide
Hemagglutinin (HA) and FLAG-epitope fusions to the Cterminus of SypA were generated using the primers listed in Table 3. The sypA-HA alleles were cloned into the Tn7 delivery plasmid pEVS107 [22] ( Table 2). The resulting plasmids were used to introduce the sypA-HA alleles at the Tn7 site of the DsypA and DsypA DsypE strains. Subsequently, the sypG overexpression plasmid pCLD56 was introduced. The resulting V. fischeri strains were streaked onto LBS agar plates containing Tc and single colonies were then cultured overnight in LBS containing Tc at 28uC with shaking. Aliquots of cells (1 mL) were spun down, washed twice with 1X PBS, and standardized to the same amounts using OD 600 measurements. Samples were lysed in 2X SDS sample buffer and resolved on SDS-PAGE gels containing 30 mM Phos-tag TM acrylamide (WAKO chemicals, Richmond, VA) and 50 mM MnCl 2 . We observed some variation in the extent to which phospho-SypA migration was retarded relative to SypA, perhaps due to aging/integrity of the Phos-tag TM acrylamide. Gels were fixed for 15 min in standard transfer buffer containing 1 mM EDTA to remove Mn 2+ from the gel. Gels were incubated for an additional 20 min in transfer buffer without EDTA. Proteins were transferred to a PVDF membrane and the proteins were detected by western blot analysis using an anti-HA antibody (Sigma-Aldrich, St. Louis, MO).
Similar approaches were used to assess SypA phosphorylation under overexpressing conditions with the exceptions that (1) two antibiotics were added (Cm and Tc) throughout to maintain the two (sypG-and sypA-overexpressing) plasmids and (2) the sypA alleles were FLAG-tagged and thus detected using an anti-FLAG antibody (Sigma-Aldrich, St. Louis, MO).

b-galactosidase Assays
The indicated strains were grown (in triplicate) with shaking in LBS containing Tc and Cm at 28uC overnight and then subcultured into fresh medium and grown for up to 24 h. Aliquots (1 mL) of cells were removed at 8 h and 24 h post-inoculation, concentrated, resuspended in Z-buffer, and lysed. The bgalactosidase activity [27] and total protein concentration [28] of each sample were assayed. b-galactosidase units are reported as units of activity per mg of protein.

Squid Colonization Assays
Experiments involving E. scolopes animals were carried out using approaches described in an Animal Component of Research Protocol (ACORP) approved by Loyola University's Institutional Animal Care and Use Committee (IACUC) (LU #107314, 201297). To perform competitive colonization assays, juvenile squid were placed into artificial seawater (Instant Ocean; Aquarium Systems, Mentor, OH) containing approximately 1,000 V. fischeri cells per mL of seawater. Juvenile squid were inoculated with an approximate 1:1 ratio of mutant and wild-type cells, and colonization was allowed to proceed for 18 h. For these assays, DsypE cells were marked with an erythromycin resistance (Em R ) cassette within the chromosome at the Tn7 site. Reciprocal experiments were also performed in which wild-type cells contained the Em R marker. The ratio of bacterial strains within the light organs of the animals was assessed through luminescence and homogenization/plating assays as described previously [11]. The competitive colonization data are reported as the Logtransformed Relative Competitive Index (Log RCI).

SypE's N-terminal Domain is Necessary and Sufficient to Inhibit SypG Biofilms
To better understand the mechanism by which SypE inhibits SypG-induced biofilm formation, we utilized a complementation approach. Briefly, we generated a set of plasmids with sypE alleles that contained or lacked the individual SypE domains (Fig. 1A). We then co-expressed sypG with the various sypE alleles in the DsypE mutant and assessed the resulting biofilm phenotypes. As previously reported, wild-type cells failed to produce SypGinduced biofilms, indicated by the smooth colony morphology on solid agar media and little to no pellicle formation in static liquid culture ( Fig. 2A and 3A) [12]. In contrast, DsypE cells exhibited dramatic SypG-induced biofilm phenotypes, including wrinkled colony formation and robust pellicle formation (Figs. 2B and 3B, respectively) [12]. Co-expression of wild-type sypE fully complemented the sypE mutant and restored inhibition of biofilms, resulting in smooth colony morphology and lack of robust pellicle formation similar to wild-type cells ( Fig. 2C and 3C, respectively).
We next examined the regulatory activity of a SypE variant (SypE DNTD ; Fig. 1A) that lacks the N-terminal 135 amino acids, and found that it failed to complement the sypE mutant, but instead permitted wrinkled colony ( These results indicate that inhibition of SypG-induced biofilms requires the N-terminal domain of SypE. We previously reported that both the SypE DNTD and sypE CTD alleles were capable of promoting biofilms induced by rscS overexpression, indicating the resulting proteins were produced and functional [19]. Western blot analyses also confirmed that the SypE DNTD and sypE CTD proteins were indeed stably produced (Fig. 2I). These results indicate that inhibition of SypG-induced biofilms requires the N-terminal domain of SypE.
To further investigate the potential inhibitory role of the Nterminal domain, we expressed SypE NTD , which contains only the 140 amino acids of the N-terminal domain of SypE (Fig. 1A). We found that SypE NTD fully complemented the sypE mutant, restoring inhibition of both wrinkled colony and pellicle formation ( Fig. 2F and 3F, respectively). Together, these results demonstrate that the N-terminal domain of SypE is both necessary and sufficient to inhibit SypG-induced biofilm formation.

Inhibitory Activity of the N-terminal Domain Requires Conserved Asparagine Residue N52A
The N-terminal domain of SypE exhibits sequence similarity to HPK (histidine protein kinase)-like serine kinases [18]. Importantly, SypE contains a conserved asparagine residue (N52) [18], which in HPK-like serine kinases participates in Mg 2+ ion coordination and the binding of ATP to the nucleotide pocket [29]. In several characterized HPK-like serine kinases, mutagenesis of this asparagine residue results in the loss of serine kinase activity [30,31]. Similarly, we found that this conserved asparagine was required for a mutant SypE variant to inhibit biofilms induced by RscS [19]. To ask whether the ability of wild-type SypE to inhibit SypG-induced biofilms also required this conserved asparagine, we assessed the inhibitory activity of a SypE mutant carrying an alanine substitution at this site (SypE N52A ). We found that expression of SypE N52A failed to complement the DsypE mutant: the colonies exhibited strong SypG-induced wrinkling and pellicle formation (Figs. 2G and 3G). Similarly, this mutation in the context of the inhibitory N-terminal domain alone (SypE NTD, N52A ) resulted in the complete loss of inhibitory activity: cells coexpressing sypE NTD, N52A and sypG exhibited both wrinkled colony morphology and pellicle formation (Figs. 2H and 3H, respectively). The failure to complement cannot be attributed to protein instability, as epitope-tagged versions of both the SypE N52A and SypE NTD, N52A proteins are stably expressed [19]. From these results, we conclude that the inhibitory activity of SypE's Nterminal, kinase domain requires conserved residue N52, consistent with that seen for other characterized HPK-like serine kinases.

Co-overexpression of SypA and SypG Permits Biofilm Formation
Our results indicate that SypE inhibits SypG-induced biofilms through the activity of its N-terminal, serine kinase domain. These findings suggest that SypE primarily functions as a kinase under SypG-inducing conditions, and likely inhibits biofilms by phosphorylating a downstream target protein. We hypothesized that SypE exerts negative control over SypG-induced biofilms by phosphorylating the syp-encoded regulator SypA, which we recently identified as a downstream target of SypE (Fig. 1B) [20]. SypA, a putative anti-sigma factor antagonist, is required for biofilm formation, and phosphorylation of SypA inhibits its activity [20]. We reasoned that if SypE prevents SypG-induced biofilms through phosphorylation of SypA, then co-overexpression of sypG and sypA may result in sufficiently high levels of SypA such that some of it could escape phosphorylation and thus inhibition by SypE. To test this, we co-overexpressed sypA and sypG from compatible plasmids in a wild-type (sypE + ) background and assessed biofilm formation. As controls, we first evaluated strains overexpressing either sypG or sypA alone, and found that biofilm formation was not induced (Fig. 4A and 4B, respectively). In contrast, cells co-overexpressing both sypG and sypA exhibited robust biofilm formation (Fig. 4C). These data indicate that providing excess SypA overcomes the inhibition of SypG-induced biofilms by SypE, and further suggest that SypA functions downstream of SypE and is likely the target for SypE's inhibitory, kinase activity. In addition, the fact that biofilms were only induced upon co-overexpression of sypG and sypA, but not when sypA was overexpressed alone, indicates that sypG is still required to induce syp transcription and the production of the Syp structural proteins (Fig. 1B).

Phosphorylation of SypA is Required for Inhibition of SypG Biofilms by SypE
Our data predicted that the inhibition of biofilm formation by SypE depends upon its ability to phosphorylate SypA. If so, then it should be possible to overcome SypE's inhibition of biofilm formation with a mutant of SypA that does not become phosphorylated. We thus assayed SypG-induced biofilm formation by cells that expressed the sypA S56A allele in single copy in the chromosome; the resulting mutant protein contains a substitution at a conserved serine required for phosphorylation, and fails to become phosphorylated either in vivo when rscS is overexpressed or in vitro [20]. We found that sypE + cells expressing SypA S56A , but not those that expressed wild-type SypA or contained the empty cassette, formed robust biofilms upon overexpression of sypG (Fig. 5). These data demonstrate that the non-phosphorylatable SypA S56A mutant is insensitive to the inhibitory activity of SypE, thus supporting the hypothesis that SypE inhibits the formation of SypG-induced biofilms via phosphorylation of SypA.

SypE Promotes Phosphorylation of SypA in sypGoverexpressing Cells
Our genetic data to date suggested that, in cells that overexpress sypG, SypE functions as a kinase to phosphorylate SypA. To assess this prediction directly, we evaluated the in vivo phosphorylation state of SypA under both biofilm-inhibitory conditions (i.e. sypGoverexpressing wild-type cells) and biofilm-permissive conditions (i.e. sypG-overexpressing DsypE cells). Briefly, we introduced an epitope-tagged wild-type sypA allele (containing a C-terminal HA tag) in single copy in the chromosome (at the Tn7 site) of either DsypA cells or DsypA DsypE cells. We then overexpressed sypG and assessed the in vivo phosphorylation state of SypA by resolving cell lysates on SDS-PAGE gels containing Phos-tag TM acrylamide. This reagent permits the separation of phosphorylated and nonphosphorylated forms of proteins by preferentially binding to and retarding the migration of phosphorylated proteins [32,33]. SypA-HA proteins were detected by western blot analysis using an anti-HA antibody. As observed in figure 6A, sypG-overexpressing wildtype cells (lane 2) exhibited a single, upper band corresponding to phosphorylated SypA (SypA,P). These data indicate that in wildtype cells overexpressing sypG, the majority of, if not all, SypA protein is in the phosphorylated state. In contrast, sypG-overexpressing DsypE cells (which are competent to produce biofilms) exhibited a single, lower band representing unphosphorylated SypA (Fig. 6A, lane 3). Complementation with a wild-type allele of sypE (SypE + ) restored SypA phosphorylation as indicated by the presence of the shifted SypA band (SypA,P) (Fig. 6A, lane 4). Finally, we confirmed that the SypA S56A mutant fails to become phosphorylated in sypG-overexpressing cells (Fig. 6A, lane 5). These results demonstrate that SypE promotes SypA phosphor-ylation under SypG-inducing conditions, and corroborate our biofilm assays indicating that, under SypG-inducing conditions, SypE functions as a kinase. Importantly, they verify that phosphorylation of SypA is critical for biofilm inhibition.
To further test the hypothesis that biofilm formation requires unphosphorylated SypA, we examined the phosphorylation state of SypA in cells co-overexpressing sypG and sypA. We previously found that the co-overexpression of sypG and sypA promotes biofilm formation in wild-type (sypE + ) cells (Fig. 4). We hypothesized that the overexpression of sypA results in an excess of SypA protein, such that a pool of SypA escapes phosphorylation by SypE and is thus active to promote SypG-induced biofilms. We found that wild-type cells expressing sypA and sypG consistently exhibited two bands: a predominant, lower band corresponding to unphosphorylated SypA and a faint, upper band corresponding to phosphorylated SypA (Fig. 6B, lane 2). These data thus demonstrate that overexpression of sypA indeed provides an excess pool of unphosphorylated SypA protein. In agreement with our previous Phos-tag TM experiments, SypA phosphorylation de-  pended upon the presence of SypE, as cells deleted for sypE exhibited only the lower SypA band corresponding to nonphosphorylated SypA (Fig. 6B, lane 3). Finally, complementation with sypE caused a subset of SypA to become phosphorylated under these conditions (Fig. 6B, lane 4). Together, these results support our earlier Phos-tag TM studies indicating that biofilm formation requires unphosphorylated, active SypA.

SypE Regulates SypG-induced Biofilm Formation Downstream of syp Activation
Although our data indicate that SypE-mediated phosphorylation of SypA inhibits biofilm formation upon sypG overexpression, it does not reveal the level at which SypE (via SypA) exerts its control. SypG regulates biofilm formation by inducing transcription of the syp locus, which is required for both biofilm formation in laboratory culture and the formation of bacterial aggregates on the surface of the squid light organ during colonization [5,12]. Therefore, one formal possibility is that SypE impacts transcription of the syp locus through SypA. To determine if SypE affects syp transcription, we utilized a sypA promoter-lacZ fusion inserted in single copy in the chromosome of the wild-type and DsypE strains. We then assessed sypA promoter activity by measuring bgalactosidase activity upon co-expression of sypG and sypE.
As previously observed [12], wild-type cells overexpressing sypG exhibited high levels of syp transcription, as indicated by the significant increase in b-galactosidase activity relative to vector control cells (Fig. 7). Compared to the wild type, cells deleted for sypE exhibited a moderate (,1.5-fold), but consistent, increase in SypG-induced syp transcription (Fig. 7). Co-expression of wild-type sypE and sypG in the DsypE mutant restored syp transcription to near wild-type levels (Fig. 7). These results suggest that SypE does have a minor effect on SypG-induced syp transcription.
To further determine whether SypE indeed controls biofilm formation at the level of syp activation, we asked whether inhibition of SypG-induced transcription required the SypE Nterminal serine kinase domain. We first tested whether expression of SypE DNTD , a sypE mutant that lacks the N-terminal domain and thus its ability to inhibit biofilm formation, impacted activation of syp transcription by SypG. Surprisingly, we found that coexpression of SypE DNTD and sypG in DsypE cells resulted in a decrease in syp transcription similar to cells expressing wild-type sypE, a result opposite to the observed biofilm phenotype (Fig. 7). These results suggested that while SypE may have a slight impact on syp transcription, it is not sufficient to account for the observed effects of SypE on SypG-induced biofilm formation. In support of this, we found that expression of the N-terminal domain alone (SypE NTD ), which is sufficient to inhibit SypG-induced biofilm formation, had no observable impact on syp transcription ( Fig. 7; compare with DsypE vector control). From these data, we conclude that the minor effect of SypE on SypG-induced syp transcription is insufficient to account for the dramatic effects on biofilm formation, and thus SypE must exert its regulatory effect downstream of syp transcriptional activation.

Deletion of sypE Confers a Competitive Colonization Advantage
The ability of V. fischeri to produce syp-dependent biofilms in laboratory culture directly correlates with the ability of the bacteria to efficiently colonize host juvenile squid [11,14,19]. However, we previously observed that the deletion of sypE in otherwise wild-type cells did not significantly impact host colonization [19,34]. We therefore questioned whether the presence of sypE would impact colonization when sypG is overexpressed. In other words, would the ability of the sypGoverexpressing sypE mutant cells to produce robust biofilms also promote colonization competence, or would the absence of sypE have no impact? To assess this question, we examined the ability of sypG-overexpressing wild-type and DsypE strains to competitively colonize juvenile squid. Newly hatched juvenile squid were inoculated with a 1:1 mixture of the two strains, one of which carried an erythromycin resistance (Em R ) cassette on the chromosome. Upon colonization, the squid were homogenized and the relative amounts of the individual strains were determined by calculating the percentage of Em R and Em S colonies. As shown in Figure 8A, DsypE cells dramatically outcompeted wild-type cells for host colonization. These data suggest that upon overexpression of sypG, deletion of sypE confers a competitive colonization advantage. To determine whether the loss of sypE was responsible for the colonization advantage observed in the DsypE strain, we complemented the DsypE strain with a wild-type allele of sypE in trans in the chromosome. Indeed, the SypE-complemented (SypE + ) cells failed to outcompete wild-type cells, when both were overexpressing sypG (Fig. 8B). These results reveal a critical role for SypE in inhibiting both biofilm formation and host colonization. Furthermore, they provide additional correlations between syp-dependent biofilm formation in culture and the ability to promote host colonization in vivo.

Discussion
In this study, we investigated the role of the SypE in the regulation of SypG-dependent biofilm formation and host colonization in V. fischeri. Our interest in this area was prompted by an earlier report demonstrating that SypE permitted RscSinduced biofilms, yet inhibited biofilms inhibited by sypG overexpression [12]. Thus, SypE appeared to function in distinct ways under the two conditions. SypE is an unusual response regulator with both serine kinase and serine phosphatase activities [18,19,20]. To better understand the role of SypE as a negative regulator of syp-dependent biofilm formation, and to determine the mechanism behind the distinct phenotypes displayed by the rscS and sypG overexpression strains, we assessed SypE function under sypG-overexpressing conditions. First, we performed a structure-function analysis to probe the regulatory activities of this novel response regulator. Through complementation studies, we observed that the N-terminal serine kinase domain is both required and sufficient to inhibit SypGinduced biofilm formation. The inhibitory activity of the isolated N-terminal domain (SypE NTD ) absolutely required residue N52, a conserved asparagine that is necessary for ATP binding and kinase activity in other characterized serine kinases [29,30]. These results parallel those observed for the role of SypE in RscSinduced biofilm formation: expression of SypE NTD alone also inhibited biofilm formation in a manner that depended on N52 [19]. Future efforts are required to determine what other residues contribute to SypE-mediated inhibition of biofilms. Unlike our results with the N-terminal domain, however, our results with respect to the C-terminal domain differed under rscS and sypGoverexpressing conditions: whereas the C-terminal domain promoted RscS-induced biofilms [19], we failed to observe any impact of this domain in promoting SypG-induced biofilm formation ( Fig. 2 and data not shown). It remains possible that the C-terminal serine phosphatase domain does promote SypGinduced biofilms, but that this impact is not readily apparent under the current experimental conditions. Indeed, SypG-mediated induction of syp transcription is stronger than that induced by RscS, and correspondingly, biofilm formation occurs at an accelerated pace by sypG-overexpressing DsypE strains relative to rscS-overexpressing strains (Morris and Visick, unpublished data). Thus, it is more difficult to assess a subtle increase in biofilm formation under the SypG conditions.
Next, we used both genetic and biochemical approaches to assess whether SypE inhibited SypG-induced biofilms by promoting the phosphorylation, and thus inactivation, of SypA, a critical regulator required for biofilm development. Co-overexpression of sypG and sypA in sypE + cells promoted biofilm formation, indicating that high levels of SypA can overcome the inhibitory activity of SypE, presumably because some SypA escapes phosphorylation (Fig. 1B and 6B). In further support of this conclusion, expression from the chromosome of a sypA mutant that does not become phosphorylated (SypA S56A ) also permitted the formation of SypG-induced biofilms in wild-type (SypE + ) cells. Finally, we found that SypA was predominantly phosphorylated under biofilm-inhibiting conditions (i.e., upon overexpression of sypG in wild-type cells), but unphosphorylated under biofilmpermissive conditions (i.e. upon overexpression of sypG in DsypE cells). The ability to evaluate the phosphorylation state of SypA in vivo provides a critical link between the biochemical proof of activity and the genetic and phenotypic analyses. Importantly, these findings demonstrate that SypE inhibits SypG-induced biofilm formation by functioning as a kinase to phosphorylate SypA (Fig. 1B). They also corroborate our previous study demonstrating that SypA is largely unphosphorylated under biofilm-permissive conditions, specifically in rscS-overexpressing wild-type cells [20].
Studies are underway to determine the mechanism by which the SypE-SypA pathway controls biofilm formation. Here, we asked whether SypE (via SypA) exerts any impact on biofilm formation through affecting syp transcription. We found that the deletion and/or overexpression of full-length sypE or a subset of mutant sypE alleles resulted in a slight, yet reproducible, impact on syp transcription. However, the observed effects on transcription were not sufficient to account for the observed biofilm phenotypes. For example, we found that a SypE mutant lacking the N-terminal domain (SypE DNTD ) and, thus, lacking inhibitory activity retained the ability to decrease SypG-induced transcription. We hypothesize that SypE's REC domain may be responsible for the observed impact on syp transcription. SypE, like SypG, appears to function below RscS in the regulatory cascade [12,19]. If these two response regulators do indeed receive signals from the same upstream sensor kinase, then the REC domains of SypE and SypG may compete for phosphorylation, resulting in a decrease in SypG activation and syp transcription. In support of this possibility, expression of SypE NTD , a SypE derivative that lacks the REC domain but retains biofilm-inhibitory activity, had no observable impact on syp transcription. Due to the lack of correlation between biofilm formation and the minor impacts of a subset of sypE alleles on syp transcription, we conclude that SypE functions at a level below syp transcription. In support of this conclusion, our preliminary studies indicate that expression of sypA, although critical for biofilm formation, exerts no impact syp transcription (Morris and Visick, unpublished data).
Finally, we found that the ability of V. fischeri to form biofilms in culture directly correlates with colonization efficiency in vivo, a result that is consistent with previous studies [11,14,19]. In particular, we observed that the enhanced ability of the DsypE strain to form biofilms upon overexpression of sypG provides these cells with a competitive colonization advantage, presumably due to enhanced aggregation outside of the symbiotic light organ. We found that this colonization advantage, similar to biofilm formation, depended upon loss of sypE as complementation with sypE in trans in the chromosome abolished the colonization advantage. Since previous studies had found no substantial colonization phenotype upon the deletion of sypE [19,34], why was SypE so critical here? Our results to date indicate that wildtype cells inactivate SypE during the early stages of squid colonization, and thus its loss would have relatively little impact on colonization initiation [19]. We speculate that the levels of SypE produced under sypG-overexpression conditions are greater than can be inactivated through the mechanisms in place. In support of this idea, we have found that SypE expressed from a multi-copy plasmid inhibits biofilm formation, even under rscSoverexpression conditions (Morris and Visick, unpublished data). In any event, the significant inhibitory effect of SypE on colonization under SypG-inducing conditions underscores the critical importance of biofilm formation in promoting symbiotic host colonization.
This study also emphasizes the strict control V. fischeri exerts over biofilm formation (Fig. 1B). The finding that SypE inhibits biofilms under SypG-inducing conditions, yet permits biofilms upon overexpression of rscS, suggests that SypE could function to prevent aberrant biofilm formation induced by the phosphorylation, and thus activation, of SypG in the absence of RscS signaling/activation. It is currently unclear under what conditions SypG may become phosphorylated, and induce syp transcription, without signaling through RscS. It remains possible that SypG may receive phosphoryl groups and become activated through other regulatory inputs in addition to RscS. Recently,  demonstrated that regulatory components within the Lux phosphorelay controlling bioluminescence also impact the regulation of syp biofilm formation. Importantly, this impact on syp biofilms appears to occur at the level of SypG and syp locus activation [35]. Thus, it is possible that, under conditions in which the Lux phosphorelay promotes SypG activation and syp induction, SypE functions as a negative regulator to inactivate SypA and prevent biofilm formation. Alternatively, SypE may be important for preventing biofilm formation under conditions in which SypG might get phosphorylated due to inadvertent ''crosstalk'' from other sensor kinases that are activated. These possibilities remain to be investigated.
In summary, this work provides further insight into the mechanism by which the regulator SypE restricts biofilm formation and host colonization by V. fischeri. Furthermore, it emphasizes the critical role of SypE in regulating biofilm formation, particularly when RscS is not activating the pathway. Together, our data suggest that SypE may function to restrict biofilms to conditions in which RscS becomes activated (i.e. upon interaction with host squid), and thus to coordinate syp biofilm production with host colonization.