Role for cheR of Vibrio fischeri in the Vibrio-Squid Symbiosis

Abstract

Upon hatching, the Hawaiian squid Euprymna scolopes is rapidly colonized by its symbiotic partner, the bioluminescent marine bacterium Vibrio fischeri. V. fischeri cells present in the seawater enter the light organ of juvenile squid in a process that requires bacterial motility. In this study, we investigated the role chemotaxis may play in establishing this symbiotic colonization. Previously we reported that V. fischeri migrates toward numerous attractants, including N-acetylneuraminic acid (NANA), a component of squid mucus. However, whether or not migration toward an attractant such as squid-derived NANA helps the bacterium to localize toward the light organ is unknown. When tested for the ability to colonize juvenile squid, a V. fischeri chemotaxis mutant defective for the methyltransferase CheR was outcompeted by the wild-type strain in co-inoculation experiments, even when the mutant was present in 4-fold excess. Our results suggest that the ability to perform chemotaxis is an advantage during colonization, but not essential.

Introduction

Over the past few decades, the study of long-term associations, or symbioses, between animals and microorganisms has become an important area of research in both the fields of microbial ecology and pathogenic microbiology (e.g., (Dethlefsen et al., 2007; Nyholm and Nishiguchi, 2008; Ruby et al., 2004)). The Vibrio-squid symbiosis has emerged as a prominent model for studying the establishment of a symbiotic relationship between animal host and microbial symbiont. The bobtail squid host, Euprymna scolopes, is found in the shallow waters off of the island of Oahu, and rapidly becomes colonized by the bioluminescent bacterium, Vibrio fischeri, shortly after hatching (for reviews, see (McFall-Ngai, 1999; Nyholm and McFall-Ngai, 2004; Stabb, 2006; Visick and McFall-Ngai, 2000). The juvenile squid’s symbiotic light organ has a physiology that permits entrance into, and permanent colonization by V. fischeri, while excluding all other species of bacteria found in the seawater.

Newly-hatched, non-colonized squid encounter V. fischeri cells in seawater being vented through the mantle cavity of the animal. Cilia present on specialized tissue appendages of the light organ help direct the bacteria towards the organ’s surface (Nyholm et al., 2000; Nyholm et al., 2002). The bacteria are thus swept into the vicinity of the light organ where they accumulate in mucus secreted on the surface of the light organ in response to bacterial exposure (Nyholm et al., 2000; Nyholm et al., 2002). To colonize their host, bacteria attached to the surface of the light organ must enter the pores, traverse through ducts and antechambers, both of which are non-permissive for colonization, before reaching the sites of colonization, or crypts, where they grow to a high density (Sycuro et al., 2006). It is in these crypts that V. fischeri begins to produce bioluminescence, which is used by the squid to avoid detection by predators (Jones and Nishiguchi, 2004).

Motility is critical for colonization, as non-motile mutants fail to colonize the cypts (Graf et al., 1994; Millikan and Ruby, 2003; Wolfe et al., 2004). Given the complex path to colonization, it seems likely that the ability to sense and move toward a chemoattractant, known as chemotaxis, may also play a role. For example, chemotaxis may be important in the ability of V. fischeri to migrate to the pores of the light organ and/or to traverse through the ducts and antechambers and ultimately reach the internal crypts. Other marine vibrio species use chemotactic motility to colonize their hosts: V. anguillarum and V. alginolyticus migrate toward mucus of their respective fish hosts (Bordas et al., 1998; Larsen et al., 2001; O’Toole et al., 1996; O’Toole et al., 1999). We have previously identified a number of attractants for V. fischeri (DeLoney-Marino et al., 2003), one of which (N-acetylneuraminic acid, NANA) is a component of the squid-derived mucus secreted on the symbiotic light organ (Nyholm et al., 2000). In addition, analysis of gene expression profiles of both free-living and symbiotic V. fischeri identified a number of transcripts that were solely expressed by planktonic cells, including two genes for methyl-accepting chemotaxis proteins (Guerrero-Ferreira and Nishiguchi, 2010), suggesting that a chemotactic attraction may help to guide the bacterium into the squid light organ, but may no longer be necessary once the microbe has established symbiotic colonization.

Chemotaxis, best-studied in Escherichia coli and Salmonella typhimurium, enables a bacterium to migrate toward an attractant by extending the time the bacterium rotates its flagella counter-clockwise (CCW), resulting in a “run” (for reviews, see (Eisenbach, 2007; Wadhams and Armitage, 2004)). These runs are interrupted continuously by short intervals of clockwise rotation (which decrease in frequency in the presence of attractant), resulting in tumbling and nearly random changes in direction. In E. coli and S. typhimurium, the signal transduction pathway that permits cells to make these comparisons consists of a group of cytoplasmic proteins which enable transmembrane chemoreceptors, known as methyl-accepting chemotaxis proteins (MCPs), to communicate with switch components of flagellar motors (reviewed in (Clausznitzer et al., 2010)). One of these signaling proteins is CheY, which interacts directly with the flagellum to cause a reversal in rotation, resulting in a tumble and permitting cells to re-orient. In V. fischeri, cheY is important for colonization: a cheY mutant was outcompeted by wild-type cells for colonization of juvenile squid (Hussa et al., 2007). These data suggested that chemotaxis may be important for symbiotic colonization, although little else is known about the role chemotaxis may play in the establishment of the V. fischeri-squid symbiosis.

Once cells such as E. coli and Salmonella recognize and respond to a stimulus, the bacteria must reset their response to ligand binding (monitored as receptor methylation) so that they may either increase or decrease the frequency of tumbling, depending on where they find themselves in the gradient of attractant (Macnab and Koshland, 1972). The methyltransferase CheR and the methylesterase CheB attenuate the signaling states by adding and removing methyl residues from MCPs, allowing the cells to continuously adapt to changes in a gradient of attractant (reviewed in (Lux and Shi, 2004)). E. coli cheR mutants cannot methylate the MCPs, and as a consequence cannot adapt: they cannot reorient their direction to detect gradients of attractants (Manson, 1992). Such chemotaxis mutants do not swim toward attractants and therefore can be used to study the role of chemotaxis in certain processes such as colonization.

The colonization defect of the cheY mutant suggests that the ability to control the chemotactic excitation response is important to symbiotic colonization by V. fischeri; however, whether or not the ability to adapt to changes in gradient of an attractant (such as encountering squid secretions) via a CheR-based attenuation of an excitation signal is also needed to establish colonization is unknown (for a review of the chemotactic signal transduction pathway, see (Szurmant and Ordal, 2004). To further explore the role of chemotaxis in the symbiotic colonization of V. fischeri with E. scolopes, we identified a gene that encodes a homolog to the CheR protein of E. coli. We subsequently constructed a cheR mutant of V. fischeri and examined the role of CheR in chemotaxis in laboratory media. Finally, we evaluated the ability of the cheR mutant to form a symbiotic colonization with juvenile E. scolopes. We conclude from these studies that V. fischeri uses a CheR protein to promote chemotaxis, likely by functioning in a manner similar to the homologous E. coli protein, and that disabling this activity significantly decreases the bacterium’s ability to compete against wild-type cells for colonization of juvenile squid.

Materials and Methods

Strains and media

Strains used in this study are listed in Table 1. V. fischeri strains were grown in seawater-tryptone (SWT) (Yip et al., 2005) or LB-Salt (LBS) (Dunlap, 1989; Graf et al., 1994) media for routine culturing. For experiments in which motility was important, we used SWT, while LBS was the medium of choice for genetic manipulation of V. fischeri. In addition, for motility experiments, we used tryptone broth seawater (TB-SW) (DeLoney-Marino et al., 2003). E. coli strains were grown in Luria Bertani (LB) medium (Davis et al., 1980). Agar was added to a final concentration of 0.25% for soft agar and 1.5% for solid agar. Where appropriate, antibiotics were added to the following final concentrations: Chloramphenicol (Cm), 1-5 μg ml⁻¹ for V. fischeri and 30 μg ml⁻¹ for E. coli; Erythromycin (Erm), 5 μg ml⁻¹ for V. fischeri and 150 μg ml⁻¹ for E. coli; Kanamycin (Kn), 100 μg ml⁻¹ for V. fischeri and 50 μg ml⁻¹ for E. coli.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain	Relevant genotype or characteristics	Reference or source
E. coli
DH5α	endA1 hsdR17 (rk−mk+) glnV44 thi-1 recA1 gyrA (NalR) relA1 Δ(lacIZYA-argF)U169 deoR (ϕ80dlacΔ[lacZ]M15)	(Woodcock et al., 1989)
CC118λpir	Δ(ara-leu) araD ΔlacX74 galE galK phoA20	(Herrero et al., 1990)
S17-1λpir	thi, pro, hsdR− hsdM + recA λ pir	(Simon et al., 1983)
V. fischeri
ES114	Wt	(Boettcher and Ruby, 1990)
ESR1	RfR	(Graf et al., 1994)
KV661	RfR flg∷Tn10luxAB ΔluxA∷erm	This study
KV961	RfR, CmR (pAM1 integrated into chromosome)	This study
KV1007	cheR ∷ erm	This study
KV1172	cheR-frameshift	This study
KV1176	cheR-frameshift attTn7∷cheV-cheR+	This study
KV1421	attTn7∷erm	(O’Shea et al., 2006)
KV1424	cheR-frameshift attTn7∷cm	This study
KV2510	VF1833 (cheY)∷pKV215 (ermR)	(Hussa et al., 2007)

Plasmid	Characteristics/Derivation	Reference or source
pAM1	pEVS79 (Stabb and Ruby, 2002) (SalI) + 2.3-kb SalI fragment from pKV119 (cheR flgB-D)	This study
pAM4	pEVS79 + flgNMA, cheVR flgB-D	This study
pAM7	pAM4 XbaI partial digest, filled-in (results in cheR-frameshift)	This study
pAM8	pAM4 XbaI partial digest, filled-in (results in loss of XbaI in the MCS)	This study
pAM9	pAM8 XbaI filled-in + ClaI/XbaI - filled-in ermR from pLMS37 (results in cheR∷erm)	This study
pCD5	pEVS107 AseI, filled-in + 2.6-kb fragment from pAM4 AseI, filled-in (results in Tn7∷cheV-cheR)	This study
pEVS104	Conjugal helper plasmid (tra, trb), KnR	(Stabb and Ruby, 2002)
pEVS107	Tn7 delivery plasmid, KnR, ErmR	(McCann et al., 2003)
pKV119	pVO8 (Visick and Ruby, 1997)(BamHI ) + 5.5 kb BglII fragment from ES114 chromosome	This study
pLMS37	pBS (EcoRV) + 1.2-kb ermR SmaI/EcoRV fragment from pKV25 (Visick and Ruby, 1998)	This study
pNDM50	Tn7 delivery plasmid, KnR, CmR	This study
pRK2013	Conjugal helper plasmid (tra, trb), KnR	(Figurski and Helinski, 1979)
pUX-BF13	Encodes Tn7 transposase (tnsA, B, C, D, E), ApR	(Bao et al., 1991)

Soft agar and motility assays

Cultures of V. fischeri were grown at 28°C to mid-exponential phase (OD₆₀₀ ≈ 0.3-0.7) in SWT. For chemotaxis studies, aliquots of approximately equal numbers of bacterial cells were inoculated near the center of soft agar motility plates and incubated at 28°C for 5 to 22 h. Soft agar motility plates were handled as described previously (DeLoney-Marino et al., 2003). Images were taken with a digital camera.

Molecular and genetic techniques

All plasmid constructions were carried out using standard molecular biology techniques with restriction and modifying enzymes obtained from New England Biolabs (Beverly, MA) or Promega (Madison, WI). Plasmids used or constructed in this study are shown in Table 1. Conjugations were performed as described previously (DeLoney et al., 2002; Visick and Skoufos, 2001).

Flagella and chemotaxis genes were cloned by complementing a non-motile transposon-insertion mutant (KV661) with a plasmid library of BglII-digested ES114 chromosomal DNA. The library was conjugated into KV661, and the resultant conjugation mixture was inoculated onto a soft agar motility plate. The plates were incubated at 28°C until motile transconjugants swam away from the inoculation site. From the motile bacteria, a plasmid (pKV119) was isolated that contained a 5.5 kb chromosomal fragment. A portion of this plasmid was subcloned into the conjugal vector pEVS79, generating pAM1. To obtain additional flanking DNA, pAM1 was recombined into the ESR1 chromosome; the chromosomal DNA was then digested with enzymes that cut outside the vector, and self-ligated. The resulting plasmid, pAM4, contained an additional 3 kb of upstream DNA.

Strain construction

The cheR frameshift mutant KV1172 was constructed with strains and plasmids as described in Fig. 2. KV1172 was complemented with a wild-type copy of the cheV-cheR region inserted in single copy at the Tn7 insertion site (att Tn7) (Bao et al., 1991) in the V. fischeri chromosome as follows. E. coli cells carrying the Tn7∷cheV-cheR-containing plasmid pCD5 served as a donor in a tetra-parental mating, which also included E. coli cells carrying the transposase-encoding plasmid pUX-BF13, E. coli cells carrying the self-transmissible plasmid pEVS104, and the V. fischeri recipient (KV1172). Similarly, antibiotic-resistant cheR mutant strain KV1424 and wild-type strain KV1421 were constructed by introducing pNDM50 (Tn7∷cm) and pEVS107 (Tn7∷erm) respectively into KV1172 and ES114 by conjugation. The resulting transconjugants were screened for loss of Kn resistance (resulting from loss of the delivery plasmid) and stable antibiotic resistance (Cm or Erm, resulting from the integration of the appropriate Tn7 cassette at the Tn7att site).

Figure 2.

Genetic manipulations involved in the construction of cheR mutants. (A) A Genetic approach to obtaining DNA flanking the cheR gene. The starting plasmid, pAM1 (i), was recombined into the chromosome of wild-type V. fischeri (ii) to generate KV961 (iii). Chromosomal DNA was isolated from KV961, digested with HindIII, ligated, and used to transform E. coli cells to chloramphenicol resistance. A plasmid, pAM4, containing cheR and about 3 kb of upstream flanking DNA, was isolated (iv). (B) Plasmid pAM9 (i), a derivative of pAM4 in which the erythromycin resistance cassette (Erm^R) was inserted into the cheR gene, was introduced into wild-type ES114 cells (ii). A mutant, KV1007 (iii), in which the cheR gene in the chromosome was exchanged with for Erm^R-disrupted gene, was isolated. (C) Plasmid pAM7 (i), a derivative of pAM4 in which cheR gene was disrupted by frame-shift mutation (indicated by the striped bar within the black arrow representing cheR), was introduced into cheR∷ Erm^R mutant KV1007 (ii). A mutant, KV1172 (iii), in which the cheR∷ Erm^R allele in the chromosome was exchanged with for cheR-frame-shift allele, was isolated. The crossed lines indicate possible recombinational cross-over events.

Southern blotting

KV1176 (derived from KV1172) was confirmed by Southern blot analysis as follows. Chromosomal DNA isolated from V. fischeri strains ES114 and KV1172 was digested with either HpaI (which cuts once within cheR), or HpaI/XbaI (which cuts twice within and twice outside cheR). DNA fragments were separated by electrophoresis, transferred onto a nylon membrane (Hybond XL, Amersham-Pharmacia Biotech, Piscataway, NJ), hybridized and detected as described previously (DeLoney et al., 2002). Purified pAM1 plasmid DNA was used to generate digoxygenin-labeled DNA complementary to the cheR gene, which served as the probe.

Colonization assays

To determine whether the cheR mutant was able to form a symbiotic association with E. scolopes, cells were inoculated into artificial seawater (Instant Ocean, Aquarium Systems, Mentor, OH) as a 1:1, 1:2, or 1:4 mixture of wild-type and cheR mutant cells to a final concentration of 2000 to 6000 cells per ml. Juvenile squid were then placed into this inoculum and analyzed for presence of bacteria in the light organ as described previously (Ruby, 1996). The actual ratio of wild-type to mutant cells was determined by plating a dilution of the homogenate onto SWT and selective media containing either Erm or Cm. The limit of detection is 14 V. fischeri cells per squid.

Animal use

Squid were handled according to the protocols approved by the Institutional Animal Care and Use Committee at Loyola University Chicago.

Results

Identification of the cheR gene of V. fischeri

To further understand the role of chemotaxis in symbiotic colonization by V. fischeri, we chose to generate and characterize a cheR mutant. We began with a search for flagella and chemotaxis mutants of V. fischeri from a library of Tn10∷luxAB transposon mutants. We further characterized one of the non-motile mutants, KV661, by complementing the motility defect with a chromosomal DNA library. The complementing construct contained, along with several other genes, a putative cheR gene, VF_1878 (Fig. 1A). We used comparative BLAST analyses to evaluate the putative CheR protein (Marchler-Bauer and Bryant, 2004; Marchler-Bauer et al., 2009; Marchler-Bauer et al., 2011). These analyses revealed that, in contrast to some other bacteria, the V. fischeri genome contained a single copy of the cheR gene. They also revealed that the predicted protein was homologous to numerous characterized and putative CheR proteins. The most closely related, with E-values of zero, were six putative CheR homologs found in related vibrio species. The alignment of the V. fischeri protein with two of these related CheR proteins, from Aliivibrio salmonicida, (95% identity, 99% positives, 0% gaps; (Hjerde et al., 2008) and Vibrio harveyi (87% identity, 93% positives, 0% gaps; Naval Research Laboratory sequencing effort GenBank CP001223-5), is shown in figure 1B. For comparison, we also analyzed a characterized CheR protein from E. coli, which yielded a much lower sequence homology with only 31% identity, 54% positives and 7% sequence gaps (Fig. 1B) (E-value of 5e-38; (Mutoh and Simon, 1986)).

Figure. 1.

A. Chromosomal arrangement of flagellin genes flgB-flgJ and chemotaxis genes cheV and cheR of V. fischeri. B. Multiple sequence alignments of the cheR genes of Vibrio fischeri ES114, Allivibrio salmonicida LFI1238, Vibrio harveyi ATCC BAA-1116, and E. coli showing conserved domains and key residues above the alignments known to be important for protein function (Shiomi et al., 2002). Solid line: CheR methyltransferase, all-alpha domain. Dashed line: AdoMet-MTase catalytic domain. Dotted line: beta-subdomain. Arrow: Residue Gly-23, which corresponds to Gly-39 of E. coli which is strictly conserved in all CheR homologs and corresponds to a crucial turn in the secondary structure of the protein. Three stars: signature sequence involved in binding to AdoMet (Malone et al., 1995). NCBI gene identification numbers: V. fischeri, YP 205261.1; A. salmonicida, YP 002263696; V. harveyi, YP 001444487; E. coli, AAA23568.1.

Importantly, our analysis revealed the presence of essential conserved domains. The protein encoded by VF_1878 contained the catalytic S-adenosylmethionine-dependent methyltransferase (SAM or AdoMet-MTase) domain common to the methyltansferase superfamily, as well as both the N-terminal all-alpha domain and beta-subdomains of the catalytic domain found exclusively in chemotaxis methyltransferases (Shiomi et al., 2002). Within those domains, a number of important residues known to be important for function in the well-studied microbe E. coli were conserved (Fig. 1B). For example, residue Gly-23 corresponds to Gly-39 of CheR of E coli, which determines a crucial structural component in the protein and is known to be strictly conserved in all CheR homologues (arrow in Fig. 1B). In addition, the (G/A)X(G/A/S)XG signature sequence, which is found in most MTase catalytic domains, was conserved (asterisks in Fig. 1B) (Shiomi et al., 2002). We conclude from these analyses that VF_1878 likely encodes a protein with a function similar to the CheR protein of E. coli.

Genetic approaches to generate a cheR mutant

We used these cloned sequences to generate cheR mutants as follows. First, we subcloned cheR and flanking DNA into a plasmid that could be used for recombination in V. fischeri, thus constructing pAM1 (Table 1). However, the cheR gene was located at one end of the cloned chromosomal DNA (in both pAM1 and its precursor), leaving no room for homologous recombination upstream of the cheR gene (Fig. 2A). We therefore devised an approach to overcome this problem. Specifically, we introduced this plasmid into V. fischeri, and sought and obtained a recombinant that integrated the plasmid into the chromosome (Fig. 2A). We then isolated chromosomal DNA from that strain, digested it with HindIII, ligated the DNA fragments and used the ligation to transform E. coli cells, selecting for chloramphenicol resistance. With this approach, we obtained plasmid pAM4, which contained cheR gene and flanking DNA, including an additional ~3 kb of DNA upstream of cheR, sufficient for subsequent gene replacement approaches.

Using pAM4, we generated two additional constructs that contained disruptions of the cheR gene. The cheR gene in pAM9 had an insertion of the erythromycin gene, while the cheR gene in pAM7 contained an unmarked frame-shift mutation. To obtain cheR mutants, we first used pAM9 to introduce the marked mutation into V. fischeri to construct strain KV1007 (Fig. 2B). Then, we used KV1007 as a recipient and pAM7 as donor DNA to replace the erythromycin cassette with the unmarked frame-shift mutation to generate KV1172 (Fig. 2C). We used this latter unmarked mutant, KV1172, in our subsequent studies as we previously determined that cells carrying the erythromycin resistance cassette exhibited slight defects in growth and motility in soft agar assays (O’Shea et al., 2006). Finally, to ensure that any phenotype we observe was due to the loss of cheR and not to a spontaneous mutation elsewhere, we introduced the cheR operon (including the upstream gene cheV) into the chromosome of KV1172 at the Tn7 site, a neutral location that permits single copy complementation, thus generating KV1176.

Migration of chemotaxis mutants in soft agar assays

To evaluate the role of cheR in chemotaxis, we used soft agar assays as described previously (DeLoney-Marino et al., 2003). Briefly, when inoculated near the center of a petri plate containing semisolid (soft) nutrient agar, motile bacteria migrate outward in concentric bands formed by chemotactic responses of the bacteria to spatial gradients of attractants resulting from metabolism and diffusion of the attractants (Wolfe and Berg, 1989). To observe the chemotactic behavior of the V. fischeri cheR mutant in soft agar, we inoculated log-phase cells of the unmarked cheR mutant along with control strains onto the surface of TB-SW soft agar. Within 3 h of incubation at 28°C, wild-type strain ES114 began to migrate beyond the point of inoculation, and within 5 to 6 h, this strain had formed the two distinct chemotaxis rings typical for V. fischeri (Fig. 3A and data not shown) (DeLoney-Marino et al., 2003). In contrast, no migration was observed for the cheR mutant after 5 to 6 hours incubation at 28°C, a phenotype that was indistinguishable from that of the cheY mutant KV2510 and the non-motile mutant KV661 (Fig. 3A). By 22 hours, a fuzzy haze of growth characteristic of diffusion in soft agar had appeared around the cheR mutants (KV1007 and KV1172) but not around the non-motile mutant KV661 (Fig. 3B). Furthermore, despite this migration, the cheR mutant failed to form concentric migrating bands. This phenotype is indicative of a chemotaxis defect (Fig. 3B). Together, these data indicated that the cheR mutants were not non-motile, but exhibited a severe defect in directed migration through soft agar.

Figure 3.

Motility phenotype of the cheR mutant. (A-B) Migration of V. fischeri in soft agar. V. fischeri cells were inoculated onto TB-SW plates and incubated at 28°C for 5h (A) or 22h (B). ES114 = parental squid isolate; KV661 = nonmotile flg⁻; KV1007 = cheR∷erm; KV1172 = cheR-frame-shift; KV1176 = cheR-frame-shift attTn7∷cheV-cheR⁺; KV2510 = cheY∷pKV215.

Our data were consistent with what is known for the well-characterized cheR mutants of E. coli. Specifically, due to an inability to methylate the MCPs, such cheR mutants have extended periods of runs and cannot bias their movement in a specific direction (Springer and Koshland, 1977; Stock et al., 1981). This same smooth swimming phenotype has been seen in cheR mutants of other bacterial species as well (Jones et al., 1992; Stephens et al., 2006), including in V. anguillarum (O’Toole et al., 1996), although in Bacillus subtilis, cheR mutants exhibit a “tumbly” phenotype (Kirsch et al., 1993). To determine if the V. fischeri cheR frameshift mutant exhibited a similar swimming behavior to that of E. coli cheR mutants, we examined the strains microscopically. When visualized for motility microscopically, the cheR mutant swam predominantly in extended runs, while the wild-type strain exhibited runs interspersed with frequent tumbles and the non-motile mutant did not display any apparent motility other than Brownian motion (data not shown). These results suggested that the inability of the mutant to migrate in TB-SW soft agar did not result from an inability to swim, but rather from a decreased ability to change the direction of its flagellar rotation and thus navigate the channels in the agar (Wolfe and Berg, 1989). When complemented with a wild-type copy of the cheVR genes, the cheR frameshift mutant migrated at a rate similar to that of the wild-type strain (Fig 3A). Thus, the cheR mutant KV1172 displayed motility and chemotactic behaviors characteristic of E. coli cheR mutant strains, and these phenotypes could be rescued with a wild-type copy of cheR. We conclude that VF_1878 encodes a CheR homolog.

Colonization by the cheR mutant

Because our data supported the identification of cheR as a chemotaxis gene with the same function that has been demonstrated for other organisms (Springer and Koshland, 1977; Stock et al., 1985), we proceeded to determine whether CheR plays an important role in colonization by V. fischeri of its symbiotic host E. scolopes. To evaluate the role of cheR, we assayed the ability of the frame-shift cheR mutant to form a symbiotic association with juveniles of the squid E. scolopes when presented in competition with a chemotaxis-proficient (wild-type) strain. Newly hatched juvenile E. scolopes were inoculated into seawater containing various ratios of parent strain to mutant. When co-inoculated with a ratio of 1:2 wild-type to mutant, wild-type bacteria represented 100% of the recovered cells from 8 of 9 animals (Fig. 4). Even when inoculated with a wild-type to mutant ratio of 1:4, 8 of 10 animals were predominately colonized by the parent strain (Fig. 4), indicating that the cheR mutant has a colonization disadvantage compared to the wild-type strain. These data demonstrate that CheR plays a critical role in symbiotic colonization and suggest that the ability to perform chemotaxis is an advantage during colonization.

Figure 4.

Colonization defect of a cheR mutant. Competition between wild-type and cheR mutant V. fischeri strains for symbiotic colonization. 9 newly-hatched squid were incubated with a mixed inoculum containing an approximate 1:2 ratio of Erm-resistant wild-type strain KV1421 to Cm-resistant cheR mutant strain KV1424 (solid circles). In a separate experiment, 10 newly-hatched squid were incubated with a mixed inoculum containing an approximate 1:4 ratio of KV1421 to KV1424 (open circles). The level of colonization achieved by each strain was determined by homogenization and plating at 20 h after the organisms were placed together. The ratio of mutant to wild-type cells extracted from individual squid is expressed as the relative competitive index (RCI) and is indicated by a circle. An RCI less than 1 indicates animals colonized predominately by the wild-type strain. Circles with an arrow indicate an RCI of <0.01. The data reported are representative of 4 different experiments.

Discussion

A number of studies have investigated the role chemotactic motility plays in colonization of animal hosts, both in pathogenic (e.g., (Bordas et al., 1998; Foynes et al., 2000; Hawes and Smith, 1989; Larsen et al., 2001; O’Toole et al., 1996; O’Toole et al., 1999; Terry et al., 2005; Yao and Allen, 2006) and symbiotic associations (e.g., (Bashan and Holguin, 1994; Miller et al., 2007; Yost et al., 1998)). While the requirement of normal motility for symbiotic colonization of juvenile squid by V. fischeri is well established (Graf et al., 1994; Millikan and Ruby, 2002; Nyholm et al., 2000), only a few studies have investigated the role chemotaxis may play in establishing the Vibrio-squid symbiosis (Guerrero-Ferreira and Nishiguchi, 2010; Hussa et al., 2007). A cheY chemotaxis mutant of V. fischeri was found to exhibit a competitive colonization defect compared to wild type (Hussa et al., 2007), suggesting that the ability to control the excitation response to a chemical gradient is important to symbiotic colonization by V. fischeri. In this work, we characterized the ability of a cheR chemotactic adaptation mutant of the marine bacterium V. fischeri to perform chemotaxis in laboratory culture and investigated the role of this behavior during symbiotic association with the Hawaiian squid E. scolopes.

We identified the cheR gene during our sequence analysis of a flagella gene cluster. The predicted protein exhibited extensive sequence identity with CheR proteins of other vibrio species and contained the three domains commonly found in CheR homologues. We subsequently mutated cheR by frame-shift mutation. The mutant exhibited defects similar to that of E. coli cheR mutants: smooth swimming in liquid culture and an inability to migrate proficiently through soft agar. Finally, we determined that a V. fischeri cheR chemotaxis mutant exhibits a competitive disadvantage compared to the parental wild-type strain to colonize juvenile squid even when present in 4-fold excess over the wild-type strain.

Previously, two possibilities were suggested to account for the defect of the cheY mutant in colonization (Hussa et al., 2007). One was that the mutant failed to sense and respond to a gradient of attractant that would direct the cells to the crypts. The second was that the smooth swimming phenotype of the cheY mutant resulted in the cells getting “stuck” in the mucus, similar to their getting stuck in the soft agar medium. For the cheR mutant, both explanations are possible. We hypothesized that, if the cheR mutant were getting stuck in the mucus, it might colonize poorly even when presented to the squid alone (rather than in competition). To test this idea, we performed several single strain inoculation experiments, and found that the cheR mutant could colonize when presented alone. However, the ability of this mutant to colonize was highly variable: in one experiment, 4 of 11 animals inoculated with the cheR mutant became colonized to high levels within 24 hours while 7 of 7 wild-type-inoculated animals became colonized; in another experiment, we observed 7 of 10 animals became colonized with the cheR mutant while 9 of 10 wild-type-inoculated animals became colonized. These results do not permit us to conclusively distinguish between the two possibilities. However, because the cheR mutant could colonize within a reasonable period of time, we favor the former hypothesis, that the cheR mutant colonizes poorly because it fails to respond appropriately to an attractant. Although the identification of a hypothetical squid-produced attractant for V. fischeri remains unknown, a number of possibilities exist, including an amino acid (e.g., serine), a peptide(s) present inside the light organ (Graf and Ruby, 1998) and a component of the squid-secreted mucus (e.g. NANA, (DeLoney-Marino et al., 2003)). Marine pathogens such as V. anguillarum and V. alginolyticus migrate to fish mucus (Bordas et al., 1998; Larsen et al., 2001; O’Toole et al., 1996; O’Toole et al., 1999), although the glycan sugar components of the mucus do not serve as attractants for V. anguillarum (O’Toole et al., 1999). Finally, it is not unprecedented that a mutant could simply be inefficient at reaching the crypt spaces: a poorly motile mutant defective for the flaA flagellin gene exhibited both a significant delay in colonization initiation and an attenuated ability to colonize the crypts, despite being able to aggregate on the light organ surface and migrate into the pores (Millikan and Ruby, 2004). Thus, the cheR mutant may be slightly less efficient in its ability to reach the deep crypts to colonize, and therefore gets outcompeted by parental strains of the bacterium in co-inoculation experiments.

In any case, the results presented here demonstrate a requirement for chemotactic motility--at the very least, the ability to perform the characteristic run and tumble behaviors necessary for migrating up a gradient of attractant during symbiotic colonization. These results, in combination with previous studies indicating the stringent requirement for normal motility, further support the notion that V. fischeri must actively participate to initiate symbiotic colonization. The identity of a hypothetical chemoattractant(s) during colonization of E. scolopes by V. fischeri remains to be determined.