INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Many bacterial species are motile by means of flagellar propulsion (reviewed in references 5, 32, and 33). Powered by a rotary motor, the flagellum acts as semirigid helical propeller, which is attached via a flexible coupling, known as the hook, to the basal body. The basal body consists of rings and rods that penetrate the membrane and peptidoglycan layers. Associating with the basal body and projecting into the cytoplasm is a structure termed the C ring, which contains the switch proteins and acts as the core, or rotating part, of the motor. Maintenance of a flagellar motility system is a sizable investment with respect to cellular economy in terms of the number of genes and the energy that must be committed to gene expression, protein synthesis, and flagellar rotation. As a result, flagellar systems are highly regulated. A hierarchy of regulation has been elucidated for peritrichously flagellated Escherichia coli and Salmonella enterica serovar Typhimurium (26, 27, 30). This scheme of control couples gene expression to assembly of the organelle. The pyramid of expression possesses three classes, or tiers, of genes. Genes in each class must be functional in order for expression of the subsequent class to occur. Class 1 genes, flhD and flhC, encode the master transcriptional activators of class 2 flagellar gene expression. The flhDC operon is controlled by a ⁷⁰ promoter and a number of global regulatory factors (28). The majority of the class 2 flagellar genes encode components of the flagellar export system and the basal body (21). One class 2 gene encodes an alternative sigma factor devoted to recognition of flagellar genes (44). Flagellar class 3 operons are positively controlled by the flagellar ²⁸ factor and negatively regulated by FlgM, an anti-sigma factor (45). The anti-sigma factor is retained within the cell until the flagellar basal body and hook are completed (18, 29). At that time, FlgM is exported, and ²⁸ becomes free to direct expression of class 3 genes encoding flagellin subunits, hook-associated, motor, and chemotaxis signal transduction proteins. There are additional intricacies to this cascade, e.g., transcriptional classes within classes and translational modulation coupled to basal body assembly, as well as linkage between cell division and flagellar production (1, 22, 31, 48, 49). The regulatory hierarchy established for E. coli and S. enterica serovar Typhimurium serves as the paradigm for peritrichous flagellar systems of other bacteria.

The other well-characterized set of flagellar genes and scheme of flagellar control are those of Caulobacter crescentus (reviewed in references 46 and 65). In this organism, flagellation and cell division are strikingly coupled. On cell division, the daughter cell is motile and propelled by a polar flagellum, while the mother cell is nonmotile and stalked. DNA replication is repressed in the motile cell until later in the cell cycle when that cell differentiates to a new stalked cell. Many of the genes required for flagellar biosynthesis are homologs of E. coli and S. enterica serovar Typhimurium genes; however, the flagellar hierarchy differs between C. crescentus and the enteric bacteria. The flagellar genes of C. crescentus are organized in four levels of expression with two assembly checkpoints: completion of the MS-ring-switch export complex and completion of the basal body-hook structures. Genes at the bottom of the hierarchy are transcriptionally regulated from ⁵⁴ promoters. The master transcriptional regulator at the top of the hierarchy is CtrA, a member of the response regulator family of two component signal transduction systems, and this regulator controls the initiation of DNA replication, DNA methylation, cell division, and flagellar biogenesis (11).

The flagella of V. parahaemolyticus are of particular interest because this organism possesses two flagellar systems: a peritrichous (or lateral) one that is expressed when the bacterium is on a surface or in viscous environments and a polar system that is expressed continuously, i.e., when the bacterium is grown in liquid or on surfaces (reviewed in reference 39). Thus, under some conditions the bacterium simultaneously assembles two distinct flagellar organelles. Prior genetic analysis suggests that the gene systems are distinct and that no structural or assembly components are shared; mutants isolated with defects in swarming translocation are competent for swimming motility in liquid, and swimming-defective mutants remain proficient for swarming. Energy for rotation of the two kinds of flagella derives from different sources. The sodium motive force powers rotation of the polar flagellum, and the proton motive force drives rotation of the lateral flagella. Some of the chemotaxis genes have been demonstrated to be shared by the two motility systems (53). In addition to its propulsive role in swimming, the polar flagellum is believed to act as a tactile sensor informing the bacterium of contact with surfaces. Conditions that inhibit rotation of the polar organelle induce the alternative, lateral motility system. In this work, we elucidate the genes and gene organization involved in the polar motility system. Until now the circuitry of a polar flagellar system, apart from C. crescentus, has not been traced. This work should provide a foundation for gaining insight into the flagellar organelle and regulation of flagellar gene expression for a number of polarly flagellated bacteria, including Pseudomonas aeruginosa, Vibrio cholerae, and other marine Vibrio species.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and growth conditions. V. parahaemolyticus strains were cultured at 30°C. The strains used in this work are derivatives of V. parahaemolyticus BB22 (4). Strain LM1017 contains a mutation in the lateral flagellar hook gene and is unable to swarm (42). Strains were routinely propagated in heart infusion (HI) broth, which contained 25 g of HI broth (Difco) and 20 g of NaCl per liter. Marine broth 2216 (Difco) (28 g per liter) was filtered after autoclaving to remove precipitate. Solidified swarming medium was prepared by adding 15 g of Bacto-Agar (Difco) per liter to HI broth. Semisolid motility medium (M agar) contained 10 g of tryptone, 20 g of NaCl, and 3.25 g of agar per liter.

Genetic and molecular techniques. General DNA manipulations were adapted from the methods of Sambrook et al. (52). Transposon mutagenesis with mini-Mu lac (Tet^r) and the strategy for cloning the targeted gene have been described previously (58). The V. parahaemolyticus cosmid library was prepared by using the pLAFRII vector (40). Chromosomal DNA was prepared according to the protocol of Woo et al. (64). Southern blot analysis of restricted genomic DNA (52) was performed on Hybond-NX nylon membranes (Amersham Life Science).

Motility assays. The effect of mutations on swimming motility was assessed by examining movement in M agar. To document swimming motility, plates were inoculated with 2 µl of an overnight culture of cells normalized to an optical density at 600 nm (OD₆₀₀) of 2.0. Plates were incubated and photographed using a Kodak Digital Imaging System.

Immunoblot analysis. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was conducted as described previously (40). Resolving gels contained 12% acrylamide. Gels were transferred to polyvinylidene fluoride membrane (Immobilon-P; Millipore Corp.) in buffer containing 12.5 mM Tris base, 96 mM Glycine, and 20% methanol for 90 min at 30 V. After blocking in TBST buffer (10 mM Tris-Cl [pH 8], 0.15 M NaCl, 0.05% Tween 20) containing 5% nonfat dry milk, blots were incubated in TBST buffer with antiflagellar antibodies. The production of antibodies to polar and lateral flagellins has been described previously (34, 42). The secondary antibody was anti-rabbit immunoglobulin conjugated to horseradish peroxidase (Amersham Life Sciences). It was incubated with the blot at a dilution of 1:20,000 in TBST for 1 h. Development of the immunoblot utilized the chemiluminescent Super Signal substrate (Pierce) according to manufacturer's instructions.

Primer extension analysis. RNA was prepared with Trizol reagent (Gibco-BRL/Life Technologies, Grand Island, N.Y.) according to the manufacturer's protocol. Broth-grown cells were harvested in late exponential phase (OD₆₀₀ = 1.0). Plate-grown cells were harvested in cold 0.3 M sucrose after 5 to 7 h of growth. Primer extension analysis was performed as described previously (38) by use of the avian myeloblastosis virus reverse transcriptase (Promega, Madison, Wis.).

Sequence analysis. Sequence determination on both strands was performed by the DNA Core Facility of the University of Iowa. Sequence assembly and detection of potential rho-independent transcriptional terminators were accomplished by using the Genetics Computer Group (GCG) software package. Searches for homology were performed at the National Center for Biotechnology Information with the BLAST network service (2). Multiple sequence alignments were performed by using the CLUSTAL W program (62).

Nucleotide sequence accession number. The nucleotide sequences have been deposited in GenBank, and the accession numbers are U12817, AF069392, AF069391, U09005, and U06949.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Transposon mutagenesis and isolation of strains with swimming motility defects. After mini-Mu lac (Tet^r) mutagenesis of strain LM1017, a transposon bank containing approximately 15,000 mutants was screened for mutants with defects in the polar motility system. Strain LM1017 contains a lux operon fusion to the lateral flagellar hook gene; therefore, there is no contribution from the lateral flagellar system to the motility of this strain (42). Strain LM1017 expresses the lfgE::lux fusion when grown on solidified medium and is luminescent. All of the mutants with swimming motility defects produced as much light on plates as the parental strain LM1017 produced, were unable to swarm over surfaces, and failed to synthesize lateral flagellin. Strains that appeared nonmotile or poorly motile in semisolid motility (M) agar potentially possessed defects in polar flagellar structure or assembly, motor function, or chemotaxis.

Phenotypic analysis of mutants. The majority of nonmotile mutants of E. coli are nonflagellated (Fla) due to the nature of feedback control built into the hierarchy of gene expression (66). Loss-of-function mutations in only two genes (motA and motB) yield the Mot phenotype, which is a flagellated but paralyzed cell. Insertion of the torque-generating components of the motor into the membrane is not required for assembly of the E. coli flagellar organelle, and expression of mot genes occurs at the final stage in the hierarchy of expression (57). The phenotypes of V. parahaemolyticus motility mutants differed from E. coli. Forty mutants were segregated into four phenotypic classes: class 1, Fla mutants were nonmotile in semisolid M agar and in the light microscope and failed to produce flagellin in immunoblots (26%); class 2, Mot mutants were nonmotile in M agar and in the microscope but produced flagellin antigen levels equivalent to the wild-type strain (1%); class 3, Che mutants appeared nonmotile in M agar but motile in the light microscope and produced wild-type levels of flagellin (39%); and class 4, Mot^± mutants showed limited radial expansion in M agar after prolonged incubation and produced detectable, but low levels of Fla antigen (34%).

View larger version (37K): [in this window] [in a new window]   FIG. 1.   Swimming motility of mini-Mu mutant strains in M agar with tetracycline. All strains were derivatives of strain LM1017. Plates A and B contain the strains indicated in the top row at the left and were incubated at 30°C for 10 and 24 h, respectively. Plates C and D contain the strains indicated in the lower row on the left and were incubated at 30°C for 12 and 24 h, respectively. Strain LM5053 was not inoculated in plates B or D. The control strain LM5053 was tetracycline-resistant and exhibited wild-type motility.

View larger version (72K): [in this window] [in a new window]   FIG. 2.   Western immunoblot analysis of polar flagellin production. Blots were reacted for 2 h with antiserum (1:1,000 dilution) directed against polar flagellins (Fla). The strain numbers are indicated above the lanes. The polar flagellins are similar in molecular size and comigrate in the resolving gel system used. An antiserum-reactive, nonflagellin band serves as a control for the amount of whole cells loaded in each lane.

Identification of polar flagellar genes: physical organization and predicted function. The tetracycline resistance from mini-Mu and flanking chromosomal DNA was cloned from some of the mutants of each class and used as a probe to retrieve cosmids from a library of V. parahaemolyticus DNA. Each cosmid contained inserts of approximately 25 kb of DNA. Cosmids were used as probes for Southern blots containing restricted chromosomal DNA of the mutant strains. The cosmids revealed perturbations of the restriction pattern due to transposon insertion and allowed segregation of the mutants into linkage groups. DNA from mutants failing to show perturbations on Southern analysis was used to prepare subsequent substrates for cloning to retrieve additional loci. The initial sequence was obtained from the Mu-derived, tetracycline-resistant clones, and the nucleotide information obtained was used to continue sequencing on both strands of the cosmid clones.

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Organization of polar flagellar gene system. Arrows indicate the direction of transcription and the extent of coding sequence for each gene. The filled circles indicate the 54 class, and the open circles indicate the 28 class of promoters that have been mapped by primer extension analysis. The promoter region for flaC is unusual and is indicated by the asterisk. The filled boxes indicate predicted rho-independent transcriptional terminators.

Correlation of genotype with phenotype. Sequence information coupled with restriction patterns using Southern analysis allowed assignment of lesions to specific gene intervals. The majority of the chemotaxis-defective mutants analyzed showed transposon-induced perturbations that placed the insertion defects within che clusters in region 1 or region 2. A minority (1%) of nonmotile V. parahaemolyticus mutants displayed the Mot phenotype, and three mutants were determined to contain mutations in novel motor genes, motX and motY (35, 36). A correlation of the genotype of the Fla and Mot^± mutants examined in Fig. 1 and 2 with phenotype is presented in Table 1. Strains LM5040, LM5043, LM5045, and LM5046 produced no flagellin, and the mini-Mu insertions in these strains mapped to intervals within the flhBA locus, which encodes components of the flagellar export pathway. One other insertion in the flhBA locus was detected. The phenotype and genotype of this strain, LM5051, was puzzling until the precise mutation was cloned and sequenced. LM5051 was partially motile and produced levels of flagellin comparable to wild-type levels. Cloning and sequencing of the mini-Mu insertion of this strain revealed that the transposon was inserted into the intergenic region between flhA and flhF. Nonmotile strain LM5042 also produced as much flagellar antigen as the wild type and contained a defect in the region encoding hook-associated-like proteins (the flgKL interval). The phenotype of LM5042 matched other strains with insertions in flgK and flgL that were previously created by allelic exchange (37). All of the mutants in the Mot^± class mapped in the flgB-flgH interval, which encodes hook and basal body components. Thus, mutants with defects in genes encoding assembly apparatus fail to produce a flagellum or synthesize flagellins, whereas mutants with defects in many of the genes encoding structural parts of the basal body, but not hook-associated proteins, seem to be able to occasionally synthesize a functional polar flagellum.

Six polar flagellin genes. Prior work identified genes encoding four polar flagellin subunits that were organized in tandem in two distinct loci, flaBA and flaCD (37). Further sequencing of the flagellin-encoding loci revealed two additional flagellin genes, flaF (located upstream of flaB transcription) and flaE (located downstream of flaD transcription). Thus, the present total number of genes encoding the structural subunits of the polar flagellar filament is six. A comparison of their relatedness to each other and to the lateral flagellin is shown in Table 3. Their location, homology, and genetic analysis suggest that these are polar genes; however, none of the flagellin genes appears to be essential for polar filament formation (37).

flaFBA flaCD

View larger version (83K): [in this window] [in a new window]   FIG. 4.   Primer extension analysis of flagellin gene transcription. RNA was prepared from the wild-type strain BB22 after growth in HI broth (B) or on HI plates (P). The amount of RNA in each lane was as follows: 1, 2 µg (B); 2, 2 µg (P); 3, 2 µg (B); 4, 5 µg (P); 5, minus RNA; 6, minus RNA; and 7, 2 µg (B). Lanes g, a, t, and c correspond to the dideoxy nucleotide used in the sequencing reactions. Sequence and primer extension products were generated with flaA-, flaB-, or flaD-specific primers. The primers were flaA (5'-GCTGTGCGGTCATCGCAGAAACG-3'), flaB (5'-GTGTTTAATTCACTGCCATG-3'), and flaD (5'-CGGTCATCGCTGATACGTTAGTG-3').

Expression of flaC is unique. Figure 5 shows the primer extension reactions that were initiated using an oligonucleotide specific for flaC. A major product was obtained (lane 1, indicated by arrow), and the upstream sequences do not resemble the upstream regions of the other ²⁸-like flagellin promoters (Table 4). Moreover, the product was only observed in RNA prepared from plate-grown wild-type cells and not in RNA from wild-type cells grown in broth (lane 1 versus 2). Expression of flaC appeared to be surface dependent. The control panel, labeled flaD, demonstrates that equivalent amounts of plate- and broth-derived RNA were used. The control reactions used the same RNA preparations and a flaD-specific primer (lanes 1 and 2 compared to 4 and 5, respectively). The flaD transcript was expressed in the wild-type strain in liquid and on surfaces. We have shown previously that plate-grown cells are starved for iron (41). To examine whether the environmental signal controlling flaC expression might be the result of iron starvation or some other nutrient condition, RNA was prepared from the wild-type cells that were grown in 2216 marine broth, which is growth limiting for phosphate and iron (40, 41). No flaC-dependent primer extension product was observed under this condition, although a flaB-dependent transcript could be detected (lane 7 versus lane 10). Primer extension reactions using RNA prepared from HI broth and plate cultures that were cultured in parallel to the 2216 broth culture reproduced the surface-induced flaC product (lanes 8 and 9 versus lanes 11 and 12). There is also a ladder of large primer extension products (most evident in lanes 7 to 9), suggesting some basal level of readthrough transcription originating with the upstream flgK operon.

View larger version (108K): [in this window] [in a new window]   FIG. 5.   Primer extension analysis of flaC transcription. The arrows indicate the surface-dependent flaC primer extension product. RNA was prepared from the wild-type strain BB22 or LM1017 after growth in HI broth (B), on HI plates (P), or in 2216 marine broth (2216B). Approximately 2 µg of total RNA was used in each primer extension reaction. Reactions 1 to 3 and 7 to 9 were primed with a flaC-specific oligonucleotide. Reactions 4 to 6 were primed with a flaD-specific oligonucleotide. Reactions 10 to 12 were primed with a flaB-specific oligonucleotide. Lanes: 1, BB22 (P); 2, BB22 (B); 3, LM1017 (P); 4, BB22 (P); 5, BB22 (B); 6, LM1017 (P); 7, BB22 (2216B); 8, BB22 (B); 9, BB22 (P); 10, BB22 (2216B); 11, BB22 (B); and 12, BB22 (P). Lanes g, a, t, and c correspond to the dideoxy nucleotide used in the sequencing reactions. Sequence and primer extension products were generated with flaC-, flaD-, or flaB-specific primers. The flaC-specific primer was 5'-CTGTTACAGCCATTTTGCTCTCC-3'.

Other flagellar and chemotaxis promoters. To establish a basis for the polar flagellar regulatory hierarchy, the transcriptional start sites for a number of other promoters were determined. The basic strategy targeted genes that possessed significant upstream, intergenic, or noncoding sequence (usually >50 bp). Figure 6 shows the primer extension products for motA, motX, and motY. Table 4 shows the tabulation of the upstream sequence information for all of the transcriptional start sites that have been determined. Both motA and motX possess upstream sequences that resemble the ²⁸-dependent promoters, whereas the sequences upstream of motY appear to resemble a potential ⁵⁴-dependent promoter (TGGCAC n5 TTGC, containing an invariant 24 GG motif and a conserved 12 GC motif [56]). All mot transcripts were expressed in broth- and plate-grown cells. Table 4 also shows the transcriptional start site and upstream sequences for six other genes. The promoters for cheV and flgM appear to be ²⁸ dependent, and fliE, flgB, flgK, and flhA appear to be ⁵⁴ dependent. The evidence suggests that flgM is also transcribed from an upstream promoter because of the presence of faint ladders of abortive primer extension products that extend to the top of the sequencing gel. There is one other relatively large intergenic gap (175 bp) in region 2, which occurs between fliJ and fliK. Using fliK-derived primers, a ladder of prominent primer extension products was observed on sequencing gels. This evidence suggests that fliK is transcribed from upstream sequences as part of an operon. Primer extension has not proved suitable for transcriptional analysis of flaK and flaL because prominent ladders of extension products are obtained. We hypothesize that some of the products may be the result of multiple species of RNA (i.e., multiple promoter control) and some may be caused by premature termination due to RNA secondary structure (i.e., a potential rho-independent terminator is found between flaK and flaL).

View larger version (92K): [in this window] [in a new window]   FIG. 6.   Primer extension analysis of mot gene transcription. RNA was prepared from the wild-type strain BB22 after growth in HI broth (B) or on HI plates (P). The amount of RNA in each lane was as follows: 1, 2 µg (P); 2, 12 µg (P); 3, minus RNA; 4, 2 µg (P); 5, minus RNA; 6, minus RNA; 7, 2 µg (B); and 8, 5 µg (P). Lanes g, a, t, and c correspond to the dideoxy nucleotide used in the sequencing reactions. Sequence and primer extension products were generated with motA-, motX-, or motY-specific primers. The oligonucleotide primers were motA (5'-CCACCGATCAAACCTATTAGGGTTGC-3'), motX (5'-CAGTAACAGTGAAGCAGCCACTG-3'), and motY (5'-GTTATCAGCCATTTATTCATC-3').

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The compilation of the repertoire of polar flagellar and chemotaxis genes of V. parahaemolyticus represents a wealth of useful information pertinent to flagellar assembly, sheath formation, polar placement, and perhaps even the connection of flagellation with the cell cycle. In addition, because the polar flagellum of V. parahaemolyticus appears to act a mechanosensor (34), an understanding of polar flagellar structure and regulation is critical for gaining insight into the mechanism of surface sensing and swarmer cell development.

Flagellar assembly. Flagella are assembled via a type III export pathway. No consensus flagellar export signal has been defined, although a number of models have been proposed, and it seems likely that classes of sequentially exported proteins exist (8). Since V. parahaemolyticus can simultaneously assemble the lateral and polar flagella, it is an ideal system to study type III secretion determinants and the specificity of export. The sequence divergence observed between the N terminus of many (but not all) of the predicted polar V. parahaemolyticus structural proteins, as well as for the predicted chaperone-like molecules and FlgM, and components of the lateral V. parahaemolyticus system and flagellar systems of other bacteria may not only allow discernment of classes of export substrates but will also provide a system for testing potential signals.

Sheath formation and the basal body complex. Little is known about the formation or function of flagellar sheaths, which are found in many bacteria, including marine Vibrio species, V. cholerae, Bdellovibrio bacteriovorus, and Helicobacter pylori (reviewed in reference 59). These sheaths appear to be extensions of the cell outer membrane, although their composition suggests that the sheath forms a distinct, stable membrane domain. Moreover, some evidence suggests that the polar basal body structure differs from peritrichously inserted basal bodies. Two models for the basal organelle of polar flagella have been derived from electron microscopy studies of V. cholerae (sheathed), Campylobacter fetus (unsheathed), Bdellovibrio bacteriovorus (sheathed), and Wolinella succinogenes (unsheathed) (12, 13, 61). Regardless of whether the flagellum is sheathed or unsheathed, all of the studies report the existence in the basal body complex of a large convex disk situated below the outer membrane. We have also seen this disk with V. parahaemolyticus (unpublished observation). One model suggests that the disk is the P-ring equivalent (acting as the bushing associated with peptidoglycan) and that the L ring (lipopolysaccharide associated) is not present. The second model places the large disk between the P and L rings. Thus, the identity of the genes encoding basal body parts of a polar flagellum is of interest. We have found a locus encoding the V. parahaemolyticus genes for basal body and hook components. Genes for both the L and the P ring exist, providing genetic support for the second model. Whereas the V. parahaemolyticus P ring displays homology with other P rings over the full length of the protein, the N terminus of the V. parahaemolyticus L ring (FlgH) is divergent. The L ring protein of S. enterica serovar Typhimurium has been shown to be a lipoprotein and postulated to anchor the basal body in the outer membrane (54). Perhaps the nature of this protein is key to understanding differences between sheathed and unsheathed flagella. Such a possibility awaits further analysis, particularly biochemical elucidation of the N terminus of V. parahaemolyticus FlgH.

Polar flagellar placement. One pair of genes not found in E. coli includes flhF and flhG (region 2). The flhF gene was first discovered in B. subtilis, where it was demonstrated to be required for motility (7). A nonpolar, null mutation in flhF produced nonmotile cells lacking flagella. FlhF shows homology to FtsY, which is a GTP-binding protein involved in the signal recognition particle targeting pathway. Intriguingly, the V. parahaemolyticus polar homolog contains an insertion of ~170 aa that is not found in other homologs. The inserted domain shows homology to a eukaryotic sodium channel. It is tempting to speculate that the insertion is unique for sodium-type flagella because all of the FlhF sequences deposited in GenBank are derived from organisms that possess proton-driven flagella. Located 15 bp downstream of V. parahaemolyticus flhF is flhG, which encodes a protein that shows homology to MinD, a membrane ATPase involved in septum site determination. Perhaps FlhF and FlhG work as a pair to determine site selection of flagellar insertion. The FlhF-FlhG pair is found in a number of polarly flagellated bacteria. In P. aeruginosa, a mutation in flhG was recently shown to increase the number of polar flagella (9). It should be noted that the gene encoding ²⁸ is immediately downstream of flhG; in fact, translation appears to be coupled for the coding regions of flhG and fliA overlap by 10 bp.

Chemotaxis. Mutations in two distinct loci produced chemotaxis-defective strains. Possessing different kinds of upstream controlling elements, the two che clusters appear to occur in different classes of the hierarchy. One locus in region 2 encodes most of the major cytoplasmic chemotaxis proteins, i.e., CheY, CheZ, CheA, CheB, and CheW. It seems likely that these genes are under the control of ⁵⁴ since they are very tightly linked to each other in a potential operon initiating with flhA. The second cluster, which occurs in region 1, encodes CheB and a hybrid CheY-CheW protein, similar to CheV of B. subtilis (51). Transcription of cheV clearly initiates at ²⁸-type promoter sequences. Although we know that che mutations in region 2 affect polar and lateral motility (53) and that che lesions in region 1 perturb polar motility, the roles that region 1 che genes play in modulating lateral motility remain to be determined. Perhaps the region 1 che genes are dedicated to the polar system.

Additional ORFs. Three additional ORFs were found within the region 2 che locus. Two encode potential proteins of unknown function and the third encodes a protein that resembles Soj of B. subtilis and other chromosome-partitioning ATPases. In B. subtilis, Soj plays a role in cell cycle progression by coupling chromosome segregation to development (55). It seems curious that a Soj-like protein exists within a flagellum-chemotaxis operon and that this particular arrangement is found in other bacteria, e.g., P. aeruginosa and P. putida. Perhaps these novel ORFs will provide the key for a similar linkage between cell division and flagellation or development.

The polar flagellar hierarchy. Analysis of motility mutant phenotypes provides some insight into the hierarchy of V. parahaemolyticus polar flagellar gene control and assembly. We have previously shown that mutants with defects in any of the four polar motor genes produce flagella, whereas mutants with defects in the switch genes do not (6). Switch genes are known to be required for flagellar assembly, rotation, and chemotaxis (66). The switch genes are found in region 2 along with other genes known to participate in the flagellar assembly and export pathway. Mutants with defects in the flhBA interval, which encodes components of the export apparatus, displayed the same phenotype as switch mutants, i.e., they were nonmotile, nonflagellated, and unable to synthesize flagellin. Most of these genes in region 2 are tightly linked. Precedence for large motility operons has been established in other bacteria, e.g., Borrelia burgdorferi (15). Primer extension analysis identified a potential ⁵⁴-dependent promoter preceding fliE. We postulate genes in the fliE-flhB region constitute a large flagellar operon. Downstream of flhB and preceding flhA, there is a relatively large intergenic region (230 bp). Primer extension analysis identified a promoter region, and these sequences also resembled the canonical ⁵⁴-dependent promoter.

Summary. Thus, there are common themes in polar flagellar gene organization and regulation, but there also appear to be unique variations. The differences may reflect the lifestyle of each organism. For example, the organization of the multiple polar flagellin genes in V. parahaemolyticus is similar to loci in V. anguillarum and V. cholerae (24, 43). In these organisms, only one specific flagellin is required for motility, and its expression is under ⁵⁴-dependent control, whereas the other flagellin genes are dispensable and require ²⁸. The critical flagellin gene of V. anguillarum and V. cholerae is most equivalent with respect to gene location and the predicted protein sequence to V. parahaemolyticus flaC. The flaC gene is also under different environmental control from the other V. parahaemolyticus polar flagellins, which have ²⁸-dependent promoters. However, this gene is not essential for motility, and its regulation is not directed by ⁵⁴. The promoter structure of the flaC flagellin-encoding gene is unusual, and expression is controlled in a surface-dependent manner. What this means with respect to polar flagellar function and regulation and in the context of growth on surfaces and swarmer cell differentiation remains to be investigated.