A Family of Six Flagellin Genes Contributes to the Caulobacter crescentus Flagellar Filament

Abstract

The Caulobacter crescentus flagellar filament is assembled from multiple flagellin proteins that are encoded by six genes. The amino acid sequences of the FljJ and FljL flagellins are divergent from those of the other four flagellins. Since these flagellins are the first to be assembled in the flagellar filament, one or both might have specialized to facilitate the initiation of filament assembly.

Caulobacter crescentus is a dimorphic bacterium that differentiates as part of its normal cell cycle. Each cell division results in a stalked cell and a motile swarmer cell. The swarmer cell has a single flagellum at one pole that is responsible for motility. The flagellum is similar in structure to enteric flagella except that the filament is composed of multiple flagellin proteins (6, 10, 21, 35, 38). The hook-proximal portion of the flagellar filament consists of a 60-nm segment containing a 29-kDa flagellin (6). The next segment is 1 to 2 μm in length and consists of a 27-kDa flagellin along with increasing amounts of a 25-kDa flagellin at its distal end. The remaining 2 to 10 μm of the filament contains the 25-kDa flagellin. Thus, the flagellar filament consists of at least three different flagellar proteins in a precise arrangement.

In addition to the three flagellins found in the filament, stationary-phase wild-type cells and some flagellar mutants synthesize a 22-kDa flagellin (15, 34). The 22-kDa flagellin is not found in functional flagellar filaments and may result from improper processing of the 25-kDa flagellin (22, 34).

Multiple flagellin genes have been identified in two unlinked clusters in the C. crescentus genome (11, 34). The alpha gene cluster includes the fljJ, fljK, and fljL genes that encode the 29-, 25-, and 27-kDa flagellins, respectively. The gene order in this region is flaY-flaE-fljJ-fljK-fljL-flaF-flbT-flBA-flaG (30, 32). Thus, the alpha cluster of flagellin genes is embedded in a cluster of other flagellar genes. The beta flagellin gene cluster is approximately 1,000 kb from the alpha cluster (8). No other flagellar genes are present at this locus.

The flagellin gene family.

We determined the nucleotide sequences of five of the six flagellin genes in the alpha (fljK and fljL) and beta (fljM, fljN, and fljO) regions using the plasmids described in Table 1. The nucleotide sequence of the sixth gene, fljJ, had been determined previously (11), and errors in this sequence were corrected using nucleotide sequence information produced by The Institute for Genome Research. Each region encodes a set of three independently transcribed flagellin genes. A comparison of these sequences to the C. crescentus genome sequence being produced by The Institute for Genome Research indicated that no additional flagellin genes were present elsewhere in the genome. Thus, C. crescentus has six flagellin genes in two clusters of three genes.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Genotype	Derivation or reference
Caulobacter crescentus
CB15	Wild type	28
NA1000	syn-1000	9
SC276	flbT650	14, 23
SC280	flaG133	30, 33
SC295	flgI141	14, 19
SC508	fliQR153	14, 41
SC514	fla-158	34
SC1032	flbD198::Tn5 str-152	7, 29
SC1055	rpoN610::Tn5 proA103 str-140	7, 2
SC1117	flgH194::Tn5 str-152	7, 5
SC1131	fliM196::Tn5 cysB102 str-142	7, 40
SC3794	fla-158 syn-1000	Phage SC514 × NA1000
SC3845	flaG133 syn-1000	Phage SC280 × NA1000
Escherichia coli
NC2159	S17-1(pflbT::lacZ)	J. W. Gober
NC2160	S17-1(pfljK::lacZ)	39
NC2161	S17-1(pfljL::lacZ)	39
NC2255	S17-1(pfljN::lacZ)	This study
NC2264	S17-1(pfljO::lacZ)	This study
NC2272	S17-1(pfljM::lacZ)	This study
S17-1	pro r− m+recA zzz::RP4-2	36
Plasmids
S14	Cosmid containing fljJKL region	31
pTE10	1.4-kb S2-S3 fragment in pBluescript SK−	This study
pTE11	0.9-kb S2-E fragment in pBluescript SK−	This study
pTE13	0.7-kb S2-B1 fragment in pBluescript SK−	This study
pTE14	0.9-kb S2-E fragment in pBluescript SK+	This study
pTE15	0.6-kb S2-P1 fragment in pBluescript SK−	This study
pTE16	1.4-kb B2-B3 fragment in pBluescript SK−	This study
pTE17	1.0-kb B3-B4 fragment in pBluescript SK−	This study
pTE19	0.2-kb P1-B1 fragment in pBluescript SK−	This study
pTE20	1.4-kb B1-B2 fragment in pBluescript SK−	This study
pTE21	0.6-kb S4-B3 fragment in pBluescript SK−	This study

To facilitate comparison of members of the flagellin gene family, the nucleotide sequences of the flagellin genes were aligned. There were 455 variable sites in the 822-bp sequence. Most of the variable sites were found in the fljJ gene, which encodes the 29-kDa flagellin. The most similar pair of genes, fljM and fljN, contained 46 differences (5.6%), indicating that there has been considerable nucleotide divergence among all of the genes. Furthermore, transversions outnumbered transitions 25 to 21, suggesting either that the mutational process was approaching saturation or that selection was occurring. A neighbor-joining tree indicated that the three beta flagellin genes clustered with fljK, suggesting that all four genes encode 25-kDa flagellins (Fig. 1).

FIG. 1.

Neighbor-joining tree indicating phylogenetic relationships among flagellin genes. The same configuration was obtained by maximum-parsimony analysis. Flagellin gene phylogenies were determined using Test Version 4.0b3a of PAUP written by D. L. Swofford.

A comparison of the derived amino acid sequences of the flagellin genes demonstrated considerable variation among the flagellin proteins, with FljJ differing at approximately half of the amino acid positions (Fig. 2). The two most similar proteins, FljM and FljN, differed at 14 (5%) of 273 amino acid positions. A neighbor-joining tree identical to that obtained with the nucleotide sequences was obtained when the amino acid sequences were compared. This result provides further evidence that the three beta flagellin genes encode 25-kDa flagellins. This result is consistent with previous observations that the alpha region deletion mutants SC507 and PC7810 produce 25-kDa flagellins but not the 27- or 29-kDa flagellin (25, 34).

FIG. 2.

Comparison of derived amino acid sequences of the six flagellin genes. Dashes indicate deletions in the amino acid sequence relative to one or more of the other sequences. Dots indicate identity with the reference amino acid sequence.

N-terminal amino acid sequences of purified flagellins.

The 25- and 27-kDa flagellins were dissociated from intact flagellar filaments as described by Johnson et al. (16) and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. When the N-terminal amino acid sequences of these purified flagellins were determined, we found that the sequence of the 27-kDa flagellin was identical to that predicted from the sequence of the fljL gene (Fig. 3). However, the N-terminal amino acid sequence of the 25-kDa flagellin appeared to be a mixture of two proteins, one encoded by the fljK gene and one encoded by the fljM or fljN gene or both. Thus, the flagellar filament contains multiple 25-kDa flagellins in addition to the 27- and 29-kDa flagellins. No evidence was obtained for the presence of the FljO flagellin. However, the fljO promoter is expressed at approximately one-fourth of the combined rate of the fljM and fljN promoters (see Table 2 footnote). Consequently, it is possible that the FljO flagellin was present in the flagellar filament at levels too low to be detected.

FIG. 3.

N-terminal amino acid sequences of purified flagellin proteins. An amino acid pair, e.g., A/S, indicates that two amino acids were found at the same position, suggesting that the protein sample consists of a mixture of two proteins.

TABLE 2.

Relative levels of transcription from flagellin promoters^a

Strain	Avg transcription level ± SD
	pfljM	pfljN	pfljO	pfljK	pfljL	pflbT
Wild type	1.0	1.0	1.0	1.0	1.0	1.0
rpoN mutant	0.44 ± 0.02	0.22 ± 0.04	0.47 ± 0.12	0.015 ± 0.006	0.004 ± 0.001	0.36 ± 0.04
flbD mutant	0.37 ± 0.02	0.57 ± 0.11	0.77 ± 0.03	0.048 ± 0.03	0.013 ± 0.001	0.48 ± 0.12
fliM mutant	0.51 ± 0.02	0.73 ± 0.13	0.63 ± 0.03	0.98 ± 0.19	0.015 ± 0.004	0.70 ± 0.05
flgI mutant	1.21 ± 0.02	1.11 ± 0.12	0.98 ± 0.05	1.18 ± 0.16	1.00 ± 0.03	1.20 ± 0.22
fliQ mutant	1.12 ± 0.04	1.07 ± 0.13	0.77 ± 0.01	1.41 ± 0.11	0.025 ± 0.001	1.04 ± 0.20
flgH mutant	0.35 ± 0.10	0.74 ± 0.05	0.87 ± 0.04	0.80 ± 0.12	0.51 ± 0.04	0.56 ± 0.15
flbT mutant	0.48 ± 0.03	0.50 ± 0.06	0.89 ± 0.05	0.08 ± 0.02	0.059 ± 0.006	0.91 ± 0.05

^aAll of the values shown are averages of three independent determinations. The actual values (Miller units) in a wild-type genetic background were as follows: pfljM, 3,830; pfljN, 4,910; pfljO, 2,353; pfljK, 25,990; pfljL, 27,047; pflbT, 23,778. 

Flagellar filaments containing multiple flagellins have been found in a variety of other eubacterial genera, including Yersinia (18), Vibrio (20, 24), Helicobacter (17), Campylobacter (12), Agrobacterium (4), and Rhizobium (1, 27). However, in contrast to that in the C. crescentus flagellar filament, the physical arrangement of these multiple flagellins within the corresponding flagellar filaments is not known. In C. crescentus, the most divergent flagellin, FljJ, is found in small quantities in the region closest to the flagellar hook (6). Thus, FljJ is used to initiate flagellin subunit assembly and then other flagellins are assembled once a short segment is completed. These observations suggest that the divergent FljJ flagellin is a specialized flagellin that is designed to facilitate the first steps of filament assembly and attachment to the hook. Similarly, the moderately diverged FljL flagellin is located between the FljJ flagellin and the 25-kDa flagellins that form the main part of the filament. Thus, it may serve as an adapter between the two disparate flagellins. Mutants that do not produce the FljJ and FljL flagellins assemble a shortened flagellum and exhibit reduced motility (25, 33). Thus, an abnormal flagellar filament is produced in the absence of these two flagellins.

Origin of the 22-kDa flagellin.

The origin of the 22-kDa flagellin protein has been a mystery since it was first observed (13). To determine whether the 22-kDa flagellin was derived from one or more of the six flagellin genes, we purified the 22-kDa flagellin from the insoluble red material produced by the SC3845 mutant strain cells as described by Smit et al. (37) and determined its N-terminal amino acid sequence using an automated amino acid sequencer (Beckman Coulter, Fullerton, Calif.). The results were identical to those obtained with the 25-kDa flagellins (Fig. 3), indicating that the 22-kDa flagellin was derived from at least two of the 25-kDa flagellin genes. To confirm this result, we isolated the 22-kDa flagellin from SC3794, a strain that contains a deletion of the alpha region. The N-terminal sequence of this 22-kDa flagellin was identical to that of the FljM and FljN flagellins, and the sequence corresponding to the deleted fljK gene was not present. This result proves that both the fljK gene and at least one of the beta region genes are the source of the 22-kDa flagellins.

Regulation of flagellin gene expression.

Previous studies have demonstrated that the fljK and fljL genes are expressed from ς⁵⁴ promoters and that they require both the RpoN protein, encoding ς⁵⁴, and the FlbD activator protein for expression (2, 29). However, inspection of the regions upstream from the beta cluster flagellin genes revealed no sequences that resemble a ς⁵⁴ promoter. This result is consistent with the observation that flagellin gene expression does occur at a low level in rpoN and flbD mutants (26; J. Malakooti and B. Ely, unpublished data). To determine experimentally whether ς⁵⁴ is required for expression of the beta cluster flagellin genes, we cloned each of the promoters in front of a lacZ reporter gene and measured β-galactosidase levels in various genetic backgrounds (3). Expression of the fljMNO genes was only 10 to 20% of that measured with the fljK gene in a wild-type background (Table 2). However, when expression in an rpoN mutant background was compared, the fljM, fljN, and fljO promoters were expressed at 20 to 50% of the rate measured for their expression in wild-type cells. In contrast, the fljK and fljL genes were expressed at less than 2% of the rate measured in wild-type cells. Similar results were obtained when flagellin promoter expression was measured in a flbD mutant (Table 2). Thus, neither RpoN nor FlbD is required for fljMNO expression, and expression of these beta region flagellar genes can account for the residual flagellin synthesis observed in rpoN and flbD mutants. As with fljK, expression of the fljMNO promoters was not affected by the presence of other flagellar mutations (Table 2).

Previous studies have demonstrated that flbT mutants overproduce the 25-kDa flagellin proteins (15, 30). The FlbT protein must be involved in flagellin mRNA turnover, since flagellin mRNA is more stable in an flbT mutant strain than in a wild-type strain (23). When expression of the fljM, fljN, and fljO promoters was measured in an flbT mutant background, expression was 45 to 90% of that measured in wild-type cells. Since expression from the fljK and fljL promoters is reduced 10-fold in flbT mutants, the beta cluster flagellin genes are likely to be responsible for most of the overproduction of flagellin observed in flbT mutants.

Nucleotide sequence accession numbers.

The flagellin gene nucleotide sequences in Fig. 2 have been assigned GenBank accession numbers AF089835 and AF040268.