Characterization of Chemotactic Responses and Flagella of Hyphomicrobium Strain W1-1B

Abstract

Motile swarmer cells of Hyphomicrobium strain W1-1B displayed positive chemotactic responses toward methylamine, dimethylamine, and trimethylamine but did not display significant chemotactic responses towards methanol and arginine. Electron micrographs of negatively stained intact flagellar filaments indicated a novel striated surface pattern. The flagella were composed of two proteins of 39 and 41 kDa. Neither protein was a glycoprotein as determined by Schiff’s staining and by enzyme immunoassay. Protein fingerprints visualized from silver-stained polyacrylamide gels and Western blots of protease-digested samples indicated that the two proteins were similar but not identical. Monoclonal antibodies prepared to the complex flagella of Rhizobium meliloti cross-reacted with the striated flagella of Hyphomicrobium strain W1-1B; however, these antibodies did not cross-react with smooth-surface flagella. These results suggest that complex and striated flagella possess homologous epitope regions.

Hyphomicrobia are appendaged organisms with differential life cycles. During their life cycles, nascent swarmer cells are motile by means of a single subpolar flagellum, and as the cells mature, they shed the flagella and form a stalk. Virtually nothing is known about the regulation of cellular differentiation, motility, and chemotaxis in hyphomicrobia. The purpose of this work was to characterize the flagella and chemotactic responses of a Hyphomicrobium strain. Chemotactic sensing of substrate-rich zones enhances swarmer cell movement toward boundary layers, and flagella may facilitate the subsequent attachment to surfaces.

Bacteria and growth conditions.

Hyphomicrobium strain W1-1B used in this study was originally isolated from a water well sample and was identified as a Hyphomicrobium sp. on the basis of phenotypic and phylogenetic characterizations (11). Swarmer cells used for chemotaxis experiments were grown in medium 337 (4), which contains 0.5% (wt/vol) methylamine, dimethylamine, or trimethylamine as the sole C, N, and energy source. Static cultures were incubated at 22 ± 2°C for 72 h. The cells were harvested by centrifugation and washed twice with 337 buffer at pH 7.0 (consisting of medium 337 minus any carbon source). Washed cells were resuspended in 337 buffer.

Chemotaxis to methylated amines.

Chemotaxis was measured by capillary assays in chemotaxis chambers as described by Palleroni (7). The attractants tested were methylamine, dimethylamine, trimethylamine, methanol, and arginine. Each assay was replicated three to four times. All chemotaxis assays were conducted at 22 ± 2°C for 60 min. After incubation, samples were plated onto agar-solidified medium 337 supplemented with 0.5% methylamine by using a model DU spiral plater (Spiral Systems, Inc., Cincinnati, Ohio). The plates were incubated at 22 ± 2°C for 7 days before colony counting.

When Hyphomicrobium strain W1-1B was grown on methylamine, dimethylamine, or trimethylamine, swarmer cells were motile and exhibited a positive chemotactic response toward the respective substrates. Concentration-response curves which illustrate the degrees of responses to various concentrations of the chemoattractants are shown in Fig. 1.

FIG. 1.

Concentration-response curves of Hyphomicrobium strain W1-1B. The cells were grown with methylamine (A), dimethylamine (B), or trimethylamine (C). The responses were determined with methylamine (•), dimethylamine (▪), or trimethylamine (▴) as an attractant. The values (with I bars indicating standard deviations) represent accumulations of cells in response to the chemoattractant. Background cell accumulations (in the absence of attractant) varied from 5.5 × 10³ ± 2.5 × 10³ to 1.2 × 10⁴ ± 3.1 × 10³ cells per capillary.

Table 1 summarizes the peak response ratios and threshold values for all experimental conditions tested. Methylamine-grown cells proved to be the most chemotactically responsive. The relative chemotactic responses of methylamine-grown cells to methylamine and dimethylamine were 43- and 39-fold over background, indicating that these compounds are strong attractants (Fig. 1; Table 1). Dimethylamine- and trimethylamine-grown cells exhibited moderate chemotaxis to all three methylated amines.

TABLE 1.

Compilation of chemotaxis data for Hyphomicrobium strain W1-1B grown with methylated amines

Growth substrate	Attractant	Peak response (mM)	Chemotactic response ratioc			Threshold concn (mM)
			A	B	C
Methylamine	Methylamine	10	17	43	4.2	3 · 10−3
	Dimethylamine	0.1	15	39	33	3 · 10−4
	Trimethylaminea	NAb	NA	NA	NA	NA
	Arginine	NA	≤1.0	≤1.0	≤1.0	NA
	Methanol	NA	≤1.0	≤1.0	≤1.0	NA
Dimethylamine	Methylamine	1.0	3.7	8.1	3.9	5 · 10−3
	Dimethylamine	1.0	9.0	12	7.1	2 · 10−4
	Trimethylamine	1.0	5.8	7.2	1.6	2 · 10−2
	Arginine	NA	≤1.0	≤1.0	≤1.0	NA
	Methanol	NA	≤1.0	≤1.0	≤1.0	NA
Trimethylamine	Methylamine	1.0	6.4	13	4.5	1 · 10−4
	Dimethylamine	1.0	8.5	19	7.5	3 · 10−4
	Trimethylamine	1.0	17	17	7.8	2 · 10−5
	Arginine	NA	≤1.0	≤1.0	≤1.0	NA
	Methanol	NA	≤1.0	≤1.0	≤1.0	NA

^aThe trimethylamine data for methylamine-grown cells were not consistent between different trials because of poor response, and therefore the chemotactic response ratio and the threshold response were not calculated. 

^bNA, not applicable. 

^cIndicated are the ratios for the attractant concentrations that are 1 order of magnitude lower (A) or higher (C) than the concentration at which the peak response (B) occurs. The definitions by Mesibov and Adler (6) for threshold concentration, peak response, and peak concentration are used here. 

Arginine and methanol failed to elicit a statistically significant response from cells of Hyphomicrobium strain W1-1B grown on any of the methylated amines (Table 1). This result was unexpected because Hyphomicrobium strain W1-1B is able to use arginine as the sole source of N, and amino acids have been shown to be among the strongest chemoattractants for a variety of chemotactic bacteria (1). Methanol is an excellent source of C for hyphomicrobia, and growth on methanol results in high yields of growth. However, there is no known methanol-specific receptor or uptake system in microorganisms, perhaps because methanol is rapidly diffused into cells. Hyphomicrobium strain W1-1B, like many other chemotactic bacteria, may have evolved chemotactic responses to compounds that can serve as both C and N sources, such as the methylated amines.

Hyphomicrobium strain W1-1B was grown aerobically with methanol for chemotaxis experiments, but the methanol-grown swarmer cells were not motile during any phase of growth. However, flagella could be readily isolated from these cells. The addition of various concentrations of methanol to motile, methylamine-grown cells did not inhibit motility, indicating that methanol does not directly inhibit motility as it does for Escherichia coli (5). Methanol-grown swarmer cells were unaffected by the addition of culture supernatant from motile, methylamine-grown cells, indicating that there is no extracellular, soluble factor responsible for inducing motility in this organism.

Flagellar structure.

To obtain flagellum preparations, Hyphomicrobium strain W1-1B was grown in 10-liter fermentor batches (1% inoculum) containing medium 337 with 0.5% (wt/vol) methylamine or 0.5% (vol/vol) filter-sterilized methanol as the sole source of C and energy. The fermentors were incubated at 22 ± 2°C with aeration but no stirring for 72 h. The cells were harvested by tangential filtration (Millipore Pellicon filter), pelleted, and resuspended in 337 buffer. Flagella were removed by shearing cells in a blender at the maximum speed for 15 s. Cells were observed microscopically to ascertain the loss of motility and then separated from the free flagella by centrifugation at 10,000 × g and 4°C for 12 min. Cell debris was removed from the flagella by centrifugation of the supernatant at 15,000 × g and 4°C for 12 min. Flagella were pelleted by ultracentrifugation and then resuspended in HEC buffer (containing 10 mM HEPES at pH 7.0, 10 μM EDTA, and 0.2 mM CaCl₂) and stored at 4°C. The protein contents of flagellum preparations were determined with a Micro BCA Protein Assay Reagent kit (Pierce Chemical Co., Inc., Rockford, Ill.), with bovine serum albumin as a standard in the HEC buffer.

Flagellum preparations were purified by sucrose density gradient centrifugation with a 30 to 40 to 50% step gradient and separated by ultracentrifugation at 100,000 × g at 4°C for 18 h. The sucrose density gradients were fractionated by passing them through a UV detector set at 280 nm. The refractive indexes of the fractions were measured with a refractometer (Bausch & Lomb, Inc., Rochester, N.Y.).

To observe the ultrastructures of flagella, a diluted suspension of flagella was placed onto a Formvar, carbon-coated grid and negatively stained with uranyl acetate. Samples were observed with a Zeiss 10 transmission electron microscope.

Electron micrographs of Hyphomicrobium strain W1-1B flagellum preparations are shown in Fig. 2. The flagellar filaments did not appear to have the cross-hatched surface pattern typical of complex flagella, such as those of Rhizobium meliloti, Rhizobium lupini, and Bradyrhizobium japonicum (10). However, the surfaces of the filaments appeared to be of a striated type rather than of the smooth-surface type characteristic of the surfaces of flagella of E. coli and Salmonella spp. (2).

FIG. 2.

Transmission electron micrograph of flagella of Hyphomicrobium strain W1-1B following purification by density gradient centrifugation. (A) Purified flagella; (B) flagellar hook (arrow). Bars, 100 nm.

To further analyze the structures of the flagella, flagellin proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Analysis by SDS-PAGE demonstrated the presence of two flagellin monomers, of 39 and 41 kDa, in the purified Hyphomicrobium flagellum preparations (Fig. 3A, lane 1). At least two flagellins have also been reported for Bdellovibrio bacteriovorus, Halobacterium halobium, Campylobacter coli, Caulobacter crescentus, and R. meliloti (8). In contrast, the filaments of E. coli and Salmonella typhimurium are made up of a single species of flagellin.

FIG. 3.

SDS-polyacrylamide gel (A) and Western blot (B) of purified flagella of Hyphomicrobium strain W1-1B previously grown with either methylamine (lane 1) or methanol (lane 2).

The two proteins associated with the flagella of Hyphomicrobium strain W1-1B were tested for glycosylation with a First Choice Glycan Detection kit (Boehringer Mannheim Corp., Indianapolis, Ind.) to visualize antibody-glycoprotein complexes and by Schiff’s staining procedure (3). In both cases, transferrin was used as a positive control and creatinase was used as a negative control. Results from both the Schiff’s staining and enzyme immunoassay indicated that neither the 39- nor the 41-kDa flagellin monomer was glycosylated (data not shown).

Western blotting was performed to examine the antigenic surface structures of Hyphomicrobium. Proteins blotted onto nitrocellulose membranes were exposed to monoclonal antibodies specific to R. meliloti 7201 flagella and then visualized by the ProtoBlot II AP System with Stabilized Substrate, Mouse (Promega Corp., Madison, Wis.). The monoclonal antibodies were prepared at the Monoclonal Antibody Facility at The Ohio State University (Columbus) by using sucrose-gradient-purified flagellum preparations from R. meliloti RMB7201 as the antigen.

Western hybridization of the two Hyphomicrobium flagellar proteins showed that both the 39- and 41-kDa proteins were cross-reactive with the monoclonal antibody against R. meliloti flagella (Fig. 3B). This monoclonal antibody also cross-reacted with the flagella of Bradyrhizobium japonicum but not with those of Agrobacterium tumefaciens, Pseudomonas aeruginosa PAO1, or the atrazine-degrading soil isolate M91-3 (9) (Fig. 4). These results indicate that homology exists between the complex flagella of R. meliloti RMB7201 and B. japonicum and the flagella of Hyphomicrobium strain W1-1B. Because the monoclonal antibodies were prepared against whole, intact flagella of R. meliloti, the common antigen may be related to surface structure.

FIG. 4.

Western blot of flagellum preparations from several bacteria with monoclonal antibody prepared against whole, intact flagella from R. meliloti RMB7210. Lane 1, R. meliloti; lane 2, Hyphomicrobium strain W1-1B; lane 3, P. aeruginosa; lane 4, A. tumefaciens; lane 5, B. japonicum; lane 6, M91-3.

Flagella prepared from cells grown with methanol and with methylamine were analyzed by SDS-PAGE and with Western blots that were prepared with the R. meliloti RMB7201 antiflagellar monoclonal antibody (Fig. 3). Comparable proteins were detected in both flagellum preparations, although the methanol-grown cells lacked motility. Interestingly, the 39-kDa protein was more abundant in the flagella of methanol-grown cells while the 41-kDa protein was more abundant in the flagella of methylamine-grown cells. Flagellar preparations from methanol-grown cells cross-reacted with the monoclonal antibody, indicating that the proteins were comparable to those from the flagella of cells grown in methylamine.

The two proteins associated with Hyphomicrobium strain W1-1B flagella were compared to each other and to flagellin proteins of R. meliloti for degrees of similarity with a Protein Fingerprinting System kit (Promega Corp.). Proteins were separated by electrophoresis, and gels were processed according to the directions of the manufacturer (Promega Corp.). Gels were stained by a modified silver stain method (1a) or transferred to a nitrocellulose membrane by Western blotting with a Bio-Rad transblot apparatus. The blotted proteins were probed with the monoclonal antibody specific to R. meliloti RMB7201 flagella.

Fingerprints of the 39- and 41-kDa flagellin proteins of methylamine-grown Hyphomicrobium strain W1-1B and of the flagellin protein from R. meliloti RMB7201 were prepared by digestion with alkaline protease, endoproteinase Lys-C, and endoproteinase Glu-C. Only the digests with Glu-C resulted in the complete digestion of the proteins to peptide fragments, yielding reproducible fingerprints (Fig. 5). The fingerprints of the Hyphomicrobium flagellins as visualized on the Western blot were more similar than were the fingerprints from SDS-PAGE, suggesting homology between the epitope regions. However, the fingerprints of the two Hyphomicrobium proteins were different enough to indicate that the proteins were not identical. The two Hyphomicrobium flagellins showed more similarity to each other than either did to R. meliloti flagellin, in spite of the monoclonal antibody cross-reactivity.

FIG. 5.

Endoproteinase Glu-C digestion of the two flagellum-associated proteins of Hyphomicrobium strain W1-1B and a flagellum preparation of R. meliloti. (A) SDS-polyacrylamide gel; (B) Western blot. Lane 1, 41-kDa flagellin of Hyphomicrobium; lane 2, 39-kDa flagellin of Hyphomicrobium; lane 3, flagellum preparation of R. meliloti.

Conclusions.

This work shows that Hyphomicrobium strain W1-1B is chemotactic to methylated amines that can serve as both C and N sources for this methylotrophic bacterium. Chemotactic responses, which have not been previously described for any Hyphomicrobium spp., were exhibited during the motile phase of the organism’s differential life cycle, Hyphomicrobium strain W1-1B has complex, striated flagella which posses epitope regions homologous to those of the flagella of R. meliloti. The production of nonfunctional flagella by methanol-grown cells suggests an unusual structure-function relationship that warrants further studies of the regulatory mechanisms of the motility and changes with life cycle in this methylotrophic organism.