Identification and Localization of Flagellins FlaA and FlaB3 within Flagella of Methanococcus voltae

Abstract

Methanococcus voltae possesses four flagellin genes, two of which (flaB1 and flaB2) have previously been reported to encode major components of the flagellar filament. The remaining two flagellin genes, flaA and flaB3, are transcribed at lower levels, and the corresponding proteins remained undetected prior to this work. Electron microscopy examination of flagella isolated by detergent extraction of whole cells revealed a curved, hook-like region of varying length at the end of a long filament. Enrichment of the curved region of the flagella resulted in the identification of FlaB3 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and N-terminal sequencing, and the localization of this flagellin to the cell-proximal portion of the flagellum was confirmed through immunoblotting and immunoelectron microscopy with FlaB3-specific antibodies, indicating that FlaB3 likely composes the curved portion of the flagella. This could represent a unique case of a flagellin performing the role of the bacterial hook protein. FlaA-specific antibodies were used in immunoblotting to determine that FlaA is found throughout the flagellar filament. M. voltae cells were transformed with a modified flaA gene containing a hemagglutinin (HA) tag introduced into the variable region. Transformants that had replaced the wild-type copy of the flaA gene with the HA-tagged version incorporated the HA-tagged version of FlaA into flagella which appeared normal by electron microscopy.

While they are functionally similar, archaeal flagella have characteristics not typically seen in bacterial flagella (37). Archaeal flagellins are synthesized with a leader peptide (3, 18), which is cleaved prior to the incorporation of the flagellin into the filament, similar to the processing of type IV pilins before incorporation into the pilus (34). A Methanococcus maripaludis protein (FlaK) possessing preflagellin peptidase activity has recently been reported (2). Archaeal flagellins show sequence similarity to type IV pilins at the N termini of the mature proteins (8) and do not demonstrate homology to bacterial flagellins. Archaeal flagella are thinner in diameter (10 to 13 nm [4, 16, 33]) than bacterial flagella (20 nm [17]) and are always composed of multiple flagellins, which are often glycosylated (6, 20, 24). Additionally, a search of completely sequenced archaeal genomes failed to identify genes homologous to any genes coding for structural proteins involved in bacterial flagellation (7). This includes, but is not limited to, genes encoding the hook, rod, or ring proteins. All of these characteristics suggest that the structural components of the archaeal flagella are composed of unique, archaeon-specific proteins, possibly fulfilling the same function as those present in bacterial flagella, and that the mode of assembly is likely distinct as well.

Methanococcus voltae is a marine organism possessing more than 70 flagella on the cell surface. As is typical of archaeal flagella, M. voltae flagella are composed of multiple flagellins (18). There are four flagellin genes found within two transcriptional units in the M. voltae chromosome, with the first transcriptional unit containing a single flagellin gene, flaA. The second transcriptional unit includes the three remaining flagellin genes (flaB1, flaB2, and flaB3) and the downstream cotranscribed flagellar accessory genes flaCDEFGHIJ (18; N. A. Thomas and K. F. Jarrell, unpublished data). Purified flagella were shown to be composed of two major proteins, flagellins FlaB1 and FlaB2, with molecular masses corresponding to 33 and 31 kDa, respectively (18). Prior to the work presented in this study, the remaining two flagellins (FlaA and FlaB3) remained undetected.

Within the flagellated archaea, little work has been done to address the universal presence of multiple flagellins. In Halobacterium salinarum, five flagellin genes are arranged in two different loci. Two flagellin genes (flgA1 and flgA2) are arranged in tandem at one locus, with the remaining three genes (flgB1, flgB2, and flgB3) clustered in another locus, and all five corresponding gene products have been identified within isolated flagella (9). The estimated lengths of the mRNAs indicate that the flagellin genes within each locus are cotranscribed, but the transcripts did not include the accessory genes as seen in M. voltae and other methanogens (10, 38). It was recently determined that the majority of the accessory genes observed in M. voltae are present in H. salinarum next to the flgB locus but are transcribed from a distinct promoter and in the opposite direction (29). Mutant studies with H. salinarum have shown that both loci are required for fully motile cells and that the A flagellins compose the majority of the filament, while the B flagellins make up the cell-proximal portion of the filament (36).

In M. voltae, insertional inactivation of flaA resulted in the production of flagella that appear to be similar to wild-type flagella, although the cells were less motile. The need for this minor flagellin for maximum motility may indicate a specialized role for this flagellin and demonstrates that the four flagellins are not simply interchangeable (14). Insertional inactivation of flaB2 or flaH resulted in nonflagellated cells, although the necessity for specific genes within the transcriptional unit could not be determined due to polar effects on the downstream cotranscribed genes. However these studies indicate that this gene family is crucial for flagellum production in M. voltae (14, 39).

This paper reports the identification of the two remaining flagellins (FlaA and FlaB3) of M. voltae as structural components of the flagella and attempts to address the spatial organization of these flagellins within the flagellum. While FlaA appears to be distributed throughout the filament, FlaB3 appears to localize specifically proximal to the cell surface, perhaps composing the curved hook-like regions next to the basal bodies of isolated intact flagella (flagellar filaments with attached basal structures). If this region is similar in function to the bacterial hook, this would be the first reported case of a flagellin fulfilling this role.

MATERIALS AND METHODS

Bacteria and growth conditions.

M. voltae PS (obtained from G. D. Sprott, National Research Council of Canada, Ottawa, Ontario, Canada) was grown in Balch medium 3 at 37°C under an atmosphere of CO₂-H₂ (20:80) in 1-liter bottles modified to accept serum bottle stoppers as previously described (19). The Balch medium 3 was supplemented with puromycin (7.5 μg/ml) when necessary for growth of mutants. Escherichia coli strains (Table 1) were grown in Luria-Bertani medium (31) supplemented with ampicillin (100 μg/ml) and chloramphenicol (30 μg/ml) when necessary.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Description	Reference or source
Strains
M. voltae PS	Wild type	G. D. Sprott
E. coli
DH5α	Laboratory K-12 strain for in vitro cloning	Novagen
BL21(DE3)/pLysS	Expression host for T7 promoter-based plasmids with pLysS for tight repression	Novagen
Plasmids
pET23a+	Expression vector with T7 promoter and sequence for C-terminal polyhistidine tag	Novagen
pKJ328	pET23a+ carrying PCR-amplified internal region of flaB3 as NdeI/XhoI fragment	This study
pKJ348	pET23a+ carrying flaA-HA from pKJ337 as NdeI/blunt fragment	This study
pKJ350	pET23a+ carrying PCR-amplified internal region of flaA as NdeI/XhoI fragment	This study
pPAC60	Methanococcal integration vector, Ampr	A. Klein
pKJ318	Modified pPAC60 with BglII/EcoRV linker in MCS2	This study
pKJ323	Modified pKJ318 containing M. voltae flaB1 and native promoter as KpnI/SmaI fragment	This study
pKJ337	Modified pKJ323 containing HA-tagged version of flaA in MCS2 as BglII/EcoRV fragment	This study
pKJ42	XL Bluescript carrying 3-kb EcoRI flaA-flaB1 fragment	18
pKJ60	XL Bluescript carrying 3.3-kb PstI fragment from M. voltae flagellin gene region	18

DNA isolations and manipulations.

Chromosomal DNA of M. voltae was isolated according to the method of Gernhardt et al. (11). Plasmid DNA was isolated from E. coli by using Qiaprep miniprep kits (Qiagen, Chatsworth, Calif.). All restriction enzyme digestions and ligations were carried out according to the manufacturer's instructions. DNA samples were separated on 0.8% (wt/vol) agarose gels run in Tris-acetate buffer (31). DNA fragments were purified from agarose gels by using the Prep-A-Gene matrix (Bio-Rad, Hercules, Calif.).

Creation of pKJ337 integration vector.

Plasmid pPAC60 (Table 1; Fig. 1) was modified through the insertion of a double-stranded linker into the second multiple cloning site (MCS2). This linker was created by using two complementary oligonucleotides, 12100 (5′ AGCTTAGATCTGATATCCTGCA 3′) and 12101 (5′ GGATATCAGATCTA 3′) (BglII-EcoRV sites are underlined). An aliquot of the double-stranded linker with HindIII-PstI-compatible ends (in boldface) was ligated into pPAC60 that had been cleaved with HindIII and PstI, and this created plasmid pKJ318 (Fig. 1). pKJ318 was further modified by the insertion of the native flaB1 promoter and the beginning of the flaB1 gene from M. voltae, corresponding to bp 1817 to 2485 (GenBank accession number M72148). This was obtained through a PCR involving primers 12102 (5′ GGGGTACCTAGACCCTACATAACCAGG 3′), including a 5′ KpnI site (underlined), and 12103 (5′ CAGTACTAATTCGAATAGCC 3′), with plasmid pKJ60 (Table 1) as the template. The resulting amplification product was cloned into pKJ318 as a KpnI-blunt fragment. The resulting plasmid (pKJ323 [Fig. 1]) was further modified by the cloning of flaA with an incorporated hemagglutinin (HA) tag (flaA-HA) into the BglII-EcoRV sites in MCS2. The creation of the flaA-HA gene was done according to the PCR strategy outlined in Fig. 1. The first 400 bp of upstream region and the beginning of flaA were linked to the first 15 bp of the HA tag through a PCR using primers 14170 (5′ GAAGATCTCATATGAAAGTAAAAGAGTTTATGAATAAC 3′), incorporating BglII and NdeI restriction sites (underlined), and 14169 (5′ CGGACGTCATAAGGATAAGCTTTTGAACCGGTTACATC 3′). The last 6 bp of this part of the HA tag also encode an AatII site (underlined). The remainder of flaA and some of the downstream region (700 bp) were linked to the last 18 bp of the HA tag, overlapping the previous product by inclusion of the 6 bp encoding the AatII site (underlined), through a PCR using primers 14168 (5′ CGGACGTCCCAGATTATGCATGGAATAACGGTGCAATT 3′) and 14171 (5′ CGGATATCTTTAGGTATATGATTAGG 3′), incorporating an EcoRV site (underlined), with plasmid pKJ42 (Table 1) (18) as the template. The two resulting PCR products were then cleaved with AatII and ligated together, and a final PCR was performed using primers 14170 and 14171, with the ligation product as the template. This final PCR product (flaA-HA) was cloned into the BglII and EcoRV sites in pKJ323, resulting in pKJ337 (Fig. 1).

FIG. 1.

Creation of the methanococcal integration vector pKJ337. The constructs are based on pPAC60, a gift from A. Klein. pKJ318 incorporates a BglII site and an EcoRV site into MCS2. pKJ323 replaces the beginning of MCS1 with the native promoter and beginning of flaB1. pKJ337 includes flaA-HA in MCS2. flaA-HA was created as shown and described in Materials and Methods. P-sl, M. voltae S-layer promoter; T-mcr, methylreductase terminator; pac, puromycin acetylase cassette; P-hmva, M. voltae histone-like promoter.

pKJ337 was digested with BglII and EcoRI, purified with a Qiagen PCR purification column (Qiagen), and resuspended in 20 mM HEPES buffer (pH 7.4) and was used to transform M. voltae, as described below. Puromycin-resistant transformants were screened for incorporation of the HA tag through amplification of flaA with primers 14170 and 14171, using chromosomal DNA as the template, and digestion of the resulting PCR product with AatII.

flaA-HA was excised from pKJ337 as an NdeI/EcoRV fragment and cloned into pET23a+ for overexpression in E. coli BL21(DE3)/pLysS. Induction was done according to the instructions for the Novagen pET system, with IPTG (isopropyl-β-d-thiogalactopyranoside) (Life Technologies, Burlington, Ontario, Canada) added to a final concentration of 0.4 mM.

Southern hybridization.

Chromosomal DNA was isolated from wild-type M. voltae and puromycin-resistant transformants, digested with restriction enzymes, electrophoresed, and transferred to nylon membranes (Boehringer, Mannheim, Germany) by a downward capillary transfer method (31). A digoxigenin-labeled flaA probe was generated through the amplification of flaA by PCR and the incorporation of digoxigenin-UTP by random priming as recommended by the manufacturer (Boehringer). Southern hybridizations were performed as previously described (39).

Transformations.

Transformations into CaCl₂-competent E. coli were done as described by Sambrook and Russell (31). Transformations of M. voltae were done by using a liposome delivery method (25) with modifications as described by Thomas et al. (39).

Overexpression and purification of the variable region of minor flagellins.

The internal variable regions of flaA and flaB3 were amplified and cloned for overexpression. The internal region of flaA (flaAv; bp 668 to 1123; GenBank accession number M72148) was amplified using primers 16094 (5′ GGAATTCCATATGAAGATGTTTCAAACATCCGG 3′), incorporating an NdeI site (underlined), and 16095 (5′ CCGCTCGAGATTATACTCTGGCAAAATTGCACC 3′), incorporating an XhoI site (underlined). The internal region of flaB3 (flaB3v; bp 3602 to 3884; GenBank accession number M72148) was amplified using primers 12342 (5′ GGAATTCCATATGATAACAGGTCACAGTGTTGACC 3′; NdeI site underlined) and 12343 (5′ CCGCTCGAGAGGGTGTTGGTCACTTTGACC 3′; XhoI site underlined). These primer sets allowed directional cloning of the PCR products as NdeI/XhoI fragments into pET23a+, creating plasmids pKJ350 and pKJ328 for expression of the C-terminal His-tagged polypeptides FlaAv and FlaB3v, respectively. Overexpression and purification of the variable regions of the minor flagellins were done under denaturing conditions as previously described (38). Protein concentrations were determined by using the Bradford binding assay (Pierce, Rockford, Ill.) with bovine serum albumin as the standard.

Production of polyclonal antibodies.

Purified polypeptides FlaAv and FlaB3v were used as antigens to raise polyclonal antibodies specific to the minor flagellins. An initial injection containing 250 μg of antigen was delivered subcutaneously along with the adjuvant Quill A (100 μg) (Cedarlane Laboratories Limited, Hornby, Ontario, Canada) into 1-year-old White Leghorn chickens. Boosts of the same amount of antigen were given subcutaneously on days 14, 28, and 42. A final injection was given on day 65, and antibodies were isolated from eggs laid at least 1 week following the final injection. Antibodies were produced by RCH Antibodies (Sydenham, Ontario, Canada).

Isolation of M. voltae flagella.

M. voltae cells (3 liters) were harvested by centrifugation at 5,900 × g for 15 min and resuspended in 10 mM Tris-HCl (pH 7) with 2% NaCl, 0.28% MgCl₂, and 0.35% MgSO₄·7H₂O. The intact cells were lysed through the addition of the nonionic detergent OP-10 (Nikko Chemicals Co. Ltd., Tokyo, Japan) to a final concentration of 1% (vol/vol), with DNase-RNase added to reduce viscosity, and incubated for 30 min at room temperature with occasional inverting. This sample was centrifuged at 5,900 × g for 15 min, and the supernatant was incubated on ice with shaking for 1 h with the addition of a precipitation solution (1 M NaCl, 20% [wt/vol] polyethylene glycol) (40) to a final concentration of 10% (vol/vol). Pellets obtained by centrifugation at 7,800 × g for 10 min were resuspended and placed on a KBr gradient as previously described (15). The flagella formed an isolated band, which was desalted by further centrifugation in distilled water. The relative amount of cell-proximal flagellar proteins was enriched by modifying the above-described protocol; a brief shearing (30 s) of intact cells in a blender (Waring Products Co.) prior to the isolation of the flagella allowed for separate isolation of the flagellar filaments and the remaining cell-proximal portion (flagellar stubs).

SDS-PAGE.

All samples were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (23), and the gels were stained with Coomassie brilliant blue G250-perchloric acid solution and destained in water as previously described (6).

Immunoblotting.

Immunoblotting with chicken antibodies was performed as described previously (38). Peroxidase-linked mouse monoclonal anti-HA antibodies (Roche Molecular Biochemicals, Laval, Quebec, Canada) were used at a dilution of 1:2,000. All immunoblots were developed with a chemiluminescent kit according to the instructions of the manufacturer (Roche Molecular Biochemicals).

N-terminal sequencing.

Following resolution by SDS-PAGE, protein samples were transferred to Immobilon-P as described above. After transfer, the membrane was washed with water and then stained briefly with Biosafe Coomassie brilliant blue G250 stain according to the instructions of the manufacturer (Bio-Rad). Sequencing was done by D. Watson (National Research Council of Canada, Ottawa).

Observations of polymorphic transition of flagella.

Isolated flagella were observed at high intensity with a dark-field microscope (Olympus CH-2) and recorded on a VTR through a CCD camera (Panasonic BL200) attached to the microscope.

Electron microscopy.

Samples were negatively stained with 2% phosphotungstic acid (pH 7.0 or 4.5) and observed with a JEM-1200EXII electron microscope (JEOL, Tokyo, Japan). Micrographs were taken at an accelerating voltage of 80 kV. Intact cells were fixed with 4% glutaraldehyde prior to staining.

RESULTS

Isolation of intact flagella from M. voltae.

An M. voltae cell has more than 70 flagellar filaments extending from a side of the round cell body. This arrangement differs from the polar tufts seen on Spirillum species (26), in that M. voltae flagella do not make a tuft but separately spread out. Since archaeal cells do not have a peptidoglycan layer, addition of detergents resulted in cell lysis, leaving “intact” flagella (i.e., flagella with some attached basal structure). We found the nonionic detergent OP-10 most effectively lysed cells, and thus it was employed to purify flagella thereafter.

Electron microscopic observations of isolated flagella.

Electron microscopy revealed that intact flagella stained with 2% phosphotungstic acid (pH 7.0) retained a curved region at the cell-proximal end of the flagellum (Fig. 2A). The hook-like portion has a sharp curvature characteristic of the bacterial hook, compared to the slow curvature of the filament. This hook-like region was much shorter than the filament region but varied in length to a much greater degree than seen in Salmonella hooks (13). The junction between the hook and filament, which is often obvious in bacterial flagella, was not distinguishable in the M. voltae filaments.

FIG. 2.

Electron micrographs of isolated flagella. (A) Intact flagella isolated from whole cells by using OP-10 detergent. (B) Flagellar filaments isolated from whole cells by shearing in a Waring blender. (C) Flagellum stubs isolated by shearing whole cells in a Waring blender to remove flagellar filaments, followed by OP-10 detergent isolation. Bar, 100 nm.

Polymorphism of M. voltae flagella.

Polymorphism of flagella is the change in helical shape according to solvent conditions (21). Most bacterial flagellar filaments studied so far have shown polymorphism (21, 32). We examined the polymorphism of M. voltae flagella under different pH conditions.

At neutral pH, flagella appeared in a coiled shape as observed in Fig. 3B. At acidic pHs of between 3 and 5 (Fig. 3A), they looked normal (left handed and 1.2 μm in pitch), which could effectively work as a propeller. At pHs of lower than 3, the flagella turned into a straight form and then gradually disappeared, indicating the depolymerization of flagella. At alkaline pHs of between 11 and 13 (Fig. 3C and D), they appeared in a semicoiled form, with a diameter slightly smaller than that of the coiled form. Overall, archaeal flagella are polymorphic in a manner similar to that for bacterial flagella. Although the sequence homology between the two types of flagellins is low, the physical properties and the structural organization of flagella appear to be similar.

FIG. 3.

Polymorphism of M. voltae flagella. Intact flagella were isolated by using the detergent OP-10 and observed with a dark-field microscope under various conditions. (A) pH 3 to 5. (B) pH 5 to 11. (C and D) pH 11 to 13.

Physicochemical properties of flagellins.

Evaluation of the physicochemical properties of flagellar homologs directly from their amino acid sequences (http://www.expasy.ch/tools/protparam.html) has proven to be very useful in analyses of bacterial flagella proteins and those involved in the needle complex of type III secretion (1). In that evaluation, parameters such as molecular size (in number of amino acids), pI (isoelectric point), instability index, aliphatic index, and grand average of hydropathy have been used. It was found that in both flagella and needle complexes, extracellular structures such as the hook and filament are more stable than the cytoplasmic components. Since homologs of the bacterial hook and basal body components have not been identified in any sequenced genome from archaea, it was thought that analysis of the physicochemical properties of the multiple archaeal flagellins (Table 2) might help in the predictions of their functions. The molecular size of archaeal flagellins is a little more than 200 amino acids, which is among the lowest values for either bacterial flagellins or hook proteins (FlgE), suggesting that archaeal flagellins have close to the minimum requirement for filament formation. The isoelectric point is acidic for all four archaeal flagellins, a common feature of bacterial flagellins and hook proteins. Interestingly both FlaA and FlaB2 contain cysteine residues, which is not the case for bacterial flagellins. The instability index of the archaeal flagellins is relatively low, indicating that these flagellins are stable, with FlaB1 being the most stable and FlaA being the least stable. The aliphatic index and grand average of hydropathy values indicate how hydrophobic the protein of interest could be. The range of aliphatic index values for the archaeal flagellins is 88 to 103, while that for bacterial flagellins is 75 to 100 and that for FlgE is 70 to 85. In a comparison of the four flagellins of M. voltae, FlaB3 stands out as being larger and having a more acidic pI.

TABLE 2.

Physicochemical properties of M. voltae flagellins

Flagellin	Molecular size (no. of amino acids)	pI	Preflagellin N terminus	Mature N terminus	Instability index	Aliphatis index	Grand avg of hydropathy
FlaA	219	6.31	MKVKEFMNNKKG	ATGVGTLI	42.91	88.54	−0.075
FlaB1	218	6.11	MNIKEFLSNKKG	ASGIGTLI	27.58	97.94	0.208
FlaB2	216	6.57	MKIKEFMSNKKG	ASGIGTLI	37.59	90.32	0.050
FlaB3	239	5.03	MLKNFMKNKKG	AVGIGTLI	30.45	103.14	0.065
SALTYa FliC	489	4.76			22.83	83.29	−0.414
BACSUb FliC	304	4.97			29.33	91.61	−0.414

^aSALTY, Salmonella enterica serovar Typhimurium flagellin.

^bBACSU, Bacillus subtilis flagellin.

Together the results indicate that archaeal flagellins have physicochemical properties similar to those of bacterial flagellins and hook protein, indicating that the overall structures of these filaments are well conserved to serve as a motility organelle.

Localization of minor flagellin FlaB3 in flagellar filament.

In order to enrich for minor structural proteins present in the anchoring structure or composing the curved region of the flagella, M. voltae cells were sheared in a Waring blender prior to being subjected to the flagellum isolation procedure. This removed the major portion of the filament and enriched for shortened flagellar stubs attached to the cell. Purification of the flagellar stubs resulted in a preparation of short filaments, many with attached hooks (Fig. 2C). Flagellar filaments sheared from the cell surface were concentrated through centrifugation. This sample contained short pieces of filament with few hooks (Fig. 2B). Examination of the sheared flagellar filaments and the isolated flagellum stubs by SDS-PAGE resulted in the detection of a protein band in the flagellum stubs not previously seen in preparations of intact flagella. This previously unseen band migrates slower than the major flagellins. N-terminal sequence analysis of this band resulted in its identification as FlaB3 (Fig. 4A, lane 3) due to the unique presence of valine at position 2. In FlaB1 and FlaB2, the second amino acid is a serine residue, and in the mature FlaA protein, the second amino acid is threonine (Table 2). There is a degree of variability in the shearing process from run to run, which is likely dependent on the concentration of the cells and the amount of time shearing. This variability can been seen as fluctuations in the amount of FlaB3 relative to the amount of the major flagellins in both the flagellar filament fraction and the flagellar stubs in different experiments (data not shown).

FIG. 4.

Localization of the minor flagellins of M. voltae within the flagellum structure. Intact flagella were isolated by using OP-10 detergent. The flagellar filaments were isolated by shearing intact cells in a Waring blender, followed by centrifugation. The cell-proximal portions of the flagella (stubs) were isolated by using OP-10 detergent to lyse intact cells following the shearing processes. Lane descriptions apply to both panels. Molecular masses are given on the left in kilodaltons. (A) Coomassie blue staining after SDS-PAGE of intact flagella (lane 1), flagellar filaments (lane 2), and flagellar stubs (lane 3). The band identified as FlaB3 is indicated with an arrow, and the N-terminal sequence of the mature protein is provided. (B) Immunodetection of flagellin localization patterns within flagella. Intact and fractionated flagella were subjected to immunoblotting with polyclonal antibodies raised against purified recombinant flagellins or regions of the flagellins as described in Materials and Methods. The antibodies used are indicated on the right.

To raise antibodies specific to FlaB3 and avoid cross-reactivity with the other flagellins due to the highly conserved nature of the N termini of archaeal flagellins, a 283-bp section of the internal variable region of flaB3 (bp 3602 to 3884; GenBank accession number M72148) was cloned into pET23a+, resulting in the creation of pKJ328. Overexpression resulted in the production of a polypeptide migrating near the predicted molecular mass of 9.5 kDa (Fig. 5A), with a His₆ tag at the C terminus. This internal fragment of FlaB3 (FlaB3v) was affinity purified and used as the antigen in the creation of FlaB3-specific antibodies. The cross-reactivity seen with antibodies raised with complete flagellins as the antigen is avoided (Fig. 5B). Immunoblots were done with anti-FlaB3v antibody against intact flagella, as well as sheared flagellar filaments and flagellar stubs that were isolated following shearing. All flagellar samples were loaded with similar amounts of the major flagellins, as determined by Coomassie blue staining (Fig. 4A). There appeared to be a greater amount of FlaB3 in the flagellar stubs relative to the amount of major flagellins (Fig. 4B), which suggests that FlaB3 localizes proximal to the cell surface, in support of the data for filaments and flagellar stubs obtained by Coomassie blue staining after SDS-PAGE.

FIG. 5.

Creation of antibodies specific to the unique internal variable regions of two M. voltae flagellins, FlaA and FlaB3. Numbers on the left indicate molecular masses in kilodaltons. (A) Overexpression of His-tagged internal variable regions of FlaA (lane 2) and FlaB3 (lane 3) from a T7-based expression system. E. coli cultures carrying the internal variable region of the corresponding gene in pET23a+ were grown to an optical density at 600 nm of 0.6 and then induced with 0.4 mM IPTG. Lane 1, E. coli carrying the vector as a background control. (B) The specificity of flagellin-specific antibodies is seen through immunoblotting. Intact flagella were isolated by using OP-10 detergent and probed in immunoblots with antibodies raised to either overexpressed intact flagellin FlaB2 (lane 2) or the overexpressed internal variable region of FlaA (lane 1) or FlaB3 (lane 3). Equivalent amounts of intact flagella were visualized by Coomassie blue staining (lane 4).

Immunoelectron microscopy.

In order to identify the hook portion as being composed of flagellin B3, we attempted immunoelectron microscopy with anti-FlaB3v antibodies. With a high concentration of polyclonal antibodies, bacterial flagellar structures make aggregates as seen with the hook (28) and with FlgD (27). Although we have not seen any such large aggregates, occasional side-by-side aggregation of the hook portions was observed (Fig. 6), indicating that the antibody weakly reacted with the surface of this portion.

FIG. 6.

Immunoelectron microscopy of M. voltae flagella with anti-FlaB3v antibody. Intact flagella were incubated with anti-FlaB3v antibody and negatively stained with 2% (wt/vol) phosphotungstic acid. Bar, 100 nm.

Localization of flagellin A.

In order to detect the remaining flagellin, FlaA, two distinct approaches were used. The first approach involved the insertion of a 9-amino-acid HA tag (YPYDVPDYA) into the variable region of FlaA. The epitope tagging of FlaA with the HA tag presents a new way to detect archaeal proteins in vivo. This tagging allows for the use of commercial antibodies to avoid cross-reactivity with other similar proteins and has been successful in labeling of eukaryotic proteins (35). In this instance the HA tag was inserted into the internal variable region of the protein. It is thought that the conserved and hydrophobic amino and carboxy termini of the flagellins are involved in subunit interactions, similar to the case for bacterial flagellins (17). As such, insertion of the HA epitope was targeted to the internal region between amino acids 132 and 133 in the hope of avoiding interference with subunit interaction and flagellum function.

Following creation of pKJ337, flaA-HA was excised and inserted into pET23a+, creating plasmid pKJ348. Overexpression of FlaA-HA by E. coli BL21(DE3)/pLysS carrying pKJ348 was induced, and whole-cell lysates were used in immunoblotting to ensure that the FlaA-HA protein was detectable using anti-HA antibodies. FlaA-HA was detectable only in lysates of induced cells and appeared to migrate at 29 kDa as determined by immunoblotting, which is larger than the predicted molecular mass of 24 kDa for FlaA (Fig. 7A).

FIG. 7.

Incorporation of the HA tag into FlaA. (A) Immunoblots with anti-HA antibodies. Lane 1, whole-cell lysates of uninduced E. coli Bl21(DE3)/pLysS carrying pKJ348; lane 2, whole-cell lysates of induced E. coli BL21(DE3)/pLysS carrying pKJ348; lane 3, OP-10-isolated intact flagella from wild-type M. voltae; lane 4, OP-10-isolated intact flagella from M. voltae expressing FlaA-HA. Numbers on the left indicate molecular masses in kilodaltons. (B) Identification of successful incorporation of flaA-HA by PCR and AatII digestion. Lane 1, 100-bp ladder; lane2, flaA-HA positive control; lane 3, flaA PCR product from wild-type M. voltae digested with AatII; lane 4, flaA PCR product from puromycin-resistant M. voltae transformant digested with AatII.

The 3-kb linear fragment (BglII-EcoRI) of pKJ337 containing flaA-HA and the beginning of flaB1 was transformed into M. voltae. A control transformation lacking the linear fragment of pKJ337 did not result in any puromycin-resistant transformants. Puromycin-resistant transformants of pKJ337 were screened for the replacement of the wild-type flaA gene with the HA-tagged version of flaA. Successful incorporation of the linear fragment of pKJ337 containing the flaA-HA gene occurred through two homologous recombination events, with the first occurring within the flaA gene, upstream of the HA tag, and the second occurring within the flaB1 promoter or the beginning of the flaB1 gene in order to retain the puromycin transacetylase cassette (pac) and confer puromycin resistance to the successful transformants.

Chromosomal DNA was isolated from puromycin-resistant transformants and used as the template in a PCR that resulted in the amplification of the flaA gene, with the resulting PCR product digested by AatII. Since the incorporation of the HA tag sequence results in the addition of an AatII restriction enzyme site that is not present in wild-type flaA, cleavage of the PCR product indicates successful incorporation of the HA tag. The PCR products from two of five puromycin-resistant transformants screened were cleaved by AatII, while the remaining transformants retained the wild-type form of flaA, indicating that the first recombination event occurred within flaA but downstream of the HA tag (Fig. 7B). Southern blotting was used to verify that the linear fragment of pKJ337 had incorporated only once in the chromosome and had incorporated into the correct location, resulting in only one copy of FlaA in the chromosome (Fig. 8). Northern blotting demonstrated that the incorporation of the HA-tagged version of flaA did not affect transcription of flaB1 or any of the downstream genes (data not shown).

FIG. 8.

Confirmation of the insertion of the 3-kb linear fragment of pKJ337 into the chromosomal copy of flaA. (A) Southern blot analysis of TaqI-digested genomic DNAs of wild-type M. voltae and a FlaA-HA-expressing strain. A flaA probe was used to determine whether the wild-type flaA had been replaced by the linear fragment of pKJ337 containing flaA-HA and the pac cassette. Lane 1, λ HindIII ladder (numbers indicate base pairs); lane 2, TaqI-digested wild-type M. voltae chromosomal DNA; lane 3, TaqI-digested FlaA-HA M. voltae chromosomal DNA. (B) Schematic diagram depicting wild-type M. voltae chromosomal DNA and FlaA-HA M. voltae chromosomal DNA following the incorporation of the linear fragment of pKJ337 through two homologous recombination events. The restriction maps are not drawn to scale. The bars below the restriction maps correspond to the DNA fragments (sizes in kilobases) identified with the flaA probe in the Southern blot experiment shown in panel A.

HA-tagged FlaA mutant.

The motility of the HA-tagged mutants was confirmed through light microscopy, and electron microscopy demonstrated that the number and length of the flagella containing the FlaA-HA flagellin appeared to be similar to those of the wild-type flagella (Fig. 9). Immunoblots of intact flagella isolated with OP-10 resulted in the detection of FlaA-HA, which was not detected in the OP-10-isolated wild-type flagella (Fig. 7A).

FIG. 9.

Electron micrographs of M. voltae wild-type (top) and FlaA-HA-expressing (bottom) cells. Cells were fixed in 4% glutaraldehyde prior to being negatively stained with 2% phosphotungstic acid (pH 7.0). Bar, 1 μm.

Attempts at shearing the flagella containing FlaA-HA were unsuccessful. Despite the facts that the flagella appeared to be normal by electron microscopy and that the cells appeared to have normal motility when visualized by light microscopy, the shearing of the cells prior to flagellum isolation resulted in difficulties in isolation of the cell-proximal fraction of the flagella. While use of the commercial HA-specific antibodies proved to be effective in detecting the HA-tagged FlaA in Western blots, immunoelectron microscopy with this antibody proved inconclusive.

The second approach for detection of FlaA in fractionated flagella involved overexpression and purification of the internal variable region of FlaA (Fig. 5A) in a manner similar to that for FlaB3v. This allowed for the creation of specific anti-FlaA antibodies, which showed a lack of cross-reactivity with the other flagellins (Fig. 5B). Immunoblotting of intact flagella, flagellar filaments, and cell-proximal stubs with anti-FlaAv Ab suggested that FlaA was present throughout the flagellum structure (Fig. 4B).

DISCUSSION

There are many differences between archaeal flagella and bacterial flagella, with one of these being the universal presence of multiple flagellins within archaeal flagella. The requirement for multiple flagellins and their spatial organization remain largely unaddressed. While multiple flagellins are occasionally found in bacterial flagella, this is not typical. When more than one flagellin is present, the spatial organization of the bacterial flagellins is varied. Helicobacter pylori, Caulobacter crescentus, and Rhizobium meliloti each have multiple flagellins organized into spatially distinct regions of the filament (5, 22, 30). The opposite is found in Campylobacter coli, which has both flagellins intertwined along the entire length of the filament. While both flagellins are not required for motility, they are both required for fully active flagella (12).

Within the domain Archaea, spatial organization of the multiple flagellins has previously been examined only in H. salinarum. Insertional inactivation of both the flgA and flgB loci resulted in cells that were unable to form fully functional flagella (36). Inactivation of the flgB locus resulted in the formation of flagella that appeared to be similar to the wild type but were significantly impaired in motility. This phenotype lead to the proposal that the A flagellins are the major components of the filament and the B flagellins may be parts of the filament proximal to the basal body, such as terminators, anchors, or hook-associated proteins, and that this would explain the decrease in motility (36). Inactivation of the flgA loci resulted in the production of short curved filaments, distributed over the cell surface instead of at the normal polar location. These cells were also determined to be less motile. Additionally, inactivation of the flgA2 gene resulted in the production of straight flagellar filaments. The multicomponent nature of the flagella of H. salinarum is explained by the need for both A flagellins for spiral flagella and involvement of the B flagellins proximal to the cell surface, involved in full function (36).

The isolation of M. voltae flagella by extraction of membranes with OP-10 detergent and the subsequent staining with 2% phosphotungstic acid (pH 7.0) led to the observation of curved regions on a majority of the filament ends, which superficially resemble bacterial hooks. The ability to visualize this curved region led to shearing of the filaments from the cells to enrich for cell proximal proteins, in order to identify the protein composing this hook-like region. The shearing and immunoblotting data indicated that FlaB3 localizes proximal to the cell surface relative to the amount of the major flagellins FlaB1 and FlaB2 and the minor flagellin FlaA. While the immunoblotting results do indicate the presence of a small amount of FlaB3 in the flagellar filaments, this is likely due to the variable nature of the shearing.

The immunoblotting results suggest that FlaA is a minor component distributed throughout the entire flagella. An insertional deletion of flaA in M. voltae has previously been created (14). The resulting strain possesses flagella that look similar to wild-type flagella, as determined by electron microscopy. Motility studies determined that while the flaA mutant was still motile, the flagella were functioning below wild-type levels (14). These initial results correspond with the results seen in H. salinarum in that the minor flagellins can be deleted while still resulting in flagella that appear superficially similar to the wild type yet have reduced function.

While the labeling of FlaA with the HA tag did not aid in immunoelectron microscopy studies, it was successful for immunoblotting detection of FlaA-HA within intact flagella. While epitope tagging has previously been proven to be successful for tracking intracellular localization of proteins (35), this is the first reported case of epitope tagging in vivo in archaea. Additionally, the incorporation of the HA tag within FlaA did not interfere either with the assembly of the flagella or with the motility. This gives indirect support to the theory that it may be the N terminus that is involved in polymerization of the flagellin subunits (37) and also identified a region of the flagellin where other insertions or deletions may be attempted without affecting assembly.

As the localization data suggest that FlaB3 localizes cell proximal, it is possible that this flagellin composes the curved hook-like region. There is a lack of homologs in sequenced genomes of archaea to the hook protein or hook-associated proteins of bacterial flagella, leaving the protein composing this curved region unidentified. That FlaB3 was the only protein experiencing a noticeable increase in relative amount following removal of the majority of the flagellar filament does suggest that it may comprise the majority of the curved region. This lack of a hook protein distinct from flagellins, as observed in bacteria, may explain two observations reported here for M. voltae flagella. One is the lack of a distinct boundary between the hook and filament, as seen with hook-associated proteins in bacterial flagella. The other is the complete dissociation of the archaeal flagella by acid dissociation under conditions where the hook remains intact in bacterial flagella (data not shown).

Furthermore, Cruden et al. (4) cite a great variability in the length of the curved region in archaeal flagella (72 to 265 nm), which is not seen in bacterial flagella (13). If the curved region is made up of flagellins, there may not be a distinct boundary to this region as seen in bacterial flagella, where the HAP proteins separate the hook proteins from the flagellins, but instead a gradient of minor flagellins to major flagellins as one moves along the flagellum, away from the cell. An alternative explanation is that the use of archaeal flagellins in this region may negate the need for tight regulation of the hook length, as seen in bacterial flagella (13).

This is the first reported identification of the minor flagellins of M. voltae within the flagella. These minor flagellins are transcribed at lower levels and may have specialized roles within the flagellum ultrastructure. This corresponds to the results seen by Tarasov et al. (36), where the minor flagellins, expressed from the B loci, are required for wild-type levels of motility. It is unclear if the curved region in archaeal flagella fulfills the same function as the hook does in bacterial flagella, i.e., the transfer of torque from the motor to the filament. If this is the case, this would be the first reported occurrence of a flagellin functioning in such a manner.