INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Flagellation is widespread among members of the domain Archaea, and it appears to be a unique motility system distinct from that of the other major prokaryotic lineage, the Bacteria (18). Significantly, no homologs of bacterial flagellins or any bacterial flagellum-related structural proteins (hook, rods, rings, hook-associated proteins, etc.) have been found in the sequences of the complete genomes of the flagellated archaea Methanococcus jannaschii (8), Archaeoglobus fulgidus (27), and Pyrococcus horikoshii (24). Flagellation in the archaea has best been studied in the mesophilic methanogen Methanococcus voltae (3, 17, 21, 22). M. voltae has been shown to possess four highly similar flagellin genes (flaA, flaB1, flaB2, and flaB3 [21]) which are followed immediately by a number of downstream open reading frames at least some of which have been shown to be cotranscribed with the flagellins (21, 47). Homologs of these downstream genes have been found in the immediate vicinity of the flagellin genes in the complete genome sequences of A. fulgidus (27), P. horikoshii (24), and M. jannaschii (8). None of these genes are found in the complete genome of the nonflagellated methanogen Methanobacterium thermoautotrophicum (38).

While four flagellin genes are found in M. voltae and all are transcribed to various degrees (21), only two flagellin bands are detected by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of KBr gradient-purified flagellar filaments (22). Evidence provided by mutational (17), transcriptional (21), and N-terminal (23) analyses have implicated the product of flaB2 as one of the major flagellins. These flagellins migrate in SDS-PAGE as proteins of approximately 31 and 33 kDa. This size is approximately 10 kDa larger than predicted from the gene sequences (FlaB2 is predicted to be 22.8 kDa [21]) and suggests a posttranslational modification such as glycosylation, as found in numerous other archaeal flagellins (4, 39, 50). While no posttranslational modifications have been described for the flagellins of M. voltae and the flagellins do not stain positively for glycoprotein by the periodic acid-Schiff or thymol sulfuric acid staining procedure (18), the deduced flagellin amino acid sequences do possess a number of the carbohydrate attachment consensus sequences Asn-X-Ser/Thr (21), which suggests that the flagellins may have a type of carbohydrate attachment that resists detection by conventional gel staining methods. Evidence from mutant studies suggests that the fully unmodified form of M. voltae flagellin may migrate at approximately 20 kDa (17). We therefore anticipated the migration of flagellin FlaB2 to be closer to that expected for unmodified flagellins when expressed in heterologous hosts in which flagellin modification is less likely to occur.

On an amino acid level, the flagellins of the archaea show no homology to bacterial flagellins but instead are highly similar to bacterial type IV pilins in their N termini (13). Like type IV pilins, archaeal flagellins are translated with short leader peptides which are cleaved before the flagellins are incorporated into the flagellum filament (1, 21). Here, we describe experiments which generate a source of M. voltae FlaB2 preflagellin in the bacterial host Escherichia coli, in which the leader peptide remained uncleaved. Also, we have expressed the flagellin gene in Pseudomonas aeruginosa, which, unlike most E. coli strains, possesses a type IV pilin system (43). Expression in this system allowed us to test the possibility that archaeal flagellin is recognized as a substrate by the type IV pilus system.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Microbial strains 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 in 1-liter bottles modified to accept serum bottle stoppers as previously described (22). Cultures were shaken at 100 rpm under an atmosphere of CO₂-H₂ (20%-80%) for 36 to 48 h. E. coli strains (Table 1) were grown at 37°C in Luria-Bertani (LB) medium (36) supplemented with ampicillin (100 µg/ml; Sigma Chemical Co., Mississauga, Ontario, Canada) when necessary. P. aeruginosa strains (Table 1) were grown at 37°C in tryptic soy broth (Difco Laboratories, Detroit, Mich.) supplemented with carbenicillin (200 µg/ml; Sigma Chemical Co., St. Louis, Mo.) and tetracycline (200 µg/ml; Sigma) when necessary.

Molecular techniques. PCR was performed on a MiniCycler (MJ Research, Watertown, Mass.), using either Pwo DNA polymerase (Boehringer Mannheim) or Taq DNA polymerase and reagents from Life Technologies, Gibco-BRL (Burlington, Ontario, Canada) under the following conditions: 95°C for 5 min; 30 cycles of 95°C for 1 min, 55°C for 30 sec, and 72°C for 1 min; and a final cycle of 72°C for 5 min. For amplification of the M. voltae flagellin gene flaB2, plasmid pKJ41 with a 1.3-kb EcoRI fragment carrying all but the extreme 5' 15 bp (21) was used as the template for amplification by primers designed to reconstruct the 5' end of the gene and to generate NdeI sites at the start codon and downstream of the stop codon [(forward, flaB2(N)for (5'GGAATTCCATATGAAAATAAAAGAATTCATGAGTAACAAAAAAGGTGC); reverse, flaB2(N)rev (5'GGAATTCCATATGTCACTATTGTAATTGAACTACTTTTGAATCG)]. In addition to the TAG stop codon, a TGA stop codon was incorporated prior to the NdeI site to ensure proper translation termination. The PCR-reconstructed flagellin gene was cloned into NdeI-digested, dephosphorylated pT7-7, which was then transformed into CaCl₂-competent E. coli DH5. pT7-7 with the insert in the correct orientation with respect to the T7 promoter was purified and then digested with XbaI. A 725-bp fragment containing flaB2 and a ribosome binding site eight nucleotides upstream from the start codon (Fig. 1) was gel purified with Prep-A-Gene (Bio-Rad Laboratories, Mississauga, Ontario, Canada) and cloned into XbaI-digested pT7-7 as described above. Cells harboring pT7-7 with the insert in the opposite direction with respect to the T7 promoter were further used as a negative control. The XbaI fragment was also cloned into pUCP18 and transformed into P. aeruginosa strains rendered competent by three successive washes in ice-cold 0.15 M MgCl₂ (35). Plasmids and host strains for induction in P. aeruginosa were kindly provided by John S. Mattick, University of Queensland, Brisbane, Queensland, Australia.

View larger version (92K): [in this window] [in a new window]   FIG. 1.   Specificity of polyclonal antibody raised in chickens against the FlaB2 flagellin of M. voltae. Lanes: 1 and 3, wild-type M. voltae whole cells; 2 and 4, M. voltae P-2 (17), a nonflagellated mutant in which the flagellin gene family has been disrupted by insertional mutagenesis; M, Bio-Rad prestained SDS-PAGE standards (low range): phosphorylase b (103 kDa), bovine serum albumin (81 kDa), ovalbumin (47.7 kDa), carbonic anhydrase (34.6 kDa), soybean trypsin inhibitor (28.3 kDa), and lysozyme (19.2 kDa). Lanes 1 and 2 are Coomassie blue stained; lanes 3 and 4 are immunoblots. For the immunoblot, primary antibody (chicken immunoglobulin Y) was used at a dilution of 1:5,000. Secondary antibody was used at a dilution of 1:50,000. Immunoblot lanes contained 1/100 of protein loaded into Coomassie blue-stained lanes. Note the strong, specific reaction with the flagellins in the wild-type cells and lack of reaction against the total cell protein of the flagellin mutant.

Production of polyclonal antibodies. FlaB2 was expressed in E. coli DH5 and separated by SDS-PAGE. The band corresponding to FlaB2 was excised. A 1-ml volume containing macerated acrylamide pieces containing FlaB2 (approximately 500 µg) and the adjuvant Quill A (100 µg; Cedarlane Laboratories Limited, Hornby, Ontario, Canada) was injected subcutaneously into 1-year-old white leghorn chickens. Boosts of the same mixture of FlaB2 and adjuvant were given subcutaneously on days 10, 20, and 34. A final injection was given on day 65, and antibodies were isolated from the eggs laid at least 1 week following the final injection. Antibodies were produced by RCH Antibodies, Sydenham, Ontario, Canada.

Induction of M. voltae flagellin synthesis. E. coli and P. aeruginosa were grown overnight as described above, then inoculated at 1% (vol/vol) into 25 ml of fresh medium, and grown to an optical density at 600 nm (OD₆₀₀) of 0.5 to 0.6. Isopropylthio--D-galactoside (IPTG; Life Technologies) was added to 2 mM, and growth was allowed to continue. For growth analyses, 0.3-ml samples were removed at 30-min intervals for spectrophotometric analysis. For detection of induction products, whole cells were prepared for SDS-PAGE as described by Laemmli (30). Samples were electrophoresed at 150 V for 1.5 h (Hoefer Mighty Small; Hoefer Scientific, San Francisco, Calif.). Bands were then visualized by staining with Coomassie brilliant blue G-250, followed by destaining in water in a microwave oven (12) or by immunoblotting (48) using a polyclonal antibody raised in chickens against FlaB2 flagellin and shown to be specific (Fig. 1) as a primary antibody. The secondary antibody was a horseradish peroxidase-conjugated rabbit anti-chicken (Jackson Immunoresearch Laboratories, West Grove, Pa.) used at a 1:50,000 dilution. Blots were developed with a chemiluminescence kit from Boehringer Mannheim Canada (Laval, Quebec, Canada) according to the manufacturer's instructions.

In vivo and in vitro labeling of flagellin. The procedure followed for specific labeling of plasmid encoded proteins was similar to that of Watson et al. (49). E. coli strains were inoculated at 1% (vol/vol) into 25 ml of fresh medium as described above. At an OD₆₀₀ of between 0.5 and 0.6, 0.2-ml aliquots of cells were harvested at 6,000 × g for 5 min and then washed in 1.5 ml of M9 salts (36). Cells were centrifuged and resuspended in 1 ml of M9 salts supplemented with 0.5% methionine assay medium (Difco) in a plastic microcentrifuge tube and incubated at 37°C for 1 h. All subsequent incubations were at 37°C unless specified. Cells were induced by addition of IPTG to 2 mM and incubated a further 30 min. A duplicate culture of E. coli harboring pKJ91 was also processed without induction. Rifampin (Boehringer Mannheim) was then added to 200 µg/ml, and incubation continued for a further 30 min. Then 10 µCi of [³⁵S]methionine (Redivue [³⁵S]methionine; >1,000 Ci/mmol; Amersham Life Science, Inc., Oakville, Ontario, Canada) was added, and incubation continued for 10 min. Cells were harvested as described above, resuspended in 200 µl (total volume) of SDS-PAGE sample buffer, heated to 100°C for 5 min, and electrophoresed as described above. Gels were then fixed in 7% acetic acid at 4°C overnight, incubated for 30 min in En³Hance (DuPont, NEN Research Products, Boston, Mass.), dried at 80°C for 1 h in a model 543 gel drier (Bio-Rad), and exposed to Kodak X-Omat AR film (Fisher Scientific).

Cell fractionation. E. coli KJ91 was grown at 37°C overnight in LB containing 100 µg of ampicillin per ml and subcultured at 1% (vol/vol) inoculum into medium of the same composition. When the culture reached an OD₆₀₀ of 0.6, IPTG was added to a final concentration of 2 mM to induce synthesis of flagellin. After a further incubation of 2.5 h, the cells were pelleted and washed in 100 mM Tris-HCl (pH 6.8) containing 10% (wt/vol) sucrose. The cell pellet was resuspended in 100 mM Tris-HCl (pH 6.8) containing 18% (wt/vol) sucrose, 5 mM EDTA, and 100 µg of lysozyme per ml and incubated at room temperature for 30 min. The cells were collected at 8,000 × g/10 min, and the pellet was resuspended in 100 mM Tris-HCl (pH 6.8) containing DNase and RNase. The viscous solution was sonicated on ice twice for 30 s each time. Unbroken cells were removed by centrifugation at 5,000 × g/10 min. The resulting supernatant was centrifuged at 20,000 × g/30 min to collect the crude particulate fraction. The supernatant was centrifuged at 80,000 × g/90 min, and the resulting supernatant from this spin was considered the cytoplasm fraction. The crude particulate fraction was layered over a preformed step sucrose gradient as described by Sprott et al. (40). The membrane fraction was collected by syringe and washed once with distilled water. Material which pelleted through the sucrose gradient was also collected and washed once with distilled water.

Preflagellin peptidase assay. The assay for preflagellin peptidase activity in M. voltae was based on the prepilin peptidase assay developed for P. aeruginosa (44). The substrate for the assay was the crude particulate fraction of induced KJ91 (overproducing M. voltae FlaB2 with leader peptide) isolated as described above. The source of preflagellin peptidase was a M. voltae membrane fraction. M. voltae cells were grown for 18 h and harvested in a microcentrifuge. The pellet from a total of 6 ml of cells was resuspended in 100 µl of growth medium, and then the cells were lysed by the addition of 1 ml of distilled water. The resulting envelopes were collected by centrifugation for 10 min in an Eppendorf centrifuge. This pellet was resuspended in 100 µl of distilled water and used as the source of preflagellin peptidase. All isolations and assays were performed aerobically. The assay conditions were 6 µl of E. coli KJ91 membranes (12 µg of protein), 1 µl of M. voltae membranes (3 µg of protein), 1 µl of 0.5% cardiolipid (Sigma), and 2 µl of reaction buffer (5× reaction buffer was 125 mM HEPES buffer [pH 7.5] containing 2.5% Triton X-100). The reaction was started with the addition of the M. voltae fraction. At each time point, the 10-µl samples were mixed with 15 µl of electrophoresis sample buffer and immediately boiled for 5 min. For detection of peptidase activity, 2 µl of each sample was examined by Western blotting using a 1/10,000 dilution of the chicken polyclonal antibody raised to the M. voltae flagellin as primary antibody.

N-terminal sequence analysis. Proteins to be sequenced were resolved by SDS-PAGE as described above and transferred onto Immobilon P (Millipore, Bedford, Mass.) as previously described (48). The membrane was briefly stained with 0.1% Coomassie blue R250, destained in 50% methanol, and rinsed thoroughly with distilled water. Protein bands were sequenced by David Watson, National Research Council of Canada.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Induction of E. coli strains. E. coli BL21(DE3)/pLysS as well as the same strain carrying pT7-7, pKJ91, or pKJ139 were grown to log phase, induced, and analyzed by both optical density and light microscopy (data not shown). The strain carrying pKJ91(pT7-7 with the flaB2 gene in the same orientation as the T7 promoter) began to lyse approximately 1 h postinduction, as determined both by a decrease in OD₆₀₀ and by light microscopy, which revealed spheroplasts and cell debris. When the same strain was processed without induction, no lysis was observed. Induction of the control strains did not result in cell lysis. SDS-PAGE analysis revealed an induction product from the strain carrying pKJ91 migrating at approximately 26.5 kDa detected by both Coomassie brilliant blue staining and immunoblotting with antisera raised to the FlaB2 flagellin of M. voltae (Fig. 2). This induction product was detected only in E. coli harboring pT7-7 with flaB2 in the same orientation as the T7 promoter. Also, the 26.5-kDa band was not detected when this strain was examined uninduced. N-terminal sequencing of this induced protein revealed the sequence MKIKEFMSNK, which indicated that it was the product of flaB2 and that the short leader peptide recognized by M. voltae was not cleaved from the protein by E. coli.

View larger version (80K): [in this window] [in a new window]   FIG. 2.   Inducible synthesis of flaB2 in E. coli BL21(DE3)/pLysS. Lanes: 1 and 5, cells with no added plasmid, induced; 2 and 6, cells carrying pKJ139, induced; 3 and 7, cells carrying pKJ91, induced; 4 and 8, cells carrying pKJ91, uninduced; M, molecular weight standards. Lanes 1 through 4 are Coomassie blue stained; lanes 5 through 8 are immunoblots. Immunoblot lanes contain 1/100 of the protein loaded into the Coomassie blue-stained lanes.

View larger version (88K): [in this window] [in a new window]   FIG. 3.   Localization of FlaB2 in fractionated E. coli cells carrying pKJ91, 2.5 h after induction with IPTG. Lanes: 1 and 5, whole-cell lysate; 2 and 6, cytoplasmic fraction; 3 and 7, membrane fraction from sucrose gradient; 4 and 8, pellet from sucrose gradient; M, molecular weight standards. Lanes 1 through 4 are Coomassie blue stained; lanes 5 through 8 are immunoblots generated by using a primary antibody dilution of 1:5,000. The immunoblots contained 1/100 of the protein used for the Coomassie blue-stained gel.

Preflagellin peptidase activity. One of the potential uses for the overexpressed archaeal preflagellin is as substrate for the detection and purification of archaeal preflagellin peptidase. A simple assay system, based on that developed for detection of prepilin activity, involving mixing of membranes of E. coli overproducing the preflagellin, with M. voltae membranes, in the presence of cardiolipid and a buffer containing Triton X-100, clearly showed the effectiveness of the overexpressed preflagellin in detecting archaeal preflagellin peptidase activity. Western blot analysis of the preflagellin peptidase assay samples (Fig. 4) demonstrated the appearance, with time, of an additional cross-reactive band that migrated more quickly than the preflagellin in SDS-PAGE. This additional band increased in intensity with time, and its size corresponded to that expected for the processed flagellin. The N-terminal sequence determined for this smaller, cross-reactive band was Ala-Ser-Gly-Ile-Gly-Thr-Leu/Gly-Ile-Val-Phe, indicating that it was indeed the product of FlaB2 after cleavage of the 12-amino-acid leader sequence.

View larger version (24K): [in this window] [in a new window]   FIG. 4.   Demonstration of preflagellin peptidase activity in M. voltae membranes. The reaction tubes contained 6 µl of crude membranes from induced E. coli KJ91, 1 µl of M. voltae membranes, 2 µl of 125 mM HEPES buffer (pH 7.5) containing 2.5% Triton X-100, and 1 µl of 0.5% cardiolipid. The reaction was started by the addition of M. voltae membranes. Samples were removed at 0, 2, 10, 20, and 30 min and then immediately mixed with 15 µl of electrophoresis sample buffer and boiled for 5 min; 2-µl aliquots were examined by immunoblotting using a primary antibody dilution of 1:10,000. Lanes: 1, 0 min; 2, 2 min; 3, 10 min; 4, 20 min, 5, 30 min. Positions of 28.3- and 19.2-kDa markers indicated in lane M.

Induction of P. aeruginosa strains. Due to the presence of short leader peptides in both the type IV pilin superfamily and the flagellins of M. voltae, as well as the sequence similarity of archaeal flagellins and type IV pilins at their N termini (13), we attempted to express the flaB2 gene in a host which possessed a type IV pilin system. This would allow us to determine if the type IV prepilin peptidase could process the methanococcal preflagellin and to see if archaeal flagellin could substitute for type IV pilin and be assembled into a surface structure. To this end, native FlaB2 as well as two P. aeruginosa pilin-M. voltae FlaB2 fusions were tested (Fig. 5).

View larger version (12K): [in this window] [in a new window]   FIG. 5.   Alignment of N-terminal sequences of the M. voltae FlaB2 flagellin, the P. aeruginosa PilA pilin, and the products of the two fusion constructs with their leader peptides intact. The position of leader peptide cleavage shown represents that known for the M. voltae and P. aeruginosa proteins and predicted for the fusions. Shown in boldface are amino acid residues of interest in the fusion proteins: the +1 position change of the mature M. voltae FlaB2 to a phenylalanine from an alanine in both fusions, and the change in the +5 position of the mature FlaB2 from an glycine to a glutamic acid in the fusion encoded by pKJ248.

View larger version (47K): [in this window] [in a new window]   FIG. 6.   Expression of flaB2 and the pilin-flagellin fusion product in P. aeruginosa wild-type and prepilin peptidase mutant cells. Lanes: V, M. voltae; M, molecular weight markers; 1 and 8, P. aeruginosa PAK; 2 and 9, P. aeruginosa PAK pilD mutant; lanes 3 and 10, P. aeruginosa PAK carrying pUCP18; 4 and 11, P. aeruginosa PAK carrying pKJ119; 5 and 12, P. aeruginosa PAK carrying pKJ118; lanes 6 and 13, P. aeruginosa PAK pilD mutant carrying pKJ156; 7 and 14 P. aeruginosa PAK carrying pKJ156. Lanes 1 to 7 are Coomassie blue stained; lanes 8 to 14 are immunoblots generated by using a primary antibody dilution of 1:10,000. Immunoblots contained 1/10 the protein used for the Coomassie blue-stained gel.

View larger version (46K): [in this window] [in a new window]   FIG. 7.   Expression of flaB2 and the pilin-flagellin fusion product with the additional position +5 change to glutamic acid in P. aeruginosa wild-type and prepilin peptidase mutant cells. Lanes: M, molecular weight markers; 1, wild-type P. aeruginosa PAK cells; 2, P. aeruginosa PAK pilD mutant cells; 3, PAK carrying pUCP19; 4, PAK pilD mutant carrying pUCP19; 5, PAK carrying pKJ118; 6, PAK pilD mutant carrying pKJ248; 7, PAK carrying pKJ248; 8, PAK carrying pKJ156; V, M. voltae cells. A primary antibody dilution of 1:5,000 was used in the blots.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The purpose of the experiments described in this report was twofold. First, we wished to develop a reliable source of archaeal flagellin which still had the signal sequence attached. This was shown to be a substrate for the M. voltae preflagellin peptidase and will be useful for further analysis of this enzyme. Second, we wished to design experiments which would test the possibility of the archaeal flagellin being treated as a type IV pilin in heterologous expression studies.

We have expressed the M. voltae native flaB2 flagellin gene in both E. coli and P. aeruginosa. Much better production of FlaB2 was obtained in E. coli than in P. aeruginosa (compare Fig. 3, lane 1, to Fig. 6, lane 5), probably due to the codon usage. M. voltae with a G+C content of 31% utilizes a number of codons that are rarely used in E. coli. Recently, it was observed that poor expression of M. jannaschii proteins in E. coli could be overcome by using E. coli strains carrying a plasmid expressing three rare tRNA species (26). In P. aeruginosa, with an even higher G+C content than E. coli, this codon usage could be expected to be even more problematic.

Sequencing of the N terminus of the expressed protein confirmed that it was the product of the M. voltae flaB2 gene and that the leader peptide was not cleaved by E. coli, and presumably also not by P. aeruginosa. Surprisingly, the overexpressed flagellin migrated at 26.5 kDa, 5 to 7 kDa smaller than the flagellins isolated from purified M. voltae flagella (22). There is much evidence to suggest that the native unmodified forms of archaeal flagellins usually migrate at about 20 kDa (4, 17, 31, 39). Despite expression of this flagellin in bacterial hosts, it appears that some modification may be occurring which results in a product of approximately 26.5 kDa. Particularly interesting is the fact that this modification results in identical migration of the expressed flagellin in both E. coli and P. aeruginosa. While the nature of the putative modification in the bacterial hosts is not known, it has very recently been shown that the flagellins of a-type strains of P. aeruginosa (including preliminary data on strain PAK) are glycosylated (7). The nature of the glycosylation in these pseudomonads is still unknown; the authors believe an O link likely, although there are a number of N-glycosylation sequons (Asn-X-Ser/Thr) in the a-type flagellins (7). The pilins of some P. aeruginosa strains have also been shown to be glycosylated (9). Clearly, P. aeruginosa strains do possess the capability of protein glycosylation. Other examples of posttranslational modifications of bacterial flagellins include glycosylation (10, 34), methylation (14, 20), and phosphorylation (25).

The use of a membrane fraction containing this preflagellin product was shown to be useful in the detection of preflagellin peptidase activity in cell membranes of M. voltae, using methodologies similar to that developed for prepilin peptidase assays in P. aeruginosa (42). The assay system should also be useful in monitoring the purification of the preflagellin peptidase in future work. Since searches of the complete genomes reported for flagellated archaea do not detect homologs to prepilin peptidases, it may be that the preflagellin peptidase represents a new distant relative of known leader peptidases. The overexpressed preflagellin may also act as substrate in the detection of protein fractions responsible for the putative flagellin posttranslational modifications in M. voltae. As well, this product could be used to adsorb flagellin-binding proteins (e.g., flagellin-specific chaperones) from cell extracts. It has been suggested that cytoplasmic chaperones may be required to prevent nonspecific aggregation of the flagellins via their very hydrophobic N termini (18). Cytoplasmic ATP-dependent flagellin-binding proteins have been isolated from both methanogens and extreme halophiles (18, 29).

Type IV pilins from a variety of bacteria show a high degree of sequence homology for approximately the N-terminal 30 amino acids following the signal peptide (11). Thereafter there is little amino acid conservation among many different type IV pilins. Yet in at least certain cases (5, 16, 33), heterologous pilins with this limited stretch of amino acid homology can be recognized by the pilus accessory proteins and assembled into a pilus in P. aeruginosa. We have previously reported that archaeal flagellins have sequence homology to type IV pilins at the extreme N-terminal 20 to 25 amino acids of the mature protein (13). We have also presented other data consistent with an assembly mechanism for archaeal flagella which is different from that for bacterial flagella and which may be similar to that of pili (18). Recently, we reported that a gene, flaI, located just downstream of the flagellins and which may be cotranscribed with the flagellins shares homology with the ATP-binding pilT homologs required for synthesis of a fully functional type IV pilus (2). A homolog of flaI is also found in the immediate vicinity of the flagellin genes in M. jannaschii (8), A. fulgidus (27), and P. horikoshii (24).

P. aeruginosa, while possessing a type IV prepilin peptidase, was not able to cleave the native leader peptide from the M. voltae flagellin. This inability may be in part due to the fact that the pilins of P. aeruginosa are expressed with a leader peptide of only 6 amino acids, compared to 12 for M. voltae FlaB2. When the leader peptide of P. aeruginosa pilin was fused to the coding region for the M. voltae mature flagellin, cleavage of the leader peptide from the fusion protein appeared to occur, as determined by the shift in mobility in SDS-PAGE of this expression product as compared to that of the native protein. Interestingly, no expression product could be detected in mutants which lack the P. aeruginosa prepilin peptidase, presumably because in the absence of the peptidase activity, the protein is not stable and is quickly degraded in the cell. However, a second fusion in which the +5 position of the flagellin portion was changed to glutamic acid was detected readily in a PilD mutant. Comparison of the migration patterns of the full-length methanogen preflagellin (mature flagellin plus 12-amino-acid leader peptide) with the +5 glutamic acid fusion in the wild type and PilD mutant indicated that the intact preflagellin migrated slower than the fusion in the PilD mutant, which migrated slower than the fusion in the wild type. This indicated that the fusion product was being processed by the wild-type cells and that the enzyme responsible for this processing was the prepilin peptidase. Whether the prepilin peptidase was also methylating the N-terminal phenylalanine after cleavage of the leader peptide as it does in the pilin system is not known. These results indicate that the nature of the mature flagellin or pilin sequence downstream of the cleavage site is unimportant in discrimination by the leader peptidase. This supports the conclusions of Strom and Lory (41) and MacDonald et al. (32), who looked at the effect of different amino acid substitutions in P. aeruginosa pilin on leader peptide cleavage and pilus assembly.

Since the leader peptide of the fusion protein was cleaved, we reasoned that the product would have access to the pilus assembly apparatus and so examined the surface of cells for assembly of filaments similar in diameter to archaeal flagella. This was performed in a pilA strain (in which the pilA mutation also would reduce the competition of the flagellin for the prepilin peptidase) to aid in discrimination between archaeal flagella and the similarly sized native pili of P. aeruginosa. No archaeal filaments were detected in cells synthesizing either the FlaB2 fusion or the FlaB2 fusion with the additional +5 glutamic acid change. Strom and Lory (41) found that substitution of valine for glutamic acid at position +5 resulted in the synthesis of P. aeruginosa pilin that was processed by the leader peptidase but nevertheless was not assembled into a pilus structure. Since the +5 position in FlaB2 was also changed to glutamic acid from glycine, this is not the reason for the failure to assemble a structure on the cell surface. This observation suggests that flagellin assembly requires other specific accessory proteins from M. voltae. A number of such putative flagellum accessory genes have been identified in the flagellin multicistronic unit, and their products bear no significant similarity to other proteins previously described (2). Finally, it is also possible that further posttranslational modification of the flagellins is necessary before the subunits will assemble into the macromolecular structure. For example, glycosylation of flagellins appeared to be necessary for stable filament formation in M. deltae (4).

The experiments described in this report have resulted in the production of a source of archaeal preflagellin, which acts as a substrate for preflagellin peptidase. This enzyme may be specific for the archaeal flagellin system analogous to the prepilin peptidase of the type IV pilus system. One of the defining differences between the bacterial and the archaeal flagella systems is the presence of signal peptides on archaeal flagellins, a situation not described for bacterial flagellins. The possible exception is FlaA of Spirochaeta aurantia (6), but this protein represents the surface protein of the layered flagella and not the filament core proteins, and FlaA shows no homology to any other flagellins studied. Such a fundamental difference between the flagella of the two prokaryotic domains likely manifests itself in a mode of assembly of archaeal flagella clearly distinct from that of bacterial flagella, in which the new flagellin monomers travel down the hollow filament and incorporate at the distal tip (19). Continued study of this unique motility structure in this laboratory is focusing on the role of the archaeal specific accessory proteins in the structure and assembly of the archaeal flagellum.