ATPase Activity and Multimer Formation of PilQ Protein Are Required for Thin Pilus Biogenesis in Plasmid R64*

Plasmid R64 pilQ gene is essential for the formation of thin pilus, a type IV pilus. The pilQproduct contains NTP binding motifs and belongs to the PulE-VirB11 family of NTPases. The pilQ gene was overexpressed with an N-terminal His tag, and PilQ protein was purified. Purified His tag PilQ protein displayed ATPase activity with aV max of 0.71 nmol/min/mg of protein and aKm of 0.26 mm at pH 6.5. By gel filtration chromatography, PilQ protein was eluted at the position corresponding to 460 kDa, suggesting that PilQ protein forms a homooctamer. To analyze the relationship between structure and function of PilQ protein, amino acid substitutions were introduced within several conserved motifs. Among 11 missense mutants, 7 mutants exhibited various levels of reduced DNA transfer frequencies in liquid matings. Four mutant genes (T234I, K238Q,D263N, and H328A) were overexpressed with a His tag. The purified mutant PilQ proteins contained various levels of reduced ATPase activity. Three mutant PilQ proteins formed stable multimers similar to wild-type PilQ, whereas the PilQ D263N multimer was unstable. PilQ D263N monomer exhibited low ATPase activity, while PilQ D263N multimer did not. These results indicate that ATPase activity of the PilQ multimer is essential for R64 thin pilus biogenesis.

Type IV pili have been identified in various Gram-negative pathogens such as Neisseria gonorrhoeae, Vibrio cholerae, Pseudomonas aeruginosa, and enteropathogenic Escherichia coli (1). The contribution of type IV pili to virulence lies primarily in their ability to promote the attachment of the pathogens to various receptors of host cells during colonization. Type IV pili are also required for bacterial locomotion known as twitching motility (2) as well as for the social gliding motility of myxobacteria (3). Proteins with extensive sequence similarity to type IV pilins were also found to be the essential components of protein secretion systems of Gram-negative bacteria and the DNA uptake systems of Gram-positive bacteria (4). Gram-negative bacteria such as Klebsiella oxytoca, Erwinia chrysanthemi, P. aeruginosa, and Xanthomonas campestris have protein secretion systems, called type II secretion systems or the main terminal branch of the general secretory pathway, which export extracellular proteins including pullulanase, pectinase, cellulase, and toxins outside of the outer membrane (5)(6)(7). The genes involved in piliation and protein export are related through amino acid sequence similarity among the structural subunits and conserved assembly proteins (4).
Of the various genes involved in the biogenesis of type IV pili and type II secretion systems, putative NTPases, including PulE of K. oxytoca (8), PilB of P. aeruginosa (9), OutE of E. chrysanthemi (6), and EpsE of V. cholerae (10), collectively known as the PulE-VirB11 family, are highly conserved. They are also found in type IV secretion systems including the conjugation systems of IncP plasmids RP4 and R751, and IncW plasmid R388 (11), the T-DNA transfer systems of agrobacterial Ti and Ri plasmids (12), and the Ptl protein secretion system of Bordetella pertussis (13). The encoded proteins contain several conserved motifs: Walker box A and B involved in ATP binding (14), an Asp box, and a His box (see Fig. 4A). In addition to these motifs, a group of PulE-VirB11 family NT-Pases, related to type IV pilus biogenesis systems and type II secretion pathways, contain two CXXC motifs and an Arg box (8). Many of the PulE-VirB11 family NTPases have been shown to be essential for piliation, extracellular protein secretion, transformation, and conjugal transfer (15). Several PulE-VirB11 family NTPases have been purified and characterized (10, 16 -18). Genetic studies have also been performed for the PulE-VirB11 family NTPases (8, 19 -21). In addition, TrwD and TrbB proteins of plasmids R388 and RP4, respectively, and HP0525 protein in the cag pathogenicity island of Helicobacter pylori have been shown to exhibit NTPase activity (17,22). IncI1 plasmids R64 and ColIb-P9 form two types of sex pili, a thick rigid pilus and a thin flexible pilus. The thin pilus is required only for conjugation in liquid media (23). A 16-kb 1 DNA segment, including the traA-D and pilI-V genes, of the 54-kb R64 transfer region is required for the formation of thin pilus (24). R64 and ColIb-P9 thin pili belong to the type IV pilus family based on the amino acid sequence similarities between various pil products and proteins related to the formation of other type IV pili. The products of the traBC and pilK-V genes were shown to be essential for the formation of thin pilus on the cell surface (25,26). R64 and ColIb-P9 thin pili function in the formation of donor-recipient cell aggregates in liquid matings. Specificity of recipient cells is determined by seven C-terminal segments of PilV adhesins, which are switched through shufflon multiple inversion (27,28). The R64 pilQ product shares amino acid sequence similarity with PulE-VirB11 family NTPases (24).
In this work, the R64 pilQ gene product, essential for thin pilus formation, was overproduced and purified. Purified PilQ protein formed homooctamers and displayed ATPase activity. Several amino acid substitutions were introduced into the pilQ gene within conserved motifs. Effects of pilQ mutations on thin pilus formation, ATPase activity, and multimer stability were examined.
Conjugal Transfer-E. coli NF83 or BL21 (DE3) cells harboring pKK661, pKK641AЈ pilQ1 and pilQ plasmid, and E. coli TN102 cells were used as donor and recipient cells, respectively. Donor cells were grown to the log phase and mixed with recipient cells in the stationary phase. After standing for 90 min at 37°C, the mixture was plated at various dilutions onto LB plates containing chloramphenicol and nalidixic acid. Since pKK661, pKK641AЈ pilQ1 and pilQ plasmid complement each other, donor cells harboring pKK661, pKK641AЈ pilQ1 and pilQ plasmid transmitted pKK661 carrying oriT into recipient cells (23). Transfer frequency is presented as the ratio (expressed as percentage) of the number of transconjugants to that of donor cells. In the case of BL21 (DE3) harboring pEQ11, LB medium containing 0, 10, 50, or 100 M IPTG was used.
Construction of Plasmids-Construction of plasmids and other methods of DNA manipulation were performed as described previously (33).
To construct the T7 promoter overexpression system of the pilQ gene, an NdeI site was introduced at the initiation codon of the pilQ gene by PCR using appropriate primers and pKK702a DNA as a template. The NdeI-BamHI fragment of the PCR product was inserted into the NdeI-BamHI site of pET28b to give pEQ28. The XbaI-BamHI fragment of pEQ28 was inserted into the XbaI-BamHI sites of pUC119 and pET11a to give pKK702h and pEQ11, respectively.
The pilQ mutants were constructed by the PCR-mediated site-directed mutagenesis method using pKK702h as a template (34). The resultant mutants are as follows: F209S (TTT to TCT), I219T (ATA to ACA), T234I (ACC to ATC), K238Q (AAA to CAA), D263N (GAT to AAT), D304N (GAT to AAT), H328A (CAC to GCC), C375A (TGT to GCT), C423A (TGT to GCT), R432A (AGA to GCA), and I490V (ATC to GTC). To construct the ⌬340 deletion mutant, the NdeI-AccI fragment of pKK702h was introduced into the HincII site of pUC119 by blunt end ligation after treatment with Klenow fragment. To construct mutant pilQ overexpression plasmids, the XbaI-BamHI fragments of pKK702h with mutant pilQ genes were inserted into the XbaI-BamHI site of pET28b.
Purification of PilQ Protein-An overnight culture of BL21 (DE3) cells harboring pEQ28 was diluted 50-fold in LB medium (500 ml) containing kanamycin and incubated at 37°C with shaking. When the A 600 of the culture reached ϳ0.5, IPTG was added to a final concentration of 1 mM. After 1 h, the induced cells were harvested by centrifugation and used for the purification of PilQ protein.
All purification steps were carried out at 4°C. IPTG-induced cells were resuspended in 50 ml of TS buffer (20 mM Tris-HCl (pH 8.0) and 100 mM NaCl) and disrupted by a French pressure cell. Cell extract was centrifuged at 100,000 ϫ g for 30 min. The supernatant was applied to a Talon Co 2ϩ affinity column (CLONTECH). After washing the column with TS buffer containing 10 mM imidazole, bound proteins were eluted with TS buffer containing 300 mM imidazole. The PilQ-containing fraction was dialyzed against 20 mM Tris-HCl (pH 7.5) and loaded onto a MonoQ ion exchange column (Amersham Pharmacia Biotech) equilibrated with 20 mM Tris-HCl (pH 7.5). The column was washed with 20 mM Tris-HCl (pH 7.5), and then bound protein was eluted with a linear gradient (from 0 to 1.0 M) of NaCl in 20 mM Tris-HCl (pH 7.5). The PilQ-enriched fractions were applied to a Sephacryl S-200 gel filtration column (Amersham Pharmacia Biotech) equilibrated with buffer G (20 mM Tris-HCl (pH 7.5) and 100 mM NaCl). Peak fractions of PilQ protein were concentrated; dialyzed against a solution containing 20 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM DTT, 100 mM NaCl, and 30% glycerol; and stored at Ϫ80°C.
Determination of ATPase Activity-ATPase activity of His tag PilQ protein was estimated by the thin layer chromatography (TLC) method (35). Reactions were carried out at 30°C for 30 min in 40 l of buffer A (50 mM MES (pH 6.5), 5 mM MgCl 2 , 25 mM KCl, and 1 mM DTT) containing 1 mM [␥-32 P]ATP (0.9 mCi/mmol) and 2 M His tag PilQ protein. Aliquots (1 l) from reaction mixtures were spotted onto polyethyleneimine-cellulose plates (Merck) and developed in 1 M formic acid and 0.5 M LiCl. The spots of free phosphate and ATP were visualized and quantitated by a BAS 2000 image analyzer (Fuji Film). The effects of pH on ATPase activity were examined using MES buffer for pH 5.0 -7.0 and Tris buffer for pH 7.5-9.0.
To determine the kinetics of PilQ ATPase activity, a spectrophotometric method (36) was performed with slight modifications. Reactions were carried out at 37°C in a mixture (1 ml) containing 50 mM MES buffer (pH 6.5), 5 mM MgCl 2 , 25 mM KCl, 1 mM DTT, 3 mM phosphoenolpyruvate, 0.25 mM NADH, 5 units of pyruvate kinase, 10 units of lactate dehydrogenase, 6 g of His tag PilQ protein, and various concentrations of ATP, and decrease in absorbance at 340 nm was measured using a UV160 spectrophotometer (Shimazu). K m and V max values were calculated from Eadie-Scatchard plots.
Preparation of Anti-PilQ Protein Antibody-Four doses of the purified His tag PilQ protein were injected subcutaneously into the back of a New Zealand White rabbit in an emulsion containing complete Freund's adjuvant. Anti-PilQ antibody was purified by affinity chromatography using PilQ-conjugated ECH Sepharose 4B (Amersham Pharmacia Biotech).
Preparation of Thin Pilus Fraction-Crude thin pilus fraction was prepared as described previously (37). E. coli cells harboring pKK641AЈ pilQ1 and pilQ plasmids with or without mutations were grown to late log phase with shaking at 37°C. The culture medium was centrifuged at 140,000 ϫ g for 1 h. The pellet is referred to as the thin pilus fraction.
Subcellular Fractionation-Cells harboring pKK641AЈ or pKK702a were lysed by EDTA-lysozyme and fractionated to periplasm, cytoplasmic, and crude membrane fraction as described previously (38). Crude membrane fraction was separated to outer membrane and inner membrane by the three-step sucrose gradient procedure (39). Amounts of PilQ protein in all fractions were determined by Western blot analysis using anti-PilQ antibody.
Gel Filtration Chromatography-Size and stability of PilQ multimer were analyzed by gel filtration chromatography. The purified wild-type and mutant His tag PilQ proteins (500 g) were applied to a Superose 6 gel filtration column (Amersham Pharmacia Biotech) equilibrated with buffer B (20 mM Tris-HCl (pH 7.0), 100 mM NaCl, and 5 mM MgCl 2 ) with or without 1 mM ATP. Protein concentration and ATPase activity were determined by a protein assay kit (Bio-Rad) and by the TLC method, respectively. The following proteins were used as molecular mass standards: ferritin (450 kDa), catalase (240 kDa), and bovine serum albumin (68 kDa).
Trypsin Digestion of Wild-type and Mutant PilQ Proteins-To analyze ATP-induced conformational change of the PilQ protein, trypsin sensitivity of His tag PilQ protein was compared in the presence and absence of 1 mM ATP. Two g of the purified His tag PilQ protein with or without mutations was digested on ice with 0.3 g of trypsin in 20 l of buffer A with or without 1 mM ATP. After 30 min, the reaction was stopped by the addition of 1 l of 4 mM phenylmethylsulfonyl fluoride and 5 l of 20% trichloroacetic acid. The precipitates of acid-denatured proteins were collected by centrifugation, washed with acetone, and dissolved in SDS sample buffer. The samples were applied to SDS-PAGE (15% gel), and then tryptic fragments of PilQ protein were detected by Western blot analysis using anti-PilQ antibody.

RESULTS
Overexpression and Purification of PilQ Protein-To study the characteristics of the R64 pilQ product, the pilQ gene was cloned into pET28b to give pEQ28 (Fig. 1A). pEQ28 carried a modified pilQ gene in which a stretch of six histidine residues was attached at its N terminus. To examine the effects of the PilQ N-terminal His tag on PilQ activity, the complementation activity of pEQ28 for the pilQ1 frameshift mutation in R64 liquid matings was estimated. For that purpose, pEQ11, an Amp r derivative of pEQ28, was constructed. E. coli strain BL21 (DE3) harboring pKK661, pKK641AЈ pilQ1 and pEQ11, and strain TN102 were used as donor and recipient cells, respectively (Table I). The transfer frequency of pKK661 from donor cells harboring pKK661 and pKK641AЈ pilQ1 and was less than 0.0001%. When pKK702a carrying the wild-type pilQ gene on pUC119 was introduced into donor cells, the transfer frequency was increased to 0.5-1.0%. When pEQ11 was introduced into donor cells, the transfer frequency was also increased to wild-type levels in the presence of 0 -100 M IPTG. These results suggest that the presence of the His tag at the N terminus of PilQ protein does not affect PilQ activity. Therefore, His tag PilQ protein was purified and used for further analyses to characterize its enzymatic and biochemical features.
Upon IPTG induction, E. coli BL21 (DE3) cells harboring pEQ28 overproduced a 58-kDa protein (data not shown). Since this protein was not produced in the absence of induction and the apparent molecular mass of the overproduced protein is consistent with the calculated value of 59,995 Da, the 58-kDa protein is most likely to be His tag PilQ protein. After ultracentrifugation of cell extract, most overproduced PilQ protein was recovered in the soluble fraction (Fig. 1B, lane S). The PilQ protein was purified by Co 2ϩ affinity chromatography (Fig. 1B, lane E). The peak fraction was further purified by MonoQ anion exchange and Sephacryl S-200 gel filtration chromatography (Fig. 1B, lane P). SDS-PAGE followed by silver staining indicated that the purity of His tag PilQ protein was over 98% (data not shown). Approximately 2.5 mg of purified His tag PilQ protein was obtained from a 500-ml culture.
The purified His tag PilQ protein was used to immunize a rabbit. The purified anti-PilQ antibodies specifically reacted against a 58-kDa PilQ protein present in whole cell lysate from E. coli cells harboring pKK641AЈ or pKK702a.
Subcellular Localization of PilQ Protein-Subcellular fractionation experiments were performed to determine localization of PilQ protein within cells. E. coli cells harboring pKK641AЈ or pKK702a were separated into cytoplasmic, periplasmic, inner membrane, and outer membrane fractions. Proteins in every fraction were separated by SDS-PAGE and analyzed for PilQ protein by Western blotting using anti-PilQ antibody. PilQ protein was recovered almost exclusively in the cytoplasmic fraction of cells harboring pKK641AЈ or pKK702a (Fig. 2). These results indicate that the PilQ protein is a cytoplasmic protein.
Detection of ATPase Activity-To detect the ATPase activity of the purified His tag PilQ protein, thin layer chromatography (TLC) analysis was performed. When the purified PilQ protein was incubated with [␥-32 P]ATP, release of a P i radiochromatographic spot was observed on a TLC plate in a time-and PilQ concentration-dependent manner (data not shown), indicating that the PilQ protein contains ATPase activity.
The effects of different reaction conditions on His tag PilQ ATPase activity were examined (Table II). When PilQ ATPase reactions were carried out under different pH conditions using MES (pH 5-7) and Tris (pH 7.5-9) buffers, maximum ATPase activity was observed around pH 6.5, while either increasing or decreasing the pH from 6.5 significantly lowered the ATPase activity. The PilQ ATPase required divalent cations since EDTA, a divalent cation chelator, strongly inhibited ATPase activity. The optimal concentration of Mg 2ϩ required for ATPase activity was ϳ5 mM, whereas a similar concentration of Ca 2ϩ resulted in lower activity. The optimal salt concentration was ϳ25 mM for both KCl and NaCl, while KCl produced slightly higher ATPase activity than NaCl. Therefore, further ATPase assays were performed with 5 mM MgCl 2 , 25 mM KCl and at pH 6.5. Kinetics of ATP-hydrolyzing activity of PilQ protein was estimated by the spectrophotometric method. As derived from the Eadie-Scatchard plot (data not shown), the apparent K m for ATP was 0.26 mM, and the V max value was 0.71 nmol/min/mg of protein.
Inhibition of ATP Hydrolysis with NTPs or dNTPs-To examine whether His tag PilQ protein is able to bind to NTPs or dNTPs other than ATP, the inhibition of PilQ ATPase by NTPs or dNTPs was estimated. A standard ATPase assay with [␥-32 P]ATP was performed in the presence of 5-fold excess of unlabeled NTPs or dNTPs. Unlabeled ATP apparently inhibited PilQ ATPase by 79% as expected (Fig. 3). PilQ ATPase inhibition by dATP and dGTP were 55% and 30%, respectively, while inhibition by GTP, CTP, UTP, dCTP, and dTTP was less than 20%. These results indicate that PilQ protein exhibits the highest affinity to ATP followed by dATP and dGTP. PilQ affinity to other NTPs or dNTPs is low. It is of great interest whether these NTPs or dNTPs are hydrolyzed by PilQ protein.
[␣-32 P]dATP hydrolysis by PilQ protein was demonstrated by TLC analysis (data not shown).
Construction of the Mutant pilQ Genes-In the previous sections, we have demonstrated that R64 PilQ protein contains ATPase activity. Hence, it was of interest to test whether the ATPase activity of PilQ protein is required for R64 thin pilus biogenesis. For that purpose, several amino acid substitutions were introduced within the conserved motifs of the His tag pilQ genes, including the NTP-binding motifs, by the PCR method. In summary, 11 missense and 1 deletion mutants were obtained (Fig. 4A, Table III).
Intracellular production of mutant His tag PilQ protein was analyzed by Western blotting using anti-PilQ antibody (Fig.  4B, top panel). E. coli cells carrying missense pilQ genes produced mutant PilQ proteins at the same levels as that of the wild-type PilQ protein. Extracellular thin pilus production was analyzed using anti-pilin antibody (Fig. 4B, bottom panel). The culture media of E. coli cells harboring pKK641AЈ pilQ1 and mutant pilQ plasmids were ultracentrifuged to recover thin pili detached from the cells. Cells harboring pilQ F209S, T234I, K238Q, D263N, H328A, C375A, and C423A mutants produced very little or no thin pili, whereas cells harboring the remaining missense pilQ mutants produced similar amounts of thin pili as the wild type.
The activity of mutant PilQ proteins for thin pilus biogenesis was also determined as transfer frequency in liquid matings. The effects of the addition of the mutant pilQ plasmids on the transfer frequency in liquid mating from donor cells harboring pKK661 and pKK641AЈ pilQ1 were estimated. E. coli NF83 donor cells harboring pKK661, pKK641AЈ pilQ1 and pKK702h transmitted pKK661 into the recipient cells by conjugation at a frequency of 1.6%, while those harboring only pKK661 and pKK641AЈ pilQ1 did not (Table III, pilQ activity). Different levels of recovery in the transfer frequency were observed by introducing various mutant pilQ genes into donor cells. Transfer frequencies were recovered to the wild-type levels by the pilQ I219T, D304N, R432A, and I490V mutants. For the pilQ F209S, T234I, K238Q, H328A, C375A, and C423A mutants, various levels of recovery in transfer frequencies were observed. For the pilQ D263N and ⌬340 mutants, no recovery was observed. These results are consistent with the results of extracellular thin pilus production (Fig. 4B) and indicate that the pilQ F209S, T234I, K238Q, D263N, H328A, C375A, C423A, and ⌬340 mutant are defective in R64 thin pilus biogenesis and subsequently in R64 liquid matings.
Transdominance of the Mutant pilQ Genes over Wild-type pilQ Gene-Addition of pKK702h (multicopy wild-type His tag pilQ) to donor cells harboring pKK661 and pKK641AЈ had little effect on the transfer frequency (Table III, Transdominance). The effects of introduction of multiple copies of mutant pilQ genes to donor cells harboring pKK661 and pKK641AЈ on the transfer frequency were studied to test whether the His tag pilQ mutants were dominant negative over the wild-type pilQ gene. The introduction of pilQ T234I, K238Q, D263N, and H328A mutants into donor cells harboring pKK661 and pKK641AЈ decreased the transfer frequency, indicating the transdominant character of these mutants. All of the four mutants exhibited reduced or no pilQ activity during liquid matings.
ATPase Activity of Mutant PilQ Proteins-The products of four mutant His tag pilQ genes (pilQ T234I, K238Q, D263N, and H328A), which exhibited reduced transfer frequency and transdominant character in liquid matings, were overproduced and purified by the same procedure as for wild-type His tag PilQ protein (Fig. 1B). ATPase activity of mutant PilQ proteins was determined. All of the mutant PilQ proteins exhibited reduced ATPase activity (Table III, ATPase activity). The ATPase activity of PilQ D263N was particularly low. Residual  transfer frequencies in these mutants were roughly dependent on their residual ATPase activity. These results suggest that reduced ATPase activity in the pilQ T234I, K238Q, D263N, and H328A mutants results in their reduced transfer frequency. Thus, it may be concluded that the ATPase activity of PilQ protein is required for R64 thin pilus biogenesis and consequently for R64 liquid matings. The products of the other mutant pilQ genes (pilQ F209S, C375A, C423A, and ⌬340) with reduced transfer frequency were also overproduced. However, these mutant proteins could not be purified due to their insolubility.
ATP Induced Conformational Change in Mutant PilQ Proteins-The PilQ T234I, K238Q, D263N, and H328A mutant proteins exhibited reduced ATPasae activity. It is important to test whether these mutant proteins are defective in ATP binding. Conformational changes of wild-type and mutant PilQ proteins caused by ATP binding were examined by trypsin digestion. By trypsin digestion of wild-type PilQ protein, only two major fragments (36 and 29 kDa) were produced in the absence of ATP, while two additional minor fragments (32 and 23 kDa) were produced in the presence of 1 mM ATP (Fig. 5). Conformational change of PilQ protein by ATP-binding may result in distinct proteolytic fragment patterns. Distinct proteolytic patterns in the absence and presence of ATP were detected for PilQ T234I, D263N, and H328A proteins. However, for PilQ K238Q, the Walker box A mutant, the 29-kDa fragment was produced even in the absence of ATP. These results suggest that PilQ K238Q mutant is unable to bind to ATP, whereas PilQ T234I, D263N, and H328A mutants are able to bind to ATP despite residual ATPase activity.

Characteristics of the products of various mutant pilQ genes
The pilQ activity was estimated as transfer frequency of pKK661 from E. coli NF83 donor cells harboring pKK661, pKK641AЈ pilQ1 and pKK702h with or without His tag pilQ mutations in liquid matings, and is shown as a percentage of transconjugants relative to donor cells. Transdominance of mutant His tag pilQ genes was estimated as transfer frequency of pKK661 from donor cells harboring pKK661, pKK641AЈ, and pKK702h with or without His tag pilQ mutations. Mutant His tag PilQ proteins were purified from induced E. coli BL21 (DE3) cells harboring pEQ28 with the indicated mutations. ATPase activity of the purified mutant His tag PilQ proteins was determined by TLC method. The relative ATPase activity of the mutant PilQ proteins to that of the wild type is indicated as a percentage in parentheses. were eluted at a position corresponding to a multimer larger than a monomer. The precise size and stability of PilQ multimers were determined by Superose 6 gel filtration chromatography in the presence or absence of 1 mM ATP. The wild-type His tag PilQ protein was eluted at a position corresponding to 460-kDa irrespective of the presence or absence of 1 mM ATP (Fig. 6A). Since the molecular mass of the PilQ monomer was shown to be 58 kDa, wild-type PilQ protein most likely forms a homooctamer in the presence and absence of 1 mM ATP. These results also indicate that the PilQ octamer is stable in the presence or absence of 1 mM ATP. The ATPase activity of each fraction roughly corresponds to its protein concentration. Three mutant PilQ proteins (T234I, K238Q, and H328A) formed stable multimers with reduced ATPase activity in the presence and absence of 1 mM ATP (data not shown). In contrast, PilQ D263N multimer was unstable especially in the absence of ATP, since PilQ D263N protein was eluted as a mixture of multimers and monomers (Fig. 6B). The PilQ D263N monomer exhibited weak ATPase activity, while its multimer did not. When concentrated PilQ D263N monomer fraction was rechromatographed on Superose 6, PilQ D263N protein was again eluted as a mixture of multimers and monomers, indicating that PilQ D263N protein is in equilibrium with multimer and monomer forms (data not shown). These results indicate that the region containing the PilQ D263N mutation is involved in stable multimer formation.

DISCUSSION
In the present work, we have overproduced the R64 pilQ gene encoding a PulE-VirB11 family NTPase with an N-terminal His tag. His tag PilQ protein was purified to near homogeneity (Ͼ98%). Purified PilQ protein displayed ATPase activity with a V max of 0.71 nmol/min/mg of protein and a K m of 0.26 mM at pH 6.5. The V max and K m values of PilQ ATPase are similar to those reported for PulE-VirB11 family NTPases such as R388 TrwD, RP4 TrbB, and H. pylori cag HP0525 (17,22).
PilQ protein was shown to be located in the cytoplasmic fraction irrespective of the presence or absence of the other Pil proteins. These results differ from those of K. oxytoca PulE protein, which was shown to be located in the inner membrane (8). R388 TrwD was located in the cytoplasmic and outer membrane fraction (17). PilQ ATPase may function at or near the inner membrane, since it is involved in the formation of thin pilus, a cell envelope-associated appendage. Association of PilQ and the other Pil components may be weak.
Optimum pH for PilQ ATPase was different from those of R388 TrwD, RP4 TrbB, and H. pylori cag HP0525 (17,22). The pH optimum of PilQ was located at approximately 6.5, while that of TrwD and HP0525 ATPase and TrbB dATPase FIG. 5. Protease digestion of mutant PilQ proteins. Two g of purified wild-type and mutant PilQ proteins were digested on ice with 0.3 g of trypsin for 30 min in the presence (ϩ) or absence (Ϫ) of 1 mM ATP. Trypsin was inactivated by the addition of 4 mM phenylmethylsulfonyl fluoride and 20% trichloroacetic acid. The samples were subjected to SDS-PAGE (15%), and then PilQ fragments were analyzed by Western blotting using anti-PilQ antibody. Lane N represents wild-type PilQ protein without trypsin. Intact PilQ protein and minor proteolytic fragments produced only in the presence of ATP are marked by arrow with PilQ and asterisk, respectively, at the right side. Molecular size markers are indicated at the left. was located at approximately 9.0 -9.5. Differences in substrate affinity were found among PilQ, TrbB, HP0525, and TrwD NTPases. PilQ, HP0525, and TrwD exhibited the highest activity for ATP, while TrbB exhibited maximum activity for dATP. Preferred substrates for PilQ are ATP, dATP, and dGTP, whereas those for TrbB are dATP, GTP, and ATP, and those for HP0525 are ATP, dATP, CTP, and dCTP. Furthermore, TrbB protein displayed distinct optimum pH for ATPase and GTPase activities (22). The biological significance of differences in substrate affinity remains to be elucidated.
PilQ protein formed homooctamers either in the presence or absence of ATP. PilQ homooctamer formation was observed even in the absence of Mg 2ϩ (data not shown). Formation of homohexameric ring structure was reported for TrbB, HP0525, and TrwD (22,40). TrbB hexamer was formed in the presence of Mg 2ϩ and dATP or dADP, while TrbB tetramer was formed in the absence of Mg 2ϩ and dATP. In contrast, HP0525 hexamer was formed irrespective of the presence or absence of Mg 2ϩ and dATP. The requirement of NTPs for multimer formation of various PurE-VirB11 NTPases may depend on differences in their protein structure.
Mutation analyses have been performed on various PulE-VirB11 family NTPases (8, 19 -21). In the present report, we have introduced several amino acid substitutions into the pilQ gene within the Walker box A, B, Asp box, His box, Arg box, and CXXC motifs. Two Walker box A mutants, PilQ T234I and PilQ K283Q, failed to produce thin pilus and exhibited low transfer frequencies. Purified mutant proteins displayed reduced ATPase activities. PilQ K283Q affected ATPase activity more severely than PilQ T234I. A trypsin digestion assay revealed that PilQ K238Q lost ATP binding activity, while PilQ T234I retained it. The failure of ATP binding of PilQ K238Q may be due to the loss of positive charge required for the direct interaction with ␤ and ␥ phosphates of ATP resulting from the substitution of lysine by glutamine. Similar Walker box A mutants of PulE and TrbB also displayed severe effects on pullulanase secretion and conjugation, respectively (8,22). Walker box B mutant PilQ D304N had no effects on thin pilus formation and transfer frequency.
Asp box mutant PilQ D263N exhibited the most severe effects on thin pilus biogenesis and liquid matings. The purified PilQ D263N protein displayed very low ATPase activity. The PilQ D263N octamer was unstable, especially in the absence of ATP. His box mutant PilQ H328A failed to produce thin pilus and exhibited low transfer frequency. The purified PilQ H328A protein displayed 43% ATPase activity.
Four mutant genes (pilQ T234I, K238Q, D263N, and H328A) exhibited dominant negative character over the wild-type allele. The dominant negative character of these mutants suggests that mixed octamers can be formed from the mutant and wild-type PilQ proteins. RP4 TrbB, H. pylori HP0525, and R388 TrwD form hexameric ring structures. Krause et al. (40) postulated that such ring structures may aid a repetitive step/ process of NTP hydrolysis. Incorporation of mutant monomers into PilQ octamers may inhibit the function of the octamer as a whole. Hence, the observed transdominance of PilQ mutants over the wild-type allele supports the above postulation. In this regard, the finding that the monomer form of the PilQ D263N mutant exhibited residual ATPase activity, is of particular interest.
The four mutant proteins (PilQ T234I, K238Q, D263N, and H328A) displayed varying levels of residual ATPase activities (2-43% of wild-type). In contrast, the amounts of thin pilus produced by these mutants were very low as were their transfer frequencies (less than 1.5% of the wild-type). This situation is different in RP4 TrbB mutants (22). Some TrbB mutants exhibiting very low dATPase activity still exhibited wild-type levels of conjugation activity. Dependence of the R64 thin pilus biogenesis system on PilQ is stronger than that of RP4 system on TrbB.
Since the two CXXC motifs and the Arg box are notably conserved among PulE-VirB11 family NTPases of type IV pilus biogenesis systems and type II secretion systems, mutations were introduced into these motifs in PilQ. Two CXXC motif mutants, PilQ C375A and C423A, failed to produce thin pili and exhibited low transfer frequencies. Overexpression of these mutants resulted in insoluble PilQ mutant proteins. Insolubility of these mutants may be due to the failure of correct folding. Cysteine residues in PilQ CXXC motifs may be responsible for intramolecular disulfide bond formation, as was suggested for K. oxytoca PulE (21). Lack of transdominance may be related to the insolubility or abnormal conformation of these mutant proteins, as they were unable to form mixed octamers with wildtype protein. Arg box mutant PilQ R432A carried pilQ activities similar to the wild type, although this arginine residue in the Arg box is highly conserved in many PulE-VirB11 family NTPases of type IV biogenesis systems and type II secretion systems.
In conclusion, ATPase activity and multimer formation of PilQ protein are essential for R64 thin pilus biogenesis. However, further investigation is required to reveal the precise role(s) of R64 PilQ ATPase in the biogenesis of R64 thin pilus.