| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 280, Issue 26, 24839-24848, July 1, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶
From the
Division of Infectious Diseases, Department of Medicine, University of Maryland School of Medicine, Baltimore, Maryland 21201 and the Department of Molecular Biology and The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, California 92037
Received for publication, January 7, 2005 , and in revised form, April 18, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
S binds to BfpE77-114, whereas hexameric BfpD-ADP binds to BfpE39-76. Interestingly, the N terminus of BfpC and BfpE77-114 together increased the ATPase activity of hexameric BfpD over 1200-fold to a Vmax of 75.3 µmol of Pi min-1 mg-1, which exceeds by over 1200-fold the activity of other PulE family members. This augmented activity occurred only in the presence of Zn2+. We conclude that allosteric interactions between BfpD and BfpC and BfpE dramatically stimulate its ATPase activity. The differential nucleotide-dependent binding of hexameric BfpD to BfpE39-76 and BfpE77-114 suggests a model for the mechanism by which BfpD transduces mechanical energy to the biogenesis machine. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
EPEC is a major cause of severe infantile diarrhea and a leading cause of infant mortality in developing countries (18). The BFP of EPEC is a particularly good system to study Tfp biogenesis because all of the genes necessary for BFP biogenesis are known. The 14-gene bfp operon contains 13 genes required for BFP biogenesis and function (4, 19-26). The bfpA gene encodes prebundlin (27, 28), which is processed into bundlin (the pilin protein) by the protein product of bfpP (the prepilin peptidase (19)). The precise functions of the other components of the biogenesis machinery are unknown. The bfpB gene encodes an outer membrane lipoprotein of the secretin family (20) that is hypothesized to form a pore in the outer membrane through which the BFP is extruded. The bfpD and bfpF genes encode putative cytoplasmic nucleotide-binding proteins BfpD and BfpF (22, 23), which are thought to provide energy for BFP biogenesis and retraction, respectively. The bfpE gene encodes a polytopic cytoplasmic membrane protein with four transmembrane domains, a large cytoplasmic N terminus, two periplasmic domains, and a small cytoplasmic domain (26). We recently demonstrated that the N terminus of BfpE interacts with BfpD and the cytoplasmic N terminus of the bitopic cytoplasmic membrane protein BfpC, which also interacts with BfpD (29). Together, BfpC and BfpE recruit BfpD to the membrane (29).
BfpD is a member of the PulE superfamily of putative nucleotide-binding proteins involved in bacterial type II and IV secretion, Tfp assembly, natural competence, and assembly of archeal flagellae (17). These proteins contain several conserved sequence motifs, including Walker A and B boxes involved in ATP binding and hydrolysis (30). However, ATPase activity has been confirmed for only a few of them, mostly from type IV secretion systems (31-38). Furthermore, this ATPase activity has been relatively weak, between 0.71 and 61 nmol of inorganic phosphate (Pi) released min-1 mg-1 of protein. Mutation of key residues in the Walker A box of PulE superfamily members abolishes activity of the cognate systems, for example, PulE (39, 40), EpsE (41), OutE (42), PilB (43), and BfpD (4). Therefore, these proteins most likely share the function of supplying energy from ATP hydrolysis to drive organellar assembly or substrate translocation.
The characterization of the ATPase activity of BfpD is crucial for understanding its role in BFP biogenesis and mechanism of action. In this paper we describe a detailed study of the ATPase activity of BfpD in which we address its activity in the context of the BFP biogenesis machine, i.e. in the presence of components of the machinery with which it interacts. The results of our studies suggest a model for conversion of chemical energy into mechanical energy to power extraction of bundlin from the cytoplasmic membrane for assembly into BFP filaments.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
pRPA302, respectively, as described previously (29). Briefly, DH5pRPA302 and BL21(DE3)pLysS,pRPA405 cultures were grown at 37 °C to an optical density at 600 nm of 
0.5, His-tagged protein expression was induced with 1 mM isopropyl 1-thio-
-D-galactopyranoside and cultures were grown for an additional 2 and 5 h, respectively. Cells were harvested by centrifugation for 20 min at 4,000 x g and lysed in a French press in 10 mM imidazole, 300 mM NaCl, 50 mM NaH2PO4, pH 8.0. Lysates were centrifuged twice (10,000 x g, 10 min) and BfpC and BfpD were purified by affinity chromatography using nickel resin (Qiagen). BfpC was further purified on a HiTrap Chelating HP 1-ml column (Amersham Biosciences) in 20 mM NaH2PO4, 500 mM NaCl, pH 7.4, and then dialyzed against 40 mM Tris-HCl, 150 mM NaCl, pH 7.0. Meanwhile, BfpD was further purified on a precalibrated HiPrep 16/60 Sephacryl S-300 gel filtration column (Amersham Biosciences) in 20 mM Tris-HCl buffer, pH 7.6, containing either 100 mM NaCl alone, 100 mM NaCl and 10 mM MgCl2, 100 mM NaCl and 1 mM ATP or 100 mM NaCl, 10 mM MgCl2, and 1 mM ATP. For use in ATPase assays, BfpD was purified in 40 mM Tris-HCl, 150 mM NaCl, 1 mM ADP, pH 7.0. Finally, the proteins were concentrated in Centricon YM-10 tubes (Millipore), quantified by the method of Gill and von Hippel (44) using protein extinction coefficients, and analyzed by SDS-PAGE and Western immunoblotting or silver staining. The gel filtration column was calibrated with blue dextran (2,000 kDa), thyroglobulin (669 kDa), ferritin (440 kDa), catalase (232 kDa), aldolase (158 kDa), bovine serum albumin (67 kDa), ovalbumin (43 kDa), chymotrypsinogen A (25 kDa), and ribonuclease A (13.7 kDa; Amersham Biosciences). The Stokes radius (R) of each form of BfpD was determined from a plot of the partition coefficient (Kav) versus log R of standards as previously described (45). Kav was calculated using the equation Kav = (Ve - Vo)/(Vt - Vo), where Ve is the elution volume, Vo is the void volume, and Vt is the bed volume.
Zonal Sedimentation Analysis
Purified BfpD (20 µg) was applied to the top of a 2.5-ml 20-40% sucrose gradient formed in 20 mM Tris-HCl buffer, pH 7.6, 100 mM NaCl, 10 mM MgCl2, with 1 mM ATP as appropriate. Gradients were centrifuged at 245,000 x g for 12 h at 4 °C. Fractions (100 µl) were collected, precipitated with trichloroacetic acid, and subjected to SDS-PAGE and immunoblotting. Molecular mass standards run in parallel gradients were visualized by SDS-PAGE and Coomassie staining. Sedimentation coefficient (s20,w) values were determined by sedimentation relative to protein standards.
Molecular mass values were calculated from the equation molecular mass = 6
Nas20,w
20,w/1 - 
20,w in which N is Avogadro's number, a is the Stokes radius, 20,w is the viscosity of water at 20 °C, is the partial specific volume of the protein calculated from its amino acid composition, and 20,w is the density of water at 20 °C as described (45).
Dynamic Light Scattering
Dynamic light scattering (DLS) measurements were carried out using a DynaPro 99 photometer (Protein Solutions). BfpD (0.2-0.6 mg ml-1) in 20 mM Tris-HCl buffer, pH 7.5, 100 mM NaCl, 10 mM MgCl2, containing 1 mM ATP, 1 mM ATPS, or 1 mM ADP was measured at 20 °C for 5 min. Data were analyzed with Dynals software.
Electron Microscopy
Negative Staining—Carbon-coated copper grids were floated on 5-µl drops of 100 µgml-1 BfpD in 20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.1 mM EDTA, 10 mM MgCl2, 1 mM ATPS. After 5 min the sample was wicked off, and grids were floated consecutively on 3 drops (60 µl) of 3% uranyl acetate, with blotting in between. Grids were floated on the third drop for 1 min, then blotted and air-dried. Images were recorded on a Philips/FEI CM100 electron microscope operating at 100 keV in low dose mode at x52,000 magnification and a defocus of 500 nm.
Cryo-electron microscopy—Five-µl drops of BfpD were applied to Quantifoil holey grids (Quantifoil Micro Tools GmbH) that had been glow-discharged in the presence of amyl amine. After 1 min the grids were blotted for 2.5 s and then plunged into an ethane slush cooled with liquid nitrogen, using the Vitrobot (FEI). Grids were transferred to a Gatan 626 cold stage (Gatan), and images were recorded on a Philips/FEI CM120 electron microscope operating at 120 keV in low dose mode at x50,000 magnification and a defocus of 1 µm.
ATPase Assay
The ATPase activity of hexameric BfpD, purified in the presence of ADP and the absence of MgCl2, was measured as the Pi released by the hydrolysis of ATP using a colorimetric malachite green assay (46). ATPase activity was determined from the initial rate of Pi release, which was linear with time over a period of at least 1 h (Fig. 2B). Assays were carried out in 40 mM Tris-HCl, pH 7.0, 150 mM NaCl, 10 mM ZnCl2, and 5% glycerol at 37 °C in a 50-µl reaction volume. The metal dependence of the ATPase activity of BfpD was examined by substituting the ZnCl2 in the reaction buffer with MnCl2, CaCl2, CuCl2, MgCl2, FeCl2, CoCl2, CdCl2, and NiCl2. In experiments to determine the influence of BfpC and BfpE on ATPase activity, different combinations of the N terminus of BfpC (6:1 molar ratio of BfpC to BfpD), BfpE39-76 (6:1 molar ratio of BfpE to BfpD), and BfpE77-114 (3:1 molar ratio) were incubated with hexameric BfpD at 37 °C for 30 min prior to initiation of the reactions.
Reactions were initiated by the addition of ATP to a 5 mM final concentration and allowed to proceed for 0, 5, 10, 15, 20, 25, and 30 min. For substrate saturation experiments, ATP was varied from 0.1 to 10 mM. Reactions were terminated and released Pi quantified by the addition of 800 µl of freshly prepared color reagent (0.034% malachite green, 0.1% Triton X-100, and 1.05% ammonium molybdate in 1 M HCl; filtered to remove insoluble material). After 1 min at room temperature, color development was stopped by addition of 100 µl of 34% citric acid. Samples were incubated at room temperature for 30 min, and then the absorbance at 660 nm was measured. Values were calibrated against KH2PO4 standards and corrected for Pi released in the absence of BfpD and the absence of ATP. Data from substrate saturation experiments, where ATP concentration was varied, were fit by nonlinear regression to the Michaelis-Menten equation with Origin software (Microcal) and the parameters Km and Vmax were calculated. Similar values were obtained using Lineweaver-Burk and Eadie-Hofstee plots.
Yeast Two-hybrid Experiments
We used the Matchmaker GAL4 two-hybrid system 3 (Clontech) following the protocol suggested by the manufacturer. The cloning of DNA fragments encoding BfpD and the N terminus of BfpE into pGBKT7 (GAL4 DNA-binding domain vector) and pGADT7 (GAL4 activation domain vector) has been described previously (29). In addition, PCR-amplified fragments encoding amino acids 1-38, 39-76, 77-114, 1-76, and 39-114 of the N terminus of BfpE (for primers see Table I) were cloned into these vectors. Saccharomyces cerevisiae strain AH109 was co-transformed with these constructs by the lithium acetate procedure and a minimum of 30 transformants for each vector combination were used to test protein interactions. The transformants were plated on medium containing X--Gal and lacking tryptophan, leucine, histidine, and adenine and assayed for -galactosidase activity to detect and quantify interactions.
|
S or ADP. Peptide corresponding to BfpE39-76 (0.018 mM) or BfpE77-114 (0.009 mM) in 20 mM Tris-HCl buffer, pH 7.6, containing 100 mM NaCl, 10 mM MgCl2 and either 1 mM ATPS or 1 mM ADP was loaded into the sample cell. BfpD (0.12 mM), in the same buffer as the peptide, was loaded into the injection syringe and titrated into the sample cell. A typical titration series involved 35 injections of 3 µl of titrant. The baseline was allowed to stabilize for 210 s between injections to permit adequate time for reaction and equilibration. All experiments were performed at 37 °C and the cell was stirred continuously at 280 rpm. For control experiments to determine the heats of dilution of BfpD, the peptides were substituted by buffer. The heat produced for each injection of BfpD into peptide or buffer was obtained by integration of the area under each peak of the titration plots with respect to time. The heats of reaction were calculated by subtraction of the integrated heats of dilution of BfpD from the heats corresponding to the injection of BfpD into peptide. Titration data, corrected for heat of dilution, were analyzed with Microcal Origin using a single-site model.
Fluorescence Measurements
The fluorescent nucleotide analogues 2',3'-O-trinitrophenyl (TNP)-ATP and TNP-ADP were used to study nucleotide binding to BfpD. Experiments were performed at 37 °C using a spectrofluorometer with bandwidths of 4 nm. Monomeric BfpD (0.72 µM) was incubated for 30 min with various concentrations of TNP-nucleotide in 20 mM Tris-HCl buffer, pH 7.6, containing 100 mM NaCl and 10 mM MgCl2. The fluorescence emission was measured at 535 nm upon excitation at 408 nm (total fluorescence). Enhancement of fluorescence (
F) was calculated as the difference between total fluorescence and the fluorescence measured in incubations of TNP-nucleotide in buffer. The enhanced fluorescence in the presence of BfpD, proportional to the quantity of BfpD bound TNP-nucleotide, was plotted as a function of TNP-nucleotide concentration. Data were fitted to the equation F = (Fmax x [S])/(Kd + [S]), where F is the change in fluorescence intensity of TNP-nucleotide at a concentration [S], Fmax is the maximum change in fluorescence intensity, and Kd is the dissociation constant. Fitting was performed using Origin and values for Fmax and Kd were obtained.
The stoichiometry of TNP-nucleotide binding to BfpD was derived using the method of Stinson and Holbrook (47). Briefly, 1/(1 - ) was plotted against [S]/, where is the fractional saturation of the TNP-nucleotide-binding sites (F/Fmax). A straight line with a slope of 1/Kd and an x-intercept of the concentration of TNP-nucleotide-binding sites was obtained. The stoichiometry of binding was then calculated by dividing the x-intercept with the BfpD concentration.
Proteinase K Digestion
All protease digestions were carried out at room temperature in 20 mM Tris-HCl buffer, pH 7.6, 100 mM NaCl, 10 mM MgCl2, containing 1 mM ATPS or 1 mM ADP as appropriate. Hexameric BfpD (5 µg), purified in the presence of ATPS or ADP, was preincubated with BfpE39-76 (5 µg), BfpE77-114 (5 µg), or neither of these peptides at room temperature for 10 min. Proteolysis was initiated by addition of proteinase K to give final concentrations of 0, 0.05, 0.1, and 1 mg ml-1 and the reaction was allowed to proceed for 30 min. Proteinase K digestion was terminated by adding Laemmli buffer (48) and boiling the samples for 10 min. Finally, proteolytic products were subjected to SDS-PAGE and detected by silver staining.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
S, TNP-ATP, ADP, and TNP-ADP, suggesting that nucleotide binding, but not its hydrolysis, is required for BfpD hexamerization.
To accurately determine the molecular masses of the two BfpD forms, parameters obtained from gel filtration (Stokes radii) and zonal sedimentation (sedimentation coefficients) were analyzed (Table II). These indicate that the BfpD in the slower peaks was monomeric BfpD (molecular mass = 60 kDa), whereas that in the faster peaks was hexameric (molecular mass = 369 kDa). In addition, DLS analysis of the high molecular mass form of BfpD in the presence of ATP, ATPS, or ADP showed that BfpD is a hexamer (Table III).
|
|
11.5 nm. Furthermore, the rings in the electron micrographs of frozen-hydrated BfpD (Fig. 1, C-E) display 6-fold symmetry, suggesting a hexamer structure consistent with the gel filtration, zonal sedimentation, and DLS results. Taken together, these findings demonstrate that BfpD assembles to form hexamers in the presence of ATP, ATPS, TNP-ATP, ADP, and TNP-ADP. Hexameric BfpD Is an ATPase—We detected the ATPase activity of BfpD using a colorimetric malachite green assay to measure the amount of Pi released upon incubation with ATP. BfpD purified in 40 mM Tris-HCl buffer, pH 7.0, containing 150 mM NaCl and 1 mM ADP was used for these ATPase assays. Use of ADP instead of ATP in the BfpD purification produces hexameric BfpD and has the advantage of eliminating high background Pi levels. ADP does not inhibit BfpD ATPase activity as in assays with increasing concentrations of ADP (up to 1000-fold molar excess over the final ADP concentration in the reactions) and BfpD purified in 40 mM Tris-HCl buffer, pH 7.0, containing 150 mM NaCl and 1 mM ATP there was no effect on BfpD activity (data not shown).
|
When hexameric BfpD protein was incubated with ATP, release of Pi occurred in a time- and BfpD concentration-dependent manner (Fig. 2, B and C), indicating that hexameric BfpD has ATPase activity. That this activity was because of BfpD rather than a contaminating ATPase is demonstrated by the fact that preparations obtained by an identical purification scheme from cells containing plasmid lacking bfpD did not display ATPase activity. Interestingly, monomeric BfpD only began to display ATPase activity after 30 min (data not shown). Using gel filtration chromatography, we found that monomeric BfpD hexamerizes in the presence of nucleotide, such as the ATP added to initiate ATP hydrolysis (data not shown). Therefore, we believe that the ATPase activity exhibited by monomeric BfpD was actually that of nascent hexamers. Finally, BfpD did not hydrolyze any other nucleotides (ADP, AMP, GTP, CTP, and UTP) tested in substrate specificity experiments (data not shown).
To determine the Michaelis-Menten constants Vmax and Km of the ATP hydrolyzing activity of hexameric BfpD, ATPase activity was measured over a range of ATP concentrations, under initial rate conditions. The dependence of the rate of ATP hydrolysis on ATP concentration exhibited typical Michaelis-Menten behavior (Fig. 2D) and showed a linear relationship in Eadie-Hofstee and Lineweaver-Burk plots (data not shown). The Vmax values obtained from these experiments (Table IV) are consistent with the results shown in Fig. 2, A-C. In conclusion, under optimal conditions, the ATPase activity was calculated as 62.2 ± 2.2 nmol released Pi min-1 mg-1 of protein (mean ± S.E.). Similar results were obtained using an assay coupling regeneration of ATP to oxidation of NADH (data not shown). Because hexameric, but not monomeric, BfpD displays ATPase activity, it appears that BfpD ATPase activity is activated upon hexamerization.
|
-Gal and lacking tryptophan, leucine, histidine, and adenine and quantified by measuring -galactosidase activity. We found reporter gene activation when fusion protein pairs murine p53 and simian virus 40 (SV40) large T-antigen (positive control), BfpD and BfpE1-114 (full-length cytoplasmic N terminus), BfpD and BfpE1-76, BfpD and BfpE39-114, BfpD and BfpE39-76, and BfpD and BfpE77-114 were expressed (Fig. 3). These interactions occurred irrespective of which protein was fused to the activation and which to the DNA-binding domains of GAL4. The remaining transformants did not grow on medium containing X--Gal and lacking tryptophan, leucine, histidine, and adenine and exhibited low levels of -galactosidase activity, similar to the negative control expressing human lamin C and SV40 large T-antigen fusion proteins (Fig. 3). Western analysis showed that the transformants expressed the Bfp fusion proteins at comparable levels (results not shown). Furthermore, S. cerevisiae AH109 cells co-transformed with each construct individually and either "empty" pGBKT7 or pGADT7 grew on the appropriate selective medium, i.e. they did not grow on medium containing X--Gal and lacking tryptophan, leucine, histidine, and adenine and grew as white colonies on medium containing X--Gal and lacking tryptophan and leucine (results not shown). This confirms that the individual fusion proteins cannot activate the reporter genes alone and that the interactions we observed were because of direct interactions between the Bfp domains of the fusion proteins. Thus, it appears that the N terminus of BfpE contains two sites to which BfpD binds, one in BfpE39-76 and one in BfpE77-114. However, these data do not distinguish whether these sites represent two independent BfpD-binding sites or whether these sites cooperate to form one binding site.
|
S (a non-hydrolysable ATP analogue) and ADP. Large exothermic enthalpies were observed in titrations of BfpD-ATPS into BfpE77-114 and BfpD-ADP into BfpE39-76. The titration heat was converted to heat mol-1 as a function of molar ratio and the data were fitted with one set of identical binding site models to give the association constant (K), stoichiometry (molar ratio of BfpD/peptide, n), and enthalpy change (H) upon peptide-BfpD complex formation (49). These parameters show that one molecule of hexameric BfpD-ADP strongly interacts with six molecules of BfpE39-76 (K = 1.33 x 107 ± 2.6 x 105 M-1, Kd = 75 nM, n = 0.162 ± 2.62 x 10-4, H = -1.02 x 106 ± 242 cal/mol) and that one molecule of hexameric BfpD-ATPS strongly interacts with three molecules of BfpE77-114 (K = 1.39 x 107 ± 2.04 x 105 M-1, Kd = 72 nM, n = 0.334 ± 3.88 x 10-4, H = -3.36 x 105 ± 577 cal/mol). Importantly, hexameric BfpD-ATPS showed no measurable interaction with BfpE39-76 and hexameric BfpD-ADP showed no interaction with BfpE77-114 (Fig. 4), even when peptide concentrations were up to 100 times higher (data not shown). Furthermore, we were unable to detect interactions between monomeric (nucleotide-free) BfpD and BfpE39-76 or BfpE77-114 (data not shown). Similarly, no interaction could be detected between monomeric BfpD and the cytoplasmic domain of BfpC (data not shown). In contrast, hexameric BfpD has previously been demonstrated to bind this region of BfpC by ITC (29). To summarize, these results show differential binding of hexameric BfpD to BfpE39-76 and BfpE77-114 depending on the phosphorylation of the nucleotide, i.e. hexameric BfpD binds to BfpE39-76 in the presence of ADP and to BfpE77-114 in the presence of ATPS. Interestingly, these data suggest that up to six molecules of BfpE can bind to BfpD that contains ADP, whereas a maximum of three molecules of BfpE can bind to BfpD containing ATP. In addition, the low Kd for these interactions suggest that BfpD-BfpE39-76 and BfpD-BfpE77-114 complexes are likely to be biologically relevant.
|
|
S to BfpE77-114 and BfpD-ADP to BfpE39-76 reflects the stoichiometry of binding of BfpD to ATPS and ADP, respectively, we used the non-hydrolysable nucleotide analogues TNP-ATP and TNP-ADP to ascertain the parameters of BfpD binding to nucleotides. TNP-nucleotides are weakly fluorescent in aqueous solution, however, their fluorescence is greatly enhanced in a hydrophobic environment such as the nucleotide-binding site of a protein. Upon excitation at 408 nm, each TNP-nucleotide in solution exhibited a characteristic fluorescence emission with a maximum at 558 nm; addition of monomeric BfpD markedly increased the fluorescence and blue shifted the maximum to 535 nm (data not shown). When monomeric BfpD was incubated with the TNP-nucleotides, both TNP-ATP and TNP-ADP bound to BfpD with saturable enhancement in the fluorescence emission (Fig. 5). The data were analyzed as described by Stinson and Holbrook (47) and the Kd and the concentration of TNP-nucleotide-binding sites were obtained from the reciprocal of the slope and the x-intercept, respectively, of the resulting plots (data not shown). BfpD bound TNP-ATP (Kd 340 nM) with higher affinity compared with TNP-ADP (Kd 760 nM). The stoichiometry of binding (molar ratio of TNP-nucleotide/BfpD), calculated by dividing the concentration of TNP-nucleotide-binding sites with the BfpD concentration, was 1.07 for TNP-ATP and 0.98 for TNP-ADP. These results indicate that one molecule of monomeric BfpD binds one molecule of either of these TNP-nucleotides. Because monomeric BfpD hexamerizes in the presence of either TNP-ATP or TNP-ADP, we can assume that one molecule of hexameric BfpD binds six molecules of the TNP-nucleotides, each BfpD monomer of the hexamer binding one TNP-nucleotide molecule. Thus, the inability of hexameric BfpD to bind more than three BfpE77-114 molecules is not because of an inability of each subunit to bind ATP.
|
S bind one molecule of BfpE77-114. If this is the case, it suggests that binding of one monomer of BfpD-ATPS to BfpE77-114 induces conformational changes that preclude binding of adjacent monomers. To examine the effect of BfpE binding on the conformation of BfpD occupied by either ATPS or ADP, BfpD in the presence of peptides of BfpE39-76 or BfpE77-114 was compared with BfpD in the absence of these peptides with respect to susceptibility to proteinase K proteolysis. Fig. 6 shows silver-stained gels of two representative proteinase K digestion experiments in which hexameric BfpD, purified in the presence of ATPS (top panel) or ADP (bottom panel), was incubated alone and together with BfpE39-76 or BfpE77-114 and subjected to limited proteinase K proteolysis. Note that BfpE39-76 and BfpE77-114 migrate too rapidly to be seen on the gels. BfpD-ATPS and BfpD-ADP alone were relatively resistant to proteinase K. The presence of BfpE39-76 and BfpE77-114 did not appear to alter the resistance of BfpD-ATPS and BfpD-ADP, respectively, to proteinase K. In contrast, BfpE77-114 significantly increased the sensitivity of BfpD-ATPS to digestion by proteinase K. In addition, BfpE39-76 modestly but reproducibly increased the sensitivity of BfpD-ADP to proteolysis. These results show that BfpE39-76 and BfpE77-114 only had a detectable effect on the proteinase K sensitivity of the form of BfpD with which our ITC data demonstrates they interact. Furthermore, they suggest that BfpE39-76 and BfpE77-114 induce changes in the conformation of BfpD-ADP and BfpD-ATPS, respectively, upon binding, which account for the observed increases in sensitivity to proteinase K digestion. However, consistent with our hypothesis, BfpD-ADP and BfpD-ATPS undergo different conformational changes as they display differing proteinase K sensitivities.
BfpD Binding to the N Terminus of BfpC and BfpE77-114 Dramatically Increases Its ATPase Activity—Because hexameric BfpD interacts with the N terminus of BfpC, BfpE39-76, and BfpE77-114 (Ref. 29, and this study), we investigated the influence of these proteins on ATPase activity. Hexameric BfpD was preincubated with various combinations of the N terminus of BfpC, BfpE39-76, and BfpE77-114 at a molar ratio of BfpC or BfpE to BfpD of 6:1, 6:1, and 3:1, respectively, based on reaction stoichiometry determined by ITC, and the ATPase activity was assayed. The results in Fig. 7A demonstrate that the proteins with which hexameric BfpD-ATPS interacts, i.e. the N terminus of BfpC and BfpE77-114, had a dramatic effect on the ATPase activity of hexameric BfpD. The N terminus of BfpC and BfpE77-114 together stimulated the ATPase activity more than 1200-fold. The N terminus of BfpC alone increased the ATPase activity to 58.0 µmol of released Pi min-1 mg-1 of protein and BfpE77-114 further increased it to 75.6 µmol of released Pi min-1 mg-1 of protein. Consistent with these results, the N terminus of BfpC and BfpE77-114 together increased the Vmax of BfpD over 1200-fold, with the N terminus of BfpC alone increasing it almost 1000-fold (Table IV). In addition, the N terminus of BfpC reduced the Km 6-fold (Table IV), indicating an increase in the affinity of BfpD for ATP in the presence of BfpC. The stimulation of ATPase activity only occurred in the presence of Zn2+ (Fig. 7B), highlighting the importance of utilizing optimal reaction conditions. Although the total Zn2+ content of E. coli is in the millimolar range, evidence suggests the free Zn2+ concentration is femtomolar (50). Therefore, we examined the ATPase activity of BfpD with and without BfpC and BfpE across a broad range of Zn2+ concentrations. Remarkably, under both conditions there was only a 2% change in activity over Zn2+ concentrations that spanned many orders of magnitude. Thus, the BfpC- and BfpE-mediated enhancement of BfpD activity still occurred in the presence of trace levels of Zn2+ (data not shown). In the presence of other divalent cations, ATPase activity was enhanced only 2-fold by the N terminus of BfpC, BfpE39-76, and BfpE77-114. Interestingly, BfpE77-114 alone and BfpE39-76 did not stimulate ATPase activity (Fig. 7A). Furthermore, the N terminus of BfpC and BfpE77-114 had similar allosteric effects on the delayed ATPase activity of monomeric BfpD (data not shown). In conclusion, the N terminus of BfpC and BfpE77-114 induce conformational changes that greatly enhance the ATPase activity of hexameric BfpD.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
S, TNP-ATP, ADP, and TNP-ADP). A combination of gel filtration, zonal sedimentation, DLS, and electron microscopy confirmed that BfpD forms ring-shaped hexamers in the presence of nucleotide (ATP, ATPS, and ADP). The ability to form hexameric rings has been reported for other PulE superfamily members (33, 34, 37, 52-54) and seems to be a common characteristic of members of this superfamily. These hexameric ring assemblies are reminiscent of those formed by nucleotide-dependent molecular motors such as helicases and F1-ATPases (52, 54, 55), and membrane fusion ATPases such as N-ethylmaleimide-sensitive fusion protein and p97 (56, 57). The ATPase activity measured for purified hexameric BfpD (Vmax of 62.2 nmol of Pi released min-1 mg-1 of protein) is similar in magnitude to the highest values previously measured for other members of the PulE superfamily (ranging from mg 0.71 to 61 nmol of Pi released min-1 mg-1 of protein). We considered the weak ATPase activities of these PulE superfamily members to be because of the absence of stimulatory proteins of the cognate systems and/or the reaction conditions used. After optimizing our reaction conditions, we tested whether components of the BFP biogenesis apparatus that interact with BfpD stimulate its ATPase activity. We have previously shown that BfpD interacts with the cytoplasmic N termini of BfpC and BfpE (BfpE1-114) and is recruited to the cytoplasmic membrane by both BfpC and BfpE (29). Here we further define the BfpD-binding sites of BfpE and show that BfpD binds to one of two sites in BfpE depending on the stage of the catalytic cycle. The ATP-bound form of BfpD binds with high affinity (Kd = 72 nM) to BfpE77-114 and the ADP-bound form of BfpD binds with high affinity (Kd = 75 nM) to BfpE39-76. We found that the N terminus of BfpC and BfpE77-114 together stimulate the ATPase activity of BfpD more than 1200-fold to a Vmax of 75.3 µmol of Pi released min-1 mg-1 of protein. This activity exceeds that of any member of the family studied thus far by over 1200-fold. Moreover, to our knowledge BfpD is the first component of a Tfp biogenesis machinery or similar system whose ATPase activity has been shown to be modulated by other components of the system. The N terminus of BfpC alone increased the ATPase activity almost 1000-fold to a Vmax of 58.7 µmol of Pi released min-1 mg-1 of protein. However, BfpE77-114 only increased ATPase activity in the presence of the N terminus of BfpC. Furthermore, BfpE39-76 had no detectable effect on BfpD ATPase activity, a result that is not surprising, as BfpD-ATP does not interact with BfpE39-76. The binding of BfpC and BfpE to BfpD induce conformational changes in each of the proteins as indicated by protease sensitivity experiments conducted in this and prior studies (29). The allosteric effects of BfpC and BfpE on BfpD ATPase activity indicate that BfpC and BfpE indirectly play a major role in supplying the BFP biogenesis machine with the energy it requires for its assembly, bundlin polymerization. and/or transmembrane transport. Not only do they recruit BfpD to the biogenesis machinery, they also dramatically increase its ATPase activity. In this way, the ATPase activity of BfpD is regulated to be maximal at the site where energy is needed for BFP formation, making the process of producing BFP energy efficient and minimizing wasted energy expenditure by the free cytoplasmic enzyme. Interestingly, monomeric BfpD preparations exhibited delayed ATPase activity that was greatly increased by the N terminus of BfpC and BfpE77-114 but not BfpE39-76. This ATPase activity was most likely that of BfpD hexamers formed in the presence of the ATP added in the assay, because monomeric BfpD does not interact with the N terminus of BfpC, BfpE39-76, or BfpE77-114. Our data suggest that nucleotide binding induces formation of BfpD hexamers, which are selectively recognized, recruited, and stimulated by membrane-bound BfpC and BfpE, thus maximizing the energy supplied to the BFP biogenesis machine.
We found that hexameric BfpD binds to a maximum of three molecules of BfpE77-114 in the presence of ATP, but can bind to six molecules of BfpE39-76 in the presence of ADP. This difference in stoichiometry is not because of differences in nucleotide saturation, as BfpD is able to bind six molecules of either ATP or ADP. Rather these data suggest that occupancy of the ATP-activated BfpD hexamers by BfpE occurs at every other subunit and induces a conformational change that precludes binding of adjacent subunits.
Hexameric BfpD exhibited the highest ATPase activity in the presence of Zn2+ and the dramatic stimulation of its activity by the N terminus of BfpC and BfpE77-114 occurred only in the presence of Zn2+. The crystal structure of a truncated form of the closely related type II secretion ATPase EpsE (missing ninety N-terminal amino acids) of V. cholerae has been solved, revealing a molecule with four domains, N2, C1, CM, and C2, and inferring a missing N1 domain. Electron density most consistent with the presence of a Zn2+ ion was detected in the CM domain, coordinated by four cysteines. This location is remote from the active site. An alignment of BfpD and EpsE reveals that these cysteine residues are conserved. It is not clear therefore, whether Zn2+ is catalytically relevant or merely structurally relevant for either EpsE or BfpD. For example, Zn2+ binding might be necessary for conformational changes that allow binding of BfpE or BfpC. However, this hypothesis is not supported by our ITC data, which were acquired in the presence of Mg2+, not Zn2+. Therefore, we propose that the Zn2+ ion might perform a critical catalytic role for BfpD and perhaps for other members of the PulE ATPase family. In our assays, we purified hexameric BfpD in buffer lacking divalent metal ions and measured ATPase activity in the presence of buffer containing various divalent metal ions. We cannot be certain that trace quantities of other metals were not present in any of the buffers used. However, we found using inductively coupled plasma optical emission spectroscopy that the enzyme purified in the absence of Zn2+ already contained an equimolar ratio of Zn2+ (data not shown). Therefore, the fact that the addition of Zn2+ to the enzyme, which already contains Zn2+, stimulated activity more than 600-fold higher than any other metal, suggests that the Zn2+ plays a catalytic role in addition to a structural role. Furthermore, if the role of Zn2+ were merely conformational, then the structural analogue Cd2+ would be expected to be a reasonable substitute. Instead the activity in the presence of Cd2+ was less than 1/1000 of that in the presence of Zn2+ when BfpC and BfpE77-114 were also present. Thus, we propose that BfpD and perhaps other PulE ATPases use Zn2+ to perform a catalytic role in ATP hydrolysis, which to our knowledge is unprecedented for ATPases. However, Zn2+ is known to catalyze hydrolysis of phosphate bonds for other enzymes, for example, the hydrolysis of phosphodiester bonds by endonuclease IV (58), and therefore such a role is clearly plausible.
Based on the findings presented here, we envision the role of BfpD in BFP biogenesis is analogous to that of p97 and N-ethylmaleimide-sensitive fusion protein in homotypic membrane fusion and organelle biogenesis in that it transduces mechanical force to the BFP biogenesis machine possibly for the extraction of bundlin from the cytoplasmic membrane. This theory is supported by the following findings: (i) nascent bundlin is an integral cytoplasmic membrane protein (59) and therefore must be extracted from the membrane for pilus biogenesis; (ii) the N terminus of BfpE is involved in interactions with other essential Bfp proteins that are necessary for BFP formation (29); (iii) hexameric BfpD exhibits high ATPase activity in the presence of other essential components of the BFP biogenesis machine, consistent with the idea that it is a molecular motor; and (iv) hexameric BfpD displays differential binding to two sites of the N terminus of BfpE depending on the phosphorylation of its associated nucleotide. This last point indicates that the relative position of BfpD and BfpE change upon ATP hydrolysis. Thus, the resulting chemical energy can be converted to mechanical energy to change the position of BfpE and the proteins to which it binds. Fig. 8 illustrates the proposed mode of action of BfpD, showing the changes it induces in the cytoplasmic membrane subassembly of the BFP biogenesis machine upon hydrolysis of ATP. The figure is simplified to show the action of only a BfpD monomer, but presumably occurs simultaneously at alternating monomers. Hexameric BfpD-ATP interacts with the cytoplasmic N termini of BfpC and BfpE, which also interact with each other. BfpD-ATP binds to BfpE77-114, the domain of the N terminus of BfpE closest to the cytoplasmic membrane. Conformational changes in BfpD upon binding to the N terminus of BfpC and BfpE77-114 greatly increase its ATPase activity. Subsequently, the BfpD-ADP resulting from ATP hydrolysis shifts to BfpE39-76 while remaining bound to the N terminus of BfpC. It seems unlikely that BfpD moves to its new binding site in BfpE as its movement is restricted by its interaction with the N terminus of BfpC. Therefore we propose that the N terminus of BfpE is raised allowing BfpD-ADP to interact with BfpE39-76. This movement may force the most distal part of the N terminus of BfpE into the cytoplasmic membrane and perhaps into the periplasm. The interaction between the N terminus of BfpC and BfpE39-114 presumably limits the distance the N terminus of BfpE is pushed through the cytoplasmic membrane. According to this model, the BfpD motor energizes BfpE to act as a piston, whereas BfpC serves as a scaffold protein, anchoring BfpD and the N terminus of BfpE in place. Preliminary studies indicate that other essential Bfp proteins bind to BfpE,77-114,2 which could be transported across the cytoplasmic membrane in association with the N terminus of BfpE. In this way, BfpD provides the necessary mechanical force for recruitment and assembly of Bfp proteins into the BFP biogenesis machine, which results in the extraction of nascent bundlin from the cytoplasmic membrane and incorporation into the growing pilus filament. Finally, BfpD is replenished with ATP and released from BfpE39-76 and the cytoplasmic membrane subassembly of the BFP biogenesis machine relaxes, allowing the distal part of the N terminus of BfpE to slip back through the cytoplasmic membrane. Whereas further experiments are necessary to test this model of energy transduction during BFP biogenesis, the insights into the function and mode of action of BfpD we report here increase our understanding of the mechanism of Tfp biogenesis machines and have important implications for similar systems that transport biological macromolecules across membranes. We propose that this model is relevant to the entire PulE superfamily. Orthologues of BfpE (GspF) are highly conserved in systems that include PulE superfamily members (17). Although BfpC orthologues are not common, GspL and GspM proteins, which span the cytoplasmic membrane, interact with each other and recruit PulE family ATPases to the cytoplasmic membrane (41, 60), are widespread and likely play roles similar to that carried out by BfpC. Furthermore, the unique properties of BfpD suggest that drug therapy targeting the ATPase activity of PulE superfamily proteins and their interactions with proteins of their cognate systems could be designed to disrupt these macromolecular machines and the virulence properties they control.
|
|
FOOTNOTES |
|---|
¶ To whom correspondence should be addressed: Division of Infectious Diseases, Dept. of Medicine, University of Maryland School of Medicine, HSF II, 20 Penn St., Baltimore, MD 21201. Tel.: 410-706-7560; Fax: 410-706-8700; E-mail: mdonnenb{at}umaryland.edu.
1 The abbreviations used are: Tfps, type IV pili; BFP, bundle-forming pilus; DLS, dynamic light scattering; EPEC, enteropathogenic Escherichia coli; H, enthalpy change; ITC, isothermal titration calorimetry; Kav, partition coefficient; s20,w, sedimentation coefficient; SV40, simian virus 40; TNP, 2',3'-O-trinitrophenyl; X-Gal, 5-bromo-4-chloro-3-indolyl--D-galactopyranoside; Mops, 4-morpholinepropanesulfonic acid; ATPS, adenosine 5'-O-(3-thiotriphosphate).
2 L. J. Crowther, A. Daniel, and M. S. Donnenberg, unpublished results.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) and TNP-ADP (
) and the fluorescence enhancement (
