UreE-UreG Complex Facilitates Nickel Transfer and Preactivates GTPase of UreG in Helicobacter pylori*♦

Background: UreE and UreG play important roles in urease activation. Results: Nickel binding to UreG and its interactions with UreE as well as their roles in Ni2+ transfer were studied. Conclusion: UreG is a specific nickel-dependent GTPase; formation of (UreE)2-(UreG)2complex is a prerequisite for Ni2+ transfer from UreE to UreG. Significance: This study provides novel information on Ni2+ transfer among metallochaperones UreE, UreG, and HypA. The pathogenicity of Helicobacter pylori relies heavily on urease, which converts urea to ammonia to neutralize the stomach acid. Incorporation of Ni2+ into the active site of urease requires a battery of chaperones. Both metallochaperones UreE and UreG play important roles in the urease activation. In this study, we demonstrate that, in the presence of GTP and Mg2+, UreG binds Ni2+ with an affinity (Kd) of ∼0.36 μm. The GTPase activity of Ni2+-UreG is stimulated by both K+ (or NH4+) and HCO3− to a biologically relevant level, suggesting that K+/NH4+ and HCO3− might serve as GTPase elements of UreG. We show that complexation of UreE and UreG results in two protein complexes, i.e. 2E-2G and 2E-G, with the former being formed only in the presence of both GTP and Mg2+. Mutagenesis studies reveal that Arg-101 on UreE and Cys-66 on UreG are critical for stabilization of 2E-2G complex. Combined biophysical and bioassay studies show that the formation of 2E-2G complex not only facilitates nickel transfer from UreE to UreG, but also enhances the binding of GTP. This suggests that UreE might also serve as a structural scaffold for recruitment of GTP to UreG. Importantly, we demonstrate for the first time that UreE serves as a bridge to grasp Ni2+ from HypA, subsequently donating it to UreG. The study expands our horizons on the molecular details of nickel translocation among metallochaperones UreE, UreG, and HypA, which further extends our knowledge on the urease maturation process.

Helicobacter pylori is a pathogenic bacterial species that was discovered in human stomach and duodenal mucous membrane (1). This pathogen is responsible for gastritis, peptic ulcer, and even stomach cancer (2). To survive in the acidic environment of the human stomach, H. pylori produces large amounts of urease to turn urea to carbon dioxide and ammonia, which neutralizes the stomach acid (3). The physiological activities for survival in an extreme acidic environment consume huge amounts of energy, and thus H. pylori produces [Ni,Fe]hydrogenase to oxidize molecular hydrogen to obtain sufficient energy (4). Thus, the survival and successful colonization of H. pylori in human stomach largely rely on the proper function of two enzymes, i.e. urease (5) and [Ni,Fe]-hydrogenase (6). Maturation of both urease and [Ni,Fe]-hydrogenase refers to the assembly of nickel-containing active centers (7,8). The utilization of nickel must be tightly controlled in H. pylori as a plethora of nickel causes serious cell damage to H. pylori (9). Therefore, H. pylori has developed an elaborate system to tightly regulate the cellular nickel homeostasis from uptake, storage, and delivery to efflux through biosynthesis of a series of metalloproteins and chaperones (10), such as nickel storage protein Hpn and Hpnl (11)(12)(13)(14), which are only produced by this pathogen, and HspA (15,16), which has a distinct His-Cys-rich C terminus with nickel binding ability.
The proper assembly of the metal-containing active site of [Ni,Fe]-hydrogenase depends on a group of accessory proteins (HypA, -B, -C, -D, -E, -F) (8,17,18). HpHypA consists of zinc binding and nickel binding domains (22,23), and binds nickel with an affinity at a micromolar level (20). HpHypB is classified into the P-loop GTPase family with its enzymatic activity being activated by potassium (24). It is commonly believed that HypA and HypB are involved in the nickel insertion into the large subunit of hydrogenase (19 -21) via the formation of HypA-HypB complex (20,25). In addition, SlyD, a protein that binds Ni 2ϩ at its C terminus (28), might also participate in hydrogenase maturation through interaction with HypB (29,30). The IF (insert-in-flap) domain of SlyD is responsible for the SlyD-HypB interaction, facilitating nickel transfer from the C terminus of SlyD to HypB in H. pylori (31), but stimulates nickel release from HypB in Escherichia coli (29). Interestingly, both HypA and HypB are also involved in urease maturation (19,26,27), indicating a cross-talk between urease and hydrogenase maturation processes.
Maturation of urease involves the proper insertion of two nickel ions to its active site and requires at least four accessory proteins (UreE, -F, -G, -H) (9). In addition, accessory proteins for [Ni,Fe]-hydrogenase were also reported to be involved in the process (19,27). UreE can form a ternary complex with HypA, consisting of one UreE dimer and one HypA monomer.
Such an interaction facilitates nickel translocation from HypA to UreE and enhances urease activity (32).
Maturation of urease in H. pylori is not well understood. The formation of apo-urease-(UreG/UreF/UreH) was thought to be a crucial step for nickel insertion to urease (33,34). It was also demonstrated that UreE-UreG complexation plays an important role in the maturation of apo-urease; however, the molecular mechanism of nickel transfer between the chaperones UreE and UreG is not fully understood. Previously, UreG from H. pylori was demonstrated to bind Zn 2ϩ , instead of Ni 2ϩ (35). Recently, the structure of HpUreG in an UreG/F/H complex revealed the metal binding site, consisting of two cysteines and two histidines, which are located at the interface between two UreG monomers. The nickel binding induces GTP-dependent dimerization of UreG (36). A previous isothermal titration calorimetry study showed that H. pylori UreE and UreG form a (UreE) 2 -(UreG) 2 complex (i.e. one UreE dimer with one UreG dimer, or as 2E-2G) and the Zn 2ϩ but not Ni 2ϩ binding stabilized the complex (37), which is quite peculiar given the role of UreE and UreG in nickel delivery. Moreover, a previous study showed that in Klebsiella aerogenes, a (UreE) 2 -UreG (i.e. 2E-G, one UreE dimer with one UreG monomer) complex was observed (38). It is not clear how nickel transfer is achieved through the interaction of UreE and UreG.
In this study, we have overexpressed and purified UreG and UreE from H. pylori 26695. By using biochemical, biophysical, and molecular biology approaches, we characterize Ni 2ϩ binding properties as well as GTPase activity of UreG. We also carried out detailed studies on the molecular mechanism of the interaction between UreE and UreG. We demonstrate that UreE and UreG form a 2E-2G complex, which preactivates GTPase of UreG through enhancement of GTP binding to facil-itate Ni 2ϩ transfer. The roles of UreE, UreG, and HypA in nickel delivery are examined and discussed.

EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases and T4 DNA ligase were purchased from New England Biolabs. Primers were synthesized by Life Technologies and are listed in Table 1. Phusion High-Fidelity DNA polymerase was obtained from Finnzymes. KOD Hot Start DNA polymerase was obtained from Toyobo Life Science. All chromatography columns and FPLC system were from GE Healthcare. All chemical reagents were purchased from Sigma, unless stated otherwise. All solutions were prepared with Milli-Q water (18.2 megaohms).
Protein Expression and Purification-UreE and its variants were overexpressed and purified similarly as described previ- The following steps were carried out at 4°C unless stated otherwise. Cells were harvested by centrifugation (4,000 ϫ g, 30 min) and resuspended in buffer A (20 mM Hepes containing 500 mM NaCl, pH 7.5) with 1 mM PMSF as an enzyme inhibitor. After cell lysis by sonication, supernatant was separated from pellets by centrifugation (16,000 ϫ g, 30 min). The supernatant was further filtered through Millex-HA filter (0.45 m, 33 mm, Millipore) and loaded onto a 5-ml HisTrap column (GE Healthcare) preloaded with nickel ion and pre-equilibrated by buffer A containing 50 mM imidazole. Five column volumes of buffer A with 50 mM imidazole were used to wash away the unspecific binding impurities. The His 6 -UreG protein was eluted by buffer A supplemented with 300 mM imidazole. The eluted protein fraction was collected and loaded into a 5-ml HiTrap desalting column (GE Healthcare) pre-equilibrated by buffer B (20 mM Hepes, 100 mM NaCl, pH 7.5) to remove excess imidazole for thrombin cleavage. Fusion protein was incubated with ϳ50 NIH units of thrombin at 25°C for 3 h with gentle shaking to cleave His 6 tag from free UreG. Trace amounts of uncleaved fusion protein were removed by loading the digested protein sample to the 5-ml HisTrap column again. Fraction of flowthrough was collected and incubated with 20 mM EDTA and 1 mM DTT at 4°C overnight to obtain the apo-form UreG protein, which was further purified by gel filtration using a HiLoad 16/60 Superdex 75 column (GE Healthcare) equilibrated with buffer C (20 mM Hepes, 300 mM NaCl, pH 7.2) containing 500 M tris(2-carboxyethyl) phosphine (TCEP) 2 as a reducing agent. Elution fractions of purified protein were collected and concentrated to ϳ2 ml. UreG variants were purified similarly. Protein concentration was determined by the BCA protein assay kit (Novagen), and metal contents of the purified proteins were determined by inductively coupled plasma-mass spectrometry (Agilent 7500a).
UV-visible Spectroscopy-Nickel binding to UreG was monitored by UV-visible spectroscopy. All UV-visible spectra were recorded on a Cary 50 UV-visible spectrometer using a 1-cm quartz cuvette at ambient temperature (ϳ25°C). UV-visible spectra were scanned with a rate of 360 nm/min from 600 to 240 nm. Apo-UreG (20 M) was freshly prepared in 20 mM Hepes, 100 mM NaCl, 500 M TCEP, pH 7.2, with or without supplementation of 100 M GTP/GDP and 1 mM Mg 2ϩ . Spectrum of the same buffer was used for baseline correction. Aliquots of nickel chloride stock solution (10 mM) were titrated into UreG sample with gentle mixing, and UV-visible spectra were recorded after a 10-min incubation at room temperature.
For competition between the Ni 2ϩ and Zn 2ϩ experiments, UreG (20 M) was freshly prepared in 20 mM Hepes, 100 mM NaCl, 500 M TCEP, pH 7.2, in the presence of 40 M Ni 2ϩ or Zn 2ϩ , 100 M GTP, and 1 mM Mg 2ϩ , and subsequently, Zn 2ϩ or Ni 2ϩ stock solutions were titrated into the above protein solution. The UV titration curves were fitted to the Ryan-Weber nonlinear equation (39) where I stands for UV absorbance intensity; I max stands for maximal UV absorbance; C p and C m refer to the total concentrations of proteins and ligands respectively; and K d is the dissociation constant.
To examine the nickel transfer between UreE and UreG, Ni-UreE (or its variants Ni-CUreE and Ni-UreE-R101A) was Analytic Gel Filtration Chromatography and Light Scattering-Analytical gel filtration chromatography was performed on an Ä KTA FPLC system using a Superdex 75 or 200 30/100 GL column (GE Healthcare) precalibrated with the low molecular weight gel filtration calibration kit (GE Healthcare) at 4°C with a flow rate of 0.5 ml/min. Typically, 40 M apo-UreG was preincubated with 1.5 molar eq of Ni 2ϩ /Zn 2ϩ , GTP/GDP, Mg 2ϩ , or their combinations for 15 min and subjected to analysis. The column was preequilibrated with Hepes buffer (20 mM Hepes, 100 mM NaCl, 200 M TCEP, pH 7.2). The molecular mass of the protein complex in the elution peaks was further measured by a multiple-angle laser light scattering detector (miniDAWN light scattering detector, Wyatt Technology) or a by DynaPro plate reader (Wyatt Technologies), and the data were analyzed by Astra version 5.3.4.18 (Wyatt Technology) or by the Dynamics software (Wyatt Technologies).
To investigate the interaction among HypA, UreE, and UreG, 50 M apo-UreG was preincubated with 50 M apo-UreE and 300 M Zn-HypA in the absence or presence of 75 M GTP and MgSO 4 (HypA protein was used in excess due to low absorption at 280 nm). The fractions of elution peaks were collected and subjected to 15% SDS-PAGE assay to determine the components of peaks.
GTPase Activity Assay-GTPase activity of UreG was determined by the Malachite Green phosphate assay kit (Abcam). To compare the effect of K ϩ , HCO 3 Ϫ , and KHCO 3 on the GTPase activity of Ni-UreG, GTPase assay was carried out with the supplementation of KCl/NH 4 Cl, NaHCO 3 , or KHCO 3 / NH 4 HCO 3 at concentrations of 0, 0.5, 1, 5, 10, 20, 30, 40, 50, 100 mM, and the pH of the buffer was gently adjusted to be 7.5. GTP hydrolysis was triggered by the addition of Ni-UreG (10 M) into KCl/NaHCO 3 /KHCO 3 /NH 4 HCO 3 buffer with incubation at 37°C for 20 min. To compare the GTPase activities of apo-UreG, Zn-UreG, and Ni-UreG at different pH, the GTPase assay was carried out in a series of buffers with different pH values supplemented with 10 mM KCl/NaHCO 3 /KHCO 3 . After incubation at 37°C for 40 min, the free phosphate from hydrolysis of GTP determined by the kit was used to calculate the percentage of GTP turnover. Similarly, to compare the effect of HypA and UreE on GTPase activity of UreG, apo-UreG was prepared and supplemented with Ni-HypA, Ni-UreE, or Ni-HypA-UreE (10 M for each protein) in GTPase assay buffer (20 mM Hepes, 100 mM NaCl, 1 mM MgCl 2 , 100 M GTP, 1% glycerol, pH 7.5). Hydrolysis of GTP was triggered by the addition of 10 mM KHCO 3. After incubation at 37°C for 40 min, the free phosphate from hydrolysis of GTP was determined. For the time course reaction, a series of 400 l of 10 M UreG protein samples was prepared in GTPase assay buffer and was incubated with 10 mM KHCO 3 at 37°C. For each time point, 50 l of reaction mixtures was taken out to allow the free phosphate from hydrolysis of GTP to be determined. Those solutions containing 100 M GTP but without UreG were also incubated as blank controls to eliminate the self-hydrolysis of GTP. To determine the enzyme kinetic parameters (K m , k cat ) of UreG, similar experiments were carried out but with the substrate concentration varied. A series of 50 l of 10 M UreG protein samples containing various GTP concentrations ranging from 10 to 500 M was prepared in GTPase assay buffer with 10 mM KHCO 3 . After incubation at 37°C for 40 min, the amounts of free phosphate in the reaction mixtures were determined. The Michaelis-Menten equation was used to obtain kinetic parameters by nonlinear fitting.
Urease Activity Assay-pET-UreA2H⌬G was transformed into E. coli strain KMl603 (BL21(DE3) ⌬slyD::kan (a gift from Prof. A. R. Davidson, University of Toronto)). The bacterial cells were cultured in Luria Broth with supplementation of 100 g/ml ampicillin at 37°C and induced at 25°C overnight with 0.2 mM isopropyl ␤-D-1-thiogalactopyranoside when A 600 reached 0.8. Cells were collected and washed with 50 mM Hepes buffer, pH 7.5, lysed by sonication, and centrifuged to obtain supernatant. About 50 l of lysate was mixed with 10 M Ni-UreG/apo-UreG or 5 M Ni 2ϩ . After supplementation of 250 l of urea buffer (50 mM Hepes, 25 mM urea, pH 7.5) with or without 10 mM KHCO 3 , the mixtures were incubated at 37°C for 1 h to activate urease. Urease activity was measured by the amounts of ammonia released using the phenol-hypochlorite assay. Total protein concentration of lysate was determined by the BCA protein assay kit (Novagen), and the unit of urease activity was defined as nmol of ammonia produced per min/mg of total protein.

RESULTS
Ni 2ϩ Binding Property of HpUreG-UreG has been previously believed to be a specific Zn 2ϩ chaperone with no Ni 2ϩ binding capacity (27), whereas the report studying the UreG/ UreF/UreH complex clearly showed that UreG binds Ni 2ϩ (36). To further investigate the nickel binding property of HpUreG, the recombinant H. pylori apo-UreG was titrated with Ni 2ϩ and monitored by UV-visible spectroscopy. The addition of Ni 2ϩ to freshly prepared apo-UreG (20 M) in 20 mM Hepes containing 1 mM Mg 2ϩ leads to no observable new absorption bands (data not shown), indicative of no binding of Ni 2ϩ to the protein under such a condition. However, upon supplementation of 100 M GTP (Fig. 1A) but not GDP (Fig. 1B), the stepwise titration of Ni 2ϩ induces the appearance and gradual increases of a peak centered at ϳ337 nm, assignable to the (S)(Cys)3 Ni(II) ligand-to-metal charge transfer transition. Also, the intensities of this band are leveled off at a molar ratio of [Ni 2ϩ ] to [UreG] of over 0.5, indicating that each monomer of UreG binds 0.5 molar eq of Ni 2ϩ , in good agreement with a previous report that each UreG dimer binds one Ni 2ϩ (36). By nonlinearly fitting the plot of absorption at 337 nm versus Ni 2ϩ concentrations to the Ryan-Weber equation, the dissociation constant (K d(Ni-UreG) )is determined to be 0.36 Ϯ 0.05 M. A similar titration experiment was carried out except in the absence of Mg 2ϩ , which shows that UreG still binds Ni 2ϩ , but with a dissociation constant of ϳ10-fold larger (3.50 M), and complete saturation of UreG by Ni 2ϩ is achieved at ϳ2 molar eq of Ni 2ϩ to the protein under this condition (Fig. 1C).
A previous x-ray absorption spectroscopy study showed that the coordination sphere for zinc is formed by residues Cys-66 and His-68 from each of the monomers of HpUreG dimer (40). To investigate whether Cys-66 and His-68 are also involved in nickel binding, we constructed, overexpressed, and purified two UreG variants, UreG-C66A and UreG-H68A, and similarly carried out Ni 2ϩ titration experiments. Mutation of either Cys-66 or His-68 to alanine abolishes Ni 2ϩ binding as judged from the disappearance of the absorption peak at 337 nm (data not shown), suggesting that both Cys-66 and His-68 are essential for nickel binding.
Competitive binding of Zn 2ϩ and Ni 2ϩ to UreG in the presence of GTP was further studied. The addition of Zn 2ϩ to Ni 2ϩ -UreG has little effect on the absorption peak at 337 nm (data not shown), indicating that Zn 2ϩ cannot replace Ni 2ϩ from UreG in the presence of GTP. In contrast, titration of Ni 2ϩ into Zn 2ϩ -UreG solution results in the appearance of the absorption peak at 337 nm, a characteristic band for nickel binding to sulfur (Fig. 1D), suggesting the displacement of Zn 2ϩ by Ni 2ϩ from the protein. These results demonstrate that HpUreG has a higher binding affinity toward Ni 2ϩ than Zn 2ϩ in the presence of GTP and Mg 2ϩ .
Potassium Bicarbonate Boosts GTPase Activity of Ni 2ϩ -HpUreG-It was demonstrated that GTPase of Ni-HpUreG dimer is stimulated by bicarbonate (HCO 3 Ϫ ) (36), whereas the HpUreG counterpart, HpHypB, shows a higher GTPase activity in solution containing potassium (K ϩ ) (24), implying the potential roles of K ϩ and HCO 3 Ϫ in GTP hydrolysis. To verify this hypothesis, GTPase assay was carried out in a buffer containing 1 mM MgSO 4 supplemented with KCl, NaHCO 3 , KHCO 3 , or NH 4 HCO 3 at concentrations of 0, 0.5, 1, 5, 10, 20, 30, 40, 50, 100 mM, and the pH of the buffer was adjusted carefully to be 7.5. Consistent with a previous study (35), Ni 2ϩ -UreG does not exhibit GTPase activity in the absence of HCO 3 Ϫ (data not shown), whereas the activity of Ni 2ϩ -UreG rises slightly with increasing amounts of HCO 3 Ϫ and is saturated at only ϳ0.7 nmol/min/mg of protein ( Fig. 2A). Similarly, supplementation of gradient amounts of either K ϩ ( Fig. 2A) or ammonia (NH 4 ϩ ) (data not shown) results in elevated GTPase activities, which saturate at ϳ1.1 nmol/min/mg of protein. Interestingly, the GTPase activity of Ni 2ϩ -UreG is boosted upon the addition of potassium bicarbonate (KHCO 3 Ͼ 10 mM) or ammonium bicarbonate (data not shown) in a similar concentration gradient to a relatively high level with a rate of ϳ1.6 nmol/min/mg of protein. These results indicate that HpUreG may employ potassium bicarbonate or ammonium bicarbonate as GTPase elements to achieve full GTPase activity. Upon GTP hydrolysis, Ni 2ϩ was gradually released from Ni-UreG as judged from gradual losses of the ligand-to-metal charge transfer transition at 337 nm (data not shown).
The effect of pH on the GTP hydrolysis by HpUreG was also investigated in the absence or presence of 10 mM KHCO 3 /K ϩ / HCO 3 Ϫ . Both apo-UreG and Zn-UreG show negligible GTPase activity at the pH range studied (Fig. 2, C and D). Upon supplementation of either K ϩ (10 mM) or HCO 3 Ϫ (10 mM) to the buffers, low levels of activity were detected for Ni-UreG and  increased with pH from 5.0 to 8.5, with only ϳ10 -20% of GTP turnover for Ni-UreG. Further increases in pH values lead to decreases in GTPase activity (Fig. 2B). However, in the presence of KHCO 3 , the catalytic activity of Ni-UreG is enhanced dramatically with ϳ40% of GTP turnover, with the highest activity attained at pH 7.5-8 (Fig. 2B).
We further explored the role of GTPase activity of UreG in urease maturation. Urease activation assay was performed by transforming the plasmid pET-UreA2H⌬G containing the full set of urease genes ureABIEFH (except ureG) into E. coli. UreG protein or Ni 2ϩ was added into cell lysate in the urease activation assay. In the absence of KHCO 3 , only weak urease activity (Ͻ100 units) was detected in all cases, whereas in the presence of KHCO 3 , ϳ5-fold increases in urease activity were detected for Ni-UreG but not for Ni 2ϩ or apo-UreG (Fig. 2E), confirming that GTP hydrolysis by UreG is essential for urease activation.
Formation of (UreE) 2 -UreG or (UreE) 2 -(UreG) 2 Complexes-Complexation of UreE with UreG under near physiological conditions was examined by gel filtration chromatography at pH 7.2. As shown in Fig. 3A, apo-UreE and apo-UreG are eluted at ϳ13.5 and 13.9 ml, respectively (Fig. 3A, lines 1 and 2), corresponding to molecule masses of ϳ40 and 23 kDa, i.e. an UreE dimer and an UreG monomer, respectively. Mixing of equimolar amounts of apo-UreE and apo-UreG resulted in the disappearance of UreE peak and the appearance of a new peak at ϳ12.5 ml with a molecular mass of 62 Ϯ 1.6 kDa as determined by light scattering, which can be assigned to (UreE) 2 -UreG complex (2E-G), consisting of a UreE dimer and a UreG monomer (Fig. 3A, line 3). Similarly, we also investigated the effect of Ni 2ϩ , Zn 2ϩ , GTP, GDP, or their combinations on the complexation of UreE and UreG. Unexpectedly, neither Ni 2ϩ nor Zn 2ϩ perturbs the formation of UreE-UreG complex with 2E-G complex being observed in in all cases (Fig. 3A). We further carried out similar experiments but with the supplementation of Mg 2ϩ (60 M) to the reaction buffer. Surprisingly, supplementation of GTP and Mg 2ϩ (60 M for each) leads to the emergence of a single peak at ϳ11.8 ml with a molecular mass of 83 Ϯ 1.4 kDa by light scattering, suggesting the formation of a complex of a UreE dimer and a UreG dimer, i.e. 2E-2G, (Fig.  3B, lines 5, 7, and 9). In conclusion, we show that neither Ni 2ϩ nor Zn 2ϩ perturbs the formation of 2E-G or 2E-2G complexes. Importantly, both GTP and Mg 2ϩ are indispensable for the formation of 2E-2G complex, whereas a 2E-G complex was observed under all conditions without both GTP and Mg 2ϩ .

Arg-101 of UreE and Cys-66 of UreG Stabilize 2E-2G
Complex-Crystal structures of UreE from H. pylori and other species (41-45) reveal a similar homodimer architecture for apo-UreE, consisting of an N-terminal (NUreE) and a C-terminal domain (CUreE) with the latter housing the metal binding site and being responsible for specific HypA-UreE interaction (32). To examine which domain of UreE is responsible for UreE-UreG interaction, both NUreE and CUreE proteins (40 M) were overexpressed and purified as described previously (32) and then incubated with 1 molar eq of UreG, and the potential protein complexes were monitored by analytic gel filtration chromatography. Surprisingly, upon mixing of either NUreE or CUreE with UreG, no new peaks corresponding to the complexes such as UreG-NUreE or UreG-CUreE were observed, even in the presence of GTP and Mg 2ϩ (data not shown), implying that the scaffold formed between NUreE and CUreE may be essential for the UreG-UreE interaction.
To identify the residues of UreE participating in UreE-UreG interaction, we constructed a series of UreE variants according to two criteria. First, the residues involved in or around the metal binding site were considered in view of their potential roles in UreE-UreG interaction for nickel delivery to urease. Second, highly conserved residues in the C-domain were identified by the alignment of UreE sequences from several species via Clustal W (46). On the basis of these criteria combined with examination of the crystal structure of HpUreE (41,42), residues for mutagenesis were selected as follows: Asn-100, Arg-101, His-102, His-152, and residues 158 -70. Both Asn-100 and Arg-101 are conserved residues and close to the metal binding site. Both His-102 and His-152 are involved in nickel binding. Residues 158 -170 are close to the metal binding site upon nickel binding to His-152 and are likely responsible for HypA-UreE interaction (32).
UreE variants were prepared and incubated with wild-type UreG prior to gel filtration analysis. As shown in Fig. 4 (black  curves), all mutants of UreE show negligible effect on the for- mation of 2E-G complex, whereas the formation of 2E-2G complex is not affected by mutations of Asn-100, His-102, His-152, and ⌬158 -170 (Fig. 4, lines 1, 3, 4, and 5) in the presence of GTP and Mg 2ϩ . However, mutation of Arg-101 of UreE to alanine leads to the majority of its mixture with UreG being eluted as 2E-G complex even in the presence of GTP and Mg 2ϩ , indicating that residue Arg-101 may play an important role in the stabilization of 2E-2G complex (Fig. 4, line 2, green curve).
Interestingly, the addition of excess Ni 2ϩ (120 M) in the buffer can prompt the mixture of UreE-R101A with UreG to be eluted out as 2E-2G complex (Fig. 4, line 2, broken line). Two UreG mutations, UreG-C66A and UreG-H68A, were also prepared to examine their roles on the interaction between UreE and UreG. H68A from UreG shows little effect on the formation of both 2E-G and 2E-2G in the absence or presence of GTP and Mg 2ϩ , respectively (Fig. 4, line 7), similar to the wild-type UreG. However, for C66A mutant, mixing UreE and UreG-C66A apo-proteins results in the formation of 2E-G complex both in the absence or in the presence of GTP and Mg 2ϩ , but not 2E-2G complex (Fig. 4, line 6), implying that Cys-66 plays a crucial role on the stabilization of 2E-2G complex. Unlike UreE-R101A-UreG, even in the presence of excess Ni 2ϩ , no elution peak corresponding to 2E-2G complex can be observed for the UreE-UreG-C66A mixture (data not shown).
Formation of 2E-2G Complex Is a Prerequisite for Ni 2ϩ Transfer from UreE to UreG-To monitor nickel translocation between UreE and UreG, Ni-UreE was titrated stepwise into UreG (25 M) in the presence of GTP (50 M) and Mg 2ϩ (1 mM). A typical absorption at 337 nm appeared and increased its intensity with increasing amounts of Ni-UreE, suggesting that Ni 2ϩ is transferred from UreE to UreG (Fig. 5A) as Ni-UreE itself exhibits no UV absorption in this region (data not shown). In contrast, when apo-UreE was stepwise added into Ni-UreG under the same condition, no perturbation on the peak corresponding to nickel binding to UreG was observed (Fig. 5B), indicating that nickel transfer between UreE and UreG is unidirectional.
To examine whether Ni 2ϩ transfer is achievable upon UreE-UreG interaction, we carried out similar titration experiments under the same condition except using Ni-CUreE. The stepwise addition of Ni-CUreE does not induce the peaks at 337 nm (Fig.  5C), indicative of no Ni 2ϩ transfer between CUreE and UreG. To examine whether Ni 2ϩ transfer is achievable upon the formation of 2E-G complex, we carried out similar experiments using Ni-UreE-R101A. The addition of Ni-UreE-R101A to UreG in the presence of both GTP and Mg 2ϩ also led to poor nickel transfer from UreE to UreG (Fig. 5D). In parallel, the stepwise addition of Ni-UreE to UreG in the buffer without Mg 2ϩ in which 2E-G complex is formed induces only very weak peaks at 337 nm (Fig. 5E), indicative of an inefficient Ni 2ϩ transfer between UreE and UreG upon the formation of 2E-G com-  plex. However, the further addition of Mg 2ϩ (100 M) into the reaction buffer enhanced the peak intensities at 337 nm dramatically (Fig. 5E, broken line), implying that the Ni 2ϩ transfer between UreE and UreG has been restored once 2E-2G complex is formed. Taken together, we demonstrate that the formation of 2E-2G complex is crucial for nickel translocation from UreE to UreG. We further examined the nickel transfer between UreE and UreG by analytical gel filtration chromatography. A series of 40 M UreE samples loaded with different molar ratios of nickel was incubated with 80 M apo-UreG in a buffer containing 100 M GTP and 1 mM MgSO 4 for 1 h to allow the potential nickel translocation and then subjected to gel filtration chromatography (Fig. 5F). With an increase of nickel loaded to UreE samples, the intensities of monomer peaks of UreG (ϳ13.9 ml) decrease accompanied by the increases in the intensities of UreG dimers (ϳ13 ml) in the presence of both GTP and Mg 2ϩ , confirming that Ni 2ϩ has been transferred from UreE to UreG because UreG presents as a monomer in the absence of Ni 2ϩ , but presents as a dimer in the presence of higher molar ratios of Ni 2ϩ (36).
UreE Enhances GTPase Activity of UreG-To investigate the biological significance of UreE-UreG interaction, GTPase activity of UreG in the presence of UreE was examined. Similar to Ni 2ϩ -UreG (Fig. 6A, black curve), the amounts of GTP turnover increase with the addition of equimolar amounts of UreE loaded with increasing amounts of Ni 2ϩ (0, 0.3, 0.6, and 1.0 molar eq), suggesting that nickel is transferred from UreE to UreG, which enhances the enzyme activity (Fig. 6A). Surprisingly, the addition of 1 molar eq of Ni-UreE to UreG results in much more efficient activity of UreG (with ϳ65% of GTP turnover) than that of the same amounts of Ni 2ϩ -UreG in the absence of UreE (with ϳ30% of GTP turnover) (Fig. 6A, black  curve). Moreover, incubation of increasing amounts of apo-UreE (0, 0.3, 0.6, and 1.0 molar eq) with Ni 2ϩ -UreG (10 M) also enhances its activity (Fig. 6B), indicating that there may be an additional channel to elevate the GTPase activity of UreG by UreE besides nickel donation.
The GTPase activity of Ni-UreG upon supplementation of apo-UreE was further determined by fitting the saturation curve to the Michaelis-Menten equation (Fig. 6C) to give rise to k cat of 1.6 Ϯ 0.4 ϫ 10 Ϫ6 s Ϫ1 , K m of 27.9 Ϯ 1.8 ϫ 10 Ϫ6 M, and k cat /K m of ϳ57.3 ϫ 10 Ϫ3 M Ϫ1 s Ϫ1 . In comparison, the activity of the same amounts of Ni-UreG alone was also determined with a k cat of 0.61 Ϯ 0.1 ϫ 10 Ϫ6 s Ϫ1 , K m of 82.7 Ϯ 17.3 ϫ 10 Ϫ6 M, and a 8-fold lower enzyme efficiency (k cat /K m ϭ ϳ7.4 ϫ 10 Ϫ3 M Ϫ1 s Ϫ1 ). The lower K m for Ni-UreG and UreE mixture indicates that UreG has a higher affinity for GTP in the protein complex than UreG alone.
The effect of UreE on GTP binding to UreG was further investigated by analytical gel filtration chromatography. Different molar equivalents of UreE (0, 0.5, 1.0, and 1.5 molar eq) were incubated with UreG (40 M) in the presence of 40 M GTP and then subjected to gel filtration analysis. As shown in Fig. 6D, increases in UreE concentrations lead to the peak intensities of 2E-2G complex increasing, accompanied by decreases in the intensity of the GTP peak, indicating that UreE enhances GTP binding to UreG due to the formation of 2E-2G complex. Taken together, UreE promotes the enzyme efficiency of UreG not only by donation of nickel but also by serving as a structural scaffold for UreG to recruit GTP.
UreG Extracts Ni 2ϩ from HypA via UreE-We have shown previously that nickel is translocated from HypA to UreE via a direct protein-protein interaction (32). Here, we demonstrate that nickel is transferred from UreE to UreG through the formation of 2E-2G complex (Fig. 5). It is not well understood how nickel is delivered among these three proteins. HypA, UreE, and UreG were mixed and examined by analytic gel filtration, and the elution fractions were collected and further examined by 15% SDS-PAGE. As shown in Fig. 7A, in the absence of GTP and Mg 2ϩ , the protein mixture is eluted out as two major peaks at ϳ12.2 and ϳ14.0 ml (Fig. 7A, blue curve), with molecular weights corresponding to 2E-A complex, i.e. dimer of UreE complexed with monomer of HypA and UreG protein, respectively, implying that UreE dimer tends to bind HypA monomer rather than UreG under this condition. By contrast, in the presence of GTP and Mg 2ϩ , almost all UreG protein is eluted out with UreE as the 2E-2G complex, as evidenced by the observation of a peak at ϳ11.8 ml corresponding to a molecular mass of ϳ80 kDa. Another peak that eluted at ϳ14.8 ml corresponds to free HypA protein (Fig. 7A, orange curve), indicating that UreE dimer prefers to bind two UreG monomers instead of HypA in the presence of GTP and Mg 2ϩ . Apparently, no stable peaks with molecular weights corresponding to either HypA-UreE-UreG or HypA-UreG complexes were observed under both conditions (data not shown). The nickel translocation among  HypA, UreE, and UreG was further examined by GTPase assay and UV-visible spectroscopy. As shown in Fig. 7B, in the GTPase assay, apo-UreG displayed a poor GTPase activity with only ϳ4% of GTP turnover, whereas mixing of Ni-HypA with apo-UreG leads to the activity of UreG increased slightly with ϳ19% of GTP turnover, which may be attributable to UreG binding to the nickel ions dissociated from HypA. On the other hand, Ni-UreE enhances UreG activity dramatically with ϳ65% of GTP turnover. As expected, the addition of apo-UreE into the mixture of Ni-HypA and apo-UreG boosts the GTPase activity of UreG to a level as high as that of the Ni-UreE and UreG mixture, further verifying that UreG can acquire Ni 2ϩ from HypA through UreE. This was further confirmed by subsequent UV titration experiments. When Ni-Zn-HypA was stepwise added into apo-UreG, peaks corresponding to nickel binding to UreG were observed with nearly negligible intensities (Fig. 7C), indicating that nickel transfer between HypA and UreG is inefficient, whereas stepwise titration of a mixture of Ni-Zn-HypA and UreE into apo-UreG led to significant increases in the intensities of the absorption peaks at 337 nm (Fig. 7D), suggesting that nickel was efficiently transferred from HypA to UreG. However, when UreE was present at very low concentrations in the above experiments, Ni 2ϩ transfer was also inefficient, suggesting that UreE serves as a stoichiometric mediator for Ni 2ϩ transfer from HypA to UreG (data not shown).

DISCUSSION
The maturation of urease in H. pylori involves nickel delivery to the metallo-active center and is highly dependent on the cooperation of at least four urease accessory proteins, i.e. UreE, UreG, UreF, and UreH. In addition, hydrogenase accessory pro-teins HypA and HypB were also found to be crucial for activation of urease, with the former donating nickel ions to UreE by a specific protein-protein interaction (32). Among these chaperones, the nickel binding properties of UreE, UreF, HypA, and HypB have been well studied. Both HpHypA and HpHypB bind Ni 2ϩ with dissociation constants of micromolar levels (20,22,24). HpUreF dimer binds two Ni 2ϩ via His-229 and Cys-231 with a dissociation constant of 6.4 Ϯ 0.4 M (47). HpUreE binds to Ni 2ϩ via at least two His-102 residues from each monomer, with an additional residue (His-152) possibly also participating in the binding, and only one Ni 2ϩ ion binds per UreE dimer with a dissociation constant of 0.15 M (37). In contrast, there has been a lack of characterization of the nickel binding property of HpUreG. A previous study by isothermal titration microcalorimetry showed that HpUreG binds to Zn 2ϩ specifically, but not to Ni 2ϩ , with a dissociation constant of 0.33 M (35). However, a nickel-bound dimer of UreG has been observed in the study of a complex structure of UreG/F/H (36).
In this study, we demonstrate by UV-visible spectroscopy that HpUreG binds Ni 2ϩ only in the presence of GTP and Mg 2ϩ with a stoichiometry of ϳ0.5 Ni 2ϩ ions bound per UreG monomer, and a dissociation constant of ϳ0.36 M. In the absence of Mg 2ϩ , UreG exhibits a 10-fold lower binding affinity to Ni 2ϩ (K d ϳ3.6 M) and requires excess Ni 2ϩ to achieve full saturation (I max ϳ0.2), suggesting a potential communication between the nickel binding site and GTP binding site of UreG, which belongs to the family of magnesium-dependent enzymes (24).
It has been shown that both Ni 2ϩ and Zn 2ϩ bind UreG through coordination of Cys-66 and His-68 (40). Although the binding affinity of UreG to Ni 2ϩ determined in the present study is similar to that of Zn 2ϩ to HpUreG (K d ϭ 0.33 M) reported previously by isothermal titration calorimetry (35), our Ni 2ϩ -Zn 2ϩ competition experiments reveal that in the presence of GTP and Mg 2ϩ , UreG exhibits a higher affinity toward Ni 2ϩ than Zn 2ϩ , implying that UreG is a specific nickel binding chaperone. The discrepancy in the binding affinity obtained might be due to the different conditions and methods used. The UreE-UreG interaction and its role in nickel transfer have been investigated in this study. Our combined gel filtration chromatography and light scattering data show that apo-UreE and apo-UreG proteins form a 2E-2G complex only when both GTP and Mg 2ϩ are present; otherwise a 2E-G complex is formed. Neither Ni 2ϩ nor Zn 2ϩ perturbs the interaction between UreE and UreG (Fig. 3), which is in contrast with a previous study indicating that complexation of UreE with UreG gives rise to 2E-2G in the absence of GTP and Mg 2ϩ and that such a complex is specifically stabilized by excess Zn 2ϩ but not Ni 2ϩ (37). The reason for the discrepancy is not clear; however, considering that UreG is a GTPase, investigation of UreE and UreG interaction in the presence of GTP/Mg 2ϩ should provide more physiologically relevant information. Indeed, UreG from H. pylori does not associate with Ni 2ϩ in the absence of GTP (Fig. 1).
Further biophysical and molecular biology studies show that UreE serves as a structural scaffold to recruit UreG; deletion of either N-domains or C-domains of UreE abolishes the binding. Moreover, replacement of residue Arg-101 of UreE or Cys-66 of UreG to Ala results in disruption of the formation of 2E-2G complex (Fig. 4), implying their critical roles in the stabilization of 2E-2G complex. Interestingly, excess Ni 2ϩ can restore 2E-2G complex for UreE-R101A-UreG but not for UreE-UreG-C66A. This suggests that Ni 2ϩ might also play a certain, although not crucial, role in the stabilization of 2E-2G complex possibly by stabilizing the UreG dimer, which binds to GTP. Although the crystal structure of the UreE-UreG complex is not available, our mutagenesis studies imply that the protein-protein interfaces may be located around the metal binding sites of both UreE and UreG. A previous structural model of 2E-2G complex also revealed that metal binding sites of UreE and UreG "grazed" each other (37); such a structural arrangement might allow Ni 2ϩ translocation readily. Similarly, the metal binding site of UreG is buried in the known UreG/UreF/UreH complex (36), implying that UreE and UreF/H complex may compete with each other for UreG in the process of urease maturation.
Our combined UV-visible spectroscopy and gel filtration chromatography data reveal that Ni 2ϩ is transferred from UreE to UreG but not vice versa and that such a process occurs only when UreE and UreG form the 2E-2G but not 2E-G complex (Fig. 5), because mixing of UreG and a Ni 2ϩ -UreE variant with Arg-101 substituted by Ala abolishes the nickel transfer from UreE and UreG. Moreover, when 2E-G complex is formed in the absence of Mg 2ϩ , no efficient Ni 2ϩ transfer was observed between UreE and UreG (Fig. 5), indicative of the requirement of 2E-2G but not 2E-G complex for nickel delivery.
Previously, we demonstrated that nickel is translocated via a specific HypA-UreE interaction (32). Therefore, we investigated the roles of these three nickel chaperones in urease maturation process. We show that nickel is transferred from HypA to UreG via UreE, which acts as a stoichiometric instead of a catalytic mediator (Fig. 7) because inefficient Ni 2ϩ transfer was observed when UreE was present at low concentration.  F F F F F H   G G G G G G G G G  G G G G G G G G G G   F  G G G   F  G G G G G G G G G   H  G G G G G G G G G G G   F F F F F F F F F F F F F