The assembly of the plant urease activation complex and the essential role of the urease accessory protein G (UreG) in delivery of nickel to urease

Urease is a ubiquitous nickel metalloenzyme. In plants, its activation requires three urease accessory proteins (UAPs), UreD, UreF, and UreG. In bacteria, the UAPs interact with urease and facilitate activation, which involves the channeling of two nickel ions into the active site. So far this process has not been investigated in eukaryotes. Using affinity pulldowns of Strep-tagged UAPs from Arabidopsis and rice transiently expressed in planta, we demonstrate that a urease–UreD–UreF–UreG complex exists in plants and show its stepwise assembly. UreG is crucial for nickel delivery because UreG-dependent urease activation in vitro was observed only with UreG obtained from nickel-sufficient plants. This activation competence could not be generated in vitro by incubation of UreG with nickel, bicarbonate, and GTP. Compared with their bacterial orthologs, plant UreGs possess an N-terminal extension containing a His- and Asp/Glu-rich hypervariable region followed by a highly conserved sequence comprising two potential HXH metal-binding sites. Complementing the ureG-1 mutant of Arabidopsis with N-terminal deletion variants of UreG demonstrated that the hypervariable region has a minor impact on activation efficiency, whereas the conserved region up to the first HXH motif is highly beneficial and up to the second HXH motif strictly required for activation. We also show that urease reaches its full activity several days after nickel becomes available in the leaves, indicating that urease activation is limited by nickel accessibility in vivo. Our data uncover the crucial role of UreG for nickel delivery during eukaryotic urease activation, inciting further investigations of the details of this process.

Nitrogen is the mineral nutrient that plants require in the highest amounts and that often limits plant growth. Although urea accounts today for Ͼ50% of the N-fertilizer used worldwide (International Fertilizer Industry Association), only little is known about the direct utilization of fertilizer urea by plants.
Generally it is assumed that urea is fully degraded in the soil by microbes and that plants mainly take up ammonium and nitrate. However, urea can be actively transported and assimilated by plants (1)(2)(3)(4), indicating its significance for plant nutrition. Apart from external sources, urea also originates within plants as a product of arginine catabolism (5). The hydrolysis of urea to ammonia and carbonic acid is facilitated by urease, a nickel metalloenzyme found not only ubiquitously in plants (6) but also in many fungi and bacteria (7). The urease activation process, which is best characterized in Klebsiella aerogenes and Helicobacter pylori (8), requires the carbamylation of a lysine residue in the active site followed by incorporation of two nickel ions. In vivo, the three urease accessory proteins (UAPs) 2 UreD, UreF, and UreG are necessary for metallocenter assembly (9), and a series of complexes of UAPs with apo-urease were shown to exist (10). A fourth UAP, UreE, facilitates the activation process by acting as a nickel metallochaperone (8,11). Crystal structure data of a UreD-UreF-UreG complex from H. pylori showed at atomic resolution how two UreD (called UreH in this organism) interact with dimeric UreF to form a docking site for a UreG dimer (12). UreE delivers nickel to UreG (13,11), which is channeled via UreG, UreF, and finally UreD to urease (14). UreG acts as GTPase during this process but it is still not entirely clear why GTP hydrolysis is necessary.
In plants, orthologs of UreD, UreF and UreG were identified and shown to be required and sufficient for urease activation (15,16). A UreE homolog was not found in plants, but it was suggested that plant UreGs might additionally possess UreE function (17)(18)(19). This hypothesis was based on the observation that plant UreGs possess a His-and Asp/Glu-rich N terminus not present in the bacterial orthologs that might be involved in the binding and delivery of Ni 2ϩ ions. A similar role has been proposed for bacterial HypB proteins involved in maturation of Ni-Fe-hydrogenases. They possess GTPase activity and contain a His-rich region at the N terminus, which binds nickel ions (20).
In this study we analyzed the interactions between plant urease and its UAPs in detail and present a model for the formation of the activation complex. Urease activation was investigated in vitro using Strep-tagged urease-UAP complexes as well as in vivo expressing a series of N-terminal deletion variants of UreG in a ureG mutant background. Our data demonstrate that UreG is crucial for nickel delivery during the urease activation process, which requires several days in vivo.

Interaction of plant urease and its accessory proteins
To obtain a complete interaction map of a plant urease and its accessory proteins (UAPs), the respective Arabidopsis cDNAs were cloned into binary vectors for transient expression in Nicotiana benthamiana. For each UAP, two constructs were prepared allowing the expression of either untagged or N-terminally Strep-tagged proteins for affinity purification using StrepTactin Macroprep beads. A urease clone in pXCS-HAStrep (V13; Ref. 21) was constructed for the expression of the untagged enzyme. Employing this construct, only small amounts of protein could be generated. To optimize urease expression, bases immediately upstream of the AUG start site were modified to better match the start-site consensus of dicot plants (22). However, expression was not improved by this modification (supplemental Fig. S1). A significant boost of protein quantity could be achieved by expressing urease from a new vector termed pXCScpmv-HAStrep (V69) containing modified 5Ј-and 3Ј-UTRs of RNA-2 from the cowpea mosaic virus that act as translational enhancers (23; supplemental Fig. S1).
Urease and UAPs were co-expressed in N. benthamiana in different combinations with generally one of the UAPs Streptagged at the N terminus for affinity purification. Successful complementation of UAP mutants with the respective Streptagged UAPs in transgenic Arabidopsis plants showed that the tagged versions of these proteins are functional (supplemental Fig. S2). Purified and co-purified UAPs as well as co-purified urease were visualized in Western blots with anti-UAP antisera, newly developed by us, and a commercial anti-urease antibody. The new antisera specifically detected the respective UAPs (supplemental Figs. S2 and S3).
In the absence of urease, direct binding of UreD to UreF was observed (Fig. 1, lanes 18 and 19), demonstrating the formation of a UreD-UreF complex. UreG on the other hand was binding neither UreD (lanes 16 and 17) nor UreF (lanes 14 and 15) alone but was co-purified when expressed with UreD and UreF together (lanes [11][12][13]. These results show that UreG can interact with a UreD-UreF complex forming a UreD-UreF-UreG complex. When studying the binding of urease to the UAPs, a direct interaction with urease was only observed for UreD but not for UreF or UreG (lanes 20 -22). A direct interaction between urease and UreD had already been shown for the rice proteins (16). Co-purification of urease with tagged UreF was only possible in the presence of UreD (lane 10) and with tagged UreG only in the presence of UreD and UreF (lane 4).
These findings in dicotyledonous Arabidopsis prompted us to investigate the urease activation complex in rice as an important representative of the monocots. From leaves of N. benthamiana transiently expressing rice urease and the three rice UAPs (with UreD tagged), a urease-UreD-UreF-UreG complex could be purified (supplemental Fig. S4). Because orthologs of urease and of the UAPs are encoded by the genomes of all plants sequenced to date (16), it appears likely that the urease-UreD-UreF-UreG activation complex is universal in plants.
An assembly model derived from the co-purification data ( Fig. 2) suggests that there are two main routes toward the urease-UreD-UreF-UreG complex: (i) binding of UreD to urease followed by UreF and finally UreG or (ii) the formation of a UreD-UreF-UreG complex by binding of UreG to preformed UreD-UreF, which then can interact with urease. The formation of a urease-UreD-UreF complex by urease and UreD-UreF represents a third intermediate assembly route.
To assess if a preformed UreD-UreF-UreG complex interacts with urease as proposed in this model, such a complex was purified from N. benthamiana as described before but left bound on the StrepTactin matrix and then incubated with a leaf extract expressing only urease (Fig. 3). The matrix was washed, and the proteins were eluted with biotin. Urease could be recovered after incubation with a matrix loaded with UreD-UreF-UreG (Fig. 3, lane 5) but not in a control experiment with empty matrix (Fig. 3, lane 6) demonstrating that urease can bind to a preformed UreD-UreF-UreG complex. Figure 1. Protein interactions of A. thaliana UAPs and urease. Urease (U) and the accessory proteins UreD, UreF, and UreG (D, F, and G) were co-expressed in leaves of N. benthamiana as indicated. N-terminally Strep-tagged variants of the UAPs (the respective tagged protein is labeled with S) and its interaction partners were affinity-purified from extracts and detected on Western blots. Lower panel, clarified leaf extracts (input). Upper panel, after affinity purification. The experiment was repeated three times, and a representative repeat is shown.

Characterization of the plant urease activation complex Urease activation in vitro
The peat substrate used for growing N. benthamiana employed for transient expression usually did not supply sufficient nickel for urease activation in planta. Therefore, it was possible to purify inactive urease-UAP complexes from these plants. To investigate whether the urease-UreD-UreF-UreG complex is competent to activate urease, it was purified, and several in vitro activation assays were performed. In a first approach, the complex was incubated in a time course with high concentrations of nickel, bicarbonate, and GTP known to be required for bacterial urease activation (24), and then the resulting urease activity was assessed. The activation of urease was rapid, and the activity did not further increase during the time course (supplemental Fig. S5). Inclusion of EDTA in the urease assay showed that initially the activity was sensitive to EDTA and that an EDTA-insensitive urease activity was formed gradually over several hours. The kinetics of this urease activation was not accelerated, and the total activity was suppressed by the addition of GTP (supplemental Fig. S5B).
A similar GTP-independent activation was also described for the in vitro activation of urease from K. aerogenes but was considered to be physiologically irrelevant because of the high nickel and bicarbonate concentrations required (24). We hypothesized that a positive effect of GTP on the activation might only occur at lower concentrations of NiCl 2 and NaHCO 3 , which by itself (without GTP) would not be sufficient for full activation. To first identify concentrations leading to intermediate activation, a range of bicarbonate and nickel concentration was tested for activation of the purified urease-UreD-UreF-UreG complex. Mid-range activation was found with 2 mM NaHCO 3 and 5 M NiCl 2 (Fig. 4, upper and middle panels). With these concentrations set, the activation assay was carried out in the presence of either 0.15 mM or 1.5 mM GTP, respectively. Surprisingly, the addition of GTP to the assay mix did not accelerate the activation of urease but, rather, inhibited it in a dosage-dependent manner (Fig. 4, lower panel). Because magnesium ions were provided in a 2-fold excess of GTP, nickel chelation by the nucleotide as an explanation for the inhibition appears unlikely. Magnesium alone without GTP had no effect on the activation kinetics.
Next it was tested whether the observed activation was at all dependent on the presence of UreF and UreG. Urease in complex with either UreD, UreF, and UreG or only UreD and UreF or with UreD alone was purified and incubated with 5 M NiCl 2 , 2 mM NaHCO 3 , and 5 mM urea. The urease activities resulting from urease activation were normalized by the relative amount of purified urease protein present in each assay (Fig. 5). Urease from the urease-UreD and urease-UreD-UreF complexes was activated to a higher extent than urease from the urease-UreD-UreF-UreG complex. When these samples were subjected to in-gel urease activity-staining, all three setups resulted in an active urease with identical electrophoretic mobility as found in the control, which was an extract from leaves of the The UreD-UreF-UreG S complex was affinity-purified without eluting the proteins from the matrix, and an aliquot was analyzed (lane 4). The clarified extract from urease-expressing leaves was incubated either with purified UreD-UreF-UreG S complex still attached to Streptactin beads or with fresh Streptactin beads (control). After washing, the respective affinity-bound proteins were eluted and detected (lanes 5 and 6) by Western blot analysis. The experiment was repeated two times, and one of the repeats is shown.

Characterization of the plant urease activation complex
Arabidopsis wild type (Fig. 5, inset). However, active urease detectable by in-gel assays was obtained only after prolonged incubation with NiCl 2 and NaHCO 3 (Ͼ16 h) indicating that urease folding into the native form, which maintains its activity during electrophoresis, required several hours. Although prolonged in vitro activation generated an EDTAresistant urease enzyme (supplemental Fig. S5), showing the same electrophoretic mobility as the urease activated in planta (Fig. 5), this activation did not resemble the activation process in vivo where all three UAPs were shown to be required (15). Despite substantial efforts, we were unable to generate conditions where UreF and UreG promoted the activation in the presence of GTP and limiting amounts of nickel and bicarbonate. UreG even had an inhibitory effect (Fig. 5), maybe because it partially blocked the diffusion of nickel and bicarbonate toward the active site of urease.

UreG can acquire in vitro activation competence in vivo
Because GTP and UreG were not required in our in vitro activation trials, we speculated that we did not purify an activation-competent species of UreG and were unable to generate such a species in vitro. This prompted us to further investigate UreG and its proposed role in nickel delivery during activation (18,17), especially since it was recently suggested that UreG from soybean might not be involved in nickel delivery (19). UreG was expressed alone in N. benthamiana plants, which were either left unfertilized or were fertilized with NiCl 2 . The peat substrate employed in our experiments usually did not supply sufficient nickel for urease activation, making nickel fertilization mandatory for obtaining active urease in planta. Desalted extracts of UreG-expressing plants were mixed with desalted extracts of plants expressing the urease-UreD-UreF complex (again with or without Ni 2ϩ fertilization), incubated for the indicated times, and then urease activity was monitored in these mixtures (Fig. 6A). Urease activation was only observed when UreG from nickel-fertilized plants was present in the mixtures. UreG from unfertilized plants was not activationcompetent, and it was irrelevant whether extracts containing the urease-UreD-UreF complex came from fertilized or unfertilized plants. This indicated that UreG must come into contact with nickel in vivo to be converted into the activation-competent form. Alternatively, a component of the desalted extract from Ni 2ϩ -fertilized plants expressing UreG is required for activation. To distinguish these two possibilities, Strep-tagged UreG was purified and mixed either with a urease-UreD-UreF-containing extract or with a purified urease-UreD-UreF complex (purified via Strep-tagged UreD). Urease activation occurred in all cases where UreG was exposed to nickel in vivo, even when purified proteins were combined (Fig. 6B), showing for the first time a UreG-dependent activation of a eukaryotic urease with purified proteins. Note that neither nickel nor bicarbonate nor GTP was added to the activation mixture. These data demonstrate that extract components were not required but that activation competence lay within the UreG protein itself after coming into contact with nickel in vivo.

Plant UreG plays a role in nickel delivery to urease
It was noted before that plant UreG proteins possess an N-terminal extension in comparison to their bacterial orthologs (17,18). This extension contains a His-and Asp/Glu- 20 min 40 min 60 min 90 min 120 min Figure 6. UreG-dependent in vitro activation of urease. A, nickel-fertilized and non-fertilized N. benthamiana plants (ᮍ or ᮎ) were used to transiently express Arabidopsis UreG or to co-express urease together with UreD and UreF (UreG and UreD were Strep-tagged). Clarified and desalted extracts of these plants (G and UDF) were mixed with each other or with the extract of a control plant (C) not expressing any proteins and incubated for the indicated times. After incubation urea was added, and the urease activity was determined. B, the experiment was in part repeated and expanded mixing clarified and desalted extracts (index: x) and purified proteins (index: p). For purification, N-terminally Strep-tagged UreG (for G p ) and UreD (for UDF p ) were employed. Activation assays were carried out in triplicate (individual data points shown as open circles) in both experiments (error bars ϭ S.D.).

Characterization of the plant urease activation complex
rich stretch immediately downstream of the start codon, which is hypervariable in sequence and length (16), followed by ϳ40 amino acids that are highly conserved in plant UreGs comprising two similar HXH motifs: HSHD and H(S/T)H(E/D) beginning at amino acid 38 and 46 in the Arabidopsis sequence, respectively (Fig. 7A).
A sequence of 15 amino acids rich in His is also found at the C terminus of the K. aerogenes UreE, which shows no further sequence resemblance to plant UreGs. UreE dimers of K. aerogenes can bind up to six Ni 2ϩ , four of which are bound by the C terminus. However, the C terminus is not required for urease activation in vivo even at low nickel availability (25). An internal Ni 2ϩ -binding site at the dimer interface involving two His, one from each monomer, is conserved in bacterial UreEs (26) and is required for UreE activity (27). More His residues, in some cases organized in an HXH motif, at the C terminus of UreE proteins from different origins are also involved in Ni 2ϩ coordination (28). If plant UreG comprises UreE functionality, we hypothesized that the variable His-and Asp/Glu-rich sequence may be involved in this function as well as the two HXH motifs. These contain all invariant His residues in the conserved region of the plant-specific N-terminal UreG extension. Four N-terminal deletion mutants were generated, (i) eliminating only the hypervariable region of 26 amino acids (mutant ⌬N26), (ii) expanding the deletion into the first HXH motif (mutant ⌬N41), (iii) including both HXH motives (mutant ⌬N49), and (iv) eliminating the entire plant-specific N-terminal extension (mutant ⌬N67). Constructs containing a full length UreG as well as the four deletion variants, all driven by a strong 35 S promoter, were transformed into the ureG-1 knock-out line (15). To assess dose-dependent effects, lines with stronger and weaker expression of the corresponding variants were selected. UreG-⌬N67 was undetectable in the transgenic lines probably because the protein was unstable; therefore, these lines were not included in further experiments.
To assess the ability of the UreG variants to activate urease in a nickel-dependent manner, the urease activity of 4-week-old plants (just before bolting) grown on peat soil was assessed before and 4 days after nickel fertilization (Fig. 7B). Low-level urease activity was observed in the wild type even before the nickel treatment. As in the ureG mutant, the activity was near to zero (at background level) without nickel addition in the complementation lines containing the ⌬N41and ⌬N49-UreG variants. However, a statistically significant (at p Ͻ 0.05 level) difference in the wild type was only obtained for the ureG mutant. Nickel treatment led to a significant rise in urease activity in all genotypes except the ureG mutant and the two ⌬N49-UreGexpressing lines. With the amount of nickel added, the wild type reached ϳ30% of the maximal possible activity (15) showing that activation was not complete. The UreG variants were intentionally assessed at sub-optimal nickel availability to reveal a potential role of the plant-specific N terminus of UreG in nickel efficiency. The ⌬N26-UreG variant still activated urease but appeared to be less efficient when expressed at near wild-type level (⌬N26 -2), indicating that the hypervariable His-and Asp/Glu-rich N terminus boosts the efficiency of UreG but is not strictly required. The variant lacking the hypervariable region and the first HXH motif (⌬N41) still retains activation competence albeit at strongly reduced efficiency, whereas the ⌬N49-UreG variant loses the capacity to activate urease even when strongly overexpressed. These data demonstrate that the conserved part of the plant-specific N-terminal UreG extension plays a major role for UreG functionality.

Urease activation in vivo
Activation of K. aerogenes urease in vitro at near physiological conditions with regard to nickel and bicarbonate concentrations is a slow process requiring up to 10 h (29). Similar results

Characterization of the plant urease activation complex
were obtained for in vivo activation, which even after 4 h of nickel supplementation is not completed (30). For plant ureases nothing is known about the kinetics of the in vivo activation. To study this process in a time-resolved manner, Arabidopsis plants were grown hydroponically for 5 1 ⁄ 2 weeks without nickel. After supplementing the nutrient solution with 10 M NiCl 2 (no nickel in the control), plants were harvested in a time course, and urease activity and nickel concentration were assessed. Both remained unchanged in the unfertilized controls (supplemental Fig. 6, gray lines). In fertilized plants, the initial nickel concentration was only ϳ10% of the final concentration of ϳ0.02 nmol mg Ϫ1 fresh weight reached after 4 days (Fig. 8).
The initial nickel concentration was sufficient to activate ϳ40% of the maximum urease activity of ϳ17 milliunits mg Ϫ1 protein reached at the end of the experiment (Fig. 8), which is similar to literature values of full urease activation in Arabidopsis Col-0 (15). Within the first 4 h, the leaf nickel concentration doubled and was increased 6 times after 24 h, whereas the urease activity remained unchanged during the first 12 h and only slightly increased after 24 h (Fig. 8). Even after 4 days only 80% of the maximum activity was reached. The amount of urease protein was not affected by nickel availability (supplemental Fig. 4). These data show that in vivo the process of urease activation is rather a matter of days than of hours.

Discussion
The interaction of eukaryotic UAPs with urease have so far not been investigated in detail. Only for UreD and urease from rice has an interaction been reported (16). The crystal structure of a UreD-UreF-UreG complex from H. pylori was recently solved revealing that UreD (called UreH in this organism) interacts with UreF, which forms dimers. Upon interaction of UreH with UreF, a binding site for a dimer of UreG is formed on the UreF surface (12). Our data for Arabidopsis are consistent with such a model of UAP interactions. That plant UreGs can form dimers and are structurally similar to UreGs from bacteria has been shown for the protein from soybean (19). The amino acid sequences of the UreD and UreF proteins are only weakly conserved between plants and bacteria. Nonetheless, amino acids Arg-179 and Tyr-183 of H. pylori UreF, which are important for interaction with UreH (12), are conserved as Arg/Lys and Tyr/ Phe, respectively, in UreF proteins from plants (16). The residues in UreG and UreF of H. pylori and K. aerogenes involved in the interaction between both proteins (31,12) are as well mostly conserved in plant sequences, although UreF is only 19% identical between H. pylori and Arabidopsis (UreG: 40% identity). For example, for the amino acids Tyr-48 and Arg-250 in UreF of H. pylori, which are absolutely required for the interaction with UreG, the residues Phe and Arg are found in all plant UreF sequences (Phe-35 and Arg-236 in UreF of Arabidopsis; Ref. 16). Our co-purification data together with the particular conservation of residues important for protein-protein interactions suggest that the plant UAPs likely form a (UreD-UreF-UreG) 2 complex as found in H. pylori and K. aerogenes (32). In line with this are computational analyses indicating that plant and bacterial UAPs are structurally conserved (19,16). Although different models exist on how the UAPs interact with ureases of different quaternary structure for activation (32,12,8), there also appears to be a significant structural conservation of the urease-UAP activation complex across kingdoms. This is supported by the ability of soybean UreF to complement a UreF mutant of Schizosaccharomyces pombe (33) and by partial complementation of the K. aerogenes urease activation machinery lacking UreG by UreG from potato (18).
Interestingly in H. pylori, UreG dissociates from (UreD-UreF-UreG) 2 in the presence of GTP and nickel. However by binding nickel and GTP, a UreG dimer is formed that acquires the capability to interact with the urease-UreH-UreF complex leading to urease activation in the absence of nickel and GTP in solution (12). Similarly, Arabidopsis UreG activated in vivo in nickel-fertilized plants is capable of transferring Ni 2ϩ to a urease-UreD-UreF complex in vitro leading to urease activation without the addition of any nickel and bicarbonate and in the presence of EDTA (Fig. 6). Although the preformed UreD-UreF-UreG complex can interact with urease ( Fig. 3) as was also shown in K. aerogenes (34,32), the activation of urease at least in H. pylori may rather occur via the interaction of a urease-UreD-UreF complex with an activation-competent UreG dimer. UreGs from different sources bind at least one nickel ion per monomer (35,36,19,31) consistent with the idea that a UreG dimer would be capable of delivering the two Ni 2ϩ ions required in each active center of urease. It is also possible that a nickel-and GTP-charged UreG dimer delivers only one nickel then dissociates into monomeric UreG, which is followed by a second round of charged UreG dimer-binding and nickel delivery. This is an attractive model because there is a single metal-binding site at the interface between the UreG dimers (12) that is conserved in bacterial and plant UreGs (16).
Bacterial UreG can be made activation-competent in vitro (24,12). In contrast, we were unable to generate activationcompetent UreG in vitro by the incubation with nickel, bicarbonate, and GTP (also ATP was tested). That plant UreG hydrolyzes GTP in principle was shown for the soybean enzyme (19). However, Arabidopsis UreG isolated from nickel-sufficient plant cells (Fig. 6) or from Escherichia coli (15) was competent for activation. This may be explained by the requirement for a further protein factor needed to deliver the nickel to plant

Characterization of the plant urease activation complex
UreG, which was not present in vitro. In H. pylori, HypA and HypB, which are associated with metallocenter biosynthesis of Ni-Fe hydrogenases, also play a role in urease activation. It was shown that HypA binds to a UreE dimer to deliver nickel. In the presence of Mg-GTP, a UreG dimer interacts with (UreE) 2 , displacing HypA and acquiring nickel from UreE (11,37). It is possible that plant UreGs, which may unite bacterial UreG and UreE functions in one protein, obtain activation competence only if the nickel is delivered from a protein factor. Alternatively, a protein factor may be required to assist UreG folding into the activation-competent state because UreG was shown to be an intrinsically disordered protein (38,36). UreE of K. aerogenes contains a His-rich C terminus, which is not required for urease activation in vivo (25), and K. aerogenes urease can be partially activated even in the absence of UreE (30,25). The deletion of the His-Asp/Glu-rich part of the N terminus of Arabidopsis UreG also had no strong effect on the functionality of the protein (Fig. 7). However, at a lower level of overexpression (line ⌬N26 -2), urease activity was significantly lower than in the wild type when limiting nickel had been supplied. The data indicate that the hypervariable N terminus of the protein may serve to enrich metals (nickel). A functional UreE requires a conserved His involved in Ni 2ϩ -binding at the dimer interface (26,27). The deletion of both HXH motifs comprising all conserved His in the plant-specific N terminus of Arabidopsis UreG (mutant ⌬N49) abolished urease activation in vivo despite strong overexpression of the variant protein (Fig.  7). We conclude that the plant-specific N terminus of UreG not only boosts urease activation, as does UreE, but is absolutely required, at least at the limiting availability of nickel in our experiment. When only the first HXH motif was removed (mutant ⌬N41), the protein remained partially functional after nickel addition, indicating that the N terminus of UreG up to this point may be involved in efficient nickel delivery to the core function of UreG.
Full urease activation in vivo requires several days after nickel addition ( Fig. 8) but not because the urease protein is unstable if it is not activated (supplemental Fig. S6). This is similar in K. aerogenes where urease is produced at normal levels when UAPs are missing (9). Also in bacteria, urease activation is not immediate but requires several hours after nickel becomes available (30). An explanation may be that the kinetics of nickel delivery to the urease-UAP complex might be governed by the diffusion rate of nickel-carrying proteins within the cell. Due to the greater size of eukaryotic cells, this process may take longer there than in prokaryotes.

Plant material and cultivation
Arabidopsis thaliana, ecotype Col-0, and the ureG-1 mutant line (GK294B06; 15) were used. Arabidopsis and N. benthamiana plants were cultivated as described before (21,15). Transient expression in N. benthamiana was performed as described by Werner et al. (39) but adjusting the optical densities of the helper Agrobacterium strain containing the p19 silencing inhibitor construct to 0.1 and of all other agrobacteria strains to 0.25 each. In co-purification experiments, equal agro-bacteria load was ensured by adding agrobacteria expressing urate oxidase (40), which is unrelated to urease activation, when agrobacteria for the expression of distinct components of the urease-UAP complex were omitted.

Affinity purification
For investigating the urease activation complex, proteins were affinity-purified according to Werner et al. (39), with slight changes in the washing procedure. The affinity matrix was washed only once but with 1.5 ml of wash buffer.

Characterization of the plant urease activation complex
For the binding assay of urease to the UreD-UreF-UreG complex, 1.5 g and 0.75 g of leaf material expressing urease and the complex were extracted, respectively (1:2 ratio with extraction buffer) and clarified (20,000 ϫ g, 4°C). 1.5 ml of supernatant containing the complex was incubated for 20 min (rotation wheel, 4°C) with 60 l Strep-Tactin Macroprep (50% suspension). The beads were washed twice with 1 ml and 0.5 ml of wash buffer, respectively. The supernatant of 3 ml-containing urease was split into two halves to which either 60 l of fresh Strep-Tactin Macroprep or the beads with bound complex were added. The slurries were incubated for 20 min, and the matrices were washed once with 1.5 ml of wash buffer and then exposed to 150 l of elution buffer.

In vitro activation of urease
Affinity-purified urease-UAP complexes were incubated in 25 mM HEPES, pH 8.0, 0 -10 mM NaHCO 3 , 0 -10 M NiCl 2 , 0 -1.5 mM GTP, and 5 mM urea. The reaction components were incubated at 50°C, and the reaction (volume 300 l) was started with the addition of purified complex. For urease assay, samples of 20 l were placed into 80 l of water in a time course, and 25 l of phenol nitroprusside followed by 50 l of hypochloride solution were added. The resulting color was quantified colorimetrically at 639 nm (6).
For urease activation experiments in mixed N. benthamiana leaf extracts, plants expressing Strep-tagged UreG or co-expressing urease, UreD, and UreF (with UreD Strep-tagged) or non-infiltrated control plants were employed. Conditionally, plants were fertilized with 1 mM NiCl 2 (6 ml per pot) 1 day before infiltration for transient expression. Clarified extracts were desalted on 5-ml HiTrap G25 Sephadex columns (GE Healthcare) equilibrated with 25 mM HEPES, pH 8.0, 75 mM NaCl, 2.5 mM EDTA, and 2 mM DTT. Extracts were mixed and incubated at 50°C for the indicated times followed by the quantification of urease activities. Each incubation was repeated three times.

Generation of polyclonal antisera
AtUreD, AtUreF, OsUreD, and OsUreF were expressed as maltose-binding protein fusion proteins in E. coli BL21 (DE3) cells from the pMAL-c5X vector (New England BioLabs). Cells were grown in 100 ml of LB medium and induced by isopropyl-␤-D-thiogalactoside after reaching an optical density of 0.5. Cells were harvested 3 h after induction, and cell pellets were dissolved in column buffer (20 mM Tris-HCl, pH 7.5; 200 mM NaCl; 1 mM EDTA) at a ratio of 2.5 ml per 100 ml cell culture. A spatula tip of lysozyme powder and 1.5 l of DNase I (10 mg ml Ϫ1 ) were added. Cells were incubated at room temperature for 30 min, sonicated four times (Branson S150-D; level 3; 15 s each), and centrifuged (20 min, 20,000 ϫ g, 4°C). AtUreD, AtUreF, and OsUreD formed inclusion bodies, whereas OsUreF was soluble.
The supernatant-containing OsUreF was loaded on a 1-ml amylose matrix (New England BioLabs) pre-equilibrated with 5 ml of column buffer and washed 5 times with 2 ml of column buffer per wash. OsUreF was eluted by the repeated addition (10 times) of 0.75 ml of elution buffer (column buffer ϩ 10 mM maltose). After examining purity and concentration of the pro-tein on SDS gels, the two fractions with the highest protein content were pooled and desalted using a 5-ml HiTrap column (GE Healthcare).
The inclusion body pellets of AtUreD, AtUreF, and OsUreD were washed by resuspending them in 1.25 ml of column buffer each and incubating in an ice-cold ultrasonic bath (5 min) followed by the addition of 12.5 l of Triton X-100 per ml and centrifugation (20 min; 20,000 ϫ g; 4°C). The washing step was repeated, and inclusion body pellets were resuspended in 300 l of SDS-PAGE loading buffer. Proteins were separated on porous SDS gels and processed for antisera production as described by Dahncke and Witte (43).
The untagged and truncated AtUreG variant ⌬N67 was expressed in E. coli from pET30nco-CTH by isopropyl-␤-Dthiogalactoside induction as described above. The soluble protein was purified using a 13% SDS gel column on a Model 491 PrepCell (Bio-Rad) system. Fractions containing the purified protein were precipitated with methanol chloroform.

Characterization of the plant urease activation complex
extract was clarified (20,000 ϫ g, 15 min, 4°C) and 100-l desalted using spin columns equilibrated with half-strength extraction buffer without AEBSF and DTT. Urease activity from three replicates per line was determined with colorimetric detection (6). The experiment was repeated with plants that had been fertilized with 1 ml of 100 M NiCl 2 applied to the soil 4 days before harvest near the center of each rosette (3 ml of Ni 2ϩ solution per pot). The expression of the UreG variants was monitored by Western blot using the antibody raised against ⌬N67-UreG. For each line the sample for the blot consisted of a pool of extracts of all replicates.

Urease activation and nickel content in hydroponically grown plants
Col-0 was grown in a hydroponic growth system (Araponics) for 12 h in the light (22°C) and 12 h in the dark (20°C) in modified Hoagland medium containing 1. After centrifugation (20,000 ϫ g, 15 min, 4°C), 100 l of the supernatant were desalted as described in Cao et al. (16), and urease activity was determined, whereas the rest was stored for Western blotting and nickel analysis. To determine the nickel content, 100 l of plant extract were mixed with 900 l of double-distilled H 2 O. After the addition of 1 ml of diluted nitric acid (2.3%), samples were analyzed with an inductively coupled plasma mass spectrometer (ICP-MS; 7500CX, Agilent).

Data analysis
Statistical analyses were performed with the GraphPad Prism software package.
Author contributions-C.-P. W. and T. M. designed the experiments, which were performed by T. M. with the exception of the experiment shown in Fig. 7, which was performed by C.-P. W. and A. Z. The manuscript was written by C.-P. W. and T. M. with revision by A. Z. All authors reviewed the results and approved the final version of the manuscript.