Functional and Biochemical Characterization of Cucumber Genes Encoding Two Copper ATPases CsHMA5.1 and CsHMA5.2*

Background: Plant copper HMA5-like ATPases have not been biochemically characterized yet. Results: Cucumber copper ATPases CsHMA5.1 and CsHMA5.2 are tonoplast monovalent copper transporters differentially regulated by copper availability. Conclusion: HMA5-like ATPases may contribute to vacuolar sequestration of copper excess in plant cells. Significance: Orthologous transporters are differently displayed to achieve organismal copper homeostasis. Plant copper P1B-type ATPases appear to be crucial for maintaining copper homeostasis within plant cells, but until now they have been studied mostly in model plant systems. Here, we present the molecular and biochemical characterization of two cucumber copper ATPases, CsHMA5.1 and CsHMA5.2, indicating a different function for HMA5-like proteins in different plants. When expressed in yeast, CsHMA5.1 and CsHMA5.2 localize to the vacuolar membrane and are activated by monovalent copper or silver ions and cysteine, showing different affinities to Cu+ (Km ∼1 or 0.5 μm, respectively) and similar affinity to Ag+ (Km ∼2.5 μm). Both proteins restore the growth of yeast mutants sensitive to copper excess and silver through intracellular copper sequestration, indicating that they contribute to copper and silver detoxification. Immunoblotting with specific antibodies revealed the presence of CsHMA5.1 and CsHMA5.2 in the tonoplast of cucumber cells. Interestingly, the root-specific CsHMA5.1 was not affected by copper stress, whereas the widely expressed CsHMA5.2 was up-regulated or down-regulated in roots upon copper excess or deficiency, respectively. The copper-induced increase in tonoplast CsHMA5.2 is consistent with the increased activity of ATP-dependent copper transport into tonoplast vesicles isolated from roots of plants grown under copper excess. These data identify CsHMA5.1 and CsHMA5.2 as high affinity Cu+ transporters and suggest that CsHMA5.2 is responsible for the increased sequestration of copper in vacuoles of cucumber root cells under copper excess.

amine oxidase, lysyl oxidase, and ferroxidase. In plants, the crucial metabolic processes such as photosynthesis, respiration, hormone signaling, and the response to redox perturbations require copper. However, only minute amounts of copper are available in the cells due to the high toxicity of copper excess (1). The critical components of the copper homeostasis mechanism in plants include membrane transporters responsible for the delivery of copper ions into different cellular compartments to supply copper-dependent activities as well as for the rapid efflux or sequestration of toxic copper excess (2).
Among the several families of heavy metal transporters that have been identified in plants, P 1B -ATPases (heavy metal ATPases, HMAs 2 ) are proteins involved in the delivery of essential metals, including copper, to intracellular compartments or in the detoxification of plant cells from heavy metal excess (3)(4)(5)(6)(7). Based on previous studies on yeast, bacterial and human P 1B -ATPases, as well as on sequence comparisons, the family was divided into two groups as follows: one including transporters of monovalent cations, copper/silver; and the second transporting divalent cations, cadmium/lead/zinc/cobalt (3,4). However, based on the conserved residues in transmembrane helices 6 -8, P 1B -ATPases have been further classified into five subgroups designated as follows: P IB1 , containing Cu ϩand Ag ϩ -ATPases; P IB2 , with Cd 2ϩ -, Zn 2ϩ -, and Pb 2ϩ -AT-Pases; P IB3 , with Cu ϩ /Cu 2ϩ -ATPases; P IB4 , including Co 2ϩ -ATPases; and P IB5 , containing P IB -type ATPases of unknown specificity (5).
In model dicot and monocot plants, Arabidopsis thaliana and Oryza sativa, eight (AtHMA1 to AtHMA8) and nine (OsHMA1 to OsHMA9) P 1B -ATPases have been identified. Based on current knowledge, copper transport has been attributed to plant HMA1 (subgroup P 1B4 ) and HMA5-8 (subgroup P 1B1 ) P 1B -ATPases. Members of the P 1B4 subgroup include broad-specificity transporters of metals. A. thaliana HMA1 was shown to be a Ca 2ϩ /heavy metal pump localized in the chloroplast envelope, which delivers copper from the cytosol into the chloroplast stroma and detoxifies chloroplasts from Zn 2ϩ under zinc excess (8,9). A similar protein from barley (Hordeum vulgaris), HvHMA1, was shown to be involved in mobilization of zinc and copper from plastids under zinc or copper deficiency or from the intracellular compartments of the mineral storage site of grains (aleurone cells) during grain filling and germination (10). In comparison, the P 1B1 subgroup of plant P 1B -ATPases includes proteins that are highly specific for monovalent Cu ϩ (11,12). The first characterized plant HMA, AtHMA7 (RAN1), is a major component of the plant hormone ethylene signaling pathway because it delivers copper to the secretory pathway to supply the formation of functional copper-dependent ethylene receptors (13,14). Chloroplast-localized HMA6 (PAA1) and HMA8 (PAA2) are responsible for the import of copper to the chloroplasts to provide the cofactor for copper-dependent activities, antioxidant enzyme CuZn-superoxide dismutase and plastocyanin, involved in photosynthesis (12,15,16). Contrary to AtHMA6 -AtHMA8, which deliver copper to the target proteins, AtHMA5 appears to be involved in detoxification of Arabidopsis roots from copper excess (17). A similar function has been recently reported for the homologous protein OsHMA5 in rice (18). The athma5 knock-out mutant is hypersensitive to copper, whereas rice knock-out lines for OsHMA5 accumulated more copper in roots and less in shoots, when compared with wild type plants (17,18). In addition, OsHMA5 was shown to be localized in the plasma membrane (18). These observations suggest that in plants HMA5 is involved in root-to-shoot translocation of copper and thus protects plant roots from copper toxicity (17). In contrast to A. thaliana, rice possess two homologs of HMA5 proteins OsHMA5 and OsHMA4 (19); however, the function of the latter one is still unknown. Multiple orthologs of HMA5 proteins have also been found in other plants as follows: four in the dicot plants Populus trichocarpa (PtHMA5.1, PtHMA5.2, PtHMA5.3, and PtHMA5.4) and Vitis vinifera (VvHMA5.1, VvHMA5.2, VvHMA5.3, and VvHMA5. 4), and three in the monocot Sorghum bicolor (SbHMA5.1, SbHMA5.2, and SbHMA5.3) and the lycophyte Selaginella moellendorffii (SmHMA5.1, SmHMA5.2, and SmHMA5.3) (19). Some of the HMA5 isoforms are very closely related and close together on the same chromosomes (VvHMA5.1, VvHMA5.2, and VvHMA5.3; PtHMA5.1 and PtHMA5.3); hence, they could have arisen as a result of a gene duplication event (19). Because none of the identified duplicate genes have been studied yet, the relevance of these extra Cu-ATPases is not known.
The recent screening of the cucumber genome revealed the presence of two genes encoding two putative proteins homologous to HMA5-like P 1B -ATPase (20). Based on the homology to AtHMA5, the proteins were initially designated as CsHMA5A and CsHMA5B (20). To make the nomenclature of cucumber HMAs consistent with the annotation of the reported isoforms of other plant HMAs (19), we renamed CsHMA5A and CsHMA5B as CsHMA5.1 and CsHMA5.2, respectively. In this work, we investigated the function, regulation, and biochemical properties of both cucumber HMA5-like pumps. Using yeast to produce the functional CsHMA5.1 and CsHMA5.2 proteins, we provide the first enzymatic character-ization of HMA5-like copper ATPases in plants and show that CsHMA5.1 and CsHMA5.2 act as high affinity vacuolar Cu ϩ transporters. In addition, we provide evidence for different regulation of the transcript and protein levels of CsHMA5.1 and CsHMA5.2 in roots under different availability of copper. Taken together, the data suggest different biological roles of cucumber CsHMA5.1 and CsHMA5.2 and reveal differences between the function of homologous HMA5 proteins from cucumber, rice, and A. thaliana.

Experimental Procedures
Plant Material and Growth Conditions-Cucumber (Cucumis sativus var. Krak) plants were grown in hydroponics as described earlier (21). To induce copper deficiency, following germination the seedlings were grown for 2 weeks in nutrient solutions completely deprived of copper. To impose copper toxicity, 2-week-old cucumber seedlings were transferred to the fresh nutrition media supplemented with 20 M CuCl 2 and grown for the next 72 h.
Isolation of Vacuolar Membranes from Yeast Cells and Plant Roots-Exponentially growing yeast cells expressing CsHMA5.1 or CsHMA5.2 or transformed with empty vector (negative control) were used for preparation of vacuolar membranes. At first, total microsomes were prepared according to a previously described protocol (26). Subsequently, vacuolar membranes were separated from total membrane proteins by ultracentrifugation in the sucrose gradient essentially as described earlier (27). The presence of CsHMA5.1 and CsHMA5.2 in yeast vacuole was confirmed by immunolocalization assay with specific antibodies raised against cucumber proteins and by using a fluorescence microscope (Axio Imager M1, Carl Zeiss) equipped with a ϫ100 oil immersion objective. For visualization of yeast vacuolar membranes, cells were labeled for 20 min at 30°C in 6.4 M FM4-64 (Molecular Probes, Eugene, OR) in SC-U/Glu medium. Following washing with water, cells were further cultured in SC-U/Glu medium for 30 min.
The roots (30 g) from 2-week old cucumber plants growing under different copper availability were harvested for the preparation of tonoplast and other membrane-enriched fractions, essentially as described earlier (20,21). Protein content was estimated according to Bradford (28), with BSA as a standard.
Assay for CsHMA5.1 and CsHMA5.2 ATPase Activities in Yeast Vacuolar Membranes-ATPase activity was measured as the rate of released inorganic phosphate as described by Ames (29). The reaction media contained 100 g of vesicle protein, 5 mM Tris-MES (pH 7.6), 0.1 mM sodium azide, 0.5 mg ml Ϫ1 saponin, different concentrations of metals, and 300 nmol of bafilomycin to avoid the background resulting from the activity of vacuolar V-ATPase. 300 M Na 2 SO 3 or 1 mM BCS (copper ϩspecific chelator) was included in the assay containing copper. Following a 10-min-long incubation at room temperature, reactions were initiated by 30 mM Mg-ATP and carried out for 30 min at 30°C. The rate of P i released due to HMA activity was calculated from the difference between the activity measured with or without the specific P-ATPase inhibitor Na 2 VO 4 (1 mM). The background hydrolytic activities measured in membranes prepared from yeast transformed with empty vector were subtracted from the results obtained in membranes containing cucumber CsHMA5.1 or CsHMA5.2.
Assay for CsHMA5.1 and CsHMA5.2 Transport Activities in Vacuolar Membranes Isolated from Yeast Cells or Plant Roots-ATP-dependent copper transport was assayed in reaction media containing 5 mM Tris-MES (pH 7.6), 0.3 M sorbitol, 0.1 mM sodium azide, different concentrations of CuCl 2 , 300 M Na 2 SO 3 , or 1 mM BCS and 300 nmol of bafilomycin. The membranes (100 g of protein) were incubated with the reagents for 5 min at room temperature, and then 1 mM MgSO 4 and 1 mM ATP were added to initiate ATPase-mediated transport of copper into tonoplast vesicles. After 5 min, tonoplast was filtered through nitrocellulose membrane (0.45 m, Millipore) and washed with the copper-free buffer to remove metal unspecifically bound to the membranes. Following washing, 3 M HCl was used to elute copper accumulated inside vesicles. Intravesicular copper content was determined using AAS 3300 (PerkinElmer Life Sciences). HMA-dependent copper transport was calculated from the difference between the transport activity measured with or without Na 2 VO 4 (1 mM). The background transport activities measured in membranes prepared from yeast transformed with empty vector were subtracted from the results obtained in membranes containing cucumber CsHMA5.1 or CsHMA5.2.
sRT-PCR and Real Time PCR-Total RNA were extracted from various cucumber organs with TRI Reagent (Sigma).
Reverse transcription was carried out as described previously (30). Semiquantitative RT-PCR (sRT-PCR) and quantitative PCR experiments were performed according to previously reported procedures (31). The primers used to analyze the expression of cucumber genes were carefully designed in Light-Cycler Probe Design software 2 (Roche Applied Science) to ensure the high specificity of PCRs. The forward primer 5Ј-TAT TAA CGA CGA CGA GGC CA-3Ј and reverse primer 5Ј-GAA GAT AGA GGA TTT GAG CCC-3Ј for CsHMA5.1 were designed in the UTR5Ј (position 53 to 37 bp upstream of the start codon) and N-terminal (position 92-112 bp downstream of the start codon) regions of the gene, respectively. The forward primer 5Ј-ATT AAC CTC AGT TCG ATC ATC C-3Ј and reverse primer 5Ј-GCT GGG TAT TTG GGC ATA GA-3Ј for CsHMA5.2 were both designed in the UTR5Ј region (positions 190 to 169 and 69 to 48 bp upstream of the start codon, respectively) of the cucumber gene. The forward primer 5Ј-GTG CTT TCT TTC TGG AAT GC-3Јand reverse primer 5Ј-TGA ACC TCG TCA AAT TTA CAC A-3Ј were used to amplify the reference gene encoding the Clathrin Adaptor Complex Subunit (CACS, accession number GW881874). The sRT-PCRs were run using Marathon polymerase (A&A Biotechnology) according to following conditions: initial denaturation at 94°C for 30 s, followed by 30 (CsHMA5.1 and CsHMA5.2) or 25 (CsCACS) cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 60 s. The quantitative PCRs were run in LightCycler 480 (Roche Applied Science) and started with initial denaturation at 95°C for 30 s, followed by 45 cycles of denaturation at 95°C for 10 s, annealing at 60°C for 10 s, and extension at 72°C for 12 s. This was followed by a melt cycle that consisted of a stepwise increase in temperature from 72 to 99°C. The PCR products were sequenced to confirm primer specificity. The relative quantification analysis (⌬⌬C T method) was performed using the LightCycler 480 software, Version 1.5.
Sequence Analyses-The hydropathy profile and location of putative transmembrane helices was determined by TMHMM Server2.0 and visualized using TMRPres2D software. Multialignment and homology estimation between cucumber HMA sequences was performed using ClustalW. The phylogenetic tree of plant HMA5-8 proteins was constructed using MEGA6.0 software (32) and the Maximum Likelihood method with the bootstrap (1000 replicates). The sequences from A. thaliana and O. sativa were retrieved from TAIR and TIGR databases, respectively. The previously annotated (19) sequences from Chlamydomonas reinhardtii, Physcomitrella patens, P. trichocarpa, S. bicolor, S. moellendorffii, and V. vinifera were retrieved from customized databases (GenBank TM , TAIR, Phytozome Version 9.1). The PCR-amplified CsHMA5.1 and CsHMA5.2 sequences have been submitted to the Gen-Bank TM data library and are available under accession numbers KJ818254 and KJ818255.
Statistical Analysis-Each experiment was repeated three to six times using material prepared from different plants or yeast cultures. Values are expressed as means Ϯ S.D. Data were analyzed using one-way analysis of variance followed by Tukey's test. Statistical analysis of the data were performed using paired or unpaired Student's t test; significance was accepted when p Ͻ 0.05. The data from ATPase activity assay and metal transport assay were fitted to the Michaelis-Menten equation using GraphPad Prism software (GraphPad Software, Inc.).

Results
Sequence Analysis of Cucumber CsHMA5s-The two genes coding for HMA5-like proteins in cucumber are located on different chromosomes of the cucumber Chinese long genome. CsHMA5.1, which is more closely related to AtHMA5, was found on chromosome 5 (ACHR02000005.1), whereas CsHMA5.2 is located on chromosome 4 (ACHR02000004.1). HMA5s are not the only plant Cu ϩ -ATPases that have been multiplied. The genes encoding HMA6s and HMA7s are also present in multiple copies in some plants, but the number of their copies is lower when compared with HMA5s ( Fig. 1). Phylogenetic analysis of plant monovalent copper ATPases revealed that they form two clearly distinct phylogenetic clusters, one including HMA6s and HMA8s and the other containing HMA7s and HMA5s, which originated from a common ancestral protein related to the present HMA7 proteins from the microalga C. reinhardtii (CrHMA7.1 and CrHMA7.2) (Fig. 1). Interestingly, HMA5 genes are absent in green alga C. reinhardtii, suggesting  that they originated from the HMA7-like ancestor after the emergence of vascular land plants. Hence, HMA5s are most closely related to HMA7-like pumps. Indeed, when compared with other cucumber Cu ϩ -transporting ATPases, CsHMA5.1 and CsHMA5.2 are the most similar to CsHMA7 (40 and 44%, respectively) and show low sequence similarity to CsHMA6 (16 and 17%, respectively) and CsHMA8 (21 and 26%, respectively). In addition, the in silico predictions (PlantLoc, TargetP servers) of subcellular localization based on sequence comparisons to the already fixed locations suggest that different cucumber Cu ϩ -ATPases might be targeted to different cellular membranes. CsHMA6 and CsHMA8 contain chloroplast-targeting peptides and localize with the highest probability to chloroplasts (data not shown), consistent with the chloroplast localization of their Arabidopsis homologs AtHMA6 and AtHMA8 (15,16). A similar chloroplast-targeting sequence was identified in CsHMA5.2; however, this protein is predicted to localize with much higher probability in Golgi and vacuolar membranes (data not shown). In comparison, CsHMA5.1 was strongly predicted to localize mainly in the tonoplast or endoplasmic reticulum (with lower probability), whereas the closest homolog of CsHMA5s, CsHMA7, might be targeted with the same probability to different intracellular membranes, including endoplasmic reticulum, Golgi, and tonoplast (data not shown). These results are consistent with the phylogenetic analysis and the level of sequence homology between cucumber Cu ϩ -ATPases and suggest that CsHMA5.1 and CsHMA5.2 localize to the intracellular membranes rather than to the plasma membrane.
Maintenance and the expansion of HMA5 genes indicate that HMA5s play important functions in different plants. However, only AtHMA5 and OsHMA5 have been functionally characterized to date, so the function of additional HMA5 isoforms in plants is not known. Using complementary DNA (cDNA) of cucumber roots (cv. Krak) and primers designed for heterologous expression in yeast, the full-length coding regions of CsHMA5.1 and CsHMA5.2 were amplified by PCR, sequenced, and deposited in the GenBank TM database. cDNA sequences of CsHMA5.1 and CsHMA5.2 isolated from cucumber Krak were identical to the relative sequences generated from genomic contigs of cucumbers Chinese long and Borszczagowski (data not shown). They encode two putative proteins of 973 and 926 amino acids, respectively (20). Protein sequence comparisons revealed that CsHMA5.1 shows 58% similarity to CsHMA5.2 and that both proteins show a comparable level of similarity (67 and 65%, respectively) to the homologous protein from A. thaliana (AtHMA5). Nevertheless, the phylogenetic analysis suggests that CsHMA5.1 and AtHMA5 are more closely related (Fig. 1). Both orthologs, CsHMA5.1 and CsHMA5.2, possess the features typical for P-ATPases (conservative motif DKTGTLT with an Asp residue that is phosphorylated and dephosphorylated during the catalytic cycle) and P 1B -ATPases (eight transmembrane-spanning domains, CPC(X) 6 P motif, and locus HP) (Fig. 2, A and B). In addition, the putative copper-binding motifs CXXC (two motifs in the N termini), YN (TMDVII), and MXXS (TMDVIII), which are characteristic for copper ATPases (33,34), were identified in CsHMA5.1 and CsHMA5.2 ( Fig. 2A), suggesting that cucumber pumps are putative copper transporters.
Organ and Subcellular Localization of CsHMA5.1 and CsHMA5.2-The function of heavy metal transporters is strongly determined by their tissue and subcellular localization.

Characterization of Cu-ATPases CsHMA5.1 and CsHMA5.2
The organ expression pattern of CsHMA5.1 and CsHMA5.2 was determined in vegetative organs of 2-week-old cucumbers, including roots, hypocotyls, cotyledons, and leaves, using sRT-PCR and semiquantitative real time PCR. sRT-PCR was also used to establish the CsHMA5.1 and CsHMA5.2 expression level in inflorescence and fruits. Fig. 3, A and B, shows that the CsHMA5.1 transcript was present only in cucumber roots. In contrast, CsHMA5.1 mRNA was detected in all vegetative organs and inflorescences, suggesting different physiological roles for CsHMA5.1 and CsHMA5.2 in cucumber. Root expression of CsHMA5.2 was ϳ3-fold higher when compared with CsHMA5.1. Nevertheless, the level of CsHMA5.2 expression in vegetative organs was the highest in hypocotyls and leaf petioles. The subcellular localization of CsHMA5.1 and CsHMA5.2 in cucumber roots was further investigated by immunostaining with two specific antibodies raised against the peptides from the deduced amino acid sequences of the N-terminal domains of CsHMA5s ( Fig. 2A). Based on the amino acid composition, the molecular masses of CsHMA5.1 and CsHMA5.2 were expected to be close to 105 and 98 kDa, respectively. Western blot analysis of different cucumber membranes fractionated on a sucrose gradient showed that CsHMA5.1 and CsHMA5.2 colocalize with the marker protein for vacuolar membrane (vacuolar PPase) and not with the marker protein for plasma membrane (H ϩ -ATPase). The antibodies reacted with the ϳ100-kDa proteins and did not cross-react with other proteins (Fig. 4). Subcellular localization of CsHMA5.1 and CsHMA5.2 was dissimilar to OsHMA5, a plasma membrane-localized homologous transporter in rice (18).
ATPase and Transport Activity of CsHMA5.1 and CsHMA5.2-Copper ATPases characterized so far, CopA, ATP7A, ATP7B, and PAA1/HMA6, required monovalent copper for maximal activation of their ATPase activity (12,(35)(36)(37)(38). To establish whether copper or other metals are essential for CsHMA5.1mediated and CsHMA5.2-mediated ATP hydrolysis, we studied the ATPase activity in vacuolar membranes isolated from the ⌬ycf1 strain expressing CsHMA5.1-GFP or CsHMA5-GFP. The strain lacking the ATP-dependent ABC transporter YCF1 was used to avoid background activity resulting from YCF1mediated ATP hydrolysis. The immunostaining of vacuolar membranes isolated from yeast expressing CsHMA5.1-GFP or CsHMA5-GFP with anti-CsHMA5.1 or anti-CsHMA5.2 antibodies, respectively, revealed the high specificity of antibodies and confirmed that both cucumber proteins were present in vacuolar membranes of yeast cells (Fig. 5, A and B). The antibodies reacted with the ϳ130-kDa proteins corresponding to CsHMA5-GFP fusions (ϳ100 kDa of CsHMA5 ϩ ϳ27 kDa of GFP). Cu-ATPase activities of CsHMA5.1 or CsHMA5.2 were assayed in the presence of DTT, Na 2 SO 3 , BCS, and cysteine to establish which form of CuCl, or copper thiolate, is required for ATP hydrolysis mediated by cucumber pumps. DTT and Na 2 SO 3 act as reducing agents generating Cu ϩ in the assay, whereas BCS chelates monovalent copper, and thus the amount of available Cu ϩ in the reactions containing BCS  instead of Na 2 SO 3 is markedly reduced (12). Cysteine was previously shown to positively affect P 1B -ATPase activity, probably through the binding of metal and forming a chelated complex interacting with the binding domains of a heavy metal transporter (35,39,40). The assay confirmed that CsHMA5.1 and CsHMA5.2 are copper ATPases, as expected from their homology to other characterized copper P 1B1 -ATPases. Both pumps were activated by copper, and the effect of metal was markedly increased by the reducing agent DTT or Na 2 SO 3 (Fig. 5C). In contrast, BCS almost completely abolished the stimulatory effect of copper on ATP hydrolysis by CsHMA5.1 and CsHMA5.2, indicating that the monovalent Cu ϩ rather than divalent Cu 2ϩ is the putative substrate for cucumber proteins (Fig. 5C). The maximal activation of CsHMA5.1 and CsHMA5.2 was observed upon addition of cysteine to the media containing both the reducing agent and copper, implicating the importance of copper thiolate formation for the interaction of the metal with both transporters (Fig. 5C). Further ATPase activity assays, including other metal ions, revealed that CsHMA5.1 and CsHMA5.2 are also activated by Ag ϩ , although to a lesser extent (ϳ2-or 3-fold less when compared with the copper-induced activity of CsHMA5.1 and CsHMA5.2, respectively). Other heavy metals, including Zn 2ϩ , Cd 2ϩ , Mn 2ϩ , Ni 2ϩ , and Co 2ϩ , were unable to significantly activate CsHMA5.1 (Fig. 5D), whereas CsHMA5.2 was also slightly (ϳ10% of copper-induced activity) activated by Cd 2ϩ (Fig. 5E). Similar to copper, the maximum rate of Ag-ATPase activity of CsHMA5.1 and CsHMA5.2 was observed in the presence of Cys in the assay. Both Cu-ATPase and Ag-ATPase activities of CsHMA5.1 and CsHMA5.2 were increased ϳ2and 3-fold by Cys, respectively (Fig. 5, D and E). The kinetic properties of Cu ϩ and Ag ϩ interaction with CsHMA5.1 and CsHMA5.2 were studied by measuring the dependence of ATPase activity on metal concentrations (Fig. 6, A and B). The rate of ATPase hydrolysis mediated by CsHMA5.1 and CsHMA5.2 increased with the increase of metals in the media up to 5 M (copper) or 10 M (silver), whereas higher copper or silver concentrations had an inhibitory effect on both pumps (Fig. 6, A and B). In the range of stimulatory metal concentrations, CsHMA5.1 showed a different 2-fold lower affinity for Cu ϩ (K m ϭ 1 Ϯ 0.11 M) when compared with CsHMA5.2 (K m ϭ 0.5 Ϯ 0.08 M). In contrast, both proteins showed lower but similar affinity for Ag ϩ (K m ϭ 2.5 Ϯ 0.21 M and 2.4 Ϯ 0.76 M for CsHMA5.1 or CsHMA5.2, respectively) (Fig. 6, A and B). Interestingly, Cu-ATPase activity of CsHMA5.2 was slightly higher, when compared with CsHMA5.1, although the Ag-ATPase activity of CsHMA5.1 was significantly higher than the relevant activity of CsHMA5.2 (Fig. 6). Hence, the mechanisms of phosphorylation and dephosphorylation of both proteins resulting from copper and silver binding and transport could be slightly different. The K m values for Cu ϩ were significantly above those expected to be encountered within the intracellular environment, so it is likely that some chelating molecules supply copper to CsHMA5.1 and CsHMA5.2 in cucumber root cells.
To confirm that the Cu ϩ -dependent activation of CsHMA5.1 and CsHMA5.2 is related to the ATP-dependent transport of copper across the membranes, copper uptake was measured in vacuolar membrane vesicles isolated from the ⌬ycf1 strain expressing CsHMA5.1-GFP and CsHMA5.2-GFP. Similarly to ATPase activities, copper transport by CsHMA5.1 or CsHMA5.2 was also dependent on metal concentration and increased with the increase of copper in reaction media up to 5 M (Fig. 7). Higher copper concentrations (10 M, 50 M) had an inhibitory effect on CsHMA5.1 and CsHMA5.2 transport activities (Fig. 7). When compared with CsHMA5.1, the copper transport activity of CsHMA5.2 was higher at the same metal concentrations. Similar to Cu-ATPase activity, the rate of copper transport by CsHMA5.1 or CsHMA5.2 followed normal Michaelis-Menten saturation kinetics over the range of 0 -5 M copper concentration (Fig. 7). The apparent K m values for CsHMA5.1-mediated and CsHMA5.2-mediated copper uptake into vacuolar membranes were close to 1 or 0.5 M, respectively (Fig. 7), as expected from the K m calculation for Cu-ATPase activities of cucumber copper pumps. These findings indicate that copper-induced activation of CsHMA5.1 and CsHMA5.2 is strongly correlated with copper transport across vacuolar membranes.

and CsHMA5.2 in Yeast Confers Tolerance to Copper and Silver and Increases Copper Accumulation within Yeast
Cells-To confirm the ability of CsHMA5.1 and CsHMA5.2 to transport copper and silver in vivo, the fusion proteins CsHMA5.1-GFP and CsHMA5.2-GFP were expressed in the yeast mutant strain ⌬ace1 that is unable to grow in high concentrations of copper and silver. Ace1 encodes a copper-dependent transcriptional activator of genes encoding proteins involved in copper detoxification and redox homeostasis in yeast, including the metallothionein proteins Cup1 and Crs5 as well as SOD1 (41)(42)(43)(44). Therefore, disruption of Ace1 renders yeast cells extremely sensitive to copper excess. The growth of mutants transformed with the empty vector pUG35 or vectors carrying CsHMA5.1 or CsHMA5.2 was monitored on the control and metal-supplemented media. The presence of CsHMA5.1-GFP and CsHMA5.2-GFP fusions in the vacuolar membranes of the mutant strain was confirmed by fluorescence microscopy (Fig. 8A). The copper-sensitive mutant phenotype was fully complemented by CsHMA5.1 or CsHMA5.2 (Fig. 8B). Both proteins also conferred increased resistance of the same yeast to silver (Fig. 8B). These data suggest that cucumber CsHMA5s contribute to copper and silver detoxification in yeast cells. To investigate further the mecha- The letters indicate the significant differences between CsHMA5.1 or CsHMA5.2 activities determined at different concentrations of metals (a) or between CsHMA5.1 and CsHMA5.2 activities measured at the same concentration of metal (b). Inset, the rate of ATP hydrolysis mediated by CsHMA5.1 or CsHMA5.2 versus metals concentrations measured in the presence of 0 -5 M copper (A) or 0-10 M silver (noninhibitory metal concentrations) in reaction media. The apparent K m value of ϳ1 or 2.5 M was calculated for the Cu-ATPase activity and Ag-ATPase activity, respectively, of CsHMA5.1. The copper-and silver-dependent activities of CsHMA5.2 had a K m value close to 0.5 or 2.4 M for copper and silver, respectively. V, the rate of ATPase activity (g P i ϫ mg Ϫ1 protein ϫ h Ϫ1 ). nism of CsHMA5.1-mediated and CsHMA5.2-mediated tolerance of ⌬ace1 to copper, the accumulation of copper was compared in mutant strains transformed with an empty vector or with vectors carrying cucumber genes. Analysis of metal content in cells growing in liquid medium supplemented with 100 M CuCl 2 showed that yeast expressing CsHMA5.1 or CsHMA5.2 accumulated a higher amount of copper during 8 h of exposure to metal than the strain transformed with the empty vector (Fig.  8C), suggesting that CsHMA5-mediated copper tolerance is related to the enhanced intracellular sequestration of copper rather than the increased copper efflux out of the yeast cells. This finding was in agreement with the vacuolar localization of CsHMA5s in yeast.
CsHMA5.1 and CsHMA5.2 Are Differentially Regulated by Copper Stress-Both CsHMA5.1 and CsHMA5.2 are expressed in cucumber roots. To determine how copper stress affects the level of CsHMA5.1 and CsHMA5.2 transcripts, quantitative PCR was performed using the bulk of roots from plants grown upon copper excess or copper deficiency. As shown in Fig. 9A, the expression of CsHMA5.1 was not significantly affected by any of the treatments. Similarly, the level of CsHMA5.1 protein in tonoplast membranes isolated from the same roots was not significantly changed upon copper stress (Fig. 9B). In contrast, CsHMA5.2 transcript and protein levels were markedly reduced or elevated upon copper deficiency or copper excess, respectively. These findings suggest that CsHMA5.1 is a constitutive root-specific vacuolar copper transporter unaffected during copper stress, whereas CsHMA5.2 functions in copper detoxification under copper excess.
ATP-dependent Copper Transport Operates in Tonoplast Vesicles from Cucumber Root Cells-Immunolocalization of CsHMA5.1 and CsHMA5.2 in the tonoplast as well as the effect of protein expression in yeast indicates that both ATPases play a role in active copper transport across the vacuolar membrane of cucumber cells. Indeed, ATP-dependent copper transport activity was detected in tonoplast vesicles prepared from the roots of cucumbers (Fig. 9C). Moreover, the rate of active copper accumulation within vacuolar membranes was significantly dependent on the availability of copper in the nutrient solution. Specifically, the ATP-dependent copper transport in vesicles prepared from plants grown under copper toxicity or copper deficiency was over 4-fold higher or almost 5-fold lower, respectively, when compared with the control (Fig. 9C). These results were consistent with the changes in CsHMA5.2 transcript and protein levels under copper stress, suggesting that CsHMA5.2 plays a role in the increased vacuolar copper sequestration in cucumber roots in conditions of copper excess.

Discussion
So far, only two genes encoding plant HMA5-like P 1B -ATPases have been functionally characterized as follows: OsHMA5 from rice and AtHMA5 from A. thaliana, respectively (17,18,45). They are expressed predominantly in pericycle cells of roots and are involved in Cu ϩ compartmentalization and detoxification, probably through active copper loading to the xylem of roots and other organs (17,18,45). Moreover, rice OsHMA5 has been localized in the plasma membrane, confirming the predicted role of plant HMA5s in copper export from root cells (18). Although AtHMA5 has not been localized yet, the knock-out of the gene encoding the Arabidopsis transporter results in increased hypersensitivity of plants to copper and enhanced accumulation of the metal in roots under copper excess, suggesting that AtHMA5 may be involved in copper efflux from root cells, and therefore it could also localize to the plasma membrane (17). However, neither OsHMA5 nor AtHMA5 has been subjected to biochemical and kinetic studies confirming their roles in copper distribution in plant cells. Although Arabidopsis possesses only one HMA5 pump, multiple genes encoding homologous proteins have been identified in other plants, including rice; however, the relevance of these extra Cu-ATPases is not known (19).
This work provides the first functional and biochemical characterization of plant HMA5-like P 1B -ATPases, CsHMA5.1 and CsHMA5.2 from cucumber, and it reveals that orthologous transporters in plants are differently displayed to achieve organismal copper homeostasis.
Contrary to plasma membrane-localized OsHMA5, CsHMA5.1 and CsHMA5.2 are associated with the vacuolar membrane, indicating that they function in vacuolar metal sequestration rather than in metal efflux out of the cells. However, different expression of CsHMA5.1 and CsHMA5.2 in cucumber organs suggests different biological roles for the two cucumber HMA5s. When studied in yeast vacuolar membranes, CsHMA5.1 and CsHMA5.2 were activated in the presence of copper and silver, similarly to homologous proteins in bacteria (CopA), humans (ATP7A and ATP7B), and A. thaliana (PAA1) (12,(35)(36)(37)(38), confirming that they are putative copper ATPases. Interestingly, the maximal copper-dependent ATPase activity of CsHMA5.1 and CsHMA5.2 required the presence of Cys and reducing agents (DTT and Na 2 SO 3 ) in the assay medium. Similar Cys-induced ATPase activity was previously reported for the Cu-ATPase CopA (A. fulgidus) and Cd-ATPases ZntA and AtHMA2 from Escherichia coli and A. thaliana, respectively (35,39,40). It has already been proposed that metal-thiolate complexes rather than metal ions are the substrates for heavy metal ATPases (39,40,46). Using a yeast two-hybrid assay, Andres-Colas et al. (17) demonstrated that the metal-binding domains of AtHMA5 interact with Arabidopsis ATX1-like copper chaperones, confirming that chelated copper rather than ionic Cu ϩ is a substrate for the copper pump. Here, we show that copper thiolate is also a putative substrate for the cucumber homolog of HMA5 pumps. Contrary to Cys, the positive effect of reducing agents on copper ATPase activity probably results from their action on the copper redox state (47,48). Reducing agents appear to be necessary to keep copper in a reduced state, a form activating phosphorylation of copper ATPases. Hence, both DTT and Na 2 SO 3 had a similar positive effect on CsHMA5.1 and CsHMA5.2 activities, probably by maintaining copper in a monovalent form (12). A strong negative effect of the Cu ϩ chelator BCS on CsHMA5.1 and CsHMA5.2 activities confirmed that Cu ϩ is required for cucumber protein activity. Interestingly, in the absence of reducing agents the copper-mediated activation of CsHMA5.1 and CsHMA5.2 was still significant when compared with the copper-deprived control assay, suggesting that in the conditions of our experiment copper was predominantly present in the Cu ϩ form. Indeed, recent studies on AtHMA6/PAA1 in membranes isolated from Lactococcus lactis demonstrated that a majority (70%) of copper in the reaction medium was rapidly reduced (in less than 5 min) by the buffer and isolated membranes (12).
The apparent K m values of CsHMA5.1 and CsHMA5.2 for Cu ϩ , estimated using ATPase and a transport assay, were slightly different (close to 1 and 0.5 M, respectively) and suggest that CsHMA5.2 binds copper more efficiently under copper-limiting conditions. Detailed studies on human ATP7B revealed that copper-mediated activation of ATPase is a result of copper binding to autoinhibitory CXXC motifs, inducing conformational changes within the protein that enable ATP hydrolysis (49,50). The copper ATPases from other organisms, including plants, are probably regulated by the same autoinhibitory mechanism. As shown in Fig. 2, two CXXC motifs are present in CsHMA5.1 (CSAC and CNSC) and CsHMA5.2 (CSAC and CTSC). The presence of different amino acids in one of the two motifs (asparagine in CsHMA5.1 instead of threonine in CsHMA2) could affect copper binding and thus Cu-ATPase activity of both pumps. Nevertheless, the apparent K m values of CsHMA5s for copper were very similar to those previously reported for human ATPases ATP7B and ATP7A (1 and 2.5 M, respectively) (37), for CopA from E. coli and from A. fulgidus (5 and 3.9 M, respectively) (35,36), and for plant AtHMA6/PAA1 (0.5 M) (12).
In comparison with Cu ϩ , the affinities of both cucumber pumps for Ag ϩ were very similar and lower (K m ϳ2.5 M), suggesting that the mechanism of Cu ϩ and Ag ϩ binding by these proteins might be different. Ag ϩ is not a physiological substrate for Cu-ATPases; hence, its stimulatory effect on activity of copper pumps probably results from the chemical (electropositive, singly charged cations) similarity between Ag ϩ and Cu ϩ ions (12). However, the difference in the Ag ϩ (0.126 nm) and Cu ϩ (0.096 nm) ionic radii could affect metal binding and thus ATPase activity of CsHMA5s. Differences in affinity to Cu ϩ and Ag ϩ were previously shown for CopA from Archaeoglobus fulgidus (35) and AtHMA6/PAA1 from A. thaliana (12), whereas CopA from Bacillus subtilis showed similar K m values for both ions (51). In addition, the rate of CsHMA5.1-mediated and CsHMA5.2-mediated ATP hydrolysis was higher in the presence of Cu ϩ than upon addition of an equal concentration of Ag ϩ to the reaction media. The phosphorylation of AtHMA6/PAA1 was also significantly (over 2-fold) higher in the presence of copper than in the presence of silver (12). In contrast, CopA from A. fulgidus was activated four times faster by Ag ϩ than by Cu ϩ (35), whereas CopA from B. subtilis was activated by both ions with similar efficiency (51). It has already been evidenced that copper ATPases from various organisms might behave similarly or differently with copper and silver. Moreover, CopA from A. fulgidus is mainly present in the ion-binding phosphorylated form in the presence of copper or in a metal-free phosphorylated form in the presence of silver (35). In contrast, the ion-binding phosphorylated form of AtHMA6/PAA1 is prevalent in the presence of either copper or silver (12).
Although the hydrolytic activities of CsHMA5.1 and CsHMA5.2 were copper-or silver-dependent, the highest concentration of metals resulted in the inhibition of ATP hydrolysis. A similar inhibitory effect of higher copper concentrations was observed when the ATP-dependent copper transport was assayed. Excessive copper also inhibited the ATPase activity of prokaryotic copper ATPases (CopB from Enterococcus hirae and A. fulgidus) as well as human ATP7A and plant AtHMA6/ PAA1 (11,12,38,52,53). High concentrations of metal ions could prevent rapid dissociation of the transported ions from the metal-binding site of P 1B -type ATPase and thus reduce the enzyme turnover rate (12). However, copper-mediated inhibition of P 1B -ATPase may result from the negative effect of metal excess on the protein structure, on the active sites of the enzyme, or on the membrane lipids. Excessive copper was found to inhibit protein activity by inducing protein precipitation or by inducing conformational changes at several key residues of proteins (54,55).
The ability of CsHMA5.1 and CsHMA5.2 to transport silver and copper was confirmed by studying copper and silver tolerance of yeast expressing cucumber proteins. Both CsHMA5s increased yeast tolerance to copper and silver through intracellular copper sequestration, indicating that they actually transport Cu ϩ and Ag ϩ in vivo. Similarly, AtHMA5, AtHMA6/PAA1, and OsHMA5 have been shown to participate in intracellular copper sequestration when expressed in the ⌬ccc2 yeast strain, lacking the Golgi-resident endogenous copper Cu ϩ /Ag ϩ -ATPase Ccc2p (12,18,45). According to available data, P 1B1 -ATPases participate in the maintenance of copper homeostasis either through the import of copper into the cell (PCC7942 and PCC6803 in cyanobacteria Synechococcus and Synechocystis, respectively) and intracellular compartments to supply copper to copper-dependent enzymes (ATP7A, ATP7B, Ccc2p, AtHMA6/PAA1, AtHMA8/PAA2, and AtHMA7/RAN1) or through the active efflux of copper out of the cell (CopA, OsHMA5) (12,18,56,57). In addition, our study provides the first evidence for the involvement of P 1B1 -ATPases in the active influx of copper into vacuoles. The subcellular localization of CsHMA5.1 and CsHMA5.2 and increased tolerance of yeast expressing cucumber pumps to copper clearly indicate the function of both proteins in detoxification of cells from copper through the vacuolar sequestration of copper excess. The expression of homologous proteins in A. thaliana and rice was up-regulated by excess copper, confirming the putative function of AtHMA5 and OsHMA5 in copper detoxification (17,18). Interestingly, CsHMA5.1 and CsHMA5.2 expression and protein levels were differentially affected by copper stress. However, CsHMA5.1 was not affected by different copper availability, whereas CsHMA5.2 was significantly up-regulated or down-regulated upon copper excess or deficiency, respectively. Similarly, the ATP-dependent copper transport into tonoplast membranes isolated from cucumber roots was also strongly increased under copper excess and markedly reduced upon copper deficiency. These findings indicate that CsHMA5.2 might be responsible for the increased vacuolar copper sequestration in cucumber root cells under copper toxicity and reveal different regulation of the two CsHMA5 isoforms by copper. Hence, we propose that the functional diversity of CsHMA5.1 and CsHMA5.2 comes from variation in their regulatory mechanisms (different expression patterns) and possibly from some structural differences between both proteins (different affinities for copper and different ATPase activities).