Zinc transporters, ZnT5 and ZnT7, are required for the activation of alkaline phosphatases, zinc-requiring enzymes that are glycosylphosphatidylinositol-anchored to the cytoplasmic membrane.

Numerous proteins are properly folded by binding with zinc during their itinerary in the biosynthetic-secretory pathway. Several transporters have been implicated in the zinc entry into secretory compartments from cytosol, but their precise roles are poorly understood. We report here that two zinc transporters (ZnT5 and ZnT7) localized in the secretory apparatus are responsible for loading zinc to alkaline phosphatases (ALPs) that are glycosylphosphatidylinositol-anchored membrane proteins exposed to the extracellular site. Disruption of the ZnT5 gene in DT40 cells decreased the ALP activity to 45% of that in the wild-type cells. Disruption of the ZnT7 gene lowered the ALP activity only by 20%. Disruption of both genes markedly decreased the ALP activity to <5%. Overexpression of human ZnT5 or ZnT7 in DT40 cells deficient in both ZnT5 and ZnT7 genes recovered the ALP activity to the level comparable to that in the wild-type cells. The inactive ALP protein in DT40 cells deficient in both ZnT5 and ZnT7 genes was transported to cytoplasmic membrane like the active ALP protein in the wild-type cells. Thus both ZnT5 and ZnT7 contribute to the conversion of apo-ALP to holo-ALP.

Zinc is an essential trace element required as a structural component of a large number of proteins such as metalloenzymes and transcription factors (1,2). Because zinc can not freely pass the membrane of cells and intracellular organelles, zinc transporter proteins have been thought to play a critical role in regulating the zinc concentration in the cytosol and lumen of organelles. Recently, a number of zinc transporters have been cloned in vertebrates (3)(4)(5)(6). Based on their sequence homology, and structural and functional properties, they have been assigned to two metal-transporter families: ZIP 1 (ZRT/ IRT-related protein) family and CDF (cation diffusion facilitator) family (3)(4)(5)(6)(7)(8). ZIP family transporters function in zinc influx into cytosol from extracellular sites and the lumen of intracellular organelles. CDF family transporters facilitate zinc efflux from cytosol and mobilize the cytosolic zinc into intracellular organelles. This mobilization of zinc into organelles may supply zinc to zinc-dependent proteins in situ and may also contribute to zinc storage and/or detoxification. To date, eight CDF members, named ZnT1-8 (zinc transporter [1][2][3][4][5][6][7][8], have been characterized in mammals (4 -6, 8, 9). Of these mammalian CDF members, ZnT5-7 are located in the Golgi apparatus and implicated in the entry of zinc into the Golgi lumen (10 -12). ZnT5-7, if not all, may be responsible for loading zinc to secretory, membrane-bound, or organelle-resident proteins that require zinc for manifestation of their biological activities, but evidence supporting this assumption has been lacking.
There are a number of secretory or membrane-bound enzymes that require zinc as an essential component; e.g. matrix metalloproteinases (13), angiotensin-converting enzyme (14), carboxypeptidases (15), and alkaline phosphatases (ALPs) (16). ALPs are a group of glycosylphosphatidylinositol (GPI)-anchored membrane proteins exposed to the extracellular site (17). In mammals, four types of ALPs, tissue-nonspecific ALP (TNAP), placental ALP (PLAP), intestinal ALP (IAP) and germ cell ALP, have been identified. They contain two tightly bound zinc ions per polypeptide in their active site (17), and the binding of zinc likely occurs in the biosynthetic-secretory pathway before reaching their final destination. Dimerization yields the active enzyme (18).
DT40 is a chicken B lymphocyte-derived cell line with a unique property that the homologous recombination activity is high, and thereby genes of interest can be disrupted at a high frequency (19,20). Furthermore, as described later, it contains counterparts of mammalian ZnT1-7 genes, and most of their mRNAs are expressed. DT40 cells also express TNAP at the extracellular site of their cytoplasmic membrane. Using DT40 cells and ALPs, here we describe the first evidence that both ZnT5 and ZnT7 are required for loading zinc to ALPs in the biosynthetic-secretory pathway.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Fetal calf serum (FCS, Trace Scientific Ltd.) was heat-inactivated. HeLa and HEp2 cells were maintained in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% FCS at 37°C. DT40 cells (gift from Dr. Shunichi Takeda) were maintained in RPMI1640 (Nacalai) supplemented with 10% FCS, 1% chicken serum (JRH), and 10 M 2-mercaptoethanol (Sigma) at 39.5°C. Zincfree serum was made according to the instruction manual by treating 50 ml of serum (mixture of FCS and chicken serum in 10:1 by volume ratio) with Chelex-100 resin (Bio-Rad). The resin was removed by filtration of the Chelex-treated serum. DNA transfection into DT40 cells was carried out according to the methods described elsewhere (21). Briefly, 1 ϫ 10 7 cells were suspended in 0.5 ml of cold phosphate-buffered saline (PBS) containing 30 g of linearized plasmid DNA. DNA was electroporated into the cells with a GenePulser apparatus (Bio-Rad) at 550 V and 25 microfarads. After electroporation, cells were transferred into 20 ml of fresh medium and incubated for 24 h. Then, cells were resuspended into 80 ml of fresh medium containing appropriate drugs and divided into four 96-well plates. After 7 days, drug-resistant clones were transferred to 24-well plates and cultured for 2 days. Then the clones were transferred to 100-mm dishes (10-ml cultures). After culture for 2 days, genomic DNA was extracted and analyzed by Southern blotting. The concentrations of drugs used for screening were 30 g/ml for Blasticidin-S (Calbiochem), 2 mg/ml for G418 (Nacalai), 0.5 g/ml for puromycin (Sigma), 1 mg/ml for L-histidinol (Sigma), and 0.3-1 mg/ml for Zeocin (Invitrogen). Over five independent clones per transfectant were established.
Plasmids Construction-The ϳ9-kb chicken ZnT5 (cZnT5) or ϳ11-kb cZnT7 gene was amplified with a pair of gene-specific primers using DT40 genomic DNA as a template by LA Taq (TAKARA). Each amplified DNA fragment was used to map the restriction sites. The long or short arm was PCR-amplified and subcloned upstream or downstream of the drug-resistant genes (Bsr r , His r , and Neo r ) under the control of ␤-actin promoter (drug selection marker cassettes, gift from Dr. Shunichi Takeda). Plasmids to express human ZnT5 (hZnT5), cZnT5, or hZnT7 were constructed by inserting each cDNA into the pA-Puro vector (22). The PLAP cDNA was made by ligating the fragment of secreted alkaline phosphatase gene prepared from pSEAP2-Basic vector (Clontech) with the fragment of RT-PCR-amplified PLAP cDNA encoding C-terminal region of PLAP. The PLAP cDNA was inserted into a pA-Zeocin vector to construct PLAP expression plasmid. The pA-Zeocin vector was constructed by replacing the Puro r gene in the pA-Puro vector with the Zeo r gene. All plasmids were linearized with appropriate restriction enzymes prior to electroporation.
Fluorescent in Situ Hybridization Analysis-Fluorescent in situ hybridization analysis was performed as described previously (23). Briefly, chromosome spreads were generated from DT40 cells after thymidine synchronization and bromodeoxyuridine incorporation. The cZnT5 or cZnT7 genomic DNA was labeled with biotin-16-dUTP (Roche Applied Science) by nick translation and used as a probe. After in situ hybridization, the hybridized probe was detected using fluorescein isothiocyanate-conjugated avidin (Roche Applied Science).
Southern Blotting-Twenty micrograms of genomic DNA prepared from DT40 cells were digested with appropriate restriction enzymes (shown in Fig. 2) for 24 h. After digestion, DNA was electrophoresed on agarose gel and transferred to nitrocellulose membrane filter in 0.4 N NaOH. The membrane was hybridized to the radiolabeled DNA probes (shown in Fig. 2) and exposed to an imaging plate (Fuji). Radioimages were obtained using a BAS 2000 Bio imaging analyzer (Fuji).

RNA Preparation and Northern
Blotting-Total RNA was extracted from DT40 cells using Sepasol I (Nacalai). The RNA (20 g) was electrophoresed on agarose gel and transferred to nitrocellulose membrane filter in 20ϫ SSC buffer. The membrane was hybridized to the appropriately radiolabeled cDNA probes. Radioimages were obtained as described above.
Preparation of Membrane Fraction-DT40 cells (ϳ2 ϫ 10 7 ) were collected and washed with PBS. These cells were resuspended in 2 ml of cold homogenizing buffer (0.25 M sucrose, 20 mM HEPES, and 1 mM EDTA) and homogenized with 20 strokes of a 7-ml Dounce homogenizer. ALPs firmly bind with zinc, and therefore the presence of 1 mM EDTA in the homogenizing buffer does not affect the enzyme activity. To remove the nucleus, the homogenate was centrifuged at 2,200 rpm for 10 min at 4°C. The post-nuclear supernatant was centrifuged at 12,000 rpm for 60 min at 4°C. The pellet, which contained cytoplasmic and organelle membranes, was lysed in ALP lysis buffer (10 mM Tris-HCl, pH 7.5, 0.5 mM MgCl 2 , and 0.1% Triton X-100), and stocked at Ϫ70°C until use. For fractionation of the organelles, the membrane was centrifuged on a 0.7-1.7 M sucrose gradient as described previously (10) and separated into 10 fractions. The concentration of protein was determined with a protein assay kit (Bio-Rad) using bovine serum albumin as a standard Immunoblot Analysis-Membrane proteins were mixed with 6ϫ SDS sample buffer and then incubated at 37°C for 10 min, or total cellular proteins were lysed in the same buffer and then boiled at 100°C for 5 min before electrophoresis. Proteins were separated with electrophoresis through 8% SDS-polyacrylamide gels. After transfer of proteins to a nitrocellulose membrane (Hybond-ECL, Amersham Biosciences), the blot was blocked with blocking solution (5% skim milk and 0.1% Tween 20 in PBS) and then incubated with anti-ZnT5 (1:2000 dilution), anti-GM130 (BD Transduction Laboratories; 1:250 dilution), anti-SEC23 (Santa Cruz Biotechnology; 1:100 dilution), anti-EEA1 (BD Transduction Laboratories; 1:1000 dilution), anti-␥1-adaptin (Santa Cruz Biotechnology; 1:100 dilution), anti-calreticulin (Affinity Bioreagents; 1:1000 dilution), anti-PLAP (GenHunter; 1:2000 dilution), or anti-calnexin (Stressgen; 1:1000 dilution) antibody in blocking solution. Horseradish peroxidase-conjugated secondary antibodies (Amersham Biosciences) were used at a 1:3000 dilution, and Super Signal Chemiluminescent Substrate (PIERCE) was used for the detection. The intensity of bands was measured using the Image Gauge (Fuji).
Immunofluorescence Staining-Immunofluorescence for ZnT5 was performed as described previously (10). Briefly, HeLa or HEp2 cells were fixed with 4% paraformaldehyde or methanol at Ϫ20°C. The cells were incubated with anti-ZnT5 antibody followed by goat anti-mouse IgG and donkey anti-goat IgG conjugated with Alexa 488 or 594 (Molecular Probe). Immunofluorescence for PLAP in DT40 cells was performed using the same procedures as for ZnT5 except for the fixation step. DT40 cells were harvested, washed with PBS, and incubated on coverslips coated with poly-L-lysine for 30 min following the fixation with methanol at Ϫ20°C for 20 min. The anti-PLAP antibody (Santa Cruz Biotechnology) was used at a dilution of 1:100. The stained cells were observed under a confocal laser-scanning microscope (Fluoview, Olympus).
Measurement of ALP Activity-The membrane protein (1 g) or total cellular protein (10 g) lysed in ALP lysis buffer was preincubated for 10 min at room temperature. A 100-l substrate solution (2 mg/ml p-nitrophenyl phosphate in 1 M diethanolamine buffer, pH 9.8, containing 0.5 mM MgCl 2 ) was added. After incubation for 10 min at room temperature, p-nitrophenol released by TNAP was measured by the absorbance at 405 nm. For measurement of PLAP activity, samples were kept at 65°C for 30 min to inactivate TNAP. Shrimp ALP (Roche Applied Science) was used as a standard.
PI-PLC Digestion and Immunoprecipitation-DT40 cells (2 ϫ 10 6 ) expressing PLAP were incubated in culture medium for 3 h and collected. These cells were washed with PBS and incubated with or without 0.2 unit of phosphatidylinositol-specific phospholipase C (PI-PLC) (Molecular Probes) in 0.5 ml of PBS for 1 h at 4°C. The cell suspension was centrifuged for 10 min at 4°C, and the resultant supernatant was subjected to immunoprecipitation with anti-PLAP antibody conjugated to agarose (Sigma) to concentrate PLAP detached from cell surface by PI-PLC digestion. The PLAP-complex beads were treated with 6ϫ SDS sample buffer at 100°C for 5 min, electrophoresed as described above, and immunoblotted.

Subcellular
Localization of ZnT5 in HeLa and HEp2-We previously identified ZnT5 as a zinc transporter that was abundantly expressed in pancreatic ␤ cells and was associated with ZnT5 and ZnT7 Are Essential for ALP Activity insulin granules in the cells, which suggested that ZnT5 transports zinc into insulin granules to form an insulin hexamer containing two zinc ions (10). Messenger RNA of ZnT5, however, was detected ubiquitously (10,24,25), indicating that ZnT5 possesses other functions. When ZnT5 was overexpressed, its localization overlapped with that of the Golgi marker protein, GM130 (10), or Golgi 58K protein (24). Because localization of the endogenous ZnT5 remains to be determined, we immunostained the cells using antibodies against hZnT5 and various marker proteins of organelles. As shown in Fig. 1A, ZnT5 was found in the cytoplasmic vesicles and perinuclear portion in HeLa and HEp2 cells. The immunostaining against marker proteins, however, failed to reveal the specific marker that was completely superimposed with ZnT5 (data not shown). Then, we fractionated HeLa-derived organelles by sucrose gradient centrifugation, and the distribution of ZnT5 was compared with that of various organelle markers. Fig. 1B shows the immunoblot of proteins in the fractionated organelles. ZnT5 was mainly detected in fractions 5 and 6, which was similar to Sec23, a marker protein of COPII vesicles that bud from endoplasmic reticulum (ER) and carry cargo to the cis-Golgi, and GM130, a marker of Golgi apparatus. Distribution of ZnT5 differed from that of EEA1 (endosomes), ␥1-adaptin (AP-1 vesicles), and calreticulin (ER). AP-1 vesicles bud from the trans side of the Golgi apparatus, which faces cytoplasmic membrane where secretion occurs. The antibody against ␥1adaptin used here recognizes two types of ␥1-adaptin with different sizes. Because ER produces vesicles with various densities during preparation, its marker protein (calreticulin) was detected in most fractions. The same results were obtained when another marker, calnexin, was used (data not shown). These results indicate that the endogenous ZnT5 is concentrated in early compartments of the secretory pathway such as COPII-coated vesicles and Golgi apparatus.
Expression of ZnT Members in DT40 Cells-To find whether the chicken-derived DT40 cell line expresses the orthologues of mammalian CDF family members, we examined their mRNA expression in DT40 cells. Comparison of the chicken expressed sequence tag sequences with mammalian ZnT1-8 revealed nucleotide fragments homologous to ZnT1-7 but not to ZnT8 (Table I). Detection of mRNA in DT40 cells by RT-PCR confirmed expression of cZnT1-7 except for cZnT3 (Table I), indicating that DT40 is suitable for studying ZnT members.
Generation of DT40 Cells Deficient in ZnT5-To investigate the cellular function of ZnT5, we generated ZnT5-deficient DT40 cells. Because fluorescent in situ hybridization analysis revealed that the cZnT5 gene resides in a monosome (data not shown), ZnT5-deficient DT40 cells are referred to as ZnT5 Ϫ cells. Fig. 2A shows the strategy for disruption of ZnT5 gene and data to confirm its absence in ZnT5 Ϫ cells. As described later, we examined DT40 cells deficient in ZnT5 and/or ZnT7 genes, and therefore the strategy for disruption of cZnT7 gene, including data confirming its absence, is also shown in Fig. 2B. Chicken ZnT7 resides in the somatic chromosome (data not shown). Disruption of cZnT5 or cZnT7 did not give apparent detrimental effects; ZnT5 Ϫ and ZnT7 Ϫ/Ϫ cells proliferate at a rate comparable to that of the wild-type cells, and there is no difference between mutant and wild-type cells in the sensitivity to zinc toxicity. Messenger RNA of cZnT5 encodes a 770-amino acid protein with 88% identity to hZnT5, whereas that of cZnT7 encodes a 378-amino acid protein with 83% identity to hZnT7 (see accession number AY703476 for cZnT5 cDNA and AY703477 for cZnT7 cDNA).
Four distinct genes encoding ALP have been identified in mammals, whereas two of these forms, IAP and TNAP, have been reported in birds (26). In DT40 cells, the TNAP mRNA alone was detected by RT-PCR and ALP activity in DT40 cells was completely inhibited by levamisole, a specific inhibitor of TNAP, which indicated that DT40 cells exclusively express TNAP of ALPs. TNAP activity was assayed by the use of membrane in the post-nuclear fraction. TNAP activity was severely reduced in the wild-type DT40 cultured for 18 h in the zinc-free medium containing Chelex-treated FCS, and the activity was fully restored when zinc was supplemented to the zinc-free medium (Fig. 3A, upper panel). There was no change in TNAP mRNA content at any zinc concentrations tested (Fig.  3A, lower panel). Little activity of TNAP was detected in the membrane prepared from the cells cultured in the zinc-free medium and washed with zinc-containing buffer. This result excludes the possibility that zinc-deprived TNAP of the cells cultured in the zinc-free medium quickly recovers its activity by re-binding with zinc in the washing buffer. Thus TNAP   (33) and the TIGR G. gallus Gene Index web server (www.tigr.org/tigr-scripts/tgi/T_index. cgi?speciesϭg_gallus). The following sequences were used for this analysis; ZnT1 (Q9Y6M5), ZnT2 (AAB02775), ZnT3 (Q99726), ZnT4 (O14863), ZnT5 (AAM09099), ZnT6 (NP_060434), ZnT7 (AAM21969), and ZnT8 (AAM80562). These proteins are systematically renamed as SLC30A1-8 in this order. All sequences except for ZnT2 (rat) are human sequences. For detection of mRNA expression in DT40 cells, specific primers were designed for each gene and used for RT-PCR amplification.

Gene
Chicken ESTs Expression in DT40 cells Ϫ NE a ϩ, Ϫ, D, UD, and NE indicate presence, absence, detectable, undetectable, and not examined, respectively.
ZnT5 and ZnT7 Are Essential for ALP Activity activity is a good reporter of zinc metabolism in the biosynthetic-secretory pathway in DT40 cells.
Next, we examined the TNAP activity of ZnT5 Ϫ cells in comparison with that of wild-type cells. TNAP activity in ZnT5 Ϫ cells was reduced to about 45% of that in wild-type cells (Fig. 3B, upper panel) despite no significant difference in expression of TNAP mRNA (Fig. 3B, lower panel). This reduction was restored by expressing the exogenous ZnT5 cDNA from either chicken or human (about 80% or 90% activity compared with that of wild-type cells) (Fig. 3B, upper panel). Overexpression of hZnT5 in the wild-type DT40 did not increase TNAP activity (Fig. 3B, upper panel), indicating that zinc influx to secretory compartments by endogenous transporters is sufficient for the full expression of TNAP activity.
The overexpressed hZnT5 in DT40 cells was localized mainly in secretory compartments such as Golgi apparatus with immunofluorescence staining of the cells or immunoblotting after fractionation of organelles by sucrose gradient centrifugation (data not shown). Similar localization in DT40 cells was found for hZnT7 tagged with FLAG epitope (data not shown).
ZnT5 and ZnT7 Are Required for the Full Activation of TNAP-Disruption of ZnT5 gene reduced TNAP activity, but 45% of the activity remained, suggesting that other transporters may also be involved in expression of TNAP activity. Kirschke and Huang (12) previously reported that ZnT7 residing in Golgi apparatus may transport zinc into the Golgi lumen, because overexpression of ZnT7 in CHO cells accumulates zinc in Golgi apparatus in a high zinc condition. To examine contribution of ZnT7 to TNAP activity in DT40 cells, we generated ZnT7-deficient DT40 cells (ZnT7 Ϫ/Ϫ cells) (Fig. 2B) and compared their phenotypes with those of ZnT5 Ϫ cells and wild-type cells. ZnT7 Ϫ/Ϫ cells showed lower TNAP activity than wild-type cells (80% of the wild-type), but this reduction was smaller than that found in ZnT5 Ϫ cells (Fig. 4A, upper panel). Thus contribution of ZnT7 to TNAP activity appears to be minor. TNAP activity was severely decreased in DT40 cells deficient in both ZnT5 and ZnT7 (ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells) (Fig. 4A, upper panel). The TNAP activity in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was restored to 75 and 85% of that in the wild-type cells by expression of human ZnT5 and ZnT7, respectively. In all cases, the level of TNAP mRNA was unchanged (Fig. 4A, lower panel). ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells showed slightly slow growth; doubling time for ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was 9.5 h, while that for wild-type cells was 7.3 h. Microscopic morphology of ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was normal and the sensitivity to a high zinc concentration was similar to that of the wild-type cells. Like the wild-type DT40 cells (see Fig. 3A, upper panel), the TNAP activity in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells restored by expressing hZnT5 or hZnT7 disappeared when cultured in the zinc-free medium, but the addition of zinc to the medium maintained the activity (Fig. 4B,  upper panel). Levels of mRNAs for hZnT5 and hZnT7 expressed in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells were unchanged under the culture conditions tested (Fig. 4B, lower panel). The addition of cobalt or manganese to the zinc-free medium had no effects (data not shown). We conclude from these results that both ZnT5 and ZnT7 are required for expression of the full TNAP activity.
ZnT5 and ZnT7 Contribute to the Conversion of Apo-ALP to Holo-ALP-Marked reduction of TNAP activity in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells (Fig. 4A, upper panel) raises two possibilities; TNAP protein may not be present in this mutant cell although its mRNA is present at a normal level (Fig. 4A, lower  panel) or TNAP protein may be present as an inactive enzyme. To examine which is the case, we used PLAP as a reporter protein, because the antibody against PLAP was available but that against chicken TNAP was not, and DT40 cells lack PLAP. Moreover, TNAP is heat-unstable, whereas PLAP is heat-resistant; at 65°C, only TNAP is completely inactivated and thereby the exogenous PLAP activity expressed in DT40 cells could be clearly distinguished from the endogenous TNAP activity (Fig. 5A).
We established both wild-type and ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells that stably express PLAP mRNA at a similar level (Fig. 5C) and examined PLAP activity and its protein level. Consistent with the endogenous TNAP, PLAP activity expressed in the wildtype DT40 cells was also dependent on zinc in the culture medium (Fig. 5B). PLAP activity was reduced in the zinc-free medium, but PLAP was less susceptible to zinc deficiency than TNAP; 25% of PLAP activity remained after culture in the zinc-free medium, whereas TNAP activity was severely reduced (see Fig. 3A). PLAP protein may have a slow turnover rate so that the active enzyme bound with zinc before cell culture in the zinc-free medium remains to exhibit PLAP activity. It is also possible that PLAP protein has a high affinity to zinc as compared with that of TNAP and therefore some active PLAP molecules can be formed even in zinc-free medium by binding with zinc that would exist at a very low concentration. The PLAP activity was fully restored in the cells cultured in the medium supplemented with zinc (Fig. 5B).
When PLAP was expressed in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells, its activity was very low (Fig. 5D). It is noted that immunoblot analysis indicates that PLAP protein was present in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells at a level similar to that in wild-type cells (Fig. 5E). Expression of hZnT5 or hZnT7 in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells restored PLAP activity to the level comparable to that in wild-type cells (Fig. 5D) without affecting PLAP protein level (Fig. 5E). These results suggest that PLAP in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells fails in formation of the active conformation due to deficiency of zinc in the biosynthetic-secretory pathway. The same would be true for the endogenous TNAP.
Subsequently, we examined the localization of PLAP in the wild-type and ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells. Immunofluorescence analysis showed that PLAP protein was mainly localized to cytoplasmic membrane in both cells, but some vesicles were also positive (Fig. 6A). This localization of PLAP was ensured by treating the cells with PI-PLC that releases GPI-anchored protein from the cell surface. PLAP released from the cell surface by PI-PLC digestion was detected by immunoblot analysis after immunoprecipitation with anti-PLAP antibody. As shown in Fig. 6B, the amount of PLAP released from ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was similar to that released from the wild-type cells. PLAP released from the wild-type cells showed enzyme activity, but that from ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells did not (Fig. 6C), indicating that PLAP was transported to plasma membrane despite having little activity.
Lastly, we investigated whether the addition of zinc into the culture medium restores PLAP activity in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells, because the culture medium contains low micromolar level of zinc derived from FCS (27), and this level may be insufficient for formation of the active enzyme. When the wild-type cells and ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells were cultured for 18 h in the medium containing 50 M ZnSO 4 , a slight increase of PLAP activity was seen in both cells but the enzyme activity of ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was much less than that of the wild-type cells (Fig. 6D). Taken together, both ZnT5 and ZnT7 are essential for zinc incorporation into ALPs in secretory pathway to convert apo-ALPs to holo-ALPs. DISCUSSION CDF family members transport cytosolic zinc to the extracellular domain or into the lumens of intracellular organelles (3)(4)(5)(6)8). Of these members, ZnT5-7 have been shown to reside in the Golgi apparatus (10 -12), suggesting that they are responsible for the zinc entry into the lumens thereby loading zinc to proteins in the biosynthetic-secretory pathway, but this hypothesis has not been verified. In this study, we focused on ZnT5 and ZnT7, because they are highly homologous in their cation efflux domains (pfam01545) that are essential for zinc FIG. 4. ZnT5 and ZnT7 are required for the full activation of TNAP. A, ZnT5 and ZnT7 are required for TNAP activity. TNAP activity was measured using the membrane protein prepared from the indicated cells (WT, ZnT5 Ϫ , ZnT7 Ϫ/Ϫ , ZnT5 Ϫ ZnT7 Ϫ/Ϫ , hZnT5-overexpressed ZnT5 Ϫ ZnT7 Ϫ/Ϫ , hZnT7-overexpressed ZnT5 Ϫ ZnT7 Ϫ/Ϫ , and ZnT5 Ϫ ZnT7 Ϫ/Ϫ transfected with vector alone). Transfection with vector alone is shown as vec. Lower panels show expression of TNAP and ␤-actin mRNAs. B, hZnT5-and hZnT7-dependent restoration of TNAP activity in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells requires zinc. Cells were cultured in normal medium (N), zincfree medium containing Chelex-treated FCS (C), or zinc-free medium supplemented with 2 M ZnSO 4 (Cϩ2) for 18 h, and the membrane protein was assayed for TNAP activity. Expression of hZnT5 or hZnT7 mRNA was not changed in culture conditions (lower panels). Each value in A and B is the mean Ϯ S.D. of triplicate experiments, and the panels for Northern blotting in A and B show the representative results. transport, although ZnT5 has a unique and very long aminoterminal domain whose function is unknown (5,8,25). DT40 cells were examined for their usefulness to investigate the role of ZnT5 and ZnT7 in the secretory pathway. Fortunately, this cell line expressed both transporters and TNAP, a member of the ALP family, which is anchored to the extracellular domain of cytoplasmic membrane via GPI (17). ALPs contain tightly bound zinc essential for the enzyme activity, and zinc is thought to be loaded in the biosynthetic-secretory pathway. The activity of ALPs, therefore, can be a reporter of zinc metabolism in this pathway. Clear dependence of the TNAP activity in DT40 on zinc in the culture medium indicated that this cell line would be competent for our study.
A partial decrease of the TNAP activity in DT40 deficient in ZnT5 or ZnT7 and marked loss of the activity in the mutant lacking both genes clearly indicate that both transporters are required for the full expression of TNAP activity. Disruption of the ZnT5 gene caused a partial decrease of TNAP activity, and the residual activity was 45% of that in the wild-type DT40 cells. Disruption of ZnT7 gene resulted in a partial but small decrease; 80% of the TNAP activity remained. These results suggest that the total zinc-mobilizing activity by ZnT5 and ZnT7 is sufficient for the full activation of TNAP molecules produced in DT40 cells. This notion is supported by the fact that overexpression of ZnT5 in the wild-type cells does not increase TNAP activity. Partial responsibility of ZnT5 in the activation of TNAP appears to hold true in mammals. TNAP is crucial for postnatal bone mineralization (28). The ZnT5 Ϫ/Ϫ mouse shows osteopenia due to impairment of osteoblast maturation, and partial reduction of TNAP activity is seen in osteoblasts differentiated in vitro from precursor cells prepared from this mouse (24).
DT40 cells express ZnT5 and ZnT7 that both function to activate TNAP. Such redundant expression is not necessarily seen in all types of mammalian cells; ZnT5 is ubiquitously expressed, but ZnT7 is less widely expressed (10,12,25). Although the physiological significance of the redundant expression of zinc transporters is not known, another transporter(s) may substitute for ZnT7 in ZnT5-positive but ZnT7-negative cells. The best candidate is ZnT6, because it resides around the trans-Golgi network (11). However, ZnT6 lacks the His-rich region in the cytoplasmic loop, a characteristic of ZnT members, and two His residues in transmembrane domains, which FIG. 5. PLAP exogenously expressed in DT40 cells. A, the exogenous PLAP activity expressed in DT40 cells was clearly distinguished from the endogenous TNAP activity. The total cellular protein was prepared from the wild-type DT40 cells or the wild-type DT40 cells expressing PLAP. The protein preparations were incubated at 65°C for 30 min (black bar) or not (white bar). The ALP activity was assayed after cooling the preparations at 4°C for 10 min and is shown as milliunits (designated mU) per milligram cellular protein. B, exogenously expressed PLAP activity is dependent on zinc in culture medium. The wild-type DT40 cells expressing PLAP were cultured for 18 h in normal medium (N), zinc-free medium containing Chelex-treated FCS (C), or zinc-free medium supplemented with 2 or 50 M ZnSO 4 (Cϩ2 and Cϩ50, respectively). PLAP activity of the total cellular protein was assayed and is shown as milliunits per milligram cellular protein. C, Northern blot analysis for the detection of PLAP mRNA in WT, ZnT5 Ϫ ZnT7 Ϫ/Ϫ , hZnT5-overexpressed ZnT5 Ϫ ZnT7 Ϫ/Ϫ , and hZnT7overexpressed ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells. Messenger RNA of ␤-actin is shown as a control. D, PLAP activity in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells is significantly reduced but recovered by the expression of either hZnT5 or hZnT7. PLAP activity of WT, ZnT5 Ϫ ZnT7 Ϫ/Ϫ , hZnT5-overexpressed ZnT5 Ϫ ZnT7 Ϫ/Ϫ , and hZnT7-overexpressed ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was assayed. E, expression level of PLAP protein in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was almost the same as in wild-type cells. 20 g of total cellular protein was loaded in each lane. Calnexin is shown as a control. Each value in A, B, and D is the mean Ϯ S.D. of triplicate experiments.
FIG. 6. PLAP in ZnT5 ؊ ZnT7 ؊/؊ cells is transported to cytoplasmic membrane despite having little activity. A, immunostaining of PLAP in WT (upper panels) or ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells (lower panels). These cells were fixed with methanol at Ϫ20°C and stained with anti-PLAP antibody, followed by second and third antibodies conjugated with Alexa 488. B, PLAP is transported to cytoplasmic membrane in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells as it is in WT. Cells expressing PLAP were washed with PBS and incubated with (ϩ) or without (Ϫ) PI-PLC for 1 h at 4°C. PLAP anchored to cytoplasmic membrane is released into the supernatant. The aliquot of the supernatant was subjected to immunoprecipitation with anti-PLAP antibody, followed by immunoblot analysis. C, PLAP activity in the supernatant was measured and is shown as milliunits per 0.5 ml of supernatant (sup). D, PLAP activity in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells could not be recovered by the addition of zinc into culture medium. PLAP activity was measured using total cellular protein from WT or ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells cultured for 18 h in the presence (Zn) or absence (Ϫ) of 50 M ZnSO 4 . Each value in C and D is the mean Ϯ S.D. of triplicate experiments.
are highly conserved among all ZnT members and thought to be important for transporter function, have been altered to Leu and Phe (5). These structural features suggest that ZnT6 may not be as efficient as ZnT5 and ZnT7. This may be supported by our observation that TNAP activity is largely lost in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells, although they express ZnT6.
In this study, we examined two ALPs as reporter proteins of zinc metabolism in the secretory pathway. There are many zinc proteins that travel in the biosynthetic-secretory pathway. To know whether ZnT5 and ZnT7 also participate in zinc loading to these proteins, further studies using these proteins as reporters are needed. In this respect, GPI-phosphoethanolamine transferase (GPI-PET), a member of the ALP superfamily enzymes (29), which adds phosphoethanolamine groups to GPI anchors in ER and requires zinc (30), is of interest. The activity of PLAP expressed in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was significantly reduced, but the PLAP protein was exposed to the outside of cytoplasmic membrane via GPI anchor as it is in the wild-type cells. Among the various possible explanations for the activity of GPI-PET in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells, one is that there is one or more unidentified zinc transporters in the ER. It is also possible that zinc in the ER is supplied from Golgi apparatus through the retrieval pathway. If this is the case, zinc in the ER of ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells may originate from that transported into Golgi by ZnT6. Affinity of GPI-PET to zinc may be very high and therefore GPI-PET can bind with zinc that may be present at a low concentration in ER in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells, resulting in the formation of active conformation of the enzyme.
Instead of TNAP, PLAP expressed in DT40 cells successfully acted as a reporter that conveyed zinc metabolism in the biosynthetic-secretory pathway. Surprisingly, like the active PLAP in the wild-type DT40, the inactive PLAP protein in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells was transported to cytoplasmic membrane where it was anchored via GPI. It is believed in eukaryotes that misfolded or incompletely assembled proteins are eventually degraded by the ER-associated degradation pathway after transported back into the cytoplasm, or lysosomal pathway (31). The inactive PLAP in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells may be misfolded or incompletely assembled through unsuccessful binding with zinc and would be a target of cellular degradation pathways. At the present time we have no evidence for misfolding of PLAP, but if we assume this is the case, the result that PLAP in ZnT5 Ϫ ZnT7 Ϫ/Ϫ cells is transported to the cytoplasmic membrane may indicate that ZnT5 and/or ZnT7 are responsible not only for activation of secretory zinc proteins but also for zinc transport to keep the degradation pathways fully active. Very recently, Ellis et al. (32) reported that Saccharomyces cerevisiae MSC2, which is highly homologous to the C-terminal portion of ZnT5 and ZnT7, is one route of zinc entry into ER in yeast. The msc2 mutants are defective in ER-associated degradation pathway under zinc-limited conditions. Redundant expression of zinc transporters has made it difficult to verify the precise role of individual transporters. Be-cause DT40 has high homologous recombination activity and expresses most of the ZnT members, this cell line is expected to be useful for elucidating the cellular functions of zinc transporters, including the molecular mechanism of zinc transport and interplay of transporters.