Functional characterization of a novel mammalian zinc transporter, ZnT6.

We describe ZnT6, a new member of the CDF (cation diffusion facilitator) family of heavy metal transporters. The human ZNT6 gene was mapped at 2p21-22, while the mouse Znt6 was localized to chromosome 17. Overexpression of ZnT6 in both wild-type yeast and mutants that are deficient in cytoplasmic zinc causes growth inhibition, but this inhibition is abolished in mutant cells with high cytoplasmic zinc. ZnT6 may function in transporting the cytoplasmic zinc into the Golgi apparatus as well as the vesicular compartment, as evidenced by its overlapping intracellular localization with TGN38 and transferrin receptor in the normal rat kidney cells. We also demonstrate that the intracellular distributions of ZnT6 as well as ZnT4 are regulated by zinc in the normal rat kidney cells. The results from this report, combined with those from other studies, suggest that the intracellular zinc homeostasis is mediated by many ZnT proteins, which act in tissue-, cell-, and organelle-specific manners.

Zinc is an essential trace element required for the structural stability of a variety of proteins involved in transcription and protein trafficking as well as for the catalytic activity of metalloenzymes such as pancreatic carboxypeptidases, alkaline phosphatase, various dehydrogenases, and superoxide dismutase (1). In mammals, zinc is absorbed through the brush border of small intestinal mucosa from diet and transported through blood to the tissues and cells where zinc is needed (2).
Four ZnT proteins, ZnT1-ZnT4, have been identified. In general, the ZnT proteins are predicted to have similar protein structures with features including six transmembrane domains and a histidine-rich cytoplasmic loop between transmembrane domain IV and V. Znt1 is expressed in many tissues, and the protein resides on the plasma membrane. Overexpression of ZnT1 in zinc-sensitive cells confers zinc resistance probably by direct export of zinc out cells (3). ZnT2 and ZnT3 are more similar to each other than they are to ZnT1. They are located on the vesicular membranes and involved in transporting zinc from the cytoplasm into vesicles. The expression of Znt2 has been demonstrated in tissues including small intestine, kidney, seminal vesicles, testis, and placenta. However, the expression of Znt3 was detected only in brain and testis (4,5). ZnT4 is more related to ZnT2 and ZnT3 than ZnT1, and it appears to be expressed ubiquitously with abundant expression in brain, mammary gland, and small intestine. ZnT4 confers zinc resistance to ⌬zrc1, a yeast strain that has a defect in a vacuolar zinc transporter (ZRC1) (zinc resistance conferring), suggesting that it is involved in sequestration of cytoplasmic zinc into vacuoles when expressed in yeast (6). In addition, ZnT4 is deficient in the lethal milk mouse mutant in which pups of any genotype suckled on homozygous lethal milk mothers die of zinc deficiency before weaning (12,13). The zinc level in the milk of homozygous lethal milk animals is about 50% that of normal animals (14,15) demonstrating that ZnT4 plays a crucial role in depositing the cytoplasmic zinc into the secretory vesicles in the lactating mammary glands.
Yeast mutants of zinc metabolism have provided useful tools to analyze the function of mammalian zinc transporters. Recent studies have been shown that at least five zinc transporters of Saccharomyces cerevisiae are involved in yeast zinc homeostasis. The ZRT1 and ZRT2 genes encode the high and low affinity zinc importers, respectively (16,17). The ZRC1 and ZRT3 genes encode the zinc transporters responsible for zinc movement from the cytoplasm to vacuoles and from vacuoles to the cytoplasm, respectively (18,19). And the MSC2 (meiotic sister-chromatid recombination) gene encodes a zinc transporter that may act in transporting zinc out of the endoplasmic reticulum/nucleus (20).
Here we describe the identification and characterization of a sixth member of the ZnT family, which we designate as ZnT6. We provide evidence that ZnT6 functions as a zinc transporter responsible for relocating the cytoplasmic zinc into the trans Golgi network (TGN) as well as the vesicular compartment. We also show that the intracellular localizations of ZnT6 and ZnT4 in the normal rat kidney (NRK) cells are regulated by zinc.
Yeast Strains and Culture Conditions-The yeast null-mutant of ⌬zrt1 was a gift from Dr. David Eide at University of Missouri-Columbia, Columbia, MO. The yeast deletion strains ⌬zrt3 and ⌬msc2 were purchased from ResGen (Invitrogen). ⌬zrc1 was kindly provided by D. Conklin (Cold Spring Harbor Laboratories). Wild-type yeast BY4741 (MATa his3D1 leu2D0 met15D0 ura3D0) was purchased from ResGen. Yeast strains were grown in synthetic defined medium supplemented with auxotrophic requirements and either 2% glucose or 2% galactose and 1% raffinose.
Plasmid Construction-The entire open reading frame (ORF) sequence of mouse Znt6 was PCR-amplified using primers with a HindIII site incorporated immediately before the methionine codon and an XbaI site immediately after the stop codon. The mouse EST clone AA386648 was used as the template for the PCR amplification. The plasmid pY-ZnT6 was generated by digestion and insertion of the amplified sequence into the HindIII and XbaI sites of a yeast expression vector, pYES2 (Invitrogen). The expression plasmid, pcDNA5/ZnT6-Myc, which was used to transfect CHO cells, was made by several cloning steps. First, the Znt6 ORF sequence lacking a stop codon was generated by PCR using the primers with a HindIII site incorporated immediately before the methionine codon and a SacII site incorporated into the stop codon. The resulting PCR fragments were digested with respective restriction enzymes designed in the primer pair and inserted into the HindIII and SacII sites of pEGFP-N1 (CLONTECH) to create pEGFP/ ZnT6. Second, the HindIII and SmalI fragment containing the Znt6 ORF was excised from pEGFP/ZnT6 and cloned into the HindIII and EcoRV sites of pcDNA3.1/myc-His vector (Invitrogen). Finally, the Hin-dIII and AgeI (blunted with Klenow fragment) fragment in which the Myc epitope was fused in frame to the C-terminal of the Znt6 ORF was isolated from the resulting plasmid, pcDNA3.1/ZnT6-Myc, and cloned into pcDNA5/FRT (Invitrogen). The final construct, pcDNA5/ZnT6-Myc, was sequenced and transfected into CHO cells alone with pOG44 (Flp recombinase) using LipofectAMINE Plus reagent (Invitrogen). The stable clones were selected by culturing the transfected cells in the medium containing 100 g/ml of hygromycin B (Invitrogen). A negative control CHO cell line was also obtained by transfecting the vector, pcDNA5/FRT, into CHO cells.
Northern Blot Analysis-Northern blot with mRNAs derived from different tissues of C57BL/6J was prepared as described in Huang and Gitschier (6). The blot was probed with a 32 P-labeled full-length ORF DNA fragment of Znt6 in ExpressHyb hybridization solution (CLON-TECH) and washed according to the manufacturer's instructions.
Antibodies-Rabbit anti-ZnT6 and anti-ZnT4 polyclonal antibodies were raised against synthetic peptides (amino acids 446 -460 of ZnT6 and amino acids 93-110 of ZnT4) and affinity-purified (Pierce). The epitopes used in generation of rabbit anti-ZnT6 and anti-ZnT4 polyclonal antibodies were found to be unique to the respective proteins in a BLAST search of the SWISSPROT data base. A monoclonal anti-Myc epitope antibody was purchased from Stressgen. A monoclonal anti-TGN38 antibody was purchased from Transduction Laboratories. A peroxidase-conjugated goat anti-rabbit antibody was purchased from Pierce. An Alexa 488-conjugated goat anti-rabbit and an Alexa 594conjugated anti-mouse antibody were purchased from Molecular Probes.
Western Blot Analysis-The brain, lung, liver, kidney, heart, and small intestine tissues were isolated from a C57BL/6J mouse and homogenized in the lysis buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Triton X-100, 10% SDS, and 0.5% deoxycholate. One mini proteinase inhibitor tablet (Roche Molecular Biochemicals) was added into 10 ml of the lysis buffer just before use. The homogenized tissues were then heated at 100°C for 5 min and centrifuged at 4°C for 10 min. The supernatants were collected and quantitated using Bio-Rad Protein Assay. Proteins of 100 g of the brain, lung, liver, kidney, heart, and small intestine extracts were separated on a 4 -20% Tris-glycine Ready Gel (Bio-Rad) and transferred to a nitrocellulose membrane (Bio-Rad). ZnT6 was detected as described previously (21) using an affinity-purified anti-ZnT6 antibody (1:250 dilution) followed by a peroxidase-con-jugated secondary antibody (1:2500). ZnT6 was visualized by using ECL kit (Amersham Biosciences). The cell lysate from the vector-and pcDNA5/ZnT6-Myc stably transfected CHO cells were prepared as follows. The CHO cell lines were cultured in the 150-mm Petri dishes for 48 h, washed with 1ϫ PBS twice and collected in 1ϫ PBS. Cells were pelleted, resuspended, and swelled in a lysis buffer (50 mM Tri-HCl, pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, and 1% deoxycholate) for 5 min. 70 l of 100 mM phenylmethylsulfonyl fluoride (ICN Biomedicals Inc.), 2 l of 2 mg/ml pepstatin A (Roche Molecular Biochemicals), and one mini-proteinase inhibitor tablet (Roche Molecular Biochemicals) were added into 7 ml of the lysis buffer just before use. Cells were then homogenized in a Dounce homogenizer with a B-pestle, and the cell lysate was incubated on ice for 15 min. The cell lysate was then put through a syringe with a 21-guage needle attached for four times. The SDS-PAGE loading buffer was added to the cell lysate, and the mixture was heated at 100°C for 10 min. The DNA and insoluble proteins were pelleted by centrifugation at 25,000 ϫ g at 4°C for 20 min in a microcentrifuge. The supernatant was collected, and the protein concentration was determined using Bio-Rad Protein Assay. 50 g of proteins were separated on a 4 -20% Tris-glycine Ready Gel (Bio-Rad) and transferred to a nitrocellulose membrane (Bio-Rad). ZnT6-Myc fusion protein was detected using an anti-Myc antibody (2 mg/ml, Stressgen) followed by a peroxidase-conjugated secondary antibody (1:2000). ZnT6-Myc fusion protein was visualized by using ECL kit (Amersham Biosciences).
Immunofluorescence Microscopy-Immunofluorescence analysis was performed essentially as described (21). NRK cells were cultured in slide chambers for 24 or 48 h, fixed with 4% paraformaldehyde, and permeabilized with 0.4% saponin (Sigma). Where indicated, the cells were treated with ZnSO 4 for 2 h prior to the fixation. The cells were subsequently stained with an affinity-purified anti-ZnT6 (1:20 dilution) or anti-ZnT4 (1:25 dilution) antibody followed by an Alexa 488-conjugated goat anti-rabbit antibody (1:250 dilution). In the brefeldin A treatment and colocalization studies, the NRK cells were costained with an anti-ZnT6 and an anti-TGN38 (1:200 dilution) antibodies followed by an Alexa 488-conjugated goat anti-rabbit antibody and an Alexa 594conjugated goat anti-mouse antibody (1:250 dilution), respectively. Photomicrographs were obtained by Nikon Eclipse 800 microscope with a digital camera.

RESULTS
Identification of Znt6 -Search of the EST data bases with the amino acid sequence of mouse ZnT4 uncovered several mouse and human EST clones predicted to encode a protein similar to the members of the ZnT family. The mouse EST clone AA386648 was purchased from ResGen (Invitrogen) and fully sequenced. This clone contains a single open reading frame encoding a 460-amino acid protein with a calculated molecular mass of 51 kDa. Based on the predicted amino acid sequence similarity to the members of the ZnT family (32%, 33%, and 33% amino acid identities to rat ZnT2, mouse ZnT3, and mouse ZnT4, respectively), we designated this novel gene as Znt6 (Slc30a6) and the gene product as ZnT6. ZnT6 shares many features described previously for the ZnT family, including six predicted transmembrane domains, cytoplasmic N-and C-terminal domains as well as a loop region between transmembrane domain IV and V (Fig. 1). However, in contrast to other ZnT proteins, the ZnT6 loop lacks multiple histidine residues, the potential ligands for zinc, while retaining serine residues, a finding reminiscent of CZCD, a prokaryotic transporter of zinc, cadmium, and cobalt (22) (Fig. 1). Another distinguishing feature of ZnT6 is its longer C terminus.
The human ZNT6 gene encodes the same size protein (460 amino acids) as that of mice and has 92% amino acid identity to it. Data base analysis revealed that two BAC clones (accession nos. AL121653 and AL121658) from human chromosome 2p21-22 contain the ZNT6 gene. The ZNT6 gene has at least 14 exons, and the genomic sequence spans about 56 kb. The locus of the mouse ZnT6 sequence is on chromosome 17, a region with homology by synteny (The Mouse Genome Sequencing Consortium). No human or mouse disease of suspected zinc metabolism maps to these loci.
By Northern blot analysis, two ZnT6 transcripts of 2.4-and 1.7-kb are revealed in mouse tissues, with relatively abundant expression in liver and brain (Fig. 2a). The expression of the 1.7-kb transcript is higher than that of the 2.4-kb in all tissues examined. The two sizes are reflected in two groups of EST clones, derived from mammary gland, skin, and embryo and differing only in their 3Ј-untranslated region. Comparison to the genomic sequence showed that two different polyadenylation signals are utilized. The Western blot analysis confirmed that a single ZnT6 protein of the predicted size (51 kDa) is produced in mouse brain and a much fainter signal is detected in small intestine and kidney (Fig. 2b), which are consistent with the levels of transcription observed on the Northern blot (Fig. 2a). A single protein band of slightly higher molecular mass (ϳ55 kDa) was detected in lung, indicating that ZnT6 may be posttranslationally modified in lung tissue (Fig. 2b). No ZnT6 protein band was detected in liver and heart on the Western blot although high level of the ZnT6 mRNA was found in liver on the Northern blot (Fig. 2). The discrepancy between the levels of ZnT6 mRNA and protein expression in lung (less mRNA detected but more protein expressed) and liver (abundant mRNA detected but no protein detected) suggests that a posttranscriptional mechanism may play a role in tissue-specific expression of the ZnT6 protein.
The specificity of the affinity-purified polyclonal anti-ZnT6 antibody was assessed by Western blots of the protein extract from a mouse brain probed with a preimmune serum, and the total protein extract from the CHO cells that were stably transfected with a plasmid expressing ZnT6-Myc fusion protein probed with a monoclonal anti-Myc antibody (Stressgen). No protein of the predicted size (ϳ51 kDa) was detected in brain when the preimmune serum was used (Fig. 2b). A protein band of the predicted size (ϳ55 kDa) was detected in the CHO cells expressing ZnT6-Myc fusion protein, whereas no protein was detected in the CHO cells transfected with vector alone (Fig.  2b). Taking together, the results indicate that the affinitypurified anti-ZnT6 antibody specifically reacts with the endogenous ZnT6 protein in the mouse tissues.
ZnT6 Acts as a Functional Zinc Transporter-We asked whether ZnT6 has the ability to transport zinc indirectly by looking at the effect of ZnT6 expression in a series of yeast mutants that are defective in zinc metabolism. First, we tested whether ZnT6 can relieve the sensitivity to zinc found in the yeast mutant ⌬zrc1, in which zinc accumulates in the cytoplasm due to a defect in a vacuolar zinc transporter ZRC1 (Fig.  1). Whereas ZnT4 was previously shown to rescue the growth of the yeast mutant in high zinc medium (6), ZnT6 failed to complement the ⌬zrc1 defect under similar conditions (data not shown). This finding suggests that ZnT6 may not be involved in the transport of the cytoplasmic zinc into the vacuoles, which largely reduces the cytoplasmic zinc concentration in cells.
We then asked whether expression of ZnT6 might exacerbate the cytoplasmic zinc deficiency seen in three yeast mutants: ⌬zrt1, which is defective in the yeast high affinity zinc uptake protein, and ⌬zrt3 and ⌬mcs2, which have defects in relocation of zinc from the vacuoles and endoplasmic reticulum/nucleus to the cytoplasm, respectively (16,19,20). In each mutant, overexpression of ZnT6 resulted in poor to no growth in the inducible media, whereas the mutants transformed with vector alone grew well (Fig. 3a). Furthermore, the inhibitory effect of ZnT6 on yeast growth was observed in the wild-type yeast strain (BY 4741) (Fig. 3b). These findings are consistent with the hypothesis that ZnT6 transports zinc out of the cytoplasm into either an intracellular pool or out of the cell, leading to a severe diminution in the cytoplasmic zinc and concomitant growth retardation. Indeed, under inducible media, ZnT6 did not inhibit the growth of ⌬zrc1 mutant cells (Fig. 3a).
To further test this hypothesis, we asked whether addition of zinc might counter the inhibitory effect of ZnT6 on cell growth. Looking at both wild-type and ⌬zrt1 strains, growth inhibition can be partially alleviated by adding 1.0 mM ZnCl 2 in the inducible media ( Fig. 3b and data not shown). However, higher zinc concentrations were no more effective (data not shown). The effect of zinc was specific: addition of iron (0.35 mM FeSO 4 ), copper (0.1 mM CuSO 4 ), cobalt (0.5 mM CoSO 4 ), or manganese (1.0 mM MnCl 2 ) did not help to restore growth in either wildtype or ⌬zrt1 strains expressing ZnT6 (data not shown).
Immunofluorescence Analysis of ZnT6 in NRK Cells-As each of the previously described ZnT proteins appears to serve a specific cellular zinc transport need, we asked by immunofluorescence analysis whether ZnT6 lies in a particular subcel-lular compartment in mammalian cells. In the cultured rat NRK cells, ZnT6 displayed a perinuclear location as well as a punctate/tubular pattern throughout the cytoplasm of the cells and underlying the plasma membrane. Localization of ZnT6 was compared with that of TGN38, a known TGN protein and to the TfR, a plasma membrane protein that recycles to the TGN shortly after internalization via recycling endosomes. As shown in Fig. 4, A, D, and G, the perinuclear staining of ZnT6 is completely coincident with that of TGN38, strongly suggesting that ZnT6 is associated with TGN. The tubular/punctuate staining of ZnT6 partially overlaps with that of transferrin receptor, whereas the perinuclear staining of both proteins completely overlaps, indicating that ZnT6 may reside in the vesicles in addition to the recycling endosomes in the NRK cells (data not shown). Fig. 4, B-I confirms that ZnT6 localizes to the TGN, as an addition of a fungal macrocyclic lactone, brefeldin A (BFA), known to disrupt the Golgi compartment, also disrupts ZnT6 with the same time course as that of TGN38. After 30 min of BFA treatment, the fluorescence perinuclear staining of ZnT6 completely disappears and the signal has diffused into the cytoplasm (Fig. 4B). Meanwhile, BFA treatment causes the fluorescence signal of TGN38 to diffuse into the cytoplasm and form a strong juxtanuclear spot (Fig. 4E). Incubation of BFAtreated cells in BFA-free media for 60 min resulted in nearly complete restoration of the normal localization of both ZnT6 and TGN38 (Fig 4, C, F, and I). These BFA-induced changes and the recovery of the changes by the fresh medium are FIG. 2. Expression of mouse Znt6. a, Northern blot assay. mRNA was purified from tissues derived from C57BL/6J strain. 2 g of poly(A) ϩ RNA was loaded in each lane. The Northern blot was probed with the ORF region of the mouse Znt6 cDNA. The standard RNA molecular mass marker is indicated on the right side of the blot. b, Western blot assays. Either 100 g of protein extracts isolated from different mouse tissues including brain, kidney, lung, heart, liver, and small intestine or 50 g of total proteins isolated from pcDNA5/FRT-or pcDNA5/ZnT6-Myctransfected CHO cells were separated on a 4 -20% Tris-glycine Ready Gel (Bio-Rad). The native ZnT6 and ZnT6-Myc fusion proteins were detected by a polyclonal anti-ZnT6 antibody and a monoclonal anti-Myc antibody, respectively. The specific proteins were visualized by a peroxidaseconjugated anti-rabbit or anti-mouse antibodies followed by application of an ECL kit (Amersham Biosciences). The pre-immune serum was used as a negative control for the anti-ZnT6 antibody. The total protein extract from the vector-transfected CHO cells was used as a negative control for the anti-Myc antibody. Protein marker sizes are shown as indicated.
characteristic features for the proteins that associate with Golgi compartment.
The Intracellular Location of ZnT6 Is Regulated by Zinc-We next asked whether the expression or distribution of ZnT6 was influenced by zinc. No changes in mRNA or protein abundance were elicited in the NRK cells cultured in the medium containing 200 M ZnSO 4 for 24 h (data not shown). However, by immunofluorescence analysis, we detected zinc-induced trafficking of ZnT6 at 2 and 24 h after ZnSO 4 treatment in the NRK cells. The panels on the left (A-D) in Fig. 5 show the staining for ZnT6 in the NRK cells exposed to various ZnSO 4 concentrations. After 2 h, the perinuclear capping started to diffuse into the cytoplasm and the punctate/tubular signals were enhanced and distributed toward the periphery of the cells with the addition of an increased amount of zinc (Fig. 5, B-D). We were unable to detect a clear delineation of the plasma membrane staining with the increased zinc concentration in the media.
To investigate whether this zinc-induced protein trafficking is a common phenomenon for the ZnT proteins, we studied the localization of ZnT4 under the same conditions. ZnT4 has been shown to reside in the TGN as well as in the endosomal compartment (Fig. 5E) (10). 2 As shown in Fig. 5, although ZnT4 does redistribute from its perinuclear location to the cytoplasm, it is less sensitive to zinc than ZnT6, with dramatic effects seen only at 200 M ZnSO 4 . Our data demonstrate that although the punctate staining patterns differ between ZnT6 and ZnT4, indicating they may reside in two different vesicular compartments, both proteins are localized in the TGN (Fig. 5, A and E and data not shown), and their intracellular localizations are regulated by extracellular zinc concentration.

DISCUSSION
In this paper, we have functionally characterized the sixth member of the ZnT family. We conclude that ZnT6 is a zinc transporter that translocates the cytoplasmic zinc into the TGN as well as the vesicular compartments based on the evidences from our studies. First, ZnT6 is closely related to several previously characterized ZnT members from mammals, yeast, and bacteria (Fig. 1). Second, ZnT6 was able to suppress the FIG. 3. Effect of ZnT6 and extracellular zinc on the growth of yeast. a, effect of ZnT6 on the growth of yeast mutants of zinc metabolism. ⌬zrt1, ⌬zrt3, ⌬mcs2, and ⌬zrc1 yeast strains were transformed with either the vector, pYES, or the ZnT6-containing plasmid, pY-ZnT6. Yeast cells were grown on the synthetic defined-ura medium plates either with 2% glucose or with 2% galactose and 1% raffinose at 30°C for 3 days. b, effect of extracellular zinc on the growth of wild-type yeast expressing ZnT6. Wild-type yeast cells (BY4741) were transformed with either pYES or pY-ZnT6 plasmid. Yeast cells were grown on the inducible synthetic defined-ura medium plates (2% galactose and 1% raffinose) supplemented with either 0 or 1.0 mM ZnCl 2 at 30°C for 3 days. growth of several yeast mutants that are defective in accumulation of the cytoplasmic zinc. Third, the higher cytoplasmic zinc content was able to alleviate ZnT6-transformed ⌬zrc1 from growth inhibition by ZnT6, and the extracellular zinc was able to partially rescue the growth inhibition by ZnT6 in wild-type, ⌬zrt1, and ⌬zrt1/⌬zrt2 yeast strains (Fig. 3. and data not shown). Last, ZnT6 is localized predominantly in the TGN compartment in NRK cells in the normal culture condition. However, its distribution from the TGN to the vesicular compartment in NRK cells is regulated by extracellular zinc ions in the culture medium.
The zinc-induced redistribution of ZnT proteins is reminiscent of the findings with a copper export protein described for the Menkes disease protein (MNK), whose relocation from the TGN to the plasma membrane is induced by elevated copper, which in turn exports excess intracellular copper (23). The zinc-induced protein trafficking of ZnT4 is consistent with a functional role of ZnT4 in the lactating mammary gland as zinc is deficient in the milk of homozygous lethal milk mothers. ZnT4 may transport zinc into the TGN vesicles, and zinc may leave the cells through the secretory pathway. We speculate that ZnT6 functions in two major aspects of intracellular zinc homeostasis. First, ZnT6 may be involved in transporting the cytoplasmic zinc into the TGN vesicles, where it may be incorporated into zinc-requiring proteins. In this situation, the function of ZnT6 overlaps with that of ZnT4. The partial overlap of cellular localization of ZnT6 and ZnT4 may explain the lack of phenotypes in the other organs in the lethal milk mutant. Second, ZnT6 may play a role in zinc export based on its intracellular distribution, its zinc-induced intracellular redistribution, and its ability to diminish the cytoplasmic zinc when overexpressed in yeast cells. ZnT6 may cycle between the TGN and the plasma membrane both in basal and elevated cytoplasmic zinc concentration. Perhaps we did not detect a clear delineation of the plasma membrane staining because ZnT6 may be constantly recycled. Alternatively, ZnT6 may transport zinc into the TGN and the cytoplasmic vesicles, which could leave the cell through the secretory pathway.
A key aspect of the function of ZnT6 in regulating the cytoplasmic zinc levels in cells may be its zinc-induced relocalization from the TGN to the cytoplasmic vesicles and then removal of zinc from cells either by the export function of ZnT6 on plasma membrane or by the secretory pathway because the expression of ZnT6 is not regulated by zinc at the transcriptional level or at the translational level in vitro. The molecular basis for the internalization of ZnT6 from the plasma membrane might be the di-leucine motifs at amino acid positions 363 and 364 as well as 401 and 402 of ZnT6 (23,24). The C-terminal di-leucine motif has been demonstrated to mediate copper-induced trafficking of the MNK from the TGN to the plasma membrane, followed by return to the TGN via clathrincoated vesicles. Further mutagenesis studies are required to demonstrate the importance of these motifs in the intracellular distribution and zinc-induced redistribution of ZnT6 (23,25).
The multiple histidine residues in the loop region between transmembrane IV and V in the ZnT proteins have been hypothesized to act for a zinc binding domain. Our preliminary data indicate that the multiple serine residues in the loop region of ZnT6, which lacks multiple histidine residues in that region, are critical for the function of ZnT6 in vivo. The serine residues may coordinate the zinc binding with the histidine residue at the C-terminal end of the ZnT proteins (His-300 in ZnT6) to facilitate zinc cross-membrane trafficking. Further studies are under way to elucidate the mechanism of how zinc transporters relay zinc across membranes and where the energy comes from.
The presence of translationally inactive ZnT6 mRNA in liver is reminiscent of the expression pattern of ZnT3 (5). ZnT3 mRNA is abundantly expressed in adult testis of mice, as detected by Northern blot analysis and reverse transcription-PCR, while the ZnT3 protein is undetectable by Western blot analysis. It has been shown that the ZnT3 transcripts in testis are predominantly associated with monoribosomes, suggesting that this mRNA may not be translated efficiently (5). A modified ZnT6 protein with slightly higher molecular mass was observed in lung. In addition, the amount of the ZnT6 protein detected in lung is substantial as compared with its low level of transcripts. Many potential protein kinase C and CK2 phosphorylation and N-glycosylation sites are predicted in the ZnT6 protein using the ProfileScan Server (hits.isb-sib.ch). Phosphorylation and/or glycosylation of ZnT6 may increase the stability of the ZnT6 protein in lung. The functional difference between the native ZnT6 and modified ZnT6 proteins in intracellular zinc transportation remains unknown.
It has become clear from our studies and others that zinc is transported into and out of a variety of intracellular compart- FIG. 5. Effects of zinc on the intracellular localizations of ZnT6 and ZnT4. The NRK cells were grown for 48 h in the DMEM media containing 10% FBS. Cells were then treated with ZnSO 4 for the indicated concentration in the serum-free DMEM media for 2 h. The endogenous ZnT6 and ZnT4 proteins were detected by the affinity-purified antibodies against the C terminus of ZnT6 and the N terminus of ZnT4, respectively, followed by an Alexa 488-conjugated anti-rabbit antibody. Normal predominant perinuclear staining of ZnT6 (A) and ZnT4 (E) was observed in the cells without ZnSO 4 treatment. The extensive punctate/tubular staining of ZnT6 in the cytoplasm was observed in the cells treated with 30 (B), 100 (C), or 200 M (D) ZnSO 4 . Similar to the zinc effects on ZnT6 localization, zinc causes normal perinuclear staining of ZnT4 to diffuse into the cytoplasm but at higher ZnSO 4 concentration (F-H). ments, and there appear to be dedicated zinc transporters for each event. The findings of the present study extend our understanding of what zinc transporters do in cells, how the expression of zinc transporters regulate, and how the transporters get to their various compartments.