Slc39a14 Gene Encodes ZIP14, A Metal/Bicarbonate Symporter: Similarities to the ZIP8 Transporter

Abstract

The mouse and human genomes contain 14 highly conserved SLC39 genes. Viewed from an evolutionary perspective, SLC39A14 and SLC39A8 are the most closely related, each having three noncoding exons 1. However, SLC39A14 has two exons 4, giving rise to Zrt- and Irt-related protein (ZIP)ZIP14A and ZIP14B alternatively spliced products. C57BL/6J mouse ZIP14A expression is highest in liver, duodenum, kidney, and testis; ZIP14B expression is highest in liver, duodenum, brain, and testis; and ZIP8 is highest in lung, testis, and kidney. We studied ZIP14 stably retroviral-infected mouse fetal fibroblast cultures and transiently transfected Madin-Darby canine kidney (MDCK) polarized epithelial cells. Our findings include: 1) ZIP14-mediated cadmium uptake is proportional to cell toxicity, but manganese is not; 2) ZIP14B has a higher affinity than ZIP14A toward Cd²⁺ (K_m = 0.14 versus 1.1 μM) and Mn²⁺ uptake (K_m = 4.4 versus 18.2 μM); 3) ZIP14A- and ZIP14B-mediated Cd²⁺ uptake is most inhibited by Zn²⁺, and next by Mn²⁺ and Cu²⁺; 4) like ZIP8, ZIP14A- and ZIP14B-mediated Cd²⁺ uptake is dependent on extracellular ${\text{HCO}}_3^{-}$; 5) like ZIP8, ZIP14 transporters are localized on the apical surface of MDCK-ZIP cells; and 6) like ZIP8, ZIP14 proteins are glycosylated. Tissues such as intestine and liver, located between the environment and the animal, show high levels of ZIP14; given the high affinity for ZIP14, Cd²⁺ is likely to act as a rogue hitchhiker—displacing Zn²⁺ or Mn²⁺ and entering the body to cause unwanted cell damage and disease.

Cadmium is a toxic nonessential divalent cation classified by International Agency for Research on Cancer (Lyon, France) as a “Category I” human lung carcinogen. Acute doses lead to immediate damage in the central nervous system, lung, bone, gastrointestinal tract, liver, placenta, developing embryo, ovary, and testis (Waisberg et al., 2003; Zalups and Ahmad, 2003). Long-term exposure to low Cd²⁺ doses primarily causes renal proximal tubular metabolic acidosis and osteomalacia (renal Fanconi syndrome); Cd²⁺ is eliminated very slowly from the body and thus accumulates (predominantly in kidney, less in liver) as a total body burden with age.

Sources of Cd²⁺ include cigarette smoke, contaminated soil (around metal-smelting operations), and polluted foods such as shellfish. Curiously, tobacco plants take up and concentrate large amounts of Cd²⁺. Dump sites for toxic waste contain Ni²⁺-Cd²⁺ batteries; these metals often leach into the groundwater, ultimately entering the human food chain. Cd²⁺ is ranked seventh among the “Top 20 Hazardous Substances Priority List” by the Agency for Toxic Substances and Disease Registry and the U.S. Environmental Protection Agency (Fay and Mumtaz, 1996). People at highest risk for Cd²⁺-induced lung cancer and chronic nephropathy include cigarette smokers, women having low body-iron stores, persons on a habitual diet rich in high-fiber foods or contaminated shellfish, and malnourished populations (Järup et al., 1998; Järup, 2002, 2003; Waalkes, 2003).

For the past 80 years, Cd²⁺ uptake into mammalian cells has been presumed to take place before Cd²⁺-mediated disease can occur. One possibility for Cd²⁺ influx into mammalian cell cultures is by way of Ca²⁺ channels (Shibuya and Douglas, 1992; Hinkle and Osborne, 1994; Olivi and Bressler, 2000; Bergeron and Jumarie, 2006). SLC11A2 has a preference for Fe²⁺ but also transports Pb²⁺ and Cd²⁺ (Bressler et al., 2004); for this transporter, Cd²⁺ influx is proton-dependent (Bressler et al., 2004). Using SLC11A2 knockdown studies in human intestinal Caco-2 cells (Bannon et al., 2003), proton-dependent Cd²⁺ transport has been demonstrated. Other studies have suggested that SLC11A2 participates in Cd²⁺ transport in gastrointestinal enterocytes (Elisma and Jumarie, 2001; Tallkvist et al., 2001; Park et al., 2002) and renal distal tubular cells (Olivi et al., 2001; Park et al., 2002). Consistent with these studies, Cd²⁺ transport in Xenopus laevis oocytes expressing human SLC11A2 shows Michaelis-Menten kinetics with a K_m of 1.04 ± 0.13 μM (Okubo et al., 2003).

Still, no definitive genotype-phenotype association (i.e., a particular gene attributed to Cd²⁺-induced disease) was reported until 2005. Recent studies have shown such a relationship: allelic differences in the mouse Slc39a8 gene are responsible for striking variability in risk of Cd²⁺-induced testicular necrosis among different inbred mouse strains (Dalton et al., 2005), as well as Cd²⁺-induced acute renal failure (Wang et al., 2007). Slc39a8 encodes the ZIP8 transporter, which undoubtedly also transports an essential divalent cation. Although Mn²⁺ was shown to be the best inhibitor of ZIP8-mediated Cd²⁺ uptake and has a very low K_m value for ZIP8-mediated uptake (2.2 μM) (He et al., 2006), we believe that zinc (Zn²⁺) cannot be ruled out as a substrate for ZIP8. Indeed, in studies with X. laevis oocytes (Liu et al., 2008), K_m values for Cd²⁺ and Zn²⁺ were recently shown to be ∼0.48 and ∼0.26 μM, respectively. Cd²⁺ probably displaces Mn²⁺ or Zn²⁺ and operates as a rogue hitchhiker, entering cells via ZIP8 and subsequently contributing to cell death, cancer, and other diseases (He et al., 2006; Wang et al., 2007).

The ZIP8 transporter is a metal/bicarbonate symporter (He et al., 2006). ZIP8 is localized on apical surfaces (He et al., 2006) of several cell types: between the blood and vascular endothelial cells of the testis (Dalton et al., 2005; Wang et al., 2007), and between the glomerular filtrate and proximal tubular epithelial cells of the kidney (Wang et al., 2007). ZIP8 therefore seems to act as a “gatekeeper” for maintaining intracellular Zn²⁺ and perhaps Mn²⁺ homeostasis. On the other hand, however, ZIP8 can serve as a means for bringing unwanted environmental Cd²⁺ into the organism. In terms of evolution, the closest neighbor to ZIP8 is ZIP14, encoded by the SLC39A14 gene. The present study provides the complete characterization of this gene and its ZIP14 mRNAs and proteins and for the first time demonstrates the significance of ZIP14 in Cd²⁺-mediated cell damage.

Materials and Methods

Bioinformatics Tools

Protein sequences and their respective ZIP domains were obtained from the NCBI website (http://www.ncbi.nlm.nih.gov). ClustalW (http://www.ebi.ac.uk/clustalw/index.html) was used to construct a multiple alignment and phylogenetic tree of human and mouse ZIP proteins. Only the ZIP domain sequences of each protein were “joined” to each other and used to construct this tree. To assess the topology and transmembrane (TM) domain sequences of ZIP proteins, each protein sequence was submitted to the online program MINNOU (membrane protein identification, without explicit use of hydropathy profiles and alignments) at http://minnou.cchmc.org/ to detect the putative TM regions (Cao et al., 2006). Only the TM domain sequences were used in generating an artificial sequence for each protein. Percentage identity was then assessed, based on only the pair-wise alignment of these artificial sequences.

Chemicals

All divalent cations, plus Fe³⁺, were purchased as chloride salts from Thermo Fisher Scientific (Waltham, MA). The remainder of the chemicals was purchased from Sigma (St. Louis, MO). ¹⁰⁹CdCl₂ [710 mCi/mg (1 mCi = 37 mBq) in 0.1 M HCl] was purchased from GE Healthcare (Chalfont St. Giles, Buckinghamshire, UK), ⁵⁴MnCl₂ [7734 mCi/mg in 0.5 M HCl] was from Perkin-Elmer Life and Analytical Sciences (Waltham, MA), and ⁶⁵ZnCl2 [140 mCi/mg in 0.1 M HCl] from the National Laboratory of Oak Ridge (Oak Ridge, TN). Uptake medium was a modified version of Hanks' balanced salt solution (HBSS), as detailed (He et al., 2006).

Cell Cultures

Mouse fetal fibroblast (MFF) cells (Dalton et al., 2005) or Madin-Darby canine kidney (MDCK) cells (American Type Culture Collection; Manassas, VA) were cultured at 37°C in 5% CO₂ in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA) plus 10% fetal bovine serum (Hyclone, Logan, UT). All details of the culture medium have been described previously (Dalton et al., 2005; He et al., 2006).

Quantitative Polymerase Chain Reaction Analysis

Total RNAwas isolated from six tissues of untreated C57BL/6Jmice. Total RNA (2.5 μg) was used as a template for reverse transcription in 20 μl and then primed with oligo(dT) using the SuperScript III first-strand kit (Invitrogen), according to the manufacturer's recommendations. The plasmids pBlueScript-ZIP14A-HA, pBlueScript-ZIP14B-HA, and pXFRM-ZIP8-HA were used to establish calibration curves, which were then used to quantify the ZIP14A, ZIP14B, and ZIP8 mRNA copy numbers, respectively, in each tissue. The absolute copy number of each plasmid can be calculated as the “copy number = mass (in grams) × (6.023 × 10²³) / molecular weight. To prepare the calibration curve, we first did a serial dilution (10⁻¹, 10⁻², 10⁻³, 10⁻⁴, 10⁻⁵, 10⁻⁶, 10⁻⁷, 10⁻⁸ μg/PCR reaction) for each plasmid; each point was analyzed in duplicate using the same PCR settings as we had used in analyzing the experimental samples. The average ΔC_t value at each point was then used to plot the linear calibration curves [log₁₀ (mass) versus ΔC_t]. Finally, the linear regression equation and correlation value r (to reflect goodness of fit) were determined. Each ΔC_t value from an experimental sample could be converted to the absolute copy number of transcripts, based on this equation. PCR primers for mRNA were as follows: for Z14A: forward, 5′-TTCCTCAGTGTCTCACTGATTAA-3′; reverse, 5′-GGAAAGGGCGTTAGAGAGC-3′ (PCR product is 142 bp); for Z14B∷ forward, 5′-CATTGAAGTATGGGGGTACGGT-3′; reverse, 5′-ATGAAGTAGAGCAGGAGCCTCT-3′ (PCR product is 122 bp). PCR primers for the ZIP8 mRNA were: forward, 5′-GCAACAATTTTGCTCCCAAT-3′; reverse, Zip8-R 5′-TCCCTATGGAGATGTTTCTGTG-3′ (PCR product is 291 bp).

The range of calibration curves was linear from 10³ to 10⁷ μg of plasmid (r = 0.999 for ZIP14A, 0.998 for ZIPB and 0.995 for ZIP8). Quantitative reverse transcription-PCR reactions were performed in triplicate in a mixture containing 1× Power SYBR Green PCR Master Mix (Bio-Rad Laboratories, Hercules, CA) on a DNA engine2 real-time PCR system (MJ Research, Watertown, MA). Using 1 μl from step 1, the 20-μl reaction (0.125 μg of total RNA) was heated to 95°C for 10 min and immediately cycled 40 times through a denaturing step at 95°C for 15 s, an annealing at 58°C for 30 s, and an elongation step at 72°C for 45 s. Melting curve analyses were performed after PCR amplification to ensure that a single product with the expected melting curve characteristics was obtained—as preliminarily determined during primer tests.

Cloning of the ZIP14 cDNAs and Delivery into MFF or MDCK Tet-off Cells

Oligo-dT-primed reverse transcription was carried out with C57BL/6J mouse liver total RNA. Primers for amplification began at the ZIP14 cDNAs' start codon and ended at the stop codon; a consensus Kozak sequence at the start site was included for efficient expression. For ZIP14A and ZIP14B, restriction sites were added at the 5′ (BamHI) and 3′ (ApaI) ends of the coding sequence for cloning into the pRevTRE vector (Invitrogen). Restriction sites for ZIP8 and ZIP4, procedures for inserting a hemagglutinin (HA) tag in-frame at the C terminus and infection of MFF cell cultures with a retrovirus encoding the Tet-off receptor (Bergwitz et al., 2000) have been detailed previously (Dalton et al., 2005; He et al., 2006). These procedures resulted in stable rvLUC (as the ZIP-absent luciferase control), rvZIP14Aha, rvZIP14Bha, rvZIP8ha, and rvZIP4ha MFF cell cultures. Transiently transfected MDCK cells are denoted by “MDCK-” at the beginning.

Determination of Cd²⁺ and Mn²⁺ Uptake

The procedures for measuring radiolabeled Cd²⁺ and Mn²⁺ uptake have been detailed previously (Dalton et al., 2005; He et al., 2006).

Cell Survival after Cd²⁺ or Mn²⁺ Treatment

The methods for measuring cell viability, using the 3-(4,5-dimethlythiazol-2-yl)-2,5-diphenyl tetra[b]zolium bromide assay, have been described previously (Dalton et al., 2005; He et al., 2006).

Inhibition of Cd²⁺ Uptake by Other Metal Ions

The procedures for measuring competitive inhibition of Cd²⁺ uptake by divalent and trivalent cations have been detailed previously (Dalton et al., 2005; He et al., 2006).

K_m Determination

Cells grown in HBSS were treated with ¹⁰⁹CdCl₂, ⁵⁴MnCl₂ or ⁶⁵ZnCl₂ at six different concentrations, as detailed previously (He et al., 2006). The uptake data at the different concentrations were entered into the Enzyme Kinetics Module, an add-on software program of SigmaPlot (Systat Software, Inc., San Jose, CA).

Dependence on pH, Other Ions, Energy, and Temperature

The procedures for measuring these parameters have been detailed previously (He et al., 2006).

Dependence on ${\text{HCO}}_3^{-}$ for Cd²⁺ Uptake

The procedures for measuring ${\text{HCO}}_3^{-}$ dependence, including inhibition by 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS), a well known competitive inhibitor of ${\text{HCO}}_3^{-}$ transporters, have been described previously (He et al., 2006).

Z-Stack Confocal Microscopy for Distinguishing between Apical and Basolateral Location

MDCK cells were seeded onto cover slips in a 24-well plate. Next day, when the cells reached 90% confluence, the cells were transfected according to the manufacturer's protocol for Lipofectamine 2000 (Invitrogen) with the combination of the following plasmids: pRevTet-Off and pRevTRE-ZIP14Aha or pRevTet-Off and pRevTRE-ZIP14Bha. Some wells were transfected with the plasmid encoding green fluorescent protein-NBC1, known to localize to the basolateral surface. The rest of the protocol for carrying out Z-stack confocal microscopy in transiently transfected confluent monolayers of MDCK polarized epithelial cells has been detailed (Li et al., 2005; He et al., 2006). In general, 0.4- to 1.0-μm fixed interval cuts were carried out, and 20 to 30 images were generated as a gallery. The images (x-y projections) and corresponding z-lines (x-z or y-z projections) and tangential cuts were subsequently obtained.

Western Immunoblot Analysis of ZIP14 Protein Glycosylation

Peptide N-glycosidase F (PNGase F; New England Biolabs, Ipswich, MA) is an amidase that cleaves between the innermost N-acetylglucosamine and asparagine residues of high mannose, hybrid, and complex oligosaccharides from N-linked glycoproteins. The procedures for performing Western immunoblots, with and without PNGase F, to assess ZIP14 glycosylation, have been described previously (He et al., 2006).

Statistical Analysis

Statistical significance between groups was determined by analysis of variance between each group and/or Student's t test. All assays were performed in duplicate or triplicate and repeated at least twice. Statistical analyses were performed with the use of SAS statistical software (SAS Institute Inc., Cary, NC). The determinations of K_m and V_max values for ZIP-mediated metal uptake, metal competition studies, and EC₅₀ values (concentrations at which cell survival was 50% of maximum), were determined using Sigma Plot (SPSS Inc., Chicago, IL).

Results

Evolutionary Analysis and Bioinformatics

The solute-carrier (SLC) gene superfamily comprises 46 families and 360 putatively functional transporter genes in the human genome (Hediger et al., 2004). No homology is shared among the 46 families; however, at least 20–25% amino acid sequence identity (i.e., a specific “signature sequence”) is shared by members within any SLC family. Figure 1A shows the phylogenetic tree of the SLC39 family. Mouse and human have 14 SLC39 genes; each of the 14 genes in mouse has an ortholog in human that is highly conserved. It can be seen from the dendrogram that ZIP14 and ZIP8 are most closely related to one another and distantly related to all of the other 12 members of this family.

Fig. 1.

A, nearest-neighbor joining (NNJ)-generated phylogenetic tree of the 14 mouse and 14 human ZIP domains of the SLC39 gene-encoded proteins. These amino acid fragments were ligated, aligned, and then compared to determine percentage identity. In terms of evolution, the mouse Slc39a14 and Slc39a8 and human SLC39A14 and SLC39A8 members (large open arrows) are most closely related and diverged from one another (small arrow) sometime after the land animal-sea animal split ∼425 million years ago (see Discussion). B, structure of the mouse Slc39a14 genomic gene (introns not drawn to scale) illustrating the alternatively spliced exons 4, which encode the ZIP14A and ZIP14B proteins, respectively. Similar to Slc39a8 (Dalton et al., 2005), the Slc39a14 gene also has three alternatively spliced nontranslated exons 1. Of the nine exons (rectangles), the coding region is closed, and the 5′ and 3′ UTRs are open. The human SLC39A14 gene has the same highly conserved structure as the mouse gene, except that exons 1b and 1c have not been identified (see Table 1 and text).

Further analysis of the mouse Slc39a14 gene structure (Fig. 1B) revealed that this gene is highly similar to the mouse Slc39a8 in that they both have nine exons, as well as three noncoding exons 1 that denote alternative transcription start sites. The translation open reading frame begins in exon 2 and ends in exon 9. Between mouse and human, each internal exon has very similar, if not identical, lengths, and the exon-intron junctions are highly conserved (Table 1). Between mouse and human, SLC39A14 exons 3, 4, 6, 7, and 8 have identical bp numbers, whereas human SLC39A14 exon 5 is 3 bp longer. Between mouse and human, SLC39A8 exons 3, 4, 5, 7, and 8 have an identical number of base pairs, whereas mouse Slc39a8 exon 6 is 6 bp longer. The mouse and human ZIP14 transporter proteins contain 489 and 492 amino acids, respectively.

TABLE 1.

Comparison of the mouse and human SLC39A14 and SLC39A8 gene structures

All sequences are based on NCBI m37 mouse assembly (Apr 2007; strain C57BL/6J) and NCBI 36 assembly of the human genome (Nov 2005). From the 5′-most (exon 1c) start site to the 3′-most nucleotide of exon 9, the four genes from left to right span 92,502, 57,942 [?], 64,860, and 88,148 bp, respectively. For mouse Slc39a14, the genomic 5′ to 3′ sequence is exon 1c, intron 1c, exon 1b, intron 1b, exon 1a, intron 1a, exon 2, intron 2, exon 3, intron 3, exon 4a, intron 4a, exon 4b, intron 4b, exon 5, intron 5, … … exon 9.

	Mouse Slc39a14		Human SCL39A14		Mouse Slc39a8		Human SLC39A8
Number
	Exon	Intron	Exon	Intron	Exon	Intron	Exon	Intron
	bp		bp		bp		bp
1c	139	32,340	[?]	[?]	256	1486	184	801
1b	110	29,383	[?]	[?]	146	632	139	429
1a	131	15,130	104	37,287	37	156	26	159
2	279	2037	285	3329	451	22,325	748	28,613
3	187	604	187	1449	163	6725	163	8062
4(a)	170	4688	170	4664	170	1409	170	2324
4(b)	170	2625	170	2573
5	120	584	123	866	123	465	123	507
6	189	117	189	115	171	25,955	165	36,237
7	208	923	208	1370	208	197	208	197
8	185	1837	185	1716	185	2018	185	4296
9	516		3122		1582		1527
Totala	1993a	90,268	[4573 bp]	[53,369 bp]	3309	61,368 bp	3473	84,510
	1964				3199		3428
	1985				3090		3315

[?], not detected in the human genome.

^aThe three total lengths of exons denote the mRNA transcripts using exons 1c, 1b, or 1a, respectively (which are found in the dbEST database).

Note that, so far, none of the mouse or human genome databases has thoroughly cataloged the alternatively spliced exons 1 of mouse Slc39a14 or mouse or human SLC39A8; this is probably because of the shortcomings of present-day “gene-finder” software programs. By comparing the dbEST with genome databases, we were able to find mRNAs that included the exons 1 of mouse Slc39a14 and of mouse and human SLC39A8. We discovered no evidence, however, of alternatively spliced exons 1 for the human SLC39A14 gene (Table 1). Either these exons 1 exist but have not yet been captured in the dbEST or these exons 1b and 1c became no longer useful in the human ancestral branch and diverged into sequences that are no longer detectable.

During the ZIP14 cDNA cloning process and while studying the dbEST, we also discovered exons 4A and 4B; when alternatively spliced, this resulted in ZIP14A, ZIP14B, and ZIP14AB cDNAs (Fig. 1B). The ZIP14AB transcript has a disrupted open reading frame and would encode a truncated protein with only 157 amino acids; therefore, we did not study ZIP14AB further. Mouse Slc39a14 exons 4A and 4B are both 170 bp long (Table 1) and share 67% nucleotide identity (Fig. 2). Transcripts of ZIP14A and ZIP14B encode two different proteins, both having 489 amino acids but molecular masses of 53,754 and 53,962 Da, respectively. These two proteins differ only in the 57-amino acid region encoded by the alternative exons 4; of the 57 amino acids, 37 are identical and 20 are different. This segment includes almost all of TM1, the intracellular loop between TM1 and TM2, and almost all of TM2 (Fig. 2). Yet differences in exon 4 seem not to change the topology of the two transporter proteins. Percentage identity in the ZIP domain sequences, between mouse and human, is 92% for ZIP14 and 96% for ZIP8. Mouse ZIP14A and ZIP14B are both 73% identical to ZIP8 in the ZIP domain (Fig. 2). In contrast, for example, mouse ZIP14A and ZIP14B share 41% identity with mouse ZIP4 and 15% identity with mouse ZIP1.

Fig. 2.

Alignment of mouse ZIP8 (top), ZIP14A (middle), and ZIP14B (bottom) proteins. The only differences between ZIP14A and ZIP14B occur in the alternatively spliced exon 4, which results in 57 amino acids (from Val151 through Glu207). ZIP8 is 73% identical to ZIP14A, and ZIP8 is 73% identical to ZIP14B (in the ZIP domain only). ZIP14A is 67% identical to ZIP14B in exon 4. Boxes show the N terminus putative membrane localization signal. Arrows denote possible cleavage sites to render the mature membrane-bound proteins. Asterisks indicate potential glycosylation sites. The eight putative transmembrane (TM) regions are underlined and denoted as “TM1” through “TM8” (labeled in each case at the left end). The precise amino acid at which a TM domain begins or ends reflects the software program and could be off by a residue or two; there are no experimental data to support this. The mouse and human ZIP14 transporter proteins comprise 489 and 492 amino acids, respectively. In contrast, the mouse and human ZIP8 proteins contain 462 and 461 amino acids, respectively. Black-on-white denotes nonsimilar residues; blue-on-cyan denotes a consensus residue derived from a block of similar residues at a given position; black-on-green denotes a consensus residue derived from the occurrence of >50% of a single residue at a given position; red-on-yellow denotes a consensus residue derived from a completely conserved residue at a given position; green-on-white denotes a residue weakly similar to a consensus residue at a given position.

Tissue Distribution of ZIP14 versus ZIP8 mRNA

Q-PCR analysis of C57BL/6J mice revealed ZIP14A mRNA levels that were highest in liver > duodenum > kidney > testis > brain = lung and ZIP14B mRNA levels highest in liver = duodenum > brain = testis > kidney = lung (Fig. 3). In contrast, ZIP8 mRNA levels were highest in lung = testis > kidney ≫ liver > brain > duodenum. The ZIP14A/ZIP14B ratio is maximal (7.8-fold) in liver, next highest (∼4-fold) in kidney and duodenum, and lowest (0.7-fold) in brain. Total ZIP14 transcripts are 39- and 18-fold greater than ZIP8 transcripts in duodenum and liver, respectively, whereas ZIP8 transcripts are 12- and 5-fold greater than total ZIP14 transcripts in lung and testis, respectively.

Fig. 3.

Comparison of ZIP14A, ZIP14B, and ZIP8 mRNA levels in six tissues of the untreated C57BL/6J mouse. Values are expressed as means ± S.E. (n = 3 mice).

Effect of Metal Uptake on Cell Survival

Note that rvZIP14A cells transported both Cd²⁺ and Mn²⁺ substantially better than rvZIP14B cells (Fig. 4A). In control rvLUC cells, Cd²⁺ or Mn²⁺ uptake was very low during the 60-min experiment. Consistent with the Fig. 4A data, the ZIP14A transporter (Table 2) exhibited a 3- to 5-fold greater V_max for Cd²⁺ and Mn²⁺ uptake than the ZIP14B transporter. The V_max for Cd²⁺ uptake by ZIP8 was not statistically different from that by ZIP14A and more than 3-fold higher than that by ZIP14B (Table 2). On the other hand, the V_max for Mn²⁺ uptake by ZIP8 was 15 times less than that by ZIP14A and 4.5 times less than that by ZIP14B. Differences in maximum rates of transport (V_max) could very easily reflect differences in expression of the constructs or differences in trafficking of the transporter to the cell surface. In contrast, neither of these cellular processes influenced the K_m value.

Fig. 4.

A, comparison of Cd²⁺ versus Mn²⁺ uptake kinetics by retroviral (rv) stably transfected MFF cultures in HBSS. Cells were exposed to 0.25 μM Cd²⁺ (spiked with ¹⁰⁹CdCl₂) or 0.25 μM Mn²⁺ (spiked with ⁵⁴MnCl₂) for 60 min at 37°C. rvLUC, control cells having no ZIP protein. B, semi-log plot of cell survival as a function of increasing doses of Cd²⁺ or Mn²⁺, in four rvZIP cell lines versus rvLUC control cells. Cultures were exposed to the indicated concentrations of Cd²⁺ or Mn²⁺ for 32 h at 37°C in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. After Cd²⁺ exposure, the rvZIP8, rvZIP14A, and rvZIP14B lines are significantly different (P < 0.001) from the rvZIP4 and rvLUC lines at all Cd²⁺ concentrations between 0.6 and 30 mM. After Mn²⁺ exposure, the rvZIP8, rvZIP14A, and rvZIP14B lines are significantly different (P < 0.05) from the rvZIP4 and rvLUC lines at all Cd²⁺ concentrations between 0.3 and 10 mM. Cell survival was monitored using the 3-(4,5-dimethlythiazol-2-yl)-2,5-diphenyl tetrasodium bromide assay. Values are expressed as means ± S.E. (n = 3 wells; two experiments done in different weeks).

TABLE 2.

Comparison of transport kinetics of ZIP14A-, ZIP14B-, and ZIP8-mediated Cd²⁺ versus Mn²⁺ uptake

Cells were grown in Dulbecco's modified Eagle's medium until 60 min before the experiment, at which time the medium was changed to HBSS. All measurements were made under conditions of the initial linear-uptake rate. Values are expressed as means ± S.E.

	Cadmium			Manganese
	ZIP14A	ZIP14B	ZIP8a	ZIP14A	ZIP14B	ZIP8a
Vmax (pmol/min/mg protein)	113 ± 12	25 ± 3	92.1 ± 10	1140 ± 100	330 ± 29	73.8 ± 6
Km (mM)	1.1 ± 0.02	0.14 ± 0.02	0.62 ± 0.07	18.2 ± 2	4.4 ± 0.5	2.2 ± 0.3

^aZIP8 uptake kinetics was reported previously (He et al., 2006).

In terms of K_m values, ZIP14B showed a ∼8-fold greater affinity for Cd²⁺ and ∼4-fold greater affinity for Mn²⁺ than ZIP14A (Table 2). ZIP8 showed almost 2-fold greater affinity for Cd²⁺ than ZIP14A and 4 times less affinity than ZIP14B. ZIP8 showed 8-fold greater affinity for Mn²⁺ than ZIP14A and 2-fold greater affinity for Mn²⁺ than ZIP14B.

The survival curves for cells exposed to Cd²⁺ were not significantly different among rvZIP14A, rvZIP14B, and rvZIP8 cells (Fig. 4B). In contrast, rvLUC control cells and rvZIP4 cells displayed at least 20-fold greater cell survival—because of their relative inability to take up Cd²⁺ (He et al., 2006). On the other hand, increasing Mn²⁺ concentrations showed no more than 2-fold differences in cell survival for rvZIP8, rvZIP14A, and rvZIP14B, compared with that for rvZIP4 or rvLUC cells (Fig. 4B), even though the difference in Mn²⁺ uptake by ZIP14A or ZIP14B was very substantial, compared with that by control cells (Fig. 4A).

Metal-Mediated Competitive Inhibition of Cd²⁺ Uptake

Fig. 5 shows that Zn²⁺ was by far the best inhibitor of Cd²⁺ uptake by both rvZIP14A and rvZIP14B cells. Tied for second-best inhibitor were Mn²⁺ and Cu²⁺. No significant inhibition of Cd²⁺ uptake—by either rvZIP14A or rvZIP14B cells—was seen with Cs²⁺, Fe²⁺, or Fe³⁺ ions (Fig. 5).

Fig. 5.

Metal cation competition for Cd²⁺ uptake in rvZIP14A (Z14A) versus rvZIP14B (Z14B) cells. ¹⁰⁹CdCl₂ was added to make a final Cd²⁺ concentration of 0.25 μM; the competing metal cations (as chloride salts) at concentrations of 0, 1, 4, or 16 μM were added at the same time as Cd²⁺, and the cells were incubated at 37°C for 20 min, after which Cd²⁺ accumulation was determined. Zn²⁺ was most inhibitory (*, P = 0.0007), with Mn²⁺ and Cu²⁺ as second most inhibitory (†, P < 0.01). No significant inhibition was seen with Cs²⁺, Fe²⁺, or Fe³⁺ (P > 0.05). Two-tailed P values were calculated using Student's t test with 4 degrees of freedom.

Other Parameters Affecting ZIP14-Mediated Cd²⁺ Uptake

Cd²⁺ uptake by rvZIP14A or rvZIP14B cells was maximal at 37°C and maximal when (extracellular) pH of the transport medium was 7.5 (data not shown), demonstrating that ZIP14 transporters are not proton-coupled. ZIP14A- and ZIP14B-mediated Cd²⁺ transport was strongly inhibited by cyanide, indicating a dependence on an energy source, such as ATP. Cd²⁺ influx was not affected by any level of Cl⁻, Na⁺, or K⁺ ions added (data not shown), suggesting that none of these ions is coupled to Cd²⁺ transport. These findings are identical to those found with rvZIP8 cells (He et al., 2006).

Dependence of ZIP14-Mediated Cd²⁺ Uptake on ${\text{HCO}}_3^{-}$ Ion

We found that Cd²⁺ uptake by rvZIP14A or rvZIP14B cells was dependent on extracellular ${\text{HCO}}_3^{-}$ concentrations, compared with control rvLUC cells (Fig. 6A). Significantly less Cd²⁺ uptake was seen at 0 and 1 mM ${\text{HCO}}_3^{-}$ than at ≥2 mM. By 4 mM ${\text{HCO}}_3^{-}$, Cd²⁺ influx had clearly reached its maximum. The 0 mM ${\text{HCO}}_3^{-}$ in rvZIP14 cells in HBSS (Fig. 6A) is actually not free of ${\text{HCO}}_3^{-}$, because there is exogenous ${\text{HCO}}_3^{-}$ present from dissolved CO₂ in the air, as well as the ${\text{CO}}_2{/\text{HCO}}_3^{-}$ derived from cellular metabolism. From the Henderson-Hasselbach equation of this buffer system, ${\text{pH}=\text{pK}+\text{log}([\text{HCO}}_3^{-}]/0.03\times {\text{pCO}}_2)$, one can calculate that putatively “ ${\text{HCO}}_3^{-}$-free” medium at 37°C and pH 7.5 actually contains 171 μM ${\text{HCO}}_3^{-}$; this is likely the reason why, at 0 mM ${\text{HCO}}_3^{-}$ HBSS, we are still able to observe substantial amounts of both ZIP14A- and ZIP14B-mediated Cd²⁺ transport, which are much greater than that seen in rvLUC cells (Fig. 6A).

Fig. 6.

Dependence of Cd²⁺ uptake on ${\text{HCO}}_3^{-}$ in the medium. A, Cd²⁺ uptake over a range from 0 to 30 mM ${\text{HCO}}_3^{-}$ levels in HBSS. The black bars denote Cd²⁺ uptake in rvLUC control cells. At 0 mM ${\text{HCO}}_3^{-}$ (*, P < 0.01) and at 1 mM ${\text{HCO}}_3^{-}$ (†, P < 0.05), Cd²⁺ uptake by both rvZIP14A and rvZIP14B cells was significantly different from that seen at ${\text{HCO}}_3^{-}$ concentration between 2 and 30 mM. B, Cd²⁺ uptake as a function of increasing DIDS concentrations. Cd²⁺ uptake by rvLUC cells was subtracted from that by the rvZIP14A or rvZIP14B cells. The transport medium contains nominal ${\text{HCO}}_3^{-}$ concentrations (171 μM). The DIDS solution was prepared fresh in dimethyl sulfoxide and added to the uptake medium 30 min before addition of the Cd²⁺. In A and B, cells were incubated with 0.25 μM Cd²⁺, spiked with ¹⁰⁹CdCl₂, for 20 min at 37°C. At all ${\text{HCO}}_3^{-}$ concentrations and all DIDS concentrations, the pH was maintained at ∼7.4, and the balance of anions and cations was carefully controlled for, as detailed in He et al. (2006).

DIDS is a well known inhibitor of ${\text{HCO}}_3^{-}$-dependent transporters (Cabantchik and Greger, 1992). To further confirm the dependence of Cd²⁺ uptake on ${\text{HCO}}_3^{-}$, we added DIDS to 171 μM ${\text{HCO}}_3^{-}$-containing HBSS medium before adding Cd²⁺ (Fig. 6B). In the absence of DIDS (far left), Cd²⁺ uptake by rvZIP14A and rvZIP14B was ∼180 and ∼155 pmol/min/mg protein, respectively (after subtracting Cd²⁺ uptake by rvLUC cells). At the very low levels of 0.125 and 0.25 mM, DIDS inhibited Cd²⁺ uptake in rvZIP14 cells by ∼28 and ∼44%, respectively. On the other hand, at 1.0 mM DIDS—a concentration regularly used to inhibit 4 mM ${\text{HCO}}_3^{-}$ uptake—DIDS inhibited Cd²⁺ uptake close to 100% in both rvZIP14A and rvZIP14B cells.

Membrane Localization and Characterization of ZIP14 Transporters

Both ZIP14A and ZIP14B proteins were shown to be localized exclusively to the apical surface of MDCK polarized epithelial cell monolayers; this can best be seen in the apical plane of the x-y projections and the z-y planes traversing from basal to apical, as well as the tangential cut (Fig. 7, right column and merged middle column). An apical-surface control, lectin (red), and a basolateral-surface control, NBC1 (green), were included in these confocal studies.

Fig. 7.

Z-stack confocal microscopy of MDCK-cell monolayers. Green fluorescent protein-NBC1 expression (top three rows); ZIP14Aha expression (next four rows); ZIP14Bha expression (bottom four rows). Two days after transfection, the confluent cells were incubated in Chelex 100 medium for 1 h, fixed with 3% formaldehyde for 20 min, and then permeabilized with 0.1% Triton X-100 for 4 min and blocked with 10% fetal bovine serum-containing PBS medium for 1 h. All of this was done at room temperature. The α-HA (Bethyl Laboratories, Montgomery, TX), at 1:500 dilution, was incubated at 4°C overnight with cells in 1% bovine serum albumin containing PBS. The next day, the primary antibody solution was removed and the cells were washed three times with PBS for 5 min each time. The secondary antibody Alexa 488-α-rabbit (causing green fluorescence) (Invitrogen) was incubated with the cells at room temperature for 1 h. Cell monolayers were then costained with either a membrane marker (F-actin) phalloidin-tetramethylrhodamine (Invitrogen) or specifically an apical membrane marker peanut agglutinin lectin (Invitrogen) at 1/250 dilution for 20 min at room temperature and then mounted for confocal analysis. The red color (F-actin or lectin) in the left column of panels, plus the green color (NBC1, ZIP14A, ZIP14B) in the right column of panels, are merged together in the middle column of panels. The x-y plane for NBC1 expression (top) was taken through the middle, whereas the x-y projections for ZIP14A (middle) and ZIP14B (bottom) were taken both through the apical and basal surfaces of the MDCK cells, as labeled. The Z-Y plane is at a right angle to the x-y plane and illustrates the basal→apical orientation of the polarized epithelial cells. The “Cut” refers to a tangential plane.

Figure 8 illustrates our study with PNGase F. We used HA-tagged proteins because so far we lack an efficient anti-ZIP antibody for Western blots. Glycosylated ZIP14A proteins (third lane) appear as 60- and 70-kDa faint bands; after PNGase F treatment, the native ZIP14A protein is close to its calculated 53,754-Da size (fourth lane). Glycosylated ZIP14B proteins (fifth lane) appear as a light doublet at ∼68 to 72 kDa; after PNGase F treatment, native ZIP14B shows up as two bands, the smaller one close to its calculated 53,962-Da size (sixth lane). A positive control, glycosylated ZIP4 proteins (seventh lane), appear as faint bands in the 80- to 90-kDa range; after PNGase F treatment, the native ZIP4 is close to its calculated 68.4-kDa size (last lane at right).

Fig. 8.

Western immunoblot of proteins from control rvLUC, rvZIP14Aha, rvZIP14Bha, and rvZIP4ha—with or without PNGase F treatment. All three ZIP proteins are tagged with hemagglutinin; thus, the α-HA antibody was used. Rat β-actin protein was used to control for lane-loading. Markers (kilodaltons) are shown on the left.

Discussion

Evolutionary Analysis

The mouse Slc39a14 and Slc39a8 genes are far more closely related to one another than to any of the other 12 family members (Fig. 1A). Although both genes have nine exons and three alternative exons 1, Slc39a14 differs from Slc39a8 in that it has two exons 4 (Fig. 1B). Excluding additional splice sites in the 3′ untranslated region, this means at least six Slc39a14 transcripts are possible, if one considers all combinations of alternative splicing in exons 1 plus exons 4. These different splice variants might dictate temporal, tissue, or cell-type specificity. This alternative exon splicing pattern of the mouse Slc39a14 gene was recently reported (Liuzzi et al., 2005) but wrongly attributed to exon 6. From an evolutionary standpoint, these data indicate that, after the vertebrate SLC39A14 and SLC39A8 genes diverged, an exon 4 duplication event occurred in the SLC39A14 gene but not in the SLC39A8 gene. This probably happened as a stochastic event and then became fixed in the mammalian ancestral genome, because having exons 4A plus 4B in two different ZIP14 transporters must have conferred a selective advantage (i.e., greater success in fertility, reproduction, viability, or defense against the environment).

Between mouse and human, each internal exon has very similar, if not identical, lengths, and the exon-intron junctions are highly conserved (Table 1). The only differences are in multiples of three base pairs; this is commonly seen during evolution of different species, wherein an extra amino acid or two is added at the beginning or the end of an exon (in one species but not the other), but the final result is to maintain the open reading frame of the mRNA.

Since the human-rodent split ∼65 million years ago, we conclude that no SLC39 genes have been lost or gained between these two species. Takefugu rubripes (pufferfish) has one SLC39 gene that shows similarly high homology to mouse (and human) SLC39A14 and SLC39A8; this finding indicates that either SLC39A14 or SLC39A8 diverged from the other (via a gene duplication event) after the land animal-sea animal split ∼425 million years ago (Fig. 1A), after which some selective advantage compelled both transporter genes to remain in the land-animal branch.

Tissue Distribution of ZIP14 versus ZIP8 mRNA

Of six tissues examined, we found C57BL/6J mouse ZIP14A mRNA levels to be twice as high in liver as in duodenum, and kidney third highest, with half that found in duodenum; ZIP14B mRNA levels were closely similar but ranked as liver > duodenum > brain > testis > kidney > lung (Fig. 3). In the only other study we could find on this topic (Liuzzi et al., 2006), two CD-1 mice were pooled, and ZIP14 mRNA by Q-PCR analysis was highest in duodenum > jejunum > liver = heart > kidney > spleen = pancreas. Thus, it appears that ZIP14 expression is highest in the liver, gastrointestinal tract, kidney, and heart. Whether these relative differences between ZIP14A and ZIP14B transcripts in different organs have any functional significance has not yet been determined.

We found mouse ZIP8 mRNA levels to be twice as high in lung and testis, relative to that in kidney (Fig. 3, right). Northern blot analysis of untreated C57BL/6J mice (Wang et al., 2007) agrees with the present Q-PCR data. We have also found that mouse yolk sac and placenta contain substantial levels of both ZIP14 and ZIP8 mRNA (B. Wang and D. W. Nebert, manuscript in preparation). An earlier extensive study via Northern blot (Begum et al., 2002) showed that human ZIP8 mRNA levels are most abundant in the pancreas > lung > placenta > liver = thymus ≫ spleen = testis = ovary = small intestine. Although we might conclude that human ZIP8 expression is highest in pancreas, lung, placenta, and liver, these data indicate that there might be important species differences between mouse and human in ZIP8 expression in other tissues. Expressed-sequence tags for ZIP14 and ZIP8 cDNAs are ubiquitous, however, and both can be found in more than 30 mouse and human tissues and cell types (http://www.ncbi.nlm.nih.gov).

Metal Uptake and Cell Survival

Intriguingly, Fig. 4 shows that Cd²⁺ uptake parallels cell damage, whereas Mn²⁺ uptake does not. We believe the difference between the cell toxicity profile for Cd²⁺ and Mn²⁺ is related to the ability of Cd²⁺ to bind up irreversibly the intracellular thiols (by a two-electron mechanism); Mn²⁺ lacks this electrophilic property.

ZIP14B and ZIP8 show the highest affinities for Cd²⁺ and for Mn²⁺ (Table 2). Given the tissues where these transporters are expressed at the highest levels, it is likely that when environmental Cd²⁺ (usually at very low levels) enters the body, by ingestion or inhalation, these two transporters would take it up at the interface between the environment and the organism's portal-of-entry tissues.

We attempted also to study Zn²⁺ influx but found that Zn²⁺ uptake in rvZIP14 cells, as had been found in rvZIP8 cells (He et al., 2006), was problematic. In rvZIP cells, Zn²⁺ uptake kinetics could not be quantified—as far as calculating reasonable V_max or K_m values. This is most probably due to the obscuring effects of too many other Zn²⁺ transporters on the surface of mammalian cells in culture. We have therefore studied Zn²⁺ transport in X. laevis oocytes, which are known to have negligible amounts of interfering transporters (Bossi et al., 2007) compared with any transporter cRNA being microinjected. Indeed, preliminary experiments with ZIP8-injected oocytes have determined the very low K_m values of 0.26 ± 0.09 and 0.48 + 0.08 μM for Zn²⁺ and Cd²⁺ uptake, respectively (Liu et al., 2008). We expect similar low K_m values in ZIP14 cRNA-injected oocytes.

Metal-Mediated Competitive Inhibition of Cd²⁺ Uptake

The best inhibitor of ZIP14-mediated Cd²⁺ uptake is Zn²⁺, followed by Mn²⁺ and Cu²⁺ (Fig. 5). Consistent with this finding, ZIP8-mediated Cd²⁺ uptake is most inhibited by Zn²⁺ in X. laevis oocytes (Liu et al., 2008). The inhibitory data for Mn²⁺ (Fig. 5) are similar to what was found for ZIP8 in MFF cultures (He et al., 2006). Finding Cu²⁺ as a significant inhibitor of ZIP14-mediated Cd²⁺ uptake (Fig. 5) is a surprise, because Cu²⁺ did not inhibit ZIP8-mediated Cd²⁺ uptake (He et al., 2006).

We did not find Fe²⁺ to be an inhibitor of Cd²⁺ influx (Fig. 4). This observation contradicts a recent observation (Liuzzi et al., 2006) suggesting that ZIP14 mediates nontransferrin-bound Fe²⁺ into cells. Those Fe²⁺ uptake experiments were carried out in cultured human embryonic kidney 293H cells, Sf9 insect cells, and AML12 mouse hepatocytes, whereas our experiments were carried out in MFF cultures; we do not understand why there should be this discrepancy. In the end, ZIP14-mediated Fe²⁺ uptake in the intestine or liver will need to be examined in the intact animal.

Dependence of ZIP14-Mediated Cd²⁺ Uptake on ${\text{HCO}}_3^{-}$ Ion

As was found with ZIP8 (He et al., 2006), ZIP14-mediated Cd²⁺ uptake is dependent on extracellular ${\text{HCO}}_3^{-}$ levels, and very low concentrations of DIDS are highly effective at blocking Cd²⁺ uptake when extracellular ${\text{HCO}}_3^{-}$ is 171 μM (Fig. 6). These data indicate that ${\text{HCO}}_3^{-}$ is essential for ZIP14-mediated Cd²⁺ influx.

As was found for ZIP8, both ZIP14A and ZIP14B transporters can thus be regarded as ${\text{Cd}}^{2+}{/\text{HCO}}_3^{-}$ symporters. Apparently, any differences in exon 4 between these two proteins do not affect this absolute dependence on the ${\text{HCO}}_3^{-}$ anion. It has been demonstrated recently in X. laevis oocytes that the ${\text{Cd}}^{2+}{/\text{(HCO}}_3^{-}{)}_2$ or ${\text{Zn}}^{2+}{/\text{(HCO}}_3^{-}{)}_2$ complex transported by ZIP8 is electroneutral and that the ZIP8 transporter is largely internalized under conditions of replete Zn²⁺, whereas it undergoes trafficking to the cell surface under conditions of Zn²⁺ depletion (Liu et al., 2008).

Membrane Localization and Characterization of ZIP14 Transporters

We have shown that the membrane-bound ZIP14A and ZIP14B transporters are localized to the apical surface (Fig. 7) and generally glycosylated (Fig. 8). These findings are similar to what has been found with ZIP8 (Dalton et al., 2005; He et al., 2006) and ZIP4 (Dufner-Beattie et al., 2003, 2004; Liuzzi et al., 2004; Huang et al., 2006). In contrast, the ZIP5 transporter has been localized to the basolateral surface (Wang et al., 2004). ZIP14 has four potential glycosylation sites: Asn75, Asn85, Asn100, and Asn455; ZIP8 has two potential N-linked glycosylation sites, Asn40 and Asn88 (Fig. 2).

Concluding Remarks

The present study has shown that the mouse Slc39a14 and Slc39a8 genes are very similar, both with nine exons and three alternatively spliced noncoding exons 1; Slc39a14 has the additional characteristic of two alternatively spliced exons 4. We have also shown that ZIP14A and ZIP14B transporters share many properties with ZIP8: Cd²⁺ and Mn²⁺ uptake with K_m values in the micromolar range, rendering cells sensitive to Cd²⁺ toxicity; inhibition of Cd²⁺ uptake by Mn²⁺; ${\text{HCO}}_3^{-}$ dependence for divalent cation uptake; localization to the apical membrane of polarized epithelial MDCK cells; and glycosylation of the transporter proteins. An observation with the ZIP14 transporter, not seen with ZIP8, is the inhibition of Cd²⁺ uptake by Zn²⁺ and Cu²⁺; however, inhibition of ZIP8-mediated Cd²⁺ uptake by Zn²⁺ and Cu²⁺ has been demonstrated in X. laevis oocytes (Liu et al., 2008).

Differences in tissue-specific expression were also seen between ZIP14 and ZIP8. We found ZIP14A + ZIP14B transcript levels to be highest in the liver, duodenum, and kidney, whereas ZIP8 levels are highest in lung, testis, and kidney. It seems likely that both ZIP14 and ZIP8 play important roles in Zn²⁺ homeostasis. Their presence in cells at the interface—between the lumen and epithelial cells of the small intestine, and between inhaled air and alveolar cells of the lung—could be essential in combating Zn²⁺ loss during important critical life processes such as inflammation and immune function. Unfortunately, the location and high affinity for Cd²⁺ of the ZIP14A, ZIP14B, and ZIP8 transporters can also serve as the means for bringing unwanted environmental Cd²⁺ into these tissues.

Abbreviations

ZIP

Zrt- and Irt-related protein

TM

transmembrane domain

HBSS

Hanks' balanced salt solution

MFF

mouse fetal fibroblast

MDCK

Madin-Darby canine kidney

PCR

polymerase chain reaction

bp

base pair(s)

HA

hemagglutinin

DIDS

4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid

PNGase F

peptide N-glycosidase F [peptide-N⁴-(N-acetyl-β-glucosaminyl)asparagine amidase]

SLC

solute carrier

dbEST

database for expressed sequence tags

rvZIP14

vZIP8, rvZIP4, rvLUC, retrovirally infected MFF cells expressing ZIP or firefly luciferase

NBC1

sodium-bicarbonate-1 cotransporter

vMDCK-ZIP14

MDCK-ZIP8, transiently transfected MDCK cells expressing ZIP or firefly luciferase

PBS

phosphate-buffered saline