Identification of the mammalian Na,K-ATPase 3 subunit.

We have isolated and characterized cDNA clones encoding the human and rat Na,K-ATPase beta3 subunit isoform. The human cDNA encodes a polypeptide of 279 amino acids that exhibits primary sequence and secondary structure similarities to Na,K-ATPase beta subunit isoforms. Sequence comparisons showed that the human beta3 subunit closely resembles the beta3 subunit of Xenopus laevis (59% amino acid identity) and is less similar to the human Na,K-ATPase beta1 and beta2 subunits (38% and 48% amino acid identity, respectively). By analyzing the segregation of restriction fragment length polymorphisms among recombinant inbred strains of mice, we localized the beta3 subunit gene to murine chromosome 7. Northern blot analysis revealed that the beta3 subunit gene encodes two transcripts that are expressed in a variety of rat tissues including testis, brain, kidney, lung, stomach, small intestine, colon, spleen, and liver. Identification of the mammalian beta3 subunit suggests an even greater potential for Na,K-ATPase isoenzyme diversity than previously realized.

Na,K-ATPase is a membrane-associated enzyme responsible for the active transport of Na ϩ and K ϩ in most animal cells. By coupling the hydrolysis of ATP to the movement of Na ϩ and K ϩ ions across the plasma membrane, the enzyme produces the electrochemical gradients that are the primary energy source for the active transport of nutrients, the action potential of excitable tissues, and the regulation of cell volume (1,2).
In all tissues from which Na,K-ATPase has been isolated, the enzyme has been found to consist of two subunits present in equimolar amounts. The ␣ subunit is a polypeptide of ϳ100 kilodaltons (kDa) which contains the binding sites for ATP and cardiac glycosides such as digoxin and ouabain (3). The ␤ subunit is a glycosylated polypeptide of molecular mass ϳ50 -60 kDa (2). The function of the ␤ subunit has yet to be elucidated.
The ␣ and ␤ subunits of Na,K-ATPase are each encoded by multigene families. Three ␣ subunit and two ␤ subunit genes have been localized to different chromosomes in the mouse (4,5), and cDNA clones encoding three rat ␣ (␣1, ␣2, ␣3) and two ␤ (␤1, ␤2) subunit isoforms have been characterized (6 -9). Substantial differences in the tissue and cell specificity of expression have been found for each ␣ and ␤ subunit. ␣1 subunit polypeptides have been detected in virtually all rat tissues (10). In contrast, ␣2 subunits are expressed predominantly in brain, heart, lung (10), and skeletal muscle (11), while ␣3 subunits are abundant in tissues of neural origin (10,12,13). Expression of ␤1 subunit polypeptides has been detected in brain, heart, and kidney (10), whereas ␤2 subunits are expressed predominantly in brain (14), pineal gland (12), and photoreceptor cells (13). Within the central nervous system, expression of the ␣3 subunit is restricted exclusively to neurons, while ␤2 subunit expression appears to be astrocyte-specific (15,16). The identification of multiple ␣ and ␤ subunit isoforms has raised questions regarding the extent of Na,K-ATPase isoenzyme complexity. RNA hybridization (15) and immunoblotting analyses (16) suggest that ␣ and ␤ subunit association is promiscuous, and that all six ␣/␤ subunit combinations are likely to exist.
Several lines of evidence have suggested the potential existence of additional ␤ subunit isoforms. Antibody probes have failed to detect expression of ␤1 or ␤2 subunits in several rat tissues, despite the presence of ␣ subunits in these tissues (12,14). A putative ␤3 subunit isoform has been identified in the toad Xenopus laevis (17). However, in the absence of identifiable mammalian ␤3 and Xenopus ␤2 subunit sequences, it is unclear whether the Xenopus ␤3 subunit actually represents the amphibian homolog of the mammalian ␤2 subunit.
In a search of the GenBank TM EST 1 (expressed sequence tag) data base, we identified a cohort of cDNA clones that encode the human Na,K-ATPase ␤3 subunit. The ␤3 subunit gene has been localized to mouse chromosome 7. This gene is expressed in a wide variety of rat tissues including brain, kidney, and liver. The identification of a mammalian ␤3 subunit suggests the possible existence of 9 distinct Na,K-ATPase isoenzymes.

EXPERIMENTAL PROCEDURES
Isolation and Characterization of cDNA Clones-Two human ESTs with sequence homology to the Xenopus Na,K-ATPase ␤3 subunit were identified in a search of the GenBank TM EST data base. BLAST analysis (18) identified additional human ESTs (I.M.A.G.E. Consortium, Lawrence Livermore National Laboratory) with overlapping sequence homology. One of these clones (I.M.A.G.E Consortium clone 139611) was obtained from Research Genetics (Huntsville, AL), sequenced in its entirety, and found to encode a full-length ␤3 subunit. A cDNA fragment encoding a portion of the ␤3 subunit was isolated from rat placental RNA by reverse transcriptase-mediated polymerase chain reaction (PCR). First strand cDNA was synthesized from 1 g of total rat placental RNA. A (dT) 17 adapter (50 pmol) served as primer for Moloney murine leukemia virus reverse transcriptase (Stratagene) as described previously (19). PCR was performed using the following primer mixtures. * This work was supported in part by National Institutes of Health Grants HL-39263 and GM-49023 (to R. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) U51478. corresponding to amino acids 173-180 of the human ␤3 polypeptide. In the primer sequence, I specifies inosine. PCR was carried out with Taq polymerase (Perkin-Elmer) for 40 cycles (1 min at 95°C, 1 min at 37°C, and 2 min at 72°C) using buffers supplied by the distributor. Buffers contained 1.5 mM magnesium. PCR products were inserted into the EcoRI site of Bluescript (Stratagene). Both the rat and human cDNAs were sequenced by dideoxynucleotide chain termination sequencing using a Sequenase kit (United States Biochemicals). DNA sequences were analyzed using the programs from the University of Wisconsin Genetics Computer Group (20).
RNA Blot Hybridization-RNA was extracted from rat tissues using Trizol Reagent (Life Technologies, Inc.) according to the method described by Chomczynski and Sacchi (21). Conditions for electrophoresis, transfer, and hybridization were as described previously (19). A 320base pair segment of the rat reverse transcriptase-mediated PCR fragment was radiolabeled by the random priming method (22) using a High Prime DNA Labeling Kit (Boehringer Manheim) and used as probe. Blots were washed to a final stringency of 0.1 ϫ SSC (1 ϫ SSC is 0.15 M NaCl, 15 mM sodium citrate), 0.1% SDS at 65°C, and exposed to Kodak XAR-5 film at Ϫ80°C with an intensifying screen.
Chromosomal Localization-DNA samples from recombinant inbred (RI) strains of mice (A/J, C57BL/6J) were purchased from the Jackson Laboratory (Bar Harbor, ME). Genomic DNA was digested to completion with MspI, separated on a 1% agarose gel, transferred to a nylon filter membrane (Hybond-N, Amersham), and hybridized as described previously (19). Blots were washed to a final stringency of 0.2 ϫ SSC, 0.5% SDS at 65°C for 45 min, then exposed to Kodak XAR film at Ϫ80°C with an intensifying screen. The strain distribution pattern of the MspI polymorphism in RI strains derived from A/J ϫ C57B1/6J crosses (AXB and BXA) was determined as described previously (5,19) and is presented in Table II.

Identification and Characterization of ␤3 Subunit cDNA
Clones-In a search of the GenBank TM EST data base, we identified several human cDNAs having homology to Na,K-ATPase ␤ subunits. A subset of these clones appeared to encode a Na,K-ATPase ␤ subunit (␤3) distinct from either the Na,K-ATPase ␤1 or ␤2 subunits. Both strands of a full-length human cDNA (I.M.A.G.E. Consortium clone 139611) were subjected to dideoxynucleotide sequencing. The complete nucleotide se-FIG. 1. Nucleotide and deduced amino acid sequences of human and rat ␤3 subunits. The nucleotide sequences of the human and rat ␤3 subunits are shown above the deduced amino acid sequences. Rat sequences derived from PCR primers are depicted by lowercase letters. Amino acids are numbered to the right of the sequence beginning with the initiating methionine. The segment of the rat ␤3 subunit cDNA that was used to probe rat tissue RNA is underlined.  (20). Sequences were obtained from GenBank™ or SwissProt data bases and are as described in Fig. 4 quence of the open reading frame and the deduced amino acid sequence of the protein is shown in Fig. 1. The human ␤3 cDNA has a 27-nucleotide 5Ј-untranslated region, an 840-nucleotide open reading frame, and a 507-nucleotide 3Ј-untranslated sequence. The nucleotide sequence of the coding region of the human cDNA exhibits 40% homology with the human Na,K-ATPase ␤2 subunit (9). The open reading frame encodes a 279-amino acid polypeptide with a molecular mass of 31,639 daltons. As shown in Table I, the amino acid sequence of the ␤3 polypeptide exhibits 59% identity with the ␤3 subunit of X. laevis (17), 38% identity with the human Na,K-ATPase ␤1 subunit (23), and 48% identity with the human ␤2 subunit (9). We therefore conclude that the human cDNA encodes a mammalian Na,K-ATPase ␤3 subunit isoform. PCR was used to generate a portion of the corresponding rat cDNA. The nucleotide and deduced amino acid sequence of the rat PCR product is presented in Fig. 1. The rat ␤3 subunit PCR product corresponds to amino acids 26 through 180 of the human polypeptide. In the region between the PCR primers (amino acids 34 -172), the rat and human ␤3 subunits exhibit 83% nucleotide and 80% amino acid sequence identity, indicating that the PCR product is likely to represent a segment of the rat ␤3 subunit. 2 The extent of amino acid sequence divergence exhibited by the rat and human ␤3 subunit contrasts with that observed for the corresponding region of the ␤1 (95% identity between rat and human) and ␤2 (97% identity) subunits. Sequence comparisons indicate that the Na,K-ATPase ␤3 subunit is more divergent among species than either the ␤1 or ␤2 subunits.
Structure of the Human Na,K-ATPase ␤3 Subunit-A comparison of the amino acid sequence of the human ␤3 subunit with the human ␤2 subunit is shown in Fig. 2. The human ␤3 subunit consists of 279 amino acids, whereas the human ␤2 subunit is composed of 290 amino acid residues. Of the 279 residues compared, 48% positions were occupied by identical residues while 18% were occupied by favored substitutions. The asparagine residues marked with asterisks represent potential sites of N-linked glycosylation. There are two such sites in the human ␤3 subunit compared to seven in the ␤2 subunit. The positions of the N-linked glycosylation sites in the human ␤3 subunit are not strictly conserved with any of the N-linked glycosylation sites in the human ␤1, human ␤2, or Xenopus ␤3 subunit isoforms. There are 6 cysteine residues (positions 128, 144, 154, 170, 191, and 250) within the proposed extracellular domain of the ␤3 subunit. In the computer-aligned sequences, these cysteine residues appear to be highly conserved among Na,K-ATPase and H,K-ATPase ␤ subunits. A putative transmembrane segment is located between residues 39 and 66. Of the 28 amino acid residues in this region, 21 are identical and 6 are conservative substitutions between the ␤2 and ␤3 subunits.
Hydropathy profiles of the human Na,K-ATPase ␤1, ␤2, and ␤3 subunit sequences obtained by using the algorithm of Kyte and Doolittle (24) are shown in Fig. 3. This analysis predicts that all three ␤ subunit isoforms contain a highly charged cytoplasmic amino terminus followed by a single hydrophobic transmembrane segment of 28 amino acids and a large extracellular carboxyl-terminal domain. A comparison of the amino acid sequence of the human ␤3 subunit with other Na,K-ATPase ␤ subunits is presented in Fig. 4 and Table I. A computer-assisted analysis of sequence similarities (Fig. 4) shows that the human ␤3 subunit clusters with the ␤3 subunits of Xenopus and Bufo marinus (25). These sequences are clearly distinct from clusters containing Na,K-ATPase ␤1 and ␤2 subunit isoforms. Interestingly, the human ␤3 subunit sequence is most closely related to that of the chicken ␤ subunit described as ␤2 (61% amino acid sequence identity) (26). The fact that the chicken ␤2 subunit sequence clusters with ␤3 subunit sequences (Fig. 4) suggests the possibility that the chicken sequence actually represents the avian ␤3 subunit.
Expression of ␤3 Subunit mRNA in Rat Tissues-A panel of adult rat tissues was examined for the presence of Na,K-ATPase ␤3 subunit mRNA. The pattern of expression of ␤3 subunit mRNA is shown in Fig. 5. The Na,K-ATPase ␤3 subunit gene encodes two transcripts, ϳ1.6 and ϳ1.8 kb in size. Of the rat tissues analyzed, ␤3 subunit mRNA was detected at highest levels in testis. Transcripts of the ␤3 subunit gene were present in much lower abundance in brain, kidney spleen, and lung, and at even lower levels in stomach, colon, and liver. No ␤3 subunit transcripts were detectable in heart. Transcripts of the ␤3 subunit gene were also present in placenta and mammary gland of pregnant rats and in the brain of 19-day-old fetal rats (data not shown).
Chromosomal Localization of the Na,K-ATPase ␤3 Subunit Gene-We have used segregation of restriction fragment length polymorphisms (RFLPs) among recombinant inbred (RI) strains of mice to identify the chromosomal location of the mouse gene encoding the Na,K-ATPase ␤3 subunit (Atp1b3). 3 Mouse genomic DNA sequences were identified by hybridizing Southern blots to radiolabeled full-length human ␤3 subunit cDNA. As shown in Fig. 6, this probe hybridized to three major genomic fragments, 4.3, 3, and 1.5 kb long, in MspI-digested

FIG. 2. Comparison of the human ␤3
and ␤2 subunits. The deduced amino acid sequence of the human ␤3 subunit is shown above the human ␤2 subunit. Ellipses in either sequence allow optimal alignment for amino acid insertions/deletions. Asterisks denote asparagine residues which are possible sites of N-linked glycosylation. The putative transmembrane-spanning domains are underlined. Identical residues are shaded. Conserved cysteines are indicated by arrowheads. Amino acids are numbered above the sequence.
A/J DNA, and two fragments, 4.3 and 3 kb long, in MspIdigested C57BL/6J DNA. The 4.3-kb and 3-kb fragments are common to the two strains, whereas the 1.5-kb fragment is specific to A/J mouse DNA. Segregation analysis of the MspI RFLP in AXB and BXA RI strains (Table II) reveals linkage of Atp1b3 with markers on the proximal portion of mouse chromosome 7. Among the 27 strains examined, there were three crossovers between Atp1b3 and Gpi1 (27) and five recombination events between Atp1b3 and Tam1 (28). The maximum likelihood estimate places Atp1b3 at a distance of 6.2 centimorgans (cM) distal to Gpi1 (95% confidence interval 1.2-20.6 cM) and 12 cM proximal to Tam1 (95% confidence interval 3.5-32.4 cM) (29). DISCUSSION We have identified cDNAs and genomic sequences encoding a third mammalian member of the Na,K-ATPase ␤ subunit gene family. The genes for the Na,K-ATPase ␤1 and ␤2 subunits have previously been localized to mouse chromosomes 1 and 11, respectively (4, 5). Here we have mapped the ␤3 subunit gene to mouse chromosome 7. The fact that each ␤ subunit gene is located on a different mouse chromosome suggests that correction events such as gene conversion did not participate in the evolution of the ␤ subunit gene family.
The classification of Na,K-ATPase ␤ subunits as ␤2 or ␤3 has been difficult to resolve, primarily because in all species so far examined, only two ␤ subunit isoforms have been identified. The putative Xenopus ␤3 subunit (17) exhibits greater sequence divergence from the mammalian ␤2 subunit than do the mammalian and Xenopus ␤1 subunits. The same is true for the chicken non-␤1-like subunit, which was classified as ␤2 based on its tissue-specific expression pattern (26). The identification of a mammalian ␤3 subunit helps clarify the evolutionary relationships among ␤ subunit isoforms. Sequence comparisons among Na,K-ATPase and H,K-ATPase ␤ subunits show three distinct clusters containing Na,K-ATPase ␤1, ␤2, or ␤3 subunit isoforms, and a fourth representing the H,K-ATPase ␤ subunit. This type of analysis predicts that the chicken ␤2 subunit is actually a ␤3 subunit, as it clusters with the family of Na,K-ATPase ␤3 subunits. In this context, it should be noted that a recently cloned ␤ subunit from B. marinus bladder (30) clusters with Na,K-ATPase ␤2 subunit isoforms. It will clearly be interesting to determine whether a ␤2 subunit homolog can also be identified in Xenopus and avian species.  3. Hydropathy profiles of human ␤1, ␤2, and ␤3 subunits. Hydropathy values were obtained by using the algorithm and hydropathy values of Kyte and Doolittle (24). Hydrophobic regions are above the center line and hydrophilic regions are below.
FIG. 4. Sequence relatedness of Na,K-ATPase ␤ subunit isoforms. Dendrogram analysis of Na,K-ATPase amino acid sequences was obtained using the PileUp program of the GCG software package (20). The scale at the bottom indicates percent identity between ␤ subunit isoforms. The dotted line represents the partial rat ␤3 sequence obtained by reverse transcriptase-mediated PCR. The following sequences were obtained from the SwissProt or GenBank sequence data bases: rat ␤1 (atnb_rat), human ␤1 (atnb_human), chicken ␤1 (atnb-_chick), X. laevis ␤1 (U17061), B. marinus ␤1 (atnb_bufma), eel Anguilla anguilla ␤1 (X76109), rat HK␤ (athb_rat), chicken HK␤ (L08047), rat ␤2 (atnc_rat), mouse ␤2 (atnc_mouse), human ␤2 (atnc_human), B. marinus ␤ (bladder) (Z25812), chicken ␤2 (atnc_chick), X. laevis ␤3 (atnd_xenla), B. marinus ␤3 (atnd_bufma), zebrafish ␤ (X89722). Northern blot analysis indicates that ␤3 subunit mRNA sequences are expressed in a variety of rat tissues including testis, brain, kidney, spleen, stomach, small intestine, colon, lung, and liver. Highest expression levels were detected in testis, whereas substantially lower levels of ␤3 subunit transcripts were present in brain. The low level of ␤3 subunit expression in rat brain is somewhat surprising since the Xenopus ␤3 (17), Bufo ␤3 (25), and chicken ␤2 (the apparent avian ␤3 subunit homolog) (26) subunits appear to be expressed predominantly in neural tissue. In contrast, ␤2 subunits are abundant in mammalian brain (9,14), while ␤2 subunit homologs have not yet been detected in amphibian or avian brain. ␤1 subunits are also expressed in mammalian and non-mammalian brain (10,25). Thus, it is possible that in non-mammalian species, ␤1 and ␤3 are the predominant ␤ subunits of brain, whereas in mammalian species, ␤1 and ␤2 subunits are the predominant neural isoforms. In mammalian brain, ␤1 subunits are localized exclusively in neurons whereas ␤2 subunits are glia-specific (16). It will be interesting to determine whether ␤3 subunit expression patterns within brain exhibit regional or cellular differences among species. It is noteworthy that we detect ␤3 subunit mRNA in kidney, a tissue thought to express only ␣1 and ␤1 subunit isoforms (10). Our results raise the possibility that kidney may also contain Na,K-ATPase isoenzymes composed of ␣1 and ␤3 subunits. The development of ␤3 isoform-specific antisera will allow us to determine whether ␤3 subunit polypeptides are present in kidney and whether ␤3-containing isoenzymes exhibit differences in cellular distribution compared with ␤1-containing isoenzymes in this tissue. A further issue that should be raised in this context is the recent identification of a putative fourth Na,K-ATPase ␣ subunit isoform (31). This ␣ subunit isoform is expressed primarily in testis, raising the possibility that the predominant Na,K-ATPase of testis may be an ␣4/␤3-containing isoenzyme.
As a step toward our goal of elucidating the molecular evolution of the cation-motive transporter ␤ subunit gene family, we determined the chromosomal position of the mouse gene encoding the Na,K-ATPase ␤3 subunit. By analyzing the segregation of RFLPs in RI strains of mice, the ␤3 subunit gene (Atp1b3) was found to segregate with markers (Gpi1 and Tam1) localized to the proximal region of mouse chromosome 7. These markers constitute part of a highly conserved linkage group on mouse chromosome 7 and human chromosome 19. We therefore speculate that the human ␤3 subunit gene is likely to be located on human chromosome 19. In the mouse, the Na,K-ATPase ␤1, ␤2, and ␤3 subunits are encoded by separate genes located on three different chromosomes (4,5). Chromosomal dispersion and tissue-specific expression of the ␤ subunit genes suggests that each gene may encode a ␤ subunit with distinct functional properties.
Several lines of evidence suggest that Na,K-ATPase ␣/␤ subunit interaction is promiscuous and that six possible ␣/␤ subunit-containing isoenzymes are likely to be formed. Cellular localization studies have provided evidence for the existence of isoenzymes composed of ␣1/␤1 and ␣3/␤1 subunit combinations in neurons and ␣2/␤1 and ␣2/␤2 subunit combinations in astroglia (16). Isoenzymes composed of ␣3/␤2 subunit combinations have been detected in pineal gland (12) and photoreceptors (13), while ␣2/␤2-containing isoenzymes appear to be expressed in skeletal muscle (32). The identification of the mammalian Na,K-ATPase ␤3 subunit raises the possibility of nine potential ␣/␤ subunit-containing isoenzymes. It will now be important to determine whether different combinations of ␣ and ␤ subunits may be related to specific functional roles of the Na,K-ATPase in different tissues and cell types.  Green et al. (27), and Tam1 alleles were typed by Nesbitt et al. (28). The Tam1 type of strain BXA26 is unavailable. Crossing-over events are identified by an x. BXA strains   1  2  4  5  6  8  10  11  13  14  19  20  23  1  2  4  7  8  11  12  13  14  16  17  24  25  26 Gpi1