Cloning of the H,K-ATPase ,f3 Subunit TISSUE-SPECIFIC EXPRESSION, CHROMOSOMAL ASSIGNMENT, AND RELATIONSHIP TO Na,K-ATPase p SUBUNITS*

We have isolated cDNA clones encoding the bovine and rat gastric H,K-ATPase B subunit. A bovine abom- asum Xgt 11 cDNA library was screened with a mono- clonal antibody raised against the rabbit H,K-ATPase 0 subunit. A single positive phage clone containing an -900-base pair cDNA insert was identified as reactive with the antibody. The identity of the cDNA was estab- by comparing the deduced amino acid sequence with sequences of cyanogen bromide of the porcine H,K-ATPase @

We have isolated cDNA clones encoding the bovine and rat gastric H,K-ATPase B subunit. A bovine abomasum Xgt 11 cDNA library was screened with a monoclonal antibody raised against the rabbit H,K-ATPase 0 subunit.
A single positive phage clone containing an -900-base pair cDNA insert was identified as reactive with the antibody. driving force for HCI secretion into the stomach (1). H,K-ATPase couples the electroneutral exchange of extracellular K' and intracellular H' to the hydrolysis of ATP (2), thereby generating an extremely high transmembrane proton gradient (3). The enzyme has been shown to consist of a major polypeptide component of M, -100,000 (4) that contains the catalytic site for ATP hydrolysis (5). cDNA clones encoding the catalytic subunit of the rat and hog stomach H,K,-ATPase have been isolated and characterized (6,7). The deduced amino acid sequence of the H,K-ATPase shares a number of structural similarities with other cation transport ATPases, including the Na,K-ATPase (Y subunit (8,9), the Ca'+-ATPase (lo), and the H+-ATPase of yeast and Neurosporu (11, 12).
These similarities suggest that P-type ion transport proteins arose from a common evolutionary ancestor. We have recently characterized an abundant microsomal glycoprotein (gp 60-80) which is found in H,K-ATPase-containing membranes of several animal species (13). A number of lines of evidence suggest that this glycoprotein may represent a @like subunit of the gastric H,K-ATPase.
First., lectin affinity chromatography and immunoprecipitation reveal that gp 60-80 is noncovalently associated with the H,K-ATPase and remains associated with the catalytic subunit even after exposure to low concentrations of SDS. Second, gp 60-80 biosynthesis and H,K-ATPase activity appear concomitantly with the development of HCl secretion in the frog (14). Third, gp 60-80 shares several structural features with Na,K-ATPase /3 subunits including (a) similarity in size of core proteins, (6) association with the cell membrane, (c) the presence of several N-linked glycosylation sites, and (d) a stoichiometry of 1:l with the catalytic subunit (13).
Here we describe the isolation and characterization of cDNA clones encoding bovine and rat stomach gp 60-80 utilizing two complementary approaches. First, a monoclonal antibody raised against rabbit gp 60-80 was used to screen a bovine abomasum Xgtll cDNA library. We identified an antibody-reactive recombinant phage clone containing an -900 base pair cDNA insert which encodes a portion of bovine gp 60-80. Second, primers based upon sequences obtained from the bovine cDNA and cyanogen bromide fragments of porcine gp 60-80 were used to amplify a fragment of rat stomach gp 60-80 cDNA using the rapid amplification of cDNA ends (RACE) procedure (15). The PCR-generated and bovine cDNAs were then used to screen a rat stomach cDNA library, and several full-length cDNA clones were isolated and characterized.
Analysis of the amino acid sequence deduced from rat gp 60-80 cDNA shows a striking degree of primary sequence and secondary structure similarity to Na,K-ATPase p subunit isoforms. These results suggest that gp 60-80 cDNA encodes a p subunit of the gastric H,K-ATPase. The availability of a molecular probe for the H,K-ATPase /3 subunit should facilitate an understanding of the role of this polypeptide in H,K-ATPase biogenesis and enzyme function.  Table II.

Isolation and Characterization
of gp 60-80 cDNA Clones-A bovine abomasum Xgtll cDNA library was screened with a monoclonal antibody (16) raised against rabbit gp 60-80. One clone was identified as immunoreactive through three successive rounds of antibody screening and plaque purification. As shown in Table I, the deduced amino acid sequence of a region of the bovine cDNA corresponds very well to the sequence of a CNBr fragment (Peptide A, Table I) derived from deglycosylated porcine gp 60-80. Identity was observed at 8 of 9 positions. These results provide direct evidence that the bovine cDNA encodes a portion of gp 60-80. A fragment of the bovine gp 60-80 cDNA insert was used as a probe to rescreen the library and two additional clones were isolated. However, these clones failed to extend the cDNA in the 5' direction. Furthermore, PCR amplification of the bovine cDNA library failed to generate a 5' extension product (data not shown). Taken together, these results suggest that the 5' end of gp 60-80 cDNA is unlikely to be represented in the library.
We next used PCR and RACE (15) to amplify gp 60-80 cDNAs from rat stomach mRNA according to the strategy schematized in Fig. 1. To carry out 3' RACE, a fully degenerate 5' primer oligonucleotide (sense primer) was synthesized that encodes a portion of the amino acid sequence of porcine gp 60-80 CNBr peptide B (Table I). The 3' (antisense) primer was a 35 base oligonucleotide composed of a stretch of 17 dT residues and an adaptor sequence (15). Total RNA was isolated from rat stomach, and first strand cDNA was synthesized using the 3' oligonucleotide as primer. The cDNA fragment spanning the primers was then amplified by PCR using the sense and adaptor primers (15), and the first strand rat stomach cDNA as template according to Fig. 1. The PCR products were analyzed by Southern blotting using radiolabeled bovine gp 60-80 cDNA, and a DNA band of the expected size (-1.15 kb) was identified as reactive with the probe. The amino acid sequence deduced from this PCR-generated cDNA agreed very well with the sequence predicted from bovine gp 60-80 cDNA (data not shown) and contained regions which were strikingly similar to the sequences of porcine gp 60-80 CNBr peptides A and B and V8 proteolytic fragments C and D (27), as shown in Table I. Portions of the 3' RACE and bovine gp 60-80 cDNAs were then used to screen a rat stomach XZAP cDNA library. cDNA clones were selected on the basis of positive hybridization to both probes. Of twelve clones isolated, five contained the complete open reading frame for rat stomach gp 60-80. No sequence differences were observed of the Gastric H,K-ATPase p Subunit between the PCR-generated cDNA and cDNA clones isolated from the rat stomach library.
Structure of Rat Stomach gp 60-BO-The complete nucleotide sequence of rat gp 60-80 cDNA and the deduced amino acid sequence of the predicted protein are shown in Fig. 2. The open reading frame defines a protein of 294 amino acids with a molecular weight of 33,689. The predicted protein starts at nucleotide position 176 and terminates at nucleotide position 1057, followed by 426 nucleotides of 3'-untranslated sequence and a poly(A) tail. The hydropathy profile (Fig. 3) indicates that rat gp 60-80 contains a polar cytoplasmic amino terminus followed by a single hydrophobic transmembrane domain of 27 amino acids and a 228-residue-long extracellular carboxyl-terminal domain including seven potential N-linked glycosylation sites (asterisks). gp 60-80 is Related to Na,K-ATPase p Subunit Isoforms-The deduced amino acid sequence of rat gp 60-80 was compared with all sequences currently in the National Biomedical Research Foundation (NBRF) database (Release 24.0). Interestingly, rat gp 60-80 showed significant similarity only to Na,K-ATPase p subunit isoforms. A comparison of the amino acid sequence of rat gp 60-80 with the rat Na,K-ATPase @l (28) and /32 (29) subunits is sL., In in Fig. 4 whereas the rat 82 and /31 subunit consist of 290 and 304 amino acid residues, respectively. There are 6 cysteine residues (positions 131, 152, 162, 178, 201, and 266) within the presumed extracellular domain of gp 60-80. In the computer aligned sequences, these cysteine residues appear to be highly conserved among gp 60-80 and the Na,K-ATPase /31 and p2 subunits. The asparagine residues marked with asterisks ( Fig.   2) represent potential sites of N-linked glycosylation. There are seven such sites in gp 60-80, seven in the rat p2 subunit, and three in the rat /31 subunit polypeptide. One of the predicted N-linked glycosylation sites found in gp 60-80 (POsition 161) is exactly conserved relative to one of the predicted N-linked glycosylation sites found in the rat /31 and p2 subunits, whereas a second (position 193) is located 3 residues from a predicted N-linked glycosylation site which is conserved between the pl and p2 subunits. A third asparagine residue (position 255) found in gp 60-80 is also exactly conserved relative to the position of a predicted N-linked glycosylation site found in the /31 and 82 subunits. However, this asparagine residue is not contained within a consensus glycosylation sequence (19). A putative transmembrane segment is located between residues 40 and 66 in gp 60-80 and 40 and 67 in the Na,K-ATPase p2 subunit. Of the 27 amino acid residues compared in this region, 11 are identical and 9 are conservative substitutions (30) between gp 60-80 and the 82 subunit. When analyzed by the method of Chou and Fasman (31), the predicted secondary structures of gp 60-80 and the Na,K-ATPase pl and p2 subunits appear to be virtually identical. Taken together, these results indicate that rat gp 60-80 is a polypeptide related to the Na,K-ATPase /3 subunits. We have therefore termed this polypeptide the H,K-ATPase p subunit.
The H,K-ATPase p Subunit Is a Glycosylated Polypeptide Expressed Exclusively in Stomach-A panel of rat tissues was examined for the presence of H,K-ATPase fi subunit mRNA. The pattern of expression of /3 subunit mRNA is shown in Fig. 5 (@ panel). Of the rat tissues analyzed (brain, heart, lung, kidney, liver, spleen, and stomach), @ subunit mRNA was detected only in stomach. The H,K-ATPase p subunit gene encodes two transcripts, -1.7 and -3.7 kb in size. The 1.7-kb mRNA is the predominant species and is -20-fold more abundant than the 3.7-kb transcript. When the blot was reprobed with a cDNA specific for a portion of the H,K-ATPase catalytic (a) subunit (Fig. 5, a!  species -4.0 and -4.2 kb in size were detected which appeared to be expressed exclusively in stomach. These results indicate that the H,K-ATPase catalytic ((Y) and /3 subunit genes exhibit an identical tissue-specific pattern of expression.
To determine the tissue distribution of H,K-ATPase /3 subunit polypeptides, we probed Western blots of rabbit cellular and rat microsomal membrane fractions with the antigp 60-80 monoclonal antibody (16). As shown in Fig. 6A, the antibody reacted with a broad band of -60 to -80 kDa in rabbit fundus. In contrast, H,K-ATPase fl subunits were undetectable in a variety of other rabbit tissues including duodenum, intestine, proximal and distal colon, liver, pancreas, and brain. We also used the monoclonal antibody to probe a Western blot of crude rat microsomes (Fig. 6B). A broad brand of 60-80 kDa was detected in rat stomach, whereas the antibody failed to show immunoreactivity with kidney, brain, and heart microsomes. These results suggest that H,K-ATPase fi subunits are expressed exclusively in stomach of at least two animal species.
We next analyzed the species distribution of the H,K-ATPase p subunit. A Western blot of microsomes prepared from hog, cow, rabbit, rat, and mouse stomach was probed with the anti-gp 60-80 monoclonal antibody. The results are shown in Fig. 7. In each species, the mature form of the @ subunit migrated as a broad band with an apparent molecular weight of 60,000-80,000. Digestion of the /3 subunit with Nglycanase F revealed that in each species, the /3 subunit core protein migrated with an apparent molecular weight of 32,000. This is very close to the value of 33,689 for the rat fl subunit predicted from cDNA cloning. (In the cow, rabbit, and mouse lanes, several intermediate bands were generated, indicating that the enzyme did not digest to completion.) These data suggest that the H,K-ATPase /3 subunit is expressed in stomach of a wide variety of mammals.

Chromosomal
Localization of the H,K-ATPase p Subunit Gene-We have used segregation of restriction fragment length polymorphisms among RI strains of mice to identify the chromosomal location of the mouse gene encoding the p subunit of the gastric H,K-ATPase (Atp4b).' Mouse genomic DNA sequences were identified by hybridizing Southern blots to radiolabeled 3' RACE-generated rat H,K-ATPase fl subunit cDNA. As shown in Fig. 8  Strain-specific alleles are abbreviated as AK (AKR/J) and L (C57L/J). Strain-specific alleles for mouse chromosome 8 markers  (32) and  (32) were published previously. The Xmu-26 type of strain AKXL-5 is not available. Crossing over events are identified by an x. LOCUS AKXL strains 5 6 7  8  9  12  13  14  16  17  19  21  24  25  28  29  37  38 Atp46 . AK AK L AK port protein subunits. The positions of the 6 cysteine residues in the extracellular portion of the rat H,K-ATPase /3 subunit are highly conserved relative to the positions of 6 cysteine residues in the rat Na,K-ATPase bl and 62 subunits. These residues could play an important role in maintaining fi subunit structure, possibly forming several folded subdomains each cross-linked by disulfide bonds (36). The H,K-ATPase @ subunit, like the Na,K-ATPase pl and /?2 subunits (25), is a glycosylated polypeptide.
Two of the predicted N-linked glycosylation sites within the rat H,K-ATPase /3 subunit are highly conserved relative to the positions of two potential Nlinked glycosylation sites in the rat Na,K-ATPase Pl(28) and fi (29) subunits. These structural similarities strongly indicate that the H,K-ATPase and Na,K-ATPase /3 subunits evolved from a common ancestral gene.
Chromosomal mapping experiments have demonstrated that the Na,K-ATPase /31 and 02 subunits are encoded by separate genes located on murine chromosomes 1 and 11, respectively (26,37). We have now localized the H,K-ATPase @ subunit gene to mouse chromosome 8. The identification of separate chromosomal loci for each fi subunit gene suggests that there are likely to be cis-acting control elements that determine the tissue-specific pattern of expression for each gene. In addition, the fact that each p subunit gene is located on a different chromosome suggests that correction mechanisms such as gene conversion did not participate in the evolution of the /3 subunit gene family, Our data indicate that the H,K-ATPase @ subunit is expressed in the stomach of a wide variety of mammalian species including rat, rabbit, hog, cow, and mouse. The polypeptide has also been detected in purified frog gastric microsomes (14), suggesting its potential existence in all vertebrates capable of gastric HCl secretion. Hybridization analysis reveals that the H,K-ATPase p subunit gene is expressed exclusively in rat stomach, and encodes two distinct mRNA transcripts (Fig. 5). All of the /3 subunit cDNA clones we isolated appear to represent a single class of cDNA, and N-glycanase F digestion of rat stomach microsomes (Fig. 7) indicates only one form of p subunit core polypeptide.
These results suggest that the two B subunit mRNA species are likely to differ in untranslated regions and encode only one form of B subunit polypeptide.
Interestingly, the H,K-ATPase (Y subunit gene also appears to encode two mRNA transcripts.
We do not yet know whether these (Y and p subunit mRNAs arise by differential splicing, utilization of alternative polyadenylation signals or whether the larger @ subunit mRNA represents an unprocessed nuclear precursor. Identification of the gene encoding the p subunit of the gastric H,K-ATPase raises important questions regarding the functional significance of /3 subunit isoforms. On one hand, the existence of a /3 subunit for the H,K-ATPase argues in favor of a common function for H,K-ATPase and Na,K-ATPase p subunits. For example, /3 subunits of the H,K-and Na,K-ATPases may play similar roles in cation transport, ATPase activity, and/or proper membrane orientation of their respective holoenzymes (36). Alternatively, chromosomal dispersion, amino acid sequence divergence, and tissue-specific expression of the @ subunit genes suggest that the polypeptide encoded by each gene may have properties selected in response to different physiological demands. Consideration of the biology of the oxyntic cell may provide some insight into this question. In the oxyntic cell, H,K-ATPase is sorted to the apical tubulovesicular membrane compartment, whereas the Na,K-ATPase is located in the basolateral membrane (reviewed in Ref. 38). Thus it is possible that determinants involved in targeting H,K-ATPase and Na,K-ATPase to their proper membrane destinations are sequences residing within their respective /3 subunits. A genetic approach could prove very powerful in studying a phenomenon of this type. The fact that both H,K-ATPase and Na,K-ATPase are expressed in the oxyntic cell raises an additional point of interest. Are there control mechanisms which govern the specific association of H,K-ATPase and Na,K-ATPase p subunits with their respective catalytic (Y subunits? Distribution of H,K-ATPase and Na,K-ATPase activities to separate membrane compartments (38) is consistent with the view that the H,K-ATPase /3 subunit does not associate with the Na,K-ATPase (Y subunit (and vice versa). It will clearly be of interest to determine how the sequence of each /3 subunit leads to specificity regarding interaction with the H,K-or Na,K-ATPase cx subunit. The construction of chimeric cDNA molecules between H,K-ATPase and Na,K-ATPase @ subunit cDNAs should permit identification of sites within a given p subunit that interact with the corresponding cy subunit. This strategy should also be useful for identification of other functional domains within H,K-and Na,K-ATPase p subunits.