INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

CELLS CONTROL THE FIDELITY of secretory protein biosynthesis before transport through the Golgi complex, assuring that proteins in distal compartments of the secretory pathway are conformationally intact and targeted to the correct cellular compartment. To a major extent, cells achieve this by synthesizing proteins that form multimeric complexes and by permitting only fully assembled multimers to exit the endoplasmic reticulum (ER). In many cases, unassembled and ER-retained subunits are degraded rapidly by a pre-Golgi degradation system (for review see Ref. 6).

Our laboratory has studied the subunit assembly and transport from the ER to the plasma membrane of the heterodimeric - cation-transporting ATPases expressed in Xenopus laevis oocytes (7, 11, 14). The multi-membrane-spanning -subunit (relative mol mass ~100 kDa) is the catalytic subunit that hydrolyzes ATP and undergoes the E₁-to-E₂ conformational transition during ion translocation. The glycosylated -subunit (relative mol mass ~45-80 kDa) is required for the plasma membrane expression of functional pumps (for review see Ref. 6), and it may influence the apparent K affinity (12, 14). Well characterized in vertebrates are the Na-K-ATPase, a ubiquitous component of the plasma membrane, and the gastric H-K-ATPase, expressed only in the parietal cell of the stomach mucosa. The -subunits share ~65% amino acid sequence identity (17), whereas the -subunits are less well conserved (24).

In the parietal cell of the stomach, the H-K-ATPase resides in the apical membrane and subapical tubulovesicles (20), whereas the Na-K-ATPase is basolateral, as it is in most polarized epithelia (18). Gottardi and Caplan (9) have identified apical targeting domains in the - and -subunits of the gastric H-K-ATPase (HK and HK, respectively), suggesting a mechanism for the correct plasma membrane localization of the molecule. Arrival in the apical or basolateral membrane, however, must be preceded by appropriate subunit assembly in the ER. The available data indicate that assembly can be promiscuous, with Na-K-ATPase -subunits (NK) and gastric HK forming heterodimers and functional pumps in the plasma membrane (5, 10, 14, 16). Clearly, however, the parietal cell must extend greater control over the expression of mixed pumps.

To better understand these control mechanisms, we expressed gastric HK or HK in combination with NK or Na-K-ATPase -subunits (NK) in Xenopus oocytes and analyzed their ability to assemble and to support the expression of functional H-K or Na-K pumps at the cell surface. With this approach, we wanted to test whether Xenopus oocytes lack the parietal cell-specific control factors for correct assembly and transport of cation pumps. Our previous work has involved the functional expression of rabbit gastric HK in oocytes along with Xenopus NK (10) or Xenopus gastric HK (17). To have a strictly homologous Xenopus system, we isolated a cDNA encoding Xenopus HK and generated an antibody against this protein. Using these tools and the Xenopus oocyte expression system, we show that, like rabbit gastric HK, the Xenopus HK assembles with Xenopus HK as well as NK in the ER. However, - heterodimers formed with Xenopus HK were not expressed at the cell surface due to early degradation. This may suggest a mechanism by which the parietal cell can assure that only appropriate - heterodimers are routed to the trans-Golgi network for trafficking to either the apical or basolateral membranes. Additionally, this presents a novel paradigm for the ER exit of a multisubunit protein. Unassembled HK is capable of rapidly leaving the ER, whereas -assembled HK is not transported to the cell surface.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Cloning of a cDNA encoding the Xenopus laevis gastric HK. A cDNA fragment of Xenopus HK was generated by PCR, using as template cDNA prepared from Xenopus stomach mRNA with oligo(dT) priming. The degenerate oligonucleotide primers used in the PCR were a sense primer encoding the amino acid sequence Pro-Asp-Tyr-Gln-Asp-Gln with an added EcoR I site at the 5' end (CGGAATTCCNGAYTAYCARGAYCA) and an antisense primer encoding the amino acid sequence His-Tyr-Phe-Pro-Tyr-Tyr with an added BamH I site at the 5' end (CCGGATCCRTARTANGGRAARTARTG). A 480-bp fragment was isolated, subcloned, sequenced, and found to encode the appropriate region of Xenopus HK.

Xenopus HK and HK antibodies. A glutathione S-transferase (GST) fusion protein containing the carboxy-terminal 100 amino acids of Xenopus HK (Fig. 1) was constructed in the vector pGEX-2T (Pharmacia). A BamH I site, in frame with the GST open reading frame, was introduced by PCR into the Xenopus HK cDNA before the codon for Pro-195. Bacterially produced fusion proteins were affinity-purified using glutathione-Sepharose 4B (Pharmacia), recovered by elution with glutathione, and used to immunize rabbits. The specificity of the antiserum was tested by immunoprecipitation of digitonin extracts of metabolically labeled Xenopus HK following cRNA injection into oocytes (see below). Preimmune serum and immune serum preabsorbed on the GST-Xenopus HK fusion protein used as antigen did not immunoprecipitate any labeled proteins (data not shown). On the other hand, both immune serum preabsorbed on GST and immune serum without treatment immunoprecipitated a core-glycosylated ER form (~50 kDa) and a fully glycosylated post-Golgi form (60-75 kDa) of the Xenopus HK.

View larger version (62K): [in this window] [in a new window]   Fig. 1.   A: primary structure and tissue distribution of Xenopus gastric H-K-ATPase -subunit (HK). Isolated cDNA coding for Xenopus HK is ~1.1 kb long and contains a 295-codon open reading frame with start codon beginning at nucleotide 22. Shown is an alignment of amino acid sequences of Xenopus laevis (X) and rabbit (rab) gastric HK. Identical amino acids are indicated by dashes and conservative substitutions by dots. Single transmembrane domain is underlined in bold. The 7 putative glycosylation sites are boxed. Two additional, less conservative consensus glycosylation sequences NDT and NPT are also present. A glutathione S-transferase fusion protein that was used to generate antibodies against HK contained the carboxy-terminal 100 amino acids (underlined). B: Northern blot analysis of tissue distribution of Xenopus HK. Total RNA (7.5 µg) from different tissues of adult X. laevis was hybridized with cDNA probes. Top: RNA hybridized with a Xenopus gastric -subunit probe. Bottom: same filter reprobed with a Xenopus elongation factor 1 (EF1) probe. mRNA encoding EF1 is a major transcript at midblastula transition in Xenopus (15). Lanes are labeled with source of RNA; st 14/15 RNA is from stage 14-15 embryos. Migration of 18S RNA is indicated at right.

Protein expression in Xenopus oocytes and immunoprecipitations. Oocytes were obtained from X. laevis as previously described (7). Oocytes were injected with the indicated amounts of -subunit cRNA alone or in combination with -subunit cRNA, metabolically labeled in modified Barth's solution (MBS) containing 0.6 µCi/ml [³⁵S]methionine, and chased in MBS containing 10 mM unlabeled methionine. Digitonin extracts were prepared, and immunoprecipitations were performed under denaturing or nondenaturing conditions as described (7). In some instances, immunoprecipitated -subunits were digested with endoglycosidase H (Calbiochem-Novabiochem, La Jolla, CA) (13). SDS-PAGE, fluorography, and laser densitometry were performed as previously described (7).

⁸⁶Rb uptake. Three days after cRNA injection of oocytes, ⁸⁶Rb uptake was measured as previously described (13). The assay solution used throughout was (in mM) 90 NaCl, 1 MgCl₂, 0.33 Ca(NO₃)₂, 0.41 CaCl₂, 5 BaCl₂, and 10 HEPES (pH 7.4). For H-K-ATPase transport measurements, oocytes were preincubated for 15 min in an assay solution containing 10 µM ouabain, which completely inhibits the endogenous oocyte Na-K-ATPase but has no effect on H-K-ATPase activity (17). After preincubation, oocytes were incubated for 12 min in an assay solution containing 5 µCi/ml ⁸⁶RbCl (Amersham) and 0.5 mM KCl for H-K-ATPase transport measurements or 5 mM KCl for Na-K-ATPase transport measurements. Oocytes were washed, and the ⁸⁶Rb uptake in single oocytes was determined by scintillation counting.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Primary structure and tissue distribution of the Xenopus gastric HK. We characterized recently a Xenopus gastric HK after coexpression with rabbit gastric HK in Xenopus oocytes (17). To test whether functional H-K pumps could be formed in a strictly homologous system different from parietal cells, we isolated a cDNA encoding Xenopus gastric HK (GenBank accession no. BankIt165014 AF042812). Xenopus HK shares 55% amino acid sequence identity with rabbit HK, with two striking differences (Fig. 1A). One area of difference is at the carboxy terminus, where the Xenopus HK extends eight amino acid beyond that of the rabbit HK. A second difference is in the cytoplasmic amino terminus, where a tyrosine-based endocytosis motif has been identified in mammalian HK (4). In Xenopus and chicken HK, however, phenylalanine replaces this tyrosine.

Xenopus HK does not support the functional expression of HK or NK. To test the ability of Xenopus HK to produce functional H-K-ATPase - complexes in Xenopus oocytes, we expressed Xenopus HK or rabbit HK together with Xenopus or rabbit HK and compared the expression of functional H-K pumps at the cell surface by ⁸⁶Rb uptake measurements. As previously shown, coexpression of rabbit HK with rabbit (10) or Xenopus (17) HK led to a significant, approximately fivefold increase in ⁸⁶Rb uptake compared with that measured in oocytes expressing HK alone (Fig. 2A). Surprisingly, coexpression of Xenopus HK with Xenopus or rabbit HK did not result in a significant change in the H-K pump activity compared with that in HK-expressing oocytes. Similarly, Xenopus HK coexpressed with Xenopus NK did not increase Na-K pump activity at the cell surface compared with that found in oocytes expressing Xenopus HK alone (Fig. 2B).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Expression of Xenopus gastric HK in Xenopus oocytes does not produce functional H-K or Na-K pumps at cell surface. Oocytes were injected with indicated combinations of cRNAs for different -subunits of N-K-ATPase (NK; 11 ng) or H-K-ATPase (HK; 11 ng) and HK or -subunit of N-K-ATPase (NK) [2 ng rabbit (r) HK, 0.5 ng Xenopus HK, or Xenopus NK]. Three days after injection, 86Rb uptake into oocytes was measured as described in METHODS. A: 86Rb uptake mediated by expressed H-K pumps. Shown is mean of 2 experiments performed on 2 different batches of oocytes. In each experiment, 16-20 oocytes were measured for each condition. Values obtained in oocytes expressing rabbit HK-rabbit HK complexes were set to 1 and amounted to 14.3 ± 0.5 and 12.8 ± 0.4 pmol · min1 · oocyte1, respectively, in 2 experiments. B: 86Rb uptake mediated by expressed Na-K pumps. Values obtained in oocytes expressing Xenopus NK-Xenopus NK complexes were set to 1 and amounted to 129.2 ± 15.5 pmol · min1 · oocyte1. Data are means ± SE (n = 20).

Xenopus HK can associate with but not stabilize HK or NK. The lack of an increased pump expression after coexpression of Xenopus HK with either HK or NK could be an indication that Xenopus oocytes lack mechanisms for the sorting of H-K-ATPases vs. Na-K-ATPases that parietal cells need to express H-K pumps in the correct cellular compartment. However, it is also possible that the results obtained were only due to inefficient translation of the injected cRNA or to lack of assembly of the newly synthesized Xenopus HK with the -subunits. To test the two latter possibilities, we compared the biosynthesis and assembly of metabolically labeled Xenopus or rabbit HK expressed alone or together with HK or NK in oocytes. After cRNA injection, oocytes were subjected to a 16-h pulse with [³⁵S]methionine and various chase periods, digitonin extracts were prepared, and the expressed proteins were immunoprecipitated under nondenaturing conditions that preserve subunit interaction. As previously observed (10), rabbit HK was well expressed in oocytes and when expressed alone was immunoprecipitated mainly in its core glycosylated form after a pulse period (Fig. 3A, lane 1) and in its fully glycosylated form after various chase periods (lanes 2 and 3). Although rabbit HK could associate with endogenous oocyte NK, as reflected by coimmunoprecipitation with a HK antibody (lanes 1-3), it was synthesized in large excess over the endogenous -subunit. Therefore our data confirm that rabbit HK is able to be transported to the plasma membrane without association with -subunits. Rabbit HK associated efficiently with coexpressed rabbit (lane 10) and Xenopus (lane 4) HK or NK (lane 7) and typically stabilized the HK (lanes 5, 6, 11, and 12) and to a somewhat lesser extent the NK (lanes 8 and 9), which in an unassembled form are degraded completely during the chase period (1). This result reflects the higher specificity of HK for HK than for NK and explains the previously observed difference in the cell surface expression of functional HK-HK and NK-HK complexes (14).

View larger version (89K): [in this window] [in a new window]   Fig. 3.   Xenopus HK assembles with but does not stabilize HK and NK. Oocytes were injected with HK (0.25 ng Xenopus HK, 2.0 ng rabbit HK) cRNA alone or in combination with HK (8 ng) or NK (7 ng) as indicated. After metabolic labeling for 16 h and chase periods of 0, 24, or 48 h, digitonin extracts were prepared and subjected to immunoprecipitation under nondenaturing conditions with an anti-rabbit HK monoclonal antibody (A) or a Xenopus HK antiserum (B). Core-glycosylated (cg) and fully glycosylated (fg) HK and NK, coprecipitated exogenous -subunit (), and endogenous -subunit ( endo) are indicated.

View larger version (40K): [in this window] [in a new window]   Fig. 4.   Carbohydrate processing of Xenopus HK expressed alone or with HK. Oocytes were injected with Xenopus HK cRNA (0.25 ng) alone or in combination with HK cRNA (8 ng), metabolically labeled for 16 h, and chased for indicated times. Digitonin extracts were denatured and immunoprecipitated with Xenopus HK antiserum. One-half of eluted proteins were digested with endoglycosidase H (Endo H) before SDS-PAGE. Nonglycosylated (ng), core-glycosylated, and fully glycosylated forms of Xenopus HK are indicated.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Assembly and functional maturation of endogenous -subunit with Xenopus HK. Oocytes were injected or not with Xenopus HK (0.25 ng) or rabbit HK (2 ng). A: immunoprecipitation of endogenous -subunits. Oocytes were labeled for 16 h and chased for 40 h. Extracts were denatured and immunoprecipitated with NK antiserum. B: 86Rb uptake. After cRNA injection, oocytes were incubated for 3 days before 86Rb uptake measurements. Data are means ± SE (n = 20). * P < 0.5 vs. noninjected control oocytes.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In this study, we document an unusual characteristic of Xenopus HK expressed in Xenopus oocytes that may be indicative of specific mechanisms governing the selective formation or targeting of H-K-ATPase and Na-K-ATPase - complexes in the stomach.

Our data show that Xenopus HK expressed in Xenopus oocytes can indeed associate with rabbit or Xenopus HK and NK but does not permit the formation of functional pumps at the cell surface due to rapid degradation of the - complexes. The Xenopus HK behaves in this respect differently from all other -subunits so far tested in the Xenopus oocyte system, including Xenopus, rat, and human NK and Bufo nongastric HK and rabbit gastric HK. Although heterologous assembly between NK and HK or HK and NK was found to be less efficient than homologous assembly, -assembly was always accompanied by a complete or at least partial stabilization of the coexpressed -subunit and a corresponding increase in the number of functional pumps at the cell surface (for review see Ref. 6).

A possible explanation for the observation that rabbit as well as Xenopus HK expressed in Xenopus oocytes can be stabilized and form functional H-K pumps with rabbit HK but not with Xenopus HK could involve the presence of a molecular signal in the Xenopus HK, which might be important for a tissue-specific control of the stable formation or targeting of H-K-ATPase - complexes in the Xenopus stomach but which cannot be interpreted correctly in the Xenopus oocyte expression system. At first sight, two domains in the Xenopus HK could be of interest in this respect. Analysis of the carboxy termini of NK and HK has revealed that the last 10 amino acids may form an amphipathic -strand that exposes on one side a hydrophilic domain and on the other side a continuous hydrophobic domain that is important for subunit assembly (2). Xenopus HK differs from all other known HK as well as from NK by an extension of eight amino acids, which could be necessary for a tissue-specific control of subunit assembly.

A second domain in the Xenopus HK that could be involved in the particular behavior of this protein when expressed in Xenopus oocytes, is the cytoplasmic amino terminus. All HK so far identified, with the exception of chicken and Xenopus HK, contain a tyrosine-containing sequence in the amino terminus that is a reversed version of the motif that is responsible for transferrin receptor internalization (3). Recently, it was shown in transgenic mice that the Phe-Arg-His-Tyr motif in the mammalian HK is necessary for endocytosis of H-K pumps and termination of acid secretion in the stomach (4). In addition, tyrosine-based motifs have been shown to be responsible for targeting to various endosomal compartments or lysosomes (for review see Ref. 22).

It is not known whether the signals that are involved in these processes are similar in amphibia or birds and in mammals, but it could be that the corresponding Phe-Arg-Arg-Phe or the Phe-Gly-Arg-Phe sequences in chicken or Xenopus HK have functions similar to those of the tyrosine-based motifs in mammals. If this is the case, the signal present in Xenopus gastric HK might specifically mediate targeting to the tubulovesicular structure in the Xenopus stomach cells. Due to the lack of these structures in Xenopus oocytes, the newly synthesized Xenopus H-K-ATPase - complexes might be sorted to lysosomes or another degradation compartment.

Finally, it is interesting to note that there is degradation only of assembled - complexes but not of Xenopus HK expressed alone in Xenopus oocytes. This result indicates that a putative targeting signal present on Xenopus HK is exposed only after the -subunit associates with the -subunit, due either to a conformational change or to targeting to a particular ER subcompartment that is involved in the recognition of the signal.