Molecular Characterization Of V-Type H⁺-ATPase (B-Subunit) In Gills Of Euryhaline Crabs And Its Physiological Role In Osmoregulatory Ion Uptake

Together with P-ATPase and F-ATP synthase, the proton-translocating vacuolar-type ATPase (V-ATPase) is one of three major classes of the so-called ion-motive ATPases in eukaryotic cells (Forgac, 1989; Harvey, 1992; Wieczorek et al., 1999). The V-ATPase is composed of two structural protein domains, the cytoplasmic V₁ domain and the membrane-spanning V₀ domain (Nelson, 1992). The V₁ domain, with a molecular mass of approximately 570 kDa, comprises the catalytic sector and is composed of at least eight different subunits (A, B, C, D, E, F, G, H) with subunits A (67–73 kDa) and B (55–60 kDa) as the catalytic and regulatory subunits, respectively. The stoichiometry found in the V₁ complex of yeast is A₃B₃C₁D₁E₁F₁G₃H₁ (Forgac, 2000). The membrane-embedded V₀ domain, with a molecular mass of approximately 250 kDa, contains the proton-translocating structure. It is composed of at least four different subunits (M100, M40, M9.7, M17) with a probable stoichiometry of M100₁M40₁M9.7₁M17₆ (Wieczorek et al., 2000).

The diversity of cellular functions of the V-ATPase is attributed to its properties of generating both electrochemical and pH gradients across endo- and exomembranes of eukaryotes by utilizing ATP to move protons unidirectionally against steep H⁺ gradients. V-ATPases are involved in acidifying endosomes, lysosomes and other compartments of the cellular vacuolar system, where they play a role in the complex mechanisms of endo- and exocytosis. In proton-excreting cells such as osteoclasts or renal intercalated cells, the V-ATPase is distributed in high densities in the apical membrane of these polarized cells (Nelson et al., 1992). In midgut goblet cells of larvae of the lepidopteran Manduca sexta, apically situated V-ATPase energizes transepithelial secondary transport processes, including amino acid absorption, by producing a high luminal pH (Wieczorek et al., 1999).

In freshwater-inhabiting species such as trout (Lin and Randall, 1993), toad (Larsen et al., 1992) and crayfish (Zare and Greenaway, 1998), the activity of V-ATPase is thought to complement Na⁺/K⁺-ATPase in energizing osmoregulatory ion uptake from highly diluted media. Such a role of gill V-ATPase has been explored in previous investigations employing the extremely euryhaline Chinese mitten crab Eriocheir sinensis. Tracer-flux experiments on isolated perfused gills (Gilles et al., 1988; Riestenpatt et al., 1995), electrophysiological studies on split gill lamellae (Onken and Putzenlechner, 1995) and immunohistochemical experiments (Putzenlechner, 1994) suggest that the activity of an apically situated V-ATPase is necessary to regulate internal hemolymph osmolarity during exposure of E. sinensis to low external salinity.

In this study, we investigated whether the V-ATPase may also be involved in ion uptake in organisms with moderate osmoregulatory abilities such as the shore crab Carcinus maenas, in which the lower tolerance of external salinity is limited to 8 ‰ (Siebers et al., 1982). Weihrauch and Towle (1997) found the first molecular evidence for the expression of a V-ATPase in the gills of the shore crab by cloning a partial mRNA sequence of the B-subunit. To explore in more detail the presence and function of the V-ATPase in C. maenas, E. sinensis and other euryhaline crab species, molecular biological techniques, immunohistochemical approaches and experiments employing isolated perfused gills were performed. We present direct molecular evidence for the expression of a V-ATPase (B-subunit) in the gills of Carcinus maenas, Eriocheir sinensis, Callinectes sapidus, Chasmagnathus granulatus and Cancer irroratus. Both immunolocalization of the B-subunit in Carcinus maenas gills and mRNA expression analysis of this subunit in Carcinus maenas and Eriocheir sinensis support the suggestion that a V-ATPase is involved in osmoregulatory ion uptake in the extremely euryhaline Chinese crab but not in the moderately euryhaline shore crab.

For experiments conducted in Maine or Illinois, green shore crabs Carcinus maenas (L) were collected from the intertidal zone at Oak Point, Trenton, Maine, or were supplied by the Marine Biological Laboratory (Woods Hole, MA, USA). For experiments conducted in Germany, Carcinus maenas were obtained from Kiel Bay, Baltic Sea. Carcinus maenas were maintained in recirculating aquaria at 15–16 °C in 35 or 10 ‰ salinity and were fed cleaned squid or bovine heart 2–3 times weekly. Chinese mitten crabs Eriocheir sinensis (H. Milne Edwards) were obtained from the Eider River (approximately 0.5 ‰ salinity) in Germany and were maintained at 16 °C in running fresh water or in recirculating aquaria at 35 or 20 ‰ salinity. The gills were removed from E. sinensis and stored in RNAlater (Ambion) for transport to Illinois. Blue crabs Callinectes sapidus (Rathbun) were obtained from the Waukegan (IL, USA) Oriental Food Store or from Gulf Specimen (Panacea, FL, USA) and were kept in recirculating aquaria at 20 °C in 35 or 5 ‰ salinity. South American estuarine crabs Chasmagnathus granulatus (Dana) were obtained from the coast of Argentina, through the University of Buenos Aires, and were kept in small aquaria at 20 °C in 2–3 ‰ salinity. Rock crabs Cancer irroratus (Say) were obtained by divers at the Mount Desert Island Biological Laboratory and were kept briefly in flowing natural sea water at approximately 12 °C. Crabs were anesthetized by immersion in ice prior to removal of the gills.

Total RNA was extracted from gill tissue under RNAse-free conditions (Chomczynski and Sacchi, 1987) using materials obtained from Promega Corporation. Reverse transcription of poly(A⁺) mRNA was initiated with oligo(dT) primer and Superscript II reverse transcriptase (Gibco-BRL). Amplification of putative B-subunit sequences was achieved using the polymerase chain reaction (PCR; Sigma REDTaq) using degenerate and specific primers designed with the assistance of Primer Premier software. Following separation on 1 % agarose gels, PCR products were extracted from gel slices (Qiagen QiaQuick) and sequenced directly by the dideoxynucleotide method (Sanger et al., 1977) either manually on 5 % polyacrylamide gels or automatically at the Marine DNA Sequencing Center of Mount Desert Island Biological Laboratory. Fragment sequences were assembled using DNASTAR software and were analyzed for open reading frames using DNASIS. A search of GenBank using the BLAST algorithm (Altschul et al., 1997) revealed close matches with previously published sequences. Multiple alignments were accomplished with ClustalW (http://pbil.ibcp.fr/ANTHEPROT) and GeneDoc software (http://www.psc.edu/biomed/genedoc/), and a phylogenetic tree was generated with TreeView (http://taxonomy.zoology.gla.ac.uk/rod/rod.html).

A crude membrane fraction was prepared from anterior and posterior gills of Carcinus maenas acclimated to 10 ‰ salinity, and western blot analysis was performed according to the methods of Lucu and Flik (Lucu and Flik, 1999), using as primary antibody a monoclonal antibody against yeast B-subunit (Molecular Probes). Visualization of putative B-subunit protein was accomplished with the Opti-4CN system (BioRad Laboratories).

Shore crabs were acclimated to 10 ‰ sea water. Posterior and anterior gills were immersed in and perfused with 20 ‰ sea water for several minutes. Subsequently, the gills were perfused with a fixation solution composed of a 1:1 mixture of 20 ‰ sea water and a solution containing 8 % paraformaldehyde, 0.35 mol l⁻¹ sucrose, 150 mmol l⁻¹ NaCl, 0.2 mol l⁻¹ sodium cacodylate buffer (pH 7.3). After 5 min of perfusion, the external 20 ‰ sea water was replaced with fixation solution, and the gills were perfused for another 10 min. Individual gill lamellae were dissected and kept in fixation solution for 1 h. Cryofixation was performed by a method similar to that described previously (Tokuyasu, 1980). The lamellae were incubated in a solution containing 2.3 mol l⁻¹ sucrose in 0.1 mol l⁻¹ sodium cacodylate buffer (pH 7.3) for at least 2 h. Small pieces of tissue from the thick region of the lamellae containing the ion-transporting cells were mounted on aluminum rods and frozen in liquid Freon cooled with liquid nitrogen. Semi-thin (0.5 μm) cryosections were obtained as described previously (Ziegler, 1997) using a Leica Ultracut S microtome equipped with an FCS cryochamber, glass knives and an antistatic device (Diatome) at temperatures of −70 °C. Sections were transferred to poly-L-lysine-covered glass slides with a droplet of 2.3 mol l⁻¹ sucrose in 0.1 mol l⁻¹ cacodylate buffer (pH 7.3). The sucrose was washed off with phosphate-buffered saline (PBS), and the sections were treated successively with 0.05 mol l⁻¹ glycine in PBS for 15 min, 1 % SDS in PBS for 5 min, PBS three times for 5 min, 0.001 % streptavidin for 15 min to block endogenous biotin, PBS for 5 min, 0.1 % biotin to block the biotin-binding sites of the streptavidin, PBS twice for 5 min and, finally, blocking solution (BS) containing 5 % bovine serum albumin (BSA) and 0.1 % cold-water fish gelatine (Biotrend) in PBS for 15 min. Small droplets (12 μl) of primary antibody (anti-yeast V-ATPase B-subunit, Molecular Probes) at 2.5 μg ml⁻¹ in incubation solution (IS) containing 1 % BSA and 0.1 % cold-water fish gelatine in PBS were placed on the sections and incubated overnight. Control sections were incubated in IS without primary antibody. The sections were washed three times in 0.1 mol l⁻¹ phosphate buffer containing 0.5 mol l⁻¹ NaCl (PBS-500) and once in PBS, incubated for 1 h in biotinylated secondary antibody [anti-mouse F(a,b′)₂ fragments from sheep; Amersham], washed three times in PBS-500 for 5 min and once in PBS for 5 min. Sections were treated with TNB-blocking-solution (NEN) for 30 min, incubated with horseradish peroxidase (HRP)-linked streptavidin (NEN) at a dilution of 1:100 in TNB blocking solution for 1 h, and washed three times in PBS-500 for 5 min and once in PBS for 5 min. The sections were incubated in the fluorescent HRP substrate Tyramid-CY3 (NEN) at a dilution of 1:40 in amplification buffer (NEN) for 10 min in the dark, washed three times in PBS-500 for 5 min and once in PBS for 5 min. Subsequently, the sections were mounted in 80 % glycine, 20 % PBS plus 2 % N-propyl-gallate (to retard fading) and examined with a Zeiss Axiophot microscope. Micrographs were taken on Kodak Elite 400 film.

Relative levels of B-subunit mRNA in gill RNA extracts pooled from three animals per treatment were estimated using quantitative reverse transcriptase (RT)-PCR by incorporating biotin-dUTP during logarithmic amplification according to previously published methods (Towle et al., 1997). Single-tube duplex PCR was performed using primer pairs for both V-ATPase B-subunit and arginine kinase (Table 1), the latter serving as an unchanging control with respect to salinity and gill type (Kotlyar et al., 2000). The biotinylated products were separated on 1 % agarose gels, transferred to nylon membranes and visualized with the PhotoTope detection system (New England Biolabs). Intensities of B-subunit products were compared with those for arginine kinase mRNA by digitizing the bands using Un-Scan-Itgel software (Silk Scientific).

Gills were perfused according to previously published methods (Siebers et al., 1985; Weihrauch et al., 1998) using symmetrical saline solutions containing (in mmol l⁻¹): 248 NaCl, 5 CaCl₂, 5 KCl, 4 MgCl₂, 2 NaHCO₃ and 2.5 Tris at pH 7.8. During perfusion at a flow rate of 0.135 ml min⁻¹, transepithelial potential differences (PD_te) were monitored using a millivolt meter (Keithley, model 197) connected to the perfusate and bath solutions via Ag/AgCl electrodes.

To identify expression of V-type H⁺-ATPase in gill epithelium of the shore crab Carcinus maenas, we designed degenerate oligonucleotide primers (HATF2, HATR4; Table 1) on the basis of conserved regions in published cDNA sequences coding for the B-subunit of the V-ATPase in other species. Starting with an initial cDNA sequence of 360 nucleotides identified via BLAST search of GenBank as a putative B-subunit fragment (Weihrauch and Towle, 1997), we designed additional degenerate and specific sense and antisense primers (Table 1) to amplify and sequence the entire B-subunit cDNA from Carcinus maenas, using the RACE technique and AUAP primer (5′-GGCCACGCGTCGACTA-GTAC-3′ (Gibco-BRL) to obtain the 5′ and 3′ regions of the cDNA.

Amplification of the 3′-untranslated region (UTR) employing the primer combination HATF1770/AUAP gave two distinct cDNA fragments of different sizes (Fig. 1). The possibility that these products resulted from genomic DNA contamination was discounted by observing that no double bands appeared after PCR amplification with any of the oligonucleotide primers targeted towards regions other than the 3′-UTR. The complete cDNA sequences of both isoforms – the larger isoform called the L-form (2520 bp; GenBank Accession Number AF247971), the shorter isoform called the K-form (2446 bp; GenBank Accession Number AF189779) – are identical except for the insertion of 74 nucleotides at position 1870–1944 in the L-form. Both forms contain an identical open reading frame coding for a protein of 489 amino acid residues (Fig. 2) with a deduced molecular mass of 54.5 kDa. PCR reactions employing the specific primer combination HATF1550/HATR1960 to amplify sequences from both isoforms revealed that both the L-form and the K-form are expressed in anterior and posterior gills of brackish-water and seawater-adapted crabs. It also appears that the L-form is more highly expressed than the K-form in all tissues (Fig. 1B).

In both B-subunit isoforms, the cDNA sequence contains a motif with an adenine residue at the −3 position from the start codon, resembling a Kozak initiation sequence (Kozak, 1991) with an adenine instead of a guanine residue at the +4 position (ACAATGA; Fig. 2). In the 3′-UTR 20 base pairs (bp) upstream from the poly(A⁺) site, a polyadenylation signal (AATAAA) was found; this is the most frequently observed polyadenylation signal in eukaryotic mRNAs (Graber et al., 1999). In the 3′-UTR, several adenosine-and uridine-rich elements (AREs) with the sequence ATTTA were found. The abundance of ARE sequences in the 3′-UTR appears to be inversely related to the lifetime of eukaryotic mRNAs (Chen and Shyu, 1995). An additional ATTTA motif was found within the additional 74 bases of the L-form. This 74-base segment appears to represent an unspliced intron, carrying the typical splicing consensus motifs with AGCT at the 5′ splicing site, TACTGAC within the sequence and AGG at the 3′ splicing site (Padgett et al., 1986) (Fig. 2). Retention of the 74-nucleotide insert with its additional ARE may relate to regulation of mRNA degradation rates, but the functional significance of the two isoforms is completely unknown.

Employing the putative B-subunit nucleotide sequence in a BLAST search of GenBank (Altschul et al., 1997) revealed extremely high similarities to sequences all identified as the B-subunit of a V-ATPase. The derived Carcinus maenas B-subunit amino acid sequence showed a slightly higher similarity to the so-called ‘brain-type’ isoform B2 from human and cow (82 % identity) than to the ‘kidney-type’ isoform B1 (78 % identity; Fig. 3A). However, a repeated motif (GGGGGN)₃ found in human ‘brain-type’ B-subunit cDNA approximately 72 bp upstream from the putative polyadenylation signal (Bernasconi et al., 1990) is not present in the crab 3′-UTR region. Alignment of the predicted Carcinus maenas B-subunit amino acid sequence with representatives of invertebrate, plant and fungal species revealed a high degree of sequence conservation, ranging from 71 to 89 % (Fig. 3B).

The degenerate primers HATF2 and HATR4 were used to amplify putative fragments of B-subunit cDNA from several crab species differing in their degree of euryhalinity, including Callinectes sapidus (GenBank Accession Number AF189780), Eriocheir sinensis (GenBank Accession Number AF189782), Chasmagnathus granulatus (GenBank Accession Number AF189783) and Cancer irroratus (GenBank Accession Number AF189781). These sequences were used as query sequences for gapped BLAST searches (Altschul et al., 1997) and were all identified as B-subunit cDNAs with high bit-scores greater than 200. A phylogenetic analysis of the B-subunit cDNA sequences revealed the expected relationship between these species, with Carcinus and Callinectes in the family of the Portunidae, Eriocheir and Chasmagnathus in the family of the Grapsidae, and Cancer relatively distant from the other species in the family of the Cancridae (Fig. 4).

To determine whether the predicted B-subunit protein can be found in the shore crab, western blot analysis was performed employing membrane fractions from anterior and posterior gills from brackish-water-adapted (10 ‰ salinity) Carcinus maenas. After SDS–PAGE electrophoresis and transfer to nitrocellulose membranes, a monoclonal antibody against yeast 60 kDa B-subunit cross-reacted with proteins from the gill membrane fraction, producing two distinct bands with calculated sizes of approximately 59 kDa and 48 kDa (Fig. 5). Since the molecular mass of the B-subunit predicted from the Carcinus maenas cDNA sequence was calculated to be 54.5 kDa, we assume that the 59 kDa band on the western blot represents the B-subunit from crab gill. The smaller 48 kDa band is probably a proteolytic product of the B-subunit, even though aprotinin was included in the medium used to prepare the membrane fraction (Lucu and Flik, 1999). The monoclonal antibody against yeast 57 kDa B-subunit probably recognizes the corresponding target in Carcinus maenas because the amino acid sequences of the two proteins are 76 % identical (Nelson et al., 1989).

To investigate the subcellular location of the V-ATPase protein in epithelial cells of Carcinus maenas gill, the monoclonal antibody against the B-subunit was used to probe semi-thin cryosections (0.5 μm) from anterior and posterior gills of brackish-water-acclimated shore crabs. As analyzed by light microscopy, labeling by the antibody was found throughout the cytoplasmic compartment of both pillar cells and epithelial cells (Fig. 6). No particular accumulation of V-ATPase was detected at the apical membrane; indeed, the region adjacent to the apical membrane appeared to be deficient in V-ATPase reactivity. The distribution of V-ATPase in posterior gills, believed to function primarily in osmoregulatory ion transport, was similar to that in anterior gills, thought to function predominantly in gas exchange (Taylor and Taylor, 1992).

To clarify whether V-ATPase gene expression is affected by osmoregulatory challenge, B-subunit mRNA abundance in gills of brackish-water-(10 ‰ salinity) and seawater-(35 ‰ salinity) adapted crabs was compared with the mRNA abundance of arginine kinase, whose mRNA expression and enzyme activity are not affected by salinity change (Kotlyar et al., 2000). The results of this semi-quantitative analysis using biotinylated PCR products in the linear phase of amplification (Towle et al., 1997) show that the method is quite sensitive to template availability, easily revealing differences in mRNA abundance of 25 % or less (Fig. 7A). The estimated relative abundance of V-ATPase B-subunit mRNA in anterior gill was shown to be somewhat higher than in posterior gills in both brackish-water-and seawater-adapted Carcinus maenas, with increasing differences noted in low salinity (Fig. 7B,C). These results contrast with the relative expression of B-subunit mRNA in gills of Chinese crab E. sinensis, in which posterior gills of crabs acclimated to salinity extremes (35 ‰ sea water and fresh water) demonstrate higher levels of B-subunit mRNA than anterior gills (Fig. 7B,C).

To explore the possible involvement of V-ATPase in osmoregulatory ion uptake at a physiological level in Carcinus maenas, isolated posterior gills from brackish-water-adapted animals were perfused and bathed with identical hemolymph-like salines, and the transepithelial potential differences (PD_te) were monitored with and without the application of bafilomycin A₁, a specific inhibitor of the V-ATPase (Bowman et al., 1988). Under control conditions, a PD_te of −5.9±0.7 mV (mean ± S.E.M., N=5) was measured with respect to the apical bath, indicating active uptake of NaCl across the gill epithelium (Lucu, 1990). After symmetrical application of 1 μmol l⁻¹ bafilomycin, no change in PD_te could be observed. A reduction in PD_te of approximately 50 % resulted when Na⁺/K⁺-ATPase was blocked by application of 5 mmol l⁻¹ ouabain (in the presence of bafilomycin) in the basolateral saline. After wash-out, a partial recovery of the PD_te was detected, indicating that the gills had not been damaged by the treatments (Fig. 8).

Electrophysiological and tracer flux studies on osmoregulatory ion-uptake mechanisms in Carcinus maenas have led to the suggestion that in this crab the activity of the basolateral Na⁺/K⁺-ATPase (Towle and Kays, 1986) is the major driving force for electrogenic NaCl uptake across the gill epithelium. Both net NaCl influx and Cl⁻-dependent short-circuit current (I_sc) are dramatically inhibited by basolateral application of ouabain, a specific inhibitor of the Na⁺/K⁺-ATPase (Lucu and Siebers, 1987; Riestenpatt et al., 1996). The failure to identify the presence of a V-ATPase in Carcinus maenas gills by enzyme activity assays (Towle and Harding, 1993) questioned the importance of a proton pump in osmoregulatory ion uptake in this moderately euryhaline species. Now that we have molecular evidence that the B-subunit of the catalytic V₁ complex of the V-ATPase is indeed expressed in Carcinus maenas gills (Weihrauch and Towle, 1997), the role of this proton pump in transbranchial NaCl uptake has to be reconsidered.

Analysis of the deduced amino acid sequence of the Carcinus maenas V-ATPase B-subunit revealed a higher homology to the so-called ‘brain-type’ or B₂ isoform than to the ‘kidney-type’ or B₁ isoform (Fig. 3). In mammalian kidney, the ‘brain-type’ isoform is found in proximal tubules but not in intercalated cells. In contrast, the ‘kidney-type’ isoform is not abundant in proximal tubules but is abundant in intercalated cells (Nelson, 1991), which are specialized in H⁺ transport and where the V-ATPase resides in high densities in a polarized distribution in the apical plasma membrane (Brown et al., 1992). Thus, the Carcinus maenas gill B-subunit resembles a mammalian isoform that is not part of an apically localized V-ATPase. But, even if the B-subunit is a very highly conserved protein (multiple alignment of the Carcinus maenas amino acid sequence with B-subunit sequences from other phyla showed 78–82 % identities to vertebrate sequences, 77–80 % identity to plant sequences, 76 % identity to fungal sequences and 59 % identity to Archaea A-ATPase B-subunit), the function and the location of the crustacean gill V-ATPase probably differ from those in mammalian tissues. Further evidence of high sequence homology was obtained by western blot analysis of a Carcinus maenas gill membrane fraction in which a monoclonal antibody against yeast V-ATPase B-subunit detected a 59 kDa protein (Fig. 5), which probably represents the crab B-subunit whose molecular mass deduced from the cDNA sequence was calculated to be 54.5 kDa.

To localize the V-ATPase B-subunit within the gills of Carcinus maenas, antibody against the yeast B-subunit was employed in semi-thin cryosections of the so-called ‘dark areas’ representing highly differentiated transporting epithelium (Taylor and Taylor, 1992). These antibodies successfully cross-reacted with the B-subunit in the sternal epithelium of another crustacean, the land-inhabiting isopod Porcellio scaber (Ziegler et al., 2000). In Carcinus maenas, the V-ATPase B-subunit was found throughout the cytoplasm, but not in the apical membrane, in epithelial and pillar cells of both anterior and posterior gills (Fig. 6), suggesting vesicular rather than apical localization. The distribution of the V-ATPase throughout the gill epithelium is in contrast to the findings in seawater-acclimated Eriocheir sinensis, in which the V-ATPase is situated predominantly in the apical membrane of pillar cells (Putzenlechner, 1994). Our localization results indicate that, in Carcinus maenas gills, the V-ATPase probably serves in acidification of intracellular vesicles rather than in generating an electrochemical proton gradient across the apical membrane that would be expected if the V-ATPase were to function in osmoregulatory NaCl transport (Larsen et al., 1992; Lin and Randall, 1993; Zare and Greenaway, 1998).

It cannot be excluded, however, that the cytoplasmic distribution of the B-subunit may be related to the regulation of V-ATPase activity by disassembly and reassembly of the V₁ and V₀ domains (Kane, 1995; Merzendorfer et al., 1997; Sumner et al., 1995). During molting or starvation of Manduca sexta larvae, the V-ATPase situated in the apical membrane of midgut goblet cells and energizing active K⁺ secretion is inactivated by disassembling of the cytoplasmic V₁ domain from the membrane-associated V₀ domain (Wieczorek and Harvey, 1995). Therefore, during inactivation, a major proportion of the V₁ is found throughout the cytoplasm (Gräf et al., 1996; Sumner et al., 1995). Since the B-subunit is part of the V₁ complex, detection of this subunit in the cytoplasm of Carcinus maenas gills using specific antibodies could lead to inaccurate interpretations of our findings.

Comparison of the relative expression of the B-subunit in Carcinus maenas gills reveals a higher abundance of the B-subunit mRNA in the respiratory anterior gills than in the osmoregulatory posterior gills, with an anterior/posterior ratio of approximately 2 for shore crabs acclimated to low salinity and 1.1 for shore crabs acclimated to full-strength sea water (Fig. 7). This is of particular interest considering the greater metabolic activities of the posterior gills, especially in diluted media, as demonstrated by the sevenfold higher activity of carbonic anhydrase (Böttcher et al., 1995) and the approximately 3.5-fold higher activity of Na⁺/K⁺-ATPase (Siebers et al., 1982) following a reduction of external salinity. The enhanced abundance of V-ATPase B-subunit mRNA in anterior gills may relate to our previous observation that the anterior gills are more proficient in active ammonia excretion than the posterior gills (Weihrauch et al., 1998). Expression analysis shows that, in the osmoregulatory posterior gills of Carcinus maenas, mRNA abundance decreases slightly following acclimation from full-strength sea water to brackish water (Fig. 7), a trend that is in opposition to the increase in the rates of net NaCl uptake expected for Carcinus maenas acclimated to low salinity, indicating once more that the V-ATPase is probably not involved in osmoregulatory ion uptake in this species.

In contrast, the Chinese crab Eriocheir sinensis shows a substantially higher abundance of V-ATPase B-subunit mRNA in the posterior gills than in the anterior gills in animals acclimated to full-strength sea water or fresh water (Fig. 7). These data are in line with enzyme activity measurements of bafilomycin-A₁-sensitive ATPase in posterior and anterior gills of Eriocheir sinensis (Putzenlechner, 1994) and are consistent with the ion-transporting ability of posterior gills in this species. Strongly hyperosmotic in fresh water and slightly hypo-osmotic in sea water (Mantel and Farmer, 1983), the Chinese crab exhibits a wider range of osmoregulatory capacity than Carcinus maenas. V-ATPase mRNA abundance in posterior gills was lowest at 20 ‰ salinity, where the NaCl concentration of hemolymph most closely matches that of the medium (Onken, 1999). It is not known whether V-ATPase enzyme activity is also reduced under these conditions. The specific activity of the V-ATPase in seawater-acclimated Eriocheir sinensis was reported to be approximately three times higher than in freshwater-acclimated animals (Putzenlechner, 1994), whereas the relative V-ATPase mRNA expression measured in this study was somewhat higher in fresh water than in sea water (Fig. 7). A possible explanation for this discrepancy could be a different translation rate or stability of the mRNA, which may lead to different protein abundance, or the V-ATPase may be regulated by assembly/disassembly as noted above. In any case, the substantial levels of V-ATPase mRNA and enzyme activity in the posterior gills of seawater-acclimated animals may reflect a possible involvement of the V-ATPase in NaCl excretion, as suggested for the Malpighian tubules of insects (Maddrell and O’Donnell, 1992).

Measurements of transepithelial potential difference (PD_te) across symmetrically perfused crab gills have been interpreted extensively to indicate active ion transport. The steady-state PD_te of Carcinus maenas and Eriocheir sinensis posterior gills is altered by inhibitors of Na⁺ transport, including ouabain (an inhibitor of Na⁺/K⁺-ATPase) and amiloride (an inhibitor of the Na⁺/H⁺ exchanger and the epithelial Na⁺ channel) (Lucu and Siebers, 1986; Pequeux and Gilles, 1988; Siebers et al., 1985; present study). In addition, Cl⁻ channel blockers applied to the basolateral side depolarize the epithelium of Carcinus maenas gills (Siebers et al., 1990). However, PD_te of isolated Carcinus maenas gills is not affected by the V-ATPase inhibitor bafilomycin, indicating that the components of active NaCl transport remain intact under these conditions. In contrast, active ammonia excretion by Carcinus maenas gills is inhibited by 66 % by bafilomycin at 1 mol l⁻¹, the same concentration employed in the present experiments (Weihrauch et al., 2000).

In conclusion, the V-type H⁺-ATPase of Eriocheir sinensis gills appears to possess a suite of properties that suggests its intimate involvement in osmoregulatory ion transport. However, several lines of evidence, both molecular and physiological, indicate that this is not the case for Carcinus maenas, a moderate osmoregulator. The salinity minimum of 8 ‰ observed for survival of Carcinus maenas may result in part from the lack of participation of V-ATPase in mediating NaCl uptake from more dilute medium. It will be interesting to discover whether freshwater-tolerant crab species other than Eriocheir sinensis utilize V-ATPase in this important function.

The authors are indebted to Davin O’Connell and Jennifer Litteral, Mount Desert Island Biological Laboratory, for providing the rock crabs Cancer irroratus and to Griselda Genovese, University of Buenos Aires, for providing Chasmagnathus granulatus. The authors thank Christine Smith of the Marine DNA Sequencing Center at Mount Desert Island Biological Laboratory and also Ute Postel of Biologische Anstalt Helgoland for tissue preparations. The research was supported by the National Science Foundation (IBN-9807539) and the Foster G. McGaw Foundation.