Antibodies to Mammalian and Plant V-Atpases Cross React With the V-Atpase of Insect Cation-Transporting Plasma Membranes

An alkali metal ion pump is present on the apical plasma membrane of many insect K⁺-and Na⁺-transporting epithelial cells. Its localization on gastrointestinal epithelia such as the midgut, Malpighian tubules, salivary glands and rectum was deduced from fluid analysis and electrical voltage measurements in vivo (e.g. Ramsay, 1953; Maddrell, 1971) and flux measurements in vitro (e.g. Harvey and Nedergaard, 1964). Further localization to ion-transporting apical membranes was based on microelectrode impalements (e.g. Moffett and Koch, 1988) and X-ray microanalysis of tissues (Dow et al. 1984). Electron microscopy reveals the presence of F₁-like particles, portasomes, on all of these membranes (see Harvey, 1980; Thurm and Küppers, 1980).

In the insect normally used as a model, Manduca sexta, the suspected K⁺-transporting membrane, the goblet cell apical membrane (GCAM), was isolated using ultrastructural features, especially the presence of portasomes, as guides (Cioffi and Wolfersberger, 1983). Analysis of purified midgut membrane fractions demonstrated that K⁺-stimulated ATPase activity is restricted to GCAM (Wieczorek et al. 1986). GCAM ATPase was solubilized and shown to be a vacuolar-type ATPase (V-ATPase) on the basis of its subunit composition and inhibitor sensitivities (Schweikl et al. 1989). The GCAM V-ATPase pumps protons towards the goblet cavity (Wieczorek et al. 1989), creating a large electrical potential difference (PD) but no proton gradient (Chao et al. 1991). The PD drives electrogenic potassium/proton antiport, resulting in secondary active potassium transport towards the cavity. Thus, the GCAM K⁺ pump consists of a protonmotive V-ATPase in parallel with an electrogenic K⁺/nH⁺ antiporter (Wieczorek et al. 1991).

The F₁-like particles are thought to couple ATP hydrolysis to the creation of cationic electrochemical gradients in these insect epithelia (see Harvey et al. 1981, 1983b). Antibodies to the midgut GCAM V-ATPase label the portasome-studded apical membranes in light and electron micrographs of midgut and Malpighian tubules of M. sexta larvae (Klein et al. 1991) and of sensory sensilla of Antheraea pernyi adults (Klein and Zimmermann, 1991).

The amino acid sequences of V-ATPase subunit polypeptides are more than 70% conserved between archaebacteria, higher plants and mammals (Nelson, 1989). If this high sequence-homology extends to insects then antibodies raised against plant and mammalian V-ATPases should cross react with insect V-ATPases. We studied cross reactions of antisera against several xenic V-ATPases and used antiserum against a plant tonoplast V-ATPase to probe sections of M. sexta midgut and Malpighian tubule cells. The results support the working hypothesis that GCAM and Malpighian tubule portasomes are V-ATPase particles.

Rabbit antiserum against the holoenzyme of Kalanchoe daigremontiana tonoplast was a gift from Dr Hans Peter Haschke, Technical University of Darmstadt, Germany. Monospecific antisera against the 72 and 39 kDa subunits of V-ATPase from bovine chromaffin granules were gifts from Dr Nathan Nelson, Roche Institute of Molecular Biology, Nutley, New Jersey. Monospecific antisera to beet root 57 and 70kDa subunits were gifts from Dr A. B. Bennett, Department of Vegetable Crops, University of California, Davis. Colloidal gold particles (5 nm) were conjugated to goat anti-rabbit antibody and used as a secondary probe for immunocytochemical investigations.

Partially purified goblet cell apical membranes were prepared from M. sexta midguts (Cioffi and Wolfersberger, 1983). The V-ATPase was solubilized and purified from the partially purified membranes (Schweikl et al. 1989; Wieczorek et al. 1990).

Tissue was dissected from the midgut and Malpighian tubules of feeding, fifthinstar larvae, 3–4 days after ecdysis from the fourth instar. It was fixed by inflation with cold paraformaldehyde/lysine/periodate fixative (PLP; McLean and Nakane, 1974) for 1 h or by rapid freezing in liquid propane at –196°C (Howard and O’Donnell, 1987). Dehydration of PLP-fixed material was carried out in a graded alcohol series. Frozen tissue was transferred to acetone at –80°C for freeze substitution. Tissue was embedded in Lowicryl HM-20 resin (Polysciences) at 4°C and polymerized by ultraviolet irradiation (Altman et al. 1984).

SDS extracts of purified GCAM V-ATPase were separated into subunits by SDS-polyacrylamide gel electrophoresis (Laemmli, 1970). The proteins were then blotted onto nitrocellulose or Immobilon (Millipore) membranes (Wieczorek et al. 1991). The blots were blocked in 3% gelatin in Tris-buffered saline (TBS: 20 mmol l⁻¹ Tris, 0.5 moll⁻¹ NaCl, pH 7.5) for Ih. The blots were probed for between 1 h and overnight with antibodies to subunits of xenic V-ATPases and washed (twice) for 10min in TBS-Tween 20 (0.3% w/v). A secondary probe of goat anti-rabbit IgG, conjugated with alkaline phosphatase (Sigma), was diluted 1:1000 with TBS-1 % gelatin and used to identify reactive bands. After 1-3 h of incubation with secondary antibody, the blots were washed (twice) for 10 min in TBS-Tween. The antibody complex was visualized using Bio-Rad kit no. 1706460. For comparison with the immunoprobed blots, a duplicate gel was run concurrently and stained conventionally with Coomassie Blue or the blot was stained with Amido Black.

For immunocytochemical localization of tissue V-ATPase, sections 1 μm thick were dried onto gelatin-covered slides and blocked in a buffer containing bovine serum albumin (BSA) solution (1 % w/v) in TBS for 30 min. These sections were probed with antibodies to beet root V-ATPase using pre-immune serum as control. The immunolocalization procedure was carried out at 24°C in a humid chamber. Pre-immune serum or serum containing the antibody was diluted 1:100 with TBS-BSA. After a 12h incubation period the slides were rinsed twice, with stirring, in TBS-Tween. Tissues were incubated for 1h with a 1:50 dilution of secondary antibody in TBS-BSA after which the slides were rinsed twice for 10 min with TBS-Tween and then briefly rinsed in deionized water. Light microscopic visualization of the colloidal gold IgG conjugate was augmented by silver enhancement (Intense II kit; Janssen Life Science Products, Piscataway, NJ). Control sections were stained lightly with Toluidine Blue for tissue differentiation.

Polyclonal antibodies directed against V-ATPases purified from K. diagremontiana tonoplast and bovine chromaffin granules cross-reacted with lepidopteran midgut GCAM ATPase in immunoblots (Fig. 1). Antibodies against the holoenzyme from tonoplasts reacted with the 67, 56, 40, 28 and 20kDa subunits of midgut ATPase (Fig. 1, lane B). An antiserum against the 72kDa subunit of chromaffin granule ATPase, thought to contain the catalytic site, reacted with the corresponding 67kDa subunit of midgut ATPase (Fig. 1, lane C). An antiserum against the 39 kDa subunit of chromaffin granule ATPase reacted with the 43 kDa subunit of midgut ATPase (Fig. 1, lane D). Polyclonal antibodies to the 57 kDa subunit of beet root V-ATPase cross-reacted with the M. sexta V-ATPase 56 kDa subunit (Fig. 2, lane C). The 70kDa beet root antibody failed to cross-react with GCAM 67 kDa subunit whereas the 72 kDa chromaffin granule antibody did react with this subunit; this difference may be due to technical problems. Finally, antibodies to oat root tonoplast V-ATPase cross-reacted with the M. sexta V-ATPase 56 kDa subunit (Fig. 2, lane E).

Sections of midgut probed with antibodies to the beet root 57 kDa V-ATPase subunit showed clear immunolabelling of the apical membrane of goblet cells from both anterior and posterior midgut in light micrographs (v, Figs 3A and 4A). The labelling was similar in both freeze-substituted tissue (Fig. 3) and fixed tissue (Fig. 4). The cells are so large (up to 100 μm tall) that there can be little doubt that the localization of the immunolabelling in light micrographs corresponds to the region of the portasome-studded apical membranes. The intensity of this immunocytochemical labelling suggests that V-ATPase must be highly concentrated in GCAM. By comparison, the GCAM of control tissues, which were exposed to pre-immune serum, was unlabelled (Figs 3B and 4B). Strong immunolabelling by antibody to the 57 kDa subunit was also seen on the apical plasma membrane of the Malpighian tubule cells (arrowhead Fig. 4A). Nonspecific staining, characteristic of a rabbit polyclonal serum, when used for probing insect tissue, was also apparent in basal lamellae (l), connective tissue (ct) and the glycocalyx of the midgut brush border (bb), especially in the apical regions. A slight staining could be seen in nuclear material (n) (Fig. 3B). Tangential sections through the brush border of PLP-fixed material revealed dense staining of the glycocalyx and moderate nonspecific staining of the brush border itself (Fig. 4B). It is likely that most of the nonspecific labelling involves glycoproteins. Moreover, since background staining was not seen to such an extent in the freeze-substituted tissue, it may be attributable to the PLP fixation method, which is especially effective in preserving carbohydrate side chains (McLean and Nakane, 1974). Thus, the background labelling may be due, in part, to antibody reactions induced by the fixation method.

The immunobiochemical cross reactions between polyclonal antibodies raised against the 57 kDa subunit of beet root tonoplast and oat root tonoplast and the 72 kDa subunit of chromaffin granule V-ATPase and the corresponding 56 kDa and 67 kDa subunits from GCAM V-ATPase (Figs 1 and 2), as well as the other cross reactions reported here, support the hypothesis that GCAM ATPase shares many epitopes with distantly related, xenic V-ATPases. The beet root antibodies preferentially labelled M. sexta goblet cell apical membrane from both anterior (Fig. 3) and posterior (Fig. 4) midgut. This immunocytochemical evidence supports the immunobiochemical evidence in demonstrating the kinship between plant and insect V-ATPases. Klein et al. (1991) have already shown, by light microscopy, that monoclonal antibodies against several subunits of the M. sexta midgut V-ATPase specifically label the goblet cell apical membrane of M. sexta posterior midgut cells and, by electron microscopy, that monoclonal antibodies to the 67 kDa subunit of the midgut V-ATPase also label the same membrane. The labelling of this membrane with a xenic antibody suggests that the plant V-ATPase contains the epitope responsible for the labelling. The present results show, for the first time, specific labelling of anterior midgut GCAM. Like posterior midgut, anterior midgut transports K⁺ (Cioffi and Harvey, 1981) and has portasomes on its GCAM (Cioffi, 1979). The beet root antibodies also label the apical membrane of M. sexta Malpighian tubule cells (Fig. 4), supporting the hypothesis (Klein et al. 1991) that a V-ATPase energizes the cation transport that underlies fluid secretion by these organs.

The high sequence homology between distantly related V-ATPases leads one to expect that the M. sexta antibodies of Klein et al. (1991) and the beet root antibody used in this study would label the V-ATPase-containing endomembranes of midgut cells as well as they label the portasome-studded plasma membranes. Since the M. sexta goblet cell apical membrane develops from an intracellular vacuole (Hakim et al. 1988), the V-ATPase of M. sexta endomembranes could be similar to that of goblet cell apical membrane. The slight intensification of cytoplasmic stain in this study is consistent with labelling of V-ATPases in endomembranes. Klein et al. (1991) suggest that the faint labelling of goblet cell cytoplasm by monoclonal antibodies against midgut V-ATPase might be explained by a reaction with specific epitopes of prospective GCAM ATPase while the enzyme was en route from its site of synthesis to GCAM. The intense labelling of portasome-studded apical membranes by antibodies against V-ATPases in both studies suggests that the concentration of the enzyme in these plasma membranes is much higher than that in endomembranes. This conclusion is supported by biochemical evidence. SDS-PAGE gels of purified goblet cell apical membranes (Harvey et al. 1983a; Wieczorek et al., 1991) contain major components corresponding to the subunits of solubilized V-ATPase (Schweikl et al. 1989; Wieczorek et al. 1990), implying that the V-ATPase is a major constituent of GCAM.

This paper, like that of Klein et al. (1991), supports the hypothesis that certain insect apical plasma membranes are energized by a primary, electrogenic, protonmotive V-ATPase (Wieczorek et al. 1991). The proton-motive V-ATPase, by itself, can only impose an electrical PD across the membrane (Al-Awqati, 1986). The presence of carriers, ion channels and water movements determines the physiological utilization of the generated PD. In lepidopteran midgut a large PD alkalizes the lumen (Dow, 1984; Dow and Harvey, 1988; Dow and O’Donnell, 1990; Chamberlin, 1990; Chao et al. 1991) and drives amino acid/cation symport from lumen to cells (Hanozet et al. 1989; Hennigan and Wolfersberger, 1989). In Malpighian tubules, net K⁺ transport drives fluid secretion (Maddrell, 1971) by a process that is inhibited by bafilomycin A₁ (Bertram et al. 1991). Midgut V-ATPase antibodies label the apical membrane of auxiliary cells in A. pernyi sensory sensilla (Klein and Zimmermann, 1991). The pump PD across this membrane modulates the receptor current (Thurm and Kiippers, 1980).

The presence of F₁-like particles (Gupta and Berridge, 1966; Smith, 1969) first suggested that goblet cell apical membrane might be responsible for active K⁺ transport across the lepidopteran midgut epithelium (Anderson and Harvey, 1966) and guided the isolation of this membrane (Cioffi and Wolfersberger, 1983). Similar particles are present on the apical membrane of Malpighian tubule and sensory sensilla auxiliary cell apical membrane (see references in Harvey, 1980; Thurm and Kiippers, 1980), Neurospora crassa and plant vacuoles (Bowman et al. 1989; Klink and Lüttge, 1991; Morré et al. 1991; Taiz and Taiz, 1991), chromaffin granules (Schmidt et al. 1982), guinea pig brain synaptic vesicles (Stadler and Tsukita, 1984) and kidney acidifying membranes (Brown et al. 1987). In many of these cases the particle-containing membrane has V-ATPase activity and labels with antibodies to the V-ATPase. Considering the homology of the B and A subunits of V-ATPase with the α and β subunits of F-ATPase and the proposed similarity of V-ATPase and F-ATPase structure (e.g. Nelson, 1991), it is not surprising that portasomes appear to be V] ATPase particles just as ‘lollipops’ are Fi ATPase particles.

We thank Dr Hans Peter Haschke, Technical University of Darmstadt, Dr Nathan Nelson, Roche Institute of Molecular Biology, and Dr A. B. Bennett, University of California, Davis, for the antibodies described in the Materials and methods section. We also thank Dr Heven Sze, University of Maryland, for screening a nitrocellulose blot of partially purified GCAM ATPase with rabbit polyclonal monospecific antiserum against the V-ATPase of oat root tonoplast and Dr Helmut Wieczorek for critical discussions and helpful suggestions on this manuscript. This work was supported in part by NIH Research Grant ROI Al 22444, DFG Research Grant Wi698 and EEC Research Grant SC1-CT90-0480.