Stimulus-induced Phosphorylation of Vacuolar H+-ATPase by Protein Kinase A*

Eukaryotic vacuolar-type H+-ATPases (V-ATPases) are regulated by the reversible disassembly of the active V1V0 holoenzyme into a cytosolic V1 complex and a membrane-bound V0 complex. The signaling cascades that trigger these events in response to changing cellular conditions are largely unknown. We report that the V1 subunit C of the tobacco hornworm Manduca sexta interacts with protein kinase A and is the only V-ATPase subunit that is phosphorylated by protein kinase A. Subunit C can be phosphorylated as single polypeptide as well as a part of the V1 complex but not as a part of the V1V0 holoenzyme. Both the phosphorylated and the unphosphorylated form of subunit C are able to reassociate with the V1 complex from which subunit C had been removed before. Using salivary glands of the blowfly Calliphora vicina in which V-ATPase reassembly and activity is regulated by the neurohormone serotonin via protein kinase A, we show that the membrane-permeable cAMP analog 8-(4-chlorophenylthio)adenosine-3′,5′-cyclic monophosphate (8-CPT-cAMP) causes phosphorylation of subunit C in a tissue homogenate and that phosphorylation is reduced by incubation with antibodies against subunit C. Similarly, incubation of intact salivary glands with 8-CPT-cAMP or serotonin leads to the phosphorylation of subunit C, but this is abolished by H-89, an inhibitor of protein kinase A. These data suggest that subunit C binds to and serves as a substrate for protein kinase A and that this phosphorylation may be a regulatory switch for the formation of the active V1V0 holoenzyme.

Vacuolar type H ϩ -ATPases (V-ATPases) 3 are the most versatile proton pumps, being common to all eukaryotic organisms, and are found in endomembrane systems and in the plasma membrane (1)(2)(3). V-ATPases are multi-subunit transporters composed of a catalytic ATP-hydrolyzing V 1 complex (Ϸ550 kDa), which resides on the cytoplasmic side of the membrane, and a membrane-bound proton-translocating V 0 complex (Ϸ250 kDa). V-ATPase-dependent proton pumping is essential for cellular pH homeostasis and creates an electrochemical proton gradient that energizes secondary transport mechanisms in a wide variety of organelles and membrane systems. Acidification of organelles by V-ATPase activity is crucial to various cellular processes such as neurotransmitter uptake into synaptic vesicles, intracellular protein trafficking, and the secretion and activation of lysosomal enzymes for protein processing and degradation (4 -7). Located in the plasma membrane of specialized cells, V-ATPases are involved in processes such as cation secretion, bone resorption, renal acidification, and osmoregulation (8 -16). With respect to this diversity of function, mutations in genes encoding V-ATPase subunits obviously lead to several diseases, e.g. osteopetrosis (17) or renal tubular acidosis (18).
Several mechanisms have been proposed for the regulation of V-ATPase activity (3). The most prominent and physiologically relevant mechanism is the reversible disassembly of the V-ATPase holoenzyme into its V 1 and V 0 complexes as discovered in the midgut of the tobacco hornworm Manduca sexta and in the yeast Saccharomyces cerevisiae (19,20). In both systems, a nutrient drop induced by glucose deprivation in yeast or the cessation of feeding because of molting or starvation in the insect lead to the reversible disassembly of the functional V-ATPase holoenzyme into its inactive V 1 and V 0 complexes (21,22). Although more recent additional examples support the notion that the reversible disassembly/reassembly is a widely used mechanism for the regulation of V-ATPase activity (23)(24)(25), the signaling cascades that trigger the association/dissociation process remain elusive.
The salivary glands of the blowfly Calliphora vicina have served as a model system for the analysis of the regulation of V-ATPase reassembly and activation. V-ATPase activity in these tubiform glands is under the control of the hormone serotonin (5-hydroxytryptamine, 5-HT). In the presence of 5-HT, the V 1 complex reallocates within minutes to the apical membrane and reassociates with the V 0 complex, thus leading to the active holoenzyme (23). Recent results demonstrate that a 5-HT-induced increase in intracellular cAMP induces V-ATPase reassembly (26) and that protein kinase A (PKA) is the downstream target of cAMP in this scenario. 4 These results lead to the assumption that one or more subunits of the V-ATPase become phosphorylated via PKA and that this phosphorylation may act as a trigger for holoenzyme formation in the blowfly salivary gland and possibly other systems.
Regarding the reversible assembly/disassembly, the V 1 subunit C is unique among V-ATPase subunits because it is released to the cytosol upon dissociation of the holoenzyme into its V 1 and V 0 complexes (21,27). Subunit C is an elongated molecule (28) that appears to bridge the V 1 with the V 0 complex (29). Moreover, subunit C binds to actin filaments, and this interaction may be involved in stabilizing the proton pump in its assembled state (30,31). These properties make subunit C suited to the control of V-ATPase reassembly state and to the mediation of the relevant cellular signals. Here we test the above-mentioned hypothesis and demonstrate that V-ATPase subunit C becomes phosphorylated by PKA.

EXPERIMENTAL PROCEDURES
Animals and Preparation-M. sexta (Lepidoptera, Sphingidae) was reared at 27°C under long day conditions (16 h light) at the University of Osnabrück. Blowflies (C. vicina) were reared at 25°C under a light-dark cycle of 12 h:12 h at the University of Potsdam. At 1-3 weeks after the eclosion of the flies, the abdominal portions of their salivary glands were dissected in physiological saline containing 128 mM NaCl, 10 mM KCl, 2 mM MgCl 2 , 2 mM CaCl 2 , 2.7 mM sodium glutamate, 2.7 mM malic acid, 10 mM D-glucose, and 10 mM Tris-HCl, pH 7.2.
Antibodies-Monospecific polyclonal antibodies directed against the recombinant subunit C from M. sexta were produced in guinea pigs (serum 488 -1; Ref. 27). On Western blots of blowfly salivary gland homogenates, these antibodies identified a single intense band at 42 kDa (Figs. 7 and 8) corresponding approximately to the molecular mass of Manduca subunit C (27). Guinea pig polyclonal antibodies against ␥-amino butyric acid (GABA) were from Biotrend (Cologne, Germany), rabbit antiserum against the Drosophila PKA catalytic subunit was from Daniel Kalderon (Columbia University, New York, NY), mouse antibody E7 against ␤-tubulin (32) was obtained from the Developmental Studies Hybridoma Bank (University of Iowa, Iowa City, IA), alkaline phosphatase-conjugated antiguinea pig antibodies were from Sigma, and horseradish-peroxidase-conjugated anti-guinea pig antibodies were from Jackson ImmunoResearch (West Grove, PA).
Preparation of the V 1 Complex, the V 1 V 0 Holoenzyme and the Recombinant Subunit C-The V 1 complex as well as the V 1 V 0 holoenzyme were purified from the posterior midgut of fifth instar larvae of M. sexta according to published procedures (33,34) with the modification that all buffers contained the protein phosphatase inhibitors fluoride (20 mM) and vanadate (1 mM). V 1 complex without subunit C was prepared by gel chromatography in the presence of 25% methanol as described by Vitavska et al. (31). The V-ATPase subunit C from M. sexta was expressed in Escherichia coli BL 21 cells by using the pET-16b expression system from Novagen (EMD Biosciences, Madison, WI) and was purified by nickel-nitrilotriacetic acid affinity chromatography as described previously (27). Protein determination was also performed as described previously (35).
Overlay Blots with Recombinant Subunit C-Samples containing 1.5 and 6 g of PKA catalytic subunit were transferred onto a nitrocellulose membrane via the slot-blot technique; transfer was controlled by Ponceau S staining. After being blocked for 1 h at room temperature with 3% gelatin, the membrane was incubated with 0.35 M recombinant subunit C from M. sexta for 1 h in the presence of 1% gelatin in TNNT buffer (20 mM Tris-HCl, pH 7.5, 0.5 M NaCl, 0.02% NaN 3 , and 0.05% Tween 20). Nonspecifically bound subunit C was removed by washing the membrane in TNNT buffer three times for 5 min each. For labeling of bound subunit C, the membrane was incubated for 1 h with anti-C antibodies at a dilution of 1:2,000 in 1% gelatin/TNNT buffer. After three washes of 5 min each in TNNT buffer, the membrane was incubated for 1 h with alkaline-phosphatase-conjugated anti-guinea pig antibodies at a dilution of 1:30,000 in 1% gelatin/TNNT buffer. The membrane was then washed as above, and the color reaction was performed with 0.34‰ nitro blue tetrazolium, 0.18‰ 5-bromo-4chloro-3-indolyl phosphate in 50 mM Tris-HCl, pH 9.5, 0.1 M NaCl, and 50 mM MgCl 2 . Another membrane without incubation with recombinant subunit C was used as a negative control.
Phosphorylation ) and Coomassie Blue staining, the gel was exposed to a phosphoscreen and finally analyzed by a phosphorimaging device (Molecular Dynamics, Sunnyvale, CA).
In nonradioactive experiments, phosphoproteins were detected by the Pro-Q Diamond phosphoprotein stain. First, 8 M V 1 complex from M. sexta was incubated with or without 0.7 M PKA catalytic subunit in PKA buffer for 3 h at 30°C. To separate the PKA catalytic subunit from the V 1 complex, the samples were loaded onto a discontinuous sucrose gradient (34) in 16 mM Tris-HCl, pH 8.1, 0.32 mM EDTA, 0.2 M NaCl, 9.6 mM ␤-mercaptoethanol and centrifuged at 4°C for 1.5 h at 310,000 ϫ g. The fractions containing the V 1 complex without the PKA catalytic subunit were collected and concentrated by precipitation with 20% trichloroacetic acid. The pellets were resuspended, heated in SDS sample buffer for 45 s at 95°C, and loaded onto a gel. To control the phosphorylation reaction, 6 M recombinant subunit C from M. sexta was incubated with or without 0.7 M PKA catalytic subunit for 1 h at 30°C. After SDS-PAGE as above, the gels were analyzed by Pro-Q Diamond phosphoprotein stain and afterward by Coomassie Blue staining.
To check whether subunit C can be phosphorylated after its reassociation with the V 1 complex, the V 1 complex without subunit C was incubated with a 10-fold excess of recombinant subunit C for 12 h at 4°C. V 1 complex saturated with subunit C (V 1 C) was collected by gel chromatography on a Superdex 200 column. Then 5 M V 1 C complex was incubated with 0.5 M catalytic subunit of PKA for 2 h at 30°C and afterward subjected to SDS-PAGE. The gels were stained with Pro-Q Diamond to detect phosphoproteins and afterward with Coomassie Blue.
Isolated blowfly salivary glands were homogenized on ice in sample buffer A (20 mM Tris-HCl, pH 7.8, 0.1 M KCl, 10 mM MgCl 2 , 10 mM Na 2 ATP, 5 mM EGTA, 1% Triton X-100), supplemented with protease inhibitor cocktail according to the manufacturer's instructions and centrifuged for 15 min at 20,000 ϫ g at 4°C to remove cell debris. The resulting supernatant was divided into pools of equal volumes, supplemented with the respective test reagent(s) (50 M 8-CPT-cAMP, antibodies), and incubated for 15 or 30 min at 37°C. The preparations were mixed with sample buffer B (250 mM Tris-HCl, pH 6.8, 5% SDS, 10% ␤-mercaptoethanol, 10% glycerol) to give a final SDS concentration of 1% and were heated for 5 min at 70°C. After SDS-PAGE (14.4% T, 0.4% C), phosphoproteins were stained with ProQ Diamond and imaged. Subsequently, the proteins were stained with Coomassie Blue.
Two-dimensional Electrophoresis-Two-dimensional electrophoresis was carried out with Mini-PROTEAN 2D and Mini-PROTEAN II cell from Bio-Rad. Isoelectric focusing was performed at 20°C in denaturating 5% polyacrylamide tube gels (8 M urea, 2% Triton X-100, 2% Servalyt 3/10) with a pH 3-10 gradient. Before the loading of samples, the pH gradient was established by pre-electrophoresis: for 10 min at 200 V, 15 min at 300 V, and 15 min at 400 V. Isolated blowfly salivary glands were incubated at room temperature for 5 min with the respective test reagent(s) (30 nM 5-HT, 100 M 8-CPT-cAMP, and 50 M H-89) diluted in physiological saline; a control group was bathed in physiological saline only. Subsequently, the glands were homogenized on ice in sample buffer A and centrifuged for 30 min at 120,000 ϫ g at 4°C. The supernatants were mixed with sample buffer B to give a final SDS concentration of 1.4% and heated for 5 min at 70°C. For the alkylation of cysteine residues, iodacetamide was added to give a final concentration of 6%, and the solution was incubated for 45 min at 30°C. The samples were diluted in sample buffer C (8 M urea, 2% CHAPS, 2% Servalyt 3/10, 0.0025% bromphenol blue) to give a final SDS concentration of 0.24% to substitute SDS by CHAPS. Proteins (from an equivalent of 5 glands/assay) were separated by the following voltage profile: 10 min at 500 V, 3.5 h at 750 V, and 1 h at 1000 V. Tube gels were rinsed in equilibration buffer (62.5 mM Tris-HCl, pH 6.8, 2.3% SDS,10% glycerol, 0.0025% bromphenol blue) just before SDS-PAGE (10.3% T, 0.3% C). The proteins were electrotransferred onto polyvinylidene difluoride membranes and probed by Western blot analysis as described previously (23).

Interaction of the V-ATPase Subunit C with the Catalytic
Subunit of Protein Kinase A-To test our assumption that subunit C can be phosphorylated by PKA, we initially chose the V-ATPase from midgut plasma membranes of the tobacco hornworm, M. sexta, a useful model system for the investigation of this family of proteins. First, we checked whether a direct interaction of subunit C with the catalytic subunit of PKA could be detected by overlay blotting. We used the catalytic subunit of bovine heart PKA (with more than 80% amino acid sequence identity with the respective Drosophila enzyme) as an appropriate substitute for the insect PKA. After loading the PKA catalytic subunit onto nitrocellulose and incubation with or without recombinant V-ATPase subunit C from M. sexta, strong immuno-signals developed on the membrane incubated with subunit C as compared with the control membrane ( Fig. 1). Interaction of the PKA catalytic subunit with the M. sexta subunit C was also confirmed by overlay blots after SDS-PAGE (not depicted). These results support the conclusion that the V-ATPase subunit C binds to the catalytic subunit of PKA and thus might also be a physiologically relevant substrate for phosphorylation.
Phosphorylation of Subunit C by Protein Kinase A-Phosphorylation was investigated by incubation of subunit C with the catalytic subunit of PKA in the presence of [␥-32 P]ATP. Fig. 2 demonstrates that, after SDS-PAGE and phosphorimaging analysis, subunit C is labeled with 32 P in a PKA-dependent manner. The upper band at ϳ42 kDa represents subunit C with a molecular mass of 44 kDa as deduced from the cDNA sequence, whereas the lower band corresponds to the PKA catalytic subunit, which has a molecular mass of about 40 kDa and is known to become autophosphorylated (36).
Phosphorylation of subunit C does not exclude the possibility that other V-ATPase subunits are also phosphorylated, especially in view of a recent report that a WNK (with no K (lysine)) kinase from Arabidopsis phosphorylates not only subunit C but also the V 1 subunits A, G, and either B or H (37). Therefore, we tested whether other subunits of the V 1 complex could also be phosphorylated by PKA. After incubation of the  Coomassie staining, the gel was exposed to a phosphoscreen and finally analyzed by phosphorimaging. Different amounts of subunit C were loaded per lane onto the gel (denoted above the image). The arrowheads indicate the PKA catalytic subunit that becomes autophosphorylated. V 1 complex with and without the PKA catalytic subunit, sucrose density gradient centrifugation was performed, and fractions containing only the V 1 complex were analyzed by SDS-PAGE. Fig. 3 shows gels after Coomassie staining and after Pro-Q Diamond phosphoprotein staining. The latter stain has, compared with the use of radiolabeled ATP, the advantage that phosphorylated proteins can be detected without prior incubation with a protein kinase and ATP. Thus, Fig. 3 also demonstrates that the recombinant subunit C, which had been expressed in E. coli, is not phosphorylated a priori and remains nonphosphorylated after incubation with ATP in the absence of the PKA catalytic subunit (Fig. 3B). In the presence of the PKA catalytic subunit, however, a strong phospho-signal could be observed at subunit C (Fig. 3B). In the V 1 complex purified from M. sexta, no subunit including subunit C was found to be phosphorylated, either in the absence or in the presence of the PKA catalytic subunit (Fig. 3B). Therefore, we suggest that the V 1 complex purified from the midgut cytosol does not contain phosphorylated subunits and that subunits cannot be phosphorylated by PKA as long as they are part of the complex. This also applies to subunit C, which usually occurs in substoichiometric amounts in the isolated V 1 complex (22).
However, there could be a qualitative difference between the substoichiometric subunit C that remains bound to the V 1 complex purified from M. sexta midgut and the subunit C that dissociates upon enzyme disassembly. Hence we reassociated the recombinant subunit C with the V 1 complex from which we had removed before all of subunit C. The resultant V 1 C complex was then incubated with the catalytic subunit of PKA. Indeed, subunit C in this complex could be phosphorylated by PKA (Fig. 4), thus confirming our suspicion. Thus, the free cytosolic subunit C as well as subunit C reassociated with the V 1 complex both appear to be competent for phosphorylation by PKA.
After dissociation of the V 1 V 0 holoenzyme, the majority of subunit C occurs freely in the cytosol (30). Therefore, it might be expected that the activation of PKA leads predominantly to the phosphorylation of the free subunit C, which then would have to bind to the V 1 complex in the phosphorylated state. To examine whether phosphorylated subunit C has this ability, we incubated the V 1 complex with either phosphorylated or unphosphorylated subunit C. Already after 30 min of incubation a significant amount of subunit C was found to be part of the V 1 complex, and no difference could be observed between phosphorylated and unphosphorylated subunits C (Fig. 5).
Next we checked whether the V 1 V 0 holoenzyme purified from M. sexta midgut contained phosphorylated subunits. As Fig. 6 (lane 2) demonstrates, this appeared not to be the case. Incubation of the holoenzyme with the catalytic subunit of PKA led to a phosphorylated band with a molecular mass similar to that of subunit C (Fig. 6B, lane 1). However, after removal of PKA by gradient centrifugation it became clear that the catalytic subunit of PKA, but not subunit C was phosphorylated. Thus, subunit C in the holoenzyme cannot be phosphorylated by PKA.
PKA-dependent Phosphorylations in Calliphora Salivary Glands-To examine whether V-ATPase subunit C becomes phosphorylated by PKA within intact cells, we used salivary   Unbound subunit C was removed by centrifugation in a sucrose gradient, and after SDS-PAGE, the fractions containing reassociated V 1 C complexes were analyzed by Pro-Q Diamond phosphoprotein staining (B) and finally by Coomassie staining (A). Lane S, molecular mass marker as in Fig. 4. V 1 C-P, V 1 complex reassociated with phosphorylated subunit C; V 1 C, V 1 complex reassociated with unphosphorylated subunit C. glands of the blowfly C. vicina. In this system, V-ATPase assembles and becomes activated within minutes after exposure to 5-HT (10,23), and the cAMP/PKA signaling cascade mediates the effect of 5-HT on the V-ATPase (26). 4 Experiments with entire glands and subsequent SDS-PAGE and Pro-Q Diamond staining of phosphoproteins indicated that stimulation with 5-HT or the membrane-permeable cAMP analog 8-CPT-cAMP led to a slight increase in phospho-signal at 42 kDa, a position corresponding to V-ATPase subunit C (data not depicted). Based on the results above that free subunit C but not subunit C in the holoenzyme became phosphorylated by PKA, we hypothesized that this 42-kDa phospho-signal should become more intense by using a cellular fraction with an increased amount of free subunit C. Therefore, glands were homogenized on ice and centrifuged, and the resulting supernatant was divided into two aliquots. One sample was supplemented with 50 M 8-CPT-cAMP, whereas the other served as a control. Fig. 7 shows that incubation with 8-CPT-cAMP for 30 min at 37°C caused a dramatic increase in phosphoprotein signal in the molecular mass region of 42 kDa. However, although this result is in line with our assumption that V-ATPase subunit C can be phosphorylated by PKA in vivo, we cannot exclude, at this point, the possibility that the prominent phospho-signal at 42 kDa results from other proteins.
PKA-dependent Phosphorylation of Subunit C in Calliphora Salivary Glands-To determine whether the phospho-signal at 42 kDa represents phosphorylated V-ATPase subunit C, we tested the effect of antibodies against subunit C on the emergence of the phospho-signal. As shown in Fig. 8, 8-CPT-cAMP led to a strong phospho-signal at 42 kDa. Preincubation of the samples with the antibodies largely abolished the 8-CPT-cAMP-induced increase in the 42-kDa phospho-signal. Other phospho-signals that increased in an 8-CPT-cAMP-dependent manner, however, were unaffected by the antibodies, demonstrating the specificity of the assay. As a further control, the assay was performed in the presence of polyclonal antibodies against an unrelated antigen, viz. GABA. Under these conditions, the 8-CPT-cAMP-induced increase in the 42-kDa phospho-signal was as intense as in the absence of antibodies. We thus conclude that the 42-kDa band that becomes phosphorylated in a PKAdependent manner represents the V-ATPase subunit C. Two-dimensional electrophoresis was undertaken as an alternative method to probe the phosphorylation of subunit C by PKA in intact salivary glands of C. vicina. Under control conditions as shown in Fig.  9A, an anti-C immuno-signal in the 42-kDa range was detected after two-dimensional electrophoresis. The signal ran from the edge of the gel near the application point of pH 10 down to a slightly lower pH. We assume that this signal resulted from subunit C, which, on the one   hand, had precipitated after entrance into the gel and, on the other hand, appeared in the range of its putative alkaline isoelectric point (calculated isoelectric points of subunit C from M. sexta, Drosophila melanogaster, and Aedes aegypti, respectively: 8.32, 7.81, and 8.44). A second slight immuno-signal, which was observed in the more acidic region, disappeared completely when salivary glands were pretreated with the specific PKA inhibitor H-89 (38). Incubation with 8-CPT-cAMP led to the increase of this immuno-signal and to the appearance of an additional spot at even lower pH (Fig. 9B). The same result was obtained when salivary glands were stimulated with 5-HT (Fig. 9C), although the intensity of this additional spot was lower than that in samples stimulated with 8-CPT-cAMP. Pre-treatment with H-89 prevented the development of these signals in both cases. Fig. 9E demonstrates that the stimulus-induced differences in immuno-signal were not due to or accompanied by changes in the total amount of subunit C within the glands. These results corroborate our hypothesis that a 5-HT stimulus, followed by an increase in cAMP concentration, leads to a PKA-dependent phosphorylation of the V-ATPase subunit C in C. vicina salivary glands and suggest further that subunit C has more than one phosphorylation site. Finally, phosphorylation of subunit C is a reversible process because the additional immuno-signals disappeared after washout of 5-HT (Fig. 9D).

DISCUSSION
This is the first demonstration that the V-ATPase subunit C binds the PKA catalytic subunit and is a substrate for this protein kinase. Remarkably, the phosphorylation of subunit C in salivary glands of C. vicina correlates with the hormone-induced reassembly and activation of the V-ATPase in the apical membrane of the secretory epithelial cells. We therefore suggest that phosphorylation of subunit C is not only an initial event in the association process of the V-ATPase complexes but also a key regulatory mechanism of V-ATPase activity.
Vacuolar H ϩ -ATPases are evolutionary related to ATPsynthases (F-ATPases) found in bacteria, mitochondria, and chloroplasts (39). The overall structure of V-and F-ATPases is characterized by a peripheral ATP-binding complex and a proton-translocating integral membrane complex (2, 40 -42). Both the ATP-binding subunits and the proteolipid subunits display a high sequence homology between V-and F-ATPases. Therefore, a function can be assigned to V-ATPase subunits by homology to those of the F-ATPase. The V-ATPase subunit C, however, is exceptional because it has no equivalent among the F-ATPases (41). Thus, subunit C may lend properties to V-ATPases distinguishing them from F-ATPases. Aside from its actin binding properties (30,31), subunit C may be a likely candidate for the control of V-ATPase disassembly/reassembly, a mode of regulation unique to V-ATPases (43). The position of subunit C in the V-ATPase holoenzyme, close to subunit a of the V 0 complex and to subunits E and G of the V 1 complex (29,44), make it ideally suited to regulate interactions between the V 1 and V 0 complexes and thus to control the formation of active V-ATPase holoenzymes. This suggestion is strengthened by the finding that disassembly of the V 1 V 0 holoenzyme appears to coincide with the dissociation of subunit C from the V 1 complex (21,27).
Here, we demonstrate, for the first time, that subunit C of the V-ATPase is a target for the cAMP/PKA messenger system. This intracellular signaling cascade has been shown to control V-ATPase activity in several systems (45)(46)(47)(48)(49) including C. vicina salivary glands (26). 4 Using recombinant V-ATPase subunit C from M. sexta and the bovine PKA catalytic subunit, we have shown that these proteins can interact physically and that subunit C is phosphorylated by PKA. Because of the high amino acid sequence identity of mammalian and insect PKA catalytic subunits, our findings imply that the PKA-dependent phosphorylation of subunit C is not an artifact but reflects a real property of the M. sexta subunit C. Binding of the PKA catalytic subunit to subunit C may also occur in C. vicina salivary glands, as indicated by the parallel redistribution of these proteins upon cell stimulation (Fig. 10). In nonstimulated secretory inactive cells, subunit C and PKA are distributed in the cytoplasm, whereas the major amount of V-ATPase protein is dissociated into V 1 and V 0 complexes (26). Following a 5-HT stimulus, V-ATPase subunit C 4 and the PKA catalytic subunit 5 both become enriched at the highly enfolded apical membrane. Binding of these proteins to each other leads to subunit C phosphorylation, as demonstrated by our assays with M. sexta subunit C and the bovine PKA catalytic subunit. Similarly, incubation of intact salivary glands with the cAMP analog 8-CPT-cAMP or the hormone 5-HT leads to a PKA-dependent increase in subunit C phosphorylation.
Two-dimensional electrophoresis of Calliphora salivary glands has revealed three different spots on the membrane for subunit C, possibly representing nonphosphorylated, monophosphorylated and bi-or oligophosphorylated states of this protein. Subunit C may thus have at least two phosphorylation sites for PKA. Likewise, the WNK kinase can phosphorylate Arabidopsis subunit C at multiple sites in vitro (37). The identity of PKA phosphorylation sites in subunit C is as yet unknown. By using bioinformatic algorithms (KinasePhos, NetPhosK, GPS), several putative phosphorylation sites for PKA can be identified in subunit C of M. sexta, D. melanogaster and other organisms, but the results vary dependent on the program used.
What is the physiological consequence of subunit C phosphorylation? In nonstimulated glands, phosphorylated subunit C is present in negligible amounts. Similarly, only a minor fraction of V-ATPase molecules is assembled and active under these conditions (10,23,26). Upon exposure to saturating concentrations of 5-HT or 8-CPT-cAMP (10, 50), a fraction of subunit C becomes phosphorylated via PKA. Likewise, saturating 5-HT levels induce a reassembly of the majority of V 1 and V 0 complexes to the active V 1 V 0 holoenzyme (23,26). Moreover, subunit C becomes phosphorylated within minutes after exposure to 8-CPT-cAMP or 5-HT and thus within a similar time scale as that for the PKA-dependent reassembly and activation of the V-ATPase on the apical membrane (10,26). Finally, PKA activity is absolutely required for 5-HTdependent regulation of V-ATPase holoenzyme reassembly. 4 Because our data demonstrate that subunit C is the only V-ATPase component that becomes phosphorylated by PKA and phosphorylation can only occur on subunit C that is free or bound to V 1 complex, it may be concluded that this phosphorylation of subunit C is an initial and essential event in V 1 V 0 holoenzyme reassembly.
In this scenario, the following model seems possible. Upon hormonally induced activation of PKA via cAMP, PKA phosphorylates subunit C, most of which is found occurring freely in the cytosol. Subunit C, V 1 complex, and the catalytic subunit of PKA then become enriched on the apical membrane. This process may be mediated by the binding capability of subunit C as well as subunit B in the V 1 complex to actin (30,31,51), and by a stimulus-induced reorganization of the actin filament system, as shown for Malpighian tubules of the yellow fever mosquito A. aegypti (52). The resulting enhanced concentration of the V 1 complex and of subunit C near the apical membrane increases the probability of their physical interaction and, in addition, the interaction with the membrane-bound V 0 complex. We suggest that the phosphorylation of subunit C is an important step that facilitates the reassociation of the reaction partners to an active V 1 V 0 holoenzyme. Upon reassembly, subunit C becomes dephosphorylated by an as yet unidentified phosphatase within the secretory cells. Our finding that C subunit in the holoenzyme is dephosphorylated suggests that dephosphorylation does not directly lead to V-ATPase disassembly. The signals for disassembly and inactivation of the V-ATPase remain enigmatic. The redistribution of the PKA catalytic subunit together with its substrate subunit C to the apical membrane (Fig. 10) may fasten C subunit phosphorylation and V-ATPase reassembly and thus support the maintenance of a large number of V-ATPase molecules in the holoenzyme state as long as the 5-HT stimulus persists.
Although the reversible disassembly/reassembly is a widespread mode of V-ATPase regulation, we do not know whether phosphorylation by PKA is a universal property of subunit C, especially if we take into consideration that subunit C is not a highly conserved V-ATPase subunit. Although the amino acid sequence identities within insects are 80% and more, the M. sexta subunit C is clearly less similar to mammalian subunits C and is only 35% identical to the yeast protein.  Cross-sections through salivary glands were triple-labeled with antibodies against actin (blue), subunit C (green), and PKA-C (red) by procedures as described previously (26). Labeling for actin reveals infoldings of the apical membrane (arrowheads). Asterisks indicate the lumen of the gland. Bar, 25 m. no information about the sites of PKA-dependent phosphorylation, we cannot determine whether these residues are conserved among subunit C orthologs. Last but not least, because subunit C has been shown to occur in a tissue-specific mode in various isoforms (53), we are left with the possibility that subunit C heterogeneity between organisms, and even within one organism, contributes to differences in V-ATPase regulation, either by the involvement of different PKA isoforms or even by other protein kinases.