Activation of PMCA by calmodulin or ethanol in plasma membrane vesicles from rat brain involves dissociation of the acetylated tubulin/PMCA complex

Authors


C. H. Casale, Departamento de Biología Molecular, Facultad de Ciencias Exactas, Físico-Químicas y Naturales, Universidad Nacional de Río Cuarto, Río Cuarto, 5800 Córdoba, Argentina
Fax: +54 358 467 6232
Tel: +54 358 467 6422
E-mail: ccasale@exa.unrc.edu.ar

Abstract

We have recently shown that acetylated tubulin interacts with plasma membrane Na+,K+-ATPase and inhibits its enzyme activity in several types of cells. H+-ATPase of Saccharomyces cerevisiae is similarly inhibited by interaction with acetylated tubulin. The activities of both these ATPases are restored upon dissociation of the acetylated tubulin/ATPase complex. Here, we report that in plasma membrane vesicles isolated from brain synaptosomes, another P-type ATPase, plasma membrane Ca2+-ATPase (PMCA), undergoes enzyme activity regulation by its association/dissociation with acetylated tubulin. The presence of acetylated tubulin/PMCA complex in membrane vesicles was demonstrated by analyzing the behavior of acetylated tubulin in a detergent partition, and by immunoprecipitation experiments. PMCA is known to be stimulated by ethanol and calmodulin at physiological concentrations. We found that treatment of plasma membrane vesicles with these reagents induced dissociation of the complex, with a concomitant restoration of enzyme activity. Conversely, incubation of vesicles with exogenous tubulin induced the association of acetylated tubulin with PMCA, and the inhibition of enzyme activity. These findings indicate that activation of synaptosomal PMCA by ethanol and calmodulin involves dissociation of the acetylated tubulin/PMCA complex. This regulatory mechanism was shown to also operate in living cells.

Abbreviations
PMCA

plasma membrane Ca2+-ATPase

PMV

plasma membrane vesicle

TSA

Trichostatin A

Tubulin, the main protein constituent of microtubules, is a soluble cytosolic protein which has also been found associated with membranes. In neural and non-neural cells, the membrane localization of tubulin was previously reported to be due, at least in part, to its association with Na+,K+-ATPase [1–4]. In Saccharomyces cerevisiae, membrane tubulin is bound to H+-ATPase [5]. In all cases, the association of tubulin with ATPase results in inhibition of the enzyme activity. Conversely, when the ATPase/tubulin complex is dissociated, ATPase activity is restored. For example, treatment of cells with Na+,K+-ATPase activators induced dissociation of the complex and stimulated enzyme activity [3,4]. A similar effect was observed in S. cerevisiae using glucose as an activator of H+-ATPase [5].

When tubulin is bound to the ATPase, it behaves as a hydrophobic compound and is found in the detergent phase after partition with Triton X-114. By contrast, free tubulin is recovered in the hydrophilic phase [6]. These observations were the basis for a method used in our previous studies (and this study) to estimate the amount of acetylated tubulin/ATPase complex by measuring the amount of acetylated tubulin present in the detergent phase. Immunoprecipitation was also a useful technique to characterize and quantify the complex under different circumstances. The presence of an acetyl group on Lys40 of the alpha chain has recently been shown to be an absolute requirement for tubulin to associate with Na+, K+-ATPase [7].

Na+,K+-ATPase and H+-ATPase are both members of the P-type ATPase family, and plasma membrane Ca2+-ATPase (PMCA), a calmodulin-regulated P-type ATPase that also belongs to this family has a key role in the control of intracellular Ca2+ [8]. Like the other P-type pumps, PMCA contains ∼ 10 transmembrane segments, including both terminal ends, which are exposed to the cytosol. PMCA is encoded by at least four different genes resulting in four basic isoforms, one of which, isoform 4 (PCMA4), has been found in rat synaptosomes [9]. The central portion of the PMCA molecule contains the catalytic domain, which is homologous with those of other family members [8]. We investigated the possible presence of an acetylated tubulin/PMCA complex in the plasma membrane. Here, we report the regulation of PMCA by acetylated tubulin, and the effects of ethanol and calmodulin, which have previously been described as stimulators of PMCA activity, on the association/dissociation of acetylated tubulin/PMCA complex.

Results

Presence of acetylated tubulin/PMCA complex in plasma membrane vesicles

A plasma membrane vesicle (PMV) preparation was analyzed by western blotting using staining with antibodies to total PMCA, PMCA4 isoform and acetylated tubulin. These proteins were present in PMVs (Fig. 1A, ‘PMV’). After partition in Triton X-114, total PMCA and PMCA4 isoform were detected uniquely in the detergent phase (Fig. 1A, ‘Deterg’), whereas most acetylated tubulin was found in the detergent phase with a small fraction in the aqueous phase (Fig. 1A, ‘Aqueous’). To investigate the possible association of acetylated tubulin with PMCA, we performed immunoprecipitation experiments using the corresponding antibodies linked to Sepharose beads. As a control of immunoprecipitation specificity, phosphatidylinositol 3-kinase mAb (an irrelevant antibody) bound to Sepharose beads was also used. Figure 1B (lane 1) shows that PMCA, PMCA4 isoform and acetylated tubulin are present in solubilized membranes prior to immunoprecipitation. When solubilized membranes were immunoprecipitated with 6-11B-1 Ig bound to Sepharose beads, total PMCA, PMCA4 isoform and acetylated tubulin were detected in the precipitated material (Fig. 1B, lane 2). These proteins were also detected by immunoprecipitation with 5F10 Ig (Fig. 1B, lane 3). However, none of these proteins was precipitated with antibody to phosphatidylinositol 3-kinase bound to Sepharose beads (lane 4). The supernatant fractions of the immunoprecipitation experiments were also investigated. As shown in Fig. 1C (lane 2), when anti-(acetylated tubulin) Ig–Sepharose beads were used as the precipitant, part of the PMCA isoform and total PMCA remained in the soluble state, although acetylated tubulin did not. By contrast, when anti-PMCA Ig–Sepharose beads were used (lane 3), neither PMCA isoform nor total PMCA was detected in the supernatant fractions, although part of the acetylated tubulin was. Because samples loaded in each lane were represented the same amount of solubilized membrane, we are able to calculate (from densitometric scanning of three independent experiments) that ∼ 35% of the PMCA isoform (and total PMCA) and 50% of the acetylated tubulin present in PMVs are not part of the acetylated tubulin/PMCA complex. These results indicate that acetylated tubulin and PMCA form part of the same complex inserted into PMVs and that PMCA4 is one of the isoforms present in the complex.

Figure 1.

 Acetylated tubulin/PMCA complex is present in PMVs from rat brain. (A) PMVs were analyzed by western blotting by staining on separate lanes with 5F10 Ig, anti-PMCA4 Ig and 6-11B-1 Ig (Ac-tubulin). Another sample of the same membrane preparation was partitioned in Triton X-114 to determine hydrophobic acetylated tubulin and hydrophilic tubulin, as described in Experimental procedures. Aliquots of the detergent phase (Deterg) and hydrophilic phase (Aqueous) were subjected to western blotting and stained with 5F10 Ig (upper), anti-PMCA4 Ig (middle) and 6-11B-1 Ig (lower). (B) PMVs solubilized with 0.5% Triton X-100 were immunoprecipitated with Sepharose beads linked to 6-11B-1 Ig (lane 2), 5F10 Ig (lane 3) or anti-(phosphatidylinositol 3-kinase) Ig (lane 4). Immunoprecipiated materials were analyzed. Lane 1, detergent-solubilized PMVs prior to immunoprecipitation. (C) Supernatant fractions of the immunoprecipitation experiment described in (B). In all cases, volumes of the analysed samples were calculated to be representative of the same amount of PMVs.

The influence of ethanol and calmodulin on the acetylated tubulin/PMCA complex

Ethanol and calmodulin have been reported to activate PMCA in rat brain synaptosomes [10–12]. P-ATPases (Na+,K+-ATPase and H+-ATPase) are known to be activated by effectors that disrupt the corresponding acetylated tubulin/ATPase complex. We investigated whether ethanol and/or calmodulin induce activation of Ca2+-ATPase via the mechanism observed for H+- and Na+,K+-ATPases, that is, dissociation of the tubulin/Ca2+-ATPase complex. For this purpose, we exposed PMVs to various concentrations of ethanol and calmodulin (using experimental conditions described by other authors) [10,13] and determined the amount of hydrophobic acetylated tubulin by measuring the tubulin/PMCA complex. Using partitioning in Triton X-114 and western blotting we also determined PMCA activity by measuring its 45Ca2+ uptake. The amount of acetylated tubulin in the detergent phase in PMVs decreased gradually as the concentration of ethanol or calmodulin increased (Fig. 2A,B). The immunoblots in Fig. 2 (upper) show that the decreased intensity of the acetylated tubulin bands in the detergent phase is not an artifact but is due to a consistent increase in the amount of this protein in the aqueous phase. At 0.6% ethanol or 60 nm calmodulin ∼ 30% of acetylated tubulin remained in the membranes. The observed diminution in hydrophobic acetylated tubulin was assumed to be due to dissociation of the acetylated tubulin/PMCA complex which renders hydrophilic tubulin. This was confirmed by immunoprecipitation experiments (see below). However, PMCA activity gradually increased to > 190% of the control value in response to ethanol or calmodulin treatment. Thus, similar to our previous results with other P-ATPases, Ca2+-ATPase activity increased gradually as the acetylated tubulin/PMCA complex dissociated, suggesting that ethanol or calmodulin stimulates enzyme activity via dissociation of the complex. That the observed decrease in acetylated tubulin in the detergent phase as ethanol or calmodulin concentration increased was due to modification of the acetylation pattern or a change in antibody affinity, rather than dissociation of the complex, was ruled out because incubation of PMVs with 0.8% ethanol or 80 nm calmodulin followed by western blotting of these vesicles (without Triton X-114 partition) produced no change in the intensity of the acetylated tubulin bands (results not shown).

Figure 2.

 Effect of ethanol and calmodulin on the quantity of acetylated tubulin/PMCA complex and PMCA activity. PMVs (0.25 mg protein·mL−1) were incubated in transport buffer for 20 min at 37 °C in the presence of various amounts of (A) ethanol or (B) calmodulin. The amount of hydrophobic acetylated tubulin, as a measure of the tubulin/PMCA complex, was then determined by the Triton X-114 partition method and subsequent western blot analysis; PMCA activity was determined by measuring 45Ca2+ uptake (see Experimental procedures). Immunoblots of aliquots of the detergent and aqueous phases representative of the same amount of PMVs and stained with anti-(acetylated tubulin) Ig are shown in the upper panels. Acetylated tubulin bands corresponding to detergent phases were scanned. Absorbance values and PMCA activities are shown in the lower panels. Values are mean ± SD from three independent experiments.

Physical dissociation of the acetylated tubulin/PMCA complex by ethanol or calmodulin treatment was also assessed by immunoprecipitation. PMVs were treated with 0.6% ethanol, washed by sedimentation/resuspension, solubilized with detergent and immunoprecipitated with anti-(acetylated tubulin) Ig–Sepharose beads or anti-PMCA Ig–Sepharose beads. The precipitated and soluble fractions were immunoblotted and revealed using anti-PMCA Ig and anti-(acetylated tubulin) Ig. Comparison of the absorbance value for PMCA in the input material (Fig. 3A, lane 1) with that for PMCA precipitated by anti-(acetylated tubulin) Ig–Sepharose beads (Fig. 3A, lane 3) reveals that ∼ 66 ± 15% of the PMCA in the membrane is associated with acetylated tubulin (mean ± SD from three independent experiments). The presence of an acetylated tubulin band in the soluble fraction after immunoprecipitation with anti-PMCA Ig–Sepharose beads (Fig. 3A, lane 4), even when no PMCA remains in that fraction (Fig. 3A, lane 4), indicates that part of the acetylated tubulin in membranes is not associated with PMCA. Importantly, ethanol treatment led to less PMCA being precipitated with anti-(acetylated tubulin) Ig. This can be seen by comparing PMCA bands precipitated from membranes treated (Fig. 3B, lane 3) and not treated (Fig. 3A, lane 3) with ethanol. This suggests that ethanol treatment induced dissociation of the PMCA/acetylated tubulin complex. The same conclusion can be drawn by comparing the amounts of acetylated tubulin precipitated by anti-PMCA Ig from membranes that were treated (Fig. 3B, lane 5) or not treated (Fig. 3A, lane 5) with ethanol. From densitometric scanning of those bands, it can be calculated that ethanol induced the dissociation of ∼ 60 ± 8% of the complex (mean ± SD from three independent experiments).

Figure 3.

 Effect of ethanol on the interaction between PMCA and acetylated tubulin in PMVs. PMVs were incubated in the absence (A) or presence (B) of 0.6% ethanol for 20 min at 37 °C, and washed by centrifugation. Pelleted vesicles were dissolved in Triton X-100 and immunoprecipitated with anti-(acetylated tubulin) Ig (Anti ac-tub–Sepharose) or anti-PMCA Ig (anti-PMCA–Sepharose) as described in Experimental procedures. Input (IN), supernatant (S) and pellet (P) fractions were analyzed by SDS/PAGE followed by immunoblotting with an antibody against acetylated α-tubulin or PMCA. In all cases, volumes of analyzed samples were calculated to be representative of the same amount of PMVs. Only areas corresponding to relevant bands from a typical experiment are shown.

When membranes pre-treated with various concentrations of calmodulin were solubilized with detergent and immunoprecipitated with anti-(acetylated tubulin) Ig bound to Sepharose beads, the amount of PMCA precipitated decreased as the calmodulin concentration increased (Fig. 4). At between 40 and 50 nm calmodulin, ∼ 50% of the complex was dissociated (lower). This is consistent with the percentage activation of PMCA determined at 50 nm calmodulin (Fig. 2B), reinforcing the idea that calmodulin induces dissociation of the complex and consequent activation of ATPase activity.

Figure 4.

 Dissociation of acetylated tubulin/PMCA complex by calmodulin. PMVs were incubated in the presence of the indicated concentrations of calmodulin for 20 min at 37 °C and subsequently solubilized (without prior washing of membranes) by the addition of Triton X-100 (see Experimental procedures). Aliquots were immunoprecipitated with anti-(acetylated tubulin) Ig bound to Sepharose beads. (A) Typical immunoblots of precipitated (P) and soluble (S) fractions were revealed with 5F10 Ig (upper) or 6-11B-1 Ig (lower). (B) Absorbance values corresponding to PMCA and acetylated tubulin detected in the precipitated fractions shown in (A). Values are mean ± SD from three independent experiments.

Effect of calmodulin on PMCA activity and acetylated tubulin/PMCA complex after ethanol treatment

Ethanol and calmodulin are known to have additive stimulatory effects on PMCA [9,12,13]. From a mechanistic point of view, it is of interest to determine whether the decrease in the quantity of acetylated tubulin/PMCA complex determined separately for each effector was additive (Fig. 2). PMVs were treated with 0 or 0.6% ethanol and aliquots were incubated in the presence of increasing concentrations of calmodulin, followed by the immediate determination of PMCA activity and the amount of acetylated tubulin remaining in the membrane (by measuring the acetylated tubulin/PMCA complex). The stimulatory effects of calmodulin and ethanol on PMCA activity were additive (Fig. 5A). Acetylated tubulin bands corresponding to the complex quantified under each experimental condition and the densitometric values for these bands are shown in Figs 5B,C, respectively. When PMVs were treated with both effectors, the amount of complex was significantly less than when the effectors were tested separately. Treatment with individual effectors (0.6% ethanol or 72 nm calmodulin) resulted in ∼ 37% non-dissociated complex (Fig. 5C) and ∼ 90% stimulation of PMCA activity (Fig. 5A). After the addition of calmodulin to ethanol-treated PMVs, PMCA activity was stimulated to a higher degree than expected based on the degree of dissociation of the complex. Treatment with 0.6% ethanol and subsequently with 72 nm calmodulin resulted in a reduction of complex quantity to 9% of control values (Fig. 5C), and stimulation of ∼ 300% relative to the PMCA activity of untreated PMVs (Fig. 5A). This stimulation cannot be explained solely by dissociation of the complex. As mentioned above, PMCA that does not form part of the complex makes up ∼ 33% of total (Fig. 1). If this 33% represents enzyme in the active state, then maximum stimulation could be achieved when the remaining 67% molecules were dissociated. This would be a maximal stimulation of ∼ 200%. Therefore, the observed stimulation of 300% resulting from ethanol and calmodulin treatment would involve an additional activating mechanism.

Figure 5.

 Additive effects of ethanol and calmodulin. PMVs (70 μg protein in 275 μL final volume) were incubated in the presence or absence of 0.6% ethanol for 5 min at 37 °C. Calmodulin was then added to the indicated concentrations, incubation continued for 10 min and PMVs were washed by centrifugation/resuspension. Aliquots were separated for PMCA activity assay (A) and for western blot analysis with anti-(acetylated tubulin) Ig (B). (C) Densitometry values (mean ± SD) of immunoblots from three independent experiments similar to that in (B). The absorbance value for the acetylated tubulin band corresponding to PMVs treated with 0% ethanol and 0% calmodulin was arbitrarily set at 100%.

PMCA activity in PMVs is inhibited by exogenous acetylated tubulin

Because dissociation of the acetylated tubulin/PMCA complex results in stimulation of PMCA activity (Fig. 2A,B), active PMCA may be inhibited by the addition of acetylated tubulin. To test this possibility, we incubated PMVs with 0.6% ethanol to dissociate the complex and washed the PMVs to eliminate released acetylated tubulin. Tubulin-free PMVs were then incubated with purified brain tubulin containing two different proportions (differing by a factor of four) of the acetylated isoform. As shown in Fig. 6B, PMCA activity decreased as PMVs were incubated with increasing concentrations of exogenous tubulin. The proportion of acetylated tubulin correlated directly with the degree of inhibition, indicating that acetylated tubulin is the isoform that interacts with the enzyme. After eliminating excess exogenous tubulin by centrifugation, estimation of the acetylated tubulin/PMCA complex was performed by solubilizing membranes with detergent followed by immunoprecipitation with 6-11B-1 Ig bound to Sepharose beads. Western blots of precipitated and soluble fractions revealed with 5F10 Ig and 6-11B-1 Ig showed that, in the absence of exogenous tubulin, PMCA was mostly not associated with acetylated tubulin because PMCA remained in the soluble fraction following immunoprecipitation (Fig. 6A). This was expected because the acetylated tubulin/PMCA complex was dissociated when PMVs were previously treated with ethanol. Figure 6A also shows that as PMVs were incubated with increasing concentrations of exogenous tubulin, the amount of PMCA increased in the precipitates with a corresponding decrease in the soluble fractions. Taken together, these results indicate that association of acetylated tubulin with PMCA leads to inhibition of the enzyme activity. Western blot analysis of purified tubulin (containing a low proportion of the acetylated isotype) used in this experiment showed it to be free of PMCA (Fig. 6C).

Figure 6.

 Effect of exogenous tubulin on PMCA activity and on tubulin/PMCA complex. PMVs (0.25 mg protein) pretreated with 0.6% ethanol and washed by centrifugation/resuspension to eliminate tubulin dissociated from the complex were incubated for 30 min at 37 °C in a final volume of 1 mL transport buffer, in the presence of various amounts of purified tubulin preparations containing a low or high proportion of the acetylated isotype. We checked that under these incubation conditions tubulin is not assembled into microtubules. After incubation, samples of PMVs that were incubated with preparations containing a low proportion of acetylated tubulin were centrifuged at 100 000 g for 20 min at 37 °C to eliminate excess exogenous tubulin and the pellets resuspended in the original volume with NaCl/Tris containing 0.5% Triton X-100 and immunoprecipitated with anti-(acetylated tubulin) Ig bound to Sepharose beads. Precipitated (P) and supernatant (S) fractions were immunoblotted and revealed with 5F10 Ig and 6-11B-1 Ig (A). Samples incubated with tubulin containing low (○) and high (•) proportions of acetylated tubulin were processed to determine PMCA activity (B). PMCA activity in the absence of exogenous tubulin was 40.6 ± 2 pmol Ca2+·min−1·mg−1 protein. Values are mean ± SD from three experiments. (C) Amount of PMCA in brain membranes (Control) and in 50 μg of purified tubulin preparation containing low proportion of the acetylated isotype.

Effect of tubulin-interacting drugs on PMCA activity and the quantity of acetylated tubulin/PMCA complex

Tubulin is the structural monomer of microtubules. We examined the effects of taxol and nocodazole on PMCA activity and the acetylated tubulin/PMCA complex because these compounds are known to stabilize or disintegrate microtubules. PMVs were incubated in the presence or absence of taxol or nocodazole, followed by determination of PMCA activity as 45Ca transport activity; the amount of acetylated tubulin complex in the detergent phase after partition in Triton X-114 was also measured. Taxol decreased the amount of complex and stimulated PMCA activity (Table 1). Interestingly, following treatment of PMVs with taxol, calmodulin treatment did not stimulate PMCA activity further (data not shown). Nocodazole partially dissociated the tubulin/PMCA complex and inhibited PMCA activity. Inhibition was presumed to result from an intrinsic property of nocodazole separate from its complex-dissociating capacity. This was tested by determining the effect of nocodazole on PMCA activity in PMVs pretreated with ethanol or calmodulin. Treatment of PMCA with ethanol or calmodulin for 30 min resulted in an ∼ 90% increase in enzyme activity. By contrast, when nocodazole was added following ethanol or calmodulin treatment, PMCA activity measured after 30 min incubation was reduced to 20–27% (Table 2). Thus, nocodazole is effectively an inhibitor of PMCA.

Table 1.   Effect of tubulin-interacting drugs on plasma membrane Ca2+-ATPase (PMCA) activity and quantity of acetylated tubulin/PMCA complex. Plasma membrane vesicles (PMVs) (120 μg protein) were incubated in 500 μL transport buffer in the absence (control) or presence of 50 μm nocodazole or 5 μm taxol. PMVs were incubated for 30 min at 37 °C and washed twice by centrifugation/resuspension to eliminate tubulin that was dissociated by the effectors. Aliquots (200 μL) were processed to quantify tubulin/PMCA complex by partition with Triton X-114, and to determine PMCA activity as described in Experimental procedures. Data are mean ± SD from three independent experiments.
Treatment of PMVsTubulin/PMCA complex (% of control)PMCA activity (% of control)
None (control)100100
+Nocodazole 33 ± 13 17 ± 11
+Taxol 34 ± 6 189 ± 20
Table 2.   Inhibitory activity of nocodazol on plasma membrane Ca2+-ATPase (PMCA) activity. Plasma membrane vesicles (PMVs) (120 μg protein) were incubated in 500 μL transport buffer in the absence (control) or presence of 72 nm calmodulin or 0.6% ethanol for 30 min at 37 °C. Aliquots (200 μL) were processed to determine PMCA activity as described in Experimental procedures. In other samples, after 30 min incubation with calmodulin or ethanol, nocodazole (50 μm, final concentration) was added and incubation continued for 30 min, followed by PMCA activity assay. Data are mean ± SD from three independent experiments.
ConditionPMCA activity (% of control)
Control (no treatment)100
+ calmodulin190 ± 13
+ ethanol185 ± 22
+ calmodulin + nocodazole 27 ± 09
+ ethanol + nocodazole 20 ± 12

Effect of ethanol treatment on the calcium-pumping activity of PMCA in living cells

To study the physiological relevance of PMCA activity regulation based on association/dissociation of the PMCA/tubulin complex, we used the Fura-2 method to estimate the Ca2+ concentration in living cells and how the concentration varied following ethanol treatment. We used CAD cells from a mouse brain tumor that proliferate in serum-containing medium, but stop dividing and differentiate into neurons when placed in serum-free medium. These cells contain little or no acetylated tubulin [7]. It is known that treatment of these cells with Trichostatin A (TSA; a non-specific inhibitor of deacetylases) leads to a significant increase in acetylated tubulin [7]. We suspected that this incresase in tubulin acetylation correlates with acetylated tubulin/PMCA complex formation and inhibition of enzymatic activity, because a similar effect on another P-type ATPase has been described previously [7]. We therefore measured PMCA activity (as calcium-pumping activity) in cells treated and not treated with TSA. The acetylated microtubule content of TSA-treated CAD cells increased significantly (Fig. 7A). In cells not treated with TSA, acetylated tubulin was absent and only PMCA was precipitated by anti-PMCA Ig bound to Sepharose beads, regardless of whether cells were treated with ethanol (Fig. 7B). In TSA-treated cells, PMCA and acetylated tubulin were both precipitated, indicating the presence of PMCA/acetylated tubulin complex in the membrane. When these cells were treated with ethanol, acetylated tubulin was not found in the precipitate, indicating dissociation of the complex (Fig. 7B).

Figure 7.

 Effect of ethanol on PMCA activity and acetylated tubulin/PMCA complex in CAD cells in culture. (A) CAD cells were grown to 70% confluence on coverslips and treated for 6 h with (+TSA) or without (−TSA) 5 μm TSA, and acetylated microtubules were visualized by immunofluorescence using mAb 6-11B-1. (B) CAD cells were grown on 10 cm dishes, treated (or not) with TSA as described in (A), and incubated for 20 min in transport buffer containing (+) or not (−) 0.6% ethanol. After elimination of incubation buffer, cells were solubilized with NaCl/Tris-Triton and immunoprecipitated with anti-PMCA Ig bound to Sepharose beads as described in Experimental procedures. Typical immunoblots of precipitated materials revealed with anti-PMCA Ig or anti-(acetylated tubulin) Ig (Ac-tub) are shown. (C) Intracellular calcium as a function of incubation time was estimated in CAD cells (±TSA) as relative fluorescence intensity, using Fura-2AM as the indicator (for details see Experimental procedures). Calcium ionophore A23187 (final concentration 3 μm) and ethanol (final concentration 0.6%) were added at the times indicated by arrows. This experiment was performed in the absence (continuous lines) or presence (scattered points) of 2 mm sodium vanadate, a potent inhibitor of P-type ATPases, added at time zero.

TSA-treated and non-treated cells were analyzed for internal Ca2+ concentration and its variation after the addition of effectors. To determine whether the variation in Ca2+ concentration was due to P-type ATPases, measurements were carried out in the presence or absence of sodium vanadate (a potent inhibitor of P-type ATPases). The Ca2+ concentration was lower in cells lacking acetylated tubulin (−TSA) than in cells containing acetylated tubulin (Fig. 7C; compare −TSA and +TSA, time zero). We ascribe this difference to a higher calcium-pumping activity in cells in which PMCA was not inhibited (due to the absence of acetylated tubulin). Addition of A23187 (a calcium ionophore) did not modify the Ca2+ concentration in cells not treated with TSA in the absence of vanadate (continuous line). This seemingly unexpected result may be explained by the high PMCA activity in these cells (as it is not associated with acetylated tubulin) which counteracts the influx of calcium due to the ionophore. This was supported by the finding that when PMCA was inhibited by vanadate (scattered points in Fig. 7C, −TSA), addition of A23187 increased the internal Ca2+ concentration. Subsequent addition of 0.6% ethanol did not modify the Ca2+ concentration. This is compatible with the observation that no tubulin/PMCA complex is present in cells not treated with TSA (Fig. 7B). In TSA-treated cells, addition of A23187 resulted in an increased internal Ca2+ concentration even in the absence of vanadate (Fig. 7C, +TSA), presumably because PMCA was inhibited by its association with acetylated tubulin. This high cytoplasmic Ca2+ concentration decreased abruptly upon the addition of ethanol in the absence of vanadate. By contrast, the addition of ethanol had no effect in the presence of vanadate, indicating that a P-type ATPase was involved in the decrease in Ca2+ concentration. Even when, due to the complexity of living cells and the non-specificity of vanadate, other explanations can be drawn, these results coincide exactly with our presumption that ethanol induces dissociation of the acetylated tubulin/PMCA complex with a consequent activation of PMCA.

ATPase activities are crucial in the reception/transmission of signals at the membrane level. Endogenous activators of these cation pumps, for example adducin in the sodium pump [14] and calmodulin for PMCA, are therefore important factors in the regulation of signaling pathways. In this context, acetylated tubulin (or acetylated microtubules?) is the first described endogenous ATPase inhibitor.

Discussion

The plasma membrane Ca2+ pump removes Ca2+ from the cell during intracellular signaling. Calcium, an early-response second messenger, plays a key role in a number of physiological processes including cell proliferation, differentiation and apoptosis [15–17]. PMCA can be activated by several factors, including acidic phospholipids, proteolysis, calmodulin and ethanol. Our findings show that PMCA is partially associated with acetylated tubulin and this association results in inhibition of its activity, as estimated by Ca2+-transport. Several pieces of evidence support the existence of an acetylated tubulin/PMCA complex in the membrane. (a) PMCA and acetylated tubulin are present in isolated membranes. (b) When PMVs are solubilized in detergent and partitioned in Triton X-114, PMCA and acetylated tubulin partition to the detergent phase even though acetylated tubulin is a hydrophilic protein that partitions in the aqueous phase. (c) When membranes are solubilized with Triton X-100 and subsequently immunoprecipitated with 6-11B-1 Ig bound to Sepharose beads, PMCA precipitates in addition to acetylated tubulin. PMCA does not precipitate under these conditions if Sepharose beads are bound to an irrelevant antibody; this rules out the possibility that PMCA was detected in the precipitate as an artifact or because it was bound to a sedimentable structure rather than to acetylated tubulin. (d) When membranes are solubilized with Triton X-100 and subsequently immunoprecipitated with 5F10 Ig bound to Sepharose beads, acetylated tubulin precipitates in addition to PMCA. (e) When acetylated tubulin-depleted membranes (by ethanol treatment) are solubilized with detergent and immunoprecipitated with anti-(acetylated tubulin) Ig bound to Sepharose beads, PMCA does not precipitate. However, when membranes have been incubated previously with purified exogenous tubulin (Fig. 6), PMCA does precipitate. The need for the presence of exogenous tubulin for PMCA to sediment indicates that a complex forms between PMCA and tubulin. (f) The complex is not present in membranes from cells lacking acetylated tubulin, however, it appears when the cells have acetylated tubulin (Fig. 7B). (h) Other P-type ATPases (sodium and proton pumps) have been shown to interact with acetylated tubulin [2,3,5].

Considering that in a classical PMV preparation ∼ 40% of the vesicles are of the inside-out type, the increase in complex formation after the addition of exogenous tubulin should be smaller (Fig. 6) and the ability of calmodulin to induce complex dissociation should be less pronounced (Figs 2 and 5). It is possible that the percentage of inside-out vesicles obtained by other authors is smaller than that obtained by us because of differences in the experimental conditions used. Another possibility is that almost all inside-out recircularized vesicles have acetylated tubulin, whereas inside-in circularized vesicles do not; in fact, there are no proteomic studies about different recircularized vesicle populations.

It should be noted that it is not clear from these experiments whether PMCA interacts directly with tubulin or via intermediary compounds. We use the term ‘tubulin/PMCA complex’ for simplicity. Preliminary results from our laboratory in relation to the sodium pump indicate that acetylated tubulin interacts directly with a cytoplasmic fragment of the ATPase (G. G. Zampar, M. E. Chesta, N. L. Chanaday, N. M. Díaz, A. Carbajal, C. H. Casale & C. A. Arce, unpublished data). This may also be the case for PMCA.

When the tubulin/PMCA complex is dissociated, tubulin no longer partitions into the detergent phase, but rather into the aqueous phase. Two lines of evidence support the conclusion that the decrease in acetylated tubulin in the detergent phase upon incubation of PMVs with ethanol or calmodulin (Fig. 2A,B) is due to dissociation of the tubulin/PMCA complex. First, acetylated tubulin partitions into the detergent phase when bound to Na+,K+-ATPase and into the aqueous phase when dissociated from the complex [1,2]. Second, the decrease in acetylated tubulin in the detergent phase was correlated with a reduction in the amount of complex in the immunoprecipitate. That dissociation of this complex (decrease in acetylated tubulin in the detergent phase) by ethanol or calmodulin results in stimulation of PMCA activity (Fig. 2A,B) indicates both that the interaction of tubulin with PMCA inhibits enzyme activity and that the hydrophobic behavior of acetylated tubulin is due to its association with a hydrophobic compound (PMCA or a hydrophobic complex containing PMCA). Dissociation of the complex by ethanol and calmodulin was confirmed by immunoprecipitation experiments (Figs 3 and 4). Again, although these experiments show that acetylated tubulin forms part of the same complex as PMCA, we have no direct evidence that the two molecules interact directly with each other. PMCA has been shown to interact with various membrane proteins, forming complex arrays [18–20]. Our results suggest that acetylated tubulin is a typical cytoplasmic component capable of interacting with such a membrane complex, and that association/dissociation of the complex regulates PMCA calcium-transporting activity.

Two other findings support the view that PMCA is associated with acetylated tubulin, and is thereby inactived: (a) when the complex was dissociated by taxol (Table 1), PMCA activity was stimulated; and (b) the exogenous addition of acetylated tubulin inhibited PMCA activity in PMVs (Fig. 6). Stimulation of PMCA by taxol seems to proceed via a mechanism similar to that for calmodulin, because the subsequent treatment of PMVs with 72 nm calmodulin had no additional effect (data not shown).

Most experiments in this study indicated that the interaction of acetylated tubulin with PMCA results in the inhibition of enzyme activity, and that dissociation of the complex reactivates the enzyme. A possible alternative explanation is that calmodulin or ethanol dissociates the complex, and that PMCA activation results, not from dissociation of the complex, but from the direct influence of these effectors on the PMCA molecule. However, the fact that three different chemicals (ethanol, calmodulin and taxol) dissociate the tubulin/PMCA complex, and coincidentally activate PMCA, supports the idea that activation proceeds via dissociation of the tubulin/PMCA complex. The inhibition of PMCA activity seen when the complex was dissociated by nocodazole (Table 1) seems to contradict this. However, the inhibition seen in this case was shown to be due to an inhibitory effect of nocodazole itself (Table 2).

Even when activation of purified PMCA by calmodulin independent of other proteins has been reported [21], it should be noted that the systems we used in this study, PMVs and whole cells, are more complex than purified PMCA. In effect, PMVs could contain, in addition to PMCA, multiple components such as membrane proteins or tubulin (or microtubules) and others cytoplasmic elements. Therefore, it is possible that calmodulin exerts a double activating effect on PMCA by dissociating the acetylated tubulin/PMCA complex and directly activating PMCA. This is in line with the apparently excessive stimulation of PMCA activity (300%) in relation to the amount of acetylated tubulin/PMCA complex dissociated when PMVs were treated with ethanol and calmodulin (Fig. 5). Another interesting possibility is that calmodulin exerts its stimulating effect directly on PMCA and that the ‘activated state’ of PMCA is not favorable for interaction (low affinity) with acetylated tubulin, so that this tubulin species dissociates from the complex.

The association of acetylated tubulin with PMCA and the simultaneous inhibition of the enzyme, or conversely the stimulation of enzyme activity due to dissociation of the complex, represents a novel mechanism for regulating PMCA activity and consequently the calcium concentration of the cell. The microtubular system (tubulin and/or microtubules) is clearly involved in this mechanism because nocodazole and taxol affect the integrity of the PMCA/acetylated tubulin complex and the ATPase activity (Tables 1 and 2). However, additional studies are needed to obtain a detailed description of this scenario.

The association/dissociation of PMCA and acetylated tubulin and the corresponding inhibition/stimulation of PMCA activity seem to operate in living cells. In effect, the cytoplasmic calcium concentration in CAD cells was altered by ethanol treatment accompanied by events consistent with the proposed mechanism. According to this, cells lacking acetylated tubulin could not form the PMCA/tubulin complex and this was seen to be the case (Fig. 7A,B). However, the complex was shown to be present in the membranes of cells containing acetylated tubulin (Fig. 7A,B). Furthermore, in cells in which PMCA was in an inhibited state due to its interaction with acetylated tubulin, the cytoplasmic calcium concentration (Fig. 7C, zero time) was higher than in cells having PMCA in an non-inhibited state (because it is not interacting with acetylated tubulin). In the presence of sodium vanadate, the A23187 ionophore was able to increase the calcium concentration and further treatment with ethanol did not result in a change in the concentration (Fig. 7C, −TSA). This was expected on the basis that there was no PMCA to activate because it is not interacting with acetylated tubulin. In cells with the complex (Fig. 7C, +TSA), following the increase in calcium concentration by A23187, ethanol treatment induced an abrupt decrease in concentration, coincident with enzyme activation due to dissociation of the complex. The blockade of this decrease in calcium concentration by vanadate demonstrates that the rapid efflux of calcium was due to a P-type ATPase. It is clear that treatment of cells with TSA led to increased tubulin acetylation which correlated with the inhibition of PMCA, but TSA, as a non-specific inhibitor of deacetylases, is also able to induce changes in transcriptional activities and therefore, the observed effects could be indirect.

We have previously shown that the acetylated isotype of tubulin is required for interaction with the sodium pump [7]. Here, we demonstrated that the same tubulin isotype is necessary for interaction with the calcium pump and that the isoform PMCA4 is involved in the complex (Fig. 1). It is possible that acetylated tubulin interacts and regulates the activity of all P-ATPases. With this aim, we are currently trying to identify the isoforms that form complexes for each of the P-ATPases, to characterize more precisely the ‘ATPase/acetylated tubulin complex’, and to identify the interacting domains of the ATPase and acetylated tubulin.

Experimental procedures

Materials

Triton X-114, ATP, anti-mouse IgG conjugated with peroxidase, calmodulin, mouse mAb (ascites fluid) 6-11B-1 specific for acetylated tubulin, mAb 5F10 specific for PMCA and mAb JA9 specific for PMCA4 were from Sigma Chemical Co (St Lousi, MO, USA) and mAb PI3-kinase p110 (D-4) was from Santa Cruz Biochnology (Santa Cruz, CA, USA). 45CaCl2 (12 mCi·mg−1) was from Perkin-Elmer Life Science (Wellesley, MA, USA) and Fura-2AM was from Molecular Probes (Eugene, OR, USA).

Cell culture

CAD cells were cultured in Dulbecco’s modified Eagle’s medium/F12 (1 : 1; Sigma) supplemented with 10% fetal bovine serum (Carlsbad, CA, USA) at 37 °C in an air/CO2 (19 : 1) incubator. The culture medium was renewed every 48 h.

Immunofluorescence

CAD cells were grown on coverslips and fixed with anhydrous methanol at −20 °C. Samples were rehydrated, incubated with 2% BSA in NACl/Pi for 60 min, and stained by indirect immunofluorescence using mouse mAb 6-11B-1 (dilution 1 : 600) in NaCl/Pi containing 1% BSA. Fluorescein-conjugated anti-mouse IgG (dilution 1 : 400) was used as secondary antibody. Coverslips were mounted on FluorSave and observed for epifluorescence with a confocal Zeiss LSM microscope.

Isolation of brain plasma membrane vesicles

Isolation of plasma membrane vesicles from rat brain was based on the method of Michaelis et al. [22], and modified for optimal results. All procedures and treatments for handling animals were reviewed and approved by the Comité de Etica of CONICET (res. number 1806/04). Seven rat brains (∼ 10 g) were homogenized at 4 °C in 10 vol of 10 mm Hepes/KOH, pH 7.4, 0.32 m sucrose, 0.5 mm MgSO4, 0.1 mm phenylmethanesulfonyl fluoride, 2 mm 2-mercaptoethanol. The homogenate was centrifuged at 1500 g for 10 min and the supernatant was centrifuged at 20 000 g for 20 min. The resulting pellet was resuspended in homogenization buffer to obtain a protein concentration of 10 mg·mL−1 in ∼ 6 mL. Samples of 1 mL were layered onto a discontinuous gradient containing 3 mL of 40% (w/v) sucrose and 3 mL of 20% sucrose and centrifuged at 63 000 g for 45 min at 2–4 °C. The synaptosome fraction was obtained at the interface, washed in 25 vol of 10 mm Hepes/KOH, pH 7.4, centrifuged at 20 000 g for 30 min, collected in the pellet and resuspended in 10 mm Hepes/KOH, pH 7.4 to give a protein concentration of ∼ 14 mg·mL−1. An aliquot (1 mL) of the synaptosome fraction was incubated for 40 min at 4 °C, in 100 vol of lysis buffer (10 mm Hepes/KOH, pH 7.4, 1 mm EDTA, 2 mm 2-mercaptoethanol) with continuous stirring. The lysate was centrifuged at 20 000 g for 30 min to give a pellet containing PMVs. This fraction was resuspended in 2 mL of 10 mm Hepes/KOH, pH 7.4 (final protein concentration 5 mg·mL−1) and stored at −70 °C until use.

PMCA activity assay

PMCA activity was determined using the 45Ca2+ transport assay as described previously [8]. The reaction mixture contained PMVs (50–60 μg protein) in 300 μL transport buffer [50 mm Tris/HCl, pH 7.3, 100 mm KCl, 95 μm EGTA, 5 mm NaN3, 400 nm thapsigargin, 20 mm sodium phosphate, 2.5 mm MgCl2, and 45CaCl2 (1 × 105 dpm·nmol−1) to obtain a free Ca2+ concentration of 10 μm]. Free Mg2+ and Ca2+ concentrations were calculated using the program described by Fabiato and Fabiato [23]. After preincubation for 5 min at 37 °C, the reaction was initiated by the addition of 2 mm ATP. After 5 min, the reaction was stopped by filtering the samples through a 0.45 μm filter, and 45Ca2+ taken up by the vesicles was determined using a liquid scintillation counter. PMCA-mediated 45Ca2+ uptake was calculated as the difference in 45Ca2+ uptake between samples incubated in the presence versus absence of ATP.

Determination of calcium concentration in the cytoplasm of living cells

This was performed using the Fura-2 method of Jeremic et al. [24] with some modification. CAD cells were grown as described above and resuspended in 5 mm Hepes, pH 7.4, containing 135 mm NaCl, 5.4 mm KCl, 1.8 mm MgCl2, 10 mm d-glucose (buffer A). Cells were washed twice with buffer A by centrifugation (1000 g for 5 min at 4 °C) and resuspended to a density of 1 × 109 cells·mL−1 in buffer A containing 10 μm Fura-2AM. The suspension was incubated for 30 min at 30 °C with mild agitation. Cells were then washed twice with ice-cold buffer A (without Fura-2AM) and resuspended at a density of 1 × 109 cells·mL−1. For fluorescence measurements, each suspension was loaded into a cuvette (final density 5 × 107 cells·mL−1) in buffer A containing 140 μm CaCl2, 100 μm EGTA and 1 μm thapsigargin, and placed in a thermostated (30 °C) Fluoromax-3 Jobin Yvon-Horiva spectrofluorimeter. Excitation was at 340 and 380 nm, and emission was measured at 510 nm. The Fura-2 fluorescence response to intracellular calcium concentration ([Ca2+]i) was calibrated from the ratio of 340/380 nm fluorescence values after subtraction of background fluorescence of the cells at 340 and 380 nm, as described by Grynkiewicz et al. [25].

Isolation and determination of acetylated tubulin/PMCA complex

Acetylated tubulin/PMCA complex was isolated into Triton X-114 phase as described previously with slight modification [26]. Briefly, PMVs (50 μg protein) were washed once with NaCl/Tris (50 mm Tris/HCl buffer, pH 7.4, containing 150 mm NaCl) and immediately solubilized in 1 mL NaCl/Tris containing 1% Triton X-100. After 30 min at 0 °C, the preparation was centrifuged at 100 000 g for 15 min and Triton X-114 was added to the supernatant fraction (1% final concentration). For phase separation, the preparation was warmed for 5 min at 37 °C and centrifuged at 600 g for 5 min. The detergent-rich lower phase containing acetylated tubulin/PMCA complex was washed once with NaCl/Tris. Aliquots were subjected to electrophoresis and immunoblotting to determine acetylated and total tubulin.

Electrophoresis and immunoblotting

Proteins were separated by SDS/PAGE on 10% polyacrylamide slab gels [27], transferred to nitrocellulose and reacted with mouse mAb 6-11B-1 (dilution 1 : 1000) to determine acetylated tubulin [28], mouse mAb 5F10 (dilution 1 : 300) to determine PMCA [29] or mouse mAb JA9 (dilution 1 : 1000) to determine PMCA4 isoform [9]. The nitrocellulose sheet was reacted with anti-mouse IgG conjugated with peroxidase. Intensities of tubulin bands were quantified by scion imaging software.

Tubulin preparation

Two brain tubulin preparations containing different proportions of the acetylated isotype were obtained as described previously [2]. These preparations, referred to as tubulin of low and high acetylated isotype content, differ approximately fourfold in the proportion of acetylated tubulin.

Preparation of antibody linked to Sepharose

Monoclonal antibodies 6-11B-1, 5F10 and phosphatidylinositol 3-kinase p110 were covalently bound to CNBr-activated Sepharose 4B as described by Hubbert et al. [30], with slight modification. Sepharose beads were washed with a 100 vol excess of 0.001 m HCl at 21 °C. The resulting packed beads (1 mL) were mixed with 2.5 mg protein of ascites fluid containing 6-11B-1 antibody (or another antibody) in 1 mL coupling buffer (0.5 m NaCl containing 0.2 m NaHCO3, pH 8.2). The mixture was agitated on a platform rocker for 4 h at 21 °C and loaded into a small chromatographic column. Unbound antibodies were removed by washing with 5 mL coupling buffer. Antibody–Sepharose beads were transferred to a beaker and suspended in 1 mL of coupling buffer containing 0.2 m glycine to block unreacted Sepharose sites. The mixture was agitated for 2 h at 21 °C and unbound glycine was removed by washing the beads with 10 mL of coupling buffer. The resulting antibody-coupled Sepharose was washed with 1.5 mL of 0.01 m Tris/HCl, pH 8, containing 0.14 m NaCl and 0.025% NaN3, and stored at 4 °C until use (maximum 2 days).

Immunoprecipitation procedure

PMVs (300 μL, 5 mg protein·mL−1) were solubilized with NaCl/Tris containing 0.5% Triton X-100 (NaCl/Tris-Triton) and centrifuged to eliminate residual insoluble material. Aliquots (0.3 mL) were mixed with 0.15 mL of packed antibody [anti-(acetylated tubulin) Ig or anti-PMCA Ig–Sepharose beads] and incubated for 4 h at 20 °C. Samples were centrifuged and the precipitated material was washed five times with NaCl/Tris-Triton. Fractions (50 μL) of packed beads were resuspended in 50 μL Laemmli sample buffer, heated at 50 °C for 15 min and centrifuged. Aliquots (20 μL) of soluble fractions were subjected to SDS/PAGE. A control was run in parallel using mAb phosphatidylinositol 3-kinase–Sepharose instead of mAb 6-11B-1 or 5F10 antibody–Sepharose.

Protein determination

Protein concentration was determined by the method of Bradford [31].

Acknowledgements

We thank Dr S. Anderson for the English editing. This work was supported by grants from Agencia Nacional de Promoción Científica y Tecnológica de la Secretaría de Ciencia y Tecnología del Ministerio de Cultura y Educación en el marco del Programa de Modernización Tecnológica (BID 802 OC/AR), Consejo Nacional de Investigaciones Científicas y Técnicas (Conicet), Agencia Córdoba Ciencia (Gobierno de la Provincia de Córdoba), Secretaría de Ciencia y Técnica de la Universidad Nacional de Córdoba y Secretaría de Ciencias de la Universidad Nacional de Río Cuarto.

Ancillary