INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Ca²⁺ plays an important role in plant growth, development, and signal transduction. Many of the functions mediated by Ca²⁺ depend upon a fine-tuning of its concentration in different organelles. Therefore, the spatial regulation of Ca²⁺ concentration is essential for the proper functioning of plant cells. Ca²⁺ pumps and Ca²⁺/H⁺ antiporters appear to be the main mechanisms by which plant cells accumulate Ca²⁺ in the organelles (Fox and Guerinot, 1998; Sze et al., 2000). Ca²⁺ pumps are vital in regulating the cytosolic Ca²⁺ concentration in plant cells, and are widely distributed on membranes, including the plasma membrane (PM), vacuole, and endoplasmic reticulum (ER; Sze et al., 2000).

In animal cells, the Golgi apparatus is also able to accumulate Ca²⁺. This organelle has been reported to contain two Ca²⁺ pumps, one sensitive and the other insensitive to thapsigargin (Taylor et al., 1997; Pinton et al., 1998; Rojas et al., 2000). Studies in yeast (Saccharomyces cerevisiae) have also shown that a Ca²⁺ pump is located at the Golgi apparatus, and, interestingly, mutations in the gene encoding this Ca²⁺ pump lead to alterations in the glycosylation pattern of secreted proteins (Rudolph et al., 1989). In plants, even though some studies suggested that Ca²⁺ pump(s) may be located in the Golgi apparatus (Canut et al., 1993; Logan and Venis, 1995), no further evidence has been provided yet that this organelle may play a role in Ca²⁺ uptake in plant cells.

Plant Ca²⁺ pumps have been classified in two different classes based on biochemical studies: their comparison with animal Ca²⁺ pumps, and the effect of inhibitors and activators on their transport activities. These are: type IIA (ER-type) Ca²⁺-ATPases, and type IIB (PM-type) Ca²⁺-ATPases (Sze el al., 2000). However, regardless of their classification (ER or PM type), none of the pumps described until now have been shown to be sensitive to low concentration of the animal sarco-endoplasmic reticulum Ca²⁺ ATPase inhibitor thapsigargin.

In this paper, we analyzed the presence of a Ca²⁺ pump in the plant Golgi apparatus by carrying out Ca²⁺ uptake assays on subcellular fractions from pea (Pisum sativum) etiolated epicotyls. Upon identifying a Ca²⁺ active uptake in Golgi fractions, we performed functional (kinetic and pharmacological) studies of the active Ca²⁺ transport in an enriched vesicle fraction of the organelle. Here, we provide evidence of a Ca²⁺ active accumulation into Golgi apparatus vesicles driven by a Ca²⁺ pump that is inhibited in the nanomolar range by thapsigargin, a feature not previously found in other plant Ca²⁺-ATPases. Finally, a thapsigargin-sensitive Ca²⁺ active uptake was also detected in cauliflower (Brassica oleracea) Golgi fractions, suggesting that this calcium uptake mechanism is not pea specific.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

An ATP-Dependent Ca²⁺ Uptake Activity Colocalizes with the Golgi Apparatus

To investigate whether the plant Golgi apparatus contains an active Ca²⁺ accumulation mechanism, we measured the ⁴⁵Ca²⁺ uptake in subcellular fractions obtained from the elongating zone of etiolated pea epicotyls. To separate Golgi vesicles from other organelles (in particular ER membranes), we used Suc step gradients (0.25/1.1/1.3 M) as described in "Materials and Methods." This procedure allowed a good separation of the Golgi marker enzymes xyloglucan fucosyltransferase (XG-FucTase; Fig. 1C; Wulff et al., 2000) and UDPase (Fig. 1E; Orellana et al., 1997), from ER (Fig. 1B), PM (1D), and mitochondrial (Fig. 1F) markers; however, the tonoplast marker -TIP showed a broad distribution and some degree of overlapping with the Golgi marker was observed (Fig. 1G). Measurement across the gradient of the ATP-dependent Ca²⁺ uptake showed a narrow peak of activity that comigrated with the distribution of the Golgi marker XG-FucTase (Fig. 1A). In addition, a broad peak of ATP-dependent Ca²⁺ uptake activity was detected toward denser fractions that contained ER and PM membranes.

View larger version (37K): [in this window] [in a new window]   Figure 1.   Active Ca2+ uptake in subcellular fractions of etiolated pea epicotyls. A, 45Ca2+ uptake by subcellular fractions measured in a reaction mixture containing 1 µM free Ca2+ in the presence () and in the absence () of ATP. B, NADH cytochrome c (Cyt c) reductase activity insensitive to antimycin A. C, XG-FucTase activity. D, PM-ATPase activity measured in the presence () and in the absence () of 70 µM vanadate. E, UDPase activity measured in native gels. F, Cyt c oxidase activity. G, Detection of -tonoplast intrinsic protein (-TIP) in subcellular fractions using western blots. H, Suc concentration and total protein across the gradient, determined by refractometry and the bicinchoninic acid method, respectively. The results are the average of three independent gradients measured by triplicates. The average relative activity and its deviation are plotted. The relative activity was standardized by setting the highest value obtained from the measurements at 100% relative activity. All subsequent values were then adjusted accordingly and the deviation was calculated. The highest value on each case correspond to: A, 1.95 pmol min1 µL1; B, 196 nmol min1 µL1; C, 0.52 pmol min1 µL1; D, 6.13 nmol min1 µL1; and F, 3.92 µmol min1 µL1.

Golgi apparatus membranes were separated from tonoplast by centrifugation in a discontinuous 0.67/1.3 M Suc gradient as described in "Materials and Methods." As shown in Figure 2, this procedure allowed a significant separation of the XG-FucTase and UDPase activities (Fig. 2, C and E) from the tonoplast marker -TIP (Fig. 2F). Although the XG-FucTase and UDPase activities were located toward denser fractions, -TIP was detected toward light density fractions. Overexposure of the western blots showed some minor amount of -TIP present at high-density fractions. This may correspond to -TIP associated to the Golgi apparatus, probably in transit to the tonoplast (Jauh et al., 1999). Measurement of ATP-dependent Ca²⁺ uptake showed that most of the activity was detected around the more dense fractions of the gradient (1.14 g mL¹) that also contained XG-FucTase and UDPase activities (Fig. 2). In contrast, fractions containing tonoplast did not significantly accumulate Ca²⁺ under our experimental conditions. The distribution of the Ca²⁺ uptake activity did not follow the same pattern of XG-FucTase and UDPase activities distribution; however, this result could be explained by the compartments found within the Golgi apparatus (Dupree and Sherrier, 1998) that can lead to differential patterns of enzymes distribution upon separation in density gradients. These findings strongly suggest that Golgi apparatus vesicles contain an active Ca²⁺ uptake mechanism. Because Ca²⁺ uptake by the tonoplast was not significant, to avoid the loss of Golgi vesicles due to the fractionation procedure, most of the following experiments were done in an enriched Golgi fraction that contained some level of tonoplast.

View larger version (32K): [in this window] [in a new window]   Figure 2.   Separation of Golgi apparatus membranes and active Ca2+ uptake from tonoplast. A, ATP-dependent Ca2+ uptake in subcellular fractions measured in the presence of 1 µM free Ca2+. B, Suc concentration across the gradient determined by refractometry. C, XG-FucTase activity. D, Total protein across the gradient. E, UDPase activity measured in native gels. F, Detection of -TIP using western blots. The measurements were done in triplicate. The average relative activity and its deviation were plotted. The relative activity was standardized by setting the highest value obtained from the measurements at 100% relative activity. All subsequent values were then adjusted accordingly and the deviation was calculated. The highest value on each case correspond to: A, 0.314 pmol min1 µL1; and C, 0.13 pmol min1 µL1.

The Active Ca²⁺ Uptake Mechanism Has a High Affinity for Ca²⁺ and Is Inhibited by Cd²⁺ and Cu²⁺

To functionally characterize this Ca²⁺ transport system, ⁴⁵Ca²⁺ uptake assays (Fig. 3) were carried out using an enriched Golgi apparatus-containing vesicle fraction. As shown in Figure 3A, Ca²⁺ uptake was stimulated by ATP and was negligible in its absence. The Ca²⁺ accumulated in the vesicles was only partially released (50%) by the Ca²⁺ ionophore A23187 (Fig. 3B), probably because Ca²⁺ binds to lumenal macromolecules. The kinetic parameters of pea Golgi apparatus active Ca²⁺ transport were determined measuring the initial rates of ⁴⁵Ca²⁺ uptake at different free Ca²⁺ concentrations (Fig. 3C). The calculated apparent K_m for Ca²⁺ was 209 nM (pCa 6.68) and the rate of maximum Ca²⁺ uptake was 2.5 nmol mg¹ min¹.

View larger version (9K): [in this window] [in a new window]   Figure 3.   Active Ca2+ uptake by Golgi vesicles. A, Time course of Ca2+ incorporation into Golgi apparatus vesicles measured in the incubation buffer containing 300 nM free Ca2+ in the presence () and in the absence () of ATP. B, Time course of Ca2+ uptake and release by the Ca2+ ionophore A23187. , Control; , ionophore. Free Ca2+ concentration was 210 nM. Values are mean ± SE. The arrows indicate the time when the ionophore was added. C, Ca2+ dependence of active Ca2+ uptake in Golgi apparatus-enriched fraction. Ca2+ uptake initial rates were measured at different free Ca2+ concentrations that were estimated with the WinMaxC 2.05 computer program (Chris Patton, Hopkins Marine Station, Stanford University, CA). The experimental points were fitted by a nonlinear fitting program (GraphPad Prism 2.0, GraphPad Software, Inc., San Diego). Apparent Km value of 0.21 µM (pCa 6.68) was obtained from by interpolation in nonlinear fit. Experiments were performed in triplicate. Values are mean ± SE.

To test the interaction of other divalent cations with the pea Golgi Ca²⁺ pump we carried out experiments of Ca²⁺ uptake in the presence of Mn²⁺, Cd²⁺, Co²⁺, Cu²⁺, and Fe²⁺ (Fig. 4). As shown in Figure 4, Cu²⁺ and Cd²⁺ strongly inhibited Ca²⁺ uptake, whereas Mn²⁺ and Fe²⁺ had only a slight effect even at concentrations as high as 8.7 µM. On the other hand, Co²⁺ did not inhibit the pump. These results suggest that Mn²⁺ and Fe²⁺ did not appreciably compete with Ca²⁺ for the pump Ca²⁺-binding sites.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Ca2+ uptake in the presence of other divalent cations. The 45Ca2+ uptake by the pea Golgi vesicles was measured at 25°C for 1.5 min, in an incubation buffer containing 870 nM free Ca2+ (4 times pump Km), 2 mM ATP, and 3 mM MgCl2. Experiments were carried out in the absence and in the presence of increasing concentrations (870 nM, 1.47 µM, 4.35 µM, and 8.70 µM) of each of the following divalent cations: Cu2+, Co2+, Mn2+, Fe2+, and Cd2+. Values of the Ca2+ uptake at 1.5 min in the presence of different concentrations of the divalent cations are expressed as a percentage of the Ca2+ uptake measured in the absence of other cations (6.3 ± 0.2 nmol mg1).

Ca²⁺ Uptake in Golgi Apparatus Is Inhibited by Vanadate and Cyclopiazonic Acid But Not Stimulated by Calmodulin

To identify the Ca²⁺ uptake mechanism, we tested the effect of some compounds that have been widely used to characterize Ca²⁺ uptake processes (Inesi and Sagara, 1994; Sze et al., 2000). Because the Golgi apparatus is involved in active membrane-trafficking processes with the ER (Boevink et al., 1998), an organelle where Ca²⁺ uptake process is well characterized, we performed pharmacological studies in parallel using both Golgi apparatus and ER vesicle fractions. When Golgi apparatus vesicles were incubated in the presence of 200 µM vanadate, the calcium uptake was greatly reduced (Table I), suggesting that the Ca²⁺ transport activity was driven by a P-type Ca²⁺-ATPase. As expected, this compound also inhibited active Ca²⁺ uptake in ER vesicles.

The effect of cyclopiazonic acid, a specific inhibitor of animal SERCA-type (Inesi and Sagara, 1994) and (IIA) plant Ca²⁺-ATPases (Sze et al., 2000) was also studied. As shown in Table I, the addition of this compound at 100 nmol mg¹ protein completely inhibited active Ca²⁺ accumulation by both Golgi apparatus and ER vesicles, suggesting that the Golgi apparatus contains a SERCA-type or a (IIA) Ca²⁺-ATPase. This result also indicates that Ca²⁺ uptake driven by a hypothetical H⁺/Ca²⁺ antiporter present in Golgi apparatus was negligible under our experimental conditions. This result was confirmed by the finding that both bafilomycin A and carbonyl cyanide 3-chlorophenylhydrazone did not affect Ca²⁺ uptake (not shown).

It has been reported that calmodulin stimulates plant PM-type (IIB) Ca²⁺ ATPase activity present in ER (Hsieh et al., 1991; Chen et al., 1993; Hwang et al., 1997; Hong et al., 1999), PMs (Thomson et al., 1993; Askerlund, 1997; Bonza et al., 2000), and tonoplast (Gavin et al., 1993; Malmstrom et al., 1997), but does not activate plant ER-type (IIA) Ca²⁺ pumps located in ER (Liang and Sze, 1998). We found that 1 µM calmodulin had no effect on the active Ca²⁺ uptake by Golgi apparatus vesicles (Table I). In contrast, active Ca²⁺ accumulation in ER vesicles increased 2- to 3-fold in the presence of 1 µM calmodulin.

The Golgi Apparatus Ca²⁺ Pump Is Sensitive to Thapsigargin

Thapsigargin is a specific inhibitor of most animal intracellular SERCA-type Ca²⁺ pumps present in the sarcoplasmic/ER (Sagara and Inesi, 1991; Treiman et al., 1998) and Golgi apparatus (Taylor et al., 1997; Zhong and Inesi, 1998; Rojas et al., 2000). However, to our knowledge, no Ca²⁺ pump sensitive to this compound has been described in plant cells. In pea stems, thapsigargin had a clear effect on the active calcium uptake of subcellular fractions that correlated with the Golgi marker (Fig. 5). In addition, thapsigargin inhibited Ca²⁺ uptake by Golgi apparatus vesicles, whereas thapsigargin did not block the Ca²⁺ uptake by pea ER vesicles (Fig. 6). To further characterize the effects of this inhibitor on the Golgi apparatus Ca²⁺ pump, we studied the concentration dependence of transport inhibition by thapsigargin. Results showed that the apparent inhibitor concentration causing 50% inhibition (I₅₀) value for thapsigargin was 88 nM at 600 µg protein mL¹ (Table II).

View larger version (20K): [in this window] [in a new window]   Figure 5.   Effect of thapsigargin on the Ca2+ uptake by subcellular fractions from etiolated pea stems. Organelles were separated as described in Figures 1A or 2B, and the uptake of 45Ca2+ by subcellular fractions was measured in a reaction mixture containing 1 µM free Ca2+ in the presence () and absence of 2 µM thapsigargin (). The lines named Golgi in A and B indicate the distribution in the gradient of the Golgi marker XG-FucTase. The line named ER in A indicates the distribution in the gradient of the ER marker, NADH Cyt c reductase activity insensitive to antimycin A. The activity of the ER marker was negligible in B. THG, Thapsigargin.

View larger version (10K): [in this window] [in a new window]   Figure 6.   Thapsigargin inhibits the uptake of Ca2+ in Golgi but not in ER vesicles. Uptake of Ca2+ was measured using Golgi apparatus-enriched vesicles (A) and ER-enriched vesicles (B). After 1 min of incubation, 2 µM thapsigargin (THG; ) or the vehicle () were added to the incubation medium and the Ca2+ uptake was determined for another 4 min. The free Ca2+ concentration in the medium was 300 nM. Values are mean ± SE.

A Thapsigargin-Sensitive Ca²⁺ Uptake Mechanism Is Also Present in Cauliflower

To test whether a thapsigargin-sensitive Golgi Ca²⁺ pump activity was also present in other plants, we measured Ca²⁺ uptake in subcellular fractions obtained from the floral meristem of cauliflower. To this end, we fractionated the organelles in a four-step Suc gradient where, in addition to a pellet, four membranes bands were observed and collected. Ca²⁺ uptake, the effect of thapsigargin, and different marker enzymes were measured. The results (Fig. 7) showed that the distribution of a thapsigargin-sensitive Ca²⁺ uptake activity correlated with Golgi markers and not with ER, PM, tonoplast, and mitochondrial markers.

View larger version (36K): [in this window] [in a new window]   Figure 7.   A thapsigargin-sensitive Ca2+ pump is present in Golgi-enriched cauliflower subcellular fractions. Membrane fractions (1-5, depicted in I) were obtained from cauliflower as described in "Materials and Methods." Ca2+ uptake was measured in the presence and absence of thapsigargin (A). The thapsigargin-sensitive Ca2+ uptake component is shown in C. The Golgi markers Latent UDPase and XG-FucTase are shown in E and G, respectively. B, NADH Cyt c reductase activity insensitive to antimycin A (ER marker). D, Vanadate-sensitive PM-ATPase activity (PM marker). F, Detection of -TIP in membrane fractions using western blots. H, Cyt c oxidase activity. I, Suc concentration across the gradient determined by refractometry. The lines named 1 through 5 indicate the interfaces from where the membrane fractions were collected. J, Total protein measured in the membrane interfaces.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

In the present study, we functionally identified and characterized a novel Ca²⁺ pump activity present in enriched Golgi apparatus vesicles isolated from the elongation zone of etiolated pea epicotyl. We carried out a detailed subcellular fractionation of pea etiolated pea epicotyl, and measured Ca²⁺ uptake activity across the gradient. The results show that the main Ca²⁺ uptake activity present in etiolated pea stems strongly correlated with the Golgi apparatus marker enzyme XG-FucTase. The measurement of another enzymatic marker for this organelle, the Golgi UDPase, using native gels, was also in good agreement with the distribution of XG-FucTase. This confirms the correlation between this Ca²⁺ uptake activity and the Golgi markers. Subcellular fractionation studies also show that this Ca²⁺ pump activity was not localized at the ER or in the PM, and it was clearly separated from the tonoplast marker -TIP (Jauh et al., 1999). Measurements of a thapsigargin Ca²⁺ uptake activity in subcellular fractions from cauliflower floral meristem suggested that Golgi apparatus from other plants might have Ca²⁺ pumps with similar characteristics. The pea Golgi apparatus active Ca²⁺ accumulation showed high affinity for Ca²⁺ (a K_m of 209 nM). This activity was almost completely inhibited by vanadate and cyclopiazonic acid and was not stimulated by calmodulin, indicating that it was driven by a P-type calcium pump rather than by a Ca²⁺/H⁺ antiporter. Ca²⁺-ATPases with these properties have been found in several endomembranes in different plant tissues and have been classified as type IIA (ER type) to indicate that they share similar characteristics to the SERCA animal Ca²⁺-ATPases (Sze et al., 2000). The calcium uptake was not inhibited by Mn²⁺, suggesting that this cation does not compete with the uptake of Ca²⁺. Therefore, in contrast to other plant Ca²⁺ pumps, this Ca²⁺ pump would not transport Mn²⁺. The inhibition caused by Cd²⁺ and Cu²⁺ could be due to competition for the Ca²⁺ uptake; however, the decrease in Ca²⁺ transport may be also explained by a lock of the pump, or alternatively due to alterations in the redox state of key residues of the pump (Zhang et al., 1990).

One of the most striking findings of this work was the fact that the Golgi apparatus Ca²⁺ pump was inhibited by thapsigargin, the most widely used specific inhibitor of SERCA-type Ca²⁺-ATPases, with an apparent I₅₀ of 88 nM at a protein concentration of 600 µg mL¹. This apparent I₅₀ value could be lower if less protein is used in the assay because Sagara and Inesi (1991) reported that the thapsigargin concentration dependence of Ca²⁺ transport inhibition is an apparent function of the concentration of protein in the reaction mixture. To our knowledge, this is the first report of a plant Ca²⁺ pump inhibited by lownanomolar concentrations of this compound. Other studies in plants had shown no effect of thapsigargin on Ca²⁺ pumps (Liang and Sze, 1998), or an effect at high concentrations that are nonselective (Thomson et al., 1994). In the animal Golgi apparatus membrane, the presence of two Ca²⁺ pumps has been reported: one sensitive and the other insensitive to thapsigargin (Taylor et al., 1997; Pinton et al., 1998; Rojas et al., 2000). Our results suggest that the plant Golgi-localized Ca²⁺ pump may have some relationship to the thapsigargin-sensitive animal Golgi calcium pump.

Studies on ER-localized Ca²⁺ pumps in other plants suggested the presence of two classes of Ca²⁺ pumps classified as IIB (PM type) and IIA (ER type; Liang and Sze, 1998; Hong et al., 1999). These two types of pumps are thapsigargin insensitive. Our results indicate that Ca²⁺ uptake by pea ER vesicles is thapsigargin insensitive, stimulated in the presence of calmodulin, and inhibited by cyclopiazonic acid. These results suggest that pea ER contain the IIA and IIB Ca²⁺ pumps, which are different from the Golgi-localized Ca²⁺ pump.

Recently, Liang et al. (1997) showed that the Arabidopsis Ca²⁺-ATPase ECA1 was able to complement the yeast mutant defective in Golgi (Pmr1) and vacuolar (Pmc1) Ca²⁺ pumps. Even though most of the ECA1 protein fractionated along with the ER, they suggest that some of this protein could be localized in the Golgi. Because ECA1 is not inhibited by thapsigargin, we think it is unlikely that the pea ECA1-orthologous gene encodes the pea Golgi Ca²⁺-ATPase that we have identified in this work.

In this work, we have kinetically and pharmacologically characterized this Ca²⁺-ATPase. This work sets the stage whereby a molecular analysis of this Golgi-localized thapsigargin-sensitive Ca²⁺ ATPase activity can be molecularly characterized. Future investigation will reveal whether this Ca²⁺ pump is orthologous to one of the members of the P-type ion pump subfamilies recently identified in the Arabidopsis genome (Axelsen and Palmgren, 2001).

What is the role of this Golgi Ca²⁺ pump? There are several potential roles for the Ca²⁺ accumulated by the pump in the Golgi apparatus lumen. One possibility is that similar to what occurs in yeast, calcium may play a role in protein glycosylation, and secretion of both proteins and polysaccharides. This may suggest that the role of Ca²⁺ in protein glycosylation and secretion of proteins/polysaccharides may be an evolutionary conserved mechanism among yeast, mammals, and plants. In addition, Ca²⁺ is likely to participate in the fusion and fission of vesicles that are in transit in the secretory pathway. On the other hand, the Golgi may serve as a calcium store, like the ER and vacuole. Eventually, this Ca²⁺ may be released upon activation of signaling pathways. Ca²⁺ may also be required by lumenally located enzymes that use it as a cofactor. Alternatively, Ca²⁺ may also bind to polysaccharides that are synthesized in the Golgi and then secreted to the cell wall, like pectins. All of these hypotheses about the role of Ca²⁺ in the Golgi, as well as the molecular nature of the Ca²⁺ transporters, remain to be analyzed. Identification of the gene(s) encoding for this (these) transporter(s) will aid in clarify the putative roles of the thapsigargin-sensitive Ca²⁺ pump located in the Golgi apparatus of etiolated pea epicotyls and cauliflower.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material

Pea (Pisum sativum var Alaska) seedlings were grown in moist vermiculite for 7 to 8 d in the dark at 25°C. Stem segments of 1 to 3 cm were excised from the elongating epicotyls, and kept on ice until homogenization. Cauliflower (Brassica oleracea) was obtained from local markets.

Subcellular Fractionation of Pea Stems and Preparation of Golgi and ER Vesicles

Pea stem segments (40-70 g) were minced by hand using razor blades in the presence of 1 volume of 0.5 M Suc (Suc), 0.1 M KH₂PO₄ (pH 6.65), 5 mM MgCl₂, and 1 mM dithiothreitol (added fresh). After the tissue was completely chopped, it was homogenized for 3 min in a mortar. Golgi apparatus vesicles were obtained by step Suc gradient centrifugation following the method described by Muñoz et al. (1996). All isolation procedures were done on ice. The tissue homogenate was filtered through Miracloth (Calbiochem-Novabiochem, San Diego) and centrifuged at 1,000g for 5 min. The supernatant was layered on 8 mL of 1.3 M Suc cushion and centrifuged at 100,000g for 90 min. The upper phase was removed without disturbing the interface fraction and Suc layers of 1.1 and 0.25 M were overlaid on the membrane pad. The Suc gradient was then centrifuged at 100,000g for 100 min. Fractions of 1 mL were collected from the top of the gradient and immediately used in enzymatic and ⁴⁵Ca²⁺ uptake assays. To prepare Golgi and ER vesicles for transport assays, fractions containing the maximum enzymatic activity for Golgi and ER markers were collected. These membrane fractions were collected and diluted with cold distilled water to twice the volume. Both suspensions were centrifuged at 100,000g for 50 min. The pellets were gently resuspended using a Dounce homogenizer in a buffer containing 0.25 M Suc, 1 mM MgCl₂, and 10 mM Tris-HCl (pH 7.5). The vesicles were aliquoted and stored at 70°C until used.

Golgi Apparatus and Vacuole Membrane Separation

To further separate Golgi apparatus vesicles from the vacuole contaminating vesicles we took the membrane interface collected from the 0.25/1.1 M Suc interface of the discontinuous gradient previously described, and loaded on top of a 5-mL 1.3 M Suc cushion. Then, 10 mL of 0.67 M Suc and 5 mL of 0.25 M Suc were laid on top of the membranes. The gradient was centrifuged for 90 min at 100,000g. One-milliliter fractions were collected and immediately used for Ca²⁺ uptake assays, enzymatic determinations, and western blotting.

Cauliflower Subcellular Fractionation

The floral meristem was homogenized in 1 volume of 0.5 M Suc, 0.1 M KH₂PO₄ (pH 6.65), 5 mM MgCl₂, and 1 mM dithiothreitol (added fresh) using a blender. The homogenate was filtrated through Miracloth and centrifuged at 1,000g for 5 min. The supernatant was layered on 8 mL of 1.3 M Suc cushion and centrifuged at 100,000g for 90 min. The membranes located on top of the 1.3 M layer were collected, adjusted to 1.1 M Suc, and layered on top of a new 1.3 M Suc cushion. Then, 0.74 and 0.25 M Suc solutions were layered respectively on top of the 1.1 M layer; then, the Suc gradient was centrifuged at 100,000g for 100 min. After centrifugation, four membrane bands and a pellet were observed throughout the gradient. These were collected, spun down at 100,000g, and resuspended in a buffer containing 0.25 M Suc, 1 mM MgCl₂, and 10 mM Tris-HCl (pH 7.5). The membrane fractions were named 1 through 4 relative to their position in the tube, from top to bottom, respectively. Fraction 5 corresponds to the pellet resuspended in the same solution.

Subcellular Marker Assays

XG-FucTase activity was measured as a Golgi marker in the presence of 0.1% (v/v) Triton X-100, as described by Wulff et al. (2000). NADH Cyt c reductase insensitive to antimycin A (ER marker) and Cyt c oxidase (mitochondrial marker) were assayed as described by Briskin et al. (1987). Vanadate-sensitive ATPase activity (PM marker) was measured as described by Lanzetta et al. (1979). As a vacuole membrane marker, we used a polyclonal antibody (kindly donated by Dr. John C. Rogers, Institute of Biological Chemistry, Washington State University, Pullman, WA) against the pea tonoplast intrinsic protein -TIP. Immunoblotting using anti--TIP antibody was carried out as described by Jauh et al. (1999). Proteins from subcellular fractions were denatured in sample buffer at 60°C, separated by SDS-PAGE, and transferred onto polyvinylidene difluoride membranes. The anti--TIP antibody was used at 0.1 ng mL¹. Goat anti-rabbit IgG conjugated to peroxidase (diluted 1:10,000 [v/v]) was used as secondary antibody. Latent UDPase and UDPase activity in native gels (Golgi markers) were measured as described by Orellana et al. (1997). Proteins were measured with the bicinchoninic acid method (Pierce Chemical, Rockford, IL) with bovine serum albumin as standard.

Ca²⁺ Uptake Assays

Ca²⁺ uptake by Golgi apparatus and ER fractions was assayed measuring ⁴⁵Ca²⁺ accumulation by the vesicles using a filtration assay. Experiments were performed using triplicates each time, and each experiment was done at least twice. Either Golgi apparatus or ER vesicles were suspended (0.3 or 0.6 mg mL¹ final concentration) in a buffer solution containing 100 mM KCl and 20 mM MOPS-Tris (pH 7.2). Free Ca²⁺ concentration in the incubation buffer was fixed at desired values using variable amounts of CaCl₂ (containing 4.5 mCi ⁴⁵Ca²⁺ mmol¹ CaCl₂, NEN Life Science Products, Boston) and EGTA. Free Ca²⁺ concentrations were calculated by means of the WinMAXC 2.05 computer program. Vesicles were incubated in the buffer at 25°C for 1 min and transport was initiated (time zero) by adding a mixture of ATP and MgCl₂ to final concentrations of 2 and 3 mM, respectively. Samples of 250 µL in triplicate or quadruplicate were taken at different times and filtered through 0.45-µm nitrocellulose membranes (Millipore, Bedford, MA). The filters were immediately rinsed with 5 mL of ice-cold buffer (100 mM KCl and 20 mM MOPS-Tris [pH 7.2]), then dried and counted in a scintillation counter. ATP-dependent Ca²⁺ uptake (nmol mg¹ min¹) was defined as the difference between the ⁴⁵Ca²⁺ retained in the filters following incubations in the presence and absence of ATP-Mg.

When needed, orthovanadate (200 µM), cyclopiazonic acid (100 nmol mg protein¹), thapsigargin (2 µM), calmodulin (1 µM), carbonyl cyanide 3-chlorophenylhydrazone (5 µM), Cd²⁺, Cu²⁺, Co²⁺, Fe²⁺, Mn²⁺, or the Ca²⁺ ionophore A23187 (10 µM) were added to the incubating solution.