Regulation of Intra-Golgi Membrane Transport by Calcium*

Calcium cations play a critical role in regulating vesicular transport between different intracellular membrane-bound compartments. The role of calcium in transport between the Golgi cisternae, however, remains unclear. Using a well characterized cell-free intra-Golgi transport assay, we now show that changes in free Ca2+ concentration in the physiological range regulate this transport process. The calcium-chelating agent 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid blocked transport with an IC50 of approximately 0.8 mm. The effect of 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid was reversible by addition of fresh cytosol and was irreversible when performed in the presence of a Ca2+ ionophore that depletes calcium from lumenal stores. We demonstrate here that intra-Golgi transport is stimulated by low Ca2+concentrations (20–100 nm) but is inhibited by higher concentrations (above 100 nm). Further, we show that calmodulin antagonists specifically block intra-Golgi transport, implying a role for calmodulin in mediating the effect of calcium. Our results suggest that Ca2+ efflux from intracellular pools may play an essential role in regulating intra-Golgi transport.

Although Ca 2ϩ is the most abundant cation in vertebrates, eukaryotic cells sequester Ca 2ϩ efficiently, mainly by uptake into intracellular stores, and thus display low cytosolic concentrations of about 100 nM (for reviews see Refs. 1 and 2). Fast and short (transient) increases in the cytosolic Ca 2ϩ concentration play a pivotal role in many physiological processes, and the dynamic characteristics of these pools are regulated in a well defined manner. Calcium sequestration into the ER 1 lumen depends on Ca 2ϩ -ATPases, known as sarco/endoplasmic reticulum ATPases (for reviews see Ref. 2), all of which share the common property of being selectively inhibited by thapsigargin, a tumor-promoting sesquiterpene lactone (3). Calcium handling by the ER is also controlled by inositol 1,4,5-triphosphate and ryanodine receptors (4 -7).
Although much is known about the ER as an intracellular Ca 2ϩ storage pool, far less is known about calcium sequestration by the Golgi apparatus. Recent studies documented the presence of a sarco/endoplasmic reticulum adenosine triphos-phatase isoform and inositol 1,4,5-triphosphate receptors on the Golgi and showed that they dynamically regulate calcium homeostasis of this organelle (8). Increasing evidence indicates the involvement of Ca 2ϩ cations in different intracellular transport steps. For example, Ca 2ϩ requirement was demonstrated for ER to Golgi transport and for the assembly of the nuclear envelope (9,10). Moreover, recent studies showed that both homotypic vacuolar fusion and fusion between early endosomes depend on Ca 2ϩ and calmodulin (CaM), the ubiquitous calcium effector (11)(12)(13). However, the role played by Ca 2ϩ in intra-Golgi transport remains unclear. The cell-free intra-Golgi transport assay was reported to be insensitive to the Ca 2ϩ chelator EGTA (14), suggesting that Ca 2ϩ is not required for intra-Golgi transport, whereas based on evidence from semiintact cells, Schwaninger et al. (9) suggested that this cation is required at a late stage of transport within the Golgi.
In this study we have characterized the requirement for calcium in transport within the Golgi. We present evidence that intra-Golgi transport require low free Ca 2ϩ concentrations within the physiological range and that above a critical level (100 nM) Ca 2ϩ has the reverse effect and inhibit transport. We thus show that free Ca 2ϩ is required in a very narrow concentration range for this process and suggest a dual role for Ca 2ϩ in intra-Golgi transport. In addition, we provide evidence that links CaM to the stimulatory effect of Ca 2ϩ on transport.

EXPERIMENTAL PROCEDURES
Preparation of Rat Brain Cytosol-Rat brain cytosol was prepared essentially as described previously for bovine brain cytosol (15).
Intra-Golgi Transport Assay-The standard intra-Golgi transport was performed as described previously (16). A standard intra-Golgi transport assay (25 l) contained 2.5 mM Hepes buffer, pH 7.0, 15 mM KCl, 2.5 mM magnesium acetate, 0.4 Ci of UDP-[ 3 H]N-acetylglucosamine, 5 l of a 1:1 mixture of donor and acceptor Golgi membranes (2-3 g of protein), 0.25 l (30 ng) of His 6 ␣SNAP, 10 M palmitoylcoenzyme A, and ATP and UTP regenerating systems. Transport activity was dependent upon addition of cytosol, and in this study 27.5 g of rat brain cytosol was added to achieve optimum conditions for the transport activity. The transport reactions were incubated at 30°C for 2 h. [ 3 H]N-Acetylglucosamine incorporated into VSV-G protein was determined as described previously (17). Each of the transport assay experiments shown in this study represent at least three independent assays performed in duplicate. (Where standard errors are absent the results represent 3 independent experiments with duplicates. Standard errors did not exceed 5%.) Glycosylation Assay-For the glycosylation assay, "wild type donor" membranes were prepared as described by Taylor et al. (18). Briefly Golgi membranes were isolated from wild type Chinese hampster ovary cells infected with VSV, after which the isolated "wild type donor" membranes were treated with N-ethylmaleimide (1 mM) for 15 min on ice, at which time dithiothreitol (2 mM) was added to quench any remaining N-ethylmaleimide. The glycosylation assay was performed under identical conditions as those described for the standard transport assay.
Free Ca 2ϩ Concentration Measurements-The free cation concentrations were determined using the Bound and Determined (BAD) software (33).

RESULTS
To address the role of Ca 2ϩ in intra-Golgi transport we tested the effects of EGTA and BAPTA, two Ca 2ϩ chelators, on the well characterized cell-free intra-Golgi transport assay (16). Addition of increasing amounts of BAPTA inhibited transport with an IC 50 of approximately 0.8 mM, whereas EGTA showed only a minimal effect (Fig. 1A). The inhibition of transport observed in the presence of BAPTA is not due to chelation of Mg 2ϩ because a similar pattern was observed when BAPTA was added in the presence of 3 mM MgCl 2 . When BAPTA was added to the cell-free transport assay in the presence of an equivalent molar concentration of Ca 2ϩ , no inhibition was observed (Fig. 1B). The different effects of these chelators on intra-Golgi transport cannot be explained by a different cation preference, because BAPTA and EGTA bind Ca 2ϩ with a similar affinity at neutral pH. However, BAPTA has Ͼ100-fold higher ion association and dissociation rates than EGTA (19). Thus, processes that are dependent on fast and local changes in Ca 2ϩ concentration will be sensitive to BAPTA but not to EGTA (20,21).
The effect of BAPTA on intra-Golgi transport was fully re-versible when the Golgi membranes were reisolated and incubated with fresh untreated cytosol (Fig. 1C). However, when the membranes were incubated with BAPTA in the presence of 100 M ionomycin, a Ca 2ϩ ionophore that causes release of Ca 2ϩ from lumenal stores, fresh cytosol failed to restore transport. These results are consistent with the notion that low and transient Ca 2ϩ effluxes, possibly from the Golgi membrane, are essential for intra-Golgi transport.
To exclude the possibility that the inhibition of the assay signal by BAPTA resulted from decreased glycosylation activity of GlcNAc transferase rather than from inhibition of transport, we used a glycosylation assay that determines GlcNAc activity when membrane transport is inactivated (for details see "Experimental Procedures"). As shown in Fig. 1D, neither BAPTA nor EGTA affected the glycosylation of VSV-G protein, indicating that BAPTA indeed exerts its effect on the process of intra-Golgi transport.
To examine at which stage BAPTA acts in inhibiting intra-Golgi transport, we examined the effect of BAPTA when combined with two other inhibitors of transport (15,22). In the experiment described in Fig. 2, the transport assay was either FIG. 1. Local calcium is required for intra-Golgi transport. A, effect of calcium chelators is shown. Increasing amounts of BAPTA and EGTA were added as indicated, to the intra-Golgi transport assay in the presence or absence of 3 mM Mg 2ϩ . B, BAPTA inhibition is blocked by calcium. BAPTA (2 mM) was added to the transport assay in the presence or absence of molar equivalence CaCl 2 as indicated. C, efflux of calcium is required for intra-Golgi transport. Golgi membranes were incubated (15 min, 25°C) with either transport buffer containing BAPTA (5 mM) or BAPTA (5 mM) and ionomycin (100 M) as indicated. The membranes were reisolated, supplemented with transport buffer in the presence or absence of rat brain cytosol, and assayed for transport using standard conditions. D, calcium chelators do not affect N-acetylglucosamine transferase I activity. The glycosylation assay was performed using wild type VSV Golgi membranes (see "Experimental Procedures") in the absence or presence of either 5 mM BAPTA or 5 mM EGTA. terminated at different time points by placing the reaction on ice, or at these time points the inhibitors BAPTA, anti-SBP56 antibodies, or GTP␥S were added as indicated, and the reaction was allowed to proceed for 2 h. Control samples treated with buffer only were incubated likewise at 30°C until the end of the 2-h incubation period and served to represent 100% of transport activity. All three inhibitors, when added at the onset of the reaction, produced 90% inhibition of transport. The reaction became resistant to BAPTA after the inhibition by GTP␥S and anti-SBP56 antibodies, indicating that Ca 2ϩ is required late in the transport process, possibly at the fusion stage.
It has been demonstrated that different organelles, including the Golgi apparatus, accumulate Ca 2ϩ in their lumen (8). To test the role of the lumenal Ca 2ϩ pool in intra-Golgi transport, Golgi membranes were incubated in the presence of increasing concentrations of either ionomycin (a Ca 2ϩ ionophore) or thapsigargin (a tumor-promoting sesquiterpene lactone that binds with high affinity and irreversibly inhibits all sarco/endoplasmic reticulum adenosine triphosphatase pumps). These two compounds are known to reduce lumenal Ca 2ϩ levels selectively from intracellular organelles. As shown in Fig. 3, both ionomycin and thapsigargin inhibited intra-Golgi transport with IC 50 values of 35 and 50 M, respectively, with no effect on glycosylation by GlcNAc transferase. These results clearly indicated that lumenal Ca 2ϩ was essential for this process.
Having demonstrated that release of Ca 2ϩ from intracellular stores may play a role in regulating transport, we examined the concentration of free Ca 2ϩ required for intra-Golgi transport. For that purpose we added increasing CaCl 2 concentrations to the cell-free transport assay in the presence of 5 mM BAPTA or EGTA (Fig. 4, A and B). Interestingly addition of 100 nM free Ca 2ϩ caused the assay signal to return to almost maximal levels (88% of the control determined in the absence of any Ca 2ϩ chelator), but addition of higher free Ca 2ϩ concentrations significantly inhibited the cell-free transport assay. When tested, neither Mn 2ϩ nor Cu 2ϩ was able to substitute Ca 2ϩ (Fig. 4A). We then tested the inhibitory effect of Ca 2ϩ using EGTA as cation chelator. Under these conditions the stimulatory effect of Ca 2ϩ could not be observed. However, as found in the presence of BAPTA, free Ca 2ϩ concentrations above 100 nM inhibited transport, reaching a maximum effect at about 200 nM free Ca 2ϩ (Fig. 4B). These results clearly demonstrated that within the physiological range, Ca 2ϩ plays a dual role in regulating intra-Golgi transport. Furthermore, the ability of BAPTA but not EGTA to inhibit transport by chelating Ca 2ϩ suggests that the effect of Ca 2ϩ in regulating membrane fusion is localized to the vicinity of the membrane.
A number of studies using endosomal fusion (12) or yeast vacuole homotypic fusion (11) have demonstrated a requirement for the cytosolic Ca 2ϩ effector, CaM. We therefore examined the involvement of CaM in intra-Golgi transport. For that purpose, increasing concentrations of two specific CaM inhibitors, W7 or trifluoperazine dimaleate, were added to the cellfree transport assay. As shown in Fig. 5A, 25 M either W7 or trifluoperazine dimaleate inhibited up to 90% of the total transport activity with an IC 50 of about 10 M. Addition of W5, a much weaker CaM antagonist, however, only partially inhibited transport (IC 50 Ͼ 250 M) thus indicating that the inhibition observed in the presence of the different antagonists is CaM-mediated. Notably, W7 or trifluoperazine dimaleate failed to significantly inhibit the glycosylation of VSV-G protein as determined by the glycosylation assay (data not shown).
To further verify the involvement of CaM in this process, we performed a two-stage transport assay in which Golgi membranes were first treated with or without W7 (30 min on ice) in transport assay conditions. The membranes were then reisolated, washed, and tested for transport activity in the presence of fresh cytosol (Fig. 5B). Addition of fresh cytosol to the control membranes recovered most of their transport activity, whereas W7-treated membranes could restore up to 90% of transport only when purified CaM was added together with the fresh cytosol. Hence, CaM appears to be involved in intra-Golgi transport. These results also indicate that the low levels of CaM in the cytosol are insufficient to reactivate the Golgi membranes after treatment with W7. CaM-dependent transport was strongly inhibited by BAPTA (Fig. 5B) indicating that Ca 2ϩ is required for the CaM activation of transport. DISCUSSION It has been well demonstrated that Ca 2ϩ regulates vesicular transport between a number of different intracellular or-

FIG. 2. Calcium is required for a late stage of intra-Golgi transport.
A standard intra-Golgi transport assay was carried out at 30°C for 2 h. At the indicated time points, anti-SBP56 antibodies (1 g, circles), GTP␥S (50 M, diamonds) or BAPTA (4 mM, triangles) were added, and the reaction was terminated after 2 h. The progression of transport in the absence of inhibitors was measured by transferring samples to ice at the indicated time points (squares).

FIG. 3. Thapsigargin and ionomycin inhibit intra-Golgi transport specifically.
Thapsigargin (Tg) and ionomycin were added as indicated to a standard transport assay and to a glycosylation assay (see "Experimental Procedures").
ganelles, but its involvement in intra-Golgi transport remains poorly understood. Here, we have investigated the role Ca 2ϩ plays in this process by using the well established intra-Golgi transport assay. We found that low cytosolic Ca 2ϩ concentrations (approximately 100 nM) are optimal for intra-Golgi transport whereas higher Ca 2ϩ concentrations inhibit this process.
Moreover, it appears that Ca 2ϩ efflux from intracellular stores (possibly from the Golgi complex itself) is essential for intra-Golgi transport. Our results indicated that Ca 2ϩ played a role in late stages of transport, possibly in the fusion of vesicles with their target membrane. Finally we demonstrate that CaM is involved in intra-Golgi transport.
Several studies have documented that Ca 2ϩ plays an important role in regulated exocytosis of secretory granules and synaptic vesicles (reviewed in Refs. 23 and 24). More recently it was reported that Ca 2ϩ participates in other intracellular transport events, including ER to Golgi transport (25), assembly of nuclear membrane (10), transcytotic vesicle fusion (26), vacuolar membrane fusion (11), and fusion between endosomes (12,13). The requirement for Ca 2ϩ in intra-Golgi transport has remained unclear, however, because early reports using a cellfree intra-Golgi transport assay demonstrated no effect of EGTA (14), whereas in semi-intact cells, EGTA was able to inhibit this transport process (9). In agreement with the early reports, we now show that intra-Golgi transport is indeed resistant to EGTA. However, this transport step is highly sensitive to the fast acting Ca 2ϩ chelator, BAPTA. This phenomenon of sensitivity to BAPTA but resistance to EGTA has been observed for other systems of membrane fusion, including homotypic vacuolar fusion (11) and endosomal fusion (13). This suggests that transient, probably local, changes in Ca 2ϩ concentration mediate the fusion process. Indeed, our results (Fig. 3) are consistent with the notion that intra-Golgi transport depends upon Ca 2ϩ efflux from inner membrane stores. Several studies have indicated that the Golgi complex may function as a Ca 2ϩ storage organelle (8,(27)(28)(29). Thus it is feasible that Ca 2ϩ from the lumen of the Golgi directly participated in regulating intra-Golgi transport.
In most cells the free Ca 2ϩ concentration rests at approxi- FIG. 4. Regulation of intra-Golgi transport by calcium. A, CaCl 2 , MnCl 2 , and CuCl 2 solutions were mixed with BAPTA (5 mM final concentration) and added to the intra-Golgi transport assay. The final free Ca 2ϩ , Mn 2ϩ , and Cu 2ϩ concentration values, as calculated using the PC-BAD4 software (33), are presented as the Log[cation]. Linear presentation of free Ca 2ϩ concentrations in the intra-Golgi transport assay is shown in the upper right inset. B, CaCl 2 solution was mixed with EGTA (5 mM final concentration) and added to the intra-Golgi transport assay. The final free Ca 2ϩ concentration values were calculated as for BAPTA.

FIG. 5. Involvement of calmodulin in intra-Golgi transport. A,
Golgi membranes were incubated in a transport assay condition, in the presence of increasing amounts of trifluoperazine dimaleate, W7, or W5 as indicated. B, Golgi membranes were incubated (30 min on ice) in a transport assay condition in the presence or absence of 20 M W7. The membranes were then washed with 250 mM KCl and resuspended in transport buffer with supplements as indicated, and the intra-Golgi transport activity was determined after 2 h incubation at 30°C. mately 100 nM, and following activation of cellular signaling pathways Ca 2ϩ levels are elevated about 1000-fold (30,31). We have found that intra-Golgi transport is optimal in the presence of about 100 nM free Ca 2ϩ , but that higher Ca 2ϩ concentrations strongly inhibit transport. This may imply that constitutive membrane transport is optimal when the cytosolic Ca 2ϩ concentration remain at resting levels, but is inhibited upon elevation of cytosolic Ca 2ϩ . Because exocytosis depends critically on high Ca 2ϩ levels, this may provide a mechanism for the cell to conserve the fusion machinery for the transient and more immediate requirement of exocytosis at the expense of constitutive transport.
What are the downstream targets of Ca 2ϩ in this system? It has recently been demonstrated that the Ca 2ϩ -binding protein CaM plays an important role in fusion between yeast vacuoles (11). It was postulated that CaM acts in this system at a late step of the fusion reaction, probably after docking of the transport vesicle with its target membrane (11). CaM antagonists were found to inhibit endosome fusion in vitro (12). The strong inhibition of intra-Golgi transport by the CaM antagonist reported here suggests that CaM may regulate various intracellular fusion processes in a Ca 2ϩ -dependent manner. Recently Mayer and co-workers (32) suggested that protein phosphatase 1 may be involved in late stages of vacuolar fusion, possibly in a CaM-dependent manner. Future studies will determine whether protein phosphatase 1 or another target represents the effector of Ca 2ϩ in constitutive membrane transport and how this effector acts on the transport machinery.