HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 49, 47756-47764, December 6, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/49/47756    most recent M209199200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Takeda, Y. Articles by Kasamo, K. Search for Related Content PubMed PubMed Citation Articles by Takeda, Y. Articles by Kasamo, K. Social Bookmarking                      What's this?

In Vitro Fusion of Plant Golgi Membranes Can Be Influenced by Divalent Cations^*

Yuichi Takeda and Kunihiro Kasamo

From the Research Institute for Bioresources, Okayama University, 1-20-2 Chuo, Kurashiki 710-0046, Japan

Received for publication, September 9, 2002, and in revised form, October 2, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The fusogenic activity of plant Golgi membranes was studied in a cell-free system by assaying lipid mixing and content leakages of fluorescence probes. Golgi membranes from mung bean (Vigna radiata L.) hypocotyl cells fused to liposomes in the absence of any cytosolic proteins and nucleotides. It was demonstrated that the fusion was mediated by integral membrane protein(s), and was influenced by divalent cations (mM). Mg²⁺, Ca²⁺, and Mn²⁺ ions enhanced the lipid mixing by reducing repulsive forces between membranes. In the content leakage assay, Mg²⁺ ions also showed a stimulative effect. However, other divalent cations were inhibitory. It is suggested that the fusion system of Golgi membranes comprises at least two components: one that mediates the formation of fusion intermediates prior to pore opening, and one that mediates the subsequent processes. The latter must be sensitive to divalent cations at millimolar concentrations. The fusion of Golgi and biological membranes was induced by divalent cations. We speculated about the biological role of the fusion system studied here.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Vesicular transport plays an essential role in the intracellular transport of macromolecules like proteins, lipids, and polysaccharides. Upon arrival at the target organelle, a vesicle fuses with the organelle membrane and releases its aqueous contents into the lumen, or the outside of the cell in the case of exocytosis.

The mechanisms of all intracellular membrane fusion events in eucaryotic cells are considered to be founded on the so-called "SNARE¹ hypothesis," proposed by Rothman et al. (1) and modified thereafter (2-4). In the model, numerous factors are involved in the fusion process at several steps. Small GTPases of the Rab/Ypt family and tethering factors play roles in vesicle tethering and docking interactions (5-8). Thereafter, the pairing of vesicular and target membrane-SNAREs occurs, which should result in tight docking and facilitate fusion (9-11). Other factors like N-ethylmaleimide-sensitive factor, soluble N-ethylmaleimide-sensitive factor attachment protein, protein phosphatase (12), calmodulin and Ca²⁺ ions (13, 14), V₀-subunits of V-type ATPase in yeast vacuole homotypic fusion (15), and GTPases of the Rho family (16, 17) are also needed for the process. It is generally accepted that intracellular membrane fusion requires the tethering and pairing of some protein factors.

Several studies have shown that membrane protein-dependent fusion without the tethering or pairing of fusion factors can occur in intracellular membranes, by utilizing artificial membranes. Rat liver Golgi membranes (18), reticulocyte endocytic vesicles (19), rat brain synaptosomes (20), and Golgi and smooth endoplasmic reticulum (ER) membranes of rabbit liver (21) fused with liposomes in cell-free systems without any cytosolic proteins or nucleotides. Rat liver ER membranes fused with liposomes at lower pH, dependent on a 50-kDa glycoprotein in ER membrane (22, 23). The Golgi apparatus of perforated CHO-K1 fibroblasts fused with liposomes in an ATP-dependent manner (24). Sea urchin exocytotic granules fused with liposomes as well as themselves in a Ca²⁺ ion-dependent and NEM-sensitive manner (25). It is sufficient for these fusogenic peptides to reside on only one of the two membranes to cause fusion, like viral fusion peptides. The peptides that mediate fusion with artificial membranes would also be involved in intracellular fusion processes and the trafficking of membrane proteins and lipids. However, except for the fusogenic 50-kDa glycoprotein of rat liver ER (22, 23), little is known about the mechanism of fusion induced by such peptides and most of them have not yet been identified, whereas much information about SNAREs, Rab/Ypts, and various other factors has been accumulated during the last decade.

Such a peptide-dependent fusogenic activity was also found in plant Golgi membranes isolated from mung bean (Vigna radiata L.) hypocotyls, by lipid mixing and content leakage assays in a cell-free system utilizing fluorescence probes. In the present study, the effect of divalent cations was examined. The results suggest that the fusion to an intermediate prior to pore opening and the subsequent processes are mediated by different components of the fusion system of plant Golgi apparatus.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Preparation of Asolectin Liposomes

Solutions of purified soybean asolectin (Wako, Tokyo, Japan) were made in chloroform. Lipids were dried into thin films under N₂. The lipids were suspended in 10 mM Mes-Tris (pH 7.2) containing 135 mM KCl (K-buffer), and sonicated with a bath-type sonicator, which produces small unilamellar vesicles (26).

Plant Materials and Preparation of Golgi- and Other Membrane-enriched Fractions

Seeds of the mung bean (V. radiata L.) were hydrated in tap water and cultivated in darkness at 29 °C for 2.5 days.

Excised hypocotyl sections (150 g fresh weight) were homogenized in 300 ml of 0.25 M sorbitol, 50 mM Mops-KOH (pH 7.6), 5 mM EGTA, 1 mM DTT, 1.5% (w/v) polyvinylpolypyrrolidone, 0.4% (w/v) bovine serum albumin, 0.4% (w/v) casein, 1 mM phenylmethylsulfonyl fluoride, and 2.5 mM K₂S₂O₅. The 10,000-150,000 × g microsomal pellet was suspended in sorbitol suspension buffer containing 0.25 M sorbitol, 1 mM EGTA, 1 mM DTT, and 10 mM Mes-Tris (pH 7.3), then loaded on top of discontinuous sucrose-density gradients consisting of 10/18/26/32/40% (w/w) sucrose in 1 mM EGTA, 1 mM DTT, and 10 mM Mes-Tris (pH 7.3), and centrifuged for 2 h at 90,000 × g. The tonoplast (TP)-enriched fraction at the 0.25 M sorbitol/10%, ER at the 18/26%, Golgi at the 26/32%, and plasma membrane (PM) plus mitochondria (Mt) at the 32/40% sucrose interfaces were collected and diluted with K-buffer supplemented with 1 mM DTT, then centrifuged at 150,000 × g for 20 min.

All membrane vesicles were finally suspended in K-buffer supplemented with 1 mM DTT. The procedures for the membrane preparation were conducted on ice or at 4 °C. Isolated membranes were used for the experiments without freezing or stored at 80 °C.

Protein Determination, Enzyme Assays, SDS-PAGE, and Immunoblotting

Protein content was determined using a Bio-Rad DC protein assay kit with bovine serum albumin as the standard.

The marker enzyme assays in the presence of 0.015% (w/v) Triton X-100, the lipid extraction, and the analysis were performed as described (27-29).

SDS-PAGE was carried out on 10% polyacrylamide gels, as described (29). After electrophoresis, the gel was subjected to electrophoretic transfer on a polyvinylidene difluoride membrane. Immunostaining was performed essentially as described by Herman et al. (30).

Fusion Assays

Lipid Mixing-- For the preparation of asolectin liposomes containing octadecylrhodamine B (R18) (Molecular Probes, Eugene, OR), dissolved in an ethanolic stock solution, was premixed with the lipids in chloroform to give a final concentration of 4 mol % to total lipids.

Golgi membranes labeled with R18 were prepared as follows; 5 mg of protein of Golgi membrane was suspended in 5 ml of K-buffer, 75 nmol of R18 was added to the suspension, and the mixture was incubated on ice for 1 h in the dark. Unincorporated R18 was removed on a Sephadex G-75 column (19, 31), equilibrated with K-buffer.

The final concentration of R18 in Golgi membranes was evaluated by solubilization with chloroform/methanol/0.1 N HCl and measuring R18 fluorescence (31), and estimated to be 6-7 mol % with respect to total membrane lipids.

Lipid mixing of membranes was assayed by de-quenching of R18 fluorescence incorporated in asolectin liposomes or Golgi membranes (31). 0.2 µmol of phospholipid (PL) of asolectin liposomes containing R18 or 0.4 mg of protein of Golgi membranes labeled with R18 was incubated in a fluorescence cuvette in K-buffer. The reaction was initiated by the injection of R18-free membrane vesicles (0.4 mg of protein) or R18-free liposomes (0.4 µmol of PL), respectively. The final volume was 2 ml. The changes in fluorescence were monitored at 37 °C with a spectrofluorophotometer (model RF-5300PC, Shimadzu, Kyoto, Japan) at excitation and emission wavelengths of 560 and 590 nm, respectively. Finally, Triton X-100 was added with a final concentration of 0.2% (w/v).

Content Leakage-- Asolectin liposomes were made in a solution of 41 mM calcein (Sigma) or 25 mM 1-aminonaphthalene-3,6,8-trisulfonic acid (ANTS) (Molecular Probes) in K-buffer. Unencapsulated calcein was removed as described by Kobayashi and Pagano (24), except that the dialysis was done against K-buffer. At such a high concentration, self-quenching of calcein can occur (24). Unencapsulated ANTS was removed using a Sephadex G-75 column. In content leakage assays, asolectin liposomes encapsulating calcein (0.4 µmol of PL) or ANTS (0.2 µmol of PL) were incubated in a fluorescence cuvette in K-buffer, and then membrane vesicles (0.4 mg of protein for calcein, 0.2 mg for ANTS) were added to initiate the reaction. The final volume for the incubations was 2 ml. In the ANTS assay, 20 mM N,N'-p-xylylenebis(oyridinium bromide) (DPX) (Molecular Probes) was pre-included in the reaction mixture. The changes in fluorescence intensity were monitored at 37 °C. Finally, Triton X-100 was added (final 0.2%, w/v). The excitation and emission wavelengths were 490 and 520 nm for calcein and 386 and 515 nm for ANTS, respectively.

In both the lipid mixing and the content leakage assays, in most cases, the fluorescence intensity at time 0 was set at 0%, and that after addition of Triton X-100, at which the infinite and/or maximum probe dilution should occur, was taken as 100% for the scale calibration.

Treatment of Golgi Membranes

Golgi membranes were incubated with trypsin (1:5, w/w, trypsin:membrane protein) at 37 °C for 20-30 min, and then soybean trypsin inhibitor (1.5-fold of trypsin) was added. For the control sample, trypsin and the inhibitor mixed beforehand were added.

Golgi membranes were incubated in 10 mM Mes-Tris (pH 7.2) containing 1 M KCl and stood for 15 min on ice, then utilized for assays. The amount of KCl in the incubation mixture for the fusion assays was adjusted to give the same final concentration (135 mM).

NEM-treated Golgi membranes were prepared by incubation in K-buffer containing 1 mM NEM for 15 min on ice, and then stopped by subsequent addition of 2 mM DTT. The control was prepared by the addition of a mixture of NEM and DTT.

Note that control samples for treatments with trypsin and NEM produced the same results as untreated samples.

Golgi membranes were heated at 90 °C for 5-10 min, and then kept on ice until measurements.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Characterization of Membrane Fractions-- Organelle membranes were fractionated from mung bean hypocotyl cells utilizing discontinuous sorbitol/sucrose gradient centrifugation. Assays using marker enzymes were performed to characterize the prepared membrane fractions. UDPase activity was used as a Golgi marker (27, 28, 32). NADH-cytochrome c reductase insensitive to antimycin A and cytochrome c oxidase were employed as markers for the ER and Mt, respectively (32). VO₄-inhibited K⁺-stimulated Mg²⁺-ATPase activity at pH 6.5 was used as a PM marker, and NO₃-inhibited Mg²⁺-ATPase activity at pH 8.0 as a TP marker (33). Table I shows that these marker enzymes were most active in their original membrane fractions.

As shown in Fig. 1, the membrane fractions were also characterized by immunoblotting with specific antibodies: VM23, a TP marker (34), and PM H⁺-ATPase were most abundant in their original fractions. BiP, an ER marker (35), was abundant in both the ER and Golgi fractions, but slightly more so in the ER fraction. JIM 84, a Golgi marker (36), antibody stained several polypeptide bands of over 60-200 kDa, most of which were most abundant in the Golgi fraction.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Immunoblots of membrane fractions obtained with specific antibodies. The membrane fractions were characterized by SDS-PAGE and immunoblotting with anti-VM23, anti-BiP, JIM 84, and anti-maize PM H+-ATPase antibodies. The levels of the markers in lanes were densitometrically quantified and compared. The normalized level is given at the bottom of each band.

TP can be obtained in high purity by sorbitol/sucrose density gradient centrifugation (28). The ER and Golgi fractions showed considerable cross-contamination, and the latter also contained PM. We tried several methods for the fractionation. However, it proved difficult to prepare a Golgi membrane-enriched fraction with higher purity from mung bean. Nevertheless, it could be concluded that these organelle membranes were most abundant in their own fractions.

Lipid Mixing Assay Shows a Protein(s)-dependent Fusion System Exists in Plant Golgi Membranes-- Fig. 2 shows the fusion of asolectin liposomes containing 4 mol % R18 with various membranes. The incorporation of R18 into membranes results in a concentration-dependent self-quenching of the fluorescence (31). An increase in fluorescence intensity should indicate a release from self-quenching of the probe caused by the mixing of lipids in membranes upon fusion. Thus, R18 can be used as a probe for membrane fusion (19, 22, 23, 25, 37-39).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   The interaction of R18-incorporated asolectin liposomes and biological membranes. R18-free membrane vesicles of various fractions were added to the incubation mixture of the R18 liposomes at 0 min. The fluorescence was calibrated.

Additions of membranes of the Golgi, ER, and PM plus Mt fractions resulted in de-quenching of fluorescence, whereas the fluorescence only slightly increased on incubation with TP. The de-quenching of R18 fluorescence may have occurred as a result of the fusion of the liposomes and these membranes. Initial rates and fluorescence (%) at 10 min of de-quenching were determined to characterize the fusion kinetics. The initial rate was calculated from the slope of the steepest part at time = 0. Because the de-quenching signals of R18 were noisy, 0 and 100% fluorescence were determined from the means of fluorescence intensity of the last 1-5 s and 1-2 min after the addition of the substrates, respectively. Fluorescence (%) values during the fusion reaction were obtained from the means of the previous and the last 10 s at the indicated time. The initial rate would reflect the net fusogenic activity, and the extent of fusion/lipid mixing, the ratio of vesicles available for fusion, could be evaluated and compared from the fluorescence (%). Table II summarizes the analytical results of the de-quenching curves shown in Fig. 2. The fusion efficiencies of both initial rate and fluorescence (%) at 10 min were highest in the Golgi fraction, and considerably low in the TP fraction. In the case of the ER and PM plus Mt fractions, de-quenching was distinctly observed, but to less of an extent than in the Golgi. The relative efficiencies in Table II are nearly in good agreement with the levels of the Golgi markers, UDPase activity in particular, in the fractions (Fig. 1, Table I).

In most cases, the increase of fluorescence upon fusion was nearly exponential within the first 15 min. However, it tended to increase slowly as well as steadily after 15 min (data not shown). As the de-quenching curves could not be necessarily exponentially fitted, thereafter, we adopted the fluorescence (%) at 15 min to evaluate and compare the extent of fusion/lipid mixing.

The effects of various treatments were examined. As shown in Fig. 3A, the incubation of Golgi membranes in 1 M KCl solution did not affect the de-quenching. However, the trypsin treatment and heating completely abolished it. These results suggest the involvement of (an) integral protein(s) in Golgi membranes, and a dilution of lipid upon membrane fusion rather than a spontaneous transfer of the probes to unlabeled membranes. Incubation with NEM did not affect the de-quenching either, indicating no contribution by cysteine residues. The fusion occurred upon the mere mixing of the liposomes and biological membranes, and did not require any cytosolic proteins or nucleotides such as ATP, GDP, and GTP.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Characterization of fusion of the liposomes and Golgi membranes. The assays were performed as in Fig. 2. A, effect of various treatments of Golgi membranes on the liposome-Golgi fusion. Golgi membranes were treated with 1 M KCl, trypsin, 1 mM NEM, or heated, then mixed with the R18 liposomes. The initial rate and fluorescence (%) at 15 min were compared with those of the control samples. Untreated Golgi membranes were utilized as the controls for 1 M KCl-treated and heated samples. Error bars indicate S.D. (n  3). The initial rate for the trypsin-treated sample was 0.0 ± 0.0 (%/min). B, effect of buffer components on the liposome-Golgi fusion. The fusion reaction was performed in 10 mM Mes-Tris (pH 7.2) containing 135 mM KCl (control) or the same concentration of NaCl, or 0.25 M sucrose or mannitol.

The assays were performed in buffer containing 135 mM KCl. As shown in Fig. 3B, substitution of NaCl for KCl in the reaction mixture rendered nearly the same results, whereas the fusion efficiencies were reduced by ~40-60% when 0.25 M sucrose or mannitol was utilized as an osmotic stabilizer instead of KCl. The screening of charges on the membrane surface by the cations would be effective in increasing membrane fusion.

Divalent Cations Can Modify the Kinetics of Lipid Mixing of Asolectin Liposomes and Golgi Membranes-- It was found that divalent cations could influence the membrane fusion. As shown in Fig. 4A, when MgSO₄ (0.1-5 mM) was added to the reaction mixture, the fusion was enhanced in a concentration-dependent manner. The utilization of MgCl₂ instead of MgSO₄ produced the same results (data not shown), indicating that Mg²⁺ ions caused the increase in de-quenching of R18. The divalent cations were not necessarily indispensable for the fusion because the addition of EDTA had no effect.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Modification of the liposome-Golgi fusion by divalent cations. Assays were performed as described in Fig. 2. A, effect of an increasing concentration of MgSO4 or EDTA on the fusion. MgSO4 at 0.1-5 mM or EDTA at 1 mM was included in the incubation before the reaction was started. B, effect of 1 mM Mg2+, Ca2+, and Mn2+ ions on the fusion. Untreated or trypsin-treated Golgi membranes were incubated with the R18 liposomes in the absence or presence of 1 mM MgCl2, CaCl2, or MnCl2. The de-quenching curves of trypsin-treated membranes in the presence of the cations were subtracted from the curves of untreated. Typical curves are presented. The controls for trypsin-treated samples showed the same de-quenching profiles as the untreated samples (data not shown).

As shown in Table III, other divalent cations also enhanced the de-quenching with the exception that Hg²⁺ ions were inhibitory, whereas the same concentration of monovalent cations did not. The initial rate was more markedly enhanced by divalent cations than the extent of fusion/lipid mixing (fluorescence, %). Because the de-quenching was only slightly increased by the presence of divalent cations in the case of trypsin-treated and heated Golgi membranes (Table III), the increase in fusion caused by the divalent cations would be the result of the increased activity of the fusion protein(s) in the system for the most part, although a slight passive transfer of R18 and/or spontaneous fusion might have occurred. De-quenching curves of trypsin-treated samples in the presence of 1 mM Mg²⁺, Ca²⁺, or Mn²⁺ ions are subtracted from the curves of untreated to remove the effects at long time of slowly increasing fluorescence. The resultant curves are shown in Fig. 4B, indicating that the fusion was distinctly enhanced by the divalent cations.

View this table: [in this window] [in a new window]   Table III Effect of divalent and monovalent cation chlorides on the kinetic characteristics of lipid mixing of R18-incorporated liposomes and Golgi membranes Values ± S.D. are given from n  3 different preparations.

The fluorescence of the reaction mixture before and after the addition of Triton X-100 was not affected by the cations tested. The lipid composition of the membranes remained unchanged under that condition in the presence of these cations, although slight self-digestion of PLs by endogenous membrane-bound phospholipase D (29, 40) occurred in the presence of 1 mM CaCl₂, but the total PL content remained unchanged (data not shown).

Divalent cations have a marked effect to reduce negative electrostatic potential on the membrane surface, which leads to a decrease in aquatic and electrostatic repulsive forces between membranes. The effect at 1 mM of divalent cations on the enhancement of fusion was greatest in the following order: Mn²⁺ > Ca²⁺ > Mg²⁺ >/ Ba²⁺  Sr²⁺ (Table III). This order is the same as that of the association constants for the phosphatidylglycerol complexes, and the enhancement by the divalent cations was in proportion to the constants (41). The divalent cations with the higher association constant would bind more tightly to the surfaces of both biological membranes and liposomes, and more effectively reduce the repulsive forces. Therefore, it is considered that the fusion peptide(s) of the fusion system on Golgi membranes is more accessible to liposomes in the presence of divalent cations, and would not be directly activated by them. In other words, the substrate concentration of the protein(s) was raised by the divalent cations, which should result in an increase of fusogenic activity.

Exceptionally, the de-quenching was markedly decreased by Hg²⁺ ions, suggesting that the fusogenic activity of the peptide(s) was directly inhibited by Hg²⁺ ions.

Although divalent cations like Mg²⁺ and Ca²⁺ ions alone can adhere and fuse lipid bilayers, especially ones composed of substantial amounts of phosphatidylethanolamine and phosphatidylserine (42-46), lipid mixing and/or transfer of the probes between asolectin liposomes did not occur even in the presence of the divalent cations ( 3 mM), although Mn²⁺ ions at 3 mM caused a mere slight de-quenching (data not shown).

Content Leakage Assays Indicate That the Fusion System Comprises at Least Two Components-- Fig. 5 shows the results of content leakage assays utilizing calcein (A) and ANTS/DPX (B and C). They can be used for monitoring fusion processes of artificial membranes and biological membranes (24, 47).

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Content leakage assays. At the time indicated by open arrowheads, Golgi membrane vesicles were added to the incubation of asolectin liposomes encapsulating 41 mM calcein (A) or 25 mM ANTS (B and C). The fluorescence was monitored, and then 0.2% Triton X-100 was added (filled arrowheads). A, assays utilizing the calcein liposomes. CoCl2 at 75 µM was added at the time indicated by arrows. Traces a and c, untreated Golgi membranes; trace b, trypsin-treated. (Inset) The same recording of trace c on a reduced scale around 0 min. B, effect of divalent cations on the leakage of ANTS liposomes following interaction with Golgi membranes. Untreated Golgi membranes were mixed with the liposomes in the absence (+none) or presence of 3 mM MgCl2, 1 mM CaCl2, or 1 mM MnCl2. The quenching of leaked ANTS by DPX in the reaction mixture was monitored, and the scale was calibrated. Trace a, trypsin-treated Golgi membranes without divalent cations. Typical profiles are presented. C, removal of divalent cations by EDTA during the reaction of the ANTS liposomes and Golgi membranes. MgCl2 or MnCl2 at 1 mM was included in the incubation beforehand, and EDTA at 2 mM was added at the time indicated by an arrow.

When asolectin liposomes encapsulating a high concentration of calcein were incubated with Golgi membranes, the fluorescence rapidly increased, showing that dilution of the probe occurred (Fig. 5A, trace a). As no phospholipase activity was detected under the conditions (data not shown), and the leakage was markedly inhibited by trypsin treatment (Fig. 5A, trace b), but not affected by 1 M KCl and NEM treatments (data not shown), it is considered that an integral protein(s) on the Golgi membrane caused leakage of the content of the liposomes. It is suggested that the protein(s) in the fusion system that caused the lipid mixing (Figs. 2-4) induced the leakage upon membrane fusion. Most likely, fusion pore opening and expansion would occur. The aqueous probes in the liposomes would have leaked through the pore and diluted.

Calcein can be quenched by divalent cations like Co²⁺ and Mn²⁺ ions (48). When CoCl₂ was added prior to the Golgi membranes, the fluorescence decreased slightly and then further on the addition of the membranes (Fig. 5A, trace c, inset). The first decrease may be caused by quenching of the dye that leaked from the liposomes by passive diffusion. The second would be induced by release of the aqueous content upon membrane fusion. The isolation of membranes from plant cells by sucrose high density gradient centrifugation sometimes results in the loss of the permeability barrier of the membranes (27). Therefore, it is plausible that the Golgi membrane vesicles are permeable to divalent cations, and the calcein encapsulated in the liposomes was quenched by Co²⁺ ions that entered the lumen of the vesicles upon fusion. Penetration by calcein of Golgi membranes fused with the liposomes is also possible.

Because divalent cations can act as quenchers of calcein, their effect on Golgi-liposome fusion was investigated by content leakage assay utilizing ANTS and DPX, which are also used for content mixing assays to monitor membrane fusion (42, 49, 50). Both are water-soluble, and DPX efficiently quenches ANTS fluorescence (42).

As the isolated Golgi membrane vesicles were rather leaky, it was difficult to load the membrane vesicles with either of the probes. We have found that comparatively large molecules like ATP could penetrate the bilayer of membrane vesicles that were still able to form a pH gradient.² Therefore, ANTS was encapsulated in asolectin liposomes, and DPX was contained in the reaction mixture. After the initial level of fluorescence reached a plateau, Golgi membranes were added to the mixture (set at time = 0). The fluorescence intensity at 0 min was set at 0% leakage, and that after addition of 0.2% Triton X-100 at 100% for the scale calibration (Fig. 5B). The addition of Golgi membranes resulted in a decrease in ANTS fluorescence, indicating that leakage of ANTS occurred upon membrane fusion, and ANTS and/or DPX would have penetrated the Golgi membranes that fused with the liposomes. The kinetic analysis was also performed in the same way as in the lipid mixing assay (Tables II and III), and the results are shown in Table IV. The initial rate would also reflect the net activity of fusogenic protein(s) that would mediate the pore opening process, and leakage (%) at 15 min was also used to evaluate and compare the extent of fusion/leakage.

View this table: [in this window] [in a new window]   Table IV Kinetic analysis of the leakage of liposomes containing ANTS on incubation with Golgi membranes in the absence or presence of divalent cation chlorides Values ± S.D. are given from n  3 different preparations. The initial rate and fluorescence (%) at time = 15 min were determined and normalized.

Utilizing this ANTS/DPX assay, the effect of the divalent cations tested in Table III on the Golgi-liposome fusion was examined. In the preliminary experiments, these divalent cations did not affect the fluorescence of free ANTS or the quenching by DPX, except that Hg²⁺ ions could not be used as they generated white nebular structures with DPX.

As shown in Fig. 5B and Table IV, Mg²⁺ ions enhanced the leakage, presumably as a result of the increase in fusion monitored by the R18 assay (Fig. 4, Table III), as expected.

Surprisingly, Mn²⁺ ions decreased the leakage (Fig. 5B and Table IV), despite their being highly stimulative of lipid mixing (Fig. 4B, Table III). At 1 mM, SrCl₂ and BaCl₂ also decreased the leakage, although less so than MnCl₂ (Table IV). Addition of 1 mM Ca²⁺ ions affected the leakage slightly (Fig. 5B, red line), but 2 mM ions distinctly inhibited the leakage (Table IV). These results suggest that divalent cations other than Mg²⁺ blocked the later fusion processes involving fusion pore opening and expansion, whereas they enhanced the preceding process where lipid mixing of leaflets and the dilution of lipidic fluorescence probes should occur. This suggestion was supported by the result shown in Fig. 5C. The addition of EDTA during the reaction in the presence of Mn²⁺ ions resulted in a marked leakage of ANTS (Fig. 5C, green line), indicating that lots of liposomes were semi-fused and accumulated on the surfaces of Golgi membrane vesicles by Mn²⁺ ions, which can block the later processes, and a huge number of completion of fusion events involving the accumulated liposomes occurred in a very short time upon removal of Mn²⁺ ions by EDTA. On the other hand, little or no such effect of EDTA on the leakage was observed in the absence of divalent cations (data not shown) or in the presence of Mg²⁺ ions (Fig. 5C, blue line), indicating an absence of such accumulated liposomes.

In this assay, we could not exclude the possibility that divalent cations bound to membranes decreased the permeability of ANTS and DPX, and reduced the leakage. However, EDTA hardly affected the leakage in the presence of Mg²⁺ ions and in the absence of divalent cations, as described. This indicates that the removal of bound Mg²⁺ ions and the remaining bound divalent cations, not chelated by EGTA in the tissue homogenization medium (see "Experimental Procedures"), from the membranes should not affect the leakage. Furthermore, the inhibitory effect of the divalent cations on the leakage was not in proportion to the binding constants for the artificial membranes (Table IV) (41). Therefore, it is unlikely that divalent cations bound to the membranes altered the permeability of ANTS and DPX.

In assays with both calcein and ANTS/DPX, the control for trypsin treatment rendered the same results as the untreated samples, and the leakage was not affected by the 1 M KCl and NEM treatments and monovalent cations tested in Table III. The membrane fraction of Golgi showed the greatest leakages (data not shown). Leakages did not occur when the liposomes were incubated with probe-free liposomes, even in the presence of divalent cations.

In Vitro Fusion of Golgi Membranes with Biological Membranes Was Induced by Divalent Cations-- As shown in Fig. 6A, the fusion of Golgi membranes and asolectin liposomes and the stimulation by divalent cations were also observed when Golgi membranes labeled with R18 were utilized. The Golgi membranes could also fuse to liposomes made of TP lipids, although the fusion efficiency was lower. This would not be the result of the curvature effect because the fusion of large unilamellar vesicles prepared by reverse-phase evaporation (51) from asolectin was nearly the same as that of the small unilamellar vesicles (data not shown). MacDonald (52) reported that the self-quenching of R18 in artificial membranes was enhanced by cholesterol. Therefore, it is possible that TP sterols, consisting of 28 mol % mung bean TP lipids (40), suppressed the de-quenching of R18 upon lipid mixing.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Lipid mixing of Golgi membranes labeled with R18. The reaction was initiated by the addition of R18-free liposomes (A) or TP (B and C). A, fusion of the Golgi membranes with liposomes made of asolectin (with 3 mM MgCl2 or without) or lipids extracted from TP. B, fusion of the Golgi membranes with TP biological membranes induced by divalent cations. Divalent cations were added at the time indicated by arrows (final concentrations, 1 mM and 3 mM). C, effect of trypsin-treatment of the R18-incorporated Golgi membranes on the fusion of biological membranes. MnCl2 was added (an arrow, final 3 mM).

Golgi membranes did not fuse with TP (Fig. 6B, time = 0-2 min) or biological membranes of other fractions (data not shown) following a simple incubation of two membranes. However, the lipid mixing was induced by the addition of divalent cations (Fig. 6B, time  2 min), although the later processes would be hindered in the case of Mn²⁺ and Ca²⁺ ions, in the same way as in the liposome-Golgi fusion (Figs. 4B and 5B, Tables III and IV). The Golgi membranes also fused with membranes of Golgi as well as ER, PM, plus Mt fractions in the presence of the divalent cations (data not shown). This would also have been caused by the fusion system on Golgi membranes. As shown in Fig. 6C, trypsin treatment of the Golgi membranes labeled with R18 seemed to virtually abolish the protein(s)-dependent fusion induced by the divalent cations. In the fusion of biological membranes, spanning domains of membrane proteins would not markedly enhance the self-quenching of R18 as Stegmann et al. (53) reported that proteins in the PM of Sup T1 cells did not enhance the quenching at low concentrations of R18. The induction of fusion events by the divalent cations would also be caused by the decrease in repulsive forces between biological membranes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In this study, the membrane fusion processes of Golgi membranes were examined by lipid mixing assays utilizing R18 (Figs. 2-4 and 6) and content leakage assays utilizing calcein and ANTS/DPX (Fig. 5). The results show that a system that can nonspecifically fuse to other lipid bilayers certainly exists in plant Golgi membranes.

The membranes of ER and PM plus Mt fractions also showed fusogenic activity (Fig. 2, Table II). However, taking account of the level of the Golgi markers (Fig. 1, Table I), it is considered that the fusion of ER and PM plus Mt to the liposomes would be the result of the cross-contamination of Golgi membranes in these fractions, not to fusion of the membranes themselves with the liposomes.

It was reported that in vitro fusion of rabbit liver Golgi membranes and reticulocyte endocytic vesicles with artificial membranes was not affected by Ca²⁺ ions at 1-5 mM (19, 21). However, it was demonstrated in the present study that the fusion of mung bean Golgi membranes can be influenced by divalent cations (Figs. 4-6, Tables III and IV). The experimental results suggest that this fusion system has at least two components that have different sensitivities to divalent cations. One mediates the binding of the lipid bilayer to a fusion intermediate prior to pore opening, the other, the subsequent processes involving fusion pore opening and expansion. The activity of the former is enhanced by divalent cations as a result of a reduction in repulsive forces (Figs. 4 and 6, Table III), whereas that of the latter is directly inhibited by them with the exception of Mg²⁺ ions (Fig. 5B, Table IV). It is possible that more than two integral membrane proteins are involved in the fusion system. It is interesting that the latter fusion process was blocked by Mn²⁺, Sr²⁺, and Ba²⁺ ions, which merely exist in the cells, and non-physiological concentrations of free Ca²⁺ ions (Fig. 5B, Table IV). Although plant cells contain a relatively high content of Mg²⁺ and Ca²⁺ ions in the cytoplasm (54, 55), the cytoplasmic concentration of free Mg²⁺ ions has been estimated to be 0.4 mM in mung bean (54) whereas that of free Ca²⁺ is usually in the order of nanomolar or micromolar (55, 56). Because the cytoplasmic concentration of free Mg²⁺ ions can be in the order of millimolar (54), Mg²⁺ ions would not block the latter fusion process.

Among divalent cations, it is already known that Ca²⁺ ions play a major role in membrane fusion (13, 14) and exocytosis in plants (57, 58). Ca²⁺ ions could act as a signaling factor at concentrations in the order of micromolar, and render a stimulative effect for fusion. However, a concentration of 10 µM was not sufficient to influence the fusion in the present study (data not shown). Therefore, the influence of Ca²⁺ ions on the fusion of Golgi membranes (Figs. 4B, 5B, and 6B) is independent of their potential role as a signaling factor.

In the lipid mixing and content leakage assays, some disagreements in the values of the activity (initial rates) and extent of fusion (fluorescence and leakage, %) were seen. In the calcein assay of untreated samples (Fig. 5A, trace a), the initial rate and leakage (%) at 15 min were calculated to be 51.0%/min and 47.5%, respectively, when the scale was calibrated. These values were distinctly greater than the fusion efficiencies in the R18 and ANTS/DPX assays (Tables II-IV).

In the ANTS/DPX assay (Table IV), the enhancement of the initial rate and leakage (%) by 3 mM Mg²⁺ ions was approximately 2.3 and 1.6 times more effective than by 1 mM Mg²⁺ ions, respectively. However, in the lipid mixing assay, 3 mM Mg²⁺ ions were 1.7 and 1.3 times more effective for the fusion efficiencies, respectively, than 1 mM Mg²⁺ ions (Table III). Likewise, 1 mM Mn²⁺ ions was 2.4 times more effective than 1 mM Mg²⁺ ions for leakage (%) at around 10 min (EDTA added) (Fig. 5C), whereas the same concentration of Mn²⁺ ions was 1.6 times more effective than Mg²⁺ ions in the lipid mixing at 10 min (Fig. 4B), although the more simulative effect of Mn²⁺ than Mg²⁺ ions, evaluated by the addition of EDTA in the case of the ANTS/DPX assay, is consistent in both assays.

These discrepancies would derive from the differences in the experimental approach. The self-quenching of calcein would not be relieved completely on the addition of Triton X-100 in the experimental conditions. The precise reason for the discrepancies of R18 and ANTS/DPX assays remains to be elucidated.

The decrease in repulsive forces between membranes would facilitate not only their approach but also fusion if the membranes have fusogenic properties. In this study, fusion between biological membranes was induced if the repulsive forces were reduced by divalent cations (Fig. 6B), whereas fusion of Golgi membranes with liposomes occurred even in the absence of divalent cations (Figs. 2-5 and 6A) or even if KCl was absent from the buffer (Fig. 3B). Biological membranes contain proteins and polysaccharides in addition to the lipid bilayer. The existence of surface peptides projecting from the bilayer including peripheral proteins and polysaccharides of glycoproteins would cause, in particular, an increase of aquatic repulsive force as well as structural obstacles, which would hinder the approach of membranes. Therefore, the reduction of repulsive forces by divalent cations would be required for the fusion peptide(s) in Golgi membranes to become accessible to the lipid bilayer of other biological membranes.

What is the biological role of the fusion system of Golgi membranes observed in this study? It is natural to consider that the system would take part in fusion events in the cells. Because the Golgi apparatus of plant cells does not fragment and reassemble during mitosis as in mammalian cells (59), the fusion system would be operating as part of a secretory pathway and function in the fusion of transport vesicles. However, unlike viral and other cellular fusion peptides (60), the fusion peptide(s) alone would not suffice to induce membrane fusion in the cells because the Golgi membrane seemed to poorly fuse with biological membranes in the cell-free system even in the presence of Mg²⁺ ions (Fig. 6B). Therefore, it will be necessary to bring transport vesicles close enough together for the fusion peptide(s) on Golgi membranes to access the lipid bilayer of the vesicular membranes and induce fusion. It is possible that the fusion peptide(s) cooperate(s) with other fusion factors at the Golgi apparatus, e.g. with SNARE proteins in the terminal step (15).

An alternative possibility is that the fusion system plays a role in membrane contact for the transport of membrane lipids (61), as suggested for the 50-kDa fusogenic glycoprotein of rat liver ER membrane (23). It is possible that the fusion system enables the Golgi apparatus to temporally fuse to the lipid-enriched microdomain of ER to import newly synthesized lipids.

The next aim is to identify the proteins comprising the fusion system of the Golgi apparatus, and to attempt to explore the biological relevance of the described phenomena. It would be especially interesting to resolve the linkage of the component that processes fusion intermediates and the other that mediates the subsequent processes. This would provide more detailed information about intracellular membrane fusion. It is also of interest to identify in which part of the Golgi apparatus, the cis-, medial-, and/or trans-Golgi, the fusion system is located.

	ACKNOWLEDGEMENTS

Anti-VM23, anti-BiP, JIM 84, and anti-maize PM H⁺-ATPase antibodies were kindly provided by Dr. M. Maeshima (Nagoya University, Nagoya, Japan), Dr. I. Hara-Nishimura (Kyoto University, Kyoto, Japan), Dr. C. Hawes (Oxford University, Oxford, United Kingdom), and Dr. H. Matsumoto (Research Institute for Bioresources, Okayama University, Kurashiki, Japan), respectively.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed. Current address: Molecular Membrane Biology Laboratory, RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan. Fax: 81-48-462-4679; E-mail: y-takeda@postman.riken.go.jp.

Published, JBC Papers in Press, October 3, 2002, DOI 10.1074/jbc.M209199200

² Y. Takeda and K. Kasamo, unpublished data.

	ABBREVIATIONS

The abbreviations used are: SNARE, soluble N-ethylmaleimide-sensitive factor attachment protein receptor; ER, endoplasmic reticulum; Mes, 4-morpholineethanesulfonic acid; TP, tonoplast; PM, plasma membrane; Mt, mitochondria; R18, octadecylrhodamine B; PL, phospholipid; ANTS, 1-aminonaphthalene-3,6,8-trisulfonic acid; DPX, N,N'-p-xylylenebis(oyridinium bromide; DTT, dithiothreitol; NEM, N-ethylmaleimide; Mops, 4-morpholinepropanesulfonic acid.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				V. J. Starai, N. Thorngren, R. A. Fratti, and W. Wickner Ion Regulation of Homotypic Vacuole Fusion in Saccharomyces cerevisiae J. Biol. Chem., April 29, 2005; 280(17): 16754 - 16762. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/49/47756    most recent M209199200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Takeda, Y. Articles by Kasamo, K. Search for Related Content PubMed PubMed Citation Articles by Takeda, Y. Articles by Kasamo, K. Social Bookmarking                      What's this?