Vacuolar-Type H+-ATPases Are Associated with the ... - NCBI

Plant Physiol. (1994) 106: 1313-1324

Vacuolar-Type H+-ATPases Are Associated with the Endoplasmic Reticulum and Provacuoles of Root Tip Cells’ Eliot M. Herman, Xuhang Li, Robert T. Su, Paul Larsen, H e i 4 Hsu, and Heven Sze* Department of Botany, University of Maryland, College Park, Maryland 20742 (X.L., P.L., H.S.); Plant Molecular Biology Laboratory (E.M.H.) and Florist and Nursery Crops Laboratory (H.-t.H.), United States Department of Agriculture, Agricultural Research Service, Beltsville, Maryland 20705; and Division of Research Grants, National Institutes of Health, Bethesda, Maryland 20892 (R.T.S.)

To understand the origin of vacuolar H+-ATPases (V-ATPases) and their cellular functions, the subcellular location of V-H+ATPases was examined immunologically in root cells of oat seedlings. A V-ATPase complex from oat roots consists of a large peripheral sector (V,) that includes the 70-kD (A) catalytic and the 60-kD (6)regulatory subunits. The soluble V1 complex, thought to be synthesized in the cytoplasm, i s assembled with the membrane integral sector (V.) at a yet undefined location. In mature cells, VATPase subunits A and B, detected in immunoblots with mqnoclonal antibodies (Mab) (7A5 and 2E7), were associated mainly with vacuolar membranes (20-22% sucrose) fractionated with an isopycnic sucrose gradient. However, in immature root tip cells, which lack large vacuoles, most of the V-ATPase was localized with the endoplasmic reticulum (ER) at 28 to 31% sucrose where a major ER-resident binding protein equilibrated. The peripheral subunits were also associated with membranes at 22% sucrose, at 31 to 34% sucrose (Golgi), and in plasma membranes at 38% sucrose. lmmunogold labeling of root tip cells with Mab 2E7 against subunit B showed gold particles decorating the ER as well as numerous small vesicles (0.1-0.3 pm diameter), presumably provacuoles. The immunological detection of the peripheral subunit B on the ER supports a model in which the V1 sector is assembled with the V ,, on the ER. These results support the model in which the central vacuolar membrane originates ultimately from the ER. The presence of V-ATPases on several endomembranes indicates that this pump could participate in diverse functional roles.

In plants, acidification of the vacuolar compartment by the V-ATPase is essential to or involved in many diverse functions (Sze et al., 1992a). Depending on the tissue, the stage of development, and the signals received, these functions include osmoregulation, transport and storage of ions and metabolites, signal transduction, storage and turnover of proteins, and storage of secondary metabolites, defense proteins, and pigments (Boller and Wiemken, 1986; Martinoia, 1992). Vacuoles are dynamic, prominent organelles. Undifferentiated and immature plant cells often possess proportionately more cytoplasm that contains an extensive endo-

membrane system and numerous small provacuoles. As cells mature, these small provacuoles merge to form a central vacuole that can occupy up to 80 to 90% of the intracellular space in differentiated cells. V-ATPase from plants is a member of the V-type ATPases widely distributed among many eukaryotes (Nelson, 1992), including animal (Forgac, 1992; Gluck, 1992), fungal (Anraku et al., 1992; Bowman et al., 1992; Kane and Stevens, 1992), and plant cells (Sze et al., 1992a). The characteristic feature of this type of H+ pump is its sensitivity to nanomolar levels of bafilomycin, but not to vanadate or azide, which are inhibitors of the P- or F-type ATPases, respectively. In plants, the anion-sensitive V-ATPase is directly stimulated by 10 mM chloride and inhibited by 10 to 50 mM nitrate (Sze, 1985). The purified enzyme complex from several plants contains 7 to 10 different subunits. A peripheral sector, or Vl, is composed of 5 to 6 different subunits that are solubilized from the membrane by KI (Parry et al., 1989; Ward and Sze, 1992a). The VI complex includes three copies each of the nucleotide-binding catalytic (approximately 70 kD) and the regulatory (approximately 60 kD) subunits, also known as subunits A and B, respectively. The membrane integral sector, or V,, which forms the proton-conducting pathway, is made up of six copies of the 16-kD proteolipid together with 1 to 3 other subunits (Sze et al., 1992a). In addition to elucidating the complex structure of VATPase, biochemical and molecular studies have provided evidence that plants contain a family of V-type ATPases. The first clue came from the diversity and variations in the subunit composition of the purified enzyme. For example, V-ATPase purified from red beet and barley contained a prominent 100kD integral subunit (Parry et al., 1989; DuPont and Morrisey, 1992); however, this subunit was absent from a purified and Abbreviations: BiP, a major ER-resident binding protein; FBS, fetal bovine serum; GERL, Golgi-ER-lysosome system; Mab, monoclonal antibody(ies); PM, plasma membrane; subunit A, the approximately 70-kD catalytic subunit of the V-ATPase; subunit 8, the approximately 60-kD regulatory subunit of the V-ATPase; TBST, Trisbuffered saline with Tween; TPCK, N-tosyl-L-phenylalanine chloromethyl ketone; V-ATPase, vacuolar H+-pumping adenosine triphosphatase; VI, the peripheral complex of the V-ATPase, which consists of subunits A and B plus three or four other subunits; VM23 or TIP, major tonoplast intrinsic protein of approximately 23 to 25 kD; V,, the membrane integral complex of the V-ATPase.

This work was supported in part by National Science Foundation grant DCB-90-06402 and by Maryland Agricultural Experiment Station (MAES) Hatch Project MD-J-151 to H.S. and by FY93 MAES competitive grant to E.M.H. and H.S. (contribution No. 8797, article NO. A-6585). * Corresponding author; fax 1-301-314-9082. 1313

1314

Herman et al. '

transport-competent ATPase from oat (Avena sativa L.) roots (Ward and Sze, 1992b). Molecular studies clearly show that multiple genes encode the major subunits: subunit A (Starke et al., 1991), subunit B (Berkelman et al., 1994), and the 16kD proteolipid (Lai et al., 1991). These results support the idea that there are several isoforms or subtypes of V-ATPases; however, it is not clear how the subtypes differ from one another in their cellular function and regulation. One working model is that the subtypes are spatially, temporally, or developmentally regulated, and that there may be organellespecific isoforms (Lai et al., 1991; Gogarten et al., 1992). Very little is known about the subcellular distribution of the V-ATPase from plants. Vacuoles purified from Kalanchoe leaves (Smith et al., 1984) or red beet roots (Bennett et al., 1984) possess an anion-sensitive ATPase activity. This finding led to the conclusion that the anion-sensitive ATPase was associated with the vacuolar membrane (Bennett et al., 1984; Mandala and Taiz, 1985). Consequently, this pump is often referred to as a vacuolar membrane marker. However, fractionation of microsomes showed that the V-ATPase is broadly distributed in linear Suc or dextran gradients. Several laboratories have suggested that the pump is associated with the ER and Golgi, based on co-migration of V-ATPase activity with ER (Churchill et al., 1983; Hager and Biber, 1984) and Golgi marker enzymes (Chanson and Taiz, 1985; Ali and Akazawa, 1986). Furthermore, immunoblotting with V-ATPase antibodies suggested that this enzyme is also found on purified clathrin-coated vesicles (Depta et al., 1991). These results would indicate that the V-ATPases have a broader distribution and physiological role than was previously thought. The most direct approach to verify the subcellular distribution of membrane proteins is by immunogold EM. In one study using polyclonal antibodies to subunit A, the VATPase was located on the Golgi, the vacuole, and, surprisingly, the PM of corn coleoptiles (Hurley and Taiz, 1989). Although V-ATPases may be located on several endomembranes in plants, as in animals (Marquez-Sterling et al., 1991), it is not clear where this complex is synthesized and assembled. The current model based on work with yeast mutants defective in one V-ATPase subunit would suggest that the V, and VI complex are independently synthesized (Kane, 1992; Kane and Stevens, 1992). Like other multisubunit membrane complexes (Hurtley and Helenius, 1989), the V, complex is thought to be synthesized and assembled in the ER. However, evidence indicates that the VI subunits are independently synthesized and assembled in the cytosol. The VI is then assembled with the V, sector at a yet undefined location, resulting in a functional H+-ATPase in the vacuolar membrane. For example, yeast mutants in which the gene encoding the 16-kD proteolipid was disrupted were still capable of synthesizing the subunits of the VI; however, the VI remained in the cytosol and did not associate with the membrane (Kane and Stevens, 1992). Alternatively, if a gene encoding subunit A or B of the VI complex was disrupted, the major V, subunits (95 and 16 kD) were synthesized and associated to the membrane. Thus, an intact V, complex is required before V, can be attached to the membrane. To understand better the cellular roles of V-ATPases and as first step in the study of the biogenesis of this proton pump, we have examined the subcellular distribution of

Plant Physiol. Vol. 106, 1994

vacuolar ATPases in the roots of oat seedlings. Monocot roots serve as an ideal system in which to study the dev4opment and formation of vacuoles. Undifferentiated and meristematic cells are located at the root tip, and differentiated cells, like the epidermis, cortex, and procambium tissues, appear a few millimeters behind the tip. Furthermore, there is substantial background information about oat root, which has heen used extensively for studying PM and V-ATPases after membrane fractionation (Hodges and Leonard, 1974). Usin,; a Mab against subunit B for immunological assays, we show that the peripheral subunit B is located on the ER and ER fractions, as well as on Golgi-derived vesicles and provacuoles. The results support a model in which the V-ATPase is sy ithesized and assembled at the ER and in which the vacuolar membrane originates from the ER. The presence of V-ATPases on several endomembranes suggests that these H+ punips could play diverse roles in the growth and developmení of plant cells. MATERIALS AND METHODS Plant Material

Oat (Avena sativa L. var Ogle) seeds were germinated in the dark over an aerated solution of 0.5 mM Caso4.Roots or root tips were harvested after 4 d of growth. Fractionation of Subcellular Membranes

Membrane vesicles were prepared using differential and linear density gradient centrifugation according to the procedures of Ward and Sze (1992a) and Hager et al. (1991) with some modifications. Mature root sections (approximately 0.5 g fresh weight) or root tips (approximaíely 0.1 g fresh weight) were homogenized with a mortar and pestle in 50 mM Hepes-bis-tris propane, pH 7.4, 250 mM sorbitol, 6 mM EGTA, 1 mM DTT, 0.1 mM PMSF, and 0.05 mM TPCK at a medium-to-fresh tissue ratio of 6 or 20 mL/g fcr mature roots or root tips, respectively. Mature roots refer to approximately 4-cm sections excised 1 cm from the tip, and root tips were 1- to 2-mm tip sections. The homogenate was then strained through four layers of cheesecloth. The debris and mitochondria were removed by centrifugation at 13,OOOg for 15 min. The supematant was layered onto a 17 to 45% continuous Suc gradient (12 mL) over a I-mL 45% Suc cushion. The SUCgradient solution contained 25 m14 Hepesbis-tris propane, 2 mM EGTA, 1 mM DTT, 0.1 mM PMSF, and 0.05 mM TPCK, pH 7.4. After centrifugation at 110,OOOg for 16 h (Beckman SW28, at maximum radius), 0.7-mL fractions were collected and diluted to less than 10% Suc in gradient solution. Proteins in each fraction were directly prwipitated with 20% TCA (final concentration) and washed with cold acetone. Protein concentration was determined using BioRad protein assay solution. SDS-PACE, Ag Stain, and lmmunoblot

Protein in each fraction was precipitated with 213% TCA (final concentration), washed with 100% acetone, and then solubilized in 60 pL of sample buffer containing 62.5 mM

Vacuolar H+-ATPase Associated with the ER and Provacuoles Tris-HC1, pH 6.8, 10% (v/v) glycerol, 8 M urea, 5% 2mercaptoethanol, and 0.002% (w/v) bromphenol blue. An equal volume fraction (typically one-fifth of 60 FL containing 2-4 r g of protein) was loaded on each lane. The proteins were separated on an 11%acrylamide gel, 15 X 20 cm, at 7 mA per gel ovemight at 15OC. One gel was silver stained for proteins. After electrophoresis, gels were soaked in 25 II~MTris, pH 8.3, 192 r m Gly, and 20% methanol for at least 10 min. Proteins were blotted onto Immobilon-P (Millipore, Bedford, MA) at constant voltage of 100 V for 2 h at 4OC using a BioRad blotting apparatus. The Immobilon-P was incubated in TPBS (PBS with 0.1% Tween-20) containing 5% dry milk and 1%protease-free BSA (Sigma) for 1 h, and washed three times with TPBS for 5 min. The membrane was then incubated with Mab or polyclonal antibodies diluted with TPBS containing 1% BSA for 1 h and washed as above. The membrane was probed with either goat anti-mouse IgG (Sigma) or goat anti-rabbit IgG (Calbiochem,San Diego, CA) conjugated to alkaline phosphatase, and color was developed with 5-bromo-4-chloro-3-indoylphosphate and nitroblue tetrazolium (Sigma). Antibodies used in these experiments include Mab 2E7 and 7A5 against the 60- and 70-kD subunits of the V-ATPase (Ward et al., 1992), and polyclonal antibodies anti-holo-V-ATPase(Ward and Sze, 1992a), anti-BiP (M. Chrispeels, San Diego, CA), anti-plasma membrane H+-ATPase (R.T. Leonard, Riverside, CA), and anti-VM23 (M. Maeshima, Sapporo, Japan). Mab and Ascites Fluid

Mab against the two major peripheral subunits of the oat V-ATPase were generated initially in mice as described by Ward et al. (1992). Hybridoma supematants of Mab 2E7 and 7A5 (both IgGJ at dilutions of 1:lOO to 1:200 were used in westem blots to detect subunits B and A, respectively. We have used Mab 7A5, instead of Mab 7D2, to detect subunit A because 7A5 reacts with all plants tested, whereas 7D2 is oat specific (Ward, 1991). Because immunolocalization studies using the hybridoma supematants showed very weak labeling, we generated ascites fluid. The hybridoma cell lines 2E7 and 7A5 were grown in RPMI medium containing 20% fetal calf serum. Cells (lo6) were then washed with serum-free RPMI and introduced peritoneally into pristine-primed BALB/c mice (Hsu and Lawson, 1985). Two to 4 weeks after transplantation of the hybridoma cells, ascites fluid was collected and further purified with a protein G-Sepharose column (Harlow and Lane, 1988). The reactivity of each fraction with membrane-bound ATPase was tested by westem blotting. Fractions containing an antibody titer of 1:10,000 were pooled and stored at -2OOC in the presence of 0.1% Na azide. The labeling pattern of 7A5 was less specific, so 2E7 alone was used for immunolocalization studies. Immunofluorescent Staining of V-ATPase

Immunofluorescent labeling of oat roots was performed using the method described by Baskin et al. (1992). Root tips, 2 to 3 mm in length, were cut from 4-d-old seedlings and

1315

transferred to Eppendorf tubes containing 4% paraformaldehyde in 15 m Hepes buffer, pH 7.2, with 15 II~MKC1. After fixation for 1 h, roots were rinsed thoroughly with 15 m KCl in 15 m Hepes, pH 7.2. After dehydration stepwise in ethanol, roots were embedded in methacrylate (Ladd Research Industries, Inc., Burlington, VT) (Baskin et al., 1992). Sections approximately 2 pm thick were placed on polylysine (Sigma) coated slides in preparation for immunofluorescent labeling. All slides were camed through the following staining regimen. Slides with tissue sections were placed in acetone for 15 min to remove the embedding plastic. After rehydrating with PBS (pH 7.2) for 15 min, the slides were put in two changes of 0.1% Na borohydride in PBS for 10 to 15 min to remove residual fixative and then rinsed with PBS. Sections were blocked with 1%BSA in PBS for 10 to 15 min and then blotted to remove the BSA. Sections were then incubated with purified ascites fluid of 2E7 (1:l ddution) (30 pL/section) ovemight at 4OC in a humid chamber. To test for autofluorescence and unspecific staining of secondary antibody, control sections were treated with BSA/PBS. After three rinses with PBS, all the slides were blotted and then incubated with 30 pL of secondary antibody, sheep antimouse IgG conjugated with Texas Red (1:lO dilution in 1% BSA/PBS) (Jackson Immunoresearch Laboratories, Inc., West Grove, PA). All sections except the autofluorescence control were rinsed three times in PBS and stained for 30 min with 30 pL of Hoechst 33258 (5 pg/mL in PBS, Molecular Probes, Eugene, OR), which binds DNA. After rinsing with PBS, all the sections were mounted in Mowiol 4-88 (Calbiochem) with a coverglass and sealed with nail polish. Sections were viewed with a Zeiss IM 35 microscope adapted for incident light fluorescent microscopy using a Zeiss lOOX achromat (numerical aperture = 1.3) objective lens. Sections were illuminated for fluorescence with a 75-W xenon arc lamp. An IR interference filter was placed in the optical path to narrow the excitation bandwidth. Irradiation with UV/violet light (approximately 365 nm) to excite the Hoechst dye and at 530 to 585 nm to excite the Texas Red was produced from standard filter packages (Carl Zeiss, New York, NY). Images were photographed with an Olympus OM-2N camera using T-Max 3200 film and the film was developed with T-Max developer (Eastman Kodak, Rochester, NY). lmmunolocalization with the Electron Microscope

Immunocytochemical analysis was performed as described by Herman and Melroy (1990). Specifically, oat root tips (apical 2 mm) of 4-d-old oat seedlings were fixed in 4% formaldehyde, 2% glutaraldehyde, 0.1 M phosphate buffer (pH 7.4) at 7OC. Parallel aldehyde-fixed roots were postfixed in 1%OsO, for 2 h at room temperature to provide samples for ultrastructural analysis. The cells and tissues were dehydrated in a graded ethanol series and embedded in hardgrade L.R. White resin. Ultrathin sections mounted on nickel grids were blocked with 10% FBS in TBST (10 m Tris-C1, pH 7.4, 0.15 M NaC1, 0.05% Tween-20). The grids were incubated in either purified IgG 2E7 (1:5 dilution) or control monoclonal IgG diluted in FBS/TBST. The 2E7 Mab was purified from the ascites fluid by affinity chromatography on

Herman et al.

1316

protein G-Sepharose (Harlow and Lane, 1988). The control Mab IgG consisted of an antibody directed at a thiol protease expressed in maturing soybean seeds that was also purified by protein G-Sepharose chromatography (Kalinski et al., 1992). The grids were then washed in TEST and indirectly labeled with anti-mouse IgG coupled to 10 nM colloidal gold (Ted Pella, Inc., Redding, CA) diluted 1:2 in FBS/TBST for 10 min at room temperature. The grids were washed in TEST and distilled water and then stained in 5% uranyl acetate for 30 min. For conventional EM, tissues were fixed in osmium, sectioned, and stained with 5% uranyl acetate for 30 min and lead citrate (33 mg/mL) for 10 min. The grids were examined and photographed with Hitachi H300 and H500 electron microscopes.


weight of roots was homogenized and the post-mitochondrial supernatant was directly separated with a Sue gradient (12 mL) containing 2 HIM EGTA (Hager and Biber, 1984). The polypeptide profile indicated that the subcellular membranes had been separated during gradient fractionation (see below). In mature roots (without tip) that contain cells with fully differentiated vacuoles, subunits A and B of the V-ATPase were associated with several endomembranes at Sue concentrations ranging from 20 to 32% (Fig. IB). The highest levels were detected in membranes at 20 to 22% Sue, where mature vacuolar membranes equilibrate (e.g. Bennett et al., 1984). Significant levels of the peripheral subunits were also detected at 28 to 31% Sue, where ER membranes equilibrated. Using anti-BiP, we show that BiP, a 78-kD ER lumen chaperone, was predominantly associated with membranes of 28 to 31% Sue (Fig. 1A). Interestingly, we have frequently observed low levels of V-ATPase subunits at 38 to 41% Sue, where PMs equilibrate, as detected by the PM H+-ATPase of 100 kD (Fig. 1A). In general, the V-ATPase distributed in parallel with the H+-PPase (not shown) and VM23 (Fig. 1C). VM23, or TIP, is a major integral protein of approximately 23 kD (Johnson et al., 1989; Maeshima, 1992) that mediates water transport across vacuolar membranes (Maurel et al., 1993). Two immunoreactive polypeprides (Fig. 1C) may be indicative of two isoforms of VM23 (Maeshima, 1992) as observed for TIP in legume seeds (Johnson et al., 1989). As a first step in determining the origin of V-ATPases, we examined the distribution of the V-ATPase in immature,

RESULTS

Distribution of V-ATPase in Membranes Separated on a Linear Sue Gradient To determine the subcellular distribution of V-ATPases, membranes from 4-d-old seedling roots were separated with a linear isopycnic Sue gradient, and Mab were used to detect V-ATPase subunits in western blots. Previously, we have shown that Mab 2E7 and 7A5 specifically react with the peripheral subunits of approximately 60 and 70 kD, in either native membranes or in the purified enzyme from oat roots (Ward et al., 1992). To minimize aggregation of membrane vesicles to one another, approximately 0.1 or 0.5 g fresh

MATURE ROOTS 70

SUCROSE

45

44

41

38

36

35

34

33

32

31

30

29

28

27

25

22

20

17

14

11

10

PM-ATPase —

BiP —

70 kD —

B

60 kD —

VM23 —

10

12

14

16

18

20

FRACTION NO. Figure 1. Peripheral subunits A and B of V-ATPases isolated from mature oat roots are located on vacuolar membranes and other endomembranes. The post-mitochondrial supernatant from approximately 0.5 g fresh weight of mature root sections was separated with a 17 to 45% Sue gradient (12 ml). After centrifugation and fractionation (0.7 ml each), each fraction was diluted to 10% Sue and the proteins were precipitated with TCA and washed in acetone. Equal-volume fractions (one-fifth), or approximately 2 to 6 jig of protein, were analyzed by SDS-PACE (11% acrylamide). Results are from one experiment representative of two. A, Markers of the PM and ER equilibrate at 38 to 41% and 28 to 30% Sue, respectively. Gels were blotted to Immobilon P. Western blots were probed with a mixture of polyclonal anti-PM-ATPase (1:2000 dilution) and anti-BiP (1:30). B, Subunit A (70 kD) and subunit B (60 kD) of V-ATPase are associated mainly with vacuolar membranes (20-22%) and ER (28-31%). Western blots were probed with a mixture of Mab 2E7 (1:100 dilution) to B and Mab 7A5 (1:50) to A. C, VM 23 co-migrates with V-ATPase and is associated with the vacuolar membrane and ER. Western blot was probed with polyclonal anti-VM23 (1:3000 dilution).

Vacuolar H+-ATPase Associated with the ER and Provacuoles

A

SILVER

SUCROSE kD

45

41 43

35 37

33 34

31 32

29 30

26 27

22 24

17 19

13 14

9 11

9

220 — 105 —

75 —

46 —

27 —

1317

icant levels were detected in membranes at 22% Sue (mature vacuoles) and at 31 to 34% Sue, where Golgi marker IDPase distributed (Hodges and Leonard, 1974). These results indicate that peripheral subunits of V-ATPase were associated primarily with the ER of root tips. The distribution of the major peripheral subunits, A and B, reflected the presence of a fully assembled holoenzyme. Immunostaining with polyclonal antibodies to the purified enzyme (Ward and Sze, 1992a) confirmed that the major subunits of the V-ATPase, including the 95-, 70-, 60-, 48-, 36-, and 16-kD polypeptides, were associated with membranes equilibrating at 23 to 35% Sue (Fig. 3A). These results resemble the broad distribution of vanadate-insensirive or nitrate-inhibited ATPase activity (24-32% Sue) previously

Immunoblot, anti-ATPase (Root tips) 19 — 16 — SUCROSE 1*1

g

10

12

14

16

18

20

22 44

FRACTION NO. "

SUCROSE PM-ATPasc—

BiP—

41

43

-

37

35

34

33

32

31

30

29

27

26

36

34

30

26

23

2O

16

13

12

kDa - 105

95-

ROOT TIPS

B

41

kDa

- 75

24

22

19

17

14

13

11

9

9

7060-

- . « . _ . - _ _ _ _

—_ _ — _ w * . _ — -

_

70 kl> —

60 kD — 10

12

I-I

20

FRACTION NO.

Figure 2. Subunits A and B of V-ATPases were distributed in several endomembranes, especially the ER of oat root tips. The postmitochondrial supernatant from 0.1 g fresh weight of root tips (2 mm) was separated with a 17 to 45% Sue gradient (12 ml). Fractions of 0.7 ml each were precipitated with TCA, washed with acetone, and solubilized in sample buffer. Equal-volume fractions (one-fifth, containing 2-4 ,ug of protein) were then separated by SDS-PACE (11% acrylamide). Results are from one experiment representative of two. A, Silver-stained gel shows separation of membrane polypeptides. B, Top, ER and PM markers peak at 27 and 41% Sue, respectively, but the PM-ATPase is also associated with the ER of root tips. Western blot was probed with a mixture of polyclonal antibodies to the PM ATPase (1:2000) and ER BiP (1:30). Bottom, Both subunits A and B of V-ATPases are abundant in the ER as well as other endomembranes. Western blot was probed with a mixture of Mab2E7 (1:100 dilution) and Mab 7A5 (1:50).

differentiating cells that lack large vacuoles. On a fresh weight basis, root tips were highly enriched in V-ATPases by 5- to 10-fold relative to mature root sections. Therefore, only approximately 0.1 g fresh weight of root dps 2 mm in length were homogenized, and the post-mitochondrial supernatant was separated on a Sue gradient. The major peripheral subunits of 70 and 60 kD were predominantly associated with ER at 27 to 30% Sue (Fig. 2B, lower panel); however, signif-

19-

- 19

16-

- 16

2

4

6

8

W

12

14

16

18

20

22

FRACTION NO. PM Fraction

B Sucrose It]

ER

Fraction

44

41

36

34

30

26

23

20

16

13

2

4

6

8

10

12

14

16

18

20

60-

Fraction No.

Figure 3. Distribution and diversity of V-ATPase holoenzyme in membranes from oat root tips. Membranes were fractionated with a Sue gradient and separated by SDS-PAGE as in Figure 2. A, Western blot was probed with polyclonal antibodies to the VATPase holoenzyme (1:300 dilution). Blot shows the presence of a putative 95-kD subunit and the diversity of the approximately 36to 38-kD subunits between vacuolar membranes at 20 to 22% Sue and endomembranes at 24 to 26% and at 28 to 32% Sue. The polypeptide of approximately 65 kD is not part of the purified VATPase complex (Ward and Sze, 1992a). Fractions 17 to 22 contained solubilized proteins that were dissociated from the membrane. B, Fractions from the same experiment were western blotted and probed with Mab 2E7. Subunit B appeared in the PM fraction (38-41% Sue) as well as other endomembranes.

1318

Herman et al.


Immunolocalization of V-ATPase by Light and EM

We verified the subcellular distribution of the V-ATPase by immunocytochemical localization of V-ATPases in root tip cells. Initial attempts to use 2E7 tissue culture supernatant proved to be unsuccessful in immunoelectron microscopy due to the low concentration of antibody present in these preparations. So we generated ascites fluid of 2E7 and 7A5 to subunit B and A, respectively, and purified the specific IgG by affinity chromatography with immobilized protein G. Purified ascites fluid 2E7 and 7A5 reacted specifically with subunits B and A, respectively, in western blots at 1:5000 dilutions. However, in preliminary tests of immunoelectron microscopy, 2E7 labeled endomembrane compartments more effectively than 7A5 and was therefore used for all subsequent studies. Immunofluorescent Labeling of Root Tip Cells

Cellular levels of the V-ATPase were visualized in semithick longitudinal sections of root tips that were labeled with mouse ascites 2E7 followed by anti-mouse IgG conjugated to Texas Red. Small, brightly fluorescent aggregates were observed residing in the perinuclear region of the cells in the cortex (Fig. 4A) and protoderm. The labeled structures appeared to be morphologically similar to small vacuoles observed in cells of root tips (Fig. 5). The large central nucleus

PROCAMBUM

CORTEX

Figure 4. Immunofluorescent staining by 2E7, a Mab to subunit B of V-ATPases, of perinuclear regions in root tip cells of 4-d-old oat seedlings. Root tips were fixed, embedded in methacrylate, sectioned, and stained with mouse ascites 2E7 and anti-mouse IgC conjugated to Texas Red. A, Bright aggregates of Texas Red fluorescence appear around the nuclei and in the cytoplasm of cortical cells. B, Location of nuclei is detected by Hoechst 33258 fluorescence in the same cells as in A. C, Autofluorescence and nonspecific binding of anti-mouse IgG conjugated to Texas Red is relatively low. Control sections were treated as above, except the primary antibody 2E7 was omitted. All the figures were photographed, exposed, developed, and printed at the same time under identical conditions. Bar = 10 jim.

observed in oat roots and in corn coleoptiles (Churchill et al., 1983; Hager and Biber, 1984). Based on co-migration with enzyme markers of the ER and Golgi, it has been suggested that the V-ATPase could be associated with the tonoplast, as well as the ER and Golgi membranes (Churchill et al., 1983; Hager and Biber, 1984). Furthermore, the immunoblot with Mab 2E7 (Fig. 3B) showed V-ATPase association with the PM fraction (38-41% Sue). Interestingly, V-ATPases associated with various endomembranes showed some diversity in their subunit composition, especially in the size of a 36-to 38-kD polypeptide (Fig. 3A).

Figure 5. Light micrograph showing a longitudinal section of the root tip of 4-d-old oat seedling. Roots were fixed in glutaraldehyde, embedded in Epon, sectioned, and stained with toluidine blue O. Dark material outside the protoderm (PD) is the mucilagenous layer. Bar = 5 Mm-

Vacuolar H+-ATPase Associated with the ER and Provacuoles

in the same section could be observed by fluorescence from Hoechst 33258 dye, which binds DNA (Fig. 4B). Control sections exhibited little autofluorescence (Fig. 4C). Little or no labeling with 2E7 was detected in root sections taken approximately 1 cm above the tip (not shown), a region containing mature differentiated cells with large central vacuoles. Immunogold EM

Immunofluorescence microscopy had insufficient resolution to identify the subcellular structures that contain VATPase, so additional immunocytochemical assays were undertaken at the EM level. The distribution of endomembrane V-ATPases may parallel the ontogeny of vacuoles, so we initially analyzed root tip cells after aldehyde and OsO4 fixation by conventional EM. At low magnification, the undifferentiated cells of the root had a dense cytoplasm and contained numerous small vacuoles (approximately 1 /tm diameter) as indicated by light microscopy (Fig. 5). At higher magnification as seen with EM, cells contained extensively elaborated ER and abundant Golgi dictyosomes and were filled with numerous cytoplasmic vesicles (Figs. 6 and 7). The Golgi consisted of eight or so closely appressed cisternae that exhibit polar differences in electron density from cis to trans

1319

(trans being the most electron dense) (Fig. 6). Electron-translucent vesicles of 0.1 to 0.3 nm diameter were associated with the trans face of the Golgi, apparently resulting from budding at this face (Fig. 6). We suggest that these Golgi-derived structures are putative provacuoles. In some instances there was apparent direct continuity between the putative provacuoles and tubular elements radiating from the Golgi apparatus (Fig. 6). The fusion of small vacuoles is widely believed to mediate the formation of the large central vacuole. Previous investigations have suggested that vacuoles form as a result of fusion by extended trans Golgi cisternae or branched ER tubules that sequester and displace the cytoplasm (Marty et al., 1980; Amelunxen and Heinze, 1984). Examination of the trans Golgi region and adjacent cytoplasm indicated that vesicles or membrane tubules aggregated and apparently fused to form extended structures (Fig. 6). Vesicles oriented in single file appeared to be in the process of fusion (Fig. 6). Numerous small vesicles were frequently seen surrounding an enlarging vacuole of 1 to 2 nm in diameter (Fig. 7). The membrane between the adjacent provacuoles appeared to be internalized within the vacuole as a result of fusion. The membrane within the vacuole seemed to be partially degraded (Fig. 7, Iv). Immunolocalization of the peripheral 60-kD subunit with

Figure 6. Interrelationship between provacuoles (Pv) associated with the Colgi apparatus (G) and the aggregation and fusion of the provacuoles. Provacuoles exhibit direct continuity with the medial and trans cisternae of the Colgi apparatus (arrowheads). Provacuoles immediately adjacent to the Colgi apparatus are oriented in a single-file array and appear to be in the process of fusing. Also shown is an elongated, flattened membrane that is continuous with the provacuole membranes (arrows). Bar = 0.5 ^m.

Herman et al.

1320


Figure 7. Association of provacuoles (Pv) with vacuoles (V). Vacuolar enlargement appears to result from the fusion of provacuoles that completely surround the vacuole shown. The process of provacuole fusion to the vacuole appears to result in the sequestration of membranes within the vacuole. Iv, Intravacuolar material.

ER

8b 0.5pm

Figure 8. Immunogold localization of peripheral subunit B of the V-ATPase in the ER. In a, the gold particles are distributed throughout the segments of the ER. In contrast, b shows an example where the gold particles are bound to only part of the continuous segments of ER, indicating that there may be spatial differences in V-ATPase content within the ER. Small vacuoles (V) are also labeled with gold particles.

Vacuolar H^-ATPase Associated with the ER and Provacuoles purified Mab 2E7 in root tip cells resulted in gold particle labeling of small vesicles, provacuoles, and ER (Figs. 8 and 9). Control Mab IgGi directed against a thiol protease p34 (Kalinski et al., 1992) did not label oat root cells (not shown). The ER was identified by its characteristic cisternal structure and associated ribosomes (Fig. 8). Both the ER and vesicle label appeared to be associated with the membrane. The ER appeared to exhibit spatial variation of gold label on contiguous ER cisternae, suggesting that there might be regional specialization of the ER with respect to vacuole formation (Fig. 8). Other immunocytochemical assays demonstrated labeling of small vesicles and various sizes of provacuoles and vacuoles (Fig. 9). The smallest vesicles (0.1-0.2 /tm diameter) appeared to be similar in size and morphology to the putative provacuoles that were associated with the Golgi (Fig. 6). The relative labeling density was low, probably because the Mab 2E7 recognizes a single epitope that might not always be correctly oriented. Additional gold label was seen inside the vacuoles (Fig. 9). This could be due to VATPases on membranes that were internalized during fusion of the provacuoles and subsequently degraded (Fig. 7). This is consistent with the proposed mechanism of vacuole enlargement by fusion of provacuoles and internalization/degradarion of excess tonoplast as seen in reformation of coty-

Figure 9. Immunogold localization of the peripheral subunit B of the V-ATPase on small vacuoles (V) and provacuoles (Pv). Note gold particles labeling a membranous structure within the vacuole (arrow).

1321

ledon vacuoles from protein bodies (Melroy and Herman, 1991). The immunocytochemical evidence for subunit B on ER, endomembrane compartments, and provacuoles is consistent with the results obtained from biochemical fractionation (Figs. 1-3). DISCUSSION

To understand the relationship between the ontogeny of plant vacuoles and the origin of the V-ATPase, we have used a specific Mab, 2E7, to determine the subcellular location of V-ATPases in root tip cells. Previous ultrasrrucrural studies have indicated that vacuoles are ultimately derived from the ER. According to one model, small vacuoles are thought to originate from a tubular membrane network at the trans face of the Golgi system called GERL (Marty et al., 1980). Although the term "GERL" is no longer in use, enclosure of the cytoplasm by this membrane network, followed by lateral expansion and fusion of the membrane tubules, seals off the enclosed cytoplasm from the cell cytoplasm. The sequestered cytoplasm is then digested by autophagy to form a small vacuole. The subsequent fusion of many small vacuoles results in the formation of a large central vacuole. Another model suggests that smooth ER tubules form a circle at the site to be occupied by the tonoplast (Amelunxen and Heinze, 1984; Hilling and Amelunxen, 1985). The tubules fuse into relatively flat sacks that dilate to form small vacuoles. Continued fusion of small vacuoles with nearby ER tubules and subsequent dilation results in the formation of large vacuoles. The cytoplasm enclosed by the smooth ER ring is gradually displaced and excess membrane can be contained within the vacuole. According to this model, the entire smooth ER becomes the tonoplast. Our immunological results (Figs. 69) are in general consistent with either of these two models; however, the distinction between smooth ER and the GERL network cannot be made. V-ATPases are associated with purified mature vacuoles (Smith et al., 1984; Sze, 1985); however, it is not clear where this pump is synthesized and assembled. There are several possibilities. V-ATPases could be assembled at the mature vacuole, provacuoles, trans Golgi-derived vesicles, or even at the ER. Using both cell fractionation and immunolocalization techniques, we have shown that V-ATPases are associated with the ER in immature cells of the root tip (Figs. 2 and 8). In these cells, higher levels of V-ATPase subunits were associated with the ER fraction and Golgi-derived vesicles than with the mature vacuoles (Fig. 2). Conversely, in mature differentiated cells, the V-ATPases were mostly located on the mature vacuolar membrane (Fig. IB). These results support the idea that vacuoles are initially derived from the ER membrane (Ameluxen and Heinze, 1984; Hortensteiner et al., 1992) and that the V-ATPase originated at the ER. Conventional EM suggests that vacuole precursors (i.e. provacuoles) could originate from the trans Golgi network (Fig. 6). Much of the V-ATPases were associated with small vesicles (0.1-0.3 pm) that may be derived from the Golgi (Fig. 9). Immunogold labeling on the Golgi dictyosomes was not observed in oat roots. However, the same 2E7 antibody did label the Golgi from sycamore suspension-cultured cells that were prepared by the rapid freeze/freeze substitution

1322

Herman et ai.

method (G.F. Zhang and A. Staehelin, personal communication). The V-ATPase was recently shown to be associated with purified Golgi membranes from maize roots (Oberbeck et al., 1994). The Golgi apparatus has been shown to mediate the deposition of soluble proteins in the vacuole; however, there is far less information on the role of Golgi in mediating the traffic of tonoplast proteins. The Golgi trafficking inhibitors brefeldin A or monensin impede the progression of newly synthesized soluble proteins to the vacuole, but not the vacuolar membrane-bound proteins (Gomez and Chrispeels, 1993). Thus, soluble and membrane-bound proteins use different mechanisms of transport to reach the vacuole. The tonoplast integral protein aTIP has been localized in the Golgi (Melroy and Herman, 1991), suggesting that some tonoplast proteins do progress through the Golgi. Taken together, these results support the idea that Golgi membranes possess vacuolar-type proton pumps. Some of the pumps may be in transit to the vacuole, whereas others may be resident pumps that acidify Golgi compartments. Mature vacuoles are formed as a result of fusion of many small vacuoles; however, the events leading to the recognition, docking, and subsequent fusion are not understood. As two vesicles fuse, the membranes that divide the two compartments appear to be intemalized and degraded to form one compartment (Fig. 7). This intemalization and degradation may be necessary for the reduction of membrane surface area to volume ratio as the vacuole enlarges. We suggest that V-ATPase subunits sometimes observed inside an enlarging vacuole could result from the fusion of two or more vesicles (Fig. 9). This interpretation is consistent with the results with negatively stained vesicles in which the peripheral knobs of V-ATPases of one vesicle are observed facing the knobs of a neighboring vesicle (Ward and Sze, 1992a). This could explain the inaccessibility of approximately 50% of V-ATPase peripheral subunits to KI washing of native membranes (Rea et al., 1987). Because the fusion of biological membranes can be facilitated by Ca2+,an acidic vesicle interior, ion movement, and osmotic swelling due to water uptake (Wilshut and Hoekstra, 1984; Lucy and Ahkong, 1986), these results imply that endomembrane V-ATPases could play a role in membrane fusion and therefore vacuole ontogeny. This idea is supported by the observation that small GTP-binding proteins (probably involved in vesicle trafficking and fusion) associate with pancreatic microsomes when the intravesicular pH is acidified by a V-ATPase (Zeuzem et al., 1992). Interestingly, low levels of V-ATPase subunits were consistently detected on plasma membrane fractions separated with a Suc gradient. PM-associated V-ATPase was relatively more prominent in root tip cells (Figs. 2 and 3B) and could be enhanced in roots grown in alkaline medium (X. Li, data not shown). Although we do not understand the role of a PM-associated V-ATPase yet, it is possible that V-ATPases are associated with secretory vesicles of immature cells, which actively secrete new wall materials (such as pectins and hemicelluloses) as well as the mucilagenous layer that protects the root tip (Fig. 5). We did not, however, detect VATPases on the PM by immunogold labeling, perhaps because the V-ATPases appeared on the PM transiently and therefore the antigen levels were relatively low. Very little is known about the biosynthesis and assembly


of V-ATPases in plants or other eukaryotes. According to the model derived from studies in yeast (Kane and Stevens, 1992), the V, complex is synthesized and assembled at the rough ER. We have some evidence supporting this idea. When the cDNA encoding an oat 16-kD subunit is transcribed and translated in vitro, the major integral protein is inserted stably into microsomal membranes (Szi? et al., 1992b). In yeast, evidence shows that the V1 is syrthesized and assembled in the cytosol; however, it is unclear whether the VI is attached to the V, at the Golgi or the vacuole (Kane and Stevens, 1992). Our results showing subunit B on the ER (Figs. 2 and 8) would strongly indicate that the peripheral subunit B or the V1 complex is attached to the V, 0'1the ER from oat roots. Thus, the fully assembled V-ATPase on the ER could (a) directly become part of the vacuole membrane during vacuole formation or (b) be sorted via the Golgi to specific endomembrane destinations or (c) both. This model for the assembly of VI with V, at the ER is supported by recent experiments in both plants and animals using different approaches. Purified membrane fra ztions of the ER, Golgi apparatus, and clathrin-coated vesicles from maize roots reacted positively with antibodies ta the VATPase on western blots. Because both peripheral (70, 57, and 44 kD) and integral (16 kD) subunits were present, these results suggest that the V-ATPase was already asseinbled at the ER (Oberbeck et al., 1994). In pulse-chase experiments of bovine kidney epithelial cells, Myers and Forgac (1393) bovine showed that the VI complex was associated wi'h membranes at 15OC or in the presence of brefeldin A. Since brefeldin A disrupts the Golgi and perturbs ER-Golgi traffic and 15OC temperatures stop traffic between the ER arid Golgi, their results would suggest that the VI is assembled with the V, in the ER. These results raise several significant questions: (a) Is the ER-bound V-ATPase functionally active? According to membrane fractionation studies, the V-ATPase associaíed with smooth ER and low-density endomembranes in oat or com seedlings is a functional proton pump. Membranes equilibrating between 24 and 28% Suc show both vanadatesensitive and nitrate-sensitive ATPase activity (ChL rchill et al., 1983) and are active in proton pumping (DuPont et al., 1982; Hager and Biber, 1984) or in ApH-driven Ca transport (Schumaker and Sze, 1985). (b) Are all the endomvmbrane V-ATPases destined for the mature vacuole or are they sorted to specific destinations? We do not know yet, but we suspect that there is active sorting to maintain compartment diversity. Endomembranes are dynamic and can participate in diverse functions, like cell plate formation (fusion of Golgi and ER vesicles), provacuole fusion to form central vacuoles, and secretion of wall components during cell wall sjgnthesis. Because of the diverse roles attributed to endomembranes, it is very likely that compartments are functionally distinct. If so, there might be organelle- or compartment-specific isoforms of V-ATPases that could differ in their subunit composition and be independently regulated (Gogarten et al., 1992). A provocative indication of subunit diversity is observed among subunits of 36 to 38 kD between membranes of densities corresponding to 20 to 23% and 28 to 32% Suc (Fig. 3A). In summary, the plant V-ATPases are located not only on

Vacuolar H+-ATPaseAssociated with the ER and Provacuoles mature vacuoles but also on many endomembrane compartments of rapidly proliferating, differentiating cells. In oat root tips, V-ATPases are associated with the ER, Golgi-derived vesicles, provacuoles, tonoplast, and possibly secretory vesicles and the PM. These results suggest that the V-ATPase complex originates at the ER and is then sorted to other endomembranes. Therefore, the V-ATPase is not a reliable marker for the vacuole in immature cells. However, the VATPase appears to be a better marker of the tonoplast in mature cells, where there is little ER and Golgi, and the VATPase is associated mainly with the mature vacuolar membrane (Smith et al., 1984). The broad distibution of VATPases in plant cells, as in other eukaryote cells (MarquezSterling et al., 1991; Nolta et al., 1993), suggests that this pump is involved in organelle biogenesis and in diverse functions of the endomembrane system, including synthesis, folding a n d assembly of proteins, cytosolic Ca regulation, osmoregulation, vesicle trafficking, and membrane fusion. Future studies to identify the various isoforms of V-ATPases and to immunolocalize isoform-specific V-ATPases are one approach to developing an understanding of the roles and significance of this endomembrane proton pump. ACKNOWLEDCMENTS We thank M. Chrispeels (University of Califomia, San Diego), R.T. Leonard (Universityof Califomia, Riverside), and M. Maeshima (Hokkaido University, Sapporo, Japan) for providing antibodies to BiP, PM-ATPase, and VM23, respectively. We also thank Stephen Wolniak for his advice about immunofluorescence microscopy and Todd Cooke (University of Maryland, College Park) for stimulating discussions regarding plant structure and function. Received May 26, 1994; accepted July 14, 1994. Copyright Clearance Center: 0032-0889/94/106/1313/12.

LITERATURE CITED Ali MS, Akazawa T (1986) Association of H+-translocatingATPase in Golgi membrane system from suspension cultured cells of sycamore. Plant Physiol81: 222-227 Amelunxen F, Heinze U (1984) Zur Entwicklung der vacuole in Testa-Zellen des Leinsamens. Eur J Cell Biol 35: 343-354 Anraku Y, Umemoto N, Hirata R, Ohya Y (1992) Genetical and cell biological aspects of the yeast vacuolar H+-ATPase.J Bioenerg Biomembr 2 4 395-405 Baskin TI, Busby CH, Fowke LC, Sammut M, Gubler F (1992) Improvements in immunostaining samples embedded in methacrylate: localization of microtubules and other antigens throughout developing organs in plants of diverse taxa. Planta 187: 405-413 Bennett AB, ONeill SD, Spanswick RM (1984) H+-ATPaseactivity from storage tissue of Beta vulgaris. Plant Physiol74 538-544 Berkelman T, Houtchens KA, DuPont FM (1994)Two cDNA clones encoding isoforms of the B subunit of the vacuolar ATPase from barley roots. Plant Physiol104 287-288 Boller T, Wiemken A (1986) Dynamics of vacuolar compartmentation. Annu Rev Plant Physiol37: 137-164 Bowman BJ, Vazquez-Laslop N, Bowman EH (1992) The vacuolar ATPase of Neurospora crassa. J Bioenerg Biomembr 2 4 361-370 Chanson A, Taiz L (1985) Evidence for an ATP-dependent proton pump on the Golgi of com coleoptiles. Plant Physiol78 232-240 Churchill KA, Holaway 8, Sze H (1983) Separation of two types of electrogenic H+-pumping ATPases from oat roots. Plant Physiol 7 3 921-928 Depta H, Holstein SEH, Robinson DG, Lutzelschwab M, Michalke

1323

W (1991) Membrane markers in highly purified clathrin-coated vesicles from Cucurbita hypocotyls. Planta 1 8 3 434-442 DuPont FM, Bennett AF, Spanswick RM (1982) Localization of a proton-translocating ATPase on sucrose gradients. Plant Physiol 7 0 1115-1119 DuPont FM, Morrisey PJ (1992) Subunit composition and CaATPase activity of the vacuolar ATPase from barley roots. Arch Biochem Biophys 294 341-346 Forgac M (1992) Structure and properties of the coated vesicle H+ATPase. J Bioenerg Biomembr 2 4 341-350 Gluck SL (1992) The structure and biochemistry of the vacuolar H+ATPase in proximal and dista1 urinary acidification. J Bioenerg Biomembr 2 4 351-360 Gogarten JP,Fishmann J, Braun Y, Morgan L, Styles P, Taiz SL, DeLapp K, Taiz L (1992) The use of antisense mRNA to inhibit the tonoplast H+-ATPasein carrot. Plant Cell4 851-864 Gomez L, Chrispeels MJ (1993) Tonoplast and soluble vacuolar proteins are targeted by different mechanisms. Plant Cell 5 1113-1 124 Hager A, Biber W (1984)Functional and regulatory properties of H+ pumps at the tonoplast and plasma membrane of Zea mays coleoptiles. 2 Naturforsch 39c: 927-932 Hager A, Debus G, Edel H-G, Stransky H, Serrano R (1991) Auxin induces exocytosis and the rapid synthesis of a high tumover pool of plasma membrane H+-ATPase.Planta 1 8 5 527-537 Harlow ED, Lane D (1988) Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Herman EM, Melroy DL (1990) Electron microscopic immunocytochemistry in plant molecular biology. Plant Mo1 Biol Manual 813 1-24 Hilling B, Amelunxen F (1985) On the development of the vacuole. 11. Further evidence for endoplasmic reticulum origin. Eur J Cell Biol18 195-200 Hodges TK, Leonard RT (1974) Purification of a plasma membrane bound ATPase from plant roots. Methods Enzymol32: 392-406 Hortensteiner S, Martinoia E, Amrhein N (1992) Reappearance of hydrolytic activities and tonoplast proteins in the regenerated vacuole of evacuolated protoplasts. Planta 187: 113-121 Hsu HT, Lawson RH (1985) Comparison of mouse monoclonal antibodies and polyclonal antibodles of chicken egg yolk and rabbit for assay of camation etched ring virus. Phytopathology 75: 778-783 Hurley D, Taiz L (1989) Immunological localization of the vacular H+-ATPasein maize root tip cells. Plant Physiol89 391-395 Hurtley SM, Helenius A (1989) Protein oligomerization in the endoplasmic reticulum. Annu Rev Cell Biol5 277-307 Johnson KD, Herman EM, Chrispeels MJ (1989) An abundant, highly conserved tonoplast protein in seeds. Plant Physiol 91: 1006-1013 Kalinski AJ, Melroy DL, Dwivedi RS, Herman EM (1992) A soybean vacuolar protein p34 related to thiol proteases which is synthesized as a glycoprotein precursor during seed maturation. J Biol Chem 267: 12068-12076 Kane PM (1992) Biogenesis of the yeast vacuolar H+-ATPases.J Exp Bioll72 93-103 Kane PM, Stevens TH (1992) Subunit composition, biosynthesis, and assembly of the yeast vacuolar proton translocating ATPase. J Bioenerg Biomembr 2 4 383-393 Lai S, Watson JC, Hansen JN, Sze H (1991) Molecular cloning and sequencing of cDNAs encoding the proteolipid subunit of the vacuolar H+-ATPase from a higher plant. J Biol Chem 266 16078-16084 Lucy JA, Ahkong QF (1986) An osmotic model for the fusion of biological membranes. FEBS Lett 199 1-11 Maeshima M (1992) Characterization of the major integral protein of vacuolar membrane. Plant Physiol98: 1248-1254 Mandala S, Taiz L (1985) Proton transport in isolated vacuoles from com coleoptiles. Plant Physiol78: 104-109 Marquez-Sterling N, Herman IM, Pesacreta T, Arai H, Terres G, Forgac M (1991) Immunolocalization of the vacuolar type H+ATPase from clathrin-coated vesicles. Eur J Cell Biol 5 6 19-33 Martinoia E (1992) Transport processes in vacuoles of higher plants. Bot Acta 1 0 5 232-245

1324

Herman et al.

Marty F, Branton D, Leigh RA (1980) Plant vacuoles. In NE Tolbert, ed, The Biochemistry of Plants, Vol I. Academic Press, New York, pp 625-658 Maurel C, Reizer J, Schroeder JI, Chrispeels MJ (1993) The vacuolar membrane protein ?-TIP creates water specific channels in Xenopus oocytes. EMBO J 12: 2241-2247 Melroy D, Herman EM (1991) TIP, an integral membrane protein of the soybean seed storage vacuole, undergoes developmentally regulated membrane insertion and removal. Planta 184: 113-122 Myers M, Forgac M (1993) Assembly of the peripheral domain of the bovine vacuolar H+-adenosine triphosphatase. J Cell Physiol 156 35-42 Nelson N (1992) Structural conservation and functional diversity of V-ATPases. J Bioenerg Biomembr 2 4 407-414 Nolta KV, Padh H, Steck TL (1993) An immunological analysis of the vacuolar proton pump in Dictyostelium discoideum. J Cell Sci 105 849-859 Oberbeck K, Drucker M, Robinson DG (1994) V-type ATPase and pyrophosphatase in endomembranes of maize roots. J Exp Bot 4 5 235-244 Parry RV, Turner JC, Rea PA (1989) High punty preparations of higher plant vacuolar H+-ATPasereveal additional subunits. J Biol Chem 264 20025-20032 Rea PA, Griffith CJ, Manolson MF, Sanders D (1987) Irreversible inhibition of H+-ATPase of higher plant tonoplast by chaotropic anions. Biochim Biophys Acta 904 1-12 Schumaker KS, Sze H (1985) A CaZ+/Hfantiport system driven by the proton electrochemicalgradient of a tonoplast H+-ATPasefrom oat roots. Plant Physiol79 1111-1117 Smith JA, Uribe EG, Ball E, Heuer S, Luttge U (1984) Character-


ization of the vacuolar ATPase activity of Crassulxean acid metabolism plant Kalanchoe daigremontiana. Eur J Biochem 141: 415-420 Starke T, Kinkilla TP, Gogarten JP (1991) Two separate genes encode the catalytic 70 kD V-ATPase subunit in Psilotum and Equisetum. Z Naturforsch 46c: 613-620 Sze H (1985) H+-translocating ATPases: advances using membrane vesicles. Annu Rev Plant Physiol36: 175-208 Sze H, Ward JM, Lai S (1992a) Vacuolar H+-ATPases fmm plants: structure, function and isoforms. J Bioenerg Bioniembr 2 4 371-381 Sze H, Ward JM, Laí S, Perera I (1992b) Vacuolar-type €I+-translocating ATPases in plant endomembranes: subunit organization and multigene families. J Exp Biol 172 123-135 Ward JM (1991) Vacuolar H+-ATPase from oat roots: p.arification, reconstitution, dissociation and reassembly. PhD dissertation. University of Maryland, College Park, MD Ward JM, Reinders A, Hsu H-T, Sze H (1992) Dissociation and reassembly of the vacuolar H+-ATPase complex from oat roots. Plant Physiol99 161-169 Ward JM, Sze H (1992a) Subunit composition and organization of the vacuolar H+-ATPasefrom oat roots. Plant Physiol9'k 170-179 Ward JM, Sze H (1992b) Proton transport activity of the purified vacuolar 13'-ATPase from oats. Plant Physiol 9 9 925-5'31 Wilschut J, Hoekstra D (1984) Membrane fusion: from liposomes to biological membranes. Trends Biochem Sci 9 479-483 Zeuzem S, Zimmerman P, Schulz I(1992) Association of a 19- and a 21-kDa GTP-binding protein to pancreatic microsomal vesicles is regulated by the intravesicular pH established by i i vacuolartype H+-ATPase.J Membr Biol 125: 231-241