Characterization of the peribacteroid membrane ATPase of lupin root nodules

ARCHIVES

OF BIOCHEMISTRY

AND BIOPHYSICS

Vol. 264, No. 2, August 1, pp. 564-573,1938

Characterization

of the Peribacteroid Membrane of Lupin Root Nodules’

ATPase

NEIL M. DOMIGAN, KEVIN J. F. FARNDEN: JOHN G. ROBERTSON,* AND BRIAN C. MONK Department of Biochemistry, University of Otago, Dunedin, New Zealand, and *Applied Biochemistry Division, DSIR, Palmerston North, New Zealand Received December 8,1987, and in revised form March 24,1938

Peribacteroid membranes can be isolated in essentially pure form from 20-day lupin root nodules by osmotic shock of the purified membrane enclosed bacteroids. The ATPase (EC 3.6.1.3) associated with this membrane has an acid pH optimum (5.25) and is specific for ATP (Mg-ATP Km = 0.16 mM). The enzyme activity requires magnesium or manganese ions, is slightly stimulated by the cations potassium and rubidium, and is inhibited by vanadate, diethylstilbestrol, N,N’-dicyclohexylcarbodiimide, fluoride, molybdate, and calcium. Molybdate and fluoride sensitivity do not in this case indicate the presence of significant nonspecific phosphatase activity. The ATPase is not inhibited by oligomycin, azide, or the soluble carbodiimide 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide. In some respects the lupin peribacteroid membrane ATPase appears to differ from the plasma membrane ATPase of other plants. o 1988 Academic press. I~~.

The legume-Rhizobium symbiosis requires genetic, metabolic, and structural complementarity between the infected root cell and the intracellular nitrogenfixing bacteroid. A feature of the symbiosis is the shielding of aspects of the operation of the bacteroid from the plant cytoplasm. The structure which provides the necessary compartmentation is the peribacteroid membrane. This membrane places the intracellular bacteroid in the equivalent of a suitably modified extracellular environment (1). Incompatibility between the host and the microbial symbiont, or mutation in either partner, may affect the proper development and persistence of the peribacteroid membrane and render the root nodules ineffective (2,3). The peribacteroid membrane appears to derive from the root cell plasma mem’ Supported by a DSIR research contract. B.C.M. is a Senior Research Fellow of the Medical Research Council of New Zealand. ’ To whom correspondence should be addressed. 0003-9861/f&3$3.00 Copyright Q 1988 by Academic Press, Inc. All rights of reproduction in any form reserved.

brane that encapsulates the bacterium as it is invaginated into the target root cell (1, 4). After bacterial penetration the peribacteroid membrane is thought to be synthesized by the root cell at rates consonant with intracellular rhizobial division and with membrane turnover. It is possible to purify peribacteroid membranes from lupin root nodules (5). However the study of the composition, structure, genetic specification, and function of the peribacteroid membrane is at a preliminary stage. Robertson et al. (5) used electron microscopic techniques to characterize the lupin peribacteroid membrane as an asymmetric membrane with a high lipid to protein ratio and noted the presence of an ATPase activity in the preparation. Verma and colleagues used a histochemical approach to demonstrate the association of an ATPase activity with the soybean peribacteroid membrane (3) and more recently presented biochemical evidence to suggest that this ATPase may 564

PERIBACTEROID

MEMBRANE

ATPase OF LUPIN ROOT NODULES

be related to the proton pumping ATPase of the root plasma membrane (6). In this report we describe the biochemical characterization of the lupin peribacteroid membrane-associated ATPase. The general properties of the peribacteroid membrane ATPase are similar but not identical to plant root plasma membrane and fungal plasma membrane proton pumping ATPases (7,8). This is consistent with the view that the peribacteroid membrane may be related to the plant plasma membrane in some limited aspects of form and function, but that the peribacteroid membrane is also a highly differentiated structure in its own right. MATERIALS

AND METHODS

Plant Material Surface sterilized seeds (Lupinus angustifolius L., variety Uniwhite), were germinated at 25°C for 2 days in the dark and then transferred to trays of sterile pumice. Each seed was inoculated with 1.0 ml of Rhizdxium lupini suspension and allowed to grow for about 20 days under controlled conditions, as previously described (9), prior to the harvesting of root nodules.

Preparaticm of the Peribacteroid Membrane Fraction The peribacteroid membrane fraction was prepared as previously described by Robertson et al (5), with minor modifications. In brief, 8-10 g of 19- to 21-day postinoculation root nodules were picked into, washed with, and resuspended in crush medium A at 0-4’C which contained 16% sucrose (w/v), 5 mM DTE: and 50 mM Tris-acetic acid, pH 7.2, at a ratio of 4 ml of medium per gram of tissue. All subsequent steps, unless otherwise specified, were carried out at 0-4’C with media buffered with Tris-acetic acid to pH 7.2. All sucrose concentrations are given as percentage (w/v) and solutions were corrected to the indicated concentration using a refractometer. The

aAbbreviations used: Chaps, 3-[(a-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; DCCD, N,N’-dicyclohexylcarbodiimide; DES, diethylstilbestrol; DTE, dithioerythritol; EDAC, l-ethyl3-(3-dimethylaminopropyl)carbodiimide; EGTA, ethylene glycol bis(@-aminoethyl ether) N,N’-tetraacetic acid, PAGE, polyacrylamide gel electrophoresis; PBM, peribacteroid membrane; SDS, sodium dodecyl sulfate.

565

nodules were ground using a mortar and pestle and the brei was filtered through a single layer of Miracloth (Chicopee Mills Inc, New York). The filtrate was centrifuged at 2509 for 5 min to remove cellular debris and the supernatant was centrifuged at 8000~ for 5 min in a Sorvall SS-34 rotor. The pelleted material, which included the bacteroids, was resuspended in crush medium A and layered over a 10 ml 55% sucrose pad buffered with 56 rnrdTris. After centrifugation at 8000~for 20 min in a Sorvall HS-4 swinging bucket rotor the bacteroids were recovered from the 16/55% sucrose interface and adjusted in a final volume of 10 ml to 30% sucrose (in 50 m?dTris buffer). This mixture was layered over 10 ml of 35% sucrose in 50 mM Tris and centrifuged at 16,OOOg for 15 min in a Beckman SW-27 rotor. The pellet of purified bacteroids was osmotically shocked by resuspension in 10 ml of 10 rnrd Tris buffer and layered on a 10 ml 35% sucrose pad buffered with 10 rnM Tris. The bulk of the osmotically shocked bacteroids were pelleted by centrifuging at 16,OOOgfor 15 min in the SW-27 rotor. The peribacteroid membrane fraction was recovered at the interface and the supernatant above the interface, adjusted in a final volume of 20 ml to 8% sucrose in 10 mM Tris buffer, layered over 12 ml of 12% sucrose in 10 mM Tris, and centrifuged at 60,000~for 30 min in the SW-27 rotor. The pellet was resuspended in 10 ml 47% sucrose and overlaid with 3 ml 45% sucrose followed by a linear gradient of 45-20% sucrose, all including 10 mM Tris buffer. This gradient was centrifuged to equilibrium (90,OOOgfor 15 h) and the peribacteroid membrane fraction, which could be readily visualized as a discrete band centering on 33% sucrose, recovered and diluted with 20 ml 10 mM Tris and pelleted by centrifugation at 75,OOOg for 30 min (Beckman Ti 50.2 rotor). The pellet of peribacteroid membranes was resuspended in 4% sucrose in 10 mM Tris-acetate, pH 7.2, at a protein concentration of about 0.75 mg/ml and stored under liquid nitrogen until required for analysis.

A TPase Assags Method A. ATP (Tris salt, vanadate free, Sigma) hydrolyzing activity was determined by detection of released phosphate using a minor modification of a single-step procedure that measures the formation of the phosphomolybdate complex (10). The reaction mix contained in 310 pl up to 10 ng of protein, 5 mM ATP, 5 mM MgCla unless otherwise specified and 100 mM Tris-acetate buffer at the desired pH (usually pH 5.25). After incubating for the appropriate period at 27”C, 1 ml of the ice-cold molybdate reagent was added to 300 pl of the reaction mix and the absorbance at 350 nm was determined 30 s later. All reactions were carried out in duplicate and reaction blanks that were terminated at t = 0 were used to subtract background levels of free phosphate. Inhibi-

566

DOMIGAN ET AL.

tors were dissolved in the assay mix at the indicated concentrations and appropriate levels of ethanol were included in blanks where necessary. Method B. A l-al aliquot of [cY-~*P]ATP (3000 Ci/ mmol) was added to the standard reaction mix described in method A. After incubation at 27’C for an appropriate period, l-r1 aliquots of the reaction mixture were spotted onto PEI thin-layer chromatography paper (Polygram CEL 300). Solutions of 20 mM ATP, ADP, and AMP were prepared in 0.5 mM NaOH and l-g1 aliquots of these solutions were spotted on the chromatography paper to act as standards. After the spots dried, the chromatogram was developed with 0.75 M KHxPO( buffered to pH 3.5 with HCI. The standards were located by their uv absorbance while the location of the radiolabeled nucleotides was determined by autoradiography using Cronex X-ray film. Each labeled nucleotide was excised, placed in a scintillation vial, 1 ml Ha0 added, and quantitated by measurement of Cerenkov emissions detected in a Philips PW 4700 liquid scintillation counter.

FIFO-ATPase (data not shown). Bacteroid envelope inner membranes form a broad band from 30 to 42% sucrose in similar gradients (5). The response of the peribacteroid membrane ATPase to NaF and Na azide at other pHs is presented later in Fig. 5. This peribacteroid membrane fraction gives a characteristic polypeptide profile on SDS-PAGE (to be published elsewhere), which differentiates this membrane from lupin root plasma membrane and the inner envelopes from the bacteroids of free-living rhizobia. General Properties of the Peribacteruid Membrane ATPase

Reaction rates. The specific activity of the peribacteroid membrane ATPase measured at pH 5.25 and 27°C in the presence of 5 mM ATP and 5 InM MgClz ranged, Protein Determination in eight preparations, between 30 and 70 Protein was determined by the RNA coprecipitapmol Pi releasedjmg protein/h, with an tion method of Polachek and Cabib (11) with bovine average of 45 pmol Pi released/mg proserum albumin used as a standard. tein/h. This activity remained stable for several months, provided the peribacteRESULTS roid membrane preparations were stored under liquid nitrogen. ATPase activity Isolation of a Peribacteroid Membranewas measured by two independent methAssociated ATPase ods. The rate of ATP hydrolysis in assay Peribacteroid membranes were isolated system A, which measures the rate of Pi from membrane-enclosed bacteroids using release, was linear and this rate remained the method of Robertson et al. ((5) and see constant for at least 40 min. The system Materials and Methods). In the final gra- gives a proportionate increase in the dient of this procedure, the peribacteroid amount of ATP hydrolyzed in response to membranes band between 28 and 35% su- added peribacteroid membrane up to at crose (5). ATPase activity, assayed at pH least 5 pg per assay. In the alternative ra5.25 (Fig. l), was found in such a gradient dioisotope assay system (system B, Fig. 2), in a peak banding between 28 and 38% which measures the rate of [w~‘P]ADP sucrose, with most of the activity recorded formation from [w~~P]ATP, a linear rate between 32 and 35% sucrose. This activity (27 pmol/h/mg protein) was observed was at least 50% inhibited by NaF (5 mM) throughout a 40 min time course, although and 75% inhibited by vanadate (200 PM) a slight decline in rate was observed by 60 (Fig. 1). min. The two assay systems gave compaPeribacteroid membranes were recov- rable rates of ATP hydrolysis during the ered from preparative sucrose density 40-min assay period. In addition, analysis gradients from the fractions that banded of the radioactive degradation products of between 28 and 36% sucrose. Fractions [cx-~~P]ATP demonstrated that there was pooled from this region were essentially negligible formation of AMP compared insensitive to Na azide (Fig. l), an inhibiwith ADP and that the rate of ADP fortor of the bacterial inner membrane en- mation closely correlated with the loss of ergy transducing ATPase (12) or oligomy- ATP (Fig. 2). This result shows that neitin, an inhibitor of the mitochondrial ther nonspecific phosphatase nor apyrase

PERIBACTEROID

0

5

MEMBRANE

10

ATPase

15 Fractkm

OF LUPIN

20

ROOT NODULES

25

567

30

number

FIG. 1. Peribacteroid membrane-associated ATPase activity. The crude peribacteroid membrane fraction, separated from osmotically shocked bacteroids by centrifuging over a 35% sucrose pad, was subjected to flotation density gradient centrifugation to equilibrium as described under Materials and Methods. The fractionation system comprised the crude peribacteroid membrane fraction resuspended in 10 ml of 47% sucrose, overlaid with 3 ml of 45% sucrose and 22 ml of a 45-20% sucrose gradient. The gradient was fractioned into l-ml fractions and lOO-~1 aliquots were assayed for protein content and ATPase activity in the absence of inhibitors (w) or in the presence of the following inhibitors: 1 mM Na azide (Cl), 5 mM NaF (0), or 200 PM Na vanadate (0). The percentage sucrose density (-) of each fraction was determined using a refractometer.

activity contribute significantly to the rates of ATP hydrolysis and Pi liberation observed in the two assay systems. Magnesium dependence and substrate specijhity. At constant peribacteroid membrane concentration in assay system A and in the presence of 5 mru MgClz, the ATPase activity responded to added ATP in a typical Michaelis-Menten fashion, giving a V,,, of about 70 pmol Pi released/mg protein/h and half V,,, at 1 mM added ATP (Fig. 3). The requirement of the enzyme for magnesium ions was tested as follows. In the presence of saturating levels of ATP (5 mM) but in the absence of magnesium ions, the ATPase was essentially inactive (4 pmol Pi liberated/ mg protein/h). Addition of 5 mM Mg2+ to the assay gave a saturable 15-fold stimulation of ATPase activity (63 kmol Pi liberated/m@; protein/h), with half-maximal stimulation observed at 0.3 mM Mgz+.

The nucleotide specificity of the ATPase is shown in Fig. 4. ATP was clearly the preferred substrate, while the other adenine nucleotides, ADP and AMP, were hydrolyzed eight- and ninefold slower, respectively. This result confirms the observations reported with assay system B in Fig. 2. The nucleotide triphosphates CTP, UTP, and GTP were even more poorly hydrolyzed by the enzyme and were not recognized as natural substrates under these assay conditions. High levels of ADP appeared to competitively inhibit ATPase activity with 50% inhibition occurring at about 10 mM ADP (inset, Fig. 4). The phosphatase substrate p-nitrophenylphosphate was very slowly hydrolyzed, indicating that the peribacteroid membrane isolated exhibits little if any phosphatase activity under these assay conditions. Glucose 6-phosphate was not hydrolyzed at a significant rate either.

568

DOMIGAN ET AL.

Lorrof

ATP

Productionof

Production

ADP

of AMP I

0

10

. 30

20

Time

. 40

I 50

60

(mh)

FIG. 2. Time course and identification of nucleotides produced by the peribacteroid membrane ATPase. Peribacteroid membrane ATPase activity was measured in assay system B by inclusion of 5 pCi of [(Y-q]ATP in a standard assay mixture containing 7.5 pg peribacteroid membrane protein, 5 mM ATP, 5 mM MgCls in 100 mM Tris-acetate buffer, pH 5.25. The percentage distribution of each nucleotide recovered (ATP, ADP, and AMP) from l-p1 aliquots, determined by chromatographic separation on PEI paper, was used to calculate the total amount of each nucleotide produced or depleted, after taking into account the endogenous levels of labeled ADP and AMP at t = 0.

pH optimum. The peribacteroid membrane ATPase has an acidic pH optimum, with a peak of NaF-sensitive and Na aside-insensitive activity at pH 5.25 (Fig. 5). Significant ATPase activity was observed only in the narrow range of pH 5-6.5 with a slight increase in the proportion of azide sensitivity observed at pH 6.5,

Substrate

FIG. 4. Substrate specificity of the peribacteroid membrane ATPase. The indicated substrates at 5 rnM were incubated for 30 min in assay system A with 5 pg peribacteroid membrane protein, 5 mM MgCla, and 100 mM Tris-acetate, pH 5.25. The specific activities are presented from duplicate determinations from a typical experiment. The inset details the inhibitory effect of the indicated concentration of ADP on peribacteroid membrane ATPase activity assayed in the same way in the presence of 5 mM ATP.

in comparison with values at pH 6 and below. The bacterial inner membrane ATPase is known to exhibit a broad pH optimum between pH 6 and pH 10 (12). The slight increase in aside sensitivity at pH 6.5 might be interpreted as low-level contamination of the peribacteroid mem-

Total

b

ATPase

Azide Insensitive ATPsse

2 3 f

, 25

OW 4

3

5

6

7

NaF ~nsenritive ATPare

6

9

16

PH [

ii, 0

2

4

6

6

10

ATP(mM)

FIG. 3. ATP concentration dependence of the peribacteroid membrane ATPase. Assays were conducted using assay system A which included 5 mM MgCl* in 100 mM Tris-acetate buffer, pH 5.25.

FIG. 5. pH dependence of the peribacteroid membrane ATPase. The profiles were generated using 5-gg aliquots of peribacteroid membrane protein in assay system A in the presence of 5 mre ATP. 5 rnM MgClx, and 100 mM Tris-acetate buffer at the indicated pH. Where indicated NaF and Na aside were included at 10 and 1 mM, respectively.

PERIBACTEROID

MEMBRANE

ATPase

brane preparation by the bacterial inner envelope. However the lack of detectable ATPase activity above pH 7 is contraindicative of this possibility. In addition the ATPase activity at pH 6.5 was almost fully sensitive to NaF.

OF LUPIN

membrane proton pumping ATPases (7), was the only anion tested that significantly (20%) inhibited Mg-ATPase activity. The ionophores gramicidin S and monensin had no stimulatory effect on the ATPase activity over the pH range

The Efects of Ions cmATPase Activity Table I illustrates the effects of a wide variety of anions and cations on the ATPase activity of the peribacteroid membrane assayed at pH 5.25. Of the divalent cations tested (Table IA) only magnesium and manganese stimulated ATPase activity, with 5 mM manganese providing 50% greater stimulation than the same concentration of magnesium. The ability of manganese to stimulate the system was independent of the presence of magnesium. Cobalt, calcium, and zinc ions failed to stimulate ATP hydrolysis and 5 mM of these divalent cations strongly inhibited ATPase activity in the presence of magnesium. The peribacteroid membrane ATPase was inhibited by added Ca2+, using assay system A (see Materials and Methods), over a concentration range of 0.1-1.25 mM Ca2+ with 50% inhibition at 0.875 mM added Ca2+ (0.312 mM free Ca2’, 0.442 mM Ca-ATP). The addition of 10 InM EGTA to the assay mixture in the absence of any added Ca2+gave a 15% stimulation of the Mg’+-dependent ATPase activity and significantly reduced the inhibitory action of any added Ca2+. The monovalent cations rubidium, potassium, sodium, and cesium at 50 InM provided slight (10-20s) stimulation of Me-ATPase activity (Table IB). Lithium and ammonium ions had no effect on the ATPase activity. The slightly stimulatory effect of KC1 is clearly dependent on the potassium ion as the other potassium halides (KBr and KI) gave identical stimulation of Mg-ATPase (Table IC). Potassium sulfate and potassium nitrate gave a similar stimulatory effect, and indicated that the peribacteroid membrane ATPase was insensitive to the presence of 50 mM nitrate. Nitrate is regarded as a specific inhibitor of the tonoplast ATPase (7). Thiocyanate (50 InM), an inhibitor of plasma

569

ROOT NODULES

TABLE

I

EFFECTS OF IONS ON PERIBACTEROID MEMBRANE ATPase ACTIVITY~ A.

Effects of divalent cations on peribacteroid membrane ATPase activity ATPase activity (pmol Pi released/h/mg

Species added (5 mM)

MgClz MnCla CoClz CaClc ZnClz B.

No magnesium added

0.6 + 0.1 49 +4

32k 2 53 2 5

7

10 + 2

6+2 7f2

+1

6.9 f 0.9 7.2 f 0.1

Effects of monovalent cations on peribacteroid membrane ATPase activity Species added (50 mM) None KC1 RbCl NaCl CsCl LiCl NH&l

C.

+5 mM magnesium

protein)

ATPase activity (pm01 Pi released/ h/mg protein)

32 +l 37 +1 38k 2 37 +2 35 +2 31 + 2

32 f 2

Effects of anions on peribacteroid ATPase activity Species added (50 mM) None KC1 KBr KI KNOa KzSO4

KSCN

membrane

ATPase activity (pm01 Pi released/ h/mg protein)

32 f2 38.2 * 0.4 38.0 + 0.6 37 +2 38.0 + 0.9 38 +3 26 +4

“All assays were conducted in assay system A at pH 5.25 in the presence 5 mbf MgClz unless otherwise indicated.

570

DOMIGAN

5.25-7.0 (data not shown). Ammonium chloride (10 InM) also failed to modify MgATPase activity over the same pH range (data not shown). Inhibitors of Peribacteroid MembraneAssociated ATPase Activity Studies employing radioactive ATP in assay system B with ATP at 0.05 mM and 15 pg PBM (Fig. 6A) showed that the peribacteroid membrane Mg-ATPase was sensitive to vanadate (1 mM), molybdate (100 PM), and NaF (5 mM) but was relatively

A

loo-

I

NaF (1mM)

(5mM)

Mel (4mm)

ET AL.

insensitive to sodium azide (1 mM). These results were confirmed and expanded using the nonisotopic A assay system. Figure 6B shows that NaF at 10 mM gave more than 90% inhibition of the ATPase, with 50% inhibition occurring at about 1 mM NaF. Vanadate and molybdate inhibited the ATPase activity at even lower concentrations (Figs. 6C and 6D). Vanadate gave 50% inhibition of the Mg-ATPase at 50 PM while molybdate gave a comparable level of inhibition at 400 PM. Assays at pH 6.25 showed that the activity of the ATPase was reduced by about 40%.

B 3> 60 K

Azlde (W)

h i 5

30

r” f i [5

15

450 L 0

2

4

6

6

10

NaF(mM)

C 5

60.

1

':[

a

k 0.50

0.25 Vsnedste

0.75 (mM)

1.0

iq

, 0.25

, 0.50 Molybdate (mM)

,

,

0.15

1.0

FIG. 6. Inhibitors of the peribacteroid membrane ATPase. (A) Peribacteroid membranes (15 pg) were added to assay system B containing 1 &i [&?PjATP in 0.05 mM ATP and incubated for 20 min in the presence of the indicated inhibitors. The proportion of the total radiolabel recovered in each nucleotide was estimated after chromatography of l-p1 aliquots of the reaction mixture on PEI paper. (B) NaF inhibition of peribacteroid membrane ATPase activity. Peribacteroid membranes were assayed in the presence of the indicated concentrations of NaF in assay system A at pH 5.25. (C) Vanadate inhibition of peribacteroid membrane ATPase activity. Peribacteroid membranes were assayed in the presence of the indicated concentrations of Na vanadate in assay system A at pH 5.25. (D) Molybdate inhibition of peribacteroid membrane ATPase activity. Peribacteroid membranes were assayed in the presence of the indicated concentrations of Na molybdate in assay system A at pH 5.25.

PERIBACTEROID

MEMBRANE

ATPase

Molybdate (1 mM) or vanadate (5 mM) inhibited the ATPase a further 75% at this pH. The lipophilic carbodiimide DCCD (at 10 and 50 mM) had no effect on the peribacteroid membrane ATPase when added in the presence of Me. If, however, peribacteroid membranes were incubated with 50 PM DCCD in the absence of Me for 30 min prior to Mg-ATPase assay, the enzyme was inhibited by at least 30%. In contrast, the water-soluble carbodiimide EDAC (5 mM) had no inhibitory effect on the ATPase under similar conditions. These results suggest that the DCCD binding site, by analogy with the FIFO-ATPase, may involve a buried carbonyl group in a proton translocating channel (13). Me is thought to interfere with the interaction between DCCD and this type of site in proton translocating ATPases. The hydrophilic EDAC molecule would not have access to such a hydrophobic site. Diethylstilbestrol, an inhibitor of plasma membrane proton translocating ATPases (14), significantly inhibited the peribacteroid membrane Mg-ATPase at low concentrations, with 50% inhibition observed at ‘75 PM DES (data not shown).

OF LUPIN

571

ROOT NODULES

the peribacteroid membrane in an essentially noncompetitive manner, indicating binding of vanadate at a site other than the active site of the enzyme. Identical results were obtained when Mn-ATP was used as substrate instead of Mg-ATP. The Mg-ATP complex is in fact believed to be the substrate for plant plasma membrane ATPases (8) and if this is the case with the peribacteroid membrane enzyme, the data of Fig. 7 can be replotted using Mg-ATP concentrations calculated by the procedure of Robertson et al (12) to give a K, of Mg-ATP of 0.16 mM. Molybdate similarly acted as a noncompetitive inhibitor of the peribacteroid membrane-associated ATPase. DISCUSSION

The location of individual ATPase activities in root cells is complicated by both the diversity of ATPases in these cells and limitations on the physical separation of individual membrane types. For example three types of electrogenic ATPase can be detected in root cells and differentiated on the basis of their response to a range of inhibitors (7). These include the mitochondrial azide- and oligomycin-sensitive FIFo -energy transducing ATPase, the

Vanudute Inhibition of Perbcteroid Membrane Mg-ATPase Activity and K, of Mg-ATP Kinetic analysis of vanadate inhibition of plasma membrane proton translocating ATPases has provided useful information leading to the proposal of reaction mechanisms for these enzymes (15). A similar analysis of the peribacteroid membrane ATPase is described in Fig. ‘7. The assays were carried out using 1:l ratios of Me:ATP and the data were subjected to the Hanes-Woolf transformation to determine the K, and V,,,,, of the enzyme in the presence of 10 and 100 PM vanadate. This study shows that while vanadate altered the V,,, of the enzyme (43.5, 27.0, and 16.1 pmol Pi liberatedimg protein/h at 0, 10, and 100 pM vanadate) no significant change was observed in the K, (0.31, 0.28, and 0.34 mM at 0, 10, and 100 pM vanadate). Thus vanadate appears to act on

-1

0

1

2 3 ATP(mM)

4

5

FIG. 7. Kinetic analysis of vanadate inhibition of peribacteroid membrane ATPase activity. Peribacteroid membrane ATPase activity was measured in assay system A at pH 5.25 in 1:l ratios of Mg:ATP up to 5 mM and in the presence of 0 mM (m), 10 PM (0), and 100 pM (0) Na vanadate. The data were subjected to the Hanes-Woolf transformation and the V,,,, and Km at each inhibitor concentration were estimated by regression analysis.

572

DOMIGAN

tonoplast nitrate-sensitive ATPase, and the plasma membrane vanadate-sensitive ATPase. This situation becomes more complex with the infection of root cells by rhizobia during the production of root nodules. The bacteroids are marked by an azide-sensitive energy transducing ATPase activity associated with the bacterial inner envelope (12) and are surrounded by the peribacteroid membrane. This last membrane system, on the basis of electron microscopic evidence, appears to be derived, at least initially, from the plasma membrane. In order to biochemically analyze and understand the structure and function of the peribacteroid membrane it is desirable to purify the peribacteroid membrane free from other intracellular membranes and to define suitable enzymatic markers specific to this important intracellular boundary. Peribacteroid membranes can be released from membrane-enclosed bacteroids by osmotic shock and purified by flotation density gradient centrifugation(s). The ATPase activity of the peribacteroid membrane fraction recovered by this procedure does not appear to include significant contamination by the FIFo energy transducing ATPase of rhizobia or plant mitochondria because it is insensitive to azide and oligomycin. Tonoplast ATPase activity can be ruled out because the peribacteroid membrane ATPase is insensitive to nitrate ions (16). In contrast to these ATPases, the peribacteroid membrane ATPase has an acidic pH optimum (5.25), exhibits a high level of specificity for Mg-ATP or Mn-ATP, is very sensitive to vanadate, diethylstilbestrol, and sodium fluoride, and shows moderate sensitivity to molybdate, DCCD, and thiocyanate. In addition, the ATPase is noncompetitively inhibited by vanadate, whether Mg-ATP or Mn-ATP is the substrate. Many of these characteristics appear compatible with the idea that the peribacteroid membrane ATPase may be similar to, or a derivative of, the root plasma membrane proton pumping ATPase (6-8). For example, plasma membrane proton pumping ATPases are characterized by an aspartylphosphate intermediate, and the

ET AL.

formation of such an intermediate is noncompetitively inhibited by vanadate. However some of the properties of the lupin peribacteroid membrane ATPase are at variance with the enzyme from soybean (6), an enzyme postulated to be a plasma membrane-like proton pumping ATPase. The pH optimum of the lupin enzyme is even more acidic than that reported by Blumwald et al. (6) for the soybean peribacteroid membrane ATPase (pH 6.0). In addition the lupin peribacteroid membrane ATPase is only weakly stimulated (20%) by potassium ions whereas the soybean enzyme is up to 50% stimulated by this cation (6). This difference could mean that the peribacteroid membrane preparations described in this paper are not sealed and are therefore permeable to cations. This view is supported by the observations that Gramicidin S, an ionophore which abolishes membrane potentials, and ammonia, which readily crosses membranes, both fail to stimulate the lupin peribacteroid membrane ATPase activity. The ATPase activity of the lupin peribacteroid membrane may differ further by exhibiting sensitivity to molybdate (Ki = 400 PM) at pH 5.25 and 6.25. Molybdate is regarded as a pH-independent inhibitor of nonspecific phosphatase activity (1’7) and Blumwald et al. (6) routinely included molybdate (100 PM) in their assays of the soybean peribacteroid membrane ATPase. Under our assay conditions the peribacteroid membrane fraction fails to hydrolyze the acid phosphatase substrate pnitrophenylphosphate, or the adenine nucleotides ADP and AMP, or to convert significant amounts of ATP to AMP (indicative of pyrophosphatase activity). Conversely, in an isotopic assay, we have demonstrated that ATP was nearly quantitatively converted by the peribacteroid membrane-associated ATPase to ADP. Finally, magnesium and manganese ions, which are both activators of the peribacteroid membrane ATPase, barely stimulate lupin acid phosphatase activity (18). We therefore argue that molybdate is a moderate pH-independent inhibitor of the lupin peribacteroid membrane ATPase and that appropriate experiments must be

PERIBACTEROID

MEMBRANE

ATPase

conducted before assigning molybdatesensitive ATPase activity to the presence of nonspecific phosphatase(s). Whether this characteristic and the other differences discussed above are unique to the peribacteroid membrane ATPase of lupin is not known, since the plasma membrane ATPase of lupin has not been characterized. However, studies reported by Mellor et al. (19) have also noted that the soybean peribacteroid membrane ATPase active in the pH range 6.0-8.0 is sensitive to molybdate inhibition. The peribacteroid membrane ATPase appears to be an intrinsic membrane protein which is not readily solubilized. A variety of detergents (octylglucoside, Triton X-100, Zwittergent, N-lauryl sarcosine, Chaps, and SDS) have failed to solubilize the active enzyme (N. M. Domigan, unpublished results). The enzyme either remains insoluble or, if extracted into the soluble phase, is inactivated. As yet we have had no success in reconstituting the “extracted” activity. Inhibition of the ATPase with DCCD, which occurs only in the absence of magnesium, and the failure of the soluble carbodiimide EDAC to inhibit the activity, imply that the ATPase is closely associated with the lipid bilayer. By analogy with the proteolipid subunit of the mitochondrial FIFOATPase, the DCCD binding site may be a hydrophobic domain which carries a carboxyl group essential for ion or proton transport (13). The ability of the ATPase to bind DCCD may provide a useful approach to the identification of the polypeptide components of the molecule. The close association of the ATPase with the lipid bilayer, its spectrum of specific inhibitors, and its sensitivity to micromolar levels of calcium suggest that the peribacteroid membrane-associated ATPase will prove a valuable marker of the peribacteroid membrane, even though the polypeptide composition of the enzyme is to date unknown. However, if the ATPase is indeed similar to the proton translocating ATPases of plant and fungal plasma membranes (8), then it would be reasonable to predict that it would include a subunit of about 100,000 Da.

OF LUPIN

ROOT NODULES

573

ACKNOWLEDGMENTS We thank Julia McNaughton for expert technical assistance and Brenda van Turnhout for preparation of the manuscript.

REFERENCES 1. VERMA, D. P. S., KAZAZIAN, V., ZOGBI, V., AND BAL, A. K. (1978) J. Cell Biol. 78,919-936. 2. VERMA, D. P. S., (1982) in The Molecular Biology of Plant Development (Smith, H., and Grierson, D., Eds.), pp. 437-466, Blackwell Scientific, London. 3. VERMA, D. P. S., AND NADLER, K. (1984) in Genes Involved in Plant Microbe Interactions (Verma, D. P. S., and Hohn, T., Eds.), pp. 57-93, Springer-Verlag, New York. 4. ROBERTSON, J. G., LY~ETON, P., BULLIVANT, S., AND GRAYSTON, G. F. (1978) J. Cell Sci. 30, 129-149. 5. ROBERTSON, J. G., WARBURTON, M. P., LYMETON, P., FORDYCE, A. M., AND BULLIVANT, S. (1978) J. CeU BioL30,151-174. 6. BLUMWALD, E., FORTIN, M. G., REA, P. A., VERMA, D. P. S., AND POOLE, R. J. (1985) Plant Physiol 78,665-672. 7. SZE, H. (1984) Phvsiol. Plant. 61,683-691. 8. SERRANO, R. (1984) Cum: Top. Cell. Regul. 23, 87-126. 9. MCCORMACK, D. K., FARNDEN, K. J. F., AND BoLAND, M. J. (1982) Arch. B&hem Biophys. 218,561-571. 10. BENCINI, D. A., WILD, J. R., AND O’DONOVAN, 132.254-258. G. A. (1983) And Bioc&m 11. POLACHEK, I., AND CABIB, E. (1981) And Bioc/&em 117,311-314. 12. ROBERTSON, J. G., THOMAS, J. E., GOWING, L. R., AND BOLAND, M. J. (1984) Arch. Biochem.

Biophys. 232,337-347. 13. SOLIOZ, M. (1984) Trends Biochem Sci 9.309-312. 14. BALKE, N. E., AND HODGES, T. K. (1979) Plant

Physiol 63,53-56. 15. GOFFEAU, A., AMORY, A., VILLALOBO, A., AND DUFOUR, J-P. (1982) Ann& N.Y. Ad Sci. 402.91-98. 16. POOLE, R. J., BRISKIN, D. P., KRATKY, Z., AND JOHNSTONE, R. M. (1984) Plant Physiol. 74, 549-556. 17. GALLAGHER, S. R., AND LEONARD, R. T. (1982) Plant Physiol 70,1335-1340. 18. NEWMARK, M. Z., AND WENGER, B. S. (1960) Arch B&&em. Biophys. 89,110-117. 19. MELLOR, R. B., CHRISTENSEN, T. M. I. E., BASSARAB, S., AND WERNER, D. (1985) 2. Naturfor&L c 40,73-79.