THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 266, No. 30, Issue of October 25, pp. 20470-20475,1991

Printed in U.S.A.

Identification of an Autoinhibitory Domain in the C-terminal Region of the Plant Plasma Membrane H+-ATPase*

(Received for publication, March 14,1991)

Michael G. PalmgrenSjlI, Marianne SommarinS, Ramon Serranoj, and Christer LarssonS From the $Department of Plant Biochemistry, University of Lund, Box 7007, S-220 07 Lund, Sweden and the $European Molecular Biology Laboratory, Postfwh 10.2209,D-6900 Heidelberg, Germany

Proteolytic (trypsin) treatment removes a small ter- minal segment from the 100-kDa plant plasma mem- brane H+-ATPase. This results in activation of H+ pumping across the plasma membrane, suggesting that an inhibitory domain is located in one of the terminal regions of the enzyme (Palmgren, M.G., Larsson, C., and Sommarin, M. (1990) J. Biol. Chem. 266, 13423- 13426). In order to identify the origin of the fragment released by trypsin, polyclonal antibodies were raised against the first 66 amino acids (N-terminal region), the last 99 amino acids (C-terminal region), and a portion of 160 amino acids in the central part of the enzyme as deduced from one of the H+-ATPase genes (PMA2) of Arubidopsis thdiuna. The native, 100-kDa H+-ATPase was recognized by all three antisera in Western blots. By contrast, the -90-kDa polypeptide appearing after trypsin treatment was only recognized by the antisera against the N-terminal and central region, but not by the antiserum against the C-terminal region, suggesting that the inhibitory domain is located in this part of the enzyme. To more closely determine the position of the inhibitory domain, three peptides representing conserved parts of the C-terminal region were synthesized (residues 861-888, 912-943, and 936-949 of the Arubidopsis (PMA2) sequence). Only one of the peptides (residues 861-888) affected H+ pumping by the trypsin-activated (-90-kDa) enzyme. This peptide of 28 amino acids inhibited H’ pumping with an ICao of about 15 PM, suggesting that the auto- inhibitory domain is located within the corresponding part of the C-terminal region.

The plant plasma membrane H+-ATPase plays a central role in plant physiology. This enzyme is a member of the P type family of cation pumps (Pedersen and Carafoli, 1987). Using ATP as the energy source, the plasma membrane H+- ATPase pumps protons across the plasma membrane from the cytoplasm to the plant cell exterior. Acidification of the cell wall is thought to induce cell wall plasticity, thus allowing

* This work was supported by grants from the Danish Agricultural and Veterinary Research Council (to M. G. P.), the Danish Natural Science Research Council (to M. G. P.), the Danish Research Acad- emy (to M. G. P.), the Swedish Council of Forestry and Agricultural Research (to M. S.), the Swedish Natural Science Research Council (to M. S. and C. L.), and the Carl Tesdorpf Foundation (to C. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

(1 A European Molecular Biology Organization long term fellow. To whom correspondence should be addressed: European Molecular Biology Laboratory, Meyerhofstrasse 1, Postfach 10.2209, D-6900 Heidelberg, Germany.

turgor-driven cell wall expansion. Alkalinization of the cyto- plasm may be one of the factors triggering cell division. The electrochemical gradient of protons that develops across the plasma membrane drives nutrient uptake through H+ sym- ports and channel proteins (Serrano, 1989, 1990). Since the activity of this enzyme is essential for plant nutrient uptake and growth, it is subject to regulation by plant growth factors such as plant hormones and light, which alter the activity of the plasma membrane H+-ATPase in vivo (Serrano, 1989, 1990). However, the mechanism for this regulation is un- known.

Many proteins undergo posttranslational modification. These modifications are often reversible and regulate the biological activity of the protein. Examples are the phos- phorylation and dephosphorylation of proteins by protein kinases and phosphoprotein phosphatases. Regulation of pro- teins by irreversible modification can be exerted by, e.g. specific proteolysis that occurs frequently in biological sys- tems. A common feature for both types of modifications is that the enzyme involved often has an inhibitory domain that is removed (temporarily or permanently) by the effectors when the enzyme is activated. We have previously shown that proteolytic removal of a terminal 7-kDa segment from the oat root plasma membrane H+-ATPase activates H+ pumping across the plasma membrane, which led us to suggest that an inhibitory domain is located in one of the terminal regions of the enzyme (Palmgren et al., 1990a). Using antibodies specific to the N- and C-terminal regions, respectively, we have now localized the inhibitory domain to the C-terminal region. Using synthetic peptides corresponding to three conserved parts of the C-terminal region, we have more closely mapped the inhibitory domain to a 28-amino acid stretch. This auto- inhibitory domain may be the ultimate target for hormones and toxins that function as regulators of H’ pumping across the plant plasma membrane.

MATERIALS AND METHODS

Reagents-All chemicals were of highest commercially available grade. Trypsin, soybean trypsin inhibitor, chymotrypsin, pyruvate kinase (solution in glycerol), and lactate dehydrogenase (solution in glycerol) were obtained from Boehringer Mannheim. Trypsin-cby- motrypsin inhibitor was from Sigma.

Plant Material-Four-week-old sugar beet plants (Beta vulgaris L.) were kindly supplied by Hilleshog AB, Sweden. Plants were maintained in soil in a greenhouse with supplementary light (23 watts m-’, 350-800 nm; Pbilips G/86/2 HPLR 400 W). Leaves of 6-8-week- old plants were used. Oat (Auena satiua L. cv. Rhiannon) was grown hydroponically in the dark for 8 days and the roots harvested and treated as described earlier (Sommarin et al., 1985). Arabidopsis thulium plants were grown in soil as above with full nutrient supply. Four-week-old plants (everything above ground) were used.

Plasma Membranes-Plasma membranes were purified from a microsomal fraction by aqueous polymer two-phase partitioning (Larsson et al., 1987) as modified for sugar beet leaves (Palmgren et

20470

Autoinhibitory Domain of the Plant Plasma Membrane H+-ATPase 20471

al., 1990~) and oat roots (Palmgren and Sommarin, 1989). Archidopsis plasma membranes were prepared as for sugar beet leaves.

Protease Treatment-Plasma membranes (60 pg of protein in 24 pl) in 25 mM Mops-BTP,' pH 7.5, 4 mM ATP-BTP, 5 mM EDTA- BTP, 2 mM dithiothreitol, and 250 mM sucrose were mixed with an equal volume of 25 mM Mops-BTP, pH 7.5,5 mM EDTA-BTP, 2 mM dithiothreitol, and 250 mM sucrose containing 2.4 pg of trypsin. Incubation was at 20 "C, and proteolysis was stopped by addition of 24 pg of soybean trypsin inhibitor in 24 p1 of the medium used for trypsin above. With chymotrypsin, proteolysis was stopped by adding trypsin-chymotrypsin inhibitor (otherwise as for trypsin). In controls, protease inhibitor was added before the plasma membranes, or pro- tease and protease inhibitor were omitted.

Proton Pumping-Proton uptake into the vesicles was monitored as the absorbance decrease at 495 nm of the ApH probe acridine orange (Palmgren, 1990). The H+-ATPase reaction mixture (910 pl; 20 "C) containing acridine orange was added to the protease incuba- tion mixture (72 pl, including inhibitor; see above) to give the follow- ing final concentrations: 10 mM Mops-BTP, pH 7.0, 140 mM KC1, 1 mM EDTA, 1 mM dithiothreitol, 2 mM ATP-BTP, and 20 pM acridine orange. Proton pumping was initiated by addition of 4 mM MgCl, (15 pl) to the assay mixture. The rate of H+ accumulation was estimated from the initial slope of absorbance quenching of acridine orange.

ATPase Assay-ATPase activity and H+ pumping were monitored simultaneously in the same cuvette (Palmgren and Sommarin, 1989; Palmgren, 1990). One mM phosphoenolpyruvate, 0.25 mM NADH, 60 pg/ml pyruvate kinase, and 30 pg/ml lactate dehydrogenase were included in the H+-ATPase reaction mixture above. Released ADP coupled to the oxidation of NADH was measured at 340 nm, whereas H+ pumping was measured as acridine orange absorbance quenching at 495 nm.

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis-To the protease incubation mixture (72 pl, including inhibitor; see above), 14 pl containing 5 mg/ml tosyl-L-lysine chloromethyl ketone, 1 mg/ ml leupeptin, and 20 mM p-aminobenzamidine was added. After 5 min, 14 p1 of 10 mM phenylmethylsulfonyl fluoride (in ethanol) was added. After another 5 min, 33 pl of this mixture (20-50 pg of membrane protein) was solubilized in sodium dodecyl sulfate at 20 'C for 15 min and subjected to slab gel electrophoresis essentially ac- cording to Laemmli (1970) (total monomer concentration, 8%; cross- linking, 2.7%).

Generation of Antibodies-Domains of the cloned H+-ATPase gene (Pardo and Serrano, 1989a) were expressed in Escherichia coli, and specific polyclonal antibodies were raised against the gene products. The N-terminal region of the H+-ATPase gene was obtained from the cDNA clone (Pardo and Serrano, 1989a) as a XhoI fragment of 140 base pairs. It was subcloned with the right orientation into the Sal1 site of the expression vector pEXl (Stanley and Luzio, 1984). This produced an in-frame fusion of the cro-lac2 gene with the coding region for amino acids 6-51. The central part of the H+-ATPase gene was obtained as a Hind11 fragment of 930 base pairs. The first HindIII site is present in the cDNA clone, and the second HindIII was present in the polylinker of the plasmid in a deletion clone generated during sequencing (Pardo and Serrano, 1989a). After blunt- ending with the Klenow fragment of DNA polymerase, the fragment was subcloned into the S m I site of the expression vector pEX1 (Stanley and Luzio, 1984). This produced an in-frame fusion of the cro-lac2 gene with the coding region for amino acids 340-650. The production of an in-frame fusion with the coding region for amino acids 851-949 from the C-terminal region of the Hf-ATPase gene has been described elsewhere (Parets-Soler et al., 1990). Purification of the three fusion proteins from the inclusion bodies of the bacteria and rabbit immunization were performed as described in the PEXFIT manual (Genofit, Geneva, Switzerland).

Western Blot Analysis-After sodium dodecyl sulfate-polyacryl- amide gel electrophoresis as above (20-50 pg of protein/lane) the polypeptides were electrophoretically transferred to an Immobilone polyvinylidene &fluoride transfer membrane (Millipore Corp.) and reacted with rabbit antisera to the fusion proteins described above, using standard procedures.

Protein-Protein was measured essentially as in Bearden (1978) with bovine serum albumin as standard.

' The abbreviations used are: Mops, 4-morpholinepropanesulfonic acid; BTP, 1,3-bis(tris(hydroxymethyl)methylamino)propane.

RESULTS

Orientation of Plasma Membrane Vesicles-Plasma mem- branes isolated by aqueous polymer two-phase partitioning are mainly right-side-out (cytoplasmic side-in) vesicles (Lars- son et al., 1984). However, when subjected to freezing and thawing, some vesicles become everted (Palmgren et al., 1990~). The plasma membrane preparations used in this study were stored in liquid nitrogen until use and were freeze/ thawed at least once. They were therefore a mixture of 30- 50% inside-out (cytoplasmic side-out) and 50-70% right-side- out vesicles, with the relative proportions depending on spe- cies. Only the inside-out vesicles hydrolyze ATP in the ab- sence of detergent and pump H+, since the active site of the H+-ATPase is located on the cytoplasmic surface of the mem- brane (Serrano, 1989).

Effect of Proteases on the H+-ATPase Activity-Treatment of plasma membrane vesicles from leaves of sugar beet with trypsin markedly increased the rate of ATP hydrolysis and H+ pumping (Figs. 1 and 2). Stimulation of H+ pumping was more than 7-fold under optimal conditions (Figs. lA and M), whereas the rate of ATP hydrolysis was always stimulated to a lesser degree (2-3-fold; Figs. 1B and 2.4). Similar experi- ments using chymotrypsin, a protease with a substrate spec- ificity much different from that of trypsin, revealed a similar pattern in stimulating H+ pumping and ATPase activity of the sugar beet H+-ATPase (data not shown) in agreement with earlier results with oat root plasma membranes (Palm- gren et al., 1990a). Proton pumping was also stimulated in plasma membrane vesicles from A. thalianaplants (see below). In the latter case, however, only chymotrypsin produced ac- tivation.

Western blotting revealed that the main band at 100 kDa

nlperlcln

11

FIG. 1. Stimulatory effect of trypsin on H+ pumping and ATP hydrolysis of sugar beet plasma membrane vesicles. The plasma membranes were treated with trypsin (50 pg/ml) for 16 min, and then proteolysis was stopped by the addition of trypsin inhibitor (500 pg/ml). Controls were without trypsin, or trypsin and trypsin inhibitor were mixed before addition to the plasma membranes. Additions of M$+ and nigericin (0.1 pg/ml) are indicated by arrows. A, proton uptake into the vesicles (50 pg/ml membrane protein) was measured as the absorbance decrease at 495 nm of the ApH probe acridine orange. B, ATP hydrolysis was measured by coupling the production of ADP to oxidation of NADH ( A A Q I 0 ) .

20472 Autoinhibitory Domain of the Plant Plasma Membrane H+-ATPase

0' I I 0 10 20 30

B Trypsin exposure, min

kDa 0 4 11 30

91 $

FIG. 2. A, effect of trypsin on H+ pumping (0) and ATP hydrolysis (0) of sugar beet plasma membrane vesicles. Control activities (100%) were 0.38 AA4gS/min/mg of membrane protein for H' pumping and 0.43 pmol of ADP/min/mg for ATP hydrolysis. B, identification of the plasma membrane H'-ATPase after trypsin treatment using Western blotting. A rabbit antiserum to a fusion protein between the N-terminal region of the plasma membrane H+-ATPase and P-galac- tosidase was used (see "Materials and Methods"). The membranes used were a mixture of 30% inside-out (cytoplasmic side-out) and 70% right-side-out vesicles, making only a limited fraction of the H+- ATPase polypeptide accessible to proteolytic attack.

12 34 5 6 7 8

L n C aa6-51 aa 339-650 aa 851-949

. FIG. 3. Schematic secondary structure of the plant plasma

membrane H+-ATPase. 1-8 are hydrophobic stretches proposed to constitute transmembrane a-helices. The hydrophilic stretches used to generate the antisera against the N-terminal region (aa 6-51), central part (aa 339-650), and C-terminal region (aa 851-949) are indicated.

representing the plasma membrane H'-ATPase diminished to -50% during incubation of the membranes with trypsin, and a new band of slightly lower molecular weight appeared, followed by a band at 91 kDa (Fig. 2B). During prolonged incubation with trypsin in the absence of ATP, the 91-kDa band rapidly disappeared (data not shown), but in the pres- ence of ATP, the 91-kDa band was relatively stable (Fig. 2B). The changes in H' pumping and ATPase activity correlated with the changes in the 91-kDa band (Fig. 2). Similar results were obtained using plasma membrane vesicles from oat roots (Palmgren et al., 1990a) and A. thuliam plants (see below). These data suggest that activation of H' pumping across the plasma membrane is caused by the removal of a small terminal fragment from the H+-ATPase. This fragment should be exposed on the same side of the membrane as the active site, i.e. on the cytoplasmic side, and the fact that it is removed from only 50% of the H+-ATPase molecules reflects the proportion of cytoplasmic side-out vesicles (see above).

Identification of the Proteolytic Fragment-To identify the origin of the fragment released by trypsin, polyclonal antibod- ies were generated against the first 55 amino acids (N-termi- nal region), the last 99 amino acids (C-terminal region), and a portion of 150 amino acids in the central part of the H+- ATPase, as deduced from the AHA3 gene of A. thaliana (Fig. 3). These antibodies reacted with the plasma membrane H+- ATPase from sugar beet, A. thulium plants, and oat roots (Fig. 4).

All three antisera recognized the native 100-kDa H'- ATPase polypeptide (Fig. 4). This band often appeared as a double band also in controls without protease. This may either be the result of endogenous proteases, or it may reflect true isozymes differing slightly in migration, as demonstrated for

Central C-term. loop

N-term. region region

0 15 0 15 0 15 kDa

0 30 0 30 0 30 Arabidopsis (+ 70%) anan *Jlm e 1 0 0

. ,&& f

0 16 0 16

FIG. 4. Western blot analysis of proteolytic cleavage prod- ucts from the plasma membrane H+-ATPase of sugar beet leaves, A. t h ~ l i ~ n ~ plants, and oat roots. Trypsin (chymotrypsin for Arabidopsk) treatment was for 0-30 min, as indicated above each lane. The proteolytic products were labeled with antisera raised against the central part of the H'-ATPase, the C-terminal region, and the N-terminal region, respectively. The apparent molecular mass of the native H'-ATPase is 100 kDa, and those of the proteolytic products are as indicated. Percent stimulation of H' pumping upon proteolytic treatment were as indicated in parentheses.

Na'/K'-ATPase isozymes (Sweadner, 1989). Only the anti- sera against the N-terminal region and the central part rec- ognized the 90-93-kDa band (the apparent molecular mass is dependent on species) appearing after protease treatment. Thus, the 90-93-kDa band was not recognized by the anti- serum against the C-terminal region, suggesting that this region was removed. Considering the size of the fragment released by the protease treatments, it seems that essentially the entire hydrophilic C-terminal region after the last trans- membrane a-helix is removed (see Figs. 3 and 5).

Effect of Synthetic Peptides on the H+-ATPase Activity- To more closely determine the position of the inhibitory domain, three synthetic peptides corresponding to conserved parts in the C-terminal region (Fig. 5) were prepared, and their ability to inhibit H+ pumping by the trypsin-activated enzyme was investigated. Only one of the peptides, represent- ing residues 861-888, affected H+ pumping. This peptide inhibited H+ pumping with a IC5o of about 15 p~ (Fig. 6). The inhibition was time-dependent, with maximal effect ob- tained after more than 1 h of incubation (Fig. 6). Proton pumping by the native enzyme, i.e. the enzyme not activated by trypsin, was not affected (data not shown). The two other peptides, representing residues 912-943 and 936-949 in the C terminus of the H+-ATPase, did not affect the proteolytically activated H'-ATPase up to a concentration of 80 pM. All three peptides are hydrophilic. Peptides 861-888 (inhibitory) and 912-943 (without effect) are of about the same length (28 and 33 residues, respectively), contain about 30% charged residues (peptide 861-888: 5 positive, 4 negative; peptide 912- 943: 8 positive, 5 negative), and are both predicted by the method of Chou and Fasman (1978) to form an a-helical structure. This suggests that the direct inhibition of the H+- ATPase by peptide 861-888 was due to specific structural features.

Similarities between Activation by Lysophospholipids and Activation by Protease Treatment-We have earlier shown that lysophospholipids stimulate both H' pumping and ATP hydrolysis of oat root plasma membrane vesicles. Lysophos- pholipids increase V,,,.,, as well as the affinity for ATP, but ATP hydrolysis is much less stimulated than H' pumping (Palmgren et al., 1988; Palmgren and Sommarin, 1989), sim-

Autoinhibitory Domain of the Plant Plasma Membrane H+-ATPase 20473 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Sp.875-ILSESAGFDR.HNGKP.KESRNQRSIEDLWALQRTSTRHEKGDA Nc.877-ILQDSVGFDNLMHGKSPKGNQKQRSLEDFWSLQRVSTQHE SC,~~~-EMSTSEAFDRWNGKPMKEKKSTRSVEDFWQRVSTQHEKET

LHA1,847-SGKAWDLVLEQRIIVTRKKDFGKELRELQWAHAQRTLHGLQVPDP.KIFSETTNFNELNQLAEEAKRRREIARLRELHTLKGHVESWKLKGLDlETI~SYTVU LHA2,847-SGRRWDLVLEQRIAFTRKKDFGKEQRELQWAHAQRTLHGLQVPDT.KLFSEATNFNELNQ~EEAKRRI\EIARQRELHTLKGH~SWK~KGLDlETl~SYWU

Np,847-SGRI\WDLVLEQRIAFTRKKDFGKEQRELQWAHAQRTLHGLQVPDT.KLFSEATNFNELNQLAEEAKRRREIARQRELHTLKGHVESVVKLKGLDIETI~SYTV AHA1,847-SGKRWASLFDNRTAFTTKKDYGlGEREAQWAQAQRTLHGLQPKED~IFPEKGSYR€LSEIAEQAKR~ElARLRELHTLKGHVESVAKLKGLDlDTAGHHYTV AHA2,847-SGKRWLNLFENKTAFTMKKDYGK~EREAQWALAQRTLHGLQPKEA~lFPEKGSYRELSElAEQAKRRREIARLRELHTLKGHVESWKLKGLDIETPS.HYTV AHA),847-AGTAWKSIIDNRTAFTTKQNYGIEEREAQWAHAQRTLHGLQNTETANWPERGGYRELSEIANQAKRRREIARLRELHTLKGHVESWKLKGLDlETAG.HYTV ..................... ......................... f t..... ......*....... f ....I.........

AFTMKKDYGKEEREAQWALAQRTLHGLQ ........ .........

peptides . * . e .

AKRRREIARLRELHTLKGHVESWKLKGLDIET KLKGLD1ETPS.HYW

p861-888 p912-943 p936-949

FIG. 5. Amino acid sequences of the hydrophilic C-terminal region of plant and fungal H+-ATPases. The deduced amino acid sequences of the hydrophilic C-terminal region after the last putative transmembrane a- helix of plasma membrane H+-ATPases from the fungi Schizosacchnromyces pombe (Sp; Gislain et al., 1987), N. crassa (Nc; Hager et al., 1986), and Saccharomyces cerevisine (Sc; Serrano et al., 1986) are compared with those of the plasma membrane H+-ATPases from the higher plants Lycopersicon esculentum (isoforms LHAl and LHAP; Ewing et al., 1990), Nicotiana plumbaginifolia (Np; Boutry et al., 1989) and A. thnliana (isoforms AHAl, AHAP, and AHA3; Pardo and Serrano, 1989b). Asterisks in the upper line indicate identity in all H+-ATPases, whereas asterisks in the lower line indicate identity in the H+-ATPases from plants only. The sequences of the synthetic peptides (p861-888,912-943, and 936-949) are shown in the bottom line.

0 10 20 30 40 Peptide, pM

FIG. 6. Effects of synthetic peptides representing conserved domains in the C-terminal region on H+ pumping of trypsin- activated H'-ATPase in sugar beet leaf plasma membrane vesicles. Plasma membranes (60 pg of protein) were incubated with trypsin (2.4 pg) for 16 min. After stopping the reaction with trypsin inhibitor (24 pg), the plasma membranes were preincubated with synthetic peptides (p861-888, open symbols; p912-943 and 936-949, closed symbols) for the indicated period of time, and then H+ pumping was assayed. The amino acid sequences of the peptides are shown in Fig. 5. Control activity (100%) was 2.4 AA495/min/mg of membrane protein.

ilarly to what is observed upon proteolytic activation (Figs. 1 and 2) (Palmgren et al., 1990a). These similarities suggest that proteases and lysophospholipids activate the H'-ATPase by a similar mechanism. To test this hypothesis, plasma membrane vesicles were treated with lysophosphatidylcholine and the detergent Brij 58. Lysophosphatidylcholine stimu- lated ATP hydrolysis severalfold, whereas Brij 58 (a detergent that permeabilizes the vesicles but neither activates nor in- hibits the H'-ATPase (Palmgren et al., 1990b)) only increased the activity about %fold, corresponding to the amount of permeabilized right-side-out vesicles (Fig. 7A, and see above). The following addition of trypsin stimulated ATP hydrolysis of plasma membranes treated with Brij 58, whereas no further stimulation was observed after activation with lysophospha- tidylcholine. Furthermore, ATP hydrolysis reached the same final level after both lysophospholipid and protease activation (Fig. 7B). A possible mechanism for the activating effect of lysophospholipids could be that they either activate an endog- enous protease, or by binding to the C-terminal region, make it susceptible to protease attack. However, lysophosphatidyl- choline did not stimulate proteolytic degradation of the H+- ATPase in the absence of trypsin as revealed by Western blotting (Fig. 7C). The following trypsin treatment, however, degraded all H'-ATPase in both the Brij 58- and the lyso- phosphatidylcholine-treated vesicles and left the 91-kDa poly-

Delrrgent Trypsin exposure

C L K h l rn,"

~ Trypsin + Trypsin I 15' 15' 30' 30' 15'

kDa Convol LFC Brij lLpc Brij LPC Brij Convol

FIG. 7. A, effects of the detergents Brij 58 (0) and lysophospha- tidylcholine (0, Lyso-PC) on ATP hydrolysis of sugar beet plasma membrane vesicles. B, effect of trypsin on ATP hydrolysis of the detergent-treated vesicles. The concentration of both detergents was 350 pg/ml. C, Western blot of control and detergent-treated vesicles before and after exposure to trypsin using an antiserum raised against the N-terminal region of the H+-ATPase (compare with Fig. 4). Detergent concentrations were 350 pg/ml. LPC, lysophosphatidylcho- line; Brij, Brij 58.

peptide as the only prominent species (Fig. 7C). In this case, no 100-kDa H+-ATPase was left since, in the detergent solu- bilized vesicles, all H'-ATPase molecules were available for proteolytic attack (compare with the control, Fig. 7C, far right). Similar results were obtained with oat root plasma membranes (data not shown). Thus, activation of the H+- ATPase by proteolytic removal of the C terminus and by lysophosphatidylcholine treatment are very similar both quantitatively ( VmaX) and qualitatively (decreased K, for ATP), and these activations are not additive (Fig. 7). This suggests that activation by lysophosphatidylcholine proceeds by the same mechanism, namely displacement of the inhibi- tory domain in the C-terminal, although this removal in the case of lysophosphatidylcholine involves a noncovalent mod- ification of the enzyme.

DISCUSSION

We have previously shown that proteolytic removal of a terminal 7-kDa segment from the oat root plasma membrane H+-ATPase activates H' pumping across the membrane, which led us to suggest that an inhibitory domain is located in one of the terminal regions of the enzyme (Palmgren et aL, 1990a). Using sequence-specific antibodies and synthetic pep-

20474 Autoinhibitory Domain of the Plant Plasma Membrane H+-ATPase

tides, we have now obtained evidence that the inhibitory domain is located within a 28-amino acid stretch of the C- terminal region.

The effect of trypsin on the activity of the plant plasma membrane H+-ATPase varies with the presence and absence of ATP or ADP, suggesting that sites within the polypeptide that are accessible to tryptic cleavage are depending on the conformational state of the enzyme (Palmgren et al., 1990a). The related H+-ATPases from yeast and Neurospora crussa plasma membranes respond to trypsinolysis in a way very similar to that of the plant enzyme. In the absence of sub- strate, trypsin treatment results in rapid inactivation of both the yeast (Perlin and Brown, 1987) and the Neurospora (Ad- dison and Scarborough, 1982) enzyme. However, in the pres- ence of substrate, trypsin treatment results in enhanced rates of ATP hydrolysis, which at least in the case of the yeast enzyme, also seems to be associated with concomitant in- creases in proton transport (Perlin and Brown, 1987). The effect of trypsinolysis on proton pumping exhibited by the Neurospora enzyme has not yet been described.

In the present study, it has not been possible to identify the site(s) of tryptic cleavage within the C terminus of the plant H+-ATPase polypeptide because the high number of different polypeptides in the plasma membrane preparation does not allow direct sequencing of proteolytic products. How- ever, the tryptic cleavage sites of the Neurospora plasma membrane H+-ATPase has recently been identified (Mandala and Slayman, 1988; Hennessey and Scarborough, 1990). All of the tryptic sites are localized to the termini of the ATPase polypeptide. In the absence of ligands, three cleavage sites at the N terminus are accessible to trypsin (Lys-24, Lys-36, and Arg-73). Cleavage at Arg-73 is greatly reduced in the presence of ligands, and an additional site at the C terminus (Arg-900) becomes accessible. Cleavage at this new C-terminal site removes a segment of only 20 amino acids from the enzyme. It was suggested by Mandala and Slayman (1988) that inac- tivation of the enzyme by trypsin in the absence of ligands was due to cleavage at Arg-73, whereas the losses of smaller portions of the N and/or C termini may bring about an increase in enzyme activity.

The tryptic cleavage sites of the plasma membrane H+- ATPase from yeast plasma membranes have not been iden- tified. Interestingly, however, an inhibitory domain has been localized to the C-terminal region of the yeast H+-ATPase (Portillo et al., 1989). The activity of the yeast H+-ATPase is regulated by glucose metabolism (Serrano, 1983). Removal (by deletion at the gene level) of 11 amino acids from the C terminus of the protein results in an enzyme in glucose- starved cells with kinetic characteristics identical with the wild-type enzyme activated by growing the cells on glucose (Portillo et al., 1989). This suggests that a part of the C- terminal region of the yeast H+-ATPase functions as a regu- latory domain in vivo (Portillo et al., 1989).

No part of the C-terminal region (the region after the last transmembrane a-helix) is conserved in all the members of the P type ATPase family. However, the inhibitory peptide (residues 861-888 of the Arabidopsis AHA2 sequence) corre- sponds to a part of the C-terminal region with large sequence homology among H'-ATPases of higher plants. There is also some homology (20%) with a part of the C-terminal region of fungal H+-ATPases (Fig. 5). Although the C-terminal region of plant enzymes appears to be considerably longer than that of the fungal enzymes, the domain sharing some homology between the plant and the fungal H+-ATPases is within the same distance from the last putative transmembrane segment. The fungal and the plant enzymes are regulated in vivo by

different modulators, and this might be reflected in the dif- ferences between their C-terminal regions.

A regulatory function has been ascribed to a C-terminal domain of also another P type ATPase, namely the erythro- cyte plasma membrane Ca2+-ATPase. In this ATPase, an amino acid stretch of about 30 amino acids located in the C- terminal region of the protein contains an internal inhibitor of the pump (Brandt et al., 1988; Enyedi et al., 1989). By interaction with the Ca2+-ATPase itself, the internal inhibitor functions as a constraint and induces a state of the enzyme with low Vmax and low affinity for Ca2+. Following binding to Ca2+/calmodulin, the constraint is released. This produces the activated form of the pump with high Vmax and high affinity for Ca2+. The activity of the plasma membrane Ca2+-ATPase is stimulated by mild treatment with trypsin and chymotryp- sin, which removes the C-terminal part of the enzyme. Fol- lowing protease treatment, synthetic peptides corresponding to the inhibitory sequence inhibit the proteolytically activated enzyme (Enyedi et al., 1989,1991). In this manner, the effects of synthetic peptides on the plasma membrane Ca2+-ATPase closely resemble the effects on the plant plasma membrane H+-ATPase described in the present study.

If an internal inhibitor of the plasma membrane H+-ATP- ase constrains this molecule, an important question arises. To which part of the enzyme does the inhibitor bind? The possibilities include the ATP-binding site, the phosphoryla- tion site, and the H+-binding site(s). The same phenotype of the yeast H+-ATPase as that of the ATPase with an 11-amino acid deletion in its C terminus is exhibited by a mutation (Ala-547 --., Val) near the proposed ATP-binding site of the enzyme (Cid and Serrano, 1988). Therefore, the C-terminal inhibitory domain of the yeast H+-ATPase seems to be inter- acting with the ATP-binding site, an interaction that is somehow modulated during growth on glucose. In the eryth- rocyte Ca2+-ATPase, the inhibitory domain has been sug- gested to interact with an amino acid stretch that constitutes a putative Ca2+-binding site next to the inhibitory sequence (Enyedi et al., 1989). When the C-terminal region of the plant plasma membrane H+-ATPase is removed by proteolysis, H+ pumping is stimulated to a higher degree than ATP hydrolysis (Figs. 1 and 2; Palmgren et al., 1990a). Therefore, an attractive hypothesis is that the inhibitory domain of the H+-ATPase interacts with a site involved in H+ binding. It is possible that only in a fraction of the H+-ATPase population, hydrolysis of ATP is coupled to a trans-membrane movement of protons. Proteolysis and lysophospholipids could increase the fraction of ATPases actually pumping protons. They might also cause a more tight coupling between ATP hydrolysis and H' trans- location in individual ATPase molecules. Alternatively, the acridine orange signal is not linearly related to changes in the ATPase activity. This is less likely since the acridine orange absorbance change is closely related to changes in ATPase activity when altering the vanadate concentration, ATP con- centration, and pH (Palmgren and Sommarin, 1989).

Many regulated proteins contain a regulatory domain in either their N- or C-terminal regions. Examples are enzymes such as protein kinases (Soderling, 1990) and adenylate cy- clase (Heideman et al., 1987), structural proteins such as ankyrin (Hall and Bennett, 1985), and receptors such as the yeast a-factor receptor (Reneke et al., 1988). The inhibitory interaction between these domains and the catalytic site is modulated by a variety of means, such as binding of effector molecules, phosphorylation, partial prabeolysis, or truncation 'at 'the gene level.

The observation that the plant plasma membrane H+- ATPase contains an internal inhibitor in the C-terminal

Autoinhibitory Domain of the Plant Plasma Membrane H+-ATPase 20475

mRNA spllclng eItern8tlve

proteolysis phorphoryletlon effector blndlng

J

FIG. 8. Possible modes of regulation of the plasma mem- brane H+-ATPase involving an autoinhibitory domain. The presence of an internal inhibitor (0) in the C-terminal region of the enzyme implies that a wide array of regulatory mechanisms seem possible. Different isoforms may vary in their C-terminal region and exhibit different basal activities. Isoforms can be encoded for by separate genes (a , b) or can be the result of alternative splicing of mRNA from one gene. Posttranslationally, the inhibitory domain can be removed, either irreversibly by proteases or reversibly by phos- phorylation, or as a result of noncovalent effector binding. N , N terminus; C, C terminus; P , hypothetical phosphorylation site.

region of the polypeptide implies that there are several pos- sible ways of regulating the enzyme (Fig. 8). Proteolysis could simply remove such a built-in inhibitor. Introduction of neg- ative charge by protein-kinase mediated phosphorylation or modification of the lipid environment by lysophospholipids could result in conformational changes of the inhibitory do- main of the H'-ATPase polypeptide similarly resulting in activation of the pump. Thus, this autoinhibitory domain may be the ultimate target for hormones and toxins that function as regulators of H+ pumping across the plant plasma mem- brane.

Acknowledgments-We are grateful to Ann-Christine Holmstrom, Inger Rohdin, and Adine Karlsson for excellent technical assistance. We also thank Dr. J. M. Pardo for the preparation of antibodies.

REFERENCES

Addison, R., and Scarborough, G. A. (1982) J. Biol. Chem. 257,

Bearden, J. C., Jr. (1978) Biochim. Biophys. Acta 533, 525-529 Boutry, M., Michelet, B., and Goffeau, A. (1989) Biochem. Biophys.

Brandt, P., Zurini, M., Neve, R. L., Rhoads, R. E., and Vanaman, T.

Chou, P. Y., and Fasman, G. D. (1978) Advances Enzymol. 47, 45-

Cid, A., and Serrano, R. (1988) J. Biol. Chem. 263,14134-14139

10421-10426

Res. Commun. 162,567-574

C. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 2914-2918

148

Enyedi, A., Vorherr, T., James, P., McCormick, D. J., Filoteo, A. G., Carafoli, E., and Penniston, J. T. (1989) J. Bwl. Chem. 264,12313- 12321

Enyedi, A., Filoteo, A. G., Gardos, G., and Penniston, J. T. (1991) J. Bwl. Chem. 266,8952-8956

Ewing, N. N., Wimmers, L. E., Meyer, D. J., Chetelat, R. T., and Bennett, A. B. (1990) Plant Physiol. 94,1874-1881

Gislain, M., Schlesser, A., and Goffeau, A. (1987) J. Biol. Chem. 262, 17549-17555

Hager, K. M., Mandala, S. M., Davenport, J. W., Speicher, D. W., Benz, E. J.. and Slavman. C. W. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,7693-7697

- . . .

Hall. T. G.. and Bennett. V. (1985) J. Biol. Chem. 262.10537-10545 Heideman,'W., Casperson, G. F., 'and Bourne, H. R. (1987) J. Biol.

Hennessev. J. P.. Jr.. and Scarborough. G. A. (1990) J. Biol. Chem. Chem. 262,7087-7091

265,532-537 . , - ,

Laemmli. U. K. (1970) Nature 227.680-685 Larsson, C., Kjellbom; P., Widell, S:, and Lundborg, T. (1984) FEBS

Larsson, C., Widell, S., and Kjellbom, P. (1987) Methods Enzymol.

Mandala, S. M., and Slayman, C. W. (1988) J. Biol. Chem. 263,

Palmgren, M. G. (1990) Plant Physiol. 94,882-886 Palmgren, M. G., and Sommarin, M. (1989) Plant Physwl. 90,1009-

1014 Palmgren, M. G., Sommarin, M., Ulvskov, P., and J~irgensen, P. L.

(1988) Physiol. Plant. 74, 11-19 Palmgren, M. G., Larsson, C., and Sommarin, M. (1990a) J. Biol.

Chem. 265,13423-13426 Palmgren, M. G., Sommarin, M., Ulvskov, P., and Larsson, C. (1990b)

Biochim. Biophys. Acta 1021,133-140 Palmgren, M. G., Askerlund, P., Fredrikson, K., Widell, S., Som-

marin, M., and Larsson, C. (1990~) Plant Physwl. 92,871-880 Pardo, J. M., and Serrano, R. (1989a) J. Bwl. Chen. 264,8557-8562 Pardo, J. M., and Serrano, R. (1989b) in Plant Membrane Transport:

The Current Position (Dainty, J., de Michelis, M. I., Marr6, E., and Rasi-Cladogno, F., eds) pp. 499-500, Elsevier, Amsterdam

Parets-Soler, A., Pardo, J. M., and Serrano, R. (1990) Plant Physiol.

Perlin, D. S., and Brown, C. (1987) J. Bwl. Chem. 262,6788-6794 Pedersen, P. L., and Carafoli, E. (1987) Trends Biochem. Sci. 12,

Portillo, F., de Larrinoa, I. F., and Serrano, R. (1989) FEBS Lett,

Reneke, J. E., Blumer, K. J., Courchesne, W. E., and Thorner, J.

Serrano, R. (1983) FEBS Lett. 156, l l -14 Serrano, R. (1989) Annu. Rev. Plant Physwl. Plant Mol. Biol. 400,

Serrano, R. (1990) in The Plant Plasma Membrane-Structure, Func- tion and Molecular Biology (Larsson, C., Mbller, I. M., eds), pp. 127-153, Springer-Verlag, Berlin, Heidelberg

Serrano, R., Kielland-Brandt, M. C., and Fink, G. R. (1986) Nature

Soderling, T. R. (1990) J. Biol. Chem. 265,1823-1826 Sommarin, M., Lundborg, T., and Kylin, A. (1985) Physiol. Plant.

Stanley, K. K., and Luzio, J. P. (1984) EMBO J. 3, 1429-1434 Sweadner, K. J. (1989) Biochim. Biophys. Acta 988,188-220

Lett. 171, 271-276

148,558-568

15122-15128

93,1654-1658

146-150

247,381-385

(1988) Cell 55,221-234

61-94

319,689-693

65, 27-32