JOURNAL OF BONE AND MINERAL RESEARCH Volume 5, Number 6, 1990 Mary Ann Liebert, Inc., Publishers

Biochemical Characterization of an Electrogenic Vacuolar Proton Pump in Purified Chicken Osteoclast Plasma Membrane Vesicles PETRUS J. BEKKER' and CAROL V. GAY'.*

ABSTRACT A well-characterized chicken osteoclast plasma membrane vesicle preparation manifested Mgz+-dependent ATP hydrolyzing activity of 0.213 pmol inorganic phosphate released per mg protein per minute ( n = 7). The Mg2+dependence showed a high-affinity component with a Khlg of 1.293 pM and Vmax of 0.063 pmol Pi per mg protein per minute, and a low-affinity component with a KhlSof 297.6 pM and a Vma, of 0.232 pmol P, per mg protein per minute. The MgZ+-ATPaseactivity was inhibited by N,N'-dicyclohexylcarbodiimide (DCCD, 0.2 mM, 50.7%), N-ethylmaleimide (0.5 mM, 34.6%), nolinium bromide (1 mM, 29.9%), 4,4'-diisothiocyano-2,2'-stilbene sulfonic acid (DIDS, 1 mM, 45.1%), and p-chloromercuribenzoic acid (PCMB, 0.1 mM, 33.8%). Sodium orthovanadate (Na3V04)at 1 pM had no effect but caused 29.5% inhibition at 1 mM. Na' could substitute for K+without loss of activity, NO,- caused 19.5% inhibition when substituted for CI-, and acetate replacement of CI- resulted in 36.4% stimulation of Mg2+-ATPase.ATP, GTP, ITP, CTP, and ADP were all hydrolyzed effectively. DCCD (0.2 mM), NEM (0.5 mM), nolinium bromide (1 mM), and DIDS (50 pM) almost completely abolished proton transport as measured spectrofluorometrically by acridine orange quenching. Na3V0, (1 mM) had no effect, and duramycin (80 pg/ml) inhibited transport 52.7%. K+ replacement of Na+caused a 79.2% increase in initial proton transport rate. NO,- and acetate substitution of CI- resulted in a 46.1 and 55.7% decrease in transport, respectively. ATP supports transport far more effectively than the other nucleotides tested. ADP was ineffective. Experiments using the potassium ionophore, valinomycin, indicated that the proton pump functions electrogenically, with CI- most likely cotransported by an anion transporter. The proton pump also seems to have at least one anion-sensitive site, elucidated by experiments in the presence of NO,- and CI-.

INTRODUCTION of an active osteoclast have a measured pH of 5 or less.".2) Until quite recently, the mechanism of this acidification was unknown. Baron et al.(3)demonstrated that an antibody to a 100 kD lysosomal protein, which cross-reacted with the gastric H+,K+-ATPase,also reacted with an osteoclast ruffled border membrane protein. Akisaka and Gay'4' showed the presence of a vanadate-resistant, duraHE RESORPTION LACUNAE beneath the ruffled border

T

mycin-sensitive Mg2+-ATPase on the osteoclast ruffled border membrane in an enzyme histochemical study. Tuukkanen and Vaananer~'~' reported that a proton pump similar to the gastric H+,K+-ATPaseis involved in bone resorption, since omeprazole, a H+,K+-ATPaseinhibitor, was found to decrease basal as well as prostaglandin E2 (PGEJ- and parathyroid hormone (PTH)-stimulated, 45Ca2+release from cultured, prelabeled neonatal mouse calvaria. Anderson et a1.(6) found that omeprazole caused M a 17% decrease in osteoclast acidity when present at

'Department of Molecular and Cell Biology, Pennsylvania State University, University Park, PA 16802. 'Department of Poultry Science, Pennsylvania State University, University Park, PA 16802.

569

BEKKER AND GAY

570

in cell culture on the basis of fluorescence measurements at 624 nm using acridine orange. Although omeprazole is considered a specific gastric H+,K+-ATPaseinhibitor,(') its effect on the vacuolar proton pump has not been studied extensively. This has made it difficult to sort out the precise nature of the proton pump in osteoclasts. In this regard it should be mentioned that omeprazole is only active in an acid a condition that can be troublesome to mimic in the test tube but may well apply in a bone's1 or culture system. Recenty, Ghiselli et al.(91and Blair et al.(lO1found that a polyclonal antibody raised against bovine kidney vacuolar H+-ATPase specifically bound to 31, 56, and 70 kD proteins present in a microsomal preparation of chicken osteoclasts. This preparation was also shown to support proton transport, which was sensitive to N-ethylmaleimide (NEM), 7-chloro-4-nitrobenz-2-oxa-1,3-diazole (NBD-CI), and ZnZ' but insensitive to vanadate and oligomycin.('O1 Vaananen et al.'"' showed that antibodies to two subunits of the vaculoar ATPase of Neurospora crassa bound to the osteoclast ruffled border membrane and small cytoplasmic vesicles. The purpose of the present study was to biochemically characterize the Mgz+-ATPase activity and capacity for proton pumping found in a purified plasma membrane vesicle suspension. Proton transport was demonstrated spectrofluorometrically using acridine orange quenching as an indicator of transport. The effect of various inhibitors on transport, substrate specificity, and ion dependence was elucidated. It is concluded that acidification of resorption lacunae is achieved by active electrogenic proton transport mediated by a vacuolar proton pump.

MATERIALS AND METHODS Nolinium bromide was kindly provided by Norwich Eaton Pharmaceuticals, Inc. (Norwich, NY 13813). Duramycin was a gift from Dr. 0. Shotwell, United States Department of Agriculture, Agriculture Research Service (Peoria, IL 61604). Acridine orange was purchased from Polysciences, Inc. (Warrington, PA 18976-2590). Other chemicals were from Sigma Chemical Co. (St. Louis, MO), except sodium chloride, potassium nitrate, magnesium sulfate, and trichloroacetic acid (Baker Inc., Phillipsburg, NJ) and potassium chloride and sucrose (Fisher Scientific, Fair Lawn, NJ). All chemicals were of analytic grade.

Plasma membrane vesicle preparation Twelve to fourteen 3-week-old chicks (Hubbard-Peterson strain), which were maintained on a low-calcium (0.3%) diet for 2 weeks prior to sacrifice, were used for each experiment. After decapitation, the tibiae were removed immediately and placed in Tyrode's buffer, pH 7.27.4, on ice. Osteoclasts were isolated from the endosteal surfaces of longitudinally split bones as described previously.(12)Briefly, the bone marrow was removed by forceps and flushing with a Pasteur pipette in Tyrode's buffer.

After trypsinization (20 minutes, 0.03%, 37"C), the bones were rinsed with Tyrode's buffer and placed in Tyrode's buffer plus 5% fetal bovine serum on ice. Scraping of the endosteal surfaces with a rubber policeman to remove osteoclasts was followed by sequential filtration of the cell suspension (250, 105, 74, and 53 pm meshes). After a Percoll purification step, the cells were homogenized and the plasma membrane fraction isolated by density gradient centrifugation in a Sorvall RC5C centrifuge, using an SS34 rotor.'") It was previously determined by electron microscopy and biochemical marker enzyme analyses that the uppermost fraction of the gradient is vesicular in nature and enriched 7.25- and 11.87-fold in the plasma membrane marker enzymes 5'-nucleotidase and Na+,K+ATPase, respectively. The starting cell population consisted of 77% osteoclasts, with the remainder mainly erythrocytes, However, the needle homogenization technique caused mainly osteoclast disruption.

Characterization of Mg2+-ATPase activity The assay conditions for the Mg2+-ATPaseactivity were essentially similar to those described by Sallman et al.(I3' The medium contained 150 mM KCI, 6 mM MgClz, 30 mM 2-[N-morpholino]ethanesulfonicacid (Mes), 30 mM Tris (pH 7), 0.5 mM ouabain, and 10 pg/ml of oligomycin. [Ethylenebis(oxyethylenenitrilo)]tetraacetic acid (EGTA, 0.5 mM) was included to chelate free calcium and block the Ca"-ATPase also present in the same plasma membrane fraction.(12.14) Plasma membrane vesicles (10-15 pg protein) were added to the medium in a total volume of 0.5 ml. After 15 minutes preincubation at 37"C, the reaction was initiated by adding MgATP (3 mM, final concentration). The reaction was allowed to proceed for 20 minutes at 37°C and was terminated by the addition of ice-cold trichloroacetic acid. Inorganic phosphate release was determined spectrophotometrically by the method of Fiske and Subbarow.'"' Mg'+-ATPase activity was expressed as pmol inorganic phosphate released per mg protein per h. Protein determinations were performed according to Lowry et al.'I6) The effect of increasing concentrations of N, N'-dicyclohexylcarbodiimide (DCCD), N-ethylmaleimide (NEM), sodium orthovanadate (Na3V04),nolinium bromide, 4,4'-diisothiocyano-2,2'-stilbene disulfonic acid (DIDS), and p chloromercuribenzoic acid (PCMB) on the Mg"-ATPase activity was assessed. The result of potassium substitution by sodium, as well as chloride substitution by acetate and sulfate, nitrate and sulfate, or nitrate alone, was also determined. Other experiments were performed to determine the nucleotide specificity of the enzyme. The H+,K+/Na+ ionophore, nigericin (1 pM), and the protonophore, carbonylcyanide rn-chlorophenylhydrazone (CCCP) at 10 pM, were included in some experiments. Finally, the magnesium dependency of the ATPase activity was determined.

Proton transport assay Measurements of proton transport were performed using the fluorescent weak base, acridine orange (1 pM), in

571

AN ELECTROGENIC PROTON PUMP IN CHICKEN OSTEOCLASTS

a reaction medium containing 150 mM KCI, 6 mM MgCL, 6 mM Mes, 6 mM Tris (pH 7.3), 1 mM EGTA, 1 mM ouabain, and 20 pg/ml oligomycin. To 60 p1 isolated vesicles (10-40 pg protein) in 0.25 M sucrose and 5 mM 4-(2hydroxyethy1)-1-piperazineethane sulfonic acid (HEPES), pH 7.3, 60 pl of the medium was added and mixed in a quartz microcuvette (lOS.25O-QS, Hellma, Forest Hills, = 493 nm; NY). Changes in the fluorescent signal (A,, A, = 526 nm) were continuously monitored with a SPEX FLUOROLOG (1681 0.22m) spectrofluorometer, Data were collected by computer and saved for subsequent analysis. MgATP, 2 mM final concentration, was added to the cuvette to initiate transport. The pH of the MgATP solution was adjusted to 7 with 2 M Tris before use. Nigericin, a H+,K+/Na+ionophore, was added to the cuvette at 1 pM final concentration to release accumulated protons (and acridine orange). All additions to the 120 pI vesicle suspension in the cuvette constituted 1% or less of the total volume. Transport data were quantitated by calculating the slope (A~uorescence/Atime)of the initial 150-250 s after MgATP addition. The effect of DCCD (0.2-0.5 mM), NEM (0.5 mM), Na,VO, (1 pM and 1 mM), nolinium bromide (0.1 and 1 mM), duramycin (40 and 80 pg/ml), and DIDS (10, 50, and 100 pM) on proton transport was assessed. Potassium was replaced by sodium and chloride by nitrate or acetate in some experiments. The nucleotide specificity of the H' pump was assessed by replacing ATP with GTP, ITP, CTP, or ADP. The electrogenicity of the pump was investigated by determining the effect of valinomycin, a K' ionophore, on H' transport.

RESULTS The results from typical experiments showing the effect of increasing concentrations of DCCD, NEM, and Na,VO, on Mgz*-ATPase activity are presented in Fig. 1. The average basal activity was 0.213 pmol Pi per mg protein per minute (n = 7). Dose responses to nolinium bromide, DIDS, and PCMB are presented in Fig. 2. Table 1 summarizes the degree of inhibition of the MgZ+-ATPase by these inhibitors. In Table 2 the results of ion substitutions on the Mg'+ATPase activity are presented. The stimulation (36.4%) in the presence of K acetate and MgSO., was due to acetate, not sulfate, stimulation of Mg2+-ATPaseactivity, since no activation was seen with SO,'- alone (results not shown). In Fig. 3 the Mg2+dependence of the ATPase activity can be seen. High-affinity (KMg = 1.293 pM, V,, = 0.063 pmol P, per mg protein per minute) and low-affinity ( K M= ~ 297.6 pM, V,,, = 0.232 pmol Pi per mg protein per minute) components were identified by Eadie Hofstee analysis. Inclusion of the H',K+/Na+ ionophore, nigericin (1 pM), and the protonophore, CCCP (10 pM), in the reaction medium stimulates Mg2+-ATPaseactivity 25.3 and 21.5%, respectively (not shown). MgATP addition to the vesicle suspension containing 1

110 100

90 80

70 60

*

DCCD NEM N%V4

-.-

50

+

40

+I

30 0

I

'

-6.0

-5.0

-4.0

-3.0

log [inhibitor] FIG. 1. The effect of increasing concentrations of DCCD, NEM, and Na3V03 on the Mg'+-dependent ATPase activity observed in isolated osteoclast plasma membrane vesicles. Results are expressed as percentages relative to activity observed in the absence of inhibitors and are averages from at least two separate experiments run in duplicate for each inhibitor tested.

1101

1

15041

-+

\

PCMB

6o

0

'

-7.0

.

'

-6.0

. -5.0 '

'

'

4.0

.

'

-3.0

log [inhibitor] FIG. 2. Mg"-ATPase activity in the presence of increasing concentrations of nolinium bromide, DIDS, and PCMB. Results are expressed as percentages of activity relative to the control activity (without inhibitors) and were calculated from at least two separate experiments run in duplicate.

BEKKER AND GAY

572

TABLE1. EFFECTOF INHIBITORSON THE Mg’+-ATPase ACTIVITY OF OSTEOCLAST PLASMA MEMBRANE VESICLES Inhibitor None DCCD NEM Na3V04 Na3V04 Nolinium bromide DIDS PCMB

Concentration

Yo Activity

-

100.0 49.3 65.4 102.1 70.5 70.1 54.9 66.2

0.2 mM 0.5 mM 1 PM 1 mM 1 mM 1 mM 0.1 mM

Number of experiments 7

3 2 3 2 2 2

TABLE 2. EFFECT OF ION SUBSTITUTIONS ON Mgz+-ATPaseACTIVITY Substitution Control Na’ for K’ NO,- for C1Acetate and SO4,- for CI-

Ionic composition

Yo Activity (number of experiments)

KCI (150 mM), MgCI, (6 mM) NaCl (150 mM), MgC1, (6 mM) KNO, (150 mM), Mg(NO,), (6 mM) K acetate (150 mM), MgS04 (6 mM)

100.0 95.0 (2) 80.5 (2) 136.4 (2)

pM acridine orange resulted in a rapid decrease in the fluo-

rescent signal (Fig. 4). Subsequent addition of the protonophore nigericin (1 pM) resulted in a rapid increase in fluorescence. DCCD and nolinium bromide completely inhibited H’ transport, and duramycin was partially inhibitory (Fig. 4). DIDS was also inhibitory, whereas Na,VO, did not inhibit transport (Fig. 5 ) . The fluorescent signal obtained in the presence of NEM showed a linear decrease even before MgATP was added (Fig. 6A). However, after MgATP addition, the rate of this decrease was only slightly changed from the baseline (Fig. 6A). The percentage transport activity remaining, as calculated from slopes before and after MgATP addition, is listed in Table 3. It was confirmed that a photochemical reaction of NEM with acridine orange was responsible for this decrease, since NEM addition to a 1 pM acridine orange solution caused an immediate linear decrease in the prior stable fluorescent signal (Fig. 6B). A summary of inhibitor effects on the initial rate of proton transport is shown in Table 3. Na’ substitution of K’ resulted in an increase in the initial rate of fluorescence quenching (Fig. 7A and Table 4). Acetate replacement of C1- resulted in a substantial decrease in the initial transport rate (Table 4). Both partial and full NO,- replacement of C1- also caused a decrease in transport rate (Fig. 7B and Table 4). A quantification of these results can be seen in Table 4. The effectiveness of ATP replacement by GTP, ITP, CTP, and ADP to support proton transport is quantitated in Table 5 . The hydrolyzing activity with various substrates is also shown in Table 5 . After transport was initiated by MgATP, the addition

of valinomycin caused a 60% decrease in transport rate (Fig. 8).

DISCUSSION Apart from the H+,K+-ATPasein the stomach parietal cell, which is responsible for acidification of the gastric c o n t e n t ~ , ( ~ ’ and - ’ ~ )the mitochondrial proton pump, which acts as an ATP synthase,‘zo2L1 a third class of proton pump has been identified, the vacuolar proton pump.(22-24) The gastric pump is an El-E, type of enzyme, sensitive to vanadate and dependent upon K+.“7.L91 It is insensitive to mitochondrial ATPase inhibitors, like oligomycin and azide, sulhydryl reagents, like NEM, and carboxylic acid group inhibitors, like DCCD.‘”) The mitochondria1 pump is inhibited by oligomycin at low concentrations (10 pg/ml or less) and also by DCCD (in the low micromolar range). It is, however, insensitive to vanadate and NEM.(25)The vacuolar pump has been identified in a wide variety of organelles, for example, endosomes,(26-z8) chromaffin granule~,(29-31)lysosomes, ( 3 2 - 3 4 ) plant t o n o p l a s t ~ , ‘ ~and ~~~~) yeast vacuoles.‘”7)The latter type of pump, which is electrogenic,”* 3 4 was also identified as the enzyme responsible for renal urinary acidification. ( 3 9 . 4 0 ) Although it is called vacuolar, this term may prove to be inaccurate. In addition to the kidney and bladder, the osteoclast may also be an exception, since it appears to manifest an electrogenic H’ pump of the “vacuolar” type on its plasma membrane and, more specifically, on the ruffled border, as demonstrated by Akisaka and Gay,(4)Ghiselli et

573

AN ELECTROGENIC PROTON PUMP IN CHICKEN OSTEOCLASTS

24 4

22

LOW

20 18 16 14

-

12

-

.. t

I

L\,

: : / I

,j y-/

vu

4

A I 100

” 0”

-7.0

-6.0

-5.0 -4.0

-3.0

-2.0

-1.0

1 200

I

300

I* 4M)

I 500

Time lsecl

log [MgCI*l FIG. 3. A representative experiment of the ATPase activity in the presence of increasing concentrations of MgCI,. The high- and low-affinity components are indicated and show a K Mof~ 1.293 pM and Vmaxof 0.063 pmol Pi per mg protein per minute and a KMg of 297.6 pM and V,,, of 0.232 pmol Pi per mg protein per minute, respectively, as calculated from Eadie-Hofstee plots of two separate experiments ran in duplicate.

B Blair et a1.,‘lo)Vaananen et al.,[l1)and Bekker and Gay(41) and also in the present report.

I 500

Time lsecl

Mg2’-ATPme activity The vacuolar H+-ATPase activity is usually demonstrated as a Mg”-dependent ATPase.(” ’’ ” ”) H owever, since many enzymes, like Na+,K+-ATPase, Ca”-ATPase, and mitochondrial ATPases, can hydrolyze ATP under these conditions, it is rather nonspecific. For this reason, ouabain, EGTA, and oligomycin were included in the assay media to inhibit these enzymes. In the present study ATPase assays were performed in parallel with proton transport studies. This provides important information concerning the H’ pump, which is responsible for acidification of resorption lacunae. Inhibition of the Mg2+-ATPaseactivity by DCCD in the high micromolar range is typical of a vacuolar H+-ATPase. NEM inhibition is used to distinguish the vacuolar from mitochondrial ATPase. Orthovanadate does not inhibit the Mg2+-ATPasein the micromolar range, which excludes the formation of an acyl phosphate intermediate of the observed Mg”-ATPase enzyme. This sets the enzyme apart from the gastric H+,K+-ATPase, which is highly sensitive to 1 pM o r t h o ~ a n a d a t e . “ ~ . ’Since ~ ) a variety of A T P hydrolyzing enzymes are probably present in the vesicle preparation, partial inhibition by orthovanadate at 1 mM may be explained by inhibition of one of these enzymes. In this

Time lsecl

FIG. 4. Changes in acridine orange fluorescence (A,, = 493 nm, A, = 526 nm) observed after MgATP (2 mM) addition ( - ) and nigericin (1 pM) addition ( I) to show protonophore-mediated proton release. The effect of preincubation for approximately 5 minutes in (A) DCCD (0.2 mM), (B) nolinium bromide (1 mM), and (C) duramycin (80 pg/ml) on transport can be seen.

BEKKER AND GAY

i

‘c. CONTROL

10’0

zw

300

400

0

100

1

I

I

200

300

400

Time lsecl

Time lsecl

Time lsecl

Time lsecl

FIG. 5. The effects of (A) 1 mM Na,V04 and (B) 10 pM DIDS are shown. The points of addition of MgATP, 2 mM ( w ) , and nigericin, 1 pM ( I), are indicated.

regard, it may be mentioned that a second ATPase, which was sensitive to vanadate, was purified from clathrincoated vesicle and chromaffin granule membranes,(43)but these enzymes do not support proton transport. The incomplete inhibition by the agents listed in Table 1 is not uncommon with vesicle preparations and is also best explained by the fact that other enzymes are also present. (13.22.31.44) Inhibition of the Mg2+-ATPaseby nolinium bromide (in the high micromolar range) was surprising. This relatively unknown inhibitor was previously shown to inhibit the stomach H + , K + - A T P ~ S ~and ( ~ ~alveolar .~~) ameloblast H+-ATPase.(*’)The effect of nolinium bromide on other vacuolar H+-ATPases is not known. Therefore, nolinium bromide inhibition may not be an exclusive characteristic of the stomach type of H+-ATPase.DIDS inhibition at 300 pM, but not at 50-100 pM, is important, as is discussed later. Another inhibitor, PCMB, is also a sulfhydry1 group reagent, which was shown by Akisaka and Gay in a previous enzyme histochemical report to inhibit the ruffled border Mg2+-ATPaseactivity.(4)

I/

I

I

500

600

FIG. 6. (A) When NEM (0.5 mM) was present in the incubation medium, a linear decrease in fluorescence was observed prior to MgATP addition. After MgATP addition there was little change in the rate of fluorescence decrease. The points of addition of MgATP, 2 mM (-), and nigericin, 1 pM ( r ) , are indicated. (B) Quenching of acridine orange fluorescence, probably resulting from a photochemical reaction, occurred with addition of 0.5 mM NEM ( y ) to a 1 gM acridine orange solution without vesicles.

Potassium is not essential for Mg2+-ATPase activity, since replacement with sodium did not affect it (Table 2). This is another finding that sets the Mg2+-ATPaseactivity described here apart from the stomach H+,K+-ATPase.Nitrate replacement of chloride caused about a 20% decrease in Mg2+-ATPaseactivity. This observation is also discussed later in relation to the effect of nitrate on proton transport. Acetate substitution of chloride caused an increase in Mg2+-ATPaseactivity. Since no stimulation by acetate was observed in the proton transport assay, acetate probably activates another enzyme(s) in the vesicle preparation that has ATP hydrolyzing activity. The presence of high- and low-affinity components of the Mg2+-ATPaseactivity with regard to MgCI, also supports the idea that more than one ATP hydrolyzing en-

575

AN ELECTROGENIC PROTON PUMP IN CHICKEN OSTEOCLASTS

TABLE3. EFFECTOF INHIBITORS ON PROTON TRANSPORT^

Agent

Concentration

Yo Activity

-

100.0 6.2 11.7 102.6 108.0 32.9

None DCCD NEM Na,VO., Na3V04 Nolinium bromide Nolinium bromide DIDS DIDS DIDS Duramycin Duramycin

0.2 mM 0.5 mM 1 PM 1 mM 0.1 mM 1 mM 10 pM 50 pM 100 pM 40 pg/ml 80 pg/ml

4.5

42.3 12.8 5.1 76.9 47.3

Number of experimenis 2 3 4

3 4 5 2 1 1

2 2

aData are expressed as percentages of the initial transport rate of controls, which contain no inhibitors.

zyme is detected with the ATPase assay. Gluck et aI.(39'reported a K M of ~ 700 pM for turtle bladder H'-ATPase, and Schneider determined a K M of~ 200 pM using a rat liver lysosomal preparation.'32' All nucleotides tested were hydrolyzed effectively (Table 5 ) . However, since proton transport occurred far more effectively with ATP than with other nucleotides (see Table 5 ) , the nonspecificity of the enzymatic activity is probably due to the presence of other enzymes in the preparation. This discrepancy is commonly observed with vesicle preparations. (33.481 The increase in Mg2+-ATPaseactivity seen with the protonophores, nigericin and CCCP, substantiates two imporA tant points: first, that inside-out sealed vesicles were present in the membrane preparation (which was also confirmed with the SDS latency studies performed previTime lsecl o ~ s I y ( ' ~and ) ) , second, that proton transport occurred with the addition of ATP to the membrane preparation. The argument here is that the protonophores permeabilize the U A vesicle membrane for protons and prevent charge accumuKND, Pr '.,..'~.,~~,~,~~~~,"/,:~,~~~~.,~ 1r-M lation inside the vesicles. This leads to an increased enzyme i, turnover.

H

/'

m

I

?i

k

KCI

-

orange is a membrane-permeable weak base that follows I

I

I

I

I

1

BEKKER AND GAY

576

TABLE 4. PROTON TRANSPORT IN THE PRESENCE OF DIFFERENT IONSTHAT REPLACED K' OR Cl-a % Activity (number

Substitution Control Na' for K' NO,- for CINO,- PIUS CIAcetate and SO,'- for CI-

Ionic composition

of experiments)

75 mM KCI, 3 mM MgCI, 75 mM NaCI, 3 mM MgC1, 75 mM KNO,, 3 mM Mg(N0J2 37.5 mM KCI, 37.5 mM KNO,, 3 mM MgCL 75 mM K acetate, 3 mM MgSO,

100.0 179.2 (3) 53.9 (2) 48.4 (2)

44.3 (3)

=Resultsare expressed as percentage activity compared to initial transport rate in the control medium, which contains KC1 and MgCI,.

TABLE5 . COMPARISON OF THE RELATIVE EFFECTIVENESS OF HYDROLYSIS OF VARIOUS SUBSTRATES AS WELLAS THEIRABILITY TO SUPPORT PROTONTRANSPORT^ Substrate

activity

Number of experiments

% Hydrolyzing activity

Number of experiments

Na2ATP GTP ITP CTP ADP

100 8.4 22.1 27.1 1.5

4 2 3 2 2

100.0 112.8 116.4 99.4 122.6

2 2 2 2 2

% Transport

aResults are expressed as percentage activity compared to activity observed with Na,ATP as substrate. Calculations for transport activity were made based on initial transport rates.

lease of accumulated protons from the vesicles and acri- have a K+-sensitivesite like the gastric H+,K+-ATPase,and dine orange is deprotonated and leaves the vesicles, with a therefore the mechanism of inhibition of proton transport resultant increase in the fluorescent green signal. in the osteoclast vesicles is not clear at present. Relatively high concentrations of DCCD were necessary Duramycin is an antibiotic that was shown to inhibit the to completely inhibit proton transport (more than 0.2 clathrin-coated vesicular H' pump, and this inhibition was mM), whereas the mitochondria1 proton pump is blocked mentioned as a possible discriminative characteristic of the by micromolar concentrations. This substantiates the pres- endosomal and the lysosomal H' pump,'"' the latter being ence of a vacuolar proton pump in the vesicles. DCCD acts insensitive to duramycin. on the proton pore of the H' transport complex and also Nevertheless, duramycin inhibited transport only parhas a direct inhibitory effect on the h o l o e n ~ y m e . ' ~ ~ ~ ~ ' )tially. In the previous histochemical report from this laboNolinium bromide, or [2-(3,4-dichlorophenyl amino)- r a t o ~ y , ' duramycin ~) was also shown to cause inhibition of quinolizium bromide], inhibits histamine-stimulated the Mg2+-ATPase situated on the ruffled border. Addigastric acid secretion by bullfrog gastric mucosa.L451 In the tional fluorescence microscopic evidence by Hunter et al. latter report, it was shown that nolinium bromide was ef- indicated that duramycin caused a decrease in intracellular fective from both the secretory and the nutrient side of an osteoclastic acidity. ( 5 2 isolated chambered mucosal patch. The mechanism of inOrthovanadate does not affect H' transport, even at 1 hibition, when nolinium bromide was applied to the secre- mM. This provides strong evidence that the proton pump tory (luminal) side of the gastric mucosa, was thought to described here is not an E,-E, enzyme, a finding corrobobe interference at a high-affinity potassium-sensitive site. rated by other studies.'4.10,11' Binding of potassium to this site enhanced ATP hydrolysis The strong inhibition (87.2%) of proton transport obby the pump. Nolinium bromide inhibited proton trans- served with DIDS at 50 pM provides supportive evidence port 50 and 80% at 0.5 and 1 mM, respectively, in this re- for the presence of an anion transporter in the same vesicle port. The mechanism of inhibition from the nutrient side is membranes as the H' pump, because at this concentration unknown. there was no inhibition of H+-ATPaseactivity (see Fig. 2). Our data showed that nolinium bromide inhibited trans- At higher concentrations of DIDS (0.3 mM), the H+port partially at 0.1 mM and almost completely at 1 mM. ATPase is also inhibited, since DIDS was shown to act diHowever, the vacuolar proton pump does not seem to rectly on the e n ~ y m e . ' l ~ .It' *should ~ be mentioned that this

AN ELECTROGENIC PROTON PUMP IN CHICKEN OSTEOCLASTS

fl

.-ln> 4-

C

.-C 4-

is

i

ii

a

0

Time lsecl

FIG. 8. The potassium ionophore valinomycin (2.5 pM) was added after transport was initiated by MgATP, and the transport rate decreased 60% based on slope calculations before and after valinomycin addition. The points of addition of MgATP, 2 mM ( w ) , valinomycin, 2.5 pM ( T ) , and nigericin, 1 yM ( r ) , are indicated.

anion transporter is not necessarily a separate entity from the H' p ~ m p . ' ~ ~ , ~ " ) Although fluorescence measurements in the presence of NEM are somewhat complicated by the photochemical reaction between NEM and acridine orange, the rates of decrease in the fluorescent signal, before and after ATP addition, were almost identical (Fig. 6A). It is therefore concluded that NEM is a strong inhibitor of proton transport, a finding reported by several other laboratories. ( 9 ~ 1 0 . 1 3 ~ 2 6 . 2 7 . 3 1 . s 4 1 The finding that Na' replacement of K' led to an increase in the initial rate of H' transport after MgATP addition is not understood. Further investigation is currently underway. The possibility of an Na+,H' exchanger in the vesicle membrane must be considered. When C1- was replaced by nitrate or acetate, transport decreased substantially. This may indicate that the transporter, which is responsible for anion cotransport, does not transport nitrate and acetate as effectively as chloride. On the other hand, both nitrate and acetate are believed to be membrane permeable. (551 Therefore, another explanation for the relative ineffectiveness of nitrate and acetate as substitutes of chloride may be that the anion transporter is specific for chloride and that lipid permeation of nitrate and acetate is relatively slow compared to the facilitated diffusion of chloride. These results also substantiate the idea that the H' transport described here is electrogenic. The lower transport observed in the presence of nitrate appears to be due to a direct inhibitory effect of nitrate on the H' pump, since the Mg"-ATPase activity is partially inhibited by nitrate (Table 2) and since transport is substantially decreased when both C1- and NO,- are present (Table 4). This finding indicates that, although C1- was present to collapse the membrane potential generated by accumulation of protons intravesicularly, nitrate prevented

577

the optimal functioning of the H' pump by a direct inhibitory action. This direct inhibitory effect of nitrate has been reported by other investigator^.'^^ 53' It is worth mentioning that nitrate seems to exert its effect at an intravesicular site on the proton pump.(s3)It must therefore penetrate the membrane to reach this sensitive site on the enzyme. Although ATP functions as the preferred substrate of the proton pump, it can also utilize CTP and ITP to a small extent and, to a much lesser extent GTP, as substrates. ADP did not support transport (Table 5). This finding is not in complete agreement with recent results published by Blair et al.(*Os61 We do not have an explanation, except that in the present study fluorescence was the basis for transport measurement but Blair et al. used acridine orange absorbance measurements. An increase in proton transport rate in the presence of the K' ionophore valinomycin usually supports the idea of electrogenicity of the proton pump; that is, continual enzyme turnover results in the development of a positive intravesicular membrane potential. When vesicles are loaded with K' before transport experiments are performed, the addition of valinomycin, before or after ATP-initiated H' transport leads to an increase in transport rate."' l9 s 4 ) However, if K' preloading is not performed but K' is present extravesicularly, valinomycin treatment after ATP addition should decrease the proton transport rate. The reason is that K' influx occurs with valinomycin addition to the vesicle suspension and the positive charge carried by potassium inhibits H' transport. Our finding that valinomycin addition slowed down H+ transport is therefore in line with the latter argument. The vesicle preparation used in this study was prepared in the absence of K', and K' was only added just prior to transport assay. Also, the argument is only valid if K' membrane permeability is low, as shown in several reports.(2834 40 491 Lee and coworkers also showed that valinomycin causes a decrease in H' transport if vesicles are not K' loaded.(571 Although the vacuolar proton pump described here is manifested on the ruffled border membrane of osteoclasts, it is believed that its immediate origin is from the numerous small vesicles beneath the ruffled border in active osteoclasts and scattered throughout the cytoplasm in "resting" osteoclasts. ( 5 8 1 These small vesicles in resting or detached osteoclasts may serve as a reservoir for the ruffled border membrane and also for proton pumps.(") An alternative explanation is that the ruffled border H' pump is supplied by lysosomes, since it was shown by Baron et al. that a 100 kD lysosomal protein shared antigenicity with a ruffled border protein."' There seems to be close similarity among the various kinds of vacuolar H' pumps. It is therefore difficult at this stage to be decisive, and further investigations are required. The absolute dependence of H' transport on C1- found recently by Blair et a1.(s6)is in contrast to our observations. Since lipid membranes are usually not impermeable to anions like acetate and nitrate,(5s1partial transport should be observed with these anions even in the absence of an anion transporter. Absolute dependence on chloride rather suggests the presence of a chloride-sensitive site on

BEKKER AND GAY

578

the H+-ATPase itself, as observed with the purified chromaffin granule A T P ~ s ~ ' ~and ~ . ' the ~ ' synaptosomal en~yme.'~ We ~ )cannot exclude the possibility of an anionsensitive site@); however, based on our studies with DIDS, an anion transporter is probably also present in the ruffled border membrane. The absence of immunoreactivity on the plasma membrane of osteoclasts, using rabbit antiserum to a synthetic peptide corresponding to the COOH terminus of murine band 3 protein (anion exchanger), was recently reported by Kellokumpu et a1.I6O' This is in contrast to findings reported by Teti et al.,(61'which indicated the presence of a chloride-bicarbonate exchanger on the plasma membrane of osteoclasts. This issue deserves more investigation before the mechanisms involved in bone resorption can be understood.

ACKNOWLEDGMENTS This work was presented at the ASCB meeting in November 1989. We thank Dr. David J. Hurley for assistance with the spectrofluorometer, Nancy L. Kief and Virginia R. Gilman for assistance in the cell isolation procedure, and also Dr. Osamu Fukushima for invaluable discussions. This work was supported by NIH Grant DE04345 and NASA Grant NAGW-1196 to the Center for Cell Research at PSU.

REFERENCES 1. Fallon MD 1984 Bone resorbing fluid from osteoclasts is

2.

3.

4.

5.

6.

7.

8.

acidic-an in vitro micropuncture study. In: Cohn DV, Fujita T, Potts JT Jr, Talmage RV (eds) Endocrine Control of Bone and Calcium Metabolism, Vol. 8A. Elsevier Science Publishers, Amsterdam, pp. 144-146. Silver IA, Murrils RJ, Etherington DJ 1988 Microelectrode studies on the acid microenvironment beneath adherent macrophages and osteoclasts. Exp Cell Res 175:266-276. Baron R, Neff L, Louvard D, Courtoy PJ 1985 Cell-mediated extracellular acidification and bone resorption: Evidence for a low pH in resorption lacunae and localization of a 100-kD lysosomal membrane protein at the osteoclast ruffled border. J Cell Biol 101:2210-2222. Akisaka T, Gay CV 1986 Ultracytochemical evidence for a proton-pump adenosine triphosphatase in chick osteoclasts. Cell Tissue Res 245507-512. Tuukkanen J, Vaananen HK 1986 Omeprazole, a specific inhibitor of H+,K+-ATPase,inhibits bone resorption in vitro. Calcif Tissue Int 38:123-125. Anderson RE, Woodbury DM, Jee WSS 1986 Humoral and ionic regulation of osteoclast acidity. Calcif Tissue Int 39: 252-258. Lorentzon P, Jackson R, Wallmark B, Sachs G 1987 Inhibition of (H' + K+)-ATPase by omeprazole in isolated gastric vesicles requires proton transport. Biochim Biophys Acta 897:41-5 1. Im WB, Sih JC, Blakeman DP, McGrath J P 1985 Omeprazole, a specific inhibitor of gastric (H'-K+)-ATPase, is a H+-

activated oxidizing agent of sulfhydryl groups. J Biol Chem 260:2591-2597. 9. Ghiselli R, Blair H, Teitelbaum S, Cluck S 1987 Identification of the osteoclast proton pump (abstract). J Bone Min Res 2(Suppl. 1):275. 10. Blair HC, Teitelbaum SL, Ghiselli R, Cluck S 1989 Osteoclastic bone resorption by a polarized vacuolar proton pump. Science 245:855-857. 1 1 . Vaananen HK, Roininen I, Parvinen E-K, Sundquist K, Hentunen T, Lakkakorpi P, Tuukkanen J, Wallmark B 1989 Localization and characterization of vacuolar ATPase in osteoclasts. J Bone Min Res 4(Suppl. l):S213. 12. Bekker PJ, Gay CV 1989 Characterization of a Ca2+-ATPase in osteoclast plasma membrane. J Bone Min Res 5557-567. 13. Sallman AL, Lubansky HJ, Taylor Z, Arruda JAL 1986 Plasma membrane proton ATPase from human kidney. Eur J Biochem 157:547-551. 14. Bekker PJ, Gay CV 1988 The plasma membrane Ca"ATPase of chicken osteoclasts is a calcium pump. J Cell Biol 107:129a. 15. Fiske CH, Subbarow Y 1925 The colorimetric determination of phosphorus. J Biol Chem 66:375-400. 16. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ 1951 Protein measurement with the fohn phenol reagent. J Biol Chem 193:265-275. 17. Ganser AL, Forte JG 1973 K'-stimulated ATPase in purified microsomes of bullfrog oxyntic cells. Biochim Biophys Acta 307: 169-190, 18. Sachs G, Faller LD, Rabon E 1982 Proton/hydroxyl transport in gastric and intestinal epithelia. J Membr Biol 64:123135, 19. Soumarmon A, Lewin MJM 1986 Gastric (H+,K+)-ATPase. Biochimie 68:1287-1291. 20. Cross RL 1981 The mechanism and regulation of ATP synthesis by F,-ATPases. Annu Rev Biochem 50:681-714. 21. Futai M, Kanazawa H 1983 Structure and function of proton-translocating adenosine triphosphatase (FoF,):Biochemical and molecular biological approaches. Microbiol Rev 47: 285-3 12. 22. Schneider DL 1987 The proton pump ATPase of lysosomes and related organelles of the vacuolar apparatus. Biochim Biophys Acta 8951-10. 23. Nelson N 1987 The vacuolar proton-ATPase of eukaryotic cells. Bioessays 7:251-254. 24. Njus D, Kelley PM, Harnadek GJ 1986 Bioenergetics of secretory vesicles. Biochim Biophys Acta 853:237-265. 25. Mellman I, Fuchs R, Helenius A 1986 Acidification of endocytic and exocytic pathways. Annu Rev Biochem 59663-700. 26. Stone DK, Xie X-S, Racker E 1983 An ATP-driven proton pump in clathrin-coated vesicles. J Biol Chem 258:40594062. 27. Galloway CJ, Dean GE, Marsh M, Rudnick G, Mellman I 1983 Acidification of macrophage and fibroblast endocytic vesicles in vitro. Proc Natl Acad Sci USA 803334-3338. 28. Xie X-S, Stone DK, Racker E 1983 Determinants of clathrincoated vesicle acidification. J Biol Chem 258:14834-14838. 29. Cidon S, Nelson N 1983 A novel ATPase in the chromaffin granule membrane. J Biol Chem 258:2892-2898. 30. Dean GE, Nelson PJ, Rudnick G 1986 Characterization of native and reconstituted hydrogen ion pumping adenosinetriphosphatase of chromaffin granules. Biochemistry 2349184925. 31. Cidon S, Nelson N 1986 Purification of N-ethylmaleimidesensitive ATPase from chromaffin granule membranes. J Biol Chem 261:9222-9227. 32. Schneider DL 1977 Membranous localization and properties

AN ELECTROGENIC PROTON PUMP IN CHICKEN OSTEOCLASTS of ATPase of rat liver lysosomes. J Membr Biol34247-261. 33. Schneider DL 1981 ATP-dependent acidification of intact and disrupted lysosomes. J Biol Chem 256:3858-3864. 34. Ohkuma S, Moriyama Y, Takano T 1982 Identification and characterization of a proton pump on lysosomes by fluorescein isothiocyanate-dextran fluorescence. Proc Natl Acad Sci USA 79:2758-2762. 35. Manolson MF, Rea PA, Poole RJ 1985 Identification of 30-(4-benzoyl)benzoyladenosine 5'-triphosphate- and N,N'-dicyclohexylcarbodiimide-binding subunits of a higher plant H'-translocating tonoplast ATPase. J Biol Chem 260:1227312279. 36. Randall SK, Sze H 1986 Properties of the partially purified tonoplast H'-pumping ATPase from oat roots. J Biol Chem 261~1364-1371. 37. Uchida E, Ohsumi Y, Anraku Y 1985 Purification and properties of H'-translocating, Mg"-adenosine triphosphatase from vacuolar membranes of Saccharoniyces cerevisiae. J . Biol Chem 260:1090-1095. 38. Cidon S , Ben-David H, Nelson N 1983 ATP-driven proton fluxes across membranes of secretory organelles. J Biol Chem 258:11684-11688. 39. Cluck S, Kelly S, Al-Awqati Q 1982 The proton translocating ATPase responsible for urinary acidification. J Biol Chem 257:9230-923 3. 40. Youmans SJ, Worman HJ, Brodsky WA 1983 ATPase activity and ATP-dependent proton translocation in plasma membrane vesicles of turtle bladder epithelial cells. Biochim Biophys Acta 730:173-177. 41. Bekker PJ, Gay CV 1989 Characterization of a proton pump in chicken osteoclast plasma membrane vesicles. J Cell Biol 10942a. 42. Xie X-S,Stone DK, Racker E 1989 Purification of a vanadate-sensitive ATPase from clathrin-coated vesicles of bovine brain. J Biol Chem 264:1710-1714. 43. Apps DK, Percy JM 1987 The H+-translocating ATPase of chromaffin granule membranes. Ann NY Acad Sci 493:178188. 44. Lubansky HJ, Arruda JAL 1985 Plasma membrane protonATPase of a turtle bladder epithelial cell line. J Biol Chem 260~4035-4040. 45. Nandi J, Wright MV, Ray TK 1983 Mechanism of gastric antisecretory effects of nolinium bromide. Gastroenterology 85~938-945. 46. Fujimoto K, Ogawa KS, Ogawa K 1986 Gastric K+-stimulated p-nitrophenylphosphatase cytochemistry. Histochemistry 84:600-608. 47. Sasaki T, Tadokoro K, Yanagisawa T, Higashi S, Garant PR 1988 H+-K+-ATPaseactivity in the rat incisor enamel organ during enamel formation. Anat Rec 221:823-833. 48. Turrini F, SaboliC I, Zimolo 2, Moewes B, Burckhardt G 1989 Relation of ATPases in rat renal brush-border rnembranes to ATP-driven H' secretion. J Membr Biol 107:l-12. 49. Kaunitz JD, Gunther RD, Sachs G 1985 Characterization of

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

579

an electrogenic ATP and chloride-dependent proton translocating pump from rat renal medulla. J Biol Chem 260:1156711573. Sun S-2, Xie X-S, Stone DK 1987 Isolation and reconstitution of the dicyclohexylcarbodiimide-sensitiveproton pore of the clathrin-coated vesicle proton translocating complex. J Biol Chem 262:14790-14794. Arai H, Berne M, Forgac M 1987 Inhibition of the coated vesicle proton pump and labeling of a 17,000-dalton polypeptide by N,N'-dicyclohexylcarbodiimide. J Biol Chem 262: 11006-11011. Hunter SJ, Gay CV, Kief NL 1988 The effects of duramycin, ouabain and colchicine on intracellular acidity of isolated osteoclasts. J Bone Min Res 3(Suppl. I):S175. Moriyama Y, Nelson N 1987 The purified ATPase from chromaffin granule membranes is an anion-dependent proton pump. J Biol Chem 262:9175-9180. Cluck S, Al-Awqati Q 1984 An electrogenic proton-translocating adenosine triphosphatase from bovine kidney medulla. J Clin Invest 73:1704-1710. Smallwood JI, Waisman DM, Lafreniere D, Rasmussen H 1983 Evidence that the erythrocyte calcium pump catalyzes a Caz+:nH+exchange. J Biol Chem 258:11092-11097. Blair H, Koziol C, Mead R, Cluck S, Teitelbaum S , Schlesinger P 1989 Osteoclast ruffled membranes contain proton pump that is tightly coupled to anion (CI-) permeability. J Bone Min Res 4(Suppl 1):S263. Lee HC, Breitbart H, Forte JG 1980 The functional role of K+-ATPase in proton transport by gastric microsomal vesicles. Ann NY Acad Sci 341:297-311. Fukushima 0, Bekker PJ, Gay CV 1989 An enzyme histochemical investigation of osteoclasts and their related cells in chick tibia metaphyses. Anat Rec submitted. Moriyama Y , Nelson N 1987 Internal anion binding site and membrane potential dominate the regulation of proton pumping by the chromaffin granule ATPase. Biochem Biophys Res Commun 149:140-144. Kellokumpu S, Neff L, Kellokumpu-Jamsa S, Kopito R, Baron R 1988 A 115-kD polypeptide immunologically related to erythrocyte band 3 is present in Golgi membranes. Science 242: 1308-1 3 11. Teti A, Blair H, Kahn A, Konsek J, Koziol C, ZamboninZallone A, Teitelbaum S, Schlesinger P 1987 Intracellular pH regulation of isolated osteoclasts by chloride/bicarbonate exchange. J Bone Min Res 2(Suppl. 1): 276a.

Address reprint requests to: Dr. Carol V. Gay 468A North Frear Laboratory Pennsylvania State University University Park, PA 16802 Received for publication September 14, 1989; in revised form January 18, 1990: accepted January 20, 1990.