Inhibition of (Na,K)-ATPase by tetravalent vanadium.

Vanadyl, the tetravalent state of vanadium and a divalent cation, VO2+, was a relatively powerful inhibitor of highly purified membrane-bound sodium and potassium ion transport adenosine triphosphatase. The sensitivity of the ATPase activity to vanadyl characteristically correlated positively with the specific activity of the enzyme preparation. Inhibition ranged from nearly complete inhibition at less than 5 microM vanadyl for some of the purest fractions (specific activity approximately 45 mumol/min/mg of protein) to no observable inhibition at 300 microM vanadyl in one crude preparation of the enzyme with a specific activity of 10 mumol/min/mg of protein. The level of free vanadyl was reduced by incubation with these membranes, but this reduction was not sufficient to account for the low sensitivity to vanadyl observed in crude preparations. A reduction in specific activity by partial inactivation of a sensitive preparation by treatment with FeCl3 and ascorbate reduced its sensitivity to vanadyl. Anionic ligands of the enzyme, vanadate or ATP, increased the rate of recovery from inhibition after chelation of free vanadyl. At pH 6.1, the inhibition was characteristically fully reversible (t1/2 approximately 10 min), whereas at pH 8.1 it was stable for hours. The degree and stability of enzyme inhibition by vanadyl increased for several hours during incubation of the vanadyl-enzyme mixture, and at pH 6.1 the properties of the inhibitor itself also changed with time. Preincubation of the ion at that pH for 5 h before addition of the enzyme produced a more stable inhibition. The time- and pH-dependent changes in the degree and stability of enzyme inhibition probably relate to the complex chemistry of the vanadyl ion in solution.

Inhibition of (Na,K)-ATPase by Tetravalent Vanadium* (Received for publication, December 8, 1983) Paula North$ and Robert L. Post8 From the Department of Physwlogy, Vanderbilt University Medical School, Nashville, Tennessee 37232 Vanadyl, the tetravalent state of vanadium and a divalent cation, V02+, was a relatively powerful inhibitor of highly purified membrane-bound sodium and potassium ion transport adenosine triphosphatase. The sensitivity of the ATPase activity to vanadyl characteristically correlated positively with the specific activity of the enzyme preparation. Inhibition ranged from nearly complete inhibition at Iess than 5 PM vanadyl for some of the purest fractions (specific activity -45 Pmol/min/mg of protein) to no observable inhibition at 300 NM vanadyl in one crude preparation of the enzyme with a specific activity of 10 pmol/min/mg of protein. The level of free vanadyl was reduced by incubation with these membranes, but this reduction was not sufficient to account for the low sensitivity to vanadyl observed in crude preparations. A reduction in specific activity by partial inactivation of a sensitive preparation by treatment with FeCls and ascorbate reduced its sensitivity to vanadyl.
Anionic ligands of the enzyme, vanadate or ATP, increased the rate of recovery from inhibition after chelation of free vanadyl. At pH 6.1, the inhibition was characteristically fully reversible (tU -10 min), whereas at pH 8.1 it was stable for hours. The degree and stability of enzyme inhibition by vanadyl increased for several hours during incubation of the vanadylenzyme mixture, and at pH 6.1 the properties of the inhibitor itself also changed with time. Preincubation of the ion at that pH for 5 h before addition of the enzyme produced a more stable inhibition. The timeand pH-dependent changes in the degree and stability of enzyme inhibition probably relate to the complex chemistry of the vanadyl ion in solution.
The potential of vanadyl ion, the tetravalent state of vanadium, as a regulator of the activity of (Na,K)-ATPase' has only recently been suggested. Vanadate ion, containing vanadium in the pentavalent oxidation state, is a potent inhibitor of the ATPase ( K I -0.5 MM), with striking pharmacologic effects on cardiac contractility, blood pressure, and urine flow (1). In addition, externally applied vanadate has insulin-like *This work was supported by Grant 5 R01 HL-01974 from the National Heart, Lung, and Blood Institute of the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Research fellow of the American Heart Association, Tennessee Affiliate during part of this work.
8 To whom correspondence should be addressed. effects on glucose transport (2), glucose oxidation (3), and on glycogen synthetase and phosphorylation of the insulin receptor (4). However, Cantley and co-workers (5, 6) have shown that vanadate is reduced in red cells to vanadyl ion by intracellular glutathione following uptake through an anion transport system. Although free vanadyl ion is oxidized to vanadate in a few minutes in aqueous solution exposed to air at neutral pH, the tetravalent state appears to be stable when complexed with intracellular proteins, such as hemoglobin (5, 6) or smaller molecules (3, 7). The difficulties inherent in preventing oxidation of vanadyl at neutral pH have hampered study of the biochemical actions of vanadyl in uitro. Based on transient kinetics of vanadyl interaction with (Na,K)-ATPase, Cantley and Aisen (5) concluded that vanadyl is at most a less potent inhibitor of the enzyme than vanadate. Since vanadyl, rather than vanadate, is the major form of intracellular vanadium, we have taken a closer look at its interaction with (Na,K)-ATPase using an anaerobic assay system that avoids inhibition of the enzyme by contaminating vanadate.

EXPERIMENTAL PROCEDURES
Materials-Vanadyl sulfate (vanadium oxysulfate) was obtained from Alfa Inorganics, Inc. (Beverly, MA), and ammonium vanadate was from Fisher. VOSO, and NHIVOB solutions were prepared at 10 mM concentration in 3 mM HCl or 1 mM Tris, respectively, and stored at room temperature. All buffers were from Sigma. Disodium ATP was from Boehringer Mannheim.
Preparation and Assay of the Enzyme-Membrane-bound (Na,K)-ATPase was purified from the outer medulla of hog kidney by the zonal gradient centrifugation method of Jorgensen (8). Ten to fifteen fractions were collected from the gradient and stored at 0 "C in 30% glycerol (w/v), 20 mM imidazole HCl (pH 7.5), 0.7 mM EDTA, and 0.7 mM dithioerythritol; in some cases the enzyme was washed free of EDTA and dithioerythritol by centrifugation before use. The ATPase activity was stable for more than 6 months and was 99.9% inhibitable by ouabain. The specific activity of the collected fractions ranged from 10 to 44 pmol of Pi/min/mg of protein at 37 "C in 30 mM imidazole/glycylglycine (pH 7.4), 100 mM NaC1, 25 mM KCl, 4 mM Na2ATP, 3.9 mM MgC12, 0.2 mM EDTA, and 0.2 mg/ml of fatfree bovine serum albumin. Activity in the presence of 0.33 mM ouabain was subtracted from that in its absence. All (Na,K)-ATPase activities reported are in units of micromoles of Pi released per min under these conditions. The protein concentration of the (Na,K)-ATPase was estimated by the method of Bradford (9) with bovine serum albumin as a standard. The values were multiplied by 1.144 to correct for the ratio of the color factors of the two proteins as estimated in this laboratory by amino acid analysis.' Inorganic phosphate split from ATP in the ATPase assay was estimated on a Technicon autoanalyzer by the method of Hegyvary et al. (10).
Membrane-bound (Na,K)-ATPase was prepared from the outer medulla of dog or guinea pig kidney by two modifications of the rapid method of Jorgensen (8). The two procedures differed primarily in their method of dissection and SDS/protein ratio. The first used visual dissection of the outer medulla based on its red color and an SDSlprotein ratio (w/w) of 0.32-0.37. It produced an enzyme with a specific activity of 10-20 units/mg of protein. In the second method, the dissection was done by scraping away the cortex from the firmer medulla, and the SDS/protein ratio was 0.43. This improved method yielded a (Na,K)-ATPase specific activity of 20-40 units/mg of protein. ATPase activity was measured as described above, and the enzyme was stored at 0 "C in 30% glycerol containing either 24 mM imidazole, 16 mM HEPES, 0.5 mM EDTA, and 2 mM dithioerythritol (pH 7.0) or 15 mM imidazole/MOPS and 0.1 mM EDTA (pH 7.1). All membrane preparations were stored at a protein concentration of 2.2-5.5 mg/ml.
Estimation of Vanudyl Inhibition of (Na,K)-ATPase-To minimize oxidation of vanadyl to vanadate, which occurs in air above pH 3, Warburg flasks containing a buffered suspension of the enzyme in the body of the flask and separate solutions of VOS04 and EDTA in two side arms were equilibrated under prepurified NP at 37 "C for 1 h before mixing of the components. The solution of VOSO, in 1 mM HC1 (0.4 ml) was then tipped into the enzyme suspension (2.0 ml) to start the reaction. Subsequently the vanadyl ion free in solution was chelated by tipping in the EDTA solution (0.4 ml). After 1 min or more the flask was opened and a solution of Mg/ATP containing bovine serum albumin (0.4 ml) was added to start the ATPase reaction. In this way the recovery from vanadyl inhibition could be observed. Aliquots (0.5 ml) were removed at five successive time intervals after ATP addition for determination of inorganic phosphate released by means of the automated assay described above. Less than 20% of the ATP was split. ATPase activity in a duplicate flask containing 0.33 mM ouabain was subtracted from the activity in its absence to estimate the (Na,K)-ATPase activity. The reaction medium was buffered with PIPES/Tris at pH 6.1 or 6.5 or with Tris/ glycylglycine at pH 8.1 and contained 160 mM NaCI, 0.8 mM KCI, 5 mM Na2ATP, 4 mM MgC12, 0.3-1.0 mM EDTA, 0.1 mg/ml of fat-free bovine serum albumin, and 0-400 p~ voso4. These ionic conditions were chosen to minimize possible inhibition of the enzyme by vanadate formed by inadvertent oxidation of vanadyl ion (see under "Results"). The coefficient of variation of enzymatic activity in replicate flasks estimated under these conditions was 4.0%. Unless otherwise stated, the Na+ was present initially with the enzyme, and the K+ was added together with the Mg/ATP. However, in some cases NazATP or one or more of the ions and ammonium vanadate (NH4V03) were added with or shortly following the EDTA in order to assess their effect on the rate of recovery from vanadyl inhibition (cf. figure legends). Complete chelation of free vanadyl by EDTA was verified by demonstrating that addition of the EDTA to the enzyme suspension before addition of the vanadyl completely prevented inhibition. At low pH (pH 6.1) EDTA was a relatively effective chelator also of vanadate, completely preventing vanadate inhibition of the enzyme (measured in the presence of 80 mM NaCl and 80 mM KCI) at EDTA/vanadate ratios above 25. However, chelation of vanadate by EDTA could be made insignificant when necessary (for instance when the effect of vanadate on the vanadyl recovery rate was measured) by increasing the pH to 6.5 and lowering the EDTA/vanadate ratio to 3.
The data for the inhibited samples in most cases were normalized against control data in order to provide a format that was independent of the rate in the control samples and that corrected the test data for an occasional small curvature in the control activity curve. The concentrations of released Pi from the control flasks were fitted to an equation of the form [Pi] = at + bt', where a and b are fitted constants and t is time. For each concentration of Pi released by the inhibited enzyme a time was calculated from the inverse of the equation. This time was that which would have been required for the control enzyme to release the same amount of Pi. The ratio of this time to the actual time was taken as the ratio of Pi released in the inhibited sample to that in the control. The control data are represented on a graph by a straight line, and the data from the inhibited samples are plotted as a fraction of these normalized control values. Data points without normalization are shown in Figs. 4A, 6, and 8.
To estimate the initial inhibition present at the time of ATP addition it was assumed that a fraction, F, of the enzyme was inhibited at zero time and that this inhibition subsequently disappeared as a single exponential function of time, t, with a rate constant, k, i.e.
[Pi]/E = t -F(le-")/k, where E is the activity of the uninhibited enzyme. Trial values were inserted for F and k, and the error was computed as the sum of the square of each difference between a data point and its computed value divided by the value of the data point.
Then the values of F and k were adjusted repeatedly by successive approximations until the minimum error was found. Since two constants were estimated from five data points and since the validity of the assumptions could not be tested independently, this method of data analysis is no more than a heuristic device.
Inactivation by Ascorbate and FeCL-The membrane-bound (Na,K)-ATPase was inactivated by a modification of the method used by Levine (11) for glutamine synthetase. Purified enzyme (0.5 mg of protein/ml; 40 units/mg of protein) was incubated at 37 "C in a 30/ 12 mM imidazole/MOPS buffer (pH 7.2) containing freshly prepared 15 mM ascorbate (neutralized with Tris) and 0.1 mM FeC13 in a final volume of 1.0 ml. After 70 or 240 min of incubation, 4-p1 aliquots were removed (in quadruplicate) for assay of (Na,K)-ATPase activity under standard conditions as described above, and the remainder of the mixture was diluted with 9 volumes of cold 20/5 mM imidazole/ MOPS, 1 mM EDTA (pH 7.4) to stop the reaction. The diluted enzyme was centrifuged at 100,000 X g. , overnight in a Beckman Ty 65 rotor at 2 "C, resuspended in 0.25 ml of 15/15 mM imidazole/ MOPS, 0.1 mM EDTA, 30% (w/v) glycerol (pH 7.1) and stored at 0-4 "C. The resuspended pellet was assayed for protein concentration and (Na,K)-ATPase activity as described earlier. A control sample was treated in the same way, but in the absence of ascorbate and FeCI3, with less than 7% loss of specific activity. The samples treated with ascorbate and FeC4 for 70 or 240 min lost 65 or 99.1% of (Na,K)-ATPase activity (per mg of protein), respectively. The degree of inactivation was stable for at least 2 weeks after dilution with the imidazole/MOPS/EDTA buffer.
Measurement of the Rate of Vanudyl Oxidation-Oxidation of the blue vanadyl ion at pH 6.1 or 8.1 was monitored by observing the disappearance of the characteristic vanadyl absorbance at 766 nm. The extinction coefficient at this wavelength was approximately 17.6 M" cm". It was not possible to measure the vanadyl absorbance directly during incubation at these pH values since a cloudy precipitate of vanadyl hydroxide appeared. Therefore, aliquots taken at various time intervals after addition of voso4 (in 3 mM HCI) to the buffer (PIPES/Tris at pH 6.1 or Tris/glycylglycine at pH 8.1) were first acidified to pH 2 by addition of HzS04 to stabilize the tetra-and pentavalent oxidation states in their cationic forms (V02+ and VO: , respectively).
Conversion of vanadyl to vanadate during incubation under air or a nitrogen atmosphere was also monitored by separating the tetravalent (V02+) and pentavalent (VO:) species as they exist in acid (pH 5 2) after various incubation times at pH 6.1. This was done by ion exchange chromatography following acidification of the samples to pH 2 with H2S04. After a brief centrifugation to remove the PIPES buffer, which precipitated at pH 2, the ions were applied to a Dowex AG 50W-X8 column pre-equilibrated with 10 mM H,SO,, and the vanadate and vanadyl were eluted serially with 0.6 and 3 N H2S04, respectively. The eluted peaks of the ions were detected using a flowcell spectrophotometer measuring absorbance at 280 nm. The relative proportions of each ion were then estimated based on the approximate extinction coefficients of vanadate (260 M" cm") and vanadyl (100 M-' cm") at 280 nm.
Estimation of Vanadyl Binding-Total binding of vanadyl to the membrane-bound enzyme preparation was estimated in an indirect manner by first allowing vanadyl to complex with the preparation and then separating free and bound vanadyl by filtration. Free vanadyl in the filtrate was oxidized quantitatively to vanadate, and the resulting concentration of vanadate was determined enzymatically by reincubation with a fresh preparation of the enzyme as described below. The enzyme (0.035 mg of protein/ml) in 0.16 M NaCI, 25 mM PIPES, 32 mM Tris (pH 6.1) was incubated in a Warburg flask at 37 "C under NP in the presence or absence of 100 p~ voso4 for 20 min. In order to assay the degree of enzyme inhibition, a 0.25-ml aliquot was quickly removed from the flask using a Hamilton syringe through a capillary polyethylene tube and was transferred to a fresh 0.16 M NaCI, 0.5 mM EDTA, 25 mM PIPES, and 32 mM Tris (pH flask containing 2.55 ml of a solution to give final concentrations of 6.1) with or without 0.33 mM ouabain. Three minutes later a solution of Mg/ATP containing bovine serum albumin and KC1 (to give 0.8 mM K+) was added to start the ATPase reaction as described previously, and aliquots were taken at 4-min intervals to determine inorganic phosphate released. Concurrently, the remainder of the vanadyl-enzyme mixture was filtered rapidly under vacuum through a 0.45-p Millipore filter. The filtrate was brought to pH 8 with Tris and bubbled with air overnight to assure complete oxidation of the vanadyl to vanadate. The concentration of vanadate (and, therefore, original vanadyl) was assayed by adding a 0.12-ml aliquot of the filtrate to 2.48 ml of a fresh suspension of the enzyme in 80 mM NaCl, 80 mM KCI, 4 mM MgC12, 40 mM imidazole, 40 mM MOPS (pH 7.2). The degree of inhibition observed 15 min later after addition of 4 mM Na2ATP (in 0.4 ml) was then compared to that produced by standard vanadate concentrations (1-4 g~) . The degree of enzyme inhibition varied linearly with the logarithm of vanadate concentration in this range.

RESULTS
Vanadyl in micromolar concentrations was a powerful and reversible inhibitor of highly purified membrane-bound (Na,K)-ATPase (Fig. 1). It was possible to study inhibition of (Na,K)-ATPase activity by vanadyl in effective isolation from inhibition by contaminating vanadate by working under a nitrogen atmosphere and in the presence of low concentrations of K+ (0.8 mM) and free M P (4 mM with 5 mM ATP). Under these ionic conditions (Na,K)-ATPase activity was about 20% of that in the standard system, and inhibition by 0.1 mM NH4V03 was less than 5%. EDTA was added to the vanadyl-enzyme mixture in order to chelate free vanadyl ion just prior to addition of Mg/ATP. This allowed estimation of the rate of recovery from vanadyl inhibition as well as the initial degree of inhibition (see under "Experimental Procedures"). For the data shown in Fig. 1, the initia1,degree of inhibition after a 15-min incubation with 50 p~ vanadyl was close to loo%, and the half-time for disappearance of the inhibition was about 6 min. The degree and stability of (Na,K)-ATPase inhibition increased with the time of incu- bation of the vanadyl-enzyme mixture (Fig. 2).
The sensitivity of (Na,K)-ATPase activity to vanadyl varied with the degree of enzyme purity. Among different fractions from a given zonal preparation of the enzyme, increasing sensitivity to vanadyl correlated roughly with increasing epecific activity (Fig. 3) and, therefore, with position in the sucrose density gradient (denser fractions possessing both higher specific activity and higher sensitivity to vanadyl). Typically, the best fractions of most zonal preparations of the enzyme (specific activity, 30-44 units/mg of protein) were inhibited 50% after a 30-min incubation with about 10 p~ vanadyl, with complete inhibition occurring at higher vanadyl concentrations. However, no inhibition was observed at 300 FM vanadyl in one crude enzyme preparation with a specific activity of 10 units/mg (Fig. 4A). Some preparations showed an atypical response to vanadyl that was characterized by relatively high sensitivity (KI < 5 p~) and high stability of inhibition (Fig. 4B). For enzyme fractions of high specific activity, a 30-min incubation at pH 7.1 with various concentrations of vanadyl showed Michaelis-Menten kinetics. A few early experiments indicated more complex kinetics with cruder poorly sensitive enzyme preparations (not shown).
We were interested in determining whether the variation of vanadyl sensitivity with enzyme purity reflected some intrinsic difference in the state of the enzyme in the different preparations or rather simply a reduction of the free vanadyl concentration by the cruder preparations. This reduction might be due to binding of vanadyl species. A nonspecific effect was suggested by early experiments showing reduction of vanadyl sensitivity with increase in the concentration of membrane (enzyme) protein in the reaction mixture (not shown). In order to measure the concentration of free vanadyl after incubation with the enzyme preparation, we separated free from bound vanadyl by filtration, quantitatively oxidized Vanadyl Inhibition of (Na, K)-ATPase the free vanadyl in the filtrate to vanadate, and enzymatically assayed the resulting concentration of vanadate by reincubation with a fresh preparation of the enzyme (see under "EXperimental Procedures"). Using this assay, we determined that during incubation of a relatively crude preparation of the membrane-bound enzyme (0.035 mg of protein/ml; 18.9 units/ mg) with 100 p~ VOs04, approximately 900 nmol of the vanadyl were bound per mg of protein, thus lowering the concentration of free vanadyl from 100 to 68 p~. At this concentration of free vanadyl ion, the enzyme was 58% inhibited relative to controls. Note that other enzyme preparations of higher specific activity would have been completely inhibited by even lower concentrations of total vanadyl (cf. Figs. 1  and 4B). Accordingly, we concluded that incubation with a crude enzyme preparation could significantly lower the concentration of free vanadyl ion but that this was not sufficient t o account for the reduced sensitivity to vanadyl seen in crude preparations. The latter, therefore, seemed to be an intrinsic property of the enzyme in the cruder preparations. Attempts to estimate the effect of adding a crude preparation to a pure one on the vanadyl sensitivity of the latter were unsatisfactory.
In order to examine the correlation between low enzyme specific activity and low sensitivity to vanadyl in an isolated enzyme preparation, we compared the vanadyl sensitivity of a high specific activity preparation and that of the same preparation after partial inactivation by the method applied by Levine (11) to glutamine synthetase. In this method incubation of the enzyme in the presence of ascorbate and FeC13 for either 70 or 240 min produced 65% or nearly complete (>99%) inactivation of (Na,K)-ATPase activity, respectively. The response of the 65% inactivated enzyme to 10 p~ vanadyl was compared to that of the untreated enzyme and to a 65:35 mixture (based on protein) of the completely inactivated enzyme and the untreated enzyme. The initial degree of inhibition of the 65% inactivated enzyme was much lower than that of either the untreated enzyme or the 65:35 mixture of the completely inactivated and untreated enzymes (Table   I). A possible interpretation of these results in terms of interaction between inactive and active enzyme molecules within the membrane bilayer is offered under "Discussion." Since pH has complex effects on the chemistry of vanadyl ion in solution (12), we compared vanadyl inhibition of (Na,K)-ATPase activity at pH values of 6.1 and 8.1. Comparisons between pH values were made for single preparations of the enzyme in order to avoid the variability of different enzyme preparations with regard to vanadyl inhibition. Characteristically inhibition was fully reversible after several minutes at pH 6.1, whereas at pH 8.1 it was stable for hours (Fig.  5 ) . In addition, the initial inhibition was slightly greater at pH 6.1 than at pH 8.1 (Fig. 5). Control experiments confirmed that EDTA was an effective chelator of vanadyl at both pH values, even after a 60-min preincubation of the VOSO, at those pH values without enzyme (data not shown). Thus we assume that a true rate of recovery from vanadyl inhibition Reduction in sensitivity to vanadyl produced by partial inactivation with ascorbate/FeCL Three enzyme systems were made from the following three preparations: untreated enzyme, 65% inactivated enzyme, and completely (>99%) inactivated (dead) enzyme. The first two systems consisted of the first two preparations at the same protein concentration. The third system consisted of a 65:35 mixture (with respect to protein) of completely inactivateduntreated preparations at the same total protein concentration. Thus the specific activity of the mixture was the same as that of the 65% inactivated enzyme. Vosod (10 p M ) was incubated with each system for 15 min at pH 6.5 under N P as described under "Experimental Procedures" before addition of EDTA to chelate the free vanadyl followed by Mg/ATP 6 min later to start the test of activity. Five successive aliquots were taken at 6-min intervals for bThe degree of initial inhibition in the presence of vanadyl is expressed as a percentage of the activity of the same system when measured without vanadyl, extrapolated back to the start of ATP hydrolysis (see under "Experimental Procedures"). e Half-times for recovery from vanadyl inhibition that are greater than 75 min represent 197% stable inhibition within the time frame of these measurements.
dThe mixture contained 1.42 pg of protein/ml of the untreated enzyme and 2.64 gg of protein/ml of the dead enzyme. under these conditions was observed. Earlier experiments with different preparations of the enzyme at an intermediate pH (pH 7.1) indicated two components of inhibition corresponding to the patterns observed at pH values of 6.1 and 8.1. About 65% of the inhibition was reversible in less than 30 min, but a residual component of stable inhibition was present several hours after addition of 1 mM EDTA (data not shown).
At pH 6.1 the properties of the vanadyl ion itself changed with time. Preincubation of the ion at that pH for 5 h before addition of enzyme produced a more stable inhibition (Fig.  6), mimicking that seen at pH 8.1 without the preincubation (cf. Fig. 5).
The tetravalent vanadyl species is stable as the divalent cation, V02+, below pH 3, but becomes increasingly susceptible to oxidation and hydroxylation as pH increases (12). We detected no oxidation of vanadyl (0.4 mM) to vanadate after 2 h of incubation at 37 "C under a N2 atmosphere at pH 6.1. In this measurement the relative proportions of the vanadyl and vanadate species were estimated by stabilizing both oxidation states by acidification, then separating them by Dowex chromatography (see under "Experimental Procedures"). Vanadyl and vanadate were monitored in the column effluent by their ultraviolet absorption. We also monitored the rate of oxidation of vanadyl solutions in air at 37 "C by observing the disappearance of the characteristic blue vanadyl absorbance at 766 nm. In a 4.2 mM vanadyl solution, half of the vanadyl was oxidized in air at pH 6.1 in 60 min, whereas at pH 8.1 half was oxidized in 12 min (Fig, 7). The rate of oxidation was dependent on the vanadyl concentration, being over twice as rapid at pH 6.1 in a 0.4 mM vanadyl solution as compared to a 4.2 mM solution (Fig. 7). In another experiment, enzymatic assay of the conversion of 5 HM vanadyl to vanadate at The effect of other ligands of the (Na,K)-ATPase on vanadyl inhibition was tested. We found that after vanadyl was allowed to complex with the enzyme preparation, a 10-min incubation with 4 mM ATP in the presence of 0.5 mM EDTA (before addition of Mg2+ to start ATP hydrolysis) significantly lowered the degree of inhibition compared to controls preincubated with EDTA alone (Fig. 8). Addition of 0.1 mM ammonium vanadate to the vanadyl-enzyme complex 2 min after chelation of the free vanadyl with EDTA and 13 min prior to addition of ATP to start the test of activity also resulted in a lower degree of initial inhibition and a higher rate of recovery from vanadyl compared to controls (Fig. 9). Initial inhibition refers to that extrapolated back to the start of the activity test. Acceleration of recovery from vanadyl inhibition by vanadate was favored by K' and Mg2+ more than by Na+. It was readily observable in the presence of K+ (0.8 mM) and M e (4 mM) without Na+, but was small or nonexistent in the presence of Na' (160 mM) alone (Fig. 9). Na+ or K+ alone with or without Mg2+ did not appear to modify vanadyl inhibition in the absence of vanadate (Fig. 9).

DISCUSSION
Vanadyl was a relatively potent inhibitor of membranebound kidney (Na,K)-ATPase producing nearly complete inhibition at a concentration of less than 5 I.LM in some highly purified preparations of the enzyme. It was not possible to the enzyme (8.6 pg of protein/ml; 25 units/mg of protein) was incubated in the buffer for 5 min in the absence of VOSO,, then EDTA (1 mM) was added, followed 1 min later by Mg/ATP. U , the enzyme was incubated with 50 PM VOSO, for 5 min before addition of EDTA and Mg/ATP as described above. A---A, the vanadyl ion was preincubated for 5 h in the solution buffered at pH 6.1, but without enzyme. Enzyme was then added and allowed to equilibrate with the vanadyl for 5 min before addition of EDTA and Mg/ATP. X-X, the enzyme and 50 PM VOSO, were incubated together for 5 h at pH 6.1 before addition of EDTA and Mg/ATP as described above.  Figs. 3 and 4). Reactions were performed under NP as described under "Experimental Procedures" with the following modifications. The enzyme (10 pg of protein/ml; 30 units/mg of protein) was incubated for 30 min at pH 6.5 in a 25/38 mM PIPES/Tris buffer containing 160 mM NaCl and 0.8 mM KC1 in the presence (---) or absence (-) of 30 p~ VOS04. EDTA (0.5 mM) was then added either alone (0) or mixed with Na2ATP to give an ATP concentration of 5 r n~ (X). Ten minutes later a solution of either MgCI, (X) or Mg/ ATP (0) was added to start ATP hydrolysis, giving in each sample final concentrations of 5 mM ATP, 4 mM M&12, 160 mM NaCl, 0.8 mM KC1, and 0.1 mg/ml of bovine serum albumin. The data points shown are averages of duplicate or triplicate determinations. Fig. 2 that inhibition did not saturate at 100% initial inhibition (which would show complete occupancy of one site), but that further inhibition developed a t longer times of incubation as indicated by a slower rate of recovery (showing a second mechanism of inhibition). The dependence of vanadyl inhibition on enzyme purity was not simply due to a greater reduction of the free vanadyl concentration by cruder enzyme preparations. The variation between the response of different enzyme preparations to vanadyl may reflect some difference in the intrinsic state of the enzyme in these preparations.

) . Also note in
The time-and pH-dependent changes in the degree and stability of (Na,K)-ATPase inhibition can be attributed to between pH 6 and 8 and below a concentration of about 0.3 mM is the monovalent anion, H2VO; (13). The state of hydroxylation of the vanadyl ion near neutral pH is much less certain, although vanadyl exists as the divalent cation V02+ in strong acid conditions and as a monovalent anion, VO(HO);, in highly alkaline conditions (12). Intermediate hydroxide forms of vanadyl presumably exist, but neither these nor their rates of equilibrium have been clearly defined. Hydroxylation of the divalent cation might be responsible for the observed time-and pH-dependent behavior of vanadyl as an inhibitor of (Na,K)-ATPase. Other potentially responsible factors include vanadyl polymerization or the formation of a vanadyl-vanadate complex. The latter, for instance, would be expected to be more pronounced at pH 8.1 than at pH 6.1 under an imperfect Nz atmosphere due to increased susceptibility of vanadyl to oxidation at the higher pH (cf. Fig. 7). Lindquist et al. (14) found that vanadate, produced by oxidation of vanadyl, increased the solubility of vanadyl. Perhaps related to the formation of a vanadyl-vanadate complex is our observation that vanadyl was more protected from oxidation at higher concentrations where the formation of such a complex would be kinetically favored ( Fig. 7 and text).
Two anionic ligands of (Na,K)-ATPase, vanadate and ATP, each increased the rate at which the enzyme recovered from vanadyl inhibition. The effect of these ions was not due to complexing of free vanadyl since they were added to the system together with or shortly following EDTA. Vanadate significantly increased the rate of recovery from vanadyl inhibition in the presence of ligands favoring the E2 conformation of the enzyme (K+ and M e ) , but not in the presence of Na+, which stabilizes the E l conformation. (The E l and E2 conformations of the enzyme are defined by multiple criteria, cf. Ref. 15.) Vanadate binding to (Na,K)-ATPase is enhanced by Mg2+ and K+ ions and appears to stabilize the enzyme in the E2 state (16). Since Na+, K+, and M e did not appear to modify vanadyl inhibition in the absence of vanadate (cf. Fig. 9), it seems that the effect of these ligands on vanadate-mediated reversal of vanadyl inhibition was due to an influence on the association of vanadate rather than that of vanadyl with the enzyme. The stimulatory effect of ATP on the recovery from vanadyl inhibition was observed in the presence of 160 mM Na' and 0.8 mM K' and in the absence of M e . Under these conditions ATP binds to the enzyme a t a high affinity site (17).
Vanadyl ion forms stable complexes with several enzymes. The electron paramagnetic resonance of vanadyl demonstrated metal binding site conformational states in carboxypeptidase A (18) and transferrin (19). The ion inhibits alkaline phosphatase (20). Complexes of both vanadate and vanadyl with uridine inhibit ribonuclease (14).
The potential of vanadyl as a physiologic inhibitor of the native cellular (Na,K)-ATPase requires further scrutiny since the inhibitory potency of vanadyl was much diminished in crude enzyme preparations. However, we are encouraged that inhibition by vanadyl may be useful for study of the mechanism of the enzyme since binding of other ligands (vanadate or ATP) or partial inactivation by chemical treatment affected vanadyl inhibition. The data shown in Table I indicate that the low sensitivity of crude enzyme preparations may be mimicked by a highly purified enzyme preparation which has been lowered in specific activity by treatment with ascorbate/ FeC13. The mechanism of this inactivation is not known. However, if the treatment inactivates in an all-or-none fashion, so that 65% inactivation represents complete inactivation of 65% of the originally active enzyme molecules, then an interesting inference can be drawn from Table I. The presence of inactive (Na,K)-ATPase molecules within the membrane bilayer may impair the response of active enzyme molecules to vanadyl.