(Received for publication, January 23, 1997, and in revised form, April 16, 1997)
From the Laboratory of Membrane Biology, Neuroscience
Center, Massachusetts General Hospital,
Charlestown, Massachusetts 02129
Na,K-ATPase activity has been demonstrated to be
regulated by a variety of hormones in different tissues. It is known to
be directly phosphorylated on its -subunit, but the functional
effects of protein kinases remain controversial. We have developed a
sensitive, antibody-based assay for detection of the level of
phosphorylation of the
1-isoform of rat Na,K-ATPase at the serine
residue that is most readily phosphorylated by protein kinase C (PKC)
in vitro, Ser18. By stimulation of endogenous
PKC and inhibition of phosphatase activity, it was possible to
consistently obtain a very high stoichiometry of phosphorylation (close
to 0.9) in several types of intact cells. This demonstrates the
accessibility and competency of the site for endogenous
phosphorylation. The cells used were derived from rat (NRK 52E, C6, L6,
and primary cultures of cerebellar granule cells, representing
epithelial cells, glia, muscle cells, and neurons). In the presence of
the phosphatase inhibitor calyculin A, full phosphorylation was
preserved during subsequent assays of enzyme activity in
vitro. Assay of the hydrolysis of ATP in NRK and C6 cells,
however, indicated that there was no significant effect of
phosphorylation on the Vmax of the Na,K-ATPase
or on the apparent affinity for Na+. Any regulatory effect
of PKC on sodium pump activity thus must be lost upon disruption or
permeabilization of the cells and is not a direct consequence of enzyme
alteration by covalent phosphorylation of Ser18.
The Na,K-ATPase is a membrane protein responsible for maintaining
the electrochemical gradients of Na+ and K+ in
almost all mammalian cells. Since many physiological events in the
cells depend on the concentrations of these ions, the mechanisms regulating Na,K-ATPase activity are of great importance. Direct phosphorylation and dephosphorylation of Na,K-ATPase by protein kinases
and protein phosphatases have been proposed to be mechanisms for
modulation of sodium pump work. Since many hormones are known to affect
activities of protein kinases or phosphatases, the reversible phosphorylation of the Na,K-ATPase molecule may present a missing link
between hormone action and changes in enzyme activity. The -subunit
of Na,K-ATPase has been shown to be a substrate for Ca2+-,
phospholipid-dependent protein kinase
(PKC)1 both in vitro (1-6, 17)
and in intact cells (5, 7-10), and two PKC phosphorylation sites have
been identified for rat
1 in vitro (6, 11). Both sites
(Ser11 and Ser18) are located at the N terminus
of the
-subunit. Using direct sequencing of phosphorylated ATPase,
we showed that Ser18 was the major phosphorylation site
with a commercially available mixture of
,
, and
isoforms of
PKC.
A physiological role for protein kinase C regulation of Na,K-ATPase is controversial, however. Although there are reports of inhibition of enzyme activity as a direct result of phosphorylation in the dogfish shark, duck salt gland, and rat kidney enzymes (2, 6, 12), most of the literature deals with indirect evidence that PKC stimulation by phorbol esters can change Na,K-ATPase activity in intact cells or membrane preparations. Moreover, both activation (13, 14) and inactivation (2, 7, 9, 15, 16) of the Na,K-ATPase have been reported upon stimulation of PKC. Our attempts to detect an effect of PKC phosphorylation on purified rat kidney enzyme showed no change in ATPase activity (17).
In this study we developed a nonradioactive antibody-based assay to
answer three central questions: 1) is the same major site used by
endogenous PKC in intact cells as by exogenous PKC in vitro?; 2) what is the balance between PKC activity and activity of phosphatases responsible for 1 dephosphorylation?; and 3) does
activation of PKC and phosphorylation of Na,K-ATPase in cells result in
any change in ATPase activity or Na+ affinity?
Purified rat Na,K-ATPase
was prepared by the method of Jørgensen (18), using the medulla
from frozen kidneys obtained from PelFreez Biologicals (Rogers, AR).
Enzyme preparations were stored in buffer containing 20 mM
Tris (pH 7.2), 320 mM sucrose, 1 mM EDTA. The
specific activity of Na,K-ATPase was 1200 µmol/h/mg. Protein
concentration was determined by the method of Lowry using bovine serum
albumin as a standard (19). PKC from rat brain was purchased from
Boehringer Mannheim GmbH Biochemica (Germany). The PKC preparations
used consist of -,
-, and
-isoforms (information provided by
the supplier).
Purified Na,K-ATPase
at a concentration of 10-50 µg/ml was preincubated with PKC (1.5 µg/ml) for 3 min at 30 °C in buffer containing 10 mM
Tris phosphate (pH 7.4), 5 mM magnesium acetate, 500 µM CaCl2, 100 nM phorbol
12-myristate 13-acetate (PMA; Sigma), and 80 µg/ml phosphatidyl
serine (in a sonicated suspension) (Sigma, P7769). The reaction was
started by the addition of [-32P]ATP (1800 cpm/pmol)
to a final concentration 60 µM. Phosphorylation proceeded
for 30 min at 30 °C and was stopped with an equal volume of
electrophoresis sample buffer containing 125 mM Tris, pH
6.8, 4% SDS, 20% glycerol, and bromphenol blue (Laemmli sample
buffer). Control samples were prepared the same way except that instead of PKC, an equal volume of PKC storage buffer (50% glycerol, 0.02% Tween 20, 20 mM Tris-HCl, 0.5 mM EDTA, 0.5 mM EGTA, 5 mM dithiothreitol, pH 7.5) was
added. Electrophoresis and transfer to nitrocellulose were performed as
described below. Phosphoproteins were visualized by overnight
autoradiography of the dried nitrocellulose on Kodak XAR-5 film.
Densitometry was performed with a Molecular Dynamics 300A laser
densitometer.
Electrophoresis was performed as described previously (20), using 10% polyacrylamide and the Bio-Rad mini-ProteanII apparatus. Electrophoretic transfer of proteins was performed in buffer containing 20 mM Tris, 150 mM glycine, and 20% methanol at 0.15 A overnight.
For immunostaining, nitrocellulose membranes were first blocked with 20 mM Tris-HCl buffer (pH 7.5) containing 150 mM
NaCl and 0.5% Tween 20 (TBS-Tween) for 30-40 min at room temperature. Incubations with the primary and horseradish peroxidase-conjugated secondary antibodies were performed in the same buffer for 1.5 and
1 h, respectively. Membranes were extensively washed with TBS-Tween after each step and finally incubated with luminol reagent (Pierce). Membranes were then exposed to Kodak XAR-5 film for 0.5-20
min. Three different Na,K-ATPase antibodies were used in this study:
monoclonal antibody McK1, which binds near the N terminus of the
1-subunit (21); monoclonal antibody 6F, which has an epitope
approximately 40 amino acids away from the McK1 epitope (22); and
polyclonal antiserum K1, which was raised against purified rat kidney
Na,K-ATPase
1-subunit and is known to recognize polypeptide
fragments from different parts of the molecule (N-terminal and
C-terminal halves) (21, 23). All three antibodies demonstrated similar
sensitivity (luminescence per unit of antigen) when used at the
following dilutions: 1:5,000 for McK1, 1:400 for 6F, and 1:2,000 for
K1. For sequential staining of the same blot, the monoclonal antibodies
were stripped off by 30-min incubation in 10% acetic acid, 40%
methanol and extensively washed with TBS-Tween. To assess the retention
of antigen, blots were restained with 0.1% Amido Black. There was no
significant loss of Na,K-ATPase after the stripping procedure. Antibody
6F was obtained from the Developmental Studies Hybridoma Bank (Iowa
City, IA).
Three rat cell lines (NRK-52E (normal rat kidney), C6 glioma, and L6 skeletal muscle myoblasts) were obtained from the American Type Culture Collection and maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and antibiotics (100 IU/ml penicillin, 100 µg/ml streptomycin) in a 5% CO2, 95% air atmosphere. NRK and C6 cells were used in phosphorylation experiments immediately upon reaching 90-100% confluence. L6 cells were allowed to fuse and form myotubes in Dulbecco's modified Eagle's medium containing 2% fetal bovine serum for 7-10 days. It should be noted that the 52E subclone of NRK has an epithelial phenotype and is different from the NRK fibroblastic subclone commonly used in virology (24).
Primary neuronal cultures were prepared from the cerebellum of 7-day-old rats as described (25). In brief, 10 cerebelli were cleaned of meninges, chopped with razor blades, and then treated with 0.025% (w/v) trypsin for 10 min in Ca2+- and Mg2+-free Dulbecco's phosphate-buffered saline. The reaction was stopped by adding 2% fetal bovine serum. The cells were pelleted and resuspended in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 2 mM L-glutamine, 30 mM glucose, and 24.5 mM KCl. The cell suspension was filtered through a mesh with a pore size of 75 µm (Falcon cell strainer) and seeded onto polylysine-coated 35-mm tissue culture dishes (Falcon). After 15 min at 37 °C, the medium containing unattached cells was removed, and fresh medium was added. To prevent proliferation of nonneuronal cells, cytosine arabinoside (10 µM) was added to the culture medium 24 h after seeding. Cells were incubated at 37 °C in a 5% CO2, 95% air atmosphere for 7-8 days prior to phosphorylation experiments.
In Vivo PhosphorylationMonolayers of the cells grown in 35- or 60-mm tissue culture dishes were preincubated for 1.5-2 h at 37 °C in a 5% CO2, 95% air atmosphere in Puck's buffer (Sigma) supplemented with 0.25 mM CaCl2, 0.5 mM MgSO4, 1 mM NaH2PO4, 22 mM NaHCO3, and 10 mM Hepes. Then drugs (PMA, calyculin A) or their vehicle (ethanol) were added at the concentrations indicated in the figure legends, and the cells were incubated for 20-30 min at 37 °C. The buffer was discarded, and dishes with the attached cells were immediately placed on ice. 0.2-0.5 ml of 100 °C 10 mM Tris-HCl buffer (pH 7.4) containing 1% SDS was then added to each dish. The cells were scraped and clarified by centrifugation for 30 min at 40,000 rpm at 10 °C. Supernatants were collected, diluted 1:1 with Laemmli sample buffer, and loaded on gels.
Cell Membrane PreparationsIntact cells were incubated to enhance phosphorylation as described above. After the incubation buffer was discarded, 1-2 ml of ice-cold hypotonic buffer containing 10 mM Tris phosphate (pH 7.2), 5 mM magnesium acetate, 1 mM EGTA, 1 mM EDTA, and 100 nM calyculin A was added to each dish. Dishes were incubated on ice for 15-20 min and then scraped, homogenized, and centrifuged for 30 min at 40,000 rpm, 4 °C. The supernatants were discarded, and the pellets were washed once in the same hypotonic buffer. The final pellets were resuspended in buffer containing 20 mM Tris (pH 7.2), 320 mM sucrose, 1 mM EDTA, and 100 nM calyculin A.
Na,K-ATPase Activity MeasurementsFirst, the test tube assay for the hydrolysis of ATP was performed as described (17). The released 32Pi was quantified by the formation of a phosphomolybdate complex. The reaction was started by the addition of 20-40 µg of NRK or C6 membranes in 0.5 ml of buffer containing 50 mM Tris-HCl (pH 7.4), 3 mM MgCl2, 3 mM NaATP, 1 mM EGTA, 1 mM EDTA, 5 mM NaN3, 115 mM NaCl, 5 mM KCl, and 100 nM calyculin A, with or without 2 mM ouabain. The reaction proceeded for 15-20 min at 37 °C and then was stopped with 0.5 ml of quenching solution (1 N sulfuric acid, 0.5% ammonium molybdate). Then 1 ml of isobutanol was added, and the phosphomolybdate complex was extracted into the organic phase by 10 s of vortex mixing, followed by separation of phases by centrifugation in a table top centrifuge. Aliquots of the upper phase were counted with a scintillation counter. Second, a spectrophotometric assay in which the generation of ADP is coupled to the oxidation of NADH and monitored continuously by absorbance at 340 nm was performed as described (17), except that 3 mM sodium azide and 100 nM calyculin A were present in the assay medium. Third, determination of total Na,K-ATPase activity in perforated cultured cells was performed as described (26). After phosphorylation, monolayers of cells were washed briefly with 10 mM Tris-EGTA (pH 7.4) containing 100 nM calyculin A and placed in the same solution with the addition of 0.03 mg/ml of alamethicin. The cells were incubated for 10 min at room temperature, scraped, and homogenized briefly with a Dounce homogenizer (loose pestle) to disperse the cells, and 50-µl aliquots were used in activity assays (approximately 50-100 µg of total protein/aliquot). The assay was performed as described above for the test tube assay.
Previously, we showed that phosphorylation of Na,K-ATPase by
PKC occurs at two sites on the N terminus of the rat 1-subunit: Ser11 and Ser18 (Fig. 1) (11). A
total stoichiometry of 0.9 mol/mol was found for rat
1 when purified
enzyme was incubated with purified kinase. Direct sequencing of
phosphorylated protein showed that 75-80% of the phosphate
incorporated was bound to Ser18 and only 20-25% to
Ser11. The conclusion was that Ser18 is the
main PKC phosphorylation site in rat, and neither of the sites was
fully phosphorylated, unless some preexisting, stable phosphorylation
interfered with the incorporation of 32P. Interestingly,
Ser18 is absent in Na,K-ATPase
-subunit sequences from
many other species; that was shown to be the reason for much lower
levels of phosphorylation (0.15 mol/mol) in pig and dog
1-subunits.
Earlier, monoclonal antibody McK1 was found to bind to the amino acid
sequence DKKSKK, which includes Ser18 (Fig. 1) (21, 22).
This antibody does not recognize Na,K-ATPase from pig and dog, which
have the shorter sequence DKKK at that location. Since the McK1 epitope
and the major site for PKC phosphorylation include the same amino acid
residues, we hypothesized that phosphorylation of Ser18
might affect McK1 binding, making it possible for this antibody to
probe the level of PKC phosphorylation of Na,K-ATPase in cell cultures
and tissues. As shown in Fig. 2 below, this is
successful under appropriate experimental conditions.
First, experiments were performed to assess the sensitivity and linearity of antibody signals from McK1 and two antibodies used as controls. Since Na,K-ATPase is not very abundant in cell cultures, a high sensitivity method is required to detect antigen-antibody binding. Horseradish peroxidase-luminol staining of Western blots is known to be several orders of magnitude more sensitive than traditional staining with colorigenic substrates, and it is possible to scan the images obtained to quantify and compare their optical densities. Caution is needed in using this method for quantifying the amount of antigen in different samples. The limitations of the assay are imposed both by the luminescent method itself and by the binding characteristics of the antibody. At higher concentrations of antigen the luminescent signal becomes nonlinear and finally reaches a plateau (due either to exhaustion of luminol reagent or x-ray film saturation), while at lower antigen concentrations it may drop abruptly below the detection level. There is a linear zone that can be shifted by adjusting the time of exposure of the blot to x-ray film. We found that if the amounts of antigen in samples differ more than 3-fold, accurate quantitation is not possible unless calibration curves are made in each experiment. Since cell cultures contain only small amounts of Na,K-ATPase, such samples are more likely to face the problem of having the signal below detection level rather than reaching saturation.
To test whether the McK1 antibody can be used to detect PKC
phosphorylation, we compared binding of the antibody to phosphorylated and nonphosphorylated Na,K-ATPase (Fig. 2A). Three different
concentrations of purified rat kidney Na,K-ATPase were used.
Na,K-ATPase was incubated in phosphorylation buffer with (lanes
4-6) or without (lanes 1-3) exogenous PKC and
[-32P]ATP. Samples were subjected to SDS-gel
electrophoresis and transferred to nitrocellulose membrane.
Phosphoproteins were visualized by autoradiography and then
immunostained with monoclonal antibody McK1 using the horseradish
peroxidase-luminol method. Phosphorylation of Na,K-ATPase by PKC
dramatically decreased the binding of McK1 antibody. Thus,
Ser18 is an important amino acid residue in the McK1
epitope, and covalent modification by phosphorylation prevents antibody
binding. The assay is consequently based on detecting how much
unmodified Ser18 is left after PKC phosphorylation.
Incorporation of 32P as detected by scintillation counting
was linear at all concentrations of Na,K-ATPase used (not shown), while
32P autoradiography (not shown) and McK1 binding (Fig.
2B) deviated at higher antigen concentration. Since the
signal for the phosphorylated sample in lane 4 of Fig.
2A ("unknown") fell in the linear zone of the McK1
binding curve, we were able to calculate the amount of unmodified
Na,K-ATPase in this sample. The result was 0.13 µg, which is 26% of
the 0.5 µg used in the phosphorylation experiment. These data
correlate very well with our previous finding that approximately 75%
of Ser18 becomes modified after PKC phosphorylation
in vitro (11).
PKC Phosphorylation in Intact CellsTo assess the
generality of our findings, four different types of rat cells were used in this study: 1) NRK (clone 52E); 2) C6 glioma; 3) L6 skeletal muscle
myotubes; and 4) primary cultures of cerebellar granule neurons. All of
these cells express the
1-isoform of Na,K-ATPase. PKC
phosphorylation was determined by quantifying McK1 binding relative to
control antibody. Intact cells were incubated with or without protein
kinase and phosphatase modulators. Fig. 3A shows the result of an experiment with NRK cells. Treatment of the
cells with PMA significantly decreased McK1 binding, while the
combination of PMA and the protein phosphatase inhibitor calyculin A
practically abolished it, meaning that most of the Na,K-ATPase became
phosphorylated at Ser18. Calyculin A alone was ineffective,
suggesting that in the absence of stimuli, the activity of PKC in the
NRK cell line is low. Prolonged treatment with PMA (24 h), which is
known to down-regulate PKC, did not change the basal level of
Na,K-ATPase phosphorylation (results not shown). 6F binding was
unaffected by any of the treatments, as expected from the observation
that its epitope is approximately 40 amino acids away (22). Fig.
3B summarizes the results of phosphorylation experiments
with all four cell culture types. Binding of McK1 antibody was
normalized to binding of 6F or K1 antibody. PMA treatment alone
increased phosphorylation of Na,K-ATPase in NRK and C6 cells (as
detected by the loss of McK1 binding). It was less effective in
cerebellar neurons and practically ineffective in L6 cells. The
combination of PMA and calyculin A caused a dramatic increase in the
phosphorylation level of Na,K-ATPase in all four types of cell
culture.
Protein Phosphatase Activity in Cell Cultures
The effect of calyculin A in potentiating the effect of PMA on the level of phosphorylation of Ser18 showed that protein phosphatases are very active in the cell cultures tested. Treatment of intact cells with 1 µM PMA alone was able to bring about 50% phosphorylation of Na,K-ATPase at most. Attempts to increase phosphorylation by raising the concentration of PMA up to 10 µM and/or prolonging the incubation time were unsuccessful (results not shown). The combination of 1 µM PMA with 100 nM calyculin A caused 70-90% PKC phosphorylation of Ser18 in all experiments. The level of phosphorylation was unaffected by a 10-fold decrease in PMA concentration, indicating that PKC was fully activated at the lower PMA concentration (0.1 µM), while the same decrease in calyculin A concentration (final concentration, 10 nM) resulted in 15-20% loss of phosphorylation (data not shown). Thus, to reach the highest level of phosphorylation, the activity of protein phosphatases must be controlled with an appropriate concentration of phosphatase inhibitor.
Another major concern is retention of the phosphate introduced by PKC.
To study the functional consequences of phosphorylation it is important
both to obtain the highest level of phosphorylation possible and to
preserve the phosphorylation during subsequent experimental
manipulations. The stability of PKC phosphorylation in perforated C6
cells and membrane preparations isolated from NRK and C6 cells was
investigated in the conditions of ATPase activity assays. We found that
in ATPase activity assay medium in the absence of calyculin A, almost
complete dephosphorylation of the Ser18 site occurred after
15-30 min of incubation at 37 °C (Fig.
4A). The presence of 100 nM of
calyculin A throughout the assay prevented dephosphorylation (Fig.
4B). The result indicates that activity of protein
phosphatase in the cells tested is very high and is not removed during
cell permeabilization or membrane isolation. It can be seen that there
was little change in the phosphorylation level of the control samples;
this was true of both cell lines, further supporting the conclusion
that both have similar, low levels of basal phosphorylation.
Functional Consequences of PKC Phosphorylation at Ser18
Two different kinds of preparations were used
in activity experiments: C6 cells perforated with alamethicin by the
method of Askari and co-workers (26) and membrane preparations isolated from cultured C6 and NRK cells. Intact cells were incubated with 1 µM PMA and 100 nM calyculin A (phosphorylated
samples) or with calyculin A only (controls), and all of the subsequent
steps were performed in the presence of 100 nM calyculin A
to preserve phosphorylation of Ser18. The level of
phosphorylation obtained was checked in every experiment. Almost
stoichiometric phosphorylation was obtained in the PMA-treated cells
(Fig. 5). To test whether the difference between control and phosphorylated sample was exaggerated because the McK1 signal dropped below the detection level, we loaded two different amounts (20 and 40 µg) of NRK membranes on the gel. McK1 staining was normalized
against K1 staining, and the percentage of unmodified Ser18
was calculated in phosphorylated samples (lanes 2 and
4) relative to that in control samples (lanes 1 and 3). The results were, respectively, 10 and 12%,
confirming that the difference in the signals was not exaggerated.
Activities of control and phosphorylated samples were compared in
maximal turnover conditions using hydrolysis of
[-32P]ATP in test tubes or, in a few experiments, the
coupled ATPase assay (Table I). No significant effect
(p > 0.05) of phosphorylation was found for membrane
preparations of NRK or C6 cells or for preparations of permeabilized C6
cells, despite approximately 90% phosphorylation of Ser18.
Table I lists the means and S.D. values, the p value for
each group, and the 95% confidence interval. Given the observed S.D. values, statistical analysis showed that we would have 80% power to
detect increases or decreases in activity of 13, 9, or 24%, respectively, in the groups given in Table I. This estimates the
accuracy of the assay and puts a quantitative limit on the conclusion
that there was effectively no change in Na,K-ATPase activity.
|
Alteration of Na+ affinity has been hypothesized to mediate
hormonal effects on Na,K-ATPase activity (14, 27, 28), and effects of
PKC have been observed on Na,K-ATPase conformational equilibria (6). We
used the coupled assay to measure Na,K-ATPase activity at different
concentrations of Na+ to determine whether phosphorylation
of Ser18 resulted in a change in Na+ affinity
(Fig. 6). Again, there was no significant effect as detected with control and phosphorylated preparations of NRK membranes. Calyculin A by itself did not affect the activity of Na,K-ATPase in any
of the assays used.
The Na,K-ATPases of various animal species differ both in
the sites where PKC phosphorylates the -subunit and in the recorded maximal stoichiometries of phosphorylation. We have used the rat
1
Na,K-ATPase because it is possible to obtain a stoichiometry of
32P incorporation close to 1.0 at a defined site near the N
terminus. Under these conditions, we hoped to have the best chance of
finding a rigorous correlation between phosphorylation and any
alteration in the activity of the enzyme. No effect of phosphorylation
was seen in our prior work using purified Na,K-ATPase and purified PKC
(17). That left open the possibility that the site phosphorylated in vitro was not used physiologically, however.
Ser18 is unique to rat
1, and it is reasonable to
question whether it is a legitimate phosphorylation site in
vivo, or whether it becomes accessible only upon some disruption
of protein structure that occurs during enzyme purification. Here we
have demonstrated with four different types of rat cell cultures that,
in intact cells stimulated with phorbol esters, the site is robustly
phosphorylated when protein phosphatase activity is inhibited with
calyculin A.
There is now little doubt that PKC phosphorylation of Na,K-ATPase
-subunit can take place in intact cells and tissues. Direct PKC
phosphorylation of Na,K-ATPase
-subunit was demonstrated in OK
(opossum kidney) cells (7); in normal and diabetic rat sciatic nerve
(8); in COS-7 cells expressing Bufo marinus
-subunit (5,
10); in rat choroid plexus (9); and in NRK, C6, and L6 cell lines and
primary cultures of rat cerebellar granule neurons (present study). In
the case of the rat choroid plexus stimulated with phorbol 12, 13-dibutyrate or serotonin, phosphopeptide maps indicated that the same
sites were used as with purified enzyme (9), consistent with the
results presented here. The mapping technique used could not resolve
Ser11 from Ser18, however, leaving open the
possibility that only Ser11 was phosphorylated in
vivo. The identity of the sites phosphorylated in intact cells or
tissues has been even less clear in other work. In transfected COS-7
cells expressing B. marinus
-subunit, a substantial
background of unidentified basal phosphorylation activity was
stimulated by phorbol ester even in mutations of the Thr15
and Ser16 PKC phosphorylation sites, suggesting that there
may be additional sites (5). In similar cultures, activation of
-adrenergic or muscarinic cholinergic receptors that normally
increase PKC activity through Ca2+ entry and diacylglycerol
release instead increased Na,K-ATPase phosphorylation at the protein
kinase A site (Ser943), in a process inhibitable by H-89, a
protein kinase A inhibitor (10). In other cases the site(s) of
endogenous PKC action on the Na,K-ATPase
-subunit have not been
identified (7, 8). With the potential for complex regulatory
interactions, the interpretation of experimental results should be
easier when the phosphorylation of specific sites can be explicitly
measured. An assay that makes it possible to assess phosphorylation
directly in every experiment should make quantitative assessment of
cause and effect more feasible.
The binding of McK1 requires unphosphorylated Na,K-ATPase, and it thus provides a good assay to demonstrate near complete phosphorylation, since it sensitively detects the fraction that has not reacted with the kinase. It is more difficult to make conclusions about the basal level of phosphorylation if it is low, however, since a small reduction in antibody binding could be hard to detect. The available indirect evidence suggests that the basal phosphorylation of Na,K-ATPase in the cell lines tested was low. Down-regulation of PKC by prolonged incubation with phorbol ester did not increase binding of McK1 antibody, and the phosphatase inhibitor calyculin A alone failed to significantly decrease it. In addition, there was little increase in McK1 binding when incubation of membranes at 37 °C allowed phosphatase action. For example, in Fig. 4A, if there had been significant basal phosphorylation in the control membranes (lane 1), then after permitting phosphatase action during prolonged incubation in assay medium (lane 3) there would have been an increase in McK1 binding. PKC activity appears to contribute very little to the resting level of Na,K-ATPase phosphorylation at Ser18. In rat choroid plexus, the basal level of phosphorylation was found to be 6-7% of that obtained after stimulation of PKC with phorbol 12,13-dibutyrate (9). Assuming that stoichiometric phosphorylation was obtained with phorbol 12,13-dibutyrate, a basal level on the order of 6-7% (or less) would be unlikely to be detected in our assay.
The lability of the phosphate attached to Ser18 in medium without phosphatase inhibitors makes it unlikely that there is much phosphorylation left in conventional membrane preparations, and thus it is unlikely that prior studies employing the McK1 antibody to detect Na,K-ATPase on Western blots are compromised by the new observation that phosphorylation prevents antibody binding. Whether or not any phosphate on Ser18 would be preserved during tissue fixation is a separate question, however. It is conceivable that McK1 staining of tissue sections could be altered by the phosphorylation state at the time of fixation, a possibility that remains to be tested.
Phosphorylation by PKC at Other SitesMost recent reports
have demonstrated the existence of PKC phosphorylation sites close to
the N terminus of the Na,K-ATPase -subunit. There is marked sequence
divergence at this location, making it necessary to identify
phosphorylation sites in each species individually. The identified
sites include Ser11 and Ser18 for rat
1 (6,
11); Ser11 for sheep (5), pig, and dog
1 (11) (deduced
because it is the only serine or threonine in the short phosphorylated
N-terminal segment that was removed by trypsinolysis); and
Thr15 and Ser16 in Bufo
1 (5).
Bufo Ser16 is the homolog of mammalian
Ser11 (Fig. 1). It is curious that the stoichiometry of
phosphorylation of Ser11 and its homolog seldom exceeds
0.15 (Refs. 5 and 11; by inference in Ref. 1), while that of
Ser18 reaches 0.7-0.9 (Ref. 11 and present study) or
0.5-0.65 (6). A higher level of phosphorylation of Na,K-ATPase by PKC
has been reported for only one other Na,K-ATPase, the dogfish shark
rectal gland enzyme, where a stoichiometry of 2.0 was found (2).
Unfortunately, the sequence of that Na,K-ATPase is not known, nor is
the location of the sites (both serine and threonine) of
phosphorylation.
The species difference in maximum observed phosphorylation
stoichiometry for Ser11 and Ser18 has a
correlate in the structure of the phosphorylation site. The high
stoichiometry in rat Ser18 corresponds with a surrounding
sequence that resembles a typical consensus sequence for PKC: basic
residues at one or two positions on either side of the candidate
Ser/Thr (29). The low stoichiometry typical of mammalian
Ser11 or Bufo Ser16 corresponds with
a surrounding sequence with no Lys or Arg residues (Fig. 1). PKCII
is autophosphorylated at sites that also are not typical PKC consensus
sites (30). The autophosphorylation sites do not have any clear
homology to the Ser11 site on Na,K-ATPase, but they have in
common with it several nearby proline, glycine, and glutamate residues.
It remains to be determined whether Ser11 can be
phosphorylated to a higher level by endogenous kinases in cells. If it
cannot, then any functional consequences expected from its
phosphorylation must be quantitatively limited compared with the
consequences of phosphorylation at sites that can be fully
phosphorylated.
An antibody-protein binding site typically encompasses a 600-Å2 surface area and involves contacts with 20 or more residues, not necessarily adjacent in the linear sequence (31). The mapping of the epitope for McK1 makes it clear that Ser18 is a crucial contact; even the single substitution of Ala18 in mouse and human greatly decreases antibody affinity, and the deletion of Lys17 and Ser18 abolishes binding (21). Whether Ser11 is also contacted by the antibody cannot be determined without co-crystallization and an atomic structure, but the failure of McK1 to bind to a chimera containing amino acids 1-13 but lacking amino acids 14-18 suggests that Ser11 is not an important part of the antibody binding site (22). Consequently, we cannot say anything about the phosphorylation state of Ser11 in these experiments. It is plausible that phosphate incorporation occurs at both serines after stimulation of PKC in intact cells.
There is some possibility of PKC phosphorylation at other sites, since
there are more than 30 potential consensus sequences for PKC
phosphorylation. It has been reported that PKC produces phosphothreonine as well as phosphoserine in dog kidney enzyme (1),
although the dog 1 sequence does not have threonine near the N
terminus. The maximum observed stoichiometry in that study was 0.15. A
phosphorylation site was mapped to an internal location in duck salt
gland enzyme by selective proteolysis (4); previous work by the same
group had shown phosphate incorporation into both serine and threonine,
implying that there were at least two sites (3). The maximum observed
stoichiometry was 0.3. For rat
1 phosphorylated in vitro,
no phosphorylation was detected at any location other than the N
terminus, as determined by proteolytic fingerprinting (11). In the
present study we of course would not have detected phosphorylation at
alternative sites even if it occurred as a consequence of activation of
specific PKC isoforms in intact cells. In any event, any such
phosphorylation did not have detectable consequences on enzyme
activity.
The observed physiological level of phosphorylation represents the balance between kinase and phosphatase activity, and in fact the second messenger-mediated regulation of phosphatase activity has been implicated in the physiological regulation of Na,K-ATPase (28, 32). In our experiments with four different kinds of cells, we found that PMA alone was able to increase phosphorylation at Ser18 by 50% at most, while in combination with calyculin A phosphorylation reached 70-90%. The result suggests strong phosphatase activity in all four cell cultures. Interestingly, in L6 myotubes PMA alone was practically ineffective in increasing phosphorylation (similar to observations in normal rat sciatic nerve (8)), but the application of the phosphatase inhibitor calyculin A alone changed the phosphorylation of Na,K-ATPase by 20-30%. Thus, it is possible that in L6 cells there was some basal phosphorylation at Ser18 that is continuously removed by phosphatase.
It is not obvious why PKC activities in the cell cultures tested appeared so low, while protein phosphatases were highly active. It is conceivable that there is little PKC activation because of relatively low hormone levels in culture medium with 10% fetal serum or because some pathways are inactive due to the absence of contacts with other cell types present in normal tissues. A similar imbalance between protein kinase and protein phosphatase activities was found by others in isolated rat adipocytes where protein kinase activities were measured directly (33).
In different cells and tissues, PKC and phosphatase activities may be
regulated independently, affecting the observed level of Na,K-ATPase
phosphorylation. A significant level of basal phosphorylation of
Na,K-ATPase has been seen in normal rat sciatic nerve (8). Since
okadaic acid also increased phosphorylation of Na,K-ATPase approximately 3-fold, the authors concluded that the phosphorylation state of the -subunit results from a balance between a strong phosphatase activity and the activity of PKC. Basal phosphorylation in
primary cultures of cerebellar granule neurons appears to be low,
however (Fig. 3). This contradicts a prediction that dephosphorylation of a high basal level could explain the increase in Na,K-ATPase activity seen when such cultures are treated with glutamate (34). The
mechanism of the effect of glutamate on Na,K-ATPase activity remains to
be determined.
Despite a high stoichiometry of phosphorylation of
Ser18 in vitro, in our prior study we did not
see any effect on Na,K-ATPase activity (11, 17). This appears to
contradict several other reports. First, there was 50% inhibition of
purified dogfish shark enzyme reported at a stoichiometry of 2.0 (2),
although species differences or phosphorylation at the second site
could obviously be responsible. More importantly, 30% inhibition was
reported in purified rat renal cortex Na,K-ATPase at a stoichiometry of 0.5, accompanying an effect on conformation affecting apparent affinity
for K+ (6). There have also been reports of PKC-mediated
inhibition of Na,K-ATPase activity correlated with phosphorylation of
in cells or tissues: 28% in rat choroid plexus (9) and 30% in OK
cells (but not in LLC-PK1 cells) (7). Modulatory effects on
Na+ affinity have been suggested in other work to be
important for Na,K-ATPase regulation (27, 28, 35). Nevertheless, we
failed to see any statistically significant effect on ATP hydrolytic rate in C6 or NRK membranes or in permeabilized C6 cells in maximal velocity conditions or on Na+ affinity.
We consider it likely that the observation of no effect under conditions where the level of phosphorylation is carefully monitored is a more significant result than the observation of inhibition, because there are alternative ways to obtain inhibition. These alternatives include activation of phospholipase A2 and the release of arachidonic acid metabolites, activation of cross-talking pathways that result in Na,K-ATPase phosphorylation at other sites, and modulation of phosphatase activity, to name a few that can be documented. In rat kidney-proximal tubules, Na,K-ATPase inhibition caused by activation of PKC can be inhibited by ethoxyresorufin, an inhibitor of the cytochrome P450-dependent monooxygenase pathway, and mepacrine, an inhibitor of phospholipase A2 (36, 37). A phospholipase A2-mediated source of lipid-derived inhibitors was also implicated in a study showing that PKC inhibited proximal tubule Na,K-ATPase in customary experimental conditions (somewhat hypoxic) but stimulated it when solutions were oxygenated (14). Cross-talking pathways include not only the activation of protein kinase A seen in COS-7 cells (10), but also Na,K-ATPase inhibition mediated by increases in NO and cGMP seen in rat renal medullary tissue slices (38). The kinase-mediated regulation of calcineurin, a protein phosphatase, has been proposed to impact on the regulation of rat proximal tubule Na,K-ATPase by either protein kinase A or PKC (28). Although it may be argued that inhibition has been seen with purified rat renal Na,K-ATPase (6), in that case enzyme was purified from renal cortex, from which it is difficult to obtain very high specific activities. The enzyme purified from rat renal medulla in our prior work had a 2-3 times higher specific activity and thus may have had lower levels of contaminants that mediate Na,K-ATPase inhibition. The absence of inhibition mediated by other pathways in the cell cultures tested here may reflect their quiescent physiological status or their adaptation to continuous cell division.
The lack of a direct effect of phosphorylation on Na,K-ATPase activity leaves a central question unanswered: why does PKC phosphorylate the enzyme, and what are the real consequences for the cell? The existence of alternative pathways for Na,K-ATPase inhibition or activation does not mean that the documented phosphorylation has no physiological significance, only that a more complex mechanism may be involved. Phosphorylation may affect interaction of Na,K-ATPase with other intracellular modulators. Examples of such mechanisms are known. Phosphorylation of the calmodulin-binding domain of the plasma membrane Ca2+-pump by PKC reduced its interaction with calmodulin (39). Phosphorylation of Ser38 of the Ca2+-ATPase of cardiac sarcoplasmic reticulum (SERCA2a) by Ca2+/calmodulin-dependent kinase was originally thought to enhance the Vmax of Ca2+ uptake up to 60% (40). A recent report from the same laboratory did not confirm the conclusion, however, finding instead that phosphorylation of the SERCA2a inhibitor phospholamban increased the apparent affinity of SERCA2a for Ca2+, presumably through dissociation of phosphorylated phospholamban from Ca2+-ATPase (41). Meanwhile, the physiological role of phosphorylation of SERCA2a at Ser38 as well as of phosphorylation of Na,K-ATPase at Ser18 remains to be elucidated.
We thank Elkan Halpern (Department of Radiology, Massachusetts General Hospital) for the statistical analysis of Na,K-ATPase activity.