To obtain information about its function, single mutations were introduced and the mutants were expressed in yeast and analysed for enzymatic activity, ion recognition, and a/b subunit i
Trang 1The highly conserved extracellular peptide, DSYG(893–896), is a
critical structure for sodium pump function
Susanne Becker*, Heike Schneider* and Georgios Scheiner-Bobis
Institut fu¨r Biochemie und Endokrinologie, Fachbereich Veterina¨rmedizin, Justus-Liebig-Universita¨t Giessen, Germany
The peptide sequence DSYG(893–896) of the sheep sodium
pump a1 subunit is highly conserved among all K+
-trans-porting P-type ATPases To obtain information about its
function, single mutations were introduced and the mutants
were expressed in yeast and analysed for enzymatic activity,
ion recognition, and a/b subunit interactions Mutants of
Ser894 or Tyr895 were all active Conservative
phenylalan-ine and tryptophan mutants of Tyr895 displayed properties
that were similar to the properties of the wild-type enzyme
Replacement of the same amino acid by cysteine, however,
produced heat-sensitive enzymes, indicating that the
aromatic group contributes to the stability of the enzyme
Mutants of the neighbouring Ser894 recognized K+
with altered apparent affinities Thus, the Ser894fiAsp
mutant displayed a threefold higher apparent affinity
for K+ (EC50¼ 1.4 ± 0.06 mM) than the wild-type
enzyme (EC50¼ 3.8 ± 0.33 mM) In contrast, the mutant
Ser894fiIle had an almost sixfold lower apparent affinity for K+(EC50¼ 21.95 ± 1.41 mM) Mutation of Asp893
or Gly896 produced inactive proteins When an anti-b1 subunit immunoglobulin was used to co-immunoprecipitate the a1 subunit, neither the Gly896fiArg nor the Gly896fiIle mutant could be visualized by subsequent probing with an anti-a1 subunit immunoglobulin On the other hand, co-immunoprecipitation was obtained with the inactive Asp893fiArg and Asp893fiGlu mutants Thus, it might be that Asp893 is involved in enzyme conformational transitions required for ATP hydrolysis and/or ion translo-cation The results obtained here demonstrate the import-ance of the highly conserved peptide DSYG(893–896) for the function of a/b heterodimeric P-type ATPases
Keywords: Na+/K+-ATPase; a/b subunit interactions; immunoprecipitation; ouabain binding; thermal stability
The sodium pump (Na+/K+-ATPase, EC 3.6.3.9) is an
a/b-oligomeric enzyme embedded in the plasma membrane
of animal cells The enzyme hydrolyzes ATP to transport
three Na+ions out of the cell and two K+ions into the cell
Although ATP binding and ion occlusion seem to be tightly
connected to the a subunit, the overall catalytic activity,
defined as ATP-driven ion transport, requires the
inter-action of both a and b subunits of the enzyme
The subunits are known to interact with each other at
extracellular sites On the a subunit, a stretch of 26 amino
acids localized within the peptide loop that connects M7 and
M8 membrane-spanning domains (hereafter denoted L7/8)
was first identified as being important for interactions with
the b subunit [1] In a more detailed study, using a yeast
two-hybrid system, the SYGQ(894–897) sequence from
the 26-amino acid peptide was identified as an essential
component for a/b subunit interactions following replacement with four alanine residues: SYGQ(894– 897)fiAAAA(894–897) [2] (Note that the peptide number-ing corresponds to the sheep Na+/K+-ATPase a1 subunit.) Using the same system, either Ser894 or Tyr895 were identified as being essential for a/b interactions [3] Besides being important for interactions with the b subunit, the L7/8 region of the a subunit also seems to
be involved in ion translocation or recognition This is supported by the results of various investigations involving either enzymatic analysis of a subunit mutants [4] or metal ion-catalysed oxidative cleavage of a and b subunits, resulting in the loss of Rb+occlusion [5]
Recognition of K+or Na+was also found to depend on the b subunit structure, a finding that was based on subunit-substitution studies involving various chimeras between the sodium pump b subunit and the gastric proton pump b subunit [6,7] Thus, besides its role in stabilization of the
a subunit and its function as a vehicle for bringing the a subunit from the endoplasmic reticulum to the plasma membrane [8], the b subunit may be directly involved in ion recognition or transport
To better understand a/b interactions and their involve-ment in ion transport, we introduced mutations within the DSYG(893–896) sequence of the L7/8 peptide of the sheep a1 subunit, which is highly conserved among the hetero-dimeric P-type ATPases The mutant a1 subunits were expressed in yeast and investigated with respect to their enzymatic properties and their interaction with the coex-pressed b subunit The results described here demonstrate
Correspondence to G Scheiner-Bobis, Institut fu¨r Biochemie und
Endokrinologie, Fachbereich Veterina¨rmedizin,
Justus-Liebig-Uni-versita¨t Giessen, Frankfurter Str 100, D-35392 Giessen, Germany.
Fax: +49 641 9938189, Tel.: +49 641 9938180,
E-mail: Georgios.Scheiner-Bobis@vetmed.uni-giessen.de
Abbreviations: Na + /K + -ATPase, sodium- and potassium-activated
adenosine triphosphatase; NaCl/P i -T, phosphate-buffered saline
containing 0.1% (v/v)TweenTM20.
Enzymes: Na + /K + -ATPase (EC 3.6.3.9).
*Note: Both authors contributed equally to the scientific work
presented here.
(Received 9 June 2004, revised 21 July 2004, accepted 26 July 2004)
Trang 2that single mutations within the highly conserved
DSYG(893–896) peptide influence enzyme activity, enzyme
stability, interactions with K+, and assembly of the a and
b subunits
Experimental procedures
Vectors and strains
The shuttle vectors YhNa1, GhNb1 and pCGY1406ab,
used for the expression of a1 and b1 subunits of Na+/K+
-ATPase in the yeast Saccharomyces cerevisiae, have been
described previously [9,10] The pBluescriptTM KS
II+ (Stratagene, La Jolla, CA, USA) was used for the
introduction of mutations and the amplification of DNA in
Escherichia colistrain DH5aF¢ (Life Technologies,
Eggen-stein, Germany) Conditions for cell growth and media
compositions have been described previously [9]
Introduction of mutations
Mutations of Asp893, Ser894 and Gly896 were introduced
by inverse PCR [11] using, as a template, the plasmid
LPSKH5-7 [4] (a derivative of the pBluescriptII KS+) and
appropriate amplification primers (Roth, Karlsruhe,
Ger-many) shown in Table 1 PCR reaction mixtures of 100 lL
contained 1 lMmutation primer, 1 lMreverse primer, 1 ng
of the plasmid LPSKH5-7, 1.5 mMMgCl2, 2 U Tfl DNA
polymerase (Promega, Madison, WI, USA), 0.2 mMeach
dNTP, and the appropriate amount of buffer provided by
the supplier After 20 PCR cycles, the amplification product
was isolated by agarose gel electrophoresis, treated with T4
DNA polymerase (Promega) to remove the dA overhang
produced by the Tfl DNA polymerase, and recircularized by
the use of T4 DNA ligase (MBI Fermentas, Vilnius,
Lithuania) After amplification in E coli, the plasmids were
tested by restriction analysis with AseI (all restriction
enzymes purchased from MBI Fermentas), and DNA sequencing was carried out according to Sanger et al [12] using T7 DNA polymerase (Amersham Life Science, Little Chalfont, Bucks., UK) and [35S]dATP (ICN Radiochemi-cals, Irvine, CA, USA) A 589 bp MunI/BglII fragment, fully sequenced to exclude additional unintended mutations, was removed from the LPSKH5-7 plasmid and inserted into the MunI/BglII site of the yeast expression vector, pCGY1406a The pCGY1406ab vectors now carrying the desired mutations in the a1 subunit cDNA (Table 1) and the wild-type cDNA for the sodium pump b1 subunit [9] were used to transform yeast [13]
Mutations of Tyr895 were introduced by PCR using the QuikChangeTMSite-Directed Mutagenesis Kit (Stratagene Europe, Amsterdam, the Netherlands) and the primers shown in Table 1 The template plasmid was pBluescriptTM
KS II+ containing, in the multiple cloning site, a 1528 bp BglII/AflII fragment of the a1 subunit cDNA The protocol
of the provider was used for the amplification of the mutant cDNA After complete automatic sequencing, the 1528 bp BglII/AflII fragments carrying the desired mutations were ligated back into the yeast vector and used for yeast transformations
Isolation of membranes containing native or mutant sodium pumps
The methods involved in the isolation of membranes from yeast cells, and for the preparation of SDS-treated microsomes enriched in the sodium pump, have been described previously in great detail [14,15] The Na+/
K+-ATPase activity in the isolated fractions was deter-mined by a coupled spectrophotometric assay in the presence or absence of 1 mM ouabain [9,16] The protein concentration of the microsomal preparations was deter-mined by the method of Lowry [17] using BSA as a standard
Table 1 Primers used for mutations.
Primers used for mutations by inverse PCR
Wild type 5 ¢- GTGGAGGACAGCTATGGGCAGCAG - 3 ¢
Asp893fiArg 5 ¢- GTGGAG CG CAGCTATGGGCAGCAG - 3 ¢
Asp893fiGlu 5 ¢- GTGGAGGA G AGCTATGGGCAGCAG - 3 ¢
Asp893fiAla 5 ¢- GTGGAGG C CAGCTATGGGCAGCAG - 3 ¢
Ser894fiAsp 5 ¢- GTGGAGGAC GA CTATGGGCAGCAG - 3 ¢
Ser894fiIle 5 ¢- GTGGAGGACA T CTATGGGCAGCAG - 3
Gly896fiArg 5 ¢- GTGGAGGACAGCTAT A GGCAGCAG - 3 ¢
Gly896fiIle 5 ¢- GTGGAGGACAGCTAT ATC CAGCAG - 3 ¢
Second primer for all of the above 5 ¢- GTC A TT A ATCCAACGGTCATCCCA - 3 ¢ a
Primers used for mutations by PCR and the QuikChangeTMSite-Directed Mutagenesis Kit
Wild type 5 ¢- GTGGAGGACAGCTATGGGCAGCAGTGG - 3 ¢
Tyr895fiCys 5 ¢- GTGGAGGACAGCT G TGGGCAGCAGTGG - 3 ¢
Second primer 5 ¢- CCACTGCTGCCCA C AGCTGTCCTCCAC - 3 ¢
Tyr895fiPhe 5 ¢- CGATGTGGAGGACAGCT T TGG C CAGCAGTGGACCTATG - 3 ¢ Second primer 5 ¢- CATAGGTCCACTGCTG G CCA A AGCTGTCCTCCACATCG - 3 ¢ b
Tyr895fiTrp 5 ¢- CGATGTGGAGGACAGCT GG GGGCAGCAGTGGACC - 3 ¢ Second primer 5 ¢- GGTCCACTGCTGCCC CC AGCTGTCCTCCACATCG - 3 ¢
a
Silent mutations produce a diagnostic restriction site for AseI.bThe silent mutation produces a diagnostic restriction site for EaeI.
Trang 3Co-immunoprecipitation of a1 and b1 subunits
by an antibody against b1 subunits
Unless otherwise specified, all of the following steps were
carried out at 4C A total of 1 mg of microsomes enriched
in Na+/K+-ATPase [9] was centrifuged for 20 min at
13 000 g and subsequently suspended in 1.5 mL of a buffer
consisting of 50 mMTris/HCl, pH 7.5, 150 mMNaCl, 1 mM
Na2EDTA, 0.2% BSA (w/v), 1% Triton X-100 (w/v),
2 mM dithiothreitol, 0.2 mMphenylmethanesulfonyl
fluor-ide, 0.5 mgÆmL)1 leupeptin and 0.7 mgÆmL)1 pepstatin
Thereafter, 0.5 lL of the anti-b1 immunoglobulin (Alexis
Corporation, Gru¨nberg, Germany) was added to the
solu-tion and the reacsolu-tion was allowed to proceed for 5 h under
continuous, gentle shaking Then, 20 lL of protein G–
Sepharose 4B (Sigma, Deisenhofen, Germany) was added
and the incubation extended for an additional 16 h The
solution was then centrifuged at 900 g for 5 s and the
supernatant removed by aspiration The pellet was then
subjected to a washing routine involving alternating steps of
careful suspension of the complex in the extraction buffer
followed by centrifugation at 900 g for 5 s at 4C, as
described in the protocol of Tamkun & Fambrough [18] The
Sepharose beads were first washed three times in 1.5 mL of
buffer A [150 mMNaCl, 50 mM Tris/HCl, pH 7.5, 1 mM
Na2EDTA, 0.5% Triton X-100 (w/v)], then once with
1.5 mL of a solution consisting of 300 mM NaCl, 50 mM
Tris/HCl, pH 7.5, 0.1% SDS (w/v), 0.1% Triton X-100
(w/v), then once with 1.5 mL of 1MNaCl, 50 mMTris/HCl,
pH 7.5, 0.5% Triton X-100 (w/v), twice with 1.5 mL of
buffer A, and finally with 1.5 mL of 1% Triton X-100 (w/v)
Thereafter, the antigen/antibody/Sepharose bead
com-plex was suspended in 20 lL of sample buffer consisting
of 250 mM Tris/HCl, pH 6.8, 10% SDS (w/v), 10%
2-mercaptoethanol (v/v), 1 mgÆmL)1 Coomassie Brilliant
Blue and 25% glycerol (v/v), and heated for 2 min at
100C The Sepharose beads were pelleted by
centrifuga-tion and 20 lL of the supernatant was mixed with 20 lL of
8Murea and heated for 5 min at 70C Proteins in 20 lL
of this mixture were separated by electrophoresis on a
polyacrylamide gel containing 10% polyacrylamide and
0.3% N,N¢-methylene-bisacrylamide [19] The gel was then
equilibrated for 30 min in 0.1% SDS (w/v), 12.5 mMTris,
96 mM glycine, 20% methanol (v/v), pH 8.4, and blotted
onto a nitrocellulose membrane (Schleicher & Schuell,
Dassel, Germany) for 2 h at 2.5 AÆcm)2[20]
The nitrocellulose membrane was first blocked overnight
in NaCl/Pi(PBS) containing 0.1% Tween 20 (v/v) (NaCl/
Pi-T) and 5% (w/v) nonfat dried milk, then washed three
times (10 min each wash) with NaCl/Pi-T, and subsequently
incubated for 2 h with either an anti-a1 mAb (diluted
1 : 300 in NaCl/Pi-T) or an anti-b1 mAb (diluted 1 : 1000
in NaCl/Pi-T), both raised in mouse (Alexis Corporation)
The detection of antibody-bound a1 or b1 subunits was
carried out by incubating the nitrocellulose membrane for
90 min with an alkaline phosphatase-conjugated
anti-mouse IgG (SeroTec, Oxford, UK; diluted 1 : 2500 in
NaCl/Pi-T) and subsequently adding the alkaline
phos-phatase substrates 5-bromo-4-chloro-3-indoyl-phosphate
(Molecular Probes, Eugene, OR, USA) for the detection
of a1 subunits, or Nitro Blue tetrazolium (Serva,
Heidel-berg, Germany) for the detection of b1 subunits, according
to the corresponding protocols of the providers The chromogenic reaction was interrupted by adding 20 mM EDTA in NaCl/Pi-T
Metabolic labelling and immunoprecipitation
of a1 subunits Single yeast colonies were grown in 5 mL of selective minimal medium for 16 h at 30C to mid-logarithmic phase [i.e an attenuance (D), at 600 nm, of 0.5–1] A total of 2.5· 106cells, corresponding to a D600 of 2.5, were then centrifuged at room temperature for 5 min at 1800 g and washed with sterile H2O This procedure was repeated Thereafter, cells were suspended in 1.25 mL of minimal medium and incubated at 30C for 60 min with continu-ous, gentle shaking Cells were pelleted again, as described above, and subsequently suspended in 1.25 mL of the minimal medium now containing 100 lCiÆmL)1 of [35S]methionine Incubation was allowed to proceed with gentle shaking at 30C for a further 30 min
Labelled cells were centrifuged, as described above, and washed with 1 mL of ice-cold water After a second wash, the pelleted cells were suspended in 100 lL of an extraction buffer comprising 50 mMTris/HCl, pH 7.4, 150 mMNaCl,
10 mM MgCl2, 1 mM EDTA, 10% glycerol (w/v), 0.2% BSA (w/v), 2 mM dithiothreitol, 0.2 mM phenylmethane-sulfonyl fluoride, 0.5 mgÆmL)1and 0.7 mgÆmL)1pepstatin (phenylmethanesulfonyl fluoride, leupeptin and pepstatin were obtained from Boehringer Ingelheim, Heidelberg, Germany) Then, 100 lL of glass beads (0.25–0.3 mm in diameter) were added and the cells were broken by 10 bursts
of 20 s of vigorous mixing at the highest speed in a vortex mixer, each time followed by a 40 s cooling phase on ice The supernatant was then transferred to a new vial and the glass beads washed twice with 100 lL of the extraction buffer The combined supernatants ( 300 lL) were cen-trifuged at 7500 g for 20 min at 4C to remove debris After adjusting the radioactivity in the supernatants with extraction buffer to be the same in a final volume of 750 lL, 0.5 lL of an anti-a1 immunoglobulin (Alexis Corporation) was added and the solution was incubated with gentle shaking for 16 h at 4C Subsequently, 25 lL of a protein
G + agarose suspension (Sigma) were added and incuba-tion was continued for another 4 h
Thereafter, the antigen/antibody/protein-G + agarose complex was sedimented at 4C by centrifugation for 5 s
at 900 g, and the supernatant was removed by aspiration The pellet was then washed, essentially as described above, for the immunoprecipitation using the anti-b immuno-globulin
After the last wash, the antigen/antibody/protein
G + agarose complex was equilibrated with 20 lL of a buffer containing 125 mMTris/HCl, pH 6.8, 4Murea, 5% SDS (w/v), 5% 2-mercaptoethanol (v/v), 12.5% glycerol (v/ v), 0.5% of a solution of 0.1% ethanol saturated with bromophenol blue solution (v/v), and heated at 70C for
15 min Then, the protein G + agarose was removed by centrifugation at 900 g for 5 s at 4C and solubilized proteins in the supernatant were separated by SDS/PAGE
on gels containing 10% polyacrylamide and 0.3% N,N¢-methylene-bisacrylamide, prepared according to Laemmli [19] After electrophoresis, proteins in the gel were stained
Trang 4with Coomassie Brilliant Blue (Serva) and, after drying,
exposed for 2 days at)80 C to a Kodak X-Omat X-ray
film
Immunodetection of wild-type and Tyr895 mutant
a1 subunits by Western blotting
A total of 50 lg of SDS-extracted yeast membrane proteins
[9], containing either native or mutant Na+/K+-ATPase,
was suspended in 10 lL of loading buffer and separated by
SDS/PAGE following established protocols [19]
Mem-brane extracts from untransformed yeast served as the
negative control Protein was then transferred onto
nitro-cellulose membranes following the instructions provided by
the commercially available ECL Western blotting system
PRN 2180 kit (Amersham Pharmacia Biotech, Freiburg,
Germany) Following the same protocol, the a1 or b1
subunit of the Na+/K+-ATPase was detected using specific
antibodies (Alexis Corporation) raised in mice, each used
at a dilution of 1 : 2500 The secondary antibody was a
horseradish peroxidase-coupled anti-mouse IgG provided
by the kit
Binding of [3H]ouabain under various conditions
To obtain a relative affinity for ATP, a total of 250 lg of
microsomal protein isolated from yeast cells expressing
either wild-type or mutant sodium pumps was incubated at
30C for 5 min in a mixture containing 10 mMTris/HCl,
pH 7.4, 50 nM [3H]ouabain, 50 mM NaCl, 5 mM MgCl2
and various concentrations of ATP (Tris salt) The total
volume of each sample was 250 lL Thereafter, the protein
was pelleted by centrifugation at 13 000 g for 2 min,
washed twice with H2O at 4C, and dissolved in 200 lL
of 1MNaOH by incubation at 80C for 10 min After the
addition of 200 lL of 1M HCl, the samples were mixed
with 3.5 mL of scintillation cocktail (Roth) and counted for
radioactivity
To obtain a relative affinity of the enzyme and its mutants
for Na+, this last experiment was performed using a
constant concentration of 100 lM ATP (Tris salt) and
varying the concentration of Na+ Before the microsomes
were used, however, they were washed twice in 1 mL of
10 mM Tris/HCl, pH 7.4, to remove any Na+ from the
microsomes storage buffer, which was composed of 25 mM
imidazole/1 mM Na2EDTA, pH 7.4 [9,15] All other
con-ditions were unchanged
To obtain a relative affinity for K+, the enzyme or its
mutants were incubated in 10 mM Tris/HCl, pH 7.4, for
60 min with 50 nM [3H]ouabain, 5 mM phosphate (Tris
salt), 5 mMMgCl2and various concentrations of KCl The
other conditions were as described above
Thermal stability of wild-type Na+/K+-ATPase and
mutants
This experiment was carried out according to a previously
published protocol [21] Briefly, a total of 125 lg of
microsomal protein was incubated on ice to serve as a
control An equivalent amount of protein was heated for
5 min at 50C Then, both samples were incubated for an
additional 15 min on ice followed by 30 min at 30C with
5 mMphosphate (Tris salt), 5 mMMgCl2, 10 mMTris/HCl,
pH 7.4, and 50 nM [3H]ouabain The total volume was
500 lL Bound radioactivity was determined as described above
Results
Na+/K+-ATPase activity Yeast membrane preparations contain endogenous ATP-ases Unlike the mammalian sodium pump, these are ouabain insensitive Therefore, in order to distinguish yeast endogenous ATPases from heterologously expressed Na+/
K+-ATPase, ATPase activity was determined in SDS-extracted membrane preparations in the presence or absence
of 1 mM ouabain In Fig 1A, it can be seen that no significant Na+/K+-ATPase activity was detected in mem-brane preparations from cells expressing any of the Asp893
or Gly896 mutants The same applied for membranes from nontransformed cells A significant, ouabain-sensitive ATPase activity was, however, detected in membrane preparations from cells expressing the wild-type enzyme
Fig 1 Na+/K+-ATPase activity in yeast membrane preparations (A) Inactive mutants (B) Active mutants Na+/K+-ATPase activity was determined, as described in the Experimental procedures, by a coupled spectrophotometric assay as the ouabain-sensitive fraction of all ATPase activity that is present in the preparations All results represent the mean ± SD of three independent experiments Membrane prep-arations from cells expressing either Asp893fiArg or Gly896fiIle did not display any Na + /K + -ATPase-specific activity The unit 1 mU is defined as the enzymatic activity that hydrolyzes 1 nmol of ATP in
1 min at 37 C NT, nontransformed yeast cells *Significantly lower (P < 0.05) than the activity obtained with the wild-type preparation.
Trang 5and all of the Ser894 or Tyr895 mutants (Fig 1B) While the
Ser894 mutants and Tyr895fiCys or Tyr895fiPhe mutants
displayed ATPase activities comparable to that of the
wild-type enzyme, the activity of the Tyr895fiTrp mutant was
significantly reduced
Coimmunoprecipitation of a1 and b1 subunits
by an anti-b1 immunoglobulin
As the mutations are all localized within a peptide sequence
of the a subunit that has been shown to interact with the b
subunit [1,3], it was important to evaluate whether the
observed loss of ATPase activity shown in Fig 1A was
caused by the loss of a/b interactions
Following a well-established protocol for
co-immuno-precipitation of a1 and b1 by using an anti-b1
immunoglobulin [18], it was possible to demonstrate
co-immunoprecipitation of the wild-type a1 with the b1
subunit (Fig 2A) Co-immunoprecipitation was also
observed with the inactive Asp893fiArg and Asp893fiGlu
mutants of the a1 subunit (Fig 2A) In contrast, a1 did not
co-immunoprecipitate with b1 when membrane
prepara-tions from yeast expressing the inactive forms of the a1
subunit (Gly896fiArg or Gly896fiIle) were used (Fig 2A)
The b1 subunits, however, were found, as expected, in all
immunoprecipitates except in membranes from
nontrans-formed cells (Fig 2B), verifying that the absence of
Gly896fiArg or Gly896fiIle in Fig 2A was not caused
by the lack of b1 subunit expression or by degradation of
this protein Nevertheless, the quantities of the b1 subunits
detected varied among the lanes, indicating either different levels of expression or variations in experimental recovery Hence, in order to obtain a value that relates to the abundance of co-immunoprecipitated a1 subunits to the precipitated b1 subunits, the protein bands in Fig 2B were analysed by densitometry using the image analysis system of Biostep (Jahnsdorf, Germany) By setting the abundance
of b1 subunits in lane 4 (Asp894fiGlu) of Fig 1B to 100%, the b1 abundance in lane 1 (wild type), lane 3 (Asp893fiArg), lane 5 (Gly896fiArg) and lane 6 (Gly896fiIle) were 91%, 40%, 54% and 51%, respectively Lane 2 (nontransformed cells) was considered to represent the background signal
In an analogous way, the a1 subunit detected in lane 1 (wild type) of Fig 2A was set to represent the 100% value
In comparison, the relative abundance of the Asp893fiArg mutant a1 subunit (lane 3) only accounted for 47% of this value The equivalent value for the Asp894fiGlu (lane 4) mutant was 90% Lane 2 was set to indicate the back-ground No protein bands were detected in lanes 5 and 6, containing the Gly896 mutants, by the auto-detect function
of the software program
The values obtained from the densitometric scans were used to define the stoichiometry between co-immunopre-cipitated a1 and b subunits by forming a quotient between the relative abundance of these subunits For all the a1/b heterodimers that co-immunoprecipitated (wild type, Asp894fiArg/b subunit, or Asp894fiGlu/b subunit), the quotient of a1 abundance to b1 abundance was approxi-mately 1, indicating proportional expression and recovery levels
Detection of [35S]methionine-labelled wild-type and mutant a1 subunits expressed in yeast
Our inability to co-immunoprecipitate the Gly896fiArg and Gly896fiIle mutants of the a1 subunits (Fig 2A) might have been caused by a lack of expression Therefore, yeast cells transformed with plasmids coding for Asp893fiArg, Asp893fiGlu, Gly896fiArg or Gly896fiIle mutants of the a1 subunit of the sodium pump were metabolically labelled with [35S]methionine and subsequently used to isolate membrane fractions Mutant or wild-type a1 subunits were isolated from this mixture by immunoprecipitation with an anti-a1 immunoglobulin As shown in Fig 3, it is apparent Fig 2 Coimmunoprecipitation of inactive mutant a1 subunits with an
antibody against b1 subunits (A) As with the wild-type a1 subunit (lane
1), the inactive Asp893fiArg (lane 3) and Asp893fiGlu (lane 4)
mutants co-immunoprecipitate with the b1 subunits, indicating that
a/b assembly is not affected by these two mutations In contrast, the
Gly896fiArg (lane 5) or Gly896fiIle (lane 6) mutants do not
co-immunoprecipitate with the b1 subunits The co-co-immunoprecipitated
subunits were visualized in a Western blot using an antibody against
the a1 subunit as a primary antibody and an alkaline
phosphatase-conjugated anti-immunoglobulin G (IgG) as a secondary antibody,
with 5-bromo-4-chloro-3-indoyl-phosphate as a chromogenic alkaline
phosphatase substrate (B) All b1 subunits precipitate as an antigen/
antibody/protein G–Sepharose complex and can be detected in the
Western blot using an antibody against b1 subunits as a primary
antibody and the same secondary antibody mentioned above The
chromogenic substrate of alkaline phosphatase used here was Nitro
Blue tetrazolium Neither a1 nor b1 subunits were visualized in
membranes from nontransformed yeast cells (lane 2).
Fig 3 Immunoprecipitation of metabolically labelled, inactive mutant a1 subunits [35S]Methionine-labelled proteins were immunoprecipi-tated by an anti-a1 immunoglobulin, as described in the Experimental procedures After SDS/PAGE, labelled proteins were detected by autoradiography Wild-type (lane 1) or mutant a1 subunits (Asp893fiArg, lane 3; Asp893fiGlu, lane 4; Gly896fiArg, lane 5; Gly896fiIle, lane 6) were present in the membrane preparations from transformed cells A labelled protein of 110 kDa was not found in membranes from nontransformed cells (lane 2).
Trang 6that a labelled protein of 110 kDa was found only in
membrane preparations from yeast cells expressing either
the wild-type or the mutant a subunits As the sodium pump
a1 subunit displays a relative molecular mass of 110 kDa,
and because no similar protein was detected in membrane
preparations from nontransformed cells (lane 2), it is very
likely that the precipitated proteins are the wild-type a1
subunits and its mutants Furthermore, these results
dem-onstrate that the a1 subunits are expressed at approximately
the same level
Immunodetection of the Tyr895 mutants by Western
blotting
The Tyr895 mutants are all active (Fig 1) Nevertheless,
the Tyr895fiTrp mutant displayed a significantly
de-creased activity when compared to the wild-type enzyme
and to the mutants Tyr895fiCys or Tyr895fiPhe In
order to evaluate whether this difference was caused by
different expression levels, SDS-extracted membranes of
cells expressing either of these mutants or the wild-type
enzyme were probed in a Western blot with antibodies
against the a1 or b subunit of the sodium pump
As shown in Fig 4, the abundance of the Tyr895fiTrp
mutant is considerably reduced when compared to the
abundance of the wild-type enzyme and the other
mutants Comparison of the signals for the wild-type
and the Tyr895fiCys and Tyr895fiPhe mutants indicates
similar expression of these enzymes, as determined by
optical densitometry using the image analysis system of
Biostep, described above Using the same system, the
quotient of a1 abundance to b1 abundance was
approxi-mately 1 for all the a1/b heterodimers that were detected
in this Western blot (wild type/b subunit, Tyr895fiCys/b
subunit, Tyr895fiPhe/b subunit or Tyr895fiTrp/b
sub-unit), indicating proportional expression and recovery
levels
Binding of [3H]ouabain as a function of ATP concentration
In the presence of ATP, Na+, Mg2+and [3H]ouabain, the sodium pump a1 subunit forms a stable [phosphoen-zymeÆ[3H]ouabain] complex that can be easily measured [4] Figure 5 shows the binding of [3H]ouabain to yeast membrane preparations as a function of the ATP concen-tration in the presence of Na+and Mg2+ After 5 min of incubation, [3H]ouabain binding was detectable with mem-branes containing the wild-type enzyme or either of the Ser894 mutants The EC50 for ATP was 0.77 ± 0.11 lM for the wild-type enzyme and 0.46 ± 0.10 lM or 0.94 ± 0.11 lM for the Ser894fiAsp and Ser894fiIle mutant enzymes, respectively
Similar results were obtained with the Tyr895 mutants (Table 2) These values and the values obtained with the Ser894 mutants are all in good agreement with KDvalues determined for ATP binding to sodium pumps from mammalian tissues [22]
Binding of [3H]ouabain as a function of Na+ concentration
When [3H]ouabain binding was measured as a function of the Na+ concentration in the presence of 100 lM ATP,
Na+-enhanced [3H]ouabain binding to the wild-type enzyme, with an EC50 of 1.26 ± 0.38 mM, was observed (Fig 6) The corresponding values determined with mem-branes containing either the Ser894fiAsp or Ser894fiIle mutants were 1.46 ± 0.58 mM and 1.56 ± 0.58 mM, respectively, and indicate the action of Na+on cytosolic sites [23,24] No specific binding was seen under these conditions with membranes from nontransformed cells The Tyr895 mutants displayed similar sensitivities towards Na+ (Table 2)
Fig 4 Immunodetection of active Tyr895 mutants The Western blot
experiment demonstrates reduced expression for the Tyr895fiTrp a1
(A) and b1 subunits (B), thus explaining the reduced Na+/K+-ATPase
activity observed with membrane preparations containing this mutant.
The wild-type, Tyr895fiCys or Tyr895fiPhe a1 (A) or b1 subunits (B)
were expressed at comparable levels The multiple protein bands reflect
various glycosylation states of the b1 subunit [9].
Fig 5 Binding of [3H]ouabain as a function of the ATP concentration Yeast membranes from cells expressing either the wild-type (h) or the Ser894fiAsp (s) and Ser894fiIle (.) mutants were incubated for
5 min with 50 n M [ 3 H]ouabain, 50 m M NaCl, 5 m M MgCl 2 and var-ious concentrations of ATP (Tris salt; see the Experimental procedures for details) ATP promotes [3H]ouabain binding to wild-type and mutant enzymes with similar EC 50 values of 1 l M
Trang 7Binding of [3H]ouabain as a function
of K+concentration
In order to obtain a value for the relative affinity for K+of
the wild-type enzyme and the Ser894fiAsp or Ser894fiIle
mutants, the binding of [3H]ouabain to these enzymes in the
presence of phosphate and Mg2+was measured as a
function of the K+concentration Under these conditions,
K+caused a reduction in [3H]ouabain binding to the wild-type enzyme, showing an EC50of 3.8 ± 0.33 mM(Fig 7) The corresponding value obtained with the Ser894fiAsp mutant was 1.4 ± 0.06 mM A similar experiment carried out with the Ser894fiIle mutant yielded an EC50 of 21.95 ± 1.41 mM, and the Tyr895 mutants revealed K+ sensitivities comparable to that of the wild-type enzyme (Table 2)
Table 2 Properties of the Tyr895 mutants.
Mutant
EC 50 for ATP
(l M )
EC 50 for Na +
(m M )
EC 50 for K +
(m M )
Detection of a and b subunits in the Western blot
Na + /K + -ATPase activity (mUÆmg protein)1)
Tyr895fiCys 0.72 ± 0.3 1.44 ± 0.11 3.74 ± 1.14 + + 13.16 ± 1.95
Tyr895fiPhe 1.07 ± 0.15 1.36 ± 0.71 1.75 ± 0.78 + + 13.30 ± 1.61
Tyr895fiTrp 0.8 ± 0.2 1.87 ± 0.31 2.78 ± 0.59 + + 7.60 ± 1.89
Fig 6 Binding of [3H]ouabain as a function of Na+concentration (A)
Yeast membranes from cells expressing the wild-type (h), the
Ser894fiAsp (s) or the Ser894fiIle (.) mutants were incubated for
5 min with 50 n M [ 3 H]ouabain, 5 m M MgCl 2 , 100 l M ATP (Tris salt)
and various concentrations of Na+ The latter promotes [3H]ouabain
binding to wild-type and mutant enzymes (B) The double reciprocal
plot of the values obtained above reveals that Na + promotes ouabain
binding to the wild-type and mutant enzymes with similar EC 50 values
of 1.5 m
Fig 7 Inhibition of [ 3 H]ouabain binding by K + Yeast membranes from cells expressing the wild-type (h), the Ser894fiAsp (s) or the Ser894fiIle (.) mutants were incubated in 10 m M Tris/HCl, pH 7.4, for 60 min with 50 n M [ 3 H]ouabain, 5 m M phosphate (Tris salt), 5 m M MgCl 2 , and various concentrations of KCl The other conditions were
as described above In all cases, K+leads to a reduction of ouabain binding (B) By plotting the reciprocal binding of [ 3 H]ouabain against the K+concentration, the EC 50 for K+can be obtained from the intercept of the straight lines with the abscissa.
Trang 8Thermal stability of the Tyr895 mutants
The binding of [3H]ouabain to membrane preparations that
have been preheated at 50C can serve as a measure for the
stability of the a/b heterodimer that forms the catalytically
active enzyme [21] After preheating the wild-type sodium
pump at 50C for 5 min, the enzyme was capable of
binding only 62.0 ± 8.0% of the ouabain that was bound
by the same control enzyme that had been incubated for the
same length of time on ice (Fig 8) The Tyr895 mutants
displayed a similar behaviour Thus, after preheating, the
Tyr895fiPhe mutant bound 67.6 ± 16.0% of the ouabain
bound by the control, while the corresponding value with
the Tyr895fiTrp mutant was 51.4 ± 5.9% Therefore,
these two conservative mutants did not convert into
more temperature-sensitive forms The nonconservative
Tyr895fiCys mutant, however, was able to bind only
36.5 ± 6.5% of the ouabain that was bound by the
unheated control, indicating a higher thermal sensitivity
than the wild-type enzyme or its conservative Tyr895
mutants (Fig 8) This result, which shows a significantly
lower value than that obtained with the wild-type enzyme,
points towards an involvement of the aromatic group in
enzyme stabilization (Fig 8)
Discussion
The extracellularly localized peptide of the sodium pump a
subunit that connects the M7 and M8 membrane-spanning
domains (L7/8) contains 26 amino acids that are important
for assembly with the b subunit [1] Corresponding peptides
of other K+-transporting ATPases seem to play a
compar-able role Results from studies of chimeric constructs
formed by replacing the 26 amino acids of the rat a3
subunit (Asn886)Ala911) with the corresponding region
from either the gastric (Gln905-Val930) or the distal (Asn908-Ala933) colon H+/K+-ATPase of the rat demon-strated interaction of the chimeras with the H+/K-ATPase
b subunits and helped to identify Val904, Tyr898, and Cys908 of the sodium pump a3 subunit as important for assembly with the b subunit [21,25] In addition, the L7/8 loop confers sensitivity towards specific inhibitors of P2-type ATPases, as shown using chimeric constructs between the
a subunits of Na+/K+-ATPase and the gastric H+/K+ -ATPase [26] This segment was also demonstrated to be important for ion conduction, as shown for Asp884 and Asp885 mutants [4], to affect interactions with Na+or K+,
as demonstrated with a/b heterohybrids [7,27], or to be associated with the loss of Rb+occlusion, as shown by
Cu2+-catalysed oxidative cleavage near His875 [5] Finally, the importance of the L7/8 loop is underlined by the high degree of homology of the 26-amino acid peptide seen in all
K+-transporting P2-type ATPases (Fig 9) The sequence DSYG(893–896) is – with one exception seen in Hydra – absolutely conserved in all of the K+-transporting P2-type ATPases, indicating that some important role for this peptide was conserved in the course of evolution Thus, to investigate the function of this tetrapeptide, single mutations were introduced within this area of the a1 subunit The investigation of these mutants revealed that each of the altered amino acids had an impact on the enzyme proper-ties, although in somewhat different ways In general, however, we can distinguish between catalytically inactive and catalytically active mutants
Catalytically inactive mutants All mutants of Asp893 and Gly896 were inactive (Fig 1) Lack of expression might be one plausible reason for the lack of measurable activities Alternatively, for some of the mutants it could be that a1 and b1 subunit interactions were disturbed, as all of the mutations are within the 26-amino acid peptide that was found to be important for assembly [1]
A possible change in expression caused by the mutations was investigated by applying an immunoprecipitation protocol after metabolic labelling of the wild-type and mutant a1 subunits The immunoprecipitation method was preferred over a standard Western blot to prevent possible degrada-tion of the mutant a1 subunits, which are known to be a target for proteases unless they form heterodimers with the b subunit [28,29] This experiment, however, verified that all Asp893 and Gly896 mutants were expressed similarly to that
of the wild-type a1 subunit (Fig 3) Based on that result, the loss of activity is not caused by the lack of expression
A possible loss of a/b interaction because of the mutations was investigated by applying a co-immunopre-cipitation protocol, described previously [18], which verified
a loss in a/b interactions when Gly896 is replaced with either Arg or Ile This may provide an explanation for the lack of any detectable enzymatic activity with these two mutants, as formation of the a/b-heterodimer is a presupposition for
Na+/K+-ATPase activity [9]
Although rather unusual, single mutations that entirely change, or even obliterate, protein–protein interactions are not that uncommon An arginine residue was found to
be absolutely essential for oligomerization of ribulose-1, 5-bisphosphate carboxylase [30] A similar experience was
Fig 8 Thermal stability of the wild-type ATPase and the Tyr895
mutants When the wild-type sodium pump is heated for 5 min at
50 C, the enzyme binds only 62.0 ± 8.0% of the [ 3
H]ouabain bound
by the unheated control The Tyr895fiPhe and the Tyr895fiTrp
mutants behave similarly Ouabain binding to the nonconservative
Tyr895fiCys mutant, however, is only 36.5 ± 6.5% of that obtained
with the unheated control under these conditions This result is
signi-ficantly lower (*P < 0.05) than that obtained with the wild-type
enzyme For all measurements n ¼ 3, error bars represent ± SD.
Trang 9also observed with sodium pump/proton pump hybrids
where the amino acids Tyr898, Val904, and Cys908 of the
a3 subunit of the sodium pump (corresponding to Tyr901,
Val907 and Cys911 of the sheep a1 subunit used in the
current investigation) were found to be important for
assembly with the b subunit [21]
On the other hand, co-immunoprecipitation was
ob-tained with the inactive Asp893fiArg and Asp893fiGlu
mutants that was similar in relative amount to the
co-immunoprecipitated wild-type a1 subunit (Fig 2) Thus, the
very strong negative effects of the Asp893 mutations on
enzyme activity cannot be explained by a lack of association
of a and b subunits One possibility is that the mutation of
this amino acid results in reduced affinity for ouabain, and
therefore binding of [3H]ouabain was not detectable under
the experimental conditions applied in this investigation
Although this explanation cannot be excluded, two facts
speak against it: first, in the coupled spectrophotometric
assay, preincubation with 1 mMouabain did not result in
different ATPase activities for the assays performed in the
presence or absence of the glycoside Unless ouabain
sensitivity is completely abolished by the mutation, 1 mM
ouabain should be sufficient to detect some inhibition
Second, comparison of the primary sequences reveals that
the aspartic acid investigated is highly conserved in all K+
-transporting P-type ATPases, regardless of whether or not
they bind ouabain (Fig 8) This latter fact suggests that this
highly conserved aspartic acid is unlikely to be directly
involved in ouabain binding As Asp893 is within an area of
the protein that interacts with the b subunit [1,2], and
because the b subunit has been shown not only to influence
enzyme properties [6,26,27] but also to be absolutely essential for catalytic activity [9,31], it might be that Asp893 is involved in enzyme conformational transitions required for ATP hydrolysis and/or ion translocation This assumption is difficult to investigate further, however, because in various experiments that were not shown here
it was not possible to detect any partial activities that are typical for the sodium pump [32] with any of the Asp893 mutants
Catalytically active mutants Mutants of Ser894 and Tyr895 were all active (Fig 1) This was a rather unexpected result, because, by using the two-hybrid system, previous reports had identified these amino acids as being critical for a/b assembly and for enzyme activity [3] In addition, interactions of the Tyr895 mutants with cations or ATP were not altered when compared with the properties of the wild-type enzyme (Table 2) The fact, however, that in Hydra the amino acid corresponding to Tyr895 is a phenylalanine (Phe910; Fig 9) indicates that although Tyr895 might not be a critical amino acid, the presence of an aromatic group might be important for enzyme stability This seems to be the case, as the Tyr895fiCys mutant displayed a significantly higher thermal sensitivity than the wild-type enzyme or the Tyr895fiPhe and Tyr895fiTrp mutants (Fig 8) The Ser894 mutants were all active and displayed, in most cases, properties similar to those of the wild-type enzyme Their interactions with ATP or Na+ were essentially unaffected (Figs 5 and 6) Nevertheless, the interactions of
Fig 9 Comparison of primary structures The 26-amino acid peptide of the sodium pump a1 subunit that is known to be important for the assembly with b subunits is compared with equivalent areas of other K + -transporting P-type ATPases The Asp893, Ser894 and Gly896 residues investigated here (underlined) correspond to highly conserved aspartic acid, serine and glycine residues present in all K+-transporting P-type ATPases Similar amino acids are not found in the Ca2+ATPases or in the Na+ATPases The Tyr895 is replaced with a phenylalanine (Phe910) in Hydra *The numbering here takes into consideration the deduction of a 5-amino acid propeptide.
Trang 10the mutants Ser894fiAsp and Ser894fiIle with K+were
clearly different from these of the wild-type enzyme As
shown in Fig 6, K+ reduces [3H]ouabain binding to
membranes containing the wild-type enzyme, with an EC50
of 3.8 ± 0.33 mM The membranes containing the
Ser894fiAsp mutants, however, display an EC50 of
1.4 ± 0.06 mM, an almost threefold-higher relative affinity
towards K+than that of the wild-type enzyme (Fig 6) In
the case of the Ser894fiIle mutant, the relative affinity for
K+was 21.95 ± 1.41 mM, which is about six times lower
than the affinity of the wild-type enzyme
How can mutation of Ser894 affect the interactions of the
enzyme with K+? Here, too, one can only assume that this
highly conserved serine stabilizes a structure of the protein
important for enzyme interactions with K+on the
extra-cellular surface The fact that replacement of the serine
hydroxyl group by the aspartic acid carboxyl group, with its
higher dipole character, results in an enzyme with higher
apparent affinity for K+, whereas replacement by the
nonpolar isoleucine results in an enzyme with lower affinity
for the cation, could indicate (but does not necessarily
require) a direct involvement of Ser894 in the process
involved in the uptake of K+from the extracellular milieu
A different explanation should also be considered: because
the b subunit is known to influence the interactions of the
enzyme with K+and because Ser894 is within a sequence
known to interact with the b subunits, it is possible that
mutations of this amino acid influence interaction between
the subunits sufficiently enough to affect K+recognition
without being absolutely essential for activity Based on the
present data, it is not possible to discern between these two
possibilities
The results obtained here demonstrate a variety of
functional implications associated with the highly conserved
peptide DSYG(893–896) While Tyr895 appears to be
rather neutral for the enzyme properties, Ser894 is involved
– directly or indirectly – in enzyme interactions with K+
Also, Asp893 and Gly896 were proven to be absolutely
essential for activity Nevertheless, while mutations of
Gly896 demonstrate that this amino acid is critical for
assembly between a and b subunits, the results obtained
with the Asp893 mutants demonstrate that loss of activity
cannot be entirely explained on the basis of structural
disturbances leading to the loss of a/b subunit interactions
Together with the results of previous investigations,
show-ing the involvement of amino acids from within the L7/8
loop in Na+conduction [4] and studies showing the loss of
Rb+occlusion after Cu2+-catalysed cleavage of the L7/8
peptide [5], the study presented here supports and underlines
the importance of the L7/8 peptide for the function of a/b
heterodimeric P-type ATPases by addressing the functional
role of single amino acids from the highly conserved peptide
DSYG(893–896)
Acknowledgements
The authors thank E A Martinson for reading the manuscript and
R A Farley for the generous gift of the vectors YhNa1 and GhNb1.
This work was supported by the Deutsche Forschungsgemeinschaft,
Sche 397/5-1 and 397/5-2 S B was supported through the
Gradu-iertenkolleg Molekulare Biologie und Pharmakologie of the
Justus-Liebig-University Giessen.
References
1 Lemas, M.V., Hamrick, M., Takeyasu, K & Fambrough, D.M (1994) 26 amino acids of an extracellular domain of the Na,K-ATPase alpha-subunit are sufficient for assembly with the Na,K-ATPase beta-subunit J Biol Chem 269, 8255–8259.
2 Colonna, T.E., Huynh, L & Fambrough, D.M (1997) Subunit interactions in the Na,K-ATPase explored with the yeast two-hybrid system J Biol Chem 272, 12366–12372.
3 Fambrough, D.M., Huynh, L & Hwang, B (2000) The Na,K-ATPase a-b subunit assembly site In Na/K-ATPase and Related ATPases (Taniguchi, K & Kaya, S., eds), pp 103–106 Elsevier Science BV, Tokyo.
4 Schneider, H & Scheiner-Bobis, G (1997) Involvement of the M7/ M8 extracellular loop of the sodium pump a subunit in ion transport Structural and functional homology to P-loops of ion channels J Biol Chem 272, 16158–16165.
5 Shimon, M.B., Goldshleger, R & Karlish, S.J (1998) Specific
Cu 2+ -catalyzed oxidative cleavage of Na,K-ATPase at the extracellular surface J Biol Chem 273, 34190–34195.
6 Eakle, K.A., Kabalin, M.A., Wang, S.G & Farley, R.A (1994) The influence of b subunit structure on the stability of Na+/K+ -ATPase complexes and interaction with K+ J Biol Chem 269, 6550–6557.
7 Eakle, K.A., Lyu, R.-M & Farley, R.A (1995) The influence of
b subunit structure on the interaction of Na+/K+-ATPase complexes with Na+ A chimeric a subunit reduces the Na+ dependence of phosphoenzyme formation from ATP J Biol Chem 270, 13937–13947.
8 Geering, K (1991) Posttranslational modifications and intracel-lular transport of sodium pumps: importance of subunit assembly Soc Gen Physiol Ser 46, 31–43.
9 Horowitz, B., Eakle, K.A., Scheiner-Bobis, G., Randolph, G.R., Chen, C.Y., Hitzeman, R.A & Farley, R.A (1990) Synthesis and assembly of functional Na,K-ATPase in yeast J Biol Chem 265, 4189–4192.
10 Muller-Ehmsen, J., Juvvadi, P., Thompson, C.B., Tumyan, L., Croyle, M., Lingrel, J.B., Schwinger, R.H., McDonough, A.A & Farley, R.A (2001) Ouabain and substrate affinities of human
Na+-K+-ATPase a1/b1, a2/b1, and a3/b1 when expressed sepa-rately in yeast cells Am J Physiol Cell Physiol 281, C1355– C1364.
11 Hemsley, A., Arnheim, N., Toney, M.D., Cortopassi, G & Galas, D.J (1989) A simple method for site-directed mutagenesis using the polymerase chain reaction Nucleic Acids Res 17, 6545–6551.
12 Sanger, F., Nickler, S & Coulson, A.R (1977) DNA sequencing with chain terminating inhibitors Proc Natl Acad Sci USA 74, 5463–5467.
13 Ito, H., Fukuda, Y., Murata, K & Kimura, A (1983) Transfor-mation of intact yeast cells treated with alkali cations J Bacteriol.
153, 163–168.
14 Horowitz, B., Hensley, C.B., Quintero, M., Azuma, K.K., Putnam,
D & McDonough, A.A (1990) Differential regulation of Na,K-ATPase alpha 1, alpha 2, and beta subunit mRNA and protein levels by thyroid hormone J Biol Chem 265, 14308– 14314.
15 Scheiner-Bobis, G & Farley, F.A (1994) Subunit requirements for the expression of functional sodium pumps in yeast cells Biochim Biophys Acta 1193, 226–234.
16 Scheiner-Bobis, G & Schoner, W (1985) Demonstration of a
Mg2+-induced conformational change by photoaffinity labelling
of the high affinity ATP-binding site of (Na++ K+)-ATPase with 8-azido-ATP Eur J Biochem 152, 739–746.
17 Lowry, O.H., Rosenbrough, N.J., Farr, A.L & Randall, R.J (1951) Protein measurement with the folin phenol reagent J Biol Chem 193, 265–275.