Download Subunit Isoform of X,K-ATPase in Human Skeletal Muscle

Biochemical and Biophysical Research Communications 277, 430 – 435 (2000)
doi:10.1006/bbrc.2000.3692, available online at http://www.idealibrary.com on
Immunochemical Demonstration of a Novel ␤-Subunit
Isoform of X,K-ATPase in Human Skeletal Muscle
Nikolay B. Pestov,* ,† Tatyana V. Korneenko,† Hao Zhao,* Gail Adams,*
Mikhail I. Shakhparonov,† and Nikolai N. Modyanov* ,1
*Department of Pharmacology, Medical College of Ohio, Toledo, Ohio 43614; and †Shemyakin and Ovchinnikov Institute
of Bioorganic Chemistry, Russian Academy of Sciences, Moscow 117871, Russia
Received September 21, 2000
Recently we have identified mRNA encoding a hitherto unknown mammalian X,K-ATPase ␤-subunit expressed predominantly in muscle tissue (Pestov, N. B.
et al. (1999) FEBS Lett. 456, 243–248). Here we demonstrate the existence of the predicted protein, designated as ␤ m (␤ muscle), in human adult skeletal muscle
membranes using immunoblotting with ␤ m-specific antibodies generated against recombinant polypeptide
formed by extramembrane ␤ m domains. The electrophoretic mobility of ␤ m was shown to be abnormally
low due to the presence of Glu-rich sequences. In contrast to mature forms of other known X,K-ATPase
␤-subunits, carbohydrate moiety of ␤ m is sensitive to
endoglycosidase H and appears to be composed of
short high-mannose or hybrid N-glycans. This finding
argues in favor of an intracellular location of ␤ m in
human skeletal muscle. © 2000 Academic Press
Key Words: electrophoretic mobility; N-glycosylation; immunodetection; Na,K-ATPase; protein expression; skeletal muscle; X,K-ATPase.
The X,K-ATPase family consists of several structurally related ion pumps which differ in their catalytic
properties and physiological functions (for reviews, see
(1–3)). This diversity is generated by combinations of
␣- and ␤-subunit isoforms. In mammals, four isoforms
of Na,K-ATPase, one gastric H,K-ATPase, and one
non-gastric H,K-ATPase isoform are known among the
catalytic X,K-ATPase ␣-subunits. Until recently, their
␤-subunit partners were thought to include three isoforms for the Na,K-ATPase and one for the gastric
H,K-ATPase, whereas a ␤-subunit specific for nongastric H,K-ATPase remains unknown.
Abbreviations used: ␤ m, muscle specific X,K-ATPase ␤-subunit;
Endo H, endoglycosidase H; PMSF, phenyl methyl sulphonyl fluoride; PNGase F, peptide-N-glycosidase F; RT-PCR, reverse
transcription-polymerase chain reaction.
1
To whom correspondence should be addressed. Fax: 1-419-3832871. E-mail: [email protected].
0006-291X/00 $35.00
Copyright © 2000 by Academic Press
All rights of reproduction in any form reserved.
In our previous paper we described homologous human and rat cDNAs encoding a novel X,K-ATPase
␤-subunit (4). RT-PCR showed that the corresponding
gene has a highly specific expression pattern being
active only in skeletal muscle and at a low level in
heart, and silent in other tissues tested. To emphasize
this remarkable tissue specificity that is in sharp contrast with other ␤-subunit isoforms the novel protein
was named as muscle-specific X,K-ATPase ␤-subunit,
or ␤ m.
Mammalian X,K-ATPase ␤-subunit proteins are
readily discernable by the presence of specific conserved sequence motifs, e.g., Y(Y/F)PYYGK. The topology of the proteins features one transmembrane
stretch and a C-terminal ectodomain that is modified
by N-glycosylation at several sites (5, 6). The deduced
primary structure of ␤ m is similar to other isoforms in
these respects, but the protein is unique in the X,KATPase family due to an extension of its N-terminus
that is rich in glutamic acid residues (4). In addition,
␤ m provides the only known example of alternative
splicing among the mammalian X,K-ATPase ␤-subunit
mRNAs. Two ␤ m transcripts that differ in the presence
or absence of a 12-bp sequence are created through the
use of two donor 5⬘-sites in intron 2 (4). These variants
encode two proteins of 357 and 353 amino acid residues
that differ in four residues in the junction between
cytoplasmic and intramembrane domains.
Here we report the production of a recombinant ␤ m
protein and specific antibodies, the demonstration of
the existence of the predicted ␤ m protein in human
skeletal muscle membranes, and evidence that ␤ m is
different from other X,K-ATPase ␤-subunits in the glycosylation status.
MATERIALS AND METHODS
DNA constructs. Conditions of RT-PCR were described before (7)
with the exception that Pfu DNA polymerase (Promega) was used
instead of Taq. PCR products of primers PU (caatgcctgsgtggcagaaattgcagatc) and TU (agatgatcggctgtcttagatgtcc) representing frag-
430
Vol. 277, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
ments from Asn80 to C-end (Thr357) were cloned into pCR-Script
Amp vector using PCR-Script Amp cloning kit (Stratagene). Among
the colonies, constructs pH3, pH21, and pH31 were identified that
comprised alternatively spliced transcripts A, B, and a transcript
lacking exon 3 (hereafter called C), respectively (Fig. 1). To restore
the complete reading frame, product of primers VEKO1 (cttkgcccactgaacagccatgag) and BS1 (ggcgaagggtctgatcataac) (corresponding to
Met1- Ala160) was cloned as above and the PstI-BamHI cassette
from this construct was introduced in the plasmids pH3, pH21, and
pH31 to give constructs pHP3, pHP21, and pHP31, respectively.
These plasmids contained the ORFs (1071 bp, variant A; 1059 bp
variant B; 942 bp, variant C) plus 129 bp from 3⬘-UTR, and 19 bp
from 5⬘-UTR. For protein expression, full-length coding region of
pHP31, was amplified with primers F2SWHBGL (gaacaagatctagaaggcarctccggtcca) and BCHINR (tgaaaagcttargtttctatgttcagggt), digested with BglII plus HindIII and cloned at BamHI/HindII sites of
vector pQE30 (Qiagen) to give an expression construct, pH3130.
Preparation of recombinant protein and ␤ m-specific antibodies.
JM109 Escherichia coli cells harboring the pH3130 plasmid were
grown at 37°C in LB medium supplemented with 100 mg/l ampicillin, 5 mM MgCl 2 and 5 mM CaCl 2 until A 600 0.5. Then 1 mM
isopropyl-␤-D-thiogalactopyranoside was added and growing continued for 4 h. The cells were pelleted and the transmembrane moietydeficient ␤ m protein was purified by immobilized metal affinity chromatography as described before (8). For immunization of rabbits a
previously described schedule (9) was modified so that a
nitrocellulose-absorbed antigen (see (10) for preparation) was used
for booster injections. Affinity purification of antibodies was performed using the antigen blotted onto PVDF membrane essentially
as described (11).
Tissues and preparation of microsomes. Tissues were from 50 –
60-year-old female donors. Samples of human skeletal and heart
muscles were provided by Cooperative Human Tissue Network (Columbus, OH) and brain from Harvard Brain Tissue Resource Center
(Belmont, MA). Tissue samples (3 g) were frozen in liquid nitrogen,
crushed with a hammer and homogenized with a Polytron homogenizer at 12000 rpm for 1 min in buffer A (20 ml 30 mM Tris–HCl (pH
8.0), 2 mM EDTA, 100 ␮M PMSF, 0.1% methanol, 5 mM benzamidine, 1 ␮M leupeptin, 1 ␮M pepstatin, 1 ␮M bacitracin). The homogenate was centrifuged at 9000g for 20 min, and the microsomes were
pelleted at 150,000g for 30 min. The pellet was washed with the
buffer A supplemented with 200 mM NaCl, resuspended in the buffer
A containing 10% sucrose and stored at ⫺70°C.
Deglycosylation and Western blotting. Microsomes containing 60
␮g protein were incubated at 100°C for 5 min with 0.5% SDS and 1%
␤-mercaptoethanol and placed on ice. Then 1% octyl glucoside, 100
␮M PMSF, 1* protease inhibitor cocktail (Sigma) were added and the
samples were incubated either with 50 mM sodium citrate (pH 5.5)
plus 1000 u Endo H f (New England Biolabs) or with 50 mM sodium
phosphate (pH 7.5) plus 500 u PNGase F (New England Biolabs) at
37°C for 1 h. Control samples were incubated without the glycosidases in 50 mM sodium phosphate. The Laemmli sample loading
buffer (12) was added and the samples were incubated again at
100°C for 5 min. Then 10 ␮g protein was electrophoresed (12) and
blotted onto PVDF membrane (Amersham-Pharmacia). The membrane was blocked in 5% lowfat milk in TBS supplemented with
Megablock (Cel Associates) diluted 1:500. Then the membrane was
consequently incubated with the affinity purified anti-␤ m rabbit antibodies, biotin conjugated anti-rabbit antibodies (AmershamPharmacia) and streptavidin-peroxidase (Sigma, S9420) for 1 h each
with thorough washes in TBS with 0.1% Tween 20 between incubations. The immunoblots were visualized with a chemiluminescent
substrate (ECL⫹Plus, Amersham-Pharmacia). The same reagents
were used for control experiments except that antibodies against
Na,K-ATPase ␤2-subunit or affinity purified rabbit antibodies
against a non-related hexahistidine-tagged protein were substituted
for the anti-␤ m antibodies. 10 kDa protein ladder (Life Technologies)
and broad range standards (New England Biolabs) were used for
calculation of apparent molecular masses.
RESULTS
In our previous work we reported the primary structure of ␤ m cDNA as determined by direct sequencing of
overlapping RT-PCR products (4). Here we have reconstructed full-length ORFs of alternatively spliced transcripts of human ␤ m. Cloning of the PCR product of
primers PU and TU allowed the identification of variants A and B (containing or lacking a 12-bp insert,
respectively). In addition, despite the use of a gelpurified band for cloning, we also identified another
variant with the 12-bp insert but without a 129-bp
sequence corresponding to exon 3 (this exon encodes
the transmembrane domain of ␤ m) (Fig. 1).
It should be stressed that this new variant (termed
C) is unlikely to be a PCR or cloning artifact because
the deletion exactly coincides with exon-intron boundaries. Moreover, RT-PCR with primers PU and TU
produced additional very weak bands of lower molecular masses when using RNA from some human skeletal
muscle samples (results not shown). Obviously, variant
C is a rare species among the ␤ m transcripts, and can
be thought of as a product of aberrant exon skipping.
Quite probably, this variant has no functional importance because the deletion of exon 3 ablates the only
transmembrane domain that is functionally indispensable in X,K-ATPase ␤-subunits.
To produce recombinant ␤ m proteins in E. coli, we
attempted to clone the ORFs into an expression vector,
pQE30 that allows production of proteins with
N-terminal hexahistidine tag. Only variant C was successfully cloned into that vector (resulting construct
pH3130) while no recombinant colonies were obtained
for variants A and B. This is possibly due to their
severe toxicity to bacteria (a frequent phenomenon for
recombinant proteins with hydrophobic segments (9)).
However, the toxicity problem was also significant for
the C-variant: induction of protein expression from the
construct pH3130 resulted in a rapid halt of cell
growth. Only trace amount of protein was purified from
such cells. Fortunately, it was found that the toxic
effect could be diminished by the addition of calcium
and magnesium in the medium. In this case only a
minor reduction of growth rate after induction was
observed and about 2.5 mg protein per liter of culture
was purified by immobilized metal affinity chromatography. Such a yield is low in comparison with usual
yields of non-toxic proteins (9), however, the amount of
the ␤ m protein obtained was sufficient to raise polyclonal antibodies with their subsequent affinity purification.
The mobility of the purified recombinant protein
(variant C) in SDS–PAGE corresponds to an apparent
molecular mass of 49 kDa (Fig. 2), that is much higher
431
Vol. 277, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Amino acid sequence of human X,K-ATPase ␤ m subunit (alternative splice variants A, B, and C), and nucleotide sequence of the
alternative splicing sites. The sequences shown represent the complete primary structure of the variant A, and portions of variants B and
C containing deletions of amino acid residues (indicated by dashes). The corresponding part of cDNA sequence is shown above the deduced
amino acid sequences. Nucleotides and amino acids are numbered to the right of the sequences beginning with the initiating ATG and Met,
respectively. Consensus sequences of potential sites of N-glycosylation are underlined. Positions of the introns are indicated by arrows and
lowercase letters and flanked by exon numbers.
than calculated from its deduced sequence (38.36 kDa).
However, the abnormal electrophoretic mobility corresponding to an apparently higher molecular mass is
FIG. 2. Electrophoretic analysis of purified recombinant X,KATPase ␤ m-subunit (variant C with N-terminal hexahistidine tag).
The protein was purified from E. coli as described under Materials
and Methods and subjected to SDS–PAGE (12% polyacrylamide gel
with Coomassie blue R staining). Lane 1, molecular mass markers
(10 kDa protein ladder, Life Technologies); lane 2, purified recombinant X,K-ATPase ␤ m-subunit (variant C with N-terminal hexahistidine tag).
frequently observed for acidic proteins, especially those
with Glu-rich domains. To give just one example, Glurich cyclic nucleotide-gated channel from rod photoreceptors migrates as 240 kDa instead of 155 kDa calculated from cDNA (13).
Microsomes prepared from three different human
tissues, skeletal muscle, heart and brain, were analyzed by Western blotting with the affinity purified
antibodies against the recombinant ␤ m protein (Fig. 3).
The tissue choice was dictated by previous results on
RT-PCR screening of many tissues (4) where ␤ m transcripts were detected only in skeletal and heart muscles. Brain microsomes were taken as a negative control. A specific band was observed in skeletal muscle
(Fig. 3, lanes 4 – 6) while no specific bands were observed in other tissues (Fig. 3, lanes 1–3 and 7– 8).
Thus, the existence of ␤ m protein in human skeletal
muscle microsomes was demonstrated, while no protein was detected in microsomes from human heart
and brain.
In order to characterize ␤ m protein glycosylation status, the microsome proteins were treated with glycosidases, Endo H and PNGase F. The electrophoretic mo-
432
Vol. 277, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Immunoblotting of microsome proteins from different
human tissues with antibodies against recombinant X,K-ATPase
␤ m-subunit and deglycosylation shift assay. Membrane proteins were
treated or not treated with Endo H or PNGase F, resolved on a 12%
polyacrylamide SDS gel and analyzed by Western blotting with
affinity purified antibodies against recombinant ␤ m protein as described under Materials and Methods. Positions of molecular mass
markers (10 kDa protein ladder, Life Technologies) are shown to the
left. Lanes 1–3, heart; lanes 4 – 6, skeletal muscle; lanes 7– 8, brain.
Lanes 2, 5, Endo H treated samples; lanes 3, 7, 8, PNGase F treated
samples. *A nonspecific band that is produced also in negative control experiments due to binding of secondary antibodies.
bility of the detected ␤ m protein without deglycosylation
corresponds to an apparent molecular mass of 57
kDa (Fig. 3, lane 4). Treatment with either Endo H or
PNGase F shifted the band mobility indicating that
N-glycans on human skeletal muscle ␤ m are sensitive
to these glycosidases (Fig. 3, lanes 5 and 6). Addition of
␣-L-fucosidase to enhance the action of the glycosidases
did not change the mobility shifts (results not shown).
The electrophoretic mobility of the detected ␤ m protein
deglycosylated with PNGase F corresponds to an apparent molecular mass of 52 kDa (Fig. 3, lane 6). The
Endo H treated protein (Fig. 3, lane 5) migrates
slightly more slowly so that its molecular mass is 0.5–
1.0 kDa higher than that of the PNGase F treated
protein.
The apparent molecular mass of the deglycosylated
␤ m protein from human skeletal muscle microsomes (52
kDa) is much higher than calculated values for proteins deduced from cDNA (41.6 kDa, 41.0 kDa, and
36.9 kDa for variants A, B, and C, respectively). However, the difference between the mobilities of the recombinant ␤ m protein (Fig. 2) and the protein detected
in skeletal muscle (Fig. 3, lane 4) corresponds to about
3 kDa and is close to the difference of their calculated
molecular masses of 3.24 kDa (the recombinant ␤ m
protein contains 312 amino acid residues from variant
C and additional 11 residues encoded by the vector).
Therefore, the observed mobility of human skeletal
muscle ␤ m corresponds to the value that should be
expected from the mobility of the recombinant protein.
The detected ␤ m protein (Fig. 3, lanes 4 – 6) can correspond to either A or B splicing variants, or a mixture
of them because the difference of their molecular
masses (0.65 kDa) is too small for adequate resolution
in SDS–PAGE. On the other hand, the mobility of
variant C, if it exists in the tissues, is expected to
correspond to 47.3 kDa and this indicates that the
detected protein does not correspond to the variant C.
In addition, we have excluded the possibility that the
bands observed (Fig. 3) were generated by crossreaction of our antibodies with a major skeletal muscle
membrane-associated protein, the 53-kDa glycoprotein, that has a similar electrophoretic mobility and
similar shifts after Endo H or PNGase F treatments
(14, 15). Specific monoclonal antibodies were used for
this purpose (14, 15). It was found that the 53-kDa
glycoprotein migrates close to ␤ m but a difference in
their mobilities is readily seen when the samples are
run in parallel (results not shown).
The results of Western blotting can be summarized
as follows: (i) ␤ m protein is present in human skeletal
muscle microsomes; (ii) ␤ m protein was not detected in
microsomes from human heart and brain; (iii) ␤ m protein has an abnormal electrophoretic mobility in SDS–
PAGE; (iv) detected ␤ m corresponds to either A or B
splicing variants, but not to variant C; (v) N-glycans of
␤ m from human skeletal muscle are sensitive to Endo H
and PNGase F; and (vi) the electrophoretic mobility
shift after deglycosylation indicates that the overall
amount of carbohydrate residues is about 5 kDa.
DISCUSSION
Here we have demonstrated for the first time that
the previously identified mRNA (4) produces the corresponding protein in human skeletal muscle.
From RT-PCR studies of many rat and human tissues it was found that the ␤ m gene expression is restricted to muscle tissue, the transcripts are abundant
in skeletal muscle and are also present in heart, albeit
at a much lower level (4). Immunodetection of the ␤ m
protein undertaken in this work was positive only in
case of skeletal muscle membranes which produced
strong signals in Western blotting with affinity-purified
antibodies against recombinant ␤ m. Brain, negative by
RT-PCR, was negative also in the immunoblotting as
expected, and no ␤ m protein was detected in heart. The
latter most possibly indicates that ␤ m protein content is
very low in adult human heart, below the detection
limit even with the use of a sensitive chemiluminescent
substrate. This implies that if heart ␤ m is uniformly
distributed in the organ then it is logical to suggest
that ␤ m has no functional importance in adult human
heart due to its extremely rare abundance. On the
other hand, its expression might be restricted to a
small population of particular cells (like Na,K-ATPase
␣3 isoform in sites of impulse transmission (16)). The
expression ␤ m in heart may be also upregulated under
certain conditions. However, there is no evidence at the
433
Vol. 277, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
moment for such possibilities, and thus ␤ m should be
regarded as a skeletal muscle-specific protein.
Despite the fact that ␤ m has an abnormal electrophoretic mobility, it is possible to say that the molecular mass of skeletal muscle ␤ m corresponds to that
determined from cDNA because the difference between
the mobilities of the purified recombinant and detected
deglycosylated proteins closely matches the calculated
value. Thus the primary structure of authentic ␤ m from
human skeletal muscle was deduced correctly (4),
though the existence of minor modifications cannot be
excluded.
The most surprising finding was that the difference
of electrophoretic mobilities between intact and deglycosylated forms is unexpectedly small. The other
known X ⫹,K ⫹-ATPase ␤-subunits are heavily glycosylated proteins, migrating as fuzzy bands in SDS–
PAGE. This usually makes them more difficult to detect by Western blotting and deglycosylation is usually
required for high sensitivity of detection in tissues (17,
18). To give another example, in our control experiments glycosylated Na,K-ATPase ␤2-subunit was detected only in brain (which has a very high content of
this isoform) and in all tissues tested after PNGase
treatment (not shown). Hence, our results (Fig. 3) may
be questioned if other ␤ m glycoforms exist in skeletal
muscle but were not detected due to the band broadening during electrophoretic separation. This possibility cannot be rigorously excluded, however, in this case
the signal from the deglycosylated band should be
stronger than from the intact glycoprotein. In fact,
densitometry of the images showed that the situation
is just the reverse: the band from the intact glycoprotein is slightly stronger than the Endo H or PNGase F
treated proteins (results not shown). This indicates
that other glycoforms, if present, can constitute only a
tiny proportion of total ␤ m, or they contain PNGase
F-resistant chains, which are uncommon in mammals.
Most likely, the ␤ m protein in human adult skeletal
muscle has a glycosylation pattern unusual for the
class of X ⫹,K ⫹-ATPase ␤-subunits. Most of oligosaccharides on other ␤-subunit isoforms are processed to
the complex type and are resistant to Endo-H (5, 6, 19,
20, 21).
How many of the four theoretical N-glycosylation
sites in ␤ m are occupied by sugar chains? Because the
Endo-H and PNGase F treated proteins have a small
but readily detectable difference in their apparent molecular masses and because Endo H leaves only one
N-acetylglucosamine residue from the chitobiose core,
our guess is that ␤ m is glycosylated at all four sites. The
overall sugar content can be estimated about 5 kDa
(Fig. 3) and this imposes constraints on the length of
the sugar chains. Together with the fact of their sensitivity to Endo-H, it is possible to suggest that ␤ m from
adult human skeletal muscle bears short highmannose or hybrid chains.
Other X,K-ATPase ␤-subunit isoforms are plasma
membrane proteins and most of their N-glycans are
processed to the complex type during transport
through the Golgi system, as most glycoproteins that
are destined to the cell surface. Only small quantities
of high-mannose precursors of the ␤-subunits could be
detected (20). Because ␤ m seems to possess only short
high-mannose or hybrid chains it is reasonable to hypothesize that the detected ␤ m protein is not translocated to the plasma membrane but resides in an intracellular compartment of human skeletal muscle cells
such as sarcoplasmic reticulum or cis-Golgi.
It is also possible that skeletal muscle has a pool of
immature ␤ m whose processing can be stimulated by
some exogenous signals. It was shown that insulin
treatment or exercise induces translocation of ␣2, ␤1
and ␤2 Na,K-ATPase subunits from intracellular
stores to the cell surface in the skeletal muscle (22–24).
This process may also engage ␤ m and its translocation
may be synchronized with the maturation of its sugar
chains.
Which catalytic ␣-subunits are associated with ␤ m in
skeletal muscle? This problem awaits future investigations, however, it should be noticed that due to its
unique structural features (acidic N-terminus and
atypical glycosylation) ␤ m cannot be a priori classified
as a Na,K-ATPase ␤-subunit. The possibility that it
associates with a different P-ATPase cannot be rejected at present especially because there may exist a
hitherto unknown muscle-specific X,K-ATPase catalytic ␣-subunit.
ACKNOWLEDGMENTS
This work was supported by National Institutes of Health Grants
HL-36573 and GM-54997 and by the Russian Foundation for Basic
Research Grants 00-04-48153 and 98-04-48408. We thank Dr. D.
MacLennan for antibodies to 53-kDa glycoprotein and Dr. P. Martı́nVasallo for antibodies to Na,K-ATPase ␤2-subunit. We are grateful
to Dr. A. Askari for his valuable comments on the manucript and to
M. Tillekeratne for her excellent assistance.
REFERENCES
434
1. Blanco, G., and Mercer, R. W. (1998) Am. J. Physiol. 275, F633–
F650.
2. Béguin, P., Hasler, U., Beggah, A., and Geering, K. (1998) Acta
Physiol. Scand. 643, 283–287.
3. Modyanov, N. N., Adams, G., Pestov, N. B., Yu, C., Tillekeratne,
M., Korneenko, T. V., Shakhparonov, M. I., Crambert, G., Horisberger, J.-D., and Geering, K. (2000) In Na/K-ATPase and Related ATPases (Taniguchi, K., and Kaya, S., Eds.), Elsevier, in
press.
4. Pestov, N. B., Adams, G., Shakhparonov, M. I., and Modyanov
N. N. (1999) FEBS Lett, 456, 243–248.
5. Geering, K. (1990) J. Membr. Biol. 115, 109 –121.
6. Chow, D.-C., and Forte, J. C. (1995) J. Exp. Biol. 198, 1–17.
7. Pestov, N. B., Romanova, L. G., Korneenko, T. V., Egorov, M. V.,
Kostina, M. B., Sverdlov, V. E., Askari, A., Shakhparonov, M. I.,
and Modyanov, N. N. (1998) FEBS Lett. 440, 320 –324.
Vol. 277, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
8. Pestov, N. B., Gusakova, T. V., Kostina, M. B., and Shakhparonov, M. I. (1996) Bioorg. Khim. 22, 664 – 670.
9. Pestov, N. B., Korneenko, T. V., Kostina, M. B., and Shakhparonov, M. I. (1999) Bioorg. Khim. 25, 505–512.
10. Knudsen, K. A. (1985) Anal. Biochem. 147, 285–288
11. Rucklidge, G. J., Milne, G., Chaudhry, S. M., and Robins, S. P.
(1996) Anal. Biochem. 243, 158 –164.
12. Laemmli, U. K. (1970) Nature 227, 680 – 685.
13. Körschen, H. G., Illing, M., Seifert, R., Sesti, F., Williams, A.,
Gotzes, S., Colville, C., Müller, F., Dosé, A., Godde, M., Molday,
L., Kaupp, U. B., and Molday, R. S. (1995) Neuron 15, 627– 636.
14. Campbell, K. P., MacLennan, D. H. (1981) J. Biol. Chem. 256,
4626 – 4632.
15. Leberer, E., Charuk, J. H., Clarke, D. M., Green, N. M.,
Zubrzycka-Gaarn, E., and MacLennan, D. H. (1989) J. Biol.
Chem. 264, 3484 –3493.
16. Zahler, R., Sun, W., Ardito, T., Zhang, Z. T., Kocsis, J. D., and
Kashgarian, M. (1996) Circ. Res. 78, 870 – 879.
17. Wang, J., Schwinger, R. H., Frank, K., Muller-Ehmsen, J.,
Martı́n-Vasallo, P., Pressley, T. A., Xiang, A., Erdmann, E., and
McDonough, A. A. (1996) J. Clin. Invest. 98, 1650 –1658.
18. Arystarkhova, E., and Sweadner, K. J. (1997) J. Biol. Chem. 272,
22405–22408.
19. Tamkun, M. M., and Fambrough, D. M. (1986) J. Biol. Chem.
261, 1009 –1019.
20. Chow, D.-C., and Forte, J. G. (1993) Am. J. Physiol. 265, C1562–
C1570.
21. Tyagarajan, K., Townsend, R. R., and Forte, J. G. (1996) Biochemistry 35, 3238 –3246.
22. Hundal, H. S., Marette, A., Mitsumoto, Y., Ramlal, T., Blostein,
R., and Klip, A. (1992) J. Biol. Chem. 267, 5040 –5043.
23. Lavoie, L., Roy, D., Ramlal, T., Dombrowski, L., Martı́n-Vasallo,
P., Marette, A., Carpentier, J. L., and Klip, A. (1996) Am. J.
Physiol. 270, C1421–C1429.
24. Juel, C., Nielsen, J. J., and Bangsbo, J. (2000) Am. J. Physiol.
Regul. Integr. Comp. Physiol. 278, R1107–R1110.
435