Fast-twitch Skeletal Muscle Sarcoplasmic Reticulum Research Articles

Roles of Long-range Electrostatic Domain Interactions and K+ in Phosphoenzyme Transition of Ca2+-ATPase

Sarcoplasmic reticulum Ca(2+)-ATPase couples the motions and rearrangements of three cytoplasmic domains (A, P, and N) with Ca(2+) transport. We explored the role of electrostatic force in the domain dynamics in a rate-limiting phosphoenzyme (EP) transition by a systematic approach combining electrostatic screening with salts, computer analysis of electric fields in crystal structures, and mutations. Low KCl concentration activated and increasing salt above 0.1 m inhibited the EP transition. A plot of the logarithm of the transition rate versus the square of the mean activity coefficient of the protein gave a linear relationship allowing division of the activation energy into an electrostatic component and a non-electrostatic component in which the screenable electrostatic forces are shielded by salt. Results show that the structural change in the transition is sterically restricted, but that strong electrostatic forces, when K(+) is specifically bound at the P domain, come into play to accelerate the reaction. Electric field analysis revealed long-range electrostatic interactions between the N and P domains around their hinge. Mutations of the residues directly involved and other charged residues at the hinge disrupted in parallel the electric field and the structural transition. Favorable electrostatics evidently provides a low energy path for the critical N domain motion toward the P domain, overcoming steric restriction. The systematic approach employed here is, in general, a powerful tool for understanding the structural mechanisms of enzymes.

Autosomal dominant Brody disease cosegregates with a chromosomal (2;7)(p11.2;p12.1) translocation in an Italian family.

Brody disease is a rare muscle disorder characterized by exercise-induced impairment in muscle relaxation, due to a markedly reduced influx of calcium ions in the sarcoplasmic reticulum. A subset of autosomal recessive families harbour mutations in the ATP2A1 gene, encoding the fast-twitch skeletal muscle sarcoplasmic reticulum Ca(2+) ATPase (SERCA1). Rare autosomal dominant families have been described, in which ATP2A1 was excluded as the causative gene, further supporting genetic heterogeneity. We report four individuals from a three-generation Italian family with a clinical phenotype of Brody disease, in which linkage analysis excluded ATP2A1 as the responsible gene. The disease cosegregates in an autosomal dominant fashion with an apparently balanced constitutional chromosome translocation (2;7)(p11.2;p12.1), suggesting a causal relationship between the rearrangement and the phenotype. FISH analysis using YAC and PAC clones as probes refined the breakpoint regions to genomic segments of about 164 and 120 kb, respectively, providing a possible clue to pinpoint the location of a novel gene responsible for this rare muscle disorder.

Protein Kinase A Transgenes

The β-adrenergic receptor/adenylyl cyclase/protein kinase A (PKA) axis is the central signaling pathway that serves to stimulate cardiac function. It is classically perceived as a linear signaling cascade (Figure), and cAMP is thought to be the second messenger responsible for the positive inotropic, chronotropic, and lusitropic effects of catecholamines. New data gained from a series of transgenic animals overexpressing various elements of this axis show that this system has a far more complex makeup than previously thought. Transgenic models of the cardiac β-adrenergic receptor/PKA system. The signaling cascade is represented as a simple linear chain. Proteins that are detrimental when overexpressed are shown in black; those causing beneficial effects upon overexpression are depicted in white. PLB indicates phospholamban. It has been known for about two decades that the β-adrenergic receptor system is dysfunctional in heart failure.1 The loss of receptor responsiveness has been attributed to a reduced β-receptor number, particularly of the β1-subtype, plus a functional desensitization of the remaining receptors—presumably mediated via increased activity of G protein–coupled receptor kinases.2 The extent of the receptor downregulation and the functional loss correlate with the severity of the disease.3,4 This phenomenon has long been regarded as detrimental to the compromised function of the failing heart, and it has been proposed that restoration of β-adrenergic receptor responsiveness might be a strategy to improve cardiac function. However, recent clinical studies and observations on a panel of transgenic animals suggest that the opposite is the case. Sustained stimulation or overexpression of cardiac β1-adrenergic receptors leads to hypertrophy, fibrosis, and heart failure,5–7 and blockade of β-adrenergic receptors with several antagonists—bisoprolol, carvedilol, and metoprolol—improves survival in heart failure patients.8 Thus, the loss of β-adrenergic receptor responsiveness in heart failure is more likely to represent a protective mechanism than …

A novel interaction mechanism accounting for different acylphosphatase effects on cardiac and fast twitch skeletal muscle sarcoplasmic reticulum calcium pumps

In cardiac and skeletal muscle Ca 2+ translocation from cytoplasm into sarcoplasmic reticulum (SR) is accomplished by different Ca 2+-ATPases whose functioning involves the formation and decomposition of an acylphosphorylated phosphoenzyme intermediate (EP). In this study we found that acylphosphatase, an enzyme well represented in muscular tissues and which actively hydrolyzes EP, had different effects on heart (SERCA2a) and fast twitch skeletal muscle SR Ca 2+-ATPase (SERCA1). With physiological acylphosphatase concentrations SERCA2a exhibited a parallel increase in the rates of both ATP hydrolysis and Ca 2+ transport; in contrast, SERCA1 appeared to be uncoupled since the stimulation of ATP hydrolysis matched an inhibition of Ca 2+ pump. These different effects probably depend on phospholamban, which is associated with SERCA2a but not SERCA1. Consistent with this view, the present study suggests that acylphosphatase-induced stimulation of SERCA2a, in addition to an enhanced EP hydrolysis, may be due to a displacement of phospholamban, thus to a removal of its inhibitory effect.

SERCA1a can functionally substitute for SERCA2a in the heart

We recently generated a transgenic (TG) mouse model in which the fast-twitch skeletal muscle sarcoplasmic reticulum (SR) Ca2+-ATPase (SERCA1a) is overexpressed in the heart. Ectopic overexpression of SERCA1a results in remodeling of the cardiac SR containing 80% SERCA1a and 20% endogenous SERCA2a with an approximately 2.5-fold increase in the total amount of SERCA protein (E. Loukianov et al. Circ. Res. 83: 889-897, 1998). We have analyzed the Ca2+ transport properties of membranes from SERCA1a TG hearts in comparison to control hearts. Our data show that the maximal velocity of SR Ca2+ transport was significantly increased ( approximately 1.9-fold) in TG hearts, whereas the apparent affinity of the SERCA pump for Ca2+ was not changed. Addition of phospholamban antibody in the Ca2+ uptake assays increased the apparent affinity for Ca2+ to the same extent in TG and non-TG (NTG) hearts, suggesting that phospholamban regulates the SERCA1a pump in TG hearts. Analysis of SERCA enzymatic properties in TG hearts revealed that the SERCA pump affinity for ATP, the Hill coefficient, the pH dependence of Ca2+ uptake, and the effect of acidic pH on Ca2+ transport were similar to those of NTG hearts. Interestingly, the rate constant of phosphoenzyme decay (turnover rate of SERCA enzyme) was also very similar between TG and NTG hearts. Together these findings suggest that 1) the SERCA1a pump can functionally substitute for SERCA2a and is regulated by endogenous phospholamban in the heart and 2) SERCA1a exhibits several enzymatic properties similar to those of SERCA2a when expressed in a cardiac setting.

Enhanced myocardial contractility and increased Ca2+ transport function in transgenic hearts expressing the fast-twitch skeletal muscle sarcoplasmic reticulum Ca2+-ATPase.

In this study, we investigated whether the fast-twitch skeletal muscle sarco(endo)plasmic reticulum Ca2+ transport pump (SERCA1a) can functionally substitute the cardiac SERCA2a isoform and how its overexpression affects cardiac contractility. For this purpose, we generated transgenic (TG) mice that specifically overexpress SERCA1a in the heart, using the cardiac-specific alpha-myosin heavy chain promoter. Ectopic expression of SERCA1a resulted in a 2.5-fold increase in the amount of total SERCA protein. At the same time, the level of the endogenous SERCA2a protein was decreased by 50%, whereas the level of other muscle proteins, including calsequestrin, phospholamban, actin, and tropomyosin, remained unchanged. The steady-state level of SERCA phosphoenzyme intermediate was increased 2.5-fold, and the maximal velocity of Ca2+ uptake was increased 1.7-fold in TG hearts, demonstrating that the overexpressed protein is functional. Although the basal cytosolic calcium signal was decreased by 38% in TG cardiomyocytes, the amplitude of cytosolic calcium signal was increased by 71.8%. The rate of calcium resequestration was also increased in TG myocytes, which was reflected by a 51.6% decrease in the normalized time to 80% decay of calcium signal. This resulted in considerably increased peak rates of myocyte shortening and relengthening (50.0% and 66.6%, respectively). Cardiac functional analysis using isolated work-performing heart preparations revealed significantly faster rates of contraction and relaxation in TG hearts (41.9% and 39.5%, respectively). The time to peak pressure and the time to half-relaxation were shorter (29.1% and 32.7%, respectively). In conclusion, our study demonstrates that the SERCA1a pump can functionally substitute endogenous SERCA2a, and its overexpression significantly enhances Ca2+ transport and contractile function of the myocardium. These results also demonstrate that the SERCA pump level is a critical determinant of cardiac contractility.

Sarcolipin Regulates the Activity of SERCA1, the Fast-twitch Skeletal Muscle Sarcoplasmic Reticulum Ca2+-ATPase

The 31-amino acid proteolipid, sarcolipin (SLN), is associated with the fast-twitch skeletal muscle sarcoplasmic reticulum Ca2+-ATPase (SERCA1). Constructs of human and rabbit SLN and of rabbit SLN with the FLAG epitope at its N terminus (NF-SLN) or its C terminus (SLN-FC) were coexpressed with SERCA1 in HEK-293 T-cells. Immunohistochemistry was used to demonstrate colocalization of NF-SLN and SERCA1 in the endoplasmic reticulum membrane and to demonstrate the cytosolic orientation of the N terminus of SLN. Coexpression of native rabbit SLN or NF-SLN with SERCA1 decreased the apparent affinity of SERCA1 for Ca2+ but stimulated maximal Ca2+ uptake rates (Vmax). The N terminus of SLN is not well conserved among species, and the addition of an N-terminal FLAG epitope did not alter SLN function. Anti-FLAG antibody reversed both the inhibition of Ca2+ uptake by NF-SLN at low Ca2+ concentrations and the stimulatory effect of NF-SLN on Vmax. Addition of the FLAG epitope to the highly conserved C terminus decreased the apparent affinity of SERCA1 for Ca2+ relative to native SLN and decreased Vmax significantly. Mutations in the C-terminal domain showed that this sequence is critical for SLN function. Mutational analysis of the transmembrane helix, together with the additive regulatory effects of coexpression of both SLN and phospholamban (PLN) with SERCA1, provided evidence for different mechanisms of interaction of SLN and PLN with SERCA molecules. Ca2+ uptake rates in sarcoplasmic reticulum vesicles, isolated from rabbit fast-twitch muscle (tibialis anterior) subjected to chronic low frequency stimulation, were reduced by approximately 40% in 3- and 4-day stimulated muscle, with a marginal increase in apparent affinity of SERCA1 for Ca2+. SERCA1 mRNA and protein levels were unaltered after stimulation. In contrast, SLN mRNA was decreased by 15%, and SLN protein was reduced by 40%. Reduced SLN expression could explain the decrease in SERCA1 activity observed in these muscles and might represent an early functional adaptation to chronic low frequency stimulation.

Characterization of the Gene Encoding Human Sarcolipin (SLN), a Proteolipid Associated with SERCA1: Absence of Structural Mutations in Five Patients with Brody Disease

Sarcolipin (SLN) is a low-molecular-weight protein that copurifies with the fast-twitch skeletal muscle sarcoplasmic reticulum Ca2+ATPase (SERCA1). Genomic DNA and cDNA encoding human sarcolipin (SLN) were isolated and characterized and theSLNgene was mapped to chromosome 11q22–q23. Human, rabbit, and mouse cDNAs encode a protein of 31 amino acids. Homology of SLN with phospholamban (PLN) suggests that the first 7 hydrophilic amino acids are cytoplasmic, the next 19 hydrophobic amino acids form a single transmembrane helix, and the last 5 hydrophilic amino acids are lumenal. The cytoplasmic and transmembrane sequences are not well conserved among the three species, but the lumenal sequence is highly conserved. Like SERCA1, SLN is highly expressed in rabbit fast-twitch skeletal muscle, but it is expressed to a lower extent in slow-twitch muscle and to an even lower extent in cardiac muscle, where SERCA2a and PLN are highly expressed. It is expressed in only trace amounts in pancreas and prostate.SLNandPLNgenes resemble each other in having two small exons, with their entire coding sequences lying in exon 2 and a large intron separating the two segments. Brody disease is an inherited disorder of skeletal muscle function, characterized by exercise-induced impairment of muscle relaxation. Mutations in theATP2A1gene encoding SERCA1 have been associated with the autosomal recessive inheritance of Brody disease in three families, but not with autosomal dominant inheritance of the disease. A search for mutations in theSLNgene in five Brody families, four of which were not linked toATP2A1,did not reveal any alterations in coding, splice junction or promoter sequences. The homozygous deletion of C438 in the coding sequence ofATP2A1in Brody disease family 3, leading to a frameshift and truncation following Pro147in SERCA1, is the fourthATP2A1mutation to be associated with autosomal recessive Brody disease.

Characterization of cDNA and Genomic DNA Encoding SERCA1, the Ca 2+-ATPase of Human Fast-Twitch Skeletal Muscle Sarcoplasmic Reticulum, and Its Elimination as a Candidate Gene for Brody Disease

Genomic DNA and cDNA encoding human SERCA1, the Ca 2+-ATPase of fast-twitch skeletal muscle sarcoplasmic reticulum (the ATP2A1 gene on chromosome 16p12), were isolated and characterized. The cDNA encodes 994 amino acids. The genomic DNA is 26 kb long and contains 23 exons, one of which can be alternatively spliced. The locations of each of the exon/intron boundaries are the same as those previously identified in the rabbit ATP2A1 gene. Brody disease is an inherited disorder of skeletal muscle, characterized by exercise-induced impairment of muscle relaxation. It has been postulated to result from a deficiency in SERCA1. In a search for the genetic basis of Brody disease, the coding sequence of the ATP2A1 gene in one Brody patient and the full-length sequences of two SERCA1 cDNAs in two other, unrelated Brody patients were compared with normal ATP2A1 sequences. In all three cases, the coding and splice junction sequences were normal, indicating that the forms of Brody disease manifested in these three patients are not caused by mutations in the coding or splice junction regions of the ATP2A1 gene.

Deletion of NH 2 − and COOH-terminal sequences destroys function of the Ca 2+ ATPase of rabbit fast-twitch skeletal muscle sarcoplasmic reticulum

Deletion mutants of the Ca 2+ ATPase of rabbit fast-twitch skeletal muscle sarcoplasmic reticulum (SERCA1a) were constructed and expressed in COS-1 cells. The mutants were expressed at levels 7- to 15-fold lower than the wild-type and were inactive. In vitro transcription-translation-insertion experiments showed that deletion of transmembrane sequences M 1 and M 2, but not of M 8, M 9, M 10 or the NH 2 − terminal 30 amino acids inhibited the stable insertion of the enzyme into the membrane. Thus there was no correlation between loss of function and membrane insertion. A signal sequence for membrane insertion may exist in M 1 and M 2.

Cytoplasmic Ca2+ does not inhibit the cardiac muscle sarcoplasmic reticulum ryanodine receptor Ca2+ channel, although Ca2+-induced Ca2+ inactivation of Ca2+ release is observed in native vesicles

Single channel properties of cardiac and fast-twitch skeletal muscle sarcoplasmic reticulum (SR) release channels were compared in a planar bilayer by fusing SR membranes in a Cs(+)-conducting medium. We found that the pharmacology, Cs+ conductance and selectivity to monovalent and divalent cations of the two channels were similar. The cardiac SR channel exhibited multiple kinetic states. The open and closed lifetimes were not altered from a range of 10(-7) to 10(-3) M Ca2+, but the proportion of closed and open states shifted to shorter closings and openings, respectively. However, while the single channel activity of the skeletal SR channel was activated and inactivated by micromolar and millimolar Ca2+, respectively, the cardiac SR channel remained activated in the presence of high [Ca2+]. In correlation to these studies, [3H]ryanodine binding by the receptors of the two channel receptors was inhibited by high [Ca2+] in skeletal but not in cardiac membranes in the presence of adenine nucleotides. There is, however, a minor inhibition of [3H]ryanodine binding of cardiac SR at millimolar Ca2+ in the absence of adenine nucleotides. When Ca(2+)-induced Ca2+ release was examined from preloaded native SR vesicles, the release rates followed a normal biphasic curve, with Ca(2+)-induced inactivation at high [Ca2+] for both cardiac and skeletal SR. Our data suggest that the molecular basis of regulation of the SR Ca2+ release channel in cardiac and skeletal muscle is different, and that the cardiac SR channel isoform lacks a Ca(2+)-inactivated site.

The epitope for monoclonal antibody A20 (amino acids 870-890) is located on the luminal surface of the Ca2(+)-ATPase of sarcoplasmic reticulum.

The epitope for monoclonal antibody A20 was mapped to amino acids 870-890 of the Ca2(+)-ATPase of rabbit fast-twitch skeletal muscle sarcoplasmic reticulum. The antibody did not react with the epitope in intact sarcoplasmic reticulum vesicles but reacted with the epitope when the vesicles were solubilized with the detergent C12E8 or made permeable by incubation in a hypotonic medium. By contrast, antibody A52, which binds to a cytoplasmic epitope consisting of amino acids 657-672, reacted with the Ca2(+)-ATPase in vesicular, permeabilized vesicular, and C12E8-solubilized states. These results clearly demonstrate that antibody A20 binds to a luminal epitope and provide the first demonstration that a specific segment of the Ca2(+)-ATPase is located on the luminal surface of the sarcoplasmic reticulum. These results are consistent with, and support, our model for folding of the Ca2(+)-ATPase (Brandl, C. J., Korczak, B., Green, N. M., and MacLennan, D. H. (1986) Cell 44, 597-607) in which residues 657-672 were proposed to form part of the cytoplasmic nucleotide binding domain, while residues 870-890 were proposed to form a luminal loop between proposed transmembrane sequences M7 and M8.

Regulation of the Skeletal Sarcoplasmic Reticulum Ca2+Pump by Phospholamban in Reconstituted Phospholipid Vesicles

Phospholamban is the regulator of the Ca(2+)-ATPase in cardiac sarcoplasmic reticulum (SR). The mechanism of regulation appears to involve inhibition by dephosphorylated phospholamban, and phosphorylation may relieve this inhibition. Fast-twitch skeletal muscle SR does not contain phospholamban, and it is not known whether the Ca(2+)-ATPase isoform from this muscle may be also subject to regulation by phospholamban in a similar manner as the cardiac isoform. To determine this we reconstituted the skeletal isoform of the SR Ca(2+)-ATPase with phospholamban in phosphatidylcholine proteoliposomes. Inclusion of phospholamban was associated with significant inhibition of the initial rates of Ca2+ uptake at pCa 6.0, and phosphorylation of phospholamban by the catalytic subunit of cAMP-dependent protein kinase reversed the inhibitory effects on the Ca2+ pump. Similar effects of phospholamban were also observed using phosphatidylcholine:phosphatidylserine proteoliposomes, in which the Ca2+ pump was activated by the negatively charged phospholipids (24). Regulation of the Ca(2+)-ATPase appeared to involve binding with the hydrophilic portion of phospholamban, as evidenced by cross-linking experiments, using a synthetic peptide that corresponded to amino acids 1-25 of phospholamban. These findings suggest that the fast-twitch isoform of the SR Ca(2+)-ATPase may be also regulated by phospholamban, although this regulator is not expressed in fast-twitch skeletal muscles.

Molecular Cloning and Expression of cDNA Encoding the 53,000-Dalton Glycoprotein of Rabbit Skeletal Muscle Sarcoplasmic Reticulum

The 53-kDa glycoprotein of rabbit skeletal muscle sarcoplasmic reticulum was purified by lentil lectin affinity chromatography and preparative polyacrylamide gel electrophoresis and partially sequenced. Polyclonal and monoclonal antibodies were raised against the 53-kDa glycoprotein and found to cross-react with the 160-kDa glycoprotein. A combination of antibody and synthetic oligonucleotide screening was used to isolate a cDNA encoding the 53-kDa glycoprotein of rabbit fast-twitch skeletal muscle sarcoplasmic reticulum. The cDNA encodes a protein of 453 amino acids with Mr of 52,421 and a 19-residue amino-terminal signal sequence. The deduced sequence contains two potential glycosylation sites and is largely hydrophilic. The presence of a glycine-rich sequence in the glycoprotein with homology to mononucleotide binding domains supports earlier observations that the glycoprotein binds ATP with high affinity. Although two sequences appear to be hydrophobic on a hydropathy plot, they are not sufficiently long nor sufficiently hydrophobic to qualify unambiguously as transmembrane sequences. The glycoprotein, like calsequestrin, was shown to be inaccessible to trypsin in intact sarcoplasmic reticulum. It can be eluted from the sarcoplasmic reticulum by extraction with [ethylenebis(oxyethylenenitrilo)]tetraacetic acid under hypotonic conditions. Thus, the glycoprotein appears to be localized entirely in the lumen of the sarcoplasmic reticulum and to be associated with the inner membrane surface through Ca2+-dependent mechanisms. Cotransfection of COS-1 cells with cDNAs encoding the glycoprotein and the Ca2+-ATPase led to expression of both proteins with a common localization in the microsomal fraction. The Ca2+ pumping activity of the microsomes isolated from transfected cells was unaltered by the presence of the glycoprotein. Thus the glycoprotein does not appear to modulate Ca2+-ATPase function.

Mutation of aspartic acid-351, lysine-352, and lysine-515 alters the Ca2+ transport activity of the Ca2+-ATPase expressed in COS-1 cells.

Full-length cDNAs encoding neonatal and adult isoforms of the Ca2+-ATPase of rabbit fast-twitch skeletal muscle sarcoplasmic reticulum were expressed transiently in COS-1 cells. The microsomal fraction isolated from transfected COS-1 cells contained immunoreactive Ca2+-ATPase and catalyzed Ca2+ transport at rates at least 15-fold above controls. No differences were observed in either the rates or Ca2+ dependency of Ca2+ transport catalyzed by the two isoforms. Aspartic acid-351, the site of formation of the catalytic acyl phosphate in the enzyme, was mutated to asparagine, glutamic acid, serine, threonine, histidine, or alanine. In every case, Ca2+ transport activity and Ca2+-dependent phosphorylation were eliminated. Ca2+ transport was also eliminated by mutation of lysine-352 to arginine, glutamine, or glutamic acid or by mutation of Asp351-Lys352 to Lys351-Asp352. Mutation of lysine-515, the site of fluorescein isothiocyanate modification in the enzyme, resulted in diminished Ca2+ transport activity as follows: arginine, 60%; glutamine, 25%; glutamic acid, 5%. These results demonstrate the absolute requirement of acylphosphate formation for the Ca2+ transport function and define a residue important for ATP binding. They also demonstrate the feasibility of a thorough analysis of active sites in the Ca2+-ATPase by expression and site-specific mutagenesis.

Structure of the rabbit fast-twitch skeletal muscle Ca2+-ATPase gene.

We have isolated two genomic clones which together encode the Ca2+-ATPase of rabbit fast-twitch skeletal muscle sarcoplasmic reticulum. One of the two 16.5 kilobase (kb) genomic inserts in the lambda phage vector Charon 4A contains 23 exons extending from the polyadenylation site at the 3' end of the ATPase gene to within 38 nucleotides of the translation initiation codon in the 5' exon. An overlapping genomic insert of 16.5 kb contains the remainder of the 5' exon and a further 8 kb of upstream sequence. S1 nuclease mapping and primer extension analysis of the 5' end of the Ca2+-ATPase mRNA indicate that the transcription initiation site is located 185 base pairs (bp) upstream of the translation initiation codon. A "TATA" box (CA-TAAA) was found at position -30 and the sequence CCAAT was found at position -78 relative to the transcription initiation site. In a previous study (Brandl, C. J., de Leon, S., Martin, D. R., and MacLennan, D. H. (1987) J. Biol. Chem. 262, 3768-3774) cDNAs for neonatal and adult forms of the fast-twitch Ca2+-ATPase were shown to encode different carboxyl-terminal sequences, presumably as a result of alternative splicing. We have now found that these different DNA sequences encoding different carboxyl-terminal sequences are located in different exons. Exon boundaries of the Ca2+-ATPase gene did not correlate well with proposed domain boundaries for the Ca2+-ATPase protein. The locations of exon/intron boundaries were only partially conserved between the Ca2+-ATPase gene and a Na+/K+-ATPase gene (Ovchinnikov, Y. A., Monastyrskaya, G. S., Broude, N. E., Allikmets, R. L., Ushkaryov, Y. A., Melkov, A. M., Smirnov, Y. V., Malyshev, I. V., Dulubova, I. E., Petrukhin, K. E., Gryshin, A. V., Sverdlov, V. E., Kiyatkin, N. I., Kostina, M. B., Modyanov, N. N., and Sverdlov, E. D. (1987) FEBS Lett. 213, 73-80) and they did not follow closely the boundaries of amino acid sequences that are highly conserved among a group of ion transport ATPases.