Reconstitution and purification of the sodium- and chloride-coupled gamma-aminobutyric acid transporter from rat brain

The sodium- and chloride-coupled gamma-aminobutyric acid transporter from rat brain has been highly purified. Synaptic plasma membranes from rat brain were extracted with cholate in the presence of 10% ammonium sulfate. The soluble extract was incorporated into liposomes consisting of asolectin and crude brain lipids. Brain lipids markedly enhanced the transport activity. The resulting proteoliposomes catalyzed sodium- and chloride-coupled gamma-aminobutyric acid transport which, in the presence of internal potassium, was greatly (up to 20-fold) stimulated by valinomycin. Using this transport of the reconstituted system as an assay, the transporter was purified by the following steps. The cholate extract was fractionated by ammonium sulfate. The activity was not precipitated by 50% but could be precipitated by 70% ammonium sulfate. The cholate and ammonium sulfate were removed on a Sephadex G-50 column. Subsequently, the transporter was partially purified on DEAE-cellulose in a mixture of Triton X-100 and octyl glucoside. The active fractions were chromatographed on a hydroxylapatite column in the presence of Triton X-100. Although the increase in specific activity was only up to 100-fold, this was due to partial inactivation. The actual purification was at least 1000-fold. The purified transporter exhibited the same features of the synaptic plasma membrane vesicles, namely dependence on sodium and chloride, electrogenicity, and a similar affinity. The sodium dodecyl sulfate gel pattern indicated that a major protein ran as a 24-kDa band. This band may represent the gamma-aminobutyric acid transporter.     

Purification and identification of the functional sodium- and chloride-coupled gamma-aminobutyric acid transport glycoprotein from rat brain

Using the reconstitution conditions developed recently (Radian, R., and Kanner, B. I. (1985) J. Biol. Chem. 260, 11859-11865) we have now purified the sodium- and chloride-coupled gamma-aminobutyric acid transporter from rat brain to apparent homogeneity. A partially purified transporter preparation was passed over wheat germ agglutinin-Sepharose 6MB and non-bound proteins were washed away. The transport activity, as expressed upon reconstitution of the protein into liposomes, was eluted by a solution containing Triton X-100 and N-acetylglucosamine. The specific transport activity was increased almost 400-fold over that of the crude extract. Taking into account an approximately 2.5-fold inactivation during the lectin column chromatography, the actual purification is about 1000-fold. Sodium dodecyl sulfate-polyacrylamide electrophoresis of the active fractions revealed one band of 80 kDa and small amounts of a band which ran at an apparent molecular mass of 160 kDa. The ratio between the two could be experimentally changed such as, for instance, by lyophilization. Polyclonal antibodies were prepared against the 80-kDa band which also cross-reacted with the 160-kDa band, indicating that the latter apparently represents a dimer form of the first. Using Protein A-Sepharose Cl-4B and the antibody against the 80-kDa band, we were able to quantitatively immunoprecipitate the potential gamma-aminobutyric acid transport activity from a crude transporter preparation. The pure transporter preparation exhibited the same features of the transporter in synaptic plasma membrane vesicles, namely dependence on sodium and chloride, electrogeneity, affinity, and efflux and exchange properties. We conclude that the 80-kDa band represents the gamma-aminobutyric acid transporter.     

Solubilization and reconstitution of the sodium-and-chloride-coupled glycine transporter from rat spinal cord

Synaptic membranes from rat spinal cord were solubilized in the presence of 2% sodium cholate, phospholipids and 15% ammonium sulphate. The soluble extract was incorporated into liposomes consisting of asolectin and crude rat brain lipids. Reconstitution of the functional transporter protein was achieved by removal of detergent by gel filtration. Several parameters proved to be important for optimal reconstitution efficiency: (a) the lipid composition of the liposomes, (b) the type of detergent, and (c) the phospholipid/protein and detergent/protein ratio during reconstitution. In the reconstituted system, the transport of glycine showed a specific activity about twice that of native vesicles. The ionic dependence of the transport, the inhibitory effect of nigericin in the presence of external sodium and the stimulatory effect of valinomycin in the presence of internal potassium on glycine transport were preserved and more clearly observed in the reconstituted system. These results indicate that, in this preparation, the glycine transporter protein retains the same features displayed in the synaptic plasma membrane vesicles, namely dependence on sodium and chloride, electrogenicity and inhibitor sensitivity.     

Glucose transport in vesicles reconstituted from Saccharomyces cerevisiae membranes and liposomes.

Glucose transport activity was reconstituted into liposomes by the freeze-thaw-sonication procedure from unextracted Saccharomyces cerevisiae membranes and preformed phospholipid liposomes. Fluorescence-dequenching measurements with octadecylrhodamine B chloride (R18)-labeled membranes showed that the yeast membrane lipids are diluted by the liposome lipids after the freeze-thaw-sonication procedure. At lipid-to-protein ratios greater than 75:1, vesicles with single transporters were formed. Reconstituted specific activity was increased at least twofold if the liposomes contained 50 mol% cholesterol. A further increase in specific activity, from 3- to 10-fold, was achieved by fractionation of the membranes on a Renografin gradient before reconstitution. Examination of the fractions from the Renografin gradient by sodium dodecyl sulfate-gel electrophoresis showed a parallel enrichment of glucose transport activity and a number of proteins including one with an apparent Mr of ca. 60,000, which might be the glucose transporter. Finally, preliminary kinetic analysis of glucose transport activity in vesicles reconstituted at a high lipid-to-protein ratio gave a Vmax of ca. 2.8 mumol/mg of protein per min at 23 degrees C and a Km of ca. 8 mM. The latter value corresponds to the kinase-independent, low-affinity component of glucose transport observed in wild-type cells.     

Low-affinity gamma-aminobutyric acid transport in rat brain

The low-affinity (Km = 100-200 microM) gamma-aminobutyric acid (GABA) transporter from membrane vesicles from rat brain has been characterized and found to be in many aspects similar to the well-known sodium- and chloride-coupled high-affinity gamma-aminobutyric acid transporter (Km = 2-4 microM). Influx by this system is sodium and chloride dependent and stimulated by an interior negative membrane potential. Steady-state levels obtained by both systems are lowered by the sodium channel openers veratridine and aconitine. However, while the channel blocker tetrodotoxin fully reverses this inhibition with the high-affinity system, this is not the case for its low-affinity counterpart. Furthermore, the toxin from the scorpion Androctonus australis Hector inhibited high-affinity transport only. Efflux of gamma-aminobutyric acid taken up by the high-affinity system displayed a Km of about 100 microM. Exchange catalyzed by the low-affinity system was observed in the absence of external sodium and chloride. Furthermore, both activities copurified in the fractionation procedure developed to purify the high-affinity transporter. All these observations are consistent with the idea that both activities are manifestations of only one gamma-aminobutyric acid transporter. The high-affinity binding site represents the extracellular and the low-affinity site the cytosolic aspect of the transporter. In addition, it was found that right-side-out synaptosomes also contain a low-affinity GABA transporter. This apparently represents a different transport protein.     

Volume enlargement and recovery of Na+-dependent amino acid transport in proteoliposomes derived from Ehrlich ascites cell membranes

Na+-dependent amino acid transport can be reconstituted from solubilized Ehrlich cell plasma membranes by addition of asolectin vesicles, gel filtration, and a freeze-thaw cycle. Removal of phosphatidic acid (approximately 10% of the total lipid) by Ba2+ precipitation reduces the efficiency of reconstitution of Na+-dependent amino acid transport by approximately 73% and decreases intravesicular volume of the proteoliposomes by approximately 43%. The loss of transport activity is not due to exclusion of specific proteins during reconstitution. The phosphatidic acid-free liposomes are less permeable and require more time to attain an equilibrium distribution of solute. Transport activity and intravesicular volume can be restored to Ba2+-precipitated asolectin proteoliposomes by addition of egg-phosphatidic acid during reconstitution. The extent of recovery of transport activity is proportional to the change in intravesicular volume and depends on the amount of phosphatidic acid present. Replacement of phosphatidic acid with 20% phosphatidylserine or phosphatidylglycerol leads to increases in intravesicular volume with little or no increase in amino acid transport. Generation of phosphatidic acid in situ by treatment of Ba2+-precipitated proteoliposomes with phospholipase D also restored transport. The observed increase in transport activity (9-fold) is accompanied by a 46% increase in intravesicular volume, presumably caused by vesicle fusion. Phosphatidic acid is also required for successful reconstitution of Na+-dependent amino acid transport from pure phosphatidylcholine:phosphatidylethanolamine (1:1) mixtures with only a small change (approximately 16%) in intravesicular volume. The results provide evidence for both indirect and direct effects of phosphatidic acid on reconstitution of Na+-dependent amino acid transport. The indirect effects occur through enlargement of intravesicular volume, large vesicles showing higher rates of transport. However, there is also evidence to indicate a specific effect of phosphatidic acid on the Na+-dependent amino acid transporter, since other acidic lipids may change intravesicular volume without a commensurate change in transport activity.     

Reconstitution and partial purification of the sodium and chloride-coupled glycine transporter from rat spinal cord

The (Na+ + Cl-)-coupled glycine transporter has been solubilized from rat spinal cord with 2% cholate and purified 6-7-fold using Wheat Germ Agglutinin-Sepharose 4B. Transport activity - as determined upon reconstitution of the fraction into liposomes - was retained on the column and eluted by N-acetylglucosamine. When the glycoprotein fraction was depleted of the N-acetylglucosamine and applied to a second round of lectin-chromatography, the glycine transport activity was retained and again could be eluted by the sugar. The transporter activity reconstituted from the glycoprotein fraction retains the same features displayed in the synaptic plasma membrane vesicles, namely an absolute dependence on sodium and chloride, electrogenicity and efflux and exchange properties. These observations indicate that the (Na+ + Cl-)-coupled glycine transporter is a glycoprotein.     

Electrogenic uptake of gamma-aminobutyric acid by a cloned transporter expressed in Xenopus oocytes.

GAT-1, a gamma-aminobutyric acid (GABA) transporter cloned from rat brain, was expressed in Xenopus oocytes. Voltage-clamp measurements showed concentration-dependent, inward currents in response to GABA (K0.5 4.7 microM). The transport current required extracellular sodium and chloride ions; the Hill coefficient for chloride was 0.7, and that for sodium was 1.7. Correlation of current and [3H]GABA uptake measurements indicate that flux of one positive charge occurs per molecule of GABA transported. Membrane hyperpolarization from -40 to -100 mV increased the transport current approximately 3-fold. The results indicate that the transport of one molecule of GABA involves the co-transport of two sodium ions and one chloride ion.     

Reconstitution into proteoliposomes and partial purification of the Golgi apparatus membrane UDP-galactose, UDP-xylose, and UDP-glucuronic acid transport activities.

Previous studies in vitro on proteoglycan biosynthesis from our laboratory have shown that nucleotide sugar precursors of all the sugars of the linkage oligosaccharides (xylose, galactose, and glucuronic acid) and of the glycosaminoglycans (N-acetylglucosamine, N-galactosamine, and glucuronic acid) are transported by specific carriers into the lumen of Golgi vesicles. More recently, we also reported the reconstitution in phosphatidylcholine liposomes of detergent-solubilized Golgi membrane proteins containing transport activities of CMP-sialic acid and adenosine-3'-phosphate-5'-phosphosulfate. We have now completed the successful reconstitution into liposomes of the Golgi membrane transport activities of UDP-galactose, UDP-xylose, and UDP-glucuronic acid. Transport of these nucleotide sugars into Golgi protein proteoliposomes occurred with the same affinity, temperature dependence, and sensitivity to inhibitors as observed with intact Golgi vesicles. Preloading of proteoliposomes with UMP, the putative antiporter for Golgi vesicle transport of these nucleotide sugars, stimulated transport of the nucleotide sugars by 2-3-fold. Transport of UDP-xylose into Golgi protein proteoliposomes was dependent on the presence of endogenous Golgi membrane lipids while that of UDP-galactose and UDP-glucuronic acid was not. This suggests a possible stabilizing or regulatory role for Golgi lipids on the UDP-xylose translocator. Finally, we have also shown that detergent-solubilized Golgi membrane translocator proteins can be partially purified by an ion-exchange chromatographic step before successful reconstitution into liposomes, demonstrating that this reconstitution approach can be used for the biochemical purification of these transporters.     

Purification and reconstitution of the sodium- and potassium-coupled glutamate transport glycoprotein from rat brain.

The sodium- and potassium-coupled L-glutamate transporter from rat brain has been purified to near homogeneity by reconstitution of transport as an assay, assuming that inactivated and active transporters cochromatograph. The purification steps involve lectin chromatography of the membrane proteins solubilized with 3-[(3-chloramidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), fractionation on hydroxylapatite, and ion-exchange chromatography. The specific activity is increased 30-fold. The actual purification is higher since 3-5-fold inactivation occurs during the purification. The efficiency of reconstitution was about 20%. The properties of the pure transporter are fully preserved. They include ion dependence, electrogenicity, affinity, substrate specificity, and stereospecificity. Sodium dodecyl sulfate-polyacrylamide electrophoresis revealed one main band with an apparent molecular mass of around 80 kDa and a few minor bands. Comparison of polypeptide composition with L-glutamate transport activity throughout the fractionation procedure reveals that only the 80-kDa band can be correlated with activity. The GABA transporter, which has the same apparent molecular mass (Radian et al., 1986), is separated from it during the last two purification steps. Immunoblot experiments reveal that the antibodies against the GABA transporter only reacted with fractions exhibiting GABA transport activity and not with those containing the glutamate transporter. We conclude that the 80-kDa band represents the functional sodium- and potassium-coupled L-glutamate transporter.     

Characterization of the ATP transporter in the reconstituted rough endoplasmic reticulum proteoliposomes

Adenosine triphosphate (ATP) transporter from rat liver rough endoplasmic reticulum (RER) was solubilized and reconstituted into phosphatidylcholine liposomes. The RER proteoliposomes, resulting from optimizing some reconstitution parameters, had an apparent K(m) value of 1.5 microM and a V(max) of 286 pmol min(-1) (mg protein)(-1) and showed higher affinity for ATP and a lower V(max) value than intact RER (K(m) of 6.5 microM and V(max) of 1 nmol). ATP transport was time- and temperature-dependent, inhibited by 4, 4'-diisothiocyanostilbene-2,2'-disulfonic acid, which is known as an inhibitor of anion transporters including ATP transporter, but was not affected by atractyloside, a specific inhibitor of mitochondrial ADP/ATP carrier. The internal and external effects of various nucleotides on the ATP transport were examined. ATP transport was cis-inhibited strongly by ADP and weakly by AMP. ADP-preloaded RER proteoliposomes showed a specific increase of ATP transport activity while AMP-preloaded RER proteoliposomes did not show the enhanced overshoot peak in the ATP uptake plot. These results demonstrate the ADP/ATP antiport mechanism of ATP transport in rat liver RER.     

Reconstitution of the glucose transporter from rat skeletal muscle

As a step in the purification and characterization of the glucose transporter from rat skeletal muscle, we have reconstituted glucose transport activity in liposomes. Plasma membranes were prepared from skeletal muscle which display D-glucose reversible binding of cytochalasin B (10 pmol sites/mg protein; KD = 0.3 microM). Older rats gave a slightly lower specific activity and much lower yield of sites per g muscle than young rats. Glucose transport activity was reconstituted into liposomes by the freeze-thaw procedure using either plasma membranes directly or cholate-extracted membrane proteins; the latter gave a 50% higher specific activity. The reconstituted transport activity was stereospecific, saturable, and inhibited by cytochalasin B, phloretin, and mercuric chloride. The optimum cholate concentration for extraction and reconstitution of transport activity was about 1.5%, and the highest specific activity of reconstituted transport was seen only at low ratios of protein to lipid in the reconstitution. Chromatography on agarose lentil lectin and agarose ethanethiol doubled both the specific activity of reconstituted transport and the fraction of glucose uptake which was stereospecific. In all of these respects the results were similar to our results with the bovine heart transporter (T. J. Wheeler and M. A. Hauck, Biochim. Biophys. Acta 818, 171-182 (1985)). Our findings suggest that further purification procedures developed for the heart transporter may be applicable to the skeletal muscle transporter as well.     

Purification of the sodium- and chloride-coupled glycine transporter from central nervous system.

We have recently developed a reconstitution assay which allows the rapid determination of sodium- and chloride-dependent glycine transport activity of many fractions (López-Corcuera, B., and Aragón, C. (1989) Eur. J. Biochem. 181, 519-524). In this paper we report the purification of the sodium- and chloride-coupled glycine transporter from pig brain stem. Transporter is solubilized from plasma membrane vesicles with 2% cholate and purified by sequential chromatography on phenyl-Sepharose, wheat germ agglutinin-Sepharose, and hydroxylapatite columns, followed by a 5-20% sucrose density gradient fractionation. Taking into account the inactivation suffered by the transporter, a final increase in specific activity of about 450-fold is achieved. Although two polypeptides with apparent molecular masses of 100 and 37 kDa are progressively enriched during the chromatographic steps, only the 100-kDa band comigrates with transport activity along the density gradient. This band is finally isolated to apparent homogeneity in the active fractions. We conclude that the 100-kDa band represents the glycine transporter. Finally, the pure transporter can be reconstituted into liposomes, retaining the absolute dependence on sodium and chloride gradients, the electrogenicity, the glycine affinity, the substrate specificity, and the sensitivity to group-selective modifiers characteristic of the native transporter.     

Molecular heterogeneity of the gamma-aminobutyric acid (GABA) transport system. Cloning of two novel high affinity GABA transporters from rat brain.

cDNA clones encoding two novel gamma-aminobutyric acid (GABA) transporters (designated GAT-2 and GAT-3) have been isolated from rat brain, and their functional properties have been examined in mammalian cells. The transporters display high affinity for GABA (Km approximately 10 microM) and exhibit pharmacological properties distinct from the previously cloned neuronal GABA transporter (GAT-1). Both transporters require sodium and chloride for transport activity. The nucleotide sequences of GAT-2 and GAT-3 predict proteins of 602 and 627 amino acids, respectively, which can be modeled with 12 transmembrane domains, similar to the topology proposed for other cloned neurotransmitter transporters. Localization studies indicate that both transporters are present in brain and retina, while GAT-2 is also present in peripheral tissues. The cloning of these transporter genes from rat brain reveals previously undescribed heterogeneity in GABA transporters.     

Partial purification of the sodium- and potassium-coupled L-glutamate transport glycoprotein from rat brain

The sodium- and potassium-coupled L-glutamate transporter from rat brain has been solubilized with cholate and 10-20-fold purified using Wheat Germ Agglutinin-Sepharose 4B. Transport activity--as determined upon reconstitution of the fraction into liposomes--was retained on the column and eluted by N-acetylglucosamine. When the glycoprotein fraction was depleted of the N-acetylglucosamine and applied to a second round of lectin-chromatography, the L-glutamate transport activity was retained and again could be eluted by the sugar. The transporter activity reconstituted from the glycoprotein fraction exhibited the same features as that in synaptic plasma membranes, including electrogenicity, an absolute dependence on external sodium and internal potassium, affinity and stereospecificity. Furthermore, efflux and exchange properties of the reconstituted preparation were also unchanged by the solubilisation and lectin-chromatography. These observations indicate that the sodium- and potassium-coupled L-glutamate transporter is a glycoprotein and is predominantly reconstituted in the 'right-side-out' conformation.     

Purification of the mitochondrial carnitine carrier by chromatography on hydroxyapatite and celite

The carnitine carrier from rat liver mitochondria has been extracted with Triton X-100 ad partially purified by chromatography on hydroxyapatite and celite. During purification the activity of the carrier was monitored by functional reconstitution into liposomes. The purified fraction is 250-fold enriched with respect to the N-ethylmaleimide-sensitive carnitine/carnitine transport activity. The substrate specificity and the inhibitor sensitivity of carnitine transport in liposomes resemble closely those described for the transport of carnitine in mitochondria.     

Asymmetric reconstitution of the glucose transporter from Ehrlich ascites cell plasma membrane: role of alkali cations

Gel chromatography of solubilized Ehrlich cell plasma membranes and preformed asolectin vesicles coupled to a freeze-thaw cycle results in the reconstitution of 3-O-methyl-D-glucose transport. The transport activity of the liposomes formed is critically dependent on the cation present during reconstitution. Liposomes formed in K+ show high levels of carrier-mediated 3-O-methyl-D-glucose uptake (495 pmol/min/mg protein) while those formed in Na+ do not (33 pmol/min/mg protein). The inactivity in Na+ is not due to a diminished incorporation of glucose transporter nor is it due to carrier molecules reconstituted with a different orientation from those in K+ liposomes. Instead, the low glucose transport level in Na+ liposomes is related to the small size of vesicles formed with Na+. A second freeze-thaw cycle in K+ causes a two- to threefold increase in the available intravesicular volume of Na+ liposomes and results in an eightfold increase in carrier-mediated 3-O-methyl-D-glucose uptake. K+ liposomes, treated in an identical manner, show only a twofold increase in uptake. The glucose transporter was identified as a protein with a molecular mass range of 44.7 to 66.8 kDa, by the D-glucose-inhibitable photoincorporation of [3H]cytochalasin B. The carrier protein is inserted in reconstituted vesicles in a nonrandom manner with at least 80% of the molecules oriented with the cytoplasmic domain accessible to the external medium. In contrast, the neutral Na+-dependent amino acid transport system appears to be randomly reconstituted.     

Neither amino nor carboxyl termini are required for function of the sodium- and chloride-coupled gamma-aminobutyric acid transporter from rat brain.

Antibodies were raised against synthetic peptides corresponding to several regions of the rat brain gamma-aminobutyric acid (GABA) transporter. According to our model, this glycoprotein has 12 transmembrane alpha-helices with both amino and carboxyl termini located in the cytoplasm. The antibodies recognized the intact transporter on Western blots. Upon papain treatment, a reconstitutively active transporter can be isolated upon lectin chromatography (Kanner, B. I., Keynan, S., and Radian, R. (1989) Biochemistry 28, 3722-3728). The papainized transporter runs on sodium dodecyl sulfate-polyacrylamide gels as a broad band with an apparent molecular mass between about 58 and 68 kDa as compared to 80 kDa for the untreated transporter. The transporter fragment was recognized by all the antibodies, except for that raised against the amino terminus. Pronase cleaves the transporter to a relatively sharp 60-kDa band, which reacts with the antibodies against the internal loops but not with either the amino- or the carboxyl-terminal. This pronase-treated transporter, upon isolation by lectin chromatography, was reconstituted. It exhibits full GABA transport activity. This activity exhibits the same features as the intact system including an absolute dependence on sodium and chloride as well as electrogenicity. We conclude that the amino- and carboxyl-terminal parts of the transporter, possibly including transmembrane alpha-helices 1, 2, and 12, are not required for the transport function.     

Mammalian Glucose Transporter Activity Is Dependent upon Anionic and Conical Phospholipids

The regulated movement of glucose across mammalian cell membranes is mediated by facilitative glucose transporters (GLUTs) embedded in lipid bilayers. Despite the known importance of phospholipids in regulating protein structure and activity, the lipid-induced effects on the GLUTs remain poorly understood. We systematically examined the effects of physiologically relevant phospholipids on glucose transport in liposomes containing purified GLUT4 and GLUT3. The anionic phospholipids, phosphatidic acid, phosphatidylserine, phosphatidylglycerol, and phosphatidylinositol, were found to be essential for transporter function by activating it and stabilizing its structure. Conical lipids, phosphatidylethanolamine and diacylglycerol, enhanced transporter activity up to 3-fold in the presence of anionic phospholipids but did not stabilize protein structure. Kinetic analyses revealed that both lipids increase the k_cat of transport without changing the K_m values. These results allowed us to elucidate the activation of GLUT by plasma membrane phospholipids and to extend the field of membrane protein-lipid interactions to the family of structurally and functionally related human solute carriers.     

Radiation inactivation studies on the rabbit kidney sodium-dependent glucose transporter

Rabbit kidney cortical brush-border membrane vesicles were irradiated in the frozen state with increasing doses of high energy electrons from a Van de Graaff generator. Sodium-dependent D-glucose and L-alanine transport showed a simple exponential loss of activity with increasing radiation dosage. Target size calculation based on these data gives estimates of 1.0 X 10(6) daltons for the glucose transporter and 1.2 X 10(6) daltons for the alanine transporter. A highly purified glucose transport protein extracted from rabbit kidney cortex was similarly irradiated both before and after reconstitution into liposomes. The target size of this purified glucose transporter was 343,000 daltons, based on inactivation of transport. The intensity of the major 165,000-dalton sodium dodecyl sulfate-gel electrophoresis band of this preparation was decreased by radiation. The decrease in staining intensity was dose-dependent, yielding a target size of 298,000 daltons, in situ. We propose that the purified glucose transporter reconstituted into liposomes is a tetramer comprised of 85,000-dalton subunits.