The membrane domain of 3-hydroxy-3-methylglutaryl-coenzyme A reductase confers endoplasmic reticulum localization and sterol-regulated degradation onto beta-galactosidase

A hybrid gene has been constructed consisting of coding sequence for the membrane domain of the endoplasmic reticulum protein 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase linked to the coding sequence for the soluble enzyme Escherichia coli beta-galactosidase. Expression of the hybrid gene in transfected Chinese hamster ovary cells results in the production of a fusion protein (HMGal) which is localized in the endoplasmic reticulum. The fusion protein contains the high-mannose oligosaccharides characteristic of HMG-CoA reductase. Importantly the beta-galactosidase activity of HMGal decreases when low density lipoprotein is added to the culture media. Therefore, the membrane domain of HMG-CoA reductase is sufficient to determine both correct intracellular localization and sterol-regulation of degradation. Mutant fusion proteins which lack 64, 85, or 98 amino acid residues from within the membrane domain of HMG-CoA reductase are found to be localized in the endoplasmic reticulum and to retain beta-galactosidase activity. However, sterol-regulation of degradation is abolished.     

3-Hydroxy-3-methylglutaryl-coenzyme A reductase of the sea urchin embryo. Deduced structure and regulatory properties

3-Hydroxy-3-methylglutaryl-coenzyme A reductase activity is developmentally regulated in the sea urchin Strongylocentrotus purpuratus (Woodward, H. D., Allen, J. M. C., and Lennarz, W. J. (1988) J. Biol. Chem. 263, 2513-2517). To study the structural and regulatory properties of this enzyme, we isolated and sequenced a 3-kb cDNA encoding the sea urchin embryo reductase. The deduced amino acid sequence of this cDNA predicted a protein structure consisting of a hydrophobic N-terminal region containing seven potential membrane-spanning domains and a somewhat less hydrophobic C-terminal domain joined by a hydrophilic linker region. Comparison with reductase from mammalian sources revealed that the N-terminal membrane domain and the C-terminal cytoplasmic domain exhibited high sequence similarity, whereas the domain that linked these two showed little or no sequence similarity. We investigated the possibility that sterols or sterol derivatives might be involved in the marked change that occurs in the level of reductase activity over development. Enzyme activity and reductase mRNA levels measured in extracts from embryos cultured in the presence of cholesterol, 25-hydroxycholesterol, dolichol, or mevalonic acid were found to be virtually unchanged as compared to control embryos. Similar experiments with mevinolin, a competitive inhibitor of reductase, failed to show a drug-induced change in enzyme or mRNA level. Thus, despite structural similarities the sea urchin embryo enzyme differs markedly from the mammalian enzyme with respect to regulation, since its level is neither repressed by sterols nor induced by mevinolin. Moreover, it appears unlikely that sterols or sterol derivatives play a role in the striking change in the level of this enzyme that occurs during development.     

Three genes encode 3-hydroxy-3-methylglutaryl-coenzyme A reductase in Hevea brasiliensis: hmg1 and hmg3 are differentially expressed

The enzyme 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGR) catalyses an important step in isoprenoid biosynthesis in plants. In Hevea brasiliensis, HMGR is encoded by a small gene family comprised of three members, hmg1, hmg2 and hmg3. We have previously described hmg1 and hmg2 (Plant Mol Biol 16: 567–577, 1991). Here we report the isolation and characterization of hmg3 genomic and cDNA clones. In comparison to hmg1 which is more highly expressed in laticifers than in leaves, the level of hmg3 mRNA level is equally abundant in laticifers and leaves. In situ hybridization experiments showed that the expression of hmg3 is not cell-type specific while hmg1 is expressed predominantly in the laticifers. Primer-extension experiments using laticifer RNA showed that hmg1 is induced by ethylene while hmg3 expression remains constitutive. The hmg3 promoter, like the promoters of most house-keeping genes, lacks a TATA box. Our results suggest that hmg1 is likely to encode the enzyme involved in rubber biosynthesis while hmg3 is possibly involved in isoprenoid biosynthesis of a housekeeping nature.     

3-Hydroxy-3-methylglutaryl-coenzyme A reductase from Haloferax volcanii: purification, characterization, and expression in Escherichia coli.

Prior work from this laboratory characterized eukaryotic (hamster) and eubacterial (Pseudomonas mevalonii) 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductases. We report here the characterization of an HMG-CoA reductase from the third domain, the archaea. HMG-CoA reductase of the halobacterium Haloferax volcanii was initially partially purified from extracts of H. volcanii. Subsequently, a portion of the H. volcanii lovastatin (formerly called mevinolin) resistance marker mev was subcloned into the Escherichia coli expression vector pT7-7. While no HMG-CoA reductase activity was detectable following expression in E. coli, activity could be recovered after extracts were exposed to 3 M KCl. Following purification to electrophoretic homogeneity, the specific activity of the expressed enzyme, 24 microU/mg, equaled that of homogeneous hamster or P. mevalonii HMG-CoA reductase. Activity was optimal at pH 7.3. Kms were 66 microM (NADPH) and 60 microM [(S)-HMG-CoA]. (R)-HMG-CoA and lovastatin inhibited competitively with (S)-HMG-CoA. H. volcanii HMG-CoA reductase also catalyzed the reduction of mevaldehyde [optimal activity at pH 6.0; Vmax 11 microU/mg; Kms 32 microM (NADPH), 550 microM [(R,S)-mevaldehyde]] and the oxidative acylation of mevaldehyde [optimal activity at pH 8.0; Vmax 2.1 microU/mg; Kms 350 microM (NADP+), 300 microM (CoA), 470 microM [(R,S)-mevaldehyde]]. These properties are comparable to those of hamster and P. mevalonii HMG-CoA reductases, suggesting a similar catalytic mechanism.     

IP3 receptor activity is differentially regulated in endoplasmic reticulum subdomains during oocyte maturation

Fertilization competency results from hormone-induced remodeling of oocytes into eggs. The signaling pathways that effect this change exemplify bistability, where brief hormone exposure irrevocably switches cell fate. In Xenopus, changes in Ca(2+) signaling epitomize such remodeling: The reversible Ca(2+) signaling phenotype of oocytes rapidly adapts to support irreversible propagation of the fertilization Ca(2+) wave. Here, we simultaneously resolved IP(3) receptor (IP(3)R) activity with endoplasmic reticulum (ER) structure to optically dissect the functional architecture of the Ca(2+) release apparatus underpinning this reorganization. We show that changes in Ca(2+) signaling correlate with IP(3)R redistribution from specialized ER substructures called annulate lamellae (AL), where Ca(2+) release activity is attenuated, into IP(3)R-replete patches in the cortical ER of eggs that support the fertilization Ca(2+) wave. These data show: first, that IP(3)R sensitivity is regulated with high spatial acuity even between contiguous ER regions; and second, that drastic reorganization of Ca(2+) signaling dynamics can be driven by subcellular redistribution in the absence of changes in channel number or molecular or familial Ca(2+) channel diversity. Finally, these results define a novel role for AL in Ca(2+) signaling. Because AL are prevalent in other scenarios of rapid cell division, further studies of their impact on Ca(2+) signaling are warranted.     

Assembly of the human T cell receptor-CD3 complex takes place in the endoplasmic reticulum and involves intermediary complexes between the CD3-gamma.delta.epsilon core and single T cell receptor alpha or beta chains

Functionally mature human T lymphocytes express a cell-surface receptor for antigen (T cell receptor (TCR)-CD3) composed of at least six polypeptides (TCR-alpha and -beta; T3-gamma, -delta, -epsilon, and -zeta). Immature thymocytes and variants of T cell lines lacking one of the TCR.CD3 polypeptide chains fail to express surface receptor and accumulate the other chains intracellularly. Here we show that the assembly of the TCR.CD3 complex within the endoplasmic reticulum (ER) began with a core of CD3-gamma, -delta, and -epsilon to which TCR-alpha and -beta bound. A recently described intracellular protein, CD3-omega, participated in the assembly since it was found to be associated with the free TCR-alpha or -beta chains or with the CD3 chains. CD3-omega dissociated as TCR.CD3 complexes were formed in the ER. Association of non-disulfide-linked TCR-alpha and -beta chains with CD3 was detected before that of disulfide-bridged TCR-alpha/beta heterodimers. These data suggest that during assembly, the association of TCR-alpha and -beta chains with the CD3 complex precedes the formation of a TCR-alpha/beta dimer. The existence of intermediates consisting of CD3-gamma, -delta, and -epsilon chains and a single TCR-alpha or -beta chain was also confirmed by using a series of variant T cell lines lacking the TCR-beta or -alpha chain, respectively. Once the single TCR-alpha and -beta chains were associated with CD3, disulfide linkages were formed, and a 70-kDa form of the TCR was detected within the ER. This intracellular precursor of the TCR.CD3 complex was subsequently processed into the mature 90-kDa TCR as the TCR.CD3 complex passed through the Golgi apparatus. Assembly of the TCR.CD3 complex is a rather rapid process, whereas export from the ER occurs at a slow rate. After 1 h, 75% of the receptor complex remained within the ER.     

Inactive 3-hydroxy-3-methylglutaryl-coenzyme A reductase in broken cell preparations of various mammalian tissues and cell cultures.

Preincubation of broken cell preparations from a variety of tissues and cell cultures resulted in an apparent increase in the level of 3-hydroxy-3-methylglutaryl-CoA reductase activity. However, apparent activation of the reductase in mouse liver, hepatomas and primary liver cell cultures was attributed largely to the loss, during the preincubation period, of an interfering enzyme, 3-hydroxy-3-methylglutaryl-CoA lyase. Among non hepatic cells and tissues (which did not contain appreciable lyase activity) the proportion of latent reductase was high in sonicates of fetal brain and in L cells and was independent of the level of total enzyme activity present. Activation of the reductase was blocked by hydroxymethylglutaryl-CoA and NADPH as well as by KF so that activation did not occur under the conditions of the enzyme assay. The enzyme was activated slowly at 4 degrees C, so that partial activation of the latent form occurred during isolation of the microsomal fraction by differential centrifugation. The reductase present in sonicates of cells with either a high or low proportion of the latent enzyme was inactivated by incubation with ATP and Mg2+. Suppression of reductase activity in L cell cultures by treatment with 25-hydroxycholesterol and an age-related decline in brain enzyme activity did not involve reversible conversion of the reductase to an inactive form.     

Independent regulation of 3-hydroxy-3-methylglutaryl-coenzyme A reductase and chylomicron remnant receptor activities in rat liver.

Treatment of rats with pharmacological doses of oestrogen resulted in a 3-fold decrease in the activity of hepatic 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMG-CoA reductase) and a 4-fold increase in saturable binding of 125I-labelled chylomicron remnants to liver membranes in vitro. Intragastric administration of mevalonolactone to rats did not affect the capacity of the liver membranes to bind to labelled chylomicron remnants even though there was a substantial decrease in the activity of HMG-CoA reductase. Similar results were obtained after cholesterol feeding. Simultaneous treatment of rats with cholestyramine and compactin increased hepatic HMG-CoA reductase activity 6-fold. However, liver membranes derived from these animals showed no change in their capacity to bind to labelled chylomicron remnants in vitro. Administration of mevalonolactone to the cholestyramine/compactin-treated animals also failed to produce a change in remnant-binding capacity. Although administration of mevalonolactone alone produced a significant 3-fold decrease in the activity of hepatic HMG-CoA reductase it was unable to suppress significantly the increase in enzyme activity caused by treatment with cholestyramine and compactin.     

The Alzheimer's disease-associated presenilins are differentially phosphorylated proteins located predominantly within the endoplasmic reticulum.

BACKGROUND: Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by the deposition of extracellular senile plaques composed of amyloid beta-peptide (A beta). Whereas most cases of AD occur sporadically, about 10% of AD cases are inherited as a fully penetrant autosomal dominant trait. Mutations in the recently cloned Presenilin genes (PS-1 and PS-2) are by far the most common cause of early onset familial AD. MATERIALS AND METHODS: Cellular expression of endogenous and overexpressed PS proteins was analyzed by immunocytochemistry and metabolic labeling followed by immunoprecipitation. In vivo phosphorylation sites of PS proteins were analyzed by extensive mutagenesis. RESULTS: PS-1 as well as PS-2 proteins were localized predominantly within the endoplasmic reticulum (ER). However, small amounts of the PS proteins were detected within the Golgi compartment, where they colocalize with the beta-amyloid precursor protein (beta APP). The PS-2 protein was found to be highly phosphorylated, whereas very little phosphorylation was observed for PS-1. The selective phosphorylation of PS-2 occurs exclusively on serine residues. In vivo phosphorylation of PS-2 was mapped to serine residues 7, 9, and 19 within an acidic stretch at the N terminus, which is absent in PS-1. casein kinase (CK)-1 and CK-2 were shown to phosphorylate the N terminus of PS-2 in vitro. CONCLUSIONS: The majority of PS proteins were detected in the ER where little if any proteolytic processing of beta APP was reported. ER retention of PS proteins might occur by intramolecular aggregation. Small amounts of PS proteins were also detected in the Golgi where they colocalized with beta APP. This might suggest that potential interactions between PS proteins and beta APP could occur within the Golgi. Selective phosphorylation of PS-2 proteins within the acidic domain missing in PS-1 indicates differences in the biological functions and regulation of the two highly homologous proteins.     

Developmental regulation of alpha beta T cell antigen receptor expression results from differential stability of nascent TCR alpha proteins within the endoplasmic reticulum of immature and mature T cells.

The alpha beta T cell antigen receptor (TCR) that is expressed on most T lymphocytes is a multisubunit transmembrane complex composed of at least six different proteins (alpha, beta, gamma, delta, epsilon and zeta) that are assembled in the endoplasmic reticulum (ER) and then transported to the plasma membrane. Expression of the TCR complex is quantitatively regulated during T cell development, with immature CD4+CD8+ thymocytes expressing only 10% of the number of surface alpha beta TCR complexes that are expressed on mature T cells. However, the molecular basis for low TCR expression in developing alpha beta T cells is unknown. In the present study we report the unexpected finding that assembly of nascent component chains into complete TCR alpha beta complexes is severely impaired in immature CD4+CD8+ thymocytes relative to their mature T cell progeny. In particular, the initial association of TCR alpha with TCR beta proteins, which occurs relatively efficiently in mature T cells, is markedly inefficient in immature CD4+CD8+ thymocytes, even for a matched pair of transgenic TCR alpha and TCR beta proteins. Inefficient formation of TCR alpha beta heterodimers in immature CD4+CD8+ thymocytes was found to result from the unique instability of nascent TCR alpha proteins within the ER of immature CD4+CD8+ thymocytes, with nascent TCR alpha proteins having a median survival time of only 15 min in CD4+CD8+ thymocytes, but > 75 min in mature T cells. Thus, these data demonstrate that stability of TCR alpha proteins within the ER is developmentally regulated and provide a molecular basis for quantitative differences in alpha beta TCR expression on immature and mature T cells. In addition, these results provide the first example of a receptor complex whose expression is quantitatively regulated during development by post-translational limitations on receptor assembly.     

The signal recognition particle receptor alpha subunit assembles co-translationally on the endoplasmic reticulum membrane during an mRNA-encoded translation pause in vitro.

Many proteins, including the alpha subunit of the signal recognition particle receptor (SR alpha), are targeted within the cell by poorly defined mechanisms. A 140 residue N-terminal domain of SR alpha targets and anchors the polypeptide to the endoplasmic reticulum membrane by a mechanism independent of the pathway involving the signal recognition particle. To investigate the mechanism of membrane anchoring, translation pause sites on the SR alpha mRNA were used to examine the targeting of translation intermediates. A strong pause site at nucleotide 507 of the mRNA open reading frame corresponded with the shortest nascent SR alpha polypeptide able to assemble on membranes. An mRNA sequence at this pause site that resembles a class of viral -1 frameshift sequences caused translation pausing when transferred into another mRNA context. Site-directed mutagenesis of the mRNA greatly reduced translation pausing without altering the polypeptide sequence, demonstrating unambiguously a role for this mRNA sequence in translation pausing. SR alpha polypeptides synthesized from the non-pausing mRNA were impaired in co-translational membrane anchoring. Furthermore, co-translational membrane assembly of SR alpha appears to anchor polysomes translating SR alpha to membranes.     

Regulated degradation of 3-hydroxy-3-methylglutaryl-coenzyme A reductase in permeabilized cells.

We have studied the regulated degradation of the enzyme 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase within the endoplasmic reticulum in cells permeabilized with digitonin. Using Chinese hamster ovary cells transfected with a plasmid encoding HMGal, a chimeric protein containing the membrane domain of HMG-CoA reductase coupled to beta-galactosidase, we have demonstrated mevalonate and sterol-stimulated loss of beta-galactosidase activity. In pulse-chase experiments we have demonstrated mevalonate-stimulated degradation of both HMGal and HMG-CoA reductase. The rate of mevalonate-stimulated degradation observed in permeabilized cells tends to be slightly slower than that observed in intact cells treated with mevalonate and is dependent upon incubation of cells with mevalonate prior to permeabilization. The degradation process measured in this report extends a previous report of HMG-CoA reductase degradation in digitonin-permeabilized cells (Leonard, D. A., and Chen, H. W. (1987) J. Biol. Chem. 262, 7914-7919) by mimicking key physiological features of the in vivo process, including: stimulation by regulatory molecules, specifically mevalonate and sterols; inhibition by cycloheximide; and inhibition by an inhibitor of neutral cysteine proteases.     

Characteristics of rat liver microsomal 3-hydroxy-3-methylglutaryl-coenzyme A reductase.

A procedure for the preparation of rat liver microsomal fractions essentially devoid of contaminating lysosomes is described. When this preparation was examined by immunoblotting with a rabbit antiserum to rat 3-hydroxy-3-methylglutaryl-CoA reductase, a single band corresponding to an Mr of 100000 was observed. No evidence was found for glycosylation of rat liver-3-hydroxy-3-methylglutaryl-CoA reductase. Native rat liver microsomal 3-hydroxy-3-methylglutaryl-CoA reductase differs from the purified proteolytically modified species in that it displays allosteric kinetics towards NADPH.     

Endoplasmic reticulum degradation of a subunit of the asialoglycoprotein receptor in vitro. Vesicular transport from endoplasmic reticulum is unnecessary.

The H2a subunit of the human asialoglycoprotein receptor is rapidly degraded from the endoplasmic reticulum (ER) when expressed in CHO15B cells. We have reconstituted ER degradation of H2a in semipermeable cells. At least the initial step in degradation (a proteolytic cleavage inhibited by N alpha-p-tosyl-L-lysine chloromethyl ketone and L-1-tosylamido-2-phenylethyl chloromethyl ketone) can occur in vitro in the presence of guanosine 5'-3-O-(thio)triphosphate or in the absence of ATP and postnuclear supernatant, conditions that do not allow vesicular transport of subunit H1 from the ER to the Golgi. We conclude that vesicular transport from the ER is not required for ER degradation of H2a to occur and thus that it takes place in the ER itself.     

Identification of the catalytically important histidine of 3-hydroxy-3-methylglutaryl-coenzyme A reductase.

We identify His381 of Pseudomonas mevalonii 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase as the basic residue functional in catalysis. The catalytic domain of 20 HMG-CoA reductases contains a single conserved histidine (His381 of the P. mevalonii enzyme). Diethyl pyrocarbonate inactivated the P. mevalonii enzyme, and hydroxylamine partially restored activity. We changed His381 to alanine, lysine, asparagine, and glutamine. The mutant proteins were overexpressed, purified to homogeneity, and characterized. His381 mutant enzymes were not inactivated by diethyl pyrocarbonate. All four mutant enzymes exhibited wild-type crystal morphology and chromatographed on substrate affinity supports like wild-type enzyme. The mutant enzymes had low catalytic activity (Vmax 0.06-0.5% that of wild-type enzyme), but Km values approximated those for wild-type enzyme. For wild-type enzyme and mutant enzymes H381A, H381N, and H381Q, Km values at pH 8.1 were 0.45, 0.27, 3.7, and 0.71 mM [(R,S)-mevalonate]; 0.05, 0.03, 0.20, and 0.11 mM [coenzyme A]; 0.22, 0.14, 0.81, and 0.62 mM [NAD+]. Km values at pH 11 for wild-type enzyme and mutant enzyme H381K were 0.32 and 0.75 mM [(R,S)-mevalonate]; 0.24 and 0.50 mM [coenzyme A]; 0.15 and 1.23 mM [NAD+]. Both pK values for the enzyme-substrate complex increased relative to wild-type enzyme (by 1-2.5 pH units for pK1 and by 0.5-1.3 pH units for pK2). For mutant enzyme H381K, the pK1 of 10.2 is consistent with lysine acting as a general base at high pH. His381 of P. mevalonii HMG-CoA reductase, and consequently the histidine of the consensus Leu-Val-Lys-Ser-His-Met-Xaa-Xaa-Asn-Arg-Ser motif of the catalytic domain of eukaryotic HMG-CoA reductases, thus is the general base functional in catalysis.     

Soluble aggregates of the human PiZ alpha 1-antitrypsin variant are degraded within the endoplasmic reticulum by a mechanism sensitive to inhibitors of protein synthesis.

Greater than 85% of the transport-impaired PiZ variant of human alpha 1-antitrypsin is retained within transfected mouse hepatoma cells and is subjected to intracellular degradation (Le, A., Graham, K., and Sifers, R.N. (1990) J. Biol. Chem. 265, 14001-14007). The retained protein undergoes a discrete size reduction that results from the modification of its endoglycosidase H-sensitive oligosaccharides and is inhibited by 1-deoxymannojirimycin. Metabolic poisons and inhibitors of protein synthesis perturb the intracellular degradation of the retained protein but do not affect its size reduction. The ability of metabolic poisons to influence the degradation of the PiZ variant in cells treated with brefeldin A indicates that export of the macromolecule from the endoplasmic reticulum (ER) is not the energy-dependent component of its degradation. Subcellular fractionation experiments have verified that both the size reduction and degradation of the retained PiZ variant occur within the rough ER. Finally, sedimentation velocity centrifugation analysis of radiolabeled cell extracts has indicated that approximately 80% of the PiZ variant consists as soluble aggregates immediately after its synthesis. An inability to detect more extensive aggregation during the retention period supports our previous conclusion that only a small fraction of the macromolecules actually form large insoluble aggregates (Graham, K.S., Le, A., and Sifers, R.N. (1990) J. Biol. Chem. 265, 20463-20468). Overall, these findings indicate that soluble aggregates of the PiZ variant are degraded within the ER by a mechanism sensitive to inhibitors of protein synthesis.     

Cross-linking of human T cell receptor proteins: association between the T cell idiotype beta subunit and the T3 glycoprotein heavy subunit

Bifunctional cross-linking reagents DSP, DSS, and BSOCOES were used to cross-link 125I-surface-labeled viable T lymphocytes. The cross-linked cells were solubilized in Nonidet-P40, immunoprecipitated with anti-Ti (monoclonal antibody T40/25) or anti-T3 (monoclonal antibodies UCHT-1 or OKT3), and analyzed by SDS-PAGE. With all three cross-linkers, the intact cross-linked products obtained with monoclonal antibody T40/25 from HPB-ALL cells were 20-30 kd heavier than the Ti dimer (Mr 80,000). When the DSP cross-linked product was isolated using either anti-Ti or anti-T3 monoclonal antibodies and then cleaved, bands having molecular weights identical with both the Ti and T3 subunits were obtained. The two-dimensional SDS-PAGE analysis (nonreducing followed by reducing conditions) of the DSS and BSOCOES cross-linked products revealed the specifically cross-linked bands to have Mr 40,000 and Mr 28,000. These data indicate that the Ti molecule and the T3 molecule are spatially associated on the cell surface and suggest the predominant association is between the Ti beta subunit (Mr 40,000) and the T3 heavy subunit (Mr 28,000).     

Domain structure of 3-hydroxy-3-methylglutaryl coenzyme A reductase, a glycoprotein of the endoplasmic reticulum

We present and evaluate a model for the secondary structure and membrane orientation of 3-hydroxy-3-methylglutaryl coenzyme A reductase, the glycoprotein of the endoplasmic reticulum that controls the rate of cholesterol biosynthesis. This model is derived from proteolysis experiments that separate the 97-kilodalton enzyme into two domains, an NH2-terminal membrane-bound domain of 339 residues and a COOH-terminal water-soluble domain of 548 residues that projects into the cytoplasm and contains the catalytic site. These domains were identified by reaction with antibodies against synthetic peptides corresponding to specific regions in the molecule. Computer modeling of the reductase structure, based on the amino acid sequence as determined by molecular cloning, predicts that the NH2-terminal domain contains 7 membrane-spanning regions. Analysis of the gene structure reveals that each proposed membrane-spanning region is encoded in a separate exon and is separated from the adjacent membrane-spanning region by an intron. The COOH-terminal domain of the reductase is predicted to contain two beta-structures flanked by a series of amphipathic helices, which together may constitute the active site. The NH2-terminal membrane-bound domain of the reductase bears some resemblance to rhodopsin, the photoreceptor protein of retinal rod disks and the only other intracellular glycoprotein whose amino acid sequence is known.