Membrane topology of the high-affinity L-glutamate transporter (GLAST- 1) of the central nervous system

The membrane topology of the high affinity, Na(+)-coupled L-glutamate/L- aspartate transporter (GLAST-1) of the central nervous system has been determined. Truncated GLAST-1 cDNA constructs encoding protein fragments with an increasing number of hydrophobic regions were fused to a cDNA encoding a reporter peptide with two N-glycosylation sites. The respective cRNA chimeras were translated in vitro and in vivo in Xenopus oocytes. Posttranslational N-glycosylation of the two reporter consensus sites monitors the number, size, and orientation of membrane- spanning domains. The results of our experiments suggest a novel 10- transmembrane domain topology of GLAST-1, a representative of the L- glutamate neurotransmitter transporter family, with its NH2 and COOH termini on the cytoplasmic side, six NH2-terminal hydrophobic transmembrane alpha-helices, and four COOH-terminal short hydrophobic domains spanning the bilayer predicted as beta-sheets.     

Transmembrane topology of the secretory Na+-K+-2Cl- cotransporter NKCC1 studied by in vitro translation

The secretory Na(+)-K(+)-2Cl(-) cotransporter NKCC1 is a member of a small gene family of electroneutral salt transporters. Hydropathy analyses indicate that all of these transporters have a similar general structure consisting of large hydrophilic N and C termini on either side of a central, relatively well conserved, hydrophobic domain. Programs that predict the transmembrane topology of polytopic membrane proteins identify 10-12 putative membrane-spanning segments (MSSs) in this hydrophobic domain; but to date, there is little experimental data on the structure of this region for any of these transporters. In this report, we have studied the transmembrane topology of NKCC1 using an in vitro translation system designed to test the membrane insertion properties of putative MSSs (Bamberg, K., and Sachs, G. (1994) J. Biol. Chem. 269, 16909-16919). Fusion proteins consisting of putative NKCC1 MSSs inserted either (i) between an N-terminal cytosolic anchor sequence and a C-terminal reporter sequence containing multiple N-linked glycosidation sites or (ii) between an N-terminal signal anchor sequence and the same glycosidation flag were expressed in the presence of canine pancreatic microsomes. The glycosidation status of the reporter sequence, which indicated its luminal or extraluminal location in the microsomes, was then used to characterize the signal anchor or stop transfer activity of the inserted MSSs. The results of this experimental analysis yielded a topology scheme consisting of 12 membrane-spanning segments, two pairs of which apparently form rather tight hairpin-like structures within the membrane.     

Interferon-induced Transmembrane Protein 3 Is a Type II Transmembrane Protein

The interferon-induced transmembrane (IFITM) proteins are a family of small membrane proteins that inhibit the cellular entry of several genera of viruses. These proteins had been predicted to adopt a two-pass, type III transmembrane topology with an intracellular loop, two transmembrane helices (TM1 and TM2), and extracellular N and C termini. Recent work, however, supports an intramembrane topology for the helices with cytosolic orientation of both termini. Here we determined the topology of murine Ifitm3. We found that the N terminus of Ifitm3 could be stained by antibodies at the cell surface but that this conformation was cell type-dependent and represented a minority of the total plasma membrane pool. In contrast, the C terminus was readily accessible to antibodies at the cell surface and extracellular C termini comprised most or all of those present at the plasma membrane. The addition of a C-terminal KDEL endoplasmic reticulum retention motif to Ifitm3 resulted in sequestration of Ifitm3 in the ER, demonstrating an ER-luminal orientation of the C terminus. C-terminal, but not N-terminal, epitope tags were also degraded within lysosomes, consistent with their luminal orientation. Furthermore, epitope-tagged Ifitm3 TM2 functioned as a signal anchor sequence when expressed in isolation. Collectively, our results demonstrate a type II transmembrane topology for Ifitm3 and will provide insight into its interaction with potential targets and cofactors.     

A method for determining the in vivo topology of yeast polytopic membrane proteins demonstrates that Gap1p fully integrates into the membrane independently of Shr3p

The general amino acid permease (Gap1p) of Saccharomyces cerevisiae is an integral membrane protein that contains 12 hydrophobic regions predicted to be membrane-spanning segments. A topological reporter construct, encoding an internal 53-amino acid peptide of invertase (Suc2p) containing three Asp-X-Ser/Thr glycosylation sites, was inserted in-frame into the hydrophilic NH(2)- and COOH-terminal domains and each of the 11 hydrophilic loops that separate the 12 hydrophobic segments of Gap1p. The resulting 13 gene sandwich fusion proteins were expressed in a gap1Delta null mutant strain; 9 of these retain amino acid transport activity and are folded and correctly targeted to the plasma membrane. The glycosylation state of each of the fusion proteins was monitored; the results indicate that all 12 hydrophobic segments of Gap1p span the membrane, and the NH(2) and COOH termini are cytoplasmically oriented. These results were independently tested by isolating sealed right-side-out microsomes from sec12-1 strains expressing six different Gap1p constructs containing functional factor Xa protease cleavage sites. The pattern of factor Xa protease cleavage was found to be consistent with the presence of 12 membrane-spanning domains. Gap1p exhibited the same membrane topology in strains lacking Shr3p; therefore, Gap1p fully integrates into the ER membrane independently of this permease-specific packaging chaperone.     

Topology of the C-terminal fragment of human presenilin 1

Mutations of human presenilin 1 (PS1) have been genetically linked to early-onset familial Alzheimer's disease. PS1 contains 10 hydrophobic regions (HRs) sufficiently long to be alpha-helical membrane spanning segments. Most previous topology studies agree that the N-terminus of PS1 is cytosolic and HRs 1-6 span the membrane but HR 7 does not. However, whether HRs 8 and 9 are membrane spanning segments remains controversial. Here we study the topology and biogenesis of this region of PS1 using a reporter gene fusion approach, where portions of the PS1 sequence containing possible membrane spanning segments were fused up- or downstream of a reporter sequence whose translocation into the endoplasmic reticulum could be monitored via its glycosylation. We provide strong evidence, supported by cysteine accessibility studies in full-length PS1, that HRs 8 and 9 are indeed membrane spanning and that the integration of HR 8 into the membrane is dependent on the presence of HR 9. We also explain how our results reconcile previous apparently divergent conclusions regarding the topology of HRs 8 and 9.     

Targeting proteins to the lumen of endoplasmic reticulum using N-terminal domains of 11beta-hydroxysteroid dehydrogenase and the 50-kDa esterase.

Previous studies identified two intrinsic endoplasmic reticulum (ER) proteins, 11beta-hydroxysteroid dehydrogenase, isozyme 1 (11beta-HSD) and the 50-kDa esterase (E3), sharing some amino acid sequence motifs in their N-terminal transmembrane (TM) domains. Both are type II membrane proteins with the C terminus projecting into the lumen of the ER. This finding implied that the N-terminal TM domains of 11beta-HSD and E3 may constitute a lumenal targeting signal (LTS). To investigate this hypothesis we created chimeric fusions using the putative targeting sequences and the reporter gene, Aequorea victoria green fluorescent protein. Transfected COS cells expressing LTS-green fluorescent protein chimeras were examined by fluorescent microscopy and electron microscopic immunogold labeling. The orientation of expressed chimeras was established by immunocytofluorescent staining of selectively permeabilized COS cells. In addition, protease protection assays of membranes in the presence and absence of detergents was used to confirm lumenal or the cytosolic orientation of the constructed chimeras. To investigate the general applicability of the proposed LTS, we fused the N terminus of E3 to the N terminus of the NADH-cytochrome b5 reductase lacking the myristoyl group and N-terminal 30-residue membrane anchor. The orientation of the cytochrome b5 reductase was reversed, from cytosolic to lumenal projection of the active domain. These observations establish that an amino acid sequence consisting of short basic or neutral residues at the N terminus, followed by a specific array of hydrophobic residues terminating with acidic residues, is sufficient for lumenal targeting of single-pass proteins that are structurally and functionally unrelated.     

Identification of two transmembrane regions and a cytosolic domain of rat mitochondrial glycerophosphate acyltransferase

The topography of rat glycerophosphate acyltransferase (GAT) in the transverse plane of the mitochondrial outer membrane (MOM) was investigated. Computer analysis of the amino acid (aa) sequence derived from rat mitochondrial GAT cDNA (GenBanktrade mark accession nos. and ) predicts the presence of two possible transmembrane domains (aa 473-493 and 574-594) separated by an 80-aa stretch (aa 494-573). To determine the actual orientation of the native protein, we prepared anti-peptide antibodies to three regions: one in between (aa 543-559) and the other two (aa 420-435 and 726-740) flanking the two putative transmembrane regions. Both immunoreaction and immunoprecipitation experiments employing intact and solubilized mitochondria indicate that regions on the N- and C-terminal sides of the transmembrane regions are sequestered on the inner surface of the MOM, while the region between the transmembrane domains is present on the cytosolic face of the MOM. Additionally, two green fluorescent protein (GFP) fusion proteins consisting of full-length GAT fused to GFP at either the C terminus or inserted 115 amino acids from the N terminus were also constructed to determine the orientation of the N and C termini. COS-1 cells expressing these fusion proteins were fractionated to obtain mitochondria. Protease digestion of intact and solubilized COS-1 cell mitochondria revealed that the GFP domains of these fusion proteins are sequestered on the inner side of the MOM. The present findings indicate that GAT is a dual-spanning, transmembrane protein adopting an inverted "U" conformation in the transverse plane of the MOM, where the N and C termini are sequestered on the inner surface of the MOM, while aa 494-573 are exposed on the cytosolic surface of the MOM.     

The retrieval function of the KDEL receptor requires PKA phosphorylation of its C-terminus

The KDEL receptor is a Golgi/intermediate compartment-located integral membrane protein that carries out the retrieval of escaped ER proteins bearing a C-terminal KDEL sequence. This occurs throughout retrograde traffic mediated by COPI-coated transport carriers. The role of the C-terminal cytoplasmic domain of the KDEL receptor in this process has been investigated. Deletion of this domain did not affect receptor subcellular localization although cells expressing this truncated form of the receptor failed to retain KDEL ligands intracellularly. Permeabilized cells incubated with ATP and GTP exhibited tubular processes-mediated redistribution from the Golgi area to the ER of the wild-type receptor, whereas the truncated form lacking the C-terminal domain remained concentrated in the Golgi. As revealed with a peptide-binding assay, this domain did not interact with both coatomer and ARF-GAP unless serine 209 was mutated to aspartic acid. In contrast, alanine replacement of serine 209 inhibited coatomer/ARF-GAP recruitment, receptor redistribution into the ER, and intracellular retention of KDEL ligands. Serine 209 was phosphorylated by both cytosolic and recombinant protein kinase A (PKA) catalytic subunit. Inhibition of endogenous PKA activity with H89 blocked Golgi-ER transport of the native receptor but did not affect redistribution to the ER of a mutated form bearing aspartic acid at position 209. We conclude that PKA phosphorylation of serine 209 is required for the retrograde transport of the KDEL receptor from the Golgi complex to the ER from which the retrieval of proteins bearing the KDEL signal depends.     

Structure of cytochrome b5 and its topology in the microsomal membrane

The complete amino acid sequence of human and chicken liver microsomal cytochrome b5 was determined. The amino termini of cytochrome b5 from four other mammalian species were examined in order to determine their complete covalent structure. As in the rat species, cytochrome b5 preparations from man, rabbit, calf and horse had an acetylated alanine as the first residue. In contrast, the pig cytochrome had alanine at the amino terminus. The amino terminus of the chicken cytochrome b5 was also unmodified, and extended three residues absent in the mammalian species. In order to investigate whether the carboxy-terminal segment of cytochrome b5 is located on the cytosolic or the luminal side of the microsomal membrane, rabbit liver microsomes were treated with trypsin and subjected to gel filtration and high-pressure liquid chromatography. The nonpolar peptide isolated from these microsomes lacked the terminal hexapeptide, indicating that when cytochrome b5 is bound to intact microsomes, the carboxy terminus is located on the cytosolic side of the membrane and does not extend in the lumen of the endoplasmic reticulum.     

Positive Charges Determine the Topology and Functionality of the Transmembrane Domain in the Chloroplastic Outer Envelope Protein Toc34

The chloroplastic outer envelope protein Toc34 is inserted into the membrane by a COOH-terminal membrane anchor domain in the orientation N_cyto-C_in. The insertion is independent of ATP and a cleavable transit sequence. The cytosolic domain of Toc34 does not influence the insertion process and can be replaced by a different hydrophilic reporter peptide. Inversion of the COOH-terminal, 45-residue segment, including the membrane anchor domain (Toc34Cinv), resulted in an inverted topology of the protein, i.e., N_in-C_cyto. A mutual exchange of the charged amino acid residues NH₂- and COOH-proximal of the hydrophobic α-helix indicates that a double-positive charge at the cytosolic side of the transmembrane α-helix is the sole determinant for its topology. When the inverted COOH-terminal segment was fused to the chloroplastic precursor of the ribulose-1,5-bisphosphate carboxylase small subunit (pS34Cinv), it engaged the transit sequence–dependent import pathway. The inverted peptide domain of Toc34 functions as a stop transfer signal and is released out of the outer envelope protein translocation machinery into the lipid phase. Simultaneously, the NH₂-terminal part of the hybrid precursor remained engaged in the inner envelope protein translocon, which could be reversed by the removal of ATP, demonstrating that only an energy-dependent force but no further ionic interactions kept the precursor in the import machinery.     

The juxtamembrane sequence of cytochrome P-450 2C1 contains an endoplasmic reticulum retention signal

The N-terminal signal anchor of cytochrome P-450 2C1 mediates retention in the endoplasmic reticulum (ER) membrane of several reporter proteins. The same sequence fused to the C terminus of the extracellular domain of the epidermal growth factor receptor permits transport of the chimeric protein to the plasma membrane. In the N-terminal position, the ER retention function of this signal depends on the polarity of the hydrophobic domain and the sequence KQS in the short hydrophilic linker immediately following the transmembrane domain. To determine what properties are required for the ER retention function of the signal anchor in a position other than the N terminus, the effect of mutations in the linker and hydrophobic domains on subcellular localization in COS1 cells of chimeric proteins with the P-450 signal anchor in an internal or C-terminal position was analyzed. For the C-terminal position, the signal anchor was fused to the end of the luminal domain of epidermal growth factor receptor, and green fluorescent protein was additionally fused at the C terminus of the signal anchor for the internal position. In these chimeras, the ER retention function of the signal anchor was rescued by deletion of three leucines at the C-terminal side of its hydrophobic domain; however, deletion of three valines from the N-terminal side did not affect transport to the cell surface. ER retention of the C-terminal deletion mutants was eliminated by substitution of alanines for glutamine and serine in the linker sequence. These data are consistent with a model in which the position of the linker sequence at the membrane surface, which is critical for ER retention, is dependent on the transmembrane domain.     

Membrane topology of the hepatitis C virus NS2 protein

The hepatitis C virus (HCV) NS2 protein is a hydrophobic protein. Previous studies indicate that this protein is an integral membrane protein, which is targeted to the membrane of the endoplasmic reticulum (ER) by the signal sequence located in its preceding p7 protein. In this report, we demonstrate that the membrane association of NS2 is p7-independent and occurs co-translationally. Further deletion-mapping studies suggest the presence of two internal signal sequences in NS2. These two internal signal sequences, which are located within amino acids 839-883 and amino acids 928-960, could target the alpha-globin reporter, a cytosolic protein, to the membrane compartments in HuH7 hepatoma cells. The presence of multiple signal sequences for its membrane association suggests that NS2 has multiple transmembrane domains. The glycosylation studies indicate that both amino and carboxyl termini of NS2 are located in the endoplasmic reticulum lumen. Based on these results, a model for the NS2 membrane topology is presented.     

Topology of the hydrophobic membrane-bound components of the histidine periplasmic permease. Comparison with other members of the family.

The membrane-bound complex of periplasmic permeases comprises two hydrophobic proteins which have been hypothesized to be integral membrane-spaninning proteins. We have investigated the topological organization of the hydrophobic components of the Salmonella typhimurium histidine permease, HisQ and HisM. Both proteins are digested by trypsin and proteinase K when either inside-out or right-side-out membrane vesicles are used. Therefore, these proteins are exposed to both surfaces of the membrane. Digestion with carboxypeptidase and binding studies with antibodies directed against the carboxyl terminus of HisQ and HisM have localized their carboxyl termini to the inside surface of the cytoplasmic membrane. Aminopeptidase digestion suggests periplasmic localization of their amino termini. Alkaline phosphatase fusions to HisQ and HisM indicate the existence of five spanners in both proteins. The periodicity and orientation of spanners and loops in HisQ and HisM match those of the five carboxyl-terminal spanners of MalF, the only other hydrophobic component of the periplasmic permeases for which topological information is available. An alignment of the sequences of all known hydrophobic components of periplasmic permeases is presented which indicates clear conservation of secondary structure and some conservation of primary sequence. The structural conservation of the components is discussed, and a role for a hydrophilic loop containing a conserved sequence (the EAA loop) is proposed.     

Targeting and topology in the membrane of plant 3-hydroxy-3-methylglutaryl coenzyme A reductase.

The enzyme 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR) catalyzes the synthesis of mevalonate. This is the first committed step of isoprenoid biosynthesis. A common feature of all known plant HMGR isoforms is the presence of two highly conserved hydrophobic sequences in the N-terminal quarter of the protein. Using an in vitro system, we showed that the two hydrophobic sequences of Arabidopsis HMGR1S function as internal signal sequences. Specific recognition of these sequences by the signal recognition particle mediates the targeting of the protein to microsomes derived from the endoplasmic reticulum. Arabidopsis HMGR is inserted into the microsomal membrane, and the two hydrophobic sequences become membrane-spanning segments. The N-terminal end and the C-terminal catalytic domain of Arabidopsis HMGR are positioned on the cytosolic side of the membrane, whereas only a short hydrophilic sequence is exposed to the lumen. Our results suggest that the plant HMGR isoforms known to date are primarily targeted to the endoplasmic reticulum and have the same topology in the membrane. This reinforces the hypothesis that mevalonate is synthesized only in the cytosol. The possibility that plant HMGRs might be located in different regions of the endomembrane system is discussed.     

Biogenesis and topology of the transient receptor potential Ca2+ channel TRPC1

The TRPC ion channels are candidates for the store-operated Ca(2+) entry pathway activated in response to depletion of intracellular Ca(2+) stores. Hydropathy analyses indicate that these proteins contain eight hydrophobic regions (HRs) that could potentially form alpha-helical membrane-spanning segments. Based on limited sequence similarities to other ion channels, it has been proposed that only six of the eight HRs actually span the membrane and that the last two membrane-spanning segments (HRs 6 and 8) border the ion-conducting pore of which HR 7 forms a part. Here we study the biogenesis and transmembrane topology of human TRPC1 to test this model. We have employed a truncation mutant approach combined with insertions of glycosylation sites into full-length TRPC1. In our truncation mutants, portions of the TRPC1 sequence containing one or more HRs were fused between the enhanced green fluorescent protein and a C-terminal glycosylation tag. These chimeras were transiently expressed in the human embryonic cell line HEK-293T. Glycosylation of the tag was used to monitor its location relative to the lumen of the endoplasmic reticulum and thereby HR orientation. Our data indicate that HRs 1, 4, and 6 cross the membrane from cytosol to the ER lumen, that HRs 2, 5, and 8 have the opposite orientation, and that HR 3 is left out of the membrane on the cytosolic side. Our results also show that the sequence downstream of HR 8 plays an important role in anchoring its C-terminal end on the cytosolic side of the membrane. This effect appears to prevent HR 7 from spanning the bilayer and to result in its forming a pore-like structure of the type previously envisioned for the TRPC channels. We speculate that a similar mechanism may be responsible for the formation of other ion channel pores.     

Expression of fusion proteins of the nicotinic acetylcholine receptor from mammalian muscle identifies the membrane-spanning regions in the alpha and delta subunits

We have investigated the topology of the alpha and delta subunits of the nicotinic acetylcholine receptor (AChR) from mammalian muscle synthesized in an in vitro translation system supplemented with dog pancreatic microsomes. Fusion proteins were expressed in which a carboxy-terminal fragment of bovine prolactin was attached downstream of each of the major putative transmembrane domains, M1-M4 and MA, in the AChR subunits. The orientation of the prolactin domain relative to the microsomal membrane was then determined for each protein by a proteolysis protection assay. Since the prolactin domain contains no information which either directs or prevents its translocation, its transmembrane orientation depends solely on sequences within the AChR subunit portion of the fusion protein. When subunit-prolactin fusion proteins with the prolactin domain fused after either M2 or M4 were tested, prolactin-immunoreactive peptides that were larger than the prolactin domain itself were recovered. No prolactin-immunoreactive peptides were recovered after proteolysis of fusion proteins containing prolactin fused after M1, M3, or MA. These results support a model of AChR subunit topology in which M1-M4, but not MA, are transmembrane domains and the carboxy terminus is extracellular.     

Subcellular localization and membrane topology of the melon ethylene receptor CmERS1

Ethylene receptors are multispanning membrane proteins that negatively regulate ethylene responses via the formation of a signaling complex with downstream elements. To better understand their biochemical functions, we investigated the membrane topology and subcellular localization of CmERS1, a melon (Cucumis melo) ethylene receptor that has three putative transmembrane domains at the N terminus. Analyses using membrane fractionation and green fluorescent protein imaging approaches indicate that CmERS1 is predominantly associated with the endoplasmic reticulum (ER) membrane. Detergent treatments of melon microsomes showed that the receptor protein is integrally bound to the ER membrane. A protease protection assay and N-glycosylation analysis were used to determine membrane topology. The results indicate that CmERS1 spans the membrane three times, with its N terminus facing the luminal space and the large C-terminal portion lying on the cytosolic side of the ER membrane. This orientation provides a platform for interaction with the cytosolic signaling elements. The three N-terminal transmembrane segments were found to function as topogenic sequences to determine the final topology. High conservation of these topogenic sequences in all ethylene receptor homologs identified thus far suggests that these proteins may share the same membrane topology.     

Topology of the non-structural rotavirus receptor glycoprotein NS28 in the rough endoplasmic reticulum.

The rotavirus non-structural glycoprotein (NS28), the receptor for the virus core during budding into the lumen of the rough endoplasmic reticulum (RER), is 175 amino acids long and possesses an uncleaved signal sequence and two amino-terminal glycosylation sites. Utilizing one of three potential hydrophobic domains, the protein spans the membrane only once, with the glycosylated amino-terminal region oriented to the luminal side of the ER and the carboxy-terminal region to the cytoplasmic side. To localize sequences involved in translocation of NS28, we constructed a series of mutations in the coding regions for the hydrophobic domains of the protein. Mutant protein products were studied by in vitro translation and by transfection in vivo. In transfected cells, all mutant forms localize to the ER, and none are secreted. In vitro, each of the three hydrophobic domains is able to associate with microsomes. However, glycosylation and proteolysis of wild-type and mutant forms of NS28 indicates that the wild-type protein is anchored in the membrane only by the second hydrophobic domain, leaving approximately 131 residues exposed on the cytoplasmic side for receptor - ligand interaction.     

Evidence for a COP-I-independent transport route from the Golgi complex to the endoplasmic reticulum

The cytosolic coat-protein complex COP-I interacts with cytoplasmic 'retrieval' signals present in membrane proteins that cycle between the endoplasmic reticulum (ER) and the Golgi complex, and is required for both anterograde and retrograde transport in the secretory pathway. Here we study the role of COP-I in Golgi-to-ER transport of several distinct marker molecules. Microinjection of anti-COP-I antibodies inhibits retrieval of the lectin-like molecule ERGIC-53 and of the KDEL receptor from the Golgi to the ER. Transport to the ER of protein toxins, which contain a sequence that is recognized by the KDEL receptor, is also inhibited. In contrast, microinjection of anti-COP-I antibodies or expression of a GTP-restricted Arf-1 mutant does not interfere with Golgi-to-ER transport of Shiga toxin/Shiga-like toxin-1 or with the apparent recycling to the ER of Golgi-resident glycosylation enzymes. Overexpression of a GDP-restricted mutant of Rab6 blocks transport to the ER of Shiga toxin/Shiga-like toxin-1 and glycosylation enzymes, but not of ERGIC-53, the KDEL receptor or KDEL-containing toxins. These data indicate the existence of at least two distinct pathways for Golgi-to-ER transport, one COP-I dependent and the other COP-I independent. The COP-I-independent pathway is specifically regulated by Rab6 and is used by Golgi glycosylation enzymes and Shiga toxin/Shiga-like toxin-1.     

Membrane topography of human phosphatidylethanolamine N-methyltransferase

In liver, phosphatidylethanolamine is converted to phosphatidylcholine through a series of three sequential methylation reactions. Phosphatidylethanolamine N-methyltransferase (PEMT) catalyzes each transmethylation reaction, and S-adenosylmethionine is the methyl group donor. Biochemical analysis of human liver revealed that the methyltransferase activity is primarily localized to the endoplasmic reticulum and mitochondria-associated membranes. Bioinformatic analysis of the predicted amino acid sequence suggested that the enzyme adopts a polytopic conformation in those membranes. To elucidate the precise membrane topography of PEMT and thereby provide the basis for in-depth functional characterization of the enzyme, we performed endoproteinase-protection analysis of epitope-tagged, recombinant protein. Our data suggest a topographical model of PEMT in which four transmembrane regions span the membrane such that both the N and C termini of the enzyme are localized external to the ER. Two hydrophilic connecting loops protrude into the luminal space of the microsomes whereas a corresponding loop on the cytosolic side remains proximate to the membrane. Further support for this model was obtained following endoproteinase-protection analysis of mutant recombinant PEMT derivatives in which specific protease cleavage sites had been genetically engineered or ablated.