Membrane Topology Mapping of Vitamin K Epoxide Reductase by in Vitro Translation/Cotranslocation

Abstract

Vitamin K epoxide reductase (VKOR) catalyzes the conversion of vitamin K 2,3-epoxide into vitamin K in the vitamin K redox cycle. Recently, the gene encoding the catalytic subunit of VKOR was identified as a 163-amino acid integral membrane protein. In this study we report the experimentally derived membrane topology of VKOR. Our results show that four hydrophobic regions predicted as the potential transmembrane domains in VKOR can individually insert across the endoplasmic reticulum membrane in vitro. However, in the intact enzyme there are only three transmembrane domains, residues 10–29, 101–123, and 127–149, and membrane-integration of residues 75–97 appears to be suppressed by the surrounding sequence. Results of N-linked glycosylation-tagged full-length VKOR shows that the N terminus of VKOR is located in the endoplasmic reticulum lumen, and the C terminus is located in the cytoplasm. Further evidence for this topological model of VKOR was obtained with freshly prepared intact microsomes from insect cells expressing HPC4-tagged full-length VKOR. In these experiments an HPC4 tag at the N terminus was protected from proteinase K digestion, whereas an HPC4 tag at the C terminus was susceptible. Altogether, our results suggest that VKOR is a type III membrane protein with three transmembrane domains, which agrees well with the prediction by the topology prediction program TMHMM. Vitamin K epoxide reductase (VKOR) catalyzes the conversion of vitamin K 2,3-epoxide into vitamin K in the vitamin K redox cycle. Recently, the gene encoding the catalytic subunit of VKOR was identified as a 163-amino acid integral membrane protein. In this study we report the experimentally derived membrane topology of VKOR. Our results show that four hydrophobic regions predicted as the potential transmembrane domains in VKOR can individually insert across the endoplasmic reticulum membrane in vitro. However, in the intact enzyme there are only three transmembrane domains, residues 10–29, 101–123, and 127–149, and membrane-integration of residues 75–97 appears to be suppressed by the surrounding sequence. Results of N-linked glycosylation-tagged full-length VKOR shows that the N terminus of VKOR is located in the endoplasmic reticulum lumen, and the C terminus is located in the cytoplasm. Further evidence for this topological model of VKOR was obtained with freshly prepared intact microsomes from insect cells expressing HPC4-tagged full-length VKOR. In these experiments an HPC4 tag at the N terminus was protected from proteinase K digestion, whereas an HPC4 tag at the C terminus was susceptible. Altogether, our results suggest that VKOR is a type III membrane protein with three transmembrane domains, which agrees well with the prediction by the topology prediction program TMHMM. The K vitamins, phylloquinone (K1), menaquinones (K2), and menadione (K3), are a family of structurally similar, fatsoluble, 2-methyl-1,4-naphthoquinones. The main function of vitamin K is to act as a co-factor for the γ-glutamyl carboxylase that catalyzes the post-translational carboxylation of specific glutamic acid to γ-carboxyglutamic acid (Gla) 1The abbreviations used are: Gla, γ-carboxyglutamic acid; VKOR, vitamin K epoxide reductase; VKORC1, vitamin K epoxide reductase complex subunit 1; TM, transmembrane; HPC4, peptide epitope comprising residues; EDQVDPRLIDGK, ER endoplasmic reticulum; RM, canine rough microsomes; MES, 2-(N-morpholino)ethanesulfonic acid, hemisodium salt; CHAPS, (3-[(3-cholamidopropyl)di-methylammonio]-1-propane sulfonate); Endo H, endoglycosidase H; PVDF, polyvinylidene difluoride. of variety of vitamin K-dependent proteins (1Presnell S.R. Stafford D.W. Thromb. Haemostasis. 2002; 87: 937-946Crossref PubMed Scopus (83) Google Scholar). Members of the vitamin K-dependent protein family include coagulation factors (factor II, VII, IX, X) as well as several other proteins that function in bone metabolism (2Price P.A. Annu. Rev. Nutr. 1988; 8: 565-583Crossref PubMed Scopus (151) Google Scholar) and signal transduction (3Manfioletti G. Brancolini C. Avanzi G. Schneider C. Mol. Cell. Biol. 1993; 13: 4976-4985Crossref PubMed Scopus (532) Google Scholar). Concomitant with γ-glutamyl carboxylation, the reduced form of vitamin K (vitamin K hydroquinone) is converted to vitamin K 2,3-epoxide, which must be converted back to vitamin K hydroquinone for the reaction to continue because of limited vitamin K amounts in vivo (4Furie B. Bouchard B.A. Furie B.C. Blood. 1999; 93: 1798-1808Crossref PubMed Google Scholar). This cyclic conversion of vitamin K establishes a redox cycle known as the vitamin K cycle (5Saxena S.P. Israels E.D. Israels L.G. Apoptosis. 2001; 6: 57-68Crossref PubMed Scopus (29) Google Scholar). VKOR is responsible for the conversion of vitamin K 2,3-epoxide into vitamin K and is highly sensitive to inhibition by coumarin drugs, such as R,S-warfarin (4-hydroxy-3-(3-oxo-1-phenylbutyl-coumarin)), the most commonly prescribed oral anticoagulant. Warfarin inhibition of VKOR reduces the availability of reduced vitamin K, which reduces the rate of carboxylation and leads to partially carboxylated, inactive vitamin K-dependent proteins. Since its discovery in 1970 (6Bell R.G. Matschiner J.T. Arch. Biochem. Biophys. 1970; 141: 473-476Crossref PubMed Scopus (61) Google Scholar), numerous futile attempts to purify the enzyme were reported (7Mukharji I. Silverman R.B. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 2713-2717Crossref PubMed Scopus (25) Google Scholar, 8Lee J.J. Principe L.M. Fasco M.J. Biochemistry. 1985; 24: 7063-7070Crossref PubMed Scopus (17) Google Scholar, 9Wallin R. Guenthner T.M. Methods Enzymol. 1997; 282: 395-408Crossref PubMed Scopus (12) Google Scholar, 10Cain D. Hutson S.M. Wallin R. J. Biol. Chem. 1997; 272: 29068-29075Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar, 11Begent L.A. Hill A.P. Steventon G.B. Hutt A.J. Pallister C.J. Cowell D.C. J. Pharm. Pharmacol. 2001; 53: 481-486Crossref PubMed Scopus (18) Google Scholar). Attempts to understand the mechanism underlying warfarin-sensitive vitamin K epoxide reduction have been somewhat more successful (8Lee J.J. Principe L.M. Fasco M.J. Biochemistry. 1985; 24: 7063-7070Crossref PubMed Scopus (17) Google Scholar, 12Silverman R.B. Nandi D.L. Biochem. Biophys. Res. Commun. 1988; 155: 1248-1254Crossref PubMed Scopus (65) Google Scholar, 13Preusch P.C. FEBS Lett. 1992; 305: 257-259Crossref PubMed Scopus (28) Google Scholar, 14Thijssen H.H. Janssen Y.P. Vervoort L.T. Biochem. J. 1994; 297: 277-280Crossref PubMed Scopus (22) Google Scholar, 15Wallin R. Hutson S.M. Cain D. Sweatt A. Sane D.C. FASEB J. 2001; 15: 2542-2544Crossref PubMed Scopus (83) Google Scholar). Recently, the gene encoding VKOR was identified independently by our laboratory (16Li T. Chang C.Y. Jin D.Y. Lin P.J. Khvorova A. Stafford D.W. Nature. 2004; 427: 541-544Crossref PubMed Scopus (598) Google Scholar) and that of Johannes Oldenburg (17Rost S. Fregin A. Ivaskevicius V. Conzelmann E. Hortnagel K. Pelz H.J. Lappegard K. Seifried E. Scharrer I. Tuddenham E.G. Muller C.R. Strom T.M. Oldenburg J. Nature. 2004; 427: 537-541Crossref PubMed Scopus (995) Google Scholar). The enzyme was designated VKOR by us and VKORC1 by Oldenburg and co-workers, who assume that there is another sub-unit still to be identified (either view may be correct). VKOR is a 163-amino acid integral membrane protein with a mass of 18.2 kDa. To understand the structure/function relationships that define VKOR activity, it is necessary to understand its membrane topology, i.e. its specific number of TM segments and their location relative to the cytoplasm or endoplasmic reticulum. Several computer programs to predict the topology of membrane proteins are available (18Moller S. Croning M.D. Apweiler R. Bioinformatics. 2001; 17: 646-653Crossref PubMed Scopus (860) Google Scholar, 19Nilsson J. Persson B. von Heijne G. FEBS Lett. 2000; 486: 267-269Crossref PubMed Scopus (93) Google Scholar, 20Melen K. Krogh A. von Heijne G. J. Mol. Biol. 2003; 327: 735-744Crossref PubMed Scopus (174) Google Scholar, 21Taylor P.D. Attwood T.K. Flower D.R. Nucleic Acids Res. 2003; 31: 3698-3700Crossref PubMed Scopus (34) Google Scholar). If the location of one of the termini is known, the best current program correctly predicts the topology of 65–70% of those membrane proteins with known crystal structure (20Melen K. Krogh A. von Heijne G. J. Mol. Biol. 2003; 327: 735-744Crossref PubMed Scopus (174) Google Scholar). The reliability of a given topology prediction is increased if the predictions from a number of different programs agree (19Nilsson J. Persson B. von Heijne G. FEBS Lett. 2000; 486: 267-269Crossref PubMed Scopus (93) Google Scholar, 20Melen K. Krogh A. von Heijne G. J. Mol. Biol. 2003; 327: 735-744Crossref PubMed Scopus (174) Google Scholar). Moreover, using C-terminal reporter fusions, experimental examination of the topology is reliable (22Kim H. Melen K. von Heijne G. J. Biol. Chem. 2003; 278: 10208-10213Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar). In this study, we report an experimental determination of the membrane topology of VKOR. We first determined the topology of VKOR using multiple prediction programs. These predictions were then tested by in vitro translation/cotranslocation of a series of VKOR truncations an N-linked reporter Our results suggest that VKOR is a type III membrane protein with three transmembrane The N terminus of VKOR in the of the endoplasmic and the C terminus in the cytoplasm. The of VKOR, type III membrane is which its into the V. FEBS Lett. 2001; PubMed Scopus Google Scholar). The main of our in vitro experiments were by experiments with microsomes derived from insect cells expressing full-length VKOR. the and the insect were from The in vitro was obtained from The and were obtained from acid proteinase K, and were obtained from and the were from MES, CHAPS, and were from was from and Endo were from membrane and protein were from the and were from was from were prepared as G. Methods Enzymol. PubMed Scopus Google Scholar). of VKOR with the at the N or C gene encoding VKOR was by using (16Li T. Chang C.Y. Jin D.Y. Lin P.J. Khvorova A. Stafford D.W. Nature. 2004; 427: 541-544Crossref PubMed Scopus (598) Google Scholar) as the and was into the and the was into the in vitro by to The tag with the acid of was obtained by the and by to the and the insert by the at the of the This was C-terminal tag with the acid of was obtained by the and by to In this of the to the insert results in the of the at the of the This is C-terminal tag was by first the insert and then the to to of VKOR with at the N was to the the acid at of the VKOR from The was used as the to C-terminal truncations of VKOR. comprising a and of the was used as for the a and to the C terminus of the predicted TM domains were used as for the of the predicted TM domains and the reporter tag were into the of VKOR with at the C C-terminal tag with the of was by using the as and a were into the and the was to by to VKOR truncations of the predicted TM were then by from the with the The encoding the of VKOR a the were the as were into to different VKOR truncations with the reporter tag at the C The acid residues and the VKOR truncations and the was obtained by the of and was by using the and a of the of VKOR at acid with a was by a a with of the of VKOR from and the used for VKOR The obtained were with to the of C-terminal tag with a of which the was to the VKOR of by to The used is the as and the of the C-terminal of VKOR and the of the tag with a The was to by the used for in vitro are in of the VKOR with HPC4 at the N or C in acid peptide of HPC4 epitope S. P.J. J. Biol. Chem. 1988; Full Text PDF PubMed Google Scholar) was to the N or C terminus of full-length VKOR. The HPC4-tagged VKOR was into the of the insect of The of the were by the at the of at In and in vitro the were at the to the C terminus of the VKOR or at the to the C terminus of the C-terminal VKOR was by and by to the The reaction was at for at to the of In vitro using and using were as J. S.M. Jin D. Stafford D.W. Blood. 2000; PubMed Google Scholar). The were at for in a of of of of a acid and of at of at and of with or of The were and with of and of Endo and of were to a of the and for at to the were with of with acid, and in of of HPC4-tagged VKOR in and of VKOR VKOR with a HPC4 tag at the N or C termini were in insect from cells expressing or were prepared as Stafford D.W. Jin D.Y. J. B. R. S.M. J. Biol. Chem. 2000; Full Text Full Text PDF PubMed Scopus Google Scholar). protein were to of proteinase K was to a of freshly prepared in the or of as the The were for and by to a of digestion, the microsomes were with and the were in and to and was to the using MES, and from the in vitro were to the proteins were to a membrane and was of VKOR and were to and to a membrane as were with by a and the protein were using the of the of VKOR by membrane topology of VKOR was predicted using different topology prediction programs with the The results are in I. can be of the programs predict three TM domains in VKOR, predicts and predicts four TM The programs that predict the location of the C terminus agree that the C terminus of VKOR is located in the cytoplasm. programs predict that the first TM residues The TM predictions are somewhat for the different prediction programs. The TM domains in VKOR predicted by the programs used are 10–29, 101–123, and been reported that the reliability of topology predictions is increased if different topology prediction the prediction (19Nilsson J. Persson B. von Heijne G. FEBS Lett. 2000; 486: 267-269Crossref PubMed Scopus (93) Google Scholar, 20Melen K. Krogh A. von Heijne G. J. Mol. Biol. 2003; 327: 735-744Crossref PubMed Scopus (174) Google Scholar). it is that VKOR three TM domains with the C terminus located in the cytoplasm and N terminus located in the ER shows the membrane topology prediction and the of VKOR by the most commonly used prediction program TMHMM. predicts that VKOR TM domains, 10–29, 101–123, and 127–149, with the of TM predicted by other programs by is in of membrane topology of VKOR by different membrane topology prediction programs topology of VKOR was predicted by different programs using the The four TM domains are the TM to the acid of location of in the are the of the predicted TM as the location of in the are the of the predicted TM as the in a VKOR with the von Heijne G. J. Biol. Chem. 1993; Full Text PDF PubMed Google Scholar) was used in this study to experimentally the membrane topology of VKOR. N-linked was at the N or C termini of the full-length VKOR or VKOR to the predicted membrane topology of VKOR, only acid residues the first TM Moreover, if the predicted TM at as predicted by and only C-terminal acid residues from the been reported that the is residues from the membrane von Heijne G. J. Biol. Chem. 1993; Full Text PDF PubMed Google Scholar, I. A. G. C. von Heijne G. J. Mol. Biol. PubMed Scopus Google Scholar). we acid to the from the The acid of the tag is and the C-terminal tag is The were to the N or C terminus of full-length VKOR and to of the VKOR truncations In we a C-terminal tag at the C terminus of full-length VKOR To the predicted TM of 75–97 or is the TM we the predicted TM of 75–97 in the of In we a tag the to the C terminus of VKOR the used are in In of the VKOR shows the of the in vitro of the full-length VKOR and full-length VKOR in the and of can be VKOR, is in the of RM, and the is sensitive to Endo which N-linked This that the N terminus of VKOR is located in the ER In C-terminal VKOR or is that the C terminus of VKOR is located in the cytoplasm. These results suggest that VKOR is a type III membrane protein with the of and it an number of TM This agrees with the topology model predicted by of the topology prediction programs in I. In of the potential in the VKOR and is this is with the prediction of which that is located the cytoplasm of the and is located the of the of the TM by in to the there are four potential TM segments in VKOR sequence. C-terminal truncations of VKOR were of the predicted potential TM The N-linked reporter tag was to the N terminus or the C terminus of these truncations shows the in vitro translation/cotranslocation results obtained from these VKOR truncations in the and of and the topological can be was was that the first predicted TM 10–29, can function as a of type III membrane protein. of and were that the N and C termini are located in the ER with the with the of this that the predicted TM can function as a sequence. However, of and which one more predicted TM are that have the topological as This the of the predicted TM of in the of as the signal sequence. TM of as a TM the predicted TM of as a TM we the potential TM of 75–97 in the of in the C-terminal tag can be in the of RM, that the C terminus is located in the of with the of this that the predicted TM of as a TM in the In a to which of the potential TM domains, 75–97 or 101–123, as the TM in the VKOR a reporter tag with the of was to the C terminus of This the the predicted TM to the von Heijne G. J. Biol. Chem. 1993; Full Text PDF PubMed Google Scholar, I. A. G. C. von Heijne G. J. Mol. Biol. PubMed Scopus Google Scholar), this tag be if as the in this the other if 75–97 as a TM this tag be there is the and the TM in was in the of This that the predicted TM is the TM in the VKOR 75–97 of the of N and C of VKOR in the topological of the N and C termini of VKOR in we full-length VKOR with a HPC4 epitope at and VKOR in insect cells of the VKOR, of the prepared microsomes from the cells expressing these were to proteinase K in the or of and to The protein were to a membrane and with the and by in proteinase K of the a protected by an that was sensitive to proteinase K by This the location of the HPC4 tag the N terminus of the VKOR. In was microsomes expressing were with proteinase K. This is with the location of the HPC4 tag the C terminus of the VKOR. These suggest that VKOR a topological of that is in with the in vitro translation/cotranslocation the membrane topology, i.e. of its TM segments and their relative to the is an first in the function of integral membrane (17Rost S. Fregin A. Ivaskevicius V. Conzelmann E. Hortnagel K. Pelz H.J. Lappegard K. Seifried E. Scharrer I. Tuddenham E.G. Muller C.R. Strom T.M. Oldenburg J. Nature. 2004; 427: 537-541Crossref PubMed Scopus (995) Google Scholar) that there are three TM domains in VKORC1, whereas (16Li T. Chang C.Y. Jin D.Y. Lin P.J. Khvorova A. Stafford D.W. Nature. 2004; 427: 541-544Crossref PubMed Scopus (598) Google Scholar) that there may TM domains the programs used for the membrane topology Recently, Biochem. Sci. 2004; Full Text Full Text PDF PubMed Scopus Google Scholar) that the TM there are four TM domains in VKOR. of the predictions were topology prediction in this study we to this to the most TM we used different computer programs to predict the membrane topology of VKOR. of programs agree that VKOR three TM domains with the N terminus located in the and C terminus located in the the programs that predict three TM domains agree their program predicts TM domains in VKOR, whereas another predicts We experiments to which of the four potential TM domains are To this we the in vitro von Heijne G. J. Biol. Chem. 1993; Full Text PDF PubMed Google Scholar, I. A. G. C. von Heijne G. J. Mol. Biol. PubMed Scopus Google Scholar). In the of canine rough full-length VKOR as well as VKOR truncations with are by the in vitro translation/cotranslocation These results suggest that VKOR is a type III membrane protein with its N terminus in the ER This was by the that an HPC4 tag at the N terminus of full-length VKOR in intact insect microsomes was protected from the membrane was first with In the with that the first predicted TM is the location of the N terminus of VKOR and the of the first TM is The of the predicted TM domains, residues 75–97 and 101–123, is the and C-terminal of and are in the in vitro canine rough This that the C termini of truncations are located in the of the ER the first predicted TM is the of these results is that the predicted TM is in these results there is to assume that it is a TM in However, to the prediction residues have a of as a TM residues 75–97 we tested in vitro as a TM by residues from In this the C-terminal tag is still in the of this that the predicted TM can function as an TM these results suggest that in vitro the and the predicted TM can function as TM is in and the intact the topology prediction programs that residues have a of a TM residues To this we a of that have a that the was If the predicted TM is a TM then this be because of the and TM (4Furie B. Bouchard B.A. Furie B.C. Blood. 1999; 93: 1798-1808Crossref PubMed Google Scholar, S.P. Israels E.D. Israels L.G. Apoptosis. 2001; 6: 57-68Crossref PubMed Scopus (29) Google Scholar). in was in the of RM, that the predicted TM is in This is with the that residues and 75–97 the TM in and in the intact protein. The of in vitro of full-length VKOR that the tag can be the C-terminal tag these it is still that the C terminus in the and that the to the tag at the C terminus of full-length VKOR is because of its to the membrane or to von Heijne G. PubMed Scopus Google Scholar). we a tag to the C terminus of full-length VKOR If the C terminus was then the tag the from the membrane to However, with this was in the of we that the C terminus of intact VKOR is the of the The predicted TM as a TM in intact VKOR. Further evidence that the C terminus of VKOR in the cytoplasm was by experiments an tag was to the C terminus of full-length VKOR microsomes were prepared from insect cells expressing and were with proteinase K. This tag was by proteinase whereas a tag was protected Altogether, our suggest that there are three TM domains 101–123, and in VKOR and that the N terminus of VKOR is located in the of the whereas the C terminus is located in the cytoplasm Our results agree with the membrane topology predicted by most of the computer prediction programs. The of VKOR, type III membrane is which its into the of are whereas residues are located in the and residues are in the ER it that the residues in the of Moreover, it that the is in the cytoplasm or the membrane is a that and hydrophobic regions and and of the four of VKOR. and Biochem. Sci. 2004; Full Text Full Text PDF PubMed Scopus Google Scholar) that these were of the of VKOR. However, in their four TM the and is in the of the whereas in our experimental model this is the of the vitamin K-dependent carboxylase in the of the it is that the of VKOR be in the cytoplasm or the membrane as vitamin K is and Biochem. Sci. 2004; Full Text Full Text PDF PubMed Scopus Google Scholar) that a was at of the of VKOR. The is the in a number of and protein S. S. K. Res. Rev. 2004; 24: PubMed Scopus Google Scholar). In with this there is the of in the of VKOR L.A. Hill A.P. Steventon G.B. Hutt A.J. Pallister C.J. Cowell D.C. J. Pharm. Pharmacol. 2001; 53: 481-486Crossref PubMed Scopus (18) Google Scholar, 12Silverman R.B. Nandi D.L. Biochem. Biophys. Res. Commun. 1988; 155: 1248-1254Crossref PubMed Scopus (65) Google Scholar, 14Thijssen H.H. Janssen Y.P. Vervoort L.T. Biochem. J. 1994; 297: 277-280Crossref PubMed Scopus (22) Google Scholar, R. Sane D.C. Hutson S.M. Thromb. Res. 2002; Full Text Full Text PDF PubMed Scopus Google Scholar, B.A. A. J. C. Biochem. J. 1992; PubMed Scopus Google Scholar, P.C. Res. Commun. 8: PubMed Scopus Google Scholar, R. J. Biochem. PubMed Scopus Google Scholar, P.C. Haemostasis. Google Scholar, J.J. Fasco M.J. Biochemistry. PubMed Scopus Google Scholar) and that the in a hydrophobic R. Sane D.C. Hutson S.M. Thromb. Res. 2002; Full Text Full Text PDF PubMed Scopus Google Scholar). Recently, Sane D.C. Hutson S.M. Wallin R. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar) reported that of or in to VKOR This is with the as the redox for VKOR. to our the is located the TM and is in the In our results suggest that VKOR is a type III membrane protein with three TM The of the is located the cytoplasm of the ER The topology model a of the of the acid residues of VKOR the ER which can be used as the for its function The that a is a TM may an for the in the of VKOR D. Hutson S.M. Wallin R. J. Biol. Chem. 1997; 272: 29068-29075Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). its may to of We for We for