Vol.
187,
No.
September
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
16, 1992
Pages
SITE-DIRECTED
MUTAGENESIS
OF AN
APO Ej(GLU,+LYS) TO LOW LbMing
COMMUNICATIONS
DENSITY
AND
APOLIPOPROTEIN ITS
LIPOPROTEIN
Dong’, Taku Yamamura’*,
1180-1186
E MUTANT,
BINDING RECEPTORS
Shoji Tajima’
and Akira
Yamamoto’
‘Department of Etiology and Pathophysiology, National Cardiovascular Center Research Institute, 5-7-l Fujishiro-dai, Suita, Osaka 565, Japan 21nstitute for Protein Research, Osaka University, Osaka, Japan Received
August
12,
1992
Apo ES(Glu,-+Lys) is a naturally occurring apolipoprotein E (apo E) mutant found in patients with hyperlipoproteinemia and atherosclerosis. It has been shown to have a high affinity for low density lipoprotein (LDL) receptors. In this study, mutant apo E.5 was produced by Chinese hamster ovary cells by means of an in vitro site-directed mutagenesis technique, and its LDL receptor binding activity was assessed. The apo E5 obtained from gene expression bound more readily to the LDL receptor than did plasma apo E3. The concentrations required for 50% competitive binding of 1251-labeledLDL to the LDL receptors were 58.9 rig/ml for plasma apo E3 and 25.7 rig/ml for the expressed apo E5. The expressed apo E5 displayed 229% normal binding. This result is highly consistent with that obtained with plasma apo E5, which showed 217% normal binding. Although the experimental apo E isoproteins contained more sialic acid than plasma apo E, the extent of sialylation had no effect on the receptor binding of apo E. 0 1992Academic Press, 1°C. SUMMARY.
Apolipoprotein
E (apo E) plays an important role in lipoprotein
metabolism through its
binding to receptors on the cell surface [I]. Apo E has three common isoforms: apo E2, apo E3 (the parent form), and apo E4 [2,3]. In terms of primary structure, apo E4 and apo E2 differ from apo E3 by one amino acid residue. Apo E3 has cysteine at residue 112 and arginine at residue 158, while apo E4 has arginine at residue 112 and apo E2 has cysteine at residue 158
[4,5]. In addition to the three major isoforms, isoelectric focusing of apo E has revealed several * To whom correspondence should be addressed. Abbreviations used are: Apo E, apolipoprotein E; CHO, Chinese hamster ovary; dhfr, dihydrofolate reductase; DMPC, dimyristoylphosphatidylcholine; D’TT, dithiothreitol; LDL, low density lipoprotein; SDS, sodium dodecyl sulfate; VLDL, very low density lipoprotein. 0006-291X/92
Copyright All rights
$4.00
0 1992 by Academic Press. Inc. of reproduction in any form reserved.
1180
Vol.
187,
No.
BIOCHEMICAL
2, 1992
minor isoforms,
AND
which result from posttranslational
BIOPHYSICAL
modification
RESEARCH
COMMUNICATIONS
of sialic acids [2,3].
The
significance of sialylated forms of apo E is not clear. It has been demonstrated that apo E4 has the same binding activity as apo E3 to LDL receptors, whereas that of apo E2 is low [6]. In some cases, the homozygosity
of apo E2 leads to type III hyperlipoproteinemia
several novel apo E mutants have been associated with lipoprotein abnormalities. have a low affinity
[7].
Also,
All of these
for LDL receptors in comparison with apo E3, and their differences
in
binding activity correspond to their structural differences [8-l 11. We previously identified two mutant apo E, apo E5 and apo E7, which are associated with hyperlipoproteinemia
and atherosclerosis
[12,13]. The mutation site of apo E5 is at residue 3,
where lysine replaces glutamic acid [14]. The asialo apo ES purified from plasma lipoproteins showed enhanced receptor binding activity [15]. In the present study, apo ES(Glu,+Lys)
was
produced by means of an in vitro site-directed mutagenesistechnique, and the binding activity of the expressedapo E5 to LDL receptorswas determined.
MATERIALS
AND METHODS
Site-directed mutagenesis of apo ES: The sequenceof the mutagenic oligonucleotide is as follows (the mismatchedbaseis underlined):5’-CCGC’ITGCnCACCTTGG-3’. For production of this mutant apo E, human apo E cDNA coding apo E3 (generously donated by Mitsubishi Kasei Kogyo, Tokyo, Japan) cloned in Ml3 (Toyobo, Osaka, Japan) was mutagenized in vitro by the site-directedmutagenesistechniquedescribedby Kunkel [16]. DNA sequencingverified the mutation. Expression
of apo E: Mutant Chinese hamster ovary (CHO) cells lacking dihydrofolate reductase(dhfr) [17], kindly provided by Dr. Chasin (Columbia University, New York), were used ashost cells. The plasmid pKCRH2PL (also donated by Mitsubishi Kasei Kogyo, Tokyo, Japan) was used to expressthe apo E5 [18]. The HindIII/EcoRI fragment coding apo E5 was ligated into the HindIII/EcoRI site of the expressingplasmid. 5x10’ cells of CHO-dhfr were transfected with 30 pg of the plasmid containing apo E5 cDNA and 2 pg of dhfr as described by Chen and Okayama [19]. After overnight culturing, the cells were split at 1:lO and subculturedin an a-medium (GIBCO, Grand Island, NY), supplementedwith 10%dialyzed fetal calf serum. Initial transformantswere grown in the samemedium, with gradedconcentrations (0.05 pM, 0.5 pM, and 5 pM) of methotrexate, an inhibitor of dhfr, to amplify the apo E5 gene. The synthesisof apo E was monitored by SDS-polyacrylamide gel electrophoretic analysisof the ?i-methionine-labeled proteins from the medium [20], as well as by isoelectric focusing. The apo E secreted by CHO cells into the medium was measured by enzyme-linked immunosorbent assay for selection of cell colonies expressing higher levels of apo E. The cDNA coding apo E3 was subjectedto the procedure describedabove. After stable cell lines had been established,the cells of apo E from medium: were cultured in a protein-free medium (S-clone, Sanko, Tokyo, Japan) at 37 ‘C for 48 hours. The medium was collected, dialyzed against 5 mM NH,HCO, and then delipidated. The delipidated proteins were fractionated by heparin affinity chromatography [12]. For further purification of apo E, the heparin-bound material was dissolved in 0.1 M Tris-HCl (pH 7.4) containing 6 M guanidine hydrochloride, 0.01% EDTA, and 1% 2-mercaptoethanol and chromatographedon a SuperdexTM200 gel permeationcolumn 1.6 cm x 60 cm (PharmaciaLKB
Purification
1181
Vol.
187,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Biotechnology, Uppsala, Sweden), equipped with a fast protein liquid chromatography apparatus (Pharmacia LKB Biotechnology, Uppsala, Sweden). The chromatography was performed at a flow rate of 20 ml per hour with 0.1 M Tris-HCl (pH 7.4) containing 4 M guanidine hydrochloride, 0.01% EDTA, and 0.1% 2-mercaptoethanol. Isolation of asialo apo E3 from serum: Serum was obtained from a subject with the E3/3 phenotype, and very low density lipoprotein (VLDL) was separated by ultracentrifugation. Asialo apo E3 was isolated by preparative isoelectric focusing on a solubilizable polyacrylamide gel, as described previously [Is]. Receptor binding assay: Apo E.dimyristoylphosphatidylcholine (DMPC, Sigma, St. Louis, MO) was prepared as described by Rall et al. [7]. Human ‘251-labeled LDL was prepared by the method of Fielding et al. [21]. Normal human fibroblasts were cultured and the competitive binding of apo E.DMPC against ‘251-labeled LDL was assayed at 4 ‘C on ice, as detailed by Innerarity et al. [22].
RESULTS The cDNA coding apo E5 was constructed, and its DNA sequencing verified the mutation (data not shown). pKCRH2PL,
The apo E5 cDNA
thus constructed was ligated to the expression vector,
and was over expressed in CHO cells lacking dhfr.
SDS-polyacrylamide
gel
electrophoretic identification of the synthesized apo E5 yielded one main band and several minor bands (Figure 1A). The main band appeared to represent intact apo E5. Its estimated molecular weight was slightly lower than that of apo E3 synthesized by THPl cells. The same observation
A
B
THPl E5 0
36-
apoE
THPl E5 ?
-
E-V E-N E-III E-II
-,
:I/,
Figure 1. Identification of apo ES expressed by CHO cells. The CHO cells transfected with cloned apo E5 cDNA were incubated in a medium containing 35S-methionineat 37 “C for 5 hours. THPI cells were cultured in the same manner. The medium was collected and the secreted apo E was immunoprecipitated with mono-specific anti-apo E antibodies. The precipitates were analyzed by SDS-polyacrylamide gel electrophoresis (A) and isoelectric focusing (B). The bands below the intact apo E on SDSelectrophoresis are thought to be degradation products of apo E. The presence of multiple bands on isoelectric focusing indicates that the apo E5 isoproteins synthesized by CHO cells contain more sialic acid residues than does plasma apo E. The position of asialo apo E5 in the focusing gel is indicated by “E-V”. 1182
Vol.
187,
No.
2, 1992
BIOCHEMICAL
*
1
AND
BIOPHYSICAL
B
2 0
Mr x 10m3 97.6-
--I
66.2’
’ ”
1 w
RESEARCH
2 -
3 -
-
Figure 2. SDS-polyacrylamide gel electrophoresis and isoelectric focusing of purified apo ES, with a polyacrylamide gel (B).
COMMUNICATIONS
E-V E-H
of apo E5 purified from or without neuraminidase
the medium treatment,
(A)
on
The molecular markers were identified, and purified apo E5 appeared as a single band under elecuophoresis on an SDS-polyacrylamide gel (A). The following isoelectric focusing patterns are shown: (1) plasma apo VLDL of an E3/3 phenotype, (2) purified apo E5, and (3) purified apo E5 treated with neuraminidase. The enzyme treatment was carried out as described previously [ 131.
was made in the caseof plasmaapo E3 and apo E5 [ 131. The minor bandsbelow the intact apo E may have beendegradationproducts of apo E, sincethey reacted with mono-specific anti-apo E antibodies. The isoelectric focusing pattern of the expressedapo E is shown in Figure 1B. Below the position of asialo apo E5 were multiple bands,which were thought to be sialylated forms of apo E5. Expressedapo E purified from the medium showed a single band on SDSpolyacrylamide gel &CtrOphOTeSiS
(Figure 2A). For verification that the multiple bandsbelow
the asialo apo E5 band shown by isoelectric focusing were sialylated forms of apo E5, the purified apo E5 was treated with neuraminidase. As shown in Figure 2B, the bandsmigrated from the basic to the acidic side of the gel after enzyme treatment, indicating that a component of the attachedsialic acid had beenremoved. Comparedwith plasmaapo E, the apo E produced by CHO cells contained more sialic acid, and this provided an opportunity to compare the receptor binding activities of the two types of apo E. Apo E was recombined with DMPC vesicles and the receptor binding activities were estimatedby measuringthe competition of apo E.DMPC againstthe binding of 12’I-labeledLDL to LDL receptors. Figure 3 showstheseresults. Expressedapo E5 had a higher binding activity than plasma apo E3. The concentration of apo E at which 50% of “‘I-labeled LDL was displaced was 58.9 rig/ml for plasma apo E3 and 25.7 rig/ml for the expressedapo E5. The expressedapo E5 binding was 229% that of the normal apo E3. Finally, we compared the binding activities of apo E3 produced by gene expression and asialo apo E3 purified from plasma. They showedthe identical binding affinity with whole apo E isoproteinsobtainedfrom 1183
Vol.
187,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
80
60
l
E5 I 0.05
I 0 Apo
E.DMPC
complexes
Figure 3. Comparison of binding to human fibroblast (0) and apo ES obtained from gene expression (0).
I 0.10 @g/ml)
LDL
receptors
by plasma
apo
E3
The fibroblasts were incubated at 4 “C for 2 h in a medium containing 2 Kg/ml of “‘Ilabeled LDL and variousconcentrations of apoE.DMPC complexes,asindicated. Eachpoint represents the averageof duplicatedishes.The insetshowsa logit-logplot of the bindingdata
that wasusedto determinethe 50%displacement of ‘251-labeled LDL. The concentrationof apo E at which 50%of ‘251-labeled LDL wasdisplacedwas58.9rig/mlfor plasmaapoE3 and25.7 rig/ml for expressed apoE5.
a subject with the E3/3 phenotype (data not shown). These results indicate that the apo E produced in our expression system does not differ functionally from plasma apo E. Also, sialylation of apo E had no effect on receptor binding. DISCUSSION We discoveredapo E.5in patientswith hyperlipoproteinemiaand atherosclerosis[ 12,131and demonstratedthat asialo apo E5 purified from plasma binds to LDL receptors with twice the affinity of apo E3 [ 151. Recently, Wardell et al. [23] also reported that an approximately equal mixture of apo E5 and normal apo E3 exhibited 188% of normal binding. In the present experiment, the cDNA coding apo E5 was prepared by site-directed mutagenesis,and
a fairly
large amount of apo E5 was obtained. The expressedapo E5 showedenhancedbinding to LDL receptors relative to that of plasma apo E3, which is in good agreement with the behavior displayed by asialo apo E5, which showed217% of normal binding [15]. These data support our hypothesis that enhanced metabolism of apo ES-containing lipoproteins results in hypercholesterolemiathrough down-regulationof LDL receptorsand acceleratedconversionfrom VLDL to LDL [15]. There are two types of polymorphism found in plasmaapo E. One resultsfrom amino acid substitution, and the other from various degreesof sialylation [2,3]. CHO cells are known to causesialylation of secretory and transmembraneproteins [24,25]. Isoelectric focusing showed 1184
Vol.
187,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
that apo E isoproteins synthesized by CHO cells contained more sialic acid than plasma apo E. In the present study, the apo E3 obtained from gene expression showed no difference in receptor binding from asialo apo E3 and whole apo E obtained from a subject with the E3/3 phenotype. The binding activity of highly sialylated apo E5 also was similar to that of plasma asialo apo E5. These results indicate that sialylation has no effect on apo E binding to the LDL receptor. The mutation site of apo E5 was not located in its receptor binding domain which has been demonstrated to
be
functionally important.
The finding that plasma apo E5 exhibited the higher
affinity of binding indicates that a structural change even at a position other than the receptor binding region can also affect receptor binding. It would be interesting and useful to use the in vitro site-directed mutagenesis technique to produce an apo E mutant with a mutation site near this region or across the whole apo E molecule, and to investigate the relationship between the structure and function of apo E. We prepared the cDNA coding mutant apo E5 by site-directed mutagenesis and obtained a fairly large quantity of apo E5 from gene expression.
Although the apo E5 expressed by CHO
cells contained more sialic acid than plasma apo E5, it displayed the same binding activity. Similar results were obtained in the case of apo E3. These data confirm that expressed apo E functions similarly to plasma apo E. Combined with naturally occurring mutant apo E, apo E mutants produced by gene expression would be useful in further investigating the significance of the apo E molecular structure in lipoprotein metabolism. ACKNOWLEDGMENTS We wish to thank Dr. Y. Miyake, National Cardiovascular Center Research Institute, for her many helpful suggestions. This work was supported in part by a grant from The HMG-CoA Reductase Research Fund, and by a Grant-in-Aid from The Research Foundation for Cancer and Cardiovascular Disease, Japan. REFERENCES 1. 2. 3.
4. 5. 6. 7. 8. 9.
Mahley, R.W., and Innerarity, T.L. (1983) B&him. Biophys. Acta 737, 192-222. Zannis, V.I., and Breslow, J.L. (1981) Biochemistry 20, 1033-1041. Zannis, V.I., and Breslow, J.L., Utermann, G., Mahley, R.W., Weisgraber, K.H., Havel, R.J., Goldstein, J.L., Brown, M.S., Schonfeld, G., Hazzard, W.R., and Blum, C. (1982) J. Lipid Res. 23, 911-914. Weisgraber, K.H., Rail, S.C., Jr., and Mahley, R.W. (1981) J. Biol. Chem. 256,9077-9083. Rall, S.C., Jr., Weisgraber, K.H., and Mahley, R.W. (1982) J. Biol. Chem. 257,4171-4178. Weisgraber, K.H., Innerarity, T.L., and Mahley, R.W. (1982) J. Biol. Chem. 257, 25182521. Utermann, G., Hees, M., and Steinmetz, A. (1977) Nature 269, 604-607. Rall, S.C., Jr., Weisgraber, K.H., Innerarity, T.L., and Mahley, R.W. (1982) Proc. Natl. Acad. Sci. USA 79, 4696-4700. Rall, S.C., Jr., Weisgraber, K.H., Innerarity, T.L., Bersot, T.P., and Mahley, R.W. (1983) J. Clin. Invest. 72, 1288-1297. 1185
Vol.
10. 11. 12. 13. 14. 15. 16. 17. 18.
19. 20. 21. 22. 23. 24. 25.
187,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Wardell, M.R., Brennan, S.O., Janus, E.D., Fraser, R., and Carrell, R.W. (1987) J. Clin. Invest. 80, 483-490. Mahley, R.W. (1988) Science 240, 622-630. Yamamura, T., Yamamoto, A., Hiramori, K., and Nambu, S. (1984) Atherosclerosis 50, 159-172. Yamamura, T., Yamamoto, A., Sumiyoshi, T., Hiramori, K., Nishioeda, Y., and Nambu, S. (1984) J. Clin. Invest. 74, 1229-1237. Tajima, S., Yamamura, T., and Yamamoto, A. (1988) J. Biochem. 104, 48-52. Dong, L.-M., Yamamura, T., and Yamamoto, A. (1990) Biochem. Biophys. Res. Commun. 168, 409-414. Kunkel, T. (1985) Proc. Natl. Acad. Sci. USA 82, 488-492. Urlaub, G., and Chasin, L.A. (1980) Proc. Natl. Acad. Sci. USA 77, 4216-4220. Mishina, M., Kurosaki, T., Tobimatsu, T., Morimoto, Y., Noda, M., Yamamoto, T., Terao, M., Lindstrom, J., Takahashi, T., Kuno, M., and Numa, S. (1984) Nature 307,604608. Chen, C., and Okayama, H. (1987) Mol. Cell. Biol. 7, 2745-2752. Tajima, S., Yamamura, T., Menju, M., and Yamamoto, A. (1989) J. Biochem. 10.5, 249253. Fielding, C.J., Vlodavsky, I., Fielding, P.E., and Gospodarowicz, D. (1979) J. Biol. Chem. 254, 8861-8868. Innerarity, T.L., Pitas, R.E., and Mahley, R.W. (1979) J. Biol. Chem. 254, 4186-4190. Wardell, M.R., Rail, S.C., Jr., Schaefer, E.J., Kane, J.P., and Weisgraber, K.H. (1991) J. Lipid Res. 32, 521-528. Scahill, S.J., Devos, R., Heyden, J.V., and Fiers, W. (1983) Proc. Natl. Acad. Sci. USA 80, 4654-4658. McCormick, F., Trahey, M., Innis, M., Dieckmann, B., and Ringold, G. (1984) Mol. Cell. Biol. 4, 166-172.
1186
187,
No.
September
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
16, 1992
Pages
SITE-DIRECTED
MUTAGENESIS
OF AN
APO Ej(GLU,+LYS) TO LOW LbMing
COMMUNICATIONS
DENSITY
AND
APOLIPOPROTEIN ITS
LIPOPROTEIN
Dong’, Taku Yamamura’*,
1180-1186
E MUTANT,
BINDING RECEPTORS
Shoji Tajima’
and Akira
Yamamoto’
‘Department of Etiology and Pathophysiology, National Cardiovascular Center Research Institute, 5-7-l Fujishiro-dai, Suita, Osaka 565, Japan 21nstitute for Protein Research, Osaka University, Osaka, Japan Received
August
12,
1992
Apo ES(Glu,-+Lys) is a naturally occurring apolipoprotein E (apo E) mutant found in patients with hyperlipoproteinemia and atherosclerosis. It has been shown to have a high affinity for low density lipoprotein (LDL) receptors. In this study, mutant apo E.5 was produced by Chinese hamster ovary cells by means of an in vitro site-directed mutagenesis technique, and its LDL receptor binding activity was assessed. The apo E5 obtained from gene expression bound more readily to the LDL receptor than did plasma apo E3. The concentrations required for 50% competitive binding of 1251-labeledLDL to the LDL receptors were 58.9 rig/ml for plasma apo E3 and 25.7 rig/ml for the expressed apo E5. The expressed apo E5 displayed 229% normal binding. This result is highly consistent with that obtained with plasma apo E5, which showed 217% normal binding. Although the experimental apo E isoproteins contained more sialic acid than plasma apo E, the extent of sialylation had no effect on the receptor binding of apo E. 0 1992Academic Press, 1°C. SUMMARY.
Apolipoprotein
E (apo E) plays an important role in lipoprotein
metabolism through its
binding to receptors on the cell surface [I]. Apo E has three common isoforms: apo E2, apo E3 (the parent form), and apo E4 [2,3]. In terms of primary structure, apo E4 and apo E2 differ from apo E3 by one amino acid residue. Apo E3 has cysteine at residue 112 and arginine at residue 158, while apo E4 has arginine at residue 112 and apo E2 has cysteine at residue 158
[4,5]. In addition to the three major isoforms, isoelectric focusing of apo E has revealed several * To whom correspondence should be addressed. Abbreviations used are: Apo E, apolipoprotein E; CHO, Chinese hamster ovary; dhfr, dihydrofolate reductase; DMPC, dimyristoylphosphatidylcholine; D’TT, dithiothreitol; LDL, low density lipoprotein; SDS, sodium dodecyl sulfate; VLDL, very low density lipoprotein. 0006-291X/92
Copyright All rights
$4.00
0 1992 by Academic Press. Inc. of reproduction in any form reserved.
1180
Vol.
187,
No.
BIOCHEMICAL
2, 1992
minor isoforms,
AND
which result from posttranslational
BIOPHYSICAL
modification
RESEARCH
COMMUNICATIONS
of sialic acids [2,3].
The
significance of sialylated forms of apo E is not clear. It has been demonstrated that apo E4 has the same binding activity as apo E3 to LDL receptors, whereas that of apo E2 is low [6]. In some cases, the homozygosity
of apo E2 leads to type III hyperlipoproteinemia
several novel apo E mutants have been associated with lipoprotein abnormalities. have a low affinity
[7].
Also,
All of these
for LDL receptors in comparison with apo E3, and their differences
in
binding activity correspond to their structural differences [8-l 11. We previously identified two mutant apo E, apo E5 and apo E7, which are associated with hyperlipoproteinemia
and atherosclerosis
[12,13]. The mutation site of apo E5 is at residue 3,
where lysine replaces glutamic acid [14]. The asialo apo ES purified from plasma lipoproteins showed enhanced receptor binding activity [15]. In the present study, apo ES(Glu,+Lys)
was
produced by means of an in vitro site-directed mutagenesistechnique, and the binding activity of the expressedapo E5 to LDL receptorswas determined.
MATERIALS
AND METHODS
Site-directed mutagenesis of apo ES: The sequenceof the mutagenic oligonucleotide is as follows (the mismatchedbaseis underlined):5’-CCGC’ITGCnCACCTTGG-3’. For production of this mutant apo E, human apo E cDNA coding apo E3 (generously donated by Mitsubishi Kasei Kogyo, Tokyo, Japan) cloned in Ml3 (Toyobo, Osaka, Japan) was mutagenized in vitro by the site-directedmutagenesistechniquedescribedby Kunkel [16]. DNA sequencingverified the mutation. Expression
of apo E: Mutant Chinese hamster ovary (CHO) cells lacking dihydrofolate reductase(dhfr) [17], kindly provided by Dr. Chasin (Columbia University, New York), were used ashost cells. The plasmid pKCRH2PL (also donated by Mitsubishi Kasei Kogyo, Tokyo, Japan) was used to expressthe apo E5 [18]. The HindIII/EcoRI fragment coding apo E5 was ligated into the HindIII/EcoRI site of the expressingplasmid. 5x10’ cells of CHO-dhfr were transfected with 30 pg of the plasmid containing apo E5 cDNA and 2 pg of dhfr as described by Chen and Okayama [19]. After overnight culturing, the cells were split at 1:lO and subculturedin an a-medium (GIBCO, Grand Island, NY), supplementedwith 10%dialyzed fetal calf serum. Initial transformantswere grown in the samemedium, with gradedconcentrations (0.05 pM, 0.5 pM, and 5 pM) of methotrexate, an inhibitor of dhfr, to amplify the apo E5 gene. The synthesisof apo E was monitored by SDS-polyacrylamide gel electrophoretic analysisof the ?i-methionine-labeled proteins from the medium [20], as well as by isoelectric focusing. The apo E secreted by CHO cells into the medium was measured by enzyme-linked immunosorbent assay for selection of cell colonies expressing higher levels of apo E. The cDNA coding apo E3 was subjectedto the procedure describedabove. After stable cell lines had been established,the cells of apo E from medium: were cultured in a protein-free medium (S-clone, Sanko, Tokyo, Japan) at 37 ‘C for 48 hours. The medium was collected, dialyzed against 5 mM NH,HCO, and then delipidated. The delipidated proteins were fractionated by heparin affinity chromatography [12]. For further purification of apo E, the heparin-bound material was dissolved in 0.1 M Tris-HCl (pH 7.4) containing 6 M guanidine hydrochloride, 0.01% EDTA, and 1% 2-mercaptoethanol and chromatographedon a SuperdexTM200 gel permeationcolumn 1.6 cm x 60 cm (PharmaciaLKB
Purification
1181
Vol.
187,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Biotechnology, Uppsala, Sweden), equipped with a fast protein liquid chromatography apparatus (Pharmacia LKB Biotechnology, Uppsala, Sweden). The chromatography was performed at a flow rate of 20 ml per hour with 0.1 M Tris-HCl (pH 7.4) containing 4 M guanidine hydrochloride, 0.01% EDTA, and 0.1% 2-mercaptoethanol. Isolation of asialo apo E3 from serum: Serum was obtained from a subject with the E3/3 phenotype, and very low density lipoprotein (VLDL) was separated by ultracentrifugation. Asialo apo E3 was isolated by preparative isoelectric focusing on a solubilizable polyacrylamide gel, as described previously [Is]. Receptor binding assay: Apo E.dimyristoylphosphatidylcholine (DMPC, Sigma, St. Louis, MO) was prepared as described by Rall et al. [7]. Human ‘251-labeled LDL was prepared by the method of Fielding et al. [21]. Normal human fibroblasts were cultured and the competitive binding of apo E.DMPC against ‘251-labeled LDL was assayed at 4 ‘C on ice, as detailed by Innerarity et al. [22].
RESULTS The cDNA coding apo E5 was constructed, and its DNA sequencing verified the mutation (data not shown). pKCRH2PL,
The apo E5 cDNA
thus constructed was ligated to the expression vector,
and was over expressed in CHO cells lacking dhfr.
SDS-polyacrylamide
gel
electrophoretic identification of the synthesized apo E5 yielded one main band and several minor bands (Figure 1A). The main band appeared to represent intact apo E5. Its estimated molecular weight was slightly lower than that of apo E3 synthesized by THPl cells. The same observation
A
B
THPl E5 0
36-
apoE
THPl E5 ?
-
E-V E-N E-III E-II
-,
:I/,
Figure 1. Identification of apo ES expressed by CHO cells. The CHO cells transfected with cloned apo E5 cDNA were incubated in a medium containing 35S-methionineat 37 “C for 5 hours. THPI cells were cultured in the same manner. The medium was collected and the secreted apo E was immunoprecipitated with mono-specific anti-apo E antibodies. The precipitates were analyzed by SDS-polyacrylamide gel electrophoresis (A) and isoelectric focusing (B). The bands below the intact apo E on SDSelectrophoresis are thought to be degradation products of apo E. The presence of multiple bands on isoelectric focusing indicates that the apo E5 isoproteins synthesized by CHO cells contain more sialic acid residues than does plasma apo E. The position of asialo apo E5 in the focusing gel is indicated by “E-V”. 1182
Vol.
187,
No.
2, 1992
BIOCHEMICAL
*
1
AND
BIOPHYSICAL
B
2 0
Mr x 10m3 97.6-
--I
66.2’
’ ”
1 w
RESEARCH
2 -
3 -
-
Figure 2. SDS-polyacrylamide gel electrophoresis and isoelectric focusing of purified apo ES, with a polyacrylamide gel (B).
COMMUNICATIONS
E-V E-H
of apo E5 purified from or without neuraminidase
the medium treatment,
(A)
on
The molecular markers were identified, and purified apo E5 appeared as a single band under elecuophoresis on an SDS-polyacrylamide gel (A). The following isoelectric focusing patterns are shown: (1) plasma apo VLDL of an E3/3 phenotype, (2) purified apo E5, and (3) purified apo E5 treated with neuraminidase. The enzyme treatment was carried out as described previously [ 131.
was made in the caseof plasmaapo E3 and apo E5 [ 131. The minor bandsbelow the intact apo E may have beendegradationproducts of apo E, sincethey reacted with mono-specific anti-apo E antibodies. The isoelectric focusing pattern of the expressedapo E is shown in Figure 1B. Below the position of asialo apo E5 were multiple bands,which were thought to be sialylated forms of apo E5. Expressedapo E purified from the medium showed a single band on SDSpolyacrylamide gel &CtrOphOTeSiS
(Figure 2A). For verification that the multiple bandsbelow
the asialo apo E5 band shown by isoelectric focusing were sialylated forms of apo E5, the purified apo E5 was treated with neuraminidase. As shown in Figure 2B, the bandsmigrated from the basic to the acidic side of the gel after enzyme treatment, indicating that a component of the attachedsialic acid had beenremoved. Comparedwith plasmaapo E, the apo E produced by CHO cells contained more sialic acid, and this provided an opportunity to compare the receptor binding activities of the two types of apo E. Apo E was recombined with DMPC vesicles and the receptor binding activities were estimatedby measuringthe competition of apo E.DMPC againstthe binding of 12’I-labeledLDL to LDL receptors. Figure 3 showstheseresults. Expressedapo E5 had a higher binding activity than plasma apo E3. The concentration of apo E at which 50% of “‘I-labeled LDL was displaced was 58.9 rig/ml for plasma apo E3 and 25.7 rig/ml for the expressedapo E5. The expressedapo E5 binding was 229% that of the normal apo E3. Finally, we compared the binding activities of apo E3 produced by gene expression and asialo apo E3 purified from plasma. They showedthe identical binding affinity with whole apo E isoproteinsobtainedfrom 1183
Vol.
187,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
80
60
l
E5 I 0.05
I 0 Apo
E.DMPC
complexes
Figure 3. Comparison of binding to human fibroblast (0) and apo ES obtained from gene expression (0).
I 0.10 @g/ml)
LDL
receptors
by plasma
apo
E3
The fibroblasts were incubated at 4 “C for 2 h in a medium containing 2 Kg/ml of “‘Ilabeled LDL and variousconcentrations of apoE.DMPC complexes,asindicated. Eachpoint represents the averageof duplicatedishes.The insetshowsa logit-logplot of the bindingdata
that wasusedto determinethe 50%displacement of ‘251-labeled LDL. The concentrationof apo E at which 50%of ‘251-labeled LDL wasdisplacedwas58.9rig/mlfor plasmaapoE3 and25.7 rig/ml for expressed apoE5.
a subject with the E3/3 phenotype (data not shown). These results indicate that the apo E produced in our expression system does not differ functionally from plasma apo E. Also, sialylation of apo E had no effect on receptor binding. DISCUSSION We discoveredapo E.5in patientswith hyperlipoproteinemiaand atherosclerosis[ 12,131and demonstratedthat asialo apo E5 purified from plasma binds to LDL receptors with twice the affinity of apo E3 [ 151. Recently, Wardell et al. [23] also reported that an approximately equal mixture of apo E5 and normal apo E3 exhibited 188% of normal binding. In the present experiment, the cDNA coding apo E5 was prepared by site-directed mutagenesis,and
a fairly
large amount of apo E5 was obtained. The expressedapo E5 showedenhancedbinding to LDL receptors relative to that of plasma apo E3, which is in good agreement with the behavior displayed by asialo apo E5, which showed217% of normal binding [15]. These data support our hypothesis that enhanced metabolism of apo ES-containing lipoproteins results in hypercholesterolemiathrough down-regulationof LDL receptorsand acceleratedconversionfrom VLDL to LDL [15]. There are two types of polymorphism found in plasmaapo E. One resultsfrom amino acid substitution, and the other from various degreesof sialylation [2,3]. CHO cells are known to causesialylation of secretory and transmembraneproteins [24,25]. Isoelectric focusing showed 1184
Vol.
187,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
that apo E isoproteins synthesized by CHO cells contained more sialic acid than plasma apo E. In the present study, the apo E3 obtained from gene expression showed no difference in receptor binding from asialo apo E3 and whole apo E obtained from a subject with the E3/3 phenotype. The binding activity of highly sialylated apo E5 also was similar to that of plasma asialo apo E5. These results indicate that sialylation has no effect on apo E binding to the LDL receptor. The mutation site of apo E5 was not located in its receptor binding domain which has been demonstrated to
be
functionally important.
The finding that plasma apo E5 exhibited the higher
affinity of binding indicates that a structural change even at a position other than the receptor binding region can also affect receptor binding. It would be interesting and useful to use the in vitro site-directed mutagenesis technique to produce an apo E mutant with a mutation site near this region or across the whole apo E molecule, and to investigate the relationship between the structure and function of apo E. We prepared the cDNA coding mutant apo E5 by site-directed mutagenesis and obtained a fairly large quantity of apo E5 from gene expression.
Although the apo E5 expressed by CHO
cells contained more sialic acid than plasma apo E5, it displayed the same binding activity. Similar results were obtained in the case of apo E3. These data confirm that expressed apo E functions similarly to plasma apo E. Combined with naturally occurring mutant apo E, apo E mutants produced by gene expression would be useful in further investigating the significance of the apo E molecular structure in lipoprotein metabolism. ACKNOWLEDGMENTS We wish to thank Dr. Y. Miyake, National Cardiovascular Center Research Institute, for her many helpful suggestions. This work was supported in part by a grant from The HMG-CoA Reductase Research Fund, and by a Grant-in-Aid from The Research Foundation for Cancer and Cardiovascular Disease, Japan. REFERENCES 1. 2. 3.
4. 5. 6. 7. 8. 9.
Mahley, R.W., and Innerarity, T.L. (1983) B&him. Biophys. Acta 737, 192-222. Zannis, V.I., and Breslow, J.L. (1981) Biochemistry 20, 1033-1041. Zannis, V.I., and Breslow, J.L., Utermann, G., Mahley, R.W., Weisgraber, K.H., Havel, R.J., Goldstein, J.L., Brown, M.S., Schonfeld, G., Hazzard, W.R., and Blum, C. (1982) J. Lipid Res. 23, 911-914. Weisgraber, K.H., Rail, S.C., Jr., and Mahley, R.W. (1981) J. Biol. Chem. 256,9077-9083. Rall, S.C., Jr., Weisgraber, K.H., and Mahley, R.W. (1982) J. Biol. Chem. 257,4171-4178. Weisgraber, K.H., Innerarity, T.L., and Mahley, R.W. (1982) J. Biol. Chem. 257, 25182521. Utermann, G., Hees, M., and Steinmetz, A. (1977) Nature 269, 604-607. Rall, S.C., Jr., Weisgraber, K.H., Innerarity, T.L., and Mahley, R.W. (1982) Proc. Natl. Acad. Sci. USA 79, 4696-4700. Rall, S.C., Jr., Weisgraber, K.H., Innerarity, T.L., Bersot, T.P., and Mahley, R.W. (1983) J. Clin. Invest. 72, 1288-1297. 1185
Vol.
10. 11. 12. 13. 14. 15. 16. 17. 18.
19. 20. 21. 22. 23. 24. 25.
187,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Wardell, M.R., Brennan, S.O., Janus, E.D., Fraser, R., and Carrell, R.W. (1987) J. Clin. Invest. 80, 483-490. Mahley, R.W. (1988) Science 240, 622-630. Yamamura, T., Yamamoto, A., Hiramori, K., and Nambu, S. (1984) Atherosclerosis 50, 159-172. Yamamura, T., Yamamoto, A., Sumiyoshi, T., Hiramori, K., Nishioeda, Y., and Nambu, S. (1984) J. Clin. Invest. 74, 1229-1237. Tajima, S., Yamamura, T., and Yamamoto, A. (1988) J. Biochem. 104, 48-52. Dong, L.-M., Yamamura, T., and Yamamoto, A. (1990) Biochem. Biophys. Res. Commun. 168, 409-414. Kunkel, T. (1985) Proc. Natl. Acad. Sci. USA 82, 488-492. Urlaub, G., and Chasin, L.A. (1980) Proc. Natl. Acad. Sci. USA 77, 4216-4220. Mishina, M., Kurosaki, T., Tobimatsu, T., Morimoto, Y., Noda, M., Yamamoto, T., Terao, M., Lindstrom, J., Takahashi, T., Kuno, M., and Numa, S. (1984) Nature 307,604608. Chen, C., and Okayama, H. (1987) Mol. Cell. Biol. 7, 2745-2752. Tajima, S., Yamamura, T., Menju, M., and Yamamoto, A. (1989) J. Biochem. 10.5, 249253. Fielding, C.J., Vlodavsky, I., Fielding, P.E., and Gospodarowicz, D. (1979) J. Biol. Chem. 254, 8861-8868. Innerarity, T.L., Pitas, R.E., and Mahley, R.W. (1979) J. Biol. Chem. 254, 4186-4190. Wardell, M.R., Rail, S.C., Jr., Schaefer, E.J., Kane, J.P., and Weisgraber, K.H. (1991) J. Lipid Res. 32, 521-528. Scahill, S.J., Devos, R., Heyden, J.V., and Fiers, W. (1983) Proc. Natl. Acad. Sci. USA 80, 4654-4658. McCormick, F., Trahey, M., Innis, M., Dieckmann, B., and Ringold, G. (1984) Mol. Cell. Biol. 4, 166-172.
1186
