- Email: [email protected]
Human apo A-I in transgenic mice is more efficient in activating lecithin: cholesterol acyltransferase than mouse apo A-I
BB,
ELSEVIER
Biochimica et BiophysicaActa 1254 (1995) 217-220
etBiochi]ic~a BiophysicaA~ta
Rapid Report Human apo A-I in transgenic mice is more efficient in activating lecithin:cholesterol acyltransferase than mouse apo A-I Elina Golder-Novoselsky, Alex V. Nichols, Edward M. Rubin, Trudy M. Forte * Life Sciences Division, Lawrence Berkeley Laboratory, University of California, 1 Cyclotron Road, Berkeley, CA 94720, USA
Received 26 October 1994; accepted2 November1994
Abstract
This study shows that, in control and transgenic mice, there is a parallel increase in LCAT activity and plasma apo A-I concentrations during postnatal development. We also demonstrate that human apo A-I is a much more efficient activator (1.6-fold) of mouse LCAT activity than mouse apo A-I. We propose that the differences in amino acid sequence between human and mouse apo A-I may account for the higher LCAT activity with human apo A-I. Keywords: Transgenicmouse; LCAT activity;Apo I; (Mouse); (Human)
High-density lipoprotein (HDL) plays a central role in extracellular lipid transport and participates in the process of reverse cholesterol transport [1] by promoting cholesterol efflux from peripheral cells [2]. In this process, unesterified cholesterol is removed from peripheral cells and is converted to cholesteryl esters by the enzyme lecithin:cholesterol acyltransferase (LCAT) and returned to the liver for catabolism. LCAT is activated by apo A-I on the surface of HDL particles. In the adult C57B1/6 mouse, the HDL particle distribution consists of a single population of particles with a diameter of approx. 9.8 nm [3,4]. However, transgenic mice expressing the human apo A-I gene exhibit a bimodal distribution with major peaks at 9.1 and 10.9 nm [3,5]. The latter distribution approximates the bimodal apo A-I without apo A-II HDL profile observed in human plasma. The human apo A-I transgenic mice have HDL containing predominantly human apo A-I and trace amounts of mouse apo A-I, and, unlike control mice, are resistant to diet-induced atherosclerosis [3]. We have recently shown that HDL distribution during late prenatal and early postnatal development in both control and transgenic mice was consistently different from that of adult mice and was skewed to smaller sized particles [6]. 2 days before delivery, a 7.5 nm particle was
* Corresponding author. Fax: + 1 (510) 486 4750. 0005-2760/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0005-2760(94)00230-4
the predominant species in both control and transgenic mice. In control mice 3 days after delivery, a monodisperse peak, consisting of 9.0 nm particles, was noted which gradually increased in size to 9.8 nm, the adult size, by day 13.3 days after delivery, HDL of transgenic mice exhibited a bimodal distribution in which the particles were smaller, 7.5 and 8.9 nm, than those of adult transgenic mice. With increasing time after delivery, the bimodal pattern showed larger particles and developed into the adult pattern by 17 days after birth. The stepwise increment in size of control and transgenic HDL particles was associated with increased plasma concentrations of mouse and human plasma apo A-I. We previously suggested that the predominance of smaller particles during the neonatal period might be indicative of reduced LCAT activity [6]. In the present studies, we examined whether differences in LCAT activity during development may contribute to the observed differences in HDL profiles between control and human apo A-I transgenic mice. Since human apo A-I is the predominant apolipoprotein on HDL in transgenic mice, we investigated the effectiveness of human vs. mouse apo A-I in activating endogenous mouse LCAT. LCAT activity was determined by the substrate independent proteoliposome assay in which the rate of cholesteryl ester formation was measured. This assay utilized proteoliposomes prepared from purified human a n d / o r mouse apo A-I, [14C]cholesterol and lecithin according to method
E. Golder-Novoselsky et al. /Biochimica et Biophysica Acta 1254 (1995) 217-220
218
of Chen and Albers [7]. Both human and mouse apo A-I were incorporated into liposomes in equivalent amounts and proteoliposomes containing 0.22 m g / m l were use in the present study. Electron microscopy was used to evaluate the structural properties of human and mouse apo A-I proteoliposomes in order to establish that there were no major physical differences between the two types of proteoliposomes. In both cases the proteoliposomes consisted of small vesicles which were similar to one another in size where human apo A-I proteoliposomes were 34.6 _ 11 nm diameter and mouse apo A-I proteoliposomes, 35.3 + 13 nm. Plasma from transgenic and control mice was used as the source of LCAT during the following developmental time points: 3, 8, 13, 17 and 30 days after birth. LCAT activity during development was compared with mouse and human apo A-I concentrations in mouse plasma. The latter were determined by radial immunodiffusion using mouse and human apo A-I specific antibodies [8]. LCAT activity in transgenic and control mouse plasma during development, using mouse proteoliposomes as substrate, is shown in Fig. 1. LCAT activity is relatively low for both transgenic and control mice 3 days postnatally. In transgenic mice, there is a steady increase in LCAT activity up to 17 days at which time an apparent plateau is reached. The latter parallels the appearance of the adult bimodal HDL distribution noted in transgenic mice 17 days after birth [6]. LCAT activity in control mice increases more modestly and is significantly lower than that of transgenic mice at 13 and 17 days. However, LCAT activity in control mice continues to increase up to 30 days at which time it is similar to that of transgenic mice. The development-associated increase in LCAT activity in both transgenic and control mice suggests an increase in LCAT mass since Albers [9] et al. have previously demonstrated a
0 Control [ ] Transgenic
C) #
-
,
5
.
,
10
-
.
15
.
,
20
.
,
25
-
,
30
.
35
Age (days)
Fig. 1. Time course of developmental changes in LCAT activity (percent cholesteryl ester formed per hour) in plasma of transgenic and control mice. Plasma LCAT activity in transgenic and control plasma is determined by using proteoliposomes containing mouse apo A-I. Open squares represent LCAT activity in transgenic mice; open circles represent LCAT activity in plasma of control mice. Data represent mean and standard deviation of four experiments. * Differences between human and mouse apo A-I are significant at P < 0.05.
direct relationship between mass and activity. Changes in LCAT activity in both groups of mice appear to parallel plasma concentrations of human apo A-I (for transgenics) and mouse apo A-I (for controls), as shown in Fig. 2. Although speculative, higher LCAT mass in transgenics compared with controls at 13 and 17 days suggests that the human apo A-I gene may directly or indirectly increase transcription of the mouse LCAT gene. The parallel increases in apolipoprotein concentration (Fig. 2) and LCAT activity (Fig. 1) may contribute to the increased HDL particle size previously noted during devel-
180.0 160.0 140.0
120.0 --'2
100.0
"0
a E
80.0 60.0 40.0
20.0 0.0 3
8
13
17
30
Age (days)
Fig. 2. Plasma apo A-I concentrations ( m g / d l ) in human apo A-I transgenic mice and control mice during development. Open bars represent mouse apo A-I in control mice. Filled bars represent human apo A-I in transgenic mice and shaded bars represent endogenous mouse apo A-I in transgenic mice.
E. Golder-Novoselsky et al./ Biochimica et Biophysica Acta 1254 (1995) 217-220
uJ
(~ 3
3
8
13
17
30
Age (days)
Fig. 3. The efficiency of human apo A-I proteoliposomes versus mouse apo A-I proteoliposomes in activation of LCAT in transgenic mouse plasma as a function of development. Shaded bars represent LCAT activity (percent cholesteryl ester formed per hour) using human apo A-I; open bars represent mouse apo A-I. Data represent mean and standard deviation of four experiments. *Differences between human apo A-I proteoliposomes and mouse apo A-I proteoliposomes is significant at P < 0.05.
opment in both transgenic and control mice. However, in transgenic mice the increase in LCAT activity and human apo A-I concentration is not only associated with increased particle size, but also with bimodality within the HDL distribution (peaks at 9.1 and 10.9 nm) as suggested by our earlier studies. This bimodality is not likely to be attributed to endogenous mouse apo A-I since this protein is extremely low in plasma of transgenic mice. Using density gradient ultracentrifugation techniques we found that, in transgenic mice, mouse apo A-II, which is reduced approx. 50% [10], is preferentially distributed in the smaller, more dense (d1.121-1.137 g / m l ) particles; this preferential association of apo A-II with smaller HDL particles may contribute to the accumulation of the 9.1 nm particles in the transgenic mouse plasma. In transgenic mice, we found that the more buoyant dl.078-1.109 g / m l fraction, which possesses 10.7 nm particles, is enriched in human apo A-I and cholesteryl ester. This suggested that human apo A-I may be more efficient than mouse apo A-I in the activation of LCAT, thus contributing to the formation of larger core-containing HDL in transgenic mice. To directly address the question of efficiency of LCAT activation, human and mouse apo A-I-containing proteoliposomes were constructed and incubated with transgenic mouse plasma. In Fig. 3 we demonstrate, for the first time, that there is a significant difference in the ability of mouse and human apo A-I to activate mouse LCAT. Human apo A-I proteoliposomes are much
219
more efficient (1.6-fold) in activating mouse plasma LCAT than mouse apo A-I proteoliposomes. Our observation suggests that apo A-I amino acid sequences that are important in LCAT activation are not homologous between human and mouse apo A-I. The report by Januzzi et al. [11] comparing primary amino acid sequences of human and mouse apo A-I indicates a 68% sequence homology between mouse and human apo A-I. Apo A-I structure and function studies using monoclonal antibodies [12,13] have demonstrated that the region between amino acid residues 90-120 is the putative site for LCAT activation. Site-directed mutagenesis studies [14] identified an additional site associated with enzyme activation in the 148-186 region. A recent report [13] has suggested that amino acid sequence 135-146 region is also involved in LCAT activation. Upon close examination of human and mouse apo A-I amino acid sequences, one can detect variability in the sequences of all three regions believed to play a role in LCAT activation. Region 90-120 shows approx. 84% homology between species, while regions 135-146 and 148-186 have only 58% and 54% homology, respectively. Recent observations by Gong et al. [15] on structural differences between mouse apo A-I and human apo A-I revealed differences in lipid binding affinity; mouse apo A-I formed less stable complexes than human apo A-I. These differences may influence the conformation of mouse apo A-I, thus rendering it less efficient in the activation of endogenous LCAT. Our results suggest that the bimodal distribution of HDL in human apo A-I transgenic mice may depend, not only on increased lipid binding affinity of human apo A-I [15], but also on human apo A-I's ability to function as a more efficient LCAT activator than mouse apo A-I, thus generating larger, cholesteryl ester-rich HDL. The large, cholesteryl ester-containing HDL in transgenic mice is also indicative of more efficacious reverse cholesterol transport. The latter may help to explain decreased susceptibility to atherosclerosis in human apo A-I transgenic mice fed an atherogenic diet.
Acknowledgements We wish to thank Elaine Gong for kindly providing purified mouse apo A-I. This work was supported by the National Institutes of Health Program Project Grant HL18574 through the US Department of Energy Contract No. DE-AC03-76SF00098. E. G.-N. was supported by an American Heart Association of California Affiliate Predoctoral Fellowship.
References [1] Glomset, J.A. (1968) J. Clin. Invest. 58, 368-379. [2] Glomset, J.A. (1972) in Blood Lipids and Lipoproteins (G. Nelson,
220
E. Golder-Novoselsky et al. / Biochimica et Biophysica Acta 1254 (1995) 217-220
ed.), pp. 745-787, John Wiley and Sons, New York. [3] Rubin, E.M., Krauss, R.M., Spangler, E.A., Verstuyff, J.G. and Cliff, S.M. (1991) Nature 353, 265-267. [4] Walsh, A., Ito, Y. and Breslow, J.L (1989) J. Biol. Chem. 254, 6488-6494. [5] Chajek-Shaul, T., Hayek, T., Walsh, A. and Breslow, J.L. (1991) Proc. Natl. Acad. Sci. USA 88, 6731-6735. [6] Golder-Novoselsky, E., Forte, T.M., Nichols, A.V. and Rubin, E.M. (1992) J. Biol. Chem. 267, 20787-20790. [7] Chen, C.H. and Albers, J.J. (1982) J. Lipid Res. 23, 680-691. [8] Rubin, E.M., Ishida, B.Y., Cliff, S.M. and Krauss, R.M. (1991) Proc. Natl. Acad. Sci. USA 88, 434-438. [9] Albers, J.J., Chen, C.H. and Lacko, A.G. (1986) Meth. Enzymol. 129, 763-790.
[10] Schultz, J.R., Verstuyft, J.G., Gong, E.L., Nichols, A.V. and Rubin, E.M. (1993) Nature 365, 762-764. [11] Januzzi, J.J., Azrolan, N., O'Connell, A., Aalto-Setala, K. and Breslow, J.I. (1992) Genomics 14, 1081-1088. [12] Banka, C.L., Bonnet, D.J., Black, A.S., Smith, R.S. and Curtiss, L. K. (1991) J. Biol. Chem. 266, 23886-23892. [13] Meng, Q.-H., Calabresi, L., Fruchart, J.-C. and Marcel, Y.L. (1993) J. Biol. Chem. 268, 16966-16973. [14] Minnich, A., Collet, X., Roghani, A., Cladaras, C., Hamilton, R., Fielding, C. and Zannis, V.I. (1992) J. Biol. Chem. 267, 1655316560. [15] Gong, E.L., Tan, C.S., Shoukry, M.I., Rubin, E.M. and Nichols, A.V. (1994) Biochim. Biophys. Acta 1213, 335-342.
ELSEVIER
Biochimica et BiophysicaActa 1254 (1995) 217-220
etBiochi]ic~a BiophysicaA~ta
Rapid Report Human apo A-I in transgenic mice is more efficient in activating lecithin:cholesterol acyltransferase than mouse apo A-I Elina Golder-Novoselsky, Alex V. Nichols, Edward M. Rubin, Trudy M. Forte * Life Sciences Division, Lawrence Berkeley Laboratory, University of California, 1 Cyclotron Road, Berkeley, CA 94720, USA
Received 26 October 1994; accepted2 November1994
Abstract
This study shows that, in control and transgenic mice, there is a parallel increase in LCAT activity and plasma apo A-I concentrations during postnatal development. We also demonstrate that human apo A-I is a much more efficient activator (1.6-fold) of mouse LCAT activity than mouse apo A-I. We propose that the differences in amino acid sequence between human and mouse apo A-I may account for the higher LCAT activity with human apo A-I. Keywords: Transgenicmouse; LCAT activity;Apo I; (Mouse); (Human)
High-density lipoprotein (HDL) plays a central role in extracellular lipid transport and participates in the process of reverse cholesterol transport [1] by promoting cholesterol efflux from peripheral cells [2]. In this process, unesterified cholesterol is removed from peripheral cells and is converted to cholesteryl esters by the enzyme lecithin:cholesterol acyltransferase (LCAT) and returned to the liver for catabolism. LCAT is activated by apo A-I on the surface of HDL particles. In the adult C57B1/6 mouse, the HDL particle distribution consists of a single population of particles with a diameter of approx. 9.8 nm [3,4]. However, transgenic mice expressing the human apo A-I gene exhibit a bimodal distribution with major peaks at 9.1 and 10.9 nm [3,5]. The latter distribution approximates the bimodal apo A-I without apo A-II HDL profile observed in human plasma. The human apo A-I transgenic mice have HDL containing predominantly human apo A-I and trace amounts of mouse apo A-I, and, unlike control mice, are resistant to diet-induced atherosclerosis [3]. We have recently shown that HDL distribution during late prenatal and early postnatal development in both control and transgenic mice was consistently different from that of adult mice and was skewed to smaller sized particles [6]. 2 days before delivery, a 7.5 nm particle was
* Corresponding author. Fax: + 1 (510) 486 4750. 0005-2760/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0005-2760(94)00230-4
the predominant species in both control and transgenic mice. In control mice 3 days after delivery, a monodisperse peak, consisting of 9.0 nm particles, was noted which gradually increased in size to 9.8 nm, the adult size, by day 13.3 days after delivery, HDL of transgenic mice exhibited a bimodal distribution in which the particles were smaller, 7.5 and 8.9 nm, than those of adult transgenic mice. With increasing time after delivery, the bimodal pattern showed larger particles and developed into the adult pattern by 17 days after birth. The stepwise increment in size of control and transgenic HDL particles was associated with increased plasma concentrations of mouse and human plasma apo A-I. We previously suggested that the predominance of smaller particles during the neonatal period might be indicative of reduced LCAT activity [6]. In the present studies, we examined whether differences in LCAT activity during development may contribute to the observed differences in HDL profiles between control and human apo A-I transgenic mice. Since human apo A-I is the predominant apolipoprotein on HDL in transgenic mice, we investigated the effectiveness of human vs. mouse apo A-I in activating endogenous mouse LCAT. LCAT activity was determined by the substrate independent proteoliposome assay in which the rate of cholesteryl ester formation was measured. This assay utilized proteoliposomes prepared from purified human a n d / o r mouse apo A-I, [14C]cholesterol and lecithin according to method
E. Golder-Novoselsky et al. /Biochimica et Biophysica Acta 1254 (1995) 217-220
218
of Chen and Albers [7]. Both human and mouse apo A-I were incorporated into liposomes in equivalent amounts and proteoliposomes containing 0.22 m g / m l were use in the present study. Electron microscopy was used to evaluate the structural properties of human and mouse apo A-I proteoliposomes in order to establish that there were no major physical differences between the two types of proteoliposomes. In both cases the proteoliposomes consisted of small vesicles which were similar to one another in size where human apo A-I proteoliposomes were 34.6 _ 11 nm diameter and mouse apo A-I proteoliposomes, 35.3 + 13 nm. Plasma from transgenic and control mice was used as the source of LCAT during the following developmental time points: 3, 8, 13, 17 and 30 days after birth. LCAT activity during development was compared with mouse and human apo A-I concentrations in mouse plasma. The latter were determined by radial immunodiffusion using mouse and human apo A-I specific antibodies [8]. LCAT activity in transgenic and control mouse plasma during development, using mouse proteoliposomes as substrate, is shown in Fig. 1. LCAT activity is relatively low for both transgenic and control mice 3 days postnatally. In transgenic mice, there is a steady increase in LCAT activity up to 17 days at which time an apparent plateau is reached. The latter parallels the appearance of the adult bimodal HDL distribution noted in transgenic mice 17 days after birth [6]. LCAT activity in control mice increases more modestly and is significantly lower than that of transgenic mice at 13 and 17 days. However, LCAT activity in control mice continues to increase up to 30 days at which time it is similar to that of transgenic mice. The development-associated increase in LCAT activity in both transgenic and control mice suggests an increase in LCAT mass since Albers [9] et al. have previously demonstrated a
0 Control [ ] Transgenic
C) #
-
,
5
.
,
10
-
.
15
.
,
20
.
,
25
-
,
30
.
35
Age (days)
Fig. 1. Time course of developmental changes in LCAT activity (percent cholesteryl ester formed per hour) in plasma of transgenic and control mice. Plasma LCAT activity in transgenic and control plasma is determined by using proteoliposomes containing mouse apo A-I. Open squares represent LCAT activity in transgenic mice; open circles represent LCAT activity in plasma of control mice. Data represent mean and standard deviation of four experiments. * Differences between human and mouse apo A-I are significant at P < 0.05.
direct relationship between mass and activity. Changes in LCAT activity in both groups of mice appear to parallel plasma concentrations of human apo A-I (for transgenics) and mouse apo A-I (for controls), as shown in Fig. 2. Although speculative, higher LCAT mass in transgenics compared with controls at 13 and 17 days suggests that the human apo A-I gene may directly or indirectly increase transcription of the mouse LCAT gene. The parallel increases in apolipoprotein concentration (Fig. 2) and LCAT activity (Fig. 1) may contribute to the increased HDL particle size previously noted during devel-
180.0 160.0 140.0
120.0 --'2
100.0
"0
a E
80.0 60.0 40.0
20.0 0.0 3
8
13
17
30
Age (days)
Fig. 2. Plasma apo A-I concentrations ( m g / d l ) in human apo A-I transgenic mice and control mice during development. Open bars represent mouse apo A-I in control mice. Filled bars represent human apo A-I in transgenic mice and shaded bars represent endogenous mouse apo A-I in transgenic mice.
E. Golder-Novoselsky et al./ Biochimica et Biophysica Acta 1254 (1995) 217-220
uJ
(~ 3
3
8
13
17
30
Age (days)
Fig. 3. The efficiency of human apo A-I proteoliposomes versus mouse apo A-I proteoliposomes in activation of LCAT in transgenic mouse plasma as a function of development. Shaded bars represent LCAT activity (percent cholesteryl ester formed per hour) using human apo A-I; open bars represent mouse apo A-I. Data represent mean and standard deviation of four experiments. *Differences between human apo A-I proteoliposomes and mouse apo A-I proteoliposomes is significant at P < 0.05.
opment in both transgenic and control mice. However, in transgenic mice the increase in LCAT activity and human apo A-I concentration is not only associated with increased particle size, but also with bimodality within the HDL distribution (peaks at 9.1 and 10.9 nm) as suggested by our earlier studies. This bimodality is not likely to be attributed to endogenous mouse apo A-I since this protein is extremely low in plasma of transgenic mice. Using density gradient ultracentrifugation techniques we found that, in transgenic mice, mouse apo A-II, which is reduced approx. 50% [10], is preferentially distributed in the smaller, more dense (d1.121-1.137 g / m l ) particles; this preferential association of apo A-II with smaller HDL particles may contribute to the accumulation of the 9.1 nm particles in the transgenic mouse plasma. In transgenic mice, we found that the more buoyant dl.078-1.109 g / m l fraction, which possesses 10.7 nm particles, is enriched in human apo A-I and cholesteryl ester. This suggested that human apo A-I may be more efficient than mouse apo A-I in the activation of LCAT, thus contributing to the formation of larger core-containing HDL in transgenic mice. To directly address the question of efficiency of LCAT activation, human and mouse apo A-I-containing proteoliposomes were constructed and incubated with transgenic mouse plasma. In Fig. 3 we demonstrate, for the first time, that there is a significant difference in the ability of mouse and human apo A-I to activate mouse LCAT. Human apo A-I proteoliposomes are much
219
more efficient (1.6-fold) in activating mouse plasma LCAT than mouse apo A-I proteoliposomes. Our observation suggests that apo A-I amino acid sequences that are important in LCAT activation are not homologous between human and mouse apo A-I. The report by Januzzi et al. [11] comparing primary amino acid sequences of human and mouse apo A-I indicates a 68% sequence homology between mouse and human apo A-I. Apo A-I structure and function studies using monoclonal antibodies [12,13] have demonstrated that the region between amino acid residues 90-120 is the putative site for LCAT activation. Site-directed mutagenesis studies [14] identified an additional site associated with enzyme activation in the 148-186 region. A recent report [13] has suggested that amino acid sequence 135-146 region is also involved in LCAT activation. Upon close examination of human and mouse apo A-I amino acid sequences, one can detect variability in the sequences of all three regions believed to play a role in LCAT activation. Region 90-120 shows approx. 84% homology between species, while regions 135-146 and 148-186 have only 58% and 54% homology, respectively. Recent observations by Gong et al. [15] on structural differences between mouse apo A-I and human apo A-I revealed differences in lipid binding affinity; mouse apo A-I formed less stable complexes than human apo A-I. These differences may influence the conformation of mouse apo A-I, thus rendering it less efficient in the activation of endogenous LCAT. Our results suggest that the bimodal distribution of HDL in human apo A-I transgenic mice may depend, not only on increased lipid binding affinity of human apo A-I [15], but also on human apo A-I's ability to function as a more efficient LCAT activator than mouse apo A-I, thus generating larger, cholesteryl ester-rich HDL. The large, cholesteryl ester-containing HDL in transgenic mice is also indicative of more efficacious reverse cholesterol transport. The latter may help to explain decreased susceptibility to atherosclerosis in human apo A-I transgenic mice fed an atherogenic diet.
Acknowledgements We wish to thank Elaine Gong for kindly providing purified mouse apo A-I. This work was supported by the National Institutes of Health Program Project Grant HL18574 through the US Department of Energy Contract No. DE-AC03-76SF00098. E. G.-N. was supported by an American Heart Association of California Affiliate Predoctoral Fellowship.
References [1] Glomset, J.A. (1968) J. Clin. Invest. 58, 368-379. [2] Glomset, J.A. (1972) in Blood Lipids and Lipoproteins (G. Nelson,
220
E. Golder-Novoselsky et al. / Biochimica et Biophysica Acta 1254 (1995) 217-220
ed.), pp. 745-787, John Wiley and Sons, New York. [3] Rubin, E.M., Krauss, R.M., Spangler, E.A., Verstuyff, J.G. and Cliff, S.M. (1991) Nature 353, 265-267. [4] Walsh, A., Ito, Y. and Breslow, J.L (1989) J. Biol. Chem. 254, 6488-6494. [5] Chajek-Shaul, T., Hayek, T., Walsh, A. and Breslow, J.L. (1991) Proc. Natl. Acad. Sci. USA 88, 6731-6735. [6] Golder-Novoselsky, E., Forte, T.M., Nichols, A.V. and Rubin, E.M. (1992) J. Biol. Chem. 267, 20787-20790. [7] Chen, C.H. and Albers, J.J. (1982) J. Lipid Res. 23, 680-691. [8] Rubin, E.M., Ishida, B.Y., Cliff, S.M. and Krauss, R.M. (1991) Proc. Natl. Acad. Sci. USA 88, 434-438. [9] Albers, J.J., Chen, C.H. and Lacko, A.G. (1986) Meth. Enzymol. 129, 763-790.
[10] Schultz, J.R., Verstuyft, J.G., Gong, E.L., Nichols, A.V. and Rubin, E.M. (1993) Nature 365, 762-764. [11] Januzzi, J.J., Azrolan, N., O'Connell, A., Aalto-Setala, K. and Breslow, J.I. (1992) Genomics 14, 1081-1088. [12] Banka, C.L., Bonnet, D.J., Black, A.S., Smith, R.S. and Curtiss, L. K. (1991) J. Biol. Chem. 266, 23886-23892. [13] Meng, Q.-H., Calabresi, L., Fruchart, J.-C. and Marcel, Y.L. (1993) J. Biol. Chem. 268, 16966-16973. [14] Minnich, A., Collet, X., Roghani, A., Cladaras, C., Hamilton, R., Fielding, C. and Zannis, V.I. (1992) J. Biol. Chem. 267, 1655316560. [15] Gong, E.L., Tan, C.S., Shoukry, M.I., Rubin, E.M. and Nichols, A.V. (1994) Biochim. Biophys. Acta 1213, 335-342.