Eur. J. Biochem. 230, 567-575 (1995) 0 FEBS 1995

Tissue-specific expression of the human gene for lecithin : cholesterol acyltransferase in transgenic mice alters blood lipids, lipoproteins and lipases towards a less atherogenic profile Anja MEHLUM’, Bart STAELS 2.3, Nicolas DUVERGER4, Anne TAILLEUX’, Graciela CASTRO’, Catherine F’IEVET’, GCrald LUC2,h, Jean-Charles FRUCHART’, Gunilla OLIVECRONA5, Grethe SKRETTING’, Johan AUWERXZ,3 and Hans PRYDZ’ ’ Biotechnology Centre of Oslo, University of Oslo and the Animal Department, Rikshospitalet, Oslo, Norway * U.325 INSERM, DCpartement d’ AthCrosclCrose, Institut Pasteur de Lille, France

Laboratoire de Biologie des RCgulations chez les Eucaryotes, DCpartement d’Ath6rosclCrose, Institut Pasteur de Lille, France Atherosclerosis Department, C. R. V. A. RhBne-Poulenc Rorer SA, Vitry sur Seine, France

’ Department of Medical Biochemistry and Biophysics, University of Umeb, Sweden ‘ Centre Universitaire de Mesures et d’Analyses, FacultC de Pharmacie de Lille, France

(Received 31 March 1995) - EJB 95 0517/1

Lecithin :cholesterol acyltransferase (LCAT) is a key enzyme in the reverse cholesterol pathway but its role in lipid metabolism is still unclear. We have generated mice transgenic for a 7-kb genomic DNA fragment comprising the 6 exons and 5 introns of the LCAT gene with 1932 bp of 5‘ flanking and 908 bp of 3‘ flanking sequences. One line had integrated about 30 copies and expressed about 40-fold increased LCAT activity in a human test system. The expression showed correct tissue specificity of the human LCAT gene. Increased LCAT activity resulted in a decrease of plasma triacylglycerols below 50% of fasting controls. This reduction was seen in all lipoprotein fractions. Lipoprotein lipase activity did not change significantly, whereas hepatic triacylglycerol lipase increased markedly. Plasma total cholesterol was similar in fasting transgenic and control mice, but low-density lipoprotein and very low-density lipoprotein cholesterol were reduced to about 50 %. High-density lipoprotein cholesterol increased about 20 %, accompanied by a correspondingly increased size and a higher cholesterol efflux-stimulating activity of transgenic LCAT high-density lipoprotein . Both apolipoprotein A-I and A-I1 plasma concentrations increased in transgenic mice. Plasma triacylglycerol and cholesteryl ester fatty acid distribution showed an increased proportion of palmitic acid, whereas oleic, linoleic and arachidonic acid decreased, thus resembling more closely the human situation.

Overexpression of the human LCAT gene provokes major changes in plasma lipoprotein and apolipo- protein concentrations, resulting in a less atherogenic plasma lipoprotein profile through a reduction in atherogenic and an increase in anti-atherogenic lipoproteins.

Keywords. Atherosclerosis ; triacylglycerol ; cholesterol ester; transgenic mice ; high-density lipoprotein.

Transgenic mice have provided interesting and useful models for the study of various disease processes. Several transgenic animal models have been established for the study of genes in- volved in atherogenesis (Breslow, 1994; Shimada et al., 1993; Liu et al., 1994; Hedrick et al., 1993). Lecithin:cholesterol acyltransferase (LCAT) is a 67-kDa single-chain plasma glyco- protein and a key enzyme in the reverse cholesterol pathway (Glomset, 1972). It catalyses the transfer of an acyl chain mainly from the sn-2 position of phosphatidylcholine to cholesterol, re- sulting in the production of lysophosphatidylcholine and cholest- eryl esters (Glomset, 1972).

Correspondence to H. Prydz, Biotechnology Centre of Oslo, Univer-

Fax: +47 22 69 41 30. Abbreviations. LCAT, lecithin :cholesterol acyltransferase; apo, apo-

lipoprotein ; HDL, high-density lipoprotein; LDL, low-density lipopro- tein; VLDL, very low-density lipoprotein.

Enzymes. Lecithin : cholesterol acyltransferase (EC 2.3.1.43) ; lipo- protein lipase (EC 3.1.1.34); triacylglycerol lipase (EC 3.1.1.3); glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.21).

Note. The first three named authors contributed equally to this work.

sity of Oslo, GaustadallCen 21, N-0371 Oslo, Norway

LCAT is activated by apolipoprotein (apo) A-I, the major protein component of plasma high-density lipoproteins (HDL) and consequently esterifies primarily HDL cholesterol (Fielding, 1972). By maintaining the ratio of free to esterified cholesterol in plasma lipoproteins, LCAT may promote the efflux of cellular cholesterol from peripheral cells to HDL, the first step of the reverse cholesterol transport pathway leading to cholesterol elimination by the liver. This enzyme plays therefore a crucial role in regulating the distribution of unesterified cholesterol be- tween cell membranes and plasma lipoproteins.

The physiological importance of plasma LCAT activity and cholesterol esterification is well illustrated by the pathological changes observed in patients with partial or complete LCAT de- ficiency. In normal subjects, LCAT is responsible for the matura- tion of nascent HDL, the transformation of small to large HDL, and the removal of cholesterol and phospholipids from apoB- containing lipoproteins. Mutations in the LCAT gene give rise to different phenotypes. We (Skretting et al., 1992; Skretting and Prydz, 1992) and others (Funke et al., 1991; Klein et al., 1992, 1993a,b; Assman et al., 1991; Funke et al., 1993) have shown that fish eye disease (Carlson and Philipson, 1979; Carlson,

568 Mehlum et al. (Eur: J. Biochem. 230)

1982) and human classical LCAT deficiency (Gjone and Norum, 1968) are caused by different defects in the same gene. In its most severe form the LCAT deficiency phenotype includes renal failure, anemia, corneal opacity, lipoprotein abnormalities and premature atherosclerosis (Norum et al., 1989). The concentra- tion, composition and shape of all lipoprotein subclasses are ab- normal, and mature HDL is absent from plasma. The renal fail- ure and corneal opacities result from lipid deposition. In addi- tion, erythrocyte membranes of patients have an abnormally high content of cholesterol and phosphatidylcholine.

While partial or complete LCAT deficiency is clearly associ- ated with decreased HDL concentrations, almost nothing is known concerning the effects of increased plasma LCAT activ- ity. We wanted to analyse the effects of LCAT overexpression on lipoprotein metabolism. Therefore, transgenic mice overex- pressing human LCAT have been created (Mehlum and Prydz, 1994). In this paper we characterize the expression of the LCAT transgene and the changes in plasma lipoprotein metabolism in these transgenic mice. We demonstrate that expression of human LCAT in mice results in significant changes in lipid metabolism and plasma lipoprotein levels through a reduction of VLDL and LDL and an increase in HDL cholesterol. Furthermore, the tria- cylglycerol levels in fasting as well as in fed transgenic mice are markedly reduced compared to those in correspondingly treated controls. Overexpression of LCAT alters the expression of he- patic triacylglycerol lipase whereas lipoprotein lipase does not change significantly. The increase in hepatic triacylglycerol li- pase and the possible differences in the structure of the human and murine LCAT active sites may all contribute to the alter- ations observed in the plasma lipoprotein and fatty acid distribu- tion.

MATERIALS AND METHODS

Transgenic animals. The cDNA and genomic DNA for LCAT have been cloned by us (Rogne et al., 1987; EMBL ac- cession no. X51966) and McLean et al. (1986a,b). The LCAT gene was cloned and sequenced from a human cosmid library. The cosmid clone was mapped and most of it sequenced (Larsen et al., 1993) using an EMBL prototype sequencer with fluores- cent primers as described (Voss et al., 1992). A clone was pre- pared by ligating an EcoRI-Hind111 fragment to a SmaI-SmaI fragment in their common BglII site. From this clone a FspI- EugI fragment of total length 7 kb, comprising the six exons of the LCAT gene (Larsen et al., 1993), 1932 bp of upstream se- quence and 908 bp of downstream sequence (Fig. 1 B), was puri- fied by agarose electrophoresis and subsequently treated with GenecleaP before microinjection into fertilized eggs from C57B1/6 mice using standard methodology (Hogan et al., 1986). The eggs were reimplanted into pseudopregnant NMRI foster mothers.

Blood and tissue samples. The transgenic animals used in these studies were 14-32-week-old males hemizygous for the LCAT insert. Control mice were non-transgenic litter mates or other C57BV6 mice of similar age. The animals were weighed before sacrifice. Tissue samples were immediately frozen in li- quid nitrogen and kept at -80°C until use. Blood samples were collected from the vena cava in anesthetized mice after an over- night fast if not otherwise stated. For the analyses of cholesterol and other lipids, blood was drawn into 10 p1 0.5 M EDTA pH 8.5, 50 mg . ml-I gentamycin sulfate and 0.1% sodium azide. Blood samples, drawn from mice 15 min after injection of 1000 IU . kg-' heparin intraperitoneally, were centrifuged at 11 000 g for 10 min and the plasma used for assay of lipoprotein lipase and hepatic triacylglycerol lipase activities. Serum from

untreated transgenic and control mice was obtained by centrifu- gation in the same way and used for preparation of HDL for cholesterol efflux. LCAT activity was measured in serum and in citrated plasma. All P values were obtained by the t-test.

Preparation and analysis of RNA and DNA. Total cellular RNA was isolated from various tissues of transgenic and control mice by the method described by Chomczynski and Sacchi (1987) and mRNA isolated using paramagnetic oligo(dT) beads (Jakobsen et al., 1990). LCAT expression was monitored using a full-length human LCAT cDNA (Rogne et al., 1987). Other cDNA probes used were rat and murine apoA-I, apoA-11, apoA- IV, apoB, apoE and human and mouse lipoprotein lipase and hepatic triacylglycerol lipase. A rat apoC-I11 cDNA fragment spanning nucleotides +54 to +356 was cloned from rat liver by reverse transcription and PCR-amplification (sense primer: ATG CAG CCC CGA ATG CTC CTC ATC GTG GCC; anti-sense primer: TCA CGG CTC AAG AGT TGG TGT TGT TAG TTG GTC CTC AGG). Sequence analysis revealed complete identity to the previously reported rat apoC-I11 cDNA sequence (Haddad et al., 1986). Probes for human p-actin and glyceraldehyde-3- phosphate dehydrogenase were used to control the amount of RNA applied. All cDNA probes were labelled by random prim- ing (Feinberg and Vogelstein, 1983). Northern blots were per- formed as described by Fourney et al. (1988) and Staels et al. (1 989). Autoradiograms were analyzed by quantitative scanning densitometry using BioRad GS670 or Shimadzu CS 9000 densi- tometers as described (Staels et al., 1989).

Genomic DNA was prepared from mouse tails by the method of Laird et al. (1991). For Southern blots 10 pg DNA was regularly used. DNA probes were labelled by random prim- ing (Feinberg and Vogelstein, 1983).

LCAT activity, cholesterol esterification rate, apolipo- protein, lipid and lipoprotein measurements. LCAT activity was determined using an exogenous proteoliposome substrate as described by Chen and Albers (1982). Under the relevant conditions, this assay was linear with time and enzyme concen- trations for over 20 min (data not shown) and the LCAT activity was determined at 10 and 20 min. The esterification rate of cho- lesterol was determined as described by Stokke and Norum (1971). Plasma lipids (cholesterol, phospholipids and triacyl- glycerols) and HDL cholesterol were measured colorimetrically (490 nm) using a microtiter plate reader (Bio-Teck Instruments EL 311) and reagents from Bio-Merieux (Marcy-L'Etoile, France). Mouse apoA-I and apoA-I1 concentrations were mea- sured by immunonephelometry using specific polyclonal anti- bodies. Lipoprotein fractions were isolated by sequential ultra- centrifugation as described (Brousseau et al., 1994). The specific protein composition of particles was analyzed by a 4-20% acrylamide gradient SDSPAGE (Laemmli, 1970). Plasma lipo- protein distribution was assayed by gel filtration chromatogra- phy (Duverger et al., 1993). Fasting plasma aliquots were ap- plied to a Superose 6 HR 10/30 prepacked column (Pharmacia), chromatographed at a flow rate of 0.4 ml . min-' in 10 mM Tris/ HC1 pH 7.4, 0.15 M NaCl and 1 mM EDTA and monitored at 280 nm. This system allows separation of the three major lipo- protein classes: VLDL, LDL and HDL. LCAT activity, protein, cholesterol and triacylglycerol concentrations were determined in the eluted fractions. Cholesteryl ester in the various fractions was calculated according to the formula cholesteryl ester = 1.67 (total cholesterol-free cholesterol). HDL fractions used for cho- lesterol efflux studies were isolated from serum by gel filtration chromatography using Superose 6 and 12 HR columns in series. The column fractions were assayed for their content of choles- terol (Boehringer Mannheim) and total protein (Lowry et al., 1951).

Mehlum et al. (EM J. Biochem. 230) 569

Cellular cholesterol efflux studies. Cellular cholesterol ef- flux was determined using rat Fu5AH hepatoma cells as de- scribed previously (Moya et al., 1994). Briefly, the cells were plated in 2 ml minimal essential medium containing 5% calf serum on 2.4-cm multiwell plates at a density of 20000/ml. Af- ter 48 h [3H]cholesterol (1 pCi/well) was added to the culture medium and cellular cholesterol was labelled for a further 48 h. To allow equilibration of the label, the cells were rinsed and incubated for 24 h in minimal essential medium with 0.5 % BSA. For determination of cholesterol efflux, the cells were washed with phosphate-buffered saline and incubated at 37°C for 1 or 2 h with 50 pg . ml-' of a HDL fraction isolated by gel filtration chromatography. At the end of the incubation, medium was re- moved and centrifuged. Cell monolayers were washed three times with phosphate-buffered saline and harvested in 0.5 ml 0.1 M NaOH. Cellular and medium radioactivity were measured and used to calculate the percentage of cholesterol efflux.

Fatty acid composition. The fatty acid composition of plasma triacylglycerols, phospholipids and cholesteryl esters was determined by gas chromatography. Plasma lipids were ex- tracted by addition of CHCl,/CH,OH (2 : 1, by vol.) according to Folch et al. (1957). The lipid extract was dried under nitrogen, dissolved in heptane and separated by thin-layer chromatogra- phy using hexane/diethylether/acetic acid (80 : 20 :2, by vol.). Lipids were revealed by iodine vapour and cholesteryl esters, triacylglycerols and phospholipids were identified by simulta- neous migration of standards. Each fraction was scraped and methanolyzed with methanolic H,SO, for 2 h at 70°C. The fatty acid methyl esters were extracted with heptane and analyzed by gas chromatography on a Varian 3400 chromatograph (Sunny- vale, CA) equipped with a flame ionization detector and a capil- lary WCOT fused Silica 88 column containing cyanopropylpo- lysiloxane (Chrompack, Middelburg, The Netherlands). A gradi- ent temperature from 160" to 210°C at 3°C . min-' and a N, pressure of 80 kPa were used. Identification of fatty acids was achieved using commercially available fatty acid methyl esters.

Determination of lipoprotein lipase and hepatic triacyl- glycerol lipase activity. Lipoprotein lipase was measured using an emulsion containing Intralipid (10 %) into which 3H-labelled trioleoylglycerol was incorporated by sonication (Bengtsson- Olivecrona and Olivecrona, 1992). Heat-inactivated rat serum was used as source of apolipoprotein C-11. Hepatic triacylglyc- erol lipase activity was measured with a gum-arabic-stabilized emulsion of trioleolylglycerol containing 1 M NaCl to inhibit any lipoprotein lipase activity in the sample (Bengtsson- Olivecrona and Olivecrona, 1992). Hepatic triacylglycerol lipase activity in the lipoprotein lipase assay was supressed by treat- ment of plasma samples with rabbit immunoglobulin G against it before the assay (Peterson et al., 1986); the antibody was shown to inhibite more than 80% of the activity in plasma samples. The remaining lipase activity in the assay after treat- ment of the sample with antibody was taken as lipoprotein lipase. Lipase activities are expressed as mU . ml-I plasma when 1 mU corresponds to the amount catalyzing the release of 1 nmol fatty acid . min-'.

Size of HDL particles. HDL particles were isolated from serum of non-fasting transgenic and control animals by flotation at a density of 1.210 g . ml-'. The size distribution of particles was analysed by electron microscopy of negatively stained prep- arations (Forte and Nordhausen, 1986). The diameters of at least 200 particles/micrograph were determined with an ocular mi- crometer (primary magnification 60 000 x, photographic magni- fication 5X). Per sample, 5- 10 micrographs were analysed blindly. The molecular mass of HDL particles from transgenic mice was also compared to that of HDL from control mice by non-denaturing gel electrophoresis on precast ultrathin poly-

acrylamide gradient gels (8 -25 %, Phast-Gel, Pharmacia) in 112 mM acetate, 112 mM Tris pH 7.4. The samples (1 pg pro- tein) were applied using loading applicators of 1 p1. The buffer strips (Phast-Gel native buffer strips, Pharmacia) were 2% agar- ose in 0.88M L-alanine and 0.25 M Tris pH 8.8. The electropho- resis was performed in a Phast-System apparatus (Pharmacia) at 400 V for 40 min at 15°C. Gels were stained with Phast-Blue R350 (0.2% mass/vol.) in 10% (by vol.) acetic acid solution. The staining and destaining steps were automatically performed using the Phast-System development unit. Thyroglobulin (665 kDa), ferritin (440 m a ) , catalase (232 m a ) , lactate dehy- drogenase (140 kDa) and albumin (67 kDa) (high-molecular- mass calibration mixture, Pharmacia) were used as calibrating proteins on each gel.

RESULTS

Transgenic animals. The LCAT gene was cloned and se- quenced from a human cosmid library. The fragment injected had a total length of 7 kb, comprising the six exons of the LCAT gene (Larsen et al., 1993), 1932 bp of upstream sequence and 908 bp of downstream sequence (Fig. 1 B). Since one of our purposes was to study the effect of LCAT overexpression on the susceptibility to atherosclerosis, we chose to use the C57B1/6 mice to produce transgenics. This strain is susceptible to devel- opment of atherosclerotic lesions at the base of the aorta (Paigen et al., 1990; Stewart-Philips and Lough, 1991).

Five transgenic mice were born. The transgenic state was verified by Southern blot analysis of a BamHI digest of tail DNA using a 1.6-kb FspI-BglII fragment from the upstream region as a probe (Fig. 1). BamHI cuts only once in each copy of the insert. The band at about 7 kb shows the expected full length of the human LCAT insert obtained when tandemly in- serted gene copies are cut once internally. The endogenous hu- man gene (lane 4) gives a band of 2.7 kb as expected. Other bands are flanking sequences or may result from the star activity of BamHI (George et al., 1980). All the analyses so far have been performed with one transgenic line with the full-length hu- man LCAT gene in about 30 copies integrated into one site. Similar analyses of tail DNA using BglII fully confirmed this interpretation. No segregation was observed at this locus during 2.5 years of breeding.

Characteristics of the LCAT transgenic line. The LCAT transgenic mice have been bred for 31 months. The first litter from the founder is still healthy at this age. A properly con- trolled life-length experiment is underway. The body masses for transgenic and normal mice have been similar at all stages of development. Litter sizes have been normal. The pattern of the transgene has remained unchanged over 2 years (Fig. 1A) through more than 100 meioses. The hybridization pattern ob- tained using genomic DNA from the transgenic mice suggests that the copies (about 30) of the injected gene are not inserted in a simple head to tail tandem pattern.

Expression of human LCAT mRNA. LCAT is a liver-specific enzyme in humans, in rodents and rhesus monkeys, minor ex- pression also being detected in brain and testes (Warden et al., 1989; Smith et al., 1990). We isolated RNA from various tissues (muscle, brain, lung, kidney, testes, small intestine, spleen, liver, heart). The transgenic LCAT mice expressed the human LCAT gene (as well as endogenous mRNA) primarily in the liver (Fig. 2). The expression was enhanced about 30-fold when com- pared to mRNA from control mice. Longer exposures revealed a weak expression in heart, adipose tissue (Fig. 2) and in brain

570 Mehlum et al. ( E m J. Biochem. 230)

1 2 3 4 A (kb)

6.1 - 4.3 - 3.4 - 2.6 - 2.3 - 2.2 - 1.2 -

0.5 -

B EcoRl Smal &Ill Hindlll

I I l l Smal

Eagl r? n ,

0 7449

probe - 1 Kb

Fig. 1. Southern blot of tail DNA from transgenic and control mice (A) and human LCAT genomic DNA fragment injected into mouse oocytes (B). (A) Tail DNA (10 pg) was digested with BamHI and ap- plied in a 0.8 % agarose gel, blotted onto a nylon membrane and hybrid- ized with the FspI-BgAI fragment. The filter was hybridized at 68°C in 105 mM NaCl, 10.5 mM sodium citrate, 1 % SDS, 0.2% pyrophosphate, 10 mM EDTA, 20 mM sodium phosphate pH 7, 10% dextran sulfate, 100 pg . ml-' salmon sperm DNA and washed in 150 mM NaCI, 15 mM sodium citrate, 1 % SDS (68°C). A trace of pBR325 sequences was in- cluded in the probe before labelling by random priming. Lane 1, pBR325 marker; lane 2, tail DNA from transgenic mice; lane 3, tail DNA from normal mouse; lane 4, human leukocyte DNA. (B) LCAT genomic DNA was cloned from a human cosmid library by using a 1-kb fragment of LCAT cDNA. The previously unreported 5' upstream sequence was de- termined (EMBL accession no. X51966). The DNA fragment used for injection was prepared by ligating the EcoRI-Hind111 fragment and the SmaI-SmaI fragment of LCAT. The two clones were ligated in the BglII site. DNA was purified by agarose gel electrophoresis and subsequently treated with Geneclean. The FspI-BglII fragment used to probe South- em blots i s indicated.

and testes (not shown). These results indicate that the 7-kb frag- ment contains the essential cis-acting elements necessary for correct expression in the liver.

LCAT activity and cholesterol esterification rate. The endog- enous LCAT activity in normal and transgenic mice was deter- mined with an exogenous proteoliposome substrate containing human apoA-I. In control mice the activity was 106 +- 15 nmol .

NON-TRANSGENIC TRANSGENIC

Li H A K SI Te Li H A K SI Te

-

LCAT

- GAPDH

Fig. 2. Tissue-specific expression of the human LCAT transgene. Northern blot of mRNA from transgenic and control mice of the same litter at the age of 6 weeks. The poly(A)-rich mRNA fraction was iso- lated from 20 pg total RNA from liver (Li), heart (H), adipose tissue (A), kidney (K), total small intestine (SI) and testis (Te). The mRNA fractions were run on denaturing gel, blotted onto a nylon membrane and hybridized with full-length human LCAT cDNA. Exposure time 24 h. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Table 1. Plasma triacylglycerols in lipoprotein fractions of male mice transgenic for LCAT. The values are mean? SEM of data from 11 mice in each group. Percentages and P values refer to comparison be- tween transgenic mice and the corresponding controls; ns., not signifi- cant.

Mice Frac- Triacylglycerol content P ~- tion

control transgenic

mg . dlY (% control)

Fasting VLDL 69.1 ? 8.2 34.4 2 8.8 (49.8) <0.02 LDL 25.7 ? 2.1 8.5 t 1.0 (33.1) <0.001 HDL 7.4 2 1.3 4.0 2 2.2 (54.1) n.s. total 102.0 2 8.6 46.8 2 8.8 (45.9) <0.001

Non-fasting total 127.2 2 12.6 101.0 2 12.4 (79.4) n.s.

(ml . h)-' and in the transgenic mice 4431 +216 nmol . (ml . h)-'. The apparent LCAT activity in mice hemizygous for the human LCAT gene was thus about 40-fold increased compared to that of the endogenous murine enzyme. One caveat for this comparison is the possibility that murine LCAT is less activated by human apoA-I than its human counterpart. The estimate of 40-fold is therefore a maximum value, suggesting that the true enzyme activity ratio may correspond to the copy number of the transgene. By contrast, using endogenous substrate the plasma cholesterol esterification rate in the transgenic mice was only half of that in control mice: control 1 0 6 2 3 pmol . (ml . h)-'; transgenic 53 2 3 pmol . (ml . h)-'.

Plasma lipoproteins and apolipoprotein levels. Overexpres- sion of human LCAT induced a strong perturbation of lipid and apolipoprotein metabolism in mice. Surprisingly, the most pro- nounced alteration in plasma lipid levels was a > 50 % reduction in plasma triacylglycerols (Table 1). Triacylglycerol concentra- tions were reduced in the three major lipoprotein subfractions in transgenic mice (Table I). Plasma total cholesterol concentra- tions did not vary significantly between fasting transgenic and control inice whereas plasma free-cholesterol concentrations were decreased to 83% of control (Table 2), but the difference was not statistically significant (P = 0.2). In non-fasting transgenic mice, however, total cholesterol was increased by 69% compared to controls (Table 2). This increased total choles-

Mehlum et al. (EUK J. Biochem. 230) 571

Table 2. Cholesterol in lipoprotein fractions of male mice transgenic for LCAT. The values of total and free cholesterol and cholesteryl ester are mean t SEM. Percentages and P values refer to comparison between transgenic mice and the corresponding controls; N = number of animals examined for control (c) or transgenic (t) mice. Amounts of cholesteryl esters were calculated as described in Materials and Methods.

Mice N for Fraction Cholesterol content

C t control

P

transgenic

mg . dl-' (% control)

Fasting total cholesterol

free cholesterol

cholesteryl ester

Non-fasting total cholesterol

free cholesterol

11 9 VLDL LDL HDL plasma

VLDL LDL HDL plasma

VLDL LDL HDL plasma

7 6 HDL plasma

3 2 plasma

6.9 2 0.6 21.6 t 2.9 72.7 2 2.4

101.2 2 5.2 6.1 2 0.5 9.9 t 1.3

16.1 t 0.9 32.1 t 2.0 1.3 t 0.8

19.5 2 3.1 94.5 2 3.1

116.2 t 6.5

76.2 2 3.7 81.7 t 2.8 11.6

3.0? 0.8 (43) 13.2 2 0.9 (61) 87.1 t 3.4 (120)

103.2 ? 4.0

2.8 t 0.6 6.5 ? 0.5

17.3 ? 0.9 26.6 ? 1.2 0.2 t 0.4

11.1 t 0.9 116.6 2 4.4 127.9 t 4.9

130.2 ? 11.1 (171) 137.8 2 9.3 (169) 17.4

<0.005 <0.02 C0.005

<0.005 <0.05

<0.05 <0.005 <0.01 <0.005 <0.005

<0.01 <0.005

terol in non-fasting transgenic mice was essentially caused by an increase in esterified HDL cholesterol. Although total choles- terol levels in plasma from fasting transgenic and control mice were similar, cholesterol distribution in transgenic mice was dif- ferent from control: VLDL-C and LDL-C were reduced (43 % and 61% of control, respectively) whereas HDL-C (120% of control) was increased (Table 2). Correspondingly, cholesteryl esters in VLDL and LDL were significantly lower, whereas HDL cholesteryl esters had increased compared to controls (Table 2). Plasma apoA-I concentrations increased from 9954 mg , dl-' to 114k6 mg . dl-' (mean? SEM; P < 0.05) in control and transgenic mice, respectively. ApoA-I1 plasma levels were 54 2 2 mg . dl-' and 4 3 k 2 mg . dl-' in transgenic and control mice, respectively (P < 0.0001).

Distribution of human LCAT in plasma and HDL size. LCAT distribution in plasma was analyzed by gel filtration chromatog- raphy followed by determination of the LCAT activity in eluted fractions (Fig. 3). Essentially all LCAT activity was found in the HDL fraction, indicating that LCAT is primarily associated with HDL. The molecular mass of HDL isolated from transgenic mice was higher (mean 275 kDa) than that of controls (mean 222 kDa) as determined using non-denaturing gradient gel elec- trophoresis (Fig. 4), corresponding to an increase of 24%. The band of about 80 kDa did not bind anti-apoA-I or anti-apoA-11. Its nature remains unknown.

Measurements of HDL particles in electron micrographs confirmed that the transgenic particles were larger than controls (15.2 t 2.9 nm versus 12.6 ? 0.7 nm; m e a n t SD, i.e. their diam- eter was 2 nm wider, corresponding to about 35% volume increase.

Cholesterol efflux. Having demonstrated that all detectable LCAT activity was connected with the HDL fraction (Fig. 3), we wanted to measure the cholesterol-efflux-stimulating potential of HDL from transgenic LCAT mice. Fu5AH hepatoma cells were incubated for different periods of time with 50 pg . ml-' of HDL

, 35 I 50 7

+Cholesterol - LCAT 8Ctlvlty

+Cholesterol - LCAT 8Ctlvlty

20 25 30 35 40 45 50

Time (min) Fig. 3. Plasma LCAT distribution in human LCAT transgenic mice. Plasma lipoproteins were separated by gel filtration chromatography using a Superose 6 HR 10/30 column at a flow rate of 0.4 ml . min-' and monitored at 280 nm. LCAT activity (4-) and cholesterol concen- trations (0) were measured in the eluted fractions as described under Materials and Methods.

isolated from transgenic or control mouse serum. Although HDL from both transgenic and control mice promoted cholesterol ef- flux from cells, such promotion by transgenic mouse HDL was significantly higher after both 1 and 2 h of incubation (Fig. 5). Apart from the increase in LCAT activity and in the amount of cholesteryl ester, the relative composition of the HDL particles was essentially identical. The effect of the HDL fraction from the transgenic mice therefore suggests that the presence of LCAT is closely linked to cholesterol removal from the plasma membrane.

Fatty acid composition of plasma phospholipids, triacylglyc- erols and cholesteryl esters. The concentration as well as the distribution of fatty acids in plasma phospholipids were similar

572 Mehlum et al. ( E m J. Biochern. 230)

Table 3. Fatty acid distribution profiles of plasma triacylglycerols and cholesteryl esters in three control and three transgenic mice. Each value is expressed as a percentage of total fatty acids in triacylglycerols and cholesteryl esters respectively; n.d. = not detectable.

Fatty acid Amount in

triacy lgl ycerols

controls transgenic

cholesteryl esters

controls transgenic

%

Myristic (14:O) Palmitic (1 6 : 0) Palmitoleic (16: 1) Stearic (18:O) Oleic [ 18 : 1 (9)] Linoleic [18:2(9, 12)] y-Linoleic [18:3(6, 9, 12)] Arachidic (20 : 0) Icosanoic [20: 1(14)] Arachidonic [20 : 4(5, 8, 11, 14)] Icosapentanoic [20:5(5, 8, 11, 14, 17)] Docosahexanoic [22:6(4, 7, 10, 13, 16, 19)l

Saturated Monosaturated Polyunsaturated

0.6 i 0.5 29.4 i 3.6 2.3 2 2.9 2.9 i 0.3

36.3 i 2.9 24.2 i 1.5 0.3 0.1 2 0.1 0.5 ? 0.2 0.7 i 0.1 1.0 i 0.4 1.7 i 0.3

32.9 i 4.0 39.1 t 2.7 28.0 i 1.6

0.1 i 0.6 40.5 i 1.5 2.9 t 0.9 5.7 i 0.7

28.2 i 2.3 20.3 i 1.7 n.d. 0.1 i 0.2 0.1 ? 0.2 0.2 i 0.3 0.7 i 0.9 0.7 i 1.2

47.0 i 1.6 31.2 i 1.9 21.8 i 0.4

0.9 i 0.4 10.8 t 1.8 4.6 i 0.5 1.3 i 0.2

23.1 i 4.1 43.0 i 2.5 0.1 i 0.2 0.2 i 0.4

n.d. 14.1 i 1.8 0.6 i 0.5 1.3 t 0.7

13.2 i 2.3 27.7 i 3.7 59.1 i 4.3

0.7 i 0.1 16.7 i 1.4 5.2 2 1.3 1.2 ? 0.2

18.7 i 1.9 43.3 i 4.4 n.d. 0.1 t 0.1

n.d. 7.9 i 1.2 3.1 i 3.9 3.0 t 3.6

18.7 i 1.6 23.9 i 3.2 57.4 i 4.8

669 - 440 - 232 - 140 -

67 -

1 2 3 4 Fig. 4. Overexpression of human LCAT in transgenic mice increases plasma HDL size. The density interval 1.063-1.21 g . ml-l was used for isolation of HDL by ultracentrifugation. HDL (20 pg) from transgenic and control mice were applied to a non-denaturing 8-25% gradient gel electrophoresis (see Materials and Methods). Lane 1, molec- ular mass maskers; lane 2, human HDL; lane 3, control mouse HDL; lane 4, transgenic mouse HDL.

in transgenic and in control mice, both containing a high propor- tion of saturated fatty acids (data not shown). The distribution of plasma fatty acids in triacylglycerols and cholesteryl esters was, however, clearly different when transgenic and control mice were compared. In transgenic mice, triacylglycerols con- tained significantly increased proportions of palmitic acid, whereas oleic acid and linoleic acid were decreased (Table 3). This was reflected by a higher proportion of saturated and a lower proportion of monounsaturated and polyunsaturated fatty acids in triacylglycerols from transgenic mice. The percentage of palmitic acid was also increased in plasma cholesteryl esters from transgenic mice, whereas the percentage of oleic acid was decreased. The percentage of linoleic acid was similar in transgenic and in control mice. Most interestingly, in transgenic mice the percentage of arachidonic acid was almost half of that in controls, whereas the alterations in icosapentanoic and doco- sahexanoic acids were not significant.

TIME (hours)

Fig. 5. Cholesterol efflux from Fu5AH cells. Cells loaded with choles- terol were incubated for 1 or 2 h with 50 pg . ml-' of HDL isolated from transgenic (-O-) and control (-0.) mouse serum by gel filtration chromatography (see Materials and Methods). The percentage of choles- terol efflux was calculated by measuring radioactivity in medium and cells. The data are the means 2 SD of three experiments, each in tripli- cate.

Apolipoprotein and lipase gene expression and activity in transgenic LCAT mice. The metabolic fate of plasma lipids and lipoproteins depends largely upon the interaction of their protein constituents, the apolipoproteins, with various cellular receptors and enzymes. Therefore, the changes in plasma lipids and lipo- proteins observed in transgenic mice were investigated to dis- cover whether they could be related to differences in the expres- sion of the different apolipoproteins and lipoprotein and hepatic lipases. Liver mRNA levels of two of the major HDL apolipo- proteins, apoA-I and apoA-11, were similar between control and transgenic mice. In contrast to the liver, intestinal apoA-I gene expression was decreased by 5 5 % in transgenic animals. ApoA- IV mRNA decreased to 52% in the liver and to 78% in the intestine. Interestingly, liver and intestinal apoC-111 mRNA levels decreased by 25% and 28%, respectively. No significant changes in apoB, apoE and lipoprotein lipase mRNA levels were observed between transgenic and control mice.

Mehlum et al. (Em J. Biochern. 230) 573

Lipoprotein lipase activity in plasma from heparinized male transgenic mice was reduced by 12% (n = 10) but this differ- ence was not significant. Hepatic triacylglycerol lipase activity in plasma increased on average 57% (control 75.1 mU . ml-I, transgenic 117.6 mU . I&', n = 16, P<O.O5). In liver triacyl- glycerol lipase mRNA showed a clear increase, exactly corre- sponding to the plasma enzyme activity increase.

DISCUSSION

The transgenic mice described in this paper harbour the en- tire human LCAT gene including 1.9 kb upstream and 0.9 kb downstream sequences. These mice express human LCAT mRNA essentially correctly with regard to tissue specificity (Warden et al., 1989). Recent data (Skretting et al., unpublished) suggest that regulation of mRNA degradation is of major impor- tance for the tissue-specific expression of LCAT protein. Al- though the exact cis-acting elements required for this and other kinds of regulation contributing to the liver-specific expression are not yet clearly identified, they must be contained within the injected fragment. Furthermore, the level of expression of hu- man LCAT mRNA in the liver, as well as the plasma LCAT activity, correspond well to the number of gene copies intro- duced in the transgenic mice. This suggests that locus control elements may be present in the injected fragment. Alternatively, the correct expression of the transgene in this line may be due to a fortunate position effect.

Plasma LCAT activity (which is measured under zero-order conditions using an exogenous substrate containing human apoA-I as a co-factor and therefore reflects the level of plasma LCAT protein) is approximately 40-fold increased in the transgenic mice. The plasma cholesterol esterification rate, which is measured using endogenous substrate, is 2-fold lower in fasting transgenic mice compared to fasting controls. This result could be the consequence of a combination of factors. The plasma cholesterol esterification of endogenous substrate not only depends on plasma LCAT, but also on the activation poten- tial of plasma co-factors for LCAT, and on substrate availability (Fielding et al., 1972; Soutar et al., 1975). The changes observed in plasma lipid and lipoprotein concentrations, as well as in the plasma cholesteryl ester fatty acid profiles in transgenic mice, demonstrate that human LCAT is indeed activated by mouse apoA-I. A more likely explanation for the low esterification is then that LCAT activity is subject to substrate inhibition (Dobia- sova et al., 1983) or lack of available free cholesterol in a sui- table compartment. Human LCAT transgenic mice HDL par- ticles have an increased size and a higher HDL cholesteryl ester content, factors which may contribute to the observed reduced cholesteryl esterification rate in these mice. Large HDL particles (e.g. subfraction 2) are competitive inhibitors of the LCAT reac- tion (Barter et al., 1984). Finally, in humans, the cholesteryl es- ters produced are rapidly transferred to VLDL and LDL through the action of cholesteryl ester transfer protein. In mice, however, this transfer does not occur due to a lack of activity of this pro- tein in plasma and saturation of HDL with cholesteryl esters is therefore reached more rapidly.

The expression of human LCAT results in major changes in the fatty acid distribution profile of plasma cholesteryl esters, the product of plasma LCAT action. Indeed, cholesteryl esters in transgenic mouse plasma contain a higher proportion of satu- rated fatty acids (e.g. palmitic acid), whereas unsaturated fatty acids (e.g. oleic and arachidonic acid) are decreased. Substrate specificity of human and mouse LCAT has been shown to differ due to a larger active site of mouse LCAT, which allows access of more bulky fatty acids such as arachidonate (Grove and Po-

wall, 1991). The fatty acid distribution profile of cholesteryl es- ters in human LCAT transgenic mice therefore resembles more closely the human situation. This indicates furthermore that hu- man LCAT is active in mouse plasma and is the basis of the observed changes in fatty acid profiles.

Human LCAT transgenic mice exhibit major changes in plasma lipoprotein profiles and concentrations. As expected HDL cholesterol increases ; this is accompanied by an increased HDL particle size and ratio of cholesteryl estedfree cholesterol. Moreover, HDL from transgenic mice more efficiently promotes cellular cholesterol efflux, which is the first step in the removal of excess cholesterol from peripheral tissues. Most surprisingly, plasma triacylglycerols decrease in all lipoprotein fractions. Sev- eral mechanisms can be invoked to explain the large effect of human LCAT on plasma triacylglycerols. First, it is conceivable that the high levels of LCAT protein interfere sterically in plasma triacylglycerol metabolism. However, this seems un- likely since plasma LCAT activity is exclusively present in HDL but not in VLDL or LDL. Second, the effects of LCAT overex- pression on triacylglycerol metabolism may be indirect and me- diated by secondary effects on the expression of several key proteins implicated in triacylglycerol metabolism. Although pe- ripheral tissue lipoprotein lipase activity did not differ signifi- cantly between transgenic and control mice, the hepatic triacyl- glycerol lipase activity was clearly increased. This increased ac- tivity may contribute to an accelerated catabolism of plasma tria- cylglycerols, since the proportionally largest decrease in triacyl- glycerols is observed in the HDL fraction, the preferred lipopro- tein substrate for hepatic triacylglycerol lipase (Jansen et al., 1980). Furthermore, plasma triacylglycerols contain higher pro- portions of saturated fatty acids and lower amounts of both mo- nounsaturated and polyunsaturated fatty acids. This profile may reflect an increased hepatic triacylglycerol lipase activity, since triacylglycerols containing polyunsaturated fatty acids are largely protected from the action of lipoprotein lipase but not hepatic triacylglycerol lipase (Melin et al., 3 991). In addition to altered lipase activities, the decrease in liver and intestinal apoC- 111 expression in LCAT transgenic mice may contribute to the lowered plasma triacylglycerol concentrations. Clinical and transgenic animal studies have indicated that plasma triacylglyc- erol concentrations are proportional to plasma apoC-I11 concen- trations and liver apoC-I11 gene expression (Curry et al., 1980; Schonfeld et al., 1979; Stocks et al., 1979; Le et al., 1988; Ito et al., 1990). ApoC-I11 appeared to impair the clearance of triacylglycerol-rich lipoproteins due to interference with apoE- mediated cellular uptake of such particles. Finally, hepatic and/ or intestinal secretion of triacylglycerol-rich lipoproteins may be lowered in LCAT transgenic mice, which may be correlated to a decreased liver and/or intestinal synthesis of apoA-IV and apoA-I.

These major changes in the expression of lipid-related genes in mice overexpressing human LCAT are remarkable. On the one hand, intestinal apoA-I and liver apoA-IV and apoC-111 gene expression is decreased in LCAT transgenic mice. Previous studies have shown that hepatic apoA-I expression is relatively refractory to regulation. On the other hand, hepatic triacylglyc- erol lipase mRNA levels are increased in the transgenic animals. The exact mechanism behind these changes is unclear at present. However, it is tempting to speculate that hepatic triacylglycerol lipase expression is regulated by changes in intracellular choles- terol metabolism, as has been shown previously in the human hepatoma cell line, Hep G2 (Busch et al., 1990), and hence these changes in its gene expression may be secondary to the alter- ations in plasma lipids due to the increased LCAT action in these mice.

574 Mehlum et al. (Eur. J. Biochem. 230)

The changes in plasma lipid and lipoprotein profiles in LCAT transgenic mice result in a less atherogenic plasma lipo- protein profile. First, overexpression of human LCAT provokes a decrease in the plasma concentrations of the atherogenic apoB- containing lipoproteins. Second, HDL cholesterol increases, which is accompanied by an increased HDL particle size and higher amounts of plasma apoA-I. Third, HDL from transgenic LCAT mice has a higher cholesterol-efflux-promoting activity compared to non-transgenic HDL. It remains to be seen whether these favourable changes protect such mice against the develop- ment of atherosclerosis-like lipid lesions inducible in non- transgenic C57B1/6 mice by feeding an atherogenic diet and whether increased LCAT expression leads to longevity. Experi- ments to answer these questions have been started.

In conclusion, the results of our study show that the 7-kb genomic DNA fragment used to generate transgenic mice con- tains the cis-acting elements necessary and sufficient for tissue- specific expression. The product of the human LCAT gene ap- pears to be active in a murine environment. The transgenic LCAT mice have significantly lower triacylglycerol and VLDL and LDL cholesterol and increased HDL cholesterol, resulting in what may be a less atherogenic plasma lipoprotein profile. Our observations suggest new roles for the LCAT enzyme in lipid metabolism. Additional studies are required to analyse this matter further.

We are grateful to Dr Tor Arne Hagve and Vigdis Bjerkeli for initial analyses of lipids and lipases, respectively, Dr GTO Smistad for electron microscopy, Dr Jan P. Blomhoff for helpful discussion, Dr Peter Hjorth for advice, Dr George Rothblat for the gift of FuSAH hepatoma cells and to Dr Patrice Denkfle for his support. The excellent technical assis- tance of Annie Tbaikhi and Brigitte Lacroix, and the secretarial assis- tance of Ingunn Schult, is gratefully acknowledged. This work was sup- ported by grants from The Research Council of Norway and The Norwe- gian Council for Cardiovascular Diseases to Hans Prydz and by grants from Fondation pour la Recherche Medical and Merck, Sharpe and Dohme to Johan Auwerx.

REFERENCES Assman, G., von Eckardstein, A. & Funke, A. (1991) Lecithin:choles-

terol acyltransferase deficiency and fish-eye disease, Cum Opin. Lip- idol. 2, 110-117.

Barter, P. J., Hopkins, G. J., Gorjatschko, L. & Jones, J. M. E. (1984) Competitive inhibition of plasma cholesterol esterification by human high-density lipoprotein-snbfraction 2, Biochim. Biophys. Acta 793,

Bengtsson-Olivecrona, G. & Olivecrona, T. (1992) in Lipoprotein analy- sis. A practical approach (Converse, C. & Skinner, E. R., edsj pp. 169-185, IRL Press, Oxford.

Breslow, J. L. (1994) Insights into lipoprotein metabolism from study in transgenic mice, Annu. Rev. Physiol. 56, 797-810.

Brousseau, T., Clavey, V., Bard, J. M. & Fruchart, J. C. (1994) Sequential ultracentrifugation micromethod for separation of serum lipoproteins and assay of lipids, apolipoproteins and lipoprotein particles, Clin. Chem. 39, 960-964.

Busch, S. J., Barnhart, R. L., Martin, G. A,, Flanagan, M. A. &Jackson, R. L. (1990) Differential regulation of hepatic triglyceride lipase and 3-hydroxy-methylglutaryl-CoA reductase gene expression in a hu- man hepatoma cell line, HepG2, J. Biol. Chem. 265, 22474-22479.

Carlson, L. A. & Philipson, B. (1979) Fish eye disease. A new familiar condition with massive corneal opacities and dyslipoproteinaemia, Lancet 2 , 921 -923.

Carlson, L. A. (1982) Fish eye disease: a new familiar condition with massive corneal opacities and dyslipoproteinaemia, Eur. J. Clin. In- vest. 12, 41-53.

Chen, C. & Albers, J. J. (1982) Characterization of proteoliposomers containing apoprotein A-I: a new substrate for the measurements of 1ecithin:cholesterol acyltransferase activity, J. Lipid Res. 23, 680- 691.

152- 160.

Chomczynski, P. & Sacchi, N. (1987) Single-step method for RNA isola- tion by acid guanidiniurn thiocyanate-phenol-chloroform extraction, Anal, Biochem. 162, 156-159.

Curry, N. D., McConathy, W. J., Fesmire, J. D. & Alaupovic, P. (1980) Quantitative determination of human apolipoprotein C-I11 by electroimmunoassay, Biochirn. Biophys. Acta 617, 503 -513.

Dobiasova, M. (1983) Lecithin :cholesterol acyltransferase and the regu- lation of endogenous cholesterol transport, Adv. Lipid Res. 20, 107- 194.

Duverger, N., Rader, D., Duchateau, P., Fruchart, J. C., Castro, G. & Brewer, H. B. Jr (1993) Biochemical characterization of the three major subclasses of lipoprotein A-I preparatively isolated from hu- man plasma, Biochemistry 32, 12372-12379.

Feinberg, A. P. & Vogelstein, B. (1983) A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity, Anal. Biochem. 132, 6-13.

Fielding, C. J., Shore, V. G. & Fielding, P. E. (1972) A protein cofactor of lecithin : cholesterol acyltransferase, Biochem. Biophys. Res. Com- mun. 46, 1493-1498.

Folch, J., Lees, M. & Sloane Stanley, G. H. (1957) A simple method for the isolation and purification of total lipids from animal tissues, J. Biol. Chem. 226, 497-509.

Forte, T. M. & Nordhausen, R. W. (1986) Electron microscopy of nega- tively stained lipoproteins, Methods Enzymol. 128, 442-457.

Foumey, R. M., Miyakoshi, J., Rufus, S. D. I11 & Paterson, M. C. (1988) Northern blotting: Efficient RNA staining and transfer, Focus 10, 5-7.

Funke, H., von Eckardstein, A., Pritchard, P. H., Albers, J. J., Kastelein, J. J. P., Droste, C. & Assman, G. (1991 j A molecular defect causing fish eye disease: An amino acid exchange in lecithin-cholesterol acyltransferase (LCAT) leads to the selective loss of a-LCAT activ- ity, Proc. Natl Acad. Sci. USA 88, 4855-4859.

Funke, H., von Eckardstein, A., Pritchard, P. H., Homby, A. E., Wie- busch, H., Corradino, M., Hayden, M. R., Dachet, C., Jacotot, B., Gerdes, U., Faergeman, O., Albers, J. J., Colleoni, N., Catapano, A., Frohlich, J. & Assman, G. (1993) Genetic and phenotypic hetero- geneity in familiar 1ecithin:cholesterol acyltransferase (LCAT) defi- ciency, J. Clin. Invest. 92, 479-485.

George, J., Blakesley, R. W. & Chirikjian, J. G. (1980) Sequence-specific endonuclease BamHT. Effect of hydrophobic reagents on sequence recognition and catalysis, J. Biol. Chem. 255, 6521 -6524.

Gjone, E. & Norum, K. R. (1968) Familiar serum cholesterol ester defi- ciency. Clinical study of a patient with a new syndrome, Acta Med. Scand. 180, 107 - 11 2.

Glomset, J. (1972) in Blood lipids and lipoproteins (Nelson, G., ed.) pp. 745-787, Wiley, New York.

Grove, D. & Powall, H. J. (1991) Comparative specificity of plasma lecithin :cholesterol acyltransferase from ten animal species, Lipids 26,416-420.

Haddad, I. A., Ordovas, J. M., Fitzpatrick, T. & Karathanasis, S. K. (1986) Linkage, evolution, and expression of the rat apolipoprotein A-11, C-111, and A-IV genes, J. Biol. Chem. 261, 13268-13277.

Hedrick, C. C., Castellani, L. W., Warden, C. H., Puppione, D. L. & Lusis, A. J. (1993) Influence of mouse apolipoprotein A-I1 on plasma lipoproteins in transgenic mice, J. Biol. Chem. 268, 20676-20682.

Hogan, B., Constantini, F. & Lacy, E. (1986) in Manipulating the mouse embryo: a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Ito, Y., Azrolan, N., O’Connell, A., Walsh, A. & Breslow, J. L. (1990) Hypertriglyceridemia as a result of buman apo CIII gene expression in transgenic mice, Science 249, 790-793.

Jakobsen, K. S., Breivold, E. & Homes, E. (1990) Purification of rnRNA directly from crude plant tissues in 15 minutes using magnetic oli- godT microspheres, Nucleic Acids Res. 18, 3669.

Jansen, H. A., van Tol, A. & Hulsmann, W. C. (1980) On the metabolic function of heparin-releasable liver lipase, Biochem. Biophys. Res. Commun. 92, 53-59.

Klein, H.-G., Lohse, P., Pritchard, P. H., Bojanovski, D., Schmit, H. & Brewer, H. B. Jr (1992) Two different alleic mutations in the lecithin- cholesterol acyltransferase gene associated with the fish eye syn- drome, J. Clin. Invest. 89, 499-506.

Klein, H.-G., Lohse, P., Duverger, N., Albers, J. J., Rader, D. J., Zech, L. A., Santamarina-Fojo, S. & Brewer, H. B. Jr (1993a) Two dif-

Mehlum et al. (Eul: J. Biochem. 230) 575

ferent alleic mutations in the 1ecithin:cholesterol acyltransferase (LCAT) gene resulting in classical LCAT deficiency: LCAT (tyrX3+ stop) and LCAT (tyrtS6-+asn), J. Lipid. Res. 34, 49-58.

Klein, H.-G., Santamarina-Fojo, S . , Duverger, N., Clerc, M., Dumon, M.-F., Albers, J. J., Marcovina, S . & Brewer, H. B. Jr (1993b) Fish eye syndrome: A molecular defect in the lecithin-cholesterol acyltransferase (LCAT) gene associated with normal a-LCAT-spe- cific activity, J. Clin. Invest. 92, 479-485.

Laemmli, U. K. (1970) Cleavage of structural proteins during the assem- bly of the head of bacteriophage T4, Nature 227, 680.

Laird, P. W., Zijderveld, A., Linders, K., Rudnicki, M. A., Jaenisch, R. & Berns, A. (1991) Simplified mammalian DNA isolation procedure, Nucleic Acids Res. 19, 4293.

Larsen, F., Solheim, J., Kristensen, T., Kolstg, A,-B. & Prydz, H. (1993) A tight cluster of five unrelated human genes on chromosome 16q22.1, Hum. Mol. Genet. 2, 1589-1595.

Le, N.-A,, Gibson, J. C. & Ginsberg, H. N. (1988) Independent regula- tion of plasma apoprotein C-I1 and C-Ill concentrations in very low- density and high-density lipoproteins : implications for the regulation of the catabolism of these lipoproteins, J. Lipid Res. 29, 669-677.

Liu, M.-S., Jirik, F. R., LeBoeuf, R. C., Henderson, H., Castellani, L. W., Luisis, A. J., Ma, Y., Forsythe, I. J., Zhang, H., Kirk, E., Brunzell, J. D. & Hayden, M. R. (1994) Alteration of lipid profiles in plasma of transgenic mice expressing lipoprotein lipase, J. Biol. Chem. 269,

Lowry, 0. H., Rosebrough, N. J., Fan, A. L. & Randall, R. J. (1951) Protein measurement with the Folin phenol reagent, J. Biol. Chem.

McLean, J., Fielding, C., Drayna, D., Dieplinger, H., Baer, B., Khor, W., Menzel, W. & Lawn, R. (1986a) Cloning and expression of human lecithin-cholesterol acyltransferase cDNA, Proc. Nut1 Acad. Sci. USA 82, 2335-2339.

McLean, J., Wion, K., Drayna, D., Fielding, C. & Lawn, R. (1986b) Human lecithin-cholesterol acyltransferase gene : complete gene se- quence and site of expression, Nucleic Acids Res. 14, 9397-9406.

Mehlum, A. & Prydz, H. (1994) Mice transgenic for the human LCAT gene, Clin. Genet. 46, 92.

Melin, T., Qi, C., Bengtsson-Olivecrona, G., Akesson, B. & Nilsson, A. (1991) Hydrolysis of chylomicron polyenic fatty acid esters with lipoprotein lipase and hepatic lipase, Biochia. Biophys. Acta 1075,

Moya, M., Atger, V., Paul, J. L., Foumier, N., Moatti, N., Giral, P., Friday, K. E. & Rothblat, G. (1994) A cell culture system for screen- ing human serum for ability to promote cellular cholesterol efflux, Arteriosclel: Thromb. 14, 1056- 1065.

Norum, K. R., Gjone, E. & Glomset, J. A. (1989) in The metabolic basis of inherited disease (Scriver, C. R., Beaudet, A. L., Sly, W. S . & Valle, D., eds) pp. 1181-1194, McGraw-Hill, New York.

Paigen, B., Ishida, B., Verstuyft, J., Winters, R. & Albee, D. A. (1990) Atherosclerosis susceptibility differences among progenitor of re- combinant inbred strains of mice, Atherosclerosis 10, 316-323.

Peterson, J., Bengtsson-Olivecrona, G. & Olivecrona, T. (1986) Mouse preheparin plasma contains high levels of hepatic lipase with low affinity for heparin, Biochim. Biophys. Acta 878, 65-70.

11 417-11 424.

193, 265-275.

259-266.

Rogne, S . , Skretting, G., Larsen, F., Myklebost, O., Mevig, B., Carlson, L. A,, Holmquist, L., Gjone, E. & Prydz, H. (1987) The isolation and characterisation of a cDNA clone for human lecithin: cholesterol acyltransferase and its use to analyse the genes in patients with LCAT deficiency and fish eye disease, Biochem. Biophys. Res. Com- mun. 148, 161-169.

Schonfeld, G., Georges, P. I., Miller, J., Reilly, P. & Witztum, J. (1979) Apolipoprotein C-I1 and C-111 levels in hypolipoproteinemia, Metab- olism 28, 1001 -1009.

Shimada, M., Shimano, H., Gotoda, T., Yamamoto, K., Kawamura, M., Inaba, T., Yazaki, Y. & Yamada, N. (1993) Overexpression of human lipoprotein lipase in transgenic mice, J. Biul. Chem. 268, 17924- 17929.

Skretting, G., Blomhoff, J. B., Solheim, J. & Prydz, H. (1992) The ge- netic defect of the original Norwegian 1ecithin:cholesterol acyltrans- ferase deficiency families, FEES Lett. 309, 307-310.

Skretting, G. & Prydz, H. (1992) An amino acid exchange in exon I of the human lecithin :cholesterol acyltransferase (LCAT) gene is asso- ciated with fish eye disease, Biochem. Biophys. Res. Commun. 182,

Smith, K. M., Lawn, R. M. & Wilcox, J. N. (1990) Cellular localization of apolipoprotein D and lecithin: cholesterol acyltransferase mRNA in rhesus monkey tissues by in situ hybridization, J. Lipid Res. 31,

Soutar, A. K., Gamer, C. W., Baker, H. N., Sparrow, J. T., Jackson, R. L., Gotto, A. M. & Smith, L. C. (1975) Effect of the human plasma apolipoproteins and phosphatidylcholine acyl donor on the activity of lecithin :cholesterol acyltransferase, Biochemistry 14, 3057 - 3063.

Staels, B., Auwerx, J., Chan, L., van Tol, A,, Rosseneu, M. & Verhoeven, G. (1989) Influence of development, estrogens, and food intake on apolipoprotein A-I, A-11, and E mRNA in rat liver and intestine, J. Lipid Res. 30, 1137-1145.

Stewart-Phillips, J. L. & Lough, J. (1991) Pathology of atherosclerosis in cholesterol-fed, susceptible mice, Atherosclerosis 90, 21 1-218.

Stocks, J., Holdsworth, G. & Galton, D. J. (1979) Hypertriglyceridaemia associated with an abnormal triglyceride-rich lipoprotein carrying excess apolipoprotein C-11-2, Lancet ii, 667-671.

Stokke, K. T. & Norum, K. R. (1971) Determination of 1ecithin:choles- terol acyltransferase in human blood plasma, Scand. J. Clin. Lab. Invest. 27, 21 -27.

Voss, H., Wiemann, S . , Wirkner, U., Schwager, C., Zimmermann, J., Stegemann, J., Erfle, H., Hewitt, N. H., Rupp, T. & Ansorge, W. (1992) Automated DNA sequencing system resolving 1000 bases with fluorescein -15-*dATP as internal label, Methods Mol. Cell. Biol. 3, 153-155.

Warden, C. H., Langner, C. A,, Gordon, J. I., Taylor, B. A,, McLean, J. W. & Lusis, J. L. (1989) Tissue-specific expression, developmental regulation and chromosomal mapping of the 1ecithin:cholesterol acyltransferase gene, J. Biol. Chem. 264, 21 573-21 581.

583-587.

995-1003.