Distribution of the lipolysis stimulated receptor in adult and embryonic murine tissues and lethality of LSR–/– embryos at 12.5 to 14.5 days of gestation Samir Mesli 1 , Sandrine Javorschi 1, †, Annie M. Be ´ rard 1 , Marc Landry 2 , Helen Priddle 3 , David Kivlichan 3 , Andrew J. H. Smith 3 , Frances T. Yen 4 , Bernard E. Bihain 4 and Michel Darmon 1 1 Laboratoire de Biochimie et de Biologie Mole ´ culaire, Universite ´ Victor Se ´ galen Bordeaux 2, France; 2 INSERM E358, Universite ´ Victor Se ´ galen Bordeaux 2, France; 3 Gene Targeting Laboratory; Center for Genome Research, University of Edinburgh, Scotland; 4 Laboratoire Me ´ decine et The ´ rapeutique Mole ´ culaire, Vandoeuvre-les-Nancy, France The lipolysis stimulated receptor (LSR) recognizes apo- lipoprotein B/E-containing lipoproteins in the presence of free fatty acids, and is thought to be involved in the clearance of triglyceride-rich lipoproteins (TRL). The distribution of LSR in mice was studied by Northern blots, quantitative PCR and immunofluorescence. In the adult, LSR mRNA was detectable in all t issues teste d exce pt muscle and h eart, and was abundant in liver, lung, intestine, kidney, ovaries and testes. During embryogenesis, LSR mRNA was detectable at 7.5 days post-coitum (E7) and increased up to E17 in parallel to p rothrombin, a liver marker. In adult liver, immunofluorescence experiments s howed a s taining at the periphery of hepatocytes as well as in fetal liver at E12 and E15. These results are i n agreement with the assumption that LSR is a plasma membrane receptor involved in the clear- ance of lipoproteins by liver, and suggest a possible r ole in steroidogenic organs, lung, i ntestine and kidney). To explore the role of LSR in vivo,theLSR gene was inactivated in 129/ Ola ES cells by removing a gene segment containing exons 2–5, and 129/Ola-C57BL/6 m ice b earing the deletion were produced. Although heterozygotes appeared normal, LSR homozygotes were not viable, with the exception of three males, while the total progeny of genotyped wild-type and heterozygote pups was 345. Mortality of the homozygote embryos was observed between days 12.5 and 15.5 o f ges- tation, a time at which their liver was much smaller than that of their littermates, indicating that the expression of LSR is critical for liver and embryonic development. Keywords: lipoprotein receptors; Northern-blot; quantita- tive PCR; immunofluorescence; gene-targetting. Lipids, absorbed exogenously by the intestine and synthe- sized endogenously by the liver, are secreted into the circulation as lipoproteins for their transport to tissues, where they are used mainly for membrane synthesis, steroidogenesis and fat storage. Dietary cholesterol, phos- pholipids, triglycerides (TG) and fat-soluble vitamins absorbed by the intestine after a meal are transported by chylomicrons into lymph, then into blood. Lipoprotein lipase (LPL), anchored to the surface of capillary endothe- lium, hydrolyzes TG of chylomicrons into free fatty acids (FFA) that are taken u p by the underlyi ng muscle and adipose tissues. Chylomicron remnants are then taken up by the liver [1]. Transport of lipids to tissues is achieved by very low density lipoproteins (VLDL) and low density lipo- proteins (LDL). Excess cholesterol is removed from the peripheral cells b y high density lipoproteins ( HDL) that are able to return it to the liver for excretion via the LDL receptor (LDLR) or the scavenger receptor class BI (SR-BI) path- ways. In the same way, HDL are also involved in the delivery of cholesterol to certain tissues, mainly steroidogenic organs. Apolipoprotein (apo) B and E containing-VLDL and chylomicron remnants bind with high affinity to the LDLR and the LDL receptor related protein (LRP) that mediates endocytosis of both particles. However, another p lasma membrane lipoprotein receptor genetically distinct from the LDLR an d L RP, c alled t he lipolysis-stimulated receptor (LSR) may also be involved in the clearance of TRL [2,3]. The LSR was originally identified by its ability to bind LDL in the presence of FFAs [4]. LSR polypeptides (85 and 115 k Da) were identified by ligand b lotting in the presence of oleate in fibroblasts isolated from a patient with familial hypercholesterolemia [2]. T hree bands of 90, 115, and 240 k Da were found when solubilized rat liver me mbrane proteins were used as a substrate [ 5]. W hen a ntibodies inhibiting LSR function were u sed for Western blotting, the Correspondence to Y. M. Darmon, Universite ´ Victor Se ´ galen Bor- deaux 2, Laboratoire de Biochimie et de Bi ologie Mole ´ culaire, Zone Nord – Case 49–146, Rue Le ´ o Saignat, 33076 Bordeaux Cedex, France. Fax: + 33 5 5 7 57 1397, Tel.: + 33 5 57 57 15 79. E-mail: darmon@u-bordeaux2.fr Abbreviations: apo, apolipoprotein; FFA, free fatty acids; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HDL, high density lipo- proteins; LDL, low density lipoproteins; LDLR, low density lipo- protein receptor; LRP, low density lipoprotein receptor related protein; LSR, lipolysis stimulated receptor; SR-BI, scavenger receptor BI; TG, triglycerides; TRL, triglyceride-rich lipoproteins; VLDL, very low density lipoproteins. Present address : Invitrogen C orp. 1610 Faraday Avenue, Carlsbad, CA 92008, U SA. (Received 1 April 2004, revised 6 May 2004, accepted 19 May 2004) Eur. J. Biochem. 271, 3103–3114 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04223.x same three bands were detected. Molecular cloning of the LSR allowed the authors to identify putative t ranslation products of 58.3, 63.8, and 65.8 kDa. The combination of various techniques suggested that the receptor was a multimer of subunits associated through disulfide bridges [5]. Several characteristics of LSR suggest that it might represent a significant element for the clearance of TRL: (a) LSR is able t o bind lipoproteins containing apoB and apoE; (b) LSR displays high affinity for T RL; (c) LSR b inding is inhibited by lactoferrin, receptor associated prote in ( RAP), and apoCIII, all reported to have a hyperlipemic effect in animals [2,3] [6,7]; (d) the apparent number of LSR binding sites expressed a t the surface of hepatocytes correlates negatively with plasma triglyceride levels measured in the postprandial stage [3]. The present work was u ndertaken to determine the distribution of LSR mRNA and protein in murine organs, and whether this distribution was c ompatible with the alleged role of this new receptor as a lipoprotein receptor. It was found that LSR was not only expressed in liver (adult and fetal), but also in steroidogenic organs (ovaries, testes, and adrenal glands), lung, intestine, kidney and brain. To explore further the role of LSR, the gene was inactivated in ES c ells and a strain of transgenic LSR knockout mice was established. However, from a total progeny of 345 mice derived from intercrossing LSR heterozygote (LSR+/–) animals, only three viable homozygote (LSR–/–)animals were obtained, so that a comprehensive description of their phenotypic defects was impossible to produce. Most LSR–/– mutants di e in utero between embryonic days 12.5 (E12.5) and E15.5. At E14.5, LSR–/– mutant mice livers were found to be much smaller than that of their littermates. Therefore, inactivation of LSR appears to be lethal at the embryonic stage, probably secondary to liver involution. Materials and methods Animals used for expression studies Normal C57Bl/6, 129/Sv, and M F1 mice were obtained from CERJ (Le G enest Saint-Isle, France). They w ere housed in a specific pathogen-free animal facility on a 12-h light : 12-h dark cycle, with free access to food and water. The research protocol was in accordance with French Ministry of Agriculture , section o f Health a nd Animal Protection (approval 04476). Northern blots Mouse embryo and adult multiple tissue Northern blots were performed with nylon membranes blotted to gels loaded with 2 lg mRNA per lane (Clontech, Saint-Quentin en Yvelines, France). They were prehybridized for 30 min at 68 °Cin Express Hyb TM hybridization solution ( Clontech) and then hybridized for 2 h at 68 °C with t he same solution supple- mented with the appropriate radiolabeled cDNA probes. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe (500 bp fragment) was prepared by No tIandEcoRI digestion o f the murin e cDNA inserted in PT7T3d plasmid (IMAGE clone 113843, UK HGMP Resource Centre, Cambridge, UK). The LSR probe (full length 2 kb insert) was prepared by EcoRI digestion o f the murine cDNA inserted in pGEMT-easy 5Zf(–) (a gift from Genset, La Jolla, CA, USA). Probes were labeled by decanucleo tide-mediated incorporation of [ 32 P]dCTP[aP] (Ambion, Montrouge, France). Blots were rinsed three times with 2· NaCl/Cit, 0.05% SDS at room temperature for 30 min and washed twice with 0.1· NaCl/Cit, 0.1% SDS at 50 °Cfor40min with agitation. Autoradiography was performed by expo- sure for 2 h in a PhosphorImager (Molecular Dynamics, Amersham–Pharmacia–Biotech, Orsay, France). Real-time RT-PCR Mouse tissues were pooled from 4 to 5 mice on a standard diet. Samples were immediately put into Trizol (Gibco BRL, Cergy-Pontoise, France)andstoredat)80 °C pending RNA isolation. Total RNAs were isolated accord- ing to the manufacturer’s instructions. The amount of RNA was determined by measuring absorption at 260 nm. The quality o f the isolated RNA was controlled by the 260/ 280 n m ratio (1.8–2.0). cDNAs were obtained by reverse-transcription of 1 lg total RNAs prepared from C 57BL/6 mouse tissues. RNAs were first treated by RQ1 RNase-Free DNase (Promega, Charbonnie ` res, France). First strand cDNA synthesis was performedina20lL mixture using the GeneAmp RNA PCR kit (Applied Biosystems, Courtaboeuf, France). For some tissues, total cDNAs were also obtained from Clontech. Specific primers and TaqMan probes were designed u sing the PRIMER EXPRESS 1.0 software (Applied Biosystems) and synthesized by Genset (Paris, France). Each probe was double-labeled with the fluorescent reporter dye, 6-carb- oxyfluorescein ( FAM), c ovalently linked to the 5 ¢-end of the probe and t he quencher dye, 6-carboxytetramethylrhodam- ine(TAMRA),attachedtothe3¢-end. Quantitative PCR was performed in 96-well reaction plates with optical caps. Fluorescence was followed continuously for each r eaction. Real-time quantitative RT-PCR analyses w ere performed in an ABI P RISM 5700 sequence detection system i nstrument (Applied Biosystems). The reaction mixture contained an amount of cDNA corresponding to 100 ng of reverse- transcribed total RNA, 300 n M sense and antisense p rimers (except for GAPDH, 120 n M of each) and 200 n M probe in a final volume of 25 lL using the TaqMan PCR mix (Applied Biosystems). Relative quantitation of a given gene w as calculated after normalization to 18S ribosomal RNA amount for tissues from which RNAs were i solated ( liver, ovaries, adrenal g lands, testes, intestine, brain, muscle), or GAPDH amount for tissues for which total cDNA were purchased (liver, lung, kidney, heart). Individual C T values are means of duplicate measurements. Delta C T were converted to arbitrary values with the f ormula: arbitrary units ¼ 2 )dC T · 10 6 assuming an efficiency of amplification of 100%. Results are expressed as the mean of two experiments. The complete list of gene-specific primers and probes can be found in Table 1 . It must be noted that the quantitative P CR was d esigned to d etect the sum of all transcripts of LSR. Antibodies and immunocytochemistry The a nti-LSR Ig used for this study was a gift from Genset (La Jolla, CA, USA) [5]. The antiserum raised in New 3104 S. Mesli et al.(Eur. J. Biochem. 271) Ó FEBS 2004 Zealand rabbits was a ble to inhibit the in vitro binding of LDL to LSR preparations. In Western blots or immuno- precipitations it recognized the same bands that were identified by ligand blotting ( 90, 115 and 240 kDa) and by Western blotting. Negative controls were prepared by substituting the anti-LSR serum by nonimmunized or irrelevant rabbit sera. In some experiments, tissues from surviving male homozygous LSR knockout mice were included as negative controls. Adult mouse tissues and post implantation embryos (E12.5 and E15.5) were fixed in 4% paraformaldehyde overnight, freezed in Tissue-Tek (Labonord, Templemars, France) and se ctioned with a cryotome (10 lm sections) onto Superfrost coated slides (Labonord). Tissues were i ncubated for 30 min in phosphate-buffered saline (NaCl/P i ) containing 0 .5% bovine s erum albumin (BSA), washed 3 · with NaCl/P i and incubated in anti-LSR serum (1 : 10 in NaCl/P i /BSA) for 2 h at 37 °C. The sections were washe d in NaCl/P i and i ncubated w ith fluorescein- conjugated goat anti-rabbit IgG (Molecular Probes/Inter- chim, Montluc¸ on, France) 1 : 200 in NaCl/P i -BSA for 1 h at 37 °C. The sections were the n washed 3· in NaCl/P i before mounting in Prolong TM Antifade (Molecular Probes/Inter- chim). Slides were examined with a Leica photomicroscope using appropriate filter systems. P hotographs were taken on Kodak films (Amersham–Pharmacia–Biotech). Gene targeting of the LSR gene and generation of LSR deficient mice The murine (C57BL/6) LSR gene contains 10 coding exons with an open reading frame of 1782 nuclotide long encoding a peptide of 594 amino acids (F. T. Yen & B. E. Bihain, unpublished results). A129/Ola mouse genomic lambd a 2001 library was screened with a full length LSR cDNA probe to isolate cloned DNA for the targeting v ector construction. Several o verlapping phage clo nes, which together covered the most part of the gene, were isolated and inserts sequenced. This sequence (GenBank AY376636) contained the first eight exons and ends 19 bp before the end of exon 9 of the LSR gene; it lacks all of intron 9 and exon 10. Altogether this sequence lacks the portion coding for the last 17 amino a cids of LSR. A r eplacement targeting vector (map, Fig. 1). was designed subsequently to create a null allele by deletion of an internal region of the gene between the 5¢-end of exon 2 and the 3¢-end of exon 5 and its substitution with a reporter (b-galactosidase) and selection marker (neomycin resistance). This vector was comprised of left and right homology arms, which consisted of 2.65 kb of cloned genomic DNA sequence containing the 5¢- part of exon 2, intron 1, exon 1 and its 5¢-flanking noncoding sequence, and 2.7 kb of sequence containing the 3 ¢-part of exon 5 and 3¢-flanking intron 5 s equence, respectively. These were i nserted into a pBluescript plasmid, with the e xon 2 and 5 sequences joined via a BamHI linker. The reporter/selection cassette TAG3/IRES lacz/ SV40pA/MC1neo/pA [8] was inserted into the BamHI linker site. A MC1-tk dimer cassette [ 9] was appended to the end of the 5¢-homology arm at a SalI site for negative selection [10]. The vector was linearized with NotI, and E14TG2a embryonic stem c ells cultured according to standard conditions [11] were electroporated and selected in G418 and gancic lovir. R esistant ES cell clones were picked in to 96-well plates, and replica plated subsequently for freezing and DNA preparation. ES cell clone DNAs were screened by Southern blot analysis using HindIII digestion and hybridization with probes fl anking and external to the vector homology arms (Fig. 1). Clones targeted correctly at both 5¢-and3¢-sides were detected at a frequency of 12%. Targeted ES cells we re injected into C57BL/6 b lastocysts and t he resulting male chimeras subsequently test-crossed with C 57BL/6 females. Germline transmission from chimeras derived with t wo independ ent targeted clones was confirmed in agouti coat colored Table 1. Sequences of primers and probes used for real-time PCR with the TaqMan system. NC, sequences no t communicated by Perkin Elmer. mRNA Upstream primer (5¢fi3¢) Probe (5¢fi3¢) Downstream primer (5¢fi3¢) Amplicon size (bp) LSR atgcgtcctccctatgggtac tggagactttgacaggaccagctcagttg acctgggagctgtggcc 71 (exons 6–7) LDLR X64414 ctgtccccccaagacgtg caagtgcatctccccgcagtttgtgt ccatctaggcaatctcggtctc 102 (430–531of 4467) LRP AF074265 gtcccattggctttgagctc tcgaggagagcggatatcagacgcatatc gccacattgttgttgtttgtttc 124 (1926–2049 of 5521) SRB1 U37799 tgatgatgaccttggcgct caccatgggccagcgtgcttt gggaagcatgtctgggagg 131 (520–650 of 1785) ApoB M35186 cgtgggctccagcattcta ccaatggtcgggcactgctcaa tcatttctgcctttgcgtcc 65 (771–835 of 2354) ApoE D00466 attacctgcgctgggtgc tgaccaggtccaggaagagctgca gtcagttcttgtgtgacttgggag 79 (134–212 of 936 CDS) Apo A1 X64262 gacactctgggttcaaccgttagt ctgcaggaacggctgggccc ttcctctaggtccttgttcatctcc 126 (268–393 of 924) Prothrombin X52308 tacatagacgggcgcatcg agggctgggacgctgagaagggtat aaaaagcatcacctgccagg 72 (1084–1155 of 2031) Ubiquitin X51703 ggtggctattaattattcggctg attcccagtgggcagtgatggcattac gggcaagtggctagagtgca 75 (1010–1084 of 1172) GAPDH NC NC NC  190 18S rRNA NC NC NC  200 Ó FEBS 2004 LSR gene and protein expression (Eur. J. Biochem. 271) 3105 test-cross offspring by Southern blot analysis of DNA obtained from tail biopsy. Male and female test-cross offspring heterozygous for the null allele were intercrossed to o btain formal proof of the creation of the null allele and for preliminary phenotypic asse ssment. The LSR deficient strain was maintained by back-crossing heterozygous males with C57BL/6, 129/Sv and MF1 females a t e ach generation. Mice at back-cross generati on 1–6 were intercrossed to provide t he homozygous null, heterozygous null and wild- type mice used in the analyses d escribed herein. Animal breeding and experiments were c arried out in accordance with the European Communities Council Directive of 24 November 1986. Genotyping of LSR and neo genes by PCR For PCR, genomic DNA from embryos and adult m ouse tails was extracted by proteinase K digestion, isolated using the Genomic DNA Purification Kit (Promega, Charbonnie ` - res, France) and p recipitated with ethanol. PCR primers were selected to generate a product specific for either the wild-type or the mutant LSR allele. The wild-type LSR allele was diagnosed by a 773-bp PCR product g enerated by a f orward primer located in e xon 4 (5¢-CAGGACC TCAGAAGCCCCTGA-3) and a reverse primer located in exon 5 (5¢-AACAGCACTTGTCTGGGCAGC-3¢). This region of the LSR gene is deleted in the mutant allele. The Fig. 1. Generation of the LSR null allele. (A) Structure of the mouse LSR gene (top), the linearized LSR targeting vector (middle) and the targeted allele (bottom) resulting from replacement recombination. The null allele was created by deletion of a 9.8 kb in ternal region of the gene from the beginning of exon 2 to the end of exon 5 and its substitu tion with a b-galactosidase/neomycin phosphotransferase reporter/selection cassette. Dashed crosse s indicate the recombination cross-over positio ns between homologous vector and c hromosomal sequence. Chrom osomal and cloned genomic DNA sequence is shown by a thick black line (for intron and flanking nonc oding sequence) and by black rectangle s (for exon sequence), the reporter/positive selection cassette by IRES laczpA and grey (loxP/MC1neopA loxP )1 ) rectangles, the HSV thymidinekinasenegativeselection cassette (MC1tk dimer) by a rectangle and p Bluescript plasmid sequence by a thin black line. Sites fo r HindIII restriction enzyme (H)areindicated by small arrows and t he sizes of relevant restriction fragments in th e wild-type and targeted allele are shown by dotted lines. The targeted allele was identified by HindIII digestion and hybridization with the 5¢-and3¢-flanking probe fragments (striped rectangles) to detect the indic ated size fragments. (B) Southern blot analysis of Hin dIII-digested genomic DNA prepared from 96-well p lates of G418+ G ancyclovir resistant ES cell clones derived from transfection w ith the LSR targeting vector. The digested DNA and a kHindIII marker was resolved o n a 0.6% agarose gel, blotted to positively charged nylon membrane and hybridized with 25 ng of 3¢-probe and 25 ng of kHindIII marker. The hyb ridized blots were exposed to Ko dak XOMAT film overnight at )80 °C. The 3¢-probe detects a 10.5 kb HindIII fragment for the w ild-type allele and a 13 kb fragment in a targeted a llele. 3106 S. Mesli et al.(Eur. J. Biochem. 271) Ó FEBS 2004 targeted mutant allele was detected by the presence of the neo gene. Two couples of neo primers have been used during the course of this work: (forward: 5¢-GGCGCCCGG TTCTTTTTGTCA-3¢ and r everse: 5 ¢-TTGGTGGTCG AATGGGCAGGT-3¢ giving a p roduct o f 281 bp) and (forward: 5¢-GAGGATCTCGTCGTGACCCATG-3¢ and reverse: 5¢-GAGGAAGCGGTCAGCCCATT-3¢ giving a product of 179 bp). For the wild-type LSR gene, c onditions were (94 °C for 30 s, 63 °Cfor1min,72°Cfor30s;35 cycles. For the neo gene, PCR conditions were: 95 °Cfor 30 s, 68 °C for 1 min, and 72 °C for 30 s; 33 cycles. In both cases, PCR cycles were preceded by 10 min at 95 °Cand endedby7minat72°C. Results To obtain insight on the possible function(s) of LSR, we first determined the tissue d istribution of i ts mRNA in organs of adult mice in comparison with that of mRNA o f other lipoprotein r eceptors o r apolipoproteins. W e a lso determined the amount of LSR mRNA at different time points of embryonic development. Northern blots Figure 2A shows a Northern blot of selected adult murine tissues hybridized with an LSR probe. As expected from results obtained in the rat [5], a 2.1 kb band was observed in liver. A faint but clean band was also observed in testis and kidney. Hybridization with a GAPDH probe showed unequal loading of the c ommercial membrane and partic- ularly that the liver lane was overloaded. Quantitation of the amount of LSR mRNA was thus performed after normal- ization of the radioactivity of the LSR bands to that of the 1.35 kb GAPDH band. Data showed that testes and kidney contained, respectively, 63% and 48% of the signal present in liver. Figure 2 B shows a Northern blot containing mRNA from whole embryos at stages E7, E11, E15, and E17 hybridized with an LSR probe and reprobed with a GAPDH cDNA. The 2.1 kb LSR band was detected at all stages. Again, loading of the lanes w as unequal making direct quantification difficult. As in the case of adult tissues, we normalized LSR bands to the corresponding 1.35kb GAPDH bands. Ratios were approximately equal at all stages, indicating that the LSR expression level was of the same order of magnitude between E7 and E 17. Real-time quantitative RT-PCR In a first selection of tissues (liver, o varies, adrenal glands, testes, intestine, brain and muscle), LSR mRNA was extracted as described in Materials and methods. Results obtained by real-time quantitative RT-PCR were normal- ized to the amount of 18S ribosomal RNA (Fig. 3A and Table 2). Quantitative PCR was also performed on lung, kidney and heart s amples, but in that case the starting material was commercially available total cDNA. For those tissues, data were normalized to the amount of GAPDH mRNA (Fig. 3B and Table 3). Liver c DNAs were obtained from both the mRNA extracted in our laboratory and from the commercial source in order to allow us to compare the two sets of experiments. Figure 3A and Table 2 show that LSR mRNA is very abundant in liver, as expected from t he Northern blot analysis. We also found a significant expression in ovaries and testes (respectively 62.8%, and 21.7% of liver), but the Fig. 2. Northern blots of adult murine t issues (A) and whole embryos (B) mRNAs. E7, E11, E15, E17: em bryo stages (days post-coitum). The LSR probe reveals a 2.1 kb band, and the GAPDH probe a 1.35 kb band and (in some tissues) a 1.2 kb band. Ó FEBS 2004 LSR gene and protein expression (Eur. J. Biochem. 271) 3107 amount in adrenal glands was only 4 % of that o f liver. A substantial expression was found in intestine and brain (respectively 41.9% and 15.9% of liver). The expression in muscle was very low (0.5% of liver). Figure 3B and Table 3 show that LSR mRNA is rather abundant in lung and kidney (55.8% and 11.8% of liver) but barely detectable in heart. The distribution of several gene mRNA involved in lipoprotein metabolism w as stu died co mparatively as an attempt to get some insight into the possible functions of LSR. The tissue distribution of LDLR mRNA (Tables 2 and 3) is not very different from that of LSR mRNA, w ith t wo notable exceptions: ( a) it was more a bundant in ad renal glands and o varies than in liver (respectively 200%, and 180% of liver); (b) it was abundant in muscle (41.2% of liver). The pattern of expression of SR-BI mRNA (Tables 2 and 3) was rather different from that of LSR mRNA: ( a) it was extremely abundant in adrenal glands and ovaries (respect- ively 46.5-fold and 18-fold the amount present in liver); (b) expression in testes, brain and muscle was rather abundant (respectively 1 15%, 122.2% and 55.8% of the amount present in liver). The tissue distribution of LR P mRNA in adult mice (Tables 2 and 3 ) was also very different from that of LSR: its amount in ovaries, adrenal glands, brain and muscle was higher than that of liver (respectively 410%, 1 90%, 180%, and 250% of the amount present in liver). Although RT-PCR arbitrary units do not reflect precisely tr ue message amounts , d ue to the different amplification efficiencies for different gene targets, t aken altoget her, the results suggest that the amount of LSR messengers in liver is higher than that of the o ther receptors here described. It must be noted that Fig. 3A and B have different scales because one was normalized to 18S ribosomal RNA and the other to GAPDH mRNA. Several mRNA species were used as con trols for tissue- specific expression. As expected, prothrombin mRNA was almost exclusively expressed in liver; apoA1 and apoB mRNA were expressed mainly in liver but also in intestine; apoE mRNA was predominant in liver but abundant in all tissues; u biquitin and GAPDH mRNA were ubiquitous, and showed important variations of expression from one tissue to another. The expression of LSR was also studied by quantitative PCR during mouse embryonic d evelopment. cDNAs from whole embryos at E7, E11, E15, E17 stages were used as starting material and results were normalized to GAPDH mRNAs. Figure 3C shows that LS R was d etectable at E7, became more abundant at E11 (fourfold increase) and maintaining these increased levels until E17. This pattern of expression seems t o parallel liver growth as a similar t ime- course was observed for prothrombin. Table 4 shows that in contrast, LDLR and SR -BI mRNA had d ifferent time- courses with a higher amount at E7, followed by a decrease at E11 and an increase at E15. LRP showed a t ime-course similar to t hat of LSR excep t for a decrease at E17; apoA1, apoB and ubiquitin showed a time-course similar to that of LRP. Immunofluorescence To localize the LSR receptor itself, different murine tissues were studied by indirect immunofluorescence with an anti- LSR antiserum. To avoid misinterpretations due to back- ground, two normal rabbit sera were systematically included in the labeling experiments. Moreover, tissues from LSR knockout mice were also tested with the anti-LSR anti- serum. Figure 4A,B shows the presence in adult liver of a strong specifi c signal at the periphery of hepatocytes. This staining pattern is compatible with the previously described localization of LSR at the plasma membrane l evel [5]. The presence of LSR could also be detected in fetal liver cells in E12 and E15 embryos (data not shown). A faint but specific staining was detected in k idney Fig. 4E. The s ignal w as observed in the kidney cortex, mainly at the level of glomerules. Fig. 3. Quantitation of LSR mRNA by real-time PCR i n adult murine tissues ( A,B) and w hole post-implantation embryos (C). Data were normalized to 18S ribosomal RNA (A) and to GAPDH mR NA (B, C). dC T were converted to arbitrary values by the following formula: 2 )dC T · 10 6 . Liver (L), o varies (o), adrenal glands (a), testes (t), intes- tine (i), brain (b) and muscle (m), l ung (lu), kidn ey (k) and heart (h). E7, E11, E 15, E17: Embryo stages (days post-coitum). 3108 S. Mesli et al.(Eur. J. Biochem. 271) Ó FEBS 2004 Knock-out of the LSR gene Mice with one LSR allele inactivated did not show any detectable defect. Their size, weight, adiposity, plasma glucose, cholesterol, triglycerides, phospholipids, nonesteri- fied fatty acids, free glycerol, as w ell as their lipoprotein profile were similar to those of their wild-type littermates. Animals bearing two i nactivated LSR alleles (LSR–/–) show an embryonic lethality between E12.5 and E15.5. As an attempt to define the reason for the embryonic lethality of LSR–/– embryos, timed matings were set up and resulting embryos examined and genotyped (Table 5 and Fig. 5). Up to E12.5, LSR–/– mice were obtained in numbers compatible with Mendelian ratios, and macro- scopic e xamination of t he whole litters showed that all embryos were alive and had no observable anomalies. But at E15.5, genotyping did not show the presence of viable homozygote embryos. Resorbed embryos were numerous at E14.5/15.5 and the majority were most probably LSR–/–, but we were not able to genotype them because of DNA degradation. At E14.5, some litters contained LSR–/– embryos. Their only c onstant defect was a reduction in liver size (Fig. 6A); in some embryos, the liver was reduced to a punctiform red spot (not shown). Histological sections of E14.5 LSR–/– embryos showed that the cell density was lower than in the wild-type littermates. Spaces devoid of cells were observed, but no specific cellular abnormalities or absence of certain cell types were ob served in the liver of the mutants. For example, megacaryocytes, although rare (Fig. 6 E), c ould be found in LSR–/– embryos (not shown). LSR–/– embryos had other anomalies but they were not constant: a general white coloration, while the LSR+/+ and LSR+/– littermate e mbryos had a pinkish hue; superficial hemmorrhages (Fig. 6A), superficial detachment of the skin (Fig. 6A); a smaller size than their littermates and finally some of them were obviously dead. Interestingly homozygote embryos did not show an overall d evelopmen- tal delay, as s hown by limb bud, eye and facial development (Fig. 6 A). During the last three years, no viable adult LSR–/– was obtained by intercrossing LSR+/– mice. However in the very first litters obtained b y intercrossing male LSR+/– derived from t he chimeras with female LSR+/– derived Table 2. Quantitation of LSR, LDLR, SR-BI, LRP, apoA1, apoB, apoE, ubiquitin, prothrombin,andGAPDH mRNAs by real-time PCR in a first set of adult murine tissues. For e ach gene, results were normalized t o 18S ribosomal RNA. dC T wereconvertedtoarbitrary values by the following formula: 2 )dC T · 10 6 . mRNA Tissue Liver Ovaries Adrenal glands Testis Intestine Brain Muscle LSR 710 446 28.5 154 298 113 3.61 LDLR 83.4 150 167 10.7 68.8 63.8 34.4 SRBI 77.8 1390 3620 89.7 16.5 95.1 43.4 LRP 22.2 91 41.7 5.5 9.8 40.6 55.8 ApoAI 244 0.79 0.04 0.06 57.1 0.04 0.06 ApoB 16900 0.3 5.5 1.5 1630 0.4 0.1 ApoE 83000 4830 7490 899 854 13900 1190 Ubiquitin 478 1840 2780 6050 343 3830 1490 Prothrombin 9490 0.5 1.8 0.4 9.8 0.1 0.8 GAPDH 1780 5910 3700 321 321 19800 51100 Table 3. Quantitation of LSR, LDLR, SR-BI, LRP, apoA1, apoB, apoE, ubiquitin, and prothrombin by real-time PCR in a second set of adult murine tissues. For each gene, results were normalized to GAPDH mRNA. DC T were converted to arbitrary values by the following formula: 2 )dC T · 10 6 . mRNA Tissue Liver Lung Kidney Heart LSR 67500 37700 7980 303 LDLR 15400 13300 2630 141 SRBI 3740 20200 144 95.6 LRP 538 1150 911 125 ApoAI 3380 46.6 4.86 3.58 ApoB 86600 593 380 4.76 ApoE 15 700 000 717 000 93400 63400 Ubiquitin 2060 000 6770 000 966 000 63700 Prothrombin 901 000 6300 128 4.63 Table 4. Quantitation of LSR, LDLR, SR-BI, LRP, apoA1, apoB, apoE, ubiquitin an d prothrombin by real-time PCR in whole post- implantation embryos. For each g ene, results w ere n ormalized to GAPDH mRNA. DC T were converted to arbitrary values by the fol- lowing formula: 2 )dC T · 10 6 . E mbryo stages (days post-coitum.): E7, E11, E15, E17. mRNA Age E7 E11 E15 E17 Adult Liver LSR 1310 5680 7190 7810 67 500 LDLR 21 900 2880 4780 2560 15 400 SRBI 14 100 4160 6430 9620 37 400 LRP 9230 16 900 25 600 7440 538 ApoA1 5.82 20.6 93.8 76.7 3380 ApoB 32 3190 25 600 5340 86 600 ApoE 12 600 30 800 277 000 578 000 1 5700 000 Ubiquitin 979 000 1 580 000 3 030 000 1 060 000 2 060 000 Prothrombin 13.6 2670 40 900 114000 901 000 Ó FEBS 2004 LSR gene and protein expression (Eur. J. Biochem. 271) 3109 from the fi rst generation, or intercrossing male and female LSR+/– mice derived from the first generation, three viable LSR–/– mice (all males, two from one litter, and one from another) were obtained. They had no morphological defects except that one of them seemed to have no testes. They were smaller than their littermates: the 9-month weight of LSR–/– was 30.7 ± 0.2 vs. 39.3 ± 2.1 g for their wild-type litter- mates (P<0.02). Continual matings for 3 months demon- strated that these mice were sterile. As one of the LSR–/– mice died spontaneously and the others became s ick (lethargic), we killed these two animals for necropsy and collection of organ samples; they both showed a limited amount of fat and one of them actually had n o testes, but no other anatomical defect was detected. To explore whether the genetic bac kground cou ld infl uence t he viability of LSR–/– mice, we backcrossed the mutations in two inbred strains (C57BL/6 and 129/Sv) and an outbred strain (MF1); we also intercrossed heterozygotes o f C57BL/6 and 129/ Sv back- grounds, but no viable LSR–/– mice were obtained. Discussion In this study, we used No rthern b lotting, real-time PCR and immunofluorescence microscopy to examine the expression of LSR in the adult mouse and during development. In the adult, the highest levels of LSR expression were found in liver as expected from results obtained in the rat [5]. Several reports published by Bihain and colleagues [2–5] have provided circumstantial evidence for a role of LSR in the Fig. 4. Immunolocalization of LSR in liver and kidney. (A,B,C,E) anti-LSR serum, (D,F) normal rabbit serum. Specific staining i s observed at the periphery of hepatocytes o f adult liver (A,B). No signal was observed i n liver from one LSR–/– mouse (C). Specific staining is also found in kidney cortex ( E) principally at the level of glomeruli (arrows). No staining was detected in liver (D) and kidney (F) t reated with normal rabbit serum. Bar in (F) relates to 20 lm (A,C,D,E,F) an d 6 lm(B). Table 5. Embryos obtained by intercrossing LSR+/– mice. Em bryos were genotyped as shown in Fig. 6. Living LSR–/– embryos were found a t E10.5 and E12.5 and were apparently normal. LSR–/– embryos could not be found at E15.5. At E14.5, living LSR–/– embryos were found but had all a small liver (Fig. 6A). Moreover, one litter contained two d ead LSR–/– embryos with a punctiform liver. Genetic background Age No. of embryos No. of litters LSR mutants +/+ +/– –/– C57BL/6 E10.5 10 1 3 6 1 C57BL/6 E12.5 7 1 1 4 2 C57BL6/)129/Ola E12.5 25 3 4 14 7 C57BL/6 E14.5 10 1 3 7 0 C57BL/6 E14.5 8 1 3 3 2 a C57BL/6–129/Ola E14.5 10 1 4 6 0 MF1 E14.5 22 2 5 11 6 C57BL/6–129/Ola E15.5 13 1 7 6 0 a Dead embryos. 3110 S. Mesli et al.(Eur. J. Biochem. 271) Ó FEBS 2004 clearance of TRL by the liver. The LDLR and the LRP have both been shown t o be involved in t he removal of chylomicron remnants by the liver [12,13]. The facts that mice with an isolated inactivation of the LDLR show no increase in circulating TG [14], and that the lack of L DLR in humans does not lead to a pathological change in the metabolismofdietaryfat[15]suggestthat(an)other receptor(s) play(s) the major part in TRL clearance by the liver. Mo reover, Rohlmann et al. [16] demonstrated that the absence of LRP expression in the livers of LDLR-deficient mice resulted in a large elevation in the plasma concentra- tion of cholesterol and TG that were carried in apo B48- containing lipoproteins resembling remnants. Nevertheless, in LDLR-deficient mice the increase in TG levels was much smaller than that obtained in RAP overexpression experi- ments [17]. The authors concluded that the most probable explanation is t hat RAP-sensitive receptors suc h as LSR [7] could be involved in TRL clearance. Actually, our real-time PCR data indicate that in liver, LSR mRNA is expressed as well as LDLR and LRP mRNA, suggesting that the newcomer receptor could indeed play an important role in TRL clearance by the liver. Moreover, the abundance of LSR i n liver contrasts with its almost complete absence in skeletal muscle and heart. In that respect LSR differs strikingly from the VLDL r eceptor which is very abundant in these l atter tissues and is involved together with LPL in fatty acid uptake by striated muscle [18]. Thus, our data indicate that LSR could be specialized in the uptake of TRL by liver as suggested by its d iscoverers [5]. The demonstra- tion of this hypothesis would require an analysis of the lipoprotein phenotype of a sufficient number o f adult LSR–/– mice (for instance after d esigning a liver-specific inducible inactivation of the LSR gene). Recent s tudies have suggested that cholesterol plays a crucial role in specific processes during embryonic develop- ment. Cholesterol deficiency during embryogenesis can be caused by defects in apolipoproteins, enzymes or cell- surface r eceptors that are potentially involved in cellular lipoprotein uptake, either by cells of the yolk sack or the placenta or by the embryo itself [19]. We have studied LSR expression during late embryogenesis in comparison with other lipoprotein receptors which are known to play a n important role in embryonic development, and with prothrombin, a liver-specific marker. Due to unequal loading of the lanes, Northern blots were not sensitive enough to show significant changes in LSR expression between E7 and E17. However real-time PCR showed that LSR mRNA is detectable at E7, becomes abundant at E11 (fourfold increase) and remains practically constant until E17. This can be attributed to liver organogenesis which follows a similar time-course [20]. Moreover, LSR protein was detected by immunofluorescence in dissected fetal livers of E12 and E15 mice. Although our real-time PCR data show that all lipoprotein receptors tested follow r oughly similar time-courses between E11 and E17, LSR and LDLR are probably the on ly receptors, among those tested to b e present in fetal liver in substantial amounts. Actually previous reports show that (a) the LDLR is present in rat liver from E19 fetuses at 19% of the adult level; (b) hepatic LRP is still low at 1 9 days of gestation (only 6% of the adult level) [21] and (c) SR-BI is not detectable in embryonic liver until stage E17 [22]. The increased SR-BI mRNA synthesis that we observed between E11 and E15 is probably due to adrenal gland organogenesis [22]. Fetal liver has been shown to synthesize and export into t he fetal circulation about one- half of the cholesterol required f or heart, lung and kidney development [21, 23]. The early e xpression of LSR in fetal liver sugge sts that this receptor could play a role in the uptake of lipoproteins during embryogenesis, a process that cannot be effected by SR-BI at this stage [22]. The scarcity of LSR messages at E7 contrasts with t he high expression at that stage o f t he other lipoprotein receptors which are involved in exchanges between the embryo and extraem- bryonic an d maternal tissues. For example, SR-BI present on the apical surfaces of visceral endodermal is thought to provide cholesterol to extraembryonic cells for s torage until it can be subsequently transferred to the embryo [22]. Whatever the importance of LSR for post-implantation embryo viability, it must be noted that the abundance of its mRNA increases dramatically during adulthood (Fig. 3). It would be interesting to determine whether it is suckling, like in the case of LRP [21], or weaning, as in the case of LDLR [24], which triggers the increase of LSR seen in liver, as such an induction would be consistent with a role o f LSR in chylomicron remnant metabolism. The lethality of LSR–/– embryos that we observed occurs around E12.5–14.5, a period which is concomittant Fig. 5. Genotyping of embryos obtained by intercrossing LSR+/– mice. PCR of diagnostic LSR (773 bp) and neo (180 bp) ge ne regions is described in Materials and me thods. LSR –/– embryos (upper 5, 6) are present at E12.5 but missing at E15.5. Ó FEBS 2004 LSR gene and protein expression (Eur. J. Biochem. 271) 3111 with the appearance of specific hepatocyte function in the fetus [25–27]. The fact that LSR–/– embryos’ livers were smaller (and sometimes punctiform) at E14.5 but normal size at E12.5 i ndicates that an atrophy of the liver occurred after a first period of apparently normal growth. Histological sections o f E14.5 LSR–/– embryo s actually showed that the cell density was lower than in the wild- type littermates; moreover spaces devoid of cells were observed in the mutants. Future studies using time-specific markers of liver development will be conducted to compare t heir time-course in the mutant and the wild- type mice. If the primary effect of the m utation indeed affects liver de velopment, otherdefectssuchasthesmaller size and even lethality of the embryo can be explained by ischemia, as liver is the major hematopoietic organ at that stage. Inactivations of some lipoprotein receptor genes, for instance LRP and gp330/megalin, have also been found to result in embryonic lethality by various mechanisms [28,29]. We found transcripts for all the lipoprotein r eceptors tested, including LSR, in steroidogenic organs such as adrenal glands, testes and ovaries. This is in agreement with results showing SR-BI to be highly expressed in steroido- genic tissues [30,31] which are the sites of the h ighest s pecific activity for selective HDL cholesterol uptake in rodents [32]. Nevertheless, LSR and SR-BI were expressed differently, more in reproductive organs than in a drenal glands fo r LSR, and inversely for SR-BI. The specific abundance of LSR mRNA in testes suggests that this receptor c ould play an important role in this organ. However its implication in steroidogenesis is questionable a s SR-BI, which has also been detected in testes, seems to mediate phagocytosis of apoptotic spermatogenic cells by Sertoli cells after recogni- tion of surface phosphatidylserine [33,34]. In ovaries and Fig. 6. Gross morphology and liver histology of atypicalLSR–/– mutant embryo compared to a wild-type littermate. Lateral views of E14.5 embryos (A,B). The LSR–/– embryo (A) shows a reduction of liver size (white arrow) and displays hemorrhages (black arrow) contrasting w ith an anemic color. I t also shows a detachment of dorsal skin ( small arrows). The wild-type littermate (B) sh ows a liver of normal size (white arrows); its skin has a pinkish hue, distinct subcutaneous vessels and does not show detach ment. Note that the overall development of the mutant is not dif- ferent from that of the wild-type as shown by embryo size, and limb bud and eye stages. Histological sections of livers of E14.5 em- bryos stained with hematoxylin and eosin (C,D,E,F). Note the presence of large inter- cellular spaces (arrow) in the liver of the mutant (C,E) contrasting with the normal architecture of the wild-t ype liver (D,F). In addition, megacaryocytes were very rare in the mutant liver while they were easily found in the wild-type liver [see arrow on view (F)]. Bar in F relates to 160 lm(C,D)and40lm ( E,F). 3112 S. Mesli et al.(Eur. J. Biochem. 271) Ó FEBS 2004 [...]... hypothesized function in lipid transport in these organs Unfortunately, the scarcity of viable LSR–/– adult mice did not allow us to obtain definitive information on the role of LSR in lipid and lipoprotein metabolism The production of a conditional knock-out will be necessary to explore this question Further studies are also required to understand the mechanisms of liver involution and lethality in LSR–/–. .. (1993) A comparison of the roles of the low density lipoprotein (LDL) receptor and the LDL receptorrelated protein/alpha 2-macroglobulin receptor in chylomicron remnant removal in the mouse in vivo J Biol Chem 268, 15804– 15811 Ishibashi, S., Brown, M.S., Goldstein, J.L., Gerard, R.D., Hammer, R.E & Herz, J (1993) Hypercholesterolemia in low density lipoprotein receptor knockout mice and its reversal... for targeting mutations to non-selectable genes Nature 336, 348–352 Smith, A.G (1991) Culture and differentiation of embryonic stem cells J Tiss Cult Meth 13, 89–94 Willnow, T.E., Goldstein, J.L., Orth, K., Brown, M.S & Herz, J (1992) Low density lipoprotein receptor- related protein and gp330 bind similar ligands, including plasminogen activator-inhibitor complexes and lactoferrin, an inhibitor of chylomicron... maintains alveolar patency Its synthesis is critically dependent on the availability of fatty acids A variety of receptors, including SR-BI [39], LDLR [40], LRP and gp330/megalin [41,42], are present on alveolar cells and are able to bind lipoproteins and to participate to surfactant synthesis Moreover there is evidence that VLDL in the presence of lipoprotein lipase (LPL), provide the free fatty acid substrate... apoE-containing lipoproteins and cell surface apoE would provide cholesterol to ovarian and adrenocortical cells for steroidogenesis Our real-time PCR data show that LSR, LDLR, SR-BI and LRP messages are all very abundant in lung, i.e in the range of liver or more Lung surfactant is a surface tension lowering mixture of lipids and hydrophobic proteins that lines the alveolar surface and maintains alveolar... of specific apolipoprotein C-III isoforms on the binding of triglyceride-rich lipoproteins to the lipolysis- stimulated receptor J Biol Chem 272, 31348–31354 Troussard, A.A., Khallou, J., Mann, C.J., Andre, P., Strickland, D.K., Bihain, B.E & Yen, F.T (1995) Inhibitory effect on the lipolysis- stimulated receptor of the 39-kDa receptor- associated protein J Biol Chem 270, 17068–17071 Nehls, M., Kyewski,... lipoprotein receptor Science 271, 518–520 31 Landschulz, K.T., Pathak, R.K., Rigotti, A., Kriegerj, M & Hobbs, H.H (1996) Regulation of scavenger receptor, class B, type I, a high density lipoprotein receptor, in liver and steroidogenic tissues of the rat J Clin Invest 98, 984–995 32 Glass, C., Pittman, R.C., Civen, M & Steinberg, D (1985) Uptake of high-density lipoprotein-associated apoprotein A-I and. .. apoB/E receptors [45] Moreover, several pathological disorders are accompanied by lipid deposition into glomeruli [46] The presence of LSR in glomerular cells might provide an additional pathway for explaining lipoprotein uptake in normal and pathological glomerular cells The abundance of LSR in fetal and adult liver as well as in steroidogenic organs and organs such as lung or kidney adds further evidence... expression of elements of hepatic cholesterol metabolism in the rat J Lipid Res 36, 641–652 22 Hatzopoulos, A.K., Rigotti, A., Rosenberg, R.D & Krieger, M (1998) Temporal and spatial pattern of expression of the HDL receptor SR-BI during murine embryogenesis J Lipid Res 39, 495–508 23 Jurevics, H.A., Kidwai, F.Z & Morell, P (1997) Sources of cholesterol during development of the rat fetus and fetal... N., Andre, P., Grosset, J.M., Bougueleret, L., Dumas, J.B., Guerassimenko, O & Bihain, B.E (1999) Molecular cloning of a lipolysis- stimulated remnant receptor expressed in the liver J Biol Chem 274, 13390–13398 Mann, C.J., Troussard, A.A., Yen, F.T., Hannouche, N., Najib, J., Fruchart, J.C., Lotteau, V., Andre, P & Bihain, B.E (1997) Inhibitory effects of specific apolipoprotein C-III isoforms on the binding . Distribution of the lipolysis stimulated receptor in adult and embryonic murine tissues and lethality of LSR–/– embryos at 12.5 to 14.5 days of gestation Samir. determine the distribution of LSR mRNA and protein in murine organs, and whether this distribution was c ompatible with the alleged role of this new receptor