Differences in lipid distribution and expression of peroxisome proliferator-activated

Differences in lipid distribution and expression of peroxisome proliferator-activated receptor gamma and lipoprotein lipase genes in torafugu and red seabreamGen Kaneko a ,⇑,Toshihiro Yamada a ,Yuna Han a ,Yuki Hirano a ,Anurak Khieokhajonkhet a ,Hirohito Shirakami a ,Reiko Nagasaka b ,Hidehiro Kondo c ,Ikuo Hirono c ,Hideki Ushio a ,Shugo Watabe a ,daDepartment of Aquatic Bioscience,Graduate School of Agricultural and Life Sciences,The University of Tokyo,Yayoi,Bunkyo,Tokyo 113-8657,Japan bDepartment of Food Science and Technology,Tokyo University of Marine Science and Technology,Minato,Tokyo 108-8477,Japan cDepartment of Marine Biosciences,Graduate School of Marine Science and Technology,Tokyo University of Marine Science and Technology,Minato,Tokyo 108-8477,Japan dDepartment of Marine Biochemistry,School of Marine Biosciences,Kitasato University,Sagamihara,Kanagawa 252-0373,Japana r t i c l e i n f o Article history:Received 17July 2012Revised 12November 2012Accepted 3January 2013Available online 19January 2013Keywords:AdipocytesIn situ hybridization Myosepta Pagrus major Takifugu rubripes Tissue distributiona b s t r a c tLipid content is one of the major determinants of the meat quality in fish.However,the mechanisms underlying the species-specific distribution of lipid are still poorly understood.The present study was undertaken to investigate the mechanisms associated with lipid accumulation in two species of fish:torafugu (a puffer fish)and red seabream.The lipid content of liver and carcass were 67.0%and 0.8%for torafugu,respectively,and 8.8%and 7.3%for red seabream,respectively.Visceral adipose tissue was only apparent in the red seabream and accounted for 73.3%of its total lipid content.Oil red O staining confirmed this species-specific lipid distribution,and further demonstrated that the lipid in the skeletal muscle of the red seabream was mainly localized in the myosepta.We subsequently cloned cDNAs from torafugu encod-ing lipoprotein lipase 1(LPL1)and LPL2,important enzymes for the uptake of lipids from blood circulation system into various tissues.The relative mRNA levels of peroxisome proliferator-activated receptor gamma (PPAR c )and the LPLs of torafugu were determined by quantitative real-time PCR together with their coun-terparts in red seabream previously reported.The relative mRNA levels of PPAR c and LPL1correlated closely to the lipid distribution of both fish,being significantly higher in liver than skeletal muscle in torafugu,whereas the highest in the adipose tissue,followed by liver and skeletal muscle in red seabream.However,the relative mRNA levels of LPL2were tenfold lower than LPL1in both species and only correlated to lipiddistribution in torafugu,suggesting that LPL2has only a minor role in lipid accumulation.In situ hybridiza-tion revealed that the transcripts of LPL1co-localized with lipids in the adipocytes located along the myosepta of the skeletal muscle of red seabream.These results suggest that the transcriptional regulation of PPAR c and LPL1is responsible for the species-specific lipid distribution of torafugu and red seabream.Ó2013Elsevier Inc.All rights reserved.1.IntroductionThe major sites of lipid accumulation in fish are the liver,skel-etal muscles,and adipose tissues.However,the distribution of lipid among these tissues is highly species-specific.For species such as torafugu Takifugu rubripes and Japanese flounder Paralichthys oliva-ceus ,lipid is predominantly located in the liver (Ando et al.,1993).Other species such as red seabream Pagrus major and amberjack Seriola dumerili store lipid in both the liver and skeletal muscle (Ando et al.,1993),whereas Japanese sardine Sardinops melanostic-tus stores more than 40%of body lipids in its subcutaneous adipose tissue (Shibata and Kayama,1989).Since lipid content is one of the major determinants of fish meat quality,the molecular basis for li-pid accumulation in skeletal muscle has been extensively investigated.A previous study has shown that the lipid content of skeletal muscle is positively correlated to the plasma triacylglycerol (TAG)levels in several fish species including torafugu,Japanese flounder,and red seabream (Ando and Mori,1993).In plasma,TAGs are mainly secreted from liver and intestine as the compo-nents of lipoproteins to transport lipids from these organs to peripheral tissues.It is thus likely that lipids in fish skeletal muscle are mainly transported from the liver and intestine via the blood stream.Meanwhile,synthesis of lipid in skeletal muscle would be of limited significance,owing to the liver acting as the main site of de novo synthesis of fatty acids in fish (Gnoni and Muci,1990;Lin et al.,1977).However,the molecular mechanisms regulating lipid transport and its contribution to the species-specific lipid distribution in fish are still poorly understood.Peroxisome proliferator-activated receptors (PPARs)are ligand-dependent transcription factors which regulate lipid metabolism.In mammals,PPARs are activated by specific ligands including fatty acids and their derivatives,and promote the transcription of genes0016-6480/$-see front matter Ó2013Elsevier Inc.All rights reserved./10.1016/j.ygcen.2013.01.003Corresponding author.Fax:+8158418166.E-mail address:agkaneko@mail.ecc.u-tokyo.ac.jp (G.Kaneko).related to lipid synthesis,the uptake of lipid from the blood stream into the tissues,and b-oxidation(Tontonoz and Spiegelman,2008). To date,three isoforms of PPAR—PPAR a,PPAR c,and PPAR b/d—have been identified.The PPAR c isoform is important for lipid synthesis and uptake,whereas the remaining two isoforms are mainly in-volved in b-oxidation.Since the protein levels of PPAR c have been shown to increase in association with the adipogenic differentiation of salmon primary preadipocytes(Vegusdal et al.,2003),it is possi-ble that the adipogenic functions of PPAR c are conserved infish. However,several conflicting reports exist on the function of PPAR c in the lipid metabolism offish.For example,the transcripts encod-ing PPAR c have been found in various tissues of torafugu(Maglich et al.,2003),plaice Pleuronectes platessa,and gilthead seabream Spa-rus aurata(Leaver et al.,2005),whereas they are restricted to the white and brown adipose tissues of mice(Tontonoz and Spiegel-man,2008).In addition,fatty acids and mammalian PPAR c-specific ligands failed to enhance the transcriptional activity of PPAR c in plaice,gilthead seabream(Leaver et al.,2005),torafugu(Kondo et al.,2007),and medaka Oryzias latipes(Kondo et al.,2010).Conse-quently,the precise role offish PPAR c is still largely undetermined.Lipoprotein lipase(LPL)[E.C.3.1.1.34]is an important enzyme for lipid uptake,the transcription of which is regulated by PPAR c in mice(Wang and Eckel,2009).LPL hydrolyzes TAGs circulating in the blood stream as components of lipoproteins into free fatty acids(FFAs)and2-monoacylglycerol.The resulting FFAs can be incorporated into tissues,and stored as TAG after re-esterification, or used as an energy source via b-oxidation.LPL activity has report-edly been highest in the adipose tissue of both rainbow trout Oncorhynchus mykiss(Albalat et al.,2006)and gilthead sea bream (Albalat et al.,2007a),suggesting the expression of LPL is responsi-ble for lipid accumulation in these species.The present study was undertaken to further investigate the role of PPAR c and LPL in the species-specific lipid distribution offish.2.Materials and methods2.1.FishTorafugu specimens(25–130g)used for the oil red O staining of frozen sections were reared in a3ton-tank under a12-h light/dark cycle at25°C.The rearing density was approximately100individ-uals per tank.Fish were fed with commercial pellets(Otohime; Marubeni Nisshin Feed Co.Ltd.,Tokyo,Japan)ad libitum once a day.Torafugu specimens(500–750g)cultured in Ehime Prefecture, Japan,purchased from a commercial dealer(Fish Interior,Tokyo, Japan),and acclimated in laboratory aquariums at20°C for several days under the above conditions were used for the measurement of lipid content,the oil red O staining offillets and the quantitative real-time PCR.Red seabream(4kg)cultured in an outdoor seawater pond in Kanagawa Prefectural Fishery Experimental Station,Japan,were collected in February2008and used for the oil red O staining of frozen sections,while specimens(600g)cultured in Ehime Prefec-ture,Japan,purchased from the Fish Interior,and acclimated at 20°C for several days under the above conditions were used for the measurement of lipid content,the oil red O staining offillets, quantitative real-time PCR,and in situ hybridization.2.2.Measurement of lipid contentThe lipid content of carcasses and the livers of torafugu(n=3) were determined using the conventional Bligh and Dyer method (Bligh and Dyer,1959).The chloroform fractions were combined and completely evaporated using a rotary evaporator and vacuum pump.The lipid content of each organ was determined gravimetrically.The same approach was used for the carcasses,liv-ers,and visceral adipose tissues of the red seabream(n=3).2.3.Oil red O stainingFillets of torafugu and red seabream(500–750g)werefixed in phosphate buffered saline(PBS)containing4%paraformaldehyde (PFA)at4°C for1week(n=3),and stained with0.3%oil red O in a mixture of isopropanol and distilled water(6:4v/v).Thefillets were then washed with distilled water and their images captured using a digital camera.For the oil red O staining of the frozen sections,skeletal muscle, liver,skin,intestine,and gills of torafugu and red seabream(n=2 and3,respectively)werefixed in4%PFA/PBS at4°C overnight.Tis-sues were washed twice with0.2M phosphate buffer for30min. Subsequently,the tissues were sequentially treated with0.2M phosphate buffer solutions containing10%,20%,and30%sucrose, for30min each,and frozen-sectioned at a thickness of16l m. The sections were briefly washed in distilled water,immersed in 60%isopropanol for1min,and stained for15min with the oil red O solution.The sections were lightly counterstained with hematoxylin.Images were captured using a MVX10microscope (Olympus,Tokyo,Japan).2.4.cDNA cloning of torafugu LPLsA BLASTP search was run against the JGI Fugu genome database v4.0(</Takru4/Takru4.home.html>) using the amino acid sequence of human LPL(DDBJ/EMBL/ GenBank accession number NM_000237)as a probe.Based on a preliminary phylogenetic analysis of the genes screened,gene-specific primers trLPL1_5UTR,trLPL1_3UTR,trLPL2_5UTR,and trLPL2_3UTR were designed to clone the cDNAs of LPL1and LPL2 from torafugu(Supplementary Table1).Total RNA was extracted from the liver and skeletal muscle of torafugu using an RNeasy lipid tissue mini kit(Qiagen,Hilden,Germany).First strand cDNA was synthesized using approximately1l g of the total RNA and Super-Script III reverse transcriptase(Invitrogen,Carlsbad,CA,USA).For LPL1,PCR was carried out in a10l L reaction mixture con-taining approximately0.5l g offirst strand cDNA synthesized from liver total RNA,2pmol trLPL1_5UTR and trLPL1_3UTR primers, 1l L of10ÂPCR buffer for KOD-plus-neo(Toyobo,Osaka,Japan), 1.6nmol dNTP mix,1.2mM MgSO4,and0.1U of KOD-plus-neo DNA polymerase(Toyobo).Amplification was performed with the initial denaturation at94°C for2min followed by35cycles of 94°C for30s,temperature gradient from55to65°C for30s and 72°C for1.5min.The PCR products were mixed,diluted10times with sterilized water,and used as a template for the second PCR. The second PCR was carried out by the same method except for the total reaction volume of20l L and temperature gradient from 59to66°C.For LPL2,thefirst PCR was carried out as described above using primers trLPL2_5UTR and trLPL2_3UTR(Supplementary Table1) under the condition at94°C for2min followed by35cycles of 94°C for30s,temperature gradient from57to65°C and72°C for1.5min.The PCR products at the annealing temperature of 65°C were diluted10times with sterilized water and used as a template for the subsequent reamplification.The second PCR was performed in a reaction mixture containing1l L of the template, 5pmol of trLPL2_5UTR and trLPL2_3UTR primers,2l L of10ÂPCR buffer,4nmol dNTP mix and0.2U of Ex Taq DNA polymerase (Takara,Otsu,Japan).The PCR condition was at94°C for2min followed by30cycles of94°C for30s,65°C for30s and72°C for1min.Thefinal extension step was performed at72°C for 5min.The PCR products were cloned into pBlueScript KS(+)vector (Stratagene,La Jolla,CA,USA)or a pGEM-T vector(Promega,52G.Kaneko et al./General and Comparative Endocrinology184(2013)51–60Madison,WI,USA),and sequenced according to the method de-scribed previously(Furukawa et al.,2004).2.5.Sequence analysesA multiple sequence alignment of the amino acid sequences was produced using Clustal W(Thompson et al.,1994).The phylo-genetic tree was constructed using the maximum likelihood meth-od of MEGA5software(Tamura et al.,2011)with the best-fit substitution model(WAG+G).Bootstrap resampling analysis from 1000replicates was used to evaluate the internal branches.Signal peptide regions and putative N-glycosylation sites were predicted using SignalP version 3.0(<http://www.cbs.dtu.dk/services/Sig-nalP/>)and Gene Runner version3.05(Hasting Software,Hasting, USA),respectively.The exon–intron structures of the torafugu LPL1and LPL2were determined using the Spidey program (</IEB/Research/Ostell/Spidey/>).2.6.Quantitative real-time PCRTotal RNA extraction and the synthesis offirst strand cDNA were performed as described in Section2.4.Quantitative real-time PCR was performed using the7300realtime PCR system(Applied Biosystems,Foster City,CA,USA)and the TaqMan fast advanced master mix(Applied Biosystems),according to the manufacturer’s protocol.Gene-specific primers and TaqMan probes(Supplemen-tary Table2)were designed using Primer express software v2.0 (Applied Biosystems)from the DNA nucleotide sequences of tora-fugu LPL1and LPL2identified in the present study,and the se-quences of torafugu PPAR c(Kondo et al.,2007)and b-actin (Venkatesh et al.,1996)characterized in previous studies.RNA extraction from the liver,skeletal muscle,and visceral adipose tis-sue of red seabream,as well as the synthesis offirst strand cDNA and the quantitative real-time PCR analyses were performed as de-scribed previously(Hirano et al.,2011),using the gene-specific pri-mer pairs shown in Supplementary Table2.The relative mRNA levels were determined by the comparative Ct method using the b-actin gene as internal controls.Data were analyzed by Student’s t-test and one-way analysis of variance(ANOVA),followed by the Tukey–Kramer post hoc test using JMP7.0.2software(SAS Institute, Inc.,Cary,NC,USA).2.7.In situ hybridizationDIG-labeled RNA probes were synthesized using a DIG RNA labeling kit(Roche,Mannheim,Germany)corresponding to a 1906–2074bp DNA fragment of the red seabream LPL1gene (AB243791).Muscle blocks of approximately7mm on a side were fixed in4%PFA/PBS overnight and washed twice with PBS contain-ing0.1%(v/v)Tween-20(PBST).The muscle blocks were then sequentially treated with50%(v/v)methanol/PBST for5min,30% (v/v)methanol/PBST for5min,and PBST for5min.Protease K digestion was performed in PBST containing200l g/mL protease K(Takara)at room temperature for30min.After washing with PBST,the samples were prehybridized in the hybridization buffer (50%formamide,5ÂSSC,0.1%Tween-20,5l g/mL yeast RNA, and50l g/mL heparin)for1h and then hybridized using the DIG-labeled RNA probes at65°C overnight.The hybridized RNA probes were detected using alkaline phosphatase conjugated with an anti-DIG antibody and nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate(Roche).3.Results3.1.Lipid contentLipid content of the carcass,liver,and visceral adipose tissue of torafugu and red seabream are shown in Table1.Since torafugu has no apparent visceral adipose tissue,only the values of red sea-bream are shown for this organ.The lipid content of the torafugu carcass was0.8±0.1%,which was significantly lower than that of the red seabream:7.3±0.4%(Student’s t-test,P<0.0001).In con-trast,the lipid content of the liver of torafugu was significantly higher than that of red seabream;the precise values being 67.0±4.4%and8.8±2.7%,respectively(P<0.0001).The visceral adipose tissue of red seabream had the highest lipid content of 73.5±3.5%.The hepatosomatic indices for torafugu and red sea-bream were calculated to be10.2±1.3%and2.2±0.1%,respec-tively.All of these values were in agreement with those of previous reports(Ando et al.,1993;Koizumi and Hiratsuka,2009).Total amount of lipid in the carcass,liver,and visceral adipose tissue were calculated from the lipid content and tissue weight. In line with the values for lipid content described above,the tora-fugu carcass and liver contained less(P<0.0001)and more (P<0.05)lipid than that of red seabream,respectively.Of the three organs,red seabream contained the highest quantity of lipid in the carcass,suggesting that skeletal muscle is the primary site of lipid accumulation in this species despite the high lipid content of the visceral adipose tissue.3.2.Oil red O stainingThefillets of torafugu exhibited very little staining with oil red O(Fig.1A),whereas those of the red seabream had clearly stained myosepta,the layers of connective tissue in skeletal muscle (Fig.1F).This pattern of staining is consistent with Atlantic salmon Salmo salar(Zhou et al.,1995)and yellowtail Seriola quinqueradiataTable1Absolute and relative lipid contents of the carcass,liver and adipose tissue of torafugu and red seabream.Torafugu Red seabreamNo.1No.2No.3Mean±SD No.1No.2No.3Mean±SDLipid content(%)Carcass0.90.70.80.8±0.1**7.07.77.37.3±0.4 Liver71.262.467.367.0±4.4**11.8 6.58.08.8±2.7 Adipose tissue––––75.269.375.473.3±3.5Lipid weight(g)Carcass 1.8 1.3 2.4 1.8±0.6**24.628.727.627.0±2.1 Liver47.233.543.941.5±7.2* 1.60.8 1.1 1.2±0.4 Adipose tissue––––13.78.68.510.3±3.0 Total49.034.846.343.4±7.539.938.137.238.4±1.4Lipid weights were calculated by%lipid content x tissue weight.*P<0.05.**P<0.0001.G.Kaneko et al./General and Comparative Endocrinology184(2013)51–6053(Thakur et al.,2003).The oil red O staining of liver sections revealed that both torafugu and red seabream accumulate lipid in their liver(Fig.1B and G).The size of lipid droplets appeared lar-ger in torafugu than red seabream,possibly a reflection of the high lipid content in the liver of torafugu.Sections of torafugu skeletal muscle showed no lipid accumulation,whereas those from red sea-bream contained lipids in the intercellular spaces between muscle fibers(Fig.1C and H).Lipid accumulation under the mucous mem-branes of the intestine was observed for bothfish(Fig.1D and I). However,neither torafugu nor red seabream showed any apparent lipid accumulation in the connective tissues under their skin (Fig.1E and J).It was not possible to prepare sections of the red seabream adipose tissue due to its high lipid content.3.3.cDNA cloning of torafugu LPL genesIn silico screening using the amino acid sequence of human LPL as a probe resulted in the identification of four LPL-like genes on three scaffolds:two on scaffold135,one on scaffold50,and one on scaffold14.Since preliminary phylogenetic analysis revealed that the two genes on scaffold135were orthologous to LPL from other animals,these two genes were selected and cloned using primers designed from the sequences of the50-and30-untranslated regions in the JGI Fugu genome database(Supplementary Table1).PCR using primers trLPL1_5UTR and trLPL1_3UTR at an anneal-ing temperature of66°C amplified a single DNA fragment consist-ing of1675bp,later identified as the LPL1cDNA(Fig.2A).The FAGBHCIDJEseabream.Fillets(A,F)and frozen sections of liver(B,G),skeletal muscle(C,H),intestine(D,I)andbars represent1cm in A and F,and200l m in the other panels.(For interpretation of thethis article.)54G.Kaneko et al./General and Comparative Endocrinology184(2013)51–60genomic organization of torafugu LPL1was determined using the cDNA sequence and Fugu genome database(Fig.2B).The torafugu LPL1comprised11exons spanning approximately4kb of the geno-mic DNA,and no splicing variant was found.PCR using the primers trLPL2_5UTR and trLPL2_3UTR at an annealing temperature of59.9°C amplified two DNA fragments of different length,1702and2359bp(Fig.2A).These fragments were la-ter identified as LPL2a and LPL2b cDNAs,respectively.Analysis of the genomic organization of the torafugu LPL2revealed that a657bp se-quence encoded by thefirst intron of torafugu LPL2remained inserted between thefirst and second exons of the LPL2b cDNA(Fig.2B).The exon–intron boundary sequences of thefirst intron followed the GT-AG rule for a splicing site.Interestingly,this insertion sequence contained a stop codon near the50-terminus.Since thefirst exon of the torafugu LPL2encodes only33amino acids,the deduced amino acid sequence of LPL2b lacked several important regions including the catalytic triads and both the lipid and heparin binding sites.Phylogenetic analysis showed that the deduced amino acid se-quences of LPL1and LPL2a clearly separated into the LPL1and LPL2clades,respectively(Fig.3).Both LPL2a from torafugu and LPL2from red seabream were assigned to a monophyletic group orthologous to the one containing both mammalian LPL andfish LPL1.Similar analysis using the predicted amino acid sequences of the genes on scaffolds50and14revealed a monophyletic rela-tionship with the endothelial lipase(EL)and hepatic lipase(HL) clades,respectively.The deduced amino acid sequence of the torafugu LPLs con-tained putative signal peptides,N-glycosylation sites,and several other functional features assigned by comparison to the human LPL(Raisonnier et al.,1995),including sites corresponding to cofactor activation,catalytic triads,heparin binding sites,lipid binding sites,and conserved cysteine residues involved in the for-mation of disulfide bridges(Fig.4).Interestingly,the LPL2s from torafugu and red seabream had more amino acid substitutions in their cofactor activation sites and heparin binding sites than LPL1s. The torafugu LPL1showed only45%and47%amino acid identity to the LPL2s from torafugu and red seabream,respectively,but58–84%amino acid identity to LPLs from other animals.The DNA nucleotide sequences of torafugu LPL1and LPL2a were registered to DDBJ/EMBL/GenBank with accession numbers AB735414and AB735415,respectively.3.4.Relative mRNA levels of PPAR c,LPL1and LPL2The relative mRNA levels of PPAR c,LPL1,and LPL2in the liver, skeletal muscle,and visceral adipose tissue were measured by quantitative real-time PCR.In torafugu,the transcripts of PPAR c were detected only in the liver,reflecting the liver-specific lipid distribution of this species(Fig.5A).However,in red seabream, the transcripts of PPAR c were found in all the tissues examined, and a one-way ANOVA failed to detect any significant differences among their relative mRNA levels(Fig.5D,P=0.07).For torafugu,the relative mRNA levels of LPL1were significantly higher in the liver than in the skeletal muscle(Fig.5B),but for red seabream,the highest values were obtained from the visceral adi-pose tissue,followed by the liver and skeletal muscle(Fig.5E).In contrast,the transcripts of LPL2had a distribution similar to PPAR c,being detected only in the liver of torafugu(Fig.5C),but in all the tissues tested in red seabream with the statistically sig-nificant difference between the mRNA levels in the liver and skel-etal muscle(Fig.5F,P<0.05).Overall,the relative mRNA levels of LPL2were lower than those of LPL1in bothfish.In addition,ourG.Kaneko et al./General and Comparative Endocrinology184(2013)51–6055quantitative real-time PCR primers for torafugu LPL2were de-signed to amplify both the LPL2a and LPL2b cDNAs.Since only the LPL2a mRNA would be translated to the functional LPL2pro-tein,the protein level of LPL2might be markedly lower than that of LPL1in torafugu.These results suggest that LPL1,rather than LPL2,has a greater role in lipid accumulation in thesefish.The relative mRNA levels of red seabream LPL1and LPL2were gener-ally consistent with those from a previous study(Oku et al.,2006a).Another set of experiments was performed to investigate the variation of internal control gene among different tissues.Two internal control genes,b-actin and elongation factor1a(EF1a) genes,were used to ensure the accuracy of results(Supplementary Data1).The mRNA levels of torafugu b-actin showed little variation in the liver and skeletal muscle,whereas those of tora-fugu EF1a gene was about fourfold higher in the liver than skeletal muscle(Fig.S1).In red seabream,the mRNA levels of b-actin and56G.Kaneko et al./General and Comparative Endocrinology184(2013)51–60EF1a genes were similar to each other,but higher in adipose tissue than liver and skeletal muscle.These results suggest that the b-ac-tin gene is a suitable internal control gene in the above analysis, although the mRNA levels of target genes might be underestimated in adipose tissue.3.5.Localization of LPL1transcripts in the skeletal muscle of red seabreamIn situ hybridization with samples of the skeletal muscle of red seabream was used to localize the transcripts of LPL1,which were found to accumulate in the cells along the myosepta(Fig.6A). These cells had a large diameter up to100l m and a globular shape (Fig.6B),both of which are typical features of adipocytes.However,it was noted that not all myosepta were coupled with the adipo-cytes expressing LPL1transcripts.Further investigation using mus-cle blocks from a cross sectional view revealed that the transcripts of LPL1were also expressed in the pericellular connective tissues between musclefibers(Fig.6C).We could not observe any hybrid-ization signals of PPAR c and LPL2genes in the skeletal muscle of red seabream possibly due to their low expression levels(data not shown).4.DiscussionThe present study investigated the lipid distribution in torafugu and red seabream using conventional weighing and staining meth-ods.The lipid content of the liver,skeletal muscle,andvisceral G.Kaneko et al./General and Comparative Endocrinology184(2013)51–6057adipose tissue have been determined by the weighing method for bothfish(Ando et al.,1993;Koizumi and Hiratsuka,2009).Our re-sults for the liver and visceral adipose tissue were comparable to the previous reports.The lipid content of the carcasses,which consisted of skeletal muscle,skin,and the connective tissues under the skin,were also similar to those of skeletal muscle in previousthe skeletal muscle of red seabream.(A)Dorsolateral view of the red seabream skeletal muscle.(arrowheads).(B)Adipocytes along the myosepta showing the expression of LPL1transcripts.(C)connective tissues between musclefibers where the transcripts of LPL1were detected.The scale barsreferences to colour in thisfigure legend,the reader is referred to the web version of this article.)58G.Kaneko et al./General and Comparative Endocrinology184(2013)51–60reports.These results reflect the lack of an apparent subcutaneous adipose tissue in either of thesefish(Fig.1E and J).Together with the results of oil red O staining newly reported here,it was re-vealed that torafugu accumulates lipid predominantly in its liver, whereas red seabream has a more dispersed distribution of lipid within its visceral adipose tissue,liver,and myosepta in skeletal muscle.However,the presence of typical adipocytes was not de-tected in any of the frozen sections of skeletal muscle,even those from red seabream(Fig.1C and H).This probably resulted from the large size of the red seabream adipocytes,which measured approximately100l m in diameter in this study(Fig.6)and 60l m in a previous study(Ji et al.,2003),compared to a thickness of16l m in the sample sections.As a consequence of this limita-tion,muscle blocks of about7mm on a side were selected for in situ hybridization with the LPL1gene in red seabream instead of conventional tissue sections(see below).Such species-specific patterns of lipid distribution prompted the investigation of the expression of genes possibly related to lipid accumulation.Quantitative real-time PCR revealed a strong corre-lation between the tissue distribution of lipids and the relative mRNA levels of PPAR c and LPL1(Fig.5).Furthermore,the tran-scripts of LPL1were expressed in the adipocytes along the myosep-ta of red seabream(Fig.6).This observation suggests that LPL1in the cells along the myosepta promotes the uptake of lipids from the circulatory system,leading to lipid accumulation in skeletal muscle of red seabream.In contrast,torafugu expresses less PPAR c and LPL1in this location,and thus accumulates less lipid in its skeletal muscle.It is therefore likely that differences in lipid trans-port,rather than lipid synthesis,account for the species-specific li-pid distribution infish.The results of the present study also indicate that the skeletal muscle can be considered to be a mixture of at least two different types of tissue:muscle cells and the adipo-cytes along the myosepta.Other reports assessing the skeletal muscle from a variety offish species have also found evidence for the expression of genes associated with lipid accumulation, e.g.adiponectin in rainbow trout(Kondo et al.,2011),PPAR c in gilthead seabream(Leaver et al.,2005),LPL in common carp (Cheng et al.,2009),and leptin in Atlantic salmon(Ronnestad et al.,2010).However,to our knowledge,this is thefirst report to visualize the adipocytes infish skeletal muscle by in situ hybrid-ization.It is therefore interesting to observe the expression sites of other adipogenic genes in skeletal muscle in detail by in situ hybridization or immunostaining using muscle blocks or thick tis-sue sections.Alternatively,isolation of the adipocytes in skeletal muscle would enable us to investigate the expression of adipogenic genes by high-sensitive methods such as quantitative real-time PCR and the next generation sequencing.These approaches are especially useful for genes of low expression levels including PPAR c and LPL2.The present study also demonstrated that the relative mRNA levels of LPL2were markedly lower than those of LPL1for both torafugu and red seabream.The latter results were consistent with a previous study(Oku et al.,2006a).Although these results suggest that LPL1is the major LPL isoform infish,the role of LPL2in lipid accumulation requires further examination for several reasons. Firstly,thesefish-specific LPL isoforms have several substitutions in their cofactor and heparin binding sites,and thus might be ex-pected to have different functions from LPL1.Secondly,LPL activity is regulated at least at three levels—transcriptional,post-transcrip-tional,and post-translational(Wang and Eckel,2009)—and thus the low levels of LPL2mRNA do not necessarily infer the low enzy-matic activity.Thirdly,the mis-splicing of introns,as observed in torafugu LPL2,has also been reported in human LPL in association with a hereditary LPL deficiency disease(Gotoda et al.,1991;Holzl et al.,1994).Lastly,the expression of the LPL isoforms would be regulated in a species-specific manner.The mRNA levels of LPL2is recently reported to be higher than those of LPL1in gilthead sea-bream(Benedito-Palos et al.,2012).Further investigation of these isoforms is necessary to assess their contributions to lipid accumulation.The results of the quantitative real-time PCR,which detected transcripts of PPAR c in the skeletal muscle of red seabream,were in contrast to a previous study using RT-PCR that did not detect any expression of PPAR c(Oku and Umino,2008).This discrepancy probably resulted from the different number of PCR cycles used:30 and40for the RT-PCR and real-time PCR,respectively.Indeed,a similar study which used40cycles of quantitative real-time PCR was also able to detect the presence of PPAR c transcripts in the skeletal muscle of gilthead seabream(Leaver et al.,2005).The expression offish LPL and PPAR c is known to be regulated by nutritional status and several hormones such as insulin and growth hormone(GH)in a tissue-specific manner.Insulin,which is generally secreted by pancreatic b-cells in response to feeding, stimulated the transcription of LPL in adipose tissue of gilthead seabream(Albalat et al.,2007b)and stromal-vascular cells derived from adipose tissue of red seabream(Oku et al.,2006b).In myo-cytes of rainbow trout,insulin and GH activated LPL transcription, whereas these hormones showed little and inhibitory effects on PPAR c transcription,respectively(Cruz-Garcia et al.,2011).Taken together with our presentfindings,the distinct lipid distribution in torafugu and red seabream might be attributed to the sensitivity to insulin and GH of the adipocytes in skeletal muscle,or their pro-genitor cells.Further characterization of these cells would expand our understanding of the species-specific lipid distribution infish.In conclusion,a strong correlation was found between lipid dis-tribution and the relative mRNA levels of PPAR c and LPL1for both torafugu and red seabream.Furthermore,the transcripts of LPL1 co-localized with lipids in adipocytes along the myosepta in the skeletal muscle of red seabream.These results suggest that high expression of PPAR c and LPL1promotes lipid accumulation in skel-etal muscle of red seabream thereby accounting for the distinct pattern of lipid distribution in torafugu and red seabream. AcknowledgmentsWe express sincere gratitude to Prof.T.Kaneko,The University of Tokyo,for his kind advice regarding the oil red O staining of tis-sue sections.We would also like to thank Dr.S.Akimoto,Kanagawa Prefectural Fisheries Experimental Station,for his help in the sam-pling of red seabream.This study was partly supported by the Min-istry of Education,Science,Sports and Culture,Grant-in-Aid for Young Scientists(B),Nos.20780153and22780190,and Grant-in-Aid for Scientific Research(A),No.21248027.This work was also partly supported by the Program for Promotion of Basic Re-search Activities for Innovative Biosciences.Appendix A.Supplementary dataSupplementary data associated with this article can be found, in the online version,at /10.1016/j.ygcen.2013.01.003.ReferencesAlbalat,A.,Saera-Vila,A.,Capilla,E.,Gutierrez,J.,Perez-Sanchez,J.,Navarro,I., 2007a.Insulin regulation of lipoprotein lipase(LPL)activity and expression in gilthead sea bream(Sparus aurata).Comp.Biochem.Physiol.Part B Biochem.Mol.Biol.148,151–159.Albalat,A.,Saera-Vila,A.,Capilla,E.,Gutiérrez,J.,Pérez-Sánchez,J.,Navarro,I., 2007b.Insulin regulation of lipoprotein lipase(LPL)activity and expression in gilthead sea bream(Sparus aurata).Comp.Biochem.Physiol.Part B Biochem.Mol.Biol.148,151–159.G.Kaneko et al./General and Comparative Endocrinology184(2013)51–6059。