LYCAT, a homologue of C. elegans acl-8, acl-9, and acl-10, determines the fatty acid composition of phosphatidylinositol in mice.

Mammalian phosphatidylinositol (PI) has a unique fatty acid composition in that 1-stearoyl-2-arachidonoyl species is predominant. This fatty acid composition is formed through fatty acid remodeling by sequential deacylation and reacylation. We recently identified three Caenorhabditis elegans acyltransferases (ACL-8, ACL-9, and ACL-10) that incorporate stearic acid into the sn-1 position of PI. Mammalian LYCAT, which is the closest homolog of ACL-8, ACL-9, and ACL-10, was originally identified as a lysocardiolipin acyltransferase by an in vitro assay and was subsequently reported to possess acyltransferase activity toward various anionic lysophospholipids. However, the in vivo role of mammalian LYCAT in phospholipid fatty acid metabolism has not been well elucidated. In this study, we generated LYCAT-deficient mice and demonstrated that LYCAT determined the fatty acid composition of PI in vivo. LYCAT-deficient mice were outwardly healthy and fertile. In the mice, stearoyl-CoA acyltransferase activity toward the sn-1 position of PI was reduced, and the fatty acid composition of PI, but not those of other major phospholipids, was altered. Furthermore, expression of mouse LYCAT rescued the phenotype of C. elegans acl-8 acl-9 acl-10 triple mutants. Our data indicate that LYCAT is a determinant of PI molecular species and its function is conserved in C. elegans and mammals.

plementary Table I ), in which the members share four conserved AGPAT motifs that are involved in substrate binding and catalysis ( 15,16 ) (supplementary Fig. I , motif I-IV). In mammals, the AGPAT family consists of at least 16 members. Among these members, lysocardiolipin acyltransferase (LYCAT), also known as LCLAT1 and ALCAT1, is the closest homolog of C. elegans ACL-8, ACL-9, and ACL-10 (supplementary Table I ) ( 14 ). ACL-8, ACL-9, ACL-10 and mammalian LYCAT possess highly conserved amino acids in the AGPAT motifs, which are unique in LYCAT/ACL-8, -9, -10 subfamily members, but not in other AGPAT family members (supplementary Fig. 1B , amino acids indicated in blue). Acyltransferases with these highly conserved amino acids are evolutionarily conserved in various species including human, zebrafi sh, and C. elegans , but not in yeast.
ACL-8, ACL-9, and ACL-10 belong to the 1-acylglycerol-3-phosphate O -acyltransferase (AGPAT) family ( 14 ) (sup- Three pairs of daughter cells are shown with brackets. In wild-type worms, all seam cells divide parallel to the A-P axis (B). In acl-8 acl-9 acl-10 triple mutants, the seam cell division is abnormally oriented relative to the A-P axis (C; asterisks). The parallel orientation of seam cell divisions is restored in acl-8 acl-9 acl-10 triple mutants expressing mouse LYCAT under the control of the epithelial-specifi c dpy-7 promoter (D). E-G: Seam cells at the adult stage visualized by scm::gfp . Merged fl uorescence and differential interference contrast images are shown. The letters (a), (b) and (c) correspond to those of (A). E: Wild-type. The anterior daughter cells lose the expression of scm::gfp , and the posterior daughter cells adopt the seam cell fate again and continue to express scm::gfp in all three lineages. In acl-8 acl-9 acl-10 triple mutants, the seam cell division becomes symmetric (both daughter cells adopted the seam cell fate) or is reversed (the fates of daughter cells are the opposite of what they are in the wild-type) (F), and the defects in asymmetric cell-fate determination are rescued by expression of mouse LYCAT (G). Scale bars are 20 m.
labeled Chemicals (St. Louis, MO). Rhizopus arrhizus lipase and phospholipase A 2 from honey bee venom were purchased from Sigma-Aldrich. DEAE column was purchased from Wako Pure Chemical Industries (Osaka, Japan).

Preparation of sn-2-acyl lysophospholipids
Each sn -2-acyl-1-lysophospholipid ( sn -2-acyl lysophospholipids) was prepared as follows using dioleoyl PC, dioleoyl PE, 1-palmitoyl-2-oleoyl PS, bovine liver PI, or egg yolk PG. Each phospholipid (1 µmol) dissolved in 1 ml of diethyl ether was incubated with 100 µl of 100 mM CaCl 2 , 600 µl of 50 mM Trismalate, pH 5.7, and 400 µl of enzyme solution containing 15 mg of Rhizopus arrhizus lipase for 1 h at room temperature while stirring vigorously. After the incubation, the reaction was terminated by adding 1 ml of methanol. Remaining phospholipids and liberated fatty acids were removed by three extractions with 4 ml of diethyl ether-petroleum ether (1:1, v/v). sn -2-acyl lysophospholipid present in the lower layer was extracted by the method of Bligh and Dyer ( 22 ) and further purifi ed by TLC on silica gel 60 plates (Merck Biosciences, Darmstadt, Germany) in chloroform-methanol-acetic acid (65:25:13, v/v). The area of silica gel corresponding to sn -2-acyl lysoPI, and lysophosphatidylglycerol (lysoPG) in addition to lysoCL in vitro (18)(19)(20). Because C. elegans homologs of LYCAT specifi cally determines the fatty acyl chain at the sn -1 position of PI in vivo ( 12 ), here we examined if LYCAT determines the fatty acid composition of PI in mammals. We generated LYCAT-defi cient mice, analyzed the fatty acid composition of phospholipids in the tissues, and found that LYCAT defi ciency markedly altered the fatty acid composition of PI, but not that of other phospholipids, in all tissues analyzed. We propose that LYCAT determines the fatty acid composition of PI via fatty acid remodeling and that its function is evolutionarily conserved in C. elegans and mammals.  ) mice. A: Schematic representation of the LYCAT genetargeting strategy. Homologous recombination of the targeting vector with the wild-type allele results in the replacement of 0.5-kilobase of the LYCAT genomic sequence containing exon 4 with the neo cassette, deleting the AGPAT motifs II and III that are essential for acyltransferase activity in AGPAT family members ( 15,16 ). A thymidine kinase gene (TK) was used for negative selection of clones with random integration of the targeting vector. B: PCR-based genotyping of genomic DNA from LYCAT

Materials
The orientation and location of PCR primers are indicated with arrows in (A). C: Western blot analysis using an anti-mouse LYCAT antibody, YN1. Total protein extracts of liver from LYCAT +/+ and LYCAT Ϫ / Ϫ mice were subjected to SDS-PAGE, blotted onto a PVDF membrane, and stained with YN1.
was linearized and electroporated into RENKA ES cell lines (TransGenic Inc., Kumamoto, Japan) derived from the C57BL/6 mouse substrain. Stable clones were selected for G418 resistance.
Out of 569 G418-resistant clones, three clones were correctly targeted, as confi rmed homologous recombination by Southern blot analysis with 5 ′ , 3 ′ , and neo probes. Blastocysts obtained by aggregation methods of ICR morula embryos and three different ES cells to recipient uterus as after 3 days of pseudopregnancy. Chimeric males were mated to C57BL/6 females and offspring animals (F1) were genotyped by Southern blot analysis. F2 homozygous mutant mice were then generated by intercrossing F1 heterozygous males and females. Mice were genotyped by PCR analyses of genomic DNA isolated from tail biopsies. The following primers were used for genotyping: primer 1, 5 ′ -GTG GTA CTG CAC ACC TTT GAT CC-3 ′ ; primer 2, 5 ′ -CAA AAT GGC TCT TAT CAT CCT TCC-3 ′ ; and primer 3, 5 ′ -GAA CAG AGT ACC TAC ATT TTG AAT GG-3 ′ . The wild-type allele is detected with primers 1 and 2, yielding a 356 bp product, whereas the mutant allele is detected with primers 1 and 3, yielding a 498 bp fragment. The animals used were LYCAT +/+ (wild-type) and LYCAT Ϫ / Ϫ (homozygous) mice in the same C57BL/6 background between 13 and 20 weeks old. The mice were kept in a temperature-controlled environment with a 12 h light and 12 h dark cycle and received a standard diet and water ad libitum.

Generation of monoclonal antibodies against mouse LYCAT
A recombinant mouse LYCAT (amino acid numbers 78-298 of mouse LYCAT, GenBank accession number NM_001081071) that was expressed and purifi ed by Escherichia coli pCold TF expression system (TaKaRa, Japan) was injected into the hind foot pads of WKY/ Izm rat strain by using Freund's complete adjuvant. The enlarged medial iliac lymph nodes were used for cell fusion with mouse myeloma cells, PAI. In the present study, the established monoclonal antibody, named YN1, was used for Western blotting and immunocytochemistry at 1:2,000 and 1:100 dilutions, respectively.

Western blot
Murine tissues were homogenized in quadruple volumes (w/v) of SET buffer with protease inhibitors (0.5 mM phenylmethylsulfonyl fl uoride, 2 µg/ml pepstatin, 2 µg/ml leupeptin, 2 µg/ml aprotinin). After centrifugation at 1,000 g for 10 min at 4°C, the supernatants were used as the total protein extracts. The protein concentrations of samples were determined by the bicinchoninic lysophospholipid was scraped off the plates. sn -2-acyl lysophospholipid was reextracted by methanol and immediately used for acyltransferase assays.

Acyltransferase assay
Mouse tissues were pulverized under liquid nitrogen and homogenized in quadruple volumes (w/v) of SET buffer (10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 250 mM sucrose) with protease inhibitors (5 g/ml pepstatin, leupeptin, and aprotinin). After centrifugation at 2,330 g for 20 min at 4°C, the resulting supernatant was further centrifuged at 105,000 g for 60 min. The resulting pellet (microsomal fraction) was resuspended in homogenizing buffer (50 mM potassium phosphate buffer (pH 7.0) containing 0.15 M KCl, 0.25 M sucrose) and used for the enzyme assay. Acyl-CoA: sn -2-acyl lysophospholipid acyltransferase assay and acyl-CoA: sn -1-acyl lysophospholipid acyltransferase assay were performed essentially as described previously ( 12 ), except that the amount of microsomal protein was 10 µg and the incubation temperature was 37°C. Because sn -2-acyl lysophospholipid is known to easily isomerize to sn -1-acyl lysophospholipid, it is possible that the acyl donor is incorporated into sn -1-acyl lysophospholipid and sn -2acyl lysophospholipid. Thus, an accurate measure of sn -2-acyl lysophospholipid acyltransferase activity could only be obtained by determining the position that had been acylated. To check the positional specifi city, the radiolabeled product was treated with bee venom phospholipase A 2 (supplementary Fig. II ). The distribution of radioactivity between the resultant sn -1-acyl lysophospholipid and free fatty acid was assessed after TLC ( 12 ).

Generation of LYCAT-defi cient mice
Mouse LYCAT genomic DNA sequences were cloned from a C57BL/6-derived genomic DNA and subcloned into the pPNT vector. The targeting vector contains a PGK-neo cassette with 2.3 kb of LYCAT homologous regions upstream of exon 4 and with 5.9 kb homologous regions downstream of exon 4. The targeting vector Fig. 3. Acyltransferase activity that incorporates stearic acid into the sn -1 position of PI in the membrane fractions from liver, heart, skeletal muscle, and brain. [ 14 C]stearoyl-CoA and sn -2-acyl lysoPI were used as the acyl donor and acceptor, respectively. * P < 0.01, ** P < 0.001. Data represent the mean ± SEM of triplicate measurements. The acyltransferase reaction mixtures contained 80 µM sn -2-acyl lysophospholipid, 12.5 µM [ 14 C]stearoyl-CoA, and 10 µg of liver microsomal protein in a total volume of 0.8 ml assay buffer. Each 14 Clabeled phospholipid produced after the incubation was treated again with phospholipase A 2 . The distribution of radioactivity between the resultant lysophospholipid and free fatty acid was taken as a measure of acylation at the sn -1 position ( sn -2-acyl lysophospholipid acyltransferase activity) ( a ) and the sn -2 position ( sn -1-acyl lysophospholipid acyltransferase activity) ( b ), respectively. Data represent the mean ± SEM of triplicate measurements.  CII-test Wako kit (Wako Pure Chemical Industries, Osaka, Japan). Plasma triglyceride and cholesterol levels were measured using a Triglyceride E-test Wako kit and cholesterol E-test Wako kit (Wako Pure Chemical Industries, Osaka, Japan), respectively. To measure hepatic triglyceride contents, livers were homogenized and lipids were extracted by the method of Bligh and Dyer ( 22 ) and then suspended in a buffer containing Triton-X100. The triglyceride content was measured using a Triglyceride E-test Wako kit.

RESULTS
Mouse LYCAT is a functional homolog of C. elegans acl-8 , acl-9 , and acl- 10 We previously reported that C. elegans acl-8 acl-9 acl-10 triple mutants are defective in the divisions of specialized epithelial cells called seam cells ( 12 ). In the wild-type, 10 seam cells (H0-2, V1-6 and T in Fig. 1A ) are present on each side of a newly hatched worm and divide asymmetrically in a stem cell-like manner during the larval stage. Both daughter cells just after the division express a seam cell marker, scm::gfp ( Fig. 1B ). Then, an anterior daughter acid (BCA) assay (Pierce). Each sample (20 µg protein/lane) was subjected to SDS-PAGE and immunoblotting. The following primary antibodies were used: anti-mouse LYCAT monoclonal antibody, anti-mouse LPIAT1 monoclonal antibody, and anti-GAPDH monoclonal antibody (6C5, Calbiochem).

Phospholipid analysis
Lipids of each tissue were extracted by the method of Bligh and Dyer ( 22 ). Phospholipids were separated from total lipids by one-dimensional TLC on silica gel 60 plates in chloroform-methanol-acetic acid (65:25:13, v/v). The area of silica gel corresponding to each phospholipid (PC, PE, PG, CL, and PI+PS) was scraped off the plates. The PI+PS fraction was reextracted, separated by TLC in chloroform-methanol-formic acid-water (60:30:7:3, v/v), and the areas of silica gel corresponding to PI and PS were scraped off the plates. Isolated phospholipids were methylated with 2.5% H 2 SO 4 in methanol. The resulting fatty acid methyl esters were then extracted with hexane and subjected to GC-MS analysis as described previously ( 23 ). Liquid chromatographyelectrospray ionization mass spectrometry (LC-ESIMS) analysis was performed as described previously ( 12 ) using 1,2-dipalmitoyl PI (16:0/16:0-PI), as internal standards. Phospholipids in the mouse liver were quantifi ed as follows. Each phospholipid (PC, PE, PI, PS, PG, and CL) was separated from total lipids by TLC as described above. Each lipid spot on the TLC plate was scraped off the plate, and the spots of each phospholipid were then extracted from silica gel by chloroform-methanol-water (1:2:0.8, v/v). The amount of inorganic phosphate in each phospholipid was quantifi ed as described previously ( 24 ).
were grown on coverslips for 24 h. To achieve mitochondrial staining, MEFs were incubated in the growth medium with MitoTracker Red CM-H 2 XMRos (Molecular Probes) and fi xed with 4% paraformaldehyde for 15 min at room temperature. After washing with PBS, MEFs were permeabilized for 10 min with 0.5% Triton X-100 in PBS. MEFs were then incubated with the anti-mouse LYCAT antibody YN1. As for endoplasmic reticulum (ER) staining, fi xed and permeabilized cells were incubated with the rabbit anti-calnexin polyclonal antibody (1:500) (StressGen Biotechnologies, Victoria, BC, Canada) and YN1. Confocal images were obtained using a Zeiss LSM510META confocal microscope system (Carl Zeiss MicroImaging, Thornwood, NY).

Blood biochemistry and measurement of plasma and liver lipid levels
Serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were measured using a Transaminase dpy-7 promoter in acl-8 acl-9 acl-10 triple mutants. Expression of mouse LYCAT restored the parallel orientation of seam cell divisions in acl-8 acl-9 acl-10 triple mutants ( Fig.  1D ). The angle between the A-P axis and the direction of cell division fell within 10 degrees in most seam cells in LYCAT-expressing mutants. Expression of mouse LYCAT also effectively rescued the defects in cell-fate determination of the daughter cells in acl-8 acl-9 acl-10 triple mutants ( Fig. 1G ). These results indicate that LYCAT is a functional homolog of C. elegans acl-8 , acl-9 , and acl-10 .

LYCAT defi ciency causes reduced acyltransferase activity toward the sn -1 position of PI and PG in mouse liver
We generated LYCAT-defi cient mice by gene targeting. A LYCAT-targeting construct was designed to replace exon cell fuses with a major epithelial syncytium and loses the expression of scm::gfp , and a posterior daughter cell assumes the seam cell fate again and continues to express scm::gfp ( Fig. 1E ). Seam cell division is oriented parallel to the anterior-posterior axis (A-P axis) as judged by scm::gfp , which is expressed in the nuclei of both daughter cells ( Fig. 1B ). In acl-8 acl-9 acl-10 triple mutants, however, the seam cell division is randomly oriented relative to the A-P axis ( Fig. 1C ) and the cell-fate determination became symmetric (both daughter cells adopted the seam cell fate) or was reversed (the fates of daughter cells were the opposite of what they are in the wild-type) ( Fig.  1F ). To determine whether mammalian LYCAT can substitute for acl-8 , acl-9 , and acl-10 in C. elegans , we expressed mouse LYCAT under the control of the epithelial-specifi c C. elegans acl-10 shows acyltransferase activity that incorporates stearic acid into the sn-1 position of PI in an in vitro assay ( 12 ). To assess the contribution of LYCAT to the acyltransferase activity toward the sn -1 position of phospholipids, we measured [ 14 C]stearoyl-CoA: sn -2-acyl lysophospholipid acyltransferase activity using the liver membrane fraction of LYCAT Ϫ / Ϫ mice. Because sn -2-acyl lysophospholipid is known to easily isomerize to sn -1-acyl lysophospholipid, it is possible that the [ 14 C]acyl donor is incorporated into both sn -1-acyl lysophospholipid and sn -2-acyl lysophospholipid (supplementary Fig. II ). To check the positional specifi city, [ 14 C]phospholipid produced after the incubation was treated again with phospholipase A 2 (for details, see Materials and Methods). The amount of radioactivity of the resultant lysophospholipid was taken as a measure of acylation at the sn -1 position ( sn -2-acyl lysophospholipid acyltransferase activity) and the amount of radioactivity of the free fatty acid was taken as a measure of acylation at the sn -2 position ( sn -1-acyl lysophospholipid acyltransferase activity). The sn -1-and sn -2-acyl lysophospholipid acyltransferase activities toward various lysophospholipids in the liver microsomes from wild-type and LYCAT Ϫ / Ϫ mice are summarized in Table 1 . At least 90% of [ 14 C] stearoyl-CoA was incorporated into the sn -1 position of all the lysophospholipids tested ( Table 1 ), indicating that sn -2-acyl lysophospholipid acyltransferase activity was predominant under the present assay conditions. In the liver from LYCAT Ϫ / Ϫ mice, sn -2-acyl lysoPI acyltransferase (LPIAT) activity was reduced by 64% (from 8.10 ± 0.11 to 2.92 ± 0.22 nmol/min/mg protein) and lysoPG acyltransferase (LPGAT) activity was reduced by 56% (from 10.45 ± 0.05 to 4.57 ± 0.17 nmol/min/mg protein). The sn -2-acyl lysoPE acyltransferase (LPEAT) activity was also reduced, but only slightly. In contrast, sn -2-acyl lysoPC acyltransferase (LPCAT) and lysoPS acyltransferase (LPSAT) activities were unchanged in the liver from LYCAT Ϫ / Ϫ mice. These results indicate that LYCAT signifi cantly contributes to acyltransferase activity toward the sn -1 position of PI and PG in mouse liver.

Reduction of acyltransferase activity toward the sn -1 position of PI in various tissues of LYCAT ؊ / ؊ mice
As described above, LYCAT was largely responsible for the acyltransferase activity that incorporates stearic acid into the sn -1 position of PI in the liver ( Table 1 ). We also examined acyltransferase activity toward the sn -1 position of PI in other tissues such as heart, skeletal muscle, and brain. As shown in Fig. 3 , a signifi cant portion of the activity was decreased in these tissues from LYCAT Ϫ / Ϫ mice. Especially in the heart and skeletal muscle, LYCAT was the main acyltransferase that incorporates stearic acid into the sn -1 position of PI.

LYCAT defi ciency causes a major change in the fatty acid composition of PI
To determine the role of LYCAT in phospholipid metabolism in vivo, we analyzed the fatty acid composition of phospholipids in tissues from LYCAT Ϫ / Ϫ mice. GC-MS analysis revealed that LYCAT defi ciency signifi cantly altered the fatty acid composition of PI in all the tissues examined including liver, heart, skeletal muscle, and brain ( Fig. 4 ). In LYCAT Ϫ / Ϫ mice, the amounts of stearic acid in PI of the liver, heart, and skeletal muscle were reduced by 42%, 36%, and 66% compared with the amounts in wildtype mice, respectively ( Fig. 4A-C ). The amount of arachidonic acid, which is a major fatty acid in the sn -2 position of PI, was also reduced in these tissues, though the degree of reduction was less than that of stearic acid. Conversely, the amounts of palmitic acid (16:0), oleic acid (18:1n-9), cis-vaccenic acid (18:1n-7), and linoleic acid (18:2n-6) in PI in these tissues increased. In the brains of LYCAT Ϫ / Ϫ mice, the change in the fatty acid composition of PI was 0.60 ± 0.07 to 0.26 ± 0.03 nmol/min/mg protein) ( Table 1 ). This observation is consistent with the previous data that small interfering RNA (siRNA)-mediated knockdown of LYCAT in HeLa cells results in reductions of acyltrans- Fig. 9. Effect of LYCAT defi ciency on acyltransferase activity that incorporates arachidonic acid into the sn -2 position of PI. A: Acyltransferase activities were measured using the membrane fractions from liver, heart, skeletal muscle, and brain. [ 14 C]arachidonoyl-CoA and sn -1-acyl lysoPI were used as the acyl donor and acceptor, respectively. Data represent the mean ± SEM of triplicate measurements. B: Expression level of LPIAT1 in the liver was measured with Western blot analysis using an anti-mouse LPIAT1 monoclonal antibody. Total protein extracts of liver from LYCAT +/+ and LYCAT Ϫ / Ϫ mice were subjected to SDS-PAGE, blotted onto a PVDF membrane, and stained with anti-LPIAT1 antibody FT10. GAPDH was used as a loading control. acid ( Fig. 4A-C ). Arachidonic acid is the predominant fatty acid at the sn -2 position of PI in mammals and is incorporated into PI by LPIAT after de novo synthesis (5)(6)(7)(8).
We fi rst investigated whether LYCAT contributes to acyltransferase activity that incorporates arachidonic acid into the sn -2 position of PI in liver, heart, skeletal muscle, and brain. Acyltransferase assays were performed using [ 14 C]arachidonoyl-CoA and sn -1-acyl lysoPI as substrates.
As shown in Fig. 9A , [ 14 C]arachidonoyl-CoA: sn -1-acyl LPIAT activity was not altered signifi cantly in all the LYCAT Ϫ / Ϫ tissues examined. Furthermore, the expression level of LPIAT1, a major acyltransferase that incorporates arachidonic acid into the sn-2 position of PI ( 11 ), was not changed in LYCAT Ϫ / Ϫ liver ( Fig. 9B ). These results indicate that LYCAT does not contribute to acyltransferase activity that incorporates arachidonic acid into the sn -2 position of PI in mouse tissues.
LYCAT is localized to the ER Cao et al. ( 17 ) reported that fl ag-tagged mouse LYCAT transfected into COS-7 cells displayed a perinuclear and punctate pattern that colocalized well with an ER marker but not with a mitochondrial marker. However, using the same type of cells, another study recovered recombinant fl ag-tagged mouse LYCAT from mitochondria and mitochondria-associated membranes but not from the microsomes ( 26 ). In the present study, we established a mouse LYCAT-specifi c monoclonal antibody YN1 and used it to analyze the subcellular localization of endogenous LYCAT. YN1 did not stain LYCAT Ϫ / Ϫ MEFs, indicating that YN1 specifi cally recognized mouse LYCAT ( Fig. 10 ). Wild-type MEFs were costained with MitoTracker Red dye (a mitochondrial selective probe) or were costained with an anticalnexin antibody that recognizes the calnexin ER-specifi c protein and the LYCAT antibody. We found that anti-LY-CAT staining overlapped substantially with anti-calnexin but not with MitoTracker ( Fig. 10 ), suggesting that endogenous LYCAT is localized to the ER but not to the mitochondria.

DISCUSSION
Mammalian LYCAT is reported to possess acyltransferase activity toward the sn -2 position of anionic lysophospholipids including lysoPI, lysoPG, and lysoCL in vitro (17)(18)(19)(20). In this study, we examined the fatty acid composition of each phospholipid from various tissues of LYCAT Ϫ / Ϫ mice, and showed that LYCAT determines the fatty acid composition of PI in vivo. 1-Stearoyl-2-arachidonoyl (18:0/20:4) PI is a major molecular species of PI in mammals. The amount of stearic acid in PI was greatly reduced in all the tissues examined from LYCAT Ϫ / Ϫ mice. Acyltransferase activity that incorporates stearic acid into the sn -1 position of PI was also reduced in the tissues from LYCAT Ϫ / Ϫ mice, whereas acyltransferase activity that incorporates arachidonic acid into the sn -2 position of PI was not altered. 1-Stearoyl-2-arachidonoyl PI is thought to be formed through fatty acid remodeling that is regulated by phospholipase A and lysoPI acyltransferase ( 5-10, 27, milder than that in the liver, heart, and skeletal muscle, and the amount of arachidonic acid in PI was not affected ( Fig. 4D ). Similarly, the amount of arachidonic acid in PI was not affected in the testis of LYCAT Ϫ / Ϫ mice ( Fig. 5 ). Meanwhile, the amount of stearic acid in PI of the testis was reduced by 66% and the amounts of palmitic acid (16:0) and cis-vaccenic acid (18:1n-7) in PI increased. In contrast to the remarkable change in the fatty acyl chains of PI, no drastic changes were observed in the fatty acid composition of other phospholipids such as PC, PE, PS, PG, and CL ( Fig. 4 ). One exception was a relatively large change in the fatty acid composition of PG in skeletal muscle; i.e., the amount of cis-vaccenic acid (18:1n-7) was reduced and instead, the amount of oleic acid (18:1n-9) increased considerably in LYCAT Ϫ / Ϫ mice ( Fig. 4C ). These results indicate that LYCAT mainly determines the fatty acid composition of PI in vivo.
We next analyzed the molecular species of PI in LYCAT Ϫ / Ϫ mice by LC-ESIMS. In the liver of LYCAT Ϫ / Ϫ mice, 18:0/20:4-PI, the major molecular species of PI, was signifi cantly decreased, and conversely, 16:0/20:4-PI, 16:0/18:2-PI, and 18:1/18:2-PI were increased compared with those in the liver of wild-type mice ( Fig. 6 ). The ratio of PI to total phospholipids was signifi cantly reduced by 35% compared with the ratio in wild-type mice as judged by quantifi cation of phospholipid phosphorus ( Fig. 7 ). On the other hand, the content and molecular species of PC and PE were not affected in LYCAT Ϫ / Ϫ mice ( Fig. 7 and data not shown). These results confi rm the notion that LYCAT plays an important role in determining the molecular species of PI in mice. A defi ciency in the reacylation reaction of lysoPI may reduce the PI content.

LYCAT defi ciency does not affect acyltransferase activity that incorporates arachidonic acid into the sn -2 position of PI
In the liver, heart, and skeletal muscle of LYCAT Ϫ / Ϫ mice, the amount of arachidonic acid in PI was also reduced, though the degree of reduction was less than that of stearic 28 ). Thus, LYCAT should serve as a lysoPI acyltransferase that catalyzes the incorporation of stearic acid into the sn -1 position of PI in the fatty acid remodeling pathway in vivo. Furthermore, mammalian LYCAT can substitute for the C. elegans homologs acl-8 , acl-9 , and acl-10 because expression of mouse LYCAT effi ciently rescued the seam cell defects observed in acl-8 acl-9 acl-10 triple mutants, in which the fatty acid composition of the sn -1 position of PI is selectively altered ( 12 ). Our immunocytochemical analysis using MEFs revealed that endogenous LYCAT is mainly localized to the ER, suggesting that the fatty acid remodeling of PI by LYCAT occurs in the ER membrane.
In the membrane fraction of LYCAT Ϫ / Ϫ mouse liver, acyltransferase activity that incorporates stearic acid into the sn -1 position of PI was reduced to one-third of that in wild-type mice, whereas acyltransferase activities that incorporate stearic acid into the sn -1 position of other phospholipids such as PC, PE, and PS, were not reduced appreciably in the liver of LYCAT Ϫ / Ϫ mice, which suggests that one or more enzymes other than LYCAT contribute to the activities. In fact, the amounts of stearic acid in phospholipids such as PE and PS were not changed in LYCAT Ϫ / Ϫ mice. In addition, acyltransferase activity that incorporates stearic acid into the sn -1 position of PI remained in the membrane fraction of LYCAT Ϫ / Ϫ mice, indicating the existence of other acyltransferase(s) that incorporates stearic acid into the sn -1 position of PI. Acyltransferase activity toward lysoPG was reduced signifi cantly in the membrane fraction of LYCAT Ϫ / Ϫ mouse liver. Indeed, we observed a relatively large change in the fatty acid composition of PG in the skeletal muscle of LYCAT Ϫ / Ϫ mice. These results indicate that LYCAT partially determines the fatty acid composition of PG, at least in certain organs.
In this study, we found that the amount of arachidonic acid in PI was also reduced in the liver, heart, and skeletal muscle of LYCAT Ϫ / Ϫ mice. Acyltransferase activity that incorporates arachidonic acid into the sn -2 position of PI was not reduced in these tissues. Moreover, the expression level of LPIAT1, a major acyltransferase that incorporates arachidonic acid into the sn-2 position of PI ( 11 ), was not changed. Thus, the reduction of the amount of arachidonic acid in PI of LYCAT Ϫ / Ϫ mice may be secondary to the reduction of stearic acid at the sn -1 position of PI. One possible explanation for the reduced arachidonic acid in PI is that LPIAT1 preferentially incorporates arachidonic acid when stearic acid is attached at the sn -1 position of lysoPI. In LYCAT Ϫ / Ϫ mice, stearic acid attached to PI is replaced by other fatty acids, such as palmitic acid and oleic acid, which may lead to reduced incorporation of arachidonic acid into PI.
Unlike the liver, heart, and skeletal muscle, the brain and testis of LYCAT Ϫ / Ϫ mice did not show any signifi cant changes in the amount of arachidonic acid in PI. It is interesting to note that these tissues possess CDS1 ( 29,30 ), one isoform of CDP-diacylglycerol synthase, which catalyzes the synthesis of CDP-diacylglycerol, a direct precursor of PI, from phosphatidic acid (PA). CDS1 prefers arachidonic acid-containing PA as a substrate ( 29 ) and therefore, CDS1 could have a role in increasing the arachidonic acid content of PI to compensate for the LYCAT defi ciency in these tissues.
Recently, Li et al. ( 26 ) reported that the fatty acid composition of CL was altered in LYCAT Ϫ / Ϫ heart under highfat diet conditions. In their study, however, the fatty acid composition of PI was not presented. The present results show that the fatty acid composition of PI was altered signifi cantly in LYCAT Ϫ / Ϫ mice, whereas that of CL was less affected under normal diet conditions. So far, we have not observed any apparent abnormalities in LYCAT Ϫ / Ϫ mice under normal diet conditions. However, it is interesting that insulin-stimulated Akt phosphorylation was significantly enhanced in tissues from LYCAT Ϫ / Ϫ mice under high-fat diet conditions ( 26 ). PI plays important roles in intracellular signaling cascades, such as the PI3K/Akt pathway through distinct phosphorylated derivatives of the inositol head group ( 1,2 ). Further analysis of LYCAT Ϫ / Ϫ mice should reveal the biological signifi cance of PI molecular species and/or the fatty acid remodeling of PI in mammals.