Novel Lysophospholipid Acyltransferase PLAT1 of Aurantiochytrium limacinum F26-b Responsible for Generation of Palmitate-Docosahexaenoate-Phosphatidylcholine and Phosphatidylethanolamine

N-3 polyunsaturated fatty acids (PUFA), such as docosahexaenoic acid (DHA, 22:6n-3), have been reported to play roles in preventing cardiovascular diseases. The major source of DHA is fish oils but a recent increase in the global demand of DHA and decrease in fish stocks require a substitute. Thraustochytrids, unicellular marine protists belonging to the Chromista kingdom, can synthesize large amounts of DHA, and, thus, are expected to be an alternative to fish oils. DHA is found in the acyl chain(s) of phospholipids as well as triacylglycerols in thraustochytrids; however, how thraustochytrids incorporate DHA into phospholipids remains unknown. We report here a novel lysophospholipid acyltransferase (PLAT1), which is responsible for the generation of DHA-containing phosphatidylcholine and phosphatidylethanolamine in thraustochytrids. The PLAT1 gene, which was isolated from the genomic DNA of Aurantiochytrium limacinum F26-b, was expressed in Saccharomyces cerevisiae, and the FLAG-tagged recombinant enzyme was characterized after purification with anti-FLAG affinity gel. PLAT1 shows wide specificity for donor substrates as well as acceptor substrates in vitro, i.e, the enzyme can adopt lysophosphatidylcholine, lysophosphatidylethanolamine, lysophosphatidylserine and lysophosphatidylinositol as acceptor substrates, and 15:0/16:0-CoA and DHA-CoA as donor substrates. In contrast to the in vitro experiment, only lysophosphatidylcholine acyltransferase and lysophosphatidylethanolamine acyltransferase activities were decreased in plat1-knockout mutants, resulting in a decrease of 16:0-DHA-phosphatidylcholine (PC) [PC(38∶6)] and 16:0-DHA-phosphatidylethanolamine (PE) [PE(38∶6)], which are two major DHA-containing phospholipids in A. limacinum F26-b. However, the amounts of other phospholipid species including DHA-DHA-PC [PC(44∶12)] and DHA-DHA-PE [PE(44∶12)] were almost the same in plat-knockout mutants and the wild-type. These results indicate that PLAT1 is the enzyme responsible for the generation of 16:0-DHA-PC and 16:0-DHA-PE in the thraustochytrid.

Studies on the benefit of n-3 PUFA to human health have led to the use of fish oils in supplements and medicine; however, a recent increase in the global demand of n-3 PUFA and decrease in fish stocks mean that an alternative source to fish oils is needed [9]. Furthermore, it was reported that dietary n-3 PUFAs administered in phospholipid (PL) form were superior to those in triacylglycerol form for maintaining a healthy metabolic profile [10]. Fish oils include n-3 PUFA-containing triacylglycerols but not PLs, and, thus, a source of n-3 PUFA-containing PLs is required.
Although several lines of evidence have indicated the biological significance of DHA, the metabolism and functions of DHAcontaining PLs in vivo remains unknown. One reason for lack of understanding of these mechanisms has been attributed to the absence of suitable model organisms for the synthesis of DHA and DHA-containing PLs. Although yeasts and Escherichia coli have greatly contributed to the accumulation of knowledge on lipid metabolism, neither can produce DHA. Thus, alternative model organisms that can produce DHA are strongly needed to further investigate the metabolism and functions of DHA and DHAcontaining PLs.
Thraustochytrids, which are unicellular marine heterotrophic protists belonging to the Chromista kingdom, can produce large amounts of n-3 PUFAs, and, thus, are expected to be an alternative to fish oils [11,12]. We have elucidated the structure of DHA-containing PLs in thraustochytrids [13], developed methods for thraustochytrid gene manipulation [14,15], and revealed their n-3 PUFA synthesis pathways [16,17]; however, how thraustochytrids synthesize DHA-containing PLs has yet to be elucidated.
In the present study, we identified a new lysophospholipid (LPL) acyl transferase (LPLAT) in the thraustochytrid, Aurantiochytrium limacinum F26-b, and named PL AcylTransferase 1 (PLAT1). PLAT1 shows wide specificity for donor substrates as well as acceptor substrates in vitro, i.e, the enzyme can adopt lysophosphatidylcholine (LPC), lysophosphatidylethanolamine (LPE), lysophosphatidylserine (LPS), and lysophosaphtidylinositol (LPI) as acceptor substrates and 15:0/16:0-CoA and DHA-CoA as donor substrates. However, disruption of the PLAT1 gene (plat1) in A. ]. This indicates that PLAT1 is the enzyme responsible for the generation of 16:0-DHA-PC and 16:0-DHA-PE, which are major PL species in the thraustochytrid. This is the first report on the identification of LPLAT in thraustochytrids, and the results obtained provide an insight into DHA metabolism in thraustochytrids, which are promising industrial microorganisms for the production of DHA and DHA-containing PLs. This study also suggests that A. limacinum F26-b is a favorable model organism for the metabolism and functions of DHA and DHA-containing PLs.

Materials
All acyl-CoAs were purchased from Avanti Polar Lipids (Alabaster, AL). [1-14 C]palmitoyl-LPC (55 mCi/mmol) was obtained from Perkin-Elmer Life Sciences (Walthama, MA). Synthetic complete medium and the yeast nitrogen base were obtained from MP Biomedica (Morgan Irvine, CA). The yeast overexpression vector pYES2/CT and S. cerevisiae InvSc2 were purchased from Life Technologies Japan Ltd. (Tokyo, Japan). All other chemicals were obtained from either Sigma (St. Louis, MO) or Wako (Osaka, Japan). The sequences of primers used in this study are listed in Table S1.

Identification of strain F26-b
The sequences were aligned with multiple-alignment method by using CLUSTAL_X [18] and refined manually. The alignment consists of 1,902 base pairs including gaps finally. The phyloge-netic analyses were performed using MEGA 5.2 [19]. A neighborjoining (NJ) phylogenetic tree was constructed using the distance matrix on the basis of the Tamura-Nei model [20]. Pairwise deletion was used for the gap treatment, and the statistical robustness of each branch was tested by a bootstrap analysis with 1,000 replications.

Culture of strain F26-b
A. limacinum F26-b was grown in GY medium (3% glucose and 1% yeast extract in 50% artificial sea water) with or without 0.1% vitamin mixture (vitamin B 1 200 mg, vitamin B 2 1 mg, vitamin B 12 1 mg/100 ml distilled water) at 25uC for the period indicated. Cells were harvested by centrifugation at 3,000 rpm for 5 min. Potato dextrose agar (PDA) plates containing hygromycin (50% potato dextrose and 2% agar in 50% artificial sea water containing 2 mg/ml hygromycin) were used to select plat1-knockout mutants.
Cloning of plat1 from A. limacinum F26-b LPLATs were searched for the genome database of A. limacinum ATCC MYA-1381 (http://genome.jgi.doe.gov/ pages/blast.jsf?db = Aurli1) using known LPLAT sequences as a query. One of the hits, scaffold_4:802352-804647, showed an open reading frame (ORF) homologous to that of human LPCAT1 and contained four motifs conserved for LPLATs. We then picked up the sequence as a putative LPLAT gene of thraustochytrids.
The putative ORF was obtained from the genomic DNA from A. limacinum F26-b by PCR using primers 1 and 2 (Table S1). The amplified 1,938-base pair PCR product was cloned into the TA cloning vector pGEM-T Easy vector system (Promega). The insert was then sequenced using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) and 3130 Genetic Analyzer (Applied Biosystems). The gene and predicted protein were named plat1 and PLAT1, respectively. The transmembrane motif was determined using the software TMHMM. A phylogenetic tree was drawn based on pairwise comparisons of the amino acid sequences of AGPAT (LPAAT), diacylglycerol acyltransferase 2 (DGAT2), and MBOAT family members using CLUSTALW, the DNA Data Bank of Japan (DDBJ) (http://clustalw.ddbj.nig.ac. jp/top-j.html).

Expression of PLAT1 in S. cerevisiae
To construct the plat1 expression vector, primers were designed to amplify the plat1 ORF. The FLAG epitope (DYKDDDDK) was attached to the N terminus of PLAT1 by PCR using primers 3 and 4 (Table S1). The fragments were then released by EcoRI and NotI double digestion and inserted into the yeast expression vector pYES2/CT, which is designed to work under the control of the inducible promoter GAL1.
The yeast strain InvSc2 was transformed with the expression construct using the lithium acetate method, and transformants were selected on minimal medium lacking uracil. Transformants were first grown on minimal medium containing 2% D-glucose, and then cultured on minimal medium containing 2% D-galactose at 30uC.

Preparation of yeast cell lysates
After the expression of plat1 was induced with D-galactose, cells were harvested, re-suspended in 1 ml of ice-cold 20 mM Tris-HCl, pH 7.5, containing 240 mM sucrose and 0.2 M PMSF, and crushed with 0.5 mm glass beads by a BEAD BEATER (Biospec Inc.) 5 times for 30 s. After centrifugation at 3,000 rpm for 10 min, the supernatant was subjected to further centrifugation at 100,0006g for 1 h by an ultracentrifuge. The supernatant and pellet were collected separately as cytosolic and microsomal fractions, respectively. The protein concentration of each fraction was measured with a BCA assay kit using bovine serum albumin as a standard.

Western blot analysis
Ten micrograms of total cellular protein was separated by 10% SDS-PAGE and transferred to a Hybond P nitrocellulose membrane (GE Healthcare). The membrane was blocked with 5% skim milk in TBS, incubated with the M2 anti-FLAG mouse monoclonal antibody in TBS, washed with TBS, and incubated with horseradish peroxidase (HRP)-linked anti-mouse IgG (GE Healthcare) in TBS. After being washed, the membranes were immersed in the peroxidase stain kit solution (Nacalai Tesque Inc.).

Purification of PLAT1 by immunoprecipitation
FLAG-tagged PLAT1 (FLAG-PLAT1) was purified using the ANTI-FLAG M2 Affinity Gel and 3X FLAG peptide (Sigma Aldrich, St. Louis, MO) according to the manufacturer's instructions. Briefly, 200 mg of the yeast cell lysate, which did or did not contain FLAG-PLAT1 (mock transfectant), was diluted in 20 mM Tris-HCl, pH 7.5, containing 0.2 M PMSF (wash buffer). This sample was subjected to FLAG resin, kept at 4uC overnight, and the supernatant was removed. The resin was washed three times with 0.5 ml wash buffer, and suspended in 100 ml of wash buffer.

Assay for LPLAT activity
LPLAT activity was measured based on the transfer of fatty acid from fatty acyl-CoA to lysoPLs (LPLs). To measure LPCAT activity, various fatty acyl-CoAs were incubated with [1-14 C]palmitoyl-LPC and the [ 14 C]PC that formed was quantified. The reaction mixture contained 1 mM EDTA, 0.01% sodium cholate, 5 mM [1-14 C]palmitoyl LPC (50,000 dpm/nmol), 25 mM of the respective acyl-CoA, and 10 mg of cell protein (cell lysate) or 10 ml of purified FLAG-PLAT1 in 100 ml of 100 mM Tris-HCl, pH 7.5. When the Ca 2+ dependency of the enzyme was measured, EDTA was replaced with 1 mM EGTA or 1 mM CaCl 2 . The reaction mixture was incubated at 30uC for 20 min and the reaction was then stopped by adding CHCl 3 /CH 3 OH (2:1, v/v). Total lipids were extracted and applied to a thin layer chromatography (TLC) plate, which was developed with CHCl 3 /CH 3 OH/H 2 O (65/25/ 4, v/v/v). The radioactivity of the corresponding bands was quantified using a FLA 5100 Bio-imaging analyzer (GE Healthcare). Other LPLAT activities were measured using the method described for the LPCAT assay except that [1-14 C]palmitoyl-CoA and a non-radioactive acceptor (LPA, LPS, LPE, or LPI) were used instead of non-labeled fatty acyl CoA and [1-14 C]palmitoyl LPC, respectively. Each assay contained 25 mM of LPL, 1 mM EDTA, 0.01% sodium cholate, 5 mM [1-14 C]palmitoyl-CoA (50,000 dpm/nmol), and 10 mlof purified FLAG-PLAT1 in 100 ml of 100 mM Tris-HCl, pH 7.5.

Construction of the plat1 knockout vector
The promoter region of the Thraustochytrium aureum ubiquitin gene was used to drive the expression of a selection marker, the hygromycin-resistant gene, as described in [16]. The terminator region of the SV40 virus coat protein gene was used to terminate gene transcription. The linear gene construct, which contained the promoter, the hygromycin resistant gene, and terminator, is referred to hereafter as the hygromycin-resistant cassette ( Figure  S1A, B).
Primers 5 and 6 (Table S1) were designed to insert the hygromycin-resistant cassette into plat1 ORF by homologous recombination, by which BglII and SalI sites were added to the middle of the coding region. Amplified PCR products were digested and linearized with BglII and SalI. The linear fragment was treated with E. coli alkaline phosphatase (BAP) (TOYOBO Inc.) to avoid self-ligation before ligation to the hygromycinresistant cassette, which was liberated from pT-HygR by digestion with BglII and SalI. The newly constructed plasmid (pKO-PLAThygR) was used as a template for PCR to prepare the linear fragment. The linear plat1-KO targeting vector is illustrated in Figure S1A.

Transformation of A. limacinum F26-b with the plat1-KO targeting vector
A. limacinum F26-b was grown in 3 ml of GY medium at 25uC for 3 days. Cells were collected by centrifugation at 3,000 rpm for 5 min, and washed with distilled water. Cell pellets were resuspended in Nucleofector solution (Lonza) to a final concentration of 5610 6 cells/100 ml. Five micrograms of the KO targeting vector was added to 100 ml of the cell suspension. The mixture was transferred into a 1-mm-gap cuvette. The cuvette was set on a GENE PULSER II (BioRad), and pulsed twice (0.75 kV, 50 mF, 50 V. The cells were added to 1 ml of fresh GY medium and incubated at 25uC overnight. Finally, all cells were transferred to hygromycin-containing PDA agar plates.

Southern blot analysis
Wet cell pellets from 50 ml of the A. limacinum F26-b culture were washed with sterile water and suspended in 10 ml of lysis buffer (50 mM NaCl, 10 mM EDTA, and 0.5% SDS in 20 mM Tris-HCl pH 8.0) containing 0.2 mg/ml proteinase K. The suspension was incubated at 55uC overnight, 10 ml of RNase A (100 mg/ml) was then added, and the incubation continued at 37uC for 2 h. Ten mg of genomic DNA was digested overnight at 37uC with HindIII. The digested DNA was fractionated on a 1% agarose gel by electrophoresis and transferred to a Hybond-N+ nylon membrane (GE Healthcare).
The probe was synthesized using a DIG DNA Labeling Kit (Roche Applied Science) according to the manufacturer's instructions and primers 7 and 8 (Table S1). The membrane was incubated at 47uC with the probe and signals were detected with a DIG Nucleic Acid Detection Kit (Roche Applied Science).

Nano ESI-MS analysis
Chip-based nanoESI-MS analysis was performed using 4000Q TRAP with a chip-based ionization source, TriVersa NanoMate (Advion BioSystems, Ithaca, NY, USA). The ion spray voltage was set at 1.25 kV, gas pressure at 0.3 pound per square inch (psi), and flow rates at 200 nL/min. The scan range was set at m/z 400-1100, de-clustering potential at 100 V, collision energies at 50-70 V, and resolutions at Q1 and Q3, ''unit''. The mobile phase composition was chloroform/methanol containing 0.1% ammonium formate (1/2, v/v). Total lipids were directly subjected to flow injection and selectivity was analyzed by precursor ion scanning of PC and PE from individual parental molecular species [21]. d18:1/12:0-sphingomyelin was added to the sample as an internal standard.

Identification of strain F26-b
The strain F26-b was isolated from fallen leaves of Rhizophora mucronata collected at Ishigaki Is., Okinawa, Japan. The neighbor-joining tree of 18S rRNA gene clearly shows that the strain F26-b forms a monophyletic group with the originally descripted ex-type strain (ATCC MYA 1381) of A. limacinum [22,23] and four related strains that is strongly supported by the highest bootstrap value (100%) (Figure 1). The strain F26-b was identified as A. limacinum based on this phylogenetic position and the microscopic morphological features (not shown).

Cloning of plat1 from A. limacinum F26-b
To identify the genes encoding LPLATs responsible for generating DHA-containing PLs in thraustochytrids, we searched the A. limacinum ATCC MYA-1381 genome database for sequence homology with previously reported LPLATs (Figure 2). We found seven candidate sequences in the database and designated them as PLAT1,7, all of which belong to the AGPAT (LPAT) family. In this study, we cloned the PLAT1 gene (plat1) from the genomic DNA of A. limacinum F26-b, which is genealogically related to A. limacinum ATCC MYA-1381 ( Figure 1).
The putative ORF of plat1 encoded a protein of 646 amino acid residues with a calculated molecular mass of 73.9 kDa. PLAT1 contained two transmembrane domains and four AGPAT motifs found in LPCATs [22,25,27]. Three EF hand motifs and a KK motif, which could serve as a calcium-binding motif and ERretaining motif, respectively, were located at the C terminus ( Figure 3). According to a phylogenetic tree based on the NCBI Blast search, PLAT1 was classified into the AGPAT (LPAAT) family, which includes mouse LPCAT1, LPCAT2 and GPATs with PLAT1 being the most homologous to mouse LPCAT1 ( Figure 2).

Characterization of PLAT1
In order to characterize it, PLAT1 was expressed in S. cerevisiae as a fusion protein with a FLAG epitope at the N terminus. After the induction of PLAT1 expression by D-galactose, the yeast cell lysate was subjected to Western blotting using the M2 anti-FLAG antibody ( Figure 4A). FLAG-PLAT1 showed a 74-kDa band on the blot consistent with the molecular mass predicted from the ORF of PLAT1. When the cell lysate was subjected to centrifugation to separate the cytosolic and microsomal fractions, FLAG-PLAT1 was mainly detected in the latter ( Figure 4B). FLAG-PLAT1 possessed LPCAT activity when [ 14 C]LPC and 16:0-CoA were used as an acceptor and donor, respectively ( Figure 4C). Although PLAT1 had three EF hand calcium-binding motifs (Figure 3), Ca 2+ was not necessary for LPCAT activity ( Figure 4D).
Generation and analysis of the plat1-disrupted mutants of A. limacinum F26-b PLAT1 exhibited wide specificity for donors ( Figure 5A) as well as acceptors ( Figure 5B, C) in vitro when the FLAG-tag purified enzyme, expressed in S. cerevisiae, was used for the assay. To examine the in vivo specificity of PLAT1 in A. limacinum F26-b, plat1 was disrupted in the thraustochytrid by homologous recombination using hygromycin as a selection marker ( Figure  S1A). Southern blotting showed that plat1 was disrupted by the hygromycin-resistant gene ( Figure S1B, C). The plat1-disrupted mutants of A. limacinum F26-b showed no obvious changes in cell growth and glucose consumption under the conditions used ( Figure S2A, B).
A decrease in LPCAT activity was observed in plat1-disrupted mutants when activity was measured using LPC and 15:0-CoA or DHA-CoA as substrates; LPCAT activities were 90% and 80% lower than the wild-type when 15:0-CoA ( Figure 6A) and 16:0-CoA (data not shown) were used, respectively, and 50% lower than the wild-type when DHA-CoA was used ( Figure 6B). These results suggest that the acylation of LPC with saturated fatty acids was mainly catalyzed by PLAT1, while that with DHA was catalyzed by PLAT1 and other LPCATs in A. limacinum F26-b. The various LPLAT activities of plat1-disrupted mutants were examined using different LPLs and [ 14 C]16:0-CoA. As shown in Figure 6C, the activities of not only LPCAT, but also LPEAT were significantly decreased after the disruption of the plat1 gene; however, those of LPSAT and LPAAT were not, which indicated that PLAT1 is likely to function in vivo as LPCAT and LPEAT, but not other LPLATs. LPIAT activity was found to increase in plat1-knockout mutants under the conditions used ( Figure 6C

Discussion
The major fatty acids of A. limacinum F26-b are 15:0/16:0 and DHA; the former occupies approximately 40% of total fatty acids, and the latter, approximately 30% when the strain was cultured in GY medium [13]. However, most of 15:0 was replaced by palmitic acid (16:0) with the addition of vitamin B12. The odd-chain fatty acid (15:0) could also be generated from the fatty acid synthase  (FAS) pathway using propionyl-CoA as a primer instead of malonyl-CoA in the absence of vitamin B12 [24,25].
In general, PLs are synthesized from Kennedy pathway ( Figure 7A) and the fatty acids at sn-2 are released from PLs through the actions of PLA 2 , and PLs are re-generated by transferring fatty acids from acyl-CoA to LPLs by LPLATs. This remodeling of the fatty acyl chains of PLs was proposed by Lands in the 1950s and has been named Lands cycle ( Figure 7A) [26]. The LPLATs involved in Lands cycle were previously unknown, but may have recently been identified independently by several laboratories [27][28][29][30][31][32].
LPLATs are divided into two major groups based on their primary structures and characteristic motifs; AGPAT (LPAAT) and MBOAT (Figure 3) [33][34][35]. The enzymes belonging to AGPAT family have four well-conserved motifs [33]. The MBOAT family enzymes also possess the conserved motifs that are postulated to be important for the activity, although the motifs of MOBAT are totally different from those of AGPAT [36,37]. (A) FLAG-PLAT1 was purified using the ANTI-FLAG M2 Affinity Gel and 3X FLAG peptide to remove the endogenous LPLs as described in Materials and Methods. Each assay contained 25 mM acyl-CoA, 1 mM [1-14 C]palmitoyl-LPC, and purified enzyme (10 ml of gel suspension) in 100 ml of 100 mM Tris-HCl, pH 7.5, containing 1 mM EDTA and 0.01% sodium cholate. The reaction was conducted at 30uC for 20 min and terminated by the addition of 500 ml of CHCl 3 /CH 3 OH (2:1, v/v). The reaction mixture was loaded on the TLC plate and developed with CHCl 3 /CH 3 OH/H 2 O (65/25/4, v/v/v). The radioactivity of the band corresponding to PC on the TLC plate was quantified by a FLA 5100 Bio-imaging analyzer. Data represent the mean 6 SD (n = 3). (B) Each assay contained 1 mM of LPLs, 1 mM [1-14 C]palmitoyl-CoA, and the purified enzyme (10 ml of gel suspension) in 100 ml of 100 mM Tris-HCl, pH 7.5, containing 1 mM EDTA and 0.01% sodium cholate. Mock represents the experiment using the enzyme prepared from the lysate of yeast cells harboring the empty vector. The assay was conducted at 30uC for 20 min and terminated by the addition of 500 ml of CHCl 3 / CH 3 OH (2:1, v/v). The reaction mixture was loaded on the TLC plate, developed with CHCl 3 /CH 3 OH/H 2 O (65/25/4, v/v/v), and analyzed with a FLA 5100 Bio-imaging analyzer. The typical TLC is presented here. (C) The enzyme reaction was performed by the same procedure as shown in (B). The reaction products were applied on TLC plates and the radioactivity of the band corresponding to each PL was quantified by a FLA 5100 Bio-imaging analyzer (n = 3). Data represent the mean 6 SD. doi:10.1371/journal.pone.0102377.g005 Among the LPLATs reported to date, several enzymes were found to adopt PUFA-CoA as the donor substrate. LPIAT, which catalyzes the incorporation of arachidonic acid (AA, 20:4n-6) and EPA into LPI from the corresponding acyl donor, was isolated from Caenorhabditis elegans [38]. This was the first LPLAT that was shown to be capable of catalyzing the incorporation of PUFAs into PLs; however, C. elegans cannot synthesize DHA [39]. PLAT1, which is composed of 646 amino acid residues, possesses four conserved AGPAT motifs and three EF hand calcium-binding motifs. These structural features closely resemble those of mouse/human LPCAT1 and LPCAT2. When plat1 was expressed in yeast cells, the LPCAT activity of recombinant PLAT1 was completely independent of Ca 2+ , similar to that of recombinant human LPCAT1 [30]. In contrast, LPCAT2 only exhibited lysoPAF acyltransferase activity in the presence of Ca 2+ [32].
The specificity of PLAT1 toward LPL acceptors in vitro was different to those of LPCAT1 and LPCAT2. Apart from LPCAT activity, PLAT1 also exhibited LPEAT, LPAAT, LPIAT, and LPSAT activities in vitro. In contrast, LPCAT1 exhibited LPAAT, LPGAT, and lysoPAF-AT activities, while LPCAT2 possessed a strong lysoPAF-AT activity in addition to primary LPCAT activity [32]. The comparison of the specificities of PLAT1 and other known LPCATs toward acyl-CoAs and LPLs is summarized in Figure 7 B, C.
To clarify the in vivo functions of PLAT1 in A. limacinum F26b, plat1-knockout mutants were generated in this study. Recombinant PLAT1 showed the LPLAT activities toward not only LPC but also LPE, LPI, LPA and LPS under in vitro enzyme assay ( Figure 5C). However, only LPCAT and LPEAT activities were significantly lower in the plat1-knockout mutants than in the wildtype ( Figure 6C). This result may indicate that other LPLATs compensate the activities of LPIAT, LPAAT and LPSAT in vivo. Interestingly, nano ESI-MS analysis indicated that the amounts of 16:0-DHA-PC and 16:0-DHA-PE, but not DHA-DHA-PC and DHA-DHA-PE, were lower in plat1-knockout mutants than in the wild-type ( Figure 6D). It is worth noting that neither 16:0-16:0-PC nor 16:0-16:0-PE was detected by nano ESI-MS in A. limacinum F26-b. These results indicate that PLAT1 may be central to the generation of 16:0-DHA-PC and 16:0-DHA-PE, which are major PL species in A. limacinum F26-b [13]. In this context, it is suggested that PLAT1 preferentially transfers 16:0 from 16:0-CoA to DHA-LPC (LPE); however, the enzyme could also adopt DHA-CoA as a donor substrate if the acceptor substrate is a 16:0-LPC (LPE) in vivo.
This study promotes understanding of the global lipid metabolism of thraustochytrids, and may accelerate the molecular breading of promising microorganisms for the production of DHA and DHA-containing phospholipids.  LPLs (C) was compared with those of the LPCATs reported so far. The darker color represents higher activity toward the substrates indicated. White (blank) and ND show the absence of activity and not determined, respectively. The specificities of LPCATs are mainly referred from [27]. doi:10.1371/journal.pone.0102377.g007 vitamin B 2 : 1 mg, vitamin B 12 : 1 mg/100 ml distilled water) at 25uC for the periods indicated. A small sample of the culture was withdrawn and the optical density at 600 nm was measured after suitable dilution. The glucose content of the culture supernatant was measured with a glucose CII-test (Wako, Japan). Data represent the mean 6 SD (n = 3). (TIF) Table S2 Ratios of PL species in wild-type and plat1-disrupted mutants. Values are the ratios of 16:0-DHA-PLs to DHA-DHA-PLs, calculated from the data of nanoESI/MS analyses. WT, wildtype of A.limacinum F26-b; KO1,KO3, three different plat1-disrupted mutants obtained from transfection of the wild-type with a KO construct containing the HygR gene as a marker, as shown in Figure S1. (DOCX)