Analysis of Δ12-fatty acid desaturase function revealed that two distinct pathways are active for the synthesis of PUFAs in T. aureum ATCC 34304.

Thraustochytrids are known to synthesize PUFAs such as docosahexaenoic acid (DHA). Accumulating evidence suggests the presence of two synthetic pathways of PUFAs in thraustochytrids: the polyketide synthase-like (PUFA synthase) and desaturase/elongase (standard) pathways. It remains unclear whether the latter pathway functions in thraustochytrids. In this study, we report that the standard pathway produces PUFA in Thraustochytrium aureum ATCC 34304. We isolated a gene encoding a putative Δ12-fatty acid desaturase (TauΔ12des) from T. aureum. Yeasts transformed with the tauΔ12des converted endogenous oleic acid (OA) into linoleic acid (LA). The disruption of the tauΔ12des in T. aureum by homologous recombination resulted in the accumulation of OA and a decrease in the levels of LA and its downstream PUFAs. However, the DHA content was increased slightly in tauΔ12des-disruption mutants, suggesting that DHA is primarily produced in T. aureum via the PUFA synthase pathway. The transformation of the tauΔ12des-disruption mutants with a tauΔ12des expression cassette restored the wild-type fatty acid profiles. These data clearly indicate that TauΔ12des functions as Δ12-fatty acid desaturase in the standard pathway of T. aureum and demonstrate that this thraustochytrid produces PUFAs via both the PUFA synthase and the standard pathways.

droplets; thus, they are under consideration as an alternative industrial source of n-3 PUFAs.
Two distinct pathways for the production of PUFAs have been proposed in thraustochytrids: the polyketide synthase-like (PUFA synthase) pathway, which occurs in several marine bacteria ( 7 ), and the desaturase/elongase (standard) pathway, which occurs widely in eukaryotes ( 6,8 ). Gene clusters in the PUFA synthase pathway have been isolated from Schizochytrium (now reclassifi ed as Aurantiochytrium ), in which the disruption of one gene in the synthase pathway resulted in the loss of DHA and n-6 docosapentaenoic acid (C22:5 n-6, C22:5 ⌬ 4, 7, 10, 13, 16 ), indicating that these PUFAs are produced solely by the PUFA synthase pathway ( 9,10 ). Several genes encoding fatty acid desaturases and elongases, which may be involved in the standard pathway, have been isolated from thraustochytrids (11)(12)(13). However, the genes encoding the ⌬ 12-and ⌬ 15fatty acid desaturases, which are key enzymes in the standard pathway, have not been identifi ed in thraustochytrids. Thus, it is unclear whether the standard pathway is responsible for the production of PUFAs in thraustochytrids.
In this study, we demonstrate, for the fi rst time, that two distinct pathways for the synthesis of PUFAs are active in T. aureum . This report is the fi rst to describe the disruption of a fatty acid desaturase gene in thraustochytrids. Thus, this study opens the door for elucidating these entire biosynthetic pathways and the biological functions of PUFAs in thraustochytrids and facilitates the genetic modifi cation of thraustochytrids for the production of benefi cial PUFAs.

Materials
Synthetic oligonucleotides were obtained from Hokkaido System Science (Hokkaido, Japan) and GeneNet (Fukuoka, Japan). D-(+)-Glucose and dry yeast extract were purchased from Nacalai Tesque (Kyoto, Japan). Restriction enzymes and a Ligation-Convenience Kit were purchased from Nippon Gene (Tokyo, Japan). The antibiotics neomycin (G418), hygromycin B, and blasticidin were purchased from Nacalai Tesque. SEALIFE was obtained from Marinetech (Tokyo, Japan). All other reagents were of the highest purity available.

Strains and culture
T. aureum ATCC 34304 was obtained from the American Type Culture Collection (USA). This strain was maintained on potato dextrose agar (PDA) plates (0.8% potato dextrose and 1.2% agar in 50% artifi cial sea water, ASW). The thraustochytrid was cultivated on GY medium consisting of 3% glucose, 1% yeast extract, and 1.75% SEALIFE.

Molecular cloning of Tau ⌬ 12des from T. aureum ATCC 34304
T. aureum was grown at 25°C in GY medium. Cells in the late logarithmic growth phase were harvested by centrifugation (3,500 × g , 4°C, 10 min), and the genomic DNA was extracted. The primers were designed based on our local genome database. The open reading frame (ORF) of the predicted ⌬ 12-fatty acid desaturase was amplifi ed with the forward primer Tw3-F1 and the reverse primer Tw3-R1. The primer sequences are listed in Supplementary Table I. PCR was then performed using these prim-ers with T. aureum genomic DNA as a template in a master mix that included LA Taq DNA polymerase (Takara Bio Inc., Shiga, Japan). The amplifi ed PCR products were purifi ed and cloned into the pGEM -T Easy Vector (Promega, Tokyo, Japan) and sequenced. The full-length genomic DNA clone encoding a ⌬ 12fatty acid desaturase was named Tau ⌬ 12des.

Expression of the tau ⌬ 12des in yeast
The ORF of the tau ⌬ 12des was amplifi ed by PCR using a 5 ′ primer containing a Hin dIII site (Tw3-Hind 3-F) and a 3 ′ primer containing an Xba I site (Tw3-Xba 1-R) and genomic DNA as a template (98°C/20 s, 60°C/30 s, 72°C/1.5 min, 30 cycles). The PCRamplifi ed Tau ⌬ 12des ORF was digested with Hin dIII and Xba I and then purifi ed and cloned into the same sites in pYES2/CT (Invitrogen, Carlsbad, CA). The resulting Tau ⌬ 12des expression vector, designated pYTau ⌬ 12Des, was introduced into S. cerevisiae INVSc1 (Invitrogen) using the lithium acetate method ( 14 ). The transformants were selected by plating on synthetic agar plates lacking uracil (SC-ura). S. cerevisiae transformants harboring the tau ⌬ 12des were cultured in SC-ura medium containing 2% glucose at 25°C for 3 days and then cultured for an additional 1 day in SC-ura medium containing 2% galactose. The cells were collected by centrifugation at 3,500 × g for 10 min.

Western blotting of FLAG-tagged Tau ⌬ 12des
The FLAG tag sequence was inserted immediately after the initiation codon of the Tau ⌬ 12des gene by PCR. The PCR was conducted using a forward primer containing the FLAG tag sequences (TD12d-FLAG-F, 5 ′ -GG A AGC TT A TG G ATT ACA AGG ATG ACG ATG ACA AGT GCA AG G TCG ATG-3 ′ ) and the reverse primer Tw3-Xba1-R; the underlining and italics here indicate the Hin dIII site and the FLAG tag sequence, respectively. The PCR fragment was cloned directly into the yeast expression vector pYES2/CT and subsequently introduced into S. cerevisiae by the method described above. After the incubation of the transformants in SC-ura medium, the proteins were extracted, and a Western blotting assay was performed as described previously ( 15 ). Briefl y, 10 g of protein was loaded onto a 10% SDS-PAGE gel and transferred to a PVDF membrane (0.45 m) using a Bio-Rad Trans-Blot SD Cell. The membrane was incubated with 5% (w/v) skim milk in TBS buffer containing 0.1% Tween 20 (Tween-TBS) for 1 h at room temperature with constant agitation. After three washes with Tween-TBS, the membrane was incubated at room temperature for 3 h with an anti-DYKDDDDK tag monoclonal antibody (1:5,000; Wako, Osaka, Japan). The membrane was then washed with Tween-TBS three more times and incubated for 3 h at room temperature with an HRP-conjugated anti-mouse IgG [H+L] goat antibody (Nacalai Tesque; 1:10,000). The membrane was again washed thrice with Tween-TBS. Protein expression was visualized using a peroxidase staining kit (Nacalai Tesque; 1:20).

Targeted disruption of the tau ⌬ 12des in T. aureum
The tau ⌬ 12des -disruption mutants were generated by homologous recombination. Because T. aureum is apparently diploid, two different markers were used for the disruption of the gene in the two different alleles. The disruption constructs consisted of either the Hyg r or Bla r expression cassette sandwiched between the 1,001-bp 5 ′ -and 3 ′ -fl anking sequences of the tau ⌬ 12des . First, the 5 ′ -and 3 ′ -fl anking sequences were amplifi ed using the TD12d-up-F and TD12d-up-R primers and the TD12d-down-F and TD12d-down-R primers, respectively. Next, these amplifi ed fragments were connected by fusion PCR and cloned into the pGEM -T Easy Vector. The Hyg r and Fig. 1. Alignment and phylogenetic tree of Tau ⌬ 12des. A: alignment of the deduced amino acid sequence of Tau ⌬ 12des with the sequences of diatom and picophytoplankton ⌬ 12-fatty acid desaturases. Tau ⌬ 12des and ⌬ 12-fatty acid desaturases from different origins were aligned using ClustalW 1.81 and the alignment was shaded in ESPript 2.2 (http://espript.ibcp.fr/ESPript/cgi-bin /ESPript.cgi). Identical and similar amino acid residues are shown as white letters on a black background and in bold face with a black box, respectively. The histidine boxes that are commonly conserved in membrane fatty acid desaturases are underlined. Tps ⌬ 12des, Thalassiosira pseudonana ⌬ 12-fatty acid desaturase (XP_002292071); Ptr ⌬ 12des, Phaeodactylum tricornutum ⌬ 12-fatty acid desaturase (3503348AJJ); Tau ⌬ 12des, T. aureum ATCC34304 ⌬ 12-fatty acid desaturase (this study); and Msp ⌬ 12des, Micromonas sp.  harboring tau ⌬ 12des . Gas chromatograms showing the FAMEs extracted from S. cerevisiae transformed with (A) an empty vector, pYES2/CT (mock transformants) and (B) tau ⌬ 12des -containing vector, pYTau ⌬ 12Des (Tau ⌬ 12des transformants). The detector voltages were shifted from 1.20 to 1.50 kV at the retention time of 22.50 min. The arrows indicate the new peak observed in the Tau ⌬ 12des transformants (B). The cells were cultured in uracil-defi cient SC medium containing 2% glucose at 25°C for 3 days and then cultured for an additional day in uracildefi cient SC medium containing 2% galactose with or without exogenous fatty acids. Fatty acids were added to the culture with 0.1% Tergitol. The fatty acids were extracted from freeze-dried cells and subjected to GC analysis as described in Materials and Methods. Bla r expression cassettes were cloned into the Bgl II site of the vector. The ubiquitin promoter and SV40 terminator were cloned from T. aureum ATCC 34304 and the pcDNA 3.1 Myc-His vector (Invitrogen), respectively. The hyg r and bla r were obtained from pcDNA 3.1/Hygro (Invitrogen) and pTracer-CV/Bsd/lacZ (Invitrogen), respectively. The primers used for the PCR amplifi cation of these sequences are listed in Supplementary Table I. Homologous recombination was performed using the modifi ed split marker system ( 16 ). The disruption construct was separated into 5 ′ -and 3 ′ -fragments by PCR and introduced into T. aureum cells by microprojectile bombardment. Cells in the logarithmic growth phase were harvested by centrifugation (3,500 × g , 4°C, 10 min) and spread on a PDA plate (15 × 60 mm) without antibiotics. The bombardment was performed in a PDS-1000/He Particle Delivery System (Bio-Rad) with DNA-coated gold microcarriers according to the manufacturer's instructions using the following bombardment conditions: pressure, 1,100 psi; target distance, 6 cm; vacuum, 26 inches Hg. Gold particles (0.6 m in diameter) were coated with the disruption construct according to the manufacturer's instructions. The fi rst allele of Tau ⌬ 12des was replaced with the disruption construct containing the Hyg r expression cassette (fi rst allele knock-out construct), and the second allele was replaced with the construct containing the Bla r expression cassette (second allele knock-out construct). After the bombardment, the plate was incubated at 25°C for 3 h, after which the cells were collected and suspended in GY medium followed by respreading on a PDA plate containing antibiotics. The transformants were selected by their ability to grow on PDA plates containing hygromycin B or hygromycin B plus blasticidin. The concentrations of hygromycin B and blasticidin in the PDA plates were 2 mg/ml and 0.2 mg/ml, respectively.

Complementation of the tau ⌬ 12des -disruption mutants with the tau ⌬ 12des
To express the tau ⌬ 12des in the tau ⌬ 12des -disruption mutants, the Neo r /Tau ⌬ 12des construct (see Fig. 5A ) was prepared. For the control experiment, the tau ⌬ 12des with the ubiquitin promoter/terminator was omitted from the expression construct (Neo r construct; see Fig. 5B ). The ubiquitin terminator was obtained from T. aureum ATCC 34304. The codons of Neo r were adjusted to match the codon usage of T. aureum ATCC 34304. The primers used for the PCR amplifi cation are listed in Supplementary Table I and in a previous report ( 15 ). The expression construct was introduced into T. aureum cells by the method described above. The cells were incubated on a PDA plate at 25°C for 3 h, after which the colonies were collected and spread on a PDA plate containing G418 at 2 mg/ml. After incubation at 25°C for 7 days, any colonies that appeared on the plates were regarded as putative transformants. The T. aureum transformants were cultured in GY medium containing G418 at 2 mg/ml at 25°C for 5 days. The cells were collected by centrifugation at 3,500 × g for 10 min.

Genomic PCR and Southern blot hybridization
Genomic PCR was performed using the Hyg-F and Hyg-R primers for the amplifi cation of the hyg r , the Bla-F and Bla-R primers for the bla r , the forward primer ub pro-Tw3-F with the reverse primer ub term-Tw3-R for the tau ⌬ 12des , and the forward primer 2F with the reverse primer pUC18-R for the Neo r / Tau ⌬ 12des construct. For Southern blot hybridization, 1.5 g of genomic DNA was digested with restriction enzymes at 37°C overnight. The digested DNA was separated on a 0.7% agarose gel and transferred onto a Hybond-N + membrane (GE Healthcare, Tokyo, Japan). The membrane was hybridized with a probe prepared using the DIG DNA Labeling Kit (Roche Diagnostics K.K., Mannheim, Germany). The probes were amplifi ed with the KO up-probe-F1 and KO up-probe-R1 primers (for the 5 ′ -fl anking region), the KO down-probe-F3 and KO down-probe-R3 primers (for the 3 ′ -fl anking region), and the TD12d-probe-F1 and TD12d-probe-R1 primers (for the Neo r / Tau ⌬ 12des construct). The genomic DNA hybridized with each probe was detected with the anti-Digoxigenin-AP Fab fragment and an NBT/BCIP stock solution (Roche Diagnostics K.K.).
primers for the amplifi cation of Hyg r cDNA, the Bla-F and Bla-R primers for the Bla r cDNA, the 3F and 4R primers for the Neo r cDNA, and the ub pro-Tw3-F and ub term-Tw3-R primers for the Tau ⌬ 12des cDNA.

Growth curve and dry cell weight
Precultured cells (2.5 ml) were inoculated into 250 ml GY medium in a 500 ml fl ask. After incubating the culture at 25°C with Detection of Hyg r , Bla r , Neo r , and Tau ⌬ 12des mRNA by RT-PCR Total RNA was prepared from transformants grown in GY medium containing appropriate amounts of antibiotics with Sepasol RNA I Super (Nacalai Tesque), an RNeasy Mini Kit (QIAGEN, Tokyo, Japan), and DNaseI (Takara Bio Inc.) and used to produce fi rst-strand cDNA with PrimeScript TM Reverse Transcriptase (Takara Bio Inc.). PCR was performed using the Hyg-F and Hyg-R shaking at 150 rpm, the absorbance measurements were performed at a wavelength of 600 nm with an Ultrospec 3000 spectrophotometer. Optimal density at 600 nm is below 0.1 at the start for culture. Spectrophotometric readings of the optimal density were taken every hour. The dry cell weight was determined by transferring 10 ml of the culture to a preweighed centrifuge tube and then centrifuging at 3,500 × g for 10 min. The cell pellet was washed twice with 50% ASW and once with distilled water. The washed cell pellets were freeze-dried and weighed.

Fatty acid analysis
Precultured cells were incubated in a 50-ml fl ask containing 25 ml of GY medium at 25°C for 5 days with shaking at 150 rpm.
The harvested cells were washed twice with 50% ASW and once with distilled water. The preparation and extraction of fatty acid methyl esters (FAMEs) were performed as described previously ( 15 ). The resulting FAMEs were analyzed by GC using the method described previously ( 15 ). The FAMEs were also subjected to GC-MS using a Shimadzu GC-MS QP-5000 (Shimadzu Co., Kyoto, Japan) equipped with a capillary column (DB-1, 0.25 mm i.d. × 30 m, fi lm thickness 0.25 m; Agilent). The column temperature was programmed to increase from 160°C to 260°C at 4°C/min. The injection-port temperature was 250°C. Using lignoceric acid (C24:0) as an internal standard, the FAME samples were analyzed and quantifi ed based on their peak areas on the chromatogram relative to the peak area of the internal standard. Picolinyl esters Fig. 4. GC of FAMEs from wild-type (A) and tau ⌬ 12des -disruption mutants (B). The cells were cultured in a GY medium containing ampicillin at a concentration of 0.1 mg/ml at 25°C for 5 days. The fatty acids were extracted from freeze-dried cells and subjected to GC as described in Materials and Methods. Endogenous substrates for Tau ⌬ 12des, C17:1 ⌬ 9 , C18:1 ⌬ 9 , and C19:1 ⌬ 9 are shown in the square, and C18:2 ⌬ 9, 12 shown in the circle.

Molecular cloning of a ⌬ 12-fatty acid desaturase from T. aureum ATCC 34304
Several fatty acid desaturase genes have been cloned from thraustochytrids (11)(12)(13); however, a ⌬ 12-fatty acid desaturase gene has not been cloned from these organisms. In this study, we isolated a putative ⌬ 12-fatty acid desaturase (Tau ⌬ 12des) gene from T. aureum ATCC 34304, as described in Materials and Methods. The gene, named tau ⌬ 12des , contains a 1,185-bp ORF encoding a putative protein of 395 amino acids. The deduced amino acid sequence of the tau ⌬ 12des exhibits a high degree of identity with ⌬ 12-fatty acid desaturases found in diatoms and picophytoplankton, such as those from Thalassiosira pseudonana (41%) (XP_002292071), Micromonas sp. (44%) (XP_002507091), and Phaeodactylum tricornutum (41%) (3503348AJJ) (the number in parentheses indicates the sequence identity relative to Tau ⌬ 12des) ( Fig. 1A ). Three histidine boxes, which are conserved in almost all membrane-bound fatty acid desaturases, are found in the deduced amino acid sequence of Tau ⌬ 12des ( Fig. 1A , underlined), whereas the cytochrome b 5 motif, a characteristic of front-end desaturases, is not present in the enzyme.

Phylogenetic analysis of Tau ⌬ 12des
The ⌬ 12-and ⌬ 12/ ⌬ 15-fatty acid desaturases have been classifi ed into the following groups based on sequence similarity: a fungal and protozoan group, a plant group, a cyanobacterial group, and a chloroplast-localized plant group. The evolutionary relationships among Tau ⌬ 12des and other ⌬ 12-and ⌬ 12/ ⌬ 15-fatty acid desaturases were examined in a phylogenetic analysis. Although Tau ⌬ 12des was not clustered with any group, it was most closely related to the ⌬ 12-fatty acid desaturase found in the diatom P. tricornutum ( Fig. 1B ). prepared from the FAMEs as described previously ( 15 ) were subjected to GC-MS using the equipment described above. The column temperature was programmed to increase from 240°C to 260°C at 2.5°C/min, hold at 260°C for 15 min, and then increase to 280°C at 2.5°C/min.

Lipid extraction and the separation of lipid classes
Precultured cells were incubated in a 500-ml fl ask containing 200 ml of GY medium at 25°C for 5 days with shaking at 150 rpm. The cells were harvested by centrifugation at 3,000 × g for 10 min and washed twice with 50% ASW and once with distilled water. The total lipids were extracted using the Folch method ( 17 ) after freeze-drying the cells.
The separation of the total lipids into neutral lipid, glycolipid, and phospholipid fractions using a Sep-Pak Plus Silica cartridge (2 ml) and TLC analysis was performed as described previously ( 18 ). The FAMEs in each fraction were prepared and analyzed by GC as described above.

Exploring the specifi city of Tau ⌬ 12des expressed in the budding yeast S. cerevisiae
To elucidate the specifi city of Tau ⌬ 12des activity, a Tau ⌬ 12des expression construct (pYTau ⌬ 12Des) and an empty-control construct (pYES2/CT) were separately introduced into the S. cerevisiae strain INVSc1, and the fatty acid compositions of these transformants were analyzed by GC using their corresponding FAMEs. The peak corresponding to the LA (18:2 ⌬ 9, 12 ) methyl ester standard was found in the GC spectra of the pYTau ⌬ 12Des transformants ( Fig. 2B ) but not in those of the mock transformants ( Fig. 2A ). GC-MS analysis of the newly generated peak in the pYTau ⌬ 12Des transformants revealed the presence of a molecular ion ( m/z 294) and fragment ions identical to those of the LA methyl ester standard (supplementary Fig. IA, B ). These results indicated that endogenous OA Fig. 5. Molecular characterization of Tau ⌬ 12des revertants. Neo r /Tau ⌬ 12des construct (A) and Neo r construct (B) were separately injected to the tau ⌬ 12des -disruption mutants, and the transformants were obtained were designated Tau ⌬ 12des revertants and Neo r transformants (control). The tau ⌬ 12des and neo r were driven with thraustochytrid-derived ubiquitin promoter/terminator and EF-1 ␣ promoter/terminator, respectively. The primers for fusion PCR of these constructs are shown below each construct. C: Genomic PCR showing the Neo r and Neo r /Tau ⌬ 12des constructs. D: Southern blot hybridization using a Tau ⌬ 12desspecifi c probe. E, F: RT-PCR amplifying Neo r mRNA (E) and Tau ⌬ 12des mRNA (F). M, Hin dIII digestion / X174 Hin cII digestion marker; N, negative control (wild-type T. aureum ); C1-C3, Neo r transformants (mock transformants); T1-T3, Neo r /Tau ⌬ 12des transformants; P1, positive control (Neo r construct); P2, positive control (Neo r /Tau ⌬ 12des construct). These procedures are described in detail in Materials and Methods.

Western blotting of FLAG-tagged Tau ⌬ 12des expressed in yeast
We examined the expression of Tau ⌬ 12des at the protein level when expressed in S. cerevisiae . Yeast cells expressing FLAG-tagged Tau ⌬ 12des were lysed and fractionated into microsomal and cytosolic fractions followed by analysis with Western blotting using an anti-DYKDDDDK tag antibody. A 45.3 kDa protein band was detected in the cell lysate and the microsomal fractions but not in the cytosolic fraction (supplementary Fig. II ). This molecular weight was consistent with that estimated from the deduced amino acid sequence of Tau ⌬ 12des with a FLAG tag. These results indicate that Tau ⌬ 12des is classifi ed as a microsomal fatty acid desaturase.

Generation of tau ⌬ 12des -disruption mutants
To address the question of whether Tau ⌬ 12des is involved in the standard pathway in T. aureum , the tau ⌬ 12des was disrupted in the thraustochytrid by homologous recombination using a disruption construct containing hyg r or bla r as a marker gene fl anked with the 5 ′ and 3 ′ sequences of the Tau ⌬ 12des genomic locus (supplementary Fig. III ). Because T. aureum ATCC 34304 appears to be diploid, two loci harboring the tau ⌬ 12des should be disrupted by different marker genes to create a full deletion mutant. Transformants grown on GY medium containing hygromycin B (fi rst-allele disrupted mutants) or hygromycin B plus blasticidin (fi rst-and second-allele disrupted mutants) were subjected to genomic PCR and RT-PCR to confi rm the disruption of the tau ⌬ 12des . The 1,026-bp and 399-bp PCR products (corresponding to the hyg r and bla r , respectively) were detected in the fi rst-and second-allele disrupted mutants ( tau ⌬ 12des -disruption mutants) but not in the wild-type strain ( Fig. 3A a, Ab ). In contrast, a 1,185-bp PCR product (corresponding to tau ⌬ 12des ) was amplifi ed in the wild-type strains and the fi rst-allele disrupted mutants but not in the fi rst-and second-allele disrupted mutants ( Fig. 3A c). Furthermore, RT-PCR revealed that transcripts of the hyg r (1,026-bp) and the bla r (399-bp), but not the tau ⌬ 12des , were present in fi rst-and second-allele disrupted mutants, whereas the transcript of tau ⌬ 12des (1,185-bp) was detected in both the wild-type strains and in the fi rst-allele disrupted mutants ( Fig. 3A d, Ae, Af). Transcripts of the hyg r , but not the bla r , were detected in the fi rst-allele disrupted mutants, and no transcripts of the hyg r or bla r were detected in the wild-type strain. Southern blot hybridization using the DIG-labeled 5 ′upstream and 3 ′ -downstream regions of the tau ⌬ 12des as the probes was conducted to further characterize the tau ⌬ 12des -disruption mutants ( Fig. 3B ). When hybridized with the 5 ′ -upstream-specifi c probe, a single 2,028-bp band was detected in the wild-type strain, whereas 5,880-and 5,253-bp bands, corresponding to the two disruption constructs containing each marker gene, were detected in the fi rst-and second-allele disrupted mutants, respectively ( Fig. 3C ). Hybridization with a 3 ′ -downstream-specifi c probe resulted in the generation of single a 2,334-bp band in the wild-type strain, whereas a single 1,496 bp band was detected in the fi rst-and second-allele disrupted mutants ( Fig. 3D ). These results clearly indicate that the tau ⌬ 12des was disrupted by homologous recombination with the two marker genes and that T. aureum is diploid under the growth conditions used.

Characterization of the tau ⌬ 12des -disruption mutants from the perspective of fatty acid biosynthesis
The compositions of the fatty acids in the wild-type strain and the tau ⌬ 12des -disruption mutants were analyzed by GC using their methyl ester derivatives. The picolinyl esters, prepared from the FAMEs, were also analyzed by GC-MS to identify each fatty acid (data not shown). In contrast to the wild-type, the mutants had no LA (C18:2 ⌬ 9, 12 ), the major product of Tau ⌬ 12des, whereas they accumulated a signifi cant amount of OA (C18:1 ⌬ 9 ), the major substrate for Tau ⌬ 12des ( Fig. 4A , B ; Table 1 ). Importantly, the downstream derivatives of LA in the standard pathway also decreased drastically in tau ⌬ 12desdisruption mutants, except for DHA, which was instead slightly increased. Furthermore, the C17:1 ⌬ 9 and C19:1 ⌬ 9 contents signifi cantly increased in the tau ⌬ 12desdisruption mutants, indicating that these odd-chain fatty acids are substrates for Tau ⌬ 12des. The loss of ⌬ 12-fatty acid desaturase activity was also confi rmed by the metabolic labeling of mutants using 14 C-oleoyl-CoA (supplementary Fig. IVA ); no 14 C-LA was found in the mutants, in contrast to the wild-type. The accumulation of OA and the decrease of LA and its downstream PUFAs in the standard pathway were observed not only in the total fatty acid fraction but also in each lipid class (i.e., neutral lipids, phospholipids, and glycolipids) of the tau ⌬ 12desdisruption mutants ( Table 2 ). Despite the signifi cant changes in the fatty acid profi les in the total fatty acid and complex lipid fractions, no difference was observed in cell growth between the wild-type strain and the tau ⌬ 12des -disruption mutants under our cultivation conditions (Supplementary Fig. V ).

Restoration of the fatty acid profi le in revertants of the tau ⌬ 12des -disruption mutants
To complement the tau ⌬ 12des in the tau ⌬ 12desdisruption mutants, a Neo r /Tau ⌬ 12des-expression construct ( Fig. 5A ) was injected into the disruption mutants by microprojectile bombardment. As a control, a Neo r -expression construct was injected into other tau ⌬ 12des -disruption mutants ( Fig. 5B ). Transformants grown on GY medium containing G418 were selected as transformants and subjected to genomic PCR to determine whether a full-length Neo r / Tau ⌬ 12des DNA was integrated into the genome of the transformants. As shown in Fig. 5C , a 5,306-bp PCR product (corresponding to the size of the Neo r /Tau ⌬ 12des construct [ Fig. 5A] ) was detected in the transformants harboring Neo r /Tau ⌬ 12des DNA (tentatively designated revertants), whereas a 2,717-bp PCR product (corresponding to the size of the Neo r construct [ Fig. 5B] ) was detected in the transformants harboring the Neo r control construct (KO/ neo r ). Southern blot hybridization using a Tau ⌬ 12des DNA probe confi rmed that the tau ⌬ 12des was integrated into the genomes of the revertants ( Fig. 5D ). Furthermore, RT-PCR revealed that transcripts of the neo r (835 bp) and the tau ⌬ 12des (1,185 bp) were present in the revertants, whereas the transcript of the neo r , but not the tau ⌬ 12des , was detected in the control ( Fig. 5E, F ). These results clearly indicate that tau ⌬ 12des was integrated into the genome of the revertants and then transcribed into Tau ⌬ 12des mRNA.
The fatty acid compositions of the revertant and KO/ neo r were analyzed by GC using their respective FAMEs. In contrast to the KO/ neo r , the fatty acid profi le of the revertants was restored to that of the wild-type strain (i.e., the levels of OA, LA, and its downstream PUFAs in the revertants were similar to those of the wild-type strains) ( Fig. 6A , B ; Table 3 ). Additionally, in vivo labeling with 14 C-oleoyl-CoA demonstrated the restoration of the ⌬ 12-fatty acid desaturase activity in the revertants (supplementary Fig. IVB ). These results clearly indicate that the change in the fatty acid profi le in the tau ⌬ 12des -disruption mutants was due to the loss of function of tau ⌬ 12des.
It is shown in this study that Tau ⌬ 12des is the ⌬ 12-fatty acid desaturase involved in the standard pathway and that this enzyme is primarily responsible for the conversion of OA into LA in T. aureum . Furthermore, DHA was found to be produced in T. aureum primarily independently of the standard pathway, possibly via the PUFA synthase pathway. In conclusion, two working pathways for the production of PUFAs in T. aureum were revealed through the analysis of a native ⌬ 12-fatty acid desaturase.

DISCUSSION
Thraustochytrids, belonging to the protist kingdom Stramenopila, are microorganisms that constitute a promising alternative to fi sh oils as an industrial source of PUFAs. The fatty acid profi les differ among the different thraustochytrid genera ( 19 ). These different PUFA profi les may indicate the presence of different PUFA biosynthetic pathways in the various thraustochytrids. Several lines of evidence suggest the occurrence of two different pathways involved in the biosynthesis of PUFAs in thraustochytrids. The fi rst, which is found in several marine bacteria, is the polyketide synthase-like pathway (the PUFA synthase pathway), comprising reiterative cycles including condensation, reduction, dehydration, and isomerization steps, with each step catalyzed by different enzymes. Three functional ORFs of the PUFA synthase pathway have been identifi ed in Schizochytrium (now reassigned to Aurantiochytrium ) ( 9, 20, 21 ). Lippmeier et al. (10) suggested that the PUFA synthase pathway is the sole system responsible for PUFA production in Schizochytrium because the disruption of an ORF of a PUFA synthase led to the loss of PUFAs in the thraustochytrids, which became PUFA-dependent auxotrophs. The other pathway, which is found in many organisms, including mammals, is the desaturase/elongase pathway (the standard pathway), comprising a series of alternating desaturation and elongation steps starting with saturated fatty acids that are produced in an FAS pathway. Although several desaturase and elongase genes have been cloned and characterized in thraustochytrids (11)(12)(13)22 ), the direct evidence that such enzymes are operative in the standard pathway has not been obtained. In this study, we demonstrated that the standard pathway is functional in T. aureum ATCC 34304 by disrupting the gene encoding a ⌬ 12-fatty acid desaturase, which is a key enzyme in the standard pathway for the production of n-3 and n-6 PUFAs.
In this study, we generated disruption mutants of tau ⌬ 12des by replacing two tau ⌬ 12des alleles with two different marker genes. The disruption construct was composed of the 5 ′ and 3 ′ regions of the tau ⌬ 12des as homologous recombination sites and an antibiotic-resistance gene ( hyg r or bla r ) as a marker gene ( Supplementary  Fig. III ). Molecular analysis of the tau ⌬ 12des -disruption mutants showed that the tau ⌬ 12des ORFs of two alleles were replaced by hyg r or bla r ( Fig. 3 ). This result indicates that T. aureum is diploid, at least under the conditions used in this study. In contrast, Schizochytrium sp. ATCC 20888 appeared to be haploid ( 10 ).
Unexpectedly, the tau ⌬ 12des -disruption mutants of T. aureum were indistinguishable from the wild-type strain in morphology and cell growth under the conditions used in this study ( Supplementary Fig. V ). However, the disruption of the tau ⌬ 12des led to a dramatic change in the fatty acid profi le, in which an increase of OA (C18:1 ⌬ 9 ) was observed in combination with the disappearance of LA (C18:2 ⌬ 9, 12 ) ( Fig. 4 ; Table 1 ). Furthermore, the tau ⌬ 12desdisruption mutants showed decreased levels of the n-6 and n-3 PUFAs that are downstream of LA in the standard pathway. In contrast, DHA levels were slightly increased in the disruption mutants. These results demonstrate that Tau ⌬ 12des functions in the standard pathway for the production of PUFAs, whereas DHA is primarily produced by a nonstandard pathway in T. aureum , possibly by the PUFA synthase pathway. However, we observed that the disruption of the PUFA synthase gene in T. aureum resulted in a marked decrease in DHA but not in other PUFAs, such as LA, ARA (C20:4 ⌬ 5, 8, 11, 14 ) and EPA (C20:5 ⌬ 5, 8, 11, 14, 17 ). A small amount of DHA was still present in the PUFA synthase-disrupted mutants, suggesting that DHA is produced not only by the PUFA synthase pathway but also by the standard pathway (data not shown). Neither Tau ⌬ 12des nor PUFA-synthase disruption mutants of T. aureum were auxotrophs, in contrast to PUFA-synthase mutants of Schizochytrium sp ( 10 ). We observed the accumulation of C17:1 ⌬ 9 and C19:1 ⌬ 9 in the tau ⌬ 12des -disruption mutants, and this accumulation was eliminated by introducing tau ⌬ 12des into the disruption mutants. This result indicates that Tau ⌬ 12des also accepts odd-chain fatty acids as substrates. Chang et al. (23) identifi ed odd-chain PUFAs in thraustochytrids and suggested that these PUFAs are synthesized through the standard pathway. The accumulation of C17:1 ⌬ 9 and C19:1 ⌬ 9 in tau ⌬ 12des -disruption mutants supports their hypothesis. The enzymes involved in the PUFA synthase pathway are cytosolic proteins, and the products are released from the synthetic machinery as free fatty acids ( 20 ). In contrast, the membrane desaturases accept a wide range of acyl substrates ( 24,25 ). Therefore, we expected that the fatty acid profi les of complex lipids from the tau ⌬ 12desdisruption mutants would be somewhat different from those of the wild-type strain, as shown in fi lamentous fungus ( 26 ). However, no difference was observed between the wild-type strain and the disruption mutants in their fatty acid profi les of neutral lipids, phospholipids, and glycolipids ( Table 2 ). Thus, it is possible that the fatty acids produced via the standard pathway are acyl-CoA forms that are directly incorporated into different complex lipids by various acyltransferases. In other words, the desaturases and elongases that constitute the standard pathway of the thraustochytrid accept the CoA forms of fatty acids as substrates. Another possibility is that the fatty acids in each lipid class are remodeled in T. aureum after their incorporation into complex lipids. Further studies are necessary to elucidate the acceptor specifi city of Tau ⌬ 12des in vitro.
In conclusion, we present direct evidence that Tau ⌬ 12des functions in the standard pathway and is responsible for generating certain PUFAs in T. aureum , although DHA is primarily produced through the PUFA synthase pathway ( Fig. 7 ). The results of this study also indicate that LA and its downstream products from the standard pathway are not necessary for the normal growth and morphology of T. aureum under our conditions because suffi cient DHA is generated through the PUFA synthase pathway.