Identification and Functional Analysis of Delta-9 Desaturase, a Key Enzyme in PUFA Synthesis, Isolated from the Oleaginous Diatom Fistulifera

Oleaginous microalgae are one of the promising resource of nonedible biodiesel fuel (BDF) feed stock alternatives. Now a challenge task is the decrease of the long-chain polyunsaturated fatty acids (PUFAs) content affecting on the BDF oxidative stability by using gene manipulation techniques. However, only the limited knowledge has been available concerning the fatty acid and PUFA synthesis pathways in microalgae. Especially, the function of Δ9 desaturase, which is a key enzyme in PUFA synthesis pathway, has not been determined in diatom. In this study, 4 Δ9 desaturase genes (fD9desA, fD9desB, fD9desC and fD9desD) from the oleaginous diatom Fistulifera were newly isolated and functionally characterized. The putative Δ9 acyl-CoA desaturases in the endoplasmic reticulum (ER) showed 3 histidine clusters that are well-conserved motifs in the typical Δ9 desaturase. Furthermore, the function of these Δ9 desaturases was confirmed in the Saccharomyces cerevisiae ole1 gene deletion mutant (Δole1). All the putative Δ9 acyl-CoA desaturases showed Δ9 desaturation activity for C16∶0 fatty acids; fD9desA and fD9desB also showed desaturation activity for C18∶0 fatty acids. This study represents the first functional analysis of Δ9 desaturases from oleaginous microalgae and from diatoms as the first enzyme to introduce a double bond in saturated fatty acids during PUFA synthesis. The findings will provide beneficial insights into applying metabolic engineering processes to suppressing PUFA synthesis in this oleaginous microalgal strain.


Introduction
Biodiesel fuel (BDF) has attracted considerable attention over the past decade as a renewable and biodegradable fuel alternative to fossil fuels. Commercially available BDFs are produced from a variety of terrestrial plants, including soybean, rapeseed, sunflower, castor seed, jatropha and palm oil. Terrestrial plants potentially have a negative impact on food supply. Furthermore, they have lower oil yield per area than oleaginous microalgae [1,2,3]. Based on the reasons, recently, oleaginous microalgae have been intensively studied as non-food biomass and high-triacylglyceride (TAG) producer for efficient BDF production [1,4].
BDF is a series of fatty acid methyl esters (FAMEs) generated by transesterification of TAG from feedstocks [5]. The physical and chemical properties of FAMEs are determined by its acyl composition, with respect to both carbon chain length and the number of double bonds. As the degree of unsaturation of fatty acids in FAMEs particularly affects the oxidative stability of BDF [6], the unsaturated fatty acid content in BDF is a primary limitation to its commercial use [7]. BDFs from soybean, sunflower and grape seed contain high levels of polyunsaturated fatty acids (PUFAs), resulting in poor oxidative stability [8,9]. On the other hand, BDFs from rapeseed, olive, corn, almond and high oleic sunflower oils show superior BDF properties because of their high content of monounsaturated [8]. Microalgal TAG mainly consists of short and saturated fatty acids, however, non-negligible quantities of long-chain PUFAs, such as methyl linolenate (C18:3), eicosapentaenoic acid (EPA; C20:5) or docosahexaenoic acid (DHA; C22:6) are also involved [10,11].
Toward addressing the above issue, breeding efforts have been done in terrestrial plants. The PUFA contents of TAG have been successfully reduced by the suppression of desaturase gene expression using RNA interference (RNAi) system in soybean, cotton seed and brassica seed [12,13,14]. By contrast, in microalgae, genetic modifications of FAME profiles have been hampered by the limited knowledge available concerning the fatty acid synthesis pathway (including PUFA synthesis) and/or by difficulties in the genetic engineering approach [15,16]. Among eukaryotic microalgal groups, diatoms are well-established in terms of genomic and transgenic capabilities. Furthermore, the enzymes involved in fatty acid synthesis have been primarily identified in a model diatom, Phaeodactylum tricornutum [17,18,19].
In P. tricornutum, v3, D5, D6 and D12 desaturases were responsible for PUFA synthesis [18,19]. The end-product of the pathway is EPA. However, among various desaturases, the function of D9 desaturase from diatom has not been determined, although the enzyme plays a key role in PUFA synthesis as the first enzyme to introduce a double bond into saturated fatty acids [19].
A marine oleaginous diatom, Fistulifera sp. used in this study, has been recognized as a potential candidate for BDF production [20] because of its exceedingly high levels of intracellular TAGs (60% w/w) and its rapid growth. High-cell-density cultivation and outdoor mass cultivation of Fistulifera sp. have been demonstrated in flat-type photobioreactors [21], and column-type and racewaytype bioreactors [22]. In this strain, the major fatty acids are palmitate (C16:0; 30-40% of total fatty acids), palmitoleate (C16:1; 40-50%) and eicosapentaenoic acid (EPA, C20:5; 4-20%) as a PUFA. Recently, genetic transformation for this strain was performed [23]. Metabolic engineering with the gene manipulation technique is a promising approach to decrease the PUFA content in TAG. One of the targets for genetic transformation was D9 desaturase because they may play a key role in fatty acid (and subsequent TAG) synthesis [12,13,14].
In this study, we report the screening of D 9 desaturase genes in the oleaginous diatom Fistulifera and their functional characterization by expression in the yeast Dole1 mutant. Through the comparison of the isolated D 9 desaturases with those from other diatoms, unique features of D 9 desaturase genes in Fistulifera sp. were determined. To our knowledge, this is the first study to confirm the function of D 9 desaturases in diatoms and also in oleaginous microalgae.
Isolation of D 9 desaturase Genes from Fistulifera sp.
To obtain the putative D 9 desaturase genes of Fistulifera sp., a homology search using BlastX was performed with reference to Delta-9 Desaturases from Oleaginous Diatoms PLOS ONE | www.plosone.org the 19,859 genes from the draft genome sequence of Fistulifera sp. [26]. The full-length cDNAs of putative D 9 desaturase genes were obtained by 59-and 39-RACE using a Smarter RACE cDNA amplification kit (Clontech, Palo Alto, CA, USA). Partial sequences of these genes predicted by the AUGUSTUS program were used for designing gene-specific primers to amplify the 59 and 39 ends of the target genes (Table S1). The PCR products were cloned into the pCR-Blunt II-TOPO vector (Invitrogen). The fulllength cDNA sequences were assembled based on the 59-and 39-RACE fragments.
Functional Characterization of D 9 Desaturases in the Yeast Dole1 Mutant For functional characterization, 4 D 9 desaturase genes (fD9desA, fD9desB, fD9desC, and fD9desD) with the Kozak sequence [33] in front of the start codon were cloned into the yeast expression vector pYES2.1/V5-His-TOPO (Invitrogen) ( Table S1). The yeast Dole1 mutant was transformed with plasmid DNA with a polyethylene glycol/lithium acetate protocol [34]. The yeast cells harboring the control pYES2.1/V5-His/lacZ were used as a negative control. All transformants were selected by uracil prototrophy on a selective dropout media (SD) plate lacking uracil. For functional expression, SD medium containing 2% (w/v) galactose, 1% Tergitol Type NP-40 (Invitrogen), and 500 mM C16:1 or C18:1 fatty acids was inoculated with the pYES2.1FsDES9 transformants and grown at 20uC for 96 h in a water bath shaker. Cell pellets were sequentially washed with 1% Tergitol Type NP-40 and 0.5% Tergitol Type NP-40, freezedried, and subject to fatty acid analysis.

Complementation Assay in the Yeast Dole1 Mutant
Each transformant harboring the plasmid for the expression of D 9 desaturases was suspended in distilled water and adjusted to an OD 600 of 1 and 0.1. The resulting 2.5 mL yeast solution was spotted on an SD agar plate lacking uracil but containing 2% galactose and (a) no fatty acids; (b) 500 mM C16:1 fatty acid and 1% Tergitol Type NP-40; or (c) 500 mM of the C18:1 fatty acid and 1% Tergitol Type NP-40. Cell growth was evaluated after 72 h at 30uC to examine the unsaturated fatty acid requirement for the growth of yeast Dole1 mutant transformants.

Fatty Acid Analysis Using GC/MS
The freeze-dried yeast cells were directly transmethylated with 1.25 M hydrochloric acid in methanol (1 h at 100uC) to prepare the FAMEs. The FAMEs were extracted in n-hexane and analyzed  Table 2. Substrate specificity analysis of D 9 desaturases (fD9desA, fD9desB, fD9desC, and fD9desD) from Fistulifera sp. on the basis of expression in the yeast Dole1 mutant (n = 3).

Results
Sequence Analysis of Putative D 9 Desaturase Genes from Fistulifera sp. The 6 D 9 desaturase candidate genes (Gene ID: g5394, g10778, g10781, g12958, g19483 and g19486) were identified with the BlastX algorithm from all 19,859 genes of this strain as the query sequence from a non-redundant protein sequences database [Manuscript in preparation]. Because the predicted D 9 desaturases from draft genome sequence seemed to be partial ORFs due to the lack of the conserved motifs of histidine box and cytochrome b 5 domain, the full-length sequences of putative acyl-CoA D 9 desaturase cDNA were sequenced from the products obtained by rapid amplification of cDNA ends (RACE) PCR. The cDNAs containing predicted gene regions were verified to be 996 bp for g10778 and g19483 and 1,002 bp for g10781 and g19486; these were designated as fD9desA, fD9desB fD9desC, and fD9desD, respectively (Fig. 1). The fD9desA nucleotide sequences exhibit high identity with the fD9desB (96%), and the fD9desC had 93% identity with the fD9desD. The amino acid sequences of fD9desA and fD9desB were identical, while fD9desC and fD9desD showed the differences of 2 amino acid residues. The 2 proteins encoded by g5394 and g12958 had 49% and 47% identity, respectively, with D 9 desaturase from the plant Asclepias syriaca (GenBank accession no. AAC49719.1) [35]. The 4 proteins encoded by g10778, g10781, g19483 and g19486 showed 78%, 69%, 72% and 71% identity, respectively, with the putative D 9 desaturase from a diatom, P. tricornutum (EEC47008.1). A phylogenetic tree of D 9 desaturase amino acid sequences from different organisms, prepared using ClustalW, showed that the g5394 and g12958 genes appeared in the cluster of the plastidial acyl-ACP desaturase group, while the remaining genes were categorized in acyl-CoA D 9 desaturase groups localizing in the endoplasmic reticulum (ER) (Fig. 2). As the PUFA synthesis occurs in the ER of eukaryotic cells [19], these 4 genes in the acyl-CoA D 9 desaturase groups were further investigated. In model strains of pennate and centric diatoms, P. tricornutum (EEC47008), and Thalassiosira pseudonana (EED91785 and EED86245), ER acyl-CoA desaturases and plastidal acyl-ACP desaturases were also identified by a BLAST search using the predicted genes from Fistulifera sp. as the query sequence (Table 1).
A unique histidine motif is generally found in D 9 desaturases [36]. Four D 9 desaturases genes in Fistulifera sp. possessed 3 histidine motifs (HxxxxH, HxxHH, and HxxHH) involved in the coordination of the di-iron center as the active site of desaturation (Table 1). On the other hand, the cytochrome b 5 domain, which is expected to transfer electrons away from the histidine motif, was not observed in the identified D 9 acyl-CoA desaturase. These unique motifs of D 9 desaturases were compared with those in 2 diatoms, P. tricornutum and T. pseudonana [37]. These enzymes from the 3 diatoms were well conserved, possessing the histidine sequences without the cytochrome b 5 domain. The findings concerning the lack of cytochrome b 5 in desaturases is in agreement with those of a previous report about alternative desaturation activity in the absence of the cytochrome b 5 domain in 2 other desaturases from T. pseudonana [38]. Furthermore, the results of truncation and disruption experiments for the cytochrome b 5 domain D 9 desaturases have suggested that this domain is not strictly required for fatty acid desaturation [39].
Localization of general D 9 desaturases in the ER membrane are predicted from N-terminal amino acid sequences [19]. Three algorithms for subcellular targeting prediction, TargetP 1.1, SignalP 4.0 and HECTAR, were used in this study. The D 9 desaturase candidates were predicted to have neither N-terminal signal sequences nor internal cleavable signal sequences. All 4 D 9 desaturases were expected to possess 2-5 transmembrane domains, according to a prediction by TMHMM 2.0 and HMMTOP 2.0 (Table 1). This prediction supports the existence of transmembrane helices in D 9 desaturases from Fistulifera sp., although further analysis is needed to confirm the number of transmembrane regions.

Functional Characterization of Putative D 9 Desaturases in the Yeast Dole1 Mutant
In order to confirm the function of D 9 desaturases, 4 putative genes were cloned in the protein expression vector, pYES2.1/V5-His-TOPO, and transformed in the yeast Dole1 mutant. Dole1 is a D 9 desaturase knockout mutant that requires supplementation with C16:1 or C18:1 fatty acids to grow. When the putative D 9 desaturases work properly in the synthesis of monounsaturated C16:1 and/or C18:1 fatty acids, the transformants can grow on an agar plate without supplementation with exogenous unsaturated fatty acids. The transformed yeast Dole1 mutant with the control vector (in the absence of any desaturase gene), serving as a negative control, did not grow in the absence of fatty acid supplementation ( Fig. 3(a), no UFA). The INVSc-1 cells possessing the native ole1 gene in the genome, serving as a positive control, grew well in the absence of fatty acid supplementation ( Fig. 3(b), no UFA) because the cell can generate these essential desaturated fatty acids endogenously. The fD9desA, fD9desB, fD9desC, and fD9desD genes failed to complement the yeast ole1 mutation in the absence of 16:1/18:1 fatty acid supplementation (Fig. 3(c)-(f), no UFA). Additionally, the growth of transformants with 16:1/18:1 fatty acid supplementation was also confirmed. In a previous study, a D 9 desaturase (D9-3) from the fungus Mortierella alpina failed to complement the yeast Dole1 mutant when transformants were grown in the absence of monounsaturated fatty acids supplementation [40]. It is likely that the D9-3 desaturase from M. alpina and D 9 desaturases from Fistulifera sp. did not have sufficient activity for the complementation in yeast, respectively.
Next, to further investigate the in vivo function and specificity of these D 9 desaturase genes, the fatty acid profiles of the yeast Dole1 transformant-carrying gene expression vector with the D 9 desaturase genes were evaluated by gas chromatography and mass spectrometry (GC/MS) analysis (Table 2). Two D 9 desaturases from Fistulifera sp., fD9desC and fD9desD, showed activity for C16:0 as a substrate (0.6% and 0.3% desaturation to C16:1, respectively) (100 6 product/substrate+product) during growth with C18:1 supplementation (Table 2) but did not show activity for C18:0. The remaining 2 D 9 desaturases, fD9desA and fD9desB, showed similar activities to each another with C16:0 as a substrate (0.8% and 1.2% desaturation to C16:1, respectively). However, the ideal substrate of fD9desA and fD9desB was C18:0 converted to C18:1 (6.9% and 5.5% desaturation to C18:1). The desaturation efficiencies of fD9desA and fD9desB for C18:0 were significantly higher than those for C16:0. The detected value of desaturation by D 9 desaturases from Fistulifera sp in yeast was similar to that of house cricket (Acheta domesticus) (5% desaturation to 18:1) [41]. These heterogeneous gene expressions and functional analyses in the model organism may cause the negative effect for the activity, exhibiting these relatively low activities.

Discussion
Six D 9 desaturase genes were identified in Fistulifera sp. by bioinformatic analysis. These genes were categorized as 4 ER acyl-CoA desaturases (fD9desA, fD9desB, fD9desC, and fD9desD) and 2 plastidial acyl-ACP desaturases (g5394 and g12958). By way of comparison, P. tricornutum has only 1 ER acyl-CoA desaturase and 1 plastidial acyl-ACP desaturase, and T. pseudonana has 2 ER acyl-CoA desaturases and 1 plastidial acyl-ACP desaturase (Table 1). These results indicate that Fistulifera sp. has more D 9 desaturase genes than other diatoms do. The higher number of D 9 desaturase genes in Fistulifera sp. suggests that the oleaginous strain may have well-developed genome organization for fatty acid metabolism to enable considerable accumulation of TAG endogenously. The identified ER acyl-CoA desaturases show different substrate specificities ( Table 2). Two D 9 desaturases (fD9desA and fD9desB) converted both C16:0 and C18:0 to C16:1 and C18:1, respectively, and the other 2 D 9 desaturases (fD9desC and fD9desD) converted C16:0 to C16:1. The sequences of the D 9 desaturases showed the presence of highly conserved histidine cluster motifs [42] (Fig. 2), and a difference of 1 amino acid residue in the first histidine motif box was observed among these 4 D 9 desaturases. The former group of D 9 desaturases (fD9desA and fD9desB) has the HRLWSH sequence, and the latter group of D 9 desaturases (fD9desC and fD9desD) has the HRLWAH sequence, with the serine substituted with alanine. In other words, D 9 desaturase pairs from Fistulifera sp. have the same histidine motifs and desaturation specificities ( Table 2). In the oil-producing fungus M. alpina, a different amino acid sequence in the first histidine box was found in 3 D 9 desaturases. Different D 9 desaturases with different sequences in the first histidine box showed varying specificities for fatty acid desaturation [40]. Therefore, the difference in amino acid sequences is considered to reflect the substrate specificity of D 9 desaturases from Fistulifera sp. P. tricornutum has only 1 ER acyl-CoA desaturase with the amino acid sequence of HRLWSH in the first histidine motif, while T. pseudonana has 2 ER acyl-CoA desaturases with the same amino acid sequence (HRLWSH; Table 1). On the basis of these results, we hypothesized that all the D 9 desaturases in both diatoms possess the same substrate specificity, although a detailed functional analysis should be performed to confirm this hypothesis. If it is confirmed, this would indicate that the varying specificities of D 9 desaturases from Fistulifera sp. may be related to the specific fatty acid metabolism in this strain. In addition, the D 9 desaturation activity was assayed for the substrates of 16:0 and 18:0. In the case of mouse D 9 desaturase, for wide range of saturated fatty acids from 12:0 to 19:0, the desaturation activity was detected [43]. To fully confirm the functional specificity of D 9 desaturase from Fistulifera sp. for various substrates, further analysis should be provided in the future.
In conclusion, 4 ER D 9 acyl-CoA desaturase and 2 plastidial D 9 acyl-ACP desaturase genes were identified from the oleaginous diatom Fistulifera sp., and gene homologs were also observed in other diatoms. Among the D 9 desaturase genes, four D 9 acyl-CoA desaturases showed desaturation activity of the saturated fatty acids of C16:0 and/or C18:0 in complementation assays using the yeast Dole1 mutant. This study is the first functional confirmation of D 9 desaturase from diatom and from oleaginous microalgae. While the in vivo function of the D 9 desaturase genes in the Fistulifera sp. should be separately addressed, this study provides information that will help in the regulation of the fatty acid profiles in the oleaginous microalgae.

Supporting Information
Table S1 List of primers used for this study. (DOCX)