Introduction

Oleaginous yeasts, such as Lipomyces starkeyi, Yarrowia lipolytica, Cryptococcus curvatus, and Rhodosporidium toruloides, are a potential cost-effective means of producing microbial oils without competing with food production. These microorganisms are able to store intracellular lipids to approximately 60% of their cell dry weight (Ageitos et al. 2011; Kosa and Ragauskas 2011; Donot et al. 2014). Oleaginous yeasts boast several advantages for lipid production over other oleaginous microorganisms, such as molds or algae, due to their unicellular form and high growth rates. In addition, oleaginous yeasts are able to produce lipids from a large variety of renewable substrates, including sugars derived from non-edible biomasses. The genetic engineering of oleaginous yeasts is a powerful tool for optimizing both the quantity and quality of produced lipids (Wang et al. 2013; Shi and Zhao 2017). For example, strains of Y. lipolytica were engineered to produce the ω-3 fatty acid, eicosapentaenoic acid (EPA) (Xue et al. 2013). L. starkeyi, which produces triacylglycerols (TAGs) to more than 70% of its dry cell weight (Angerbauer et al. 2008), is one of the most attractive and well-studied oleaginous yeasts. The complete genome of L. starkeyi is available from the Joint Genome Institute (JGI) website (http://genome.jgi.psf.org/Lipst1_1) (Riley et al. 2016). In addition, several genetic engineering tools for L. starkeyi have already been developed, including transformation (Calvey et al. 2014), multicopy integration and expression of heterologous genes (Oguro et al. 2015), and efficient gene targeting for gene deletion and integration using non-homologous end-joining-deficient strains (Oguro et al. 2017).

Fatty acids play vital roles in the functions of cytoplasmic and organelle membranes, and in carbon source storage. In L. starkeyi, major fatty acid species are saturated fatty acids with 16 carbon atoms (palmitic acid, C16:0) and monounsaturated (single double bond) fatty acids containing 18 carbon atoms (oleic acid, C18:1). In the most studied model yeast, S. cerevisiae, palmitic acid (C16:0) is synthesized from acetyl-CoA and NADPH by acetyl-CoA carboxylase and fatty acid synthase. Monounsaturated fatty acids are synthesized from saturated fatty acids by Δ9 fatty acid desaturase Ole1p (Stukey et al. 1989; Martin et al. 2007; Henry et al. 2012). Fatty acid elongases convert 16-carbon fatty acids to 18-carbon fatty acids (Toke and Martin 1996; Schneiter et al. 2000).

In addition to palmitic acid (C16:0) and oleic acid (C18:1), L. starkeyi contains small amounts of polyunsaturated fatty acids, C18:2 (linoleic acid) and C18:3 (Α-linolenic acid), abbreviated here as LA and ALA, respectively. In L. starkeyi and many other yeast species, second and third double bonds are introduced by Δ12 and Δ15 fatty acid desaturases, respectively. Previously, the enzymatic activity of L. starkeyi Δ12 fatty acid desaturase was measured using microsomal system extracted from L. starkeyi cells (Lomascolo et al. 1996). This study yielded the following information: (1) the optimal pH was between pH 7 and 8, (2) thermal stability was low, (3) the enzyme was inhibited by Hg2+ and activated by Mg2+, Mn2+, and Zn2+, and (4) oleoyl-CoA was the preferred substrate (Lomascolo et al. 1996). However, the gene(s) encoding Δ12 fatty acid desaturase has (have) not been identified. The identification of Δ12 fatty acid desaturase gene(s) is essential for the modification of lipid content, the production of polyunsaturated fatty acids, and the physiological analyses of unsaturated fatty acids in L. starkeyi cells. Some Δ12 and Δ15 fatty acid desaturases have been identified in yeasts, plants, and animals (Watanabe et al. 2004; Wei et al. 2004; Kainou et al. 2006; Wei et al. 2006; Matsuda et al. 2012; Radovanovic et al. 2014; Sangwallek et al. 2014; Angelis et al. 2016; Lee et al. 2016; Sun et al. 2016). In addition, some fatty acid desaturases are reportedly bifunctional, exhibiting both Δ12 and Δ15 desaturase activities (Damude et al. 2006; Zhou et al. 2011; Yan et al. 2013; Cui et al. 2016). Both Δ12 and Δ15 fatty acid desaturases are key enzymes for the production of ω-3 polyunsaturated fatty acids, including EPA and docosahexaenoic acid (DHA).

In this study, we identified and characterized Δ12 and Δ12/Δ15 bifunctional fatty acid desaturases, termed LsFad2 and LsFad3, respectively, in L. starkeyi. Genes encoding these desaturases were not essential for L. starkeyi growth, but they contributed significantly to the production of polyunsaturated fatty acids, especially at low temperatures. The overexpression of LsFAD2 or LsFAD3 induced an accumulation of LA or ALA, respectively, indicating that these genes encoded key enzymes for the production of polyunsaturated fatty acids including ω-3 polyunsaturated fatty acids.

Materials and methods

Strains and media

The L. starkeyi and S. cerevisiae strains used in this study are listed in Table 1. L. starkeyi strains were grown under aerobic conditions at 30 or 20 °C for the indicated time in GY (5% glucose and 0.5% yeast extract) or YPD (2% glucose, 2% peptone, and 1% yeast extract) medium. When necessary, 50 μg/ml zeocin (Nacalai Tesque, Kyoto, Japan) and/or 100 μg/ml hygromycin B (Wako Pure Chemical Industries, Osaka, Japan) were added to the YPD medium. Escherichia coli DH5α competent cells were purchased from Nippon Gene (Tokyo, Japan).

Table 1 Yeast strains used in this study
Full size table

Genetic manipulation

Transformation of L. starkeyi was performed as described previously (Oguro et al. 2015), and transformants were selected on YPD medium containing 100 μg/ml of hygromycin B. Total DNA was extracted using the GenTLE (Yeast) High Recovery system (Takara Bio, Shiga, Japan). Total RNA was extracted using the NucleoSpin RNA system (Takara Bio) after mechanical cell disruption.

Expression of LsFAD2 and LsFAD3 genes in S. cerevisiae

pL1091-5/DGA1∆N (DGA1∆N encodes Dga1p with 29 N-terminal amino acids deleted) was constructed as described previously (Kamisaka et al. 2013). The L. starkeyi ∆12 desaturase gene (termed LsFAD2) (protein ID: 68481) was amplified by polymerase chain reaction (PCR) with a high-fidelity DNA polymerase (PrimeSTAR GXL; Takara Bio) using L. starkeyi cDNA as a template with the following primers: 5′-GCAAGCTTATGTCCACAATAACATAC-3′ (the HindIII site is underlined) and 5′-ATGCGGCCGCTTACTGAGCCCTTCTT-3′ (the NotI site is underlined). The amplified product was excised as a HindIII-NotI fragment and used to construct pL1177-2/LsFAD2. The L. starkeyi ∆12/∆15 desaturase gene (termed LsFAD3) (protein ID: 310193) sequence was codon-optimized for expression in S. cerevisiae and synthesized by GenScript (Piscataway, NJ, USA). This was then excised as a HindIII-NotI fragment and used to construct pL1177-2/LsFAD3. The correctness of the constructed vectors was verified by DNA sequencing.

S. cerevisiae BY4741 ∆snf2 mutant was purchased from Invitrogen Life Technologies (Carlsbad, CA, USA) (Table 1). The yeast cells were transformed by the lithium acetate method (Ito et al. 1983) using a transformation kit (Zymo Research; Orange, CA, USA). The ∆snf2 mutant was transformed by plasmids pL1091-5/ DGA1∆N (Kamisaka et al. 2013), pL1177-2 harboring LsFAD2 or LsFAD3 (described above), and pL2137-26 (Kainou et al. 2006). pL1091-5, pL1177-2, and pL2137-26 are yeast expression vectors with 2-μm replication origins, an ADH1 promoter, and a yeast selection marker (URA3, LEU2, HIS3, respectively) (Kainou et al. 2006). The transformed yeast cells were grown under aerobic conditions at 20 °C for 7 days in nitrogen-limited SD medium containing 0.17% Bacto-yeast nitrogen base without amino acids and ammonium sulfate (Wako Pure Chemical Industries), 5% glucose, 0.25% ammonium sulfate, and 20 μg/ml methionine. In some experiments, 0.2 mg/ml linoleic acid was added to the medium with 0.25% Tergitol NP-40.

Overexpression of LsFAD2 and LsFAD3 genes in L. starkeyi

The construct for the overexpression of LsFAD2 or LsFAD3 is shown in Fig. 1a. pKS-18S-hph-LsFAD2 and pKS-18S-hph-LsFAD3, which were used for the overexpression of LsFAD2 and LsFAD3 genes in L. starkeyi, were constructed by the insertion of artificially synthesized DNA between the PmeI and AvrII sites of pKS-18S-hph (Oguro et al. 2015). The hph gene encoded hygromycin B phosphotransferase. First, we designed a DNA sequence containing an LsTDH3 promoter region (PTDH3), the LsFAD2 or LsFAD3 gene, and a TDH3 terminator region (TTDH3), in which PmeI and AvrII sites were added to the 5′ and 3′ ends, respectively. The synthesized DNA sequences for the overexpression of LsFAD2 or LsFAD3 are shown in the Supplemental Materials, Table S1. These synthetic DNA fragments were created and cloned in pUC57 by GENEWIZ (South Plainfield, NJ, USA). Second, pKS-18S-hph plasmid was amplified by PCR using PrimeSTAR Max Polymerase (Takara Bio) with the following primers: 5′-ATTGCCCTAGGAACTAGCTCAAGGGACGTGCTATTCCCAC-3′ (the AvrII site is underlined) and 5′-ATTGCGTTTAAACATGTAGCGGGTGGTGATGGTGGAAC-3′ (the PmeI site is underlined). The resulting PCR products and synthesized DNA fragments were digested with AvrII and PmeI and ligated using Ligation High (TOYOBO, Osaka, Japan). The resultants were termed pKS-18S-hph-PTDH3-LsFAD2-TTDH3 and pKS-18S-hph- PTDH3-LsFAD3-TTDH3, respectively. The Δlslig4 strain was transformed with pKS-18S-hph-PTDH3-LsFAD2-TTDH3 or pKS-18S-hph-PTDH3-LsFAD3-TTDH3 digested by ApaI, and transformants were selected on YPD medium containing 100 μg/ml hygromycin B.

Fig. 1
figure 1

Construction of a overexpression and b deletion of LsFAD2 and LsFAD3 genes. PTDH3 and TTDH3 indicate the promoter and terminator regions of the L. starkeyi TDH3 gene, respectively. Homologous recombination is indicated by dotted lines

Full size image

Construction of LsFAD deletion mutants

The construct for the deletion of the LsFAD2 or LsFAD3 gene is shown in Fig. 1b. Two DNA fragments of 5′- and 3′-non-coding regions of the LsFAD2 gene (1033 and 1159 bp, respectively) were amplified by PCR using the genomic DNA of L. starkeyi and the following primers: 5′-GCTGGGTACCGGGCCCTAAAGTACAGAGCATCGTTTG-3′ and 5′-CGAAGTCGTTTAATTGGCGATGCTCTGCACTGGGTG-3′ (for the 5′-non-coding region of the LsFAD2 gene) and 5′-GTGGTGGGCGATGTATTTAATTAATTTAATTTATAACATTGTTTTAACACA-3′ and 5′-CTCGAGGGGGGGCCCGGACACAGACGGATCAAGCGTCTTGACT-3′ (for the 3′-non-coding region of the LsFAD2 gene), respectively. Two DNA fragments of the 5′- and 3′-non-coding regions of the LsFAD3 gene (922 and 1039 bp, respectively) were amplified by PCR using the genomic DNA of L. starkeyi and the following primers: 5′-CTGGGTACCGGGCCCGTAGAGATCTCACAATTGCTGTCAGAC-3′ and 5′-CGAAGTCGTTTAATTTGTTGTTCTAGTTCAGATATTCCG-3′ (for the 5′-non-coding region of the LsFAD3 gene) and 5′-GTGGTGGGCGATGTAGCGAGATCCTATCTTAGACTCTTCCAC-3′ and 5′-CTCGAGGGGGGGCCCAAGAATAATCTCCAAGGAATGAACATATAGCTAT-3′ (for the 3′-non-coding region of the LsFAD3 gene), respectively. The hph gene cassette containing the LsTHD3 promoter and terminator regions was amplified by PCR using pKS-18S-hph (Oguro et al. 2015) as a template with the following primers: 5′-AATTAAACGACTTCGCCAAACGGAGACCATGCATTC-3′ and 5′-TACATCGCCCACCACTACCACCTTGTTTCAACAA-3′. pBluescript KS (+) vector was amplified by PCR using pKS-18S-hph as a template with the following primers: 5′-GGGCCCCCCCTCGAGGTCGACGGTATCGATAA-3′ and 5′-GGGCCCGGTACCCAGCTTTTGTTCCCTTTAGTG-3′. These PCR fragments (pBluescript KS (+) vector, approximately 1 kbp of the 5′- and 3′-non-coding regions of LsFAD2 or LsFAD3 genes, hph gene cassette) were linearized using an In-Fusion HD cloning kit (Takara Bio). The resultant plasmid vectors were termed pKS-LsFAD2::hph and pKS-LsFAD3::hph, respectively.

Next, pKS-LsFAD2::hph and pKS-LsFAD3::hph were digested by ApaI and PvuII, respectively, and introduced into L. starkeyi Δlslig4 to construct LsFAD2 and LsFAD3 deletion mutants.

Lipid analysis

The direct transmethylation of all fatty acid residues in yeast cells was carried out using methanolic 10% (v/v) HCl and methylene chloride, as described previously (Kamisaka et al. 2006). Aliquots of yeast cultures were taken at the indicated times and collected by centrifugation at 1400×g for 5 min. Cell pellets were washed once with distilled water and dried at 105 °C for 3 h. The dried pellets were then resuspended in 1 ml 10% methanolic HCl and 0.5 ml methylene chloride. After incubating at 60 °C for 3 h, 1 ml of saturated NaCl and 1 ml of hexane were added to the mixture. The resulting fatty acid methyl esters were analyzed on a gas chromatography-mass spectrometer (GC-MS) (QP2010 SE; Shimadzu, Kyoto, Japan) equipped with a DB-225MS capillary column (30 m × 0.25 mm i.d., Agilent Technologies, Santa Clara, CA, USA). Total fatty acids were quantified using heptadecanoic acid methyl ester as an internal standard.

Real-time quantitative reverse transcription PCR

L. starkeyi cells (Δlslig4, LsFAD2- and LsFAD3-overexpressing strains) were pre-cultured in YPD medium at 30 °C for 3 days and inoculated into GY medium. The cultures were incubated at 150 rpm, at 20 or 30 °C for 2 or 4 days. After incubation, cells were collected by centrifugation at 11,000×g for 2 min and washed with distilled water. Total RNA was extracted using a NucleoSpin RNA system (Takara Bio) and reverse transcription PCR was performed using a PrimeScript II 1st-strand cDNA Synthesis Kit (Takara Bio). cDNA was synthesized from 100 ng of total RNA.

Real-time quantitative PCR (qRT-PCR) was performed using LightCycler 480 SYBR Green I Master and LightCycler 480 II (Roche, Basel, Switzerland). qRT-PCR conditions were as follows: 95 °C for 5 min, followed by 45 cycles of 95 °C for 10 s and 60 °C for 10 s. The melting program was 95 °C for 5 s and 65 °C for 1 min. Primer pairs used were as follows: 5′-GCCGGAGTACACAATCAAAGA-3′ and 5′-TAAGAGAGCGGTCGAAGCAG-3′ for the LsFAD2 gene, 5′-AGGGCGCGACAGCTACTAT-3′ and 5′-ATGATGCTGTGGAACAGGTG-3′ for the LsFAD3 gene, and 5′-ACCCAGATTGTCTTTGAGACG-3′ and 5′-GACAGCCTGGATGGAGACA-3′ for the LsACT1 gene (ID of JGI data base 67392). Gene expression data from qRT-PCR were analyzed using the ΔΔCt method (Livak and Schmittgen 2001). Gene expression ratios were normalized to actin (LsACT1 gene) using the 2−ΔΔCt method and determined by the following equations: ΔCt = Ct (target gene) − Ct (actin).

Nucleotide sequence accession number

The synthesized DNA sequence of LsFAD3 for the heterologous expression in S. cerevisiae was deposited in DDBJ/EMBL/GenBank under the accession number LC387462.

Results

Identification of putative ∆12 desaturases in L. starkeyi

Two putative ∆12 desaturases of L. starkeyi, termed LsFad2 and LsFad3, were identified by searching the BLAST database (https://blast.ncbi.nlm.nih.gov/) using amino acid sequences of Fusarium verticillioides ∆12 desaturase (GenBank accession no. ABB88515.1, Damude et al. 2006) and ∆12/∆15 bifunctional desaturase (GenBank accession no. ABB88516.1, Damude et al. 2006). The protein IDs of LsFad2 and LsFad3 were 68481 and 310193 in the JGI database, respectively. The length of the LsFAD2 and LsFAD3 genes was 1263 and 1330 bp, respectively, and both genes contained one putative intron region (LsFAD2: 54 bp, LsFAD3: 52 bp). The LsFAD2 and LsFAD3 genes encoded proteins containing 402 and 425 amino acids, respectively. LsFad2 showed a low amino acid sequence similarity to F. verticillioides ∆12 fatty acid desaturase (identity, 50%) and ∆12/∆15 bifunctional fatty acid desaturase (identity, 44%). Likewise, LsFad3 showed low amino acid sequence similarity with F. verticillioides ∆12 (identity, 44%) and ∆12/∆15 bifunctional (identity, 48%) fatty acid desaturases (Table 2). The LsFad2 and LsFad3 also showed low similarity with each other (identity, 50%). Among these four proteins, there is some consensus in the amino acid sequences, as indicated by the asterisk in Fig. 2. Both LsFad2 and LsFad3 contained three histidine-rich motifs, H-box 1 to 3 (Fig. 2), which are conserved in many fatty acid desaturases that act as potential ligands for non-heme iron atoms and are essential for desaturase activity (Shanklin et al. 1994; Avelange-Macherel et al. 1995; Shanklin et al. 1997; Shanklin and Cahoon 1998). Based on our analyses of the deduced amino acid sequences of LsFad2 and LsFad3 using TMHMM Server ver. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/) (Möller et al. 2001), LsFad2 and LsFad3 were predicted to have more than three and five transmembrane helices, respectively (Fig. 2). We also compared the deduced amino acid sequences of LsFad2 and LsFad3 to those of other fatty acid desaturases, including Pichia pastoris, Mortierella alpina, and Caenorhabditis elegans (Table 2). Both LsFad2 and LsFad3 showed similarities with ∆12 and ∆12/∆15 fatty acid desaturases isolated from yeast and fungi but showed little similarity with C. elegans ∆12 fatty acid desaturase. Since it is not possible to fully characterize the desaturase activities of LsFad2 and LsFad3 based only on their deduced amino acid sequences, both proteins were expressed in S. cerevisiae and L. starkeyi in order to obtain empirical activity data.

Table 2 Homology of LsFad2 and LsFad3 with other desaturases
Full size table
Fig. 2
figure 2

Sequence alignment of L. starkeyi LsFad2 and LsFad3 and F. verticillioides ∆12 fatty acid desaturase (FvD12D) and bifunctional ∆12/15 fatty acid desaturase (FvD12/15D). Amino acid sequences were compared using ClustalW (DNA Data Bank of Japan; http://www.ddbj.nig.ac.jp/) using default settings. Histidine-rich motifs (H-box) are indicated in bold and circled. Putative transmembrane helices are underlined

Full size image

Characterization of LsFad2 and LsFad3 using a heterologous expression system in S. cerevisiae

To characterize the enzymatic specificity of LsFad2 and LsFad3, these proteins were expressed in S. cerevisiae and cultured with or without LA, as described in the “Materials and methods” section. The main fatty acids in S. cerevisiae are palmitoleic acid (C16:1) and oleic acid (C18:1). Polyunsaturated fatty acids, including LA and ALA, were not detected in control cells (Table 3). S. cerevisiae cells that overexpressed LsFAD2 accumulated LA with reduced levels of oleic acid. In contrast, S. cerevisiae cells overexpressing LsFAD3 produced both LA and ALA. These results indicated that LsFAD2 and LsFAD3 encoded ∆12 fatty acid desaturase and ∆12/∆15 bifunctional fatty acid desaturase, respectively. S. cerevisiae control cells cultured in a medium containing LA were shown to contain LA in their lipids, indicating that additive fatty acids were absorbed and employed in the synthesis of lipids. In media containing LA, S. cerevisiae cells overexpressing LsFAD3 converted added LA to ALA and accumulated significantly more ALA than did cells overexpressing LsFAD3 in the absence of LA in the initial culture medium (Table 3). These results indicated that added LA was converted to ALA by LsFad3.

Table 3 Fatty acids composition of S. cerevisiae cells overexpressing LsFAD2 or LsFAD3
Full size table

Overexpression and deletion of LsFAD2 and LsFAD3 genes in L. starkeyi

Next, we constructed L. starkeyi strains that overexpressed or lacked LsFADs, as described in the “Materials and methods” section. The ∆lslig4 strain, which is a non-homologous end-joining-deficient strain, exhibited high homologous recombination efficiency (Oguro et al. 2017) and was used as a parent (control) strain in this study. The LsFAD2 and LsFAD3 genes were overexpressed using the LsTDH3 promoter (PTDH3) and a multicopy integration system, which introduced genes into the ribosomal DNA locus (Oguro et al. 2015). Using qRT-PCR and the genomic DNA of each overexpression strain, we confirmed that cells overexpressing LsFAD2 contained two copies of PTDH3-LsFAD2-TTDH3 cassettes in their genomes. Cells overexpressing LsFAD3 had four copies of PTDH3-LsFAD3-TTDH3 cassettes in their genomes (data not shown). The mRNA levels of the LsFAD2 gene in cells overexpressing the LsFAD2 gene and cultured at 20 or 30 °C for 2 days were approximately 4- to 5-fold higher than those of control cells (Fig. 3a). The mRNA levels of the LsFAD3 gene in cells overexpressing LsFAD3 were approximately 15- to 21-fold higher than those of control cells (Fig. 3b). The expression levels of LsFAD2 and LsFAD3 were higher during the log-phase of growth (cultured for 2 days) than in the stationary phase (cultured for 4 days) in cells overexpressing LsFAD2 and LsFAD3, respectively.

Fig. 3
figure 3

Expression levels of LsFAD2 and LsFAD3 genes. a Expression levels of LsFAD2 in control cells (empty bars) and cells overexpressing LsFAD2 (shaded bars) that had been cultured at 20 or 30 °C for 2 or 4 days. b Expression levels of LsFAD3 in control cells (empty bars) and cell overexpressing LsFAD3 (shaded bars) cultured at 20 or 30 °C for 2 or 4 days. The expression level of the actin gene was defined as unity. Error bars indicate standard deviation, n = 3

Full size image

In addition to palmitic and oleic acids, L. starkeyi (∆lslig4) cells produced LA (15.4%) and a small amount of ALA (0.2%) at 30 °C (Table 4). In cells cultured at 30 °C, the LA content was dramatically increased by the overexpression of LsFAD2, and the production of LA and ALA was completely abolished by deletion of the LsFAD2 gene. In contrast, at 20 °C, ∆lsfad2 cells produced only small amounts of both LA and ALA, suggesting that LsFad3 contributed to the production of polyunsaturated fatty acids at low temperatures. Cells overexpressing LsFAD3 accumulated more ALA compared to control cells. The ALA content of the LsFAD3 overexpressing strain cultured at 20 °C was approximately 3-fold higher than that of the same strain cultured at 30 °C.

Table 4 Fatty acids composition of L. starkeyi strains: LsFAD2 or LsFAD3 overexpressing strains (+LsFAD2 or +LsFAD3), and lsfad2 or lsfad3 deletion mutants (Δlsfad2 or Δlsfad3).
Full size table

Cells cultured at 20 °C contained higher levels of both LA and ALA than cells cultured at 30 °C (Table 4). In particular, the ALA content of control cells cultured at 20 °C was approximately 60-fold higher than that of control cells cultured at 30 °C. These results suggest that L. starkeyi produces polyunsaturated fatty acids as an adaptation to low temperatures. However, the expression levels of both LsFAD2 and LsFAD3 genes were not enhanced under low temperature conditions (Fig. 3). This is discussed in greater detail below.

Discussion

Fatty acid desaturases are attractive for the production of polyunsaturated fatty acids, such as ω-3 fatty acids. Although L. starkeyi is one of the most studied oleaginous yeasts, the genes encoding its fatty acid desaturases have not been reported. In this study, we identified and characterized two novel fatty acid desaturases, a ∆12 fatty acid desaturase (LsFad2) and a ∆12/∆15 bifunctional fatty acid desaturase (LsFad3), using heterologous expression in the model yeast S. cerevisiae and overexpression and deletion analyses of genes encoding these desaturases in L. starkeyi. In the L. starkeyi Δlsfad2 and LsFAD3-overexpressing S. cerevisiae cells, a large proportion of LA was converted to ALA (Tables 3 and 4). The high conversion efficiency of LA to ALA by LsFad3 (L. starkeyi Δlsfad2 cells 75%, LsFAD3 overexpression S. cerevisiae cells 82%), which was calculated as [ALA / (ALA + LA) × 100%] of cells cultured at 20 °C, suggested that LsFad3 may convert LA to ALA without completely releasing its substrate after conversion of oleic acid to LA. The same phenomenon was seen in other bifunctional Δ12/Δ15 fatty acid desaturases, including a fatty acid desaturase isolated from F. moniliforme (Damude et al. 2006). Both LsFad2 and LsFad3 contributed to the production of LA and ALA under low temperature conditions. Interestingly, although the expression levels of these fatty acid desaturase genes were not enhanced at low temperatures (Fig. 3), both LA and ALA accumulated under those conditions. Previously, Lomascolo et al. reported a low thermal stability for L. starkeyi ∆12 desaturase (Lomascolo et al. 1996). We hypothesize that the contradiction between the expression levels of fatty acid desaturase genes and the amount of polyunsaturated fatty acids is the reason for the observed thermal stability of LsFad2 and LsFad3, which would be inactivated shortly after their expression at higher temperatures. In yeast, polyunsaturated fatty acids play vital roles in stress tolerance in low temperatures (Angelis et al. 2016), alkaline pH (Yazawa et al. 2009), and other conditions. As the most studied yeast, S. cerevisiae, lacks polyunsaturated fatty acids; we think that L. starkeyi strains that have modified their fatty acid compositions are useful for analysis of the physiological roles of polyunsaturated fatty acids in eukaryotic microbes.

Our ongoing efforts focus on the biochemical analysis of LsFad2 and LsFad3, including their substrate specificity and temperature stability, and the physiological importance of polyunsaturated fatty acids in L. starkeyi. It was reported that ∆12 desaturases of L. starkeyi and Candida lipolytica exhibit desaturase activity toward oleoyl-CoA (Lomascolo et al. 1996; Horváth et al. 1991), whereas the substrates of ∆12 desaturases and ∆15 desaturases are phosphatidylcholine in plants (Miquel and Browse 1992; Ohlrogge and Browse 1995). Detailed biochemical analyses of LsFad2 and LsFad3 are necessary to understand their physiological roles. In addition, we expect that both LsFad2 and LsFad3 will be applicable to the bioproduction of polyunsaturated fatty acids, including ω-3 fatty acids.