Introduction

Production of biofuels as a clean, renewable, environment-friendly, and sustainable source of alternative energy has recently attracted substantial research interests. Biofuels derived from microorganisms, including yeasts, are gaining more attention due to several advantages in comparison to plant-derived oils. These include short life cycle of the microorganisms, reducing production time, and the fact that microbial growth can be independent of climate and season. Previous studies of lipid production in the oleaginous yeast Yarrowia lipolytica showed that its fatty acids’ profile is similar to vegetable oils, suggesting potential use as a replacement for plant oils in biofuel production (Beopoulos et al. 2011). The baker’s yeast Saccharomyces cerevisiae is a well-suited unicellular organism used in various industrial and biotechnological processes, including production of bioethanol and pharmaceutical components. In addition to its robustness and feasibility for upscaling and commercial processes, the S. cerevisiae genome and metabolic pathways are also well-documented. These properties make the model S. cerevisiae yeast suitable for the strain development and pathway engineering (Dyer et al. 2002; Nielsen and Jewett 2008; Veen and Lang 2004). Currently, S. cerevisiae is applied as a host for the production of fatty acid-derived biofuels and fuel chemicals to replace vegetable oils or animal fats which are food sources with high costs for biofuel production (Papanikolaou et al. 2002; Zhou et al. 2014).

S. cerevisiae can utilize a wide variety of substrates including fatty acids as a source of carbon (Schüller 2003). Extensive studies have demonstrated that manipulation of genes, either overexpression or deletion of genes in the fatty acid synthetic pathway (FAA1, FAA4, or ACC1), alters accumulation and composition of free fatty acids (FFAs) which are important precursors for the biosynthesis of triacyglycerol (TAG), a starting material in the transesterification reaction for the production of fatty acid methyl/ethyl esters (FAMEs/FAEE (biodiesel)), and fatty acy-CoA for production of fatty alcohols and alka(e)nes in S. cerevisiae (Runguphan and Keasling 2014; Scharnewski et al. 2008), oleaginous yeasts (Haddouche et al. 2011; Sitepu et al. 2014), or Escherichia coli (Janssen and Steinbüchel 2014; Lennen and Pfleger 2013). In addition to FFAs, S. cerevisiae accumulates storage lipids in the form of TAGs at approximately 10 % of dry cell weight or less. Overexpression of genes encoding fatty acyl synthetases (FAS1 and FAS2), acetyl-CoA carboxylase (ACC1), and diacyl-glycerol acyltransferase (DGA1), as well as the deletion of isocitrate dehydrogenases IDH1 and IDH2, in S. cerevisiae strains expressing ATP-citrate lyase (ACL1), are shown to enhance TAG levels (Chen et al. 2014; Li et al. 2014; Runguphan and Keasling 2014; Tang et al. 2013). Furthermore, disruption of the SNF2 gene, encoding the catalytic subunit of the SWI/SNF chromatin remodeling complex, strongly boosts TAG accumulation, suggesting an extensive involvement of transcriptional control over TAG and fatty acid biosynthesis (Kamisaka et al. 2006). Alternatively, to enhance fatty acid biosynthesis, blocking the pathway of fatty acid degradation has also been shown to increase FFA accumulation (Chen et al. 2014). In S. cerevisae, fatty acid breakdown is mediated by the β-oxidation pathway, responsible for breaking down fatty acids to generate acetyl-CoA (Kohlwein et al. 2013).

Zinc cluster proteins form a major class of transcription regulators in the yeast S. cerevisiae. They contain a Zn2Cys6 binuclear cluster DNA-binding motif with the consensus sequence of Cys-X2-Cys-X6-Cys-X5–12-Cys-X2-Cys-X6–8-Cys. Zinc cluster transcriptional regulators are associated with many cellular processes, such as sugar, amino acid, carbon, and nitrogen metabolism (MacPherson et al. 2006). Other zinc cluster regulators control expression of genes involved in the metabolism of non-fermentable carbon sources or in lipid metabolism (Turcotte et al. 2010). For example, Upc2 plays a primary role in regulating the expression of ergosterol biosynthetic genes and has a secondary role in anaerobic sterol uptake and in controlling expression of DAN/TIR cell wall gene, encoded mannoprotein (Abramova et al. 2001; Hartman et al. 2001). Zinc cluster transcriptional regulators Oaf1 and Pip2 regulate genes involved in β-oxidation in response to the presence of oleate (Rottensteiner et al. 1996). The Sut1 regulator activates genes, involved in sterol uptake under anaerobic conditions, and is involved in hypoxic gene expression (Bourot and Karst 1995; Ness et al. 2001). Adr1 is a Cys2His2 regulator that is also important for yeast growth on fatty acids and is involved in the regulation of gene-encoding peroxisomal proteins, including FOX2 and POT1 (Gurvitz et al. 2000). Cells lacking Adr1 show no growth on media containing fatty acids as a sole carbon source. Recently, Tog1 regulator of zinc cluster proteins was found to induce expression of β-oxidation genes in response to a shift from glucose to oleate (Thepnok et al. 2014).

The objective of this study was to characterize the involvement of a lesser-known zinc cluster protein Asg1 (also called Yil130w) in fatty acid metabolism. Phenotypes of the Δasg1 strain on fatty acids and oils as sole carbon sources and in the presence of oxidative agents were examined. Expression levels of selected genes related to fatty acid utilization and binding enrichment of Asg1 at the promoters of these genes were investigated during growth on oleate. High-performance liquid chromatography (HPLC) analysis was also employed to identify lipid classes in the Δasg1 cells grown in glucose or following the glucose-oleate shift. These analyses should allow for a better understanding of the contribution of the Asg1 regulator to the control of lipid utilizing genes as well as its effect on fatty acid accumulation.

Materials and methods

Yeast strains

The S. cerevisiae zinc cluster deletion Δasg1 strain was derived from the wild-type strain FY73 (MATa; his3-Δ200; ura3–52) which is isogenic to S288C (Winston et al. 1995). The FY73 and the Δasg1 strains, kindly provided by B. Turcotte (McGill University, Canada), were used for phenotypic, gene expression, and lipid analyses. The obtained Δasg1 strain was previously constructed by disrupting a part of the open reading frame (ORF) of ASG1 gene, encoding zinc cluster protein by using the PCR method with HIS3 as a marker for selection and confirmed via Southern blot analysis, as performed and described (Akache et al. 2001; Baudin et al. 1993). Again, the resulting Δasg1 strain was verified by PCR using primers specific for HIS3 marker with a pair of primers located downstream of the ASG1 ORF with the oligo primer 5′-TTGCAGTTTATCACCATTAT and the primer 5′TTACTCTTGGCCTCCTCTAG (located in the 5′-end of the HIS3 marker) (data not shown). The S. cerevisiae strains BY4742 (MATα; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0) and the isogenic strain Δasg1 (MATα; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0; yil130w::kanMX4) were obtained and used (Winzeler et al. 1999) while the W303 strain (MATα, leu2–3112, trp1–1, can1–100, ura3–1, ade2–1, his3-Jl,15) and the corresponding Asg1-Myc tagged strain used in ChIP analysis, were also kindly provided by B. Turcotte. The W303 strain expressing N-terminally Myc tagged-Asg1 was constructed and verified by J. Drolet as previously described (Drolet 2007; Schneider et al. 1995).

Phenotypic analysis of the yeast strains on lipid-containing media and in the presence of oxidative agents

To screen for the involvement of the zinc cluster Asg1 in the utilization of fatty acid as a sole carbon source, a phenotypic analysis of zinc cluster deletion strains Δasg1 (in FY73, BY4742, and W303 backgrounds), and the revertant ASG1 strain (FY73) was performed. For construction of the ASG1-revertant strain, the pRS316-ASG1 plasmid was constructed using oligos:

ASG1F:5′-CGCGGATCCCCGTAGGAGGAGAGTCTGGACC and ASG1R:5′-AAA GCGGCCGCTCATTCAGAGGGGTAATTTAAAG for PCR amplification of the promoter region (1000 bp upstream of the ATG codon) and the ORF of ASG1 gene with high fidelity polymerase (Thermo Fisher Scientific, NY, USA). The amplified fragment was inserted into the vector pRS316 (Sikorski and Hieter 1989) to obtain the pRS316-ASG1 which was checked for correct insertion at BamHI and NotI site using restriction digest with the enzymes BamHI and NotI (New England Biolabs, Ipswich, MA, USA). After, the wild-type FY73 and the Δasg1 strains were transformed with either the empty vector pRS316 or pRS316-ASG1 and checked for growth on SD-Ura plates, containing 0.67 % yeast nitrogen base without amino acids (Himedia Laboratories, Mumbai, India), 2 % dextrose (Himedia Laboratories, Mumbai, India), and supplement of yeast synthetic dropout medium without uracil (Sigma-Aldrich, Rehovot, Israel). After, cells were grown in yeast extract/peptone/dextrose (YPD) medium containing 1 % yeast extract, 2 % peptone, and 2 % dextrose (Himedia Laboratories, Mumbai, India) and incubated at 30 °C overnight. They were then serially diluted with distilled water to an optical density (OD600) of 0.1, 0.04, 0.0062, and 0.0015, respectively. Ten microliters of each dilution was spotted onto YP plates containing different fatty acids or oils as sole carbon sources at a final concentration of 0.125 % oleic acid, 0.125 % linoleic acid, 0.125 % palmitic acid, 0.125 % stearic acid, 10 mg/ml palm oil, or 10 mg/ml sunflower oil. Fatty acids and oils were emulsified in 0.5 % Tween 80 (LabChem, Zelienople, PA, USA).

To examine the sensitivity of yeast zinc cluster deletion strains to oxidative stress, cells were challenged with H2O2 and menadione. Mid-log phase cells (OD600 of 0.8) were exposed to menadione solubilized in DMSO, for 1 h at 30 °C and 100 rpm. Cells were serially diluted and spotted on YPD plates, followed by incubation at 30 °C for 2–3 days to monitor the growth. To compare the growth of the Δasg1 strain in FY73 and in BY4742 and W303 backgrounds, cells were again prepared as previously described prior to being spotted on plates containing either glucose or oleate as a sole source of carbon or in the presence of 4.0 or 5.0 mM H2O2. For cell viability assay, viability of the wild-type cells and the Δasg1 strains were determined, using the colony counting method on YPD agar plates following exposure to H2O2 and menadione. Yeast cell cultures were grown overnight in YPD medium at 30 °C with shaking and then diluted into fresh media to a final OD600 of 0.1. Cells were grown until the mid-log phase and then exposed to 0.4 mM menadione in DMSO, for 1 h at 30 °C with shaking. Following the exposure to menadione, cells were serially diluted and spread onto YPD agar plates and incubated at 30 °C for 2–3 days. The viability of cells was calculated by comparing the number of colonies formed from non-exposed and exposed cultures. These results were normalized using values from the wild-type strain and presented as a percentage of viable cells.

Culture conditions for lipid analysis

Yeast cell cultures were grown in YPD medium at 30 °C with shaking overnight and then diluted to an OD600 of 0.1 in 100 ml of YPD and regrown until reaching the exponential growth phase (OD600 of 0.6), then divided into half, harvested by centrifugation, and washed twice with distilled water. Cells were resuspended in YP medium containing 0.125 % oleic acid emulsified with Tween 80 and allowed to grow for an additional 3 h. To examine lipid composition in cells during the stationary phase of growth, cells were inoculated at OD600 of 0.1 in YPD medium and incubated for 96 h, harvested, and washed at least twice with deionized water prior to being used for analysis of lipid composition.

Lipid extraction

Harvested cells from 50 ml of culture were mixed with 10 ml of methanol and disrupted with glass beads by vortexing. Chloroform was then added to the cell suspension (chloroform/methanol 2:1 (v/v)) and stirred for 1 h. The extract was filtered, mixed with 10 ml of 0.345 % MgCl2 and centrifuged. Afterwards, the upper layer was removed by aspiration. Ten milliliters of 2 M KCl/methanol (4:1 (v/v)) was then added to the organic phase. Samples were again centrifuged, and the upper aqueous layer was again aspirated. The organic phase was washed twice with 10 ml of chloroform/methanol/water (3:48:47 (v/v)). The solvent was evaporated in a rotary evaporator at 55 °C and 200 mbar. The lipid film was weighed and dissolved in toluene.

HPLC analysis

HPLC was performed to analyze the lipid composition from yeast cells. Extracted lipids that were dissolved in toluene were injected into a HPLC system which was equipped with a Sedex 75 Evaporative Light Scattering Detector (ELSD; Sedere, Alfortville, France). Detector temperature was set at 30 °C, and N2 gas pressure was set at 2 bar. Data was collected and processed by CSW32 HPLC software (DataApex Ltd., Prague, Czech Republic). Samples were analyzed in a 100 Å Phenogel column (300 mM × 7.8 mM ID, 5 μm) (Phenomenex, Torrance, CA, USA). The column and injector were put in an oven set at 65 °C. The mobile phase comprised of 100 % toluene and 0.25 % acetic acid. Flow rate for the mobile phase was 1.0 ml/min.

Gene induction and RT-qPCR

Yeast cells were cultured in 5 ml of YPD medium at 30 °C with shaking overnight. Cell cultures were diluted to an OD600 of 0.1 in YPD medium and grown until reaching the OD600 of 0.6. Cultures were then divided in half, and cells were grown either in 2 % glucose or 0.125 % oleic acid induction conditions for an additional 3 h. Cells were washed with diethylpyrocarbonate (DEPC) water, and pellets were used for RNA extraction, purification, and reverse transcription quantitative PCR (RT-qPCR) analysis. The wild-type and ∆asg1 strains were grown as described above. Total RNA was isolated by the hot acid phenol method and purified using a Rneasy Mini Kit (Qiagen, Hilden, Germany), and cDNA synthesis employed a Super Script™ III first-strand synthesis kit (Life Technologies, Carlsbad CA, USA). RT-qPCR was performed using an Mx3005P QPCR system with MxPro QPCR software for analysis (Agilent Technologies, Santa Clara CA, USA). The reaction mixture contained Brilliant II SYBR Green QPCR Mix (Kapa Biosystems, Wilmington, MA, USA). DNA sequences of the oligonucleotides used for RT-qPCR analysis are given in Table 1. The relative quantification of each transcript was calculated by the 2-ΔΔCt method (Livak and Schmittgen 2001) and normalized using the ACT1 (actin) gene.

Table 1 DNA sequences of primers used in RT-qPCR and ChIP analysis
Full size table

Chromatin immunoprecipitation assays

Yeast cells from wild-type W303 and Myc-Asg1 tagged strains were grown as previously described for gene induction. The chromatin immunoprecipitation (ChIP) assays were performed as described (Larochelle et al. 2006; Soontorngun et al. 2007) with minor modifications. Equal amounts of whole-cell extracts from each sample were incubated with anti-Myc antibody (12CA5) (Roche, Mannheim, Germany) coupled with magnetic beads (Dynabeads, Oslo, Norway). Following immunoprecipitation and cross-link reversal, DNA was purified and used for QPCR analysis with an ABI 7500 Real-time PCR system (Applied Biosystems, Foster City, CA, USA) with the gene-specific oligos listed in Table 1. The binding enrichments were calculated using the 2-ΔΔCt method (Livak and Schmittgen 2001), and the values were normalized with the non-tagged strain.

Results

Impaired growth on fatty acids and increased sensitivity to oxidative stress of the Δasg1 strain

Akache and co-workers previously demonstrated that some zinc cluster deletion strains, including a strain with deletion in the YIL130W (ASG1) gene, display impaired growth on glycerol and lactate (Akache et al. 2001). Here, we further tested the ability of the zinc cluster ∆asg1 strain (the FY73 genetic background) to utilize fatty acids and oils including oleic, linoleic, palmitic, and steric acids as well as palm and sunflower oils as a sole source of carbon. Our phenotypic analysis revealed that the ∆asg1 strain grows poorly when these compounds are used as a sole carbon source (Fig. 1a), suggesting a role of Asg1 in non-fermentable carbon utilization of fatty acids. To test whether the observed defective phenotypes were specific to the S. cerevisiae FY73 background, we also examined growth of the Δasg1 strain in BY4742 and W303 backgrounds by growing cells on oleate and linoleate as a sole source of carbon and in the presence of H2O2, the oxidant commonly produced during the β-oxidation of fatty acids (Kunau and Hartig 1992) The results showed that the growth of the Δasg1 (BY4742 and W303 genetic backgrounds) strains on oleate and linoleate is normal unlike what is observed for the Δasg1 (FY73 background) strain (Fig. 1b). Since oxidation of fatty acids is known to produce H2O2 , it could result in cellular oxidative stress and reduced growth rate and cell viability (Hatem et al. 2014; Kurita 2003; Minard and McAlister-Henn 1999). The effect of ASG1 deletion on sensitivity to oxidative stress was also tested with oxidative agents such as H2O2 and menadione. Menadione is commonly used as a source of reactive oxygen species (ROS). This quinone compound when reduced to a semiquinone produces various ROS such as superoxide ion, hydrogen peroxide and hydroxyl radical (Stohs and Bagchi 1995). Here, the results demonstrated that the ∆asg1 strain shows impaired growth in the presence of 3.0, 4.0, or 5.0 mM H2O2 and upon exposure to menadione at a final concentration of 0.6, 0.8, and 1 mM (Fig. 1a). Interestingly, the sensitivity to H2O2 was found to be similarly impaired in all three genetic backgrounds tested. Overall, the results suggest specific growth impair on fatty acids and common hypersensitive phenotype of the Δasg1 strains among these S. cerevisiae genetic backgrounds tested (Fig. 1b).

Fig. 1
figure 1

Phenotypes of the ∆asg1 strain (FY73 background) during growth on YP-plates containing a either glucose, oleate, linoleate, palmitate, sterate or palm, and sunflower oils as a sole source of carbon or in the presence of oxidizing agents H2O2 or menadione at various concentrations as shown. b Growth of the ∆asg1 strains in different genetic backgrounds (FY73, BY4742, and W303) on glucose or oleate as a sole source of carbon and in the presence of 4.0 or 5.0 mM of H2O2. c Growth of the ASG1 reverant strain (FY73 background) on glucose or oleate as a sole source of carbon and in the presence of 4.0 or 5.0 mM of H2O2. Representative pictures of cells from three independent experiments are shown

Full size image

To confirm that Asg1 in the FY73 background contributes to the defective phenotypes on fatty acids and during H2O2 exposure, a centromeric vector pPRS316-ASG1 which contains the original ASG1 promoter and the open reading frame of ASG1 gene was constructed and transformed into the Δasg1 (FY73) strain for the phenotype analysis via spot tests. The results showed that growth of the Δasg1 (FY73) is rescued on oleate or linoleate as a sole carbon source and in the presence of oxidizing agent H2O2 (Fig. 1c), supporting the contribution of Asg1 in the utilization of these two fatty acids and increased tolerance to H2O2 stress. To confirm results of the spot test, cell viability assay was also performed. It was observed that the ∆asg1 strain displays increased sensitivity to 2.0 or 3.0 H2O2 and 0.4 mM menadione stress with 32.9, 12.6, and 18.4 % viability as compared to the wild-type FY73 strain, respectively (data not shown). Both spot and viability assays revealed an increased sensitivity of the ∆asg1 strain to the oxidative agents (Figs. 1a and data not shown), suggesting an important role of Asg1 during oxidative metabolism and in cellular defense against harmful oxidative agents.

Asg1 regulates expression of lipid utilization genes

The β-oxidation, the glyoxylate shunt, and the gluconeogenesis are important processes involved in lipid degradation and generation of important metabolites during non-fermentative metabolism (Schüller 2003; Turcotte et al. 2010). Since Asg1 encodes a putative transcriptional regulator in the zinc cluster protein family, we hypothesized that Asg1 may play a role in controlling expression of genes involved in fatty acid breakdown, β-oxidation, or other downstream pathways that are required for the utilization of fatty acids. To test this potential role of Asg1, RT-qPCR analysis was performed to determine the expression of key oleate utilizing genes that encode peroxisomal transporter (PXA1), triacylglycerol lipase (TGL3), β-oxidation (POX1, FOX2 and POT1), glyoxylate cycle (ICL1), or gluconeogenic (PCK1) enzymes. The expression of these genes was examined in the Δasg1 strain in comparison to the wild-type FY73 strain, following the glucose-oleate shift. Expression levels of these genes were enhanced in response to the oleate shift, in agreement with previous reports (Karpichev and Small 1998; Karpichev et al. 1997). Interestingly, deletion of ASG1 reduced the expression level of POX1 and POT1 by approximately 2.6-fold, 3.7-fold for MDH2, and 5.2-, 4.7-, and 6.7-fold for TGL3, PXA1, and FOX2, respectively (Fig. 2). Likewise, the expression of ICL1 and PCK1 was almost absent in the Δasg1 strain compared to the wild-type FY73 strain (Fig. 2). The transcript levels were also measured to include the known regulatory genes CAT8, ADR1, OAF1, and PIP2 in the Δasg1 strain. However, the results showed that following oleate induction, deletion of ASG1 gene does not alter the expression levels of these regulatory genes with the relative expression levels (Δasg1/WT) of 0.7-, 1.1-, 0.9-, and 1.3-fold for CAT8, ADR1, OAF1, and PIP2, respectively, as compared to the wild-type strain (data not shown). This suggests that the Asg1 regulator may directly activate genes, involved in fatty acid utilization and breakdown as well as those in the glyoxylate cycle and gluconeogenesis.

Fig. 2
figure 2

Expression of Asg1 target genes during the oleate induction. mRNA levels were measured using RT-qPCR analysis. The mRNA fold enrichments were calculated, using the comparative ΔΔCt method to determine the relative mRNA levels of Asg1 putative targets genes in the Δasg1 versus the wild-type strain. ACT1 was used to normalize mRNA levels in each sample (Livak and Schmittgen 2001)

Full size image

Since the Asg1 was required to activate some key genes involved in the β-oxidation and the gluconeogenesis, we further investigated whether the Asg1 regulator is directly bound to the promoters of these genes. Standard ChIP assays were performed using a Myc-tagged Asg1 under conditions in which cells are shifted from glucose- to oleate-containing medium. A substantial binding enrichment of Asg1 was found at promoters of the genes necessary for the utilization of non-fermentable carbon sources. Binding enrichment of the zinc cluster transcriptional regulator Asg1 was observed at the promoters of POX1, FOX2, and POT1 genes, encoding enzymes in β-oxidation, PCK1 and MDH2 genes, encoding key enzymes phosphoenolpyruvate carboxykinase and malate dehydrogenase, respectively, of the gluconeogenesis (Fig. 3). Binding of Asg1 was observed at positions upstream of the ATG start codons on promoters of POX1 (between −112 and −469 bp), FOX2 (between −102 and −430 bp), POT1 (between (−23 and −346 bp), PCK1 (between −450 and −150 bp), and MDH2 (between −500 and −300 bp) relative to the ATG codon. In contrast, there was no binding enrichment on ACT1 gene (Fig. 3). Thus, the Asg1 activator was enriched at promoters of genes in central pathways for the utilization of non-fermentable compounds, and not exclusively genes for oleate utilization.

Fig. 3
figure 3

The binding enrichment of Asg1 on the promoters of genes involved in oxidative metabolism as identified using ChIP analysis of the Myc-Asg1 during the glucose-oleate shift. Asg1 is enriched on the promoters of PCK1, FOX2, POT1, POX1, and MDH2 genes. The negative control ACT1 was not enriched for Asg1 binding and used for normalization

Full size image

Increased lipid content and percentage abundance of free fatty acids in the ∆asg1 strain

Yeast cells are known to accumulate neutral lipids mainly as TAG in the stationary phase of growth. TAG is an essential lipid for maintaining energy homeostasis and is a precursor or a buliding block for the synthesis of structural lipids and other important components of membranes. It is also used to support rapid growth and division of cells (Czabany et al. 2007). During the exponential phase of growth, when cells divide rapidly, TAGs are continuously consumed. However, in the stationary phase, cell division decreases, and TAG is stored (Sandager et al. 2002). Since our RT-qPCR analysis indicated that Asg1 is involved in the regulation of TGL3, a gene for triacylglycerol lipase (Fig. 2), we questioned whether the Δasg1 strain’s ability to break down TAG was impaired. Since defective breakdown of TAG would alter the lipid content and composition, the neutral lipid content in wild-type and ∆asg1 cells grown in glucose-containing medium during the stationary phase of growth (96 h) were determined. Lipid content in the wild-type FY73 strain was found to be approximately 0.011 g/g of lipid per cell dry weight whereas the Δasg1 strain contained 0.043 g/g of lipid per cell dry weight after 96 h of cultivation in glucose (Table 2). This was approximately a 4-fold increase in lipid content (Table 2), confirming that the Δasg1 strain accumulated more lipid than in the wild-type strain. Our results suggested that the inability to breakdown TAGs is one contributing factor for the elevated total lipid content observed for the Δasg1 strain.

Table 2 Lipid content during the stationary phase (96 h) of cell growth in glucose and the profiles of neutral lipids (TAG + SE) and free fatty acid (FFA) percentage abundance of the wild-type and Δasg1 strains (FY73 background)
Full size table

The lipid profile of the ∆asg1 strain was also analyzed using HPLC analysis. The TAG and SE were found as major neutral lipids. As expected, results indicated that the percentage abundance of TAG and SE of the wild-type FY73 glucose-grown cells (the exponential growth phase) is lower than when cells are grown to the stationary phase (Table 2). In contrast, a similar trend was found for the percentage abundance of TAG and SE of the glucose-grown Δasg1 cells in both phases of growth (Table 2). A similar trend was found for the percentage abundance of free fatty acids (FFA) (Table 2), indicating that the Δasg1 strain is unable to break down triacylglyceride. For the wild-type FY73 strain, during the exponential phase of growth, the FFA percentage abundance was found to be similar for glucose-grown and glucose-oleate shifted cells (Table 2). Interestingly, FFA percentage abundance of the ∆asg1 strain following the glucose-oleate shift was found to be approximately 2-fold higher than that of wild-type cells, cultured in glucose-containing medium (Table 2), supporting that yeast cells that are unable to utilize fatty acids as a sole carbon source because they display compromised β-oxidation activity and therefore accumulate FFA.

Discussion

Here, the transcriptional role of the zinc cluster transcription factor Asg1 in fatty acid utilization was shown by RT-qPCR and ChIP analyses which demonstrates that Asg1 regulator induces the expression of several oleate utilizing genes namely the PXA1, POX1, FOX2, POT1, ICL,1 and PCK1 genes and binds to POX1, FOX2, POT1, MDH2, and PCK1 promoters in response to oleate induction, respectively (Figs. 2 and 3). A model for Asg1 function in controlling oleate utilization is provided in Fig. 4. Many Asg1 target genes encode proteins that are required for effective degradation of fatty acids, and some of which are localized in the peroxisomes. First, Asg1 regulates expression of the PXA1 gene (Fig. 2) encoding a peroxisomal transporter that facilitates the passage of fatty acids and metabolites across the peroxisomal membrane, allowing for cross-talk between this organelle and other subcellular compartments. Deletion of PXA1 and PXA2 genes results in impaired fatty acid oxidation and poor growth on the medium, containing long-chain fatty acids as a sole carbon source, even though the genes for the β-oxidation pathway remain intact (Beopoulos et al. 2011). In addition, activation of medium-chain fatty acids also occurs in the peroxisome through the action of acyl-CoA synthetase. Acyl-CoA substrates are subsequently oxidized to trans-2-enoyl-CoA by the peroxisomal acyl-CoA oxidase Pox1 (also called Fox1) whose expression is also controlled by the Asg1 regulator (Figs. 2 and 3). A by-product of this reaction is excess H2O2 which can damage cellular macromolecules through toxic ROS. Whenever the concentration of ROS exceeds the antioxidant buffering capacity, cells experience oxidative stress, leading to cell death (Costa and Moradas-Ferreira 2001). Antioxidant systems thereby play important roles in neutralizing H2O2 and ROS and in repairing damaged macromolecules. Importantly, it is conserved in all genetic backgrounds that Asg1 is involved in stress response to oxidizing agent such as H2O2, an interesting role for future detail examination. Asg1 may act to coordinate cellular responses to tolerate oxidative stress, a common phenomenon during non-fermentative metabolism and oxidation of fatty acids.

Fig. 4
figure 4

Role of the zinc cluster transcriptional regulator Asg1 in S. cerevisiae. Following the glucose-oleate shift, the regulator Asg1 acts to activate the expression of genes involved in the pathways of β-oxidation (PXA1, POX1, FOX2, and POT1), gluconeogenesis (PCK1 and MDH2), glyoxylate cycle (ICL1) and triacylglycerol lipase (TGL3). It is also required for oxidative stress response against some oxidizing agents. Deletion of ASG1 in the FY73 background results in increased lipid accumulation, namely FFAs (during exponential phase of growth in oleate-containing medium) and TAGs (during stationary phase of growth in glucose-containing medium). Both FFAs and TAGs are major sources for fatty acid- and triacylglycerol-derived biofuels. The relative expression levels and binding enrichment of Asg1 target genes were provided in the bracelets, respectively. ND no data

Full size image

Continuing to the β-oxidation process, trans-2-enoyl-CoA is subsequently processed to 3-ketocyl-CoA by Pot1 peroxisomal oxoacyl thiolase (Erdmann 1994), another target of Asg1 (Figs. 2 and 3), to yield acetyl-CoA and C2-shortened acyl-CoA which can then enter either the glyoxylate or the tricarboxylic acid (TCA) cycles to generate ATPs and metabolic intermediates. Asg1 is also involved in activation of the glyoxylate ICL1 gene (Figs. 2 and 3) for conversion of glyoxylate and acetyl-CoA to malate (Fernandez et al. 1992) and at least two key gluconeogenic genes MDH2 and PCK1 (Figs. 2 and 3) for generation of glucose-6-phospate, important precursors for several cellular pathways (Minard and McAlister-Henn 1991; Valdes-Hevia et al. 1989). Regarding triacylglerol degradation, the enzyme triacylglycerol lipases (Tgl3–5) act to hydrolyze TAGs to provide diacyl-glycerol and free fatty acids (Athenstaedt and Daum 2003) and are essential for membrane lipid biosynthesis, as well as for energy production (Beopoulos et al. 2011). Cells lacking these TGL genes are unable to utilize TAGs, and lipids are accumulated inside the cells as lipid bodies, composed predominantly of neutral lipid TAGs (Dulermo et al. 2013). Expression of TGL3 as well as other fatty acid-utilizing genes partially depends on the Asg1 regulator (Fig. 2) which also explains the altered level of TAG and fatty acid contents observed in the Δasg1 strain (Table 2). Interestingly, Asg1 is enriched in promoter regions that contain cis-acting oleate-responsive elements (OREs), important for the induction of β-oxidation as well as carbon source responsive elements (CSREs) of the key gluconeogenic gene PCK1, suggesting a direct role in the up-regulation of their expression (Figs. 2 and 3, (Gurvitz and Rottensteiner 2006; Hiltunen et al. 2003)). The increased FFA percentage abundance of the ∆asg1 strain also corelates well with the decreased expression of fatty acid β-oxidation pathway which prevents the conversion of FFAs to acetyl-CoA (Table 2 and Fig. 2).

Our data shows that defective transcriptional control of β-oxidation genes is one contributing factor for observed poor growth phenotypes of the Δasg1 strain on fatty acids and oils. The deletion of ASG1 results in poor growth on several fatty acids (in FY73 background) and renders the cells sensitive to oxidative agents (Fig. 1). Since the impaired utilization on oleate and linoleate of the Δasg1 is specific to the FY73 background, it implies that the observed defective growth is partly contributed by deletion of ASG1 gene while the other causes also possibly contribute. Deletion in the ASG1 gene is tolerated differently by laboratory strains of S. cerevisiae, being viable in the BY4742 and W303 backgrounds on oleate and other non-fermentable carbon sources (Fig. 1 and data not shown) but dead in the FY73 (S288C-derived) background, suggested that the lethality of ASG1 deletions in the FY73 background may be suppressed by an allele specific to the BY4742 or W303 background. A similar example of the observed phenotypic variation has been elegantly demonstrated, using the synthetic genetic array analysis and mapping approach to identify suppressor of the lethality linked to the Cbk1 kinase signaling pathway (Jorgensen et al. 2002). In fact, several genes whose products buffer one another or impose on the same essential pathway have been uncovered in S. cerevisiae (Tong and Boone 2006). Second, impaired peroxisomal proliferation and morphology may also be possible as shown by S. cerevisiae the Δpip2, the Δtog1, and the Δadr1 strains with deletion in genes encoding activators of β-oxidation and peroxisomal genes. The Δpip2 mutant lacking the peroxisomal activator Pip2 displays a reduction of peroxisomal content with many lipid droplets in the cells and low expression level of PMP27 that is required for incorporating peroxisomal membrane proteins (Rottensteiner et al. 1996). Similarly, the Δtog1 strain in the FY73 background shows reduced numbers of peroxisomes when cells are shifted from glucose to oleate (Thepnok et al. 2014). Lower expression of PEX11 encoding peroxin, for peroxisome proliferation, has also been observed for the Δadr1 strain (Gurvitz et al. 2001). Also, the mitochondrial stability may differ for the each yeast backgrounds, and this could influence growth and phenotypic outcomes. Thirdly, the poor growth of the ∆asg1 strain (FY73 background) on some fatty acids and hypersensitivity to oxidizing agents of the Δasg1 strain in the FY73 background may be partly due to the toxicity from the excessive accumulation of FFAs. It has been reported that high FFAs promote the formation of harmful reactive intermediates (ROS) and ceramide (Listenberger et al. 2003; Sorger and Daum 2003) or reduced expression levels or enzymatic activity of oxiding enzymes. Lastly, Asg1 appears to have a redundant function in fatty acid or non-fermentable carbon utilization with other regulators in the regulatory network of non-fermentable carbon metabolism in S. cerevisiae, including Oaf1, Pip2, Znf1, Tog1, Cat8, Sip4, Adr1, and Rds2 (Gurvitz and Rottensteiner 2006; Soontorngun et al. 2012; Tangsombatvichit et al. 2015; Thepnok et al. 2014; Turcotte et al. 2010) as shown by some overlapping target genes (Figs. 2 and 3) which explains for functional redundancy. Nevertheless, a phenotypic screening of Candida albicans mutants reveals that C. albicans ASG1 is also necessary for growth on non-fermentative carbon sources (Coste et al. 2008).

Furthermore, the inability of the Δasg1 strain (FY73 background) to break down fatty acids provides some important advantages, suggesting a new approach to increase the content of FFAs in addition to disruption of individual genes of β-oxidation pathway which has been previously demonstrated (Li et al. 2014). Importantly, the Δasg1 strain grows normally in glucose (Fig. 1) and contains approximately four times higher lipid yield and double FFA percentage abundance compared to the wild-type FY73 strain under glucose conditions (Table 2). Thus, this strain in the FY73 background is more suitable for glucose-based lipid production. Despite the poorer growth of the Δasg1 strain in oleate-containing plates (Fig. 1), this strain accumulates extracellular FFAs (Table 2). This may be beneficial in bioremediation for treating and recycling industrial wastewater, containing high FFA content. The yeast with FFA/lipid accumulation could also be harnessed for microbial lipid extraction (Haddouche et al. 2011; Runguphan and Keasling 2014; Scharnewski et al. 2008; Sitepu et al. 2014). In support, previous research on some mutants of S. cerevisiae and Y. lipolytica also demonstrate that yeast can accumulate a large amount of lipids in cells which are disrupted in POX1-6 genes, encoding peroxisomal acyl-CoA oxidases. Resulting Δpox1-6 deletion strain lacks the ability to utilize fatty acids as a sole carbon source which leads to increased cellular lipid accumulation (Beopoulos et al. 2008). Inactivation of TGL3 and/or TGL4 has been shown to double the amount of lipids and increase the accumulation of TAGs in Y. lipolytica (Dulermo et al. 2013). Some studies have successfully applied the Yarrowia yeast for the production of microbial lipids from industrial waste, promising a renewable and low-cost source of starting raw materials (Azocar et al. 2010; Cheirsilp and Louhasakul 2013; Koch et al. 2014; Louhasakul and Cheirsilp 2013). Increased understanding of Asg1’s function in lipid metabolism of the model yeast S. cerevisiae may offer new strategies for metabolic engineering and construction of oleaginous yeast strains lacking Asg1 homologs for fatty acid- or triacylglycerol-derived biofuel production. Despite the complexity of lipid metabolism and complicated cellular processes, genetic manipulation in S. cerevisae merits further investigation.