Skip to main content
Download PDF

Metabolic engineering of Rhodotorula toruloides IFO0880 improves C16 and C18 fatty alcohol production from synthetic media

Microbial Cell Factories volume 21, Article number: 26 (2022) Cite this article

Abstract

Background

The oleaginous, carotenogenic yeast Rhodotorula toruloides has been increasingly explored as a platform organism for the production of terpenoids and fatty acid derivatives. Fatty alcohols, a fatty acid derivative widely used in the production of detergents and surfactants, can be produced microbially with the expression of a heterologous fatty acyl-CoA reductase. Due to its high lipid production, R. toruloides has high potential for fatty alcohol production, and in this study several metabolic engineering approaches were investigated to improve the titer of this product.

Results

Fatty acyl-CoA reductase from Marinobacter aqueolei was co-expressed with SpCas9 in R. toruloides IFO0880 and a panel of gene overexpressions and Cas9-mediated gene deletions were explored to increase the fatty alcohol production. Two overexpression targets (ACL1 and ACC1, improving cytosolic acetyl-CoA and malonyl-CoA production, respectively) and two deletion targets (the acyltransferases DGA1 and LRO1) resulted in significant (1.8 to 4.4-fold) increases to the fatty alcohol titer in culture tubes. Combinatorial exploration of these modifications in bioreactor fermentation culminated in a 3.7 g/L fatty alcohol titer in the LRO1Δ mutant. As LRO1 deletion was not found to be beneficial for fatty alcohol production in other yeasts, a lipidomic comparison of the DGA1 and LRO1 knockout mutants was performed, finding that DGA1 is the primary acyltransferase responsible for triacylglyceride production in R. toruloides, while LRO1 disruption simultaneously improved fatty alcohol production, increased diacylglyceride and triacylglyceride production, and increased glucose consumption.

Conclusions

The fatty alcohol titer of fatty acyl-CoA reductase-expressing R. toruloides was significantly improved through the deletion of LRO1, or the deletion of DGA1 combined with overexpression of ACC1 and ACL1. Disruption of LRO1 surprisingly increased both lipid and fatty alcohol production, creating a possible avenue for future study of the lipid metabolism of this yeast.

Background

Long-chain alcohols are widely used as constituents of cosmetics, surfactants, and detergents [1]. These compounds have traditionally been produced chemically, through hydrolysis and hydrogenation of plant or animal fats, or directly synthesized by oligomerization of ethylene followed by oxidation of the resulting olefin [2]. In the United States, synthesis from ethylene is the predominant method [2]; however, this process requires the use of non-renewable petroleum feedstocks as well as the highly energy-intensive production of ethylene by steam cracking at 750–950 °C [3]. The use of microbes as cellular factories for chemical conversion has been increasingly investigated for many chemical products including fuels [4], pharmaceuticals [5], and various other value-added bioproducts [6] from renewable feedstocks at approximately ambient temperature and pressure, creating the potential for significant energy savings and reduced carbon emissions compared to traditional chemical synthesis approaches.

Fatty alcohols are natively produced in many organisms as a component of natural waxes by reduction of fatty acids or fatty acyl-CoA by carboxylic acid reductases (CARs) or fatty acyl-CoA reductases (FARs), respectively [1]. These fatty alcohol-producing genes have been expressed in a variety of heterologous microbial hosts including Escherichia coli [7], Synechocystis species [8], Saccharomyces cerevisiae, Yarrowia lipolytica [9], and Rhodotorula toruloides [10], enabling fatty alcohol production by microbial fermentation. Strikingly, fatty alcohol titers and yields remain far below those obtained for related compounds such as lipids and fatty acids [11, 12]. Significant metabolic rewiring of genetically tractable organisms such as E. coli and S. cerevisiae has led to noteworthy increases in the fatty alcohol titers obtained, with E. coli producing 6.3 g/L with the deletion of thioesterase gene tesCB, lactate dehydrogenase gene ldhA, acetate kinase gene ackA, and phosphate acetyltransferase gene pta [13]. S. cerevisiae produced 6.0 g/L of fatty alcohols using FAR from Mus musculus after 9 distinct genome edits were combined including targets for deletion and overexpression, and replacing the negative regulator of the GAL1 promoter used for FAR expression with a positive regulator [14]. Oleaginous yeasts such as Y. lipolytica and R. toruloides are promising candidate hosts for fatty alcohol bioproduction owing to their high levels of acyl-CoA production, which is natively used to produce lipids. However, metabolic engineering efforts in these less domesticated yeasts are less sophisticated, with Y. lipolytica producing 5.8 g/L of fatty alcohols using FAR from Marinobacter aquaeolei (MaFAR) after overexpression of DGA1, mutagenesis of the Mga2 regulator, and expression of 2 copies of FAR [9]. Although production of 8 g/L fatty alcohols has been reported from growth of R. toruloides on YP-sucrose media in a fed-batch fermentation also using MaFAR [10], no metabolic engineering efforts towards production of fatty alcohols have been reported to date, likely due to the lack of advanced gene editing tools compared to E. coli and S. cerevisiae. Nevertheless, the impressive results achieved so far in oleaginous yeasts demonstrate their potential as production hosts for fatty acid derivatives. With additional metabolic engineering work, the fatty alcohol titers produced in these yeast species can doubtlessly be further increased.

Recently, the genetic engineering toolkit for R. toruloides has been improved significantly with development of CRISPR gene editing [15,16,17] and RNA interference [18], and characterization of many promoters [19,20,21]. In this study, we applied the CRISPR/Cas9 gene editing system we previously developed in combination with overexpression of native and heterologous genes to create and explore a broad selection of fatty alcohol-producing upregulation and knockout mutants in R. toruloides IFO0880. We observed that improving the supply of precursor metabolites, blocking the formation of TAGs, and disrupting genes potentially involved in fatty alcohol degradation all improved the fatty alcohol titer of the producing strain. Beneficial modifications were then explored combinatorially at the culture tube and bioreactor scale, and lipidomic analysis of engineered strains shed light on the differing roles of the acyltransferases DGA1 and LRO1 in the lipid metabolism of R. toruloides.

Results

Selection of promoter and FAR gene for fatty alcohol production

Various FAR genes have been tested for production of fatty alcohols in different yeast species. D’Espaux and coworkers compared the previously used Marinobacter aqueaolei VT8, Tyto alba, and Mus musculus FARs in S. cerevisiae, finding that MmFAR1 yielded the highest titer [14], while in Yarrowia lipolytica, Cordova and coworkers compared FAR genes from Apis mellifera, Homo sapiens, Arabidopsis thaliana, and M. aqueaolei, finding that FAR from M. aqueaolei (maqu_2220 or MaFAR) showed the best performance [9]. In R. toruloides, FARs from M. aqueaolei, T. alba, A. thaliana, A. mellifera, G. gallus, and A. domesticus were recently compared, with MaFAR also giving the highest titer in this species [22]. Owing to its high activity in both R. toruloides and Y. lipolytica, MaFAR was selected for use in this study. It is possible this enzyme functions particularly well in oleaginous yeasts as its preferred substrates are the highly plentiful C16 and C18 acyl-CoA [23].

The GAPDH, or TDH3, promoter is known to be a strong promoter in the model yeast S. cerevisiae and the native version has frequently been used for heterologous gene expression in R. toruloides [10, 24, 25]. While a recent study of constitutive promoters in R. toruloides identified various promoters stronger than pGAPDH such as the transcription elongation factor promoter (pTEF1) and the adenine nucleotide transporter promoter (pANT1) in a GFP expression assay, when pANT1 and pGAPDH were compared for fatty alcohol production in another recent study in R. toruloides, pGAPDH-driven expression resulted in a four-fold greater titer of 50 mg/L [19, 22]. 1-hexadecanol and 1-octadecenol constituted approximately 25% each of the total fatty alcohols, while the remaining 50% were 1-octadecanol. We compared pTEF1- and pGAPDH-driven MaFAR expression and similarly found pGAPDH to give a higher titer of fatty alcohols (Fig. 1). Therefore, pGAPDH was used to drive MaFAR expression in all subsequent experiments.

Fig. 1
figure 1

Fatty alcohol titers resulting from pTEF1- and pGAPDH-driven MaFAR expression in R. toruloides IFO0880. Error bars represent ± standard deviation of biological duplicates. Presence of an asterisk indicates statistically significant difference in titer (p < 0.05, student t-test)

Full size image

Optimization of CRISPR/Cas9 gene editing in R. toruloides IFO0880

The creation of knockout strains has long been an essential step for the metabolic engineering of microbial organisms, typically with the goal of preventing carbon flux from traveling through undesirable pathways, increasing the flux available for the intended product [1]. Homologous recombination (HR)-based gene disruption was previously used in R. toruloides to delete PEX10 with the goal of improving lipid production [12]. Due to the very low efficiency of HR in the wild-type R. toruloides, the nonhomologous end-joining (NHEJ)-associated gene KU70 was first disrupted. Subsequently an antibiotic marker with 1 kb homology arms targeting the PEX10 locus was transformed and PEX10 was disrupted with 27% efficiency.

CRISPR/Cas9-mediated gene editing is advantageous compared to HR-based strategies in that only a 20 bp guide RNA (gRNA) must be cloned rather than 1 kb homology arms, and this editing can be performed in a wild-type R. toruloides strain rather than a KU70-deficient strain, which is known to suffer from much lower transformation efficiency [12]. We previously reported CRISPR/Cas9 gene editing in the R. toruloides strain NP11 with > 95% efficiency in the deletion of the carotenogenic reporter gene CAR2 [15], with pPGK1-driven SpCas9 expression and gRNA expression driven by a fusion 5S rRNA-tRNAArg promoter. We attempted to directly transfer this system into R. toruloides IFO0880, which has become the better characterized strain due to various tool development, omics and modeling studies [19, 26,27,28], but found there was no gene editing activity. The PGK1 promoter elements of the strains NP11 and IFO0880 are 92% similar and the two strains overall share ~ 95% genetic similarity [25]. However, our results suggest the strains are dissimilar enough that promoter elements are not always interchangeable between them.

CRISPR/Cas9-mediated gene editing was also previously reported in the R. toruloides strain IFO0880, using genetic elements from this strain, with pGAPDH-driven SpCas9 expression and tRNATyr-driven gRNA expression, albeit with a lower efficiency of 46% in the deletion of CAR2 [16]. Reasoning that the problem with our design was related to the use of promoter elements from NP11, we next attempted to improve the efficiency of our system by testing a variety of medium (pGAPDH, p27) and strong (pTEF1, pANT1, and p17, driving expression of an unknown gene [19]) promoters driving SpCas9 expression and either tRNATyr or 5S rRNA-tRNATyr-driven gRNA expression in the knockout of CAR2. We first randomly integrated the Cas9 expression cassette to IFO0880, then selected two colonies from each Cas9 transformation to receive the gRNA expression cassette subsequently. In contrast to FAR expression, Cas9 gene editing activity benefitted substantially from replacing pGAPDH with stronger promoters such as pTEF1, pANT1, and p17, while the medium strength p27 resulted in similar efficiency as pGAPDH (Table 1). In most cases both gRNA expression systems performed similarly. The combination of pANT1 and 5S-tRNATyr showed 100% editing efficiency in both colonies screened and was chosen to drive Cas9 and gRNA expression, respectively, for subsequent metabolic engineering efforts. The sequences for the codon-optimized SpCas9, MaFAR, and all other synthetic genes used in this study are shown in Additional file 1: Table S1.

Table 1 CAR2 disruption efficiency obtained from combinatorial optimization of Cas9 and gRNA promoters for high-efficiency gene disruption, screening 2 random Cas9 integrants per Cas9 promoter tested
Full size table

Metabolic engineering to improve fatty alcohol production in R. toruloides

To create a platform strain for our engineering efforts, a combined pGAPDH-MaFAR and pANT1-SpCas9 expression cassette was randomly integrated to the R. toruloides IFO0880 genome. Random integration mutants were screened for FAR activity, and Cas9 activity in the deletion of CAR2 as reported previously [14, 15], and a clone (subsequently referred to as 880CF) was isolated with 99% + efficiency for CAR2 deletion upon transformation of a CAR2-targeting gRNA and producing a fatty alcohol titer of 50 mg/L in glass culture tubes with a 10% dodecane overlay. Subsequent engineering was performed by randomly integrating either a gRNA expression cassette targeting the beginning of an ORF for NHEJ-mediated gene disruption, or a p17-driven overexpression cassette to 880CF. G418 was used as the selection marker for the combined Cas9 and MaFAR integration, while hygromycin and nourseothricin were used for gRNA and overexpression cassette integrations, respectively. A schematic overview of the strain engineering process used in this study is provided in Fig. 2. To mitigate the issue of variable expression resulting from random integration, three colonies were screened for each overexpression target, the fatty alcohol titers were measured and the clone with the highest titer (corresponding to the most optimal expression level for beneficial overexpression targets) was preserved and tested in replicate for the final comparison. For deletions, genomic DNA at the target locus (a predicted high-efficiency cut site in the first 5–10% of the gene ORF) was sequenced and clones containing a frame-shift mutation were considered to have the target gene disrupted. gRNAs used, their gene deletion efficiencies, and the specific gene mutations observed are described in Additional file 1: Table S2.

Fig. 2
figure 2

Schematic overview of the strain engineering process employed in this study. SpCas9 and MaFAR were co-integrated to create a platform strain, 880CF, followed by integration of gRNA cassettes to create knockout mutants and gene overexpression cassettes for overexpression strains. neo gene encoding aminoglycoside 3'-phosphotransferase (G418 resistance), hpt gene encoding hygromycin phosphotransferase (hygromycin resistance), nat gene encoding nourseothricin acetyltransferase (nourseothricin resistance)

Full size image

Many genetic targets and strategies have been explored for the metabolic engineering of different yeast species for fatty alcohol production. For the engineering of R. toruloides, we tested various approaches that have shown prior success in yeast including upregulating the production of precursor metabolites, blocking pathways that compete for metabolic flux, reactivating acyl chains sequestered in lipids, and blocking fatty alcohol degradation pathways [9, 14] (Fig. 3). For overexpression targets, the native genomic copy of a gene was used unless otherwise noted. Mycocosm proteinIDs are provided for each native gene targeted in Additional file 1: Table S3 and a full list of strains created in this study is provided in Additional file 1: Table S4. The strains created were compared in glass culture tubes containing synthetic complete media with a 10% dodecane overlay inoculated to an initial OD600 of 0.1. The fatty alcohol concentration in the overlay was measured by GC-FID after 6 days of growth (Fig. 4).

Fig. 3
figure 3

Diagram of the roles metabolic engineering targets explored in this study play within the central carbon metabolism of R. toruloides. G6PD glucose 6-phosphate dehydrogenase, ACL1 ATP-citrate lyase, ACC1 acetyl-CoA carboxylase, LRO1 lecithin cholesterol acyltransferase, DGA1 diacylglycerol acyltransferase, TGL2 triglyceride lipase, FAA1 fatty acyl-CoA synthetase, ARE1 acyl-CoA:sterol acyltransferase, FAR fatty acyl-CoA reductase, PXA1 peroxisomal ABC-transporter, FAO fatty alcohol oxidase

Full size image
Fig. 4
figure 4

Fatty alcohol titers produced from single mutant strains derived from 880CF and tested in glass culture tube fermentation. For genes with multiple identified homologs, the Mycocosm proteinID is provided. Error bars represent ± standard deviation of biological triplicates. Presence of an asterisk indicates statistically significant improvement compared to 880CF (p < 0.05, student t-test). ACC1 acetyl-CoA carboxylase gene, ACL1 ATP-citrate lyase gene G6PD1 R. toruloides glucose-6-phosphate dehydrogenase gene, Yl G6PD Y. lipolytica glucose-6-phosphate dehydrogenase gene, Yl ME1 gene encoding Y. lipolytica malic enzyme, DGA1 diacylglycerol acyltransferase gene, LRO1 phospholipid:DAG acyltransferase gene, ARE1 acyl-CoA:cholesterol acyltransferase gene, FAA1 fatty acyl-CoA synthase gene, TGL2 triglyceride lipase gene, FAO fatty alcohol oxidase gene, PXA1 peroxisomal ABC transporter gene, AMPD1 AMP deaminase gene, SCD1 stearoyl-CoA desaturase gene

Full size image

Increasing production of precursor metabolites: MaFAR produces fatty alcohols through the reduction of fatty acyl-CoA chains, consuming two molecules of NADPH per fatty alcohol molecule produced [23]. Fatty acyl-CoA chains are produced by the fatty acid synthase complex, consuming one malonyl-CoA molecule and two NADPH molecules per two-carbon extension of the growing acyl-CoA chain [1]. Therefore, we tested overexpression of the native ACL1 to increase cytosolic acetyl-CoA production, overexpression of the native ACC1 to increase malonyl-CoA production from acetyl-CoA, and overexpression of three NADPH-producing genes (R. toruloides glucose-6-phosphate dehydrogenase gene (G6PD1), Y. lipolytica G6PD1, and Y. lipolytica malic enzyme gene (ME1)).

ACC1 and ACL1 overexpression resulted in notable 1.8-fold and 3.7-fold increases in the fatty alcohol titer, respectively. Overexpression of the native G6PD1 or Y. lipolytica G6PD1 and ME1 did not significantly increase the titer.

Blocking or reversing lipid formation: In wild-type R. toruloides, as in other oleaginous yeasts, the typical destination for fatty acyl-CoA chains is storage in lipid droplets as triacylglycerides (TAGs) [29]. We attempted to block this competing pathway through the disruption of two genes involved in TAG formation, DGA1 and LRO1. DGA1 catalyzes the transfer of an acyl chain from an acyl-CoA molecule to a diacylglyceride (DAG), while LRO1 forms TAGs by transferring an acyl chain from a phospholipid to a DAG [30]. We additionally tested the disruption of ARE1 encoding an acyltransferase involved in sterol esterification [31].

Disruption of DGA1 and LRO1 increased the fatty alcohol titer by 2.3-fold and 4.4-fold, respectively, indicating that these acyltransferases (or in the case of LRO1, possibly an upstream protein as acyl-CoA is not thought to be a direct substrate) are noteworthy competitors with MaFAR for acyl-CoA molecules. ARE1 disruption, however, did not increase the fatty alcohol titer. Due to the promising results of DGA1 and LRO1 disruption, simultaneous knockout of both genes was attempted by integrating a single construct with expression cassettes for gRNAs targeting both genes and sequencing the resulting colonies at both gene loci. However, only single knockouts, and a small number of very slow-growing colonies unable to be sequenced were obtained, suggesting the DGA1ΔLRO1Δ double mutation may be lethal, or strongly detrimental to growth in R. toruloides. As these two genes are expected to account for the majority of TAG formation [30], one possible explanation for the failure to obtain the double mutant is that the inability to form TAGs confers a severe fitness defect in R. toruloides.

An alternative approach we explored was to reactivate acyl chains that have been sequestered in TAGs, first by hydrolyzing TAGs with triglyceride lipase (TGL2) to form free fatty acids, then reactivating the free fatty acids with acyl-CoA synthetase (FAA1). As overexpression of both genes may be necessary for this strategy to work, FAA1 and TGL2 overexpression cassettes driven by the ANT1 and TEF1 promoters, respectively, were cloned to a single plasmid and integrated simultaneously. However, overexpression of these two genes did not increase the fatty alcohol titer.

Blocking fatty alcohol degradation: The elimination of fatty alcohol degradation pathways has been shown to improve fatty alcohol titers in S. cerevisiae and Y. lipolytica [9, 14]. Three IFO0880 genes were identified with homology to Y. lipolytica fatty alcohol oxidase gene (FAO1). Furthermore, a popular strategy for increasing lipid production which has also proved effective in improving fatty alcohol titers is to disrupt β-oxidation, either by disrupting peroxisome biogenesis through deletion of PEX10 [32, 33], or by interfering with peroxisomal uptake of fatty alcohols or fatty acyl-CoA by deleting the peroxisomal transporter genes PXA1 and PXA2 [14]. While PEX10 is reported to be essential in R. toruloides [27] (and supporting this, we were unable to obtain any viable PEX10Δ mutants), two genes were identified with homology to S. cerevisiae PXA1 and PXA2 and successfully targeted for disruption. However, disruption of the three FAO genes and both PXA1-like genes did not significantly improve the fatty alcohol titer.

Other targets: Overexpression of stearoyl-CoA desaturase gene (SCD1 or OLE1) [34] and AMP deaminase gene (AMPD1) [33] have both been shown to increase lipid production in Y. lipolytica, while SCD1 overexpression also improved fatty alcohol production in S. cerevisiae [14]. AMPD1 is involved in the initiation of lipid biosynthesis, reducing the mitochondrial AMP concentration to inhibit isocitrate dehydrogenase, resulting in the accumulation of citrate, which can be converted to cytosolic acetyl-CoA by ACL1 following export from the mitochondria [35]. SCD1 converts stearoyl-CoA to oleyl-CoA, relieving stearoyl-CoA-mediated allosteric inhibition of ACC1, allowing for more malonyl-CoA formation in the presence of long-chain acyl-CoA molecules [34]. However, we found that neither of these approaches increased the fatty alcohol titer in R. toruloides.

Combinatorial exploration of validated single targets: Although we were unable to disrupt the two most promising deletion targets, LRO1 and DGA1, in the same strain, both single knockouts were evaluated in combination with the two best overexpression targets, ACC1 and ACL1, individually and in tandem (Fig. 5). Combination of DGA1Δ with ACC1 overexpression was found to increase the fatty alcohol titer from the initial 106 mg/L to 206 mg/L. However, all of the LRO1Δ strains with ACC1 or ACL1 overexpression had lower fatty alcohol titers than the starting strain’s titer of 202 mg/L.

Fig. 5
figure 5

Fatty alcohol titers produced from double and triple mutant strains derived from 880CF and tested in glass culture tube fermentations. Error bars represent ± standard deviation of biological triplicates. Presence of an asterisk indicates statistically significant improvement compared to the parent strain (p < 0.05, student t-test). ACC1 acetyl-CoA carboxylase gene, ACL1 ATP-citrate lyase gene, DGA1 diacylglycerol acyltransferase gene, LRO1 phospholipid:DAG acyltransferase gene

Full size image

Bioreactor fermentation of fatty alcohols

Three of the best-performing DGA1Δ-based strains (DGA1Δ, DGA1Δ-ACC1, and DGA1Δ-ACC1-ACL1) and LRO1Δ-based strains (LRO1Δ, LRO1Δ-ACL1, and LRO1Δ-ACC1-ACL1) were tested in 250 mL Eppendorf DASbox Mini Bioreactors with a 20% dodecane overlay in week-long fed-batch fermentations with glucose concentrations restored to 50 g/L each day for the first five days (Fig. 6, Table 2). As in the culture tube-scale tests, the LRO1Δ strain performed better than its derivatives with additional overexpression targets, while the DGA1Δ strain benefitted from ACC1 and ACC1-ACL1 overexpression. The highest fatty alcohol titer and yield of 3.7 g/L and 0.024 g/g glucose, respectively, were obtained in 880CF-LRO1Δ. In each experiment, the peak fatty alcohol titer was reached no later than day 7. Each strain showed similar growth characteristics, reaching an OD600 between 25 and 30 after approximately 48 h of growth. In most cases, major fatty alcohol secretion began at this point and continued until ~ 48 h after the glucose feed was stopped.

Fig. 6
figure 6

Fatty alcohol titers produced, glucose consumed, and OD600 values in bioreactor fermentations of the strains A 880CF-LRO1Δ, B 880CF-LRO1Δ-ACL1, C 880CF-LRO1Δ-ACC1-ACL1, D 880CF-DGA1Δ, E 880CF-DGA1Δ-ACC1, and F 880CF-DGA1Δ-ACC1-ACL1

Full size image
Table 2 Peak fatty alcohol titers and yields resulting from pulse-fed batch fermentation in Eppendorf DASbox Mini Bioreactors for 7 days
Full size table

Lipidomic analysis of DGA1Δ and LRO1Δ Rhodotorula strains

Our finding that LRO1Δ improves fatty alcohol production more than DGA1Δ in R. toruloides stands in contrast to a report in S. cerevisiae, where DGA1 deletion was highly beneficial but LRO1 deletion abolished fatty alcohol production entirely [14]. It is surprising that LRO1Δ would improve fatty alcohol production so significantly, as the reaction LRO1 is known to catalyze (reversible transfer of an acyl chain from a phospholipid to a DAG) does not directly use acyl-CoA as a substrate, unlike DGA1 which consumes acyl-CoA to convert DAGs into TAGs.

To investigate and compare the mechanisms by which these knockouts lead to improved fatty alcohol production in R. toruloides, a lipidomic study was performed comparing the strains 880CF, 880CF-LRO1Δ, and 880CF-DGA1Δ after six days of growth on synthetic media. Most strikingly, the DGA1Δ mutant has greatly reduced TAG accumulation, as expected, while the LRO1Δ mutant has increased TAG and DAG accumulation (Fig. 7a, b). This result suggests that DGA1 is likely a net TAG producer and direct competitor with MaFAR for acyl-CoA molecules as expected, while either LRO1 is a net TAG consumer under our experimental conditions, moving acyl chains from the acylglycerol population to lyso-phospholipids, or the deletion of LRO1 has increased the overall carbon flux entering lipid biosynthesis.

Fig. 7
figure 7

Intracellular concentrations of the lipid species A TAG; B DAG; C PC, PE, LPC and LPE; and D CL as determined by lipidomics, in the strains 880CF, 880CF-DGA1Δ, and 880CF-LRO1Δ. Error bars represent ± standard deviation of biological triplicates. PE phosphatidylethanolamine, PC phosphatidylcholine, LPC lyso-phosphatidylcholine, DAG diacylglyceride, TAG triacylglyceride, LPE lyso-phosphatidylethanolamine, CL cardiolipin. Presence of an asterisk indicates statistically significant improvement compared to 880CF (p < 0.05, student t-test)

Full size image

No shifts in the balances between phosphatidylcholine (PC) and lyso-PC (LPC), or phosphatidylethanolamine (PE) and lyso-PE (LPE) were observed (Fig. 7c), and no other major lipid species were significantly depleted in the LRO1Δ mutant (Additional file 1: Table S5) suggesting the extra carbon entering the DAG and TAG pools in this mutant must have come from outside the lipid network. A comparison of the carbon consumption of the three strains showed 880CF-LRO1Δ having the highest glucose consumption after six days of growth, at 14 g/L compared to 10.5 g/L consumed by 880CF (Additional file 1: Fig. S1). Therefore, we consider the most likely explanation for the increased fatty alcohol production and DAG/TAG production of 880CF-LRO1Δ to be a general increase in sugar uptake and carbon flux entering the lipid biosynthesis pathway, although the exact mechanism causing this increase in flux warrants further study. Finally, we also observed that cardiolipin (CL) levels in the LRO1Δ strain were significantly lowered (Fig. 7d), suggesting disruptions to phosphatidylglycerol metabolism may have occurred as a result of the knockout, although the decrease in this membrane lipid species alone is not enough to account for the much larger increase in TAGs and DAGs.

Discussion

While R. toruloides has been explored as a production host for a variety of molecules, fewer efforts have aimed to engineer its metabolism and improve product titers, particularly with strategies incorporating the still-difficult process of gene disruption. Zhang and coworkers identified a number of overexpression targets capable of improving lipid production in R. toruloides, including ACC1, DGA1, ME1, and SCD1, and investigated the knockout of a single gene, PEX10, although its knockout did not increase the lipid titer [12, 25]. However, numerous R. toruloides strains developed to produce other bioproducts such as terpenoids [36,37,38], fatty acid ethyl esters [39], and, until this study, fatty alcohols [10, 22] have not been optimized further than the expression of the heterologous enzyme responsible for the formation of the final product. Therefore, the development of a workflow for performing sophisticated metabolic engineering as well as the identification of broadly useful metabolic engineering targets in this non-model yeast is of significant importance to the research community.

In this study, a CRISPR/Cas9 tool based on previous reports [15, 16] was used in conjunction with constitutive gene overexpression to create an array of single, double, and triple overexpression/deletion mutants in a fatty alcohol-producing strain derived from IFO0880. Notably, the use of CRISPR/Cas9 for the generation of knockouts greatly improved the throughput of the R. toruloides strain engineering process, allowing the efficient creation of eight knockout strains for characterization. In screening the engineered strains, we found that increasing the supply of precursor metabolites in the form of acetyl-CoA and malonyl-CoA through overexpression of ACL1 and ACC1 and reducing competition for fatty acyl-CoA molecules from the lipid biosynthetic pathway through disruption of DGA1 and LRO1 proved to be the most effective strategies for improving the fatty alcohol titer. Increasing the production of the reducing cofactor NADPH, used for acyl-CoA biosynthesis and the reduction of acyl chains to fatty alcohols, was less effective, suggesting NADPH supply may not be limiting for this product in R. toruloides. Disruption fatty alcohol oxidase-like genes (identified based on homology to Y. lipolytica FAO1) and peroxisomal transporter PXA1-like genes had no effect, as did overexpression of SCD1, AMPD1, and FAA1 + TGL2, and disruption of ARE1.

While LRO1 and DGA1 were selected as knockout targets based on similar reasoning (attempting to increase availability of acyl-CoA chains by preventing their incorporation into TAGs), the significantly differing properties of the resulting knockout strains highlight the different metabolic roles played by these enzymes in R. toruloides. Most notably, while the DGA1Δ mutant had largely abolished TAG formation compared to the wild-type strain, the LRO1Δ mutant surprisingly had elevated DAG and TAG levels. This stands in contrast to Y. lipolytica, where DGA1Δ and LRO1Δ were found to moderately lower TAG levels individually and a DGA1Δ LRO1Δ double mutant showed a more substantial decrease in TAG [30]. An additional report in Y. lipolytica demonstrated that in this yeast species, overexpression of LRO1 increased lipid production [40]. Our finding in R. toruloides implicates DGA1 as the major producer of TAGs, while the role of LRO1 is less certain. For future studies of lipid production in R. toruloides, both overexpression and knockout of LRO1 may be informative to explore. The fact that the LRO1Δ mutant showed a greater improvement in fatty alcohol production than the DGA1Δ mutant is partially explained by the increased glucose consumption of the LRO1Δ mutant, although the LRO1Δ mutant also showed a higher carbon yield in culture tube and bioreactor fermentations.

Combining the beneficial overexpression of ACL1 and ACC1 was found to further increase the fatty alcohol titer in the DGA1Δ strain background, up to 3.1 g/L in bioreactor fermentation, while the LRO1Δ mutant’s titer of 3.7 g/L in the bioreactor was not further improved. This discrepancy may be due to different characteristics of the starting strains, with 880CF-DGA1Δ producing about half as much fatty alcohol as 880CF-LRO1Δ. As both of these knockouts are reasoned to improve the fatty alcohol production by redirecting fatty acyl-CoA flux from TAG formation to fatty alcohol formation, 880CF-LRO1Δ likely has a higher level of fatty acyl-CoA available as substrate for the FAR. Overexpression of ACC1 and ACL1 is intended to increase production of fatty acyl-CoA, but availability of this precursor may have been a limiting factor for fatty alcohol production only in 880CF-DGA1Δ.

Cordova and coworkers were able to produce 5.8 g/L fatty alcohols in a Y. lipolytica strain lacking β-oxidation (MFE1Δ) and peroxisome biogenesis (PEX10Δ), but containing DGA1 overexpression [9]. While disruption of peroxisomal transporters was ineffective in our study and a PEX10Δ mutant could not be obtained (the gene has been reported to be essential in IFO0880 [27]), in the future a more comprehensive elimination of β-oxidation, such as by deletion of MFE1, may further improve fatty alcohol production. The high fatty alcohol titer in a DGA1-overexpressing Y. lipolytica strain is interesting, but it may be that DGA1Δ rather than overexpression would further improve the titer of fatty alcohols as strains with native DGA1 expression levels or DGA1Δ were not compared. In our study and D’Espaux and coworkers’ study in S. cerevisiae, DGA1Δ improved fatty alcohol production [14]. Also in agreement between these studies is the beneficial effect of overexpression of ACC1, while upregulation of unsaturated fatty-acyl production to potentially relieve feedback inhibition of saturated fatty acyl-CoA on the fatty acid synthase complex [1] was more beneficial in S. cerevisiae (OLE1 overexpression) than R. toruloides (SCD1 overexpression), resulting in a fourfold increase in fatty alcohol titer compared to the parent strain. In contrast, LRO1Δ was effective in promoting R. toruloides’s fatty alcohol production but detrimental to S. cerevisiae’s. The difference in response to SCD1 overexpression may be attributable to the different fatty acid profiles of S. cerevisiae and R. toruloides: S. cerevisiae produces predominantly saturated C16 fatty acids and alcohols, while saturated and unsaturated C18 fatty acids and alcohols are major components of the R. toruloides product spectrum [10, 14]. This difference could confound the effect of SCD1 overexpression in numerous ways, such as through additional post-translational regulation of the SCD1 protein, or through the presence of much greater amounts of saturated C18-CoA substrate for the enzyme to act on to meaningfully shift the fatty acid profile.

Production of fatty alcohols has been previously reported in R. toruloides through heterologous expression of MaFAR without further engineering, with 8 g/L achieved in a bioreactor on YP-sucrose with 5 g/L tergitol and no organic overlay [10] and 1.7 g/L achieved in a bioreactor on synthetic complete media with 0.1 g/L of the nonionic surfactant tergitol and a 20% vol/vol dodecane overlay to improve fatty alcohol extraction [22]. This study, in comparison, achieved a maximum titer of 3.7 g/L in a bioreactor on synthetic complete media without tergitol and with a 20% vol/vol dodecane overlay in a LRO1Δ strain. Improved titers were also identified in DGA1Δ as well as ACC1 and ACL1 overexpressing strains identifying all of these targets as well as LRO1Δ as relevant to the fatty alcohol overproduction phenotype in R. toruloides. Combination of these metabolic engineering targets with other strategies such as addition of nonionic surfactants or different media formulations may enable additional increases in fatty alcohol productivity from R. toruloides. Furthermore, as the gene editing toolkit for R. toruloides becomes more advanced, the use of selection marker recycling or an episomal plasmid will facilitate the creation of strains with a larger number of modifications and further improved titers.

Conclusions

In this study, a Cas9-FAR-expressing strain of R. toruloides IFO0880 was created and used as a platform organism for the exploration of 16 overexpression and deletion metabolic engineering targets. Several promising targets were explored combinatorially at the culture tube and bioreactor scale, and the best-performing strain, harboring LRO1Δ, produced 3.7 g/L fatty alcohols from synthetic media. In contrast to findings in other yeast, DGA1 and LRO1 knockouts were both found to increase fatty alcohol production in R. toruloides, while a lipidomic survey showed that DGA1 is the predominant TAG-producing protein in this yeast. Based on these findings, the role of LRO1 in R. toruloides lipid biosynthesis and fatty acyl-CoA consumption appears to differ from other yeasts. Further investigation may lead to new insights about the lipid metabolism of R. toruloides.

Methods

Strains and media

The R. toruloides strain IFO0880 was grown at 30 °C, 250 rpm. YPD media (10 g/L yeast extract, 20 g/L peptone, 20 g/L glucose) was used for routine handling of cells. For selection or maintenance of transformants, 200 µg/mL G418 (KSE Scientific, Durham, NC), 100 µg/mL nourseothricin (Gold Biotechnology, St. Louis, MO), or 50 µg/mL hygromycin (Gold Biotechnology, St. Louis, MO) was supplemented. Fatty alcohol production was measured in synthetic complete medium (1.7 g/L yeast nitrogen base (BD, Franklin, NJ), 5 g/L ammonium sulfate, 0.78 g/L complete synthetic mixture (MP Biomedicals, Santa Ana, CA), 40 g/L glucose) adjusted to pH 7 using sodium hydroxide. Culture tube-scale fatty alcohol fermentation experiments were performed by preculturing the yeast for 48 h in SC media, then inoculating 5 mL of SC media in glass culture tubes to an OD600 of 0.1, with a 10% dodecane overlay containing 100 mg/L of pentadecane as internal standard and grown for 6 days.

Standard cloning was performed in the Escherichia coli strain NEB10β (New England Biolabs, Ipswich, MA). Cells were grown on Luria Broth (LB) medium at 37 °C, 250 rpm with 100 µg/mL ampicillin or 50 µg/mL kanamycin. Multi-fragment cloning of 20 kb and larger plasmids was performed in S. cerevisiae BY4741 using DNA assembler [41], with growth on YPD or synthetic uracil dropout media (1.7 g/L yeast nitrogen base, 5 g/L ammonium sulfate, 0.78 g/L complete synthetic mixture without uracil, 20 g/L glucose).

Gene synthesis

Genes were synthesized by Twist Bioscience (San Francisco, CA) following codon optimization by the JGI BOOST tool, set to use the most frequent codon for each amino acid as they occur in R. toruloides [42].

Plasmid construction

The plasmid pRTN was constructed for heterologous expression or native gene overexpression from the E. coli elements of pUC19 (pMB1 origin, ampicillin resistance), the S. cerevisiae elements of pRS426 (2µ origin and URA3 selection marker), the strong R. toruloides p17 promoter, GFP, the T35S terminator, and a R. toruloides nourseothricin resistance cassette from pGI2 [25] using DNA assembler. The promoter was subsequently replaced if needed using AflII and MfeI digestion, and expressed gene was replaced using MfeI and SpeI digestion. Promoters and native genes were PCR amplified from IFO0880 genomic DNA extracted using lithium acetate/SDS/heat lysis [43], while heterologous genes were synthesized by Twist Bioscience. Multi-gene expression plasmids were pieced together from either the pRTN or NM9 [15] backbones (for nourseothricin or G418 resistance, respectively) and the appropriate promoter and gene elements using DNA assembler. To prevent complications in the assembly, unique terminators (Tnos, T35S, and Ttub) were used for each gene and the selection marker.

The plasmid pRTH was constructed for gRNA cloning and expression from the E. coli elements of pUC19 (pMB1 origin, ampicillin resistance), the S. cerevisiae elements of pRS426 (2µ origin and URA3), a gRNA expression cassette with the IFO0880 5S rRNA, tRNATyr, 2 BsaI sites, the S. cerevisiae SUP4 terminator, and a R. toruloides hygromycin resistance cassette from pZPK-PGPD-HYG-Tnos [44] using DNA assembler. gRNAs were cloned by digesting pRTH with BsaI and ligating two annealed and 5’-phosphorylated oligos with a forward strand 5’ GGGA and reverse strand 5’ AAAC overhang. For attempted double deletions, 2 tandem gRNA expression cassettes were synthesized as a gBLOCK and cloned into the same BsaI restriction sites using Golden Gate assembly. A list of primers used in this study is provided in Additional file 1: Table S6, and a list of plasmids created is provided in Additional file 1: Table S7.

Transformation of R. toruloides

R. toruloides was transformed using heat shock as has been described previously [16]. Briefly, a colony was picked and cultured overnight in YPD supplemented with an appropriate antibiotic if necessary. The overnight culture was used to inoculate a shake flask with 25 or 50 mL of YPD (for up to 5 or 10 transformations, respectively) to an OD600 of 0.2 and cultured for 4 h. The cells were collected by centrifugation, washed with water twice, and mixed with 240 µL PEG3350 (Sigma Aldrich, St. Louis, MO), 36 µL 1 M lithium acetate (Sigma Aldrich, St. Louis, MO), 50 µL of 2 mg/mL salmon sperm DNA (Sigma Aldrich, St. Louis, MO), and 1–2 µg of linear DNA dissolved in 40 µL of water. The cells were incubated with shaking in the transformation mixture for 30 min at 30 °C. Then, 34 µL of DMSO was added to the mixture, which was briefly vortexed, and heat shocked at 42 °C for 15 min. The cells were collected, washed once with water, resuspended in 2 mL YPD, and allowed to recover overnight. The cells were then collected and spread to YPD agar plates supplemented with the appropriate antibiotic. Transformation efficiencies of ~ 102 CFU/µg DNA were typically observed.

Genetic manipulation of R. toruloides

All genetic modifications to R. toruloides were made by random integration of linear PCR or restriction digestion fragments using heat shock transformation. After cloning of the gene or gRNA expression cassette to be transformed to a plasmid as described above, fragments were prepared either through PCR amplification (with the primers ZPK F/R, or gRNA F/R, respectively) from their plasmid, or (for cassettes longer than 7 kb) through excision of this fragment by digestion with suitable restriction enzymes, followed by spin column purification. Overexpression cassettes were created by expressing either the R. toruloides genomic copy of a gene (for endogenous genes) or a synthetic codon optimized gene (for heterologous genes) with the strong p17 promoter. Deletion mutants were created by transforming a single gRNA targeting the first 10% of the target gene ORF to generate frame shift mutations following NHEJ DNA repair. gRNAs were designed using the Benchling gRNA tool. For verification of deletion mutants, genomic DNA was extracted using lithium acetate/SDS/heat lysis [43], PCR amplified at the target locus and sequenced. For gene activation targets, genomic DNA was extracted using the same method and integration was verified using colony PCR.

Bioreactor fermentation

For bioreactor experiments, a seed culture was grown for 48 h in a culture tube in SC media, then transferred entirely to 50 mL of media in a shake flask for another 24 h. The cells were collected, washed once with water, and used to inoculate 100 mL of SC media to an OD600 of 1 in 250 mL Eppendorf DASbox Mini Bioreactors (Eppendorf, Hamburg, Germany) with 20 mL of dodecane containing 100 mg/L pentadecane. Bioreactors were set to maintain a pH of 7, agitation of 1200 rpm, temperature of 30 °C, and air flow rate of 6 standard liters per minute. Antifoam 204 was added to control foaming as necessary (typically 1 drop per reactor). Each day, samples were collected to measure OD600, fatty alcohol titer, and glucose concentration. Glucose was added to restore the concentration to 50 g/L for the first five days.

Analytical methods

Culture tube and bioreactor fermentation experiments were performed with a 10% or 20% dodecane overlay, respectively, with 100 mg/L pentadecane added as an internal standard. For quantification of fatty alcohols, 200 µL or 1 mL (for culture tube and bioreactor, respectively) was removed and centrifuged to separate the dodecane layer. 10 µL of dodecane was mixed with 90 µL of ethyl acetate and analyzed on an Agilent 8860 GC-FID with a 30-m DB-5 column (Agilent, Santa Clara, CA) with a temperature ramp of 70 °C for 3 min, increase at 20 °C/minute to 320 °C, hold 320 °C for 1 min. The total fatty alcohol titer was calculated as the sum of 1-hexadecanol, 1-octadecanol, and oleyl alcohol titers.

Glucose consumption was measured using an Agilent 1260 Infinity HPLC with a refractive index detector (RID) and H+ column (Rezex ROA-Organic Acid; Phenomenex, Torrance, CA). The column and detector were run at 50 °C and 0.6 mL/min of 0.005 N H2SO4 was used as the mobile phase.

Lipidomics

Cells were cultured in pH7-adjusted SCD media with a 10% dodecane overlay in glass culture tubes under the same conditions as for the fatty alcohol fermentations and cell pellets were collected after 6 days of growth following inoculation. Lipidomics analysis of both the reference and mutant yeast strains was performed using a two-step chloroform–methanol extraction as described elsewhere [45, 46]. Briefly, harvested cells were washed with 150 mM ammonium bicarbonate (ABC) followed by cell lysis with zirconium glass beads. Lipids were extracted from 1 OD unit of cell lysate, to which an internal lipid standard mix was added, in a 2-step method. First extraction was performed with 1 mL of 15:1 chloroform–methanol after which the chloroform layer was aspirated, dried, and then resuspended in an infusion solvent. The second extraction was performed on the remaining aqueous layer with 1 mL of 2:1 chloroform–methanol and again the chloroform layer was aspirated, dried, and resuspended in an infusion solvent. Both resuspended extracts were then infused into a Q-Exactive mass spectrometer via the Advion Triversa Nanomate in the nano-electrospray mode. The infusion solvent used for the dried lipid extract from the 15:1 extraction contained 7.5 mM ammonium formate in a mix of chloroform–methanol–propanol 1:2:4 (v/v) whereas the infusion solvent for the dried extract from the 2:1 extraction contained 0.05% methylamine in a mix of chloroform–methanol 1:5 (v/v) [46].

Identification and quantification of lipid species and classes from the lipidomic data generated by the mass spectrometer was performed using an in-house Python script that employed the pymzml library. Raw data files were input to the pipeline after conversion to the open source mzML format. The data analysis pipeline consisted of scan averaging, offline calibration, deisotoping and identification, quantification, and quantification across replicates. The list of internal standards and their absolute amounts used for spike-in are listed in Additional file 1: Table S8. The code used for analysis is available upon request.

Availability of data and materials

All data and materials produced in this study are available upon reasonable request.

References

  1. Krishnan A, McNeil BA, Stuart DT. Biosynthesis of fatty alcohols in engineered microbial cell factories: advances and limitations. Front Bioeng Biotechnol. 2020;8:1385.

    Google Scholar 

  2. Noweck K, Grafahrend W. Fatty alcohols. In: Ullmann’s encyclopedia of industrial chemistry. Weinheim, Germany: Wiley-VCH Verlag GmbH & Co. KGaA; 2006.

  3. Zimmermann H, Walzl R. Ethylene. In: Ullmann’s encyclopedia of industrial chemistry. Weinheim, Germany: Wiley-VCH Verlag GmbH & Co. KGaA; 2009.

  4. Mohd Azhar SH, Abdulla R, Jambo SA, Marbawi H, Gansau JA, Mohd Faik AA, et al. Yeasts in sustainable bioethanol production: a review. Biochem Biophys Rep. 2017;10:52–61.

    PubMed  PubMed Central  Google Scholar 

  5. Paddon CJ, Keasling JD. Semi-synthetic artemisinin: a model for the use of synthetic biology in pharmaceutical development. Nat Rev Microbiol. 2014;12(5):355–67.

    CAS  PubMed  Google Scholar 

  6. Hong KK, Nielsen J. Metabolic engineering of Saccharomyces cerevisiae: a key cell factory platform for future biorefineries. Cell Mol Life Sci. 2012;69(16):2671–90.

    CAS  PubMed  Google Scholar 

  7. Fatma Z, Jawed K, Mattam AJ, Yazdani SS. Identification of long chain specific aldehyde reductase and its use in enhanced fatty alcohol production in E. coli. Metab Eng. 2016;37:35–45.

    CAS  PubMed  Google Scholar 

  8. Kaczmarzyk D, Cengic I, Yao L, Hudson EP. Diversion of the long-chain acyl-ACP pool in Synechocystis to fatty alcohols through CRISPRi repression of the essential phosphate acyltransferase PlsX. Metab Eng. 2018;45:59–66.

    CAS  PubMed  Google Scholar 

  9. Cordova LT, Butler J, Alper HS. Direct production of fatty alcohols from glucose using engineered strains of Yarrowia lipolytica. Metab Eng Commun. 2020;10: e00105.

    PubMed  Google Scholar 

  10. Fillet S, Gibert J, Suárez B, Lara A, Ronchel C, Adrio JL. Fatty alcohols production by oleaginous yeast. J Ind Microbiol Biotechnol. 2015;42(11):1463–72.

    CAS  PubMed  PubMed Central  Google Scholar 

  11. Yu T, Zhou YJ, Huang M, Liu Q, Pereira R, David F, et al. Reprogramming yeast metabolism from alcoholic fermentation to lipogenesis. Cell. 2018;174(6):1549-1558.e14.

    CAS  PubMed  Google Scholar 

  12. Zhang S, Ito M, Skerker JM, Arkin AP, Rao CV. Metabolic engineering of the oleaginous yeast Rhodosporidium toruloides IFO0880 for lipid overproduction during high-density fermentation. Appl Microbiol Biotechnol. 2016;100(21):9393–405.

    CAS  PubMed  Google Scholar 

  13. Liu Y, Chen S, Chen J, Zhou J, Wang Y, Yang M, et al. High production of fatty alcohols in Escherichia coli with fatty acid starvation. Microb Cell Fact. 2016;15(1):129.

    PubMed  PubMed Central  Google Scholar 

  14. d’Espaux L, Ghosh A, Runguphan W, Wehrs M, Xu F, Konzock O, et al. Engineering high-level production of fatty alcohols by Saccharomyces cerevisiae from lignocellulosic feedstocks. Metab Eng. 2017;42:115–25.

    PubMed  Google Scholar 

  15. Schultz JC, Cao M, Zhao H. Development of a CRISPR/Cas9 system for high efficiency multiplexed gene deletion in Rhodosporidium toruloides. Biotechnol Bioeng. 2019;116(8):2103–9.

    CAS  PubMed  Google Scholar 

  16. Otoupal PB, Ito M, Arkin AP, Magnuson JK, Gladden JM, Skerker JM. Multiplexed CRISPR-Cas9-based genome editing of Rhodosporidium toruloides. mSphere. 2019;4(2):00099–119.

    Google Scholar 

  17. Jiao X, Zhang Y, Liu X, Zhang Q, Zhang S, Zhao ZK. Developing a CRISPR/Cas9 system for genome editing in the basidiomycetous yeast Rhodosporidium toruloides. Biotechnol J. 2019;14(7): e1900036.

    PubMed  Google Scholar 

  18. Liu X, Zhang Y, Liu H, Jiao X, Zhang Q, Zhang S, et al. RNA interference in the oleaginous yeast Rhodosporidium toruloides. FEMS Yeast Res. 2019;19:31.

    Google Scholar 

  19. Nora LC, Wehrs M, Kim J, Cheng JF, Tarver A, Simmons BA, et al. A toolset of constitutive promoters for metabolic engineering of Rhodosporidium toruloides. Microb Cell Fact. 2019;18(1):117.

    PubMed  PubMed Central  Google Scholar 

  20. Johns AMB, Love J, Aves SJ. Four inducible promoters for controlled gene expression in the oleaginous yeast Rhodotorula toruloides. Front Microbiol. 2016;7:1666–76.

    PubMed  PubMed Central  Google Scholar 

  21. Wang Y, Lin X, Zhang S, Sun W, Ma S, Zhao ZK. Cloning and evaluation of different constitutive promoters in the oleaginous yeast Rhodosporidium toruloides. Yeast. 2016;33(3):99–106.

    CAS  PubMed  Google Scholar 

  22. Liu D, Geiselman GM, Coradetti S, Cheng Y, Kirby J, Prahl J, et al. Exploiting nonionic surfactants to enhance fatty alcohol production in Rhodosporidium toruloides. Biotechnol Bioeng. 2020;117(5):1418–25.

    CAS  PubMed  PubMed Central  Google Scholar 

  23. Willis RM, Wahlen BD, Seefeldt LC, Barney BM. Characterization of a fatty acyl-CoA reductase from Marinobacter aquaeolei VT8: a bacterial enzyme catalyzing the reduction of fatty acyl-CoA to fatty alcohol. Biochemistry. 2011;50(48):10550–8.

    CAS  PubMed  Google Scholar 

  24. Partow S, Siewers V, Bjørn S, Nielsen J, Maury J. Characterization of different promoters for designing a new expression vector in Saccharomyces cerevisiae. Yeast. 2010;27(11):955–64.

    CAS  PubMed  Google Scholar 

  25. Zhang S, Skerker JM, Rutter CD, Maurer MJ, Arkin AP, Rao CV. Engineering Rhodosporidium toruloides for increased lipid production. Biotechnol Bioeng. 2016;113:1056–66.

    CAS  PubMed  Google Scholar 

  26. Dinh HV, Suthers PF, Chan SHJ, Shen Y, Xiao T, Deewan A, et al. A comprehensive genome-scale model for Rhodosporidium toruloides IFO0880 accounting for functional genomics and phenotypic data. Metab Eng Commun. 2019;9: e00101.

    PubMed  PubMed Central  Google Scholar 

  27. Coradetti ST, Pinel D, Geiselman GM, Ito M, Mondo SJ, Reilly MC, et al. Functional genomics of lipid metabolism in the oleaginous yeast Rhodosporidium toruloides. Elife. 2018;7: e32110.

    PubMed  PubMed Central  Google Scholar 

  28. Kim J, Coradetti ST, Kim YM, Gao Y, Yaegashi J, Zucker JD, et al. Multi-omics driven metabolic network reconstruction and analysis of lignocellulosic carbon utilization in Rhodosporidium toruloides. Front Bioeng Biotechnol. 2021;8: 612832.

    PubMed  PubMed Central  Google Scholar 

  29. Li Y, Zhao Z, Bai F. High-density cultivation of oleaginous yeast Rhodosporidium toruloides Y4 in fed-batch culture. Enzyme Microb Technol. 2007;41:312–7.

    Google Scholar 

  30. Athenstaedt K. YALI0E32769g (DGA1) and YALI0E16797g (LRO1) encode major triacylglycerol synthases of the oleaginous yeast Yarrowia lipolytica. Biochim Biophys Acta Mol Cell Biol Lipids. 1811;10:587–96.

    Google Scholar 

  31. Yang H, Bard M, Bruner DA, Gleeson A, Deckelbaum RJ, Aljinovic G, et al. Sterol esterification in yeast: a two-gene process. Science (80-). 1996;272(5266):1353–6.

    CAS  Google Scholar 

  32. Xue Z, Sharpe PL, Hong SP, Yadav NS, Xie D, Short DR, et al. Production of omega-3 eicosapentaenoic acid by metabolic engineering of Yarrowia lipolytica. Nat Biotechnol. 2013;31(8):734–40.

    CAS  PubMed  Google Scholar 

  33. Blazeck J, Hill A, Liu L, Knight R, Miller J, Pan A, et al. Harnessing Yarrowia lipolytica lipogenesis to create a platform for lipid and biofuel production. Nat Commun. 2014;5(1):1–10.

    Google Scholar 

  34. Qiao K, Imam Abidi SH, Liu H, Zhang H, Chakraborty S, Watson N, et al. Engineering lipid overproduction in the oleaginous yeast Yarrowia lipolytica. Metab Eng. 2015;29:56–65.

    CAS  PubMed  Google Scholar 

  35. Beopoulos A, Cescut J, Haddouche R, Uribelarrea JL, Molina-Jouve C, Nicaud JM. Yarrowia lipolytica as a model for bio-oil production. Prog Lipid Res. 2009;48(6):375–87.

    CAS  PubMed  Google Scholar 

  36. Yaegashi J, Kirby J, Ito M, Sun J, Dutta T, Mirsiaghi M, et al. Rhodosporidium toruloides: a new platform organism for conversion of lignocellulose into terpene biofuels and bioproducts. Biotechnol Biofuels. 2017;10(1):1–13.

    Google Scholar 

  37. Zhuang X, Kilian O, Monroe E, Ito M, Tran-Gymfi MB, Liu F, et al. Monoterpene production by the carotenogenic yeast Rhodosporidium toruloides. Microb Cell Fact. 2019. https://doi.org/10.1186/s12934-019-1099-8.

    Article  PubMed  PubMed Central  Google Scholar 

  38. Kirby J, Geiselman GM, Yaegashi J, Kim J, Zhuang X, Tran-Gyamfi MB, et al. Further engineering of R. toruloides for the production of terpenes from lignocellulosic biomass. Biotechnol Biofuels. 2021;14(1):101.

    CAS  PubMed  PubMed Central  Google Scholar 

  39. Zhang Y, Peng J, Zhao H, Shi S. Engineering oleaginous yeast Rhodotorula toruloides for overproduction of fatty acid ethyl esters. Biotechnol Biofuels. 2021;14(1):115.

    CAS  PubMed  PubMed Central  Google Scholar 

  40. Amalia L, Zhang YH, Ju YH, Tsai SL. Enhanced lipid production in Yarrowia lipolytica Po1g by over-expressing lro1 gene under two different promoters. Appl Biochem Biotechnol. 2020;191(1):104–11.

    CAS  PubMed  Google Scholar 

  41. Shao Z, Zhao H, Zhao H. DNA assembler, an in vivo genetic method for rapid construction of biochemical pathways. Nucleic Acids Res. 2009;37(2):1–10.

    CAS  Google Scholar 

  42. Oberortner E, Cheng JF, Hillson NJ, Deutsch S. Streamlining the design-to-build transition with build-optimization software tools. ACS Synth Biol. 2016;6(3):485–96.

    PubMed  Google Scholar 

  43. Lõoke M, Kristjuhan K, Kristjuhan A. Extraction of genomic DNA from yeasts for PCR-based applications. Biotechniques. 2011;50(5):325–8.

    PubMed  PubMed Central  Google Scholar 

  44. Lin X, Wang Y, Zhang S, Zhu Z, Zhou YJ, Yang F, et al. Functional integration of multiple genes into the genome of the oleaginous yeast Rhodosporidium toruloides. FEMS Yeast Res. 2014;14(4):547–55.

    CAS  PubMed  Google Scholar 

  45. Ejsing CS, Sampaio JL, Surendranath V, Duchoslav E, Ekroos K, Klemm RW, et al. Global analysis of the yeast lipidome by quantitative shotgun mass spectrometry. Proc Natl Acad Sci. 2009;106(7):2136–41.

    CAS  PubMed  PubMed Central  Google Scholar 

  46. Klose C, Tarasov K. Profiling of yeast lipids by shotgun lipidomics. Methods Mol Biol. 2016;1361:309–24.

    CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We thank Christopher Rao for the gifts of the strain IFO0880 and the plasmids pGI2 and pGI2_880_ACC.

Funding

This work was funded by the DOE Center for Advanced Bioenergy and Bioproducts Innovation (U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research under Award Number DE-SC0018420). Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the views of the U.S. Department of Energy.

Author information

Authors and Affiliations

  1. Department of Chemical and Biomolecular Engineering, Carl R. Woese Institute for Genomic Biology, University of Illinois at Urbana-Champaign, Urbana, IL, 61801, USA

    J. Carl Schultz, Shekhar Mishra, Emily Gaither, Andrea Mejia & Huimin Zhao

  2. U.S. Department of Energy Center for Bioenergy and Bioproducts Innovation (CABBI), Urbana, IL, 61801, USA

    J. Carl Schultz, Shekhar Mishra, Emily Gaither, Hoang Dinh, Costas Maranas & Huimin Zhao

  3. Department of Chemical Engineering, Pennsylvania State University, University Park, PA, 16802, USA

    Hoang Dinh & Costas Maranas

  4. Departments of Chemistry, Biochemistry, and Bioengineering, University of Illinois at Urbana-Champaign, Urbana, IL, 61801, USA

    Huimin Zhao

Authors
  1. J. Carl Schultz
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Shekhar Mishra
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Emily Gaither
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Andrea Mejia
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Hoang Dinh
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Costas Maranas
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Huimin Zhao
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Huimin Zhao.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

None declared.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1. Figure S1.

Glucose consumption of 880CF, 880CF-DGA1Δ, and 880CF-LRO1Δ. Table S1. DNA sequences of heterologous, codon-optimized genes evaluated in this study. Table S2. List of gRNAs used in this study. Table S3. List of Mycocosm proteinIDs for R. toruloides IFO0880 gene targets in this study. Table S4. List of R. toruloides strains used in this study. Table S5. Results of lipidomic comparison of 880CF, 880CF-DGA1Δ, and 880CF-LRO1Δ. Table S6. List of primers used in this study. Table S7. List of plasmids used in this study. Table S8. List of lipidomics internal standards and their absolute amounts used for spike-in.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Schultz, J.C., Mishra, S., Gaither, E. et al. Metabolic engineering of Rhodotorula toruloides IFO0880 improves C16 and C18 fatty alcohol production from synthetic media. Microb Cell Fact 21, 26 (2022). https://doi.org/10.1186/s12934-022-01750-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12934-022-01750-3

Keywords

Microbial Cell Factories

ISSN: 1475-2859

Contact us