Current advances in alteration of fatty acid profile in Rhodotorula toruloides: a mini-review

Microbial lipids are considered promising and environmentally friendly substitutes for fossil fuels and plant-derived oils. They alleviate the depletion of limited petroleum storage and the decrement of arable lands resulting from the greenhouse effect. Microbial lipids derived from oleaginous yeasts provide fatty acid profiles similar to plant-derived oils, which are considered as sustainable and alternative feedstocks for use in the biofuel, cosmetics, and food industries. Rhodotorula toruloides is an intriguing oleaginous yeast strain that can accumulate more than 70% of its dry biomass as lipid content. It can utilize a wide range of substrates, including low-cost sugars and industrial waste. It is also robust against various industrial inhibitors. However, precise control of the fatty acid profile of the lipids produced by R. toruloides is essential for broadening its biotechnological applications. This mini-review describes recent progress in identifying fatty synthesis pathways and consolidated strategies used for specific fatty acid-rich lipid production via metabolic engineering, strain domestication. In addition, this mini-review summarized the effects of culture conditions on fatty acid profiles in R. toruloides. The perspectives and constraints of harnessing R. toruloides for tailored lipid production are also discussed in this mini-review. Graphical abstract


Introduction
Owing to the depletion of petroleum storage and the growing world population, lipids derived from plants and animals have been exploited as an alternative to petroleum for producing biofuel and oleochemicals (Chemat et al. 2021;Ndiaye et al. 2019;Riazi et al. 2020;Shaah et al. 2021). However, production limitations (long harvesting time and unstable production affected by geographical and climate factors) and environmental issues (environmental pollution and competence of arable land for food crops) of plant-and animal-based lipids have forced scientists to seek new alternatives (Singh et al. 2020a;Meijaard et al. 2020;Toldrá-Reig et al. 2020). To date, much attention has been paid to the development of microbial lipids because of their short harvesting time and fewer environmental concerns (Bao et al. 2021;Liu et al. 2021c;Aamer Mehmood et al. 2021;Uthandi et al. 2021).
Rhodotorula toruloides is a non-conventional red yeast, also known as Rhodosporidium toruloides, R. rubescens, R. glutinis, or R. gracilis, that belongs to the subphylum Pucciniomycotina within the division Basidiomycota (Abeln and Chuck 2021;Koh et al. 2014). This strain is an excellent microbial lipid producer and has received great attention owing to its outstanding lipogenic ability. R. toruloides can accumulate up to 70% of its dry cellular weight as lipid content under nutrient-limiting conditions (nitrogen, phosphate, and sulfate) (Wang et al. 2018;Wu et al. 2010Wu et al. , 2011. It is also well known for its robustness, as it can produce lipids by utilizing a wide range of substrates from low-cost industrial by-products (Zhao et al. 2022;Wen et al. 2020;Saini et al. 2020;Park et al. 2018b). Numerous studies have described ways to improve lipid production in R. toruloides, including robust mutant selection, bioprocess optimization, and metabolic engineering (Zhao et al. 2022;Wen et al. 2020;Saini et al. 2020;Park et al. 2018b). In addition, available genome sequence (Dinh et al. 2019;Martín-Hernández et al. 2021;Zhu et al. 2012), characterized lipid metabolism (Jagtap et al. 2021), and accessible genetic tools (Bonturi et al. 2022;Jiao et al. 2019;Otoupal et al. 2019) for R. toruloides have also been established and developed. Two efficient and time-saving transformation methods by using Lithium-acetate (Tsai et al. 2017) and electroporation (Liu et al. 2017) were established to shorten the experimental time of the traditional method via Agrobacterium tumefaciens-mediated transformation (ATMT) (Lin et al. 2014). In addition, RNAi machinery was also proven to be functional for gene downregulation in R. toruloides . Genome editing toolboxes such as CRISPR-Cas9 system were also reported to achieve single or multi-genes editing in R. toruloides Otoupal et al. 2019;Schultz et al. 2019). All these advanced genetic tools and knowledge accelerated the improvement of the control of lipid production in R. toruloides and broaden its applications.
Microbial lipids rich in triacylglycerols are promising and attractive alternatives to existing petroleum-, plant-, and animal-based lipids. A triacylglycerol molecule consists of three fatty acid (acyl) chains attached to an alcohol glycerol backbone (Yoshinaga 2021). Carbon length, desaturation level, and position determine the physicochemical and biological properties of fatty acid (FA) chains attached to triacylglycerols (Cook and McMaster 2002;Temkov and Muresan 2021;Lee et al. 2022;Saini et al. 2021;Falomir-Lockhart et al. 2019 . In addition, a small proportion of FAs, such as myristic acid (MA, C14:0, 2% of TFA), arachidic acid (AA, C20:0, less than 0.5% of TFA), docosanoic acid (DA, C22:0, less than 0.5% of TFA), and tetracosanoic acid (TA, C24:0, less than 1% of TFA), have also been reported in R. toruloides (Krikigianni et al. 2022;Zhang et al. 2022a;Liu et al. 2021a). Therefore, R. toruloides-derived lipids could potentially be exploited as feedstock for plant oil substitutes in different applications, such as biodiesel, biolubricants, cosmetics, nutritional supplements, plastics, and coating materials (Carmona-Cabello et al. 2021;Papadaki et al. 2018;Adrio 2017;Lopez-Huertas 2010;Yu et al. 2014). For example, lipids rich in OA (over 70% of TFA) are preferred for biodiesel production due to their appropriate fluidity and stability during storage (Graef et al. 2009). Lipids with high saturated FA profiles are desired as equivalents to coconut (PA, 23-30%; SA, 32-37%; OA, 30-37%; LA, 2-4% of TFA) and cocoa (PA, 26%; SA, 35%; OA, 33%; LA, 3% of TFA) butter in food industries (Lipp et al. 2001;Papanikolaou and Aggelis 2010). Moreover, lipids rich in polyunsaturated fatty acids (PUFA) such as docosahexaenoic acid (C22:6, DHA) and eicosapentaenoic acid (C20:5, EPA) (16% and 17% of TFA) are considered as substitutes to fish oils (Lee et al. 2019). Lipids containing conjugated linoleic acids (CLAs, a group of isomers of LA), γ-linolenic acid (GLA, C18:3), and nervonic acid (NA, C24:1), which are relatively low in their sources, are considered as valuable lipids because of their clinical benefits (Szczepańska et al. 2022). However, commercially desired lipids are often produced at low concentrations or with inappropriate FA compositions in microbial lipids, which may lead to complex downstream processes and result in high production costs (Ochsenreither et al. 2016;Barta et al. 2021). Therefore, harnessing R. toruloudes for specific FA-rich lipid production is a vital issue for industrialization. Most studies have focused on maximizing the lipid-producing ability of R. toruloides (Zhao et al. 2022;Wen et al. 2020). Hence, few review papers have summarized and discussed strategies for controlling FA composition in R. toruloides. Further, a thorough discussion of the culture conditions that affect FA profiles in R. toruloides has not yet been done. In this mini-review, we summarize the advances in the control and production of tailored lipids and the effects of culture conditions on FA profiles in R. toruloides (Table 1).

Harnessing de novo FA synthesis pathways in R. toruloides
The FA composition of yeasts is precisely synthesized and regulated by FA synthases (Fass), elongases (Elos), and desaturases (Fads) (Heil et al. 2019;Singh et al. 2020b;Takaku et al. 2020;Matsuzawa et al. 2020). FA synthase, an essential enzyme in FA synthesis, catalyzes the reaction between acetyl-CoA and malonyl-CoA to produce C16-C18 FAs (Heil et al. 2019). R. toruloides Fas (RtFas) consists of two subunits (α-and β-subunits), which are encoded by FAS1 (RtFAS1) and FAS2 (RtFAS2), to form a multifunctional enzyme (Fischer et al. 2015;Zhu et al. 2012) (Fig. 1). RtFas1 contains acyltransferase and enoyl reductase domains. RtFas2 contains phosphopantetheinyl transferase, ketoacyl synthase, ketoacyl reductase, two acyl carrier proteins, malonyl/palmitoyl transferase, and dehydratase domains (Fig. 1). Liu et al. (2019) demonstrated that knocking down RtFAS1 and RtFAS2 using RNA interference decreased lipid content, but did not change the FA profile in R. toruloides NP11. Heterologous expression of RtFas in Saccharomyces cerevisiae also showed no effect on the FA composition (Zhou et al. 2016). These studies imply that FAS may not be a good candidate for manipulating FA composition in R. toruloides. However, enzyme engineering of the KS domain of Fas in the oleaginous yeast Yarrowia lipolytica enhanced the content of medium chain FAs (Rigouin et al. 2017), which suggests that engineering Fas may be applied in R. toruloides to modify the FA composition.
Once malonyl-CoA is elongated to the carbon length of the acyl chain to C16, C16 acyl-CoA is released from the barrel-shaped Fas in R. toruloides. The released C16 acyl-CoA is further modified by FA Elos and Fads in R. toruloides (Zhu et al. 2012) (Fig. 2). Fads catalyze the desaturation of FAs by introducing double bonds into C-C bonds, which generate monounsaturated FAs (MUFAs) and polyunsaturated FAs (PUFAs) Wang et al. 2022a). Elos mediate the elongation of FAs by adding extra carbon atoms to the acyl chains. All desaturases and elongases exhibit different substrate specificities and spatial and temporal characteristics (Cerone and Smith 2022;Szczepańska et al. 2022). Therefore, elucidation of the functions of Fads and Elos could assist in the control of FA composition and production of customized lipids in R. toruloides.
The functions of Elos have been well studied in the model yeast S. cerevisiae and other yeasts used in the industry (Kihara 2012;Uemura 2012;Rigouin et al. 2018;Matsuzawa et al. 2020); however, few studies have elucidated the functions of Elos in R. toruloides. A patent (Fillet et al. 2016) reported that overexpression of RtELO1 significantly enhanced the OA content to 70% of the TFA in R. toruloides. In another patent (Liu et al. 2018), overexpression of RtELO1 led to an increase in OA. In addition, disruption of RtELO2 did not affect FA profiles (Liu et al. 2018). Conversely, studies on RtFads have also been conducted. R. toruloides Δ9-desaturase (RtOle1) catalyzes the conversion of SA to OA by introducing a cis-double bond at the Δ9 position of acyl-CoA. OA, one of the most essential and abundant FA in R. toruloides, provides appropriate fluidity and excellent thermal and oxidative stability for oleochemical applications such as biodiesel, biolubricants, and hydraulic fluids (Wang et al. 2022b). OA production is determined by RtOle1, which is a rate-limiting enzyme in LA and ALA biosynthesis (Nagao et al. 2019). Hence, RtOle1 plays a vital role in controlling the FA profiles. Overexpression of RtOle1 in R. toruloides have been shown to effectively increase OA content up to 70% and only slightly increase POA content (Tsai et al. 2019). Liu et al. (2021b) suggested that RtOle1 has a strong substrate preference for stearoyl-CoA over palmitoyl-CoA. In addition, blocking the downstream desaturation of OA into LA by disrupting RtFAD2 could also increase the OA content from 30 to 60% TFA. LA and ALA are essential FAs that humans cannot synthesize. Hence, lipids with high LA and ALA contents have high value as nutraceuticals in the human diet (Chen and Liu 2020). Liu et al. (2021b) also demonstrated that R. toruloides FAD2 and FAD4 encode Δ12/Δ15 desaturase and ∆9/Δ12/Δ15 desaturase, respectively, which produce LA and ALA from OA by introducing C-C double bonds at the Δ12 and Δ15 positions of FAs. Additionally, the co-overexpression of RtOLE1 and RtFAD2 has been applied to produce LA-rich lipids . Nevertheless, downregulation of RtOLE1 expression by elevated LA and ALA was observed in R. toruloides . Furthermore, two regulatory elements, ORE1 and ORE2, upstream of RtOLE1 have been reported (Liu et al. 2021b). The authors suggested that   (Liu et al. 2015). Taken together, these studies show that manipulating the endogenous FA synthesis pathway in R. toruloides is an effective strategy for tailoring lipid production. Furthermore, the regulation of the native FA synthesis pathway should also be considered for a rational design to enhance specific FA production.

Metabolic engineering by introduction of exogenous genes for specific high-value lipid production
In addition to manipulating the native FA synthesis pathway, overexpression of genes from plants or other fungi was attempted in R. toruloides to produce specific high-value lipids. Disrupting native aldehyde dehydrogenase (ALD1) and overexpressing FAD2 from Mortierella alpina and FAD3 from Aleurites fordiiek led to a total ALA content of up to 49% of TFA (Liu et al. 2018  . GLA is a valuable omega-6 FA predominantly present in plants in relatively low amounts. GLA has anti-inflammatory properties and has applications in the treatment of several diseases (Liu et al. 2021b). Furthermore, Liu et al. (2021b) succeeded in producing GLA by disrupting native RtFAD2 and co-overexpressing M. alpine FAD2 and FAD6. The OA and GLA contents were further increased to 60.1% and 27.3% of TFA, respectively, in 2-L bioreactors.
FAs with carbon lengths of 20 or more are known as verylong-chain FAs (VLCFAs). VLCFAs, such as erucic acid (EA, C22:1) and NA, are potentially renewable feedstocks for the production of plastics, cosmetics, resins, nylon, and lubricants (Fillet et al. 2017). Additionally, NA has the potential to be used in the treatment of several neurological diseases (Liu et al. 2021a). Hence, EA and NA are highvalue FAs. However, R. toruloides cannot naturally produce either EA or NA. To produce VLCFA-rich lipids, Fillet et al. (2017) introduced various plant-derived 3-ketoacyl-CoA synthases (KCSs) into R. toruloides and successfully produced EA and NA. By increasing the KCS copy number and co-overexpressing plant-derived ELOs, the engineered strain produced more EA and NA, which accounted for 27% of TFA in 7-L bioreactors.

Domestication of R. toruloides strains for tailored lipid production
Although metabolic engineering is a powerful strategy for harnessing R. toruloides for tailored lipid production, lack of information of the genetic background is the main limitation for strain engineering. Therefore, other strategies, such as adaptive laboratory evolution (ALE) and mutagenesis by physical/chemical methods, are also popular for isolating mutants with the desired traits for tailored lipid production in R. toruloides (Wen et al. 2020). The mutation rate is one of the main differences between ALE and mutagenesis using physicochemical methods. ALE domesticates parental strains with simultaneous mutation rates by growing several generations under selective pressure conditions (Phaneuf et al. 2020). Mutagenesis by physical and chemical methods commonly introduces a high mutation rate by random DNA double-strand breakage with UV irradiation and chemical agents. Although mutagenesis by physical and chemical methods is an effective way to introduce mutations into the genome, isolated mutants usually possess undesired defects (Arora et al. 2020). Owing to the advantages and disadvantages of the two methods, researchers have employed ALE and mutagenesis by physical and chemical methods in R. toruloides. Wu et al. (2018) attempted to overcome the high temperatures that occur during fermentation by isolating thermotolerant R. toruloides DMKU3-TK16. The thermotolerant strain L1-1 was isolated under heat stress at 37°C using the ALE method. The isolated strain exhibited improved growth and lipid productivity at 37°C and a high OA content (90%) when cultured at 37°C. Liu et al. (2021d) attempted to improve the tolerance of R. toruloides to toxic lignocellulosic hydrolysates for lipid production. The isolated strains showed improved growth in media containing lignocellulosic biomass hydrolysate. According to their data, the evolved strains CH4 and CH5 showed improved OA and LA content (calculated). However, most evolved strains did not show significant FA profile alterations compared to the parental strain (R. toruloides NRRL Y-1091). Mutagenesis by UV has also been used for strain improvement in R. toruloides. Yamada et al. (2017) attempted to improve the lipid-producing ability of R. toruloides NBRC 8766 by UV irradiation and mutant selection under ethanol and H 2 O 2 or cerulenin stress. Two strains with improved lipid production (8766 2-31 M and 8766 3-11C) were isolated. The FA profiles of these two strains showed considerably increased PA content (12.7% and 13.5%) with decreased MA and LA contents. Another study adopted a similar strategy of UV mutagenesis followed by mutant selection under ethanol and H 2 O 2 or LiCl stress (Guo et al. 2019). The research team obtained two strains (R-ZL2 and R-ZL13) with improved lipid production from R. toruloides strain 2.1389. The R-ZL2 and R-ZL13 strains exhibited higher saturated FA profiles with increased PA (23 and 24%) and SA (12 and 10%) content, respectively. These two studies demonstrated that UV mutagenesis combined with stress selection could effectively domesticate R. toruloides to produce FA-rich lipids. Zhang et al. (2016a) obtained mutants of R. toruloides NP11 using atmospheric and room temperature plasma (ARTP) and nitrosoguanidine (NTG) methods. However, they did not apply stress selection after mutagenesis. According to their results, the isolated mutant XR-2 did not show a significant change in the FA profile compared with the parental strain R. toruloides NP11. In contrast, Qi et al. (2020) applied stress selection using inhibitory lignocellulosic hydrolysates after ARTP. Mutant RM18 exhibited improved ALA content (5.5%, 1.6 times more than the parental strain). These studies (Yamada et al. 2017;Guo et al. 2019;Zhang et al. 2016a;Qi et al. 2020) imply that stress selection after physical or chemical mutagenesis is necessary for screening mutants for desired FA production.

Culture conditions affect FA profiles in R. toruloides
Numerous strategies, such as optimizing culture conditions and waste utilization, have been applied to enhance lipid production in R. toruloides to produce economically competitive lipids (Zhao et al. 2022). However, the effects of each strategy on the FA profiles of lipids varied. Therefore, we have consolidated the progress obtained and provided a wide scope in the following sections. The information can help design rational bioprocesses to enhance specific FArich lipid production and assist in tailored lipid production.

Nitrogen, phosphate, and sulfate starvation
Nitrogen (N), phosphate (P), and sulfate (S) starvation can promote lipid accumulation in R. toruloides (Wang et al. 2018;Wu et al. 2010Wu et al. , 2011, and nitrogen limitation is the most adapted method. Different strategies affect lipid synthesis via different pathways, resulting in different metabolic fluxes. Therefore, the N, P, and S concentrations also affected the FA profiles. Ye et al. (2021) found that, when using sucrose as the sole carbon source and adjusting the C/N ratio to over 80 with ammonium sulfate or ammonium nitrate in the culture media, the FA contents of PA and ALA significantly increased in cells. Increasing the C/P ratio with glucose as a carbon source led to an increase in the OA content from 50 to 67% (Wu et al. 2010). In a study of S-limitation (Wu et al. 2011), lipids produced under increased C/S conditions favored saturated FAs (MA, PA, and SA) up to 60%. According to these studies, the strategies used to trigger lipid production should focus on enhancing the specific FA content.

Carbohydrates
Carbohydrates significantly affect FA profiles by activating key metabolic genes involved in gluconeogenesis, the glyoxylate cycle, and the tricarboxylic acid cycle Patel et al. 2015). Most microbes possess a unique carbon catabolite mechanism. Microbes metabolize sugars sequentially because glucose represses the utilization of other sugars Patel et al. 2015). Therefore, strategies using a primary carbon substrate or a defined ratio of carbon substrates have been reported for R. toruloides. Wiebe et al. (2012) demonstrated that R. toruloides could utilize C5 (arabinose and xylose) and C6 (glucose) sugars for lipid production. The lipids produced using arabinose and xylose as the sole carbon sources showed increased PA and LA contents compared to glucose, indicating that C5 (xylose and arabinose) sugars significantly changed the FA profiles. However, the FA profiles of the lipids produced using mixed sugars of glucose, arabinose, and xylose did not change significantly compared to those produced using glucose alone. Lignocellulosic biomass hydrolysate is a promising sustainable feedstock for microbial lipid production. After pretreatment and saccharification of lignocellulosic biomass, the complex links among lignin, cellulose, and hemicellulose are broken down, thus releasing fermentable sugars, such as glucose, xylose, or arabinose (Saini et al. 2022b). Osorio-González et al. (2019) demonstrated that R. toruloides strains had higher PA and SA contents when using C5 hydrolysate than C6 hydrolysate. Therefore, selecting carbohydrates and monitoring the carbohydrate content in the media during fermentation is vital for designing a rational fermentation process for desired FA production.

Crude substrates
In addition to using a single carbon source or defined carbon source, non-monomeric substrates containing abundant carbon sources from industrial, agricultural, municipal solid, and biomass wastes have been extensively studied and employed for lipid production in R. toruloides owing to their low cost (Abeln and Chuck 2021;Zhao et al. 2022).
Crude glycerol (CG) is a widely available product in the biodiesel industry. Approximately 10% (w/w) of glycerol is generated in every batch of biodiesel and is considered a waste (Lee et al. 2014;Uprety et al. 2016). Uprety et al. (2017 indicated that R. toruloides cells grown in media containing CG contained nearly 40% saturated FA (PA and SA), and a reduction in OA (from 60 to 47% of TFA) was observed compared to cells grown with pure glycerol. CG usually consists of glycerol and other impurities, such as glycerol, soap, methanol, ash, FA methyl esters (FAME), and salt.  reported that the decrease in OA in cells resulted from various impurities, particularly soap. Therefore, impurities in CG should be noted for tailored lipid production.
R. toruloides can utilize volatile FAs (VFAs), such as acetic acid (C2), propionic acid (C3), and butyric acid (C4) (Gao et al. 2017) for lipid production. VFAs can be obtained through the degradation of organic waste biomass by anaerobic digestion (Park et al. 2018a;Llamas et al. 2020), and thus VFAs can be exploited as cheap carbon sources for lipid production in yeasts (Park et al. 2018a, b;Llamas et al. 2020). However, different VFAs caused different FA changes in R. toruloides. R. toruloides grown in media with acetic acid (C2) as the sole carbon source generates high OA-lipids (nearly 50% of TFA) (Huang et al. 2016;Krikigianni et al. 2022). When propionic acid (C3) was used as the sole carbon source, the lipids had a high content of saturated FAs (65.9% of TFA) and a low MUFA content of 16% (Krikigianni et al. 2022). Notably, when propionic acid C3 was used, odd-chain FAs (margaric acid, C17:0, and heptadecenoic acid, C17:1) were generated, accounting for 30.8% of TFA. When fermented brewer's spent grain contained mixed VFAs, the production of odd-chain FAs (ODFAs) was not observed, indicating that VFA composition influenced the FA profiles in R. toruloides. ODFAs are unusual and rare FAs that are found in natural sources. They are valuable FAs because they are associated with several health benefits, such as regulating allergies, psoriasis, and autoimmune disorders, and reducing the risk of metabolic disorders (Dąbrowski and Konopka 2022). Therefore, using VFAs as a carbon source for lipid production in R. toruloides could alter the native FA profile and produce unusual FA.
Plant-derived essential oils (EOs) can also alter the FA profiles of oleaginous microbes Uprety et al. 2022). EOs are volatile and hydrophobic compounds mainly composed of terpenic hydrocarbons and oxygenated derivatives. EOs can impact the carbon flux toward a specific type of FAs inside microbes. Hence, EOs have been applied to alter the FA profile of R. toruloides . The EOs from clove-, cinnamon-, pine-, and orange-supplemented growth media for R. toruloides significantly increased the saturated FA content to 29%, 15%, 14%, and 36% of TFA, respectively. Furthermore, the EOs from orange also elevated PA content to 41% of TFA . Therefore, EOs are good inducers of FA production in R. toruloides.

Trace metal salts
Trace metal salts, such as zinc sulfate, copper sulfate, ferric chloride, manganese sulfate, and magnesium sulfate have also been commonly used as micronutrients in growth media to promote microbial growth (Saini et al. 2022a, b). These micronutrients serve as co-factors for several enzymes involved in lipid synthesis and desaturation pathways, and affect their catalytic activities (Singh et al. 2016;Ma et al. 2009;Romero et al. 2018). Furthermore, these metals may alter the catalytic activities of desaturase enzymes, resulting in alterations in the FA composition of lipids (Cai et al. 2020). Metal supplementation with wood hydrolysate as a carbon source causes the production of unusual PUFAs in R. toruloides (Saini et al. 2022a, b). When cells were grown in media supplemented with Mg 2+ , R. toruloides cells produced eicosatrienoic acid (ETA, C20:3) (0.03% of TFA). Similarly, an unusual DHA (0.05% of TFA) was produced by R. toruloides in media supplemented with Cu 2+ . In addition, adding Zn 2+ , Fe 2+ , and Cu 2+ increased the content of GLA, which is also an unusual FA in R. toruloides, from 0.07% to 0.22%, 0.2%, and 0.12% of TFA, respectively. EA, DHA, and GLA are essential FA in humans; hence, they are highvalue compounds. These results indicated that the addition of specific trace metals induced the production of highvalue PUFAs in R. toruloides.

Oxygen
Oxygen availability is a critical parameter, especially during high cell density and large-scale fermentation, because of the unequal oxygen distribution in the microenvironment of the vessels, which affects cell growth. Oxygen is also a key molecule involved in FA desaturation. Fads introduce C-C double bonds into acyl chains by reducing oxygen (Sperling et al. 2003). Because these are oxygendependent reactions, low oxygen availability results in a higher degree of lipid saturation (Abeln and Chuck 2021). In addition, oxygen levels affect the activation of crucial genes, such as S. cerevisiae OLE1 in the FA saturation pathway through transcriptional factors (Romero et al. 2018;Burr et al. 2016). Therefore, oxygen is essential for the desaturation of FA by Fas. Choi et al. (1982) observed that the degree of unsaturation was significantly affected by the specific oxygen uptake rate of R. toruloides. Parreira et al. (2015) also confirmed that the PUFAs fraction increased as oxygen availability increased. These results indicated that oxygen availability is associated with PUFA synthesis in R. toruloides. Furthermore, the oxygen-activation pathways of Fas should be confirmed in R. toruloides.

Temperature
The growth temperature greatly influences various life processes in organisms, including gene transcription, metabolic activity, cell growth, nutrient absorption, cell survival, energy production, and membrane fluidity (Price and Sowers 2004;Fonseca et al. 2019). Microbial cells can modulate membrane fluidity to adapt to environmental temperature. Temperature influences membrane fluidity; decreasing temperature results in reduced viscosity and a more rigid membrane, whereas increasing temperature leads to increased viscosity and a looser membrane (Fonseca et al. 2019;Renne and Kroon 2018). Hence, microbial cells change their FA composition in the membrane lipids to maintain membrane fluidity. An increase in unsaturated FAs content reduces the melting point and increases membrane fluidity (Chen et al. 2022). On the other hand, an increase in saturated FAs content elevates their melting point and decreases membrane fluidity (Mejía-Barajas et al. 2018).
A growth temperature of 30°C is typically adopted for lipid production in R. toruloides. However, higher and lower temperatures significantly influence the FA profile of R. toruloides. The FA profile of the lipids in R. toruloides cells grown at 37°C increased the content of saturated FAs (PA and SA) (30% from 17%), which led to a higher saturated level than that at 30°C . Notably, the increase in saturated FA at 37°C may result from reduced RtFad2 desaturase activity ). These reports suggest that temperature affects Fad activity and alters membrane FA composition, which may help cells maintain membrane fluidity. In contrast, when R. toruloides cells were grown at 15°C, the content of PUFAs, including LA and ALA, increased significantly compared to that at 25°C. The LA and ALA contents increased from 22 to 35% and from 8 to 21%, respectively (He et al. 2015). He et al. (2015) also demonstrated that the mRNA transcription level of a putative FA desaturase was five-fold higher in cells grown at 15°C than in cells grown at 25°C. In Y. lipolytica, the transcriptional factor Mga2, encoded by YALI0B12342g, may be associated with the upregulation of FAD2 at low temperatures (Tezaki et al. 2017). Therefore, elucidating the mechanism of low-temperature-induced FAD2 expression is necessary for R. toruloides to produce PUFAs.

pH
The commonly set pH for lipid production in R. toruloides is 5-6 (Li et al. 2006;Yang et al. 2014;Zhang et al. 2016b;Zeng et al. 2017). However, different medium pH values influence the FA profiles at different levels (Hall and Ratledge 1977). Similar scenarios were first observed in R. toruloides IIP-30 (formerly known as R. glutinis IIP-30) in the pH range of 3-6 (Johnson et al. 1992). Dias et al. (2016) also reported that medium pH values slightly altered FA profiles in R. toruloides. A slight MUFA percentage (OA) increased in the pH range 4.0-5.5. The PUFA percentages (LA and ALA) decreased at pH 4.5-5.5. Although the alteration in FA profiles did not show strong pH dependency, a specific pH range could assist in specific FA-rich lipid production based on their studies.

Light
Rhodotorula toruloides is a promising strain for carotenoid production. The carotenoid and lipid biosynthesis pathways are related because acetyl-CoA molecules are shared as precursors (Bruder et al. 2019). As a critical photo-inducer, light can promote carotenoid production by increasing microbial growth and activity of enzymes essential for carotenoid biosynthesis (Frengova and Beshkova 2009). In addition, photo-induced reactive oxygen species (ROS) may be associated with the activation of carotenoid synthesis pathways through the transcription factor bZIP in R. toruloides (Lin et al. 2017). Pham et al. (2020) demonstrated that R. toruloides cells grown under light conditions showed higher POA (two-fold) and ALA (four-fold) contents of TFA compared to those grown under dark conditions. Their results indicate that light could be a helpful factor for controlling FA profiles in R. toruloides.

Electro-fermentation
Electro-fermentation (EF) is a promising technique for improving the performance of bioprocesses that has been applied to increase lipid production in R. toruloides (Arbter et al. 2019). EF with the redox mediator Neutral Red (NR) caused a significant shift in the FA profile from OA to PA, resulting in a highly saturated FA composition (from 37 to 50%) in R. toruloides (Arbter et al. 2019). EF with NR is an excellent technique for producing highly saturated lipids as an alternative source of plant lipids such as coconut oil. In addition, EF with different media and strains may be applicable to R. toruloides.

Conclusion
Numerous review articles have summarized the different strategies for maximizing lipid production in R. toruloides. However, few have summarized and compared the impact of each method on FA composition. Thus, this mini-review describes strategies and advanced progress for controlling the FA composition in R. toruloudes through metabolic engineering. In addition, this mini-review also summarized the parameters of culture conditions such as substrate utilization and bioprocess control that have impacts on FA profile in R. toruloides. R. toruloides exhibits flexibility in producing lipids that are rich in saturated FA, MUFA, PUFA, and even unusual FAs. Therefore, R. toruloides is a promising platform for producing specific FA lipids. Further studies are necessary to reveal the mechanisms underlying the described strategies and findings to precisely control FA synthesis and bioprocess in R. toruloides for specific FA lipid production.
Author contributions WCC, HK and KF discussed the content and direction. WCC wrote the main manuscript text and prepared figures and table.
Funding Open access funding provided by Osaka University. This study was supported in part by the NEDO Project (Project code: P20011) of the Ministry of Economy, Trade, and Industry (METI), Japan, and by the COI-NEXT Program (Grant number: JPMJPF2106) of Japan Science and Technology (JST).

Data availability
The data supporting the findings of this mini-review are available in the references cited within this article.

Competing interests
The authors declare no competing interests.

Conflict of interest
The authors declare that they have no conflicts of interest.
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:// creat iveco mmons. org/ licen ses/ by/4. 0/.