The Production of Lipids Using 5-Hydorxymethy Furfural Tolerant Rhodotorula graminis Grown on the Hydrolyzates of Steam Pretreated Softwoods

Acid catalyzed (SO2) steam pretreated softwoods inevitably contain furans such as 5-hydorxymethy furfural (HMF) and furfural, which are derived from the respective degradation of component hexoses and pentoses. As these materials are known to be inhibitory to fermentation, six oleaginous yeasts were grown on corn steep liquor (CSL) medium containing HMF (0.2%) and furfural (0.1%) to assess their resistance to possible inhibition and its possible influence on lipid production. R. graminis showed the highest tolerance to HMF (0.2%) and furfural (0.1%) when they were added individually to the CSL medium. However, when both HMF (0.2%) and furfural (0.1%) were added together, this inhibited the growth of R. graminis. Subsequent evaporation of the CSL medium successfully removed furfural from the CSL medium and increased the sugar concentration. However, the residual concentration of HMF (0.4%) still inhibited R. graminis growth. To try to improve HMF tolerance, R. graminis was slowly acclimatized in medium containing HMF (0.4%) and was eventually able to produce 1.8 g/L of lipids after four days of growth in the HMF containing medium. This was close to the same amount of lipid produced as when R. graminis was grown in the CSL medium without HMF and furfural. This indicated that an acclimatization strategy is a promising way to enhance lipids production when R. graminis is grown on the hydrolyzates of SO2-catalyzed steam pretreated lignocellulosic substrates.


Introduction
The production of biodiesel and jet fuel from lignocellulosic materials is one way of reducing the world's dependence on fossil fuels [1,2]. Currently, bio-and-renewable diesels are mainly produced from animal and plant lipids such as tallow, used cooking oil, and soy, rape, and palm oils [3]. Biodiesel is predominantly produced from triacylglycerols (TAGs), and of the various microorganisms that produce TAG, the oleaginous yeasts have been shown to accumulate as much as 50% of the cell dry weight as intracellular lipids [4,5]. Although lignocellulosic materials have the potential to act as the substrates for growing oleaginous yeasts, the recalcitrance of biomass means that typically, some form of pretreatment is required prior to enzymatic hydrolysis of cellulose and hemicellulose [6,7]. SO 2 -catalyzed steam pretreatment is a promising pretreatment process due to the limited use of chemicals, relatively low levels of energy, and the recovery of cellulose and hemicellulose in a fermentable form [6]. After the SO 2 -catalyzed steam pretreatment, lignocellulosic substrates are recovered as water insoluble fractions, while a part of cellulose and hemicellulose, was depolymerized into oligomeric or monomeric sugars that dissolved in water soluble fractions [6,8]. In addition to monomeric sugars, the water soluble fractions contain 5-hydorxymethy furfural (HMF) and furfural, which was dehydrated from hexoses and pentoses, respectively [8]. It was reported that furfural and HMF inhibited the growth of microorganisms in a manner of decreasing enzymatic activities such as alcohol dehydrogenase, aldehyde dehydrogenase, and pyruvate dehydrogenase [9]. In addition, furfural was reported to induce the accumulation of reactive oxygen species in cells which includes damage to mitochondrial and vacuole membranes, the actin cytoskeleton, and nuclear chromatin [10]. Although the production of lipids using oleaginous yeasts from AFEX and SPORL pretreated lignocellulosic biomass has been reported [11][12][13], the use of the hydrolyzates of SO 2 -catalyzed steam pretreated lignocellulosic biomass for the production of lipids by oleaginous yeasts have not been reported. In this study, to produce biodiesel from the hydrolyzates of SO 2 -catalyzed steam pretreated lignocellulosic biomass using oleaginous yeast, the effect of HMF and furfural on the growth of six oleaginous yeasts were compared. Of the six strains used in this study, Rhodotorula graminis, which had higher tolerance to the inhibitors compared with other strains was selected and the tolerance of R. graminis to HMF (0.4%) was attained using acclimation. To evaluate the availability of the HMF (0.4%) tolerant R. graminis, the productivity of lipids by the strain in the presence of HMF (0.4%) was examined.

Yeast Screening
For preculture, a loopful of cells from an isolated colony on YM agar plate was aseptically transferred to 50 mL of YM medium (3 g/L yeast extract; 5 g/L hipolypepton; 3 g/L malt extract; 10 g/L glycerol) in a 300-mL Erlenmeyer flask. The culture was incubated at 26 • C with shaking at 100 rpm until OD 600 was reached to 10. For oleaginous yeasts screening, 50 mL of corn steep liquor (CSL) medium (35 g/L glucose; 15 g/L xylose; CSL 8 g/L) with or without 5-hydroxymetyl furfural (HMF, 0.2%) and furfural (0.1%) were prepared in a 100-mL Erlenmeyer flask. The aliquots of preculture was inoculated in the CSL medium and incubated at 26 • C with shaking at 100 rpm. Initial OD 600 of 0.1 was used for all fermentation experiments. All flask culture experiments were performed in triplicate.

Analytical Methods
Biomass was measured as the optical density at 600 nm (OD 600 ) using a UV mini-1240 spectrophotometer (Shimadzu, Japan). HMF and furfural contents in the medium was measured by HPLC (JASCO LC-2000 plus series equipped with UV-2075 UV-vis detector) using an inert ODS-3 reverse phase column (4.6 mm × 150 mm L, GL sciences Inc. Japan). The chromatographic zones were visualized under UV light (254 nm). The mobile phase was methanol/water (2/1, v/v) with a flow rate of 1.0 mL/min. Cell dry weight and total lipids in the cell were measured according to the reported methods [14].

Acclimation of R. Graminis to HMF(0.4%)
Preculture of R. graminis incubated in the YM medium was inoculated into 50 mL of the CSL medium with HMF (0.4%) in a 300-mL Erlenmeyer flask. The culture was incubated at 26 • C with shaking at 100 rpm and the OD 600 of the medium was measured.

Screening of Oleaginous Yeasts for the Production of Lipids in the Presence of HMF and Furfural
The growth rates of six oleaginous yeasts, T. oleaginosus, R. graminis, two L. starkeyi, L. spencer-martinsiae, and Y. lipolytica, which were reported as the promising strains for producing lipids from the pretreated lignocellulosic biomass [5], were compared to determine the strain that is adequate for the production of lipids from the hydrolyzates of SO 2 -catalyzed steam pretreated lignocellulosic biomass. It was known that the presence of inhibitors (HMF, furfural), monomeric sugar concentrations, and sugar compositions affected the production of lipids from the pretreated lignocellulosic biomass [5]. For the inhibitors, when lignocellulosic biomass is pretreated under acidic condition, cellulose and hemicelluloses are hydrolyzed into monomeric sugars such as glucose and xylose and further dehydrates into HMF and furfural, respectively [15]. The produced HMF and furfural decrease the growth of microorganisms in a manner of enzyme activity inhibition [9] and the generation of active oxygen [10]. When beetle-killed lodgepole pine was pretreated by SO 2 -catalyzed steam pretreatment at medium severity, the hydryzates contained 0.2% of HMF and 0.1% of furfural, respectively [8]. The HMF concentration was increased to 0.3% at high severity that decreased the fermentability of Saccharomyces cerevisiae, while furfural concentration was almost 0.1% even the severities of SO 2 -catalyzed steam pretreatment were increased [8]. Other factors affecting the growth of oleaginous yeasts were the concentration and compositions of sugars [5,16]. It was reported that when Douglas-fir was SO 2 -catalyzed steam pretreated, the water soluble fraction contained about 3.5% of glucose and 1.5% of xylose, respectively [17]. In this study, to observe the applicability of the oleaginous yeasts for the production of lipids from the hydrolyzates of SO 2 -catalyzed steam pretreated lignocellulosic biomass at medium severity, the effects of HMF (0.2%) and furfural (0.1%) on the growth of six oleaginous yeasts were compared in the corn steep liquor (CSL) medium supplemented with glucose (3.5%) and xylose (1.5%). CSL, a by-product of the corn wet milling industry, was used for the nitrogen source in the medium due to its containing a rich complement of important nutrient to support robust microorganism growth, and it is thought to be inexpensive [18,19]. The effects of HMF and furfural on the growth of the oleaginous yeasts were different. The OD 600 of T. oleaginosus with HMF (0.2%) was decreased by 23% compared with that without inhibitors after seven days. In addition, the presence of furfural (0.1%) completely inhibited the growth of T. oleaginosus ( Figure 1). Of the six oleaginous yeasts used in this study, it seemed that R. graminis was the most promising strain for the production of lipids from the hydrolyzates of the SO 2 -catalyzed steam pretreated lignocelluloses, because the growth of R. graminis was not affected by HMF (0.2%). The OD 600 of R. graminis in the presence of HMF (0.2%) reached to 11.7 after seven days, which was the same OD 600 as without the inhibitors ( Figure 2). The growth of R. graminis was reported to be decreased by 35% when R. graminis was inoculated in the minimal medium in the presence of HMF (0.15%) [20], while the growth of R. graminis with HMF (0.2%) was not decreased in the CSL medium ( Figure 2). This suggests that the CSL medium was preferable for R. graminis growth even in the presence of HMF. Another reason that R. graminis was thought to be preferable for producing lipids from the hydrolyzates of SO 2 -catalyzed steam pretreated lignocellulosic biomass is that although induction time was observed, the OD 600 of R. graminis in the presence of furfural (0.1%) was almost same as the OD 600 without inhibitors ( Figure 2). The OD 600 of R. graminis in the presence of either HMF (0.2%) or furfural (0.1%) was not decreased after seven-days incubation, but R. graminis did not grow when both furfural (0.1%) and HMF (0.2%) were added in the CSL medium ( Figure 2). Two L. starkeyi and L. spencer-martinsiae species showed similar growth patterns ( Figure 3). The presence of HMF (0.2%) or furfural (0.1%) in the CSL medium increased the induction period and the presence of both HMF and furfural in the CSL medium inhibited the growth of the strains. It was reported that L. starkeyi could grow in up to 0.2 g/L furfural and 0.63 g/L HMF, respectively, in chemically defined medium [21]. As three Lipomyces species grew even in the presence of 0.1% furfural and 0.2% HMF, respectively (Figure 3), it is likely that CSL medium is preferable for the growth of Lipomyces species in the presence of furfural and HMF. Of the six oleaginous yeasts, only Y. lypolytica grew in the presence of both HMF (0.2%) and furfural (0.1%) after induction period ( Figure 4). However, the OD 600 of Y. lypolytica even without the inhibitors was lower than those of other five strains. Thus, R. graminis was thought to be the promising strain for the production of lipids using the hydryzates from SO 2 -catalyzed steam pretreated lignocellulosic biomass, because the growth of the strain in the presence of either HMF (0.2%) or furfural (0.1%) was almost same as the OD 600 without the inhibitors after seven-days incubation ( Figure 2). However, R. graminis could not grow when both HMF (0.2%) and furfural (0.1%) were presented in the medium. The removal of both HMF and furfural or the removal of furfural from the medium was required to produce lipids from the hydryzates of SO 2 -catalyzed steam pretreated lignocellulosic biomass.
Sustainability 2020, 12, x FOR PEER REVIEW 4 of 9 medium [21]. As three Lipomyces species grew even in the presence of 0.1% furfural and 0.2% HMF, respectively (Figure 3), it is likely that CSL medium is preferable for the growth of Lipomyces species in the presence of furfural and HMF. Of the six oleaginous yeasts, only Y. lypolytica grew in the presence of both HMF (0.2%) and furfural (0.1%) after induction period ( Figure 4). However, the OD600 of Y. lypolytica even without the inhibitors was lower than those of other five strains. Thus, R. graminis was thought to be the promising strain for the production of lipids using the hydryzates from SO2-catalyzed steam pretreated lignocellulosic biomass, because the growth of the strain in the presence of either HMF (0.2%) or furfural (0.1%) was almost same as the OD600 without the inhibitors after seven-days incubation ( Figure 2). However, R. graminis could not grow when both HMF (0.2%) and furfural (0.1%) were presented in the medium. The removal of both HMF and furfural or the removal of furfural from the medium was required to produce lipids from the hydryzates of SO2catalyzed steam pretreated lignocellulosic biomass.   Sustainability 2020, 12, x FOR PEER REVIEW 4 of 9 medium [21]. As three Lipomyces species grew even in the presence of 0.1% furfural and 0.2% HMF, respectively (Figure 3), it is likely that CSL medium is preferable for the growth of Lipomyces species in the presence of furfural and HMF. Of the six oleaginous yeasts, only Y. lypolytica grew in the presence of both HMF (0.2%) and furfural (0.1%) after induction period ( Figure 4). However, the OD600 of Y. lypolytica even without the inhibitors was lower than those of other five strains. Thus, R. graminis was thought to be the promising strain for the production of lipids using the hydryzates from SO2-catalyzed steam pretreated lignocellulosic biomass, because the growth of the strain in the presence of either HMF (0.2%) or furfural (0.1%) was almost same as the OD600 without the inhibitors after seven-days incubation ( Figure 2). However, R. graminis could not grow when both HMF (0.2%) and furfural (0.1%) were presented in the medium. The removal of both HMF and furfural or the removal of furfural from the medium was required to produce lipids from the hydryzates of SO2catalyzed steam pretreated lignocellulosic biomass.

Removal of Furfural from CSL Medium
Furfural (0.1%) was more toxic for R. graminis compared with HMF (0.2%), because furfural (0.1%) increased induction time (Figure 2), while the growth rate of R. graminis with HMF (0.2%) was the same as that without inhibitors. Thus, it was thought that removing furfural was more efficient compared with removing HMF (0.2%) from the medium. To remove furfural from the CSL medium, rotary-evaporator was used, as furfural is known to form azeotrope with water. It was reported that rotary-evaporation could remove 96% of furfural dissolved in water in the presence of other inhibitors (HMF, acetic acid, formic acid, and levulinic acid) when the initial concentration of each inhibitor was 0.1 g/L, while HMF was concentrated almost proportionately with concentration factor [22]. In this study, although the concentration and the kinds of inhibitors were different from the previous study [22], furfural was successfully removed by two-times concentration using rotaryevaporation. In addition, increasing of the monomeric sugars concentrations in the medium was possible by rotary-evaporation. Thus, it was shown that the evaporation of the hydryzates from the SO2-catalyzed steam pretreatment would be a beneficial way to remove furfural and increase sugar concentrations.

Acclimating of R. graminis for HMF(0.4%)
The removal of furfural and increasing of monomeric sugar concentration were attained by the concentration of the CSL medium by rotary-evaporation. However, HMF concentration was simultaneously increased from 0.2% to 0.4%, as HMF is not volatile compound. Although R. graminis had a higher HMF tolerance compared with the other strains (Figures 1-4), R. graminis did not grow in the medium in the presence of HMF (0.4%). This result was the same as the previous research in which wild type of R. graminis could grow in up to 0.35% of HMF [20]. It was assumed that acclimation was one of the best ways to produce lipids from the medium containing 0.4% of HMF, because the acclimation was effectively used for other oleaginous yeasts, L. tetrasporoous and Y. lipolytica, which increased the productivity of lipids due to the decreasing of induction time from SPORL enzymatic hydrolyzates containing HMF (0.06%) [12]. In addition, when S. cerevisiae was inoculated in a medium containing HMF and furfural, it started to grow when these inhibitors were reduced to their less toxic alcohol derivatives during the growth lag time [10]. Thus, it was considered that obtaining HMF (0.4%) tolerant R. graminis could be possible with longer incubation time in the CSL medium with HMF (0.4%) due to acclimation. As the result, after 140 h incubation of R. graminis

Removal of Furfural from CSL Medium
Furfural (0.1%) was more toxic for R. graminis compared with HMF (0.2%), because furfural (0.1%) increased induction time (Figure 2), while the growth rate of R. graminis with HMF (0.2%) was the same as that without inhibitors. Thus, it was thought that removing furfural was more efficient compared with removing HMF (0.2%) from the medium. To remove furfural from the CSL medium, rotary-evaporator was used, as furfural is known to form azeotrope with water. It was reported that rotary-evaporation could remove 96% of furfural dissolved in water in the presence of other inhibitors (HMF, acetic acid, formic acid, and levulinic acid) when the initial concentration of each inhibitor was 0.1 g/L, while HMF was concentrated almost proportionately with concentration factor [22]. In this study, although the concentration and the kinds of inhibitors were different from the previous study [22], furfural was successfully removed by two-times concentration using rotary-evaporation. In addition, increasing of the monomeric sugars concentrations in the medium was possible by rotary-evaporation. Thus, it was shown that the evaporation of the hydryzates from the SO 2 -catalyzed steam pretreatment would be a beneficial way to remove furfural and increase sugar concentrations.

Acclimating of R. graminis for HMF(0.4%)
The removal of furfural and increasing of monomeric sugar concentration were attained by the concentration of the CSL medium by rotary-evaporation. However, HMF concentration was simultaneously increased from 0.2% to 0.4%, as HMF is not volatile compound. Although R. graminis had a higher HMF tolerance compared with the other strains (Figures 1-4), R. graminis did not grow in the medium in the presence of HMF (0.4%). This result was the same as the previous research in which wild type of R. graminis could grow in up to 0.35% of HMF [20]. It was assumed that acclimation was one of the best ways to produce lipids from the medium containing 0.4% of HMF, because the acclimation was effectively used for other oleaginous yeasts, L. tetrasporoous and Y. lipolytica, which increased the productivity of lipids due to the decreasing of induction time from SPORL enzymatic hydrolyzates containing HMF (0.06%) [12]. In addition, when S. cerevisiae was inoculated in a medium containing HMF and furfural, it started to grow when these inhibitors were reduced to their less toxic alcohol derivatives during the growth lag time [10]. Thus, it was considered that obtaining HMF (0.4%) tolerant R. graminis could be possible with longer incubation time in the CSL medium with HMF (0.4%) due to acclimation. As the result, after 140 h incubation of R. graminis with HMF (0.4%), increasing of OD 600 was observed. The aliquot of the medium was inoculated on the YMA with HMF (0.4%) and then the strain was incubated in the CSL medium with HMF (0.4%). Although the OD 600 of wild type R. graminis was 1.3 times higher than that of HMF (0.4%) tolerant R. graminis after seven-days incubation, the OD 600 of the two strains were almost same until four-days incubation ( Figure 5). The previous study showed that the growth rate of R. graminis was decreased by about 70% in the presence of HMF (0.35%) [20]. In this study, the growth rate of the HMF (0.4%) tolerant R. graminis was similar to that of wild type ( Figure 5), showing that acclimation was a useful way for obtaining oleaginous yeast even at higher HMF concentration (0.4%). Further study is required to reveal the mechanisms for the acclimation of R. graminis to HMF (0.4%), but one possible mechanism would be the increasing of alcohol dehydrogenase activity that reduces HMF to 5-hydroxymethylfurfuryl alcohol [23]. with HMF (0.4%), increasing of OD600 was observed. The aliquot of the medium was inoculated on the YMA with HMF (0.4%) and then the strain was incubated in the CSL medium with HMF (0.4%).
Although the OD600 of wild type R. graminis was 1.3 times higher than that of HMF (0.4%) tolerant R. graminis after seven-days incubation, the OD600 of the two strains were almost same until four-days incubation ( Figure 5). The previous study showed that the growth rate of R. graminis was decreased by about 70% in the presence of HMF (0.35%) [20]. In this study, the growth rate of the HMF (0.4%) tolerant R. graminis was similar to that of wild type ( Figure 5), showing that acclimation was a useful way for obtaining oleaginous yeast even at higher HMF concentration (0.4%). Further study is required to reveal the mechanisms for the acclimation of R. graminis to HMF (0.4%), but one possible mechanism would be the increasing of alcohol dehydrogenase activity that reduces HMF to 5hydroxymethylfurfuryl alcohol [23].

Production of Lipids Using 2 Times Concentrated CSL Medium
The total lipids of acclimated R. graminis and wild type R. graminis were compared after fourdays incubation in the two-times concentrated CSL medium with HMF (0.4%) and CSL medium without HMF, respectively ( Figure 5). The acclimated R. graminis produced 1.8 g/L of lipids after four days when the strain was incubated in two-times concentrated the CSL medium by rotaryevaporation which contained 0.4% of HMF, while wild type R. graminis produced 2.0 g/L of lipids after four days in the CSL medium without HMF and furfural. Previous study showed that the lipids concentration produced by R. graminis was increased by 1.4 times when 0.15% of HMF was added into the medium [20]. In this study, the increasing of lipids concentrations was not observed when 0.4% of HMF was added in the CSL medium. One of the possible reasons could be that the higher HMF concentration used in this study negatively affected the biosynthetic pathway of lipids. However, considering the result that wild type R. graminis did not grow in the CSL medium containing HMF (0.4%) after four-days incubation (data not shown), the use of HMF (0.4%) tolerant R. graminis was shown to be effective for the production of lipids from the hydrolyzates of the SO2catalyzed steam pretreatment of lignocellulosic biomass.

Production of Lipids Using 2 Times Concentrated CSL Medium
The total lipids of acclimated R. graminis and wild type R. graminis were compared after four-days incubation in the two-times concentrated CSL medium with HMF (0.4%) and CSL medium without HMF, respectively ( Figure 5). The acclimated R. graminis produced 1.8 g/L of lipids after four days when the strain was incubated in two-times concentrated the CSL medium by rotary-evaporation which contained 0.4% of HMF, while wild type R. graminis produced 2.0 g/L of lipids after four days in the CSL medium without HMF and furfural. Previous study showed that the lipids concentration produced by R. graminis was increased by 1.4 times when 0.15% of HMF was added into the medium [20]. In this study, the increasing of lipids concentrations was not observed when 0.4% of HMF was added in the CSL medium. One of the possible reasons could be that the higher HMF concentration used in this study negatively affected the biosynthetic pathway of lipids. However, considering the result that wild type R. graminis did not grow in the CSL medium containing HMF (0.4%) after four-days incubation (data not shown), the use of HMF (0.4%) tolerant R. graminis was shown to be effective for the production of lipids from the hydrolyzates of the SO 2 -catalyzed steam pretreatment of lignocellulosic biomass.

Conclusions
Six oleaginous yeasts were compared to produce lipids from the hydrolyzates of SO 2 -catalyzed steam pretreated lignocellulosic biomass using the CSL medium with HMF (0.2%) and furfural (0.1%). The presence of both HMF (0.2%) and furfural (0.1%), which was inevitably contained in the hydrolyzates of SO 2 -catalyzed steam pretreated lignocellulosic biomass, decreased the growth of the oleaginous yeasts showing that the removal of the inhibitors from the medium was required. Of the six oleaginous strains used in this study, R. graminis was the most promising strain for higher tolerant to the inhibitors. The removal of furfural from the CSL medium was conducted by rotary-evaporation because furfural was more toxic for R. graminis.
However, HMF concentration was increased from 0.2% to 0.4% by rotary-evaporation that resulted in the growth inhibition of R. graminis. As the result of acclimation, HMF (0.4%) tolerance R. graminis was obtained, and this strain produced lipids as much as that produced by wild type of R. graminis without inhibitors. This result suggests that the acclimation is an effective way to obtain HMF tolerant R. graminis and this strain is applicable for the production of lipids from the hydrolyzates of the SO 2 -catalyzed steam pretreated lignocellulosic biomass. Further study is required to observe the productivity of lipids by the HMF (0.4%) tolerance R. graminis from the concentrated hydrolyzates of SO 2 -catalyzed steam pretreated lignocellullosic biomass such as lodgepole pine and Douglas-fir.