Utilization of Sugarcane Bagasse as a Substrate for Lipid Production by Aurantiochytrium sp.

which are useful in the fields of health foods, cosmetics, fine chemicals, and biofuels. Lignocellulosic biomass, which is abundant and cheap, is a promising feedstock for producing cheaper bulk and high-value-added products using Aurantiochytrium sp. However, the steam explosion of lignocellulosic biomass for efficient enzymatic saccharification generates substances that inhibit the growth of microorganisms. In this study, the inhibitory activities of these by-products on the growth and lipid production of Aurantiochytrium sp. were investigated. Aurantiochytrium sp. was found to be highly sensitive to furfural and vanillin and moderately sensitive to 5-hydroxymethylfurfural and syringaldehyde. Washing steam-exploded bagasse with water, followed by activated charcoal treatment, significantly reduced furfural, which was a major inhibitory component in the saccharified solution.

have been reported to inhibit the growth and material production of microorganisms 15,16 . Therefore, it is necessary to evaluate the sensitivity of biocatalysts to such by-products and establish an efficient procedure for removing harmful compounds. This study investigated the inhibitory effect of substances generated by the steam explosion of sugarcane bagasse on the growth and lipid production of Aurantiochytrium limacinum SR21 14 and examined pretreatment methods to reduce the inhibitory substances.

Preparation of saccharification liquid from sugarcane
bagasse Sugarcane bagasse containing 50 water was blasted once by the steam explosion at 3 MPa for 5 min. Sugarcane bagasse soaked in 0.5 sulfuric acid was blasted under the same conditions. Steam-exploded bagasse or steam-exploded bagasse washed once with water was treated with 20 mg-protein/g bagasse of the commercial enzyme complex of cellulase and hemicellulase at pH 5.5, 50 and 120 rpm for 7 days.

Component analysis of saccharification liquid
The glucose concentration in the saccharification liquid was measured using a glucose GO assay kit Sigma-Aldrich and a VICTOR NIVO multimode microplate reader PerkinElmer, Waltham, MA, USA . Furans, phenolic compounds, and organic acids were measured using a highperformance liquid chromatography system 1260 Infinity, Agilent Technology, Santa Clara, CA, USA equipped with a silica-based reversed-phase column YMC Carotenoid, 4.6 250 mm, YMC, Kyoto, Japan . A mobile phase consisting 0.05 phosphoric acid/acetonitrile 85:15, v/v was used at a flow rate of 1 mL/min. All compounds were detected by a diode array detector at 210 nm and identified by comparing their retention times with those of the reference chemicals.

Evaluation of cell growth
The growth of A. limacinum SR21 in the culture fluid was determined by measuring the optical density at 590 nm OD 590 . The half-maximal inhibitory concentration IC 50 of each chemical was defined as the concentration at which OD 590 was half that of the control.

Lipid extraction and fatty acid analysis
The cells of the strain SR21 from 1 mL of culture were washed with distilled water and vigorously vortexed with 1 mL of tert-butylmethylether/methanol 2:1, v/v and glass beads 0.5 mm using a bead crusher µT-12 Taitech, Aichi, Japan . Five hundred microliters of distilled water and 50 µg of arachidic acid as an internal standard were added, and the mixed solutions were vortexed for a short time. The organic layer was collected by centrifugation at 12,000 g for 10 min and transferred into a screw cap test tube. Methanolysis of fatty acids was performed by adding 10 methanolic hydroxide and heating at 60 for 90 min. Fatty acid methyl esters FAMEs were extracted with n-hexane. A gas chromatography system GC2025, Shimadzu, Kyoto, Japan equipped with a capillary column TC-70, GL Science, Tokyo, Japan was used to analyze the FAMEs. FAMEs were identified by comparing their retention times with those of 37-Compounds Spelco, Bellefonte, PA, USA and quantified by comparing the peak area with that of arachidic acid.

Statical analysis
Student s t-test was used to determine the differences in experimental values between the control medium and the medium containing inhibitors. p 0.05 was considered significant.

Growth and fatty acid production of A. limacinum
SR21 cultured in the medium containing saccharified solutions The saccharified solutions, Sol-exp, Sol-exp-W, and Solexp-S, were prepared by enzymatic saccharification of bagasse blasted by steam explosion, bagasse blasted and washed with water, and bagasse pretreated with 0.5 sulfuric acid followed by steam explosion Fig. 1A , respectively. Sol-exp, Sol-exp-W, and Sol-exp-S contained 98.8 g/ L, 110.5 g/L, and 80.5 g/L of glucose, respectively. A. limacinum SR21 was cultured in the GPY medium containing each saccharified solution Fig. 1B . The cell growth in the Sol-exp medium containing glucose at 50 and 80 g/L were OD 590 values of 22.9 4.6 and 22.7 1.6, which were 24.6 and 24.8 lower than those in the control GPY medium, respectively. The total fatty acid production also decreased by 3.0 1.9 and 2.6 0.8 g/L, which were 28.6 and 41.1 lower than control, respectively. In addition, the cell growth and total fatty acid production in the Sol-exp-S medium containing glucose at 50 g/L was OD 590 of 0.6 0.01 and 0.03 0.02 g/L, which were 97.9 and 99.2 lower than those of control medium, respectively. These results indicate the presence of some substances inhibiting the growth and fatty acid production of strain SR21 in Sol-exp and Sol-exp-S. The particularly strong inhibitory activity of Sol-exp-S suggests that the formation of these inhibitors may be enhanced in the presence of sulfuric acid. In contrast, strain SR21 cultured in a medium containing Sol-exp-W did not significantly reduce cell growth and fatty acid production. Hence, the growth inhibitors contained in the steam-exploded bagasse were presumed to be watersoluble.

The sensitivity of A. limacinum SR21 against sub-
stances generated by the heat treatment of lignocellulosic biomass The inhibitory substances generated by the pretreatment of lignocellulosic biomass 17 19 were examined to determine the sensitivity of strain SR21. As shown in Fig. 2A, the presence of 60 mM 5-hydroxymethylfurfural completely inhibited growth and fatty acid production. A significant inhibitory effect was also observed with the addition of furfural at concentrations higher than 15 mM Fig. 2A . In phenol derivatives from lignin Fig. 2B , syringaldehyde at 30 mM showed significant inhibition, and vanillin at con-  centrations higher than 15 mM completely inhibited growth. Vanillic acid and coniferyl alcohol had moderate effects. As for organic acids Fig. 2C , acetic acid and formic acid are also major by-products generated by the heat treatment of lignocellulosic biomass and have been reported to reduce the growth of Saccharomyces cerevisiae 20 and Trichoderma reesei 21 . However, these organic acids did not show any inhibitory activity against SR21. The addition of acetic acid tended to increase fatty acid production, which suggested to be assimilated, at least in part, as a nutrient.

Identification of the toxic substances in Sol-exp and
Sol-exp-S HPLC analysis of Sol-exp and Sol-exp-S revealed three major peaks with retention times consistent with those of furfural, 5-hydroxymethylfurfural, and acetic acid Figs. 3B, 3D . Their concentrations were 4.8, 0.7, and 13.7 mM in Sol-exp and 11.2, 2.9, and 51.8 mM in Sol-exp-S, respectively, whereas these substances were not detected in Solexp-W data not shown . Among these substances, furfural was expected to be the main growth inhibitor according to the sensitivity test of A. limacinum SR21 Fig. 2A . To remove inhibitory substances, activated charcoal was added to Sol-exp and Sol-exp-S at 5 w/v and agitated for 30 min. This resulted in a significant reduction in furfural, 5-hydroxymethylfurfural, and acetic acid Figs. 3C, 3E , whereas 92-94 of glucose remained data not shown . The strain SR21 was cultured in media containing activated charcoal-treated Sol-exp or Sol-exp-S with 80 g/ L-glucose and showed growth levels similar to those cultured in the control GPY medium Fig. 4B . The fatty acid production in the media with the charcoal-treated Sol-exp was 7.1 0.5 g/L, which was 88 higher than that of control. The presence of a substance that promotes fatty acid synthesis in Sol-exp was presumed.

Discussion
Fermentative production of lipids from unutilized biomass is important to ensure the sustainability and economic viability of the production process, and some efforts are underway to utilize biomass, such as non-edible plants 9 , food waste 7, 22 24 , and industrial waste 25 27 for lipid production by Aurantiochytrium sp. The utilization of lignocellulosic biomass for lipid production by Aurantiochytrium sp. is meaningful in expanding the field of lipid product application. In this study, the inhibitory effects of typical by-products generated by the heat, acid, and alkali treatment of lignocellulosic biomass 13, 15 on the growth and fatty acid production of A. limacinum SR21 were investigated. In addition, a procedure to remove substances from the saccharified solution prepared from sug-arcane bagasse was developed.
The dehydration of hexose and pentose in hemicellulose generates furan derivatives such as furfural and 5-hydroxymethylfurfural 28,29 . Furan derivatives are major microbial growth inhibitors in lignocellulosic hydrolysates and have been reported to inhibit glycolysis, induce oxidative stress, damage cell membrane homeostasis, and repress protein translation 30 32 . The strain SR21 was also sensitive to furfural IC 50 7.5 mM and 5-hydroxymethylfurfural IC 50 Fig. 3 HPLC analysis of Sol-exp and Sol-exp-S before and after activated charcoal treatment. Sol-exp and Sol-exp-S were treated with 5 activated charcoal for 30 min. The substances contained in Sol-exp B , activated charcoal-treated Sol-exp C , Solexp-S, and activated charcoal-treated Sol-exp-S were separated on YMC Carotenoid column using a mobile phase of phosphate buffer pH 7.5 / acetonitrile 85:15, v/v at a flow rate of 1 mL/min. All the compounds were detected by monitoring the absorbance at 210 nm and identified by comparing their retention times with those of reference chemicals A . The peaks of acetic acid, 5-hydroxymethylfurfural, and furfural are indicated by an asterisk, a pound sign, and a dagger, respectively. Fig. 2A , and fatty acid production was significantly inhibited by furfural. Furfural is reported to reduce the activity of malic enzyme and citrate lyase, which are key enzymes for fatty acid production in the oleaginous yeast, Trichosporon fermentans 33 . Furfural was speculated to inhibit the fatty acid biosynthesis of the strain SR21 by the same mechanism since these enzymes play important roles in the fatty acid production of Aurantiochytrium sp. 34,35 .

mM
Phenolic compounds generated by degrading lignin in plants under acidic conditions 36 have been reported to inhibit the growth and substance production of fungi and microalgae by harming enzyme activities and homeostasis of the cell membrane 37 . Among the 4-hydroxybenzoic acid and its derivatives tested in this study, chemicals other than 4-hydroxybenzoic acid inhibited growth and lipid production of SR21. Since vanillin showed significant inhibitory activity Fig. 2B , it was necessary to reduce its concentration in the hydrolysate. In addition, the result that the inhibitory activity of vanillin IC 50 8.3 mM is higher than that of vanillic acid IC 50 27 mM , an oxide of vanillin, is consistent with a report that phenols with aldehyde groups show a higher inhibitory effect on the growth of S. cerevisiae than phenols with carboxyl groups 38 .
Organic acids such as acetic acid, formic acid, and levuric acid have been reported to inhibit yeast cell growth by lowering the intracellular pH or inducing the accumulation of reactive oxygen species 39,40 . In contrast, the inhibitory effects of these organic acids were marginal or negligible under the conditions tested in this study Fig. 2C . Moreover, with acetic acid at concentrations higher than 50 mM, an increased fatty acid content was observed. Strain SR21 has been reported to grow and produce lipids using up to 40 g/L acetic acid as the sole carbon source 26,41 . Therefore, acetic acid in the saccharified solution is pre-sumed to positively affect growth and the fatty acid production.
As a result of culturing the strain SR21 in the medium containing the hydrolysates prepared from hydrothermally treated sugarcane bagasse, remarkable growth inhibition was observed in the medium containing Sol-exp and Solexp-S compared with the GPY medium containing glucose at the same concentration, whereas Sol-exp-W showed no growth inhibitory effect Fig. 1B . Furfural, 5-hydroxymethylfurfural, and acetic acid were identified as major byproducts in Sol-exp and Sol-exp-S Figs. 3B, 3D . Considering the sensitivity of SR21 to each compound Fig. 2 , furfural was expected to be the primary inhibitor of cell growth and fatty acid production. Sol-exp-W, which contained only trace amounts of furfural and 5-hydroxymethylfurfural, showed no inhibitory effect. Therefore, washing steam-exploded bagasse with water effectively removes growth-inhibiting substances. In addition, enriching the recovery of inhibitory components in the saccharified solution should also be considered because washing treatment requires wastewater disposal. The by-products in Sol-exp and Sol-exp-S, including furfural, were effectively removed by the activated charcoal treatment Figs. 3C, 3E . Activated charcoal treatment has advantages over other procedures, such as distillation and ion exchange 42 in terms of removal efficiency and energy cost. Furfural absorbed on activated charcoal can be collected using organic solvent 43 and used as a material for lubricating oil and solvent in the chemical industry. When Sol-exp and Sol-exp-S were treated with activated charcoal, cell growth and lipid productivity were equal to or higher than those in the control Fig. 4 . In particular, charcoal-treated Sol-exp induced significantly higher fatty acid productivity, suggesting sugarcane bagasse hydrolysate contains substances that promote not only cell growth 44 but also lipid production, such as plant hormones 45,46 .

Conclusion
The inhibitory effect of substances generated by heat treatment of lignocellulosic biomass on the growth and lipid production of Aurantiochytrium sp. was elucidated for the first time. Furfural was speculated as the primary harmful substance in the saccharified solution of sugarcane bagasse, and efficient lipid production was possible by removing it by water-washing and activated charcoal treatment. These results provide an efficient strategy for preparing a saccharified solution from lignocellulosic biomass suitable for lipid fermentation by Aurantiochytrium sp. and contribute to the development of bioprocesses that convert lignocellulosic biomass to valuable lipids.