Microbial oil production by Mortierella isabellina from corn stover under different pretreatments

Mixed culture of Trichoderma reesei and Aspergillus niger was employed to accomplish on-site cellulase production where cellulases were applied directly to the enzymatic hydrolysis of pretreated corn stover. We comprehensively compared the five pretreatments including sodium hydroxide, steam explosion, aqueous ammonia, lime and diluted sulfuric acid in the enzymatic hydrolysis and then in the whole processes from corn stover to single cell oil (SCO). The results were not completely but roughly the same. However, it is conclusive that sodium hydroxide pretreatment was the best one. The process with sodium hydroxide pretreatment could produce 23.5 g dry cell biomass harboring 13.7 g single cell oil (SCO) from 146.2 g corn stover. The total yields of cell biomass and SCO were 0.161 and 0.094 g g−1 corn stover, respectively. This process was proven the most efficient. Through this work, we established efficient SCO production process from corn stover.


Introduction
As the second largest producer of corn stover in the world, 1 China produces up to 0.3 billion metric tons of corn stover annually. However, the corn stover is oen burned, causing serious environmental problem. Nonetheless, it is an important resource that could be utilized to produce valuable products in an environment-friendly and sustainable way. This is good for China's environment protection and rural economy.
Microbial lipids, regarded as single cell oils (SCO), are more renewable than plant oils or animal fats. They have much shorter production period, don't compete with food production and are much easier to scale up. Many microorganisms can accumulate high content of lipids, among which Mortierella isabellina is able to accumulate more than 80% lipids of its dry cell biomass. [2][3][4] M. isabellina's excellence in SCO production and its good tolerance towards the inhibitors derived from the pretreatment of lignocelluloses suggest that M. isabellina is a strong candidate for SCO production from cheap agricultural wastes. 5,6 Before enzymatic hydrolysis and subsequent fermentation, agricultural residues should be pretreated to crack down the recalcitrance for better enzymatic digestibility and more efficient conversion. Pretreatment is one of the most expensive single unit operation in lignocellulose-based bioreneries. 7,8 It is not only one of the major contributor to high production costs, but also affects the subsequent processes owing to the inhibitors form during pretreatments. Hitherto many different pretreatments have been developed, including physical, chemical, physico-chemical, and biological methods, etc 9, 10 . Different pretreatments have different mechanisms to break the recalcitrant lignocellulosic biomass. However, those mechanisms could be summarized as follows: removing lignin, reducing crystallinity of cellulose, hydrolyzing hemicellulose, disrupting linkage between hemicellulose and lignin, increasing specic surface areas of substrates and so on. Furthermore, different lignocellulosic biomasses have different physico-chemical characteristics, 8 therefore a suitable pretreatment technology should be established for a specic lignocellulose material to increase efficiency of lignocellulose-based bioprocesses.
The biggest challenge to the SCO production from agricultural wastes is its high production cost, making it economically uncompetitive. The priority for the researches is how to reduce the costs by increasing the bioconversion efficiency. In order to achieve it, we sought efficient pretreatment and adopted the concept of on-site cellulase production since it is well known that pretreatment, enzymatic hydrolysis and cellulase production are the most expensive single unit operations in lignocellulose-based bioreneries. 7 In this work, the leading pretreatments including steam explosion pretreatment, sodium hydroxide pretreatment, diluted sulfuric acid pretreatment, lime pretreatment and aqueous ammonia pretreatment were applied and compared in the bioprocess from corn stover to SCO to look for the most suitable pretreatment for corn stover in the context of SCO production. Though some of them had been compared in the enzymatic hydrolysis of corn stover, 11 it's the rst we compared them in the SCO production by M. isabellina from cron stover. On-site cellulase production by the mixed culture of Trichoderma reesei and Aspergillus niger 6,12,13 was carried out to realize cellulase autarky due to its advantages as described by Culbertson et al. 14 such as cost-saving. Another purpose of this work is to establish an efficient process from corn stover to SCO.

Corn stover and pretreatments
Corn stover was collected from Kaifeng area, Henan Province, in the autumn of 2014. It was dried in the sun and cut to a suitable size (5-10 cm) before storage at room temperature. Prior to the pretreatments except steam explosion, corn stover was milled using a laboratory hammer mill to a size less than 2 mm. The materials aer pretreatments were washed using tap water and distilled water until neutral pH. Then they were collected and stored at 4 C. The compositions of the corn stover materials before and aer pretreatments are listed in Table 1.
2.1.1 Sodium hydroxide pretreatment. Corn stover with the size of less than 2 mm was soaked in 2% NaOH aqueous solution with a solid-to-liquid ratio of 1 : 8 (g:mL), mixed thoroughly and kept at 120 C for 30 min. 11, 15 2.1.2 Steam explosion pretreatment. Corn stover with the size of 5-10 cm was steam exploded under the following conditions: temperature 200 C, pressure 1.6 MPa, pressure maintained time 7 min, substrate loading 100 g (dry material). 12, 13 2.1.3 Aqueous ammonia pretreatment. Corn stover with the size of less than 2 mm was mixed with 10% NH 4 OH aqueous solution with a solid-to-liquid ratio of 1 : 10 (g:mL) and incubated in a shaker at 26 C for 24 h. Then it was ltered and washed with tap water 3 times. Subsequently, it was further treated with 0.3 mol L À1 HCl at 100-108 C for 1 h. 11 2.1.4 Lime pretreatment. Corn stover with the size of less than 2 mm was mixed with calcium hydroxide with a ratio of 1 : 2.5 (g:g). Then it was added with distilled water with a water versus corn stover ratio of 10 : 1 (g:g) and kept at 120 C for 4 h under non-oxidative condition. 11 2.1.5 Diluted sulfuric acid pretreatment. Corn stover with the size of less than 2 mm was mixed with 1.5% H 2 SO 4 aqueous solution with a solid-to-liquid ratio of 1 : 10 (g:mL) and maintained at 106-108 C for 6 h. 11 2.2 Mixed culture of T. reesei and A. niger and relevant media T. reesei and A. niger were pre-cultured in the seed medium for 36 h and 48 h, respectively. Mixed culture of T. reesei and A. niger was performed in 250 mL Erlenmeyer asks with a working volume of 50 mL (90% the fermentation medium for the mixed culture and 10% the total inoculums of T. reesei and A. niger). The delay time of A. niger inoculation and the inoculum ratio of T. reesei versus A. niger were 48 h and 5 : 1, respectively. 12,13 The seed medium consisted of 10 g L À1 glucose, 1 g L À1 peptone, 5 mL Mandels nutrients salts solution, 2.5 mL 1 M citrate buffer, 0.05 mL Mandels trace elements solution, 0.1 g L À1 Tween 80.
The media were autoclaved at 121 C for 20 or 30 min (20 min for the medium without pretreated corn stover and 30 min for the media with pretreated corn stover).

Enzymatic hydrolysis of pretreated corn stover
The enzymatic hydrolysis of pretreated corn stover was performed in 250 mL Erlenmeyer asks with a working volume 50 mL containing 2.5 mL 1 M citrate buffer solution (pH 4.8), 100 g L À1 pretreated corn stover (dry material), 25 FPIU g À1 glucan the cellulase from the mixed culture described above which was harvested aer 5 d fermentation, and a supplementary amount of distilled water to make up 50 mL. Aer the crude enzyme in the fermentation broth collected by centrifugation (3000 rpm, 5 min) were added and mixed evenly, asks were incubated in an orbital shaker (140 rpm) at 50 C for 48 h. Samples were taken at 2, 4, 6, 12, 24, 48 h during enzymatic hydrolysis for further analyses.

Fermentation of enzymatic hydrolysates by M. isabellina
M. isabellina spores on the PDA slants were suspended with sterile water and counted by Fisher Scientic hemacytometer. The fermentation medium (50 mL) in 250 mL Erlenmeyer asks was inoculated with 1 Â 10 8 spores and incubated at 28 C with a shaking of 160 rpm. Samples were taken once a day to monitor the growth of M. isabellina and the lipid accumulation during the fermentation.
Fermentation medium was composed of the enzymatic hydrolysate, 0.5 g L À1 (NH 4 ) 2 SO 4 , 0.5 g L À1 yeast extract, 0.5 g L À1 , 7 g L À1 KH 2 PO 4 , 2 g L À1 Na 2 HPO 4 , 1.5 g L À1 The initial pH was adjusted to 5.5. The enzymatic hydrolysate was autoclaved at 121 C for 30 min and the mixture of salts was sterilized by ltering through 0.22 mm membrane (Millipore, MA, USA). Then they were blended before use.

Analytical methods
2.5.1 Component analysis of raw and pretreated corn stover. The components of raw and pretreated corn stover, which were glucan, xylan and lignin, were analyzed using the method recommended by National Renewable Energy Laboratory (NREL), i.e. the NREL/TP-510-42618 procedure.
2.5.2 Enzymatic activities of cellulases. Filter paper activity (FPA) was measured using the standard method recommended by the International Union of Pure and Applied Chemistry (IUPAC) and expressed in international unit (IU). 16 One IU of FPA (FPIU) was dened as the mount of enzyme required for releasing 1 mmol reducing sugars in 1 min.
b-Glucosidase activity (BGA) was assayed using the method adapted from the standard FPA determination. 16 The substrate was rNPG (r-nitrophenyl-b-D-1,4-glucopiranoside) (Sigma-Aldrich, St. Louis, MO, USA) instead of cellobiose. The reaction mixture consisted of 0.1 mL the enzyme solution diluted properly and 0.9 mL 5 mM rNPG solution. The reaction was conducted at 50 C with a shaking of 80 rpm for 10 min and stopped with 2 mL 1 M Na 2 CO 3 . The product r-nitrophenol was measured spectrophotometrically at a wavelength of 400 nm. One IU of BGA was dened as the amount of enzyme which can produce 1 mmol r-nitrophenol in 1 min.
The measurements of cellobiohydrolase activity (CBA) and endoglucanase activity (EGA) were similar to the FPA assay 16 except for the substrates, the reaction time, and the denition of enzymatic activity unit. The substrates for assaying CBA and EGA were microcrystalline cellulose PH101 and carboxymethyl cellulose, respectively, both of which were purchased from Sigma-Aldrich, St. Louis, MO, USA. They were used in terms of 1% (w/v) suspensions. The reaction time was 30 min. One Unit (1 U) of CBA and EGA was dened as the amount of enzyme required for producing 1 mg reducing sugars in 1 h.
The assay of xylanase activity (XLA) was similar to the standard method for FPA determination 16 except for the substrate and the reaction time, which were birch xylan (Sigma-Aldrich, St. Louis, MO, USA) and 30 min, respectively. Birch xylan was used in terms of 1% (w/v) suspension. One IU of XLA was dened as the mount of enzyme required for releasing 1 mmol reducing sugars in 1 min.
2.5.3 Analysis of sugars. Sugars were measured by highperformance liquid chromatography (HPLC) using an Agilent 1100 system equipped with Bio-Rad Aminex HPX-87H (300 mm Â 7.8 mm). Deionized and degassed water was used as the mobile phase at a ow rate of 0.6 mL min À1 . The column temperature was xed at 55 C. The eluate was detected by a refractive index detector.
2.5.4 Determination of microbial lipids. The microbial cells ($30 mg biomass in 1 mL water solution), 4.5 mL methanol, and 1 mL tridecanoic acid (internal standard) were added into a tube. The tube was capped and vortexed for 30 s. Add 0.2 mL 12 M H 2 SO 4 into the tube and mix it using vortex. The tube was kept in 85 C water bath for 15 min for esterication. Immediately the tube was cooled down with tap water. Add 2 mL H 2 O and mix using vortex. Add 2 mL hexane and mix again for fatty acid methyl esters (FAME) extraction. The hexane layer was moved to vial for analysis. The FAMEs were determined using capillary gas chromatography. A SP-2560, 100 m Â 0.25 mm Â 0.20 mm capillary column (Supelco) was installed on the Hewlett Packard 5890 gas chromatograph instrument equipped with a Hewlett Packard 3396 Series II integrator and a 7673 controller, a ame ionization detector, and split injection (Agilent Technologies Inc., Santa Clara, CA, USA).

Pretreatment of corn stover
Corn stover was pretreated using steam explosion pretreatment, sodium hydroxide pretreatment, diluted sulfuric acid pretreatment, lime pretreatment and aqueous ammonia pretreatment, respectively. The compositions before and aer pretreatments are listed in Table 1 These pretreatments are widely used in lignocellulose-based bioreneries. 9,10 In subsequent work, therefore, we compared these pretreatments in SCO production from corn stover by M. isabellina.

Cellulase production from pretreated corn stover
The pretreated materials, steam exploded corn stover (SECS), SHPCS, DSAPCS, LPCS and AAPCS, were used as substrate and inducer for cellulase production by the mixed culture of T. reesei and A. niger as reported by the previous work. 12,13 The results of cellulase production are shown in Fig. 1. It was found that SHPCS had the best induction for cellulase production, leading to the highest FPA which was 3.85 AE 0.23 FPIU mL À1 . This suggests that sodium hydroxide pretreatment changed corn stover to let it be the best inducer. Besides, SHPCS induced the second highest BGA, 1.08 AE 0.17 IU mL À1 , which was just slightly lower than the biggest BGA from SECS, 1.16 AE 0.15 IU mL À1 . This inconspicuous difference could not prove anything. Therefore, we chose the SHPCS as the best substrate and inducer for cellulase production by the mixed culture of T. reesei and A. niger.
These cellulases were applied to the subsequent experiments of enzymatic hydrolysis respectively, i.e. the cellulase induced by SHPCS was used in the enzymatic hydrolysis of SHPCS, the cellulase induced by SECS was used in the enzymatic hydrolysis of SECS, and so forth. The rst reason is that we hope the whole process from corn stover to SCO just uses one kind of pretreatment. More than one pretreatments employed in the process signies more single unit operations, more complicacy, and higher production cost. The second reason is that use of lignocellulosic biomass as substrate to induce cellulase production has an increased enzymatic hydrolysis specicity for the substrate itself than others. 14,17,18 This is so-called on-site enzyme production for lignocellulose-based bioreneries. 6,14

Enzymatic hydrolysis of pretreated corn stover
The pretreated materials, SECS, SHPCS, DSAPCS, LPCS and AAPCS, were enzymatically hydrolyzed by the cellulase induced by and produced from themselves. The results are shown in Fig. 2. The enzymatic hydrolysis of SHPCS by the cellulase produced from SHPCS was found to be the most efficient, achieving the highest yield of 87.9 AE 2.5%. It produced 61.5 g L À1 glucose and 14.8 g L À1 xylose aer 48 h enzymatic hydrolysis from 100 g L À1 dry SHPCS. The second highest yield, 81.5 AE 1.2%, was from the enzymatic hydrolysis of SECS by the cellulase produced from SECS. This process produced 47.6 g L À1 glucose and 6.5 g L À1 xylose aer 48 h enzymatic hydrolysis from 100 g L À1 dry SECS. Next in line was the enzymatic hydrolysis of AAPCS by the cellulase produced from AAPCS which led to a yield of 75.2 AE 2.7%, i.e. 44.8 g L À1 glucose and 8.3 g L À1 xylose from 100 g L À1 dry AAPCS. The penultimate, the enzymatic hydrolysis of LPCS by the cellulase produced from LPCS, gave rise to a yield of 66.5 AE 1.8%, producing 39.2 g L À1 glucose and 14.7 g L À1 xylose from 100 g L À1 dry LPCS. The process with the lowest yield of 48.0 AE 2.1% was the enzymatic hydrolysis of DSAPCS by the cellulase produced from DSAPCS, which released 31.4 g L À1 glucose and 4.1 g L À1 xylose from 100 g L À1 dry DSAPCS.
Sodium hydroxide pretreatment was proven here to be the most suitable for pretreating corn stover using enzymatic  hydrolysis yield as criterion, which led to the highest yield. This is probably because this pretreatment had the highest lignin removal, 11 making the substrate most enzymatically digestible. 9,10,19 Diluted sulfuric acid pretreatment mainly depolymerized hemicelluloses but had no effect on lignin, thus resulting in the highest lignin content. Therefore, it makes sense that it caused the lowest yield. Then we compared the effect of pretreatment on the fermentation process for SCO production.

Fermentation of enzymatic hydrolysate for SCO production
The enzymatic hydrolysates were fermented by M. isabellina to produce SCO. Only glucose and xylose were taken into consideration because other sugars were not fermentable or trace. The fermentation results are shown in Fig. 3.
The hydrolysates from the differently pretreated corn stover contained different concentrations of glucose and xylose which led to different time or productivities for the fermentation processes (Fig. 3a-e). It's in the same pattern that during the all the fermentation processes, M. isabellina fermented glucose rst and xylose since then. This is normal because of glucose effect. As shown in Fig. 1a, M. isabellina spent 12 d depleting 61.5 g L À1 glucose and 14.8 g L À1 in the enzymatic hydrolysate of SHPCS and it produced 23.5 AE 1.4 g L À1 cell biomass (dry cell weight, the same below) encompassing 13.7 AE 1.0 g L À1 SCO. In the fermentation of the SECS enzymatic hydrolysate containing 47.6 g L À1 glucose and 6.5 g L À1 xylose (Fig. 3b), M. isabellina produced 15.7 AE 1.1 g L À1 cell biomass encompassing 8.9 AE Fig. 3 Time courses of SCO production by M. isabellina in the enzymatic hydrolysates of SHPCS (sodium hydroxide pretreated corn stover) (a) containing 61.5 g L À1 glucose and 14.8 g L À1 xylose, SECS (steam exploded corn stover) (b) containing 47.6 g L À1 glucose and 6.5 g L À1 xylose, AAPCS (aqueous ammonia pretreated corn stover) (c) containing 44.8 g L À1 glucose and 8.3 g L À1 xylose, LPCS (lime pretreated corn stover) (d) containing 39.2 g L À1 glucose and 14.7 g L À1 xylose, and DSAPCS (diluted sulfuric acid pretreated corn stover) (e) containing 31.4 g L À1 glucose and 4.1 g L À1 xylose. The SCO production was performed in 250 mL Erlenmeyer flasks with 50 mL of the fermentation medium inoculated with 1 Â 10 8 spores and incubated in a rotary shaker at 160 rpm and 28 C. Data shown are average values of triplicate samples and error bars are standard deviations. Fig. 4 Comparison of different processes from corn stover to SCO involving different pretreatments including sodium hydroxide pretreatment (a), steam explosion pretreatment (b), aqueous ammonia pretreatment (c), lime pretreatment (d) and diluted sulfuric acid pretreatment (e). SHPCS: sodium hydroxide pretreated corn stover; SECS: steam exploded corn stover; AAPCS: aqueous ammonia pretreated corn stover; DSAPCS: diluted sulfuric acid pretreated corn stover; LPCS: lime pretreated corn stover. Detailed information is given here to facilitate the comparison. 0.6 g L À1 SCO aer 9 d fermentation. Fig. 3c shows the fermentation of the AAPCS enzymatic hydrolysate containing 44.8 g L À1 glucose and 8.3 g L À1 xylose and M. isabellina produced 14.4 AE 0.6 g L À1 cell biomass encompassing 7.5 AE 1.0 g L À1 SCO aer 9 d fermentation. Fig. 3d presents the fermentation of the LPCS enzymatic hydrolysate involving 39.2 g L À1 glucose and 14.7 g L À1 xylose, which resulted in 13.9 AE 0.8 g L À1 cell biomass and 7.8 AE 0.4 g L À1 SCO aer 10 d fermentation. In the fermentation of DSAPCS enzymatic hydrolysate that had 31.4 g L À1 glucose and 4.1 g L À1 xylose (Fig. 3e), M. isabellina accumulated 10.3 AE 0.9 g L À1 cell biomass harboring 5.6 AE 0.4 g L À1 SCO aer 9 d fermentation.
It was inconclusive here that sodium hydroxide pretreatment was the best approach just based on the fermentation of hydrolysates for SCO production because those fermentation processes were incomparable when the fermentable sugars were in different concentrations. Therefore, we would compare the entire processes from corn stover to SCO and the relevant details to nd out the most suitable pretreatment among the pretreatment methods we adopted in the work for SCO production from corn stover by M. isabellina. Fig. 4 illustrates the technological process ows from corn stover to SCO and the details of each process are given in Table  2. As shown in Fig. 4, we developed different processes from corn stover to SCO involving different pretreatments. It was found that sodium hydroxide pretreatment used the smallest amount of starting material, 146.2 g corn stover, but produced the largest amounts of cell biomass and SCO, 23.5 g and 13.7 g respectively. This means Process a had the highest yields and productivities, as listed in Table 2. Thus, it's conclusive that sodium hydroxide pretreatment was the best suited for SCO production from corn stover by M. isabellina among the ve pretreatment methods compared in this work. The superiority of sodium hydroxide pretreatment over the others was ascribed to its most potent capability of lignin removal and less formation of inhibitors. 11,20 Lignin hinders the contact between cellulase and its substrate cellulose, as well as the contact between hemicellulases and its substrate hemicellulose. 21 In addition, the presence of lignin causes nonproductive enzyme adsorption. 19,22 The inhibitors derived from pretreatments are inhibitory or even toxic to downstream enzymes and organisms. 9,10 Thus, it is reasonable that sodium hydroxide pretreatment performed the best. These ve pretreatments are widely used in pretreating lignocellulosic materials, 6,8,9,11,15,[23][24][25] indicating that these are leading pretreatment technologies. Sodium hydroxide pretreatment, therefore, should be the desirable option when carrying out the production of lignocellulosic SCO.

Comparison of the processes from corn stover to SCO
In the comparison of the processes from corn stover to SCO, it was found that the proper order from good to bad was sodium hydroxide pretreatment, steam explosion pretreatment, lime pretreatment, aqueous ammonia pretreatment, and diluted sulfuric acid pretreatment. This order is not exactly the same as that concluded from the enzymatic hydrolysis experiment (Fig. 2). Though aqueous ammonia pretreatment was obviously better than lime pretreatment in the enzymatic hydrolysis, the result was contrary in the fermentation for SCO production where the yields of Process d with lime pretreatment were higher than those of Process c with aqueous ammonia pretreatment. This suggests that enzymatic hydrolysis yield could not adopted as single criterion to evaluate pretreatment and the effects on subsequent bioprocesses, although the orders were grosso modo the same-sodium hydroxide pretreatment the rst and diluted sulfuric acid pretreatment the last. Hence, we should take both enzymatic hydrolysis and fermentation into consideration and, more importantly, emphasize the whole process from raw material to product. The contrary results of lime pretreatment and aqueous ammonia pretreatment in enzymatic hydrolysis and in the whole process may be caused by the different recovery rates of pretreatments and different contents of glucan and xylan aer pretreatment. Moreover, the inhibitors derived from pretreatment and their inuence on subsequent bioprocesses 23,26 may fuzz up the results. Though we washed pretreated corn stover until neutral pH, there were still some remanent inhibitors (data not shown) which could be released gradually during enzymatic hydrolysis and cause detrimental impact on fermentation. Further work should be conducted to elucidate the phenomenon. This work highlights sodium hydroxide pretreatment as the best among the ve pretreatments. Moreover, the efficient process from corn stover to SCO was established successfully here. In the future, we will compare it with novel or more advanced pretreatments [27][28][29] in a holistic way, and improve it for more perfect pretreatment of corn stover. 30,31

Conclusions
Mixed culture of T. reesei and A. niger was employed to achieve cellulase autarky, which was proven more efficient than other cellulases in the enzymatic saccharication of corn stover. The ve pretreatments inducing sodium hydroxide pretreatment, steam explosion pretreatment, aqueous ammonia pretreatment, lime pretreatment and diluted sulfuric acid pretreatment were compared comprehensively in the SCO production by M. isabellina from corn stover. Among the ve pretreatments, sodium hydroxide pretreatment was proved to be the best one, leading to the highest yield of SCO and the highest process efficiency from corn stover.

Conflicts of interest
There are no conicts of interest to declare.