Fermentation of hexoses and pentoses from hydrolyzed soybean hull into ethanol and xylitol by Candida guilliermondii BL 13

Abstract

This work investigated the ability of a recently isolated strain of Candida guilliermondii to convert hexoses and pentoses obtained from acid-enzymatic soybean hull hydrolysates into ethanol and, in smaller amounts, into xylitol. Operational conditions and media formulation were optimized concerning ethanol production using experimental designs (Plackett-Burman and Central Composite Design). Results showed that C. guilliermondii BL 13 was capable of growing in non-supplemented, non-detoxified biomass hydrolysates, and the best culture conditions were determined to be 28 °C, pH 5.0, and 10⁹ CFU mL^-1 of inoculum size. Ethanol productivity reached 1.4 g L^-1 h^-1, and maximal yields of 0.41 g g^-1 were obtained, representing 80.4 % of the expected theoretical yields, whereas small amounts of xylitol were also produced. These results suggest that C. guilliermondii BL13 is a potentially useful yeast strain to be applied in second-generation ethanol production from lignocellulosic biomass based on its natural capacity to metabolize C-5 and C-6 sugars.

Keywords:
Lignocellulosic Biomass; Fermentation; Ethanol production; Xylitol production; Candida guilliermondii

INTRODUCTION

Lignocellulosic biomass materials represent an abundant and renewable source of carbohydrates that can be used to produce chemical specialties and biofuels of high-added value through biochemical processes, which is a promising alternative to oil-based products (Martínez, 2012). The production of ethanol from sugary substrates, such as sugarcane in Brazil and maize starch in the USA, is economically sound and has been the main technology to obtain this biofuel. However, there is a large amount of lignocellulosic biomass, consisting of cellulose, hemicellulose, and lignin, which could be used to obtain second-generation ethanol (Kuhad et al., 2011). Second-generation ethanol might contribute to reduce negative environmental impacts of fossil fuels, at the same time avoiding the food versus fuel competition for arable land (Cassman and Liska, 2007). However, in order to achieve these objectives, compatible costs of biomass preparation (i.e., the liberation of sugars from the lignocellulosic matrix), and the use of yeast strains capable of converting both hexoses and pentoses into ethanol, are necessary.

Soybean (Glycine max) is the world´s most cultivated oilseed, with the United States of America and Brazil as its main producers, representing 35 % and 27 % of the global production share, respectively (USDA, 2015). Grain hulls represent the major by-product of the soybean processing industry, representing approximately 8 % to 10 % (mass fraction) of the whole seed (Gnanasambandan and Proctor, 1999). The insoluble carbohydrate fraction of soybean hull cell walls consists of 40 % of cellulose and 27 % of hemicellulose (Cassales et al., 2011). In a typical hydrolysis process, the hemicellulose is depolymerized into pentoses (predominantly xylose), whereas cellulose is mainly broken into glucose (Chen et al., 2011).

Considering that hemicellulose is the second most abundant polymer in lignocellulosic materials, the conversion of biomass to ethanol only becomes economically feasible if both pentoses and hexoses are converted into this alcohol. This conversion, in second-generation ethanol production, depends on the ability of microorganisms to ferment the different sugars resulting from hydrolysis. Pentoses cannot be fermented by wild-type strains of Saccharomyces cerevisiae, the most extensively used microorganism for ethanol production, whereas yeasts such as Scheffersomyces (Pichia), Candida, and Pachysolen have been used for the xylose conversion. However, Scheffersomyces and Pachysolen tannophilus, for example, require a fine control of the supply of dissolved oxygen to produce ethanol or to assimilate nutrients and form biomass, which is generally difficult to control in the fermentation process (Lin et al., 2012; Fu and Peiris, 2008). Alternatively, other alcohols, such as xylitol and butanol, can be produced by some of these yeasts.

Xylitol is a natural five-carbon alcohol that has been highly valued for food and pharmaceutical applications because of its sweetening power, as a dental cavities reducer, and as insulin-independent carbohydrate source for diabetic-patient support (Mussatto et al., 2006). However, the production of xylitol by yeasts is highly depended on several parameters, such as the type of biomass used, the hydrolysis conditions, and fermentation parameters such as pH, substrate concentration, size of inoculum, and medium composition (Granström et al., 2001). Different species of Candida have been used for xylitol production. Candida tropicalis W103, for example, was able to convert xylose into xylitol under aerobic and anaerobic conditions in a synthetic medium (Cheng et al., 2014). In another study, Candida guilliermondii FTI 20037 was studied concerning the influence of the toxic compounds present in brewers spent grain hydrolysate on xylose-to-xylitol bioconversion (Mussatto and Roberto, 2008). The same strain was used to ferment rice straw hydrolysate with a high xylose concentration into xylitol (Mussatto and Roberto, 2004a). However, the literature concerning the use of C. guilliermondii for ethanol production on lignocellulosic hydrolysates as substrates is scarce. In a previous work, we tested C. guillermondii NRRL Y-2075, a culture-collection strain, to ferment an acid-hydrolyzed soybean hull medium, presenting high osmotic pressure (Schirmer-Michel et al., 2008). The strain was able to convert glucose into ethanol, showing reasonable yields (Y _P/S = 0.46 g g^-1), but low productivities concerning this alcohol (Q _P = 0.24 g L^-1 h^-1).

In this context, the present work aimed at evaluating the biotechnological bioconversion of sugars obtained from soybean hull hydrolysate (SHH) into ethanol, using a recently isolated strain of Candida guilliermondii (also known as Meyerozyma guilliermondii) in the fermentation process. This strain was isolated from a natural environment and has never been used in bioprocesses before. In order to optimize ethanol production, experimental designs were developed to define the important nutrients and the best environmental conditions for the fermentation using SHH as substrate. Xylitol production was also evaluated as a by-product of the fermentation, but its production was not optimized. To our knowledge, this is the first report in the literature describing the optimization of fermentation conditions for ethanol production using a C. guilliermondii strain on a lignocellulosic biomass hydrolysate.

MATERIALS AND METHODS

Microorganism and cell maintenance

Candida guilliermondii BL 13 was used in this research. This strain was isolated from environment-discharged piles of rotten rice hulls. This strain was selected because it has shown the ability to metabolize C-5 and C-6 sugars and to produce ethanol. The isolated strain was identified comparing the ITS1 and ITS4 amplicon DNA sequences with GenBank databases (access number JQ425356.1), producing 100 % of homology with Candida guilliermondii species. Stock cultures were maintained on YM agar containing (in g L^-1): yeast extract, 3; malt extract, 3; peptone, 5; glucose, 10; agar, 20. For long-term stock storage, cells were kept frozen at -20 °C in a suspension of 50 % glycerol (volume fraction) and 50 % of cultivation medium.

Soybean hull hydrolysate preparation

The Solae Company kindly supplied the soybean hulls (Esteio, Brazil). The composition of this biomass, including sugar-degradation products such as hydroxymethylfurfural and furfural, are described elsewhere (Cassales et al., 2011). The complete optimized procedure for acid-enzymatic hydrolysis of the soybean hulls was carried out following methodologies described by Hickert et al., (2014). The hulls were milled to a particle size smaller than 1 mm in diameter and the acid hydrolysate (SHH) was obtained by acid diluted hydrolysis of this material in an autoclave (Phoenix, Brazil) at 1 atm and 121 °C, in a solution of 1 % (v/v) concentrated sulfuric acid (98 % v/v), using a solid-liquid proportion (m/m) of 1.0:8.8, and a reaction time of 40 min. This procedure was based on a previously published experimental design (Cassales et al., 2011). This hydrolysate was used in the Plackett-Burman and CCD designs.

After the acid hydrolysis, liquid and solid fractions were separated by filtration and the liquid fraction was vacuum-concentrated at 60 °C in order to increase the final sugar concentration. The pH was adjusted to 5.0 with NaOH. The solid fraction was washed with tap water until neutral pH was reached. Enzymatic hydrolysis of cellulose present in this solid fraction was performed using a cellulolytic enzyme complex produced by Penicillium echinulatum strain S1M29, which was obtained from the mutant strain 9A02S1 (Dillon et al., 2011). The enzymatic hydrolysis was carried out using a solid-liquid ratio (mass fraction, dry matter) of 1:20 in citrate phosphate buffer (pH 4.8) and 15 FPU g^-1 dry matter of enzyme, incubated on an orbital shaker at 120 rpm, 50 ºC for 72 h (Hickert et al., 2014). Both fractions of hydrolysates were mixed, vacuum-concentrated at 70 °C in order to increase the sugar concentration, and autoclaved at 121 ^oC for 30 min. The amount of toxic compounds (inhibitors of microbial growth), formed during hydrolysis in the final soybean hull acid-enzymatic hydrolysate (SH-AEH), was determined by HPLC analysis. The pH was adjusted to 5.0 with solid pellets of NaOH. Both hydrolysates (SHH and SH-AEH) were used in the fermentations without any detoxification (removal of furans and acetic acid), apart from the loss during the final sterilization in the autoclave. Table 1 presents the complete composition of SHH and SH-AEH.

Table 1
Final composition of acid soybean hull hydrolysates (SHH) and acid-enzymatic soybean hull hydrolysates (SH-AEH) used for fermentations.

Component (g L-1)	SHH	SH-AEH
Glucose	17.75 ± 0.10	42.1 ± 0.69
Xylose	18.25 ± 0.22	15.3 ± 0.13
Arabinose	4.95 ± 0.52	5.83 ± 0.32
Furfural	0.02 ± 0.01	0.08 ± 0.02
HMF	0.07 ± 0.01	0.58 ± 0.01
Acid Acetic	1.10 ± 0.08	2.10 ± 0.12
Osmotic pressure (mOsm kg-1)	2.149	2,916

Inocula preparation and fermentation conditions in orbital shaker

Pre-inocula were prepared by seeding one isolated colony of yeast cells from plates into 250 mL Erlenmeyer flasks containing 50 mL of synthetic medium composed of (in g L^-1): yeast extract, 3; malt extract, 3; peptone, 5; glucose, 10. Cultivations were carried out in an orbital shaker (Marconi MA 830, Brazil) at 180 rpm, 28 °C for 24 h and cells were subsequently recovered by centrifugation (3 000 g, 15 min). The cell pellet was washed with sterile distilled water, resuspended in culture medium and inoculated into culture flasks (10 % v/v).

The experiments using SHH as substrate were carried out in 250 mL Erlenmeyer flasks covered using cotton-wool plugs, containing 60 mL of SHH and incubated in an orbital shaker at 180 rpm for 72 h, with temperature varying according to the CCD (23 ºC to 33 ºC). Samples were collected during cultivation to determine biomass, ethanol, glycerol, and residual sugars concentrations in the broth. The experiments using SH-AEH as substrate were carried in 500 mL Erlenmeyer flasks, filled with 120 mL of this hydrolysate, and incubated in an orbital shaker under the best conditions obtained in the CCD (2.5).

Plackett-Burman design

Plackett-Burman (PB) design was used to screen and evaluate the effects of four complex nutrients (peptone, yeast extract, corn steep liquor, and Tween 80) on the bioconversion of SHH into ethanol by C. guilliermondii BL 13. The PB design consisted of randomized experiments of eight fermentations plus three replicates at the central point to evaluate the reproducibility of the experimental procedure. The variables were chosen because they would represent the addition of a nitrogen source (peptone); growth cofactors (yeast extract and corn steep liquor, which is an inexpensive nutrient source); and a source of unsaturated fatty acids (Tween 80) in order to avoid physiological impairments due to oxygen limitations. For each variable, the higher level (+1), the central point, and the lower level (-1) of the components were tested (Table 2). Student’s t-test was performed to determine the significance of each variable. Significant positive effects were considered when the reported p-values were lower than 0.05.

Table 2
Levels of the real and the codified values of independent variables tested in the Plackett-Burman design to test supplementation of SHH.

Independent variables (g L-1)	Range of levels
	Low -1	Central point 0	High 1
Yeast Extract	0	2.0	4.0
Corn Steep Liquor	0	2.0	4.0
Peptone	0	2.5	5.0
Tween 80	0	0.2	0.4

Central composite design

A 2³ full factorial central composite design (CCD), with four replicates at the central point and six axial points resulting in a total of eighteen experiments, was used to investigate the fermentation conditions of temperature, pH, and inoculum size for ethanol production by C. guilliermondii BL 13 on SHH. The coded and uncoded values are presented in Table 3 and were chosen based on the literature (Silva and Roberto, 2001; Cunha-Pereira et al., 2011; Mussatto and Roberto, 2004b). Experimental data were analyzed using a second-order response surface regression procedure.

Table 3
Levels of the real and the codified values of independent variables used in the central composite design to test fermentation on SHH.

Independent variables		Range of levels
	Symbols	-1.68	-1	0	1	1.68
pH	X 1	4.16	4.5	5.0	5.5	5.84
Temperature (°C)	X 2	23	25	28	31	33
Inoculum (Log CFU mL-1)	X 3	5	6	7	8	9

Fermentation using acid-enzymatic soybean hull

In order to improve the conversion of total sugars present in the soybean hull into ethanol, an acid-enzymatic hydrolysate (SH-AEH) was used for fermentations under the conditions that were optimized in the CCD. The experimental procedure was identical to the cultivations described for SSH, except for the medium, which was SH-AEH (composition defined in the section “Soybean hull hydrolysate preparation”). These experiments were conducted in triplicates.

Analytical methods

Glucose, xylose, arabinose, ethanol, and xylitol concentrations were determined by HPLC (Shimadzu, Japan) equipped with a refractive index detector and Bio-Rad HPX-87H column (300 x 7.8 mm) using 5 mM sulfuric acid as eluent at 45 °C, flow rate of 0.6 mL min^-1 and sample volumes of 20 μL. Furfural and HMF were determined by HPLC using an UV detector (at 276 nm) and a Nucleosil C18 column (250 x 4.6 mm) at 24 ^oC, using acetonitrile-water (2:8) containing 10 g L^-1 acetic acid as eluent, and a flow rate of 1.1 mL min^-1.

Biomass was estimated as viable cells, using CFU (colony forming units, expressed as CFU mL^-1) plated in yeast morphology agar (YMA) medium. Results were statistically evaluated using the Statistica 7.0 software.

Calculation of fermentation parameters

Ethanol conversion yield (Y _P/S ) was calculated as the ratio between the highest ethanol concentration produced and the sugars consumed (glucose and xylose; difference between the initial and residual sugar concentrations). Ethanol productivity (Q _P ) was defined as the ratio between the highest ethanol concentration and the respective time of fermentation, which was at 12 h of fermentation in all experiments.

RESULTS AND DISCUSSION

Analysis of supplementation requirements using PB design

The components yeast extract, corn steep liquor, peptone, and Tween 80 were screened at the confidence level of 90 % based on their effects and the results for ethanol productivity and yields are shown in Table 4. All coefficients exhibited high p-values for ethanol yield and productivity (Table 5), implying that they had no significant effects on ethanol production, productivity, and yield. Similar results were observed in the fermentation of brewers spent grain hydrolysate by Candida guilliermondii FTI 20037 aiming at conversion of xylose to xylitol (Mussatto and Roberto, 2008). These authors reported that the hydrolysate supplementation with calcium chloride, ammonium sulfate, and rice bran extract did not show any significant effect on xylose-to-xylitol conversion.

Table 4
Plackett-Burman design matrix showing the effects of SHH supplementation with yeast extract, corn steep liquor, peptone, and Tween 80 on ethanol productivity (Q_p) and yields (Y_P/S).

Assays	Yeast Extract (g L-1)	Corn Steep Liquor (g L-1)	Peptone (g L-1)	Tween 80 (g L-1)	YP/S (g g-1)	QP (g L-1 h-1)
1	+1 (4)	-1 (0)	-1 (0)	+1 (0.4)	0.35	0.13
2	+1 (4)	+1 (4)	-1 (0)	-1 (0)	0.32	0.13
3	+1 (4)	+1 (4)	+1 (5)	-1 (0)	0.41	0.16
4	-1 (0)	+1 (4)	+1 (5)	+1 (0.4)	0.28	0.11
5	+1 (4)	-1 (0)	+1 (5)	+1 (0.4)	0.47	0.15
6	-1 (0)	+1 (4)	-1 (0)	+1 (0.4)	0.31	0.12
7	-1 (0)	-1 (0)	+1 (5)	-1 (0)	0.32	0.11
8	-1 (0)	-1 (0)	-1 (0)	-1 (0)	0.44	0.15
9	0 (2)	0 (2)	0 (2.5)	0 (0.2)	0.37	0.15
10	0 (2)	0 (2)	0 (2.5)	0 (0.2)	0.39	0.14
11	0 (2)	0 (2)	0 (2.5)	0 (0.2)	0.39	0.15

Table 5
Plackett-Burman effect estimated values of ethanol productivity (Q_P) and yields (Y_P/S).

Response variable	Factor	Effect	p-value
QP	Mean/Interaction	0.1373	0.0000
	Yeast Extract	0.0144	0.2887
	Corn Steep Liquor	-0.0036	0.7830
	Peptone	-0.0015	0.9087
	Tween 80	-0.0110	0.4081
YP/S	Mean/Interaction	0.3697	0.0000
	Yeast Extract	0.0483	0.2734
	Corn Steep Liquor	-0.0675	0.1429
	Peptone	0.0125	0.7657
	Tween 80	-0.0181	0.6669

The results obtained in this work suggest that essential nutrients are already present in the hydrolysate as this substrate is rich in nitrogen and has several minerals in its chemical composition (Cassales et al., 2011), fulfilling the requirements of the yeast metabolism. Thus, the addition of these nutrients could be eliminated for the subsequent experiments, at the same time showing that the use of soybean hull hydrolysate is an interesting substrate for fermentation processes because it does not require expensive supplementations.

Optimization of fermentation conditions

A central composite design (Table 6) was carried out to evaluate the effect of temperature, pH, and inoculum size on ethanol productivity and yield during the fermentation of SHH by C. guilliermondii BL 13. Four assays at the central point were carried out to estimate the experimental error needed for the analysis of variance. The linear, quadratic, and interaction effects for ethanol productivity and yield were predicted using the analysis of variance (ANOVA) aiming at the lowest p-values for the regression model, shown in Table 7. A quadratic model was estimated and the less significant values were removed in order to obtain the lowest p-values for the regression model. The data provided by the model equations indicates that 74.20 % (R²=0.7420) of the ethanol productivity, and 74.22 % (R²=0.7422) of ethanol yield can be explained by the model, showing that the model provides good predictions of the experimental results. It was observed that the model is highly significant, both for ethanol productivity (p = 0.000004) and for yield (p = 0.003). The reduced models for ethanol productivity and yield are presented in equations (1) and (2) , respectively:

$$Q_P=0.1361-0.0519X_3$$ (1)

$$Y_{P/S}=0.3281-0.01X_1+0.0106X_2-0.0264X_2^2-0.0117X_3^2-0.0187X_2{X}_3$$ (2)

where X₁, X₂, and X₃ are the coded values of pH, temperature, and inoculum size, respectively.

Table 6
Central composite design to optimize fermentation conditions of SHH using C. guilliermondii BL 13 and the experimental values obtained for ethanol productivity (Q_P) and yields (Y_P/S).

Assay	pH	Temperature (°C)	Inoculum (CFU mL-1)	QP (g L-1 h-1)	YP/S (g g-1)
1	-1 (4.5)	-1 (25)	-1 (106)	0.10	0.28
2	+1 (5.5)	-1 (25)	-1 (106)	0.10	0.26
3	-1 (4.5)	+1 (31)	-1 (106)	0.11	0.34
4	+1 (5.5)	+1 (31)	-1 (106)	0.11	0.33
5	-1 (4.5)	-1 (25)	+1 (108)	0.19	0.29
6	+1 (5.5)	-1 (25)	+1 (108)	0.18	0.26
7	-1 (4.5)	+1 (31)	+1 (108)	0.20	0.27
8	+1 (5.5)	+1 (31)	+1 (108)	0.20	0.26
9	-1.68 (4.16)	0 (28)	0 (107)	0.07	0.35
10	+1.68 (5.84)	0 (28)	0 (107)	0.11	0.36
11	0 (5.0)	-1.68 (23)	0 (107)	0.15	0.33
12	0 (5.0)	+1.68 (33)	0 (107)	0.21	0.30
13	0 (5.0)	0 (28)	-1.68 (105)	0.02	0.34
14	0 (5.0)	0 (28)	+1.68 (109)	0.24	0.30
15	0 (5.0)	0 (28)	0 (107)	0.12	0.25
16	0 (5.0)	0 (28)	0 (107)	0.12	0.27
17	0 (5.0)	0 (28)	0 (107)	0.12	0.27
18	0 (5.0)	0 (28)	0 (107)	0.12	0.30

Table 7
Regression coefficients estimated by ANOVA for ethanol productivity as a function of temperature (T), pH, and inoculum size (I) in the CCD.

Response	Source	Coefficient	p-value	Regression
Variable				p-value	F	R²
QP	Mean/Interaction	0.1361	0.0000	0.000004	46.014	0.7420
	Linear I	0.0519	0.000004
YP/S	Mean/Interaction	0.3281	0.0000	0.003	6.9086	0.7422
	Linear pH	-0.0100	0.1097
	Linear T	0.0106	0.0952
	Quadratic T	-0.0264	0.0009
	Quadratic I	-0.0117	0.0195
	T x I	-0.0187	0.0297

The models suggest that, within the chosen range, ethanol productivity is independent of pH and temperature. The response surfaces described by the model (Table 7) to estimate ethanol productivity based on the independent variables temperature (X₂) and inoculum size (X₃) is shown in Figure 1. For ethanol productivity, the increase in inoculum size produced the greatest effect. Laopaiboon et al. (2007), who studied the effects of initial cell concentrations (10⁶, 10⁷, and 10⁸ CFU mL^-1) on ethanol production in batch fermentations of S. cerevisiae in sweet sorghum juice, reported that, although the final ethanol concentration would remain the same, there would be a faster substrate consumption and increased ethanol productivity proportionally to higher initial cell concentrations.

Fig. 1
Response surface plots for ethanol productivity (Q_P, g L^-1 h^-1) as a function of inoculum size and temperature.

The response surfaces obtained for ethanol yield described by the model are depicted in Figure 2a and 2b. C. guilliermondii BL 13 produced higher yields of ethanol at pH below 4.5. The individual effects of the initial pH have been well documented in the literature for several yeast strains described within the genus Candida, and some strains of C. guilliermondii for xylitol production (Tamburini et al., 2015; Felipe et al., 1997), but not for ethanol. Regarding the variable temperature, Phisalaphong et al. (2006) studied its effect on ethanol fermentation by S. cerevisiae M30. The authors reported that ethanol yields were slightly higher when increasing the isothermal control from 30 ºC to 33 ºC, but above this range negative effects on cell metabolism and ethanol production were observed.

Fig. 2
Response surface plots for ethanol yields (Y_P/S, g g^-1) as a function of (A) inoculum size and pH; and (B) inoculum size and temperature.

The interaction between temperature and inoculum size suggests that, at low inoculum size, higher temperatures have higher positive response and, inversely, increasing inoculum size, lower temperatures will be significantly better to increase ethanol yields. For instance, Silva and Roberto (2001) studied the combined effects of initial xylose concentration and inoculum size on xylitol production by C. guilliermondii growing in rice straw hydrolysate. The authors reported the optimum xylose concentration and inoculum size to be 82 and 3 g L^-1, respectively. The authors did not comment on ethanol production. Powchinda et al. (1997) demonstrated that, for S. cerevisiae fermentation of sugars, the increase in inoculum size increases ethanol yields up to a critical amount of cells because there is a better utilization of sugars by yeast. However, high cell densities can adversely affect mass and energy transfer in culture broths and increase cell-to-cell interactions, negatively affecting metabolism and ethanol production (Jarzebski et al., 1989; Laluce et al., 2009).

In this work, the highest ethanol productivity would be dependent of inoculum size (1 × 10⁹ CFU mL^-1), independent of the pH and temperature, whereas best ethanol yields would be achieved for pH and inocula sizes around their minimal values and temperature between 28 °C and 29 °C. Although larger inoculum sizes would increase ethanol productivity, they would negatively affect ethanol yields, probably because of more energy being channeled for cell maintenance. Similar results were reported by Yamada et al., (2011), who evaluated the effect of the initial inoculum concentration on ethanol production from brown rice hydrolysate by cultures of S. cerevisiae.

In order to validate the models predicted by the CCD, experiments were carried out in triplicates under the conditions representing the maximal response for ethanol productivity, which was 0.24 g L^-1 h^-1, using an inoculum size of 1×10⁹ CFU mL^-1, pH 5.0 and temperature of 28 °C. An experimental mean value of 0.21 ± 0.015 g L^-1 h^-1 ethanol productivity was obtained, close enough to the predicted value in the CCD, validating the response model.

Some authors reported the production of xylitol by several strains of C. guilliermondii (Mussatto et al., 2006; Silva and Roberto, 2001; Roberto et al., 1996). However, there are only few reports concerning ethanol production by this yeast, and in none of them the optimization of fermentation conditions was performed. This fact reflects the metabolic preference of this yeast to ferment xylose to xylitol instead of ethanol under specific conditions of oxygen limitation. Nevertheless, in this work, it was possible to show reasonable ethanol productivities (0.21 g L^-1 h^-1) by C. guilliermondii BL 13. These results compared well with another report for this yeast, C. guilliermondii strain NRRL Y-2075, which showed ethanol productivities of 0.12 g L^-1 h^-1 when it was cultivated in non-detoxified concentrated SHH containing 0.8 g L^-1 of glucose and 16 g L^-1 of xylose in shaker flasks (Schirmer-Michel et al., 2008).

Ethanol production under optimal conditions in SH-AEH

Assuming that the results of C. guilliermondii BL 13 fermentation of SHH were interesting from the point of view of ethanol production, it was decided to test this yeast in a richer medium, containing sugars from the hemicellulose and cellulose fractions of soybean hull, which was designated SH-AEH, obtained by the enzymatic and acid hydrolyses of this biomass (section 2.2). The kinetics of sugar consumption, ethanol, and xylitol production for C. guilliermondii BL 13 growing in SH-AEH are shown in Figure 3. Fermentation of SH-AEH proceeded vigorously during the first 12 h, with all glucose consumed with a corresponding ethanol production of 16.8 g L^-1 and a high ethanol productivity of 1.4 g L^-1 h^-1, and yield of 0.41 g g^-1. Exponential growth lasted for about 6 to 12 h of cultivation reaching a maximum biomass formation of 1.45×10¹⁰ CFU mL^-1. The xylitol productivity was 0.05 g L^-1 h^-1 and the xylose to xylitol conversion was 0.46 g g^-1. Efficiency of xylose utilization by this yeast was 67 % in SH-AEH. Comparatively, a xylose consumption rate of 69 % was observed for Candida guilliermondii FTI 20037 growing in semi-defined medium containing 85 g L^-1 of xylose and supplemented with (NH₄)₂SO₄, CaCl₂.2H₂O, and 10 % of rice bran extract (Mussatto et al., 2006). Hickert et al., (2014) cultivated S. cerevisiae, S. arborariae, and C. shehatae in a similar soybean hull hydrolysate as used in this work, in a shaker, and reported that the tested yeasts lost their viability after 24 h of cultivation and only 50 % of glucose present in the medium was consumed, whereas xylose remained intact.

Because the culture conditions in a shaker, as conducted in this work, were under a limited supply of oxygen, it would be expected that the ethanol produced should be consumed along with xylose in diauxic kinetics after glucose depletion, a fact clearly observable in Figure 3. Similar fermentation profiles were reported for other yeast species cultivated in lignocellulosic biomass hydrolysates (Schirmer-Michel et al., 2008; Cunha-Pereira et al., 2011; Hickert et al., 2013). Using C. guilliermondii strain FTI 20037 in supplemented enzymatic hydrolysate of sugarcane bagasse, which was de-lignified and pretreated by hydrothermal processing, Silva et al. (2011) reported an ethanol production of 20.5 g L^-1 in 28 h, in shaker cultures growing at 200 rpm and 30 ºC.

Fig. 3
Time course of substrate consumption, ethanol, and xylitol production by C. guilliermondii BL13 cultivated on soybean hull acid-enzymatic hydrolysate. Experiments were carried out in orbital shaker at 28 °C, pH 5.0, 180 rpm. Glucose (□); xylose (○); arabinose (∆); ethanol (▲); xylitol (■). Results are the mean of triplicates.

Roberto et al. (1996) evaluated xylitol production by C. guilliermondii strain FTI 20037 in rice straw hydrolysate. High initial cell density did not show a positive effect in this bioconversion, and increasing the initial cell density from 0.67 g L^-1 to 2.41 g L^-1 had a detrimental effect on the rate of xylose utilization and xylitol accumulation, resulting in xylitol yields of 0.47 g g^-1 and 51 % efficiency of consumption in 72 h of cultivation. These results are consistent with results obtained in this work, where xylose was slowly converted into xylitol, even after all glucose had been metabolized, resulting in a yield of 0.46 g g^-1 and conversion efficiency of 50 % in 72 h of cultivation.

The final composition of the enzymatic hydrolysate (SH-AEH) contained small amounts of HMF, furfural, and acetic acid, all considered toxic compounds for cell metabolism, and an osmotic pressure of 2 916 mOsm kg^-1. However, these factors did not influence yeast metabolism in this work. Comparatively, the growth of C. guilliermondii FTI 20037 was inhibited when cultivated in concentrated acid-hydrolyzed soybean hull showing osmotic pressures of 2 950 mOsm kg^-1 and 2.8 g L^-1 of toxic compounds combined (Schirmer-Michel et al., 2008).

Table 8 presents a general comparison of data for ethanol production using several agro industrial byproducts or residues, including starchy and sugary biomass for comparison, because the last-mentioned are the standard raw materials in the ethanol industry. Comparing the experiments of this work with those using starch, sweet sorghum, brown rice, and cassava pulp (Laopaiboon et al., 2007; Yamada et al., 2011; Kosugi et al., 2009), the results are promising, considering that SH-AEH was not supplemented ((Laopaiboon et al., 2007; Yamada et al., 2011; Kosugi et al., 2009), and because genetically modified or adapted yeast strains were not used ((Laopaiboon et al., 2007; Yamada et al., 2011; Kosugi et al., 2009) in the fermentation. Results for SH-AEH, compared with other lignocellulosic biomass such as corn stalk and stovers, soybean hull, rice straw, and sugarcane bagasse (Schirmer-Michel et al., 2008; Silva et al., 2011; Li et al., 2011; Yang et al., 2011; Chandel et al., 2013), show better ethanol volumetric productivity (1.4 g L^-1 h^-1), even when pretreatment and enzymatic hydrolysis were used in combination with inhibitors-adapted or ethanol-tolerant strains (productivities not higher than 0.89 g L^-1 h^-1) (Li et al., 2011; Yang et al., 2011). In the frame of the concept of biodistilleries, soybean hull could be considered a valuable resource.

Table 8
Comparison among several feedstocks used for the fermentation process to obtain ethanol.

Feedstock	Pretreatment	Strains	Ethanol concentration (g L-1)	Y P/S (g g-1)	Q P (g L-1 h-1)	Reference
Traditional sources of fermentation
Starch/Sugar
Sweet sorghum	-	S. cerevisiae	100	0.42	1.67	Laopaiboon et al. (2007)
Brown rice	-	S. cerevisiae	28.8	0.43	1.2	Yamada et al. (2011)
Cassava pulp (5 %)	Hydrothermal reaction and enzymatic hydrolysis	S. cerevisiae MT8-1/pGA11	18.6	0.50	0.77	Kosugi et al. (2009)
Second-generation ethanol
Lignocellulose
Soybean hull	Acid and enzymatic hydrolysis	C. guilliermondii	16.8	0.41	1.4	This work
Corn stalk	Steam exploded and enzymatic hydrolysis	P. stipitis	42.15	0.45	0.89	Yang et al. (2011)
Corn stover	Steam exploded and enzymatic hydrolysis	S. cerevisiae	43.21	0.47	0.72	Li et al. (2011)
Soybean hull	Acid hydrolysis	C. guilliermondii	5.78	0.53	0.24	Schirmer-Michel et al. (2008)
Rice straw	Acid hydrolysis	P. stipitis	18.7	0.37	0.39	Silva et al. (2010)
Sugarcane bagasse	OAFEX and	C. shehatae	4.83	0.28	0.20	Chandel et al. (2013)
	Enzymatic hydrolysis	S. cerevisiae	6.6	0.46	0.47

CONCLUSION

In this work, the possibility was demonstrated of using SHH and SH-AEH as substrates for ethanol production without the addition of any synthetic nutrients to the media. When using SHH, ethanol productivity and yield could be improved by optimizing temperature, pH, and inoculum size. The strain C. guilliermondii BL 13 proved to be an efficient converter of hexoses to ethanol and, to a lesser extent, xylose to xylitol. Using the enzymatic hydrolysate under the optimal conditions resulted in reasonably high ethanol productivities. Therefore, further optimization of the culture conditions of C. guilliermondii BL 13 on lignocellulosic hydrolysates might be carried out in future experiments, at the same time testing it in co-cultures with other ethanogenic microorganisms, under different oxygen conditions. Results obtained in the present study are promising in terms of product yields and volumetric ethanol productivity to grant further scaling-up studies of such a process.

ACKNOWLEDGEMENTS

The authors wish to thank CNPq and CAPES (Brazil) for their financial support for the project through scholarships.

REFERENCES

• Cassales, A.R., Souza-Cruz, P.B., Rech, R. and Ayub, M.A.Z., Optimization of soybean hull acid hydrolysis and its characterization as a potential substrate for bioprocessing. Biomass and Bioenergy, 35, 4675-4683 (2011).
• Cassman, K.G., and Liska, A.J., Food and fuel for all: realistic or foolish. Biofuel, Bioproducts and Biorefining, 1, 18-23 (2007).
• Chandel, A.K., Antunes, F.F.A., Anjos, V., Bell, M.J.V., Rodrigues, L.N., Singh, O.V., Rosa, C.A., Pagnocca, F.C. and Silva, S.S., Ultra-structural mapping of sugarcane bagasse after oxalic acid fiber expansion (OAFEX) and ethanol production by Candida shehatae and Saccharomyces cerevisiae Biotechnology for Biofuels, 6, 4 (2013).
• Chen, W.H., Xu, Y.Y., Hwang, W.S. and Wang, J.B., Pretreatment of rice straw using an extrusion/extraction process at bench-scale for producing cellulosic ethanol. Bioresource Technology, 102, 10451-10458 (2011).
• Cheng, K., Jing, W., Lin, Z. and Zhang, J., Aerobic and sequential anaerobic fermentation to produce xylitol and ethanol using non-detoxified acid pretreated corncob. Biotechnology for Biofuels , 7, 166 (2014).
• Cunha-Pereira, F., Hickert, L.R., Senhem, N.T., Rosa, C.A., Souza-Cruz, P.B. and Ayub, M.A.Z, Conversion of sugars present in rice hull hydrolysates into ethanol by Spathaspora arborariae, Saccharomyces cerevisiae, and their co-fermentations. Bioresource Technology , 102, 4218-4225 (2011).
• Dillon, A.J.P., Bettio, M., Pozzan, F.G., Andrighetti, T. and Camassola, M., A new Penicillium echinulatum strain with faster cellulase secretion obtained using hydrogen peroxide mutagenesis and screening with 2-deoxyglucose. Journal of Applied Microbiology, 111, 48-53 (2011).
• Felipe, M.G.A., Vitolot, M., Manchilha, I.M. and Silva, S.S., Fermentation of sugar cane bagasse hemicellulosic hydrolysate for xylitol production: effect of pH. Biomass and Bioenergy, 13, 11-14 (1997).
• Fu, N. and Peiris, P., Co-fermentation of a mixture of glucose and xylose to ethanol by Zymomonas mobilis and Pachysolen tannophilus World Journal of Microbiology and Biotechnology, 24, 1091-1097 (2008).
• Gnanasambandan, R. and Proctor, A., Preparation of soy hull pectin. Food Chemistry, 65, 461-467 (1999).
• Granström, T., Ojamo, H. and Leisola, M., Chemostat study of xylitol by Candida guilliermondii Applied Microbiology and Biotechnology, 55, 36-42 (2001).
• Hickert, L.R., Cunha-Pereira, F., Souza-Cruz, P.B., Rosa, C.A. and Ayub, M.A.Z., Ethanogenic fermentation of co-cultures of Candida shehatae HM 52.2 and Saccharomyces cerevisiae ICV D254 in synthetic medium and rice hull hydrolysate. Bioresource Technology, 131, 508-514 (2013).
• Hickert, L.R., Cruz, M.M., Dillon, A.J.P., Fontana, R.C., Rosa, C.A. and Ayub, M.A.Z., Fermentation kinetics of acid-enzymatic soybean hull hydrolysate in immobilized-cell bioreactors of Saccharomyces cerevisiae, Candida shehatae, Spathaspora arborariae, and their co-cultivations. Biochemical Engineering Journal, 88, 61-67 (2014).
• Jarzebski, A.B., Malinowski, J.J. and Goma, G., Modeling of ethanol fermentation at high yeast concentrations. Biotechnology Bioengineering, 34, 1225-1230 (1989).
• Kosugi, A., Kondo, A., Ueda, M., Murata, Y., Vaithanomsat, P., Thanapase, W., Arai, T. and Mori, Y., Production of ethanol from cassava pulp via fermentation with a surface engineered yeast strain displaying glucoamylase. Renewable Energy, 34, 1354-1358 (2009).
• Kuhad, R.C., Gupta, R., Khasa, Y.P., Singh, A. and Zhan, P., Bioethanol production from pentose sugars: current status and future prospects. Renewable and Sustainable Energy Reviews, 15, 4950-4962 (2011).
• Laluce, C., Tognolli, J.O., Oliveira, K.F., Souza, C.S. and Morais, M.R., Optimization of temperature, sugar concentration, and inoculum size to maximize ethanol production without significant decrease in yeast cell viability. Applied Microbiology and Biotechnology , 83, 627-637 (2009).
• Laopaiboon, L., Thanonkeo, P., Jaisil, P. and Laopaiboon, P., Ethanol production from sweet sorghum juice in batch and fed-batch fermentations by Saccharomyces cerevisiae World Journal of Microbiology and Biotechnology, 23, 1497-1501 (2007).
• Li, Y., Gao, K., Tian, S., Zhang, S. and Yang, X., Evaluation of Saccharomyces cerevisiae Y5 for ethanol production from enzymatic hydrolysate of non-detoxified steam-exploded corn stover. Bioresource Technology, 102, 10548-10552 (2011).
• Lin, T.H., Huang, C.F., Guo, G.L., Hwang, W.S. and Huang, S.L., Pilot-scale ethanol production from rice straw hydrolysates using xylose-fermenting Pichia stipitis Bioresource Technology, 116, 314-319 (2012).
• Martínez, M.L., Sanchez, S. and Bravo, V., Production of xylitol and ethanol by Hansenula polymorpha from hydrolysates of sunflower stalks with phosphoric acid. Industrial Crops and Products, 40, 160-166 (2012).
• Mussatto, I.S. and Roberto, I.C., Kinetic behavior of Candida guilliermondii yeast during xylitol production from highly concentrated hydrolysate. Process Biochemistry, 39, 1433-1439 (2004a).
• Mussatto, I.S. and Roberto, I.C., Optimal experimental condition for hemicellulosic hydrolyzate treatment with activated charcoal for xylitol production. Biotechnology Progress, 20, 134-139 (2004b).
• Mussatto, I.S., Silva, C.J.S.M. and Roberto, I.C., Fermentation performance of Candida guilliermondii for xylitol production on single and mixed substrate media. Applied Microbiology and Biotechnology, 72, 681-686 (2006).
• Mussatto, I.S. and Roberto, I.C., Establishment of the optimum initial xylose concentration and nutritional supplementation of brewer’s spent grain hydrolysate for xylitol production by Candida guilliermondii Process Biochemistry, 43, 540-546 (2008).
• Phisalaphong, M., Srirattana, N. and Tanthapanichakoon, W., Mathematical modeling to investigate temperature effect on kinetic parameters of ethanol fermentation. Biochemical Engineering Journal, 28, 36-43 (2006).
• Powchinda, O., Delia-Dupuy, M.L. and Strehaiano, P., Alcoholic fermentation from sweet sorghum: some operating problems. The 9^th Annual Meeting of the Thai Society for Biotechnology and the 2^nd JSPS-NRCT-DOST-LIPP-VCC Seminar on Biotechonology: An Essential Tool for Future Development. Thailand (1997).
• Roberto, I.C., Sato, S. and Mancilha, I.M., Effect of inoculum level on xylitol production from rice straw hemicellulose hydrolysate by Candida guilliermondii Journal of Industrial Microbiology, 16, 348-350 (1996).
• Schirmer-Michel, A.C., Flôres, S.H., Hertz, P.F., Matos, G.S. and Ayub, M.A.Z., Production of ethanol from soybean hull hydrolysate by osmotolerant Candida guilliermondii NRRL Y-2075. Bioresource Technology, 99, 2898-2904 (2008).
• Silva, A., Mussatto, S.I. and Roberto, I., The influence of initial xylose concentration, agitation, and aeration on ethanol production by Pichia stipitis from rice straw hemicellulosic hydrolysate. Applied Biochemistry and Biotechnology, 162, 1306-1315 (2010).
• Silva, C.S.M. and Roberto, I.C., Optimization of xylitol production by Candida guilliermondii FTI 20037 using response surface methodology. Process Biochemistry, 36, 1119-1124 (2001).
• Silva, V.F.N., Arruda, P.V., Felipe, M.G.A., Gonçalves AR and Rocha GJM, Fermentation of cellulosic hydrolysates obtained by enzymatic saccharification of sugarcane bagasse pretreated by hydrothermal processing. Journal of Industrial Microbiology and Biotechnology, 38, 809-817 (2011).
• Tamburini, E., Costa S., Marchetti M.G., Pedrini, P., Optimized production of xylitol from xylose using a hyper-acidophilic Candida tropicalis Biomolecules, 5, 1979-1989 (2015).
• USDA - United States Department of Agriculture. Foreign Agriculture Service: Production, Supply and Distribution Online. Oilseeds, 2014. http://www.fas.usda.gov/psdonline Accessed October 2015.
  » http://www.fas.usda.gov/psdonline
• Yamada, R., Yamakawa, S.I., Tanaka, T., Ogino, C., Fukuda, H. and Kondo, A., Direct and efficient ethanol production from high-yielding rice using a Saccharomyces cerevisiae strain that express amylases. Enzyme and Microbial Technology, 48, 393-396 (2011).
• Yang, X., Zhang, S., Zuo, Z., Men, X. and Tian, S., Ethanol production from the enzymatic hydrolysis of non-detoxified steam-exploded corn stalk. Bioresource Technology, 102, 7840-7844 (2011).

Publication Dates

• Publication in this collection
  Oct 2017

History

• Received
  04 Jan 2016
• Reviewed
  15 Apr 2016
• Accepted
  25 May 2016