Conversion of rice straw to bio-based chemicals: an integrated process using Lactobacillus brevis
- First Online:
- Cite this article as:
- Kim, JH., Block, D.E., Shoemaker, S.P. et al. Appl Microbiol Biotechnol (2010) 86: 1375. doi:10.1007/s00253-009-2407-8
- 701 Downloads
Commercialization of lignocellulosic biomass as a feedstock for bio-based chemical production is problematic due to the high processing costs of pretreatment and saccharifying enzymes combined with low product yields. Such low product yield can be attributed, in large part, to the incomplete utilization of the various carbohydrate sugars found in the lignocellulosic biomass. In this study, we demonstrate that Lactobacillus brevis is able to simultaneously metabolize all fermentable carbohydrates in acid pre-processed rice straw hydrolysate, thereby allowing complete utilization of all released sugars. Inhibitors present in rice straw hydrolysate did not affect lactic acid production. Moreover, the activity of exogenously added cellulases was not reduced in the presence of growing cultures of L. brevis. These factors enabled the use of L. brevis in a process termed simultaneous saccharification and mixed sugar fermentation (SSMSF). In SSMSF with L. brevis, sugars present in rice straw hydrolysate were completely utilized while the cellulase maintained its maximum activity due to the lack of feedback inhibition from glucose and/or cellobiose. By comparison to a sequential hydrolysis and fermentation process, SSMSF reduced operation time and the amount of cellulase enzyme necessary to produce the same amount of lactic acid.
KeywordsLactobacillus brevis Carbon catabolite repression Simultaneous carbohydrate utilization Lignocellulosic biomass SSMSF
Commercial lactic acid production largely relies on bacterial fermentation of glucose or maltose derived from starch biomass (Datta et al. 1995). However, use of lignocellulosic biomass has some additional advantages (Galbe and Zacchi 2002). Lignocellulosic biomass is an inexpensive, abundant, and underutilized renewable feedstock derived from a number of sources including agricultural and forestry residues, food and beverage processing wastes, municipal wastes, pulp and paper industry wastes, and dedicated energy crops. The available lignocellulosic biomass in US is estimated at 200 million tons and is increasing (Bashir and Lee 1994). In addition to the large supply of biomass, utilization of lignocellulose can augment waste treatment and provide economical benefit to society in terms of reducing solid wastes, carbon dioxide, and greenhouse gases (Hatti-Kaul et al. 2007).
Lignocellulosic biomass is a complex polymer composed of cellulose, hemicelluloses, and lignin. One of the major challenges in using lignocellulosic biomass as a feedstock is the inherent heterogeneity of its sugar composition. Cellulose is a glucose polymer with β(1→6) glycoside linkage and hemicellulose is a heteropolymer containing hexose and pentose sugars. To maximize product yield and productivity from lignocellulose, the complete utilization of mixed sugars is essential (Saha 2003). In addition, lignocellulosic hydrolysate contains cell growth inhibitors such as polyphenolics and furfurals (Nilvebrant et al. 2001; Bjorklund et al. 2002). Ideally, the saccharification of lignocellulosic materials and fermentation of mixed sugars could be integrated into a single processing step in order to increase the productivity and to reduce the requirement of enzymes for hydrolysis. An approach taken to achieve this goal involves both strain and process development. The ideal strain would be able to consume all sugars present simultaneously and operate at the optimal temperature conditions needed for cellulase activity. Moreover, the strain would need to be resistant to any and all inhibitors present in lignocellulose (Houghton et al. 2006).
Many microorganisms are able to utilize the pentose sugars derived from lignocellulose; however, in most cases, pentose catabolism is suppressed by glucose also present in the mixture (Stulke and Hillen 1999). This phenomenon is termed carbon catabolite repression (CCR). When a CCR-positive strain is used for bio-based chemical production from mixed sugar substrates, the overall process design is restricted and productivity is reduced because of the sequential utilization of the sugars. Recently, we demonstrated that L. brevis exhibits an atypical utilization of mixed sugars, with no apparent CCR, and is able to simultaneously consume all sugars that can be utilized as a single carbon source (Kim et al. 2009).
In this study, L. brevis was examined as a fermentative host under the conditions of simultaneous saccharification and fermentation of acid-pretreated rice straw in a fed-batch mode. The fermentation kinetics were evaluated for simultaneous mixed sugar fermentation. The impact of potential inhibitory compounds from acid-pretreated rice straw on lactic acid production by L. brevis was also investigated. Finally, enzyme hydrolysis and mixed sugar fermentation by L. brevis was integrated and optimized. The integrated process, simultaneous saccharification and mixed sugar fermentation (SSMSF), was performed in batch and fed-batch mode with rice straw hydrolysate.
Materials and methods
Bacterial strains, culture media and conditions
L. brevis IFO 3960 was obtained from the Institute for Fermentation, Osaka (Osaka, Japan). L. brevis NRRL 1834 and NRRL 1836 were obtained from the Agricultural Research Service culture collection (Peoria, IL, USA). L. brevis ATCC 14869 (type strain) were purchased from the American Type Culture Collection (Manassas, VA, USA). MRS medium (15.0 g l−1 of bactopeptone, 5.0 g l−1 of yeast extract, 2.0 g l−1 of ammonium citrate, 5.0 g l−1 of sodium acetate, and 2.0 g l−1 of dipotassium phosphate) with 20.0 g l−1 of glucose was used for the cell growth and maintenance. In this study, glucose was not included in MRS unless stated. Carbon sources were prepared separately and mixed with inoculum. The initial pH of the medium was set at 6.0 and the temperature was maintained at 37 °C. Fermentations were initiated by adding a 5% (v/v) inoculum.
Preparation of rice straw hydrolysate
Composition of acid-pretreated rice straw
41.6 ± 0.9
10.6 ± 2.2
(25.7 ± 2.2)f
4.4 ± 1.3
(10.6 ± 0.9)f
4.9 ± 0.4
(11.8 ± 0.4f
21.5 ± 0.9
(51.9 ± 0.9)f
Two different kinds of enzyme-hydrolyzed rice straw were prepared for the inhibition study. Acid-pretreated rice straw was suspended in the 0.05 M sodium acetate (pH 5.0) and the insoluble fraction was separated by centrifugation. The insoluble pellet, mainly cellulose fiber and silica particles, was transferred to filter paper (Whatman paper No.1) and washed twice with distilled water followed by the ethanol. After washing, the insoluble fraction was hydrolyzed in 0.05 M sodium acetate buffer (pH 5.0) with 50 filter paper unit per gram substrate (FPU g-substrate−1) of cellulase (Spezyme CP; Genencor International, Palo Alto, CA, USA) and 50 cellobiose unit per gram substrate (CBU g-substrate−1) of cellobiase (Novozyme 188; Novozymes, Franklinton, NC, USA) at 50 °C for 24 h (Vlasenko et al. 1997). This hydrolysate was designated ‘washed hydrolysate’. In a second treatment, acid-pretreated rice straw was simply hydrolyzed without filtration in 0.05 M sodium acetate buffer with 50 FPU g-substrate−1 of cellulase and 50 CBU g-substrate−1 of cellobiase at 50 °C for 24 h. This hydrolysate was designated ‘unwashed hydrolysate’. Since free xylose and arabinose were washed away during filtration step, ‘washed hydrolysate’ contains only glucose (bound as cellulose) as a potential fermentation carbon source. Conversely, ‘unwashed hydrolysate’ contained all the components from rice straw including glucose, xylose, arabinose, phenolic derivatives from lignin and solid particles.
Enzyme hydrolysis of acid-pretreated rice straw for fermentation studies
The acid-pretreated rice straw was suspended in 0.05 M sodium acetate buffer (pH 5.0) with 50 FPU g-substrate−1 and 50 CBU g-substrate−1 of cellulase and cellobiase, respectively. After enzyme hydrolysis, MRS medium was added and L. brevis was inoculated. Fermentation was carried out with 100 g-dry mass l−1 of acid-pretreated rice straw using a BioFlo 3000 Bioreactor (New Brunswick Scientific, Edison, NJ, USA). Temperature was maintained at 37 °C and pH was controlled at 6.0 by 10 N NaOH. Agitation rate was set at 100 rpm without aeration.
Simultaneous saccharification and mixed sugar fermentation
A fed-batch operation of SSMSF was carried out in a BioFlo 3000 Bioreactor (New Brunswick Scientific, Edison, NJ, USA) with 2.5 L of initial volume. In MRS medium, 80 g-dry mass l−1 of acid-pretreated rice straw was suspended and fermentation was initiated by adding a 2% (v/v) inoculum with 10 FPU g-substrate−1 of cellulase and 10 CBU g-substrate−1 of cellobiase. Sixty grams (dry mass) per liter of acid-pretreated rice straw was added directly into the fermentation for second round of fermentation. The fermentation pH and temperature were maintained at 6.0 by 10 N NaOH and 37 °C, respectively. To measure the substrate and the end-product concentrations, samples were taken from fermentation broth and solid materials were removed by centrifugation at 13,000 rpm (Eppendorf, Hamburg, Germany). The supernatant was filtrated by 0.2 μm syringe filter (Millipore, Billerica, MA, USA) for further analysis by HPLC.
Enzyme stability tests
To examine the stability of enzyme mixture during L. brevis cell growth, 800 FPU and 800 CBU of cellulase and cellobiase were added to 450 ml of modified MRS medium adjusting the final volume to 500 ml. To prevent feedback inhibition of cellulase or cellobiase by glucose, 20 g l−1 of xylose was used to support L. brevis growth instead of glucose. Culture medium was taken every 2 h and optical density of cells and total enzyme activity were measured as described previously by Vlasenko and coworkers with the following modification (Vlasenko et al. 1997). After removal of cells by centrifugation, 1.0 ml of supernatant was mixed in 3.0 ml of 0.05 M sodium acetate buffer (pH 5.0) with the addition of 0.4 g of filter paper. The final enzyme and substrate concentrations were 1.44 FPU (or CBU) g-substrate−1 and 100 g l−1 substrate, respectively. Hydrolysis was performed at 50 °C. One hundred microliters of supernatant was taken from this sample every 30 min for a 2 h period. The reaction was stopped by boiling, and the glucose concentration of supernatant was measured by HPLC. The total enzyme activity was defined as an initial glucose production rate (g-glucose l−1 h−1). All experiments were performed in triplicate.
Analysis of substrate and fermentation end-products
The concentrations of substrates and fermentation end-products were analyzed by HPLC (Shimadzu, Kyoto, Japan) using a BioRad HPX-87H column (BioRad, Hercules, CA, USA). One milliliter of fermentation broth was centrifuged at 13,000 rpm for 10 min and the supernatant was transferred to a new microcentrifuge tube prior to analysis. For HPLC analysis of supernatant, the BioRad HPX-87H column was heated at 65 °C and a refractive index (RI) detector was used for the identification of substrate(s) and product(s). As a mobile phase, 0.01 N H2SO4 was used, and the flow rate was 0.6 ml min−1. The standard deviation of the HPLC concentration determinations for any substrate or end-product was within ±5 mM. To determine cell density, the cell pellet was resuspended in the same volume of deionized water and the optical density (OD) was measured using a Beckman DU 7400 spectrophotometer (Beckman, Fullerton, CA, USA) at 600 nm.
Calculation of kinetic values
The effects of substrate composition: co-utilization of xylose, arabinose, and glucose
Kinetic parameters and substrate consumption ratio of simultaneous fermentation of glucose, xylose and arabinose mixture
Specific cell growth ratea (h−1)
Specific total carbohydrate consumption rateb (nM OD−1 h−1)
Used substrate ratioc
0.81 ± 0.07
0.82 ± 0.26
0.82 ± 0.38
Although L. brevis simultaneously utilized these three sugars without apparent catabolite repression, the preferences between sugars were clearly observed. As shown in Table 2, the molar ratios of substrate consumed were 1.0:1.8 for glucose/xylose and 1.0:2.7 for glucose/arabinose indicating the preference for pentose sugars over glucose. The sugar consumption ratios were maintained in the glucose/xylose/arabinose mixture showing the ratio of 1.0:1.3:2.4 (glucose/xylose/arabinose). Meanwhile, the specific consumption rates of total sugars were similar in any composition of sugar mixtures (Table 2).
The cell growth, specific substrate consumption and specific product formation rates in different concentration of glucose and xylose
Carbon source (mM)
Specific growth ratea (h−1)
Specific substrate consumption rateb (mM substrate OD−1 h−1)
Specific product formation rateb (mM product OD−1 h−1)
4.69 ± 1.39
4.85 ± 1.69
1.14 ± 0.87
4.60 ± 1.58
1.87 ± 0.59
3.08 ± 1.11
4.15 ± 1.76
3.31 ± 1.69
1.21 ± 0.57
1.94 ± 1.05
3.02 ± 1.15
4.51 ± 2.05
3.51 ± 1.88
1.34 ± 0.66
1.90 ± 1.00
3.11 ± 1.18
4.54 ± 1.96
3.57 ± 1.72
1.20 ± 0.33
4.39 ± 1.68
3.17 ± 0.75
3.14 ± 0.86
0.04 ± 0.08
The impact of temperature on the growth of L. brevis
The effect of temperature on the cell growth, substrate utilization and product formation rates
Specific cell growth ratea (h−1)
Specific substrate consumption rateb (mM glucose OD−1 h−1)
Specific product formation ratec (mM lactate OD−1 h−1)
0.28 ± 0.02
0.44 ± 0.14
0.28 ± 0.16
0.28 ± 0.18
0.24 ± 0.05
0.50 ± 0.01
0.18 ± 0.05
0.32 ± 0.01
0.29 ± 0.06
0.36 ± 0.12
0.26 ± 0.08
0.32 ± 0.15
0.32 ± 0.08
0.83 ± 0.32
0.29 ± 0.08
0.75 ± 0.37
The effect of temperature on the product yields of heterofermentation by L. brevis strains
The change of two-carbon metabolic flux implies that L. brevis had an increased requirement for NADH regeneration at the higher temperature. Redirection of metabolic flux from acetate production to ethanol production is expected to result in a reduction of ATP production. Instead, L. brevis accelerated the catabolic activity of the cell, and presumably ATP production, at the higher growth temperature since both the specific substrate consumption and product formation rates were increased up to 2.6-fold. However, the cell growth rate of L. brevis decreased at the higher temperature suggesting ATP produced from a central metabolism was used for cell maintenance rather than cell growth.
The impact of temperature on enzymatic hydrolysis
To determine the total activity of cellulase and cellobiase mixtures at different temperatures, glucose production rates during hydrolysis were calculated (Fig. 2b). The initial glucose production rates were 8.0 g l−1 h−1 at 50 °C and 4.1 g l−1 h−1 at 30 °C and 40 °C. Although the initial glucose production rate at 50 °C was two times higher than at 30 °C and 40 °C, a significant drop was observed in enzyme activity due to the inhibition by glucose produced. After 4 h of initial hydrolysis, the glucose production rate decreased to 1.0 g-glucose l−1 hr−1 at all three temperatures.
The effect of rice straw hydrolysate: potential growth inhibitors
Plant cell hydrolysate contains inhibitors of microbial growth including soluble phenolic compounds derived from lignin and acid-induced reaction products of sugars such as furfural or 5-hydroxymethylfurfural (Larsson et al. 2000; Zaldivar et al. 2000; Martinez et al. 2001). Moreover, rice straw possesses large amounts of silica particles that can physically damage microbial cells during the vigorous mixing of a fermentation. Several pretreatment processes have been developed to eliminate such inhibitors and to improve fermentability of hydrolyzed biomass, however, limitations still exist in a large-scale process (Larsson et al. 1999; Nilvebrant et al. 2001; Bjorklund et al. 2002; Persson et al. 2002). These physical and chemical inhibitors can be removed by filtration followed by the acid hydrolysis of rice straw. However, additional filtration step makes the overall process more complicated and expensive. In addition, the pentose sugars from hemicellulose are removed during this process. Given that pentose content up to 50% of available carbohydrates in rice straw hydrolysate, a loss of pentose sugar should not be ignored for the fermentation of lignocellulosic biomass.
Enzyme stability during the fermentation
Sequential hydrolysis and fermentation of acid-pretreated rice straw
Simultaneous saccharification and mixed sugar fermentation in batch mode
Enzyme hydrolysis of cellulosic fiber and fermentation of lignocellulosic sugars were then integrated into a single process termed “simultaneous saccharification and mixed sugar fermentation”. SSMSF was performed in the batch mode at the 200 ml scale with 100 g-dry mass l−1 of acid-pretreated rice straw. To take advantage of the integrated process, 10 FPU g-substrate−1 of cellulase and 10 CBU g-substrate−1 of cellobiase were added in MRS medium instead of 50 FPU (or CBU) g-substrate−1 was used in the SHF of acid-pretreated rice straw. Temperature was maintained at 37 °C and pH was set at 6.0 initially.
Although the 40 °C was determined to be optimum temperature for the integrated process (as discussed above), the SSMSF was performed at 37 °C due to the unique characteristics of the initial rice straw hydrolysate medium. When the acid-pretreated rice straw was resuspended in MRS, the initial medium was a viscous semi-solid paste due to the high concentration of cellulose fiber and silica. At low agitation rate, these characteristics impacted temperature control and permitted temperature fluctuations above 40 °C which negatively impacted L. brevis cell growth (data not shown). As a consequence, 37 °C was determined to be a better set point when working with this unusual substrate thereby preventing temperature fluctuations rising above 40 °C.
The control experiment that was performed without inoculation exhibited inefficient hydrolysis of cellulose fiber due to the sub-optimum temperature, improper mixing, and, most likely, feedback inhibition (Fig. 6b). During 80 h of hydrolysis, glucose concentration increased only 14.7 g l−1.
A stoichiometric calculation derived from the lactic acid produced during 34 h of SSMSF indicated that, at a minimum, 21.7 g l−1 of glucose was consumed even though the observed change of glucose concentration was only 0.1 g l−1. During the same time frame, only 9.2 g l−1 of glucose was produced from enzyme hydrolysis alone indicating that the SSMSF produced 2.4-fold more glucose than the enzyme hydrolysis alone (Fig. 6c).
Fed-batch operation of SSMSF with acid-pretreated rice straw
Estimated renewable biomass in the US is approximately 200 million dry tons/year. As such, lignocellulosic biomass is a most promising substitute for petroleum in chemical and fuel production. However, several technical hurdles limit the use of lignocellulosic biomass as a commercial substrate. The inherent complexity and heterogeneity of lignocellulose complicate the fermentation strategies centered on complete utilization of all sugars present. As a consequence, product yields are lower than that obtained from starch biomass substrates. In order to overcome these critical barriers in commercialization, several conditions need to be satisfied. Mixed sugars from lignocellulosic biomass have to be utilized efficiently and completely in order to maximize the overall yield. Hydrolysis of substrate and fermentation processes should be integrated to reduce the operation time and increase the efficiency of hydrolysis. In order to increase the final product concentration, the fermentation process needs to be operated in fed-batch mode. Considering the efficacy of process design and its operation to meet those conditions, each sugar in the mixture needs to be utilized simultaneously by the fermentation host microorganism. In addition, the microbial strain should be resistant to the potential inhibitors in lignocellulosic hydrolysate to avoid the additional pretreatment steps. Finally, where possible, the use of cellulase and cellobiase should be minimized since the enzymes are one of the dominant cost factors for the fermentation of lignocellulosic biomass.
Various researchers have sought to identify or engineer strains capable of efficient utilization of the pentose derived from hemicellulose (Du Preez and Van Der Walt 1983; Zhang et al. 1995; Ho et al. 1998; Jeffries 2006; Rodrigues et al. 2008). In most cases, however, glucose suppressed the consumption of pentoses showing a typical sequential utilization of mixed sugars (Laplace et al. 1993; Lawford et al. 1998, 2000; Mohagheghi et al. 1998). Dien et al. developed an E. coli strain which completely lacks carbon catabolite repression (CCR) by deletion mutation of glucose PTS genes (ptsG; Dien et al. 1999, 2001, 2002). Recently, we demonstrated that L. brevis was able to simultaneously metabolize a large number of sugar substrates including xylose, arabinose, and galactose without the typical hierarchical carbohydrate consumption pattern (Kim et al. 2009).
While the molecular underpinnings for this relaxed control of sugar consumption remain to be determined, L. brevis met the demands as an optimal strain for the lignocellulosic utilization as any fermentable sugars in the media could be metabolized simultaneously without suppression by glucose. Moreover, L. brevis preferred the pentose sugar over glucose (Table 2 and Table 3). Meanwhile, the specific L. brevis growth and total carbohydrate consumption rates remain constant regardless of sugar composition, or their initial concentration ratio, suggesting L. brevis maintains a maximum metabolic activity when growing on mixed sugars similar to that of a single sugar like glucose.
We noted that washing the acid-pretreated rice straw hydrolysate to eliminate the soluble inhibitors prolonged the lag time of cell growth, likely due to removal of pentose sugars, and reduced the lactic acid production. By comparison, unwashed hydrolysate supported vigorous growth indicating L. brevis is resistant to such inhibitors. The natural resistance to the potential inhibitors in rice straw hydrolysate provides a strong advantage of L. brevis as a host microorganism of lignocellulosic fermentation. Others have noted that L. brevis was resistant to hops—the plant antibacterial compound (Sakamoto et al. 2001, 2002)—and attributed the resistance to specific ABC transporters (Suzuki et al. 2002). Given that L. brevis is often found in lignocellulose-rich environments, such as plant materials and fermented foods or beverages, it is likely this species has evolved/acquired such resistance to dominate these niches.
For the efficient fermentation of lignocellulose, the integration of both enzymatic hydrolysis and fermentation steps offers a possible increase of the productivity combined with a reduction in enzyme requirement. Others have examined concurrent operation of enzymatic hydrolysis of cellulose fiber or starch biomass, and the fermentation of resultant glucose (Nikolic et al. 2009; Ou et al. 2009; Trovati et al. 2009). However, these simultaneous saccharification and fermentation (SSF) approaches necessarily use feedstock that could be broken down into glucose alone. Lignocellulosic fermentations are more problematic because of the additional pentose sugars which, in the case of L. brevis, can be simultaneously fermented. The synergistic effects of integration can only be achieved when two processes, hydrolysis and fermentation, are compatible. For example, the catalytic activity of hydrolysis should not jeopardize, nor be jeopardized by, the bacterial fermentation. As shown in Fig. 4, the stability of cellulase and cellobiase were maintained throughout the fermentation of L. brevis and the fermentation was not influenced by the presence of enzymes in the medium.
Not surprisingly, the optimum conditions for the fermentation of L. brevis are different from those for enzyme hydrolysis of lignocellulosic starting material. While MRS medium at pH 6.0 was compatible with the hydrolysis of acid-pretreated rice straw (data not shown), temperature strongly impacted both the hydrolysis and the fermentation processes. The initial hydrolysis rates at 40 °C and 30 °C were reduced to a half of that observed at 50 °C, a temperature that does not support growth of L. brevis. Nonetheless, the feedback inhibition by glucose and cellobiose produced a larger effect on the hydrolysis process than the temperature shift away from 50 °C. Even at 50 °C, total enzyme activity decreased exponentially approaching only 20% of initial activity after first 4 h of hydrolysis (Fig. 2b). Thus, removal of feedback inhibition by simultaneous saccharification and fermentation at a lower temperature appears to compensate for the reduction of enzyme hydrolysis rate. Indeed, SSMSF produced 2.4-fold more glucose than the simple enzyme hydrolysis at 37 °C (Fig. 6c). In a similar vein, the use of a non-optimal (high) fermentation temperature for L. brevis resulted in a decreased cell growth rate yet produced a better outcome in terms of the bio-based chemical production. Both specific product formation and substrate consumption rates increased at 40 °C and the yield of ethanol, which is currently used as fuel chemical, increased as well.
A critical factor for the commercial use of lignocellulosic biomass is the high cost of enzymatic hydrolysis. Reduction of this enzyme requirement can be achieved in two different ways. Removal of feedback inhibition enables maximum activity throughout the process and thus minimizes the need for additional enzyme input. In addition, enzyme stability throughout the fermentation decreases the total amount of enzyme required during the fed-batch operation of SSMSF. As shown in Fig. 6, the lactic acid production during the batch SSMSF ceased after 36 h while the glucose concentration increased continuously for additional 44 h in the similar pattern to that obtained from direct hydrolysis without cell growth. This enzyme stability within the fermentation process may enable subsequent fed-batch operation of SSMSF without additional enzyme supplement (Fig. 7).
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.