Effect of Ethanol Washing on Enzymatic Hydrolysis of OPLP
The ethanol extractives content (9.64%) in OPLP-UW were much higher than those in the untreated biomass (1.18%) (Table 2). The ethanol extractives were reduced to 0.79% in OPLP after ethanol washing (Table 2). During the organosolv pretreatment, lignin was depolymerized and a significant amount of 𝛽-o-4 linkages was cleaved which was catalyzed by acids [24,25]. The depolymerized lignin was precipitated on the surface of the wood fibers and it can be largely removed by ethanol washing [22,26]. Lai et al. have reported that the ethanol washed extractives were similar to ethanol organosolv lignin (EOL) by 13C-NMR [22]. To examine the effect of ethanol washing on enzymatic digestibility of OPLP, the pretreated substrates with and without ethanol washing were hydrolyzed and compared under the SHF and SSF conditions (Fig. 1). The results showed the 72h hydrolysis yield of OPLP-W and OPLP-UW was similar (90%) (Fig. 1a). The addition of precipitated organosolv lignin (0.3g) also did not change the hydrolysis yield of OPLP-W. It indicated that ethanol washing did not have any positive or negative effects on substrates digestibility at the SHF conditions (50 °C and pH 4.8). Previous studies reported that EOL from loblolly pine had a negative effect on enzymatic hydrolysis of OPLP, in which the enzyme loading was 5 FPU/g glucan [27]. High enzyme loading (25 FPU/g glucan) in this study probably overcome the negative effect of EOL on enzymatic hydrolysis. It reduced the effect of nonproductive binding between cellulase and lignin by providing sufficient enzyme active sites [28]. The residual lignin adopted on the additional active sites offered by extra enzyme compared to low enzyme loading and resulted in the negligible negative effect of residual lignin on enzymatic hydrolysis. Enzyme dosage below 10 FPU/g cellulose is usually considered as low enzyme loading [29,30]. The National Renewable Energy Laboratory (NREL) in the United States set the enzyme loading to 19-33 mg protein/g-cellulose (equivalent to 15-20 FPU/g cellulose) when building the ethanol cost evaluation model, which is the normal range for bioconversion of lignocellulosic biomass [31,32]. A minimum cellulase (Celluclast 1.5) loading of 32 mg protein/ g cellulose is required for efficient hydrolysis (70% glucan conversion) of organosolv-pretreated lodgepole pine [33]. A slightly high enzyme loading (25 FPU/ g glucan) applied in this study is to minimize the rate-limiting effect of enzymatic hydrolysis during the SSF process. Therefore, by eliminating the enzyme hydrolysis impact on SSF, the effect of extractable lignin on ABE production could be explored. However, the practice of reducing enzyme loading could be carried out in the future upon obtaining a better understanding of how the extractable lignin affects ABE production in SHF and SSF processes.
While for the enzymatic hydrolysis of OPLP-W and OPLP-UW at the SSF conditions (35 °C and pH 6), the 72 h hydrolysis yield of OPLP-W and OPLP-UW was 82.5% and 73.9% respectively (Fig.1b), which were lower than those at the SHF conditions (50°C and pH 4.8). Notably, the OPLP-UW had even lower hydrolysis yield compared to OPLP-W at the test conditions. The lower temperature (35 °C) also resulted in the lower initial hydrolysate rate (Fig. 1). According to the adsorption kinetics of cellulase on the cellulose and lignin, the decrease of temperature reduced the adsorption of cellulase on both cellulose and lignin, however, the reduction was much more considerable for cellulose than lignin [28,34]. The lignin had a higher affinity to cellulase than cellulose [28,35]. These factors made more cellulase adsorbed on lignin rather than cellulose and the negative effect of extractable lignin on enzymatic hydrolysis was exhibited at the lower temperature. In addition to the lower temperature, higher pH (6.0) could be another main reason for lower hydrolysis yield and initial rate. Similar results have been reported that higher pH (6.0) reduced the hydrolysis yield of organosolv pretreated loblolly pine at 10 FPU/g glucan [36]. The pH increase from 5 to 7 could result in less adsorption of cellulase on cellulose substrate [34] and decrease the enzyme activity, pH 4.8 has long been suggested for cellulase enzymatic hydrolysis [37]. Adding CaCO3 to control pH in the hydrolysis solution could also contribute to the decrease in enzyme activity. It is reported that the inhibition of CaCO3 to enzymatic hydrolysis possibly caused by non-productive enzyme binding on CaCO3 particles and deactivation of enzyme resulting from enzyme aggregation by dissociated calcium ion [38,39]. The observation that the removal of extractable lignin from the examined substrate (softwood) by ethanol washing improved the enzymatic hydrolysis appears not to agree with the effect of EOL from hardwood, but is consistent with the effect of EOL from softwood by Lai et al [22,40]. They reported a contrasting effect of hardwood and softwood organosolv lignin, where EOL from hardwood enhanced enzyme hydrolysis and EOL from softwood inhibited enzymatic hydrolysis. Huang et al. investigated the reason why the lignin from two types of sources exerted opposite effects [27]. They found a strong correlation between hydrophobicity and zeta potential of EOL and enzymatic hydrolysis yield, indicating the stimulation or inhibition effect of lignin is controlled by the combination of hydrophobicity and zeta potential.
Effect of Ethanol Washing on ABE production in SHF processes
Under SHF conditions, the ethanol washing showed no effect on the 72h hydrolysis yield of organosolv pretreated loblolly pine at current enzyme loading (25 FPU/g glucan). However, the subsequent effect on ABE fermentation of the hydrolysates from ethanol washed substrates is unknown. Therefore, three enzymatic hydrolysates from OPLP-UW, OPLP-W and OPLP-W/EOL (plus precipitated EOL) were compared in ABE fermentation (Fig. 2). It was observed that butanol production from the OPLP-UW hydrolysate was 8.16 g/L with a yield of 0.14 g/g at 96 h, and the residual glucose was 5.06 g/L (Fig. 2a). The initial glucose consumption rate (within 36 h) was low at 0.30 g/L/h. The organism began to accumulate butyric acid at 24 h and acetic acid at 48 h. Butyric acid peaked (5.81 g/L) at 72 h and then gradually decreased to 3.89 g/L at 96 h. Butanol production began late at 36 h in the fermentation. The acetone and ethanol reached 2.22 and 1.52 g/L at 96 h, respectively. While for the ABE fermentation with the OPLP-W (Fig. 2b), the initial glucose consumption rate (within 36 h) was fast at 0.69 g/L/h, but glucose consumption and ABE production ceased at 48 h. The initial glucose concentration was nearly 50 g/L. The organism began to accumulate butyric acid at 12 h and quickly reached 6.23 g/L at 36 h and did not decrease further. Butanol production began from 36 h but stopped at 48 h with a low concentration of 1.69 g/L. The butyric acid and acetic acid were 6.44 g/L and 4.24 g/L at 48 h and then leveled off. The residual glucose was 19.72 g/L at 48 h and did not change further. It indicated ABE production from the OPLP-W hydrolysate suffered an “acid crash”, in which solventogenesis was initiated but the metabolic activity (glucose consumption, acid production, and ABE production) ceased within a short time (Fig. 2b). The butyric acid production rate (2.92 mM/h, between 12-36 h) was much higher than that (0.59 mM/h) in ABE fermentation with OPLP-UW hydrolysate. The toxic butyric acid was generated quickly inside cells and inhibited solventogenesis and ceased the ABE production. It has been suggested “acid crash” occurs in pH-uncontrolled ABE fermentation when undissociated acids exceed 57-60 mM [8]. In this study, pH was controlled by CaCO3 and the pH was kept in the range of 5 to 6 over the fermentation time. The concentration of the total acids reached 144 mM at 48 h, which included undissociated acids and dissociated acids. It has been proposed previously that the high concentration of dissociated acids rather than undissociated acids are responsible for the inhibition of solventogenesis at some ABE fermentation [8]. The comparison of ABE fermentation with OPLP-W and OPLP-UW hydrolysates indicated that the metabolism of the organism could be altered by ethanol-washing or the presence of extractable lignin (after pretreatment). We hypothesized that extractable lignin (similar to EOL) can inhibit the glucose consumption and acid production rates thus prevent the “acid crash” in ABE fermentation. To test this hypothesis, precipitated EOL from organosolv pretreatment was added into ABE fermentation of the OPLP-W hydrolysates (Fig. 2c). The initial glucose consumption rate (within 36 h) was fast at 0.52 g/L/h, but glucose consumption and ABE production continued until 84 h. The organisms began to produce butyric acid at 12 h and increased to 4.67 g/L at 36 h and reached 5.76 g/L at 48 h, then decreased due to the shift from acidogenic phase to solventogenic phase. Butanol production began from 36 h and reached 7.60 g/L at 96 h. The acetic acid was 4.17 g/L at 60 h and then leveled off. The acetone and ethanol reached 2.23 and 0.73 g/L at 96 h, respectively. The residual glucose was 4.31 g/L at 96 h, which was similar to that from the ABE fermentation of OPLP-UW hydrolysates. The results demonstrated that the presence of extractable lignin could lower the metabolic rate and prevent the “acid crash” in ABE fermentation. Different approaches have been suggested previously to prevent “acid crash” by pH controlling or lowering the metabolic rate. Lowing yeast extract concentrations (0.05 g/L) in the medium resulted in higher ABE production of 134 mM, low sugar uptake and acid product rates [8]. Overexpressing aldehyde/alcohol dehydrogenase and CoA-transferase in Clostridium beijerinckii was able to prevent “acid crash” and increase butanol production [41]. Syngas fermentation with Clostridium carboxidivorans at a low temperature has been reported to enhance butanol production by lowering metabolic rates at 25 °C [11]. In this study, we found the inhibitory extractable lignin could be potentially effective to prevent the “acid crash” in ABE fermentation by lowering the glucose uptake and acid production rates.
Effect of Ethanol Washing on ABE Production in SSF Processes
ABE production with OPLP-UW and OPLP-W in SSF was compared (Fig. 3 and Table 1). In both cases, ABE fermentation suffered “acid crash” after 60 h, and butanol, ethanol, and acetone production ceased. However, the ABE production recommenced at 96 h for OPLP-W. Specifically for OPLP-UW, acetic acid and butyric acid quickly reached 2.97 and 3.21 g/L at 24 h, respectively. The butanol reached 1.22 g/L at 24 h. The glucose concentration reached 15.11 g/L at 24 h and it was much lower than the initial glucose concentration in the SHF process. For OPLP-W, acetic acid and butyric acid reached 2.92 and 3.09 g/L at 24h, respectively, which are similar to those in OPLP-UW. The butanol reached 1.89 g/L at 24 h. The glucose concentration (20.75 g/L) was 37% higher than that in OPLP-UW at 24 h. This suggested that ethanol washing significantly increased the hydrolyzability of OPLP-W as comparing to OPLP-UW, which provided more initial glucose in SSF process. Cells produced more butanol (3.92 g/L) and less butyric acid (2.00 g/L) from OPLP-W than that from OPLP-UW (2.08 g/L butanol and 2.63 g/L butyric acid) at 60 h. This indicated that initial sugar concentration significantly affected the solvent and acid production, cells appear to produce more acids and fewer solvents when the initial sugar concentration is low. A similar observation has been reported previously, where only 2.93 g/L of solvents were produced from 20 g/L of glucose as compared to 8.77 g/L of solvents from 40 g/L of glucose [42].
In addition, the metabolic activity including acid production and ABE production ceased at 60 h for OPLP-UW, the glucose uptake probably also ceased (Fig. 3a). The total acid concentration was 76 mM at 60 h and did not change until 183 h. The residual extractable lignin not only inhibited the enzymatic hydrolysis but also inhibited the microbial fermentation. Unexpectedly for OPLP-W, the solventogenesis and glucose uptake recommenced at 96 h. All the glucose was consumed, and the final butanol concentration reached 9.29 g/L at 183 h. The total ABE concentration reached 15.74 g/L. During the phase of metabolic inactivity (60-96 h), the total acid concentration had slowly decreased from 66 to 61 mM [8]. Similarly, ABE recommencement after “acid crash” has been reported on pure glucose fermentation before, when the total undissociated acids dropped below a threshold of 55 mM. It should be noticed that the final acetone concentration reached 5.79 g/L, which was much higher than those in the SHF process. The results indicated that residual extractable lignin in OPLP-UW inhibited ABE fermentation and potentially intensified “acid crash” in SSF processes. Comparing the ABE fermentation in SHF and SSF processes, the effect of residual extractable lignin was beneficial in SHF on ABE production by slowing the glucose consumption in ABE fermentation at high initial glucose concentration (50 g/L), but it became unfavorable in SSF due to its inhibition on both enzymatic hydrolysis and ABE fermentation with low initial sugar concentration (around 0 g/L). In SHF processes, high sugar concentration (50 g/L) was available for fast acidogenesis. The presence of extractable lignin in OPLP-UW hydrolysate inhibited microbial metabolic activity and decreased the metabolic rate. Subsequently, the “acid crash” was avoided in OPLU-UW and OPLP-W/EOL hydrolysates. In this case, extractable residual lignin helped ABE fermentation in SHF processes. In SSF processes, the low initial sugar concentration resulted in an “acid crash” for both OPLP-UW and OPLP-W substrates after 60 h. The inhibition of extractable lignin on enzymatic hydrolysis of OPLP-UW made it even less favorable for ABE production due to the lower sugar concentration. The inhibition of extractable lignin on microbial metabolic activity further intensified the “acid crash” for OPLP-UW. This suggested that inhibitory extractable lignin could deep “acid crash” in low sugar concentration for ABE production. Without the presence of extractable lignin in OPLP-W, the butyric and acetic acids were slowly consumed in the “acid crash” phase, which in turn enabled the solventogenesis and glucose uptake to recommence at 96 h. Therefore, it is essential to remove extractable lignin of substrates for ABE production in SSF processes. And a higher initial sugar concentration is needed to prevent the “acid crash” in SSF processes.
The residual lignin was observed to aid ABE production in SHF process but hinder the ABE production in SSF process. It is considered to affect the occurrence of “acid crash” together with initial sugar concentration. However, the threshold of sugar concentration resulted in “acid crash” in both SHF and SSF is not clear and of interest. In the meantime, the presence of lignin levels is also a critical variable affecting the onset of “acid crash”. The combination effect of lignin and initial sugar concentration was also examined in the next part of experiment with the addition of prehydrolysates. Under the test experiment, it is estimated the initial sugar concentration between 5-20 g/L could potentially avoid the acid crash in SHF or SSF processes. The initial sugar concentration in SHF or SSF could be changed by varying solid loading. It is speculated that the ABE production from OPLP-W might be higher than OPLP-UW in SHF process when the solid loading is lower than the current study. Also, the “acid crash” might be avoided by increasing the solid loading of OPLP-W and improving the enzymatic hydrolysis in SSF process. The improvement of enzymatic hydrolysis could be achieved by increasing enzyme dosage or adding additives.
Table 1 Acetone-butanol-ethanol fermentation in SHFand SSF a
|
SHF
|
|
SSF
|
|
OPLP-UW
|
OPLP-W
|
OPLP-W/EOL
|
|
OPLP-UW
|
OPLP-W
|
OPLP-W/PH b
|
Residual glucose (g/L)
|
5.06±0.13
|
19.42±0.51
|
4.30±0.25
|
|
30.93±0.02
|
1.36±0.33
|
0.59±0.36
|
Butanol (g/L)
|
8.16±0.53
|
1.69±0.25
|
7.60±0.39
|
|
2.13±0.05
|
9.29±0.21
|
10.51±0.18
|
Butanol Yield (g/g)
|
0.14±0.01
|
0.03±0.00
|
0.13±0.01
|
|
0.04±0.00
|
0.16±0.00
|
0.15±0.00
|
ABE (g/L)
|
11.89±0.12
|
2.66±0.33
|
10.56±0.22
|
|
3.65±0.05
|
15.74±0.33
|
18.29±0.22
|
ABE Yield (g/g)
|
0.20±0.00
|
0.04±0.01
|
0.18±0.00
|
|
0.06±0.00
|
0.27±0.01
|
0.26±0.01
|
Butyric Acid (g/L)
|
3.89±0.41
|
6.52±0.07
|
4.41±0.50
|
|
2.62±0.31
|
1.21±0.07
|
1.68±0.04
|
Acetic Acid (g/L)
|
3.99±0.31
|
4.25±0.05
|
4.13±0.48
|
|
2.74±0.20
|
1.71±0.06
|
1.80±0.01
|
Acid Crash
|
No
|
Yes
|
No
|
|
Yes
|
Yes c
|
No
|
|
|
|
|
|
|
|
|
|
a Data are presented as the final point in fermentation processes, SHF, 96h; SSF for OPLP-UW and OPLP-W, 183h; SSF for OPLP-W/PH, 132h. The value was presented as mean value ± standard deviation.
b OPLP-W/PH: ethanol washed OPLP with detoxified prehydrolysates.
c Fermentation recommenced after acid crash
Effect of Adding Detoxified Prehydrolysates on ABE fermentation in SSF processes
Considering the removal of residual extractable lignin in SSF gave the best ABE production (Table 1), detoxified prehydrolysates was supplemented into OPLP-W in SSF processes. Two-Step detoxification has been used to detoxify the prehydrolysates from organosolv pretreatment [43]. The final butanol and ABE concentration in OPLP-W with prehydrolysate (OPLP-W/PH) reached 10.51 g/L and 18.29 g/L, respectively. No “acid crash” occurred in this case (Fig. 4a). It suggested that adding detoxified prehydrolysates into SSF processes alleviated the “acid crash” for ABE fermentation due to the increase of initial sugar concentration. The acetone and ethanol reached 6.13 and 1.53 g/L at 96 h, respectively. The residual glucose was only 0.59 g/L at 132 h. The organism began to produce butyric and acetic acids at 12 h. Butyric acid peaked (2.09 g/L) at 24 h and then decreased to 1.14 g/L at 36 h and gradually increased to 1.68 g/L at 132 h. The butyric acid maximum concentration was 35% less than that from OPLP-W without prehydrolysates in the previous SSF process. Butanol production from OPLP-W/PH began at 24 h and slowed at 48 h, but it recommenced at 60 h quickly. The ABE fermentation from OPLP-W/PH completed within 96 h, which was 60-hour shorter than that from OPLP-W without prehydrolysates in the SSF process (Fig. 3b). These results indicated the addition of sugars from prehydrolysates potentially prevented the “acid crash” and helped the ABE fermentation in the SSF processes.
After adding prehydrolysates into OPLP-W, the initial glucose, xylose, mannose, galactose concentration in the aqueous phase was 2.06, 4.09, 3.08 and 2.42 g/L, respectively (Fig 4b). The released glucose reached 12.99 g/L at 12 h and then decreased to 9.93 g/L at 36h. After that, the glucose concentration increased to 12.15 g/L at 60 h, then decreased again quickly to 1.06 g/L at 96 h. The glucose concentration at 12 h was 13% lower than that without prehydrolysates in the previous SSF (Fig. 3b). This probably was caused by the inhibition resulting from residual undetoxified inhibitors in prehydrolysates. It is reported that lignin-derived aromatic compounds induced inhibition or complete inactivation of enzymes [20]. The initial sugars in the prehydrolysates could also inhibit the enzymatic hydrolysis. Previously, sugar inhibition on cellulases and beta-glucosidase has been reported on enzymatic hydrolysis of softwood substrates [44]. The solventogenic clostridia are capable of using both hexose and pentose as carbon source for ABE production. It was observed mannose was used firstly and quickly and followed by xylose. All the available sugars were assimilated by Clostridium at the end of fermentation, leaving an insignificant amount of residual sugars. The results suggested the prehydrolysates could reduce the ABE fermentation time in SSF processes, all the C5 and C6 could be consumed for ABE production and the “acid crash” could be potentially avoided.
In this study, the ABE production from OPLP-W and OPLP-UW was carried out in two different fermentation processes. Overall, the SSF process gave a higher ABE production (15.74 g/L) compared to the SHF process (11.89 g/L) due to the removal of glucose inhibition in SSF process. Although fermentation time for the SSF (156 h) was longer than the SHF (96 h), the whole time of SHF was identical to SSF if the time for enzymatic hydrolysis (72 h) was taken into consideration. The residual extractable lignin showed a significant effect on ABE fermentation on OPLP-W and OPLP-UW. In SHF process, it prevented the “acid crash” by slowing the microbial metabolism and increased the ABE yield from 0.04 g/g (OPLP-W) to 0.20 g/g (OPLP-UW). However, in the SSF process, whereas the initial sugar concentration was low, the presence of residual extractable lignin intensified the “acid crash” and the ethanol washed substrate (OPLP-W) resulted in higher ABE production than that from OPLP-UW (15.74 g/L vs. 3.65 g/L). Also, the addition of prehydrolysates to OPLP-W further improved the ABE fermentation by prevention of “acid crash” and gave the highest ABE titer of 18.29 g/L.