Structure dependent toxicity of lignin phenolics and PEG detoxification in VHG ethanol fermentation

Abstract

The inhibition of lignin phenolics on fermenting microbes has become one of the major barriers in developing an economically viable process for cellulosic ethanol production. In this study, the toxicity to yeast cells and the inhibition to the very high gravity (VHG) fermentation of the phenolic compounds were investigated in the absence or presence of polyethylene glycol (PEG). It was found that the inhibitory effects of phenolic compounds on VHG ethanol fermentation depend on the activity of their hydroxyl (–OH) hydrogen. 250 g L⁻¹ PEG-1000 detoxified 2.0 g L⁻¹ guaiacol in the fermentation broth, and boosted the ethanol concentration from 131 g L⁻¹ to 173 g L⁻¹. The inhibitory effect of 5.0 g L⁻¹ guaiacol on ethanol fermentation was also alleviated, and the ethanol concentration was increased from 51 g L⁻¹ to 151 g L⁻¹ after detoxification with 250 g L⁻¹ PEG-1000. The ¹H-NMR of hydroxyl group (–OH) of phenolic compounds in PEG revealed the role of hydrogen bonding formation on the in situ detoxification mechanism of PEG, and the order in the strength of the intermolecular hydrogen bond between phenolic compounds and PEG. Furthermore, the kinetics of VHG ethanol fermentation in the presence of phenolic compounds were determined. The obtained kinetic model (phenolic compounds inhibitory effect) fits well the kinetics of ethanol production from lignocellulosic hydrolyzates using batch VHG ethanol fermentation technology.

---

Introduction

Fuel ethanol has received considerable attention as an environmentally friendly transportation fuel. The economics of fuel-ethanol production are significantly influenced by the cost of raw materials and energy consumption in ethanol distillation.^1–3 Lignocellulosic biomass has been intensively studied for the production of fuel ethanol due to the abundant carbohydrate content of the broadly distributed non-food feedstock and the low cost. However, serious challenges remain to be addressed in the development of a fermentation technology for economical production of cellulosic ethanol due to lower ethanol titer and lower tolerance of fermentation microbes to toxic compounds.^4,5 Meanwhile, very high gravity (VHG) ethanol fermentation technology has gained considerable attention most recently as it offers the advantages of increased ethanol titer,⁶ which reduces energy consumption for ethanol distillation.⁷ This technology helps in the conversion of lignocellulosic feedstock to ethanol with improved economics. However, the current VHG ethanol fermentation suffers from incomplete glucose utilization and decreased fermentation rate by subjecting yeast cells to high osmotic stress and severe ethanol inhibition in yeast cells.⁸ To address the problem of stuck fermentation resulting from severe ethanol inhibition in yeast cells, we have developed a new method by using polyethylene glycol (PEG) to improve the viability and tolerance of industrial S. cerevisiae yeast (starch-base) cells with boosted ethanol fermentation performance under VHG fermentation conditions.⁹

Pretreatment of lignocellulosic biomass generates toxic compounds (including furans, weak organic acids, and phenolic compounds) that are known to inhibit the fermenting microbes with reduced ethanol productivity.^10,11 To achieve a high concentration of fermentable sugars for VHG ethanol fermentation, the lignocellulosic biomass loading ratio must be increased to a considerable level, therefore resulting in the accumulation of inhibitors to a much elevated level.^12–14 Accumulated phenolic compounds may create stronger inhibition than furans and organic acids as a result of their relatively high contents.¹⁵ It has also been reported that phenolic compounds are the most toxic compounds among the various inhibitors even in small amount,^12,16 with the low molecular weight phenolic compounds being the most toxic.^17,18

To alleviate the inhibition of phenolic compounds on ethanol productivity, several types of detoxification processes have been employed, such as extensive pretreatment to remove phenolic compounds or improvement of the tolerance of yeast strains to phenolic compounds from lignocellulosic hydrolyzates. One approach is the adsorption of activated charcoal to remove the phenolic compounds from hydrolyzate.¹⁹ Immobilized laccase significantly reduced the amount of toxic phenolic compounds in the xylan rich fraction (XRF) by polymerization and enhanced the ethanol productivity.²⁰ The surfactant-based detoxification method achieved more than 90% removal of phenolic compounds from simulated hydrolysate and corn stover hydrolysate.²¹ Those and other reported methods carry limitations including sugar loss, and additional filtration steps. An in situ detoxification strategy by PEG exo-protection of fermentable microbe demonstrated the advantages of alleviated phenol inhibition without additional detoxification steps and lowered sugar loss.²²

While various phenolic compounds may be generated from a pretreatment of lignocellulosic biomass prior to the hydrolysis of cellulose and the fermentation of the sugars, the relative toxicity of the different phenolic compounds differing in the position of the substituent group to fermentation yeasts has not been reported. Among the phenolic compounds, guaiacol is a typical compound from lignin (guaiacyl lignin), and it has been proposed as a major fermentation inhibitor in hydrolysate. So, we select guaiacol as an appropriate model compound to study PEG detoxification in greater details. Thus, the objectives of this work are to study (1) the effects of phenolic compounds on VHG ethanol fermentation in the absence or presence of PEG-1000, (2) the detoxification mechanism of PEGs for phenolic compounds by using ¹H-NMR spectroscopy, (3) the kinetics of VHG ethanol fermentation in the presence of guaiacol and phenolics, and their kinetic parameters.

Results and discussion

Comparison of the inhibitory effects of guaiacol and phenol on VHG ethanol fermentation

We first investigated the inhibition of 2.0 g L⁻¹ guaiacol on VHG ethanol fermentation in great details. Fig. 1a and b show the glucose conversion and ethanol concentration profiles in the presence of 2.0 g L⁻¹ guaiacol. For the control experiment, the end of glucose fermentation was reached in 72 h without inhibitor, 90% of glucose conversion and ethanol in 160 g L⁻¹ were obtained. The VHG fermentation productivity was evidently inhibited when 2.0 g L⁻¹ guaiacol was present in the fermentation broth. Even after extending the fermentation period to 96 h, the produced ethanol concentration in 131 g L⁻¹ and the glucose conversion in 75% were lower than that of control. However, when 250 g L⁻¹ PEG-1000 was added to the same fermentation broth, the ethanol concentration and glucose conversion were pronouncedly improved and became comparable with that of the control in 64 h. Furthermore, the glucose was nearly completely converted in 96 h. The maximum ethanol concentration was increased to 173 g L⁻¹ which was substantially higher than the ethanol concentration ceiling (160 g L⁻¹) of the control. During the VHG fermentation, only 6% yield of glycerol and 1% yield of acetic acid were detected. Glycerol is biocompatible for yeast cell, and the acetic acid is not considered as a toxicant because of its relatively low content (only about 1.2 g L⁻¹). Obviously, PEG-1000 in 250 g L⁻¹ not only detoxified 2.0 g L⁻¹ guaiacol in the fermentation broth, but also improved the tolerance of the yeast cells to a high concentration of ethanol, as reported in a previous study.⁹ As a result, the ethanol concentration was boosted from 160 g L⁻¹ to 173 g L⁻¹. In comparison, a previous study²² has shown that 250 g L⁻¹ PEG-1000 helped to recover the ethanol concentration to that of control (160 g L⁻¹) by detoxifying 2.0 g L⁻¹ phenol, but not sufficient to break the high ethanol concentration ceiling (160 g L⁻¹).

Fig. 1 The inhibitory effects of guaiacol on glucose conversion and ethanol concentration in the presence and absence of PEG-1000. (a) The inhibitory effects of 2.0 g L⁻¹ guaiacol on glucose conversion. (b) The inhibitory effects of 2.0 g L⁻¹ guaiacol on ethanol concentration. (c) The inhibitory effects of 5.0 g L⁻¹ guaiacol on glucose conversion. (d) The inhibitory effects of 5.0 g L⁻¹ guaiacol on ethanol concentration. Fermentation conditions: 398 g L⁻¹ glucose, approximately 5 × 10⁸ cells per mL, 250 g L⁻¹ of PEG-1000, 33 °C, 160 rpm, and pH of 4.3.

The inhibitory effect of 5.0 g L⁻¹ guaiacol on VHG ethanol fermentation with and without PEG-1000 was also evaluated. The glucose conversion and the produced ethanol concentration curves are shown in Fig. 1c and d. It can be seen that 5.0 g L⁻¹ guaiacol has seriously inhibited VHG fermentation productivity. Even after extending the fermentation period to 96 h, only 31% glucose conversion and 51 g L⁻¹ ethanol were obtained. Although adding 250 g L⁻¹ PEG-1000 improved fermentation productivity, the obtained 151 g L⁻¹ ethanol and 84% glucose conversion are still inferior to the control (90% glucose conversion with 160 g L⁻¹ ethanol). The results indicate that 250 g L⁻¹ PEG-1000 has largely, although not fully, detoxified 5.0 g L⁻¹ guaiacol. In comparison, when 5.0 g L⁻¹ phenol was present in the VHG fermentation broth, glucose fermentation was almost completely inhibited; only 10% glucose conversion and 16 g L⁻¹ ethanol were obtained (data shown in Fig. S1†). Even though adding 250 g L⁻¹ PEG-1000 showed a substantially improved glucose conversion (72%) and ethanol concentration (129 g L⁻¹), the inhibitory effect appeared stronger by phenol than by guaiacol. The results imply that 250 g L⁻¹ PEG-1000 can't afford sufficient detoxification for 5.0 g L⁻¹ phenol inhibition on VHG fermentation.

To determine if the ethanol concentration could be increased to above the control ceiling of 160 g L⁻¹ by increasing PEG-1000 concentration in the presence of 5.0 g L⁻¹ guaiacol or phenol, we performed the experiments at increasing PEG-1000 concentration in ethanol fermentation (Fig. S3†). It was found that the ethanol concentration was first increased and then decreased with the increasing PEG-1000 concentration. The maximum ethanol concentration (151.0 g L⁻¹) is lower than 160 g L⁻¹, indicating that ethanol concentration cannot be increased to more than 160 g L⁻¹ by optimizing PEG-1000 concentration. This may be attributed to the excess of PEG-1000 in the fermentation system, which could affect the vitality of yeast cells.

The above results indicate that the phenolic compounds have different inhibitory effects on VHG ethanol fermentation at the same weight concentration. The above stuck VHG ethanol fermentation can be explained by the severe inhibition of phenolic compounds on fermentation microorganisms, and the phenolic compounds at the same weight concentration may exhibit different toxicity behaviour to yeast cells.

Yeast cell mortality from VHG ethanol fermentation processes

To confirm the toxicity of different phenolic compounds to yeast cells, the effects of guaiacol and phenol (range 1.0–5.0 g L⁻¹) on cell mortality were also investigated in the absence of PEG in a 48 h period.²³ Compared to the cell mortality of the control (Fig. 2a),⁹ the yeast cell mortality is lower than the control when guaiacol concentration is at or below 3.0 g L⁻¹ (Fig. 2b). This observation may be ascribed to the increased lag time from the addition of guaiacol. The higher cell mortality than the control (Fig. 2b) can be attributed to the toxicity of phenol and guaiacol compounds in sufficient amount to kill the yeast cells. Fig. 2b shows that the cell mortality with guaiacol is lower than that of phenol in each concentration. As we had expected, the toxicity of guaiacol to yeast cells is lower than that of phenol at the same weight concentration. To get more insight on the cause of phenolic compounds inhibition on ethanol productivity, the yeast cell mortality in 48 h was compared for guaiacol in the concentration range of 1.0–5.0 g L⁻¹ (Fig. 2c) before and after adding PEG-1000.²³ It can be seen that the yeast cell mortality increased with increasing guaiacol concentration when guaiacol concentration was over 2.0 g L⁻¹ (Fig. 2c). The presence of 5.0 g L⁻¹ guaiacol resulted in 83% yeast cell death after 48 h fermentation period. However, when PEG-1000 was supplemented in the fermentation broth, the yeast cell mortality maintained an almost constant value and showed no remarkable increase with the increasing guaiacol concentration after 48 h. Therefore, the yeast cell mortality results clearly indicate that the tolerance of the yeast cell to guaiacol was boosted when PEG-1000 was present in the fermentation media. The added PEG-1000 may have provided exo-protection to the yeast cells.²² The exo-protected yeast cells by PEG-1000 showed significantly improved ethanol productivity in the presence of guaiacol (see Fig. 1). The yeast cell mortality in the presence of phenol also showed similar trends as that in the presence of guaiacol (data shown in Fig. S2†).

Fig. 2 Comparison of yeast cell mortality. (a) The cell mortality of the control; (b) yeast cell mortality at different guaiacol or phenol concentrations from VHG fermentation at 48 h; (c) yeast cell mortality with or without PEG-1000 at different guaiacol concentrations from VHG fermentation at 48 h. Fermentation conditions: 398 g L⁻¹ glucose, approximately 5 × 10⁸ cells per mL, 250 g L⁻¹ of PEG-1000, 33 °C, 160 rpm, and pH of 4.3.

The different phenolic compounds showed different inhibitory effects on VHG fermentation and the toxicity behavior to yeast cells at the same weight concentration. These results suggest that the inhibitory effect of phenolic compounds may be dependent on their molecular structure. It is therefore important to investigate the relationship of the phenolic compound structures with their inhibitory effects on the fermentation performance.

The relationship of different phenolic compound structures with their inhibitory effects on VHG ethanol fermentation

In order to understand the relationship of different phenolic compound structures with their inhibitory effects on VHG ethanol fermentation, we further compared the inhibition effects of different phenolic compounds on glucose fermentation at the same molar concentration (0.04 mol L⁻¹, equivalent to guaiacol in 5.0 g L⁻¹). As shown in Fig. 3a, the order of the inhibitory effect by the phenolic compounds was determined as follows based on the order of experimental glucose conversion and ethanol concentration: m-methoxyphenol > phenol > o-methoxyphenol (guaiacol) ≈ p-methoxyphenol > 2,6-dimethoxyphenol. It is important to note that the inhibitory effect of phenolic compounds is dependent on the structure of the phenolic compounds given the same molar concentration. The activity of the hydroxyl hydrogen as determined by the molecular structure of the phenolic compounds appears to play a major role. The activity of hydroxyl hydrogen in phenolic compounds depends both on the position of methoxy group (–OCH₃) in the phenyl ring and the number of the methoxy group. The methoxy group is a substituent with both electron-inductive withdrawing effect and conjugative electron-donating effect. For m-methoxyphenol, the methoxy group (–OCH₃) has a stronger inductive electron-withdrawing effect, which resulted in a higher activity of hydroxyl hydrogen in m-methoxyphenol than that in phenol. For guaiacol, p-methoxyphenol and 2,6-dimethoxyphenol, the methoxy group (–OCH₃) exhibit conjugative electron-donating effect, which resulted in their activity of hydroxyl hydrogen weaker than that in phenol. Furthermore, 2,6-dimethoxyphenol has two methoxy groups (–OCH₃) in the o-position of phenyl ring, so the activity of hydroxyl hydrogen in 2,6-dimethoxyphenol is lowest compared to that of other phenolics. The order of the activity of hydroxyl hydrogen in phenolic compounds is as follows: m-methoxyphenol > phenol > guaiacol ≈ p-methoxyphenol > 2,6-dimethoxyphenol, which is in good accordance with their inhibitory effects. Fig. 3a also shows that the ethanol formation did not proportionally decrease with glucose conversion in response to the toxicity of the phenolic compounds. This observation may be ascribed to variation in the glycerol and acetic acid produced during the metabolic pathways for the consumption of glucose with the phenolics.

Fig. 3 The inhibitory effects of different phenolic compounds (0.04 mol L⁻¹) on VHG ethanol fermentation. (a) The VHG ethanol fermentation in the absence of PEGs. (b) The VHG ethanol fermentation in the presence of 250 g L⁻¹ PEG-1000. Fermentation conditions: 398 g L⁻¹ glucose, approximately 5 × 10⁸ cells per mL, 250 g L⁻¹ of PEG-1000, 33 °C, 160 rpm, and pH of 4.3.

Although the phenolic compounds have shown different inhibitory effects on glucose fermentation in the absence of PEGs, remarkably, the ethanol concentration and glucose conversion were pronouncedly improved and became indistinguishable in glucose fermentation when 250 g L⁻¹ PEG-1000 was added to the fermentation broth (Fig. 3b). The difference among the phenolic compounds of different structures on VHG ethanol fermentation disappeared in the presence of 250 g L⁻¹ PEG-1000. Therefore, the inhibitory effect of different phenolic compounds on VHG fermentation were alleviated by addition of 250 g L⁻¹ PEG-1000, and the ethanol production have reached the same level.

Detoxification mechanism of PEG for phenolic compounds

A preliminary detoxification mechanism of PEG to phenol involving intermolecular hydrogen bond has been proposed based on qualitative ¹H-NMR spectroscopy.²² In this work, we studied the modes of hydrogen bonding between different phenolic compounds and PEGs (Fig. 4). According to the chemical structures, PEG is expected to be good hydrogen bond acceptor (HBA) due to the unpaired electrons in the oxygen of the ether linkage (–O–). While phenolic compounds are good hydrogen bond donor (HBD) because the phenolic OH can donate protons to form hydrogen bonds. The OH proton of phenolic compounds can interact with the ether oxygen bond (–O–) of PEG to form an intermolecular hydrogen bond²⁴ (as shown in Fig. 4), and also can form intramolecular hydrogen bond with ortho-methoxy group (–O–) (Fig. 4a within the green circle).

Fig. 4 Modes of different hydrogen bonding between phenolic compounds and PEGs. (a) The OH proton of guaiacol forming both intramolecular H-bonding (in green circle) with o-methoxy ether (–O–), and intermolecular hydrogen bonding with the ether oxygen (–O–) of PEG. (b) The OH proton of p-methoxyphenol interacts with the ether oxygen (–O–) of PEG, forming intermolecular hydrogen bond.²⁴

We further conducted quantitative ¹H-NMR spectra of phenolic OH proton in the presence of PEG-1000 to verify the role of hydrogen bonding formation on the detoxification mechanism of PEG. There are strong intermolecular H-bonds between PEG with H₂O besides phenolic compounds in VHG fermentation broth, which may cause confusion with intermolecular H-bonds between PEG with phenolic compounds. In order to understanding the in situ detoxification mechanism of PEG for phenols, it is necessary to exclude the effect of H₂O. Therefore, we performed ¹H-NMR experiments in the presence of PEG-1000 and without H₂O, in poor H-donating solvent CDCl₃. The PEG amounts in Fig. 5 were much lower than that in tests corresponding to Fig. 1. Guaiacol and p-methoxyphenol were taken as representative phenolic compounds to probe the formation of intermolecular hydrogen bonding systematically. In general, the magnetic resonance of H-bonded protons moves downfield compared to non-hydrogen-bonded protons.²⁵ Fig. 5a shows that the chemical shift of OH proton of guaiacol moved toward downfield and the line width became broadened gradually with the increase of molar ratio of PEG-1000 to guaiacol. The ¹H-NMR peak broadening has been attributed to the increase of association degrees of intermolecular hydrogen bonding, which also results in increased relaxation rate R₁ (=1/T₁) of guaiacol molecules.²⁶ Fig. 5b also shows that the chemical shift of p-methoxyphenol OH proton moved to downfield and the line width became broadened when the molar ratio of PEG-1000 to p-methoxyphenol is 0.04; and no OH proton signal is observed when the molar ratio of PEG-1000 to p-methoxyphenol is 0.12. The reason may be due to the lack of competitive intramolecular hydrogen bond (see Fig. 4b), the strong association degrees of intermolecular hydrogen bonding enhanced relaxation rate R₁ (=1/T₁) of p-methoxyphenol molecules.²⁶ The intermolecular hydrogen bonding of the OH proton, which causes the concentration dependence of the OH chemical shift, is determined by a number of factors; the activity of the OH proton, the hydrogen bonding ability of the substituent, steric hindrance of the OH proton by the substituent (for ortho substituted phenols), and the strength of any intramolecular hydrogen bond. The influence of these competing factors can be seen in guaiacol. The intermolecular hydrogen bond competes with the intramolecular interaction. For the guaiacol solution, a much higher PEG concentration than that for p-methoxyphenol is needed to increase the relaxation rate R₁ (=1/T₁) sufficient enough to make the NMR peak due to the H bonding totally disappear.²⁷ Overall, all of the active H (–OH) in phenolic compounds can be fixed by intermolecular hydrogen bond at appropriate molar ratio of PEG to phenolic compounds. Therefore, the inhibitory effects of phenolic compounds on VHG fermentation can be alleviated by supplementing PEG for improved ethanol concentration and glucose conversion. The ¹H-NMR spectral results further support our proposed detoxification mechanism.²²

Fig. 5 ¹H-NMR spectra of the OH proton of phenolic compounds in CDCl₃. (a) ¹H-NMR spectra of the OH proton of guaiacol in different molar ratio of PEG-1000 to guaiacol. (b) ¹H-NMR spectra of the OH proton of p-methoxyphenol in different molar ratio of PEG-1000 to p-methoxyphenol. (c) ¹H-NMR spectra of the OH proton of different phenolic compounds in PEG-1000 (molar ratio of PEG-1000 to phenolic compounds is 0.04 : 1).

The above results suggest that the strength of intermolecular hydrogen bonds between PEGs and the phenolic compounds depends on the structure of phenolic compounds. The ¹H-NMR spectroscopy by varying the ratio of PEGs to the phenolic compounds enabled us to differentiate the strength of the H bond donors among the phenolic compounds. In order to verify the relationship between the strength of intermolecular hydrogen bond and the structure of phenolic compounds, we further compared the OH proton chemical shifts of phenolic compounds with different number of methoxy groups in different positions on the benzene ring before or after addition PEG-1000 in CDCl₃ (Fig. 5c). The OH proton signal of m-methoxyphenol disappeared after adding PEG-1000 in 0.04 molar ratio to the phenolic compound. This observation may be attributed to the enhanced relaxation rate R₁ (=1/T₁) of m-methoxyphenol molecules due to the formation of the strongest intermolecular hydrogen bonding.²⁶ Other phenolic OH proton only showed downfield shift and ¹H-NMR signal broadening when PEG-1000 was added at the same ratio. The strength of the intermolecular hydrogen bond is dependent on the shift distance of OH proton peak from that of the phenolic compounds in the absence of PEG. The ¹H-NMR spectra of the OH proton of the phenolic compounds in the presence of PEG-1000 indicate that the order in the strength of the intermolecular hydrogen bond between phenolic compounds and PEGs is as follows: m-methoxyphenol > phenol > p-methoxyphenol > guaiacol > 2,6-dimethoxyphenol. These results have important significance for the selection of the detoxification reagents in the design of the detoxification fermentation process.

The kinetics of ethanol production in the absence and presence of PEG-1000

Mathematical models have been developed to fit the rate of product formation. There are various factors that affect the product formation kinetics including the substrate limitation, substrate inhibition, product inhibition, as well as the inhibition of toxic components.²⁸ The kinetics of VHG ethanol fermentation using Saccharomyces yeast with or without PEGs was fitted to the logistic model.⁹ In this work, the kinetics of VHG ethanol fermentation using Saccharomyces yeast with phenolic compounds was also fitted to the logistic model (4), as given below, where ν is specific rate of ethanol formation (g L⁻¹ h⁻¹), ν_max is maximum specific rate of ethanol formation (g L⁻¹ h⁻¹), P is ethanol concentration (g L⁻¹), and P_max, ethanol concentration above which cells do not produce ethanol (g L⁻¹).

Fig. 6a shows the ethanol profiles during batch fermentation for various guaiacol concentrations. It can be seen that ethanol concentration decreases drastically with the increase in guaiacol concentration. Ethanol concentration was less than 51.0 g L⁻¹ when the guaiacol content is 5.0 g L⁻¹. For ethanol production, significant lag times are observed in experiments with guaiacol. This can be attributed to the effect of guaiacol inhibition on the yeast growth. However, when 250 g L⁻¹ PEG-1000 was added to the same fermentation broth, the ethanol concentration was pronouncedly improved. Fig. 6b shows the corresponding results of batch glucose fermentation using PEG-1000 for various guaiacol concentrations. The ethanol concentration decreases slightly with increased guaiacol amounts. Even if the guaiacol content reached 5.0 g L⁻¹, the ethanol concentration was still 151.0 g L⁻¹ at 96 h. These results can be attributed to PEG-1000 alleviated guaiacol inhibition on the yeast growth. As seen from Fig. 6, the kinetic model is in good agreement with the experimental data. The kinetic model of VHG ethanol fermentation⁹ was shown to be suited for VHG ethanol fermentation with phenol (data shown in Fig. S4†).

Fig. 6 Experimental and model profiles of VHG ethanol fermentation and estimated P_max and ν_max with different concentration of guaiacol. (a) Without PEG-1000. (b) With PEG-1000. Lines represent model predictions and symbols represent experimental data. (■), 1.0 g L⁻¹; (▲), 2.0 g L⁻¹; (◀), 3.0 g L⁻¹; (♦), 4.0 g L⁻¹; (●), 5.0 g L⁻¹. (c) Correlation between P_max and guaiacol concentration. (d) Correlation between ν_max and guaiacol concentration. Fermentation conditions: 398 g L⁻¹ glucose, approximately 5 × 10⁸ cells per mL, 250 g L⁻¹ of PEG-1000, 33 °C, 160 rpm, and pH of 4.3.

Effect of phenolic compounds concentration on P_max and ν_max

The data shown in Fig. 6a and b are used to estimate the parameters P_max (g L⁻¹) (ethanol concentration above which cells do not produce ethanol) and ν_max (g L⁻¹ h⁻¹) (the highest specific rate of ethanol formation) of the modified logistic model for glucose fermentation. Based on our previous work,⁹ when there are no inhibitors in the medium, the ν_max and P_max are 8.78 g L⁻¹ h⁻¹ and 162 g L⁻¹, respectively. The estimated parameters P_max and ν_max of ethanol as a function of the guaiacol concentration are shown in Fig. 6c and d. A pronounced inhibition effect of guaiacol (1.0 g L⁻¹) on P_max (156 g L⁻¹) was observed. Beyond guaiacol concentration of 2.0 g L⁻¹, an inhibitory effect on ν_max (5.66 g L⁻¹ h⁻¹) was observed. Overall, there is a linear correlation between guaiacol concentration (x) and P_max (y) or ν_max at guaiacol concentration range of 1.0–5.0 g L⁻¹ with or without PEG-1000. The P_max and ν_max decrease with increasing guaiacol concentration. The different slopes for P_max are −24.0 and −7.40, and for ν_max are −1.03 and −0.366 without or with PEG-1000, respectively. The pronounced guaiacol inhibition effect for ethanol production is reflected by the relatively lower P_max and ν_max without PEGs compared to those with PEG-1000. The inhibition effect of phenol on P_max and ν_max showed similar trends as to that of guaiacol (data shown in Table S1†). The results indicate that the inhibition on P_max and ν_max is dependent on inhibitor (phenolic compounds) concentration used in the fermentation medium. The inhibition effect of inhibitor (phenolic compounds) on P_max and ν_max may be eliminated in whole or in part depending on the ratio of inhibitor to PEGs.

Experimental

Organisms and chemicals

The microorganism used for fermentation was Saccharomyces cerevisiae in the form of dry yeast (thermal resistant) (Angel Yeast Company Ltd, Yichang, China), which was named Angel Super-Alcohol Active Dry Yeast (starch base). The yeast was kept at 4 °C during storage, and was weighed and directly added to specified fermentation media as received right before each use. D-Glucose (C₆H₁₂O₆·H₂O, AR), ethanol (99 wt%), phenol (98 wt%), guaiacol (99 wt%), m-methoxyphenol (98 wt%), p-methoxyphenol (98 wt%), 2,6-dimethoxyphenol (98 wt%) and polyethylene glycol (PEG-1000), were purchased from Sinopharm (China). Sulphuric acid (98 wt%) and methylene blue (90 wt%) were provided by a local supplier. Ultrapure water was produced by a Milli-Q Integral 5 system. All other chemicals were of analytical quality. Dry yeast was activated in ultrapure water at 34 °C for 20 min.

Measurement of viable cell density

The yeast viability was measured in 48 h fermentation period according to the methylene-violet staining procedure.²³ Diluted sample (100 μL) containing cells was mixed with 100 μL of a methylene blue solution. After 20 min staining, the numbers of viable (living) cells and of total cells were counted under a microscope (Nikon Ci-L). The cell mortality is calculated according to the following equations:

VHG ethanol fermentation of glucose

The VHG ethanol experiments were performed in 250 mL Erlenmeyer flasks sealed with a fitted rubber stopper. The activated Saccharomyces cerevisiae cells were transferred directly without additional incubation step to prepare fermentation medium. Fermentation medium was prepared using 398 g L⁻¹ glucose, 0–5 g L⁻¹ phenolic compounds, 250 g L⁻¹ PEG-1000, the total cell count was approximately 5 × 10⁸ cells per mL in ultrapure water. No additional nutrients were applied during fermentation. The pH of all fermentation broth was adjusted to 4.3 using H₂SO₄ solution. For reference, sugar solution with no inhibitor and PEGs was used as controls. The fermentation was performed in batch mode and the temperature was controlled at 33 °C. During fermentation, the flasks were placed on a rotary shaker (ZWY-240) at 160 rpm. All the experiments were conducted in duplicate with the average and standard deviation shown in figures.

HPLC analysis

The sample of each fermentation broth was diluted with deionized water, and filtered through a 0.22 μm filter. The glucose and ethanol concentrations of fermentation samples were quantified using a high performance liquid chromatography (HPLC). An Agilent 1260 Series HPLC system equipped with a refractive index detector was used. Ion exchange columns (HPX-87H, 300 × 7.7 mm) were used in series. The column and detector temperatures were maintained at 65 °C and 50 °C, respectively, with 5 mM H₂SO₄ as the mobile phase at 0.6 mL min⁻¹. The glucose conversion was calculated based on initial glucose and consumed glucose. The concentration of ethanol was calculated based on the water volume in the fermentation broth. Data analyses were performed using the Agilent Chemstation software and Microsoft Excel. The glucose conversion, ethanol concentration were calculated according to the following equations:

NMR spectroscopy analysis

NMR spectra were measured in CDCl₃ on a 400 MHz instrument and recorded at the following frequencies: proton (¹H, 400 MHz). ¹H-NMR chemical shifts were reported in ppm using tetramethylsilane (TMS, δ (ppm) = 0.00 ppm) as the internal standard.

Fermentation model and parameter estimation

The model for ethanol productivity on glucose was a modified logistic model. The parameters in the model were evaluated by using Matlab 7.5. Its parameter estimation feature seeks to minimize the residual sum of squares between the model predicted values and the experimental values.

Conclusions

Results of this study revealed that the inhibitory effect of phenolic compounds on VHG ethanol fermentation and the toxicity to yeast cells is dependent on the activity of hydroxyl hydrogen in the phenolic compounds. The ¹H-NMR spectroscopic study of the phenolic compounds with and without PEG revealed the in situ detoxication mechanism of PEG for VHG ethanol fermentation in the presence of phenolic compounds. The predicted kinetic model was suitable to describe the rate of VHG ethanol fermentation in the presence of high toxic compounds. This study is of great importance for the understanding of the in situ detoxication process.

Acknowledgements

This work was supported by the Chinese Government “Thousand Talent” program funding; the CAS/SAFEA International Partnership Program for Creative Research Teams.

References

1. J. F. Wu, S. M. Lastick and D. M. Updegraff, Nature, 1986, 321, 887–888.
2. A. E. Farrell, R. J. Plevin, B. T. Turner, A. D. Jones, M. O'Hare and D. M. Kammen, Science, 2006, 311, 506–508.
3. M. Enquist-Newman, A. M. Faust, D. D. Bravo, C. N. Santos, R. M. Raisner, A. Hanel, P. Sarvabhowman, C. Le, D. D. Regitsky, S. R. Cooper, L. Peereboom, A. Clark, Y. Martinez, J. Goldsmith, M. Y. Cho, P. D. Donohoue, L. Luo, B. Lamberson, P. Tamrakar, E. J. Kim, J. L. Villari, A. Gill, S. A. Tripathi, P. Karamchedu, C. J. Paredes, V. Rajgarhia, H. K. Kotlar, R. B. Bailey, D. J. Miller, N. L. Ohler, C. Swimmer and Y. Yoshikuni, Nature, 2014, 505, 239–243.
4. A. T. Hendriks and G. Zeeman, Bioresour. Technol., 2009, 100(1), 10–18.
5. N. Wei, J. Quarterman, S. R. Kim, J. Cate and H. Y. Jin, Nat. Commun., 2013, 4, 2580.
6. F. W. Bai, W. A. Anderson and M. Moo-Young, Biotechnol. Adv., 2008, 26, 89–105.
7. C. G. Liu, N. Wang, Y. H. Lin and F. W. Bai, Biotechnol. Biofuels, 2012, 5(1), 61.
8. P. Chan-u-tit, L. Laopaiboon, P. Jaisil and P. Laopaiboon, Energies, 2013, 6, 884–899.
9. X. Liu, W. Xu, C. Zhang, P. Yan, S. Jia, Z. Xu and Z. C. Zhang, RSC Adv., 2014, 4, 52299–52306.
10. C. Felby, L. G. Thygesen, J. B. Kristensen, H. Jørgensen and T. Elder, Cellulose, 2008, 15, 703–710.
11. S. Helle, D. Cameron, J. Lam, B. White and S. Duff, Enzyme Microb. Technol., 2003, 33, 786–792.
12. E. Palmqvist and B. Hahn-Hȁgerdal, Bioresour. Technol., 2000, 74, 25–33.
13. R. Koppram, E. Tomas-Pejo, C. Xiros and L. Olsson, Trends Biotechnol., 2014, 32, 46–53.
14. H. Jorgensen, J. Vibe-Pedersen, J. Larsen and C. Felby, Biotechnol. Bioeng., 2007, 96(5), 862–870.
15. T. Y. Mills, N. R. Sandoval and R. T. Gill, Biotechnol. Biofuels, 2009, 2(1), 26.
16. H. B. Klinke, A. B. Thomsen and B. K. Ahring, Appl. Microbiol. Biotechnol., 2004, 66(1), 10–26.
17. J. Buchert, J. Puls and K. Poutanen, Appl. Biochem. Biotechnol., 1989, 20–21, 309–318.
18. T. A. Clark and K. L. Macki, J. Chem. Technol. Biotechnol., 1984, 34, 101–110.
19. S. I. Mussatto and I. C. Roberto, Biotechnol. Lett., 2001, 23, 1681–1684.
20. D. Ludwig, M. Amann, T. Hirth, S. Rupp and S. Zibek, Bioresour. Technol., 2013, 133, 455–461.
21. P. B. Dhamole, B. Wang and H. Feng, J. Chem. Technol. Biotechnol., 2013, 88, 1744–1749.
22. X. Liu, W. Xu, L. Mao, C. Zhang, P. Yan, Z. Xu and Z. C. Zhang, Sci. Rep., 2016, 6, 20361.
23. K. A. Smart, K. M. Chambers, I. Lambert, C. Jenkins and C. A. Smart, J. Am. Soc. Brew. Chem., 1999, 57, 18–23.
24. Y. Zhang, X. Xu, Y. Zhang and J. Li, Biotechnol. Bioprocess Eng., 2011, 16, 930–936.
25. E. Arunan, G. R. Desiraju, R. A. Klein, J. Sadlej, S. Scheiner, I. Alkorta, D. C. Clary, R. H. Crabtree, J. J. Dannenberg, P. Hobza, H. G. Kjaergaard, A. C. Legon, B. Mennucci and D. J. Nesbitt, Pure Appl. Chem., 2011, 83, 1619–1633.
26. K. Yoshida, A. Kitajo and T. Yamaguchi, J. Mol. Liq., 2006, 125, 158–163.
27. R. J. Abraham and M. Mobli, Magn. Reson. Chem., 2007, 45, 865–877.
28. P. Thangprompan, A. Thanapimmetha, M. Saisriyoot and L. Laopaiboon, Appl. Biochem. Biotechnol., 2013, 171, 294–314.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20693j

---