Biotechnol. J.

2014, 9

DOI 10.1002/biot.201300367

www.biotechnology-journal.com

Research Article

Flocculating Zymomonas mobilis is a promising host to be engineered for fuel ethanol production from lignocellulosic biomass
Ning Zhao1, Yun Bai1, Chen-Guang Liu1, Xin-Qing Zhao1, Jian-Feng Xu2 and Feng-Wu Bai1,3
1 School

of Life Sciences and Biotechnology, Dalian University of Technology, Dalian, China Biosciences Institute, Arkansas State University, AR, USA 3 School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China
2 Arkansas

Whereas Saccharomyces cerevisiae uses the Embden-Meyerhof-Parnas pathway to metabolize glucose, Zymomonas mobilis uses the Entner-Doudoroff (ED) pathway. Employing the ED pathway, 50% less ATP is produced, which could lead to less biomass being accumulated during fermentation and an improved yield of ethanol. Moreover, Z. mobilis cells, which have a high specific surface area, consume glucose faster than S. cerevisiae, which could improve ethanol productivity. We performed ethanol fermentations using these two species under comparable conditions to validate these speculations. Increases of 3.5 and 3.3% in ethanol yield, and 58.1 and 77.8% in ethanol productivity, were observed in ethanol fermentations using Z. mobilis ZM4 in media containing ~100 and 200 g/L glucose, respectively. Furthermore, ethanol fermentation by the flocculating Z. mobilis ZM401 was explored. Although no significant difference was observed in ethanol yield and productivity, the flocculation of the bacterial species enabled biomass recovery by cost-effective sedimentation, instead of centrifugation with intensive capital investment and energy consumption. In addition, tolerance to inhibitory byproducts released during biomass pretreatment, particularly acetic acid and vanillin, was improved. These experimental results indicate that Z. mobilis, particularly its flocculating strain, is superior to S. cerevisiae as a host to be engineered for fuel ethanol production from lignocellulosic biomass.

Received 23 AUG 2013 Accepted 31 OCT 2013 Accepted article online 06 NOV 2013

Keywords: Biofuels Ethanol fermentation White/industrial biotechnology Yield Zymomonas mobilis

1 Introduction
Global concern about the sustainable supply of crude oil and the environmental deterioration caused by the overconsumption of petroleum-derived products, particularly transportation fuels, makes the development of alternatives essential. Biofuels from renewable biomass resources represent one alternative, and have been intensively studied since the 1970s following the oil crisis [1, 2]. At present, fuel ethanol is the dominating biofuel, and is produced mainly by Saccharomycescerevisiae strains via the Embden-Meyerhof-Parnas (EMP) pathway. In this pathway two moles of ATP are produced per mole of glucose consumed, leading to a significant accumulation of biomass at the expense of ethanol yield [3]. In contrast to S. cerevisiae, Zymomonas mobilis metabolizes glucose

Correspondence: Prof. Feng-Wu Bai, School of Life Sciences and Biotechnology, Dalian University of Technology, 2 Linggong Rd., Dalian 116023, China E-mail: fwbai@sjtu.edu.cn Additional correspondence: Dr. Xin-Qing Zhao, School of Life Sciences and Biotechnology, Dalian University of Technology, 2 Linggong Rd., Dalian 116023, China E-mail: xqzhao@dlut.edu.cn Abbreviations: DCW, dry cell weight; ED, Entner-Doudoroff; EMP, EmbdenMeyerhof-Parnas; FBRM, focused-beam reflectance measurement; HG, high gravity; HMF, hydroxymethylfurfural; LG, low gravity; ORP, oxidoreduction potential; VHG, very high gravity

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnology Journal
www.biotechnology-journal.com

Biotechnol. J. 2014, 9

via the Entner-Doudoroff (ED) pathway, in which only one mole of ATP is produced per mole of glucose consumed. Consequently, less biomass is accumulated during ethanol fermentation, which in theory could improve ethanol yield [4]. In addition, the relatively small bacterial cells provide a high specific surface area for sugar uptake [5], which could improve ethanol productivity. Although these advantages have been speculated, Z.mobilis has never been used for fuel ethanol production industrially due to its narrow substrate spectrum of glucose, fructose and sucrose only [4]. Glucose is too costly for fuel ethanol production, and starch-based feedstocks such as corn meal and cassava chips are used, which are liquefied by amylase into dextrins, followed by the hydrolysis of the dextrins by glucoamylase. While glucose is the major sugar produced during the saccharification process, other sugars such as maltose and maltotriose are also presented in the mash, which can be fermened by S.cerevisiae [6] but not Z.mobilis. Ethanol yield from fructose and sucrose is much lower in Z.mobilis, since a significant level of sorbitol is accumulated during the fermentation [4]. Therefore, Z.mobilis is not suitable for fuel ethanol production from sugar- and starch-based feedstocks. Sugar-based feedstocks such as sugarcane juice or molasses are geographically limited to tropical and subtropical areas such as Brazil, and starch-based feedstocks are major food sources. Neither sugar- nor starch-based feedstocks are regarded as sustainable for large-scale production of fuel ethanol. Lignocellulosic biomass resources, particularly agricultural residues such as corn stover and rice and wheat straw, are alternative feedstocks for sustainable production of fuel ethanol [7]. The major carbohydrates of lignocellulosic biomass are hemicelluloses and cellulose, which can be hydrolyzed into hexose and pentose sugars. Since glucose is the only hexose sugar released from the hydrolysis of the cellulose component, the disadvantage of the narrow substrate spectrum with Z.mobilis would not be problematic for ethanol production from lignocellulosic biomass. In addition, Z.mobilis can be readily engineered to provide pentose metabolic pathways for co-fermentation of hexose and pentose sugars, which has been explored since the 1990s [8, 9]. The completion of the genome sequence of Z.mobilis and elucidation of its metabolic network [10, 11] will help in developing robust strains from this species for fuel ethanol production from lignocellulosic biomass. Recently, DuPont established a pilot plant for cellulosic ethanol production in Tennessee, USA, in which a recombinant Z. mobilis strain is being tested. Although the progress seems exciting, how significant the improvement on ethanol yield and productivity by Z.mobilis compared to S.cerevisiae would be has not yet been addressed under comparable conditions. We performed ethanol fermentations using Z.mobilis and S.cerevisiae with media containing same glucose concentrations. Moreover, the advantages of flocculating Z. mobilis were further

explored, since the morphological change not only facilitates biomass recovery by cost-effective sedimentation instead of centrifugation, but also enhances tolerance to environmental stresses such as toxic byproducts released during the pretreatment of lignocellulosic biomass.

2 Materials and methods

2.1 Strains, media and seed culture
Z.mobilis ZM4 (ATCC 31821), which has been intensively studied for ethanol fermentation, was used in the study. The industrial strain S.cerevisiae ADY (Angel Yeast CO., LTD, China) used for fuel ethanol production was selected as the control to study the improvements in ethanol yield and productivity of Z.mobilis under comparable fermentation conditions. To address the challenge in biomass recovery with Z. mobilis, which needs high-speed centrifugation involving heavy capital investment on centrifuges and intensive energy consumption on centrifuge running, the flocculating strain Z.mobilis ZM401 (ATCC 31822) was selected, and compared with Z.mobilis ZM4 to highlight its advantages in biomass recovery by sedimentation. In addition, the tolerance of the Z. mobilis strains to inhibitory byproducts released during the pretreatment of lignocellulosic biomass was also studied. Medium comprising 20 g/L glucose, 10 g/L yeast extract and 2g/L KH2PO4 was used for the seed culture of Z. mobilis ZM4 and ZM401 [12]. For S. cerevisiae ADY propagation, medium comprising 30 g/L glucose, 5 g/L yeast extract and 3g/L peptone was used [13]. Media containing 110 or 210g/L glucose, respectively, were used for ethanol fermentations so that after inoculation, initial glucose within the fermentor was controlled at ~100g/L for low-gravity (LG) fermentation and ~200g/L for high-gravity (HG) fermentation. The media were supplemented with 10 g/L yeast extract for ethanol fermentation by Z. mobilis ZM4 and ZM401, and 5 g/L each of yeast extract and peptone for ethanol fermentation by S.cerevisiae ADY. In addition, very-high-gravity (VHG) medium containing 300g/L glucose was used for feeding the fedbatch fermentation by Z. mobilis ZM4. All media were sterilized at 121C for 30min before inoculation. A loop-full of stock culture was inoculated into 250mL Erlenmeyer flask containing 100mL medium for seed culture. Although Z. mobilis produces ethanol through the ED pathway under anaerobic conditions, this species also possesses an oxidative phosphorylation pathway for growth, and no significant difference was observed in its metabolite profiles at the exponential growth phase [5, 14]. Thus, seed cultures for the Z. mobilis ZM4 and ZM401 strains and S. cerevisiae ADY were performed overnight at 30C and 150 rpm to the mid-exponential growth phase under microaerobic conditions by sealing the flasks with regular gauze plugs.

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnol. J. 2014, 9
www.biotecvisions.com

2.2 Batch and fed-batch ethanol fermentations

The harvested seed culture was inoculated into a fermentor (KoBioTech KF-2.5L, Korea) containing 1L medium for batch fermentation. The inoculation level was controlled at 20% for Z.mobilis ZM4 and ZM401, and 10% for S.cerevisiae ADY. The fed-batch fermentation by Z.mobilis ZM4 was initiated by inoculating the fermentor containing 0.5L LG medium supplemented with 20g/L yeast extract, and 0.5L VHG medium, divided into two lots, was fed into the fermentator when glucose was depleted. Ethanol fermentations were performed at 30C and 150rpm without aeration, and the pH was maintained automatically at 4.5 by adding 2N NaOH.

2.3 Tolerance to inhibitory byproducts

Furfural, hydroxymethylfurfural (HMF), acetic acid and vanillin were added to the LG medium to study their impact on ethanol fermentations by Z.mobilis ZM401 and ZM4. The amount of supplementation was decided based on their concentrations in the hydrolysate of corn stover donated by COFCO, a large fuel ethanol producer in China and also a pioneer in developing bioethanol from lignocellulosic biomass. Three concentrations of each supplement were selected: 0.57, 0.86 and 1.15 g/L for furfural; 0.62, 0.93, and 1.24 g/L for HMF; 6.30, 8.40 and 10.50 g/L for acetic acid; and 0.50, 0.75 and 1.00 g/L for vanillin. The ethanol fermentations were carried out in flasks at 30C and 150rpm under microaerobic conditions. Furthermore, 0.57g/L furfural, 0.62g/L HMF, 6.30g/L acetic acid and 0.50 g/L vanillin were added simultaneously to the LG medium to study their synergistic effect on ethanol fermentations by Z.mobilis ZM401 and ZM4 in the fermentor. When acetic acid was added, the pH of the media was adjusted to 6.0 using 2N KOH, which was similar to overliming neutralization by Ca(OH)2 that has been recommended for industrial applications [15]. Finally, for ethanol fermentation with Z.mobilis ZM401 and ZM4, we tested a COFCO hydrolysate of corn stover that had been prepared using the dilute-acid steam-explosion pretreatment and hydrolysis of the cellulosic component by the cellulases Ctec3 (developed by Novozymes), and contained 80.0g/L glucose, 33.0g/L xylose, 0.49g/L furfural, 0.28g/L HMF and 0.66g/L acetic acid. Triplicate experiments were performed for all experiments, and statistical errors are reported with the experimental results.

pyruvic acid were quantified by HPLC (Waters 2707, Waters 1525 binary HPLC pump;detectors: Waters 2998 photodiode array detector and refractive index detector; column: Bio-Rad HPX-87H ion exclusion column, 300 7.8 mm; eluant: 0.005mol/L H2SO4 at a flow rate of 0.4mL/min). Another byproduct, acetoin, of ethanol fermentation by Z. mobilis was analyzed by GC (Agilent 6890 series GC system) equipped with flame ionization detector and capillary column (0.25mm diameter), with isobutanol as the internal standard. The oxidoreduction potential (ORP) was monitored in situ during ethanol fermentations using an ORP electrode (Pt4805-DPAS-SC-K8 12 mm250 mm, Mettler Toledo). For ethanol fermentation using flocculating Z. mobilis ZM401, a focused-beam reflectance measurement (FBRM) probe was inserted into the fermentor to monitor the size distribution of the bacterial flocs [16].

3 Results and discussion

3.1 Ethanol fermentations by Z. mobilis ZM4 and S. cerevisiae ADY
The time courses of glucose consumption, ethanol and biomass accumulation, and ORP profiles during ethanol fermentations by Z.mobilis ZM4 and S.cerevisiae ADY were monitored and recorded (Fig.1). Major byproducts produced during the fermentations were analyzed (Table1) to evaluate the ethanol fermentation processes. Only 2.62g(DCW)/L biomass was accumulated during ethanol fermentation by Z.mobilis ZM4 under the LG fermentation condition, much lower than the 7.17g(DCW) L1 accumulated during ethanol fermentation by S.cerevisiae ADY. An increase of 3.5% in ethanol yield was achieved. Moreover, Z. mobilis ZM4 consumed glucose much faster, and ethanol fermentation was completed in 12.5h (compared to the 20.0h required by S.cerevisiae ADY); ethanol productivity was increased by 58.1%. However, under the HG fermentation conditions, a lag phase of ~20h was observed with Z.mobilis ZM4, during which glucose uptake was extremely slow, and almost no ethanol was produced. After that, exponential growth was initiated and glucose utilization was facilitated. No similar phenomenon was observed with ethanol fermentation by S. cerevisiae ADY, indicating that Z. mobilis ZM4 was not as tolerant to osmotic stress as S.cerevisiae ADY. As a result, although only 3.30g(DCW)/L biomass was accumulated by Z. mobilis ZM4, compared to 7.13g(DCW)/L using S.cerevisiae ADY, no significant difference was observed in ethanol yield and productivity for the two species. From the viewpoint of bioprocess engineering, fedbatch fermentation that is initiated with LG medium and followed by feeding VHG medium can effectively alleviate substrate inhibition. This strategy was thus employed to

2.4 Analytical methods

Dry cell weight (DCW) was used for biomass measurement. Glucose and ethanol was analyzed as described previously [16]. Byproducts of ethanol fermentations by Z. mobilis and S. cerevisiae, including glycerol, acetic acid, lactic acid, citric acid, succinic acid, sorbitol and

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnology Journal
www.biotechnology-journal.com

Biotechnol. J. 2014, 9

Figure 1. Ethanol fermentations by S. cerevisiae ADY () and Z. mobilis ZM4 (o) and ZM401 () in media containing ~100 g/L glucose (AD) and ~200 g/L glucose (EH). Triplicate experiments were performed, and their average values are illustrated. The error bars represent standard deviations. Arrows indicate the feeding of the concentrated medium in the fed-batch fermentation. Error analysis indicates the significance of the experimental results.

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnol. J. 2014, 9
www.biotecvisions.com

Table 1. Ethanol fermentation results of Z. mobilis ZM4 and S. cerevisiae ADYa)

Fermentation parameters

LG fermentation with ~100 g/L glucose S. cerevisiae ADY Batch Z. mobilis ZM4 97.00.71 1.710.00 0.250.00 0.140.03 46.60.66 2.870.08 0.300.10 0.500.10 0.030.00 0.500.08 0.470.30 2.700.28 0.860.20 0.780.50 0.090.02 0.030.00 96.9 44.9 2.62 12.5 3.59 0.08 2.96 1.37 0.03 0.46 60.4 30.2 7.3 2.1

HG fermentation with ~200 g/L glucose S. cerevisiae ADY Batch 2020.09 1.250.12 0.300.03 4.000.06 94.01.71 7.430.21 4.000.88 0.960.07 0.220.07 5.400.10 0.100.03 1.100.09 0.120.06 0.100.03 0.030.00 198 92.3 7.13 50.0 1.85 0.02 0.56 0.26 0.04 0.47 60.6 30.3 6.1 3.0 Z. mobilis ZM4 Batch 2000.00 1.540.41 0.270.02 4.000.03 93.52.12 3.570.71 0.600.20 0.000.01 0.160.05 1.200.20 0.210.02 3.600.10 0.500.21 0.860.08 0.070.02 0.040.00 196 92.0 3.30 50.0 1.84 0.02 1.19 0.56 0.02 0.47 61.2 30.6 4.2 4.0 Fed-batch 1041.41 1.640.22 0.300.71 0.170.01 1000.02 5.800.14 0.550.00 0.060.01 0.090.02 0.690.07 0.120.02 1.100.50 0.350.01 0.550.05 0.040.02 0.030.00 204 98.6 5.50 30.0 3.29 0.03 1.24 0.60 0.03 0.48 63.0 31.5 2.3 3.2

Initial glucose (g/L) Initial ethanol (g/L) Initial biomass (g/L) Residual glucose (g/L) Ethanol (g/L) Biomass (g/L) Major byproducts (g/L) Glycerol Pyruvate Acetate Lactate Citric acid Succinate Sorbitol Acetoin Ethyl acetate Isoamyl alcohol Acetaldehyde Glucose consumed (g/L) Ethanol produced (g/L) Biomass accumulated (g/L) Fermentation time (h) Ethanol productivity (g/L per h) Specific growth rate (h1) Specific glucose uptake rate (h1) Specific ethanol productivity (h1) Biomass yield Ethanol yield Carbon flux distribution (%) Ethanol CO2 Byproducts detected Others

1020.71 1.450.09 0.250.01 0.140.00 46.90.06 7.430.21 2.100.20 1.300.20 2.700.30 1.700.02 1.700.03 0.140.08 0.100.04 0.030.00 102 45.4 7.17 20.0 2.27 0.05 0.71 0.32 0.07 0.45 58.0 29.0 8.5 4.5

a) For the fed-batch process, the fermentation was initiated with the LG medium containing ~100 g/L glucose, and concentrated medium containing ~300 g/L glucose was fed, as illustrated in Fig. 1, to guarantee that the amount of glucose fermented was the same as that in the HG fermentation. Concentrations of glucose, ethanol, biomass and byproducts from triplicate experiments were measured and averages with standard deviations recorded. All other parameters were calculated based on niM(Byproduct)i 2MEthanol these averages. Molar carbon flux to ethanol and detected byproducts was calculated as: 100 %, respectively, where 100 % and 6MGlucose 6MGlucose MEthanol, M(Byproduct)i and MGlucose represent molar concentrations of ethanol, byproduct i produced and glucose consumed, and ni represents the number of carbon atoms in the molecule of byproduct i. Molar carbon flux to CO2 associated with ethanol production was estimated as 50% of that to ethanol production. All remaining carbon flux was directed to undetected byproducts, biomass and CO2 associated with all byproduct formation and biomass synthesis.

address the osmotic stress exerted on Z.mobilis ZM4 by the HG medium under batch fermentation conditions. The VHG medium was fed at 10 and 20h, when glucose within the fermentor was almost depleted. As can be seen, no lag phase for cell growth and ethanol fermentation was observed for Z.mobilis ZM4 under the fed-batch conditions, and the fermentation time was shortened to 30 h, from 50 h required for the batch fermentation. Although more biomass (5.50 g(DCW)/L) was accumulated with the fed-batch fermentation process, Z. mobilis

ZM4 produced 98.6 g/L ethanol from 204.0 g/L glucose consumed, increasing the ethanol yield to 0.483, equivalent to 94.6% of the theoretical maximum of 0.511. This represents an increase of 3.3% in ethanol yield compared to that achieved with ethanol fermentation by S.cerevisiae ADY. This result indicates that the fed-batch process is an effective strategy for overcoming the disadvantage of Z.mobilis ZM4 in tolerating osmotic stress, and demonstrates its advantage in improved ethanol yield and productivity for more efficient fuel ethanol production.

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnology Journal
www.biotechnology-journal.com

Biotechnol. J. 2014, 9

ORP reflects the net balance of reducing and oxidizing power within culture and fermentation systems, and thus directly reflects the status of cell growth and product formation, particularly under microaerobic and anaerobic fermentation conditions [17]. Since Z.mobilis generates less ATP and reducing power from the ED pathway, the ORP level during ethanol fermentation by this species should be higher than that monitored with ethanol fermentation by S.cerevisiae, particularly under the HG fermentation condition. This was validated by the quantitative comparison of the ORP profiles monitored during ethanol fermentations by Z.mobilis ZM4 and S.cerevisiae ADY shown in Figs.1D and H.

3.2 Major byproducts formation

In addition to biomass accumulation, byproducts formation is another factor that compromises ethanol yield. Whereas the major byproducts produced during ethanol fermentation by S. cerevisiae have been characterized, including glycerol, ethyl acetate, acetaldehyde, and higher alcohols [18], reports on the byproducts formed during ethanol fermentation by Z. mobilis are very limited. Recently, metabolic profiling was developed for Z.mobilis, and acetate, lactate, citrate, succinate and acetoin were predicted as potential major byproducts [14]. To validate these speculations and provide more insights on the improvement of ethanol yield observed with Z. mobilis ZM4, the major byproducts produced during ethanol fermentation by the two strains were analyzed, and illustrated in Table1.

In addition to the glycerol, pyruvate, lactate, citric acid, succinate, isoamyl alcohol and acetaldehyde detected in ethanol fermentation by S. cerevisiae ADY, two unique byproducts, sorbitol and acetoin, were produced during ethanol fermentation by Z. mobilis ZM4, but no ethyl acetate was detected. Mass balance was further performed for evaluating carbon molar flux distributions during ethanol fermentations by Z. mobilis ZM4 and S.cerevisiae ADY. Under the batch LG fermentation condition, carbon molar flux to ethanol and CO2 during ethanol production was 87.0% for the EMP pathway in S.cerevisiae and 90.6% for the ED pathway in Z.mobilis. Although the improvement in carbon molar flux to ethanol and CO2 was not significant with Z.mobilis ZM4 under the batch HG fermentation condition, because of the generation of more byproducts such as sorbitol to address the osmotic stress [19], this disadvantage was overcome effectively under the fed-batch HG fermentation condition, and a carbon molar flux directed to ethanol and CO2 was as high as 94.5%. These experimental results clearly confirmed the improved ethanol production yield of the bacterial species.

3.3 Ethanol fermentation by the flocculating Z. mobilis ZM401

The flocculation of microbial cells not only makes them self-immobilized within fermentors creating a high cell density and improving productivity (demonstrated in ethanol fermentation by the flocculating yeast [20]), it also facilitates biomass recovery since cell flocs can settle

Figure 2. Flocculation of Z. mobilis. (A) Morphology of ZM401 observed with SEM. (B) Chord length distribution of the bacterial flocs detected by FBRM during ethanol fermentation in medium containing ~200 g/L glucose. Li: Chord length at the midpoint of channel bin i (m), and ti: chord counts per second. (C, D) Sedimentation performance of the flocculating and regular strains ZM401 and ZM4.

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnol. J. 2014, 9
www.biotecvisions.com

down quickly from fermentation broth. This has been in practice in breweries for a long time [21]. Therefore, ethanol production by Z. mobilis ZM401, a flocculating strain, was explored and compared with that observed with the regular Z.mobilis ZM4. The results indicated no significant difference between the two bacterial strains, except that, under the HG fermentation conditions, Z. mobilis ZM401 completed ethanol fermentation ~5 h earlier (data not shown). This might be due to the improved ethanol tolerance caused by the morphological change, which was observed using scanning electron microscopy (SEM) and characterized by the FBRM system (Figs.2A and B). The sedimentation performance of the bacterial flocs was further qualitatively characterized (Fig. 2C) and compared to that observed with the nonflocculating strain ZM4 (Fig. 2D). The findings clearly indicate that the flocculation of Z. mobilis ZM401 can make biomass recovery by sedimentation viable. Sedimentation is the most economically competitive strategy from the viewpoint of engineering, compared to centrifugation that requires heavy capital investment for the centrifuges and intensive energy consumption for centrifuge operation.

3.4 Tolerance to inhibitory byproducts

Theoretically, tolerance to environmental stresses can be improved when microbial cells flocculate, due to enhanced quorum sensing [22]. The flocculation of Z. mobilis is expected to improve its tolerance to inhibitory byproducts presented in the hydrolysate of lignocellulosic biomass, such as furfural and HMF from degradation of pentose and hexose sugars, acetic acid from acetylated hemi-

celluloses, and vanillin from lignin [23]. These major inhibitory byproducts were added to the LG medium containing ~10% glucose to study their effect on ethanol fermentations by the flocculating Z.mobilis ZM401 and the regular Z.mobilis ZM 4. The sugar content of the LG medium was equivalent to the total sugars in the hydrolysate of corn stover donated by COFCO for this research. Table2 summarizes the experimental results. While no difference was observed with fermentation performance under the furfural and HMF supplementation conditions, the flocculation of Z.mobilis did improve its tolerance to the inhibition exerted by acetic acid and vanillin, two major byproducts released from side reactions associated with the degradation of hemicelluloses and lignin [23]. More glucose was consumed by the flocculating Z.mobilis ZM401 (97.4 and 63.0g/L) than the non-flocculating Z.mobilis ZM4 (67.5 and 34.0g/L) when 10.5g/L acetic acid and 1.00g/L vanilin were supplemented. Consequently, more ethanol was produced by Z. mobilis ZM401 than that by Z.mobilis ZM4 (46.2 and 25.3g/L vs. 23.9 and 11.3g/L). The improved ethanol production was in accordance with the increased biomass accumulation, 1.30 and 0.90g(DCW)/L by Z.mobilis ZM401 vs. 1.08 and 0.60g(DCW)/L by Z.mobilis ZM4. Moreover, fermentation time was shorted from 48 to 30 h for Z. mobilis ZM401 when 10.5g/L acetic acid was supplemented. The synergistic effects of furfural, HMF, acetic acid and vanillin on ethanol fermentations by Z.mobilis ZM401 and ZM4 are illustrated in Figs.3AC. Although no significant impact on cell growth was observed, a significant improvement in tolerance to these byproducts was observed with Z.mobilis ZM401 due to the morphological change associated with the flocculation. As can be seen,

Table 2. Ethanol fermentations by Z. mobilis ZM4 and ZM401 from the simulated medium supplemented with furfural, 5-HMF, acetic acid and vanillina)

Glucose consumed (g/L) ZM4 Furfural (g/L) 0.69 0.92 1.15 HMF (g/L) 0.62 0.93 1.24 Acetic acid (g/L) 6.30 8.40 10.50 Vanillin (g/L) 0.50 0.75 1.00 99.90.00 99.60.30 97.80.17 99.90.00 99.90.00 97.70.16 99.90.00 97.60.19 67.51.00 104.00.00 104.00.00 34.02.00 ZM401 99.90.00 99.80.10 97.90.15 99.90.00 99.90.00 97.90.20 99.90.00 99.90.00 97.40.13 104.00.00 104.00.00 63.03.00

Ethanol produced (g/L) ZM4 42.40.48 41.42.50 41.70.85 47.60.92 41.90.56 39.80.84 46.00.72 44.00.99 23.93.67 47.33.00 43.21.20 11.13.40 ZM401 47.00.30 45.31.24 45.12.30 48.00.50 46.40.42 44.60.80 48.90.38 48.30.24 46.20.98 49.61.30 47.43.00 25.35.30

Biomass accumulated (g(DCW)/L) ZM4 ZM401 1.450.21 1.200.09 1.250.07 2.050.20 1.900.05 1.750.08 1.900.04 1.850.15 1.080.10 1.560.09 1.200.08 0.600.10 1.400.09 1.100.07 1.100.15 2.000.05 1.800.09 1.700.10 1.800.05 1.600.09 1.300.20 1.550.07 1.100.06 0.900.04

Fermentation time (h) ZM4 ZM401 24 24 24 24 24 24 24 24 48 24 24 48 24 24 24 24 24 24 24 24 36 24 24 48

a) All data were collected from triplicate experiments and averages with standard errors were recorded.

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnology Journal
www.biotechnology-journal.com

Biotechnol. J. 2014, 9

Figure 3. Ethanol fermentations by Z. mobilis ZM4 (o) and ZM401 () in the simulated medium containing ~100 g/L glucose, supplemented with 0.5 g/L vanillin, 0.5 mL/L furfural, 0.5 mL/L 5-HMF and 6.0 mL/L acetic acid (AC), and in the hydrolysate of corn stover donated by COFCO (D, E). Triplicate experiments were performed, and average values are illustrated. The error bars represent standard deviations.

50.1 g/L ethanol was produced by Z. mobilis ZM401 at 48h compared to 48.4g/L ethanol produced at 54h by Z. mobilis ZM4. This indicates a significant increase in the ethanol productivity by Z.mobilis ZM401, from 0.89 to 1.04g/L/h, an increase of 16.8%. Similar results were also observed during ethanol fermentation from the hydrolysate of corn stover donated by COFCO (Figs.3D and E). Developing robust strains is a prerequisite for economic fuel ethanol production from lignocellulosic biomass. To make complete use of pentose and hexose sugars released from hemicelluloses and cellulose, two major strategies have been developed (depending on the type of host strains S.cerevisiae or Z.mobilis) for engineering

them with pentose metabolic pathways [24]. Until now, more effort has been made in engineering S.cerevisiae, with genes encoding xylose reductase and xylitol dehydrogenase from Pichia stipitis to enable xylose to be metabolized to xylulose. The xylulose is then directed to the native pentose phosphate pathway, facilitated by the overexpression of xylulokinase gene, for ethanol production by the EMP pathway. However, an imbalance of cofactors leads to an accumulation of xylitol, which significantly compromises ethanol yield [2527]. In contrast, only one enzyme xylose isomerase is needed to be engineered into Z.mobilis for xylose to be metabolized to xylulose [8, 28], which can ultimately solve this bottleneck with strain development. Taking into account high

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnol. J. 2014, 9
www.biotecvisions.com

ethanol yield and productivity, Z. mobilis could outperform S.cerevisiae for ethanol production from lignocellulosic biomass. In addition, the flocculating strain of this species presents more advantages in cost-effective biomass recovery and improved tolerance to inhibitory byproducts released during biomass pretreatment. With the completion of the genome sequencing of Z. mobilis ZM401 [29], more fundamental knowledge on its flocculation and improved tolerance to toxic byproducts will be gained through comparative genome analysis. This could thus provide a better host strain for genetic modifications for more efficient production of fuel ethanol from lignocellulosic biomass.

Funding for this research was provided in part by the National Natural Science Foundation of China, grant number 21276038, and the National High-Tech R&D Program, grant number 2012AA021205. The corn stover hydrolysate used in this research was kindly donated by Dr. F. Li at COFCO, which is gratefully appreciated. The authors declare no conflict of interest.

5 References
[1] Bungay, H. R., Biomass refining. Science 1982, 218, 643646. [2] Ragauskas, A. J., Williams, C. K., Davison, B. H., Britovsek, G. et al., The path forward for biofuels and biomaterials. Science 2006, 311, 484489. [3] Bai, F. W., Anderson W. A., Moo-Young M., Ethanol fermentation technologies from sugar and starch feedstocks. Biotechnol. Adv. 2008, 26, 89105. [4] Sprenger, G. A., Carbohydrate metabolism in Zymomonasmobilis: A catabolic highway with some scenic routes. FEMS Microbiol. Lett. 1996, 145, 301307. [5] Swings, J., De Ley, J., The biology of Zymomonas. Bacteriol. Rev. 1977, 41, 146. [6] Kelsall, D. R, Lyons T. P., Grain dry milling and cooking procedures: Extracting sugars in preparation for fermentation, in: Jacques, K. A., Lyons, T. P., Kelsall, D. R. (Eds.), The alcohol textbook, Nottingham University Press, 2003, pp. 921. [7] Tilman, D., Socolow, R., Foley, J. A., Hill, J. et al., Energy. Beneficial biofuels the food, energy, and environment trilemma. Science 2009, 325, 270271. [8] Zhang, M., Eddy, C., Deanda, K., Finkelstein M. et al., Metabolic engineering of a pentose metabolism pathway in ethanologenic Zymomonasmobilis. Science 1995, 267, 240243. [9] Deanda, K., Zhang, M., Eddy, C., Finkelstein M. et al., Development of an arabinose-fermenting Zymomonasmobilis strain by metabolic pathway engineering. Appl. Environ. Microbiol. 1996, 62, 4465 4470. [10] Seo, J. S., Chong, H., Park, H. S., Yoon, K. O. et al., The genome sequence of the ethanologenic bacterium Zymomonasmobilis ZM4. Nat. Biotechol. 2005, 23, 6368. [11] Lee, K. Y., Park, J. M., Kim, T. Y., Yun, H. et al., The genome-scale metabolic network analysis of Zymomonas mobilis ZM4 explains physiological features and suggests ethanol and succinic acid production strategies. Microb. Cell Fact. 2010, 9, 94. [12] Goodman, A. E., Rogers, P. L., Skotnicki, M. L., Minimal medium for isolation of auxotrophic Zymomonas mutant. Appl. Environ. Microbiol. 1982, 44, 496498. [13] Ge, X. M., Zhang, L., Bai, F. W., Impacts of yeast floc size distributions on their observed rates for substrate uptake and product formation. Enzyme Microb. Technol. 2006, 39, 289295. [14] Yang, S., Tschaplinski, T. J., Engle, N. L., Carroll, S. L., Transcriptomic and metabolomic profiling of Zymomonasmobilis during aerobic and anaerobic fermentations. BMC Genomics 2009, 10, 34. [15] Jennings, E. W., Schell, D. J., Conditioning of dilute-acid pretreated corn stover hydrolysate liquors by treatment with lime or ammonium hydroxide to improve conversion of sugars to ethanol. Bioresour. Technol. 2011, 102, 12401245. [16] Ge, X. M., Zhao, X. Q., Bai, F. W., Online monitoring and characterization of flocculating yeast cell flocs during continuous ethanol fermentation. Biotechnol. Bioeng. 2005, 90, 523531.

4 Concluding remarks
For the first time, ethanol fermentations by S.cerevisiae and Z.mobilis have been studied under comparable conditions to validate the hypotheses that Z.mobilis would lead to an improved ethanol yield due to the unique ED pathway with less ATP produced in the bacterial species, and enhanced ethanol productivity due to higher glucose uptake associated with the high specific surface area of the much smaller bacterial cells. In addition, ethanol fermentation performance of the flocculating Z. mobilis ZM401 was further studied and compared with that observed with the regular non-flocculating Z. mobilis ZM4. Based on experimental results, we conclude that compared to S.cerevisiae ADY, a more than 3% increase in ethanol yield can be achieved with ethanol fermentation by Z. mobilis ZM4. Moreover, ethanol productivity can also be improved by Z. mobilis, depending on fermentation processes applied to the system. Under the batch LG fermentation condition, a more than 50% increase in ethanol productivity was observed with Z.mobilis ZM4 compared to that with S.cerevisiae ADY. When the fed-batch strategy is applied to the HG fermentation, much higher ethanol productivity is obtained, which could lead to saving in capital investment on the production facilities. Also, under HG fermentation conditions, the improved ethanol titer in the fermentation broth could save energy consumption for downstream ethanol distillation and reduce the discharge of distillage from the distillation system. In addition, not only can the bacterial flocs be conveniently recovered by cost-effect sedimentation, instead of high-speed centrifugation required by non-flocculating bacteria involving intensive capital investment and energy consumption, but also their tolerance to inhibitory byproducts released during the pretreatment of lignocellulosic biomass is improved. These merits make the flocculating Z.mobilis a suitable host to be engineered with pentose metabolic pathways for fuel ethanol production from lignocellulosic biomass.

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biotechnology Journal
www.biotechnology-journal.com

Biotechnol. J. 2014, 9

[17] Liu, C. G., Xue C., Lin, Y. H., Bai, F. W., Redox potential control and applications in microaerobic and anaerobic fermentations. Biotechnol. Adv. 2013, 31, 257265. [18] Russell, I., Understanding yeast fundamentals, in Jacques, K. A., Lyons, T. P., Kelsall, D. R. (Eds.), The alcohol textbook, Nottingham University Press, 2003, pp. 85120. [19] Loos, H., Kramer, R., Sahm, H., Sprenger, G. A., Sorbitol promotes growth of Zymomonasmobilis in environments with high concentrations of sugar: Evidence for a physiological function of glucosefructose oxidoreductase in osmoprotection. J. Bacteriol. 1994, 176, 76887693. [20] Zhao X. Q., Bai, F. W., Yeast flocculation: New story in fuel ethanol production. Biotechnol. Adv. 2009, 27, 849856. [21] Verstrepen, K. J., Derdelinckx, G., Verachtert, H., Delvaux, F. R., Yeast flocculation: What brewers should know. Appl. Microbiol. Biotechnol. 2003, 61, 197205. [22] Rosche, B., Li, X. Z., Hauer, B., Schmid, A. et al., Microbial biofilms: A concept for industrial catalysis? Trends Biotechnol. 2009, 27, 636643. [23] Klinke, H. B., Thomsen, A. B., Ahring, B. K., Inhibition of ethanol-producing yeast and bacteria by degradation products produced dur-

[24]

[25]

[26]

[27]

[28]

[29]

ing pre-treatment of biomass. Appl. Microbiol. Biotechnol. 2004, 66, 1026. Aristidou, A., Penttil, M., Metabolic engineering applications to renewable resource utilization. Curr. Opin. Biotechnol. 2000, 11, 187198. van Vleet, J. H., Jeffries, T. W., Yeast metabolic engineering for hemicellulosic ethanol production. Curr. Opin. Biotechnol. 2009, 20, 300 306. Zhao, X. Q., Zi, L. H., Bai, F. W., Lin, H. L. et al., Bioethanol from lignocellulosic biomass. Adv. Biochem. Eng. Biotechnol. 2012, 128, 2551. Lee, S. H., Kodaki, T., Park, Y. C., Seo, J. H., Effects of NADH-preferring xylose reductase expression on ethanol production from xylose in xylose-metabolizing recombinant Saccharomyces cerevisiae. J. Biotechnol. 2012, 158, 184191. Dien, B. S., Cotta, M. A., Jeffries, T. W., Bacteria engineered for fuel ethanol production: current status. Appl. Microbiol. Biotechnol. 2003, 63, 258266. Zhao, N., Bai, Y., Zhao, X. Q., Yang, Z. Y. et al., Draft genome sequence of the flocculating Zymomonas mobilis strain ZM401 (ATCC 31822). J. Bacteriol. 2012, 194, 70087009.

10

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim