Bioresource Technology 129 (2013) 321–328

Contents lists available at SciVerse ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Enhanced butanol production by modulation of electron flow in Clostridiumacetobutylicum B3 immobilized by surface adsorption

Dong Liu 1, Yong Chen 1, An Li, Fengying Ding, Tao Zhou, Ying He, Bingbing Li, Huanqing Niu,Xiaoqing Lin, Jingjing Xie, Xiaochun Chen, Jinglan Wu, Hanjie Ying ⇑State Key Laboratory of Materials-Oriented Chemical Engineering, College of Biotechnology and Pharmaceutical Engineering, Nanjing University of Technology, Nanjing, China

h i g h l i g h t s

" NAD(P)H that had escaped from the fermentation as H2 limited the butanol yield." NAD(P)H regulation in C. acetobutylicum B3 by methyl viologen increased the butanol yield by 37.8%." C. acetobutylicum B3 cells were immobilized on a cotton towel by surface adsorption." The butanol tolerance of the immobilized cells was significantly improved." An average of 15.6 g/L butanol was achieved within 12 h in repeated batch fermentation.

a r t i c l e i n f o

Article history:Received 20 September 2012Received in revised form 17 November 2012Accepted 20 November 2012Available online 29 November 2012

Keywords:Butanol toleranceClostridium acetobutylicumImmobilizationMetabolic fluxNAD(P)H

0960-8524/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.biortech.2012.11.090

⇑ Corresponding author. Tel./fax: +86 25 86990001E-mail address: yinghanjie@njut.edu.cn (H. Ying).

1 These authors are equally contributed to this work

a b s t r a c t

The objective of this study was to improve butanol yield and productivity by redox modulation andimmobilization of Clostridium acetobutylicum B3 cells. Stoichiometric network analysis revealed thatNAD(P)H that had escaped from the fermentation as H2 limited the butanol yield and led to the accumu-lation of oxidation byproducts, e.g., acetone. Methyl viologen was used as an electron carrier to divert theelectron flow away from H2 production and to reinforce the NAD(P)H supply. Butanol yield was increasedby 37.8% with severely diminished acetone production. Immobilization of the cells by adsorption onto afibrous matrix improved their butanol tolerance and production rate. An average of 15.6 g/L butanol wasachieved within 12 h with a solvent productivity of 1.88 g/L/h in repeated batch fermentation. To ourknowledge, this is the highest solvent productivity with a relatively high butanol titer produced by a Clos-tridium strain in batch fermentation.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The bioproduction of acetone, butanol and ethanol (ABE) by sol-ventogenic clostridia, such as Clostridium acetobutylicum, was oncethe second largest biotechnological industry in the world (Jonesand Woods, 1986) and has attracted renewed interest for severaleconomic and environmental reasons in recent years. ABE produc-tion by C. acetobutylicum strains in a batch culture is characterizedby two distinct phases, the acidogenic phase and the solventogenicphase. Acetic and butyric acids are predominantly produced in theacidogenic phase, followed by the solventogenic phase whereinpartial acids are reassimilated and solvent formation is observed.Butanol, acetone and ethanol are typically produced in an approx-imate ratio of 6:3:1 (w/w). Butanol, an important industrial

ll rights reserved.

.

.

chemical and excellent alternative to gasoline, is the preferred sol-vent and attracts the highest price (Green, 2011; Jang et al., 2012).To increase the metabolic flux towards butanol and reduce byprod-uct formation, genetic modifications of the acids and acetone for-mation pathways have been explored. Regarding the acetonebranch, antisense RNA against ctfB was employed to improve thebutanol:acetone ratio. However, the mutant strain also exhibitedreduced butanol production (Tummala et al., 2003). Similarly, amutant of C. acetobutylicum EA2018 with the adc gene disruptedby insertion of the group II intron produced small amounts of ace-tone but also significantly lower butanol titers as compared to theparent strain (Jiang et al., 2009). In addition, impairment of theacetone pathway in these mutant strains resulted in largeramounts of acetate accumulating in the growth medium (Lütke-Eversloh and Bahl, 2011). Inactivation of the pta gene involved inacetate formation did not achieve improved butanol productionin C. acetobutylicum ATCC824 (Green et al., 1996). In the degener-ated strain M5, butanol and ethanol production was restored by

322 D. Liu et al. / Bioresource Technology 129 (2013) 321–328

expressing the alcohol dehydrogenase gene adhE, although largeamounts of acetate and butyrate accumulated. To reduce acid pro-duction in M5, an acetate kinase knockout strain was constructedwith adhE expression. However, it still produced large amountsof acids with reduced butanol production compared to the M5adhE expressing strain (Sillers et al., 2008). This suggests that theefficacy of disruption of the acids and acetone pathways to im-prove butanol yield is not as good as was expected.

In addition to low butanol yield, low butanol titers and lowsolvent productivities are another two major challenges for ABEfermentation (Green, 2011). Butanol is a very hydrophobic com-pound. It is toxic and has been shown to disrupt membrane-linkedfunctions, lower intracellular ATP levels and inhibit sugar uptake(Alsaker et al., 2010). The inhibitory threshold of wild Clostridiumis generally considered to be about 13 g/L and the growth of C.acetobutylicum ATCC824 would be completely inhibited at a buta-nol concentration of 14 g/L (Nielsen and Prather, 2009; Ounineet al., 1985). Therefore, most wild C. acetobutylicum strains onlyproduce about 10–13 g/L butanol at an initial glucose concentra-tion of 60 g/L. Since low butanol titers result in high recovery costs,some chemically mutated or genetically engineered C. acetobutyli-cum strains have been created to improve their butanol tolerance.These strains can produce butanol at titers as high as 15 g/L (Jianget al., 2009; Shen et al., 2011). Although this has partially improvedthe butanol titer, low solvent productivity is still an important is-sue that hampers the feasibility of the ABE fermentation.

The present study focused on the redox balance of the meta-bolic networks in C. acetobutylicum B3 to address the low butanolyield. To further overcome the problems of low butanol titer andproductivity, cells were immobilized onto a fibrous matrix (cottontowel) by surface adsorption to improve their butanol toleranceand applied to biobutanol production in a repeated batch fermen-tation model.

2. Methods

2.1. Organism and culture conditions

C. acetobutylicum B3 (CGMCC No. 5234) was used in this study. Itwas previously isolated from soil and was selected after UV muta-genesis. It was grown anaerobically at 37 �C in an anaerobic chamber(Bug Box, Ruskinn Technology, Leeds, UK). Cultures of C. acetobutyl-icum B3 were grown in solid reinforced clostridial medium (RCM) forroutine growth, and modified P2 medium (P2 medium containing10 g/L glucose as the sole carbohydrate) for seed culture.

2.2. Fermentation

Fermentation experiments were performed anaerobically in P2medium (glucose 60 g/L; K2HPO4 0.5 g/L; KH2PO4 0.5 g/L; CH3-

COONH4 2.2 g/L; MgSO4�7H2O 0.2 g/L; MnSO4�H2O 0.01 g/L; NaCl0.01 g/L; FeSO4�7H2O 0.01 g/L; p-aminobenzoic acid 1 mg/L; thia-mine 1 mg/L; biotin 0.01 mg/L) (Baer et al., 1987) at 37 �C.

Methyl viologen was purchased from Sigma–Aldrich as thedichloride hydrate and was prepared as a sterile stock solution inthe oxidized state (colorless) and added as necessary to a final con-centration of 0.4 mM before fermentation. Generally, fermentationwas carried out in 500-mL Duran bottles with 300 mL of workingvolume. Each bottle was inoculated with 30 mL of a 12-h-old seedculture. The bottles have a tube connected to a 0.22-um syringefilter for gas release and a tube inserted deep in the culture forsample taking. Samples were taken with autoclaved syringes andthen the tube was closed with a clamp. All experiments were car-ried out in triplicate and the mean value was calculated.

For fermentation with immobilized cells, 18 g of dried cottontowel was sterilized by autoclaving at 121 �C for 15 min in a Duranbottle and added to the 300 mL of medium before fermentation toimmobilize the cells by surface adsorption.

The repeated batch fermentations using the immobilized cellswere carried out in 2-L stainless steel columns packed with 120 gof cotton towel. The columns were sterilized at 121 �C for 30 minbefore use. 600 mL of 12-h-old seed culture were inoculated to a10-L stirred tank fermentor containing 6 L P2 medium and allowedto grow for 20 h at 37 �C without agitation. Then the culture wastransferred to a stainless steel column and circulated through thecotton towel at a pumping rate of 25 mL/min to allow the cells tobe immobilized onto the fibrous matrix. After about 15 h of contin-uous circulation, the fermentation broth was replaced with fresh P2medium to start the repeated batch fermentations. The medium cir-culation rate was maintained at 35 mL/min via a peristaltic pumpand the temperature was controlled via connection of the jacketedcolumn to a warm water bath at 36 �C. When a batch ended, 3 Lfresh P2 medium were fed to the bottom of the column to replacethe fermentation broth and rinse the fibrous matrix. Then the med-ium was circulated again, and a new batch fermentation wasstarted with the cells that remained adsorbed to the fibrous matrix.

2.3. Measurement of the intracellular NAD(P)H concentration

Intracellular concentrations of NAD(P)H were determined bythe enzyme cycling method of Bernofsky and Swan (Bernofskyand Swan, 1973) with modifications. Generally, 2.8 mL of samplewere taken and cells were collected and dissolved in 0.5 mL0.2 M NaOH. The cell lysate was heated at 60 �C for 30 min to spe-cifically extract NAD(P)H and to decompose NAD(P)+, then cooledto 0 �C and centrifuged.

350 uL ddH2O, 200 uL 0.1 M HCl, 100 uL Tris–HCl (1 M, pH 7.8),200 uL of the alkaline extract, 100 uL 4.2 mM MTT, 150 uL 16.6 mMPES and 100 uL ethanol for the determination of NADH, or 100 uL60 mM glucose 6-phosphate for the determination of NADPH, weresequentially added to a test tube and kept at 37 �C for 5 min in thedark. The reaction was started by adding 10 uL of alcohol dehydro-genase (660 units/mL, for NADH) or glucose 6-phosphate dehydro-genase (70 units/mL, for NADPH) and incubated at 37 �C for 30 minin the dark. The absorbance at 570 nm was determined. The sameprocedure was followed for NAD(P)H standards.

2.4. Butanol tolerance

Butanol tolerance was assessed by two procedures. First, to testthe tolerance of the growing culture, butanol at different concen-trations was added simultaneously with the bacterial inoculumto P2 medium. Growth was determined spectrophotometricallyas the optical density at 660 nm (OD660nm). To determine the den-sity of cells immobilized on the fibrous bed, the culture was carriedout in a 15-mL screw-cap bottle with a 6-mL working volume con-taining 0.36 g of cotton towel. The bottle contents were sonicatedat 4 �C for 5 min (using a Bransonic 220; Branson Co., Shelton,Conn.) to remove and disperse the immobilized cells (Hoffmanet al., 2005). Second, cells were grown in P2 medium at 37 �C for12 h. Then, butanol was directly added to a final concentration (to-gether with the concentration of butanol produced in the culture)of 12 or 15 g/L. The survival rates at different time intervals weredetermined by viable plate counts. Experiments were carried outin triplicate and the mean value was calculated.

2.5. Scanning electron microscopy of planktonic and immobilized cells

For scanning electron microscopy, the cells were collected inthe stationary phase. Standard methods were used to fix

D. Liu et al. / Bioresource Technology 129 (2013) 321–328 323

planktonic cells to slides as described previously (Kubota et al.,2008). For observation of immobilized cells, a piece of cotton towelwas harvested and rinsed twice in saline before freeze-drying.

2.6. Analyses

Glucose, acetate, butyrate and lactate concentrations weredetermined by HPLC analysis (Agilent 1100 series, Hewlett–Pack-ard, USA), using an Aminex HPX-87H ion exclusion column(300 � 7.8 mm; Bio-Rad Laboratories, Hercules, CA, USA), with a5.0 mM H2SO4 solution used as the mobile phase (0.6 mL/min) at50 �C and a refractive index detector. Acetone, ethanol, butanoland acetoin were analyzed using gas chromatography (7890A, Agi-lent, Wilmington, DE, USA) equipped with a flame ionization detec-tor (FID) and an Agilent HP-INNOWAX column (0.25 mm � 60 m).

To determine hydrogen production, the Duran bottles were in-verted instead and the hydrogen production was ascertained usingthe gas capture method (Cappelletti et al., 2011).

3. Results and discussion

3.1. Metabolic network stoichiometry analysis

The metabolic pathways are depicted in Fig. 1. Details on thesemetabolic pathways can be found in a reference (Desai et al., 1999).Elementary steps were combined to form the overall reactions

Fig. 1. Metabolic flux distributions in the planktonic and immobilized cells with or withoMV addition; (C) immobilized cells without MV addition; (D) immobilized cells with M

numbered in Fig. 1. They consist of 17 species and 16 pathwayfluxes. The species can be segregated into metabolic intermediateswhose accumulation is set to zero at the end of fermentation andexchangeable components (end products including glucose andbiomass) whose accumulation can be directly measured (seeAppendix A). The stoichiometric matrix can be obtained from theequations that describe the metabolic networks of C. acetobutyli-cum. However, the rank of the stoichiometric matrix is only 15,which is less than the pathway fluxes to be determined, resultingin a singularity that prevents calculation of a unique set of fluxes.This singularity is inevitable as acetone production is coupled tothe reassimilation of either acetate or butyrate. To solve this prob-lem, some workers developed a nonlinear constraint based on thein vivo kinetics and selectivity of the CoA transferase to relate theacetate and butyrate uptake fluxes (Desai et al., 1999). However,we had no interest in solving the singularity here, since it has noeffect on calculating the metabolic flux distribution of the endproducts in the networks based on the carbon and NAD(P)Hbalance.

The accumulation of the end products can be determined fromEqs. (19) and (21) derived from the network stoichiometry inAppendix A. Accumulation of any two of the ten products, glucose(Glu), biomass, acetoin (Acoi), acetone (Acon), lactate (Lac), molec-ular hydrogen (H2), acetate (Ac), butyrate (But), ethanol (EtOH) andbutanol (BuOH), can be calculated from the accumulation of theothers. In this work, besides the biomass, the accumulation of ace-tone was calculated and compared to the measured value. This was

ut MV addition. (A) Planktonic cells without MV addition; (B) planktonic cells withV addition.

A

B

C

Fig. 2. Effects of methyl viologen on H2 yield (A), intracellular content of NADH (B),and NADPH (C) in C. acetobutylicum B3.

324 D. Liu et al. / Bioresource Technology 129 (2013) 321–328

because the measured acetone production may be far less than theactual value in the open system used for hydrogen measurementdue to its high volatility.

2rAcoi þ 2rAc þ 2rBut þ 4rAcon � 0:873rBiomass ¼ rH2 ð19Þ

It should be noted that Eq. (19) is significant for understandingthe metabolism of C. acetobutylicum. Though it is derived from thenetwork stoichiometry, it has a biological meaning in view of theNAD(P)H balance. In butanol production by C. acetobutylicum,when one molecule of acetoin, acetate, butyrate or acetone is gen-erated from glucose, 2, 2, 2 or 4 molecules of NAD(P)H, respec-tively, are concomitantly produced. Apart from that used forbiomass synthesis, all NAD(P)H produced during the formation ofthe oxidation products escaped in the form of H2 via the hydroge-nase reaction. Products with equal degrees of reduction to glucose,such as butanol and ethanol, are absent from Eq. (19), suggestingthey can be produced from glucose with 100% molar yield theoret-ically. However, the release of H2 causes a deficiency of the reduc-ing equivalents NAD(P)H used for ethanol and butanol production.This was why a large quantity of oxidation products (especiallyacetate and acetone) is commonly present at the end of fermenta-tion, resulting in a low butanol yield. In other words, the NAD(P)Hthat has escaped as H2 is balanced by an accumulation of oxidationproducts. This statement seems quite reasonable since the butanolyield was always effectively enhanced with little production ofacetate and acetone when the hydrogenase was inhibited (Grupeand Gottschalk, 1992; Meyer et al., 1986). However, when the ace-tate or acetone pathway was disrupted, there was still a largequantity of oxidation products (e.g. acetone and acetate) and buta-nol production could not be improved (Jiang et al., 2009; Kuit et al.,2012; Lee et al., 2008). Therefore, it seems that the hydrogenasenode, directly regulating the NAD(P)H balance, plays a moreimportant role in butanol production than the acetate or acetonesynthesis pathway.

3.2. Modified electron flow for enhanced butanol yield

According to the metabolic stoichiometry analysis, it is theNAD(P)H that escapes in the form of H2 that is responsible forthe low butanol yield. In C. acetobutylicum, pyruvate is convertedto acetyl–CoA by pyruvate–ferredoxin oxidoreductase (PFOR) withconcomitant conversion of oxidized ferredoxin to its reduced form,FdH2. FdH2 is liable to be oxidized by hydrogenase to release H2.Another enzyme, ferredoxin–NAD(P) reductase, competes withhydrogenase to oxidize FdH2 and convert NAD(P) to NAD(P)H,but less efficiently (Gheshlaghi et al., 2009). To divert the electronflow away from H2 production, an electron carrier methyl viologen(MV) was added to the culture initially. Methyl viologen, the com-mercial herbicide ‘‘paraquat’’, can accept electrons from PFOR andtransfer them to NAD(P)+ via the ferredoxin–NAD(P)+ reductase(Peguin et al., 1994). Thus, the electron flow may be redirected toform NAD(P)H and used for butanol production.

During the exponential growth of C. acetobutylicum B3, FdH2

was predominantly oxidized by hydrogenase to release hydrogenwith concomitant accumulation of acetate and butyrate. Afterentering the solventogenesis phase, the hydrogen yield graduallydeclined and FdH2 was used to generate NAD(P)H to reduce acetateand butyrate with concomitant production of solvents. However,the average hydrogen yield at the end of fermentation was stillas high as 1.16 mol/mol glucose, indicating that 60% of the FdH2

produced via PFOR was converted to hydrogen (Fig. 2). Whenmethyl viologen was added, the hydrogen yield decreased dramat-ically to 0.60 mol/mol glucose, 50% lower than the control. Theelectron flow from FdH2 to NAD(P)+ was increased from 0.53 to1.2 mol/mol glucose and both the intracellular NADH and NADPHcontents were elevated significantly. The butanol yield was

strongly favored by the redirected electron flow and elevatedNAD(P)H levels, and increased by 37.8% relative to the control(Figs. 1 and 2). Meanwhile, the yield of oxidation product was ex-tremely low with maximum production of acetone of no more than1.4 g/L, and the butanol:acetone ratio was successfully increasedfrom 2.6 to 10.8 (Fig. 3).

Though the yield of butanol was significantly increased uponaddition of MV, the titer of butanol was not substantially improved(from 11.9 to 12.8 g/L), with 14.5 g/L residual glucose. Butanol istoxic and has been shown to disrupt membrane-linked functions,lower intracellular ATP levels and inhibit sugar uptake (Alsaker etal., 2010). The inhibitory threshold of wild Clostridium is generallyconsidered to be about 13 g/L and the growth of C. acetobutylicumATCC824 would be completely inhibited at a butanol concentrationof 14 g/L (Nielsen and Prather, 2009; Ounine et al., 1985). So, it isreasonable to assume that the low butanol titer can mainly beattributed to the susceptibility of C. acetobutylicum B3 to butanol.To further improve the butanol titer, C. acetobutylicum B3 was

Fig. 3. The fermentation performance of planktonic C. acetobutylicum B3 culturewith MV addition.

A

B

D. Liu et al. / Bioresource Technology 129 (2013) 321–328 325

immobilized by surface adsorption to improve its butanol tolerance(see below).

Fig. 4. Comparison of the butanol tolerance of planktonic and immobilized C.acetobutylicum B3 cells. (A) The cell density under different initial concentrations ofbutanol; (B) the survival of planktonic and immobilized B3 cells challenged withhigh concentrations of butanol.

3.3. Improved butanol tolerance of C. acetobutylicum cells immobilizedby surface adsorption

Dispersed cells can come together and form polymicrobialaggregates attached to a solid surface. This cell-solid surface con-tact has been recognized as providing significant structure modu-lations and phenotype changes in the anchored cells, one ofwhich is an increased resistance to antimicrobial agents (Mahand O’Toole, 2001). Here, a fibrous matrix (cotton towel) was ap-plied to immobilize C. acetobutylicum B3 cells by surface adsorp-tion. It seems that the cells tended to propagate on the fibrousmatrix rather than in the bulk medium, as was indicated by thelow turbidity of the culture with cotton towel (data not shown).SEM images demonstrated that the surface adsorbed cells were lo-cated in aggregates and effectively immobilized by the extracellu-lar polymeric substances (EPSs) they produced. Unlike theaggregates of surface adsorbed cells, the morphology of the plank-tonic cells showed that the cells grew separately from each other,free of EPS cover.

The butanol tolerance of the C. acetobutylicum B3 cells immobi-lized on the fibrous matrix was characterized. When no butanolwas added to the culture, the OD660nm of the immobilized cellscontinued to increase over 21 h, whereas the OD660nm of the plank-tonic cells began to decrease as early as 15 h. Addition of 8 g/Lbutanol initially had little effect on the growth of the immobilizedcells, as was indicated by the comparable OD660nm, but the maxi-mum cell density of the planktonic cells was decreased by 22%and began to decline at an earlier time. No growth of the plank-tonic cells was observed when 12 g/L butanol was added initially,whereas the immobilized cells showed steady growth althoughat a lower cell density (Fig. 4). These results demonstrated thatthe fibrous matrix significantly improved the butanol resistanceof the growing cells. The survival of the immobilized cells exposedto high butanol concentrations for 2 h was dramatically enhanced,and was about 3 orders of magnitude higher than that of the plank-tonic cells (Fig. 4). Thus, the immobilization of C. acetobutylicum B3by adsorption to the fibrous matrix rendered the cells more resis-tant to butanol.

Recent work has indicated that, when cells attach to a solid sur-face, they exhibit a different pattern of gene expression. Somegenes are repressed or activated, and the cellular structure (e.g.plasma membrane composition) is modulated, leading to a resis-

tant phenotype (Flemming and Wingender, 2010). When C. acet-obutylicum B3 cells were detached from the fibrous matrix, theirsurvival rate was lower than that of the EPS-immobilized cells afterexposure to butanol, but still much higher than that of the plank-tonic cells (data not shown), suggesting the extracellular polymericsubstance (EPS) produced by the cells when attached to the fibrousmatrix could also confer resistance. The EPS plays an importantrole in exclusion of toxic substances and maintenance of a highlyhydrated microenvironment (Flemming and Wingender, 2010).Since butanol is a very hydrophobic compound, the hydratedmicroenvironment maintained by the EPS may be of especialimportance for ability of the cells to resist butanol toxicity.

3.4. Enhanced butanol yield and productivity by the surface-adsorbedcells with modified electron flow

Modulation of the electron flow in the surface-adsorbed cellshad a marked effect on butanol production. An average of 14.9 g/L butanol was obtained with MV addition, which was well abovethe inhibitory threshold of planktonic cells, indicating the im-proved butanol tolerance of the adsorbed cells. The metabolic fluxdistributions of the planktonic and immobilized cells with or with-out MV addition were calculated and are summarized in Fig. 1.Immobilization of the cells by surface adsorption led to an in-creased release of electrons (H2 production), and the yield ofoxidation product acetone was 37.1% higher than that produced

A

B

Fig. 5. Repeated batch fermentations by immobilized C. acetobutylicum B3 cells in a2-L stainless steel column with medium circulation. (A) Solvent production; and (B)pH and cell density in the bulk medium.

326 D. Liu et al. / Bioresource Technology 129 (2013) 321–328

by the planktonic cells. However, when MV was added to divert theelectrons away from H2 release in the immobilized cells, the yieldsof oxidation products (acetate, acetone and acetoin) were reducedand the butanol yield was significantly increased by 31.3%. Ethanolproduction was also increased dramatically, from 1.3 g/L to 2.1 g/L.Unlike the planktonic cells, MV addition to the immobilized cellculture did not result in incomplete sugar consumption. Therefore,a high butanol titer was achieved with an improved yield by mod-ulation of the electron flow in the immobilized cells.

The surface-adsorption immobilization of the cells not only al-lowed a high butanol titer but also accelerated the fermentationprocess dramatically. The typical fermentation time for the plank-tonic cells in P2 medium was about 80 h, whereas the fermentationtime for the surface-adsorbed cells was no more than 40 h. When

Table 1Comparison of fermentation performance in butanol production.

Culture Technique

C. acetobutylicum ATCC 55025 Immobilized in fibrous bed, continuous fermentationC. beijerinckii BA101 Adsorbed onto clay brick, continuous fermentationE. coli Recombinant, batch fermentationC. acetobutylicum EA2018 NTG mutagenesis, batch fermentationC. beijerinckii BA101 Addition of sodium acetate, batch fermentationC. acetobutylicum B3 Adsorbed on fibrous matrix, repeated batch fermentati

the fermentations were scaled up to 2 L in stainless steel columnswith medium circulation, the fermentation time dropped further toless than 14 h and, thus, the productivity was dramatically en-hanced. In comparison, when the planktonic cell fermentationswere carried out in the 2-L stainless steel columns, the fermenta-tion performance showed no significant improvement. Further-more, the immobilized cells can be reutilized for repeated batchfermentations. In the repeated batch fermentations, the averagebutanol yield and productivity were further improved. TheOD660nm of the fermentation broth decreased significantly afterthe 4th batch and then remained at about 0.2–0.7 throughoutthe fermentation. Compared to the cell density in the planktoniccell fermentation, whose maximum OD660nm was 5–6, this celldensity was much lower, indicating the fibrous matrix providedan excellent immobilization of the cells. The first 10 batches areshown in Fig. 5. Excitingly, an average of 15.6 g/L butanol can beobtained within 12 h with a total solvent productivity of1.88 g L�1 h�1, compared to 11.9 g/L butanol and a productivity of0.23 g L�1 h�1 for the planktonic cell fermentation without MVaddition. The repeated batch fermentation can be operated steadilyfor a long time (at least 4 weeks).

In addition to the improved butanol tolerance of the surface ad-sorbed cells, the existing cells remaining in the fibrous matrix atthe start of a new batch and their already established butanol-pro-ducing capacity might contribute to the accelerated fermentationrate and increase the carbon flux to butanol rather than biomasssynthesis. Furthermore, the quantity of cells in the fibrous matrixunder high butanol concentrations in the late fermentation phasewas much larger than the quantity of cells in the planktonic cellfermentation (Fig. 4), accounting for the accelerated fermentationrate as well. The cells adsorbed onto the unpretreated fibrous ma-trix were immobilized in a natural way (immobilized by self-pro-duced EPS). Thus, the cells were likely renewed continuously andmaintained a stable long-term performance. Similarly, C. acetobu-tylicum cells immobilized into poly(vinyl alcohol) cryogel were alsodemonstrated to be applicable for multiple reuses in the ABE fer-mentation of pretreated microalgae biomass (Efremenko et al.,2012).

To our knowledge, an average of 12 h is the shortest fermenta-tion time ever reported for a batch fermentation. The results ob-tained in this study are compared with the best results reportedin continuous or batch fermentations with different techniquesin Table 1. Continuous production with immobilized cells has tra-ditionally been developed to improve butanol productivity. Indeed,the productivity was usually dramatically enhanced and muchhigher than that of the planktonic cell fermentation. The highestproductivity ever reported for a continuous production was as highas 15.8 g/L/h (Qureshi et al., 2000), which was about 50 times high-er than the typical productivity in planktonic cell fermentation.However, this impressively high productivity was achieved at a rel-atively low butanol concentration of only about 5 g/L, which wouldincrease the recovery costs. Until now, the highest butanol concen-tration achieved in a fermentation broth was 20.9 g/L, which was

Performance References

Butanoltiter(g/L)

Solventproductivity(g/L/h)

Solventyield(g/g)

�5 4.6 0.42 Huang et al. (2004)�5 15.8 0.38 Qureshi et al. (2000)15 0.23 0.34 Shen et al. (2011)14.5 0.33 0.3 Jiang et al. (2009)20.9 0.59 0.41 Chen and Blaschek (1999)

on 15.6 1.88 0.38 This work

D. Liu et al. / Bioresource Technology 129 (2013) 321–328 327

obtained in a batch fermentation with the well known hyper-buta-nol-producing strain, another species of clostridium, C. beijerinckiiBA101. In spite of the use of this strain, the productivity and yieldof batch fermentation were usually relatively low. In this study, thesimple, natural immobilization of C. acetobutylicum B3 cells and itsapplication to butanol production by repeated batch fermentationwas demonstrated to generate both a relatively high butanol titer(yield) and high solvent productivity, and thus was predicted tobe economically favorable.

4. Conclusions

The NAD(P)H released in the form of H2 was responsible for thelow butanol yield in C. acetobutylicum. Methyl viologen was used todivert the electron flow away from H2 formation, and the butanolyield was successfully increased by 37.8%. Immobilization of C.acetobutylicum B3 by adsorption onto a fibrous matrix renderedthe cells more resistant to butanol. An average of 15.6 g/L butanolwas obtained within 12 h with a productivity of 1.88 g/L/h in therepeated batch fermentation. The results demonstrated thatNAD(P)H manipulation in C. acetobutylicum and surface-adsorptionimmobilization were effective in improving butanol production.

Acknowledgements

This work was supported by a Grant from the National Out-standing Youth Foundation of China (Grant No.: 21025625); theNational High-Tech Research and Development Program of China(863) (Grant No.: 2012AA021200); the National Basic ResearchProgram of China (973) (Grant No.: 2011CBA00806); the NationalNatural Science Foundation of China, Youth Program (Grant No.:21106070); Program for Changjiang Scholars and Innovative Re-search Team in University (Grant No.: IRT1066); Jiangsu ProvincialNatural Science Foundation of China (Grant No.: SBK 201150207)and Project Funded by the Priority Academic Program Develop-ment of Jiangsu Higher Education Institutions (PAPD).

Appendix Appendix. A

Details on the metabolic pathways can be found in a reference(Desai et al., 1999). Elementary steps were combined to form theoverall reactions. NADH was used as the single pool of reducingequivalents when fluxes were calculated.

Accumulation of the metabolic intermediates was set to zero:

rNADH ¼ 2r1 � r2 þ r6 � 2r7 � 2r10 � 2r15 � 0:873r16 ¼ 0 ð1Þ

rpyr ¼ 2r1 � r2 � 2r3 � r4 ¼ 0 ð2Þ

rFdH2 ¼ r4 � r5 � r6 ¼ 0 ð3Þ

rAcCoA ¼ r4 � r7 � r8 � 2r9 þ r12 ¼ 0 ð4Þ

rBurlCoA ¼ r10 � r11 þ r13 � r15 ¼ 0 ð5Þ

rAcAcCoA ¼ r9 � r10 � r12 � r13 ¼ 0 ð6Þ

rAcAc ¼ r12 þ r13 � r14 ¼ 0 ð7Þ

Product accumulation:

rGlu ¼ �r1 � r16 ð8Þ

rLac ¼ r2 ð9Þ

rAc ¼ r8 � r12 ð10Þ

rBut ¼ r11 � r13 ð11Þ

rAcon ¼ r14 ð12Þ

rAcoi ¼ r3 ð13Þ

rEtOH ¼ r7 ð14Þ

rBuOH ¼ r15 ð15Þ

rH2 ¼ r5 ð16Þ

rBiomass ¼ r16 ð17Þ

To eliminate the intermediate fluxes and obtain the relationshipbetween the product fluxes, a linear operation on the above equa-tions was performed as follows:

ð1Þ � ð2Þ þ ð3Þ � 2 � ð4Þ � 2 � ð5Þ � 4 � ð6Þ � 4 � ð7Þ

yielding the following equation:

2r3 � r5 þ 2r8 þ 2r11 � 2r12 � 2r13 þ 4r14 � 0:873r16 ¼ 0 ð18Þ

In view of the product accumulation equations, Eq. (18) can bewritten in the form,

2rAcoi þ 2rAc þ 2rBut þ 4rAcon � 0:873rBiomass ¼ rH2 ð19Þ

Similarly, another linear operation was performed as follows:

ð2Þ þ ð4Þ þ 2 � ð5Þ þ 2 � ð6Þ þ 2 � ð7Þ

yielding the following equation:

2r1� r2� 2r3� r7� r8� 2r11þ r12þ 2r13� 2r14� 2r15 ¼ 0

ð20Þ

Also, Eq. (20) can be written in the form,

rLacþ2rAcoiþ rEtOHþ rAcþ2rButþ2rAconþ2rBuOHþ2rBiomass ¼�2rGlu

ð21Þ

According to Eqs. (19) and (21), accumulation of any two of theten end products can be calculated from the others.

References

Alsaker, K.V., Paredes, C., Papoutsakis, E.T., 2010. Metabolite stress and tolerance inthe production of biofuels and chemicals: gene-expression-based systemsanalysis of butanol, butyrate, and acetate stresses in the anaerobe Clostridiumacetobutylicum. Biotechnol. Bioeng. 105, 1131–1147.

Baer, S.H., Blaschek, H.P., Smith, T.L., 1987. Effect of butanol challenge andtemperature on lipid composition and membrane fluidity of butanol-tolerantClostridium acetobutylicum. Appl. Environ. Microbiol. 53, 2854–2861.

Bernofsky, C., Swan, M., 1973. An improved cycling assay for nicotinamide adeninedinucleotide. Anal. Biochem. 53, 452–458.

Cappelletti, B.M., Reginatto, V., Amante, E.R., Antonio, R.V., 2011. Fermentativeproduction of hydrogen from cassava processing wastewater by Clostridiumacetobutylicum. Renew. Energy 36, 3367–3372.

Chen, C.K., Blaschek, H.P., 1999. Acetate enhances solvent production and preventsdegeneration in Clostridium beijerinckii BA101. Appl. Microbiol. Biotechnol. 52,170–173.

Desai, R.P., Nielsen, L.K., Papoutsakis, E.T., 1999. Stoichiometric modeling ofClostridium acetobutylicum fermentations with non-linear constraints. J.Biotechnol. 71, 191–205.

Efremenko, E.N., Nikolskaya, A.B., Lyagin, I.V., Senko, O.V., Makhlis, T.A., Stepanov,N.A., Maslova, O.V., Mamedova, F., Varfolomeev, S.D., 2012. Production ofbiofuels from pretreated microalgae biomass by anaerobic fermentation withimmobilized Clostridium acetobutylicum cells. Bioresour. Technol. 114, 342–348.

Flemming, H.C., Wingender, J., 2010. The biofilm matrix. Nat. Rev. Microbiol. 8, 623–633.

Gheshlaghi, R., Scharer, J.M., Moo-Young, M., Chou, C.P., 2009. Metabolic pathwaysof clostridia for producing butanol. Biotechnol. Adv. 27, 764–781.

Green, E.M., 2011. Fermentative production of butanol—the industrial perspective.Curr. Opin. Biotechnol. 22, 1–7.

Green, E.M., Boynton, Z.L., Harris, L.M., Rudolph, F.B., Papoutsakis, E.T., Bennett, G.N.,1996. Genetic manipulation of acid formation pathways by gene inactivation inClostridium acetobutylicum ATCC 824. Microbiology (UK) 142, 2079–2086.

328 D. Liu et al. / Bioresource Technology 129 (2013) 321–328

Grupe, H., Gottschalk, G., 1992. Physiological events in Clostridium acetobutylicumduring the shift from acidogenesis to solventogenesis in continuous culture andpresentation of a model for shift induction. Appl. Environ. Microbiol. 58, 3896–3902.

Hoffman, L.R., D’Argenio, D.A., MacCoss, M.J., Zhang, Z., Jones, R.A., Miller, S.I., 2005.Aminoglycoside antibiotics induce bacterial biofilm formation. Nature 436,1171–1175.

Huang, W.C., Ramey, D.E., Yang, S.T., 2004. Continuous production of butanol byClostridium acetobutylicum immobilized in a fibrous bed bioreactor. Appl.Biochem. Biotechnol. 115, 887–898.

Jang, Y., Malaviya, A., Cho, C., Lee, J., Lee, S.Y., 2012. Butanol production fromrenewable biomass by clostridia. Bioresour. Technol. 123, 653–663.

Jiang, Y., Xu, C., Dong, F., Yang, Y., Jiang, W., Yang, S., 2009. Disruption of theacetoacetate decarboxylase gene in solvent-producing Clostridiumacetobutylicum increases the butanol ratio. Metab. Eng. 11, 284–291.

Jones, D.T., Woods, D.R., 1986. Acetone–butanol fermentation revisited. Microbiol.Rev. 50, 484–524.

Kubota, H., Senda, S., Nomura, N., Tokuda, H., Uchiyama, H., 2008. Biofilm formationby lactic acid bacteria and resistance to environmental stress. J. Biosci. Bioeng.106, 381–386.

Kuit, W., Minton, N.P., López-Contreras, A.M., Eggink, G., 2012. Disruption of theacetate kinase (ack) gene of Clostridium acetobutylicum results in delayedacetate production. Appl. Microbiol. Biotechnol. 94, 729–741.

Lee, S.Y., Park, J.H., Jang, S.H., Nielsen, L.K., Kim, J., Jung, K.S., 2008. Fermentativebutanol production by Clostridia. Biotechnol. Bioeng. 101, 209–228.

Lütke-Eversloh, T., Bahl, H., 2011. Metabolic engineering of Clostridiumacetobutylicum: recent advances to improve butanol production. Curr. Opin.Biotechnol. 22, 1–14.

Mah, T.F., O’Toole, G.A., 2001. Mechanisms of biofilm resistance to antimicrobialagents. Trends Microbiol. 9, 34–39.

Meyer, C.L., Roos, J.W., Papoutsakis, E.T., 1986. Carbon monoxide gasing leads toalcohol production and butyrate uptake without acetone formation incontinuous cultures of Clostridium acetobutylicum. Appl. Microbiol. Biotechnol.24, 159–167.

Nielsen, D.R., Prather, K.J., 2009. In situ product recovery of n-butanol usingpolymeric resins. Biotechnol. Bioeng. 102, 811–821.

Ounine, K., Petitdemange, H., Raval, G., Gay, R., 1985. Regulation and butanolinhibition of D-Xylose and D-Glucose uptake in Clostridium acetobutylicum. Appl.Environ. Microbiol. 49, 874–878.

Peguin, S., Goma, G., Delorme, P., Soucaille, P., 1994. Metabolic flexibility ofClostridium acetobutylicum in response to methyl viologen. Appl. Microbiol.Biotechnol. 42, 611–616.

Qureshi, N., Schripsema, J., Lienhardt, J., Blaschek, H., 2000. Continuous solventproduction by Clostridium beijerinckii BA101 immobilized by adsorption ontobrick. World J. Microb. Biot. 16, 377–382.

Shen, C.R., Lan, E.I., Dekishima, Y., Baez, A., Cho, K.M., Liao, J.C., 2011. Driving forcesenable high-titer anaerobic 1-butanol synthesis in Escherichia coli. Appl.Environ. Microb. 77, 2905–2915.

Sillers, R., Chow, A., Tracy, B., Papoutsakis, E.T., 2008. Metabolic engineering of thenon-sporulating, non-solventogenic Clostridium acetobutylicum strain M5 toproduce butanol without acetone demonstrate the robustness of the acid-formation pathways and the importance of the electron balance. Metab. Eng. 10,321–332.

Tummala, S.B., Welker, N.E., Papoutsakis, E.T., 2003. Design of antisense RNAconstructs for downregulation of the acetone formation pathway of Clostridiumacetobutylicum. J. Bacteriol. 185, 1923–1934.