Bioresource Technology 155 (2014) 8–14

Contents lists available at ScienceDirect

Bioresource Technology

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

Enhanced poly(c-glutamic acid) fermentation by Bacillus subtilis NX-2immobilized in an aerobic plant fibrous-bed bioreactor

0960-8524/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.biortech.2013.12.080

⇑ Corresponding author at: College of Food Science and Light Industry, NanjingUniversity of Technology, Nanjing 211816, PR China. Tel./fax: +86 25 58139433.

E-mail address: fengxiaohai@njut.edu.cn (X. Feng).

Zongqi Xu a,b, Xiaohai Feng a,b,⇑, Dan Zhang a,b,c, Bao Tang a,b, Peng Lei a,b, Jinfeng Liang a,b, Hong Xu a,b

a State Key Laboratory of Materials-Oriented Chemical Engineering, Nanjing University of Technology, Nanjing 210009, PR Chinab College of Food Science and Light Industry, Nanjing University of Technology, Nanjing 211816, PR Chinac School of Agriculture and Biology, Shanghai Jiaotong University, Shanghai 200240, PR China

h i g h l i g h t s

� Aerobic fibrous-bed bioreactor using bagasse as immobilized carrier was designed.� The time of immobilization fermentation is 48 h while free-cell 72 h.� PGA yield of 71.21 g/L was achieved by repeated fed-batch fermentation in APFB.� We evaluated the stability of the APFB system.� The changes in the cells of free and immobilized cell fermentation were compared.

a r t i c l e i n f o

Article history:Received 16 October 2013Received in revised form 16 December 2013Accepted 18 December 2013Available online 27 December 2013

Keywords:Bacillus subtilis NX-2Aerobic plant fibrous-bed bioreactorPoly(c-glutamic acid)Repeated fed-batch fermentationCell membrane permeability

a b s t r a c t

To enhance poly(c-glutamic acid) (PGA) production, a novel aerobic plant fibrous-bed bioreactor (APFB)was constructed for immobilized fermentation. Based on the analysis of the kinetics of immobilized-cellfermentation using the APFB and conventional free-cell fermentation, immobilized-cell fermentationexhibited more efficient PGA production. Furthermore, repeated fed-batch cultures for PGA productionwere conducted to evaluate the stability of the APFB system. Average final PGA concentration and pro-ductivity of 71.21 ± 0.83 g/L and 1.246 ± 0.008 g/L/h were respectively achieved by cells immobilized inbagasse during APFB, which was reused eight times over a period of 457 ± 18 h. Analysis of the membranephospholipids and the key enzyme activities indicated that APFB-adapted cells had better productivitythan original cells. Thus, this study demonstrated the significant potential of the APFB culture systemin future industrial applications.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Poly(c-glutamic acid) (PGA) is a homopolymer that consists ofD- and L-glutamic acid units connected by c-amide linkages pro-duced by microbes (Shih and Van, 2001). PGA has been widely uti-lized in a broad range of industrial fields, such as cosmetics, food,medicine, agriculture, and water treatment, for its nontoxic, biode-gradable, and excellent environmental characteristics (Ashiuchi,2010). PGA has recently been successfully used as fertilizer syner-gists, biodegradable fibers, drug carriers, and biological adhesives(Sonaje et al., 2010; Tajima et al., 2011; Zeng et al., 2013).

The increasing applications of PGA has caused extensive re-search to be conducted with a focus on increasing PGA fermenta-tive production using a variety of approaches, such asoptimization of the fermentation process (Jeong et al., 2010),

agitation speed-shift control strategies (Zhang et al., 2011a), andthe addition of oxygen vectors or organic acid to free cell fermen-tation (Zhang et al., 2012a,b). However, PGA fermentation pro-cesses still encounter problems, such as low final PGAconcentration, low productivity, and low reactor productivity(Bajaj and Singhal, 2010, 2011). Thus, a more effective fermenta-tion method is essential. The immobilization method, which hasa number of excellent characteristics, such as biomass and produc-tivity improvement as well as prevention of cell loss, is recognizedas a practical tool for the enhancement of process stability andactivity of microorganisms involved in fermentation systems(Zhang and Yang, 2009). The production process of a series ofchemicals, including xanthan gum (Yang et al., 1996), enzymes(Bai and Yang, 2005), and organic acids (Zhang and Yang, 2009),could be effectively improved by the immobilized cell technique.Moreover, this method can also enhance the endurance capabilityof strain to substrates as well as the quality products (Tay andYang, 2002). Notably, sugarcane bagasse, which is an abundant,

Z. Xu et al. / Bioresource Technology 155 (2014) 8–14 9

inexpensive, and renewable lignocellulosic biomass, is an accept-able support material for cell immobilization (Plangklang et al.,2012). Therefore, sugarcane bagasse immobilization might be asuitable method to improve PGA fermentation.

In this study, a novel aerobic plant fibrous-bed bioreactor(APFB) was constructed for the production of high-viscosity PGA.This research aims to improve PGA production from glucose fur-ther by using Bacillus subtilis NX-2 immobilized in the APFB. More-over, a fed-batch strategy was applied to improve PGA productionbase on the analysis of the kinetics of immobilized-cell fermenta-tion, and the stability of the APFB system was evaluated. In addi-tion, the membrane fatty acid composition and enzyme activityof cells in both free and immobilized cell fermentation were alsoinvestigated.

2. Methods

2.1. Microorganism

B. subtilis NX-2, a PGA production strain, was originally isolatedfrom a soil sample in previous study (Xu et al., 2005). The strainwas deposited in the China General Microbiological Culture Collec-tion Center with the accession number CGMCC No. 0833.

2.2. Culture and medium

B. subtilis NX-2 was first inoculated into a 500 mL flask with80 mL of seed medium, which was composed of 20 g/L of glucose,15 g/L of glutamate, 5 g/L of yeast extract, 2 g/L of K2HPO4 � 3H2O,and 0.1 g/L of MgSO4. The strain was then aerobically incubatedat 32 �C for 16 h with shaking at 220 rpm. The fermentation med-ium contained 60 g/L of glucose, 50 g/L of glutamate, 5 g/L of(NH4)2SO4, 2 g/L of K2HPO4 � 3H2O, 0.1 g/L of MgSO4, and 0.03 g/Lof MnSO4.

2.3. Free-cell fermentation

Free-cell fermentation was performed in a 7.5 L bioreactor (Bio-Flo 110, New Brunswick Scientific, USA). The seed culture (2%, v/v)was then transferred to the bioreactor containing 4.5 L of basalmedium and was incubated at 32 �C with an agitation speed con-

Fig. 1. Structure of th

trolled at 400 rpm. The pH was automatically controlled at7.0 ± 0.1 by adding 2 N NaOH or 2 N HCl, and the aeration ratewas maintained at 1.2 vvm.

2.4. APFB construction

To improve the final concentration of PGA, APFB was employedto immobilize the cells of B. subtilis NX-2. The APFB system con-sisted of a 7.5 L bioreactor (BioFlo 110, New Brunswick Scientific,USA) and immobilized cells in jacketed glass column (height60 cm and tubing ID 4 cm) connected through a recirculation loopwith a peristaltic pump (Fig. 1). The temperature of the glass cell-immobilized column was controlled at 32 �C by a super thermo-static water bath, and the consistency was retained with the useof a stirred-tank fermenter. The bagasse was chopped into smallpieces (1–10 mm) and then dried. Thereafter, 30 g of bagasse wasloaded onto the glass cell-immobilized column. Stainless steel wiremesh was placed on top and at the bottom of the column to keepthe bagasse in place and to avoid leakage. Air pipes were installedat the bottom of the stirred-tank fermenter and glass column.

The empty bioreactor system was sterilized before use by auto-claving at 121 �C for 20 min and then cooled. The bioreactor sys-tem was then filled with 4.5 L of fresh fermentation medium andautoclaved for another 20 min. The APFB was then ready to be usedfor batch, fed-batch, and repeated fed-batch fermentation.

2.5. Immobilized-cell fermentation

For batch fermentation, PGA was produced in the APFB with aninitial glucose concentration of 60 g/L. Culture conditions were thesame as those for free-cell fermentation. Fed-batch fermentationwas also conducted in the APFB with a total glucose concentrationof 90 g/L. Once the residual glucose in both culture media was ex-hausted (below 10 g/L), the feeding solution (500 g/L glucose) wasadded to the bioreactor using a peristaltic pump to maintain theglucose level at approximately 10 g/L. For the repeated fed-batchfermentation experiment in the APFB, eight repeated fed-batch fer-mentations were applied to investigate the stability of PGA produc-tion by the APFB. The total glucose concentration of each batch was90 g/L. Cells immobilized on the surface and the apertures of thefibrous bagasse were used as the inoculums for the second batch.

e APFB system.

10 Z. Xu et al. / Bioresource Technology 155 (2014) 8–14

As soon as the residual glucose in the culture was exhausted (be-low 5 g/L), the fermentation broth was removed, and the sameamount of fresh culture medium was supplied. Samples were ta-ken out for analysis at regular time intervals throughout thefermentation.

2.6. Analytical methods

The biomass of free cells (m1) was determined from 10 mL ofcell suspension that was harvested by centrifugation at 10,000�gand 4 �C for 20 min, then washed with distilled water, and driedat 80 �C to a constant weight. To determine the biomass of immo-bilized cells (m2), the dry quality (m3) of immobilized bagassematerial was weighed before fermentation, the immobilized ba-gasse fiber collected from APFB was dried at 80 �C to a constantweight (m4), then the biomass of immobilized cells should be(m2 = m4 �m3), and the total biomass was (m1 + m2). The glucoseand L-glutamic acid concentrations remaining in the broth weremeasured enzymatically using a biosensor (SBA-40C, ShandongAcademy of Sciences, China). The specific glucose consumptionrate (qs, /h) and specific PGA formation rate (qp, /h) were estimatedfrom residual glucose concentration (s, grams per liter) and PGAproduction (p, grams per liter) (Mao and Zhong, 2004). The PGAconcentration was determined using a gel permeation chromatog-raphy (GPC) system with an RI-10 refractive-index detector and aSuperpose™ 6 column (Shimadzu Co.). Samples were passedthrough 0.22 lm filters and then treated according to the detailedprocedure described in the product information. Thereafter,50 mM of NaCl aqueous solution-acetonitrile (4:1, v/v) was usedas a mobile phase at a flow rate of 0.7 mL/min. The amount ofPGA was calculated from the peak area of the GPC measurementsusing purified PGA as a standard, and the average molecularweight of PGA was estimated by GPC (Birrer et al., 1994). Phospho-lipid extraction and analysis were conducted based on a previousreport (Mikhaleva et al., 1999). At the end of repeated fed-batchfermentation, the cells from the bagasse and subculture were col-lected and then cultured in the 80 mL seed medium at 32 �C to theexponential phase of phospholipids for cell membrane analysis.Fatty acids in phospholipids were prepared and analyzed as previ-ously described (Wollenweber and Rietschel, 1990), and palmito-leic acid (C16:1), oleic acid (C18:1), myristic acid (C14:0), cetylicacid (C16:0), and stearic acid (C18:0) were used as standards (Sin-opharm Chemical Reagent, AR, China). All experiments were con-ducted in triplicate.

Fig. 2. PGA production by B. subtilis NX-2 with free-cell (a) and immobilized cell (b)cultured by batch process. The initial glucose concentration was 60 g/L.

2.7. Assay of enzyme activity

The adapted and original cells cultured in the seed medium tothe exponential phase were collected, washed, and resuspendedin 100 mM Tris-HCl buffer (pH 7.5). The cell suspension was thensonicated at 400 W for 10 min to break the cell walls. The cell deb-ris was removed by centrifugation at 10,000�g for 1 h. The cell ex-tracts were kept at a low temperature using ice before beingassayed for enzyme activity. Isocitrate dehydrogenase (ICDH), glu-tamate dehydrogenase (GDH), and 2-ketoglutarate dehydrogenase(ODHC) activity in cell extracts was measured according to themethod described by Shirai et al. (2005). Enzyme activities weredetermined using an enzyme microplate reader (MK3, Thermo,USA) based on the appearance or disappearance of NADH orNADPH at 340 nm. One unit of activity was defined as the activityof enzyme catalyzing the formation of 1 lmol of NADPH per min.Total protein concentration was determined using the Bradford as-say (Bradford, 1976). Every experiment was repeated thrice, andthe experimental errors on enzyme activities were less than 5%.

2.8. Scanning electron microscopy

Scanning electron microscope was used to detect free andimmobilized cells (Hitachi S-3000 N, Japan). The samples werefixed by adding 2.5% (w/v) glutaraldehyde for 24 h at 4 �C. Thesamples were then cleaned thrice with potassium phosphate buf-fer. The samples were processed thrice with 50%, 70%, 80%, and90% (v/v) ethanol for 15 min and with 100% (v/v) ethanol for30 min. Tert-butyl alcohol was also used to dry the samples. Thesamples were coated with 10 nm of gold membrane and thenscanned and photographed using a scanning electron microscopeat 7 kV.

3. Results and discussion

3.1. Kinetics of free-cell and APFB immobilized-cell batch fermentation

The typical kinetics of batch fermentations with free cell in a7.5 L bioreactor and immobilized-cell in the APFB is shown inFig. 2. In the free-cell fermentation mode, 60 g/L glucose was com-pletely exhausted after 72 h (Fig. 2a). The maximum PGA concen-tration reached 32.86 ± 0.58 g/L with a PGA productivity of0.456 ± 0.002 g/L/h. By contrast, the glucose was exhausted after48 h of fermentation, resulting in 34.74 ± 0.61 g/L of PGA, with aproductivity of 0.723 ± 0.003 g/L/h in immobilized-cell fermenta-tion. These results are higher than those of free-cell fermentation

Fig. 3. Comparison of specific glucose consumption rate (a) and specific PGAformation rate (b) of B. subtilis NX-2 in batch free-cell and immobilized-cellfermentation.

Fig. 4. Fed-batch fermentation (a) and repeated fed-batch fermentation (b) byimmobilized B. subtilis NX-2 cells in the APFB. The initial glucose concentration was60 g/L, and the sugar concentration was maintained at 10 g/L during the fed-batchfermentation.

Z. Xu et al. / Bioresource Technology 155 (2014) 8–14 11

(32.86 g/L PGA with a productivity of 0.456 g/L/h). Meanwhile, theglucose conversion efficiency increased by 5.66% compared withthat of free-cell fermentation. In particular, the fermentation timeshortened from 72 to 48 h (Fig. 2b). More efficient production ofPGA was evidently achieved in the APFB than in free-cellfermentation.

A free cell biomass of 6.34 ± 0.27 g/L was obtained in the APFBimmobilized-cell fermentation. Moreover, a large amount of cellswas adsorbed onto the surface of the bagasse with a high cell den-sity for PGA fermentation according to the scanning electronmicrographs. The immobilized-cell fermentation in the APFB gen-erated a total biomass of 7.43 ± 0.39 g/L, which was 52.57% higherthan that of the free-cell fermentation (4.87 g/L). Previous studieshave shown that high cell density in the immobilized bioreactor,which could improve the tolerance of strain toward substratesand products, is better for PGA fermentation (Chen et al., 2012).

The specific glucose consumption rate (qs) and specific PGA for-mation rate (qp) of free-cell and immobilized-cell batch fermenta-tion were analyzed to understand the PGA fermentation processbetter in B. subtilis NX-2. As shown in Fig. 3, the qs and qp of B. sub-tilis NX-2 in immobilized-cell fermentation was higher than that infree-cell fermentation throughout the fermentation process. Themaximum qs (at �24 h) in batch immobilized-cell fermentationwas 1.26 times that in free-cell fermentation. The qp in batch

immobilized-cell fermentation reached its maximum at �24 h,and its maximum qp was 77.04% higher than that in free-cell fer-mentation. The results indicate that immobilized-cell batch fer-mentation is better for both glucose consumption and PGAproduction.

3.2. Fed-batch fermentation in the APFB

When the glucose was exhausted, the results indicated that thestain in the immobilized-cell batch fermentation still had PGA pro-duction capacity (Fig. 3b). Thus, a fed-batch fermentation strategywith glucose feeding in the APFB was proposed in order to increasePGA production further (Fig. 4a). As shown in Fig. 4a, an initial glu-cose concentration of 60 g/L and an additional 30 g/L was graduallyfed to the fermentation medium for PGA production. All of the glu-cose (90 g/L) was exhausted after 68 h of fermentation, resulting inthe production of 66.38 ± 0.67 g/L PGA. Compared with immobi-lized-cell batch fermentation, the maximum concentration andglucose conversion efficiency of PGA improved by 91.08% and27.46%, respectively. Meanwhile, the PGA productivity with fed-batch fermentation reached 0.976 ± 0.005 g/L/h, which was34.99% higher (0.976 vs. 0.723 g/L/h) than that with batch fermen-tation in the APFB. The fermentation results showed that thefed-batch strategy achieved better fermentation results than batchfermentation in the immobilized-cell fermentation, which wasconsistent with the results of feed control strategy in fibrous-bedbioreactor (Jiang et al., 2009).

12 Z. Xu et al. / Bioresource Technology 155 (2014) 8–14

3.3. Repeated fed-batch fermentation in the APFB

To maximize PGA production and evaluate the long-term per-formance of the APFB system, repeated fed-batch fermentationfor PGA production was investigated (Fig. 4b). Almost no lag phasewas observed, and faster substrate consumption rate was recorded,except in the first batch. The cells grew rapidly within 6 h afterreplacement with fresh medium. The first batch finished at68 ± 3 h when glucose had almost been used up (below 5 g/L).The second batch lasted for 59 ± 2 h, and subsequent batches lastedfor 55 ± 2 h each. The free cell biomass reached a maximum of9.01 ± 0.32 g/L in the first batch and then dropped and fluctuatedin the next batches, reaching an average concentration of approx-imately 7.38 ± 0.29 g/L in eight cycles.

The PGA yield from all eight batches fluctuated slightly from66.42 ± 0.71 to 74.53 ± 0.85 g/L, with an average of 71.21 ± 0.83 g/L. The PGA productivity varied from 0.976 ± 0.005 to1.355 ± 0.008 g/L/h, with an average of 1.246 ± 0.008 g/L/h. The re-sults demonstrate that PGA production in the APFB was highlystable.

3.4. Analysis of parameters under different fermentation cultureconditions

To compare the four fermentation culture modes better, the fer-mentation parameters of different fermentation modes were ana-lyzed. The free cell biomass and total biomass in four culturemodes were assessed (Table 1), and large amount of cells wereimmobilized onto the fibers to facilitate PGA fermentation (Zhanget al., 2010). Among the four cultures, the repeated fed-batch fer-mentation in the APFB showed the highest productivity and PGAyield of 1.246 ± 0.008 g/L/h and 71.21 ± 0.83 g/L, respectively. Incontrast to free-cell fermentation, the fermentation time was evi-dently shortened by immobilized-cell fermentation in the APFB.Moreover, a higher PGA productivity (0.723 vs. 0.456 g/L/h) wasachieved in batch immobilized-cell fermentation. Thus, more effi-cient PGA production was achieved in the APFB than free-cell fer-mentation. Fed-batch fermentation achieved a 34.99% higher PGAproductivity (0.976 vs. 0.723 g/L/h) than immobilized-cell batchfermentation, suggesting that more efficient PGA production canbe achieved using the fed-batch strategy. In repeated fed-batch fer-mentation, PGA was produced immediately after the initiation of anew batch, and the production reached a maximum average levelof 71.21 ± 0.83 g/L, 7.23% higher (71.21 vs. 66.38 g/L) in therepeated fed-batch fermentation mode than in the fed-batchfermentation. Furthermore, the yield as at a relatively high levelcompared with the values reported in the literature (Bajaj andSinghal, 2009,2010; Jeong et al., 2010). The repeated fed-batchfermentation achieved the best fermentation results among thefour different fermentation culture modes.

Table 1Analysis of parameters under different fermentation culture conditions.

Method Free cell batch fermentation AFB

Batch fer

Number of batches 1 1Total sugar per batch (g/L) 60 60Fermentation time per batch (h) 72 48Biomass of free cells (g/L) 4.87 ± 0.23 6.34 ± 0.Total biomass (g/L) 4.87 ± 0.23 7.43 ± 0.Max. PGA concentration (g/L) 32.86 ± 0.58 34.74 ± 0Molecular weights of PGA (�105 Da) 14.85 ± 0.04 14.86 ± 0PGA productivity per batch (g/L/h) 0.456 ± 0.002 0.723 ± 0c-PGA yield on total sugar (g/g) 0.548 ± 0.004 0.579 ± 0

a Average value for batches 1–8.

Numerous attempts have been made to improve PGAproduction, including the use of oxygen vectors or organic acid(Zhang et al., 2012a,b), fermentation modes and new bioreactors.However, the cell density, final PGA concentration, and productiv-ity remained low. It was reported that a higher PGA productivity of0.563 g/L/h could be achieved by two-stage control strategy duringthe batch fermentation (Zhang et al., 2011a). However, its PGA pro-ductivity was much lower than the value of 1.246 g/L/h observed inthe repeated fed-batch fermentation by using the APFB in the pres-ent study. The continuous parallel small scale shaken bioreactorsystem (COSBIOS) was an effective bioreactor used for PGA produc-tion of continuous cultivations (Wilming et al., 2013), however, theshaken culture method that the COSBIOS used was only suitablefor small-scale fermentation, which prevented production scale-up. By contrast, the APFB system could easily overcome theaforementioned problems. Moreover, the use of APFB for immobi-lized-cell production of PGA presented some advantages, such theabsence of lag phase during batch fermentation, high PGAproductivity and yield, and efficient continuous operation withoutrepeated sterilization and inoculation.

The molecular weights of PGA under different culture condi-tions had no significant difference between each other and main-tained a relatively stable level (Table 1). Similar results werefound that the molecular weights of PGA had no obvious changeswhen the strain was cultured under the same agitation and med-ium component (Bajaj and Singhal, 2011).

3.5. Analysis of the membrane fatty acid composition and enzymesactivities at 2-oxoglutarate branch point of APFB-adapted and originalcells

The consumption of extracellular glutamate and the secretion ofPGA were stimulated during the repeated fed-batch fermentationin the APFB. In addition, previous studies proved that microorgan-isms, such as Escherichia coli (Casey and Ingledew, 1986) and Sac-charomyces cerevisiae (Ingram, 1976), could increase theirtolerance to environmental stresses by regulating their membranelipid composition (Denich et al., 2003). Moreover, key enzymeactivities of substrates and final products of cells in fibrous-bedbioreactors are affected by stressors (Suwannakham and Yang,2005). Thus, the membrane fatty acid composition and enzymesactivities around the 2-oxoglutarate branch were analyzed in thestudy.

Compared with original cells, the content of myristic acid C14:0,cetylic acid C16:0, and stearic acid C18:0 increased by 0.9%, 1.1%,and 1.9%, respectively, which might improve the fluidity of cellmembrane in immobilized-cell fermentation (Table 2). By contrast,the content of unsaturated fatty acid and oleic acid C18:1 de-creased from 26.3% to 13.8% in the cell membrane of APFB-adaptedcells, respectively. Meanwhile, the palmitoleic acid C16:1 content

mentation Fed-batch fermentation Repeated fed-batch fermentation a

1 890 9068 57.13

27 8.17 ± 0.31 7.38 ± 0.3039 10.58 ± 0.51 11.69 ± 0.57.61 66.38 ± 0.67 71.21 ± 0.83.05 14.84 ± 0.03 14.87 ± 0.04.003 0.976 ± 0.005 1.246 ± 0.008.005 0.738±0.006 0.791±0.005

Table 2Changes in phospholipids in the cell membrane of APFB-adapted and original cells.

Myristic acid C14:0 (%) Cetylic acid C16:0 (%) Stearic acid C18:0 (%) Palmitoleic acid C16:1 (%) Oleic acid C18:1 (%)

Original cells 3.0 ± 0.12 18.2 ± 0.74 13.7 ± 0.47 19.48 ± 0.69 26.3 ± 0.81APFB-adapted cells 3.9 ± 0.13 19.3 ± 0.68 15.6 ± 0.52 15.24 ± 0.58 13.8 ± 0.49

Z. Xu et al. / Bioresource Technology 155 (2014) 8–14 13

decreased by 4.24% compared with that in the original cells. Previ-ous studies found that either the rise in the saturated fatty acids orthe decrease in the unsaturated fatty acids would decrease the per-meability of the cell membrane (Suwannakham and Yang, 2005),which suggests that the APFB-adapted cells have lower cell mem-brane permeability leading to greater tolerance of the substratesand final product. The results are consistent with those in previouspapers in that cell tolerance could be improved by immobilized-cell fermentation (Suwannakham and Yang, 2005; Zhang et al.,2011b).

The accumulation of 2-oxoglutarateis is influenced by specificactivities of ICDH and ODHC. Concretely, ICDH catalyzes the reac-tion from isocitrate to 2-oxoglutarate while ODHC catalyzes theoxidative decarboxylation of 2-oxoglutarate to succinyl-CoA.Besides, 2-oxoglutarate can also be converted by GDH toglutamate, the precursor of PGA (Zhang et al., 2011a). The specificactivity of ICDH and GDH in the APFB-adapted cells increased by34.11% (16.59 � 10�3 vs. 12.37 � 10�3 U/mg) and 9.15% (1.67 vs.1.53 U/mg), respectively, compared with the original cells. How-ever, the specific activity of ODHC in the APFB-adapted cells(0.77 � 10�3 U/mg) was nearly constant with that of original cells(0.78 � 10�3 U/mg). Either the increase in the ICDH and GDH activ-ity or decrease in the ODHC activity could result in the accumula-tion of intracellular glutamate (Wu et al., 2008). Moreover, theincrease in ICDH and GDH activity in the cells during immobi-lized-cell fermentation could improve PGA production. Similar re-sults were found in immobilized fermentation for organic acid andthe production for other biopolymers (Suwannakham and Yang,2005; Jin et al., 2013).

The APFB-adapted cells have lower cell membrane permeabilityand higher key enzyme activities than original cells, thus causingAPFB-adapted cells to have greater tolerance for substrates and fi-nal product and higher biosynthesis activity of PGA. These condi-tions may be the reasons why the consumption of substrates andPGA productivity were improved during APFB fermentation.

4. Conclusion

In this study, a novel APFB system was used for the effectiveenhancement of PGA production by B. subtilis NX-2. A high averagePGA concentration of 71.21 g/L was achieved in the repeat fed-batch immobilized fermentation. The bagasse in APFB was an effi-cient immobilization material, thus, high cell density fermentationfor PGA was achieved by APFB. Furthermore, the APFB exhibitedexcellent stability for PGA production by repeat fed-batch fermen-tation. This study indicates that this method is potentially usefulfor PGA production at an industrial scale.

Acknowledgements

This work was funded by the National Basic Research Programof China (973) (Nos. 2013CB733603, 2013CB733504), NationalHigh Technology Research and Development Program of China(863) (No. 2013AA020301), Jiangsu Provincial basic researchprogram (Natural Science Foundation) (No. SBK200910195), theNational Key Technology R&D Program (2011BAD23B04) andGraduate Student Innovation Project of Jiangsu Province (No.CXZZ13_0463).

References

Ashiuchi, M., 2010. Occurrence and biosynthetic mechanism of poly-gamma-glutamic acid. In: Hamano, Y. (Ed.), Amino-acid Homopolymers Occurring inNature. Microbiology Monographs, New York, pp. 77–93.

Bai, Y.L., Yang, S.T., 2005. Biotransformation of R-2-hydroxy-4-phenylbutyric acidby D-lactate dehydrogenase and Candida boidinii cells containing formatedehydrogenase coimmobilized in a fibrous bed bioreactor. Biotechnol. Bioeng.92, 137–146.

Bajaj, I.B., Singhal, R.S., 2009. Enhanced production of poly (c-glutamic acid) fromBacillus licheniformis NCIM 2324 by using metabolic precursors. Appl.Biochem. Biotechnol. 159, 133–141.

Bajaj, I.B., Singhal, R.S., 2010. Effect of aeration and agitation on synthesis ofpoly(gamma-glutamic acid) in batch cultures of Bacillus licheniformis NCIM2324. Biotechnol. Bioprocess Eng. 15, 635–640.

Bajaj, I., Singhal, R., 2011. Poly (glutamic acid) – an emerging biopolymer ofcommercial interest. Bioresour. Technol. 102, 5551–5561.

Birrer, G.A., Cromwick, A.M., Gross, R.A., 1994. C-Poly(glutamic acid) formation byBacillus licheniformis 9945: physiological and biochemical studies. Int. J. Biol.Macromol. 16, 265–275.

Bradford, M.M., 1976. A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dye binding.Anal. Biochem. 72, 248–254.

Casey, G.P., Ingledew, W.M., 1986. Ethanol tolerance in yeasts. Crit. Rev. Microbiol.13, 219–280.

Chen, F., Feng, X.H., Xu, H., Zhang, D., Ouyang, P.K., 2012. Propionic acid productionin a plant fibrous-bed bioreactor with immobilized Propionibacteriumfreudenreichii CCTCC M207015. J. Biotechnol. 164, 202–210.

Denich, T.J., Beaudette, L.A., Lee, H., Trevors, J.T., 2003. Effect of selectedenvironmental and physico-chemical factors on bacterial cytoplasmicmembranes. J. Microbiol. Methods 52, 149–182.

Ingram, L.O., 1976. Adaptation of membrane lipids to alcohols. J. Bacteriol. 125,670–678.

Jeong, J.H., Kim, J.N., Wee, Y.J., Ryu, H.W., 2010. The statistically optimizedproduction of poly(c-glutamic acid) by batch fermentation of a newlyisolated Bacillus subtilis RKY3. Bioresour. Technol. 101, 4533–4539.

Jiang, L., Wang, J.F., Liang, S.Z., Wang, X.N., Cen, P.L., Xu, Z.N., 2009. Butyric acidfermentation in a fibrous bed bioreactor with immobilized Clostridiumtyrobutyricum from cane molasses. Bioresour. Technol. 100, 3403–3409.

Jin, Y.X., Shi, L.H., Kawata, Y., 2013. Metabolomics-based component profiling ofHalomonas sp. KM-1 during different growth phases in poly(3-hydroxybutyrate) production. Bioresour. Technol. 140, 73–79.

Mao, X.B., Zhong, J.J., 2004. Hyperproduction of cordycepin by two-stage dissolvedoxygen control in submerged cultivation of medicinal mushroom Cordycepsmilitaris in bioreactors. Biotechnol. Prog. 20, 1408–1413.

Mikhaleva, N.I., Santini, C.L., Giordano, G., Nesmeyanova, M.A., Wu, L.F., 1999.Requirement for phospholipids of the translocation of the trimethylamine N-oxide reductase through the Tat pathway in Escherichia coli. FEBS Lett. 463,331–335.

Plangklang, P., Reungsang, A., Pattra, S., 2012. Enhanced bio-hydrogen productionfrom sugarcane juice by immobilized Clostridium butyricum on sugarcanebagasse. Int. J. Hydrogen Energy 37, 15525–15532.

Shih, I.L., Van, Y.T., 2001. The production of poly-(c-glutamic acid) frommicroorganisms and its various applications. Bioresour. Technol. 79, 207–225.

Shirai, T., Nakato, A., Izutani, N., Nagahisa, K., Shioya, S., Kimura, E., Kawarabayasi,Y., Yamagishi, A., Gojobori, T., Shimizu, H., 2005. Comparative study of fluxredistribution of metabolic pathway in glutamate production by twocoryneform bacteria. Metabol. Eng. 7, 59–69.

Sonaje, K., Chen, Y.J., Chen, H.L., Wey, S.P., Juang, J.H., Nguyen, H.N., Hsu, C.W., Lin,K.J., Sung, H.W., 2010. Enteric-coated capsules filled with freeze-dried chitosan/poly(c-glutamic acid) nanoparticles for oral insulin delivery. Biomaterials 31,3384–3394.

Suwannakham, S., Yang, S.T., 2005. Enhanced propionic acid fermentation byPropionibacterium acidipropionici mutant obtained by adaptation in a fibrous-bed bioreactor. Biotechnol. Bioeng. 91, 325–337.

Tajima, T., Ueno, S., Yabu, N., Sukigara, S., Ko, F., 2011. Fabrication andcharacterization of poly-c-glutamic acid nanofiber. J. Appl. Polym. Sci. 122,150–158.

Tay, A., Yang, S.T., 2002. Production of L(+)-lactic acid from glucose and starch byimmobilized cells of Rhizopus oryzae in a rotating fibrous bed bioreactor.Biotechnol. Bioeng. 80, 1–12.

Wilming, A., Begemann, J., Kuhne, S., Regestein, L., Bongaerts, J., Evers, S., Maurer,K.H., Büchs, J., 2013. Metabolic studies of c-polyglutamic acid production inBacillus licheniformis by small-scale continuous cultivations. Biochem. Eng. J. 73,29–37.

14 Z. Xu et al. / Bioresource Technology 155 (2014) 8–14

Wollenweber, H.W., Rietschel, E.T., 1990. Analysis of lipopolysaccharide (lipid A)fatty acids. J. Microbiol. Methods 11, 195–211.

Wu, Q., Xu, H., Shi, N.N., Yao, J., Li, S., Ouyang, P.K., 2008. Improvement of poly(c-glutamic acid) biosynthesis and redistribution of metabolic flux with thepresence of different additives in Bacillus subtilis CGMCC 0833. Appl. Microbiol.Biotechnol. 79, 527–535.

Xu, H., Jiang, M., Li, H., Lu, D.Q., Ouyang, P.K., 2005. Efficient production of poly(c-glutamic acid) by newly isolated Bacillus subtilis NX-2. Process Biochem. 40,519–523.

Yang, S.T., Lo, Y.M., Min, D.B., 1996. Xanthan gum fermentation by Xanthomonascampestris immobilized in a novel centrifugal fibrous-bed bioreactor.Biotechnol. Prog. 12, 630–637.

Zeng, W., Li, W., Shu, L., Yi, J.Y., Chen, G.G., Liang, Z.Q., 2013. Non-sterilizedfermentative co-production of poly(c-glutamic acid) and fibrinolytic enzyme bya thermophilic Bacillus subtilis GXA-28. Bioresour. Technol. 142, 697–700.

Zhang, A., Yang, S.T., 2009. Engineering propionibacterium acidipropionici forenhanced propionic acid tolerance and fermentation. Biotechnol. Bioeng. 104,766–773.

Zhang, Y., Feng, X.H., Xu, H., Yao, Z., Ouyang, P.K., 2010. e-Poly-L-lysine productionby immobilized cells of Kitasatospora sp. MY 5-36 in repeated fed-batchcultures. Bioresour. Technol. 101, 5523–5527.

Zhang, D., Xu, Z.Q., Xu, H., Feng, X.H., Li, S., Cai, H., Wei, Y., Ouyang, P.K., 2011a.Improvement of poly(c-glutamic acid) biosynthesis and quantitative metabolicflux analysis of a two-stage strategy for agitation speed control in the culture ofBacillus subtilis NX-2. Biotechnol. Bioprocess Eng. 16, 1144–1151.

Zhang, Y., Cong, W., Shi, S.Y., 2011b. Repeated fed-batch lactic acid production in apacked bed-stirred fermentor system using a pH feedback feeding method.Bioprocess Biosyst. Eng. 34, 67–73.

Zhang, D., Feng, X.H., Li, S., Chen, F., Xu, H., 2012a. Effects of oxygen vectors on thesynthesis and molecular weight of poly(c-glutamic acid) and the metaboliccharacterization of Bacillus subtilis NX-2. Process Biochem. 47, 2103–2109.

Zhang, H.L., Zhu, J.Z., Zhu, X.C., Cai, J., Zhang, A.Y., Hong, Y.Z., Huang, J., Huang, L., Xu,Z.N., 2012b. High-level exogenous glutamic acid-independent production ofpoly-(c-glutamic acid) with organic acid addition in a new isolated Bacillussubtilis C10. Bioresour. Technol. 116, 241–246.