Production of pentocin 31-1 by high cell density Lactobacillus pentosus 31-1 repeated batch cell recycle fermentations

Pentocin 31-1 was produced by Lactobacillus pentosus 31-1, isolated from the traditional Chinese fermented Xuan-Wei Ham. The wide range of antimicrobial activity especially to Listeria spp showed its potential use as food preservative. The study of pentocin 31-1 production was carried out in repeated batch cell recycle fermentations in 5 L bioreactor and up to 50 L pilot bioreactor. Batch fermentation of L. pentosus 31-1 gave bacteriocin production and productivity of 1395.3 IU/mL and 66.4 IU/(mL h), respectively. The higher bacteriocin production (2186.2 and 2004.5 IU/mL, respectively) and productivity (260.7 and 237.4 IU/(mL h), respectively) was measured in repeated batch cell recycle fermentations in 5 and 50 L bioreactors. Biomass (4.8 and 5.8 g/L, respectively) were also increased with repeated batch cell recycle fermentations in 5 and 50 L bioreactors, which were equivalent to 1.48-and 1.8-fold increases compared with the biomass (3.2 g/L) obtained from batch fermentation. Fermentations were successfully scaled up from 5 to 50 L because there were no significant differences (p<0.05) in production and productivity of bacteriocin between the two volumes by using the same strain and fermentation conditions. These results suggest a reasonable solution to develop large-scale production of both pentocin 31-1 and L. pentosus 31-1 in food industry.


INTRODUCTION
Bacteriocins are ribosomally synthesized peptides with a bactericidal activity against closely related bacteria and other bacteria of health or spoilage significance (Jack et al., 1995;Cleveland et al., 2001).In recent years, there are considerable interests in bacteriocins owing to their potential use in food preservation, the use of the bacteriocin-producing lactic acid bacteria in starter cultures or use as adjunct starter cultures (Chen and Hoover, 2006).Pentocin 31-1, a class IIa bacteriocin produced by Lactobacillus pentosus 31-1, was isolated from the traditional China fermented Xuan-Wei Ham and has a wide range of antimicrobial activity against both Grampositive and Gram-negative bacteria, including Listeria spp., Staphylococcus spp., Bacillus spp., Lactobacillus spp., Streptococcus spp., Pediococcus spp., and Escherichia spp.(Liu et al., 2008).The pH-resistant and heatstable characteristics of pentocin 31-1 showed its wide potential use as food preservative (Liu et al., 2008).Pentocin 31-1 has been shown to extend shelf life of chillstored nonvacuum-tray-packaged meat, and the bacteriocin-producing strain L. pentocus 31-1 has been considered as a novel functional starter culture or coculture *Corresponding author.E-mail: lipinglan420@126.com.Tel: +86 10 62737664.Fax: 86 10 62738678.
for sausage fermentation (Liu et al., 2010;Zhang et al., 2010).Based on the previous study, both of the use of bacteriocin and the producer strain are of particular interest to the food industry since they may help to ensure the microbial safety of the food products.
The limiting factor in using bacteriocin as food preservatives is their low yield during production (Bertrand et al., 2001).Since high biomass density proved to be beneficial for bacteriocin production (Yang and Ray, 1994;Parente and Ricciardi, 1999), it is likely that bacteriocin production could be enhanced in a system with high cell density.Some researchers attempted to increase cell concentration by using immobilized cell reactors, and it was demonstrated that bacteriocin production produced in this reactor was much higher than that of free-cell fermentations (Wan et al., 1995;Bertrand et al., 2001).With the development of membrane technology in the last few years, membrane processes have been increasingly used to separate microbial cells from the culture medium.Typical membrane bioreactors combine a continuous fermentor and a cross-flow microfiltration module enabling continuous cell recycle and removal of metabolic substances.In such recycle systems, the cell concentration can also reach high levels, leading to improved productivities.Membrane-based cell recycle fermentation has been proved to be efficient in producing primary and secondary metabolites such as lactic acid, xylitol, and ethanol (Bae et al., 2004;Nishiwaki and Dunn, 1999;Oh et al., 2003;Roca and Olsson, 2003).In addition, repeated batch fermentation gave several ad-vantages of timesaving processes for cleaning, sterilization and inoculation.There are also many approaches of repeated operations tried to increase the productivity of amino acids, ethanol, erythritol, xylitol, and crude glycerol (Hermann, 2003;Koh et al., 2003;Lu et al., 2003;Bae et al., 2004;Rywińska and Rymowicz, 2010).The advantages of integrated repeated-cycle batch cultures with immobilizedcell fermentation have been demonstrated for nisin (Bertrand et al., 2001) but there is no study concerning the productions of bacteriocin by integrating repeated batch fermentation with membrane-based cell recycles.
Appropriate exploitation of both cell biomass and bacteriocin can result in further improvement of the profitability of the fermentation process.In this study, repeated batch fermentation coupled with cell-recycle were developed to improve both biomass and bacteriocin production in modified MRS medium (mMRS) by using L. pentosus 31-1 and the scale was raised from a laboratory scale (5 L fermentor) to a pilot scale (50 L fermentor).

Microorganisms and media
L. pentosus 31-1, which produce pentocin 31-1, was isolated from Chinese tradition ham and maintained at -20°C in MRS containing 10% glycerol (De Man et al., 1960).Two successive inoculation at 10% (v/v; 12 h of growth at 30°C without pH control) were transferred into the seed cultivation broth before inoculating into the bioreactor (5%, v/v).For each fermentation experiment, the initial biomass did not exceed 0.7 g/L as dry weight and the initial cell concentration was 4-8×10 7 cfu/mL.
L. plantarum pl-2 was used as the indicator strain in the pentocin 31-1 assay.Cultures were stored in MRS containing 10% glycerol (De Man et al., 1960).

Batch fermentation of L. pentosus 31-1
Batch fermentation of L. pentosus 31-1 was carried out in a 5 L bioreactor (BIOTECH, Baoxing Biological Equipment Co., China) with a working volume of 3 L.The seed culture was prepared in a 500-mL flask containing 150 mL mMRS broth for 12 h at 30°C.Then, the seed culture was transferred to 5 L fermentor and followed by cultivation for 24 h at 30°C with uncontrolled pH.The foam formation was suppressed by adding Antifoam 289 (Sigma Chemical Co.,) during the initial stage of batch fermentation.The agitation was adjusted to 150 rpm to keep the culture homogenous.During the cultivation, bacteriocin activity, viable cell count, biomass, lactose consumption, lactic acid production and pH were analyzed.

Repeated batch cell recycle fermentations in 5 and 50 L bioreactors
The repeated batch cell recycle fermentations were carried out in the same 5 L bioreactor with a working volume of 3 L at 30°C, 5% (v/v) inoculum and 150 rpm, with uncontrolled pH.Two peristaltic pumps (Longer precision pump company, China) and a cross-flow type external polyethersulfone membrane (Beijing Jinxia super-filter equipments Co., Ltd, China) having the molecular cut-off of 10,000 Da were installed to filtrate culture broth and return the cells into the fermentor.One peristaltic pump was used for the recirculation between the fermentor and the external membrane, and the other one was connected to the permeate line of external membrane flux control.The apparatus (bioreactor and piping) was sterilized by steam, and the polyethersulfone membrane was prepared by soaking it in 5 mg/mL hydrogen peroxide for 2 h and then washed in sterile water.For the first 8 h, the cultivation was conducted in batch mode, then half of the culture liquid was pumped out and then the same volume of fresh fermentation medium was pumped in.After a short time (0.5 h) interval of pump circulation, this procedure was repeated 8 h later.In the whole 24 h fermentation process, no cell bleed was introduced in the system (total cell recycle).
In a 50 L fermentor (BIOTECH, Baoxing Biological Equipment Co., China), the first seed culture was prepared in a 500-mL flask in the manner as previously described.For the second seed culture, the first seed of 30 mL was inoculated to a 5 L jar containing 3.0 L of mMRS medium and then cultivated for 12 h at 30°C.The second seed culture of 1 L was transferred to the 50 L fermentor containing 20 L fermentation medium.The repeated batch cell recycle fermentations were carried out at 30°C with 150 rpm.The culture pH was controlled at 6.5 by automatic addition of 6 mol/L NaOH.
The membrane-based recycle fermentation system was used as described in 5 L bioreactor.After 8 h of batch fermentation, half of the culture liquid was replaced with the fresh mMRS medium with no cell bleed and after a 1.5 h interval of pump circulation the procedure was repeated 8 h later as it described in 5 L bioreactor.The time each pump circulation takes in 50 L bioreactor was longer than that in 5 L bioreactor for more culture liquid needed to be exchanged, and the whole fermentation time was 26 h.

Analysis Pentocin 31-1 activty
Production of pentocin 31-1 was determined by the nisin standard curve method, as described by Bober (Bober and Demirci, 2004).To eliminate the antimicrobial effect of lactic acid, the pH of the supernatants was adjusted to 6.0 with sterile 1 mol/L NaOH and the bacteriocin activity was expressed in IU/mL.

Bacterial growth and biomass production
Viable counts of L. pentosus 31-1 were determined by standard plate count on MRS agar after serial tenfold dilutions in sterile sodium chloride solution (0.85%, w/v) and expressed as colonyforming units (cfu/mL).
Biomass concentration was measured by cell dry mass determination by filtration of a 10 mL sample of the fermentation medium through 0.45 µm pore size filters and subsequent drying at 80°C for 24 h.

Lactose and lactic acid concentration
Lactose and lactic acid concentrations were measured by HPLC, as described by Aguirre-Ezkauriatza (Aguirre-Ezkauriatza et al., 2010).Samples were filtered through 0.45 µm pore size filters and diluted in distilled water as appropriate.

Statistics analysis
Each experiment was repeated twice and each determination was done in duplicate.The data were examined by analysis of variance (ANOVA) using SPSS 13.0 ( SPSS, Chicago, IL, USA) at a level of significance of p<0.05.Graphs of the fermentations were plotted using origin 8.0.

Batch fermentation
Figure 1 shows the fermentation profiles of L. pentosus 31-1 with batch fermentation in 5 L bioreactor.During the first 8 h of cultivation, the cell biomass increased to 3.1 g/L and then the culture passed to the stationary growth phase.Most of bacteriocin was accumulated in the medium during the stationary growth phase.Lactose consumption was initiated after 4 h and led to biosynthesis of lactic acid, which accumulated at 4.9 g/L at the end of the fermentation.The maximum biomass and bacteriocin pro-duction reached 3.2 g/L and 1395.3IU/mL respectively after 21 h, corresponding to a viable count of 9.2 cfu/mL in logarithm.From these results, the calculated volumetric productivity of bacteriocin was 66.4 IU/(mL h).

Repeated batch cell recycle fermentations in 5 and 50 L bioreactors
The fermentation profiles of L. pentosus 31-1 with repeated batch cell recycle fermentations in 5 L bioreactor are depicted in Figure 2. Cell-recycle fermentation equipped with polyethersulfone membrane was performed to return L. pentosus 31-1 cells to the bioreactor.L. pentosus 31-1 cells cannot penetrate the polyethersulfone membrane with 10,000 of molecular cut-off, but the metabolites like lactic acid as well as bacteriocin could effuse from the membrane.During the first 8 h of batch operation, a bacteriocin activity 176.0 IU/ml and a biomass concentration 3.4 g/L were achieved, while the lactose concentration decreased from 30 to 19 g/L and the lactic acid concentration added up to a level of 1.6 g/L.In the first batch cell recycle fermentation cycle, the sudden drops of bacteriocin and lactic acid were observed because of pumping out of 50% of the culture media and supplementing an equal volume of fresh fermentation medium.The initial bacteriocin activity and lactose concentrations for this cycle were 99.9 IU/mL and 26.9 g/L respectively.Within a period of 8 h, following bacteriocin production rates much higher to those observed for batch operation, and the bacteriocin activity rose to a peak of 2186.2IU/mL.The biomass continued to increase and reached 4.3 g/L at the end of this cycle.In a second batch cell recycle fermentation cycle, a similar sudden drop curve of bacteriocin and lactic acid were observed.The initial bacteriocin activity and lactose concentrations for this cycle were 920.1 IU/mL and 26.7 g/L respectively.After 7 h, the following trends the previously observed for bacteriocin and biomass production.The second recycle fermentation resulted in 2106.0IU/mL bacteriocin activity and 4.8 g/L biomass.The calculated volumetric productivity of bacteriocin in the first and second batch cell recycle fermentation cycle were 260.8 and 169.4 IU/(mL h) respectively, which were significantly higher than the values observed from batch fermentation.The maximum bacteriocin activity (2186.2IU/mL) obtained from the repeated batch cell recycle was about two times greater than that from the batch fermentation.The results show that extracting lactic acid from fermentation broth and feeding fresh medium to the reactor could increased both the yield of bacteriocin and the bacterial growth.
Laboratory-scale (5 L) L. pentosus 31-1 fermentation was increased to pilot (50 L) scale fermentation and the fermentation profiles are shown in Figure 3. Similar curves of bacterial growth, bacteriocin activity and biomass production were noted during the repeated batch cell recycle fermentations.In the first and second batch cell recycle fermentation cycles, the maximum bacteriocin activity were 2004.5 and 1982.2IU/mL respectively, correspon-ding to volumetric productivity of 237.4 IU/(mL h) and 166.6 IU/(mL h).Biomass showed a steady increase during the fermentation process and it increased to a maximum
Bacteriocin activities reached their maximum levels after 21 h of fermentation and did not declined rapidly in the following 3 h, however, our results contrasted those reported by Ghalfi et al. (2007) and Callewaert and De Vuyst (2000), who showed that bacteriocin activity de-clined sharply after reaching the maximum production.
The decrease of bacteriocin activity may be attributed to absorbing producer cells or degradation by specific protease and the former has been proven by most bacte-riocins (Meghrous et al., 1992;Parente et al., 1994;Van't Hul and Gibbons, 1996;Parente and Ricciardi, 2008).Since adsorption of pentocin 31-1 to cells is maximal at pH 5.0 and decreases at lower pH (Zhang et al., 2007), it is not surprising that no reduction of bacteriocin activity wasobserved in batch fermentation without pH control.
Lactose consumption occurs sharply when bacteriocin and lactic acid accumulates in the culture medium in the batch fermentation which indicates that mostly lactose was used to support bacteriocin and lactic acid production, thus providing enough lactose to the bacterium may improve bacteriocin production.On the other hand, Lactobacillus spp.growth is inhibited by lactic acid (Youssef et al., 2005) and growth limitation would also restrict bacteriocin production.In addition, the bacteriocin producing cells may be sensitive to their own bacteriocin due to the fact that the genes for bacteriocin producing and immunity are regulated and transcribed simultaneously for some bacteriocin producers (Abee, 1995).When nutrient components necessary for bacteriocin production become limited for bacteriocin production, bacteria will become more sensitive and results in a decrease of viable cell numbers.Therefore, repeated batch cell recycle operation was characterized because it was believed that it could remove lactic acid, bacteriocin and other inhibitory end products, supply fresh medium and sustain the productivity of a batch operation.
Membrane fermentation have been found to be successful in several fermentation studies to consistently produce higher biomass concentrations and increased product concentration as well as production rates like citric acid, lactic acid and mannitol (Xu et al., 2006;Racine and Saha, 2007;Rymowicz et al., 2010) but only a few reports exist on the production of antimicrobial proteins such as nisin and other bacteriocins.Moreover, the results obtained are far from being encouraging.Bhugaloo-Vial (Bhugaloo-Vial et al., 1997) described bacteriocin production in continuous fermentation with ceramic membrane was greater than that obtained from batch cultures, but the bacteriocin productivity was 2.8 to 10-fold lower than that obtained with batch cultures and the ceramic membrane was fouled after 102 h of conti-nuous fermentation.Nisin production using bench-scale repeated-batch fermentation of L .lactis in biofilm reactors with novel plastic composite supports are equivalent to suspended-cell control fermentation during either pH condition (Bober and Demirci, 2004).In our case, repea-ted batch cell recycle fermentations in 5 L bioreactor with a cross-flow type external polyethersulfone membrane can greatly increase both bacteriocin production and productivity compared with batch fermentation and the membrane was not fouled or blocked.A very high bacteriocin activity production (2186 IU/mL 16.5 h, 2106 IU/mL 24 h) was measured, which was approximately 2-fold higher than the maximum production (1395 IU/mL, 21 h) obtained for batch fermentation, and the maximum bacteriocin productivity of 260.8 IU/(mL h) was obtained in repeated batch cell recycle fermentations, which were equivalent to 3.4fold increases compared with the maximum productivity 76.8 IU/(mL h) obtained in batch fermentation.The increase of pentocin 31-1 production during repeated batch cell recycle fermentations correlated with increased cell production, which was more than 1.5-fold higher than that measured during batch fermentation (Table 1).Our results confirmed the previous hypothesis: the removal of the lactic acid and other inhibitory end products and the supply of fresh medium resulted in a significant prolongation of the exponential phase of growth and yielded higher bacteriocin production compared with batch fermentation.Moreover, the repeated batch cell recycle fermentations have advantages offered by continuous processes, for example, the decrease in the expenditure of sterilization and preparation of a bioreactor and inoculums preparation, and hence, increase the economical efficiency of pentocin 31-1 biosynthesis.
Scale up of bacteriocin production at a 50 L bioreactor validated the results obtained at laboratory scale.As shown in Table 1, the maximum bacteriocin production and productivity in the first and second cycle using the 50 L bioreactor were similar to that of 5 L bioreactor.However, cultivation of L. pentosus 31-1 in 50 L bioreactor yielded significantly (p<0.05)higher biomass than in 5 L bioreactor when the same medium and fermentation strategies were used.These results may attribute to the control of the pH at 6.5 in the 50 L bioreactor.Enhancement of growth by LAB by controlling the pH depending on the bacteriocin and the producer organism is well established (Kaiser and Montville, 1993;Yang and Ray, 1994).In spite of the different maximum cell counts, the similar bacteriocin production suggesting that there are other factors affecting bacteriocin production and maximization of growth does not necessarily result in maximization of bacteriocin production.The non-linear relationship between bacteriocin production and growth has also been observed for nisin and plantaricin C (Kim et al., 1997;Bárcena et al., 1998).Depletion of the energy source, lactose, at the end of each cycle fermentation or lactic acid accumulation may be the main factor for the restricted bacteriocin production.The complexity of bacteriocin biosynthesis and regulation had been used to explain the phenomenon (Parente and Ricciardi, 1999) but more studies are needed to find out the causes.
During the fermentation of repeated batch cell recycle fermentations in either 5 L or 50 L bioreactor, neither growth nor bacteriocin reached its maximum level after 24 and 26 h respectively of fermentation.Therefore, the production of L. pentosus 31-1 as well as bacteriocin could be further increased by increasing of the cyclic number or cultivation time.In addition, for optimal bacteriocin production, the energy source and other nutrients  should be sufficiently available during the whole fermentation in each cycle.An appropriate feeding strategy that matching the feeding rate of the limiting substrate with the rate of its utilization by the bacteria should maximize volumetric bacteriocin production, especially feeding energy source such as lactose at variable rate that are high enough to provide the necessary nutrients.

Conclusions
To our knowledge, this study is the first controlled repeated batch cell recycle application of pentocin 31-1 production.Our results clearly show that bacteriocin production can be increased significantly with repeated batch cell recycle fermentations compared with batch fermentation.In addition, the fermentation medium with high concentrations of pentocin 31-1 contains high cell biomass, which could be used as a biopreservative ingredient for many food applications.High yield of bacteriocin production from the laboratory scale to the pilot scales was successfully achieved; suggesting that the commercial production of pentocin 31-1 by repeated batch cell recycles fermentation holds great potential.
significantly different (p<0.05) if followed by a different lowercase letter.

Table 1 .
Fermentation parameters from duplicate Lactobacillus pentosus 31-1 in batch and repeated batch cell recycle fermentations.