Production of lactic acid using a new homofermentative Enterococcus faecalis isolate

Lactic acid is an intermediate-volume specialty chemical for a wide range of food and industrial applications such as pharmaceuticals, cosmetics and chemical syntheses. Although lactic acid production has been well documented, improved production parameters that lead to reduced production costs are always of interest in industrial developments. In this study, we describe the production of lactic acid at high concentration, yield and volumetric productivity utilizing a novel homofermentative, facultative anaerobe Enterococcus faecalis CBRD01. The highest concentration of 182 g lactic acid l−1 was achieved after 38 h of fed-batch fermentation on glucose. The bacterial isolate utilized only 2–13% of carbon for its growth and energy metabolism, while 87–98% of carbon was converted to lactic acid at an overall volumetric productivity of 5 g l−1 h−1. At 13 h of fermentation, the volumetric productivity of lactate production reached 10.3 g l−1 h−1, which is the highest ever reported for microbial production of lactic acid. The lactic acid produced was of high purity as formation of other metabolites was less than 0.1%. The present investigation demonstrates a new opportunity for enhanced production of lactic acid with potential for reduced purification costs.


Introduction
Lactic acid (C 3H6O3, MW 90.08) is a chiral, three-carbon carboxylic acid recognized as GRAS (generally regarded as safe) by the US Food and Drug Administration because of its unique physicochemical properties and the diverse applications in food and other chemical industries (Narayanan et al., 2004). The world production of lactic acid in 2009 was 258 000 tons, with a projected growth of 329 000 tons by year 2015, and 367 000 tons by 2017 (Global Industry Analysts Inc, 2011). The estimated consumption of lactic acid in the US alone is 30 million pounds per year with a potential annual demand of 5.5 billion pounds, and the global consumption of lactic acid is expected to increase rapidly in the near future (Bajpai, 2013). This growth would mainly come from the lactic acid consumption in chemical applications. For example, the use of polylactic acid and ethyl lactate is expected to expand 19% per year (Jarvis, 2001;Garlotta, 2002;Södergård and Stolt, 2010). With a current price of $1.5/kg for a 88% purity food-grade product (www .icispricing.com), lactic acid has the potential to become a very large volume, commodity-chemical intermediate (Bozell and Petersen, 2010).
Lactic acid can be produced chemically using petrochemical feedstocks such as lactonitrile (Narayanan et al., 2004). However, the chemical synthesis produces a racemic mixture of D(−)-lactic acid and L(+)-lactic acid, which is not suitable for the synthesis of polylactic acid, as the optical purity of lactic acid is one of the major factors that determine the physical properties of its polymer. In contrast, about 90% of current commercial lactic acid is obtained via biological fermentation of sugars (Hofvendahl and Hahn-Hagerdal, 2000;Zhou et al., 2003). Microbial fermentation of sugars produces stereospecific D(−)-lactic acid or L(+)-lactic acid depending on the strains used. For example, Lactobacillus, Bacillus, Rhizopus, Streptococcus and Enterococcus produce L(+)lactic acid, while Leuconostoc and La. vulgaricus produce D(−)-lactic acid (Park et al., 2010). L(+)-lactic acid is the preferred component of many food and industrial applications that is currently produced via biological fermentation utilizing lactic acid bacteria or fungi such as Rhizopus. Some recombinant yeast strains have also been used to produce lactic acid from various carbon feedstocks (Porro et al., 2008). However, the yield and productivities of fungal and yeast strains are very low compared with lactic acid bacteria. In addition, the mycelial morphology of fungal strains affects the viscosity of the fermentation medium and can cause blockage during fermentation (Maas et al., 2006;Zhang et al., 2007;Ilmen et al., 2013). Therefore, these organisms require further gene manipulations and process development for the industrial production of lactic acid.
Recently, we have isolated a bacterial strain with high potential for lactic acid production. This strain was identified as Enterococcus faecalis and partially characterized. Subsequently, its full genome was sequenced, and the isolate E. faecalis CBRD01 was deposited with the American Type Culture Collection with a patent deposition designation PTA-12846 (unpublished results). Here we describe the abilities of E. faecalis CBRD01 to produce L(+)-lactic acid from glucose at high yields, titres and productivity with a minimum by-product formation at both flask and bioreactor scales.

Morphology and biochemical characteristics of isolate CBRD01
Enterococcus faecalis CBRD01 belongs to Gram-positive elongated Cocci with cell size of approximately 1.2-1.5 μm. CBRD01 was able to grow anaerobically under dark conditions at temperature between 30 and 45°C, with a temperature optimum of 37°C. CBRD01 was able to utilize trehalose, lactose, ribose, saccharose, melizitose, cellobiose, mannose, mannitol, sorbitol and inositol ( Table 1). The strain did not show oxidase, catalase, alkaline phosphatase, β-galactosidase, aminopeptidase and urease activities. The cellular fatty acid compositional analysis indicated that the strain CBRD01 had a match with E. faecalis (data not shown).

Production of lactic acid by E. faecalis CBRD01 at flask scale
In order to examine the production of lactic acid, the strain E. faecalis CBRD01 was cultured in 100 ml serum bottles with 50 ml of LA5 medium containing three different concentrations of glucose (28.79, 56.13 and 110.22 mM) at 37°C and 150 r.p.m. in an orbital incubator shaker under anaerobic condition. The cells were inoculated at 0.1 ± 0.01 Optical Density at 600 nm (OD600), and fermentation was continued for 12 h. The results revealed that a maximum yield of lactate was obtained at 110.22 mM glucose (Table 2). On average, 27-34 mM of glucose was consumed regardless of the initial concentrations of glucose. However, a reverse relationship was observed between the initial glucose concentration and glucose consumption after 12 h of incubation. The glucose consumption at lower starting concentrations (28.79 and 56.13 mM) was 93% and 60% respectively. This was 3.3-and 2.1-fold, respectively, higher than that at 100.22 mM glucose (28%). The lactate concentration varied from 49 to 61 mmol l −1 . In theory, 1 mol of glucose yields a maximum of 2 mol of lactate under anaerobic condition. Accordingly, the glucose consumption at 110.22 mM glucose concentration was 30.55 mM. The lactate titre was 60.21 mmol l −1 after 12 h, thus yielding 1.97 mol mol −1 glucose, which is equivalent to 98.54% of the theoretical maximum. However, it should be noted that lactate production ceased after 12 h, although the experiment was carried out for 24 h.

Production of lactic acid by E. faecalis CBRD01 in batch process at bioreactor scale
To investigate the potential of E. faecalis CBRD01 for lactic acid production at bioreactor scale, the strain was cultured under controlled conditions at a constant pH of 7.0 ± 0.2 in a 1 l glass jar bioreactor (DASGIP) with a working volume of 0.5 l for 24 h. Results are shown in Fig. 1. The initial glucose concentration was 94 mM. In 24 h of the bioreactor culture, E. faecalis CBRD01 Glucose produced 172.8 mmol lactate l −1 , which is 2.87-fold higher than the titre obtained from the flask culture. The biomass production reached a maximum of 0.64 g l −1 at 12 h (from 0.07 g l −1 at 0 h) and then decreased to 0.62 g l −1 at 24 h (Fig. 1A). Both the specific glucose uptake rate and specific lactate production rate peaked at 12 h of cultivation ( Fig. 1B), with the former being 0.57% of the latter. At 12 h, a maximum volumetric productivity of 10 mmol −1 l −1 h −1 was attained (Fig. 1C). Therefore, to better understand the behaviour of CBRD01, it was of interest to discuss its fermentation profile in two phases (phase 1, 0-12 h; phase 2, 12-24 h). The total glucose consumption at 24 h was 94.06 mmol l −1 (Table 3). Although the lactate yield in the first half of cultivation (0-12 h) was 88%, it improved to 98% in the second (12-24 h), resulting in an overall yield of 91.86 % of theoretical maximum of glucose. The μmax, estimated between 0 and 3 h, was 0.59 h −1 . The maximum specific production rate (qmax) of lactate, estimated between 0 and 12 h (phase 1), was 66.72 mmol g cdw −1 h −1 . In phase 2 (12-24 h), qmax decreased to 15 mmol g cdw −1 h −1 .
Lactate was the major product of the fermentation process with E. faecalis CBRD01. Other metabolites, such as acetate and formate, were produced in quantities of less than 4% in the 1st phase and less than 0.6% in the 2nd phase (Table 4). Analysis of the carbon material balance revealed that 101.16% of glucose carbon was recovered in phase 1 (0-12 h) and 98.9% in phase 2 (12-24 h). In phase 1, 87.9% of glucose carbon was converted as lactate, while 6.2% of carbon was directed towards biomass formation. However, no carbon was utilized for biomass in the second phase, while 0.6% of carbon was utilized for acetate and 0.3% for formate, thus yielding 98% of lactate in the second phase. The formate and acetate co-metabolites were produced in very low quantities of less than 1 g l −1 . The electron balance, calculated for the substrates and products, indicated 0.99-1.0 (Table 4). This means that the electrons, released during glucose oxidation, were completely recovered in the form of products. Based on the carbon distribution of glucose into metabolites (no metabolites were omitted during analysis), a carbon metabolic pathway for E. faecalis CBRD01 is proposed in Fig. 2, suggesting that this strain utilizes a homolactic fermentative pathway to produce lactic acid.   A. Glucose consumption, biomass and lactate production. B. Specific rates of glucose uptake and lactate production. C. Volumetric productivity of lactate.

Production of lactic acid by E. faecalis CBRD01 in fed-batch process at bioreactor scale
To further evaluate the potential of E. faecalis CBRD01 for lactic acid production, a fed-batch process was carried out in a 1 l glass jar bioreactor with a 0.5 l working volume of LA5 medium at pH 7.0 and 37°C for 87 h under anaerobic conditions. The initial glucose concentration was 86.4 mM. During the fermentation period, a total of 825 mM of glucose was added in 14 batches (Fig. 3). The fed-batch process was initiated by adding 3.0 g l −1 of E. faecalis CBRD01 cells, grown under anaerobic conditions, to the culture medium. The total glucose consumption at the end of the fermentation process was 823 mmol l −1 . The strain produced 973 mmol l −1 lactate in 87 h. This translates into a lactate yield of 59% from glucose, with an average productivity of 1.0 g l −1 h −1 (Fig. 3). However, the lactate yield in the first phase (0-12 h) was above 91%. When compared with the batch production of lactic acid after 24 h ( Fig. 1), the fed-batch process increased lactic acid production approximately threefold (532 mM).

Fed-batch bioreactor production of lactic acid by E. faecalis CBRD01at high cell density culture
In an attempt to increase the lactic acid titres, fed-batch experiments were performed using a high cell density   a. The standard deviation of the measurements was less than 3% for the substrate and the metabolites. b. Degree of reduction (γi), which was calculated using the formula: γi = (4a + b − 2c + 6e + 5f)/a, where a denotes carbon, b denotes hydrogen, c denotes oxygen, e denotes sulfur, and f denotes phosphorus (Gustafsson et al. 1993). c. Degree of reduction (γi) balance, which was calculated as γi for total products/γi for substrate (expressed in mol l −1 ). d. g l −1 .
e. An average molecular weight of 24.6, which corresponds to an average cell with a molecular formula of CH1.8 O0.5 N0.2. The average ash content of 8% was deduced from the actual cell dry mass (Stephanopoulos et al., 1998). f. Optically pure L(+)-lactic acid. g. Total carbon in biomass and metabolites, Total carbon in glucose −1 × 100. culture (22 g l −1 ) and LA6 culture medium in a 1 l glass jar bioreactor (0.3 l working volume). Lactic acid production was carried out under anaerobic conditions at pH 7.0 and 37°C for 38 h (Fig. 4). At the end of the fermentation process, E. faecalis CBRD01, grown in LA6 medium, was able to produce 2022 mmol lactate l −1 , which is equivalent to 182.1 g l −1 . The total glucose consumption was 210.01 g with an overall lactate yield of 86.7% from the theoretical maximum on glucose. The microbial biomass, retained at the end of fermentation, was 16.86 g cdw l −1 . About 811 mM of glucose was consumed in 13 h (before glucose addition) to produce 1484 mmol lactic acid l −1 , which represents a 91.5% yield of the theoretical maximum on glucose or 1.83 mol lactate mol −1 glucose. The volumetric productivity accounted between 0 and 13 h was 114 mmol −1 l −1 h −1 (10.3 g l −1 h −1 ), which is the highest volumetric productivity ever reported for microbial production of lactic acid. As the energy requirement for maintaining high cell density is very high, there was no net cell growth. The biomass production remained unchanged for 13 h and thereafter declined due to the dilution effect that occurred with the glucose spiking.

Discussion
The production of lactic acid for commercial applications requires the use of less fastidious, but robust microorganisms that can meet the industrial needs for high titre, high productivity and high yield of lactate of at least 100 g l −1 , 2.5 g l −1 h −1 and 80% respectively (Werpy and Petersen, 2004). As purification costs comprise a significant portion of the overall production costs, the purity of food grade (80-90% purity) and pharmaceutical grade (> 90%), lactic acid determines its price (John et al., 2007). Therefore, to minimize the production and purification costs, there is a need of new microbial isolates with enhanced capabilities for high-purity lactic acid production.
Here, we describe a wild-type bacterial strain, recently isolated and identified as E. faecalis CBRD01. This strain is different from other known lactic-producing  ferment the previous carbon sources to lactic acid (Manero and Blanch, 1999). Enterococcus faecalis CBRD01was able to produce high yields and titres of lactic acid with minimal impurities on glucose. The response of CBRD01 to different glucose concentrations is of particular interest because of the direct impact of parameters such as substrate inhibition, viscosity of culture medium, by-product formation on the cellular metabolisms and product formation (Booth, 1985;Hall et al., 1995). Initial glucose concentrations of up to 110 mM had no adverse effect on cell growth of CBRD01 as μ max (estimated between 0 and 3 h) remained constant (0.59 h −1 ) at 29 and 56 mM glucose, and slightly increased (0.64 h −1 ) at 110 mM glucose. Acetate and formate were the only cometabolites detected at concentrations of less than 1 mM, suggesting that the formation of lactate by-products and their inhibitory role on glucose fermentation was negligible. However, the increase in the initial glucose concentrations from 29 to 110 mM did not improve glucose consumption. In fact, at 110 mM glucose, the total glucose consumption was lower than at 56 mM glucose. Furthermore, the rate of glucose uptake did not change substantially and remained in the range of 19.9-23.9 mmol g −1 cdw h −1 at all three glucose concentrations. This suggests that (i) the CBRD01 cells were not under stress at high glucose concentrations (110 mM), and (ii) the cellular metabolism of lactate production was not influenced by the glucose concentration. The latter was evident from the lactate yield that was in the range of 90-99% at all three glucose concentrations.
While examining the pH of the culture broth after 12 h of fermentation, it was noticed that pH decreased to 5.3 ± 0.3 from the initial pH 7.0, regardless of the initial glucose concentrations. The pH drop, which was due to the formation of lactic acid, could be responsible for the cease of further biomass and lactate production in E. faecalis CBRD01. In order to understand the effect of pH on improving lactate production, a batch process was conducted under pH-controlled conditions in a 1 l glass jar bioreactor. When the pH of the culture medium was maintained at 7.0 ± 0.2 throughout fermentation, 173 mmol lactate l −1 was produced from 94 mmol glucose l −1 , which is equivalent to a 92% lactate yield on glucose carbon. Hence, the culture pH had a major impact on glucose metabolism and lactate production.
Although the batch process showed a 2.9-fold improvement in lactate production compared with shake flask experimental results, the supply of initial glucose was seen as a limiting factor for increased lactate production. Therefore, a fed-batch process was designed to better understand the lactate production potential of E. faecalis CBRD01. In a separate study, the effect of inoculum load was examined (data not shown). It was observed that at a given lactate titre, the use of higher inoculum densities (on dry weight) reduced the fermentation time to reach that titre. Alternatively, a higher lactate titre can be attained at the same fermentation time. For example, a fed-batch process with 3 g l −1 inoculum increased lactate production to 973 mmol l −1 from 823 mM of glucose in 87 h. When accounted at 24 h, the lactate titre was 541 mmol l −1 , and the total glucose consumed was estimated at 338 mmol l −1 . This corresponds to a lactate yield of 80% on glucose, which is threefold higher than the batch production of lactate after 24 h. However, at 13 h of fed-batch fermentation, the biomass concentration declined to 2.85 g l −1 from 3.0 g l −1 . Although the culture medium was maintained at glucose concentrations in excess of 60 mM, the carbon to cell growth and biomass maintenance with CBRD01 were low. This necessitated addition of complex nitrogen sources such as yeast extract to maintain biomass concentration. To further improve the lactate production in E. faecalis CBRD01, fermentation parameters including medium composition, initial glucose concentrations and inoculum density were varied and modified. CBRD01 growth and lactate production were examined as a function of these parameters (data not shown). Results suggested the possibility of improving lactate production by using a high cell density culture of 22 g l −1 cells and a high glucose concentration of up to 1 M on an altered medium LA6. Under those conditions, CBRD01 produced 2.02 mol lactate l −1 (equivalent to 182 g lactate l −1 ) at an overall lactate yield of 87% on glucose carbon. The overall volumetric rate of lactate production was estimated to be 56 mmol l −1 h −1 (equivalent to 5 g lactate l −1 h −1 ). A similar strain, E. faecalis RKY1 was previously reported to ferment glucose at an optimal concentration of 150 g l −1 to lactic acid with a final titre of 144 g l −1 , at an optimal productivity of 5.1 g l −1 h −1 (Yun et al., 2003;Wee et al., 2004;2006a,b). In comparison, the present investigation demonstrates a higher lactate titre at a similar volumetric productivity.
Recently, two Bacillus strains were used to produce lactic acid (Ou et al., 2011;Meng et al., 2012). High lactic acid titres were obtained on glucose and xylose. For example, a thermotolerant B. coagulans 36D1 strain produced 182 g l −1 of lactic acid with a carbon yield of 92.3% on glucose following a fed-batch fermentation for 261 h. The same strain could also produce lactic acid at 163 g l −1 from 86.3 g l −1 of xylose, thus yielding 87.3% on xylose carbon equivalents. However, the volumetric productivities of lactic acid production estimated from the fedbatch process with both glucose and xylose were less than 0.84 g l −1 h −1 (Ou et al., 2011). In another report (Meng et al., 2012), an alkaliphilic Bacillus sp. WL-S20 was shown to produce a maximum of 225 g l −1 of lactic acid in a fed-batch process at a productivity rate of 1.04 g l −1 h −1 , while a single-pulse feeding of glucose yielded a maximum of 180 g l −1 of lactic acid at 1.61 g l −1 h −1 after 112 h of fermentation. Both Bacillus strains were very efficient in producing lactic acid at high concentrations, comparable or exceeding those of E. faecalis CBRD01, but at threefold to fivefold lower volumetric productivity rates.
Enterococcus faecalis CBRD01, believed to use a homolactic fermentative pathway to produce L(+)-lactic acid, utilized 2-13% of carbon for its growth and energy metabolism, with the rest of the carbon converted to lactic acid at an overall volumetric productivity of 5 g l −1 h −1 . Formation of other metabolites such as acetate and formate was less than 0.1%. The fermentation characteristics and lactic acid-producing capabilities of CBRD01, as presented in this work, can certainly compete with existing published literature. Following further optimization, E. faecalis CBRD01 may become a strong candidate for industrial production of lactic acid.
A 0.1 ml of lactate-producing microbial culture was placed onto agar plates, containing selective medium, and incubated under anaerobic condition. In total, 55 different microbial strains were isolated from the soil samples (data not shown). Cultures that displayed rapid growth on glucose with lactic acid as the primary metabolite were resorted for further isolation and verified for purity by streaking on medium solidified with agar. The most promising isolate with high lactateproducing capability on glucose was referred to as CBRD01 and used in this study. The strain CBRD01 was biochemically characterized (Table 1) and identified as E. faecalis by DSMZ-Germany (http://www.dsmz.de/). This strain was referred to as E. faecalis CBRD01.

Flask cultivation of E. faecalis CBRD01
To examine the production of lactic acid, the strain E. faecalis CBRD01 was cultured in 100 ml serum bottles with 50 ml working volume in an orbital incubator shaker (Innova R42, Eppendorf, USA) at 150 r.p.m. under anaerobic condition. The lactic acid medium (LA5 medium) contained the following components per litre of deionized water: NH4Cl, 1.6 g; MgSO4·7H2O, 0.25 g; NaCl, 1.0 g; FeCl3·6 H2O, 2.5 mg; Na2S2O3·5 H2O, 1.0 g; yeast extract, 3.0 g; Pfennig and Lippert's trace element solution (PL 9), 1.0 ml. The pH of the culture medium was adjusted to pH 7.0 with 100 mM potassium phosphate buffer. Three different concentrations of glucose, 25 mM (4.5 g l −1 ), 50 mM (9 g l −1 ) and 100 mM (18 g l −1 ), were used in the fermentation studies. The cells were inoculated at 0.1 ± 0.01 OD600, and fermentation was carried out at 37°C for 12 h. Samples were withdrawn periodically to determine cell mass, residual glucose and metabolites.

Bioreactor cultivation of E. faecalis CBRD01
A 1 l glass jar bioreactor (DASGIP-FB04CS, Eppendorf, Germany) with a working volume of 500 ml of LA5 medium containing 100 mM potassium phosphate buffer (pH 7.0) was used. Glucose was added at 86.4 mM. The dissolved oxygen was expelled by flushing the reactor with nitrogen gas (99.99%) for 30 min at 1.0 vol vol −1 min −1 . The strain was grown overnight and then inoculated at an OD600 of 0.2 into the medium. The temperature and agitation rate were maintained at 37°C and 250 r.p.m. respectively. The pH was maintained at 7.0 ± 0.1 with 2.5 N NaOH and 2.5 N HCl. Samples were withdrawn periodically to determine the biomass, residual substrate and metabolite concentration. Fermentation was continued for 72 h. For fed-batch fermentation, concentrated glucose (4.0 M) was added intermittently. The glucose content in the medium was monitored periodically using high-performance liquid chromatography (HPLC), and its concentration was maintained above 60 mM until 60 h.
An improved fed-batch process was carried out at 37°C in a 1 l glass jar bioreactor with a 0.3 l working volume at high cell density using LA6 medium, which contained the following components per litre of deionized water: yeast extract, 18 g; tryptone, 9.0 g; dipotassium phosphate, 9.38 g; monopotassium phosphate, 1.73 g; Lippert's trace element solution, 1.0 ml. After sterilization, the fermenters were flushed with N2 gas for 1 h to ensure oxygen-free environment. Fermentation was initiated by adding E. faecalis CBRD01 cells, grown under anaerobic conditions, at 22 g l −1 . Glucose was added at 0.86 M to the fermenter. The pH of LA6 was maintained at pH 7.0 ± 0.2 using 10 N NaOH and 5.0 N HCl solution throughout fermentation. Experiments were conducted for 38 h.

Analytical methods
Cell concentrations were measured in a 10 mm path length cuvette using a UV-2450 double-beam spectrophotometer (Shimadzu, Kyoto, Japan) at 600 nm. One unit of absorbance at 600 nm corresponded to 0.3 g cell dry weight (cdw) l −1 . Concentrations of glucose and fermentation metabolites (lactate, acetate, formate) were determined by the method described previously using Shimadzu LC20 HPLC (Shimadzu) equipped with a refractive index detector (model−RID-10A, Shimadzu) (Talluri et al., 2013). Briefly, the supernatants, obtained by centrifugation of the culture samples at 10 000 × g for 10 min, were filtered through the Nylon-membrane (Cole-Parmer, Vernon Hills, IL, USA) and eluted through a 300 × 7.8 mm Aminex HPX-87H (Bio-Rad, Hercules, CA, USA) column at 60 o C using 5.0 mM H2SO4.

Calculation of fermentation parameters
The maximum specific growth rate (μmax; h −1 ) was estimated by plotting the total cell concentration against time in a loglinear plot. The slope of the straight line during exponential growth was used to find the average specific growth rate. Specific rates (glucose uptake rate or lactate production rate) were calculated by dividing the volumetric rates by the time average concentration of cells (mmol g cdw −1 h −1 ). Consumption of glucose and production of lactate were used to calculate the volumetric rates or productivities (mmol l −1 h −1 ). Lactate yield on glucose (mol mol −1 ) was estimated by dividing lactate concentration (mol) by glucose consumption (mol). Biomass yield (g cdw g glucose −1 ) was estimated from the cell mass produced (on cell dry weight) per gram of glucose consumed.