Ann Microbiol (2017) 67:763–778 https://doi.org/10.1007/s13213-017-1306-y

ORIGINAL ARTICLE

Characterization of D-lactic acid, spore-forming bacteria and Terrilactibacillus laevilacticus SK5-6 as potential industrial strains

Budsabathip Prasirtsak1 & Sitanan Thitiprasert2 & Vasana Tolieng 2 & Suttichai Assabumrungrat3 & Somboon Tanasupawat4 & Nuttha Thongchul2

Received: 4 July 2017 /Accepted: 11 October 2017 /Published online: 24 October 2017 # Springer-Verlag GmbH Germany and the University of Milan 2017

Abstract In this study, we screened and isolated D-lactic acid- (inoculum age) and proper concentration of cell mass (inocu- producing bacteria from soil and tree barks collected in lum size), T. laevilacticus rapidly converted glucose into D- Thailand. Among the isolates obtained, Terrilactibacillus lactate under anaerobic conditions, resulting in a high final laevilacticus SK5-6 exhibited good D-lactate production in lactate titer (102.22 g/L), high yield (0.84 g/g), and high pro- the primary screening fermentation (99.27 g/L final lactate ductivity (2.13 g/L⋅h). When the process conditions were titer with 0.90 g/g yield, 1.38 g/L⋅h, and 99.00% D-enantiomer shifted from an aerobic to an anaerobic environment, unlike equivalent). Terrilactibacillus laevilacticus SK5-6 is a Gram- other lactate-producing bacteria, the mixed acid fermentation positive, endospore-forming, homofermentative D-lactate pro- route was not activated in the culture of T. laevilacticus SK5-6 ducer that can ferment a wide range of sugars to produce D- during the fermentation stage when some trace oxygen still lactate. Unlike the typical D-lactate producers, such as remained. Our study demonstrates the excellent characteristics catalase-negative Sporolactobacillus sp., T. laevilacticus of this isolate for D-lactate production; in particular, a high SK5-6 possesses catalase activity; therefore, a two-phase fer- product yield was obtained without byproduct formation. mentation was employed for D-lactate production. During an Based on these key characteristics of T. laevilacticus SK5-6, aerobic preculture stage, a high-density cell mass was rapidly we suggest that this isolate is a novel D-lactate producer for use obtained due to aerobic respiration. When transferred to the in industrial fermentation. fermentation stage at the correct physiological stage Keywords Terrilactibacillus laevilacticus . Two-phase The GenBank/EMBL/DBBJ accession number for the 16S rRNA of fermentation . D-lactic acid . Catalase positive . strain SK5-6 is LC222555; for strain BRY67-1, LC222556; for strain Homofermentative BRY67-2, LC222557; for strain BRY67-3, LC222558; for strain NK44-2, LC222559; for strain SP43-2, LC222560.

* Nuttha Thongchul Introduction [email protected] Lactic acid is widely used in the pharmaceutical and chemical 1 Program in Biotechnology, Faculty of Science, Chulalongkorn industries as well as for many other applications including, for University, Phayathai Road, Wangmai, Pathumwan, Bangkok 10330, example, as an acidulant, a flavor enhancer, and a food pre- Thailand servative. The relatively recent global awareness of the large 2 Institute of Biotechnology and Genetic Engineering, Chulalongkorn consumption of non-renewable plastic products has driven the University, Phayathai Road, Wangmai, Pathumwan, Bangkok 10330, Thailand development of compostable plastics from the renewable re- sources as an alternative. Polylactic acid (PLA) is one of the 3 Department of Chemical Engineering, Faculty of Engineering, Chulalongkorn University, Phayathai Road, Wangmai, Pathumwan, more promising bio-based plastics drawing interest in research Bangkok 10330, Thailand and development. During its early developmental stage, PLA 4 Department of Microbiology and Biochemistry, Faculty of was synthesized from the optically pure L-lactic acid owing to Pharmaceutical Sciences, Chulalongkorn University, Phayathai the availability of L-lactic acid in the market. The first gener- Road, Wangmai, Pathumwan, Bangkok 10330, Thailand ation of PLA products has limited uses since their mechanical 764 Ann Microbiol (2017) 67:763–778 and thermal properties were not competitive with the available Material and methods commodity plastics. The mechanical and thermal properties of PLA can be improved by incorporating the stereoblock struc- Sample collection ture that yields an increasing melting temperature. Both opti- cally pure L-andD-lactic acid are required to synthesize the A total of 76 samples of soil and tree barks were collected stereoblock PLA. However, unlike L-lactic acid with its wide from December 2014 to August 2015 in the northern, eastern, range of applications, the production of D-lactic acid is cur- and central provinces of Thailand (3 samples from Chiangmai, rently limited (Li et al. 2013). 5fromPhitsanulok,11fromRayong,4fromSuphanburi,5 Lactic acid is currently produced via microbial fermenta- from Saraburi, 8 from Ayutthaya, 33 from Nakhonpathom, tion under mild conditions and low energy consumption. and 7 from Bangkok). The isolate SK5-6 was screened from Microbial fermentation utilizes low-cost, renewable feed- a soil sample collected at the Faculty of Pharmaceutical stocks to produce a racemic mixture or an optically pure iso- Sciences, Chulalongkorn University, Bangkok. mer of lactic acid (L- and D-lactic acid), depending on the microbes employed (Wee et al. 2006;Johnetal.2009;Xu Bacterial isolation et al. 2010). D-lactic acid is only produced in some bacteria, including Lactobacillus delbrueckii, Leuconostoc sp., and Natural samples (approx. 200–300 mg) were collected from Sporolactobacillus sp., as a metabolic response to environ- soil and tree barks. To selectively revive the growth of bacte- mental stress (Panesar et al., 2010; Tashiro et al. 2011; ria, we transferred a portion of the collected sample (0.25 Mimitsuka et al. 2012;Lietal.2013;Zhaoetal.2014;Bai g)into 5 mL sterile enrichment medium [composition per liter: et al. 2016). It has been noted that D-lactate production with 10 g glucose, 5 g yeast extract, 5 g peptone, 0.25 g KH2PO4, the aforementioned strains usually results in poor fermentation and 0.25 g K2HPO4, supplemented with 10 mL salt solution performance, which has led to genetic engineering being (composition per 10 mL: 400 mg MgSO4⋅7H2O, 20 mg employed for strain development to specifically produce op- MnSO4⋅5H2O, 20 mg FeSO4⋅7H2O,and20mgNaClpH tically pure D-lactate; however, the genetically modified mi- 4.4), pH 4.4], followed by heat shock at 80 °C for 10 min. crobes have limited industrial applications due to biosafety The suspension was subsequently incubated at 37 °C until regulation limits and genetic instability (Feng et al. 2014; growth was observed and then transferred onto a glucose– Tsuge et al. 2014;Baeketal.2016). Therefore, wild strains yeast extract–peptone (GYP) agar (composition per liter: are still preferred in industrial D-lactate production facilities. 10 g glucose, 5 g yeast extract, 5 g peptone, 0.25 g

Previous screening studies in Thailand have isolated novel KH2PO4,0.25gK2HPO4, 20 g agar, 5 g CaCO3,10mLsalt D-lactate-producing strains from natural samples. solution) plate containing CaCO3 as the acid production indi- Thamacharoensuk et al. (2015)recentlyisolatedtwoD-lactate cator. The pH of the GYP agar medium was adjusted to 6.80. producers from tree barks which they subsequently identified The GYP agar plate was incubated at 37 °C until colonies as the novel species Sporolactobacillus spathodeae and developed. The colonies that formed in a large clear zone were S. shoreae. Prasirtsak et al. (2016) isolated the novel bacterial transferred onto a new GYP agar plate for single colony puri- genus Terrilactibacillus from soil samples collected in fication. A single colony was transferred into a new GYP agar Thailand, which they later identified as Terrilactibacillus slant. To maintain the activity of the selected colony, the col- T laevilacticus NK26-11 , a catalase-positive D-lactate producer. ony was subcultured into a new GYP agar slant every 10 days. This species has attracted researchers’ attention due to its good The single colony was ultimately transferred into a sterile fermentation performance (high yield, high productivity, and skim milk solution and kept at −80 °C for long-term storage. high optical purity). Based on these findings, Thailand would appear to be a good source of natural material for bacterial Phenotypic characterization screening. The aim of the study reported here was to obtain robust D-lactate producers through a process of bacterial The bacterial isolates were cultured in a phenotypic medium screening of natural samples and subsequent isolation, and (composition per liter: 10 g glucose, 10 g yeast extract, 10 g identification of D-lactate producers. Among the D-lactate pro- peptone, 10 g CH3COONa, 5 mL salt solution) at 37 °C under ducers screened in this study, we identified one as anaerobic conditions for phenotypic characterization. Cell T. laevilacticus SK5-6, a novel genus, which exhibited a good form and cell size were examined under the microscope. The fermentation performance in the primary screening. Little is Hucker–Conn modification method was used for Gram stain- known about D-lactate production by this genus. ing (Tanasupawat et al. 1992. 1998). Spore formation was Consequently, we report here the preliminary fermentation examined by scanning electron microscopy. The effects of optimization for D-lactate production by T. laevilacticus temperature, initial pH, and NaCl concentration on growth SK5-6 and propose that this novel genus is a promising D- were observed in the culture grown in the phenotypic medi- lactate-producing strain for use in industrial fermentation. um. Catalase and oxidase activities, nitrate reduction, and Ann Microbiol (2017) 67:763–778 765 hydrolysis of arginine and starch were tested following the silicone stopper and placed into a W-zip pouch containing methods previously described in Tanasupawat et al. (1998). AnaeroPack-Anaero (Mitsubishi Gas Chemical)]. At the end Additional biochemical characteristics of the isolates were of the fermentation, the sample was collected for analyses of determined after incubation for 2 days at 37 °C under aerobic the remaining glucose, D-lactate, byproducts, and optical pu- conditions using API 50 CH strips (bioMérieux, Marcy- rity of D-lactate. l'Étoile, France) according to the manufacturer’sinstructions. For the catalase-positive isolate, the bacterial suspension was inoculated in preculture medium (pH 6.8) containing

Genotypic characterization (per liter) 10 g glucose, 15 g yeast extract, 4 g NH4Cl, 0.5 g KH2PO4,0.5gK2HPO4, 20 mL salt solution, and 5 g CaCO3 DNAwas extracted from the cells grown in the GYP broth for and incubated at 37 °C, 200 rpm for 5 h in a flask plugged with 2 days and purified using the method described by Saito and a C-type silicone stopper. Thereafter, 25 mL of the preculture Miura (1963). The 16S rRNA gene of the isolate was ampli- broth was transferred into 25 mL of the sterile fermentation fied and the PCR product purified and sequenced as previous- medium containing (per liter) 240 g glucose and 80 g CaCO3. ly described (Saito and Miura 1963; Tanasupawat et al. 2004). The culture was then incubated at 37 °C for another 72 h in a The sequence of the isolate was aligned with the selected flask plugged with a T-type silicone stopper and placed into a sequences obtained from GenBank using CLUSTAL_X ver- W-zip pouch containing AnaeroPack-Anaero (Mitsubishi Gas sion 1.81. The alignment was manually edited to remove gaps Chemical). At the end of the fermentation, the sample was and ambiguous nucleotides before the construction of a phy- collected for analyses of the remaining glucose, D-lactate, logenetic tree with the neighbor-joining method and the byproducts, and optical purity of D-lactate. maximum-likelihood method using MEGA version 6 (Felsenstein 1981;SaitouandNei1987;Tamuraetal. Effects of inoculum size, oxygen, and mixing on growth 2013). The confidence values of the individual branches in of T. laevilacticus SK5-6 during the preculture stage the phylogenetic tree were determined using the bootstrap analysis of Felsenstein (1985) based on 1000 replications. The operating conditions during both the preculture and fer- The values for the sequence similarity among the closest mentation stages of the catalase-positive isolate strains were determined using the EzTaxon server (Kim T. laevilacticus SK5-6 were further optimized in flask cultiva- et al. 2012). tion. The growth profile during the preculture stage was de- termined. T. laevilacticus SK5-6 was subcultured onto fresh Primary screening for D-lactate-producing isolates new GYP agar slants and incubated at 37 °C for 24 h to prepare the bacterial suspension. The preculture medium The bacterial isolates were categorized into two groups based was inoculated with the bacterial suspension [optical density on the presence of catalase activity in the primary screening at 600 nm (OD600) of 30-40, where OD600 of 1 is equivalent to for D-lactate production. Each isolate was subcultured onto a 0.12 g/L cell dry weight] of T. laevilacticus SK5-6 at different fresh new GYP agar slant and incubated anaerobically at 37 inoculum sizes (0.5, 1, and 2%). The culture was incubated in °C for 48 h. To prepare the bacterial suspension, we first an Erlenmeyer flask at 37 °C under different mixing (no transferred 2 mL of sterile preculture medium into the culture mixing and mixing at 200 rpm) and gas permeability (to gen- slant and then thoroughly mixed the contents of the slant. The erate an aerobic/anaerobic culture environment) conditions. bacterial suspension was then inoculated in a preculture flask To manipulate the air permeability of the culture flask in order containing 48 mL sterile preculture medium. to initiate an aerobic condition, we plugged the flask with a C- For the catalase-negative isolate, we inoculated 2 mL of the type silicone stopper to allow good air permeability. To gen- bacterial suspension into the preculture medium (pH 6.8) con- erate anaerobic conditions, we plugged the flask with a T-type taining (per liter) 10 g glucose, 5 g yeast extract, 5 g peptone, silicone stopper and placed it into a W-zip pouch containing

0.25 g KH2PO4,0.25gK2HPO4,10mLsaltsolution,and5g AnaeroPack-Anaero (Mitsubishi Gas Chemical). Samples CaCO3. The culture was incubated at 37 °C under anaerobic were taken at 1-h intervals for 8 h for the analyses of cell mass, conditions for 26 h [in a flask plugged with a T-type silicone the remaining glucose, and product formation. stopper and placed into a W-zip pouch containing AnaeroPack-Anaero (Mitsubishi Gas Chemical, Tokyo, Effect of the preculture seed on D-lactate fermentation Japan)], following which 1 mL of the preculture broth was transferred into 49 mL of sterile fermentation medium (com- Terrilactibacillus laevilacticus SK5-6 was subcultured onto position per liter: 120 g glucose, 10 g yeast extract, 5 g pep- GYP agar slants and incubated at 37 °C. The bacterial suspen- tone, 0.25 g KH2PO4,0.25gK2HPO4, 10 mL salt solution, sion of OD600 of 30-40 was prepared for inoculation into the 80 g CaCO3). The culture was incubated at 37 °C under an- preculture medium at 1% inoculum size. The preculture flask aerobic conditions for 72 h [in a flask plugged with a T-type was incubated at 37 °C under different mixing (no mixing and 766 Ann Microbiol (2017) 67:763–778 mixing at 200 rpm) and gas permeability [using a C-type high-performance liquid chromatography (HPLC). silicone stopper or a T-type silicone stopper and placing the Fermentation samples were centrifuged, filtered through a flask in a W-zip pouch containing AnaeroPack-Anaero PTFE (hydrophilic) membrane, and diluted with double de- (Mitsubishi Gas Chemical)] conditions. The preculture time ionized water. For analyses of the remaining glucose, total was varied between the mid log phase (4 h) and the late log lactic acid (both L-andD-lactic acid), and acetic acid, 15 μL phase (5 h). Following preculture,, the preculture broth was diluted particle-free samples were automatically injected (SIL- transferred into the fermentation medium at 50% inoculum 10A HPLC autosampler (Shimadzu Corp., Kyoto, Japan) into size. The fermentation culture was incubated at the same tem- an organic acid analysis column (Aminex HPX-87H ion ex- perature for 48 h under varied mixing (no mixing and mixing clusion organic acid column; 300 mm × 7.8 mm; Bio-Rad, at 150 rpm) and gas permeation [with either a C-type silicone Hercules, CA), and maintained at 45 °C in a column oven stopper or a T-type silicone stopper and placing the flask in a (CTO-10A; Shimadzu Corp.). H2SO4 (0.005 M) was pumped W-zip pouch containing AnaeroPack-Anaero (Mitsubish Gas through the system at a flow rate of 0.6 mL/min (LC-10Avp Chemical) conditions. Samples were taken every 12 h for System Controller; Shimadzu Corp.). A refractive index de- analyses of OD600 readings, the remaining glucose, and lactate tector (RID-10A; Shimadzu Corp.) was used to detect the and byproduct formation. organic compounds. The standards containing 0–2 g/L of each component (glucose, lactate, and acetate) were injected as D-lactate fermentation by T. laevilacticus SK5-6 in a 5-L references to determine sample concentration. The chromato- stirred fermenter gram peak area was used as the comparison basis in determin- ing the concentration. To determine the optical purity of lactic The fermentation platform of T. laevilacticus SK5-6 was pre- acid, 5 μL diluted particle-free samples were automatically liminarily determined in a 5-L stirred fermenter. A 24-h GYP injected into a chiral column (Sumipack, Sumichiral agar slant was used to prepare the bacterial suspension. The OA5000) and maintained at 40 °C. CuSO4 (0.001 M) was bacterial suspension (1% inoculum size) was inoculated in a used as the eluent at a flow rate of 1.0 mL/min. A UV detector preculture flask plugged with a C-type silicone stopper. The was used to detect the lactate isomers at 254 nm. The stan- preculture flask was incubated at 37 °C, 200 rpm for 4 h. After dards containing 0–2g/LofD-andL-lactic acid were injected that, the preculture flask was transferred into a 5-L stirred as references to determine sample concentration. fermenter that contained 2.5 L of the sterile preculture medi- Product yield (Yp/s) was determined from the ratio of the um at 10% inoculum size. The fermenter was operated at 37 product formed to carbon substrate consumed during the fer- °C, agitated at 300 rpm, with 1.0 vvm air. After 3 h, 0.5 L of mentation stage. Volumetric productivity was defined as the the fermentation medium containing (per liter) 720 g glucose total amount of product formed per unit volume per time. The and480gCaCO3 was added into the fermenter. Aeration was optical purity of the D-lactate was defined from the peak areas stopped and the agitation speed was varied at 200 and 300 rpm of the chromatogram as follows (Zhao et al. 2014). to obtain the optimal D-lactate production rate. Samples were D−lactate−L−lactate taken every 6 h for 48 h for analyses of cell mass, remaining ¼ Â % Optical purity − þ − 100 glucose, and lactate and byproduct formation. D lactate L lactate

Analytical methods

A sample of fermentation broth was centrifuged at 10,000 g Results and discussion for 5 min to separate the cell-free supernatant from the cell mass. The supernatant was collected for further analyses of the Isolation, identification, characterization, and screening of remaining glucose, lactic acid, byproducts, and the optical D-lactic acid producers purity of D-lactate. The cell mass was acidified with 1 M

HCl to remove any insoluble CaCO3 remaining in the sample. A total of 116 bacterial isolates were obtained from 76 natural The acidified sample was centrifuged and the particles resus- samples, of which only six, including SK5-6, SP43-2, NK44-2, pended in deionized water for the OD reading. BRY67-1, BRY67-2, and BRY67-3, produced D-lactic acid. The

Spectrophotometry was used to determine the OD600 of the neighbor-joining phylogenetic tree based on the 16S rRNA gene cell mass present in the fermentation broth. Cell mass concen- sequences suggested that the isolate SK5-6 was 100% related to tration was then calculated from the correlation between the novel Terrilactibacillus laevilacticus NK26-11T (= T T T OD600 and cell dry weight (where an OD600 of 1 is equivalent LMG27803 = TISTR2241 = PCU335 ) previously reported to 0.12 g/L). by Prasirtsak et al. (2016). The isolate SP43-2 demonstrated a The concentration of products formed and glucose sub- high similarity in 16S rRNA sequence to Sporolactobacillus strate remaining during the fermentation were analyzed using nakayamae subsp. nakayamae DSM11696T (99.6%) (Yanagida Ann Microbiol (2017) 67:763–778 767 et al. 1997). The other isolates, including NK44-2, BRY67-1, mixing to allow a good gas diffusion and homogeneity in the BRY67-2, and BRY67-3, had similar percentages of 99.2– culture. Without mixing, a similar growth rate was obtained 99.7% to Sporolactobacillus laevolacticus DSM442T from the cultures incubated under anaerobic/aerobic condi- (Hatayama et al. 2006). (Table 1;Fig.1). tions (Fig. 2). With mixing at 200 rpm, the culture incubated The morphological and biochemical characterization re- under the aerobic conditions yielded a slightly higher growth vealed that all six isolates were Gram-positive, facultative rate, resulting in a higher final cell concentration when the anaerobic, spore-forming, rod-shaped bacteria. None were growth reached stationary phase (5 h). With sufficient mixing, able to hydrolyze starch and arginine. All isolates lacked ox- the lag phase was shortened and the growth rate was higher. idase and catalase with the exception of SK5-6 which exhib- Therefore, comparing the two parameters studied, mixing ited a positive result for catalase activity. As a result, oxygen would appear to have had a stronger effect on the growth of was not restricted in the cultivation of SK5-6 while the others SK5-6. In addition, a higher growth rate and maximum cell required an anaerobic environment for growth. All six isolates dry weight were achieved from the culture grown under the produced D-lactate from glucose via a homofermentative aerobic conditions. In the presence of oxygen, SK5-6 gener- route. They grew at 20–40 °C, at a pH of 5.5–8.5. They tol- ated more ATP via glycolysis, citric acid cycle (TCA), and the erated a salt concentration of up to 3%. These isolates utilized electron transport chain (ETC). which eventually resulted in a various substrates in addition to glucose (Table 1). higher growth rate, although the difference between the Based on the results of the primary screening fermentation, all growth rate under anaerobic and aerobic conditions, respec- six isolates exhibited a high D-lactate production compared with tively, was not dramatically different. the type strain T. laevilacticus NK26-11T previously identified by The effects of inoculum size were also investigated (Fig. 3). our research group (Table 2). All six isolates also produced D- It was found that a small inoculum size (0.5%) resulted in a lactate from glucose at high yields of approximately 0.83–1.00 g/ longer lag phase (Fig. 3a) and that, conversely, increasing the g glucose. Productivity in flaskculturewasinanacceptable inoculum size to 1 and 2%, respectively, shortened the lag range of 1.06–1.42 g/L⋅h, with a sufficiently high final titer of phase. However, with respect to the log phase, the growth rate 76.01–102.43 g/L from an initial glucose concentration of 120 during the log phase of SK5-6 was similar for all three inoc- g/L. While isolates SK5-6, SP43-2, and NK44-2 yielded D-lac- ulum sizes studied. Changing the culture from an aerobic to an tate of a high optical purity, the D-lactate produced by isolates anaerobic environment did not affect cell growth to any great BRY67-1, BRY67-2, and BRY67-3 had an optical purity of < degree regardless of inoculum size. Glucose consumption 99.0% D-enantiomer equivalent (ee), thereby necessitating a (Fig. 3b) was consistent with the growth profile. The forma- more complicated recovery and purification process to acquire tion of lactate was observed during the preculture stage (Fig. a sufficiently higher purity than 99.0% for the polymer-grade 3c), indicating that the expression of lactate dehydrogenase lactic acid. As a result, fermentation optimization by these latter for lactate production when the metabolic flux shifted toward three isolates was not continued. Comparison of isolates SK5-6, the anaerobic route. As shown in Fig. 3a, c, lactate started to SP43-2, and NK44-2 showed that SK5-6 exhibited strong cata- accumulate when growth entered the log phase. At a small lase activity while the other two isolates were catalase negative. inoculum size (0.5%), lactate formation appeared when Therefore, it was possible to implement a two-phase fermenta- growth was in mid log phase; at a larger inoculum size (1 tion process (aerobic preculture and subsequent anaerobic fer- and 2%), lactate started to form from the early log phase. At mentation) using SK5-6, with high cell densities of SK5-6 ob- the 2% inoculum size, a high concentration of lactate (8.89– tained during the aerobic preculture stage, which enabled the 10.60 g/L) accumulated in the preculture broth; this acted as a production of D-lactate during the anaerobic fermentation stage stressor on the cells and later had diverse effects on lactate at a high production rate (Qin et al. 2010;Maetal.2014). formation during the fermentation stage. At the 1% inoculum Terrilactibacillus laevilacticus is a novel bacterial genus capable size used to inoculate the preculture flask, a shortened lag of producing D-lactate, as previously identified by Prasirtsak et al. phase was observed with rapid growth, high cell mass con- (2013) and Prasirtsak et al. (2016), but to date optimization of the centration, and slight lactate formation. This indicates that fermentation process by this bacterial genus has not been report- inoculation of the new medium with the proper density of cells ed. Therefore, we attempted to determine the key process param- induced high biosynthetic rates within a short time frame eters that modify D-lactate fermentation by this novel genus. (Egervarn et al. 2007). Based on these results, we considered the 1% inoculum size to be suitable for use in further optimi- Determination of the correct physiological stage zation studies. of preculture seed for D-lactate fermentation Not only did a sufficiently high concentration of the inoc- by T. laevilacticus SK5-6 ulum enhance the metabolic rate, the proper physiological stage also played a role in modifying the performance of We next investigated the metabolic response of SK5-6 under SK5-6 during fermentation. Zagari et al. (2013)reportedthat both aerobic and anaerobic conditions, as well as the effect of the inoculum promptly metabolized glucose to D-lactate at a 768

Table 1 Selected characteristics of six D-lactate-producing isolates and their closely related type strains.

Characteristic Isolate SK5-6 Terrilactibacillus Isolate SP43-2 Sporolactobacillus Isolate NK44-2 Isolate BRY67-1 Isolate BRY67-2 Isolate BRY67-3 Sporolactobacillus laevilacticus NK26-11T nakayamae subsp. laevolacticus nakyamae DSM1196T DSM442T

Cell shape Rod Rod Rod Rod Rod Rod Rod Rod Rod Spore formation + + + + + + + + + Temperature range (°C) 25–40 20–45 20–40 20–40 20–40 20–40 20–40 20-40 20–40 NaCl range (%) 0–30–30–30–30–30–30–30–30–2 pH range 5.5–8.5 5.0–8.5 4.0–9.0 4.5–8.5 4.0–8.5 4.0–9.0 4.0–9.0 4.0–9.0 4.0–9.0 Catalase activity + + ------Lactate isomer produced DD DD DDDDD Acid production from: Glycerol + + ------Erythritol + + ------

D-Galactose ++ -- ++++W D-Glucose ++ ++ +++++ D-Fructose ++ ++ +++++ D-Lyxose ++ ------D-Mannose ++ ++ +++++ D-Saccharose ++ ++ +++++ D-Turanose ++ -w ----- D-Ribose ------D-Xylose ------n irbo 21)67:763 (2017) Microbiol Ann Origin Soil - Bark - Bark Bark Bark Bark -

+, Positive; -, negative; W, weakly positive – 778 Ann Microbiol (2017) 67:763–778 769

Fig. 1 Neighbor-joining tree, 77 Sporolactobacillus laevolacticus BRY67-1 (LC222556) based on 16S rRNA gene Sporolactobacillus laevolacticus BRY67-2 (LC222557) Sporolactobacillus laevolacticus NK44-2 (LC222559) sequences, showing the 100 phylogenetic relationship of the Sporolactobacillus laevolacticus BRY67-3 (LC222558) 56 isolated strains SK5-6, SP43-2, Sporolactobacillus laevolacticus IAM 12321T (D16270) NK44-2, BRY67-1, BRY67-2, 98 Sporolactobacillus kofuensis M-19T (AB374517) BRY67-3, and related species. Sporolactobacillus nakayamae subsp. racemicus DSM 16324T (AJ698860) Bar represents sequence 59 100 Sporolactobacillus nakayamae subsp. nakayamae SP43-2 (LC222560) divergence (1%) (color figure Sporolactobacillus nakayamae subsp. nakayamae DSM 11696T (AJ634663) online) NRIC 1133T (AB362770) 100 Sporolactobacillus inulinus 100 Sporolactobacillus terrae DSM 11697T (AJ634662) Sporolactobacillus spathodeae BK117-1T (AB917309) QC81-06T (AB374522) 98 Sporolactobacillus putidus BK92T (AB917308) 96 Sporolactobacillus shoreae T 51 Sporolactobacillus vineae SL153 (EF581819) 99 Sporolactobacillus pectinivorans GD201205T (KP340805) 100 Pullulanibacillus naganoensis ATCC 53909T (AB021193) 64 Pullulanibacillus uraniitolerans UG-2T (AM931441) Tuberibacillus calidus 607T (AB231786) 100 Terrilactibacillus laevilacticus NK26-11T (AB841307) Terrilactibacillus laevilacticus SK5-6 (LC222555) 64 Bacillus solimangrovi GH2-4T (KC616733) KMM 3737T (AY228462) 97 Bacillus algicola CVS-14T (AM292417) 74 Salirhabdus euzebyi Bacillus shackletonii LMG 18435T (AJ250318) 99 Bacillus subtilis DSM 10T (AJ276351)

0.01 high rate during the log phase. In our study, we observed the the high growth rate could have compensated for with a high effect of inoculum size on D-lactate production in the culture cell concentration. This finding suggested that the seed train during the mid log phase (4 h) and during the late log phase development for SK5-6 was rather flexible and effective as it when the cell concentration was higher (5 h). The results could occur at times between the mid log and the late log showed that there was no difference in D-lactate production phases. performance from both inoculum seeds studied. This result can be explained by fact that the specific growth rate of Investigation of the metabolic behavior of T. laevilacticus SK5-6 at the mid log phase was higher than that at the late SK5-6 during the fermentation stage log phase when cells were approaching the stationary phase, although the cell concentration at the mid log phase was about The preculture seed obtained during the preculture stage under half that at the late log phase (Figs. 2, 3a). More specifically, different culture conditions was used as inoculum for the

Table 2 D-lactic acid production from six isolates selected in the primary screening compared with the type strain Terrilactibacillus laevilacticus NK26-11

Isolate D-lactate Remaining glucose (g/L)

Final concentration (g/L) Optical purity (%ee) Yield (g/g) Productivity (g/L⋅h)

NK26-11T 60.52 ± 0.44 96.44 ± 0.68 0.50 ± 0.01 0.84 ± 0.02 0.00 ± 0.00 SK5-6 99.27 ± 0.71 99.00 ± 0.70 0.90 ± 0.00 1.38 ± 0.00 9.74 ± 0.14 SP43-2 102.43 ± 2.71 100.00 ± 0.00 0.85 ± 0.01 1.42 ± 0.02 0.00 ± 0.00 NK44-2 90.84 ± 2.57 100.00 ± 0.00 0.83 ± 0.01 1.26 ± 0.01 10.93 ± 0.08 BRY67-1 101.12 ± 0.72 98.76 ± 0.70 1.00 ± 0.01 1.40 ± 0.01 18.75 ± 0.13 BRY67-2 83.68 ± 0.59 98.30 ± 0.70 0.86 ± 0.01 1.16 ± 0.01 22.84 ± 0.16 BRY67-3 76.01 ± 1.07 98. 56 ± 1.39 0.85 ± 0.01 1.06 ± 0.01 30.41 ± 0.43

The fermentation was conducted in a shake flask culture with an initial glucose concentration of 120 g/L Values in table are presented as the mean ± standard deviation (SD) 770 Ann Microbiol (2017) 67:763–778

Fig. 2 Effect of mixing [no 1 mixing (no) vs. mixing at 200 rpm no:anaerobic no:aerobic 200:anaerobic 200:aerobic 0.9 (200)] and oxygen (anaerobic vs. aerobic conditions) on growth of 0.8 isolate SK5-6 during the preculture stage in the flask 0.7 culture (color figure online) 0.6

0.5 lmass(g/L) l 0.4 Ce 0.3

0.2

0.1

0 012345678 Time (h)

fermentation medium for D-lactate production. The metabolic fermentation medium contained only high concentrations of response of SK5-6 under different cultivation conditions and glucose and CaCO3. The fermentation operated at 150 rpm its production performance during the fermentation stage are under aerobic conditions, then the cells entered their lag phase shown in Table 3. It is clear that mixing and anaerobic condi- as they required the expression of catalase activity to break tions were necessary for D-lactate production, as indicated by down hydrogen peroxide (H2O2). As a result, the growth rate the higher final lactate titer, yield, and productivity and lower was low when the metabolic pathway was shifted from the glucose levels remaining in the fermentation broth. No signif- anaerobic preculture seed to the aerobic fermentation with icant difference (P ≤ 0.05) in D-lactic acid production was sufficient mixing. observed when the fermentation was carried out at 150 rpm With respect to the effect of preculture seed being trans- under anaerobic conditions, although varied preculture seeds ferred to the fermentation stage, changing the inoculum age were used as inoculum for the fermentation stage (Table 3,see from the mid log to late log phases did not cause a dramatic footnotes a and b). The high yield indicates that SK5-6 suc- change in the fermentation performance during the fermenta- cessfully utilized glucose for the production of lactate during tion stage conducted under the same conditions (Table 3,see the anaerobic respiration process when oxygen was limited, footnote a). The aerobic and anaerobic preculture seeds also unlike the process previously reported during the preculture demonstrated a similar performance (Table 3, see footnote a). stage where mixing played a significant role in promoting More interestingly, no byproduct was found during the fer- growth while aerobic/anaerobic conditions appeared to have mentation stage, indicating that SK5-6 converted glucose to less effect on growth. During the fermentation stage, anaero- D-lactate through homofermentation. This preliminary finding bic conditions had a positive effect on lactate production irre- indicates that high lactate production can be achieved when gardless of mixing (Fig. 2; Table 3). the fermentation is performed at 150 rpm under anaerobic When the fermentation medium was inoculated with the conditions. The preculture seed of SK5-6 transferred to the aerobic preculture seed (both at the mid log and late log fermentation stage can be prepared under both aerobic and phases) and the fermentation stage proceeded under aerobic anaerobic conditions. Therefore, we can conclude that the conditions with mixing (Table 3, see footnote c), cells grew fermentation process using SK5-6 is relatively simple com- rapidly and approached a stationary phase within 24 h pared with conventional lactate production which usually (Table 3). These cells subsequently entered the declining leads to problems in terms of byproduct formation, low yield, phase, resulting in lower cell dry weight compared with that and low productivity. obtained under the other fermentation conditions (data not shown). As a result, low lactate production was obtained. Effect of agitation on D-lactate fermentation Transferring the anaerobic preculture seed (both at the mid by T. laevilacticus SK5-6 in the 5-L stirred fermenter log and late log phases) to the fermentation stage with mixing and aeration (Table 3, see footnote d) caused adverse effects The time course kinetics of SK5-6 during the fermentation on cell growth which resulted in low lactate production stage are shown in Fig. 4. We found that increasing the agita- (Table 3). Cell mass slowly increased (data not shown). One tion speed from 200 to 300 rpm significantly enhanced lactate explanation is that anaerobic cells being transferred into the production. With adequate mixing, glucose was rapidly Ann Microbiol (2017) 67:763–778 771

Fig. 3 Effect of inoculum size a (0.5,1,2%)onthegrowth, 1.2 glucose consumption, and lactate aerobic, 0.5 % anaerobic, 0.5 % formation of isolate SK5-6 during the preculture stage in flask 1 aerobic, 1 % anaerobic, 1 % culture under aerobic and anaerobic conditions. The aerobic, 2 % anaerobic, 2 % cultures were incubated at 37 °C, 0.8 200 rpm under either aerobic or anaerobic conditions. a Cell mass, b glucose concentration, c lactate 0.6 concentration (color figure online) Cell mass (g/L) 0.4

0.2

0 012345678 Time (h) b 12 aerobic, 0.5 % anaerobic, 0.5 %

10 aerobic, 1 % anaerobic, 1 % aerobic, 2 % anaerobic, 2 % 8

6 se (g/L) o uc l G 4

2

0 012345678 Time (h) c 12 aerobic, 0.5 % anaerobic, 0.5 %

10 aerobic, 1 % anaerobic, 1 %

aerobic, 2 % anaerobic, 2 % 8

6 tate (g/L) c La 4

2

0 012345678 Time (h) consumed via the anaerobic fermentative pathway, resulting in the fermentation culture at 300 rpm (2.066 g/L) was almost increasing cell mass and increasing levels of lactate as the fourfold higher than that in the culture at 200 rpm (0.564 g/L). growth-associated product. The final cell mass obtained from This subsequently resulted in high lactate production in the 772 Ann Microbiol (2017) 67:763–778

Table 3 D-lactate production during the fermentation stage in relation to preculture seed and cultivation conditions during the fermentation stage

Preculture seedsa Fermentation At 48 h Fermentation performance conditionsb Cell mass (g/L) Glucose (g/L) Lactate (g/L) Yield (g/g) Productivity Optical purity (g/L⋅h) (%ee)

Mid log (4 h) and No mixing, aerobic 1.445 ± 0.012 32.97 ± 3.31 67.54 ± 0.94 0.75 ± 0.04 1.41 ± 0.02 98.76 ± 1.40 aerobic conditions No mixing, anaerobic 1.798 ± 0.001 16.63 ± 0.92 85.46 ± 0.89 0.79 ± 0.03 1.78 ± 0.02 98.87 ± 0.35 150 rpm, aerobic 0.450 ± 0.004c 17.05 ± 2.09 66.84 ± 2.22 0.62 ± 0.03 1.39 ± 0.05 98.22 ± 0.14 150 rpm, anaerobic 1.475 ± 0.011 2.74 ± 3.88 102.22 ± 4.48 0.84 ± 0.00 2.13 ± 0.09 99.64 ± 0.07 Mid log (4 h) and No mixing, aerobic 1.552± 0.001 14.57 ± 0.09 84.62 ± 3.26 0.78 ± 0.01 1.76 ± 0.07 98.68 ± 0.14 anaerobic conditions No mixing, anaerobic 1.546 ± 0.004 22.55 ± 3.56 81.05 ± 1.21 0.81 ± 0.05 1.69 ± 0.03 98.69 ± 0.28 150 rpm, aerobic 0.372 ± 0.004d 44.79 ± 3.74 35.43 ± 3.08 0.46 ± 0.06 0.74 ± 0.06 98.60 ± 0.35 150 rpm, anaerobic 1.289 ± 0.003 0.00 ± 0.00 101.76 ± 5.81 0.82 ± 0.06 2.12 ± 0.12 99.63 ± 0.07 Late log (5 h) and No mixing, aerobic 1.609 ± 0.001 22.63 ± 1.51 71.98 ± 1.75 0.71 ± 0.04 1.50 ± 0.04 99.32 ± 0.14 aerobic conditions No mixing, anaerobic 1.756 ± 0.001 37.87 ± 1.17 71.88 ± 1.42 0.84 ± 0.01 1.50 ± 0.03 99.37 ± 0.21 150 rpm, aerobic 0.663 ± 0.003c 94.60 ± 2.10 18.98 ± 2.22 0.69 ± 0.04 0.40 ± 0.05 98.85 ± 0.35 150 rpm, anaerobic 1.356 ± 0.005 6.88 ± 0.44 99.22 ± 1.01 0.86 ± 0.01 2.07 ± 0.02 99.69 ± 0.28 Late log (5 h) and No mixing, aerobic 1.463 ± 0.002 12.12 ± 0.64 69.90 ± 0.45 0.62 ± 0.01 1.46 ± 0.01 99.35 ± 0.14 anaerobic conditions No mixing, anaerobic 1.498 ± 0.006 19.65 ± 0.49 79.67 ± 0.99 0.71 ± 0.03 1.66 ± 0.01 99.26 ± 0.07 150 rpm, aerobic 0.759 ± 0.018d 92.43 ± 1.83 18.68 ± 0.33 0.61 ± 0.07 0.39 ± 0.01 99.80 ± 0.58 150 rpm, anaerobic 1.262 ± 0.002 6.61 ± 0.48 101.02 ± 0.91 0.85 ± 0.02 2.10 ± 0.02 99.67 ± 0.25

Values in table are presented as the mean ± SD a There was no significant difference at P ≤ 0.05 in D-lactic acid production when preculture conditions were varied (age and oxygen) during fermentation carried out with 150 rpm (mixing) under anaerobic conditions b Changing the fermentation conditions [mixing (no mixing/200 rpm) and oxygen (aerobic/anaerobic)] resulted in significant changes at P ≤ 0.05 in cell growth, glucose consumption, and D-lactic acid production c Cell mass during the aerobic fermentation at 150 rpm reached the maximum concentration at 24 h (data not shown), followed by a subsequent drop in cell mass, resulting in the low cell concentration d Cell mass slowly increased because of the long lag phase when shifting from anaerobic preculture to aerobic fermentation fermentation system at 300 rpm (final titer of 92.60 g/L, yield culture. It has been reported that the microaeration condition of 0.94, and productivity of 1.93 g/L⋅h). On the other hand, a generated by sufficient mixing enhanced the cell growth of low mixing rate (200 rpm) rather limited the bioconversion Lactococcus lactis MG1363 (Nordkvist et al. 2003; Neves process; therefore, a lower glucose consumption rate resulted, et al. 2005). It has also been reported that combining mixing with with a large amount of residual glucose (58.49 g/L, only half aeration resulted in better growth of L. lactis than no mixing was consumed during fermentation) and subsequently low (Ishizaki and Ueda 1995; Pedersen et al. 2008). In typical lactate lactate production (final titer of 46.05 g/L, with a yield of fermentation, a calcium base is used to control pH. Without 0.71 and productivity of 0.96 g/L⋅h). From the fermentation adequate mixing, heterogeneity is generated in the culture broth profiles shown in Fig. 4, it should be noted that the pH was due to CaCO3 and cell sedimentation, eventually affecting the successfully maintained during the fermentation by adding overall process kinetics (Qin et al. 2010;Maetal.2014).

CaCO3 at the beginning of the fermentation stage. The results A comparison of the preliminary fermentation optimization suggested that with sufficient mixing, the homogeneous sys- from the primary screening of the flask and fermenter culture tem generated the proper heat and mass transport which in revealed that under our preliminary optimized process condi- turn facilitated the bioconversion of glucose to D-lactate tions, the catalase-positive strain T. laevilacticus SK5-6 per- (Aguirre-Ezkauriatza et al. 2008; Ibrahim et al. 2010; formed promisingly in terms of D-lactate production. We there- Watanabe et al. 2012). fore considered this strain to be a novel industrial producer in Mixing is another key parameter controlling the fermentation addition to the typical catalase-negative Sporalactobacillus sp. performance as it modifies both broth homogeneity and transport (Table 4). We believe that further process optimization can pro- in the fermentation system. With sufficient mixing, glucose up- vide greater improvement in the production performance of SK5- take rate was rapid, resulting in a large pool of pyruvate being 6 and the subsequent effective fermentation platform for large- converted to D-lactate by lactate dehydrogenase (Hou et al. scale production operations. 2000). Mixing is typically employed for gas distribution in the Ann Microbiol (2017) 67:763–778 773

Fig. 4 Effect of mixing during a the fermentation stage of isolate 180 7 SK5-6 cultivated under anaerobic Glucose Lactate Cell pH conditions in the 5-L stirred 6 150 fermenter. a 200 rpm, b 300 rpm (color figure online) 5 /L)

120 H p e(g t

4 L), / ta

ac 90 L

, 3 mass (g se l

uco 60 Cel

Gl 2

30 1

0 0 0 6 12 18 24 30 36 42 48 b Time (h) 180 7 Glucose Lactate Cell pH 6 150

5 /L)

120 H p e(g t

4 L), / ta

ac 90 L

, 3 mass (g se l

uco 60 Cel

Gl 2

30 1

0 0 0 6 12 18 24 30 36 42 48 Time (h)

Terrilactibacillus sp. as a promising industrial D-lactate acidic pH (Sanders et al. 2003). Currently, Sporolactobacillus sp. producing isolate is known as the key producer of D-lactate in industrial fermenta- tion processes. Sporolactobacillus sp. can be found in soil, tree The origin of the major D-lactate producers previously reported in bark, and fermented food and feed. Previous studies have report- the literature as well as those isolated in this study is shown in ed the isolation of Sporolactobacillus inulinus from chicken feed Table 5.ItappearsthatD-lactate producers are distributed in and the isolation of S. nakayamae subsp. nakayamae, limited sources and that their preferred environment is a rather S. nakayamae subsp. racemicus, S. lactosus, S. kofuensis,and

Table 4 Preliminary optimization of fermentation conditions for D-lactate production by isolate SK5-6 resulted in improved fermentation performance.

Optimized conditions Time (h) Residual glucose (g/L) Final lactate (g/L) Optical purity (%ee) Yield (g/g) Productivity (g/L⋅h)

Flask culture Primary screening 72 9.74 ± 0.14 99.27 ± 0.71 99.00 ± 0.70 0.90 ± 0.00 1.38 ± 0.02 Flask culture Aerobic preculture, anaerobic 48 2.74 ± 3.88 102.22 ± 4.48 99.64 ± 0.07 0.84 ± 0.00 2.13 ± 0.09 fermentation Fermenter culture Aerobic preculture, anaerobic 48 19.20 ± 0.03 92.60 ± 0.65 99.56 ± 0.07 0.84 ± 0.00 1.93 ± 0.01 fermentation agitated at 300 rpm

Values in table are presented as the mean ± SD where appropriate 774 Ann Microbiol (2017) 67:763–778

Table 5 Origin of the major D-lactate producers reported in the literature and those isolated in this study

Sample origin Microorganism References

Chicken feed, soil, fermentation starter Sporolactobacillus inulinus Yanagida et al. 1997 S. nakayamae subsp. nakayamae S. nakayamae subsp. racemicus S. lactosus S. kofuensis S. terrae Vineyard soil S. vineae Chang et al. 2008 Spoiled orange juice S. putidas Fujita et al. 2010 Bark of Spathodea campanulata P. Beauv. S. spathodeae Thamacharoensuk et al. 2015 Bark of Shorea talura Roxb S. shoreae Thamacharoensuk et al. 2015 Bark of Lagerstroemia floribunda Jack ex Blume. S. terrae CU68-2 Prasirtsak et al. 2013 Bark of Samanea saman S. nakayamae subsp. nakayamae SP43-2 This study Bark of Musa sapientum Linn. S. laevolacticus NK44-2 This study Bark of Sterculia foetida L. S. laevolacticus BRY67-1 This study Bark of Sterculia foetida L. S. laevolacticus BRY67-2 This study Bark of Sterculia foetida L. S. laevolacticus BRY67-3 This study Soil Terrilactibacillus laevilacticus NK26-11T Prasirtsak et al. 2016 Soil T. laevilacticus SK5-6 This study

S. terrae from soil samples and fermentation starters (Yanagida enzymes catalase and superdismutase that remove these toxic et al. 1997). Sporolactobacillus vineae and S. putidas were found compounds, respectively, while strict anaerobes usually lack in vineyard soil in Korea and in spoiled orange juice, respectively these enzymes and are therefore restricted to an oxygen-free (Chang et al. 2008; Fujita et al. 2010). Sporolactobacillus sp. is a environment (Hassan and Fridovich 1977;Pedersenetal. Gram-positive, facultatively anaerobic, motile, and endospore- 2008). forming microbe. This genus lacks the activities of oxidase and Based on its biochemical characterization, our novel catalase. In addition, it cannot reduce nitrate. The members of Terrilactibacillus isolate exhibits catalase activity, thus this bacterial genus produce D-lactate from glucose and other confirming that the bacteria in this genus can survive in the carbon substrates depending on the specific characteristics of presence of oxygen—unlike species of the genus the type strains belonging to this genus (Yanagida et al. 1997; Sporolactobacillus. Since bacteria belonging to the genus Fujita et al. 2010). Cultivation of Sporolactobacillus is restricted Terrilactibacillus possess catalase activity, they generate more under anaerobic conditions when the oxygen is limited because ATP and grow rapidly in the presence of oxygen, which sub- of the lack of catalase activity and a cytochrome system sequently results in a high cell mass concentration and lactate (Holzapfel and Botha 1988; Watanabe et al. 2012), which results productivity (Watanabe et al. 2012; Prasirtsak et al. 2016). in low ATP generation and a subsequently low metabolic rate It should be noted that Terrilactibacillus sp. is also a Gram- compared to aerobic culture. positive and endospore-forming D-lactate producing microbe; Respiratory metabolism results in the conversion of glu- therefore, it can survive exposure to harsh conditions, such as cose into ATP via glycolysis, the TCA, and oxidative phos- lowpH(Sandersetal.2003). As seen in Table 1, phorylation. The ATP is generated in this metabolic process Terrilactibacillus sp. consumed a wide range of carbon sugars, because of the activity of the electron transport system. As the especially glycerol, for D-lactate production, while electrons are passed down through the ETC, each molecule in Sporolactobacillus sp. demonstrated a limited ability to con- the chain alternates between its reduced and oxidative forms. vert several sugars. The ability of Terrilactibacillus sp. to me- Cytochrome oxidase, which appears at the end of the chain, tabolize glycerol is a consequence of the catalase activity pos- catalyzes the oxidation of the last cytochrome molecule by sessed by these bacteria. Hassan and Fridovich (1977)claimed molecular oxygen. This final process results in the reduction that oxygen is involved in the consumption of polyhydroxy of the molecular oxygen to water. The incomplete reduction of alcohols by some lactate producers as in their study 1 mole of oxygen often leads to the production of H2O2 or the superox- oxygen was required to convert 1 mole of glycerol substrate - ide free radical (O2 ), both of which are toxic to the cell. into 1 mole of lactate and 1 mole of H2O2. Many aerobic or facultatively anaerobic cells produce the Ann Microbiol (2017) 67:763–778 775

Table 6 D-lactate production by the potential strains reported in the literature and in this study

Strain Operation pH control Lactate (g/L) Yield (g/g) Productivity Optical purity (%ee) References (g/L⋅h)

Sporolactobacillus sp. CASD Batch CaCO3 80–90 0.86–0.97 1.77 - Zhao et al. 2010

Sporolactobacillus sp. CASD Fed-batch CaCO3 207 0.93 3.80 99.3 Wang et al. 2011

S. laevolacticus JCM2513 Continuous Ca(OH)2 67.3 0.98 11.20 - Mimitsuka et al. 2012

Sporolactobacillus sp. Y2-8 Batch CaCO3 122 0.81 1.00 99.1 Zheng et al. 2012

S. laevolacticus DSM442 Fed-batch CaCO3 144.2 0.96 4.13 99.3 Li et al. 2013

Lactobacillus coryniformis Flask NH4OH 186.4 0.85 3.11 - Nguyen et al. 2013

Klebsiella pneumoniae Fed-batch NH4OH 142.1 0.82 2.96 100 Feng et al. 2014 ATCC25955 Escherichia coli HBUT-D Pilot scale Ca(OH)2 127 0.93 6.35 99.5 Liu et al. 2014

Sporolactobacillus inulinus Y2-8 Fed-batch NH4OH 218.8 - 1.65 >99.0 Zhao et al. 2014

Sporolactobacillus sp. Y2-8 Batch CaCO3 125.0 - 1.39 99 Sun et al. 2015

S. inulinus NBRC13595 Batch CaCO3 189.0 0.94 5.25 >98 Tadi et al. 2017

S. inulinus YBS1-5 Fed-batch CaCO3 107.2 0.85 1.19 99.2 Bai et al. 2016

Terrilactibacillus laevilacticus Flask CaCO3 102.2 0.84 2.13 99.6 This study SK5-6 T. laevilacticus SK5-6 Batch CaCO3 96.6 0.84 1.93 99.6 This study

Based on the taxonomical and biochemical characteristics this phenomenon in the two-phase fermentation process with mentioned here, we conclude that our novel Terrilactibacillus SK5-6 for D-lactate production, which strongly indicates the sp. shows a promising performance as an industrial D-lactate- robustness of SK5-6 as a suitable industrial-scale D-lactate producing strain. producer.

Establishing a D-lactate fermentation platform Comparing D-lactate production performance by Terrilactibacillus sp. by Terrilactibacillus sp. with other species

Understanding the metabolic behavior of the microbe is man- D-Lactate production has been reported in a number of bacte- datory in fermentation process optimization to achieve the ria, including Sporolactobacillus inulinus, S. laevolacticus, targeted production performance. In this study, we have iden- Lactobacillus delbrueckii,andLeuconostoc sp. Among these tified the key parameters that controlled the growth and syn- genera, Sporolactobacillus sp. has been claimed to be the thesis of metabolites during the cultivation of SK5-6 and suc- promising candidate for industrial fermentation. Tashiro cessfully introduced a two-phase fermentation process for cul- et al. (2011) reported that fermentation by wild-type strains, tivating SK5-6 under aerobic conditions and fermenting homo such as thermotolerant L. delbrueckii subsp. lactis QU41, D-lactate under anaerobic conditions. The major advantage of yielded 87.4 g/L D-lactate with a high optical purity of this two-phase fermentation procedure is that the high cell 99.9%ee using MRS medium containing 100 g/L glucose. density obtained during the preculture stage accelerates the Mimitsuka et al. (2012) conducted batch fermentation with anaerobic conversion of glucose to lactate during the fermen- S. laevolacticus JCM2513 and reported a final titer of D-lactate tation stage. However, previous studies have reported draw- of 55.7 g/L, with a yield of 0.89 g/g and productivity of 0.25 backs of this process when the process condition is switched g/L⋅h. from an aerobic to an anaerobic environment (Qin et al. 2010; Comparison of the production performance of D-lactate to Ma et al. 2014). It has been observed that byproducts often that obtained from the well-established L-lactate production appear as the metabolic pathway shifts from aerobic respira- process revealed that the final titer along with yield and pro- tion to the mixed acid fermentation, when a high glucose ductivity were still relatively low using D-lactate, thus making concentration is present, with the remaining oxygen available the former process economically uncompetitive. Li et al. at the initial stage of the fermentation. Expelling the remaining (2013) claimed that the poor efficiency of D-lactate fermenta- oxygen from the preculture stage by flushing the fermenter tion was the result of inhibition due to high substrate concen- with inert gas to obtain a strictly anaerobic condition is con- tration. In batch fermentation by S. laevolacticus DSM442 sidered to be cumbersome. Nevertheless, we did not observe using a high initial glucose concentration (>100 g/L) 776 Ann Microbiol (2017) 67:763–778 supplemented with cottonseed as the nitrogen source, these robust, high-cell-density cell mass of SK5-6 for D-lactate pro- authors found that lactic acid production was inhibited. duction during the anaerobic fermentation stage without Improved fermentation efficiency was obtained by a fed- byproducts. This resulted in high yield and high productivity. batch process, with an increasing final titer of 144.4 g/L, yield The results of this research therefore confirm the potential of of 0.96 g/g, productivity of 4.13 g/L⋅h, and optical purity of using Terrilactibacillus sp. as a D-lactate producer. 99.3%ee (Li et al. 2013). D-lactate production by S. inulinus Y2-8 in batch fermentation in a fibrous bed bioreactor using Acknowledgements Financial supports from the Grant for corn flour hydrolysate was reported by Zhao et al. (2014). In International Research Integration: Research Pyramid, Ratchadapiseksomphot Endowment Fund (GCURP_58_01_33_01) and that process, a high yield of 0.98 g/g, with a productivity of Thailand Research Fund via the Distinguished Research Professor Grant 1.62 g/L⋅h and optical purity of 99.0%ee was obtained. In (DPG5880003) are gratefully acknowledged. Research facility support another study, Bai et al. ( 2016)usedfed-batchD-lactate fer- through the Chulalongkorn Academic Advancement into its 2nd mentation by S. inulinus YBS1-5 with simultaneous utiliza- Century Project (CUAASC) is also appreciated. tion of cottonseed meal and corncob residue to produce a high Compliance with ethical standards D-lactate concentration (107.2 g/L), high productivity (1.19 g/L.h), and high yield (0.85 g/g). These authors also reported Conflict of interest The authors declare no conflict of interest. an optical purity of D-lactate of 99.2%ee. As seen in Table 6,wildstrains,suchasSporolactobacillus Ethical approval This article does not contain any studies with human sp. and Lactobacillus sp., have been found to exhibit slightly participants or animals performed by any of the authors. low productivity in batch and fed-batch processes (de Franca et al. 2009). Improved production performance was obtained during continuous culture incorporated with the separation References system (Koutinas et al. 2014). For example, by incorporating an electrodialysis membrane unit with the fermentation sys- Aguirre-Ezkauriatza EJ, Galarza-Gonzalez MG, Uribe-Bujanda AI, Rios- tem, an improved productivity of 8 g/L⋅h was obtained (Min- Licea M, Lopez-Pacheco F, Hernandez-Brenes CM, Alvarez MM (2008) Effect of mixing during fermentation in yoghurt manufactur- tian et al. 2005). Improved productivity has also been obtained ing. J Dairy Sci 91:4454–4465 in metabolically engineered strains of Escherichia coli and Baek SH, Kwon EY, Kim YH, Hahn JS (2016) Metabolic engineering Klebsiella pneumoniae (Feng et al. 2014; Tsuge et al. 2014; and adaptive evolution for efficient production of D-lactic acid in – Baek et al. 2016). Nevertheless, the engineered strains are Saccharomyces cerevisiae. Appl Microbiol Biotechnol 100:2737 2748 usually of limited use, especially in large-scale operations, Bai Z, Gao Z, Sun J, Wu B, He B (2016) D-lactic acid production by due to biosafety regulations, need for large investment in pro- Sporolactobacillus inulinus YBS1-5 with simultaneous utilization duction facilities, and genetic instability of the strains. As a of cottonseed meal and corncob residue. Bioresour Technol 207: result, wild strains are still preferable. Comparing the produc- 346–352 Chang YH, Jung MY, Park IS, Oh HM (2008) Sporolactobacillus vineae tion performance of SK5-6 to that of the other D-lactate pro- sp. nov., a spore-forming lactic acid bacterium isolated from vine- ducers, as reported in Table 6, we suggest that with further yard soil. Int J Syst Evol Microbiol 58:2316–2320 process optimization and fermentation platform development, Egervarn M, Lindmark H, Roos S, Huys G, Lindgren S (2007) Effects of SK5-6 can become a promising industrial strain for D-lactate inoculum size and incubation time on broth microdilution suscepti- fermentation. bility testing of lactic acid bacteria. Antimicrob Agent Chemother 51:394–396 Felsenstein J (1981) Evolutionary trees from DNA sequences: a maxi- mum likelihood approach. J Mol Evol 1:368–376 Conclusion Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791 Feng X, Ding Y, Xian M, Xu X, Zhang R, Zhao G (2014) Production of In this study, D-lactate producers were isolated from natural optically pure D-lactate from glycerol by engineered Klebsiella samples collected in Thailand. Among the six isolates which pneumoniae strain. Bioresour Technol 172:269–275 produced D-lactate, T. laevilacticus SK5-6 showed the best de Franca FP, de Jesus AM, Oliveira FJS (2009) Enhancement of lactic characteristics in comparison to other D-lactate producers pre- acid fermentation by Lactobacillus delbrueckii ATCC6949 using – viously reported in the literature. Strain SK5-6 demonstrated sugarcane molasses. Can J Pure Appl Sci 3:773 778 Fujita R, Mochida K, Kata M, Goto K (2010) Sporolactobacillus putidus the ability to utilize several carbon substrates for D-lactate sp. nov., an endospore-forming lactic acid bacterium isolated from production. The optical purity of D-lactate was sufficiently spoiled orange juice. Int J Syst Evol Microbiol 60:1499–1503 high for polymer-grade specification (>99.5%ee). Unlike Hassan HM, Fridovich I (1977) Physiological function of superoxidase Sporolactobacillus sp., the most common D-lactate-producing dismutase in glucose-limited chemostat cultures of Escherichia coli. J Bacteriol 130:805–811 strain, SK5-6 produced catalase; thus, oxygen was not restrict- Hatayama K, Shoun H, Ueda Y, Nakamura A (2006) Tuberibacillus ed during the cultivation. Using a simple two-phase fermen- calidus gen. nov., sp. nov., isolated from a compost pile and reclas- tation process, we introduced an aerobic preculture to generate sification of Bacillus naganoensis Tomimura et al. 1990 as Ann Microbiol (2017) 67:763–778 777

Pullulanibacillus naganoensis gen. nov., comb. nov. and Bacillus Prasirtsak B, Tanasupawat S, Boonsombat R, Kodama K, Thongchul N laevolacticus Andersch et al. 1994 as Sporolactobacillus (2013) Characterization of lactic acid producing bacteria from Thai laevolacticus comb. nov. Int J Syst Evol Microbiol 56:2545–2551 sources. J Appl Pharm Sci 3:33–38 Holzapfel WH, Botha SJ (1988) Physiology of Sporolactobacillus strains Prasirtsak B, Thongchul N, Tolieng V, Tanasupawat S (2016) isolated from different habitats and the indication of in vitro antag- Terrilactibacillus laevilacticus gen. nov., sp. nov., isolated from soil. onism against Bacillus species. Int J Food Microbiol 7:161–168 Int J Syst Evol Microbiol 66:1311–1316 Hou R, Chen Z, Yi X, Bi-An J, Xu G (2000) Catalytic reaction mecha- Qin JY, Wang XW, Zheng ZJ, Ma CQ, Tang HZ, Xu P (2010) Production nism of L-lactate dehydrogenase: an ab initio study. Sci China Ser B of L-lactic acid by a thermophilic Bacillus mutant using sodium Chem 43:587–599 hydroxide as neutralizing agent. Bioresour Technol 101:7570–7576 Ibrahim SB, Rahman NAA, Mohamad R, Rahim RA (2010) Effects of Saito H, Miura KI (1963) Preparation of transforming deoxyribonucleic agitation speed, temperature, carbon and nitrogen sources on the acid by phenol treatment. Biochim Biophys Acta 72:619–629 growth of recombinant Lactococcus lactis NZ9000 carrying domain Saitou N, Nei M (1987) The neighbor-joining method: a new method for 1 of aerolysin gene. Afr J Biotechnol 9:5392–5398 reconstructing phylogenetic trees. Mol Biol Evol 4:406–425 Ishizaki A, Ueda T (1995) Growth kinetics and product inhibition of Sanders ME, Morelli L, Tompkins TA (2003) Sporeformers as human Lactococcus lactis IO-1 culture in xylose medium. J Ferment probiotics: Bacillus, Sporolactobacillus,andBrevibacillus. Compr Bioeng 80:287–290 Rev Food Sci Food Saf 2:101–110 John RP, Anisha GS, Nampoothiri KM, Pandey A (2009) Direct lactic Sun J, Wang Y, Wu B, Bai Z, He B (2015) Enhanced production of D- acid fermentation: focus on simultaneous saccharification and lactic lactic acid by Sporolactobacillus sp. Y2-8 mutant generated by at- acid production. Biotechnol Adv 27:145–152 mospheric and room temperature plasma. Biotechnol Appl Biochem Kim OS, Cho YJ, Lee K, Yoon SH, Kim M, Na H, Park SC, Jeon YS, Lee 62:287–292 H, Yi H, Won S, Chun J (2012) Introducing EzTaxon-e: a prokary- Tadi SRR, Arun EVR, Limaye AM, Sivaprakasam S (2017) Enhanced otic 16S rRNA gene sequence database with phylotypes that repre- production of optically pure D(-) lactic acid from nutritionally rich sent uncultured species. Int J Syst Evol Microbiol 62:716–721 Borassus flabellifer sugar and whey protein hydrolysate based- Koutinas AA, Vlysidis A, Pleissner D, Kopsahelis N, Garcia IL, Kookos fermentation medium. Biotechnol Appl Biochem 64:279–289 IK, Papanikolaou S, Kwan TH, Lin CSK (2014) Valorization of Tamura K, Stecher G, Peterson D, Filipski A, Kumar S (2013) MEGA6: industrial waste and byproduct streams via fermentation for the pro- molecular evolutionary genetics analysis version 6.0. Mol Biol Evol duction of hemicals and biopolymers. Chem Soc Rev 43:2587–2627 30:2725–2729 Li Y, Wang L, Ju J, Yu B, Ma Y (2013) Efficient production of polymer- Tanasupawat S, Ezaki T, Suzuki KI, Okada S, Komagata K, Kozaki M grade D-lactate by Sporolactobacillus laevolacticus DSM442 with (1992) Characterization and identification of Lactobacillus pentosus agricultural waste cottonseed as the sole nitrogen source. Bioresour and Lactobacillus plantarum strains from fermented foods in Technol 142:186–191 Thailand. J Gen Appl Microbiol 38:121–134 Liu Y, Gao W, Zhao X, Wang J, Garza E, Manow R, Zhou S (2014) Pilot Tanasupawat S, Okada S, Komagata K (1998) Lactic acid bacteria found scale demonstration of D-lactic acid fermentation facilitated by in fermented fish in Thailand. J Gen Appl Microbiol 4:193–200 Ca(OH)2 using a metabolically engineered Escherichia coli. Tanasupawat S, Thawai C, Yukphan P, Moonmangmee D, Itoh T, Adachi Bioresour Technol 169:559–565 O, Yamada Y (2004) Gluconobacter thailandicus sp. nov., an acetic Ma KD, Maeda T, You HY, Shirai Y (2014) Open fermentative produc- acid bacterium in the α-Proteobacteria. J Gen Appl Microbiol 50: tion of L-lactic acid with high optical purity by thermophilic Bacillus 159–167 coagulans using excess sludge as nutrient. Bioresour Technol 151: Tashiro Y,Kaneko W, Sun Y,Shibata K, Inokuma K, Zendo T, Sonomoto 28–35 K (2011) Continuous D-lactic acid production by a novel Mimitsuka T, Na K, Morita K, Sawai H, Minegishi S, Henmi M, Yamada thermotolerant Lactobacillus delbrueckii subsp. lactis QU 41. K, Shimizu S, Yonehara T (2012) A membrane-integrated fermen- Appl Microbiol Biotechnol 89:1741–1750 tation reactor system: its effects in reducing the amount of sub-raw Thamacharoensuk T, Kitahara M, Ohkuma M, Thongchul N, materials for D-lactic acid continuous fermentation by Tanasupawat S (2015) Sporolactobacillus shoreae sp. nov. and Sporolactobacillus laevolacticus. Biosci Biotechnol Biochem 76: Sporolactobacillus spathodeae sp. nov., two spore-forming lactic 67–72 acid bacteria isolated from tree barks in Thailand. Int J Syst Evol Min-tian G, Koide M, Gotou R, Takanashi H, Hirata M, Hano T (2005) Microbiol 65:1220–1226 Development of a continuous electrodialysis fermentation system Tsuge Y, Kawaguchi H, Sasaki K, Tanaka T, Kondo A (2014) Two-step for production of lactic acid by Lactobacillus rhamnosus.Process production of D-lactate from mixed sugars by growing and resting Biochem 40:1033–1036 cells of metabolically engineered Lactobacillus plantarum.Appl Neves AR, Pool WA, Kok J, Kuipers OP, Santos H (2005) Overview on Microbiol Biotechnol 98:4911–4918 sugar metabolism and its control in Lactococcus lactis—the input Wang L, Zhao B, Li F, Xu K, Ma C, Tao F, Li Q, Xu P (2011) Highly from in vivo NMR. FEMS Microbiol Rev 29:531–554 efficient production of D-lactate by Sporolactobacillus sp. CASD Nguyen CM, Choi GJ, Choi YH, Jang KS, Kim JC (2013) D-andL-lactic with simultaneous enzymatic hydrolysis of peanut meal. Appl acid production from fresh sweet potato through simultaneous sac- Microbiol Biotechnol 89:1009–1017 charification and fermentation. Biochem Eng J 81:40–46 Watanabe M, van der Veen S, Nakajima H, Abee T (2012) Effect of Nordkvist M, Jensen NBS, Villadsen J (2003) Glucose metabolism in respiration and manganese on oxidative stress resistance of LACTOCOCCUS LACTIS MG1363 under different aeration conditions: Lactobacillus plantarum WCFS1. Microbiology 158:293–300 requirement of acetate to sustain growth under microaerobic condi- Wee YJ, Kim JN, Ryu HW (2006) Biotechnological production of lactic tions. Appl Environ Microbiol 69:3462–3468 acid and its recent applications. Food Technol Biotech 44:163–172 Panesar PS, Kennedy JF, Knill CJ, Kosseva M (2010) Production of L(+) Xu TT, Bai ZZ, Wang LJ, He BF (2010) Breeding of D(-)-lactic acid high lactic acid using Lactobacillus casei from whey. Braz Arch Biol producing strain by low-energy ion implantation and preliminary Technol 53:219–226 analysis of related metabolism. Appl Biochem Biotechnol 160: Pedersen MB, Garrigues C, Tuphile K, Brun C, Vido K, Bennedsen M, 314–321 Mollgaard H, Gaudu P, Gruss A (2008) Impact of aeration and heme- Yanagida F, Suzuki KI, Kozaki M, Komagata K (1997) Proposal of activated respiration on Lactococcus lactis gene expression: identifica- Sporolactobacillus nakayamae subsp. nakayamae sp. nov., subsp. tion of a heme-responsive operon. J Bacteriol 19:4903–4911 nov., Sporolactobacillus nakayamae subsp. racemicus subsp. nov., 778 Ann Microbiol (2017) 67:763–778

Sporolactobacillus terrae sp. nov., Sporolactobacillus kofuensis sp. Zhao T, Liu D, Ren H, Shi X, Zhao N, Chen Y, Ying H (2014) D-lactic nov., and Sporolactobacillus lactosus sp. nov. Int J Syst Bacteriol acid production by Sporolactobacillus inulinus Y2-8 immobilized in 47:499–504 fibrous bed bioreactor using corn flour hydrolyzate. J Microbiol. Zagari F, Jordan M, Stettler M, Broly H, Wurn FM (2013) Lactate me- Biotechnol 24:1664–1672 tabolism shift in CHO cell culture: the role of mitochondrial oxida- Zheng L, Bai Z, Xu T, He B (2012) Glucokinase contributes to glucose tive activity. New Biotech 30:238–245 phosphorylation in D-lactic acid production by Sporolactobacillus Zhao B, Wang L, Li F, Hua D, Ma C, Ma Y, Xu P (2010) Kinetics of D- inulinus Y2-8. J Ind Microbiol Biotechnol 39:1685–1692 lactic acid production by Sporolactobacillus sp. strain CASD using repeated batch fermentation. Bioresour Technol 101:6499–6505