Efficient Production of Optically Pure d-Lactic Acid from Raw Corn Starch by Using a Genetically Modified l-Lactate Dehydrogenase Gene-Deficient and α-Amylase-Secreting Lactobacillus plantarum Strain

Abstract

In order to achieve direct and efficient fermentation of optically pure d-lactic acid from raw corn starch, we constructed l-lactate dehydrogenase gene (ldhL1)-deficient Lactobacillus plantarum and introduced a plasmid encoding Streptococcus bovis 148 α-amylase (AmyA). The resulting strain produced only d-lactic acid from glucose and successfully expressed amyA. With the aid of secreting AmyA, direct d-lactic acid fermentation from raw corn starch was accomplished. After 48 h of fermentation, 73.2 g/liter of lactic acid was produced with a high yield (0.85 g per g of consumed sugar) and an optical purity of 99.6%. Moreover, a strain replacing the ldhL1 gene with an amyA-secreting expression cassette was constructed. Using this strain, direct d-lactic acid fermentation from raw corn starch was accomplished in the absence of selective pressure by antibiotics. This is the first report of direct d-lactic acid fermentation from raw starch.

Poly-lactic acid (PLA) is an important agro-based plastic that can be produced from inexpensive, renewable, and abundantly available biomass resources, including starchy materials. These resources have advantages over limited oil- and fossil-based sources, as they do not result in any net carbon dioxide release to the atmosphere (7). Recently, stereocomplex PLA, which is composed of both poly-l- and -d-lactic acid, has been attracting much attention due to its high thermostability. Stereocomplex-type polymers show a melting point (ca. 230°C) that is approximately 50°C higher than that of the respective single polymers (8). Therefore, d-lactic acid, in addition to l-lactic acid, which has been the focus of production to date, is of significant importance.

Lactic acid bacteria (LAB) are promising microorganisms for the efficient production of lactic acid from various sugars, such as glucose, sucrose, and lactose. However, when starchy materials are used as a carbon source, they must be saccharified by physicochemical and enzymatic treatment because most LAB cannot utilize starchy materials directly (13). This makes the whole process less economically viable. Therefore, many researchers have examined the direct production of lactic acid from starchy materials by using wild amylolytic LAB (ALAB) (6, 24, 25) or genetically modified amylase-producing LAB (15, 16). Although d-lactic acid has been produced by fermentation from pretreated substrates such as rice starch (5) and by simultaneous saccharification and fermentation from cellulose (23), there have been no reports on the direct production of d-lactic acid from starchy materials. This is due to a lack of d-lactic acid-producing ALAB and difficulties in gene manipulation of d-lactic acid-producing LAB, such as Lactobacillus delbrueckii (22).

We focused on Lactobacillus plantarum, which is an industrially important strain due to its environmental flexibility and its ability to assimilate a wide range of carbohydrates (9). In recent years, several gene manipulation methods for Lactobacillus plantarum have been established (18, 19). Moreover, the complete genome sequence has been decoded for L. plantarum NCIMB 8826 (9). Based on whole-genome analysis, L. plantarum possesses two types of lactate dehydrogenase (LDH), l-LDH and d-LDH, which convert pyruvate into l- and d-lactic acid, respectively. Ferain et al. (4) reported that chromosomal deletion in the ldhL1 gene of L. plantarum NCIMB 8826 provoked an absence of l-LDH activity and produced d-lactic acid from glucose.

In the present study, to produce d-lactic acid directly from starch, we constructed an l-LDH-deficient, α-amylase-secreting L. plantarum strain. The engineered strain expressed α-amylase from Streptococcus bovis 148 (AmyA) (20) and efficiently degraded raw starch with the aid of a C-terminal starch-binding domain (11). Using this strain, we achieved the direct and efficient fermentation of optically pure d-lactic acid from raw corn starch.

MATERIALS AND METHODS

Bacterial strains and media.

The bacterial strains used in this study are listed in Table 1. Escherichia coli VE 7108 (12) was used for pG⁺host9-based DNA manipulation (10). It was grown in Luria-Bertani (LB) medium containing 250 μg/ml erythromycin and 10 μg/ml kanamycin at 37°C. L. plantarum NCIMB 8826 and its derivative were grown in MRS broth (Difco Laboratories, Detroit, MI) or MRS broth containing 25 μg/ml erythromycin at 37°C. For solid media, 1.5% (wt/vol) agar was added to the media described above.

TABLE 1.

Strains, plasmids, and oligonucleotide primers used in this study

Strain, plasmid, or primer	Relevant phenotype and description or sequence (5′-3′)a	Source or reference
Strains
Escherichia coli strains
VE7108	Kanamycin resistance marker; contains the wild-type repA plasmid gene (not thermosensitive); host for pG+host9 DNA manipulation	12
VE6838	Kanamycin resistance marker; VE7108 carrying pG+host9	12
Lactobacillus plantarum strains
WT	NCIMB 8826 WT strain	NCIMB
ΔldhL1	NCIMB 8826 ΔldhL1 strain	This study
ldhL1::amyA	NCIMB 8826 strain with ldhL1 gene replaced with amyA-secreting expression cassette	This study
Plasmids
pCUSαA	Vector for secreting α-amylase; erythromycin resistance marker	16
pG+host9	Thermosensitive replicon; erythromycin resistance marker	10
pGh9-ΔldhL1	Vector for deletion of ldhL1 gene of NCIMB 8826; erythromycin resistance marker	This study
pGh9-ldhL1::amyA	Vector for replacement of ldhL1 gene of NCIMB 8826 with amyA-secreting expression cassette; erythromycin resistance marker	This study
Oligonucleotide primers
ldhL1-up_F	CCGCTCGAGAATCCGGTGATTAGTTGCCG
ldhL1-up_R	ACGCGTCGACGAGCTCAGATCTAATAAGTCATCCTCTCGTAGTGAAAA
ldhL1-down_F	CCCGATATCGTCGACTCATTTCATACGATTAAATGTATGATGAACG
ldhL1-down_R	GGACTAGTAACAGCGTTTTGGTCTAAACTTACTTT
ldhL1-up_seq	AAATCTTCTCACCGTCTTGAAAAG
ldhL1-down_seq	TGGTTAAATCTTTTTTAGTATTGATTGTAATACC

^aFor primers, restriction enzyme sites are underlined.

Plasmid construction.

The oligonucleotides and plasmids used in this study are summarized in Table 1. PCR was carried out using KOD-Plus polymerase (Toyobo Co. Ltd., Osaka, Japan). The plasmid for disruption of the l-LDH gene (ldhL1) was constructed as follows. The 1,000-bp upstream region from the start codon of ldhL1 and the 1,000-bp downstream region from the stop codon of ldhL1 were amplified by PCR from the genome of L. plantarum NCIMB 8826, using oligonucleotide primers ldhL1-up_F plus ldhL1-up_R and ldhL1-down_F plus ldhL1-down_R, respectively. The resulting fragments were digested with SalI and ligated. Using the ligated fragment (2,000 bp) as a template, the same fragment was amplified by PCR using oligonucleotide primers ldhL1-up_F and ldhL1-down_R. The amplified fragment was digested with XhoI and SpeI and subsequently inserted into the XhoI and SpeI sites of the plasmid pG⁺host9 (10). The resulting plasmid was designated pGh9-ΔldhL1 (Fig. 1a). The plasmid to replace ldhL1 with an amyA-secreting expression cassette was constructed as follows. The expression cassette of amyA consisted of a clpC UTLS promoter, which consists of the clpC core promoter and an untranslated leader sequence (14), the amyA signal sequence, the mature region of amyA, and the terminator of the conjugated bile acid hydrolase gene (cbh) (2) and was obtained from pCUSαA (Fig. 1b) (16) by digestion with BglII and SacI. It was subsequently inserted into the BglII and SacI sites of plasmid pGh9-ΔldhL1 (Fig. 1a). The resulting plasmid was designated pGh9-ldhL1::amyA.

FIG. 1.

(a) Strategy of double-crossover ldhL1 deletion and substitution utilizing thermosensitivity of pG⁺host plasmids. Ts ori, temperature-sensitive replication origin; Em^r, erythromycin resistance gene; up, 1,000-bp upstream region of ldhL1 gene; down, 1,000-bp downstream region of ldhL1 gene. (b) Schematic illustration of α-amylase secretion plasmid. P-clpCUTLS, clpC UTLS promoter; S.S., signal sequence of amyA gene; Tcbh, terminator of cbh gene; Rep, replication origin; Em^r, erythromycin resistance gene. The amyA-secreting expression cassette is obtained by digestion with BglII and SacI. (c) Agarose gel electrophoresis after gene deletion and integration and schematic illustration of amplified regions by PCR. Lanes and/or illustrations: M, marker DNA with base pairs indicated; 1, WT; 2, ΔldhL1 strain; 3, ldhL1::amyA strain.

Disruption and substitution of ldhL1 gene of L. plantarum.

Disruption and substitution of the ldhL1 gene of L. plantarum NCIMB 8826 were carried out using pG⁺host plasmid-based double-crossover homologous integration as described by Biswas et al. (1). pGh9-ΔldhL1 was introduced into L. plantarum NCIMB 8826 by electroporation, as described previously (15), and then, utilizing the thermosensitivity of the pG⁺host9 plasmid, ΔldhL1 and ldhL1::amyA mutants were obtained according to the scheme illustrated in Fig. 1a. L. plantarum possessing pGh9-ΔldhL1 was cultivated at 42°C under selective conditions by addition of antibiotics, and the first recombination event (integration) occurred through the “up” region, after which pGh9-ΔldhL1 was integrated into the L. plantarum chromosome. Integrants were cultivated at 28°C under nonselective conditions, and the second recombination event (excision) occurred. ΔldhL1 mutants were obtained as the strain underwent gene excision through the “down” region, as excision through the up region restored the parental chromosome structure. ldhL1 substitution with the amyA-secreting expression cassette was carried out by the same procedure, using plasmid pGh9-ldhL1::amyA. Deletion and substitution of the ldhL1 gene were confirmed by PCR using primers ldhL1-up_seq and ldhL1-down_seq, which are forward and reverse primers and anneal the upstream region (bp 477 to 500) and downstream region (bp 467 to 500) of ldhL1, respectively. pCUSαA (16) was then introduced into both the L. plantarum NCIMB 8826 wild-type (WT) strain and the ΔldhL1 strain for secretion of AmyA. The resulting transformants were designated WT/pCUSαA and ΔldhL1/pCUSαA, respectively.

Fermentation experiments using a 2.0-liter bioreactor.

All fermentation experiments were performed in a 2.0-liter bioreactor (Able & Biott, Tokyo, Japan) with a 700-ml working volume. The fermentor, containing liquid modified MRS medium (MRS without glucose, beef extract, sodium acetate, and Tween 80), was heat sterilized (121°C, 15 min). Next, heat-sterilized glucose solution or nonsterilized raw corn starch (Wako Pure Chemical Industries, Ltd., Osaka, Japan) was added to the fermentor (to a final concentration of 100 g/liter). Erythromycin (in the case of fermentation using the WT/pCUSαA and ΔldhL1/pCUSαA strains) or cycloheximide (in the case of fermentation using the ldhL1::amyA strain) was then added to a final concentration of 25 or 10 μg/ml, respectively. Subsequently, 10 M H₂SO₄ was added to the medium to adjust the pH to 5.5, and 7 ml of inoculum (adjusted to an optical density at 600 nm [OD₆₀₀] of 10 with sterile distilled water) was added to the fermentor. Prior to this addition, the inoculum was grown in MRS medium and subcultured at regular intervals (12 h) to stabilize the growth rate. The temperature was maintained at 37°C, the agitation speed was kept at 100 rpm, and the pH was kept at approximately 5.5 (±0.03) by the automatic addition of 10 M NH₃ solution. After incubation, the culture was regularly harvested and subjected to analysis.

Growth of L. plantarum NCIMB 8826 cells was monitored by measuring the OD₆₀₀ or counting viable cells by the pour plate method, using bromocresol purple plate count agar (Nissui Pharmaceutical Co., Tokyo, Japan). The glucose concentration was measured using the Glucose CII test (Wako Pure Chemical Industries, Ltd., Osaka, Japan). Total sugar concentrations of raw corn starch were determined by colorimetric assay based on the phenol-sulfuric acid reaction described by Dubois et al. (3) as the glucose equivalent. Lactic acid concentrations were measured by high-performance liquid chromatography, as described previously (15). The optical purity of produced lactic acid was measured using a BF-5 biosensor (Oji Scientific Instruments, Hyogo, Japan). For the detection of l-lactic acid, an l-lactic acid enzyme electrode was used, and for the detection of d-lactic acid, a d-lactic acid enzyme electrode and d-lactic acid kits were used in accordance with the manufacturer's instructions. The optical purity of lactic acid is defined as follows: optical purity (%) = |d-lactic acid concentration − l-lactic acid concentration|/(d-lactic acid concentration + l-lactic acid concentration) × 100. α-Amylase activity in the culture supernatant was measured as described previously (16), using an α-amylase measurement kit (Kikkoman Co., Chiba, Japan).

RESULTS

Construction of ldhL1-deficient and substitution strains.

Construction of L. plantarum mutants having the ldhL1 gene disrupted or replaced with an amyA-secreting expression cassette was carried out (Fig. 1a). On agarose gel electrophoresis, the ldhL1-deficient strain showed a smaller band due to lack of the ldhL1 gene, while the ldhL1 substitution strain showed a larger band (Fig. 1c) than that of the WT strain. These shifts corresponded with the expected changes (−963 and +1,621 bp, respectively). These results confirmed the deletion or replacement of the ldhL1 gene. Further confirmation was carried out by DNA sequencing analyses (data not shown).

Lactic acid fermentation from glucose by WT/pCUSαA and ΔldhL1/pCUSαA strains.

In order to confirm the effects of ldhL1 gene deletion on fermentative performance and heterologous protein production, lactic acid fermentation from glucose was carried out using the WT/pCUSαA and ΔldhL1/pCUSαA strains. Although one of the two LDHs was disrupted in the ldhL1-deficient strain, similar cell growth (Fig. 2a), glucose consumption, and lactic acid production (Fig. 2b) were observed in both WT/pCUSαA and ΔldhL1/pCUSαA. Production of AmyA was confirmed by Western blotting for both strains (data not shown). In ΔldhL1/pCUSαA, a slightly higher maximum α-amylase activity (0.672 U/ml) than that of WT/pCUSαA (0.505 U/ml) was observed (Fig. 2a). After 36 h of fermentation, both strains consumed all sugar and produced 86.0 and 86.6 g/liter of lactic acid, respectively. These results suggest that deletion of the ldhL1 gene has little effect on fermentative performance and that the ldhL1-deficient mutant is capable of producing heterologous protein similarly to the WT strain.

FIG. 2.

Lactic acid fermentation from glucose by L. plantarum NCIMB 8826 WT (closed symbols) and ΔldhL1 (open symbols) strains harboring pCUSαA. (a) OD₆₀₀ of culture (triangles) and α-amylase activity in culture supernatant (diamonds). (b) Glucose (squares) and lactic acid (circles) concentrations. Data points represent means and standard deviations for three independent experiments.

We then analyzed the optical purity of the produced lactic acid. The lactic acid produced by WT/pCUSαA consisted of approximately the same amounts of both isomers, and the optical purity of the produced lactic acid was 12.6% (Table 2). In contrast, the lactic acid produced by ΔldhL1/pCUSαA consisted mainly of d-lactic acid, and the optical purity was very high, at 99.7% (Table 2).

TABLE 2.

Various parameters in lactic acid fermentation^a

Parameter	Glucose fermentation		Starch fermentation
	WT/pCUSαA	ΔldhL1/pCUSαA	WT/pCUSαA	ΔldhL1/pCUSαA	ldhL1::amyA strain
Max volumetric productivity (g/liter/h)	4.51 ± 0.06	4.54 ± 0.44	4.52 ± 0.14	3.86 ± 0.38	2.06 ± 0.21
Yield (g lactic acid produced per g of total sugar consumed)	0.89 ± 0.03	0.89 ± 0.03	0.83 ± 0.03	0.85 ± 0.03	0.86 ± 0.02
Optical purity of produced lactic acid (%)	12.6 ± 3.11	99.7 ± 0.04	4.0 ± 0.74	99.6 ± 0.04	99.2 ± 0.37

^aThe values represent means ± standard deviations for three independent experiments.

Lactic acid fermentation from raw corn starch, using the WT/pCUSαA and ΔldhL1/pCUSαA strains.

Encouraged by these findings, we carried out lactic acid fermentation from raw corn starch, using the WT/pCUSαA and ΔldhL1/pCUSαA strains. As shown in Fig. 3a, both WT/pCUSαA and ΔldhL1/pCUSαA showed similar α-amylase activities in culture supernatants, and these activities reached 0.505 and 0.714 U/ml, respectively, at 18 h of incubation. With the aid of α-amylase activity, both strains were able to grow in raw starch as the sole carbon source (Fig. 3a), while the corresponding plasmid-free strains could not (data not shown). After 18 h of fermentation, viable cell counts reached 7.12 × 10⁹ and 7.17 × 10⁹ CFU/ml, respectively. After 48 h of fermentation, WT/pCUSαA and ΔldhL1/pCUSαA consumed 87.1 and 86.2 g/liter sugar and produced 72.6 and 73.2 g/liter lactic acid, respectively (Fig. 3b). In addition, the optical purity of lactic acid produced by ΔldhL1/pCUSαA had an extremely high value (99.6%), while that produced by WT/pCUSαA had a low value (4.0%) (Table 2). These results confirm direct and efficient d-lactic acid fermentation from raw corn starch by ΔldhL1/pCUSαA.

FIG. 3.

Lactic acid fermentation from raw corn starch by L. plantarum NCIMB 8826 WT (closed symbols) and ΔldhL1 (open symbols) strains harboring pCUSαA. (a) Viable cell counts in culture (triangles) and α-amylase activity in culture supernatant (diamonds). (b) Total sugar (squares) and lactic acid (circles) concentrations. Data points represent means and standard deviations for three independent experiments.

Lactic acid fermentation from raw corn starch by ldhL1::amyA strain.

In order to conduct stable expression of amyA without selective pressure by antibiotics to maintain the plasmid, chromosomal ldhL1 replacement with an amyA-secreting expression cassette was carried out. Using the ldhL1::amyA strain, fermentation from raw corn starch was then performed. As shown in Fig. 4a, the integrated amyA gene was successfully expressed and exhibited its activity. The maximum α-amylase activity of ldhL1::amyA was 0.052 U/ml (Fig. 4a). The ldhL1::amyA strain was also able to grow using raw starch as the sole carbon source (Fig. 4a and b) and efficiently produced lactic acid (Fig. 4b). After 72 h of fermentation, the ldhL1::amyA strain consumed 72.2 g/liter sugar and produced 61.9 g/liter lactic acid. In addition, the optical purity of the produced lactic acid had a high value (99.2%) (Table 2). These results confirm the successful d-lactic acid fermentation from raw corn starch by the ldhL1::amyA strain under cost-efficient, nonselective conditions.

FIG. 4.

Lactic acid fermentation from raw corn starch by L. plantarum NCIMB 8826 ldhL1::amyA strain. (a) Viable cell counts in culture (closed triangles) and α-amylase activity in culture supernatant (open diamonds). (b) Total sugar (closed squares) and lactic acid (open circles) concentrations. Data points represent means and standard deviations for three independent experiments.

DISCUSSION

The aim of this study was to achieve the direct fermentation of optically pure d-lactic acid from biomass resources in order to reduce the production costs of PLA. We initially investigated the construction of d-lactic acid-producing LAB from gene manipulation-facilitated d,l-lactic acid-producing strains, as gene manipulation of d-lactic acid-producing LAB is known to be difficult (22). Ferain et al. (4) reported that chromosomal deletion of the ldhL1 gene of L. plantarum has little effect on fermentative kinetics and that the ΔldhL1 strain exclusively produces d-lactic acid with 20 g/liter glucose as a carbon source. These features prompted us to use L. plantarum as a host strain. However, the mass production of lactic acid from high-concentration substrate or biomass resources, including starchy materials, and heterologous protein expression using the ΔldhL1 strain have not been attempted. Therefore, we attempted fermentation from high-concentration glucose (100 g/liter) and the expression of amyA.

As shown in Fig. 2, disruption of ldhL1 had little effect on fermentative kinetics and ΔldhL1/pCUSαA exclusively produced d-lactic acid (99.7% optical purity) (Table 2), even with high concentrations of glucose. Moreover, with expression of amyA (Fig. 2b), ΔldhL1/pCUSαA showed a higher maximum α-amylase activity (0.672 U/ml) than that of WT/pCUSαA (0.505 U/ml). It has been reported that LDH is a strongly expressed protein in lactobacilli such as Lactobacillus casei and Lactobacillus helveticus (17, 21). One possible explanation is that in the L. plantarum ΔldhL1 strain, the absence of strongly expressed l-LDH provokes the increased expression of other proteins. Therefore, the ΔldhL1 strain is considered to be a useful host for heterologous protein production, the same as the WT strain.

Subsequently, lactic acid fermentation from raw corn starch was carried out. As with fermentation from glucose, ΔldhL1/pCUSαA exclusively produced d-lactic acid (99.6% optical purity) (Table 2). Despite a low initial fermentation rate compared to that of glucose fermentation, which was attributed to the initial degradation of starch, the maximum volumetric productivity of lactic acid in starch fermentation (3.86 g/liter/h) was comparable to that in glucose fermentation (4.54 g/liter/h) by ΔldhL1/pCUSαA (Table 2). The yields of lactic acid (g of lactic acid per g of consumed sugar) were also approximately the same between starch (0.85) and glucose (0.89) fermentation by ΔldhL1/pCUSαA (Table 2). These values are comparable to those in previous reports describing direct l-lactic acid fermentation or a mixture of l- and d-lactic acid fermentation from raw starch, using wild ALAB, S. bovis 148 (0.88) (13), and L. plantarum A6 (0.84) (6). Based on these results, direct and efficient d-lactic acid fermentation from starch was successful. The L. plantarum ΔldhL1 strain we developed is an attractive host for d-lactic acid production from biomass resources, such as not only starchy materials but also cellulosic or hemicellulosic materials, in the future.

In an effort to test industrial applications, a strain with the ldhL1 gene replaced with the amyA-secreting expression cassette was constructed. Since this ldhL1::amyA strain does not require selective culture with antibiotics to retain the plasmid and possesses amyA in the chromosome, it is considered to be advantageous with regard to both cost and genetic stability compared to ΔldhL1/pCUSαA. As shown in Fig. 4, successful d-lactic acid fermentation was achieved using the ldhL1::amyA strain, with a high yield (0.86 g per g of consumed sugar) and high optical purity (99.2%) (Table 2). However, the maximum volumetric productivity was lower (2.06 g/liter/h) (Table 2) than that of ΔldhL1/pCUSαA (3.86 g/liter/h), and this may be due to low α-amylase activity (14-fold lower than that of ΔldhL1/pCUSαA). An increase of the chromosomal copy number of amyA and an improved expression system will improve the α-amylase activity and increase lactic acid productivity. We also confirmed by Western blotting that no degradation of AmyA was present after 36 h, which suggests that the decreased α-amylase activity was caused by inactivation of AmyA. Unfortunately, the cause of this inactivation of AmyA is uncertain, but stabilization of AmyA is a promising method for improving lactic acid productivity.

In conclusion, our results successfully demonstrated the direct and efficient fermentation of optically pure d-lactic acid from raw corn starch by ldhL1-deficient AmyA-secreting L. plantarum. The results obtained in this study strongly indicate that L. plantarum ldhL1-deficient and substitution strains are useful as hosts for d-lactic acid production from biomass resources.