Metabolic Engineering of a Glycerol-Oxidative Pathway in Lactobacillus panis PM1 for Utilization of Bioethanol Thin Stillage: Potential To Produce Platform Chemicals from Glycerol

Abstract

Lactobacillus panis PM1 has the ability to produce 1,3-propanediol (1,3-PDO) from thin stillage (TS), which is the major waste material after bioethanol production, and is therefore of significance. However, the fact that L. panis PM1 cannot use glycerol as a sole carbon source presents a considerable problem in terms of utilization of this strain in a wide range of industrial applications. Accordingly, L. panis PM1 was genetically engineered to directly utilize TS as a fermentable substrate for the production of valuable platform chemicals without the need for exogenous nutrient supplementation (e.g., sugars and nitrogen sources). An artificial glycerol-oxidative pathway, comprised of glycerol facilitator, glycerol kinase, glycerol 3-phosphate dehydrogenase, triosephosphate isomerase, and NADPH-dependent aldehyde reductase genes of Escherichia coli, was introduced into L. panis PM1 in order to directly utilize glycerol for the production of energy for growth and value-added chemicals. A pH 6.5 culture converted glycerol to mainly lactic acid (85.43 mM), whereas a significant amount of 1,3-propanediol (59.96 mM) was formed at pH 7.5. Regardless of the pH, ethanol (82.16 to 83.22 mM) was produced from TS fermentations, confirming that the artificial pathway metabolized glycerol for energy production and converted it into lactic acid or 1,3-PDO and ethanol in a pH-dependent manner. This study demonstrates the cost-effective conversion of TS to value-added chemicals by the engineered PM1 strain cultured under industrial conditions. Thus, application of this strain or these research findings can contribute to reduced costs of bioethanol production.

INTRODUCTION

Thin stillage (TS) is a major fermentation residue from the dry-grind process of bioethanol production. Its downstream treatment, including making a condensed form for animal feed, is an energy- and cost-intensive process (1), while its utilization as animal feed is hardly a high-value application of this industrial by-product and hence does not significantly reduce the cost of ethanol production. TS contains a variety of complex nutrients, including various carbohydrates, minerals, and amino acids. Glycerol (up to 2%) is generated in TS during bioethanol production regardless of feedstock (i.e., sugar cane or corn) (2). Glycerol is at a higher reduced state than carbohydrates (i.e., fermentable sugars). The reduced nature of glycerol allows synthesis of fuels and other reduced chemicals at higher yields with minimal by-products compared to common sugars (e.g., glucose and xylose); however, the high degree of reduction in glycerol makes it difficult for microorganisms to utilize under anaerobic conditions. A few bacteria, including Klebsiella, Enterobacter, and Clostridium, have been exploited to bioconvert glycerol to industrially relevant chemicals (e.g., 1,3-propanediol [1,3-PDO], 3-hydroxypropionic acid, butanol, and ethanol) (3–7). Since these organisms are classified as opportunistic pathogens, industrial applications of these strains may pose issues, particularly regarding worker biosafety.

Lactobacilli have been long considered generally recognized as safe (GRAS) microorganisms. Because of their small genomes and relatively simple metabolic pathways, lactobacilli have been intensively utilized as cell factories for the production of platform chemicals, including lactic acid, 2,3-butanediol, 1,3-PDO, ethanol, food flavoring agents, sweeteners, and complex polysaccharides, through metabolic engineering approaches (8). Lactobacillus panis PM1 is a natural 1,3-PDO-producing organism originally isolated from TS (9). This strain achieves primary carbohydrate fermentation through substrate-level phosphorylation via the 6-phosphogluconate/phosphoketolase (6-PG/PK) pathway wherein NADH is mainly recycled back to NAD⁺ through the production of lactic acid and ethanol (Fig. 1). In the presence of glycerol, the glycerol-reductive pathway (glycerol to 1,3-PDO) is a major NADH recycling system used in conjunction with the ethanol production pathway. The consumption of the common cofactor, NADH, is shifted from ethanol to 1,3-PDO production, yielding one extra ATP per glucose (10). However, glycerol metabolism in L. panis is not connected to the central metabolic (6-PG/PK) pathway, and thus strain PM1 cannot utilize glycerol as a sole carbon source for energy. In a previous paper, we demonstrated that TS is a good culture medium for 1,3-PDO fermentation by L. panis PM1, providing a final 1,3-PDO concentration of 16.23 g/liter and yield of 0.72 g/g from TS supplemented to 30.0 g/liter glucose and 25.0 g/liter glycerol; however, low indigenous fermentable sugar content in TS (i.e., sugars are depleted during ethanol fermentation) is the main bottleneck for the cost-effective glycerol bioconversion by the wild-type PM1 strain (11). Therefore, the absence of a glycerol-oxidative pathway (glycerol to glyceraldehyde 3-phosphate) indeed hampers a wide range of industrial applications by this strain.

FIG 1.

Schematic diagram of the glycerol metabolic pathway in the engineered PM1 strain. Dashed lines represent metabolically engineered pathways. Abbreviations: ADH, alcohol dehydrogenase; AK, acetate kinase; ALDH, acetaldehyde dehydrogenase; DhaB, glycerol dehydratase; DhaT, 1,3-propanediol dehydrogenase; ENO, enolase; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GlpD, glycerol 3-phosphate dehydrogenase; GlpF, glycerol facilitator; GlpK, glycerol kinase; LDH, lactate dehydrogenase; PDH, pyruvate dehydrogenase; PGK, phosphoglycerate kinase; PGM, phosphoglyceromutase; PK, pyruvate kinase; PTA, phosphotransacetylase; TPI, triosephosphate isomerase; YqhD, NADPH-dependent aldehyde reductase.

In an effort to develop a cost-effective approach to processing glycerol based on utilization of TS, we metabolically engineered L. panis PM1 for the production of more-valuable platform chemicals from glycerol in TS. In this article, we describe the successful expression of an artificial glycerol-oxidative pathway, composed of the glycerol facilitator (GlpF), glycerol kinase (GlpK), glycerol 3-phosphate dehydrogenase (GlpD), triosephosphate isomerase (TPI), and NADPH-dependent aldehyde reductase (YqhD) genes of Escherichia coli, in L. panis PM1. The engineered strain can produce value-added chemicals, including lactic acid, ethanol, and 1,3-PDO, from TS without the need for additional fermentable sugar and nutrient supplementation (e.g., glucose, yeast extract, beef extract, peptone, and vitamin B₁₂).

MATERIALS AND METHODS

Thin stillage source and liquid stillage preparation.

Thin stillage (TS) remaining after bioethanol production was obtained from Pound-Maker Agventures Ltd. (Lanigan, SK, Canada). Organic components of four batches of TS produced by this company were analyzed by ¹H NMR and high-pressure liquid chromatography (HPLC), showing dextrin (8.47 to 11.65 g/liter), maltotriose (0.14 to 1.10 g/liter), maltose monohydrate (0.03 to 1.05 g/liter), glycerol (2.39 to 7.87 g/liter), ethanol (0.23 to 1.31 g/liter), lactic acid (5.07 to 7.41 g/liter), and acetic acid (0.56 to 1.14 g/liter) as major components (12). After centrifugation (6,084 × g, 20 min, 15°C) by a Sorvall RC-5C centrifuge (Du Pont Instruments, Hoffman Estates, IL, USA), solid matter was removed from the TS, and the resultant liquid part (liquid stillage [LS]) was sterilized at 121°C for 20 min.

Bacterial strains and growth conditions.

Lactobacillus panis PM1 and its engineered strains were cultured in commercial MRS medium (BD, Franklin Lakes, NJ, USA) and modified MRS (mMRS) medium (or LS), respectively. The mMRS medium consisted of 5 g yeast extract, 10 g peptone, 10 g meat extract, 2 g K₂HPO₄, 2 g ammonium citrate, 5 g sodium acetate, 100 mg MgSO₄·7H₂O, 50 mg MnSO₄, and a defined concentration of glycerol per liter. The cultures were incubated at 37°C under microaerobic conditions, unless otherwise stated. For acidic media, the initial pH of 6.5 was adjusted to a pH of 4.5 through the addition of 10 N HCl. Air-tight 15-ml tubes, filled to the two-thirds level, were incubated under static conditions to establish microaerobic conditions. Escherichia coli TOP 10 (Invitrogen, Carlsbad, CA, USA) was used for cloning and preparation of the target plasmid and cultured in LB medium at 37°C with vigorous shaking. The concentrations of erythromycin (Em) for selective plating were 10 μg/ml and 300 μg/ml for L. panis PM1 and E. coli, respectively.

General DNA techniques, plasmid construction, and bacterial transformation.

DNA work was carried out according to standard protocols, described in reference 13. Restriction endonucleases were purchased from Fermentas (Waltham, MA, USA), and digestions were performed according to the recommendations of the manufacturer. A Bio-Basic gel extraction kit (Bio-Basic Inc., Markham, ON, Canada) was used to isolate specific restriction and PCR fragments from agarose gels. DNA ligations and subsequent transformations into competent E. coli TOP 10 were carried out according to standard protocols (13). The genes encoding GlpFK, GlpD, TPI, and YqhD were amplified using the genomic DNA of E. coli JM109 as the template. For PCR amplification, a 50-μl PCR mixture was prepared using 0.5 μM (each) primers (Table 1), 1.25 U Pfu DNA Polymerase (Fermentas), 1× Pfu buffer with MgCl₂, and 0.2 mM deoxynucleoside triphosphates (dNTPs). The DNA template was denatured for 3 min at 95°C and was then PCR amplified using a thermal cycler (FTGene-5D; Techgene, Burlington, NJ, USA) over 30 cycles of 30 s at 95°C, 30 s at 55 to 60°C, and 72°C for 1 to 3 min. After the final cycle, the mixtures were further incubated for 5 min at 72°C. PCR fragments for GlpFK (introduced restriction sites, SalI/XbaI), GlpD (introduced restriction sites, XbaI/SacI), TPI (introduced restriction sites, NcoI/SacI), and YqhD (introduced restriction sites, SacI) were consecutively cloned into pUC18 to yield pUC18-GlpFK, pUC18-GlpFKD, pUC18-GlpFKD-TPI, and pUC18-GlpFKD-TPI-YqhD, respectively. The GlpFKD-TPI (4.7-kb) and GlpFKD-TPI-YqhD (5.9-kb) fragments from each construct were amplified using fGlpFK and rM13 primers, digested with NotI and SalI restriction endonucleases, and cloned into the L. panis PM1 expression vector pCER-EGFP (14) to yield the pCER-GlpFKD-TPI and pCER-GlpFKD-TPI-YqhD plasmids (Table 2 and Fig. 2), respectively. These plasmids were then transformed into L. panis PM1 by the electroporation method, as described previously (15).

TABLE 1.

Primers used in this study

Primer	Restriction enzyme(s)	Tm (°C)a	Nucleotide sequence (5′ →3′)b	Target gene(s)	Function
fGlpFK	SalI/NotI	54	ACGCGTCGACGCGGCCGCATGAGTCAAACATCAACCTTG	GlpF and GlpK genes	Glycerol facilitator and glycerol kinase
rGlpFK	XbaI	59	TGCTCTAGATTATTCGTCGTGTTCTTCCCA
fGlpD	XbaI	60	GCTCTAGAGAAGGAGAATGGAAACCAAAGATCTGATTGTG	GlpD gene	Glycerol 3-phosphate dehydrogenase
rGlpD	SacI/NcoI	63	TCCGAGCTCCCATGGTTACGACGCCAGCGATAACCT
fTPI	NcoI	56	CATGCCATGGGAGAAAGAATGCGACATCCTTTAGTGATG	TPI gene	Triosephosphate isomerase
rTPI	SacI	59	TCCGAGCTCTTAAGCCTGTTTAGCCGCTTC
fYqhD	SacI	51	TCCGAGCTCGAAGGAGAATGAACAACTTTAATCTGCAC	YqhD gene	NADPH-dependent aldehyde reductase
rYqhD	SacI	64	TCCGAGCTCTTAGCGGGCGGCTTCGTATAT
rM13	SalI	45	ACGCGTCGACCAGGAAACAGCTATGAC		pUC/M13 sequencing reverse primer

^aT_m, melting temperature.

^bUnderlined and bolded primer sequences represent restriction enzyme sites and artificial ribosome binding sites, respectively.

TABLE 2.

Plasmids used in this study

Plasmid	Relevant features	Source or reference
pUC18	Cloning vector, Apr, 2.7 kb	Invitrogen
pUC18-GlpKF	pUC18 derivative carrying the GlpF and GlpK genes from E. coli JM109	This study
pUC18-GlpKFD	pUC18 derivative carrying the GlpF, GlpK, and GlpD genes from E. coli JM109	This study
pUC18-GlpKFD-TPI	pUC18 derivative carrying the GlpF, GlpK, GlpD, and TPI genes from E. coli JM109	This study
pUC18-GlpKFD-TPI-YqhD	pUC18 derivative carrying the GlpF, GlpK, GlpD, TPI, and YqhD genes from E. coli JM109	This study
pCER-EGFP	Lactobacillus panis PM1 expression vector, in which the EGFPa gene was expressed under the control of PorfX promoter, Emr Cmr, 6.0 kb	14
pCER-GlpKFD-TPI	pCER-EGFP derivative, in which the EGFP gene was replaced with the GlpKFD-TPI fragment, Emr Cmr, 10.7 kb	This study
pCER-GlpKFD-TPI-YqhD	pCER-EGFP derivative, in which the EGFP gene was replaced with the GlpKFD-TPI-YqhD fragment, Emr Cmr, 11.9 kb	This study

^aEGFP, enhanced green fluorescent protein.

FIG 2.

Construction of the glycerol artificial operons. Black regions shown between each gene represent artificial ribosomal binding sites.

pH-uncontrolled and -controlled batch fermentation.

Batch fermentation (pH uncontrolled) was performed in 50-ml conical tubes containing 40 ml of mMRS medium. The concentration of glycerol in mMRS was set at 160 mM, and sodium acetate was at 0 or 50 mM. The fermentation tests were conducted at 37°C. Batch fermentation (pH controlled) was performed in a 5-liter fermentor (Bioflow III; New Brunswick Scientific [NBS] Co. Inc., Edison, NJ, USA) with 3 liters of LS (initial glycerol concentration, 160 mM), into which 50 ml mid-log-phase culture was inoculated. The fermentation temperature was adjusted to 37°C, pH was controlled at 6.5 (for lactic acid fermentation) or 7.5 (for 1,3-PDO fermentation) with 5 N NaOH, and agitation was set at 80 rpm to mix the pH control agent.

Determination of glycerol and end products.

The optical density at 600 nm (OD₆₀₀) was measured with a DU 800 spectrophotometer (Beckman Coulter, Mississauga, ON, Canada). After centrifugation, the supernatant was filtered through a 0.22-μm-pore-size filter and stored at −20°C for HPLC analysis. To quantify the concentration of glycerol, acetic acid, lactic acid, 3-hydroxypropionaldehyde (3-HPA), 1,3-PDO, and ethanol, samples were analyzed on an organic acid column (HPX-87H; Bio-Rad, Hercules, CA, USA) using an HPLC system equipped with a refractive index detector (RID G1362A, 1100 series; Agilent Technologies, Palo Alto, CA, USA). The operating conditions necessary to optimize peak separation were determined by the method described in the column manual with minor modifications. Filtered culture medium (40 μl) was loaded onto the column and eluted with 5 mM sulfuric acid at a flow rate of 0.6 ml/min at 55°C for 30 min.

Preparation of crude culture extracts.

Lactobacillus panis PM1 and its engineered cells grown in mMRS containing different carbon sources (glycerol or glucose) or LS at pH 4.5 to 7.5 were disrupted by sonication, as described previously (16). Crude culture extract was obtained by centrifugation for 10 min at 16,160 × g, and protein concentration was determined using a protein assay kit (Bio-Rad) with bovine serum albumin (BSA) as a standard.

Enzyme assays.

Acetaldehyde dehydrogenase (ALDH) and alcohol dehydrogenase (ADH) assays were spectrophotometrically evaluated at 37°C by measuring the changes in absorbance at 340 nm (ε^NADH = 6,220 M⁻¹ cm⁻¹). One unit of activity corresponds to the generation or consumption of one micromole of NAD(P)H per min. For ALDH (acetaldehyde to acetic acid) activity, the reaction mixture consisted of 50 mM potassium carbonate (pH 9.0), 5 mM dithiothreitol (DTT), 0.2 mM NADP⁺, and 0.1 mg protein. The mixture was preincubated for 10 min at room temperature, and the reaction was started by the addition of 10 mM acetaldehyde (in methanol). ADH (acetaldehyde to ethanol) was measured in 100 mM potassium carbonate (pH 9.0), 5 mM DDT, 0.2 mM NADH, and 0.1 mg protein. After 10 min of preincubation at room temperature, the reaction was started by addition of 10 mM acetaldehyde (in methanol). Glycerol dehydratase (DhaB) and 1,3-propanediol dehydrogenase (DhaT or YqhD) activities were determined by the 3-HPA quantification method. Briefly, for DhaB (glycerol to 3-HPA), the reaction mixture consisted of 100 mM potassium phosphate (pH 7.0), 10 mM glycerol, 20 μmol/liter coenzyme B₁₂, and 0.1 mg protein (17). The reaction mixture was incubated at 37°C for 20 min, and the enzyme reaction was terminated by the addition of the same volume of 100 mM citrate solution. Quantification of 3-HPA in the reaction solution was carried out using a colorimetric method (18). For DhaT or YqhD (1,3-PDO to 3-HPA), the assay mixture contained 2 mM NAD⁺ (for DhaT) or NADP⁺ (for YqhD), 100 mM 1,3-PDO, 30 mM ammonium sulfate, and 0.1 mg protein in 100 mM potassium carbonate buffer at pH 9.0 (10). The assay was carried out at 37°C for 20 min, and reaction termination and 3-HPA quantification were conducted as described above. In these determinations, one unit of activity corresponds to the production of one millimole of 3-HPA per min. Specific activities of the enzymes were expressed as units per milligram of protein.

Statistical analysis.

For determinations of end product concentrations and enzyme activities, data were presented as mean values calculated from at least two independent experiments. The specific activities of DhaB, DhaT, and YqhD were analyzed by two-way analysis of variance with Bonferroni posttests using GraphPad Prism, version 5.0, software (GraphPad Software, Inc., San Diego, CA, USA). A P value of <0.05 was considered significant.

Nucleotide sequence accession numbers.

The sequence data of NAD(P) transhydrogenase alpha subunit (KJ756519 and KJ756520) and beta subunit (KJ756518) were deposited in GenBank.

RESULTS AND DISCUSSION

Metabolic engineering strategy for a glycerol-oxidative pathway.

The glycerol metabolic pathway of L. panis PM1 is limited to a reductive pathway, and thus glycerol utilization is restricted to NADH reoxidation (for 1,3-PDO production) rather than as an energy source (10). This inability implies that TS contains little readily utilizable carbon for L. panis PM1 growth (11). For more-cost-effective biotechnological applications of TS, a glycerol-oxidative pathway was constructed in L. panis PM1 to enable glycerol to serve as the sole carbon source for energy and growth (Fig. 1 and 2). Escherichia coli possesses two glycerol-oxidative pathways, GlpK/GlpD and glycerol dehydrogenase (DhaD)/dihydroxyacetone kinase (DhaK), but uses mostly the DhaDK route under anaerobic conditions (19); however, the DhaDK route did not work properly in L. panis PM1 under microaerobic conditions (data not shown) since DhaK is a phosphotransferase system-dependent kinase (20) and group III lactobacilli generally tend to lack activity of this system (21). Alternatively, the GlpKD route, which contains an ATP-dependent kinase, was chosen as the backbone of the artificial glycerol-oxidative pathway. This oxidative pathway consisted of GlpF, GlpK, and GlpD, which are required to import glycerol and to oxidize the imported glycerol to dihydroxyacetone phosphate (DHAP). DHAP would then be converted into glyceraldehyde 3-phosphate (GAP) via TPI activity as part of the typical Embden-Meyerhof (EM) pathway. However, we previously showed that the absence of TPI activity in L. panis PM1 resulted in an accumulation of DHAP to the extent that its toxicity influenced cell growth and eventually led to early growth cessation (14). Consequently, the TPI gene from E. coli was cloned into an artificial glycerol-oxidative pathway to prevent DHAP buildup and toxicity. This artificial operon, GlpFKD-TPI, was subcloned into the L. panis PM1 expression vector pCER-EGFP (14). In this plasmid, the expression of the GlpFKD-TPI fragment was controlled under the PorfX promoter that was derived from the sakacin P regulon of Lactobacillus sakei LTH 673 (22) and was developed as an inducible promoter (by sakacin P) for Lactobacillus sakei and Lactobacillus plantarum (23). However, the basal activity of this promoter was quite high in L. sakei and L. plantarum (23), suggesting potential high expression in L. panis PM1 without the need for specific induction. Previously, we reported on the successful expression of the exogenous TPI gene in L. panis PM1 using the same expression system in the absence of its inducer (14).

The resultant plasmid was transformed into the wild-type PM1 strain, and the subsequent recombinant strain (PM1-pCER-GlpFKD-TPI) was cultured in mMRS supplemented with glycerol (24 mM) to test whether glycerol could be used as the sole carbon source (Fig. 3). Strain PM1-pCER-GlpFKD-TPI consumed 10 times more glycerol than the control strain containing the empty plasmid (PM1-pCER) over the first 24 h, indicating that the artificial operon functioned to metabolize glycerol. However, the growth of this engineered strain ceased after 24 h and only 17 mM glycerol was consumed, resulting in a final cell density (OD₆₀₀) of only 0.5 (Fig. 3).

FIG 3.

Growth and glycerol consumption of wild-type and recombinant PM1 strains. Strains PM1-pCER (solid lines) and PM1-pCER-glpFKD-TPI (dashed lines) were cultured in mMRS medium containing 24 mM glycerol. Symbols: circles, cell density (OD₆₀₀); triangles, glycerol concentration.

End product analyses revealed that most of the consumed glycerol was directed toward lactic acid (data not shown) and not 1,3-PDO production, which is a typical NAD⁺ regeneration route in wild-type L. panis PM1 (10), indicating that NADH formed from the new artificial pathway (i.e., by the GlpD reaction) was not recycled sufficiently through the indigenous glycerol reductive pathway. During anaerobic fermentation, the overall redox balance within the cell is a critical determinant governing overall end product profiles (2). How NADH is recycled was likely responsible for the poor growth and low glycerol consumption rates shown in Fig. 3, suggesting the necessity for an additional, or supplemental, NADH recycling route.

Construction of a supplemental NADH recycling system.

The reduced growth and glycerol utilization of PM1-pCER-GlpFKD-TPI suggested that the NADH recycling system did not function as we hypothesized, and therefore a supplemental NADH recycling system appeared necessary to enable TS utilization. Accordingly, a new NADH recycling system was constructed by subcloning the E. coli YqhD gene into the glycerol-oxidative pathway, generating pCER-GlpFKD-TPI-YqhD (Fig. 2). E. coli YqhD is an NADPH-dependent aldehyde reductase, and its activity contributes to a variety of reactions involving butyraldehyde, glyceraldehyde, malondialdehyde, isobutyraldehyde, methylglyoxal, propanealdehyde, acrolein, furfural, glyoxal, 3-HPA, glycolaldehyde, acetaldehyde, and acetol (24). It is expected that YqhD can process acetic acid in the medium into ethanol in conjunction with NADH-dependent ADH, oxidizing NAD(P)H. It would also be beneficial to produce more 1,3-PDO, as we previously reported that the expression of the exogenous YqhD gene in L. panis PM1 could compensate for the reduced DhaT activity caused by 3-HPA accumulation, resulting in higher 1,3-PDO production (11). The final recombinant strain, PM1-pCER-GlpFKD-TPI-YqhD, was cultured in mMRS supplemented with glycerol (160 mM) as the sole carbon source to test how the YqhD gene could contribute to overall NADH recycling and glycerol consumption. Strain PM1-pCER-GlpFKD-TPI-YqhD achieved 4 times greater growth on glycerol than did strain PM1-GlpFKD-TPI, reaching a final cell density (OD₆₀₀) of 2.0.

To detect the activities of both ADH and ALDH, which comprised the new NADH recycling system, strain PM1-pCER-GlpFKD-TPI-YqhD was cultured in mMRS medium supplemented with 160 mM glycerol as the sole carbon source for 24 h. As a reference, strain PM1-pCER (which cannot use glycerol as a carbon source) was cultured in mMRS with 60 mM glucose. The specific activities of the ADH enzyme observed in strains PM1-pCER and PM1-pCER-GlpFKD-TPI-YqhD were 13.59 ± 0.24 and 5.95 ± 0.64 units/mg protein, respectively. No detectable activity of the ALDH enzyme was observed in strain PM1-pCER, whereas 3.19 ± 0.03 units/mg protein was detected for strain PM1-pCER-GlpFKD-TPI-YqhD. This observation indicated that the ALDH activity originated from the exogenous YqhD gene. Obviously, the YqhD system is dependent on the recycling of NAPDH. Hummel (1999) reported the production of (R)-phenylethanol by a new NADPH-dependent ADH coupled with an NADP-dependent formate dehydrogenase in Lactobacillus kefiri (25). The draft genome of L. panis PM1 (M. C. Haakensen, V. Pittet, D. A. S. Grahame, D. R. Korber, and T. Tanaka, unpublished data) revealed a candidate gene for NAD(P) transhydrogenase (EC 1.6.1.2; GenBank accession numbers KJ756519 and KJ756520 [alpha subunit] and KJ756518 [beta subunit]). While the exact NADPH recycling system has not been identified in L. panis PM1, these enzymes (or unidentified dehydrogenases in many anabolic and auxiliary pathways) may contribute to reduction of NADP⁺ to NADPH and thereby assist with the YqhD reaction. Moreover, our previous work has shown that the cloning of YqhD caused no adverse effects in L. panis PM1 (11), indicating that NADPH was regenerated in the unidentified peripheral systems.

With the new NADH recycling system in place, NADH produced from the oxidative glycerol pathway could now be reoxidized via an acetic acid-to-ethanol route by the activities of YqhD and the indigenous ADH, consuming NAD(P)H (Fig. 1). As a result, the presence of acetic acid, a component of the MRS medium (from 5 g/liter sodium acetate), positively affected the cell growth of strain PM1-GlpFKD-TPI-YqhD in the mMRS medium containing glycerol as the sole carbon source. In the presence of acetic acid, strain PM1-pCER-GlpFKD-TPI-YqhD increased glycerol consumption by 75% and lactic acid production by 77% compared to the culture in the absence of acetic acid (Table 3). In the absence of acetic acid, the new recycling system could not function due to substrate limitation, and consequently the glycerol reductive pathway was the main NAD⁺ regeneration route. This shift in NADH recycling could force the recombinant strain to consume glycerol with the production of 1,3-PDO, leading to carbon loss of glycerol for ATP generation via pyruvate production (Fig. 1). The acetic acid-to-ethanol route, on the other hand, could allow most glycerol to be used for lactic acid production in the presence of acetic acid, generating one net ATP per glycerol. Thus, a greater ATP yield in the presence of acetic acid could be a factor for the improved growth. However, the growth and glycerol consumption of strain PM1-GlpFKD-TPI-YqhD halted during the mid-log to late log phase, even though sufficient acetic acid for the new recycling system remained in the culture medium (approximate cell OD₆₀₀, 2.0). Table 4 shows the negative effect of acidic culture conditions on the new NADH recycling system that was introduced as a part of the recombinant plasmid. In pH 6.5 medium, strain PM1-GlpFKD-TPI-YqhD consumed more glycerol (2.9-fold) and acetic acid (6.9-fold) and produced more lactic acid (3.8-fold) and ethanol (7.3-fold) than in the pH 4.5 medium. Lactic acid was a major glycerol metabolism end product in the engineered strain, and its accumulation during fermentation caused the pH of the medium to drop from the initial 6.5 to below 5.0 within 24 h. Thus, growth-related acidification negatively affected the new NADH recycling system, which eventually resulted in the cessation of growth and glycerol consumption during the mid-log to late log phase.

TABLE 3.

Effects of acetic acid on the NADH recycling system^a

Acetic acid content in mediumb	OD600	Glycerol consumed (mM)	End product concn (mM)
			Lactic acid	Acetic acid	Ethanol
50 mM	1.53 ± 0.07	57.24 ± 0.77	46.15 ± 0.38	29.80 ± 0.30c	29.02 ± 0.28
None	0.59 ± 0.01	32.79 ± 0.53	26.05 ± 0.89	6.07 ± 0.08	11.70 ± 0.08

^aValues are means ± standard errors.

^bStrain PM1-pCER-glpFKD-TPI-YqhD was cultured in mMRS medium containing 160 mM glycerol with or without acetic acid for 24 h under pH-uncontrolled conditions.

^cThe initially available acetic acid amount (50 mM) was reduced to 29.80 mM.

TABLE 4.

Effects of pH on the NADH recycling system^a

Initial medium pHb	Glycerol consumed (mM)	End product concn (mM)
		Lactic acid	Acetic acidc	Ethanol
6.5	59.47 ± 0.23	56.19 ± 0.31	22.62 ± 0.46	37.77 ± 1.65
4.5	20.59 ± 0.46	14.98 ± 0.08	46.04 ± 0.56	5.14 ± 0.08

^aValues are means ± standard errors.

^bStrain PM1-pCER-glpFKD-TPI-YqhD was cultured in mMRS medium containing 160 mM glycerol for 24 h under pH-uncontrolled conditions.

^cThe initially available acetic acid amount under each pH condition was 50 mM.

pH-controlled batch culture of the recombinant PM1 in LS.

Since the growth-related acidification negatively affected the engineered NADH recycling system, a pH-controlled batch culture method was adopted to prevent low-pH growth limitation. Considering acetic acid consumption via the new NADH recycling system of strain PM1-GlpFKD-TPI-YqhD (acetic acid-to-ethanol route), 4 g/liter sodium acetate (approximately 40 mM acetic acid) was added to LS. End product formation and glycerol consumption by the recombinant strain were monitored over 4 days. Eighty-three percent of the initially available glycerol (160 mM) in LS was consumed within 72 h during culture under the pH 6.5 condition, producing 85.43 mM lactic acid with a yield of 0.61 mol/mol (Fig. 4A). Over the same period, acetic acid decreased by 20.65 mM, which led to an increase in ethanol (82.16 mM); i.e., ethanol production exceeded the decrease in acetic acid. These data indicated that the acetic acid-to-ethanol route was insufficient for NAD⁺ regeneration during the rapid glycerol consumption period (from 48 h to 72 h). Ethanol production is a typical NAD⁺ regeneration route, and the pyruvate-to-ethanol route (via acetyl coenzyme A) has previously been reported in L. panis PM1 (15). Thus, a part of the pyruvate formed from glycerol could be directed to ethanol production due to the limited capacity of lactate dehydrogenase for NAD⁺ regeneration (Fig. 1) (14). Lactic acid production was severely reduced after 72 h of culture, along with 3-HPA accumulation (26.84 mM). The compound 3-HPA is the sole intermediate of the glycerol reductive pathway, and its toxicity to cell growth has been well-studied in various bacteria (10, 26, 27). It is suggested that 3-HPA accumulation was a major factor hampering high yields during lactic acid fermentation.

FIG 4.

pH-controlled batch fermentation of the engineered PM1 strain in the liquid stillage. The engineered strain PM1-pCER-glpFKD-TPI-YqhD was cultured at pH 6.5 (A) and 7.5 (B) for lactic acid and 1,3-PDO fermentation, respectively. The concentrations of glycerol (white circles), ethanol (black circles), acetic acid (white squares), lactic acid (black squares), 3-HPA (white down-pointing triangles), and 1,3-PDO (white diamonds) are provided on the y axis. Note: the difference in the initial amounts of lactic acid in panels A and B was due to different batches of thin stillage.

In previous results, we mentioned the negative effect of acidic pH on glycerol consumption, whereas mild alkaline conditions were a significant factor in determining the fate of glycerol utilization by the engineered strain. In pH 7.5-controlled fermentation, lactic acid was not produced, and instead, larger amounts of 1,3-PDO were produced than under the pH 6.5 condition, with a concentration of 59.96 mM and a yield of 0.49 mol/mol (this molar ratio was calculated as 1,3-PDO produced over glycerol consumed) (Fig. 4B). In addition, acetic acid (added for NADH recycling) increased from 50.68 to 127.63 mM during fermentation and was accompanied by an increase in ethanol (83.22 mM). The specific activities of DhaB, DhaT, and YqhD (for 3-HPA-to-1,3-PDO conversion) were determined from the engineered strain cultured at pH 6.5 and 7.5 for 48 h to further elucidate how 1,3-PDO production was affected by changes in culture pH. At pH 7.5, the specific activities of DhaT and YqhD were significantly higher than those at pH 6.5 (P < 0.001), whereas DhaB activities under the two pH conditions were comparable (Fig. 5). These results could explain the higher accumulation of 3-HPA, which would further repress DhaT, eventually resulting in little 1,3-PDO production (10) under the pH 6.5 condition. In contrast, under the pH 7.5 condition, YqhD was centrally involved in the conversion of 3-HPA to 1,3-PDO in conjunction with DhaT, and the pyruvate pool formed by the new glycerol-oxidative pathway was directed to acetic acid and ethanol for extra ATP formation and NADH recycling, respectively, suggesting that pH can shift the NAD⁺ regeneration routes (Fig. 1).

FIG 5.

Effects of pH on the glycerol reductive pathway. The specific activities of DhaB, DhaT, and YqhD were measured from 48-h samples obtained during pH-controlled fermentation. White and gray bars represent pH 6.5 and 7.5 conditions, respectively. ***, P < 0.001.

Recently, the search for new technologies capable of producing commodity chemicals from crude glycerol has been renewed due to the rising production of bioethanol and biodiesel (2, 28). For the production of several chemicals, including lactic acid and 1,3-PDO, microbial fermentation has been considered to be a safer alternative, i.e., mild pressure, temperature, and few toxic intermediates, compared with conventional chemical methods (8, 29). Recently, Lactobacillus diolivorans, which belongs to the same lactobacillus group (III) as L. panis PM1, successfully produced a large amount of 1,3-PDO (75 g/liter) in a fed-batch fermentation system (30). In their study, the authors used crude glycerol from biodiesel production and lignocellulosic hydrolysate for 1,3-PDO production to reduce the material costs, instead of using purified glycerol and glucose. However, such a high 1,3-PDO concentration was obtained using expensive MRS-based media, vitamin B₁₂ supplementation, and pretreatment by physicochemical and enzymatic methods for lignocellulosic hydrolysate. TS is the largest global source of waste glycerol and contains complex nutrients, and several studies have focused on the utilization of TS for the production of value-added chemicals, such as 1,3-PDO by L. panis PM1 (11) and butanol by Clostridium pasteurianum DSM 525 (31). Djukic-Vukovic et al. (32) reported that large amounts of lactic acid (42 g/liter) were produced by Lactobacillus rhamnosus immobilized on zeolite, using TS supplemented with 50 g/liter glucose as a medium. However, these attempts required the supplementation of a large amount of fermentable sugars or expensive nutrient sources (e.g., yeast extract).

To the best of our knowledge, there have been no reports showing the production of platform chemicals solely from TS. Our attempts show more-economic conditions in TS utilization than those of other previous studies, suggesting the feasibility of value-added chemical production by L. panis PM1 under simulated industrial conditions. Although TS was a possible culture medium for the engineered PM1 strains, relatively low levels of glycerol (up to 2%) in TS could be a limitation for large amounts of value-added chemical production. The supplementation of additional low-price glycerol to TS-based medium could be a simple solution to this limitation. However, the accumulation of toxic intermediates (e.g., 3-HPA or DHAP) and the instability of the rolling-circle replicon used for the artificial operon expression are other issues that arise during prolonged fermentation. Thus, to achieve our final goal, (i) cloned enzymes may need to be controlled by separate promoters for fine-tuning of their activities, which would reduce the accumulation of the toxic intermediates; (ii) a theta-type replicon suitable for L. panis strain PM1 should be identified for the stability of the artificial operon; and (iii) the inoculum volume and feeding system for additional glycerol (e.g., continuous or fed-batch method) need to be optimized to reduce fermentation time and to maximize productivity. Further studies are being conducted to address these issues in our research group.

In our present study, to reduce the costs of fermentable substrates and nutrient sources, genetic engineering strategies were applied to L. panis strain PM1. An artificial glycerol-oxidative pathway was able to direct glycerol in equimolar amounts to GAP, which was converted to lactic acid by the phosphoketolase pathway, and the exogenous YqhD contributed to direct 1,3-PDO production and the NADH recycling system in the recombinant strain. In addition, ethanol was a major end product regardless of whether lactic acid or 1,3-PDO fermentation by the recombinant PM1 strain occurred, providing a cost-efficient option to recover more ethanol from TS in bioethanol plants. Overall, introduction of an artificial glycerol-oxidative pathway and NADH recycling system enabled the engineered strain to use glycerol as the sole energy source and to produce ethanol, lactic acid, and/or 1,3-PDO from TS. Our findings clearly demonstrate the significant feasibility of direct platform chemical production from TS without the need for expensive fermentable sugar and nutrient source supplementation.