Glycerol and environmental factors: effects on 1,3-propanediol production and NAD+ regeneration in Lactobacillus panis PM1

Authors

  • T.S. Kang,

    1. Department of Food and Bioproduct Sciences, College of Agriculture and Bioresources, University of Saskatchewan, Saskatoon, SK, Canada
    Search for more papers by this author
  • D.R. Korber,

    1. Department of Food and Bioproduct Sciences, College of Agriculture and Bioresources, University of Saskatchewan, Saskatoon, SK, Canada
    Search for more papers by this author
  • T. Tanaka

    Corresponding author
    • Department of Food and Bioproduct Sciences, College of Agriculture and Bioresources, University of Saskatchewan, Saskatoon, SK, Canada
    Search for more papers by this author

Correspondence

Takuji Tanaka, Department of Food and Bioproduct Sciences, College of Agriculture and Bioresources, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK S7N 5A8, Canada. E-mail: Takuji.tanaka@usask.ca

Abstract

Aims

This study was conducted to understand the influences of fermentation factors in NADH recycling and mechanisms of 1,3-propanediol (1,3-PDO) production in Lactobacillus panis PM1.

Methods and Results

We conducted metabolite analyses, qRT-PCR of the glycerol reductive pathway [glycerol dehydratase (DhaB) and 1,3-PDO dehydrogenase (DhaT)] and DhaT activity assays at different pH, temperature and initial glycerol concentrations. The supplementation of 150 mmol l−1 glycerol caused a shift in NADH flux from ethanol to 1,3-PDO production, whereas 300 mol l−1 glycerol negatively affected the regeneration of NAD+ via 1,3-PDO production. This retardation decreased transcription levels and specific activities of DhaT. The decreased DhaT activity eventually caused the shutdown of 1,3-PDO production. Temperature and pH did not significantly affect the specific activity of DhaT, whereas expression of genes for DhaB and DhaT was activated under acidic conditions. Moreover, fresh glucose addition after its depletion could not restart the glycerol reduction, but increased ethanol production.

Conclusions

Those environmental factors affect 1,3-PDO production in different ways through changing the expression level of enzymes and shifting the NAD+ regeneration pathways.

Significance and Impact of the Study

Our findings elucidated a key element to optimize 1,3-PDO production by Lact. panis PM1, which potentially improves 1,3-PDO manufacturing efficiencies.

Introduction

A key feature of lactobacilli metabolism is efficient carbohydrate fermentation coupled with substrate-level phosphorylation through the production of lactate and/or ethanol. To adapt to various environmental conditions, lactobacilli shift their metabolism, which leads to significantly different end product profiles (Axelsson and Lindgren 1987). In homolactic fermentation, NADH formed during early glycolysis is mainly reoxidized by lactate dehydrogenase, thereby producing lactate as the only end product, whereas heterolactic fermentation achieves primary carbohydrate fermentation through substrate-level phosphorylation in the 6-phosphogluconate/phosphoketolase (6-PG/PK) pathway wherein NADH is mainly recycled back to NAD+ through the production of lactate and ethanol (Kandler 1983; Stiles and Holzapfel 1997) (Fig. 1). The difference in the manner by which NADH is recycled is closely related to energy metabolism. The homofermentative pathway generates two moles of ATP per mole of glucose, whereas the heterofermentative pathway yields one ATP per glucose (Poolman 1993; Zaunmuller et al. 2006). Besides this main route of metabolic energy production, heterofermentative lactic acid bacteria (LAB) can use external electron acceptors, such as citrate, fructose and oxygen, to reoxidize NADH in conjunction with the ethanol production pathway, and the phosphoketolase/external electron acceptor pathways can, for example, yield one or two extra ATP per glucose (Zaunmuller et al. 2006).

Figure 1.

Schematic diagram of glycerol and glucose metabolic pathways and a possible erythritol formation route. Glucose metabolism via the 6-phosphogluconate pathway produces two NADH, and the fate of these NADH is finely regulated through the production of various end products, including acetate lactate, ethanol, 1,3-PDO and erythritol. DhaB and DhaT represent glycerol dehydratase and 1,3-propanediol dehydrogenase, respectively.

Lactobacillus panis PM1 belongs to heterofermentative lactobacilli and is of interest due to its ability to produce 1,3-propanediol (1,3-PDO) during anaerobic glycerol fermentation (Khan et al. 2013). Although glycerol cannot be used as a sole carbon source by Lact. panis PM1, glycerol fermentation plays an important role as an NAD+ regeneration route. Lactobacillus panis PM1 ferments glycerol via the reductive pathway where glycerol is first converted to 3-hydroxypropionaldehyde (3-HPA) catalysed by glycerol dehydratase (DhaB), after which 3-HPA is reduced to 1,3-PDO by 1,3-propanediol dehydrogenase (DhaT) (Fig. 1). In this process, DhaT plays a key role in the regeneration of NAD+. However, it has been reported that this enzyme is sensitive to its substrate, 3-HPA (Hao et al. 2008; Celinska 2010). Therefore, the accumulation of this compound triggers decreased DhaT activity, which in turn accelerates further build-up of 3-HPA in Klebsiella pneumoniae, Citrobacter freundii and Enterobacter agglomerans (Barbirato et al. 1997; Celinska 2010).

Various factors affecting the fine regulation of the DhaBT pathway have been reported for Klebsiella, Citrobacter and Lactobacilli (Borch et al. 1991; Biebl et al. 1999; Torino et al. 2001; Hao et al. 2008; Celinska 2010). For example, aeration can influence glycerol metabolism differently in these bacteria. While aerobic conditions are more favoured than anaerobic conditions for Kl. pneumoniae (Hao et al. 2008), Lact. panis PM1 primarily produces 1,3-PDO under anaerobic or microaerobic conditions (Khan et al. 2013). Another influential factor is pH. In acidic media, Kl. pneumoniae produces 1,3-PDO, along with 2,3-butanediol, as characteristic end products. We have also demonstrated that alkaline pH-stimulated Lact. panis PM1 produces acetic acid and 1,3-PDO under specific conditions (Grahame et al. 2013). Carbon source is another crucial factor. Glycerol metabolism is stimulated by the existence of glucose and lactate in resting cells of Lact. reuteri (Luthi-Peng et al. 2002).

In spite of the importance of the glycerol reductive pathway for NADH recycling, the factors affecting this pathway have not yet been fully explored. Therefore, the metabolic profiles of 1,3-PDO and ethanol were examined at different pH, temperature and initial glycerol concentrations to improve our understanding of NADH recycling and mechanisms of 1,3-PDO production in Lact. panis PM1. The observed metabolic shifts were also further examined through qRT-PCR analysis of the glycerol reductive pathway and determination of DhaT activity, demonstrating how Lact. panis PM1 changes its metabolic profiles in response to environmental factors.

Materials and methods

Bacterial strains and growth conditions

Lactobacillus panis PM1 was cultured at 37°C under microaerobic conditions using commercial MRS medium (BD, Franklin Lakes, NJ, USA) until late log phase, and 1% of this preculture (v/v) was used to inoculate fresh modified MRS (mMRS) medium. The mMRS media consisted of 10 g glucose, 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 litre. Cultures were incubated at 37°C under microaerobic conditions, unless otherwise stated. Air-tight 15-ml tubes, filled to the two-thirds level, were incubated under static conditions to establish a microaerobic culture atmosphere. To determine the effects of initial glycerol concentration on 1,3-PDO production and DhaT activity, Lact. panis PM1 was cultured in mMRS containing 0, 150 or 300 mmol l−1 glycerol, and to elucidate the effects of pH and temperature on NAD+ regeneration and DhaT activity, Lact. panis PM1 was cultured at four combinations of temperature (30 or 37°C) and pH (4·5 or 6·5) in mMRS containing 150 mmol l−1 glycerol. For acidic media, the initial pH 6·5 was adjusted to a pH of 4·5 through the addition of 10-N HCl. These pH values were chosen as they reflected the optimal growth condition (pH 4·5) of Lact. panis PM1 (Khan et al. 2013), as well as the pH of the general lactobacilli culture medium, MRS (pH 6·5).

Preparation of crude extracts

For enzyme assays, Lact. panis PM1 cells, grown as described above, were harvested by centrifugation and washed with 100 mmol l−1 phosphate buffer (pH 7·0), and the cells in pellets were disrupted using sonication (three times for 1 min with a 3-min rest interval at output level 2, Sonifier 450; Branson, CT, USA) using the same buffer (Kang et al. 2013). Crude extract was obtained by centrifugation for 10 min at 14 000 g, and protein concentration was determined using the Protein Assay Kit (Bio-Rad, Hercules, CA, USA) with bovine serum albumin (BSA) as a standard.

Enzyme assay

DhaT activity in crude cellular extract was determined at 340 nm (εNADH = 6220 mol l−1 cm−1) (Johnson and Lin 1987). The assay mixture contained 2 mmol l−1 NAD+, 100 mmol l−1 1,3-PDO and 30 mmol l−1 ammonium sulfate in 100 mmol l−1 potassium carbonate buffer at pH 9·0. The assay was carried out at 37°C for 10 min. In these determinations, one unit of activity corresponds to the generation of 1 μmol of NADH per min. Specific activity was expressed as units per milligram of protein.

Determination of glucose and metabolic end products

Culture optical density at 600 nm was measured as an index of growth 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 glucose, organic acids and ethanol, samples were analysed on an organic acid column (HPX-87H; Bio-Rad) using an HPLC system equipped with a refractive index detector (RID G1362A, 1100 series; Agilent Technologies, Palo Alto, CA, USA). Operating conditions were determined by the method described in the column manual with minor modifications. Filtered culture medium (40 μl) was loaded on the column and eluted with 5 mmol l−1 sulfuric acid at a flow rate of 0·6 ml per min at 55°C for 30 min.

Quantification of H2O2 and 3-HPA

Hydrogen peroxide in culture media was determined as described previously by Kang et al. (2013), and 3-HPA quantification was carried out using a colorimetric method (Doleyres et al. 2005).

RNA extraction and reverse transcription

RNA was extracted as described previously by Kang et al. (2013). First-strand cDNA was synthesized using reverse primers of the selected genes (Table 1) according to Kang et al. (2013).

Table 1. Sequences of primers used for qRT-PCR in this study
Target geneFunctionPrimerTm (°C)Nucleotide sequence (5′→3′)
16S rRNA 16S ribosomal RNAf16S58tggcccaactgatatgac
r16S58ctctcatgcacgttcttctt
dhaB Glycerol dehydratasefGDH60tggttggttcacaaactgga
rGDH60ccggcatcgaaatacaactt
dhaT 1,3-PDO dehydrogenasefPDDH60gtgccggcattattttgact
rPDDH60tgtggtcaacggaaacgtaa

Quantitative RT-PCR

Real-time PCR amplification was performed in a CFX96 real-time detection system (Bio-Rad) using the SsoFast EvaGreen Supermix (Bio-Rad). The total volume of the PCR master mixture was 20 μl, to which cDNA template equivalent to 25 ng RNA starting material and 0·5 μmol l−1 of each primer (Table 1) were added. PCR amplification was initiated at 95°C for 30 s followed by 40 cycles of 95°C for 5 s and 60°C for 10 s. Amplification was followed by a melt-curve analysis between 65 and 95°C using a 0·5°C increment. All sample and primer combinations were assessed in three biological replicates with two technical replicates per biological replicate. A no-template control was used for the negative control PCR, and PCR specificity and product detection were verified by examining the temperature-dependent melting curves of the PCR products and ethidium bromide staining on 1% agarose gels. For relative gene expression, the math formula method, using the 16S rRNA gene for normalization, was performed as described by Livak and Schmittgen (2001). The steps for calculating gene expression ratios were as follows:

display math
display math
display math

Normalized expression ratio of dhaB(test) and dhaT(test) = math formula.

The RT-PCR data were processed using CFX Manager Software (Bio-Rad).

Statistical analysis

For determinations of end product concentration and DhaT enzyme activities, data were presented as mean values (±SEM) calculated from at least three independent experiments. Differences of final end products were analysed by the t-test (Mann–Whitney test) for two groups, or the one-way anova test (Kruskal–Wallis test) for three groups, using GraphPad Prism 5.0 software (GraphPad Software, Inc., San Diego, CA, USA). A significance level of < 0·05 was considered significant.

Results

The effects of initial glycerol concentrations on 1,3-PDO production

External glycerol was used as an electron acceptor by Lact. panis PM1, which resulted in the production of 1,3-PDO (Table 2). This reduction reaction also caused a shift in the NAD+ regeneration system. At 48 h, cells grown in 150 mmol l−1 glycerol media produced 87·52 mmol l−1 1,3-PDO along with 41·71 mmol l−1 ethanol that represented the primary NAD+ regeneration system, whereas in the absence of glycerol, 48-h Lact. panis PM1 culture yielded 53·57 mmol l−1 ethanol. Although the growth of Lact. panis PM1 was not affected by up to 8% (870 mmol l−1) of glycerol (Khan et al. 2013), the production of 1,3-PDO was negatively affected by high glycerol (300 mmol l−1) concentration (Table 2). When glycerol was increased from 150 to 300 mmol l−1, the final yield of 1,3-PDO in the 300 mmol l−1 glycerol media was reduced from 87·52 to 64·72 mmol l−1, while ethanol production was increased from 41·71 to 48·08 mmol l−1. This result demonstrated that glycerol fermentation was in competition with ethanol production and thus served as an alternate route for NAD+ regeneration in Lact. panis PM1, in agreement with our previous findings (Khan et al. 2013).

Table 2. The effect of initial glycerol concentrations on the expression of dhaB and dhaT and on the yield of end products
Glycerol (mmol l−1)aRelative gene expression levelMetabolite production (mmol l−1)
dhaB dhaT 1,3-PDOEthanolLactateSuccinate
  1. a

    Lactobacillus panis PM1 was cultured in mMRS media containing 0, 150, 300 mmol l−1 glycerol for 24 h (for qRT-PCR) and 48 h (for HPLC); 0 mmol l−1 glycerol was used for control condition for relative gene expression by qRT-PCR.

01·00 ± 0·391·00 ± 0·3053·57 ± 0·9456·21 ± 3·0110·97 ± 0·02
1501·11 ± 0·411·25 ± 0·3587·52 ± 0·5241·71 ± 0·3758·39 ± 0·2310·91 ± 0·02
3000·79 ± 0·130·17 ± 0·0464·72 ± 0·4648·08 ± 0·6759·91 ± 0·2411·58 ± 0·36

The effects of glucose supplementation on 1,3-PDO production

The reduction in glycerol to 1,3-PDO and ethanol is absolutely dependent on the availability of reducing equivalents from glucose metabolism (Khan et al. 2013). Therefore, the effect of glucose supplementation on NADH flux after glucose depletion was expected to increase both 1,3-PDO and ethanol production. Lactobacillus panis PM1 was cultured in mMRS media containing 55 mmol l−1 glucose and 300 mmol l−1 glycerol at 30°C under microaerobic conditions, and this first-batch culture reached the stationary phase within 48 h. Additional glucose (32 and 24 mmol l−1) was then fed directly into the media after 48- and 96 h culture, respectively, where available glucose had already been depleted. Interestingly, metabolite production profiles indicated that the glucose supplementation after initial glucose depletion (at 48 h) did not increase 1,3-PDO production despite added glucose was used up within 24 h after feeding. After 120 h culture, ethanol and acetate production increased by 42 and 56% of the 48 h culture values to concentrations of 76·06 and 96·62 mmol l−1, respectively. The sum of ethanol and acetate was almost equal (on a molar basis) to the added glucose (total 56 mmol l−1; Fig. 2), that is, one glucose yielded one acetyl phosphate, the intermediate at the fork of metabolic pathways to acetate and ethanol.

Figure 2.

The effect of glucose supplementation on further 1,3-PDO production. Lactobacillus panis PM1 was cultured in mMRS (pH 6·5) containing 55 mmol l−1 glucose and 300 mmol l−1 glycerol at 30°C. Arrows represent additional glucose supplementation (32 mmol l−1 at 48 h and 24 mmol l−1 at 96 h). (image_n/jam12291-gra-0006.png) OD600; (image_n/jam12291-gra-0003.png) glucose; (image_n/jam12291-gra-0001.png) lactate; (▲) glycerol; (image_n/jam12291-gra-0005.png) acetate; (image_n/jam12291-gra-0002.png) 1,3-PDO and (image_n/jam12291-gra-0004.png) ethanol.

The effects of initial pH and culture temperature on 1,3-PDO production and other end products

The pH (4·5 or 6·5) significantly affected both 1,3-PDO and ethanol production during culture at 37°C (< 0·05), where the lower pH condition decreased the production of 1,3-PDO and ethanol by 12% (87·52–77·45 mmol l−1) and 44% (41·71–23·52 mmol l−1), respectively (Table 3). At 37°C, we observed that Lact. panis PM1 growth was almost double of that observed at 30°C under either pH conditions, while the lower temperature (30°C) yielded more 1,3-PDO at each pH (12% more at pH 6·5 and 23% more at pH 4·5) compared with culture at 37°C. The lowest yields of 1,3-PDO and ethanol were observed during the optimal growth condition (37°C and pH 4·5). In our other studies (Kang et al. 2013; Khan et al. 2013), other NAD+ regeneration routes (i.e. pyruvate to lactate, citrate to succinate and molecular oxygen to hydrogen peroxide) were reported to maintain redox balance in Lact. panis PM1. The range of lactate yield was comparable among all conditions (from 52·11 to 58·39 mmol l−1), indicating that the NADH pool should be similar in size among the tested conditions [glucose (55 mmol l−1) was used to produce similar amounts of lactate with consumption of NADH, Table 3]. Significant differences in hydrogen peroxide accumulation were not observed in any of the samples (from 6·2 to 4·1 μmol l−1, > 0·05, Table 3). However, pH and temperature affected succinate and erythritol pathways. The production of succinate decreased under pH 4·5 conditions, whereas the final yields of erythritol significantly increased under acidic conditions (< 0·05) (Table 3).

Table 3. The effect of initial pH and culture temperature on the expression of dhaB and dhaT and on the yield of end products
ConditionsaRelative gene expression levelbSpecific Activityb,cMetabolite productionb (mmol l−1 or μmol l−1)d
dhaB dhaT DhaT1,3-PDOEthanolLactateSuccinateErythritolH2O2
  1. a

    Lactobacillus panis PM1 was cultured in mMRS media containing 150 mmol l−1 glycerol at each pH and temperature condition.

  2. b

    Samples for qRT-PCR, enzyme activity and HPLC were withdrawn from the 37°C cultures after 24 h and from the 30°C cultures after 48 h.

  3. c

    μmol min−1 mg−1.

  4. d

    Unit of H2O2 is μmol l−1 and the others are mmol l−1.

  5. e

    Control condition for relative gene expression by qRT-PCR.

pH 6·5/37°Ce1·00 ± 0·101·00 ± 0·5416·88 ± 0·4087·52 ± 0·5241·71 ± 0·3758·39 ± 0·2311·38 ± 0·021·04 ± 0·015·8 ± 0·8
pH 6·5/30°C3·07 ± 0·101·68 ± 0·1015·00 ± 0·5399·54 ± 0·5340·10 ± 1·5852·11 ± 0·0611·49 ± 0·080·99 ± 0·024·1 ± 0·6
pH 4·5/37°C15·33 ± 1·7416·10 ± 0·1017·51 ± 0·1177·45 ± 0·6823·52 ± 0·4655·41 ± 0·2710·31 ± 0·041·23 ± 0·036·2 ± 0·6
pH 4·5/30°C8·49 ± 0·968·28 ± 0·5415·43 ± 0·59100·78 ± 0·4834·80 ± 0·6055·93 ± 0·3110·15 ± 0·051·56 ± 0·064·6 ± 1·0

Transcription levels and enzyme activities of 1,3-PDO pathway

To further evaluate the effects of pH, temperature and glycerol concentration on the regeneration of NAD+ through the glycerol-1,3-PDO reductive pathway, candidate genes for DhaB and DhaT were identified through analysis of the annotated Lact. panis whole-genome sequence data (M.C. Haakensen, V. Pittet, D.A.S. Grahame, D.R. Korber and T. Tanaka, unpublished data). The expression levels of these two genes were examined by qRT-PCR under the conditions examined for metabolite analyses presented in the previous sections, that is, different pH (4·5 and 6·5), temperature (30 and 37°C) and glycerol concentrations (0, 150 and 300 mmol l−1). The different glycerol concentrations resulted in similar expression levels of dhaB under all conditions. The expression of dhaT occurred at similar levels under both control (0 mmol l−1 glycerol) and 150 mmol l−1 glycerol conditions; however, its expression in the cultures amended with 300 mmol l−1 glycerol was one-sixth that of the control condition (Table 2). In addition, these transcriptional data were well correlated with DhaT enzyme assay results at 24 h (Fig. 3). Specific activities of this enzyme were comparable among all conditions until 12-h culture; however, DhaT activity under 300 mmol l−1 glycerol did not increase at 24 h, in contrast to the other two assay conditions (0 and 150 mmol l−1 glycerol). While DhaT activities varied, the yields of 1,3-PDO and ethanol at qRT-PCR sampling time (24 h culture) were comparable among all conditions (1,3-PDO: 0, 19·39 and 19·69 mmol l−1; and ethanol: 22·45, 21·78 and 19·35 mmol l−1 in 0, 150 and 300 mmol l−1 glycerol samples, respectively). In contrast to glycerol concentration, pH significantly influenced the expression of the two genes, as shown in Table 3. Under the optimal growth condition (37°C and pH 4·5), 15-fold higher expression of dhaB and dhaT was observed relative to the control condition (37°C and pH 6·5). However, the specific activity of DhaT was not induced by the acidic conditions and was thus comparable among all conditions (Table 3), indicating high glycerol concentration to be the main controlling factor of DhaT activity.

Figure 3.

The effect of glycerol concentration on DhaT activity. Lactobacillus panis PM1 was cultured in mMRS (pH 6·5) containing 55 mmol l−1 glucose and 0 (circles), 150 (squares) or 300 (triangles) mmol l−1 glycerol, respectively. DhaT activity was measured after 6, 12 and 24 h of culture.

Discussion

Only a few lactobacillus strains, including Lact. panis PM1, Lact. buchneri, Lact. reuteri, Lact. hilgardii and Lact. dilovorans, have been reported to be 1,3-PDO producers, and the reductive pathway in those heterofermentative strains plays a key role in the production of 1,3-PDO from glycerol (Biebl et al. 1999; Saxena et al. 2009; Celinska 2010; Khan et al. 2013). Our earlier studies proposed that the conversion of glycerol to 1,3-PDO was an important auxiliary pathway to maintain the redox balance in Lact. panis PM1 (Khan et al. 2013). However, more detailed information of the factors influencing NADH recycling during end product formation is necessary in order that Lact. panis PM1 be used for glycerol conversion as part of engineered biofuel waste management applications.

As Lact. panis PM1 belongs to the group III LAB (Khan et al. 2013), NADH recycling under anaerobic conditions depends on the production of lactic acid and ethanol through the 6-PG/PK pathway (Veiga-da-Cunha and Foster 1992; Luthi-Peng et al. 2002; Pedersen et al. 2004). Our HPLC analyses showed that the presence of an external source of glycerol shifted NAD+ regeneration from ethanol production to the reduction of glycerol to 1,3-PDO (Table 2), decreasing ethanol production by 22% in the presence of 150 mmol l−1 glycerol. In other words, the decrease in ethanol production provided surplus NADH that was then utilized for the production of 1,3-PDO. In our previous study, we reported that extremely high glycerol concentrations actually decreased the conversion of glycerol to 1,3-PDO and that the ratio of glucose to glycerol was a key factor for optimal 1,3-PDO production (Grahame et al. 2013). Our present results agree with this decrease, as providing higher glycerol concentration (300 mmol l−1) altered the route of NADH recycling, consequently reducing the yield of 1,3-PDO by 26% and increasing ethanol production by 15% compared with the production levels seen at 150 mmol l−1 glycerol. This result demonstrated that both ethanol and 1,3-PDO production are in competition for the same NADH pool and that at a high glycerol concentration (300 mmol l−1), part of the NADH pool was redirected to the production of ethanol. This observation also confirmed that the conversion of glycerol to 1,3-PDO acted as an auxiliary pathway for the NAD+ regeneration system.

In the presence of glycerol under microaerobic conditions, NADH from initial glucose metabolism was directed to 1,3-PDO and ethanol production. However, the availability of NADH was not a determining factor in shifting between ethanol and 1,3-PDO. We observed that glucose depletion caused a shutdown of the reductive pathway for 1,3-PDO production, and NADH flux was directed to ethanol production after providing additional glucose, that is, 1,3-PDO production did not resume (Fig. 2). The yield of 1,3-PDO increased until initial glucose depletion at 48 h (20 mmol l−1 at 24 h and 160 mmol l−1 at 48 h), and total glycerol consumption at each time point was higher than these amounts (32 mmol l−1 at 24 h and 205 mmol l−1 at 48 h). Similar results were observed in our previous research where this gap (between glycerol consumed and 1,3-PDO produced) increased at high glycerol concentration (52 and 24 mmol l−1 in 326 and 163 mmol l−1 glycerol media, respectively) along with a reduced 1,3-PDO conversion rate (65 and 81% in 326 and 163 mmol l−1 glycerol media, respectively) (Khan et al. 2013). These observations are indicative of the accumulation of the intermediate 3-HPA (the only intermediate in the reducing pathway).

Our qRT-PCR results revealed that the expression of dhaB and dhaT was mainly activated under acidic conditions, whereas dhaT expression was repressed at the high glycerol concentration (300 mmol l−1) (Tables 2 and 3). These data matched the result of a low final 1,3-PDO yield in the presence of high concentrations of glycerol. At 24 h, glucose consumption and ethanol and 1,3-PDO production values were similar at all glycerol concentrations (0, 150 and 300 mmol l−1; except 0 mmol l−1 glycerol, where 1,3-PDO was not observed). As shown in Fig. 3, the activities of the DhaT enzyme were similar at each glycerol concentration until 12 h of culture. However, at 24 h, the transcription level and specific activity of this enzyme in mMRS medium supplemented with 300 mmol l−1 glycerol were reduced by 83 and 50%, respectively, from than observed in mMRS alone (Table 2 and Fig. 3). This observation clearly indicated that decreased DhaT activity, which was negatively affected by the progress of growth, was a main reason for reduced 1,3-PDO production in the presence of 300 mmol l−1 glycerol compared with 150 mmol l−1 glycerol. Also, considering that significant amounts of 3-HPA were not detected by our quantification method from media of all glycerol concentrations (data not shown), the accumulation of 3-HPA inside of the cells offers a logical explanation for the decreased transcription level and activity of DhaT enzyme, which eventually negatively impacted 1,3-PDO production and necessitated the redirection in NADH recycling.

However, total redox balance cannot be completely explained solely with the above 1,3-PDO and ethanol producing routes. The total amounts of 1,3-PDO and ethanol production were similar among the treatments pH 6·5–37°C, pH 6·5–30°C and pH 4·5–30°C, ranging from 129·23 to 135·58 mmol l−1, whereas, under the optimal condition (pH 4·5–37°C), less 1,3-PDO and ethanol production was observed (100·97 mmol l−1), indicating a shift of NADH flow through ethanol and 1,3-PDO production and leaving some NADH untouched under the optimal condition. However, the yield of lactate and hydrogen peroxide were comparable among all conditions, and succinate production decreased during culture under pH 4·5 conditions (Table 3), suggesting those pathways did not reoxidize the remaining NADH resulting from decreased ethanol or 1,3-PDO production. These results suggest the existence of an alternative, unexplored NADH recycling route.

Temperature and pH were also significant factors affecting the yield of 1,3-PDO. Although the lowest 1,3-PDO yield was observed when the optimal growth condition was employed, the expression of the dhaB and dhaT was found to be highest. Higher 1,3-PDO production was observed at 30°C cultures compared with 37°C cultures, showing no significant changes in transcription levels of those two genes. Furthermore, the tested environmental factors did not significantly affect the specific activity of the DhaT enzyme (Table 3). It therefore appears that the decreased production of 1,3-PDO under the optimal condition was not associated with reduced activity of DhaT enzyme, in contrast to the high glycerol (300 mmol l−1) situation. The different growth rates seen at 37 and 30°C do offer a clue. Generally, the activities of NADH-generating enzymes from the 6-PG/PK pathway (i.e. glucose-6-phosphate and 6-phosphogluconate dehydrogenases) largely exceed the activities of the NADH-reoxidizing enzymes from the ethanol pathway (i.e. acetaldehyde and alcohol dehydrogenases) in typical glucose fermentation by heterofermentative LAB (Veiga-da-Cunha et al. 1993; Richter et al. 2001, 2003; Zaunmuller et al. 2006). Therefore, at the high growth rate observed under the optimal condition, an excess NADH pool formed by rapid glucose metabolism and limited capacities of the ethanol and 1,3-PDO pathways to process this excess should force a partial shift of NADH reoxidation towards other pathways, even though the transcription level of the glycerol reductive pathway was up-regulated.

The erythritol pathway has been observed in various heterofermentative LAB as a main alternate route for the limited NADH reoxidation of ethanol pathway during rapid growth (Veiga-da-Cunha et al. 1993; Stolz et al. 1995; Richter et al. 2001). Although Lact. panis PM1 could not ferment erythritol as a sole carbon source (Khan et al. 2013), it was detected as a minor end product (Table 3). The final yields of erythritol increased under acidic conditions, unlike the production of ethanol or succinate, suggesting a possible role of erythritol production as a minor NADH-reoxidizing route in Lact. panis PM1. Different routes of NAD+ regeneration have been reported for many end-product-yielding pathways (e.g. mannitol, erythritol and glycerol), and multiple combinations of those pathways, in conjunction with the typical ethanol pathway, are used by heterofermentative LAB according to their environmental circumstances (Richter et al. 2003). Therefore, these multiple combinations, including the erythritol pathway, could be used to complement the decreased NAD+ regeneration capacity seen during ethanol and 1,3-PDO production in Lact. panis PM1 grown under the optimal growth condition.

Acknowledgements

This study was supported by the Saskatchewan Agriculture Development Fund and Agricultural Bioproducts Innovation Program of Agriculture and Agri-Food Canada. We thank Drs. Low and Qiu of the University of Saskatchewan for providing access to HPLC and qRT-PCR instrumentation, respectively.

Conflict of Interest

The authors declare no conflict of interest.

Ancillary