Introduction

Large amounts of raw glycerol are formed as the main by-product in biodiesel production and constitutes approx. 10% (w/w) of the total biodiesel generated (Johnson and Taconi 2007). This surplus raw glycerol has not only greatly disturbed the market for traditional glycerol in terms of price but has also created a significant environmental problem because it cannot be discharged directly into the environment without treatment (da Silva et al. 2009). Thus, considerable efforts have been directed towards development of methods to refine glycerol from a low-cost feedstock into industrially valuable materials including fuels, building blocks, and bioactive substances.

Glycerol is a non-fermentable by most microorganisms except for a group of bacteria including Bacillus, Clostridium, Enterobacter, Klebsiella and Lactobacillus species. Recently, Gonzalez and colleagues demonstrated anaerobic fermentation of glycerol by Escherichia coli and thus offered a new metabolic platform for production of fuels and chemicals (Dharmadi et al. 2006; Gonzalez et al. 2008; Murarka et al. 2008). Moreover, the cost of ethanol production from glycerol was almost 40% lower compared with production from corn-derived sugars in terms of feedstock demand and operational costs (Yazdani and Gonzalez 2007). These observations encouraged the engineering of E. coli for the efficient fermentative conversion of glycerol to ethanol and co-products (Yazdani and Gonzalez 2008; Durnin et al. 2009).

The methylotrophic yeast, Hansenula polymorpha, is one of the most important non-conventional yeasts used in industrial processes such as heterologous protein production (Gellissen 2002). In addition, H. polymorpha has the potential to be useful in ethanol production owing to its ability to ferment xylose, which is derived from the hemicellulose component of biomasses (Ryabova et al. 2004; Guerra et al. 2005). In the present study, we evaluated the potential of H. polymorpha as a metabolic platform for conversion of glycerol to ethanol.

Materials and methods

Strains, plasmids, and media

Escherichia coli DH5α was used as the host strain for genetic cloning. Zymomonas mobilis ZM4 was the source of genes encoding pyruvate decarboxylase (pdc) and aldehyde dehydrogenase II (aldB). Klebsiella pneumoniae MGH78578 was the source of genes encoding glycerol dehydrogenase (dhaD) and dehydroxyacetone kinase (dhaKLM). The methylotrophic yeast Hansenula polymorpha DL1-L was used for production of ethanol from glycerol. pGEM-T-Easy (Promega) was employed for cloning and sequencing of amplified DNA fragments. pYHSA161 (Heo et al. 2003) was used to express genes in H. polymorpha. E. coli and H. polymorpha were grown in LB (yeast extract, 0.5%; Bacto-tryptone, 1%; and NaCl, 1%; all w/v) and YPD medium (yeast extract, 1%; peptone, 2%; glucose, 2%; all w/v) supplemented with leucine (5 g l−1), respectively. Ampicillin (50 μg ml−1) and G418 (0.5–2 mg ml−1) were used for selection of transformants of E. coli and H. polymorpha, respectively.

Construction of plasmids pYH-pdc-adhB, pYH-dhaDKLM, and pYH-pdc-adhB-dhaDKLM

A schematic representation of the strategy used to construct pYH-pdc-adhB, pYH-dhaDKLM, and pYH-pdc-adhB-dhaDKLM is shown in Fig. 1. The primers used in amplification of DNA sequences by PCR are listed in Supplementary Table 1. The 0.8 and 0.3 kb DNA sequences, including the promoter (pGAP) upstream of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the terminator sequence (tT) of alcohol oxidase (AOX) of H. polymorpha, were amplified using pYHSA161 (Heo et al. 2003) as a template. Amplified DNA sequences were cloned into pGEM-T-Easy vector generating pGEM-pGAP-tT, followed by DNA sequencing to confirm the absence of any errors in the DNA polymerization. The 1.8 and 1.15 kb ORF of pdc and adhB were amplified from Z. mobilis chromosomal DNA using the primers as indicated in Supplementary Table 1. DNA sequences encoding dhaD (1.1 kb), dhaK (1.1 kb), dhaL (0.6 kb), and dhaM (1.4 kb) were amplified from K. pneumoniae chromosomal DNA using primers listed in Supplementary Table 1. Amplified DNA fragments were cloned into appropriate restriction sites of pGEM-pGAP-tT, followed by DNA sequencing to confirm the absence of any errors. To generate pYH-pdc-adhB, the NotI and SpeI–NheI fragments, including pdc (pGEM-pdc) and adhB (pGEM-adhB) were serially inserted into the NotI and SpeI sites of pYHSA161, respectively. Plasmids pYH-dhaDKLM and pYH-pdc-adhB-dhaDKLM were also similarly constructed as shown in Fig. 1.

Fig. 1
figure 1

Schematic representation of the strategy used for construction of pYH-pdc-adhB and pYH-pdc-adhB-dhaDKLM. Abbreviations for restriction enzymes: B: BamHI, Bg: BglII, Nh: NheI, No: NotI, S: SmaI, Sp: SpeI, X: XhoI, Xb: XbaI. Kle is the Klenow fragment

Full size image

Screening and cultivation of H. polymorpha transformants expressing pdc and adhII genes

Hansenula polymorpha was transformed using the lithium acetate and DMSO method (Hill et al. 1991), and the resulting Leu+ transformants harboring pYH-pdc-adhII or pYH-pdc-adhII-dhaDKLM were stabilized as previously described (Sohn 1997).

Hansenula polymorpha strains were cultivated in 250 ml flasks with 50 ml YP medium, including glucose or glycerol as carbon source. Growth of the cells was monitored from the OD600 values. Cells were subjected to enzyme activity assays and the culture broth was analyzed for relevant metabolites. All results presented reflect data from three independent experiments.

Enzyme activity assays

The activities of pyruvate decarboxylase (Pdc) and aldehyde dehydrogenase (AdhII) were measured using the method of Postma et al. (1989), whereas the activity of glycerol dehydrogenase (DhaD) was estimated by the method of Slininger et al. (1983). One unit of enzyme activity was the amount of enzyme consuming 1 μmol substrate per min. Protein was determined using a protein assay kit (Bio-Rad), with BSA as a standard.

Metabolite analysis

The levels of residual glycerol and ethanol in culture broth were determined by HPLC equipped with a refractive index detector and an organic acid analysis column (300 × 78 mm; Aminex HPX-87H; Bio-Rad) held at 65°C. The mobile phase was 5 mM H2SO4 at 0.8 ml min−1.

Results

Construction of a recombinant H. polymorpha strain transformed with pdc and adhB genes from Z. mobilis

Ethanol is produced in H. polymorpha from the reduction of pyruvate by pyruvate decarboxylase and alcohol dehydrogenase. Overexpression of the pyruvate decarboxylase gene (PDC1) from H. polymorpha or Kluyveromyces lactis resulted in increased production of ethanol in H. polymorpha using xylose as substrate (Ishcuhk et al. 2008). We investigated whether overexpression of genes encoding proteins involved in the ethanol production could also result in an increase in ethanol yield from glycerol. To this end, the recombinant plasmid pYH-pdc-adhB (Fig. 1) was prepared, in which expression of the genes pdc and adhB, encoding pyruvate decarboxylase and aldehyde dehydrogenase II from Z. mobilis, respectively, was driven by the constitutive GAPDH promoter (pGAP) and terminated by the AOX terminator (tAOX). Plasmid pYH-pdc-adhB was linearized and successfully transformed into H. polymorpha DL1-L (Leu−). Transformants were selected by leucine prototrophy (Leu+) on minimal medium without leucine, and stabilized by serial cultivation in such medium. To induce possible multiple integration of the relevant genes into chromosomal DNA, stabilized transformants were grown on minimal medium, without leucine, but supplemented with G418 at either 0.5, 1, or 2 mg ml−1. The resulting recombinant yeast strain was termed Hp DL1-L/pYH-pdc-adhB. The presence of the required inserts in the genome of the recombinant strain was confirmed by PCR (Fig. 2).

Fig. 2
figure 2

Confirmation of chromosomal integration of adhB (lanes 1, 7, and 13; 1.15 kb), pdc (lanes 2, 8, and 14; 1.8 kb), dhaD (lanes 3, 9, and 15; 1.1 kb), dhaK (lanes 4, 10, and 16; 1.1 kb), dhaL (lanes 5, 11, and 17; 0.6 kb) and dhaD (lanes 6, 12, and 18; 1.4 kb) genes in recombinant H. polymorpha strains, as shown by PCR amplification. Lanes 1–6, HpDL1-L/pYH-pdc-adhB; lanes 7–12, HpDL1-L/pYH-pdc-adhB-dhaDKLM; lanes 13–18, HpDL1-L. M: molecular size marker

Full size image

Enhanced ethanol yield from glycerol by action of the recombinant H. polymorpha strain Hp DL1-L/pYH-pdc-adhB

The recombinant yeast grew similarly to the control strain with glycerol at 2% (v/v) as sole carbon source (Fig. 3). The ethanol yield was 3.3-fold higher (2.74 g l−1) than the parent strain (0.83 g l−1), although ethanol production was still low. Consistent with these results, the activities of Pdc and AdhII were 121 and 170% higher, respectively, in Hp DL1-L/pYH-pdc-adhB than in Hp DL1-L (Table 1).

Fig. 3
figure 3

Cell growth (closed symbols) and ethanol production (open symbols) of recombinant H. polymorpha strains. HpDL1-L, circles; HpDL1-L/pYH-pdc-adhB, squares; HpDL1-L/pYH-pdc-adhB-dhaDKLM, triangles

Full size image
Table 1 Activities (enzyme units) of Pdc, AdhII, and DhaD in recombinant H. polymorpha strains
Full size table

Effect of overexpression of the dhaD and dhaKLM genes of Klebsiella pneumoniae on glycerol utilization and ethanol yield

In K. pneumoniae, anaerobic utilization of glycerol is catalyzed by glycerol dehydrogenase (DhaD) and dehydroxyacetone kinase (DhaKLM). To increase ethanol yield from glycerol, dhaD and dhaKLM were overexpressed in the recombinant yeast strain. The recombinant plasmid pYH-pdc-adhB-dhaDKLM was prepared by inserting dhaDKLM between the constitutive GAPDH promoter (pGAP) and the AOX terminator (tAOX), together with pdc and adhB (Fig. 1), and transformed into H. polymorpha. A transformant was selected and stabilized as described above.

Enzyme analysis showed that DhaD activity was 54 and 44% higher in Hp DL1-L/pYH-pdc-adhB-dhaDKLM than in Hp DL1-L and Hp DL1-L/pYH-pdc-adhB, respectively (Table 1). Glycerol utilization by these strains reflected the respective levels of enzyme activity (data not shown). Also, ethanol yield from glycerol was slightly increased (to 3.1 g l−1) after expression of dhaDKLM in the recombinant yeast strain (Fig. 3).

Discussion

Ethanol production from glycerol occurs in some microbial strains, such as Klebsiella and Enterobacter species, in which the cellular redox balance during the oxidative process of glycerol was retained by a coupled reductive process producing 1,3-propanediol. The highest yield of ethanol from glycerol (17 g l−1) was reported during fed-batch fermentation of K. pneumoniae (Cheng et al. 2007). Recently, fermentative utilization of glycerol has been reported in E. coli in which glycerol is usually assimilated via an aerobic metabolic pathway (Durnin et al. 2009). The production of ethanol from glycerol attained a maximum level of approximately 19 g l−1 in an engineered E. coli strain. In the present study, we constructed a recombinant H. polymorpha yeast strain expressing the pdc and adhB genes from Z. mobilis. This recombinant strain demonstrated a 3.3-fold increase in ethanol yield, from glycerol, in a flask experiment (final concentration of ethanol: 2.74 g l−1).

Glycerol is dissimilated by distinct metabolic pathways in microorganisms. The relevant respiratory pathways include that involving the action of glycerol kinase (GlpK) and glyceraldehyde-3-phosphate dehydrogenase (GlpD/GlpABC) and the fermentative pathway catalyzed by glycerol dehydrogenase (DhaD) and dehydroxyacetone kinase (DhaKLM). Durnin et al. (2009) reported efficient stimulation of ethanol production from glycerol in E. coli by overexpression of genes involved in the fermentative pathway. Introduction of dhaD and dhaKLM from K. pneumoniae also resulted in an increase in glycerol utilization and ethanol yield in the recombinant yeast strain.

Although production of ethanol from glycerol increased in H. polymorpha after appropriate genetic engineering, the present yield is lower than that of bacteria such as K. pneumoniae and E. coli. However, we expect that a recombinant yeast strain will be useful in the development of a industrial process for ethanol production by metabolic and process engineering because a critical advantage of yeast compared with bacteria, is high-level ethanol tolerance.