Applied Microbiology and Biotechnology

, Volume 73, Issue 5, pp 1017–1024

Fermentation of 1,3-propanediol by a lactate deficient mutant of Klebsiella oxytoca under microaerobic conditions

Authors

  • Guang Yang
    • State Key Laboratory for AgrobiotechnologyChina Agricultural University
    • Department of Microbiology and Immunology, College of Biological SciencesChina Agricultural University
  • Jiesheng Tian
    • State Key Laboratory for AgrobiotechnologyChina Agricultural University
    • Department of Microbiology and Immunology, College of Biological SciencesChina Agricultural University
    • State Key Laboratory for AgrobiotechnologyChina Agricultural University
    • Department of Microbiology and Immunology, College of Biological SciencesChina Agricultural University
Biotechnological Products and Process Engineering

DOI: 10.1007/s00253-006-0563-7

Cite this article as:
Yang, G., Tian, J. & Li, J. Appl Microbiol Biotechnol (2007) 73: 1017. doi:10.1007/s00253-006-0563-7
  • 546 Views

Abstract

Klebsiella oxytoca M5al is an excellent 1,3-propanediol (1,3-PD) producer, but too much lactic acid yielded greatly lessened the fermentation efficiency for 1,3-PD. To counteract the disadvantage, four lactate deficient mutants were obtained by knocking out the ldhA gene of lactate dehydrogenase (LDH) of K. oxytoca M5al. The LDH activities of the four mutants were from 3.85 to 6.92% of the parental strain. The fed-batch fermentation of 1,3-PD by mutant LDH3, whose LDH activity is the lowest, was studied. The results showed that higher 1,3-PD concentration, productivity, and molar conversion rate from glycerol to 1,3-PD can be gained than those of the wild type strain and no lactic acid is produced under both anaerobic and microaerobic conditions. Sucrose fed during the fermentation increased the conversion and sucrose added at the beginning increased the productivity. In fed-batch fermentation with sucrose as cosubstrate under microaerobic conditions, the 1,3-PD concentration, conversion, and productivity were improved significantly to 83.56 g l−1, 0.62 mol mol−1, and 1.61 g l−1 h−1, respectively. Furthermore, 60.11 g l−1 2,3-butanediol was also formed as major byproduct in the broth.

Introduction

As an important chemical intermediate, 1,3-PD was used as a monomer to produce polyesters, polyethers, and polyurethanes. Until now, 1,3-PD has been produced by chemical synthesis, either by the hydratation of acrolein or by the hydroformylation of ethylene oxide (Biebl et al. 1999; Zeng and Biebl 2002). Compared with chemical synthesis, microbial fermentation processes for 1,3-PD have many obvious advantages in that they use renewable resources and reduce pollution to a great extent. The 1,3-PD producers are all bacteria including enterobacteria, such as Klebsiella pneumoniae, Citrobacter freundii, and Enterbacter agglomerans; clostridia, such as Clostridium butylicum and Clostridium pasteurianum, and lactobacter, such as Lactobacillus brevis and Lactobacillus buchneri, etc. (Biebl et al. 1999; Homann et al. 1990; Zeng and Biebl 2002). The production of 1,3-PD is generally performed under anaerobic conditions using glycerol as the sole carbon source. In the metabolic reactions, glycerol is dissimilated through coupled oxidative and reductive pathways (Ahrens et al. 1998; Biebl et al. 1999; Zeng and Biebl 2002; Zeng et al. 1993). The reductive branch consists of two steps: glycerol is first dehydrated to 3-hydroxypropionaldehyde (3-HPA) and then 3-HPA is reduced to 1,3-PD under the consumption of reducing power (NADH). The reducing power and various byproducts were produced in the oxidative branch. In the actual fermentation process, a number of byproducts can be found (Biebl 2001; Biebl et al. 1998, 1999; Solomon et al. 1995; Zeng 1996; Zeng et al. 1993). The inhibition potentials of substrate and products on cell growth will impair the 1,3-propanediol production and then make the product recovery and purification a troublesome task. Multiple inhibition by glycerol and products to K. pneumoniae and C. butylicum had been studied in several investigations (Biebl 1991; Cameron et al. 1998; Cheng et al. 2005; Colin et al. 2000, 2001; Zeng et al. 1994). The organic acids and ethanol were more toxic than glycerol and 1,3-PD to both bacteria, and among the organic acids, acetic acid was most toxic to K. pneumoniae.

Many efforts have been put into the 1,3-PD fermentation processes to increase the glycerol conversion and final 1,3-PD concentration. New strains selection (Biebl et al. 1992; Papanikolaou et al. 2000; Petitdemange et al. 1995), different culture techniques applications (Biebl 2001; Boenigk et al. 1993; Günel et al. 1991; Menzel et al. 1997; Reimann and Biebl 1996; Reimann et al. 1998; Papanikolaou et al. 2000), and cofermentation glycerol with sugar (Abbad-Andaloussi et al. 1998; Biebl and Marten 1995; Saint-Amans and Soucaille 1995; Saint-Amans et al. 2001; Tong and Cameron 1992) have been practiced and achieved some improvement in a sense. Recently, fermentations under microaerobic conditions were studied (Huang et al. 2002; Chen et al. 2003; Cheng et al. 2004) and an anaerobic/aerobic combined culture was developed to improve the 1,3-PD concentration (Cheng et al. 2004). Although some progress had been made in these studies, the 1,3-PD concentration of 55–73 g l−1 was still unlikely for bulk production. Some recombinant strains of Escherichia coli have been constructed, but they can only produced 1,3-PD with very low levels from glycerol (Cameron et al. 1998; Skraly et al. 1998; Sprenger et al. 1989; Tong and Cameron 1992; Tong et al. 1991; Zhu et al. 2002). Methods of metabolic engineering and recombinant DNA technology are also used to bring out new organisms utilizing cheaper and abundant substrates such as sugar and starch (Biebl et al. 1999; Cameron et al. 1998; Nakamura and Whited 2003; Zeng and Biebl 2002). Unfortunately, no desirable results were gained until Dupont and Genencor described in a patent that a recombinant E. coli reached a final 1,3-PD concentration of 135 g l−1 using glucose as substrate (Nakamura and Whited 2003). Although it was an exciting outcome, to add vitamin B12 in the fermentation process became a new restriction to discourage it into large-scale production of 1,3-PD.

In view of the results not so desirable for 1,3-PD industrial scale production mentioned above, maybe to block the pathways leading to the byproducts of no benefit to 1,3-PD production is a simple and effective method to improve the glycerol conversion and final 1,3-PD concentration. The purpose of this work was to construct lactate-deficient mutants to make an intrinsic improvement to 1,3-PD production. In addition, the effect of sucrose as cosubstrate under microaerobic conditions was also studied to boost 1,3-PD production from extrinsic aspects.

Materials and methods

Strains and media

Klebsiella oxytoca M5al was kindly supplied by Dr. Burris, Department of Bacteria, University of Wisconsin, Madison. It was once designated as Aerobacter aerogenes or K. pneumoniae (Ohta et al. 1991). And it was used in this work not only for it was an excellent 1,3-PD producer, but also for it was free of capsular and nonpathogenic and so can be used safely. The LDH mutants were ldhA deficient by inserting the vector pGPCm (Zhao and Li 2004) into the ldhA gene of K. oxytoca M5al (DQ438981).

The media were same as Günel et al. (1991) used except the deletion of CaCl2·2H2O and CaCO3 and the glycerol concentration in fermentation media. The preculture media contained (per liter): 3.4 g K2HPO4, 1.3 g KH2PO4, 20 g glycerol, 2.0 g (NH4)2SO4, 0.2 g MgSO·7H2O, 1.0 g yeast extract, 2.0 ml Fe solution, and 1.0 ml trace element solution. Furthermore, in fermentation medium, the phosphate was reduced to 1.0 g K2HPO4 and 0.5 g KH2PO4, and Fe solution was reduced to 1.0 ml. The Fe solution contained (per liter): 5 g FeSO4·7H2O and 4 ml HCl (37%). The trace element solution contained (per liter): 70 mg ZnCl2, 20 mg CuCl2·2H2O, 0.1 g MnCl2·4H2O, 25 mg NiCl2·6H2O, 60 mg H3BO3, 35 mg Na2MO4·2H2O, 0.2 g CoCl2·2H2O, and 4 ml HCl (37%).

Culture conditions

The seed cells were grown in a 500-ml shake flask containing 100 ml media at 180 rpm and 37 °C for 14 h aerobically. The fed-batch cultivation was carried out in a 7.5 l stirring bioreactor (NBS Bioflo 110, USA) with a working volume of 4 l after 10% inoculation by volume. Temperature and agitation speed were maintained at 37 °C and 300 rpm, respectively. The pH was maintained at 7.0 by automatic addition of 5 mol l−1 KOH. N2 of 0.5 l min−1 or air of 1.6 l min−1 was sparged into bioreactor to maintain the anaerobic or microaerobic conditions. It is worth pointing out that under the microaerobic conditions, the dissolved oxygen value decreased rapidly from 100 to 0% within 2 h until the end of fermentation although air was sparged into the bioreactor through the fermentation processes but was always 0 when N2 was sparged. Eighty percent glycerol or solution contained 80% glycerol, 10% sugar, and 10% H2O was fed into bioreactor to maintain the glycerol concentration of 60 g l−1 from 6 to 40 h, and then the supplement was stopped to lead the minimum of glycerol in the final broth.

Analytical methods

The biomass concentration was measured as absorbance at 600 nm using a Beckman DU640 UV/VIS spectrophotometer.

Glycerol, sucrose, and the products 1,3-PD, 2,3-BD, lactic acid, acetic acid, succinic acid, and ethanol were determined by high performance liquid chromatography (HPLC) system (Waters 510 system, USA) with an Aminex HPX-87H Organic Acid Analysis Column (Bio-Rad, USA), using a Waters 2414 Refractive Index Detector. The column temperature was 65 °C and the detector temperature was 45 °C. A solution of 5 mmol l−1 H2SO4 was used as mobile phase at 0.8 ml min−1 flow rate.

Construction of the ldhA deficient mutants

An 805-bp segment of truncated ldhA gene of K. oxytoca M5al was polymerase chain reaction (PCR)-amplified with oligonucleotides primer1 (forward 5′-ACGGTTGCGAACGGTATGTA-3′) and primer2 (reverse 5′-AGTGGTCTCCGAAATGCTGA-3′) using total DNA from K. oxytoca M5al as template. The segment was cloned into the pGEM-T easy vector (Promega, USA) and then transferred to suicide vector pGPCm after being digested with EcoRI, resulting in vector pLDH. The pLDH vector was then transformed into E. coli SM10 (Miller and Mekalanos 1988), and the resulting strain was used as donor in conjugation with K. oxytoca M5al. Transconjugants were selected for both chloromycetin resistance from pGPCm inserting and ampicillin resistance because the K. oxytoca M5al strain is resistive to ampicillin while the E. coli SM10 is sensitive.

Assays of the insertion mutants

PCR and Southern blot analysis were used to confirm that the transconjugants were the correct insertion mutants.

Oligonucleotides primer3 (forward 5′-CATGCGCTCCATCAAGAAGA-3′) and primer4 (reverse 5′-GTGGGTCTCGCGGTATCATT-3′) were designed to amplify a 1,160-bp segment of pGPCm (from 1,627 to 2,786 bp). Total DNA from K. oxytoca M5al and the four transconjugants were used as template and the pGPCm plasmid DNA was used as positive contrast.

The 805-bp segment of truncated ldhA was used as probe and the manipulation was performed according to the instructions of DIG DNA Labeling and Detection Kit (Boehringer Mannheim, Germany). Total DNA of K. oxytoca M5al and the four transconjugants were digested by XbaI overnight, electrophoresed on a 0.8%-agarose gel, and then transferred to positively charged nylon membrane for hybridizing with the labeled probe.

Preparation of cell-free extracts and LDH activity assays

Determination of lactate dehydrogenase (LDH) activity was done as described by Tarmy and Kaplan (1968).

Cells were grown in 200 ml rich broth media (10 g tryptone, 5 g NaCl, and 1 g yeast per liter) for 12 h and were harvested by centrifugation. After washing the cell pellet with 50 ml 0.1 mol l−1 KH2PO4 buffer (pH 7.5) for two times, the cells were resuspended in 3.0 ml phosphate buffer. The cells were disrupted by ultrasonic treatment and then centrifuged at 145,000×g for 1 h at 4 °C, the precipitate containing cellar debris and unbroken cells was discarded and the supernatant was used to LDH activity assay.

LDH (EC1.1.1.28) activity was assayed using a Beckman DU640 UV/VIS spectrophotometer. The assay followed the decrease in absorbance at 340 nm as NADH was oxidized to nicotinamide adenine dinucleotide (NAD+) by pyruvate as catalyzed by LDH. The LDH assay mix solution contained the following: KH2PO4 buffer 0.1 mol l−1, pH 7.5, NADH 0.33 mmol l−1, and sodium pyruvate 30 mmol l−1. A cell-free extract of 20 μl was added into 3.0 ml mix solution to begin the reaction. One unit of enzyme activity is defined as the amount of enzyme necessary to convert 1 μmol NADH to NAD+ per minute.

Protein concentrations were determined by the method of Bradford (1976) using bovine serum albumin (BSA) as the standard.

Results

Construction and assay of the ldhA deficient mutants

Four transconjugants selected by antibiotic were named LDH1 to 4 and then PCR and Southern blot analysis confirmed that the pLDH vector had integrated into the chromosome of the strains and the ldhA gene had been deleted by the pGPCm inserting. That is, the transconjugants were the correct insertion mutants.

The PCR result was shown in Fig. 1. Segments of about 1,200 bp were amplified from the chromosome DNA of the four transconjugants as well as the pGPCm plasmid DNA but no product can be obtained from the M5al chromosome DNA. This result indicated that the pLDH vector had integrated into the chromosome of the transconjugant strains.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0563-7/MediaObjects/253_2006_563_Fig1_HTML.gif
Fig. 1

PCR analysis of total DNA from K. oxytoca M5al and the mutants

The result of Southern blot was shown in Fig. 2. Because there is one XbaI site in pGPCm while no XbaI site in the ldhA gene, the two hybridizing bands in the transconjugants DNA indicated that the ldhA gene had been divided into two segments by inserting of the pGPCm while only one band in the M5al DNA.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0563-7/MediaObjects/253_2006_563_Fig2_HTML.gif
Fig. 2

Southern blot analysis of total DNA from K. oxytoca M5al and the mutants

LDH activity assays

The LDH activities of the four mutants were from 3.85 to 6.92% of the parental strain (Table 1). This result further proved that the ldhA gene had been knocked out by insertion. The mutant LDH3, whose LDH activities is the lowest, was used in the fermentations aftermentioned.
Table 1

LDH activity of K. oxytoca M5al and the mutants

Strains

LDH activity

Protein (U/mg)

Percentage of wild type

M5al

0.8617

100

LDH1

0.0596

6.92

LDH2

0.0424

4.92

LDH3

0.0332

3.85

LDH4

0.0549

6.37

Strains were grown for 12 h aerobically at 37 °C in rich broth plus 0.4% glucose.

Comparison of fed-batch fermentations by K. oxytoca M5al and LDH3

The results of fed-batch fermentation by K. oxytoca M5al and LDH3 under anaerobic and microaerobic conditions are compared in Tables 2 and 3, respectively. The fermentation efficiency for 1,3-PD were improved significantly by LDH3 compared with K. oxytoca M5al in both cases. The indexes of 1,3-PD concentration, conversion, and productivity under anaerobic conditions were 45, 29, and 32% higher than that of K. oxytoca M5al, respectively, and were 58, 22, and 21% higher under microaerobic conditions. No lactic acid was produced in the LDH3 fermentation processes under both conditions. The results also showed that more biomass, higher 1,3-PD concentration, and productivity can be gained for both strains under microaerobic conditions but conversion was only a little lower than that of anaerobic conditions. This is the reason we carried out fermentations under microaerobic conditions below. Furthermore, a lot more 2,3-BD was produced by both strains under microareobic conditions, and it even replaced the lactic acid as the maximum byproduct in K. oxytoca M5al.
Table 2

Comparison of fed-batch fermentations by K. oxytoca M5al and LDH3 under anaerobic conditions

Strains

AGC (g l−1)

OD600

Products concentration (g l−1)

Conversion (mol mol−1)

Carbon recovery (%)

Productivity (g l−1 h−1)

PD

BD

Lac

Suc

Ace

Eth

M5al

116.37

4.75

39.14

5.27

16.73

3.77

12.31

12.26

0.41

96.38

0.63

LDH3

130.16

4.37

56.73

9.12

ND

6.67

14.23

15.89

0.53

95.16

0.83

AGC accumulative glycerol consumed, ND could not be detected by HPLC, PD 1,3-propanediol, BD 2,3-butanediol, Lac lactic acid, Suc succinic acid, Ace acetic acid, Eth ethanol

Table 3

Comparison of fed-batch fermentations by K. oxytoca M5al and LDH3 under microaerobic conditions

Strains

AGC (g l−1)

OD600

Products concentration (g l−1)

Conversion (mol mol−1)

Carbon recovery (%)

Productivity (g l−1 h−1)

PD

BD

Lac

Suc

Ace

Eth

M5al

134.03

7.16

43.26

25.37

22.13

4.29

6.42

11.56

0.36

101.3

0.86

LDH3

149.71

6.99

62.64

35.18

ND

4.08

5.33

9.65

0.50

97.95

1. 04

AGC accumulative glycerol consumed, ND could not be detected by HPLC, PD 1,3-propanediol, BD 2,3-butanediol, Lac lactic acid, Suc succinic acid, Ace acetic acid, Eth ethanol

Effect of sucrose on cell growth and products formation in fed-batch fermentation of LDH3 under microaerobic conditions

As indicated in Fig. 3, sucrose added at the beginning of fermentation led to the rapid cell growth in a short time while sucrose fed together with glycerol increased the biomass a little until the biomass all reached a close level of about 7.0 (OD600 nm) after 30 h under the conditions whether or how sucrose was added into the bioreactor.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0563-7/MediaObjects/253_2006_563_Fig3_HTML.gif
Fig. 3

Effect of sucrose to cell growth of mutant LDH3 (diamond) no sucrose added in the fermentation process, (square) sucrose fed with glycerol during the fermentation process, (triangle) sucrose added at the beginning of fermentation, (x mark) sucrose added at the beginning and also fed with glycerol during the fermentation process

Effects of sucrose on products formation are showed in Table 4. Sucrose added at the beginning of fermentation increased the productivity from 1.04 to 1.14 g l−1 h−1, while sucrose fed with glycerol improved the conversion from 0.5 to 0.58 mol mol−1. And when the two steps were combined, the 1,3-PD concentration, conversion, and productivity were all improved substantially (Table 5).
Table 4

Effect of sucrose on fed-batch fermentations by mutant LDH3

Number

AGC (g l−1)

ASC (g l−1)

Products concentration (g l−1)

Conversion (mol mol−1)

Carbon recovery (%)

Productivity (g l−1 h−1)

PD

BD

Suc

Ace

Eth

1

149.71

0

62.64

35.18

4.08

5.33

9.65

0.50

97.95

1. 04

2

137.13

18.58

66.12

52.07

4.87

5.71

4.92

0.58

96.68

1.07

3

164.89

10.67

68.13

49.64

3.96

6.37

7.28

0.50

96.75

1.14

1 no sucrose was added in the fermentation process, 2 sucrose was fed with glycerol in the fermentation process, 3 sucrose was added at the beginning of fermentation

ASC accumulative sucrose consumed, AGC accumulative glycerol consumed, PD 1,3-propanediol, BD 2,3-butanediol, Suc succinic acid, Ace acetic acid, Eth ethanol

Table 5

Comparison of fed-batch fermentations with sucrose as cosubstrate by K. oxytoca M5al and LDH3 under microaerobic conditions LDH3

Strains

AGC (g l--1)

ASC (g l--1)

Products concentration (g l--1)

Conversion (mol mol--1)

Carbon recovery (%)

Productivity (g l--1 h--1)

PD

BD

Lac

Suc

Ace

Eth

M5al

139.46

28.52

58.79

32.85

23.57

3.17

7.83

11.07

0.51

97.65

0.98

LDH3

164.67

31.72

83.56

60.11

ND

3.62

6.32

12.39

0.62

95.26

1.39

ASC accumulative sucrose consumed, AGC accumulative glycerol consumed, ND could not be detected by HLPC, PD 1,3-propanediol, BD 2,3-butanediol, Lac lactic acid, Suc succinic acid, Ace acetic acid, Eth ethanol

Comparison of fed-batch fermentations with sucrose as cosubstrate by K. oxytoca M5al and LDH3 under microaerobic conditions

More than 30 fed-batch fermentations were carried out under microaerobic conditions, and 10 g l−1 sucrose was added just before inoculation with solution contained 80% glycerol, 10% sugar, and 10% H2O fed into bioreactor after 6 h in these fermentations. The representative time courses of fed-batch fermentation by K. oxytoca M5al and LDH3 under these conditions are indicated in Figs. 4 and 5, respectively. In addition, the final fermentation results by them are compared in Table 5. The fermentation efficiency for 1,3-PD was improved obviously by LDH3 under these conditions and the 1,3-PD concentration, conversion, and productivity were 42, 22, and 42% higher than that of K. oxytoca M5al, respectively. Also, 2,3-BD produced by LDH3 almost doubled that by K. oxytoca M5al.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0563-7/MediaObjects/253_2006_563_Fig4_HTML.gif
Fig. 4

Representative time course of a K. oxytoca M5al fed-batch fermentation under microaerobic condition

https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0563-7/MediaObjects/253_2006_563_Fig5_HTML.gif
Fig. 5

Representative time course of a LDH3 fed-batch fermentation under microaerobic condition

Discussion

Lactic acid was not given much importance in former studies maybe because it was not so toxic to the cells and so little lactic acid are formed that it could be negligible to 1,3-PD yield. But in this work, lactic acid was the main byproduct of K. oxytoca M5al (Tables 2, 3, and 5). It always reached 20 g l−1 in the broth and even exceeded 40 g l−1 in some individual fermentation (data not shown). The lactate pathway will contest NADH with the 1,3-PD pathway and too much lactic acid yield will certainly reduce the conversion from glycerol to 1,3-PD (Ahrens et al. 1998; Biebl et al. 1999; Zeng and Biebl 2002; Zeng et al. 1993). So, we had to take action to prevent its production. Moreover, this action will bring facilitation and cost reduction to the downstream processes. Hereby, we constructed four mutants deficient in lactate-producing pathway and the mutant LDH3 was used in the actual fermentations. 2,3-BD was the maximum byproduct and no lactic acid was produced in this mutant, which alleviated the multiple inhibition to cell growth and led to the improvement of 1,3-PD production (Tables 2, 3, and 5).

Adding sugar as cosubstrate was a feasible way to improve the 1,3-PD conversion and had been widely used (Abbad-Andaloussi et al. 1998; Biebl and Marten 1995; Saint-Amans and Soucaille 1995; Saint-Amans et al. 2001; Tong and Cameron 1992). Although sugar could not be converted to 1,3-PD, it may be used for cell growth and regeneration of reducing power. Sucrose fed together with glycerol increased the conversion from glycerol to 1,3-PD as expected, and was mostly converted into 2,3-BD (Table 4). The effect of sucrose added at the beginning of fermentation was also studied in this work. The results (Fig. 3 and Table 4) showed that productivity was increased because cells grew rapidly in a short time and more cells worked throughout the fermentation process. When the two steps were combined, the indexes of 1,3-PD concentration, conversion, and productivity were all improved substantially (Table 5). Glucose was not used because the “glucose effect” was so obvious that bacteriolysis often happened when glucose was to be exhausted whereas glycerol could not be used by the bacteria in time.

Although the production of 1,3-PD from glycerol is generally performed under anaerobic condition, 1,3-PD could still be obtained under microaerobic or mild aerobic conditions (Huang et al. 2002; Chen et al. 2003; Cheng et al. 2004). In all these studies, more biomass, higher productivity, and almost equal or lower conversion were gained, and Huang et al. (2002) and Cheng et al. (2004) also mentioned that 2,3-BD production was enhanced under microaerobic conditions. We gained the similar results as they were described but the 2,3-BD production was much more than they gained (Table 2). 2,3-BD fermentation from sugar was generally carried out in the range pH 5.0–6.0 and the availability of oxygen was the most important operating factor affecting the fermentation (Voloch et al. 1985). Biebl et al. (1998) also reported that 2,3-BD generated at low pH of 6.5 and increased as the pH fell in glycerol fermentation by K. pneumoniae, especially when glycerol was excess; however, a great deal of 2,3-BD produced at neutral pH in this work. This may be due to three reasons: first, more substrates (glycerol and/or sucrose) and NADH flowed to 2,3-BD pathway after the lactate pathway was eliminated; second, the excess glycerol and the air flow were satisfactory for the strain to produce 2,3-BD; and the last, maybe the pH was not strict to this strain and the acetic acid accumulation was not the only factor that determined 2,3-BD formation. 2,3-BD is also known as an important chemical feedstock and liquid fuel, and it also can be used as monomer for the synthesis of various polyesters (Syu 2001; Voloch et al. 1985). Due to the special structure, it is hard and costly to go through chemical synthesis. In this work, the results of more than 80 g l−1 1,3-PD coupled with more than 60 g l−1 2,3-BD is undoubtedly an evangel to chemical industry. Although some more 1,3-PD yield would be gained theoretically when acetic acid was the only by-product than 2,3-BD as the only by-product (Zeng et al. 1993), if we take the lower toxicity to cells and 1,3-PD production and the valuable applications of 2,3-BD into account, maybe having 2,3-BD as the main product of the oxidative branch is the really preferred and profitable route to carry the microbial 1,3-PD industrialization into effect.

Acknowledgments

This work was supported by the National Key Technologies Research and Development Program of China during the 10th Five-Year Projects period (2001BA708B01-04).

Copyright information

© Springer-Verlag 2006