Applied Microbiology and Biotechnology

, Volume 89, Issue 4, pp 1119–1125

Elimination of carbon catabolite repression in Klebsiella oxytoca for efficient 2,3-butanediol production from glucose–xylose mixtures

Authors

  • Xiao-Jun Ji
    • State Key Laboratory of Materials-Oriented Chemical Engineering, College of Biotechnology and Pharmaceutical EngineeringNanjing University of Technology
  • Zhi-Kui Nie
    • State Key Laboratory of Materials-Oriented Chemical Engineering, College of Biotechnology and Pharmaceutical EngineeringNanjing University of Technology
    • State Key Laboratory of Materials-Oriented Chemical Engineering, College of Biotechnology and Pharmaceutical EngineeringNanjing University of Technology
  • Lu-Jing Ren
    • State Key Laboratory of Materials-Oriented Chemical Engineering, College of Biotechnology and Pharmaceutical EngineeringNanjing University of Technology
  • Chao Peng
    • State Key Laboratory of Materials-Oriented Chemical Engineering, College of Biotechnology and Pharmaceutical EngineeringNanjing University of Technology
  • Ping-Kai Ouyang
    • State Key Laboratory of Materials-Oriented Chemical Engineering, College of Biotechnology and Pharmaceutical EngineeringNanjing University of Technology
Applied Genetics and Molecular Biotechnology

DOI: 10.1007/s00253-010-2940-5

Cite this article as:
Ji, X., Nie, Z., Huang, H. et al. Appl Microbiol Biotechnol (2011) 89: 1119. doi:10.1007/s00253-010-2940-5

Abstract

Microbial preference for glucose implies incomplete and/or slow utilization of lignocellulose hydrolysates, which is caused by the regulatory mechanism named carbon catabolite repression (CCR). In this study, a 2,3-butanediol (2,3-BD) producing Klebsiella oxytoca strain was engineered to eliminate glucose repression of xylose utilization. The crp(in) gene, encoding the mutant cyclic adenosine monophosphate (cAMP) receptor protein CRP(in), which does not require cAMP for functioning, was characterized and overexpressed in K. oxytoca. The engineered recombinant could utilize a mixture of glucose and xylose simultaneously, without CCR. The profiles of sugar consumption and 2,3-BD production by the engineered recombinant, in glucose and xylose mixtures, were examined and showed that glucose and xylose could be consumed simultaneously to produce 2,3-BD. This study offers a metabolic engineering strategy to achieve highly efficient utilization of sugar mixtures derived from the lignocellulosic biomass for the production of bio-based chemicals using enteric bacteria.

Keywords

2,3-ButanediolKlebsiella oxytocaGlucoseXyloseCarbon catabolite repressionMutant cAMP receptor protein

Introduction

As fossil fuel supplies are becoming increasingly scarce, the bio-refinery systems that integrate biomass conversion processes and equipment to produce fuels, power, and chemicals from annually renewable resources, are at the stage of development worldwide (Li et al. 2010; Ragauskas et al. 2006). Lignocellulose derived from agricultural residues is a kind of potential low-cost feedstock for fermentative production of different types of chemicals (van Haveren et al. 2008). Hydrolysis of lignocellulose yields a mixture of sugars containing mainly glucose and xylose (Sheehan and Himmel 1999), thus it is important that the applied microorganism could utilize both substrates to maximize carbon conversion to a product. Klebsiella species belong to such organisms, and they are able to produce 2,3-butanediol (2,3-BD) from both glucose and xylose (Cheng et al. 2010; Kosaric et al. 1990; Wang et al. 2010). Recently, microbial production of 2,3-BD has attracted worldwide interest due to the wide industrial application of 2,3-BD (Ji et al. 2009a; Syu 2001; Celińska and Grajek 2009).

In our previous study, a 2,3-BD hyper-producing Klebsiella oxytoca strain was obtained by altering the mixed acid–butanediol fermentation pathway (Ji et al. 2008; 2010). And, an industrial medium containing urea as a sole nitrogen source, low levels of corn steep liquor and mineral salts as nutrition factors to obtain high 2,3-BD production through co-fermentation of glucose and xylose was developed (Ji et al. 2009b). However, when glucose and xylose were present at the same time, xylose consumption generally did not commence until glucose was depleted, i.e., sugars were sequentially consumed resulting in several exponential growth phases that were separated by intermediate lag phases, developing so-called diauxic growth (Ji et al. 2009b). This phenomenon is universally displayed among the other enteric bacteria such as Klebsiella pneumoniae and Escherichia coli (Gosset 2005; Wang et al. 2010), and it hinders the possibility of utilizing the lignocellulose hydrolysates as carbon sources for efficient chemical production. Moreover, the utilization of xylose in mixed sugar fermentation was usually delayed and was slower than fermentation of pure xylose; furthermore, it was often incomplete in some cases (Bothast et al. 1994; Moniruzzaman et al. 1996). This would be disadvantageous for the fermentative production process using the sugar mixtures derived from lignocellulose hydrolysates as substrates.

The described above trait was attributed to carbon catabolite repression (CCR) or glucose effect (Görke and Stülke 2008; Jojima et al. 2010). CCR has been studied extensively in enteric bacteria with respect to the ability of glucose to block induction of the genes for other substrate utilizations, such as xylose via the phosphoenolpyruvate-dependent phosphotransferase system (PTS) and cyclic adenosine monophosphate (cAMP; Postma et al. 1993; Roseman and Meadow 1990). In the enteric bacteria, when glucose is absent from the medium, the IIA component of the glucose-specific PTS (EIIAGlc) is mainly in its phosphorylated state, bound to an enzyme adenylate cyclase (EC 4.6.1.1) and activating its cAMP biosynthetic capacity, which leads to increase in cAMP concentration. The cAMP further binds to a cAMP receptor protein (CRP) and causes the induction of carbon catabolite repressed genes. However, if the glucose is present in the culture medium, EIIAGlc protein is mainly in its nonphosphorylated state, bound to various non-PTS permeases and inhibiting uptake of non-PTS sugars, which is called “inducer exclusion” (Gosset 2005). This response is mediated mainly by the CRP whose active form is a homodimer requiring cAMP for functioning (Saier and Ramseier 1996). In the present study, a mutant CRP, which does not require cAMP for functional dimerization, was characterized and overexpressed in a 2,3-BD producing K. oxytoca strain to eliminate the CCR. The profiles of sugar consumption and 2,3-BD production by the engineered recombinant, in glucose and xylose mixtures, were examined and showed that glucose and xylose could be consumed simultaneously to produce 2,3-BD.

Materials and methods

Strains and plasmids construction

Strains, plasmids and primers used in this study were listed in Table 1. K. oxytoca ME-XJ-8, a mutant from K. oxytoca CCTCC M207023 (China Center for Type Culture Collection), was used as the parent strain (Ji et al. 2010). The crp(in) gene encoding a mutant CRP(in) from E. coli ET25 (Eppler and Boos 1999) was polymerase chain reaction (PCR) amplified using primers PC1 and PC2 designed according to the sequence of the crp(in). The amplified crp(in) gene was inserted into a vector pMD18-T Simple, resulting in pMDT-crp(in), from which the crp(in) gene was excised by BamH I and Sal I and introduced into the BamH I-Sal I site of an expressing vector pDK7 under the control of the trp-lac (tac) promoter. Then pDK7-crp(in) was transformed into the competent cells of K. oxytoca ME-XJ-8 by standard transformation protocol using the method of electroporation (Jeong et al. 1998; Zhu et al. 2009). The chloromycetin-resistant transformants were selected, and the insert was confirmed by colony PCR, restrictive digestion, and sequencing. The confirmed clone ME-XJ-8 (pDK7-crp(in)) was designated ME-CRPin. The stability of plasmid pDK7 in K. oxytoca in the absence of antibiotics had been tested, and the plasmid could be stably maintained in the engineered strain for at least 80 generations. Therefore, the strain ME-CRPin could be cultivated in media in the absence of antibiotics, where the effect of plasmid could be minimized. Furthermore, fermentation characteristics of ME-XJ-8 and ME-CRPin were compared using data from flask experiments in the absence of antibiotics, and no obvious difference was observed.
Table 1

Bacterial strains, plasmids, and primers used in this study

Strain, plasmid or primer

Genotype, properties, or sequence

Source or reference

Strains

 K. oxytoca ME-XJ-8

Apr, parent strain

Ji et al. (2010)

 K. oxytoca ME-CRPin

Apr, Cmr, ME-XJ-8/pDK7-crp(in)

This work

 E. coli DH5α

supE44 ∆lacU169 (φ80 lacZ ∆M15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1

TaKaRa

Plasmids

 pLOI 3244

Apr, crp(in) PCR product from E. coli ET25 in pUC18 (3,813 bp)

Wood (2005)

 pDK7

Expressing vector, Cmr (4,800 bp)

Kleiner et al. (1988)

 pMD18-T Simple

Apr (2,692 bp)

TaKaRa

 pMDT-crp(in)

Apr, pMD18-T Simple derivative carrying a 1,109-bp DNA fragment containing the crp(in) gene

This work

 pDK7-crp(in)

Cmr, pDK7 derivative, where a 1,109-bp DNA fragment containing the crp(in) gene was inserted

This work

Primers

 PC1

5′-TTAGGATCCCTCCACTGCGTCAATTTTC-3′ (BamH I)

This work

 PC2

5′-TAAGTCGACCAGGAACGAGGGAGAAGAG-3′ (Sal I)

This work

The underlined letters indicate the restriction sites

Apr ampicillin resistance, Cmr chloromycetin resistance

Medium and culture methods

Luria–Bertani (LB) medium was used for the seed cultures of E. coli, K. oxytoca, and their derivatives. When K. oxytoca was subjected to electroporation, EDTA was added to the LB medium to a final concentration of 0.7 mM (Zhu et al. 2009). After the electroporation, the cells were regenerated in SOC medium (Joseph and David 2001). When necessary, ampicillin (60 μg/mL) and/or chloromycetin (12 μg/mL) were added to the medium as selection pressure.

For a seed culture preparation, a full loop of K. oxytoca from a fresh slant was inoculated into an Erlenmeyer flask (500-mL) containing 100 mL of fresh seed medium and cultivated on a rotary shaker at 200 rpm for 24 h. Seed culture (5%, v/v) was then inoculated into the glucose–xylose co-fermentation medium in which the glucose and xylose were set at three ratios (w/w, 2:1, 1:1, 1:2), representing what was present in different sources of lignocellulose hydrolysates. The composition of the fermentation medium besides the sugars is provided in our previous study (Ji et al. 2009b). Batch fermentation was carried out in a 3-L stirred fermenter (BioFlo 100; New Brunswick Scientific Co., NJ, USA) with a working volume of 2 L. All cultivations were carried out at 37 °C with the aeration rate and agitation speed at 1.0 vvm and 200 rpm, respectively. pH was controlled at 6.5 automatically by adding 3 M NaOH (for the engineered K. oxytoca strain, when OD600 reached about 0.7, 1 mmol/L isopropyl beta-d-thiogalactoside (IPTG) was added to induce expression of the recombinant protein).

Analytical methods

Biomass was determined by measuring turbidity at 600 nm with appropriate dilution using a UV–visible spectrophotometer (Lambda-25, Perkin-Elmer, USA). Dry cell weight (DCW) was estimated from the absorbance, and 1 unit of optical density was determined to be equivalent to 0.35 g DCW per liter. The substrates (glucose and xylose) and the metabolites (2,3-BD and acetoin) were analyzed by high pressure liquid chromatography as indicated in our previous study (Ji et al. 2008; Yan et al. 2009).

Results

Characterization of the crp(in) gene

The crp(in) gene amplified from E. coli ET25 located in pLOI 3244 was sequenced (GenBank accession number HM595439). The ORF of crp(in) gene is 633 bp long and the predicted amino acid sequence encompasses 210 amino acid residues. The nucleotide and deduced amino acid sequences of crp(in) gene were compared with sequences available in the National Center for Biotechnology Information GenBank database using Basic Local Alignment Search Tool. The nucleotide sequence of crp(in) and amino acid sequence of CRP(in) were aligned with four enteric bacterial origin crp and CRP and turned out to have a high level of identity (Table 2). Comparison of the nucleotide sequences of crp genes from E. coli K12, K. pneumoniae MGH 78578, K. oxytoca M5al, Klebsiella aerogenes KC1043, and crp(in) gene from E. coli ET25 revealed a high degree of identity among all five sequences (Fig. S1). The five residues in the crp(in) gene of E. coli ET25 that are different from the other crp genes might contribute to the non-cAMP dependent phenotype (A337C, C383T, G433A, A607G, T614C). In order to further investigate this deduction, the predicted amino acid sequences of each gene were compared (Fig. S2), and the results confirmed that the five positional switch sites play a vital role in the mutant phenotype. The amino acid changes, I113L, T128I, A147T, T203A, and V205A, were thus shown to confer a CRP(in) phenotype. Also, high level of identity of the amino acid sequences among the five analyzed proteins further indicates the conserved function of crp in the enteric bacteria.
Table 2

The BLAST results of crp(in) gene and its predicted amino acid sequences in NCBI

Definition

GenBank accession number

Identities

Strains

References

crp(in)

HM595439

100%

E. coli ET25

This work

crp

AC000091

99.1%

E. coli K12

Riley et al. (2006)

crp

AJ278967

85.6%

K. oxytoca M5al

Li et al. (2002)

crp

M68973

87.0%

K. aerogenes KC1043

Osuna and Bender (1991)

crp

NC009648

87.2%

K. pneumoniae MGH 78578

CRP(in)

ADK89557.1

100%

E. coli ET25

This work

CRP

AP_004432

97.1%

E. coli K12

Riley et al. (2006)

CRP

CAC07215

96.2%

K. oxytoca M5al

Li et al. (2002)

CRP

AAA25058

97.1%

K. aerogenes KC1043

Osuna and Bender (1991)

CRP

YP 001337397

97.1%

K. pneumoniae MGH 78578

NCBI National Center for Biotechnology Information, BLAST basic local alignment search tool

Altered pattern of glucose and xylose utilization in the engineered K. oxytoca

Plasmid pDK7-crp(in) was constructed as described in the “Strains and plasmids construction” section. It was then transformed into K. oxytoca ME-XJ-8 by electroporation. To determine the effect of crp(in) overexpression on xylose utilization, the engineered recombinant ME-CRPin and the parent strain ME-XJ-8 were grown in three different glucose–xylose mixtures and sugar concentrations were monitored during the cultivation. The strain harboring the cloning vector pDK7 was used as a control. The engineered recombinant ME-CRPin could metabolize both sugars simultaneously; in contrast, the parent strain ME-XJ-8 metabolized xylose only after exhaustion of the glucose (Fig. 1). As summarized in Table 3, in the case of three different glucose–xylose mixtures, expression of crp(in) could shorten the time of consumption of both sugars (up to 10 h when using 40 g/L glucose and 20 g/L xylose mixtures). This proves the elimination of diauxic growth by crp(in)-expressing recombinant.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-010-2940-5/MediaObjects/253_2010_2940_Fig1_HTML.gif
Fig. 1

Time course of cell growth and sugar consumption of K. oxytoca ME-XJ-8 (A, B, C) and its recombinant K. oxytoca ME-CRPin (D, E, F) in different glucose–xylose mixtures. A, D 40 g/L glucose plus 20 g/L xylose. B, E 30 g/L glucose plus 30 g/L xylose. C, F 20 g/L glucose plus 40 g/L xylose

Table 3

Comparison of the metabolic characteristics of K. oxytoca ME-XJ-8 and its recombinant K. oxytoca ME-CRPin in different glucose–xylose mixtures

Strain

Composition of sugar mixture (g/L)

Sugar consumption time (h)

Glucose

Xylose

Glucose

Xylose

Glucose + Xylose

ME-XJ-8

40

20

26

44

44

ME-CRPin

40

20

30

34

34

ME-XJ-8

30

30

28

48

48

ME-CRPin

30

30

30

44

44

ME-XJ-8

20

40

18

48

48

ME-CRPin

20

40

22

44

44

2,3-Butanediol production by engineered K. oxytoca cultured in glucose–xylose mixtures

Complete hydrolysis of lignocellulose such as corn stover generates a solution containing primarily glucose and xylose in the ratio of around 2:1 (w/w) (Stephanopoulos 2007). If the process of 2,3-BD production from lignocellulosic materials is to be economically attractive, effective co-fermentation of mixed sugar must be achieved (Ji et al. 2009b). Therefore, the ability of each strain to ferment a concentrated (6%, w/v) mixture of glucose and xylose (2:1, w/w) was compared. As shown in Fig. 2, ME-CRPin metabolized almost xylose simultaneously with glucose, whereas ME-XJ-8 fermented the glucose before using the xylose to synthesize 2,3-BD. In the case of the recombinant, 2,3-BD production was not affected by heterologous crp(in) expression, and the 2,3-BD yield (0.44 g/g) mirrored the total sugar consumption nearly the same as the parent strain. However, the highest 2,3-BD production of 23.9 g/L from the sugar mixtures was obtained at 34 h, 10 h earlier compared to the parent strain. This might be explained by the fact that the classic diauxie which was commonly observed in dual substrate batch fermentation of the parent strain was eliminated efficiently (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-010-2940-5/MediaObjects/253_2010_2940_Fig2_HTML.gif
Fig. 2

Time course of 2,3-butanediol production using AK. oxytoca ME-XJ-8 and its recombinant BK. oxytoca ME-CRPin in the mixture of glucose and xylose (2:1, w/w)

Discussion

The use of lignocellulose hydrolysates offers a means to increase the economics of 2,3-BD production beyond current starch-based production. However, the step utilization of mixed sugars caused by the regulatory mechanism of CCR delays the fermentation of lignocellulose hydrolysates as raw materials. To some extent, the CCR is a prevailing obstacle that must be overcome for efficient utilization of the lignocellulosic biomass (Jojima et al. 2010). Many previous efforts at relieving the CCR and eliminating diauxie have focused on mutations of the PTS (Dien et al. 2002; Hernández-Montalvo et al. 2001; Li et al. 2007; Lindsay et al. 1995; Nichols et al. 2001). Due to inactivation of ptsG gene encoding the major glucose transporter IICBGlc involved in PTS, the strains lacked the molecular machinery required for the CCR. Many of the obtained mutants did co-utilize glucose and xylose simultaneously, but this came at the expense of glucose utilization rates. To recover growth deficits in the PTS mutant strain, increased expression of a secondary glucose transporter and/or glucokinase was needed (Gosset 2005; Hernández-Montalvo et al. 2003; Snoep et al. 1994; Wang et al. 2005).

In the present study, a simple metabolic engineering strategy was developed to relieve the CCR in K. oxytoca for simultaneous consumption of glucose and xylose to produce 2,3-BD. A mutant CRP protein (CRP(in)) which is independent of cAMP was characterized and overexpressed in a 2,3-BD producing K. oxytoca. The recombinant turned to be able to consume glucose and xylose simultaneously to produce 2,3-BD in glucose–xylose mixtures. Moreover, the operating time was reduced and the 2,3-BD productivity was increased for the recombinant compared with the parent strain. To the best of our knowledge, this is the first report showing that elimination of the CCR in K. oxytoca has a prominent role in simultaneous consumption of glucose and xylose for 2,3-BD production.

The above response might be attributed by the following reasons. In the enteric bacteria, the use of sugar is transcriptionally regulated by CRP (Görke and Stülke 2008; Jojima et al. 2010). The active form of CRP is a homodimer requiring cAMP for functioning. In the present study, we have shown that the mutations of CRP(in) were located within the region of the protein known to be involved in functional dimerization with cAMP (Harman et al. 1986), and thus, conferred a cAMP-independent phenotype. Therefore, in the recombinant, the overexpressed CRP(in) did not require cAMP for functional dimerization and played the regulatory role in the absence of cAMP. Thus, it could replace the native CRP, and facilitate xylose uptake from mixtures of glucose and xylose. The CRP(in) phenotype should promote xylose uptake in the presence of glucose by activating the native xylose transporters and/or by activating other CRP-controlled promiscuous transporters capable of xylose uptake. Furthermore, in order to make the process economically feasible, the promoter of the plasmid used in present study should be exchanged, some other stress-induced (such as temperature-induced) promoters could be applied. In addition, the lactose could be used instead of IPTG to induce the CRP(in) expression for industrial production.

In the present work, it was shown that an engineered K. oxytoca harboring the CRP(in) phenotype could metabolize glucose and xylose mixtures simultaneously, which would be beneficial for efficient mixed substrates utilization. Thus, this recombinant has the potential for producing 2,3-BD directly from the inexpensive lignocellulose hydrolysates. As a consequence of the co-metabolism of glucose and xylose, the total sugars could be consumed faster, leading to higher growth rate and 2,3-BD productivity by the recombinant compared to the parent strain. This is another example showing the advantage of metabolic engineering to improve the microbial physiological functionality. The idea developed in this paper could be applied to the other similar industrial microorganisms to achieve highly efficient utilization of sugar mixtures derived from the lignocellulosic biomass.

Acknowledgements

We are grateful to Prof. Mike J. Merrick, John Innes Centre, UK, and Prof. Lonnie O. Ingram, University of Florida, USA, for providing the pDK7 and pLOI 3244 plasmids. This work was financially supported by the National Natural Science Foundation of China (Nos. 20606018 and 21006049), the Key Program of National Natural Science Foundation of China (No. 20936002), the National Basic Research Program of China (Nos. 2007CB707805, 2009CB724700 and 2011CB200906), and the Program for New Century Excellent Talents in University from the Ministry of Education of China (No. NCET-09-0157). X.-J. Ji was supported by the Innovation Fund for Doctoral Dissertation of Nanjing University of Technology (No. BSCX200808), the China Postdoctoral Science Foundation Funded Project (No. 20100471328) and the Jiangsu Planned Projects for Postdoctoral Research Funds of China (No. 1001015 C).

Supplementary material

253_2010_2940_MOESM1_ESM.doc (1.2 mb)
Supplementary Fig. S1Alignments of cloned nucleotide sequence containing coding regions for crp(in) from Escherichia coli ET25 (GenBank accession no. HM595439) and crp from E. coli K12 (NC000913), Klebsiella pneumoniae MGH 78578 (CP000647), Klebsiella oxytoca M5al (AJ278967), and Klebsiella aerogenes KC1043 (M68973) searched from GenBank data. Nucleotides differing in E. coli ET25 are surrounded with rectangles (DOC 1,259 kb)
253_2010_2940_MOESM2_ESM.doc (1.1 mb)
Supplementary Fig. S2Alignments of amino acid sequences of CRP(in) of Escherichia coli ET25 (ADK89557.1) and CRP of E. coli K12 (AP_004432), Klebsiella pneumoniae MGH 78578 (YP 001337397), Klebsiella oxytoca M5al (CAC07215) and Klebsiella aerogenes KC1043 (AAA25058). Amino acids differing in E. coli ET25 are surrounded with rectangles (DOC 1,140 kb)

Copyright information

© Springer-Verlag 2010