Applied Microbiology and Biotechnology

, Volume 95, Issue 4, pp 1001–1010

Physiological characterization of recombinant Saccharomyces cerevisiae expressing the Aspergillus nidulans phosphoketolase pathway: validation of activity through 13C-based metabolic flux analysis

  • Marta Papini
  • Intawat Nookaew
  • Verena Siewers
  • Jens Nielsen
Applied microbial and cell physiology

DOI: 10.1007/s00253-012-3936-0

Cite this article as:
Papini, M., Nookaew, I., Siewers, V. et al. Appl Microbiol Biotechnol (2012) 95: 1001. doi:10.1007/s00253-012-3936-0

Abstract

Several bacterial species and filamentous fungi utilize the phosphoketolase pathway (PHK) for glucose dissimilation as an alternative to the Embden–Meyerhof–Parnas pathway. In Aspergillus nidulans, the utilization of this metabolic pathway leads to increased carbon flow towards acetate and acetyl CoA. In the first step of the PHK, the pentose phosphate pathway intermediate xylulose-5-phosphate is converted into acetylphosphate and glyceraldehyde-3-phosphate through the action of xylulose-5-phosphate phosphoketolase, and successively acetylphosphate is converted into acetate by the action of acetate kinase. In the present work, we describe a metabolic engineering strategy used to express the fungal genes of the phosphoketolase pathway in Saccharomyces cerevisiae and the effects of the expression of this recombinant route in yeast. The phenotype of the engineered yeast strain MP003 was studied during batch and chemostat cultivations, showing a reduced biomass yield and an increased acetate yield during batch cultures. To establish whether the observed effects in the recombinant strain MP003 were due directly or indirectly to the expression of the phosphoketolase pathway, we resolved the intracellular flux distribution based on 13C labeling during chemostat cultivations. From flux analysis it is possible to conclude that yeast is able to use the recombinant pathway. Our work indicates that the utilization of the phosphoketolase pathway does not interfere with glucose assimilation through the Embden–Meyerhof–Parnas pathway and that the expression of this route can contribute to increase the acetyl CoA supply, therefore holding potential for future metabolic engineering strategies having acetyl CoA as precursor for the biosynthesis of industrially relevant compounds.

Keywords

Phosphoketolase Metabolic engineering 13C-based metabolic flux analysis Acetyl CoA supply Glycolysis 

Introduction

The phosphoketolase pathway (PHK) is a route used by some bacteria and filamentous fungi for sugar dissimilation as an alternative to the commonly used Embden–Meyerhof–Parnas pathway. This pathway is present in heterofermentative and facultatively homofermentative lactic acid bacteria, some bifidobacteria (Meile et al. 2001), and in xylose-fermenting yeast (Ratledge and Holdsworth 1985). The genes of this pathway have also been identified in members of the Aspergillus spp. (Panagiotou et al. 2008).

The first step of the route leads to the formation of glyceraldehyde-3-phosphate and acetylphosphate from the pentose phosphate pathway (PPP) intermediate xylulose-5-phosphate. This step is catalyzed by a phosphoketolase, encoded by the gene xpkA in the filamentous fungus Aspergillus nidulans. Acetylphosphate is successively converted into acetate by the enzyme acetate kinase, encoded by the gene ack (Spector 1980; Ingram-Smith et al. 2006), generating one ATP. In some bacteria, acetylphosphate is instead converted into acetyl coenzyme A (acetyl CoA) through the action of a phosphotransacetylase (Presecan-Siedel et al. 1999), encoded by the gene pta in Bacillus subtilis. In Saccharomyces cerevisiae, phosphoketolase activity has been detected only during growth on xylose but it was shown to be repressed on glucose (Evans and Ratledge 1984). Recently, the phosphoketolase gene from B. subtilis has been expressed in the proteobacterium Ralstonia eutropha, restoring the growth on fructose in a mutant deleted in the KDPG aldolase (Δeda mutant, Fleige et al. 2011), thus conferring the ability to metabolize hexoses.

To explore whether the expression of this pathway in S. cerevisiae could confer any advantage and to increase the carbon fluxes towards acetate, we reconstructed the phosphoketolase pathway by expressing the genes xpkA and ack from A. nidulans. In order to encourage the yeast to use the phosphoketolase pathway, we manipulated the NADPH pool through the overexpression of the cytosolic transhydrogenase gene sth (Chung 1970). Nonmembrane transhydrogenases are a class of enzymes that lead to increased NADH generation from NADPH and should thus force the carbon flux towards the PPP. A similar strategy, based on increasing the NADPH demand to force the flux through the PPP, was used by Toivari et al. (2010). In our work we chose to clone the gene from Azotobacter vinelandii as this was previously cloned in S. cerevisiae (Nissen et al. 2001), leading to an increased NADH generation from NADPH. Additionally, to sustain growth, we overexpressed the yeast endogenous phosphoenolpyruvate carboxykinase (PCK1, Valdés-Hevia et al. 1989), responsible for the conversion of oxaloacetate to phosphoenolpyruvate and having a major anaplerotic role (Zelle et al. 2010). As phosphoenolpyruvate is the main precursor for the biosynthesis of aromatic amino acids, we believed that overexpressing the gene coding for the enzyme catalyzing this main anaplerotic reaction could help the growth of the recombinant strain. The four genes were cloned on one single multicopy plasmid, under the control of two different glucose-induced promoters.

The physiology of the recombinant strain was characterized through batch and chemostat cultivations. The main physiological effects of PHK expression during batch cultivations are an increase in the acetate yield (YsAc) as well as a decrease in the biomass yield (Ysx).

In order to determine whether yeast can actively use this pathway or whether the minor phenotypic changes observed were due to an indirect effect of the expression of the pathway, we resolved the intracellular flux distribution based on 13C labeling. Metabolic network analysis through feeding of 13C-labeled substrate (Dauner and Sauer 2000) has been extensively used in metabolic engineering to gain insight into the metabolic state of the yeast cell (Zamboni and Sauer 2009). Flux distribution showed the active role of the PHK pathway in the S. cerevisiae strain expressing the recombinant route, thus confirming the ability of yeast to use this pathway together with glycolysis for glucose assimilation. Our work showed that the recombinant strain is potentially very attractive for the engineering of an acetate-overproducing strain. Acetate can be converted into acetyl CoA, which is the precursor for the biosynthesis of a wide range of industrially relevant compounds, e.g., 1-butanol, polyketides, polyhydroxybutyrate, isoprenoids, fatty acids, and thereof derived products.

Materials and methods

Strains and storage

The S. cerevisiae strains used in this study were CEN.PK113-7D (MATa ura3-52MAL2-8cSUC2; provided by P. Kötter, University of Frankfurt, Germany) and CEN.PK 113-5D (MATa MAL2-8cSUC2 ura3-52; P. Kötter, Frankfurt, Germany). The strains were maintained on YPD (yeast extract-peptone-dextrose) plates containing 10 g L−1 yeast extract, 20 g L−1 casein peptone, 20 g L−1 glucose, and 20 g L−1 agar. Plasmid-carrying strains were selected on synthetic dextrose agar containing 6.9 g L−1 yeast nitrogen base without amino acids (Formedium, Hunstanton, UK), 0.77 g L−1 complete supplement mixture without uracil (MP Biomedicals, Solon, OH, USA), 20 g L−1 glucose, and 20 g L−1 agar.

Plasmids and recombinant strain construction

The genes used in this study are listed in Table 1. The heterologous genes (xpkA, ack, and sth) were codon optimized for optimal expression in S. cerevisiae and synthesized by GenScript (Psicataway, NJ) starting from protein sequence. The amino acidic sequences used for optimization can be found using the accession numbers provided in the table, except for the PCK1 gene, since as it is endogenous and was not codon optimized but amplified from the genomic S. cerevisiae DNA.
Table 1

Genes expressed with respective microorganisms source and accession number to identify the amino acidic sequence used for codon optimization

Gene

Gene product

Organism

Accession number

xpkA

Phosphoketolase

A. nidulans

AN4913.3, FGSC_A4:578145-580928W

ack

Acetate kinase

A. nidulans

XP_662518.1

sth

Transhydrogenase

A. vinelandii

YP_002798658.1

PCK1

Phosphoenolpyruvate carboxykinase

S. cerevisiae

SGD S000001805

For the endogenous PCK1, the NCBI entry is shown

In order to clone the four different genes into the plasmid pSP-G2 (Partow et al. 2010), the following strategy was applied. Two of the genes (xpK and either ack) were cloned using the restrictions enzymes BglII and SpeI for xpkA and the restriction enzymes BamHI and KpnI for ack, leading to the formation of the vector A (pMPa2).

The other two genes, the endogenous S. cerevisiae PCK1 gene and the transhydrogenase gene sth, were cloned into pSP-G2 using, respectively, the restriction enzymes SacI and SpeI for PCK1 and BamHI and HindIII for sth, thus generating vector B. From vector B, the cassette containing the genes PCK1 and sth (with their respective promoters and terminators) was amplified using primers with restriction sites for the enzyme EcoRI and cloned using the MfeI site (generating sites compatible with EcoRI) into pMPa2 containing the genes xpkA and ack. The resulting vector contains the genes xpkA, ack, and sth, and PCK1 was named pMPa and is represented in Fig. 1.
Fig. 1

The plasmid pMPa carrying the genes xpKA and ack, sth, and PCK1 four genes mentioned in Table 1 and has a bidirectional promoter PPGK1 and PTEF1

All restriction enzymes, ligases, and phosphatase were purchased from Fermentas International Inc. (Burlington, Canada) or from New England BioLabs (Ipswich, MA). For PCRs, the Phusion polymerase (Finnzymes, Vantaa, Finland) was used (http://www.neb.com/nebecomm/ManualFiles/manualF-530.pdf). All the gene sequences were confirmed by sequencing. The resulting vector pMPa (xpk-ack-sth-PCK1) was transformed into the yeast strain CEN.PK 113-5D (MATa MAL2-8c SUC2 ura3-52; Kötter P. Frankfurt, Germany) using the lithium acetate method (Kawai et al. 2010).

Media and growth conditions

For batch cultivations, a previously described (Verduyn et al. 1992) mineral salts medium was used, commonly called CBS, having the following composition (per liter): (NH4)2SO4, 5 g; KH2PO4, 3 g; MgSO4·7H2O, 0.50 g; Antifoam 289 (A-5551; Sigma-Aldrich, St. Louis, MO), 0.050 mL; trace metals, 1 mL; and vitamins, 1 mL. The trace metal solution consisted of the following (per liter): EDTA (sodium salt), 15.0 g; ZnSO4·7H2O, 0.45 g; MnCl2·2H2O, 1 g; CoCl2·6H2O, 0.3 g; CuSO4·5H2O, 0.3 g; Na2MoO4·2H2O, 0.4 g; CaCl2·2H2O, 0.45 g; FeSO4·7H2O, 0.3 g; H3BO3, 0.1 g; and KI, 0.1 g. The pH of the trace metal solution was adjusted to 4.0 with 2 M NaOH prior to heat sterilization. The vitamin solution contained (per liter): biotin, 0.05 g; p-amino benzoic acid, 0.2 g; nicotinic acid, 1 g; Ca pantothenate, 1 g; pyridoxine HCl, 1 g; thiamine HCl, 1 g; and myo-inositol, 25 g. The pH of the vitamin solution was adjusted to 6.5 with 2 M NaOH. The vitamin solution was filter sterilized and stored at 4°C. This medium was supplemented with 20 g L−1 glucose. The medium used for shake flask cultivations had the same composition as described above, but the pH was adjusted to 6.5 prior autoclaving. For chemostat cultivations the feed composition was the same as described above but the final glucose concentration was 10 g L−1.

Batch cultivations

Aerobic batch cultures were performed in 1.0-L DasGip stirrer-pro® vessels with a working volume of 0.7 L. Agitation was maintained at 600 rpm using a magnetic stirrer integrated in the BioBlock®, which maintained the temperature at 30°C. The aeration was set to 0.5 L min−1. The pH of the medium was maintained at 5.0 by automatic addition of KOH 2 N. The temperature, agitation, gassing, pH, and composition of the off gas were monitored and controlled using the DasGip monitoring and control system. Dissolved oxygen was monitored with an autoclavable polarographic oxygen electrode (Mettler Toledo, Columbus, OH, USA). The effluent gas from the fermentation was analyzed for real-time determination of O2 and CO2 concentration by DasGip fedbatch pro® gas analysis systems with the off gas analyzer GA4 based on zirconium dioxide and two-beam infrared sensor.

Chemostat cultivations

The aerobic chemostat cultures were initiated only after the residual ethanol produced from the glucose consumption phase was completely depleted. During the chemostat cultivations, the medium described above was fed with a constant dilution rate of 0.01 h−1 and aeration was set to 0.5 L min−1. The working volume was kept at 0.5 L by a peristaltic effluent pump, and the pH was kept at a value of 5 by adding KOH 2 N. The exact volume was measured at the end of each cultivation. Samples were taken after a steady state (defined by constant values of CO2 and O2 in the off gas, as well as a constant biomass concentration for at least five residence periods) was achieved.

Shake flask cultivations

Cultivations were carried out in 500-mL baffled Erlenmeyer flasks with four diametrically opposite baffles. The flasks were prepared with 100 mL medium as described above. Cultures were incubated with agitation in an orbital shaker at 100 rpm, and the temperature was controlled at 30°C.

Inoculum preparation

The seed cultures for the cultivations were grown at 30°C in 500-mL shake flasks containing 100 mL of culture with agitation in an orbital shaker at 100 rpm. Precultures were used to inoculate the fermenters to a final OD600 nm of 0.05. All cultivations were performed in triplicates.

Analytical methods

Cell dry weight was measured by filtering a known volume of the culture through a predried and preweighed 0.45-μm pore size nitrocellulose filter (Supor®-450 Membrane Filters; PALL Life Sciences, Ann Arbor, MI, USA). The filters with the biomass were washed with water, dried for 15 min in a microwave oven at 150 W, and weighed again. The optical density was determined at OD600 nm using a Hitachi U-1100 spectrophotometer.

Concentrations of glycerol, ethanol, acetate, succinate, and pyruvate were analyzed by an isocratic HPLC (UltiMate® 3000 Nano/Capillary Autosamplers, Dionex) with an Aminex HPX-87H ion exchange column (Bio-Rad, Hercules, USA) at 65°C using 5 mM H2SO4 as mobile phase at a flow rate of 0.6 mL min−1. Glucose, glycerol and ethanol were measured with a refraction index detector and succinate, acetate and pyruvate with an ultraviolet–visible light absorbance detector.

Measurement of fractional labeling of intracellular metabolites

All labeling experiments were carried out during chemostat cultivations, samples were withdrawn after isotopic and metabolic steady state was assumed (Gombert et al. 2001). Upon reaching a metabolic steady state, the original medium containing 2 g L−1 naturally labeled glucose was replaced by chemically identical medium, but where the glucose was replaced by 100% d-glucose-1-13C (13C >99%; Isotec/Sigma-Aldrich).

In order to determine the labeling pattern of proteinogenic amino acids, the cells were harvested by centrifugation at an OD600 nm of 1 and the residual medium was removed by washing the pellet with water. Cell protein was hydrolyzed for 16–20 h in an oven at 105°C in 6 M HCl and dried in a heating block at 85°C for 6–8 h. Derivatization of amino acids was performed according to Zamboni et al. (2009) by resuspending the dry hydrolysate in 20 mL of dimethylformamide and derivatized by adding 20 mL of N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide with 1% tert-butyldimethylchlorosilane, Sigma-Aldrich. The mixture was incubated at 85°C for 1 hour before being analyzed on the GC-MS. The samples were analyzed by gas chromatography-mass spectrometry (GC-MS) with a DSQ II single quadrupole mass spectrometer (Thermo Scientific, Waltham, MA). Analytes from 1 μL sample were separated on an SLB-5 MS capillary column (15 m, 0.25 mm i.d., 0.25 μm film thickness; Supelco, Bellefonte, PA, USA) using helium as carrier gas at a flow rate of 1.5 mL min−1. A split/splitless injector was used in the split mode 1:10. The initial oven temperature was 160°C. The oven temperature was increased to 310°C after 1 min with a rate of 20°C/min and held for 0.5 min at 310°C. Full mass spectra were generated by scanning the m/z range within 180–550 (Zamboni et al. 2009; Villas-Boas et al. 2005).

Calculation of summed fractional labeling and metabolic flux distribution analysis

The mass isotopomer intensities of each amino acid fragment were extracted from the Xcalibur software (Thermo Scientific) and successively corrected for natural labeling abundance following the method previously described (van Winden et al. 2002; Wittmann 2007). The summed fractional labeling (SFL) of each amino acid fragments was calculated following the formula described by Gombert et al. (2001). The selected SFLs were used as inputs for mathematical modeling to quantify the fluxes in the central carbon metabolism of S. cerevisiae based on Gombert et al. (2001) and Gombert and Nielsen (2000). The fluxes were estimated using the calculations described in Grotkjaer et al. (2005) and Wiechert (2001). The calculations were performed using an in-house developed package for 13C-based flux analysis in MATLAB available at the BioMet Toolbox (http://129.16.106.142/). The model used in this work is shown in the Supplementary 1 file, indicating the reactions considered and the C atoms involved.

Results

Expression of the A. nidulans phosphoketolase pathway in S. cerevisiae

The complete A. nidulans phosphoketolase pathway (xylulose-5-phosphate to acetate) was installed in a S. cerevisiae strain by expressing the genes xpkA and ack. To try to force the carbon flux to the PPP, we expressed from the same vector the A. vinelandii transhydrogenase (sth gene) and overexpressed the endogenous phosphoenolpyruvate carboxykinase, encoded by the PCK1 gene, to attempt to sustain amino acid precursor formation. The vector carrying the four activities and named pMPa is shown in Fig. 1. Transforming S. cerevisiae CEN.PK 113-5D with pMPa generated the strain MP003, expressing the recombinant pathway represented in Fig. 2. The overexpressed recombinant activities are shown in bold: Fig. 2(B) is the phosphoketolase (xpkA gene), Fig. 2(A) the transhydrogenase (sth gene), Fig. 2(C) the acetate kinase (ack), and Fig. 2(D) represents the phosphoenolpyruvate carboxykinase (PCK1 gene).
Fig. 2

Simplified metabolic network of the recombinant strain MP003 expressing the phosphoketolase pathway. Bold arrows represent the overexpressed genes xpkA (B), ack (C), sth (A), and PCK1 (D)

Physiology of the recombinant strain MP003 during batch cultivations

In order to investigate the effects of the expression of the phosphoketolase pathway in S. cerevisiae and to compare its physiology to the wild type, we performed aerobic batch cultivations. The growth curves of both strains are shown in Fig. 3 (a for wild type and b for MP003). Under these conditions both the recombinant strain MP003 and the wild-type strain show fermentative metabolism as ethanol was produced in connection with glucose consumption. The main physiological parameters such as glucose consumption rate and yields on substrate (glucose) of the main metabolites are shown in Table 2. The specific growth rate (μmax) of MP003 is reduced, very likely due to the overexpression of recombinant and endogenous activities; in agreement with this hypothesis, the lag phase was found to be longer. The specific glucose uptake rate is also reduced, as this parameter is dependent on the growth rate. The biomass yield Ysx is lower than the wild-type strain, instead the acetate yield on substrate YsAc was found to be higher for the strain MP003. Ethanol yield on substrate YsEt does not report a significant difference between the two strains. Overall, the yield coefficients under batch conditions of the recombinant strain MP003 do not show significant differences with the wild type and the main differences observed are a reduction in the specific growth rate and a small increase in the acetate yield on glucose. The similarities in the metabolite pattern between the strains do, intuitively, suggest a partial or nonutilization of the pathway. To address this question and to verify the functionality of the PHK pathway, further physiological analysis through metabolic flux distribution was performed.
Fig. 3

Growth curve of the reference strain CEN.PK 113-7D (a) and the recombinant strain MP003 (b) during aerobic batch cultivation on minimal medium with glucose 20 g L−1 as carbon source

Table 2

Physiological parameters of the reference strain CEN.PK 113-7D and the recombinant strain MP003 expressing the PHK pathway

Strain

μmax (h−1)

Ysx (g/g Glc)

r Glc (g/g DCW/h)

YsEtOH (C mol/C mol Glc)

YsAc (C mol/C mol Glc)

WT

0.410

0.17

2.72

0.45

0.016

Sdev

0.002

0.013

0.065

0.005

0.006

MP003

0.330

0.15

2.60

0.46

0.019

Sdev

0.019

0.008

0.090

0.006

0.001

The values reported are calculated for the exponential phase during aerobic batch cultivations on minimal media with glucose 2% as carbon source

Ysx biomass yield on substrate (glucose), rGlc specific glucose consumption rate, YsEtOH yield of ethanol on substrate (glucose), YsAc yield of acetate on substrate (glucose), Sdev standard deviation

Physiology during chemostat cultivations and flux distribution based on 13C labeling

Aerobic chemostat cultivations were performed at a dilution rate of 0.1 h−1. Under this condition both the strains show a fully respiratory metabolism as no formation of fermentation products was reported. The full respiratory metabolism is also confirmed by the respiratory quotient RQ value (data not shown), which is always below 1. As shown in Table 3, biomass yield Ysx is found to be higher for the wild type.
Table 3

Biomass yield on substrate (Ysx) with standard deviation (Sdev) during chemostat cultivations in minimal medium operated at dilution rate 0.1 h−1

Strain

D (h−1)

Ysx (g/g Glc)

WT

0.11

0.46

Sdev

0.001

0.013

MP003

0.10

0.41

Sdev

0.001

0.02

To analyze the flux distribution, the carbon source in the media was switched to 13C-labeled glucose and samples were taken after metabolic and isotopic steady states were reached (five residence times). Fluxes were analyzed in chemostat cultivations because the relative flux through the PPP for this condition is generally higher than in batch cultures, leading to a probably higher impact of the phosphoketolase flux, which is directly linked to the PPP.

In Table 4 the measured and the calculated SFLs of precursor amino acids for the wild-type and MP003 strain are shown. The fluxes, based on the labeling patterns of proteinogenic amino acids, are net fluxes, normalized with respect to the glucose uptake rate (100). Figure 4 shows the flux distribution of the two strains (wild type in normal font and MP003 in bold), showing flux values through each reaction. This flux is represented in the model as a route from xylulose-5-phosphate to acetate (two steps pathway based on the activity of xpkA and ack gene products), and the quantification was made based on the method described by Thykaer and Nielsen (2007). This flux shows a value of 9.16 in MP003 and 0 in the wild type, thus proving the native incapability of a wild-type S. cerevisiae strain to possess similar reactions. The positive value found for this flux indicates the ability of the recombinant strain to channel the carbon through this pathway, therefore proving the activity of the route in the recombinant strain MP003. From flux distribution it is possible to confirm that the recombinant strain MP003 is able to use, at least partially, the phosphoketolase pathway. Interestingly, the flux towards the PPP (glucose-6-phosphate to xylulose-5-phosphate) is lower for MP003 compared to the wild type. This flux shows a value of 32.44 in the recombinant strain MP003, whereas in the wild type in agreement with previous results (Christensen et al. 2002) has a value of 44.55. The flux through the PCK1 reaction (the overexpressed activity going from oxaloacetate to phosphoenolpyruvate) is not represented in Fig. 4. It was previously shown for Corynebacterium spp. that since phosphoenolpyruvate and pyruvate show identical fractional enrichment (Marx et al. 1996; Petersen et al. 2001), it is therefore difficult to resolve this flux only using labeled glucose as carbon source. The recombinant strain also shows a slightly higher flux through the TCA cycle and a lower flux through the pyruvate node towards acetate formation.
Table 4

SFL (measured and calculated) of the derivatized fragments for reference strain CEN.PK 113-7D and the recombinant strain MP003

Fragment

Ion type

C atoms in measured fragment

Precursor

CENPK-113-7D

MP003

Measured

Calculated

Measured

Calculated

Ala260

M-57

1,2,3

PYR

35.57

35.41

37.31

34.92

Ala232

M-85

2,3

PYR

32.65

34.31

34.43

33.41

Gly288

M-15

1,2

G3P

4.55

5.70

5.52

5.69

Gly246

M-57

1,2

G3P

6.50

5.70

5.59

5.69

Gly218

M-85

2

G3P

4.01

3.23

3.26

3.04

Val288

M-57

1,2,3,4,5

PYR

73.29

79.21

73.94

78.65

Val260

M-85

2,3,4,5

PYR

70.16

75.17

70.74

73.50

Val302

f302

1,2

PYR

20.99

14.83

13.36

12.44

Leu344

M-15

1,2,3,4,5,6

PYR+AcCOA

140.65

112.75

139.62

110.25

Leu274

M-85

2,3,4,5,6

PYR+AcCOA

102.21

103.96

103.19

102.96

Leu200

M-159

2,3,4,5,6

PYR+AcCOA

104.80

103.96

104.07

102.96

Ile344

M-15

1,2,3,4,5,6

OAA+PYR

101.88

100.12

88.46

95.30

Ile274

M-85

2,3,4,5,6

OAA+PYR

100.28

94.08

87.63

91.53

Ile200

M-159

2,3,4,5,6

OAA+PYR

103.59

94.08

89.04

91.53

Pro328

M-15

1,2,3,4,5

AKG

94.66

94.65

95.80

95.26

Pro286

M-57

1,2,3,4,5

AKG

99.26

94.65

98.88

95.26

Pro258

M-85

2,3,4,5

AKG

84.29

79.62

84.80

80.59

Pro184

M-159

2,3,4,5

AKG

86.29

79.62

87.26

80.59

Ser432

M-15

1,2,3

G3P

31.91

35.41

32.80

34.92

Ser390

M-57

1,2,3

G3P

32.77

35.41

29.60

34.92

Ser362

M-85

2,3

G3P

28.20

34.31

29.49

33.41

Ser288

M-159

2,3

G3P

29.61

34.31

30.50

33.41

Ser302

f302

1,2

G3P

2.40

2.20

3.47

2.61

Thr446

M-15

1,2,3,4

OAA

57.31

53.25

Thr404

M-57

1,2,3,4

OAA

56.15

55.61

Thr376

M-85

2,3,4

OAA

49.78

49.57

Phe336

M-57

1,2,3,4,5,6,7,8,9

E4P+PEP

86.69

94.51

Phe308

M-85

2,3,4,5,6,7,8,9

E4P+PEP

85.53

86.69

94.88

94.00

Phe234

M-159

2,3,4,5,6,7,8,9

E4P+PEP

87.91

86.69

92.92

94.00

Phe302

f302

1,2

PEP

2.10

2.20

2.90

2.61

Asp460

M-15

1,2,3,4

OAA

61.05

62.54

57.86

58.55

Asp390

M-85

2,3,4

OAA

53.97

56.50

50.41

54.78

Asp316

M-159

2,3,4

OAA

54.43

56.50

53.79

54.78

Asp302

f302

1,2

OAA

15.10

14.84

10.56

8.72

Glu474

M-15

1,2,3,4,5

AKG

96.54

94.65

96.12

95.26

Glu432

M-57

1,2,3,4,5

AKG

98.40

94.65

98.61

95.26

Glu404

M-85

2,3,4,5

AKG

84.34

80.62

Glu330

M-159

2,3,4,5

AKG

67.91

67.84

Glu302

f302

1,2

AKG

28.84

37.83

26.69

38.09

Lys431

M-57

1,2,3,4,5,6

AKG+AcCOAm

120.35

123.12

Lys329

M-159

2,3,4,5,6

AKG+AcCOAm

118.89

112.10

105.21

107.69

The SFL values from the measurement were reported as the mean from triplicate experiments with standard deviation less than 5%. The nonreported values for caculted SFL indicate that those fragment were not used for the metabolic flux analysis

Fig. 4

Flux distribution of the wild-type CEN.PK 113-7D and the recombinant strain MP003 (in bold) based 13C labeling. The fluxes shown are net fluxes. Abbreviations: G6P glucose-6-phosphate, F6P fructose-6-phosphate, G3P glyceraldehyde-3-phosphate, PEP phosphoenolpyruvate, PYR pyruvate, ACA acetaldehyde, ACE acetate, ICIT isocitrate, AcCoA acetyl CoA, AKG α-ketoglutarate, FUM fumarate, OAA oxaloacetate, MAL malate, GLY glycine, SER serine, P5P pentose-5-phosphate, E4P erythrose-4-phosphate, S7P sedoheptulose-7-phosphate, CO carbon dioxide

Discussion

In our work, we expressed in S. cerevisiae the A. nidulans phosphoketolase pathway leading to the generation of acetate from the PPP intermediate xylulose-5-phosphate. To our knowledge, this is the first time that the complete fungal phosphoketolase pathway is installed in yeast to channel the carbon flow from the PPP pathway to the TCA and glyoxylate shunt.

To sustain growth and to increase the carbon flux towards the PPP, we introduced modification in the NADPH metabolism using the bacterial transhydrogenase (sth) and overexpressed the endogenous PCK1 gene. There is no direct proof of activity of these enzymes in the recombinant strain, however we could attribute the different split ration of the PPP from the glycolytic intermediate glucose-6-phosphate to a difference in cofactor imbalance probably caused by the activity of the transhydrogenase (Nissen et al. 2001).

In a study aimed at increasing ethanol yield on xylose, Sonderegger et al. (2004) expressed the Bifidobacterium lactis phosphoketolase gene in a S. cerevisiae strain engineered for xylose fermentation, showing an increased acetate formation. To reduce acetate formation and improve xylose catabolism, they coexpressed the B. subtilis phosphotransacetylase and the Entamoeba histolytica acetaldehyde dehydrogenase, which resulted in increased ethanol yield on xylose under anaerobic conditions. Phosphotransacetylases (Shin et al. 1999) are a class of bacterial enzymes that convert acetylphosphate into acetyl CoA and can therefore represent a valid alternative to the fungal pathway that uses acetate kinase, converting acetylphosphate into acetate, described in this work. We also evaluated the functionality of the phosphoketolase pathway expressing, in the second step, the B. subtilis phosphotransacetylase gene pta alternatively to the acetate kinase. However, in this case, we found no flux through the phosphoketolase pathway during growth on glucose (data not shown), probably due to the differences in thermodynamic driving forces between the two pathways.

Instead of introducing in yeast different vectors, we decided to express all the genes required from one single vector. The growth of MP003 is affected by the presence of the vector with the different recombinant activities as lower specific growth rate (μmax) as well as a longer lag phase were observed, probably in light of metabolic burden caused by the synthesis of the several recombinant activities. The biomass yield on substrate (Ysx) is, however, only minimally affected for MP003, where it is found to be lower. A slightly lower biomass yield is reported also during chemostat cultivations. The glucose consumption rate of MP003 (rGlc) is reduced, but as this parameter is dependent on the specific growth rate, this does not mean that the strain is not able to utilize glucose with the same efficiency. The acetate yield (YsAc) is moderately increased, suggesting an active role of the PHK pathway expression. As no striking physiological differences were found for the two strains, we were uncertain as to whether MP003 was using the installed route or if the changes in YsAc were due to an indirect effect of the expression of the pathway. To address this question and gain insight into the effective utilization of this route, we decided to resolve the intracellular flux distribution based on 13C labeling. The analysis of flux distribution allowed us to clarify the functionality of the recombinant route in the strain MP003. Flux distribution was determined during chemostat cultivations and demonstrated the activity of the phosphoketolase pathway, which reports a positive value (9 in MP003 and 0 in the wild type). The fact that the flux from pyruvate to acetate through acetaldehyde was found to be lower in MP003 can represent an additional indication of the activity of the PHK pathway as the lower fluxes from pyruvate to acetate made it unlikely that acetate is formed from pyruvate but could instead suggest that the acetate generated is formed through the use of the PHK pathway, despite no acetate secretion was detected during chemostat cultivations. The surplus of intracellular acetate generated from the PHK pathway could be responsible for the minor increase observed in the acetate yield during batch cultivations nevertheless could also furnish an explanation for the higher fluxes through the first steps of the TCA cycle, as glyoxylate (which unfortunately cannot be resolved with this method) and TCA cycle could be fueled by the increased levels of acetate/acetyl CoA.

The utilization of the PHK pathway does not interfere with glucose assimilation through the EMP pathway as the flux through glycolysis is not substantially changed, except for a higher flux through the phosphoglucoisomerase to balance the lower flux to xylulose-5-phosphate at the PPP branch point. The higher flux from glucose-6-phosphate to the PPP in the wild type is in agreement with the slightly higher biomass yield on substrate during chemostat cultivations.

The recombinant strain was characterized through cultivations in different modes, not showing a substantially different phenotype from the wild type, except for a slightly higher acetate yield. No remarkable physiological effects on overflow metabolism were observed, probably due to the fact that the MP003 strain did not show, based on the 13C labeling, differences in the glycolytic flux. The work presented here gives evidence of the utilization of the PHK route in S. cerevisiae during chemostat cultivation. This route contributes to increase the acetate supply during batch cultivations. As acetate can be naturally converted into acetyl CoA, this strategy is of potential interest for the biosynthesis of chemical compounds having acetyl CoA as precursor for their biosynthesis, such as isoprenoids, polyhydroxybutyrate, and polyketides.

It would be extremely interesting to observe the effects of the expression of this route in a mutant with nonfunctional glycolytic pathway in order to establish whether the introduction of the phosphoketolase route would restore glucose catabolism, as illustrated recently in R. eutropha (Fleige et al. 2011). Glycolysis is an essential pathway, and, as shown in several works, deletion of genes encoding glycolytic enzymes shows growth defect on glucose (Ciriacy and Breitenbach 1979; Papini et al. 2010; Compagno et al. 1996; Lam and Marmur 1979; Sprague 1977). As our work demonstrated the activity of this pathway, this strategy holds relevance for future works aiming at bypassing glycolysis and establishing the phosphoketolase pathway as stand-alone route to push the carbon flux to the glyoxylate shunt and to the TCA cycle through the pentose phosphate pathway and hereby avoid overflow metabolism towards ethanol.

Acknowledgments

The authors acknowledge the Knut and Wallenberg foundation, the Chalmers foundation, and the European Research Council. Saeed Shoaie is acknowledged for the technical support during flux calculations.

Supplementary material

253_2012_3936_MOESM1_ESM.docx (13 kb)
Supplementary 1DOCX 13 kb

Copyright information

© Springer-Verlag 2012

Authors and Affiliations

  • Marta Papini
    • 1
  • Intawat Nookaew
    • 1
  • Verena Siewers
    • 1
  • Jens Nielsen
    • 1
  1. 1.Department of Chemical and Biological EngineeringChalmers University of TechnologyGothenburgSweden

Personalised recommendations