Improving Product Yields on D-Glucose in Escherichia coli via Knockout of pgi and zwf and Feeding of Supplemental Carbon Sources

Abstract

The use of lignocellulosic biomass as a feedstock for microbial fermentation processes presents an opportunity for increasing the yield of bioproducts derived directly from glucose. Lignocellulosic biomass consists of several fermentable sugars, including glucose, xylose, and arabinose. In this study, we investigate the ability of an E. coli Δpgi Δzwf mutant to consume alternative carbon sources (xylose, arabinose, and glycerol) for growth while reserving glucose for product formation. Deletion of pgi and zwf was found to eliminate catabolite repression as well as the ability of E. coli to consume glucose for biomass formation. In addition, the yield from glucose of the bioproduct D-glucaric acid was significantly increased in a Δpgi Δzwf strain.

1. Introduction

In recent years, concerns over declining petroleum reserves and climate change due to atmospheric carbon dioxide accumulation have spurred significant interest in using alternative feedstocks for the manufacture of petroleum-derived products. Because it is renewable and abundant, non-food plant (lignocellulosic) biomass represents a promising alternative feedstock to crude oil. Moreover, because plants assimilate carbon dioxide during growth, the use of plant-based feedstocks could potentially slow the accumulation of carbon dioxide in the atmosphere. Recent research has focused heavily on the identification of ideal plant biomass feedstocks (Joyce and Stewart, 2012), the determination of commercially valuable, biomass-derived products (Werpy and Petersen, 2004), and the development of processes for converting plant biomass into these products of interest.

One such process that has received heavy attention is microbial fermentation. A significant portion of lignocellulosic biomass consists of fermentable sugars such as glucose, xylose, arabinose, and galactose (Joyce and Stewart, 2012), and many microbes are naturally able to convert these sugars into products of interest such as biochemicals and biopolymers (Lee, 1996). In addition, microbes can be engineered to produce a wide array of non-natural products via recombinant DNA technology (Curran and Alper, 2012). A few products, including D-glucaric acid (Moon et al., 2009) and D-gluconic acid (Rogers et al., 2006), are derived directly from glucose; however, microbial production of these products generally suffers from low yields, as a portion of the glucose feed is utilized to generate cell biomass. For products such as these that do not require further glucose metabolism through the canonical reduction pathways, eliminating the ability to utilize glucose for growth may address this limitation. We thus set out to design an E. coli production platform that utilizes an alternative carbon source such as arabinose or xylose for cell growth, reserving glucose solely for product generation to maximize yield.

The main pathways for glucose utilization in E. coli are depicted in Figure 1. Glucose enters the cell through the phosphotransferase system (PTS) encoded by ptsG and ptsHI-crr and is phosphorylated to glucose-6-phosphate in the process. Glucose-6-phosphate can then proceed through the Entner-Dudoroff Pathway via zwf or through the Embden-Meyerhoff-Parnas Pathway via pgi. The carbon in glucose-6-phosphate can also be stored as glycogen via pgm. Previous work to engineer E. coli for co-utilization of glucose and alternative carbon sources involved deleting the PTS system (Balderas-Hernández et al., 2011; Solomon et al., 2013; Wang et al., 2011). By eliminating the PTS system, catabolite repression can be eliminated, allowing simultaneous uptake of glucose and a secondary carbon source. However, these strategies also eliminate E. coli’s primary method of glucose uptake, and the cell must rely on nonspecific transporters to import glucose into the cell. Subsequent phosphorylation of glucose to glucose-6-phosphate via ATP-dependent glk is also required for glucose metabolism in PTS-deficient E. coli. Overexpression of the galactose:H⁺ symporter galP and upregulation of glk has been shown to recover wild-type growth rates in PTS-deficient strains of E. coli (Hernández-Montalvo et al., 2003). However, because glycolytic pathways remain intact in this strain, it is likely that product yields on glucose would remain low. In this study, we explore the behavior of an E. coli strain which lacks pgi and zwf and investigate this strain’s ability to produce a glucose-derived product when supplemented with L-arabinose and D-xylose, sugars which are readily available from biomass. Glycerol is also explored as a carbon source, as its price has dropped significantly in recent years due to significant increases in biodiesel production (Johnson and Taconi, 2009). Previous work has demonstrated improved yield of glucose-derived products in a Δpgi Δzwf strain supplemented with mannitol (Kogure et al., 2007; Pandey et al., 2013); however, the price of mannitol remains high relative to glycerol and biomass-derived sugars.

Figure 1.

Glucose utilization pathways in E. coli.

Materials and Methods

2.1 E. coli strains and plasmids

E. coli strains, plasmids, and oligonucleotides used in this study are listed in Table I. All molecular biology manipulations were performed according to standard practices (Sambrook and Russell, 2001). E. coli DH10B was used for transformation of cloning reactions and propagation of all plasmids. Strains M2, M2-2, and M3 were constructed by our group previously (Gonçalves et al., 2013; Shiue and Prather, 2014). Strain M2 was generated via knockout of endA and recA from E. coli MG1655. Strain M2-2 was generated via knockout of gudD and uxaC from strain M2 to prevent consumption of D-glucaric and D-glucuronic acids during D-glucaric acid production experiments. Strain M3 was derived from strain M2 via knockout of pgi and served as an intermediate strain; this strain was not further characterized in this work. Deletion of zwf from strain M3 was achieved by P1 transduction with Keio collection strain JW1841-1 as the donor (Baba et al., 2006). The λDE3 lysogen was then integrated site-specifically into this quadruple knockout strain using a λDE3 Lysogenization Kit (Novagen, Darmstadt, Germany), generating strain M4 (MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf. To prevent E. coli from consuming D-glucuronic and D-glucaric acids, both gudD and uxaC were deleted from the genome. Deletion of uxaC was performed with λ-Red mediated recombination (Datsenko and Wanner, 2000) using pKD46recA (Solomon et al., 2013). PCR primers pKD13_uxaC_fwd and pKD13_uxaC_rev (Table 1) were used to amplify the recombination cassette from pKD13 (Datsenko and Wanner, 2000), and strain M4 harboring pKD46recA was transformed with this PCR product. The kan selection cassette was cured from successful deletion mutants using FLP recombinase expressed from pCP20, generating strain M5. Similar to strain M3, strain M5 served as an intermediate strain only and was not further characterized in this work. Finally, strain M6 (MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf ΔuxaC ΔgudD) was generated using the same λ-Red mediated recombination method described above; in this case, primers pKD13_gudD_fwd and pKD13_gudD_rev were used to amplify the recombination cassette from pKD13. pRSFD-IN and pTrc-Udh were constructed by our group previously (Moon et al., 2009; Yoon et al., 2009). To construct pRSFD-IN-Udh, pRSFD-IN was first digested with XhoI, end-filled with Klenow enzyme, then digested with EcoRI-compatible MfeI. pTrc-Udh was then digested with EcoRI and SmaI, and the Udh-containing fragment was ligated into digested pRSFD-IN to generate pRSFD-IN-Udh.

Table I.

E. coli strains, plasmids, and oligonucleotides used

Name	Relevant Genotype	Reference
Strains
DH10B	F− mcrA Δ(mrr-hsdRMS-mcrBC) φ80lacZΔM15 ΔlacX74 recA1 endA1 araD139 Δ(ara, leu)7697 galU galK λ-rpsL nupG	Life Technologies (Carlsbad, CA)
JW1841-1	F−, Δ(araD-araB)567, ΔlacZ4787(::rrnB-3), λ−, Δzwf-777::kan, rph-1, Δ(rhaD-rhaB)568, hsdR514	CGSC #9537 (Baba et al., 2006)
MG1655	F− λ− ilvG− frb- 50 rph-1	ATCC #700926
M2	MG1655(DE3) ΔendA ΔrecA	(Shiue and Prather, 2014)
M2-2	MG1655(DE3) ΔendA ΔrecA ΔgudD ΔuxaC	(Shiue and Prather, 2014)
M3	MG1655 ΔendA ΔrecA Δpgi	(Gonçalves et al., 2013)
M4	MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf	This study
M5	MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf ΔuxaC	This study
M6	MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf ΔuxaC ΔgudD	This study
Plasmids
pCP20	Repa, AmpR, CmR, FLP recombinase expressed by λ pr under control of λ cI857	CGSC #7629
pKD13	R6Kγ ori, AmpR, kan	CGSC #7633
pKD46	R101 ori, repA101a, AmpR, araC, araBp-λγ-λβ-λexo	CGSC #7739
pKD46recA	R101 ori, repA101a, AmpR, araC, araBp-λγ-λβ-λexo, recA	(Solomon et al., 2013)
pRSFDuet-1	pRSR1030 ori, lacI, KanR	EMD4 Biosciences (Darmstadt, Germany)
pTrc99A	pBR322 ori, AmpR	(Amann and Brosius, 1985)
pRSFD-IN	pRSFDuet-1 with INO1 inserted into the EcoRI and HindIII sites	(Moon et al., 2009)
pTrc-Udh	pTrc99A with Udh from Pseudomonas syringae inserted into the NcoI and HindIII sites	(Moon et al., 2009; Yoon et al., 2009)
pRSFD-IN-Udh	pRSFD-IN with Udh inserted into the MfeI and XhoI sites	This study
pTrc-SUMO-MIOX	pTrc99A with SUMO-MIOX	(Shiue and Prather, 2014)
Oligonucleotides
pKD13_gudD_fwd
pKD13_gudD_rev
pKD13_uxaC_fwd
pKD13_uxaC_rev

^aAll oligonucleotides purchased from Sigma-Genosys (St. Louis, MO). Homologous sequences used for recombination are underlined.

2.2 Culture conditions

For determination of growth curves, cultures were grown in 250 mL baffled shake flasks containing 50 mL LB medium supplemented with approximately 10 g/L D-glucose, L-arabinose, glycerol, and/or D-xylose as indicated in Figures 2-5. Seed cultures were grown overnight at 30°C and inoculated to an optical density at 600 nm (OD₆₀₀) of 0.005. Cultures were incubated at 30°C, 250 rpm, and 80% relative humidity for 72 hours. To construct a growth curve, cell densities were measured at regular time intervals on a DU800 Spectrophotometer (Beckman Coulter, Pasadena, CA), with more frequent sampling during the exponential growth phase. For analysis of metabolite concentrations, samples were taken daily, centrifuged to remove cell debris, and the supernatants analyzed via high performance liquid chromatography as described in Section 2.3.

Figure 2.

Growth curves for strains M2 (top row) and M4 (bottom row) in rich medium supplemented with L-arabinose (“Ara”, triangles), glycerol (“Gly”, circles), and D-xylose (“Xyl”, squares) with D-glucose (“Glc”) absent (open points) or present (filled points). M2: MG1655(DE3) ΔendA recA; M4: MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf.

Figure 5.

Growth and D-glucaric acid (“Gla”, circles) production from D-glucose in strain M6 in lean medium. Cultures contained M6 harboring pRSFD-IN-Udh and pTrc-SUMO-MIOX and were grown in modified MOPS minimal medium supplemented with 6 g/L L-arabinose (“Ara”, triangles) and 10 g/L D-glucose (“Glc”, squares). M6: MG1655(DE3) ΔendA ΔrecA ΔgudD ΔuxaC Δpgi Δzwf.

For glucaric acid production in rich medium, cultures were grown in 250 mL baffled shake flasks containing 50 mL LB medium supplemented with 10 g/L D-glucose and 10 g/L L-arabinose, 10 g/L glycerol, or 10 g/L D-xylose. Cultures were induced at inoculation with 0.1 mM β-D-1-thiogalactopyranoside (IPTG). Ampicillin (100 μg/mL) and kanamycin (30 μg/mL) were added for plasmid maintenance. Seed cultures were grown overnight at 30°C in LB medium supplemented with 10 g/L D-glucose and 10 g/L L-arabinose, 10 g/L glycerol, or 10 g/L D-xylose and inoculated to an OD₆₀₀ of 0.005. Cultures were incubated at 30°C, 250 rpm, and 80% relative humidity for 72 hours. Adequate aeration for product formation was ensured with the use of baffled shake flasks. Samples were taken daily, centrifuged to remove cell debris, and the supernatants analyzed for metabolite concentrations.

For minimal medium experiments, a modified MOPS-buffered medium was used containing 10 g/L D-glucose, 6 g/L L-arabinose, 6 g/L NH₄Cl, 0.4 g/L K₂HPO₄, 2 mM MgSO₄, 0.1 mM CaCl₂, 40 mM MOPS, 4 mM tricine, 50 mM NaCl, 100 mM Bis-Tris, 134 μM EDTA, 31 μM FeCl₃, 6.2 μM ZnCl₃, 0.76 μM CuCl₂, 0.42 μM CoCl₂, 1.62 μM H₃BO₃, 0.081 μM MnCl₂, carbenicillin (100 μg/mL), and kanamycin (30 μg/ml). Seed cultures were started using a 1:100 dilution from LB and were grown at 30°C for 48 hours in modified MOPS. Working cultures were inoculated to an OD₆₀₀ of 0.02 and induced 32 hours after inoculation with 0.1 mM IPTG. Cultures were incubated at 30°C, 250 rpm, and 80% relative humidity for 110 hours. Adequate aeration for product formation was ensured with the use of baffled shake flasks. Samples were taken periodically, centrifuged to remove cell debris, and the supernatants analyzed for metabolite concentrations.

2.3 Determination of metabolite concentrations

D-glucose, L-arabinose, glycerol, D-xylose, and D-glucaric acid were quantified from culture supernatants using high performance liquid chromatography (HPLC) on an Agilent Series 1100 or Series 1200 instrument equipped with an Aminex HPX-87H column (300 mm by 7.8 mm; Bio-Rad Laboratories, Hercules, CA). Sulfuric acid (5 mM) was used as the mobile phase at 35°C and a flow rate of 0.6 mL/min in isocratic mode. Compounds were detected and quantified from 10 μL sample injections using refractive index and diode array detectors. Reported metabolite concentrations are the average of triplicate samples, and error bars represent one standard deviation above and below the mean value.

3. Results

In E. coli, glucose is imported into the cell and phosphorylated to glucose 6-phosphate (G6P) by the phosphotransferase system (PTS). Glucose metabolism then proceeds through two routes (Figure 1): the Embden-Meyerhoff-Parnas Pathway via phosphoglucose isomerase (pgi) or the Entner-Dudoroff Pathway via glucose 6-phosphate dehydrogenase (zwf). A third route interconverts glucose 6-phosphate and glucose 1-phosphate via phosphoglucomutase (pgm) for glycogen storage and accumulation, though flux through this node is typically extremely low (Chassagnole et al., 2002). To eliminate native consumption of glucose, both pgi and zwf were deleted from an MG1655-derived strain. Growth on M9 minimal medium supplemented with various carbon sources confirmed that the Δpgi Δzwf mutant does not grow on glucose but retains the ability to utilize other carbon sources (Supplementary Figures 1 and 2).

3.1 Behavior of a Δpgi Δzwf mutant

Cell growth was compared for strains M2 and M4, a Δpgi Δzwf mutant in rich medium supplemented with various carbon sources (Figure 2). Maximum specific growth rates for each combination of strain and carbon supplement were also calculated (Table II). As expected, growth of strain M2 was similar for all conditions tested, with similar lag phases and maximum specific growth rates. Final cell densities were lower when strain M2 was fed D-glucose, likely due to increased production of acetate (Figure 3), which has been shown to inhibit cell growth (Roe et al., 1998). Growth of strain M4 was also similar for all conditions tested with the exception of glycerol-supplemented cultures, which displayed a significant lag in growth of approximately 24 hours. We hypothesize that this lag corresponds to depletion of the metabolic precursors provided by LB and a metabolic shift towards gluconeogenic metabolism for growth on glycerol. Maximum growth rate, lag time, and final cell densities are similar for strain M4 in the presence of L-arabinose, glycerol, and D-xylose regardless of whether D-glucose was supplemented in the growth medium, indicating that substrate consumption was similar in the presence or absence of D-glucose. Overall, maximum growth rates of strain M4 were approximately 70% that of strain M2.

Table II.

Maximum growth rates of various strains in rich medium supplemented with various carbon sources.

Strain*	Alternative Carbon Source	Glucose	μmax (hr−1)
M2	L-Arabinose	−	1.00 ± 0.01
		+	1.01 ± 0.06
	Glycerol	−	0.98 ± 0.02
		+	0.97 ± 0.04
	D-Xylose	−	1.00 ± 0.03
		+	0.87 ± 0.20
M4	L-Arabinose	−	0.65 ± 0.01
		+	0.66 ± 0.02
	Glycerol	−	0.66 ± 0.05
		+	0.68 ± 0.01
	D-Xylose	−	0.70 ± 0.01
		+	0.67 ± 0.01

^*M2: MG1655(DE3) ΔendA ΔrecA; M4: MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf

Figure 3.

Carbon source and acetate (“Act”, circles) concentrations in rich medium cultures of M2 (filled points, solid line) and M4 (open points, dotted lines) supplemented with L-arabinose (“Ara”, triangles), glycerol (“Gly”, inverted triangles), and D-xylose (“Xyl”, diamonds) with D-glucose (“Glc”, squares) absent (left column) or present (right column). M2: MG1655(DE3) ΔendA ΔrecA; M4: MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf.

Concentrations of D-glucose, alternative carbon source, and acetate were measured for each strain/carbon supplement combination as a function of time (Figure 3). As expected, the presence of D-glucose prevents consumption of the alternative carbon source in strain M2 via catabolite repression. In contrast, the deletion of pgi and zwf prevents consumption of D-glucose in strain M4. Interestingly, deletion of pgi and zwf appears to eliminate catabolite repression in strain M4, as the presence of D-glucose does not preclude consumption of the alternative carbon source in this strain.

This phenomenon is perhaps due to intracellular buildup of glucose-6-phosphate. Catabolite repression is mediated by cyclic AMP (cAMP), which is synthesized by adenylate cylcase. Adenylate cylcase is activated via phosphorylation by EIIA^Glc, but this phosphorylation can only occur if EIIA^Glc has a phosphate group to donate. When glucose is being actively imported through the PTS system, EIIA^Glc donates its phosphate to the incoming glucose, resulting in a mostly unphosphorylated population of EIIA^Glc, inactive adenylate cyclase, and a low concentration of cAMP. The absence or depletion of glucose from the culture medium leads to a buildup of phosphorylated EIIA^Glc, activation of adenylate cyclase, and an increase in cAMP concentration, eventually leading to the expression of catabolite-repressed genes such as araBAD and xylAB, which are important for arabinose and xylose metabolism, respectively. Thus, catabolite repression occurs when there is active glucose flux into the cell, not simply when glucose is present in the medium. In a Δpgi Δzwf mutant, the accumulation of glucose-6-phosphate quickly eliminates glucose flux into the cell, resulting in derepression of genes normally repressed in the presence of glucose.

3.2 D-glucaric acid production in a Δpgi Δzwf mutant

D-glucaric acid, a dicarboxylic organic acid, is a naturally occurring product which has been investigated for a variety of potential applications. A biosynthetic pathway to D-glucaric acid from D-glucose has been constructed in E. coli (Moon et al., 2009). This pathway begins with glucose-6-phosphate, which is converted to myo-inositol-1-phosphate by myo-inositol-1-phosphate synthase (INO1). Myo-inositol-1-phosphate is then dephosphorylated by an endogenous phosphatase to yield myo-inositol, which is oxidized to D-glucuronic acid by myo-inositol oxygenase (MIOX). Finally, D-glucuronic acid is oxidized to D-glucaric acid by uronate dehydrogenase (Udh). Because production of D-glucaric acid requires glucose-6-phosphate, we hypothesized that the yield of D-glucaric acid could be increased significantly in a Δpgi Δzwf strain.

In supplemented LB, D-glucaric acid yield on glucose is increased in the Δpgi Δzwf mutant nearly 18-fold over an unmutated control supplemented with L-arabinose or D-xylose, while yield is increased approximately 9-fold in the Δpgi Δzwf strain supplemented with glycerol (Table III). Additionally, D-glucaric acid titers are significantly higher in the Δpgi Δzwf mutant (Figure 4 and Table III). We hypothesize that deletion of pgi and zwf results in higher glucose-6-phosphate pools, allowing INO1 to operate much closer to its maximum activity, which leads to increased flux through the D-glucaric acid pathway.

Table III.

D-glucaric acid yields on D-glucose for various carbon sources.

Strain*	Carbon	D-Glucaric Acid Titer (g/L)	Yield on D-Glucose (g/g)
M2-2	L-Arabinose in LB	0.13 ± 0.01	0.044 ± 0.002
	Glycerol in LB	0.20 ± 0.02	0.052 ± 0.009
	D-Xylose in LB	0.13 ± 0.01	0.039 ± 0.002
M6	L-Arabinose in LB	0.50 ± 0.01	0.76 ± 0.13
	Glycerol in LB	0.81 ± 0.10	0.44 ± 0.04
	D-Xylose in LB	1.19 ± 0.08	0.73 ± 0.03
	L-Arabinose in MOPS	0.40 ± 0.02	0.47 ± 0.25

^*M2-2: MG1655(DE3) ΔendA ΔrecA ΔgudD ΔuxaC; M6: MG1655(DE3) ΔendA ΔrecA Δpgi Δzwf ΔuxaC ΔgudD

Figure 4.

D-glucaric acid production (“Gla”, circles) from D-glucose in strains M2-2 and M6 in rich medium. Cultures contained strain M2-2 or M6 harboring pRSFD-IN-Udh and pTrc-SUMO-MIOX and were grown in LB supplemented with D-glucose (“Glc”, squares) and L-arabinose (“Ara”, triangles), glycerol (“Gly”, inverted triangles), or D-xylose (“Xyl”, diamonds). M2-2: MG1655(DE3) ΔendA ΔrecA ΔgudD ΔuxaC (filled points, solid line); M6: MG1655(DE3) ΔendA ΔrecA ΔgudD ΔuxaC Δpgi Δzwf (open points, dotted line).

To simulate a lean medium that might be obtained from the hydrolysis of lignocellulosic biomass, the Δpgi Δzwf strain was also tested in a modified MOPS minimal medium containing 10 g/L D-glucose and 6 g/L L-arabinose. Although the strain grew more slowly under these conditions, D-glucaric acid titers of 0.40 ± 0.02 g/L were obtained, nearly as much as observed in supplemented LB (Figure 5). The yield of D-glucaric acid from glucose was 47%; however, approximately 0.2 g/L of myo-inositol was also produced, bringing the total yield of G6P-derived products to 71%. Myo-inositol is produced from G6P as an intermediate during D-glucaric acid production and has previously been observed to build up in the culture medium under some conditions (Moon et al., 2009). Modified MOPS medium containing D-xylose and D-glucose was also tested, but no growth of the Δpgi Δzwf mutant was observed, possibly due to stronger residual catabolite repression of xylAB in minimal medium. In modified MOPS medium supplemented with D-xylose alone, M6 does not produce glucaric acid or myo-inositol, consistent with the expectation that glucaric acid can only be derived from glucose in a Δpgi Δzwf mutant (Supplementary Table 1).

4. Discussion

Traditionally, the main focus of metabolic engineering projects has been on increasing the final titer of a product of interest, and this approach has been widely successful for high-value compounds such as pharmaceutical intermediates and therapeutic proteins. However, increasing titers alone may not be sufficient for low-margin, high-volume bioproducts such as commodity chemicals. In these cases, product yield becomes an important process consideration, as raw material costs can be a large percentage of the manufacturing costs. Strategies which are able to increase product yield without sacrificing titer would be valuable tools for the metabolic engineer.

The use of renewable feedstocks such as lignocelluosic biomass for biochemical production presents an interesting opportunity for increasing the yield of biochemicals derived directly from glucose: in addition to glucose, lignocellulosic biomass contains several other fermentable sugars (e.g., xylose and arabinose) which may be used for biomass formation while reserving glucose solely for product generation. Because wild-type E. coli preferentially consumes glucose, strain engineering is necessary to shift the cell’s preference towards alternative carbon sources. In this work, we characterized the carbon source preference of a Δpgi Δzwf mutant and explored its ability to improve the yield of D-glucaric acid on D-glucose.

As expected, deletion of pgi and zwf eliminates the cell’s ability to consume D-glucose for biomass formation. Catabolite repression is eliminated in this strain as well, as the Δpgi Δzwf mutant is able to consume L-arabinose, glycerol, and D-xylose in the presence of D-glucose. Because glucose-mediated catabolite repression occurs when glucose transport into the cell is high, we believe that intracellular buildup of glucose-6-phosphate in the Δpgi Δzwf mutant leads to significantly reduced glucose transport, alleviating catabolite repression. Interestingly, introduction of the D-glucaric acid pathway, which should draw down intracellular glucose-6-phosphate pools, does not appear to affect the uptake of alternative carbon sources in the presence of D-glucose. We speculate that glucose influx in the presence of INO1 is not high enough to significantly reduce the levels of phosphorylated EIIA^Glc to result in activation of catabolite repression. Because the threshold rate of glucose import necessary for activation of catabolite repression is unknown, efforts to further increase the activity of INO1 or to introduce more active glucose consumption pathways should proceed with caution to avoid reactivation of catabolite repression.

5. Conclusions

In this work, we investigate the behavior of a Δpgi Δzwf mutant and its ability to utilize alternative carbon sources for cell growth while reserving D-glucose for product formation. This strain was able to consume L-arabinose, glycerol, and D-xylose even in the presence of D-glucose, and yields of D-glucaric acid on D-glucose were increased 9- to 18-fold in the Δpgi Δzwf strain. Additionally, product titers were also increased, as the initial D-glucaric acid pathway enzyme was no longer in competition with glycolytic enzymes for glucose-6-phosphate. Furthermore, the Δpgi Δzwf mutant exhibits similar yield increases in minimal medium, suggesting the strain’s potential in an industrial setting; however, additional investigation is necessary to fully characterize the strain’s robustness. These gains in product yield should easily translate to other bioproducts derived from D-glucose, and it is hoped that this strain will help improve the process economics of these value-added biochemicals.