OUP user menu

Myo-inositol facilitators IolT1 and IolT2 enhance d-mannitol formation from d-fructose in Corynebacterium glutamicum

Carsten Bäumchen, Eva Krings, Stephanie Bringer, Lothar Eggeling, Hermann Sahm
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01425.x 227-235 First published online: 1 January 2009


Reduction of d-fructose to d-mannitol by whole-cell biotransformation with recombinant resting cells of Corynebacterium glutamicum ATCC13032 requires the coexpression of mdh and fdh, which encode mannitol and formate dehydrogenases, respectively. However, d-mannitol formation is limited by the uptake of d-fructose in its unphosphorylated form, because additional expression of the sugar facilitator from Zymomonas mobilis resulted in a significantly increased productivity. Here we identified similarities of the myo-inositol transporters IolT1 and IolT2 of C. glutamicum to the sugar facilitator of Z. mobilis. The myo-inositol transporter genes were both individually overexpressed and deleted in recombinants expressing mdh and fdh. Biotransformation experiments showed that the presence and absence, respectively, of IolT1 and IolT2 significantly influenced d-mannitol formation, indicating a d-fructose transport capability of these transporters. For further evidence, a C. glutamicumΔptsF mutant unable to grow with d-fructose was complemented with a heterologous fructokinase gene. This resulted in restoration of growth with d-fructose. Using overexpressed iolT1, mdh and fdh, d-mannitol formation obtained with C. glutamicum was 34.2 g L−1, as opposed to 16 g L−1 formed by the strain overexpressing only mdh and fdh, showing the suitability of myo-inositol transporters for d-fructose uptake to obtain d-mannitol formation by whole-cell biotransformation with C. glutamicum.

  • whole-cell biotransformation
  • sugar transport
  • Corynebacterium glutamicum


Whole-cell biotransformation with recombinant, resting bacterial cells represents a new approach for synthesizing products of industrial interest. A number of NAD(P)H-dependent whole-cell biotransformation systems have been described expressing recombinant mono-oxygenases or oxidoreductases in Escherichia coli (Galkin et al., 1997; Bühler & Schmid, 2004; Kaup et al., 2004; Ernst et al., 2005). In addition to the expression levels of the enzymes involved and the mode of cofactor regeneration, substrate import and product export is also obviously of importance for high cellular productivity. Limitations in substrate import would lead to a situation where the substrate is present in the cell at nonsaturating concentrations, and limitations in export might block the entire system (Walton & Stewart, 2004).

We previously described recombinant strains of E. coli, Corynebacterium glutamicum and Bacillus megaterium, which are useful as biocatalysts in whole-cell biotransformation for the formation of d-mannitol from d-fructose. For biotransformation, cells expressing the genes for formate dehydrogenase (FDH) and mannitol dehydrogenase (MDH) were supplied with formate and d-fructose in phosphate buffer, i.e. biotransformation proceeded under nongrowing conditions (Fig. 1) (Kaup et al., 2004, 2005; Ernst et al., 2005; Bäumchen & Bringer-Meyer, 2007; Bäumchen et al., 2007). The strains overproduced NAD+-dependent mannitol 2-dehydrogenase from Leuconostoc pseudomesenteroides ATCC 12291 (Hahn et al., 2003) for the reduction of d-fructose to d-mannitol, NAD+-dependent FDH from Mycobacterium vaccae N10 (Galkin et al., 1995) for NADH regeneration and the glucose facilitator (GLF, glfZm) from Zymomonas mobilis (Parker et al., 1995; Weisser et al., 1995) for the uptake of d-fructose without any concomitant phosphorylation. This sugar porter belongs to the major facilitator superfamily and transports d-glucose and d-fructose along a concentration gradient across the cell membrane (Pao et al., 1998). For the whole-cell biotransformation of d-fructose to d-mannitol, the glfZm gene was overexpressed in E. coli and C. glutamicum, in both cases increasing the productivities of the strains, which verified its function in these bacteria (Parker et al., 1995; Weisser et al., 1995; Kaup et al., 2004; Bäumchen & Bringer-Meyer, 2007). Overexpression of glfZm in C. glutamicum increased the d-fructose uptake rate 5.5-fold, and d-mannitol concentrations of up to 87 g L−1 were formed (Bäumchen & Bringer-Meyer, 2007).

Figure 1

d-Mannitol production from d-fructose using a recombinant redox cycle in Corynebacterium glutamicum ATCC13032. Formate dehydrogenase from Mycobacterium vaccae N10; mannitol dehydrogenase from Leuconostoc pseudomesenteroides; IolT1 or IolT2, myo-inositol transporter from C. glutamicum.

During these studies, we found that the wild type of C. glutamicum ATCC13032 as well as a phosphotransferase system (PTS)-fructose-negative mutant (ΔptsF) that overexpressed mdh and fdh produced d-mannitol in a significant concentration of up to 16 g L−1 (Bäumchen & Bringer-Meyer, 2007). This unexpected result indicates that C. glutamicum had unphosphorylated d-fructose available in the cell for d-mannitol formation. In these strains, d-fructose was apparently transported via a PTS-independent uptake mechanism. Dominguez & Lindley (1996) postulated the presence of a d-fructose exporter because d-fructose is formed intracellularly during growth on sucrose, which is taken up by the PTS system as sucrose-6-phosphate and then split to glucose-6-phosphate and d-fructose, which, due to the absence of a fructokinase, cannot be further metabolized (Kalinowski et al., 2003). d-Fructose is excreted by an unknown export system and d-fructose is taken up again by the d-fructose-specific PTS. However, no genes encoding any transporters or permeases for d-fructose had been identified in C. glutamicum. Here we analysed the genome of C. glutamicum and identified amino acid sequences with high identities to GLF of Z. mobilis that functioned as d-fructose transporters, and we demonstrate their usefulness for the whole-cell biotransformation of d-fructose into d-mannitol.

Materials and methods

Strains and culture conditions

The bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli DH5α was grown in 15-mL test tubes containing 5 mL Luria–Bertani medium (37 °C) as described previously (Sambrook & Russell, 2000). Corynebacterium glutamicum ATCC13032 was grown in 500-mL baffled flasks containing 50 mL of brain heart infusion (BHI, Difco), CgIII medium (Menkel et al., 1989) or CgXII minimal medium (Keilhauer et al., 1993) containing 4%d-glucose or 1%, 4% or 9%d-fructose at 30 °C on an orbital shaker (Infors, Bottmingen, Switzerland) at an agitation of 120 r.p.m. Cultivation of C. glutamicum in CgXII minimal medium was preceded by overnight growth of a preculture in CgIII medium with 4% glucose, harvesting and washing (0.9% NaCl) of the cells, growth in CgXII medium, again harvesting and washing of cells and inoculation of the main cultures at an OD600 nm of 1.0. Plasmids were selected by adding 25 μg mL−1 kanamycin (pEKEx2-fdh-mdh), 5 μg mL−1 tetracyclin (pVWEx2-glf) or 10 μg mL−1 chloramphenicol (pXFK1) to the medium. In order to construct recombinant C. glutamicum ATCC13032, harbouring plasmids pEKEx2-fdh-mdh together with pVWEx2-iolT1 or pVWEx2-iolT2, cells were first transformed with pEKEx2-fdh-mdh. Strains were cultivated by inoculating a single colony into 20-mL BHI medium with the appropriate addition of antibiotics as given above. These precultures were grown overnight at 30 °C and 120 r.p.m. and were used to inoculate the main cultures to an OD600 nm of 0.5. The main cultures were grown in BHI medium in the presence of the appropriate antibiotics at 30 °C and 120 r.p.m. to an OD600 nm of 1.5, induced with isopropyl-β-d-thiogalactoside at a final concentration of 0.7 mM, incubated at 30 °C and 120 r.p.m. for 5 h and harvested at an OD600 nm of 8.

View this table:
Table 1

Strains and vectors used in this study

Strain or plasmidRelevant characteristicsReferences/source
E. coli DH5αendA1 supE44 recA1 gyrA96 relA1 deoR U169Φ80dlacZΔM15 mcrAΔ(mrr-hsdRMS-mcrBC)Grant et al. (1990)
C. glutamicum strains
ATCC 13032WT
WTΔiolT1In-frame deletion of Cgl0181 (iolT1)Krings et al. (2006)
WTΔiolT2In-frame deletion of Cgl3058 (iolT2)Krings et al. (2006)
WTΔiolT1ΔiolT2In-frame deletions of Cgl0181 and Cgl3058 (iolT1 and iolT2)Krings et al. (2006)
WTpfdh-mdh piolT1Carrying pEXEx2-fdh-mdh; pVWEx2-Cgl0181This study
WTpfdh-mdh piolT2Carrying pEXEx2-fdh-mdh; pVWEx2-Cgl3058This study
WTΔiolT1-pfdh-mdhCarrying pEKEx2-fdh-mdhThis study
WTΔiolT2-pfdh-mdhCarrying pEKEx2-fdh-mdhThis study
WTΔiolT1ΔiolT2-pfdh-mdhCarrying pEKEx2-fdh-mdhThis study
WTΔptsFptsF mutant, KanRMoon et al. (2005)
WTΔptsF-pXFK1Carrying pXFK1Moon et al. (2005)
Plasmids (simplified name)
pVWEx2TetR; C. glutamicum/E. coli shuttle vector for regulated gene expression (Ptac, lacIq, pHM1519, oriVC.g., oriVE.c.)Wendisch (1997)
pVWEx2-Cgl0181 (piolT1)TetR; C. glutamicum/E. coli shuttle vector for regulated gene expression of Cgl0181 (Ptac, lacIq, pHM1519, oriVC.g., oriVE.c.)This study
pVWEx2-Cgl3058 (piolT2)TetR; C. glutamicum/E. coli shuttle vector for regulated gene expression of Cgl3058 (Ptac, lacIq, pHM1519, oriVC.g., oriVE.c.)This study
pEKEx2-fdh-mdh (pfdh-mdh)KanR; C. glutamicum/E. coli shuttle vector for regulated gene expression of fdh and mdh (Ptac, lacIq, pBL1, oriVC.g., oriVE.c.)Bäumchen & Bringer-Meyer (2007)
pVWEx2-glfTetR; C. glutamicum/E. coli shuttle vector for regulated gene expression of glf (Ptac, lacIq, pHM1519, oriVC.g., oriVE.c.)Bäumchen & Bringer-Meyer (2007)
pXFK1CmR; pXMJ19 with a 1.2 kb scrK gene encoding fructokinase from Clostridium acetobutylicumMoon et al. (2005), Jakoby et al. (1999)

DNA preparation and transformation

Standard protocols were used for the construction, purification and analysis of plasmid DNA and for the transformation of E. coli (Sambrook & Russell, 2000). The construction of plasmids was carried out in E. coli DH5α. Transformation of C. glutamicum ATCC13032 by electroporation was performed by the method of Tauch et al. (2002) using a Gene Pulser and Pulse Controller (BioRad, Munich, Germany). Gene Pulser settings were: voltage, 2.5 kV; resistor, 200 Ω; and capacitance, 25 μF. PCR was performed with a DNA thermal cycler (T3 Thermocycler, Biometra, Göttingen, Germany) using Phusion DNA polymerase (Finnzyme, distributed by New England Biolabs, Frankfurt, Germany).

Vector construction for expression and deletion of iolT1 and iolT2 in C. glutamicum

To determine the effect of IolT1 and IolT2 on the d-fructose uptake in C. glutamicum ATCC13032 genes, Cgl0181 and Cgl3058 each were ligated into expression vector pVWEx2. For the gene Cgl0181, PCR forward primer (ECgl0181_Pst1_f) was 5′-TATACTGCAGATGGCTAGTACCTTCATTCAGGCC-3′ and the reverse primer (E_Cgl0181_Sal1_r) was 5′-ATATGTCGACTTAGTGCACCTTTCCTTTTCGGATGTC-3′. For the gene Cgl3058, PCR forward primer (ECgl3058_XbaI_f) was 5′-TATATCTAGAATGACGGACATCAAGGCCACATCAAGT-3′ and the reverse primer (E_Cgl3058_BamHI_r) was 5′-ATATGGATCCTTAAGCCTTCTTGAAGATCTGGCCGG-3′. Oligonucleotides were obtained from Operon (Cologne, Germany). Standard protocols were applied for the generation of fragments via PCR, for restriction and for ligation (Sambrook & Russell, 2000). The chromosomal mutations leading to strains C. glutamicum 13032 Δcgl0181 (=ΔiolT1) and C. glutamicum 13032 Δcgl3058 (=ΔiolT2) were performed as described by Krings et al. (2006).

Preparation of cell-free extracts and enzyme assays

Cells were harvested by centrifugation (10 000 g, 10 min, and 4 °C) and washed in 100 mM potassium phosphate buffer, pH 6.5. Cell disruption was carried out by sonication (6 min, 4 °C) and crude extracts were centrifuged at 10 000 g for 10 min at 4 °C. The supernatants were used as cell-free extracts. The activities of FDH and MDH were determined spectrophotometrically. The activity of FDH was determined in an assay mixture consisting of 200 mM sodium formate, 2 mM NAD and 100 mM potassium phosphate buffer, pH 7.0 (Schütte et al., 1976). The assay for MDH contained 200 mM d-fructose and 200 μM NADH in 100 mM potassium buffer, pH 6.0 (Hahn et al., 2003). One unit of enzyme activity (U) was defined as the amount of enzyme catalysing the conversion of 1 μmol pyridine dinucleotide per minute at 30 °C. Protein concentrations were determined by the method of Bradford (1976) using bovine serum albumin as a standard.


Cells from 400 mL BHI complex medium were harvested at OD600 nm 8 by centrifugation (10 000 g, 10 min and 4 °C) and washed twice in 100 mM potassium phosphate buffer, pH 6.5. Twenty grams of cells per liter wet weight (corresponding to 2.9 g cells L−1 dry weight) were resuspended in a reaction solution containing 500 mM d-fructose and 250 mM sodium formate in 100 mM potassium phosphate buffer, pH 6.5. The higher concentration of d-fructose was found to be necessary to maintain an outside-in sugar concentration gradient across the cell membranes to allow sugar transport by facilitated diffusion (Kaup et al., 2004). The reaction temperature was 30 °C. For pH-static conditions, a solution of 3 M formic acid/3 M d-fructose was titrated into the reaction vessel. The titration rate was controlled by automatic titration at pH 6.5. Incubation times were as indicated in the text. Samples of the biotransformation system were centrifuged (10 000 g, 10 min, 4 °C), filtered and the resulting supernatants were analysed by HPLC.

Determination of d-fructose and d-mannitol concentrations

For the quantification of d-fructose and d-mannitol, the samples obtained from biotransformation experiments were subjected to HPLC analysis. Chromatography was performed with an HPX-87C 300 × 7.8 mm column (Biorad) at 70 °C using Mili-Q water as an eluent. Formate was analysed with a Biorad HPX-87H 300 × 7.8 column at 70 °C with 6 mM H2SO4 as the mobile phase. Substances were detected by two detectors connected in series, i.e. photometrically at 190 nm and by a refractive index detector.


Identification of GLF homologues in C. glutamicum

Based on a genome sequence analysis of C. glutamicum ATCC13032 (GenBank accession no. NC_00345) via blast2 search (blastp 2.2.16) (Schaffer et al., 2001), two gene loci were identified encoding amino acid sequences with high identities to the amino acid sequence of GLF from Z. mobilis (Fig. 2). Locus Cgl0181 (NCgl0178) encodes for a gene of 1476 bp (NCBI Gene-ID 1021246). This gene produced alignments with a score (bits) of 147 and an e-value of 8e−37. The second gene identified was Cgl3058 (NCgl2953) with a size of 1527 bp (NCBI Gene-ID 1021001). It produced alignments with a score (bits) of 141 and an e-value of 8e−35. The sizes of the genes were in the same range as that of glfZm (1422 bp). Besides individual residues, such as Gly introducing turns, or charged residues such as Glu, or Arg, even stretches were largely retained, such as aminoacyl residues in position 39–46 or 98–106 (referring to Cgl0181). Interestingly, a recent report by Krings et al. (2006) showed that both genes code for the membrane-associated myo-inositol transporters: IolT1 (Cgl0181) and IolT2 (Cgl3058). The presence of at least one transporter is essential for growth on myo-inositol as the sole carbon source (Krings et al., 2006). Myo-inositol is the most abundant stereoisomer of inositol (1, 2, 3, 4, 5, 6-cyclohexanehexol) and its phosphorylated form, myo-inositol hexakisphosphate or phytic acid, is the most abundant component of organic phosphate in soil (Turner & Richardson, 2004). Many microorganisms, mostly inhabitants of soil, can utilize myo-inositol as a carbon source (Yebra et al., 2007).

Figure 2

clustalx alignment of the amino acid sequence of IolT1, IolT2 and Glf. The line above the ruler is used to mark strongly conserved positions. Three characters (‘*’, ‘:’ and ‘.’) are used: ‘*’ indicates positions that have a single, fully conserved residue; ‘:’ indicates that one of the following ‘strong’ groups is fully conserved: STA, NEQK, NHQK, NDEQ, QHRK, MILV, MILF, HY and FYW; ‘.’ indicates that one of the following ‘weaker’ groups is fully conserved: CSA, ATV, SAG, STNK, STPA, SGND, SNDEQK, NDEQHK, NEQHRK, FVLIM and HFY.

Amplification of iolT1 and iolT2 in the C. glutamicum whole-cell biocatalyst

Strain C. glutamicum pfdh-mdh expressing the genes for FDH and MDH exhibited a constant d-mannitol productivity, yielding 16 g L−1d-mannitol after 24 h of biotransformation (Bäumchen & Bringer-Meyer, 2007). During biotransformation, the cells retained their integrity as was derived from constant intracellular concentrations of the cofactors NAD+ and NADH (Bäumchen & Bringer-Meyer, 2007). As shown in Fig. 3, coexpression of the genes encoding the myo-inositol transporters IolT1 or IolT2 in the production strain led to a significant increase in d-mannitol accumulation. Strain C. glutamicum pfdh-mdh piolT1 yielded as much as 34.2 g L−1d-mannitol, while strain C. glutamicum pfdh-mdh piolT2 yielded 28.8 g L−1 after 24 h of biotransformation. This 2.1-fold and 1.8-fold increase in d-mannitol productivity, in comparison with the reference strain, is presumably due to a higher d-fructose uptake rate resulting in a higher conversion rate of d-fructose to d-mannitol. The determination of in vitro enzyme activities for FDH showed nearly equal values of 0.11 U mg−1 protein for the FDH for C. glutamicum pfdh-mdh piolT1 and 0.12 U mg−1 protein for C. glutamicum pfdh-mdh piolT2. For the MDH, an activity of 1.9 U mg−1 protein for C. glutamicum pfdh-mdh piolT1 and 2.5 U mg−1 protein for C. glutamicum pfdh-mdh piolT2 was determined. These values largely agree with the reference strain that does not overexpress one of the myo-inositol transporter genes (Bäumchen & Bringer-Meyer, 2007), thus making it unlikely that the artificial redox cycle was altered in the strains exhibiting increased performance.

Figure 3

d-Mannitol formation during biotransformation of recombinant strains of Corynebacterium glutamicum13032. Biotransformations were performed over a period of 24 h. The data represent the result of two independent biological experiments for each strain.

Generally, C. glutamicum takes up d-fructose by the phosphoenolpyruvate PTS (Moon et al., 2005). Because of the lack of a fructokinase gene in C. glutamicum (Kalinowski et al., 2003), it was assumed that d-fructose import only relies on the PTS or the d-fructose-specific EII gene. Therefore, a ΔptsF mutant is not able to grow with d-fructose because the PTS system not only imports but also phosphorylates d-fructose (Moon et al., 2005). If a C. glutamicumΔptsF mutant expressing a heterologous fructokinase gene was able to grow with d-fructose, evidence of the function of IolT1/IolT2 as d-fructose transporters would be provided. Therefore, vector pXFK1 carrying the fructokinase gene (scrK) of Clostridium acetobutylicum (Moon et al., 2005) was introduced into C. glutamicumΔptsF (Moon et al., 2005), and growth rates were determined in liquid culture in minimal medium CgXII. As shown in Table 2, the presence of fructokinase enabled growth of the ΔptsF mutant by as much as about 50% of that of the wild type in a fructose concentration-dependent manner as would be expected for a facilitator. By complementation with a fructokinase gene, the d-fructose utilization capability of the C. glutamicumΔptsF mutant was restored.

View this table:
Table 2

Restoration of the d-fructose utilization capability of a Corynebacterium glutamicumΔptsF mutant by complementation with a fructokinase gene

d-Fructose (%)C. glutamicum wild typeC. glutamicum ΔptsF pXFK1Growth ratemutant
Growth rate μ (h−1)Growth rate μ (h−1)Growth rate wild type (%)
  • The cells were grown with minimal medium and d-fructose as the carbon source.

d-Mannitol formation by iolT1 and iolT2 deletion mutants of the C. glutamicum whole-cell biocatalyst

In order to further substantiate the view that IolT1 and IolT2 mediate transport of d-fructose, marker-free deletion mutants C. glutamicum ATCC13032 ΔiolT1 and C. glutamicum ATCC13032 ΔiolT2 were transformed with vector pfdh-mdh and tested for their capabilities to form d-mannitol from d-fructose during biotransformation. Both strains produced significantly less d-mannitol in comparison with the reference strain C. glutamicum pfdh-mdh (Fig. 4). Corynebacterium glutamicumΔiolT1-pfdh-mdh produced 7.2 g L−1d-mannitol, corresponding to a decrease in productivity of 55%. The second mutant, C. glutamicumΔiolT2-pfdh-mdh, yielded a final concentration of 5.8 g L−1d-mannitol corresponding to a decrease in productivity in the same range (64%). This result indicates that lack of iolT1 or iolT2 leads to a lower d-fructose uptake rate in comparison with the reference strain C. glutamicum pfdh-mdh and again agrees with the view that IolT1 and IolT2 have import activity with d-fructose as a substrate. In addition to the single-deletion mutants, the double-deletion mutant C. glutamicumΔiolT1ΔiolT2 was made, which was transformed with pfdh-mdh, and its d-mannitol accumulation was determined. This strain showed almost no d-mannitol formation after 24 h (0.3 g L−1), further corroborating the view that both myo-inositol transporters IolT1 and IolT2 catalyse uptake of d-fructose in C. glutamicum under the conditions assayed. Therefore, two transporters for the uptake of d-fructose exist in C. glutamicum, and these transporters are independent of a PTS and operate without concomitant phosphorylation of the substrate.

Figure 4

d-Mannitol formation during biotransformation of recombinant wild-type and mutant strains of Corynebacterium glutamicum13032. Biotransformations were performed over a period of 24 h. The data represent the result of two independent biological experiments for each strain.


The Gram-positive microorganism C. glutamicum is well known as an efficient producer of amino acids on an industrial scale and its metabolism has been widely investigated over the last few decades (Sahm et al., 2000). Because of the lack of a fructokinase gene (Kalinowski et al., 2003), it was assumed that d-fructose import relies only on the phosphoenolpyruvate PTS, or the d-fructose-specific EII gene. It was therefore surprising that the wild type of C. glutamicum utilizes d-fructose independently of PTS. The existence of a glucose importer and a d-fructose exporter is postulated for C. glutamicum (Yokota & Lindley, 2005). With respect to d-fructose, an export carrier was anticipated, due to the fact that during growth on sucrose, which is taken up by the PTS, phosphorylated and then split into glucose-6-phosphate and d-fructose, d-fructose was found to accumulate transiently in the medium (Dominguez & Lindley, 1996). The experimental results of the present work clearly show that the myo-inositol transporters IolT1 and IolT2 enable d-fructose import. Both are structurally related to the glucose/fructose permease of Z. mobilis characterized as a facilitator (Parker et al., 1995; Weisser et al., 1995). Given an appropriate concentration gradient (high internal fructose), it is therefore possible that the myo-inositol transporters may also catalyse the d-fructose export known to occur during growth on sucrose (Dominguez & Lindley, 1996).

Hexoses, such as d-fructose, are frequently transported by bacteria via the highly specific PTS under concomitant phosphorylation, enabling the sugars to enter the central metabolism of the cells. Besides the presence of the PTS, additional transporters are known that import sugars in their unphosphorylated forms. For instance, Bacillus subtilis possesses a glucose permease, GlcP, that contributes up to 30% of the total glucose import of the cell (Inaoka & Ochi, 2007), and in B. megaterium the d-fructose permease FruP enables the cell to take up d-fructose without concomitant phosphorylation (Chiou et al., 2002). Zymomonas mobilis has a glucose/fructose permease, GLF, belonging to the major facilitator superfamily (Parker et al., 1995; Weisser et al., 1995), which we have already used successfully to increase the d-fructose uptake rate in C. glutamicum (Bäumchen & Bringer-Meyer, 2007). While such transporters are rare among bacteria, yeasts have a number of sugar permeases available (Heiland et al., 2000).

Overexpression of the genes coding for IolT1 and IolT2 led to a roughly twofold increase of the d-fructose uptake rate. Most probably, d-fructose uptake was favoured by the high concentration gradient of d-fructose across the cell membranes, formed by 500 mM d-fructose present in the reaction buffer and the strong intracellular d-fructose conversion rate by the overproduced MDH. Accordingly, depending on the external concentration of d-fructose, we could show restoration of growth of a PtsF mutant of C. glutamicum recombinantly expressing a fructokinase gene. Under physiological conditions, the d-fructose uptake activities of IolT1 and IolT2 might be negligible, in particular because the PTS have high affinities for their substrates (Rohwer et al., 2000; Kornberg, 2001). Because C. glutamicum cannot metabolize d-mannitol (Eggeling & Bott, 2005), the final concentration of d-mannitol formed represents the number of turnovers of the introduced redox cycle, thus allowing an estimation of the d-fructose uptake rate. For the strain coexpressing iolT1, the estimated uptake rate for d-fructose is 22.7 nmol min−1 mg−1 protein, and for the strain coexpressing iolT2, it is 19.1 nmol min−1 mg−1 protein. In comparison with the d-fructose uptake rate due to GLF, with 56.3 nmol min−1 mg−1 protein, the myo-inositol transporters possibly have a lower specificity towards d-fructose. However, overexpression of iolT1 or iolT2 led to a significant increase of d-mannitol productivity in whole-cell biotransformation, showing the usefulness of these transporters and the importance of the transport process for biotransformation. For further studies, the whole-cell biotransformation system could be used for the screening of d-fructose permeases because the overexpression of putative d-fructose transporter genes from other bacteria would lead to an increase in d-mannitol formation, indicating an elevated d-fructose import, independent of PTS. Suitable candidates for further experiments are myo-inositol transporters from other bacteria such as Bacillus sp. (Yoshida et al., 2002).


We thank Karin Krumbach for skillful technical assistance.


  • Editor: Atsushi Yokota


View Abstract