The DeoR-Type Regulator SugR Represses Expression of ptsG in Corynebacterium glutamicum

Abstract

Corynebacterium glutamicum grows on a variety of carbohydrates and organic acids. Uptake of the preferred carbon source glucose via the phosphoenolpyruvate-dependent phosphotransferase system (PTS) is reduced during coutilization of glucose with acetate, sucrose, or fructose compared to growth on glucose as the sole carbon source. Here we show that the DeoR-type regulator SugR (NCgl1856) represses expression of ptsG, which encodes the glucose-specific PTS enzyme II. Overexpression of sugR resulted in reduced ptsG mRNA levels, decreased glucose utilization, and perturbed growth on media containing glucose. In mutants lacking sugR, expression of the ptsG′-′cat fusion was increased two- to sevenfold during growth on gluconeogenic carbon sources but remained similar during growth on glucose or other sugars. As shown by DNA microarray analysis, SugR also regulates expression of other genes, including ptsS and the putative NCgl1859-fruK-ptsF operon. Purified SugR bound to DNA regions upstream of ptsG, ptsS, and NCgl1859, and a 75-bp ptsG promoter fragment was sufficient for SugR binding. Fructose-6-phosphate interfered with binding of SugR to the ptsG promoter DNA. Thus, while during growth on gluconeogenic carbon sources SugR represses ptsG, ptsG expression is derepressed during growth on glucose or under other conditions characterized by high fructose-6-phosphate concentrations, representing one mechanism which allows C. glutamicum to adapt glucose uptake to carbon source availability.

Corynebacterium glutamicum is a high-G+C content gram-positive soil bacterium which was isolated in 1957 as an l-glutamate-excreting bacterium (1, 32). It plays an important role in the large-scale biotechnological production of l-glutamate and l-lysine, with annual production of about 1.5 million tons/year and 0.85 million tons/year, respectively (21, 40). This species has become a model organism for the Corynebacterianeae, a suborder of the Actinomycetales, which also comprises Nocardiaceae and Mycobacteriaceae, and extensive knowledge about this nonpathogenic bacterium has accumulated since its discovery (summarized in reference 11).

C. glutamicum is able to grow on various carbon and energy sources, such as sugars, sugar alcohols, and organic acids (10, 12, 17, 27, 34, 42, 70, 73). In contrast to Escherichia coli and Bacillus subtilis, which show distinct catabolite repression, C. glutamicum usually coutilizes the carbon sources present in the mixtures without showing diauxic growth. Glucose is the preferred carbon source, but it has been shown to be cometabolized with, e.g., acetate (71), l-lactate (65), propionate (6), pyruvate and serine (46), protocatechuate (41), vanillate (41), and fructose (8). The only known exception is the sequential consumption of glucose before glutamate (35).

The central carbon metabolism of C. glutamicum has been characterized by enzymatic studies, carbon flux analysis, genetic analysis, and, since the complete genome sequence of ATCC 13032 is available (25), also by genome-wide studies (reviewed in references 12, 13, 59, and 73). For example, based on intensive studies on acetate metabolism and its regulation (see reference 17 and references therein), three transcriptional regulators of acetate metabolism, namely, RamA (7), RamB (16), and GlxR (29), as well as AcnR, the transcriptional repressor of the aconitase gene acn (36), have been identified. However, only a few studies investigated the sugar transport systems and their regulation on the molecular level (42, 48). Mori and Shiio (43) first mentioned the presence of a phosphoenolpyruvate-dependent sugar phosphotransferase system (PTS) in C. glutamicum. The PTS operates by phosphoryl group transfer from phosphoenolpyruvate via the general PTS components EI (encoded by ptsI) and HPr (encoded by ptsH) to the sugar-specific enzyme II permeases, which phosphorylate and transport their cognate substrates into the cell (33, 39, 48). C. glutamicum possesses the specific EII permeases EII^Glc, EII^Fru, and EII^Suc (encoded by ptsG, ptsF, and ptsS, respectively) for uptake of glucose, fructose, and sucrose, respectively (9, 27, 28, 42).

The glucose uptake in C. glutamicum appears to be regulated, as, e.g., during growth on glucose-acetate mixtures glucose uptake is reduced about twofold in comparison to growth on glucose alone (71). Since C. glutamicum lacks homologs of regulatory proteins involved in control of glucose uptake in E. coli (e.g., cyclic AMP [cAMP] receptor protein [CRP] and Mlc) and B. subtilis (e.g., CcpA, LicT, and GlcT) and since PTS components are involved in regulation of sugar uptake in many bacteria, we screened the genomic regions of the genes encoding PTS components for the presence of genes coding for potential transcriptional regulators of glucose uptake. Regarding the localization of the PTS genes in the genome of C. glutamicum, it was striking that the genes coding for the general PTS proteins (ptsI and ptsH) and for the fructose-specific EII (ptsF) lie next to two genes coding for transcriptional regulators of the DeoR family: NCgl1856 (renamed SugR in this study) and NCgl1859. Based on the assumption that one or both of these regulators is involved in the regulation of glucose uptake, we examined the role of the transcriptional regulators SugR and NCgl1859 in control of ptsG expression.

(Part of the work described here belongs to the planned dissertation of Verena Engels at the faculty of Mathematics and Natural Sciences of the Heinrich-Heine-Universität Düsseldorf.)

MATERIALS AND METHODS

Bacterial strains, plasmids, oligonucleotides, and culture conditions.

The strains and plasmids used are listed in Table 1. For growth experiments, in the first preculture, 5 ml Luria-Bertani (LB) medium (58) was inoculated from a fresh LB agar plate and incubated at 30°C and 170 rpm. After washing the cells in medium without a carbon source, the second preculture and the main culture were inoculated to an optical density at 600 nm (OD₆₀₀) of 0.5 to 1.0 in 60 ml LB medium or in 60 ml CgXII minimal medium (26), which contained 0.03 g/liter protecatechuic acid. As carbon and energy sources, 100 mM glucose, 100 mM fructose, 50 mM glucose plus 50 mM fructose, 50 mM sucrose, 300 mM K-acetate, 100 mM Na₃-citrate plus 100 mM MgCl₂, 200 mM Na-pyruvate, 50 mM maltose, 120 mM ribose and 50 mM glucose plus 150 mM K-acetate were used. Precultures and main cultures were incubated at 30°C and 120 rpm on a rotary shaker in 500-ml baffled shake flasks. When appropriate, the medium was supplemented with kanamycin (50 μg/ml) or ampicillin (50 μg/ml). Growth of C. glutamicum was followed by measuring the OD₆₀₀. The biomass concentration was calculated from the OD₆₀₀ values using an experimentally determined correlation factor of 0.25 g (dry weight) (0.25 gDW) for an OD₆₀₀ of 1 (71). For all cloning purposes, Escherichia coli DH5α was used as the host, and for overproduction of SugR, E. coli BL21(DE3) (66) was used.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Source or reference
E. coli strains
DH5α	F−thi-1 endA1 hsdR17(r− m−) supE44 ΔlacU169 (φ80lacZΔM15) recA1 gyrA96 relA1	Bethesda Research Laboratories (20)
BL21(DE3)	ompT hsdSB(rB− mB−) gal dcm(DE3)	66
C. glutamicum strains
ATCC 13032	Wild-type strain auxotrophic for biotin	1
ΔsugR mutant	In-frame deletion of the sugR gene	This study
Δ1859 mutant	In-frame deletion of the NCgl1859 gene	This study
ΔsugR Δ1859 mutant	In-frame deletion of the sugR and NCgl1859 genes	This study
Plasmids
pGEM-T	Ampr; PCR cloning vector (PT7, PSP6)	Promega, WI, USA
pGEM-T-ΔsugR	Ampr; pGEM-T with deletion fragment of 906 bp of sugR	This study
pGEM-T-Δ1859	Ampr, pGEM-T with deletion fragment of 987 bp of NCgl1859	This study
pGEM-T-SugRHis	Ampr; pGEM-T with a 786-bp fragment of the sugR gene	This study
pk19mobsacB	Kanr; E. coli/C. glutamicum shuttle vector for construction of insertion and deletion mutants in C. glutamicum (pK18 oriVEc, sacB, lacZα)	60
pk19mobsacB-ΔsugR	Kanr; pk19mobsacB with the deletion construct of the sugR gene	This study
pk19mobsacB-Δ1859	Kanr; pk19mobsacB with the deletion construct of the NCgl1859 gene	This study
pVWEx1	Kanr; PtaclacIq	51
pVWEx1-sugR	Kanr; pVWEx1 with an 807-bp fragment of the sugR gene and an artificial ribosome binding site	This study
pET16b	Ampr; overproduction of proteins with a N-terminal decahistidine tag in E. coli (pBR322 oriVEc, PT7, lacI)	Novagen
pET16b-sugRHis	Ampr; pET16b with a 786-bp fragment of the sugR gene	This study
pET2	Kanr; promoter-probe vector	69
pET2-ptsG	Kanr; pET2 with the full-length promoter fragment of 707 bp of the ptsG promoter	This study
pET2-ptsG/F4	Kanr; pET2 with promoter fragment 4 of 188 bp of the ptsG promoter	This study

Recombinant DNA work.

The oligonucleotides listed in Table 2 were obtained from Operon (Cologne, Germany) or MWG (Ebersberg, Germany). Standard methods such as PCR, restriction, and ligation were carried out as described previously (58). Plasmids from E. coli were isolated with the QIAprep spin miniprep kit (QIAGEN, Hilden, Germany). E. coli was transformed by the RbCl method (19), and C. glutamicum was transformed by electroporation (68) at 25 μF, 600 Ω, and 2.5 kV/cm (Bio-Rad Gene Pulser Xcell; Bio-Rad Laboratories, Hercules, Canada). All cloned DNA fragments were shown to be correct by sequencing (Agowa GmbH, Berlin, Germany).

TABLE 2.

Oligonucleotides used in this study

Oligonucleotide	Sequence (5′→3′)	Sequence details
		Underlined nucleotide (position in NC003450)	Sequence in bold (restriction site)	Sequence in italic
sugR-A	GCGAATTCACAAGGATTCATCTGGCATC	2037391	EcoRI
sugR-B	CCCATCCACTAAACTTAAACAGCGCTCCTCTGCGTACAT	2037832		Linker
sugR-C	TGTTTAAGTTTAGTGGATGGGCGAGAACGCGATGTAGAAGTTGTG	2038559		Linker
sugR-D	GCGGATCCCAAATTGCCACCCAACAACACCC	2039001	BamHI
sugR-k-for	GTTCGTCGCGGCAATGATTGACG	2037351
sugR-k-rev	CTCACCACATCCACAAACCACGC	2039042
1859-A	GCTCTAGAAACGCCGGTGCCCTTCAGTACA	2041268	XbaI
1859-B	CCCATCCACTAAACTTAAACATTCCGTTTGGCTGACCATGCTTAAC	2041745		Linker
1859-C	TGTTTAAGTTTAGTGGATGGGTTGCCCGACCACAACTTTCAGG	2042487		Linker
1859-D	GCGAATTCGACGCGCGTACCAGTCAACTGG	2042966	EcoRI
1859-k-for	GGCGACGACTTCGCCTGCTTG	2041190
1859-k-rev	TATCGACAGCCACGCGAACGTC	2043017
sugR-for	GCGGATCCAAGGAGATATAGATATGTACGCAGAGGAGCG	2037815	BamHI	Ribosome binding site
sugR-rev	GCTCTAGAGTAAGAATCATTCTGCAATCACAAC	2038601	XbaI
SugRshift-for	GGAACCCATATGTACGCAGAGGAGCG	2037815	NdeI
SugRshift-rev	GGGGGATCCAGAATCATTCTGCAATCACAAC	2038598	BamHI
ptsG-for	TAGTTCCTGCACCTCTTAAGAGG	1422560
ptsG-rev	CCAAGGTTTTCCAGAATATGTTGCG	1423266
ptsG-1	GCGCCTGCAGTAGTTCCTGCACCTCTTAAGAGG	1422560	PstI
ptsG-2	GCGCGTCGACCCAAGGTTTTCCAGAATATGTTGC	1423266	SalI
F2-rev	CCGAAGCGCCCGCCAAAAGC	1422734
F3-for	GTGAGGCTTGGTTTTCGTACG	1422684
F3-rev	GAATATGTCCGACCTGCATTTTTTG	1422890
F4-for	GCATAATCTGACAGTGTGTCCGTTTTC	1422834
F4-rev	GGCTCCCCCGCAATAGATTTGTGG	1423019
F5-for	CCGCTTAATTGGAAAAAAGTGTGATCC	1422968
F5-rev	CGCCGTTACGAATTGGGCC	1423148
F6-for	GGTAGGCGCTGAGGAATC	1423095
NCgl2027-for	AGCTTCAGAGGAGCCTGAAGCGG	2224999
NCgl2027-rev	CGGAGTGGGTTGATTCTGGGATG	2225188
NCgl1955-for	GTTCGCTACGTCCGAGTGATCACC	2146341
NCgl1955-rev	GTCCCCCACGATTGTTTAAAAGTC	2145944
mepA-for	GCACATCCGCTACCGAACCATC	2649369
mepA-rev	CTAGTCGACGCTTCATGTTGCGG	2649635
ptsG_30*	[IRD800]GTTGCGATGTCGTCGTCAGTTTG	1423247		Labeled with IRD800
ptsG_60*	[IRD800]GTCTGGTCCACCAAGGTTTTC	1423276		Labeled with IRD800
ptsG_90*	[IRD800]GAGTCGCACAGTGAGTCATCG	1423307		Labeled with IRD800
ptsG-Primex-rev	GCAGGGCCCACAGGATTGG	1423607
ptsS_for	CTGCAGGTGATGGACAGGTTTCATTCG	2812022
ptsS_rev	CGCCGCCAATGTCGCGCAGG	2811759
ptsF_for	CCCAGAAGAAAACACTGGAGG	2043357
ptsF_rev	CGCCGAAATCGACATCCAG	2043559
NCgl1858/1859_for	CCGGTGCCCTTCAGTACAG	2041272
NCgl1858/1859_rev	GGTGCCAGTAAAGAAGCAATTGCATG	2041777
A6	GACCTGCATTTTTTGAAAATGAAAACGG	1422880
B1	CCTTTTTGGTTTGTCTGTATTCAGC	1422901
B6	GTCGGACATATTCAAAAGTATTACCTTTTTGG	1422878
SF_1	GGTTAAGCTTTTGGCGGGCGC	1422709
SF_2	GGCGAAAATACATGGGCCTACAACG	1422733
SF_3	CGTGAAACTGGGGGATGGGG	1422767
SF_4	CGGGGACGTGATGCGGTGG	1422802
SF_5	GGTGCGGGCATAATCTGACAGTG	1422827
SR_1	GTGGTGGATCACACTTTTTTCC	1422999
SR_2	GCGGTAAGTCACCCATCATG	1422971
SR_3	GCCGGAAACCCACCCAAAAC	1422947
SR_4	GCTGAATACAGACAAACCAAAAAG	1422925
SR_5	GTAATACTTTTGAATATGTCCGACC	1422901

Construction of mutants carrying in-frame deletions of sugR or NCgl1859.

The in-frame sugR deletion mutants of C. glutamicum were constructed via a two-step homologous recombination procedure as described previously (47). The sugR up- and downstream regions were amplified using primer pairs sugR-A/sugR-B and sugR-C/sugR-D. The product of the subsequent crossover PCR was subcloned into the pGEM-T vector (Promega, WI) and was cloned as an EcoRI-BamHI fragment into pK19mobsacB vector (60). Transfer of the resulting plasmid, pK19mobsacB-ΔsugR, into C. glutamicum by electroporation and screening for the correct mutants was performed as described previously (47). Of eight clones tested by PCR (primer pair sugR-k-for/sugR-k-rev), two showed the wild-type (WT) situation and six had the desired in-frame deletion of the sugR gene. Similarly, three independent NCgl1859 deletion mutants were constructed in the WT background as well as in the ΔsugR deletion mutant by using the primers 1859-A to -D.

Homologous overexpression of sugR.

For homologous overexpression of sugR, sugR was amplified from genomic DNA of C. glutamicum ATCC 13032 (referred to below as C. glutamicum WT) by using primers sugR-for and sugR-rev and was cloned into the expression vector pVWEx1 (51). The constructed vector, pVWEx1-sugR, allows the IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible expression of sugR in C. glutamicum.

Construction of transcriptional fusions.

The ptsG promoter region (−399 to +309 [see Fig. 4A]) including the first 16 codons of ptsG as well as promoter subfragment 4 (−125 to +62 [see Fig. 4A]) was amplified with primer pairs ptsG-1/ptsG-2 and F4-for/F4-rev, respectively (Table 2), and cloned into the corynebacterial promoter-probe vector pET2 (69). As the longer fragment contained ptsG coding sequence, a peptide of 46 amino acids containing the first 16 amino acids of PtsG may be formed. However, no fusion protein between chloramphenicol acetyltransferase and PtsG is formed. The resulting vectors, pET2-ptsG and pET2-ptsG/F4, were transferred into C. glutamicum WT, the ΔsugR mutant, and the Δ1859 deletion mutant by electroporation. The full-length promoter as well as subfragment 4 were tested for promoter activity by measuring chloramphenicol acetyltransferase activity (45).

FIG. 4.

Binding of SugR^His to the ptsG promoter. (A) DNA fragments (circled 1 to 6, A6, B1, B6, and BS) used to analyze SugR binding to the ptsG promoter by gel shift assays. The numbers indicate the ends of the fragments relative to the ptsG transcription start site (+1). Oligonucleotides used for amplification of the 10 fragments via PCR are listed in Table 2. The boxes labeled B indicate the regions where potential RamB binding sites are located. The sequence at the bottom shows the region between position −58 and −81 relative to the transcriptional start site of ptsG (bold) plus 10 bases upstream and downstream. The 8-bp motif contained within the putative binding site of SugR is underlined. (B) Gel showing binding of purified SugR^His (10- to 65-fold molar excess) to fragment 1 (10 nM, 707-bp fragment). A 190-bp NCgl2027 promoter fragment (40 nM) served as a negative control. (C) Gel showing binding of purified SugR^His (25- to 100-fold molar excess) to fragments A6, B1, B6, and BS (38, 64, 53 and 102 nM; 197-, 119-, 142-, and 75-bp fragments). (D) Gel showing binding of purified SugR^His (70-fold molar excess) to fragment 4 (40 nM, 186-bp fragment). A 400-bp NCgl1955 promoter fragment (19 nM) served as a negative control. The possible effectors of SugR shown here, tested at a concentration of 20 mM, were glucose-6-phosphate (G-6-P), frucose-6-phosphate (F-6-P), and phosphoenolpyruvate (PEP).

Overproduction and purification of SugR.

For overproduction and purification of SugR carrying an N-terminal decahistidine tag, the sugR coding region was amplified using oligonucleotides that introduce an NdeI restriction site, including the start codon (SugRshift-for) and a BamHI restriction site after the stop codon (SugRshift-rev) for cloning into pET16b. For overproduction of SugR, the plasmid was transferred into E. coli BL21(DE3), and the resulting strain, E. coli BL21(DE3)(pET16b-SugR^His), was grown in 2 liters LB medium or 1 liter Overnight Express Instant TB Medium (Merck Biosciences Ltd, Germany) with 50 μg/ml ampicillin and 10 ml of 100% glycerol according to the manufacturer's recommendations. For cell extract preparation, washed (in 20 ml TNI5 buffer [20 mM Tris-HCl {pH 7.9}, 300 mM NaCl, 5% {vol/vol} glycerol, 5 mM imidazole) and thawed cells (about 5 to 6 g [wet weight]) from a 0.5- to 2-liter culture volume were resuspended in 10 ml of TNI5 buffer containing 1 mM diisopropylfluorophosphate and 1 mM phenylmethylsulfonyl fluoride. The cell suspension was passed five times through a French pressure cell (SLM Amino; Spectronic Instruments, Rochester, NY) at 150.8 MPa. After centrifugation (20 min, 5,292 × g, 4°C) and ultracentrifugation (1 to 1.5 h, 150,000 × g, 4°C), the supernatant was used to purify SugR by nickel affinity chromatography (Ni-nitrilotriacetic acid; Novagen, San Diego, CA). After washing with TNI175 buffer (which contained 175 mM imidazole), the SugR^His protein was eluted with TNI400 buffer (which contained 400 mM imidazole). Fractions with the highest protein content were pooled, and the elution buffer was exchanged against BS buffer (100 mM Tris-HCl, 20% [vol/vol] glycerol, 100 mM KCl, 20 mM MgCl₂, 1 mM EDTA, pH 7.5) using PD10 columns.

Gel shift assays.

Gel shift assays with SugR^His were preformed as described previously (72). Briefly, various concentrations of purified SugR^His were mixed with full-length promoter DNA of ptsG; with promoter DNAs of ptsS, ptsF, and NCgl1858/1859; or with promoter subfragments of ptsG in 50 mM Tris-HCl, 10% (vol/vol) glycerol, 50 mM KCl, 10 mM MgCl₂, 0.5 mM EDTA, pH 7.5 (total volume of 20 μl). Different nontarget promoter fragments were added as a negative control. The promoter DNA fragments were obtained by PCR with the primers listed in Table 2, using the following combinations: ptsG-for/ptsG-rev, ptsG-for/F2-rev, F3-for/F3-rev, F4-for/F4-rev, F5-for/F5-rev, F6-for/ptsG-rev, F3-for/SR_1, F3-for/SR_2, F3-for/SR_3, F3-for/SR_4, F3-for/SR_5, SF_1/F4-rev, SF_2/F4-rev, SF_3/F4-rev, SF_4/F4-rev, SF_5/F4-rev, SR_5/SF_5 (BS), F3-for/A6, B1/F4-rev, B6/F4-rev, ptsF-for/ptsF-rev, ptsS-for/ptsS-rev, and NCgl1858/1859-for/NCgl1858/1859-rev. After incubation for 30 min at room temperature, the samples were separated on 10 or 15% native polyacrylamide gels at a constant voltage of 170 V, stained, and photographed as described previously (72). To test for possible effectors, the protein was incubated with glucose-6-phosphate, fructose-6-phosphate, fructose-1,6-bisphosphate, phosphoenolpyruvate, dihydroxyacetone phosphate, pyruvate, l-lactate, d-lactate, acetyl coenzyme A (acetyl-CoA), and 3-phosphoglycerate (20 mM each) in the binding buffer for 15 min before addition of DNA fragment 4 and incubation for another 30 min.

Preparation of total RNA and primer extension analysis.

Exponentially growing cells were harvested, and RNA was isolated using the RNeasy system (QIAGEN, Hilden, Germany) with on-column DNase I treatment as described previously (37). Purified RNA samples were analyzed for quantity and quality by UV spectrophotometry and stored at −20°C until use. Nonradioactive primer extension analysis with 10 to 13 μg of total RNA was performed using IRD800-labeled oligonucleotides (MWG Biotech, Ebersberg, Germany) as described previously (14), except with 200 U/μl SuperScript II (Invitrogen, Karlsruhe, Germany). The template for primer extension analysis for ptsG was designed by PCR amplification with the primers ptsG-for and ptsG-Primex-rev. Sequencing reactions and primer extension products were analyzed using denaturing 4.6% (wt/vol) Long Ranger (Biozym, Hamburg, Germany) sequencing gels in a Long Read IR DNA sequencer (Licor Inc., Lincoln, NE). The transcriptional start site of ptsG was determined using three different oligonucleotides (ptsG_30*, ptsG_60*, and ptsG_90*).

DNA microarray analysis.

Generation of C. glutamicum whole-genome DNA microarrays (70), synthesis of fluorescently labeled cDNA from total RNA, microarray hybridization, washing, and gene expression analysis were performed as described previously (23, 36, 37, 54). Genes that exhibited mRNA levels that were significantly changed (P ≤ 0.05 in Student's t test) by at least a factor of 2.0 were determined in three DNA microarray experiments performed with RNA isolated from three independent cultures.

Chloramphenicol acetyltransferase assay.

For determination of chloramphenicol acetyltransferase activity, the cell pellet from a 50-ml culture volume was washed in 40 ml 0.08 M Tris-HCl (pH 7.0) buffer, centrifuged for 5 min at 5,422 × g and 4°C, resuspended in 1 ml of the same buffer, and disrupted by sonication (9 min, cycle 0.5, amplitude of 55%, on ice) (sonication processor UP200s; Hielscher Ultrasonics GmbH, Stuttgart, Germany). After centrifugation (1 to 1.5 h, 13,000 × g, 4°C), the supernatant was used for measuring the chloramphenicol acetyltransferase activity by the method of Shaw (63). Briefly, the assay mixture (1 ml) contained 90 mM Tris-HCl (pH 7.8), 0.09 mM acetyl-CoA, 0.36 mg/ml 5,5′-dithiobis-2-nitrobenzoic acid, 0.25 mM chloramphenicol, and crude extract. The formation of free 5-thio-2-nitrobenzoate was measured photometrically at 412 nm and 37°C. Protein concentrations were determined with the Bradford assay kit (Bio-Rad Laboratories, Hercules, CA) using bovine serum albumin as standard.

Determination of glucose and acetate concentrations.

Samples (1 ml) of the cultures were centrifuged for 5 min at 16,060 × g, and aliquots of the supernatant were used directly for the determination. d-Glucose and acetate were quantified enzymatically with a d-glucose/d-fructose kit and an acetic acid kit, respectively (R-Biopharm, Darmstadt, Germany), as described by the manufacturer, by comparison of the sample probes with external standards.

RESULTS

Construction and growth properties of C. glutamicum sugR and NCgl1859 deletion mutants.

The genome of C. glutamicum type strain ATCC 13032 (25) contains four genes (NCgl0105, NCgl0110, sugR/NCgl1856, and NCgl1859) encoding putative DeoR-type transcriptional regulators (5). The genes for two of these regulators, sugR and NCgl1859, are located in the vicinity of the genes for the general PTS components ptsI (NCgl1858) and ptsH (NCgl1862) as well as the gene of the fructose-specific enzyme II permease ptsF (NCgl1861) and might therefore be involved in the regulation of sugar uptake via the PTS and/or the initial consumption of the sugar-phosphates. For functional analysis, in-frame deletion mutants of sugR and NCgl1859 and a double deletion mutant were constructed by two-step homologous recombination. The resulting mutants, i.e., the ΔsugR, Δ1859 and ΔsugR Δ1859 mutants, which lack the sugR and/or the NCgl1859 coding region except for the 6 5′-terminal and the 12 3′-terminal codons, were verified by PCR analysis (see Materials and Methods).

The growth behaviors of the ΔsugR, Δ1859 and ΔsugRΔ1859 mutants on different carbon and energy sources were compared to that of the WT. Only marginal differences between the ΔsugR and Δ1859 deletion mutants, the ΔsugR Δ1859 double deletion mutant, and the WT with respect to the growth rate and the biomass yield were observed on CgXII minimal medium with 100 mM glucose, 100 mM fructose, 50 mM sucrose, 50 mM glucose plus 50 mM fructose, 300 mM K-acetate, 100 mM Na₃-citrate plus 100 mM MgCl₂, 200 mM Na-pyruvate, 50 mM maltose, or 120 mM ribose (data not shown). Thus, neither sugR nor NCgl1859 is essential for growth on these carbon sources.

Expression of a ptsG′-′cat fusion in C. glutamicum WT and the ΔsugR and Δ1859 mutants.

To test whether SugR and/or NCgl1859 exerts control on expression of the gene for the glucose-specific EII, ptsG, reporter gene fusion analyses were performed. A fusion between the ptsG promoter region and the promoterless chloramphenicol acetyltransferase gene (cat) was constructed using the promoter-probe vector pET2 (69). The resulting plasmid, pET2-ptsG, was transferred into C. glutamicum WT and the ΔsugR and Δ1859 mutants.

Expression of the ptsG′-′cat fusion in C. glutamicum WT was higher on glucose minimal medium (1.8 ± 0.3 U/mg of protein) than on LB medium (1.3 ± 0.4 U/mg of protein) or on LB medium plus glucose (1.2 ± 0.2 U/mg of protein) (Fig. 1). In C. glutamicum Δ1859(pET2-ptsG), expression of the ptsG′-′cat was comparable to that observed in C. glutamicum WT(pET2-ptsG) during growth on LB medium plus glucose (1.1 ± 0.1 U/mg of protein), while it was about 1.5-fold higher on LB medium (1.9 ± 0.1 U/mg of protein). In C. glutamicum ΔsugR(pET2-ptsG), the ptsG′-′cat fusion showed an approximately twofold-increased expression on LB medium in comparison to the WT (2.5 ± 0.2 U/mg of protein) (Fig. 1). This difference in the expression of the ptsG′-′cat fusion was observed only on LB medium and not on either LB medium plus glucose (0.8 ± 0.1 U/mg of protein) or glucose minimal medium (2.0 ± 0.2 U/mg of protein) (Fig. 1). These results indicated a role of SugR in regulation of ptsG expression.

FIG. 1.

Expression of a ptsG′-′cat fusion in C. glutamicum WT(pET2-ptsG) and C. glutamicum ΔsugR(pET2-ptsG). Expression of the ptsG′-′cat reporter gene fusion was determined by measuring the chloramphenicol acetyltransferase activity after cultivation on 50 mM maltose (M), 100 mM glucose (G), 100 mM fructose (F), 50 mM sucrose (S), 120 mM ribose (R), 200 mM Na-pyruvate (P), 300 mM K-acetate (A), 100 mM Na₃-citrate plus 100 mM MgCl₂ (C), LB medium (LB), and LB medium plus glucose (LB + G). The values represent averages and standard deviations from three independent experiments. The numbers above the bars indicate the ratio of ptsG expressions between the deletion mutant ΔsugR and the WT.

As ptsG expression was increased in the ΔsugR deletion mutant in comparison to the WT on LB medium but was comparable in both strains when glucose was present in the medium, we tested several carbon sources concerning their effects on expression of a ptsG′-′cat transcription fusion in the C. glutamicum WT and the ΔsugR deletion mutant (Fig. 1). Expression of the ptsG′-′cat transcription fusion in C. glutamicum WT depended on the carbon source. Expression of ptsG′-′cat was highest on maltose (3.4 ± 0.1 U/mg of protein), i.e., about 2-fold higher than on glucose (1.8 ± 0.3 U/mg of protein), 3.2-fold higher than on fructose (1.1 ± 0.05 U/mg of protein) or sucrose (1.1 ± 0.02 U/mg of protein), and more than 5-fold higher than on ribose (0.65 ± 0.01 U/mg of protein). On the gluconeogenic carbon sources pyruvate (0.3 ± 0.1 U/mg of protein), acetate (0.3 ± 0.1 U/mg of protein), and citrate (0.4 ± 0.4 U/mg of protein), the expression of ptsG was only 8 to 12% of the expression observed on maltose.

Some of these effects are due to the presence of SugR, but the absence of SugR had little effect on maltose (1.2-fold higher than in C. glutamicum WT), glucose (1.1-fold), fructose (1.1-fold), sucrose (1.2-fold), ribose (1.6-fold), and LB medium plus glucose (0.8-fold) (Fig. 1). In contrast, expression of the ptsG′-′cat transcription fusion was derepressed in the ΔsugR deletion mutant on LB medium and on media containing gluconeogenic carbon sources, namely, pyruvate (5.8-fold; 1.7 ± 0.1 U/mg of protein), acetate (4.3-fold; 1.2 ± 0.1 U/mg of protein), and citrate (6.7-fold; 2.8 ± 0.2 U/mg of protein).

These results show that SugR represses ptsG expression when C. glutamicum grows on media containing gluconeogenic carbon sources but lacking sugars.

Effect of SugR on growth and glucose utilization.

To determine whether control of ptsG expression by SugR affects glucose utilization in vivo, we tested whether growth of C. glutamicum on LB medium with or without 100 mM glucose and on CgXII minimal medium either with 100 mM glucose or with 50 mM glucose plus 150 mM acetate is altered by overexpression of sugR. Therefore, we constructed C. glutamicum strains carrying either a sugR expression plasmid or, as a control, the expression vector alone. When the cells were grown on LB medium alone, the growth rate and final optical density of C. glutamicum WT(pVWEx1) (μ = 0.38 ± 0.02 h⁻¹; 1.9 ± 0.5 gDW liter⁻¹) were similar to those of the sugR-overexpressing strain C. glutamicum WT(pVWEx1-sugR) (μ = 0.35 ± 0.01 h⁻¹; 1.7 ± 0.1 gDW liter⁻¹) (Fig. 2A). But when glucose was present in the medium, the sugR overexpressing strain grew significantly slower (μ = 0.45 ± 0.05 h⁻¹) than the control strain WT(pVWEx1) (μ = 0.61 ± 0.03 h⁻¹). Moreover, the sugR overexpressing strain formed less than 1/3 of the biomass formed by the control strain (2.0 ± 0.1 compared to 7.5 ± 0.6 gDW l⁻¹) and utilized less than 15 mM of the 100 mM added glucose, while the control strain consumed the added glucose almost completely (Fig. 2A). In a similar experiment, it was shown that overexpression of sugR perturbed the growth of C. glutamicum on glucose minimal medium, as the biomass formed (4.6 ± 0.8 gDW liter⁻¹), the growth rate (μ = 0.17 ± 0.04 h⁻¹), and the glucose uptake rate (29 nmol mg⁻¹ min⁻¹) were reduced compared to those of the control strain (7.3 ± 0.4 gDW liter⁻¹, μ = 0.40 ± 0.02 h⁻¹, and 93 nmol mg⁻¹ min⁻¹). Growth inhibition during the first 3 h, i.e., prior to IPTG induction, might indicate that sugR expression is not completely repressed in plasmid pVWEx1-sugR.

FIG. 2.

Role of SugR in growth of C. glutamicum on glucose-containing media. (A) Growth of C. glutamicum WT(pVWEx1) (open symbols) and WT(pVWEx1-sugR) (filled symbols) on LB medium (left panel), LB medium with 100 mM glucose (middle panel), and CgXII minimal medium containing 100 mM glucose (right panel). The cultures were induced 3 hours after inoculation by addition of 1 mM IPTG. The ODs (triangles) and the glucose concentrations (circles) are indicated. (B) Growth of C. glutamicum WT (open symbols, left panel) and C. glutamicum ΔsugR (filled symbols, left panel) as well as of C. glutamicum WT(pVWEx1) (open symbols, right panel) and WT(pVWEx1-sugR) (filled symbols, right panel) on CgXII minimal medium containing 50 mM glucose and 150 mM K-acetate. The ODs (triangles), the glucose concentrations (circles), and the K-acetate concentrations (squares) are indicated. The cultures of C. glutamicum WT(pVWEx1) and WT(pVWEx1-sugR) were induced 3 hours after inoculation by addition of 1 mM IPTG. Values are averages and standard deviations from at least two independent cultivations are indicated.

On minimal medium containing 50 mM glucose and 150 mM acetate, WT as well as WT(pVWEx1) simultaneously utilized glucose (uptake rates of 43 and 43 nmol mg⁻¹ min⁻¹) with acetate (132 and 130 nmol mg⁻¹ min⁻¹) (Fig. 2B). On the glucose-acetate mixture, WT ΔsugR grew with a growth rate similar to that of WT (0.35 and 0.33 h⁻¹) but showed two phases with respect to carbon source utilization (Fig. 2B). Within the first 9 h of cultivation, glucose was utilized by WT ΔsugR with an uptake rate higher than that by WT (59 compared to 43 nmol mg⁻¹ min⁻¹), while acetate uptake was reduced (81 compared to 130 nmol mg⁻¹ min⁻¹). After glucose was exhausted, growth slowed and only acetate was utilized (34 nmol mg⁻¹ min⁻¹). Thus, in the absence of SugR, glucose and acetate are coutilized, but the proportion between glucose and acetate is shifted towards glucose (glucose/acetate ratio of 42:58 compared to 25:75). On the glucose- and acetate-containing medium, overexpression of sugR resulted in slow growth (0.09 compared to 0.33 h⁻¹) (Fig. 2B). First, glucose and acetate were coutilized (13 and 68 nmol mg⁻¹ min⁻¹, respectively) until acetate was exhausted, and then only glucose was utilized (36 nmol mg⁻¹ min⁻¹). Overexpression of sugR thus perturbed growth on the glucose-acetate mixture and shifted the proportion of glucose and acetate coutilization towards acetate (16:84 compared to 25:75).

Taken together, the results show that overexpression of sugR has a negative influence on utilization of glucose as a carbon source for growth of C. glutamicum.

Identification of the transcriptional start site of ptsG.

In order to determine whether overexpression of sugR reduces ptsG transcript levels and to determine the transcriptional start site of ptsG, primer extension analyses were performed. In these experiments, RNAs from C. glutamicum WT grown on LB medium and on glucose minimal medium, as well as RNAs from C. glutamicum WT(pVWEx1) and C. glutamicum WT(pVWEx1-sugR) grown on glucose or acetate minimal medium, were used.

A single primer extension product was detected using three independent primers (ptsG_30*, ptsG_60*, and ptsG_90* [Table 2]) when C. glutamicum WT was grown on glucose minimal medium or on LB medium (Fig. 3A and data not shown). The transcriptional start site identified by the primer extension experiments is located 258 bp upstream of the ATG start codon of ptsG. Analysis of the promoter region (Fig. 3B) revealed a sequence (5′-TATCAT-3′) similar to the consensus −10 sequence motif, 5′-tgngnTA(c/t)aaTgg-3′ (Uppercase, conserved; lowercase, less conserved), of C. glutamicum promoters (50) but no obvious −35 region as is typical for C. glutamicum promoters.

FIG. 3.

Expression analysis and determination of the transcriptional start site of the C. glutamicum ptsG gene. (A) For primer extension analysis, 13 μg of total RNA isolated from C. glutamicum WT grown on LB medium (lane 1) and on CgXII minimal medium containing 100 mM glucose (lane 2), as well as 10 μg of total RNA from C. glutamicum WT(pVWEx1) grown on glucose (lane 3) and on CgXII minimal medium containing 300 mM acetate (lane 4) and from C. glutamicum WT(pVWEx1-sugR) grown on glucose (lane 5) and acetate (lane 6) were used. The transcriptional start site is indicated by an asterisk. The corresponding sequencing reactions (lanes T, G, C, and A) were performed using the same IRD800-labeled oligonucleotides that were used in the primer extension reactions and PCR products which covered the region of the corresponding transcriptional start site as template DNA. (B) Sequence of the intergenic region between NCgl1304 and ptsG. The start codon of ptsG is indicated in bold, the transcriptional start site is shown in bold and marked with an asterisk, the putative −10 region is highlighted in gray, and the putative RamB binding motifs are underlined.

In contrast to the empty vector control strain C. glutamicum WT(pVWEx1), nearly no ptsG primer extension product could be detected in the sugR-overexpressing strain C. glutamicum WT(pVWEx1-sugR) on glucose minimal medium. Thus, SugR negatively regulates ptsG expression in a direct or indirect manner. Only a very small amount of ptsG primer extension product was observed in the control strain when the cells were grown on acetate minimal medium, and no primer extension product could be detected in the sugR-overexpressing strain on acetate minimal medium (Fig. 3A). Taken together, the primer extension data support the assumption that expression of ptsG is dependent on the amount of SugR, which may function as a repressor of ptsG transcription.

Binding of purified SugR^His to the ptsG promoter region.

To test whether the influence of SugR on ptsG expression is direct, we assayed the binding of purified SugR to the ptsG promoter in vitro. For that purpose, in several independent experiments the SugR protein containing an amino-terminal decahistidine tag (SugR^His) was overproduced in E. coli BL21(DE3), and 0.33 to 0.44 mg of SugR^His was purified to apparent homogeneity by affinity chromatography as described in Materials and Methods.

For the gel shift assays, the DNA fragments (10 to 102 nM) were mixed with various concentrations of the SugR^His protein (0 to 10 μM) and then separated on 10 or 15% native polyacrylamide gels. As shown in Fig. 4B, SugR^His bound the full-length ptsG promoter in vitro. A complete shift was observed at a 65-fold molar excess, whereas the negative control, a 190-bp promoter fragment of NCgl2027, was not shifted by SugR^His.

Binding of SugR to various subfragments of the ptsG promoter region was tested in order to confine the SugR binding site. Therefore, the full-length promoter fragment 1 (Fig. 4A) was first divided into five fragments (named fragments 2 to 6 [Fig. 4A]), overlapping by approximately 50 bp, and tested for SugR binding. SugR bound to fragments 3 and 4 (data not shown). Gel shift assays with 10 additional subfragments of the ptsG promoter covering different parts of fragments 3 and 4 (SF1 to -5 and SR1 to -5) suggested that SugR binds to a sequence between positions −132 and −58 relative to the transcriptional start site of ptsG (data not shown). The 75-bp fragment BS covering the DNA region from position −132 to −58 was sufficient for SugR binding (Fig. 4A and C). In binding assays with fragment BS, SugR also bound to fragment B6 but did not bind to fragments B1 and A6 (Fig. 4A and C). Thus, the sequence from bp −81 to −58, which is present in fragments BS and B6 but absent from fragments B1 and A6, likely contains the SugR binding site.

To characterize the relevance of the SugR binding site in vivo, a fusion of the ptsG promoter fragment 4, which contains the transcriptional start site and the SugR binding site, to the promoterless chloramphenicol acetyltransferase gene was constructed, and its expression was tested in C. glutamicum WT and the ΔsugR mutant. Expression of the ptsG/F4′-′cat transcriptional fusion was derepressed in the absence of SugR during growth on LB medium (2.3 ± 0.7 U/mg of protein in WT compared to 5.0 ± 0.5 U/mg of protein in the ΔsugR mutant) and on acetate minimal medium (0.4 ± 0.5 U/mg of protein in WT compared to 4.9 ± 1.5 U/mg of protein in the ΔsugR mutant), but expression was comparable during growth on LB medium plus glucose regardless of the presence or absence of SugR (2.0 ± 0.1 U/mg of protein in WT compared to 1.6 ± 0.04 U/mg of protein in the ΔsugR mutant). Although under the tested conditions expression of the ptsG/F4′-′cat transcriptional fusion was generally higher in both C. glutamicum WT and the ΔsugR mutant compared to the ptsG′-′cat reporter gene fusion (Fig. 1), expression of both fusions depended on SugR in a comparable manner. Thus, ptsG promoter fragment 4 contains the sequences required for initiation and SugR-dependent control of ptsG transcription.

Fructose-6-phosphate prevents binding of SugR to the ptsG promoter.

The carbon source-dependent effects of the absence of SugR on ptsG expression suggested that the binding of SugR might be affected by an inducer molecule. To test this hypothesis and to identify such an effector molecule, we assayed binding of SugR^His protein to the ptsG promoter in the presence of several intermediates of the central metabolism (Fig. 4D). The possible effectors (20 mM each; see Materials and Methods) were incubated with the protein in the binding buffer for 15 min before addition of ptsG promoter fragment 4 (40 nM), and after incubation for another 30 min, free DNA and protein-DNA complexes were separated by 15% native polyacrylamide gel electrophoresis.

Neither glucose-6-phosphate nor phosphoenolpyruvate had an effect on the binding affinity of SugR to the ptsG promoter (Fig. 4D). Similarly, addition of dihydroxyacetone phosphate, fructose-1,6-bisphosphate, pyruvate, l-lactate, d-lactate, acetyl-CoA, or 3-phosphoglycerate did not affect formation of a SugR/ptsG promoter complex (data not shown). The inducer of SugR could be identified as fructose-6-phosphate, as at a 70-fold molar excess of protein to DNA the binding of SugR to ptsG promoter fragment 4 was inhibited almost completely at a concentration of 20 mM fructose-6-phosphate (Fig. 4D).

Comparison of the expression profiles of C. glutamicum WT and the ΔsugR mutant by use of DNA microarrays.

To investigate the effect of the sugR deletion on global gene expression and to identify further putative target genes of SugR, the transcriptomes of C. glutamicum WT and the ΔsugR mutant were compared using whole-genome DNA microarrays (70). RNA was isolated from cells grown on LB medium in the exponential growth phase. Table 3 shows those genes whose mRNA level was significantly (P < 0.05) changed by a factor of two or more in three biological replicates. Of these 25 genes, 14 genes showed a higher mRNA level and 11 genes a lower mRNA level in the ΔsugR mutant.

TABLE 3.

Genes whose average mRNA ratio was altered ≥2-fold (P ≤ 0.05) on LB medium in C. glutamicum WT compared with the ΔsugR mutant

NCgl no.a	Annotationa	Ratio (WT/ΔsugR)b
0468	LSU ribosomal protein L10P	0.49
0492	LSU ribosomal protein L22P	0.42
0627	2-Methylisocitrate synthase, prpD2	2.56
0630	2-Methylcitrate synthase, prpC2	2.61
0917	Hypothetical cytosolic protein	3.82
0959	Phosphinothricin N-acetyltransferase	0.31
0960	Allophanate hydrolase subunit 2	0.33
0962	Lactam utilization protein, LamB	0.16
0963	Transporter	0.08
0965	Transcriptional regulator, GntR family	0.49
1206	ABC-type sugar transport system, periplasmic component	0.48
1305	Glucose-specific PTS enzyme II, ptsG	0.43
1368	Acetyltransferase	6.59
1418	Hypothetical secreted protein	2.08
1857	1-Phosphofructokinase homolog	0.37
1859	Transcriptional regulator, DeoR family	0.21
1861	Fructose-specific PTS enzyme II, ptsF	0.17
2252	Hypothetical cytosolic protein	2.36
2439	Ferritin-like protein, ftn	2.01
2477	Succinyl-CoA synthetase beta chain, sucC	2.47
2552	Cysteinyl-tRNA synthetase	0.43
2553	Sucrose-specific PTS enzyme II, ptsS	0.06
2657	Phosphate acetyltransferase, pta	2.43
2787	Predicted flavoprotein involved in K+ transport	2.98
2804	DNA binding protein	2.05

^aGene numbering, gene designations, and descriptions of gene products are given according to the National Center of Biotechnology (NCBI) for the genome of C. glutamicum ATCC 13032 (GI:58036263).

^bThe mRNA ratios represent average values obtained from three DNA microarray experiments performed with RNA isolated from three independent cultures in LB medium.

The mRNA level of ptsG was about twofold increased in the ΔsugR mutant compared to the WT, verifying the previous results that ptsG is repressed by SugR. Furthermore, the PTS genes ptsS (sucrose-specific PTS IIABC component) and ptsF (fructose-specific PTS IIABC component) with the adjacent NCgl1857 (1-phosphofructokinase-like protein) (42) and NCgl1859 (transcriptional repressor DeoR family) showed higher mRNA levels in the ΔsugR mutant than in the WT. Besides genes of propionate metabolism (coding for 2-methylisocitrate synthase and 2-methylcitrate synthase) and genes of the tricarboxylic acid (TCA) cycle (coding for succinyl-CoA synthetase), a cluster of genes of unknown function, NCgl0959 to NCgl0965, showed higher mRNA levels in the ΔsugR mutant than in the WT. While it is clear that SugR regulates ptsG expression directly, the effect of SugR on expression of these genes might be either direct or indirect.

Binding of purified SugR^His to additional target promoters.

To determine whether SugR regulates expression of ptsS and the putative NCgl1859-fruK-ptsF operon directly, binding of SugR to the corresponding promoter regions was tested. As shown in Fig. 5, ptsS was shifted completely by SugR at a 50-fold molar excess, whereas the negative control, a 400-bp promoter fragment of NCgl1955, was not bound by SugR. The DNA region upstream of NCgl1859-fruK-ptsF (506 bp), but not the region upstream of ptsF (203 bp), was bound by SugR at a 70-fold molar excess. Binding of SugR to the DNA regions upstream of ptsS and NCgl1859-fruK-ptsF in vitro (Fig. 5) and their increased RNA levels in the absence of SugR (Table 3) indicate that SugR is a global regulator repressing transcription of ptsG, ptsS, and NCgl1859-fruK-ptsF.

FIG. 5.

Binding of SugR^His to the promoter fragments of ptsS and the putative NCgl1859-fruK-ptsF operon. On the left side, a gel demonstrating binding of purified SugR^His (25- to 70-fold molar excess) to the intergenic region between NCgl1858 and NCgl1859 (15 nM, 506 bp) is shown. The ptsF fragment (37 nM, 203 bp) was not bound by SugR^His. On the right side, a gel demonstrating binding of purified SugR^His (25- to 70-fold molar excess) to the ptsS promoter fragment (29 nM, 264 bp) is shown. A 400-bp NCgl1955 promoter fragment (19 nM) served as a negative control.

DISCUSSION

In this study, we have shown that the DeoR-type transcriptional regulator SugR (NCgl1856) regulates ptsG in a carbon source-dependent manner. Thus, we identified a novel transcriptional regulator of the central carbon metabolism of C. glutamicum. The inducer of SugR could be identified as fructose-6-phosphate, since of the glycolytic intermediates tested only fructose-6-phosphate inhibited binding of SugR to the ptsG promoter (Fig. 4D).

While regulation of glucose uptake in C. glutamicum is only poorly understood, in E. coli CRP and Mlc are known to be the major regulators of ptsHIcrr (coding for enzyme I, HPr, and EIIA^Glc) and ptsG (coding for EIICB^Glc) in response to the availability of carbon sources (18, 53). Transcription of ptsG is stimulated by cAMP · CRP and is repressed by Mlc (18, 30, 52, 53, 57, 76). Mlc binds upstream of ptsG and represses its transcription in the absence of glucose. During growth on glucose, EIICB^Glc occurs in its unphosphorylated form due to glucose uptake and binds to Mlc, which results in derepression of ptsG (4, 38). FIS, a nucleoid-associated protein, facilitates rapid adaptation of E. coli to different carbon and energy sources through the formation of nucleoprotein complexes either with cAMP · CRP in the presence of glucose or with Mlc in the absence of glucose (64). In addition, expression of ptsG is regulated posttranscriptionally via modulation of ptsG mRNA stability by RNase E-mediated mRNA degradation in response to the glycolytic flux (24, 31). Moreover, ptsG transcription appears to be inhibited by the response regulator ArcA, which is phosphorylated and activated by the histidine kinase ArcB under reducing conditions. Phosphorylated ArcA binds to the CRP binding site of ptsG, thus interfering with CRP activation of ptsG transcription (24). In B. subtilis, ptsGHI transcription terminates in the absence of glucose due to the inactivation of the RNA binding antitermination protein GlcT by phosphorylation by the phosphorylated EII^Glc. In the presence of the inducer glucose, phosphorylated EII^Glc transports and phosphorylates the incoming glucose. Under these conditions, the GlcT protein is not phosphorylated and binds to the ribonucleic antiterminator (RAT) sequence of the nascent ptsGHI operon RNA and prevents premature termination of transcription (61, 62, 67). Although not characterized functionally, it was found that the upstream region of ptsG from Corynebacterium diphtheriae contains a RAT-like sequence, and a GlcT-like protein (DIP1150) is encoded in its genome (49). However, the C. glutamicum genome does not code for a functional GlcT-like protein, as the putative 197-amino-acid protein encoded by NCgl2743, which shows the highest sequence similarity to amino acids 93 to 275 of C. diphtheriae DIP1150, lacks the 92 N-terminal amino acids of C. diphtheriae DIP1150, including the RNA binding domain (25). In addition, the region upstream of C. glutamicum ptsG is lacking a RAT-like sequence (49). As C. glutamicum lacks functional homologs of regulatory proteins such as Mlc, CRP, CcpA, or GlcT and as in neither E. coli nor B. subtilis was a DeoR-type regulator shown to be involved in the regulation of PTS genes, the mechanism of ptsG regulation in C. glutamicum is distinct from that in other bacteria.

SugR belongs to the transcriptional regulators of the DeoR family, which in most instances act as transcriptional repressors in sugar metabolism (44, 75). For example, DeoR of B. subtilis represses the transcription of the dra-nupC-pdp operon, which is important for utilization of deoxyribonucleosides and deoxyribose (75). Molecular genetic studies indicated that a palindromic sequence located between nucleotides −60 and −43 relative to the transcriptional start site of the dra-nupC-pdp operon as well as a direct repeat of the 3′ half of the palindrome located between the −35 and the −10 regions [5′-ATTGAA-(6)-TTCAAT-(16)-TTCAA-3′] were both required for repression of this operon by DeoR (74). DeoR of E. coli, on the other hand, represses the initiation of transcription of the deo operon, which consists of four structural genes encoding ribonucleoside- and deoxyribonucleoside-catabolizing enzymes (44). The DeoR protein of E. coli binds to the putative 16-bp palindromic sequence 5′-TGTTAGAA · TTCTAACA-3′ in either of two of the three operator sites O₁, O₂, and O_E, forming a single or double DNA loop (44). By gel shift experiments, the 75-bp ptsG promoter fragment BS was shown to be sufficient for SugR binding, and additional experiments suggested that SugR binds to a DNA region from bp −81 to −58 (Fig. 4A and C). As SugR also bound to DNA regions upstream of ptsS and the putative NCgl1859-fruK-ptsF operon, we compared the sequences for similarities by using the MEME software (http://bioweb.pasteur.fr/seqanal/motif/meme/meme.html). A partly conserved 8-bp motif was found upstream of ptsG (5′-GTCGGACA-3′ within the DNA region from bp −81 to −58 [Fig. 4A]), of ptsS (5′-TGTACAAA-3′) and of NCgl1859-fruK-ptsF (5′-TGTGCAAC-3′). Thus, it is likely that the SugR binding sites contain this 8-bp motif.

Although the DeoR proteins of E. coli and B. subtilis share little sequence similarity and the DNA sequences to which they bind are dissimilar, their binding is inhibited by the same effector, i.e., deoxyribose-5-phosphate (44, 75). Effector molecules for DeoR-type regulators are generally phosphorylated intermediates of the relevant metabolic pathways (including, for example, besides deoxyribose-5-phosphate, fructose-1-phosphate for FruR of Lactococcus lactis [2]). However, nonphosphorylated inducers also have been described, e.g., opine for AccR from Agrobacterium tumefaciens (3), fucose for FucR from Bacteroides thetaiotaomicron (22), and likely N-acetylglucosamine or galactosamine for AgaR from E. coli (55). In C. glutamicum, the sugar phosphate fructose-6-phosphate was shown to inhibit binding of SugR to the ptsG promoter, thus linking regulation of glucose uptake to the first glycolytic intermediate after the branch point to the pentose phosphate pathway (Fig. 4D).

Intracellular fructose-6-phosphate concentrations of 13 mM, 1 mM, and 5 mM were determined in C. glutamicum DM1730 during growth on glucose, sucrose, and fructose, respectively (15), and these are in the same range as the fructose-6-phosphate concentrations affecting binding of purified SugR^His to the ptsG promoter in vitro. During growth on fructose and sucrose, which enter glycolysis as fructose-6-phosphate and/or fructose-1,6-bisphosphate, slightly lower expression levels of the ptsG′-′cat fusion than on glucose were observed, but this difference was not dependent on the presence of SugR (Fig. 1). Similarly, the absence of SugR did not affect expression of the ptsG′-′cat fusion during growth on glucose, as under these conditions fructose-6-phosphate should prevent binding of SugR to the ptsG promoter. However, during growth on the gluconeogenic carbon sources pyruvate, acetate, and citrate, expression of the ptsG′-′cat fusion, which was lower than that during growth on sugars, increased four- to sevenfold in the absence of SugR (Fig. 1). Although intracellular fructose-6-phosphate concentrations have not been reported for C. glutamicum grown on gluconeogenic carbon sources, they are expected to be much lower than those during growth on sugars. Thus, the observed expression levels of the ptsG′-′cat fusion in C. glutamicum WT and the ΔsugR mutant during growth on sugars or gluconeogenic carbon sources can be explained by fructose-6-phosphate-dependent regulation by SugR.

The transcriptome comparison of C. glutamicum WT and the ΔsugR mutant during growth on LB medium (Table 3) and binding of purified SugR to DNA regions upstream of ptsS and the putative operon NCgl1859-fruK-ptsF indicate a role for SugR as a global regulator of glucose, fructose, and sucrose uptake. The transcriptome comparison also revealed direct or indirect regulation of the putative operon NCgl0959-NCgl0963 and NCgl0965, which code for proteins of unknown function and showed increased expression in the absence of SugR. In contrast, the mRNA level of, e.g., the phosphotransacetylase gene pta was lower in the absence of SugR. The pta-ack operon (mRNA levels for ack could not be determined), which also encodes acetate kinase, is required for acetate activation in C. glutamicum (56). On a glucose-acetate mixture, glucose uptake was increased in the absence of SugR due to derepression of ptsG. It is conceivable that the decreased acetate uptake observed under these conditions is due to direct or indirect regulation of pta-ack expression by SugR. Commensurate with this view, overexpression of sugR reduced glucose uptake during growth on a glucose-acetate mixture, while acetate uptake was increased.

Regulation of ptsG in C. glutamicum does not depend solely on SugR but appears to be more complex. Expression of the ptsG′-′cat fusion during growth on the sugars maltose, glucose, fructose, and sucrose was similar in the absence and presence of SugR but was twofold and fourfold higher on maltose than on glucose and fructose or sucrose, respectively (Fig. 1). Gerstmeir et al. (16) suggested a putative binding site for the global regulator RamB in the ptsG promoter region. In C. glutamicum, repression of genes of acetate metabolism by RamB was observed only during growth on glucose, i.e., in the absence of acetate (16). As the putative binding sites of RamB are located downstream of the transcriptional start site of ptsG (Fig. 3), RamB might function as a negative regulator of ptsG. It is noteworthy that expression of the transcriptional fusion containing ptsG promoter fragment 4, which lacks the putative RamB binding sites, was about twofold higher than expression of the full-length ptsG promoter fusion.

In summary, we have identified the DeoR-type regulator SugR as a transcriptional regulator of ptsG, ptsS, and NCgl1859-fruK-ptsF expression in C. glutamicum. Fructose-6-phosphate interferes with SugR binding to the promoter of ptsG, and thus ptsG is derepressed during growth on sugars or under conditions characterized by high fructose-6-phosphate concentrations. During growth on gluconeogenic carbon sources such as acetate, pyruvate, or citrate, however, ptsG is repressed by SugR. While SugR is the first transcriptional regulator of ptsG described to date for C. glutamicum, carbon source-dependent differences of ptsG expression in the absence of SugR indicate the involvement of additional regulatory systems allowing C. glutamicum to fine-tune ptsG expression according to the availability of carbon and energy sources.