Regulation of glnA expression and glutamine synthetase I activity was analyzed in Corynebacterium glutamicum. Transcription is regulated by the global repressor protein AmtR, essential for derepression of glnA transcription are GlnK and uridylyltransferase, key proteins of the C. glutamicum nitrogen regulatory system. Glutamine synthetase I activity is controlled by adenylylation/deadenylylation via adenylyltransferase. The gene encoding this bifunctional enzyme, glnE, was isolated and its function was characterized by deletion analysis. Upstream of glnE, a second gene encoding a GSI-type protein in C. glutamicum was isolated. This gene, designated glnA2, forms an operon with glnE, its transcription is not regulated and neither its deletion or overexpression showed any effect. Therefore, the physiological role of glnA2 remains unclear.
Corynebacterium glutamicum, a Gram-positive soil bacterium which belongs to the mycolic acid containing actinomycetes, is used for industrial production of amino acids, especially l-glutamate and l-lysine. One of the key steps in amino acid production and nitrogen metabolism in general is the assimilation of ammonium which can occur via two different pathways: under conditions of excess nitrogen supply, ammonium is fixed in the form of glutamate by glutamate dehydrogenase (GDH). Under nitrogen starvation, however, ammonium is assimilated to glutamine by glutamine synthetase I (GSI) and subsequently metabolized to glutamate by glutamate synthase (GOGAT). As assimilation via the GS/GOGAT pathway is very energy-intensive, a strict regulation on the level of activity and transcription depending on the amount of nitrogen is necessary. In a number of bacterial species like Escherichia coli, Streptomyces coelicolor and Thiobacillus ferrooxidans glutamine synthetase I is regulated in its activity depending on the nitrogen supply by covalent addition of an adenyl residue via an adenylyltransferase. The transcriptional regulation is complex and repression mechanisms seem to be predominant (for review see . In Bacillus subtilis for example the transcription of the gene encoding glutamine synthetase I (glnA) is repressed under nitrogen excess by the action of the specific regulator GlnR, the gene of which is organized in an operon with the glnA gene and is subject to autoregulation. Under nitrogen starvation glutamine synthetase I produces a signal which derepresses transcription of the glnRA operon . Also for S. coelicolor and Bacillus cereusglnA-specific transcriptional regulators have been identified. The gene encoding glutamine synthetase I in C. glutamicum, glnA, was isolated recently . It was shown that GSI of C. glutamicum is regulated on the level of activity and might be modified by reversible adenylylation . Additionally, Southern hybridization experiments indicated that a second glutamine synthetase-encoding gene might exist in C. glutamicum.
In C. glutamicum transcription of a considerable number of genes is repressed in nitrogen-rich medium. In a recent publication, we demonstrated that all these genes are under control of a single repressor protein, designated AmtR . Essential for derepression of these AmtR-controlled genes are the GlnK protein and uridylyltransferase. The glnK gene was originally designated glnB, according to the designation in the closely related Mycobacterium tuberculosis. Based on a number of homologs cloned in different bacteria, sequence similarity analyses, and its genetic organization this system is designated glnK now .
We were interested if glnA transcription might also be under control of AmtR and the GlnK/uridylyltransferase regulatory system. Additionally, we investigated the regulation of glutamine synthetase I on the level of activity and identified a second GSI-type protein.
2Materials and methods
2.1Strains and growth conditions
Strains and plasmids used in this study are listed in Table 1. C. glutamicum was routinely grown in BHI medium (Difco, Detroit, MI, USA) at 30°C, E. coli was grown in LB medium  at 37°C. If appropriate, antibiotics were added in standard concentrations . In order to induce nitrogen starvation, C. glutamicum cells were precultured in BHI medium, then grown overnight in minimal medium . This culture was used to inoculate fresh minimal medium to an OD600 of approximately 1 and cells were incubated until the exponential growth phase was reached (OD600 approximately 4–5). To induce nitrogen starvation, cells were transferred to minimal medium without nitrogen source .
Table 1. Strains and plasmids used in this study
Abbreviations: ApR, KmR, NxR: resistance to ampicillin, kanamycin, nalidixic acid.
pUC18 carrying a 1-kb fragment used for glnA–lacZ fusion
pUC18 carrying a 4-kb glnA–lacZ fusion
pUCglnA carrying XbaI and XhoI restriction site 3 bp downstream of the glnA start codon
pUC18 carrying a 4-kb fragment containing glnE and its flanking regions
pUC18 carrying a 2.8-kb fragment containing glnA2 and its flanking regions
pZ8-1 carrying complete glnA gene for overexpression
pZ8-1 carrying complete glnA2 gene for overexpression
2.2RNA preparation, RNA hybridization analyses and RT-PCR
Total RNA from C. glutamicum cells was prepared using the NucleoSpin RNAII kit (Macherey-Nagel, Düren, Germany). 10 ml of a C. glutamicum culture (OD600 of approximately 4–5) were centrifuged at 4°C and the cells were transferred to an 2-ml cryo vial containing 300 mg glass beads and 700 μl RA1 buffer. After disruption in an amalgamator for 30 s the suspension was centrifuged (2 min, 4°C, 14 000×g) and the supernatant was worked up with the NucleoSpin RNAII kit. For transcription analysis the RNA was spotted directly onto Nylon membranes using a Schleicher and Schüll (Dassel, Germany) Minifold I dot blotter and hybridized overnight with the appropriate RNA probe. For RT-PCR the OneStep RT-PCR Kit (Qiagen, Hilden, Germany) was used. By PCR the absence of chromosomal DNA contaminations in the total RNA used as template was verified. For the analysis of glnA2 and glnE transcription the following primers were used: 5′-GTGAATTCTGGAGCGTGATGCA-3′/5′-ATCAGCTGCAGATGAAGGCT-3′ (glnA2) and 5′-GTGGCGCTCTTGTGCGCT-3′/5′-GCTGACTATGCAACGACTGAA-3′ (glnE).
For real-time PCR a Lightcycler (Roche, Mannheim; Germany) and the following primer combinations were used: 5′-GCATAACTTGAGTGCTGTAGGG-3′/5′-GCTGGCAACATAAGACAAGG-3′ (16S rRNA), 5′-TAGCACCATACGACCAAACC-3′/5′-TGATTGGGATACGGACAGC-3′ (glnA). A first denaturation step was carried out at 95°C for 30 s, followed by 45 PCR cycles of 1 s 95°C, 10 s 60°C, and 20 s 72°C.
2.3Determination of the transcriptional start site
For the determination of the transcriptional start site of glnA the 5′–3′RACE kit from Roche Diagnostics (Mannheim, Germany) was used according to the instructions of the manufacturer. For the determination about 5 μg of RNA isolated from C. glutamicum cells grown under nitrogen excess or from nitrogen-starved cells and appropriate primers (5′-GGAAGGTCGGTGAATCGAACGT-3′; 5′-ACGAACTCGACGTTTTCATCCTTGAT-3′) were applied. The isolated and purified PCR products were cloned into pUC18 and sequenced directly using pUC18 sequencing primers 5′-AGCGGATAACAATTTCACACAGGA-3′ and 5′-GTAAAACGACGGCCAGT-3′.
About 10 ng of the DNA fragment, which was labelled with digoxigenin during PCR via DIG-labelled primers (5′-DIG-ATCAGCTGGAATTTCGGGTCCGTCAA-3′; 5′-DIG-ACGAACTCGACGTTTTCATCCTTGAT-3′), were incubated in TEK buffer (10 mM Tris–HCl, pH 8, 10 mM KCl, 1 mM β-mercaptoethanol) with raw extract from E. coli cells containing either the amtR gene on plasmid pUC18 or, for control, only the vector pUC18. Protein amounts varied from 5 to 30 μg per assay, native electrophoresis was carried out using a 4% polyacrylamide gel and run at 15 mA in 0.5×TBE buffer . After electrophoresis the DNA was blotted onto a nylon membrane (Roche Diagnostics, Mannheim, Germany). Detection of labelled DNA was carried out using the DIG labelling and detection kit (Roche Diagnostics, Mannheim, Germany) according to the instructions of the manufacturer.
2.5Determination of enzyme activity and protein concentrations
β-Galactosidase activity was assayed using the method described by Miller . Glutamine synthetase activity was determined using the ‘in vivo-like’ method . For this purpose, cell extracts were prepared by ultrasonic cell disruption followed by low speed centrifugation to remove cell debris. GS activity was determined in the presence of 1 mM MnCl2 by a coupled photometrical test. In the first reaction step, GS metabolizes glutamate, ammonium, and ATP. ADP generated in this reaction is phosphorylated by pyruvate kinase. In this reaction the donor of the phosphoryl group, phosphoenol pyruvate, is converted to pyruvate which is metabolized by lactate dehydrogenase to lactate. In this reaction NADH is consumed and the decrease of this compound is measured photometrically. The protein contents of cell extracts were determined using a modified Bradford assay (Roti Nanoquant, Roth, Karlsruhe, Germany).
2.6General molecular biology techniques
For plasmid isolation, transformation and cloning standard techniques were used . Electrocompetent cells of C. glutamicum were prepared according to . DNA sequence determination was carried out using the BigDye terminator kit and an ABI 310 automated sequencer (Applied Biosystems, Weiterstadt, Germany).
2.7Construction of unmarked deletion mutants
Chromosomal deletions of the glnA2 and glnE gene were introduced in the C. glutamicum genome according to the protocol described  using plasmid pK18mobsacB. All deletions were verified by PCR (data not shown). For glnA2 deletion a 3.4-kb fragment containing the glnA2 gene and its flanking regions was amplified by PCR using the primers 5′-TCTGGGCGAAGTGCTGCAT-3′ and 5′-GTTCTCTAGGACAGGAACGA-3′, cut with KpnI and BamHI and the resulting 2.8-kb fragment was ligated to KpnI/BamHI-restricted and dephosphorylated pUC18 DNA leading to plasmid pUCglnA2. By the PCR-based method of Imai et al. , a deletion of 1.3 kb was introduced in the glnA2 gene using the primers 5′-AGCAGCCTTGATCAGGCACT-3′ and 5′-GCGCTGAGTACAAATTCCTGT-3′. After religation which led to plasmid pUCΔglnA2 the insert containing the deletion was cut out by EcoRI/BamHI restriction and ligated to EcoRI/BamHI-restricted and dephosphorylated pK18mobsacB leading to plasmid pK18ΔglnA2 which was then used for the chromosomal deletion of glnA2 via two independent recombination events as described . For glnE deletion a 4.5-kb fragment containing the glnE gene and its flanking regions was amplified by PCR (primers used: 5′-TGCCCACTGACAACGGCGGAT-3′, 5′-TCCTCTGCACATGCACGAGCT-3′), restricted with EcoRI and BamHI and the resulting 3.5-kb fragment was ligated to EcoRI/BamHI-restricted and dephosphorylated pUC18 DNA leading to plasmid pUCglnE. Via PCR using the primers 5′-TGGGTCTCTGACAAAGCCAACGA-3′ and 5′-ATGCGGTCGAGGCTTGCGA-3′ a 2.3-kb deletion was introduced in glnE, after religation leading to plasmid pUCΔglnE, the insert, containing the deletion was cut out by EcoRI/BamHI restriction and ligated to EcoRI/BamHI-restricted and dephosphorylated pK18mobsacB leading to plasmid pK18ΔglnE. With this plasmid the chromosomal deletion of glnE was constructed as described .
2.8Construction of the glnA–lacZ fusion strain
The gene replacement method developed by Schäfer et al.  was also applied for the chromosomal insertion of the lacZ gene into the glnA gene. In the first step the lacZ gene was amplified by PCR from plasmid pLacZi (MATCHMAKER OneHybrid System, Clontech, Heidelberg, Germany). By the primers 5′-TCTAGAATGACCGGATCCGGAGCTTG-3′ and 5′-CTCGAGTTACGCGAAATACGGGCAGAC-3′ a XbaI site was added to the 5′ end and a XhoI site to the 3′ end of lacZ (shown in bold letters). By PCR a 1-kb fragment containing 500 bp up- and downstream from the glnA start codon was amplified using the primers 5′-GAATTCACGTTCCTCTGCATCCACACG-3′ and 5′-GCATGCACGCTCAAGAGCGAAGCC-3′ and cloned into pUC18 leading to plasmid pUCglnA. By PCR-based site-directed mutagenesis a XbaI and a XhoI site (shown in bold letters) were introduced 3 bp downstream of the glnA start codon using the primers 5′-TCTAGAGGTTTCAAACGCCACGGT-3′ and 5′-CTCGAGCCGGAAGAAATTGTCAAGTTC-3′. After religation which led to plasmid pUCglnAXbaIXhoI the plasmid was restricted with XbaI/XhoI, dephosphorylated and ligated to the XbaI/XhoI-restricted lacZ gene leading to plasmid pUCglnA-lacZ. From this plasmid, the glnA–lacZ fusion was cut out with EcoRI/SphI and ligated into EcoRI/SphI-restricted and dephosphorylated pK18mobsacB DNA. This plasmid was used to introduce the glnA–lacZ fusion into the genome of C. glutamicum via two independent recombinations according to the method described .
3.1Regulation of glnA transcription
The glnA gene encoding glutamine synthetase I in C. glutamicum was isolated by a PCR-based approach using degenerated primers . Up- and downstream of the glnA gene several ORFs of unknown function were found. By RT-PCR we found that none of these ORFs are cotranscribed with glnA (data not shown). 109 bp upstream from the GTG start codon of glnA the transcription start site was localized by the 5′ RACE method. This start site was independent of the nitrogen supply of the cells, indicating that only one promoter is present (data not shown).
In a previous study we showed that a global regulator of nitrogen control exists in C. glutamicum, which was designated AmtR  and a highly conserved binding motif ATCTATAGN1–4ATAG, was identified upstream of AmtR-controlled genes. We were now interested to find out, if AmtR might also control transcription of the glnA gene. In fact, we found a truncated form of the described consensus motif, the 6-bp sequence ATCTAT, upstream of glnA. To prove that AmtR binds to this motif, gel retardation experiments were carried out. A 250-bp fragment containing the putative binding sequence was amplified by PCR and incubated in a binding assay with raw extracts from E. coli cells in which the AmtR-encoding amtR gene was expressed. The same DNA fragment was incubated with raw extracts from E. coli cells containing only the vector pUC18 for control. After native polyacrylamide electrophoresis a distinct retardation of the fragment which had been incubated with the AmtR-containing cell extract was observed (Fig. 1). In contrast, the same DNA fragment, in which the motif ATCTAT had been deleted did not show a gel retardation (data not shown) proving that AmtR binds to the sequence ATCTAT. To further analyze whether glnA transcription was in fact regulated by AmtR, we isolated RNA from wild-type cells and from amtR deletion strain MJ6-18 grown under nitrogen excess and after different times of nitrogen starvation, respectively, and hybridized it with an antisense glnA probe. For control, a probe complementary to the 16S rRNA gene was used. In wild-type cells an upregulation of glnA transcription by a factor of at least two under nitrogen starvation compared with nitrogen excess was observed whereas amtR deletion strain MJ6-18 showed a constitutive glnA transcription (Fig. 2A,B). These results were verified by a real-time RT-PCR approach using a Lightcycler (Roche, Germany). As calculated from the data of this assay, the glnA transcript is three-fold increased in response to 20 min of nitrogen starvation (data not shown). As an independent proof for upregulation upon nitrogen starvation we also analyzed glnA transcription under nitrogen excess and nitrogen starvation by determination of the β-galactosidase activity using a glnA–lacZ fusion strain. This strain, ATCCglnA-lacZ, showed a constitutive β-galactosidase activity which was further increased after only 10 min of incubation in nitrogen-free medium by a factor of approximately 1.2 (Fig. 2C). These data are in accord with the results obtained by the RNA hybridizations experiments showing a considerable constitutive expression which is enhanced upon nitrogen starvation.
Recently, a GlnK-type protein and an uridylyltransferase were identified in C. glutamicum. These components are discussed to be involved in nitrogen sensing and signal transfer and are obviously essential for derepression of AmtR-controlled genes under nitrogen starvation. When measuring glnA transcription in a glnK or a glnD deletion strain under nitrogen excess and in response to nitrogen starvation, respectively, an increase in glnA transcription as in the wild-type was not observed anymore (Fig. 2A). This indicates that GlnK and uridylyltransferase, i.e. an intact nitrogen regulation cascade, are essential for derepression of glnA transcription in C. glutamicum.
3.2Isolation of the glnA2–glnE cluster
In most bacteria GSI is regulated on the level of activity via covalent modification by an adenyl residue. This modification is catalyzed by the adenylyltransferase, encoded by the glnE gene. As we know from previous studies [8,9] that glutamine synthetase of C. glutamicum is also regulated on the level of activity we analyzed if a glnE gene exists in C. glutamicum. Based on homology searches using C. glutamicum genome data a gene with high similarity to the putative M. tuberculosis glnE gene was identified. To verify that this gene encodes the adenylyltransferase which is responsible for GSI activity regulation, we constructed a glnE deletion mutant and analyzed the influence of the deletion on GSI activity (see below).
Southern blot experiments carried out during characterization of the glnA gene indicated that, as in different other bacterial species, also in C. glutamicum a second form of glutamine synthetase might exist . In fact, an open reading frame which showed high similarity to glnII gene of M. tuberculosis (65% identity of the deduced amino acid sequence) was found directly upstream of the putative glnE gene. This gene, designated glnA2, encodes a protein of 427 amino acids. Interestingly, the similarity of glnA2 to the glnA gene of C. glutamicum is only weak (34% identity) and a putative site of adenylylation is not to present. For further characterization we investigated the glnA2 regulation pattern and constructed and analyzed the resulting phenotype a deletion mutant (see below).
3.3Regulation of glnA2–glnE transcription
The tight succession of glnA2 and glnE in the C. glutamicum genome suggested a possible operon structure (Fig. 3A). By using RT-PCR with primers located in glnA2 and glnE we demonstrated that these genes are in fact cotranscribed (Fig. 3B). Based on the finding that the majority of genes in nitrogen metabolism of C. glutamicum are transcriptionally regulated, we also analyzed a putative regulation of glnA2–glnE transcription. As RNA hybridizations were not successful, we carried out RT-PCR experiments using RNA isolated from wild-type cells grown under nitrogen excess and incubated in nitrogen-free medium, respectively, using primers located within glnA2 and glnE. A constitutive glnA2 and glnE transcription independent of the nitrogen supply was shown (Fig. 3C). Additionally, when doing the same experiment with RNA isolated from amtR deletion mutant MJ6-18, no effect of the deletion was detected, ruling out the possibility of a regulation of glnA2 and glnE by AmtR (data not shown).
3.4Deletion analysis of glnA2 and glnE
To investigate the function of the putative GSII protein and adenylyltransferase deletion mutants of glnII and glnE were constructed. The corresponding strains, LNΔglnA2 and LNΔglnE, were analyzed in respect to growth and glutamine synthetase I activity. The glnA2 deletion mutant LNΔglnA2 was glutamine-prototrophic and showed no growth defects under nitrogen limitation, also no effect on glutamine synthetase activity was found in this strain. Since deletion of glnA2 gave no hints for the function of the corresponding protein, the effect of overexpression of glnA2 in the glnA deletion mutant MJ4-26 was analyzed. For this purpose the glnA2 gene and, for control, the glnA gene were cloned separately into expression vector pZ8-1 under control of a powerful tac promoter. The resulting plasmids pZ8-1glnA and pZ8-1glnA2 were transformed separately into glnA deletion mutant MJ4-26 and clones were tested for glutamine auxotrophy on minimal medium agar containing ammonium as nitrogen source with or without addition of glutamine. Whereas strain MJ4-26 pZ8-1glnA grew on both media, strain MJ4-26 pZ8-1glnA2 grew only on the glutamine-supplemented medium. In summary, the physiological function of glnA2 remains unclear.
When measuring the GS activity in cells of glnE deletion strain LNΔglnE grown under nitrogen excess and after incubation in nitrogen-free medium, respectively, no up- and downregulation in GS activity in contrast to the wild-type was visible. Also, when an ammonium pulse was given to nitrogen-starved LNΔglnE cells GS activity did not decrease, in contrast to the response of the wild-type (Fig. 4), indicating that the glnE-encoded protein is responsible for modifying GSI activity in C. glutamicum. To verify this interpretation, we tested the influence of snake venom phosphodiesterase (SVPDE), which removes adenyl residues from the GSI enzyme, on glutamine synthetase activity. When adding SVPDE to cell extract from wild-type C. glutamicum cells grown under nitrogen excess an increase in GS activity was observed whereas no effect was detectable in the case of ΔglnE mutant LNΔglnE (data not shown). Interestingly, strain LNΔglnE exhibits a higher GS activity as the wild-type when grown in nitrogen-rich medium. This indicates that GS is partially inactivated by adenylylation in the wild-type. The increase of GS activity upon nitrogen starvation in strain LNΔglnE cannot be the result of enzyme activation as in the wild-type, but is due to upregulation of glnA expression.
The regulation of glutamine synthetases in C. glutamicum was investigated on the level of gene expression and activity. In most bacteria, transcription of the glutamine synthetase I-encoding glnA gene is regulated depending on the nitrogen supply of the cell. In contrast to Gram-negative bacteria where activation mechanisms dominate, Gram-positive bacteria e.g. B. subtilis, mostly use repression mechanisms to regulate glnA transcription . In this study, we were able to show that also in C. glutamicum glnA transcription is regulated via a repression mechanism. Interestingly, this repressor, designated AmtR, does not only control glnA transcription but also regulates most other nitrogen-controlled genes in C. glutamicum, namely amt, amtB, glnK, glnD, gltB and gltD (; G. Beckers, personal communication). By gel-shift experiments, we verified that AmtR recognizes and binds to the 6-bp sequence ATCTAT in the glnA promoter. This is a truncated form of the motif found in the promoters of other AmtR-controlled genes, e.g. amt with the recognition sequence ATCTATAGN4ATAG . Due to the deletion in respect to the previously described AmtR-binding motif the repression of glnA transcription by AmtR is not very strict. RNA hybridizations, real-time PCR assays and reporter gene fusions revealed a considerable background expression of glnA which was further increased in response to nitrogen starvation. This is in strong contrast to other AmtR-controlled genes as for example amt and amtB where almost no expression is detectable under full nitrogen supply. The weak repression of glnA transcription might underline the pivotal role of GSI for anabolism even in nitrogen-rich medium. The reason for this is the high demand of glutamine for cell wall synthesis and growth in C. glutamicum. Additionally, we could show that the GlnK protein and uridylyltransferase are essential for derepression of glnA transcription as in the case of other AmtR-controlled genes (unpublished data).
A typical feature of GSI proteins in bacteria is the posttranslational adenylylation of a tyrosine residue via adenylyltransferase depending on the nitrogen status of the cell. In this study we isolated the glnE gene encoding adenylyltransferase in C. glutamicum. The absence of up- and downregulation of GSI activity in the glnE deletion strain LNΔglnE proved that the glnE gene product is responsible for GSI activity regulation. Upstream of glnE, a gene with weak identity (34% of the deduced amino acid sequence) to the glnA gene of C. glutamicum was detected. The glnA2 gene does not encode a GS II-type protein as described for streptomycetes (e.g. ) but a GSI enzyme of the α subgroup of glutamine synthetases . Under the experimental conditions tested, glnA2 had no physiological function as shown by the lack of glutamine synthetase activity in the glnA deletion mutant MJ4-26 , the lack of phenotype caused by a glnA2 deletion and by the inability of glnA2 to complement the glutamine auxotrophy of this mutant. In this respect, however, it is an interesting fact that the glnA2 gene forms an operon with the glnE gene and is constitutively expressed under various nitrogen conditions. This shows that glnA2 is not a silent gene in C. glutamicum.
The glnA2 and glnE nucleotide sequences were kindly provided by Degussa and were deposited in the EMBL Nucleotide Sequence Database under the accession number AJ310086. This work was supported by the Fonds der chemischen Industrie and the Deutsche Forschungsgemeinschaft (Grant BU894/1-1).