The ATP-dependent protease Clp plays important roles in the cell's protein quality control system and in the regulation of cellular processes. In Corynebacterium glutamicum, the levels of the proteolytic subunits ClpP1 and ClpP2 as well as of the corresponding mRNAs were drastically increased upon deletion of the clpC gene, coding for a Clp ATPase subunit. We identified a regulatory protein, designated ClgR, binding to a common palindromic sequence motif in front of clpP1P2 as well as of clpC. Deletion of clgR in the ΔclpC background completely abolished the increased transcription of both operons, indicating that ClgR activates transcription of these genes. ClgR activity itself is probably controlled via ClpC-dependent regulation of its stability, as ClgR is only present in ΔclpC and not in wild-type cells, whereas the levels of clgR mRNA are comparable in both strains. clpC, clpP1P2 and clgR expression is induced upon severe heat stress, however, independently of ClgR. Identification of the heat-responsive transcriptional start sites in front of these genes revealed the presence of sequence motifs typical for σECF-dependent promoters. The ECF sigma factor σH could be identified as being required for transcriptional activation of clpC, clpP1P2 and clgR in response to severe heat stress. A second heat-responsive but σH-independent promoter in front of clgR could be identified that is subject to negative regulation by the transcriptional repressor HspR. Taken together, these results show that clpC and clpP1P2 expression in C. glutamicum is subject to complex regulation via both independent and hierarchically organized pathways, allowing for the integration of multiple environmental stimuli. Both the ClgR- and σH-dependent regulation of clpC and clpP1P2 expression appears to be conserved in other actinomycetes.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Proteolysis in bacterial cells is mainly performed by ATP-dependent proteases. Together with several chaperones, i.e. DnaKJ-GrpE and ClpB, these proteases are integral parts of the cell's protein quality control system, which is responsible for clearing the cell of non-functional proteins (for recent reviews, see Wickner et al., 1999; Dougan et al., 2002). Most of these proteases also perform important regulatory functions by controlling the availability of transcriptional regulators, enzymes and other proteins via conditional degradation (for a recent review, see Jenal and Hengge-Aronis, 2003). Of these proteases, Clp has been most extensively studied both mechanistically and functionally. The Clp holoenzyme consists of two separate and functionally distinct subunits. The proteolytic subunits, ClpP, perform the actual hydrolysis of substrates. However, their active sites are buried within the cavity of the so-called proteolytic core formed by 14 ClpP subunits. Therefore, hexamers of ATPase subunits (ClpA, ClpC or ClpX), which are members of the Clp/Hsp100 superfamily (Schirmer et al., 1996) and associate with the core, are required in order to recognize, unfold and ultimately translocate substrate proteins into it (for a recent review, see Hlavacek and Vachova, 2002).
Obviously, the cellular concentration of Clp has to be adapted in response to environmental conditions resulting in the accumulation of non-functional, i.e. truncated, misfolded or aggregated, proteins. The underlying regulatory phenomena are well studied in enterobacteria and low G+C Gram-positive bacteria. In Escherichia coli, RNA polymerase (RNAP) containing the alternative sigma factor σ32 is responsible for heat-triggered transcriptional activation of clpX, clpP and other stress genes (for a recent review, see Arsene et al., 2000), whereas the transcriptional repressor CtsR controls expression of the clpC, clpE and clpP genes in Bacillus subtilis (Krüger and Hecker, 1998; Derre et al., 1999). The activity of σ32 as well as of CtsR is at least partially controlled at the level of protein stability. σ32 is normally directed by the DnaKJ-GrpE chaperone machinery into a degradation pathway involving several ATP-dependent proteases. However, σ32 stability increases rapidly during heat stress because the amounts of DnaKJ-GrpE available for σ32 sequestration become limited as the chaperone machinery is now required for disaggregating and refolding non-native proteins. In contrast, CtsR stability is apparently modulated via heat-triggered post-translational modification by a putative protein kinase, McsB, and subsequent ClpCP-dependent degradation (Krüger et al., 2001).
Corynebacterium glutamicum belongs to the bacterial order Actinomycetales, comprising such important genera such as Mycobacterium and Streptomyces. The C. glutamicum genome (Ikeda and Nakagawa, 2003; Kalinowski et al., 2003) contains structural genes coding for at least two regulatory Clp protease subunits, ClpC and ClpX, and two proteolytic Clp protease subunits, ClpP1 and ClpP2. The mRNA levels of the clpC, clpP1 and clpP2 genes have been shown to be increased significantly upon heat shock (Muffler et al., 2002). However, no CtsR or σ32 homologues are present in C. glutamicum, suggesting that other mechanisms than those described above are underlying this regulation. The group of Philippe Mazodier was able to show for Streptomyces lividans that expression of the clpP3P4 operon, which is unique for Streptomyces species and probably arose via duplication of the clpP1P2 operon common to all actinomycetes, is positively controlled by the transcriptional activator PopR (Viala et al., 2000). PopR activity is apparently regulated via conditional degradation by an isoform of the Clp protease, although the stimuli resulting in PopR stabilization remain elusive (Viala and Mazodier, 2002). However, the regulatory pathways controlling expression of the clpC and clpP1P2 operons common to all actinomycetes remain undisclosed. Here, we show that, in C. glutamicum, expression of these genes is under dual control by a hitherto uncharacterized transcriptional activator, designated ClgR, and the ECF sigma factor σH. Furthermore, we provide evidence that ClgR activity is controlled at the level of ClgR stability as well as by transcriptional control exerted by σH and the negative regulator of chaperone gene expression, HspR (Bucca et al., 1997; Grandvalet et al., 1997; 1999). Finally, we present data indicating that the described regulatory mechanisms are at least partially conserved in actinomycetes.
Construction and characterization of a C. glutamicum ΔclpC mutant
Corynebacterium glutamicum harbours two genes, clpP1 and clpP2, that code for proteolytic subunits of the Clp protease and are organized in a bicistronic operon. Additionally, at least two genes coding for ATPase subunits, clpC and clpX, can be found in the genome sequence (GenBank accession NC_003450). In order to characterize functionally the C. glutamicum Clp protease, we have constructed strains with in frame deletions in the clpX and clpC genes. Phenotypically, both mutants are characterized by altered cell morphologies and reduced growth rates in both minimal and rich media (data not shown). Deletion of the genes coding for the proteolytic subunits proved to be impossible, indicating an essential function of the respective proteins (data not shown). In order to identify possible Clp protease substrates that might accumulate in the clpX and clpC mutants as a result of the absence of functional ClpXP and ClpCP proteases, we compared the cytosolic protein patterns of these mutants with that of the wild type. Among the most drastic changes observed was the accumulation of two proteins with apparent molecular weights of 22.7 and 23.5 kDa in the ΔclpC deletion strain (Fig. 1A). When constructing a proteomic map of C. glutamicum (Schaffer et al., 2001), these proteins had been shown to represent the proteolytic subunits of the Clp protease, ClpP1 and ClpP2. Both proteins were approximately 10-fold more abundant in the mutant than in the wild type (Fig. 1A) and the ΔclpX deletion strain (data not shown). In contrast, the cellular concentration of other stress proteins, e.g. DnaK, GroEL, GroES and ClpB, remained essentially unchanged upon clpC deletion (data not shown). Therefore, the observed increase in ClpP1 and ClpP2 abundance is not due to the generally increased synthesis of stress-related proteins, which could have been caused by the accumulation of non-native proteins in the mutant strain lacking ClpC.
In order to test whether the increase in protein abundance is reflected by increased clpP1P2 mRNA levels in the ΔclpC mutant, we performed transcriptome analyses using whole-genome DNA microarrays containing polymerase chain reaction (PCR) probes representing 3541 C. glutamicum open reading frames (ORFs). Comparison of the transcriptional profiles of the C. glutamicum wild type and the ΔclpC mutant revealed that the levels of clpP1 and clpP2 mRNA are increased sixfold in the mutant strain, whereas those of the clpB, dnaK and groES genes are raised only slightly (Table 1). In the case of clpP1 and clpP2, these results were independently confirmed by primer extension experiments (Fig. 2A, lanes 1 and 2). Therefore, deletion of clpC obviously results in activation of clpP1P2 transcription or, alternatively, in stabilizing clpP1P2 mRNA.
Table 1. Results of DNA microarray experiments.
Average ratios of mRNA levels with standard deviation
ΔclpC vs. WT
ΔclgRΔclpC vs. ΔclpC
ΔclgR vs. WT
WT 50°C vs. 30°C
ΔclgR 50°C vs. 30°C
ΔsigH 50°C vs. 30°C
. Signal-to-noise ratios in one or more experiments ≥thinsp;2 and ≤ 3.
. NA, not applicable.
. ND, not determined, because signal-to-noise ratios in all experiments ≤ 2.
Identification of a protein binding to the clpP1P2 promoter region
If an increase in clpP1P2 transcription upon clpC deletion is responsible for the observations made, we had to postulate the presence of a transcriptional regulator binding to the clpP1P2 regulatory region in either the ΔclpC mutant or the wild-type strain, depending on the putative regulator acting as a transcriptional activator or repressor. In order to identify such a protein, we used the intergenic region between clpP1 and the tig gene located upstream (−176 to +135 bp relative to the clpP1P2 transcriptional start site as determined by primer extension; see Fig. 2A) coupled to Dynabeads® streptavidin for DNA affinity purification experiments. DNA affinity purification was performed with crude extracts from wild-type and ΔclpC cells and led to the isolation of a protein with an apparent mass of 10.7 kDa from crude extracts of the mutant but not from the wild type (Fig. 3A). Unambiguous protein identification was performed by peptide mass fingerprinting and post-source decay analysis using matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry (data not shown). The enriched protein is identical to NCBI non-redundant database entry BAB99355 from C. glutamicum. BAB99355 is predicted to be 107 residues long corresponding to a molecular mass of 11.3 kDa, which is in good agreement with the experimentally determined value. The NCBI annotation predicts the protein to be a transcriptional regulator, and the search for conserved domains (SMART; Schultz et al., 1998; Letunic et al., 2002) revealed the presence of a central DNA-binding domain (helix–turn–helix motif of XRE family-like proteins; SMART accession number SM0530) in the centre of the protein. The protein, designated ClgR (for Clp gene regulator), could only be enriched from ΔclpC cells, which correlates with high clpP1P2 expression and indicates that ClgR acts as transcriptional activator.
Regulation of clpP1P2 and clpC expression by ClgR
In order to analyse whether and how ClgR regulates expression of the clpP1P2 operon, we constructed a clgR in frame deletion mutant and compared the transcriptional profiles of the mutant with that of the wild type. As is shown in Table 1, the mRNA levels of both clpP1 and clpP2 in the ΔclgR strain were reduced to half of wild-type levels. These results could be independently verified by primer extension analysis (Fig. 2A, lanes 1 and 4), supporting the assumption that ClgR is a positive regulator of clpP1P2 expression. The abundance of the corresponding gene products was also reduced in the ΔclgR mutant (Fig. 1A). Interestingly, the mRNA level of the clpC gene was also decreased in the mutant (Table 1), suggesting that this gene might also be part of the ClgR regulon. As we could not analyse clpC expression in the ΔclpC strain using DNA microarrays, we compared clpC mRNA levels in wild-type and ΔclpC cells using primer extension. In fact, the clpC-derived primer extension signal was much stronger in the case of the clpC mutant, confirming that clpC might be part of the ClgR regulon (Fig. 2B, lanes 1 and 2). The increase in clpC promoter activity in the clpC mutant was also reflected at the protein level. The inactive ClpC derivative encoded by pJC1-clpCΔ533-931 was shown to be much more abundant in the ΔclpC background than in the wild-type strain (Fig. 1B). In contrast to clpC mRNA (Fig. 2B, lanes 1 and 4), the abundance of ClpCΔ533-931 was somewhat higher in the ΔclgR mutant than in the wild type (Fig. 1B; see below).
The effects on clpP1P2 and clpC expression of a clgR deletion in the wild type were rather weak, indicating that levels of active ClgR in the wild type are already very low, as had also been suggested by the failure to isolate any ClgR via DNA affinity purification from wild-type cells. As the abundance of active ClgR is much higher in the absence of the ClpC ATPase, deletion of clgR in the ΔclpC background should result in more drastic effects if ClgR is to be an activator of clpP1P2 and clpC expression. In fact, comparison of the transcriptional profiles of the strains ΔclpC and ΔclgRΔclpC (Table 1) as well as primer extension analysis of clpC and clpP1P2 expression in these strains (Fig. 2A and B, lanes 2 and 3 respectively) showed that the drastic increase in mRNA levels caused by clpC deletion is completely abolished upon introduction of the clgR mutation. Again, these observations are reflected at the protein level (Fig. 1A and B), although ClpCΔ533-931 levels in the double mutant are somewhat higher than in the wild type. This was also true for the ΔclgR single mutant (see above) but, in both strains, the primer extension signals specific for clpC transcription from plasmid pJC1-clpCΔ533-931 were of lower intensity than in the wild type (data not shown). We currently have no explanation for these discrepancies between relative mRNA and protein levels. These data clearly show that ClgR is involved in the activation of clpC and clpP1P2 gene expression in C. glutamicum in the artificial situation of ClpC absence and, to a lesser extent, in wild-type cells. Moreover, ClgR and ClpC appear to act within the same regulatory pathway.
Interaction of ClgR with the clpC regulatory region
In order to test whether increased clpC expression in the absence of functional ClpC is based on interaction of ClgR with the clpC regulatory region, we repeated the affinity purification approach described above with the clpC regulatory region (−222 to +51 bp relative to the clpC transcriptional start site as determined by primer extension; see Fig. 2B) coupled to Dynabeads® streptavidin. As with the clpP1P2 promoter, an 11 kDa protein could be enriched from ΔclpC but not from wild-type cells (Fig. 3B) and was confirmed to represent ClgR using mass spectrometry (data not shown).
Identification of ClgR binding sites upstream of clpC and clpP1P2
In order to determine the ClgR binding sites in front of the clpC and clpP1P2 operons, DNase I footprinting assays were performed. To that end, ClgR with a carboxy-terminal fusion to a StrepTag II (ClgR-C) was overproduced and purified from Escherichia coli BL21 (DE3), carrying pET24b-clgR-C. As we reproducibly co-purified DnaK together with ClgR-C, even in the presence of 1 M NaCl and 0.1% (v/v) NP-40 in the chromatography buffers (data not shown), we decided to overproduce ClgR-C in the E. coliΔdnaK strain BB1553 (Tomoyasu et al., 2001). This required the construction of plasmid pEKEx1-clgR-C. In order to test for functionality of ClgR-C, pEKEx1-clgR-C and the control vector pEKEx1 were introduced into C. glutamicumΔclgRΔclpC. The levels of ClpP1 and ClpP2 in the strain carrying pEKEx1-clgR-C were much higher than in the control strain even in the absence of IPTG, indicating that ClgR-C can functionally replace wild-type ClgR (data not shown).
As is shown in Fig. 4, ClgR-C protects sections of both regulatory regions from DNase I digestion. The protected region upstream of clpP1P2 stretches from −36 to −59 relative to the clpP1P2 transcriptional start site (Fig. 4A and B). In the case of clpC, protection is observed from position −120 to −140 relative to the distal clpC transcriptional start site (Fig. 4C and D). Protection of the non-template strand of the ClgR operator in front of clpP1P2 was reproducibly observed to be much weaker than that of the template strand. We do not know the reason for that as experiments were performed with the same protein preparation as in the template strand reaction and the probe quality was comparable to that of the template strand probe. Possibly, the non-template strand of ClgR operators is generally protected to a lesser degree, and this is especially accentuated in the case of the operator in front of clpP1P2 as affinity of ClgR to this operator is low compared with the operator in front of clpC (first signs of template strand protection at ClgR concentrations of 440 nM for clpP1P2 and at 44 nM for clpC; Fig. 4B and D). The localization of ClgR binding sites relative to the respective transcriptional start sites and putative promoter sequences is shown in Fig. 5. Comparison of the protected regions revealed the presence of a DNA sequence motif common to both ClgR operators. The motif, 5′-RWWCGCT-N3-RGCGWAC-3′, has an imperfectly palindromic structure, indicating that ClgR binds to its operator sites as a dimer. We have performed genome-wide searches for additional occurrences of this sequence pattern, but found none except when allowing for mismatches. Even then, most matches mapped to coding regions, and those localized in intergenic regions were not in proximity to genes related to proteolysis or stress response in general or genes with altered mRNA level in the ΔclpC/ΔclgRΔclpC microarray comparison (data not shown). Thus, we have no indications that additional members of the ClgR regulon exist in C. glutamicum.
ClgR-independent activation of clpC and clpP1P2 expression upon severe heat stress
Once the role of ClgR as a transcriptional activator of clpC and clpP1P2 expression had been established, we asked in response to which environmental stimuli activation occurs. Expression of clp genes in many bacteria is activated upon several stresses, most markedly in response to heat stress, but also upon oxidative, ethanol, cold, salt and other stresses. In the case of C. glutamicum, an increase in clpC and clpP1P2 gene expression had been shown in response to severe heat stress (temperature shift from 30°C to 50°C; Muffler et al., 2002). Therefore, we asked whether heat-triggered activation of clpC and clpP1P2 expression depends on ClgR. In order to answer that question, we compared the transcriptome of the wild-type strain when growing at ambient temperature (30°C) with that 10 min after shifting the cells to 50°C. In parallel, a similar experiment was carried out with the ΔclgR mutant. Unexpectedly, the mRNA levels of clpC, clpP1 and clpP2 increased significantly upon heat shock not only in the wild type but also in the ΔclgR mutant (Table 1). In fact, the heat-triggered increase in the respective mRNA levels was approximately twofold higher in the mutant strain than in the wild type (Table 1). Thus, ClgR is apparently not involved in the heat-triggered activation of clpC and clpP1P2 expression. We also tested whether clpC and clpP1P2 expression is increased in response to other stress situations but were unable to detect such an induction in response to salt (1 M NaCl) and oxidative stress (20 mM hydrogen peroxide) as well as treatment with ethanol (2% v/v) and puromycin (100 µg ml−1) (data not shown). Moreover, growth behaviour of the ClgR mutant upon exposure to these stresses was identical to that of the wild type (data not shown). Therefore, we are currently unable to define the physiological role of ClgR.
σH dependence of clpC and clpP1P2 expression in response to severe heat stress
In order to identify the heat-responsive transcriptional start sites in front of clpC and clpP1P2, we repeated the primer extension experiments with total RNA isolated from the wild-type strain cultivated at 30°C or 50°C respectively (Fig. 2C and D). In the case of clpP1P2, the increase in mRNA level is caused by transcripts originating at a transcriptional start site 12 bp downstream of the ClgR-dependent start site (Fig. 2C, lanes 3 and 4, and Fig. 5). In contrast, the heat-responsive transcriptional start site in front of clpC is identical to the distal ClgR-dependent start site (Fig. 2D, lanes 3 and 4, and Fig. 5). However, under conditions of severe heat stress, use of this start site does not require ClgR, as the very same primer extension signal can be observed when the ΔclgR mutant is analysed (data not shown). This indicates that the corresponding promoter can be used by RNAP holoenzymes containing different sigma factors, one active at ambient temperatures and requiring ClgR, and the other activated by severe heat stress and acting independently of ClgR.
Upstream of the heat-responsive transcriptional start sites in front of clpC and clpP1P2, putative −10 and −35 boxes are present that are very similar to those recognized by extracytoplasmatic function (ECF) sigma factors (Fig. 5). These proteins form a clearly defined subgroup of the σ70 sigma factors and owe their name to the fact that they were originally described as being responsible for transcription of genes coding for proteins with extracellular function (for reviews, see Missiakas and Raina, 1998; Helmann, 2002). The C. glutamicum genome harbours five genes coding for ECF sigma factors, namely sigC, sigD, sigE, sigH and sigM. Of the corresponding Mycobacterium tuberculosis homologues two, σH and σE, have been found to play a role in the response to heat stress, as expression of the respective structural genes is induced upon heat shock (Manganelli et al., 1999), and sigE and sigH mutants are severely attenuated with respect to survival of heat shock and other environmental stresses (Wu et al., 1997; Fernandes et al., 1999; Manganelli et al., 2001; 2002; Raman et al., 2001). σH- and σE-dependent promoters have been characterized in this organism (Manganelli et al., 2001; 2002; Raman et al., 2001), and they are very similar to those in front of clpC and clpP1P2. Therefore, we constructed strains with in frame deletions in sigE and sigH, respectively, and compared these strains with the wild type with respect to heat-triggered activation of clpC and clpP1P2 expression. In the ΔsigE mutant, the increase in mRNA levels of these genes upon severe heat stress was comparable to that in the wild-type strain (data not shown). However, deletion of sigH led to complete abolishment of heat-triggered activation of clpC and clpP1P2 transcription, as shown by DNA microarray experiments (Table 1) and primer extension analysis (Fig. 2C and D, lanes 1 and 2 respectively). Hence, σH-RNAP apparently mediates the observed effect.
Transcriptional regulation of clgR expression
The DNA microarray experiments had shown that severe heat shock also leads to a fivefold increase in clgR mRNA level. In order to confirm these results and to determine the cause of this induction, we performed primer extension experiments with RNA isolated from C. glutamicum wild-type cells cultivated at 30°C and subjected to severe heat shock for 10 min respectively. In fact, this results in higher clgR transcript levels, which is caused by the increased use of two transcriptional start sites separated by 30 bp (Fig. 2E, lanes 1 and 2, and Fig. 5). Mapping of these transcriptional starts on to the DNA sequence revealed that, upstream of clgR, putative −10 and −35 regions very similar to those of the σH-dependent promoters in front of clpC and clpP1P2 are present in appropriate spacing to the proximal start site (Fig. 5). In accordance with that, this transcriptional start is no longer used in the ΔsigH mutant under conditions of severe heat shock (Fig. 2E, lanes 3 and 4). We postulate that, in addition to clpC and clpP1P2, σH-RNAP also transcribes clgR from its proximal promoter.
Heat responsiveness of the distal clgR transcriptional start site is not affected in the ΔsigH mutant (Fig. 2E, lanes 3 and 4), and this start site is not preceded by sequence motifs with obvious similarities to promoter sequences recognized by ECF sigma factors. At positions −6 to −11, a TAAACT motif similar to the consensus of −10 regions of C. glutamicum vegetative promoters (TAnaaT; Patek et al., 1996) can be found (Fig. 5). No obvious −35 region is recognizable, although conservation of this region is generally very low in C. glutamicum (Patek et al., 1996). However, at positions −19 to −39, a palindromic sequence (5′-TTGAGTC-N5-GACTCAA-3′) with high similarity to the consensus operator of the transcriptional repressor HspR (5′-CTTGAGT-N7-ACTCAAG-3′; Grandvalet et al., 1999) is present (Fig. 5). HspR has been shown to be a negative regulator of expression of the dnaK–grpE–dnaJ–hspR regulon as well as of the clpB and lon genes in Streptomyces spp. (Bucca et al., 1997; Grandvalet et al., 1997; 1999; Sobczyk et al., 2002). HspR and corresponding operator sequences can be found in all actinomycetes (Grandvalet et al., 1999). In C. glutamicum, HspR operators are present in front of clpB and the dnaK–grpE–dnaJ–hspR operon (Grandvalet et al., 1999; S. Schaffer, unpublished results). In order to prove a possible role for HspR in the regulation of clgR expression, we constructed an hspR in frame deletion mutant and performed primer extension analyses for clgR with total RNA prepared from the ΔhspR mutant. In contrast to the results obtained with the C. glutamicum wild type and the ΔsigH mutant, the distal transcriptional start in front of clgR is used in ΔhspR cells cultivated at 30°C and 50°C (Fig. 2E, lanes 5 and 6), suggesting derepression of clgR transcription from its distal promoter under normal culture conditions. However, this is not translated into ClgR-dependent activation of clpC or clpP1P2 transcription (data not shown), probably because derepression of clgR transcription is counteracted by immediate ClpC-dependent inactivation of newly synthesized ClgR.
Post-transcriptional regulation of ClgR activity
Although in response to severe heat stress clgR expression is obviously regulated at the transcriptional level, the available experimental data suggested that the observed increase in ClgR activity upon clpC deletion resulted from post-transcriptional regulation. Active ClgR (with respect to its DNA-binding activity) could only be isolated from ΔclpC but not from wild-type cells (Fig. 3). However, the DNA microarray data showed that the level of clgR mRNA is only 1.5-fold higher in the ΔclpC background (Table 1). Quantitative reverse transcription polymerase chain reaction (RT-PCR) analysis of clgR expression confirmed this observation (Fig. 6A). Hence, ClgR activity has to be regulated mainly at the level of translation or post-translationally, i.e. by modulating the stability or DNA-binding activity of ClgR. Attempts to affinity purify ClgR-C from strains ΔclgR and ΔclgRΔclpC, both harbouring the expression vector pEKEx1-clgR-C, failed in the case of the ΔclgR strain even in the presence of IPTG, whereas clgR-C overexpression in the double mutant resulted in massive accumulation of the protein (Fig. 6B). This strongly argues against a translational control mechanism, as the clgR translation initiation signals on pEKEx1-clgR-C are derived from plasmid pET24b (for details, see Experimental procedures). Rather, ClgR activity is most probably controlled by regulating its stability in a ClpC-dependent manner. If so, degradation is not inhibited by the C-terminal StrepTag II in ClgR-C. As the addition of 10 mostly charged and bulky amino acid residues would possibly mask signals required for efficient recognition and/or degradation of ClgR by a ClpC-containing protease, the respective sequence determinants might not be localized C-terminally.
Conservation of the regulatory network controlling clpC and clpP1P2 expression in C. glutamicum in other actinomycetes
As the genome sequences of several other actinomycetes species besides C. glutamicum are available, we analysed those with respect to the possible conservation of the regulatory pathways described above. With the exception of the fully sequenced M. leprae genome, we were able to detect clgR-like genes in all actinomycetes genomes (E-values ≤ 4 × 10−14) at a conserved genomic location downstream of the ftsK and pgsA3 genes (data not shown), although the identity and number of genes interspersed between ftsK, pgsA3 and clgR varies to some degree (data not shown). As can be seen from the amino acid sequence alignment of ClgR proteins from representative members of the individual actinomycetes genera, the individual ClgR proteins are very similar to each other in the DNA-binding region (Fig. 7). The DNA-binding domain present in ClgR can also be found in several bacteriophage repressor proteins with known three-dimensional structure. Therefore, the amino acid residues within the DNA-binding domain interacting with DNA and thereby conferring recognition sequence specificity could be identified (Wintjens and Rooman, 1996). The residues in ClgR that are homologous to those shown to interact with DNA in the phage repressor proteins are highlighted in Fig. 7. With three exceptions (residues 55, 60 and 65 with respect to the C. glutamicum ClgR), they are conserved in all ClgR proteins, indicating that ClgR in these bacterial species might interact with very similar recognition sequences. In contrast, PopR from Streptomyces species, which probably arose via duplication of ClgR, displays four additional amino acid exchanges (R45K, V53I, G56P and L72V) at positions critical for DNA binding (Fig. 7). This might be responsible for the different specificities of PopR and ClgR concerning their target sequences (TCTGCC-N3-GGCARR versus WWCGCT-N3-RGCGWA). There is no apparent sequence conservation in the N- and C-terminal portions of ClgR and PopR proteins, apart from the high incidence of short-chained hydrophobic amino acids following a conserved proline residue in the extreme C-termini of some of these proteins (Fig. 7).
Nucleotide sequence analysis of the DNA regions upstream of the clpP1P2 and clpC genes revealed that putative ClgR recognition sites are also present in all sequenced Mycobacterium species, Streptomyces coelicolor and Thermobifida fusca. In Rhodococcus spp., a ClgR operator could only be found in front of clpC(Fig. 8). The central CGC-N5-GCG motif is fully conserved except for the putative ClgR operators in front clpP1P2 and clpC of Mycobacterium leprae, which lacks ClgR (Fig. 8).
The presence of ClgR proteins and ClgR operator sequences in front of clpP1P2 and clpC in other actinomycetes pointed to a conserved role for ClgR as an activator of clpC and clpP1P2 gene expression in actinomycetes. However, the lack of sequence conservation in the ClgR proteins apart from the DNA-binding domain left room for doubts with respect to conservation of function beyond mere binding to similar recognition sequences. Therefore, we amplified the M. tuberculosis clgR gene, cloned it into the expression vector pXMJ19, resulting in pXMJ19-clgR-Mtub, and introduced it into the C. glutamicumΔclgRΔclpC double mutant. Abundance of ClpP1 and ClpP2 in these strains was compared using two-dimensional electrophoresis of proteins (2D-PAGE), and both proteins were shown to be about 10-fold more abundant in the strain expressing M. tuberculosis ClgR (Fig. 1C). As this is comparable to the increase in ClpP1 and ClpP2 abundance observed upon overexpression of C. glutamicum clgR in this genetic background (data not shown), M. tuberculosis ClgR can apparently replace its C. glutamicum counterpart, at least with respect to activation of clpP1P2 expression. Therefore, based on the available experimental and in silico evidence, we propose that ClgR acts as a transcriptional activator of clpC and clpP1P2 gene expression in most actinomycetes.
Sequence motifs very similar to the putative −10 and −35 regions of the experimentally characterized σH-dependent promoters in front of C. glutamicum clpC and clpP1P2 are present upstream of clpC and clgR in all actinomycetes analysed in that respect and in front of the clpP1P2 operon of Corynebacterium species (Fig. 8). Moreover, σH-homologous proteins are encoded on the genomes of all completely sequenced actinomycetes (E-values < 10−62; data not shown), except for M. leprae, with the respective structural genes located at a conserved genomic location (data not shown).
Finally, sequences similar to the consensus HspR operator are present upstream of clgR in Corynebacterium efficiens and Corynebacterium diphtheriae, but not in other actinomycetes species, even though HspR is found in the entire group.
Although regulation of clp gene expression is well understood in enterobacteria and low G+C Gram-positive bacteria, only a very limited body of information is available concerning the corresponding regulatory phenomena in actinomycetes. Expression of the clpP3P4 operon, which is unique for Streptomyces species, has been shown to be positively controlled by the transcriptional activator PopR (Viala et al., 2000). PopR activity itself is apparently controlled via regulated proteolysis by the Clp protease, although it is not clear which conditions trigger stabilization of PopR and consequently activation of clpP3P4 expression (Viala and Mazodier, 2002). The similarities between PopR and the ClgR protein identified in this study are striking, not only with respect to primary sequence but also with respect to function, operator sequences and control of activity. However, in contrast to PopR-dependent expression of clpP3P4 (Viala et al., 2000), clpP1P2 expression at ambient temperature is not strictly dependent on ClgR. Also, the location of the ClgR binding site in front of the clpC transcriptional start site (−130 bp) is different from those of the ClgR and PopR operators in front of clpP1P2 and clpP3P4 respectively (−46 to −63 bp). We do not know whether this also implies different modes of transcriptional activation.
In the present study, we present data that clearly show that clpC and clpP1P2 gene expression in C. glutamicum is subject to complex control involving multiple regulatory pathways (Fig. 9). Expression of these genes is activated in response to severe heat stress, a process requiring the ECF sigma factor σH and probably mediated by σH-RNAP-dependent transcription of these genes. Transcription of the clgR gene is activated as well. The corresponding gene product is a transcriptional activator of clpC and clpP1P2 expression, which is very unstable because of ClpC-dependent degradation. Consequently, increased clgR transcription under conditions of severe heat stress does not translate into higher levels of ClgR and ClgR-dependent activation of clpC and clpP1P2 expression. Additional as yet unknown stimuli are required to stabilize ClgR. In summary, clpC and clpP1P2 expression is subject to complex regulation allowing signal integration via three different pathways. Signal input via σH (activation upon severe heat stress) and ClgR (stabilization in response to as yet unknown stimuli) results directly in activation of clpC and clpP1P2 expression. In contrast, signal input via activation of clgR transcription by deactivation of HspR and activation of σH is hierarchically subordinated, as it requires the presence of additional as yet unknown stimuli to be translated into increased amounts of ClpC, ClpP1 and ClpP2.
One of the main tasks to be tackled by future studies is the identification of stimuli resulting in stabilization of ClgR. All environmental stresses inducing clp gene expression in other bacteria (heat, osmotic, ethanol and oxidative stress as well as the accumulation of non-native proteins resulting from puromycin treatment) do not do so in C. glutamicum or, if so, they do it in a ClgR-independent fashion. Of course, we cannot rule out that, when applied under different regimes (different exposure times and/or severity of stress), some of these stimuli might result in higher expression of clpC and clpP1P2. We will address this question in future work and also test for the influence of additional stresses. Alternatively, it might just be impossible to mimic those environmental conditions in the laboratory that lead to ClgR stabilization in the native habitat. One intriguing way of controlling clpC and clpP1P2 expression via ClgR stability is that the block in ClpC-dependent degradation of ClgR is only possible in the presence of non-native proteins in amounts high enough virtually to saturate the ClpCP protease activity of the cell. Possibly, such conditions may not just be brought about by applying single, well-defined stresses in the laboratory. However, such conditions might be found if the proteolytic capacity of the cell is artificially downregulated, for instance by deleting genes coding for other proteases known to be involved in general protein turnover.
Our data suggest that ClgR degradation is performed by a ClpC-containing protease, probably ClpCP. The question arises as to which sequence determinants within ClgR are required for its recognition and/or degradation. ClpP-dependent degradation of the Streptomyces lividans PopR protein has been shown to depend on the presence of two carboxy-terminal alanine residues (Viala and Mazodier, 2002). The involvement of short-chained hydrophobic amino acid residues located at the last two C-terminal positions of protease substrates has also been shown for other proteins, such as the cell cycle regulator CtrA from Caulobacter crescentus and the ubiquitous SsrA tag, responsible for targeting proteins for degradation that have been translated from truncated mRNAs (Keiler et al., 1996; Domian et al., 1997). Such residues also exist in some ClgR proteins, but not in others, including those from Corynebacterium species. Therefore, at least in the case of Corynebacterium ClgR, other sequence motifs have to be responsible for its efficient degradation. These might not reside in the protein's C-terminus as the C-terminal addition of 10 amino acid residues (LEWSHPQFEK) in ClgR-C does not interfere with its degradation. We are currently investigating this issue by screening for stabilized ClgR proteins within a randomly mutagenized ClgR library.
In contrast to ClgR, the activation/deactivation pathways of σH and HspR are known to some degree. In a recent paper, the activity of σH was shown to be modulated by an antisigma factor, RshA, encoded by the gene downstream of sigH (Song et al., 2003). RshA was shown to bind to σHin vitro and in vivo, and this interaction as well as the interference with σH-dependent transcription was found to be inhibited by elevated temperatures and oxidizing conditions (Song et al., 2003). HspR activity is hypothesized to be modulated by the availability of DnaK. HspR binds to its operators only in complex with DnaK, and artificial depletion of the cellular DnaK pool results in activation of the HspR regulon (Bucca et al., 2003). Consequently, HspR is hypothesized to be inactivated by conditions resulting in the accumulation of non-native proteins, as these sequester DnaK in the course of refolding events.
Conservation of sequences typical for σECF-dependent promoters in front of clpC and clpP1P2 genes in other actinomycetes suggested the conservation of this regulatory mechanism in these organisms. However, expression of these genes was not found to be inducible by heat shock in M. tuberculosis or S. coelicolor (Viala et al., 2000; Stewart et al., 2002), although the heat shock regimes applied in the respective studies were more moderate (ΔT 8 and 10 K respectively) than those used in our study (ΔT 20 K). We used the term ‘severe heat stress’ in order to emphasize these differences. It should be noted here that, in C. glutamicum, induction of clpC and clpP1P2 expression is not as pronounced by far if the cells are subjected to a more moderate heat shock (30°C to 45°C) and not detectable when cells are transferred from 30°C to 40°C (data not shown). In contrast to clpC and clpP1P2, clgR expression is increased 11-fold upon exposure to moderate heat shock in M. tuberculosis (Stewart et al., 2002), supporting the idea that σH-dependent heat-triggered induction of clgR expression is conserved in this organism. Moreover, it confirms that increased clgR transcription alone is not sufficient for activation of clpC and clpP1P2 expression, possibly because M. tuberculosis ClgR is also an instable protein.
clgR, clpC and clpP1P2 expression is slightly (1.5- to twofold) increased in M. tuberculosis exposed to oxidative stress (5 mM diamide; Manganelli et al., 2002) and detergent stress (0.05% SDS; Manganelli et al., 2001). This activation depends on σH (oxidative stress) and σE (detergent stress) respectively (Manganelli et al., 2001; 2002), raising the question whether the putative σH-dependent promoters in front of these genes can also be used by σE if this protein is activated by appropriate environmental conditions. Experimentally determined σH- and σE-dependent promoters from M. tuberculosis are virtually indistinguishable by mere sequence inspection (Manganelli et al., 2001; 2002), and it has been shown that the σH-dependent promoter in front of clpB can also be recognized by σE- RNAP (Raman et al., 2001).
In summary, the data presented in this study give a first picture of the complex regulation of clpC and clpP1P2 expression in C. glutamicum. However, additional research especially concerning the input routes of individual environmental stresses and the pathways controlling ClgR stability needs to be performed.
Bacterial strains, media and growth conditions
Corynebacterium glutamicum was cultivated aerobically on a rotary shaker (150 r.p.m.) at 30°C in Luria–Bertani (LB) medium (Sambrook et al., 1989) with 2% (w/v) glucose or in BHIS medium [brain–heart infusion (Difco) with 0.5 M sorbitol]. In order to subject C. glutamicum strains to heat shock, 80 ml of culture was transferred to 500 ml Erlenmeyer shake flasks preheated to 50°C in a rotary shaker and then incubated at 150 r.p.m. For strain construction and maintenance, BHIS agar plates were used. E. coli DH5α and E. coli BB1553 (Tomoyasu et al., 2001) were grown aerobically on a rotary shaker (150 r.p.m.) at 37°C (DH5α) or 30°C (BB1553) in LB medium or plated onto LB agar plates [LB medium with 1.5% (w/v) agar]. E. coli BB1553 harbours a deletion in dnaK and was used for the purification of ClgR. If appropriate, antibiotics were used at the following concentrations: chloramphenicol 10 µg ml−1 (C. glutamicum and E. coli), kanamycin 25 µg ml−1 (C. glutamicum and E. coli BB1553) or 50 µg ml−1 (E. coli DH5α). IPTG was added to a final concentration of 1 mM if C. glutamicum strains carrying pXMJ19 or a derivative were cultivated (Table 2).
In frame deletion of the clpC and clgR genes in C. glutamicum was essentially performed as described previously (Niebisch and Bott, 2001). The clpC up- and downstream regions were amplified using the oligonucleotide pairs Delta clpC-1/Delta clpC-2 and Delta clpC-3/Delta clpC-5 respectively. The corresponding regions of clgR were amplified using the oligonucleotide pairs Delta clgR-1/Delta clgR-2 and Delta clgR-3/Delta clgR-4. The respective up- and downstream regions were joined using the two oligonucleotides introducing restriction sites (Delta clpC-1/Delta clpC-5 and Delta clgR-1/Delta clgR-4 respectively). The resulting cross-over PCR products were digested and cloned into respectively cut pK19mobsacB (Schäfer et al., 1994). Transformation of the pK19mobsacB derivatives into C. glutamicum, screening for the first and second recombination event as well as confirmation of the chromosomal deletion was performed as described previously (Niebisch and Bott, 2001). For construction of the ΔclgRΔclpC double mutant, the ΔclpC single mutant was used as recipient of pK19mobsacB-Delta clgR. Construction of the ΔclpX, ΔsigE, ΔsigH and ΔhspR mutants was performed as described for the ΔclgR mutant using oligonucleotides labelled respectively (see Table 3).
. In some cases, oligonucleotides were designed to introduce recognition sites for restriction endonucleases (recognition sites underlined, restriction endonucleases indicated in parentheses), complementary 21-mer sequences for generating cross-over PCR products (printed in italics), sequence tags for labelling PCR products with biotin using Biotinprimer 1 (printed in bold).
. Oligonucleotide adds three glutamate codons to the C-terminus of ClpCΔ533-931 (printed in italics), introduces an opal stop codon after codon 532 (printed in bold and italics) and leads to an exchange of an arginine codon by an aspartate codon (printed in bold).
TAT AGT CGA CTC ATT CTT CTT CCA ATC CTT CTT CCA TGT TGA GCA GGT CTG AAG (SalI)b
CTT TCC GTG GGG CAG ATT CTG
CGC GGA GAA GTG GCT CTG GAG
CTT ATA CAA CCC TTC TAG ACA AGC CG
TCT TCC TGC GCT GCT TGC AGC
TGC CCT TCC GGC GAT GTG CAA
CTG GAA CCA TGT CGT CCC AGA G
TAT ACC CGG GCA CCT GAT GAC GAT TGC GAG GTG G (XmaI)
TAT AGA ATT CTT GTC GAC CCG GAC CGA CCG GCG AC (EcoRI)
For overproduction and purification of ClgR, the region coding for clgR was amplified using oligonucleotides introducing an NdeI restriction site at the translation initiation codon (OE-ClgR-fw) and a XhoI restriction site preceding the stop codon (OE-ClgR-CT-rv). The purified PCR product was cloned into the modified expression vector pET24b-Streptag (M. Meyer and M. Bott, unpublished), and the resulting plasmid pET24b-clgR-C was transferred into E. coli BL21 (DE3).
For purification of ClgR-C from E. coli BB1553, the region coding for ClgR with StrepTag II including the vector-derived translational start signals was amplified from pET24b-clgR-C with the oligonucleotides pEK-clgR-fw and pEK-clgR-rv. The PCR product was digested with EcoRI–SalI and cloned into the respectively cut vector pEKEx1 (Eikmanns et al., 1991). The resulting plasmid pEKEx1-clgR-C was transferred into E. coli BB1553.
In order to get a stronger signal in primer extension analysis of the clpC promoter region and to analyse the abundance of the inactive ClpC derivative ClpCΔ533-931, the plasmid pJC1-clpCΔ533-931 was constructed. The promoter region together with the clpC 5′ end was amplified with the oligonucleotides clpC-HOE-fw-2 and OE-clpC-rv-3. Oligonucleotide OE-clpC-rv-3 introduces a number of amino acid exchanges and additions at the C-terminus of the truncated ClpC protein (for details, see Table 3), lowering the isoelectric point of ClpCΔ533-931 and allowing for its detection on two-dimensional gels. The PCR product was digested with SalI–XbaI and cloned into the E. coli/C. glutamicum shuttle vector pJC1 (Cremer et al., 1990).
For heterologous overexpression of the M. tuberculosis clgR, the gene was amplified using chromosomal DNA of M. tuberculosis H37Rv, kindly provided by Sabine Rüsch-Gerdes (Forschungszentrum Borstel, Germany), and oligonucleotides clgR-Mtub-fw and clgR-Mtub-rv. The PCR product was digested with XmaI–EcoRI and cloned into respectively cut pXMJ19 (Jakoby et al., 1999)
DNA microarray analyses
The generation of whole-genome DNA microarrays, extraction of total RNA from C. glutamicum, cDNA synthesis and labelling, microarray hybridization and washing as well as data analysis were performed as described previously (Lange et al., 2003).
Primer extension analysis
Non-radioactive primer extension analysis was performed using IRD800-labelled oligonucleotides (MWG Biotech). Total RNA (10 µg) was combined with 2 pmol of labelled oligonucleotide and 2 µl of 5× annealing buffer (50 mM Tris-HCl, pH 7.9, 1.25 M KCl) in a total volume of 10 µl. The reaction was heated to 65°C for 5 min and then slowly (0.5°C/2 min) cooled to 42°C in a thermocycler. In order to perform reverse transcription, 23 µl of water was combined with 10 µl of 5× reverse transcription buffer (250 mM Tris-HCl, pH 8.3, 125 mM KCl, 15 mM MgCl2), 5 µl of 100 mM dithiothreitol (DTT), 1 µl of deoxyribonucleotides (25 mM ATP, CTP, GTP and TTP), 0.5 µl of actinomycin D (5 mg ml−1 in ethanol) and 0.5 µl of SuperScript II RNase H– reverse transcriptase (200 units µl−1; Life Technologies) in that order, and the entire mix was added to the RNA immediately upon the annealing step. After incubation for 1 h at 42°C, the reaction was stopped by adding 120 µl of RNase A reaction mix (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 100 mM NaCl, 100 µg ml−1 sonicated salmon sperm DNA, 200 µg ml−1 RNase A, added freshly), incubated for 1 h at 37°C and the DNA precipitated overnight after adding 17 µl of 3 M sodium acetate (pH 5.2) and 380 µl of ice-cold ethanol. After washing the pellet with 500 µl of 70% (v/v) ethanol and air drying, the DNA was dissolved in 2 µl of water and 2 µl of formamide loading dye (Epicentre Technologies). A sample of 1 µl was then loaded onto a denaturing 4.6% (w/v) Long Ranger (Biozym) sequencing gel (separation length 42 cm) and separated in a Long Read IR DNA sequencer (Licor). The length of the reaction products was determined by running the four lanes of a sequencing reaction set-up using the same oligonucleotide as for reverse transcription alongside the primer extension products. Each transcriptional start was determined using two different oligonucleotides (clpC-PE3 and clpC-PE2 for clpC, clpP-PE1 and clpP-BS1-rv for clpP1P2, PE-clgR-1 and PE-clgR-2 for clgR).
Quantitative RT-PCR analysis of clgR expression was performed using the oligonucleotides RT-clgR-fw/RT-clgR-rv for clgR and RT-dnaE-fw/RT-dnaE-rv for the dnaE gene coding for the α-subunit of DNA polymerase and serving as an internal control. Total RNA (5 µg) was incubated with 2 pmol of each of the oligonucleotides RT-clgR-rv and RT-dnaE-rv for 10 min at 65°C in a total volume of 12 µl. After brief incubation on wet ice, cDNA synthesis was started by adding 4 µl of the supplied 5× reaction buffer, 2 µl of 100 mM DTT, 1 µl of deoxyribonucleotides (25 mM ATP, CTP, GTP and TTP) and 0.8 µl of SuperScript II RNase H– reverse transcriptase (200 units µL−1; Life Technologies). The reaction was carried out at 42°C for 90 min. An aliquot of 2 µl of the RT reaction mixture was used as template in the following PCR set-up as described recently (Schlösser et al., 2001) with 22 cycles of template denaturation, oligonucleotide annealing and extension. Reaction products were separated by electrophoresis in 2% (w/v) agarose gels and visualized upon staining with ethidium bromide.
2D-PAGE and mass spectrometry
Preparation of C. glutamicum cell-free extracts (for enrichment of ClgR and 2D-PAGE), sample preparation for 2D-PAGE, 2D-PAGE, protein detection by colloidal Coomassie staining, in-gel trypsin digestion and peptide mass fingerprinting using MALDI-TOF mass spectrometry were performed as described previously (Schaffer et al., 2001).
DNA affinity purification of ClgR
In order to enrich a DNA-binding protein interacting with the clpP1P2 and clpC regulatory regions, these regions were amplified using oligonucleotides clpP-activator-fw and clpP-activator-rv (clpP1P2; fragment length 310 bp) and clpC-BS1-fw and PclpC-ClgR (clpC; fragment length 272 bp). For biotinylation, the PCR products were reamplified using the respective forward oligonucleotides and Biotinprimer 1, which carries a biotin label at the 5′ end and is identical in sequence to the 5′ end of the reverse oligonucleotides used for the first amplification. Labelled PCR products were purified using the QIAquick PCR purification kit (Qiagen) and 220–370 pmol of labelled DNA fragments immobilized on Dynabeads® streptavidin M-280 (Dynal) as recommended by the manufacturer. For protein binding, the immobilized DNA fragments were incubated for 45 min at room temperature with 190–280 mg of cellular protein in TGED buffer [20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 10% (v/v) glycerol, 0.01% (v/v) Triton X-100, 100 mM NaCl and 1 mM freshly added DTT] with 500 µg of chromosomal DNA from C. glutamicum in a 50 ml Falcon tube with just enough shaking to prevent sedimentation of the paramagnetic beads (≈ 150 r.p.m.). Subsequently, the reaction was transferred into microcentrifuge tubes, washed once with 1 ml of TGED buffer, twice with 1 ml of TGED buffer plus 400 µg of chromosomal DNA from C. glutamicum and finally with 1 ml of TGED buffer. Proteins bound to the immobilized DNA were eluted by washing the beads twice with 350 µl of elution buffer (TGED buffer with 2 M NaCl). The eluates were pooled, concentrated and desalted with Microcon 3 microconcentrators (Millipore) and analysed by denaturing PAGE.
Overproduction and purification of ClgR-Strep Tag II
Escherichia coli BB1553 carrying pEKEx1-clgR-C was cultivated at 30°C until the OD600 reached a value between 0.3 and 0.5. Then, synthesis of the ClgR-C was induced by the addition of IPTG to a final concentration of 1 mM, and the cultures were incubated for another 2 h. Cells were washed once and resuspended in 10 ml of buffer W (100 mM Tris-HCl, pH 8, 1 mM EDTA).
After the addition of 1 mM diisopropylfluorophosphate and 1 mM phenylmethylsulphonyl fluoride (PMSF), the cell suspension was passed three times through a French pressure cell (207 mPa, SLM Aminco Spectronic Instruments). Intact cells and cell debris were removed by centrifugation (20 min, 5000 g, 4°C). The cell-free extract was subjected to ultracentrifugation (1 h, 50 000 g, 4°C), and the supernatant was applied to a StrepTactin Sepharose column with a bed volume of 2 ml (IBA). Subsequently, purification was performed as suggested by the manufacturer.
In fractions containing ClgR-C, the elution buffer was exchanged against buffer containing 20 mM Tris-HCl, pH 7.5, 5% (v/v) glycerol by gel filtration with Sephadex G-25 (PD-10 column; Amersham Pharmacia Biotech). Protein concentrations were determined with the BCA protein assay kit (Pierce) using bovine serum albumin (BSA) as standard. The purity of the protein preparation was assessed by denaturing PAGE and subsequent protein detection with Gel Code blue stain reagent (Pierce). Using this protocol, ClgR-C was purified to apparent homogeneity.
For purification of ClgR-C from C. glutamicum, the respective strains were cultivated in 100 ml of BHIS and harvested at an OD600 of about 5. Cell-free extracts were prepared as described for 2D-PAGE and subjected to StrepTactin affinity chromatography as suggested by the manufacturer, except that all buffers contained 300 mM instead of 100 mM NaCl. The eluates were pooled, concentrated and desalted with Microcon 3 microconcentrators (Millipore). One-fifth of the total eluate was then analysed by SDS-PAGE.
DNase I footprinting assays
Labelled DNA fragments were obtained by amplification with 5′ end IRD800 labelled oligonucleotides (MWG Biotech). The clpC regulatory region was amplified with oligonucleotides clpC-BS1-fw/IRD800-clpC-BS1-rv (labelled non-template strand) and IRD800-clpC-BS2-fw/clpC-BS2-rv (labelled template strand), and the clpP1P2 regulatory region with oligonucleotides clp-activator-fw/IRD800-clpP-BS1-rv (labelled non-template strand) and IRD800-clpP-FP4/clpP-BS2-rv (labelled template strand). In a total volume of 200 µl, 1.5–3 nM labelled fragments were combined with 20 µl of 5× binding buffer [100 mM Tris-HCl, pH 7.5, 5 mM EDTA, 50% (v/v) glycerol, 5 mM DTT, 0,05% (v/v) Triton X-100 and 500 mM NaCl], 10 µl of salt solution (100 mM MgCl2, 50 mM CaCl2), 0.5 µl of 1 µg µl−1 poly [d(IC)] and different amounts of ClgR (0–4.4 µM). Then, the binding reactions were incubated for 30 min at room temperature. Subsequently, 5 µl of 0.25 µg ml−1 DNase I in DNase I buffer (10 mM Tris-HCl, pH 8, 5 mM MgCl2, 5 mM CaCl2, 50 mM KCl and 1 mM DTT) was added, and digestion was allowed to proceed for exactly 2 min, stopped with 700 µl of ice-cold stop solution [645 µl of 96% ethanol (v/v), 5 µl of salmon sperm DNA (4 µg µl−1), 50 µl of saturated ammonium acetate solution], and DNA was precipitated overnight at −20°C.
After washing the DNA pellet with 500 µl of 70% (v/v) ethanol and air drying, the DNA was dissolved in 2.5 µl of water and 2.5 µl of formamide loading dye (Epicentre Technologies). The reaction products were separated as described for primer extension analysis
We are indebted to Christian Lange, Andreas Krug, Georg Sindelar, Tino Polen and Volker Wendisch (Forschungszentrum Jülich, Germany) for generously providing whole-genome DNA microarrays of C. glutamicum and for introducing us to the methods concerning their use. We are grateful to Brita Weil for technical assistance, and to B. Bukau (ZMBH Heidelberg, Germany) for providing E. coli BB1553, as well as to Sabine Rüsch-Gerdes (Forschungszentrum Borstel, Germany) for providing chromosomal DNA of M. tuberculosis H37Rv. We thank Jörn Kalinowski and Christof Larisch (Universität Bielefeld) for providing strains C. glutamicum RES167 ΔsigB and RES167 ΔsigE for initial experiments concerning identification of the sigma factor responsible for heat-triggered induction of clpC, clpP1P2 and clgR, as well as for sharing results before publication. We are especially grateful to Hermann Sahm (Forschungszentrum Jülich, Germany) for continuous support of our work.