Correspondence: Heung-Shick Lee, Department of Biotechnology and Bioinformatics, Korea University, Jochiwon, Chungnam 339-700, Korea. Tel.: +82 2 3290 3436; fax: +82 41 864 2665; e-mail: email@example.com
The Corynebacterium glutamicum whcA gene is known to play a negative role in the expression of genes responding to oxidative stress. The encoded protein contains conserved cysteines, which likely coordinate the redox-sensitive Fe–S cluster. To identify proteins which may interact with WhcA, we employed a two-hybrid system utilizing WhcA as ‘bait’. Upon screening, several partner proteins were isolated from the C. glutamicum genomic library. Sequencing analysis of the isolated clones revealed out-of-frame peptide sequences, one of which showed high sequence homology with a dioxygenase encoded by NCgl0899. In vivo analysis of protein interaction using real-time quantitative PCR, which monitors his3 reporter gene expression, demonstrated that the interaction between NCgl0899-encoded protein and WhcA was specific. The interaction was labile to oxidants, such as diamide and menadione. Based on these data, NCgl0899 was named spiA (stress protein interacting with WhcA). Physical association and dissociation of the purified His6–WhcA and GST–SpiA fusion proteins, as assayed by in vitro pull-down experiments, were consistent with in vivo results. These data indicated that the interaction between WhcA and SpiA is not only specific but also modulated by the redox status of the cell and the functionality of the WhcA protein is probably modulated by the SpiA protein.
Corynebacterium glutamicum is a Gram-positive bacteria that belongs to the order Actinomycetales, which also includes the genera Mycobacterium and Streptomyces (Ventura et al., 2007). Corynebacterium glutamicum is a remarkable organism and is capable of producing a variety of amino acids and nucleotides in large quantities (Leuchtenberger et al., 2005). Because of the industrial importance of this organism, its relevant genetic and biochemical features have been extensively characterized. Accordingly, strategies that C. glutamicum cells adopt in response to cellular stresses have attracted scientific interests in recent years.
WhiB-like genes are a class of genes that perform diverse cellular processes, such as cell division, differentiation, pathogenesis, starvation survival, and stress response (Gomez, 2000; Steyn et al., 2002; Kim et al., 2005; Geiman et al., 2006; Raghunand & Bishai, 2006; Singh et al., 2007; Choi et al., 2009). The whiB gene, which was originally identified and characterized in Streptomyces coelicolor, is a developmental regulatory gene that is essential for the sporulation of aerial hyphae (Davis & Chater, 1992). The whiB homologues are only found in the order Actinomycetales. Seven whiB homologues have been identified in the Mycobacterium tuberculosis genome and at least six are present in S. coelicolor (Soliveri et al., 2000), whereas only four are found in C. glutamicum (Kim et al., 2005). The WhiB-like proteins have four conserved cysteine residues that bind to a redox-sensitive Fe–S cluster (Jakimowicz et al., 2005; Alam et al., 2007; Singh et al., 2007; Crack et al., 2009; Smith et al., 2010), which plays a critical role in controlling protein function. In general, the cluster loss reaction followed by oxidation of the coordinating cysteine thiols that form disulfide bridges is important for activity. For example, S. coelicolor WhiD loses its Fe–S cluster upon exposure to oxygen (O2) and the apo-WhiD may play important roles in cell physiology (Crack et al., 2009). Some WhiB-like proteins may function as transcription factors, as suggested by the presence of predicted helix–turn–helix DNA-binding motif. Recently, the M. tuberculosis WhiB1 protein in its apo-form was shown to have DNA-binding activity (Smith et al., 2010). On the other hand, the WhiB-like proteins, WhiB1 and WhiB4, of M. tuberculosis are thioredoxin-like proteins and apparently function as protein disulfide reductases and probably repair oxidized proteins through thiol-disulfide exchange (Alam et al., 2007; Garg et al., 2007). Subsequently, α-(1,4)-glucan branching enzyme GlgB was identified in a yeast two-hybrid screen as one of the in vivo substrates of M. tuberculosis WhiB1 (Garg et al., 2009).
Among the four whiB-like genes of C. glutamicum, only whcE and whcA have been studied so far. The whcE gene plays a positive role in the survival of cells exposed to oxidative and heat stress (Kim et al., 2005). The whcA gene plays a negative role in the expression of genes involved in the oxidative stress response (Choi et al., 2009). As WhcE and WhcA are presumably redox-sensitive proteins with conserved cysteine residues coordinating the Fe–S cluster, the activity and functionality of the proteins are likely conveyed through interactions with other proteins. We therefore developed a two-hybrid screening system using WhcA as bait and identified several partners, among which a putative dioxygenase encoded by NCgl0899 turned out to be relevant to WhcA. According to the physiological and biochemical data, we propose a model for the whcA-mediated stress response pathway.
Materials and methods
Strains, media, and growth conditions
Escherichia coli DH10B (Invitrogen) was utilized for the construction and propagation of plasmids. Escherichia coli BL21 DE3 (Merck, Germany) was employed for the expression of whcA (His6–WhcA) and spiA (GST–SpiA) cloned into pET28a (Merck) and pGEX-4T-3 (GE Healthcare), respectively. Cells carrying the two plasmids were named HL1386 and HL1337, respectively. Strain HL1387 carrying pBT-whcA and pTRG-NCgl0899 was used in assays involving diamide. Unless otherwise stated, E. coli and C. glutamicum cells were cultured at 37 °C in Luria–Bertani broth (Sambrook & Russell, 2001) and 30 °C in MB medium (Follettie et al., 1993), respectively. Selective and nonselective media were prepared as described previously (BacterioMatch II Two-Hybrid System, Agilent Technology). Antibiotics were added at the following concentrations: 20 μg ampicillin mL−1; 10 μg tetracycline mL−1; and 30 μg kanamycin mL−1.
DNA analysis and plasmid construction
Plasmid pSL482 carrying whcA cloned into the pBT vector (Agilent Technology) was constructed by introducing the BamHI-digested fragment, which was amplified from the C. glutamicum chromosome with primers 5′-GGAATTCCATATGATGACGTCTGTGATT-3′ and 5′-CCCAAGCTTAACCCCGGCGAT-3′, into the vector. Plasmid pSL487 (pTRG-NCgl0899), which carries spiA/NCgl0899, was constructed as follows. The chromosomal gene was amplified with primers 5′-TGCCATGAGCATCCTTGACA-3′ and 5′-AAAGCACTCCCCCCAACATT-3′ and cloned into the pGEM-T-easy vector (Promega). Then, the NotI fragment was isolated and inserted into the pTRG vector. Plasmids pSL484 (pTRG-NCgl1708), pSL485 (pTRG-NCgl1938), pSL486 (pTRG-NCgl0108), and pSL488 (pTRG-NCgl1141) were constructed using the same procedure described above except the primers were as follows: pTRG-NCgl1708, 5′-GTAGAATATAACTATGCCAACACTGC-3′ and 5′-GCCATACGCAGCCTACATT-3′; pTRG-NCgl1938, 5′-AGCATGTAACTTGTCTACCCCA-3′ and 5′-CACCCGACTAACCAAACCA-3′; pTRG-NCgl0108, 5′-TTCCACCATGAACACCCCAC-3′ and 5′-ATGACGAGGCGATGTGTGG-3′; pTRG-NCgl0899, 5′-TGCCATGAGCATCCTTGACA-3′ and 5′-AAAGCACTCCCCCCAACATT-3′; pTRG-NCgl1141, 5′-AGGAGGTGCAGTACTAATGAAGGTC-3′ and 5′-CATCCGCCTAGTCCTTCCTTG-3′. The pSL487 plasmid expressing the GST–SpiA fusion protein was constructed by ligating the BamHI–XhoI fragment from pSL487 into the pGEX-4T3 vector. The pSL494 plasmid expressing the His6–WhcA fusion protein was constructed via the amplification of the whcA gene using the primers 5′-CCCAAGCTTTCATGACGTCTGTGATT-3′ and 5′-CCCAAGCTTTTAAACCCCGGC-3′, and by subsequently digesting the fragment with HindIII and ligating the DNA with the HindIII-digested pET28a vector.
Corynebacterium glutamicum (100 μL) genomic DNA (2 μg μL−1), isolated as described by Tomioka et al. (1981), was partially digested with 0.195 U of SauIIIA1 for 1 h at 37 °C. DNA fragments 1–3 kb in size were isolated and inserted into the BamHI-digested pTRG vector. The recombinants were introduced into E. coli cells and plasmids were isolated and pooled from approximately 10 000 transformants.
The BacterioMatch II Two-Hybrid System (Agilent Technology) was used according to the manufacturer's instructions. Briefly, the two plasmids, pBT and pTRG, containing the ‘bait’ and target genes, respectively, were used to simultaneously transform E. coli. Protein–protein interactions were screened based on expression of his3 and aadA, which confer histidine prototrophy (His+) and streptomycin resistance (Str+), respectively. For screening, 50 ng of each pBT-whcA and target library DNA was introduced into reporter cells and spread onto the selective media (His− and Str+). Colonies were isolated and the plasmids in the growing cells were analyzed.
Total RNA was prepared with the NucleoSpin RNA II Kit (Macherey-Nagel, Germany). cDNA conversion was carried out with DyNAmo™ cDNA Synthesis Kit (FinnZymes, Finland). Real-time quantitative PCR (RT-qPCR) was performed using a CFX96™ Real-Time PCR Detection System (Bio-Rad). Different gene expressions were normalized to the levels of 16S rRNA gene transcripts. The degree of change in expression was calculated with the method using cfx™manager software (Bio-Rad). Primers used for the quantification of the reporter gene his3 were as follows: sense primer 5′-CGCTAATCGTTGAGTGCATTG-3′; antisense primer 5′-CGCAAATCCTGATCCAAACC-3′. 16S rRNA gene transcripts were amplified with sense primer 5′-TGGGAACTGCATCTGATACTGGCA-3′ and antisense primer 5′-TCTACGCATTTCACCGCTACACCT-3′.
Protein purification and pull-down assay
The GST–SpiA fusion protein was expressed and purified using the GSTrap™ FF column (GE Healthcare), in accordance with the manufacturer's instructions. Pull-down experiments were performed with purified recombinant proteins. His6–WhcA was overexpressed, isolated, and refolded on a HisTrap™ FF column (GE Healthcare) as described in the Recombinant Protein Purification Handbook (GE Healthcare). Then, the opposite fusion protein, GST–SpiA protein (6 mg), was applied to the column. When needed, dithiothreitol, which was added to the refolding buffer, was substituted with 1 mM diamide. Eluted protein samples were analyzed by SDS-PAGE. Purified maltose-binding protein (5 mg) was applied to a Ni-NTA column with bound His6–WhcA protein and treated as described above to assess nonspecific binding.
Effect of oxidant on protein–protein interaction
HL1387 cells were grown to log phase in nonselective media, followed by diamide addition to a final concentration of 0.25–0.5 mM. After an additional 2-h incubation, transcripts were isolated and the amount of his3 mRNA was analyzed by RT-qPCR. If necessary, diamide was substituted with menadione, which was added to a final concentration of 0.5 mM.
Isolation of proteins interacting with WhcA
A BacterioMatch II Two-Hybrid System was used to search for proteins that interact with WhcA. After transformation of the reporter strain with pSL482 (pBT-whcA) and C. glutamicum target library, five clones that exhibited efficient growth on selective media were recovered, and the plasmids were isolated and sequenced. One of the plasmids contained a 243-bp fragment of the ispG gene encoding the C-terminal region (starting from the amino acid at position 151) of the 4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase. Four others contained out-of-frame genes that expressed peptide sequences. Subsequently, we searched genes whose protein products had a high homology with the peptides. ORFs NCgl1708- (hypothetical protein), NCgl0108- (NADPH-dependent dehydrogenase), NCgl0899- (dioxygenase), and NCgl1141 (nitrate reductase)-encoded proteins showed a homology with the respective peptide sequences (Table 1). To verify the interaction of the encoded proteins with WhcA, we cloned the full-length ORFs of the above five clones into the pTRG vector, introduced them into reporter cells carrying the bait vector pBT-whcA, and monitored growth on selective media. Aside from the cells carrying pTRG-NCgl1141, all others grew efficiently on the media. These protein–protein interactions were then quantified by measuring the transcript level of the reporter gene his3 by RT-qPCR. Cells carrying pTRG-NCgl0899 showed the highest transcript level, which corresponded to 37% compared with the positive control cells (Fig. 1). The transcript level for cells carrying pTRG-NCgl0108 was 25% relative to the positive control cells. In contrast, the transcript levels for cells carrying the NCgl1708- and NCgl1938-encoded proteins were close to the background level (Fig. 1). As the NCgl0899-encoded protein (now designated as SpiA) showed the strongest interaction with WhcA, we further analyzed and characterized this interaction.
Table 1. Proteins and peptide sequences screened to interact with WhcA
Reading frame of the cloned gene that produced the screened sequences.
† Identity (percent of matched amino acid sequence).
243-bp DNA encoding part of IspG
Demonstration of WhcA–SpiA protein interaction in vitro
A direct physical interaction between WhcA and SpiA was tested by a protein-binding ‘pull-down’in vitro experiment using His6–WhcA fusion that was bound on beads and incubated with the GST–SpiA fusion protein. Because the overexpressed His6–WhcA fusion protein formed insoluble inclusion bodies, it was bound to the beads using an on-column renaturation technique. Finally, protein–protein interactions were identified by SDS-PAGE. As shown in Fig. 2a, 16.7 kDa His6–WhcA and 62.2 kDa GST–SpiA coeluted together, indicating specific binding. Nonspecific binding of the GST–SpiA protein to the beads was not observed (data not shown). Purified maltose-binding protein, which was used to assess nonspecific interactions, did not bind to the bait His6–WhcA (data not shown). However, the band intensity of the GST–SpiA protein was lighter than expected (Fig. 2a), suggesting a weak protein interaction. If protein–protein interactions occurred at a 1 : 1 molar ratio, the band intensity of the GST–SpiA protein, which was three times larger than the His6–WhcA in size, should be approximately three times stronger than that of the His6–WhcA band. This discrepancy could be due simply to inefficient refolding, leaving only a fraction of the bead-bound His6–WhcA in the correct conformation. Alternatively, fractions of the refolded His6–WhcA could have lost their Fe–S cluster during the denaturation–refolding process, thus remaining in an alternative conformation that does not interact with GST–SpiA (see Discussion). Nevertheless, the pull-down assay indicated that WhcA can specifically bind the SpiA protein.
Effect of oxidant on protein–protein interaction
So far, we were able to show that WhcA interacts with SpiA via in vivo and in vitro assays. As the WhcA protein was found to play a negative role in the oxidative stress response pathway, we postulated that the protein–protein interaction could be affected by external factors, such as external redox environments. When oxidant diamide was applied to growing HL1387 cells, the interaction between WhcA and SpiA was significantly reduced to 34% relative to those of positive and negative control strains (Fig. 3a). The effect of oxidant menadione was observable but rather marginal (Fig. 3b), whereas reductant dithiothreitol was not effective at all in disrupting the protein–protein interaction (data not shown). Whereas the thiol-specific oxidant diamide specifically oxidizes sulfhydryl groups (Kosower & Kosower, 1995), the redox-cycling compound menadione exerts its toxic effects via stimulating intracellular production of superoxide radicals and hydrogen peroxide (Hassan & Fridovich, 1979). However, the redox-cycling compound is also known to drain electrons from the reductive pathways, including the thioredoxin system (Holmgren, 1979), thus inducing disulfide bond formation in cells. The differential response of the protein to diamide and menadione may suggest that the cysteine residues of the WhcA protein are involved in disulfide bond formation.
To study the effect of diamide on in vitro protein–protein interactions, the pull-down assay was performed in the presence of oxidant diamide, as described in Materials and methods. To our surprise, when diamide was present during the refolding of WhcA, coelution of the WhcA and SpiA proteins was not detected (Fig. 2b), suggesting that the WhcA protein undergoes conformational changes, probably by losing its Fe–S cluster that leads to disulfide bond formation between cysteine residues. Collectively, these data indicated that the protein interaction was modulated by cellular redox conditions. Based on these data, the ORF NCgl0899-encoded protein was named SpiA (stress protein interacting with WhcA).
The C. glutamicum WhcA has been suggested to play a negative role in the oxidative stress response pathway (Choi et al., 2009). However, it is not known how the action of WhcA is regulated. The WhcA protein appeared to contain Fe–S clusters. The primary sequence of WhcA contained a likely Fe–S cluster-binding motif consisting of four conserved cysteine residues C-X29-C-X2-C-X5-C (where X is any amino acid) (Jakimowicz et al., 2005). In addition, aerobically isolated WhcA protein was reddish-brown in color (data not shown), a characteristic feature of Fe–S cluster proteins, although the refolded protein showed a diminished color. Fe–S proteins are known to play important roles in sensing external signals as well as the intracellular redox state of microbial cells (Green & Paget, 2004). Interacting proteins may transfer signals to the WhcA protein or help the WhcA protein sense cellular redox status. The isolated protein SpiA was annotated to encode 2-nitropropane dioxygenase, which is involved in the detoxification of nitroalkanes by oxidizing compounds to their corresponding carbonyl compounds and nitrite (Kido & Soda, 1978; Gorlatova et al., 1998). The protein contains FMN or FAD and belongs to a group of NADPH-dependent oxidoreductase (Marchler-Bauer et al., 2011). In accordance with this, the purified SpiA protein was yellowish in color (data not shown). The fact that the interaction between WhcA and SpiA was affected by oxidant diamide and menadione indicated that the activity of WhcA was probably modulated by SpiA. The annotated function of SpiA as an oxidoreductase (or dioxygenase) is in agreement with this notion.
The WhiB3 protein from M. tuberculosis was shown to function as intracellular redox sensor responding to O2 through its Fe–S cluster (Singh et al., 2007). The WhiB4 protein also contains a Fe–S cluster. Upon exposure to O2, the holo-WhiB4 protein loses its Fe–S cluster and becomes active, functioning as a protein disulfide reductase. The apo-form of the protein accepts electrons either from an unidentified reductase or directly from an unidentified reductant and becomes activated (Alam et al., 2007). The active form of the protein then transfers the signal to the oxidized target proteins as a disulfide reductase (Alam et al., 2007). However, it is still not known how WhiB3 and WhiB4 proteins respond to O2. In C. glutamicum, the SpiA protein, annotated as oxygenases or oxidoreductases, might be the molecule that is involved in making the WhcA protein respond to O2. The purified M. tuberculosis WhiB1 does not respond to O2, which further supports the notion that SpiA is involved in the whcA-mediated stress response pathway. Collectively, our data suggest that the WhcA protein from C. glutamicum may function in a similar but unique fashion. Under normal growth conditions, SpiA may reduce apo-WhcA (S–S) to its holo form (Fe–S). During this process, the WhcA protein attains its Fe–S cluster, gains its ability to bind to DNA, and represses genes involved in oxidative stress response. However, under conditions of oxidative stress, the WhcA protein loses its Fe–S cluster, leading to the loss of its DNA-binding ability. Nevertheless, the DNA-binding activity of the WhcA protein has not yet been shown. To summarize, a regulatory model involving WhcA and SpiA is shown in Fig. 4.
This work was supported by a National Research Foundation grant (to H.-S. Lee) from the Korean Ministry of Education, Science and Technology (MEST 2010-0021994 Program of the NRF).