|Home | About | Journals | Submit | Contact Us | Français|
Upon exposure to alkylating agents, Escherichia coli increases expression of aidB along with three genes (ada, alkA, and alkB) that encode DNA repair proteins. While the biological roles of the Ada, AlkA, and AlkB proteins have been defined, despite many efforts, the molecular functions of AidB remain largely unknown. In this study, we focused on the biological role of the AidB protein, and we demonstrated that AidB shows preferential binding to a DNA region that includes the upstream element of its own promoter, PaidB. The physiological significance of this specific interaction was investigated by in vivo gene expression assays, demonstrating that AidB can repress its own synthesis during normal cell growth. We also showed that the domain architecture of AidB is related to the different functions of the protein: the N-terminal region, comprising the first 439 amino acids (AidB “I-III”), possesses FAD-dependent dehydrogenase activity, while its C-terminal domain, corresponding to residues 440 to 541 (AidB “IV”), displays DNA binding activity and can negatively regulate the expression of its own gene in vivo. Our results define a novel role in gene regulation for the AidB protein and underline its multifunctional nature.
Transcription regulation is one of the principal strategies used by bacteria to respond to external stimuli and to adapt to changing environments. Exposure of Escherichia coli to sublethal concentrations of alkylating agents, such as methyl methanesulfonate (MMS), stimulates the expression of four genes, ada, alkB, alkA, and aidB. The activation of these genes is known as the adaptive response and confers increased cellular resistance to the mutagenic and cytotoxic effects of alkylating agents (7, 8, 19, 28). The Ada protein is the key enzyme of this process; Ada acts both as a methyltransferase able to remove methyl groups from damaged DNA and as a transcriptional activator for the adaptive response genes (13, 19, 23). AlkA is a DNA glycosylase that catalyzes the base excision repair of alkylpurines (19). AlkB is an α-ketoglutarate-Fe(II)-dependent DNA dioxygenase that repairs 1-methyladenine and 3-methylcytosine lesions by oxidative demethylation (24). In contrast, a precise role for the product of the aidB gene in DNA repair has not been identified.
AidB is a protein of 541 amino acids (aa) that is related in sequence to the acyl-coenzyme A (CoA) dehydrogenase family (ACADs) (12); this class of enzymes uses flavin adenine dinucleotide (FAD) to catalyze the α,β-dehydrogenation of acyl-CoA conjugates (20). AidB has been reported to show weak isovaleryl-CoA dehydrogenase (IVD) activity (12, 20) and to exhibit nonspecific DNA binding activity (20). The crystal structure of AidB reveals a unique quaternary organization, a peculiar FAD active site that provides a rationale for its limited dehydrogenase activity, and a putative DNA binding site located in the C-terminal region of the protein (2).
aidB expression is regulated by several different mechanisms. aidB transcription is activated by exposure to acetate under mild acidic conditions, anaerobiosis, or in the presence of high leucine concentrations; under these conditions, aidB expression depends on the rpoS gene, encoding an alternative sigma factor active mainly during the stationary phase of growth (11, 26, 27). Ada-independent induction and IVD activity would suggest that AidB might be a multifunctional protein whose production can respond to different cellular needs.
In this work, we demonstrated that AidB shows preferential binding to a DNA region that includes the upstream sequence and the −35 box of its own promoter, PaidB. The functional significance of this specific binding was investigated by in vivo transcription assays using lacZ as a reporter gene, demonstrating that AidB is able to repress its own synthesis during normal cell growth. We also showed that the domain architecture of AidB is related to the different functions of the protein: the N-terminal region, comprising the first 439 residues (AidB “I-III”), is endowed with dehydrogenase activity, but it is unable to bind DNA. The C-terminal domain, corresponding to residues 440 to 541 (AidB “IV”), was shown to possess DNA binding capability and to function as a transcriptional repressor in vivo. Thus, our results indicate that the AidB protein is a modular transcription factor of its own promoter which requires the short C-terminal region for its regulatory function.
The bacterial strains and plasmids used in this work are listed in Table Table1.1. The ΔaidB35::tetR gene was constructed using the λ red technique as described previously (18). Primers for construction and verification of the mutant strain are listed in Table Table22.
The aidB gene, portion of aidB encoding regions I to III [aidB(I-III)], and portion of aidB encoding region IV [aidB(IV)] of E. coli K-12 were amplified from the bacterial chromosome by PCR using the primers listed in Table Table2.2. pET28a-aidB and pET28a-aidB′I-III' were constructed by cloning the corresponding genes into the pET28a(+) expression vector (Novagen), digested with BamHI and HindIII. To obtain pET28a-aidB′IV', the amplification product aidB(IV) was digested with BamHI and XhoI and cloned into the plasmid pET28a. The resulting expression vectors all contain a 6× histidine tag to allow protein purification by Ni2+ affinity chromatography. Plasmid construction was verified by automated DNA sequencing.
For complementation experiments, the AidB protein or its two mutant forms (AidB I-III and AidB IV) were placed under the control of the lac promoter (Plac). The lac promoter was amplified from genomic DNA of E. coli by PCR using Plac Fw and Plac Rv as primers (Table (Table2).2). The Plac fragment was digested with SphI and BamHI and cloned into the pET28a(+) corresponding sites, generating the vector pET28a-Plac. Then, the aidB, aidB(I-III), and aidB(IV) amplified products were digested with restriction enzymes underlined in Table Table22 and positioned downstream of the lac promoter.
The aidB, aidB(I-III), and aidB(IV) recombinant genes were expressed separately in the E. coli strain C41(DE3) (16). Recombinant cells were grown at 37°C to an optical density at 600 nm of ~0.5, at which time 0.05 mM isopropyl-beta-d-thiogalactopyranoside (IPTG) was added, and the cultures were shifted down to 25°C and allowed to grow until the optical density at 600 nm (OD600) reached 3.0. Cells were harvested by centrifugation at 5,000 × g for 15 min at 4°C, resuspended in 50 mM Na2HPO4 (pH 7.4), disrupted by passage through a French press, and centrifuged at 14,000 × g for 30 min at 4°C.
Recombinant proteins were purified by affinity chromatography on His-Select nickel affinity gel (Sigma). The lysate was loaded onto His-Select nickel affinity gel and equilibrated with equilibration buffer (50 mM Na2HPO4 pH 7.4, 0.3 M NaCl, 10 mM imidazole). After 1 min of incubation at 4°C, the matrix was collected by centrifugation at 11,000 × g for 1 min and washed 3 times with the same equilibration buffer. The recombinant proteins were eluted with buffer containing 250 mM imidazole in 50 mM Na2HPO4 (pH 7.4)-0.3 M NaCl. The protein concentration was estimated using the Bradford reagent (3) (Bio-Rad protein assay), and protein content was checked by SDS-PAGE.
Electrophoretic mobility shift assays (EMSA) were performed using PaidB as a biotin-labeled DNA probe. Sense and antisense oligonucleotides (Table (Table2)2) were annealed by incubation at 95°C for 5 min and successive gradual cooling to room temperature. Purified recombinant AidB, AidB I-III, and AidB IV were incubated with 20 ng of biotinylated DNA (UP35 PaidB) for 20 min at room temperature in 20 μl of buffer Z (25 mM HEPES, pH 7.6, 50 mM KCl, 12.5 mM MgCl2, 1 mM dithothreitol [DTT], 20% glycerol, 0.1% Triton). Protein-DNA complexes were separated on an 8% native polyacrylamide gel (29:1 cross-linking ratio) in 0.5× TBE (45 mM Tris, pH 8.0, 45 mM boric acid, 1 mM EDTA) at 200 V (20 V/cm) at room temperature. Afterwards, electrophoretic transfer to a nylon membrane was carried out in 0.5× TBE at 380 mA for 45 min, and the transferred DNA was cross-linked to the membrane with UV light. After incubation in blocking buffer for 1 h at room temperature, the membrane was incubated with streptavidin-horseradish peroxidase (HRP) conjugate (Sigma) for 30 min at room temperature. The membrane was washed and visualized with SuperSignal chemiluminescence reagent (Pierce).
Competition experiments were performed using increasing quantities (100× to 500×) of either unlabeled UP35 PaidB (used as a specific competitor) or PleuA (used as a nonspecific competitor).
In vitro transcription experiments were performed on two different DNA templates: plasmid pSL101, containing the wild-type aidB promoter, and plasmid pSL112, carrying a mutant aidB promoter with a C-T substitution at position −12. Plasmids (5 nM) and RNA polymerase holoenzyme (120 nM) (Epicentre Technology) were incubated for 20 min at 37°C in 40 mM HEPES, pH 8.0, 10 mM magnesium chloride, 200 mM potassium glutamate, 4 mM dithiothreitol, and 100 μg/ml bovine serum albumin in the absence and the presence of increasing quantities of the AidB protein (1.6 to 6.7 pmol). The elongation step was started by the addition of a prewarmed mixture containing nucleotides and heparin (final concentrations were 500 μM ATP, GTP, and CTP, 30 μM UTP, 1 μCi of [α32P]UTP, and 500 μg/ml heparin) to the template-polymerase mix and allowed to proceed for 10 min at 37°C. The reactions were stopped by the addition of 10 mM EDTA, 0.5% bromophenol blue, and 0.025% xylene cyanol. After heating to 65°C, samples were subjected to electrophoresis on a 7% denaturing polyacrylamide gel in 0.5× Tris-borate-EDTA. Transcripts were detected by exposure to X-ray film overnight at −80°C.
MV1161 (wild type) and MV5924 (ΔaidB) E. coli strains were transformed with the reporter plasmid pMV132H carrying the lacZ gene under the control of PaidB. The complementation experiments were performed by transforming MV5924 containing pMV132H with the following constructs: pET28a-Plac-aidB, pET28a-Plac-aidB′I-III', and pET28a-Plac-aidB′IV'. These bacterial cultures, grown overnight in LB medium at 37°C, were diluted 1:100 in fresh medium. Cellular pellets were collected during the exponential growth phase. The cells were resuspended in 50 mM Na2HPO4 (pH 7.4), disrupted by passage through a French press, and centrifuged at 14,000 × g for 30 min at 4°C. The supernatant was collected, and the protein concentration was determined with the Bio-Rad protein assay (3), using bovine serum albumin as a standard. β-Galactosidase activity was determined by measuring o-nitrophenyl-β-d-galactopyranoside (ONPG) hydrolysis, as described by Miller (15).
In vivo transcription from the ada promoter region was measured using both the MV1601 and MV6608 strains; to measure in vivo transcription from the alkA promoter, we used both the MV1571 and MV6607 strains.
Size exclusion chromatography was performed by using a Superdex 200 PC 3.2/30 column (for AidB I-III) and a Superdex 75 PC 3.2/30 column (for AidB IV) (GE Healthcare) equilibrated in buffer containing 50 mM Tris-HCl, 150 mM NaCl, pH 8. The molecular masses of the native proteins were estimated by comparing their retention times to those of molecular mass standards (thyroglobulin, 670,000 Da; bovine γ-globulin, 158,000 Da; chicken ovalbumin, 44,000 Da; equine myoglobin, 17,000 Da; and vitamin B12, 1,350 Da; Bio-Rad).
Isovaleryl-CoA dehydrogenase activity assays were carried out at room temperature in 200 mM phosphate buffer, pH 8.0, and using purified recombinant proteins that were dialyzed to remove imidazole. For routine assays, 2 mM isovaleryl-CoA (Sigma) was used as the substrate and 0.1 mM 2,6-dichlorophenolindophenol (DCPIP) was used as the terminal electron acceptor. The change in absorbance at 600 nm was monitored by using a Beckman DU 7500 spectrophotometer, and the enzyme activity was calculated by assuming an extinction coefficient of 20.6 mM−1 cm−1 for DCPIP (4).
Recent studies demonstrated that AidB has DNA binding activity (20). In this study, we examined if AidB has any DNA binding preference for specific sequences and found that it preferentially binds a DNA fragment containing the upstream region of its own promoter, PaidB. The aidB gene was cloned into a commercial expression vector of the pET series, and the recombinant protein was expressed as a chimeric protein bearing a 6-His tag. The expressed protein was then purified by affinity chromatography on Ni2+-agarose beads, and its homogeneity was tested by SDS-PAGE and mass fingerprinting analyses. AidB binding capabilities were investigated by electrophoresis mobility shift assay (EMSA) experiments using a biotin-labeled DNA fragment corresponding to the upstream (UP) region and the −35 box of the aidB promoter (UP35 PaidB) (Table (Table22).
The AidB preference for binding to the UP35 PaidB probe was tested by competition experiments: the AidB protein was incubated with the biotinylated UP35 DNA fragment alone or with a 100- or 500-fold excess of either unlabeled UP35 PaidB, a specific competitor, or PleuA, a promoter lacking the UP element used as a nonspecific competitor. Figure Figure11 shows that binding of AidB to labeled UP35 PaidB was indeed reversed when either a 100-fold or 500-fold excess of unlabeled UP35 PaidB was added (lanes 2 and 3), while it was not affected by competition with identical concentrations of PleuA (Lanes 4 and 5). Besides the PleuA-specific promoter, the EMSA competition assays were also performed by using a random DNA sequence with the same nucleotide content as UP35 PaidB. We observed that binding of AidB to labeled UP35 PaidB was not affected by competition with increasing quantities (up to 500-fold) of unlabeled random sequence in a manner similar to that seen by using PleuA (data not shown). Altogether, these results demonstrate that although AidB has been reported to exhibit nonspecific DNA binding activity, it shows preferential binding to a region that includes its own promoter and upstream sequences from −60 to −35 bp.
In order to determine whether the specific binding of the AidB protein to PaidB might be of any biological relevance, we next examined if AidB affects its own expression by in vivo β-galactosidase reporter assays. For these experiments, both MV1161 (wild type) and the aidB deletion mutant MV5924 (ΔaidB) E. coli strains (Table (Table1)1) were transformed with the pMV132H plasmid carrying the gene coding for the β-galactosidase under the control of the aidB promoter. These strains were grown in LB medium, and the β-galactosidase activity was monitored during exponential growth phase. Figure Figure22 demonstrates that lacZ expression from the aidB promoter was increased in the aidB deletion strain (13-fold higher) over that in the wild type, suggesting that transcription from the aidB promoter was repressed by AidB. We reintroduced AidB into the deletion mutant using an expression plasmid, and this dramatically reduced the expression of β-galactosidase to levels as low as those seen in the wild type, thus confirming that AidB has an autoregulatory function and that it represses its own synthesis during normal cell growth.
To further investigate the effect of AidB on gene expression, we tested the effects of the AidB protein in an in vitro expression experiment using E. coli RNA polymerase holoenzyme. Efficient transcription from the aidB promoter can occur only in the presence of the activator protein Ada (13); however, a C-to-T substitution at position −12 (C12T mutation) allows recognition of the promoter by RNA polymerase even in the absence of additional protein factors (9) (Fig. (Fig.3).3). Thus, we carried out in vitro transcription assays in the presence of the AidB protein using the C12T mutant promoter.
The results of these experiments are shown in Fig. Fig.33 and demonstrate that increasing concentrations of AidB inhibit the production of the aidB transcript by RNA polymerase in vitro. Transcripts from the aidB promoter are indicated by the black arrow: as expected, the aidB transcript was rather weak when the wild-type promoter was used (lane 1), while a strong transcript band was observed using the C12T mutant PaidB (lane 2). However, clear repression of transcription from the C12T mutant could be observed in the presence of increasing quantities (1.6 to 6.7 pmol) of the AidB protein (Fig. (Fig.3,3, lanes 3 to 6). In contrast, transcription from the RNA I promoter, used as an internal control (indicated by the white arrow), was clearly unaffected by the AidB protein, thus further substantiating AidB's function as an autoregulatory repressor.
To determine if this repressor activity is aidB specific, or if AidB dampens expression of all adaptive response genes, we examined if the presence or absence of AidB affected expression from either the ada or alkA promoter (Pada or PalkA) when these are fused to lacZ. These promoters possess different structures, since Pada, like PaidB, is characterized by the presence of an UP element, which is missing in the alkA promoter (10, 13). In vivo transcription from the ada promoter region was measured using both the MV1601 strain, carrying a lacZ transcriptional insertion within the chromosomal alkB gene, and the MV6608 strain, harboring the aidB mutation in the alkB::lacZ background. To measure in vivo transcription from the alkA promoter, we used the MV1571 strain, which has a chromosomal lacZ operon fused to the alkA promoter, and its aidB mutant derivative, MV6607. Figure Figure44 indicates that expression of these two adaptive response genes, driven by either Pada or PalkA, is not affected by the presence or absence of AidB. This finding demonstrates that the AidB autoregulatory function is specific only for its own synthesis and does not extend to other adaptive response genes.
Taken together, these data demonstrated that AidB is a transcriptional regulator specific for repressing transcription from its own promoter.
The AidB protein structure suggests that it is organized in four domains: domains I to III, encompassing the first 439 residues, comprise the acyl-CoA dehydrogenase activity, and domain IV comprises the DNA binding activity (2). In order to determine whether the entire protein is required for the repressor activity or if the acyl-CoA dehydrogenase domain is dispensable, we constructed plasmids that express only the dehydrogenase domain or the DNA binding domain and tested them for AidB regulatory activity in vivo. To this aim, we expressed and purified both the protein containing domains I to III (AidB I-III) and the protein containing only domain IV (AidB IV).
The two mutant proteins were then characterized by size exclusion chromatography. On the basis of its retention time and comparison to standard proteins, AidB I-III was estimated to have a molecular mass of 196 kDa; given the predicted mass value of this domain (about 50 kDa), the apparent molecular mass of 196 kDa would suggest that the AidB I-III assembles in a tetramer. In contrast, AidB IV appears to be monomeric.
Functional characterization of the two domains was performed by examining their DNA binding and enzymatic capabilities. To investigate the catalytic properties, the isovaleryl-CoA dehydrogenase (IVD) activities of AidB I-III and AidB IV were examined and compared with that of the full-length protein. Deletion of the C-terminal DNA binding domain did not impair the weak isovaleryl-CoA dehydrogenase activity of the protein (0.12 ± 0 μmol min−1 [mg protein]−1 [average of 10 measurements] for both the wild type and AidB I-III). As expected, the construct that expresses only the DNA binding domain (AidB I-III) showed no IVD activity. These findings clearly indicated that AidB I-III is responsible for its catalytic activity, as predicted by structural analyses (2).
To characterize the DNA binding properties of AidB, the two domains were assayed in vitro for DNA binding activity by gel retardation assays using biotin-labeled UP35 PaidB (Table (Table2).2). Figure Figure5A5A shows that AidB I-III has no aidB promoter binding activity (lane 2) while AidB IV is still capable of binding the aidB promoter despite the lack of domains I to III (lane 4). These experiments clearly demonstrate that domain IV alone is sufficient for the DNA binding activity of AidB.
We next expressed these two forms of AidB in a strain containing the PaidB-lacZ fusion. The ΔaidB strain was transformed with the reporter pMV132H plasmid together with either pET28a-Plac-aidB′I-III' or pET28a-Plac-aidB′IV'. E. coli cells were grown in LB medium, and the β-galactosidase activity was measured and compared with that detected in the ΔaidB strain and in the ΔaidB strain supplemented with the pET28a plasmid carrying the aidB gene. Figure Figure5B5B shows the potential regulatory effects of these two proteins and compares PaidB-lacZ expression with that seen in the wild type and the aidB deletion strain. As was seen before, introduction of full-length AidB on a plasmid reduced expression in the aidB deletion mutant to levels comparable to that seen in wild type. Introduction of AidB I-III into the aidB deletion strain resulted in expression similar to that seen with the aidB deletion itself, indicating that this form of AidB lacked any detectable autoregulatory activity. In contrast, introduction of the AidB IV form of the protein fully repressed lacZ expression from the aidB promoter, indicating that domain IV is necessary and sufficient for this autoregulatory activity.
The AidB protein belongs to the adaptive response to alkylating agents (25), which is defined by the Ada regulon. Upon exposure to alkylating agents, the Ada protein transfers methyl groups from damaged DNA to itself. Once methylated, Ada becomes able to activate the expression of the ada, alkA, alkB, and aidB genes, which confer cellular resistance to the mutagenic and cytotoxic effects of these agents (7, 8, 28).
While the biological roles of Ada, AlkA, and AlkB have been defined (21, 22), little is known of the biological function of AidB in vivo. AidB was reported to possess weak isovaleryl-coenzyme A dehydrogenase activity (12, 20) and to be a nonspecific DNA binding protein (20). The ability to bind DNA both specifically and nonspecifically is a similarity to the Ada protein, which binds DNA nonspecifically in its unmethylated form and becomes able to recognize specific target sequences upon self-methylation (23). However, in the case of AidB protein, there is no indication that an interaction with damage or a damaging agent is required for the transition between specific and nonspecific binding. Instead, this is an intrinsic property of the unmodified AidB protein, since protein isolated from untreated cells is capable of both specific and nonspecific DNA binding (Fig. (Fig.1)1) (20). In this respect, the sequence specificity of AidB binding to DNA deserves some comments. Results presented in this paper demonstrate that AidB is able to regulate expression of its own promoter but not the ada promoter, which also contains an A/T-rich region. Further investigations are needed to fully clarify AidB binding specificity, including functional experiments with other promoters containing or not containing UP elements. Furthermore, chromatin immunoprecipitation (ChIP)-chip experiments might also be designed in order to determine other potential regulatory targets.
Induction of a regulatory element that counteracts induction appears to be a common feature of other stress responses. When E. coli cells are stressed by heat to cause induction of heat shock genes, the normal sigma factor, sigma 70, is induced. It has been suggested that this increased level of sigma 70 may compete with sigma 32 for RNA polymerase binding (5), thereby contributing to the characteristic heat shock gene expression pattern in which a strong, immediate induction of heat shock genes is followed by a dampening of their expression, reducing expression to an intermediate level (6). Induction of the SOS response results in induction of lexA, which encodes the LexA repressor of SOS genes. Since LexA is cleaved and inactivated by an activated form of the RecA protein, it is thought that the elevated levels of the LexA protein result in a rapid restoration of repression as soon as LexA protein production exceeds its cleavage, thereby allowing intact, functional LexA to accumulate (14).
The biological relevance of AidB autorepression is not as universal as the repression functions triggered as part of either heat shock or the SOS response, since it affects only aidB gene expression. In the absence of alkylation, this autorepression clearly functions to further reduce the level of expression (Fig. (Fig.2).2). The need for cells to regulate the basal expression of AidB by the additional autoregulatory mechanism suggests that expression of AidB may have deleterious consequences that will require it to be kept at a minimum in normal, actively growing cells. The determination of why cells require such a complex series of regulatory mechanisms to control aidB expression will clearly require a determination of its exact function in response to alkylation damage.
Finally, we investigated the domain architecture of AidB by expressing and functionally characterizing the N- and C-terminal domains of the protein containing domains I to III and domain IV, respectively. The N-terminal region encompassing the first 439 residues (AidB I-III) was shown to exhibit the same levels of isovaleryl-CoA dehydrogenase (IVD) activity as the full-length protein.
However, the level of IVD activity observed in AidB is quite low compared to that in other acyl-CoA dehydrogenases. Human isovaleryl-CoA dehydrogenase exhibits a specific activity of 8.2 to 11.7 μmol min−1 (mg protein)−1 (1, 17), about 2 orders of magnitude higher than that of AidB. Recent structural studies revealed several unique features that distinguish AidB's FAD cavity from the ACAD active sites despite the conservation of their general properties (2). These observations provided a rationale for AidB's limited acyl-CoA dehydrogenase activity (20), suggesting that fatty acyl-CoAs are too large to fit the AidB active site and cannot in fact be the main substrates of this enzyme. The crystal structural analysis of AidB also identified a putative DNA binding site located in its C-terminal region. We found that the C-terminal domain alone (aa 440 to 541) (AidB IV) possesses DNA binding activity and can negatively regulate the expression of the aidB gene at the same level as the full-length protein.
Thus, acyl-CoA dehydrogenase activity does not appear to be involved in modulation of the DNA binding properties of the AidB protein. It is possible, however, that DNA binding might affect the enzymatic activity of the AidB protein, allowing AidB to carry out repair activity. Although this event might need a large conformational change according to the three-dimensional structure of the protein, a similar mechanism of action has been described for the AlkB protein and for its human homologues ABH2 and ABH3. These proteins are able to repair methylated bases by oxidative demethylation upon DNA binding.
This work was supported by grants from the Ministero dell'Università e della Ricerca Scientifica (Progetti di Rilevante Interesse Nazionale 2005, 2006; FIRB Rete Nazionale di Proteomica, RBRN07BMCT). Support from the National Center of Excellence in Molecular Medicine (MIUR, Rome), from the Regional Center of Competence (CRdC ATIBB, Regione Campania, Naples), and from NIH grant CA100122 to M.V. is gratefully acknowledged.
Published ahead of print on 1 October 2010.