|Home | About | Journals | Submit | Contact Us | Français|
The Escherichia coli LexA protein was used as a query sequence in TBLASTN searches to identify the lexA gene of the δ-proteobacterium Geobacter sulfurreducens from its genome sequence. The results of the search indicated that G. sulfurreducens has two independent lexA genes designated lexA1 and lexA2. A copy of a dinB gene homologue, which in E. coli encodes DNA polymerase IV, is present downstream of each lexA gene. Reverse transcription-PCR analyses demonstrated that, in both cases, lexA and dinB constitute a single transcriptional unit. Electrophoretic mobility shift assays with purified LexA1 and LexA2 proteins have shown that both proteins bind the imperfect palindrome GGTTN2CN4GN3ACC found in the promoter region of both lexA1 and lexA2. This sequence is also present upstream of the Geobacter metallireducens lexA gene, indicating that it is the LexA box of this bacterial genus. This palindrome is not found upstream of either the G. sulfurreducens or the G. metallireducens recA genes. Furthermore, DNA damage induces expression of the lexA-dinB transcriptional unit but not that of the recA gene. However, the basal level of recA gene expression is dramatically higher than that of the lexA gene. Likewise, the promoters of the G. sulfurreducens recN, ruvAB, ssb, umuDC, uvrA, and uvrB genes do not contain the LexA box and are not likely to bind to the LexA1 or LexA2 proteins. G. sulfurreducens is the first bacterial species harboring a lexA gene for which a constitutive expression of its recA gene has been described.
Geobacter sulfurreducens is an Fe(III)-reducing microorganism belonging to the δ subdivision of Proteobacteria that is able to completely oxidize organic compounds to carbon dioxide with fumarate, Fe(III), elemental sulfur, or malate as terminal electron acceptors (16). Although this organism was previously classified as strict anaerobe, recent evidence indicates that it can tolerate oxygen exposure. It has recently been demonstrated that a brief exposure to oxygen can generate reactive oxygen species (O2− and H2O2) in anaerobic bacteria (28). Furthermore, it is known that Fe(II) in combination with these oxygen species can result in DNA damage via the Fenton reaction (1, 9, 12, 18). These iron-mediated DNA damage-induced lesions seem to be repaired mainly by recombination (27, 32). Even in the absence of oxygen, other intracellular reductants, such as NADH and glutathione, can damage DNA in vitro and in vivo in the presence of iron (8, 23, 35).
Bacterial cells contain several DNA repair pathways, with those of Escherichia coli being the most studied. Among them, the SOS network is probably the most versatile DNA repair system. The E. coli SOS system contains about 40 genes that display several cellular activities, including recombination increase, error-prone DNA replication, inhibition of cell division, and prophage induction (5, 7, 34). These genes are regulated by the recA and lexA products, which also belong to this regulon (14). The E. coli LexA protein specifically binds the consensus sequence CTGTN8ACAG, called the E. coli SOS box (34). When the DNA is not injured, LexA blocks SOS gene transcription. However, when DNA damage is present, the RecA protein acquires an active conformation and binds single-stranded DNA regions produced by DNA damage-mediated inhibition of replication (26). After activation, RecA promotes the autocatalytic cleavage of LexA at the Ala84-Gly85 bond with the participation of its own Ser119 and Lys156 residues (13, 17). The LexA cleavage, whose mechanism is very similar to that of serine proteases, triggers the induction of the SOS genes. After repair of DNA, the single-stranded regions disappear and the RecA protein is no longer activated resulting in an increase in the LexA repressor level which in turn inhibits SOS gene expression.
The E. coli SOS box is also present in the promoter of DNA damage-inducible genes in many members of γ-Proteobacteria, including those belonging to the Enterobacteriaceae, Pseudomonadaceae, Vibrionaceae, and Pasteurellaceae families, among others. Nevertheless, in other γ-Proteobacteria, such as Xylella fastidiosa and Xanthomonas campestris, the LexA-binding sequence (TTAGN6TACTA) is different from that of E. coli (3). This fact indicates that a significant heterogeneity in the LexA-binding sequence exists in the γ class. In other bacterial phylogenetic groups, such as the gram-positive and α-class Proteobacteria, the sequence of the LexA-binding region also differs (CGAACRNRYGTTYC and GTTCN7GTTC, respectively) (29, 35).
Despite the fact that G. sulfurreducens lives under conditions likely to promote DNA damage, nothing is known about the DNA repair mechanisms of this microorganism. In fact, no information is available concerning the DNA repair system in any of the δ-Proteobacteria. For this reason and in order to take advantage of the fact that the G. sulfurreducens genome is being sequenced, the lexA gene of this organism was cloned, its DNA-binding sequence was identified, and the behaviors of several DNA repair-related genes were analyzed.
Bacterial strains and plasmids used in this work are listed in Table Table1.1. E. coli was grown at 37°C in Luria-Bertani medium (24). Antibiotics were added to the cultures at previously reported concentrations (24). G. sulfurreducens was cultured at 30°C in NBAF medium under strict anaerobic conditions as described previously (4), and E. coli cells were transformed with plasmid DNA as described previously (24). All restriction enzymes, PCR-oligonucleotide primers, T4 DNA ligase and polymerase, and the digoxigenin (DIG)-DNA labeling and detection kit were from Roche (Mannheim, Germany). Total G. sulfurreducens DNA was prepared by using the MasterPure complete DNA purification kit (Epicentre, Madison, Wis.).
The synthetic oligonucleotide primers used for PCR amplification are listed in Table Table2.2. To facilitate subcloning of some PCR-DNA fragments, specific restriction sites (identified in Table Table2)2) were incorporated into the oligonucleotide primers. Mutations were introduced into the G. sulfurreducens lexA1 promoter by PCR mutagenesis with oligonucleotides carrying designed substitutions (Table (Table2).2). The DNA sequence of all PCR-mutagenized fragments was determined by the dideoxy method (25) on an ALF Sequencer (Pharmacia Biotech, Uppsala, Sweden). In all cases, both DNA strands were sequenced. Southern analysis was performed as previously described (30) with an internal 513-bp fragment of the G. sulfurreducens recA gene obtained with the primers indicated in Table Table22.
The G. sulfurreducens lexA1 and lexA2 genes were identified by a TBLASTN search of its preliminary genome sequence (http://www.tigr.org) by using the E. coli LexA protein as a probe. Two independent regions of significant homology were detected, indicating the presence of two different lexA genes. In order to amplify these two genes, primers based upon their flanking regions were designed (Table (Table2).2). The resulting DNA fragments (1,075 and 710 bp for lexA1 and lexA2, respectively) were then cloned into pGEM-T to yield plasmids pUA1021 and pUA1022. Both inserts contained in these plasmids were sequenced to confirm that no mutation had been introduced during the PCR.
The plasmids pUA1025 and pUA1026 were constructed in order to create and express Glutathione S-transferase (GST)-LexA1 and -LexA2 fusion proteins. The first step in the construction of these plasmids was to amplify the two lexA genes from plasmids pUA1021 and pUA1022 by using the primers LexA1EcoRI, LexA1SalI, LexA2EcoRI, and LexA2SalI. The resulting DNA fragments were cloned into pGEM-T to produce pUA1023 and pUA1024. After excision with EcoRI and SalI, both lexA genes were inserted into the pGEX4T1 expression vector (Amersham-Pharmacia, Uppsala, Sweden) immediately downstream of the T7 promoter. The initiation codon of the LexA1 and LexA2 proteins was placed immediately downstream of the EcoRI sites in the LexA1EcoRI and LexA2EcoRI primers, such that the lexA1 and lexA2 genes could be fused to GST in frame. The inserts of pUA1025 and pUA1026 were sequenced in order to ensure that no mutations were introduced during amplification.
To overproduce the LexA-GST fusion protein, pUA1025 and pUA1026 plasmids were transformed into E. coli BL21-CodonPlus (Stratagene, La Jolla, Calif.). The two resulting BL21-CodonPlus strains were diluted in 0.5 liter of Luria-Bertani medium and incubated at 37°C until they reached an optical density at 600 nm (OD600) of 0.8. Fusion protein expression was induced at this time by the addition of IPTG (isopropyl-β-d-thiogalactopyranoside) to a final concentration of 1 mM. After incubation for an additional 3 h at 37°C, cells were collected by centrifugation for 15 min at 3,000 × g. The bacterial pellet was resuspended in phosphate-buffered saline (PBS; 10 mM Na2HPO4, 1.7 mM KH2PO4, 140 mM NaCl, 2.7 mM KCl [pH 7.4]) containing Complete Mini protease inhibitor cocktail (Roche). The resulting cell suspensions were lysed by sonication. Unbroken cells and debris were removed by centrifugation for 20 min at 19,000 × g. The supernatants containing the GST-LexA1 and GST-LexA2 fusion proteins were incubated with PBS-glutathione-Sepharose 4B beads (Amersham-Pharmacia) for 2 h at 4°C in order to affinity purify the fusion protein. The beads were then washed twice with PBS containing 0.1% Triton and three times with PBS without detergent.
The sequence Leu-Val-Pro-Arg-Gly-Ser is located immediately downstream of the GST coding sequence in the pGEX4T vector series and serves as a linker between the LexA and GST moieties of the fusion proteins. This hexapeptide is recognized by the thrombin protease, which cleaves at the Arg-Gly bond. It was therefore possible to release the G. sulfurreducens LexA proteins from the Sepharose beads by incubating a 700-ml bed volume of beads with 25 U of thrombin (Amersham-Pharmacia) in 1 ml of PBS. The supernatants containing the G. sulfurreducens LexA1 and LexA2 proteins with an additional five-amino-acid tail at their N termini (Gly-Ser-Pro-Glu-Phe) were visualized in a Coomassie blue-stained sodium dodecyl sulfate-13% polyacrylamide gel electrophoresis gel (11). Their purity was >95% (Fig. (Fig.11).
G. sulfurreducens LexA1 and LexA2-DNA complexes were detected by electrophoresis mobility shift assays (EMSAs) with the purified G. sulfurreducens LexA1 and LexA2 proteins (Fig. (Fig.1).1). DNA probes were prepared by PCR amplification with one primer labeled at its 5′ end with DIG (Table (Table2).2). Probes were purified in 2 to 3% low-melting-point agarose gels depending upon the size. DNA-protein-binding reactions (20 μl), typically containing 10 ng of DIG-DNA-labeled probe and 40 ng of purified G. sulfurreducens LexA1 or LexA2 protein, were incubated in binding buffer (10 mM HEPES-NaOH [pH 8], 10 mM Tris-HCl [pH 8], 5% glycerol, 50 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 20 μg of salmon DNA/ml, 50 μg of bovine serum albumin/ml) for 30 min at 30°C and loaded onto a 5% nondenaturing Tris-glycine polyacrylamide gel (prerun for 30 min at 10 V/cm in 25 mM Tris-HCl [pH 8.5]-250 mM glycine-1 mM EDTA). DNA-protein complexes were separated at 150 V for 1 h, followed by transfer to a Biodine B nylon membrane (Pall Gelman Laboratory). DIG-labeled DNA-protein complexes were detected by following the manufacturer's protocol (Roche). For the binding-competition experiments, a 300-fold molar excess of either specific or nonspecific unlabeled competitor DNA was also included in the mixture. All EMSAs were repeated a minimum of three times to ensure the reproducibility of the results.
The DNA damage response was induced by mitomycin C (Sigma Chemical Co., St. Louis, Mo.) as previously described (3). In brief, mitomycin C was added to mid-log-phase NBAF cultures (OD600 of 0.3) of G. sulfurreducens at a final concentration of 0.8 μg/ml. After 1.5 h of incubation at 30°C the cells were harvested, and the RNA was extracted as described below.
To determine the transcriptional organization of the G. sulfurreducens lexA and dinB genes, reverse transcriptase (Roche) was used to generate cDNA by reverse transcription (RT)-PCR with total RNA from G. sulfurreducens as a template and the pairs of primers indicated in Table Table2.2. These primers were designed to amplify a PCR product of either 736- or 947-bp if lexA1 and dinB1 or lexA2 and dinB2 genes constituted a single transcription unit, respectively.
Total RNA from G. sulfurreducens was prepared from mid-log-phase (OD600 of 0.3) cells grown in NBAF medium (4). After collection of the cells by centrifugation at 4°C for 10 min at 4,300 × g, the RNA was purified by using the RNeasy Midi Kit (Qiagen, Valencia, Calif.). The RNA was treated with RNase-free DNase I (Roche) to ensure the absence of contaminating DNA. RNA concentration and its integrity were determined by A260 measurements and 1% formaldehyde-agarose gel electrophoresis, respectively (24). In all RT-PCR experiments, the absence of contaminating DNA in RNA samples after RNase-free DNase I (Roche) treatment was confirmed by carrying out PCR amplification without the addition of reverse transcriptase.
Mitomycin C-mediated induction of several genes studied here was carried out by real-time quantitative RT-PCR analysis of total G. sulfurreducens RNA with the LightCycler apparatus (Roche) by using the LC-RNA Master SYBR Green I kit (Roche) and the primers indicated in Table Table22 according to the manufacturer's instructions. The concentration of total RNA from both treated and untreated cultures was adjusted to the same value. The amount of mRNA of each gene was determined by using a standard curve generated by the amplification of an internal fragment of the G. sulfurreducens 16S rRNA with the appropriate primers indicated in Table Table2.2. In all determinations, the amount of mRNA of G. sulfurreducens trpA gene, encoding the tryptophan synthase, was also determined. This gene was used as the negative control because it is not directly related to any DNA repair system and its expression is not DNA damage inducible (2, 10).
Genes for two proteins homologues to the LexA of E. coli were identified in the G. sulfurreducens genome when the E. coli LexA protein sequence was used as a query for TBLASTN analysis at the NCBI database (http://www.ncbi.nlm.nih.gov/Microb_blast/unfinishedgenome.html). The two homologs, designated LexA1 and LexA2, are predicted to be 201 and 203 amino acids long, respectively, and both contain the two characteristic domains of LexA repressors: the N-terminal DNA-binding domain harboring a helix-turn-helix motif and the C-terminal protease (13, 17). In addition to the Ala-Gly bond, at which the autocleavage takes place, the conserved Lys and Ser residues involved in hydrolysis are also present (Fig. (Fig.2).2). The two G. sulfurreducens LexA proteins are 68% identical and, curiously, are most closely related to Thermotoga maritima and Bacillus subtilis LexA proteins (43% identity). Not surprisingly, when G. sulfurreducens LexA1 and LexA2 are compared through CLUSTALW analysis (31) to other LexA proteins whose DNA-binding sequences have been determined (B. subtilis, E. coli, Rhodobacter sphaeroides, and X. fastidiosa), the highest degree of divergence is at the N-terminal DNA-binding domain (Fig. (Fig.22).
Downstream from both G. sulfurreducens lexA genes is an open reading frame encoding a protein with significant identity (41%) to E. coli DNA polymerase IV, the product of dinB (33) (Fig. (Fig.3A).3A). The identity between both G. sulfurreducens dinB gene homologues was 64%. The fact that the distances between lexA1 and dinB1 and between lexA2 and dinB2 were very short (50 and 296 bp, respectively) suggested that, in both cases, the lexA and dinB genes were cotranscribed. To evaluate this hypothesis, an RT-PCR analysis of total G. sulfurreducens RNA was carried out with a single set of primers designed to amplify fragments of either 736 or 947 bp (Fig. (Fig.3A)3A) if lexA1-dinB1 or lexA2-dinB2 were a transcriptional unit. The products recovered demonstrated that transcription of each one of the two lexA genes is linked to its respective dinB gene (Fig. (Fig.3B).3B). G. sulfurreducens is the first bacterial species found to have this genetic organization. In E. coli, lexD is downstream of the lexA gene, whereas recA is downstream in X. fastidiosa (3, 19), but dinB does not appear to be immediately downstream of lexA in either organism. A TBLASTN search of the Geobacter metallireducens genome (http://www.jgi.doe.gov) with the G. sulfurreducens LexA1 protein as a query revealed that the dinB gene is also immediately downstream of the lexA gene in this Geobacter species (data not shown). This fact may indicate that the presence of a single transcriptional unit comprising both lexA and dinB genes is a common characteristic of the Geobacter genus. Nevertheless, we were unable to detect a second copy of the lexA-dinB operon in G. metallireducens. The genome of this organism is not complete; thus, it is not yet certain whether a second lexA-dinB operon does not exist or its sequence is not yet available.
The fact that the regions surrounding the G. sulfurreducens lexA1 and lexA2 genes are different (Fig. (Fig.3A)3A) indicates that the origin of these two genes is unlikely to be attributable to a recent single chromosomal duplication. However, the presence of an open reading frame encoding a putative DNA invertase immediately upstream of the lexA1 gene suggests that a genetic rearrangement may have taken place. The G+C content of G. sulfurreducens lexA2 and G. metallireducens lexA, as well as that of their surrounding region, is approximatelly 65%, whereas the G+C content of the lexA1 gene is only 56%. Given this difference, the possibility that the lexA1 gene was incorporated into the G. sulfurreducens genome by horizontal transfer cannot be excluded.
All lexA genes that have been functionally studied so far are autoregulated. To analyze the possibility of LexA autoregulation in G. sulfurreducens, EMSAs with the purified LexA1 protein and the lexA1 promoter as a probe were performed. The addition of LexA1 to a fragment extending from −95 to +48 with respect to its putative translational start point of the lexA1 gene decreased its mobility (Fig. (Fig.4A).4A). This DNA-protein interaction was specific because it was abolished by an excess of unlabeled lexA1 fragment but not when an excess of nonspecific DNA was added (Fig. (Fig.4A).4A). The addition of a 169-bp fragment spanning base pairs from −104 to +65 of the lexA2 gene also abolished LexA1 binding to the LexA1 fragment. Likewise, LexA2 binds specifically to its own lexA2 promoter, as well as to the lexA1 promoter (Fig. (Fig.4B).4B). These data clearly suggest that the sequence recognized by both LexA1 and LexA2 proteins is the same and that it is present in both the lexA1 and the lexA2 promoters.
To precisely define the G. sulfurreducens LexA1- and LexA2-binding sites, serial deletions of the upstream promoter region of the lexA1 gene were generated and analyzed in EMSA experiments with the purified LexA1 protein (Fig. (Fig.5A).5A). The electrophoretic mobilities of LexA1.1, LexA1.2, and LexA1.3 were retarded in the presence of the LexA1 protein, whereas the mobility of LexA1.4 was not affected (Fig. (Fig.5B).5B). These data indicate that the LexA1-binding motif is located between the −34 and −41 positions. A search for potential binding motifs (i.e., either direct or inverted repeats) in this region revealed the presence of the perfect palindrome GGTTGACATATGTCAACC. In fact, the insertion of four cytosines into the middle of this palindrome (between the internal submotifs GACAT and ATGTC) abolished LexA1.3 fragment retardation (Fig. (Fig.5B).5B). Furthermore, a sequence very close to this palindrome is also present upstream of the lexA2 gene (Fig. (Fig.6).6). Then, using the LexA1.3 fragment, point mutations were introduced into each of the nucleotides of this palindrome, which are conserved in both lexA promoters and its surrounding bases, to determine whether they are effectively involved in the LexA1-binding reaction. The data obtained indicate that G. sulfurreducens LexA1, and consequently LexA2, binds the GGTTNNCNNNNGNNNACC motif (Fig. (Fig.6).6). An examination of the upstream region of the G. metallireducens lexA promoter revealed the presence of this sequence. All of these data led us to designate this imperfect palindrome as the LexA box of the Geobacter genus. It must be noted that other DNA-binding sequences of bacterial LexA repressors, such as those of B. subtilis and X. fastidiosa, are also imperfect palindromes in which internal bases are specifically required for LexA binding (3, 36).
The product of the E. coli recA gene is involved in both recombination-mediated and error-prone DNA repair. Furthermore, E. coli RecA is also required for LexA autocleavage after DNA damage. These data indicate that the G. sulfurreducens LexA regulon is different from that of E. coli, at least regarding the genetic organization of the lexA gene. Therefore, regulation of G. sulfurreducens recA gene expression was analyzed. This gene was identified by a TBLASTN search by using the E. coli RecA protein as a query. The G. sulfurreducens RecA protein showed maximal identity to the Myxococcus xanthus RecA2 protein (73%), another member of the δ-Proteobacteria. The G+C content of the G. sulfurreducens recA gene (63%) was similar to that of the lexA2 gene. An analysis of the sequence of the recA upstream region revealed that the imperfect palindrome required for LexA1 and LexA2 binding was not present. In agreement with this, the electrophoretic mobility of the G. sulfurreducens recA promoter was not affected by the presence of either the LexA1 or the LexA2 proteins (Fig. (Fig.7A).7A). Likewise, the addition of the unlabeled recA promoter did not abolish retardation of the lexA1 promoter in the presence of LexA1 protein (Fig. (Fig.7B).7B). Furthermore, the upstream region of the G. metallireducens recA gene also lacks the LexA recognition sequence.
As a consequence of these results, we investigated whether the G. sulfurreducens recA gene is DNA damage inducible. Figure Figure88 shows that recA gene transcription is not stimulated by the DNA-damaging agent mitomycin C, whereas this compound triggers expression of the lexA2-dinB2 transcriptional unit. These data indicate that the lexA2 gene is induced by DNA damage and that both G. sulfurreducens LexA proteins contain all amino acid residues that are involved in the RecA-catalyzed autocleavage process (14). Thus, the absence of DNA damage-mediated expression of the G. sulfurreducens recA gene cannot be attributed to a defect in the process of lexA regulated-gene induction.
It has been reported that some bacteria such as M. xanthus possess two distinct recA genes (20). To determine that this is not the case in G. sulfurreducens, Southern analysis was performed with a highly conserved internal region of the G. sulfurreducens recA gene as a probe. As shown in Fig. Fig.9,9, a single band of the expected size was detected when G. sulfurreducens chromosomal DNA was digested with three restriction enzymes. These data unequivocally confirm that the G. sulfurreducens genome contains only one copy of the recA gene. In agreement with this result, TBLASTN analysis of the G. sulfurreducens sequence data bank revealed the presence of only one RecA-like protein when E. coli RecA was used as a query.
To our knowledge, G. sulfurreducens recA is the first gene from a bacterial species possessing a LexA protein that is not DNA damage inducible. The Porphyromonas gingivalis recA gene is not stimulated by DNA damage (15), but the complete sequence of the genome of this microorganism has revealed that it lacks a lexA-like gene (http://www.tigr.org). It seems likely that the strong DNA-damaging potential of the G. sulfurreducens environment has directed an evolutionary process to attain a constitutive synthesis of RecA, which is probably the most important protein involved in the recombinational DNA repair pathway in bacteria. In fact, the transcriptional level of the G. sulfurreducens recA gene is dramatically higher than that of both the lexA2 and the trpA genes (Fig. (Fig.88).
TBLASTN searches also identified several G. sulfurreducens genes, which in E. coli belong to the LexA regulon (recN, ruvAB, ssb, umuDC, uvrA, and uvrB). However, none of their promoters contained the G. sulfurreducens LexA-binding sequence. Likewise, a search in the G. sulfurreducens sequence database by using the Findpatterns program of the Genetics Computer Group package (6) has not revealed the presence of a LexA-binding sequence upstream of any of the either identified or known open reading frames. Since the sequence of the G. sulfurreducens genome is practically finished, this finding suggests that the gene content of the LexA regulon in this organism is very narrow. In agreement with this, only the single transcriptional unit lexA-recA and a gene encoding a putative DNA modification methylase are directly under the control of the LexA protein in X. fastidiosa (2). The fact that both the LexA box and the composition of the LexA regulon differ greatly in X. fastidiosa, G. sulfurreducens, and E. coli indicates that this network has been subjected to a great deal of variability throughout bacterial evolution. Interestingly, the level of divergence is not so high in other global bacterial regulons such as fur, which controls iron uptake (21, 22). The identification of new LexA-binding sites, as well as the elucidation of LexA regulon gene contents for other phylogenetic bacterial classes, will provide further understanding of its evolution.
This work was funded by grants BMC2001-2065 from the Ministerio de Ciencia y Tecnología (MCyT) de España and 2001SGR-206 from the Departament d'Universitats, Recerca i Societat de la Informació (DURSI) de la Generalitat de Catalunya. A.R.F.D.H. and M.J. were recipients of a postdoctoral contract and a predoctoral fellowship from the MCyT and DURSI, respectively. C.N. and D.R.L. were supported by the Genomes to Life Program, U.S. Department of Energy (grant DE-FC02-02ER63446). In addition, C.N. was also the recipient of a DGAPA/UNAM postdoctoral fellowship.
We are deeply indebted to M. Coppi for help in the revision of the manuscript and to Joan Ruiz and Pilar Cortés for excellent technical assistance.