Identification of a DNA-Damage-Inducible Regulon in Acinetobacter baumannii

Abstract

The transcriptional response of Acinetobacter baumannii, a major cause of nosocomial infections, to the DNA-damaging agent mitomycin C (MMC) was studied using DNA microarray technology. Most of the 39 genes induced by MMC were related to either prophages or encoded proteins involved in DNA repair. Electrophoretic mobility shift assays demonstrated that the product of the A. baumannii MMC-inducible umuD gene (umuDAb) specifically binds to the palindromic sequence TTGAAAATGTAACTTTTTCAA present in its promoter region. Mutations in this palindromic region abolished UmuDAb protein binding. A comparison of the promoter regions of all MMC-induced genes identified four additional transcriptional units with similar palindromic sequences recognized and specifically bound by UmuDAb. Therefore, the UmuDAb regulon consists of at least eight genes encoding seven predicted error-prone DNA polymerase V components and DddR, a protein of unknown function. Expression of these genes was not induced in the MMC-treated recA mutant. Furthermore, inactivation of the umuDAb gene resulted in the deregulation of all DNA-damage-induced genes containing the described palindromic DNA motif. Together, these findings suggest that UmuDAb is a direct regulator of the DNA damage response in A. baumannii.

INTRODUCTION

In Escherichia coli and many other bacterial species, extensive DNA damage that cannot be repaired by other cellular mechanisms induces a mutagenic repair pathway known as the SOS response (1). As part of the SOS response, the RecA protein binds to the single-stranded region of the damaged DNA, where it is activated and forms a nucleoprotein filament (2). Activated RecA induces the autoprotease activity of LexA, which in the absence of activated RecA represses the expression of SOS genes by binding to specific sequences in their promoters. As cleaved LexA is unable to bind to these regulator sequences, the SOS genes are derepressed (1). The proteins encoded by SOS genes mediate DNA repair and replication in addition to pausing the cell cycle. The activities in the SOS response include those of highly error-prone polymerases (1). Once the damaged DNA is repaired, RecA no longer induces the autocleavage of LexA, which then accumulates and shuts down the SOS response (3).

The SOS genetic network is widely present in Eubacteria (4), but in Acinetobacter spp. the DNA damage response is characterized by several atypical features: (i) there is no damage-induced mutagenesis response to DNA damage, with the remarkable exception of the opportunistic pathogens Acinetobacter baumannii, Acinetobacter ursingii, and Acinetobacter beijerinckii (5–7); (ii) after DNA damage, further induction of recA does not require the RecA protein (8); (iii) none of the promoters of the DNA-damage-inducible genes of Acinetobacter spp. contain a known SOS box (7, 9, 10); (iv) a canonical LexA homologue has not been identified (8, 10, 11); and (v) a UmuD homologue in A. baumannii (UmuDAb) has been proposed as a putative indirect regulator of the DNA damage response (6). Given the clinical importance of A. baumannii, a major cause of nosocomial infections, in the present work we used DNA microarray technology to study the response of this bacterium to the DNA-damaging agent mitomycin C (MMC).

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Wild-type (WT) A. baumannii ATCC strain 17978 and its derivative recA mutant strain (12) were grown in Luria Bertani (LB) medium and incubated at 37°C with shaking at 180 rpm. For MMC treatment, an overnight culture of the corresponding strain in LB medium was diluted 1:25 in fresh medium. At the mid-exponential growth phase (optical density at 600 nm [OD₆₀₀] of 0.6), the cultures were treated with MMC (final concentration of 0.5 μg ml⁻¹) and incubated for an additional 2 h before samples were removed for RNA extraction. The corresponding controls were carried out in the same way but without the addition of MMC.

Inactivation of the A. baumannii umuD gene.

Plasmid pCR-BluntII-TOPO (Life Technologies), unable to replicate in A. baumannii, was used as a suicide vector as previously described (13). Briefly, an internal fragment of the umuDAb gene, generated by PCR using the oligonucleotides umuDAbintUP and umuDAbintRV (5′-GTTCCTGAATCTGAAGTC and 5′-GTGCAATCACAATATCGC, respectively), was inserted into pCR-BluntII-TOPO using a Zero Blunt TOPO PCR cloning kit (Life Technologies), yielding plasmid pTOPO-umuDAbint. This plasmid was introduced into kanamycin-susceptible A. baumannii ATCC strain 17978 by electrotransformation. A. baumannii ATCC 17978 (pTOPO-umuDAbint) was selected on plates containing kanamycin (50 μg ml⁻¹). Inactivation of umuDAb by insertion of pTOPO-umuDAbint was confirmed by sequencing the obtained PCR product using primers M13 reverse (5′-GCGGATAACAATTTCACACAGG) and umuDAbextUP (5′-CGAGCATGGCGGTGCACG).

RNA extraction.

Total RNA was extracted with a High Pure RNA isolation kit (Roche) according to the manufacturer's instructions. Samples were treated with DNase I (Life Technologies) and purified using an RNeasy MinElute cleanup kit (Qiagen). RNA samples were tested by PCR without reverse transcriptase, which confirmed the absence of DNA. The quality of the RNA was determined using a Bioanalyzer instrument (Bioarray, Alicante, Spain).

Microarrays.

The microarrays were designed specifically for A. baumannii strain 17978 by Bioarray Diagnóstico Genético (Alicante, Spain) using eArray (Agilent). Labeling was carried out following two-color microarray-based prokaryote analysis using Fair Play III labeling, version 1.3 (Agilent). Three independent RNA extractions per condition (biological replicates) were used for each experiment. Statistical analysis was carried out using Bioconductor, in the software package RankProd for the R computing environment. A gene was considered to be induced when the ratio of the treated to the nontreated preparation was ≥1.5 and the P value was <0.05.

RT-qPCR.

Gene expression was analyzed by reverse transcription-quantitative real-time PCR (RT-qPCR) as reported previously (14). The specific internal oligonucleotides used for each gene are listed in Table S1 in the supplemental material.

Prediction of DNA-binding motifs.

The predicted promoter-containing regions (300 bp immediately upstream of the start codon) of the genes upregulated in the MMC-treated WT strain were analyzed for conserved motifs using the program Multiple Em for Motif Elicitation (MEME) (15).

Protein techniques.

The umuDAb gene was PCR amplified from the purified genomic DNA of A. baumannii strain ATCC 17978, using the primers UmuDAbXhoI (5′-ATCGCTCGAGATGCCAAAGAAGAAAGAA) and UmuDAbBamHI (5′-AGGGATCCTTATCTCATTCGTTTGAG). The purified PCR products were then digested with XhoI and BamHI, cloned into the appropriate restriction sites in the polylinker of the pET15b expression vector (Novagen), and introduced by transformation into E. coli DH5α cells (Clontech). Recombinant plasmids were purified, sequenced (Macrogen), and used to transform strain BL21-CodonPlus(DE3)-RIL (Stratagene). Protein expression, purification, and analysis were carried out as described previously (16).

For electrophoretic mobility shift assays (EMSAs), oligonucleotide (100 bases each) pairs including the putative binding regions were designed with the desired nucleotide changes (when necessary). An A residue was added to both 3′ ends (see Table S2 in the supplemental material) to facilitate cloning of the corresponding pairs into the pGEM-T vector (Promega). DNA probes were prepared by PCR amplification using the universal forward (5′-GTAAAACGACGGCCAGT) and reverse M13 primers labeled at their 5′ ends with digoxigenin (DIG) and by purifying each product in a 2 to 3% low-melting-point agarose gel. DNA (20 nM) and protein (from 0 to 2 μM) were mixed in binding buffer (10 mM Tris-HCl [pH 8], 10 mM HEPES, 50 mM KCl, 1 mM EDTA, 5% [vol/vol] glycerol, 1 μg of bulk carrier salmon sperm DNA, 0.5 mM 1,4-dithiothreitol, and 0.1 mg of bovine serum albumin per ml). Binding reaction mixtures (20 μl) were loaded onto a 6% nondenaturing Tris-glycine-EDTA (TGE) polyacrylamide gel (500 mM Trizma base, 380 mM glycine, 2 mM EDTA, pH 8.5). DNA-protein complexes were separated at 150 V for 90 min in TGE buffer and then transferred to a Biodine B nylon membrane (Pall Gelman Laboratory). DIG-labeled DNA-protein complexes were detected according to the manufacturer's protocol (Roche). When required, promoter fragments without DIG were used as unlabeled DNA competitors. All EMSAs were repeated a minimum of three times, with reproducible results.

Microarray data accession number.

The resulting microarray data sets (MMC-treated WT versus non-MMC-treated WT) were submitted to the NCBI Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE44735.

RESULTS AND DISCUSSION

The general transcriptional response of A. baumannii to MMC.

Several genes commonly found in most bacterial species and induced as part of the SOS DNA damage response are not present within the genomes of Acinetobacter spp. The missing genes include those encoding the error-prone DNA polymerase II (PolB), the transcriptional factor σ³⁸, and several cell division proteins (SulA, FtsE, and FtsX) (5). In addition, Acinetobacter spp. lack a clear homologue of the SOS repressor LexA (5). These differences point to an unusual DNA damage response in these bacteria. Until very recently, only two DNA-damage-inducible loci had been identified in Acinetobacter spp.: recA, which functions in DNA repair and recombination, and ddrR, a gene of unknown function that seems to be present in only the Acinetobacter genus (5, 8, 10, 12). However, very recently, a semiquantitative RT-PCR analysis showed that uvrA (encoding a putative enzyme involved in nucleotide excision repair) and several umuDC genes (encoding putative proteins involved in mutagenic repair) are present in the A. baumannii genome and that they are likewise induced by DNA-damaging agents (7). In this work, we used DNA microarrays to characterize the DNA damage response in A. baumannii. Specifically, we determined the transcriptional response of A. baumannii strain 17978 to MMC exposure. The results showed that in cultures treated for 2 h with 0.5 μg of MMC ml⁻¹, the transcription of 39 genes (including recA, ddrR, uvrA, and all umuDC genes) increased ≥1.5-fold over the levels measured in untreated cultures of the same strain (Table 1). Our results agree with those reported very recently (7). In addition, data obtained in the microarray analysis were validated by RT-qPCR by using a set of genes induced in the microarray and a set whose level was not affected by MMC (data not shown). Following the treatment of A. baumannii strain 17978 with MMC, 19 genes belonging to two putative prophages were induced (Table 1). However, the lack of a reduction in the optical density of the A. baumannii cultures after MMC treatment (data not shown) suggested that both putative prophages are defective in their lytic infection abilities. The most highly expressed gene was the A1S_0408 gene, encoding a putative glutathione S-transferase (GST), part of a superfamily of enzymes that play a major role in the detoxification of many drugs (17) (Table 1). Furthermore, several genes encoding DNA metabolism proteins were induced in the MMC-treated A. baumannii WT strain 17978 (Table 1). Finally, five genes had no identifiable sequence homologues in the current sequence databases, and their functions are therefore currently unknown (Table 1).

Table 1.

Genes deregulated in MMC-treated wild-type A. baumannii

Predicted functional group and genea	Product descriptionb	Expression level (fold change)c
Prophage
A1S_1145	Putative Cro protein	2.4
A1S_1149	Putative phage-related protein	1.6
A1S_1155	Putative phage-related protein	1.5
A1S_1156	Putative phage-related protein	1.6
A1S_1581	Putative methyltransferase	1.9
A1S_1583	Hypothetical protein (putative family peptidase S24)	1.5
A1S_1587	Phage terminase EsvK2	2
A1S_1588	Phage terminase-like protein large subunit	1.9
A1S_1590	Peptidase U35 phage prohead HK97	1.9
A1S_1591	Phage major capsid protein HK97	2.8
A1S_1592	Phage head-tail adaptor	2.1
A1S_1593	Hypothetical protein (putative phage protein HK97 gp10 family)	2.3
A1S_1594	Hypothetical protein (putative phage protein)	2.2
A1S_1595	Hypothetical protein (phage tail protein)	1.8
A1S_1596	Hypothetical protein (phage tail protein)	1.8
A1S_1597	Lambda family phage tail tape measure protein	1.7
A1S_1598	Hypothetical protein (putative phage protein)	1.9
A1S_1599	Hypothetical protein (putative phage protein)	2
A1S_1600	Lysozyme	2.1
DNA repair
A1S_0636	DNA polymerase V component UmuD	3.7
A1S_1174	DNA polymerase V component UmuD	2.8
A1S_1389	DNA polymerase V component UmuDAb	2.7
A1S_1962	Recombinase A	2.1
A1S_2008	DNA polymerase V component UmuC	3.1
A1S_2015	DNA polymerase V component	1.5
A1S_2035	Hypothetical protein (putative endonuclease)	2
A1S_2036	DNA cytosine methyltransferase	1.7
A1S_2039	Hypothetical protein (DNA polymerase III subunit-like)	3.2
A1S_2586	dGTP triphosphohydrolase-like protein	2.3
A1S_3115	Hypothetical protein (putative DNA metabolism protein)	2.5
A1S_3116	Hypothetical protein (putative DNA repair SAM protein)	3.7
A1S_3295	Nucleotide excision repair component UvrA	4.3
Unknown
A1S_1143	Hypothetical protein	3.4
A1S_1226	Hypothetical protein	1.9
A1S_1388	Hypothetical protein DdrR	4.2
A1S_2033	Hypothetical protein	1.5
A1S_2038	Hypothetical protein (putative lipoprotein)	4.2
A1S_3385	Putative membrane protein	1.9
Detoxification
A1S_0408	Putative glutathione S-transferase	7.3

^aGene designation for A. baumannii 17978 strain.

^bFor hypothetical proteins, a TBLASTX search was carried out, and the best match, if found, is indicated in parentheses.

^cExpression in MMC-treated WT versus that in untreated WT. The standard deviation was <10% in all cases.

UmuDAb binds specifically to a palindromic sequence present in its promoter region.

UmuDAb, the product of the A. baumannii MMC-inducible gene A1S_1389 (Table 1), is a homologue of the error-prone polymerase accessory UmuD and is present throughout the Acinetobacter genus (11). In E. coli, the protein products of umuDC genes bind to form a UmuD₂C complex, which acts as a checkpoint inhibitor of cell division in order to support error-free repair mechanisms (18). The formation of this complex involves the RecA-mediated cleavage of UmuD to yield UmuD′ (19), which subsequently binds to UmuC to form UmuD′₂C (DNA polymerase V). However, in A. baumannii, RecA-induced autocleavage of the UmuDAb protein more closely resembles that of LexA-type repressors than UmuD-DNA polymerase V components (11). Moreover, the UmuDAb amino acid sequence contains an N-terminal extension (about 30 amino acids) that does not occur in other UmuD proteins. With these observations in mind, we analyzed the UmuDAb promoter region and found an almost perfect palindromic sequence, as a putative DNA binding motif, that is shared with an adjacent MMC-inducible gene, A1S_1388, also known as ddrR (Fig. 1). In Acinetobacter spp., ddrR differs from recA in that induction of the latter after DNA damage does not require RecA (7, 8, 10). In addition, we did not detect any similar palindromic sequences in the upstream DNA sequence of recA, and UmuDAb did not bind to the recA promoter (data not shown).

Fig 1.

Intergenic region showing the 282 bp between the ddrR (A1S_1388) and umuDAb (A1S_1389) genes in A. baumannii strain 17978. The palindromic DNA motif is underlined.

Hare et al. suggested that UmuDAb acts as a transcriptional regulator in A. baumannii through indirect mechanisms, given that there is no amino acid similarity between the DNA-binding N-terminal domains of UmuDAb and LexA (11). To further examine this possibility, we cloned the umuDAb gene of A. baumannii strain 17978 and purified its protein product (Fig. 2A) to investigate its capacity to bind to the putative promoter region shared by ddrR and umuDAb. EMSA with purified UmuDAb showed that it bound specifically to the palindromic motif contained in the shared promoter region (Fig. 2B). Furthermore, modification of any of the central (Fig. 3A, positions 2 to 6) but not the distal (Fig. 3A, positions 1 and 7) motifs abolished gel retardation of the band, as did a one- to three-base insertion after the T residue of the central motif GTAAC (Fig. 3B). This finding implied that the perfect symmetry of the palindrome is crucial for its recognition by UmuDAb.

Fig 2.

(A) SDS-PAGE analysis of lysates prepared from E. coli BL21 cells carrying the empty vector and of lysates from the same bacterial strain transformed with the recombinant construct, before (lanes 1 and 3, respectively) and after (lanes 2 and 4, respectively) isopropyl-β-d-thiogalactopyranoside induction. Lane 5 shows the purified UmuDAb protein. M, protein molecular mass markers. (B) Electrophoretic mobility of a DIG-labeled DNA fragment containing the palindromic DNA motif identified in the predicted promoters of the A. baumannii genes umuDAb and ddrR. The experiment was carried out in the presence of increasing concentrations (0 to 2 μM) of purified UmuDAb protein, with at least a 10-fold molar excess of nonspecific or specific unlabeled DNA.

Fig 3.

(A) Nucleotide substitutions in the palindromic DNA motif, including the surrounding areas, and their effects on the electrophoretic mobility of the umuDAb-ddrR predicted promoter in the presence of purified UmuDAb protein (50 nM). Modified motifs are indicated by arrows and numbers. (B) Nucleotide insertions in the palindromic DNA motif and their effects on the electrophoretic mobility of the umuDAb-ddrR predicted promoter in the presence of purified UmuDAb protein (50 nM). In both panels, the mobility of the wild-type fragment in the absence (−) or presence (+) of the same amount of purified UmuDAb protein is shown as a control.

All A. baumannii umuDC homologues and ddrR are under UmuDAb control.

To investigate the presence of additional putative common elements associated with the DNA-damage-induced mutagenesis response in A. baumannii, we analyzed the upstream regions of all MMC-induced genes (Table 1). Palindromic sequences similar to the 21-bp palindrome motif identified in the umuDAb and ddrR promoter regions were also detected in the putative promoter regions of four genes: A1S_0636, A1S_1174, A1S_2008, and A1S_2015 (Table 2). Notably, these genes encode homologues of UmuD or UmuC, both of which, as cited above, are components of the error-prone DNA polymerase V. Comparisons of these sequences yielded the consensus sequence TTGAN₄GTWACN₄TCAA (Fig. 4). We then searched the complete genomic sequence of A. baumannii strain 17978 for this 21-bp consensus motif but were unable to identify any further copies. In subsequent experiments, we found that UmuDAb bound to the four additional palindromic DNA motifs identified in A. baumannii strain 17978 (Fig. 5). These motifs were present, in all cases, in the predicted promoters of the MMC-induced genes (Table 2). In addition, none of the eight genes controlled by promoters containing this DNA motif (Table 2) were induced in the MMC-treated recA mutant derivative of A. baumannii strain 17978 (data not shown), indicating that RecA is critical for the expression of these genes following DNA damage. It is worth noting that, as previously described by Rauch et al., induction of recA after MMC treatment in a recA mutant was detected through qRT-PCR (data not shown), which indicated that, unlike in most bacteria, recA does not require RecA for its expression (8).

Table 2.

Genes belonging to the UmuDAb regulon

Genea	Homologous gene	Predicted function	Palindromic DNA motifb	Positionc
A1S_0636	umuD	DNA polymerase V component	TTGATTACGTTACGTTTTCAA	12
A1S_0637	umuC	DNA polymerase V component
A1S_1174	umuD	DNA polymerase V component	TTGAATATGTTACAAAATCAA	15
A1S_1173	umuC	DNA polymerase V component
A1S_1388d	ddrR	Hypothetical protein	TTGAAAAAGTTACATTTTCAA	218
A1S_1389	umuDAb	DNA polymerase V component	TTGAAAATGTAACTTTTTCAA	45
A1S_2008	umuC	DNA polymerase V component	TTGATTATGTTACAAATTCAA	212
A1S_2015	umuC	DNA polymerase V component	TTGAATTTGTTACGATTTCAA	169

^aGene designation for A. baumannii strain 17978. The first gene of a putative operon is given in bold.

^bConserved nucleotides are underlined.

^cNumber of bases between the regulatory sequence and the start codon of the gene.

^dA1S_1388 (ddrR) and A1S_1389 (umuDAb) are transcribed in opposite directions. The two genes share a palindromic DNA motif (Fig. 1).

Fig 4.

Consensus sequence of the A. baumannii palindromic DNA motif. E value, <0.001; P value, ≤2.04 × 10⁻¹³ in all cases.

Fig 5.

Electrophoretic mobility of DIG-labeled DNA fragments containing the A. baumannii palindromic DNA motif identified in the predicted promoter regions of the indicated genes in the presence of purified UmuDAb protein (50 nM).

In A. baumannii, like the self-cleaving serine proteases LexA and UmuD, UmuDAb undergoes a LexA-like cleavage event after DNA damage in a process requiring RecA (11). In addition, as discussed above, UmuDAb specifically binds to five similar palindromic DNA motifs, all of them present in the putative promoter region of A. baumannii DNA-damage-inducible genes (Table 2). Accordingly, to assess whether the function of UmuDAb is analogous to that of the LexA regulator in other bacteria, we constructed a umuDAb mutant and then carried out qRT-PCR to analyze all of the putative transcriptional units whose predicted promoter regions contained the palindromic DNA motif described in this work (Fig. 6). Surprisingly, the results showed the transcription of both upregulated and downregulated genes, suggesting a dual regulatory activity of UmuDAb. The expression of ddrR and all umuDC homologues, including umuDAb, increased in the umuDAb mutant, indicating that UmuDAb acts to repress these genes under normal conditions. However, the genes A1S_0636-A1S_0637 were highly downregulated in the umuDAb mutant, indicating that UmuDAb may also be able to act as a transcriptional activator (Fig. 6A). Accordingly, the basal expression of these genes (A1S_1174-A1S_1173, ddrR, umuDAb, A1S_2008, and A1S_2015) is reduced when the mutant is complemented with the pET-RA plasmid (20) carrying the WT gene (Fig. 6B). Moreover, under these conditions, expression of the genes A1S_0636-A1S_0637 is increased in concordance with a positive-regulator role for these genes (Fig. 6B). A similar dual activity of a regulator of DNA-damage-inducible genes was previously reported for the Rhodobacter sphaeroides LexA protein, which also acts as both a repressor and an activator (21).

Fig 6.

Basal expression factors of the indicated genes in the umuDAb mutant (A) and in the complemented umuDAb mutant (B). The expression factor is the ratio of the mRNA concentration of each gene from either the umuDAb mutant (A) or the complemented umuDAb mutant (B) with respect to the wild-type strain (ATCC 17978). The amount of mRNA of each gene was determined by using a standard curve generated by the amplification of an internal fragment of the A. baumannii gyrB gene (see Table S1 in the supplemental material for primer sequences). The results are the mean of two independent experiments, each carried out in duplicate. Error bars, standard deviations.

Taken together, our results provide evidence of a novel coordinated DNA damage response in Acinetobacter spp. regulated by UmuDAb, whose functional role is similar to that of the canonical LexA in other bacteria.