A WYL transcriptional regulator activates the DNA damage response pathway in Acinetobacter species

Abstract

The ability to sense DNA damage and activate DNA damage response pathways is critical for repairing DNA damage in all domains of life. The most well studied pathway for DNA damage repair (DDR) in bacteria includes the “save our souls” (SOS) pathway, although many components of this pathway are missing in the Acinetobacter clade. One aspect of DDR pathways includes the inhibition of cell division to prevent the inheritance of damaged DNA by daughter cells, the mechanism of which is also unknown in Acinetobacter. In this work, we show the cell filamentation factor AciT acts as a cell division inhibitor that confers a fitness advantage in the presence of DNA damage-inducing agents. Suppressor mutations that permit cell viability in conditions with constitutively active AciT mapped to a WYL family transcriptional regulator, DdaA, which we demonstrate acts as an activator of aciT expression. DdaA was also found to activate DDR pathway gene expression including recA via a conserved mechanism used by WYL family homologues to sense and respond to DNA damage. We further demonstrate that DdaA activates the expression of DDR pathway genes upon DNA uptake during natural transformation, identifying a new mechanism for how transcriptional control networks intersect to regulate bacterial responses in diverse environmental contexts.

Graphical Abstract

Graphical Abstract.

Introduction

Bacteria encounter DNA damage in the environment from numerous sources including ultraviolet radiation and antibiotic exposure, and the ability to sense and repair this damage is critical for survival. The most well-characterized bacterial DNA damage repair (DDR) pathway is the “save our souls” (SOS) pathway that is defined by the regulatory activity of the proteins RecA and LexA. RecA becomes activated by binding to single-stranded DNA (ssDNA) upon DNA damage, triggering the autoproteolysis of the widespread transcriptional repressor LexA [1, 2]. In the model Escherichia coli, relief of repression by LexA enables the transcription of >30 DDR genes, including RecA, to create a feedback loop that maintains DDR gene expression until DNA damage is resolved [3, 4]. While LexA is widely distributed among bacteria, many species including those found within the Acinetobacter genus do not possess a LexA homologue despite exhibiting a robust DNA damage response [5]. How species lacking LexA and other known DDR regulators sense DNA damage and activate DNA repair genes has remained unclear, prompting necessary analysis of DDR pathways in these organisms.

Previous studies have shown that Acinetobacter species possess a unique UmuD homologue that acts similarly to LexA to repress DDR genes. Canonical umuD genes encode error-prone DNA polymerases that are upregulated upon DNA damage, but Acinetobacter species have evolved a truncated version of UmuD called UmuDAb that can undergo RecA-stimulated autoproteolysis to relieve DDR gene repression [5, 6]. While recA expression also increases upon DNA damage, RNA sequencing (RNAseq) data show that UmuDAb does not transcriptionally regulate recA expression or several other significantly upregulated DDR genes [5]. These results suggest another transcriptional regulator may activate the expression of recA and genes encoding factors important for DDR.

A highly conserved characteristic of bacterial DDR pathways is the inhibition of cell division to enable DNA repair and prevent daughter cells from inheriting damaged DNA. The small protein SulA is the most well-characterized DDR-regulated cell division inhibitor, and its mechanism of action includes binding to the master cell division recruitment factor FtsZ to inhibit cell division [7, 8]. Like lexA, sulA is missing from Acinetobacter genomes, and the mechanism of DDR-activated cell division inhibition in these species has remained unknown. A previous study identified the small membrane-anchored protein AciT as a cell filamentation factor in Acinetobacter baumannii upon ciprofloxacin exposure and proposed AciT may act as a SulA analogue in the Acinetobacter genus [9]. RNAseq data show that aciT expression increases upon DNA damage, but neither RecA nor UmuDAb are responsible for aciT upregulation, making it unclear how aciT expression is activated [5]. In this study, we sought to characterize AciT regulation and identify DDR factors that contribute to its increased expression upon DNA damage. To identify such factors, we constructed a strain expressing constitutively active AciT and isolated suppressor mutants that were resistant to AciT-dependent cell death. Suppressor mutations were mapped to a gene encoding a WYL transcriptional regulator which we found activates both aciT and recA expression upon DNA damage in addition to regulating the expression of >300 other genes. While this work was ongoing, another group also identified this same WYL protein as a DDR activator in the Moraxellaceae family (which includes Acinetobacter species) and named it DdaA [10]. The data between our groups show that DdaA behaves similarly to other WYL transcriptional regulators to sense DNA damage and facilitate promoter binding and subsequent DDR gene expression. The collective data independently derived by each group offer a more comprehensive picture of DDR regulation in Acinetobacter species than either study alone and demonstrate the reproducibility and validity of the findings in both studies.

A connection between DDR pathway activation and environmental DNA uptake during natural transformation (also known as natural competence) has been reported for several species, though how these networks overlap and the functional consequences remain unclear [1, 11–14]. Environmental DNA uptake mediated by widespread, dynamic bacterial appendages called type IV pili (T4P) has recently been shown to trigger the DNA damage response pathway and upregulate recA expression in Acinetobacter baylyi [15]. Upon T4P binding and uptake, DNA becomes single-stranded upon entry into the cell cytoplasm, and if there is sufficient homology to the chromosome, it can be incorporated via homologous recombination in a process called natural transformation [16–18]. ssDNA entering the cell during natural transformation was shown to trigger the activation of the DDR pathway, but the mechanism has remained unknown until this work. We find that DdaA activates recA expression upon DNA uptake, identifying the mechanism of DDR activation during natural transformation and the overlap between these two important processes.

Materials and methods

Bacterial strains and culture conditions

Acinetobacter baylyi strain ADP1 [19, 20] was used throughout this study. Acinetobacter nosocomialis strain M2 [21] and Acinetobacter radioresistens strain LH6 [22, 23] were used where indicated. For a list of strains, see Supplementary Table S1. Acinetobacter baylyi cultures were grown at 30°C in Miller lysogeny broth (LB) medium and on agar supplemented with kanamycin (50 µg/ml), spectinomycin (60 µg/ml), zeocin (50 µg/ml), and apramycin (50 µg/ml) where appropriate. Acinetobacter nosocomialis and A. radioresistens were grown at 37°C in LB medium and on LB agar supplemented with spectinomycin (60 µg/ml) where appropriate.

Construction of mutant strains in A. baylyi

Mutants in A. baylyi were made using natural transformation as described previously [24, 25]. Mutant constructs were made by splicing-by-overlap (SOE) polymerase chain reaction (PCR) to stitch (i) ∼3 kb of the homologous region upstream of the gene of interest, (ii) the mutation where appropriate (for deletion by allelic replacement with an AbR cassette), and (iii) ∼3 kb of the homologous downstream region. For a list of primers used to generate mutants in this study, see Supplementary Table S2. The upstream region was amplified using F1 + R1 primers, and the downstream region was amplified using F2 + R2 primers. All AbR cassettes were amplified with ABD123 (ATTCCGGGGATCCGTCGAC) and ABD124 (TGTAGGCTGGAGCTGCTTC). SOE PCR reactions were performed using a mixture of the upstream and downstream regions, and middle region where appropriate using F1 + R2 primers. SOE PCR products were added with 50 µl of overnight-grown culture to 450 µl of LB in 2-ml round-bottom microcentrifuge tubes (USA Scientific) and grown at 30°C rotating on a roller drum for 3–5 h. For AbR-constructs, transformants were serially diluted and plated on LB and LB + antibiotic. For protein fusion constructs, after the 3–5 h incubation, cells were diluted and 100 µl of 10⁻⁶ dilution was plated on LB plates. Deletions were confirmed by PCR using primers ∼150 bp upstream and downstream of the introduced mutation, and fusions were confirmed by sequencing.

To construct the aciT^3xFLAG strain, primers 2313 + 2361 were used to amplify aciT upstream of the stop codon and primers 2362 + 2316 were used to amplify the region downstream of aciT with universal linkers built into the R1 and F2 primers. A second set of PCR reactions was then performed to add the 3xFLAG tag to both arms using 2313 + 1778 on the up arm template, and 1777 + 2316 on the down arm template followed by SOE PCR and transformation as described above. 

Fluorescence reporter strains were constructed at the vanAB locus using a strain containing a previously published P_tac-mRuby3 as a DNA template [26]. First, the up arm without the P_tac promoter was amplified using primers 317 + 693 and the down arm containing the mRuby3 open reading frame was amplified using primers 1670 + 176. P_recA was amplified using primers 1668 + 1669 and P_aciT was amplified using primers 2437 + 2364. Up + Down + promoter regions were then amplified by SOE PCR using primer 317 + 176 and transformed into the parent strain CE100. The P_aciTmut-mRuby3 strain was constructed using the P_aciT-mRuby3 strain as a DNA template. PCR regions for P_aciTmut mutant construction were amplified using primers 173 + 4116 for the up arm and 4117 + 176 for the down arm. The SOE PCR was then amplified using primers 173 + 176 and transformed into the parent strain CE100. Reporter strains were confirmed by sequencing.

Ectopic expression strains were constructed by placing the gene of interest under a LacI-inducible P_tac promoter at either the vanAB locus for aciT expression or the ACIAD1551 locus for ddaA expression. For the P_tac-aciT strain, we first engineered the vanAB locus to contain a LacI-inducible P_tac- promoter. The vanAB locus containing the kan^R cassette from the constitutive (no LacI) P_tac promoter published previously [26] was amplified using primers 317 + 1816 to amplify the upstream arm and primers 406 + 176 to amplify the downstream arm. The lacIq, P_tac- promoter region was amplified from a strain containing lacIq, P_tac-GFP [27] using primers 1817 + 1796. aciT was amplified using primers 3189 + 2358. The up arm containing vanAB::kan^R, lacIq, P_tac-, down arm containing ∼3 kb homology to the vanAB locus, and aciT were assembled by SOE PCR using primers 317 + 176 and transformed into the parent strain. For ectopic expression of ddaA at the ACIAD1551 locus (a frame-shifted transposase locus), we built a strain containing ∆ACIAD1551::zeo^R, lacIq, P_tac-ddaA with a previously published transcriptional terminator built in downstream of the ddaA stop codon [28]. ACIAD1551 was first replaced with a zeo^R cassette using standard ABD Ab^R replacement protocols described above using ∼3 kb arms of homology for deletion constructs with indicated primers in Supplementary Table S2. To reduce the length of SOE PCR products for easier amplification, ACIAD1551 expression constructs were amplified using ∼1 kb arms of homology (3044 for F1 and 3048 for R2). The following pieces were assembled to generate the ∆ACIAD1551::zeo^R, lacIq, P_tac-ddaA construct: The up arm containing ACIAD1551::zeo^R was amplified using primers 3044 + 234. The lacIq, P_tac- promoter region was amplified using primers 3889 + 1796 from a vanAB::kan^R, lacIq, P_tac- promoter strain, the ddaA gene was amplified using primers 3891 + 3892, and the transcriptional terminator from strain JBA163 [28] could be amplified using primers 3284 + 3046. The down arm containing homology to the ACIAD1551 locus was amplified using primers 3047 + 3048. To generate the ddaA-3xFLAG strains for western analysis, primers 3044 + 3171 was used to amplify ∆ACIAD1551::zeo^R, lacIq, P_tac-ddaA and primers 4174 + 3048 was used to amplify the downstream region from the above expression strain with C-terminal fusion linkers built into 3171 and 4174 primers. The 3xFLAG tag was then added to each PCR product using primers 3044 + 1778 for the up arm and primers 1777 + 3048 for the down arm. The SOE PCR product was amplified using primers 3044 + 3048 and transformed into the parent strain. Point mutations were built into indicated R1 and F2 primers denoted in Supplementary Table S2 targeting indicated residues using 3044 + 3048 as the F1 and R2 primers. SOE products were again amplified using primers 3044 + 3048 and transformed into the parent strain. All ectopic expression constructs were confirmed by sequencing.

Construction of mutant strains in A. nosocomialis and A. radioresistens

Mutants of A. nosocomialis and A. radioresistens were constructed using SOE PCRs similar to A. baylyi with some differences as described previously [24]. Briefly, strains were grown by shaking overnight in LB medium at 37°C. Then, 50 μl of overnight culture were subcultured into 3 ml of LB broth and grown with shaking for 2–3 h to exponential growth phase. Exponential cultures were diluted 1:100 in phosphate-buffered saline (PBS), and then 5–10 μl of diluted culture was mixed with >50 ng of transforming DNA and spotted onto the surface of transformation agar (TA; 2.5 g/l NaCl, 5 g/l tryptone, 2% agarose) in microcentrifuge tubes and incubated overnight at 30°C. The next day, cells were removed from the surface of the TA using 200 μl of PBS to resuspend surface-grown cells. The cell suspension was then serially diluted and plated on LB + antibiotic plates and grown overnight at 37°C. AbR mutants were struck out for single colony isolation. Deletions were confirmed by PCR.

Cell length measurements

Cultures of A. nosocomialis, A. radioresistens, and A. baylyi were grown overnight for 16–24 h to stationary phase in LB medium supplemented with or without mitomycin C (MMC) (Fisher BioReagents) at concentrations of 20 μg/ml for A. nosocomialis and A. radioresistens or 2 μg/ml for A. baylyi. Overnight cultures were spotted onto a coverslip under an LB 1.5% agar pad and imaged using a Nikon Ti2-E microscope using a Plan Apo 100× oil immersion objective, a Hamamatsu ORCA-Fusion Gen-III camera, and Nikon NIS Elements Imaging Software. Cell lengths were quantified using the ImageJ plug-in MicrobeJ [29, 30]. Both parent and ∆aciT mutants exhibited cell chaining phenotypes (defined as more than two cells connected by distinct septal constriction sites). The lengths of chained cells or cells within chains were not quantified, as this phenotype is AciT-independent. At least 50 cells were measured for each replicate.

Competition assays

Strains were grown for 18 h in LB broth at either 30°C (A. baylyi) or 37°C (A. nosocomialis and A. radioresistens). Cultures were normalized to an OD₆₀₀ of 1.0 in a volume of 500 µl. Parent and ∆aciT::Ab^R mutant cultures were mixed at a 1:1 ratio, serially diluted, and plated for quantitative culture on LB + appropriate antibiotic plates (to quantify the number of viable ∆aciT::Ab^R cells present) and on plain LB plates (to quantify total viable cell counts). Thirty microliters of the remaining culture mixture was subcultured into 3 ml fresh LB medium with or without MMC (20 µg/ml for A. nosocomialis and A. radioresistens, 2 µg/ml for A. baylyi) at 37°C or 30°C dependent on species for 6 h. Cultures were then serially diluted and plated for colony forming unit (CFU) counts on LB + antibiotic plates and on plain LB plates exactly the same as above. The proportion of N_∆acIT:N_total was calculated by dividing Ab^R colonies on LB antibiotic plates by the total number of colonies from LB plates without selection.

Western blotting

Cells grown with 2 µg/ml of MMC do not consistently reach high enough OD₆₀₀ readings for bulk population assays, and we thus used a lower MMC concentration for more consistent measurements for certain experiments where indicated. Overnight cultures grown with or without 1.33 µg/ml of MMC were normalized to an OD₆₀₀ of 1.0 in a volume of 1 ml, concentrated into a pellet by centrifugation, and the culture supernatant was discarded. Cell pellets were resuspended in 50 µl PBS and then mixed with an equal volume of sodium dodecyl sulphate–polyacrylamide gel electrophoresis sample buffer [250 mM Tris, pH 6.8, 40% glycerol, 8% sodium dodecyl sulphate (SDS), 0.8% bromophenol blue, and 20% β-mercaptoethanol] and denatured using a heat block set to 99°C for 10 min. Proteins were separated on a 4%–20% pre-cast polyacrylamide gel (Bio-Rad) by SDS electrophoresis, electrophoretically transferred to a nitrocellulose membrane, and probed with 1:5000 dilution of mouse monoclonal α-FLAG antibodies (Sigma) and/or a 1:12 000 dilution of mouse monoclonal α-RpoA (BioLegend) primary antibodies. Blots were washed and then incubated in a 1:10 000 dilution of goat α-mouse antibody conjugated to horseradish peroxidase secondary antibody (Sigma). Blots were washed again and then incubated with SuperSignal West Pico PLUS Chemiluminescence substrate (Thermo Fisher). Blots were then imaged using a Bio-Rad Chemidoc imaging system. Western quantification was performed using ImageJ software gel analysis tools. The RpoA and AciT protein levels for each replicate were normalized to the RpoA or AciT levels respectively from the aciT^3xFLAG strain grown without MMC to normalize for day-to-day variation in immunoblot signal strength. AciT protein levels for each strain and condition were then normalized to their respective RpoA levels.

Growth curve assays

Strains were grown overnight in LB broth at 30°C on a roller drum. Then, 1 µl of overnight cultures were subcultured into 200 µl fresh LB medium with vancomycin (MP Biomedicals) (0 µg/ml, 25 µg/ml, 50 µg/ml, 100 µg/ml, 250 µg/ml) grown in a 96-well plate in a Synergy H1 multimode plate reader (BioTek) utilizing Gen 6 software at 30°C shaking for 16 h. OD₆₀₀ was measured every 10 min.

Suppressor selection and mutation mapping

Tubes containing 3 ml LB medium supplemented with 100 µg/ml vancomycin were inoculated with ∼10⁸ cells from an overnight culture of an aciT-3xFLAG ∆umuDAb strain and grown at 30°C until cultures exhibited turbidity (∼2 days). Cultures were plated for single colony isolation on plain LB plates, and smooth, small colonies were inoculated into 3 ml LB with or without 2 µg/ml MMC (cells with constitutive AciT activity exhibit a “fuzzy” or enlarged colony morphology). The next day, overnight cultures were imaged by microscopy to qualitatively assess cell division phenotypes. Genomic DNA (gDNA) was extracted (DNeasy Blood and Tissue Kit, Qiagen) from cultures presenting reduced cell filamentation and shipped to SeqCenter for Illumina whole genome sequencing and analysis. Illumina-generated 2 × 151 bp paired-end read data was generated and used as the input for variant calling against the A. baylyi strain ADP1 reference genome [19]. Variant calling was carried out using BreSeq under default settings [31].

RNAseq experiments

Cultures were grown overnight for 16–20 h to stationary phase in LB medium supplemented with or without 2 μg/ml MMC. Two milliliters of overnight cultures normalized to an OD₆₀₀ of 1.0 were centrifuged at 18 000 × g and the supernatant was discarded. Pellets were stored at −70°C until all samples were collected, and then pellets were shipped on dry ice to SeqCenter. RNA was extracted using the ZymoBIOMICS Quick-RNA Miniprep Kit (Zymo Research) following Zymo Research recommendations. Samples were eluted into a final volume of 40 μl and RNA concentrations were determined by Qubit. Samples were DNAse treated with RNAse-free DNAse (Invitrogen). Library preparation was performed using Illumina’s Stranded Total RNA Prep Ligation with Ribo-Zero Plus kit and 10 bp unique dual indices. Sequencing was done on a NovaSeq X Plus, producing paired end 150 bp reads. Demultiplexing, quality control, and adapter trimming was performed with bcl-convert (v4.2.4). Read mapping was performed with HISAT2 [32]. Read quantification was performed using Subread’s featureCounts functionality [33]. For a summary of differentially expressed genes and their statistics refer to Supplementary Dataset 1. An interactive browser-viewable form of these data can also be found here: https://genome.ucsc.edu/s/ckellison/MMC_gene_expression.

Fluorescence reporter assays

Cultures were grown overnight for 16–20 h at 30°C to stationary phase in LB medium supplemented with or without 1.33 μg/ml MMC. The next morning, 200 µl of overnight cultures was centrifuged at 18 000 × g and the supernatant was discarded. Cell pellets were resuspended in PBS buffer, which has significantly less background fluorescence than LB medium, and fluorescence intensity was measured using a Synergy H1 multimode plate reader (BioTek) utilizing Gen 6 software. Plate reader fluorescence measurements were normalized to OD₆₀₀.

Escherichia coli and A. baylyi co-incubation experiments

Overnight cultures of A. baylyi in plain LB and E. coli in LB supplemented with 30 µg/ml gentamycin grown at 30°C for 16–20 h were subcultured into 3 ml of media and grown for 2 h at 30°C to early exponential growth phase. One milliliter of each A. baylyi culture and 3 × 1 ml of E. coli culture were centrifuged at 18 000 × g and the supernatant was discarded. Pellets from 1 ml aliquots of A. baylyi strains and E. coli were resuspended in the same 30 µl of fresh LB and spotted onto a pre-warmed LB agar plate and incubated at 30°C for 4.5 h. Cells were then lifted from the surface of LB plates using a P200 pipette tip and resuspended into 100 µl fresh LB. Resuspended cells were spotted onto a coverslip under an LB 1.5% agar pad and imaged using the same microscope set up described above. A DSRed/TRITC/Cy3 filter set was used for fluorescence microscopy to measure mRuby3 fluorescence intensity. Background fluorescence was first subtracted from images and then the ImageJ plug-in MicrobeJ was used for segmentation to outline cells based on the phase contrast images. The integrated density of fluorescence intensity for outlined cells in the fluorescence channel was measured using the built-in ImageJ measurement tool [29, 30]. A total of 30 cells were quantified for each of three independent, biological replicates.

AlphaFold modeling

A random 34 bp sequence of ssDNA was modeled with two monomers of DdaA and a 52 bp region of double-stranded DNA containing the predicted DdaA binding site in the aciT promoter (20 bp of DNA on either side of the binding motif) using the AlphaFold3 server [34]. AlphaFold3 predicted an overall high confidence structure (ipTM = 0.68, pTM = 0.72) with DdaA forming a homodimer that binds to the DdaA binding motif by conserved promoter binding residues in the wHTH domain. All visualization, figure generation, and analysis of the model was performed using ChimeraX software version 1.10.1 [35].

Analysis of DdaA binding motifs in target promoters

Promoters from several significantly downregulated genes in ∆ddaA RNAseq data were searched for putative DdaA binding sequences based on similarity to AlphaFold3-predicted DdaA binding and the Caulobacter crescentus DriD sequence (CGACNNNNNNNGTCG). In addition to P_aciT, predicted binding motifs were identified in the promoters of recA, ACIAD1170, ACIAD2477, ACIAD2480, ssb, uvrA, dgt. Predicted binding motifs were submitted to the WebLogo3 server to generate a sequence conservation logo [36].

Purification of DdaA protein

Protein purification was performed as described previously with some changes [37]. Plasmid pTB146-N-6His-SUMO-ddaA was constructed by Gibson assembly of PCR-amplified pTB146 (primers 3691 + 3692) and ddaA (primers 3748 + 3749). Then, 10 ml of an overnight culture of E. coli strain BL21(DE3) harboring plasmid pTB146-N-6His-SUMO-ddaA was subcultured into 1 l Terrific broth (25 g yeast extract, 12 g tryptone, 0.4% glucose, 2.31 g KH₂PO₄, and 12.5 g K₂HPO₄) and grown at 37°C to an OD₆₀₀ of 0.5–0.6. Expression of 6His-SUMO-DdaA was then induced by the addition of 1 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) and grown at 16°C for ∼20 h. Cells were pelleted by centrifugation at 3750 × g for 30 min at 4°C and then resuspended in 50 ml of base buffer (150 mM NaCl, 25 mM Tris, pH 8.0) containing 5 mM imidazole. Cells were lysed using a cell disruptor and then centrifuged at 20 000 × g for 20 min at 4°C. The clarified lysate was then applied to a column containing nickel-nitrilotriacetic agarose (Ni-NTA) (Qiagen). The column was washed with 25 ml of base buffer containing 25 mM imidazole and protein was then eluted with base buffer containing 250 and 500 mM imidazole and then dialyzed overnight at 4°C in 1 l base buffer containing 10% glycerol, 1 mM dithiothreitol, and purified 6His-Ulp1 protease to remove the 6His-SUMO tag from DdaA. The dialyzed contents were then incubated with Ni-NTA agarose beads using end-over-end rotation at 4°C for 1 h and then applied to a gravity flow column to collect the flow through containing purified DdaA. Final protein concentration was determined to be 1.95 µM by Bradford assay.

Electrophoretic mobility shift assays

Electrophoretic mobility shift assays (EMSAs) were performed as described previously with some changes [38]. To make P_aciT fluorescent probes, ∆vanAB::kan^R, P_aciT-mRuby3 and ∆vanAB::kan^R, P_aciTmut-mRuby3 regions were amplified by PCR using primers 317 + 176 on strains CE1293 and CE2535 respectively. Using amplified vanAB regions as a template, probes were then generated by PCR using primers 4463 + 4464 with primer 4463 containing a 6-FAM modification on the 5′ end (Sigma). Probes were cleaned using the Qiaquick PCR Purification Kit (Qiagen) and quantified by nanodrop. Thirty nanograms (0.231 pmol) of probe was incubated with 0, 2-fold (0.461 pmol), or 8-fold (1.846 pmol) molar excess protein in 1× EMSA buffer (25 mM Tris–HCl, pH 8.0, 10% glycerol, 0.1 mM EDTA, 50 mM NaCl) in 25 µl total volume at 25°C for 40 min. Five microliters of 50% glycerol was added to each sample and then samples were loaded onto a pre-cast 5% Mini-PROTEAN TBE (Tris–borate–EDTA) gel (Bio-Rad) equilibrated at 4°C with 0.5× TBE buffer (1× = 90 mM Tris base, 90 mM boric acid, 2 mM EDTA). Gels were electrophoresed at 120 V in 1× TBE buffer at 4°C for 45 min and then imaged using an Amersham Typhoon biomolecular imager using Cy2 default settings.

Statistics and reproducibility

All experiments were repeated a minimum of three times, and all attempts at replication were successful. Statistics were determined using Prism 10 software version 10.4.1.

Results

AciT is a functionally-conserved DNA damage-induced cell division inhibitor in Acinetobacter species

Cell filamentation is a common phenotypic response after DNA damage, as it allows for damaged DNA to be repaired before inheritance by daughter cells upon division. Many well-characterized species including E. coli employ the small protein SulA to inhibit cell division by binding to the essential cell division protein FtsZ and inhibiting its ability to form stable Z-rings [8, 39, 40]. SulA expression is repressed by LexA, and upon DNA damage, LexA autoproteolysis allows for sulA expression [3]. Acinetobacter species lack both SulA and LexA, yet still exhibit a robust DNA damage response phenotype, suggesting other factors may carry out these functions. A previous study identified the gene aciT as a factor in cell filamentation upon the addition of DNA damage-causing antibiotics in the pathogen Acinetobacter baumannii [9]. Naidu et al. identified aciT as phylogenetically conserved throughout the Acinetobacter clade, and proposed that it may play a similar role as SulA to inhibit cell division upon DNA damage.

To test if AciT plays a functionally-conserved role in DNA damage-induced cell division inhibition in the Acinetobacter genus, we assessed AciT activity in multiple Acinetobacter species. Naidu et al. showed the importance of AciT in cell filamentation in A. baumannii, so we chose the closely related species Acinetobacter noscomialis to represent other medically relevant Acinetobacter pathogens. A commensal strain of A. radioresistens previously isolated from chicken feces was chosen to represent a distant Acinetobacter relative [22, 41], and the laboratory model A. baylyi was chosen as a nonpathogenic environmental representative. Acinetobacter species, and especially A. baylyi, are known for their high rates of natural transformation due to DNA uptake by dynamic T4P [24, 42–45]. Using natural transformation, we replaced aciT in each species with an antibiotic resistance cassette (FDQ49_10 080 in A. nosocomialis M2; DOM24_13 925 in A. radioresistens LH6; ACIAD3565 in A. baylyi ADP1). DNA damage was induced by the addition of the DNA-damaging antibiotic mitomycin C (MMC). The parent strain of each species filamented in the presence of MMC, although to varying degrees (Fig. 1A and B). Notably, the more extremely filamented cells were difficult to accurately measure with quantitative image analysis software and are underrepresented in these data. Populations of all species exhibited heterogeneous cell phenotypes when grown with MMC with the range between the shortest and longest cell lengths spanning 26.76 µm for A. nosocomialis, 13.57 µm for A. radioresistens, and 20.32 µm for A. baylyi. Acinetobacter nosocomialis cells were the longest on average at 6.29 µm, and A. radioresistens cells were the shortest averaging 3.56 µm in length. ∆aciT mutants were comparable to the parent strain in all species in the absence of MMC. However, all ∆aciT mutants showed a significant defect in filamentation in the presence of MMC compared to the parent strain (Fig. 1A and B). Interestingly, although ∆aciT mutants no longer filament in the presence of MMC, cells still exhibit a statistically significant increase in cell length compared to cells grown without MMC, suggesting there may be an additional AciT-dependent mechanism of cell division inhibition (Fig. 1A and B). These data demonstrate that AciT is a functionally-conserved cell division inhibitor induced by DNA damage.

Figure 1.

AciT is a functionally conserved DDR-activated cell division inhibitor. (A) Representative microscopy images of indicated strains and conditions. Scale bars, 5 µm. (B) Cell length measurements of strains shown in panel (A). (C) Competition assays of parent versus ∆aciT in indicated species grown ± MMC. Data show the proportion of CFUs of ∆aciT mutants recovered out of the total number of CFUs recovered from indicated conditions. Bars indicate the mean ± standard error of the mean (SEM). Each data point represents an independent, biological replicate. Statistics were determined using Sidak’s multiple comparisons test. ****P <.0001; ***P <.001; **P <.01; *P <.05; ns, not significant.

The conservation of AciT throughout the Acinetobacter genus suggests there is strong selective pressure to regulate the inhibition of cell division upon DNA damage, leading us to wonder whether the presence of AciT confers a fitness advantage during DDR. We thus competed parent versus ∆aciT strains in the presence of MMC and quantified the number of viable cells (colony forming units, or CFUs) of each strain following 6 h of incubation. Significantly fewer aciT mutants were recovered in competition experiments with each species, showing that AciT-mediated inhibition of cell division provides a fitness advantage in DNA-damaging conditions (Fig. 1C). These data indicate that AciT-mediated cell division inhibition is an important aspect of the Acinetobacter DDR pathway.

Mutations in a WYL transcriptional regulator suppress AciT-mediated cell division inhibition

Because AciT plays an important role in DDR, we next sought to identify regulators of AciT function. To identify candidate regulators, we employed the species A. baylyi due to its high rates of natural transformation and consequent robust genetic tractability [42]. Although Acinetobacter species possess no known LexA homologues, the Acintobacter-conserved protein UmuDAb was previously shown to function similarly in repressing DDR genes in the absence of DNA damage [6, 46]. However, RNAseq datasets from wildtype and ∆umuDAb mutant strains demonstrate aciT transcription is independent of UmuDAb function [5]. In line with these data, deletion of umuDAb does not cause cell filamentation demonstrating that AciT expression is not repressed by this regulator (Fig. 2A and Supplementary Fig. S1) [5]. Notably, ∆umuDAb mutants exhibit slight but significant cell elongation similar to ∆aciT mutants in the presence of MMC, suggesting that it may regulate an aspect of cell morphology independently of AciT (Supplementary Fig. S1).

Figure 2.

A C-terminal 3xFLAG tag constitutively activates AciT. (A) Representative microscopy images of indicated strains. Scale bars, 10 µm. (B) Representative western blot on indicated strains grown with or without 2 µg/ml MMC. Blot was probed with ⍺-RpoA antibody as a loading control and ⍺-FLAG antibody to probe for AciT expression. (C) Growth curves of indicated strains grown with indicated concentrations of vancomycin in 96-well plates over 12 h. Arrows indicate concentration used for selection experiments. Each dot shows the mean of three independent, biological replicates, and error bars show SEM.

To identify AciT regulatory factors, we used forward genetics to select for suppressor mutations that would inhibit AciT function. The mechanism of the cell division inhibitor SulA was originally identified by isolating suppressors in SulA overexpression strains [8]. However, while AciT overexpression induced cell filamentation, it was not sufficient to cause cell death, making it impractical for suppressor selection experiments (Supplementary Fig. S2). In attempts to tag AciT to assess protein expression via western blot analysis, we serendipitously discovered that a 3xFLAG tag appended to its C-terminus causes constitutive cell filamentation (Fig. 2A). Furthermore, deletion of ∆umuDAb in an aciT^3xFLAG background caused a hyper-filamentation phenotype that could not be explained by changes in AciT protein levels (Fig. 2A and B). This hyper-filamenting strain exhibited poorer growth compared to the parent strain, and we hypothesized that it may be more susceptible to outer membrane impermeable antibiotics which could be used in AciT suppressor experiments (Fig. 2C). In line with this hypothesis, growth assays revealed that unlike the parent strain, aciT^3xFLAG ∆umuDAb mutants are unable to grow in 100 µg/ml vancomycin (Fig. 2C).

A strain encoding aciT^3xFLAG ∆umuDAb was thus subjected to growth in liquid in the presence of 100 µg/ml vancomyin and subsequently plated in the absence of vancomycin to select for viable cells. Cell filamentation causes a large, fuzzy colony morphology on agar plates, allowing us to clearly identify AciT-defective strains. Whole genome sequencing of four small colonies from independent cultures grown in the presence of vancomycin revealed mutations within a putative transcriptional regulator annotated in the A. baylyi genome as “mdcH (ACIAD2439)” (Fig. 3A). During the course of this work, another group also identified this same protein as a transcriptional regulator in Acinetobacter species, naming it DNA damage response protein Acinetobacter  A, or “ddaA” [10]. To maintain consistency in the literature, we likewise refer to this regulator as DdaA throughout.

Figure 3.

DdaA regulates AciT expression. (A) Schematic of the domain architecture of DdaA. Numbers indicate positions in protein sequence. Green arrows indicate suppressor mutations obtained in a strain with constitutively active AciT. IS indicates location of mobile genetic insertion sequence. (B) Representative microscopy images of indicated strains in the presence and absence of MMC. Scale bars, 10 µm. (C) Cell length measurements of strains shown in panel (B). Statistics were determined using Sidak’s multiple comparisons test. ****P <.0001; ns, not significant. (D) Western blot showing AciT^3xFLAG expression with RpoA used as a loading control. Blot was probed with both ⍺-RpoA and ⍺-FLAG antibodies. (E) Levels of AciT relative to AciT abundance in the aciT^3xFLAG strain grown without MMC quantified from western blots. Bars indicate the mean ± standard deviation (SD). Each data point represents an independent, biological replicate. Statistics were determined using Sidak’s multiple comparisons test.

Homology searches revealed DdaA is a homologue of DNA sensing transcriptional regulators that contain a winged helix-turn-helix (wHTH) DNA-binding motif predicted to bind target gene promoters and WYL and WCX (WYL C-terminal extension) domains that are predicted to play a sensory role and influence promoter binding (Fig. 3A) [1, 47–49]. WYL transcriptional regulators have been shown to act as both activators and repressors of gene expression depending on the species and regulon function [47]. Three suppressor mutants contained point mutations that fell within the wHTH domain (R47Q) and the WYL domain (A229T, obtained twice independently). The fourth suppressor mutation was an insertion of an A. baylyi-encoded mobilized genetic element (annotated in Fig. 3A as IS for insertion sequence) in the wHTH-encoding region of ddaA. These mutations suggest a loss of function of DdaA is responsible for suppression of AciT-mediated cell division inhibition and suggest DdaA acts as an aciT activator. To test whether DdaA is required for AciT-dependent cell filamentation, we made a deletion strain lacking ddaA and examined cell morphologies. The ∆ddaA mutant exhibited similar defects in cell filamentation as ∆aciT mutants (Figs 1A, and 3B and C), supporting the model that DdaA acts as a transcriptional activator of aciT expression. In line with this hypothesis, western blot analysis of AciT^3xFLAG levels in the ∆ddaA strain showed a significant decrease in AciT protein abundance both in the presence and absence of MMC (Fig. 3D and E). While AciT levels in the aciT^3xFLAG strain increased ∼2.5-fold in the presence of MMC, the ∆ddaA mutant experienced no significant increase in AciT abundance under the same conditions (Fig. 3E).

DdaA is a DDR pathway transcriptional activator

Recent work has established WYL transcriptional regulators as playing important roles in stress responses including DNA damage response pathways in several species [1]. The most well-characterized WYL regulators of DDR include the PafBC complex from Mycobacteria and DriD from C. crescentus, both of which have been shown to sense ssDNA upon DNA damage to activate expression of DNA repair genes including recA [50–53]. RecA is highly expressed upon DNA damage, playing a critical role in DNA repair by relieving LexA-based transcriptional repression. Because RecA is highly upregulated upon DNA damage, but Acinetobacter species possess no known recA transcriptional activators, it seemed likely that DdaA may also activate recA expression. RNAseq experiments revealed that >300 genes were significantly differentially expressed in ddaA mutants compared to the parent strain in DNA-damaging inducing conditions, including aciT, recA, and other DDR pathway genes (Fig. 4A and Supplementary Dataset 1). Notably, the CRISPR locus was among the most significantly downregulated genes in the ∆ddaA mutant, suggesting an overlap between DDR and bacterial immunity pathways (Fig. 4A).

Figure 4.

DdaA transcriptionally activates aciT and other DDR pathway genes. (A) RNAseq plot of a ∆ddaA mutant compared to the parent strain in the presence of MMC. Each dot represents a single gene that was significantly differentially expressed in the ∆ddaA mutant. For the full list of genes and their summary statistics, see Supplementary Dataset 1. (B) Transcriptional fluorescence reporter data for indicated promoters grown in the presence and absence of MMC. Bars indicate the mean ± SD. Each data point represents an independent, biological replicate. Statistics were determined using Sidak’s multiple comparisons test. ****P <.0001; ***P <.001; *P <.05; ns, not significant.

To validate RNAseq results and verify that aciT and recA expression are indeed both activated by DdaA, we built fluorescent protein (mRuby3) transcriptional reporters fused to the promoter region of each gene with identical ribosome binding site sequences. In agreement with RNAseq data, both aciT and recA transcription was significantly reduced in the absence of DdaA when grown in MMC (Fig. 4B). Interestingly, these data show that aciT expression is also regulated by DdaA in the absence of MMC, while recA expression in the absence of MMC is independent of DdaA. Complementation of ddaA at an ectopic site within the chromosome partially restored expression of aciT and recA in the presence of MMC, confirming DdaA as a transcriptional activator of DDR pathway genes (Fig. 4B). aciT expression was also restored by complementation of DdaA in the absence of MMC, suggesting that aciT is constitutively expressed at low levels in the absence of DNA damage in agreement with western blot data (Fig. 3D). These results demonstrate DdaA transcriptionally activates several components of the DDR pathway and consequent AciT-mediated cell filamentation.

DdaA contains functionally conserved residues that activate DDR gene expression

One of the most well-characterized WYL transcriptional activators of a DNA damage response is DriD from C. crescentus [51, 54, 55]. DriD exhibits a similar domain architecture as DdaA, with an N-terminal wHTH domain that binds promoter DNA to activate gene expression and a C-terminal WYL domain that binds and senses ssDNA to facilitate promoter binding. Residues in both the N and C terminal functional domains have been shown to play important roles in each domains’ function, and ClustalW alignments show some of these residues are conserved in DdaA (Supplementary Fig. S3) [56]. Specifically, R47 and R51 in the wHTH domain of DdaA align with residues important for promoter binding and Y173 and Y200 align with residues important for sensing ssDNA (Fig. 5A) [51]. Notably, R47Q was a suppressor mutation isolated from the aciT suppressor selection experiments, supporting the model that this residue is important for DdaA function. DriD forms a homodimer that interacts with promoter DNA upon interaction with ssDNA, and AlphaFold3 predicts a similar topology for DdaA modeled with part of the promoter of AciT and ssDNA (Fig. 5A). The AlphaFold3 model confidently predicted that DdaA might form a homodimer (Supplementary Fig. S4) that binds to the sequence CGACAATATGTGTCG ∼160 bp upstream of aciT which surprisingly shares sequence homology with DriD-regulated promoters (CGACNNNNNNNGTCG) (Fig. 5B) [57]. Searches for similar sequences upstream of other DdaA-regulated genes based on our RNAseq data revealed several additional promoters contain this motif including the recA promoter (CGTCATGTTATGTCG) as well as the promoters of ACIAD1170, ACIAD2477, ACIAD2480, ssb, uvrA, and dgt. To test the importance of this sequence in DdaA function and to tease apart which domains of DdaA are important for promoter binding versus DNA-damage sensing, we employed the aciT transcriptional reporter since its expression is DdaA-dependent both in the presence and absence of DNA damage (Fig. 4B). In contrast, recA expression in the absence of DNA damage is not regulated by DdaA, making it difficult to isolate regions of DdaA that are specific to promoter binding versus DNA damage sensing in vivo. Mutating the four outer-most nucleotides CGACAATATGTGTCG (P_aciT) to ATACAATATGTGTAT (P_aciTmut) completely abrogated aciT transcription in the presence and absence of DNA damage demonstrating this sequence is essential for aciT expression (Fig. 5C). To test if DdaA directly binds to this sequence, we performed in vitro EMSAs using recombinantly produced wildtype A. baylyi DdaA (Fig. 5D). Increasing concentrations of DdaA resulted in a clear shift in the migration of the P_aciT probe, demonstrating that DdaA directly binds the aciT promoter. DdaA failed to bind P_aciTmut, showing that this motif is essential for DdaA-P_aciT binding (Fig. 5D). Because AciT function is conserved in other Acinetobacter species, we speculated that DdaA function may be similarly conserved. A scan for DdaA-binding motifs in the aciT promoters of A. nosocomialis and A. radioresistens revealed sequence conservation with the DdaA-binding sequence identified in A. baylyi (Supplementary Fig. S5A). Deletion of ddaA from either species resulted in reduced filamentation phenotypes similar to ∆aciT deletion strains grown with MMC (Supplementary Fig. S5B–E). These data support a model where DdaA activates DDR and AciT expression upon DNA damage throughout the Acinetobacter clade (Supplementary Fig. S5).

Figure 5.

DdaA regulates DDR genes in Acinetobacter species using a similar mechanism to other WYL homologues. (A) AlphaFold3 model of DdaA dimer (each monomer is colored a different shade of green) interacting with both ssDNA and its predicted DNA-binding box colored in blue in the aciT promoter. Cyan residues are predicted to be involved in ssDNA binding. Magenta residues are predicted to be involved in promoter binding. For confidence metrics of this model, see Supplementary Fig. S4. (B) Logo of conserved residues within promoters of DdaA-regulated genes determined by RNAseq and similarity to the C. crescentus DriD DNA promoter binding sequence. Underlined residues indicate mutation made to P_aciT to prevent predicted DdaA binding. (C) Transcriptional P_aciT fluorescence reporter data for indicated strains grown in the presence and absence of MMC. Bars indicate the mean ± SD. Each data point represents an independent, biological replicate. (D) EMSA using increasing concentrations of purified A. baylyi DdaA with indicated P_aciT probe. Numbers above each lane indicate molar-excess ratio of DdaA:probe. (E) Transcriptional P_aciT fluorescence reporter data for indicated strains grown in the presence and absence of MMC. Bars indicate the mean ± SD. Each data point represents an independent, biological replicate. Statistics were determined using Sidak’s multiple comparisons test. Asterisks above no MMC data indicate significance compared to the P_tac-ddaA^WT strain without MMC. Asterisks above + MMC data indicate significance compared to the P_tac-ddaA^WT strain + MMC ****P <.0001; ***P <.001; **P <.01; ns, not significant.

Given the similarities between DdaA and DriD, we wondered if conserved residues in DriD that are important for promoter binding in C. crescentus are also important for DdaA-mediated expression. DriD contains several positively charged residues that are important for promoter binding that are conserved in DdaA including the region containing residues R47 and R51 (Supplementary Fig. S3). DdaA mutations R47Q or R51Q abolished aciT expression both with and without MMC, supporting a role for these residues in binding the aciT promoter region in the presence and absence of DNA damage (Fig. 5D).

To test whether conserved ssDNA-binding residues found in DriD play a role in DNA damage sensing and response by DdaA, we mutated important ssDNA-binding tyrosine residues to create Y173S and Y200S alleles and measured aciT expression in these strains. In the absence of DNA damage, both Y173S and Y200S mutant strains expressed aciT at similar levels as wildtype DdaA, demonstrating that these mutants retain promoter-binding capability (Fig. 5D). In contrast, aciT expression in the Y173S and Y200S mutant strains in the presence of MMC was significantly reduced compared to the strain expressing wildtype DdaA, supporting a role for these residues in sensing ssDNA to promote increased transcription when DNA damage occurs (Fig. 5D). Importantly, none of these point mutations affected DdaA expression or stability (Supplementary Fig. S6). Together, these results demonstrate that DdaA shares similar functional and regulatory roles as C. crescentus DriD in promoting DDR gene expression. The evolutionary distance between C. crescentus (an Alphaproteobacterium at the phylum level) and A. baylyi (a Gammaproteobacterium) highlight that this mechanism for activating DDR pathways upstream of RecA might be a more widespread phenomenon than is currently appreciated.

DdaA activates the DDR pathway during natural transformation

It was previously reported that the high rates of DNA uptake by A. baylyi during natural transformation trigger activation of the DNA damage response pathway (and cell filamentation) through an unknown mechanism [15]. Acinetobacter baylyi employs a robust type VI secretion system (T6SS) that it uses to rapidly kill other species including E. coli, releasing large quantities of DNA into the environment that can then be internalized via T4P [58]. Incoming DNA from natural transformation enters the cytoplasm as ssDNA, and this was shown to stimulate the expression of DDR genes including recA in A. baylyi [15].

We reasoned that if DdaA senses ssDNA to activate AciT and RecA expression upon DNA damage, it might function similarly to sense incoming ssDNA from natural transformation to activate DNA repair genes and phenotypes upon internalization of high concentrations of DNA. Using the fluorescent recA transcriptional reporter (designed similarly to the one used to show natural transformation activates the DNA repair pathway previously [15]), we tested whether mutants lacking ddaA would still upregulate recA expression under similar conditions (Fig. 6A). Acinetobacter baylyi strains expressing the P_recA-mRuby3 reporter were mixed in a 1:1 ratio with E. coli, resulting in almost complete elimination of all E. coli cells within 4.5 h (Fig. 6B). The parent strain exhibited a similar increase in P_recA-mRuby3 fluorescence and increased cell elongation upon incubation with E. coli as reported previously (Fig. 6B and C) [15]. A ∆ddaA mutant exhibited no increase in fluorescence intensity, similar to a ∆comP mutant that no longer produces the T4P required for DNA uptake (Fig. 6B and C). Interestingly, when mixed with E. coli ∆ddaA mutants appear morphologically similar to ∆umuDAb mutants in the absence of DNA damage, suggesting that ssDNA entering the cell after T4P-mediated DNA uptake is sufficient to relieve UmuDAb repression of other DDR genes (Fig. 6B and C). In addition to a lack of recA expression, ∆comP mutants also appear morphologically identical both in the presence and absence of E. coli, suggesting the ∆ddaA mutant phenotype after incubation with E. coli is directly due to DNA uptake. It is possible that DdaA-activated recA expression in these experiments might arise from DNA damage introduced by failed recombination attempts with heterologous E. coli DNA rather than from DNA uptake per se. To test whether the heterologous nature of E. coli DNA could explain DdaA-activated recA expression, we incubated the same P_recA-mRuby3 strains with purified A. baylyi gDNA and performed fluorescence plate reader assays. Fluorescence profiles from strains mixed with purified gDNA exhibited the exact same trends as when mixed with E. coli: while the parent strain exhibits a significant increase in fluorescence, the ∆ddaA and ∆comP mutants remain unchanged (Fig. 6D). These data demonstrate the mechanism of natural transformation-induced stimulation of DNA repair pathway genes including aciT and recA is due to DdaA activity.

Figure 6.

DdaA activates the DDR pathway upon DNA uptake during natural transformation. (A) Schematic of experimental set up. Exponential cultures of E. coli and A. baylyi strains carrying the P_recA fluorescence reporter were co-incubated for 4.5 h before cells were resuspended in LB and imaged. (B) Representative microscopy images of indicated strains in the presence and absence of E. coli. Scale bars, 5 µm. (C) Single cell fluorescence intensity measurements of strains shown in panel (B) in the presence and absence of E. coli. (D) Transcriptional fluorescence reporter data for indicated strains containing the P_recA fluorescence reporter grown in the presence and absence of homologous gDNA. Bars indicate the mean ± SD. Each data point represents an independent, biological replicate. Statistics were determined using Sidak’s multiple comparisons test. ****P <.0001; ***P <.001; **P <.01; ns, not significant.

Discussion

In this work, we identify the WYL transcriptional regulator DdaA as a DDR activator in Acinetobacter species. As mentioned above, another group independently identified DdaA as a DDR activator in the same clade of bacteria [10]. Both groups also independently identified the DNA-binding sequence of DdaA in DDR gene promoter regions, and both groups used RNAseq to establish the DdaA regulon under DNA damage-inducing conditions. The similarities between our findings validate and highlight the reproducibility of our studies, especially as our groups arrived to the same conclusions via different methodologies.

Our study employed a suppressor mutation approach to identify DdaA as a regulator of aciT expression, and RNAseq on ddaA mutants identified ddaA-regulated genes. Analysis of promoter sequences from genes most affected by the deletion of ddaA subsequently led to the identification of a conserved promoter binding sequence that was highly similar to the C. crescentus DriD-binding region. In contrast, Song et al. discovered DdaA by first identifying the DNA binding sequence conserved in DDR-regulated genes by bioinformatic analysis. The conserved sequence was then used for DNA pull-downs followed by mass spectrometry identification, revealing DdaA as a DDR promoter-binding protein. These DNA pulldowns also identified the direct binding targets of DdaA, and in vitro analyses by both groups confirmed binding of DdaA to the same consensus sequence. Both studies developed fluorescence-based reporters to assess aciT expression when the consensus DdaA binding motif is mutated, showing the importance of this binding motif for promoting expression in vivo. Song et al. further found that DdaA likely interacts with components of RNA polymerase in vitro, supporting the model that DdaA acts as a transcriptional activator by facilitating recruitment of RNA polymerase to target promoters. Our groups both used AlphaFold3 modeling to predict DdaA forms a homodimer and bioinformatically compared the sequence of DdaA to other WYL transcriptional regulators to identify conserved residues important for function. In our work, we further tested the importance of conserved residues in DdaA in vivo, highlighting structural similarities to DriD and isolating important functional properties of DdaA domains.

In RNAseq data from both studies, the CRISPR locus was found to contain some of the most significantly downregulated genes in ddaA mutants. Song et al. explored the functional consequences of this regulation, focusing on the role of DdaA in activating phage defense systems [10]. In contrast, we explored the role of DdaA in activating the DDR response due to T4P-mediated DNA uptake during natural transformation, finding that DdaA acts as a platform for cross-talk between these two pathways. Recent work proposes that an inverse relationship between prophage maintenance and competence induction in Streptococcus pneumoniae creates an “evolutionary arms race” between prophage and competence machinery [59]. The connection provided by DdaA between the DDR pathway, natural competence, and phage defense systems may represent a strategy by which cells can fine-tune their response to broad types of genetic assault to maintain better control over genome integrity. The parallel studies between Song et al. and the work presented here together uncover a new mechanism of transcriptional regulation that intersects multiple DNA regulatory networks and provides a more holistic picture of DDR regulation in the Acinetobacter clade than either manuscript alone [10].

An interesting aspect of this work reveals population heterogeneity of cell morphologies at the single cell level. While some cells presented extreme filamentation phenotypes, others were much less filamented with lengths similar to the no MMC condition (Fig. 1). Similarly, single cells exhibited varying degrees of recA expression upon natural transformation (Fig. 6). The variability in cell lengths in all Acinetobacter species suggest there may be additional regulatory mechanisms that influence DdaA activity to promote population heterogeneity that is conserved in the Acinetobacter clade. The differences in AciT-dependent cell filamentation between different species are also intriguing. Acinetobacter nosocomialis cells exhibited the most extreme filamentation phenotype, reaching lengths > 30 µm, while A. radioresistens cells appear wider and shorter, with no cells reaching lengths >20 µm. The activity of the peptidoglycan synthesis protein PBP1a has been shown to directly contribute to cell lengths and morphologies in A. baumannii strains [60], and it is possible that differences in PBP1a activity may also influence AciT-mediated cell filamentation.

While it is now clear that AciT acts as a DDR-activated cell division inhibitor, its mechanism of cell division inhibition remains unknown. The 3xFLAG peptide appended to the C-terminus of AciT causes constitutive filamentation even in the absence of DNA damage and has proven to be a valuable tool for dissecting AciT regulation. Future work utilizing this allele and interrogating the mechanism of constitutive activation will provide insight into AciT activity. Likewise, the relationship between 3xFLAG-based constitutive activation of AciT and further activation by deletion of umuDAb will reveal important aspects of AciT regulation. Because umuDAb deletion does not affect AciT protein levels, it is possible that it represses expression of a factor important for post-translational modification of AciT. Alternatively, the subtle but significant increase in cell lengths in ∆umuDAb mutants appears to be an AciT-independent mechanism of cell elongation. It is possible that UmuDAb represses expression of an additional factor involved in cell division inhibition or elongation that could create an additive effect when combined with constitutively active AciT^3xFLAG to elicit hyper-filamentation.

The diversity of DDR pathways including the ssDNA-sensing PprI system found in Deinococcus and the DNA methylation-dependent pathway found in C. crescentus reflect the importance of SOS-independent pathways in diverse species including those like C. crescentus which also possess a classical SOS response in addition to DriD-mediated DDR activation [61, 62]. The role of WYL family of transcriptional regulators in DDR activation is becoming increasingly clear as new research reveals their widespread distribution and conservation in diverse species [47]. While DDR-associated WYL family transcriptional regulators typically activate gene expression, those involved with bacterial defense systems including the CBASS and BREX regulators are generally associated with gene repression [47, 48, 63, 64]. DdaA positively regulates the expression of both the CRISPR defense system and DDR genes, highlighting an interesting mechanism of multi-network activation. The interrogation of WYL family proteins and their role in regulating gene expression continues to provide important insight into how diverse organisms evolve regulatory networks to respond to and ameliorate cellular damage caused by environmental stresses. Detailed mechanistic understanding of bacterial DDR pathways will identify new targets to exploit for drug development to therapeutically treat medically relevant, under characterized bacterial pathogens like those within the Acinetobacter clade.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

National Institutes of Health (R35GM150916 and R35GM157138); Damon Runyon Cancer Research Foundation (DFS6023). Funding to pay the Open Access publication charges for this article was provided by NIH grant funding.

Data availability

RNA sequencing raw data are available on Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra) under accession code PRJNA1348722. A summary of RNA sequencing results is available in Supplementary Dataset 1. An interactive browser-viewable form of these data can also be found here: https://genome.ucsc.edu/s/ckellison/MMC_gene_expression.