An activator regulates the DNA damage response and anti-phage defense networks in Moraxellaceae

Abstract

DNA-damage chemicals, including many antibiotics, often induce prophage induction and phage outbreaks within microbial communities, posing a significant threat to bacterial survival. Moraxellaceae strains are clinically relevant due to their remarkable resistance to antibiotics and radiation. However, the cellular-level regulation mechanisms that underlie their DNA damage response and anti-phage defense remain extensively unexplored. Here, we report a WYL family protein, DdaA, that has replaced the ubiquitous SOS system during the evolution of Moraxellaceae. DdaA functions as an activator and directly regulates the transcriptional networks of both DNA damage response and anti-phage defense genes under conditions of DNA damage stress. Our findings elucidate a pathway that shows how these bacteria enhance their immunity under DNA damage and shed light on controlling the resistance of Moraxellaceae strains in clinical practice.

Graphical Abstract

Graphical Abstract.

Introduction

The Moraxellaceae family comprises a highly diverse group of bacteria belonging to the order Pseudomonadales and has significant clinical relevance [1]. Certain lineages of Moraxellaceae bacteria that are pathogenic to humans, such as Acinetobacter baumannii, have developed multidrug resistance [2]. Others have developed radioresistance, as exemplified by Acinetobacter radioresistens [3, 4]. It has been reported that the mortality rate associated with infections caused by antibiotic-resistant strains of A. baumannii can be as high as 70% [5]. Many antibiotics exert their antimicrobial effects by targeting DNA or DNA replication pathways and have similar mechanisms of action to common DNA damage agents such as mitomycin C (MMC), cisplatin, and ultraviolet (UV) radiation [6]. Most bacteria in the Pseudomonadales, including the clinically important pathogen genus Pseudomonas, utilize the SOS system to combat these DNA damage conditions; though Moraxellaceae is closely related to these bacteria, the DNA damage response (DDR) pathways in Moraxellaceae remain not fully understood. A long-standing question in the field pertains to the identification of DDR regulators of Moraxellaceae [7, 8], given their lack of the three known bacterial DDR systems, namely SOS [9], PprI-DdrO [10], and PafBC [11]. Various studies have reported a complex DDR network in Acinetobacter, with multiple regulators regulating unique DNA damage repair genes [12–14]. For instance, studies have identified UmuDAb and DdrR as regulators of DNA damage-induced genes in the Acinetobacter species. These regulated genes include the error-prone DNA polymerases that lead to mutagenesis, which contribute to the development of antibiotic resistance [13, 15–20]. However, the DDR system regulating canonical DNA damage repair genes in Moraxellaceae remains unclear [13].

Prophages, the dormant state of lysogenic phages that are often integrated into host chromosomes [21], can be activated by environmental stressors, including DNA damage agents such as UV [22], MMC [13], and colibactin [23]. The expression of prophage proteins facilitates prophage genome excision and replication, eventually leading to the lysis of host cells [24]. It has been frequently reported in bacteria, including Moraxellaceae strains, that prophages are induced by DNA damage agents [9, 13], posing an additional threat to cells [23, 25]. In addition, the induced phages can infect closely related strains in the same habitat, resulting in phage epidemics [26]. Bacteria have evolved several antiviral strategies, such as restriction-modification (RM) systems and clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated protein (Cas) systems [21, 27]. The upregulation of Cas genes under DNA damage conditions has been reported in Acinetobacter strains [13, 28], suggesting a possible connection between the DDR and anti-phage defense. The mechanisms by which bacteria cope with prophage induction after DNA damage remain largely unknown.

In this study, we identified a common regulatory network for DDR in Moraxellaceae. DNA pulldown assays revealed that a newly classified WYL family protein, DDR protein Acinetobacter A (DdaA), binds to the motif identified using a comparative genomics strategy. Evolutionary analyses and transcriptomic experiments verified that this protein functions as a transcriptional activator for both DDR and anti-phage defense genes in Moraxellaceae. Knockout of the ddaA gene caused a severe decrease in the survival rate of A. baumannii under DNA damage and phage infection treatments. Collectively, these studies reveal a unique DDR pathway regulated by DdaA in Moraxellaceae.

Materials and methods

The identification of the binding motif in Acinetobacter

Potential DDR genes identified in a previous study [13] (Supplementary Table S1) were downloaded from NCBI Entrez [29] using Biopython scripts [30]. A pipeline, Novel-Sample, was developed to search for potential binding motifs. DIAMOND (blastp -b10.0 -k500 -p 16) was used to search for protein homologs of each gene against the protein sequences of 88 representative Acinetobacter species. The −200 to 0 regions from the translation start site of the genes were extracted using a Biopython script, concatenated into one file, and transferred to MEME [31] (-dna -minw 12 -maxw 25 -mod zoops -evt 1e-20 -pal -p 32 -nmotifs 3).

Bacterial strains, growth conditions, and genomic sequencing

Acinetobacter radioresistens ATCC 43998 was obtained from the Shanghai Bioresource Collection Center. The phage PhAb24 and bacterial strains A. baumannii ATCC 17978, ATCC 19606, and ZWAb067 were obtained from the laboratory stock. Acinetobacter baumannii ZWAb067 was isolated from the brain tissue of a dead duck embryo, and its genomic sequence can be found in the NCBI BioProject: PRJNA1146946. Acinetobacter strains were grown at 37°C in Luria-Bertani (LB) broth. The working concentrations of kanamycin and apramycin were 50 μg/ml and 100 μg/ml, respectively.

Acinetobacter baumannii ZWAb067 genomic DNA was extracted using FastPure Bacteria DNA Isolation Mini Kit (DC103-01), which was purchased from Vazyme Biotech Co. Genomic DNA was sequenced using Illumina NovaSeq PE150 (Illumina, San Diego, CA, USA) at Beijing Novogene Bioinformatics Technology Co., Ltd. The resulting sequencing reads were assembled into a complete sequence using Unicycler release 5.0 [32].

DNA pulldown assay

Cultures of A. radioresistens ATCC 43998 and A. baumannii ATCC 17978 were grown in 500 ml of broth until the OD₆₀₀ reached 1.0. Cells were harvested by centrifugation at 8000 × g for 5 min. The resultant supernatants were discarded, and the cell pellets were resuspended in 40 ml wash buffer (20 mM Tris–HCl, pH 7.5, 2 mM ethylenediaminetetraacetic acid, 150 mM NaCl, 0.05% Triton X-100). Two milliliter aliquots were then collected in a 2-ml EP tube by centrifugation at 8000 × g for 5 min. Subsequently, the cell pellets were lysed using B-PER^® Bacterial Protein, following the instructions provided by Thermo Fisher.

The 3′-biotin-TEG-labeled single-stranded DNA (ssDNA) and complementary ssDNA were synthesized by Sangon Biotech (Shanghai), based on the promoter region of the recA gene of A. radioresistens ATCC 43998. The two DNA strands were then annealed in an annealing buffer (30 mM Tris–HCl, pH 7.5, 100 mM KCl) through thermal annealing from 90°C to 4°C. Next, 20 μl of 10-μM biotin-TEG-labeled double-stranded DNA (dsDNA) was incubated with 50 μl of MyOne™ Streptavidin C1 for 15 min. The DNA sequences are listed in Supplementary Table S2. Four hundred microliters of cell lysate and DNA-coated beads were mixed. When purified DdaA was needed for the cofactor pulldown, 15 μg of DdaA was added to the solution. After gentle mixing for 30 min, the beads were collected using magnets and washed five times with 500 μl wash buffer. Finally, the beads were resuspended in 30 μl wash buffer and analyzed by routine sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Silver staining was performed using a Fast Silver Stain Kit (Beyotime Biotechnology, China).

Mass spectrometry

The target protein band was excised from the gel, and the peptides were extracted after enzymatic digestion. The samples were then analyzed using mass spectrometry Q-Exactive at the Hangzhou Precision Medicine Research Center (Hangzhou). Protein identification was performed using the Thermo Scientific Proteome Discoverer 2.1 software.

Protein expression and purification

ddaA genes from A. radioresistens ATCC 43998, A. baumannii ATCC 17978, P. aquimaris strain KCTC 12254, H. influenzae strain 65290_NP_Hi3, C. arsenatis strain DSM 11915, and G. subterraneus strain Red1 were cloned or synthesized by Beijing Tsingke Biotech Company and inserted into the plasmid vector pET-28a backbone (Supplementary Table S3). A 6× His-tag was added at the 5′-end of ddaA for affinity chromatography. The vectors were transformed into Escherichia coli BL21 (DE3) cells for protein expression. The expression cells were grown in LB broth at 37°C, and when the OD₆₀₀ reached 0.6, isopropyl-β-d-thiogalactopyranoside was added to a final concentration of 0.2 mM. Following 5 h of incubation, the cells were harvested by centrifugation at 8000 × g for 5 min and washed with His-A buffer (150 mM NaCl, 20 mM Tris–HCl, pH 7.5, and 5% w/v glycerol). The cells were resuspended in 40 ml of His-A buffer and lysed by sonication, followed by centrifugation at 16 000 × g for 30 min to remove cell debris. The supernatant was applied to an AKTA Purifier system and loaded onto a HisTrap HP column after equilibration with the His-A buffer. After washing with His-B buffer (150 mM NaCl, 20 mM Tris–HCl, pH 7.5, 5% w/v glycerol, 50 mM imidazole), the protein was eluted with His-C buffer (150 mM NaCl, 20 mM Tris–HCl, pH 7.5, 5% w/v glycerol, 250 mM imidazole). In instances where desalting was needed for downstream experiments, the buffer was returned to His-A using a HiPrep 26/10 Desalting column.

Electrophoretic mobility shift assay

Site mutants of the recA promoter sequence from A. radioresistens ATCC 43998 labeled with 3′FAM were synthesized by Sangon Biotech (Shanghai) and annealed in an annealing buffer (30 mM Tris–HCl, pH 7.5, 100 mM KCl) through thermal annealing from 90°C to 4°C. Twenty microliters of 50 nM 3′FAM-labeled DNA (in annealing buffer) and 5 μl of 4-μM DdaA (in His-A buffer) and its homologs were mixed and incubated for 15 min. Thereafter, 5 μl of the mixture was loaded onto 12% native-PAGE and electrophoresized at 140 V for 30 min. The gels were imaged in the fluorescence mode (FAM) using a Typhoon FLA 9500 (GE). The DNA sequences are shown in Supplementary Table S2.

Generation of ddaA-knockout and complemented strains

ddaA of A. baumannii ZWAb067 was knocked out using a CRISPR–Cas-based strategy [33]. The specially designed sgRNA expression vector pSGAb-km and ssDNA were introduced into ZWAb067::pCasAb-apr electrocompetent cells. Knockout of ddaA was confirmed by Sanger sequencing, and the pCasAb and pSGAb vectors were removed using sacB selection.

DdaA of A. baumannii ATCC 17978 was knocked out using a pT18-sacB-based knockout system, following a method similar to that introduced by Amin et al. [34]. In short, the upstream and downstream sequences (∼500 bp, respectively) of ddaA (locus tag: HKO16_RS19275) from A. baumannii ATCC 17978 were cloned into the sacB+ suicide vector pT18-sacB (with a tetracycline-resistant gene) using In-Fusion^® technology to generate the ddaA-knockout vector pT18-sacB-ddaA. The product of the In-fusion cloning reaction was transformed into E. coli S17-1 λpir chemically competent cells. The colonies grown on the tetracycline LB plate were selected, and successful transformation was confirmed by sequencing the PCR products using M13-F and M13-R primers. Acinetobacter baumannii ATCC 17978 and E. coli S17-1 λpir:: pT18-sacB-ddaA were grown in LB broth with and without Tc antibiotic, respectively, till OD₆₀₀ reached 0.8. The 700 μl A. baumannii culture was mixed with 700 μl E. coli culture, and the mixed culture was incubated overnight at 37°C. The culture was spread on a tetracycline + ampicillin LB plate and incubated overnight at 37°C. The colonies on the plate were selected and transferred to liquid LB broth containing 10% sucrose for sacB selection. Successful ddaA knockout was validated by DNA sequencing of PCR product amplifying ddaA locus. To obtain the complemented strain, PCR-amplified wild-type ddaA with its promoter region was cloned into the plasmid pWH-1266 (apramycin resistance), which was then introduced into the ddaA-knockout strain by electrotransformation.

Chromatin affinity precipitation sequencing

Similar to chromatin immunoprecipitation sequencing (ChIP-seq), chromatin affinity precipitation sequencing (ChAP-seq) is a genomic technique used to identify DNA sequences bound by specific proteins. The primary distinction between these methods lies in the strategy for isolating protein–DNA complexes: ChAP-seq employs tag-based affinity purification to capture interactions involving a tagged protein of interest, whereas ChIP-seq relies on antibody-mediated immunoprecipitation to target endogenous proteins. This difference enables ChAP-seq to avoid antibody-dependent biases while maintaining high specificity for the protein of interest [35–37]. Briefly, pWH1266-DdaA, fused with an N-terminal His-tag, was introduced into the ddaA knockout strain ZWAB067. The bacterial cells were grown in LB broth at 37°C overnight. The cells were harvested by centrifugation at 8000 × g for 5 min and washed with His-A buffer (150 mM NaCl, 20 mM Tris–HCl, pH 7.5, 5% w/v glycerol). The cells were resuspended with 40 ml of His-A buffer and lysed by sonication, followed by centrifugation at 16 000 × g for 30 min to remove cell debris. The supernatant was applied to an AKTA Purifier system and loaded onto a HisTrap HP column after equilibration with His-A buffer. After washing with His-B buffer (150 mM NaCl, 20 mM Tris–HCl, pH 7.5, 5% w/v glycerol, 50 mM imidazole), the protein was finally eluted with His-C buffer (150 mM NaCl, 20 mM Tris–HCl, pH 7.5, 5% w/v glycerol, 250 mM imidazole). This solution contains the DNA that is co-precipitated with DdaA protein.

The input control DNA and DNA co-precipitated with DdaA protein were purified using the Wizard^® SV Gel and PCR Clean-Up System, following the manufacturer’s instructions. The final DNA library (generated by EzyNGS DNA Library Construction Kit) was obtained after size selection (200–400 bp, BeyoMag™ DNA Size Selection Magnetic Beads) and PCR amplification (100 ng purified DNA as template). The qualified libraries were sequenced on the Illumina NovaSeq platform to generate 150-bp paired-end reads at KAITAI-BIO.

Sequencing reads were processed using TrimmomaticPE [38] and mapped to the representative genome of A. baumannii ZWAb067 using HISAT2 [39]. DNA peak enrichment was called by MACS3 (with q-value < 0.01) [40] and visualized using the Integrative Genomics Viewer [41].

Phylogenetic relationship between DdaA and other studied WYL transcriptional regulators

A protein BLAST database encompassing all completed RefSeq bacterial genomes was constructed. We ran PSI-BLAST [42] (E-value 1e−60, iterates until convergence) against the database to identify DdaA, PafB, PafC, DriD, BrxR, and CapW homologs. Protein sequences with <260 aa or >400 aa were removed, as DdaA of A. baumannii ATCC 17978 has a length of 335 aa, PafB of M. tuberculosis H37Rv has a length of 332 aa, PafC of M. tuberculosis H37Rv has a length of 316 aa, DriD of C. vibrioides ATCC 19089 has a length of 331 aa, CapW of P. aeruginosa PA17 has a length of 299 aa, and BrxR of Acinetobacter sp. NEB 394 has a length of 288 aa. Homologous sequences were aligned using MAFFT v7.490 [43] with the default parameter and trimmed using trimAI v1.4 [44] with the default parameter. A phylogenetic tree was generated using FastTree 2.1.11 [45] with default parameters. The tree was visualized using iTOL v5 [46].

Evolutionary analysis of the DdaA DDR networks

In total, 402 genome sequences from the family Moraxellaceae and its closely related species were selected for the evolutionary analysis [based on the Genome Taxonomy Database (GTDB) release 214 [47], Supplementary Fig. S1, and Supplementary Table S4]. The genome sequences were annotated by Prokka 1.14.6 [48]. Protein orthologous groups were determined using Orthofinder 2.5.5 [49] with default parameters. The MSA of the 120 marker genes provided by the GTDB was used for the phylogenetic analysis of the 402 genomes. BMGE 1.12 [50] was used to trim the MSA with default parameters. A phylogenetic tree was generated using FastTree 2.1.11 [45] with default parameters.

To determine the DdaA orthologs, we conducted PSI-BLAST on NCBI using the DdaA from A. baumannii sequence against the nr_clustered database (re-analyzed on 22 January 2024). The “Expect threshold” and “PSI-BLAST incl. threshold” were set to 1e−60, the “Max target sequences” parameter was set to 20 000, and the “word size” was set to 3. Highly similar sequences were removed using CD-HIT (-c 0.8 -T 6 -n 2 -d 50). All Orthofinder-determined DdaA ortholog (DdaA-1 hereafter, Supplementary Text) sequences were added to the PSI-BLAST results to perform phylogenetic analysis. Protein sequences with <260 aa or >400 aa were removed, as DdaA from A. baumannii has a length of 335 aa. The sequences were aligned using MAFFT v7.490 with the default parameters and trimmed using trimAl with the automated1 mode. The phylogenetic tree was generated using FastTree 2.1.11 [45] with the default parameters. The tree was visualized using iTOL v5 [46].

To reconstruct the transfer of DdaA genes in the evolutionary history of Moraxellaceae, the protein sequences of DdaA-1 were aligned using MAFFT v7.490 and trimmed using BMGE 1.12 with default parameters. The phylogenetic tree was generated using IQ-Tree multicore version 2.0.3 [51]. The best-fit model was found as LG + R6. The support value of the phylogenetic tree was analyzed using the ultrafast bootstrap (UFBoot) feature 5000 times. ALEml_undated algorithm of the ALE [52] package was used to determine the duplications, losses, intra-LGT (transfer within the sampled genome set), and origination events of DdaA-1. The species tree of the 402 genomes and the DdaA-1 bootstrap trees generated above were used as input. The results were summarized using ALE helper (github.com/Tancata/phylo/tree/master/ALE).

Identification of genes that are regulated by DdaA, LexA, or UmuD

The promoter sequences were defined as −200 to 0 bp from the translation start site. A Biopython script was used to extract promoter sequences. All genes encoded in the 402 genomes were searched using the FIMO program [53]. A hit with a P-value lower than 5e−6 was considered to have a DdaA binding motif in the promoter, indicating that DdaA regulates the transcription of the gene. CRISPR–Cas or RM systems are subject to regulation by DdaA if a DdaA binding motif is present in the promoter of any gene belonging to these systems. The same process was applied to identify genes regulated by LexA and UmuD proteins.

CRISPR and phage determination

The CRISPR–Cas system was identified using CRISPRCasFinder 4.3.2 [54], and the spacer sequences were extracted from the result files. The prophage sequences were detected using Phispy [55] 4.2.6. BLASTN was used to identify spacer and target phage pairs. To identify more potential spacer-phage pairs, we did not restrict the target phage DNA to have a PAM sequence. Spacer sequences were used as queries, and all phage DNA sequences were used as subjects. Only results with no more than one mismatch were considered true spacer-target pairs (Supplementary Table S5).

Phenotyping of DNA damage treatments

Wild-type A. baumannii, ddaA-knockout strain, and complemented strain were grown on LB plates overnight at 37°C. Fresh colonies were picked up into liquid LB broth and incubated at 37°C until an OD₆₀₀ of 0.6 was reached. Subsequently, each 1 ml culture was transferred to a 1.5 ml EP tube and centrifuged at 4000 × g for 5 min. The supernatant was discarded. To assess cell viability, cells were resuspended in LB broth supplemented with varying concentrations of MMC or exposed to UV radiation using a CL-1000 Ultraviolet Crosslinker (snap exposure at 5–20 J/m²). After incubation at 37°C for 30 min, the cell culture was placed on an LB plate and incubated overnight at 37°C.

Transcriptome sequencing and RT-qPCR

Acinetobacter baumannii strains were grown on the LB plate overnight. Fresh colonies were picked up into liquid LB broth and incubated at 37°C until OD₆₀₀ reached 0.6. Each 2 ml culture was spread on a clean glass surface to ensure uniform exposure of the bacterial cells to UV radiation. The UV radiation group was treated with 15 J/m² of UV radiation when testing DDR or 5 J/m² when testing anti-phage defense ability, while the control group was maintained at room temperature. In the case of MMC-related experiments, fresh colonies were transferred to liquid LB broth and grown at 37°C until the OD₆₀₀ reached 0.6. The MMC was added to the experimental group culture at a final concentration of 1 μg/ml for testing DDR or 0.5 μg/ml for testing anti-phage defense ability. Lower doses of UV or MMC were used to test the anti-phage defense ability and minimize the inhibition effect on the growth of bacterial cells. The cultures were then incubated at 37°C for 15 min. Following this, the culture was immediately frozen in liquid nitrogen and stored at −80°C.

RNA sequencing procedure: The total RNA was extracted using the RNAprep Pure Cell/Bacteria Kit (RNAprep Pure, DP430) and purified using the RNAClean XP Kit (Beckman Coulter). The remaining DNA was digested using an RNase-Free DNase Set (Qiagen). Libraries were constructed using the U-mRNAseq Library Prep Kit AT4221 (KAITAI-BIO) with the Ribo-off rRNA Depletion Kit (Bacteria) (Vazyme) and then sequenced on the Illumina NovaSeq platform to generate 150-bp paired-end reads at KAITAI-BIO. Sequencing reads were processed using TrimmomaticPE [38] and mapped to the representative genome of A. baumannii ATCC 17978 (RefSeq accession: GCF_014672775.1) using HISAT2 [39]. The expression matrix was calculated using StringTie [56] and DESeq2 [57] to determine differentially expressed genes.

Reverse transcription quantitative polymerase chain reaction (RT-qPCR) procedure: RNA extraction was performed using the TransZol Up Plus RNA Kit, and the subsequent reverse transcription reaction was performed using the HiScript III RT SuperMix kit (Vazyme) according to the manufacturer’s instructions. The purified RNA was then added to an RNase-free centrifuge tube containing a gDNA wiper mix, which was subsequently incubated at 42°C for 2 min to remove any residual genomic DNA. Subsequently, a 5× Hiscript III enzyme mix was added to the mixture from the previous step. The reverse transcription reaction was performed at 37°C for 15 min, followed by 85°C for 2 min. The complementary DNA was then used as the template for RT-qPCR with gene-specific primers. Agilent Mx3005P automatically generates the Ct value. rpoB was used as the reference gene. The primer sequences used for RT-qPCR are listed in Supplementary Table S2. All RNA sequencing and RT-qPCR experiments were conducted in three biological replicates.

Quantifying promoter effectiveness using the eGFP reporter system

To examine the transcriptional activity of DdaA-regulated promoters in the presence or absence of the DdaA protein, we measured eGFP fluorescence intensity using a fluorescence microscope. pWH-1266-derived vectors were designed to contain an eGFP gene after the motif-containing promoter of aciT or uvrA from A. baumannii. For comparison, vectors containing mutated motifs were constructed. To establish a positive background control, both vectors also carried the mCherry gene following the promoter of ompA, which is not regulated by DdaA but is consistently expressed. The first vector was introduced into A. baumannii ATCC 17978 wild-type and ddaA knockout strains, whereas the second was introduced into the wild-type strain. Bacterial cells were imaged using a Nikon Ti2 fluorescence microscope. The expression of mCherry was stable and exhibited no significant differences between the transformed strains, demonstrating the same conditions in vivo. A flow cytometer was used to validate the microscopy results. The overnight bacterial culture was analyzed using a CytoFLEX flow cytometer (Beckman Coulter, Inc., USA). The data were analyzed using FlowJo, and cells with an FITC-A value higher than 1000 were considered to highly express the eGFP protein.

Phage purification

Phage PhAb24 [58] stock was added to 5 ml of exponentially growing wild-type A. baumannii ATCC 17978. The cultures were incubated overnight at 37°C. The supernatant from the overnight culture was then filtered through a 0.22-μm membrane and stored at 4°C. The titer of PhAb24 was determined using a soft-agar overlay assay with wild-type ATCC 17978.

The growth assays and plaque assays for testing A. baumannii strains’ resistance to phage infection

A series of strains originating from A. baumannii ATCC 17978 harboring the RM system from A. baumannii ATCC 19606 was constructed. To preserve the gene neighborhood of the RM system, we inserted genes from the ATCC 19606 defense locus 6 (PADLOC-DB v2.0.0) into the pSGAb backbone [33]. The ATCC 17978 wild-type (with empty pSGAb backbone), ATCC 17978 harboring the RM system from ATCC 19606 (abbreviated as strain RM-19606), strain ATCC 17978 harboring the RM system with a mutated DdaA binding motif (abbreviated as strain RM-19606 with mutated motif), and ATCC 17978 ddaA-knockout strain harboring the RM system (abbreviated as strain △ddaA RM-19606) were grown at 37°C until OD₆₀₀ reached 0.6. If needed, MMC was added to the cultures at a final concentration of 0.5 μg/ml. The cultures were incubated at 37°C for 30 min to activate the DdaA network, and 2 ml of culture was harvested by centrifugation at 7000 × g for 1 min. The supernatant was discarded, and the cell pellet was resuspended in 2 ml of LB broth. For growth assays, 200 μl of resuspended culture was transferred to 5 ml of fresh LB broth, and PhAb24 was added to the culture at a multiplicity of infection (MOI) of ∼0.02. OD₆₀₀ was automatically recorded every 5 min using a Scientz^TM microbial growth curve analyzer (MGC-200). Three biological replicates were performed. For plaque assays, 200 μl of the resuspended culture was transferred to warmed 3-ml soft agar LB broth and placed on LB agar to form an overlay layer. The plate was incubated at 37°C for 15 min to allow the surface to dry. Serially diluted PhAb24 was then placed on the surface. MMC concentrations used in the plaque assays were 0.5 μg/ml, respectively.

Assays for testing the exogenous DNA removal ability of the A. baumannii native CRISPR I-F system

To perform the plasmid interference assay, the pSGAb plasmid [33] was used as the backbone and modified with PCR-based point mutation technology to insert CRISPR–Cas target sequences, forming pTarget. pTarget contains two target DNA sequences recorded by the A. baumannii ZWAb067 native CRISPR array. The target DNA can be found in the prophage region of A. baumannii strain AB046 and the prophage region of Acinetobacter pittii strain FDAARGOS 1397; thus, we used pSG-target to mimic the invasion of the prophage induced by DNA damage. Spacer information and point mutation primers are listed in Supplementary Table S2. To perform the plasmid interference assay, the pControl and pTarget plasmids (both could provide kanamycin resistance to bacteria) were diluted to 100 ng/μl, and 1.5 μl of diluted DNA was transformed into 50 μl of electrocompetent cells of wild-type A. baumannii ZWAb067, ddaA-knockout strain, or complemented strains by electrotransformation. After incubation at 37°C for 1 h, the transformed cells were plated on kanamycin-LB agar, and the colonies were manually counted after incubation at 37°C for 16 h.

We also used qPCR to test the relative abundance of pControl to pTarget in overnight cultures. The pControl and pTarget plasmids were mixed at a 1:1 ratio with a final DNA concentration of 50 ng/μl. Electrocompetent cells of wild-type A. baumannii ZWAb067, ddaA-knockout strain, complemented strain, or wild-type A. baumannii ATCC 17978 were transformed by introducing 1.5 μl of diluted DNA into 50 μl of electrocompetent cells. The transformed cells were transferred to 5 ml of fresh LB broth and incubated at 37°C for 1 h. Kanamycin was then added to the cell cultures at a final concentration of 50 μg/ml. The cell cultures were incubated with shaking at 37°C for 16 h, and the cells were harvested by centrifugation at 16 000 × g for 5 min. The supernatant was discarded, and the cell pellet was resuspended in 2 ml of sterile ddH₂O. The cell pellet was washed again with 2 ml of sterile ddH₂O and resuspended in 500 μl of ddH₂O. The resuspended cultures were then placed in boiling water for 10 min and used as the direct qPCR template. The qPCR primers were specially designed to distinguish between pControl and pTarget, and are listed in Supplementary Table S2.

Statistical analysis

Data are presented as mean ± standard deviation (SD) from three independent biological replicates. Statistical analyses were performed using the “agricolae” R package. Significant differences between the experimental groups were analyzed using one-way analysis of variance (ANOVA), followed by the LSD test for comparisons. Statistical significance was set at P < .05.

Results

A conserved motif is involved in a putative DDR system in Acinetobacter

We compared the promoter regions of DDR genes from Acinetobacter to identify potential DDR regulatory systems. Based on previously published transcriptomic data investigating the DDR in Acinetobacter [12, 13] (Supplementary Table S1), the promoter regions of 63 genes involved in DDR were analyzed in 88 representative Acinetobacter genomes (Fig. 1A and Supplementary Table S6). A conserved and palindromic DNA motif (CGWCA-N5-TGWCG, Fig. 1B) was identified in the promoter region of >15 genes, including recA, which is essential for homologous recombination in DNA damage repair; aciT for ciprofloxacin tolerance; ssb for ssDNA binding; uvrA/C for nucleotide excision repair; gyrA/B for DNA supercoil formation; and dGTPase-ruvA-ruvB operon associated with Holliday junction processing, suggesting the existence of an unreported DDR regulation system in Acinetobacter. Moreover, the Cas gene clusters were found to contain this motif within their promoter regions, suggesting that anti-phage defense may also be regulated in this system (Supplementary Table S7).

Figure 1.

Discovery of the regulator of a putative DDR system and its binding motif in Acinetobacter. (A) Schematic bioinformatic workflow employed to identify the common binding motif in the promoter regions of DDR genes in Acinetobacter. (B) Binding motif of a potential regulator of the DDR system in Acinetobacter. Logo was generated by MEME software [31] analyzing the promoters of DDR genes in Acinetobacter. (C) DNA pull-down assay of A. radioresistens ATCC 43998 with recA promoter. The protein marker is PageRuler™ 26616 in lane 1. Lane 2 is blank. Lane 3 is the negative control, where no DNA was added to the reaction. Lane 4 shows the dsDNA pulldown results. Lane 5 shows the ssDNA pulldown results. The arrow indicates a protein band unique to lane 4. (D) EMSA of DdaA with the recA promoter motif of A. radioresistens ATCC 43998. The mutated bases are shown in Supplementary Fig. S2A. (E) DdaA ChAP-seq results from A. baumannii ZWAB067. The distribution of predicted DdaA-binding motifs and ChAP-seq peaks is labeled. The right panel exhibits the Venn diagram showing the degree of concordance between DdaA binding site predictions and ChAP-seq experimental data. (F) Maximum-likelihood tree of WYL domain-containing proteins, with each cluster designated by a distinct color. Branches of DdaA homologs are represented by gray lines, and those that have been validated by EMSA, as shown in panel (G), are indicated by red lines. Ab = A. baumannii ATCC 17978. Pa = Psychrobacter aquimaris KCTC 12254. Hi = Haemophilus influenzae 65290_NP_Hi3. Ca = Chrysiogenes arsenatis DSM 11915. Gs = Geoalkalibacter subterraneus Red1. (G) EMSA of DdaA homologs and the promoter sequence of recA from A. radioresistens ATCC 43998.

A potential regulatory protein binds specifically to this motif

To identify the regulator of this putative DDR regulator, we performed a pulldown experiment by incubating the cell lysate of the radioresistant strain A. radioresistens ATCC 43998 with beads coated with the promoter DNA of its recA gene. After electrophoretic separation of the enriched proteins, silver staining revealed a protein band in the 35–40 kDa size range in the lane where the recA promoter was added. However, it did not appear in the lane that used ssDNA or in the negative control lane where no DNA was added (Fig. 1C). Mass spectrometry results for this specific band identified a molecular weight of 38.685 kDa for this protein, corresponding to UniProt accession ID: A0A7U9PMQ2 (Supplementary Table S8). Since this protein is likely to be involved in regulating DDR genes in Acinetobacter, we hereafter refer to it as DdaA.

The specific interaction between the protein and the motif was further confirmed by in vitro electrophoretic mobility shift assays (EMSAs) using DdaA from A. radioresistens and the motif-containing promoter region of recA. The band of DdaA binding to the promoter containing an intact motif (WT) showed a significant mobility shift (Fig. 1D). Mutations in any of the four conserved base pairs in the motif (Supplementary Fig. S2A) drastically impaired or decreased binding affinity, verifying the critical role of these sites in specific interactions with DdaA (Fig. 1D).

Chromatin affinity precipitation sequencing (ChAP-seq, a method similar to ChIP-seq) [35, 36, 59] was also conducted to determine the sequence specificity of DdaA binding and identify the DdaA regulon in the A. baumannii genome. DNA co-precipitated with DdaA harbors the DdaA binding motif (Supplementary Fig. S2B and C). Although only 14 DdaA binding motifs were predicted using FIMO, 104 peaks were identified by MACS3 in the ChAP-seq experiment (Fig. 1E). Only 2 predicted DdaA binding sites (with relatively high P-values) did not show enrichment in ChAP-seq peaks, indicating overall high accuracy of DdaA binding site prediction. These differences may indicate that DdaA binds to more genomic regions in the cellular environment than predicted in silico.

DdaA homologs form a large cluster in the WYL protein family

WYL domain-containing proteins are ubiquitous in bacteria and function as regulators of bacterial immunity [60–62] or DDR [11, 63]. Our phylogenetic analysis showed that homologs of the DDR regulators DdaA, DriD, PafB, and PafC formed four independent monophyletic clusters, while homologs of the anti-phage defense regulators CapW and BrxR formed another monophyletic cluster (Fig. 1F). Similar to other class A WYL domain-containing proteins, DdaA consists of an N-terminal wHTH domain, a WYL domain, and a WCX (WYL protein C-terminal extension) domain [64], and size exclusion chromatography indicates that it forms a homodimer in solution (Supplementary Text and Supplementary Fig. S2D and E). The WYL and WCX domains located at the C-terminus of these proteins are thought to have a function in the recognition of negatively charged ligands such as ssDNA [11, 63, 65, 66].

Multiple sequence alignments of the HTH domains of these WYL domain-containing proteins, which are responsible for the recognition and binding to the promoter motifs [64], showed that the amino acid residues for base recognition, such as the two arginine residues in the α3 helix, were highly conserved in the DdaA homologs (Supplementary Fig. S3A). The adjacent positions of DdaA and DriD [63, 67] in the phylogenetic tree and their similar binding motifs (DdaA: CGWCA-N5-TGWCG, DriD: CGAC-N7-GTCG) suggest that they may share a common ancestor (Fig. 1F, Supplementary Text, and Supplementary Fig. S3B). Moreover, in contrast to the other DDR regulators in the WYL family, the first of the two arginine residues in the β4 strand and the β4/β5 loop, which are responsible for ssDNA interaction [64], was substituted with alanine in DdaA, which may lead to different sensing efficiencies or signals during regulation (Supplementary Fig. S3C). DdaA homologs from five other strains were purified and subjected to EMSA with the promoter DNA of recA from A. radioresistens. All these homologous proteins bound to the motif efficiently and exhibited a shifted band, which was weakened or vanished upon motif mutation, indicating the conservation of DdaA in specific binding to the motif (Fig. 1G).

DdaA regulation replaced the SOS system for DDR in the early divergence of Moraxellaceae

A comparison of the species tree of Moraxellaceae and the gene tree of DdaA suggests that the DdaA-based DDR regulation system was gradually established during the early divergence of Moraxellaceae. The species tree of Moraxellaceae and neighboring taxonomic lineages was extracted from the GTDB bacterial tree for analysis (Supplementary Fig. S1). A copy of the ddaA gene (DdaA-1) was horizontally transferred from a gammaproteobacterium to the common ancestor of Alcanivoracaceae and Moraxellaceae (Fig. 2A, Supplementary Figs S4 and S5, and Supplementary Text). Prior to this transfer, DdaA was a putative regulator protein of various genes located near the ddaA gene locus (Supplementary Fig. S4B). However, the functions of these regulated genes remain to be elucidated. After its acquisition, this copy of DdaA was vertically inherited in the two families of Alcanivoracaceae and Moraxellaceae (Fig. 2A and Supplementary Fig. S5). However, in Alcanivoracaceae, the basal lineages of Moraxellaceae, and the genus Moraxella, ddaA genes were sporadically lost and often replaced by copies of DdaA horizontally transferred from various bacteria (DdaA-other) (Fig. 2A, Supplementary Figs S4A and S5).

Figure 2.

Distribution and evolutionary route of DdaA, LexA, and UmuD and the genes they regulate in Moraxellaceae. (A) Phylogenomic maximum-likelihood tree of Moraxellaceae species and neighboring taxonomic lineages is displayed on the left panel. Branches were collapsed at the genus level, except for the outgroups at the family level, based on the GTDB taxonomy [47]. IQ-Tree ultrafast bootstrap values >95% are indicated by black dots on the branches. Evolutionary events of DdaA-1 are labeled on the branches. In the middle panel, the darkness of the black dots indicates the percentage of genomes in the taxonomic clade encoding the protein. The color of the ring surrounding each dot indicates the percentage of genes regulated by DdaA (red), LexA (blue), or UmuD (green) within the taxonomic clade. The right panel illustrates the correspondence between pairs of CRISPR–Cas spacers and protospacers in the prophages. The darkness of the black dots denotes the percentage of genomes encoding CRISPR–Cas or prophages. The darkness of the T bars indicates the percentage of CRISPR-prophage pairs in the total genomes within the genera. As CRISPR–Cas systems fall into many variation groups, in our research, they were clustered into different ortholog groups (OGs). OGs are labeled, and related information is provided in Supplementary Table S9. (B) Representative cases of the locations of DdaA binding motifs in the promoter region of anti-phage defense systems, including Zorya (an anti-phage defense system), RM systems, and subtypes of CRISPR–Cas systems in Moraxellaceae.

Both competition and complementarity between DdaA regulation and the SOS system in the control of DDR genes have been observed in Alcanivoracaceae and basal lineages of Moraxellaceae. In these taxa, LexA regulates essential DDR genes, thereby forming typical SOS systems, as is typical of most bacteria (Fig. 2A and Supplementary Fig. S5). In contrast, DdaA often regulates a small number of DDR- and anti-phage defense-related genes, including recA, recC, sbcC, N-6 DNA methylase, and ftsH. Notably, the essential gene recA was found to be simultaneously regulated by LexA and DdaA in some organisms, showing an evolutionary transition state [68, 69]. This finding suggests that the competition between these two regulators in controlling the DDR system may have originated from recA gene regulation.

It was not until the common ancestor of Fluviicoccus, Acinetobacter, and Psychrobacter, as well as all other genera in that monophyletic clade, came into being that the role of LexA as a global DDR regulator was completely taken over by DdaA (Fig. 2A and Supplementary Fig. S5). The DdaA regulon can regulate up to 20 genes implicated in DDR and anti-phage defense, such asRM system, Zorya system, and various subtypes of CRISPR–Cas systems (Fig. 2B and Supplementary Fig. S5). The composition of the DdaA regulon exhibits intragroup heterogeneity, varying among genus-level subgroups. For instance, gyrB, a DNA gyrase subunit, is subject to DdaA regulation in Psychrobacter but not in most Acinetobacter. Notably, in the basal clade, including Agitococcus lubricus and Fluviicoccus keumensis, DdaA is self-regulated. Self-regulation ceases after the common ancestors of the three well-developed genera of Acinetobacter, Moraxella, and Psychrobacter, at which point the DdaA regulatory network evolved by recruiting many more DDR genes into its regulon, thus becoming a regulator of the DDR.

It was reported that the UmuD protein is involved in the error-prone DNA replication process in Acinetobacter [12, 13] by regulating itself and UmuC [4, 20]. We found that UmuD homologs were frequently found in Moraxellaceae lineages (Fig. 2A, Supplementary Figs S5 and S6, and Supplementary Text), and most of them are self-regulated, but some others are regulated by DdaA. Consequently, the UmuD regulator may operate independently, as observed in A. baumannii ATCC 17978, or may function as a component of the DdaA network.

DdaA activates the transcription of DDR genes, particularly in response to DNA damage

To validate that DdaA is crucially involved in DDR, we conducted phenotype assays with wild-type and ddaA-knockout strains of A. baumannii ATCC 17978, which does not contain known RM or CRISPR–Cas systems (Supplementary Fig. S7A and Supplementary Table S10). Treated by a snap exposure to 5–20 J/m² UV radiation, the viability of the ddaA-knockout strain diminished by 10 to 10 000-fold compared with the wild-type or the complemented strain (Fig. 3A and Supplementary Fig. S7B). The dot blotting assay demonstrated that the ability of the ddaA knockout strain to remove cyclobutane pyrimidine dimers caused by UV radiation [70] was much weaker than that of the wild-type or complemented strain (Supplementary Fig. S7C). Furthermore, the viability of the ddaA-knockout strain was reduced by ∼100-fold following 30 min of exposure to 1, 2, or 4 μg/ml MMC (Fig. 3A and Supplementary Fig. S7D). In the wild-type strain, 118 genes were upregulated and 210 were downregulated following exposure to 15 J/m² UV radiation (Fig. 3B and Supplementary Table S11). In contrast, the ddaA-knockout strain exhibited a different response, with only 29 genes upregulated and 15 genes repressed after UV treatment (Fig. 3C and Supplementary Table S11), exhibiting the dramatic impact of ddaA knockout. These findings underscore the critical role of DdaA in the DDR of A. baumannii. RT-qPCR further supported the transcriptome results (Supplementary Fig. S8A).

Figure 3.

Activation effect of DdaA on DDR. (A) Viability of A. baumannii ATCC 17978-derived strains after treatment with 10 J/m² of UV radiation. Viability of A. baumannii ATCC 17978-derived strains following treatment with 1 μg/ml of MMC. Transcriptomic profiles of A. baumannii ATCC 17978 wild-type (B) and ddaA-knockout strain (C) after 15 J/m² UV radiation. The data presented represent the mean values of three biological replicates. The genes potentially regulated by DdaA are indicated. (D) Validation of DdaA as a transcriptional activator in A. baumannii ATCC 17978-derived strains. The untreated cells were imaged. PaciT = intact aciT promoter. PaciT-M = mutated aciT promoter. Green and red letters indicate the DdaA-binding motif and mutated bases, respectively. (E) DNA–DdaA complex pull-down assay. The protein marker used was PageRuler™ 26616. Lane 1 was the negative control to determine non-specific binding by residual proteins in the purification product. The promoter DNA of recA of A. radioresistens DSM 6976 and the DdaA purification product of A. baumannii ATCC 17978 were mixed and added to the reaction. Lane 3 serves as another negative control to determine the nonspecific DNA binding proteins from A. baumannii cell lysate, in which the promoter DNA of recA of A. radioresistens DSM 6976 and the cell lysate of A. baumannii ATCC 17978 were mixed and added. The arrows highlight the protein bands that were exclusively visible in lane 2. In Lane 2, the promoter DNA of recA from A. radioresistens DSM 6976, DdaA purification product, and cell lysate of A. baumannii ATCC 17978 were mixed and added to determine the protein(s) potentially interacting with DdaA.

The basal transcription of DDR genes, such as aciT, dGTPase, ruvB, uvrA, the gene of radical SAM protein, and the gene of TIGR03915-family protein, decreased significantly (>two-fold, P-value < .005) in the ddaA-knockout strain compared with the wild-type strain (Supplementary Tables S11 and S12), suggesting that DdaA is a transcription activator. The transcriptional levels of the ddaA gene itself remained unchanged after UV radiation in this study (Supplementary Table S11) and after MMC or ciprofloxacin treatments in previous studies [13, 28], indicating that the upregulation of DDR genes is not caused by the increased expression level of DdaA.

To verify the conclusion that DdaA is a transcriptional activator in vivo, a vector was designed containing an eGFP gene after the motif-containing promoter of aciT from A. baumannii was used. For comparison, a vector containing the mutated motif was constructed. Strong eGFP expression was detected exclusively in the wild-type strain containing the intact aciT promoter. However, no eGFP signal was detected in either the wild-type strain with a mutated promoter or the ddaA-knockout strain with the intact aciT promoter (Fig. 3D). The activating effect of DdaA following MMC treatment was also confirmed by flow cytometry (Supplementary Fig. S8B). These results indicate that both DdaA and its binding motif are required for its role as an activator in regulating the transcription of downstream genes in A. baumannii.

WYL domain-containing activators, namely PafBC and DriD, have been shown to facilitate the recruitment of RNA polymerases (RNAPs) to activate transcription [11, 71]. To explore whether DdaA is involved in RNAP recruitment, we used the DdaA-binding complex with the recA promoter as bait to pull down the proteins interacting with this complex from the cell lysate of A. baumannii. The results of this experiment revealed the presence of two additional bands with molecular weights of ∼40 and 180 kDa, respectively, compared to the negative controls (Fig. 3E and Supplementary Table S13). Subsequent mass spectrometry analysis indicated that the former was RNAP subunit alpha RpoA (37.2 kDa, UniProt: A0A828SFM7), and the latter was RNAP subunit beta RpoB (151.8 kDa, UniProt: A0A828SUQ7) and RNAP subunit beta RpoC (154.1 kDa, UniProt: A0A828SPY9). This outcome suggests that, similar to other WYL family proteins, DdaA may also participate in RNAP recruitment to initiate the transcription of downstream genes.

Moreover, the DdaA binding motif is located ∼50 bp upstream of the transcription start site (TSS, defined by the AcinetoCom database [72]) of numerous DNA damage repair gene operons (Supplementary Fig. S8C), which distance usually appears for activators [73]. This arrangement of the promoter is significantly different from that of PafBC-regulated promoters, where PafBC uniquely interacts with RNAP and adapts the sigma factor to recognize the PafBC-specific −26 element [11]. Thus, DdaA may not adopt the pafBC type of “sigma adaptation” to regulate the transcription of downstream genes.

DdaA enhances the immunity from RM and CRISPR–Cas systems under DNA damage stress

In the DefenseFinder RefSeq database [74], we found that promoters of 27.2% of the RM systems and 90.3% of the CRISPR–Cas systems encoded in the Acinetobacter harbor DdaA binding motifs, suggesting their regulation by DdaA (Supplementary Table S14). To determine the function of DdaA in the regulation of RM systems during phage infection, we transferred a vector containing the type I RM system from A. baumannii ATCC 19606 with its native promoter harboring the DdaA binding motif (Fig. 4A) to wild-type A. baumannii ATCC 17978, which lacks RM or CRISPR–Cas systems (Supplementary Table S10). In contrast to the wild-type A. baumannii ATCC 17978, which was sensitive to phage infection by phage PhAb24, the incorporation of the RM system enhanced its resistance to phage infection (Fig. 4B and C, blue dashed line). Moreover, the transformed strain exhibited an enhanced anti-phage defense capability when treated with MMC prior to phage infection (Fig. 4B and C, orange dashed line). This finding was consistent with the observation that MMC could promote the activation effect of DdaA, resulting in increased transcription of RM genes (Fig. 4D). Conversely, knockout of the ddaA gene or mutation in the DdaA binding motif of the RM system led to the disappearance of the protective effect against phages with or without pretreatment with MMC, which was accounted for by the unregulated transcription levels of RM genes (Fig. 4B–D). This result confirmed that both a functional DdaA and its intact binding motifs in the promoters are needed for the normal transcription of RM system genes, as well as for defense against phage infection. The transcription of this gene is upregulated under DNA damage conditions, resulting in better protection from phage infection.

Figure 4.

Validation of the regulatory role of DdaA in anti-phage defense. (A) Schematic illustration of the design of plasmid pRM-19606. The RM system and its neighboring genes, which form a defense locus against A. baumannii ATCC 19606, are shown. (B) Growth curve obtained from co-cultures of A. baumannii ATCC 17978-derived strains and phage PhAb24 at MOI = 0.02. (C) Phage plaque formation assay. Ten-fold serial dilutions of phage PhAb24 were spotted on a lawn of A. baumannii ATCC 17978-derived strains. (D) RT-qPCR results of A. baumannii ATCC 17978-derived strains. To normalize the data for each gene, the transcription of the RM-mutated motif strain was set to 1. The grouping results of the least significant difference (LSD) test (P-value = .05) are labeled. Groups sharing the same small letter (a, b, c) indicate no statistically significant differences. (E) Schematic representation of the pTarget and pControl plasmids. pControl is devoid of any sequence recorded by the CRISPR array, while pTarget contains the protospacer-adjacent motif (2 bp) and protospacer (32 bp) sequences of the prophages from A. baumannii strain AB046 (blue dot) and Acinetobacter pittii strain FDAARGOS 1397 (red dot). (F) The plasmid interference assay was designed to assess the capacity of the native CRISPR–Cas system in A. baumannii ZWAb067 to eliminate exogenous DNA. To accentuate the discrepancy in size between the colonies, the plate images were magnified by a factor of 25. (G) The relative abundance of pControl versus pTarget plasmids in overnight cultures of A. baumannii ZWAb067 was assessed using qPCR. (H) RT-qPCR results of A. baumannii ZWAb067-derived strains. To normalize for each gene, the transcription levels of the ddaA-knockout strains were set to 1. The grouping results of the (LSD test, P-value = .05) are labeled for panels (G) and (H). Groups sharing the same lowercase letter (a, b, c) indicated no statistically significant differences.

A significant proportion of Moraxellaceae species that employ the DdaA are equipped with CRISPR–Cas systems and/or prophages (Fig. 2). CRISPR spacer sequences frequently target prophage sequences in the same strain or other strains in the same genus (Fig. 2, Supplementary Fig. S5, and Supplementary Table S5). The cross-immunity exhibited by Moraxellaceae strains suggests potential cohabitation in the environment. In contrast, cross-immunity has not been observed in the basal lineages of Moraxellaceae or in the family of Alcanivoracaceae, where the SOS system is predominantly present.

To investigate the regulation of the CRISPR–Cas system by DdaA, we conducted a plasmid interference assay with two designed plasmids with (pTarget) or without (pControl) protospacer sequences that would be targeted by the native CRISPR I-F system in A. baumannii strain ZWAb067 (Fig. 4E). The successful transformation of these plasmids and evasion of the native CRISPR–Cas system cleavage would enable bacteria to form colonies on kanamycin plates. The results showed that in the wild-type strain, the transformation efficiency of pTarget was 0.16% of that of pControl, indicating a functional CRISPR–Cas system (Fig. 4F). In the ddaA-knockout strain, this ratio increased to 5.14%, suggesting that the loss of DdaA impaired the immune efficiency of the CRISPR–Cas system. The complemented strain showed no successful transformation, proving the regulatory role of DdaA in the CRISPR–Cas system. Following exposure to 5 J/m² UV radiation, wild-type cells exhibited a 100% removal rate of pTarget (corresponding to elimination of phage infection). In contrast, the ddaA-knockout strain remained at 97.5%, suggesting that UV radiation-induced DNA damage further stimulated the regulation of the CRISPR–Cas system by DdaA (Fig. 4F). Furthermore, a mixture of pControl and pTarget plasmids at a 1:1 ratio was introduced into the ZWAb067-derived strains. The relative abundance of pControl to pTarget in overnight cultures was analyzed using qPCR. An elevated pControl to pTarget ratio corresponds to a more efficient CRISPR–Cas system for removing pTarget plasmids. The ddaA-knockout strain displayed the lowest ratio at 3.48 × 10³, compared to the wild-type strain at 4.42 × 10⁴ and the complemented strain at 1.54 × 10⁵ (Fig. 4G). Furthermore, RT-qPCR analysis revealed that ddaA knockout led to significant reductions in Cas gene transcription under normal conditions or after UV radiation (Fig. 4H). These findings collectively suggest that DdaA plays a crucial role in regulating the expression of Cas genes, particularly in response to DNA damage.

Discussion

DDR systems are organized as regulons that coordinate the expression of proteins involved in numerous cellular processes in response to DNA damage, such as cell division, error-prone replication, and excision repair, ensuring the recovery and survival of cells [9, 75]. Among the three known global DDR systems in bacteria, namely SOS [9], PprI-DdrO [10], and PafBC [11], the SOS response system is the most widely distributed, as it can be found in most bacterial phyla. The SOS response is initiated when RecA is activated by ssDNA during DNA damage, promoting the autocatalytic cleavage of LexA repressor, derepressing SOS genes, and triggering DNA damage repair [9]. The PprI–DdrO system is mainly distributed in the Deinococcus–Thermus phylum. Similar to the SOS system, this system also utilizes a repression–derepression mechanism and relies on ssDNA sensing [76–78]. PafB and PafC heterodimers are transcriptional activators that regulate DDR in Actinobacteria [11, 79]. DriD is a newly defined activator that regulates multiple genes involved in DDR and cell division in the Caulobacteraceae family [67, 80]. Both PafBC and DriD work alongside the SOS system to ensure bacterial survival under DNA-damaging conditions.

In this study, we found a DDR in the family Moraxellaceae regulated by a WYL domain-containing protein, DdaA, which acts as a transcriptional activator for genes in its regulatory network (Fig. 3B–E). The knockout of ddaA influences the transcription of around 100 genes in DNA-damaging conditions, though only around 15 genes are predicted to be regulated by DdaA in Acinetobacter. Similar results can also be found in the well-studied E. coli K12 strain, where among the 138 LexA-dependent (>two-fold) induced genes, only 37 are classical SOS genes (with SOS box) [81]. We assume that the knockout of DDR regulators may result in genome instability, thus causing genome-wide transcriptional changes. During evolution, the DdaA network replaced the SOS system in the common ancestor of Fluviicoccus, Acinetobacter, Psychrobacter, and related genera of Moraxellaceae (Fig. 2A). While the three previously discovered bacterial global DDR systems control DNA damage repair genes exclusively, DdaA is involved in the regulation of both canonical DDR genes and anti-phage defense genes in Moraxellaceae (Fig. 2). Knockout of ddaA resulted in impaired transcription of downstream genes and increased mortality of Acinetobacter strains upon exposure to DNA damage agents and phage infections (Figs 3 and 4). Specifically, DdaA plays a pivotal role in enhancing protection against phage infection by regulating CRISPR–Cas and RM systems in response to DNA damage (Fig. 4).

As stated in this article, DdaA belongs to an emerging class of nucleic acid-sensing regulators in bacteria [64]. The DDR regulators PafBC and DriD are activators that use ssDNA generated upon DNA damage as the signal for their activation [11, 63, 67, 80]. Recent structural biology research suggests that the PafBC nature of the heterodimer may underlie the reason why one ssDNA can penetrate the PafBC complex [66], whereas the DriD complex interacts with two ssDNAs [67]. It has recently been reported that the three-helix bundle (3HB) domain is crucial for allosteric activation of DriD [80]. Since DdaA also harbors the 3HB domain, DdaA might share a similar allosteric activation mechanism with DriD. The two arginine residues in the β4 strand and β4/β5 loop are believed to be crucial for ssDNA interaction management in PafBC and DriD [64]. However, the shift of the first arginine residue to alanine has been shown in DdaA, which might lead to different sensing efficiencies or signals during regulation.

Additionally, the anti-phage defense system regulators CapW and BrxR have been identified as repressors, suggesting a regulatory balance between the protective and toxic effects of these abortive infection systems under normal conditions [62, 82] or during horizontal transfer of anti-phage genes [61, 83]. In our experiment, we found that DdaA acts as a transcriptional activator not only for the DDR but also for anti-phage defense systems. Notably, a significantly higher percentage of CRISPR–Cas systems (>90%) were subjected to DdaA regulation, whereas a relatively low proportion (∼30%) was observed for RM systems. Furthermore, the regulatory influence of DdaA on other anti-phage systems within the Acinetobacter genus is rare. The underlying biological mechanisms of these regulatory differences warrant further investigation. Given the ubiquity of WYL domain transcriptional regulators and their utilization of nucleic acid as their signal, the crosstalk between DDR and anti-phage defense in the same bacterial clade seems to be a plausible phenomenon that may not be limited to the Moraxellaceae family.

The function of the DdaA regulatory network may provide a selective advantage to Moraxellaceae bacteria in natural microbial communities. Many microorganisms secrete DNA-damaging chemicals, such as MMC [84], colibactin [23], bleomycin [85], and many other antibiotics [6], to compete with cohabiting bacteria. These chemicals have been shown to cause direct DNA damage and activate lytic replication of prophages in bacteria through various potential mechanisms, including those intensively studied for phage lambda [86, 87]. Under normal conditions, the phage repressor protein CI binds to the promoter of prophage genes, thereby repressing their transcription. However, when DNA damage is caused by MMC, the activated bacterial RecA protein promotes the digestion of LexA and triggers the SOS system for DDR. As the prophage repressor CI is a homolog of LexA, RecA also digests the CI repressor as a side effect and consequently activates the lytic cycle of prophages [88]. The outbreak of phages in the microbial community [23] enhances the antimicrobial effect of DNA damage agents, causing massive death of bacteria: not only of those encoding these prophages but also of their close or distant relatives infected by broad-host phages [23]. As a unique DDR regulator, DdaA in Moraxellaceae regulates both DDR and anti-phage defense genes, thus providing comprehensive protection against the dual effects of DNA damage chemicals [23] (graphical abstract).

In summary, DdaA is a transcriptional activator that belongs to a monophyletic cluster of WYL domain-containing proteins and controls both DDR and phage inhibition in the Moraxellaceae family. This unique mechanism in DDR provides Moraxellaceae strains with potential competitive advantages by resisting both DNA-damaging chemicals and the following phage outbreaks in environmental or hospital settings. Our findings reveal a previously unexpected challenge in combating the outstanding resistance of Moraxellaceae bacteria and shed light on new medication strategies to combat increasing antibiotic and phage therapy resistance [64, 89].

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This research was funded by the National Natural Science Foundation of China [32370028, 32200016, 42376113, and U22A20338] and the Zhejiang Provincial Natural Science Foundation of China [LQ23C010002, LZYQ25C010002, and LGN22C010002]. Funding to pay the Open Access publication charges for this article was provided by National Natural Science Foundation of China.

Data availability

The raw data of ChAP-seq and RNA-seq has been deposited in NCBI under the BioProject: PRJNA895108. Acinetobacter baumannii ZWAb067 genomic sequence can be found in the BioProject: PRJNA1146946. The pipeline Novel-Sample is available online at https://github.com/songshuang1996/Novel-Sample and https://doi.org/10.6084/m9.figshare.29832665.v1.