All strains and plasmids used in this study are listed in Table 1. A. baumannii strains were routinely cultured overnight (12-16 h) at 37°C in Luria Bertani (LB) broth or on LB agar plates and supplemented with antibiotics when necessary. For all biological assays (catalase, zymogram, peroxide resistance, SOD, and bacterial growth assays), swimming broth (SB) comprising 1% tryptone and 0.5% NaCl was used as the culturing medium. When necessary, SB was supplemented with 1.5% (w/v) agar to yield swimming agar (SA). For blue light treatments during bacterial culturing, a blue light LED array (peak emission wavelength, 469 nm) was used as described previously (Mussi et al., 2010; Wood et al., 2019).

Plasmid DNA was routinely isolated from E. coli using a commercial kit (Epoch Life Science). A. baumannii genomic DNA (gDNA) was isolated as previously described (Barcak et al., 1991). DNA polymerases, restriction enzymes, and T4 DNA ligase were purchased from New England Biolabs and were used according to the manufacturer’s instructions. All primers used in this study, which were either custom designed or based on sequences provided by cloning kits (Table 2), were purchased from IDT. The pCR Blunt II-TOPO and pCR8/GW/TOPO TA cloning kits were acquired from Invitrogen and used according to the manufacturer’s instructions. All DNA sequencing was carried out using automated DNA sequencing (BigDye, Applied Biosystems).

Catalase mutants were generated based on a recombineering approach designed for A. baumannii (Tucker et al., 2014). Briefly, each gene along with at least 500 nucleotides up- and downstream of each coding region was PCR amplified using Q5 DNA polymerase, 17978 wildtype gDNA, and the following primer sets: 4752 and 4753 for katA; 4760 and 4761 for katE; 4744 and 4745 for katE2; and 4756 and 4757 for katG (Table 2). The amplicons were cloned into pCR-Blunt. After successful cloning was verified by PCR screening using M13 primers, each gene was then subjected to a deletion-insertion mutation that removed more than 90% of each coding region. Deletions were generated by performing inverse PCR using the following primer pairs: 4754 and 4755 for katA; 4750 and 4751 for katE; 4746 and 4747 for katE2; and 4758 and 4759 for katG (Table 2). For selection purposes, in place of each deletion was ligated the Lactococcus lactis pIL252 ermAM gene encoding erythromycin resistance with its promoter region as described before (Fiester et al., 2016). After confirmation by PCR and automated DNA sequencing, each mutated gene was PCR amplified using Q5 DNA polymerase and the following primer sets: 4764 and 4765 for katA; 4766 and 4767 for katE; 4768 and 4769 for katE2; and 4762 and 4763 for katG (Table 2). Clean PCR products were used to transform 100 µL aliquots of electrocompetent 17978 cells carrying the pAT04 plasmid, prepared as described previously (Tucker et al., 2014). Electroporation of 1-12 µg PCR product was carried out at 2,500 V as described before (Dorsey et al., 2002), after which bacteria were allowed to recover for four h in SOC medium supplemented with 2 mM IPTG for induction of gene expression from pAT04. After recovery, bacteria were pelleted, and recombinants were selected for on LB agar containing 20 µg/mL erythromycin. Resulting colonies were screened via PCR using primers annealing outside homologous regions to confirm successful gene deletion and insertion. Primers used are numbered as follows in Table 2: 4752 and 4753 for the katA interruption; 4760 and 4761 for the katE interruption; 4744 and 4745 for the katE2 interruption; and 4756 and 4757 for the katG interruption. Mutations were finally verified by automated DNA sequencing.

Catalase genes and their predicted promoters, which were predicted with the aid of Softberry BPROM software (Solovyev and Salamov, 2011), were amplified using primers with 5’ BamHI restriction sites as follows: 4792 and 4793 for katE; 4794 and 4795 for katE2. (Table 2). After amplification, PCR products were cloned into pCR8, and the resulting plasmids (Table 1) were used to transform Top10 E. coli. Positive transformants were selected for on LB agar containing 100 µg/mL spectinomycin. Genes and predicted promoters were then subcloned from pCR8 into the E. coli-A. baumannii shuttle vector pMU368 (Table 1) by restriction cloning using BamHI. Resulting plasmids (Table 1) were used to transform Top10 E. coli competent cells according to the manufacturer’s instructions. Positive transformants were selected for on LB agar supplemented with 50 µg/mL kanamycin (Km). Successful cloning of each gene with its predicted promoter into pMU368 was confirmed by PCR screening using M13 primers and automated DNA sequencing. Complementing plasmids were introduced into respective electrocompetent A. baumannii mutant strains by electroporation (Dorsey et al., 2002); positive transformants were selected for on LB agar containing 50 µg/mL Km. The presence of complementing plasmids as independent replicons without detectable rearrangement was confirmed by isolating plasmids from A. baumannii derivatives using the hot-Triton method (Actis et al., 1993). Isolated plasmids were then transformed into E. coli Top10 competent cells, and Km resistant colonies were used to isolate plasmids that were examined by restriction analysis (Sambrook and Russell, 2001).

Overnight bacterial cultures were diluted 1:100 in SB and grown at 24°C in darkness with shaking at 150 rpm to an OD₆₀₀ of approximately 0.8. Bacteria in 1 ml of culture were pelleted by centrifugation and then lysed using the TRIzol reagent (Invitrogen). Total RNA was extracted using the Zymo Direct-zol RNA Miniprep Plus kit (Zymo Research) according to the manufacturer’s instructions. DNA was removed from RNA samples using the TURBO DNA-free kit (Invitrogen) according to the manufacturer’s instructions, after which samples were verified to be DNA-free using Taq DNA polymerase (New England Biolabs) and primers 4513 and 4514 (Table 2) for the detection of csuAB as an indicator of gDNA contamination. Total RNA was reverse transcribed using the iSCRIPT cDNA synthesis kit (Bio-Rad) according to the manufacturer’s instructions. Relative gene transcription was measured by qPCR using the iTaq Universal SYBR Green supermix (Bio-Rad) according to the manufacturer’s instructions and as previously described (Fiester et al., 2016). Primers 4685 and 4686 (Table 2) were used to measure the expression of katE. Expression of katE was normalized to that of the constitutively expressed reference gene recA, which was measured using primers 4796 and 4797 (Table 2). qPCR efficiencies were at least ~90% in all reactions. Transcriptional analysis was performed in this way for each strain using two biological replicates, the first analyzed in technical duplicate and the second analyzed in technical triplicate (n = 5).

Catalase activity was measured using an assay as previously described (Iwase et al., 2013). In brief, overnight bacterial cultures were diluted 1:100 in SB and then incubated at 24°C in darkness or under blue light illumination provided by a LED array as described above with shaking at 150 rpm to an OD₆₀₀ of approximately 0.8. Culture samples representing equal numbers of bacteria were then pelleted by centrifugation, washed in sterile phosphate-buffered saline (PBS), and resuspended in equal volumes of PBS. Catalase activity was assayed by mixing 100 μL bacteria prepared as described above with 100 μL 1% Triton X-100 and 100 μL 30% H₂O₂ (Fisher Chemical) in a glass culture tube. Total foam formation was measured 15 min after mixing and used as a proxy for catalase activity. The catalase activity of each bacterial strain in each condition (blue light treatment or darkness) was measured using three independent biological replicates, each in technical triplicate (n = 9).

For zymogram analysis, overnight bacterial cultures were diluted 1:100 in SB and then incubated at 24°C in darkness or under blue light with shaking to an OD₆₀₀ of approximately 0.8. Equal volumes of bacteria were collected by centrifugation at 6,000 x g for 10 min, resuspended in extraction buffer (50 mM Tris-HCl, pH 8.0), and were then lysed by bead beating. Lysates were clarified by centrifugation at 12,000 x g for 20 min, and protein concentration of the resulting supernatant was measured using the Bradford method (Bradford, 1976) using a commercial kit (Bio-Rad). Equal amounts of protein were loaded onto a 10% non-denaturing polyacrylamide gel and were run under non-denaturing conditions. In-gel catalase activity was visualized as previously described (Gamero-Sandemetrio et al., 2017). This analysis was carried out twice using separate biological replicates each time.

Resistance to H₂O₂ was analyzed by performing disc diffusion assays as described previously (Juttukonda et al., 2019) with some modifications. Briefly, bacterial cultures were grown at 24°C for 18 h with shaking. Then, 200 μL of bacterial culture were added to 20 mL warm SA, and equal volumes of this mixture were aliquoted into two separate 60-mm Petri dishes. After drying for 10 min, 6-mm sterile filter discs were deposited on the top of the agar, and 10 μL 9.8 M H₂O₂ was added to each disc. Plates were incubated overnight at 24°C in darkness or under blue light, and diameters of growth inhibition were measured. Disc diffusion experiments were carried out for each tested bacterial strain in each condition using three independent biological replicates (n = 3).

Bacterial growth in SB only or in the same medium supplemented with H₂O₂ was analyzed by monitoring OD₆₀₀ values over time. Overnight liquid bacterial cultures were diluted 1:100 in SB supplemented with 0 or 2.5 mM H₂O₂ and were grown with shaking and in the presence of blue light or in darkness, as indicated, at 24°C. OD₆₀₀ was measured using a spectrophotometer. All growth experiments were carried out at least twice using independent biological replicates each time.

SOD activity was measured in bacteria cultured as described for catalase activity analysis. After culturing, cells were pelleted by centrifugation (6 min at 6,000 x g), washed in PBS, and lysed by sonication using an ultrasonic cell disruptor with a microtip (Fisherbrand). Lysed samples were centrifuged for 6 min at 6,000 x g to collect cell debris, and clarified samples were assayed for SOD activity using the commercial SOD Assay Kit-WST (Dojindo Molecular Technologies Inc.) according to the manufacturer’s instructions. SOD activity was normalized to total protein concentration as measured by a Bradford assay. For each tested strain/condition, a total of 2-3 independent biological replicates in technical triplicate were analyzed.

Predicted amino acid sequences of selected proteins were downloaded from NCBI. Sequence alignments and comparisons were performed using MUSCLE (Edgar, 2004). Binding and active sites of amino acid sequences were identified using Swiss-Prot or TrEMBL sections of UniProt (UniProt Consortium, 2021). A phylogenetic tree was constructed based on a comparative genomic analysis of the catalase homologs from A. baumannii strains 17978, 19606, AB5075, and Acinetobacter sp. Ver3. Multiple alignment of the amino acid sequences of all catalase homologs was carried out using MUSCLE, implemented within the Molecular Evolutionary Genetics Analysis tool, MEGA version X (Kumar et al., 2008; Stecher et al., 2020). This tool was also used to infer protein phylogeny using the Maximum Likelihood method based on an initial tree constructed from the Neighbor Joining method. Evolutionary distances were computed using the Poisson correction method (Zuckerkandl and Pauling, 1965). A discrete Gamma distribution was used to model evolutionary rate differences among sites [5 categories (+G, parameter - 200.0000)]. Reliability of the inferred tree was tested by bootstrapping with 800 repetitions.

The statistical significance of the experimental data sets was analyzed using GraphPad Prism version 9.0 for MacOS with the appropriate statistical test applied for each experiment as described in the figure legends. P values ≤ 0.05 were considered statistically significant in these analyses. Error bars represent the standard deviation of the mean of each data set.

In the 19606 strain, BlsA is responsible for the regulation of catalase gene expression and activity in response to blue light (Muller et al., 2017). These observations prompted us to determine whether the same phenomenon is observed in the 17978 isolate and, if so, whether the BlsA residues that are necessary for the regulation of bacterial surface-associated motility, another BlsA- and light-dependent phenotype, are critical for the regulation of catalase activity. Notably, using a different means of measurement, we were able to demonstrate that catalase activity is a light-regulated activity in 17978 (Figure 1). Although there is indeed catalase activity under both conditions, this activity is regulated by light such that there is significantly higher activity when wildtype 17978 bacteria are cultured under blue light compared to in darkness. This regulation is dependent on the BlsA photoreceptor: in the isogenic blsA::aph derivative (17978.OR), light regulation is abolished, yet when this mutant is complemented in trans with the wildtype blsA allele (17978.OR.WT), the light-dependent phenotype is restored.

The genome of the 17978 strain harbors genes encoding four catalase homologs: KatA, KatE, KatE2, and KatG. All homologs include the critical residues necessary for catalase or peroxidase activity including binding sites for heme, enzymatic active sites, and metal binding sites (Supplementary Figures 1, 2). Notably, each catalase homolog of the 17978 strain shares above 99% identity with and similarity to cognate catalase homologs in both the 19606 and AB5075 strains, with the exception of KatE2, which is absent in the AB5075 strain (Table 3). More specific information about which catalase homologs and associated physiological functions are regulated by light and/or BlsA and the mechanism of said regulation is needed. This conclusion is supported by the recent report by Muller et al. (2017) that identified two proteins expressing light- and BlsA-dependent catalase activity, although their RNA-Seq data identified only one gene being differentially transcribed in response to illumination (Muller et al., 2017). Likewise, although no light-mediated regulation of catalase activity has been explored in the AB5075 strain, the role of different catalase homologs has been significantly characterized in this strain (Sun et al., 2016). Importantly, the AB5075 strain also encodes a BlsA-like homolog whose light sensing and regulatory functions remain to be determined (Gallagher et al., 2015).

Despite similarities between strains of A. baumannii, within each strain the catalase homologs are distinct in sequence and annotated function. In the 17978 strain specifically, KatE and KatE2 have the greatest similarity and identity to each other with values of 40.5% and 27.3%, respectively. The second highest similarity and identity are shared between KatA and KatE2 with similarity and identity values of 28.7% and 19.1%, respectively. Although there are significant amino acid sequence differences among the homologs, KatA, KatE, and KatE2 all possess heme-binding domains (Supplementary Figure 1), whereas KatG contains a domain belonging to the PRK15061 superfamily that confers both catalase and peroxidase activities (Supplementary Figure 2). Characterization of KatE and KatG has been particularly explored in the AB5075 isolate as previously noted (Sun et al., 2016). However, unlike the 17978 and 19606 strains, the AB5075 strain does not possess a KatE2 homolog. Notably, a clue into the significance of KatE2 with regard to the physiology of bacteria that possess this type of catalase may be taken from Acinetobacter sp. Ver3, which encodes two catalases, ^AV3KatE1 and ^AV3KatE2, that are both highly similar to 17978’s KatE2 (Figure 2). Interestingly, the ^AV3KatE1 and ^AV3KatE2 catalases are implicated in UV survival and H₂O₂ resistance in this polyextremophile (Sartorio et al., 2020). The phylogenetic relationships of the catalase homologs from these four strains are illustrated by the tree shown in Figure 2. The distinctive functions among these clades of catalase homologs indicates that the pathophysiological impact of light-mediated catalase regulation in A. baumannii is dependent on the catalase homolog being regulated.

Because 17978 cells regulate catalase activity in response to light but encode several candidate homologs, we aimed to determine which of these homologs are indeed responsible for the activity observed under our experimental conditions, i.e., which are regulated by light through BlsA. For this purpose, deletion-insertion mutations were generated in all four catalase genes using a recombineering mutagenesis approach developed specifically for A. baumannii (Tucker et al., 2014). Importantly, none of these mutations resulted in a significant growth deficiency compared to wildtype as well as among themselves (Supplementary Figure 3). The same assay described above was then used to test the catalase activity of each mutant cultured to exponential phase in either blue light or darkness. Compared to wildtype 17978 cells, mutations of katA and katG caused no significant decrease in catalase activity in mutant bacteria (17978.katA and 17978.katG, respectively) that had been cultured in either condition (Figure 3A). Regarding the katE2 mutation, the catalase activity of this strain (17978.katE2) was decreased significantly compared to wildtype only when bacteria had been cultured in darkness (Figure 3A). Nonetheless, in the absence of any of these three genes, catalase activity was still regulated by light. On the other hand, the katE deletion mutant strain (17978.katE) exhibited a complete absence of catalase activity when bacteria had been cultured in either condition (Figure 3A). Notably, we were able to complement the activity of 17978.katE to wildtype levels, albeit without the light-dependent differential response, when katE was expressed from a complementing plasmid (Figure 3B). To provide additional evidence that it is indeed KatE that BlsA is regulating in response to blue light, we carried out zymogram analyses of 17978 wildtype cells and cells of the four catalase mutant strains. According to the results as shown in Figure 3C, the analysis of soluble proteins isolated from 17978 cells under non-denaturing conditions revealed two bands representing proteins that decomposed H₂O₂, and both bands were slightly brighter in bacteria that had been cultured under blue light compared to in darkness. The soluble protein profile of the 17978.katE strain lacked the top band, indicating that the light-regulated catalase is indeed KatE. Notably, in the protein profile of the 17978.katG strain, the bottom band was lacking, indicating that this band represents KatG. Importantly, in neither the 17978.katE2 nor the 17978.katA strains were there noticeable differences in the protein profile compared to the wildtype 17978 strain (Figure 3C).

In a previous study of the 19606 strain, it has been proposed that BlsA is responsible for the light-dependent differential production of two different catalases, which were predicted to be KatE and KatG (Muller et al., 2017). Because of this previous hypothesis and because KatG is a catalase-peroxidase, we investigated whether peroxidase activity is also regulated in response to blue light. Peroxidase activity is related to H₂O₂ resistance based on the hypothesized role of this activity in detoxification of peroxy compounds (Sun et al., 2016). Therefore, we used two different measurements of peroxide resistance, disc diffusion assays and bacterial growth curve analyses, to examine the response of wildtype and mutant bacteria grown under blue light versus in darkness to H₂O₂. Interestingly, we found that under the conditions of the disc diffusion assay, there was no measurable light-dependent difference in the response of 17978 wildtype bacteria to H₂O₂ challenge (Figure 4A). In the AB5075 strain of A. baumannii, KatG is the main enzyme responsible for H₂O₂ resistance measured using the same type of assay (Sun et al., 2016). To determine whether the same trend is observed in the 17978 strain, we further analyzed the peroxide resistance of the individual catalase mutant derivatives. Notably, as with wildtype bacteria, there was no light-dependent phenotype observed in any of the mutant strains (Figure 4A). The absence of both KatE and KatE2 caused a significant negative impact on the bacterial response to peroxide, although KatE2’s absence caused the greatest effect when compared to all other strains (Figure 4A). Importantly, complementation of the katE2 gene in the cognate mutant background was successful in significantly decreasing the growth inhibition to levels comparable to those detected with the parental strain (Supplementary Figure 4). Remarkably, the absence of neither KatA nor KatG had a significant impact on peroxide resistance under the growth conditions used in the disc diffusion assay when compared with the parental strain (Figure 4A). Alternatively, the results of the growth curve analysis provided evidence of a significant, albeit slight, difference between the resistance of 17978 wildtype bacteria cultured under blue light compared to in darkness during exponential growth (Figure 4B). Compared to the wildtype strain at 8 h, all mutant strains exhibited a significant growth defect in the presence of H₂O₂, with 17978.katE exhibiting the greatest defect (Figures 4C–F). Furthermore, 17978.katE was the only strain that was completely unable to grow in the presence of 2.5 mM H₂O₂ across the entire 24 h time period during which growth was analyzed (Supplementary Figure 5; see supplemental figure for full growth data of all strains).

Taken together, these observations indicate that in the 17978 strain, 1) the catalase activity measured under our enzymatic assay conditions is light- and BlsA-regulated and mediated by KatE (Figures 1, 3), and 2) of the four catalase homologs, KatE and KatE2 play the most significant roles in the breakdown of and overall resistance to H₂O₂ (Figures 3, 4A and Supplementary Figure 5). Notably, not all catalase-related functions are light-regulated (compare Figure 3 to Figure 4A). Additionally, KatG and KatA show only residual roles in catalase activity or general H₂O₂ resistance when cells are cultured at 24°C under shaking conditions (Figures 4C, F and Supplementary Figure 5).

Once we had investigated which of the catalase homologs and associated functions are regulated by light and BlsA, we set out to analyze the role of particular BlsA residues in the regulation of catalase activity. Specifically, we analyzed the impact of three different BlsA mutations on this regulation. Two of these mutations are substitution mutations of lysine residues at positions 144 and 145 to glutamic acid residues in both cases (K144E and K145E, respectively). The third mutation is a complete truncation of the last five amino acid residues of BlsA (Δ5). We have explored the effect of these three mutations on the capacity for BlsA to regulate surface-associated motility previously (Wood et al., 2019). Notably, both lysine residues are critical for BlsA’s role as a regulator of surface-associated motility, yet the five last BlsA residues are dispensable for said regulation.

We used the assay described above to analyze the light-dependent catalase activity of the isogenic 17978 OR blsA insertion mutant (17978.OR) complemented with plasmids encoding the BlsA K144E, K145E, or Δ5 derivatives. This approach showed that OR cells producing the K145E BlsA protein (17978.OR.K145E) exhibited a differential light response similar to that displayed by cells producing the wildtype BlsA protein (17978.OR.WT), albeit not as statistically different between light and dark (Figure 5A). In contrast, such a differential response was not detected in bacteria producing the K144E protein (17978.OR.K144E). These results indicate that the K144 residue is important for BlsA’s ability to regulate KatE, which is not unexpected considering this residue is critical for BlsA’s regulation of surface-associated motility (Wood et al., 2019). What was unexpected was the catalase activity displayed by OR cells producing the Δ5 BlsA deletion derivative; this strain (17978.OR.Δ5) differentially produced catalase activity in response to illumination, but its overall catalase activity was significantly lower when compared with bacteria producing wildtype BlsA (Figure 5A). This observation is in contrast to results showing that the last 5 BlsA amino acid residues do not play a role in the expression of light-dependent BlsA-mediated 17978 surface motility responses (Wood et al., 2019). Importantly, this decrease in overall catalase activity cannot be attributed to a growth defect, as 17978.OR.Δ5 bacteria do not display any growth deficiency compared to wildtype (Supplementary Figure 6). Furthermore, the Δ5 deletion did not significantly affect the production of BlsA when overexpressed in E. coli or tested using A. baumannii total lysates prepared from wildtype and 17978.OR.Δ5 cells by western blotting as described before (Wood et al., 2019) (data not shown). Interestingly, when analyzed by reverse transcription quantitative PCR (RT-qPCR), bacteria producing the Δ5 deletion BlsA molecule exhibited significantly less katE expression compared to those producing the wildtype photoreceptor (Figure 5B).

We hypothesized that, based on the data we obtained upon analyzing catalase activity in bacteria producing the Δ5 BlsA deletion derivative, BlsA’s last five residues may somehow play a role in eliciting an oxidative stress response, which then leads to an increase in katE expression and overall catalase activity that is independent of light. To test this possibility, the overall oxidative stress response of 17978 was examined using an assay to measure the production of SOD, a marker for such response. SOD catalyzes the dismutation of the superoxide radical O 2 − to protect against both direct and indirect toxic damage. Specifically, superoxide is the primary ROS formed by photooxidation (Ziegelhoffer and Donohue, 2009). We analyzed the production of SOD in 17978 WT, 17978.OR, and 17978.OR.Δ5 bacteria cultured under blue light or in darkness. There was no significant difference in the oxidative stress responses of bacteria completely lacking BlsA compared to those producing the Δ5 BlsA deletion derivative (Figure 6). However, unexpectedly, the result of this analysis indicated that although catalase activity was higher when bacteria were cultured under blue light (Figure 1), SOD activity was downregulated in bacteria cultured under blue light (Figure 6). The fact that the absence of either the entire BlsA molecule or only its last five residues eliminated the light regulation of SOD activity observed in wildtype bacteria lends further importance to these residues that were, based on our previous study, thought to be dispensable for BlsA’s regulatory activity.