Staring at the Cold Sun: Blue Light Regulation Is Distributed within the Genus Acinetobacter

We previously showed that the opportunistic nosocomial pathogen Acinetobacter baumannii is able to sense and respond to light via BlsA, a BLUF (Blue-Light-sensing Using FAD)-domain photoreceptor protein. Here, we extend our previous studies showing that light regulation is not restricted to A. baumannii, but rather widespread within the genus Acinetobacter. First, we found that blue light modulates motility and biofilm formation in many species of the genus, including members of the Acinetobacter calcoaceticus-A. baumannii complex. In many of these species blue light acts as a key factor guiding the decision between motility or sessility at 24°C, whereas in A. baumannii, light inhibits both motility and biofilm formation. We also show that light regulation of motility occurred not only at 24°C but also at 37°C in non-A. baumannii species, contrasting the situation of A. baumannii which only shows photoregulation at 24°C. Second, we show that Acinetobacter baylyi (strain ADP1) BLUF-photoreceptors can functionally replace in vivo the A. baumannii 17978 BlsA protein and that the pathways leading to biofilm formation are inversely regulated at 24°C between these two microorganisms. Finally, we found the presence of predicted genes coding BLUF-containing proteins in all Acinetobacter sequenced genomes, even though the copy number is variable among them. Phylogenetic analysis suggests a common origin for all BLUF domains present in members of this genus, and could distinguish well-differentiated clusters that group together BLUF homologs from different species, a situation particularly clear for members of the ACB complex. Despite a role played by these BLUF domain-containing proteins in the photoregulation observed in the members of the genus Acinetobacter is a likely scenario given our findings in A. baumannii and A. baylyi, further research will contribute to confirm this possibility.


Introduction
The members of the genus Acinetobacter are strictly aerobic, oxidase negative, ubiquitous Gram-negative cocobacilli that are frequently found in the environment but also in the hospital setting, where some particular groups of the genus have been associated with outbreaks of nosocomial infections [1]. Currently, the genus comprises 27 species with valid names (www.bacterio. cict.fr/a/acinetobacter.html) and several putative species with provisional designations including nine genomic species delineated by DNA-DNA hybridization [1,2]. The Acinetobacter calcoaceticus-Acinetobacter baumannii (ACB) complex is a subgroup within the genus comprised by closely related species including A. baumannii, A. calcoaceticus, Acinetobacter nosocomialis and Acinetobacter pittii [1,2]. A. calcoaceticus is mainly an environmental microorganism rarely involved in human infections, while A. nosocomialis and A. pittii are predominantly isolated from hospitalized patients [1][2][3]. A. baumannii is the clinically most important species of the genus. During the last decades, it has emerged as one of the most common opportunistic pathogens involved in hospital-acquired infections, generally affecting immunocompromised and severely injured patients [4]. Outbreak strains of A. baumannii are generally multidrug-resistant and can readily colonize nosocomial environments and withstand unfavorable conditions such as desiccation, nutrient starvation, and antimicrobial treatment. A key determinant for the success of this pathogen in clinical settings has been proposed to be its ability to attach to and form biofilms on abiotic and biotic surfaces [5][6][7][8]. We have recently reported that this nosocomial pathogen perceives light as a stimulus governing many processes of its life cycle at low environmental temperatures [5]. In fact, we found that blue light inhibited the formation of biofilms and pellicles in A. baumannii ATCC 17978 cells cultured at 24uC in liquid broth, and enhanced the ability of the bacteria to kill the filamentous form of the eukaryotic fungus Candida albicans [5]. We also showed that A. baumannii ATCC 17978 cells ceased moving on semisolid media plates in the presence of blue light at 24uC, while they spread all over the surface of the plates in the dark. Genome sequence analysis suggested a likely candidate for the photoreceptor implicated in this process: a BLUF-domain containing protein encoded by the A1S_2225 open reading frame. Its involvement in A. baumannii photoregulation was latter confirmed by both biophysical and genetic approaches, and therefore it was designated BlsA for blue light sensing A [5]. Interestingly, light regulation was lost at 37uC, a temperature compatible with warmblooded hosts [5]. This temperature-dependency prompted speculations on the role of light sensing in the lifestyle of A. baumannii. Our current hypothesis postulates that light sensing may play a role during the pathogens environmental life, modulating its behavior outside the human body.
Many questions arise from our previous work, such as whether blue light regulation is widely distributed within the genus Acinetobacter. In such a case, relevant questions would be whether environmental species show light-mediated responses and temperature dependence similar to that described above in A. baumannii. Alternatively, blue light regulation could be restricted only to a subgroup in which A. baumannii is contained, suggesting that acquisition of light sensing genes occurred by recent horizontal gene transfer events.
Therefore, in this work we evaluated whether other species within the genus Acinetobacter also share the light-dependent responses described in A. baumannii, in particular, by studying photoregulation of motility and biofilm formation. We also provide insights into the evolution of BLUF-domains encoded within Acinetobacter genomes, shown to mediate light regulation in A. baumannii, by inferring and analyzing their phylogenetic relationships. This knowledge may contribute to our current understanding of the role, importance and evolution of light sensing and regulation in these bacterial species.

Bacterial Strains
The A. baumannii and E. coli strains, as well as the plasmids constructed and/or used in this work are listed in Table 1. The Acinetobacter strains used in this study, which include representatives of 25 validly named species, and their origin (if known) are listed in Table 2. 'A. indicus' and 'A. oleivorans' are effectively but not validly published names for single strains (CCM 7832 and DR1, respectively), and therefore are mentioned between apostrophes. Many of these strains have been reported in literature before while others, selected from our collection, have been identified at the species level by amplifed ribosomal DNA restriction analysis (ARDRA) [9]. All this information is also indicated in Table 2.

Cell Motility Experiments and Biofilm Assays
Cell motility was tested on ''swimming agarose'' (Tryptone 1%, NaCl 0.5%, agarose 0.3%; 5) or LB agarose (Peptone 1%, NaCl 1%, yeast extract 0.5%, agarose 0.3%) plates incubated in the presence or absence of light. ''Swimming agarose'' is an inherited name for plates of this composition, but it should be noted that members of the genus Acinetobacter do not perform flagellamediated swimming as they do not produce this type of cell appendage [10]. Biofilm assays were carried out in glass tubes as described before [5]. All assays were performed at least in triplicate for both light and dark conditions.
Plates or biofilm tubes were incubated at 24uC or 37uC in the dark or under light emitted by LED (light-emitting diode) arrays with an intensity of 6 to 10 mmol photons/m2/s, determined using a Li-250A Light Meter (LI-COR). Each array was built using three-LED module strips (containing three LEDs each) emitting blue, green, or red light with emission peaks centered at 462 nm, 514 nm, and 636 nm, respectively, as determined using a LI-COR LI-1800 spectroradiometer [5]. It should be noted that temperature of both the liquid or agar medium under the illumination conditions used in these experiments did not vary significantly from that measured under dark conditions or in the incubation chamber.
For quantification of biofilms assays, duplicate cultures for each sample at each condition were used. One tube was sonicated immediately for 5 s with a thin probe and the OD 600 of the culture was determined to estimate total cell biomass. The supernatant of the other tube was aspirated and rinsed thoroughly with distilled water. The cells attached to the tube walls were visualized and quantified by staining with crystal violet and solubilization with ethanol-acetone as described in ref. [11]. The OD 580 /OD 600 ratio was used to normalize the amount of biofilm formed to the total cell content of each sample tested, to avoid variations due to differences in bacterial growth under different experimental conditions. Error bars show standard error of the mean for 3 different biological replicates (n = 3).

General DNA Procedures
Genomic and plasmid DNA were isolated as described before [4], and DNA restriction and cloning experiments were carried out using standard protocols [12]. DNA sequences were determined at the DNA Sequencing Facility of the University of Maine, Orono, ME, USA; or at Macrogen (Korea).

Disc Diffusion Antibiotic Susceptibility Test
Müeller Hinton (MH) agar plates were inoculated with a culture of each tested strain, which were previously adjusted to 0.5 McFarland standard turbidity using fresh culture medium or saline solution, according to the recommendations of the National Committee for Clinical Laboratory Standards [14]. Antimicrobial commercial discs (BBL, Cockeysville, MD, USA) containing 10 mg of ampicillin, 30 mg of amikacyn, 30 mg of cefepim, 30 mg of cefotaxime, 30 mg of cefoxitin, 30 mg of cephalotin, 30 mg of chloramphenicol, 5 mg of ciprofloxacin, 10 mg of imipenem, 10 mg of gentamycin, 10 mg of meropenem, 100 mg of piperacillin or 5 mg of rifampicin were placed on the surface of plates, which were latter incubated overnight at 24uC or 37uC under blue light or darkness. The assays were performed in triplicate.

Sequence Analyses
Protein sequences were retrieved from databases at NCBI (http://www.ncbi.nlm.nih.gov) and Pfam [15], and were aligned using ClustalW version 1.7 [16]. To exclude the high sequence variability often found after BLUF domains which -in some casessupposes diversity of the accompanying effector domains [17], only the region corresponding to the BLUF domain (comprising 93-96 amino acid residues, depending on the organism) was extracted from the complete sequences and used for the alignments. Gaps were removed from the alignments using BioEdit version 7.05.3 [18].

Phylogenetic Inferences
Phylogenetic relationships were inferred from amino acid sequence alignments using the programs provided in the PHYLIP package, version 3.69 [19] (http://evolution.genetics.washington. edu/phylip.html). The maximum-likelihood method (PROTML) was used for the construction of the BLUF-domain phylogenetic tree. In all cases, confidence levels were calculated from 1,000 bootstrap resamplings (SEQBOOT) of alignments used for phylogenetic inferences by both neighbor-joining method using a Dayhoff PAM distance matrix (PROTDIST) and the parsimony (PROTPARS) method, also included in the PHYLIP software package [19].

Motility is also Regulated by Blue Light in non-A. baumannii Members of the Genus Acinetobacter
Given that light regulates motility in A. baumannii [5], we analyzed whether other species of this genus are also able to move and respond to light at 24uC. We found that A. baylyi, A. calcoaceticus, A. nosocomialis, 'A. oleivorans', A. pittii and A. tjernbergiae 17978.ORc 1499 - 17978.ORc 2125 - 17978.ORc 2129 -  A. rudis showed light-dependent regulation of motility, at least in one from three to five strains assayed (see Figure 1 and Table 2). Just as described for A. baumannii [5], motility in these bacteria was inhibited under blue light, while in the dark they spread all over the surface of the plate (Figure 1). In the case of A. baylyi and A. pittii, blue light inhibition was not absolute and bacteria were still able to move, even though to a much lesser extent than in the dark ( Figure 1). It is interesting to note that the appearance of A. pittii is different from that observed for other species such as A. baumannii or A. calcoaceticus. A. pittii did not move homogeneously from the inoculation point, but rather formed independent striations irradiating from the center. We did not observe motility in any of the  Table 2). In our previous study we showed that light regulation of motility in A. baumannii occurred at 24uC but not at temperatures associated with warm-blooded hosts such as 37uC [5]. We thus evaluated next whether any of the strains that showed photoregulation at 24uC also exhibited light regulation at 37uC and found that, in contrast to A. baumannii, all of them did respond to light at a higher incubation temperature ( Figure 1 and Table 2). It should be noted that similar results were obtained when all strains were tested using LB agarose 0.3% instead of ''swimming agarose'' under similar experimental conditions.

A. baumannii Members of the Genus Acinetobacter
We further evaluated whether biofilm formation was photoregulated in different species within the genus by studying their     Figure S1 and Table 2). Some strains of A. beijerinckii, A. brissouii, A. guillouiae, A. johnsonii, A. lwoffii, A. nosocomialis, A. pittii, A. ursingii, and A. venetianus also showed photoregulation of biofilm formation (Figure 2, Figure S1 and Table 2), but the levels of wall biofilms were lower than in the aforementioned strains (Figure 2A). It is interesting to note that even though there is photoregulation of biofilm formation in these species (similarly to A. baumannii), the amount of wall biofilms formed by non-A. baumannii species was much greater under blue light than in darkness, in contrast to A. baumannii in which larger levels are observed in the dark [5]. In non-A. baumannii species, we observed mainly the presence of wall biofilms only. The presence of pellicles (with no wall biofilm occurring at the same time) was evident only in a few strains of A. bereziniae, A. guillouiae, A. nosocomialis, A. pittii and A. radioresistens, and no light regulation was detected on them. The effect of light on biofilm formation at 37uC by the Acinetobacter strains used in this study could not be evaluated because of wide variations observed in different assays. Such a response is not surprising, as similar situations have been described before for clinical isolates of A. baumannii [5]. A summary of the biofilm properties of all the strains studied is presented in Table 2, and quantification of biofilm formation for strains showing photoregulation is shown in Figure 2 for representative cases or Figure S1.

A. baylyi: a Case Study
The fact that A. baylyi harbors four paralogous genes encoding BLUF-containing putative photoreceptors, designated ACIAD1499, ACIAD2110, ACIAD2125 and ACIAD2129 (Table 2), suggests that light might play a key role in the lifestyle of the bacterium, which could justify the abundance of these genes and their possible functional redundancy. To better characterize the response of this bacterium to light, we compared its ability to move under red or green light with its response under blue light or darkness at 24uC. We observed that green light inhibited motility, even though to a lesser extent than blue light (Figure 2). In contrast, under red light the bacteria behaved as in darkness. Thus, A. baylyi is able to sense and respond to green light in a similar way as A. baumannii [5], indicating that at least one of the four putative photoreceptors is capable to respond to green light.
We assayed next whether any of the different photoreceptors present in A. baylyi were able to rescue the lost photoregulation of motility and biofilm formation at 24uC of the A. baumannii ATCC 17978 blsA mutant (17978.OR) [5]. For this purpose, we cloned each of the A. baylyi predicted photoreceptor genes under their own promoter control in the shuttle plasmid pWH1266 to generate pWH1499, pWH2110, pWH2125 and pWH2129 (Table 1). Then, we transformed the 17978.OR mutant strain with these constructions, as well as with the empty vector, to generate 17978.ORc 1499 , 17978.ORc 2110 , 17978.ORc 2125 , 17978.ORc 2129 and 17978.OR p , respectively. Finally, we analyzed whether these strains showed photoregulation of motility or biofilm formation at 24uC.   Figure 3A shows that 17978.OR harboring the empty pWH1266 plasmid (17978.ORp) does not exhibit photoregulation of motility, spreading throughout the plate either under blue light or in the dark at 24uC. In contrast, 17978.ORc 1499 , 17978.ORc 2125 and 17978.ORc 2129 showed a tight inhibition of motility under blue light while spreading throughout the plate in the dark. Such response resembles that of the complemented 17978.ORc BlsA strain harboring the native A. baumannii blsA wildtype allele (see Figure 3A) [5]. Therefore, the corresponding A. baylyi putative photoreceptors are able to fully complement the blsA gene. Conversely, the A. baylyi photoreceptor encoded by ACIAD2110 only partially complemented the blsA mutation in ATCC 17978 showing that it is not able to fully restore BlsA functioning in A. baumannii, at least regarding motility ( Figure 3A).
We also analyzed the effects of the expression of the different A. baylyi photoreceptors on the ability of the A. baumannii ATCC 17978.OR cells to form biofilms at 24uC ( Figure 3B and C). In this case, all four A. baylyi BLUF-domain containing photoreceptors restored the original phenotype, behaving as the 17978.ORc BlsA strain when tested as described before [5].

Blue Light and Resistance to Antibiotics
As many clinically-relevant species of Acinetobacter show an outstanding ability to rapidly evolve resistance to antibiotics, reducing therefore the available therapeutic options [4], we speculated whether blue light also modulates resistance to antibiotics. For this purpose, we conducted disc-diffusion antibiotic susceptibility assays both at 24uC as well as 37uC under blue light or in the darkness, using strains A. nosocomialis ACI 32, A. pitii ACI 988, both of which show photoregulation of motility and biofilm formation, and A. haemolyticus ACI 25. Despite various antibiotics belonging to different groups such as ampicillin (b-lactam), ceftazidime (b-lactam), cephalotin (b-lactam), chloramphenicol (chloramphenicol), ciprofloxacin (fluoroquinolone), gentamycin (aminoglycoside), imipenem (carbapenem b-lactam), meropenem (carbapenem b-lactam), piperacillin (b-lactam) and rifampicyn (rifamycin) were tested, no significant differences were detected between light and dark conditions at either temperature for these strains (Table S1).
We also evaluated the effect of light on antibiotic resistance in A. baumannii strains, as this species is the most frequently recovered in clinical settings and the one that has been extensively reported to exhibit resistance to multiple antibiotics [4]. We used here strain ATCC 17978 (as it has been reported to respond to light), which is sensitive to most antibiotics; as well as strain Ab244 [4], which shows resistance to multiple antibiotics and a slight photoregulation of biofilm formation ( Figure S1). Here again, although many antibiotics belonging to different groups were tested, e.g., imipenem, meropenem, cloramphenicol, rifampicin, ampicillin, amikacyn (aminoglycoside), piperacillin, cefoxitin (b-lactam), cephalotin (b-lactam), cefotaxime (b-lactam), cefepim (b-lactam) and ceftazidime (b-lactam), no significant differences were observed that could result from differential resistance mediated by light (Table S1).

Members of the Genus Acinetobacter
We showed previously that the BLUF domain-containing protein encoded in the A. baumannii ATCC 17978 genome is an active photoreceptor that modulates different traits such as motility and biofilm formation [5], and also that the four BLUF domaincontaining proteins encoded in A. baylyi strain ADP1 are active photoreceptors able to sense light and transduce the signal in the A. baumannii genetic background modulating motility and biofilm formation in this organism. To further extend our knowledge of BLUF domain-containing proteins, we analyzed the presence and phylogenetic relationships of these domains in other members of the genus Acinetobacter, as they could likely play similar roles in other species. For this purpose, we screened the complete (or almost complete) sequenced genomes available in databases of members of the genus Acinetobacter for the presence of genes coding for BLUF-domain containing proteins, to determine if their presence is distributed in the genus and evaluate their phylogenetic relatedness. Genes coding for BLUF-containing proteins were present in all of the screened genomes, i.e. those of A. baumannii, A. baylyi, A. calcoaceticus, A. johnsonii, A. lwoffii, A. nosocomialis, 'A. oleivorans', A. pittii, A. radioresistens, and Acinetobacter sp. ATCC 27244 ( Table 2). It is worth mentioning that based on the Pasteur MLST scheme [20], as well as rpoB sequence comparisons and phenotypic analyses [21], 'A. oleivorans' DR1 is highly related to one of the two strains designated 'Between 1 and 39, being therefore also a member of the ACB complex [A. Nemec, unpublished data].
As shown in Table 2, the number of predicted BLUF-containing proteins encoded per genome in the above species fluctuates from one to six. Indeed, it is noteworthy that close species such as those comprised within the ACB complex show variability in the number of genes coding for BLUF-domain containing proteins. As seen in Table 2, A. calcoaceticus PHEA-2, A. pittii SH024 and 'A. oleivorans' DR1 encode two, while A. baumannii and A. nosocomialis RUH2624 encode one and three BLUF-domain containing proteins, respectively. Besides, A. baylyi ADP1 encodes four putative BLUF-photoreceptors while A. radioresistens SK8 and SH164 encode six and four, respectively. All of the predicted BLUF-proteins found in members of the genus Acinetobacter correspond to the most common bacterial BLUF photoreceptors, small proteins containing a flavin-binding photosensing core lacking a recognizable effector or output domain(s), such as BlsA from A. baumannii [5].
To determine the phylogenetic relationships between the BLUF domains present within the Acinetobacter genus, we retrieved 93 protein sequences corresponding to the BLUF domains of predicted and known blue-light photoreceptors of different members of this genus, and also of organisms belonging to different taxa such as a, b, c and d Proteobacteria; and from eukaryotes such as Euglenozoa and Fungi (Figure 4) [22]. In some cases such as those of Euglena and Eutreptiella, which contain two BLUF-domains in the same protein molecule, both sequences were included in the analyses. Figure 4 shows the maximum likelihood phylogenetic tree constructed from the alignments of the above sequences, whereby bootstrap values were calculated by the neighbor joining and parsimony methods. The tree clearly illustrates that all the BLUF domains present in members of the genus Acinetobacter are grouped together in a well-supported monophyletic cluster (bootstrap values of 100% and 99% by NJ and parsimony, respectively), suggesting that all of these putative photoreceptor domains share a common origin. The Acinetobacter cluster contains two major branches, B1 and B2 (Figure 4). Each branch contains at least one paralog gene from each species (similar colors in B1 and B2), with the exceptions of A. lwoffii staining with crystal violet. C. Quantification of the biofilms of cognate samples shown in B. Error bars show standard error of the mean for 3 different biological replicates (n = 3). OD 580 / 600 , optical density at 580 or 600 nm, respectively. L: light; D: Dark. doi:10.1371/journal.pone.0055059.g003 (whose paralogs are restricted to B2, light blue sequences) and all A. baumannii strains, each showing one gene coding for highly similar BLUF-domain containing proteins (blsA), also concentrated in B2.
It is worth noting that, with the only exception of Psychrobacter, the other genera of the Moraxellaceae family, i.e. Alkanindiges, Enhydrobacter, Moraxella, Oligella, Paraperlucidibaca and Perlucidibaca, do not contain genes coding for BLUF-containing proteins. Despite the presence of BLUF domains in Psychrobacter, as exemplified by the homolog present in P. cryohalolentis, this sequence does not group together with those of Acinetobacter (Figure 4), indicating significant divergence from BLUF domains present in members of this genus. Overall, no horizontal gene transfer between members of the Acinetobacter genus and other non-Acinetobacter BLUF-containing organisms is evident from the tree. Most likely, all BLUF-domain containing genes found in currently existing species of the Acinetobacter genus evolved from a aboriginal BLUF-domain containing gene already present in the common ancestor of the genus. Further analysis allows the differentiation of four clusters supported by bootstrap values (C1, C2, C3 and C4, Figure 4) that group together BLUF domain homologs from different species (particularly clear for members of the ACB complex). For example, cluster C4 within B2 groups BLUF domain sequences of members of the ACB complex such as all A. baumannii BlsA proteins, with homologs present in A. calcoaceticus, DR1, A. nosocomialis and A. pittii, which in turn share a monophyletic cluster with A. baylyi (ACIAD2110), A. radioresistens, and Acinetobacter ATCC 27244. Cluster C2 within B1 groups together another set of BLUF domain homologs also present in members of the ACB complex including A. calcoaceticus, DR1, A. nosocomialis and A. pittii, all of them composing a monophyletic cluster with A. baylyi BLUF domain (ACIAD2125). Finally, in cluster C1 (B1) are grouped together a third BLUF domain paralog present in A. nosocomialis RUH2624 sharing cluster with homologs present in A. baylyi (ACIAD1499 and ACIAD2129) and A. radioresistens. Overall, multiple duplications and differential gene losses occurring during the evolution of the different Acinetobacter lineages could explain the presence of different number of paralogs of BLUF-domain containing genes in the diverse species composing the genus. Besides, the homologs from different species grouped altogether in different clusters (C1-C4) could correspond to orthologs; i.e. they were separated by a speciation event, but also could have resulted from horizontal gene transfer of BLUFcoding genes among Acinetobacter species.
Finally, no robust bootstrap values where obtained for branches grouping non-Acinetobacter sequences, probably due to structural more than sequence similarities between the different BLUF domains.

Presence of BLUF-coding Genes in Strains Showing Photoregulation of Motility and Biofilm Formation
Finally, we analyzed whether strains that presented photoregulation of motility and biofilm formation such as A. calcoaceticus ACI 412, A. nosocomialis ACI 32 and A. pittii ACI 988 contained BLUF domain-containing genes. For this purpose, we used information derived from the genome-sequenced strains A.
calcoaceticus PHEA-2, A. nosocomialis RUH 2624 and A. pittii SH 024 to design specific primers that amplify the cognate BLUF coding genes (acc. numbers ADY82057 and ADY82317 for A. calcoaceticus; EEW98085, EEX00046 and EEX01065 for A. nosocomialis; and EFF86081 and EFF86339 for A. pittii), and investigated their presence by PCR and nucleotide sequencing. We found the presence of homologs showing 100% identity to ADY82057 and ADY82317, in A. calcoaceticus ACI 412, EEX00046 and EEX01065 in A. nosocomialis ACI 32, and EFF86081 and EFF86339 in A. pittii ACI 988 ( Table 2). We were not able to obtain an amplification product in the case of A. nosocomialis EEW98085, despite we assayed different amplification conditions.

Discussion
In this work, we show that light regulation is not restricted to A. baumannii but is rather widespread within the genus Acinetobacter. In fact, we found that blue light effectively regulates motility and biofilm formation at 24uC in many Acinetobacter species, including members of the ACB complex, such as A. calcoaceticus, A. nosocomialis and A. pittii (see Figure 1 and Table 2). Yet, in contrast to A. baumannii, in which the formation of biofilms is inhibited under blue light while stimulated in the dark, the opposite was observed in all the other species where blue light regulation was detected: biofilm formation was inhibited in the dark while stimulated under blue light. In non-A. baumannii species where both biofilm formation and motility were regulated such as A. baylyi, A. calcoaceticus, A. nosocomialis and A. pittii, low biofilm correlated with high motility under dark conditions. This response makes sense since a large amount of data support the notion that motility and biofilm formation can be expected to be mutually exclusive and counter-regulated, i.e. being sticky seems counterproductive for moving around, whereas adhesion and settling down might require reduced activity of the motility machinery [23][24][25]. Therefore, in these bacteria blue light contributes to the decision between motility and sessility and also may facilitate acclimation to different environments. In our previous work, we showed that blue light regulation in A. baumannii occurred only at low temperature, suggesting that it is important during its life in the environment perhaps allowing bacteria to sense environmental locations outside the human host [5]. However, the results presented here show that many environmental species such as A. baylyi, A. calcoaceticus and A. tjernbergiae, as well as the clinically relevant species A. nosocomialis and A. pittii displayed blue light regulation of motility also at 37uC. This differential behavior at 37uC compared to A. baumannii may result from the extra content of BLUF-domain putative photoreceptors encoded in the genomes of these non-A. baumannii species (Figure 4 and Table 2). Indeed, only one photoreceptor is encoded in the A. baumannii genome. The protein present in strain ATCC 17978, BlsA, most probably functions only at 24uC: blsA mRNA levels at 37uC are significantly lower with respect to levels at 24uC [5], and the content of BlsA protein in the cells at 37uC is negligible or null [Mussi et al., unpublished data]. Alternatively, the differential behavior at 37uC between A. baumannii and other Acinetobacter species could result from idiosyncratic differences in expression patterns of photoreceptors and/or pathways and partners modulating motility functions downstream the photosensing step in these organisms. In any case, A. baumannii might have become ''blind'' to light at 37uC because there is no positive selection to respond to this stimulus in the relative darkness of the warm-blooded host tissues.
The ability of A. baylyi's BLUF-domain containing genes to restore photoregulation at 24uC in the ATCC 17978 blsA mutant not only confirms that they encode bona fide BLUF photoreceptors, but also that they can transduce the light signal into A. baumannii motility and biofilm regulatory cascades, probably by using BlsA partner/s. In this context, A. baylyi formed large amounts of biofilms under blue light and almost negligible amounts were produced under dark conditions. It is therefore noteworthy that the opposite situation is observed when A. baylyis photoreceptors are expressed in the A. baumannii blsA mutant, restoring in all cases the wild type phenotype corresponding to this species. Therefore, both cascades in A. baumannii and A. baylyi seem to be inversely affected at 24uC, independently of the origin of the photoreceptor used.
Moreover, we found that BLUF-domain containing genes, shown to be active photoreceptors in A. baumannii and A. baylyi, are present in all completely sequenced genomes available for members of this genus, and also in strains showing photoregulation of motility and biofilm formation, such as A. calcoaceticus ACI 412, A. nosocomialis ACI 32 and A. pittii ACI 988. The variable number of genes coding for BLUF-domain containing proteins (from one to six) in the different species analyzed in this study suggests that sensing and responding to light might be of differential importance among them, probably reflecting their different lifestyles and the diversity in niches in which they thrive. Phylogenetic analysis suggests a common origin for all BLUF domains within Acinetobacter and could distinguish well-differentiated clusters that group together BLUF homologs from different species, a situation particularly clear for members of the ACB complex, which most likely correspond to groups of orthologs. The different clusters may reflect a closer phylogenetic relationship among the species with related orthologs, or the observed clustering could have resulted from horizontal gene transfer of BLUF-coding genes among Acinetobacter species. Finally, we were not able to detect photoregulation of neither motility nor biofilm formation in many species or strains of the genus Acinetobacter, despite the presence of multiple BLUF-domain containing genes (such as the 4-6 putative BLUF-domain photoreceptors present in the A. radioresistens genomes, Table 2). This is not surprising since a similar situation has been previously reported by us in the case of A. baumannii strain ATCC 19606 T [5]. Many possibilities could lead to the above results: the machineries of motility or biofilm formation may be impaired in some of these strains, or their specific components may not be expressed under the conditions studied. Still, light could be regulating other cellular processes not tested in this work. Besides, other possible explanations suppose that photoreceptor genes in these strains might not be expressed under the assayed conditions, the corresponding proteins may not be active, or downstream partners of the signaling cascade or targets of the photoreceptors involved in these phenotypes might be missing. Leaving aside particular cases, which would not be unexpected since a high genetic heterogeneity has been described for strains of some members of Acinetobacter such as A. baumannii [4,5,26], the conservation of multiple BLUF-domain containing genes as well as the fact that they have not been subjected to genetic derive, suggests that they have been maintained to play important roles in the bacterial physiology. Most probably, these putative photoreceptors could serve functions modulating other cellular processes that remain unidentified in these species. Yet, functional characterization of BLUF domain-containing proteins encoded in species of the genus Acinetobacter other than those of A. baumannii or A. baylyi still needs to be conducted to ascertain their role as photoreceptors involved in light perception in these microorganisms. Regarding other cellular processes affected by light, it is worth mentioning that we could not detect light regulation of resistance to antibiotics in different clinically-relevant species of the genus Acinetobacter. Nevertheless, further research would contribute to draw a final conclusion in this sense.
Our understanding of the signal transduction mechanisms and regulatory cascades involved in A. baumannii BlsA and its homologs present in other species of the genus Acinetobacter is still scarce and currently under study in our laboratory. The final goal is to gain a full comprehension of light regulation in relation to host's niches and lifestyles, which would perhaps need further understanding of Acinetobacter biology. Figure S1 Quantification of the biofilms produced by different strains showing photoregulation of biofilm formation within the genus Acinetobacter. Error bars show standard error of the mean for 3 different biological replicates (n = 3). OD 580 / 600 , optical density at 580 or 600 nm, respectively. (TIF) Table S1 Blue light and resistance to antibiotic. Disc difussion antibiotic susceptibility assay under blue light or in the dark at 24 or 37uC, of some strains of A. nosocomialis, A. pittii and A. baumannii which showed photoregulation of motility and/or biofilm formation. The A. haemolyticus strains analyzed in this work did not show photoregulation neither of motility nor biofilm formation, but one strain was included in this study due to the importance of this species in the clinical settings. The diameter of inhibition from three independent experiments (mm +/2 SEM of three biological replicates) is indicated. AM, ampicillin; AN, amikacyn; FEP, cefepim; CTX, cefotaxime; FOX, cefoxitin; CAZ, ceftazidime; CF, cephalotin; C, chloramphenicol; CIP, ciprofloxacin; IPM, imipenem; GM, gentamycin; MEM, meropenem; PIP, piperacilin; RA, rifampicin. (XLSX)