Characterization of BLUF-photoreceptors present in Acinetobacter nosocomialis

Acinetobacter nosocomialis is a Gram-negative opportunistic pathogen, whose ability to cause disease in humans is well recognized. Blue light has been shown to modulate important physiological traits related to persistence and virulence in this microorganism. In this work, we characterized the three Blue Light sensing Using FAD (BLUF) domain-containing proteins encoded in the A. nosocomialis genome, which account for the only canonical light sensors present in this microorganism. By focusing on a light-modulated bacterial process such as motility, the temperature dependence of light regulation was studied, as well as the expression pattern and spectroscopic characteristics of the different A. nosocomialis BLUFs. Our results show that the BLUF-containing proteins AnBLUF65 and AnBLUF46 encode active photoreceptors in the light-regulatory temperature range when expressed recombinantly. In fact, AnBLUF65 is an active photoreceptor in the temperature range from 15°C to 37°C, while AnBLUF46 between 15°C to 32°C, in vitro. In vivo, only the Acinetobacter baumannii BlsA’s ortholog AnBLUF65 was expressed in A. nosocomialis cells recovered from motility plates. Moreover, complementation assays showed that AnBLUF65 is able to mediate light regulation of motility in A. baumannii ΔblsA strain at 30°C, confirming its role as photoreceptor and in modulation of motility by light. Intra-protein interactions analyzed using 3D models built based on A. baumannii´s BlsA photoreceptor, show that hydrophobic/aromatic intra-protein interactions may contribute to the stability of dark/light- adapted states of the studied proteins, reinforcing the previous notion on the importance of these interactions in BLUF photoreceptors. Overall, the results presented here reveal the presence of BLUF photoreceptors in A. nosocomialis with idiosyncratic characteristics respect to the previously characterized A. baumannii’s BlsA, both regarding the photoactivity temperature-dependency as well as expression patterns, contributing thus to broaden our knowledge on the BLUF family.


Introduction
The ability of Acinetobacter spp. to cause disease in humans is well-recognized. While being primarily associated with pneumonia, they are also frequent causes of wound and burn infections, meningitis, urinary tract infections, and sepsis [1]. There is a growing trend for these isolates to display high levels of antibiotic resistance, with some already showing resistance to all clinically available antibiotics [2][3][4][5]. Acinetobacter nosocomialis is a member of the Acinetobacter calcoaceticus-Acinetobacter baumannii (ACB) complex [6]. While A. baumannii predominates over all other members of the ACB complex in terms of incidence, poorer clinical outcomes, and antibiotic resistance rates, A. nosocomialis has gained recognition also as a clinically relevant human pathogen [3][4][5]7]. Many potential virulence factors have been identified in A. nosocomialis and include a CTFR inhibitory factor (Cif), a protein O-glycosylation system, a type-I secretion system, a type-II secretion system, secretion of outer membrane vesicles, the OmpA protein, the CpaA protease, and quorum sensing [1,2,7].
We have previously shown that A. nosocomialis is able to sense and respond to light modulating biofilm formation and motility at 24˚C [8]. Also, we have shown that light modulates persistence, metabolism, the ability to grow under iron limiting conditions and virulence in this microorganism [9]. The genome of A. nosocomialis RUH2624 encodes three Blue Light sensing Using FAD (BLUF)-domain containing proteins, as the only canonical light sensors [8]. Extensive work performed on A. baumannii showed that this microorganism encodes only one BLUF-type photoreceptor, designated BlsA, which functions at low-environmental temperatures up to 25±1˚C and is regulated both at the transcriptional level as well as the activity of the photocycle [10,11]. Other BLUFs from Acinetobacter have been characterized based on light induced phenotypes, gene knockouts and transcriptomic analyses [12,13]. In this work, we present evidence indicating that regulation of motility by light in A. nosocomialis is maintained in a wide range of temperatures from 24 to 37˚C. Recombinant expression, purification and characterization of the different BLUF-domain containing-proteins showed that AnBLUF46 and AnBLUF65 encode stable and active photoreceptors; while, AnBLUF85 could not be characterized. Interestingly, only anbluf65 is expressed in vivo and exhibits a stable photocycle in the temperature-range at which light regulates motility in A. nosocomialis, between 15 and 37˚C, displaying thus differential characteristics with respect to A. baumannii's BlsA. Moreover, complementation assays show that AnBLUF65 is able to mediate light regulation of motility in A. baumannii ΔblsA strain at 30˚C, confirming its role as a photoreceptor and in modulation of motility by light. Spectroscopic characterization and analyses of 3D models built for these proteins provide insights into the intra-protein signaling process connecting the widely characterized BLUF photophysics with the subtle re-arrangements located in the C-terminal part of these proteins. Finally, proteomic analyses revealed that light mainly regulates proteins related to signalling and cellular metabolism.

Bacterial strains, plasmids, and media
Luria-Bertani (LB) broth (Difco) and agar (Difco) were used to grow and maintain bacterial strains. Broth cultures were incubated at the indicated temperatures either statically or with shaking at 200 rpm. by 9-light-emitting diode (LED) arrays, with an intensity of 6 to 10 μmol photons m -2 s -1 . Each array was built using 3-LED module strips emitting blue light, with emission peaks centered at 462 nm, determined using the LI-COR LI-1800 spectroradiometer [15].

Cell motility assays
Cell motility was tested on 1% tryptone, 0.5% NaCl, 0.3% agarose plates inoculated on the surface by depositing 3 μl of Tryptic Soy Broth (TSB) cultures grown to an optical density at 660 nm (OD 660 ) of 0.3. Pairs of plates (one in the dark and the other under blue light) were analyzed, at the indicated temperature. Plates were incubated until the cells of at least one of the plates of the analyzed pair just reached the edge of the dish at each given temperature. The area of plates covered with bacteria was measured with ImageJ (NIH), and then the percentage of plate coverage was calculated. Three independent experiments were performed.

Analyses of gene expression by qRT-PCR
qRT-PCR analysis were done as described in Tuttobene et al., 2019 [18], using primers listed in Table 1. Data are presented as NRQ (Normalized Relative Quantities) calculated by the qBASE method [19], using recA and rpoB genes as normalizers. The anbluf65 transcript levels of each sample were normalized to the rpoB transcript level for each cDNA sample. Relative gene expression to rpoB was calculated using the comparative 2 -ΔCT method [20]. Each cDNA sample was run in technical triplicate and repeated in at least three independent sets of samples. t-test was usedto determine whether two values were significantly different comparing data within each temperature assayed. p-values: � , p < 0.01; �� , p < 0.001.

Cloning, synthesis, overexpression and purification of AnBLUFs
In the case of gene anbluf46, a PCR product amplified from A. nosocomialis RUH2624 genomic DNA using primers 46F and 46R (Table 1), which contain NdeI and BamHI restriction sites, respectively, was cloned into pGEM1T-Easy vector and then subcloned into the NdeI and BamHI sites of pET28-TEV. In turn, anbluf65 and anbluf85 coding sequences were directly synthesized and subcloned into pET 28 a(+) (Genscript, USA). Thus, the different proteins were overexpressed as N-terminal His-tag fusions. E. coli BL21 (pLysS) cells were transformed with plasmid DNA, and then cultured in LB broth supplemented with chloramphenicol and kanamycin at 37˚C until they reached an OD 600 of 0.6 to 0.7. Overexpression of Table 1. Primers designed and used in this study. the His-tagged proteins was induced with 0.5 mM IPTG at 15˚C . After incubation for 5 h at  15˚C to avoid the formation of inclusion bodies, the cells were collected by centrifugation, frozen with liquid nitrogen and lysed by manual grinding with a mortar and pestle in the presence of lysis buffer (20 mM Tris-HCl [pH 8.0], 500 mM NaCl, 1 mM β-mercaptoethanol). Cell debris were removed by centrifugation at 20,000 x g for 30 min at 4˚C, and the supernatant was loaded onto a nickel nitrilotriacetic acid (Ni-NTA)-agarose column (Invitrogen). The column was washed sequentially with lysis buffer containing 20 mM and 40 mM imidazole, respectively, and the His-tagged protein was eluted with the same buffer containing 250 mM imidazole (elution buffer). PD-Minitrap G-25 (GE Healthcare) columns were used to desalt and exchange the buffer to 20 mM Tris-HCl (pH 8.0) and 200 mM NaCl (working buffer). Vivaspin 500 (Sartorius) centrifugal filters (cutoff of 10 kDa) were used to concentrate the purified proteins. The purity of the overexpressed His-tagged protein was confirmed by sodium dodecyl sulfate (SDS)-PAGE 16% gels [21].

Construction of a complementation plasmid
A 884-bp fragment harboring anbluf65 and its predicted promoter was PCR amplified using A. nosocomialis RUH2624 total DNA and primers Panbluf65F and Panbluf65R (Table 1), both of which were tailed with BamHI restriction sites. The amplicon was cloned into pBluescript SK-and then subcloned as a BamHI fragment into the cognate site of pWH1266. Proper construction of the complementing pWH1266 anbluf65 plasmid was confirmed by automated DNA sequencing. Plasmid DNA was electroporated into A. baumannii ATCC 17978 ΔblsA as described before [15].

Sequence analyses
Protein sequence alignments were performed using CLUSTALW (https://www.genome.jp/ tools-bin/clustalw), and the alignments were visualized with Jalview 2 [22]. AnBLUF65 and AnBLUF46 3D modeled structures were performed using the Swiss-Model workspace/GMQE [23] in a search for template mode. In all cases, the best fit corresponded to the BlsA PDB structures (6W6Z for the dark-adapted form and 6W72 for the light-adapted form). Models were visualized and manipulated using PyMOL (by Schrödinger). Ring 2.0 webserver [24] was used to explore the specific π-π stacking interactions in all protein models as well as other non-covalent interactions.

Spectroscopic measurements
Absorption spectra were registered using an Ocean Optics modular UV-Vis spectrophotometer USB2000+. The assays were done using a 5x5 mm quartz cuvette (Hellma, Germany), placed in a Quantum Northwest FLASH 300 cuvette holder connected to a Peltier-based temperature controller, containing 250 μl of air-saturated protein solution in 20 mM Tris-HCl, 200 mM NaCl, pH 8.0 buffer. Scattering effects on the absorption spectra were corrected using a|e-UV-Vis-IR Spectral Software 1.2, FluorTools (www.fluortools.com). The Light-adapted state for AnBLUFs (l-AnBLUF) was obtained by blue light irradiation of the dark-adapted form (d-AnBLUF) with a 1 W Royal Blue LED (Luxeon Star LEDs) emitting at 443 ± 20 nm. Absorbance changes followed at 510 nm (ΔA) vs time (t) were recorded during dark-light cycles at different temperatures. The light-adapted state formation time, τ lBLUF , and the back-recovery time through thermal recovery to the dark-adapted state, τ rec , were determined using the exponential Eqs 1 and 2, respectively, where A is the pre-exponential factor and ΔA 0 is the initial absorbance: Fluorescence emission spectra were recorded with a Hitachi F-2500 spectrofluorometer equipped with a Hamamatsu R-928 photomultiplier. Neutral density filters (10%T) were placed onto the excitation output beam to minimize photochemical processes during acquisition. Emission spectra were obtained by selective excitation of FAD cofactor at 460 nm. Temperature control was performed using a circulating fluid bath (Haake F3) connected to the cuvette holder.

Protein extraction
A. nosocomialis was grown in LB at 37˚C under blue light or in the dark for 24 h without shaking, and the procedure was repeated to generate three independent biological replicates. The cells were recovered by centrifugation at 7,000 g for 10 min at 4˚C. The cells were then resuspended in 500 μl of extraction buffer (25 mM Tris-HCl pH 7.0), 0.5% Tween 20, 2 mM EDTA, 5 mM, β-mercaptoethanol) and lysed by sonication while keeping samples in ice. Cell debris and non lysed cells were collected by centrifugation at 12,000 rpm for 15 min at 4˚C and the supernatants were carefully decanted into clean tubes and stored at −80˚C. Total protein content present in the supernatant was quantified using bicinchoninic acid (BCA) (Thermo Scientific, Germany). Then, 30 μg total proteins were loaded in SDS-PAGE (10% stacking gel and 5% running gel), and allowed to separate electrophoretically only 1 cm long within the separation gel. The gel was incubated for at least 3 h in fixing solution: 30% v/v ethanol, 2% H 3 PO 4 , and after being washed with MiliQ water, it was incubated for 1 h in staining solution: 18% v/v methanol, 17% w/v (NH 4 ) 2 SO 4 and 2% v/v H 3 PO 4 under vigorous shaking. Then, Coomassie G250 powder (0.5 g/L) was added and further incubated for 1 or 2 days until stained proteins were visible, which were then cut from the gel and sent to the Proteomics Core Facility of CEQUIBIEM at the University of Buenos Aires where protein digestion and Mass Spectrometry analysis were performed. Samples were resuspended in 50 mM (NH 4 )HCO 3 pH 8.0, digested overnight with sequencing-grade modified trypsin (Promega) and desalted with Zip-Tip C18 (Merck Millipore). Proteins were analyzed by nanoHPLC (EASY-nLC 1000, Thermo Scientific, Germany) coupled to a mass spectrometer with Orbitrap technology (Q-Exactive with High Collision Dissociation cell and Orbitrap analyzer, Thermo Scientific, Germany). Peptide Ionization was performed by electrospray. Data were analyzed with Proteome Discoverer 2.1 software (Thermo Scientific, Germany) for identification and area quantitation of each protein. Protein identification was performed using Acinetobacter nosocomialis strain Ab1 protein collection as reference (UP000244598; https://www.uniprot.org/uniprot/?query=proteome:UP000244598) since a fully annotated A. nosocomialis RUH2624 proteome is not currently available.
Missing value imputation method [25] was applied to the analyzed results, and Perseus software v1.6.1.3 (Max Planck Institute of Biochemistry) was used to perform the statistical tests.
Proteins showing a fold change (FC) � 1.5 between light and dark conditions and pvalue < 0.05 were considered differentially produced.

Proteomics bioinformatics
Biological functional analyses of differentially over-represented proteins detected by proteomic approaches were categorized according to their molecular function, biological process, and cellular component, by using the Blast2GO tool.

Light modulates motility in a wide range of temperatures in A. nosocomialis RUH2624
We have recently shown that in A. baumannii ATCC 17978, light modulates motility from low to moderate temperatures up to 25±1˚C through the BLUF photoreceptor BlsA [11,15]. In contrast, in A. nosocomialis RUH2624 the response to light is different. In fact, our results show that in this microorganism light inhibits motility in a wide range of termperatures from 25 up to at least 37˚C, the highest temperature tested, reaching 10-20% of plate coverage ( Fig  1A and 1B). On the contrary, the bacteria spread to cover the whole plates at all temperatures assayed in the dark (Fig 1A and 1B). Thus, in contrast to A. baumannii, in A. nosocomialis RUH2624 light modulates motility not only at environmental temperatures, but within a wide range of temperatures including those found in warm-blooded animals such as 37˚C. This differential behavior prompted us to characterize A. nosocomialis' photoreceptors, as we envisaged differential properties from the A. baumannii's BlsA. i.e., the percentage of the Petri plate area covered with bacteria, in motility plates inoculated with RUH2624 wild type and incubated at the indicated temperatures. Three independent experiments were performed in each case. The area of plates covered with bacteria was measured with ImageJ (NIH), and then the percentage of plate coverage was calculated. The mean and standard deviation is informed. Different letters indicate significant differences as determined by ANOVA followed by Tukey's multiple comparison test (p < 0.05). For those conditions at which the bacteria just reached the edge of the plate a value of 100% is informed. (C). cDNA from A. nosocomialis RUH2624 cells grown in motility plates in the presence of blue light (L) or in darkness (D) at different temperatures was used as template for quantitative real-time PCR using anbluf65 specific primers. Transcription of rpoB was used as a constitutively expressed internal control. The results are representative of three independent experiments. The mean and standard deviation are shown. t-test was used to determine whether two values were significantly different comparing data within each temperature assayed.

A. nosocomialis RUH2624 encodes three BLUF-type putative photoreceptors
Comparative sequence analyses have shown the presence of three BLUF domain-containing proteins encoded in the A. nosocomialis RUH2624 genome [8]. These putative photoreceptors received accession numbers EEW98085, EEX00046, and EEW01065 in the Genbank database and will be further referred to here as AnBLUF85, AnBLUF46 and AnBLUF65, respectively.
Sequence alignments of these three BLUF-domains containing proteins with other well characterized BLUF domains such as Slr1694 from Synechocystis sp., Tll0078 from Thermosynechococcus elongatus, AppA and BlrB from Rhodobacter sphaeroides, and BlsA from A. baumannii, confirmed the presence of 22 highly conserved residues in the N-terminus of canonical BLUF domains according to pfam04940 (Fig 2A). Several BLUF domains belonging

PLOS ONE
to different species from the genus Acinetobacter have been described so far as functional blue light photoreceptors [10][11][12]15]. Alignment of these functional Acinetobacter BLUF-photoreceptor sequences show a higher level of conserved residues in their N-terminus domain (33 residues, Fig 2B), while the analysis of all γ-Proteobacteria, including 30 Acinetobacter sequences deposited in pfam04940 displays lower conservation (14 residues, Fig 2C) in their primary structure and in some cases, insertions between the conserved Tyr and Gln (Fig 2A), though with a low occupancy level ( Fig 2B).
Phylogenetic analyses indicate that AnBLUF65 and A. baumannii's BlsA BLUF domains are grouped in a monophyletic cluster [8], thus indicating that they are orthologs. anbluf65's genomic context shares many features with blsA's (Fig 3). anbluf65's 5' upstream region is highly conserved with respect to the 5' upstream region of BlsA in A. baumannii ATCC 17978 (Fig 3), as well as with different strains of A. nosocomialis (S1 Fig). Interestingly, the presence of insertion sequence ISAba27 was detected in the intergenic region upstream of the Bof coding sequence in A. nosocomialis RUH2624 (Figs 3A and S1). Whether ISAba27 is providing a hybrid promoter or disrupting the transcription of a putative operon that includes anbluf65 is a possibility that deserves further experimentation. A. nosocomialis RUH2624 anbluf65 3' downstream region is more variable compared both with BlsA or with other strains of A. nosocomialis (Figs 3 and S1). Gene anbluf46 is located in a different genome location, surrounded by genes coding for a putative succinyl-CoA:3-ketoacid-coenzyme A transferase, a LysR transcriptional regulator, a hypothetical protein, a putative heat shock protein, a conjugal protein, and another hypothetical protein on the 5' upstream region (Fig 3A). On the 3' downstream region there are encoded: a putative protein with a metallo-β-lactamase fold, an H + / gluconate symporter, a hydroxybutyrate dehydrogenase, an AraC transcriptional regulator, and an aromatic acid/H + symporter (Fig 3A). anbluf46 does not seem to be part of an operon, as both flanking genes are encoded in the opposite direction ( Fig 3A). Comparing anbluf46 genomic context in strain RUH2624 with other strains of the same species, it is observed that the 3' downstream region is highly conserved while the 5' upstream region is variable, showing the presence of genes coding for transposases in different locations (S2 Fig). anbluf85 was only found encoded in A. nosocomialis RUH2624 and UBA873 strains (100% identity). In RUH2624, anbluf85 is flanked by a 5' upstream region coding for an acetyltransferase and a putative conjugal transfer pilus assembly protein TraB. On the 3' downstream region there are encoded a hypothetical protein, a FAM199X domain-containing protein, and a relaxase MobA/MobL ( Fig  3A). RUH2624 anbluf85 genomic context is different from that in UBA873 (S3 Fig). In the latter, only the 3' downstream region is shown as the contigs are not assembled yet (S3 Fig). Sequence comparisons indicate the presence of an anbluf85 homolog in Acinetobacter lwoffii strain M2a plasmid pAVAci116 as well as in Acinetobacter sp. ACNIH1 plasmid pACI-148e with 100% identity at nucleotide level. Interestingly, an anbluf85 homolog with 98.13% identity at the nucleotide level was also reported to be present in Acinetobacter baumannii isolate KAR [26]. Moreover, 100% identity homologs at the aminoacidic level are also present in A. baumannii strains A30, A10, and MSHR_A189, two in A. radioresistens strains 50v1 and TG29429; and in A. baumannii strains MRSN7353 and ARLG1306 showing 1 amino acid difference with AnBLUF85. Thus anbluf85 is not widely distributed but its presence is observed discretionally in some Acinetobacter species, and the possibility of horizontal gene transfer is suggested.

anBLUF65 and anBLUF46 produce stable and active photoreceptors in vitro
We evaluated next whether the three BLUF-domain containing genes encode active photoreceptors. For this purpose, genes anbluf65, anbluf46 and anbluf85 were recombinantly expressed and the resulting proteins were purified to assess their functionality. Fig 4 shows the UV-vis spectra of AnBLUF65 (92% identity with BlsA) and AnBLUF46 (45% identity with BlsA) at 15˚C in working buffer. Spectra were initially collected at this temperature to imitate conditions at which A. baumannii's BlsA was previously analized and reported [10,11]. Before blue light irradiation, both dark-adapted states of AnBLUF65 (d-AnBLUF65) and AnBLUF46 (d-AnBLUF46) showed the typical absorption band of the fully oxidized flavin corresponding to the transition S 0 !S 1 with λ max � 460 nm and shoulder approximately at 483 nm. Upon blue-LED illumination, the band is red-shifted, characteristic of the formation of the light-adapted state of BLUF proteins (l-BLUF) [27], with absorption maximum wavelength shift of 8 and 12 nm for l-AnBLUF65 and l-AnBLUF46, respectively (Fig 4A and 4C). The differential absorption spectrum (ΔA) between the light-and darkadapted states of AnBLUF65 showed positive changes at 508 and 472 nm, and negative changes at 455 and 427 nm. AnBLUF46 also showed positive changes at 511 and 476 nm, whereas negative bands were less defined (insets Fig 4A and 4C). The overall results indicate that AnBLUF65 and AnBLUF46 are active photoreceptors. Gene anbluf85 was weakly expressed and the partially purified product was barely stable precipitating at T higher than 12˚C, and thus was not further characterized.
Fluorescence emission spectra obtained by excitation at 460 nm at 15˚C for d-AnBLUF65 and d-AnBLUF46 also depicted a BLUF-like behavior, with an emission maximum at 512 and 519 nm, respectively (Fig 4B and 4D). The fluorescence bands of both proteins were blueshifted compared with free FAD in the same working buffer [10,11] confirming the existence of specific interactions between the isoalloxazine group of the flavin with the surrounding residues, and reducing solvent accessibility within the flavin binding pocket [28]. When the fluorescence spectrum of the light-adapted and the dark-adapted forms of both photoreceptors were compared, they exhibited the typical red shift of approximately 5 nm of the fluorescence maximum in addition to a loss of vibratory structure and a lower emission intensity, suggesting a more relaxed environment around the cofactor [10,[29][30][31].

AnBLUF65 and AnBLUF46's photoactivity temperature-dependency
A. nosocomialis modulates motility in response to blue light in a similar fashion as observed for A. baumannii, but in a wider range of temperatures. Table 2 summarizes the temperature dependency of visible absorption and emission fluorescent properties of flavin in purified AnBLU65 and AnBLUF46 photoreceptors. Absorption maxima of the dark-adapted forms of both proteins were blue-shifted when temperature increased, reaching in the case of d-AnBLUF46 the value of free FAD in solution (� 450 nm). The ratio between the shoulder absorbance at � 480 nm and the absorption maximum (A sh /A max ) diminished with temperature increments, as did the redshift upon blue light illumination, suggesting a distancing between the conserved Tyr to the isoalloxazine ring [32]. Fluorescence emission properties of the cofactor were significantly affected upon formation of l-AnBLUF65 at 24 and 37˚C, with a redshift in response to blue light, as described before at 15˚C or when temperature increased, implying a relaxation in the hydrogen bond network surrounding the isoalloxazine, and/or changes in the polarity in the chromophore cavity as a consequence of increased solvent accessibility. Widening of the emission spectrum full-width half-maximum (FWHM) in response to the formation of the light-adapted state for AnBLUF65 was observed at 24 and 37˚C, and this effect was also evident when temperature increased. Overall, these results indicate that the flavin environment in AnBLUF65 is affected by temperature, as it has been shown for BlsA [10,11]. In contrast, the maximum emission wavelength and FWHM for the dark-or the light-adapted states of AnBLUF46 did not show significant changes neither in response to blue light at 24 and 32˚C, nor to temperature increments. In fact, d-and l-AnBLUF46 isoalloxazine pockets appeared to be more polar than the same states of AnBLUF65. Chromophore environments in AnBLUF65 and AnBLUF46 were somewhat dissimilar and appeared to react in a different way to the formation of the light-adapted state. This fact is in good agreement with structural predictions from models of these two proteins generated using the crystallographic structure of dark-and light-adapted states of BlsA (PDBs 6W6Z and 6W72, respectively) as templates (Fig 5). Fig 5A shows the location of water molecules (blue dots) for the dark-adapted form of BlsA (d-BlsA) and the light-adapted (l-BlsA) respectively, determined by Chitrakar et al. [33]. The flavin pocket access site has more affinity and/or accessibility to water molecules in the light-adapted form, supporting our previous characterization of chromophore behavior inside BlsA cavity by the flavin fluorescence emission (Fig 5A) [10]. Modeled structure of BlsA ortholog AnBLUF65 (Fig 5B) shows that the environment of the cofactor is conserved compared with BlsA. This means that analogously to BlsA, the flavin pocket in the light-adapted form of AnBLUF65 is more solvent-accessible. In contrast, the solvent accessibility to the chromophore cavity between d-and l-AnBLUF46 is indistinguishable (Fig 5C). In addition, the flavin binding site in AnBLUF46 seems more solvent-accessible than in BlsA/AnBLUF65. Hence, the analyses of these modeled structures, based upon BlsA structures, are consistent with the differential influence of temperature and blue light.
Insets in Fig 4 show difference spectra both for AnBLUF65 and AnBLUF46 (lAnBLUF-dAnBLUF). In previous work, we have used the ΔA at 510 nm at different temperatures to follow the kinetics of formation of the light-adapted state of each protein, and in darkness, the recovery time, through cycles of blue light illumination. This kinetic profile was used to determine the quantum yield of the protein light-adapted formation, F lAnBLUF . Fig 6 presents the kinetic profiles of AnBLUF65 (Fig 6A and 6B) and AnBLUF46 (Fig 6C and 6D) tested at two temperatures. For AnBLUF65 the kinetic curves were monitored at 24˚C and 37˚C, which correspond to environmental and warm-blooded host temperatures, respectively. However, AnBLUF46 was tested at 24˚C and 32˚C, the maximum temperature at which this photoreceptor presented activity. Table 3 summarizes the light-adapted-state formation time τ lAnBLUF , and the recovery time τ rec , for both BLUFs as a function of temperature along with the photoactivation quantum yield of light-adapted state, F lAnBLUF . Temperature increments had a significant effect on the τ lAnBLUF for both proteins, with reductions of 25 and 40% for τ lAnBLUF65 and τ lAnBLUF46 , respectively. AnBLUF46 had similar values to BlsA at 24˚C (Table 3), however, τ lAnBLUF65 was larger than that of BlsA. Analyses of τ rec show that AnBLUF46 takes 15 mins to recover, nine times slower than BlsA; while AnBLUF65 was only two-fold slower than the A. baumannii protein. As expected for a thermal process, τ recAnBLUF46 was nine times faster when temperature increased to 32˚C, becoming similar to the value of τ recBlsA at 24˚C. However, this was not the case of AnBLUF65, whose τ recAnBLUF65 remained constant at 24 and 37˚C. This behavior prompted us to speculate whether this recovery time might have some physiological meaning. AnBLUF65, as evidenced

PLOS ONE
by maximum emission wavelength and full-width at half maximum (Table 2), presents conformational changes with temperature in its chromophore binding cavity upon illumination. Structural data (Fig 5) supported fluorescence behavior, however, the time to recover to the dark form is similar for both temperatures tested.
Despite the high sequence similarity between AnBLUF65 and BlsA (92%), the behavior of the former was significantly different compared with that of the latter, regarding the efficiency to respond to blue light generating the light-adapted form. In fact, F lAnBLUF65 was more than

PLOS ONE
two-fold higher than that of BlsA at 24˚C, which can be interpreted as AnBLUF65 being more efficient to form the light-adapted state. The τ lAnBLUF65 was also significantly higher than that for BlsA, although to a lesser extent, and the τ recAnBLUF65 was two-fold higher than τ recBlsA , probably because the adapted form is more stabilized in AnBLUF65 than in BlsA. Another contrasting effect between such similar proteins is that AnBLUF65 is less prone to aggregate upon temperature increments than BlsA, which shows this effect macroscopically above 30˚C, while AnBLUF65 does not aggregate at 37˚C showing only incipient turbidity in the buffer solution. While F lBlsA reaches non-detectable values at T higher than 28˚C, F lAnBLUF65 at 37˚C is 0.12±0.03. In contrast, F lAnBLUF46 was not significantly impacted between 24˚C and 32˚C. Nevertheless, AnBLUF46 lost all photoactivity at T>32˚C. These results suggest that AnBLUF65 is more efficient to form the light-adapted state than AnBLUF46 at 24˚C. Overall, evidence presented in this section indicates that AnBLUF65 is an active photoreceptor in the temperature range from 15˚C to 37˚C, while AnBLUF46 is active between 15˚C to 32˚C.

Intramolecular interactions and their role in intra-protein signaling
So far, we have analyzed these photoreceptors describing intrinsic phenomena occurring mostly in the N-terminal side of the protein, where the flavin resides in its cavity. These data have profiled AnBLUF65 and AnBLUF46 in the first part of the photo-reception process. Next, we aimed to expand our analyses to cover the subtle structural rearrangements between the dark and the light-adapted states of the protein. Intra-protein changes in hydrophobic and aromatic residue interactions driven upon illumination have been suggested by Chitrakar et al, 2020 to explain how the photo-signal might be translated from the flavin surroundings to the variable domain. Based on the similarity between BlsA and AnBLUF65, we modeled structures for both A. nosocomialis proteins, AnBLUF65 and AnBLUF46 (Fig 5), using BlsA as template. We used RING 2.0 [24] to profile all intramolecular interactions, such us π-π stacking, Hbonding, Van der Waals (VDW) interactions, and salt bridges, in AnBLUF65 and AnBLUF46, as presented in S1 Table. The analysis of π-π stacking interactions amongst aromatic residues helped us to profile aromatic networks for both proteins, in both dark-and light-adapted forms.
The first aromatic cluster found surrounds the flavin cofactor and involves the hydrophobic side of the isoalloxazine ring (ring I): the conserved Tyr, a Phe and a His for AnBLUF65 as shown in Fig 7. This His belongs to the already described motif Asp-X-Arg-His and it has been suggested that the protonation/deprotonation of the conserved Gln via His is energetically favourable [34]. In AnBLUF65, His73 forms π-π stacks with Tyr7 and with Phe49 in addition to other interactions i.e. H-bonds that have been described before. After blue light illumination this cluster does not show a significant impact on its Cα main chain or in the angle of the aromatics involved. This might be necessary to conserve the hydrophobicity in the flavin pocket [35].
The next aromatic network found is in the variable domain (C-terminal) and contains two aromatic tetrads, that are formed by six Phe in AnBLUF65 (Fig 7). Upon blue light illumination, distances between aromatic centroids in these tetrads are shortened in all cases in AnBLUF65. Phe106 loses its π-π stack interaction with Phe115, being left only with VDW interactions. Thus, the formation of the light-adapted state of AnBLUF65 requires a local rearrangement of α-helices in the C-terminal and the aromatic network in this area is strengthened.
A Tyr-Tyr motif, highly conserved in Acinetobacter BLUFs (Fig 2B) also participates in the aromatic networks described before. AnBLUF65 has Tyr44 interacting with Phe49 on the chromophore side of the protein, while Tyr43 interacts through π-π stacking with Phe128 on the other side of the β-sheet, towards the variable domain (Fig 7). The notable re-arrangement is found towards the variable domain may be possible because an Asp residue located in the loop between the Tyr-Tyr motif and the first Phe-Phe motif, rotates upon blue light illumination, losing its contact with Phe115. Thus, one way to explain these findings is that the aromatic clusters found in the variable domain might internally stabilize the structure of this protein upon the formation of the light-adapted state. In contrast, the aromatic cluster in the flavin/N-terminal side of these proteins does not suffer any alteration when changing states. Taken together, it suggests that intramolecular hydrophobic/aromatic interactions might have a role in the elasticity of the protein monomer to go back and forth between the two states.

anbluf65 is expressed in a wide range of temperatures in A. nosocomialis RUH2624 cells recovered from motility plates
The expression levels of the three BLUF-type photoreceptors encoded in A. nosocomialis RUH2624 were analyzed on cells recovered from motility plates incubated at a wide range of temperatures (25-35˚C) under blue light or in the dark by qRT-PCR as described previously [14]. We were able to detect expression of anbluf65 only, the blsA ortholog in A. nosocomialis [8]. anbluf65 was expressed at 25, 30, and 35˚C, with approximately 9, 8, and 7.5-fold higher expression in the dark than under blue light ( Fig 1C); a response previously observed for BlsA [11,15]. Moreover, expression levels were higher at lower temperatures ( Fig 1C). Thus, our results suggest that expression of the putative anbluf65 photoreceptor correlates with regulation of motility by light.

AnBLUF65 is able to mediate light regulation of motility at 30˚C in A. baumannii
In order to study whether AnBLUF65 plays a role in mediating regulation of motility by light, we have cloned its coding sequence preceded by its own promoter into the Acinetobacter replicative plasmid pWH1266, and transformed A. baumannii ATCC 17978 ΔblsA strain with this construct. We chose this strain as it is a knock-out mutant on the only BLUF photoreceptor encoded in A. baumannii, and surface motility assays have showed that it displays comparable high-motility responses on swimming agar plates incubated at 23-24˚C [11,15,36], as well as at 30 and 37˚C [11], in the presence and absence of light. The rationale was to introduce the anbluf65 gene with its own promoter and observe whether regulation of motility by light occurred with respect to the same strain carrying the empty vector at 30˚C. Our results show that introduction of the anbluf65 gene effectively resulted in regulation of motility by light in A. baumannii, at temperatures higher than 25˚C, such as 30˚C (Fig 1D). Thus, the results confirm that AnBLUF65 is a photoreceptor and is capable of mediating light regulation of motility.

Quantitative differential proteomic profiling in response to light in A. nosocomialis RUH2624 at 37˚C
In previous studies, we demonstrated that light can modulate important physiological traits related to bacterial physiology and virulence in A. baumannii, A. nosocomialis, Staphylococcus aureus and Pseudomonas aeruginosa at temperatures found in warm-blooded hosts [8,9,14,15,17,18,37]. To broaden our knowledge of the response to light in A. nosocomialis at 37˚C and possibly identify other important traits modulated by light, a quantitative comparative proteomic analysis was conducted to obtain an overall representation of the total cellular changes that occur in this microorganism at the protein level upon blue light illumination at the mentioned temperature. We aimed to use proteomics as a tool to identify those proteins that could be present in different amounts under light and dark conditions and how these proteins could affect A. nosocomialis' physiology.
In total, 38 proteins were over-represented (fold change (FC) > 1.5 and p-value < 0.05, see Tables 4 and 5) when cells were grown under blue light or dark at 37˚C (Fig 8A). Twenty proteins were accumulated in higher amounts in cells cultured under blue light with respect to darkness (Fig 8A and 8B), while eighteen proteins were found to be in a greater amount under dark state (Fig 8A and 8C). When categorized using BLAST2GO suite, these proteins were dispersed across a wide variety of functions, according to gene ontology terms (Fig 9). In the "Biological Process" and "Molecular Function" sections the majority of the over-represented proteins fell into the metabolic/cellular process and catalytic activity categories, respectively, indicating changes in the metabolism of this bacterium in response to light. Considering the category "Cellular Component", the over-represented proteins under light conditions appear mainly distributed in the plasma and outer membranes, while in darkness most of them are cytoplasmatic (Fig 8).
Proteins with increased abundance in light compared to darkness. Light regulates the synthesis of proteins involved in a wide range of cellular functions. As mentioned before, these differentially over-represented proteins include some playing roles in metabolism, stress responses, and virulence in different bacterial pathogens.
Two proteins, PurD and Alc, related to purine metabolism were over-represented (FC PurD = 1.50, FC Alc = 62.0) in the presence of blue light. PurD is a phosphoribosylamine-glycine ligase and participates in the synthesis of inosine 5´-monophosphate (IMP) [38]. Purine metabolism plays an important role in microorganisms, recycling carbon and nitrogen compounds and regulating metabolism [39]. Several previous reports have described the importance of nucleotide biosynthesis by bacteria triggering infections. For instance, certain auxotroph mutants of Salmonella, S. aureus, or S. pneumoniae were avirulent in murine infection models [40,41]. Other studies demonstrate that limiting amounts of nucleotide bases in human serum, force pathogens to rely on de novo nucleotide biosynthesis [42].
Alc, which codes for an allantoicase, is involved in (S)-allantoin degradation and catalyzes the conversion of allantoate to (S)-ureidoglycolate and urea. Previous studies have found that nitrogen controls several pathways involved in secondary metabolism [43]. In general, secondary metabolites are produced by the cells as a response to environmental cues. For example, a notable decrease in antibiotic production was observed in S. coelicolor M145 as a response to the excess of intracellular ammonium generated during degradation of allantoin when the microorganism was grown in allantoin as the sole nitrogen source [44]. Regarding proteins involved in amino acids biosynthesis, LeuD and ProA are present in higher amounts in the presence of blue light (FC LeuD = 1.62, FC ProA = 1.50). LeuD catalyzes the isomerization between 2-isopropylmalate and 3-isopropylmalate, and is part of L-leucine biosynthesis [45].
ProA (gamma-glutamyl phosphate reductase protein) participates in step 2 of the subpathway that synthesizes L-glutamate-5-semialdehyde from L-glutamate and is involved in L- proline biosynthesis [46]. There is some evidence that proline metabolism has complex roles in a variety of biological processes, including cell signaling, stress protection, and energy production [47]. Proline can also contribute to the pathogenesis of various disease-causing organisms. Besides, some organisms use proline directly for the biosynthesis of secondary metabolites with antibacterial or antifungal properties [48].
An interesting protein that is over-represented in the light state is SurA, a chaperone involved in the correct folding and assembly of outer membrane proteins (OMPs), and may act both in early periplasmic as well as in late outer membrane-associated steps of protein maturation [49]. A great variety of outer membrane proteins are porins and autotransporters that facilitate transport and other essential functions, and act as virulence factors [50]. In P. aeruginosa, a significantly lower amount of many porins were detected in the outer membrane (OM) of the conditional surA mutant, including members of the OprD family (OpdO, OpdN, OpdP, and OprD) [51]. Also, it was found that siderophore receptors were absent or less abundant in the OM upon depletion of SurA [51]. Related to this point, in the present work a TonB-dependent receptor and a biopolymer transporter ExbD, which are part of a complex that energizes specific high-affinity receptors involved in regulating iron uptake, were found in a greater amount under blue light (FC TonB = 27.35, FC ExbD = 9.27). Consistently, a more robust growth under iron-deprived conditions, i.e., in the presence of the iron chelator 2,2 0 -dipyridyl (DIP), was observed in A. nosocomialis under blue light at 37˚C while it was severely affected in the dark [9]. Therefore, the higher abundance of SurA, TonB-dependent receptor, and ExbD under blue light could contribute to the enhanced iron acquisition observed in this condition, and this is an important feature since iron is a pathogenicity determinant essential for the success of bacterial infections [52].
Other proteins such as queuine tRNA-ribosyltransferase (Tgt) and ribosomal RNA large subunit methyltransferase K/L (RlmL) showed an increase in abundance under blue light. Tgt catalyzes the exchange of a guanine 34 with the queuine precursor 7-aminomethyl-7-deazaguanine (PreQ1) in specific tRNAs containing anticodones G(guanine)-U(uracil)-N (tRNA-Asp, -Asn, -His and -Tyr), where N is one of the four canonical nucleotides [53]. It was shown that a functional Tgt is required for efficient pathogenicity of Shigella bacteria [54]. A null-mutation in the tgt gene strongly reduces translation of virF-mRNA, a transcriptional activator required for the expression of a large number of Shigella pathogenicity genes [55]. On the other hand, RlmL K/L specifically methylates the guanine in position 2445 and the guanine in position 2069 (m7G2069) of 23S rRNA. The rRNA methyltransferases (rRNA MTases), especially those acting on 23S rRNA, are associated with development of antibiotic resistance in a wide variety of bacteria, and some of them are recurrent human pathogens [56]. Inactivation of 23S rRNA MTases function has been shown to negatively affect translation and cell physiology [57].
Finally, a glutathione S-transferase was over-represented in the light condition. The glutathione S-transferases (GSTs) are a family of proteins that conjugate glutathione to the sulfur atom of cysteine in various compounds [58]. GST can bind to a variety of hydrophobic compounds endogenous and xenobiotic alkylating agents with high affinity, including carcinogens, therapeutic drugs, environmental toxins and products of oxidative stress, protecting cells from oxidative damage [59].
Taken together, these observations suggest that light could play a main role in the control of A. nosocomialis physiology at 37˚C, allowing cells to respond and adapt to environmental signals.
Proteins with increased abundance in the dark compared to light. Among the proteins with higher abundance in the dark is the phenylacetate-CoA oxygenase, PaaK subunit. Phenylacetate-CoA oxygenase is comprised of a complex composed of 5 proteins responsible for the hydroxylation of phenylacetate-CoA (PA-CoA), which is the second catabolic step in phenylacetic acid (PAA) degradation [60]. Interestingly, we have shown that A. nosocomialis RUH2624 growth is stimulated in the dark when PAA is present as the sole carbon source at 37˚C [9], supporting the notion that the PAA catabolic pathway is modulated by light. Modulation of the PAA catabolic pathway has been shown to influence A. baumannii's pathogenesis. In fact, inhibition of this pathway resulted in increased neutrophil migration to the site of infection and bacterial clearance [60,61].
Another protein that appears over-represented in the dark is the nonhydrolyzing UDP-N-acetylglucosamine 2-epimerase (FC = 1.7). This enzyme catalyzes the reversible interconversion of UDP-N-acetylglucosamine (UDP-GlcNAc) to UDP-N-acetylmannosamine (UDP-ManNAc) [62], an intermediate in the biosynthesis of a variety of bacterial capsular polysaccharides (CPSs) [63]. In several pathogenic strains, such as B. anthracis, N. meningitides and S. aureus, CPSs are important virulence factors that protect bacteria from the immune system of a host and harsh environmental conditions [64][65][66]. In this sense, the UDP-Nacetylglucosamine 2-epimerase could contribute to cell-surface polysaccharide synthesis in A. nosocomialis, protecting the cells in the absence of light. The probable potassium transport Kup protein, found in a greater quantity in dark, could also favor the adaptation to rapidly changing external conditions.

Conclusion
In this work, we show that light regulates motility in A. nosocomialis in a wide range of temperatures, which go from environmental to temperatures found in warm-blooded hosts (23 to 37˚C). This temperature dependence is different from that observed for regulation of motility by light in A. baumannii ATCC 17978, which was shown to occur only in the low to environmental temperature range (18 to 24˚C). We hypothesized that this could be due to an unequal endowment of blsA homologs, whose intrinsic characteristics may allow them to function in other temperature ranges. In this work, we show that two of the BLUF domain-containing genes present in the A. nosocomialis RUH2624 genome [8], AnBLUF65 and AnBLUF46, encode stable proteins when produced in vitro and encode active photoreceptors in the temperature range analyzed, (15-37˚C) and (15-32˚C), respectively. The fact that AnBLUF65 is the only of the three BLUF domain-containing proteins that is expressed in cells recovered from motility plates along the photo-regulatory temperature range, and the fact that it can modulate motility at temperatures higher than 25˚C in A. baumannii, strongly suggests that it participates in modulation of this physiological trait by light in A. nosocomialis. Future work including mutagenesis studies are needed to determine the role each photoreceptor plays in this microorganism. Spectroscopic characterizations of AnBLUF65 and AnBLUF46 in vitro, indicate that AnBLUF65 is more efficient to form the light-adapted state than AnBLUF46 at 24˚C. Although AnBLUF65 efficiency is negatively affected by temperature increments, the protein remains active at 37˚C, consistent with being the only blue light photoreceptor expressed in A. nosocomialis. Analysis of homology models of this protein suggested that the relative solvent accessibility to the cofactor pocket derived from flavin emission fluorescence correlates well with the tertiary structure modeled for both proteins and both states, dark and light-adapted. Characterization of intramolecular interactions upon formation of the lightadapted state suggest the presence of aromatic clusters networks on either side of the β-sheet has been described. We propose that the rupture of an Asp-X interaction via H-bonding observed upon illumination in the loop between the Tyr-Tyr and the first set of Phe-Phe motifs, allows the aromatic tetrads in the variable domain to displace and to adopt the lightadapted state of this part in AnBLUF65 and AnBLUF46 (Fig 10). Thus, the calculated strengthening of the π-π tetrads in AnBLUF65 and the displacements of these residues between the dark and the light-adapted forms in BlsA as observed by Chitrakar et al 2020, reinforce the idea that intramolecular signaling events involve, at least in part, hydrophobic/aromatic interactions. The extent to which these networks are present in other members of the genus Acinetobacter, and other microorganisms remain to be further explored.
Overall, in this work, we have characterized different BLUF photoreceptors and show that they operate at different temperature ranges. It is worth mentioning that AnBLUF65 is the only BLUF-protein found to be expressed in A. nosocomialis in motility plates, and the possibility araises that the other BLUF-photoreceptors are expressed under different conditions, such as growth media, temperatures, presence of a host, etc. The presence of AnBLUF85 only in two different A. nosocomialis strains and discretionally in other species of the Acinetobacter genus is striking, in addition to the fact that this photoreceptor is not expressed in its host.