Signature Motifs Identify an Acinetobacter Cif Virulence Factor with Epoxide Hydrolase Activity*

Background: Pathogens target airway clearance mechanisms to facilitate infection. Results: Sequence analysis reveals an Acinetobacter epoxide hydrolase (EH) that triggers loss of the cystic fibrosis transmembrane conductance regulator (CFTR). Conclusion: Homologous EH virulence factors found in a variety of opportunistic pathogens can impair CFTR, a key element of host airway defenses. Significance: EH virulence factors are potential therapeutic targets. Endocytic recycling of the cystic fibrosis transmembrane conductance regulator (CFTR) is blocked by the CFTR inhibitory factor (Cif). Originally discovered in Pseudomonas aeruginosa, Cif is a secreted epoxide hydrolase that is transcriptionally regulated by CifR, an epoxide-sensitive repressor. In this report, we investigate a homologous protein found in strains of the emerging nosocomial pathogens Acinetobacter nosocomialis and Acinetobacter baumannii (“aCif”). Like Cif, aCif is an epoxide hydrolase that carries an N-terminal secretion signal and can be purified from culture supernatants. When applied directly to polarized airway epithelial cells, mature aCif triggers a reduction in CFTR abundance at the apical membrane. Biochemical and crystallographic studies reveal a dimeric assembly with a stereochemically conserved active site, confirming our motif-based identification of candidate Cif-like pathogenic EH sequences. Furthermore, cif expression is transcriptionally repressed by a CifR homolog (“aCifR”) and is induced in the presence of epoxides. Overall, this Acinetobacter protein recapitulates the essential attributes of the Pseudomonas Cif system and thus may facilitate airway colonization in nosocomial lung infections.

Opportunistic lung infections by Gram-negative pathogens are a major source of mortality and morbidity. In patients with cystic fibrosis (CF), 2 airway mucociliary transport is impaired by loss of function in the cystic fibrosis transmembrane conductance regulator (CFTR), an ion channel essential for proper mucus hydration (1). The resulting breakdown of airway clearance permits colonization by bacteria such as Pseudomonas aeruginosa and Burkholderia cepacia complex, leading to cycles of infection and inflammation that impair respiratory function and lead ultimately to death (2). Smoking also suppresses airway epithelial CFTR (3)(4)(5)(6) and contributes to the etiology of chronic obstructive pulmonary disease, the third leading cause of mortality in the United States (7). Although the role of P. aeruginosa in chronic obstructive pulmonary disease is complex, it has been associated with exacerbations, reduced lung function, and increased mortality (8 -11). Finally, Acinetobacter species cause drug-resistant nosocomial infections in the lung and elsewhere (12,13), and include one of the clinically significant "ESKAPE" pathogens (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, P. aeruginosa, and Enterobacter species) (14,15).
In addition to exploiting weakened host immunity, these pathogens employ multiple approaches to facilitate the establishment and maintenance of airway infections. Key defensive strategies include the formation of biofilms and the development of high levels of antibiotic tolerance (16 -19). They also deploy a variety of highly sophisticated virulence factors with more specialized targets. Among these is the CFTR inhibitory factor (Cif), a recently characterized P. aeruginosa protein expressed both in the PA14 strain and in clinical isolates (20,21). Originally characterized as an extracellular factor that inhibited CFTR-mediated chloride currents across polarized epithelial monolayers (22), Cif is both directly secreted and is also packaged into outer membrane vesicles that dramatically enhance delivery to airway epithelial cells (23). Further investigation revealed that Cif acts by triggering the post-endocytic degradation of CFTR, reducing its cell-surface abundance (22). As a result, Cif has the potential to interfere with mucociliary clearance mechanisms, in concert with previously identified Pseudomonas ciliostatic virulence factors, such as pyocyanin, rhamnolipids, and phenazines (24 -26).
Sequence and structural homologies indicate that Cif adopts an ␣/␤ hydrolase-fold, a scaffold commonly associated with hydrolytic enzymes (20,27). Based on alignments to other ␣/␤ hydrolase family members, it was originally proposed to act as an epoxide hydrolase (EH). However, its sequence was found to lack several active site features conserved among canonical EH family members. These include one of the tyrosine side chains that coordinate the epoxide ring for a nucleophilic attack, which yields a covalent enzyme-substrate intermediate. They also include a cis-proline-mediated backbone conformation that helps position an attacking water molecule, and a chargerelay acidic residue C-terminal to the cap domain that helps to activate the water for a second nucleophilic attack, releasing the diol product and regenerating the active site (28). Nevertheless, detailed enzymological studies confirmed that Cif can indeed hydrolyze epoxides (27). Subsequent analysis revealed a series of structural adaptations that enable Cif to retain EH activity despite sequence variations in these active site residues (27,29). Sequence comparisons uncovered similar motifs in the genomes of other opportunistic pathogens, and sequence-structure mapping suggested that these putative EHs are likely to adopt a conformation similar to Cif (29).
In particular, homologs are found in the genomes of several members of the A. calcoaceticus/baumannii (ACB) complex, including Acinetobacter nosocomialis type strains LMG 10619 T and RUH 2624 and A. baumannii strain WC-487. Ventilatorassociated pneumonia is the most frequent nosocomial infection associated with A. baumannii (13). Here, we experimentally characterize the biochemical and cellular activities of Acinetobacter Cif ("aCif"), and demonstrate its functional equivalence to the P. aeruginosa system. These data confirm the ability of the Cif-specific motifs to identify EH proteins that inhibit CFTR abundance, and thus to flag a new potential class of virulence factors in opportunistic lung infections.

EXPERIMENTAL PROCEDURES
Plasmid Construction-Based on similarity to the cif gene of P. aeruginosa, the coding region of NCBI accession number WP_004885628 (formerly ZP_05823503) from A. nosocomialis strain RUH2624 (30) was amplified with primers RBS/13TU cif F and 6-His/13TU cif R (Table 1) using Phusion polymerase (New England Biolabs). These primers added a canonical Shine-Dalgarno sequence and a carboxyl-terminal His 6 tag to the gene for recombinant protein expression and purification from Escherichia coli. The resulting PCR product was purified using a Qiaquick PCR Purification Kit (Qiagen) and served as the template for a second round of PCR using the 70/RBS 5Ј and 70/His 6 3Ј primers to add pMQ70 plasmid overlap regions to the ends. This final PCR product was cloned into plasmid pMQ70 via Saccharomyces cerevisiae recombineering as previously described (31).
Positive clones were verified by PCR and Sanger sequencing. Mutagenesis to generate the aCif-D158S mutant construct was performed using the QuikChange Lightning Site-directed Mutagenesis Kit (Stratagene), which was again verified by Sanger sequencing.
Deletion of the Acinetobacter cifR Homolog-An in-frame deletion of the cifR gene from the chromosome of A. nosocomialis strain RUH2624 was created using construct pMQ30-cifR. First, the pMQ30-cifR plasmid was created using S. cerevisiae recombineering (31), empty pMQ30 plasmid, and the primers listed in Table 1. The pMQ30-cifR construct was electroporated into A. nosocomialis using the procedure described by Choi et al. (32) for P. aeruginosa. Exconjugants containing an inserted plasmid were selected on LB agar supplemented with 50 g/ml of gentamicin, followed by counterselection with 5% (w/v) sucrose. In-frame deletions were confirmed by colony PCR.
Protein Purification-Recombinant carboxyl-terminal His 6tagged aCif protein was expressed and secreted from E. coli, and purified by immobilized metal affinity chromatography using procedures previously described (33) for Cif.
Biochemical Characterization-Size-exclusion chromatography was carried out with a Superdex 200 HR 10/30 column (GE Healthcare) calibrated with aldolase, ovalbumin, chymotrypsinogen A, and RNase A. The size-exclusion chromatography buffer consisted of 100 mM NaCl and 20 mM HEPES (pH 7.4). Circular dichroism (CD) spectroscopy was performed with a 1-mm path length quartz cuvette containing 300 l of 10 M aCif protein in 100 mM NaCl, 20 mM HEPES (pH 7.4). Edman degradation was performed by the W. M. Keck Biotechnology Resource Laboratory (Yale University) using purified aCif protein.
EH Activity-Epoxide hydrolysis activity was first assessed for a panel of candidate epoxides using an adrenochrome reporter end-point assay as described previously (27). Briefly, samples were incubated for 1 h at 37°C in a 100 l volume 5Ј-TGCAGGTCGACTCTAGAGGATCCCCGGGTACCGAGCTTTAGTGGTGATGATGGTGATGTT-3Ј Deletion of Acinetobacter cifR containing 20 M aCif in 100 mM NaCl with 20 mM MOPS (pH 7.4 at RT) and 1% (v/v) dimethyl sulfoxide. Next, 50 l of 2 mM NaIO 4 in 90% (v/v) acetonitrile were added, and samples were incubated at room temperature for 1 h. A colorimetric signal was generated by addition of excess adrenaline-HCl and quantified by absorbance at 490 nm (A 490 ).
To determine specific activities, epoxides were incubated with varying aCif concentrations for 10 min at 2% (v/v) dimethyl sulfoxide. Samples were then incubated with 4 mM NaIO 4 for 30 min. To test the effect of phosphate, a parallel experiment was performed at the same assay pH with 20 mM sodium phosphate buffer in place of MOPS.
To quantitate product formation, standard curves were generated using 5 M aCif and 0 -2 mM of the cognate diols. After verifying linearity of the A 490 read-out in this range, various concentrations of aCif were incubated with 2 mM epoxide substrate under the conditions listed above, and specific activity was calculated in a range where the signal varied linearly with enzyme concentration. The amount of product formed was determined by linear regression analysis (Excel) of the standard curves.
Protein Crystallization and Data Collection-Purified aCif protein at 5 mg/ml in 100 mM NaCl, 20 mM HEPES (pH 7.4) was submitted to the Hauptman-Woodward Medical Research Institute High-Throughput Screening Facility for microbatch crystallization (34). Next, promising conditions were adapted to the hanging drop vapor diffusion method to obtain diffraction-quality protein crystals (33). aCif protein was mixed in a 1:1 ratio with well solution to achieve a total volume of 4 l. The well solution consisted of 100 mM KH 2 PO 4 , 100 mM sodium citrate (pH 4.0), 20% (w/v) polyethylene glycol (PEG) 4000. Prior to data collection, crystals were harvested into a cryoprotectant solution consisting of 100 mM KH 2 PO 4 , 100 mM sodium citrate (pH 4.0), 20% (w/v) PEG 4000, 20% (w/v) glycerol, and flash cooled by plunging into a liquid nitrogen bath. Oscillation data were collected at 100 K at the X6A beamline of the National Synchrotron Light Source at Brookhaven National Laboratory. Diffraction images were processed and scaled using the XDS package (35). Phases were determined by molecular replacement. The search model was generated by threading the aCif protein sequence into the experimentally determined structure of Cif (PDB code 3KD2) using SWISS-MODEL (36,37). Molecular replacement searches, model building, and iterative rounds of refinement were performed using Phenix (38,39). Manual adjustment of the model was performed using WinCoot (40), and structural images were generated using PyMOL (41).
Protein crystals for aCif-D158S were obtained by the same method used for the wild-type (WT) protein, using a well solution consisting of 250 mM KH 2 PO 4 , 100 mM sodium citrate (pH 4.0), 20% (w/v) PEG 4000, and a cryoprotectant solution consisting of 250 mM KH 2 PO 4 , 100 mM sodium citrate (pH 4.0), 20% (w/v) PEG 4000, and 20% (w/v) glycerol. The aCif-WT model (PDB entry 4MEA) was used as a molecular replacement search model to obtain phase information, and all other data collection and processing steps were performed as for aCif-WT.
Cell Culture and CFTR Biotinylation Assay-The determination of apical membrane CFTR abundance was performed using CFBE4lo-cells stably transduced with WT-CFTR (CFBE-WT) (42). The effects of Cif virulence factors on CFTR levels were monitored after applying purified Cif or aCif protein to the apical surface of polarized cells, followed by domain-selective cell-surface biotinylation using EZ-Link TM Sulfo-NHS-LC-Biotin (Pierce), pull-down using streptavidin-conjugated agarose, and quantitative Western blot for CFTR, as described previously (20,43,44).
RNA Isolation, cDNA Synthesis, and qRT-PCR-A. nosocomialias strain RUH2624 was cultured overnight at 37°C in LB and subcultured 1:100 into 5 ml of LB with 1 mM epibromohydrin, (R)-styrene oxide, or (S)-styrene oxide as indicated. Cultures were grown to an A 600 of 1.0, upon which 1 ml was removed and centrifuged briefly to pellet cells. The pellets were flash-frozen in a dry ice-ethanol bath and stored at Ϫ80°C. RNA was isolated as described previously (21). cDNA was synthesized from 1 g of purified RNA using the QuantiTect Reverse Transcription Kit (Qiagen). Quantitative real-time (qRT) PCR primers were designed to the A. nosocomialis genes cif (13TU cif F and 13TU cif R) and cifR (13TU CifR F and 13TU CifR R), as well as to the 50 S ribosomal protein L21 (13TU rplU F and 13TU rplU R) as a transcriptional control (GenBank TM accession numbers: EEX01162, EEX01160, and EEX01619, respectively; Table 1). Primers were tested for specificity by PCR of the genomic DNA. qRT-PCR was performed with SYBR Green PCR master mix (Applied Biosystems) using a CFX96 Real-Time system on a C1000 thermal cycler (Bio-Rad) for 42 cycles, followed by a melt curve to verify proper amplification. Real-time data were analyzed using CFX manager software (Bio-Rad) and expressed as picograms of input RNA of the gene of interest relative to the ribosomal transcriptional control.

RESULTS AND DISCUSSION
aCif Exhibits Cif-like Epoxide Hydrolase Activity-The putative aCif proteins from A. nosocomialis (GenBank locus EEX01162) and A. baumannii (locus EKU60908) exhibit 100% sequence identity to each other. To begin our investigation of the aCif protein, we first performed a full sequence alignment with the Cif protein from P. aeruginosa (Fig. 1). aCif contains a 24-residue N-terminal sequence predicted to act as a secretion signal by SignalP (45), corresponding to the validated signal sequence observed in Cif (33). In aCif, this sequence is followed by a 27-residue linker, and then by a 298-residue region with 36% sequence identity to the mature Cif sequence (Fig. 1).
As a basis for biochemical characterization, we cloned the cif gene from A. nosocomialis into the pBAD arabinose-inducible expression vector pMQ70 with a carboxyl-terminal His 6 tag. Recombinant aCif could be readily purified from culture supernatants of E. coli transformed with the expression vector. Furthermore, Edman sequencing of the amino terminus of purified aCif yielded a sequence of 25 EYDPNLKSIDT 35 (numbered relative to the predicted translational start site), consistent with cleavage of the predicted signal sequence during secretion. The mature aCif protein thus carries a 27-residue extension at the N terminus of the Cif homology domain.
Cif has previously been shown to form a dimer in solution (27). To test the oligomeric state of aCif, we calibrated the elution volume of aCif during analytical size-exclusion chromatography relative to a series of mass standards. The elution volume of 14.95 ml corresponds to a molecular mass of ϳ60 kDa ( Fig. 2A), substantially larger than the predicted monomeric molecular mass of 37 kDa, and directly comparable with the estimate obtained previously for Cif using analytical size-exclu-sion chromatography and sedimentation analysis. The protein is also thermally stable, with a melting temperature of 51.2°C (Fig. 2B), consistent with the requirement that it function in the potentially challenging environment outside the cell.
To test the functional similarity of the two proteins, we sought to determine whether aCif possesses EH activity. Using an adrenochrome reporter assay, we measured the ability of aCif to catalyze vicinal diol formation from a series of candidate epoxide substrates (Fig. 3A). The strongest signals were observed with three compounds: epoxycyclohexane and the (S)-stereoisomers of styrene oxide and 4-nitrostryene oxide.
We next determined specific activities for aCif as a catalyst of hydrolysis for these compounds and for (R)-styrene oxide, which showed a lower signal in the adrenochrome assay (Fig.  3A). Compared with controls, statistically significant specific activities were observed for the three expected substrates, but not for (R)-styrene oxide (Fig. 3B). Taken together, these data clearly demonstrate that aCif can catalyze the hydrolysis of multiple epoxide compounds. We therefore conclude that, like the Cif protein from P. aeruginosa, aCif is a bona fide EH.
We also calculated a background threshold for the adrenochrome signal based on the confirmed lack of aCif activity for (R)-styrene oxide (gray shading, Fig. 3A). Aside from the three substrates tested in Fig. 3B, no other epoxide showed a level of hydrolysis significantly above this background, suggesting that none of them is a substrate. Thus, the aCif active site exhibits a high level of stereoselectivity.
Interestingly, we were unable to detect significant aCif catalysis of epibromohydrin, a substrate hydrolyzed by P. aeruginosa Cif and originally used to characterize its EH activity (27,46). Direct side-by-side epibromohydrin hydrolysis assays confirmed the discrepancy (data not shown). The residues surrounding the catalytic machinery of aCif are remarkably conserved with Cif, in both identity and placement, consistent with a shared enzyme activity. However, the residues that line the entrance to the active site display more significant divergence, including backbone location as well as residue identity. Furthermore, position of Cif residue His-269, which was previously  . aeruginosa and A. nosocomialis. An initial alignment was performed using ClustalW, which was then manually edited to account for biochemical data such as signal sequence cleavage. postulated to play a role in substrate selectivity determination (33), is occupied by Glu-25, the N-terminal residue of mature aCif, suggesting that the N-terminal extension may participate in substrate selection. Thus, differential active-site accessibility may provide one explanation for observed discrepancies in the substrate profiles for these two homologous Cif EHs. aCif Can Reduce CFTR Apical Membrane Levels-The physiological activity that led to the identification of Cif was a decrease in cell-surface CFTR abundance in epithelial cells cocultured with P. aeruginosa (22). This CFTR effect does not require the presence of live bacteria. It can be mediated directly by purified protein, which permitted the initial isolation of Cif activity and the proteomic identification of Cif as the causative agent (20,27). Given the different specificity profiles of Cif and aCif, we wished to determine whether aCif could also trigger a reduction in CFTR at the apical membrane. We applied aCif or Cif to the surface of polarized human bronchial epithelial cells that were grown in culture at an air-liquid interface, and monitored cell-surface CFTR abundance by biotinylation, streptavadin capture, and immunoblotting. Both aCif and Cif exhibit a robust ability to reduce the levels of CFTR at the cell surface (Fig. 3C), with similar efficacies. These data clearly demonstrate that aCif recapitulates a key cellular effect of Cif, inhibiting the net accumulation of CFTR at the apical membrane.
Structural Analysis of aCif-aCif was identified as a putative Cif-like EH because it shares with Cif a set of active-site sequence motifs that differ from those conserved among canonical EH family members (29). To test our hypothesis directly that these motifs adopt Cif-like structures in the context of the aCif sequence, we crystallized aCif and determined its structure by x-ray diffraction analysis. Although aCif protein crystals were small and needle-like (ϳ10 -20 m thick and 100 -200 m in length), we obtained strong diffraction to 1.95-Å resolution ( Table 2). Phase information was determined by molecular replacement. Phasing quality was confirmed by composite omit maps and by clear electron density for the aCif N-terminal extension, which was not included in the search model. The refined model demonstrates excellent agreement with the experimental data, yielding a final R work /R free of 0.146/0.186. Consistent with our hydrodynamic analysis, aCif crystallized as a dimer (Fig. 4A), with one dimer per asymmetric unit in the crystal lattice. aCif exhibits the classic ␣/␤ hydrolase-fold: a ␤ sheet consisting of seven parallel and one antiparallel ␤ strand sandwiched by ␣ helices (Fig. 4A). In direct comparison, aCif and Cif also possess a high degree of structural similarity (Fig.  4B). A distance-alignment search using DALI (47) showed that the known structures closest to aCif are Cif (Z-scores Ն42), where I i (h) and I(h) values are the i-th and mean measurements of the intensity of reflection h. c R mrgd-F is a robust indicator of data quality, as described by Diederichs and Karplus (53).
To determine whether the aCif structure is more similar to canonical EH enzymes or to the Cif-like EH family, we focused our analysis on four distinguishing features of the Cif structure. The first and perhaps largest structural divergence demarcating Cif from canonical EHs is the position of the ␣5 helix. Most EH enzymes possess an active site buried within the interior of the protein with two outlets to the solvent. The shifted position of this helix in the Cif structure caps one opening of the active-site tunnel, converting the active-site tunnel into a pocket (29). This helix is in a nearly identical position in the aCif and Cif structures (blue and yellow helices in Fig. 4B). As a result, aCif also possesses only one outlet from the active site to the solvent.
The second feature involves the identity of the "ring-opening" pair of residues. These residues are responsible for binding and coordinating the epoxide oxygen within the EH active site, as well as donating a proton to assist in ring opening. Classically, EHs enzymes have utilized a pair of Tyr residues for this task (49). In contrast, Cif uses a His/Tyr pair, a combination also used to position substrates in the fluoroacetate dehalogenases (27,29). aCif shares this altered His/Tyr motif, and the active site clearly displays His and Tyr side chains aligned in a nearly identical position and orientation as the Cif ring-opening pair (Fig. 4C).
The third feature distinctive to the Cif EH is the position of the charge-relay acid of the catalytic triad. This residue projects into the active site near the junction between the ␣/␤ hydrolase core and cap domains of Cif, whereas it is commonly located within the ␣/␤ hydrolase core domain of canonical EHs (29). aCif also possess an acidic residue forming similar hydrogen bonds and at the position equivalent to the charge-relay acid of Cif (Fig. 4C).
The fourth and final feature we examined was the structural HGFG motif. This stretch of 4 amino acids enables the protein backbone to form a tight turn, which positions the backbone to participate in two hydrogen bonds thought to be essential for the enzymatic mechanism. Canonical EHs utilize a conserved HGXP motif, where "X" is any amino acid (usually aromatic), and "P" is a cis-proline (50). Both aCif and Cif utilize a second Gly residue in place of the canonical cis-Pro to cap the motif and allow the backbone to make the tight turn (*, purple trace in Fig.  4B).
Thus, despite sharing only 36% sequence identity, the noncanonical signature motifs first identified in the Cif protein fulfill stereochemically equivalent roles in the aCif protein. In particular, this observation validates the sequence-structure alignment approach used to assess structural similarity to Cif (29). It also suggests that the other candidate pathogenic EHs, which exhibit similar levels of sequence divergence, are nonetheless likely to be members of the Cif-like EH subfamily.
Differences between aCif and Cif-In addition to these fundamental similarities, there are several aspects of the aCif structure that distinguish it from Cif. The most notable variation is a 27-residue extension at the N terminus of mature aCif (red residues in Fig. 1). These amino acids form an additional loop and helix that protrude from the ␣/␤ hydrolase core and wrap around the outside of the cap domain, contributing to the dimer interface (Fig. 4A). The extension helix resides on the opposite side of the protein from the ␣5 tunnel-blocking helix, but does not appear to obstruct the remaining entrance from the solvent to the active site (Fig. 4B). Its function remains unclear.
The second marked difference observed in the aCif structure was the presence of extra electron density bound within the active site (Fig. 4C). We had previously observed a water molecule coordinated by the ring-opening pair at the corresponding position in the original Cif structure (27). Unlike Cif, aCif was crystallized in the presence of 100 mM phosphate. Because the electron density is consistent with a tetrahedral geometry (Fig.  4C), it appears most likely that it represents a phosphate group occupying the substrate-binding site. To assess the ability of phosphate to act as a competitive inhibitor of aCif EH activity, we compared the specific activities of aCif for (S)-styrene oxide in buffers containing either 20 mM phosphate or 20 mM MOPS. The difference in specific activity was not significant (Fig. 3B). Therefore, we conclude that the candidate phosphate molecule observed in the structure is most likely an artifact of the crystallization conditions, rather than a physiologically relevant competitive inhibitor. Mutation of the aCif Catalytic Nucleophile-Next, we sought to probe the catalytic activity of aCif. Based on sequence and structural comparison to Cif and other EHs, we predicted that Asp-158 serves as the catalytic nucleophile of aCif and that mutation of this residue to a Ser would ablate the EH activity. We generated this point mutant and purified aCif-D158S protein by the same method used for the WT protein. We assayed for hydrolysis of (S)-styrene oxide and observed the expected complete loss of enzyme activity (Fig. 5A).
To assess whether this loss of activity is due to localized changes in the active site or more global effects on aCif structure, we also determined the crystal structure of aCif-D158S. Refinement of the aCif crystallization conditions yielded diffraction-quality protein crystals similar in shape and size to those obtained with the WT protein. Phase information was again obtained by molecular replacement, using the aCif-WT structure as the search model. The aCif-WT and D158S models align with a root mean square deviation of 0.50 Å over 4723 atoms (Fig. 5B), confirming that the mutation does not disrupt the overall fold of the protein. The side chain of the mutant Ser at position 158 is pointed away from the active site and forms hydrogen bonds with the same backbone residues used to coordinate the Asp side chain in the WT structure (Fig. 5C). Additionally, we again found a phosphate molecule coordinated in the putative substrate binding position. We therefore conclude that Asp-158 plays an essential role in the active site, suggesting a hydrolytic mechanism shared with other EH enzymes.
Genetic Regulation of aCif-Comparison of the Acinetobacter and Pseudomonas cif operons reveals a conserved organization (Fig. 6A) that extends the structural and functional simi-   6. A, the cif operons of Pseudomonas and Acinetobacter are compared, with the genes and intergenic regions displayed to relative scale. The Acinetobacter operon does not possess a putative major facilitator superfamily (MFS) transporter gene that is found in the Pseudomonas operon. Sequence alignments were performed with ClustalW, and the percent identity was then calculated. B, qRT-PCR was performed on cultures of A. nosocomialis incubated with epoxide, or with a cifR gene deletion. Expression was not measured for cifR in the cifR deletion strain. Expression levels were compared with the values for the corresponding gene in the WT strain with buffer, using Student's unpaired t test (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; n ϭ 3). larities between the Cif and aCif proteins. In P. aeruginosa, the cifR gene encodes a transcriptional repressor that is transcribed divergently from the cif operon (Fig. 6A). CifR binds to the intergenic sequence between the cifR and morB genes and blocks transcription in both directions; repression is relieved in the presence of various epoxides (21,46). Based on its 50% sequence identity to cifR, we hypothesized that the WP_004707394 locus encodes a functionally equivalent Acinetobacter CifR ("aCifR") protein that can regulate cif expression in A. nosocomialis by the same mechanisms observed previously in P. aeruginosa.
To test this proposal, we first investigated the ability of aCifR to act as a transcriptional repressor by monitoring cif expression by qRT-PCR in WT A. nosocomialis strain RUH2624 and in a mutant strain carrying a clean deletion of the cifR gene. Consistent with our hypothesis, relative cif expression increased by nearly 2 orders of magnitude in the mutant lacking aCifR compared with the expression seen in WT (Fig. 6B, two leftmost columns).
Next, we tested the ability of epoxides to increase transcription of both cif and cifR genes, which are coordinately repressed by CifR in P. aeruginosa (46). Gene expression was monitored by qRT-PCR in WT A. nosocomialis cultured in the absence or presence of three epoxide compounds. One of these was epibromohydrin, a known activator of cif transcription in P. aeruginosa (46). However, because our substrate profiling indicated that aCif has a strong preference for (S)-styrene oxide (Fig. 3B), we also chose to examine both stereoisomers of styrene oxide, because the racemic mixture has previously been shown to induce Cif expression in P. aeruginosa (21). In each case, we found the epoxide compounds were able to induce both cif and cifR expression in A. nosocomialis by approximately 1 order of magnitude (Fig. 6B).
The two stereoisomers of styrene oxide yielded similar levels of induction, and both were modestly more effective than epibromohydrin. This suggests that all three compounds are able to interact with an effector-binding site on aCifR. The (R)-and (S)-stereoisomers of styrene oxide are likely to have similar affinities, because they should have a similar ability to cross the bacterial membrane. However, a direct comparison of potency with epibromohydrin is difficult, because it is likely to be more hydrophilic and thus may have different intracellular bioavailability. In any case, none of the candidate inducers were able to fully de-repress cif expression to the level observed in the ⌬cifR background, suggesting that physiological compounds could achieve even higher levels of activation.
Taken together, these results suggest that the Acinetobacter cifR gene encodes an epoxide-sensitive repressor capable of blocking transcription of both cifR and cif. Upon binding to an epoxide effector, the aCifR protein would then release from its DNA binding site and allow transcription to proceed at both operons. Thus, both the genetic control and biochemical functionality of the Acinetobacter cif gene are closely comparable with those of its Pseudomonas counterpart.
Concluding Remarks-Analysis of National Nosocomial Infections Surveillance (NNIS) data from 1986 to 2003 showed that A. baumannii was associated with 5-10% of ICU pneumonia cases. It was also the only pathogen in the survey that showed a significant increase in frequency of association over time (51). Due to high levels of antibiotic resistance, recent efforts have focused on the development of drugs that target Acinetobacter virulence factors (52). In the present study, we sought to determine whether Cif-like sequence motifs can correctly identify homologous epoxide hydrolases in opportunistic pathogens other than P. aeruginosa. Indeed, our evidence overwhelmingly suggests that A. nosocomialis and A. baumannii strains possess systems similar to Cif, consisting of a secreted EH enzyme that is genetically regulated by an epoxide-sensitive TetR family repressor. These systems likely enable the bacteria to reduce the abundance of CFTR in polarized human airway epithelial cells, and thus to inhibit mucociliary clearance during the establishment of airway infections. As a result, the aCif protein may represent a new therapeutic target, either alone or as part of a broader strategy to combat this emerging threat to public health.