Structure and Function of the PiuA and PirA Siderophore-Drug Receptors from Pseudomonas aeruginosa and Acinetobacter baumannii

Identification of PiuA and PirA orthologues in A. baumannii.

We (31) and others (15) previously identified the TBDRs PiuA and PirA as important to the activity of the siderophore-drug conjugates BAL30072 and MC-1 in P. aeruginosa. Since BAL30072 also shows potent activity toward A. baumannii (14, 18, 32), we attempted to identify its mechanism of action in the organism. We performed a homology search for potential orthologues of PiuA and PirA on the genome of A. baumannii reference strain ATCC 19606 (Table 1). The BLAST algorithm identified two uncharacterized open reading frames (ORFs), A1S_0474 and A1S_0980 (ATCC 17978 numbering), as the closest hits, sharing amino acid identities of 44% with PiuA and 51% with PirA of P. aeruginosa, respectively. Both genes have been shown to be preceded by putative Fur binding sites (A1S_0474, GCGAATAATAATAATTCTTATTTAT, and A1S_0980, GATATTGTTAATAATTATCATTATT) (33), suggesting regulation by iron availability via the Fur iron uptake regulator. Downstream of the piuA gene of A. baumannii, we identified an ORF (A1S_0473) showing 52% amino acid identity with piuC from P. aeruginosa and ybiX from Escherichia coli (Fig. 1). The corresponding proteins belong to the family of Fe(II)-dependent oxidoreductases (34), and their colocalization with piuA might indicate that they act on the natural substrate transported by PiuA/FiuA. In contrast, the genetic environment of the pirA region of A. baumannii is different from those in P. aeruginosa and E. coli (Fig. 1). The pirA gene in A. baumannii is preceded by nfuA, coding for an Fe-S-containing protein involved in intracellular iron utilization and oxidative-stress response (35). In P. aeruginosa, the gene for the secondary enterobactin transporter PirA is located next to a two-component system comprising the transcriptional regulator PirR (PA0929) and the sensor kinase PirS (PA0930) (36). The closest orthologue of PirA in E. coli is the ferric enterobactin transporter FepA, whose gene is embedded within the enterobactin biosynthesis region (not shown in Fig. 1). Based on this evidence, we adopted the nomenclature from P. aeruginosa and termed these proteins PiuA (A1S_0474) and PirA (A1S_0980), designated below Ab-PiuA and Ab-PirA.

PiuA and PirA contribute similarly to siderophore–beta-lactam activity in A. baumannii.

To test the involvement of PiuA and PirA in the activity of BAL30072 and MC-1 against A. baumannii, we constructed piuA (A1S_0474) and pirA (A1S_0980) gene deletion mutants in A. baumannii strain ATCC 19606. Both the piuA and the pirA mutants showed 2- to 8-fold increases in the MICs for both drug conjugates, in agreement with decreased transport in the individual mutants. The piuA-pirA double mutant showed a further increase in MICs, which was more pronounced for MC-1 than for BAL30072 (Table 2). This suggests that in the wild-type strain and under the conditions of the MIC assay, both receptors have a basal expression level and may contribute similarly to the uptake of BAL30072 and MC-1. This is in contrast to the situation in P. aeruginosa, since deletion of pirA in PAO1 has no effect on susceptibility to BAL30072 and MC-1, and the contribution of this TBDR is seen only in a piuA deletion background (15).

PiuA and PirA from A. baumannii are functional in P. aeruginosa.

To test whether the A. baumannii proteins are functional in P. aeruginosa, plasmids ppirA-Ab4.1 and ppiuA-Ab6.3, expressing, respectively, the pirA and piuA genes of A. baumannii from a constitutive or an inducible promoter, were transferred into P. aeruginosa strain PAO1 and into a piuA deletion mutant. While expression of the vector plasmids alone had no significant effect on the BAL30072 or MC-1 MIC, the expression of both piuA and pirA from A. baumannii conferred increased susceptibility to these conjugates in P. aeruginosa. As expected, susceptibility to aztreonam (ATM) was not affected by overexpression of the TBDR genes (Table 3). Similar results were obtained in a piuA deletion strain showing a 32- to 64-fold increase in susceptibility to BAL30072 and MC-1 when the pirA gene of either P. aeruginosa or A. baumannii was overexpressed. However, Ab-PiuA expression from plasmid ppiuA-Ab6.3 did not seem to confer increased susceptibility on the piuA deletion strain (Table 3). A possible explanation is that the induced expression level of Ab-PiuA in the piuA deletion strain might be insufficient to restore wild-type susceptibility.

Structure comparison of PiuA and PirA from P. aeruginosa and A. baumannii.

The sequence alignment of the PiuA and PirA orthologues revealed high conservation within the putative substrate binding domains NL1 (PiuA and PirA) and NL2 and NL3 (PirA) (Fig. 2). The structures of PiuA and PirA from P. aeruginosa and A. baumannii have been solved. As expected, these structures were similar to those of other TBDRs, sharing a two-domain organization consisting of a 22-stranded transmembrane β-barrel (Fig. 3, green ribbons) and an N-terminal plug domain (blue ribbons) folded inside the barrel.

As often observed with TBDRs, several extracellular loops were missing in the Pa-PiuA structure. Loops NL1 (83 to 90), L7 (505 to 526), and L8 (560 to 567) could not be modeled due to the high flexibility of these regions. In contrast, the equivalent extracellular loops were present in the structure of Ab-PiuA, perhaps due to the presence of a phosphate ion that coordinated with Glu73 of NL1, Asp303 of L4, Arg316 of β8, and His697 of L11. Structural comparison of Pa-PiuA and Ab-PiuA (root mean square deviations [RMSD] of 1.3 Å for 616 aligned C-α atoms) showed differences in the lengths and positions of several extracellular loops (Fig. 4; see Fig. S1 in the supplemental material). Ab-PiuA contained additional amino acids in loops L7 and L9, but fewer in L6 and L8 than Pa-PiuA. Strikingly, loop L7 of Ab-PiuA folded inside the barrel close to the top of the plug domain (Fig. 4; see Fig. S1 in the supplemental material). A mutant protein with five amino acids (EGGNP; positions 73 to 76) deleted from loop NL1 was generated in PiuA from P. aeruginosa. When expressed in strain PAO1, the NL1 deletion mutant was unable to confer hypersusceptibility to BAL30072, as observed with the PiuA wild-type protein (Table 3). We conclude that loop NL-1 is likely involved in the initial binding of BAL30072 in PiuA.

Comparison with other TBDRs showed that the closest structural relatives of PiuA are the pyochelin (FptA) and pyoverdine (FpvA) receptors of P. aeruginosa (37, 38), with an RMSD of 1.84 Å (567 aligned C-α atoms) and 1.99 Å (555 aligned C-α atoms), respectively, and the E. coli ferrichrome iron receptor FhuA, with an RMSD of 1.83 Å (600 aligned C-α atoms) (39, 40) (see Fig. S1 in the supplemental material).

As for PiuA, the extracellular loops of PirA from P. aeruginosa were highly flexible, and the structural analysis of Pa-PirA was focused on the structure of PirA from A. baumannii, where these loops where experimentally located. PirA was similar to the ferric enterobactin receptor FepA of E. coli (41), with an RMSD of 1.09 Å (618 aligned C-α atoms), as expected from the elevated sequence similarity (51% amino acid identity) (see Fig. S1 in the supplemental material). This is consistent with a previous report identifying PirA of P. aeruginosa as a secondary transporter of ferric enterobactin (36).

Selection of A. baumannii mutants with decreased susceptibility to BAL30072.

We investigated the possibility of resistance emergence in A. baumannii. Colonies of strains 307-0294 and ATCC 19606 able to grow in the presence of BAL30072 concentrations of 4 or 8 μg/ml were obtained at frequencies of 7 × 10⁻⁸ to 7 × 10⁻⁹. After determining their susceptibility profiles, we selected colonies that presented increased MICs of BAL30072 and MC-1 but unchanged MICs of ATM. Mutations in such colonies are expected to affect mechanisms specific to the siderophore-drug conjugate. Three out of 6 clones (R4, R5, and R6) obtained with strain 307-0294 and none of the 8 clones tested from strain ATCC 19606 fulfilled this criterion (Table 4). Therefore, clones R4, R5, and R6 were selected for further analysis and submitted to whole-genome sequencing. Mapping of the reads on the genome of strain 307-0294 revealed a single but distinct mutation in each of the three strains. Mutant R4 had acquired a frameshift mutation (insertion of A at position 9) in the exbD3 gene, while mutations in clones R5 (deletion of A at position 319) and R6 (insertion of A at position 243) resulted in frameshift mutations in the tonB3 gene (Table 4). Interestingly, the tonB3 gene, but not the tonB1 and tonB2 genes (see Fig. S2 in the supplemental material), has recently been shown to be induced under iron-limiting conditions (42). The TonB3-ExbB3-ExbD3 system could therefore be a likely candidate for providing energy to PiuA and PirA siderophore receptors in A. baumannii.

Decreased growth rates of spontaneous exbD3 and tonB3 mutants of A. baumannii.

We observed that the spontaneous mutants R4, R5, and R6 formed slightly smaller colonies than the wild type when grown on lysogeny broth (LB) agar plates. To assess whether the mutations in the energy-transducing TonB3 system impose a fitness cost, we compared the growth of the mutants to that of the wild-type strain in Mueller-Hinton (MH) broth. As shown in Fig. 5, the tonB3 (R5 and R6) and exbD3 (R4) mutants were characterized by a longer lag phase and a lower final cell density after 20 h growth than the wild type. When the growth medium was supplemented with Fe(II)Cl₂ at a final concentration of 20 μM, the growth phenotype of mutants R4 to R6 could be restored to reach nearly wild-type levels. The data suggest that TonB3-ExbB3-ExbD3 is important for siderophore-mediated Fe³⁺ acquisition and that the fitness of such mutants is likely to be severely decreased in vivo, where iron availability is limited.

Bacterial strains and growth conditions.

The strains and plasmids used in this study are listed in Table 1, and the primers used are shown in Table S1 in the supplemental material. A. baumannii strains were grown in double-strength yeast tryptone (dYT) broth and E. coli and P. aeruginosa in LB at 37°C with shaking (220 rpm). E. coli XL-1 Blue and DH10B were used as cloning hosts. E. coli ST18 was the donor strain for biparental matings, requiring supplementation with 50 μg/ml 5-aminolevulinic acid. Gentamicin (25 μg/ml) or kanamycin (25 μg/ml) was added for plasmid-carrying strains. MIC determinations were performed in MH broth according to CLSI guidelines (52) and were repeated on at least three different occasions.

Mutant selection and growth assay.

Spontaneous mutants of the fully sequenced A. baumannii strain 307-0294 were selected on LB agar plates supplemented with BAL30072. Individual colonies were picked and grown in the absence of antibiotic. MICs were determined, and strains with decreased susceptibility to BAL30072 were retained for further analysis. The growth of mutant strains was performed in microtiter plates in a total volume of 200 μl. Strains were grown for 7 h in LB with shaking (250 rpm) at 37°C. Cultures were diluted to an optical density at 600 nm (OD₆₀₀) of 0.015, and 4 μl was inoculated into 196 μl of MH broth. Growth was followed in a BioTek Synergy 1 plate reader at 37°C with shaking for 1 min before hourly measurement of the optical density at 600 nm. The growth curves represent the average OD₆₀₀ values for duplicate cultures.

PCR amplifications and DNA modifications.

PCR primers are listed in Table S1 in the supplemental material. All primer sequences were based on the genome of A. baumannii ATCC 19606. Screening PCRs were carried out using the DreamTaq DNA polymerase (Thermo Scientific), in accordance with the manufacturer's instructions and the optimized annealing temperature for each primer set. For screening PCRs, bacterial cells were boiled at 95°C for 5 min and subsequently pelleted at 13,000 rpm for 1 min. Phusion DNA polymerase (Thermo Scientific) was used for high-fidelity PCRs (supplemented with 5% dimethyl sulfoxide [DMSO]). Restriction digestions were performed using Thermo Scientific restriction enzymes, according to the manufacturer's instructions, at the appropriate temperature. All ligation reactions were carried out at room temperature using Thermo Scientific T4 DNA ligase. DNA purifications were performed using either the GeneJet PCR purification or the GeneJet gel extraction kit (Thermo Scientific), following the manufacturer's instructions.

Construction of pirA and piuA knockout mutants in A. baumannii.

The generation of unmarked knockout mutants was based on the protocol described by Hoang et al. (53). Briefly, DNA fragments of 500 to 700 bp were PCR amplified using primer pairs A1/A2 and B1/B2, respectively (see Table S1 in the supplemental material). For deletion of pirA, the up- and downstream regions flanking the gene were PCR amplified. For knockout of piuA, the amplified fragments were located in the 5′ and 3′ regions of the genes. After amplification, the obtained A and B fragments were gel purified, and approximately 40 ng of each fragment was used in a PCR fusion with primers A1 and B2, which share an 18-bp homologous region. The resulting fusion products were gel purified and further cloned into the suicide vectors pEX18Gm, via HindIII/EcoRI restriction sites (pirA), and pEX18Km, via SalI/EcoRI (piuA). The cloned knockout fragments were verified by sequencing. The replacement vectors were mobilized into A. baumannii via biparental conjugation, and the generation of the unmarked mutants was carried out as previously described (54). In order to confirm a gene deletion, locus-specific primers (out1 and out2) that bind up- and downstream of the gene were designed and used for amplification of the knockout region. Sequencing of the obtained PCR fragments confirmed successful deletion of the genes.

Construction of expression plasmids.

The piuA and pirA coding regions were amplified by PCR from genomic DNA of P. aeruginosa reference strain PAO1. The piuA coding region was amplified with primers piuA-Xba and piuA-Hind and cloned as a 2,540-bp XbaI-HindIII DNA fragment into the expression vector pIApX2, yielding plasmid ppiuA1.1. Deletions of residues 73 to 76 from loop NL1 (ppiuAΔL1.1) were constructed with a modified protocol of the QuikChange method (55) using primers piuA-Δ1 and piuA-Δ2. The pirA coding region was amplified as a 2,498-bp DNA fragment obtained with primers pirA-Bam and pirA-Hind2 (see Table S1 in the supplemental material). The PCR product was digested by BamHI and HindIII and cloned into pIApX2, yielding plasmid ppirA1.1. The coding regions of piuA (A1S_0474) and pirA (A1S_0980) were amplified by PCR from genomic DNA of A. baumannii strain 307-0294 (provided by T. A. Russo, SUNY, Buffalo, NY, USA), using primers listed in Table S1 in the supplemental material. The 2,436-bp-long piuA fragment was digested with EcoRI and PstI and ligated to the EcoRI- and PstI-cleaved vector pEX1.8, yielding plasmid ppiuA-Ab6.3. The 2,557-bp-long pirA fragment was cleaved with XbaI and PstI and ligated to plasmid pIApX2 cleaved with the same restriction enzymes, yielding plasmid ppirA-Ab4.1. The plasmids were transferred into P. aeruginosa by electroporation. The piuA gene of A. baumannii (piuA-Ab) was further subcloned by BamHI-PstI digestion into the A. baumannii expression vector pBAV1K-5T-gfp (56) (kindly provided by G. Wilharm, Robert Koch Institut, Werningerode, Germany), yielding plasmid ppiuA-Ab7.1. The piuA-Ab gene was further subcloned by XbaI-PstI digestion into pBAV1K-5T-gfp, yielding plasmid ppirA-Ab3.1. The PCR conditions were as follows: denaturation at 95°C for 2 min, followed by 27 cycles of 95°C for 20 s, 57°C for 30 s, and 72°C for 2 min and a final extension at 72°C for 4 min. Plasmids were introduced into P. aeruginosa and A. baumannii by electroporation. Plasmid-containing colonies were selected on LB agar supplemented with carbenicillin at 200 mg/liter (P. aeruginosa) or kanamycin at 50 mg/liter (A. baumannii).

Cloning, overexpression, and purification of PiuA and PirA from P. aeruginosa and A. baumannii.

The positions of the signal peptide cleavage sites of the proteins were predicted with Signal P4.0 (57). The coding sequence of the mature protein Pa-PiuA (PA4514) was amplified from PAO1 genomic DNA using primer pairs Pa-PiuA-F/Pa-PiuA-R (see Table S1 in the supplemental material) and Pfu DNA polymerase. The PCR product was digested by BspHI and EagI restriction enzymes and cloned into vector pET20b using restriction enzymes NcoI and EagI so that the protein was expressed with a PelB signal sequence for the outer membrane localization and a noncleavable C-terminal His₆ tag. Codon-optimized Pa-PirA (PA0931) with a C-terminal, noncleavable His₆ tag was purchased from DNA2.0 and inserted into vector pET20b using HindIII and EcoRV restriction sites.

The coding sequences of the mature proteins Ab-PiuA and Ab-PirA were amplified by PCR from A. baumannii strain ATCC 19606 using KOD DNA polymerase (Novagen) and the primer pairs Ab-PiuA-F/Ab-PiuA-R and Ab-PirA-F/Ab-PirA-R (see Table S1 in the supplemental material). Both genes were cloned with an N-terminal tobacco etch virus (TEV) protease-cleavable His₇ tag in vector pTAMAHISTEV using NcoI and BamHI restriction sites. The pTAMAHISTEV expression vector was obtained by replacing the PelB signal peptide of plasmid pEPELBHISTEV (courtesy of Huanting Liu, University of St Andrews, St Andrews, United Kingdom) with the TamA signal peptide, as described by Liu et al. (58).

Proteins were overexpressed in E. coli strain C43(DE3) or BL21(DE3). Cells were grown at 37°C in LB medium containing 100 mg/liter ampicillin to an OD₆₀₀ of 0.7 and then induced with 0.4 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at 25°C overnight.

For Pa-PiuA and Pa-PirA, harvested cells were resuspended in lysis buffer (50 mM sodium phosphate, pH 7.5, 250 mM NaCl, 10% glycerol, 100 μg/ml lysozyme, and 20 μg/ml DNase). The cells were lysed by two passages through a cell disruptor at 30,000 lb/in² (Constant Systems). Unbroken cells were removed by centrifugation (20 min at 10,000 × g), and membrane fractions were isolated by ultracentrifugation (100,000 × g; 1 h). The supernatant was discarded, and inner membrane proteins were solubilized in buffer 1 (50 mM sodium phosphate, pH 7.5, 250 mM NaCl, 10% glycerol, and 1% [wt/vol] N-lauroylsarkosine [Sigma]). The insoluble outer membrane fraction was isolated by ultracentrifugation (100,000 × g; 1 h), and outer membrane proteins were solubilized in buffer 2 (50 mM sodium phosphate, pH 8, 250 mM NaCl, and 7% [vol/vol] octylpolyoxyethylene [octylPOE] [Bachem]). Insoluble material was removed by ultracentrifugation (100,000 × g; 1 h). The supernatant was applied to a charged Histrap nickel-sepharose high-performance column (GE Healthcare) equilibrated with 1% (vol/vol) octylPOE, 250 mM NaCl, 10 mM imidazole, and 50 mM sodium phosphate, pH 8, and washed with the same buffer containing 15 mM imidazole. Protein was eluted with a linear gradient of imidazole (5 to 150 mM). Fractions containing the protein were pooled; dialyzed against 50 mM Tris, pH 8.4, and 1% octylPOE; loaded on an anion exchange column (Resource Q [GE Healthcare]) equilibrated with the dialysis buffer; and eluted with a linear gradient of 0 to 600 mM NaCl over 12 column volumes. Fractions were pooled and loaded on a Superdex S200 gel filtration column (GE Healthcare) equilibrated with 10 mM Tris, pH 8, 150 mM NaCl, and 0.45% (vol/vol) n-octyltetraoxyethylene (C₈E₄) (Bachem). Protein fractions were pooled and concentrated to 8 mg/ml.

The expression and initial purification steps for Ab-PiuA and Ab-PirA were the same as for Pa-PiuA and Pa-PirA. After applying them to a first immobilized-metal affinity chromatography (IMAC) column, imidazole was removed by dialysis (50 mM Tris, pH 8, 250 mM NaCl, 10% glycerol, 2 mM mercaptoethanol, and 1% octylPOE) before adding TEV protease to cleave the His₇ tag. The protein was applied to a second IMAC column, followed by Superdex S200 gel filtration chromatography (GE Healthcare) (10 mM Tris, pH 8, 150 mM NaCl, 0.45% C₈E₄). The protein fractions were pooled and concentrated to 10 mg/ml.

Crystallization and structure determination.

Crystals of Pa-PiuA appeared at 20°C after several days by mixing 1 μl of protein solution (8 mg/ml) with 1 μl of reservoir solution containing 32% polyethylene glycol (PEG) 400, 0.05 M Tris, pH 8.5, 0.05 M sodium sulfate, and 0.05 M lithium chloride. Pa-PirA crystals were obtained at 20°C by mixing 1.5 μl of the protein solution (7.5 mg/ml) containing 2.5 mM n-dodecylphosphocholine (Anatrace) with 1 μl of reservoir solution containing 17% PEG 3350, 0.2 M magnesium chloride, 0.1 M Tris, pH 7. The crystals were frozen with the same solution containing 25% PEG 400.

Crystals of Ab-PiuA appeared at 20°C after several days by mixing 1 μl of protein solution (10 mg/ml) with 1 μl of reservoir solution containing 1 M diammonium hydrogen phosphate, 0.1 M sodium acetate, pH 4.6. The crystals were cryoprotected by supplementing the mother liquor with 30% sucrose. Prior to crystallization, Fe³⁺-enterobactin was added to Ab-PirA at a final concentration of 1 mM before loading on a Superdex S200. Ab-PirA crystals appeared a week after mixing at 20°C 1 μl of protein solution (7 mg/ml) with 1 μl of reservoir solution containing 16% PEG 8000, 0.15 M magnesium sulfate, and 0.1 M bicine, pH 9.

Data were collected at beamlines IO4 and IO3 at the Diamond Light Source, Oxfordshire, United Kingdom, and ID14-4 at European Synchrotron Radiation Facility (ESRF). Data were processed with XIA2 (59–63). The structure of Pa-PiuA was solved by molecular replacement using the coordinates of FhuA (Protein Data Bank [PDB] code 2FCP) (39) as a model with the program PHASER (64). The structure of Ab-PiuA was solved using Pa-PiuA coordinates, and the structures of Pa-PirA and Ab-PirA were solved using FepA coordinates (PDB code 1FEP) (41). Models were adjusted with COOT (65), and refinement was carried out using REFMAC in the CCP4 program suite with TLS parameters (66). The final refinement statistics are given in Table S2 in the supplemental material. The quality of all structures was checked with MOLPROBITY (67). Figures were drawn using PyMOL (version 1.8; Schrödinger, LLC).

Whole-genome sequencing.

Samples were sequenced on an Illumina MiSeq V2 instrument. Assembly and analysis were performed off instrument using CLCBio Genomics Workbench v 6.5 (Cambridge, MA). Detection of single nucleotide polymorphisms (SNPs)/indels was accomplished through mapping to a parent reference assembly using the same parameters. Quality-based SNPs were detected at a minimum frequency of 80% using default criteria and verified by examination of de novo assembly. Annotation of parental strains was accomplished through ORF calling using Prodigal software and BLAST analysis at NCBI for regions of interest. Sequence analyses were performed using the BLAST algorithm on genomic data provided by the MIST2.2 database (mistdb.com).

Accession number(s).

Atomic coordinates and structure factors have been deposited in the Protein Data Bank (accession no. 5FOK, 5FP2, 5FP1, and 5FR8).