Heme uptake and utilization by hypervirulent Acinetobacter baumannii LAC-4 is dependent on a canonical heme oxygenase (abHemO)

Abstract

Acinetobacter baumannii is an opportunistic pathogen that causes serious infections in critically ill and immune compromised patients. The ability to acquire iron from the hosts iron and heme containing proteins is critical to their survival and virulence. The majority of A. baumannii hypervirulent strains encode a heme uptake system that includes a putative heme oxygenase (hemO). Despite reports indicating A. baumannii can grow on heme direct evidence of extracellular heme uptake and metabolism has not been shown. Through isotopic labeling (¹³C-heme) we show the hypervirulent A. baumannii LAC-4 metabolizes heme to biliverdin IXα (BVIXα), whereas ATC 17978 that lacks the hemO gene cluster cannot efficiently utilize heme. Expression and purification of the protein encoded by the A. baumannii LAC-4 hemO gene confirmed catalytic conversion of heme to BVIX. We further show inhibition of abHemO with previously characterized P. aeruginosa HemO inhibitors in a fluorescence based assay that couples HemO catalytic activity to the BVIXα binding phytochrome IFP1.4. Furthermore, the hemO gene cluster encodes genes with homology to heme-dependent extra cytoplasmic function (ECF) σ factor systems. The hemophore-dependent ECF system in Pseudomonas aeruginosa has been shown to play a critical role in heme sensing and virulence within the host. The prevalence of a hemO gene cluster in A. baumannii LAC4 and other hypervirulent strains suggests it is required within the host to adapt and utilize heme and is a major contributor to virulence.

Graphical Abstract

A. baumannii is a Gram-negative opportunistic pathogen that in recent years has become a major global threat in the clinical setting [1–3]. In the last three decades the emergence of A. baumannii as a nosocomial pathogen is largely due to the rapid increase in multi drug and extended drug resistant strains [4–7]. The Centers for Disease Control (CDC) have categorized Acinetobacter baumannii as a serious threat requiring constant public health surveillance [8]. Furthermore, the World Health Organization lists multi drug resistant A. baumannii as a priority 1 pathogen in need of new antibiotics [9]. Therefore, it is imperative that new therapeutic strategies are identified to address this growing need for new antibiotics.

A major challenge bacterial pathogens face within the host is the scarcity of iron. The ability of the host to withhold iron is critical to the innate immune response on bacterial insult [10, 11]. As such, bacteria employ a variety of strategies to acquire iron including the production of ferric (Fe³⁺)-siderophores, acquisition of ferrous (Fe²⁺)-iron by the Feo system and heme acquisition systems [12–14]. Uptake of Fe³⁺-siderophores and heme is mediated by specific outer membrane receptors that couple the proton motive force of the cytoplasmic membrane through the TonB protein. An alternative mechanism of iron-acquisition is the direct transport of Fe²⁺ by the Feo permease. The multiplicity of iron-acquisition systems highlights the adaptability required to respond to and colonize a variety of physiological niches.

A comparative genome analysis of A. baumannii clinical isolates identified a total of six gene clusters associated with iron acquisition systems [15]. These included three siderophore clusters for acinetobactin [16, 17], baumannoferrin A and B [18], and the fimsbactin A-F [19]. Two heme uptake gene clusters, heme uptake cluster 1 encoding a TonB dependent outer membrane receptor, periplasmic heme binding protein and an inner membrane ATP-binding cassette (ABC) transporter, and heme uptake cluster 2; encoding a TonB dependent receptor, an extra cytoplasmic function (ECF) sigma factor and its cognate anti-sigma factor as well as a putative heme oxygenase (hemO) (Figure 1). Several bacterial HemO enzymes have been shown to catalyze the oxidative cleavage of heme to biliverdin (BV) and CO, to release iron including those from Corynebacterium diphtheriae, Neisseria meningitidis and P. aeruginosa [20–22]. The majority of clinical strains sequenced encode the ferrous uptake system Feo, acetinobactin and baumannoferrin gene clusters along with heme uptake cluster 1 [15]. Bioinformatic analysis of sequenced A. baumannii genomes identified heme uptake cluster 2 (hemO gene cluster) in approximately 60% of clinical strains irrespective of lineage or sequencing group (Table S1) [23]. Interestingly, the fimsbactin gene cluster was found only in A. baumannii ATCC 17978. In contrast the hemO gene cluster is found in many MDR epidemic strains, including the hypervirulent A. baumannii LAC-4 strain, where it was proposed to account for the increased virulence [24, 25].

Figure 1.

Schematic arrangement of heme uptake cluster 1 (A) and the hemO gene cluster (B). Gene numbering/annotation is shown for A. baumannii LAC-4 genome (ABLAC).

In recent years, the number of sequenced A. baumannii genomes, especially those of MDR clinical isolates has increased significantly [15, 26–32]. Despite the identification of heme uptake clusters in a significant number of A. baumannii clinical strains, no studies have directly shown extracellular heme uptake and degradation. It has been shown that A. baumannii ATCC19606 and 17978 can grow on heme as an iron source in the absence of acinetobactin biosynthesis and transport, although this recovery was not robust [33]. In a separate study, A. baumannii LAC-4 was shown to have robust growth on heme that could be inhibited by Ga(III)-protoporphyrin [25]. In an effort to delineate the contributions of the respective heme uptake clusters to iron acquisition in A. baumannii clinical strains we undertook isotopic ¹³C-heme labelling LC-MS/MS to directly assay heme metabolites. We specifically chose to analyze the hypervirulent clinical isolate LAC-4 and ATC 17978 where both isolates encode the acinetobactin and baumannoferrin A-B gene clusters, and the Feo system [24, 25, 34–36]. While ATC 17978 encodes heme uptake cluster 1 it does not carry the hemO gene cluster but a unique third siderophore uptake system the fimsbactin A-F gene cluster [15, 19]. This cluster is also found in the non-pathogenic Acinteobacter baylyi ADP1 [37].

Herein we describe the first direct evidence that the MDR A. baumannii LAC-4 strain encoding the hemO gene cluster can efficiently utilize heme as an iron source, in contrast to ATC 17978 which lacks the hemO gene cluster. Furthermore, through biochemical and ¹³C-heme isotopic labeling followed by LC-MS/MS we show the uptake and oxidative cleavage of ¹³C-heme in A. baumannii LAC-4 releases iron with the production of ¹³C-BVIXα. We further characterized the purified abHemO protein confirming catalysis of heme to BVIXα. We have previously shown in Pseudomonas aeruginosa that the iron-regulated paHemO is critical to driving the metabolic flux of heme into the cell and that the BVIXβ and BVIXδ metabolites function in the feedback regulation of the heme assimilation system (has) which encodes the extra cytoplasmic function (ECF) [38],[39]. HasI on recruitment of the core RNA polymerase activates transcription of the outer membrane receptor HasR and the extracellular hemophore HasAp required for heme sensing and adaptation to heme as an iron source. We have further shown that hasAp is the most upregulated gene in acute murine lung infection and deletion of the HasR receptor, which is required for transmission of the extracellular heme signal, attenuates virulence in a murine acute lung infection [40]. Interestingly, the A. baumannii hemO gene cluster in addition to the OM receptor encodes an ECF σ/anti-σ factor system. Within the cluster are two genes encoding hypothetical proteins of unknown function. Employing hemin-agarose pull-down assays we identified ABLAC_16810 as an extracellular hemophore. Thus, the presence of the hemO gene cluster in a significant number of A. baumannii hypervirulent and epidemic strains (irrespective of lineage) likely contributes to virulence in part due to the ability to adapt within the host to utilizing heme as an iron source. We propose HemO together with the cell surface signaling system represents a novel therapeutic target and the purification and characterization of A. baumannii HemO represents a first step in developing inhibitors of heme utilization [41–43].

RESULTS AND DISCUSSION

Optimal growth of A. baumannii on heme as an iron source requires the hemO gene cluster

Many epidemic and hypervirulent A. baumannii strains have been shown to encode putative heme uptake and metabolism systems. However, the contributions of the individual heme uptake clusters has not been determined. We set out to define the contributions of the respective heme uptake clusters to the utilization of heme as an iron source. Specifically, the clinical strains LAC-4 which encodes both heme uptake clusters and ATC 17978 which only encodes heme uptake cluster 1. Consistent with the fact both strains encode the acinetobactin and baumannoferrin gene cluster, and in the case of ATC 17978 the fimsbactin cluster, both strains were able to grow in media supplemented with 2.5 μM FeCl₃ (Fig 2A). On supplementation with 1 μM heme the LAC-4 strain was able to grow on heme as an iron source, whereas ATC 17978 on heme supplementation had a similar growth curve to that in low iron conditions (Fig 2B). The data suggests the presence of the hemO gene cluster and specifically hemO is required for efficient heme utilization, despite the presence of heme uptake cluster 1. We further confirmed the expression of abHemO in the LAC-4 strain by Western blot where we observe robust abHemO protein expression at all time points in the presence of heme (Fig 2A). Although we observe expression of abHemO at the early time points in iron-restricted media, at later time points no abHemO protein is detected. As expected, the ATC 17978 strain shows no detectable abHemO, consistent with the fact it does not encode the hemO gene cluster (Fig 2B). In LAC-4 cultures supplemented with 2.5 μM FeCl₃ no abHemO protein is detected consistent with iron-regulation of the hemO gene (Fig 2 A). The iron-dependent repression of hemO suggests at the transcriptional level the gene is regulated by the ferric uptake regulator, Fur. Previous bioinformatics analysis of heme uptake cluster 1 and 2 identified a putative Fur-box upstream of several of the genes in both clusters (Figure 1) [15]. However, the fact that HemO protein levels are maintained in the presence of heme and decrease over time in the absence of heme, suggests heme independent of iron levels may positively regulate HemO expression.

Figure 2.

Growth curves and Western blot detection of HemO for A. baumannii strains LAC-4 (A), and ATC 17978 (B). Growth curves were performed in triplicate in M9 (low iron) or supplemented with 1 μM heme or 2.5 μM FeCl₃. Western blots were performed as described in the Experimental section. Lanes were loaded based on total protein (25 μg) and the RNA polymerase α subunit was run as a loading control.

Extracellular heme is metabolized to BVIXα in A. baumannii LAC-4

To confirm the A. baumannii strains containing the hemO gene cluster are able to transport and utilize extracellular heme, we supplemented A. baumannii LAC-4 with isotopically labelled ¹³C-heme as the sole iron source. We have previously shown by ¹³C-heme isotopic labeling and LC-MS/MS that we can distinguish BVIX isomers derived from both intracellular (¹²C-heme) and extracellular (¹³C-heme) heme based on their respective dominant fragment ions (Figure S1) [38, 44, 45]. LC-MS/MS analysis of A. baumannii LAC-4 supernatants from cultures supplemented with 1 μM ¹³C-heme gave a major fragment ion with an m/z of 301.1 as expected for ¹³C-BVIXα and a less intense fragment ion at 297.1 for ¹²C-BVIXα (Fig 3A). Quantification of the ¹³C-BVIXα versus ¹²C-BVIX metabolites in the extracellular media of bacterial cultures previously grown in low iron is consistent with active ¹³C-heme uptake and metabolism (Fig 3B). We assume the intracellular ¹²C-BVIXα detected in ¹³C-heme supplemented cultures reflects turnover of intracellular heme in maintaining iron homeostasis during the early iron-starved phase of growth prior to active ¹³C-heme uptake. In contrast, A. baumannii LAC-4 cultures supplemented with 200 μM FeCl₃ show only background levels of extracellular BVIXα consistent with utilization of iron by the siderophore pathways and subsequent transcriptional repression of HemO by Fur (Fig 2A). The results are consistent with A. baumannii encoding a single HemO that is required for maintenance of intracellular heme and iron homeostasis. The BVIXα regioselectivity of the A. baumannii HemO is identical to the HemO enzymes purified from Neisseria meningitidis, N. gonorrhea, and Corynebacterium diphtheria [21, 22].

Figure 3.

Heme utilization by A. baumannii LAC-4. (A). LC-MS/MS analysis of extracellular ¹³C-heme (red) and ¹²C-heme (black). A. baumannii cultures supplemented with 1 μM heme or 200 μM FeCl₃ were grown for 6 h. Following pelleting and removal of the cells BVIXα was extracted from the supernatant and subjected to LC-MS/MS as described in the Experimental section. (B) Quantification of BVIXα experiments supplemented with 1 μM heme. BVIXα values were obtained from standard curves corrected for extraction efficiency with an internal standard and OD₆₀₀. Values represent the standard deviation of three separate experiments. The indicated p value as determined by a Student’s two-tailed t-test comparing BVIXα levels derived from ¹²C-heme to those to those obtained from ¹³C-heme, where *, p 0.05. For more details see Experimental section.

Purification and characterization of the A. baumannii LAC-4 HemO

We further cloned, expressed and characterized the A. baumannii LAC-4 HemO enzyme. The purified protein has an estimated molecular weight of 23 kDa (Fig 4A). A heme binding constant (K_D) of 0.43 μM was calculated from the decrease in Trp fluorescence at 337 nm as a function of increasing heme concentration (Fig 4B). The holo-HemO complex has a Soret at 404 nm and a visible band at ~630 nm, a spectrum typical of a His-coordinated high spin ferric (Fe³⁺) heme as previously observed for other bacterial HemOs (Fig 4C) [20–22]. On addition of ascorbate, the heme Soret shifts to 407 nm with the appearance of visible bands at 539 and 578 nm typical of a ferrous-oxy (Fe²⁺-O₂) species (Fig 4C). The Fe²⁺-O₂ species decays over time yielding a spectrum with a broad absorbance band at 392 nm and loss of definition in the visible region consistent with formation of the Fe³⁺-BVIX complex (Fig 4C). HPLC analysis following acidification and extraction of the BVIX isomer confirmed oxidative cleavage and CO release of the α-meso carbon (Fig 4D).

Figure 4.

Characterization of the purified A. baumannii LAC-4 HemO (abHemO). (A) SDS-PAGE of purified abHemO. Lane 1, molecular weight markers; Lane 2, purified abHemO. (B) Plot of the decrease in fluorescence as a function of heme concentration. 1 μM HemO in 20mM Tris-HCl (pH 8.0) was titrated with heme from 0.1 to 15 μM (1 μl aliquots) as described in Experimental. The binding constant (K_d) was calculated from the decrease in Trp fluorescence intensity at 337 nm as a function of increasing heme concentration. (C) In vitro turnover of the holo-abHemO complex. Spectral changes on addition of 200 μM ascorbate to a solution containing abHemO (10 μM) in 20 mM Tris-HCl pH 7.4. The inset (x10) shows the increase in intensity of the visible bands on formation of the Fe²⁺-O₂ complex which subsequently decreases on formation of the Fe³⁺-BVIXα complex. Acidification with 3 N HCl produces the iron-free BVIXα (blue line). (D) HPLC trace of the extracted BVIXα product. The inset shows the elution profile of the BVIX isomer standards.

Coupling HemO activity to the BVIXα bacterial phytochrome IFP1.4

Ascorbate dependent turnover of the bacterial holo-HemO complex does not result in the release of Fe³⁺-BVIXα from the enzyme active site. Therefore, it is not possible to perform multiple enzyme turnovers in the presence of excess substrate as has been reported for the biliverdin reductase coupled eukaryotic heme oxygenases [46, 47]. Furthermore, the addition of a chemical reductant to non-enzymatic heme proteins such as cytochrome b₅ can lead to the coupled oxidation of heme, yielding BVIX as a non-stoichiometric by-product [48, 49]. Therefore, non-enzymatic heme degradation complicates in vitro characterization of catalytically inactive HemO mutants [50]. In an effort to overcome issues of non-specific heme degradation in the presence of chemical reductants we utilized the fluorescence properties of the previously characterized bacterial phytochrome IFP1.4 to directly assay in-cell enzymatic activity [51, 52]. IFP1.4 is a Deinococcus radiodurans phytochrome (BphP) that has been re-engineered to optimize its fluorescence excitation and emission properties on covalent linkage of BVIXα to the IFP1.4 protein. We have previously shown enzymatic activity of a P. aeruginosa HemO BVIXα producing mutant can be supported in E. coli and coupled to the fluorescent reporter IFP1.4 [42]. In this system an endogenous E. coli reductase can support HemO activity for screening potential HemO inhibitors. The advantages of the in-cell inhibitor screening assay are several including (i) a measure of the ability to cross the Gram negative cell wall, (ii) the ability to inhibit HemO (decreased fluorescence) and (iii) a measure of the toxicity of the compound in an E. coli system that does not require HemO activity for survival. On co-expression of HemO and IFP1.4 an increase in fluorescence is observed over time compared to cells expressing HemO alone (Fig 5). The in-cell assay was previously optimized so that expression of IFP1.4 is not rate limiting and a linear increase in fluorescence is observed. To confirm the utility of this in-cell assay we repeated the experiment in the presence of a previously characterized P. aeruginosa HemO iminoguanidine inhibitor (Fig 5; inset) [42]. As shown in Fig 5, the relative fluorescence of IFP1.4 increases over time as BVIXα is produced and is quenched in a concentration dependent manner in the presence of the iminoguanidine inhibitor. The in-cell assay provides a platform for the development of a high-throughput screening assay for the identification of bacterial HemO inhibitors [42].

Figure 5.

HemO activity as monitored by fluorescence of the IFP1.4-BVIXα chromophore in E. coli expressing abHemO. Following induction of abHemO and IFP1.4 cultures were incubated at 25°C with shaking and the OD₆₀₀ and fluorescence of the IFP1.4-BVIXα complex (excitation 630 nm, emission 700 nm) was monitored over 16 hours as described in the Experimental section. Relative expression was corrected for differences in growth rate (OD600) Standard deviation shown for three separate experiments.

The hemO gene cluster encodes a secreted hemophore

Interestingly, the hemO gene cluster also encodes a putative ECF σ/anti-σ factor system upstream of hemO (Figure 1). We hypothesize that this system may function similarly to those described for Serratia marcescens and P. aeruginosa [53–55]. Specifically, a secreted hemophore on heme binding and interaction with the outer membrane receptor, triggers release of a σ factor activating a signaling cascade that leads to transcriptional activation of the hemophore and outer-membrane receptor. Encoded within the hemO cluster, in addition to the ECF σ and anti-σ factor, is a putative TonB-dependent outer membrane heme receptor and two genes encoding hypothetical proteins of unknown function (Figure 1B). However, neither of the genes of unknown function have sequence similarity to previously characterized hemophores [56–59]. To determine if A. baumannii strains encoding the hemO gene cluster secrete a hemophore, we performed hemin-agarose pull downs on A. baumannii extracellular supernatants. SDS-PAGE of the hemin-agarose pull-down fractions of extracellular supernatants from A. baumannii LAC-4 and a second hypervirulent strain AB0057 that also encodes a hemO gene cluster following growth in low iron show enrichment of a protein with a molecular weight of ~26 kDa that is putative hemophore (abHph) (Fig 6). In contrast SDS-PAGE analysis of the pull-down fraction from cultures of ATC 17978 lacking the hemO gene cluster did not show a band corresponding to the putative secreted hemophore. The band at 26kDa from AB0057 and LAC-4 matches the predicted molecular weight of the hypothetical protein products of genes AB57_0988 and ABLAC_16810, respectively. Further analysis of the isolated proteins by electrospray ionization mass spectrometry (ESI-MS) confirmed the most abundant proteins in A. baumannii AB0057 and LAC-4 extracellular supernatants to be the gene products corresponding to the putative hemophore AB57_0988 and ABLAC_16810, and OmpA, an outer membrane protein critical for adhesion and biofilm formation (Table S2) [60, 61]. Although, the putative secreted hemophores from AB0057 and LAC-4 have 99% sequence identity (Figure S2), they show no similarity to the previously characterized hemophores from P. aeruginosa or S. marcescens. The secreted abHph represents a previously unidentified new class of heme binding protein.

Figure 6.

SDS-PAGE of the hemin agarose pull down fraction. Lane 1, Molecular weight markers; Lane 2, A. baumannii LAC-4 supernatant; Lane 3, A. baumannii AB0057 supernatant; 4, A. baumannii ATC 17978 supernatant. 15 μl of the 200 μl hemin agarose eluted sample was loaded in each well. The supernatants prior to the addition of hemin agarose were corrected for OD₆₀₀. See Experimental section for more details.

Conclusions

The rapid rise in A. baumannii MDR and extended drug resistant (XDR) strains highlights the need to develop new strategies to combat infection [62]. The ability to acquire iron is critical to A. baumannii virulence, and the presence of an additional heme utilization gene cluster in hypervirulent and outbreak associated strains suggested this locus may be in part responsible for the increased virulence. Indeed, our data confirms that despite the presence of heme uptake cluster 1, the ability to internalize and utilize heme is dependent on the presence of the hemO gene cluster. In P. aeruginosa we have previously shown the catalytic activity of HemO is required to drive the metabolic flux of heme into the cell despite the upregulation of the heme uptake systems [44]. Furthermore, the P. aeruginosa has system was shown to be the most upregulated operon in dual RNA-sequence analysis comparing P. aeruginosa from a murine acute lung infection to P. aeruginosa grown in vitro. The hasAp gene in particular showed a greater than 300-fold induction [40]. In the same study, a P. aeruginosa ΔhasR strain showed reduced colonization and virulence in the acute murine lung infection model when compared to the wild type strain. It is intriguing to speculate that genes encoding a hemophore-dependent ECF σ factor system within the A. baumannii hemO gene cluster are similarly important to adaptation and virulence within the host. While further investigation of the role the hemophore-dependent ECF system in heme sensing and adaptation is needed, the current studies clearly show the hemO gene cluster is required for optimal utilization of heme in A. baumannii hypervirulent strains. As heme is a significant source of iron and the majority of clinically relevant strains encode the hemO gene cluster, heme sensing and utilization represents a novel antimicrobial target. The purification and characterization of the A. baumannii HemO represents a first step in the screening, identification and development of HemO inhibitors targeting heme utilization in opportunistic Gram-negative pathogens including A. baumannii [42, 43].

Methods

Strains and plasmids

E. coli DH5α and BL21(DE3) were used for propagation and protein expression, respectively. A. baumannii clinical isolates LAC-4, AB0057, and ATC 17978 were selectively grown on Leeds Acinetobacter medium [63] or plates purchased from Hardy Diagnostics. Luria Broth (LB) or minimal medium M9 with 0.5% deferrated casamino acids were used for all liquid cultures.

Growth Curves

A. baumannii strains were first grown in Luria Broth (LB) overnight, cells were collected, washed to remove any remaining LB, and used to inoculate a fresh M9 supplemented with 0.5 % deferrated casamino acids (CAA). The cultures were allowed to grow for 3 hrs to deplete the intracellular iron at which point cultures were supplemented with either 1 μM heme or 2.5 μM FeCl₃. Growth was monitored by measuring optical density (OD) at 600 nm.

Cloning, expression and purification of HemO

The hemO gene was amplified by PCR from A. baumanniii LAC-4 genomic DNA utilizing primers AbhemO-F containing an NdeI site (5’-ATCGTCCATATGATGAATTCATC TACAG AACAAATG-3’) and AbhemO-R containing an XhoI site (5’-ACGTACCTCGAGTTATTTCAATTCATCA AG-3’). The NdeI/XhoI digested PCR fragment was cloned into NdeI/XhoI linearized pET21. Positive colonies containing the abhemO gene were selected and sequenced. The resulting pETabhemO vector was transformed into E. coli BL21(DE3) for protein expression. A single colony was used to inoculate a LB liquid culture (20 ml) containing 100μg/mL ampicillin and grown with shaking overnight at 37°C. This culture was used to seed 1 L LB-ampicillin cultures to an OD600 of 0.05 which were further grown to an OD600 of 0.5, following which protein expression was induced on addition of 1 mM IPTG (final concentration). The protein expression cultures were grown at 30°C for a further 4–5 hrs. For purification, pellets were thawed on ice and lysed with stirring in 50 mL lysis buffer (50 mM Tris-HCl (pH 8.0) containing 50 mM NaCl, 100 μM phenylmethanesulfonyl fluoride (PMSF), lysozyme, DNAse I, protease inhibitor cocktail (complete mini, Roche Diagnostics)) for 1 h. Lysate was sonicated then centrifuged at 21,000 g at 4 °C for 1 h. Purification was accomplished using ion exchange and size exclusion chromatography. Cell lysate was applied to a Q-Sepharose fast flow column (30 × 100 mm) previously equilibrated with 10 mM Tris-HCl (pH 8.0). The column was washed (10 column volumes) with the same buffer containing 50 mM NaCl and eluted over a linear salt gradient of 50–500 mM NaCl. The HemO containing fractions were pooled concentrated to 2 ml and run over a Superdex 200 prep column (10 × 300 mm) in 10 mM Tris-HCl containing 50 mM NaCl (pH8.0). The HemO fractions were pooled and dialyzed against 20 mM Tris-HCl and the resulting protein was judged to be > 98% pure by SDS-PAGE.

Western blots

Aliquots (1 ml) of A. baumannii strains were harvested at 2, 5 and 7 h. Cell pellets were resuspended in 200 μl per OD₆₀₀ of 1.0 in Bugbuster (Novagen). Cells were incubated at room temperature for 30 min with occasional agitation to ensure complete cell lysis. Total protein concentrations were determined using the BioRad RCDC Assay. Total protein (25 μg) in 2X SDS-loading buffer was run on 12% SDS-PAGE. Proteins were transferred by electrophoresis to a PVDF membrane (Bio-Rad). Membranes were blocked with blocking buffer (5% w/v skim milk in Tris-buffered saline (TBS) with 0.2% v/v Tween 20), washed and probed with 1:1000 dilution of anti-HemO primary antibodies in hybridization buffer (1% w/v skim milk in TBS with 0.2% v/v Tween 20). Polyclonal antibodies were obtained from Covance Custom Antibodies, NJ and generated from purified abHemO protein supplied by our laboratory. Antibody specificity and sensitivity was checked against serial dilution of purified HemO protein (0.010–1 μg). All experimental Western blots were run with molecular weight markers as standards. RNA polymerase α was probed as the loading control with the E. coli RNA polymerase primary antibody (BioLegend) at a dilution of 1:1000. Membranes were rinsed three times in TBS with 0.2% (v/v) Tween 20 and probed with goat-anti rabbit immunoglobin G conjugated to horseradish peroxidase (KPL) at a dilution of 1:10000 in hybridization buffer. Proteins were visualized by chemiluminescent detection using the Super-Signal chemiluminescence kit (Pierce) and Amersham hyperfilm ECL (Amersham).

Heme binding

A solution of 1 μM abHemO in 20mM Tris-HCl (pH 8.0) was titrated with heme in 1 μl aliquots from 0.1 to 15 μM. Titrations were monitored by fluorescence quenching. Samples were excited at 295 nm and fluorescence emission was monitored at 337 nm. The binding constant (K_d) was calculated from the decrease in Trp fluorescence intensity at 337 nm as a function of increasing heme concentration. Heme stocks were prepared immediately prior to use by dissolving in 0.1 N NaOH and buffered to pH 7.4 with 1 M Tris-HCl and the final concentration determined by pyridine hemochrome (pH 7.4) [64].

In vitro abHemO enzyme activity measurements

Purified abHemO was incubated with 1.5 molar equivalents of heme from freshly prepared hemin solution in 0.1 M NaOH brought up with 20 mM Tris-HCl (pH 7.4). Excess hemin was removed by passage of the holo-HemO protein over a Q-Sepharose column (1.5 cm × 1.0 cm) and the holo-HemO eluted with 250 mM NaCl in 20 mM Tris-HCl (pH 7.4). Turnover of the holo-abHemO WT was performed in the presence of ascorbate as previously described [20]. Briefly, the reactions contained 10 μM holo-abHemO in a final volume of 1 ml. Catalase (final concentration of 0.1 mg/ml) was present as a H₂O₂ scavenger. The reactions were initiated with the addition of ascorbate (1 mM) to a final concentration of 200 μM. Heme turnover was monitored spectroscopically between 300–700 nm with repeated scanning every 30 s for 30 min. To obtain the spectrum of iron free BVIXα the reaction was acidified with 3 N HCl (10 μl). BVIXα was extracted into chloroform and analyzed by HPLC as previously described [21]. Briefly, the chloroform extract was washed with distilled water (3×1 ml), and the chloroform removed by evaporation. The residue was analyzed by reverse phase HPLC on a Phenomenex Luna 5 μ, C18(2) (250 × 4.6 mm) column and eluted with 50:50 (v/v) 20 mM formic acid:acetone at a flow rate of 1 ml/min. The eluent was monitored at 376 and 685 nm and the BVIX standards eluted in the order α (27.5 min), β (34 min), δ (37.5 min), and γ (49 min).

HemO in-cell fluorescence assay

The fluorescence assays were performed as previously described with some minor modifications.[42] E. coli BL21 (DE3) were transformed with a pBADhemO and pETIFP1.4-His₆. Protein expression was performed at 37°C in LB medium containing 25 μg/ml chloramphenicol and 30 μg/ml kanamycin to an OD₆₀₀ = 0.8. IFP1.4 was induced with IPTG (final concentration 1 mM) and cells expressed at 25°C for a further 2 h. HemO was then induced on addition of arabinose (final concentration 0.02%) and 200 μL aliquots transferred to a black, clear-bottom 96-well plate. Cells were maintained at 25°C with orbital shaking simultaneously monitoring both OD₆₀₀ and fluorescence emission at 700 nm following excitation at 630 nm. Fluorescence on formation of the BVIXα-IFP1.4 complex was monitored every twenty minutes over a sixteen-hour period on a Synergy H1 hybrid microplate reader. Fluorescence change over time was corrected to account for differences in OD₆₀₀ between samples. Negative controls were performed on IPTG induction of IFP1.4 expression in cultures transformed with pETIFP1.4-His₆ plasmid only.

Isotopic labeling and LC-MS/MS

¹³C-heme was produced from the precursor [4-¹³C]–δ-aminolevulinic acid according to the method described by Rivera and Walker [65]. A. baumannii LAC-4 cultures (50 ml) were grown as described above and supplemented with either 1 μM ¹³C-heme or 2.5 μM FeCl₃ and grown for 6 hr. Culture supernatants following pelleting of the cells were first filtered through a 0.2 μm PES syringe filter and then spiked with 1 nM BVIXα dimethyl ester standard. The supernatant was acidified to pH ~3 with 10% trifluoroacetic acid (TFA). A C₁₈ Sep-Pak column (Waters) was equilibrated with 2 ml of acetonitrile, 2 ml of methanol, 2 ml of water, and 2 ml of 10% methanol in 0.1% TFA (v/v). The supernatant was loaded on the column, and washed with 4 ml of 0.1% (v/v) TFA, 4 ml of acetonitrile:0.1% TFA (v/v 20:80), 2 ml of methanol:0.1% TFA (50:50 v/v) and 450 μl of methanol. BVIX isomers were eluted with additional 600 μl of methanol, dried under vacuum and stored at −80°C prior to LC-MS/MS analysis.

For LC-MS/MS analysis samples were resuspended in 10 μl DMSO and diluted to 100 μl with 90 μl of H₂O:Acetonitrile (50:50). Samples were run over an Ascentis RP-amide 2.7 μm (C18) HPLC column (10 cm × 2.1 mm) with a flow rate of 0.4 ml/min and analyzed on a Waters TQD triple quadrupole mass spectrometer with AQUITY H-Class UPLC. Fragmentation patterns of the parent BVIXα ions derived from either ¹²C-heme or ¹³C-heme and the BVIXα dimethyl ester standard were analyzed using multiple reaction monitoring (MRM). The source temperature was set to 150 °C, the capillary voltage to 3.6 kV, and the cone voltage to 75 V. The collision energy was set to 43 for the BVIXα dimethyl ester internal standard and 34 V for ¹²C-BVIXα and ¹³C-BVIXα isomers. Standard curves for quantification were prepared with known concentrations of the BVIXα isomer. The lower and higher limit of quantification was 0.001 and 150 μM, respectively. The [4-¹³C]- δ-ALA labeling shifts the heme mass by 8 yielding distinct BVIX isomer fragmentation patterns for ¹³C-BVIXα shifted from the ¹²C-BVIX fragment by a mass of +4 (Figure S1).

Hemin agarose affinity purification

Hemin-agarose binding studies were performed as previously described with slight modification [66]. A. baumannii LAC-4, AB0057 and ATC 17987 were grown in M9 containing 0.5% deferrated casamino acids as described for the growth curves. At 4 h an aliquot (5 ml) was harvested and the extracellular medium retained. 100 μl hemin-agarose (Sigma-Aldrich) previously equilibrated in 25 mM Tris-HCl (pH 7.4) containing 100 mM NaCl was added to 5 ml of the extracellular medium. The suspension was incubated with gentle mixing for 1 h at 37^oC following which the resin was pelleted (10,000 × g for 5 min) and washed three times in 25 mM Tris-HCl (pH 7.4) containing 100 mM NaCl. Bound protein was eluted by the addition of 2% (wt/vol) SDS and 1% (vol/vol) β-mercaptoethanol in 500 mM Tris HCl (pH 6.8), boiled at 100°C for 5 min running buffer and the resin removed by centrifugation. Proteins were then analyzed on 12% SDS-PAGE.

Proteomic analysis of hemin-agarose pull-down products

Proteins retained on hemin-agarose beads were eluted with 5 % sodium deoxycholate at 95 °C. Proteins were washed, reduced, alkylated and trypsinolyzed in filter as previously described [67, 68]. Tryptic peptides were separated on a nanoACQUITY UPLC analytical column (BEH130 C18, 1.7 μm, 75 μm × 200 mm, Waters) over a 165-minute linear acetonitrile gradient (3 – 40%) with 0.1 % formic acid on a Waters nano-ACQUITY UPLC system, and analyzed on a coupled Thermo Scientific Orbitrap Fusion Lumos Tribrid mass spectrometer [69]. Survey scans were acquired at a resolution of 120,000, and precursors were selected for fragmentation by higher-energy collisional dissociation (normalized collision energy at 32 %) for a maximum 3-second cycle. Tandem mass spectra were searched against UniProt A. baumannii proteomes using Sequest HT algorithm [70]. Carbamidomethylation of cysteine and deamidation of asparagine and glutamine were treated as static and dynamic modifications, respectively. Resulting hits were validated at a maximum false discovery rate of 0.01 using a semi-supervised machine learning algorithm Percolator [71]. Proteins were quantified by measuring the MS1 peak intensities of peptide ions, whose identities were confirmed by MS2 sequencing as described above.

Supplementary Material

1

Table S1. Sequenced A. baumannii strains encoding the hemo gene cluster.

Figure S1. LC-MS/MS BVIX fragmentation patterns.

Figure S2. Sequence alignment of ABLAC_16810 and AB57_0988.

Table S2. Proteomic analysis of A. baumannii hemin agarose pull downs

ABBREVIATIONS

LC-MS/MS

liquid chromatography (LC) with tandem mass spectrometry (MS/MS)

CSS

cell surface signaling

ECF

extra cytoplasmic function

MDR

Multi-Drug Resistance

Has

heme assimilation system