The NEAT Domain-Containing Proteins of Clostridium perfringens Bind Heme

The ability of a pathogenic bacterium to scavenge iron from its host is important for its growth and survival during an infection. Our studies on C. perfringens gas gangrene strain JIR325, a derivative of strain 13, showed that it is capable of utilizing both human hemoglobin and ferric chloride, but not human holo-transferrin, as an iron source for in vitro growth. Analysis of the C. perfringens strain 13 genome sequence identified a putative heme acquisition system encoded by an iron-regulated surface gene region that we have named the Cht (Clostridium perfringens heme transport) locus. This locus comprises eight genes that are co-transcribed and includes genes that encode NEAT domain-containing proteins (ChtD and ChtE) and a putative sortase (Srt). The ChtD, ChtE and Srt proteins were shown to be expressed in JIR325 cells grown under iron-limited conditions and were localized to the cell envelope. Moreover, the NEAT proteins, ChtD and ChtE, were found to bind heme. Both chtDE and srt mutants were constructed, but these mutants were not defective in hemoglobin or ferric chloride utilization. They were, however, attenuated for virulence when tested in a mouse myonecrosis model, although the virulence phenotype could not be restored via complementation and, as is common with such systems, secondary mutations were identified in these strains. In summary, this study provides evidence for the functional redundancies that occur in the heme transport pathways of this life threatening pathogen.


Introduction
Many pathogenic bacteria require iron to replicate and successfully cause an infection in the mammalian host. Iron is used in many important biological processes including DNA synthesis, the generation of energy and protection against reactive oxygen species [1]. However, the availability of iron in the mammalian host is limited because most iron is sequestered within cells, in the form of ferritin (iron-storage protein) and heme, an iron-porphyrin complex.
were attenuated for virulence, secondary mutations presumably were responsible for this phenotype. This study highlights the importance of iron acquisition in C. perfringens JIR325 since functionally redundant uptake systems were found to be encoded by this bacterium.

Ethics statement
All virulence trials in this study were approved by the Monash University Animal Ethics Committee (Ethics number: MARP/2011/093) and were conducted in accordance with the guidelines of the Victorian State Government legislation.

Computational analyses
The accession number for the strain 13 genome sequence is BA00016.3, with the cht locus located between coordinates 291056 and 300674. Sequence similarity searches were performed using the BLAST program (www.ncbi.nlm.nih.gov/BLAST/). Pairwise identity scores of the NEAT domain amino acid sequences were obtained using the Clustal Omega multiple sequence alignment program (www.ebi.ac.uk/Tools/msa/clustalo/).

Preparation of human hemoglobin
Red blood cells were freshly collected from a volunteer donor and washed in 0.9% (w/v) NaCl with centrifugation at 1,000 × g. Following collection of the red blood cells, and at each step in the procedure, the sample was saturated with carbon monoxide to maintain the human carboxyhemoglobin (HbCO) state, which is resistant to auto-oxidation. Red blood cells were lysed under hypotonic conditions and the stroma removed by centrifugation following re-adjustment by NaCl to 0.9% (w/v). The hemolysate was dialyzed against 20 mM Tris.HCl, 0.5 mM ethylenediaminetetraacetic acid (EDTA), pH 8.2, at 4°C and loaded onto pre-equilibrated Q-Sepharose anion exchange resin (GE Healthcare). The column was developed with a gradient of 0-0.2 M NaCl over 15 column-volumes and peak fractions containing HbCO were pooled. To obtain human oxyhemoglobin (HbO 2 ), aliquots of HbCO were stirred on ice in a stream of pure O 2 under illumination using white-light CREE LEDs (Ay Up Lighting systems, Samford QLD) as a high intensity light source. Complete conversion to HbO 2 was monitored by UV-visible spectroscopy. Final globin concentrations were determined by UV-visible absorbance using the molar absorption coefficients ε 419 = 192 000 M -1 cm -1 at 419 nm for HbCO, and ε 415 = 129 000 M -1 cm -1 at 415 nm for HbO 2 .

Molecular and genetic techniques
E. coli plasmid DNA was isolated by using an alkaline lysis method as previously described [35] or a QIAgen Spin Miniprep kit according to the manufacturers' instructions. Chemically competent E. coli cells were prepared as previously described [36], unless stated otherwise. Electrocompetent C. perfringens cells were prepared as previously described [37] and electroporated with an electric pulse of 1.8 kV, resistance of 200 O and capacitance of 25 μF using an ECM630 Electro Cell Manipulator (BTX 1 Harvard Apparatus). C. perfringens genomic DNA was isolated as before [38].
Restriction endonuclease digestions were prepared according to the manufacturers' instructions (Roche Diagnostics, New England Biolabs). DNA ligations, DNA modifications and PCR were carried out as previously described [39]. The concentrations of DNA and RNA were determined using a Nanodrop 1 spectrophotometer (ThermoFisher Scientific). Nucleotide sequencing reactions were performed by Micromon (Monash University, Clayton, Australia) using the PRISM Big Dye Terminator Premix (Applied Biosystems) and the sequences analyzed using the ContigExpress™ software (Thermofisher Scientific). Oligonucleotide primers synthesized by Sigma-Aldrich. All plasmids and oligonucleotide primers used in this study are listed in S1 Table. Construction of chtDE and srt mutants C. perfringens chtDE and srt mutants were constructed by allelic exchange and Targetron technology, respectively. To generate the chtDE deletion mutant, an chtDE suicide plasmid, pJIR3929, was constructed. A 2 kb fragment containing the region directly upstream of chtD, including the first 77 bp of chtD, was PCR amplified using primers JRP4422 and JRP4410 (S1 Table). The PCR fragment was cloned into the multiple cloning site of pCR-BluntII-TOPO to obtain pJIR3926. The cloned fragment was digested with BamHI and PstI, purified using the QIAquick Gel Extraction kit (Qiagen), and subcloned into the equivalent sites upstream of the chloramphenicol resistance cassette of pJIR2783, generating pJIR3928. In a separate reaction, a 2.1 kb fragment encompassing the region directly downstream of chtE, including the last 24 bp of chtE, was PCR amplified using primers JRP3991 and JRP3992 (S1 Table). This PCR fragment was cloned into the multiple cloning site of pCR-BluntII-TOPO to generate pJIR3924. The plasmid was digested with SacI and XhoI, the 2.1 kb fragment was purified as before and sub-cloned into the equivalent sites downstream of the chloramphenicol resistance cassette of pJIR3928 to obtain pJIR3929. All recombinant plasmids were confirmed by restriction digestion and nucleotide sequencing. The plasmid pJIR3929 was used to transform the wild-type C. perfringens strain, JIR325. Chloramphenicol-resistant and erythromycin-susceptible transformants were isolated and confirmed by PCR and Southern blot hybridization. The resultant chtDE deletion mutant was designated, JIR12755.
To construct an srt Targetron mutant, the best intron insertion site (lowest p-value) predicted by the Intron Site Finder (T. Seemann, personal communications) was chosen and the corresponding primers JRP4763, JRP4764, JRP4765 and JRP3867 used to generate a 353 bp PCR product using pJIR3562 as the DNA template. The PCR fragment was cloned into the HindIII and BsrGI sites of the Targetron plasmid pJIR3566, to generate pJIR4038. This vector was used to transform JIR325. Erythromycin-resistant and chloramphenicol-susceptible transformants were isolated and confirmed as before. One of the resulting srt mutants, JIR12614, was chosen for further studies.

Complementation of chtDE and srt mutants
The chtDE mutant was complemented in cis by allelic exchange using a suicide vector, pJIR4055, which was designed to insert an erm(Q) cassette ligated to the chtD chtE coding sequence construct upstream of the promoter of the chtR gene. Briefly, the erm(Q) cassette of pJIR2715 was digested with PstI/NotI and subcloned into the corresponding sites of pJIR3926. The resulting plasmid pJIR4032, contained the 2.1 kb fragment encompassing the region upstream of chtD followed by an erm(Q) cassette. In a separate reaction, a 3.7 kb fragment encompassing chtD and chtE, including 549 bp and 158 bp of the upstream and downstream sequences, respectively, was PCR amplified using primers JRP4903 and JRP4583. The PCR fragment was digested with Asp718/BamHI and subcloned into pJIR751, to generate pJIR3996. A 2.2 kb region downstream of chtE was PCR amplified using primers JRP4901 and JRP3992, digested with XbaI/SphI and ligated in-frame to the C. perfringens chtDE sequence of pJIR3996. The resulting plasmid, pJIR4029, contained a 5.3 kb fragment of the wild-type sequence from 549 bp upstream of chtD to 650 bp downstream of the start codon of chtA. This fragment was subcloned into the SacI/PvuI site of pCR-BluntII-TOPO to generate pJIR4053.
Finally, the SacI-digested fragment of pJIR4032 containing the cloned sequence was subcloned into pJIR4053. The resulting plasmid, pJIR4055, was confirmed by restriction digestion and DNA sequencing, and then used to generate the in cis complemented chtDE strain, JIR12773.
For complementation of the srt mutant in trans, an Asp718/SphI-digested 5.2 kb fragment of pJIR4029 containing the sequence 548 bp upstream of chtD to chtA was cloned into an E. coli-C. perfringens shuttle vector, pJIR750. The srt complementation plasmid, pJIR4151, was used to transform the srt mutant to generate the in trans srt complemented strain, JIR12861.

RT-PCR analysis
RT-PCR was used to analyze the transcription of genes in the cht locus. Total RNA was isolated from JIR325 cells grown to mid-logarithmic phase under iron-limited conditions (TSB containing 100 μM 2,2-dipyridyl). Reverse transcription (RT) was carried out on 2 μg of RNA using 45 units of the AMW Reverse Transcriptase (Promega) in a 20 μL reaction, according to the manufacturer's instructions, with several modifications including the use of 0.5 μg of random primers (Promega) and incubation at 42°C for 1 h. The RT reaction was terminated by heating the sample at 99°C for 5 min. For all RT-PCR analysis, a no RT control reaction (NRT) was performed, in which the reverse transcriptase was replaced with an equal volume of nuclease-free water. PCR amplifications were performed on the RT or NRT reactions using Taq polymerase (NEB) and the primers listed in S1 Table. As PCR controls, the sample was replaced with either nuclease-free water (negative control (NC), or with JIR325 genomic DNA (positive control (DNA)).

Growth assays
C. perfringens cultures grown for 6 h in FTG media were used to inoculate 20 ml of pre-reduced TSB broth. The following day, 100 ml of fresh pre-reduced TSB broth (normal conditions), as well as pre-reduced TSB broths containing 150 μM 2,2-dipyridyl (Sigma) (iron-depleted conditions) only or supplemented with 200 μM FeCl 3 (Sigma), 40 μM commercial human Hb (Sigma), 50 μM HbCO, 25 μM and 50 μM HbO 2 or 7 μM human transferrin (Merck) in the holo-form (iron-replete conditions) were inoculated to a final turbidity at 600 nm of 0.1. Following reconstitution of the commercial lyophilized Hb, analysis by UV-visible spectrophotometry revealed absorbance peaks at approximately 405, 500 and 620 nm; comparison with published data [39] suggested this material contained predominantly ferric heme (methemoglobin (metHb)). The appropriate iron sources were added after 1 h of growth under irondepleted conditions. Cultures were incubated at 37°C and bacterial growth was monitored by measuring the turbidity at 600 nm at 1 h intervals for 8 h.

Virulence studies in a mouse myonecrosis model
The virulence of the isogenic C. perfringens strains was assessed in six-to eight-week old female BALB/c mice (purchased from the Walter and Elisa Hall Institute, Melbourne, Victoria) as previously described [40]. Mice were injected in the right hind thigh muscle with 50 μL of culture (approximately 10 9 CFU) and monitored at 30 min intervals for a total of 12 h post-infection for disease symptoms including limping, blackening of the thigh or footpad, and swelling of the footpad. Mice were euthanized by CO 2 asphyxiation when they showed severe disease in any of the symptoms other than swelling of the thigh. Statistical comparisons were carried out using the Mantel-Cox test for survival curves and the Mann-Whitney U test for survival time using GraphPad Prism software (GraphPad Software Inc).

Purification of recombinant proteins and antibody generation
ChtD, ChtE and Srt were purified as recombinant proteins containing an N-terminal His 6 -tag. The resultant proteins lacked the predicted signal sequence at the N-terminus and/or the transmembrane domain and charged C-terminus. The gene regions encoding amino acids 31-711 of ChtD, 31-191 of ChtE and 29-187 of Srt, were PCR amplified from genomic DNA of JIR325 using the specific primers listed in S1 Table. The primer pairs were flanked with NdeI/BamHI or NheI/XhoI restriction sites, and the reverse primers for the chtD and chtE gene fragments contained a stop codon (TAA). The PCR product and pET28a(+) expression vector were digested with the appropriate restriction enzymes and ligated to generate pJIR4080 (His-ChtD), pJIR4075 (His-ChtE) and pJIR4077 (His-Srt). The cloned gene fragments were confirmed by sequence analysis.
The plasmids were used to transform E. coli BL21 DE3(C43) cells to kanamycin resistance, and were grown overnight at 37°C on 2x YT agar supplemented with kanamycin. The transformants were used to inoculate 10 mL of 2x YT broth with kanamycin and grown overnight at 37°C. Overnight cultures were then subcultured (1:100 volume ratio) into 500 ml of 2x YT broth and incubated at 37°C with shaking (220 rpm) until the turbidity at 600 nm reached 0.6. Then, 0.5 mM of IPTG (isopropyl-β-D-galactopyranoside) was added to induce protein expression and the cultures incubated for an additional 4 h at 37°C. Cells were harvested by centrifugation at 9,300 x g for 10 min, washed once in phosphate buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 , pH 7.4) and resuspended in lysis buffer (20 mM Tris-HCl, 0.3 M NaCl, 10% glycerol, 5 mM imidazole, pH 7.9). The cells were lysed using an EmulsiFlex-C5 cell disrupter (Avestin Inc). The cell lysate was centrifuged at 27,000 x g for 20 min and the supernatant, which contained soluble proteins, subjected to affinity chromatography using Talon 1 resin (Clonetech Laboratories) pre-equilibrated with lysis buffer. After 2 h of incubation at 4°C, the resin was washed three times with 10 ml of lysis buffer to remove unbound proteins. His-tagged proteins were eluted from the resin with 5 ml aliquots of elution buffer (60 mM imidazole, 20 mM Tris-HCl, 0.3 M NaCl, 10% glycerol, pH 7.9) and dialyzed overnight against 1.5 L of Tris buffer (150 mM NaCl, 50 mM Tris, 0.5 mM EDTA, 20% (v/v) glycerol, pH 7.5) with constant stirring at 4°C. The concentrations of the purified proteins were measured using the Pierce bicinchoninic acid (BCA) protein assay according to the manufacturer's instructions (ThermoFisher Scientific).
Polyclonal antibodies specific to ChtD, ChtE and Srt were raised in New Zealand white rabbits by immunization with a total of 2 mg of each purified protein at the Institute of Medical and Veterinary Science, South Australia. Western blotting was used to determine the specificity and the final dilution for the ChtD (1: 80,000), ChtE (1:50,000) and Srt (1:2,500) antibodies.

SDS-PAGE and Western blotting
Proteins or cell lysates were separated on 10% (w/v) SDS-PAGE acrylamide gels and stained using Coomassie Brilliant Blue, or transferred onto a Whatman Protran nitrocellulose membrane at 100 V for 1 h using a mini Trans-Blot 1 Electrophoretic Transfer Cell (BioRad). Membranes were blocked for 1 h with 5% (w/v) skim milk (Diploma, Australia) in TBS-Tween [5 mM Tris-HCl, 15 mM NaCl, pH 7.4, 0.05% (v/v) Tween20]. The membranes were then washed twice (15 min wash followed by a 5 min wash) in TBS-Tween, incubated with a primary antibody in TBS-Tween for 1 h, washed again as before, incubated with a secondary antibody conjugated to horseradish peroxidase and washed as before. All incubations and washes were performed with shaking at room temperature. Membranes were developed with an enhanced chemiluminescent detection reagent (Amersham Life Science) in a FPM-100A film developer (FujiFilm).
Heme association using absorbance spectroscopy E. coli cells carrying the plasmids encoding the His-ChtD, His-ChtE or His-Srt proteins were grown in 1 L of 2x YT broth supplemented with kanamycin as before and induced for 17 h at 30°C. Each protein was purified as before and the His-ChtD (10 μM), His-ChtE (30 μM) and His-Srt (10 μM) proteins were diluted in Tris buffer. A hemin control sample was prepared by diluting 1 mg/mL of (bovine) hemin solution in Tris buffer (pH7.5) to a concentration of 10 μM. The absorbance of each sample between the wavelengths 250-650 nm was measured using a Cary 1 UV-Visible spectrophotometer (Agilent 1 ) and normalized using Tris buffer as a reference.

Heme detection by MALDI TOF analysis
Protein samples were diluted in 50% (v/v) acetonitrile-0.1% (v/v) trifluoroacetic acid (TFA) solution in a 1:10 ratio. The sample solution was mixed in a 1:1 ratio with a matrix solution of 10 mg/mL α-cyano-4-hydroxycinnamic acid (Laser Biolabs) in TFA solution. The mixture was spotted onto a matrix-assisted laser desorption ionization (MALDI) target plate. The mass spectrometry and tandem time-of-flight (TOF/TOF) spectra were determined using a 4700 Proteomics Analyzer (Applied Biosystems) with a mass range of 500 to 3000 Da and focus mass of 1200 Da at 1500 shots per spectra. The spectra were processed using the 4000 series explorer software version 3.0 and calibrated using the plate method calibration against the 4700 mix peptide standards (Applied Biosystems).

C. perfringens utilizes ferric chloride and human hemoglobin
To determine the ability of C. perfringens to utilize various iron sources, strain JIR325 was grown in vitro in iron-depleted TSB media supplemented with host iron sources including human Hb and holo-transferrin, as well as an inorganic iron source, FeCl 3 (iron-replete conditions). As controls, JIR325 was also grown in TSB alone (normal conditions) and in TSB containing 150 μM 2,2-dipyridyl, a chelator of free iron (iron-depleted conditions). The lack of growth of JIR325 in iron-depleted TSB indicated that iron was essential for its growth (Fig 1). The addition of 200 μM FeCl 3 to the iron-depleted media restored growth to the level observed under normal conditions. Similarly, growth was restored in the presence of 40 μM commercial Hb (predominantly, metHb), 25 μM and 50 μM HbO 2 , albeit partially (Fig 1), but was not restored in 50 μM HbCO, under the conditions tested. Lower concentrations, including 10 μM of both the commercial Hb and freshly prepared HbO 2 did not restore growth (data not shown). These findings indicated that JIR325 was capable of utilizing iron from FeCl 3 and human Hb in both the metHb and HbO 2 form. By contrast, the lack of growth of JIR325 in iron-replete media supplemented with human holo-transferrin (Fig 1) indicated that JIR325 was unable to utilize this host iron source for its in vitro growth under the conditions tested.

Identification and transcriptional analysis of the cht locus
Studies in several Gram-positive bacteria have shown that heme uptake is mediated by Isd or Isd-like systems [7,41,42]. Therefore, the cht locus of JIR325 was investigated to determine its role in heme acquisition by C. perfringens. The cht locus of JIR325 comprises eight genes (Fig  2A), and encodes components similar to those of the Isd system from S. aureus. The chtD and chtE genes of JIR325 encode putative cell surface proteins containing four and one NEAT domain(s), respectively. These NEAT domains have amino acid sequence similarity to each other (22.0% to 44.9% identity) and to most NEAT domains of S. aureus (8.2% to 21.3%) and B. anthracis (12.6% to 31.4%) (Alignments shown in S1 Fig). ChtD has regions of similarity to the Isd hemoprotein receptors of S. aureus, which include an N-terminal signal peptide, a Cterminal cell wall sorting signal consisting of a LPQTG sortase recognition motif, a hydrophobic domain and a positively charged tail (Fig 2B). ChtE also has an N-terminal signal peptide and NKESS or SEATG residues at the C-terminus, potentially for sortase recognition. The Srt protein is a putative sortase (transpeptidase) (Fig 2A) belonging to the class D sortase family, as reported in a previous study [43,44]. This protein contains an N-terminal transmembrane region (Fig 2B), suggesting that it may be localized to the cell membrane. Based on the prediction that ChtD and ChtE are cell surface proteins, it was postulated that ChtD and/or ChtE were exported through the general secretory pathway and tethered to the C. perfringens cell surface by the Srt enzyme.  Downstream of srt, three genes designated as chtA, chtB and chtC1, are predicted to encode a periplasmic binding protein, a permease and an ATP-binding protein of a putative ABC transport complex, respectively (Fig 2A). Two other ORFs downstream of chtC1, known as chtC2 and chtC3, encode putative permease/ATPase fusion proteins containing a transmembrane domain and an ATPase domain at the N-and C-terminal regions, respectively.
The genes in the cht locus are organized in an operon-like structure, with several genes either containing short intergenic regions or overlapping one another, suggesting that they are co-transcribed. RT-PCR analysis was performed across the intergenic junctions from chtD to chtC3, including the adjacent genes and the downstream gene (as negative controls). The absence of a transcript for the chtR and chtD junction indicated that these genes were not cotranscribed, as expected since these genes are divergently orientated (Fig 2C). RT-PCR of the gene junctions from chtD to chtC3 produced gene transcripts of the expected sizes (Fig 2C), whereas no product was observed for the gene junction between chtC3 and the gene immediately downstream (data not shown). These results indicate that the cht locus of JIR325 consisted of eight genes, from chtD to chtC3, and that these genes were co-transcribed.
Sequence analysis of the cht locus promoter region also indicated the presence of a putative Fur binding site, or Fur box (5'-CATTATTGATAATCATTTTCA-3'), located 33 bp upstream of the chtD start codon (Fig 2A). In many bacterial species, Fur acts as a transcriptional repressor of iron acquisition genes by controlling the promoter activity of a gene in response to iron availability [45,46].

ChtD, ChtE and Srt are expressed under iron-limiting conditions
The presence of a putative Fur box in the isd promoter region suggested that these genes may be regulated in an iron-dependent manner. Therefore, the production of proteins encoded by the cht locus was examined during growth under normal and iron-limited conditions in TSB. Western blot analysis, using specific polyclonal antibodies raised against ChtD, ChtE and Srt, detected immunoreactive bands approximately 77 kDa, 25 kDa and 20 kDa in size, corresponding to ChtD, ChtE and Srt, respectively. These bands were detected in cell lysates of JIR325 grown under iron-limited conditions (TSB + 100 μM 2,2-dipyridyl), but not from cells grown under normal conditions (TSB) (Fig 3A). These findings provide evidence that the cht locus of JIR325 is up-regulated under iron-limited conditions. RT-PCR analysis of RNA isolated from JIR325 cells grown under similar conditions confirmed that the chtD, chtE and srt genes were transcribed under iron-limited conditions, but not under normal growth conditions ( Fig 3B). These results indicate that the cht locus of JIR325 is up-regulated under iron-limiting conditions and that this iron-dependent regulation occurs at the transcriptional level.

ChtD and ChtE bind heme
To determine the role of the JIR325 cht locus in heme acquisition, the ability of the NEAT proteins ChtD and ChtE to bind heme was determined. The Srt protein was used as a negative control as it was not predicted to bind heme. The genes encoding ChtD (residues 31 to 711), ChtE (residues 31 to 191) or Srt (residues 29 to 187) were cloned into a pET28a(+) expression vector to generate N-terminal His 6 -tagged fusion proteins. The cloned region included the DNA sequence that encoded the NEAT domains (for ChtD and ChtE), but not the N-terminal signal peptide and/or transmembrane domain, to obtain soluble proteins. The cell pellets of E. coli BL21(DE3) expressing His-ChtD or His-ChtE were red in color, unlike the pellets containing the His-Srt protein, which were cream in color (Fig 4A). These findings suggest that His-ChtD and His-ChtE were able to sequester heme during their expression in E. coli, as reported in several studies [47,48]. Purification of the His-tagged fusion proteins by affinity chromatography and SDS-PAGE showed the expected sizes for His-ChtD, His-ChtE and the control protein His-Srt, at 77 kDa, 25 kDa and 20 kDa, respectively (Fig 4B).
To assay for bound heme, the absorbance spectrum from 250 to 650 nm was analyzed for each recombinant protein. An absorbance peak at~400 nm (Soret band) is a characteristic feature that is indicative of heme bound to protein [48,49]. The absorbance spectra of the His-ChtD and His-ChtE proteins revealed a Soret band at~400 nm (Fig 5A), which indicated that His-ChtD and His-ChtE had co-purified with heme. In contrast, a hemin only preparation produced a broad absorbance maximum between 350 to 400 nm [49]. The His-Srt protein, which was not expected to bind heme, did not have a Soret peak at~400 nm. These observations also indicated that the ability of His-ChtD and His-ChtE to bind heme was not due to the non-specific binding of the His 6 -tag to heme.
To confirm the co-purification of endogenous heme with His-ChtD and His-ChtE, the purified proteins were subjected to LC-MS analysis. The Tris buffer used for the purification of the recombinant proteins, and a hemin preparation, were also analyzed as negative and positive controls, respectively. Samples containing the heme species were predicted to produce a peak at 616 m/z, which is representative of a single charged heme molecule [50]. LC-MS analysis of His-ChtD and His-ChtE samples revealed a peak at 616 m/z, which was consistent with that observed for hemin (Fig 5B). These results indicate that heme was present in the His-ChtD and His-ChtE preparations. The His-Srt sample did not produce an equivalent peak, which again was consistent with the protein not containing bound heme. Additionally, the absence of a 616 m/z peak in the protein buffer alone indicated that the 616 m/z peak was only associated with a heme molecule. A 617.0 m/z peak was observed in all of the samples including the protein iron-depleted (TSB + 150 μM 2,2-dipyridyl, TSB + Dp) conditions. The chtD (576 bp), chtE (230 bp) and srt (276 bp) gene regions were PCR amplified using the appropriate primer pairs (S1 Table). DNA standard sizes (kb) are stated on the right of the gel. Lanes: RT-with reverse transcriptase (cDNA); NRT-no reverse transcriptase (RNA); NC-no template control (dH 2 0); DNA-JIR325 genomic DNA; L-PCR markers (Promega).   buffer, which did not contain heme, suggesting that this peak is non-specific and is not indicative of heme. Taken together, these observations provide evidence that ChtD and ChtE are able to sequester heme. chtDE and srt mutants are not defective in hemoglobin and ferric chloride utilization To examine the functional role of the C. perfringens JIR325 cht locus in heme acquisition, allelic exchange was used to construct an chtDE deletion mutant (JIR12755) by replacing the chtD and chtE genes with a chloramphenicol resistance cassette, cat(P). In addition, an srt mutant (JIR12614) was constructed by insertional inactivation using Targetron technology, which introduced a group II intron with an erm(B) erythromycin resistance gene into the srt gene. PCR analysis and Southern hybridization confirmed the deletion of the chtDE genes and the insertional inactivation of the srt gene in the respective mutants (data not shown).
To determine whether the cht locus was involved in the acquisition of iron from Hb, the in vitro growth of the wild-type and the isogenic chtDE and srt mutants was compared under normal, iron-depleted and iron-replete conditions containing human Hb (metHb or HbO 2 ) or FeCl 3 as the sole iron source. The growth of the chtDE and srt mutants was similar to the wildtype strain under normal conditions (Fig 6), while no growth was observed for all strains grown under iron-depleted conditions, as previously observed for the wild-type strain (Fig 1). Under iron-replete conditions, the chtDE ( Fig 6A) and srt (Fig 6B) mutants were still able to utilize FeCl 3 or human Hb (metHb or HbO 2 ) at wild-type levels. This observation indicated that under the conditions tested, the cht locus was not essential for Hb or FeCl 3 utilization in JIR325.

The inactivation of srt does not affect the cell wall-anchoring of ChtD and ChtE
The cellular localization of ChtD, ChtE and Srt was examined by separating the whole cell lysates of JIR325 into cytoplasmic and cell envelope fractions. Western blot analysis using polyclonal antibodies to ChtD, ChtE or Srt, indicated the presence of ChtD in both the cell envelope and cytoplasmic fractions (Fig 7). In contrast, the ChtE and Srt proteins were associated only with the cell envelope fraction. The presence of ChtD, ChtE and Srt in the cell envelope fraction was consistent with the predicted role of ChtD and ChtE as cell surface proteins, and the role of the sortase as a membrane-anchored enzyme.
To determine whether Srt was involved in the anchoring of ChtD and ChtE to the cell wall, the cellular localization of these proteins was analyzed in the srt mutant as before. The results showed that the localization of both ChtD and ChtE was unaltered in the srt mutant (Fig 7). Immunoreactive species were not observed in any of the cellular fractions from the srt mutant when they were probed with the Srt polyclonal antibody, confirming that Srt was not produced in the mutant strain. These results suggested that this Srt protein was not required to localize ChtD and ChtE to the cell wall of C. perfringens JIR325, under the conditions tested.

Virulence studies of the isogenic chtDE and srt mutants in a mouse myonecrosis model
To investigate the contribution of the cht locus to virulence, cells of either the wild-type strain or the isogenic mutants were injected into the right thigh muscle of BALB/c mice, using our established mouse myonecrosis model [40]. Mice infected with the chtDE or srt mutants showed reduced disease in most of the disease parameters tested (data not shown) and had an increased mean survival time (chtDE mutant, 11.6 ± 0.3 h; srt mutant 10.4 ± 0.6 h) compared to those infected with the wild-type strain, which progressively succumbed to disease from 2 to 3 h post-infection (mean survival time, 5.6 ± 0.5 h) (Mann Whitney U test, p < 0.0001) ( Fig  8A and 8B). In an attempt to validate these results, the chtDE and srt mutants were complemented in cis and the isogenic complemented strains were virulence tested using the same model. The chtDE and srt complemented strains (JIR12773 and JIR12861, respectively) were however, also attenuated for virulence (p < 0.0001). Viable cell counts were performed preand post-infection and showed that the bacterial numbers recovered from mice infected with the mutants were comparable to those infected with the wild-type strain. These results suggested that the chtDE and srt mutations were not responsible for the reduced virulence observed with the mutants.

Whole genome sequencing of the chtDE and srt mutants
To determine if the inability to complement the chtDE and srt mutants was due to spontaneous secondary mutations, whole genome sequencing of the wild-type, chtDE, and srt strains was undertaken. Among the mutations identified in the in cis complemented chtDE mutant was a single nucleotide polymorphism (SNP) in feoA, which resulted in an A44D amino acid substitution ( Table 2). The feoA gene is upstream of feoB, which in other organisms, encodes a ferrous iron transporter and is involved in bacterial virulence [51,52]. Three mutations that were common to the chtDE and srt mutants were also identified ( Table 2). These mutations resulted in the deletion of amino acid A163 in a putative ferrodoxin hydrogenase (CPE0276), a frameshift in a hypothetical protein (CPE2000) and a D443N substitution in a putative BaeS sensor histidine kinase ( Table 2). The impact of these mutations on protein function, as well as in combination with the other mutations in the chtDE isogenic strains and the srt mutant remains to be determined.

Discussion
Most bacteria have developed diverse mechanisms to acquire iron from the host since this element is critical for bacterial growth and the subsequent development of an infection [4,14]. In this study, we showed that the in vitro growth of C. perfringens strain JIR325 was supported in media containing commercially sourced human Hb (predominantly metHb), freshly prepared HbO 2 or FeCl 3 as the sole iron source (Fig 1). The ability of C. perfringens to utilize Hb, albeit at high concentrations, is in agreement with the literature on numerous bacteria, including human pathogens such as S. aureus, B. anthracis and N. meningitidis, as Hb is the major reservoir of iron in the mammalian host [7,15,53]. C. perfringens type A produces cytotoxins such as α-toxin and perfringolysin O, which are able to lyse red blood cells [25]. The production of cytolytic toxins and/or hemolysins is suggested to contribute to the release of Hb and the subsequent acquisition of heme by pathogenic bacteria [54,55].
The use of freshly prepared human Hb in this study was important since lyophilized human Hb is typically oxidized [56] and can contain denatured protein and free heme, making it unsuitable as a sole iron source, to test for microbial growth [57]. Growth on freshly prepared HbO 2 could be attributed to uptake of Fe(II) or Fe(III) heme. Thus, whilst dissociation of O 2 typically regenerates Fe(II) heme, oxidation to Fe(III) heme is favored under low O 2 conditions [58]. In contrast, CO binds Hb~200 times more strongly than O 2 [59] conferring protection against heme oxidation and dissociation from Hb [60]. Failure of HbCO to support the growth of C. perfringens could be due to lower rates of heme dissociation from Hb, and/or the absence of hemophores required to utilize this form of Fe(II) heme. Notably, most hemophores show a strong preference for binding to Fe(III) over Fe(II) heme [61,62]. The ability to utilize specific forms of Hb has previously been noted. For example, a NEAT domain of the primary Hb receptor from S. aureus, IsdB, was shown to be capable of extracting heme from metHb but not from HbCO [63]. The utilization of FeCl 3 by C. perfringens is consistent with the presence of four putative siderophore-mediated iron uptake systems, which in many pathogenic bacteria, are used to acquire ferric iron from host compounds such as ferritin, transferrin or lactoferrin [64][65][66]. In contrast, the inability of human holo-transferrin to support the in vitro growth of C. perfringens could be due to a lack of receptors or siderophores for transferrin, as reported for S. pyogenes and S. pneumoniae [67,68]. Importantly, the growth of JIR325 under iron-replete conditions indicated that the wild-type strain was unable to grow under iron-depleted conditions due to the restricted availability of iron and not because the iron chelator 2,2-dipyridyl used in the growth assay was cytotoxic to the C. perfringens cells.
The presence of multiple mechanisms for heme/iron uptake by C. perfringens suggests that it may use these systems to adapt to changes in both the availability and type of iron sources within the host, as observed for P. aeruginosa and S. aureus [42,69]. High concentrations of Hb (25 μM and 50 μM HbO 2 ) were required to effectively support the growth of C. perfringens strain, JIR325, in contrast to that reported for S. aureus (10 nM HbO 2 ) [70], suggesting that Hb may not be the preferred iron source for the in vitro growth of this organism, under the conditions tested. Furthermore, recent studies have shown that FeoB from JIR325 appears to be the major protein required for iron acquisition in this strain, since a feoB mutant had an approximately 99% reduction in total iron content compared to the wild-type strain [27].
The genes encoding two NEAT proteins, a putative sortase and a putative membrane transport system were shown to be co-transcribed in JIR325. Similar genes in the Isd system of S. aureus and B. anthracis are also encoded in an operon, with the smaller B. anthracis isd operon predicted to encode the cell surface NEAT proteins [7]. Additionally, three of the cell surface proteins (IsdA, IsdB and IsdH) of S. aureus and the heme oxygenases of S. aureus and B. anthracis are encoded as single transcriptional units [22,71]. In C. perfringens, a hemO gene, located five ORFs upstream of chtD, has been shown to encode a heme oxygenase that mediates the degradation of heme to biliverin and iron in strain 13, the parent of JIR325 [72]. These findings suggest that the C. perfringens cht locus encodes a complete heme acquisition system, although the mechanism of heme uptake may differ to that of the well-characterized Isd system of S. aureus.
Transcriptional and Western blot analyses revealed that when JIR325 was grown under iron-limited conditions there was an increase in both the transcription of the chtD, chtE and srt genes, and the production of the respective proteins (Fig 3). These results provide evidence that the expression of the cht locus is regulated by the availability of iron and are consistent with the presence of a putative Fur box in the chtD promoter region of the locus, as has been observed in the Isd systems of S. aureus and L. monocytogenes [4,6]. The expression of iron uptake genes in many bacteria is regulated by the levels of extracellular iron present in the environment and is usually repressed, in the presence of excess iron, by the binding of a Fur protein to Fur boxes in the promoter region of regulated genes [8,45]. Whilst three genes have been annotated as potential Fur homologues in strain 13 [26], it is not known which of these genes encode a functional Fur protein Absorbance spectroscopy and LC-MS analysis indicated that the NEAT proteins, ChtD and ChtE, were able to bind Fe(III) heme. The mechanism of heme binding by these proteins is not known, but studies on the S. aureus and B. anthracis hemoprotein receptors have shown that individual NEAT domains mediate heme binding [49,73,74], which suggests that the NEAT domains of ChtD and ChtE have a similar functional role. Structural and biochemical studies on the NEAT domains of IsdA from S. aureus and IsdX2 from B. anthracis indicate that heme is coordinated by an invariant tyrosine residue located on the β7 strand of the protein [11,46]. A second invariant tyrosine side chain is located four residues away in the same β strand, and is hydrogen bonded to the coordinating tyrosine side chain. These tyrosine residues are conserved in the NEAT domains of many reported heme-binding proteins [13,74,75]. However, the ChtD and ChtE NEAT domains studied here lack these invariant tyrosine residues (S1 Fig), suggesting that heme coordination in these proteins may occur by a different mechanism. The Shp heme binding protein of S. pyogenes has a unique heme binding mechanism involving two methionine residues (Met-66 and Met-153), although the protein is folded into an immunoglobulin-like β-sandwich fold similar to other NEAT domains [76]. Heme is transferred from Shp to the lipoprotein component of the heme transporter HtsA in this organism, a process facilitated through the displacement of the Met-66 and Met-153 residues of Shp by the Met-79 and His-229 residues of HtsA, respectively [77].
C. perfringens may utilize a similar heme acquisition pathway since the NEAT domain of its ChtE protein has 25% identity to Shp and the putative periplasmic binding protein, ChtA, has 32% identity to HtsA from S. pyogenes. The equivalent Met or His residues in Shp and HtsA are also conserved in ChtE and ChtA (S1 Fig). Additionally, the last NEAT domain of ChtD has 27% identity to NEAT5 of IsdX2 in B. anthracis, which is known to be essential for scavenging heme from metHb [78]. In IsdX2, this process requires an amino acid with an amine side chain within the 3 10 helix of the NEAT5 domain [78]. The presence of an arginine in ChtD at a similar position is consistent with a similar mode of interaction between ChtD and Hb, however, further studies are required to examine this hypothesis. In addition to heme binding, individual NEAT domains can also mediate the extraction of heme from Hb and the transfer of heme between NEAT domains [49]. Therefore, the presence of multiple NEAT domains in ChtD suggests that it may be able to perform multiple roles in heme acquisition, potentially to increase the efficiency of heme uptake.
The ChtD, ChtE and Srt proteins were localized to the cell envelope of JIR325, where ChtD and ChtE were found to contain potential sorting motifs (LPQTG and NKESS, respectively) at their C-terminus. This association coupled with their heme binding ability, suggests that these proteins act as receptors for hemoprotein binding and may be processed for cell wall anchoring by the membrane-anchored Srt protein. A recent biochemical study of the Srt protein of C. perfringens strain 13 indicated that it preferentially recognised the LPQTG motif [44]. Additionally, in S. aureus, B. anthracis and L. monocytogenes, the isd locus encodes a sortase B protein, which is iron-regulated and anchors a subset of cell surface proteins bearing the NPQTN, NPKTH and NAKTN (or NPKSS) sorting motifs, respectively [14,22,79,80]. Mutation of the srtB gene in L. monocytogenes resulted in the absence of a hemin binding protein, SvpA, in the peptidoglycan layer, however, a lower molecular weight SvpA species was observed in the supernatant [81]. In our system, the inactivation of srt did not affect the cell envelope localization of ChtD and ChtE in JIR325. The efficient activity of Srt has been suggested to be dependent on temperature and may require a substrate-induced conformational change for binding to LPQTG motifs [44]. Although similar conditions were used to assess the activity of the C. perfringens Srt protein, a lack of phenotype in the srt mutant suggests that ChtD may require different conditions in order to be properly anchored to the cell wall. Alternatively, ChtD and/or ChtE may be anchored to the cell wall by other putative sortases encoded by JIR325, or tethered to the cell wall by their C-terminal hydrophobic regions and short charged tails as suggested for the cell surface heme binding proteins Shp and Shr of S. pyogenes and HtaA of Corynebacterium diphtheriae [9,48,82]. Therefore, the functional role of the sortase encoded by the cht locus of JIR325 remains unknown.
Several studies have reported that the Isd system is required for the virulence of S. aureus [15,16], although this result may be dependent on the specific conditions used in the animal infection model [83]. Studies in a mouse myonecrosis model showed that the C. perfringens chtDE and srt mutants were attenuated for virulence, although the virulence defects were not restored by complementation with the respective wild-type genes (Fig 7). The reduced virulence of these mutants may be the result of the specific mutation in the Cht system, the presence of SNPs in potential virulence-associated genes, as detected by genomic sequence analysis, or a combination of these mutations. SNPs were detected in genes encoding proteins such as FeoA, an ABC transporter and/or a phosphoenolpyruvate-phosphotransferase protein, all of which have been shown to be involved in the virulence of other bacteria [70,[84][85][86][87][88]. These secondary mutations had no effect on the growth of the chtDE or srt mutants in the presence of human Hb or FeCl 3 , but they appear to have contributed to the attenuated virulence phenotype observed in the animal model.
Growth studies showed that the chtDE or srt mutants were able to utilize the commercially prepared human Hb which was most likely in the form of metHb, freshly prepared human HbO 2 and FeCl 3 as the sole iron source at similar levels to the wild-type (Fig 6). These results indicate that heme acquisition by the Cht system might be required but is not essential for the growth of JIR325 under these conditions. This result is supported by studies in S. aureus and L. monocytogenes where the deletion of several isd genes also did not produce significant growth defects in the presence of commercial heme or Hb products as the sole iron source [83,89,90]. However, the results of these studies are in contrast to those which showed that several isd genes were essential for heme utilization by S. aureus when freshly prepared human Hb was provided as the sole iron source [57,70].
Functional redundancies between iron acquisition systems have been reported in several organisms, where strains containing mutations in two different iron transport systems showed significant defects in utilizing heme/iron for its growth as compared to single mutants of an iron transport system [64,91]. C. perfringens JIR325 contains seven putative iron uptake systems [27], thus deficiencies in Hb utilization of the cht mutants produced in this study may have been compensated by one or more of the other putative heme/iron acquisition systems found within this strain.
In summary, this study has demonstrated that the C. perfringens cht locus is expressed under iron-limited conditions and that the NEAT proteins, ChtD and ChtE can bind heme. The chtD, chtE and srt genes of the cht locus are most likely required but are not essential for FeCl 3 or Hb utilization, and similarly do not appear to be essential for the growth and virulence of C. perfringens strain JIR325. Human Hb does not appear to be the major source of iron for C. perfringens strain JIR325, and as shown in other systems, one or more of the other putative heme/iron acquisition systems may have compensated for the deficiencies in Hb utilization of the cht mutants produced in this study.