|Home | About | Journals | Submit | Contact Us | Français|
Francisella species secrete a polycarboxylate siderophore that resembles rhizoferrin to acquire ferric iron. Several of the Francisella siderophore synthesis genes are contained in a Fur-regulated operon (designated fig or fsl) comprised of at least seven open reading frames (ORFs) including fur. Reverse transcriptase-PCR showed transcriptional linkage between figD and figE and between figE and figF. Mutations were constructed in four of these ORFs (figB, figC, figD, and figE) in F. novicida U112. All four of these new mutants and a F. novicida figA mutant grew at rates comparable to that of wild-type under iron-replete conditions but growth of all five mutants was stunted in iron-limiting media. When ferric rhizoferrin was added to the iron-limited media, growth of the figA, figB, figC, and figD mutants was restored to levels similar to those obtained in iron-replete media. However, this exogenously added siderophore could not rescue the figE mutant. When Chrome Azurol S assays were used to measure siderophore production, the figA, figB, and figC mutants were markedly deficient in their ability to synthesize siderophore whereas the figD and figE mutants produced siderophore at levels equivalent to the wild-type parent strain.
Interest in the metabolism and physiology of pathogenic Francisella species has been stimulated by the concern that F. tularensis subspecies tularensis (type A) and F. tularensis subspecies holarctica (type B) might be used as bioweapons (Ellis et al., 2002, Oyston et al., 2004). Both of these Francisella strains can cause tularemia in both humans and animals (Oyston, Sjostedt, A., and Titball, R. W. 2004, Petersen et al., 2005, Sjostedt 2005). To date, most studies of Francisella species have involved either the attenuated live vaccine strain (LVS) of F. tularensis (Saslaw et al., 1961a, Saslaw et al., 1961b), or F. novicida which is pathogenic for mice but essentially avirulent in humans (Oyston, Sjostedt, A., and Titball, R. W. 2004). F. novicida can be readily manipulated with respect to mutant construction (Anthony et al., 1991, Deng et al., 2006, Lauriano et al., 2003).
The data available concerning iron acquisition systems used by Francisella species are limited at present. Early studies showed that growth of F. tularensis was improved by some iron salts (Tresselt et al., 1964), and by two ferric hydroxamate siderophores (Halmann et al., 1967). This hint that F. tularensis might produce a siderophore was recently confirmed by studies from two laboratories (Deng et al., 2006, Sullivan et al., 2006). Our laboratory identified a gene cluster in F. tularensis LVS which was shown to be essential for production of a siderophore by this bacterium (Deng et al., 2006). These four genes, designated Francisella iron-regulation gene (fig) A, B, C, and D, were found to be located in an operon downstream from the fur gene and were shown to be transcriptionally linked to fur (Deng et al., 2006). These same genes have been designated as fslABCD by another laboratory (Ramakrishnan 2006, Sullivan et al., 2006). Mutant analysis showed (1) that inactivation of the figA gene resulted in the lack of production of the siderophore activity and (2) that this siderophore activity could enhance the growth of a figA mutant under iron-limited conditions (Deng et al., 2006). Other analyses (Sullivan et al., 2006) showed that this siderophore was a polycarboxylate molecule very similar if not identical to rhizoferrin, a polycarboxylate siderophore produced by Rhizopus species (Carrano et al., 1996). The outer membrane receptor for this rhizoferrin-like siderophore has not been identified to date, although at least one gene (designated figE) encoding a potential outer membrane protein receptor has been identified immediately downstream from the figABCD gene cluster [together with another gene (figF) encoding a protein of undetermined function] (Fig. 1) (Ramakrishnan et al., 2007). The figE gene was reported to be co-transcribed with figABCD (Milne et al., 2007).
In the present report, we describe the construction and phenotypic characterization of a set of F. novicida fig operon mutants. Reverse-transcriptase PCR was used to establish transcriptional linkage between the figD and figE ORFs and between figE and figF. Mutants unable to express FigA, FigB, FigC, FigD, or FigE individually were severely limited in their ability to grow under conditions of iron restriction, and only the figD and figE mutants produced levels of the Francisella siderophore equivalent to those produced by the wild-type parent strain.
F. tularensis subsp. novicida U112 (Larson et al., 1955) was used as the wild-type strain and will be referred to as F. novicida in this report. The F. novicida figA mutant used in this study has been described previously (Deng et al., 2006). F. novicida strains were grown on solidified media at 37° C in an atmosphere of 95% air-5% CO2 or in broth at 37° C. F. novicida strains were grown in two different types of media, each of which varied with regard to the availability of iron. Iron-replete medium (MH+), iron-limited medium (MH−), and iron-restricted medium (MH−/DF) have been described previously (Deng et al., 2006). For detection of siderophore production, F. novicida strains were grown in Chamberlain’s defined medium (Chamberlain 1965) with 0.002 g L −1 FeSO4 (CDM+) or without FeSO4 (CDM−). For construction of the F. novicida mutants, the growth medium was supplemented with kanamycin (15 μg mL−1). Escherichia coli TOP10 cells (Invitrogen) were used for cloning. The E. coli plasmid vector pACYC177 (New England Biolabs) was modified by inactivating the kanamycin resistance gene. The modified vector, designated pWL100, was used for all subsequent cloning efforts in this study.
Chromosomal DNA from MH+ plate-grown F. novicida cells was isolated using an UltraClean Microbial DNA Isolation Kit (Mo-Bio Laboratories). Plasmid DNA from E. coli was isolated using the Miniprep Spin Kit (Qiagen). RNA was isolated from F. novicida strains [grown in MH− or MH−/DF broth into the logarithmic phase (OD600= 0.6–0.7)] using the RNeasy Midi Kit (Qiagen).
All of the oligonucleotide primers used in this study are listed in Table 1. The same general strategy was used to construct the figB, figC, figD, and figE deletion mutants. Overlapping extension PCR (Ho et al., 1989) was used to replace most of each ORF (figB, figC, figD) with a promoterless kanamycin resistance cassette (np-kan) (Menard et al., 1993). For example, to obtain the figB mutant, a ~0.9-kb fragment corresponding to the chromosomal region 5′ from the intended deletion was PCR-amplified from F. novicida chromosomal DNA with primers B1 and B2. A second ~0.9-kb fragment corresponding to the chromosomal region 3′ from the intended deletion was amplified with primers B3 and B4. These two PCR products were then used together as the template for a second round of PCR using primers B1 and B4. The PCR-derived amplicon was then digested with BamHI and ligated into BamHI- digested pWL100 to obtain pWL100B. The np-kan cassette was PCR-amplified with primers np-kan-5′ and np-kan-3′ from pUC18K3. This fragment was digested with KpnI and ligated into KpnI-digested pWL100B to obtain pWL100B-Kan. After verification of the nucleotide sequence of the insert, pWL100B-Kan was transformed into E. coli strain HB101. Plasmid DNA extracted from E. coli HB101 was used to electroporate F. novicida cells as described (Deng et al., 2006) and transformants were selected for kanamycin resistance. The deletion in the target gene was confirmed by nucleotide sequence analysis in one transformant which was selected as the figB mutant for this study. The figC and figD mutants were constructed similarly.
A slight modification to the method described above was necessary to construct the figE mutant. A DNA segment containing the promoter region from a kanamycin resistance cartridge was inserted immediately upstream of the np-kan cassette because of difficulty in obtaining expression of kanamycin resistance from the np-kan cassette in the figE mutant. Site-directed mutagenesis was used to create a unique KpnI site immediately in front of the np-kan cassette in plasmid pWL100E-Kan. A fragment corresponding to the promoter region of the kanamycin resistance gene (Pharmacia) was PCR-amplified with primers pk-5′ and pk-3′, digested with KpnI, and ligated into KpnI-digested pWL100E-Kan to obtain pWL100E-KanC.
F. novicida chromosomal DNA was used as the template for all PCR amplifications in the following reactions. A 390-bp fragment immediately 5′ from the figA ORF was PCR-amplified using oligonucleotide primers comp1 and comp2 (Table 1). This amplicon was designated as the figA promoter. The DNA fragments containing the figC, figD, and figE ORFs were PCR-amplified with primers comp3 and comp4 (for figC), comp5 and comp6 (for figD), and comp7 and comp8 (for figE). The figA promoter fragment was attached to the 5′ end of each of these amplicons by overlapping extension PCR (Ho et al., 1989) and the resultant construct was cloned into the KpnI site of the shuttle vector pFNLTP-CAT (Deng et al., 2006). The resultant recombinant plasmids were designated as pWL101 (for figC), pWL102 (for figD) and pWL103 (for figE). Each of these three plasmids was electroporated into its respective mutant (i.e., the figC mutant, the figD mutant, and the figE mutant) as described (Deng at al., 2006). Plasmid pFNLTP-CAT was electroporated into each of these mutants as a control.
RNA was isolated from the F. novicida strain grown in MH− and MH−/DF media as described above and was further digested with DNase I (GenHunter Corp) to remove any DNA contamination. RT-PCR was performed with the SuperScript III kit (Invitrogen) according to the manufacturer’s protocol. Primers binding to regions within different ORFs were designed to amplify the regions between figA and figB, between figD and figE, and between figE and figF. RNA (3 μg) was incubated with SuperScript III reverse transcriptase and appropriate primers. A positive control reaction using F. novicida chromosomal DNA was included to assure proper primer function. The negative control used RNA as the template but did not contain reverse transcriptase. PCR amplification with Taq DNA polymerase (New England Biolabs) was carried out at 52°C for annealing and 72°C for extension.
The CAS agar assay described by Schwyn and Neilands (Schwyn et al., 1987) was modified here for use with CDM- agar. CAS-HDTMA solution with iron was added to CDM- agar as described to prepare CAS-CDM- medium (Deng et al., 2006). To measure siderophore secretion as evidenced by the production of an orange halo, the F. novicida wild-type strain was grown overnight on MH− agar and the figA, figB, figC, figD, and figE mutants were grown overnight on MH− agar containing kanamycin (12 μg mL−1). The complemented figC and figD mutants were grown overnight on MH- agar containing chloramphenicol (10 μg mL−1). One colony of each wild-type or mutant was picked from the MH− agar plate and stabbed into the CAS-CDM- plate. Incubation was carried out at 37° C in an atmosphere of 95% air-5% CO2 for 24 hrs at which time the diameter of each orange halo was measured.
Ferric rhizoferrin (purified from Rhizopus (oryzae) arrhizus) was purchased from Biophore Research Products, EMC Microcollections GmbH, Tuebingen, Germany. F. novicida wild-type and mutant strains were grown in CDM+ broth, CDM− broth, and CDM− broth containing 10−6 M ferric rhizoferrin. The optical density (OD600) was followed for 15 hrs.
Previous work from this laboratory included the characterization of figA mutants of both F. novicida and F. tularensis LVS and demonstration that these mutants were deficient in siderophore production and that the fur gene was transcriptionally linked to the figABCD gene cluster (Fig. 1A) (Deng et al., 2006). To further address the involvement of the fig operon in siderophore biosynthesis, non-polar kan cartridges were inserted into the figB, figC, and figD ORFs in F. novicida (Fig. 1B). It should be noted here that comparison of the fig locus in three Francisella species revealed dissimilar figF ORFs (Fig. 1C). In F. novicida, two mutations relative to F. tularensis SCHU S4 resulted in a fusion between figF and a downstream ORF (Fig. 1C). Therefore, a F. novicida figF mutant was not constructed in this study.
Because of a report which indicated that the figD ORF was co-transcribed with figE (Milne et al., 2007), we analyzed the transcriptional linkage between figD and figE and between figE and figF. When F. novicida cells were grown in MH− medium where iron is limiting for growth, there was no detectable transcript between figD and figE (Fig. 2, lane 9). However, when the amount of available iron was further reduced by the addition of the iron chelator deferoxamine mesylate, a transcript that spanned the figD-figE intergenic region was readily detectable (Fig. 2, lane 11). Similarly, under these iron starvation conditions, there was a readily detectable transcript linking figE and figF (Fig. 2, lane 16). Interestingly, a modest figE-figF transcript was detected in cells grown in iron-limiting MH- medium (Fig. 2, lane 14).
The figA, figB, and figC mutants all produced less siderophore in vitro than did the wild-type strain (Fig. 3A and 3B). Complementation of the figA mutant with a wild-type figA gene was reported previously (Deng et al., 2006). We were unable to construct a figB mutant carrying the wild-type figB gene in trans, for reasons that remain unknown. Complementation of the figC mutant was accomplished in the present study (Fig. 3C). Siderophore production by the figD mutant was only slightly less than that of wild-type in one set of repeated experiments (Fig. 3B; p = 0.034) and equivalent to wild-type in another set of repeated experiments (Fig. 3D; p = 0.587). Because this mutant made essentially wild-type levels of the siderophore, the provision of the figD gene in trans had little effect on siderophore production by the figD mutant (Fig. 3D). The figE mutant did not produce less siderophore than the wild-type strain (Fig. 3A and 3B). When inoculated into the iron-limited CDM- medium, the growth of the wild-type strain and all five mutants was severely limited (Fig. 4A and Fig. 4B). However, when ferric rhizoferrin was included in this medium, it restored the growth of the wild-type strain and that of the figA, figB, figC, and figD mutants to wild-type levels (Fig. 4A). In contrast, inclusion of ferric rhizoferrin in the CDM- medium had little or no effect on the very poor growth of the figE mutant (Fig. 4B). However, when the figE gene was provided in trans, ferric rhizoferrin was able to restore growth of the figE mutant (Fig. 4B).
Iron acquisition by F. tularensis appears to be important in at least two major ways. First, iron is essential for the growth and survival of the organism in vivo. Recent studies from two groups using random transposon mutagenesis of either F. tularensis LVS (Su et al., 2007) or F. novicida (Weiss et al., 2007) showed that insertions into the figA, figB, figC, or feoB genes caused reduced virulence of these mutants. These data strongly suggest that expression of these particular gene products is essential for full virulence of Francisella species. Second, we have obtained evidence that the expression of certain F. tularensis virulence genes is regulated by iron availability. More specifically, we have shown that genes located in the F. tularensis pathogenicity island involved in intracellular growth (Nano et al., 2004, Santic et al., 2005), including pdpA, pdpB, pdpC, iglA, iglB, iglC, and iglD, were all up-regulated at least two-fold under iron-restricted conditions (Deng et al., 2006). Up-regulation of some of these same gene products as a result of iron limitation in vitro was subsequently confirmed by independent proteomics analysis (Lenco et al., 2007).
In the present study, we showed that F. novicida mutants unable to express the figA, figB, or figC genes had a reduced ability to produce the Francisella siderophore (Fig. 3). Immediately downstream from the figABCD cluster is the figE-figF two-gene cluster (Fig. 1). It was reported that figE gene was up-regulated by iron limitation in F. novicida (Milne et al., 2007) and our own DNA microarray analyses with F. tularensis LVS indicated that expression of at least figE was increased 1.5-fold under conditions of iron limitation in vitro (Deng et al., 2006). The FigE protein possesses a predicted signal peptide and coiled-coil domains, and has been proposed to be member of a new protein family (Larsson et al., 2005). We and others (Milne et al., 2007) suspected that this protein could be a receptor for the Francisella siderophore, and the ferric rhizoferrin feeding experiments contained in the present study (Fig. 4B), together with the apparent lack of effect of a figE mutation on siderophore production (Fig. 3), reinforce this possibility. The FigE proteins (listed as both locus FTT0025c and SrfA) of both F. tularensis SCHU S4 and F. tularensis LVS have been localized to the outer membrane of these organisms (Huntley et al., 2007), and the fact that the predicted F. novicida FigE protein has at least 97% identity with these two proteins makes it likely that FigE is also present in the outer membrane of F. novicida.
The expression of the figD-figE intergenic region was not detected in cells grown under conditions of modest iron deprivation (Fig. 2, lane 9) but was readily apparent in cells grown under more stringent iron limitation (Fig. 2, lane 11). These data are reflected by our previous DNA microarray-based study (Deng et al., 2006) in which expression of the figABCD gene cluster was the highest of all of the genes that were up-regulated by iron limitation in F. tularensis LVS whereas expression of at least figE was only modestly increased [see Supplemental Data in (Deng et al., 2006)]. The presence of a predicted fur box between figD and figE (Fig. 1A) may have some effect on transcriptional control of the figE-figF transcript which appears to be more tightly regulated by iron availability.
This study was supported by U.S. Public Health Service grant PO1 AI55637. The authors thank Shelley Payne for very helpful advice regarding methods for the measurement of siderophore production.