Iron-dependent gene expression in Actinomyces oris.

BACKGROUND: Actinomyces oris is a Gram-positive bacterium that has been associated with healthy and diseased sites in the human oral cavity. Most pathogenic bacteria require iron to survive, and in order to acquire iron in the relatively iron-scarce oral cavity A. oris has been shown to produce iron-binding molecules known as siderophores. The genes encoding these siderophores and transporters are thought to be regulated by the amount of iron in the growth medium and by the metal-dependent repressor, AmdR, which we showed previously binds to the promoter of proposed iron-regulated genes.
OBJECTIVE: The purpose of this study was to characterize siderophore and associated iron transport systems in A. oris.
DESIGN: We examined gene expression of the putative iron transport genes fetA and sidD in response to low- and high-iron environments. One of these genes, sidD, encoding a putative Fe ABC transporter protein, was insertionally inactivated and was examined for causing growth defects. To gain a further understanding of the role of iron metabolism in oral diseases, clinical isolates of Actinomyces spp. were examined for the presence of the gene encoding AmdR, a proposed global regulator of iron-dependent gene expression in A. oris.
RESULTS: When A. oris was grown under iron-limiting conditions, the genes encoding iron/siderophore transporters fetA and sidD showed increased expression. One of these genes (sidD) was mutated, and the sidD::Km strain exhibited a 50% reduction in growth in late log and stationary phase cells in media that contained iron. This growth defect was restored when the sidD gene was provided in a complemented strain. We were able to isolate the AmdR-encoding gene in seven clinical isolates of Actinomyces. When these protein sequences were aligned to the laboratory strain, there was a high degree of sequence similarity.
CONCLUSIONS: The growth of the sidD::Km mutant in iron-replete medium mirrored the growth of the wild-type strain grown in iron-limiting medium, suggesting that the sidD::Km mutant was compromised in iron uptake. The known iron regulator AmdR is well conserved in clinical isolates of A. oris. This work provides additional insight into iron metabolism in this important oral microbe.

Iron is essential for most pathogenic bacteria to survive and persist. In the human host, most iron is scavenged by host-binding proteins such as lactoferrin. In order to acquire iron, bacteria possess iron transporters as well as iron-scavenging molecules known as siderophores. When iron is low, organisms produce siderophores that solubilize the iron. Bacteria then transport the iron–siderophore complex into the cell by transport proteins such as ATP-binding cassette (ABC) transporters (1). However, too much iron can lead to damaging oxidative stress. Therefore, the concentration of iron within the bacterium is tightly regulated. Both Gram-positive and Gram-negative bacteria possess transcriptional regulators that control expression of the genes involved in iron uptake. Gram-positive organisms, which make up a significant proportion of oral bacteria, possess regulators similar to the diphtheria toxin repressor (DtxR). The DtxR protein in Corynebacterium diphtheriae is a global iron-dependent repressor that regulates transcription from multiple promoters including the phage-encoding tox gene and at least seven other genes (2). In C. diphtheriae and other pathogens, low iron modulates the expression of surface carbohydrates, adherence to human red blood cells, and slime production (3–5).

Actinomyces oris (formerly A. naeslundii) belongs to the genus Actinomyces, which includes several species of Gram-positive, facultatively anaerobic, rod-shaped bacteria. A. oris has been referred to as a pioneer of the oral cavity, because this microbe is one of the first organisms to colonize the mouth after birth and is among the first to recolonize newly cleaned tooth surfaces (6, 7). A. oris is a numerically dominant member of the oral flora throughout life and is found in significant numbers in dental plaque, as well as on mucosal surfaces in adults (8). These organisms have been found at both healthy and diseased sites in the mouth (9). For example, A. oris peptidoglycan has been shown to contribute to bone loss in periodontal disease (10), and Actinomyces species have been isolated from bisphosphonate-related osteonecrosis of the jaw (11–13).

Despite all of these roles in important human illnesses, there is comparatively little known about A. oris due to the difficulties of conducting molecular studies with this organism. Our laboratory has shown that A. oris possesses a transcriptional repressor AmdR, which appears to mediate the metal-dependent gene expression of a broad number of genes, including those coding for known or suspected virulence factors (14). Furthermore, we have shown that A. oris produces siderophores when cells are grown in low-iron medium and that siderophore production decreased when the iron was added to the medium (14). It is likely that the siderophores produced by Actinomyces contribute to the ecology of the plaque biofilm. D'onofrio et al. (15) have shown that siderophores are important molecules in complex microbial communities and, furthermore, siderophores from one organism may be used by another microbe. The siderophores produced by A. oris may play an important role in the survival of other organisms in the dental plaque that lack siderophores, such as Streptococcus mutans.

The acquisition of metal plays a major role in microbial virulence and disease process. Our hypothesis is that the low-metal-ion environment in the human oral cavity is used as a signal to induce expression of A. oris genes involved in metal ion acquisition and transport. The purpose of our study was to characterize metal transport in A. oris.

Materials and methods

Bacterial strains, plasmids, and growth conditions

The bacterial strain A. oris MG1 and 10 Actinomyces clinical isolates (from Dr. Robert Burne at the University of Florida) were maintained on brain–heart infusion agar at 5% CO2 at 37oC. The chemically defined Actinomyces medium (ADM) was used to cultivate A. oris (16). This medium was a modified version of the ‘HP 6/A’ medium omitting iron, hemin, and manganese to allow for low-metal conditions (16). The base ADM medium was shown to have ~8 µM iron as assessed by inductively coupled plasma mass spectrometry (ICP-MS), as described below. Ferric chloride was used to supplement the medium to allow for a high-iron condition (100 µm to 1 mM). In addition, to achieve iron-limiting conditions the iron-binding compound dipyridyl (DP) was added to the growth medium at 75 and 150 µm; given that there is 8 µM iron in the base medium, this concentration of DP is sufficient to scavenge the free iron in the medium.

The Escherichia coli strain DH5α from New England Biolabs was maintained on lysogeny broth (LB) solidified with 1% agar as appropriate. The medium was supplemented with ampicillin (ApR; 100 µg mL−1), kanamycin (KmR; 50 µg mL−1), or streptomycin (SmR; 50 µg mL−1), as indicated. For growth curves, bacteria were grown in 96-well plates and optical density (OD) at 630 nm was recorded using a spectrophotometer.

RNA isolation and qRT-PCR studies

Total RNA was isolated from mid-log stage A. oris and an A. oris sidD::Km mutant using the RNA mini prep kit (Qiagen, Hilden, Germany). RNA was quantified using Nanodrop lite (Thermo Scientific, Waltham, MA). Samples were treated with DNAse (Qiagen, Hilden, Germany). For qRT-PCR experiments, 2 µg of cDNA was synthesized from RNA using the cDNA kit from Applied Biosystems (Foster City, CA) according to the manufacturer's protocol. qRT-PCR was performed using Applied Biosystems Step One applying the recommended Taqman protocol. The experiments were conducted in triplicate. The cycle threshold (Ct) value was recorded and the relative quantification of the specific gene expression was calculated using the 2−ΔΔCt method, with the 16S rRNA as the reference gene. ANOVA analysis was conducted on the biological replicates to determine if there was a significant difference in the gene expression in the different iron concentrations, and the results were analyzed using Dunnett's multiple comparisons test.

Construction of a sidD::Km mutant strain of A. oris

An A. oris mutant containing an insertion in the sidD (anae_c_1_5343) gene was constructed as follows: Primers sidF (5′-CCGAAGAAGGGATCCGTGCCC-3′) and sidR (5′-GGGCAGCGCGGATCCCTACCGGC-3′) containing engineered BamHI sites (underlined) were used to amplify a 1 kb fragment containing the sid open reading frame (ORF). This ORF was subcloned into pGEM-7 (Promega, Madison, WI), which does not replicate in A. oris (17). A kanamycin-resistant cassette KmR with NheI generated ends was PCR amplified from pJRD215 (ATCC) with the primers kanF NheI (5′-TGGGCGTCGCTTGCTAGCTCATTTCGC-3′) and kanR NheI (5′-CCCTGCAGCCAAGCTAGCTTCACGC-3′). The KmR gene was introduced into a unique NheI site in the sidD ORF. The resulting plasmid (psidKmR) was electroporated into A. oris MGI as described previously (18). Chromosomal DNA was isolated from KmR colonies and was screened for the presence of the kanamycin-resistant marker via PCR. Primers KmF71114 (5′-CAGACAATCGGCTGCTCTGA-3′) and KmR71114 (5′-GATGTTTCGCTTGGTGGTCG-3′), which corresponded to the sequences of the KmR gene and sid ORF, were used for PCR confirmation of insertion of the Km cassette.

Generation of the complemented strain

The ORF of the sidD gene was synthesized by Genewiz (South Plainfield, NJ) with the promoter (corresponding to regions 1567709-1567928 on the chromosome) located upstream of the operon driving its expression. The promoter and sidD gene were isolated by restriction digest with BamHI and EcoRI from a plasmid carrying the sidD gene and promoter (pUC57, which was synthesized by Genewiz), and subcloned into BamHI and EcoRI restriction-digested pJRD215 (18). Plasmid pJRD215 replicates in A. oris. This resulting plasmid, psidD, was electroporated into the A. oris sidD::Km mutant strain and plated onto BHI supplemented with streptomycin (50 µg/ml) and kanamycin (50 µg/ml). Plasmid pJRD215 was also electroporated into the A. oris sidD::Km mutant to serve as the vector-only control. KmR and SmR colonies were selected, and plasmids were isolated to confirm complementation. The sidD::Km/pJRD215 and sidD::Km/psidD strains were used in the growth experiments.

Identification of the amdR gene in clinical isolates

Clinical isolates of Actinomyces were obtained from Dr. Robert Burne (University of Florida). 16S rRNA sequence analysis was used to confirm that the strains were Actinomyces species. These strains may be either A. naeslundii (now oris) genospecies 1 or genospecies 2. These two strains cannot be separated based on 16S rRNA analysis alone (19). Gram staining was used to confirm the morphology. To determine if metal-dependent repressors were conserved in the clinical isolates, primers corresponding to the amdR gene (5′-AGAAACACTCCTCAAACCCCC-3′ and 5′-ATCACAGGGTGACGCCTATC-3′) were used for PCR amplification from chromosomal DNA isolated from these clinical isolates. Chromosomal DNA was isolated from the Actinomyces clinical isolates using the Wizard Genomic Purification System (Promega). PCR products were subcloned into the pCR2.1 vector (Life Technologies, Carlsbad, CA) and sequenced at the Dartmouth Sequencing Core. Deduced amino acid sequences were compared using CLUSTAL alignment analysis.

Measurement of iron in the ADM medium

Iron concentration in the Actinomyces defined medium was measured by ICP-MS as described by Moreau-Marquis (20).

Results

The Actinomyces metal-dependent repressor possesses a metal-dependent DNA-binding domain

As a first step in investigating iron-dependent gene regulation in A. oris, we examined the published genome of A. oris MG1 for the presence of a metal-dependent regulator with sequence similarity to the well-characterized DtxR family of regulators in Gram-positive organisms. The ORF encoding Actinomyces metal-dependent repressor, AmdR, was identified. The deduced amino acid sequence of AmdR contained domains found in other metal-dependent repressors, including SloR of S. mutans and DtxR of Corynebacterium diphtheriae (Fig. 1). These domains include the DNA-binding helix-turn-helix domain and the iron-dependent repressor superfamily domain, and have been described for both SloR from S. mutans (21) and DtxR from C. diphtheriae (22). Unlike S. mutans SloR, the AmdR protein does not possess a ‘FeoA’ domain (Fig. 1); this short domain is found at the C-terminus of a variety of metal-dependent transcriptional regulators (23). A recent study suggests that the FeoA domain may mediate metal binding (21). AmdR has more sequence similarity and identity to DtxR (64%/52%) and shares less similarity and identity with SloR (47%/26%).

Fig. 1

AmdR shares conserved domains with other metal-dependent repressors. Predicted domains of the deduced amino acid sequence of Actinomyces metal-dependent repressor AmdR, S. mutans SloR, and C. diphtheriae DtxR from NCBI.

fetA and sidD show increased expression in iron-limited growth medium

Our laboratory has previously determined that Actinomyces metal-dependent repressor AmdR binds to the promoter regions upstream of the operon containing the sidD gene, which codes for a component of an iron transport system (14). These studies were conducted using purified AmdR in electrophoretic mobility shift assays. We hypothesized that a number of A. oris genes were likely subject to regulation by changing iron concentrations, due to binding of the AmdR to promoter sequences when abundant iron is present. We identified genes that were likely to be involved in iron acquisition by examining the genome of A. oris, which is available via the Bioinformatics Resources for Oral Pathogens (BROP). At the time of this publication, BROP refers to A. oris MG1 as A. naeslundii MG1.

As outlined here, sequence analyses revealed that there are over a dozen genes that may be involved in iron acquisition. Two of these genes identified, fetA and sidD, are located in operons with a high degree of sequence similarity (>98%) to transporters involved in iron uptake (Fig. 2). The sidC gene encodes a predicted 37 kDa, iron–siderophore ABC transporter substrate-binding protein, whose periplasmic component is 99% identical to iron–siderophore ABC transporter substrate-binding protein from A. viscosus. The sidD gene likely encodes the iron–siderophore ATP-binding protein of the ABC transporter component containing an ATP-binding site and the ABC transporter signature motif. The SidD protein is 95% identical to the iron dicitrate ABC transporter ATP-binding protein from A. viscosus (accession number WP_009406982.1; Fig. 2a). The first gene in the fetA (anae_c_1_8278) operon encodes a predicted 32 kDa protein with 98% identity and 99% similarity to a predicted A. oris high-affinity iron transporter (accession number WP_010612841.1; Fig. 2b) with 65% identity and 78% similarity to predicted ferrous iron permease EfeU from Propionibacterium acidipropionici ATCC 4875 (accession number=WP_015071892.1). The second gene in the operon fetB (anae_c_1_8283) encodes a 46.8 kDa protein with 99% similarity and 99% identity to a predicted peptidase M75 from A. oris. The peptidase M75 family includes an insulin-cleaving membrane protease, an iron-regulated protein A (IrpA) and an iron transporter EfeO-like alginate-binding protein (NCBI). The FetB protein also shows 67% identity and 75% similarity to a predicted PbrT family lead (Pb2+) uptake porter of Actinomyces urogenitalis S6-C4.

Fig. 2

Iron transport genes are regulated in response to iron. (a) Diagram of the ABC operon, the proposed regulatory region is underlined. sidA (anae_c_1_15012) and sidB (anae_c_1_15008), ABC transporter permease; sidC (anae_c_1_5341), ABC substrate-binding component; sidD (anae_c_1_5343), ABC ATP-binding component. (b) Diagram of the operon containing high-affinity iron transporter fetA (anae_c_1_8278) and a putative peptidase M75, fetB (anae_c_1_8283). (c) Shown is the relative quantity of transcript of the sidD gene in Actinomyces defined medium (ADM), ADM with 75 µm dipyridyl (DP; a strong iron chelator) or ADM plus 100 µM FeCl3 (FeCl3). (d) Relative quantity of transcript of the fetA gene in wild-type A. oris ADM (WT ADM), wild-type A. oris in ADM with 75 µm dipyridyl (WT DP), or A. oris in ADM plus 100 µM FeCl3 (WT FeCl3). Also shown is the relative quantity of transcript of the fetA gene in the sidD mutant (sidD::Km ADM), sidD mutant in ADM with 75 µm dipyridyl (sidD::Km DP), or sidD mutant in ADM plus 100 µM FeCl3 (sidD::Km FeCl3). ‘***’ indicates significant difference (p<0.05) in gene expression as determined by ANOVA and Dunnett's multiple comparisons test. ‘****’ indicates significant difference (p<0.02) in gene expression.

Using qRT-PCR analysis, we explored the regulation of fetA and sidD genes as representative genes in each of the operons described above to determine if their expression was indeed regulated by iron availability. To perform these studies, RNA was extracted from cells grown in minimal medium (ADM) or the same medium treated with the chemical iron-chelator dipyridyl (DP), or ADM supplemented with iron chloride.

The transcripts of both the fetA and sidD genes are increased when the wild-type A. oris is grown in metal-limited medium (Fig. 2c and d), which supports our hypothesis that these genes are regulated in response to iron limitation in the growth medium. Interestingly, addition of iron chloride did not significantly change the expression of these genes compared to the unamended medium control, suggesting the ADM is an iron-replete medium for this microbe. Consistent with this idea, we showed that ADM contains ~8 µM iron. In addition, we examined the sidD::Km mutant strain for both fetA and sidD gene expression. We were unable to detect a sidD transcript in the sidD::Km mutant strain, as expected (data not shown). However, the fetA transcript was detected, but in much lower levels than observed for the wild-type strain (Fig. 2d).

Generating an iron transport mutant of A. oris

To determine if the sidD locus contributes to iron transport in A. oris, we built an insertional loss-of-function mutant carrying a kanamycin resistance marker (see Materials and Methods for details). The kanamycin resistance insertion was verified using primers for the kanamycin marker and for the sidB gene, as described in Materials and Methods.

The growth of sidD::Km was compared to a wild-type strain on both iron-replete and iron-limited media. The final growth yield of sidD::Km was reduced when compared to the wild type in the iron-depleted medium and the lag phase was extended (Fig. 3a). The sidD::Km growth in iron-replete media was very similar to the growth of the wild-type strain in the DP-treated, iron-limited medium. To determine if supplementing the growth medium with additional iron would rescue the growth defect of the of sidD::Km strain, iron chloride was added to the medium at a final concentration of up to 1 mM; the additional iron did not restore the growth of the sidD::Km mutant (Fig. 3b). The growth profile of the sidD::Km mutant in iron-replete medium was reduced to about 50% of that of the wild-type strain in late log and stationary phases.

Fig. 3

Growth of sidD::Km mutant is reduced when compared to wild type in the iron-depleted medium. Strains were grown in Actinomyces defined medium (ADM) in low iron (addition of 150 µM dipyridyl, DP) or high iron (addition of 1 mM FeCl3) in 96-well plates at 5% CO2, WT = wild-type A. oris MG1.

When the sidD gene on a plasmid was added back to the sidD::Km mutant, growth was restored in the iron-replete media to almost wild-type levels. However, when the conditions were severely iron-limited, the complemented strain only partially rescued growth; this strain did not reach the same final growth yield as well as the wild type.

Actinomyces metal-dependent repressor is conserved in clinical isolates

There is often a high degree of gene variation in microorganisms in the oral cavity (24), and as such, we wanted to explore whether the key iron regulator required to control expression of the siderophore systems outlined above, AmdR, was conserved in clinical isolates of Actinomyces spp. found in the oral cavity. Ten Actinomyces spp. strains isolated from individual patient's oral cavity at the University of Florida were kindly provided by Dr. Burne. amdR was successfully amplified by PCR (see Material and Methods for primers) from 7 of the 10 clinical isolates. The reason that we were unable to amplify the amdR ORF from the remaining isolates, even after repeated attempts, was not clear. Deduced amino acid identity varied from 92 to 99% among these AmdR homologs (Fig. 4), indicating that this gene and its product are highly conserved among this collection of isolates. GenBank Accession numbers for clinical isolates UF5–UF45 are KU162964–KU162970.

Fig. 4

Alignment of AmdR sequences from clinical isolates. Alignment of deduced amino acid sequences of cloned amdR genes from seven Actinomyces clinical isolates was performed using Clustal Omega (25). The laboratory strain A. oris MG1 (BROP.org) is labeled ‘LAB’. Percent amino acid identity and similarity to A. oris MG1 is as follows: UF5 (95/97), UF 10 (95/97), UF 12 (98/98), UF 14 (96/97), UF 22 (96/97), UF 28 (98/98), UF 45 (99/99). ‘*’ indicates identical amino acids; ‘.’ and ‘:’ indicate conservative substitution.

Discussion

ABC transporters are responsible for the uptake of iron in many bacteria (1). A. oris contains multiple putative iron transport systems (BROP.org) including the operons containing the sidABCD genes as well as the fetAB genes. Iron ABC transporters have been shown to be expressed differentially depending on the amount of iron in the environment. In a low-iron environment, expression of iron transporters in many other organisms increases due to the need of scavenging iron in the environment using siderophores (1, 26). In a high-iron environment, expression of iron ABC transporters is typically low since iron is readily available through low-affinity transporters (26).

Here, we found that the genes fetA and sidD showed significantly increased gene expression in iron-limited medium in the wild-type strain. In order to further characterize the role of these transporters in the survival of A. oris, we sought to mutate the sidD gene. This sidD::Km mutant showed a longer lag phase and lower yield in iron-depleted medium compared to the wild type; however, the addition of iron to the growth media did not restore growth of the mutant to wild-type levels. Other studies that have generated mutants of iron transport, including one report in C. diphtheriae, have found similar growth defects. That is, a mutation in the C. diphtheriae ciuA gene, encoding a putative lipoprotein component of an iron transport system, resulted in a severe defect in iron uptake and reduction of C. diphtheriae siderophore uptake (27). When the sidD gene was added back to the sidD::Km mutant in a complementation experiment, the growth was restored to wild-type levels in iron-replete medium; however, when the conditions were severely iron-limited, the complemented strain did not reach quite the same final growth yield as well as the wild type, although the growth rates of these strains were similar. The sidD::Km mutant was examined for a sidD transcript, and as predicted, no transcript was detected. We found that the fetA gene transcript in the sidD::Km mutant was decreased compared to the wild type. As noted above, the sidD::Km mutant strain grows much slower than the wild type and does not reach the same final OD, perhaps explaining the reduced fetA gene expression. Alternatively, the sidD::Km mutant may be defective in responding to iron limitation.

Mutants of iron transporters have been linked with virulence, speaking to the key role of iron acquisition during infection (28). Mutations in S. mutans sloC (part of an ABC transporter involved in iron transport) resulted in decreased virulence in an endocarditis model (29, 30). Future studies could examine the effects of the sidD defective strain of A. oris in an animal model for periodontal disease.

A. oris is an important colonizer of the oral cavity. By the age of 1 year, 90% of infants evaluated are colonized with A. oris (31). In addition, this organism is a prominent part of the periodontal flora (32). Laboratory strains are useful model organisms, and clearly metal transport is important for optimal growth of the A. oris MG1 strain in our laboratory studies; however, not all genes that are found in laboratory strains are conserved across clinical isolates. To further illustrate how important metal transport genes are, studies characterizing the genomes of clinical isolates of S. mutans have found that most of the genome varies widely; however, metal transporters and other key virulence factors are conserved (24). The 73 ‘core genes’ of S. mutans are thought to contribute to conditions that allow S. mutans to colonize and persist in the oral cavity (24). In order to gain a better understanding of the role of iron uptake and regulation in clinically relevant strains, we obtained 10 clinical isolates of Actinomyces, including A. oris. The identity of these isolates was confirmed by sequencing the 16S rRNA gene. We were able to PCR-amplify, sequence, and subclone a gene encoding AmdR from 7 out of the 10 clinical isolates. We saw a high degree of deduced amino acid sequence identity (between 92 and 99%), indicating that these metal-dependent regulators may be conserved across at least a subset of clinical isolates. Taken together with our previously published work (14), our data indicate that AmdR, and the AmdR- and iron-regulated genes, including siderophore systems, are keys for growth of this microbe in low-iron environments, and thus likely contribute to their survival in the oral cavity.