Introduction, methods, and results

Rhodococcus equi is an actinobacterial pathogen that can infect immunocompromised humans and foals by causing a fatal pyogranulomatous bronchopneumonia [1]. R. equi is distributed worldwide, being highly prevalent in farms because of its colonization of the horse intestine [2]. This pathogen is usually transmitted by inhaling R. equi-contaminated dust or respiratory particles produced by infected animals [2].

During the past few decades, a lot of effort has been focused on identifying and studying genes of R. equi that could be involved in host–pathogen interactions in search of new strategies to tackle the infections caused by these Actinobacteria. The rise of multidrug-resistant R. equi strains is making current antibiotherapies ineffective [3, 4]. In addition, any attempts to develop a vaccine against R. equi have been unsuccessful so far [5]. Because of this, hyperimmune plasma administration has been implemented as a preventative primary intervention in foals, despite of its high costs and variable efficacy [6].

It is becoming clear that the virulence associated proteins (Vaps) of R. equi are major determinants of the control of intraphagolysosomal pH during cell infection [7]. Furthermore, different members of the pVAP megaplasmids family carry specific complements of vap genes, which are essential for the intracellular survival of R. equi and they are considered the main driving factor of the host tropism of this pathogen [8].

On the other hand, bacterial proteins involved in redox homeostasis have been traditionally considered very attractive targets for the development of novel anti-infectives against many pathogens [9, 10]. Importantly, R. equi is exposed to high concentrations of reactive oxygen and nitrogen species (RONS) during phagocytosis, which may affect membrane lipids, nucleic acids, housekeeping proteins and virulence factors of the pathogen [10]. During phagocytosis, the production of RONS is triggered by the activation of NOX and iNOS proteins in the macrophage [11]. In response, R. equi resists the oxidative stress in the phagosome with catalases, superoxide dismutases, alkyl hydroperoxide reductases and thiol peroxidases [12]. Furthermore, R. equi is well equipped with protein-repairing genes encoding mycothiol and mycoredoxins (Mrx), which are only present in Actinobacteria [13]. Moreover, there are several genes in the genome of R. equi encoding proteins with thioredoxin domains. The main thioredoxin-based antioxidant system is well conserved in bacteria and, in particular, this is encoded by REQ_47340-50 in R. equi [12].

However, a detailed analysis of the R. equi genome annotation [12] revealed the presence of four genes encoding putative extracellular thioredoxins (Etrx), which were named as Etrx1 (REQ_05180), Etrx2 (REQ_08580), Etrx3 (REQ_13520) and Etrx4 (REQ_37440) (Additional file 1). All four R. equi putative Etrx proteins were aligned to extracellular thioredoxins previously studied as important virulence factors in Streptococcus pneumoniae [14,15,16] and Mycobacterium tuberculosis [17] (Additional file 1). All Etrx proteins showed a high sequence homology in the domains containing a thioredoxin-active site (WCxxC).

Furthermore, we clustered all Etrx proteins of R. equi in an evolutionary distance tree (Additional file 2). Etrx2 of R. equi was not rooted with any of the other Etrx proteins included in this analysis, whereas Etrx1 was clustered with CcsX of M. tuberculosis. Interestingly, both Etrx3 and Etrx4 were grouped with Rv0526 of M. tuberculosis, suggesting that these proteins were two orthologs of Rv0526. Rv0526 has been previously characterized as an extracellular protein anchored to the bacterial membrane in mycobacteria, but very little is known about its possible role in virulence [18].

In addition, all Etrx proteins were analysed with SignalP [19], TMHMM [20], Pfam [21] and pDomTHREADER [22] to determine their signal peptides, trans-membrane helix domains and other protein domains (Figure 1). As expected, the overall structure of Etrx3 and Etrx4 was quite similar to that from Rv0526 of M. tuberculosis (Figure 1).

Figure 1
figure 1

Protein domains of different extracellular thioredoxins. The protein sequences of extracellular thioredoxins from R. equi, S. pneumoniae and M. tuberculosis were analysed with SignalP 5.0, TMHMM server, Pfam and pDomTHREADER.

Full size image

Therefore, we analysed the genome regions carrying etrx3 and etrx4 of R. equi and M. tuberculosis or the non-pathogenic Rhodococcus erythropolis using the Artemis Comparison Tool [23] (Additional file 3). The gene cluster carrying etrx4 in R. equi was inverted but very well conserved in both M. tuberculosis and R. erythropolis (Additional file 3A). In addition, etrx3 is an ortholog of Rv0526 of M. tuberculosis (Additional file 3B). However, the synteny of the region carrying etrx3 was very poorly conserved in all three genomes analysed (Additional file 3C). In addition, this region was not acquired by horizontal gene transfer according to a previous analysis of the R. equi 103S+ genome [12]. This suggested that etrx3 might have been acquired by a duplication of etrx4 in R. equi, and the new copy of the gene was created in a region made of recurrent genomic rearrangements. Overall, our in silico analysis indicated that Etrx3 might be an extracellular thioredoxin that is unique to R. equi. Therefore, we generated an etrx3-null mutant strain to study its role in the control of the redox homeostasis of R. equi during phagocytosis.

To generate an unmarked in-frame deletion of etrx3 (REQ_13520) in R. equi 103S+ (Additional file 4), we amplified by PCR two 1.5 kbp DNA fragments corresponding to the upstream and downstream sequences of the gene (Additional file 5). The resulting amplicons were used as the DNA template of a fusion-PCR reaction to generate a 3 kbp DNA cassette harbouring an in-frame deletion of etrx3, which was cloned into pSelAct (Additional file 4) as previously described for the deletion of other genes in R. equi [10]. The resulting vector (pSelActΔetrx3—Additional file 4) was electroporated into R. equi 103S+ and its integration was verified by PCR in several apramycin resistant transformants resulting from the electroporation. The deletion of the etrx3 gene in R. equi Δetrx3 was achieved by means of a second recombination event, making use of 5-fluorocytosine counterselection, as previously described [10]. The in-frame deletion of the gene was confirmed by PCR amplification.

To complement the etrx3-null mutant with a single copy of etrx3, the gene was amplified under the control of its own promoter and cloned in the integrative plasmid pSET152, as described previously for the complementation of other gene deletions in R. equi [10]. The resulting vector (pSETetrx3—Additional file 4) was used to electroporate R. equi Δetrx3, transformants were selected by apramycin-resistance, and the integration of the vector in R. equi Δetrx3 + pSETetrx3 was confirmed by PCR (Additional file 5). All vectors produced in this study were verified by DNA sequencing.

Optical density at 600 nm (OD600) was used to establish the growth curves in trypticase soy broth (TSB) of the mutant strains produced in this study in order to discard any polar effects on their replication rate that could possibly result from genetic engineering. When compared to the wild type strain, the replication rate of both R. equi Δetrx3 and R. equi Δetrx3 + pSETetrx3 was unaltered, which facilitated the analysis of their intracellular proliferation rate during infection assays (Additional file 6). Statistical analyses were conducted using IBM® SPSS® statistics v24. One-way ANOVA and post hoc Tukey’s multiple-comparison tests were routinely employed to identify statistically significant differences across conditions in this study.

Macrophage infection assays were performed as previously described [12] using low-passage J774A.1 murine macrophages (American Type Culture Collection) cultured in Dulbecco’s Modified Eagle Medium (DMEM—Thermo-Fisher Scientific). Macrophages were infected at a multiplicity of infection of 10 with exponentially growing cultures (OD600 ≈ 1) of R. equi in TSB. The presence of the virulence plasmid pVAPA was verified by PCR in all R. equi strains tested preceding each infection assay, as previously described [10]. After 1 h of incubation, the medium was replaced with DMEM supplemented with 5 µg/mL vancomycin to kill extracellular bacteria, as previously described [24]. At different time points, cells were lysed with 0.1% Triton X-100 and serial dilutions of the lysates were spread onto trypticase soy agar (TSA) plates for colony forming unit (CFU) counting.

We infected J774A.1 cells with R. equi Δetrx3 and R. equi Δetrx3 + pSET-etrx3. In parallel, we also infected macrophages with the R. equi 103S+ wild type strain and the virulence plasmid cured derivative R. equi 103S−, which were respectively considered positive and negative controls of macrophage infection (Figure 2). Interestingly, the internalization rate of the etrx3-null mutant strain was significantly higher when compared to the wild type strain (Figure 2A). Despite of this, the R. equi Δetrx3 strain was unable to persist in the intracellular environment. In contrast, the internalization and intracellular survival of the R. equi Δetrx3 + pSET-etrx3 complemented strain was comparable to R. equi 103S+ (Figure 2). Overall, these results suggest that Etrx3 has an essential role in R. equi’s macrophage infection.

Figure 2
figure 2

Macrophage infection assays. Percentages of internalization (A) and intracellular survival at 48 h (B) in J774A.1 macrophages of the wild type R. equi 103S+ strain, the virulence plasmid cured R. equi 103S− strain, R. equi Δetrx3, and R. equi Δetrx3 + pSET-etrx3 (Δetrx3 + etrx3). Bacterial viability was calculated by quantifying the number of colony forming units (CFUs) of each strain and by normalizing these data against R. equi 103S+ CFUs. Data are expressed as mean ± SD of three technical replicates repeated in three independent experiments. One-way ANOVA and post hoc Tukey’s multiple comparison tests were performed to assess for statistical significance in relation to the wild type strain. **p-value < 0.01.

Full size image

To cast some light on the role of Etrx3 during phagocytosis, we analysed the resistance of the R. equi Δetrx3 mutant strain to different oxidative stressors as previously described [10]. Exponential growth phase cultures (OD600 = 1) were diluted 1:10 in plain TSB or in TSB supplemented with 10 mM H2O2, 5 mM NaClO, or minimum medium supplemented with free methionine sulfoxide (MetSO−) at different concentrations, and incubated at 30 °C and 220 rpm. At different time points, each culture was serially diluted and spread on TSA plates, which were incubated for 24 h at 30 °C. The number of CFUs was then quantified and results were normalized to the survival rate of R. equi 103S+. In contrast, to determine the susceptibility to DETA NONOate (a nitric oxide donor), R. equi exponential growth phase cultures (OD600 = 1) were 1:10 diluted in 10 mL of liquefied TSA (0.6% agar) at 50 °C and spread over 10 mL of settled TSA. Nitrocellulose disks were then placed on the surface of R. equi-containing TSA plates. Finally, DETA NONOate was added to paper disks at defined concentrations and plates were incubated at 30 °C for 24 h.

Our results showed that R. equi Δetrx3 was significantly more susceptible to sodium hypochlorite than the wild type strain or the R. equi Δetrx3 + pSET-etrx3 complemented strain (Figure 3). In contrast, R. equi’s resistance to nitric oxide or free methionine sulfoxide (MetSO−) was not altered in the etrx3-null mutant (Additional files 7 and 8), and its resistance to H2O2 was even increased (Additional file 9). Importantly, sodium hypochlorite is considered a source of hypochlorous acid, which is produced by a myeloperoxidase expressed in professional phagocytes such as macrophages [25, 26]. Therefore, the low survival rate of R. equi Δetrx3 within murine macrophages might be due to its high susceptibility to sodium hypochlorite.

Figure 3
figure 3

Resistance of differentR. equistrains to 5 mM NaClO. Data were normalized by the percentage of R. equi 103S+ CFUs and are expressed as mean ± SD of three technical replicates repeated in three independent experiments. One-way ANOVA and post hoc Tukey’s multiple comparison tests were performed to assess for statistical significance across conditions. *p-value < 0.05.

Full size image

To verify this, we also analysed the ratiometric response of Mrx1-roGFP2 in this context, a reengineered redox biosensor that allows to evaluate changes in mycothiol levels in response to an oxidative stressor [10]. The redox status of Mrx1-roGFP2 was measured as described before by means of confocal microscopy [10]. Interestingly, the deletion of etrx3 in R. equi delays the oxidation of Mrx1-roGFP2 caused by NaClO (Figure 4). Overall, our results suggest that Etrx3 has a role in counteracting the redox stress exerted by NaClO.

Figure 4
figure 4

Ratiometric response of Mrx1-roGFP2 biosensor expressed inR. equi103S+andR. equiΔetrx3, which were cultured in TSB supplemented with 1 mM NaClO. Data represent mean ± SD of three technical replicates repeated in three independent experiments.

Full size image

Discussion

During phagocytosis, R. equi is exposed to RONS generated by host myeloperoxidases, nitric oxide synthases and NADPH oxidases [11]. The control of the pathogen’s extracellular redox homeostasis could be essential to maintain the reduced state and activity of its secreted virulence factors. Otherwise, the reactive oxygen and nitrogen species generated by the macrophage may inactivate essential pathogen’s proteins by oxidation of their cysteine or methionine amino acids.

In Actinobacteria, the thioredoxin/thioredoxin reductase (Trx/TrxR) system act together with the mycoredoxins/mycothiol (Mrx/MSH) system to maintain the reduced state of proteins [13]. The mycoredoxins/mycothiol system restores the reduced state of cysteine residues [13]. The methionine oxidized residues could be reduced by methionine sulfoxide reductases (Msr), which are in turn reduced by a transfer of electrons from the active CxxC site of thioredoxins to the Msr disulphides. Finally, the oxidized thioredoxins are reduced by an NADPH-dependent thioredoxin reductase.

There is an increasing evidence demonstrating the essential role of thioredoxins in the virulence of many bacterial pathogens. In Listeria monocytogenes, TrxA maintains the reduced status of the master regulator of virulence PrfA and the key regulator of flagellar synthesis MogR [27]. TrxA is also essential for the intracellular induction of Salmonella pathogenicity island 2 (SPI2) type III secretion system (T3SS) and, consequently, for the intracellular replication of Salmonella enterica serovar Typhimurium [28].

In addition, several extracellular thioredoxins have been recently described with essential roles in the virulence of M. tuberculosis, S. pneumoniae or Agrobacterium tumefaciens [14,15,16,17, 29]. In S. pneumoniae, it is becoming clear that the functional paralogues Etrx1 and Etrx2 and the methionine sulfoxide reductase MsrAB2 are part of an extracellular electron pathway. This is required to maintain the redox state of methionine residues present in surface-exposed proteins that are essential for the pathogen’s survival to phagocytosis [14, 15]. However, extracellular thioredoxins may have other functions in the cell. For instance, the extracellular thioredoxin CcsX of M. tuberculosis is involved in the maturation of cytochrome c oxidase. In all cases, these extracellular redoxins are probably coupled to electron transport chains in the pathogen’s cytoplasmic membrane, which act as the source of their reducing power.

Similarly, here we describe the importance of the extracellular thioredoxin Etrx3 on the intracellular survival of R. equi, an actinobacterial pathogen causing infections that are becoming very difficult to treat due to antibacterial resistance. Overall, our data suggest that Etrx3 is essential for the survival of R. equi to phagocytosis, and that this extracellular thioredoxin is required to preserve the redox homeostasis of R. equi when the pathogen is exposed to NaClO.

However, the high resistance of R. equi Δetrx3 to H2O2 suggests that the deletion of etrx3 leads to a compensatory effect that may implicate the overexpression of other proteins involved in redox homeostasis. Similarly, a ccsX-null mutant of M. tuberculosis exhibited high resistance to H2O2 due to the overexpression of the cytochrome bd oxidase [17]. In addition, a double etrx1/etrx2-null mutant of S. pneumoniae was found to be more resistant than the wild type strain to the superoxide-generating compound paraquat [15]. Further studies are necessary to understand the role of Etrx3 in this context. Nonetheless, the high resistance to H2O2 of R. equi Δetrx3 had no impact on macrophage infection, since the etrx3-null mutant strain was still unable to survive phagocytosis (Figure 2).

On the other hand, the deletion of etrx3 did not alter R. equi’s resistance to the oxidative stress induced by free methionine sulfoxide (Additional file 8). This is in stark contrast to the susceptibility of Etrx1 and Etrx2-null mutants of S. pneumoniae to MetSO− [15], suggesting that the role of Etrx3 is not related to the reduction of Msr proteins in R. equi (encoded by REQ_01570 and REQ_20650).

Further research is required to elucidate the precise function of Etrx3. However, the essential role of this extracellular thioredoxin during macrophage infection makes the etrx3-null mutant strain an attractive candidate for the development of an attenuated vaccine. The Δetrx3 deletion strain might be able to elicit a strong immune response against R. equi since it was unable to survive phagocytosis despite of carrying a functional pVAPA virulence plasmid, which is required to generate both cell-mediated and humoral immune responses [5].