FRAMING THE ROLE OF RHODANESE-LIKE PROTEINS IN CELL REDOX BALANCE IN TWO BACTERIAL MODEL SYSTEMS

A key component of the host’s ability to survive bacterial challenge consist in the innate ability of macrophages to ingest and destroy the invading organism via various mechanisms which the most thoroughly studied is oxidative burst (Hanna et al., 1994). In response pathogens have evolved different approaches to survive the severe oxidative stress generated by the host. Genome analysis have clustered more than 14000 sequences coding for putative rhodanese like proteins. These sequences contain domains structurally similar to those of the extensively studied bovine rhodanese, and were found in more than 2100 species homogeneously distributed in all life's phyla. Proteins belonging to the rhodanese like superfamily (PFAM accession number: PF00581) are characterized by having more than 140 different architectures of the rhodanese domain, that can be present mostly alone or in tandem with another rhodanese domain, or fused to other functional domains. Although only few amino acid residues are conserved among rhodanese-like proteins, their most distinctive structural feature is the active site configuration, that contains an electronegative residue (generally cysteine) surrounded by positively charged residues. This particular architecture allows rhodanese-like proteins to bear a low pKa catalytic residue that can be the clue to explain their biological activity (Bordo et al 2001). Although the in vitro reported sulfurtransferase activity for the few characterized rhodanese-like proteins is the transfer of sulfur atom from a sulfur donor (e.g. thiosulfate) to a thiophilic acceptor (e.g. cyanide) (E.C. 2.8.1.x), in the last two decades the scientific community has started to indicate biological roles different from cyanide detoxification for rhodanese like proteins. Proposed roles for rhodanese like proteins can be summarized in two different but complementary fields. Rhodanese-like proteins can function as source of bioactive sulfur equivalents by the formation of a persulfide sulfur on a cysteine residue (R-SSH) (Cartini et al 2011), or can be involved in maintaining redox homeostasis acting as a direct or indirect scavengers of reactive oxygen species (ROS). My PhD research project was devoted to unravel the biological roles of rhodanese-like proteins using two prokaryotic model systems: the Azotobacter vinelandii and the Bacillus subtilis system. A. vinelandii is a Gram negative bacterium of the Pseudomonadaceae family in which redox balance must be carefully controlled due to its ability to fix molecular nitrogen via the molybdenum-iron-sulfur cluster enzyme nitrogenase (Setubal et al., 2009). The A. vinelandii genome possesses 14 ORFs coding for rhodanese like proteins with the tandem domain rhodanese-like protein RhdA (Gene ID: 7759697) being responsible for more than 80% of the crude extract thiosulfate:cyanide sulfurtransferase (TST) activity (Cartini et al., 2011). RhdA was widely studied in our lab from both structural and functional point of views. Starting from the evidence that the rhdA null-mutant strain (MV474) showed altered sensitivity to oxidative events (Cereda et al., 2009), I investigated the nature of the endogenous oxidative stress induced in A. vinelandii by the absence of RhdA. I found that in MV474 strain the ratio GSH/GSSG was misregulated, and the levels of lipid hydroperoxides were significantly increased, although defensive activities against oxidative stress damage were activated (e.g. upregulation of the ahpC gene, coding for Alkylhydroperoxidase C, a member of the OxyR regulon). Furthermore, rhdA expression was highly induced in the A. vinelandii strain (UW136) when the oxidative stress was performed by the incubation with the superoxide generator phenazine methosulfate (PMS). These results demonstrated that RhdA has a role in protecting A. vinelandii from oxidative damage and were the subject of a publication (Remelli et al., 2010). Therefore I addressed my study to understand how RhdA could function in protecting redox homeostasis in A. vinelandii. I studied, in vitro, the oxidation behavior of the only cysteine residue (Cys230) present in RhdA. Site directed mutagenesis showed that the Cys230 residue is mandatory for RhdA TST activity (Cartini et al., 2011). Combining results of TST activity assays, and thiol quantification by the use of the fluorescence probe monobromobimane (mBBr), I found that, after incubation with PMS, Cys230 underwent irreversible oxidation, while underwent reversible oxidation if RhdA was preincubated with thiosulfate or reduced glutathione (GSH). This latter result was taken as an indication that productive interaction with GSH, the widespread redox buffer, could be a key point on understanding RhdA biological function in A. vinelandii. The RhdA/glutathione interaction was characterized using different approaches. The ability of RhdA to interact with different glutathione species of biological relevance was studied by measuring changes of the RhdA intrinsic fluorescence due to tryptophan residues surrounding the RhdA active site. In particular, RhdA is able to strongly interact with GSH (Kd = 1.5 µM), glutathione thiyl radical (GS•; Kd = 10.4 µM), a radical form of GSH, while interaction with disulfide glutathione (GSSG) is weak (Kd = 1.4 mM). According to the RhdA active site features and to the thiol chemistry, a covalent binding with GSH was excluded by mass spectroscopy analyses and by the finding that reducing agents were unable to break the RhdA/GSH complex. Moreover, the inability of RhdA to bind GSH in the presence of high ionic strength conditions suggests that RhdA/glutathione complex is stabilized by electrostatic interactions according to the in silico model produced by docking the interaction. The ability of RhdA to interact with glutathione (especially with GS•) as well as the results that GSH level was lower in the mutant than in the wild-type strain, were considered for discerning the biological functions of RhdA. Only a small number of proteins can manage GS•, among them human glutaredoxin 1 and 2 (hGrx1; hGrx2) can catalyze the reaction between GSH and GS• leading to the formation of GSSG and a superoxide radical (Starke et al., 2003). The ability of these proteins to catalyze the oxidation of glutathione is mainly due to the peculiar feature of their active site that bear a low pKa cysteine residue (Gallogly et al., 2008) . In vitro assays, using GS•-generating mix as a substrate, demonstrate the ability of RhdA to catalyze GSSG production with a turnover number 180-fold higher compared to human glutaredoxin 1, the glutaredoxin that present higher activity (RhdA kcat: 628 s-1). The lacking in the A. vinelandii genome of genes coding for hGrx1- and hGrx2-like glutaredoxins (while genes for other glutaredoxins are present), together with the absence of a complete ascorbic acid biosynthetic pathway, increases the biological relevance of the RhdA ability to scavenge GS•, preventing further oxidation of GS• to toxic and/or unrecoverable forms. Considering that GS• is mainly produced from GSH activity of, for example, detoxification of hydroxyl radicals (OH•) produced during cellular respiration (Lushchack, 2011), it is reasonable to believe that the RhdA function in protecting redox homeostasis must be related to the maintaining of the cellular respiration rate and has been further suggested by monitoring the oxygen consumption of UW136 and MV474 crude extracts. B. subtilis is a Gram positive bacterium that has been taken as a model prokaryotic system because of its close evolutional relationship with pathogen Gram positive bacteria like B. anthracis, (which causes anthrax), B. cereus (responsible of the foodborne illness), and B. thuringiensis that is an important insect pathogen. Analysis of the B. subtilis genome allowed the identification of 5 rhodanese like proteins; 3 (YtwF, YqhL, YbfQ) are single-domain protein and 2 (YrkF and YrkH) have the rhodanese domain fused to other functional domains. Noteworthy, rhodanese like proteins having the RhdA domain architecture (i.e. two rhodanese domains organized in tandem) are not present in the B. subtilis genome. B. subtilis rhodanese-like proteins share the structural features of the rhodanese catalytic domain: a putative catalytic cysteine residue surrounded by positively-charged amino acid residues. In silico analyses on transcription factor binding sites revealed that YbfQ, YhqL and YtwF are putatively controlled by FurR homologs, that in B. subtilis are the transcription factors that controls: the iron uptake (FurR), the zinc uptake (ZurR) and the peroxide response system (PerR). Furthermore yhqL transcription is also putatively controlled by NagC that is related to the expression of proteins involved in the N-acetil-glucosammine (GlcNac) transport. These informations suggest that B. subtilis rhodanese-like proteins, or at least some of them, could be involved in the oxidative stress response system. This idea have been recently supported by transcriptome analyses in which overexpression of ybfQ B. anthracis homolog has been shown after induction of the oxidative stress by treatment with hydrogen peroxide (Pohl et al., 2011). A preliminary characterization of biological function of rhodanese-like proteins in B. subtilis was performed analyzing phenotypic changes on the rhodanese-like quadruple null mutant J1235 strain (yrkF, yhqL, ybfQ, ytwF), kindly provided by professor T. Larson, compared to the isogenic wild type strain PS832. Whereas no visible growth differences were observed in rich medium (LB), a small, but significant, growth difference was observed in Spizizen minimal medium supplemented with 0.4% sucrose. Hydrogen peroxide challenging assays, using both sublethal and lethal stressor concentrations, indicated that J1235 strain was more prone to oxidative stress damage compared to the wild type PS832 strain. It suggests that the absence of rhodanese-like proteins caused a chronic endogenous oxidative stress condition enhanced by growths on the minimal media. These results were further confirmed by the ~40% decrease of the aconitase activity, an oxidative stress sensitive Fe-S cluster enzyme that is generally used as a marker to evaluate cellular oxidative stress conditions. Furthermore J1235 strain exhibited a ~50% increase of the intracellular lipid hydroperoxide content suggesting membrane oxidation. Differently from eukaryotes and most of the Gram negative bacteria, the majority of the Gram positive bacteria have evolved different strategies to maintain intracellular redox homeostasis that don't involve the synthesis of GSH (Fahey et al., 1978). The most thoroughly studied of these alternative compounds are mycothiol (MSH) and coenzyme A (CoA) (Rawat et al., 2007; del Cardayré et al., 1998). Recent studies on the redox-sensitive thiol proteins of B. subtilis, have uncovered an unknown low molecular mass thiol in association with the transcription factor Ohr (Nicely at al., 2007) that, in eukaryotic organisms, is regulated by glutathionylation (Hermansson et al., 1990). This new molecule, named bacillithiol (BSH), shares some structural characteristics with MSH (e.g. the GlcNac and cysteine mojeties) (Newton et al., 2009), and has proven to be the B. subtilis major low molecular weight thiol (Gaballa et al., 2010). Interestingly, a decrease of both total soluble BSH (~35%) and the BSH/BSSB ratio (~40%) were observed when the B. subtilis J1235 strain, where rhodanese-like proteins are not expressed, was compared with the wild-type strain. These and other phenotype features (i.e. differences of low-molecular weight thiol levels; increase of lipid oxidation), make the condition observed for the B. subtilis J1235 strain very similar to that observed for the A. vinelandii rhdA null mutant strain MV474, suggesting that at least one of the rhodanese-like protein shares the biological function of A. vinelandii RhdA. Structural differences between RhdA and B. subtilis rhodaneses could be explained by the structural differences between GSH and BSH, that requires a different active site architecture. For the cell localization and the presence of FurR and NagC control elements, among the B. subtilis rhodanese-like proteins, YhqL is supposed to be the primary candidate that could share the same biological function with RhdA. In preliminary experiments, the yhqL single mutant JD0206 and the rhodanese-like quadruple null mutant J1235 strains showed similar results about the BSH levels, further supporting the hypothesis that YhqL is the B. subtilis orthologous of RhdA. In conclusion, during my PhD research project, I gave solid indications of a direct involvement of members of the rhodanese-like protein superfamily in protecting the cell from endogenous oxidative stress that can arise, for example, from cellular respiration. The maintenance of the cellular respiration is equivalent to cell survival and became critic, for example, for pathogens bacteria when under attack of the host cell opening, among others, to future implications of the rhodanese-like protein superfamily as a potential drug target for pathogen eradication.