Hannah M. Southam

Carbon monoxide (CO)-releasing molecules (CORMs), mostly metal carbonyl compounds, are extensively used as experimental tools to deliver CO, a biological 'gasotransmitter', in mammalian systems. CORMs are also explored as potential novel antimicrobial drugs, effectively and rapidly killing bacteria in vitro and in animal models, but are reportedly benign towards mammalian cells. Ru-carbonyl CORMs, exemplified by CORM-3 (Ru(CO) 3 Cl (glycinate)), exhibit the most potent antimicrobial effects against Escherichia coli. We demonstrate that CORM-3 releases little CO in buffers and cell culture media and that the active antimicrobial agent is Ru(II), which binds tightly to thiols. Thus, thiols and amino acids in complex growth mediasuch as histidine, methionine and oxidised glutathione, but most pertinently cysteine and reduced glutathione (GSH)protect both bacterial and mammalian cells against CORM-3 by binding and sequestering Ru(II). No other amino acids exert significant protective effects. NMR reveals that CORM-3 binds cysteine and GSH in a 1:1 stoichiometry with dissociation constants, K d , of about 5 M, while histidine, GSSG and methionine are bound less tightly, with K d values ranging between 800 and 9000 M. There is a direct positive correlation between protection and amino acid affinity for CORM-3. Intracellular targets of CORM-3 in both bacterial and mammalian cells are therefore expected to include GSH, free Cys, His and Met residues and any molecules that contain these surface-exposed amino acids. These results necessitate a major reappraisal of the biological effects of CORM-3 and related CORMs.

Medicinal leads that are also compatible with imaging technologies are attractive, as they facilitate the development of therapeutics through direct mechanistic observations at the molecular level. In this context, the uptake and antimicrobial activities of several luminescent dinuclear RuII complexes against E. coli were assessed and compared to results obtained for another ESKAPE pathogen, the Gram-positive major opportunistic pathogen Enterococcus faecalis, V583. The most promising lead displays potent activity, particularly against the Gram-negative bacteria, and potency is retained in the uropathogenic multidrug resistant EC958 ST131 strain. Exploiting the inherent luminescent properties of this complex, super-resolution STED nanoscopy was used to image its initial localization at/in cellular membranes and its subsequent transfer to the cell poles. Membrane damage assays confirm that the complex disrupts the bacterial membrane structure before internalization. Mammalian cell culture and animal model studies indicate that the complex is not toxic to eukaryotes, even at concentrations that are several orders of magnitude higher than its minimum inhibitory concentration (MIC). Taken together, these results have identified a lead molecular architecture for hard-to-treat, multiresistant, Gram-negative bacteria, which displays activities that are already comparable to optimized natural product-based leads.

Carbon monoxide (CO)-releasing molecules (CORMs) are used to deliver CO, a biological ‘gasotransmitter’, in biological chemistry and biomedicine. CORMs kill bacteria in culture and in animal models, but are reportedly benign towards mammalian cells. CORM-2 (tricarbonyldichlororuthenium(II) dimer, Ru2Cl4(CO)6), the first widely used and commercially available CORM, displays numerous pharmacological, biochemical and microbiological activities, generally attributed to CO release. Here, we investigate the basis of its potent antibacterial activity against Escherichia coli and demonstrate, using three globin CO sensors, that CORM-2 releases negligible CO (<0.1 mol CO per mol CORM-2). A strong negative correlation between viability and cellular ruthenium accumulation implies that ruthenium toxicity underlies biocidal activity. Exogenous amino acids and thiols (especially cysteine, glutathione and N-acetyl cysteine) protected bacteria against inhibition of growth by CORM-2. Bacteria treated with 30 μM CORM-2, with added cysteine and histidine, exhibited no significant loss of viability, but were killed in the absence of these amino acids. Their prevention of toxicity correlates with their CORM-2-binding affinities (Cys, Kd 3 μM; His, Kd 130 μM) as determined by 1H-NMR. Glutathione is proposed to be an important intracellular target of CORM-2, with CORM-2 having a much higher affinity for reduced glutathione (GSH) than oxidised glutathione (GSSG) (GSH, Kd 2 μM; GSSG, Kd 25,000 μM). The toxicity of low, but potent, levels (15 μM) of CORM-2 was accompanied by cell lysis, as judged by the release of cytoplasmic ATP pools. The biological effects of CORM-2 and related CORMs, and the design of biological experiments, must be re-examined in the light of these data.

A gasotransmitter is defined as a small, generally reactive, gaseous molecule that, in solution, is generated endogenously in an organism and exerts important signalling roles. It is noteworthy that these molecules are also toxic and antimicrobial. We ask: is this definition of a gasotransmitter appropriate in the cases of nitric oxide, carbon monoxide and hydrogen sulfide (H2S) in microbes? Recent advances show that, not only do bacteria synthesise each of these gases, but the molecules also have important signalling or messenger roles in addition to their toxic effects. However, strict application of the criteria proposed for a gasotransmitter leads us to conclude that the term ‘small molecule signalling agent’, as proposed by Fukuto and others, is preferable terminology.