Polynuclear ruthenium complexes are effective antibiotics against Pseudomonas 1 aeruginosa 2 3

There is an urgent need to develop new antibiotics for the treatment of infections 26 caused by drug-resistant Gram-negative bacteria. In particular, new and diverse 27 chemical classes of antibiotics are needed, as most antibiotics in clinical development 28 are derivatives of existing drugs. Despite a history of use as antimicrobials, metals and 29 metal-based compounds have largely been overlooked as a source of new chemical 30 matter for antibacterial drug discovery. In this work, we identify several ruthenium 31 complexes, ruthenium red, Ru265, and Ru360’, that possess potent antibacterial activity 32 against both laboratory and clinical isolates of Pseudomonas aeruginosa . Suppressors 33 with increased resistance were sequenced and found to contain mutations in the 34 mechanosensitive ion channel mscS-1 or the colRS two component system. The 35 antibacterial activity of these compounds translated in vivo to Galleria mellonella larvae 36 and mouse infection models. Finally, we identify strong synergy between these 37 compounds and the antibiotic rifampicin, with a dose-sparing combination therapy 38 showing efficacy in both infection models. Our findings provide clear evidence that 39 these ruthenium complexes are effective antibacterial compounds against a critical 40 priority pathogen and show promise for the development of future therapeutics. 41 42 43 44 45 46 we show that and other similar complexes have potent antibacterial activity against laboratory and clinical isolates of Pseudomonas aeruginosa . We find that P. aeruginosa shows unique susceptibility to these compounds and identify mutations that decrease RuRed susceptibility. Finally, we show that these ruthenium complexes are effective as therapeutics in several in vivo infection models and synergize with the Gram-positive antibiotic rifampicin in dose- sparing combination treatments. Our work provides a foundation for further exploration of these ruthenium complexes as therapies to treat P. aeruginosa infections.


48
The antimicrobial resistance crisis is a major threat to global public health, and our 49 ability to treat serious bacterial infections remains tenuous. The pipeline of new drugs, 50 particularly for critical priority Gram-negative pathogens, is inadequate to meet the 51 realities facing modern healthcare (1,2). For numerous reasons, including economic   (Table S1). We note that, while RuRed has been used experimentally for 118 decades, its purity can vary between suppliers and is often not reported (25). Therefore, 119 the purity of the commercially available RuRed used in this work was assessed by 120 ultraviolet-visible (UV−vis) spectroscopy (Fig. S2). Our samples showed strong 121 absorbance at 533 nm, characteristic of RuRed, and a second smaller peak at 260 nm, 122 whose identity is not known but is often observed in commercial RuRed samples as a 123 minor impurity (25). We also observed a minor peak at 360 nm which likely corresponds 124 to the Ru360, a known impurity of RuRed preparations (26-28).  To further characterize the antibacterial activity of RuRed against P. aeruginosa, we 139 isolated spontaneous resistant mutants by plating the wild-type strain on an inhibitory 140 concentration of RuRed. Mutants recovered using this method had a 4-fold increased 141 MIC compared to the parental strain ( Fig. 2A). Next, we sequenced the genomes of four 142 mutants and the parental strain to identify any genetic differences. One of the four 143 suppressor mutants ("suppressor1") contained a mutation in the gene PA4394, which 144 encodes a predicted mechanosensitive ion channel involved in osmotic tolerance called   Expression of MscS-1 K73N , but not ColS T201P , in the wild-type background increased 156 resistance of P. aeruginosa to RuRed (Fig. 2C). Interestingly, overexpression of the 157 wild-type MscS-1 allele in the suppressor1 background did not restore susceptibility to RuRed (Fig. 2D), whereas overexpression of ColS WT was able to restore sensitivity in 159 the suppressor2 background (Fig. 2E). All constructs were sequence confirmed and 160 produced similar levels of recombinant protein (Fig. S3). Thus, these results were not 161 due to a lack of protein expression but instead suggest a more complex interaction 162 between the wild-type and mutant alleles for each gene. To further investigate the interactions of these alleles in RuRed susceptibility, we 175 obtained insertionally-inactivated mutants in mscS-1 and colS from the P. aeruginosa 176 MPAO1 transposon mutant collection (32). We reasoned that this would allow us to 177 determine the contribution of the wild-type or mutant alleles in a background devoid of a 178 functional chromosomal copy of the genes. Unlike the suppressor mutants we isolated, 179 both mscS-1 and colS transposon mutants had identical RuRed MICs to the parental 180 MPAO1 strain ( Figure 3A). This indicates that the RuRed resistance seen in our 181 suppressor mutants was not simply due to production of non-functional variants of these whereas expression of MscS-1 WT or the vector control had no effect. ColS T201P 185 expression had no effect in the wild-type background ( Figure 3D), but significantly   In addition to RuRed, several related ruthenium compounds have been investigated for 214 their ability to inhibit the MCU in eukaryotic cells (33,34). Given their structural and 215 functional similarity to RuRed, and our ability to synthesize pure compounds in-house, 216 we tested the antibacterial activity of the well-characterized dinuclear ruthenium  (Table S1). We also tested 6 mononuclear and 2 221 additional dinuclear Ru compounds, which had varying levels of antibacterial activity 222 against P. aeruginosa PAO1 (Fig. S4). Like RuRed, Ru265, Ru270, and Ru360' were 223 less effective against non-P. aeruginosa bacterial species (Table S1). One exception, 224 however, was the activity of Ru360' against Acinetobacter baumannii. All tested A. 225 baumannii, including common laboratory strains and clinical isolates, were susceptible 226 to Ru360' at concentrations similar to P. aeruginosa (Table S1).    Although our in vivo studies show that these ruthenium complexes show promise as 300 treatments for P. aeruginosa infection, it is also known that some of these compounds 301 can have toxic side effects (36, 37). A common approach to ameliorate toxic side effects 302 is to identify a partner compound that reduces the dose of the toxic molecule required  (Fig. 7A). Interestingly, this synergy also extended to E. coli (Fig. S6A).

310
One possible explanation for these synergies is that the ruthenium compounds 311 physically disrupt the outer membrane to allow entry of these antibiotics (40). To 312 investigate this, we used a lysozyme lysis assay where physical disruption of the E. coli 313 outer membrane allows lysozyme to reach the peptidoglycan layer and lyse the 314 bacterial cells. Treatment with the control compound SPR741 led to nearly 60% lysis, 315 whereas RuRed treatment resulted in no change in lysis compared to the no drug 316 control (Fig. S6B). Thus, although RuRed can potentiate large scaffold antibiotics, the 317 mechanism of synergy does not seem to be through outer membrane disruption.    remain to be determined. In our previous work, we found that RuRed had no activity 375 against K. pneumoniae in LB growth media but did have potent activity against K. 376 pneumoniae grown in human serum (17). Curiously, MHB media seems to enhance the 377 activity of RuRed, as we observed an MIC in this media even for K. pneumoniae. This 378 suggests that growth media formulation influences susceptibility to these compounds.

379
For example, serum is known to increase membrane permeability which enhances 380 penetration of large scaffold drugs normally excluded by the outer membrane (48). We 381 previously identified several K. pneumoniae gene knockouts that were susceptible to 382 RuRed in LB media, such as those involved in outer membrane biogenesis and DNA 383 repair pathways, and similar pathways have been shown to affect RuRed sensitivity in 384 E. coli, further supporting the condition-specific antibacterial efficacy of RuRed (17, 49).

385
It is possible that RuRed uses an entry mechanism into the cell that is constitutively 386 active in P. aeruginosa, but only induced in other species under certain conditions. In 387 any case, all bacterial species we tested were more resistant to the ruthenium 388 compounds than P. aeruginosa, which indicates that there is some intrinsic property of 389 P. aeruginosa that sensitizes them to these compounds. The one exception was our the negatively charged dinuclear compounds tested were inactive (see Figure S3).

396
However, a more thorough structure-activity-relationship campaign with a broader range 397 of compounds should be a priority for future work.   The bacterial strains used in this work are listed in Table S2. Strains were routinely 475 cultured and maintained in LB (per liter: 10 g trypticase peptone, 5 g yeast extract, 10 g   for the two compounds being tested (67). FICI values ≤ 0.5 were considered synergistic.

512
MICs in serum were done as above but P. aeruginosa PAO1 was inoculated into 50%