Effect of methylglyoxal on multidrug-resistant Pseudomonas aeruginosa

Honey has a complex chemistry, and its broad-spectrum antimicrobial activity varies with floral source, climate, and harvesting conditions. Methylglyoxal was identified as the dominant antibacterial component of manuka honey. Although it has been known that methylglyoxal has antibacterial activity against gram-positive bacteria, including methicillin-resistant Staphylococcus aureus and vancomycin-resistant Enterococcus, there is not much information describing its activity against gram-negative bacteria. In this study, we report the effect of methylglyoxal against multidrug-resistant Pseudomonas aeruginosa (MDRP) using 53 clinically isolated strains. We also assessed the effect of deleting the five multidrug efflux systems in P. aeruginosa, as well as the efflux systems in Escherichia coli and Salmonella enterica serovar Typhimurium, on MICs of methylglyoxal. Our results indicate that methylglyoxal inhibits the growth of MDRP at concentrations of 128–512 μg/ml (1.7–7.1 mM) and is not recognized by drug efflux systems.


INTRODUCTION
Pseudomonas aeruginosa is endemic among critically ill patients, and multidrug-resistant strains are increasingly being isolated in intensive care units (Ortega et al., 2004). Because P. aeruginosa is a virulent organism susceptible to a limited number of antibiotic agents, infections caused by this organism are difficult to cure and often require combination therapy. Multidrug-resistant P. aeruginosa (MDRP) has been defined as P. aeruginosa resistant to imipenem, amikacin, and ciprofloxacin (Sekiguchi et al., 2007). The increasing resistance of P. aeruginosa is a growing threat to the clinical management of such infections (Ortega et al., 2004).
In bacteria, resistance to bactericidal agents is often associated with multidrug efflux systems, which decrease cellular drug accumulation (Nikaido, 1996). In gram-negative bacteria, systems belonging to the resistance/nodulation/division (RND) family are particularly effective in generating resistance because they form a tripartite complex with the periplasmic proteins of the membrane fusion protein family and an outer membrane channel, ensuring that drugs are pumped out directly to the external medium (Nikaido and Pages, 2012). P. aeruginosa expresses several RND-type multidrug efflux systems, including MexAB-OprM, MexCD-OprJ, MexEF-OprN, and MexXY, which are significant determinants of multidrug resistance in laboratory and clinical isolates (Poole, 2004;Piddock, 2006;Lister et al., 2009). These systems are three-component systems comprising antiporters of the RND family driven by proton motive force (MexB, MexD, MexF, and MexY), outer membrane channels (OprM, OprJ, and OprN), and periplasmic membrane fusion proteins (MexA, MexC, MexF, and MexX). These pumps function in a manner similar to AcrAB-TolC, which is the best-studied RND-type multidrug pump of Escherichia coli (Nakashima et al., 2011;Nikaido, 2011). It is necessary to develop drugs that are not recognized by the efflux pumps to prevent multidrug resistance modulated by drug efflux systems.
Honey has several antibacterial features that are distinct from classical antibiotics, including high osmolarity, low pH, and generation of hydrogen peroxide by the bee-derived enzyme glucose oxidase (Allen et al., 1991). Antibacterial phenolic components have been identified in honey (Weston et al., 1999), and an antimicrobial peptide has been discovered in a Dutch medicalgrade honey produced from an undisclosed floral source cultivated in greenhouses (Kwakman et al., 2010). Manuka honey is derived from nectar that has been collected by honey bees (Apis mellifera) foraging on a shrub known as manuka (Leptospermum scoparium) that is indigenous to New Zealand. Manuka honey is broad in spectrum, able to inhibit a diverse range of bacterial and yeast pathogens, and equally effective against multidrug-resistant bacteria (Blair et al., 2009;Henriques et al., 2010;Kwakman et al., 2011). It is used in modern wound-care formulations and has been shown to eradicate methicillin-resistant Staphylococcus aureus (MRSA) from wounds (Natarajan et al., 2001;Blaser et al., 2007;Gethin and Cowman, 2008;Visavadia et al., 2008). Clinically isolated strains of methicillin-susceptible and -resistant staphylococci were shown to be equally susceptible to manuka honey in vitro, with minimum inhibitory concentrations (MICs) reported to be <3% (v/v) [equivalent to 41,000 mg/L or 4.1% (w/v)] (Cooper et al., 1999(Cooper et al., , 2002b. Methylglyoxal was identified as the dominant active antibacterial component of manuka honey (Mavric et al., 2008;Adams et al., 2009b). Active manuka honey contains high levels of the reactive dicarbonyl methylglyoxal (Mavric et al., 2008;Adams et al., 2009a), which is formed nonenzymatically from nectar-derived dihydroxyacetone during ripening. Methylglyoxal was also found to be produced from dihydrocyacetone phosphate in E. coli, initiating a bypass of the glycolytic pathway (Cooper and Anderson, 1970). It was suggested that methylglyoxal inhibits protein synthesis by reacting with guanine residues in RNA and its precursors. It also inhibits DNA synthesis by reacting with guanine residues in DNA and its precursors (Krymkiewicz et al., 1971).
It has been known that methylglyoxal has antibacterial activity against gram-positive bacteria, including MRSA and vancomycin-resistant Enterococcus. It was also reported that methylglyoxal containing manuka honey is biocidal against S. aureus strains at a concentration of 33-66% w/v (equivalent methylglyoxal concentration, 260-530 µg/ml) (Jervis- Bardy et al., 2011). However, there is not much information describing methylglyoxal activity against gram-negative bacteria. Although it was previously reported that manuka honey is bactericidal against P. aeruginosa (Roberts et al., 2012), the effect of methylglyoxal on MDRP has been unknown. In this study, we report the antibacterial effect of methylglyoxal on MDRP using 53 clinically isolated strains. We also demonstrate that methylglyoxal is not recognized by drug efflux systems in P. aeruginosa, Salmonella enterica, and E. coli.

BACTERIAL STRAINS AND GROWTH CONDITIONS
The bacterial strains used in this study are listed in Table 1. We used P. aeruginosa PAO1 (Stover et al., 2000), S. enterica serovar Typhimurium ATCC14028s (Fields et al., 1986), and E. coli MG1655 (Blattner et al., 1997) as wild-type strains. All clinically isolated MDRP strains, which showed resistance to imipenem, amikacin, and ciprofloxacin, were kindly provided by Biomedical Laboratories, Inc. (Tokyo, Japan).

SUSCEPTIBILITIES OF MDRP STRAINS TO METHYLGLYOXAL
To evaluate the antibacterial activity of methylglyoxal against clinically isolated MDRP strains, we determined MICs using the 53 confirmed MDRP strains. The MIC of methylglyoxal for most of the MDRP strains was 512 µg/ml (Table 2), whereas the susceptibilities of these strains to imipenem, amikacin, and ciprofloxacin were different. The methylglyoxal concentration at which MDRP14 was susceptible was 128 µg/ml and that at which MDRP4, 5, 41, and 50 were susceptible was 256 µg/ml. We also tested the methylglyoxal susceptibility of the drug-sensitive wildtype strain P. aeruginosa PAO1. The MIC of methylglyoxal for PAO1 was 512 µg/ml (Table 3), which was the same that for most of the MDRP strains.

EFFECT OF DRUG EFFLUX SYSTEMS IN P. aeruginosa, E. coli, AND S. enterica TO METHYLGLYOXAL
Multidrug efflux pumps in P. aeruginosa, such as MexAB-OprM, MexCD-OprJ, MexEF-OprN, and MexXY, have been shown to be significant determinants of multidrug resistance in laboratory and clinical isolates (Poole, 2004;Piddock, 2006;Lister et al., 2009). The existence of an additional multidrug efflux system, MexHI-OpmD, was also previously reported (Sekiya et al., 2003) in this organism. Because multidrug efflux systems display the ability to transport various structurally unrelated drugs, we investigated whether methylglyoxal is exported by these drug efflux systems in P. aeruginosa. For this purpose, we measured MIC of methylglyoxal for the wild-type P. aeruginosa strain PAO1 and its efflux-deficient mutant strain PMX52 ( mexAB oprM mexCD oprJ mexEF oprN mexXY mexHI opmD). Although PMX52 was more susceptible to amikacin and ciprofloxacin than PAO1, MIC of methylglyoxal for these strains was the same. This suggests that methylglyoxal is not recognized by drug efflux systems in P. aeruginosa. To confirm whether same phenomenon could be observed in other gram-negative bacteria, we determined MICs of methylglyoxal for the efflux-deficient mutants of E. coli and S. enterica serovar Typhimurium. There are five RND-type drug efflux systems (AcrAB, AcrD, MdtABC, MdtEF, and AcrEF) in E. coli, and all of them require the TolC outer membrane channel for their function (Nishino et al., 2003). To investigate the defect of these drug efflux systems in E. coli, we measured MICs of methylglyoxal for MG1655 (wild-type), NKE1329 ( acrB acrD mdtABC mdtEF acrEF), and NKE95 ( tolC) strains. The susceptibility of NKE1329 and NKE95 to methylglyoxal was same as that of the wild-type strain, although they were more susceptible to ciprofloxacin than the wild-type strain. S. enterica serovar Typhimurium harbors at least nine drug efflux systems belonging to RND, multidrug and toxic compound extrusion, and ATP-binding cassette (ABC) superfamilies (Nishino et al., 2006). Seven of them (AcrAB, AcrEF, AcrD, MdtABC, MdsAbC, EmrAB, and MacAB) require TolC for their function (Horiyama et al., 2010). For S. enterica, we used ATCC14028s (wild-type), NKS196 ( acrAB acrEF acrD mdtABC mdsABC emrAB mdfA mdtK macAB), and NKS233 ( tolC) strains. Although NKS196 and NKS233 were more sensitive to ciprofloxacin than the wildtype strain ATCC14028s, MICs of methylglyoxal for ATCC14028s, NKS196, and NKS233 were the same. In addition to MIC determination using agar plates, we tested the effect of methylglyoxal on bacterial growth in liquid medium. The growth of E. coli (MG1655, NKE1329, and NKE9) and Salmonella (ATCC14028s, NKS196, and NKS233) strains was inhibited by methylglyoxal at a concentration of 256 µg/ml, and the growth of P. aeruginosa (PAO1 and PMX52) strains was inhibited at 512 µg/ml, which is consistent with MICs determined (Figure 1). These data suggest that methylglyoxal is not recognized by drug efflux systems in E. coli or S. enterica.
In this study, we showed that methylglyoxal equally inhibits drug-susceptible P. aeruginosa and MDRP at concentrations of 128-512 µg/ml (1.7-7.1 mM). Methylglyoxal is a key antimicrobial component of manuka honey, and manuka honey has previously been suggested as a topical treatment option for burn patients infected with P. aeruginosa (Cooper et al., 2002a). Jenkins and Cooper reported that MICs of manuka honey for MRSA and methicillin-resistant P. aeruginosa were 6-7% w/v (Jenkins and Cooper, 2012). This corresponds to 50-100 µg/ml methylglyoxal when manuka honey contains 7% of methylglyoxal. Cooper et al. also reported that MIC for E. coli is 16% w/v , which corresponds to approximately 200 µg/ml methylglyoxal. It was previously reported that methylglyoxal is the dominant antibacterial constituent of manuka honey and that MIC of methylglyoxal for E. coli and S. aureus, determined using the agar well diffusion assay, is 1.1 mM (79.3 µg/ml) (Mavric et al., 2008). Our data showed that methylglyoxal itself inhibits the growth of MDRP strains at high concentrations, suggesting that methylglyoxal activity might be enhanced when in honey solution. Further research is required to demonstrate whether methylglyoxal and manuka honey exert their antibacterial effects through a common mechanism. We also showed that methylglyoxal is not recognized by drug efflux systems in P. aeruginosa, E. coli, and S. enterica.