INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Aminoglycosides are among the few classes of antibiotics that are useful in the treatment of infections caused by Pseudomonas aeruginosa. Antagonism of aminoglycosides by divalent cations is well documented for this organism (6, 21, 41) and is regarded as one of the problems in aminoglycoside therapy. It is a general understanding that (i) divalent cations interfere with uptake of aminoglycosides, and (ii) this interference occurs at both outer and inner membranes (3, 5, 41). It was demonstrated that polycationic aminoglycosides interact with divalent cation binding sites positioned on cell surface lipopolysaccharides (1, 7). Since aminoglycosides are much larger than the native divalent cations that normally stabilize the outer membrane, their binding causes a disruption which apparently results in permeabilization of the outer membrane, facilitating uptake of aminoglycosides in the periplasmic space. Accordingly, uptake across the outer membrane was termed "self-promoted uptake" (9-11). It was suggested that divalent cations might prevent (antagonize) self-promoted uptake by competing with aminoglycosides for the binding to lipopolysaccharides (9-11).

Entry of aminoglycosides into the cytoplasm across the inner membrane involves an energy-dependent transport (3, 4). It is presumed that cations may antagonize this transport by competing with antibiotics for binding to some membrane component (3). It was also shown that cations appeared to inhibit uptake of aminoglycosides in Staphylococcus aureus and in spheroplasts of Escherichia coli, confirming the hypothesis that the inner membrane may indeed be a site for cation antagonism (3).

Until recently, aminoglycosides were among the rare classes of antibiotics for which no active extrusion due to activity of multidrug resistance (MDR) pumps (26-28, 30) had been demonstrated. This was changed by the discovery of the MDR pump AmrAB-OprA in Burkholderia pseudomallei (23). It has been shown that inactivation of either amrA or amrB resulted in a significant increase in susceptibility to various aminoglycosides. Like many previously discovered MDR pumps from gram-negative bacteria, AmrAB-OprA appears to be a three-component structure containing a transporter (AmrB) located in the cytoplasmic membrane, an outer membrane channel (OprA), and a periplasmic linker protein (AmrA), which is thought to bring the other two components into contact (37-39). This structural organization allows extrusion of substrates, such as aminoglycosides, directly into the external medium, thus bypassing the periplasm (37). More recently, two other aminoglycoside pumps were discovered: single-component AcrD in E. coli (35) and tripartite MexXY-OprM in P. aeruginosa (2, 36). In addition to aminoglycosides, the latter pump can also extrude fluoroquinolones, macrolides, and tetracyclines (2, 20). Thus, MexXY-OprM has become the fourth MDR pump to be identified in P. aeruginosa (30). Interestingly, this tripartite pump employs as its outer membrane component the protein OprM, which was previously found to function as an essential part of another MDR system, MexAB-OprM (31, 32). The gene oprM is located immediately downstream of the mexAB genes and can be transcribed either together with them or independently (33, 40). In the latter case, transcription is initiated from the promoter located inside the mexB sequence. In contrast to the mexAB genes, the mexXY genes are not located near oprM in the chromosome.

The mexXY operon appears to be negatively regulated by the upstream, divergently transcribed gene mexZ (also called amrR), since insertional inactivation of mexZ results in the increased expression of mexXY (36). It also appears that additional genes may be involved in the regulation of mexXY since several reported spontaneous aminoglycoside resistant mutants with elevated expression of mexXY isolated under laboratory conditions did not contain mutations in mexZ (36).

Importantly, the clinical relevance of the MexXY-OprM-mediated efflux of aminoglycosides has been established: elevated expression of the mexXY genes has been detected in clinical isolates of P. aeruginosa exhibiting the impermeability-type resistance, which can be defined as panaminoglycoside resistance in the absence of modifying enzymes (36). It is noteworthy that in the case of P. aeruginosa, impermeability is the most common mechanism of resistance to these antibiotics (22). While it appears that the impermeability type of resistance is a multifactorial phenomenon, the MexXY-OprM efflux pump may prove to be its essential component.

As is evident from the facts described above, multiple intrinsic mechanisms affecting susceptibility to aminoglycosides may coexist in a single cell of P. aeruginosa. In our study we sought to investigate interplay between decreased uptake and increased efflux. To do so, we studied the effects of divalent cations, Mg²⁺ and Ca²⁺, on susceptibility to aminoglycosides in the wild-type strain of P. aeruginosa and in mutants overexpressing or lacking the MexXY-OprM efflux pump. Unexpectedly, antagonism of aminoglycosides by divalent cations was seen only in the strains of P. aeruginosa that contained the functional MexXY-OprM efflux pump. We have concluded that such antagonism occurs only in the presence of active efflux of aminoglycosides.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and media. All strains used in this study (Table 1) are derivatives of PAM1020 (17). The laboratory bacterial strains used in this study are listed in Table 1. Bacterial cells were grown in L broth (1% [wt/vol] tryptone, 0.5% [wt/vol] yeast extract, 0.5% [wt/vol] NaCl) or L agar (L broth plus 1.5% agar) at 37°C. Levofloxacin was synthesized at Daiichi Pharmaceutical Co., Ltd. (Tokyo, Japan). Arbekacin was from Meiji Seika (Osaka, Japan), and netilmicin was a gift from Schering-Plough. MC-207,110 was from the synthetic compound library of Microcide Pharmaceuticals, Inc. MC-005,556 (alanine -naphthylamide [Ala-Nap]) was synthesized at Microcide Pharmaceuticals, Inc.

MIC determinations. MIC determinations were carried out in 96-well microtiter plates using a twofold standard broth microdilution method (25) in Mueller-Hinton broth (Difco). Inocula in all experiments were 10⁴ to 10⁵ cells/ml. Interactions between antibiotics and Mg²⁺, Ca²⁺, or MC-207,110 were assessed using twofold dilution schemes in both directions. Aminoglycosides or levofloxacin was dispensed alone in the first row and was combined with either cations or MC-207,110 in the remaining rows. Cations or MC-207,110 was also dispensed alone in the first column. The highest concentrations of both divalent cations and MC-207,110 were 8 mM and 64 µg/ml, respectively. Results are presented for the four concentrations of cations and MC-207,110.

MC-005,556 (Ala-Nap) uptake assays. Ala-Nap, which is not fluorescent in solution, is cleaved enzymatically inside the cells to produce highly fluorescent -naphthylamine (18, 34). The more Ala-Nap that enters the cells, the more fluorescence is produced. To assess uptake of Ala-Nap, cultures of P. aeruginosa were grown in L broth to an optical density at 600 nm of ~1, washed, and resuspended in buffer at pH 7.0 containing 50 mM K₂HPO₄ and 0.4% glucose. Assays were performed in 96-well flat-bottom black plates (Applied Scientific or Costar) in a final volume of 200 µl and were initiated by the addition of Ala-Nap to suspensions of intact cells to a final concentration of 128 µg/ml. Fluorescence was measured on an fMAX spectrofluorometer (Molecular Devices) using an excitation wavelength of 320 nm and an emission wavelength of 460 nm. To measure the effects of Mg²⁺ (MgSO₄) on the rate of Ala-Nap uptake, cells were preincubated with different concentrations of Mg²⁺ prior to the addition of Ala-Nap.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Creation of isogenic strains containing or lacking a functioning MexXY-OprM efflux pump. We have constructed a set of isogenic strains of Pseudomonas aeruginosa expressing various levels of the MexXY-OprM efflux pump. All strains have been derived from the wild-type strain PAM1020 (17). The strains PAM2378 and PAM1154 contained mexX::Hg and oprM::Hg insertions, respectively. As expected, both mutants demonstrated increased susceptibility to aminoglycosides compared to PAM1020 (Table 2), indicating that the MexXY-OprM efflux pump in the mutants was indeed nonfunctional.

Effect of divalent cations on susceptibility to aminoglycosides in strains containing or lacking a functioning MexXY-OprM efflux pump. Next, we determined the impact of various concentrations of Mg²⁺ and Ca²⁺ on the MICs of gentamicin, streptomycin, amikacin, apramycin, netilmicin, and arbekacin in the strains containing or lacking a functioning MexXY-OprM efflux pump. We unexpectedly found that both Mg²⁺ and Ca²⁺ antagonized all tested aminoglycosides only in the strains of P. aeruginosa that contained a functioning MexXY-OprM efflux pump: almost no antagonism was seen at concentrations of Mg²⁺ and Ca²⁺ up to 8 mM for the mexX::Hg or oprM::Hg derivatives of any of the mutants (Tables 2 and 3). This was the case regardless of other mutations present in these strains. For example, in the strain PAM1415, an increase in resistance to aminoglycosides was not due to elevated production of MexXY-OprM. Still, introduction of either mexX::Hg or oprM::Hg in PAM1415 prevented Mg²⁺ antagonism (Table 2). Deleting mexX::Hg or oprM::Hg was also sufficient to prevent Mg²⁺ and Ca²⁺ antagonism in PAM1147.

Effect of Mg²⁺ on uptake of Ala-Nap (MC-005,556) in strains containing or lacking a functioning MexXY-OprM efflux pump. It was important to know whether a functioning MexXY-OprM pump was required for Mg²⁺ effects on other substrates of this pump. Fluoroquinolones and tetracycline are among these substrates (2, 20); however, in these cases the interpretation of Mg²⁺ effects could be difficult, since these compounds are known to chelate divalent cations (13, 14).

View larger version (26K): [in this window] [in a new window]   FIG. 1.   Uptake of MC-005,556 (Ala-Nap) with no Mg2+ added by cells of P. aeruginosa overexpressing or lacking the MexXY-OprM efflux pump. Ala-Nap was added to intact cells of P. aeruginosa either overexpressing MexXY-OprM (PAM2394, open symbols) or lacking this pump (PAM2380, filled symbols). A similar linear increase in fluorescence with time due to intracellular hydrolysis of MC-005,556 (Ala-Nap) is seen over a wide range of Ala-Nap concentrations (32 [circles], 64 [diamonds], 128 [triangles], and 256 [squares] µg/ml) for both cell types.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Uptake of MC-005,556 in the presence of 4 mM Mg2+ by cells of P. aeruginosa overexpressing or lacking the MexXY-OprM efflux pump. Differential rates of uptake of Ala-Nap between cells overexpressing the pump (PAM2394) and those lacking it (PAM2380) are seen.

View larger version (18K): [in this window] [in a new window]   FIG. 3.   Effect of Mg2+ on uptake of MC-005,556 in intact cells of P. aeruginosa overexpressing or lacking the MexXY-OprM efflux pump. (A) Decrease in rates of fluorescence production by cells of P. aeruginosa overexpressing the MexXY-OprM pump (PAM2394) is seen in the presence of different concentrations of Mg2+. (B) No effect of Mg2+ is observed for cells of PAM2380, which lack constitutively expressed MexAB-OprM and MexXY-OprM efflux pumps.

Effect of MC-207,110 on susceptibility to aminoglycosides in strains containing or lacking a functioning MexXY-OprM efflux pump. MC-207,110 has been recently identified as an inhibitor of MexAB-OprM, MexCD-OprJ, and MexEF-OprN efflux pumps from P. aeruginosa and the AcrAB-TolC pump from E. coli (18, 34). This compound has been shown to potentiate multiple antibiotics, which are substrates of these pumps.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This work was inspired by our long-standing interest in relationships between different mechanisms of resistance to a particular antibiotic that can coexist in the same bacterial cell. Studies performed both by us and by others demonstrated that the combined presence of two mechanisms might confer either multiplicative (the fold increase in MIC due to combined presence is close to the product of individual fold increases) or additive (the fold increase in MIC is a sum of the individual fold increases) effects on drug resistance. Examples of the multiplicative case include interplay between efflux-based and target-mediated resistance to fluoroquinolones in P. aeruginosa (17) and E. coli (29) and between multicomponent and single-component efflux pumps in resistance to antibiotics, which are substrates of these pumps (15). The additive case is exemplified by the interplay between efflux pumps and -lactamases in resistance to various -lactam antibiotics (19, 24). In this particular study we sought to investigate the interplay between increased efflux and decreased uptake in resistance to aminoglycoside antibiotics in P. aeruginosa. It is noteworthy that active extrusion of aminoglycosides from bacterial cells was discovered quite recently: tripartite pumps, AmrAB-OprA (23) and MexXY-OprM (2, 36), were identified in B. pseudomallei and P. aeruginosa, respectively, and a single-component pump, AcrD, was found in E. coli (35).

To determine the effects of decreased permeability on the observed susceptibility to aminoglycosides in the strains either lacking or containing the functional MexXY-OprM pump, we decided to vary the concentration of divalent cations in the test media. This approach was based on numerous literature reports indicating that the well-known antagonistic effect between divalent cations, such as Mg²⁺ and Ca²⁺, and aminoglycosides occurs due to the interference of cations with the uptake of aminoglycosides at the level of both outer and inner membranes (3, 5, 41). Our results demonstrated that divalent cations antagonized aminoglycosides only when the MexXY-OprM efflux pump was functional. An analogous effect was seen in accumulation assays with a different substrate of the MexXY-OprM pump, Ala-Nap. Like Mg²⁺, MC-207,110, a cationic dipeptide which was previously identified as an inhibitor of Mex pumps (18, 34), antagonized aminoglycosides in a MexXY-OprM-dependent manner.

There seem to be two different explanations for the requirement of an efflux pump for the antagonism between divalent metals (or MC-207,110) and aminoglycosides. The simplest possibility is that divalent metals and the cationic compound MC-207,110 compete with cationic aminoglycosides for binding at the outer membrane, and inhibit antibiotic penetration into the periplasm. However, this inhibition of antibiotic penetration is observed only when an opposing active efflux via an MDR pump is present. In the absence of an MDR pump, the rate of accumulation of the aminoglycoside can be decreased by another cation, but this will not affect the final concentration of the antibiotic in the cytoplasm and will have no effect on the MIC. Indeed, the outer membrane barrier is functional only in the presence of MDR pumps that extrude compounds across this barrier. For example, inhibition of efflux pumps or their mutational inactivation renders cells hypersensitive to antibiotics even in the presence of an intact outer membrane. The opposite is also true. In the case of many compounds that are substrates of Mex pumps, permeabilization of the outer membrane of P. aeruginosa dramatically increases the susceptibility (16) and/or increases the rates of uptake of these compounds (8), regardless of the presence of functioning efflux pumps. These results confirmed the now generally accepted paradigm of a synergistic interaction between increased efflux due to tripartite multidrug resistance pumps, extruding antibiotics in the external medium, and decreased uptake though the outer membrane to confer resistance to some antibiotics (37).

It is noteworthy that in E. coli, which possesses its own aminoglycoside pump, AcrD, antagonism of aminoglycosides by 5 mM Mg²⁺ was not increased in the presence of AcrD compared to that in its absence (35). AcrD most probably is a single-component pump, extruding aminoglycosides in the periplasmic space. If antagonism occurs mainly at the level of the outer membrane, then one should not expect that such a pump would be essential for the manifestation of an antagonistic effect.

Another possibility is that divalent metals or MC-207,110 directly activates efflux of aminoglycosides and Ala-Nap by MexXY-OprM. In the case of Ala-Nap, we measured the effect of Mg²⁺ on its uptake kinetics. Interestingly, we did not detect the MexXY-OprM-dependent efflux of Ala-Nap when no external Mg²⁺ was added to the assay buffer: strains lacking and expressing MexXY-OprM converted Ala-Nap at the same rate. The difference between the two strains became apparent in the presence of Mg²⁺, but only because it decreased the apparent initial rate of Ala-Nap uptake in the strain containing the MexXY-OprM pump: no Mg²⁺ effect was seen in the absence of the pump.

To assess the possibility that Mg²⁺ can directly facilitate efflux of aminoglycosides, we are planning to study the effects of Mg²⁺ on accumulation kinetics of aminoglycosides using the strains lacking or overexpressing MexXY-OprM. We are also attempting to isolate mutants containing changes in the pump genes resulting in an altered response to Mg²⁺.

From a practical point of view, our findings indicate that inhibition of the MexXY-OprM efflux pump from P. aeruginosa should abolish antagonism of aminoglycosides by divalent cations regardless of its precise mechanism. This should significantly improve the clinical utility of this important class of antibiotics. We attempted to demonstrate this directly by using our recently identified efflux pump inhibitor, MC-207,110. Unfortunately, instead of synergy, we observed an antagonistic effect of MC-207,110, most probably due to a similar mechanism to that for divalent cations. Our studies indicate that identifying a MexXY-OprM inhibitor that will block the efflux of aminoglycosides will be of significant therapeutic value.