FR-31564, a new phosphonic acid antibiotic: bacterial resistance and membrane permeability

Growth medium-dependent antimicrobial activity of early stage MEP pathway inhibitors

The in vivo microenvironment of bacterial pathogens is often characterized by nutrient limitation. Consequently, conventional rich in vitro culture conditions used widely to evaluate antibacterial agents are often poorly predictive of in vivo activity, especially for agents targeting metabolic pathways. In one such pathway, the methylerythritol phosphate (MEP) pathway, which is essential for production of isoprenoids in bacterial pathogens, relatively little is known about the influence of growth environment on antibacterial properties of inhibitors targeting enzymes in this pathway. The early steps of the MEP pathway are catalyzed by 1-deoxy-d-xylulose 5-phosphate (DXP) synthase and reductoisomerase (IspC). The in vitro antibacterial efficacy of the DXP synthase inhibitor butylacetylphosphonate (BAP) was recently reported to be strongly dependent upon growth medium, with high potency observed under nutrient limitation and exceedingly weak activity in nutrient-rich conditions. In contrast, the well-known IspC inhibitor fosmidomycin has potent antibacterial activity in nutrient-rich conditions, but to date, its efficacy had not been explored under more relevant nutrient-limited conditions. The goal of this work was to thoroughly characterize the effects of BAP and fosmidomycin on bacterial cells under varied growth conditions. In this work, we show that activities of both inhibitors, alone and in combination, are strongly dependent upon growth medium, with differences in cellular uptake contributing to variance in potency of both agents. Fosmidomycin is dissimilar to BAP in that it displays relatively weaker activity in nutrient-limited compared to nutrient-rich conditions. Interestingly, while it has been generally accepted that fosmidomycin activity depends upon expression of the GlpT transporter, our results indicate for the first time that fosmidomycin can enter cells by an alternative mechanism under nutrient limitation. Finally, we show that the potency and relationship of the BAP-fosmidomycin combination also depends upon the growth medium, revealing a striking loss of BAP-fosmidomycin synergy under nutrient limitation. This change in BAP-fosmidomycin relationship suggests a shift in the metabolic and/or regulatory networks surrounding DXP accompanying the change in growth medium, the understanding of which could significantly impact targeting strategies against this pathway. More generally, our findings emphasize the importance of considering physiologically relevant growth conditions for predicting the antibacterial potential MEP pathway inhibitors and for studies of their intracellular targets.

Structure-Activity Relationships of the MEPicides: N-Acyl and O-Linked Analogs of FR900098 as Inhibitors of Dxr from Mycobacterium tuberculosis and Yersinia pestis

Despite continued research efforts, the threat of drug resistance from a variety of bacteria continues to plague clinical communities. Discovery and validation of novel biochemical targets will facilitate development of new drugs to combat these organisms. The methylerythritol phosphate (MEP) pathway to make isoprene units is a biosynthetic pathway essential to many bacteria. We and others have explored inhibitors of the MEP pathway as novel antibacterial agents. Mycobacterium tuberculosis, the causative agent of tuberculosis, and Yersinia pestis, resulting in the plague or "black death", both rely on the MEP pathway for isoprene production. 1-Deoxy-d-xylulose 5-phosphate reductoisomerase (Dxr) catalyzes the first committed step in the MEP pathway. We examined two series of Dxr inhibitors based on the parent structure of the retrohydroxamate natural product FR900098. The compounds contain either an extended N-acyl or O-linked alkyl/aryl group and are designed to act as bisubstrate inhibitors of the enzyme. While nearly all of the compounds inhibited both Mtb and Yp Dxr to some extent, compounds generally displayed more potent inhibition against the Yp homologue, with the best analogs displaying nanomolar IC50 values. In bacterial growth inhibition assays, the phosphonic acids generally resulted in poor antibacterial activity, likely a reflection of inadequate permeability. Accordingly, diethyl and dipivaloyloxymethyl (POM) prodrug esters of these compounds were made. While the added lipophilicity did not enhance Yersinia activity, the compounds showed significantly improved antitubercular activities. The most potent compounds have Mtb MIC values of 3-12 μg/mL. Taken together, we have uncovered two series of analogs that potently inhibit Dxr homologues from Mtb and Yp. These inhibitors of the MEP pathway, termed MEPicides, serve as leads for future analog development.

Illicit Transport via Dipeptide Transporter Dpp is Irrelevant to the Efficacy of Negamycin in Mouse Thigh Models of Escherichia coli Infection

Negamycin is a hydrophilic antimicrobial translation inhibitor that crosses the lipophilic inner membrane of Escherichia coli via at least two transport routes to reach its intracellular target. In a minimal salts medium, negamycin's peptidic nature allows illicit entry via a high-affinity route by hijacking the Dpp dipeptide transporter. Transport via a second, low-affinity route is energetically driven by the membrane potential, seemingly without the direct involvement of a transport protein. In mouse thigh models of E. coli infection, no evidence for Dpp-mediated transport of negamycin was found. The implication is that for the design of new negamycin-based analogs, the physicochemical properties required for cell entry via the low-affinity route need to be retained to achieve clinical success in the treatment of infectious diseases. Furthermore, clinical resistance to such analogs due to mutations affecting their ribosomal target or transport is expected to be rare and similar to that of aminoglycosides.

Profiling of defense responses in Escherichia coli treated with fosmidomycin

The mevalonate-independent isoprenoid biosynthesis pathway has been recognized as a promising target for designing new antibiotics. But pathogens treated with compounds such as fosmidomycin, a slow binding inhibitor of 1-deoxy-D-xylulose 5-phosphate reducto-isomerase, the second enzyme in this pathway, develop rapid drug resistance. In Escherichia coli, acquired resistance results mostly from inactivating the cAMP-dependent glpT transporter, thereby preventing import of the inhibitor. Such mutant strains are characterized by cross-resistance to fosfomycin, by susceptibility to efflux pump inhibitors, by disability to use glycerol 3-phosphate as a carbon source or by increased activity of the promoter controlling the expression of the glpABC regulon when grown in presence of fosmidomycin. The quite challenging task consists in conceiving new and efficient inhibitors avoiding resistance acquisition. They should be efficient in blocking the target enzyme, but should also be durably taken up by the organism. To address this issue, it is essential to characterize the mechanisms the pathogen exploits to defeat the antibiotic before resistance is acquired. Having this in mind, a 2-D Fluorescence Difference Gel Electrophoresis proteomic approach has been applied to identify defense responses in E. coli cells being shortly exposed to fosmidomycin (3 h). It seems that combined strategies are promptly induced. The major one consists in preventing toxic effects of the compound either by adapting metabolism and/or by getting rid of the molecule. The strategy adopted by the bacteria is to eliminate the drug from the cell or to increase the tolerance to oxidative stress. The design of new, but still efficient drugs, needs consideration of such rapid modulations required to adapt cell growth in contact of the inhibitor.

Antibacterial and antitubercular activity of fosmidomycin, FR900098, and their lipophilic analogs

The nonmevalonate pathway (NMP) of isoprene biosynthesis is an exciting new route toward novel antibiotic development. Inhibitors against several enzymes in this pathway are currently under examination. A significant liability of many of these agents is poor cell penetration. To overcome and improve our understanding of this problem, we have synthesized a series of lipophilic, prodrug analogs of fosmidomycin and FR900098, inhibitors of the NMP enzyme Dxr. Several of these compounds show improved antibacterial activity against a panel of organisms relative to the parent compound, including activity against Mycobacterium tuberculosis (Mtb). Our results show that this strategy can be an effective way for improving whole cell activity of NMP inhibitors.

Dxr is essential in Mycobacterium tuberculosis and fosmidomycin resistance is due to a lack of uptake

Fosmidomycin is a phosphonic antibiotic which inhibits 1-deoxy-D-xylulose 5-phosphate reductoisomerase (Dxr), the first committed step of the non-mevalonate pathway of isoprenoid biosynthesis. In Mycobacterium tuberculosis Dxr is encoded by Rv2870c, and although the antibiotic has been shown to inhibit the recombinant enzyme 1, mycobacteria are intrinsically resistant to fosmidomycin at the whole cell level. Fosmidomycin is a hydrophilic molecule and in many bacteria its uptake is an active process involving a cAMP dependent glycerol-3-phosphate transporter (GlpT). The fact that there is no glpT homologue in the M. tuberculosis genome and the highly impervious nature of the hydrophobic mycobacterial cell wall suggests that resistance may be due to a lack of cellular penetration.

RESULTS

We demonstrated that dxr (Rv2780c) is an essential gene in M. tuberculosis, since we could not delete the chromosomal copy unless a second functional copy was provided on an integrating vector. This confirmed that the intracellular target of fosmidomycin was essential as well as sensitive. We looked at the uptake of fosmidomycin in two mycobacterial species, the slow-growing pathogenic M. tuberculosis and the fast-growing, saprophytic Mycobacterium smegmatis; both species were resistant to fosmidomycin to a high level. Fosmidomycin was not accumulated intra-cellularly in M. tuberculosis or M. smegmatis but remained in the extra-cellular medium. In contrast, fosmidomycin uptake was confirmed in the sensitive organism, Escherichia coli. We established that the lack of intra-cellular accumulation was not due to efflux, since efflux pump inhibitors had no effect on fosmidomycin resistance. Finally, we demonstrated that fosmidomycin was not modified by mycobacterial cells or by extracts but remained in a fully functional state.

CONCLUSION

Taken together, these data demonstrate that fosmidomycin resistance in M. tuberculosis and M. smegmatis results from a lack of penetration of the antibiotic to the site of the sensitive target.

Double ester prodrugs of FR900098 display enhanced in-vitro antimalarial activity

Fosmidomycin and FR900098 are inhibitors of the 1-deoxy-D-xylulose 5-phosphate reductoisomerase (DXR; IspC), a key enzyme of the mevalonate-independent isoprenoid biosynthesis pathway. We have determined the in-vitro antimalarial activity of two double ester prodrugs 2, 3 in direct comparison with the unmodified FR900098 1 against intraerythrocytic forms of Plasmodium falciparum. Temporarily masking the polar properties of the phosphonate moiety of the DXR inhibitor FR900098 1 enhanced not only its oral bioavailability but also the intrinsic activity of this series against the parasites.

1-Deoxy-D-xylulose 5-phosphate reductoisomerase (IspC) from Mycobacterium tuberculosis: towards understanding mycobacterial resistance to fosmidomycin

1-Deoxy-d-xylulose 5-phosphate reductoisomerase (IspC) catalyzes the first committed step in the mevalonate-independent isopentenyl diphosphate biosynthetic pathway and is a potential drug target in some pathogenic bacteria. The antibiotic fosmidomycin has been shown to inhibit IspC in a number of organisms and is active against most gram-negative bacteria but not gram positives, including Mycobacterium tuberculosis, even though the mevalonate-independent pathway is the sole isopentenyl diphosphate biosynthetic pathway in this organism. Therefore, the enzymatic properties of recombinant IspC from M. tuberculosis were characterized. Rv2870c from M. tuberculosis converts 1-deoxy-d-xylulose 5-phosphate to 2-C-methyl-d-erythritol 4-phosphate in the presence of NADPH. The enzymatic activity is dependent on the presence of Mg(2+) ions and exhibits optimal activity between pH 7.5 and 7.9; the K(m) for 1-deoxyxylulose 5-phosphate was calculated to be 47.1 microM, and the K(m) for NADPH was 29.7 microM. The specificity constant of Rv2780c in the forward direction is 1.5 x 10(6) M(-1) min(-1), and the reaction is inhibited by fosmidomycin, with a 50% inhibitory concentration of 310 nM. In addition, Rv2870c complements an inactivated chromosomal copy of IspC in Salmonella enterica, and the complemented strain is sensitive to fosmidomycin. Thus, M. tuberculosis resistance to fosmidomycin is not due to intrinsic properties of Rv2870c, and the enzyme appears to be a valid drug target in this pathogen.

Fosmidomycin resistance in adenylate cyclase deficient (cya) mutants of Escherichia coli

Adenylate cyclase deficient (cya) mutants of E. coli K-12 were found to be resistant to fosmidomycin, a specific inhibitor of the non-mevalonate pathway, just like to fosfomycin. E. coli glpT mutants were resistant to fosfomycin and also to fosmidomycin. This fact shows that fosmidomycin was transported inside via the glycerol-3-phosphate transporter, GlpT. DNA micro-array analysis showed that the transcription of glpT and other genes concerning glycerol utilization were highly dependent on the presence of cAMP.

Cloning of a gene from Escherichia coli that confers resistance to fosmidomycin as a consequence of amplification

A gene conferring resistance to fosmidomycin (Fs) was cloned from the gene pool of a wild-type strain of Escherichia coli. The cloned DNA fragment was sequenced and shown to encode a putative polypeptide of 406 amino acids (aa) with a molecular weight of 43303. The gene mapped at 10.9 min on the E. coli chromosome and was designated fsr (fosmidomycin resistance). Maxicell analysis revealed that the Fsr protein migrated in sodium dodecyl sulfate-polyacrylamide-gel electrophoresis as a broad band of 35 kDa. A comparison between the aa sequence of Fsr and sequences in a protein database revealed 18% homology to the bacterial drug-export proteins that mediate resistance to tetracycline and chloramphenicol. Hydropathy analysis of the Fsr protein revealed twelve putative transmembrane segments. The degree of FsR of transformants depended on the number of copies of the plasmid that contained fsr. The levels of ubiquinone-8 and undecaprenyl phosphate in cells that harbored a high-copy-number plasmid that included fsr were almost the same as those in the cells without the plasmid. These results suggest that Fsr does not have any direct effect on the biosynthesis of isoprenoid in E. coli, and that the mechanism for FsR involves the efflux of the drug by a process that is facilitated by Fsr.

Activity of fosmidomycin in an in vitro model of the treatment of bacterial cystitis

The response to fosmidomycin of four strains of Escherichia coli was studied in an in vitro model of the treatment of bacterial cystitis. Three susceptible strains of E. coli responded well to relatively low concentrations of fosmidomycin: doses achieving peak concentrations of 50 or 250 mg/l suppressed bacterial growth for 13 h or more; however, when the surviving bacteria were challenged with a second dose, a reduced response was observed. When a fully resistant strain was exposed to fosmidomycin, bacterial growth was also suppressed for 13 h or more, even when the peak concentration achieved was below the conventionally determined minimum inhibitory concentration. Resistant variants which emerged after exposure to fosmidomycin were also resistant to fosfomycin in the absence of the potentiating agent, glucose-6-phosphate. In the presence of glucose-6-phosphate, complete (or partial) susceptibility to fosmidomycin and fosfomycin was retained by three of the four strains. These results suggest that fosmidomycin and fosfomycin are transported into E. coli by a similar mechanism, and that deletion of the hexose phosphate transport system does not occur following exposure to fosmidomycin in the absence of glucose-6-phosphate.

Comparison of the response of Escherichia coli to fosfomycin and fosmidomycin

The responses of Escherichia coli to fosfomycin and fosmidomycin were investigated by continuous turbidimetric monitoring of cultures exposed to the drugs and by microscopy. The activity of both agents was potentiated by glucose-6-phosphate, suggesting that they share the inducible hexose phosphate transport system in Escherichia coli, but several differences of response were also detected: the inoculum effect was much smaller with fosfomycin than with fosmidomycin; inhibition of bacterial growth occurred much more rapidly with fosfomycin than with fosmidomycin; and fosfomycin was able to induce the formation of spheroplasts much more rapidly than fosmidomycin. Stable resistance to fosfomycin and fosmidomycin was readily induced in cultures of Escherichia coli, and some resistant variants retained susceptibility (or partial suceptibility) to the other compound. These observations suggest that although fosfomycin and fosmidomycin may be transported into Escherichia coli by a similar mechanism, the intracellular target site may be different.

Synergy of fosmidomycin (FR-31564) and other antimicrobial agents

Fosmidomycin (FR-31564), a phosphonic acid derivative, was combined with cefazolin, cephalexin, ampicillin, carbenicillin, ticarcillin, gentamicin, and trimethoprim. Synergy between fosmidomycin and penicillins or cephalosporins was found for 37 to 52% of the Enterobacteriaceae tested. Synergy with trimethoprim was found against 55% of bacteria isolated, but only 17% of the strains showed synergy between formidomycin and gentamicin. Synergy between fosmidomycin and ticarcillin was shown for 35% of the Pseudomonas isolates. Cefazolin-, ampicillin-, and gentamicin-resistant isolates of various species were synergistically inhibited by fosmidomycin, as were ticarcillin- and gentamicin-resistant isolates. Antagonism was not encountered. This study illustrates another example of synergistic activity of compounds which attack different mechanisms in bacterial cells.

Fosfomycin resistance plasmids do not affect fosfomycin transport into Escherichia coli

Escherichia coli cells carrying fosfomycin resistance plasmids were able to take up fosfomycin from the medium to the same extent as plasmid-free bacteria. The antibiotic entered the plasmid-harboring cells by means of the glpT and uhp transport systems, as is the case with susceptible bacteria. Active fosfomycin could be detected in soluble extracts of cells which had previously been incubated in the presence of the antibiotic. Furthermore, fosfomycin resistance plasmids did not confer on E. coli cells resistance to the novel antibiotic FR-31564, which is incorporated by the same transport systems as fosfomycin. We conclude that, in contrast to chromosomal resistance mutants, altered transport does not play a role in the plasmid-encoded fosfomycin resistance mechanism.

Pharmacokinetics of fosmidomycin, a new phosphonic acid antibiotic

The pharmacokinetics of fosmidomycin was investigated in animals and humans after parenteral and oral dosing. In dogs the serum concentration was 54.8 microgram/ml at 0.25 h after an intravenous dose of 20 mg/kg, and the half-life was 1.14 h. Peak concentration was 41.4 microgram/ml after an intramuscular dose of 20 mg/kg and 16.6 microgram/ml after an oral dose of 40 mg/kg. In volunteers, the serum concentrations 0.25 h after dosing was 157 microgram/ml after an intravenous dose of 30 mg/kg, 12.3 microgram/ml after an intramuscular dose of 7.5 mg/kg, and 2.45 microgram/ml after an oral dose of 500 mg. More than 90% of the given dose was excreted in the 24-h urine in rats and dogs after parenteral dosing with 20 mg/kg. The 24-h urinary recovery was 45.8% of the given dose in rats after oral dosing with 100 mg/kg and 37.8% in dogs after oral dosing with 40 mg/kg. In volunteers 85.5% of the intravenous dose (30 mg/kg), 66.4% of the intramuscular dose (7.5 mg/kg), and 26.0% of the oral dose (500 mg) were excreted unchanged in the 24-h urine. In the multiple-dose study, there was no accumulation of fosmidomycin in the serum even after 21 consecutive intramuscular dosings of 1 g every 6 h or 29 consecutive 0.5-h drip infusions of 2 g every 6 h. Biliary excretion was extremely low in rats. Fosmidomycin was well distributed to the tissues of rats after parenteral and oral dosing. The lymph concentrations in dogs were nearly the same as serum concentrations. Serum protein binding was low (4% or less) to mouse, rat, dog, and human serum.

Pharmacokinetics and metabolism of fosmidomycin, a new phosphonic acid, in rats and dogs

The absorption, distribution, metabolism, and excretion of 3-(N-formylhydroxylamino) propylphosphonic acid monosodium salt (fosmidomycin), a new antibiotic, were investigated in rats and dogs after i.v. and oral dosing. After i.v. administration of 10 mg/kg of body weight, [14C]-fosmidomycin was excreted mainly in the urine (about 90% of dose within 72 h); and only a little was excreted in the expired air (14CO2) and bile of rats (less than 1% of dose), which suggested the absence of enterohepatic circulation. After oral administration of 10 mg/kg of body weight to rats, 34% and 61% of dose were excreted in the urine and faeces, respectively, suggesting about 30% gastro-intestinal absorption. No metabolites were found by autoradiography of the urine after thin layer chromatography. Radioactivity levels in the serum essentially agreed with the unchanged fosmidomycin levels determined by reverse isotope dilution method. [14C]-fosmidomycin was rapidly distributed in the tissues of rats, and was maintained in high concentration in the liver, kidneys, and bone. The serum level data after i.v. administration closely fitted a 3-compartment open model with first order kinetics after nonlinear least squares regression by NONLIN. The half-lives of the serum level curves for the early, midway, and terminal phases were: 0.13, 0.51, and 17.3 h, respectively in rats; and 0.44, 0.75, and 2.0 h, respectively in dogs.

In vitro and in vivo antibacterial activity of FR-31564, a phosphonic acid antimicrobial agent

The in vitro and in vivo activity of FR-31564 [sodium hydrogen 3-(N-hydroxyformamido)propylphosphate] against gram-positive and -negative aerobic and anaerobic bacteria was investigated and compared with that of fosfomycin, cephalexin, carbenicillin, and trimethoprim-sulfamethoxazole. The in vitro activity of FR-31564 was markedly enhanced when combined with glucose 6-phosphate or fructose 6-phosphate, but not when combined with ribose phosphate, adenosine monophosphate, or glycerol phosphate. In vitro activity of FR-31564 also was enhanced by human or horse blood, but not by human serum. The type of medium had a great effect on the minimal inhibitory concentration, with the lowest minimal inhibitory concentrations achieved on nutrient agar, 8- to 16-fold less than with Mueller-Hinton, heart infusion, or Trypticase soy agars. FR-31564 was more active than fosfomycin, cephalexin, carbenicillin, or trimethoprimsulfamethoxazole against Escherichia coli, Klebsiella pneumoniae, Proteus vulgaris, Enterobacter cloacae, E. aerogenes, and Citrobacter. It was less active than fosfomycin against Serratia marcescens and Proteus mirabilis and did not inhibit gram-positive cocci or anaerobic species. FR-31564 inhibited a number of E. coli, K. pneumoniae, and some Pseudomonas aeruginosa strains resistant to the other agents. In the presence and absence of human blood FR-31564 showed bactericidal activity, and P. aeruginosa exposed to FR-31564 for 3 h showed a 6-h lag in regrowth. FR-31564 administered by the subcutaneous route was more active in protecting mice challenged with P. aeruginosa than was fosfomycin, carbenicillin, or cefoperazone. It was as active by the oral route in protecting mice challenged with E. coli as was fosfomycin, ampicillin, cephalexin, or trimethoprimsulfamethoxazole.