Discovery at the VICB
Fighting Drug Resistance in Tuberculosis
By: Carol A. Rouzer, VICB Communications
Published: February 2, 2016
Studies of the interaction of fluoroquinolone antibacterial agents with the gyrase from M. tuberculosis reveal new insights into fighting drug resistance.
Tuberculosis is a major public health problem with 9 million new cases diagnosed and 1.6 million deaths worldwide in 2013. Contributing to this toll of morbidity and mortality is the increasing resistance of the causative organism, Mycobacterium tuberculosis, to the four-drug combination currently used as the primary mode of therapy. An important tool in the fight against drug-resistant M. tuberculosis is the fluoroquinolone class of broad-spectrum antibacterials. These drugs target bacterial type II topoisomerases, leading to DNA strand breaks and resultant genotoxicity. Most bacterial species encode two type II topoisomerases: a DNA gyrase that controls the density of supercoiling and topoisomerase IV that untangles the DNA. M. tuberculosis, however, encodes only a DNA gyrase, which carries out both functions. Fluoroquinolones have exhibited high efficacy in treating tuberculosis to date, but resistant strains of M. tuberculosis have begun to appear. This led Vanderbilt Institute of Chemical Biology member Neil Osheroff and his colleague James Berger (Johns Hopkins University) to seek a better understanding of the interaction of fluoroquinolones with M. tuberculosis DNA gyrase and to design novel molecules that target resistant strains of the bacterium [K. J. Aldred et al., (2016) Proc. Natl. Acad. Sci. U.S.A., published online January 20, DOI:10.1073/pnas.1525055113].
DNA gyrase is a heterotetramer comprising two subunits of GyrA and two of GyrB (Figure 1). Together they form a complex containing two gates, designated N and C. The catalytic cycle of the enzyme begins with binding a segment of DNA referred to as the G (or Gate) segment, which enters through the N gate. The enzyme then wraps the DNA around and binds a second segment, the T (or Transport) segment, also via the N gate. GyrA then cleaves the G segment and retains the two ends covalently bound to a critical active site tyrosine residue in each of the two subunits. Meanwhile, GyrB binds and cleaves two molecules of ATP, which provides the energy required for the enzyme to pass the T segment of DNA through the break in the G segment and out through the C gate. Finally, the enzyme rejoins the ends of the G segment. Fluoroquinolones, which are classified as topoisomerase II poisons, bind to the enzyme-DNA complex and prevent rejoining of the G segment. The result is an accumulation of DNA breaks, leading eventually to replicative stress and cell death.
FIGURE 1. Structure and mechanism of DNA gyrase. (a) The enzyme is a heterotetramer of two monomers, GyrA (blue) and GyrB (yellow). The N-terminal domains of GyrA interact with DNA, and together with GyrB, form the C gate. The C-terminal domains (light blue spheres) are attached to the N-terminal domains by a flexible linker that allows them to rotate so they can bind and wrap DNA. GyrB binds and hydrolyzes ATP and forms the N-gate. (b) Mechanism for (-) supercoiling by gyrase. The gyrase first binds a segment of DNA, referred to as the G (or Gate) segment, which enters through the N gate (1). The enzyme then wraps the DNA so that a second segment (the T or Transport segment) also enters through the N gate (2). Two molecules of ATP are bound (3). Hydrolysis of ATP then drives cleavage of the G segment and passage of the T segment through the G segment and out through the C gate (4). Then, the G segment is religated (5). Image reproduced by permission from Macmillan Publishers Ltd, from M. Nöllmann, et al., (2007) Nat. Struct. Mol. Biol., 14, 264. Copyright 2007.
Prior work had shown that fluoroquinolones bind to most type II topoisomerases through a bridge formed between a metal ion complexed to the C3/C4 ketoacid functionality of the drug (Figure 2) and four water molecules, two of which bind to a serine and an acidic residue in the GyrA active site (Figure 3). Mutation of the residues involved in water-metal ion bridge formation is a primary mechanism by which bacteria acquire fluoroquinolone resistance. The gyrase of M. tuberculosis is unusual, however, in that it possesses an alanine (A90) rather than serine at one of the sites predicted to be involved in bridge formation. This led the Osheroff and Berger lab investigators to question whether a water-metal ion bridge plays a role in fluoroquinolone binding to the enzyme. To explore this question, the investigators used wild-type M. tuberculosis gyrase and four proteins having mutations in the GyrA subunit. These included A90S, which would be predicted to form a water-metal ion bridge more efficiently than the wild-type enzyme and A90V, which might be expected to interfere with the bridge through the interposition of a larger hydrophobic residue. In addition, they used mutants of aspartate-94, D94G and D94H. Aspartate-94 occupies the acidic site predicted to be involved in water-metal ion bridge formation. Either of these mutations should disrupt the bridge. A90V, D94G, and D94H are among the most common mutations found in fluoroquinolone-resistant M. tuberculosis isolated in the clinic.
Figure 2. Structures of fluoroquinolones and fluoroquinolone derivatives used in the study. The C3/C4 ketoacid functionality that binds the metal ion in the water-metal ion bridge is highlighted in red. The C8 substituent that was found to influence activity against M. tuberculosis DNA gyrase is highlighted in blue.
FIGURE 3. Close-up view of the water-metal ion bridge formed between moxifloxacillin and topoisomerase IV from A. Baumannii. DNA is in green, enzyme residues are in gray, moxifloxacillin is in yellow, the Mg2+ ion is a light green sphere, and the water molecules are shown as red spheres. Image reproduced by permission from Macmillan Publishers Ltd, from A. Wohlkonig, et al., (2010) Nat. Struct. Mol. Biol., 17, 1152. Copyright 2010.
All four of the mutant enzymes exhibited DNA supercoiling and cleavage activity comparable to that of the wild-type enzyme. However, the A90V, D94G, and D94H mutants were resistant to the effects of two prototypical fluoroquinolones, ciprofloxacin (cipro), and moxifloxacin (moxi), as determined by their ability to induce increased DNA cleavage by the enzymes. The aspartate-94 mutants were more resistant to the effects of both drugs than were the serine-90 mutant, and all of the mutants were more resistant to cipro than to moxi. In contrast, the A90S mutant was more sensitive to the effects of the fluoroquinolones than the wild-type enzyme. These results suggest that both alanine-90 and aspartate-94 play some role in the interaction of fluoroquinolones with M. tuberculosis gyrase, possibly through the formation of a water-metal ion bridge.
Quinazolinediones containing a (3´-aminomethyl)pyrrolidinyl [3´-(AM)P] substituent at C7 (Figure 4) are effective topoisomerase II poisons through interaction between the 3´-(AM)P substituent and the enzyme rather than formation of a water-metal ion bridge. Consistently, the ability of 8-methyl-3´-(AM)P-dione to induce DNA break formation by M. tuberculosis gyrase was unaffected by the three mutants predicted to disrupt water-metal ion bridge formation. This was also true for 8-methyl-3´-(AM)P-FQ, a fluoroquinolone bearing the 3´-(AM)P substituent at C7. However, the A90S mutant was more sensitive to 8-methyl-3´-(AM)P-FQ than the wild-type gyrase, whereas this was not observed in the case of 8-methyl-3´-(AM)P-dione. As prior work had suggested an important role for a C8 substituent in the activity of fluoroquinolones, the investigators assessed the potency of 8-H-3´-(AM)P-dione and 8-H-3´-(AM)P-FQ, compounds lacking a substituent at C8. As expected, removal of the C8 methyl group resulted in a loss of activity against all enzymes tested. Together, these results suggest that a water-metal ion bridge, formed predominantly through aspartate-94, plays a role in fluoroquinolone binding to M. tuberculosis gyrase, but that substituents at C7 and C8 are also important to drug potency.
FIGURE 4. Structures of quinazolinediones and the structurally similar fluoroquinolone derivatives used in the study. The C7 3´-(AM)P substituent that is important for quinazolinedione-binding to topoisomerases is highlighted in red, and the C8 substituent is highlighted in blue.
To further test the hypothesis that water-metal ion bridge formation is involved in fluoroquinolone binding to M. tuberculosis gyrase, the investigators explored the effects of varying the concentration of Mg2+ ‑ the preferred bridge metal ion ‑ in their reaction mixtures on fluoroquinolone efficacy. They found that the Mg2+ requirement for cipro activity increased in the order of the A90S mutant, wild-type enzyme, A90V mutant. In contrast, the three enzymes exhibited a similar Mg2+ requirement for 8-methyl-3´-(AM)P-dione activity. Replacing Mg2+ with Mn2+ resulted in reduced cipro efficacy and potency with the wild-type gyrase, while the A90S mutant responded to cipro equally well in the presence of either Mn2+ or Mg2+, and the A90V mutant was completely resistant to cipro in the presence of Mn2+. These results strongly support the formation of a water-metal ion bridge in the interaction of fluoroquinolones with M. tuberculosis gyrase, although the substitution of alanine for serine at position 90 appears to have weakened the bridge.
With their increased understanding of how fluoroquinolones interact with the enzyme, the investigators next tested the hypothesis that drug activity against resistant mutant gyrases could be substantially improved. As their work indicated a role for the C8 substituent in the fluoroquinolone-gyrase interaction, they compared the efficacies and potencies of moxi and cipro analogs containing a hydrogen atom, a methyl group, or a methoxy group at that position (Figure 2). Consistent with expectations, neither compound bearing a hydrogen at this position exhibited optimal activity with the wild-type enzyme, and both compounds were essentially inactive with the A90V or D94G mutant gyrases. In the case of the cipro nucleus, the methoxy substituent at C8 conveyed the highest efficacy and potency with both the wild-type and mutant enzymes, whereas with moxi, the methyl substituent provided the greatest activity. Indeed, the 8-methyl-moxi compound was the most potent of all tested.
Competition studies between the 8-H-containing analogs and either 8-methoxy-cipro or 8-methyl-moxi revealed that the increased potency of the latter two compounds was due to a C8 substituent-dependent increase in binding affinity in the case of both the D94G and the A90V mutants. This conclusion was further supported by time-course studies that showed that the increased potency of 8-methoxy-cipro and 8-methyl-moxi correlated with increased stability of the drug/enzyme/DNA cleavage complex. These results suggest that the 8-methyl and 8-methoxy substituents increase the binding affinity of moxi and cipro, respectively, even in the absence of a water-metal ion bridge. Importantly, none of the cipro or moxi analogs tested had activity against human topoisomerase IIα.
The investigators concluded that a water-metal ion bridge plays a role in the interaction of fluoroquinolones with the M. tuberculosis gyrase. Strong support for this conclusion came from X-ray crystal structures of complexes of the gyrase with multiple fluoroquinolones, including moxi, 8-methyl-moxi, and cipro (Figure 5), which were published in an accompanying paper from the Berger lab [T. R. Blower, et al. (2016), Proc. Natl. Acad. Sci. U.S.A., published online January 20, DOI: 10.1073/pnas.1525047113]. In each case, the major contact between the inhibitor and the enzyme was a water-metal ion bridge, established primarily through interaction with aspartate-94. Lack of a serine at position 90 results in a weaker bridge than is seen with other type II topoisomerases. As a result, the binding affinity contributions from substituents at C7 and C8 may play a greater role in the case of the M. tuberculosis gyrase than in other bacterial enzymes of this class. This provides an excellent opportunity to exploit these interactions for the design of more potent compounds that are effective against some of the most common mutations associated with fluoroquinolone resistance. The work of the Osheroff and Burger labs paves the way for new drugs to meet the challenge of rapidly evolving antimicrobial resistance in the treatment of tuberculosis.
FIGURE 5. X-ray crystal structure showing the interactions of moxi (orange) and 8-methyl-moxi (green) with the active site of M. tuberculosis gyrase. The gyrase is shown in two shades of pink, and DNA bases are in light cream. Magnesium ions are shown as yellow spheres and water molecules as smaller red spheres. Aspartate-94 and alanine-90 are labeled, and the proximity of aspartate-94 to the bridging magnesium ion and its associated water molecules is readily seen. The active site magnesium ion can also be seen at a distance. Figure kindly provided by Neil Osheroff. Copyright 2016, Neil Osheroff.