Quinolone Resistance Mutations in the GrlB Protein of Staphylococcus aureus

Abstract

Two altered GrlB proteins (one with an Asp-432→Asn alteration and one with an Asn-470→Asp alteration) of Staphylococcus aureus were purified as fusion proteins to maltose-binding protein. The 50% inhibitory concentrations of levofloxacin were 14 and 3.4 μg/ml against topoisomerase IV containing GrlB proteins with alterations at positions 432 and 470, respectively. These results suggest that the alteration of Asp to Asn at position 432 may be responsible for quinolone resistance.

Fluoroquinolones have potent and broad antibacterial activities against gram-positive and -negative bacteria. However, the increase in clinical use of quinolones has led to the widespread emergence of resistance, especially in Staphylococcus aureus (13). The three known resistance mechanisms to quinolones in S. aureus are alteration of two target enzymes and overproduction of quinolone efflux protein (8). The two target enzymes of quinolones are topoisomerase IV and DNA gyrase, and the alteration of subunit A proteins (GrlA and GyrA) is reported to be responsible for quinolone resistance on the basis of genetic and enzymological evidence (1, 3–7, 9, 10, 12, 14–20, 22). We have reported that the inhibitory activities of quinolones against topoisomerase IV reconstituted with GrlA with an alteration at position 80 and/or 84 and wild-type GrlB were weaker than those against wild-type enzyme (17). While there are many reports of quinolone resistance involving GrlA and GyrA, association of the subunit B proteins (GrlB and GyrB) with quinolone resistance has been investigated only genetically for S. aureus (6, 9, 15, 17). There has been a report on altered GyrB proteins (with Asp-426→Asn and Lys-447→Glu alterations) in Escherichia coli (24). In our previous work (17), a mutation in the grlB gene [GAT(Asp)-432→AAT(Asn)] was found in a clinical isolate that was highly resistant to quinolones and that also had alterations in GrlA at position 80 and in GyrA at position 84. Recently, Fournier and Hooper (6) reported a GrlB mutation encoded at codon 470 (Asn→Asp), and the strain conferring the altered GrlB became, in comparison to the parent strain, slightly resistant to quinolones (two- to eightfold more) and susceptible to novobiocin. In this study, these two altered GrlB proteins (with Asp-432→Asn and Asn-470→Asp alterations) were purified and the inhibitory activities of quinolones against mutated topoisomerase IV were determined.

All quinolones used were synthesized at the New Product Research Laboratories I, Daiichi Pharmaceutical Co. Ltd., Tokyo, Japan. Novobiocin was purchased from Sigma-Aldrich Japan (Tokyo, Japan). The bacterial strains used in this study were S. aureus FDA 209-P and E. coli MC1061 (17). GrlA and GrlB proteins of topoisomerase IV of S. aureus FDA 209-P were purified separately as fusion proteins with maltose-binding protein (MBP) from overproducing strains of E. coli by using a protein fusion and purification system (New England Biolabs, Beverly, Mass.). Mutated grlA and grlB genes prepared by site-directed mutagenesis with Mutan-K (Takara Syuzo, Shiga, Japan) were also used for construction of expression vectors. Details of the purification procedures and the method for determination of inhibitory activities of quinolones were described previously (17). The substrate of the topoisomerase IV for determination of decatenation activity was kinetoplast DNA (Nippon Gene, Toyama, Japan), and the inhibitory activity of each drug was measured three times. The 50% inhibitory concentrations (IC₅₀s) were calculated as the drug concentrations that reduced the decatenation observed with drug-free controls by 50%.

The altered GrlB proteins with a substitution of Asn (AAT) for Asp-432 (GAT) or of Asp (GAT) for Asn-470, the altered GrlA (Ser-80→Phe), and wild-type GrlA and GrlB proteins were purified as fusion proteins with MBP. When the purified GrlB proteins were reconstituted with GrlA, decatenation activity was observed. The relative decatenation activity of the altered enzymes with one substitution in the GrlB sequence were almost the same as that of the wild-type enzyme; however, the activity of topoisomerase IV with alterations at position 80 in the GrlA sequence and at position 432 in the GrlB sequence was half that of the wild-type enzyme. Thus, the amount of the enzyme with altered GrlA and GrlB proteins producing 1 U of decatenation activity was twice that of the wild-type enzyme with GrlA and GrlB. The inhibitory activities of the quinolones and novobiocin on the decatenation activity of topoisomerase IV with or without alteration are shown in Table 1. The IC₅₀s of quinolones against the altered enzyme in which Asn (AAT) was substituted for Asp (GAT) at position 432 of the GrlB sequence were 5 to 16 times higher than those against wild-type topoisomerase IV, which suggests that the change at position 432 in GrlB resulted in low-level quinolone resistance. In contrast, the IC₅₀s of quinolones against the enzyme with an alteration in GrlB at position 470 were within twofold of those against wild-type topoisomerase IV, while the IC₅₀ of novobiocin was one-fourth of that against the strain FDA 209-P enzyme. These results reflect the MIC change reported by Fournier and Hooper (6). Among the quinolones tested, sitafloxacin (DU-6859a) showed the highest inhibitory activity against both enzymes.

TABLE 1.

Inhibitory activities of quinolones and novobiocin against topoisomerase IV

Compound	IC50 (μg/ml) for topoisomerase IV wild type or with indicated alteration(s)
	Wild typeab	GrlAa			GrlBc		GrlA (Phe-80) and GrlB (Asn-432)c
		Phe-80	Lys-84	Phe-80 and Lys-84	Asn-432	Asp-470
Levofloxacin	2.3	180	130	1,160	14.0 ± 1.3	3.4 ± 0.57	267 ± 17
Ofloxacin	3.9	260	285	1,250	41.6 ± 2.2	6.5 ± 2.45	784 ± 9.7
Sitafloxacin	0.45	4.2	13	170	2.5 ± 0.33	0.17 ± 0.02	11.6 ± 0.69
Ciprofloxacin	2.5	75	110	530	39.5 ± 6.3	4.0 ± 1.14	184 ± 32
Sparfloxacin	7.4	270	570	1,190	59.3 ± 2.8	8.9 ± 0.86	486 ± 40
Tosufloxacin	1.8	90	170	>200	8.2 ± 2.1	1.6 ± 0.01	71.8 ± 19.7
Novobiocin	16	14	10	11	16.4 ± 1.4	4.3 ± 1.86	24.8 ± 1.64

^aFrom reference 15. 

^bFor the wild type, GrlA contained Ser-80 and Glu-84 and GrlB contained Asp-432 and Asn-470. 

^cValues are expressed as means ± standard deviations. 

The inhibitory activities of the drugs against topoisomerase IV reconstituted with altered GrlA (Ser-80→Phe) and GrlB (Asp-432→Asn) are also shown in Table 1. The IC₅₀ of sitafloxacin against the double-mutated topoisomerase IV was also the lowest among those of the quinolones tested. The IC₅₀s of levofloxacin and ciprofloxacin against topoisomerase IV with a double amino acid change were 117- and 72-fold, 1.3- and 2.4-fold, and 19- and 4.5-fold higher than those against the wild-type enzyme and the enzymes with GrlA or GrlB alterations, respectively. These results confirm the association of a mutation in GrlB (Asp-432→Asn) with quinolone resistance.

The Asp-432→Asn change in the GrlB sequence is considered to correspond with the Asp-437→Asn change in GyrB that is a part of the quinolone resistance-determining region (QRDR). The QRDR in GyrA was considered to be a part of the quinolone-binding pocket that is the binding site of quinolones to the DNA-DNA gyrase complex (23). The crystal structure of the 59-kDa protein of E. coli GyrA showed that the QRDR was at the periphery of GyrA and in close proximity to Tyr-122, which is a DNA-binding site (2). Ito et al. (9) also discussed that the QRDR around the Ser-84-to-Glu-88 region in GyrA is in close proximity to Asp-437 and Arg-458 in GyrB in S. aureus. The QRDRs of DNA gyrase and topoisomerase IV are well conserved, and the mutations conferring quinolone resistance are similar. In the case of topoisomerase IV, the region around Ser-80 to Glu-84 in GrlA may be in close proximity to Asp-432; however, Asn-470 is slightly outside of the highly conserved QRDR. When the structure of the topoisomerase-DNA-quinolone complex is revealed, the exact mode of binding of quinolones to the enzyme-DNA complex will be clarified.

Novobiocin, a coumarin antibacterial agent, acts by inhibiting ATP hydrolysis by the GyrB protein. Lewis et al. reported the crystal structures of a complex between DNA gyrase B protein and novobiocin (11). From their results, the binding sites for ATP and novobiocin overlapped to some degree and mutations which confer resistance to novobiocin were localized at the periphery of the ATP-binding site (6, 11, 21). It was also reported that hypersusceptibility to novobiocin might occur if the altered topoisomerase IV (GrlB altered at position 470) had reduced affinity for ATP. The IC₅₀ of novobiocin against the altered enzyme containing Asn-432 in GrlB and Phe-80 in GrlA was slightly higher than those against topoisomerase IV with other alterations. We speculate that the tertiary structure of the enzyme with the double substitution might be changed.

In our previous studies (16–18), we found that the MICs of levofloxacin and ciprofloxacin against strain 891185, which possessed a grlB mutation at codon 432 and alterations at Phe-80 in GrlA and at Leu-84 in GyrA, were 100 and >800 μg/ml, respectively. Furthermore, the MICs of levofloxacin and ciprofloxacin for S. aureus 90-37 with the same alterations in GrlA and GyrA were 12.5 and 50 μg/ml, respectively. For 75 clinical isolates that possessed the same alterations in GrlA and GyrA but unknown alterations in GrlB and GyrB, the MICs of levofloxacin varied between 3.13 and 50 μg/ml and those of ciprofloxacin varied between 12.5 and 800 μg/ml. From these data and the possibility of the third mechanism of quinolone resistance (efflux), the association of a mutation in GrlB position 432 with quinolone resistance was within 16 times for levofloxacin and 128 times for ciprofloxacin at the MIC level.