The first mechanism-based inhibitors of APH(3')s were described by Roestamadji et al. (1995b). The compounds are derivatives of neamine and kanamycin B, in which a nitro (NO2) group replaces the amine at the 20-po-sition (Fig. 8). These suicide substrates are excellent substrates for APH(3')s but poor antibiotics. Upon phosphorylation by APH(30) enzymes at the 30-hydroxyl, the phosphoryl group, being an excellent leaving group, is
Fig. 8 Proposed mechanism of aminoglycoside kinase inhibition by 2'-nitro aminoglyco-side derivatives. Spontaneous loss of phosphate from a phospho-aminoglycoside yields an electrophilic nitroalkene. Trapping of a nucleophilic active site residue (Nuc) produces an inactivated enzyme
rapidly eliminated, generating an electrophilic nitroalkene. The reactive electrophilic intermediate can in turn capture an active-site nucleophilic amino acid side chain and form a covalent bond, irreversibly inactivating the enzymes.
A more recent derivative is 30-oxo-kanamycin A, a self-regenerating aminoglycoside, in which the hydroxyl group at 30-position is replaced by a ketone (Haddad et al. 1999) (Fig. 9). The hydrated variant of this compound is a good substrate for APH(30) enzymes. However, the phosphorylated prod-
uct is unstable and releases the inorganic phosphate in a spontaneous non-enzymatic way, regenerating the parent compound. The antibiotic is therefore not inactivated, making the resistance enzymes obsolete.
The molecular recognition of aminoglycosides in both the ribosomal Asite and all three classes of resistance enzymes is accomplished in analogous ways. In the case of APH(30) enzymes, this is illustrated by comparing the binding of paromomycin I to the 16S ribosomal RNA (Carter et al. 2000) and neomycin B to APH(3')-IIIa (Fong and Berghuis 2002). The crystal structure of the 30S ribosome in complex with different aminoglycosides (including paromomycin I) has been solved (Carter et al. 2000). Paromomy-cin I is a 4,5-disubstituted 2-deoxystreptamine aminoglycoside that can be inactivated by many APH(30) enzymes. The structure of paromomycin I closely resembles that of neomycin B; the only difference is in the functional group at the 60-position, where paromomycin I has a hydroxyl and neomycin B an amino group. Comparison of the binding of neomycin B to APH(30)-IIIa and paromomycin I to the A-site of the bacterial ribosome reveals that the conformation of the aminoglycosides and the hydrogen bond network between the aminoglycoside and their respective targets are essentially identical. However, the two complexes differ considerably in their van der Waals interactions. The face of the aminoglycoside that makes most of the van der Waals interactions with APH(3')-IIIa is opposite to that which interacts with the bacterial ribosome (Fong and Berghuis 2002) (Fig. 10). The similarities in aminoglycoside binding to the two targets explain the effectiveness of APH(3')-IIIa as a resistance factor, but more importantly, the differences in binding mechanism can be exploited in the design strategies of inhibitors
Fig. 10 The van der Waals surface of the aminoglycoside binding site of APH(3')-IIIa (left) and the bacterial ribosome (right). APH(3')-IIIa is shown with bound neomycin B, while the ribosome structure has bound paromomycin I. Atoms of the ball-and-stick illustrations of the aminoglycoside are colour coded as follows: oxygen, light grey; carbon, grey; nitrogen, black
Fig. 10 The van der Waals surface of the aminoglycoside binding site of APH(3')-IIIa (left) and the bacterial ribosome (right). APH(3')-IIIa is shown with bound neomycin B, while the ribosome structure has bound paromomycin I. Atoms of the ball-and-stick illustrations of the aminoglycoside are colour coded as follows: oxygen, light grey; carbon, grey; nitrogen, black and novel aminoglycoside variants. For example, the binding of aminoglyco-side derivatives can be blocked by modifying the corresponding face with bulky chemical moieties (Fong and Berghuis 2002; Vicens and Westhof 2003).
Was this article helpful?