Structural basis for dual nucleotide selectivity of aminoglycoside 2''-phosphotransferase IVa provides insight on determinants of nucleotide specificity of aminoglycoside kinases

Abstract

Enzymatic phosphorylation through a family of enzymes called aminoglycoside O-phosphotransferases (APHs) is a major mechanism by which bacteria confer resistance to aminoglycoside antibiotics. Members of the APH(2″) subfamily are of particular clinical interest because of their prevalence in pathogenic strains and their broad substrate spectra. APH(2″) enzymes display differential preferences between ATP or GTP as the phosphate donor, with aminoglycoside 2″-phosphotransferase IVa (APH(2″)-IVa) being a member that utilizes both nucleotides at comparable efficiencies. We report here four crystal structures of APH(2″)-IVa, two of the wild type enzyme and two of single amino acid mutants, each in complex with either adenosine or guanosine. Together, these structures afford a detailed look at the nucleoside-binding site architecture for this enzyme and reveal key elements that confer dual nucleotide specificity, including a solvent network in the interior of the nucleoside-binding pocket and the conformation of an interdomain linker loop. Steady state kinetic studies, as well as sequence and structural comparisons with members of the APH(2″) subfamily and other aminoglycoside kinases, rationalize the different substrate preferences for these enzymes. Finally, despite poor overall sequence similarity and structural homology, analysis of the nucleoside-binding pocket of APH(2″)-IVa shows a striking resemblance to that of eukaryotic casein kinase 2 (CK2), which also exhibits dual nucleotide specificity. These results, in complement with the multitude of existing inhibitors against CK2, can serve as a structural basis for the design of nucleotide-competitive inhibitors against clinically relevant APH enzymes.

FIGURE 1.

Key components of nucleoside-binding site. Residues forming the nucleoside-binding site of APH(2″)-IVa are shown. A bound adenosine molecule is shown in cyan stick representation. Key residues are divided into three regions based on secondary structure elements and color-coded as follows: the N-terminal β-strands forming the top face of the cleft are shown in orange, the linker loop is shown in blue, and the loops from the core subdomain forming the bottom face of the cleft are shown in green.

FIGURE 2.

Adenosine versus guanosine binding for APH(2″)-IVa. A, APH(2″)-IVa-adenosine complex, showing the enzyme in green graphic representation and the 2F_o − F_c electron density (gray, 1.0 σ) for the adenosine molecule in light orange stick representation. Hydrogen-bonding interactions are represented as black dashed lines. Residues that directly interact with the substrate are show in stick representation and colored light green. B, APH(2″)-IVa-guanosine complex, showing the enzyme in blue graphic representation and the 2F_o − F_c electron density (gray, 1.0 σ) for the adenosine molecule in cyan stick representation. Ordered water molecules (Wat1–Wat4) forming a solvent network are highlighted as red spheres. Residues involved in interacting with the substrate are show in stick representation and colored purple. C, structural superposition of the APH(2″)-IVa-guanosine structure (blue) onto the APH(2″)-IVa-adenosine structure (green), showing the displacement of the purine base. Hydrogen-bonding interactions between the protein and each nucleoside are shown in the color of the protein molecule. The conformation of the backbone is slightly shifted to optimize binding with the respective nucleoside. The solvent network only applies to the guanosine-bound structure because it would clash with the adenosine molecule.

FIGURE 3.

Partial multiple sequence alignment of selected APH enzymes and CK2. The aligned enzymes are ordered based on increasing preference for GTP as the second substrate. Secondary structural elements are indicated below the alignment. Residues forming the nucleoside-binding site in APH(2″)-IVa are separated into three regions. Phe-95 of APH(2″)-IVa and its corresponding amino acids in the other enzymes are highlighted by the red box. The alignment for APH(2″) enzymes was created with Clustal Omega (32). APH(3′)-IIIa and CK2 were aligned based on a manual structural alignment between representative structures (PDB accession numbers 1L8T and 1LP4) and the APH(2″)-IVa-adenosine complex. The kinetic parameters for the six enzymes were taken from literature (3, 33, 34). The K_m(ATP)/K_m(GTP) ratio for APH(2″)-IVa varies among three independent studies due to small differences in specific experimental conditions (3, 13).

FIGURE 4.

Structural basis for ATP-selectivity of APH(3′)-IIIa. A, nucleotide-binding clefts of APH(2″)-IVa-guanosine (blue with cyan ligand) and APH(3′)-IIIa-ADP (brown with dark red ligand), showing that the key residue Ala-93 for coordinating the guanine ring is positioned too far away from the base for effective interactions in APH(3′)-IIIa. B, alternative view of the nucleotide-binding clefts, showing that the conformation of residue Met-90 of APH(3′)-IIIa clashes with a solvent network that would be required for guanosine binding. The ADP molecule aligns with the adenosine in the APH(2″)-IVa-adenosine complex, which is shown in semitransparent orange stick representation.

FIGURE 5.

Structural comparison between nucleoside-bound complexes of APH(2″)-IVa and CK2. A, structural superposition of residues forming the nucleoside-binding pocket of AMPPNP-bound CK2α (yellow with red ligand) onto those of adenosine-bound APH(2″)-IVa (green with light cyan ligand). Despite significant discrepancies in the overall protein structure, this region shows strong structural conservation. B, structural superposition of GMPPNP-bound CK2α structure (yellow with red ligand) onto guanosine-bound APH(2″)-IVa (blue with dark cyan ligand). The conformation of the interdomain linker, highlighted in graphic representation, and the position of key residues, shown in stick representation, are conserved. Also conserved between the two structures is a solvent network, consisting of two water molecules for CK2 (green spheres) and three water molecules for APH(2″)-IVa (red spheres, with one overlapping and occluded by a green sphere).