The structure of a protein primer-polymerase complex in the initiation of genome replication

Abstract

Picornavirus RNA replication is initiated by the covalent attachment of a UMP molecule to the hydroxyl group of a tyrosine in the terminal protein VPg. This reaction is carried out by the viral RNA-dependent RNA polymerase (3D). Here, we report the X-ray structure of two complexes between foot-and-mouth disease virus 3D, VPg1, the substrate UTP and divalent cations, in the absence and in the presence of an oligoadenylate of 10 residues. In both complexes, VPg fits the RNA binding cleft of the polymerase and projects the key residue Tyr3 into the active site of 3D. This is achieved by multiple interactions with residues of motif F and helix alpha8 of the fingers domain and helix alpha13 of the thumb domain of the polymerase. The complex obtained in the presence of the oligoadenylate showed the product of the VPg uridylylation (VPg-UMP). Two metal ions and the catalytic aspartic acids of the polymerase active site, together with the basic residues of motif F, have been identified as participating in the priming reaction.

Figure 1

Structure of the primer protein VPg in a complex with 3D. (A) Stereo view of a sigma A weighted ∣F_o∣−∣F_c∣ electron density map at 2.9 Å resolution and contoured at 3.0σ around the VPg–UMP molecule (The VPg–UMP and ions were omitted from the phasing model). The 15 amino acids of VPg, the UMP covalently linked to the protein and the metal ions are placed inside the density in ball and stick representation colored in atom type code. Names for all residues are explicitly labeled in one letter code. (B) Details of the interactions seen in the active site of the 3D polymerase during the uridylylation reaction. The residues Pro2, Tyr3 and Ala4 of VPg are shown in sticks in red and the UMP, covalently linked to the hydroxyl group of Tyr3, in light green. The divalent cations Mn²⁺ and Mg²⁺ are shown as magenta and orange spheres, respectively, and the anomalous difference Fourier map is shown as a chicken wire in blue. The 3D amino acids involved in direct hydrogen bonds with ions and the uridylylated tyrosine are shown in ball and sticks in atom type code, and the hydrogen bonds appear as dashed lines. All residues are explicitly labeled. The predicted position of the oligo(A) template strand (dark green) was determined using the 3D–RNA template–primer complex (PDB entry 1WNE) as a guide.

Figure 2

Structure of VPg bound to the FMDV 3D polymerase. (A) Molecular surface of the 3D polymerase shown in two different views: the conventional orientation, as if looking into a right hand (left) and a side view (right). The polymerase domains: fingers, palm and thumb are colored in light-blue, dark-blue and purple, respectively. The VPg protein that binds across the RNA binding cleft is represented as a red ribbon with side chains shown in stick representation in red, the UMP molecule covalently linked to the hydroxyl group of the VPg residue Tyr3 is shown in green, and the metal ions as orange spheres. In the right panel, the fingers residues at the top of the NTP tunnel and most of the thumb domain are removed to allow the visualization of VPg. (B) Top-down views of the polymerase molecule showing the trajectory of the VPg protein (left) compared to the trajectory of RNA template–primer (right) in the structure of the complex FMDV 3D–RNA template–primer (PDB entry 1WNE; Ferrer-Orta et al, 2004). Comparisons were carried out by the structural superimposition of both polymerase complexes that gave a root mean square deviation of 0.33 Å for the superimposition of all Cα atoms. The RNA is shown as a ribbon representation in green (template chain) and yellow (primer chain). Roughly, the N-terminal moiety of the bound VPg protein occupies part of the NTP entry channel and the position of the primer, whereas the carboxy-terminus of the protein superimposes with the trajectory of the RNA duplex product. In both panels, the N-terminal residues (from 34 to 48) and residues at the top of the NTP tunnel (from 163 to 180) of 3D are omitted to better show the VPg protein.

Figure 3

VPg–3D polymerase interactions. (A) Structure of the VPg primer protein (red) with the contacting residues of the 3D polymerase shown in different colors. Four different regions of the polymerase molecule contact VPg residues E166, I167, R168, K172 and R179, belonging to motif F of fingers (orange), together with residues T407, A410 and I411 of the thumb domain (light blue), interact with the N-terminal moiety of VPg, stabilizing the conformation of Y3 in the active site cavity. In addition, residues E166, I167 of motif F (orange), K387 and R388 of motif E (dark blue) and T407, A410 and I411 of helix α13 (light blue) interact with the central part of the VPg protein. Finally, the 3D residues G216, C217 and P219, located at the beginning of helix α8 (light blue) in the fingers domain, together with the side chain of Y336 within the C motif (yellow) of the palm domain, establish hydrophobic contacts with R11 at the exit of the polymerase cavity. (B) Structure of the uridylylated VPg protein (shown in red and the linked UMP in green) with the contacting residues of the 3D polymerase shown in blue. In addition to the interactions described in (A), amino acids D245 of motif A (pink) and D338 of motif C (yellow) are placed in the correct orientation for the catalysis of the phosphodiester linkage in the active site of the 3D protein.

Figure 4

Effect of amino acids replacements in 3D and VPg on the VPg-uridylylation activity. (A) VPg1 uridylylation activity by substituted FMDV 3Ds, including one or two amino-acid replacements, relative to the wild-type enzyme. Values are the average of at least three independent experiments. The mono- and di-uridylylated band of a representative electrophoretic separation is shown. No significant differences were seen when the reaction products were treated with RNAse A before electrophoresis. Preparation of mutant 3Ds and assay procedures are described in Materials and methods. (B) Uridylylation activity of wild-type FMDV 3D on VPg1, VPg2, VPg3 and VPg1 with single amino-acid substitution. Results and procedures are as in (A).

Figure 5

(A) Structure-based alignment of the picornavirus polymerase residues that make contact with VPg. Aligned are domains for the different picornaviral polymerases whose structure is known. The strictly conserved residues are in red blocks and similar residues in red characters and blue boxes. The FMDV 3D residues interacting with the VPg primer protein are marked by green inverted triangles (B) Sequence alignment of the VPg protein of the different picornaviruses. The FMDV VPg1 residues interacting with 3D are marked by green squares. Residues of PV VPg previously shown to interact with PV 3D polymerase by yeast two-hybrid analysis (Xiang et al, 1998) are highlighted in yellow blocks.

Figure 6

Comparison of FMDV polymerase bound to VPg with other picornaviral polymerases. (A) Mapping of the conserved FMDV 3D residues that contact the VPg primer protein onto the polymerases structure (blue). The VPg-interacting residues are colored according to the different polymerase motifs: D245 of motif A is in magenta; E166, R168, K172 and R179 belonging to motif F are in orange; residues G216, C217 and P219 located in or near helix α8 in dark gray; amino acids Y336 and D338 of motif C are in yellow; and residues K387 and R388 of motif E in dark blue. The VPg protein occupying the central cleft of the polymerase is shown in red. (B) Stereoview of the structural superimposition of the coordinates of PV (green) and HRV16 (yellow) RdRPs onto the FMDV 3D–VPg complex (blue), showing the positioning of the VPg primer protein (red) in the central cavity of the polymerases. The Cα root mean square deviation is 1.24 Å for the superimposition of 362 atoms between PV and FMDV RdRPs, and 1.29, 1.25 and 1.29 Å for the superimpositions of 254, 370 and 353 Cα atoms, between FMDV and HRV1B, HRV14 and HRV16 RdRPs, respectively. The conserved VPg-interacting residues are shown as in (A). (C) Ribbon diagram of the structure of PV 3D polymerase (green) with the VPg of FMDV modeled inside (red). Two different views of the PV polymerase are shown: the conventional orientation (left) and a side view (right; the finger residues at the top of the NTP tunnel and most of the thumb domain are removed to allow the visualization of VPg). The residues that affected VPg binding (yellow) and uridylylation (blue), as determined by mutational analysis (Lyle et al, 2002), are shown as stick representation and are explicitly labeled. In the model, Tyr326 has the correct orientation to interact directly with the UMP substrate, and Lys359 is properly oriented to interact directly with the N-terminal moiety of the VPg peptide.