The vaccinia virus DNA polymerase structure provides insights into the mode of processivity factor binding

Abstract

Vaccinia virus (VACV), the prototype member of the Poxviridae, replicates in the cytoplasm of an infected cell. The catalytic subunit of the DNA polymerase E9 binds the heterodimeric processivity factor A20/D4 to form the functional polymerase holoenzyme. Here we present the crystal structure of full-length E9 at 2.7 Å resolution that permits identification of important poxvirus-specific structural insertions. One insertion in the palm domain interacts with C-terminal residues of A20 and thus serves as the processivity factor-binding site. This is in strong contrast to all other family B polymerases that bind their co-factors at the C terminus of the thumb domain. The VACV E9 structure also permits rationalization of polymerase inhibitor resistance mutations when compared with the closely related eukaryotic polymerase delta-DNA complex.

Fig. 1

Domain structure of E9. Definitions according to Liu et al.: N-terminal: 1–157, 497–523; exonuclease: 158–353, 435–496; insert 2: 354–434; palm: 524–618, 676–829; finger: 619–675; thumb: 830–1006. The same view is used throughout the figure. a Front and back view of the domain organization. b View of insert 0, insert 3, and insert 4. c View of insert 1 and insert 2. d View of the electrostatic potential of the solvent accessible surface with colors ranging from red (−3kT/e) to blue (3kT/e). The position of inserts 2 and 3 is indicated. e Conservation of residues within 29 representative sequences from the Chordopoxvirinae subfamily. Coloring calculated with ESPript as a function of the degree of conservation ranging from red for strict conservation to white. The conserved patch located in insert 3 is encircled as is insert 2, which does not show a particularly conserved surface

Fig. 2

The C-terminal domain of A20 (A20 C-ter) interacts with VACV E9. a Elution profiles from SEC are compared. A20 C-ter and E9 were either loaded separately or mixed together prior to injection onto the column. b SDS-PAGE analysis of the eluted fractions for each run. c Alignment of the last amino acids of A20 using representative viruses from each genus of the Chordopoxvirinae subfamily: VACV, CPXV (cowpox virus), MPXV (monkeypox virus), CMLV (camelpox virus), ECTV (ectromelia virus), VARV (variola virus), YKPV (yokapox virus), LSDV (lumpy skin disease virus), TANV (tanapox virus), SWPV (swinepox virus), MYXV (myxoma virus), MOCV (molluscum contagiosum virus), DPV (deerpox virus), ORFV (ORF virus), CNPV (canarypox virus). Conserved residues are in red. The predicted secondary structure using MLRC and its reliability on a 0–9 scale are shown at the bottom of the alignment (ACC). d Helical wheel projection with residues of A20 predicted to fold as an α-helix (aa 400–417). Hydrophilic residues are presented as circles, hydrophobic residues as diamonds, charged residues as triangles and pentagons. e A point mutation in A20 C-ter abrogates binding to E9. A20 C-ter-Phe414Ala was incubated with WT E9 before injection onto SEC. The eluted fractions were analyzed as in b

Fig. 3

Peptides protected by the E9 A20 C-ter interaction determined by H/D exchange MS. a Peptides from E9, identified by LC-MS/MS, are represented as bars above the primary sequence. The secondary structure, derived from the crystal structure, is shown underneath the sequence (in gray residues not seen in the structure). The level of protection of individual peptides, as determined by comparing the % D incorporation for free E9 with the one for E9 bound to A20 C-ter, is color coded according to the scale bar. Highly protected areas are shown in red, whereas peptides becoming more exposed upon complex formation are shown in blue. The highly protected region involves the α-helix of insert 3 (residues 577–590). b H/D exchange results are mapped onto the crystal structure of E9. Protected residues are color coded as in a

Fig. 4

Mutational analysis of E9 insert 3 regarding A20 C-ter binding. a, c, e Crystal structure of E9 insert 3 (gray) with mutated residues highlighted. The corresponding sequence of the α-helix of insert 3 (black letters) is shown with mutated amino acids depicted as colored letters. b, d, f Each E9 mutant was mixed with WT A20 C-ter prior to analysis by size exclusion chromatography. A 4–20% SDS-PAGE analysis of the eluted fractions for each of the runs is presented

Fig. 5

Models of polymerase holoenzymes. a Model of the VACV holoenzyme. The envelope obtained by SAXS is used to define the global outline. Components of the polymerase holoenzyme have been placed manually: a model of E9 in complex with DNA in elongation mode, the D4/A20_1–50 complex with bound DNA (A20_1–50, in violet, D4 in yellow, and DNA in blue), the SAXS ab initio model of the C-terminal fragment of A20 (magenta). Schematic view of the processivity factor binding of b E9, c archaeal and phage DNA polymerases, d HSV DNA polymerase, and e S. cerevisiae polymerase δ. Different orientations of the holoenzyme (brown) at the replication fork, where D4 binds either the newly synthesized dsDNA strand (f), or the incoming template strand before (g) or after (h) strand separation by the helicase–primase D5 (green)

Fig. 6

Analysis of drug-resistance mutations on E9. a Locations of different mutants listed in Table 2 on the model of E9 in complex with DNA in elongation mode. Domains of the polymerase are color coded as in Fig. 1a. b Analysis of the finger domain movements: in magenta, the DNA-bound structure of human pol α (PDB 5iud), in orange E9 after superposition of the palm and exonuclease domains onto the binary complex of human pol α, in blue the ternary complex of yeast pol δ with DNA and an incoming nucleotide (PDB 3iay). Upon nucleotide binding, the tip of the finger of E9 may move close to the insert 2 domain allowing a potential contact, which would occur in vicinity of the PAA resistance mutations depicted as green spheres