Crystal structure of a KSHV-SOX-DNA complex: insights into the molecular mechanisms underlying DNase activity and host shutoff

Abstract

The early lytic phase of Kaposi's sarcoma herpesvirus infection is characterized by viral replication and the global degradation (shutoff) of host mRNA. Key to both activities is the virally encoded alkaline exonuclease KSHV SOX. While the DNase activity of KSHV SOX is required for the resolution of viral genomic DNA as a precursor to encapsidation, its exact involvement in host shutoff remains to be determined. We present the first crystal structure of a KSHV SOX-DNA complex that has illuminated the catalytic mechanism underpinning both its endo and exonuclease activities. We further illustrate that KSHV SOX, similar to its Epstein-Barr virus homologue, has an intrinsic RNase activity in vitro that although an element of host shutoff, cannot solely account for the phenomenon.

Figure 1.

(A) Exonuclease assays involving the DNA substrates dsDNA and dsDNA-5′P illustrate that cleavage is only observed on duplexes in which one strand has been substituted with a 5′ phosphate group. (B) Fluorescence anisotropy assays indicate that both duplexes bind to SOX with similar affinities (12 µM for dsDNA and 6 µM for dsDNA-5′P) suggestive of largely equivalent modes of binding.

Figure 2.

(A) Cartoon of KSHV SOX and corresponding 2mFo-DFc omit map density for the DNA (magenta, contoured at 1σ) bound at the interface or ‘canyon' between the N- and C-terminal lobes (green and yellow, respectively) of the SOX molecule. The density for nucleotides at the free end of the duplex is more fragmented due to a lack of stabilizing protein–DNA interactions. (B) Sequence alignments of the conserved DNase motifs and bridge regions in the γ-herpesvirus SOX homologues from Human herpesvirus 8 (KSHV), Bovine herpesvirus 2 (BHV2), Mouse herpesvirus 68 (MHV68), Epstein–Barr virus (BGLF5), Porcine lymphotropic herpesvirus 2 (PLHV) and UL12 from the α-herpes virus HSV-1. The alignment was performed using CLUSTALW (http://www.ebi.ac.uk/Tools/clustalw2/index.html). (C) The locations of motifs II (pink), III (grey), IV (orange) and VI (blue) within the KSHV–SOX complex that out of the seven conserved sequences mediate the observed protein–DNA interactions or are involved in catalysis. The position of the partially ordered bridge (red) is also highlighted. The second SOX nuclear localization sequence (NLS2, green), that has a role in DNA binding is also shown.

Figure 3.

Schematic diagram of protein–DNA hydrogen bonds stabilizing the complex. Those depicted by dashed lines and asterisks are mediated by water molecules and those with unbroken lines, main or side chain groups. C12 and A29 (marked with red asterisks) correspond to the CA mismatch.

Figure 4.

DNA recognition. (A) Stabilization of the duplex in the region closest to the catalytic centre exclusively involves strand 1 and is dominated by residues in motif III. The protein–DNA contacts are a combination of water, main and side chain mediated hydrogen bonds involving residues K246, F249, K250 and Y373 (colour coded according to their motif locations as in Figure 2C) and the phosphate groups of G3, G4 and A5 together with a range of van der Waals interactions. mFo-DFc omit map density (red, contoured at 3σ although present at 5σ) attributable to a second metal ion (green) was also identified that is absent in the native KSHV structure. (B) The protein–DNA interactions observed for the second region of contact involves a more mixed distribution of residues over the conserved motifs compared with the first. Here residues R370 (motif VI), K318 and K320 (NLS), E297 and E300 (motif IV) mediate water, main and side chain hydrogen bonds with G31, A32 and G33 of strand 2 and A13 of strand 1. The side chain moiety of R319 protrudes into the minor groove where it makes a range of van der Waals contacts with C12, G31 and A32. (C) Mg1 and surrounding co-ordination sphere in the wild-type (left) and DNA bound (right) KSHV SOX structures. The loss of tight ligands contributed by E184 and a water molecule (disordered in our complex) results in a weaker association as a result of a shift of Mg1. In this position, Mg1 is within hydrogen bonding distance of the Nζ group of K246. (D) Schematic diagram of the classical two metal nucleophilic attack mechanism. The oxygen atom of the activated water is shown in red.

Figure 5.

(A) Stereofigure showing the active site geometry in relation to G3, the nucleotide whose phosphate group is in contact with K246 of the PD-(D/E)XK and the second metal ion, Mg2. The relative juxtaposition of G3 to Mg1 and a possible catalytic water molecule are inconsistent with a configuration that would promote cleavage at this site based on the classical two metal mechanisms. (B) Rotation of G1 out of the duplex stack such that a modelled 5′ phosphate group would be able to hydrogen bond to S146 and S219 (in the rotamer conformations observed in the native KSHV structure) that form a conserved ‘serine cluster’ (brown), is sufficient to align the scissile phosphate G2 with Mg1 and Mg2 for cleavage in a near classical two metal mechanism configuration. In this conformation, the two metal ions straddle the stereospecific oxygen O1P and the water molecule bound to K246 (A, light blue sphere) is well placed for inline attack.

Figure 6.

(A) Superposition of the KSHV DNA complex with the D203S BGLF5 mutant reveals a number of severe steric clashes involving residues G156-V159 (BGLF5) at the C-terminus of the bridge and all base pairs 5′ to G4. This is as result of repositioning of residues A160 and F161 that move ∼5.5 Å towards the C-terminal lobe of the SOX molecule. Only a denatured duplex could interact with the serine cluster (highlighted in green) and be favourably accommodated by the bridge in this alternative conformation. (B) RNase assay of wild-type KSHV SOX in the presence of 10 mM MgCl₂ or MnCl₂. (C) Fluorescence anisotropy binding assays of SOX involving single- (ssRNA-5′P) and double-stranded (dsRNA-5′P) RNA oligonucleotides. The results obtained for dsDNA-5′P are included for comparison.

Figure 7.

(A) Cartoon showing the locations of the non-catalytic host shutoff residues (red) within the complex (the DNA is shown in light blue). (B) RNase assays of all mutants (excluding T24I) known to reduce/abolish host shutoff. (C) Plasmid exonuclease assays of catalytic and non-catalytic shutoff mutants.