Specific recognition of RNA/DNA hybrid and enhancement of human RNase H1 activity by HBD

Abstract

Human RNase H1 contains an N-terminal domain known as dsRHbd for binding both dsRNA and RNA/DNA hybrid. We find that dsRHbd binds preferentially to RNA/DNA hybrids by over 25-fold and rename it as hybrid binding domain (HBD). The crystal structure of HBD complexed with a 12 bp RNA/DNA hybrid reveals that the RNA strand is recognized by a protein loop, which forms hydrogen bonds with the 2'-OH groups. The DNA interface is highly specific and contains polar residues that interact with the phosphate groups and an aromatic patch that appears selective for binding deoxyriboses. HBD is unique relative to non-sequence-specific dsDNA- and dsRNA-binding domains because it does not use positive dipoles of alpha-helices for nucleic acid binding. Characterization of full-length enzymes with defective HBDs indicates that this domain dramatically enhances both the specific activity and processivity of RNase H1. Similar activity enhancement by small substrate-binding domains linked to the catalytic domain likely occurs in other nucleic acid enzymes.

Figure 1

Hs-HBD preferentially binds to RNA/DNA hybrids. (A) EMSA of nucleic acid (n.a.) binding. Concentrations from 32 to 0.036 μM of the 1:1 mixtures of Hs-HBD and 12 bp RNA/DNA, dsRNA or DNA of equivalent sequences were resolved on a native TBE polyacrylamide gel. Bands corresponding to HBD–nucleic acid complexes are indicated by arrows or brackets. (B) Sedimentation equilibrium profiles of 1:1 mixture of ∼5 μM Hs-HBD and the 12 bp RNA/DNA (red/blue), dsRNA (red) or dsDNA (blue) at 22 k.r.p.m. and 4°C. Data (red circles) were collected at 280 nm. The solid blue lines show the slope expected for the free duplex nucleic acid, and the solid green lines the slope expected for the 1:1 protein–nucleic acid complex. Complete data analyses are detailed in the Supplemental data.

Figure 2

The structure of Hs-HBD complexed with a 12 bp RNA/DNA. (A) The content of an asymmetric unit. Each RNA/DNA hybrid (RNA in red, DNA in blue) interacts with three protein molecules. (B) Sequence alignment of HBD domains of RNases H1 from human (Hs), mouse (Mm), chicken (Gg), fruitfly (Dm), yeast (Sc) and B. halodurans (Bh). Conserved residues are highlighted in blue for DNA phosphate interaction, green for deoxyribose interaction, yellow for the hydrophobic core and purple for Arg's that stabilize the helices and RNA-binding loop. The RNA-binding loop is shown in boldface. (C) The structure of Hs-HBD. Side chains forming the hydrophobic core (mainly aromatic) and two Arg's capping the C-termini of helices are shown in sticks. (D) Electrostatic surface potential of Hs-HBD calculated using GRASP (Nicholls et al, 1991). The positively charged residues (blue) in a cluster are labelled.

Figure 3

Interactions between HBD and RNA/DNA. (A) The nucleic-acid interface. Residues interacting with the DNA strand and R35 stabilizing the RNA-binding loop are shown in ball-and-sticks with the same colour coding as in Figure 2B. Base pairs interacting with HBD are numbered 1–5. (B) The close-up view of the RNA-binding loop. The 2′-OH groups interacting with the loop are shown as red spheres. The dashed lines indicate hydrogen bonds. (C) W43 and F58 shown in green ball-and-sticks under the molecular surface closely interact with deoxyriboses of the DNA (blue) and would clash with the 2′-OH groups (red) of a modelled RNA (light pink). Ribonucleotides (R1 and R2) and deoxyribonucleotides (D3-D5) are numbered as in (A).

Figure 4

Mutagenesis studies and a model of RNA/DNA degradation by RNase H1. (A) EMSA of mutant Hs-HBDs. Mixtures of a mutant protein and the 12 bp hybrid (4 μM each) were resolved on a native TBE polyacrylamide gel. The smeared band corresponding to protein-hybrid complexes is indicated by brackets. (B) Cleavage assay of the 5′ end-labelled 36 bp RNA/DNA by RNase HC and full-length Hs-RNases H1 with mutations in HBD. The proteins are indicated on the top of the panel. The hybrid substrate has two bands. (C) A bar graph of specific activity (blue) and processivity (green) on the uniformly ³²P-labelled poly-rA/poly-dT substrate. The relative activity is normalized to WT RNase H1 (100%). The error bars in C to F represent the standard deviation of at least three independent measurements. (D) pH, (E) Mg²⁺ concentration and (F) salt concentration dependence of human (black) and mouse (red) RNase H1 activity on uniformly ³²P labelled poly-rA/poly-dT substrate. (G) A cartoon of RNA/DNA degradation by RNase H1. HBD (cyan) enhances substrate recognition and probably anchors the enzyme for multiple rounds of cleavage by RNase HC (green) leading to an increased processivity.

Figure 5

Non-sequence-specific binding of DNA, RNA/DNA and dsRNA. (A) BAF (PDB: 2BZF) with two HhH motifs (each with the first helix shown in blue and the second in green) interacting with DNA backbones across the minor groove. Most interactions are formed by the main-chain amides. (B) Hs-HBD reported in this study. (C) dsRBD from Xenopus leavis RNA-binding protein (PDB: 1DI2). It interacts with two successive minor grooves of the dsRNA, and two protein helical dipoles interact with the RNA backbones across the major groove.