Structural and biochemical characterization of CRN-5 and Rrp46: an exosome component participating in apoptotic DNA degradation

Abstract

Rrp46 was first identified as a protein component of the eukaryotic exosome, a protein complex involved in 3' processing of RNA during RNA turnover and surveillance. The Rrp46 homolog, CRN-5, was subsequently characterized as a cell death-related nuclease, participating in DNA fragmentation during apoptosis in Caenorhabditis elegans. Here we report the crystal structures of CRN-5 and rice Rrp46 (oRrp46) at a resolution of 3.9 A and 2.0 A, respectively. We found that recombinant human Rrp46 (hRrp46), oRrp46, and CRN-5 are homodimers, and that endogenous hRrp46 and oRrp46 also form homodimers in a cellular environment, in addition to their association with a protein complex. Dimeric oRrp46 had both phosphorolytic RNase and hydrolytic DNase activities, whereas hRrp46 and CRN-5 bound to DNA without detectable nuclease activity. Site-directed mutagenesis in oRrp46 abolished either its DNase (E160Q) or RNase (K75E/Q76E) activities, confirming the critical importance of these residues in catalysis or substrate binding. Moreover, CRN-5 directly interacted with the apoptotic nuclease CRN-4 and enhanced the DNase activity of CRN-4, suggesting that CRN-5 cooperates with CRN-4 in apoptotic DNA degradation. Taken together all these results strongly suggest that Rrp46 forms a homodimer separately from exosome complexes and, depending on species, is either a structural or catalytic component of the machinery that cleaves DNA during apoptosis.

FIGURE 1.

Sequence alignment of rice, human, and C. elegans Rrp46 (CRN-5) and archaeal Rrp41. Eukaryotic Rrp46 shares significant sequence identity with archaeal Rrp41. Sequences listed here are Rrp46 from Oryza sativa (oRrp46), Caenorhabditis elegans (CRN-5), and human (hRrp46), and Rrp41 from P. abyssi (PabRrp41) and S. solfataricus (SsoRrp41). The secondary structure generated from the crystal structure of oRrp46 is shown above the sequence. TT represents β-turns. Identical and similar residues are shown in red. Mutational residues constructed in this study are shown in yellow. The residues involved in homodimeric interactions in CRN-5 are shown in cyan. The alignment was generated by ClustalW (http://www.ch.embnet.org/software/ClustalW.html) and ESpript (Gouet et al. 2003).

FIGURE 2.

Recombinant and endogenous Rrp46/CRN-5 are dimeric proteins. (A) The purified recombinant Rrp46 proteins from different species were analyzed by SDS-PAGE. CRN-5 was less homogeneous with strong linked dimers that could not be dissociated into monomers on SDS-PAGE under the reduced condition. Molecular weight markers are shown as indicated. (B) Gel filtration (Superdex 200) profiles for hRrp46, oRrp46, and CRN-5. The recombinant Rrp46/CRN-5 all eluted with a size of a homodimer in a buffer of 25 mM Tris-HCl (pH 7.6), 150 mM NaCl and 1 mM β-mercaptoethanol. oRrp46 shifted into monomers with the addition of 5% PEG 3350 (oRrp46/PEG). (C) Dynamic light scattering (DLS) analysis of native Rrp46 proteins. The estimated molecular weights of each protein (in parentheses) are listed in the figure. (D) Analytical ultracentrifugation (AUC) of recombinant hRrp46 and CRN-5. The sedimentation velocities of hRrp46 and CRN-5 were recorded in a buffer of 25 mM Tris-HCl (pH 7.6), 150 mM NaCl, and 1 mM β-mercaptoethanol under 40,000 rpm ultracentrifugation. The calculated molecular weights of each protein are listed in the figure. See also Supplement Data (Supplemental Fig. S1). (E) Endogenous Rrp46 either formed homodimers, or associated with a protein complex. Human (293T) and rice (callus) cell extracts were fractionated on a Superdex 200 gel-filtration column in a buffer with (+β-ME) or without β-mercaptoethanol and analyzed by Western blotting using rabbit anti-hRrp42, anti-hRrp46 or anti-oRrp46 serum. The hRrp42 was used as a control to show that it only associated with a protein complex. The cell extract before fractionation is shown in lane L.

FIGURE 3.

Rrp46 DNA-binding and DNase activity assays. (A) Gel-retardation assays show that Rrp46 binds to double-stranded DNA. Rrp46 (0, 5, and 10 μM) were incubated with 20-mer DNA (20 nM) in the buffer of 20 mM HEPES (pH 7.0) and 5 mM EDTA. The protein–DNA complexes were indicated by a bracket marked at the right of the panel. (B) DNA cleavage assays show that only rice oRrp46 cleaved double-stranded linear DNA. Rrp46 (1 μM) was incubated with the linear dsDNA (30 nM) in the DNase reaction buffer (20 mM HEPES [pH 7.0], 100 mM NaCl, 1 mM CaCl₂, and 1 mM DTT) at 37°C for the indicated periods, followed by agarose gel electrophoresis. A control reaction without any Rrp46 added is shown in lane C. RNA-specific RNase T was used as a negative control. (C) Rice oRrp46 is a metal ion-dependant hydrolytic DNase. oRrp46 (1 μM) was incubated with linear dsDNA (30 nM) at 37°C for 20 min in the presence or absence of MgCl₂, CaCl₂, and NaH₂PO₄. All the reaction buffers contained 20 mM HEPES (pH 7.0), 100 mM NaCl, and 1 mM DTT with addition of 2 mM MgCl₂ in lane Mg, 1 mM CaCl₂ in lane Ca, 5 mM NaH₂PO₄ in lane Pi, and combinations of each of two factors in the rest of the lanes. A control reaction without metal and phosphate ion added is shown in lane Buffer. (D) Rice oRrp46 cleaves plasmid DNA. oRrp46 (0.2 μM) was incubated with 100 ng plasmid DNA at 37°C for up to 60 min in the DNase reaction buffer. TREX2, an exonucleolytic hydrolase, was used as a negative control: o, open circular; l, linear; s, super coiled. (E) Rice oRrp46 lost its DNase activity in the presence of PEG3350. oRrp46 (0.2 μM) was incubated with 100 ng plasmid DNA at 37°C for 30 min in the presence of various concentration (0%–20%) of PEG3350. Lane C is for loading control: o, open circular; l, linear; s, super coiled.

FIGURE 4.

Rrp46 RNA-binding and RNase activity assays. (A) Gel retardation assays show that only rice oRrp46 binds to RNA. In vitro-transcribed RNA (500 ng) was incubated with Rrp46 (2.5 or 10 μM) in the RNA-binding buffer (50 mM Tris, pH 8.0, 10 mM EDTA, 1 U/μL RNaseIN) on ice for 30 min. The up-shifted RNA was indicated by an arrow. (B) RNA cleavage assays show that rice oRrp46 degrades RNA substrates. The bracket marked at the right indicates the degraded RNA. Rrp46 (5 μM) was incubated with the same substrate used in Figure 4A in the RNase reaction buffer (50 mM Tris [pH 8.0], 50 mM KCl, 2 mM MgCl₂, 5 mM Na₂HPO₄, and 1U/μL RNaseIN) at 30°C for 1 h followed by agarose gel electrophoresis. (C) Rice oRrp46 is a metal ion- and phosphate-dependent RNase. oRrp46 (2 μM) was incubated with 5′-labeled 20-mer ssRNA in conditions mentioned in Figure 3C. A control reaction with both cofactors but without any Rrp46 added is shown in lane C. (D) Thinlayer chromatography (TLC) separation of the final RNA reaction products cleaved by rice oRrp46. The internal ³²P-labeled RNA substrate was incubated with oRrp46 for 60 min with or without the presence of 5 mM phosphate (−Pi and +Pi). Positions of the cofractionated unlabeled ADP and AMP are shown on the right.

FIGURE 5.

Crystal structures of the monomeric oRrp46 and dimeric CRN-5. (A) The crystal structure of the monomeric oRrp46. The catalytic residue, E160, and substrate-binding residues, K75 and Q76, in the active site are shown as a stick model. (B) Superposition of the active sites of several RNase PH proteins shows that a conserved acidic residue is located at the same position: E160 in oRrp46 (this study), E148 in CRN-5 (this study), D180 in PaRrp42 (2PO1, chain A), E174 in hRrp46 (2NN6, chain D), and E219 in hRrp42 (2NN6, chain E). The bound RNA nucleotide cocrystallized with PaRrp41 is displayed as a ball-and-stick model. (C) The crystal structure of the dimeric CRN-5. Individual monomers are shown in blue and cyan, respectively. Comparison of RNA-binding residues in RNase PH proteins. The two basic residues, R96 and R97, bind to the phosphate groups of RNA (displayed as a ball-and-stick model) in PaRrp41 (green, 2PO1, chain A). Corresponding residues in oRrp46 (Q76 and K75) are more suitable for RNA binding compared with those in the inactive enzymes of hRrp46 (V89 and A90) and hRrp42 (D100 and L101). The strong RNA-binding residues are circled in solid lines and the weak RNA-binding residues are circled in dashed lines.

FIGURE 6.

In vitro nuclease activity assays of wild-type and mutated oRrp46. (A) DNA (309 bp, 30 nM) cleavage assays by the wild type, active-site mutant (E160Q), and nucleotide-binding-site mutant (K75E/Q76E) of oRrp46 (1 μM). All the reaction conditions were identical to those in Figures 3 and 4. (B) Time-course RNA cleavage assays by the wild type, E160Q, and K75E/Q76E mutant of oRrp46. The RNA markers that varied from 1 to 20 nucleotide(s) are labeled on the left of the panel. (C) DNA gel shift assays of the wild type and K75E/Q76E mutant of oRrp46. The protein–DNA complex is indicated with an arrow. (D) RNA gel shift assays of the wild type and K75E/Q76E mutant of oRrp46. The up-shifted RNA is indicated by arrows. The bracket at the right indicates the up-shifted RNA.

FIGURE 7.

CRN-5 interacts with CRN-4 and enhances CRN-4's DNase activity. (A) His-tag pull-down assays of CRN-4 by His-tagged CRN-5. The His-tagged CRN-5 was incubated with or without CRN-4, and then the complex was pulled down in a Ni-NTA spin column. The eluted solution was analyzed by Western blotting using anti-CRN-4 (α-CRN-4) and anti-6xHis (α-6xHIS) antibodies. (B) Assays of CRN-4 DNase activity in the presence or absence of CRN-5. A linear 309-bp double-stranded DNA (30 nM) was incubated with CRN-4 (1 μM), CRN-5 (2 μM), or both together, and the DNA digests were analyzed and quantified by gel electrophoresis. CRN-4 cleaved linear dsDNA more efficiently in the presence of inactive CRN-5 (6.5% vs. 13% DNA remained).

FIGURE 8.

Electrostatic surface potential and substrate-binding sites of active and inactive RNase PH proteins. The active RNase PH proteins all have positive surfaces at the RNA-binding site (circled), including PaRrp41 (2PO1, chain A), RNase PH II domain of PNPase (3CDI), and oRrp46 (this study). The RNA bound in PaRrp41 is displayed as a stick model. On the contrary, the inactive RNase PH proteins have neutral or acidic surfaces at the RNA-binding sites (circled), including PaRp42 (2PO1, chain B), RNase PH I domain of PNPase (3CDI), and hRrp46. The color scale of the surface potential was set from −75 kT/e (red) to 75 kT/e (blue), as calculated by Pymol (DeLano Scientific LLC, http://www.pymol.org).