RNA metabolism has been found to play important roles in cell viability1 and dysregulation of mRNA life-time is implicated in several human diseases, including cancer, inflammation, and Alzheimer's disease.2 However, RNA metabolism is complicated and requires a wide variety of discrete ribonucleases (RNases) with both endo- and exo-cleaving activities to complete the turnover of aged RNA molecules. Six exo-ribonuclease superfamilies, as well as various subfamilies, were recently identified through extensive data mining.3 Oligoribonuclease (Orn) was found to belong to the DEDDh-type exo-ribonuclease superfamily, and is exclusively responsible for degrading small oligoribonucleotides of 2–5 residues in length to mononucleotides.3 It is different from other known exo-RNases of E. coli in that knockout of this gene (orn) can lead to cellular lethality.1, 4
XC847 from the plant pathogen Xanthomonas campestris pv. campestris str. 17 (Xcc) is classified as belonging to the Orn family in the Pfam database.5 It contains 194 amino acids, and shares a high degree of identity and similarity with other identified Orn3 throughout the DEDDh domains including motifs I, II, and III that are unique to Orns.3 Until now, Orns are less studied structurally. Only the crystal structure of Haemophilus influenzae Orn has been deposited in the PDB (1J9A), but its structural details have not yet been reported. Because of the importance of this type of RNase, we have solved the Xanthomonas campestri Orn (XC847) structure to a resolution of 2.1 Å. Although the overall architecture of this Orn is similar to other reported 3′–5′ DNases,6–9 with the active DEDD residues also located in a similar environment, there is one significant difference between them; the helix H in XC847 is oriented opposingly from the similar helices in all reported 3′–5′ DNases discovered so far, possibly to prevent the steric hindrance of accommodating oligoribonucleotide substrates.
Methods and Materials
The cloning of XC847 gene fragment into a pMCSG7 vector using a ligation-independent cloning approach,10 along with its expression and purification, has been described in a previous communication.11 The final target protein (194 aa) has a greater than 99% purity and contains only an extra tripeptide (SNA) at the N-terminal end. Crystallization screening was performed using the sitting-drop vapor-diffusion method in 96-well plates (Hampton Research) at 295 K. Pyramid-like crystals grew to maximum dimensions of 1 × 1 × 0.5 mm3 after three days.11 The crystals were mounted in a capillary for X-ray diffraction data collection. The Matthews coefficient for XC847 is 2.10 Å3/Da, and the estimated solvent content is 41.9%. The data collection statistics has been published before.11
The AMoRe12 program was used in the rotation and translation search and the deposited structure 1J9A was chosen as the search model. The search model was placed in a P1 cell with a = b = c = 80 Å and α = β = γ = 90°. Data between 15.0 and 3.0 Å and a Patterson radius of 20 Å were used for all rotation and translation function calculations. Because a significant solution could be found in the initial trial, the sequence of search model was replaced by XC847 to optimize the result. After rigid body refinement, the optimal solution was determined at Euler's angle α = 8.5°, β = 26.65°, γ = 50.34° and x = −0.1099, y = −0.1581, z = 0.1370 with a correlation coefficient of 54.2% and an R-factor of 48.1%. Structural refinement and model adjustment were carried out using CNS13 and XtalView/Xfit.14 Simulated annealing omit map were then used to reduce the model bias. The final electron density map can be traced from residue 8 to 186 without interruption. The final model includes 181 residues, 70 water molecules, and one magnesium ion. The Ramachandran plot reveals that 94% of the residues are in the most favored region, with an additional 5.5% in the allowed region. The only residue that deviates somewhat from the ideal ϕ/ψ angles is Asp62 (−62.5° and −143.7°, respectively), which uses its carbonyl oxygen atom to form a H-bond with the Arg66 Nε atom. The refined structure coordinates of XC847 have been deposited (PDB: 2GBZ) and the refinement statistics is listed in Table I.
Table I. Statistics of Structural Refinement of XC847
Values in parentheses are for the outmost shell.
Resolution range (Å)
Data cutoff (σF)
Completeness of used reflections (%)
Number of used reflection
Rfree test set size (%)
Number of nonhydrogen atoms
Bond lengths (Å)/bond angles (°)
The docking of U5 oligoribonucleotide into the XC847 dimer crevice was carried out using the DSModeling 1.1 program (Accelrys, USA). The U5 template was generated by the InsightII program (Accelrys, USA) and the docking started by first fixing the 3′-end nucleotide of U5 substrate into the XC847 active site.6 The direction of the U5 substrate was established as shown in Figure 1(b). The resulting complex was then energy-minimized using the CHARMm force field and the steepest descent protocol in 500 cycles to get the final model.
Results and Discussion
The final model of the XC847 Orn monomer comprises nine α-helices (αA to αI) and five β-strands (β1–β5) [Fig. 1(a)]. It folds into the co-called DnaQ fold family containing a core of five-stranded β sheet aligned in the order of β5–β4–β1–β2–β3 that are flanked by α-helices [Fig. 1(b)].3 All Orns were found to contain well conserved ExoI, ExoII, ExoIII, and ExoIV motifs that are unique to Orn.3 This is also true for XC847 [Fig. 1(a)], and such motifs were boxed in green in Figure 1(a). Overall, these motifs exhibit larger than 40% identity among the XC847 and those of E. coli, H. influenzae, M. tuberculosis, yeast, and human.3 One magnesium ion was found coordinating well with the DEDDh active site carboxylate groups of Asp15, Glu17, Asp166, and the oxygen atom of U51Op group, with a bonding distances of ∼2.12 Å. These coordination interactions were shown as a close-up view in Figure 1(c).
XC847 forms a dimer in crystal with a 2-fold noncrystallographic symmetry [Fig. 2(a)]. In addition to a disulfide bond of Cys113–Cys113′, and intermolecular hydrophobic interactions among amino acid residues Leu135, Ile140, Leu143, Val151, Leu174, Met181, and Leu184, several H-bonds and salt bridges were also observed, including the Glu142 Oε–Arg133′ Nζ, Asp24 Oδ–His120′ Nε, Thr139 Oγ–Asp136′ HN, and Phe180 CO–Leu184′ HN atom pairs. These interactions contribute to the formation of a stable homodimer. Interestingly, the surface charge distribution in this homodimer is very uneven, with the active site highly acidic [pointed by a red arrow in Fig. 2(b)], and the substrate binding site highly basic (pointed by a cyan arrow). This may partially accounts for the processive oligonucleotide cleavage mechanism as described below. Many residues from both monomers were found to interact with the modeled U5 oligonucleotide [Fig. 2(d)].
A structural homolog search was performed using the DALI server16 with the coordinates of XC847. The top homologue of the XC847 protein include the E. coli exonuclease I (1FXX),8 the human antiviral ribonuclease (1WLJ),6 the ε subunit of E. coli DNA polymerase III (1J54),7 and the nuclease subunit of the mRNA deadenylate complex (1UOC),9 with a root-mean-square deviation (RMSD) of 1.41 Å based on the Cα of 113 residues, 1.46 Å on 71 residues, 1.51 Å on 100 residues, and 1.61 Å on 93 residues, respectively [Fig. 2(c)]. Yet the sequence identity of XC847 with them is very low; with a mere 10% with 1FXX, 14% with 1WLJ, and 19.7% with 1J54, respectively. However, the DEDD residues in the active site of XC847 do overlap very well with other DEDD type DNases and RNases mentioned above, with a RMSD of only 0.25 Å for the four acidic residues between the XC847 and 1WLJ structures (data not shown). These data indicate that XC847 also exploits similar oligonucleotide cleavage mechanism17 as those employed in the DEDDh type DnaQ superfamily.7 However, one major difference between the XC847 and its structural homologue mentioned above is noted, i.e. the opposing orientation of helices H [the hinge point was marked by a dotted arrow in Fig. 2(c)]. In XC847, the spacer between helix G and helix H contains only three amino acids, and the helix H is oriented leftward (in the figure orientation). While its structural homologue mentioned above encompass longer spacers, with all αH helices orienting rightward relatively [Fig. 2(c)]. Hence the XC847 αH helix (in red) forms an angle of ∼90° with these αH helices (in yellow, green and blue, respectively), with the Cα of the first αH residue of XC847 shifting ∼15 Å relative to those in the 1WLJ, 1J54, and 1FXX structures. Since the mode of attack of Orn appears to be processive,18 we propose that the small RNA oligomers must first bind with Orn before they can be cleaved. This binding is mainly achieved by the electrostatic force [the left side in Fig. 2(b)] and the coordination force exhibited by the active site acidic residues and Mg+2 ion [the right side in Fig. 2(b)]. The shifting of αH helix away from the active site in XC847 may thus provide the room necessary for accommodating the U5 substrate. To support this point, we have modeled a XC847/U5 complex by docking the U5 substrate into the XC847 dimer crevice by fixing the 3′-end nucleotide in the active site6 and energy-minimizing the resulting complex. Consistent with our proposal, the U5 substrate can be fit into the crevice of the XC847 dimer very well [Fig. 2(d)]. Several specific salt bridge and H-bonds were observed in the modeled XC847/U5 complex, including the Asn8 Nγ–U1O1p, Arg101 Nε–U1O1p, Arg145 Nε–U1O2′H, Arg145 NH1–U1O2′H, and Lys157 Nε–U4O2′H atom pairs [Fig. 2(d)]. In contrast, the U5 substrate experiences considerable steric hindrance with the amino acid residues located in the αG–αH spacer and the αH helix of other DEDD type DNase and RNase structures (not shown).
Based on extensive analysis, four well conserved motifs for Orn have been proposed. However, the functional significance of these motifs remain unknown.3 After detailed structural studies of XC847, it becomes clear that these conserved residues may play important roles in stabilizing the Orn dimer or in interacting with the U5 substrate, in addition to the classical role of the Asp12, Glu14, Asp103, Asp167, and His161 residues in forming an active center. For examples, amino acids Leu135, Leu143, Ile140, and Leu174 were found to form a hydrophobic cluster, and side chains of Glu142/Arg133 and Thr139/Asp136 were found to form salt bridges and H-bonds in the dimer interface, while side chains of Arg145, and Lys157 were found to interact with the ribose 2′OH or phosphate oxygen atoms of the U5 substrates. The crystal structure of XC847 thus provides an outlet for further studies regarding its substrate binding and processive cleavage.
We thank the Core Facilities for protein X-ray crystallography in the Academia Sinica, Taiwan, in the National Synchrotron Radiation Research Center, Taiwan, and in the SPring-8 Synchrotron facility in Japan for assistance of X-ray data collection. The National Synchrotron Radiation Research Center is a user facility supported by the National Science Council, Taiwan, Republic of China, and the Protein Crystallography Facility is supported by the National Research Program for Genomic Medicine. This work is supported by NCS grants 94-2113-M005-003 and 95-2113-M005-018 to SHC.