|Home | About | Journals | Submit | Contact Us | Français|
Messing et al. (2009) report the homodimeric structure of the Bdellovibrio bacteriovorus RppH pyrophosphohydrolase, which hydrolyzes the mRNA 5′ triphosphate to initiate bacterial mRNA decay. These structures reveal insights into BdRppH substrate recognition and analogies to eukaryotic decapping enzymes.
All cellular messenger RNAs (mRNAs) possess distinct intrinsic half-lives and are ultimately degraded. Cells have evolved intricate strategies to stabilize as well as degrade their mRNAs through defined pathways. In eukaryotes, the 5′ end of the mRNA is protected from 5′ to 3′ exonucleolytic activity by the presence of the 5′ cap structure. In prokaryotes, the 5′ end of the newly transcribed mRNA is not further modified and retains the 5′ triphosphate. Recent work indicates that the 5′ triphosphate in prokaryotic mRNAs fulfills a protective function similar to that of the cap structure in eukaryotes (Celesnik et al., 2007). Identification of RppH (ORF176/NudH/YgdP) as a prokaryotic functional homolog of a decapping enzyme that similarly removes a diphosphate from the mRNA 5′ end to generate a 5′ monophosphate mRNA (Deana et al., 2008) further reveals the strategic conservation of mRNA decay between prokaryotes and eukaryotes (Figure 1).
RppH (ORF176/NudH/YgdP) belongs to the Nudix superfamily of enzymes that hydrolyze Nucleoside diphosphates linked to other moiety X. Members of the family, found in species from bacteria to eukaryotes, share a highly conserved 23 residues long Nudix signature motif GX5EX7REUXEEXGU where U is a hydrophobic residue and X any residue (McLennan, 2006). Messing et al. (2009) resolved the structure of the predatory bacterium Bdellovibrio bacteriovorus RppH (BdRppH) protein in the apo form and in complex with GTP. The structure reveals an asymmetric head-to-head homodimer with a 20 residue interface on each monomer mediated through hydrophobic interactions and hydrogen bonding. The co-crystal structure reveals the GTP purine ring is arranged in a syn conformation analogous to that observed with the purine ring of the adenosine in m7GpppA substrate complexed with the Xenopus laevis X29 nuclear decapping enzyme (Scarsdale et al., 2006) with both purines positioned within a similar cleft. The ring is stabilized by Phe52 and Asn136. which coordinate it in a manner equivalent to Gln184 in X29. The α- and γ-phosphates of GTP are further coordinated respectively by BdRppH Arg40 and Lys56. Moreover, as would be expected for an enzyme that hydrolyzes the triphosphate moiety at the 5′ end of an mRNA, the 3′ hydroxyl of the guanosine appears unencumbered and positioned towards the outer surface of the protein potentially providing an exit site for a linked RNA molecule. Interestingly, the authors state BdRppH lacks an RNA binding structure referred to as the Box B α-helix at the carboxyl terminus of the Nudix fold of X29, but propose a positive patch along helix α2 and α3 might serve this role. The overall similarities in homodimer formation and substrate binding suggest a mechanistic conservation between BdRppH and X29 Nudix proteins in pyrophosphate hydrolysis.
All Nudix enzymes require divalent cations for catalysis (McLennan, 2006). The ions appear to be coordinated by conserved residues in the Nudix motif. In the case of BdRppH, Mg2+ is the preferred cation. Each monomer of BdRppH binds three Mg2+ ions coordinated by residues Gly54, Glu70, Glu73, and Glu74 and multiple water molecules. In contrast, the prototypic prokaryotic MutT Nudix protein requires two divalent cations coordinated by a glycine and four glutamates including Glu53 which functions as the general base (Mildvan et al., 2005). Interestingly, unlike Glu53 in MutT, substitution of the equivalent residue in BdRppH, Glu70, has a modest effect on kcat and is unlikely the catalytic base as it also directly coordinates two cations. By extrapolation to the catalytic base at structurally equivalent positions within the GDPMH and ADPRase Nudix proteins (Mildvan et al., 2005), the direct involvement of two histidines, within loop L6 as potential residues responsible for the catalytic hydrolysis step of BdRppH were tested. However, single and double mutants of His115 and His116 ruled out a role for these histidines as the general base for the BdRppH catalytic step as well. The authors propose that flexibility of the L6 loop may enable alternative residues or a solvent hydroxyl to fulfill the catalytic function. Further mutagenesis studies are needed to precisely delimit the residues involved in the catalysis step.
Unlike the cytoplasmic Dcp2 mRNA decapping protein that only functions on a cap structure linked to an RNA moiety and lacks hydrolytic activity on a nucleotide substrate (Wang et al., 2002), BdRppH contains a broader substrate specificity and can function as a dGTPase (Steyert et al., 2008) in addition to its mRNA 5′ triphosphate pyrophosphohydrolase activity. BdRppH dual substrate hydrolytic activity is not restricted to in vitro reactions as it can complement the dGTPase deficiency in ΔmutT E. coli (Steyert et al., 2008) as well as mRNA pyrophosphohydrolase activity to restore mRNA decay in ΔrppH E. coli (Deana et al., 2008). Why E. coli have taken a specialization route to utilize two proteins, MutT and RppH for the dGTPase and mRNA pyrophosphohydrolase activities respectively, while both functions can be carried out by BdRppH in B. bacteriovorus is not apparent. Nor are the potential modulatory pathways that might exist to shunt BdRppH into a dGTPase or mRNA pyrophosphohydrolase function in cells.
Remarkably, an additional wrinkle was recently added to the pyrophosphohydrolase class of enzymes. Xiang et al. (2009) reported that removal of a pyrophosphate from an mRNA 5′ triphosphate is not restricted to bacteria nor is it restricted to Nudix proteins. Despite the lack of a discernable nuclease domain, crystal structures of the yeast Rat1 interacting protein, Rai1 and the mouse DOM3Z homolog revealed a potential octahedral catalytic pocket capable of coordinating a metal ion. Biochemical studies confirmed a catalytic activity for Rai1 and interestingly demonstrated it can hydrolyze the triphosphate group at the 5′ end of an mRNA to remove a pyrophosphate and generate a 5′ monophosphorylated mRNA, a function analogous to RppH and BdRppH. The functional significance of Rai1 pyrophosphohydrolase activity in eukaryotes where mRNAs are capped at the 5′ end in unclear although a role in 5′ end quality control was proposed by the authors. These finding further highlight the functional similarities that exist between prokaryotic and eukaryotic mRNA decay mechanisms, and emphasizes our current rudimentary understanding of the complexity of nucleases involved in mRNA decay.
The BdRppH crystal structure represents the first example of a bacterial Nudix pyrophosphohydrolase and provides important clues into its molecular mechanism and potential avenues for structure-based therapeutic antibacterial approaches. These studies raise interesting avenues for future studies including how does BdRppH accommodate and hydrolyze more than one substrate while the corresponding mammalian decapping enzymes only function on capped mRNA? Are additional proteins involved in regulating these distinct pyrophosphohydrolase activities in bacteria? For example are there 5′ end triphosphate binding proteins in prokaryotes analogous to the eIF4E cap binding protein in eukaryotes to protect the 5′ end? Do prokaryotes contain additional proteins that hydrolyze mRNA 5′ triphosphate? As our long held belief that prokaryotic and eukaryotic mRNA decay pathways are distinct begins to crumble, it is becoming more apparent that distant but analogous mechanisms have converged evolutionarily.