Removal of the mRNA 5′ cap is an important step in the regulation of mRNA stability. mRNAs are degraded by at least two distinct exonucleolytic decay pathways, one from the 5′ end, and the second from the 3′ end. Two major cellular decapping enzymes have been identified, and each primarily functions in one of the two decay pathways. The Dcp2 decapping enzyme utilizes capped mRNA as substrate and hydrolyses the cap to release m7GDP (N7-methyl GDP), while a scavenger decapping enzyme, DcpS, utilizes cap dinucleotides or capped oligonucleotides as substrate and releases m7GMP (N7-methyl GMP). In this review, we will highlight the function of different decapping enzymes and their role in mRNA turnover.
Regulation of mRNA stability has recently emerged as a critical mediator of gene expression. The steady-state accumulation of any given cellular mRNA is determined by both its synthesis and degradation. The stability of mRNAs in eukaryotes can vary from minutes to days with a median half-life of 21 min in yeast and 10 h in mammals .
The decay of mRNA is not a stochastic process and proceeds through distinct decay pathways. There are two major exonucleolytic pathways by which mRNAs are degraded in eukaryotic cells. They are either degraded from the 5′ end or from the 3′ end. Both pathways are initiated by the removal of the 3′ polyadenylated tail [2,3,4–5]. In the 5′ decay pathway, deadenylation can trigger 5′ decapping carried out by the mRNA-decapping complex Dcp1–Dcp2, in which Dcp2 is the catalytic subunit [6,7,8,9,10–11] (Figure 1). The exposed 5′ end is subjected to digestion by the 5′-exoribonuclease, Xrn1, in the cytoplasm [2,12]. In the 3′ end decay pathway, the mRNA body continues to be degraded from the 3′ end by the exosome complex following deadenylation [13,14]. The resulting capped oligonucleotide of ten nucleotides or fewer, is a substrate for the scavenger decapping enzyme, DcpS [15,16]. Therefore two classes of decapping enzymes have been identified to date, each primarily functioning in one of the two pathways.
General eukaryotic mRNA-decay pathways
Specificity of decapping
An interesting property of the two decapping enzymes is their substrate specificity: Dcp2 functions on capped RNA, but not cap structure [10,11,17], while DcpS functions on cap structure, but not capped RNA [15,16,18]. Dcp2 is an RNA-binding protein that preferentially utilizes a methylated cap linked to an RNA longer than 25 nucleotides as substrate, and hydrolyses the cap to generate m7GDP (N7-methyl GDP) and a monophosphate-terminated mRNA [8,10,11,17]. Consistent with Dcp2 being an RNA-binding protein, it binds RNA in vitro , and its activity is competed by uncapped RNA in both mammals and yeast [8,9]. Thus this mRNA-decapping enzyme recognizes its substrate by interacting with both the methylated cap structure and the RNA moiety. Conversely, the scavenger decapping enzyme is unable to utilize capped RNA as substrate and only functions on short capped oligonucleotides of fewer than ten nucleotides generated by 3′ end decay. The resulting hydrolysis products are m7GMP (N7-methyl GMP) and a 5′ diphosphate-oligonucleotide mRNA [14,16]. The N7-methyl moiety was also essential for substrate specificity, as the unmethylated cap was not capable of competing for the scavenger decapping activity and neither did it serve as a substrate for direct hydrolysis .
Biochemical and structural analysis of Dcp1 and Dcp2
Dcp2 is a member of the Nudix (nucleoside diphosphate linked to some other moiety X) hydrolase family [7,11,17,19]. The Nudix proteins are evolutionarily conserved and are present in diverse species from bacteria to mammals. The hallmark of this family is the presence of a highly conserved 23-residue Nudix motif, GX5EX7REUXEEXGU (where U is a hydrophobic residue and X is any residue) contained within a larger ∼100-amino-acid Nudix fold  (Figure 1). Both yeast and human Dcp2 contain a functional Nudix motif, as mutations in this region considerably decreased decapping activity [7,11,17].
Characterization of the human Dcp2 revealed that, in addition to the Nudix motif, two additional evolutionarily conserved regions were present (Figure 2). The first termed Box A, spans amino acids 12–49 and is essential for decapping fidelity. Truncation of Box A results in a Dcp2 protein compromised in its ability to hydrolyse the cap exclusively between the β and α phosphates to generate m7GDP. Dcp2 also cleaves between the γ and β phosphates to release m7GMP in the absence of Box A . A second evolutionarily conserved domain, termed Box B, is positioned C-terminally to the Nudix motif and located at residues 223–242. Box B was found to be essential for RNA binding, and its truncation abolished both the RNA-binding property of Dcp2 and its ability to decap RNA . A biochemical characterization of human Dcp2 revealed that a 163-amino-acid region from 94 to 257 encompassing the Nudix fold and the adjacent Box B region was sufficient to retain decapping activity. Structural analysis of Dcp2 will help elucidate the specific contribution of each domain in substrate recognition and decapping.
Human Dcp2 and DcpS decapping enzymes
The function of the Dcp1 protein is still unclear. Although Dcp1 is no longer thought to contain intrinsic decapping activity, it is believed to function as a cofactor for Dcp2 at least in yeast . Surprisingly, the human Dcp1a and Dcp1b proteins do not appear to stimulate human Dcp2-mediated decapping, and their function in mammals remains unknown [11,19,21]. Structural analysis of yeast Dcp1 demonstrated that it contains an EVH1 [Ena–VASP (vasodilator-stimulated phosphoprotein) homology] protein domain . Two conserved regions on the molecular surface of Dcp1 were identified. One corresponds to a proline-rich sequence-binding site within the EVH1 domain, and was proposed to be a protein–protein interaction site. The second conserved region was essential for the stimulatory role of Dcp1 in the Dcp2-mediated decapping, but was not required for their interaction. A large hydrophobic patch adjacent to the two sites was also identified and was shown to significantly affect decapping, but not binding to Dcp2 . Uncoupling the capacity of Dcp1 to facilitate Dcp2 decapping from its ability to interact with Dcp2 suggests a more direct role of Dcp1 in the stimulation of Dcp2 decapping. The mechanism whereby Dcp1 facilitates Dcp2 activity still remains to be determined. In this regard, analysis of the three-dimensional structure of the Dcp1–Dcp2 complex will provide more insights into how the interaction and co-ordination between these two factors control decapping.
Biochemical and structural analysis of DcpS
An activity analogous to DcpS was initially reported over 30 years ago  and was subsequently purified  and cloned . DcpS is a member of the HIT (histidine triad) superfamily of nucleotide hydrolases and transferases , all of which possess a conserved HUHUHU HIT motif, where U denotes a hydrophobic amino acid . The HIT motif is contained within a larger ∼100-amino-acid HIT fold region (Figure 2). Substitution of asparagine for the central histidine within the HIT motif abolished the hydrolytic activity of the protein, demonstrating the significance of this domain in cap hydrolysis .
Elucidation of the DcpS structure provided informative insights into the decapping mechanism of this protein. Co-crystallization of DcpS with m7GpppG substrate revealed it is an asymmetric dimer consisting of distinct N-terminal and C-terminal domains linked by a flexible hinge region . The dimer forms a simultaneous closed productive conformation on one side of the protein and an open non-productive conformation on the other . The two N-termini form into a domain-swapped dimer and are required for cap dinucleotide substrate binding and hydrolysis [15,26]. Mutagenesis of amino acid residues involved in the stabilization of the closed conformation resulted in the enhancement of decapping, suggesting that the structure is dynamic and can alternate between an open and closed state . Consistent with this proposed dynamic nature, ligand-free DcpS was shown recently to be a symmetric dimer, with the N-terminus positioned such that both sides of the structure are in the open conformation . The N-terminal domain in ligand-free DcpS was indicated to be inherently flexible and in a dynamic state ready for substrate binding and product release [27,28]. The exclusive utilization of cap dinucleotides or short capped oligonucleotides as substrate for DcpS appears to be a function of several factors. First is the 16-fold greater affinity of DcpS for cap structure relative to capped RNA . However, a more significant contributor to the utilization of the cap dinucleotide appears to be due to greater steric and entropic constraints to form a closed decapping competent conformation with long capped RNA [26,28].
The biological significance of DcpS is only beginning to be understood. It appears that its intrinsic property to only function on capped oligonucleotides, but not capped mRNA, ensures that mRNAs are not decapped prematurely. Furthermore, its ability to displace eIF4E (eukaryotic initiation factor 4E) from cap dinucleotides suggests that DcpS might function to ensure eIF4E is not sequestered by the residual cap structure that could otherwise accumulate following 3′ end decay .
Localization of the decapping enzymes
The recent identification of discrete cytoplasmic foci referred to as mRNA-processing bodies (P-bodies) in both yeast and mammalian cells containing the Dcp1 and Dcp2 decapping proteins, as well as additional factors involved in the 5′ decay pathway suggests mRNA decay can be compartmentalized [11,29,30]. The recent demonstration that the eIF4E cap binding protein, as well as its transporter protein are also contained within the P-bodies , implies that these foci could also serve as storage sites for mRNA. Interestingly, the DcpS protein is primarily nuclear and absent from the P-bodies [15,32]. It is currently difficult to reconcile the proposed cytoplasmic function of DcpS with its predominant nuclear localization, although it is possible that DcpS can shuttle between the two cellular compartments.
Summary and prospects
Cells use two different decapping enzymes, each containing a distinct highly conserved hydrolase motif to carry out decapping. The Nudix motif in Dcp2 is involved in mRNA decapping and enables rapid 5′-exoribonucleolytic decay of the RNA body, while the HIT motif in DcpS is responsible for the scavenger decapping activity following degradation of the RNA body from the 3′ end (Figure 2). Thus two distinct decapping mechanisms function in mRNA decay.
Although the two decay pathways have been presented as distinct pathways above, it is becoming increasingly evident that there is extensive interplay between them. The scavenger decapping activity, which was initially identified in the 3′ decay pathway, can also hydrolyse the m7GDP product of the Dcp2 decapping enzyme in the 5′ decay pathway [27,33]. Although the hydrolysis activity on m7GDP is considerably less robust, the resulting product is m7GMP in both cases. More recently, the DcpS activity has also been implicated as a positive mediator of 5′→3′-exoribonucleolytic activity in the 5′ decay pathway . The correlation between the two exonucleolytic decay pathways and the role of the decapping enzymes in their potential interplay will be an exciting area for future studies.
Translation UK: Focused Meeting and Satellite to BioScience2005, held at Western Infirmary, Glasgow, U.K., 21–23 July 2005. Organized and Edited by M. Bushell (Nottingham, U.K.), S. Newbury (Newcastle upon Tyne, U.K.), G. Pavitt (Manchester, U.K.) and A. Willis (Nottingham, U.K.).
This work was supported by the National Institutes of Health grant GM67005 to M.K.