Eukaryotic mRNA decapping factors: molecular mechanisms and activity

Abstract

Decapping is the enzymatic removal of 5’ cap structures from mRNAs in eukaryotic cells. Cap structures normally enhance mRNA translation and stability, and their excision commits an mRNA to complete 5’ to 3’ exoribonucleolytic digestion and generally ends the physical and functional cellular presence of the mRNA. Decapping plays a pivotal role in eukaryotic cytoplasmic mRNA turnover and is a critical and highly regulated event in multiple 5’ to 3’ mRNA decay pathways, including general 5’ to 3’ decay, nonsense-mediated mRNA decay (NMD), AU-rich element (ARE)-mediated mRNA decay, microRNA (miRNA)-mediated gene silencing, and targeted transcript-specific mRNA decay. In the yeast Saccharomyces cerevisiae mRNA decapping is carried out by a single Dcp1-Dcp2 decapping enzyme in concert with the accessory activities of specific regulators commonly known as decapping activators or enhancers. These regulatory proteins include the general decapping activators Edc1, 2, and 3, Dhh1, Scd6, Pat1, and the Lsm1–7 complex, as well as the NMD-specific factors, Upf1, 2, and 3. Here, we focus on in vivo mRNA decapping regulation in yeast. We summarize recently uncovered molecular mechanisms that control selective targeting of the yeast decapping enzyme and discuss new roles for specific decapping activators in controlling decapping enzyme targeting, assembly of target-specific decapping complexes, and the monitoring of mRNA translation. Further, we discuss the kinetic contribution of mRNA decapping for overall decay of different substrate mRNAs and highlight experimental evidence pointing to the functional coordination and physical coupling between events in mRNA deadenylation, decapping, and 5’ to 3’ exoribonucleolytic decay.

Introduction

Eukaryotic mRNAs contain a ubiquitous 5’ cap modification, 7-methyguanosine (m⁷G), that is linked to the first transcribed nucleotide via a 5’–5’ triphosphate bond [1, 2]. This cap structure impacts almost every aspect of the mRNA life cycle and is a critically important determinant of gene expression [3]. In the nucleus, the cap structure, through association with the nuclear cap-binding complex (CBP20 and CBP80), promotes efficient pre-mRNA splicing [4, 5] and mRNA nuclear export [6]. In the cytoplasm, the cap structure, through association with the translation initiation factor eIF4E, enhances mRNA stability [7, 8] and promotes efficient mRNA translation [9, 10].

Removal of the cap structure from an mRNA by the decapping enzyme, dubbed decapping, is a fundamental cellular process conserved in eukaryotic cells. Decapping commits the mRNA to complete 5’ to 3’ exoribonucleolytic digestion and essentially ends the life and function of the mRNA. Decapping plays a pivotal role in eukaryotic cytoplasmic mRNA turnover [11, 12] and is a critical event in multiple 5’ to 3’ decay pathways including general 5’ to 3’ decay [13], nonsense-mediated mRNA decay (NMD) [14], AU-rich element (ARE)-mediated mRNA decay [15, 16], microRNA (miRNA)-mediated gene silencing [17, 18], and transcript-specific targeted mRNA decay [19–22].

In most eukaryotic cells mRNA decapping is largely carried out by the Dcp2 decapping enzyme [23–26]. The complexity of the process is enhanced in mammalian cells, where mRNA decapping not only depends on Dcp2, but at least two additional Nudix hydrolases, Nudt3 and Nudt16 [27–29]. In addition to the decapping enzyme, mRNA decapping also requires the functions of specific regulators, namely the decapping activators or enhancers [30, 31]. In the yeast Saccharomyces cerevisiae, decapping of wild-type mRNAs requires the general decapping activators Dhh1 [32, 33], Scd6 [34], Pat1 [35–38], and the Lsm1–7 complex [37–39], and decapping of nonsense-containing mRNAs requires the NMD-specific Upf1–3 factors [14]. Decapping enhancers Edc1–3 in yeast [19, 40–43] and Edc4 in metazoans [15, 26, 44] also have either general or specific roles in mRNA decapping.

mRNA decapping is intimately linked to translation. Since the cap structure must bind to initiation factor eIF4E to promote mRNA translation while also serving as a substrate for the decapping enzyme to commence mRNA decay it was hypothesized that decapping competes with translation initiation and that the activation of mRNA decapping must require translational repression at initiation [45]. Consistent with this hypothesis, inhibiting translation initiation by a cis-acting structural element or trans-acting gene mutations accelerates mRNA decapping [46, 47], and the decapping activators Dhh1, Pat1, and Scd6 function in repressing mRNA translation [34, 48–50]. On the contrary, decapping of NMD substrates requires translation [51], and translation of non-optimal codons promotes Dhh1-mediated mRNA decay [52, 53], suggesting that specific events in translation elongation or termination have an active role in mRNA decapping.

Structural and biochemical studies of mRNA decapping factors over the last two decades have provided significant insights into the molecular mechanisms involved in decapping regulation. These studies have revealed the domain organization and structures for many decapping factors including the Dcp1-Dcp2 decapping enzyme [54–56], Upf1–3 [57–59], Dhh1[60], Pat1 [61, 62], Edc3 [63, 64], Scd6 [65], and the Lsm1–7 complex [66–69]; elucidated the structural basis of numerous specific molecular interactions including those between the Edc1 and Dcp1 [40, 70], Edc3, Scd6, or Pat1 and Dhh1 [65, 71, 72] or Dcp2 [73, 74], and Pat1 and the Lsm1–7 complex [66, 67, 75]; identified the molecular determinants and mechanisms involved in cap recognition [76–78], RNA-binding [79] and cap hydrolysis [80]; and revealed the conformational dynamics of the decapping enzyme [81–83] and the functions or mechanisms of action of Edc1 and Edc3 in decapping enzyme activation [73, 76, 78, 82, 84]. The structural and biochemical aspects of mRNA decapping regulation, including Dcp2 catalytic mechanisms and conformational dynamics, decapping factor functions, and protein-protein interaction networks have been summarized in several recent reviews [27, 30, 82, 85–87] and will not be addressed here.

In addition to the canonical m⁷G modification, a novel nicotinamide adenine dinucleotide (NAD⁺) cap modification has been identified in mRNAs from bacteria, yeast, and mammalian cells [88–91]. In bacteria, the NAD⁺ cap promotes mRNA stability and is removed by the NudC protein of the Nudix family [88]. In mammalian cells, the NAD⁺ cap promotes mRNA degradation and is removed by the DXO family proteins with both decapping and 5’ to 3’ exoRNase activities [90]. Interestingly, the Dox family members Rai1 and Dox1 in yeast and DXO in mammalian cells are also involved in canonical cap quality control and can remove unmethylated caps as dinucleotides [92–94]. Non-canonical mRNA decapping and cap quality control have been reviewed recently [95] and will not be further discussed in this review.

Here, we focus on mRNA decapping regulation in vivo in a model organism, the yeast Saccharomyces cerevisiae. We will first summarize the recently uncovered molecular mechanisms that control the selective targeting of the yeast decapping enzyme and then discuss the roles of specific decapping activators in controlling decapping enzyme targeting, mRNP remodeling, and monitoring mRNA translation. Further, we will discuss the kinetic contribution of mRNA decapping for overall decay of different decapping substrates and highlight experimental evidence supporting functional coordination and physical coupling between the events in mRNA deadenylation, decapping, and 5’ to 3’ exoribonucleolytic decay.

Yeast decapping factors and their interaction networks

Dcp1-Dcp2 decapping enzyme.

mRNA decapping in yeast is carried out by a single decapping enzyme consisting of two subunits, the regulatory subunit Dcp1 and the catalytic subunit Dcp2. Both subunits were isolated based on genetic screens and are essential for mRNA decapping in vivo [36, 96, 97]. Dcp2 (also named Nmd1) was also isolated as a Upf1-interacting factor in a yeast two-hybrid screen using Upf1 as bait [98]. Dcp1 is a 231-amino acid EVH1 protein [55] and Dcp2 is a 970-amino acid protein containing a 245-amino acid conserved catalytic domain at its N-terminus and a 725-amino acid largely disordered domain at its C-terminus [97, 99]. The Dcp2 N-terminal catalytic domain consists of two subdomains, an all-helical N-terminal regulatory domain (NRD) and a C-terminal Nudix domain [54]. Dcp2’s NRD is involved in cap recognition [76, 78, 83], interacts with Dcp1 [56], and is essential for mRNA decapping in vivo [56, 77, 99]. Dcp2’s C-terminal domain encodes an array of regulatory elements including an auto-inhibitory element, one binding motif for Edc3, two binding motifs for Upf1, and multiple helical leucine-rich binding motifs for Pat1 and Scd6 [73, 99, 100] (Figure 1).

Figure 1. Interaction networks of yeast decapping factors.

Protein domain-domain or domain-motif interactions were identified with the yeast two-hybrid system in vivo or by biochemical and structural experiments in vitro. Each pair of arrows indicates a demonstrated or potential direct interaction. Edc3, Upf1, Pat1, and Scd6 bind directly to the C-terminal domain of Dcp2. Edc3 and Upf1 self-associate. Upf1’s CH and helicase domains also exhibit intramolecular interactions. Upf2, Upf3, Dhh1, and the Lsm1–7 complex do not interact with Dcp1 or Dcp2, but Upf2 interacts with Upf1, Dhh1 interacts with Edc3 and Pat1, and Lsm1, Lsm2, and Lsm3 interact with Pat1. In addition, Edc1 and Edc2 bind to Dcp1, Upf3 binds to Upf2, and Scd6 also binds to Pat1. Dcp2 regulatory elements are indicated by NRD, Dcp1-binding site; E3–1, Edc3-binding motif; U1₁ and U1₂, Upf1-binding motifs; E3–2 and L₁-L₉, Scd6 or Pat1-binding motifs; and IE, the inhibitory element. Upf1 and Dhh1 RNA helicase motifs are drawn as blue bars. In Pat1, N, M, and C indicate the N-terminal, middle, and C-terminal domains, respectively; in Edc3, RB indicates the Rps28-binding motif; in Edc1 and Edc2, P indicates the proline-rich motif; in Scd6, D, F, T, R indicate the DFDF, FFD, TFG, and RG motifs, respectively.

General decapping activators Dhh1, Pat1, and the Lsm1–7 complex.

Decapping of wild-type mRNAs in the canonical 5’ to 3’ decay pathway requires prior poly(A) shortening and usually occurs when the poly(A) tails reach a length of approximately 10 nucleotides [13, 101]. General mRNA decapping requires the functions of Dhh1 [32, 33], Pat1, and the Lsm1–7 complex [35–38]. Dhh1 is a conserved DEAD-box protein with two RecA-like domains [60]. The protein is highly abundant in yeast cells and exhibits strong RNA binding activity, but weak RNA helicase activity in vitro [102]. Dhh1 shows direct interactions with Edc3 and Pat1, but no direct interaction with either Dcp1 or Dcp2 [71, 99] (Figure 1). Dhh1 also associates with Scd6 in vivo [103], but a binary two-hybrid interaction with Scd6 could not be detected [100], suggesting that the observed Dhh1:Scd6 in vivo interaction may be bridged by additional factors. Pat1 is a conserved RNA binding protein containing three distinct domains, the N-terminal, Middle, and C domains [104, 105]. The Pat1 N-M domains appear to be disordered, but the C domain forms an α-α superhelix fold similar to the ARM repeats [62, 66, 67]. Pat1 shows direct interactions with Dhh1, Scd6, Dcp2, Xrn1, and the Lsm1–7 complex: Pat1 N binds to Dhh1 [71], Pat1 N and M domains both bind to Scd6 [49], the C domain binds to Dcp2 and Xrn1 [73, 99], and the Pat1 M and C domains are both engaged in binding to the Lsm1–7 complex [49, 66, 67, 104, 106] (Figure 1). The Lsm1–7 proteins each contain a conserved Sm fold and assemble into a circular heptameric ring that binds RNA [68, 69]. The Lsm1–7 complex displays direct interaction with Pat1 [66, 67], but no interaction with either Dcp1 or Dcp2 [99]. Pat1 and the Lsm1–7 complex each have low intrinsic RNA-binding activity, but the combined Pat1/Lsm1–7 complex has strong RNA-binding activity and a preference for oligoadenylated over polyadenylated or unadenylated RNAs [106, 107].

NMD-specific decapping factors Upf1, Upf2, and Upf3.

Decapping of nonsense-containing mRNAs proceeds through a deadenylation-independent mechanism [108]. The conserved NMD factors Upf1, Upf2, and Upf3 are required for decapping of nonsense-containing mRNAs, as loss of each these factors causes the accumulation of capped nonsense transcripts [14]. NMD factors perform their functions during premature translation termination [109, 110], and decapping of nonsense-containing mRNA occurs on polyribosomes [111]. Upf1 is a superfamily I RNA helicase containing an N-terminal CH domain and a C-terminal helicase domain [112, 113], and the protein has RNA-binding, ATPase, and helicase activities [57, 114]. The Upf1 CH domain binds to Upf2 [115] and the decapping enzyme subunit Dcp2 [99], and also self-associates [116] (Figure 1), suggesting that Upf1 may perform several sequential functions during execution of NMD. Upf2 (also named Nmd2) is an acidic protein containing three tandem MIF4G domains [98, 117, 118] and Upf3 is a basic protein containing an RNA recognition motif (RRM) [119]. Upf2 binds directly to both Upf1 and Upf3 [59, 115, 120, 121]. A 157-amino acid segment from Upf2’s C-terminal region interacts with the CH domain in Upf1 and its MIF4G-3 domain interacts with the central RRM in Upf3 [115, 122] (Figure 1). Human Upf2 has been shown to possess RNA binding activity [59]. In addition, binding of Upf2 to the Upf1 CH domain reduces Upf1’s affinity for RNA, and switches Upf1’s activities from RNA clamping to unwinding [57]. Full activation of Upf1’s ATPase and helicase activities requires both Upf2 and Upf3 [123], arguing that Upf2 and Upf3 function in regulating Upf1’s activities. The latter notion is also supported by yeast experiments showing that overexpression of Upf1 can compensate for the loss of Upf2 or Upf3, but not vice versa [124].

Decapping enhancers Edc1, Edc2, and Edc3.

EDC1 and EDC2 were, respectively, isolated as high-copy suppressors of conditional dcp1 or dcp2 alleles in screens for cell growth at the mutants’ non-permissive temperature [42]. Deletion of EDC1, EDC2, or both genes had no effect on decapping of reporter mRNAs in otherwise wild-type cells. However, these deletions slowed mRNA decapping in cells compromised for decapping activity [42]. In addition, overexpression of Edc1 or Edc2 partially suppressed mRNA decapping defects caused by specific DCP1 or DCP2 mutations, and Edc1 coprecipitated with Dcp1 and Dcp2 [42]. Based on these observations, Edc1 and Edc2 were proposed to be general enhancers of mRNA decapping [42]. Edc1 and Edc2 are small basic proteins that share two distinct regions of homology [42, 125] and, in vitro, both proteins bind to RNA and stimulate the decapping activity of the decapping enzyme [126]. A 25-amino acid segment from the C-terminus of Edc1 is sufficient for this stimulation activity, and this fragment contains a proline-rich Dcp1-binding motif (Figure 1) and a YAGxxF Dcp2-activating motif [40]. Similar segments from human and S. pombe Edc1 orthologs bind to both subunits of the decapping enzyme (Dcp1 and Dcp2), and the binding by these fragments stabilizes an active conformation of the decapping enzyme and promotes its binding to mRNA substrates [78, 82, 83]. Although the biochemical functions and in vitro mechanisms of action for Edc1 and Edc2 have been well established, the exact in vivo functions for each of these factors are currently unknown. Whether these two factors target most mRNAs or just a specific mRNA subset is still an open question. It should be noted that the functions of Edc1 and Edc2 are linked to translation during heat stress and to reprograming of gene expression during carbon source switching [125, 126]. Whether the translation and reprograming functions of Edc1 and Edc2 are related to or distinct from their functions in mRNA decapping is an open question.

Initial investigations of Edc3’s role in mRNA decapping [19, 20, 43] were all prompted by one important observation made from several published proteomic data sets, namely that a protein of unknown function generated from the Yel015W locus (later designated as Edc3, Dcp3, or Lsm16) exhibited two-hybrid interactions with multiple mRNA decay factors [127–129] and copurified with the decapping enzyme [130, 131]. These initial studies all identified Edc3 as an mRNA decapping factor, but yielded two contrasting conclusions regarding the protein’s function. In one study, a synergistic decapping defect in reporter mRNAs was detected when deletion of EDC3 was combined with dcp1 or dcp2 temperature-sensitive alleles in cells grown at the permissive temperature. This result led to the proposal that Edc3 was a general decapping activator [43]. In two other studies, deletion of EDC3 led to in vivo stabilization of only two transcripts in the entire transcriptome, RPS28B mRNA and YRA1 pre-mRNA, leading to the hypothesis that Edc3 functioned as a transcript-specific decapping activator [19, 20]. Our recent experiments have identified Edc3 as a common component of multiple decapping complexes but also as a targeting component of specific decapping complexes [100] (see below), suggesting that Edc3 plays multiple roles in mRNA decapping and functions indeed as both a general and a specific decapping activator.

Edc3 is an Lsm-like protein with a long C-terminal extension. The protein is conserved in eukaryotes and exhibits a modular structure, containing an Lsm domain at its N-terminus, an FDF domain in its middle, and a YjeF-N domain at its C-terminus [43, 132, 133]. The Lsm and YjeF-N domains from metazoan Edc3 orthologs form defined folds, but their FDF domains appear to be largely disordered [63, 64, 74, 134]. The Edc3 Lsm domain binds to Dcp2 [49, 99, 135], whereas distinct motifs of the FDF domain bind to Dhh1 and the ribosomal protein Rps28b [71, 136] (Figure 1). The Edc3 YjeF-N domain self-associates and, in combination with the Lsm4 C-terminal Q/N-rich prion-like domain, promotes P-body assembly in vivo [135].

Edc3-targeted degradation of RPS28B mRNA and YRA1 pre-mRNA both occur through deadenylation-independent decapping and each of these transcript-specific decay events is a functional component of negative feed-back autoregulatory systems for the respective genes [19, 20]. RPS28B encodes a 40S subunit ribosomal protein, and YRA1 encodes an hnRNP-like protein involved in an early stage of mRNA export. Edc3-mediated RPS28B mRNA decay occurs on polyribosomes, requires translation, and is controlled by a single regulatory element located in the 3’-UTR of the RPS28B mRNA [19, 137]. This decay pathway also requires the functions of the Edc3 Lsm1 and YjeF-N domains and the Rps28 biding motif located in the Edc3 FDF domain [19, 136, 137]. In contrast to the originally proposed mechanism of RPS28B mRNA decay [19], recent experimental evidence indicates that the Rps28 protein does not directly bind to the RPS28B 3’-UTR decay-inducing element; instead, an Edc3 dimer directly engages this element [137]. Edc3p-mediated YRA1 pre-mRNA decay is functionally linked to nuclear events in pre-mRNA splicing and export [20]. In this autoregulatory mechanism, excess Yra1p inhibits YRA1 pre-mRNA splicing and commits the pre-mRNA to nuclear export. Once in the cytoplasm, YRA1 pre-mRNA is targeted by Edc3 to initiate decapping, a step rapidly followed by 5’ to 3’ exoribonucleolytic decay. In contrast to RPS28B mRNA decay, Edc3-mediated YRA1 pre-mRNA decay requires translational repression and is controlled through multiple functionally distinct and redundant regulatory elements in the YRA1 intron [41]. Two elements (EREs; Edc3 response elements) control Edc3 substrate specificity and the other three (TREs; translational repression elements) function in repressing YRA1 pre-mRNA translation. Translational repression of YRA1 pre-mRNA also requires the conserved, general mRNA export receptor Mex67/Mtr2 to ensure that the pre-mRNA is degraded exclusively by the Edc3-mediated pathway, but not by the NMD pathway [20, 41]. Surprisingly, Edc3-mediated YRA1 pre-mRNA decay requires neither the FDF domain nor the YjeF-N domain of Edc3, and the Edc3 Lsm domain is sufficient to promote efficient YRA1 pre-mRNA degradation [137]. These results suggest that the Edc3 Lsm domain is directly involved in recognition of the YRA1 pre-mRNP as a decapping substrate.

Translational repressor and decapping activator Scd6.

SCD6 was originally isolated as a high copy suppressor that rescued a conditional lethality caused by clathrin deficiency [138]. However, a direct role of Scd6 in clathrin-mediated vesicular transport has never been demonstrated. Bioinformatic analyses identified Scd6 as a conserved Lsm-like protein with a long C-terminal extension and predicted that this protein functions in RNA metabolism [132, 133]. Consistent with this prediction, Sdc6 orthologs from Saccharomyces cerevisiae, Drosophila melanogaster (Tral), Caenorhabditis elegans (CAR-1), Xenopus laevis (xRAP55), and human (RAP55) cells have since been shown to accumulate in several different types of cytoplasmic mRNP granules that involve translational repression, including P-bodies and stress granules [50, 65, 139, 140], germline polar granules and storage bodies [141–143], and neuronal transport granules [144]. Scd6 orthologs Tral, CAR-1, and xRAP55 are each key components of maternal storage mRNP complexes and co-associate with translation repressors, including Dhh1, Cup, and 4E-T in these mRNP complexes during oocyte maturation and early embryo development [139, 143, 145, 146]. In addition, Tral from Drosophila and Scd6 from S. pombe respectively bind to the Dcp1 and Dcp2 subunits of the decapping enzyme [65, 74].

Like Edc3, Scd6 is an Lsm-like protein with a modular structure, containing an Lsm domain at its N-terminus, DFDF, FFD, and TFG motifs in its middle, and RG repeats at its C-terminus [132, 133, 146]. Scd6 binds to multiple helical leucine-rich motifs in the C-terminal domain of Dcp2 (He et al., 2022), and the N and M domains of Pat1 [49] (Figure 1). The Lsm domain of human Scd6 ortholog RAP55 also binds to two distinct regions of translation repressor 4E-T, its DFDF and TFG motifs bind to the C-terminal RecA-like domain of Dhh1 ortholog DDX6, and its FFD motif engages the decapping activator EDC4 [72]. Scd6 also binds to the C-terminal region of translation initiation factor eIF4G [50]. Overexpression of Scd6 inhibits cell growth and recombinant Scd6 can block 48S PIC formation in vitro [49]. Recently, it was shown that tethering Scd6 to reporter mRNAs in yeast can trigger both Dhh1-mediated translation repression and Dcp2-mediated accelerated mRNA decapping, and that both effector functions of tethered Scd6 are dependent on its Lsm domain [34]. Further, RNA-Seq and ribosome profiling experiments indicated that Scd6 targets a small subset of endogenous mRNAs for accelerated degradation and yet another small subset of mRNAs for translation repression, and that both of these activities require the functions of Dhh1 and Dcp2 [34]. These results point to a dual role of Scd6 in yeast mRNA translation and decay, suggesting that Scd6 functions as both a translation repressor and a decapping activator, most likely for specific mRNAs. Significantly, genetic evidence indicates that Scd6 has redundant functions with Edc3. First, Scd6 binds to the sole Edc3-binding motif in the C-terminal domain of Dcp2 (He et al., 2022). Second, double deletions of SCD6 and EDC3 yield synergistic defects in cell growth and reporter mRNA decapping [147] and cause a fraction of Dcp2 to localize to the nucleus [148].

Molecular functions and regulatory mechanisms of yeast decapping factors

Negative and positive regulation of the yeast decapping enzyme.

Dcp2, the catalytic subunit of the yeast decapping enzyme, contains a 245-amino acid N-terminal catalytic domain and a 745-amino acid largely disordered C-terminal domain. Recent genetic experiments revealed that the C-terminal domain of Dcp2 plays pivotal roles in mRNA decapping and encodes critical regulatory activities that control the temporal activation and targeting specificity of the decapping enzyme [12, 99, 100]. This C-terminal domain encodes multiple regulatory elements, including a 25-amino acid proline-rich inhibitory element that auto-inhibits Dcp2 decapping activity and a set of short linear motifs that promote the independent binding of specific decapping activators. The latter include one specific binding motif for Edc3, two binding motifs for Upf1, and multiple binding motifs for Pat1 or Scd6 [73, 99, 100] (Figure 2).

Figure 2. Functional classification of yeast decapping factors.

Based principally on their interaction patterns with Dcp2, yeast decapping factors were classified into three different categories: core components, targeting components, and mRNP components. Since Edc3 self-associates and appears to have both general and specific functions in mRNA decapping, Edc3 molecules were partitioned into two different pools. In this classfication, Dcp1, Dcp2, and Edc3 function as core components of multiple decapping complexes; Edc3, Upf1, Scd6, and Pat1 function as regulatory components of decapping complexes and target the decapping enzyme to specifc mRNPs; Upf2, Upf3, Dhh1, and the Lsm1–7 complex function as specific mRNP components. Targeting component Edc3 binds the core components through Edc3 dimerization and also interacts with Dhh1 or RNA elements in specific mRNPs. Targeting components Upf1, Scd6, and Pat1 bind to the core components through distinct cis-binding elements in Dcp2 and each interacts with at least one specific mRNP component. mRNP components do not interact with Dcp1 or Dcp2, but each interacts with at least one targeting component. Edc3, Scd6, and Pat1 may have redudant functions in targeting the Dhh1-regulated mRNAs, as these three factors all interact with Dhh1. Dcp2 regulatory elements are indicated by NRD, Dcp1-binding site; E3–1, Edc3-binding motif; U1₁ and U1₂, Upf1-binding motifs; E3–2 and L₁-L₉, Scd6 or Pat1-binding motifs; and IE, the inhibitory element.

The autoinhibition imposed by the Dcp2 cis-inhibitory element is critically important for controlling the decapping enzyme’s function in vivo. Elimination of the inhibitory element leads to a constitutively activated decapping enzyme with indiscriminate decapping activity that targets both typical decapping substrates and non-decapping substrates, thus resulting in the loss of the decapping enzyme’s specificity for substrate mRNAs [12, 99]. Interestingly, elimination of the inhibitory element also partially bypasses the functional requirement of the decapping activator Edc3 in decay of its substrate YRA1 pre-mRNA [99]. Thus, Dcp2 autoinhibition provides at least two important functions for the decapping enzyme. One is to ensure its substrate specificity and to avoid accidental or indiscriminate mRNA decapping and the other is to control its timely activation by the decapping activators. The 3D structure of full-length Dcp2 was recently generated by the AlphaFold program [149]. Significantly, a 7-residue segment of the inhibitory element was predicted to bind to the catalytic domain of Dcp2, sandwiched between its NRD and Nudix domains and forming molecular interactions with two Dcp2 catalytic residues (Figure 3). This structural arrangement suggests that the inhibitory element may function by blocking the binding of Dcp2 to its mRNA substrates. The Dcp2 cis-regulatory elements control the targeting specificity and final activation of the decapping enzyme. Binding to specific elements by different decapping activators alleviates the autoinhibition imposed by the inhibitory element and promotes the assembly of target-specific decapping complexes [99, 100]. The yeast Dcp1-Dcp2 holoenzyme is thus subjected to both negative and positive regulation, and both of these regulatory mechanisms are important for the decapping enzyme’s function in vivo as elimination of the entire Dcp2 C-terminal domain causes massive deregulation, altering the levels of mRNA expression for one quarter of the transcriptome [12].

Figure 3. Dcp2 inhibitory element is predicted to bind to the catalytic center of the decapping enzyme.

The three-dimensional structural model of full-length Dcp2 was generated by the AlphaFold program [149]. Dcp2’s N-terminal 260 amino acids form folded domains, but its entire C-terminal domain including the inhibitory element is unstructured or disordered. (A) A global view of a Dcp2 structural model showing the juxtaposition of the disordered inhibitory element (IE) and the Dcp2 nudix domain. (B) A zoomed-in view of the Dcp2 structural model showing amino acid interactions between residues 363 to 369 of the inhibitory element and specific residues from the Dcp2 Nudix domain and the connecting helix between the Dcp2 NRD and Nudix domains. In A and B, residues are color-coded according their pLDDT scores and the inhibitory element is shown with stick structures. In B, the phenylalanine and proline residues of the inhibitory element are predicted to be directly engaged in interactions with the catalytic lysine135 and glutamic acid198 residues of the Dcp2 Nudix domain. The predicted hydrogen bond and cation-π interactions are shown by dotted blue or orange lines, respectively. It should be noted that residues of the inhibitory element all have pLDDT scores lower than 50 and that the predicted residue interactions between Dcp2’s inhibitory element and Nudix domain in the structural model need additional experimental validation.

Decapping activators are selective and target specific subsets of mRNAs for decapping.

Current mechanistic models of in vivo mechanisms of mRNA decapping and functions of different decapping activators are essentially all derived from the analyses of a few reporter mRNAs [13, 32, 33, 35–38, 42, 43, 108, 150, 151]. While reporter-based analyses have provided significant insights into the mechanisms of decapping and the functional requirements of different decapping activators for both wild-type and nonsense-containing mRNAs, the models derived from these reporter mRNA analyses are probably too simplistic and thus may not entirely reflect the complexities of decapping regulation of endogenous mRNAs, e.g., their wide range of decay rates [152–155]. Importantly, transcripts targeted by the same decapping activator can still be degraded by drastically different mechanisms [19, 20, 137]. Early microarray analyses and recent RNA-Seq experiments recognized the target specificity of different decapping activators [11, 12, 19, 20, 152, 156, 157]. Further analyses of the transcripts targeted by individual factors have yielded significant insights into the functions and relationships of different decapping factors and revealed, at least in yeast, that decapping factors are highly selective and each targets a specific subset of mRNAs, even for the general decapping activators originally thought to act on all wild-type mRNAs [11, 12, 31].

The NMD-specific decapping activators Upf1, Upf2, and Upf3 commonly target 907 transcripts in the yeast transcriptome [11]. A small fraction of NMD-targeted transcripts (~12%) can be classified into the known structural classes of typical NMD substrates, i.e., mRNAs encoded by genes containing nonsense mutations in their coding regions, mRNAs using frameshifting in their translation, transcripts generated from pseudogenes, mRNAs containing upstream open reading frames (uORFs), and unspliced pre-mRNAs that enter the cytoplasm as a consequence of inefficient or regulated splicing [152]. The vast majority of NMD-targeted transcripts (88%) are normal-looking protein-coding mRNAs. Bioinformatic analyses reveal that this group of transcripts generally have low average codon optimality scores and high transition probability to nonoptimal codons. In addition, they also have intrinsically low translational efficiencies and relatively high ratios of out-of-frame translation [11]. They may also have unannotated uORFs that arise from alternatively transcribed isoforms [158]. These NMD-targeted transcripts are sensitive to deletions of DCP1, DCP2, or XRN1 and show increased mRNA levels in cells harboring these gene deletions [11], indicating they are bona fide decapping substrates.

The general decapping activators Dhh1, Pat1, and Lsm1 collectively regulate 1587 transcripts in the yeast transcriptome, with each of these factors respectively targeting 1098, 940, and 955 transcripts [12]. Pat1 and Lsm1-targeted transcripts share 84% overlap and exhibit almost identical expression patterns in cells harboring deletions of the PAT1 or LSM1 genes, indicating that Pat1 and Lsm1 function together and commonly control the same set of transcripts. Dhh1-targeted transcripts share about 50% overlap with those targeted by Pat1 and Lsm1, indicating that although these three factors share partial overlap in their regulatory targets, Dhh1 has functions that are distinct from those of Pat1 and Lsm1 in mRNA decapping. Further analyses identified 556 transcripts regulated only by Dhh1, 382 transcripts regulated only by Pat1 and Lsm1, and 482 transcripts regulated by all three factors. Collectively, these observations indicate that the general decapping activators Dhh1, Pat1, Lsm1 do not act on most wild-type mRNAs in yeast cells, and instead selectively target specific subsets of yeast mRNAs. The observation that Dhh1, Pat1, and Lsm1 target a common subset of transcripts is intriguing. One possibility is that these factors target the same substrate mRNA but function at different steps that can independently trigger mRNA decapping. Another possibility is that Dhh1 or Pat1 and Lsm1 target two different mRNAs of the same substrate and promote mRNA decapping in parallel pathways. Dhh1, Pat1, and Lsm1-targeted transcripts generally all have relatively low average codon optimality scores, high ribosome occupancies, and low protein levels, indicating that these mRNAs are translated inefficiently [12]. Significantly, transcripts targeted by these three factors all show increased accumulation in cells lacking either the Dcp2 decapping or Xrn1 5’ to 3’ exoribonuclease activities [12], indicating that Dhh1, Pat1, and Lsm1-targeted transcripts are bona fide decapping substrates.

Two independent microarray analyses identified only two transcripts (RPS28B mRNA and YRA1 pre-mRNA) in the entire yeast transcriptome as Edc3 targets, suggesting that Edc3 functions as a transcript-specific decapping activator [19, 20]. Recent RNA-Seq experiments identified 83 transcripts targeted by Scd6 [34], suggesting that Scd6 may also have limited function in mRNA decapping. However, recent genetic analyses of Dcp2 cis-binding elements indicate that Edc3 also targets the decapping enzyme to Dhh1 substrates [100]. In addition, Scd6 appears to have redundant functions with Edc3 in mRNA decapping [100, 147]. These observations indicate that both Edc3 and Scd6 may have much broader roles in mRNA decapping than previously thought.

Dcp2 C-terminal cis-binding elements control the substrate specificity of the decapping enzyme by orchestrating the formation of target-specific decapping complexes.

Two-hybrid, in vitro binding, deletion, and sequence analyses thus far have identified 13 cis-binding elements in the C-terminal domain of Dcp2 [73, 99, 100, 159], including two binding motifs in the original Edc3-binding region (E3–1 and E3–2), two Upf1-binding motifs (U1₁ and U1₂), and nine helical leucine-rich motifs (HLMs, L_1–9) (Figure 2). The nine HLMs were originally proposed as Pat1-binding motifs based on the observation that eight out of these nine HLMs in isolation bind Pat1 in either yeast two-hybrid or in vitro pull-down assays [73, 99]. Recent Dcp2 element deletion analyses, however, indicate that in the context of full-length Dcp2, HLMs L₁ to L₅ are engaged in Pat1 binding, but HLMs L₆ to L₉ show no binding to Pat1 [100]. The exact contribution of each of the L₁ to L₅ motifs to Pat1 binding is currently not clear. Interestingly, Edc3 binds only to the E3–1 motif, and Scd6 binds to both the E3–1 and E3–2 motifs, as well as to the L₃ and L₅, and L₈ and L₉ motifs [100]. The pattern of Scd6 binding to Dcp2 suggests that Scd6 engages the binding motifs in these three pairs cooperatively, and that it can bind to Dcp2 either as a monomer or as a dimer with each monomer engaging three different binding motifs, one from each of the three pairs [100].

To identify the roles of different Dcp2 cis-binding elements in controlling decapping enzyme functions, He et al. recently carried out systematic deletion of these elements and analyzed their consequences for Dcp2 binding to specific decapping activators and the decay of different classes of decapping substrate mRNAs [100]. These genetic analyses defined the roles of the Dcp2 cis-binding elements in controlling decapping enzyme functions and elucidated the molecular mechanisms that control the decapping enzyme’s selective targeting to different classes of substrate mRNAs. Dcp2 cis-binding elements were found to promote largely independent binding of Edc3, Scd6, Upf1, and Pat1 to Dcp2 and to control the selective targeting of the decapping enzyme to distinct substrate mRNAs. These conclusions are supported by strong experimental genetic evidence. It was demonstrated that loss of both Upf1-binding motifs eliminated Upf1 binding to Dcp2 and caused selective partial stabilization of NMD substrates. and that loss of the Edc3-binding motif eliminated Edc3-binding to Dcp2 and caused selective partial or complete stabilization of Edc3 substrates, as well as selective partial stabilization of Dhh1 substrates. In addition, specific deletions of the HLM motifs (L_1–8, L_1–9, L_9–2, and L_9–3) eliminated Pat1 binding to Dcp2 and, when these deletions were combined with deletions of the Edc3-binding motif, caused selective additional stabilization of Dhh1 substrates. Surprisingly, despite the well established in vivo function and multiple identified in vitro activities for Pat1 in mRNA decapping [35, 37, 38, 49, 73, 75, 104], it was shown that loss of the Pat1-binding motifs L_1–5, or even all nine HLMs (L_1–9), had no effect on the accumulation of Pat1/Lsm1 substrates. One possible explanation for the latter seemly inconsistent observations is that Pat1 binding to Dcp2 HLMs L_1–5 could still target the decapping enzyme to Pat1/Lsm1 substrates and may even enhance the decapping rates for the substrate mRNAs, but the Pat1-mediated decapping step may not be rate-limiting for overall decay of those substrates.

Dcp1, the essential regulatory subunit of the decapping enzyme, forms Dcp2-bridged two-hybrid interactions with several decapping activators including Edc3, Dhh1, Upf1, Pat1, and Scd6 [99, 100]. To dissect the molecular basis for these Dcp2 bridged interactions, He et al. constructed two sets of yeast two-hybrid tester strains and analyzed each of the observed interactions in these tester strains [100] (see genetic principles summarized in Figure 4). One set of strains contained specific deletions of the EDC3, UPF1, PAT1, and SCD6 genes or a truncation of the entire DCP2 C-terminal domain. The second set of strains contained different DCP2 C-terminal truncations in both EDC3 and edc3Δ backgrounds. The first set of tester strains was designed to assess the factor requirements and the second set of tester strains was designed to evaluate the Dcp2 cis-binding element dependence for each of the observed interactions. These two-hybrid analyses revealed that Dcp2 cis-binding elements promote assembly of distinct decapping complexes. The E3–1 motif by itself could promote the formation of two dimeric Edc3-containing decapping complexes, a Dcp1-Dcp2-Edc3-Edc3 and a Dcp1-Dcp2-Edc3-Edc3-Dhh1 complex. The E3–1 motif in combination with the U1₁ or U1₂ motifs promoted the formation of two different Upf1-containing Dcp1-Dcp2-Edc3-Upf1 decapping complexes of the same composition but distinct configurations with Upf1 binding to either U1₁ or U1₂ motifs. Similarly, the E3–1 motif in combination with L₁ or L₂ could promote the formation of two different Pat1-containing Dcp1-Dcp2-Edc3-Pat1 decapping complexes of the same composition but distinct configurations with Pat1 binding to either L₁ or L₂ motifs. Further, the E3–1 motif in combination with the Scd6 binding motifs was able to promote the formation of an Scd6-containing Dcp1-Dcp2-Edc3-Scd6 decapping complex. In addition, in the absence of Edc3, the E3–1 and E3–1 motifs in combination with the Pat1-binding motifs could promote the formation of a dimeric Scd6-containing Dcp1-Dcp2-Scd6-Scd6-Pat1 complex. Collectively, these new results indicate that Dcp2 cis-binding elements control the selective targeting of the decapping enzyme by forming target-specific decapping complexes. Edc3 functions as a core component of multiple decapping complexes but also as a targeting component of specific decapping complexes. Edc3, Upf1, Pat1, and Scd6 function as regulatory subunits of the holo-decapping enzyme, controlling both its substrate specificity and enzymatic activation.

Figure 4. Genetic inferences of direct and indirect yeast two-hybrid interactions.

Two-hybrid interaction between proteins A and B was assayed in a wild-type tester strain and in a set of tester strains containing specific gene or element deletions. These dual testing systems allow differentiation of direct interactions from indirect interactions. A direct interaction is insensitive to specific gene or element deletions. In contrast, an indirect interaction is sensitive to specific gene or element deletions. The indirect bridged interactions can be classified into three general categories. In the first category, A-B interaction is bridged simply by C. In this case, A and B interact in wild-type cells, but loss of C eliminates A-B interaction. In the second category, A-B interaction is bridged by C then D or is bridged by C and indirectly dependent on D. In both cases, A and B interact in wild-type cells, but loss of C, D, or both C and D eliminates A-B interaction. In the third category, A-B interaction is bridged by C but B is competed by D. In this case, A and B show no interaction in wild-type cells, but loss of D promotes A-B interaction and loss of C or both C and D eliminates A-B interaction. Blue indicates a positive two-hybrid interaction and white indicates no interaction.

Established in vivo and in vitro functions of the decapping activators.

Decapping competes with translation initiation and thus requires a transition from a translating mRNP to a decapping mRNP. Based on these considerations, it was hypothesized that decapping activators must perform two important functions in vivo, namely repression of mRNA translation initiation and activation of the decapping enzyme [31, 45]. Consistent with this hypothesis, it has been shown that the decapping activators Dhh1, Pat1, and Scd6 can repress mRNA translation both in vivo and in vitro, and Edc3 and Pat1 can stimulate the decapping reaction in vitro. Evidence for translational repression includes observations that: i) Dhh1 and Pat1 are required for repressing general translation during glucose deprivation or amino acid starvation [48, 160], ii) overexpression of Dhh1 or Pat1 causes decreased polyribosome accumulation and reduced overall translation rates [48, 104], iii) recombinant Dhh1, Pat1, and Scd6 proteins can block 48S PIC formation in vitro [48, 49], and iv) tethering Dhh1 or Scd6 to either endogenous or reporter mRNAs represses translation of the respective mRNAs independent of their decay-promoting activities [34, 161, 162]. Evidence that these factors also have roles in the activation of mRNA decapping includes the observations that recombinant Edc1, Edc2, Edc3, Edc3 Lsm domain, and Pat1C domain can all stimulate the decapping reaction carried out by a decapping enzyme containing a 670 amino-acid C-terminal truncation of Dcp2 [23, 49]. The activation of the decapping enzyme by Edc1, Edc2 and Edc3 likely results from direct interactions between each of these factors and Dcp1, Dcp2, or both [40, 163], but the observed activation by Pat1C domain is intriguing, as truncated Dcp2 apparently lacks all of the Pat1-binding motifs [100].

The targets of the translational repression activities of Scd6, Dhh1, and Pat1 are still not well defined. The RGG domain of Scd6 binds to a C-terminal fragment of eIF4G, and, based on this observation, Scd6 was proposed to repress mRNA translation by inhibiting eIF4G activity [50]. However, in a tethered function assay, the Scd6 RGG domain is dispensable and instead its Lsm domain is required for repressing translation of a reporter mRNA, suggesting that Scd6 may target a translation factor other than eIF4G [34]. Dhh1 blocks 48S PIC assembly in vitro and its human ortholog RCK/p54 can inhibit translation mediated by a cricket paralysis virus internal ribosome entry site (IRES), which does not require any translation initiation factors. From these observations, Dhh1 was originally thought to repress mRNA translation by targeting the 40S ribosomal subunit [48]. However, Dhh1 associates preferentially with mRNAs having high levels of non-optimal codons, tethering Dhh1 causes increased accumulation of tethered reporter mRNAs in heavy polyribosome fractions, and Dhh1binds to ribosomes, indicating that Dhh1 most likely represses mRNA translation by targeting elongating ribosomes [53, 162]. The target of Pat1 in repressing mRNA translation is thus far unknown. As a note, the general functions of translation repression originally proposed for Dhh1 and Pat1 during glucose deprivation have been challenged, as loss of polyribosome accumulation can also be explained by massive mRNA degradation without evoking Dhh1- or Pat1-mediatied repression at translation initiation [164].

Structural and kinetic analyses also revealed new important functions or mechanisms of action for several decapping activators, with most of the mechanistic studies utilizing protein orthologs from S. pombe or K. lactis. Edc1 has thus been shown to bind to Dcp1 and both the NRD and Nudix domains of Dcp2, stabilize a substrate-induced active conformation, and promote both substrate binding and catalysis by the decapping enzyme [40, 78, 83]. Depending on the model organism used for analyses, Edc3 binds to either single or multiple Dcp2 HLMs [74, 99, 100]. Edc3 binding to Dcp2 alleviates the autoinhibition imposed by the protein’s inhibitory elements and enhances substrate binding by the decapping enzyme [76, 84]. Similarly, Pat1 also binds to specific HLMs in Dcp2 [73, 99, 100]. Pat1 binding to Dcp2 also alleviates decapping enzyme autoinhibition and enhances the enzyme’s substrate binding [75]. In addition, Dcp1 binds to the NRD domain of Dcp2 [56]. In the absence of substrate, Dcp1 binding to Dcp2 shifts the major population of Dcp2 to a closed conformation that occludes its RNA-binding channel. However, in the presence of substrate, Dcp1 binding to Dcp2 stabilizes the conformation around the split active site of the Dcp2 NRD domain and thus promotes both substrate binding and catalysis by the decapping enzyme [79, 83]. In considering these mechanistic results it should be noted that biochemical and structural studies of decapping activators over last two decades have mostly used short peptides, isolated domains, or truncated proteins. In addition, almost all the studies used C-terminally truncated Dcp2 that lacks either a part or the entirety of its regulatory motifs [40, 56, 76–79, 81–83]. As the loss of the Dcp2 inhibitory element yielded a constitutively activated decapping enzyme that is independent of the functions of decapping activators [12, 99], some proposed functions or mechanisms of action that have been based on biochemical and structural investigation may need additional validation.

New in vivo functions and functional order of the decapping activators.

In another approach to identifying the functions and functional relationships of yeast decapping factors, He et al. recently carried out systematic two-hybrid analyses to map the pertinent protein-protein interaction networks [99, 100]. These analyses confirmed Dcp1:Dcp2 interaction, and identified Edc3, Upf1, Pat1, and Scd6 as direct Dcp2 interactors. In contrast, the general decapping activators Dhh1 and Lsm1, or the NMD-specific decapping activators Upf2 and Upf3, were shown to lack direct Dcp1 or Dcp2 interactions. Interestingly, each of the direct Dcp2 interactors also appears to interact with at least one of the specific decapping activators that lacks direct interaction with Dcp1 or Dcp2. For example, Edc3 interacted with Dhh1 [71, 99], Upf1 interacted with Upf2 [98, 115], Pat1 interacted with Dhh1 and Lsm1 [66, 104], and Scd6 associated with Dhh1[103, 147]. Based on these observations, Edc3, Upf1, Pat1, and Scd6 are thought to be the targeting components of the decapping enzyme that recruit the enzyme to specific substrate mRNAs (Figure 2). In contrast, Dhh1, Lsm1, Upf2, and Upf3 are thought to function upstream of decapping enzyme recruitment and be specific components of targeted mRNPs and regulators of the direct Dcp2 interactors [99, 100] (Figure 2).

Recent extensive genetic analyses have revealed that, in addition to recruiting the decapping enzyme to specific substrate mRNAs, the targeting components Edc3, Upf1, Pat1, and Scd6 each appear to perform at least one additional and perhaps major function upstream of their targeted mRNPs during the decapping process. First, trans deletions of EDC3, UPF1, or PAT1 almost aways yielded stronger stabilization of their targeted mRNAs than deletion of the respective factor’s cis-binding element(s) in Dcp2 [100]. Second, in Edc3-targeted YRA1 pre-mRNA decay, loss of Edc3 appeared to be epistatic to loss of the Edc3-binding element in Dcp2, and similarly, in NMD of the CYH2 pre-mRNA or ade2–1 mRNA, loss of Upf1 appeared to be epistatic to loss of both Dcp2 Upf1-binding elements [99, 100]. Third, Edc3 was recently shown to target the decapping enzyme to Dhh1 substrates for mRNA decapping, as loss of the Edc3-binding element caused partial stabilization of Dhh1-targeted mRNAs. Importantly, trans elimination of EDC3 caused additional stabilization of Dhh1 substrate mRNAs in cells harboring Dcp2 Edc3-binding element deletions, arguing strongly that Edc3 has a function that is independent of its Dcp2 interaction in decapping of Dhh1 substrate mRNAs [100]. The exact nature of the upstream functions for Edc3, Upf1, Pat1, and Scd6 is largely unknown, but needs to be defined to achieve a mechanistic understanding of mRNA decapping regulation. That said, the targeting components Edc3, Upf1, Pat1, and Scd6, as well as the specific mRNP components Dhh1, Lsm1, Upf2, and Upf3, probably all function in some aspects of mRNP remodeling in which they promote mRNP transitions from a state still engaged in translation to a state permissible for mRNA decapping. The remodeling activities carried out by these factors may include the recruitment of additional translational repressors, dissociation of a final Pab1 molecule from shortened poly(A) tails, eviction of eIF4E from the cap structure, or release and recycling of prematurely terminating ribosomes on nonsense-containing mRNAs.

The functional order for general or NMD-specific decapping activators described here contrasts sharply with those proposed based on P-body accumulation in different yeast decapping mutants. In earlier studies, cells harboring a deletion of PAT1 exhibited weak P-body accumulation (small in number and weak in intensity), cells harboring a deletion of LSM1 exhibited strong P-body accumulation (large in number and strong in intensity), and cells harboring double deletions of PAT1 and LSM1 exhibited weak P-body accumulation [165]. Similarly, cells harboring a deletion of UPF1 exhibited weak P-body accumulation, cells harboring single deletions of UPF2 or UPF3 exhibited strong P-body accumulation and cells harboring double deletions of UPF1 and UPF2 or UPF1 and UPF3 exhibited weak P-body accumulation [166]. Based on these relationships and additional observations, Lsm1, and Upf2 or Upf3 were proposed to function at P-bodies and act downstream of the respective Pat1 or Upf1-mediated transcript targeting events [165, 166]. Because Pat1 and Upf1 both bind to the decapping enzyme and also bind respectively to the mRNP components Lsm1 and Upf2, loss of Lsm1 or Upf2 and Upf3 could affect the cycling of Pat1 or Upf1 between mRNPs and the decapping enzyme and thus may shift these two factors predominantly to the decapping enzyme bound population. In this case, the Pat1- or Upf1-associating decapping enzyme complexes are unable to target their intended specific mRNPs because of loss of either Lsm1 or Upf2 and thus may even not contain any targeted mRNAs as originally postulated.

Active recruitment of the decapping enzyme mediated by different targeting components makes distinct contributions to the overall decay of targeted mRNAs.

Dcp2 cis-binding element deletions abolish the binding of specific targeting components of the decapping enzyme and thus eliminate the active recruitment functions of the targeting factors. Recent genetic analyses of Dcp2 cis element mutants revealed the relative mRNA decay contributions of targeting factor-mediated decapping enzyme recruitment events and provided significant insights into the decay mechanisms for different decapping substrate mRNAs [100]. For the Edc3 substrate RPS28B mRNA, it was shown that loss of the Edc3-binding motif (E3–1) in Dcp2 caused complete mRNA stabilization relative to deletion of EDC3 itself. Interestingly, for the Edc3 substrate YRA1 pre-mRNA, and all the tested Dhh1 substrates, it was shown that loss of the Edc3-binding site caused only partial stabilization relative to the respective EDC3 or DHH1 deletions. Significantly, additional loss of the Pat1-binding motifs (HLMs L_1–9, L_1–8, L_9–2, or L_9–3) caused further stabilization of the latter Edc3 and Dhh1 substrates. These results indicate that both Edc3 and Dhh1 substrates can be targeted by Pat1-mediated recruitment of the decapping enzyme. However, since loss of the Pat1-binding motifs alone had no effect on these Edc3 and Dhh1 substrates, it seems that Pat1-mediated recruitment of the decapping enzyme merely functions as a backup system in the decay of Edc3 and Dhh1-regulated mRNAs. Because the combined deletions of the Edc3- and Pat1-binding motifs caused substantial stabilization of Edc3 and Dhh1 substrates, these observations argue that decapping is rate-limiting for both Edc3 and Dhh1 substrates, suggesting that the major functions of Edc3 or Pat1 in decay of these mRNAs is recruiting the decapping enzyme.

For NMD substrates, it has been shown that loss of both Upf1-binding motifs caused only marginal stabilization (2–3-fold), but deletion of the UPF1 gene caused substantial stabilization (>10-fold). These results indicate that Upf1-mediated recruitment of the decapping enzyme is likely to only make a minor contribution to the decay rate of NMD substrates, suggesting that: mRNA decapping is not a major rate-limiting step in NMD, recruitment of the decapping enzyme to NMD substrates could be largely dispensable, and Upf1 must perform one major function upstream of decapping enzyme recruitment. In support of these ideas, NMD substrates are still degraded by a decapping-dependent pathway in the absence of active recruitment of the decapping enzyme by Upf1. In addition, further deletions of the Edc3- and Pat1-binding motifs do not have significant effects on NMD substrates and eliminating the entire C-terminal domain only causes marginal stabilization of NMD substrates (He and Jacobson, 2015).

For Pat1/Lsm1 substrates, it has been shown that loss of the Pat1-binding motifs had no effect on their stability, but deletion of PAT1 caused substantial stabilization. These results indicate that the Pat1-mediated recruitment of the decapping enzyme must not make a significant contribution to the overall decay rates of Pat1/Lsm1 substrates, suggesting that: decapping must not be rate-limiting in the decay of Pat1/Lsm1 substrates, active recruitment of the decapping enzyme to Pat1/Lsm1 substrates is dispensable, and Pat1 must perform one major function upstream of the recruitment of the decapping enzyme. Consistent with this conclusion, Pat1/Lsm1 substrates are still degraded by a decapping-dependent pathway in the absence of active recruitment of the decapping enzyme by Pat1. Significantly, eliminating all of the known-binding motifs in the Dcp2 C-terminal domain also did not alter the decay of Pat1/Lsm1 substrates.

Novel functions for decapping activators Edc3, Dhh1, and Pat1/Lsm1–7.

Recent genetic analyses of Dcp2 cis-binding elements broadened the role of Edc3 in mRNA decapping and revealed that Edc3 has both general and specific functions in controlling the decapping enzyme in vivo [100]. In addition to the stabilization of the two known Edc3 substrates, RPS28B mRNA and YRA1 pre-mRNA, it was shown that loss of the Edc3-binding motif E3–1 in Dcp2 caused selective stabilization of Dhh1 substrates. In addition, deletion of the EDC3 gene causes further stabilization of Dhh1-regulated mRNAs in yeast cells harboring dcp2 alleles that lack the E3–1 motif. These observations reveal a new direct role of Edc3 in selective targeting of the decapping enzyme to Dhh1 substrates and thus broaden the range of Edc3 substrates from just two transcripts to more than one thousand additional mRNAs [12]. This direct role of Edc3 in decapping of Dhh1 substrates seems surprising but is not totally unexpected, as Edc3 interacts directly with Dhh1 [71, 99] and Dhh1 association with the decapping enzyme is also dependent on Edc3 [100]. Interestingly, in contrast to the consequences of Dcp2 E3–1 cis-element deletion, trans deletion of EDC3 in cells with a wild-type DCP2 background had no effect on Dhh1 substrates [100]. This result suggests that one additional factor can bind to the E3–1 motif and target the decapping enzyme to Dhh1-regulated mRNAs. Significantly, Scd6 was recently shown to bind to the E3–1 motif and was able to form both Edc3-containing and Edc3-lacking decapping complexes [100]. These observations suggest that Scd6 may function either cooperatively or redundantly with Edc3 in targeting the decapping enzyme to Dhh1 substrates.

Besides functioning as a unique targeting component of the decapping enzyme for Edc3 and Dhh1 substrates, Edc3 was also shown to be a common component of multiple decapping complexes [100]. This observation raises the intriguing question of Edc3 function in these different complexes. Earlier biochemical and structural studies suggested that Edc3 binding to Dcp2 alleviates Dcp2 autoinhibition imposed by its inhibitory elements thus enhancing substrate binding by the decapping enzyme [76, 84]. Recent genetic experiments indicated that the core Edc3 component must carry out two additional important functions in mRNA decapping. One function is to provide the decapping enzyme with a set of unique Edc3 binding surfaces or modules and thus endow the decapping enzyme with Edc3 targeting specificity through Edc3 homodimerization. In support of this proposition, eliminating Edc3 binding to Dcp2 caused selective stabilization of both Edc3 and Dhh1 substrates [100]. Another Edc3 function is to promote the assembly of additional target-specific decapping complexes, such as the Dcp1-Dcp2-Edc3-Upf1 and Dcp1-Dcp2-Edc3-Scd6 complexes, by remodeling Dcp2 or providing weak but specific binding surfaces for specific targeting or coupling factors. Consistent with this notion deletion of EDC3 diminishes Dcp2-bridged Dcp1:Upf1, Dcp1:Scd6, and Dcp1:Xrn1 interactions, and leads to a switch of Upf1 binding from the U1₁ to the U1₂ motif in Dcp2 [100]. The proposition that Edc3 exists both as a common core component of multiple decapping complexes and as a unique targeting component of specific decapping complexes provides a unified theory that explains the otherwise contradictory proposed functions for Edc3, i.e., that Edc3 functions as a general decapping activator (Kshirsagar and Parker, 2004) or as a transcript-specific decapping activator (Badis et al., 2004; Dong et al., 2007).

Codon optimality was recently demonstrated to be a major determinant of mRNA stability from yeast to humans [154, 167–173] and yeast Dhh1 was proposed to function as a codon optimality sensor capable of detecting slowly elongating ribosomes and targeting mRNAs with poor codon optimality for decapping and degradation [53]. Consistent with this proposed role, Dhh1 is: i) required for rapid decay of mRNAs containing high levels of non-optimal codons, ii) capable of binding to ribosomes, and iii) preferentially associated with mRNAs comprised of non-optimal codons. Further, overexpression of Dhh1 yields increased ribosome A-site occupancy over non-optimal codons relative to wild-type cells, tethered Dhh1 enhances ribosome accumulation on a reporter mRNA with non-optimal codons but not a reporter mRNA with optimal codons, and the stimulatory effect of Dhh1 on decay of mRNAs with non-optimal codons is dependent on the number of slow elongating ribosomes [53]. However, recent experiments indicate that Dhh1 is unlikely to be a direct sensor of codon optimality and, instead, one or more components of the Ccr4/Not deadenylase complex appear to have that function [52, 174]. The N-terminal domain of Not5 binds to the E-site of ribosomes with an empty A site and anchors the Ccr4/Not complex to ribosomes, indicating that Not5 acts as a sensor recognizing a specific ribosome conformation during translation elongation. In addition, Not4 immunoprecipitation selectively enriches ribosomes with non-optimal codons occupying their A sites, loss of Not5, Not4, and Not4-mediated ubiquitination of the small ribosomal subunit protein eS7 causes selective stabilization of reporter mRNAs comprised of non-optimal codons, and deletion of the Not5 N-terminal domain or loss of eS7 ubiquitination abolishes Dhh1’s association with ribosomes [52]. Ccr4 and Caf1, the two catalytic subunits of the deadenylase complex, appear to have differential activities and loss of Caf1 preferentially stabilizes mRNAs of low codon optimality [174], suggesting that the codon-sensing mechanism is linked to deadenylation. Collectively, these observations indicate that the deadenylase complex functions as a sensor of codon optimality and that Dhh1 may function as an effector that couples codon sensing by Ccr4/Not to mRNA decapping by the decapping enzyme.

For some time Pat1 and the Lsm1–7 complex (Pat1/Lsm1–7) have been thought to function as activators of decapping as loss of these factors generally causes mRNA stabilization and increased accumulation [36–39]. New insights into the roles of Pat1/Lsm1–7 emerged recently from a study of the ATG mRNAs induced during nitrogen starvation [175, 176]. Gatica et al. analyzed the role of Pat1/Lsm1–7 in autophagy-related ATG mRNA expression and demonstrated that the Pat1/Lsm1–7 complex promotes the stability of specific ATG mRNAs [177]. Loss of Pat1 or Lsm1 resulted in decreased accumulation of specific ATG mRNAs, and this decrease was restored by eliminating Ski3, a component of the cytoplasmic exosome [177]. Pat1 is mostly phosphorylated during nutrient-rich conditions [178], but a significant fraction of Pat1 is dephosphorylated during nitrogen starvation, and this dephosphorylation is required for promoting the stability of specific ATG mRNAs and for Pat1 binding to these mRNAs [177]. These observations indicate that, during nitrogen starvation-induced autophagy, dephosphorylation of Pat1 binds to specific ATG mRNAs in combination with Lsm1–7, protects these mRNAs from exosome-mediated 3’ to 5’ decay, and thus ensures normal translation of these mRNAs and robust autophagy induction. This protective function of Pat1/Lsm1–7 for specific ATG mRNAs is likely related to a previous observation that loss of Pat1 or Lsm1 results in the accumulation of reporter mRNAs that are trimmed into their 3’-UTRs [39, 179]. The observation that Pat1/Lsm1–7 can promote the stabilization of specific ATG mRNAs during nitrogen starvation is intriguing. It is possible that under this stress condition, decapping of these ATG mRNAs becomes rate-limiting due to either reduced decapping enzyme activity or reduced active recruitment of the decapping enzyme, and these ATG mRNAs can be degraded by more efficient exosome-mediated 3’ to 5’ decay pathway in the absence of Pat1/Lsm1–7. Interestingly, Pat1/Lsm1–7 also binds selectively to specific mRNAs induced by osmotic stress, and Pat1/Lsm1–7 binding to these mRNAs appears to repress their translation under both stress and non-stress conditions, avoiding hyper-sensitivity under mild osmotic stress and excessive responses under real osmotic stress [180]. Dhh1, Pat1, and Lsm1 control the levels of expression of additional stress responsive mRNAs [12, 34], suggesting that these factors may play a general and important role in cellular responses to environmental stresses.

Cytoplasmic sites of mRNA decapping action: in P-bodies or on translating ribosomes?

In yeast as well as in higher eukaryotes different decapping activators and decapping-dependent decay enzymes are often concentrated in cytoplasmic granules called P-bodies [181–184]. Although P-body formation is a conserved feature of eukaryotic cells, the precise function of this structure in cytoplasmic mRNA metabolism is currently not clear. Since trapped mRNA decay intermediates have also been detected in these structures, P-bodies were originally suggested to be sites of mRNA decapping and decay [182, 184]. In yeast, P-bodies are mostly formed when cells are subjected to specific stress conditions or when the cellular decapping or 5’ to 3’ exonucleolytic decay activities are severely compromised, and P-body formation requires non-translating mRNAs [185]. It was thus suggested that P-bodies might function as sites for storage of translationally repressed mRNAs [186]. Recent experiments identified thousands of mRNAs associated with P-bodies in yeast and human cells and suggested that P-bodies may have a general function in translation repression and mRNA storage [187, 188]. However, whether mRNP assembly into a P-body is required or otherwise important for mRNA decapping and translation repression is controversial [189, 190].

Several lines of evidence indicate that P-bodies may not be the sites of decapping and that mRNP assembly into P-bodies has no significant role in mRNA decapping. First, RNA-Seq analysis indicated that P-body-associated mRNAs are full-length and have intact 5’-ends [187]. Second, mRNA decay intermediates originally detected in P-bodies have been shown to be artifacts of the MS2 stem loops (MS2-SLs) utilized to detect them, as binding to the MS2-SLs by the MS2-coat-GPF fusion proteins impairs mRNA decay and causes substantial accumulation of the MS2-SL binding sequences [191–193]. Third, quantitative analyses indicated that P-bodies sequester at most only a few percent of decapping factors from the cytoplasm even in cells lacking mRNA decapping activity [194, 195]. Finally, it is well established that in yeast, as well as in fly and human cells, P-bodies are not required for mRNA decay [135, 196, 197] and that mRNA decapping principally occurs on polyribosomes. The latter conclusion is supported by multiple experimental results, including: i) the decapping activators Upf1–3 [198], Dhh1, Pat1, Lsm1, and Scd6 [35, 52, 162, 199, 200] are all associated with polyribosomes; ii) the decay enzymes Dcp2 and Xrn1 are also mostly polyribosome-associated [199, 201, 202]; iii) decapped wild-type and nonsense-containing mRNAs are detected on polyribosomes in wild-type cells [111, 203]; iv) when 5’ to 3’ exoribonucleolytic decay is blocked by elimination of Xrn1 >80% of decapped transcripts are associated with polyribosomes [203]; v) co-translational 5’ to 3’ decay is a widespread process that acts on most mRNAs [204] ( see below); and vi) decapping still occurs when mRNAs are trapped on polyribosomes by treating cells with cycloheximide [203, 205]. These observations indicate that, for NMD and general 5’ to 3’ decay, decapping occurs while mRNAs are still associated with translating ribosomes, and ribosomes may even play active roles in recruiting decapping complexes [52, 206]. That said, decapping may still require translational repression at initiation [31] but, in contrast to an earlier model for mRNA decapping [45], a ribosome-free mRNP state is unlikely to be a prerequisite.

Mechanistic and physical coupling between decapping and 5’ to 3’ exoribonucleolytic decay

Pat1’s C-terminal extension binds to multiple helical leucine-rich motifs in Dcp2 and a single such motif in Xrn1, observations which led to a proposal that Pat1 coordinated the decapping and exonucleolytic decay events in general 5’ to 3’ mRNA decay by first recruiting the decapping enzyme and then Xrn1 to targeted mRNAs [73]. However, as Pat1 targets only a subset of yeast transcripts [12] Pat1-mediated coupling may be limited to a small number of mRNAs and is unlikely to be a general mechanism. An alternative coupling mechanism is suggested by recent two-hybrid experiments which revealed that Xrn1 binds to Dcp2 and is directly recruited to decapping complexes by Dcp2 [100]. Several observations supported these conclusions. First, Xrn1 was demonstrated to bind to a specific internal segment of Dcp2. Second, Xrn1 was also shown to interact with both Dcp1 and Edc3, two other core components of the decapping enzyme, and each of these interactions requires the Dcp2 C-terminal domain as well as an intact Xrn1-binding region in this domain. Third, Xrn1 appeared to be a common component of Dcp1-Dcp2-Edc3-Edc3-Xrn1 and Dcp1-Dcp2-Edc3-Upf1-Xrn1 complexes. And, finally, Xrn1’s recruitment to the decapping complexes is dependent on prior Edc3 binding to Dcp2, as deletion of EDC3 eliminated both Dcp1:Xrn1 and Upf1:Xrn1 interactions [100]. These results strongly indicate that decapping and 5’ to 3’ exoribonucleolytic decay are likely to be physically and mechanistically linked. This coupling appears to be conserved over eukaryotic evolution. However, the precise mechanism of this coupling may differ for different organisms, as D. melanogaster Xrn1 binds to Dcp1 [207] and human Xrn1 binds to Edc4 in their respective decapping complexes [208]. Xrn1 binding to decapping complexes likely ensures timely 5’ to 3’ exoribonucleolytic degradation of decapped transcripts and may also serve to inhibit the catalytic activity or substrate binding of the decapping enzyme until the enzyme is properly targeted to substrate mRNAs. Consistent with this hypothesis, it was shown that Xrn1 binding to Dcp2 requires the Dcp2 inhibitory element [100] and that overexpression of Xrn1 in D. melanogaster cells can inhibit the decapping of different reporter mRNAs [207].

Mechanistic and physical coupling between decapping and deadenylation

Decapping of general mRNAs requires prior deadenylation [31]. In yeast, mRNA deadenylation is carried out by the sequential actions of two distinct multi-subunit deadenylases, the Pan2/Pan3 and Ccr4/Not complexes [101]. Early experimental evidence suggested that mRNA deadenylation and decapping are kinetically coupled, as unstable mRNAs usually have high rates of both deadenylation and decapping and, in contrast, stable mRNAs usually have low rates of deadenylation and decapping [13, 209]. In addition, blocking translation initiation by either a cis-acting structural element or trans-acting factor mutations accelerates mRNA decapping, but also accelerates the rate of mRNA deadenylation [46, 47]. Further, 3’-UTR binding proteins target specific mRNAs for rapid decay and these factors accelerate both deadenylation and decapping of their targeted transcripts [21, 22, 210]. Recent genome-wide analyses indicate that mammalian mRNAs have a broad range of deadenylation rates that can vary over 1000-fold. Interestingly, decay rates of mRNAs after their poly(A) shortening also vary broadly and are usually rapid for those that have previously undergone more rapid deadenylation [211]. Similarly, microRNAs target specific mRNAs for rapid deadenylation, but also accelerate the decay of target mRNAs after poly(A) shortening [212].

The apparent kinetic coupling between deadenylation and decapping suggests that these two events are functionally coordinated in deadenylation-dependent mRNA decay, i.e., that they are mechanistically or physically linked. Numerous results support this idea, including demonstrations that: i) tethering the Not3 component of the Ccr4/Not deadenylase complex to a reporter mRNA stimulates mRNA decapping [213] ii) four components of the Ccr4/Not complex (Not2, Not3, Not4, and Not5) appear to have functions in mRNA decapping independent of their roles in deadenylation [213, 214], iii) overexpression of the decapping activator Dhh1 suppresses the phenotypes associated with the loss of Ccr4 or Caf1, two catalytic components of the deadenylase complex [215], iv) deletion of DHH1 exhibits synthetic lethality with a not1–2 mutation or deletions of CAF1 or NOT4 [216], v) Dhh1 co-precipitates with several components of the deadenylase complex, including Not1, Ccr4, and Caf1 [32, 215, 216], vi) Dhh1’s association with translating polyribosomes is dependent on Not5 [52], vii) Dhh1 associates with more than 2,000 mRNAs and the vast majority (87%) of Dhh1-associated transcripts are also targets of Ccr4 [217], viii) decapping activator Pat1 shows two-hybrid interactions with the Not3 and Not5 components of the Ccr4/Not complex [213], ix) the transcript-specific regulator, Puf5, recruits both the deadenylase and decapping complexes to its target mRNAs [21, 218], x) tethering fly or human Pat1 orthologs to reporter mRNAs can promote accelerated deadenylation in addition to rapid mRNA decapping [219, 220], and xi) the human Dhh1 ortholog, DDX6, binds directly to the central MIF4G domain of CNOT1, and this specific binding is required for DDX6 ATPase activity and its function in microRNA-mediated translation repression [221–223]. Collectively, these observations indicate that mechanistic coupling and functional coordination between the deadenylation and decapping machineries are generally conserved features of deadenylation/decapping-dependent decay pathways in eukaryotic cells. Such coupling and coordination ensure that deadenylated mRNAs will not accumulate in normal cells and enables different mRNA deadenylation rates to impart a wide range of mRNA stabilities.

A New Model for Control of mRNA Decapping

Extensive genetic analyses of the decapping enzyme and its interactors have led He et al. to propose a new model for the control of mRNA decapping in yeast [100]. In this model (Figure 5), Dcp2 cis-binding elements in collaboration with Edc3 control selective targeting of the decapping enzyme by orchestrating the formation of distinct multi-component decapping complexes. These complexes include Dcp1, Dcp2, Edc3, and Xrn1 as shared common components that are joined by distinct decapping activators that target each complex to specific mRNA substrates. Edc3 functions as a shared common component of multiple decapping complexes, and also as a unique targeting component of dimeric Edc3-containing complexes. Xrn1 is proposed to associate with Dcp2 after Edc3 binding to Dcp2 and to function as a component of multiple, or perhaps all, decapping complexes. Upf1 and Pat1 each serve as unique mRNA-targeting components in their respective Upf1- or Pat1-containing complexes, and Scd6 likely functions as a second targeting component of Dhh1-containing complexes and may also collaborate with Pat1 to target Pat1/Lsm substrates. The targeting components Edc3, Scd6, Upf1, and Pat1 each are equipped with at least two separate binding modules, one for the Dcp1-Dcp2-Edc3 core complex and another for their targeted mRNPs, and the final assembly and decapping activation of the target-specific decapping complexes occurs on a to-be-degraded mRNP. Finally, the decapping events are physically coupled with 5’ to 3’ degradation by Xrn1.Target-specific decapping complexes other than those specific for NMD, Edc3, Dhh1, and Pat1 substrates most likely exist in yeast cells, as several Dcp2 leucine-rich motifs originally proposed for Pat1 binding actually do not contribute to Pat1 binding in the context of full-length Dcp2 and thus still lack assigned specific binding partners. Accordingly, yeast cellsmay contain many distinct decapping complexes, and the notion of a single yeast decapping enzyme appears to be too simplistic and should be replaced by the concept of a decapping enzyme family that shares the core components Dcp1 and Dcp2, analogous to the PP1 and PP2A phosphatase families involved in protein dephosphorylation [224, 225].

Figure 5. A new model for control of mRNA decapping.

In this model, Dcp2 cis-binding elements collaborating with Edc3 promote sequential assembly of target-specific decapping complexes. Edc3 first joins the Dcp1-Dcp2 complex, and Xrn1 then joins the Dcp1-Dcp2-Edc3 complex to form a Dcp1-Dcp2-Edc3-Xrn1 core complex. A pool of Edc3 molecules as well as Upf1, Scd6, and Pat1 function as unique targeting components and direct the core decapping complex to distinct substrate mRNAs. Final assembly and decapping activation occur at decapping targeted mRNPs. Adapted from reference 100.

Except for Edc4, the decapping factors described here are conserved from yeast to human [30, 85]. The molecular mechanisms uncovered in yeast cells that control the selective targeting and activation of the yeast Dcp2 decapping enzyme could thus have broad application, i.e., they are likely to also operate in human cells. In contrast to yeast Dcp2, human Dcp2 has a much shorter disordered C-terminal domain and appears to lack specific binding motifs for different decapping activators [24, 25, 27]. Interestingly, human Dcp1 and Edc4 both contain long disordered regions [226, 227], thus suggesting that the regulatory mechanisms seen in yeast have been recapitulated in principle by alternative structural arrangements in human cells. Possibly, these two factors may rewire decapping factor interactions and encode yet-to-be-identified binding motifs for specific decapping activators in their disordered regions that ultimately control human Dcp2 targeting and activation. In addition, human Dcp1, Edc3, and Edc4 all self-associate and exhibit dimeric or trimeric interactions [63, 227, 228], and human cells also appear to possess multiple mRNA decapping enzymes [27–29]. These observations suggest that human cells may have evolved additional mechanisms that control decapping enzyme targeting and that different decapping enzymes themselves may be equipped with specific mRNA targeting properties.

Future perspectives

Genetic analyses in yeast have uncovered the molecular mechanisms that control the selective targeting of the decapping enzyme to its different substrate mRNAs and revealed new roles for specific decapping activators in controlling decapping enzyme targeting and assembly of target-specific decapping complexes. The genetic experiments also indicated the functional order and relationships for different decapping activators and suggested that most decapping activators may function in monitoring mRNA translation and promoting mRNP transition from a state of active translation to a state permissive for mRNA decapping. In addition, mRNA decapping is both functionally and mechanically coupled with upstream deadenylation and downstream 5’ to 3’ exoribonucleolytic decay. Genetic experiments over the last three decades have paved the way for future mechanistic studies. It will be of great importance to elucidate the functions and mechanisms of action of different Dcp2 cis-binding elements in promoting the binding of specific targeting factors (Figure 6), define the component composition of different decapping complexes, delineate the specific functions of different components in these decapping complexes, determine the timing and mRNP states that promote mRNA association by different decapping activators, and identify the molecular events that trigger the transition from deadenylation to decapping, or from decapping to 5’ to 3’ exoribonucleolytic decay. Given the intimate relationship between mRNA translation and decapping, it is very likely that decapping activators may also function as translational repressors when the decapping enzyme becomes limiting or when the expression or activity of a specific targeting component is increased under specific physiological or environmental conditions. In addition, since several decapping factors are phosphorylated under either normal or stressful growth conditions [176–178, 229], and Dcp2 can be relocated into the nucleus under specific genetic conditions [148, 195], cytoplasmic mRNA decapping may also be regulated by posttranslational modification of the decapping machinery or by the subcellular localization of decapping enzyme components.

Figure 6. Dcp2 cis-binding elements are mostly disordered.

A model of the Dcp2 three-dimensional structure was generated by the AlphaFold program [149]. Of the 13 Dcp2 cis-binding elements, only the Edc3 and Scd6-binding motif E3–1 and the Scd6-binding motif E3–2 are predicted to form structured helical elements (A). The two Upf1-binding motifs U1₁ and U1₂ (B) and the nine leucine-rich Pat1-binding motifs L₁ to L₉ (C) are predicted to be mostly disordered. In A, B, and C, residues in the Dcp2 model are color-coded according to their pLDDT scores. The L₁, L₂, and L₉ motifs were shown to form amphipathic α-helices when they bind to the Pat1-C domain [73], suggesting that these disordered motifs form defined structures upon binding to specific interacting partners.

Abbreviations:

ARE

AU-rich element

NMD

Nonsense-mediated mRNA decay

NRD

N-terminal regulatory domain

miRNA

microRNA

NAD

nicotinamide adenine dinucleotide

Nudix

nucleoside diphosphates linked moiety X

EVH1

Ena/Vasp Homology domain 1

Lsm

like sm

CH domain

cysteine- and histidine-rich zinc-finger domain

MIF4G

middle domain of eukaryotic initiation factor 4G

RRM

RNA recognition motif

EREs

Edc3 response elements

TREs

translational repression elements

uORFs

upstream open reading frames.