Translational gene regulation in plants: A green new deal

Abstract

The molecular machinery for protein synthesis is profoundly similar between plants and other eukaryotes. Mechanisms of translational gene regulation are embedded into the broader network of RNA-level processes including RNA quality control and RNA turnover. However, over eons of their separate history, plants acquired new components, dropped others, and generally evolved an alternate way of making the parts list of protein synthesis work. Research over the past 5 years has unveiled how plants utilize translational control to defend themselves against viruses, regulate translation in response to metabolites, and reversibly adjust translation to a wide variety of environmental parameters. Moreover, during seed and pollen development plants make use of RNA granules and other translational controls to underpin developmental transitions between quiescent and metabolically active stages. The economics of resource allocation over the daily light-dark cycle also include controls over cellular protein synthesis. Important new insights into translational control on cytosolic ribosomes continue to emerge from studies of translational control mechanisms in viruses. Finally, sketches of coherent signaling pathways that connect external stimuli with a translational response are emerging, anchored in part around TOR and GCN2 kinase signaling networks. These again reveal some mechanisms that are familiar and others that are different from other eukaryotes, motivating deeper studies on translational control in plants. This article is categorized under: Translation > Translation Regulation RNA Structure and Dynamics > Influence of RNA Structure in Biological Systems RNA Interactions with Proteins and Other Molecules > Protein-RNA Interactions: Functional Implications.

Keywords: plant; regulation; signaling; translation.

Figure 1.. General overview of cytosolic mRNA translation.

Translation initiation begins with the assembly of a 48S pre-initiation complex. For this, the 40S ribosomal subunit loaded with initiation factors (1A, 1, 3, 5) docks with the ternary complex (eIF2, GTP and Met-tRNAi) to form the 43S pre-initiation complex. Meanwhile, the linear mRNA is thought to form a closed loop by virtue of PABPs on the poly(A)-tail interacting with the cap-binding complex eIF4F, consisting of eIF4G scaffold and eIF4E cap binding proteins. The resulting closed-loop structure is loaded onto the 43S pre-initiation complex likely by contacts between eIF4G and eIF3 to form the 48S scanning complex. Loading of the mRNA into the 40S is facilitated by the helicase eIF4A with assistance from the RNA binding protein eIF4B. The 48S then scans the 5’-leader of the mRNA. Upon AUG recognition by the ternary complex, eIF1, 1A and 5, GTP hydrolysis by eIF2 locks the initiator Met-tRNAi into the peptidyl (P)-site of the 40S, assuming a closed confirmation. Upon release of all other initiation factors the GTPase eIF5B catalyzes the joining of the 60S to complete the assembly of a functional 80S ribosome. Next, elongation factor eEF1A delivers a charged aminoacyl-tRNA into the A-site of the 80S; a correct codon-anticodon match triggers GTP hydrolysis by eEF1A and the release of eEF1A-GDP. Following peptide bond formation by the peptidyltransfrase activity of the ribosome elongation factor eEF2 uses GTP hydrolysis to translocate the mRNA-tRNA complex by one codon. The CDS is decoded into a protein by successive cycles of aminoacyl-tRNA binding, peptide bond formation, and translocation of the spent tRNA into the exit (E)-site. A stop codon in the A-site triggers termination by the release factor complex (eRF1-eRF3-GTP). GTP hydrolysis by eRF3 triggers its release and opens the binding for the recycling factor, ABCE1, at the A-site. This releases the polypeptide from the P-site and a final ATP hydrolysis by ABCE1 promotes the disassembly of the 80S components, to be recycled for another round of protein synthesis. For details refer to Browning and Bailey-Serres 2015.

Figure 2:. Regulation of translatomes in pollen development and seed germination.

Yellow boxes indicate translation state, abundance of polysome or monosomes and cellular RNP complexes; blue (pink) boxes indicate functional gene ontology terms that are translationally active (repressed); green boxes indicate examples of translationally repressed mRNAs corresponding to a specific developmental stage. (A) Pollen. The vertical timeline covers (pre-)meiotic (sporophytic, diploid) followed by postmeiotic (gametophytic, haploid) stages of tobacco pollen. Early stage proteomes are characterized by abundant heat shock protein chaperones (HSPs), RNA binding proteins (RBPs), cell wall loosening enzymes and protein degradation, while the later gametophytic stage is rich in glycolytic and fermentation enzymes and late embryo abundant (LEA) proteins. LEA proteins are linked to ABA signaling, cell wall synthesis and the maintenance of ROS balance. Mature, desiccated pollen grains contain a high amount of monosomes in the form of special structures called EDTA/Puromycin resistant Particles (EPPs), which are translationally inactive. EPPs contain stalled monosomes on transcripts maintained in a translationally quiescent state, to be released for translation towards the maturation phase of pollen germination. The EPP proteome contains ribosomal 60S proteins and translation initiation and elongation factors. Examples of the mature pollen EPP transcriptome include pollen-specific cell wall glycoprotein (NTP303 ortholog ATSKU5) and pollen-specific LIM domain containing protein (WLIM2B). After Hafidh et al., 2018 and Ischebeck et al., 2014. (B) Seed germination begins with a dry seed (top=early) and involves two distinct translational shifts, coincident with seed hydration and root protrusion (‘germination’ proper). During the hydration shift (0–6 h after imbibition) transcripts found abundant in the last stages of seed maturation become polysome loaded and actively translated (e.g. PM1 is translated during this phase and suppresses the translation of a glycine-rich protein). The second translational shift (26–48 h) occurs between testa rupture and root protrusion. At this stage, abundant polysomes contain transcripts related to RNA processing and modification and lipid metabolism. After Bai et al., 2017, Bai et al., 2020.

Figure 3.. Noncanonical mechanisms of translational enhancement in plant viruses.

The main protein-coding ORFs of the viral mRNA are represented by a single red box. (A) The viral genome-linked protein Vpg substitutes for a 5’ cap in certain RNA viruses by attracting eIF4E and the cap binding complex to the 5’ end of the viral RNA. (B) Shunting (upper) and reinitiation (lower schematic) of ribosomes on the long 5’ leader of pararetroviruses. For details see text. (C) An internal ribosome entry site in triticum mosaic virus relies on a critical hairpin loop and associated polypyrimidine sequence to attract eIF4G and the ribosome. (D) 3’ cap-independent translational enhancers come in a wide variety of forms. The schematic is not drawn to scale and illustrates shared concepts rather than one specific example. Common elements include 3’ hairpin structures that attract elements of the cap-binding and preinitiation complexes and interact with the 5’ leader of the viral RNA through kissing-loop basepairing interactions. For details see text.

Figure 4.. uORFs can control translation in response to metabolites.

(A) Structure of an mRNA with one main ORF (mORF, yellow) and three uORFs (blue), one of which overlaps the mORF. Start codons are highlighted with bent arrows. (B) Reporter gene translation assay designed to quantify leaky scanning versus reinitiation on an mRNA with a single uORF. As compared to the wild-type allele (I), in allele (II) the start codon of the uORF has been mutated to prevent initiation. In allele (III), the stop codon of the uORF has been mutated so as to extend the uORF into the main ORF, which precludes reinitiation at the Reporter AUG. Additional mutations may be needed to remove any downstream in-frame stops in the 5’ leader and to ensure that the uORF is out of frame with the main ORF. The rates of leaky scanning past the uORF initiation codon (LS), of initiation at the uAUG (I), of initiation-reinitiation (I/R) and of failure to reinitiate (I/NR) are calculated from the empirical reporter gene expression activities A_I, A_II and A_III as shown in the table. Numbers shown are for illustration only. (C) uORF-mediated metabolic control on the GGP (GDP-galactose phosphorylase) mRNA (after Laing et al., 2015). At a high ascorbate level, the non-canonical start codon (ACG) of the uORF is recognized and the encoded conserved peptide disallows reinitiation thus switching translation of the main ORF off. At low ascorbate the ACG codon is thought to be skipped by leaky scanning. (D) uORF-mediated metabolic control on the AdoMetDC (adenosyl-methionine decarboxylase) mRNA (after Hanfrey et al., 2005 and Uchiyama-Kadokura et al., 2014). At a low polyamine level, ribosomes initiate on uORF1, thus bypass the start of uORF2, and reinitiate on the main ORF. At a high polyamine level, uORF1 is skipped and uORF2 is translated. uORF2 codes for a conserved peptide and is not conducive to reinitiation, thus inhibiting translation of AdoMetDC by negative feedback.

Figure 5.. The expression of termination factor eRF1 is regulated by negative feedback.

(A) Structure of the eRF1 mRNA. After splicing of an intron from the 3’UTR, the 3’UTR is believed to be marked by an exon junction complex. The canonical eRF1 CDS terminates with a stop codon that is prone to translational readthrough (weak stop), which is followed by a regular in-frame stop codon downstream of the exon-exon junction. (B) Alternate outcomes during the first round of translation. When the eRF1 level is high, the canonical termination codon is recognized by eRF1. The exon junction complex remains and triggers degradation of the mRNA by NMD. When the eRF1 level is low, the weak, canonical termination codon is not recognized, translation readthrough occurs and translation stops at the downstream stop. In the process, the exon junction complex is removed thus keeping the mRNA stable and the rate of eRF1 synthesis high. After (Nyiko et al., 2017).

Figure 6.. Translational control by proteins associated with the eIF4 cap-binding complex.

CBE1 is an eIF4E binding protein that is phosphorylated in a TOR-dependent manner (Patrick et al., 2018). CERES likewise forms cap-binding complexes with eIF4E/iso4E, which lack eIF4G/iso4G (Toribio et al., 2019). Activity of the RNA helicase eIF4A is stimulated by MRF proteins, likely through phosphorylation through the TOR kinase pathway (based on Lee et al., 2017). In contrast, phosphorylation of eIF4A by cyclin-dependent kinase may inhibit eIF4A (after Hutchins et al., 2004 and Bush et al., 2016). The RNA binding protein SOAR1 inhibits translation initiation, likely by disrupting the interaction between eIF4G and poly(A) binding proteins. SOAR1 can also bind directly to mRNAs such as the one for ABI5, an abscisic acid-signaling transcription factor (after Mei et al., 2014 and Bi et al., 2019). The jasmonic acid inducible protein JIP60 in barley undergoes proteolytic cleavage upon methyl-jasmonate (MeJA) treatment, which liberates a domain resembling ribosome inactivating protein (RIP30) and an eIF4E-like domain. Uncleaved JIP60 also induces the dissociation of the 80S subunit by an unknown mechanism (after Rustgi et al., 2014). MRF: MA3-containing translation regulatory factor; CDKA: Cyclin-dependent kinase A; eIF: Eukaryotic initiation factor; SOAR1: Suppressor of the ABA Receptor-overexpressor 1; JIP: Jasmonate-induced protein; RIP: Ribosome inactivating protein. Some of the findings presented in this figure are correlative or based on circumstantial evidence.

Figure 7.. Effects of SnRK signaling on translation.

The Snf1 related kinases SnRK1 and SnRK2 are activated during energy deprivation and ABA signaling. Phosphorylation by SnRK1 of eIFiso4G contributes to the repression of translation under hypoxia in maize (note color coding). Phosphorylation of eIF4E and eIFiso4E by SnRK1 is observed in vitro. Both SnRK1 and SnRK2 also lead to phosphorylation of Raptor in the TOR complex, which may contribute to TOR inhibition during energy deprivation and ABA stress signaling. Some of the findings presented here are correlative or based on circumstantial evidence. ADH: Alcohol dehydrogenase, LBD: LOB domain-containing protein, NIP: Nodulin-26 like intrinsic protein. For details see text.

Figure 8:. Overview of the cytosolic translation regulatory network anchored on three major protein kinases, TOR, SnRKs and GCN2.

(Left) Four TOR outputs. (i) TOR triggers phosphorylation of the h subunit of eIF3 through the S6 kinase. This event supports the role of eIF3h in reinitiation after uORF translation. It is triggered by light stimulated synthesis of auxin and subsequent activation of TOR by the Rho-like small G protein (ROP2). (ii) Phosphorylation of the RNA polymerase III repressor protein, Maf1, boosts synthesis of tRNAs and rRNAs. (iii) TOR induces the phosphorylation of ribosomal protein eS6 through the S6 kinase, although the biochemical significance of this event remains unknown. (iv) TOR phosphorylation of MRF proteins through S6 kinase triggers MRF1 association with eIF4A and activates translation. (v) Sulfur deficiency depletes TOR activity, possibly via low glucose and SnRK signaling. (Middle) Interaction of SnRKs and TOR. (i) Under conditions of low energy, active SnRK1 may phosphorylate Raptor1B, which may trigger disassembly of the active TOR complex and translational inhibition. High energy signals such as trehalose-6-phosphate, glucose −6 or −1-phosphate inhibit repression of SnRK1 on TOR complex. (ii) ABA also activates SnRK1 via its negative regulator, protein phosphatase 2C (PP2C). (iii) SnRK2 is activated by ABA under abiotic stress conditions and may likewise lead to Raptor phosphorylation. SnAK is the SnRK activating kinase whose binding to SnRK is inhibited by T6P. Other phosphorylation targets of SnRK are listed in Table 1. (Right) GCN2 kinase is co-activated by GCN1 and uncharged tRNA. GCN2 phosphorylates the α subunit of the GTPase eIF2. GCN2 is activated by multiple abiotic and biotic stressors, some of which may deplete amino acids and increase deacylated tRNAs. Depletion of O-acetyl serine, a precursor of cysteine, stimulates GCN2. Reactive oxygen species (ROS) arise from the photosynthetic apparatus under excess light and activate GCN2. ROS also accompany many of the other triggers known to activate GCN2 such as herbicide and cold. Black arrows and red T-bars represent activation or inhibition, respectively. Dashed lines indicate indirect or hypothetical connections. Based on Schepetilnikov et al., 2013 and , Rodrigues et al., 2013; Nukarinen, 2016; Lee and Pai, 2017; Izquierdo et al., 2018; Ahn and Pai, 2019; Van Leene et al., 2019, Wang et al., 2018, Rodriguez et al., 2019, Dong et al., 2017, Lageix et al., 2008, Wang et al., 2018, Lokdarshi et al., 2020a, Lokdarshi et al., 2020b.