Author response: The molecular basis of coupling between poly(A)-tail length and translational efficiency

Article Figures and data Abstract eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract In animal oocytes and early embryos, mRNA poly(A)-tail length strongly influences translational efficiency (TE), but later in development this coupling between tail length and TE disappears. Here, we elucidate how this coupling is first established and why it disappears. Overexpressing cytoplasmic poly(A)-binding protein (PABPC) in Xenopus oocytes specifically improved translation of short-tailed mRNAs, thereby diminishing coupling between tail length and TE. Thus, strong coupling requires limiting PABPC, implying that in coupled systems longer-tail mRNAs better compete for limiting PABPC. In addition to expressing excess PABPC, post-embryonic mammalian cell lines had two other properties that prevented strong coupling: terminal-uridylation-dependent destabilization of mRNAs lacking bound PABPC, and a regulatory regime wherein PABPC contributes minimally to TE. Thus, these results revealed three fundamental mechanistic requirements for coupling and defined the context-dependent functions for PABPC, which promotes TE but not mRNA stability in coupled systems and mRNA stability but not TE in uncoupled systems. eLife digest Cells are microscopic biological factories that are constantly creating new proteins. To do so, a cell must first convert its master genetic blueprint, the DNA, into strands of messenger RNA or mRNA. These strands are subsequently translated to make proteins. Cells have two ways to adjust the number of proteins they generate so they do not produce too many or too few: by changing how many mRNA molecules are available for translation, and by regulating how efficiently they translate these mRNA molecules into proteins. In animals, both unfertilized eggs and early-stage embryos lack the ability to create or destroy mRNAs, and consequently cannot adjust the number of mRNA molecules available for translation. These cells can therefore only regulate how efficiently each mRNA is translated. They do this by changing the length of the so-called poly(A) tail at the end of each mRNA molecule, which is made up of a long stretch of repeating adenosine nucleotides. The mRNAs with longer poly(A) tails are translated more efficiently than those with shorter poly(A) tails. However, this difference disappears in older embryos, when both long and short poly(A) tails are translated with equal efficiency, and it is largely unknown why. To find out more, Xiang and Bartel studied frog eggs, and discovered that artificially raising levels of a protein that binds poly(A) tails, also known as PABPC, improved the translation of short-tailed mRNAs to create a situation in which both short- and long-tailed mRNAs were translated with near-equal efficiency. This suggested that short- and long-tailed mRNAs compete for limited amounts of the translation-enhancing PABPC, and that long-tailed mRNAs are better at it than short-tailed mRNAs. Further investigation revealed that eggs also had to establish the right conditions for PABPC to enhance translation and had to protect mRNAs not associated with PABPC from being destroyed before they could be translated. Overall, Xiang and Bartel found that in eggs and early embryos, PABPC and poly(A) tails enhanced the translation of mRNAs but did not influence their stability, whereas later in development, they enhanced mRNA stability but not translation. This research provides new insights into how protein production is controlled at different stages of animal development, from unfertilized eggs to older embryos. Understanding how this process is regulated during normal development is crucial for gaining insights into how it can become dysfunctional and cause disease. These findings may therefore have important implications for research into areas such as infertility, reproductive medicine and rare genetic diseases. Introduction Most eukaryotic mRNAs are polyadenylated at their 3′ ends in a process associated with transcriptional termination. In the nucleus, these poly(A) tails can facilitate mRNA nucleocytoplasmic export (Kühn and Wahle, 2004), whereas in the cytoplasm, they serve as molecular timers for mRNA decay, with their lengths becoming progressively shorter by deadenylation, which eventually leads to mRNA de-capping and turnover (Chen and Shyu, 2011; Eisen et al., 2020; Goldstrohm and Wickens, 2008). The length of a poly(A) tail can also influence mRNA translational efficiency (TE). Pioneering studies in maturing oocytes and early embryos show that lengthening of poly(A) tails through cytoplasmic polyadenylation is critical for regulating gene expression during these early stages of animal development (Richter, 1999; Sallés et al., 1994; Sheets et al., 1995). Results from these and other single-gene studies in oocytes and early embryos had led to the notion that the length of a poly(A) tail generally correlates with TE (Eckmann et al., 2011; Weill et al., 2012). More recent transcriptome-wide studies confirm a strong global relationship between tail length and TE in oocytes and early embryos (Eichhorn et al., 2016; Lim et al., 2016; Subtelny et al., 2014). However, in fish, frogs, and flies, this correlation diminishes near the time of gastrulation, and coupling between poly(A)-tail length and TE is essentially nonexistent in post-embryonic systems (Eichhorn et al., 2016; J.-E. Park et al., 2016; Subtelny et al., 2014). Thus, these global analyses reveal a developmental transition in how translation is regulated (Subtelny et al., 2014), which closely follows the long-known maternal-to-zygotic transition in transcriptional control. The existence of this transition in translational control brings to the fore mechanistic questions as to how coupling between poly(A)-tail length and TE is established in oocytes and early embryos and why this coupling disappears later in development. Cytoplasmic poly(A)-binding proteins (PABPCs) are highly conserved RNA-binding proteins in eukaryotes (Mangus et al., 2003). Although Saccharomyces cerevisiae has only one PABPC (Pab1p), most animals contain multiple paralogs that have spatially and temporally varied expression patterns (Smith et al., 2014; Wigington et al., 2014). PABPCs have high affinity to poly(A) sequences in vitro (Kd ~5 nM for A25) and require at least 12 As for efficient binding (Kühn and Wahle, 2004). Binding of PABPCs to mRNA poly(A) tails can enhance translation, but the mechanism of this enhancement is unclear. One model posits that the mRNA forms a closed-loop structure mediated by the association of the eukaryotic translation initiation factor eIF4G (a scaffolding protein) with both PABPC and the cap-binding protein eIF4E (Hinnebusch, 2014; Thompson and Gilbert, 2017; Wells et al., 1998). This association is proposed to stabilize the interaction between eIF4E and the mRNA 5′ cap and facilitate recruitment and/or recycling of ribosomes to increase translation initiation (Kahvejian et al., 2001). However, despite direct visualization of loop-like assemblies both within some cells and in an in vitro reconstituted system (Christensen et al., 1987; Wells et al., 1998), results of several studies have questioned the universality of this model among different mRNAs and biological systems (Adivarahan et al., 2018; Amrani et al., 2008; Costello et al., 2015; Rissland et al., 2017; Thompson and Gilbert, 2017). PABPCs can also influence mRNA stability, as shown in yeast. Genetic ablation of yeast Pab1p is lethal and causes lengthening of steady-state poly(A)-tail lengths (Sachs and Davis, 1989), which is attributed to pre-mature mRNA decapping and compromised deadenylation (Caponigro and Parker, 1995). Both yeast and mammalian PABPCs can interact with two mRNA deadenylation complexes PAN2-PAN3 and CCR4-NOT, and either promote or inhibit their activities in vitro (Schäfer et al., 2019; Uchida et al., 2004; Webster et al., 2018; Yi et al., 2018). Because mRNA decay is coupled to deadenylation (Decker and Parker, 1993; Eisen et al., 2020), the deadenylation-stimulatory effects of PABPC would accelerate the demise of bound mRNAs, which contrasts to other studies suggesting PABPC protects mRNAs from degradation in cell extracts (Bernstein et al., 1989; Wang et al., 1999). The dichotomous and potentially conflicting functions of metazoan PABPC examined in vitro raise the question of the extents to which PABPC might influence mRNA poly(A)-tail length and stability in metazoan cells. PABPCs are generally thought to coat mRNA poly(A) tails in the cytoplasm (Kühn and Wahle, 2004; Mangus et al., 2003). However, the stoichiometry between PABPC and poly(A) sites might vary in different biological contexts (Cosson et al., 2002; Voeltz et al., 2001), and it is unclear whether this potentially variable stoichiometry might impact gene regulation in cells. Moreover, the possibility that PABPC might influence protein synthesis by affecting either mRNA stability or TE can complicate analysis of its molecular functions in different biological systems, leaving its mechanistic roles poorly understood. Here, we uncover mechanistic requirements for coupling between poly(A)-tail length and TE observed in oocytes and early embryos, showing that this coupling and the subsequent uncoupling observed later in development rely on a context-dependent switch in the function of PABPCs. Results Limiting PABPC is required for tail length to strongly influence TE of reporter mRNAs To assay the influence of poly(A)-tail length on TE, we used an in vitro translation extract made from stage VI Xenopus laevis oocytes, where cytoplasmic polyadenylation leads to translational activation of the c-mos, cdk2 and some cyclin mRNAs (Richter and Lasko, 2011). Into this extract we added Nanoluc luciferase (Nluc) reporter mRNAs with either a short (29 nt) or a long (139 nt) poly(A) tail (Figure 1A). These mRNAs were made by in vitro transcription from DNA templates that encoded the mRNA body followed by the poly(A) tail as well as the hepatitis delta virus (HDV) self-cleaving ribozyme, which cleaved during in vitro transcription to generate not only a defined 3′ end at the desired poly(A)-tail length but also a 2′−3′-cyclic phosphate designed to inhibit undesired lengthening or shortening of the tail (Avis et al., 2012). Figure 1 with 2 supplements see all Download asset Open asset PABPC overexpression uncouples poly(A)-tail length and TE in frog oocytes. (A) Schematic of capped T7 transcripts with two different tail lengths, which were used as reporter mRNAs. Additional sequences beyond the HDV sequence are not depicted. (B) The effect of tail length on relative yields of in vitro translation of short- and long-tailed Nluc reporter mRNAs, in either frog oocyte extract (left) or rabbit reticulocyte lysate (right). The number above each bracket indicates the fold difference of the mean normalized luciferase signal (error bars, standard deviation from three technical replicates). (C) The effect of purified PABPC1 on relative yields of in vitro translation of short- and long-tailed Nluc reporter mRNAs in frog oocyte extract. Purified eGFP and PABPC1 were each added as indicated. Otherwise, this panel is as in (B). (D) Experimental scheme for serial-injection of mRNAs into frog oocytes. (E) The effect of overexpressing PAPBC1 and a PABPC1 M161A mutant on relative translation of Nluc reporter mRNAs in frog oocytes. Differential PABPC1 expression was achieved by injecting the indicated amount of mRNA in the first injection (error bars, standard deviation from three biological replicates). Otherwise, this panel is as in (B). (F) The effect of PAPBC1 overexpression on translation of reporter mRNAs with different 3′-end structures in frog oocytes. Shown are raw luciferase yields from Nluc reporters that have either a short poly(A) tail, a long poly(A) tail, a histone mRNA 3′-end stem-loop, or a Malat1 triple-helix 3′-end in oocytes overexpressing either eGFP or PABPC1 (error bars, standard deviations from three biological replicates). p values are from one-sided t-tests (n.s., not significant). For overexpression, 2.4 fmol mRNA was injected per oocyte. Figure 1—source data 1 Source data for luciferase values shown in Figure 1 and Figure 1—figure supplement 2. https://cdn.elifesciences.org/articles/66493/elife-66493-fig1-data1-v1.xlsx Download elife-66493-fig1-data1-v1.xlsx When added to the frog oocyte extracts together with a firefly luciferase (Fluc) mRNA, used to normalize for overall translation activity, the long-tailed reporter was translated substantially better than was the short-tailed reporter (Figure 1B). In contrast, the same reporter mRNAs were translated nearly equally well in rabbit reticulocyte lysate, a post-embryonic differentiated system for which no coupling between tail length and TE was expected (Subtelny et al., 2014; Figure 1B). Similar results were observed for an analogous pair of Renilla luciferase reporter mRNAs (Figure 1—figure supplement 2A). In both the oocyte and reticulocyte systems, the reporter mRNAs were stable with no detectable changes to their tail lengths (Figure 1—figure supplement 1A–C). Thus, the large difference in luciferase signal observed between the short- and long-tailed reporters in the oocyte extract was attributable to a difference in TE. These results showed that the causal relationship between longer poly(A)-tail length and greater TE observed for some maturation-specific mRNAs in frog oocytes (Sheets et al., 1995; Stebbins-Boaz and Richter, 1994) is not unique to those mRNAs, and indicated that frog oocyte extracts provide a system for probing the mechanism that couples tail length to TE. When considering the potential mechanisms for reading out tail length and promoting translation, a role for PABPC seemed plausible. For instance, translation might be sensitive to the number of PABPC molecules associated with an mRNA. In one mechanistic possibility, PABPC might be in excess over its binding sites within tails, such that tails are coated with the protein, as is generally thought to occur (Kühn and Wahle, 2004; Mangus et al., 2003), in which case, mRNAs with longer tails might be detected as those able to bind more PABPC molecules. At another mechanistic extreme, PABPC might be limiting, such that mRNAs compete with each other for PABPC binding, in which case, those with long poly(A)-tail lengths would compete more effectively and thereby preferentially benefit from any enhancement in TE that PABPC binding confers. To distinguish between these possibilities, we increased available PABPC in our oocyte extracts, reasoning that if PABPC were already coating the tails, adding more would have little effect, whereas if PABPC were limiting, adding more would diminish the competition for PABPC binding and thereby reduce the difference in TE observed between short- and long-tailed mRNAs. Accordingly, we purified recombinant Xenopus PABPC1 to near homogeneity (Figure 1—figure supplement 1D) and examined its influence when added to the in vitro translation extract derived from stage VI oocytes. As more PABPC1 was added, translation of the short-tailed reporter increased, with little change in translation of the long-tailed reporter, whereas adding equivalent amount of eGFP had little impact on translation of either reporter (Figure 1C). This concentration-dependent diminution of coupling between tail-length and TE strongly supported the hypothesis that limiting PAPBC is required for strong coupling. To investigate whether this requirement of limiting PABPC was restricted to our in vitro extracts or whether it also applied to living oocytes, we performed serial-injection experiments in oocytes. Stage VI frog oocytes were first injected with either PABPC1 mRNA or a control, and after waiting 24 hr to allow PABPC1 protein to accumulate (Figure 1—figure supplement 1E), oocytes were injected with the reporter mRNAs and assayed for luciferase activity (Figure 1D). Whereas injecting the control mRNA, eGFP, had no more influence than injecting water, injecting PABPC1 mRNA significantly reduced the extent to which poly(A)-tail length and TE were coupled (Figure 1E). Similar results were observed for an analogous pair of Renilla luciferase reporter mRNAs or when injecting ePAB mRNA rather than PABPC1 mRNA (Figure 1—figure supplement 2B). Reporter poly(A)-tail lengths did not change over the course of the experiment (Figure 1—figure supplement 1F), which indicated that the increased relative translation of the short-tailed reporter mRNA was not due to elongated poly(A) tails. Introducing additional PABPC into frog oocytes specifically improved translation of the short-tailed reporter while having little effect on translation of either the long-tailed reporter or reporters for which tails were replaced with either a stem-loop from the 3′ end of a histone mRNA (Ling et al., 2002) or a triple-helix from the 3′ end of the Malat1 non-coding RNA (Wilusz et al., 2012; Figure 1F). The observation that mRNAs required a tail to benefit from added PABPC indicated that the effects of adding PABPC were mediated in cis through tail-bound PABPC molecules, and were direct and not some secondary consequence of altering translation. Moreover, the observation that PABPC had little effect on translation of long-tailed mRNAs suggested that these mRNAs competed for the limiting endogenous PABPC so effectively that binding of additional PABPC imparted no detectable additional benefit to their translation. Introducing PABPC1(M161A), which encodes a PABPC1 mutant that is unable to bind eIF4G (Groft and Burley, 2002), also diminished coupling but did so by repressing translation of the long-tailed reporter. The reduced translation of the long-tailed reporter was presumably due to a dominant-negative effect of replacing functional endogenous PABPC molecules with defective ones. The observation that the long-tailed reporter was preferentially affected agreed with our conclusion that endogenous PABPC was limiting and preferentially binding to long-tailed mRNAs. The idea that the M161A mutant was unable to enhance translation in frog oocytes implied that the ability for PABPC to bind eIF4G and form the closed-loop structure is important for enhancing translation in this context (Wakiyama et al., 2000). In summary, our results with reporters in oocytes and oocyte extracts confirmed both the positive effect of PABPC on translation and the causal relationship between poly(A)-tail length and TE in these systems. Moreover, these results revealed that strong coupling between poly(A)-tail length and TE requires limiting PABPC. Limiting PABPC is required for tail length to strongly influence TE of endogenous mRNAs To examine the global effect of increasing PABPC on the translational regulatory regime acting in the oocyte, we monitored the relationship between tail length and TE for endogenous mRNAs of the oocytes. As expected from results of single-gene experiments in frog oocytes (Figure 1E; Sheets et al., 1995; Stebbins-Boaz and Richter, 1994) and the strong coupling between poly(A)-tail length and TE observed in both frog embryos and fly oocytes (Eichhorn et al., 2016; Lim et al., 2016; Subtelny et al., 2014), we found that poly(A)-tail length correlated strongly with TE in stage VI frog oocytes (Figure 2A). Overexpressing either PABPC1 or ePAB in these oocytes significantly diminished the coupling, with the Spearman correlation (Rs) for the relationship between tail length and TE dropping from 0.62 to 0.36 and 0.38, respectively (Figure 2A, both p = 0, modified Dunn and Clark's z-test [Diedenhofen and Musch, 2015]). In contrast, overexpressing eGFP had no significant impact on the coupling (p = 0.11), which indicated that this transcriptome-wide effect was a result of additional PABPC protein rather than a non-specific effect of adding more mRNA. Figure 2 with 1 supplement see all Download asset Open asset Increased PABPC promotes translation of endogenous short-tailed mRNAs, thereby diminishing coupling between tail length and TE. (A) The effect of PABPC on coupling between tail length and TE in frog oocytes. Shown is the relationship between TE and median poly(A)-tail length in oocytes injected with either water or mRNAs encoding either eGFP, PABPC1, or ePAB (injecting 16 fmol mRNA per oocyte). Results are shown for mRNAs from genes with ≥100 poly(A) tags. Each poly(A) tag represents a pair of sequencing reads that identify the mRNA and its poly(A) tail length. Rs is the Spearman's correlation coefficient. (B) A global increase in TE observed upon overexpressing ePAB. TE in ePAB-overexpressing oocytes is compared with TE in eGFP-expressing oocytes. Experiment was as in A, except rRNAs were not depleted, and TE values are normalized to those of mitochondrial mRNAs. Also shown are results from a co-injected short-tailed Nluc reporter and a long-tailed Fluc reporter. (C) Effect of expressing PABPC1 on protein synthesis in oocytes. At the left is a gel showing total protein after injection of either water or the indicated mRNA (injecting 4 fmol mRNA per oocyte), with or without treatment with cycloheximide (CHX), as visualized by Coomassie staining. At the right is the same gel showing protein synthesis, as visualized by incorporation of 35S-methionine and 35S-cysteine. Prominent bands presumably represent PABPC1, eGFP, and a not fully denatured form of eGFP (asterisk) expressed from injected mRNAs. (D) Quantification of the effect of expressing PABPC1 on protein synthesis in oocytes, as measured in (C) and two additional biological replicates. Only regions above the PABPC1 band, and between the PABPC1 band and the top eGFP band, were used for quantification. Values were normalized to that of the mean value from water-injected oocytes (error bars, standard deviation; p values, one-sided t-tests; n.s., not significant). (E) The preferential effect of overexpressing ePAB on the TE of short-tailed mRNAs. TE fold changes observed between ePAB-overexpressing and eGFP-expressing oocytes are plotted as a function of median tail length in eGFP-expressing oocytes. TE and tail-length values were obtained from different batches of oocytes; results are shown for mRNAs from genes with ≥100 poly(A) tags. TE values are normalized to those of mitochondrial mRNAs. (F) Effect of overexpressing ePAB on the distribution of TE values observed in frog oocytes. Shown is the TE distribution observed in ePAB-overexpressing oocytes and that observed in eGFP-expressing oocytes. Tail-length measurements in this figure were obtained using TAIL-seq. Figure 2—source data 1 Source data for values shown in Figure 2D. https://cdn.elifesciences.org/articles/66493/elife-66493-fig2-data1-v1.xlsx Download elife-66493-fig2-data1-v1.xlsx Accompanying the reduced coupling observed upon PABPC overexpression was a significant relative increase of TE for short-tailed mRNAs, an effect not observed in eGFP-expressing oocytes (Figure 2—figure supplement 1A). This TE increase was not accompanied by corresponding lengthening of poly(A) tails (Figure 2—figure supplement 1B), implying that tail-length changes did not cause these relative TE changes. To make comparisons of absolute TE changes, we repeated the ePAB-overexpression experiment but omitted rRNA depletion during sequencing library construction, thereby allowing us to normalize TE using mitochondrial mRNAs (Iwasaki et al., 2016), which were otherwise depleted by Illumina Ribo-Zero kits (Figure 2—figure supplement 1C). In this experiment, we also injected oocytes with a short-tailed Nluc mRNA reporter and a long-tailed Fluc mRNA reporter and monitored their absolute TE changes together with those of endogenous mRNAs. Most endogenous mRNAs had greater absolute TE in ePAB-overexpressing oocytes compared to eGFP-expressing control oocytes (Figure 2B). This result was consistent with 35S metabolic-labeling experiments showing that overexpression of PABPC1 but not eGFP significantly increased global protein synthesis in oocytes (Figure 2C–D). Moreover, the magnitude of the TE increase conferred by ePAB-overexpression negatively correlated with tail length, which showed that translation of short-tailed mRNAs improved substantially more than that of long-tailed mRNAs (Figure 2E), as observed for our co-injected reporters. Indeed, adding ePAB had essentially no overall effect on TE of endogenous mRNAs with the longest tails (median TE fold change = 1.06 for the 54 mRNAs with median tail lengths > 80 nt), as observed for our long-tailed reporters. The preferential improvement of TEs for short-tailed mRNAs led to not only an overall shift in TE but also narrowing of the TE distribution (Figure 2F) to more closely resemble the distributions observed in cells in which poly(A)-tail length and TE are not coupled (Subtelny et al., 2014). These results supported the hypothesis that increasing PABPC in oocytes increases the opportunity for short-tailed mRNAs to bind a PABPC molecule, thereby promoting translation. Overall, the results of our global analyses of mRNAs in frog oocytes agreed with those of reporter assays, thereby extending to endogenous mRNAs support for the conclusion that limiting PABPC plays a critical role in conferring strong coupling between poly(A)-tail length and TE. Intragenic analyses further demonstrate the importance of limiting PABPC for establishing coupling between tail length and TE Our global analysis examining the relationship between poly(A)-tail length and TE of endogenous mRNAs in oocytes differed from our reporter assays in that the comparison was made between mRNAs of different genes, which can be confounded by features other than tail length that vary between these mRNAs. To overcome this issue, we developed a high-throughput method for comparing effects on different tail-length isoforms from each gene. This approach for intragenic analyses, called PAL-TRAP (Poly(A) tail-Length profiling following Translating Ribosome Affinity Purification), resembled other TRAP approaches in that ribosomes were sparsely tagged such that their immunoprecipitation (IP) preferentially isolated mRNA isoforms associated with more ribosomes, which were inferred to be more highly translated (Chen and Dickman, 2017; Heiman et al., 2008). In a system in which poly(A)-tail length and TE were coupled, longer-tail mRNAs were expected to be associated with more ribosomes and therefore enriched in the eluate (Figure 3A), whereas in an uncoupled system, longer-tail mRNAs were not expected to be enriched in the eluate. Figure 3 with 1 supplement see all Download asset Open asset Limiting PABPC is required for intragenic coupling between poly(A)-tail length and TE. (A) The PAL-TRAP method for measuring intragenic effects of tail length on TE. Ribosomes are sparsely tagged (red stars) so that highly translated mRNAs are more likely to contain tagged ribosomes and thus be enriched in the immunoprecipitation (IP) eluate. Tail lengths of both input and eluate mRNAs were measured and compared for mRNAs of each gene. The depicted enrichment of long-tailed isoforms in the eluate indicates that poly(A)-tail length and TE are coupled, whereas no enrichment would indicate otherwise. (B) The experimental scheme of PAL-TRAP in frog oocytes. See Figure 3—figure supplement 1B–C for results from pulldowns using the eGFP-HA control. (C) Effect of overexpressing ePAB on coupling between tail length and TE, as detected after pooling PAL-TRAP results for mRNAs from different genes. Plotted are cumulative distributions of poly(A)-tail lengths in the PAL-TRAP input and eluate obtained after expressing either eGFP (left) or ePAB (right) in oocytes (injecting 4 fmol mRNA per oocyte). Median values are indicated (dashed lines) and listed in parentheses. (D) The effect of overexpressing ePAB on intragenic coupling between tail length and TE. Plotted for each mRNA isoform is the median poly(A)-tail length of mRNAs in the PAL-TRAP eluate compared to that in the input. Shown are results for mRNAs from oocytes either expressing eGFP or overexpressing ePAB (left and right, respectively). Each point represents an mRNA isoform with a unique 3′ end represented by ≥100 poly(A) tags in both input and eluate. Also indicated are results for (1) an Rluc reporter mRNA possessing a variable-length tail (reporter), which was co-injected with mRNAs expressing either eGFP or ePAB, (2) an mRNA with a variable-length tail, which was spiked into the lysate immediately before IP (spike-in), and (3) synthetic RNAs with defined tail lengths added to samples prior to library preparation (tail standards). Points for eight standards with longer tails fell outside the plot areas, as did a point representing the mRNA 3′-end from one gene (uqcrb.S) in the ePAB sample. (E) Summary of differences in median tail lengths observed between the eluate and the input of mRNA isoforms shown in (D). Box and whiskers indicate the 10th, 25th, 50th, 75th, and 90th percentiles. (F) Effect of overexpressing ePAB on intragenic tail-length distributions in frog oocytes. Shown are tail-length distributions of the reporter Rluc (left), an endogenous oocyte mRNA btg4.S (middle), and the spike-in mRNA Nluc (right) in eGFP-expressing (top) or ePAB-overexpressing (bottom) oocytes. Median tail-length values are indicated (vertical lines) and listed in parentheses. Tail-length measurements in this figure were obtained using PAL-seq v3. To implement PAL-TRAP, we first injected stage VI frog oocytes with an mRNA encoding C-terminal HA-tagged RPL3 (Chen and Dickman, 2017) and allowed time for RPL3-HA protein expression and incorporation int