Apoptosis Triggers Specific, Rapid, and Global mRNA Decay with 3′ Uridylated Intermediates Degraded by DIS3L2

mRNA Decay Occurs Early in Apoptosis and Requires MOMP

We next determined the kinetics of the mRNA decline relative to hallmarks of apoptosis. Jurkat cells treated with ActD ± αFas were harvested hourly over 4 hr and analyzed for housekeeping gene mRNA stability by qRT-PCR, cleavage of PARP-1 and eiF4G by immunoblot, phosphorylation of eiF2α by immunoblot, caspase activation by a luminescent assay, annexin V binding, and DNA fragmentation (Figures 3A–3E). ACTB and GAPDH mRNAs began to decline within 1 hr after addition of αFas and leveled off by 3 hr, whereas 7SL ncRNA was stable (Figure 3A). Nonapoptotic oxidative stress caused by arsenite (+ActD) and caspase-independent programmed cell death induced by the killer protease granzyme A did not affect mRNA levels (Figure 3A and data not shown). Caspase activation and PARP-1 cleavage began at 1 hr and were complete by 3 hr (Figures 3B and 3C). eIF2α phosphorylation and eIF4G cleavage were first detected after 2 hr (Figure 3B). Similarly, phosphatidylserine externalization and DNA fragmentation were not detected until 2 hr after αFas was added (Figures 3D and 3E). Mitochondrial depolarization began within 1 hr after αFas treatment (Figure 3F). Thus, mRNA decay occurred early in cell death coincidently with caspase activation and mitochondrial disruption.

Open in a separate window

Figure 3

mRNA Decay Begins Early in Cell Death and Requires MOMP

(A–E) Jurkat cells were treated with ActD ± αFas or arsenite for 4 hr and harvested hourly for qRT-PCR (A), immunoblot (B), caspase activity by luminescent assay (AU, arbitrary units) (C), annexin V staining by flow cytometry (D), and DNA fragmentation by electrophoresis (E). Asterisks in (B) indicate caspase cleavage products. mRNA decay was concurrent with caspase activation, but preceded DNA cleavage and phosphatidylserine externalization. Error bars represent the SEM of at least three independent experiments. Results in (B) and (E) are representative of multiple experiments, and the experiment shown in (C) was performed once.

(F) Jurkat cells were treated for 1 hr with ActD ± αFas and stained with DiIC₁(5) to measure mitochondrial depolarization. The transmembrane potential began dissipating within 1 hr after αFas treatment. Results are representative of at least three independent experiments.

(G and H) HeLa-BCL2 and HeLa-puro were treated for 6 hr with STS ± zVAD.

(G) Fractionated cells were analyzed by immunoblot (top) for cytochrome c release. Effector caspase activity was measured by luminescent reporter assay (bottom). Error bars represent the SEM of separate wells from one experiment.

(H) RNAs were assayed by qRT-PCR. Error bars represent the SEM of at least three independent experiments (*p < 0.05). Results were normalized to untreated cells. BCL2 overexpression inhibited mRNA decay to a greater extent than did zVAD. Survival (assayed by CytoTox-Glo assay performed after 6 hr of STS treatment followed by a 24 hr recovery period) was normalized to untreated cells. See also Figure S3.

Because MOMP is an early step in apoptosis (Aldridge et al., 2011), we asked whether MOMP is required for mRNA decay. Stressors such as STS activate cell death by initiating MOMP upstream of caspase activity, whereas αFas activates caspase 8 to trigger MOMP. Thus, zVAD blocks MOMP induced by αFas, but not by STS (Arnoult et al., 2003; Holler et al., 2000). We compared HeLa cells stably expressing BCL2 (HeLa-BCL2) or an empty vector (HeLa-puro). In agreement with previous studies (Arnoult et al., 2003), BCL2 overexpression blocked the release of cytochrome c and effector caspase activation caused by STS (Figure 3G). In contrast, zVAD blocked STS-triggered caspase activation, but did not block cytochrome c release or cell death. mRNA decay was rescued by BCL2 expression, but only partially by zVAD (Figure 3H). Knockdown of BAX and BAK also rescued mRNA decay to a greater extent than did zVAD (Figure S3). Thus, mRNA decay occurs early in apoptosis and depends on MOMP.

Depletion of Cellular mRNA Promotes Apoptosis

We hypothesized that the global mRNA depletion we observed would contribute to apoptosis. To test the importance of maintaining the pool of mRNAs for cell viability, we treated HeLa cells with the Pol II inhibitor α-amanitin and examined how changes in RNA levels over time (assessed by qRT-PCR; Figure 4A) correlated with cell death (assessed by annexin V staining and flow cytometry; Figure 4B). The decline in mRNA levels triggered by transcription inhibition correlated closely with (r² = 0.98; Figure 4B). We also found that transcription inhibition significantly enhanced the cytotoxic effect of STS, TRAIL, and etoposide (Figure 4C). Thus, a global loss of mRNA is incompatible with continued cell survival.

Open in a separate window

Figure 4

Cell Viability Correlates with mRNA Levels

(A and B) HeLa cells were treated with the transcriptional inhibitor α-amanitin for 20 hr and harvested at regular intervals, and RNA levels were measured by qRT-PCR (A). Cell death was measured by annexin V staining and flow cytometry, and then plotted relative to the average level of the four mRNAs analyzed (B). Labels on data points represent the time of α-amanitin treatment in hours. Cell death increased as mRNA levels dropped. Results are mean ± SEM from triplicate experiments.

(C) HeLa cells were treated with or without α-amanitin and STS, TRAIL, etoposide, or nothing (UNT) for 8 hr. Cell death was measured by annexin V staining and flow cytometry. In all cases, transcription inhibition significantly enhanced the progression to apoptosis. Error bars represent the SEM of at least three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001.

mRNA Decay Intermediates Contain Nontemplated 3′ Uridylates

To begin to understand the mechanism of mRNA decay during cell death, we searched for decay intermediates using circular rapid amplification of cDNA ends (cRACE) (Rissland and Norbury, 2009) to ligate and amplify the 5′ and 3′ ends of ACTB mRNA (Figure 5A). cRACE was performed with and without tobacco acid pyrophosphatase (TAP) pre-treatment to distinguish between capped and decapped intermediates (only decapped mRNAs can be ligated without TAP) in HeLa cells treated with STS ± zVAD for 6 hr. Using a forward primer directed to the ACTB 3′ UTR near its polyadenylation site, we amplified residual full-length ACTB mRNAs from both untreated and STS-treated apoptotic cells, but only using TAP-treated RNA (Figures S4A and S4B). There were no visible decapped decay intermediates. The cloned and sequenced cRACE products contained poly(A) tails of similar lengths independently of the treatment used (Figure S4C). However, when RNA was amplified with a PCR primer located near the stop codon in the ACTB open reading frame (ORF), we detected decay intermediates in apoptotic cell RNA even in cells treated with zVAD (Figure 5B). These intermediates were also amplified without TAP pre-treatment, indicating that at least some of the decay intermediates were decapped. When the cRACE products from STS-treated cells were cloned, the junctions that mapped exactly to the ACTB mRNA had two notable features (Figures 5C and 5D). First, their 3′ termini were within 50 nt of the ACTB stop codon, suggesting stalled decay near the ORF. Second, most junctions captured from TAP-treated RNA had 5′ termini that mapped exactly to the ACTB transcription start site (TSS), whereas all of the 5′ termini captured from TAP-untreated RNA began after the TSS. These results suggest that decay proceeds from 3′ to 5′ on deadenylated mRNAs before decapping occurs. Surprisingly, some clones had nontemplated residues between the 3′ and 5′ termini. These added bases were rich in uridylates (Figure 5E). Similar results were obtained by cRACE and sequencing of the decay intermediates of another mRNA (EEF1A; Figures S4D–S4G). Thus, nontemplated 3′ uridylates are added in the region near the stop codon during apoptotic mRNA decay. To confirm that 3′ uridylated ACTB decay intermediates were produced during apoptosis, we performed RT-PCR on RNA isolated from HeLa cells treated or not with STS using an A₁₂-adaptor for RT to prime poly(U) sequences, and nested primers targeting the ACTB ORF and the adaptor primer (Figure 5F). Novel decay products reproducibly appeared only in STS-treated cells (Figure 5G). Cloning and sequencing confirmed that these were similar to the cRACE products (data not shown).

Open in a separate window

Figure 5

3′ Uridylation of Apoptotic Decay Intermediates

(A) Schematic of the cRACE method used to identify ACTB mRNA decay intermediates. cRACE was performed on RNA from HeLa cells treated for 6 hr with STS ± zVAD.

(B–E) ACTB decay products were amplified using an ORF-targeted primer and a reverse primer in the 5′ UTR, and analyzed by gel electrophoresis (B). The ACTB fragment was gel purified, cloned, and sequenced from reactions that were treated (D) or not (C) with TAP prior to cRACE. Line diagrams (C and D) depict all ORF-primer amplified cRACE clones that mapped to the ACTB transcript, excluding clones with nontemplated 3′ extensions. The decay intermediates were truncated just 3′ of the stop codon. Decapped decay intermediates (−TAP) had shortened 5′ ends, whereas most clones derived from TAP-treated RNA had a complete 5′ sequence.

(E) Some apoptotic ACTB mRNA fragments had nontemplated 3′ tails. Shown are nontemplated RNAs obtained from TAP-treated RNA. Templated residues matching the ACTB sequence are depicted as a line, and uridylate-rich nontemplated tails are indicated. All clones in (C)–(E) were derived from one experiment.

(F and G) Amplification of poly(U)-tailed RNAs using an A₁₂-adaptor reverse RT primer.

(F) Schematic of the RT-PCR method. After RT, nested PCR was used to amplify uridylated decay products, with forward primers targeting the ACTB ORF and reverse primers targeting the adaptor sequence.

(G) Gel electrophoresis of RT-PCR products. Arrow indicates new uridylated ACTB mRNA decay fragments in apoptotic cells. The asterisk denotes products primed from oligouridylate tracts elsewhere in the ACTB 3′UTR. Results are representative of at least three independent experiments. See also Figure S4.

ZCCHC6 and ZCCHC11 Contribute to mRNA Decay and Apoptosis

We hypothesized that one or more TUTases add these non-templated uridylate-rich tails. We focused on ZCCHC6 and ZCCHC11 because they uridylate other cytosolic RNAs (Schmidt et al., 2011; Thornton et al., 2012, 2014). We transfected HeLa cells with a control siRNA or siRNAs targeting ZCCHC6 and/or ZCCHC11, or the TUTase PAPD7 (Figure 6A). Knockdown of ZCCHC6, ZCCHC11, or both reduced apoptotic 3′ uridylation of the ACTB mRNA (Figure 6B) and partially rescued mRNA levels after STS treatment (Figure 6C), whereas PAPD7 knock-down had no effect. ZCCHC6 and ZCCHC11 siRNAs, alone and in combination, reduced annexin V staining (Figure 6D) and caspase 3 cleavage and activation (Figures 6E and 6F) in response to STS. Again, PAPD7 siRNAs had no effect. Collectively, these results suggest that ZCCHC6 and ZCCHC11 collaborate to uridylate mRNAs during cell death, promoting mRNA decay and apoptosis.

Open in a separate window

Figure 6

ZCCHC6/ZCCHC11 Knockdown Inhibits Cell Death and mRNA Decay

(A) HeLa cells were transfected with control (CTL), ZCCHC6, and/or ZCCHC11 or PAPD7 siRNAs and harvested 72 hr later. RNA was then analyzed by qRT-PCR to assess knockdown.

(B) Accumulation of uridylated decay intermediates was assessed by RT-PCR with an A₁₂-adaptor primer (as in Figure 5F). Uridylated intermediates that arose after STS treatment were reduced after knockdown of ZCCHC6 and/or ZCCHC11. The arrowhead denotes new products and the asterisk denotes products primed from oligouridylate tracts in the ACTB 3′ UTR. Results are representative of three independent experiments.

(C–F) After knockdown and STS treatment, mRNA levels were assayed by qRT-PCR (C). Cells were analyzed for cell death by annexin V staining and flow cytometry (D). Caspase activation was assessed by immunoblot for caspase 3 (E) and a luminescent caspase activity assay (F). ZCCHC6 and/or ZCCHC11 knockdown partially restored mRNA levels, rescued cell death, and reduced caspase 3 cleavage and activation. Error bars represent the SEM of at least three independent experiments.

**p < 0.01; ***p < 0.001, relative to CTL knockdown.

DIS3L2 Knockdown Inhibits mRNA Decay and Cell Death

Because DIS3L2 targets 3′ uridylated pre-miRNAs (Chang et al., 2013; Ustianenko et al., 2013), we hypothesized that DIS3L2 digests the uridylated mRNA decay products we observed in cell death. We knocked down DIS3L1 (a ribonuclease in the cytosolic exosome, as control) and DIS3L2 in HeLa cells (Figure 7A) and used RT-PCR with an A₁₂-adaptor to amplify 3′ uridylated ACTB mRNA intermediates (Figure 7B). Novel uridylated ACTB products were detected in nonapoptotic cells transfected with DIS3L2 siRNA, but not in those transfected with DIS3L1 or control siRNA. The appearance of uridylated mRNA decay intermediates in living cells after DIS3L2 knockdown suggests that DIS3L2 plays a role in basal mRNA decay. After STS treatment, the uridylated ACTB mRNA intermediates greatly increased with DIS3L2 knockdown, and increased somewhat with DIS3L1 knockdown. We confirmed the increase in uridylated ACTB mRNA intermediates with DIS3L2 knockdown by cloning and sequencing cRACE products (Figures 7C and S5A–S5G). The supplemental data in Figures S5A–S5G were obtained using another DIS3L2 siRNA to verify that the effect of knocking down DIS3L2 was not due to off-target effects. Knockdown of DIS3L2, but not DIS3L1, also significantly reduced STS-mediated mRNA decay of three housekeeping genes (ACTB, GAPDH, and SDHA; Figures 7D and S5E). Knockdown of DIS3L2, but not DIS3L1, increased the mRNA half-life in STS-treated HeLa cells (Figures S5H and S5I). To verify that the role of DIS3L2 in mRNA decay was not limited to STS-treated HeLa cells, we also examined DIS3L1 and DIS3L2 knockdown in HCT116 cells treated with TRAIL (Figures S5J–S5N). DIS3L2 knockdown specifically reduced TRAIL-mediated decay of four housekeeping genes, but knockdown of DIS3L1 only affected one of them (and to a lesser extent; Figure S5K). Thus, DIS3L2 is a mediator of mRNA decay in apoptosis and may also play a role in basal mRNA decay.

Open in a separate window

Figure 7

DIS3L2 Knockdown in HeLa Cells Inhibits Cell Death and mRNA Decay

(A) HeLa cells transfected with control (CTL), DIS3L1, or DIS3L2 siRNAs were harvested 72 hr later for immunoblot, confirming robust knockdown of DIS3L1 and DIS3L2 proteins.

(B) Accumulation of uridylated decay intermediates was assessed by RT-PCR with an A₁₂-adaptor primer (as in Figure 5F). Uridylated intermediates accumulated in STS-treated cells regardless of the siRNA transfected, but accumulated to a greater extent after DIS3L2 knockdown. The arrowhead denotes new products and the asterisk denotes products primed from oligouridylate tracts in the ACTB 3′ UTR. Results are representative of at least three independent experiments.

(C) Increased uridylation was validated by cloning cRACE products (as in Figure 5). For knockdown in this experiment and Figures S5A–S5G, we used an independent set of siRNAs to verify that the results were not due to off-target effects of gene knockdown. Products with two or more nontemplated uridylates were more abundant in STS-treated cells after DIS3L2 knockdown, but the proportion of products with any nontemplated residues on the 3′ end was similar between the CTL and DIS3L2 knockdown samples.

(D–F) After knockdown and STS treatment, mRNA levels were assayed by qRT-PCR (D). Cytochrome c release was measured by cellular fractionation followed by immunoblot (E) and caspase 3 cleavage was measured by immunoblot (F). β-Actin (ACTB) was probed as a loading control and VDAC1 was probed to verify fractionation. DIS3L2 knockdown partially restored mRNA levels and reduced cytochrome c release and caspase 3 cleavage.

(G) At 72 hr after transfection with the indicated siRNAs, cells were incubated with STS or TRAIL for 3 hr or etoposide or tunicamycin for 8 hr. Cells were analyzed by annexin V staining and flow cytometry. DIS3L2 knockdown significantly reduced phosphatidylserine externalization and cell survival in response to all treatments.

(H) Cells were transfected with a plasmid expressing GFP or GFP-tagged DIS3L2 and treated 72 hr later, as in (G). Cells were analyzed by annexin V staining and flow cytometry. Ectopic DIS3L2 expression significantly increased phosphatidylserine externalization in response to STS, TRAIL, and tunicamycin.

(I) Cells were replated at low density 48 hr after knockdown and were treated 24 hr later with TRAIL or two doses of STS for 3 hr. After the drugs were removed, the cells were cultured and surviving colonies were counted. DIS3L2 knockdown significantly enhanced clonogenic survival in response to all treatments. UNT, untreated.

(J and K) siRNA-transfected cells were treated 2 days later with STS for 4 hr and protein incorporation was measured by ³⁵S incorporation followed by PAGE and autoradiography. A representative experiment (J) and the results of three independent experiments (K) are shown. Significance relative to CTL siRNA is shown.

(L) siRNA-transfected cells were treated 3 days later with STS for 4 hr and proteins were analyzed by immunoblot.

(M) MCL-1 protein levels declined significantly after 4 hr of STS treatment. The decline was partially rescued by DIS3L2, but not DIS3L1, knockdown.

(N) siRNA-transfected cells were treated 3 days later with STS for 3 hr and cell death was measured by annexin V staining and flow cytometry. BCL2L3 knockdown abolished the protective effect of DIS3L2 knockdown.

Error bars represent the SEM of at least three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001. See also Figure S5.

Next, we examined whether mRNA decay contributes to apoptosis. We compared cytochrome c release, caspase activation, annexin V staining, and clonogenic survival of HeLa or HCT116 cells transfected with control, DIS3L1, or DIS3L2 siRNAs and then treated with a variety of apoptotic stimuli. STS-induced cytochrome c release in HeLa was reduced when DIS3L2, but not DIS3L1, was knocked down (Figure 7E). Caspase 3 cleavage was detected 4 hr after STS treatment in cells treated with DIS3L1 or control siRNA, but was rescued by DIS3L2 knockdown (Figure 7F). DIS3L2 knockdown also reduced annexin V staining in HeLa cells after exposure to STS, etoposide, tunicamycin, and several doses of TRAIL (Figures 7G and S5O). Knockdown of DIS3L1 or DIS3, the homologous exonuclease in the nuclear exosome (data not shown), did not affect annexin V staining. Conversely, DIS3L2 overexpression in HeLa significantly increased annexin V staining in response to STS, TRAIL, and tunicamycin (Figure 7H). DIS3L2 knockdown also specifically and significantly enhanced the clonogenic survival of HeLa cells after TRAIL or STS treatment (Figure 7I). DIS3L2 knockdown reduced caspase activation in HeLa cells transfected with a different siRNA (Figures S5F and S5G), and caspase activation and cell death in HCT116 cells treated with TRAIL (Figures S5L–S5N). Taken together, these results show that mRNA decay, mediated by DIS3L2, strongly promotes apoptosis.

DIS3L2-Mediated mRNA Decay Inhibits De Novo Protein Synthesis during Apoptosis

It is well known that translation is arrested during apoptosis. This has been attributed to post-translational modifications of translation initiation factors. However, global mRNA decay could also have a profound effect on translation. To explore how mRNA decay and its inhibition by DIS3L2 knockdown affect translation during apoptosis, we measured global translation by incorporation of ³⁵S-methionine and ³⁵S-cysteine into HeLa cells after control, DIS3L1, or DIS3L2 knockdown, with and without STS treatment (Figures 7J and 7K). In the absence of STS, knock-down of DIS3L1 or DIS3L2 did not affect translation. Following STS treatment, translation declined by about 50% in control cells and cells knocked down for DIS3L1. However, DIS3L2 knockdown restored translation almost to the level observed in nonaptoptotic cells. Thus, global mRNA decay during apoptosis strongly inhibits new protein synthesis, and inhibition of translation depends on DIS3L2.

FISH

Adherent cells were grown in 24-well plates (Corning 3524) on 12-mm glass coverslips (VWR 89015-724) pre-treated with poly-L-lysine (Sigma P8920) for 5 min. Jurkat cells were washed, resuspended in PBS (Life Technologies 14190-144) with 0.5% BSA (Sigma A9647) and 2 mM EDTA (Life Technologies AM9260G), and then cytospun (Shandon Cytospin 3) for 4 min onto coverslips at 400 rpm before fixation. YT-Indy:721.221 immune conjugates were fixed first and then cytospun as described above. Cells were fixed for 10 min in 2% formaldehyde (Polysciences 18814) in 1× PBS and then permeabilized in pure methanol (Fisher BP1105-4) on dry ice for 10 min. Slides were washed three times for 5 min in 2× SSC. A control for each experiment was treated for 30 min in 0.1 M NaOH (Fisher SS255-1) in 2× SSC to hydrolyze all RNA. All coverslips were inverted onto a drop of hybridization buffer (10% dextran sulfate, 35% formamide, 0.3 M NaCl, 30 mM sodium citrate, 20 mM DTT) containing 1 μM 18S rRNA probe (Cy5-ACCAGACTTGCCCTCC) and 333 nM poly(A) (Cy3-dT₅₀) probe. Samples were placed in a humidified chamber and denatured for 5 min at 65°C, followed by 30 min at 45° C and at least 90 min at 42°C. The samples were washed three times for 5 min in 23 SSC at 37°C and stained with DAPI (Sigma D9542) in 2× SSC. Finally, the slides were mounted using polyvinyl alcohol (Sigma P-8136) aqueous mounting medium. Cells were imaged using an Axiovert 200M microscope (Pan Apochromat, 1.4 NA; Carl Zeiss) at 63×. Images were analyzed with SlideBook 4.2 (Intelligent Imaging Innovations). For quantification, cells were automatically identified based on the rRNA signal and background intensity was subtracted. All images shown are representative of at least three independent experiments.

cRACE

Total RNA (20 μg) was treated with the DNA-free kit (Life Technologies) per the manufacturer’s instructions. RNA was ethanol precipitated overnight and washed twice with 70% ethanol. RNA was resuspended in 11 μl of 1× TAP buffer (Epicenter {"type":"entrez-nucleotide","attrs":{"text":"T19050","term_id":"601093","term_text":"T19050"}}T19050). Half of this suspension was treated with 2 U of TAP for 1 hr at 37°C, and the other half was treated the same way except that TAP was not added. The reactions were then treated overnight at room temperature in a 100 μl volume with 10 U of T4 RNA ligase (NEB M0204S) with 1 mM ATP in 1× RNA ligase buffer. RNA was ethanol precipitated, washed, and resuspended in 20 μl dH₂O. Then 3 μl of this RNA was mixed with 1 μl of 10 mM deoxyribonucleotide triphosphates (dNTPs), 2 μl of the ACTB RT primer, and 7 μl of dH₂O. This was heated to 65°C for 5 min and cooled to 25°C. Then 7 μl of a master mix containing 4 μl 5× SuperScript III buffer, 1 μl 100 mM DTT, 1 μl SuperScript III (Life Technologies 18080-093) and 1 μl RNaseOUT (Life Technologies 10777-019) was added. The reaction was incubated at 55°C for 60 min and 70° C for 15 min before it was stored at 4° C. Then 2 μl of this reaction was mixed with 1 μl of each PCR primer (10 μM), 6 μl dH₂O, and 10 μl 2× Phusion polymerase mix (NEB M0531L), and cycled as follows: one cycle at 98° C for 30 s; 35 cycles at 98° C for 10 s, 60° C for 10 s, and 72° C for 10 s; and one cycle at 72° C for 3 min. PCR products were gel extracted and cloned using the Zero-Blunt PCR kit (Life Technologies K2750) according to the manufacturer’s instructions. The primers used are listed in Table S1. In the figures, line diagrams depict the mapped 5′ and 3′ ends; the sequence between the primers is inferred to be full-length ACTB mRNA.

RT-PCR with Adenylated Adaptor Primer to Detect Uridylated Decay Products

Total RNA (1 μg) was mixed with 3 μl of a 100 μM adaptor primer and 1 μl 10 mM dNTPs in a final volume of 13 μl. This was heated to 65° C for 5 min and cooled to 25° C. Then 7 μl of a master mix containing 4 μl 5× SuperScript III buffer, 1 μl 100 mM DTT, 1 μl SuperScript III, and 1 μl RNaseOUT was added. The reaction was incubated at 25° C for 10 min, 50° C for 30 min, and 70° C for 15 min before it was stored at 4° C. Then 2 μl of this reaction was mixed with 1 μl of each outer primer, 6 μl dH₂O, and 10 μl 2× Phusion polymerase mix, and cycled as follows: one cycle at 98° C for 30 s; 30 cycles at 98° C for 10 s, 60° C for 10 s, 72° C for 10 s; and one cycle at 72° C for 3 min. PCR was repeated using identical parameters with the inner primers. Products were cloned as above. The primers used are listed in Table S1.

Plasmids

GFP reporter constructs were kindly provided by B. Glaunsinger (Abernathy et al., 2014; Gaglia et al., 2012; Lee and Glaunsinger, 2009). GFP-DIS3L2 constructs were kindly provided by A. Dziembowski and M. Lubas (Lubas et al., 2013).

We thank Nancy Kedersha, Ashish Lal, and Aaron Deutsch for advice and technical assistance; Paul Anderson and Steve Buratowski for valuable advice; and Andrej Dziembowski and Britt Glaunsinger for plasmids. This work was supported by an NSF graduate research fellowship to M.P.T. and NIH grant AI045587 to J.L.