Genetic Characterization of smg-8 Mutants Reveals No Role in C. elegans Nonsense Mediated Decay

The nonsense mediated decay (NMD) pathway degrades mRNAs bearing premature translation termination codons. In mammals, SMG-8 has been implicated in the NMD pathway, in part by its association with SMG-1 kinase. Here we use four independent assays to show that C. elegans smg-8 is not required to degrade nonsense-containing mRNAs. We examine the genetic requirement for smg-8 to destabilize the endogenous, natural NMD targets produced by alternative splicing of rpl-7a and rpl-12. We test smg-8 for degradation of the endogenous, NMD target generated by unc-54(r293), which lacks a normal polyadenylation site. We probe the effect of smg-8 on the exogenous NMD target produced by myo-3::GFP, which carries a long 3′ untranslated region that destabilizes mRNAs. None of these known NMD targets is influenced by smg-8 mutations. In addition, smg-8 animals lack classical Smg mutant phenotypes such as a reduced brood size or abnormal vulva. We conclude that smg-8 is unlikely to encode a component critical for NMD.


Introduction
The nonsense mediated decay (NMD) pathway is an evolutionarily conserved mRNA surveillance mechanism that recognizes and degrades transcripts bearing premature translation termination codons [1,2]. In C. elegans, mRNAs that have acquired a nonsense mutation or an extended 39 untranslated region (UTR) are targeted for NMD [3]. In addition, physiological transcripts that carry an early stop codon are substrates for NMD [4,5,6]. The central players of the NMD pathway were discovered in S. cerevisiae [7] and C. elegans [1,3,8,9] and comprise SMG-2/Upf1, SMG-3/ Upf2 and SMG-4/Upf3. In worms, Drosophila and mammals, the NMD core components are modulated by additional SMG factors. Recent screens have also identified and confirmed new candidate NMD proteins in C. elegans, H. sapiens and D. rerio [10,11,12].
Elegant studies in several organisms have suggested a model in which NMD reflects a competition between SMG-2/UPF1 and Poly(A) Binding Protein (PABP) for ribosome-associated translation release factors. In mammals, the locations of mRNA splice junctions are marked by exon-junction complexes (EJC) during mRNA processing [1]. If the ribosome encounters a premature termination codon (PTC) during the pioneer round of translation, the SURF complex, (consisting of SMG-1, UPF1 and the release factors eRF1 and eRF3) interacts with the EJC, triggers UPF1 phosphorylation by SMG-1 and initiates NMD [13]. In C. elegans, EJC components are dispensable for NMD, and PTCs are instead distinguished from normal stop codons by the size of the 39UTR [10,14]. Once the PTC has been recognized, target mRNAs are destroyed by either the SMG-6 endonuclease (mammals, Drosophila and probably worms) [15,16], and/or a SMG-5/SMG-7-dependent exonuclease (mammals, yeast and probably worms) [17,18].
Recently, Yamashita and colleagues used immunoprecipation of HeLa cell lysates to identify proteins that interact with SMG-1, a phosphatidylinositol kinase-related protein kinase [19,20]. SMG-1 bound two novel, conserved proteins, FLJ23205 and FJL12886, which were renamed SMG-8 and SMG-9 [21]. This pair bound strongly to each other, and modified SMG-1 kinase activity in vitro. Inactivation of SMG-8 and SMG-9 lead to a partial stabilization of ß-globin mRNAs in mammalian cell culture, suggesting they might play a role in NMD. The authors also used RNA interference (RNAi) to inactivate C. elegans smg-8 and smg-9, and concluded that smg-8, but not smg-9, contributed to NMD in worms [21].
Here we examine C. elegans smg-8 using a newly generated mutant allele. smg-8(tm2937) contains a 272 bp deletion and a 1 bp insertion within smg-8 ( Figure 1A). This deletion encompasses 22 bp upstream of the start site, the initiator ATG and the first two exons. Using animals homozygous for this allele, we employed four approaches to investigate a possible role for smg-8 in the NMD pathway. Our findings suggest that smg-8 is unlikely to be a key component for NMD in C. elegans.

smg-8 Mutants do not Exhibit Phenotypes Associated with NMD Mutants
For our first assay, we examined two of the classical phenotypes associated with smg genes. Hodgkin and colleagues reported a reduced brood size of 174 (range 147-211) for smg-1, which was lower than the mean brood size of wild-type animals (327, range 270-373) [9]. We observed a mean brood size of 301 (range 242-367, n = 10) for smg-8(2937), similar to the mean brood size of our wild-type strain (279, range 211-328, n = 10) ( Table 1). In addition, animals bearing a mutation in a canonical smg gene have a protruding vulva due to morphological defects [9]. However, vulvae appeared normal in smg-8(tm2937) worms ( Figure 1B). These data suggest that smg-8(tm2937) animals lack two overt phenotypes associated with canonical smg mutants. smg-8 does not Show an NMD Phenotype for the Native NMD Target rpl-7a and rpl-12 For our second assay, we examined transcripts for two ribosomal proteins rpl-7a and rpl-12, which are natural NMD targets [4]. These genes each generate two alternatively spliced mRNAs, one of which contains a premature termination codon (PTC; Figure 2A). When NMD is active, the longer isoform containing the PTC is degraded and only the shorter isoform accumulates. When the NMD pathway is compromised, the isoform containing the PTC is stabilized, and both mRNA isoforms accumulate ( Figure 2A). Using RT-PCR primers that flank the PTC, it is possible to distinguish between the two transcripts [4]. smg-8 shows no NMD phenotype by this assay. When a known component of the NMD pathway, such as smg-1, smg-2 or smg-3, is mutated, the mRNA isoform containing the PTC is stabilized, generating a robust upper band ( Figure 2B). In contrast, for smg-8(tm2937), a very faint upper band was observed, comparable to that of the wild-type strain ( Figure 2B). To extend this result, we inactivated smg-8 and also smg-9 using RNAi, which reduced smg mRNA levels at least five-fold ( Figure 2G) and similar to a No Reverse Transcriptase negative control ( Figure 2H). The results were again negative for NMD ( Figure 2C). RNAi is not always robust; therefore we repeated the assay using the strain eri-6/7(tm1917), which enhances RNAi [22,23], and once more observed no NMD phenotype ( Figure 2D). Finally, to exclude the possibility that the tm2937 allele was hypomorphic, we treated smg-8(tm2937) mutants with smg-8 RNAi or smg-9 RNAi, but again we observed no NMD phenotype ( Figure 2E). Similar results were observed for rpl-12 ( Figure 3).
To quantify these data, we used RT-qPCR to measure the increase of the PTC containing isoform. We observed a fold enrichment of 15 and 38 in smg-1 and smg-3 mutants respectively, compared to the wild-type ( Figure 2F). The PTC containing isoform in smg-8 mutants remained similar to the wild-type (0.7 fold enrichment). A virtually identical result was obtained with smg-8 RNAi treatment of smg-8 or eri-6/7 mutant worms (0.7 and 0.4 fold enrichment respectively) ( Figure 2F). Together, these data reveal that inactivation of smg-8 fails to stabilize two natural NMD targets, rpl-7a and rpl-12.    [26]. RNA was analyzed as in (B). (D) An enhanced RNAi mutant strain eri-6/7 [22,23] was used and RNAi conducted as in (C). RNA was analyzed as in (B). (E) As in D, using the smg-8(tm2937) mutant strain. (F) RT-qPCR using primers flanking the PTC-containing isoform of rpl-7a, mRNA levels were calculated using the delta-delta-CT method, relative to the control gene pmp-3 [27]. Fold enrichment of the PTC mRNA was normalized to 1 for wild-type. The smg-1 and smg-3 mutants show an enrichment of 15 and 38 fold, respectively. In contrast, in smg-8 mutants, the accumulation of the PTC containing isoform is similar to wild-type (0.7 fold enrichment). smg-8 and eri-6/7 mutant worms treated with smg-8 RNAi

smg-8 is not Required for Endogenous NMD in C. elegans
As a third test for NMD, we examined an endogenous target of the NMD pathway: unc-54(r293) [3,9]. The unc-54 gene generates a muscle myosin heavy chain (MHC) in C. elegans, and the unc-54(r293) allele contains a 256 bp deletion that removes the normal 39 cleavage/polyadenylation site and most of the 39UTR [3] ( Figure 4A). This deletion causes the production of a long unc-54 mRNA transcript that terminates at a cryptic poly(A) site and renders unc-54 an NMD target. Without the MHC, unc-54 mutant worms are paralyzed ( Figure 4B). In the absence of NMD components such as smg-1, the unc-54(r293) mRNA is stabilized, wild-type protein is produced and the Unc phenotype is suppressed [3] (Figure 4B). We generated unc-54(r293); smg-8(tm2937) double mutants and observed no suppression of the paralysis phenotype ( Figure 4B), consistent with our hypothesis that smg-8 is not required for NMD in C. elegans.
In summary, we tested smg-8 for a role in NMD using four different assays: i) anatomical phenotype and brood size, ii) accumulation of natural NMD targets rpl-7a and rpl-12, iii) rescue of the paralysis phenotype caused by the endogenous NMD target unc-54(r293) and iv) GFP accumulation of the exogenous NMD target reporter myo-3::GFP. The discrepancy between our study and the results presented by Yamashita and colleagues is due in part to the use of an allele (our study) vs. RNAi [21]. In addition, we note that the effect of smg-8 inactivation on NMD in C. elegans was not robust in the Yamashita study [21]. In all of our assays, smg-8 mutants resembled wild-type worms and differed from classical smg mutants. We detected no accumulation of mRNA or protein in smg-8 mutants, even when the smg-8 mutation was combined with RNAi. We suggest that smg-8 in C. elegans is a novel, conserved gene whose function remains to be elucidated.

RNA Interference
HT115 bacteria expressing double stranded RNA targeting smg-1, smg-8 or smg-9 grown for ,8 hours at 37uC were plated using 1 mM IPTG (Sigma) and 50 mg/ml of Carbenicillin (Sigma). RNAi clones were derived from the Ahringer library [26] and verified by sequencing. Five wild-type, smg-8(tm2937) or eri-6/ 7(tm1917) worms were transferred at the L4 stage to RNAi plates and allowed to lay embryos for one day. The progeny was collected ,48 hours later, when most worms had grown at least to the L4 stage, by rinsing with water and frozen at -80uC for show 0.7 and 0.4 fold enrichment, respectively. (G) RT-qPCR to quantify smg-8 RNA. mRNA levels were calculated using the delta-delta-CT method, relative to the control gene pmp-3 [27]. Fold enrichment was normalized to 1 for wild-type. smg-8 and eri-6/7 worms treated with smg-8 RNAi show 0.3 and 0.26 fold enrichment, respectively. (H) As in (G) for wild-type animals and a negative control that lacked Reverse Transcriptase (No RT). Fold enrichment was normalized to 1 for wild-type. No RT control shows 0.3 fold enrichment. doi:10.1371/journal.pone.0049490.g002

RNA Extraction
For total RNA extraction, glass beads (Sigma) and 1 ml of Trizol Reagent (GibcoBRL) were added to frozen worm pellets. Pellets were lysed by vortex followed by chloroform extraction. RNA was precipitated with isopropanol and washed with 70% ethanol. Resuspended RNA was extracted with phenol:chloroform and precipitated with ethanol. A first-strand reaction kit (NEB) was used to perform the reverse transcriptase reaction, following the manufacturers protocol.

RT-PCR of rpl
Amplification of rpl-7a from the cDNA was performed as described in [4], and PCR product was analyzed in a 1% agarose gel. Primers used were rpl-7a-fw GACATCCAGCCAAA-GAAGGA and rpl-7a-rv AACGGTGTTTGGTCTCTTGG.

RT-qPCR
For smg-8 and rpl-7a, mRNA levels were calculated using the delta-delta-CT method, relative to the control gene pmp-3 [27]. Control mRNA was normalized to 1 and the mRNA levels are shown as relative fold change. Primers for smg-8 were smg8-fw-4348 GCTGCCAATATTTCCATCGT and smg8-rv-5165 TGACCACGGGAACATTCATA.

Brood Size
10 worms at the L4 stage were picked into individual plates at 20uC. Their progeny was counted everyday until no more progeny was generated. The number of progeny per plate was averaged (n = 10). Two worms were placed in the middle of the bacterial lawn and allowed to crawl for 45 minutes. Wild-type worms that crawl (top left) leave tracks in the lawn whereas unc-54(r293) mutants cannot move well (top right). smg-1(r861) suppresses unc-54(r293) mRNA degradation and restores movement [3] (bottom left). In contrast, smg-8 does not suppress the paralysis phenotype (bottom right). Arrowheads indicate the tracks left by paralyzed worms. doi:10.1371/journal.pone.0049490.g004 Figure 5. smg-8 does not restore expression of myo-3::GFP, an exogenous NMD target. (A) Schematic representation of the exogenous NMD GFP reporter, driven by the myo-3 promoter, which is destabilized by an amino acid sequence that marks a protein for degradation (degron), and a long 39UTR [24]. (B) smg-8 and control mutations were introduced into CL724 (myo-3::GFP) worms. The double combinations were then inspected under a fluorescent microscope. smg-1 and smg-3 mutants express high levels of GFP. In contrast, smg-8 animals photographed under the same conditions show only a slight accumulation of GFP, similar to the wild type. doi:10.1371/journal.pone.0049490.g005