Research ArticleMOLECULAR BIOLOGY

AAA+ Proteins RUVBL1 and RUVBL2 Coordinate PIKK Activity and Function in Nonsense-Mediated mRNA Decay

See allHide authors and affiliations

Science Signaling  06 Apr 2010:
Vol. 3, Issue 116, pp. ra27
DOI: 10.1126/scisignal.2000468

Abstract

Phosphatidylinositol 3-kinase–related protein kinase (PIKK) family proteins play essential roles in DNA-based and RNA-based processes, such as the response to DNA damage, messenger RNA (mRNA) quality control, transcription, and translation, where they contribute to the maintenance of genome integrity and accurate gene expression. The adenosine triphosphatases associated with diverse cellular activities (AAA+) family proteins RuvB-like 1 (RUVBL1) and RUVBL2 are involved in various cellular processes, including transcription, RNA modification, DNA repair, and telomere maintenance. We show that RUVBL1 and RUVBL2 associate with each PIKK family member. We also show that RUVBL1 and RUVBL2 control PIKK abundance at least at the mRNA level. Knockdown of RUVBL1 or RUVBL2 decreased PIKK abundance and impaired PIKK-mediated signaling. Analysis of SMG-1, a PIKK family member involved in nonsense-mediated mRNA decay (NMD), revealed an essential role for RUVBL1 and RUVBL2 in NMD. RUVBL1 and RUVBL2 associated with SMG-1 and the messenger ribonucleoproteins in the cytoplasm and promoted the formation of mRNA surveillance complexes during NMD. Thus, RUVBL1 and RUVBL2 regulate PIKK functions on two different levels: They control the abundance of PIKKs, and they stimulate the formation of PIKK-containing molecular complexes, such as those involved in NMD.

Introduction

Phosphatidylinositol 3-kinase (PI3K)–related protein kinases (PIKKs) are unconventional serine–threonine protein kinases, and their catalytic domains are homologous to the catalytic domains of phosphatidylinositol 3-kinases. The PIKK family in mammals includes the following catalytically active members, DNA-PKcs (DNA-dependent protein kinase catalytic subunit), ATM (ataxia telangiectasia mutated), ATR (ATM- and Rad3-related), mTOR (mammalian target of rapamycin), SMG-1 (suppressor with morphogenetic effect on genitalia-1), as well as the catalytically inactive member TRRAP (transformation/transcription domain–associated protein) (1, 2). PIKKs are large molecules (270 to 470 kD) with a conserved kinase domain, a FAT-C (FRAP, ATM, TRRAP C-terminal) domain, and multiple helical repeats (1, 3). ATM, ATR, TRRAP, and mTOR are evolutionarily conserved from Saccharomyces cerevisiae to Homo sapiens, whereas DNA-PKcs and SMG-1 appeared during metazoan evolution. DNA-PKcs, ATM, ATR, and SMG-1 sense the DNA damage and activate effector proteins to induce cell cycle arrest and DNA repair pathways (47). SMG-1 also detects premature termination codons (PTCs) and activates the messenger RNA (mRNA) surveillance complex to induce degradation of mRNAs with PTCs (8). mTOR senses nutrient status and controls cell growth and proliferation and protein translation (9). TRRAP is involved in transcription and DNA repair as a core component of a histone acetyltransferase (HAT) complex (10). Thus, PIKKs function in controlling accurate gene expression through their roles in surveillance of the integrity of genome and transcripts, as well as function to control protein production in response to nutrient supply.

SMG-1 is part of a multiprotein complex called SMG1C that is composed of SMG-1, SMG-8, and SMG-9 (11). SMG1C is essential for nonsense-mediated mRNA decay (NMD), an mRNA quality control mechanism that occurs in the cytoplasm and detects and degrades mRNAs with PTC, to prevent the production of potentially harmful C-terminally truncated proteins (8, 1214). In mammals, PTC recognition involves sequential remodeling of the mRNA surveillance complex on messenger ribonucleoproteins (mRNPs), which contain the nuclear cap-binding proteins CBP20 and CBP80 and the cytoplasmic poly(A)-binding protein PABPC1. During the initial round of translation, a termination complex, called the SURF complex forms at the site of a PTC-recognizing ribosome. SURF is so named because this complex contains SMG-1, Upf1, and the eukaryotic release factors (eRFs), eRF1 and eRF3. If a cytoplasmic exon-junction complex (EJC; a protein complex deposited 20 to 24 nucleotides upstream of an exon-exon junction and composed of Upf2, Upf3b, eIF4A3, Y14, Magoh, and Btz) is present downstream from the termination codon, the ribosome-SURF complex binds to the EJC to form a DECID (decay-inducing) complex, which triggers SMG-1–mediated phosphorylation of Upf1, an essential process in NMD (8, 11, 12, 15). In addition to its roles in NMD and DNA damage responses (7), SMG-1 may also have an antiapoptotic function that is triggered in response to a tumor necrosis factor–α (TNF-α) (16).

RUVBL1 and RUVBL2, also known as TIP49a and TIP49b or Pontin and Reptin, are closely related AAA+ [ATPases (adenosine triphosphatases) associated with diverse cellular activities] family proteins (17). Both RUVBL1 and RUVBL2 have ATPase and DNA helicase activities in vitro (18). Together, RUVBL1 and RUVBL2 form a double hexamer (1921) and function in various nuclear multimolecular complexes, including various chromatin-remodeling complexes, a HAT complex, and small nucleolar ribonucleoprotein (snoRNP) complexes, and the telomerase reverse transcriptase complex (2224). RUVBL1 and RUVBL2 regulate transcription by interacting with various transcription factors (23). Although RUVBL1 and RUVBL2 are predominantly nuclear proteins (25, 26), they are also detectable in the cytoplasm (27), where their functions remain largely unknown.

In this study, we show that RUVBL1 and RUVBL2 associate with all PIKKs and regulate PIKK functions through two mechanisms, one at the level of the transcripts of PIKKs and one at the level of complex formation. We show that cells deficient in RUVBL1 or RUVBL2 exhibit decreased abundance of the PIKK transcripts and the encoded proteins. We also show that RUVBL1 and RUVBL2 play a critical role in the formation of the mRNA surveillance complex during NMD, through a mechanism that is independent of the effects of RUVBL1 and RUVBL2 on the abundance of SMG-1. These results suggest that RUVBL1 and RUVBL2 coordinate distinct PIKK functions and may serve to direct PIKKs to chromatin-based complexes or RNA-based complexes.

Results

RUVBL1 and RUVBL2 associate with all PIKKs

To identify proteins that interact with SMG-1, we established a stable cell line that expresses SBP (streptavidin-binding peptide)-tagged SMG-1 (SBP-SMG-1) in a doxycycline-dependent manner. After purification of SBP-SMG-1, we identified the copurified proteins by mass spectrometry (Fig. 1A). These proteins included SMG1C subunits (SMG-8 and SMG-9) (11), NMD transacting factors (Upf1, Upf2, eIF4A3, Y14, and Magoh), a SURF-associated factor (eEF2), and molecular chaperones (Hsp90 and Hsp70). RUVBL1, RUVBL2, KIAA0406, and RPB5 were also identified as SMG-1–interacting proteins. RUVBL1 and RUVBL2 form a complex (RUVBL1/2) and function in various nuclear multiprotein complexes (2224). KIAA0406 is an uncharacterized gene encoding a putative protein of 1089 amino acids with no functional domains except for four HEAT (Huntingtin, elongation factor 3, A subunit of protein phosphatase 2A, and TOR1) repeats. Because R10H10.7, the Caenorhabditis elegans ortholog of KIAA0406, is involved in nematode NMD, we named KIAA0406 “SMG-10.” RPB5 is a subunit of RNA polymerases I, II, and III (28). We confirmed the interactions between purified SBP-SMG-1 and RUVBL1, RUVBL2, SMG-10, and RPB5 by Western blotting (Fig. 1B).

Fig. 1

Identification of proteins that interact with SMG-1 and RUVBL1. (A) Silver staining of affinity-purified (AP), streptavidin-binding peptide (SBP)–tagged SMG-1 complexes. Tet-inducible SMG-1 stable FlpIn T-REX 293 cells were treated with doxycycline (DC; 0 or 1 ng/ml) for 3 days. The cell extracts were affinity-purified with streptavidin Sepharose, and biotin-eluted fractions were separated on SDS-PAGE and silver stained. Polypeptides that yielded unambiguous mass spectrometry spectra are indicated. (B) Western blot analysis of AP, SBP-tagged SMG-1 complexes with the indicated antibodies. Input: 3, 1, 0.33, and 0.11% of the AP amount. (C) Silver staining of AP, SBP-tagged RUVBL1 complexes. Tet-inducible RUVBL1 stable FlpIn T-REX 293 cells or control cells, which express tag peptides only, were treated with DC (1 ng/ml) for 3 days. Samples were processed as in (A). Polypeptides that yielded unambiguous mass spectrometry spectra are indicated. All the identified proteins are listed in table S1. (D) Western blot analysis of AP, SBP-tagged RUVBL1 complexes with the indicated antibodies. Input: 1, 0.33, 0.11, and 0.037% of the AP amount. All experiments were performed three times, and typical results are shown.

We also investigated RUVBL1-interacting proteins with a similar strategy. Mass spectrometry identified SMG-10 and RPB5 as RUVBL1-interacting proteins, in addition to RUVBL2 and various known RUVBL1-interacting proteins (Fig. 1C and table S1). We confirmed these interactions by coimmunoprecipitation of SMG-10 and RPB5, and affinity purification of SBP-tagged SMG-10 (fig. S1, A to C). These results indicate that SMG-10 and RPB5 associate with the RUVBL1/2 complex. Consistent with these results, RUVBL1, RUVBL2, and SMG-10 interact with the N-terminal half region of SMG-1 (fig. S1D). The specificity of the interactions identified with SBP-tagged proteins was confirmed by showing that RUVBL1, RUVBL2, α-tubulin, Hsp70, SMG-10, and RPB5 failed to interact with SBP–glutathione S-transferase (GST) (fig. S1E).

The purified SBP-tagged RUVBL1 and SBP–SMG-10 complexes contained slowly migrating bands (Fig. 1C and fig. S1C; asterisk). Mass spectrometry analysis revealed that they contained two PIKK proteins, DNA-PKcs and TRRAP. Western blot analysis of the purified SBP-RUVBL1 complex confirmed this and further revealed the interactions between RUVBL1 and all PIKKs (DNA-PKcs, ATM, ATR, mTOR, and TRRAP) in addition to SMG-1 (Fig. 1D). Kinases in other families, such as PI3K p110α, γ, Akt, c-Jun N-terminal kinase 1 (JNK1), and extracellular signal–regulated kinases 1 and 2 (ERK1/2), were not detected in the purified SBP-RUVBL1 complexes (Fig. 1D). Thus, the RUVBL1/2 complex, and most likely SMG-10 and RPB5, can specifically associate with any kinase of the PIKK family.

RUVBL1 and RUVBL2 regulate the abundance of PIKKs and their downstream signals

The association of the RUVBL1/2 complex with all PIKKs suggested that they are involved in PIKK functions. Therefore, we investigated the effect of knockdown of RUVBL1 or RUVBL2 on downstream signals mediated by each PIKK (4, 5, 9, 12). Knockdown of either RUVBL1 or RUVBL2 caused a significant reduction of the other protein amount, which is consistent with previous observation and the formation of these two proteins into a hexameric complex (Fig. 2) (22). Knockdown of RUVBL1 or RUVBL2 decreased phosphorylation of direct downstream effectors of ATM, ATR, mTOR, and SMG-1. Knockdown of either RUVBL1 or RUVBL2 decreased the phosphorylation of Chk2 at residue Thr68 by ATM in response to ionizing radiation (IR) (Fig. 2A), phosphorylation of Chk1 at residue Ser345 by ATR in response to ultraviolet (UV) radiation (Fig. 2B), phosphorylation of p70 S6K at residue Thr389 by mTOR (Fig. 2C), and SMG-1–mediated phosphorylation of Upf1 at residues Ser1078 and Ser1096 (Fig. 2D).

Fig. 2

Knockdown of RUVBL1 or RUVBL2 impairs PIKK signaling. (A) Analysis of ATM activity in HCT116 cells transfected with the indicated siRNAs and then 60 hours later unexposed or exposed to 3 or 10 Gy irradiation (IR). After 1 hour of incubation, total cell lysates were analyzed by Western blotting with the indicated antibodies. (B) Analysis of ATR activity in HeLa TetOff cells transfected with the indicated siRNAs and then 60 hours later unexposed or exposed to 25 or 100 J/m2 of UV radiation. After 1 hour of incubation, total cell lysates were analyzed by Western blotting with the indicated antibodies. (C) Analysis of mTOR activity in HeLa TetOff cells transfected with indicated siRNAs. Sixty hours later, total cell lysates were analyzed by Western blotting with the indicated antibodies. Cells were cultured in the presence of serum, but no additional stimulation. (D) Analysis of SMG-1 activity in HeLa TetOff cells following the same procedures as in (C). The antibody against P-Upf1 specifically recognizes phosphorylated Upf1 residues Ser1078 and Ser1096. All experiments were performed three times, and typical results are shown.

In addition to decreased PIKK activity, we also found that knockdown of RUVBL1 or RUVBL2 decreased the abundance of PIKKs (Fig. 2) but not the abundance of other kinases, PI3K p110α, γ, Akt, JNK1, and ERK1/2 (fig. S2, A and B). The effect on the abundance of each PIKK differed: DNA-PKcs, ATM, ATR, and mTOR were reduced to less than 25% of the amounts in cells treated with the nonsilencing (NS) small interfering RNAs (siRNAs), whereas TRRAP and SMG-1 were only reduced to 40 to 50% (Fig. 3A) of their abundance in control cells. To determine how RUVBL1 and RUVBL2 regulate the abundance of PIKKs, we tested whether protein stability, translation, or transcription was affected by knockdown of RUVBL1 and RUVBL2. Treatment of cells with cycloheximide, a protein synthesis inhibitor, failed to alter the abundance of the PIKKs in cells in which RUVBL1 and RUVBL2 were knocked down (fig. S3A). None of the proteolysis inhibitors examined restored the abundance of PIKKs in RUVBL1 and RUVBL2 knockdown cells to the abundance in cells exposed to NS siRNA (fig. S3B). However, treatment of cells with 17-AAG, an inhibitor of Hsp90, reduced the abundance of all analyzed PIKKs (fig. S3C). Hsp90 is essential for the proper protein folding and the stability of many protein kinases (29). Because Hsp90 associated with SMG-1 and RUVBL1 (Fig. 1, A and C) and inhibition of Hsp90 activity reduced PIKK abundance, Hsp90 may be involved in the control of PIKK abundance by the RUVBL1/2 complex. Neither SMG-10 knockdown nor RPB5 knockdown notably altered PIKK abundance or the abundance of RUVBL1 or RUVBL2 (fig. S4).

Fig. 3

RUVBL1 and RUVBL2 regulate PIKK abundance at the mRNA level. (A) Knockdown of RUVBL1 or RUVBL2 decreases the abundance of each PIKK. HeLa TetOff cells were transfected with the indicated siRNAs. Sixty hours later, total cell lysates were analyzed by Western blotting with the indicated antibodies. Relative protein abundances compared to those in cells exposed to NS siRNA were graphed. The means ± SD from three independent experiments are shown. (B) The ATPase activities of RUVBL1 and RUVBL2 are required to regulate PIKK abundance. HeLa TetOff cells were transfected with the indicated siRNAs. Twelve hours later, the cells were transfected with rescue plasmids encoding siRNA-insensitive (ins) pSR-RUVBL1ins or RUVBL2ins wild-type (WT) or Asp-to-Asn (DN) mutants (RUVBL1ins-D302N or RUVBL2ins-D299N). Sixty hours after plasmid transfections, cells were harvested and total cell lysates were analyzed by Western blotting with the indicated antibodies. To estimate the PIKK abundance, 100, 33, and 11% of NS control samples were loaded in (A) and (B). (C) RUVBL1 and RUVBL2 affect the amount of mRNAs that encode PIKKs. HeLa TetOff cells were transfected with the indicated siRNAs. Sixty hours later, total cytoplasmic RNAs were analyzed by real-time quantitative PCR with Taqman specific probes. Relative mRNA expression was normalized to those of GAPDH and 18S rRNA, and the means ± SE (n = 3) from three independent experiments were graphed (*P < 0.05 compared to NS control). All experiments were performed three times, and typical results are shown.

To determine if the ATPase activity of RUVBL1 and RUVBL2 was essential for controlling the abundance of PIKKs, we performed rescue experiments using siRNA-resistant RUVBL wild-type or ATPase activity–deficient mutants (DN mutants: RUVBL1 D302N or RUVBL2 D299N) (30). Wild-type RUVBL1 or RUVBL2 rescued the reduced PIKK abundance, whereas the DN mutants failed (Fig. 3B). These results indicate that the ATPase activities of both RUVBL1 and RUVBL2 are required to control the abundance of PIKKs; however, the exact mechanisms by which this is achieved remains unclear.

RUVBL1 and RUVBL2 associate with various transcription factors and chromatin-remodeling complexes and thereby regulate transcription (23). Thus, RUVBL1 and RUVBL2 may regulate the transcription of PIKK-encoding genes. Real-time quantitative PCR analysis of mRNAs from RUVBL1 or RUVBL2 knockdown cells revealed that, with the exception of SMG-1, PIKK mRNAs were decreased (Fig. 3C). Significant decreases in transcript abundance were observed in the RUVBL1 and RUVBL2 knockdown cells for mRNAs encoding DNA-PKcs, ATM, ATR, and mTOR, with ATM transcripts reduced the most. TRRAP mRNA was slightly, but significantly reduced; whereas SMG-1 mRNA was not reduced. Knockdown of RUVBL1 or RUVBL2 did not reduce the abundance of the mRNA encoding PI3K p110α (fig. S2C).

RUVBL1 and RUVBL2 affect the function of SMG-1 to contribute to nonsense-mediated mRNA decay in mammalian cells

The interactions between the RUVBL1/2 complex and each PIKK suggested that the RUVBL1/2 complex may directly regulate PIKK functions. To evaluate this possibility, we focused on SMG-1 because the knockdown effect of RUVBL1 or RUVBL2 for SMG-1 abundance was not significant and we could monitor SMG-1 function at a time point when the abundance of SMG-1 was not affected by RUVBL1 or RUVBL2 knockdown (fig. S2A). SMG-1–mediated Upf1 phosphorylation was substantially reduced by RUVBL1 or RUVBL2 knockdown (Fig. 4A). Rescue experiments with siRNA-resistant RUVBL1 wild-type or the DN mutant (RUVBL1 D302N) revealed that the ATPase activity of RUVBL1 is required for SMG-1–mediated Upf1 phosphorylation (Fig. 4B). The DN mutant had a dominant-negative effect on Upf1 phosphorylation, providing additional evidence for the involvement of these ATPases in SMG-1–mediated Upf1 phosphorylation (Fig. 4B). Thus, the RUVBL1/2 complex regulates SMG-1–mediated Upf1 phosphorylation under conditions in which SMG-1 abundance is not changed.

Fig. 4

RUVBL1 and RUVBL2 are required for NMD independent of the control of the SMG-1 abundance. (A) Upf1 phosphorylation is impaired in the absence of RUVBL1 or RUVBL2. HeLa TetOff cells were transfected with the indicated siRNAs. Forty-five hours after transfections, total cell lysates were analyzed by Western blotting with the indicated antibodies. (B) The ATPase activity of RUVBL1 is required for phosphorylation of Upf1. HeLa TetOff cells were transfected with plasmids encoding an siRNA-insensitive (ins) RUVBL1ins wild-type (WT) or DN mutant (RUVBL1ins-D302N). Twelve hours later, cells were transfected with the indicated siRNAs. Forty-five hours after siRNA transfections, total cell lysates were analyzed by Western blotting with the indicated antibodies. (C) The Tet-inducible BGG-PTC or BGG-WT reporter plasmids. (D) NMD is inhibited by knockdown of RUVBL1 or RUVBL2. HeLa TetOff cells were cotransfected with the indicated siRNAs and Tet-inducible BGG-PTC plasmid. After the addition of doxycycline to repress the transcription of the reporter plasmids, total RNAs prepared at the times indicated were analyzed by Northern blotting. The quantities of BGG mRNA, normalized to GADPH signals, were plotted (t1/2, half-life). The means ± SD from three independent experiments are shown. (E) HeLa TetOff cells were cotransfected with the indicated siRNAs and Tet-inducible BGG-WT plasmid and treated and analyzed as in (D). The mean values from two independent experiments are shown. Each experiment was performed two times (E) or three times [(A), (B), and (D)], and typical results are shown.

To show that RUVBL1 and RUVBL2 are involved in NMD, RUVBL1 or RUVBL2 was knocked down in HeLa TetOff cells transfected with either a wild-type or a PTC-containing β-globin reporter (Fig. 4C), and the half-life of the β-globin reporter mRNA (after addition of doxycycline to repress transcription of the reporter gene) was evaluated by Northern blotting. Note that the cells were harvested at time points when the knockdown of RUVBL1 or RUVBL2 did not affect the abundance of SMG-1 (36 to 44 hours after siRNA transfections) (fig. S2D). Knockdown of either RUVBL1 or RUVBL2 stabilized the PTC-containing β-globin mRNAs (Fig. 4D). In contrast, neither knockdown affected the half-life of the wild-type β-globin reporter (Fig. 4E). The requirement of RUVBL1 for NMD was also confirmed by monitoring the effect of RUVBL1 knockdown on the abundance of UHG and GAS5 mRNAs (fig. S5), which are endogenous targets of NMD (11, 31). These results indicate that the RUVBL1/2 complex is required for NMD through a mechanism independent of the control of SMG-1 abundance. Because we identified SMG-10 and RPB5 as common interacting proteins of SMG-1 and RUVBL1, we also examined the effects of knockdown of SMG-10 or RPB5 on SMG-1–mediated Upf1 phosphorylation and NMD. RPB5 knockdown decreased SMG-1–mediated Upf1 phosphorylation and prolonged the half-life of PTC-containing mRNAs, suggesting that RPB5 is likely involved in NMD (fig. S6). However, knockdown of RPB5 also decreased UHG and GAS5 mRNA most likely through general transcription suppression (fig. S5). The requirement of RPB5 for NMD might provide an additional link between transcription and transcribed mRNA fate like that of other RNA polymerase II subunits (32). In contrast, SMG-10 knockdown did not significantly affect either Upf1 phosphorylation or NMD (Fig. 4D and figs. S5 and S6A).

The function of RUVBL1 and RUVBL2 in nonsense-mediated mRNA decay is evolutionarily conserved

To investigate whether the function of RUVBL1 and RUVBL2 in NMD is evolutionarily conserved, we tested whether the C. elegans orthologs of RUVBL1 and RUVBL2 (encoded by ruvb-1 and ruvb-2, respectively) are involved in C. elegans NMD. C. elegans RUVB-1 and RUVB-2 are 56 and 57% identical, respectively, to human RUVBL1 and RUVBL2. We performed RNA interference (RNAi) experiments targeting ruvb-1 and ruvb-2 and assessed C. elegans NMD with two different assays. The first assay measures phenotypic suppression of unc-54(r293), an allele that causes a motility phenotype that reflects the efficiency of NMD. The second assay measures directly the effects of ruvb-1(RNAi) and ruvb-2(RNAi) on the abundance of an endogenous NMD substrate.

unc-54 encodes a muscle myosin heavy chain, and unc-54 loss-of-function mutants are strongly paralyzed. Whereas the phenotype of most unc-54 alleles is unaffected by NMD, the motility defects of unc-54(r293) are suppressed by mutations that reduce or eliminate NMD (33, 34). To increase the sensitivity of this phenotypic assay, we introduced mutations of two other genes into an unc-54(r293) genetic background. rrf-3(pk1426) enhances the efficiency of RNAi in C. elegans (35). smg-1(cc546ts) increases our ability to detect weak NMD defects. smg-1(cc546ts) is a temperature-sensitive allele of smg-1. smg-1(cc546ts) animals are defective for NMD at 25°C, but they are competent for NMD at 20°C. smg-1(cc546ts) unc-54(r293) animals exhibit normal motility at 25°C due to phenotypic suppression of unc-54(r293) paralysis by smg-1(cc546ts), which is inactive at 25°C. They are paralyzed at 20°C as a result of NMD-dependent instability of unc-54(r293) mRNA [smg-1(cc546ts) is active at 20°C]. Presumably, because the mutant SMG-1 protein is not fully normal even at 20°C, we could detect weak NMD defects in strains containing smg-1(cc546ts) grown at 20°C, whereas weak NMD defects were not detectable in smg-1(+) genetic backgrounds. For example, when expression of smg-8 is inhibited by feeding RNAi, we observe phenotypic suppression of unc-54(r293) at 20°C, but only if the strain also contains smg-1(cc546ts) (11) (Table 1). Thus, the triple-mutant strain unc-54(r293) smg-1(cc546ts); rrf-3(pk1426) provides a sensitized genetic background in which we can observe NMD defects by scoring the motility of the animals.

Table 1

Phenotypic suppression of the motility defects of unc-54(r293) after RNAi. smg-1(+) indicates a wild-type allele of smg-1. smg-1(0) indicates a null allele that expresses no detectable SMG-1 protein. smg-1(cc546ts) is a temperature-sensitive allele of smg-1 that is defective for NMD at 25°C and competent for NMD at 20°C. All strains, except wild-type, contain rrf-3(pk1426) in addition to the indicated smg-1 and unc-54 mutations.

View this table:

We performed ruvb-1(RNAi) and ruvb-2 (RNAi) in the triple-mutant strain unc-54(r293) smg-1(cc546ts); rrf-3(pk1426) and analyzed motility at 20°C. We consistently observed improved motility (indicative of phenotypic suppression and, hence, inhibition of NMD) in worms in which either ruvb-1(RNAi) or ruvb-2(RNAi) was knocked down (Table 1). The strength of phenotypic suppression of unc-54(r293) by ruvb-1(RNAi) or ruvb-2(RNAi) is similar to but slightly stronger than that of smg-8(RNAi). Such suppression was considerably weaker than that of smg-2(RNAi). We also observed weak phenotypic suppression of unc-54(r293) after RNAi of R10H10.7, a C. elegans gene encoding SMG-10 ortholog. R10H10.7 is 29% similar to human SMG-10. These results indicate that ruvb-1(RNAi), ruvb-2(RNAi), and R10H10.7(RNAi) (termed smg-10) exhibit weak NMD defects as judged by in vivo phenotypic suppression of the unc-54(r293) motility defect.

To confirm that the phenotypic effects result from partial inhibition of NMD after ruvb-1(RNAi), ruvb-2(RNAi), or smg-10(RNAi), we investigated the abundance of alternatively spliced rpl-12 mRNAs after RNAi. rpl-12 pre-mRNA is alternatively spliced to yield two mature mRNAs, one of which does not contain a PTC [termed “rpl-12(+)” in Fig. 5A], and one of which contains a PTC [termed “rpl-12(PTC)” in Fig. 5A] (36). The efficiency of NMD is reflected in the ratio of the two spliced rpl-12 mRNAs, which are distinguishable by reverse transcription polymerase chain reaction (RT-PCR) (37). We measured the relative abundance of rpl-12(PTC) and rpl-12(+) mRNAs after treatment of unc-54(r293) smg-1(cc546ts); rrf-3(pk1426) animals at 20°C with double-stranded RNA targeting ruvb-1, ruvb-2, or smg-10 (Fig. 5B). By measuring the relative, as opposed to the absolute, abundance of each mRNA, we controlled for slight variation in the quantities of mRNA from sample to sample and in the efficiencies of RT-PCR. An increase in the ratio of rpl-12(PTC) to rpl-12(+) mRNA (termed the “PTC/+ ratio”) after RNAi indicates inhibition of NMD, with the magnitude of increase reflecting the degree of inhibition.

Fig. 5

RUVBL1, RUVBL2, and SMG-10 are involved in NMD in C. elegans. (A) Diagram of rpl-12 alternative splicing. Unproductively spliced mRNA rpl-12(PTC) is a substrate of NMD, but productively spliced mRNA rpl-12(+) is not. (B) RT-PCR analysis of rpl-12 alternatively spliced mRNAs after RNAi. Strain unc-54(r293) smg-1(cc546ts); rrf-3(pk1426) was grown at 20°C in the presence of the indicated feeding RNAi clones. smg-2(r908) is an smg-2 deletion that expresses no SMG-2 protein and is completely defective for NMD. WT strain N2 is normal for NMD. Data shown are the range of values from two independent experiments or the value for the single experiment performed with ruvb-2(RNAi).

ruvb-1(RNAi), ruvb-2(RNAi), and smg-10(RNAi) yielded an increased PTC/+ ratio relative to the empty vector (Fig. 5B). The PTC/+ ratios of ruvb-1(RNAi), ruvb-2(RNAi), and smg-10(RNAi) were considerably less than that of smg-2(RNAi) or an smg-2 deletion mutant without RNAi, but they were comparable to results obtained for smg-8(RNAi), which was consistent with previous reports (11). These results, together with those showing phenotypic suppression of the motility defect of unc-54(r293), suggest that ruvb-1(RNAi), ruvb-2(RNAi), and smg-10(RNAi) caused weak NMD defects. An important unanswered question is whether these weak NMD defects are directly due to involvement of RUVB-1, RUVB-2, and SMG-10 in C. elegans NMD or to indirect effects of RNAi targeting essential genes. It is likely that ruvb-1, as well as ruvb-2, are essential genes in C. elegans, and inhibiting their expression causes pleiotropic effects (3842). On the other hand, contributions of ruvb-1, ruvb-2, and smg-10 to C. elegans NMD may be underestimated in our experiments because of incomplete inhibition of their expression by RNAi.

RUVBL1 knockdown impairs the formation of the mRNA surveillance complex on mRNPs

To understand how the RUVBL1/2 complex is involved in SMG-1–mediated Upf1 phosphorylation, which occurs on a spliced mRNP in the cytoplasm, we investigated whether the RUVBL1/2 complex associates with such mRNPs. Two proteins that associate with capped and poly(A) tail–containing transcripts before and during the mRNA surveillance process are CBP80 and PABPC1 (11, 15). CBP80 and PABPC1 associated with RUVBL1 affinity purified from the cytoplasmic fraction only in the absence of ribonuclease (RNase) treatment, whereas the association between RUVBL1 and SMG1C (SMG-1–SMG-8–SMG-9) was not affected by RNase treatment (Fig. 6A). Thus, the association of RUVBL1 with CBP80 and PABPC1 is likely mediated by RNA, whereas the association between RUVBL1 and SMG1C is likely a protein-protein interaction.

Fig. 6

RUVBL1 and RUVBL2 are essential for the formation of mRNA surveillance complexes on mRNPs. (A) Cytoplasmic RUVBL1 associates with components of the mRNA surveillance complexes. Tet-inducible RUVBL1 stable FlpIn T-REX 293 cells or control cells were treated with doxycycline (1 ng/ml) for 3 days. The cells were fractionated into nuclear fractions (N) and cytoplasmic fractions (C). GAPDH was used as a marker for the cytoplasmic fraction, lamin C for the nuclear fraction. The cell extracts from cytoplasmic fractions were subjected to affinity purification with streptavidin Sepharose in the presence or absence of RNase. Purified RUVBL1 complexes (AP-SBP tag) were analyzed by Western blotting with the indicated antibodies. Input: 1, 0.33, 0.11, and 0.037% of the amount affinity purified. (B) RUVBL1 and RUVBL2 associate with SURF. HeLa TetOff cells were transfected with plasmids encoding HA-tagged Upf1 WT, HA-tagged Upf1-C126S (CS), or control plasmids. The scheme shows the domains of Upf1 and the position of the Upf1-C126S mutation. The cell extracts (input) were immunoprecipitated with anti-HA affinity matrix in the presence of RNase. The immunoprecipitates were analyzed by Western blotting with the indicated antibodies. (C) Knockdown of RUVBL1 impairs the interaction between Upf1 and EJC components. HeLa TetOff cells were transfected with the indicated siRNAs. Forty-five hours after siRNA transfections, the cytoplasmic cell extracts (input) were immunoprecipitated with antibodies against Upf1 in the presence of RNase. Normal mouse immunoglobulin G (IgG) was used as control IgG. The recovered immunocomplexes were analyzed by Western blotting with the indicated antibodies. (D) Knockdown of RUVBL1 impairs the interaction between Y14 and SURF components. HeLa TetOff cells were transfected with the indicated siRNAs and prepared as described for (C) except that the cytosplasmic cell extracts were immunoprecipitated with an antibody against Y14. (B to D) Input: 5, 1.67, 0.56, and 0.19% of the amount immunoprecipitated. All experiments were performed three times, and typical results are shown.

SMG-1–mediated Upf1 phosphorylation is induced by successive remodeling of the mRNA surveillance complex, involving first SURF formation on a PTC-recognizing ribosome and then subsequent formation of the DECID complex, which is defined as the ribosome-SURF-EJC complex on an mRNP (11, 15) (Fig. 7). Thus, we examined the association between the SURF and the RUVBL1/2 complexes. SMG-1 and Upf1, components of SURF, interact with EJC through Upf2, an EJC-associated protein, and a defect in the association of Upf2 with Upf1 freezes the remodeling of mRNA surveillance complex, allowing the mRNP to accumulate in the SURF state. The mutation C126S in Upf1 prevents its interaction with Upf2 and results in the accumulation of the SURF-state mRNP (11, 15). We transiently transfected hemagglutinin (HA)-tagged Upf1 (HA-Upf1) wild-type or the C126S mutant form, immunoprecipitated the tagged Upf1, and analyzed the coprecipitated proteins by Western blotting (Fig. 6B). RUVBL1 and RUVBL2 were more abundant in the Upf1-C126S immunoprecipitates than in Upf1-wild type immunoprecipitates, suggesting that the RUVBL1/2 complex associates with SURF, where it may function in the remodeling of the surveillance complex to form a DECID.

Fig. 7

RUVBL1 and RUVBL2 promote the remodeling of mRNA surveillance complexes during NMD. In mammals, PTC recognition is established by the formation of mRNA surveillance complexes called “SURF” and “DECID” that occur at sites containing a PTC-recognizing ribosome and downstream EJC during the initial round of translation. During translation termination, the transiently formed SURF complex, composed of SMG1C (SMG-1–SMG-8–SMG-9 complex), Upf1, and eRF, detects the downstream EJC and forms DECID complex, which induces SMG-1–mediated Upf1 phosphorylation. Phosphorylated Upf1 becomes a mark of PTC-mRNA and this directs mRNA decay. Through a process involving their ATPase activities, RUVBL1 and RUVBL2, together with SMG-10 and RPB5, associate with SURF and promote DECID formation. The order and the timing of the release of the ribosome and eRF are unknown. AUG, start codon; Ter, termination codon; eRF, eRF1-eRF3 complex; S8, SMG-8; S9, SMG-9; RUV1, RUVBL1; RUV2, RUVBL2; P, phosphate group; SURF, SMG-1–Upf1–eRF1–eRF3 complex.

To investigate whether the RUVBL1/2 complex is involved in the remodeling to form a DECID, we immunoprecipitated endogenous Upf1 or Y14, in the presence of RNase, from the cytoplasmic fractions of cells treated with NS, RUVBL1, or SMG-1 siRNA and analyzed the immunoprecipitates. RNase was added to avoid RNA-mediated interactions. The cells were harvested at a time point when the knockdown of RUVBL1 did not affect the abundance of SMG-1 (Fig. 6, C and D). RUVBL1 knockdown impaired the interaction between Upf1 and the EJC components, Upf2, Upf3b, eIF4A3 (Fig. 6C), and Y14 (Fig. 6D). In contrast, SMG-1 knockdown enhanced these interactions, which indicates that the defects observed in RUVBL1 knockdown were not caused by a reduction of SMG-1 abundance or function (Fig. 6, C and D). Reduced interactions between Y14 and the other SURF components, SMG-1, SMG-8, and SMG-9, were also observed in cells in which RUVBL1 was knocked down (Fig. 6D). RUVBL1 and RUVBL2 coprecipitated with Y14 and this was not affected by SMG-1 knockdown, suggesting that the RUVBL1/2 complex can associate with the EJC in an SMG-1–independent manner (Fig. 6D).

Knockdown of RUVBL1 or SMG-1, which decreased the phosphorylation of Upf1, also decreased the coprecipitation of SMG-5 and SMG-7 with Upf1 (Fig. 6C). SMG-5 and SMG-7 preferentially associate with phosphorylated Upf1 and are required for NMD (43); thus, the decrease in the interaction of these two proteins with Upf1 when RUVBL1 is knocked down provides further support for a role of the RUVBL1/2 complex in SMG-1–mediated Upf1 phosphorylation. Together, these coprecipitation experiments suggest that the RUVBL1/2 complex is involved in the formation of DECID and that there may be multiple proteins with which the RUVBL1/2 complex interacts to facilitate the association of the SURF with the EJC.

Discussion

In this study, we revealed unexpected molecular and functional links between the conserved ATPases, RUVBL1 and RUVBL2, and a family of protein kinases, PIKKs that function in cellular surveillance processes for maintaining the genome integrity and accurate and appropriate gene expression. RUVBL1 and RUVBL2, presumably acting as a complex, regulate PIKKs in at least two different ways: RUVBL1 and RUVBL2 control PIKK abundance at least at the mRNA level and RUVBL1 and RUVBL2 directly influence PIKK function. Specifically, we showed that these ATPases are directly involved in the remodeling of the mRNA surveillance complex during NMD (Fig. 7). These findings not only suggest the coordinated regulation of different PIKK functions by the RUVBL1/2 complex, but also imply interplay between processes mediated by the RUVBL1/2 complex and PIKK-mediated surveillance processes.

We demonstrated that regulation of PIKK abundance by the RUVBL1/2 complex required their ATPase activities (Fig. 3) and that knockdown of RUVBL1 or RUVBL2 impaired PIKK signaling (Fig. 2). Decreased PIKK signaling may explain some phenotypes observed in RUVBL1-depleted cells. For example, RUVBL1 knockdown increases the sensitivity of cells to a DNA-damaging agent and causes defects in the recruitment of RAD51, a downstream mediator of ATM and ATR, to the site of DNA damage (44).

The RUVBL1/2 complex regulates PIKK abundance at least at the mRNA level. Considering that RUVBL1 and RUVBL2 associate with several chromatin-remodeling or -modifying complexes and with transcription factors, such as c-Myc and E2F to regulate transcription (23, 4547), they might be involved in the transcriptional regulation of PIKKs. Consistent with this hypothesis, the transcription of the genes encoding ATM and DNA-PKcs is responsive to E2F1, which can be regulated by RUVBL1 and c-Myc (45, 46, 4850).

The RUVBL1/2 complex may also regulate PIKK abundance at the level of protein production or structural maturation. RUVBL1 knockdown did not alter PIKK protein stability (fig. S3A); however, the RUVBL1/2 complex associates with the molecular chaperone Hsp90 and its cofactors, RPAP3 and NOP17 (5153). We identified these proteins as RUVBL1- and SMG-10–interacting proteins (Fig. 1C and fig. S1C). Consistent with a possible role for Hsp90 in PIKK production, the treatment of cells with a Hsp90 inhibitor decreased the abundance of all PIKK proteins (fig. S3C). We also found that RUVBL1 associated with TELO2 (Fig. 1C), which increases the stability of PIKK protein without changing PIKK-encoding mRNA abundance (54). Further analysis is needed to clarify all of the mechanisms by which the RUVBL1/2 complex promotes PIKK abundance.

In addition to the regulation of PIKK abundance, we also revealed that the RUVBL1/2 complex interacts with all PIKKs (Fig. 1D). The physical interaction of the RUVBL1/2 complex with PIKKs suggests their direct involvement in the PIKK functions. Together with the involvement of RUVBL1 and RUVBL2 in various chromatin-related complexes, their physical interaction with PIKKs suggests an unexpected interplay and cooperation between chromatin-based processes and PIKK-mediated surveillance mechanisms, which might be important to maintain the integrity of the genome and ensure accurate gene expression (5558).

Detailed investigation of the interaction of SMG-1 and the RUVBL1/2 complex showed that the RUVBL1/2 complex plays a critical role in NMD through a mechanism that is independent of an effect on SMG-1 abundance. RUVBL1 interacts with the NMD transacting factors both in the SURF complex and EJC and these interactions are independent of RNA (Figs. 1D and 6, A, B, and D). RUVBL1 knockdown impairs the interaction between components of SURF and components of EJC, suggesting that RUVBL1 knockdown impairs DECID formation. DECID formation induces SMG-1–mediated Upf1 phosphorylation, which directs PTC-mRNA decay (11, 15). Knockdown of RUVBL1 or RUVBL2 inhibited Upf1 phosphorylation, resulting in the inhibition of NMD (Fig. 4 and fig. S5). The requirement for the ATPase activity of RUVBL1 for Upf1 phosphorylation suggests that the RUVBL1/2 complex regulates the complex formation through their ATPase activities (Fig. 4B). Because we found an interaction between the RUVBL1/2 complex and each PIKK (Fig. 1D), the RUVBL1/2 complex may also be involved in the formation of other PIKK-containing complexes. For example, RUVBL1 and RUVBL2, together with TRRAP, a PIKK, are components of the TIP60 HAT complex, and the knockdown of RUVBL1 decreased TIP60 HAT activity toward core histones (59). RUVBL1 and RUVBL2 also might influence the kinase activity of PIKKs because at least ATM is activated by TIP60 HAT-mediated acetylation in response to DNA damage (60). The RUVBL1/2 complex has been found in various nuclear DNA-protein and RNA-protein multimolecular complexes, and their knockdown or depletion leads to insufficient complex assembly and results in functional defects (22, 61, 62). We now report a role for the RUVBL1/2 complex in the formation and remodeling of a cytoplasmic RNA-protein complex. Thus, a common function of the RUVBL1/2 complex may be to control multimolecular complex formation or remodeling in various situations.

Materials and Methods

Plasmids

The pcDNA5/FRT/TO/NTAP vector was constructed by insertion of the blunted Sac I–Pvu I fragment of the pNTAP vector (Stratagene) into the Pme I site of the pcDNA5/FRT/TO vector (Invitrogen). pcDNA5/FRT/TO/NTAP-SMG-1, pcDNA5/FRT/TO/NTAP-RUVBL1, pcDNA5/FRT/TO/NTAP-SMG-10, pSR-RUVBL1, and pSR-RUVBL2 were constructed by the cloning of each complementary DNA (cDNA) fragment by standard methods. The cDNAs of RUVBL1 and RUVBL2 were mutated by site-directed mutagenesis at nucleotides 307 to 309 (RUVBL1) or 51, 52, and 54 (RUVBL2) for siRNA resistance. A U6 promoter–driven short hairpin RNA expression cassette (Promega) for RUVBL1 [siRNA sequence: siGENOME duplex D-012299-01 (Dharmacon)] or SMG-10 (siRNA sequence targeted to the 3′ untranslated region of the SMG-10 mRNA: GGTGAATTAGTTAGCCAAT) was inserted into the Pci I or Bst Z17I site of pcDNA5/FRT/TO/NTAP-RUVBL1 or pcDNA5/FRT/TO/NTAP-SMG-10, respectively. In pSR-RUVBL1-D302N and pSR-RUVBL2-D299N, the aspartic acids at residues 302 or 299, respectively, were substituted with asparagine by site-directed mutagenesis. HA-Upf1-WT, HA-Upf1-C126S, pTRE-BGG-WT, and pTRE-BGG-PTC reporter plasmids have been previously described (11, 12, 15, 43).

Antibodies

Antibodies that recognize SMG-10 and against RPB5 were generated against recombinant human SMG-10 (amino acids 773 to 853) or full-length RPB5 protein fused to GST or maltose binding protein (MBP), respectively. Affinity purification of antibodies against SMG-10 was performed following standard procedures. The antibodies against SMG-1, SMG-8, SMG-9, Upf1, Upf2, Upf3b, SMG-5, SMG-7, and phosho-Upf1 (clone 8E6 or 7D5) have been described previously (11, 12, 15, 43). Antibodies to RUVBL1 (Biomatrix and Santa Cruz), RUVBL2 (BD Transduction Laboratories and Abcam), RPB5 (Euromedex), DNA-PKcs (Bethyl), ATM (Cell signaling technology), ATR (Calbiochem), mTOR (Cell Signaling Technology), TRRAP (Bethyl), PI3K p110α (Cell Signaling Technology), PI3K p110γ (Cell Signaling Technology), Akt (Cell Signaling Technology), JNK1 (BD Biosciences), ERK1/2 (Millipore), Chk1 (Santa Cruz), P-Chk1 (Ser345) (Cell Signaling Technology), Chk2 (Cell Signaling Technology), P-Chk2 (Thr68) (Cell Signaling Technology), p70 S6K (Santa Cruz), P-p70 S6K (Thr389) (Cell Signaling Technology), eIF4A3 (ProteinTech), Y14 (Sigma), eRF3a (Abcam), PABPC1 (Abcam), CBP80 (ProteinTech), myosin IIa (Sigma), β-actin (Sigma), glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Abcam), lamin A/C (Santa Cruz), and the HA tag (3F10) (Roche) were obtained commercially.

siRNAs

The following siRNA target sequences were used: RUVBL1, siGENOME duplex D-008977-02 (Dharmacon); RUVBL2#3, siGENOME duplex D-012299-03 (Dharmacon); RUVBL2#1, siGENOME duplex D-012299-01 (Dharmacon), RUVBL2#1 siRNA was used only for the rescue experiment in Fig. 3B; SMG-10, siGENOME duplex J-014188-05 (Dharmacon); SMG-1, AAGTGTATGTGCGCCAAAGTA; mTOR, Hs_FRAP1_7_HP Validated siRNA SI03023587 (Qiagen); NS siRNA, All Star Negative Control siRNA (Qiagen).

Cell culture and transfection

HeLa TetOff cells (Clontech) and Flp In T-REX HEK293 cells (Invitrogen) were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, penicillin (100 U/ml), and streptomycin (100 μg/ml). HCT116 cells were grown in McCoy’s 5A medium (Sigma) supplemented with 10% fetal bovine serum, penicillin (100 U/ml) and streptomycin (100 μg/ml). Plasmid transfections were performed in 6- or 12-well plates or 15-cm dishes with Lipofectamine LTX (Invitrogen) according to the manufacturer’s recommendations. For the immunoprecipitation of HA-tagged proteins, transfected cells were harvested 36 hours after transfection. siRNA transfections were performed in 12-well plates or in 15-cm dishes with siLentFect (BioRad) or Hiperfect (Qiagen) according to the manufacturer’s protocol, and cells were harvested 36 to 60 hours later as indicated in the figure legends.

Real-time quantitative PCR

HeLa TetOff cells were transfected with siRNAs and harvested 60 hours later. Total RNAs were isolated with an RNeasy Plus kit (Qiagen) from cytoplasmic fractions. cDNAs were reverse-transcribed with a High Capacity cDNA Reverse Transcription kit (Applied Biosystems) and subjected to real-time quantitative PCR with iQ Supermix (BioRad) and iCycler iQ system (BioRad) according to the manufacturer’s recommendations. The following Taqman primer sets (TaqMan Gene Expression Assays, Applied Biosystems) were used: DNA-PKcs, Hs00179161_m1; ATM, Hs00175892_m1; ATR, Hs00169878_m1; mTOR, Hs01042424_m1; SMG-1, Hs00247891_m1; TRRAP, Hs01591130_m1; GAPDH, #4352934E; 18S rRNA, #4352930E. Relative mRNA expression levels were normalized to those of GAPDH and 18S ribosomal RNA (rRNA), and the means ± SE (n = 3) from three independent experiments were calculated. The two-tailed Student’s t test (Microsoft Excel 2007) was used to analyze the differences between the pairs of groups. Values were regarded significant at P < 0.05.

Affinity purification and mass spectrometry

Tet-inducible stable cells were established with the Flp In T-REX system (Invitrogen) following the manufacturer’s recommendations. Cloned stable cells were cultured in doxycycline (0 or 1 ng/ml) for 3 days and lysed with a loose-fit Potter-Elvehjem homogenizer in T buffer [20 mM Hepes-NaOH at pH 7.5, 50 mM NaCl, 0.05% Tween 20, 2.5 mM MgCl2, 0.5 mM dithiothreitol (DTT), 100 nM okadaic acid (Calbiochem), protease inhibitor cocktail (Roche), phosphatase inhibitor cocktail (Roche), and RNaseA (100 μg/ml; Qiagen)]. The soluble fractions were precleared with Sepharose 4B (Sigma) and then incubated with streptavidin Sepharose (GE Biotech) for 2 hours at 4°C with gentle rotation. After washing with RNase(−) lysis buffer, the affinity-purified protein complexes were eluted by incubation at 4°C for 30 min with RNase(−) lysis buffer containing 2 mM biotin (Sigma). The eluted fractions were separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and stained with Silver Stain Reagent II (Cosmobio) or SYPRO Ruby (BioRad), or transferred to Pro Blott membrane (Applied Biosystems) and stained with Colloidal Gold Total Protein Stain (BioRad). Excised protein bands were digested and subjected to matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry (MALDI-TOF-MS) or liquid chromatography–tandem mass spectrometry (LC-MS/MS). Peptide sequences were analyzed with the Mascot search engine (http://www.matrixscience.com/search_form_select.html).

Immunoprecipitation and Western blot analysis

HeLa TetOff cells were transfected with Lipofectamine LTX (Invitrogen) or siLentFect (BioRad), and lysed in T buffer (as above) or in NP-40 buffer [50 mM tris-HCl (pH 7.4), 150 mM NaCl, 0.4% NP-40, 2 mM MgCl2, 1 mM DTT, 100 nM okadaic acid (Calbiochem), protease inhibitor cocktail (Roche), phosphatase inhibitor cocktail (Roche), cycloheximide (100 μg/ml; Sigma), and RNaseA (100 μg/ml; Qiagen)]. The soluble fractions were subjected to preclearing and then incubated with antibodies for 1 hour at 4°C with gentle rotation. Subsequently, the soluble fractions were incubated with 30 μl of protein G Sepharose (GE Biotech) for an additional 1 hour at 4°C with gentle rotation. To immunoprecipitate HA-Upf1, 25 μl of anti-HA affinity matrix (Roche) was used. After washing with RNase(−) lysis buffer, the immunocomplexes were boiled in SDS sample buffer and analyzed by Western blotting. In experiments where anti-HA affinity matrix or Upf1-monoclonal antibody was used, proteins recovered on the matrix were eluted by incubation at 37°C for 15 min with RNase(−) lysis buffer containing HA-peptide (1 mg/ml) or Upf1 peptide (1 mg/ml), boiled with SDS sample buffer, and analyzed by Western blotting. All proteins in Western blot experiments were detected with an ECL Western blot detection kit (GE Biotech) or Lumi-Light Western blot substrate (Roche) and quantified with a Lumino-Imager, LAS-3000, and Science Lab 2001 Image Gauge software (Fuji Photo Film).

NMD reporter assay in mammalian cells

NMD analysis was performed in essentially the same manner as described elsewhere (11), except that HeLa TetOff cells (1.8 × 106) were transfected with siRNAs in combination with 1.5 μg of pTRE-BGG-WT or -PTC. Thirty-six hours after transfection, doxycycline was added at time zero to inhibit de novo transcription, and total RNAs isolated at the indicated time points were analyzed by Northern blotting with β-globin or GAPDH as control probes. The quantities of β-globin gene (BGG) mRNA normalized to GADPH signals were graphed.

RNAi, motility analysis, and RT-PCR in C. elegans

Escherichia coli HT115 (DE3) strains harboring plasmids derived from the feeding RNAi vector L4440 that express ruvb-1, ruvb-2, or control double-stranded RNAs (dsRNAs) were cultured overnight in LB medium containing carbenicillin (25 μg/ml; Sigma) and tetracycline (12.5 μg/ml; Sigma). Eighty microliters of liquid culture was spread on nematode growth medium (NGM) agar containing 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside; Sigma) and carbenicillin (25 μg/ml) and incubated at 30°C overnight to induce T7 RNA polymerase and allow dsRNA synthesis. Plates were equilibrated to 20°C, nematodes of genotype unc-54(r293) smg-1(cc546ts); rrf-3(pk1426) were added, and phenotypic suppression of unc-54(r293) was scored among the offspring as described elsewhere (33). Mutation rrf-3(pk1426) makes RNAi more efficient (35). smg-1(cc546ts) is a temperature-sensitive allele of smg-1 that is NMD defective at 25°C but NMD competent at 15°C and 20°C. Strains harboring smg-1(cc546ts) and grown at 20°C provided a sensitized genetic background in which weak NMD defects were detected. Such weak NMD defects are not detected in smg-1(+) genetic backgrounds. For RT-PCR, total RNA was extracted from mixed-stage animals with TRIzol (Invitrogen). First-strand cDNA synthesis was performed with random hexamers and SuperScript III reverse transcriptase (Invitrogen). rpl-12 mRNA was amplified with primers on either side of the alternatively spliced intron (37), and RT-PCR products were separated on 1.5% ethidium bromide–stained agarose gels.

Acknowledgments

Acknowledgments: We thank S. Getz, S. Zu, and A. Fire for the gift of smg-1(cc546ts); I. Kashima for technical support and discussions; S. Hirai and T. Hirose for helpful advice; and R. Muramatsu, Y. Bamba, Y. Okada-Katsuhata, and K. Kutsuzawa for technical support. Funding: This work was supported in part by grants from the Japan Society for the Promotion of Science (A.Y., S.O., and N.I.); the Ministry of Education, Culture, Sports, Science and Technology of Japan (S.O.); Japan Science and Technology Corporation (A.Y. and S.O.); Mitsubishi Foundation (S.O.); and the Yokohama Foundation for Advancement of Medical Science (A.Y.). N.I. is a research fellow of the Japan Society for the Promotion of Sciences. Work involving C. elegans was supported by research grant GM50933 (P.A.) from the U.S. NIH. Author contributions: N.I. and A.Y. performed experiments and analysis. A.I., R.K., H.N., and H.H. performed mass spectrometry analysis. B.S. and P.A. performed experiments and analysis of C. elegans. N.I., A.Y., P.A., and S.O. discussed the results and wrote the paper. S.O. and A.Y. supervised the study.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/3/116/ra27/DC1

Fig. S1. RUVBL1, RUVBL2, SMG-10, and RPB5 form a complex.

Fig. S2. Time course analysis of the effect of the RUVBL1 or RUVBL2 knockdown on PIKK abundance.

Fig. S3. Treatment with proteolysis inhibitors did not restore the decreased PIKK abundance caused by the knockdown of RUVBL1 or RUVBL2.

Fig. S4. Knockdown of SMG-10 or RPB5 does not have notable effects on the PIKK abundance.

Fig. S5. RUVBL1 is required for down-regulation of the natural targets of nonsense-mediated mRNA decay.

Fig. S6. RPB5 is involved in nonsense-mediated mRNA decay in mammalian cells.

Table S1. RUVBL1 interacting proteins identified by LC-MS/MS analysis of affinity-purified, SBP-tagged RUVBL1 complexes.

References and Notes

Stay Connected to Science Signaling

Navigate This Article