Regulation of RNA interference by Hsp90 is an evolutionarily ...
9
Regulation of RNA interference by Hsp90 is an evolutionarily conserved process Yang Wang a , Rebecca Mercier a , Tom C. Hobman a,b,c , Paul LaPointe a, ⁎ a Department of Cell Biology, Faculty of Medicine and Dentistry, University of Alberta, Canada b Department of Medical Microbiology & Immunology, Faculty of Medicine and Dentistry, University of Alberta, Canada c Li Ka Shing Institute of Virology, Faculty of Medicine and Dentistry, University of Alberta, Canada abstract article info Article history: Received 12 April 2013 Received in revised form 19 June 2013 Accepted 20 June 2013 Available online 1 July 2013 Keywords: Hsp90 RNAi Argonaute Dicer Yeast RNAi is a highly conserved mechanism in almost every eukaryote with a few exceptions including the model organism Saccharomyces cerevisiae. A recent study showed that the introduction of the two core components of canonical RNAi systems, Argonaute and Dicer, from another budding yeast, Saccharomyces castellii, restores RNAi in S. cerevisiae. We report here that a functional RNAi system can be reconstituted in yeast with the in- troduction of only S. castellii Dicer and human Argonaute2. Interestingly, whether or not TRBP2 was present, human Dicer was unable to restore RNAi with either S. castellii or human Argonaute. Contrary to previous re- ports, we find that human Dicer, TRBP2 and Argonaute2 are not sufficient to reconstitute RNAi in yeast when bona fide RNAi precursors are co-expressed. We and others have previously reported that Hsp90 regulates conformational changes in human and Drosophila Argonautes required to accommodate the loading of dsRNA duplexes. Here we show that the activities of both human and S. castellii Argonaute are subject to Hsp90 regulation in S. cerevisiae. In summary, our results suggest that regulation of the RNAi machinery by Hsp90 may have evolved at the same time as ancestral RNAi. © 2013 Elsevier B.V. All rights reserved. 1. Introduction RNA interference (RNAi) regulates gene expression at both the transcriptional and post-transcriptional levels in eukaryotes [1,2]. Long double-stranded RNA (dsRNA) precursors undergo one or more processing steps to generate small interfering RNAs (siRNAs) or microRNAs (miRNAs) that are ultimately incorporated into the RNA-induced silencing complex (RISC). RNaseIII proteins specifically cleave dsRNA and include multiple components of the RNAi pathway [3]. In the miRNA biogenesis pathway, primary miRNA transcripts are processed in the nucleus by the class 2 RNaseIII proteins Drosha and Pasha, into hairpin structures [4,5]. This step is critical to generate 2 nt 3′ overhangs at the ends of hairpins that can be recognized by the PAZ domain of the class 3 RNaseIII protein, Dicer, which then cleaves the double-stranded portion of the hairpin into 21–23 nucleotide dsRNAs [6–10]. Canonical Dicer proteins have been called “molecular rulers” for their ability to cleave dsRNA a fixed distance from the end [7]. The short dsRNAs generated by Dicer bind to a core component of RISC called Argonaute which, depending on the properties of the Argonaute isoform involved, cleaves or displaces the passenger strand [11–15]. This leaves the ssRNA guide strand in the binding cleft of Argonaute and comprises the functional core of the RISC. It is this ssRNA-containing Argonaute complex that targets mRNAs for cleavage or translational suppression. Despite significant advances in our understanding of the RNAi system, the minimum re- quirements for the pathway and how it is regulated are not fully understood. It is estimated that ~ 70% of mammalian genes are regulated by RNAi, but not all eukaryotes possess an RNAi system [16–18]. The budding yeast Saccharomyces cerevisiae is a powerful model organism that has been widely used in genetic and molecular studies of eukaryotic cells, however, it lacks any recognizable homologues of Argonaute or Dicer, and does not carry out RNAi-mediated silencing [18]. Recent studies show that some closely related species of budding yeasts such as Saccharomyces castellii and Kluyveromyces polysporus, are capable of RNAi-mediated silencing [19]. The Argonaute proteins in these fungi share all four identified domains with human Argonaute but also have elongated N-terminal regions, whereas S. castellii Dicer is much shorter than canonical Dicer proteins from humans and the fission yeast Schizosaccharomyces pombe (Fig. 1A and B). RNAi can be restored in S. cerevisiae by introducing only Argonaute and Dicer from either S. castellii or K. polysporus [19,20]. The reconstituted RNAi system si- lences both exogenous reporter genes and endogenous retrotransposons. This suggests that the last common ancestor of these budding yeasts pos- sessed RNAi, but it was lost in some species including S. cerevisiae. The RNAi pathway in humans is more complicated but three components, Biochimica et Biophysica Acta 1833 (2013) 2673–2681 Abbreviations: RNAi, RNA interference; GFP, green fluorescent protein; TRBP2, Tar (HIV)-RNA binding protein 2 ⁎ Corresponding author. Tel.: +1 780 492 1804. E-mail address: [email protected] (P. LaPointe). 0167-4889/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbamcr.2013.06.017 Contents lists available at ScienceDirect Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbamcr
Transcript of Regulation of RNA interference by Hsp90 is an evolutionarily ...
Biochimica et Biophysica Acta 1833 (2013) 2673–2681
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
j ourna l homepage: www.e lsev ie r .com/ locate /bbamcr
Regulation of RNA interference by Hsp90 is an evolutionarilyconserved process
Yang Wang a, Rebecca Mercier a, Tom C. Hobman a,b,c, Paul LaPointe a,⁎a Department of Cell Biology, Faculty of Medicine and Dentistry, University of Alberta, Canadab Department of Medical Microbiology & Immunology, Faculty of Medicine and Dentistry, University of Alberta, Canadac Li Ka Shing Institute of Virology, Faculty of Medicine and Dentistry, University of Alberta, Canada
Abbreviations: RNAi, RNA interference; GFP, green fl
(HIV)-RNA binding protein 2⁎ Corresponding author. Tel.: +1 780 492 1804.
E-mail address: [email protected] (P. LaPoin
0167-4889/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.bbamcr.2013.06.017
a b s t r a c t
a r t i c l e i n f oArticle history:Received 12 April 2013Received in revised form 19 June 2013Accepted 20 June 2013Available online 1 July 2013
Keywords:Hsp90RNAiArgonauteDicerYeast
RNAi is a highly conserved mechanism in almost every eukaryote with a few exceptions including the modelorganism Saccharomyces cerevisiae. A recent study showed that the introduction of the two core componentsof canonical RNAi systems, Argonaute and Dicer, from another budding yeast, Saccharomyces castellii, restoresRNAi in S. cerevisiae. We report here that a functional RNAi system can be reconstituted in yeast with the in-troduction of only S. castellii Dicer and human Argonaute2. Interestingly, whether or not TRBP2 was present,human Dicer was unable to restore RNAi with either S. castellii or human Argonaute. Contrary to previous re-ports, we find that human Dicer, TRBP2 and Argonaute2 are not sufficient to reconstitute RNAi in yeast whenbona fide RNAi precursors are co-expressed. We and others have previously reported that Hsp90 regulatesconformational changes in human and Drosophila Argonautes required to accommodate the loading ofdsRNA duplexes. Here we show that the activities of both human and S. castellii Argonaute are subject toHsp90 regulation in S. cerevisiae. In summary, our results suggest that regulation of the RNAi machinery byHsp90 may have evolved at the same time as ancestral RNAi.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
RNA interference (RNAi) regulates gene expression at both thetranscriptional and post-transcriptional levels in eukaryotes [1,2].Long double-stranded RNA (dsRNA) precursors undergo one ormore processing steps to generate small interfering RNAs (siRNAs)or microRNAs (miRNAs) that are ultimately incorporated into theRNA-induced silencing complex (RISC). RNaseIII proteins specificallycleave dsRNA and include multiple components of the RNAi pathway[3]. In the miRNA biogenesis pathway, primary miRNA transcripts areprocessed in the nucleus by the class 2 RNaseIII proteins Drosha andPasha, into hairpin structures [4,5]. This step is critical to generate2 nt 3′ overhangs at the ends of hairpins that can be recognizedby the PAZ domain of the class 3 RNaseIII protein, Dicer, which thencleaves the double-stranded portion of the hairpin into 21–23nucleotide dsRNAs [6–10]. Canonical Dicer proteins have been called“molecular rulers” for their ability to cleave dsRNA a fixed distancefrom the end [7]. The short dsRNAs generated by Dicer bind to acore component of RISC called Argonaute which, depending on theproperties of the Argonaute isoform involved, cleaves or displaces
uorescent protein; TRBP2, Tar
te).
rights reserved.
the passenger strand [11–15]. This leaves the ssRNA guide strand inthe binding cleft of Argonaute and comprises the functional core ofthe RISC. It is this ssRNA-containing Argonaute complex that targetsmRNAs for cleavage or translational suppression. Despite significantadvances in our understanding of the RNAi system, the minimum re-quirements for the pathway and how it is regulated are not fullyunderstood.
It is estimated that ~70% ofmammalian genes are regulated byRNAi,but not all eukaryotes possess an RNAi system [16–18]. The buddingyeast Saccharomyces cerevisiae is a powerful model organism that hasbeen widely used in genetic and molecular studies of eukaryotic cells,however, it lacks any recognizable homologues of Argonaute or Dicer,and does not carry out RNAi-mediated silencing [18]. Recent studiesshow that some closely related species of budding yeasts such asSaccharomyces castellii and Kluyveromyces polysporus, are capable ofRNAi-mediated silencing [19]. The Argonaute proteins in these fungishare all four identified domains with human Argonaute but also haveelongated N-terminal regions, whereas S. castellii Dicer is much shorterthan canonical Dicer proteins from humans and the fission yeastSchizosaccharomyces pombe (Fig. 1A and B). RNAi can be restored inS. cerevisiae by introducing only Argonaute and Dicer from eitherS. castellii or K. polysporus [19,20]. The reconstituted RNAi system si-lences both exogenous reporter genes and endogenous retrotransposons.This suggests that the last common ancestor of these budding yeasts pos-sessed RNAi, but it was lost in some species including S. cerevisiae. TheRNAi pathway in humans is more complicated but three components,
H. sapiens Dicer (1922 aa)
S. pombe Dicer1(1374 aa)
S. castellii Dicer1 (610 aa)
Helicase dsRBD PAZ RNase IIIa dsRBDRNase IIIb
51 602 630 722 891 1042 1276 1403 1666 1824 1849 1914
Helicase
19 517 537 628 916 1038 1083 1233
dsRBD RNase IIIa RNase IIIb
dsRBDdsRBDRNase III
119 263 274 337 537 599
NTD
H. sapiens Argonaute2 (859 aa)
S. pombeArgonaute1(834 aa)
S.castelliiArgonaute1 (1299 aa)
235 370 517 818
213 348 499 799
585 722 937 1258
PAZ
PAZ
PAZ
N
N
N
MID
MID
MID
PIWI
PIWI
PIWI
A
B
Fig. 1. Domain architecture of Dicer and Argonaute proteins. A. Dicer proteins structure with helicase (light green), dsRNA binding (orange), PAZ (red), and RNaseIII (blue) domainsin ribbon representation. B. Argonaute protein structure with N (purple), PAZ (red), Middle (orange), and PIWI (green) domains in ribbon representation. The amino acid sequenceof Argonaute and Dicer proteins was retrieved from GenBank, and the domain boundaries predicted by SMART [41,42] and UniProt (http://www.uniprot.org/).
2674 Y. Wang et al. / Biochimica et Biophysica Acta 1833 (2013) 2673–2681
Argonaute2, Dicer, and TRBP2, have been shown to be sufficient to recon-stitute the minimal human RISC complex in vitro [21].
The important role that the RNAi system plays in regulating geneexpression necessitates that RNAi itself is subject to regulation.Clues about the regulation of the RNAi machinery have come fromthe observation that the loading of dsRNA duplexes into human andDrosophila Argonaute is ATP-dependent [23]. Moreover, it has beenshown that the hydrolysis of ATP is catalyzed by the Hsc70/Hsp90chaperone machinery during formation of the RISC [23–25]. We andothers have proposed that Hsp90 facilitates a conformational changerequired to accommodate the RNA duplex into the binding cleft ofArgonaute. Importantly, the components of the Hsp90 system arehighly conserved in eukaryotes and many human Hsp90 client pro-teins can be regulated by the yeast Hsp90 system [26,27]. Onlyrecently has the Hsp90 system been identified as a key regulatorof the RNAi system in higher eukaryotes. It is not clear howeverif Hsp90 plays a role in the regulation of distantly related RNAisystems — such as that from S. castellii.
We report here that RNAi can be reconstituted in S. cerevisiaewithS. castellii Dicer and either S. castellii Ago1 or human Argonaute2. Con-trary to a recently published study however, we observed that humanDicer did not support functional RNAi with either Argonaute protein,with or without the co-factor TRBP2. Similar to RNAi in humans andDrosophila, we found that Hsp90 plays an important regulatory rolein reconstituted RNAi systems in budding yeast; consistent with thestructural conservation of Argonaute proteins.
2. Materials and methods
2.1. Plasmids and strains
Silencing constructs were expressed from plasmids pRS403-PGAL1-weakSC-GFP, pRS403-PGAL1-strongSC-GFP and pRS403-PGAL1-
hpSC-URA3 as previously described [19]. Untagged versions ofS. castelli Argonaute and Dicer were encoded by pRS404-PTEF-ScaAgo1and pRS405-PTEF-ScaDcr1 (both acquired from Addgene) as previouslydescribed [19].
We constructed the pRS406-PADH1-GFP(S65T) dual reporter plas-mid (for uracil autotrophy and GFP fluorescence) as follows. TheSacI–KpnI restriction fragment from p414ADH1 [28] was cloned intosimilarly digested pRS406 to make pRS406-PADH1. GFP(S65T) was am-plified by polymerase chain reaction (PCR) with primers designed toincorporate an upstream BamHI and downstream SalI sites. This PCRfragment was cut with BamHI and SalI and cloned into similarlydigested pRS406-PADH1. The resultant reporter plasmid was used tomeasure both URA3 and GFP expression in our studies.
The complete human Ago2 coding sequence was amplified by PCRwith SpeI (upstream) and SalI (downstream) sites and cloned in placeof S. castellii Ago1 in pRS404-PTEF-ScaAgo1 (digested with SpeI andXhoI) to make pRS404-PTEF-hAgo2.
The complete human Dicer coding sequence was amplified by PCRwith SpeI (upstream) and XhoI (downstream) sites and cloned inplace of S. castellii Ago1 in pRS405-PTEF-ScaDcr1 (digested with XbaIand XhoI) to make pRS405-PTEF-hDcr1.
pRS401-PTEF-TRBP2-natNT2 was constructed as follows. The KpnI–SacI fragment from p414TEF [28] was inserted into the similarly cutpRS401 to make pRS401PTEF. PCR-amplified TRBP2 (complete codingsequence with upstream BamHI and downstream XhoI sites) wascloned into the BamHI and XhoI sites of pRS401PTEF to makepRS401-PTEF-TRBP2. Finally, the nourseothricin (natNT2) selectablemarker from pRS41N [29] was amplified by PCR to have an EagI siteupstream of the promoter and downstream of the terminator. Thisfragment was cloned into the EagI site of pRS401-PTEF-TRBP2 tomake pRS401-PTEF-TRBP2-natNT2.
Epitope tagging of Ago and Dcr was done as follows. For humanAgo2 and S. castellii Ago1, complementary primers comprised of an
2675Y. Wang et al. / Biochimica et Biophysica Acta 1833 (2013) 2673–2681
ATG codon and the HA coding sequence was annealed to have over-hangs compatible with SpeI. This small double stranded fragmentwas cloned into the SpeI site in pRS404-PTEF-hAgo2 and pRS404-PTEF-ScaAgo1 and clones with the appropriate orientation were iden-tified by sequencing. The resultant plasmids were called pRS404-PTEF-HA-hAgo2 and pRS404-PTEF-HA-ScaAgo1. For human Dcr1, theC terminal domain of hDcr1 was amplified by PCR with a senseprimer complementary to the sequence near an internal SacII site(5′-gagagaccgcggcagcactccccgggggtcctg-3′) and an antisense primercomplementary to the hDcr1 C-terminus and harboring the myccoding sequence and a XhoI site (5′-gagactcgagctatcacagatcttcttcagaaataagtttttgttcttgggaacctgagg-3′). This fragment was cloned intothe SacII and XhoI sites of pRS405-PTEF-hDcr1 to make pRS405-PTEF-hDcr1-myc. For S. castellii Dcr1, the C terminal domain of ScaDcr1was amplified by PCR with a sense primer complementary to thesequence near an internal NheI site (5′-gagagagctagcaacatccccagtctcagtc-3′) and an antisense primer complementary to the ScaDcr1C-terminus and harboring the myc coding sequence and a XhoIsite (5′-gagactcgagctatcacagatcttcttcagaaataagtttttgttccagattgttgc-3′).
Table 1Yeast strains used in this study.
Strain Genotype
W303a MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15YW1 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS406-PADHYW2 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS406-PADHYW3 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW4 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW5 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-SC-URA3YW6 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW7 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW8 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-SC-URA3YW9 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW10 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW11 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-SC-URA3YW12 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW13 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW14 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-SC-URA3YW15 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW16 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW17 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-SC-URA3YW18 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW19 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS406-PADH1-GFP(S65T)YW20 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS406-PADH1-GFP(S65T) pRS403-PGAL1-SC-URA3YW21 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEFYW22 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS406-PADH1-GFP(S65T)YW23 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS406-PADH1-GFP(S65T) pRS403-PGAL1-SC-URA3YW24 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-weakSC-GFPYW25 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
PGAL1-strongSC-GFPYW26 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-weakSC-GFPYW27 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-strongSC-GFPYW28 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-weakSC-GFPYW29 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS403-PGAL1-strongSC-GFPYW30 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS406-PADH1-GFP(S65T) pRS403-PGAL1-weakSC-GFPYW31 MATa leu2-3,112 trp1-1 can1-100 URA3-1 ade2-1 his3-11,15 pRS404-PTEF
pRS406-PADH1-GFP(S65T) pRS403-PGAL1-strongSC-GFP
This fragment was cloned into the NheI and XhoI sites of pRS405-PTEF-ScaDcr1 to make pRS405-PTEF-ScaDcr1-myc.
All our strains were constructed fromW303a (MATa ura3-1 leu2-3,112, trp1-1, his3-11, 15, ade2-1, can1-100). Table 1 summarizes thegenotypes of all strains used. Briefly, each strain was constructed byintegration of different combinations of pRS403, pRS404, pRS405and pRS406 vectors harboring cloned elements expressed under thecontrol of TEF1 or ADH1 promoters.
2.2. Flow cytometry analysis of GFP expression
Analysis of GFP fluorescence was carried out as previously de-scribed [19,21]. Briefly, each strain was inoculated in synthetic com-plete media with 2% raffinose and grown overnight. Fresh cultureswere then seeded from the overnight cultures (500 uL to 5 mL freshmedia) and cells were grown to log-phase with either 2% glucose(non-inducing) or 2% galactose (inducing). 0.2 OD600 units of cellswere harvested by centrifugation at 960 ×g for 2 min in a micro-centrifuge tube. The cell pellet was washed two times with 1 mL
Reference
[22]1-GFP(S65T) This study1-GFP(S65T) pRS403-PGAL1-SC-URA3 This study-ScaAgo1 pRS405-PTEF-ScaDcr1 This study-ScaAgo1 pRS405-PTEF-ScaDcr1 pRS406-PADH1-GFP(S65T) This study-ScaAgo1 pRS405-PTEF-ScaDcr1 pRS406-PADH1-GFP(S65T) This study
-HA-ScaAgo1 pRS405-PTEF-ScaDcr1-myc This study-HA-ScaAgo1 pRS405-PTEF-ScaDcr1-myc pRS406-PADH1-GFP(S65T) This study-HA-ScaAgo1 pRS405-PTEF-ScaDcr1-myc pRS406-PADH1-GFP(S65T) This study
-HA-hAgo2 pRS405-PTEF-ScaDcr1-myc This study-HA-hAgo2 pRS405-PTEF-ScaDcr1-myc pRS406-PADH1-GFP(S65T) This study-HA-hAgo2 pRS405-PTEF-ScaDcr1-myc pRS406-PADH1-GFP(S65T) This study
-HA-ScaAgo1 pRS405-PTEF-hDcr1-myc This study-HA-ScaAgo1 pRS405-PTEF-hDcr1-myc pRS406-PADH1-GFP(S65T) This study-HA-ScaAgo1 pRS405-PTEF-hDcr1-myc pRS406-PADH1-GFP(S65T) This study
-HA-hAgo2 pRS405-PTEF-hDcr1-myc This study-HA-hAgo2 pRS405-PTEF-hDcr1-myc pRS406-PADH1-GFP(S65T) This study-HA-hAgo2 pRS405-PTEF-hDcr1-myc pRS406-PADH1-GFP(S65T) This study
-HA-hAgo2 pRS405-PTEF-hDcr1-myc pRS401-PTEF-TRBP2-natNT2 This study-HA-hAgo2 pRS405-PTEF-hDcr1-myc pRS401-PTEF-TRBP2-natNT2 This study
-HA-hAgo2 pRS405-PTEF-hDcr1-myc pRS401-PTEF-TRBP2-natNT2 This study
-HA-ScaAgo1 pRS405-PTEF-hDcr1-myc pRS401-PTEF-TRBP2-natNT2 This study-HA-ScaAgo1 pRS405-PTEF-hDcr1-myc pRS401-PTEF-TRBP2-natNT2 This study
-HA-ScaAgo1 pRS405-PTEF-hDcr1-myc pRS401-PTEF-TRBP2-natNT2 This study
-ScaAgo1 pRS405-PTEF-ScaDcr1 pRS406-PADH1-GFP(S65T) This study
-ScaAgo1 pRS405-PTEF-ScaDcr1 pRS406-PADH1-GFP(S65T) pRS403- This study
-HA-ScaAgo1 pRS405-PTEF-ScaDcr1-myc pRS406-PADH1-GFP(S65T) This study
-HA-ScaAgo1 pRS405-PTEF-ScaDcr1-myc pRS406-PADH1-GFP(S65T) This study
-HA-hAgo2 pRS405-PTEF-ScaDcr1-myc pRS406-PADH1-GFP(S65T) This study
-HA-hAgo2 pRS405-PTEF-ScaDcr1-myc pRS406-PADH1-GFP(S65T) This study
-HA-hAgo2 pRS405-PTEF-hDcr1-myc pRS401-PTEF-TRBP2-natNT2 This study
-HA-hAgo2 pRS405-PTEF-hDcr1-myc pRS401-PTEF-TRBP2-natNT2 This study
2676 Y. Wang et al. / Biochimica et Biophysica Acta 1833 (2013) 2673–2681
PBS before final resuspension in PBS for analysis. Cells were analyzedusing the 488 nm laser on a FACSCalibur (BD Biosciences). Data wereprocessed with CellQuest Pro (BD Biosciences) and FlowJo (Tree Star).
2.3. Growth assays
Each strain was grown in synthetic complete media with 2% raffi-nose overnight. Cells were diluted to OD600 of 1.0, and 1:10 serial dilu-tions were spotted onto the appropriate plates (SC, SC-ura, or 5-FOA;containing 2% glucose or galactose) and grown at 30 °C for 3 days.
2.4. Lysate generation and western blotting
Cell lysates were generated as previously described [30]. Proteinlysates were separated by SDS-PAGE and transferred to nitrocellulosefor immunoblotting. Myc-tagged proteins were detected with mousemonoclonal antibody 9E10 (Millipore) [31]. HA-tagged proteins weredetected with rat monoclonal antibody 3F10 (Roche). TRBP2 wasdetected with mouse monoclonal antibody 1D9 (AbCam). Rabbit an-tiserum raised against yeast β-actin was a kind gift from Gary Eitzen(University of Alberta) [30]. Human Ago2 was detected with a rabbitpolyclonal antibody raised against the PAZ domain of human Ago2.
Fig. 2. Schematic for silencing constructs directed against GFP and URA3. A. The strong silinverted repeats of a GFP fragment and forms a dsRNA hairpin structure. The productionThe weak silencing construct directed against GFP (weakSC-GFP) is comprised of two tragalactose-inducible GAL1 promoter (PGAL1) and the antisense transcript (red) is constitutiscripts anneal to form a dsRNA structure. C. The silencing construct directed against URA3forms a dsRNA hairpin structure. The production of this construct is driven by the galactosconstruct used in a previous study with human Dicer, Argonaute2 and TRBP was a singlestructure [21].
Human Dicer1 was detected with mouse monoclonal antibodyN167/7 (Millipore).
3. Results
3.1. Argonaute and Dicer from S. castellii reconstitute RNAi in S. cerevisiae
Previous studies showed that introduction of S. castellii Argonaute(ScaAGO1) and Dicer (ScaDCR1) restores RNAi in S. cerevisiae [19]. Toconfirm this result, we integrated plasmids encoding ScaAGO1 andScaDCR1 respectively into the S. cerevisiae strain, W303a. A dual-reporter plasmid, encoding GFP and URA3, as well as plasmidsencoding galactose-inducible silencing constructs against these re-porter genes were also integrated into W303a. The resultant strains(harboring different combinations of RNAi components) were usedto measure the efficiency of silencing of either GFP using flow cytom-etry, or URA3 using growth assays (Table 1).
The efficiency of GFP silencing was measured by comparing the rel-ative green fluorescence of yeast grown under non-inducing (glucose)and inducing (galactose) conditions for two different silencing con-structs [19]. Expression of the strong GFP Silencing Construct(strongSC-GFP) results in a single RNA transcript that forms a long hair-pin structure in which the ds stem region is comprised of two inverted
encing construct directed against GFP (strongSC-GFP) is a single transcript harboringof this construct is driven by the galactose-inducible GAL1 promoter (PGAL1) [19]. B.nscripts by two convergent promoters. The sense transcript (blue) is driven by thevely driven from the weaker URA3 promoter (PURA3). These two complementary tran-(SC-URA3) is a single transcript harboring inverted repeats of a URA3 fragment and
e-inducible GAL1 promoter (PGAL1) and is similar to the strongSC-GFP. D. The silencing, antisense transcript (red) that anneals directly to the GFP mRNA to form a dsRNA
2677Y. Wang et al. / Biochimica et Biophysica Acta 1833 (2013) 2673–2681
GFP segments (Fig. 2A). Expression of the weak GFP Silencing Construct(weakSC-GFP) produces a ssRNA that is perfectly complementary toanother transcript driven by the URA3 promoter at the opposite end(Fig. 2B). These two promoters drive the transcription of the same cas-sette but in opposite directions which results in the formation of a longdsRNA (Fig. 2B).
Consistent with previous reports, the S. cerevisiae strain express-ing both ScaAGO1 and ScaDCR1 exhibited RNAi-mediated silencing(Fig. 3A) [19]. The GFP fluorescence in yeast expressing ScaAGO1,ScaDCR1 and GFP was reduced when the expression of the weakSC-GFP construct was induced in galactose (Fig. 3A—blue line). Also con-sistent with previous studies, yeast harboring the strongSC-GFP con-struct did not show any GFP fluorescence regardless of inductionconditions [19]. This suggests that the small amount of transcriptionfrom the GAL1 promoter under repressing conditions (glucose) wassufficient to completely silence GFP expression. Under non-inducingconditions, the GFP fluorescence in the strain harboring the weakSC-GFP (Fig. 3A— blue line) was identical to that of a strain that was iden-tical in every sense except that it lacked a silencing construct (Fig. 3A—
filled green plot). When grown in galactose however, induction of thesilencing construct resulted in a significant reduction in GFP fluores-cence (Fig. 3A). In contrast, the strain harboring the strongSC-GFPconstruct did not exhibit detectable GFP fluorescence even underrepressing (glucose) conditions, likely due to the fact that the GAL1 pro-moter is leaky even under glucose conditions [19].
The RNAi-dependent silencing of URA3 was assayed by plating se-rial dilutions of S. cerevisiae cells on agar plates lacking uracil (-Ura),or containing uracil but also the anti-metabolite 5-fluoroorotic acid(5-FOA). The expression of the URA3 gene allows for growth on medialacking uracil but not on media containing uracil and 5-FOA, which isconverted by the URA3 gene product to the toxic 5-fluorouracil. The si-lencing construct targeting URA3 (SC-URA3) contains inverted repeatsof a URA3 gene fragment driven by the GAL1 promoter (Fig. 2C) [19].The single-stranded transcript of the hpSC-URA3 construct forms a hair-pin structure that is comparable to the strongSC-GFP construct.
The growth of yeast expressing ScaAGO1, ScaDCR1 and URA3 wasreduced on media lacking uracil compared to control (i.e. yeast lack-ing the SC-URA3 silencing construct) only when the silencing con-struct was induced with galactose (Fig. 4). Similarly, growth of this
100 101 102 103 104
100 101 102 103 104
100 101 102 103 104
100 101 102 103 104
Relative GFP Fluorescence
Cel
l Co
unt
Relative GFP Fluorescence
Cel
l Co
unt
Relative GFP Fluorescence
Cel
l Co
unt
Relative GFP Fluorescence
Cel
l Co
unt
A B Cglucose glucose
galactose galactose
ScaAgo1 + ScaDcr1 HA-ScaAgo1 + ScaDcr1-myc
Fig. 3. GFP fluorescence in yeast expressing different pairs of Argonaute and Dicer. S. cecastellii. A. Yeast expressing untagged ScaAgo1 and ScaDcr1. B. Yeast expressing HA-ScaApressing HA-HsAgo2, HsDcr1-myc and TRBP2. Each panel shows four traces for GFP fluorereporter (filled gray trace; A—YW3, B—YW6, C—YW9, D—YW18), yeast expressing the GFYW10, D—YW19), yeast harboring both the GFP reporter and the weakSC-GFP construct (GFP reporter and the strongSC-GFP construct (red trace; A—YW25, B—YW27, C—YW29, Dfluorescence intensity.
strain on media containing 5-FOA was improved relative to controlwhen galactose was present to induce expression of the silencingconstruct (Fig. 4). This shows that ScaAGO1 and ScaDCR1 are able toreduce URA3 expression by RNAi-mediated silencing. Consistentwith previous reports, our assays for both GFP fluorescence andgrowth on media lacking uracil showed that ScaAGO1 and ScaDCR1are able to reconstitute RNAi-mediated silencing of reporter genesin S. cerevisiae [19,20].
One major obstacle we encountered in the construction of ourstrains was the loss of integrated genes during successive rounds oftransformation. This was likely due to the high similarity of the back-bone of the plasmids we used, which allows for unwanted homolo-gous recombination that disrupts or discards genes of interest whileleaving the selectable markers intact. Moreover, no commercial anti-bodies are available for ScaAGO1 or ScaDCR1. To overcome this limi-tation, we added epitope tags to ScaAGO1 and ScaDCR1 to allowfor the detection of protein expression by western blot. ScaAGO1was tagged at the N-terminus with the hemaglutinin (HA) tag, andScaDCR1 was tagged at the C-terminus with the myc epitope. Impor-tantly, S. cerevisiae cells expressing the epitope-tagged version ofHA-ScaAGO1 and ScaDCR1-myc were capable of silencing both GFP(Fig. 3) and URA3 (Fig. 4). This demonstrated that the epitope tagsdo not impair the function of ScaAGO1 or ScaDCR1 and allowed us toverify the expression of these proteins in our experiments (Fig. 5).
3.2. Human Argonaute2 and S. castellii Dicer restore RNAi in S. cerevisiae
While conserved in a functional sense, there are structural differ-ences in the Dicer homologues between the RNAi systems of yeastand humans (Fig. 1). To determine the level of compatibility betweencomponents like Argonaute and Dicer, we constructed yeast strains ex-pressing different cross-species, epitope-tagged pairs of these proteins.To further characterize the role of TRBP2 in RNAi, we constructed aseries of yeast strains containing different combinations of human orS. castelli Argonaute and Dicer, with and without TRBP2 (Summarizedin Table 1). Five strains were constructed that express the followingArgonaute–Dicer pairs (with and without TRBP2): (1) HA-HsAgo2,ScaDCR1-myc; (2) HA-ScaAGO1, HsDcr1-myc; (3) HA-ScaAGO1,HsDcr1-myc, TRBP2; (4) HA-HsAgo2, HsDcr1-myc; (5) HA-HsAgo2,
Relative GFP Fluorescence
Cel
l Co
unt
Relative GFP Fluorescence
Cel
l Co
unt
100 101 102 103 104
100 101 102 103 104
100 101 102 103 104
100 101 102 103 104
Relative GFP Fluorescence
Cel
l Co
unt
Relative GFP Fluorescence
Cel
l Cou
nt
Dglucose glucose
galactose galactose
HA-HsAgo2 + ScaDcr1-myc HA-HsAgo2 + HsDcr1-myc + TRBP2
revisiae strains containing different Argonaute and Dicer genes from human and S.go1 and ScaDcr1-myc. C. Yeast expressing HA-HsAgo2 and HsDcr1-myc. D. Yeast ex-scence in yeast grown in glucose (top) or galactose (bottom). Yeast lacking the GFPP reporter but lacking a silencing construct (filled green trace; A—YW4, B—YW7, C—blue trace; A—YW24, B—YW26, C—YW28, D—YW30), and yeast expressing both the—YW31). Each strain was analyzed on a FacsCalibur flow cytometer to measure GFP
WB: α-myc
WB: α-HA
HA-ScaAgo1
ScaDcr1-myc
HA-HsAgo2
HsDcr1-myc
HsTRBP2
++---
-+
-+-
+
-+
--
+
--+
-+
+--
+
--+++
250 kDa-
130 kDa-
100 kDa-
250 kDa-
130 kDa-
No RNAi components
ScaAgo1 + ScaDcr1
--+
-++
HA-ScaAgo1 + ScaDcr1-myc
HA-HsAgo2 + ScaDcr1-myc
HA-ScaAgo1 + HsDcr1-myc
HA-HsAgo2 + HsDcr1-myc
HA-HsAgo2 + HsDcr1-myc + TRBP2
--+
-++
--+
-++
--+
-++
--+
-++
--+
-++
--+
-++
Completemedia
Media-uracil
Media+uracil+5-FOA
100 10-1 10-2 10-3 100 10-1 10-210-310010-1 10-2 10-3
W303aYW1YW2
YW3
YW4
YW5
YW6
YW7
YW8
YW9YW10YW11
YW12YW13YW14
YW15YW16YW17
YW18YW19YW20
Fig. 4. URA3 silencing in yeast expressing different combinations of Argonaute and Dicer measured with a growth assay. S. cerevisiae strains containing different Argonaute and Dicergenes from human and S. castellii. Each group has three strains; yeast lacking both the pRS406-PADH1GFP(S65T) reporter construct and the pRS403-PGAL1-SC-URA3 silencing con-struct, yeast harboring only the reporter construct, and yeast harboring both the reporter and silencing constructs. Each strain was grown on media containing uracil (left),media lacking uracil (centre), and media containing uracil and 5-FOA (right). Cells were grown overnight in appropriate media and then diluted to 1 × 108 cells per ml. 10-foldserial dilutions were prepared and 5 μl aliquots were spotted on indicated plates and grown for 2 days at 30 °C.
2678 Y. Wang et al. / Biochimica et Biophysica Acta 1833 (2013) 2673–2681
HsDcr1-myc, TRBP2. All of these strains were tested for the ability to si-lence GFP and URA3 expression. Interestingly, human Argonaute2(HA-HsAgo2) and S. castellii Dicer (ScaDCR1-myc) restored RNAi-mediated silencing of reporter genes in S. cerevisiae (Figs. 3C and 4).Unlike the Ago–Dicer pair from S. castellii, which completely silencedGFP expression with the strongSC-GFP construct even in glucose, theHA-HsAgo2/ScaDCR1-myc pair was less efficient at silencing GFP ex-pression under these conditions (Fig. 3C). In galactose however, theHA-HsAgo2/ScaDCR1-myc pair efficiently silenced GFP expressioneven in strains expressing the weakSC-GFP construct. Unexpectedly,strains expressing HA-HsAgo2, HsDcr1-myc, and TRBP2, did not exhibitRNAi activity in the GFP fluorescence assay or the URA3-dependentgrowth assays regardless of whether the strongSC-GFP or the weakSC-GFP was employed (Figs. 3D and 4). To rule out the possibility that theepitope tags fused to human HsAgo2 and HsDcr1 interfered with func-tion, we tested strains expressing untagged versions of these proteinsin theURA3 silencing assay.We verified the expression of all three com-ponents by western blot but these proteins were similarly unable to re-constitute RNAi in yeast (Supplemental Fig. 1). These data contrast witha previous study [21] which reported that expression of these humanRNAi components rendered budding yeast RNAi-competent.
WB: α-TRBP2
WB: α-actin
100 kDa-
35 kDa-
55 kDa-
Fig. 5. The detection of Argonaute, Dicer, and TRBP2 proteins by western blot. Anti-HA,anti-myc, anti-TRBP2, and anti-S. cerevisiae β-actin antibodies were used to detect pro-tein expression in yeast lysates. Strains analyzed are indicated above each lane.
3.3. Human and S. castellii Argonaute proteins are subject to Hsp90regulation
Previous work in our lab and others showed that the molecularchaperone Hsp90, regulates conformational changes in Argonaute re-quired to accommodate the loading of dsRNA duplexes [23–25].There are two Hsp90 homologues in S. cerevisiae, the constitutivelyexpressed HSC82 and the inducible HSP82, which each share highidentity with their mammalian counterparts. Importantly, humanHsp90 client proteins can be processed in vivo by the yeast Hsp90
2679Y. Wang et al. / Biochimica et Biophysica Acta 1833 (2013) 2673–2681
system. We therefore questioned whether the functions of humanand S. castellii Argonautes in the reconstituted RNAi pathwayswere regulated by Hsp90. To determine this, we employed theURA3-dependent growth assay in the presence of the potent Hsp90inhibitor, radicicol [32]. Under conditions where silencing constructswere expressed, strains expressing S. castellii Dicer and eitherhuman or S. castellii Argonaute showed reduced silencing of URA3(i.e. increased growth relative to control strain lacking a silencingconstruct) when grown on plates lacking uracil but supplementedwith radicicol (Fig. 6). This suggests that Hsp90 regulates bothhuman and S. castellii Argonaute function in RNAi.
4. Discussion
Since the discovery of RNAi, a great deal of research has focused onthe mechanism and biological significance of this conserved pathway.Clues regarding how individual steps are regulated have emerged butmuch remains unknown. In the present study, we provide evidencethat Argonaute proteins in the most rudimentary RNAi systems areregulated by the Hsp90 chaperone system. This is consistent withthe growing body of evidence suggesting that a significant conforma-tional change must occur in Argonaute proteins to accommodatedsRNAs [33–35]. Hsp90 is known to regulate similar conformationalchanges in steroid hormone receptors that allow for ligand binding[36]. While it is possible Hsp90 regulates the S. castellii Dicer that iscommon in both reconstituted systems, this seems unlikely in lightof the studies on the Drosophila RNAi system [23,24]. Specifically,these reports show that blocking Hsp90 activity does not affect
No RNAi components
ScaAgo1 + ScaDcr1
--+
-++
HA-ScaAgo1 + ScaDcr1-myc
HA-HsAgo2 + ScaDcr1-myc
HA-ScaAgo1 + HsDcr1-myc
HA-HsAgo2 + HsDcr1-myc
HA-HsAgo2 + HsDcr1-myc + TRBP2
--+
-++
--+
-++
--+
-++
--+
-++
--+
-++
--+
-++
Media+ DMSO
100 10-1 10-2 10-3
Fig. 6. Hsp90 inhibitor radicicol alleviates URA3 silencing. S. cerevisiae strains containing dstrains; yeast lacking both the pRS406-PADH1GFP(S65T) reporter construct and the pRS403yeast harboring both the reporter and silencing constructs. Each strain was grown on media(centre left), media lacking uracil with DMSO control (centre right), and media lacking urathen diluted to 1 × 108 cells per ml. 10-fold serial dilutions were prepared and 5μl aliquots
Dicer binding to, or processing of, dsRNA precursors but severely im-pairs Argonaute binding to Dicer products.
We were surprised that human Dicer, even with its cofactorTRBP2, was not able to reconstitute RNAi in S. cerevisiae with eitherS. castellii or human Argonaute given a recent report to the contrary[21]. The explanation for this discrepancy may lie with the nature ofthe silencing constructs we used in our study [19]. Both the strongand weak silencing constructs we employed are double-stranded;the weak silencing construct is comprised of two complementarytranscripts and the strong silencing construct is a single transcriptwith internal complementarity that results in a double-stranded hair-pin structure (Fig. 2A–C) [19]. In contrast, the silencing constructused with human Dicer, Argonaute2 and TRBP2 by Suk and colleaguesis a single-stranded RNA molecule that must anneal directly to theGFP mRNA to form a double-stranded structure (Fig. 2D) [21]. Thisdifference could have a profound impact on how RNAi componentsare recruited to the dsRNA structure and on how it is processed.Regardless, it seems that in our assays, human Dicer, TRBP2 andArgonaute2 are not able to process bona fide RNAi precursors to gen-erate functional RISC.
Several studies illustrate the importance of the 5′ end and the 3′overhang structures that result from cleavage of pre-miRNAs by themicroprocessor complex in the nucleus for subsequent processingby canonical Dicer proteins [7,8,10]. Canonical Dicer proteins bindto the 3′ overhang generated by microprocessor cleavage via thePAZ domain. Dicer then acts as a “molecular ruler” and cleaves at afixed distance from this PAZ domain via its two RNAseIII domainsand generates the characteristic 21–23 nucleotide dsRNA product.Importantly, the silencing constructs we used have large portions of
Media-uracil
+ DMSOMedia
+ 75µM RD
Media-uracil
+ 75µM RD
100 10-1 10-2 10-3 100 10-1 10-2 10-3 100 10-1 10-2 10-3
W303aYW1
YW2
YW3YW4
YW5
YW6
YW7
YW8
YW9
YW10
YW11
YW12YW13
YW14
YW15YW16
YW17
YW18
YW19
YW20
ifferent Argonaute and Dicer genes from human and S. castellii. Each group has three-PGAL1-SC-URA3 silencing construct, yeast harboring only the reporter construct, andcontaining uracil and DMSO control (left), media containing uracil and 75 μM radicicolcil with 75 μM radicicol (right). Cells were grown overnight in appropriate media andwere spotted on indicated plates and grown for 2 days at 30 °C.
2680 Y. Wang et al. / Biochimica et Biophysica Acta 1833 (2013) 2673–2681
non-complementary single-stranded RNA beyond the regions ofcomplementarity that may not be recognized by the PAZ domain ofhumanDicer or, if they are, are too far from the double-stranded portionof the molecule to result in cleavage. In contrast, non-canonical Dicerproteins from yeast lack a PAZ domain (Fig. 1) and are thought tobind to dsRNA substrates as homodimers [37]. Polymerization of theseDicer homodimers along the length of the dsRNA is thought to facilitatecoordinated cleavage required to generate products 21–23 nt in lengthindependent of the structure of the substrate ends. This “inside–out”cleavage mechanismmay allow for the functional versatility associatedwith non-canonical Dicer proteins as they are known to play an impor-tant role in ribosomal and spliceosomal RNA processing [38].
In contrast to Dicer proteins from yeast and humans, Argonauteproteins are structurally much more conserved. Argonaute proteinsfrom humans, S. castellii, and the fission yeast S. pombe share allfour major domains (Fig. 1A). Moreover, two recent publications[20,35] showed the crystal structures of human and K. polysporusArgonaute are very similar. Both proteins adopt a bilobed shapewith an internal cleft to bind dsRNA duplexes. Interestingly, thedifferent domains of Argonaute2 all appear to bind to miRNA in afashion that is conserved with other Argonaute homologues [33].Moreover, Argonaute2 is rendered protease resistant upon miRNAbinding suggesting that the ligand-free protein is less stable and per-haps more conformationally dynamic that the bound form [33]. Thisis important because rapid conformational sampling is one of thefew hallmarks of Hsp90 client proteins. The oncoprotein Bcr–Abl isa well characterized Hsp90 client protein for which drugs are avail-able that bind to and stabilize both the active and inactive states[39]. Intriguingly, stabilization of Bcr–Abl in either conformationweakens its interaction with Hsp90. Therefore structurally dynamic,but not necessarily misfolded, proteins are likely to be Hsp90 clients.Hormone receptors are useful comparisons to Argonaute as an Hsp90client as well. In the absence of ligand, estrogen and glucocorticoid re-ceptors form complexes with Hsp90 in the cytoplasm in a ligand-freestate. Hormone binding requires Hsp90 and results in a ligand-boundbut Hsp90-free complex that translocates to the nucleus to activatedownstream transcription [40]. Similar to what is observed with hor-mone receptors, ligand binding (miRNA in the case or Argonaute) re-quires Hsp90 and also results in stabilization of the protein.
5. Conclusions
In summary, this study presents a powerful model for studying in-dividual RNAi components in a minimal system. Furthermore, datapresented here and from other studies [23,24] suggest that Hsp90 isan evolutionarily conserved regulator of Argonaute proteins, whichmay be critical to modulate post-transcriptional gene silencing in re-sponse to cell stress.
Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbamcr.2013.06.017.
Acknowledgements
Research in the TCH laboratory is supported by a Discovery Grantfrom the Natural Sciences and Engineering Research Council. TCH isthe holder of a Canada Research Chair.
Research in the PL laboratory is supported by a Discovery Grantfrom the Natural Sciences and Engineering Research Council, and anoperating grant from the Canadian Institutes of Health Research. PLis an Alberta Innovates — Health Solutions Scholar.
References
[1] D.P. Bartel, MicroRNAs: genomics, biogenesis, mechanism, and function, Cell 116(2004) 281–297.
[2] G. Meister, T. Tuschl, Mechanisms of gene silencing by double-stranded RNA, Na-ture 431 (2004) 343–349.
[3] B. Lamontagne, S. Larose, J. Boulanger, S.A. Elela, The RNase III family: a conservedstructure and expanding functions in eukaryotic dsRNA metabolism, Curr. IssuesMol. Biol. 3 (2001) 71–78.
[4] A.M. Denli, B.B. Tops, R.H. Plasterk, R.F. Ketting, G.J. Hannon, Processing of primarymicroRNAs by the Microprocessor complex, Nature 432 (2004) 231–235.
[5] R.I. Gregory, K.P. Yan, G. Amuthan, T. Chendrimada, B. Doratotaj, N. Cooch, R.Shiekhattar, The Microprocessor complex mediates the genesis of microRNAs,Nature 432 (2004) 235–240.
[6] J.B. Ma, K. Ye, D.J. Patel, Structural basis for overhang-specific small interferingRNA recognition by the PAZ domain, Nature 429 (2004) 318–322.
[7] I.J. Macrae, K. Zhou, F. Li, A. Repic, A.N. Brooks, W.Z. Cande, P.D. Adams, J.A.Doudna, Structural basis for double-stranded RNA processing by Dicer, Science311 (2006) 195–198.
[8] J.E. Park, I. Heo, Y. Tian, D.K. Simanshu, H. Chang, D. Jee, D.J. Patel, V.N. Kim, Dicerrecognizes the 5′ end of RNA for efficient and accurate processing, Nature 475(2011) 201–205.
[9] A. Tsutsumi, T. Kawamata, N. Izumi, H. Seitz, Y. Tomari, Recognition of thepre-miRNA structure by Drosophila Dicer-1, Nat. Struct. Mol. Biol. 18 (2011)1153–1158.
[10] H. Zhang, F.A. Kolb, L. Jaskiewicz, E. Westhof, W. Filipowicz, Single processing centermodels for human Dicer and bacterial RNase III, Cell 118 (2004) 57–68.
[11] R.I. Gregory, T.P. Chendrimada, N. Cooch, R. Shiekhattar, Human RISC couplesmicroRNA biogenesis and posttranscriptional gene silencing, Cell 123 (2005)631–640.
[12] K. Kim, Y.S. Lee, R.W. Carthew, Conversion of pre-RISC to holo-RISC by Ago2 dur-ing assembly of RNAi complexes, RNA 13 (2007) 22–29.
[13] C. Matranga, Y. Tomari, C. Shin, D.P. Bartel, P.D. Zamore, Passenger-strand cleav-age facilitates assembly of siRNA into Ago2-containing RNAi enzyme complexes,Cell 123 (2005) 607–620.
[14] K. Miyoshi, H. Tsukumo, T. Nagami, H. Siomi, M.C. Siomi, Slicer function ofDrosophila Argonautes and its involvement in RISC formation, Genes Dev. 19(2005) 2837–2848.
[15] J.B. Preall, E.J. Sontheimer, RNAi: RISC gets loaded, Cell 123 (2005) 543–545.[16] R.C. Friedman, K.K. Farh, C.B. Burge, D.P. Bartel, Most mammalian mRNAs are con-
served targets of microRNAs, Genome Res. 19 (2009) 92–105.[17] B.P. Lewis, C.B. Burge, D.P. Bartel, Conserved seed pairing, often flanked by aden-
osines, indicates that thousands of human genes are microRNA targets, Cell 120(2005) 15–20.
[18] B.R. Harrison, O. Yazgan, J.E. Krebs, Life without RNAi: noncoding RNAs andtheir functions in Saccharomyces cerevisiae, Biochem. Cell Biol. 87 (2009)767–779.
[19] I.A. Drinnenberg, D.E. Weinberg, K.T. Xie, J.P. Mower, K.H. Wolfe, G.R. Fink, D.P.Bartel, RNAi in budding yeast, Science 326 (2009) 544–550.
[20] K. Nakanishi, D.E. Weinberg, D.P. Bartel, D.J. Patel, Structure of yeast Argonautewith guide RNA, Nature 486 (2012) 368–374.
[21] K. Suk, J. Choi, Y. Suzuki, S.B. Ozturk, J.C. Mellor, K.H. Wong, J.L. MacKay, R.I.Gregory, F.P. Roth, Reconstitution of human RNA interference in budding yeast,Nucleic Acids Res. 39 (2011) e43.
[22] B.J. Thomas, R. Rothstein, Elevated recombination rates in transcriptionally activeDNA, Cell 56 (4) (1989) 619–630.
[23] S. Iwasaki, M. Kobayashi, M. Yoda, Y. Sakaguchi, S. Katsuma, T. Suzuki, Y. Tomari,Hsc70/Hsp90 chaperone machinery mediates ATP-dependent RISC loading ofsmall RNA duplexes, Mol. Cell 39 (2010) 292–299.
[24] T. Miyoshi, A. Takeuchi, H. Siomi, M.C. Siomi, A direct role for Hsp90 in pre-RISCformation in Drosophila, Nat. Struct. Mol. Biol. 17 (2010) 1024–1026.
[25] J.M. Pare, N. Tahbaz, J. Lopez-Orozco, P. LaPointe, P. Lasko, T.C. Hobman, Hsp90regulates the function of Argonaute 2 and its recruitment to stress granules andP-bodies, Mol. Biol. Cell 20 (2009) 3273–3284.
[26] Y. Xu, M.A. Singer, S. Lindquist, Maturation of the tyrosine kinase c-Src as a kinaseand as a substrate depends on the molecular chaperone Hsp90, Proc. Natl. Acad.Sci. U. S. A. 96 (1999) 109–114.
[27] S.P. Bohen, K.R. Yamamoto, Isolation of Hsp90 mutants by screening for de-creased steroid receptor function, Proc. Natl. Acad. Sci. U. S. A. 90 (1993)11424–11428.
[28] D. Mumberg, R. Muller, M. Funk, Yeast vectors for the controlled expressionof heterologous proteins in different genetic backgrounds, Gene 156 (1995)119–122.
[29] C. Taxis, M. Knop, System of centromeric, episomal, and integrative vectors basedon drug resistance markers for Saccharomyces cerevisiae, Biotechniques 40 (2006)73–78.
[30] H. Armstrong, A. Wolmarans, R. Mercier, B. Mai, P. LaPointe, The co-chaperoneHch1 regulates Hsp90 function differently than its homologue Aha1 and conferssensitivity to yeast to the Hsp90 inhibitor NVP-AUY922, PLoS One 7 (2012)e49322.
[31] G.I. Evan, G.K. Lewis, G. Ramsay, J.M. Bishop, Isolation of monoclonal antibodiesspecific for human c-myc proto-oncogene product, Mol. Cell. Biol. 5 (1985)3610–3616.
[32] S.V. Sharma, T. Agatsuma, H. Nakano, Targeting of the protein chaperone,HSP90, by the transformation suppressing agent, radicicol, Oncogene 16(1998) 2639–2645.
[33] E. Elkayam, C.D. Kuhn, A. Tocilj, A.D. Haase, E.M. Greene, G.J. Hannon, L.Joshua-Tor, The structure of human argonaute-2 in complex with miR-20a, Cell150 (2012) 100–110.
[34] F. Frank, M.R. Fabian, J. Stepinski, J. Jemielity, E. Darzynkiewicz, N. Sonenberg, B.Nagar, Structural analysis of 5′-mRNA-cap interactions with the human AGO2MID domain, EMBO Rep. 12 (2011) 415–420.
2681Y. Wang et al. / Biochimica et Biophysica Acta 1833 (2013) 2673–2681
[35] N.T. Schirle, I.J. MacRae, The crystal structure of human Argonaute2, Science 336(2012) 1037–1040.
[36] P.C. Echeverria, D. Picard, Molecular chaperones, essential partners of steroid hor-mone receptors for activity and mobility, Biochim. Biophys. Acta 1803 (2010)641–649.
[37] D.E. Weinberg, K. Nakanishi, D.J. Patel, D.P. Bartel, The inside–out mechanism ofDicers from budding yeasts, Cell 146 (2011) 262–276.
[38] D.A. Bernstein, V.K. Vyas, D.E. Weinberg, I.A. Drinnenberg, D.P. Bartel, G.R. Fink,Candida albicans Dicer (CaDcr1) is required for efficient ribosomal and spliceosomalRNA maturation, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 523–528.
[39] M. Taipale, I. Krykbaeva, M. Koeva, C. Kayatekin, K.D. Westover, G.I. Karras, S.Lindquist, Quantitative analysis of HSP90-client interactions reveals principlesof substrate recognition, Cell 150 (2012) 987–1001.
[40] D. Picard, Chaperoning steroid hormone action, Trends Endocrinol. Metab. 17(2006) 229–235.
[41] I. Letunic, T. Doerks, P. Bork, SMART 7: recent updates to the protein domain an-notation resource, Nucleic Acids Res. 40 (2012) D302–D305.
[42] J. Schultz, F. Milpetz, P. Bork, C.P. Ponting, SMART, a simple modular architectureresearch tool: identification of signaling domains, Proc. Natl. Acad. Sci. U. S. A. 95(1998) 5857–5864.
