development of resistance to rnai in mammalian cells

PDFlib PLOP: PDF Linearization, Optimization, Protection

Page inserted by evaluation versionwww.pdflib.com – [email protected]

105

Ann. N.Y. Acad. Sci. 1058: 105–118 (2005). © 2005 New York Academy of Sciences.doi: 10.1196/annals.1359.019

Development of Resistance to RNAi in Mammalian Cells

ZHI-MING ZHENG, SHUANG TANG, AND MINGFANG TAO

HIV and AIDS Malignancy Branch, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892, USA

ABSTRACT: The discovery of RNA interference (RNAi) in C. elegans and inplants has revolutionized current approaches to biology and medicine. RNAi si-lences genes in a sequence-specific manner through the actions of small piecesof double-stranded RNAs (siRNAs and miRNAs). RNAi has been found as awidespread natural phenomenon in eukaryotic cells and is also being used as apowerful experimental tool to explore gene function. Most importantly, it hasmany potential therapeutic applications. Viral gene–specific siRNAs are theo-retically very promising antiviral inhibitors and have been examined in a broadrange of medically important viruses. However, many RNA viruses escapeRNAi-mediated suppression by counteracting the RNAi machinery throughmutation of the targeted region, by encoding viral suppressors, or both. DNAviruses also counteract the RNAi machinery, preferentially using viral sup-pressors. Cellular factors may also contribute to RNAi resistance; ADAR1 wasthe first cellular factor found to be responsible for editing-mediated RNAi re-sistance. Because siRNAs can be used as potent small-molecule inhibitors ofany cellular gene, the best way for a cell to maintain expression of essentialgenes for its long-term survival is to develop a program to resist the detrimentaleffects of RNAi.

KEYWORDS: RNA interference; siRNA; miRNA; gene expression; viralinhibitors; gene therapies

INTRODUCTION

RNA interference (RNAi) has recently been recognized as an important mecha-nism for controlling gene expression at a post-transcriptional level1,2 and regulatingheterochromatin formation for genome stability.3,4 RNAi-mediated post-transcrip-tional gene silencing can restrict expression of certain genes during tissue develop-ment or cell differentiation. Small RNA duplexes, usually 21 to 23 nucleotides (nts)in size, are generated by cleavage of larger double-stranded (ds) RNAs by the ribo-nuclease Dicer.5,6 These small RNA duplexes trigger RNAi by sequence-specificcleavage and degradation of the homologous single-stranded mRNA or by suppres-sion of mRNA translation.1,7,8

Address for correspondence: Zhi-Ming Zheng, HIV and AIDS Malignancy Branch, Center forCancer Research, National Cancer Institute, National Institutes of Health, 10 Center Drive,Room 10S 255, Bethesda, MD 20892, USA. Voice: 301-594-1382; fax: 301-480-8250.

[email protected]

106 ANNALS NEW YORK ACADEMY OF SCIENCES

On the basis of their origin and function, the small RNAs produced by Dicercleavage are classified into two categories: small interfering RNAs (siRNAs) andmicroRNAs (miRNAs). Both siRNAs and miRNAs are ∼21- to 23-bp duplexes bear-ing 2-nt overhangs at the 3′ ends. They differ in that siRNAs originate in the cyto-plasm from cleavage of long dsRNAs with perfect base-pairing,9,10 whereasmiRNAs are derived from cleavage of long, single-stranded primary transcripts con-taining imperfectly matched hairpin-loop structures. In general, primary miRNAtranscripts are cleaved first by the RNase III-like protein Drosha into ∼65 nt stem-loop pre-miRNAs in the nucleoplasm11,12 before being transported by Exportin-5into the cytoplasm,13 where they are further processed by Dicer into ∼21- to 23-bpmiRNA duplexes with imperfect base-pairing. For both siRNAs and miRNAs, onestrand (antisense) is then selectively incorporated into a dsRNA-induced silencing

FIGURE 1. RNAi mediated by synthetic siRNA duplexes depends on active cellgrowth. Efficiencies of synthetic p53 siRNA and non-specific (NS) siRNA duplexes on cel-lular p53 expression were assayed in CaSki cells at 2 × 105 at 20–24 hours (0 time) afterplating by transient transfection by using oligofactamine (Invitrogen). Cell samples wereprepared at 24, 48, 72, 96, and 120 hours after transfection of each siRNA (20 nM) as dia-gramed above and examined by Western blot for p53 by using anti-p53 antibody (Onco-gene). Cells with (lane 7) or without (lane 6) retransfection with either NS or p53 siRNAsat 96 hours of the initial siRNA transfection, as indicated in diagram above, were furthercompared and showed no difference in responding to p53 siRNAs at 120 h after the initialtransfection.

107ZHENG et al.: RNAi RESISTANCE IN MAMMALIAN CELLS

complex (RISC) that contains Argonaute (eIF2C), SMN3, and several otherproteins8,14,15 and guides the selection of a complementary mRNA for cleavage orfor translational inhibition. Recently, Ago 2 was shown to be a key component incleavage of the substrate.14,16 The accumulated evidence so far indicates thatsiRNAs and miRNAs program RISCs equivalently and that they produce similar ef-fects on mRNA expression despite their somewhat different origins.17,18

RNAi can be induced by the introduction of synthetic ∼21- to 23- nt siRNA du-plexes into cells, bypassing the requirement for processing of a long dsRNA medi-ated by Dicer.19 RNAi mediated by siRNAs or miRNAs artificially introduced intomammalian cells has been successfully used to knock down gene expression and tosuppress viral replication, although the outcome of individual RNAi events variesdepending on the designed siRNAs and their targeted regions. Efficient silencing ofa target using the RNAi approach also requires active cell metabolism, with the cellsundergoing exponential growth. In most transient transfection assays of syntheticsiRNA duplexes, one run of RNAi at an effective dose generally confers detectablesuppression of the target for no more than 48–72 hours (FIG. 1). Several runs ofRNAi may be needed to achieve more extensive or longer-lasting suppression. Al-though an individual siRNA’s half-life and its targeted region influence the profi-ciency of RNAi, active RISC function in cells is also necessary to make thesuppression effective. Cells under exponentially active growth both confer efficientuptake of siRNA duplexes and have active RISC function. As shown in FIGURE 1,cells at 96 hours in culture show no response to RNAi (compare lane 7 to lane 6) andthereby are not affected by siRNA-mediated p53 suppression. Although “RNAi re-

TABLE 1. Mechanisms in the development of resistance to RNAi in mammalian cells

aSee references in text.b?, Hypothetic.cSee CMV (cytomegalovirus) reference.76

RNAi resistance by:aMutation in target

regionsProduction ofsuppressors Cellular proteins

RNAviruses

Poliovirus yes

HIV-1 yes

La Crosse yes NSs

PFV-1 Tas

Influenza NS1

Reovirus σ3

DNA viruses

HBV yes

Adenovirus VA1

Vaccinia E3L

HPV16 50kDa(?)b

CMV TRS1, IRS1(?)b,c

RNA editing ADAR1


sistance” usually refers to more active processes, as discussed below, care should betaken to study RNAi and RNAi resistance in cells at the appropriate phase of growth.

Although it has been known for a while in plant virology that many plant viruseshave to counteract RNAi in plant cells in order to invade their hosts, the many reportsof the development of RNAi resistance in mammalian cells came as a surprise. Ap-proximately 22 silencing suppressors have been described in plant and insect virus-es; these have been summarized in several excellent reviews.20–23 In this report, weemphasize the emergence of siRNA resistance as a mechanism that counteractsRNAi-mediated gene silencing in mammalian cells (TABLE 1).

DEVELOPMENT OF RESISTANCE TO siRNA-MEDIATEDVIRAL GENE SILENCING IN RNA VIRUS INFECTIONS

RNA viruses have an RNA genome that can be either single-stranded or double-stranded. Viral RNA replication in infected cells is mediated by a viral RNA–dependent RNA polymerase (RDRP), which produces a dsRNA intermediate phaseeven in single-stranded RNA viruses. Because it lacks proofreading activity, this vi-ral RDRP replicates the RNA genome with low fidelity, resulting in a mutation rateon the order of ∼10−4 per incorporated nucleotide. That is, viral RDRP has an in-herent error frequency on the order of 1 in 104, depending on the virus.24,25 Thus,it is conceivable that RNA viruses under selection pressure may escape siRNA-mediated suppression through mutation of the target regions, thus becoming siR-NA-resistant viruses. This presumption has been supported by accumulating evi-dence from several elegant studies, as described below.

Using RNAi as an antiviral therapy has been considered for many viruses, includ-ing both DNA viruses and RNA viruses. Poliovirus was one of the first to be testedfor siRNA-mediated sequence-specific inhibition. HeLa cells treated with an siRNAspecific for a capsid protein or viral polymerase before infection with a virulent po-liovirus produced only 1%–3% of the virus titer observed in control-treated cells andshowed no virus plaque formation. However, the poliovirus readily escapes thesehighly effective siRNAs through unique point mutations within the targeted re-gions.26 The emergence of siRNA-resistant viruses begins 30–40 hours post-infection. Further analysis of these resistant mutants showed that RNAi recognitionis sensitive to subtle point mutations within the central region and the 3′ end of thetarget RNA. Even a single point mutation resulting in a G:U mismatch is sufficientfor the virus to escape siRNA inhibition.27 Using a pool of siRNAs to simultaneous-ly target multiple sites in the virus genome might be a good choice to prevent theemergence of resistant viruses.

RNAi has been demonstrated to be a powerful tool for suppressing HIV infectionin various transient, short-term assays.28–31 Recently, resistance to RNAi directedagainst HIV-1 has been noticed when siRNA duplexes targeting the viral tat32 ornef33 genes are stably expressed in human T cells for long-term inhibition. RNAi-resistant HIV-1 variants have been identified with deletions or substitutions in thesiRNA target regions. Through these mutations, HIV-1 removes its RNA as a targetand escapes from RNAi-mediated inhibition.

Resistance to siRNA-mediated suppression of La Crosse Virus (LACV) replica-tion has also been noticed after only 72 hours of virus infection in 293T cells pre-


treated with LACV siRNAs.34 Most of the resistant viruses recovered from culturesupernatant show single or double nucleotide changes in the targeted regions, with asignificantly high frequency in the portion of the S segment targeted by the siRNALACS103. More interestingly, a viral nonstructural protein of LACV, NSs, appearsto function as an RNAi suppressor protein. Cells expressing the NSs protein stronglysuppress RNAi directed against both an LACV M segment and a host gene, GAPDH(glyeraldehyde-3-phosphate dehydrogenase).34

The Tas protein from the retrovirus primate foamy virus type 1 (PFV-1) has re-cently been identified as a suppressor of miR-32, an miRNA of cellular origin. ThePFV-1 open reading frame 2 contains a target for miR-32. As a counterdefense tomiR-32-mediated suppression, PFV-1 encodes Tas as a silencing suppressor that en-ables host cell invasion and virus replication.35

DEVELOPMENT OF RESISTANCE TO siRNA-MEDIATEDVIRAL GENE SILENCING IN DNA VIRUS INFECTIONS

Compared to RNA viruses, which are designed to have a high rate of mutationduring their RNA genome replication, the genomes of DNA viruses are relativelystable, with a low mutation rate of 10−8 to 10−11 per incorporated nucleotide. Thus,resistance to RNAi-mediated viral gene silencing though mutation of the target se-quence has not been observed in DNA virus infection. However, siRNA may exertselection pressure on pre-existing resistant mutants, as exemplified by a recent re-port on hepatitis B virus (HBV).36 HBV is a DNA virus, but its DNA replicatesthrough a genomic RNA intermediate and utilizes a virally encoded reverse tran-scriptase. Consequently, a significant amount of diversity, similar to that seen inRNA viruses, occurs in the sequences of HBV isolates; these can be classified intoat least 8 genotypes (A–H). A mutant C genotype that contains a mutation in the 3′portion of the targeted region was isolated and was more resistant to HBVS1 shorthairpin (sh) RNA-mediated RNAi than the wild-type. This mutant can be preferen-tially selected out in the presence of HBVS1 shRNA in Huh-7 cells cotransfectedwith wild-type and mutant HBV-C at a ratio of 9:1 at day 5 post-transfection.36

Adenovirus infection may also result in inhibition of RNAi, perhaps through theactions of a virally encoded noncoding RNA, VA1. Adenovirus VA1 RNA is an∼160-nt structured RNA that is highly expressed in adenovirus-infected cells(∼108 molecules/cell) and plays important roles in increasing viral mRNAtranslation37 and stabilizing ribosome-bound mRNAs.38,39 VA1 RNA enhances viralmRNA translation by blocking activation of protein kinase R (PKR), which is usu-ally induced by viral dsRNA produced during the adenoviral replication cycle.Cullen and colleagues recently used a reporter assay to show that VA1 RNA inhibitsRNAi induced by shRNAs, but does not affect RNAi induced by synthetic siRNAduplexes. VA1 RNA-mediated inhibition of RNAi appears to occur through inhibi-tion of nuclear export of shRNA, competition for Exportin-5, and inhibition of Dicerfunction by direct binding of Dicer.40

Our recent study on shRNA-mediated oncogene silencing of human papillomavi-rus 16 (HPV16) in cervical cancer cells provides evidence on how a DNA virusmight escape RNAi.77 HPV16 contains two viral oncogenes that encode E6, whichtargets p53 for degradation, and E7, which targets pRB. The two oncogenes are tran-


scribed as a single E6E7 bicistronic RNA and are essential genes for the survival ofcervical cancer cells. Transfection of two HPV16-positive cervical cancer cell lines,CaSki and SiHa, with a synthetic siRNA duplex based on a sequence motif of 21 ntsfrom the HPV16 E6E7 bicistronic RNA effectively suppressed both the E6 and E7oncogenes, leading to stabilization of p53 and increased p21 expression. When sta-bly expressed as a short hairpin RNA in these cells under selection pressure, howev-er, the siRNA silencing of E6 and E7 expression was efficient only at early cellpassages, and became inefficient with additional cell passages despite the continuedexpression of siRNA at the same level. This loss of the siRNA function could be du-plicated in cells with stable p53 siRNA, but not in cells with stable lamin A/C siR-NA, suggesting that it is target-selective. The siRNA-resistant cells retain normalsiRNA processing, duplex unwinding and degradation of the unwound sense strand,and RISC formation, suggesting that the loss of siRNA function occurred at a laterstep. Surprisingly, the siRNA-resistant cells were found to contain a cytoplasmicprotein of ∼50 kDa that specifically and characteristically interacted with the E6E7siRNA. These data indicate that a potent siRNA targeting an essential or regulatorygene might induce a cell to develop siRNA resistance through the production of aspecific protein.

VIRAL PROTEINS FROM MAMMALIAN VIRUSESFUNCTION AS SUPPRESSORS OF RNAi INNONMAMMALIAN CELLS AND PLANTS

Several viral proteins expressed from mammalian viruses, the interferon antago-nist proteins of influenza virus NS1 (non-structural protein 1), reovirus -3, and vac-cinia virus E3L, have been examined for their suppressive effects on RNAi in non-mammalian cells.

Influenza viruses are a group of single-stranded RNA viruses and are dividedinto types A, B, and C. The influenza virus NS1 protein is a multifunctional, non-structural protein that contains three functional domains. Two of the domains are inthe N-terminal half of NS1: one is responsible for translational enhancement of theinfluenza virus mRNAs through interactions with eIF4G, a eukaryotic translationinitiation factor, and the other is essential for dsRNA-binding and antagonizing in-terferon induction by sequestering dsRNA from PKR binding and activation. Thethird domain, which inhibits mRNA processing and nuclear-cytoplasmic transportof host mRNAs, is in the C-terminal half of the protein.41 Using a plant silencingsuppression assay, two groups simultaneously demonstrated that the influenza NS1protein is an RNA silencing suppressor in plants and promotes plant virus pathoge-nicity of potato virus X in three different plant hosts.42,43 NS1 suppression of RNAirequires the dsRNA-binding domain of the protein to be functional and to interactwith the siRNAs.42 These observations were further supported by independent evi-dence from the Ding and Palese groups showing that all three NS1 proteins from theinfluenza A, B, and C viruses are RNAi suppressors in a Drosophila cell–based, no-daviral silencing screen assay. This study showed that NS1 from influenza A virusesalso suppresses RNAi in Drosophila cells through its N-terminal dsRNA-bindingdomain and its binding of siRNAs.44


Reoviruses are a group of dsRNA viruses. Reovirus outer shell polypeptide -3 isone of the best-characterized dsRNA binding proteins. Like influenza virus NS1, re-ovirus-3 carries conserved dsRNA-binding motifs and binds dsRNAs in vitro and invivo. Accordingly, reovirus-3 protein sequesters dsRNA from PKR binding andthereby prevents activation by dsRNA. When tested in plant cells, reovirus-3 showedstrong RNAi suppression, although it failed to sequester miRNA precursors.45 Nev-ertheless, the data suggest that the reovirus-3 protein is capable of counteractingRNAi-mediated gene silencing in addition to inhibiting PKR-mediated responses.

Vaccinia virus is a member of the poxvirus family and has a DNA genome thatreplicates in the cytoplasm during viral infection. The vaccinia E3L protein is adsRNA-binding protein46 that inhibits PKR by sequestering dsRNA from PKR, thuspreventing binding.47,48 The C-terminus of the vaccinia virus E3L is responsible forbinding to dsRNA and preventing it from activating the interferon pathway. A recentstudy demonstrated that the E3L protein is a functional suppressor of RNAi inDrosophila cells that inactivates the RNAi silencing–based antiviral response of thecells to flock house virus infection.44

RNA EDITING PLAYS A ROLE IN THE DEVELOPMENTOF siRNA RESISTANCE IN MAMMALIAN CELLS

Double-stranded RNA induces the homology-dependent degradation of cognatemRNA in the cytoplasm via RNAi, but it is also a target for adenosine-to-inosine (A-to-I) RNA editing by adenosine deaminases acting on RNA (ADARs). RNA editingthat affects siRNA-mediated RNAi in vitro was first reported by Chris Smith’sgroup,49 who showed that production of siRNAs could be progressively inhibitedwith increasing deamination of a long dsRNA. This initial observation was immedi-ately supported by a study in C. elegans that showed that A-to-I editing of dsRNAsderived from both transgenes and endogenous genes indeed appeared to prevent theirsilencing by RNAi.50,51 Recent studies further demonstrated a direct interaction be-tween three isoforms of ADARs and siRNA, two of which, ADAR1 and ADAR2,strongly bind siRNA without RNA editing. ADAR1p110, a short form of ADAR1via an alternative translation initiation codon, and ADAR2 also bound a 19-bp siR-NA, but their binding affinities were 15 and 50 times lower than that of ADAR1p150(a full length ADAR1), respectively. ADAR3 bound longer dsRNAs, but failed tobind the 19-bp siRNA. All ADARs that were capable of binding the 19-bp siRNA(ADAR1p150 and p110 and ADAR2) also bound siRNAs containing either 15- or23-bp dsRNA regions. Thus, the length of the siRNA determines whether the boundsiRNA is edited or held in a stable complex without a change of sequence; the criticalsize threshold appears to be 30 bp.52

The cytoplasmic full-length isoform of ADAR1 has the highest affinity forsiRNA among known ADARs, with a subnanomolar dissociation constant. Gene si-lencing by siRNA is significantly more effective in mouse fibroblasts homozygousfor an ADAR1 null mutation than in wild-type cells. This was further supported bythe suppression of RNAi in fibroblast cells overexpressing functional ADAR1, butnot in cells overexpressing mutant ADAR1 lacking double-stranded RNA-bindingdomains. The results provide convincing evidence that ADAR1 is a cellular factorthat limits the efficacy of siRNA in mammalian cells.52


OTHER FACTORS THAT MIGHT LEAD TO RNAiRESISTANCE IN MAMMALIAN CELLS

As described above, resistance to RNAi during viral infection in mammalian cellshas thus far been ascribed to two major mechanisms: mutations in the targeted re-gions and expression of suppressors (TABLE 1). One might wonder whether viruseshave also evolved mechanism(s) to counteract the initiation of the RNAi pathway,rather than to block the pathway’s intermediate components.

This hypothesis has received some preliminary support from a hepatitis delta vi-rus (HDV) study. Data from Taylor’s group indicate that HDV RNAs are resistant toDicer activity.53 Dicer cleaves RNAs that are 100% double-stranded, as well as cer-tain RNAs with extensive but <100% pairing, to release RNAs of ∼21 nt. The circu-lar 1,679-nt genome of HDV and its exact complement, the antigenomic RNA, canfold into a rod-like structure with 74% pairing, but show resistance to Dicer cleav-age. In an in vitro cleavage assay with purified recombinant Dicer, <0.2% ofunit-length HDV RNAs were cleaved, even though Dicer was able to digest frag-ments of the genome. A 66-nt hairpin RNA with 79% pairing was digested with>80% cleavage activity, and a 66-nt hairpin derived from one end of the HDV ge-nome was also digested. The data from these in vitro Dicer digestion experiments areconsistent with the results that small RNAs of ~21 nts cannot be detected duringHDV replication in hepatocytes.53 That unit-length, highly structured HDV RNAsresist Dicer cleavage suggests the some viral RNAs have evolved to prevent initia-tion of RNAi by resisting Dicer. Although HDV itself can be edited by ADAR at ad-enosine 1012 position in the antigenome,54,55 its resistance to Dicer and RNAiappears unrelated to ADAR-mediated editing.

The same group also show that the 1,679-nt genomic and antigenomic RNAs ofHDV are resistant to siRNA-mediated degradation, although its viral mRNAs aresensitive to synthetic siRNA duplexes. It was thought that the base-pairing of HDVgenomic and antigenomic RNAs within the nucleus might make them inaccessibleto siRNA. However, further studies show that HDV circular RNAs transcribed froma DNA template by host RNA polymerase II in Huh-7 cells, which have a similarmRNA structure to mRNAs with a 5′ cap and a 3′ poly A tail, were also resistant tosiRNA attack.56 The mechanism by which HDV becomes resistant to siRNA re-mains unknown.

DOES siRNA ACTIVATE THE PROTEIN KINASE PKR?

PKR, a mammalian dsRNA-dependent serine-threonine protein kinase, has an N-terminal dsRNA-binding domain that can interact non–sequence-specifically withlong stretches of dsRNA, leading to activation. The activated PKR phosphorylatesand inactivates the translation factor eIF2-, resulting in a generalized suppression ofprotein synthesis, followed by cell death via both apoptotic and nonapoptotic path-ways. Activation of this nonspecific pathway could mask any sequence-specific ef-fects that might result from the RNAi pathway. Nevertheless, successful applicationof RNAi has been reported in many mammalian tissue culture cell lines by usingchemically synthesized siRNAs that bypass activation of PKR. It has been docu-mented in the literature that a small blunt-ended dsRNA of less than 33 bp does not


activate PKR.57 In addition, early studies demonstrated that siRNA-mediated inhi-bition of gene expression in mammalian cells is independent of non-specific inter-ference pathways activated by larger dsRNAs.58 However, two recent studiesindicate that, although the RNAi mechanism itself is independent of the interferonsystem, transfection of 21-bp siRNAs or expression of shRNAs from shRNA vectorsdoes trigger an interferon response through PKR.59,60 Thus, these studies argue thatsiRNAs may have broad and complicating effects beyond the selective silencing oftarget genes when introduced into cells.

In an effort to evaluate these contradictory results and search for possible activa-tors of PKR, truncated and full-length versions of PKR were used to select dsRNAsfrom a partially structured dsRNA library containing 1011 sequences.61 This studyidentified a minimal RNA motif for activation of PKR. This motif has a hairpin witha nonconserved, imperfect 16-bp dsRNA stem flanked by 11- to 15-nt single-strand-ed tails. Boundary experiments revealed that the single-stranded tails flanking thedsRNA core provide the critical determinant for activation. A minimum of 11 nts inthe tails is needed for the activation, although the tails could be on the 5′-end, the 3′-end, or both.61 This group also compared dsRNAs with lengths of 24, 33, and 76 bpfor PKR binding and activation and found that a 24-bp dsRNA was unable to activatePKR; instead, it inhibited the kinase. Resolution of the discrepancies among the find-ings of these different study groups will require additional biochemical experiments.

CLOSING REMARKS AND PERSPECTIVES

We are entering an RNAi era. The discovery of RNAi in C. elegans62 and inplants63 has revolutionized today’s approaches to biology and medicine. Throughthe actions of small pieces of dsRNAs (siRNAs and miRNAs), RNAi-mediated genesilencing has been found to be not only a widespread natural phenomenon in eukary-otes, but also a powerful experimental tool to explore gene function. Most important-ly, it has many potential therapeutic applications. Although a great deal has beenlearned in the past few years about RNAi pathways and the functions of their com-ponents, much remains to be understood before we have a complete picture of howRNAi is regulated in mammalian cells. There is absolutely no doubt that RNAi path-ways, just like many other pathways discovered in mammalian systems, will be sub-ject to sophisticated regulation (FIG. 2).

One RNAi function in mammalian cells is thought to be a nucleic acid–based an-tiviral immunity, as evidenced by the recent discovery that cellular miR-32 mediatesantiviral activities in human cells.35 Identification of viral RNAi suppressors in med-ically important RNA and DNA viruses has heralded the beginning of explorationinto existing anti-silencing mechanisms in virus infection. Many of the RNAi sup-pressors identified so far in various RNAi reporter assays turn out to be viraldsRNA– binding proteins (DRBPs). Although fewer cellular DRBPs have been iden-tified than viral DRBPs, one may assume that they will also regulate endogenousmiRNA or siRNA functions for the cell whenever it is needed. About 300 miRNAshave been identified in the genomes of humans, viruses, and non-human eukaryoticspecies, some of which are very abundant (∼ 40,000 copies/cell).64–68 The questionis how to examine a DRBP’s suppressive function in a mammalian system. The sup-pressive functions of viral RNA suppressors in RNAi reporter assays have not been


FIG

UR

E2.

RN

Ai p

athw

ays

and

RN

Ai r

esis

tanc

e. D

espi

te th

eir d

iffe

rent

ori

gins

, bot

h si

RN

A a

nd m

icro

RN

A h

ave

sim

ilar e

ffec

ts o

n m

RN

A e

xpre

ssio

nth

roug

h R

ISC

-med

iate

d ac

tiviti

es. D

evel

opm

ent o

f R

NA

i res

ista

nce

in m

amm

alia

n ce

lls c

ould

be

beca

use

of a

ny p

ossi

ble

bloc

ks (

−) in

RN

Ai p

athw

ays.


recapitulated in any in vivo viral infections in mammals. For example, NS1 of influ-enza A virus behaves as an effective RNAi suppressor in Drosophila and plant RNAireporters, but it seems to play no role in siRNA-mediated inhibition of animal influ-enza infection.69,70

There are at least 388 known eukaryotic proteins with dsRNA-binding domains(DRBDs), 72 of which are human.71,72 More than 20 DRBPs have been identifiedand are reported to function in a diverse range of critically important roles in the cell,including some (such as Dicer and Ago-2) with essential roles in RNAi path-ways.73,74 Since proteins harboring DRBDs have been reported to interact with aslittle as 11 bp of dsRNA, an event that is independent of nucleotide sequence ar-rangement,74 it is conceivable that many cellular DRBPs might actually be involvedin switching on and off the function of a partially double-stranded miRNA that, onaverage, targets hundreds or more mRNAs through pairing of only 6–8 bases.ADAR1 is the only factor of this type that interacts with siRNAs in vitro and limitsthe efficacy of siRNAs in mammalian cells.

Perhaps the most intriguing question is why a natural siRNA has not been foundin eukaryotic cells with or without virus infection. Some RNA viruses have a dsRNAgenome and others carry a single-stranded RNA genome, but produce dsRNA inter-mediates during their replication. Theoretically, these viral dsRNAs could be sub-strates for Dicer digestion to produce siRNAs in the infected cells. However, viraldsRNAs appear to resist Dicer digestion, as exemplified by HDV. This may be be-cause production of viral siRNAs to suppress viral gene expression would be harm-ful to replication of the virus. Similarly, endogenous siRNA production from thehuman genome would be expected, considering that over 20% of human genes pro-duce antisense transcripts that might form sense-antisense pairs (dsRNA),75 but pro-duction of endogenous siRNAs from these human transcripts would not be beneficialfor human gene expression. Discovering how these dsRNAs escape Dicer digestionwill be another challenge.

Understanding viral RNAi resistance has greatly enriched our knowledge of theregulation of RNAi pathways. It has also been greatly informative to know that someviruses escape RNAi-mediated inhibition through one or more counteracting mech-anisms: mutations in the targeted region under siRNA selection pressure and/or pro-duction of RNAi suppressors. However, HPV16 seems to use another method todevelop siRNA resistance, because the proteins that appear to be involved in siRNAresistance in cells expressing HPV16 E6 and E7 are neither viral proteins norADAR1. Identification and characterization of such proteins will be extremely chal-lenging, but will definitely guide us to further understanding of how a cell counter-acts siRNA-mediated events that would adversely affect its own survival. SincesiRNAs can be used as potent small-molecule inhibitors to any cellular gene, the bestway for a cell to keep expression of those genes that are essential for its survival isto develop a program to resist the detrimental effects of RNAi.

[Competing interests: The authors state that they have no competing financialinterests.]

REFERENCES

1. DYKXHOORN, D.M., C.D. NOVINA & P.A. SHARP. 2003. Killing the messenger: shortRNAs that silence gene expression. Nat. Rev. Mol. Cell Biol. 4: 457–467.


2. NOVINA, C.D. & P.A. SHARP. 2004. The RNAi revolution. Nature 430: 161–164.3. PROVOST, P., R.A. SILVERSTEIN, D. DISHART, et al. 2002. Dicer is required for chromo-

some segregation and gene silencing in fission yeast cells. Proc. Natl. Acad. Sci.USA 99: 16648–16653.

4. VOLPE, T.A., C. KIDNER, I.M. HALL, et al. 2002. Regulation of heterochromatic silenc-ing and histone H3 lysine-9 methylation by RNAi. Science 297: 1833–1837.

5. PROVOST, P., D. DISHART, J. DOUCET, et al. 2002. Ribonuclease activity and RNA bind-ing of recombinant human Dicer. EMBO J. 21: 5864–5874.

6. ZHANG, H., F.A. KOLB, V. BRONDANI, et al. 2002. Human Dicer preferentially cleavesdsRNAs at their termini without a requirement for ATP. EMBO J. 21: 5875–5885.

7. HANNON, G.J. 2002. RNA interference. Nature 418: 244–251.8. MARTINEZ, J., A. PATKANIOWSKA, H. URLAUB, et al. 2002. Single-stranded antisense

siRNAs guide target RNA cleavage in RNAi. Cell 110: 563–574.9. HAMMOND, S.M., E. BERNSTEIN, D. BEACH & G.J. HANNON. 2000. An RNA-directed

nuclease mediates post-transcriptional gene silencing in Drosophila cells. Nature404: 293–296.

10. ZAMORE, P.D., T. TUSCHL, P.A. SHARP & D.P. BARTEL. 2000. RNAi: double-strandedRNA directs the ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals.Cell 101: 25–33.

11. LEE, Y., C. AHN, J. HAN, et al. 2003. The nuclear RNase III Drosha initiates microRNAprocessing. Nature 425: 415–419.

12. LEE, Y., K. JEON, J.T. LEE, et al. 2002. MicroRNA maturation: stepwise processing andsubcellular localization. EMBO J. 21: 4663–4670.

13. LUND, E., S. GUTTINGER, A. CALADO, et al. 2004. Nuclear export of microRNA precur-sors. Science 303: 95–98.

14. HAMMOND, S.M., S. BOETTCHER, A.A. CAUDY, et al. 2001. Argonaute2, a link betweengenetic and biochemical analyses of RNAi. Science 293: 1146–1150.

15. MOURELATOS, Z., J. DOSTIE, S. PAUSHKIN, et al. 2002. miRNPs: a novel class of ribonu-cleoproteins containing numerous microRNAs. Genes Dev. 16: 720–728.

16. RAND, T.A., K. GINALSKI, N.V. GRISHIN & X. WANG. 2004. Biochemical identificationof Argonaute 2 as the sole protein required for RNA-induced silencing complexactivity. Proc. Natl. Acad. Sci. USA 101: 14385–14389.

17. DOENCH, J.G., C.P. PETERSEN & P.A. SHARP. 2003. siRNAs can function as miRNAs.Genes Dev. 17: 438–442.

18. ZENG, Y., R. YI & B.R. CULLEN. 2003. MicroRNAs and small interfering RNAs caninhibit mRNA expression by similar mechanisms. Proc. Natl. Acad. Sci. USA 100:9779–9784.

19. ELBASHIR, S.M., J. HARBORTH, W. LENDECKEL, et al. 2001. Duplexes of 21-nucleotideRNAs mediate RNA interference in cultured mammalian cells. Nature 411: 494–498.

20. DING, S.W., H. LI, R. LU, et al. 2004. RNA silencing: a conserved antiviral immunityof plants and animals. Virus Res. 102: 109–115.

21. LI, W.X. & S.W. DING. 2001. Viral suppressors of RNA silencing. Curr. Opin. Biotech-nol. 12: 150–154.

22. VOINNET, O. 2005. Induction and suppression of RNA silencing: insights from viralinfections. Nat. Rev. Genet. 6: 206–220.

23. ZAMORE, P.D. 2004. Plant RNAi: how a viral silencing suppressor inactivates siRNA.Curr. Biol. 14: R198–R200.

24. DOUGHERTY, J.P. & H.M. TEMIN. 1988. Determination of the rate of base-pair substitu-tion and insertion mutations in retrovirus replication. J. Virol. 62: 2817–2822.

25. LEIDER, J.M., P. PALESE & F.I. SMITH. 1988. Determination of the mutation rate of aretrovirus. J. Virol. 62: 3084–3091.

26. GITLIN, L., S. KARELSKY & R. ANDINO. 2002. Short interfering RNA confers intracellu-lar antiviral immunity in human cells. Nature 418: 430–434.

27. GITLIN, L., J.K. STONE & R. ANDINO. 2005. Poliovirus escape from RNA interference:short interfering RNA-target recognition and implications for therapeuticapproaches. J. Virol. 79: 1027–1035.

28. COBURN, G.A. & B.R. CULLEN. 2002. Potent and specific inhibition of human immuno-deficiency virus type 1 replication by RNA interference. J. Virol. 76: 9225–9231.


29. JACQUE, J.M., K. TRIQUES & M. STEVENSON. 2002. Modulation of HIV-1 replication byRNA interference. Nature 418: 435–438.

30. LEE, N.S., T. DOHJIMA, G. BAUER, et al. 2002. Expression of small interfering RNAstargeted against HIV-1 rev transcripts in human cells. Nat. Biotechnol. 20: 500–505.

31. NOVINA, C.D., M.F. MURRAY, D.M. DYKXHOORN, et al. 2002. siRNA-directed inhibi-tion of HIV-1 infection. Nat. Med. 8: 681–686.

32. BODEN, D., O. PUSCH, F. LEE, et al. 2003. Human immunodeficiency virus type 1escape from RNA interference. J. Virol. 77: 11531–11535.

33. DAS, A.T., T.R. BRUMMELCAMP, E.M. WESTERHOUT, et al. 2004. Human immunodefi-ciency virus type 1 escapes from RNA interference-mediated ihhibition. J. Virol. 78:2601–2605.

34. SOLDAN, S.S., M.L. PLASSMEYER, M.K. MATUKONIS & F. GONZALEZ-SCARANO. 2005.La Crosse virus nonstructural protein NSs counteracts the effects of short interferingRNA. J. Virol. 79: 234–244.

35. LECELLIER, C.H., P. DUNOYER, K. ARAR, et al. 2005. A cellular microRNA mediatesantiviral defense in human cells. Science 308: 557–560.

36. WU, H.L., L.R. HUANG, C.C. HUANG, et al. 2005. RNA interference-mediated control ofhepatitis B virus and emergence of resistant mutant. Gastroenterology 128: 708–716.

37. THIMMAPPAYA, B., C. WEINBERGER, R.J. SCHNEIDER & T. SHENK. 1982. AdenovirusVA1 RNA is required for efficient translation of viral mRNAs at late times afterinfection. Cell 31: 543–551.

38. STRIJKER, R., D.T. FRITZ & A.D. LEVINSON. 1989. Adenovirus VA1-RNA regulatesgene expression by controlling stability of ribosome-bound RNAs. EMBO J. 8:2669–2675.

39. SVENSSON, C. & G. AKUSJARVI. 1990. A novel effect of adenovirus VA RNA1 on cyto-plasmic mRNA abundance. Virology 174: 613–617.

40. LU, S. & B.R. CULLEN. 2004. Adenovirus VA1 noncoding RNA can inhibit small inter-fering RNA and MicroRNA biogenesis. J. Virol. 78: 12868–12876.

41. KRUG, R.M., W. YUAN, D.L. NOAH & A.G. LATHAM. 2003. Intracellular warfarebetween human influenza viruses and human cells: the roles of the viral NS1 protein.Virology 309: 181–189.

42. BUCHER, E. H. HEMMES, P. DE HAAN, et al. 2004. The influenza A NS1 protein bindssmall interfering RNAs and suppresses RNA silencing in plants. J. Gen. Virol. 85:983–991.

43. DELGADILLO, M.O., P. SAENZ, B. SALVADOR, et al. 2004. Human influenza virus NS1protein enhances viral pathogenicity and acts as an RNA silencing suppressor inplants. J. Gen. Virol. 85: 993–999.

44. LI, W.X., H. LI, R. LU, et al. 2004. Interferon antagonist proteins of influenza and vac-cinia viruses are suppressors of RNA silencing. Proc. Natl. Acad. Sci. USA 101:1350–1355.

45. LICHNER, Z., D. SILHAVY & J. BURGYAN. 2003. Double-stranded RNA-binding proteinscould suppress RNA interference-mediated antiviral defences. J. Gen. Virol. 84:975–980.

46. DAVIS, M.V., H.W. CHANG, B.L. JACOBS & R.J. KAUFMAN. 1993. The E3L and K3L vac-cinia virus gene products stimulate translation through inhibition of the double-strandedRNA-dependent protein kinase by different mechanisms. J. Virol. 67: 1688–1692.

47. ROMANO, P.R., F. ZHANG, S.L. TAN, et al. 1998. Inhibition of double-stranded RNA-dependent protein kinase PKR by vaccinia virus E3: role of complex formation andthe E3 N-terminal domain. Mol. Cell. Biol. 18: 7304–7316.

48. SHARP, T.V., F. MOONAN, A. ROMASHKO, et al. 1998. The vaccinia virus E3L geneproduct interacts with both the regulatory and the substrate binding regions of PKR:implications for PKR autoregulation. Virology 250: 302–315.

49. SCADDEN, A.D. & C.W. SMITH. 2001. RNAi is antagonized by A→I hyper-editing.EMBO Rep. 2: 1107–1111.

50. KNIGHT, S.W. & B.L. BASS. 2002. The role of RNA editing by ADARs in RNAi. Mol.Cell 10: 809–817.

51. TONKIN, L.A. & B.L BASS. 2003. Mutations in RNAi rescue aberrant chemotaxis ofADAR mutants. Science 302: 1725.


52. YANG, W., Q. WANG, K.L. HOWELL, et al. 2005. ADAR1 RNA deaminase limits shortinterfering RNA efficacy in mammalian cells. J. Biol. Chem. 280: 3946–3953.

53. CHANG, J., P. PROVOST, & J.M. TAYLOR. 2003. Resistance of human hepatitis deltavirus RNAs to dicer activity. J. Virol. 77: 11910–11917.

54. CASEY, J.L. & J.L. GERIN. 1995. Hepatitis D virus RNA editing: specific modificationof adenosine in the antigenomic RNA. J. Virol. 69: 7593–7600.

55. POLSON, A.G., B.L. BASS & J.L. CASEY. 1996. RNA editing of hepatitis delta virusantigenome by dsRNA-adenosine deaminase. Nature 380: 454–456.

56. CHANG, J. & J.M. TAYLOR. 2003. Susceptibility of human hepatitis delta virus RNAs tosmall interfering RNA action. J. Virol. 77: 9728–9731.

57. MANCHE, L., S.R. GREEN, C. SCHMEDT & M.B. MATHEWS. 1992. Interactions betweendouble-stranded RNA regulators and the protein kinase DAI. Mol. Cell. Biol. 12:5238–5248.

58. CAPLEN, N.J., S. PARRISH, F. IMANI, et al. 2001. Specific inhibition of gene expressionby small double-stranded RNAs in invertebrate and vertebrate systems. Proc. Natl.Acad. Sci. USA 98: 9742–9747.

59. BRIDGE, A.J., S. PEBERNARD, A. DUCRAUX, et al. 2003. Induction of an interferonresponse by RNAi vectors in mammalian cells. Nat. Genet. 34: 263–264.

60. SLEDZ, C.A., M. HOLKO, M.J. DE VEER, et al. 2003. Activation of the interferon systemby short-interfering RNAs. Nat. Cell Biol. 5: 834–839.

61. ZHENG, X. & P.C. BEVILACQUA. 2004. Activation of the protein kinase PKR by shortdouble-stranded RNAs with single-stranded tails. RNA 10: 1934–1945.

62. FIRE, A., S. XU, M.K. MONTGOMERY, et al. 1998. Potent and specific genetic interfer-ence by double-stranded RNA in Caenorhabditis elegans. Nature 391: 806–811.

63. HAMILTON, A.J. & D.C. BAULCOMBE. 1999. A species of small antisense RNA in post-transcriptional gene silencing in plants. Science 286: 950–952.

64. CAI, X., S. LU, Z. ZHANG, et al. 2005. Kaposi’s sarcoma-associated herpesvirusexpresses an array of viral microRNAs in latently affected cells. Proc. Natl. Acad.Sci. USA 102: 5570–5575.

65. LIM, L.P., M.E. GLASNER, S. YEKTA, et al. 2003. Vertebrate microRNA genes. Science299: 1540.

66. LIM, L.P., N.C. LAU, E.G. WEINSTEIN, et al. 2003. The microRNAs of Caenorhabditiselegans. Genes Dev. 17: 991–1008.

67. PFEFFER, S., A. SEWER, M. LAGOS-QUINTANA, et al. 2005. Identification of microRNAsof the herpesvirus family. Nat. Methods 2: 269–276.

68. PFEFFER, S., M. ZAVOLAN, F.A. GRASSER, et al. 2004. Identification of virus-encodedmicroRNAs. Science 304: 734–736.

69. GE, Q., L. FILIP, A. BAI, et al. 2004. Inhibition of influenza virus production in virus-infected mice by RNA interference. Proc. Natl. Acad. Sci. USA 101: 8676–8681.

70. TOMPKINS, S.M., C.Y. LO, T.M. TUMPEY & S.L. EPSTEIN. 2004. Protection against lethalinfluenza virus challenge by RNA interference in vivo. Proc. Natl. Acad. Sci. USA101: 8682–8686.

71. LETUNIC, I., R.R. COPLEY, S. SCHMIDT, et al. 2004. SMART 4.0: towards genomic dataintegration. Nucleic Acids Res. 32: D142–D144.

72. STEFL, R., L. SKRISOVSKA & F.H. ALLAIN. 2005. RNA sequence- and shape-dependentrecognition by proteins in the ribonucleoprotein particle. EMBO Rep. 6: 33–38.

73. DOYLE, M. & M.F. JANTSCH. 2002. New and old roles of the double-stranded RNA-binding domain. J. Struct. Biol. 140: 147–153.

74. SAUNDERS, L.R. & G.N. BARBER. 2003. The dsRNA binding protein family: criticalroles, diverse cellular functions. FASEB J. 17: 961–983.

75. CHEN, J. M. SUN, W.J. KENT, et al. 2004. Over 20% of human transcripts might formsense-antisense pairs. Nucleic Acids Res. 32: 4812–4820.

76. CHILD, S.J., M. HAKKI, K.L. DE NIRO & A.P. GEBALLE. 2004. Evasion of cellular anti-viral responses by human cytomegalovirus TRS1 and IRS1. J. Virol. 78: 197–205.

77. TANG, S., M. TAO, J.P MCCOY, JR. & Z.M. ZHENG. Short-term induction and long-termsuppression of HPV16 oncogene silencing RNA interference in cervical cancer cells.Oncogene. In press.

development of resistance to rnai in mammalian cells

Documents