RNA INTERFERENCE
`(RNAi)
`NUTS & BOLTS OF RNAi
`TECHNOLOGY
`
`Edited by
`David R. Engelke, Ph.D.
`
`~ DNA Press™
`
`Alnylam Exh. (cid:20)(cid:19)(cid:23)(cid:24)
`
`

`

`Publisher: Xela Schenk
`Cover Design and Production Layout: Alex Nartea
`
`ISBN 0-9664027-8-2
`
`This book contains information obtained from
`authentic and highly regarded sources. Reprinted
`material is quoted with permission, and sources are
`indicated. A wide variety of references are listed.
`Efforts have been made to publish reliable data
`and information, but the Editor and the publisher
`can not assume responsibility for the validity of all
`materials or for the consequences of their use.
`
`This book was carefully produced.
`Nevertheless, authors and publisher
`do not warrant the
`information
`therein to be free of errors. Readers
`are advised
`to keep in mind that
`statements, data, illustrations, pro(cid:173)
`cedural details, or other printed mat(cid:173)
`ter may inadvertently be inaccurate.
`
`Neither this book nor any part may be reproduced or transmitted in any form or by any
`means, electronic or mechanical, including photocopying, microfilming, and recording,
`or liy any information storage or retricvJL system, without prior permission in writing
`from the publisher.
`
`All rights reserved (including those of translation in other languages). Authorization to
`photocopy items for internal use or personal use, or the personal or internal use of spe(cid:173)
`cific clients, may be granted by DNA Press LLC. The consent of DNA Press LLC does not
`extend to copying general distribution, for promotion, for creating new works, or for
`resale. Specific permission must be obtained in writing from DNA Press LLC for such copy(cid:173)
`ing.
`
`Direct all inquiries to DNA Press LLC. For information and mailing address see
`http://www.dnapress.net
`
`Library of Congress Control Number: 2003111224
`
`©2003 by DNA Press LLC
`
`No claim to original U.S. GovPrnmP.nt works
`International Standard Book Number ISBN 0-9664027-8-2
`Printed in the United States of America
`2 3 4 5 6 7 8 9 0
`First printing 2003
`Printed on acid-free paper
`
`ii
`
`

`

`contents
`
`Introduction .....•.•... .. .....•... .. .................. 9
`Daniel J. Coughlin & David R. Engelke
`
`Section I
`RNA Interference as an Experimental Tool in Invertebrates • . ........ 17
`
`Chapter 1
`Targeted Gene Silencing in Plants Using RNA Interference ......•... 19
`Sasha Preuss and Craig S. Pikaard
`
`Chapter 2
`RNAi in Caenorhabditis elegans . •. .••••••••.•...•...•...... 33
`Maureen M. Barr and Juan Wang
`
`Chapter 3
`Application and Analysis of RNAi in Drosoph;[a Systems . •••.•.. . ... 53
`C. Lipardi, Q. Wei, and 8. M. Paterson
`
`Section II
`Short Interfering RNA as a Tool in Vertebrate Systems:
`Synthesizing the RNA ... . •••• ....•.• •...•••••••. . ..•.•. 69
`
`Chapter 4
`Gene Silencing by Synthetic siRNA Duplexes in Mammalian
`Cell Culture .. ....•. . •••.•.... . .•. .. •.•....•••..••.. 71
`Agnieszka Patkaniowska and Thomas Tuschl
`
`Chapter 5
`Strategies for Synthesizing Small Interfering RNA (siRNA) .......... 89
`Q. Boese, WS. Marshall, and 5.A. Scaringe
`
`Chapter 6
`Generating Long dsRNA, Individual siRNA
`and siRNA Cocktails In Vitro •.....•.. . ........•••......• ,113
`D. Brown, M. W. Byrom, R. Jarvis, V. Polatta, A.M. Cheng and L.P. Ford
`
`iii
`
`

`

`Contents (continued)
`
`Section III
`Short Interfering RNA as a Tool in Vertebrate Systems:
`Applications ..........••..•.. ... ... . ...... . ...... . . 129
`
`Chapter 7
`Optimizing High Throughput RNAi-based Assays Using Transient
`Transfection of Synthetic siRNAs in Cultured Mammalian Cells ... .. . . 131
`£. Krausz, A. Grabner, A. Kroenke, C. Sachse and C.J. Echeverri
`
`Chapter 8
`Subcellular Delivery and Detection of Small Interfering RNA . •.. ... . 169
`Cynthia P. Paul
`
`Chapter 9
`Oownregulation of Cellular Genes by PCR Products Expressing
`siRNAs Mammalian Cells
`. . ... . •.•.• . •.. ........ . .. ..•.• 189
`Daniela Castanatta, John J. Rossi, and Lisa Scherer
`
`Chapter 10
`Generating Adenoviruses for shRNA or siRNA Delivery . . .. . . . ..... 205
`Beverly L. Davidson
`
`Chapter 11
`Achieving Stable, Heritable Gene Silencing in the Mouse • . .. . .. •.. 217
`M. Carmell and G. Hannan
`
`Index .. . . . . . ..... . ... .. .................... .. . . .. . 235
`
`iv
`
`

`

`SECTllil
`
`SHORT INTERFERING RNA .AS A
`TOOL IN VERTEBRATE SYSTEMS:
`SYNTHESIZING THE RNA
`
`'
`
`

`

`RNA Interference (RNAi) - Nuts & Bolts of RNAi Technology
`
`I'
`
`7-{)
`
`

`

`CHAPTER !
`
`GENE SILENCING BY SYNTHETIC
`siRNA DUPLEXES IN
`MAMMALIAN CELL CULTURE
`
`Agnieszka Patkaniowska and Thomas Tuschl
`
`Contact:
`Agnieszka Patkaniowska or Thomas Tuschl
`Laboratory for RNA Molecular Biology, The Rockefeller University,
`1230 York Avenue, Box 186, New York, NY 10021 USA
`Email: patkana@rockefeller.edu, ttuschl@rockefeller.edu
`
`-9664027-8-2/03/$0.00+$.50
`2003 by ONA Press, LLC
`
`From: RNA Interference (RNAi)~Nuts & Bolts of RNAi Technology (pp.71-88)
`Edited by: David R. Engelke, Ph.D.
`
`

`

`RNA fotcrfercnce (RNAi) - Nuts & Bolts of RNAi 1eclmology
`
`4.1 Introduction
`
`RNA interference (RNAi) ·is a sequence-specific post-transcriptional gene
`silencing mechanism, which is triggered by double-stranded RNA (dsRNA) and
`targets degradation of complementary mRNAs. Long dsRNA is cleaved by the
`endonuclease Dicer[1, 2l into 21-23 nucleotide (nt) duplexes with 2 nt 3' over(cid:173)
`hangs[3J (Figure 4- 1, Step 1) . The resulting short interfering RNA (siRNA)
`duplexes are unwound and only one strand of each duplex is incorporated into
`the RNA-induced silencing complex (RISC)f4,5J (Figure 4-1, Step 2). RISC is a
`multiple turnover RNA endonuclease complexf6J (Figure 4-1, Step 5) and one of
`its components is a member of the Argonaute protein family[5J. The single siRNA
`strand in the RISC guides the recognition of the complementary target mRNA
`and determines the position at which the mRNA is cleaved (Figure 4-1, Step 3).
`The target mRNA is cleaved within the region recognized by the antisense siRNA
`strand and the cleavage site is located opposite of the phosphodiester bond
`between 10th and 11th nt of the antisense strand counting from its 5'endl5,7l.
`The mRNA fragments deprived of either cap structure or polyA tail, both
`required for mRNA stability, are subsequently degraded by cellular nucleases
`(Figure 4-1, Step 4) . Consequently, the specific depletion of mRNA from the
`cell leads to a reduction of the corresponding protein, resulting in a knockdown
`phenotype.
`
`The discovery of siRNAs as mediators of RNAi enabled their application in
`mammalian cell biology, bypassing unspecific cellular responses triggered by
`long dsRNAs[8J. In most types of mammalian cells, dsRNA longer than 30 bp
`triggers dsRNA-activated protein kinase (PKR) and 2',5'-oligoadenylate syn(cid:173)
`thetase (2',5'-AS), leading to blockage of protein synthesis initiation and to
`RNase L mediated RNA degradation, respectively (reviewed in:[9J). These path(cid:173)
`ways are common to the interferon response which is ubiquitous in mammalian
`cells with the exception of oocytesf101 and early embryonic cells[2,11J.
`
`The first synthetic siRNA duplexes designed to mimic Dicer products were
`analyzed in Drosophila melanogaster embryo lysate(3l. Soon thereafter they
`were shown to mediate silencing of endogenous genes in mammalian ·cell
`lines[8l. Biochemical studies of the RNAi cleavage activity in Drosophila lysate
`and Hela cytoplasmic extract indicated a high conservation of the RNAi me'th(cid:173)
`anism between flies and mammals[4,5J.
`
`This chapter discusses the design and application of synthetic siRNA
`duplexes for use in mammalian tissue culture. The use of chemica'Lly synthe(cid:173)
`sized siRNA duplexes offers several advantages over alternative RNAi protocols
`presented in other chapters: (1) The sequence of synthetic siRNAs is well
`defined, in contrary to the siRNAs obtained by enzymatic prqcessing.
`
`72
`
`

`

`CHAPTER 4
`Gene Silencing by Synthetic siRNA Duplexes in Mammalian Cell Culture
`
`111111 lllllllll lllllf llll II II II 111111111111111111111111111111111111111111111 IIIIII
`
`., ____ __
`__________ .,
`111 lltlllllll 1111111111
`
`Step1 l ~~ ~------
`
`1111ll111111111111111"
`"
`
`s· •
`s· ,.
`,. ________ ., 6' • •
`nl 111111111111111111"
`
`Step 21 :w.onautea, ?
`
`Rl8C
`
`Steps
`
`Figure 4-1: Outline of the RNAi mechanism. Long dsRNA is processed by the endonu(cid:173)
`clease Dicer into siRNA duplexes (Step 1). The duplexes are unwound and protein com- ·
`ponents of RISC assemble on one of siRNA strands (Step 2). RISC targets the comple(cid:173)
`mentary mRNA (Step 3). RISC cleaves the target (T), releases the cleavage products
`(Step 4) and recycles (Step 5).
`
`73
`
`

`

`RNA lnterfenmce (RNAi) - Nuts & Bolts of RNAi Tech110Logy
`
`(2) Transfection of siRNA duplexes allows for the precise adjustment of the min(cid:173)
`imal siRNA concentration necessary to obtain a silencing effect. This is not eas(cid:173)
`ily accomplished with siRNA expression systems. (3) siRNA duplexes do not
`trigger the interferon response as observed for the siRNA expression s_ystems(121.
`
`4.2 Structure
`
`siRNA duplexes composed of two 21 nt RNA strands forming a 19 nt base(cid:173)
`paired region with 2 nt 3' overhangs were found to be optimal for mediating
`RNAi in Drosophila lysatePI and in mammalia n cell culturel81. In mammalian·
`systems, however, RNAi can be induced with a wide variety of short dsRNA
`agents, such as 18-23 bp siRNA duplexesll3,14l with or without 2 nt over(cid:173)
`hangs[1sJ or 19-29 bp short hairpin RNAs (shRNAs)l16l.
`siRNAs produced by
`terminH3,17l.
`Dicer cleavage contain 5'-phosphate and 3'-hydroxyl
`Phosphorylation of the 5'ends, however, does not seem to be crucial for cell cul(cid:173)
`ture applications[1U6,1BJ since free 5'...'.hydroxyl groups of RNA molecules are rap(cid:173)
`idly phosphorylated by endogenous kinases[4,SJ.
`
`RNAi can also be triggered by single-stranded siRNAs at high doses. Single(cid:173)
`stranded siRNAs can associate directly with the RISC protein components
`bypassing the unwinding step of the siRNA duplex(4,51. Unmodified single(cid:173)
`stranded RNAs, however, are generally too unstable for t he routine use in cell
`culture[16,18l. On the contrary, siRNA duplexes were shown to be unexpectedly
`resistant to nucleases contained in serum (stable for 48h in cell culture medi(cid:173)
`um supplemented with serum[19l) and in cytoplasmic extract (stable for at least
`2h in Hela cytoplasmic extract(51).
`
`4.3 Nucleotide Composition
`
`RISC activity is dependent on the sequence complementarity between the
`siRNA residing in the RISC and the cytoplasmic mRNA. The effect of mismatches
`depends on their nature and position with respect to the cleavage site. Single
`mismatches located in the middle of the guide siRNA are more inhibitory than
`mismatches at the.ends of the siRNA strand[l6,20·22l. Additionally, mismatches
`close to the 5'end of the antisense siRNA are more disruptive than those at the
`3'end[1B,23J. This is most likely related to the fact that the 5'end of the guide
`siRNA serves as a 1'ruler" by defining the position of mRNA cleavage[3,5J (Fig.11
`Step 3). It is probable, however, that some mismatches are better tolerated
`than othersl16l considering that RNA can support many different non-canonical
`base-pairings (e.g. G/U, U/G, C/A, A/C, U/U, C/U, U/C). The sequence of the
`antisense 3' overhang does not play a significant role in target recognitionl3J.
`In Drosophila lysate, only the 20th nt (counting from the 5'end of antisens·e
`strand) slightly contributes to the efficiency of RNAWl.
`
`74
`
`

`

`CHAPTER 4
`Gene Silencing by Synthetic siRNA Duplexes in Mammalian Cell Culture
`
`The overall G/C content of a siRNA sequence does not appear to influence
`silencing efficiency under standard siRNA application conditions, since duplex(cid:173)
`es with 30 to 80% G/C content effectively silenced various target genes[14J.
`However, it is advisable to avoid more than 3 guanosine nucleotides in a row as
`they tend to form G-quartet structuresl14l. Recent studies in Drosophila have
`suggested that the relative strength of base-pairing at the 5'end of either siRNA
`strand partially determines which strand will be incorporated into RISC(21i,s21. A
`helicase activity involved in separating siRNA strands appears to preferentially
`unwind weak base-pairs (2 hydrogen bonds, A/U or U/A) rather than strong
`ones (3 hydrogen bonds, G/C or C/G). It is suggested that the siRNA strand for
`which the 5'end is most readily unwound, preferentially accumulates into RISC.
`The complementary strand is presumably degraded. This recent data suggests
`that siRNAs should be designed to have a lower G/C content at the 5'end of the
`antisense strand than at the 5'end of the sense strand. Adaptation of these rec(cid:173)
`ommendations may increase silencing efficiency as well as minimize off-target
`effects caused by the sense strand (see Selection).
`
`4.4 Target mRNA
`
`RNAi seems to occur predominantly in the cytoplasm[4,s,2s1 and only a
`mature mRNA should be considered a target for siRNAs. Sequence information
`regarding mRNAs can be retrieved from public databases, e.g. NCBI
`(www.ncbi.nlm.nih.gov). When analyzing mRNA sequence data, it is important
`to consider the expression pattern of the gene of interest in order to assure that
`the target is present in the cells used for the experiment. One should also take
`into account the presence of any splice· variants since siRNAs complementary to
`skipped exons will not be effective. Annotated single nucleotide polymorphisms
`(SNP) should be avoided since even a single nucleotide mismatch with the tar(cid:173)
`get sequence can drastically reduce the silencing efficiency of the siRNA (see
`Nucleotide Composition). The selected siRNA may still not be effective due to
`discrepancies between the annotated and the ·actual sequence of the targetl161.
`
`The design of a powerful siRNA is still an empirical process, although with
`a reasonably high chance of success independently of the mRNA region being
`targeted (ORF or UTR)l16l. siRNA duplexes show various silencing efficiencies,
`which ~an be explained by either the nucleotide composition of the siRNA
`duplex itself or by the targeted position within the mRNA. The positional effect
`can be attributed to the inaccessibility of the target site due to the secondary
`structure of mRNA or to binding of regulatory proteins. Prediction of the RNA
`secondary structure with software such as mfold[261 does not correlate. with the
`potency of an siRNA[16,27, 28l, although some investigators like to suggest the
`oppositel29l. Experimental data on rnRNA accessibility from classical antisense
`
`75
`
`') I
`
`

`

`RNA fn1e1.ferencc (RNAi) - Nitts & Bolts of RNAi Technology
`
`or ribozyme applications may be found useful when designing siRNAs[21, 30·32).
`However, this approach is neither time nor cost effective and does not guaran(cid:173)
`tee success[28, 311.
`
`4.5 Selection
`
`Each siRNA duplex should be tested in silica to avoid silencing of multiple
`genes (off-target effects). The antisense strand of the siRNA duplex can be
`checked by a nucleotide BLAST-search using appropriate parameters
`(www.ncbi.nlm.nih.gov/BLAST) to discard siRNA sequences with a high degree
`of complementarity to mRNAs other than the target. The number of mismatch(cid:173)
`es to a non-target mRNA should be 3 or more[33l, located preferentially in the
`middle or towards the 5'end of the guide siRNA. Contiguous stretches of homol(cid:173)
`ogy extending over 15 nt should be avoided, especially at the 5'end of the
`guideltB,231,
`
`To perform a search with the query sequence of 21 nt it is necessary to
`change the filters, word size and expect value threshold for the standard BLAST
`program. Low complexity filtering should be removed since it considers most
`of the short sequence insignificant, resulting in little or no query sequence
`remaining. The word size should be set to 7 letters, the minimal value for the
`algorithm used by this search. It is necessary to increase the expect value
`threshold (e.g. to 1000) since there is a high probability for a short sequence
`to occur by chance in the database. The expect value depends ort the size of a
`database, so the threshold can be lowered when the search is restricted to a
`subset of the database (from the Description of BLAST services, NCBI). The
`parameters mentioned above are preset in the ''B LAST search for short nearly
`exact matches" option.
`
`The choice of the source database is very important, since it can consider(cid:173)
`ably shorten the time of a performed search and further facilitate analysis of
`the results. The ideal database would be a non-redundant, comprehensive col(cid:173)
`lection of mRNA sequences expressed in the cell Line or tissue of the interest.
`Unfortunately, available resources are still far from that stage. For the moment,
`the optimal choices offered by NCBI are Reference Sequences (RefSeq) and
`UniGene databases. RefSeq is a non-redundant set of for the most part curat(cid:173)
`ed sequences, containing mRNA sequences and UniGene is a collection of gene(cid:173)
`oriented clusters of rnRNA and EST sequences. To facilitate the siRNA selection
`process there is some software available offering search interfaces for the most
`unique siRNA sequences within the mRNA of interest, e.g. a non-commercial
`website of the Whitehead Institute, MIT at: http://jura.wi.mit.edu/bioc/
`siRNA/home.php.
`
`76
`
`

`

`CHAPTER 4
`Gene Silencing by Synthetic siRNA Duplexes in Mam,malian Cell Culture
`
`4.6 Verification
`
`The possibility of facing some unspecific silencing effects cannot be exclud(cid:173)
`ed even when si RNAs are carefully selected. This has been recently demon(cid:173)
`strated by microarray profiling studiesl23l. Not all off-target effects caused by
`siRNAs may be seen at the mRNA level. siRNAs can act like endogenously encod(cid:173)
`ed miRNAs, which inhibit translation by hybridizing to partially complementary
`sequences in the 3'UTR of mRNAs(34J. However, since rniRNAs usually require
`several such target sites in the mRNA, translatfonal inhibition is probably not a
`prominent off-target effect of siRNAs.
`
`Any knockdown phenotype obtained via RNAi can be due to both intended
`and off-target effects. To avoid drawing wrong conclusions from the RNAi
`experiment, two or more different siRNA duplexes targeting the same mRNA
`should be independently tested for a reproducible phenotype[33J. The ultimate
`proof for the specificity of a silencing experiment can be provided by a com(cid:173)
`plementation approach (resc.ue experiment)[33J. The silencing is considered to
`be specific, when the knockdown phenotype can be reversed by the introduc(cid:173)
`tion of a vector encoding the targeted gene. There are two approaches to a res(cid:173)
`cue experiment[35l: (1) the siRNA duplexes can be directed aga.inst 3'UTR of the
`endogenous mRNA, which is lacking in the rescue vector; .(2) silent mutations
`at the third position of a few codons can be introduced within the targeted
`region of the vectorf36l.
`In both cases, the wild-type phenotype should be
`observed when co-transfecting the rescue vector and siRNA specific for the
`endogenous copy of the gene. The knockdown phenotype should remain when
`co-transfecting the rescue vector and siRNA against a sequence common for the
`endogenous and introduced gene copy. For a well established co-transfection
`protocol see[14l. An alternative to cDNA complementation is a rescue by intro(cid:173)
`duction of the recombinant form of the depleted protein into the siRNA treat(cid:173)
`ed cells, e.g. by microi njection(371.
`
`4. 7 Transfection
`
`siRNA duplexes must be introduced into the cytoplasm in order to initiate
`silencing effects. For single cell study purposes, siRNA duplexes can be rnicroin(cid:173)
`jected into the cellsf38, 39l and the knockdown phenotype microscopically evalu(cid:173)
`ated. For studying cell populations, however, introduction of siRNAs by either
`chemical transfection or electroporation are the methods of choice.
`Transfection efficiency and the level of cytotoxicity resulting from different
`siRNA delivery methods depend on the particular cell type being used[401. For
`this reason, the method and conditions of siRNA delivery should be optimized
`individually for all new cell types under investigation .
`
`77
`
`

`

`RNA bzte1formce (RNAi) - Nuts & Bolts of RNAi Technology
`
`Chemical transfection is a convenient and widely used method for intro(cid:173)
`ducing siRNA duplexes into adherent cells. The most commonly used transfec(cid:173)
`tants
`are
`cationic
`lipids,
`such
`as Oligofectamine
`(Invitrogen),
`Lipofectamine2000 (Invitrogen) and TransIT-TKO (Mirus) or calcium phos(cid:173)
`phate[411. There are however many other new reagents that are commercially
`available. For the standard transfection protocols seel14l or follow the manu(cid:173)
`facturer's instructions. The ratio of siRNA duplex to transfection reagent is
`important for both the efficiency and toxicity of the transfection procedure.
`Excess of either of the components may have negative effects on the formation
`of the RNA/carrier complex or may result in the presence of either component
`in the non-complexed state. This may be toxic to the cells or decrease the
`uptake of siRNAs.
`
`Electroporation is well suited for cells in suspension and can be readily
`adapted for adherent cells[40-43l. This method delivers nucleic acids directly into
`the cytoplasm and does not require coupling of siRNAs with any carrier. The
`optimal parameters for the electroporation of siRNAs differ from those for plas(cid:173)
`mids by allowing the use of milder conditions that are less toxid41l.
`
`There is no universal transfection technique. Cells of different origin or
`morphology can vary with respect to cellular components involved in the
`processes that allow for the transition of siRNAs from the cell q.1lture medium
`into the RISC. The duplex may fail to cross the cellular membrane, get trapped
`in endosomal structures or remain stably associated with the transfection
`reagent. Any of the above scenarios would contribute to reduction of transfec(cid:173)
`tion efficiency.
`
`Titration of the siRNA concentration for both chemical transfection and
`electroporation is highly recommended, as it provides a graded readout of the
`silencing effect!14J. This allows the identification of th.e minimal effective
`siRNA concentration, which reduces the probability of causing off-target
`effects[20.44J. RNAi has been shown to occur at siRNA concentrations as low as
`100 pM in the cell culture medium!8l. Nevertheless, the minimal effective siRNA
`concentration as well as the kinetics of the silencing effect can vary for differ(cid:173)
`ent transfection techniques and cell types[411.
`
`Every siRNA transfection experiment needs well-defined controls. The neg(cid:173)
`ative control aims to correct for the side effects that the transfection procedure
`exerts on the cells (see Silencing Effect). The siRNA duplex serving as a nega(cid:173)
`tive control should target a sequence that is not expressed in the transfected
`cell line, e.g. luciferase, Gf P[1~l. Positive controls ~re used to assess the trans(cid:173)
`fection efficiency and thus should provide a dear and quantifiable knockdown
`phenotype (e.g. Lamin A/C). When possible, it is recommended to ,quantify the
`
`78
`
`

`

`CHAPTER 4
`Gene Silencing by Synthetic s1RNA Duplexes in Mammalian Celt Culture
`
`level of silencing for the positive control and the experimental sample by the
`same methods (see Silencing Effect). For a rough estimation of the transfec(cid:173)
`tion efficiency, it is possible to set up the positive control so that it produces
`an easy to spot phenotype that can be microscopically observed soon after
`transfection, e.g. f?i-actin depletion causing cellular blebbing[45l, EgS depletion
`resulting in loss of cell adhesionl41,45J or any other essential protein knockdown
`driving cells into apoptosis.
`
`Although cornmonly used cell transfection methods are very robust, they
`still require knowledge of some basic cell culture routine. To provide repro(cid:173)
`ducible results, cells must be systematically split at the proper cell density and
`used only during early passages. Silencing experiments should be conducted at
`constant cell culture densities. Low cell densities during the transfection pro(cid:173)
`cedure can lead to an increase in cytotoxicity. Cell densities that are too high
`may lead to low transfection efficiencies. For chemical transfections, cells
`should typically be plated 24 hours prior to transfection at a density that results
`in confluency at the time of experimental readout. When ' choosing the initial
`cell density, it is important to consider that even low toxicity reagents impair
`the growth rate. Systematic monitoring of the cell morphology by phase con(cid:173)
`trast microscopy before and after transfection provides valuable info(mation
`about the experiment.
`
`4.8 Silencing Effect
`
`lhe rate at
`The direct effect of RNAi is depletion of t he targeted mRNA.
`which it occurs depends on the initial abundance and turnover rate of the par(cid:173)
`ticular mRNA. Reduction of the corresponding protein levels depends on the
`rate of mRNA depletion, initial protein concentration and its stability. The
`onset of the silencing effect depends on the rate of release of the siRNA duplex
`into the cytoplasm (see Transfection). Silencing effects monitored at the mRNA
`level are expected to appear earlier than those at the protein level. Using
`cationic lipid-based transfection, mRNA depletion is near completion within 24
`hours, and reduction of protein level becomes readily detectable after 40
`hours[16J: For the electroporation based delivery, silencing effects may occur
`earlier, because of immediate delivery of siRNA into the cytop-Lasm(43J. The
`silencing effect may last for 96 hours or longer(14.40).
`
`For validation of the silencing effect, standard cell, molecular biology and
`biochemistry methods are used. Methods for quantification of mRNA level
`include Northern hybridization, quantitative RT-PCR, real time PCR or branched
`DNA[461. For PCR based analysis, it is recommended to position the primers more
`than 50 nt upstream of the siRNA targeting site. The mRNA from knockdown
`and control cells should be purified using a polyA-specific rnRNA isolation kits
`
`79
`
`

`

`RNA lntc1fe1·cnce (RNAi) - Nuts & Bolts of llNAi Techuology
`
`in order to remove the (sometimes stable) target RNA 5' cleavage product prior
`to the amplification reactionf16l. Immunodetection methods are used to quah(cid:173)
`titate the level of protein, and can be performed on whole cells (immunocyto(cid:173)
`chemistry, flow cytometry (FACS)) or cell lysates (Western analysis, immunopre(cid:173)
`cipitation (IP) and Enzyme Linked Immunosorbent Assay (ELISA)). Methods
`based on cell lysates rather than intact cells provide information about the
`average silencing effect for a cell population, while the whole cell readouts pro(cid:173)
`vide information about the nature of the silencing effect, i.e. what percent of
`the cell population shows the knockdown phenotype and to what extent.
`
`4.9 Chemical Modifications
`
`siRNA duplexes can be chemically modified and depending on the position
`of modification terminal and internal modification are distinguished. Terminal
`modifications are substitutions at the 5' or 3' position of a terminal ribose
`(Table 4-1). siRNAs are typically terminally modified to change their stability,
`uptake and cellular distribution or to serve as affinity tags in biochemical stud(cid:173)
`ies. Modifications designed to stabilize siRNAs against exonucleases are often
`nucleotide analogs that are c0nnected to the siRNAs via non-natural 5',5' or
`3',3' phosphodiester linkages (inverted nucleotides). Non-nucleotidic reagents
`are usually conjugated to the siRNAs via aminolinkers. When using siRNA
`duplexes linked to fluorophores it is important to keep in mind, that the cellu(cid:173)
`lar uptake of fluorescently labeled siRNAs is not necessarily correlated with their
`efficiency of gene silenci ng06l.
`
`While terminal modifications at both 5' and 3'end of the sense siRNA strand
`are very well tolerated(S,15,16.471, it is more difficult to predict the effect for the
`antisense strand. Blocking of the 5' or the 3'end of the antisense siRNA has
`interfered with the function of siRNAs[S,15,16,471. In biochemical assays, it was
`shown that the 3'end of the antisense siRNA is permissive to modification15l,
`but when some 3' modified antisense siRNAs were tested in cultured cells, they
`did not suppress the target gene[16l.
`
`Internal modifications of siRNA duplexes comprise substitutions of the 2'(cid:173)
`hydroxyl group of the ribose and modifications of the phosphate backbone
`(Table 4-2; for recent review seel48l). Nucleotides with modified bases can also
`be introd,uced[49l. Internal modifications are generally used to modulate the
`thermodynamic stability and conformation of the helix of "'siRNAs and their
`resistance against endonucleases. Among recently tested modifications are 2'(cid:173)
`deoxy, 2'-0-methyl and 2'-fluoro-2'-deoxy substitutions at the ribose and phos(cid:173)
`phorothioate modifications of the backbone.
`
`80
`
`

`

`CHAPTER 4
`Gene Silencing by Synthetic siRNA Duplexes in Mammalian Cell Culture
`
`The most commonly used internal modification is substitution of two over(cid:173)
`hanging ribonucleotides at the 3'ends of siRNAs by deoxyribonucleotides, which
`enables reducing costs of RNA chemical synthesis with no effect on RNAi effi(cid:173)
`ciency[B,14, 15,271. 2'-Deoxythymidine overhangs (TT) have become the most wide(cid:173)
`ly used since their first successful applicationf8J. The effect of modifications on
`RNAi efficiency depends on position and number of modified nucleotides with(cid:173)
`in the base-paired region of the siRNA duplex[1s,19,49,S0J. Modifications of the
`sense siRNA strand of the duplex interfere with the RNAi efficiency less than
`modifications of the antisense strand[15,49l. Also modifications located in the 5'
`half of the antisense siRNA are more disruptive than those located in its 3'
`halfl49J.
`
`2'-fluoro-2'-deoxyribose is better tolerated than 2'-O-methylribose, presum(cid:173)
`ably° because the fluoride has properties similar to the replaced oxygen in the
`2'-hydroxyl group. However, introduction of 2'-O-methyl modifications increase
`stability of siRNA duplexes in seruml15J and prolong the silencing effect in the
`cell culture[15,18l. A phosphorothioate backbone does not seem to impair the
`function of siRNAs, but the introduction of siRNAs with many sulfur-modified
`linkages may lead to cytotoxic effects[16J,
`
`By careful selection of modifications and adjustment to the specific target(cid:173)
`ing sequence it should be possible to generate fully nuclease resistant yet active
`siRNAs, which may prove useful for in vivo applications. Conditions for optimal
`pharmacokinetic and pharmacodynamic distribution of siRNAs remain to be
`established. Conjugation of non-nudeotidic moieties to the siRNAs may be use(cid:173)
`ful for modulating these parameters.
`
`Acknowledgements
`
`We thank Y. Dorsett, M. Mitkovski, M. Landthaler, G. Meister, M. Lagos(cid:173)
`Quintana and J . Martinez for critical comments and help in editing of the man(cid:173)
`uscript.
`
`81
`
`

`

`RNA lnte,ference (RNAi) - Nuts & Bolts of RNAi Technology
`
`Table 4-1.
`Terminal modifications of siRNA duplexes.
`Scale of the silencing effect as compared to the efficiency of unmodified siRNA
`duplex: ++++, > 80% of efficiency of unmodified duplex; +++, 60-80%; ++, 40-
`60%; +, 20-40%; -, modification rendering the duplex inactive.
`
`l~oqifjcacion
`sense strand termini
`
`.. l .:·. ·
`G
`~ ene sum~ng
`
`• . ~~~t~I~.·•
`,'c': -_ '
`:Sy'~t.e
`. -.... •:m • Ref~rent:~ ~ ~
`
`Hela extract
`Hela
`H.ela
`
`Hel a
`
`Hela extract
`Hel a
`Hela
`
`HaCaT
`Hela
`
`Hela
`
`[5]
`[15, 47]
`[47]
`
`[16]
`
`(5]
`[15]
`[47]
`
`[27]
`[16]
`
`(15]
`
`Hela extract
`Hela
`Hela
`Hela
`
`.
`
`[5]
`[15, 47)
`[16]
`[15]
`
`aminolinker
`
`++++
`
`puromycin or
`biotin
`fluorescei n .
`antisense strand 3' terminus
`aminolinker
`
`++++
`
`++++
`
`++++
`
`++++
`
`+++
`-
`
`++++
`
`puromycin or
`biotin
`fluorescei n
`fluorescein or
`Alexa488
`inverted deoxy
`abasic cap
`antisense strand 5' terminus
`aminolinker
`
`-
`
`fluorescein
`inverted deoxy
`abasic cap
`
`++++
`-
`
`82
`
`

`

`CHAPTER 4
`Gene Silencing by Synthetic siRNA Duplexes in Mammalian Cell Culture
`
`Table 4-2.
`Internal modifications of si RNA duplexes.
`Scale of the silencing e