RNA interference: Little RNAs, Big Impact

1
RNA interference Little RNAs, Big Impact

• Lecture 5.3
• Stephanie Minnema
• University of Calgary

2
What Well Cover

• What is RNAi/ useful terms
• Brief history of RNAi
• Biogenesis and mechanisms of action
• Applying RNAi to model systems
• Endogenous RNAi miRNA in the genome
• New frontiers for RNA

3
What is RNA Interference (RNAi)

• The Process by which dsRNA silences gene
  expression... Mittal, 2004
• Generally Post transcriptional level (PTGS)
• Degradation or translation inhibition
• Blazing hot topic...
• 368 published articles this year!
• A field with many unknowns

4
Handy RNAi Terms

• dsRNA double stranded RNA, longer than 30 nt
• miRNA microRNA, 21-25 nt.
• Encoded by endogenous genes.
• Hairpin precursors
• Recognize multiple targets.
• siRNA short-interfering RNA, 21-25 nt.
• Mostly exogenous origin.
• dsRNA precursors
• May be target specific

5
An Arbitrary Distinction?

• miRNA vs. siRNA?
• Discovered in different ways
• Similar biogenesis
• Share common pathway components
• Common pathway outcomes
• Understanding of miRNA comes from research on
  siRNA and vice versa
• Maybe current understanding does not allow us to
  distinguish them
• Some use terms interchangeably

6
A Brief History of RNAi
7
An Unexpected Result

• Late 1980s, Rich Jorgenson and group
• chalone synthase for deeper purple petunias
• Got white and variegated
• Co-suppression both endogenous and introduced
  genes silenced Napoli et al., 1990
• PTGS but what is the causative factor?
• Similar effects seen in N. crassa Quelling
  Cogoni et al., 1996

From Gura,2000.
8
Later, in the C. elegans Camp...

• Antisense RNA injection method for gene
  inactivation
• 1995 characterization of Par1 by Sue Guo
• Essential for embryo polarity
• Did antisense Par1 RNA injections
• Results in embryonic lethality
• Sense Par1 RNA injections gave same result!
• Remained a mystery...
• The basis for the sense effect is under
  investigation and will not be discussed
  further.. Guo and Kemphues, 1995

9
Some Sharp Reasoning

• Fire and Mello, 1998
• Both sense and antisense RNAs sufficient for
  silencing
• Silencing can persist, even though RNA is easily
  degraded
• RNA for silencing often generated using
  bacteriophage RNA polymerases
• Specific, but can also make ectopic transcripts
• Maybe some dsRNA in these preparations?
• Could dsRNA be mediating a new silencing
  mechanism?

10
Their Experiment

• C. elegans Unc-22 inactivation
• Null phenotype uncoordinated twitching
• Injected sense, antisense, or both into c.
  elegans gut
• dsRNA was orders of magnitude more effective than
  ssRNA
• Effective even in tiny amounts
• Unc-22 null phenotype also seen in progeny of
  injected worms
• Inactivation was due to degradation of target
  mRNA
• Coined the term RNA interference Fire et al.,
  1998

11
siRNA Identified

• 25bp species of dsRNA found in plants with
  co-suppression Hamilton and Baulcombe, 1999
• Not in other plants
• Sequence similar to gene being suppressed
• Drosophila long dsRNA triggers processed into
  21-25bp fragments Elbashir et al., 2001
• Fragments short interfering RNA (siRNA)
• siRNA necessary for degradation of target

12
Meanwhile, Back in C. elegans

• Discovery of the first miRNA, lin-4
• Non-coding, 22nt RNA
• Identified in screen for defects in timing of
  larval development
• lin-4 mutation ectopic larval stage 1-like cell
  divisions at later stages
• lin-14 mutations reciprocal phenotype, same
  regulatory pathway as lin-4
• lin-4 negatively regulates lin-14 translation
• lin-4 partially complementary to conserved sites
  in lin-14 3UTR Lee et al., 1993
• Required for negative regulation of lin-14
• lin-4 binds these sites

13
Unique Occurrence?

• No other miRNAs found for 7 years!
• Second miRNA let-7 Reinhart et al., 2000
• Non coding, 21nt RNA
• Regulates lin-14 in same way as lin-4
• Maybe miRNA is in other organisms?
• Homologues of lin-4 escaped bioinformatics
• Let-7 Homologs were easily detected Pasquinelli
  et al., 2000
• Drosophila, sea urchins, mice, humans...
• Indicates RNAi general conserved mechanism

14
An Ancient Process

• Predates evolutionary divergence of plants and
  worms Novina and Sharp, 2004
• Silencing of viruses and rogue genetic elements
• Aberrant RNAi pathway function inability to
  suppress some mobile genetic elements
• Plants Tabara et al, 1999
• C. elegans Xie et al, 2004
• Weve come a long way...
• miRNA and siRNA same mechanism
• Increasingly detailed knowledge

15
Biogenesis and Mechanism of RNAi
16
From Nature Reviews Web Focus on RNAi
17
RNAi Two Phase Process

• Initiation
• Generation of mautre siRNA or miRNA
• Execution
• Silencing of target gene
• Degradation or inhibition of translation

18
Initiation
Execution
He and Hannon, 2004
19
The RNA Components

• siRNA and miRNA
• How do they arise?
• What are their characteristics?

20
miRNA Biogenesis

• Transcribed from endogenous gene as pri-miRNA
• Primary miRNA long with multiple hairpins
• Imperfect internal sequence complementarity
• Cleaved by Drosha into pre-miRNA
• Precursor miRNA 70nt imperfect hairpins
• Exported from nucleus
• Cleaved by Dicer into mature miRNA
• 21-25nt
• Symmetric 2nt 3 overhangs, 5 phosphate groups

Novina and Sharp, 2004
21
siRNA Biogenesis

• Dicer cleaves long dsRNA into siRNA 21-25nt
• dsRNA from exogenous sources
• Symmetric 2nt 3 overhangs, 5 phosphate groups
• Evidence for amplification in C. elegans and
  plants
• Allows persistence of RNAi?

Novina and Sharp, 2004
22
The Protein Components

• What are they?
• How do they function?
• We think...

23
Drosha

• Processes pri-miRNA into pre-miRNA
• Leaves 3 overhangs on pre-miRNA
• Nuclear RNAse-III enzyme Lee at al., 2003
• Tandem RNAse-III domains
• How does it identify pri-miRNA?
• Hairpin terminal loop size
• Stem structure
• Hairpin flanking sequences
• Not yet found in plants
• Maybe Dicer does its job?

24
Dicer

• Cleaves dsRNA or pre-miRNA
• Leaves 3 overhangs and 5 phosphate groups
• Cytoplasmic RNAse-III enzyme
• Functional domains in Dicer Bernstein et al.,
  2001
• Putative helicase
• PAZ domain
• Tandem RNAse-III domains
• dsRNA binding domain
• Multiple Dicer genes in Drosophila and plants He
  and Hannon, 2004
• Functional specificity?

25
RNA Dependent RNA Polymerase (RdRP)

• RdRP activity found in plants and C. elegans
• May explain efficiency of RNAi
• Required for RNAi?
• Not found in mammals or drosophila
• RdRP deficient plants and worms... Results not
  decisive
• Random degenerative PCR Lipardi et al., 2001
• Proposed mechanism
• siRNA acts as primer for elongation on target
  mRNA
• Result more long dsRNA

26
RNA Induced Silencing Complex (RISC)

• RNAi effector complex
• Critical for target mRNA degredation or
  tranlslation inhibition
• Not well characterized 4 subunits? More?
• Activities associated with RISC
• Helicase
• Endonuclease and exonuclease Slicer (or is it
  Dicer?)
• homology seeking/RNA binding
• Preferentially incorporates one strand of unwound
  RNA Khvorova et al., 2003
• Antisense
• How does it know which is which?

27
RISC Preference for Antisense RNA

• Helps ensure specificity for target
• 5 stability of siRNA and miRNA duplex strands
  often different
• The strand with less 5 stability usually
  incorporated into RISC Schwarz et al., 2003
• Due to easier unwinding from one end?
• If strand stability is similar (rare), strands
  incorporated at similar frequency He and Hannon,
  2004

28
Argonaute (Ago)

• Consistently co-purifies with RISC Hammond et
  al., 2001
• Homology seeking activity?
• Binds siRNA and miRNA Ekwall, 2004
• Distinguishes antisense strand Novina and Sharp,
  2004
• Multiple Ago family proteins
• Different RISCs?
• Tissue specific? Developmentally regulated?
• Evidence for different RISCs Tijsterman et al.,
  2004
• Drosophila Dicer1 vs Dicer2/R2D2
• Inhibition vs. degradation Lee et al., 2004

29
Getting the Job Done

• Translational inhibition
• Transcript degradation

30
Translational Inhibition

• Imperfect match between siRNA or miRNA in RISC
  and target mRNA
• RISC usually binds 3 UTR
• Mechanism of inhibition... ????

He and Hannon, 2004
31
mRNA Degradation

• Perfect complementarity between siRNA or miRNA in
  RISC and the target mRNA
• Cleavage by RISC Slicer activity
• Could be Dicer?
• Other endo/exonucleases?
• Recruitment of other components?

Novina and Sharp, 2004c
32
Other Effects?

• RNAi process also work on transcriptional level?
  Volpe et al., 2002
• Plants, C. elegans, Drosophilia
• Via chromatin modification Mochizuki et al.,
  2002
• Heterchromatin formation machinery fairly well
  characterized
• Whats the connection?
• Argonaute
• RITS complex RNA-induced initiation of
  transcriptional silencing Verdel et al., 2004
• RITS could mediate targeted heterchromatin
  formation

33
RITS Connects RNAi and Heterchromatin Formation
Machinery
Novina and Sharp, 2004
34
Applying RNAi to Model Systems
35
Why All the Hype?

• Quick way to do loss-of-function studies
• Targeting takes long time, lots of work
• Not all loci amenable to targeting
• Cheap

36
RNAi in plants, C. elegans, Drosophila

• Introduction of dsRNA sufficient for RNAi
• In vitro transcription
• Chemical synthesis
• Remarkably straightforward C. elegans
• Feed E.coli expressing dsRNA Timmons and Fire,
  1998
• Soak them in dsRNA Tabara et al., 1998
• Common methods transfection or microinjection of
  dsRNA
• Effect lasts days
• Passed onto daughter cells/progeny

37
Great Potential

• Whole genome RNAi screening
• What do all the proteins do?
• Knock each down!
• Done in C. elegans
• 19 757 protein coding genes (predicted)
• 16 757 inactivated using RNAi
• Ravi Kamath et al., 2003
• New standard for systematic genome wide
  functional studies

38
Generation of Bacterial Feeding Library

• C. elegans primer set from Research Genetics
• 19 213 primer pairs each for protein coding gene
• Generated PCR products
• Cloned into dual promoter vector
• Both sense and antisense strands transcribed
  under induction conditions
• Result 16 757 bacterial strains
• 86.3 of predicted genes
• Remaining PCR failures, cloning failures

39
Induction of RNAi
Tuschl, 2003
40
Assayed Phenotypes Examples

• Unc uncoordinated
• Clr clear
• Prz paralyzed
• Lon long
• Mlt moulting defects
• Egl egg laying defects
• Him high incidence of males

• Emb embryonic lethal
• Ste sterile
• Gro slow growth
• Adl adult lethal
• Lvl larval lethality
• Lva larval arrest
• Bmd body morphological defects

41
Results

• 10 of targeted genes gave obvious phenotypes
• Highly conserved genes most likely to give
  aberrant phenotype
• More likely to be essential
• DNA synthesis, cell cycle control
• New genes unlikely to have detectable phenotype
• Lots of gene duplications
• Very specialized or redundant functions

42
Results

• Examined domains in proteins knocked down
• Non-lethal phenotypes from more recent domains
• Animal specific domains
• Ex Immunoglobulin-like repeats
• Non-viable phenotypes from inactivation of
  proteins with ancient domains
• Domains shared with plants and lower eukaryotes
• Domains needed for survival evolutionarily
  preserved
• Genomic clustering of genes yielding phenotypes
• Common origins or regulatory mechanisms?

43
One Step Further...

• Ashrafi et al. (2003) used same RNAi library
• Screened for particular phenotypic readout using
  a cellular marker
• Interested in fat storage regulation
• Found 417 genes involved in fat storage
• Many conserved new obesity drug targets?
  Tuschl, 2003
• Repeated RNAi with mutant lines
• Known defects in fat storage
• Allowed new genes to be placed in fat regulation
  pathways

44
How About Mammals?

• Application of RNAi to mammalian system promising
  for functional studies
• Evidence of RNAi in mammals was harder to
  establish
• Methods for RNAi not a straightforward

45
Non-Specific Silencing via Antiviral Pathway
McManus and Sharp, 2002
46
Getting Around the Problem

• Critical observation Elbashir et al., 2001
• Size matters
• siRNA (21-22nt) mediate mammalian RNAi
• Introducing siRNA instead of dsRNA prevents
  non-specific effects
• Application via transient transfection
• Dont see persistent or propagative effect as in
  C. elegans etc.
• No RdRP activity identified
• Chemically synthesized
• In vitro transcription

47
Empirical siRNA Design Rules

• 21nt long, with 2nt 3 overhangs
• Avoid introns and UTRs
• Avoid regions gt50 GC content
• Use stringent BLAST to help ensure specificity
• Limitations
• Inability to interact with RISC
• Target inaccessibility (structural constraints?)
• Instability of the siRNA

48
Still Not Too Efficient

• Usually need to design several siRNAs to get an
  effective one
• Could use a mixture of siRNAs
• Recombinant Dicer available
• Use in vitro to cleave dsRNA
• Problems
• Increased possibility of non-specific targeting
• Low effective siRNA concentration
• Dont know which siRNA is most potent

49
Rational Design of siRNA

• Arising from research on RISC assembly
• RISC contains one strand of the siRNA duplex
  Martinez et al., 2002
• Needs to be the antisense strand to find right
  target
• Can we direct preferential incorporation of the
  antisense strand into RISC?
• Observation 5 end of an siRNA strand is
  incorporated into RISC most efficiently Schwarz
  et al., 2003

50
Rational Design Points
Mittal, 2004
51
Stable RNAi in Mammals

• Vector driven methods
• Expression of sense and antisense siRNA
• Stable production of siRNA with 3 overhangs
• Expression of pre-miRNA like RNAs
• RNA that folds into hairpin loops with 3
  overhangs
• Act like pre-miRNA dicer substrates
• Some evidence for induction of interferon
  response? Bridge et al., 2003 Sledz et al.,
  2003
• Could do inducible, time, and tissue specific
  RNAi
• Therapeutic potential
• Effective delivery an issue...

52
Endogenous RNAi miRNA in the Genome
53

• miRNAs might have a general role in regulating
  gene expression in diverse developmental and
  physiological processes, and (there are)
  substantial hints that mis-regulation of miRNA
  function might contribute to human disease
• He and Hannon, 2004

54
Genome Wide miRNA Identification

• Most has been done experimentaly
• Cloning and sequencing
• Over 100 novel miRNAs identified from C. elegans,
  Drosophila, and mammals
• Highly conserved, particularly 5 end
• All from hairpin precursors
• Expected to represent 1 of predicted genes Lim
  et al., 2003
• Same as other gene families with regulatory roles
• 200-255 miRNAs in humans
• gt175 have now been experimentally confirmed
  Griffiths-Jones, 2004

55
Prediction of Vertebrate miRNA

• Method MiRscan Lim et al., 2003
• Looks for sequences that form pre-miRNA hairpin
  structures
• Sequences must also be highly conserved
• Score sequences on basis of characteristics of
  experimentally characterized miRNAs
• Revealed 188 human loci.
• Set contained 81 of 109 previously known human
  miRNAs (74)
• If 188 is 74 of total miRNA in human genome,
  then 255 miRNAs expected

56
The miRNA Registry

• Hosted by Rfam
• http//www.sanger.ac.uk/Software/Rfam/
• Searchable database of miRNAs Griffiths-Jones,
  2004
• Both validated and predicted entries
• Each entry shows a hairpin, with the mature miRNA
  indicated
• Establishes a system for miRNA annotation Ambros
  et al., 2003

57
Functional Characterization

• What does each miRNA regulate?
• Critical!
• Lewis et al., (2003) estimate average of five
  mRNA targets per miRNA
• Thousands of proteins may be regulated by miRNA

58
Prediction of miRNA Targets

• Fairly straightforward in plants Rhoades et al.,
  2002
• miRNAs almost perfectly complementary to their
  targets
• Methods search for near-perfect matches in 3UTRs
• Also look for conservation of target sites
• Many targets transcription factors
• miRNAs regulate the regulators
• Suggests major role in highly regulated processes

59
Target Prediction in Vertebrates

• Not so easy miRNAtarget not always highly
  complementary
• Method for plants tried in other organisms
• Results same as would be expected by chance
• Some knowledge to start with
• Interaction of 5 end of miRNA and the target
  most critical
• Target binding sites likely conserved
• Target binding sites common in 3 UTR

60
Target Prediction in Vertebrates

• TargetScan Lewis et al., 2003
• http//genes.mit.edu/targetscan
• Based on several criteria
• Perfect complementarity between target 3 UTR and
  most 5 heptamer of miRNA
• Conservation across species
• Favorable structural and thermodynamic duplex
  formation between target and miRNA
• Predicted 451 targets
• 400 non-redundant

61
Target Validation

• Luciferase reporter system in HeLa cells
• Tested 15 predicted targets
• 11 validated
• Long way to go...

62
New Frontiers for RNA
63
Just Scratching the Surface

• Small RNAs likely to have bigger impact on gene
  and protein regulation
• New classes of small RNAs
• Tiny non-coding RNA Ambros et al., 2003
• tncRNA 20-22nt
• Discovered in C. elegans
• Not likely generated from hairpin loops
• Not conserved among species
• Many complementary to mRNAs
• Function unknown

64
RNA as a Molecular Switch

• New class of RNA can act as a switch specifying
  cell fate
• Small Modulatory RNA smRNA Kuwabara et al.,
  2004
• Discovered in mice
• Conserved in vertebrates
• Interacts with regulatory protein
• Turns transcriptional repressor into activator

65
Fate Specification by smRNA

• Neuronal Restricted Silencing Element (NRSE)
  NRSF
• Keep neuron specific genes from being expressed
  in non neuronal cells
• In neuronal cells smRNA expressed
• Allows transcriptional activation of these neuron
  specific genes
• Mechanism Unknown
• Conformational change induced?

66
Need a Project?

• New roles for RNA add to our current paradigm for
  gene and protein regulation
• Post transcriptional and transcriptional
• Predictive methods, data management, and user
  tools will have to catch up
• Maybe well need a regulomics or RNA
  informatics specific workshop in the future...

67
For More Detailed Information

• Reference list included in your notes