Highly Specific Gene Silencing by Artificial miRNAs in Rice
Citation: Warthmann N, Chen H, Ossowski S, Weigel D, Herve P (
Highly Specific Gene Silencing by Artificial miRNAs in Rice
Norman Warthmann 0
Hao Chen 0
Stephan Ossowski 0
Detlef Weigel 0
Philippe Herve 0
Justin O. Borevitz, University of Chicago, United States of America
0 1 Department of Molecular Biology, Max Planck Institute for Developmental Biology, Tu bingen, Germany, 2 International Rice Research Institute (IRRI) , Metro Manila , Philippines
Background: Endogenous microRNAs (miRNAs) are potent negative regulators of gene expression in plants and animals. Artificial miRNAs (amiRNAs)-designed to target one or several genes of interest-provide a new and highly specific approach for effective post-transcriptional gene silencing (PTGS) in plants. Methodology: We devised an amiRNA-based strategy for both japonica and indica type strains of cultivated rice, Oryza sativa. Using an endogenous rice miRNA precursor and customized 21mers, we designed amiRNA constructs targeting three different genes (Pds, Spl11, and Eui1/CYP714D1). Upon constitutive expression of these amiRNAs in the varieties Nipponbare (japonica) and IR64 (indica), the targeted genes are down-regulated by amiRNA-guided cleavage of the transcripts, resulting in the expected mutant phenotypes. The effects are highly specific to the target gene, the transgenes are stably inherited and they remain effective in the progeny. Conclusion/Significance: Our results not only show that amiRNAs can efficiently trigger gene silencing in a monocot crop, but also that amiRNAs can effectively modulate agronomically important traits in varieties used in modern breeding programs. We provide all software tools and a protocol for the design of rice amiRNA constructs, which can be easily adapted to other crops. The approach is suited for candidate gene validation, comparative functional genomics between different varieties, and for improvement of agronomic performance and nutritional value.
Funding: This work was supported by the Max Planck Society, of which D.W. is a director, by European Community FP6 IP SIROCCO (contract
LSHG-CT-2006037900), and by the USAID-funded project entitled Development of rice biotechnology products for Asia: technical and pre-regulatory components at the
International Rice Research Institute (IRRI) coordinated by P.H. and Gerard Barry. The funders had no role in study design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
. These authors contributed equally to this work.
Analyzing naturally occurring loss-of-function alleles and
screening large collections of chemically induced or
sequenceindexed insertion mutants have proven instrumental to discover
gene functions and biological mechanisms. However, saturating
entire genomes by this approach requires very large populations
and it is restricted to specific genetic backgrounds (varieties). A
current challenge in rice research is still to unravel the functions of
all genes in the genome, which would ultimately facilitate the
identification of genes for agronomically important traits and the
linkage between gene functions and specific traits across different
varieties. Perhaps contrary to intuition, there are several examples
where loss of gene function results in improved plant performance,
such as increased yield or biotic and abiotic stress tolerance .
One prominent example in rice is sd1, which encodes a GA20
oxidase, GA20ox-2, involved in gibberellin biosynthesis.
Inactivation of Sd1 results in a semi-dwarf phenotype, a trait that triggered
the Green Revolution in rice. Since the 1960s, sd1 alleles have
been widely used to confer semi-dwarfism in modern rice cultivars
. Similarly, loss of activity of Eui1 (Elongated uppermost
internode1/CYP714D1), which encodes a cytochrome P450
monooxygenase, promotes panicle exsertion, a trait useful in male sterile
lines for hybrid seed production [7,8]. Originally identified as a
spontaneous mutant in a japonica variety, the original eui1 allele
had to be introduced into indica varieties by introgression or new
alleles had to be isolated after mutagenesis in the desired
background [9 and references therein].
Today transgene-mediated gene silencing through RNA
interference (RNAi) offers a directed way of inactivating one or
several specific genes . RNAi transgenes are dominant and
can be applied in many different genetic backgrounds for any
known gene in the genome. In addition it enables, for example,
simultaneous targeting of several sequence-related genes, including
duplicate genes in polyploids or redundantly acting homologs.
Pathogen-derived genes have also successfully been targeted, e.g.,
to achieve enhanced virus and insect resistance [2,1315].
Through tissue-specific, inducible or stimuli-responsive and partial
gene inactivation, desired mutant phenotypes may be temporally
and/or spatially restricted and modulated in severity. Although
widely used for gene discovery and validation of gene function,
one impediment to a broader use of transgene-mediated gene
silencing especially in crop improvement has been the question of
In the 1990s, several related transgenic approaches to induce
loss of gene function in plants were developed. They were later
found to all act through small silencing RNAs (sRNAs) derived
from double-stranded RNA precursors [16,17]. Surprisingly, only
few studies have systematically compared different silencing
strategies. Two of these found that hairpin RNA interference
(hpRNAi) produced more efficient silencing triggers than
separately transcribed sense and antisense RNAs [18,19]. Although
most efforts to improve sRNA-mediated gene silencing have
focused on maximal effectiveness, a recent computational analysis
from 25 plant species predicted that a majority of transcribed
regions have potential off-target effects when used in
posttranscriptional gene silencing (PTGS) . Potential off-target
effects of hpRNAi in plants arise because the large number of
sRNAs spawned by the silencing construct may target any gene
that shares perfect or near-perfect complementarity to these
sRNAs. Long hairpins are more likely to give rise to optimally
effective sRNAs, but the chance of generating sRNAs with
unwanted off-target effects increases as well.
The recently developed artificial microRNA (amiRNA)
technology exploits endogenous miRNA precursors to preferentially
generate a single specific sRNA in vivo [2,11,12,2125]. MiRNAs
in plants typically induce cleavage of the target mRNA opposite
positions 10 and 11 of the miRNA . Because the amiRNA
sequence does not have to be perfectly complementary to the
target site, it can be optimized to target only one or, alternatively,
several sequence-related genes. Genome-wide expression analyses
have shown that amiRNAs have similarly high specificity as
endogenous miRNAs [21,27]. Efficient amiRNA-mediated gene
silencing has been observed to occur in a quantitative fashion, with
stronger promoters often causing higher degrees of gene silencing
and it seems that there are few, if any non-autonomous effects
[21,22,25]. Arabidopsis miRNA precursors have been modified to
silence endogenous and exogenous target genes in the
dicotyledonous plants Arabidopsis, tomato and tobacco [2,13,2123], and
because of their specificity and versatility, amiRNA vectors are
recognized as second-generation RNAi vectors .
Gene silencing by amiRNAs has not yet been demonstrated for
a monocotyledon species, or for traits proven to be of agronomic
importance. Here, we report the successful application of
amiRNAs to agronomically relevant strains of both japonica and
indica rice, together with enabling tools for the breeding and
functional genomics communities.
To optimize expression of the amiRNA, we chose to employ an
endogenous rice miRNA precursor, a 245 bp fragment of the
osaMIR528 locus. Appropriate 21mers to replace the endogenous
miRNA and miRNA* of osa-MIR528 were designed based on the
TIGR4 rice genome annotation. Modified precursors (Figure 1)
were constitutively expressed in two rice varieties, Nipponbare
(japonica) and IR64 (indica). The method was evaluated by
targeting three different rice genes, Phytoene desaturase (Pds,
Os03g08570), Spotted leaf 11 (Spl11, Os12g38210), and Eui1
(CYP714D1, Os05g40384) (Table 1). Mutations in these genes
cause an albino phenotype (pds) , spontaneous lesion formation
in the absence of pathogens (spl11) , and elongation of the
uppermost internode at heading stage (eui1) [7,8], respectively.
Transgenic lines containing the amiRNA constructs reproduced
the previously described phenotypes for all intended target genes
in both varieties (Figure 2, Table 2, Table S1).
Phytoene desaturase (PDS) catalyzes an early step in the
carotenoid biosynthesis pathway and the absence of protective
carotenoids results in bleaching through photo-oxidation of
chlorophyll, making it a convenient gene for proof-of-principle
applications. PDS is encoded by a single copy gene in the rice
genome and silencing of Pds (AF049356) in rice by hpRNAi, using
a 470 bp gene fragment, has been reported to cause an albino
phenotype . Over 90% (39 of 42) of transgenic Nipponbare
plants (NB_pNW78) and 56% (38 of 68) of transgenic IR64 plants
(IR64_pNW78) had an albino phenotype (Figure 2, Table 2).
Many plants with severe bleaching died soon after transfer to
rooting medium or soil (Figure S1).
The original rice spl11 mutant had been identified from an EMS
mutagenized IR68 (indica) population and shows rust colored
spots from maximum tillering stage to maturity . Fifty-two%
(Nipponbare, n = 29) and 38% (IR64, n = 32) of independent
transgenic lines showed the characteristic phenotype (Figure 2,
Table 2). Both sets of transgenic plants, NB_pNW76 and
IR64_pNW76, exhibited leaf lesions at tillering stage, with the
mutant phenotype being most severe at the maximum tillering
stage. Although the observed efficiency based on the phenotypic
data seems low for IR64, we detected reduced expression of Spl11
transcript in 18 out of 20, or 90%, of primary IR64_pNW76
transformants, (data not shown). We subsequently compared T1
progeny of four independent IR64_pNW76 lines to three IR64
spl11 mutants (G5612-1-6, G5717-1-6, and D2487-3-5). In the
mutants as well as in the pNW76 T1 plants we observed the
characteristic leaf spots already at the 4-leaf stage, even though we
did not detect the mutant phenotype at this early stage in the T0
lines. We attribute the weak phenotype in T0 plants to
environmental conditions and to the effects of in-vitro
regeneration (Guo-Liang Wang, personal communication). These data
suggest that the silencing efficiency based on the phenotype of T0
lines reported in Table 2 may be an underestimate.
Eui1 encodes a cytochrome P450 monooxygenase (CYP714D1)
that inactivates gibberellin (GA) and thus modulates GA-dependent
development . It is preferentially expressed in young panicles
during heading stage and loss of EUI1 activity results in increased
availability of biologically active gibberellin in the uppermost
internode . Plants carrying eui1 mutations are taller than
wildtype plants owing to increased internode lengths, most extreme in
the uppermost internode [7,8]. Transgenic plants expressing an
amiRNA construct against Eui1 (pNW81) were taller than control
plants with significantly longer uppermost internodes and better
panicle exsertion (Figure 2, Table S1). 83% (33 of 40 independent
transformants) of IR64_pNW81 plants and 28% (9 of 32)
NB_pNW81 plants exhibited the characteristic eui1 phenotype
(Table 2). The phenotype was more pronounced in the indica
variety IR64, which is in agreement with reports on different eui1
alleles in indica and japonica varieties . Although the phenotypic
data indicate a low efficiency for Nipponbare, we found the
expression of Eui1 transcript to be clearly reduced in 15 out of 19
(79%) of Nipponbare_pNW81 plants (data not shown). This again
indicates that the actual silencing efficiency might be higher than
indicated in Table 2, but that the phenotype in T0 plants was
attenuated by environmental and in-vitro regeneration conditions.
Among the transgenic plants that exhibited the expected
phenotypes, we identified T0 plants that contained only one or
two copies of the amiRNA transgene (Figure S2), which confirms
the effectiveness of amiRNA-mediated gene silencing. For three
Figure 1. Secondary Structure of the osa-MIR528 stemloop in pNW55. Depicted is the secondary structure of the 255 bp derived from
osaMIR528 (245 bp plus the BamHI and KpNI restriction sites) as predicted by RNAfold  (Vienna RNA package, http://www.tbi.univie.ac.at/RNA) for
23uC. The sequences replaced in the different transgenes are boxed. The oligo designer in WMD2  (http://wmd2.weigelworld.org) designs the
amRNA* sequence (black) such that it pairs to the respective miRNA (red) in the same way as in osa-MIR528.
Nipponbare T0 lines that contained a single copy of pNW78, we
analyzed the progeny (T1 plants) and observed Mendelian,
dominant inheritance of the albino phenotype (3:1). PCR analysis
confirmed that the phenotype co-segregated with the transgene
(Table S2, Figure S3).
To demonstrate that the observed phenotypes are indeed due to
amiRNA mediated down-regulation of the target genes, we
analyzed the expression levels of the amiRNAs and their target
genes in selected transgenic lines for both Nipponbare and IR64.
The expected amiRNA species were readily detected by RNA blot
in every line and tissue tested, which demonstrates that the
amiRNA precursor is efficiently produced and processed in planta
(Figure 3). Using reverse transcription followed by quantitative
real time PCR (qRT-PCR), we found the abundance of the target
transcripts to be greatly reduced (Figure 3). To ensure causality
between amiRNA expression and target gene down-regulation, we
mapped the amiRNA-guided cleavage sites by RACE-PCR .
In all transgenic lines tested, the expected cleavage products were
detected (Figure 3), which also indicates that the amiRNAs are of
Spl11 has several homologs in the rice genome, which might be
problematic for conventional gene silencing strategies such as
Predicted mature miRNA sequences
Elongated upper-most internode
hpRNAi. To examine the predicted specificity of the amiRNA
against Spl11, we investigated the mRNA levels of Spl11 and its
nine closest homologs (Table S3) by multiplex qRT-PCR. We
found that only the transcript abundance of the targeted Spl11
(Os12g38210) was significantly reduced (Figure S4). This
observation not only confirms the specificity of the amiRNA, but
Figure 2. Phenotypes of plants overexpressing amiRNAs. The empty vector control is always on the left. (AD) pNW76, construct targeting
Spl11. (A) Nipponbare whole plants, (B) Nipponbare leaves, (C) IR64 whole plants, (D) IR64 leaves at maximum tillering stage. (EH) pNW78, construct
targeting Pds. (E) Nipponbare whole plants, (F) Nipponbare leaves, (G) IR64 whole plants, (H) IR64 leaves at tillering stage. (IL) pNW81, construct
targeting Eui1. (I) Nipponbare whole plants, (J) Nipponbare uppermost internode, (K) IR64 whole plants, (L) IR64 uppermost internode. The ruler in (J)
and (L) is 60 cm long.
No. of independent
also indicates that production of secondary siRNAs from the
primary target gene, which could indirectly silence related genes,
should not be a major concern, in line with the observation that so
called transitivity effects of miRNAs are rare.
We had designed two different amiRNAs for each target gene
(Table 1, Figure 4). In each case, one of these was much more
effective in both varieties in terms of producing the predicted
phenotype (Table 2) as well as reducing the RNA levels of the
target gene (data not shown). Many aspects of the processing of
plant miRNA precursors are still unknown. Some might not be
processed as intended, and the resulting amiRNA might therefore
not have the predicted sequence. Other possible explanations
include sequence properties of the amiRNA or the target site such as
target site accessibility that are not optimal. Several recent
publications describe important effects of the mRNA structure
surrounding the target site on the efficiency of siRNA-mediated
gene silencing in animals [31, 32 and references therein]. It has
hence been suggested that one should model the thermodynamics of
RNA-RNA interactions as the sum of two energy contributions: (1)
the energy necessary to open the binding site(s) and (2) the energy
gained from hybridization . We assessed the thermodynamics of
the amiRNA-mRNA interactions for our amiRNAs with their
targets using the program RNAup of the Vienna RNA package
. Two of the three ineffective amiRNAs (pNW77 and pNW82)
were found to have unfavorable target sites, as the energy required
to open the nucleotide bonds at the target sites is high. In contrast,
for pNW78 and pNW81, the probability that nucleotides at the
respective target sites are unpaired is higher and hence the effective
hybridization energy of the amiRNA to the target mRNA is better
(Figure 4). The inefficacy of amiRNA pNW75 is not easily
explained by hybridization properties. However, amiRNA pNW75
has a purine:purine mismatch (A:A) to its target mRNA at position
16 (Figure 4), which has recently been shown to have a strong
negative impact on siRNA mediated silencing in animals .
Rice, Oryza sativa, is one of the most important staple crops in
the world and the model plant for monocots. Still, the majority of
rice genes (61% of predicted loci) do not have an assigned or even
putative function (Kevin Childs, personal communication). Rice
shares excellent gene co-linearity with other monocots and nearly
all genes in important crops such as wheat, barley and maize have
a homolog in rice . Mutation and activation strategies have
been deployed at large scales in rice and other species. Identifying
candidate genes for induced mutations or Quantitative Trait Loci
(QTL), or establishing parallels between gene functions and
specific traits across different species and varieties is difficult and
No. of independent
often a rate-limiting step. Directed, specific gene silencing by
transgenes provides a powerful tool to tackle these challenges. Our
work demonstrates successful gene silencing by amiRNAs in an
indica and a japonica variety of rice. Our method is based on a
new vector derived from an endogenous rice miRNA precursor,
osa-MIR528, and can be easily adopted by the community and
expanded to other monocot crops following the experimental
procedures described. In our examples, the different mutant
phenotypes triggered by amiRNA-directed gene silencing could be
rapidly detected in T0 lines in different rice varieties, the
transgenes were stably inherited and remained active in the
progeny. Our data demonstrate that the amiRNA precursor is
efficiently processed in planta in all tissues tested and the
amiRNAguided cleavage of target transcripts is highly specific.
Although the effects of target site accessibility and the impact of
particular mismatches at certain positions on plant miRNA
efficiency have not yet been studied systematically, these
characteristics may provide a means to deliberately modulate amiRNA
efficiency and hence the degree of target gene down-regulation. Our
original WMD tool for the design of amiRNAs selected and
designed amiRNA sequences based on empirically established
criteria . The next generation, WMD2, has been modified such
that amiRNA candidates with a purine:purine mismatch to the
target mRNA at position 16 are avoided. Based on the results from
the current work, it should be possible to improve the design tool by
giving more attention to the effective binding energy. In any case,
designing several amiRNA constructs targeting different regions in
the target mRNA remains a prudent avenue. For the analysis of
small numbers of genes, the amiRNA constructs are easily produced
by modification PCR from the precursor plasmid. In addition, the
small size of the miRNA precursor stemloop should facilitate future
high throughput production by synthesis.
AmiRNAs will have many applications in rice functional
genomics. They can trigger efficient silencing of one or several
genes or specific alleles with great specificity, even for genes
essentially inaccessible to other reverse genetics techniques, e.g.,
tandem duplicated genes. Many recent QTL studies have shown
that a reduction or loss of gene function often underlies varietal
differences and agronomically important traits in rice and other
grasses . The amiRNA technology offers a way for
timeefficient modification of the expression of such genes in any
variety. This not only enables rapid knowledge transfer between
different varieties, e.g., between indica and japonica strains, which
are separated by complex incompatibility barriers , but also
the introduction of important traits for improving agronomic
performance and/or nutritional value into a broad range of
varieties. Its dominant nature and specificity will also be useful in
hybrid crop breeding, since only one of the parental lines needs to
carry the amiRNA construct and amiRNAs have the potential to
specifically reduce the activity of only one of the parental alleles.
Materials and Methods
Sequence Information and Software
Rice microRNA precursor sequences were downloaded from
miRBase/Rfam (http://microrna.sanger.ac.uk/ ), rice gene
sequences for WMD from TIGR (http://www.tigr.org/tdb/e2k1/
osa1) and 1,140,645 available rice cDNA sequences for blast
searches from NCBI (http://www.ncbi.nlm.nih.gov). Sequence
homology of miRNA precursors to available rice cDNAs was
determined by BLAST and multiple alignments were performed
using MUSCLE . The possible amiRNA candidate sequences
were generated using WMD [21,25], customized to use the
published Oryza sativa genes (TIGR v4, 62,827 sequences) as input.
The binding properties of the amiRNAs to the respective target
mRNAs (80 bp surrounding the target site) at 23uC (Figure 4)
were assessed using the program RNAup . The top 9
homologs of Spl11 (Os12g38210) in the rice genome were
identified by blasting the Spl11 CDS against Genes in TIGR
Rice Pseudomolecules-CDS using WU-BLAST 2.0 implemented
on the TIGR Rice Genome Annotation website (http://www.tigr.
org/tdb/e2k1/osa1/ . Most of them are annotated as spotted
leaf protein 11, putative (Table S3). Their percent identity to
Os12g38210 was determined by pair wise Needleman-Wunsch
alignments using the EMBOSS program needle with default
WMD2 (http://wmd2.weigelworld.org) currently supports 38
plant species including Oryza sativa (TIGR v5) and designs 21mer
sequences directed against one or several genes. WMD suggests
suitable amiRNA candidates after a two-step selection process based
on empirically established criteria for efficiency and specificity and is
improved as the knowledge on the biology of miRNAs grows
[21,25,27]. Currently, 21mers from the reverse complement of the
target transcript(s) are considered effective amiRNA candidates, if
they have an A (sometimes also U) at position 10 and display 59
instability (higher AU content at the 59 end and higher GC content
at the 39 end). The first base (position 1) will be replaced by an U
and all candidates then undergo an iterative series of in-silico
mutations at positions 1315 and 1721 followed by mappings
against all currently known cDNA sequences or gene models for the
particular species. Any transcript other than the specified target(s),
which matches the 21mer with up to 5 mismatches, is considered an
off-target, if no more than 1 mismatch in the seed-region (positions
212) is found. Candidates with one or two mismatches to the
intended target(s) and without off-targets are then ranked based on
hybridization properties and returned to the user.
The Stemloop Backbone
To select an endogenous stemloop backbone for the amiRNA
transgenes we screened all rice miRNA precursors for Oryza sativa
present in miRBase/Rfam  for several criteria: the predicted
secondary structure should form a short and simple hairpin-loop
and the miRNA* sequence should pair to the miRNA without
insertions or deletions. The precursor sequence should be present
in EST databases and it should possibly have further proof of its
existence and processing in the literature. We selected the
precursor of osa-MIR528 (accession number: MI0003201),
because of its short stemloop with a predicted simple fold-back
structure and its effective expression and processing in several rice
tissues including leaves, root and panicle . In addition, MPSS
signatures had been detected for the miRNA as well as for the
miRNA* in stem tissue of Nipponbare (http://mpss.udel.edu/
rice/) . Database searches revealed 3 EST sequences of this
precursor, accession numbers: AK099390, AK073820,
AK063857. By sequence alignment we noticed that the precursor
likely contains an intron. The stemloop of osa-MIR528 is identical
in sequence to the annotated rice gene LOC_Os03g03724 (TIGR
v5, or LOC_Os03g0129400, RAP). A 245 bp DNA fragment
including the osa-MIR528 stemloop, but avoiding the predicted
intron, was amplified from Nipponbare genomic DNA by PCR
(primers G-11491 and G-11494, Table S4, introducing BamHI
and KpnI restriction sites), and blunt end cloned into pBluescript
KS to give rise to the backbone vector pNW55. All subsequent
clones were produced by modification PCR amplifications from
We used a customized version of the original Web MicroRNA
Designer platform (WMD) to design amiRNA sequences (21mers)
based on the TIGR v4 rice genome annotation. WMD suggested
suitable amiRNA candidates based on good hybridization
properties to the target mRNAs while minimizing possible
offtarget effects to other genes in the rice genome . We selected
two 21mers (amiRNAs) per target gene (Table 1, Figure 4) to
target different sites in the target mRNA and designed appropriate
amiRNA* sequences that would in pairing to the amiRNA exactly
mimic the foldback structure of the endogenous osa-MIR528
(Figure 1). Each primary amiRNA construct was engineered
from pNW55, replacing the 21 bases of the natural osa-MIR528
miRNA as well as the partially complementary region of the
miRNA* by modification PCRs in a similar way as described
earlier , following the PCR scheme in Figure S5. All PCRs
were performed with ProofStart TM DNA Polymerase (Qiagen) in
a volume of 25 ml according to the manufacturers
recommendation and the PCR protocols are given in Table S5. For each
amiRNA construct, three fragments including the multiple cloning
sites (MCS) were PCR amplified from the template clone pNW55
using a total of six primers (Table S6). Four primers were
designed such that the miRNA and miRNA* sequences are
substituted by the desired 21 mers. Since in osa-MIR528 the
miRNA is on the forward strand, primer I contains the amiRNA
in sense orientation, primer II its reverse complement, primer III
the amiRNA* sequence in sense and primer IV the amiRNA*
sequence in antisense orientation. The plasmid information for
pNW55 has been integrated into the online WMD2 platform, and
all appropriate primer sequences, needed for customization of
pNW55, can be retrieved using the primer design function of
WMD2. For each miRNA construct three modification PCRs
were performed with primers G-4368+II, I+IV and III+G-4369 on
pNW55 as template, yielding fragments of 256, 87 and 259 bp
lengths, respectively. The three resulting fragments were gel
purified (Promega) and then fused by one PCR with the two
flanking primers G-4368 and G-4369 on a mixture of 1 ml from
each previous PCR as template. The fusion product of 554 bp was
again gel purified (Promega), cloned into pGEMH-T Easy Vector
(Promega), sequence verified, excised with BamHI/KpnI and
transferred into the multiple cloning site of the binary vector
IRS154 (a pCambia derivative, provided by Emmanuel
Guiderdoni, CIRAD, France). In IRS154, the expression of the
transgene is driven by the maize ubiquitin promoter and first
intron (GenBank accession number: S94464, basepair-positions
21993). All six amiRNA plant expression vectors were transformed
into Agrobacterium tumefaciens strains LBA4404 and EHA105.
Rice transformation and culture
Two rice varieties, Nipponbare (japonica) and IR64 (indica)
were transformed with the transgenes according to modified
protocols from references  and , respectively, and selected
on hygromycin. Control plants were obtained by transformation
with the empty binary vector IRS154. All regenerated T0
transgenic plants were genotyped for the presence of the
transgenic Hygromycin B Phosphotransferase gene (Hpt) with the
endogenous Gos5 gene (AF093635, Os05g48630) as control; see
Table S4 for primer sequences. Plants were grown in a
greenhouse in Los Ba nos, Philippines under natural photoperiod
(December 2006May 2007) at 29 uC (day) and 23 uC (night).
Small RNA blots
Total RNA was isolated from rice leaves (Spl11, Pds) and young
panicles (Eui1/CYP716D1) using Trizol (Invitrogen). Six mg of total
RNA was resolved on a 17% PAGE under denaturing conditions
and transferred onto a charged nylon membrane (Nytran SuPer
Charge, WhatmanH Schleicher & Schuell) by semi-dry blotting
. The blot was hybridized with five pmol of a radioactively
end-labeled oligo probe complementary to the mature amiRNA in
PerfectHybTM Plus (Sigma-Aldrich) hybridization buffer at 38uC
overnight. After hybridization, the blot was briefly rinsed with
26SSC, 0.2% SDS, washed twice for 20 min with 26SSC, 0.2%
SDS at 42uC and exposed to Kodak BioMax MS film for 4 h at
room temperature. The blots shown in Figure 3 were exposed
simultaneously onto the same film.
Primary transgenic lines for each amiRNA vector were selected
based on their phenotype. Total RNA was extracted from leaves at
tillering stage (Pds and Spl11 genes) or from flowering panicles at
heading stage (Eui1/CYP714D1 gene), using Trizol (Invitrogen).
The primary RT-PCR screen was performed with a one-step
RTPCR kit (Qiagen) on 40 ng total RNA, and two transgenic lines for
each transgene were further analyzed by quantitative Real Time
RT-PCR and compared  to an empty vector control plant
(Figure 3) using the primers G-13120, G-13121 (Spl11); G-14494,
G-14495 (Pds); G-13110, G-13111 (Eui1/CYP714D1). Reverse
transcription (Invitrogen) was performed with 2 mg of total RNA.
PCR amplification was carried out in the presence of the
doublestrand DNA-specific dye SYBR Green (Invitrogen-Molecular
ProbesH) and monitored in real time with the Opticon Continuous
Fluorescence Detection System (MJR). GAPDH (Os04g40950,
primers G-12061, G-12062) served as standard. All primers are
listed in Table S4; more detailed protocols and PCR conditions
are available upon request.
Cleavage site mapping
The amiRNA-guided cleavage sites were mapped by
RACEPCR . From 30 mg total RNA extracted from one plant per
transgenic line (Trizol), the mRNA fraction was enriched
(OligotexH mRNA Mini Kit, Qiagen). 59-RACE reactions
(GeneRacerTM Kit, Invitrogen) were performed on 250 ng mRNA
following the manufacturers instructions, but omitting the
dephosphorylation and de-capping steps. PCRs and nested PCRs
were performed with the GeneRacerTM primers (Invitrogen) and
the gene specific primers G-15552, G-13121 (Spl11), G-15553,
G13117 (Pds) and G-15556, G-15554 (Eui1/CYP714D1); see Table
S4. PCR products were cloned into the pCRH4-TOPO Vector
(Invitrogen) and the DNA sequence of the 59-prime-end was
determined by dideoxy sequencing between 21 and 42 clones for
each transgenic line.
Genomic DNA blot
Three transgenic lines for each transgene with reduced
expression of the target genes (assessed by RT-PCR) were
subjected to DNA blot analysis to determine the copy number of
the transgene. Fifteen mg genomic DNA of each sample was
digested with BamHI, separated on a 1% agarose gel, and
transferred onto a nylon membrane (Millipore) by capillary
blotting. Two different DIG-labeled probes (NOS terminator
and Hpt gene) were produced by PCR labeling (PCR DIG Probe
Synthesis Kit, Roche), and the blots were hybridized according to
the DIG Applications Manual for Filter Hybridization (Roche);
see Table S4 for primer sequences.
Mature seeds from three T0 pNW78 Nipponbare transgenic
lines (NB_pNW78_02, NB_pNW78_16, NB_pNW78_39, all
targeting Pds), one Nipponbare empty vector control plant, and
wild type Nipponbare were planted. After 3 weeks, the phenotype
of the T1 seedlings was observed and the segregation ratios
recorded. Six T1 transgenic plants (three plants with normal
phenotype and three albino plants) were analyzed by PCR to
confirm that the phenotype co-segregated with the transgene.
Specificity/Transitivity of the amiRNAs
AmiRNA specificity was assayed on RNA from T0 pNW76
IR64 plants. Total RNA from three independent IR64_pNW76
transgenic plants with the expected phenotype and from three
IR64 empty vector transgenic plants was extracted from flag-leaf
samples at heading stage, using Trizol (Invitrogen). After DNase I
(Promega) treatment, RNA concentration was normalized to
30 ng/ml. Transcript abundance for Spl11 and its homologs
(Table S3) was measured synchronously, using the
LabTM GeXP Genetic Analysis System (Beckman-Coulter) with
ATPase (Os04g56160) as reference gene following the
manufacturers instructions. After reverse transcription (GenomeLabTM
GeXP Start Kit, Beckman-Coulter), a multiplex PCR reaction
was conducted with Thermo-Start DNA Polymerase (ABgene).
The PCR products were then placed in the GenomeLabTM
GeXP Genetic Analysis System for capillary electrophoresis and
fragment size analysis. The RT-PCR primers used are listed in
Figure S1 Regenerated NB_pNW78 plant on regeneration
medium. A regenerated transgenic Nipponbare plant
(NB_pNW78, the transgene targeting Phytoene Desaturase) on
regeneration medium. Many transgenics with severe albino
phenotypes died later upon transfer to soil.
Found at: doi:10.1371/journal.pone.0001829.s001 (7.63 MB TIF)
Figure S2 DNA blot analysis of amiRNA transgenes. DNA blot
analysis of T0 transgenic plants with different probes. (A)
Transgenic Nipponbare plants with Nos terminator probe; (B)
Transgenic IR64 plants with Nos terminator probe; (C)
Transgenic Nipponbare plants with Hpt probe; (D) Transgenic IR64
plants with Hpt probe. Wt is wild type Nipponbare as
Found at: doi:10.1371/journal.pone.0001829.s002 (3.70 MB TIF)
Figure S3 Inheritance of an amiRNA transgene. T1 progeny
from one NB_pNW78 T0 plant that contained one single copy of
the transgene. (A) 20-day-old rice seedlings of wild type
Nipponbare (Wt) and NB_pNW78 T1 transgenic line. (B)
20day-old rice seedlings; NB_pNW78 transgenic plants show albino
leaves and delayed growth; IRS154 is the empty vector control. (C)
PCR analysis of NB_pNW78 T1 plants. The lower bands are the
127 bp PCR products of Gos5, and the upper bands are the 910
bp PCR products of Hpt primers: (1) 1 kb DNA ladder; (2) IRS154
plasmid; (3) wild type; (4) NB_IRS154 empty vector control plant.
(57) NB_pNW78 T1 plants with normal leaf color; (810)
NB_pNW78 T1 plants with albino phenotype.
Found at: doi:10.1371/journal.pone.0001829.s003 (5.71 MB TIF)
Figure S4 Gene expression analysis on Spl11 and homologs.
Gene expression profile of the targeted Spl11 (Target gene,
Os12g38210) and homologs based on multiplex RT-PCR in
transgenic IR64_IRS154 (empty vector control, black) and
IR64_pNW76 (white) plants relative to ATPase in leaves
(GenomeLabTM GeXP Genetic Analysis System, Beckman
Coulter). Homologs and primers are listed in Supplementary
Tables S3 and S7. The value of gene expression is given as the
mean of three independent transgenic plants (T0) and error bars
indicate the standard deviation between biological replicates.
Gene expression of Os08g37570 and Os06g51130 could not be
detected in agreement with the Rice MPSS database (http://mpss.
* Gene expression difference is significant according to
Students t test (P = 0.0002).
1. Nakano, M. et al. Plant MPSS databases: signature-based
transcriptional resources for analyses of mRNA and small RNA.
Nucleic Acids Res 34, D731-735 (2006).
Found at: doi:10.1371/journal.pone.0001829.s004 (0.63 MB TIF)
Figure S5 PCR Scheme and Map of pNW55.
Found at: doi:10.1371/journal.pone.0001829.s005 (0.69 MB TIF)
We thank Heike Wollmann for sharing technical expertise, Heike
Wollmann, Rebecca Schwab, Roosa Laitinen, John Bennett and Hei
Leung for discussion and critical reading of the manuscript at various
stages, and Maricris Zaidem and Virginia Laluz for technical assistance.
Seeds of IR64 spl11 mutants were kindly provided by Hei Leung,
International Rice Research Institute.
Analyzed the data: NW HC SO. Wrote the paper: DW NW HC PH.
Other: Designed research: PH. Coordinated: PH DW. Did computational
work: SO. Performed rice transformations: HC. Did molecular and plant
evaluation work: HC. Selected and cloned stemloop and amiRNAs: NW.
Did molecular and computational work: NW.
1. Song XJ , Huang W , Shi M , Zhu MZ , Lin HX ( 2007 ) A QTL for rice grain width and weight encodes a previously unknown RING-type E3 ubiquitin ligase . Nat Genet 39 : 623 - 630 .
2. Niu QW , Lin SS , Reyes JL , Chen KC , Wu HW , et al. ( 2006 ) Expression of artificial microRNAs in transgenic Arabidopsis thaliana confers virus resistance . Nat Biotechnol 24 : 1420 - 1428 .
3. Leshem Y , Melamed-Book N , Cagnac O , Ronen G , Nishri Y , et al. ( 2006 ) Suppression of Arabidopsis vesicle-SNARE expression inhibited fusion of H2O2- containing vesicles with tonoplast and increased salt tolerance . Proc Natl Acad Sci USA 103 : 18008 - 18013 .
4. Sasaki A , Ashikari M , Ueguchi-Tanaka M , Itoh H , Nishimura A , et al. ( 2002 ) Green revolution: a mutant gibberellin-synthesis gene in rice . Nature 416 : 701 - 702 .
5. Spielmeyer W , Ellis MH , Chandler PM ( 2002 ) Semidwarf (sd-1), ''green revolution'' rice, contains a defective gibberellin 20-oxidase gene . Proc Natl Acad Sci USA 99 : 9043 - 9048 .
6. Asano K , Takashi T , Miura K , Qian Q , Kitano H , et al. ( 2007 ) Genetic and molecular analysis of utility of sd1 alleles in rice breeding . Breed Sci 57 : 53 - 58 .
7. Luo A , Qian Q , Yin H , Liu X , Yin C , et al. ( 2006 ) EUI1, encoding a putative cytochrome P450 monooxygenase, regulates internode elongation by modulating gibberellin responses in rice . Plant Cell Physiol 47 : 181 - 191 .
8. Zhu Y , Nomura T , Xu Y , Zhang Y , Peng Y , et al. ( 2006 ) ELONGATED UPPERMOST INTERNODE encodes a cytochrome P450 monooxygenase that epoxidizes gibberellins in a novel deactivation reaction in rice . Plant Cell 18 : 442 - 456 .
9. Rutger JN ( 2005 ) Registration of Guichao 2 eui rice genetic stock . Crop Sci 45 : 433 .
10. Kusaba M ( 2004 ) RNA interference in crop plants . Curr Opin Biotechnol 15 : 139 - 143 .
11. Small I ( 2007 ) RNAi for revealing and engineering plant gene functions . Curr Opin Biotechnol 18 : 148 - 153 .
12. Tang G , Galili G , Zhuang X ( 2007 ) RNAi and microRNA: breakthrough technologies for the improvement of plant nutritional value and metabolic engineering . Metabolomics 3 : 357 - 369 .
13. Qu J , Ye J , Fang R ( 2007 ) Artificial miRNA-mediated virus resistance in plants . J Virol 81 : 6690 - 6699 .
14. Baum JA , Bogaert T , Clinton W , Heck GR , Feldmann P , et al. ( 2007 ) Control of coleopteran insect pests through RNA interference . Nat Biotechnol 25 : 1322 - 1326 .
15. Mao YB , Cai WJ , Wang JW , Hong GJ , Tao XY , et al. ( 2007 ) Silencing a cotton bollworm P450 monooxygenase gene by plant-mediated RNAi impairs larval tolerance of gossypol . Nat Biotechnol 25 : 1307 - 1313 .
16. Watson JM , Fusaro AF , Wang M , Waterhouse PM ( 2005 ) RNA silencing platforms in plants . FEBS Lett 579 : 5982 - 5987 .
17. Sen GL , Blau HM ( 2006 ) A brief history of RNAi: the silence of the genes . FASEB J 20 : 1293 - 1299 .
18. Wesley SV , Helliwell CA , Smith NA , Wang MB , Rouse DT , et al. ( 2001 ) Construct design for efficient, effective and high-throughput gene silencing in plants . Plant J 27 : 581 - 590 .
19. Chuang CF , Meyerowitz EM ( 2000 ) Specific and heritable genetic interference by double-stranded RNA in Arabidopsis thaliana . Proc Natl Acad Sci USA 97 : 4985 - 4990 .
20. Xu P , Zhang Y , Kang L , Roossinck MJ , Mysore KS ( 2006 ) Computational estimation and experimental verification of off-target silencing during posttranscriptional gene silencing in plants . Plant Physiol 142 : 429 - 440 .
21. Schwab R , Ossowski S , Riester M , Warthmann N , Weigel D ( 2006 ) Highly specific gene silencing by artificial microRNAs in Arabidopsis . Plant Cell 18 : 1121 - 1133 .
22. Alvarez JP , Pekker I , Goldshmidt A , Blum E , Amsellem Z , et al. ( 2006 ) Endogenous and synthetic microRNAs stimulate simultaneous, efficient, and localized regulation of multiple targets in diverse species . Plant Cell 18 : 1134 - 1151 .
23. Parizotto EA , Dunoyer P , Rahm N , Himber C , Voinnet O ( 2004 ) In vivo investigation of the transcription, processing, endonucleolytic activity, and functional relevance of the spatial distribution of a plant miRNA . Genes Dev 18 : 2237 - 2242 .
24. Zeng Y , Wagner EJ , Cullen BR ( 2002 ) Both natural and designed micro RNAs can inhibit the expression of cognate mRNAs when expressed in human cells . Mol Cell 9 : 1327 - 1333 .
25. Ossowski S , Schwab R , Weigel D ( 2008 ) Gene silencing in plants using artificial microRNAs and other small RNAs . Plant J 53 : 674 - 690 .
26. Llave C , Xie Z , Kasschau KD , Carrington JC ( 2002 ) Cleavage of Scarecrow-like mRNA targets directed by a class of Arabidopsis miRNA . Science 297 : 2053 - 2056 .
27. Schwab R , Palatnik JF , Riester M , Schommer C , Schmid M , et al. ( 2005 ) Specific effects of microRNAs on the plant transcriptome . Dev Cell 8 : 517 - 527 .
28. Miki D , Shimamoto K ( 2004 ) Simple RNAi vectors for stable and transient suppression of gene function in rice . Plant Cell Physiol 45 : 490 - 495 .
29. Zeng LR , Qu S , Bordeos A , Yang C , Baraoidan M , et al. ( 2004 ) Spotted leaf11, a negative regulator of plant cell death and defense, encodes a U-box/armadillo repeat protein endowed with E3 ubiquitin ligase activity . Plant Cell 16 : 2795 - 2808 .
30. Singh K , Multani DS , Khush GS ( 1995 ) A new spotted leaf mutant in rice . Rice Genet Newsl 12 : 192 - 193 .
31. Ameres SL , Martinez J , Schroeder R ( 2007 ) Molecular basis for target RNA recognition and cleavage by human RISC . Cell 130 : 101 - 112 .
32. Kertesz M , Iovino N , Unnerstall U , Gaul U , Segal E ( 2007 ) The role of site accessibility in microRNA target recognition . Nat Genet 39 : 1278 - 1284 .
33. Muckstein U , Tafer H , Hackermuller J , Bernhart SH , Stadler PF , et al. ( 2006 ) Thermodynamics of RNA-RNA binding . Bioinformatics 22 : 1177 - 1182 .
34. Schwarz DS , Ding H , Kennington L , Moore JT , Schelter J , et al. ( 2006 ) Designing siRNA that distinguish between genes that differ by a single nucleotide . PLoS Genet 2 : e140 .
35. Sallaud C , Gay C , Larmande P , Be`s M , Piffanelli P , et al. ( 2004 ) High throughput T-DNA insertion mutagenesis in rice: a first step towards in silico reverse genetics . Plant J 39 : 450 - 464 .
36. Doebley JF , Gaut BS , Smith BD ( 2006 ) The molecular genetics of crop domestication . Cell 127 : 1309 - 1321 .
37. Harushima Y , Nakagahra M , Yano M , Sasaki T , Kurata N ( 2002 ) Diverse variation of reproductive barriers in three intraspecific rice crosses . Genetics 160 : 313 - 322 .
38. Griffiths-Jones S , Bateman A , Marshall M , Khanna A , Eddy SR ( 2003 ) Rfam: an RNA family database . Nucleic Acids Res 31 : 439 - 441 .
39. Edgar RC ( 2004 ) MUSCLE: multiple sequence alignment with high accuracy and high throughput . Nucleic Acids Res 32 : 1792 - 1797 .
40. Ouyang S , Zhu W , Hamilton J , Lin H , Campbell M , et al. ( 2007 ) The TIGR Rice Genome Annotation Resource: improvements and new features . Nucleic Acids Res 35 : D883 - 887 .
41. Rice P , Longden I , Bleasby A ( 2000 ) EMBOSS: the European Molecular Biology Open Software Suite . Trends Genet 16 : 276 - 277 .
42. Liu B , Li P , Li X , Liu C , Cao S , et al. ( 2005 ) Loss of function of OsDCL1 affects microRNA accumulation and causes developmental defects in rice . Plant Physiol 139 : 296 - 305 .
43. Nobuta K , Venu RC , Lu C , Belo A , Vemaraju K , et al. ( 2007 ) An expression atlas of rice mRNAs and small RNAs . Nat Biotechnol 25 : 473 - 477 .
44. Herve P, Kayano T ( 2006 ) Japonica rice varieties (Oryza sativa, Nipponbare, and others) . Methods Mol Biol 343 : 213 - 222 .
45. Hiei Y , Komari T ( 2006 ) Improved protocols for transformation of indica rice mediated by Agrobacterium tumefaciens . Plant Cell Tissue Organ Cult 85 : 271 - 283 .
46. Pfaffl MW ( 2001 ) A new mathematical model for relative quantification in realtime RT-PCR . Nucleic Acids Res 29 : e45 .