Download New Construct Approaches for Efficient Gene Silencing in Plants

New Construct Approaches for Efficient Gene Silencing
in Plants
Hua Yan, Robert Chretien, Jingsong Ye, and Caius M. Rommens*
Simplot Plant Sciences, J.R. Simplot Company, Boise, Idaho 83706
An important component of conventional sense, antisense, and double-strand RNA-based gene silencing constructs is the
transcriptional terminator. Here, we show that this regulatory element becomes obsolete when gene fragments are positioned
between two oppositely oriented and functionally active promoters. The resulting convergent transcription triggers gene
silencing that is at least as effective as unidirectional promoter-to-terminator transcription. In addition to short, variably sized,
and nonpolyadenylated RNAs, terminator-free cassette produced rare, longer transcripts that reach into the flanking promoter.
These read-through products did not influence the efficacy and expression levels of the neighboring hygromycin phosphotransferase gene. Replacement of gene fragments by promoter-derived sequences further increased the extent of gene
silencing. This finding indicates that genomic DNA may be a more efficient target for gene silencing than gene transcripts.
The unidirectional and unperturbed transcription
of either genes or gene fragments from promoter to
terminator can trigger posttranscriptional silencing of
target genes. Initial expression cassettes for posttranscriptional gene silencing in plants comprised a single gene fragment positioned in either the antisense
(Shewmaker et al., 1992; McCormick et al., 2003) or
sense (van der Krol et al., 1990) orientation between
regulatory sequences for transcript initiation and termination. In Arabidopsis (Arabidopsis thaliana), recognition of the resulting transcripts by RNA-dependent
RNA polymerase leads to the production of doublestranded (ds) RNA (Dalmay et al., 2000). Cleavage of
this dsRNA by Dicer-like (Dcl) proteins such as Dcl4
yields 21-nucleotide small interfering RNAs (siRNAs;
Hamilton and Baulcombe, 1999; Dunoyer et al., 2004).
These siRNAs complex with proteins including members of the Argonaute (Ago) family to produce RNAinduced silencing complexes (Morel et al., 2002; Liu
et al., 2004). The RNA-induced silencing complexes then
target homologous RNAs for endonucleolytic cleavage.
More effective silencing constructs contain both
a sense and antisense component, producing RNA
molecules that fold back into hairpin structures
(Waterhouse et al., 1998; Smith et al., 2000). The high
dsRNA levels produced by expression of inverted repeat transgenes were hypothesized to promote the
activity of multiple Dcls. Analyses of combinatorial
Dcl knockouts in Arabidopsis supported this idea and
* Corresponding author; e-mail [email protected]; fax 208–
327–3212.
The author responsible for distribution of materials integral to the
findings presented in this article in accordance with the policy
described in the Instructions for Authors (www.plantphysiol.org) is:
Caius M. Rommens ([email protected]).
Article, publication date, and citation information can be found at
www.plantphysiol.org/cgi/doi/10.1104/pp.106.082271.
1508
also identified Dcl4 as one of the proteins involved in
RNA cleavage (Gasciolli et al., 2005; Xie et al., 2005).
Although constructs that lack a promoter element
failed to induce gene silencing (Waterhouse et al.,
1998), the requirement of termination sequences has
not been studied extensively. Here, we demonstrate
that terminators are only important for the efficacy of
conventional gene silencing constructs. A new type of
expression cassette that contains either gene or promoter fragments between oppositely oriented promoters
was found to trigger silencing at least as effectively as
conventional constructs. Construct efficacy was correlated with convergent collisional transcription that resulted in the production of pools of variably sized
RNAs.
RESULTS
The Role of Terminators in Conventional Expression
Cassettes for Gene Silencing
The requirement of terminator elements in gene
silencing was studied by retransforming a transgenic
tobacco (Nicotiana tabacum) plant that constitutively
expressed the b-glucuronidase (gus) gene (see ‘‘Materials and Methods’’) with unidirectional silencing constructs (Fig. 1). Two control constructs contained a single
305-bp gus gene fragment inserted in the antisense
(pSIM755) or sense (pSIM718) orientation between the
35S promoter of Cauliflower mosaic virus (P35S) and the
terminator of the nopaline synthase (nos) gene. Introduction of these constructs into gus tobacco triggered
low frequencies of gene silencing (3%–6%) that are
typical for antisense and sense approaches (see, for
example, Smith et al., 2000; Singh et al., 2000; Table I).
Employment of a construct containing both a sense
and antisense copy of the gene fragment between promoter and terminator (pSIM374) confirmed the much
greater frequency of gene silencing that is obtained
Plant Physiology, August 2006, Vol. 141, pp. 1508–1518, www.plantphysiol.org Ó 2006 American Society of Plant Biologists
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
New Constructs for Gene Silencing
with hairpin RNAs (Waterhouse et al., 1998; Table I). In
contrast, the use of terminator-free constructs representing antisense (pSIM758) and sense (pSIM140)
approaches did not yield any silenced plants, and
the silencing frequency of a terminator-free construct
designed to produce hairpin RNA (pSIM777) was only
one-sixth that of pSIM374 (Fig. 1; Table I).
An important signal for RNA polyadenylation is the
near upstream element (NUE) with consensus sequence 5#-AAUAAA, located between 13 and 30 nucleotides upstream of the cleavage site (Loke et al.,
2005). Elimination of the corresponding DNA sequence
in terminator elements was predicted to negatively
affect mRNA 3#-end processing and, consequently,
reduce the efficacy of silencing. This concept was tested
by substituting the nos terminator of pSIM374 with a
derivative element lacking the NUE element (mT). Use
of the resulting vector pSIM376 was, indeed, shown to
decrease the frequency of gus gene silencing if compared to the original construct (P 5 0.05; Table I). Collectively, our data demonstrate an important link between
conventional gene silencing and mRNA 3#-end processing. Effective silencing appears to be associated with
the ability of the construct to process its transcripts.
Terminator-Free Silencing Constructs Containing
Inverted Repeats
To study whether gene silencing could be established by new types of expression cassettes, we generated terminator-free constructs that contained two
oppositely oriented promoters (Fig. 1). The first convergent transcription vector pSIM717 contained two
copies of the gus gene fragment positioned as inverted
repeat between P35S and the figwort mosaic virus
(FMV) 35S promoter (PFMV). Surprisingly, retransformation of gus-expressing tobacco plants with this
construct resulted in a high frequency of gene silencing. This frequency was at least as high as that of
the corresponding conventional silencing construct
pSIM374 (P 5 0.02; Table I; data not shown). Transformation with the second vector pSIM756, which is
identical to pSIM717 except that the gus gene fragments are oriented as divergent repeat, also yielded
similar gene silencing frequencies. This finding indicates that the orientation of the inverted repeat does not
play an important role in establishing gene silencing.
Because PFMV of pSIM717 and pSIM756 was identical
to the promoter driving the original gus gene expression cassette, gene silencing could have been induced
by promoter targeting. This possibility was excluded
by evaluating vector pSIM754, which contains
PFMV operably linked to P35S. None of the tested
Figure 1. Silencing constructs. A blue arrow indicates the position and
orientation of the gus gene fragment used as trigger for gene silencing; a
green box between arrows indicates a spacer. 35S, CMV 35S promoter;
FMV, FMV35S promoter; FMVT, FMV35S promoter lacking TATA box;
Ubi, promoter of the potato ubiquitin-7 gene; T, nos gene terminator;
mT, modified nos terminator lacking a NUE element; pA, synthetic
poly(A) tail. A pink arrow depicts a fragment of the tobacco Ppo gene;
an orange box indicates the intron of the gus gene. The expression
cassette for the hpt selectable marker gene was oriented in such a way
that the terminator of this cassette would not function in 3#-end
processing of transcripts produced by the silencing constructs.
Plant Physiol. Vol. 141, 2006
1509
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
Yan et al.
Table I. Constructs for transcript-targeted gene silencing
Eight-week-old plants that were PCR positive for both gus gene and
silencing construct were assayed by histochemical analysis of at least
three leaves per plant. TER, Terminator; UD, unidirectional; CV,
convergent transcription.
Construct
TER
Approach
Target
Plants Silencing
Assayed Frequency
%
pSIM755 Y
pSIM718 Y
pSIM374 Y
pSIM758 N
pSIM140 N
pSIM777 N
pSIM376 Y (mT)
pSIM717 N
pSIM756 N
pSIM754 N
UD antisense
UD sense
UD hairpin
UD antisense
UD sense
UD hairpin
UD hairpin
CV hairpin
CV hairpin
CV promoter
gus
gus
gus
gus
gus
gus
gus
gus
gus
gus
RNA
RNA
RNA
RNA
RNA
RNA
RNA
RNA
RNA
promoter
36
34
107
29
36
36
35
147
37
38
3
6
48
0
0
8
29
62
51
0
double-transformed tobacco plants displayed a reduced level of gus expression (Table I).
The stability of pSIM717-mediated gene silencing
was assessed by assaying groups of approximately 25
T1 and T2 plants derived from three randomly chosen
double transformants that segregated for single loci
carrying the silencing construct. Progeny plants that were
PCR positive for both the gus gene and the silencing
construct displayed a similar level of gus silencing as
determined for T0 plants, indicating that terminatorfree gene silencing is, at least in the T1 and T2 generation, not gradually diminished (data not shown).
The molecular basis of pSIM717-mediated gene
silencing was studied by isolating RNA from double
transformants. This RNA was used as template for
reverse transcription (RT) PCRs with the primer combinations pr1 to pr3 and pr1 to pr4 (Fig. 2A). These
experiments demonstrated that effective silencing of
the gus gene in plants such as 717-36 was correlated
with successful amplification of both cDNAs, indicative for the production of dsRNA (Fig. 2, A and B; data
not shown). In contrast, amplification of only one of
the cDNAs in plants 717-8 (with pr1–pr3) and 717-13
(with pr1–pr4) was associated with a lack of detectable
gene silencing (Fig. 2B). Similar results were obtained
for silenced plants transformed with the pSIM717derivative vector pSIM715, which contained a larger
intron separating the two gus gene fragments (Fig. 2B;
data not shown).
RNA gel-blot analyses using a gus gene fragment
amplified with primers pr1 and pr2 showed that
convergent transcription in fully silenced plants such
as 717-12 and 717-36 (Fig. 2, C and D), and 715-19, 71538, and 715-55 (Fig. 2D) resulted in both 3- to approximately 10-fold reduced levels of gus gene transcripts
and the production of new RNAs that were produced
by the silencing construct and shorter than the distance between the two driver promoters (Fig. 2, C and
D). In contrast, a lack of detectable gene silencing in,
for instance, plants 717-7, 717-8, 717-9, 717-10, and
717-13 was linked to less than 2-fold reduced gus gene
transcript levels (Fig. 2, C and D). These plants produced only limited amounts of the new small RNAs
(Fig. 2, C and D). The production of relatively large
amounts of variably sized (approximately 0.2–1.0 kb)
transcripts in silenced plants was confirmed with a
probe specific for the silencing construct (Fig. 2E).
Silenced plants such as 715-19 and 715-38 also produced rare, approximately 1-kb transcripts that extended into P35S and could be visualized after
extended exposure (Fig. 2F). The absence of similar
transcripts extending into PFMV (Fig. 2G) suggests that
this promoter is stronger and dominant over P35S.
The above-described studies had shown that
terminator-free sense (pSIM140) and antisense
(pSIM758) constructs did not trigger gene silencing
effectively. We therefore assumed that the rare readthrough transcripts of pSIM715 would not interfere
with the expression of neighboring genes. To test this
hypothesis, we compared the efficacy of the hygromycin phosphotransferase (hptII) selectable marker gene
of pSIM717 with that of pSIM374. Two weeks after
infection with Agrobacterium strains carrying the
binary vectors, tobacco explants had developed comparable numbers of hygromycin-resistant calli (Fig. 3A).
The apparent lack of an inadvertent regulatory effect
of the terminator-free expression cassette on the neighboring hptII gene was confirmed by carrying out realtime PCR. Figure 3B shows similar levels of hptII
transcript for 6-week-old pSIM374 and 717 plants that
displayed gus gene silencing. Each group of plants
contained one individual with a lower expression level.
This variation within the groups is probably due to
position-integration effects.
We studied whether the products of collisional transcription were polyadenylated by performing RT-PCRs
on the RNA of silenced plants. Control amplifications
were carried out with an oligo(dT) reverse primer
together with pr1. As expected, this reaction yielded a
strong band that was confirmed by sequencing to represent the polyadenylated 3# region of the gus transcript
(Fig. 4). However, amplification with the oligo(dT) and
pSIM717-specific pr4 primer combination yielded only
small quantities of cDNA that visualized on agarose
gels as a faint smear (Fig. 4). Sequence analysis confirmed that none of 20 cDNAs isolated from this smear
corresponded to transcripts produced by the silencing
construct (data not shown). Similar results were obtained by sequencing weak bands amplified with
oligo-T and pr3 (Fig. 4; data not shown). Our results
indicate that collisional transcription of inverted repeats triggers effective gene silencing through the
production of variably sized and generally nonpolyadenylated RNAs.
Alternative Gene-Based Silencing Constructs
In addition to silencing constructs containing inverted repeats, we also tested expression cassettes carrying
direct repeats (Fig. 1). Two gene fragments positioned
in the same orientation (pSIM779) already proved more
1510
Plant Physiol. Vol. 141, 2006
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
New Constructs for Gene Silencing
Figure 2. RNA analysis of gus plants containing a terminator-free silencing construct. A, Diagrams of expression cassettes for the
gus gene and the silencing constructs of pSIM717 and 715. The position of introns is indicated with horizontal striped bars.
Primers (pr1–pr4) used for RT-PCR are shown as black arrows. Probes, depicted as delineated gray lines, are derived from: the gus
gene (G), P35S (1), intron of the potato Gbss gene (I; Rohde et al., 1990), and PFMV (2). For further explanations, see the legend of
Figure 1. B, RT-PCR analysis of silenced (underlined) and nonsilenced pSIM717 and pSIM715 plants. Amplified products are
shown after separation on agarose gel and ethidium bromide staining. DNA fragments amplified with primers pr1 and pr2 can be
derived from RNAs transcribed from P35S, PFMV , or both. Because pSIM717 contains a Gbss-derived intron that is shorter than that
of pSIM715, products generated with the pr1 to pr3 primers are shorter for pSIM717 plants than for pSIM715 plants. C, RNA gelblot analysis of three pSIM717 plants together with a wild-type (WT) and original gus (GUS) plant (16-h exposure). The white
arrow shows the position of the gus gene transcript. Transcripts produced by the silencing construct of plants 717-12 and 717-36
vary in size from approximately 0.2 to approximately 1.0 kb. Silenced line numbers are underlined. CV, Transcripts specific for
convergent transcription. D, RNA gel-blot hybridized with the gus gene-derived probe G (24-h exposure). WT, Untransformed;
GUS, original gus expressing plant. E, Hybridization with probe I derived from the Gbss intron (15-h exposure). F, Hybridization
with the P35S probe 1 (5-d exposure). G, Hybridization with PFMV (5-d exposure) probe 2.
effective (8%) than the single gene fragment of vector
pSIM772 (3%; Table II). An additional 4- to 5-fold increase was obtained by employing expression cassettes
containing four direct repeats (pSIM787 and pSIM1111;
Table II). These results demonstrate that convergent
transcription of four direct repeats can be as effective
as that of an inverted repeat and suggest that the
complementary RNAs produced by convergent transcription hybridize efficiently to produce dsRNA.
Terminator-free silencing constructs can also be
designed to express the intron of a target gene. The
silencing construct of vector pSIM782 contains an inverted repeat consisting of two copies of the intron of
the gus gene inserted between P35S and PFMV. Though
only 3% of plants transformed with this construct were
partially silenced, it is interesting that targeting of
transient and nuclear pre-mRNAs can, in fact, result in
silencing of the corresponding gene (Table II).
Plant Physiol. Vol. 141, 2006
1511
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
Yan et al.
Promoter Substitutions Influence Silencing Efficacy
Figure 3. Effect of convergent-transcription constructs on expression of
the neighboring hptII gene. A, Transformation frequencies for pSIM374
(gray) and 717 (white) were determined by infecting tobacco explants
with the corresponding Agrobacterium strains, and, after 2 weeks,
counting the number of calli/explant. Results are means 6 SE of at least
12 explants per experiment. B, Real-time PCR was performed to
determine relative transcript levels of pSIM374 (gray) and pSIM717
(white). Data are the mean 6 SE of three independent measurements.
Multigene Silencing Constructs
Conventional gene silencing constructs have in
some cases been used to simultaneously downregulate the expression of multiple genes (Halpin
et al., 2001). The inverted repeats of such expression
cassettes consisted of sense and antisense copies of
two or more different gene fragments. Because transcripts produced by constructs such as pSIM717 were
often shorter than the distance between the two driving promoters, we reasoned that distal gene fragments
would be less effective than centrally located fragments in triggering gene silencing. This hypothesis
was confirmed by generating two binary vectors with
sense and antisense gene fragments from both the gus
and tobacco polyphenol oxidase (Ppo) genes inserted
between two promoters (Table II). Vector pSIM774
contained the gus gene fragments in distal positions
with the fragments from the Ppo gene positioned near
the center of the expression cassette. In contrast, vector
pSIM775 contained the central spacer flanked by gus
gene fragments. Analyses of double-transformed tobacco
plants demonstrated that the frequency of gus silencing
was lower for pSIM774 than for pSIM775 (P 5 0.007;
Table II). These results demonstrate that the efficacy of
gene silencing depends, indeed, on the position of the
trigger gene fragments within the terminator-free collisional transcription (TFCT) construct. Most likely, the
outer sequences are less frequently transcribed into
RNAs that fold back into hairpins.
To test the promoter specificity of TFCT silencing
constructs, new vectors were constructed that contained a sense and antisense fragment of the gus gene
inserted between various convergent promoter combinations. Histochemical analyses of retransformed
gus plants demonstrated that the expression cassette
of pSIM771, which combines the P35S with the promoter of the potato (Solanum tuberosum) ubiquitin-7
(PUbi7) gene (Garbarino et al. 1995), triggered a high
overall frequency of gus silencing (43%) approaching
that previously determined for pSIM717 (Tables I and
II). In contrast, the use of two P35S elements in pSIM770
dramatically lowered the silencing efficacy to 8% only
(Table II). The poor efficacy of this expression cassette
may be due to methylation of P35S. Although not
studied as part of the work described here, the accessibility of P35S to methylation was frequently observed
by others (Meyer et al., 1992; Dieguez et al., 1998).
Another combination of identical promoters for TFCT
constructs was tested as well. Vector pSIM789 contains
two copies of PFMV and triggered gene silencing about
as frequently (46%) as pSIM717 (Table II; Fig. 5).
Collectively, our data demonstrate that the efficacy of
TFCT-based gene silencing is to a large degree dependent on the choice of promoter combination.
TFCT-Mediated Gene Silencing of Endogenous Genes
Expressed in Potato Tubers
We also studied whether TFCT expression cassettes
could be exploited to down-regulate the expression of
endogenous genes. For this purpose, three new vectors
targeting the potato tuber-expressed Ppo gene were
constructed (Fig. 6A). The control vector pSIM217
represented a conventional silencing approach and
Figure 4. Amplification of transcripts from plants transformed with
pSIM717. The amplified DNAs are shown after separation on agarose
gel and ethidium bromide staining. Silenced line numbers are underlined. The position of a cDNA representing the 3# end of the gus transcript is indicated with a white arrowhead. The weaker band marked
with a gray band relates to a tobacco PR52 cDNA, which contains an
inadvertent pr1 binding site at 0.5 kb upstream from the cleavage site.
1512
Plant Physiol. Vol. 141, 2006
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
New Constructs for Gene Silencing
Table II. Alternative constructs for transcript-targeted gene silencing
For explanation, see legend of Table I.
Construct
TER
Approach
Target
Plants Silencing
Assayed Frequency
%
pSIM772
pSIM779
N
N
pSIM787
N
pSIM1111 N
pSIM782
pSIM774
pSIM775
pSIM771
pSIM789
pSIM770
N
N
N
N
N
N
CV sense
CV direct
repeat
CV direct
repeat
CV direct
repeat
CV hairpin
CV hairpin
CV hairpin
CV hairpin
CV hairpin
CV hairpin
gus RNA
gus RNA
35
36
3
8
gus RNA
32
37
gus RNA
36
42
gus intron
gus-PPO RNA
PPO-gus RNA
gus RNA
gus RNA
gus RNA
35
35
35
35
35
38
3
11
39
43
46
8
comprised both a sense and antisense 154-bp fragment
of the untranslated trailer of the potato tuberexpressed Ppo gene (Rommens et al., 2004) inserted
between the promoter of the granule-bound starch
synthase (Gbss) gene and the nos gene terminator.
Vector pSIM764 was identical to pSIM217 except that
the terminator was replaced by a second copy of the
Gbss promoter. The third vector, pSIM765, contained
the two Ppo gene fragments inserted in the opposite
orientation between two convergent Gbss promoters.
Agrobacterium strains carrying the various binary
vectors were used to transform the potato variety
Ranger Russet, and three copies of 21 PCR positive
lines for each construct were allowed to set tubers in
the greenhouse. Subsequent biochemical assays of
three tubers per line demonstrated that all lines
displayed reduced Ppo activity (Fig. 6, B and C). On
average, the activity was lower in pSIM764 plants
(36.8% 6 2.5% of wild-type levels) than in pSIM217
controls (49.5% 6 3.9%; P 5 0.007). The opposite
orientation of Ppo gene fragments in pSIM765 resulted
in an average activity of 41.2% 6 3.9%.
A second gene that was targeted for silencing was
the potato tuber-expressed phosphorylase-L (PhL) gene.
Conventional silencing approaches have shown that
lowered PhL gene expression triggers reduced Glc
accumulation in cold-stored tubers (Kawchuk et al.,
1999). These results were confirmed with vector pSIM216,
which contains an inverted repeat consisting of two
untranslated leader sequences inserted between the
same regulatory elements as used for pSIM216 (Fig.
6D). The TFTC vector pSIM847 differs from pSIM216
in that the inverted repeat is positioned between the
Gbss promoter and the promoter of the potato ADPGlc pyrophosphorylase gene (du Jardin and Berhin,
1991). Transgenic tubers containing this construct
displayed a level of reduced Glc accumulation that
was, on average, at least similar to that for pSIM216
(Fig. 6E). Because Glc levels provide only an indirect
indication of the efficacy of gene silencing, we also
studied transcript levels in randomly chosen pSIM216
and pSIM847 tubers. This analysis confirmed that the
two different constructs triggered similar reductions
in PhL gene expression (Fig. 6F). In both cases, PhL
expression levels were about 20-fold lower than those
of control plants containing only the selectable marker
gene. Thus, endogenous genes can be silenced effectively through convergent transcription.
Promoter-Targeted Gene Silencing
To compare the frequencies of gene-based silencing
constructs with constructs targeting promoter sequences,
we produced two vectors containing one (pSIM1112)
and two (pSIM773) copies of PFMV, which also drives
the gus target gene, between the convergent driver P35S
and PUbi7 elements (Fig. 1). Compared to the abovedescribed vector pSIM772, which contains a single copy
of the gus gene fragment, exploitation of pSIM1112 and
pSIM773 yielded higher frequencies of gene silencing
(Tables II and III). These results indicate that promoter
sequences can be more effective than gene-derived
sequences in triggering gene silencing. Furthermore,
the much stronger activity of pSIM773 compared to
pSIM1112 demonstrates the importance of dsRNA
generation for promoter-targeted gene silencing
approaches.
Interestingly, all silenced pSIM773 plants lacked any
gus expression (Fig. 7A). This phenotype is different
from the generally partial silencing that is triggered by
gene-based silencing constructs (Fig. 7; data not shown)
and indicates that gene silencing may be accomplished
more effectively by targeting promoters than through
Figure 5. pSIM789 plants. Histochemically stained leaf punches of
plants containing both the gus gene and the silencing construct of
pSIM789. Boxed leaf samples depict stained gus positive (top) and
wild-type (bottom) plants.
Plant Physiol. Vol. 141, 2006
1513
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
Yan et al.
Figure 6. Silencing of the potato tuber-expressed genes. A, Diagram of expression cassettes pSIM217, pSIM764, and pSIM765,
designed to target the Ppo gene. PG, Promoter of the Gbss gene; T, terminator of the potato ubiquitin-3 gene. An orange arrow
indicates a Ppo gene fragment and a green box depicts the intron of the potato ubiquitin-7 gene (B) Ppo enzyme activity. Data are
shown as percentages of wild type (OD410/g 5 0.15 6 0.01) and represent the mean of three tuber measurements per plant. C,
Tubers stained for Ppo activity by pipetting 0.5 mL of catechol (50 mM) on cut surfaces. D, Expression cassettes pSIM216 and
pSIM847 aimed to reduce Glc accumulation through silencing of the PhL gene. AG, ADP-Glc pyrophosphorylase promoter. E,
Glc concentrations (milligrams per gram) in cold-stored tubers. CNTRL, Transgenic control. Data are the mean 6 SE of three
independent measurements. F, PhL gene expression as determined by quantitative real-time PCR analyses of tuber RNA. Results
are means 6 SE of three experiments.
RNA degradation. Progeny analyses showed that the
level of silencing triggered by pSIM773 is stable in at
least the T1 and T2 generations (Fig. 7A; data not
shown). The efficacy of promoter-targeted silencing is
not limited to constructs that contain the target promoters in a convergent orientation. Similar frequencies
(77%) were obtained when these promoters were placed
in the opposite (divergent) direction in pSIM1120
(Table III; Fig. 7B).
The pSIM773-derived vector pSIM788 only contained
the target promoter sequences upstream from the
TATA box. Although this promoter fragment was shown
to be nonfunctional (data not shown), pSIM788 also
triggered effective gus gene silencing (Table III; Fig.
7C). This finding demonstrates, for the first time, that
nonfunctional target promoter fragments can activate
transgene silencing if expressed to produce dsRNA.
We also found that replacement of one of the two
driver promoters of pSIM773 by a terminator element
(pSIM1101) still provided effective silencing. In this
case, some of the retransformed plants displayed only
a partial gene silencing phenotype (Table III; Fig. 7D).
This result indicates that the presence of the terminator element may have influenced the plant’s silencing
response.
DISCUSSION
Terminator-Free Gene-Based Silencing Constructs
One aspect of the work presented here relates to the
significance of a terminator element in silencing constructs. Although little is understood about the processes and sequences that govern transcriptional
termination, the nos terminator that was used in our
studies is known to contain mRNA 3#-end processing
signals required for cleavage and polyadenylation
(Depicker et al., 1982). Removal of this element greatly
reduced the efficacy of constructs that contained a promoter element driving the expression of gene-derived
Table III. Constructs for promoter-targeted gene silencing
For explanation, see legend of Table I.
Construct
TER
Approach
pSIM1112
pSIM773
pSIM788
pSIM1120
pSIM1101
N
N
N
N
Y
CV sense
CV hairpin
CV hairpin
CV hairpin
UD hairpin
Target
Plants
Silencing
Assayed Frequency
%
gus
gus
gus
gus
gus
1514
promoter
promoter
promoter
promoter
promoter
36
35
36
35
34
8
57
60
73
59
Plant Physiol. Vol. 141, 2006
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
New Constructs for Gene Silencing
sense, antisense, or inverted repeat structures. This
result implies that 3#-end processing has a positive
effect on the ability of transcripts to function as substrates for Dcl proteins. In fruit flies (Drosophila melanogaster), the Ago protein associated with siRNAdirected silencing (Ago2) was shown to play a role in
poly(A) length maintenance (Siomi et al., 2005). Depletion of Ago2 resulted in mRNA stabilization and
shortened poly(A) tails. A similar link between mRNA
3#-end processing and silencing would explain the
importance of the terminator element in conventional
constructs for gene silencing in plants.
Interestingly, the need for a terminator element is
circumvented by placing the gene-derived trigger sequences between two convergent promoters. Our results indicate that the resulting transcripts are not
polyadenylated. In addition to RT-PCR experiments,
further studies on such RNAs are currently ongoing.
The activation of gene silencing by transcripts that did
not undergo full 3#-end processing is not without
precedent. In petunia, cosuppression of the chalcone
synthase gene was correlated with an accumulation of
nonpolyadenylated chalcone synthase RNA (Metzlaff
et al., 1997). Similarly, the initiation of silencing of the
MuDR/Mu transposition system in maize (Zea mays)
coincided with nonpolyadenylated RNA production
(Rudenko et al., 2003). Thus, aberrant transcripts produced by convergent transcription, cosuppression,
and transposable elements may represent highly accessible substrates for Dcl activity, possibly after processing by factors involved in the recognition of such
RNAs. Convergent transcription constructs display
about the same efficacy as conventional promoterto-terminator constructs in suppressing target genes.
We also found that the two different construct
approaches were most effective if used to produce
dsRNA. These resemblances imply an involvement of
similar Dcl proteins in transcript cleavage.
To some extent, the organization of convergent transcription constructs resembled that of rare transformation events where two T-DNAs that contain a
conventional silencing construct integrate at the same
locus in a divergent orientation. Read-through tran-
scription might in some cases produce hairpin RNAs
that trigger gene silencing. For instance, a transgenic
tomato (Lycopersicon esculentum) line silenced for the
polygalacturonase gene contains two oppositely oriented T-DNA inserts at the same locus. RNA gel-blot
analyses of this line might suggest the presence of
mRNAs that extend beyond the terminators (Sanders
and Hiatt, 2005). However, there may be alternative
explanations for this phenomenon. By studying transgenic plants containing multiple copies of a target
gene, gene silencing appeared to be linked to copy
number rather than inverted repeats (Lechtenberg et al.,
2003). Convergent transcription constructs have been
used previously for Dicer-dependent gene silencing in
mammalian cells. In contrast to our approach, these
constructs still contained the gene-derived sequences
linked to five consecutive thymine residues that function in transcript termination (Tran et al., 2003).
The new construct approaches described here may
facilitate efforts to fine tune the regulation of gene
silencing. For instance, it may be possible to specifically silence certain genes in cold-stored potato tubers
by using a silencing construct that contains both a
cold-inducible and tuber-specific promoter. Most
RNAs that were produced through convergent transcription had sizes that are shorter than the distance
between the two convergent promoters. However,
we did detect transcripts that extended into one of
the driver promoters. This finding was unexpected
because hairpin transcripts are susceptible to Dicer,
and transcripts lacking a poly(A) tail would be predicted
to be unstable. Given the poor efficacy of terminatorfree antisense constructs such as pSIM758, the production of rare long transcripts is unlikely to inadvertently
affect the expression of neighboring genes. Indeed, we
did not find convergent transcription to lower hptII
gene expression levels in pSIM717 plants.
In addition to the use of exon-derived sequences as
trigger for gene silencing, we also demonstrated the
efficacy of constructs containing intron DNA. This
finding indicates that nuclear pre-mRNAs can function as target for degradation. A similar retention of
RNAs was associated with MuDR/Mu silencing
Figure 7. Promoter-targeted gene silencing. Representative histochemically stained leaf punches are shown for pSIM773 (A),
pSIM1120 (B), pSIM788 (C), and pSIM1101 (D). Partially silenced leaf punches are shown in a green box.
Plant Physiol. Vol. 141, 2006
1515
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
Yan et al.
(Rudenko et al., 2003). However, intron-based gene silencing appears to be not as effective as silencing
methods that employ exon sequences. The targeting of
transient pre-mRNAs may play a much more limited
role in plants than in, for instance, Plasmodium falciparum. This parasite contains introns in its var genes that
act as transcriptional silencing elements to control
antigenic variation (Gannoun-Zaki et al., 2005).
Recently, an analysis of the Arabidopsis genome identified 956 gene pairs that overlap within their 3# regions
(Jen et al., 2005). In a similar way as demonstrated here
for sense and antisense constructs, simultaneous expression of these genes might trigger DNA silencing. The
occurrence of overlapping genes is even more prominent
in organisms that contain more closely packed genes
such as yeast (Saccharomyces cerevisiae). It has been
shown that convergent transcription of the yeast Pot1
and YIL161w genes causes a small but noticeable negative effect on the level of Pot1 mRNA and nucleosome
displacement in the intergenic region (Puig et al., 1999).
Furthermore, bacteria contain more than 100 convergently transcribed gene pairs that are involved in the
regulation of a variety of biological functions including
copy number control (Wagner and Flardh, 2002).
Our results demonstrate that convergent transcription of direct repeats can also successfully downregulate gene expression. Thus, effective gene silencing
does not necessarily require the production of hairpin
RNA. This finding suggests that separately produced
sense and antisense RNAs anneal rapidly to form
dsRNA. However, direct repeat strategies have also been
used with some success in a conventional promoterterminator context (Ma and Mitra, 2002; Lechtenberg
et al., 2003). Such approaches would be expected to
produce dsRNA less effectively and suggest alternative mechanisms of gene silencing.
Terminator-Free Promoter-Based Silencing Constructs
As opposed to the frequent occurrence of partial
gene silencing that is triggered by constructs comprising gus gene fragments, complete gene knockouts
were often obtained by employing constructs that
express promoter transcripts. Our promoter-based
silencing constructs lack any sequences that represent
the target gene and are, therefore, different from
previously described constructs containing both the
promoters themselves and 13- to 34-bp sequences downstream from their transcription start (Mette et al., 1999,
2000; Kanno et al., 2004, 2005). Thus, the silencing
efficacy of these earlier constructs could have been
influenced, at least in part, by their ability to express
untranslated leader sequences. Another unique aspect
of the promoter-based expression cassettes described
here is that they are not operably linked to a terminator. The earlier expressed cassettes either contained
a terminator or were functionally associated with the
terminator of an adjacent expression cassette. The
processing of at least some of the transcripts produced
by the promoter-based silencing constructs may have
influenced the silencing response. To the best of our
knowledge, the constructs described in our study
provide the first conclusive evidence that expressed
inverted repeats only containing promoter fragments
can be used to effectively silence genes that are driven
by the target promoters. The employed fragments can
represent functional or nonfunctional promoters and
can be oriented as convergent or divergent repeat.
The convergent transcription of inverted repeats
containing target promoter sequences triggers a high
frequency of complete gene silencing. The efficacy of
this approach suggests that ds promoter RNAs activate a pathway different from that triggered by genebased methods. Based on silencing pathways that
operate at the level of the nuclear genome, promoterderived dsRNAs are likely to trigger RNA-mediated
DNA methylation (Wassenegger et al., 1994; Chan
et al., 2005). The Dcl protein involved in this process
may be Dcl3, which produces siRNA of the 24-nt size
class that target cytosine methylation (Xie et al., 2004).
Proteins involved in this de novo methylation include
cytosine methyltransferases such as Drm1, Drm2,
Met-1, and Cmt3. The latter two enzymes are involved
in the maintenance of CG, CNG, and CNN methylation, respectively (Jones et al., 2001; Lindroth et al.,
2001; Chan et al., 2005), and may explain the complete
knockout phenotypes in, for instance, pSIM773 plants.
Replacement of one of the driver promoters by a
terminator negatively affects the efficacy of gene silencing with some plants still displaying the full gene
silencing typical for the corresponding terminator-free
constructs, but others are only partially silenced.
In summary, we have shown that expression cassettes lacking a terminator but containing two convergent promoters can be used to induce effective gene
silencing. The efficacy of these constructs implies that
silencing can be activated by a pool of nonprocessed
transcripts of variable size. The use of two different
promoters makes it possible to fine tune the regulation
of gene silencing.
MATERIALS AND METHODS
Development and Use of a gus-Expressing Tobacco Line
A transfer DNA only containing an expression cassette for the gus gene
flanked by T-DNA borders was introduced into tobacco (Nicotiana tabacum) by
employing a novel marker-free transformation method (Weeks and Rommens,
2003). Explant material for retransformation was obtained from propagated
6-week-old tissue culture material derived from a single homozygous T1
plant. Silencing constructs were introduced into the T1 plant as described
previously (Rommens et al., 2004). Gus expression levels were determined by
performing at least three independent assays per plant. Histochemical and
fluorometric assays were performed as described previously (Jefferson et al.,
1987; Lin et al., 1994).
Construction of Expression Cassettes
The gus gene fragment used as trigger for gene silencing represents the
sequences at position 1292-1596 of GenBank accession number S69414. Silencing constructs were inserted into the T-DNA of a binary vector next to an
expression cassette for the hptII selectable marker gene in such a way that the
1516
Plant Physiol. Vol. 141, 2006
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
New Constructs for Gene Silencing
associated terminator of the expression cassette was not operably linked to the
silencing constructs. The PFMV element was produced synthetically and
corresponds to nucleotides 6,368 to 6,949 of accession X06166. A truncated
promoter derivative lacking TATA box and downstream sequences comprises
the sequence from 6,368 to 6,887.
RNA Analysis
TRIzol (Invitrogen) was used to isolate RNA according to the manufacturer’s recommendations. The following primers were used for RT-PCRs:
5#-CAA CGC GTA AAC TCG ACC CGA CGC GTC (pr1), 5#-ATG CAC ACT
GAT ACT CTT CAC TCC AC (pr2), 5#-TTG TTT TTG TTC ATC TGT AGC TTC
TGC (pr3), and 5#-TGG AGG AGA TGA GTA AAA GTT ACC ACG (pr4). Gelblot analyses were carried out using Hybond-N1 membranes (Amersham
Biosciences). A 294-bp gus gene fragment amplified with primers pr1 and pr2
was used as probe to visualize the transcripts produced by expression of the
gus gene and silencing constructs. Additional probes targeted an intron
derived from the potato (Solanum tuberosum) Gbss gene that separates the two
gus gene fragments of vectors such as pSIM717, amplified with 5#-GCT GCA
GAA GCT ACA GAT GAA C and 5#-GAG TAA AAG TTA CCA CGA ATT C
(256 bp); P35S, amplified with 5#-GTC AAG AGT CCC CCG TGT TCT CTC and
5#-GCT TCA TGG AGT CAA AGA TTC AAA TAG (548 bp); and PFMV,
amplified with 5#-CAT TTA GCA GCA TTC CAG ATT GGG TTC and 5#-ATT
GGT TGA GTA TCT GAT GAT CCT TC (582 bp).
To determine hptII gene expression levels, RNA was isolated from 6-weekold plants and treated with RNAse-free DNAseI (Invitrogen). Quantitative
real-time RT-PCR reactions were performed using 50 ng RNA and the
QuantiTech SYBR green RT-PCR kit (Qiagen) according to the manufacturer’s
instructions. Internal control primers (5#-TCT TAG GGC TTT CGG GTA TGC
CA and 5#-CAA CTA CGG ATA TAT AAG AGC CAA AAC T) were designed
to amplify the cytochrome oxidase gene. Primers for the hptII gene are 5#-TGG
TTG GCT TGT ATG GAG CAG CAG and 5#-TGG TCA AGA CCA ATG CGG
AGC ATA. Expression levels were defined as percentage of the average levels
found in pSIM374 plants.
PhL gene expression was assessed by isolating potato tuber RNA using the
plant RNA purification reagent (Invitrogen). Quantitative real-time RT-PCRs
were carried out using 5#-AGT GGT CGT ACC ACC GGT ATT GTG and
5#-ATG ATC AGT GAG GTC ACG ACC TGC as internal control primers to
amplify the actin gene, and 5#-ATC CCA TTA CTG AAC AAG GTG GTG and
5#-CAA TGC TCT ACC CTG CAG AAA TTC for the PhL gene. Expression
levels in transgenic pSIM216 and pSIM847 plants were defined as percentage
of the average levels of transgenic controls (pSIM401).
Ppo Assays
The levels of Ppo enzyme activity were compared with wild-type levels by
mixing the pulverized tubers (1 g) of plants that had been grown for 6 weeks in
the greenhouse for 1 h in 50 mM MOPS buffer at pH 6.5 (5 mL). After precipitation
of the solid fraction, the change of OD410 was determined over time.
Glucose Determination in Cold-Stored Potato Tubers
Tubers of plants that had been grown for 6 weeks in the greenhouse were
harvested and incubated for 1 month at 4°C. After this cold storage, Glc levels
were determined by using a Glc oxidase/peroxidase reagent (Megazyme).
ACKNOWLEDGMENTS
We are grateful to Dr. Kathy Swords for fruitful discussion and a critical
review of the manuscript, and Scott Simplot and Bill Whitacre for continued
support. We also thank Dr. William Belknap for providing the potato
ubiquitin-7 promoter. Kristine Barney, Joanna Owen, and Lynda Zhang are
acknowledged for excellent technical assistance.
Received April 17, 2006; revised May 26, 2006; accepted June 4, 2006;
published June 9, 2006.
LITERATURE CITED
Chan SW, Henderson IR, Jacobsen SE (2005) Gardening the genome: DNA
methylation in Arabidopsis thaliana. Nat Rev Genet 6: 351–360
Dalmay T, Hamilton A, Rudd S, Angell S, Baulcombe DC (2000) An RNAdependent RNA polymerase gene in Arabidopsis is required for posttranscriptional gene silencing mediated by a transgene but not by a
virus. Cell 101: 543–553
Depicker A, Stachel S, Dhaese P, Zambryski P, Goodman HM (1982)
Nopaline synthase: transcript mapping and DNA sequence. J Mol Appl
Genet 1: 561–573
Dieguez MJ, Vaucheret H, Paszkowski J, Mittelsten Scheid O (1998)
Cytosine methylation at CG and CNG sites is not a prerequisite for the
initiation of transcriptional gene silencing in plants, but it is required for
its maintenance. Mol Gen Genet 259: 207–215
du Jardin P, Berhin A (1991) Isolation and sequence analysis of a cDNA
clone encoding a subunit of the ADP-glucose pyrophosphorylase of
potato tuber amyloplasts. Plant Mol Biol 16: 349–351
Dunoyer P, Himber C, Voinnet O (2004) DICER-LIKE 4 is required for RNA
interference and produces the 21-nucleotide small interfering RNA component of the plant cell-to-cell silencing signal. Nat Genet 37: 1356–1360
Gannoun-Zaki L, Jost A, Mu J, Deitsch KW, Wellems TE (2005) A silenced
Plasmodium falciparum var promoter can be activated in vivo through
spontaneous deletion of a silencing element in the intron. Eukaryot Cell
4: 490–492
Garbarino JE, Oosumi T, Belknap WR (1995) Isolation of a polyubiquitin
promoter and its expression in transgenic potato plants. Plant Physiol
109: 1371–1378
Gasciolli V, Mallory AC, Bartel DP, Vaucheret H (2005) Partially redundant functions of Arabidopsis DICER-like enzymes and a role for DCL4
in producing trans-acting siRNAs. Curr Biol 15: 1494–1500
Halpin C, Barakate A, Askari BM, Abbott JC, Ryan MD (2001) Enabling
technologies for manipulating multiple genes on complex pathways.
Plant Mol Biol 47: 295–310
Hamilton AJ, Baulcombe DC (1999) A species of small antisense RNA in
posttranscriptional gene silencing in plants. Science 286: 950–952
Jefferson RA, Kavanagh TA, Bevan MW (1987) GUS fusions, betaglucuronidase as a sensitive and versatile gene fusion marker in higher
plants. EMBO J 6: 3901–3907
Jen CH, Michalopoulos I, Westhead DR, Meyer P (2005) Natural antisense
transcripts with coding capacity in Arabidopsis may have a regulatory role
that is not linked to double-stranded RNA degradation. Genome Biol 6: R51
Jones L, Ratcliff F, Baulcombe DC (2001) RNA-directed transcriptional
gene silencing in plants can be inherited independently of the RNA
trigger and requires Met1 for maintenance. Curr Biol 11: 747–757
Kanno T, Aufsatz W, Jaligot E, Mette MF, Matzke M, Matzke AJ (2005) A
SNF2-like protein facilitates dynamic control of DNA methylation.
EMBO Rep 6: 649–655
Kanno T, Mette MF, Kreil DP, Aufsatz W, Matzke M, Matzke AJ (2004)
Involvement of putative SNF2 chromatin remodeling protein DRD1 in
RNA-directed DNA methylation. Curr Biol 14: 801–805
Kawchuk LM, Armstrong JD, Lynch DR, Knowles NR, inventors.
December 7, 1999. Potatoes having improved quality characteristics and
methods for their production. United States Patent Application No. 5998701
Lechtenberg B, Schubert D, Forsbach A, Gils M, Schmidt R (2003) Neither
inverted repeat T-DNA configurations nor arrangements of tandemly
repeated transgenes are sufficient to trigger transgene silencing. Plant J
34: 507–517
Lin J-J, Assad-Garcia N, Kuo J (1994) Effects of Agrobacterium cell concentration on the transformation efficiency of tobacco and Arabidopsis
thaliana. Focus 16: 72–77
Lindroth AM, Cao X, Jackson JP, Zilberman D, McCallum CM, Henikoff
S, Jacobsen SE (2001) Requirement of CHROMOMETHYLASE3 for
maintenance of CpXpG methylation. Science 292: 2077–2080
Liu J, Carmell MA, Rivas FV, Marsden CG, Thomson JM, Song JJ,
Hammond SM, Joshua-Tor L, Hannon GJ (2004) Argonaute2 is the
catalytic engine of mammalian RNAi. Science 305: 1437–1441
Loke JC, Stahlberg EA, Strenski DG, Haas BJ, Wood PC, Li QQ (2005)
Compilation of mRNA polyadenylation signals in Arabidopsis revealed
a new signal element and potential secondary structures. Plant Physiol
138: 1457–1468
Ma C, Mitra A (2002) Intrinsic direct repeats generate consistent posttranscriptional gene silencing in tobacco. Plant J 31: 37–49
McCormick FP, Barton KA, Swain WF, inventors. September 9, 2003.
Effecting virus resistance in plants through the use of negative strand
RNAs. United States Patent No. 6617496
Mette MF, Aufsatz W, van der Winden J, Matzke MA, Matzke AJ (2000)
Plant Physiol. Vol. 141, 2006
1517
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.
Yan et al.
Transcriptional silencing and promoter methylation triggered by doublestranded RNA. EMBO J 19: 5194–5201
Mette MF, van der Winden J, Matzke MA, Matzke AJ (1999) Production of
aberrant promoter transcripts contributes to methylation and silencing
of unlinked homologous promoters in trans. EMBO J 18: 241–248
Metzlaff M, O’Dell M, Cluster PD, Flavell RB (1997) RNA-mediated RNA
degradation and chalcone synthase A silencing in petunia. Cell 88: 845–854
Meyer P, Linn F, Heidmann I, Meyer H, Niedenhof I, Saedler H (1992)
Endogenous and environmental factors influence 35S promoter methylation of a maize A1 gene construct in transgenic petunia and its colour
phenotype. Mol Gen Genet 231: 345–352
Morel JB, Godon C, Mourrain P, Beclin C, Boutet S, Feuerbach F, Proux F,
Vaucheret H (2002) Fertile hypomorphic ARGONAUTE (ago1) mutants
impaired in post-transcriptional gene silencing and virus resistance.
Plant Cell 14: 629–639
Puig S, Perez-Ortin JE, Matallana E (1999) Transcriptional and structural
study of a region of two convergent overlapping yeast genes. Curr
Microbiol 39: 369–373
Rohde W, Becker D, Kull B, Salamini F (1990) Structural and functional
analysis of two waxy gene promoters from potato. J Genet Breed 44: 311–315
Rommens CM, Humara JM, Ye J, Yan H, Richael C, Zhang L, Perry R,
Swords K (2004) Crop improvement through modification of the plant’s
own genome. Plant Physiol 135: 421–431
Rudenko GN, Ono A, Walbot V (2003) Initiation of silencing of maize
MuDR/Mu transposable elements. Plant J 33: 1013–1025
Sanders RA, Hiatt W (2005) Tomato transgene structure and silencing. Nat
Biotechnol 23: 287–289
Shewmaker CK, Kridl JC, Hiatt WR, Knauf V, inventors. April 21, 1992.
Anti-sense regulation of gene expression in plant cells. United States
Patent No. 5107065
Singh S, Green A, Stoutjesdijk P, Liu Q (2000) Inverted-repeat DNA: a
new gene-silencing tool for seed lipid modification. Biochem Soc Trans
28: 925–927
Siomi MC, Tsukumo H, Ishizuka A, Nagami T, Siomi H (2005) A potential link between transgene silencing and poly(A) tails. RNA 11:
1004–1011
Smith NA, Singh SP, Wang MB, Stoutjesdijk PA, Green AG, Waterhouse
PM (2000) Total silencing by intron-spliced hairpin RNAs. Nature 407:
319–320
Tran N, Cairns MJ, Dawes IW, Arndt GM (2003) Expressing functional
siRNAs in mammalian cells using convergent transcription. BMC
Biotechnol 6: 21
van der Krol AR, Mur LA, Beld M, Mol JN, Stuitje AR (1990) Flavonoid
genes in petunia: addition of a limited number of gene copies may lead
to a suppression of gene expression. Plant Cell 2: 291–299
Wagner EG, Flardh K (2002) Antisense RNAs everywhere? Trends Genet
18: 223–226
Wassenegger M, Heimes S, Riedel L, Sanger HL (1994) RNA-directed de
novo methylation of genomic sequences in plants. Cell 76: 567–576
Waterhouse PM, Graham MW, Wang MB (1998) Virus resistance and gene
silencing in plants can be induced by simultaneous expression of sense
and antisense RNA. Proc Natl Acad Sci USA 95: 13959–13964
Weeks T, Rommens CM, inventors. October 2, 2003. Refined plant transformation. World Patent Application No. 079765A2
Xie Z, Allen E, Wilken A, Carrington JC (2005) DICER-LIKE 4 functions in
trans-acting small interfering RNA biogenesis and vegetative phase
change in Arabidopsis thaliana. Proc Natl Acad Sci USA 102: 12984–12989
Xie Z, Johansen LK, Gustafson AM, Kasschau KD, Lellis AD, Zilberman
D, Jacobsen SE, Carrington JC (2004) Genetic and functional diversification of small RNA pathways in plants. PLoS Biol 2: E104
1518
Plant Physiol. Vol. 141, 2006
Downloaded from on August 9, 2017 - Published by www.plantphysiol.org
Copyright © 2006 American Society of Plant Biologists. All rights reserved.