Báo cáo khoa học: A neomorphic sgs3 allele stabilizing miRNA cleavage products reveals that SGS3 acts as a homodimer

A neomorphic sgs3 allele stabilizing miRNA cleavage products reveals that SGS3 acts as a homodimer Taline Elmayan1, Xavier Adenot1, Lionel Gissot1,2, Dominique Lauressergues1,3, Isabelle Gy1 and Herve´ Vaucheret1

1 Laboratoire de Biologie Cellulaire, Institut Jean-Pierre Bourgin, INRA, Versaille, France 2 Plateforme de Cytologie et d’Imagerie Ve´ ge´ tale, Institut Jean-Pierre Bourgin, INRA, Versailles, France 3 UMR 5546 CNRS ⁄ Universite´ Toulouse III, Surfaces Cellulaires et Signalisation chez les Ve´ ge´ taux, Castanet Tolosan, France

Keywords homodimer; neomorphic allele; RNA stabilization; RNAi; siRNA

silencing and trans-acting small

Correspondence T. Elmayan, Laboratoire de Biologie Cellulaire, Institut Jean-Pierre Bourgin, INRA (Institut National de la Recherche Agronomique), 78026 Versailles, Cedex, France Fax: +33 01 30 83 30 99 Tel: +33 01 30 83 30 29 E-mail: elmayan@versailles.inra.fr

(Received 1 October 2008, revised 20 November 2008, accepted 3 December 2008)

The putative RNA-binding protein SUPPRESSOR OF GENE SILEN- CING 3 (SGS3) protects RNA from degradation before transformation into dsRNA by the RNA-dependent RNA polymerase RDR6 during plant post-transcriptional gene interfering (siRNA) pathways. In this study, we show that SGS3 acts as a homodimer, and that the point mutation sgs3-3 impairs post-transcriptional gene silenc- ing in a dominant-negative manner through the formation of SGS3 ⁄ sgs3-3 heterodimers. Unlike complete-loss-of-function sgs3 mutants, which are impaired in the accumulation of both micro RNA-directed TAS cleavage products and mature trans-acting siRNAs, the sgs3-3 mutant overaccumu- lates TAS cleavage products and exhibits slightly reduced trans-acting siRNA accumulation. Together, these results suggest that sgs3-3 is a neo- morphic allele that shows increased RNA protective activity, resulting in decreased RNA processing by downstream post-transcriptional gene silenc- ing and trans-acting siRNA pathway components.

doi:10.1111/j.1742-4658.2008.06828.x

RNA silencing was discovered in plants more than 18 years ago during the course of plant transgenic experiments that triggered silencing of the introduced transgene and, in some cases, the homologous endo- genous genes or resident transgenes [1–5].

gene

[PTGS;

silencing

Post-transcriptional

also referred to as RNA interference (RNAi)] requires the interfering RNAs (siRNAs) [6], production of small which direct the degradation of complementary RNA [7,8]. PTGS represses the expression of endogenous genes, but, in plants, also targets viral RNAs and, as is an important and widespread viral defence such,

mechanism [9]. In Arabidopsis thaliana, the genetic sense-transgene PTGS (S-PTGS) and dissection of antiviral RNA silencing pathways has revealed ARGONAUTE 1 (AGO1), DICER-LIKE 2 (DCL2), DCL4, HUA ENHANCER 1 (HEN1), RNA-DEPEN- DENT RNA POLYMERASE 6 (RDR6), SILENC- ING DEFECTIVE 5 (SDE5) and SUPPRESSOR OF GENE SILENCING 3 (SGS3) as essential compo- nents [7,8,10]. S-PTGS and antiviral PTGS are trig- gered by the production of aberrant ssRNA, which is converted into dsRNA by RDR6 with the help of SGS3 and SDE5 [8]. These dsRNAs subsequently are

Abbreviations AGO1, ARGONAUTE 1; ARF, auxin response factor; BiFC, bimolecular fluorescence complementation; DAPI, 4¢,6-diamidino-2-phenylindole; DCL2 ⁄ 4, DICER-LIKE 2 ⁄ 4; GUS, b-glucuronidase; HEN1, HUA ENHANCER 1; miRNA, micro RNA; nat-siRNA, natural-antisense siRNA; PPR, pentatricopeptide repeat; PTGS, post-transcriptional gene silencing; RDR6, RNA-DEPENDENT RNA POLYMERASE 6; RNAi, RNA interference; SDE5, SILENCING DEFECTIVE 5; SGS3, SUPPRESSOR OF GENE SILENCING 3; siRNA, small interfering RNA; S-PTGS, sense-transgene PTGS; ta-siRNA, trans-acting siRNA; YFP, yellow fluorescent protein.

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

835

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

Results and Discussion

sgs3-3 behaves as a semi-dominant allele over SGS3

diced by DCL4 ⁄ DCL2 (two RNAse III proteins that use dsRNA as substrates) to siRNAs [7,8,11]. Like micro RNAs (miRNAs), siRNAs are methylated by HEN1 and guide the sequence-specific cleavage of complementary mRNA after loading on to AGO1 slicer, an RNAse H domain-containing protein [7,8]. Orthologues of AGO1, DCL and HEN1, and, in some cases, RDR, also function during RNAi in animals [12].

recessive.

Indeed, heterozygous

the transcripts of

the pentatricopeptide repeat

A forward genetic screen using the silenced 35S::GUS line L1 identified 59 PTGS-deficient mutants, including one sgs1, 23 rdr6, seven sgs3, 12 ago1, one hen1 and two met1 alleles, as well as 13 unclassified mutants [8]. Genetic complementation tests performed among the 46 classified mutants indicated that all mutants, except sgs3-3 one, were mutants escaped L1 PTGS at the early stages of devel- opment (Fig. 1A), but often triggered PTGS at later stages of development (data not shown), indicating that sgs3-3 behaves as a semi-dominant allele over wild-type SGS3 during PTGS. The sgs3-3 mutation causes a Glu to Ala substitution between the first and second coiled-coil domains of the SGS3 protein. By contrast, each of the other sgs3 mutant alleles is reces- sive and harbours a premature termination codon (Fig. S1A). Because sgs3-3 transcripts do not contain a premature termination codon, they do not enter the non-sense-mediated decay pathway, a fate presumed the recessive sgs3 mutants. for Indeed, sgs3-3 mutants accumulate SGS3 mRNA at wild-type levels, whereas two different sgs3 mutant alleles with premature show termination codons reduced SGS3 mRNA accumulation (Fig. S1B).

In plants, in addition to the exogenously produced siRNAs, different classes of endogenous small RNAs, miRNA, natural-antisense siRNA (nat-siRNA) and trans-acting siRNA (ta-siRNA) [13] guide the repres- sion of endogenous transcripts complementary to these small RNAs. Based on their shared require- ments for RDR6, SDE5, SGS3, DCL4, HEN1 and transgene-derived RNAs and viral RNAs AGO1, appear to activate an RNA-based silencing process similar to the ta-siRNA pathway. Indeed, after miRNA- directed AGO1 or AGO7 cleavage of non-coding transcripts transcribed from endogenous TAS loci, RDR6, SDE5, SGS3, DCL4 and HEN1 process these aberrant transcripts and produce functional ta-siRNAs [7,8,10,14–17]. Once produced, ta-siRNAs guide AGO1-mediated cleavage of endogenous com- plementary mRNA targets, which include members of (PPR), MYB and auxin response factor (ARF) gene families [18]. Null rdr6, sde5, sgs3 and dcl4 mutant plants exhibit down- ward curled leaf margins as a result of accelerated juvenile to adult vegetative phase change caused by elevated ARF3 and ARF4 transcription factor levels, which normally are repressed by TAS3 ta-siRNAs [14,19–23].

the curling of

SGS3 is a plant-specific protein that acts upstream of RDR6-dependent dsRNA production [16]. The SGS3 protein contains three known protein domains: a zinc finger XS domain, a true XS domain that probably acts as an RNA recognition motif [24] and three coiled-coil domains that probably mediate pro- tein–protein interactions [25]. Consistent with the role of SGS3 in the plant antiviral response, SGS3 is directly targeted by the tomato yellow leaf curl geminivirus PTGS suppressor protein V2 [26]. SGS3 appears to protect TAS RNA cleavage products from degradation after miRNA-guided cleavage of primary TAS transcripts, probably facilitating their conversion to dsRNA by RDR6 and their subsequent DCL4- mediated dicing into ta-siRNA [16]. However, the precise role of SGS3 during PTGS remains unclear. In this study, we present structural and functional analyses of SGS3 that reveal new insights into its mode of action.

To examine further the effect of the sgs3-3 mutation, we analysed the integrity of the PTGS and ta-siRNA pathways in a series of plants expressing various doses of endogenous SGS3 in combination with the com- plete-loss-of-function sgs3-1 or semi-dominant sgs3-3 mutations. Analyses of L1 PTGS confirmed that sgs3- 3 behaves as a dominant allele over SGS3 at early stages of development (Fig. 1A). Indeed, 3-week-old SGS3 ⁄ SGS3 and sgs3-1 ⁄ SGS3 plants were silenced, whereas SGS3 ⁄ sgs3-3 plants exhibited high b-glucuron- idase (GUS) activity, similar to the level observed in sgs3-3 ⁄ sgs3-3, sgs3-3 ⁄ sgs3-1 and sgs3-1 ⁄ sgs3-1 mutants (Fig. 1A). GUS expression in these plants correlated with the disappearance of GUS siRNAs (Fig. 1C), indicating that SGS3 ⁄ sgs3-3 plants are impaired in PTGS even though a wild-type SGS3 copy is present. Similarly, analyses of leaf margins like sgs3-3 ⁄ sgs3-3 revealed that SGS3 ⁄ sgs3-3 leaves, and sgs3-3 ⁄ sgs3-1 leaves, are curled slightly more than wild-type leaves, indicating minor ta-siRNA pathway defects (Fig. 1B). Although the effect of sgs3-3 on leaf curling was less pronounced than in sgs3-1 ⁄ sgs3-1 mutants, SGS3 ⁄ sgs3-3 and sgs3-3 ⁄ sgs3-3 leaf shape dif- fered significantly from that of SGS3 ⁄ SGS3 plants (see

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

836

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

A

4000

3000

2000

i

1000

y t i v i t c a S U G

·

0

1 – ) t o r p g µ ( · 1 – n m U L F

4

3.4

t

3

2.6*

t

2.8** 2.6***

2.3

B h g n e

i

2.2

l f

h d w

2

f

a e L

/

1

a e L = R

0

Fig. 1. sgs3-3 acts as a semi-dominant allele over SGS3. Dosage effect of 3-week-old wild-type SGS3, sgs3-3 and sgs3-1 alleles on transgene PTGS and ‘zippy’ leaf phenotype. All lines are in the Col accession and are homozygous for the L1 locus. (A) GUS activity quantification (in fluorescence units per minute per microgram of total protein). Mean relative levels ± standard error (SE) for 12 plants of each genotype are reported. (B) Zippy phenotype is reported as the average ratio of the length by width of the fifth and sixth rosettes leaves of 12 plants of each genotype. The horizontal line denotes the average wild-type leaf measurement. Means rela- tive levels ± SE are reported. The number of SGS3 or sgs3-3 gene copies is indicated. The significance of the differences from the control values (SGS3 ⁄ SGS3) were assessed by Student’s t-test (*P < 0.03; **P < 0.0002; ***P < 0.008). (C) A low-molecular-mass RNA blot was probed with DNA complementary to the internal cod- ing sequence of GUS and with a DNA oligonucleotide probe com- plementary to the TAS3 ta-siRNAs (5¢D7+). U6 snRNA hybridization served as a loading control. A high-molecular-mass RNA blot was probed with DNAs complementary to ARF4. 25S rRNA hybridization served as a loading control. Normalized values of ARF4 to 25S rRNA, and GUS siRNAs and TAS3 ta-siRNAs to U6 snRNA are indicated.

3 S G S / 3 S G S

1 - 3 s g s / 3 S G S

3 - 3 s g s / 3 S G S

1 - 3 s g s / 1 - 3 s g s

3 - 3 s g s / 3 - 3 s g s

1 - 3 s g s / 3 - 3 s g s SGS3 copy 2 0 0 1 1 0 sgs3-3 copy 0 0 2 0 1 1

C

showed

1 - 3 s g s / 1 - 3 s g s

3 - 3 s g s / 3 - 3 s g s

3 S G S / 3 S G S

1 - 3 s g s / 3 S G S

3 - 3 s g s / 3 S G S

1 - 3 s g s / 3 - 3 s g s

GUS siRNA

GUS/U6

1

0

0

0.5 0 0

TAS3

TAS3/U6

1 0.3 0.6 1 0.9 0.7

U6

ARF4

transformants

all

ARF4/25S

1 3.2 1.2 1 1.5 1.9

25S rRNA

are functional in the ta-siRNA pathway, because 82% and 92% of the plants, respectively, restored wild-type in sgs3-1 null mutants (data not leaf development shown). GUS mRNA analyses that L1 ⁄ 35S::SGS3 transformants were PTGS competent, similar to wild-type plants, indicating that overexpres- sion of wild-type SGS3 does not impair L1 PTGS (Fig. 2). In contrast, the majority of L1 ⁄ 35S::sgs3-3 transgenic plants accumulated GUS mRNA to a high level (Fig. 2), indicating that, like SGS3 ⁄ sgs3-3 hetero- zygotes, the L1 ⁄ 35S::sgs3-3 transformants are PTGS deficient. PTGS impairment in L1 ⁄ 35S::sgs3-3 trans- genic plants was not a result of cosuppression of the 35S::sgs3-3 transgene and SGS3 endogenous gene accumulated ta-siRNA because in SGS3 ⁄ sgs3-3 (Fig. 2). These results suggest that, heterozygous plants and SGS3 ⁄ SGS3 plants trans- formed with 35S::sgs3-3, the mutant sgs3-3 protein probably poisons the activity of the wild-type SGS3 protein during PTGS.

t-test in Fig. 1B), consistent with the slight reduction of TAS3 siRNAs and the increased accumulation of its ARF4 target mRNA in SGS3 ⁄ sgs3-3 and sgs3-3 ⁄ sgs3-3 plants (Fig. 1B,C).

Yeast two-hybrid assays suggest that SGS3 dimerizes

sgs3-3 protein poisons wild-type SGS3 during PTGS

Our results suggest that the mutant sgs3-3 protein poi- sons the activity of functional wild-type SGS3 protein. To test this hypothesis, we overexpressed wild-type SGS3 (35S::SGS3) and mutant sgs3-3 (35S::sgs3-3) protein in L1 plants and examined the effect on L1 PTGS. Both the 35S::SGS3 and 35S::sgs3-3 constructs

The SGS3 protein contains three coiled-coil domains that probably promote protein–protein interactions. If SGS3 acts as a dimer in the TAS and PTGS pathways, mutant sgs3-3 protein could poison the PTGS pathway by dimerizing with wild-type SGS3 protein, and conse- functional SGS3 quently reducing the amount of homodimer in SGS3 ⁄ sgs3-3 plants and 35S::sgs3-3- expressing plants. To test the capacity of SGS3 to

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

837

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

35S::SGS3

L1

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18

GUS mRNA

25S rRNA

TAS2 siRNA

U6

35S::sgs3-3

L1

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18

GUS mRNA

25S rRNA

TAS2 siRNA

U6

Fig. 2. sgs3-3 overexpression inhibits PTGS in wild-type plants. RNA gel blot analysis of GUS mRNA accumulation in an L1 untransformed plant and L1 transgenic plants overexpressing SGS3 or sgs3-3 cDNA (numbers correspond to independent primary transformants). Total RNAs were extracted from mature leaves of 3-week-old primary transgenic plants and probed with DNA complementary to GUS. 25S rRNA hybridization served as a loading control. Total RNAs were also probed with a DNA oligonucleotide complementary to TAS2 ta-siRNA (seq F) to ensure that the transgenic plants did not undergo cosuppression of endogenous SGS3 and the 35S::SGS3 or 35S::sgs3-3 transgenes. U6 RNA hybridization served as a loading control.

and

of YN(cid:2)sgs3-3

dimerize, we performed yeast two-hybrid assays and observed that wild-type full-length SGS3 protein inter- acted with itself when used either as bait or as prey, and that the three C-terminal coiled-coil domains of SGS3 were sufficient for this interaction (Fig. 3A). An interaction was not observed between SGS3 and the unrelated lamine protein (Fig. 3A). Supporting our dimerization hypothesis, full-length mutant sgs3-3 pro- tein heterodimerized with the wild-type full-length SGS3 protein and the wild-type SGS3 coiled-coil domains alone (Fig. 3A). We also observed sgs3-3 homodimerization. These results suggest that SGS3 functions as a dimer and that mutant sgs3-3 protein disrupts PTGS by heterodimerizing with wild-type SGS3 protein.

SGS3 dimerizes in planta

To test SGS3 dimerization in vivo, we performed bimo- lecular fluorescence complementation (BiFC) experi- ments. For this purpose, SGS3, sgs3-3 and sgs3-10, which encodes a truncated form of SGS3 protein

(Fig. S1), were fused translationally to the C-terminal (YC) or N-terminal (YN) portion of yellow fluorescent protein (YFP). A fluorescence signal was observed when YN(cid:2)SGS3 and SGS3(cid:2)YC were co-infiltrated in Nicotiana benthamiana leaves (Fig. 3B). Co-infiltration of YN(cid:2)SGS3 and YC(cid:2)SGS3 gave similar results (Fig. S2), whereas co-infiltration of SGS3(cid:2)YN and YC(cid:2)SGS3 or SGS3(cid:2)YN and SGS3(cid:2)YC gave a weaker signal or no detectable signal, respectively (data not shown). Dimerization was observed after sgs3-3(cid:2)YC, co-infiltration YN(cid:2)sgs3-3 and SGS3(cid:2)YC (Fig. 3B) and YN(cid:2)sgs3-3 and YC(cid:2)sgs3-3 (Fig. S2), indicating that mutant sgs3-3 protein is able to homodimerize and also hete- the rodimerize with wild-type SGS3 protein. Thus, sgs3-3 probably semi-dominant behaviour of is explained by the ability of the sgs3-3 mutant protein to poison the activity of SGS3 by forming a non-func- tional SGS3 ⁄ sgs3-3 heterodimer. We did not detect a dimerization signal with the truncated sgs3-10 mutant protein, indicating that this protein, which is truncated in the second coiled-coil domain, is unable to dimerize

Fig. 3. Wild-type SGS3 and mutant sgs3-3 proteins dimerize in yeast and in planta. (A) SGS3 and sgs3-3 interaction in yeast two-hybrid assay. Results are summarized for the interaction observed between SGS3, sgs3-3 and the coiled-coil C-terminal part of the protein in fusion with the LexA binding domain and ⁄ or the GAL4 activation domain in L40 yeast. An unrelated protein, lamine, in fusion with LexA was used as a negative control. (B) SGS3 and sgs3-3 interactions in planta. Subcellular localization of reconstructed YFP was determined in the leaf epidermis of N. benthamiana for wild-type SGS3 protein and sgs3-3 mutant protein in fusion with the N- and ⁄ or C-terminal part of YFP (YN and YC, respectively). Channels are indicated above each column. The overlay is a YFP ⁄ chlorophyll autofluorescence overlay. Scale bars, 10 lm. (C) Subcellular localization of GFP was determined in transgenic Arabidopsis lines expressing wild-type SGS3 protein and sgs3-3 mutant protein in fusion with the C-terminal part of GFP. Channels are indicated above each column. The overlay is a DAPI ⁄ GFP ⁄ chlorophyll autofluorescence overlay. Scale bars, 10 lm.

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

838

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

A

GAL4~SGS3 GAL4~sgs3-3

– –

LexA~SGS3 + +

LexA~sgs3-3 LexA~coiled-coil x3 LexA~lamine + +

+ +

B

Overlay

YFP channel

DIC channel

Chlorophyll Autofluorescence channel

YN~SGS3 SGS3~YC

YN~sgs3-3 SGS3~YC

YN~sgs3-3 sgs3-3~YC

DAPI

GFP

Chlorophyll

Merge

C

sgs3-3~GFP line2

sgs3-3~GFP line4

SGS3~GFP line1

SGS3~GFP line3

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

839

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

ize in our BiFC analyses (Fig. S2), suggesting that dimerization plays a role in the formation of these cytoplasmic foci.

Overexpression of sgs3-3 does not restore PTGS

with the same efficiency as wild-type or sgs3-3 mutant protein (Fig. S2). Because sgs3-10 behaves like a null allele, it is probable that SGS3 dimerization is required for its function during PTGS and the ta-siRNA path- way. It remains to be tested whether RNA plays a role during this dimerization.

that

sgs3-3 protein as

foci observed was

reduced for

Our BiFC analyses in N. benthamiana suggested that, when expressed under the control of the 35S promoter, SGS3 dimers localized in the cytoplasm (Fig. 3B), which is consistent with previously published data showing that SGS3 localizes in the cytoplasm [26]. To confirm this localization, we expressed transiently in N. benth- amiana and stably in Arabidopsis 35S::SGS3(cid:2)GFP and 35S::sgs3-3(cid:2)GFP which, like 35S::SGS3 and 35S::sgs3- 3, complement the zippy leaf defects of sgs3-1 mutants In N. benthamiana leaves and Arabidopsis (Fig. 4). leaves and flowers, both wild-type SGS3 and mutant sgs3-3 protein localized to the cytoplasm and endoplas- mic reticulum (Fig. 3C, and data not shown). We did not detect a fluorescence signal in the nuclei stained by 4¢,6-diamidino-2-phenylindole (DAPI). However, we cannot rule out the possibility that these proteins are present in the nucleus at a level below our detection limit. It has been reported previously that, in the cyto- plasm, SGS3 accumulates in uncharacterized foci [26]. Consistently, we frequently observed fluorescence in cytoplasmic foci for both 35S::SGS3 and 35S::sgs3-3 (Fig. 3C and data not shown). However, the number 35S::sgs3-10 of (Fig. 3C, and data not shown), which does not dimer-

Nevertheless,

35S::sgs3-3(cid:2)GFP

l

sgs3-3 GFP

GFP sgs3-3

GFP SGS3

SGS3 GFP

o C / 1 L

1 - 3 s g s / 1 1 2 3 4 1 2 3 4 5 6 1 2 3 1 2 L

GUS TAS1

TAS2

U6

TAS3 U6

Zip phenotype

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

- +

-

SGS3

Transgene Endogene

25S

The formation of SGS3 ⁄ sgs3-3 heterodimers explains how sgs3-3 could behave as a dominant allele over SGS3. However, it is unclear why sgs3-3 behaves as a null allele in PTGS, but as a hypomorphic allele in the ta-siRNA pathway. It is possible that disparities between PTGS and ta-siRNA substrate dosages account for the differential effects. Indeed, TAS genes are expressed at lower levels than 35S promoter-dri- ven transgenes. Therefore, if sgs3-3 is a hypomorphic allele, it is possible that it retains sufficient activity to partly protect TAS cleavage products from degrada- tion after miRNA-guided cleavage, but this activity is insufficient to protect GUS transcripts tar- geted by siRNA because they are much more abun- dant. To examine whether RNA protection activity is limiting in sgs3-3 mutants, we overexpressed wild- fusion type SGS3 and mutant proteins with GFP under the control of the 35S promoter in sgs3-1 mutants, and scored for restora- tion of the TAS pathway and PTGS. Although the C-terminal GFP fusion constructs 35S::sgs3-3(cid:2)GFP and 35S::SGS3(cid:2)GFP appeared to be more robust, all fusion constructs appeared to be functional for like 35S::SGS3 and the TAS pathway because, 35S::sgs3-3, they complemented the leaf phenotype of sgs3-1 mutants and restored ta-siRNA accumulation (Fig. 4). and 35S::GFP(cid:2)sgs3-3 transformants accumulated mature levels below that of wild-type plants, ta-siRNA at similar to the sgs3-3 mutant, indicating that overex- pression of the mutant sgs3-3 protein does not result in increased production of mature ta-siRNA. More- over, only the SGS3 constructs, but not the sgs3-3 constructs, complemented the L1 PTGS defects of sgs3-1 mutants (Fig. 4), indicating that sgs3-3 does not support PTGS, even when expressed at high doses. Together, these results indicate that increasing the sgs3-3 mutant protein dosage does not restore S-PTGS in sgs3-1 mutants, suggesting that sgs3-3 is probably not a hypomorphic allele.

sgs3-3 probably is a neomorphic allele

respectively. Leaf phenotype (zip)

Fig. 4. sgs3-3 overexpression does not rescue PTGS in sgs3-1 mutants. RNA gel blot analyses of high- and low-molecular-mass RNAs extracted from inflorescences of the indicated plants. High- molecular-mass RNA blots were probed with SGS3 DNA. Low- molecular-mass RNA blots were probed with GUS DNA and DNA oligonucleotides complementary to TAS1 ta-siRNA 480 (+), TAS2 ta-siRNA (seq F) and TAS3 ta-siRNAs (5¢D7+). 25S rRNA and U6 snRNA hybridizations served as loading controls for high- and low-molecular-mass blots, is indicated.

To explore the more upstream effects of sgs3-3 on the ta-siRNA pathway, we analysed the accumulation of TAS cleavage products. Complete-loss-of-function sgs3 cleavage mutations destabilize TAS1

and TAS2

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

840

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

sgs3-3 ⁄ SGS3 and sgs3-3 ⁄ sgs3-1 plants. Moreover, sgs3-1 mutants transformed with 35S::sgs3-3 accumulate TAS cleavage products at higher levels than 35S::SGS3 plants (Fig. 5B), supporting the conclusion that sgs3-3 protein hyperprotects TAS cleavage products and, by doing this, probably decreases the efficiency of down- stream ta-siRNA maturation reactions, such as those performed by RDR6 or SDE5, and, consequently, decreases mature ta-siRNA production.

In contrast with the limited effect of sgs3-3 on the ta-siRNA pathway, PTGS is impaired in the sgs3-3 mutant as in the null allele sgs3-1, suggesting that PTGS is more sensitive to SGS3 perturbation than is the ta-siRNA pathway. PTGS impairment in the sgs3- 3 mutant is not specific to the L1 transgene. Among the three sgs3 mutants recovered in another forward the cosuppressed genetic screen for suppressors of 35S::NIA2 line 2a3, one mutant (sgs3-9) harbours the same point mutation as sgs3-3 (Fig. S1A) [19]. In addi- tion, sgs3-3 and sgs3-9 are as impaired in cucumber mosaic virus silencing as is the null sgs3-1 allele [19]. Because mutant sgs3-3 protein is able to form a

products generated after miR173-guided, AGO1-medi- ated cleavage of primary TAS1 and TAS2 transcripts, and are epistatic to complete-loss-of-function rdr6 mutations, suggesting that SGS3 protects TAS cleav- age products from degradation to allow RDR6 to use these RNAs as templates for dsRNA production [16]. Consistent with this function, TAS1 and TAS2 cleav- age products accumulate at reduced levels in leaves and inflorescences of the sgs3-1 mutant (Fig. 5A). By contrast, TAS1 and TAS2 cleavage products overaccu- mulate in the sgs3-3 mutant, whereas the levels of miR173 are unchanged in sgs3-1 and sgs3-3 mutants indicating compared with wild-type plants (Fig. 5A), that the changes in TAS cleavage product accumula- tion in the sgs3-1 and sgs3-3 mutants are not a result of changes in miR173 accumulation. Rather, these results suggest that the mutant sgs3-3 protein shows increased protective activity towards TAS cleavage products and that sgs3-3 is a neomorphic allele. Lend- ing to this hypothesis, which explains the dominant behaviour of sgs3-3, TAS cleavage products not only overaccumulate in sgs3-3 ⁄ sgs3-3 plants, but also in

Inflorescences

A

Leaves 3 - 3 s g s / 3 - 3 s g s

TAS1

TAS1

pri 3′ 5′

pri 3′ 5′

TAS2

pri 3′ 5′

TAS2

pri 3′ 5′

25S

25S

miR173

U6

1 - 3 s g s / 3 S G S 3 - 3 s g s / 3 S G S 1 - 3 s g s / 3 - 3 s g s 1 - 3 s g s / 1 - 3 s g s 3 S G S / 3 S G S 1 - 3 s g s / 1 - 3 s g s 1 - 3 s g s / 3 S G S 3 - 3 s g s / 3 - 3 s g s 3 - 3 s g s / 3 S G S 1 - 3 s g s / 3 - 3 s g s 3 S G S / 3 S G S 0 1 - 3 s g s / 0 1 - 3 s g s

B

GFP sgs3-3

sgs3-3 GFP

SGS3 GFP

GFP SGS3

TAS1

pri 3′ 5′

25S

1 2 3 4 1 2 3 4 5 6 1 2 3 1 2 l o C / 1 L l o C / 1 L 1 - 3 s g s / 1 L 1 - 3 s g s / 1 L

8 0

0 1

6 1

6 1

4 1

2 1

0 1

0 1

0 1

0 1

1 1

6 0

7 0

8 0

6 0

7 0

9 0

4 0

4 0

3′ cleaved/uncleaved

. . . . . . . . . . . . . . . . . . .

9 1

8 1

8 2

7 2

6 2

6 2

2 2

9 1

7 1

9 1

1 2

3 1

3 1

5 1

3 1

6 1

8 1

4 0

4 0

5′ cleaved/uncleaved

Fig. 5. sgs3-3 exhibits enhanced TAS cleav- age product protective activity. (A) High- and low-molecular-mass RNA gel blot analyses of mature rosette leaves and inflorescences of the indicated mutant plants. (B) High- molecular-mass RNA gel blot analyses of inflorescences of the indicated transgenic sgs3-1 and controls. High-molecular-mass RNA blots were probed with DNAs comple- mentary to TAS1a and TAS2 precursors. Low-molecular-mass RNA blots were probed with DNA oligonucleotides comple- mentary to miR173. The expected migration positions of primary TAS RNA precursors (pri) and the 5¢ and 3¢ cleavage products generated after miR173-guided cleavage are indicated on the left. 25S rRNA and U6 snRNA hybridizations served as loading con- trols for high- and low-molecular-mass blots, respectively. The ratios of TAS1a 5¢ and 3¢ cleavage products to uncleaved TAS1a precursor are indicated.

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

841

. . . . . . . . . . . . . . . . . . .

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

Molecular analyses and RNA blot probes and quantification

Molecular analyses (DNA sequencing, RNA gel blot analy- sis and GUS fluorimetric assays) were performed as described previously [31]. All RNA gel blot analyses were performed using 10 lg of total RNA. TAS1a and TAS2 probes for the detection of the corresponding precursor and cleavage products were amplified by PCR using the follow- ing pairs of primers: TAS1a: 5¢-GAGACTACACAACACC CATTAC-3¢ and 5¢-AGGTTCCGCCTTTAGATCGAG-3¢; TAS2: 5¢-GTGCTTCACAATGCTCTTTC-3¢ and 5¢-AGG ACAGAATCTCCTGTCAC-3¢. 32P-labelled DNA probes were prepared as described previously [19]. 32P-end-labelled oligonucleotide probes corresponding to TAS1 (si-480(+)), TAS2 (si-F), TAS3 (5¢D7+), U6 and 25S rRNA were pre- pared as described previously [19]. RNA hybridization sig- nals were quantified using multigauge software on a Fuji (Woodbridge, CT, USA) phosphorimager.

GUS assays were performed on rosette leaves of 3-week- old plants. All RNAs were extracted from mature leaves (three to six leaves) and from inflorescences at the same stage of development. GUS, ARF4 and 25S probes have been described previously [19].

Yeast two-hybrid and BiFC cloning assays

spectrum of

two-hybrid assays, SGS3,

homodimer, like wild-type SGS3, PTGS impairment is probably not caused by a lack of dimerization. One possible reason that PTGS is impaired, whereas the ta-siRNA pathway is only slightly inhibited in the sgs3-3 mutant, is because PTGS involves a threshold RNA level that conditions the triggering of an siRNA amplification step, whereas the ta-siRNA pathway appears to lack this threshold requirement and amplifi- cation loop. Indeed, very little secondary ta-siRNAs that come from amplification were cloned in wild-type plants [27,28]. In contrast, L1 produces large amounts of siRNA, and we have reported previously that the sgs1-1 mutation that causes a twofold decrease in the transcription of the 35S::GUS transgene is sufficient to abolish L1 PTGS [29]. In the absence of a threshold requirement and amplification, slight perturbations in SGS3 activity would be translated directly into a slight change in ta-siRNA production. By contrast, during PTGS, a slight decrease in primary siRNA production could prevent the production of secondary siRNA needed for PTGS amplification. In this way, a slight decrease in SGS3 activity could exact a major negative consequence on the efficiency of PTGS. Future studies combining loss-of-function sgs3 mutants and the newly identified neomorphic sgs3-3 allele will help to reveal roles played by SGS3 and the full its interacting partners in the ta-siRNA and PTGS pathways.

Experimental procedures

Plant material

Arabidopsis plants and all mutants were grown in con- trolled growth chambers, as described previously [19]. Transformations were performed as described previously [30].

Cloning and constructs

produce

and

sgs3-3 full-length For yeast cDNAs and the C-terminal three coiled-coil domains of SGS3 (corresponding to the amino acids 463–626) were cloned into vectors pGAD3S2X and pLex10 [32] to pro- duce fusion proteins with the GAL4 activation domain and LexA DNA-binding domain, respectively. To confirm the specificity of the interaction and to reduce false-positive clones, L40 strain carrying an unrelated protein, the lamine fused to LexA, was transformed with all the constructs in fusion with the GAL4 activation domain. SGS3, sgs3-3 and sgs3-10 full-length cDNAs were cloned in the GATE- WAY(cid:2)-compatible vector pDONR207 (Invitrogen, Cergy, for RT-PCR: 5¢- France) using the following primers GGGGACAAGTTTGTACAAAAAAGCAGGCTTAATG AGTTCTAGGGCTGGTC-3¢ and 5¢-GGGGACCACTTTG TACAAGAAAGCTGGGTAATCATCTTCATTGTGAAG GC-3¢. RT-PCR was performed on sgs3-10 mutant plants to produce sgs3-10 full-length cDNA. cDNAs were trans- ferred into the BiFC GATEWAY(cid:2)-modified vector devel- oped by F. Parcy (CEA-Grenoble, France) prior to being recombined with the N- and C-terminal parts of YFP (YN 35S::YN-sgs3-3 ⁄ sgs3-10 ⁄ SGS3, to and YC) 35S::YC-sgs3-3 ⁄ sgs3-10 ⁄ SGS3 35S::sgs3-3 ⁄ sgs3- 10 ⁄ SGS3-YN, 35S::sgs3-3 ⁄ sgs3-10 ⁄ SGS3-YC destination vectors. Nicotiana benthamiana plants grown in the glass- house under 13 h light, 25 (cid:3)C day temperature and 17 (cid:3)C night temperature were used for all the agroinfiltration

To make the 35S constructs, SGS3 and sgs3-3 cDNA, corre- sponding to an SspI-HincII 3.3 kb fragment, which includes the SGS3 ORF and its 3¢ UTR (the 327 bp sequence down- stream of the STOP codon) and covers its 5¢ UTR (with 705 bp sequence upstream of the START codon), were cloned in the binary vector pBIB-HYG prior to transfer into Agrobacterium tumefaciens (C58 pmp90) and transforma- tion. QuikChange(cid:2) (Stratagene, Arcueil, France) PCR- based mutagenesis was used to create the mutant sequence encoding the E500K sgs3-3 mutant protein using the follow- ing primers: 5¢-GAACAGAACAGGGAAAAGATGGATG CACACGAC-3¢ and 5¢-GTCGTGTGCATCCATCTTTTC CCTGTTCTGTTC-3¢.

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

842

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

4 Smith CJ, Watson CF, Bird CR, Ray J, Schuch W & Grierson D (1990) Expression of a truncated tomato polygalacturonase gene inhibits expression of the endog- enous gene in transgenic plants. Mol Gen Genet 224, 477–481.

5 van der Krol AR, Mur LA, de Lange P, Mol JN &

Stuitje AR (1990) Inhibition of flower pigmentation by antisense CHS genes: promoter and minimal sequence requirements for the antisense effect. Plant Mol Biol 14, 457–466.

6 Hamilton AJ & Baulcombe DC (1999) A species of

small antisense RNA in posttranscriptional gene silenc- ing in plants. Science 286, 950–952.

experiments. The leaves were infiltrated with an overnight culture of Agrobacterium tumefaciens strain C58C1 that was resuspended to an absorbance (A) at 600 nm of 0.1 with the infiltration medium (10 mm MES, pH 5.6, 10 mm MgCl2, 200 lm acetosyringone). For co-infiltration, equal volumes of both Agrobacterium cultures were mixed before infiltration. Observations were performed 72 h after infiltra- tion. Confocal microscopy was performed on an inverted laser-scanning micro- TCS-SP2-AOBS spectral confocal (Leica Microsyste` mes SAS, Rueil-Malmaison, scope France) as previously described [33]. Samples were excited with a 514 nm argon laser (50%) with an emission band of 520–550 nm for YFP detection and 640–700 nm for chloro- phyll autofluorescence.

7 Brodersen P & Voinnet O (2006) The diversity of RNA silencing pathways in plants. Trends Genet 22, 268–280.

8 Vaucheret H (2006) Post-transcriptional small RNA

Localization of SGS3 in fusion with GFP

pathways in plants: mechanisms and regulations. Genes Dev 20, 759–771.

9 Ding SW & Voinnet O (2007) Antiviral immunity direc-

ted by small RNAs. Cell 130, 413–426.

10 Hernandez-Pinzon I, Yelina NE, Schwach F, Stud-

to

the

produce

35S::GFP-SGS3 ⁄ sgs3-3

laser

holme DJ, Baulcombe D & Dalmay T (2007) SDE5, the putative homologue of a human mRNA export factor, is required for transgene silencing and accumu- lation of trans-acting endogenous siRNA. Plant J 50, 140–148.

11 Mlotshwa S, Pruss GJ, Peragine A, Endres MW, Li J, Chen X, Poethig RS, Bowman LH & Vance V (2008) DICER-LIKE2 plays a primary role in transitive silenc- ing of transgenes in Arabidopsis. PLoS ONE 3, e1755.

cDNAs of SGS3 and sgs3-3 cloned in the GATEWAY(cid:2) compatible vector pDONR207 (Invitrogen) were trans- ferred to the binary vectors pB7WGF2 and pB7FWG2 [34] and 35S::SGS3 ⁄ sgs3-3-GFP destination vectors, respectively. sgs3-1 mutants were transformed with the resulting vec- tors. Confocal microscopy was performed on an inverted Leica TCS-SP2-AOBS spectral confocal scanning microscope, as described in [33]. Leaves of each transgenic plant stained with 0.5 mm DAPI were excited with a 488 nm argon laser (30%) with an emission band of 495– 540 nm for GFP detection and 640–700 nm for chloro- phyll autofluorescence.

12 Bartel DP (2004) MicroRNAs: genomics, biogenesis,

mechanism, and function. Cell 116, 281–297.

Acknowledgements

13 Chapman EJ & Carrington JC (2007) Specialization

and evolution of endogenous small RNA pathways. Nat Rev Genet 8, 884–896.

14 Peragine A, Yoshikawa M, Wu G, Albrecht HL &

We thank Allison Mallory for helpful discussions and critical reading of the manuscript. This work was sup- ported by a PhD fellowship from the Minister of Research to X.A. and ANR-06-BLAN-0203 grant to H.V.

Poethig RS (2004) SGS3 and SGS2 ⁄ SDE1 ⁄ RDR6 are required for juvenile development and the production of trans-acting siRNAs in Arabidopsis. Genes Dev 18, 2368–2379.

15 Vazquez F, Vaucheret H, Rajagopalan R, Lepers C,

References

Gasciolli V, Mallory AC, Hilbert JL, Bartel DP & Crete P (2004) Endogenous trans-acting siRNAs regulate the accumulation of Arabidopsis mRNAs. Mol Cell 16, 69– 79.

1 Linn F, Heidmann I, Saedler H & Meyer P (1990) Epi- genetic changes in the expression of the maize A1 gene in Petunia hybrida: role of numbers of integrated gene copies and state of methylation. Mol Gen Genet 222, 329–336.

2 Matzke MA, Primig M, Trnovsky J & Matzke AJ

16 Yoshikawa M, Peragine A, Park MY & Poethig RS (2005) A pathway for the biogenesis of trans-acting siRNAs in Arabidopsis. Genes Dev 19, 2164–2175. 17 Montgomery TA, Howell MD, Cuperus JT, Li D,

(1989) Reversible methylation and inactivation of mar- ker genes in sequentially transformed tobacco plants. EMBO J 8, 643–649.

Hansen JE, Alexander AL, Chapman EJ, Fahlgren N, Allen E & Carrington JC (2008) Specificity of ARG- ONAUTE7–miR390 interaction and dual functionality in TAS3 trans-acting siRNA formation. Cell 133, 128– 141.

3 Napoli C, Lemieux C & Jorgensen R (1990) Introduc- tion of a chimeric chalcone synthase gene into petunia results in reversible co-suppression of homologous genes in trans. Plant Cell 2, 279–289.

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

843

T. Elmayan et al.

Arabidopsis SGS3 acts as a homodimer

18 Mallory AC & Bouche N (2008) MicroRNA-directed

29 Elmayan T, Balzergue S, Beon F, Bourdon V, Daubr-

regulation: to cleave or not to cleave. Trends Plant Sci 13, 359–367.

19 Adenot X, Elmayan T, Lauressergues D, Boutet S,

emet J, Guenet Y, Mourrain P, Palauqui JC, Vernhettes S, Vialle T et al. (1998) Arabidopsis mutants impaired in cosuppression. Plant Cell 10, 1747–1758.

30 Mourrain P, Beclin C, Elmayan T, Feuerbach F,

Bouche N, Gasciolli V & Vaucheret H (2006) DRB4- dependent TAS3 trans-acting siRNAs control leaf mor- phology through AGO7. Curr Biol 16, 927–932.

Godon C, Morel JB, Jouette D, Lacombe AM, Nikic S, Picault N et al. (2000) Arabidopsis SGS2 and SGS3 genes are required for posttranscriptional gene silencing and natural virus resistance. Cell 101, 533–542.

31 Gy I, Gasciolli V, Lauressergues D, Morel JB, Gombert

20 Allen E, Xie Z, Gustafson AM & Carrington JC (2005) microRNA-directed phasing during trans-acting siRNA biogenesis in plants. Cell 121, 207–221.

21 Fahlgren N, Montgomery TA, Howell MD, Allen E,

J, Proux F, Proux C, Vaucheret H & Mallory AC (2007) Arabidopsis FIERY1, XRN2, and XRN3 are endogenous RNA silencing suppressors. Plant Cell 19, 3451–3461.

32 Elmayan T, Fromentin J, Riondet C, Alcaraz G, Blein JP & Simon-Plas F (2007) Regulation of reactive oxygen species production by a 14-3-3 protein in elicited tobacco cells. Plant Cell Environ 30, 722–732.

Dvorak SK, Alexander AL & Carrington JC (2006) Regulation of AUXIN RESPONSE FACTOR3 by TAS3 ta-siRNA affects developmental timing and patterning in Arabidopsis. Curr Biol 16, 939–944. 22 Garcia D, Collier SA, Byrne ME & Martienssen RA (2006) Specification of leaf polarity in Arabidopsis via the trans-acting siRNA pathway. Curr Biol 16, 933–938.

23 Hunter C, Willmann MR, Wu G, Yoshikawa M, de la Luz Gutierrez-Nava M & Poethig SR (2006) Trans-act- ing siRNA-mediated repression of ETTIN and ARF4 regulates heteroblasty in Arabidopsis. Development 133, 2973–2981.

33 Gissot L, Polge C, Jossier M, Girin T, Bouly JP, Kreis M & Thomas M (2006) AKINbetagamma contributes to SnRK1 heterotrimeric complexes and interacts with two proteins implicated in plant pathogen resistance through its KIS ⁄ GBD sequence. Plant Physiol 142, 931–944.

24 Zhang D & Trudeau VL (2008) The XS domain of a

plant specific SGS3 protein adopts a unique RNA rec- ognition motif (RRM) fold. Cell Cycle 7, 2268–2270. 25 Bateman A (2002) The SGS3 protein involved in PTGS

34 Karimi M, Inze D & Depicker A (2002) GATEWAY vectors for Agrobacterium-mediated plant transforma- tion. Trends Plant Sci 7, 193–195.

finds a family. BMC Bioinformatics 3, 21.

Supporting information

26 Glick E, Zrachya A, Levy Y, Mett A, Gidoni D, Belau- sov E, Citovsky V & Gafni Y (2008) Interaction with host SGS3 is required for suppression of RNA silencing by tomato yellow leaf curl virus V2 protein. Proc Natl Acad Sci USA 105, 157–161.

sgs3-10

sgs3-3

and

interactions

27 Howell MD, Fahlgren N, Chapman EJ, Cumbie JS,

The following supplementary material is available: Fig. S1. sgs3 allelic series. Fig. S2. SGS3, in planta.

This supplementary material can be found in the

online version of this article.

Sullivan CM, Givan SA, Kasschau KD & Carrington JC (2007) Genome-wide analysis of the RNA-DEPEN- DENT RNA POLYMERASE6 ⁄ DICER-LIKE4 path- way in Arabidopsis reveals dependency on miRNA- and tasiRNA-directed targeting. Plant Cell 19, 926–942. 28 Rajagopalan R, Vaucheret H, Trejo J & Bartel DP

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corre- sponding author for the article.

(2006) A diverse and evolutionarily fluid set of micro- RNAs in Arabidopsis thaliana. Genes Dev 20, 3407–3425.

FEBS Journal 276 (2009) 835–844 Journal compilation ª 2009 FEBS. No claim to original French government works

844