Seed Dormancy in Arabidopsis Is Controlled by … Dormancy in Arabidopsis Is Controlled by...

Seed Dormancy in Arabidopsis Is Controlled byAlternative Polyadenylation of DOG11

Malgorzata Cyrek2, Halina Fedak2, Arkadiusz Ciesielski, Yanwu Guo3, Aleksandra Sliwa, Lien Brzezniak,Katarzyna Krzyczmonik, Zbigniew Pietras, Szymon Kaczanowski, Fuquan Liu, and Szymon Swiezewski*

Institute of Biochemistry and Biophysics, Department of Protein Biosynthesis, Pawinskiego 5a, 02–106 Warsaw,Poland (M.C., H.F., A.C., Y.G., A.S., L.B., K.K., Z.P., S.K., S.S.); Warsaw University, Department of Chemistry,Pasteura 1, 02-093 Warsaw, Poland (A.C.); and Queen’s University Belfast, School of Biological Sciences,Belfast BT9 7BL, Northern Ireland (F.L.)

ORCID IDs: 0000-0001-7449-8081 (M.C.); 0000-0001-8764-6370 (F.L.).

DOG1 (Delay of Germination 1) is a key regulator of seed dormancy in Arabidopsis (Arabidopsis thaliana) and other plants.Interestingly, the C terminus of DOG1 is either absent or not conserved in many plant species. Here, we show that inArabidopsis, DOG1 transcript is subject to alternative polyadenylation. In line with this, mutants in RNA 39 processing complexdisplay weakened seed dormancy in parallel with defects in DOG1 proximal polyadenylation site selection, suggesting that theshort DOG1 transcript is functional. This is corroborated by the finding that the proximally polyadenylated short DOG1 mRNA istranslated in vivo and complements the dog1 mutant. In summary, our findings indicate that the short DOG1 protein isoformproduced from the proximally polyadenylated DOG1mRNA is a key player in the establishment of seed dormancy in Arabidopsisand characterizes a set of mutants in RNA 39 processing complex required for production of proximally polyadenylated functionalDOG1 transcript.

The ability to postpone germination, to align it withpermissive environmental conditions, greatly enhancesthe chances of plants survival (Koornneef et al., 2002;Penfield and King, 2009). Therefore, seed dormancy isextensively controlled by many factors, including hu-midity, temperature, and light, and is also subject toparental memory (Finch-Savage and Leubner-Metzger,2006; Chiang et al., 2011; Graeber et al., 2012; Nonogaki,2014). DOG1 (Delay of Germination 1) has been identi-fied as amajor quantitative trait locus for seed dormancy

variability among natural Arabidopsis (Arabidopsisthaliana) accessions, and dog1 T-DNA insertionalmutants exhibit reduced seed dormancy (Bentsinket al., 2006). The expression of DOG1 increases duringseedmaturation, and themRNAdisappears quickly afterimbibition, although the DOG1 protein is more stable(Nakabayashi et al., 2012). A recent report indicated thatDOG1 enforces seed dormancy by strengthening theseed coat, which occurs via modulation of the expressionof gibberellin metabolism genes (Graeber et al., 2014).The expression of DOG1 is extensively regulated, beingstrongly induced by abscisic acid and low temperatureduring seedmaturation (Chiang et al., 2011; Kendall et al.,2011). Other factors required for DOG1 expression in-clude histone-modifying enzymes such as the histoneH2B ubiquitin transferase HUB (Liu et al., 2007) and H3Lys 9 methyl transferase KYP (Zheng et al., 2012). DOG1expression is also highly dependent on transcriptionelongation factor TFIIS (Grasser et al., 2009).

The DOG1 gene is comprised of three exons, and theexon 2-exon 3 junction is subject to alternative splicing,generating four different forms of mRNA (Bentsinket al., 2006). The function of these alternatively splicedtranscript isoforms is unknown, and their relativeratio remains unchanged during seed development(Bentsinket al., 2006).Our recent analysis of the regulationofalternative splicing of DOG1 in a mutant defective inPolII elongation suggested that the rate of transcriptelongation regulates alternative splice site selectionin accordance with the kinetic coupling model andimplies that DOG1 splicing is cotranscriptional (Dolataet al., 2015).

1 This work was supported by MSHE (grant nos. IdP2011–000461and 2011/01/D/NZ8/03690 to S.S.). H.F., M.C., A.S., and Z.P. weresupported by Foundation for Polish Science Grant NumberTEAM2010–5/9. L.B. was supported by IP2011 004671. S.S. was sup-ported by EMBO Installation Grant CEBM/05/17.

2 These authors contributed equally to the article.3 Present address: Friedrich Miescher Institute, 4002 Basel,

Switzerland.* Address correspondence to [email protected] author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Szymon Swiezewski ([email protected]).

S.S. and F.L. designed the project and planned the program ofresearch; H.F., M.C., and A.S. performed the germination tests andexpression level measurements; H.F. and M.C. performed the west-ern blots; L.B. did the RACE experiment; S.S. and S.K. performed thebioinformatics analysis; A.C. and H.F. examined localization; G.Y.performed the DOG1-GFP IP; L.B. and A.S. performed the RNA sta-bility assay; K.K and Z.P. prepared the short and long DOG1 lines forcomplementation; S.S. and M.C. wrote the article.

www.plantphysiol.org/cgi/doi/10.1104/pp.15.01483

Plant Physiology�, February 2016, Vol. 170, pp. 947–955, www.plantphysiol.org � 2016 American Society of Plant Biologists. All Rights Reserved. 947 www.plantphysiol.orgon June 16, 2018 - Published by Downloaded from

Copyright © 2016 American Society of Plant Biologists. All rights reserved.

http://orcid.org/0000-0001-7449-8081

http://orcid.org/0000-0001-8764-6370

mailto:[email protected]

http://www.plantphysiol.org

mailto:[email protected]

http://www.plantphysiol.org/cgi/doi/10.1104/pp.15.01483


To our knowledge, complementation of the dog1mutation in Arabidopsis has not been achieved using aDOG1 cDNA, while the seed dormancy phenotype ofthis mutant was complemented using a genomicDOG1clone from Lepidium sativum (Graeber et al., 2014). No-tably, the L. sativum DOG1 gene lacks the exon 3 so canonly encode a short two-exonic mRNA, with no alter-natively spliced exon 2-exon 3 isoforms.

Many alternatively spliced genes are also subject toalternative polyadenylation (APA; Di Giammartinoet al., 2011). APA leads to the generation of transcriptswith different 39 ends through a series of steps cata-lyzed by components of RNA 39 processing complexeslike the Cleavage and Polyadenylation Specificity Fac-tor (CPSF) and Cleavage Stimulation Factor (CstF)complexes, and poly(A) polymerases (Gruber et al.,2014; Proudfoot, 2011; Mandel et al., 2008). APA iscommon and widespread in animals, plants, and othereukaryotic organisms (Tian et al., 2005; Pickrell et al.,2010; Shi, 2012; Sun et al., 2012; Wu et al., 2015). Inanimals, APA is involved in a range of developmentalprocesses including cell differentiation and has beenimplicated in cancer (Danckwardt et al., 2008;Mayr andBartel, 2009; Lin et al., 2012; Lianoglou et al., 2013).Similarly, in plants, APA has been shown to control keydevelopmental processes.

In one case, the nuclear RNA-binding protein FCAinteracts with FY, a component of CPSF complex, topromote the usage of a proximal polyadenylation site inits own gene, leading to the production of a nonfunc-tional RNA isoform (Simpson et al., 2003). In addition,FCA functions with CstF complex to promote proximalpolyadenylation of the noncoding antisense transcriptof FLC (Flowering Locus C), leading to suppression ofFLC expression (Liu et al., 2010). APA has also beenimplicated in the control of pathogen resistance inArabidopsis through the selection of proximal polyad-enylation sites in the RPP7 (Recognition of PeronosporaParasitica 7) gene (Tsuchiya and Eulgem, 2013).

Given the key function of DOG1 in Arabidopsis seedsurvival and the potential role of DOG1 homologs incontrolling seed dormancy in other species, it is impor-tant to understand both the mechanisms ofDOG1 locusregulation and the function of the encoded protein. Herewe describe the process of DOG1 APA that producestwo alternatively polyadenylated isoforms of theDOG1transcript. We characterize mutants in proteins thatcontrol this mechanism and show that the proximallypolyadenylatedmRNA isoform is translated in vivo andis functional as it complements the dog1 mutant, dem-onstrating that APA ofDOG1 gene in Arabidopsis playsa fundamental role in regulation of seed dormancy.

RESULTS

Mutants of RNA 39 Processing Factors Display Weak SeedDormancy Phenotype

The majority of Arabidopsis genes have multiplepolyadenylation sites (Wu et al., 2011), but the biological

consequences of the APA are in most cases unknown.FY is the Arabidopsis homolog of the yeast RNA 39processing factor Pfs2p (Simpson et al., 2003). A previousstudy demonstrated that the fy-1 mutant in Landsbergerecta background has weak seed dormancy (Jiang et al.,2012). To check if the function of FY in seed dormancycontrol is independent of genetic background, we

Figure 1. Arabidopsis mutants in RNA 39 processing complexes displayweak seed dormancy, unchanged DOG1 expression, and a low DOG1protein level. A, Germination efficiency (%) of freshly harvested seedsof fy-2 mutant. The graph shows the mean values from at least fourbiological replicates, with error bars representing SD. Each experimentwas repeated at least three times. dog1-3 mutant shown as positivecontrol. B, Germination efficiency (%) of freshly harvested seeds of RNA39 processing mutants. The graph shows the mean values from at leastfour biological replicates, with error bars representing SD. Each ex-periment was repeated at least three times. C, Relative expression levelofDOG1 in freshly harvested seeds from Col-0, dog1-3, and fy-2 plantsdetermined by RT-qPCR. Values were normalized to the level of UBCmRNA and are the means of at least three biological replicates, witherror bars representing SD.D, The fy-2mutant exhibits a decreased levelof DOG1 protein in seeds. The DOG1 protein level was determined bywestern blotting in Col-0, fy-2, dog1-3, and dog1-4mutant seed extractsusing a polyclonal DOG1 antibody. The Coomassie blue-stained gel isshown as a loading control. Four biological replicateswere analyzed forCol-0 and fy-2 plants. E, 39 RACE initiated withinDOG1 exon 3 revealsno clear defects in transcription termination. RACEwas performed usingRNA isolated from freshly harvested seeds of Col-0 and fy-2 lines. Thealignment shows the 39 end sequence of individual clones with thenumber of adenines (A) attached for each clone.

948 Plant Physiol. Vol. 170, 2016

Cyrek et al.

www.plantphysiol.orgon June 16, 2018 - Published by Downloaded from Copyright © 2016 American Society of Plant Biologists. All rights reserved.


assayed seed dormancy of fy-2 mutant in ecotypeColumbia of Arabidopsis (Col-0) background. Comparedwith the Col-0 wild type, fy-2 seeds showed very weakdormancy (Fig. 1A), demonstrating that FY plays a sig-nificant role in controlling this process in Arabidopsis.Since FY functions with other RNA 39 processing fac-tors in flowering time control (Manzano et al., 2009; Liuet al., 2010), we analyzed seed dormancy of themutantsof RNA 39 processing factors PCFS4 (PCF11P-SimilarProtein 4) and ESP1 (Enhanced Gene Silence 1). Likefy mutants, the pcsf4-1 and esp1-2 mutants showedweakened seed dormancy (Fig. 1B), suggesting thatthe observed seed dormancy defect is caused bymisregulation of RNA 39 processing of a gene or genesinvolved in seed dormancy control.Next, we examined whether the weak seed dor-

mancy phenotype of fy-2 may be due to misregulationof DOG1 transcription. Reverse transcription andquantitative PCR (RT-qPCR) analysis revealed no sig-nificant change of DOG1 mRNA level in fy-2 mutant(Fig. 1C). However, western-blot analysis using aDOG1-specific antibody revealed a consistent reduc-tion in the DOG1 protein level in fy-2 seeds whencompared with Col-0 wild type (Fig. 1D; SupplementalFig. S1). Thus, we concluded that FY is required forproper DOG1 protein expression and hypothesizedthat the fy-2 mutant may be defective in DOG1 RNAprocessing, which leads to the suppression of transla-tion. Therefore, we searched for potential DOG1 poly-adenylation defects in fy-2. Surprisingly, 39 RACEexperiments initiated from exon 3 of DOG1 showed noobvious differences in either polyadenylation site se-lection or the length of the poly(A) tail, between Col-0wild-type and fy-2 seeds (Fig. 1E).

APA of DOG1 Gene Results in Production of TwomRNA Isoforms

As we were unable to detect any defect in the use ofthe canonical polyadenylation site of DOG1 in the fymutant, we searched for additional polyadenylationsites within this gene. The most common polyadenyl-ation signal motif in plants is AAUAAA and UUGUUUpositioned 19 and 7 nt, respectively, upstream of thecleavage site (Sherstnev et al., 2012). The identificationof likely polyadenylation sites with PASPA (Ji et al.,2015), a software designed for mRNA polyadenylationsite prediction in plants, revealed three potential poly-adenylation clusters in the DOG1 gene. One clustercorresponds to the predicted full-length polyadenyla-tion site of DOG1, and the two others to additionalinternal polyadenylation sites: one in intron 1 and an-other in intron 2 (Fig. 2A). To validate these predictions,we reanalyzed published Direct RNA Sequencing-based mapping data to detect Arabidopsis polyade-nylation sites in the DOG1 locus (Supplemental Fig. S2;Sherstnev et al., 2012). We identified two distinct pol-yadenylation clusters: a distal one used to produce thepredicted full-length mRNA, which is named hereafter

as the long DOG1 (lgDOG1) form, and a proximal onematching one of the predicted internal polyadenylationsites located in intron 2 that would produce a truncatedmRNA, which we name short DOG1 (shDOG1). The

Figure 2. The DOG1 transcript is alternatively polyadenylated. A, Poly-adenylation site prediction along theDOG1 gene performed with PASPAsoftware. Diagram of DOG1 gene structure with alternatively splicedisoforms of intron 2 labeled as alpha, beta, gamma, and delta, plus theposition of the APA site. Lines above represent the fragments amplified inRT-qPCR analysis using the numbered primer sets. Below shown in scalewith the DOG1 gene structure is the diagram of the quality score repre-senting the likelihood of a polyadenylation site predicted by PASPAsoftware. B, Schematic presentation of two alternatively polyadenylatedDOG1 mRNA isoforms. C, RT-qPCR along the DOG1 gene to examinethe expression of the different mRNA isoforms in Col-0, fy-2, and dog1-3seeds. Values were normalized to the level ofUBCmRNA. Expression inmutants is normalized to the Col-0 value. Data represent the means ofthree biological replicates; error bars show SD. D, Long DOG1/shortDOG1 ratio in Col-0 andmutants of RNA 39 processing factors. RT-qPCRresults were normalized to the level ofUBCmRNA, and the expression inmutants was then normalized to the Col-0 value. Data represent themeans of three biological replicates with error bars showing SD. E, Al-ternativeDOG1 polyadenylation site selection during seed development.RT-qPCR specific for shDOG1 and lgDOG1was used to measure relativelevels of each mRNA isoform. Manually pollinated flowers were col-lected at different days after pollination (dap). Values were normalized tothe level of UBC mRNA. Data represent the means of three biologicalreplicates with error bars showing SD.

Plant Physiol. Vol. 170, 2016 949

Alternative Polyadenylation Controls Seed Dormancy


http://www.plantphysiol.org/cgi/content/full/pp.15.01483/DC1




secondpredicted internal polyadenylation site lieswithinintron 1, but analysis of RNA sequencing data revealedno indication of its usage (Supplemental Fig. S2).

In summary, the lgDOG1 transcript is comprisedof three exons and corresponds to the previously de-scribed Arabidopsis DOG1 mRNA (Bentsink et al.,2006), whereas the newly identified shDOG1 transcriptcomprises only exons 1 and 2 and therefore lacks thealternative splicing isoforms described (Fig. 2B).

Next, to quantify the differentDOG1mRNA isoformsin Col-0 wild-type and fy-2 mutant seeds, we designedsets of primers within the DOG1 locus for use in RT-qPCR analysis. Amplification with primer set 4 showedno difference in the level of lgDOG1 mRNA betweenfy-2 and wild-type seeds (Fig. 2, A and C). In contrast,use of primer set 3 demonstrated a clear reduction of theDOG1 transcript in fy-2 seeds. The primers of set 3 spanthe exon 2-intron 2 border that is specific for the shortalternatively polyadenylated shDOG1 mRNA isoform.Therefore, this result suggested a reduction in usage ofthe proximal polyadenylation site in the fy-2 mutant. Toconfirm this, we designed primers to amplify the full-length short and long DOG1 transcript isoforms andobtained similar results (Supplemental Fig. S3).

Different mRNA isoforms generated through APAoften differ in their stability (Krol et al., 2015). However,a cordycepin-dependent RNA stability assay (Goliszet al., 2013) showed that both the short and long DOG1mRNA isoforms have a similar half-life of about 1 h(Supplemental Fig. S4). This shows that usage ofDOG1proximal polyadenylation site leads to the productionof a stable transcript.

A change in the long/short DOG1 transcript ratio,similar to that observed in fy-2, was also detected inmutants of other factors required for RNA 39 process-ing, which showed weak seed dormancy (Figs. 1B and2D). This finding led us to speculate that the reductionin the shDOG1mRNA level in thesemutantsmay be theunderlying cause of their reduced seed dormancy.

Having established that theDOG1 gene is subjected toAPA, we questioned whether this process is develop-mentally regulated. DOG1 expression is tightly con-trolled during seed development and has previouslybeen shown to peak between 9 and 16 d after pollina-tion and then slowly decay (Nakabayashi et al., 2012;Zhao et al., 2015). Alternative splicing of the lgDOG1mRNA isoform has been reported to be unaffectedduring seed development (Bentsink et al., 2006). Ouranalysis of short and long DOG1mRNA levels showedthat both forms are induced during seed developmentand show a similar expression profile (Fig. 2E).

Short DOG1 mRNA Is Translated in Vivo and Produce aConserved Protein

The previously recognized long DOG1 mRNA tran-script (lgDOG1) encodes a protein of approximately 32kD. The proximally polyadenylated shDOG1 transcriptdescribed in this study codes for a slightly smaller

approximately 30-kD protein. These short and longDOG1 proteins only differ by several amino acids attheir C-terminal ends, encoded by exon 3 in lgDOG1and a sequence from intron 2 in shDOG1 (Fig. 3A). In-terestingly, a multiple sequence alignment with DOG1proteins from other species showed that the sequenceunique to shDOG1 is conserved at a similar levelcompared with that encoded by exon 2, while theamino acids unique to lgDOG1 (exon 3) are either ab-sent due to stop codons or areweakly conserved (Fig. 3A).Evolutionary conservation is often considered an indi-cation of functionality, suggesting that the shDOG1protein is functional.

Next, we determined which of the DOG1 mRNAisoforms is translated. DOG1 antibodies used previ-ously and that described here do not distinguish thelong and short DOG1 proteins, because they were raisedto peptides shared by the two isoforms (Nakabayashiet al., 2012). In addition, the small difference in proteinmass does not allow distinction between the shDOG1and lgDOG1 proteins on awestern blot: only oneDOG1band is observed in samples from freshly harvestedseeds (Fig. 1D). Therefore, we created a GFP::DOG1genomic fusion driven by theDOG1 promoter. Samplesfrom transgenic plants expressing this fusion proteinwere used in a green fluorescent protein (GFP) pull-down assay (Fig. 3B) with subsequent analysis by massspectrometry. Notably, no longDOG1-specific peptideswere identified, but we could clearly detect peptidesunique to the shDOG1 protein (Fig. 3C; SupplementalFig. S5). Although this does not exclude the possibilitythat the long DOG1 form is expressed but not detected,this result clearly showed that use of the proximalpolyadenylation site leads to the production of a shortDOG1 mRNA that is translated in vivo.

A previous study examining Arabidopsis lines consti-tutively overexpressing DOG1 suggested nuclear locali-zation of this protein (Nakabayashi et al., 2012). However,this experiment did not distinguish between the short andlong DOG1 forms. Therefore, we compared the localiza-tion of these DOG1 proteins. Both N- and C-terminal GFPfusions of the shDOG1 and lgDOG1 protein showedpredominantly nuclear localization when transientlyexpressed in Arabidopsis protoplasts (Fig. 3D).

We conclude that the short alternativelypolyadenylatedDOG1 mRNA isoform is translated in vivo, thus pro-ducing a short nuclear localized DOG1 protein. The factthat the shDOG1 is evolutionary conservedwhereas thelgDOG1 is less conserved, together with our inability tofind long DOG1 specific peptides, suggested that theshDOG1may be the functional DOG1 protein in respectto seed dormancy.

Short DOG1 Is Sufficient to Inhibit Germination

To directly test the ability of the short and long DOG1proteins to control seed dormancy in Arabidopsis, wecreated dog1-3 mutant plants with constructs carryingthe shDOG1 and lgDOG1 sequences driven by theDOG1


Cyrek et al.








promoter (Fig. 4A). Only the shDOG1 construct was ableto partially complement the weak seed dormancy phe-notype of dog1-3 (Fig. 4A), even though both showed asimilar DOG1 expression as measured by RT-qPCR(Supplemental Fig. S6). This indicated that shDOG1protein is sufficient for seed dormancy establishment.To further support this conclusion, we selected a dog1-5T-DNA insertion mutant allele. In this mutant, a T-DNAis inserted in exon 3 of theDOG1 gene (Fig. 4B). RT-qPCRanalysis with primers located at the exon 2-exon 3 junc-tion, upstream of the T-DNA insertion, confirmed thatlgDOG1 is no longer expressed in thismutant (Fig. 4C). Incontrast, levels of the shDOG1 transcript appear slightlyincreased in dog1-5. Accordingly, the dog1-5 mutantexhibited an increase in the DOG1 protein level and in-creased rather than reduced seed dormancy (Fig. 4, Dand E). These findings confirmed that the short version ofthe DOG1 transcript generated through APA is trans-lated and functional in controlling seed dormancy.

DISCUSSION

APA is widespread in eukaryotes. Whole-genomepolyadenylation site mapping has revealed that the

vast majority of Arabidopsis genes harbor APA sites,but the biological significance of this phenomenon ismostly unknown. In this study, we have shown that theArabidopsis DOG1 gene is subjected to APA. Theproximally polyadenylated short DOG1 transcript istranslated and functional in contrast to the long DOG1.Our data clearly demonstrate that expression of theshort DOG1 isoform is sufficient to promote seed dor-mancy, thus providing an example where APA is im-portant in seed development.

DOG1 mRNA Is Alternatively Polyadenylated

DOG1 is a major quantitative trait locus for seeddormancy in Arabidopsis and an important regulatorof this process in agriculturally important plants. Here,we show that the Arabidopsis DOG1 gene producestwo alternatively polyadenylated RNAs. The produc-tion of the short isoform is sensitive to mutations inRNA 39 processing complex components. RNA 39processing complexmutants with defects in selection ofthe proximal polyadenylation site onDOG1 showweakseed dormancy, suggesting that the proximal poly-adenylation is required for DOG1 function. This is

Figure 3. Comparison of the twoDOG1protein isoforms. A, Alignment ofDOG1 genomic sequences from selected plant speciesused to construct the protein sequence alignment beneath. Nucleotides and amino acids shared with Arabidopsis are highlightedby gray and red shading, respectively. Black box highlights peptide unique to shDOG1 identified below. B, Purification of DOG1from green siliques extract of plants expressing aGFP-DOG1 fusion under the control of the native promoter andCol-0 (Wt) plantsusing GFP Nano-trap beads. Isolated proteins were analyzed with western blot using GFP specific antibodies (left) and on a SDS-PAGE gel followed by silver staining (right). GFP-DOG1 band ismarkedwith an arrow. C, Purified proteinswere subjected tomassspectrometry analysis. Short DOG1 specific peptide fragmentation with mass spectrometry spectra of fragments derived,shown as intensity (y axis) sorted by mass-to-charge ratio (m/z) shown on x axis. D, Short and long DOG1 show nuclearlocalization. Arabidopsis protoplasts were transformed with constructs encoding N- and C-terminal fusions of GFP with longand short DOG1, then analyzed using a confocal microscope. The green color represents GFPand red the chlorophyll signal.






corroborated by the finding that the RNA 39 processingmutant fy-2 displays low DOG1 protein levels, but nochange or increase in the level of the full-length DOG1transcript. Taken together, these data indicate that theshort, alternatively polyadenylated DOG1 isoform isfunctional in seed dormancy regulation. The most well-characterized developmental process that involves acritical change in the polyadenylation site is the globalswitch to the usage of proximal polyadenylation sitesobserved during mammalian brain development orin tumors, which has been proposed to cause a releasefrom miRNA-mediated control due to the loss of

miRNA binding sites (Di Giammartino et al., 2011;Elkon et al., 2013). This is unlikely to be the underlyingreason for the switch in polyadenylation site of DOG1,since there is no dramatic difference in RNA stabilitybetween the long and short DOG1 isoforms.

Short DOG1 Protein Is Functional in SeedDormancy Establishment

A reanalysis of published data together with ourresults show that the DOG1 locus produces two alter-natively polyadenylated forms of DOG1 mRNA.Interestingly, the DOG1 genes in Arabidopsis lyrataand L. sativum, two species quite closely related toArabidopsis, harbor mutations, including stop codonsin exon 3, that encode the specific C-terminal region ofthe long DOG1 isoform. This shows that these speciesexpress only the short DOG1 protein isoform encodedby exons 1 and 2 plus part of intron 2.

Our complementation analysis demonstrated thatwhile the long DOG1 is unable to complement thedog1-3 mutation, the short version of DOG1 is func-tional. In addition, we identified a novel mutant dog1-5carrying a T-DNA insertion in exon 3, a region that ispresent only in lgDOG1. This mutant lacks the longversion of the DOG1 transcript, but it expresses theproximally polyadenylated shDOG1 at a higher levelthan wild type and exhibits stronger seed dormancy.This confirms that the shDOG1 isoform produced byAPA in intron 2 is sufficient for seed dormancy.

The fact that the fy-2 mutant shows a lower level ofthe DOG1 protein without any reduction in the lgDOG1mRNA strongly suggests that this transcript does notproduce a protein in this mutant. Examination of cur-rently available proteomic datasets revealed DOG1peptides. However, all of them correspond to DOG1exons 1 and 2, and none are specific to either the short orlong DOG1 protein isoforms (Baerenfaller et al., 2008;Castellana et al., 2008). Nevertheless, using targetedmass spectrometry analysis of immunoprecipitatedDOG1-GFP, shDOG1-specific peptides were clearlydetectable, but we did not identify peptides specificfor lgDOG1.

The short DOG1 protein can be derived eitherthrough the use of a proximal polyadenylation site orby splicing of the longDOG1 transcript at an alternativesplice site (DOG1-beta mRNA isoform; Bentsink et al.,2006). Analysis of splice sites in the DOG1 gene revealsweak conservation of the exon 3 acceptor splice siteused by the DOG1-beta mRNA isoform. This, togetherwith the strong DOG1 protein signal in mutant dog1-5that lacks the lgDOG1 mRNA isoforms, suggests thatthe shDOG1 protein is predominantly produced fromthe proximally polyadenylated short DOG1 transcriptisoform.

The function of the longDOG1 transcript and lgDOG1protein is currently not clear. Lack of conservation of thelgDOG1 protein specific sequence encoded by exon 3,together with the inability of lgDOG1 to complement

Figure 4. Short DOG1 is required for seed dormancy. A, Germinationrate (%) of freshly harvested seeds of the dog1-3 mutant and dog1-3complemented with long (# lg1, # lg2) and short (# sh3, # sh4) versionsof theDOG1 gene controlled by its native promoter. Data represent themeans of at least four biological replicates with error bars showing SD.B, Schematic description of DOG1 mutants: dog1-3 (SALK_000867),dog1-4 (SM_3_20886), and dog1-5 (SALK_022748). C, Relative ex-pression level of long and short DOG1 mRNAs in freshly harvestedseeds from Col-0 and dog1-5 plants determined by RT-qPCR. Valueswere normalized to the level ofUBCmRNA. Expression in themutant isnormalized to the Col-0 value. Data represent the means of three bio-logical replicates; error bars showSD.D,Germination rate (%) of freshlyharvested seeds of dog1mutants. The graph shows themeans from threebiological replicates, with error bars showing SD. The experiment wasrepeated at least three times. E, Western-blot analysis of the DOG1protein level in freshly harvested seeds from Col-0, dog1-5, dog1-4,and dog1-3 plants. The Coomassie blue-stained gel is shown as aloading control. Three biological replicates were examined for Col-0and dog1-5 plants.


Cyrek et al.



dog1 mutant, suggests that long DOG1 protein may benot functional. However, extensive examples of APAsites competition, including proximal and distal FCApolyadenylation sites (Simpson et al., 2003), imply thatthe DOG1 distal polyadenylation site might act as adecoy for transcripts that fail to terminate earlier. If so,then it could be expected that the lgDOG1 mRNA israpidly degraded. Due to technical limitations we wereunable to perform RNA stability assays on intact seeds.However, mRNA half-life measurements in seedlingsshowed no clear difference between the stability of thelong and short DOG1 transcripts.It is interesting to note that the dog1-5 mutant shows

up-regulation of both the proximally polyadenylatedshDOG1 mRNA isoform and the DOG1 protein,suggesting that the lgDOG1 mRNA may act as a neg-ative regulator of shDOG1 mRNA expression.In summary, our findings provide several lines of

evidence demonstrating that DOG1 nascent transcriptis subjected to APA and that the short form of thistranscript is functional in seed dormancy establishment.Given that DOG1 is subject to elaborate transcriptionalelongation control, it would be interesting to test whetherpolyadenylation site selection is controlled by the rateof PolII elongation onDOG1, as is the case for splice siteselection (Dolata et al., 2015).

MATERIALS AND METHODS

Plant Materials and Growth Conditions

Arabidopsis (Arabidopsis thaliana) plants were grown on soil in a greenhousewith a 16-h-light/8-h-dark photoperiod at 22°C/18°C. For expression analysisduring seed development, flowers were emasculated and pollinated the fol-lowing day. Plant material was collected at the indicated times representingdays since manual pollination. The termination complex mutants used in thisstudy have been described previously: fy-2 (Simpson et al., 2003), esp1-2(SALK_078793), and pcsf4-1 (SALK_102934c; Xing et al., 2008). Col-0 was usedas wild-type control. The DOG1 T-DNA insertion mutants used were: dog1-3(SALK_000867; Bentsink et al., 2006), dog1-4 (SM_3_20886), and dog1-5(SALK_022748).

Germination Assay

About 100 to 150 freshly harvested Arabidopsis seeds were sown on bluegerminationpaper(Anchor) supportedbytwolayersof thickfabricsaturatedwithwater. Plateswere sealed and transferred to a growth chamberwith 16-h-light/8-h-dark at 22°C/18°C. Plateswere photographed daily for 7 d and germinatingseeds counted. At least three biological replicates were performed for eachexperiment. Data show a time point when the dog1-3 mutant seed had fullygerminated. Seed dormancy strength is strongly dependent on external con-ditions during seed maturation (Kendall et al., 2011); we therefore always in-cluded a set of controls: dog1-3 and wild type in all experiments. Small externalcondition fluctuations between separate experiments may therefore explain thedifference in seed germination strength betweenCol-0 plants in Figure 1,A and B.

RNA Extraction, cDNA Synthesis, and PCR Analysis

Total RNA was isolated using a phenol-chloroform extraction procedure(Shirzadegan et al., 1991). The RNAwas then treated using a TURBODNA-freekit (Life Technologies) according to the manufacturer’s protocol. The efficiencyof DNA removal was monitored by PP2A PCR. Before RT, the RNA was ex-amined by electrophoresis on a 1.2% agarose gel to determine its quality andquantified using a Nanodrop 2000 spectrophotometer. For cDNA synthesis,2.5 mg of RNA were used with oligo(dT) primers and a RevertAid First Strand

cDNA Synthesis kit (Fermentas). Diluted cDNA (10-fold) was used in qPCR(LightCycler 480; Roche) with SYBR Green mix (Roche). A UBC gene was usedfor normalization (Czechowski et al., 2005).

Protein Extraction

Freshly harvested Arabidopsis seeds (30 mg) were frozen in liquid nitrogenand pulverized in 100 mL of acetone. The suspension was centrifuged for 1 minat 1,000g. The pelleted material was rinsed with 96% ethanol and centrifugedagain as above. The pellet was then resuspended in 50% Percoll and centrifugedfor 10 min at 1,000g. Pelleted material was again washed in 96% ethanol andcentrifuged for 1 min at 1,000g. The final pellet was resuspended in 8 M urea foranalysis by SDS-PAGE.

Western Blotting

Proteins were extracted from freshly harvested seeds as described above.The protein concentration was determined using the Bradford method(Bradford, 1976), and samples were analyzed by standard SDS-PAGE. Proteinswere then electrophoretically transferred to Hybond P membrane (Amersham)using a Trans-Blot Cell (Bio-Rad). A synthetic peptide (MGSSSKNIEQAQDSY)corresponding to amino acids 1 to 15 of DOG1 was synthesized and usedto produce an affinity-purified DOG1-specific rabbit polyclonal antibody(Eurogentec), available fromAgrisera (AS15 3032). DOG1-specific (Eurogentec)and Goat Anti-Rabbit IgGH&L (HRP; Abcam b7090) antibodies were also usedfor western blotting. Western blot was developed using a 1:1 mixture ofSuperSignal West Pico and Femto Chemiluminescent Substrates (Pierce).

GFP-DOG1 Immunoprecipitation and MassSpectrometry Analysis

pDOG1::GFP::DOG1g vector was constructed using primers shown inSupplemental Table S1 and a pCambia5408 modified to carry a Basta resistanceunder NOS promoter resulting in nativeDOG1 gene with GFP inserted at ATG.Plant extracts from a mixture of green and mature siliques of transgenic plantswere used for immunoprecipitation with GFP Nano-trap beads (ChromoTek).Mass spectrometry followed by analysis using Mascot software was performedas described previously (Dolata et al., 2015).

Protoplast Transformation

The following plasmid constructswere prepared in the vectors pSAT6-eGFP-C1 and pSAT6A-eGFP-N1 using the short and longDOG1 cDNAs: 35S∷sDOG1-YFP, 35S∷YFP-sDOG1, 35S∷lgDOG1-YFP, and 35S∷YFP-DOG1. To maximizethe efficiency of protoplast isolation, protoplasts were isolated from the leavesof 20-d-old late flowering Col-0 with backcrossed active FRI (Frigida) from SF2ecotype (Michaels and Amasino, 1999) and transformed with 30 mg of purifiedconstruct DNAs according to a previously described method (Wu et al., 2009).Transformed protoplasts were incubated overnight under continuous light at22°C prior to analysis by confocal microscopy.

Confocal Microscopy Imaging

Fluorescence imagesof transformedArabidopsisprotoplastswerevisualizedusing aNikon EZ-C1 laser scanning microscope mounted on an inverted NikonTE 2000E epifluorescence microscope. Excitation was at 514 nm, and theemission signal was collected between 525 and 590 nm for YFP fluorescence andbetween 622 and 700 nm for chlorophyll autofluorescence. Untransformedprotoplasts were examined as a negative control.

39 RACE

RACE was performed as described previously (Swiezewski et al., 2009).Total RNA from seedlings was ligated to 39 RACE oligo adapter (Gene Racerkit; Invitrogen) using T4 RNA Ligase 1 (New England Biolabs). After phenol/chloroform extraction, RNA was precipitated and subjected to RT usingadapter-specific primer and Superscript III (Invitrogen). 39 ends of DOG1mRNAwere amplified using adapter specific primer and DOG1 specific RACEprimer. PCR products were purified from agarose gel and cloned to pJet1.2vector and sequenced using plasmid specific primers in both directions.






RNA Stability Assay

Cordycepin-dependent RNA stability assay was performed as previouslydescribed (Golisz et al., 2013). In brief, Col-0 plants were grown on half-strengthMurashige and Skoog medium containing 1% Suc in a 16-h-light/8-h-darkcycle at 22°C. About 100 2-week-old seedlings were transferred to 40 mL ofbuffer (1 mM PIPES, pH 6.25, 1 mM trisodium citrate, 1 mM KCl, 15 mM Suc) andincubated for 30 min with shaking. Cordycepin was then added to a finalconcentration of 150 mg/mL and vacuum infiltration performed for 30 s. Ateach time point thereafter, seedlings representing approximately 0.5 g werecollected and frozen in liquid nitrogen. Samples were analyzed in triplicate.RNA extractionwas performed as described above, and RT-qPCR analysis withEIF4A and At3g45970 was used as a control showing, respectively, high andlow mRNA stability level (Golisz et al., 2013).

dog1-3 Complementation

For complementation with lgDOG1, dog1-3 mutant was transformed with aDOG1 genomic fragment with the sequence of the second intron deleted. Forcomplementation with shDOG1, the dog1-3 mutant was crossed with Col-0plants expressing a construct containing the genomic fragment spanningDOG1exon 1-intron 1 and exon 2. TheDOG1 genes in both constructs are driven by theendogenous DOG1 promoter. Two independent transformants were analyzedfor each construct.

Accession Numbers

The sequences of genes examined in this study can be found in TheArabidopsis Information Resource data library under the following accessionnumbers: DOG1 (At5g45830), FY (At5g13480), ESP1 (At1g73840), and PCSF4(At4g04885).

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. Western blot and Coomassie-stained gel used forFigure 1D.

Supplemental Figure S2. DOG1 polyadenylation site identified usingDirect RNA sequencing data (Sherstnev et al., 2012).

Supplemental Figure S3. RT-PCR to examine the expression of full-lengthshDOG1 and lgDOG1 mRNA transcripts in Col-0 and fy-2 seeds.

Supplemental Figure S4. RNA stability assay of shDOG1 and lgDOG1transcripts.

Supplemental Figure S5. Characterization of short DOG1-specific peptides.

Supplemental Figure S6. DOG1 expression in freshly harvested seedsof Col-0 (wild type) and dog1-3 or dog1-3 mutants complementedwith lgDOG1 (# lg1, # lg2) or shDOG1 (# sh3, # sh4) expressingconstructs.

Supplemental Table S1. List of primers.

ACKNOWLEDGMENTS

We thank Maciej Kotli�nski for help with analysis of mass spectrometry dataand Krzysztof Kokoszka for excellent technical assistance. We are also verygrateful to Dorota Konopka for critical reading of the manuscript and all mem-bers of the Swiezewski laboratory for stimulating discussions.

Received September 18, 2015; accepted November 25, 2015; published Novem-ber 30, 2015.

LITERATURE CITED

Baerenfaller K, Grossmann J, Grobei MA, Hull R, Hirsch-Hoffmann M,Yalovsky S, Zimmermann P, Grossniklaus U, Gruissem W, Baginsky S(2008) Genome-scale proteomics reveals Arabidopsis thaliana gene modelsand proteome dynamics. Science 320: 938–941

Bentsink L, Jowett J, Hanhart CJ, Koornneef M (2006) Cloning of DOG1, aquantitative trait locus controlling seed dormancy in Arabidopsis. ProcNatl Acad Sci USA 103: 17042–17047

Bradford MM (1976) A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dyebinding. Anal Biochem 72: 248–254

Castellana NE, Payne SH, Shen Z, Stanke M, Bafna V, Briggs SP (2008)Discovery and revision of Arabidopsis genes by proteogenomics. ProcNatl Acad Sci USA 105: 21034–21038

Chiang GCK, Bartsch M, Barua D, Nakabayashi K, Debieu M, Kronholm I,Koornneef M, Soppe WJJ, Donohue K, De Meaux J (2011) DOG1 expres-sion is predicted by the seed-maturation environment and contributes togeographical variation in germination in Arabidopsis thaliana. Mol Ecol 20:3336–3349

Czechowski T, Stitt M, Altmann T, Udvardi MK, Scheible W-R (2005)Genome-wide identification and testing of superior reference genes fortranscript normalization in Arabidopsis. Plant Physiol 139: 5–17

Danckwardt S, Hentze MW, Kulozik AE (2008) 39 end mRNA processing:molecular mechanisms and implications for health and disease. EMBO J27: 482–498

Di Giammartino DC, Nishida K, Manley JL (2011) Mechanisms andconsequences of alternative polyadenylation. Mol Cell 43: 853–866

Dolata J, Guo Y, Kołowerzo A, Smoli�nski D, Brzy _zek G, Jarmołowski A,�Swie _zewski S (2015) NTR1 is required for transcription elongationcheckpoints at alternative exons in Arabidopsis. EMBO J 34: 544–558

Elkon R, Ugalde AP, Agami R (2013) Alternative cleavage and polyade-nylation: extent, regulation and function. Nat Rev Genet 14: 496–506

Finch-Savage WE, Leubner-Metzger G (2006) Seed dormancy and thecontrol of germination. New Phytol 171: 501–523

Golisz A, Sikorski PJ, Kruszka K, Kufel J (2013) Arabidopsis thaliana LSMproteins function in mRNA splicing and degradation. Nucleic Acids Res41: 6232–6249

Graeber K, Nakabayashi K, Miatton E, Leubner-Metzger G, Soppe WJJ(2012) Molecular mechanisms of seed dormancy. Plant Cell Environ 35:1769–1786

Graeber K, Linkies A, Steinbrecher T, Mummenhoff K, Tarkowská D,Ture�cková V, Ignatz M, Sperber K, Voegele A, de Jong H, et al (2014)DELAY OF GERMINATION 1 mediates a conserved coat-dormancymechanism for the temperature- and gibberellin-dependent control ofseed germination. Proc Natl Acad Sci USA 111: E3571–E3580

Grasser M, Kane CM, Merkle T, Melzer M, Emmersen J, Grasser KD(2009) Transcript elongation factor TFIIS is involved in arabidopsis seeddormancy. J Mol Biol 386: 598–611

Gruber AR, Martin G, Keller W, Zavolan M (2014) Means to an end:mechanisms of alternative polyadenylation of messenger RNA precur-sors. Wiley Interdiscip Rev RNA 5: 183–196

Ji G, Li L, Li QQ, Wu X, Fu J, Chen G, Wu X (2015) PASPA: a web serverfor mRNA poly(A) site predictions in plants and algae. Bioinformatics31: 1671–1673

Jiang S, Kumar S, Eu Y-J, Jami SK, Stasolla C, Hill RD (2012) TheArabidopsis mutant, fy-1, has an ABA-insensitive germination phenotype.J Exp Bot 63: 2693–2703

Kendall SL, Hellwege A, Marriot P, Whalley C, Graham IA, Penfield S(2011) Induction of dormancy in Arabidopsis summer annuals requiresparallel regulation of DOG1 and hormone metabolism by low temper-ature and CBF transcription factors. Plant Cell 23: 2568–2580

Koornneef M, Bentsink L, Hilhorst H (2002) Seed dormancy and germination.Curr Opin Plant Biol 5: 33–36

Krol J, Krol I, Alvarez CPP, Fiscella M, Hierlemann A, Roska B, FilipowiczW(2015) A network comprising short and long noncoding RNAs and RNAhelicase controls mouse retina architecture. Nat Commun 6: 7305

Lianoglou S, Garg V, Yang JL, Leslie CS, Mayr C (2013) Ubiquitouslytranscribed genes use alternative polyadenylation to achieve tissue-specific expression. Genes Dev 27: 2380–2396

Lin Y, Li Z, Ozsolak F, Kim SW, Arango-Argoty G, Liu TT, Tenenbaum SA,Bailey T, Monaghan AP, Milos PM, et al (2012) An in-depth map ofpolyadenylation sites in cancer. Nucleic Acids Res 40: 8460–8471

Liu F, Marquardt S, Lister C, Swiezewski S, Dean C (2010) Targeted 39processing of antisense transcripts triggers Arabidopsis FLC chromatinsilencing. Science 327: 94–97

Liu Y, Koornneef M, Soppe WJJ (2007) The absence of histone H2Bmonoubiquitination in the Arabidopsis hub1 (rdo4) mutant reveals arole for chromatin remodeling in seed dormancy. Plant Cell 19: 433–444


Cyrek et al.










Mandel CR, Bai Y, Tong L (2008) Protein factors in pre-mRNA 39-endprocessing. Cell Mol Life Sci 65: 1099–1122

Manzano D, Marquardt S, Jones AME, Bäurle I, Liu F, Dean C (2009) Alteredinteractions within FY/AtCPSF complexes required for Arabidopsis FCA-mediated chromatin silencing. Proc Natl Acad Sci USA 106: 8772–8777

Mayr C, Bartel DP (2009) Widespread shortening of 3’UTRs by alternativecleavage and polyadenylation activates oncogenes in cancer cells. Cell138: 673–684

Michaels SD, Amasino RM (1999) FLOWERING LOCUS C encodes anovel MADS domain protein that acts as a repressor of flowering. PlantCell 11: 949–956

Nakabayashi K, Bartsch M, Xiang Y, Miatton E, Pellengahr S, Yano R,Seo M, Soppe WJJ (2012) The time required for dormancy release inArabidopsis is determined by DELAY OF GERMINATION1 proteinlevels in freshly harvested seeds. Plant Cell 24: 2826–2838

Nonogaki H (2014) Seed dormancy and germination-emerging mechanismsand new hypotheses. Front Plant Sci 5: 233

Penfield S, King J (2009) Towards a systems biology approach to under-standing seed dormancy and germination. Proc Biol Sci 276: 3561–3569

Pickrell JK, Marioni JC, Pai AA, Degner JF, Engelhardt BE, Nkadori E,Veyrieras J-B, Stephens M, Gilad Y, Pritchard JK (2010) Understandingmechanisms underlying human gene expression variation with RNAsequencing. Nature 464: 768–772

Proudfoot NJ (2011) Ending the message: poly(A) signals then and now.Genes Dev 25: 1770–1782

Sherstnev A, Duc C, Cole C, Zacharaki V, Hornyik C, Ozsolak F, MilosPM, Barton GJ, Simpson GG (2012) Direct sequencing of Arabidopsisthaliana RNA reveals patterns of cleavage and polyadenylation. NatStruct Mol Biol 19: 845–852

Shi Y (2012) Alternative polyadenylation: new insights from global analyses.RNA 18: 2105–2117

Shirzadegan M, Christie P, Seemann JR (1991) An efficient method forisolation of RNA from tissue cultured plant cells. Nucleic Acids Res19: 6055

Simpson GG, Dijkwel PP, Quesada V, Henderson I, Dean C (2003) FY isan RNA 39 end-processing factor that interacts with FCA to control theArabidopsis floral transition. Cell 113: 777–787

Sun Y, Fu Y, Li Y, Xu A (2012) Genome-wide alternative polyadenylation inanimals: insights from high-throughput technologies. J Mol Cell Biol 4:352–361

Swiezewski S, Liu F, Magusin A, Dean C (2009) Cold-induced silencing bylong antisense transcripts of an Arabidopsis Polycomb target. Nature462: 799–802

Tian B, Hu J, Zhang H, Lutz CS (2005) A large-scale analysis of mRNApolyadenylation of human and mouse genes. Nucleic Acids Res 33:201–212

Tsuchiya T, Eulgem T (2013) An alternative polyadenylation mechanismcoopted to the Arabidopsis RPP7 gene through intronic retrotransposondomestication. Proc Natl Acad Sci USA 110: E3535–E3543

Wu F-H, Shen S-C, Lee L-Y, Lee S-H, Chan M-T, Lin C-S (2009) Tape-Arabidopsis Sandwich - a simpler Arabidopsis protoplast isolationmethod. Plant Methods 5: 16

Wu X, Liu M, Downie B, Liang C, Ji G, Li QQ, Hunt AG (2011) Genome-wide landscape of polyadenylation in Arabidopsis provides evidencefor extensive alternative polyadenylation. Proc Natl Acad Sci USA 108:12533–12538

Wu X, Zeng Y, Guan J, Ji G, Huang R, Li QQ (2015) Genome-wide char-acterization of intergenic polyadenylation sites redefines gene spaces inArabidopsis thaliana. BMC Genomics 16: 511

Xing D, Zhao H, Xu R, Li QQ (2008) Arabidopsis PCFS4, a homologue ofyeast polyadenylation factor Pcf11p, regulates FCA alternative pro-cessing and promotes flowering time. Plant J 54: 899–910

Zhao M, Yang S, Liu X, Wu K (2015) Arabidopsis histone demethylases LDL1and LDL2 control primary seed dormancy by regulating DELAY OFGERMINATION 1 and ABA signaling-related genes. Front Plant Sci 6: 159

Zheng J, Chen F, Wang Z, Cao H, Li X, Deng X, Soppe WJJ, Li Y, Liu Y (2012)A novel role for histone methyltransferase KYP/SUVH4 in the control ofArabidopsis primary seed dormancy. New Phytol 193: 605–616





Seed Dormancy in Arabidopsis Is Controlled by … Dormancy in Arabidopsis Is Controlled by...

Documents

Transcript of Seed Dormancy in Arabidopsis Is Controlled by … Dormancy in Arabidopsis Is Controlled by...