ARTICLE IN PRESS

Journal of Plant Physiology ] (]]]]) ]]]—]]]

KEYWORDArabidopscDNA-AFLPGerminatiGibberelliTransposon

0176-1617/$ - sdoi:10.1016/j.

Abbreviationmorphism; cDN�CorrespondE-mail addr

www.elsevier.de/jplph

cDNA-AFLP analysis of seed germination inArabidopsis thaliana identifies transposons andnew genomic sequences

Juana G. de Diegoa, F. David Rodrıgueza, Jose Luis Rodrıguez Lorenzo,Philippe Grappinb, Emilio Cervantesc,�

aDepartamento de Bioquımica y Biologıa Molecular, Edificio Departamental, Campus Miguel de Unamuno, Universidadde Salamanca, 37007, Salamanca, SpainbLaboratoire de Biologie de Semences, INRA, 78026, Versailles Cedex, FrancecIRNA-CSIC, Apartado 257, 37080, Salamanca, Spain

Received 30 November 2004; accepted 26 April 2005

Sis;;on;c acid;

ee front matter & 200jplph.2005.04.032

s: AFLP, amplificatioA, complementary DNing author. Tel.: 923 2ess: ecervant@usal.es

SummaryA cDNA-AFLP experiment was designed to identify and clone nucleotide sequencesinduced during seed germination in Arabidopsis thaliana. Sequences corresponding toknown genes involved in processes important for germination, such as mitochondrialbiogenesis, protein synthesis and cell cycle progression, were isolated. Othersequences correspond to Arabidopsis BAC clones in regions where genes have notbeen annotated. Notably, a number of the sequences cloned did not correspond toavailable sequences in the databases from the Arabidopsis genome, but insteadpresent significant similarity with DNA from other organisms, for example fish species;among them, some may encode transposons. A number of the sequences isolatedshowed no significant similarity with any sequences in the public databases.Oligonucleotides derived from these new sequences were used to amplify genomicDNA of Arabidopsis. Expression analysis of representative sequences is presented. Thiswork suggests that, during germination, there may be a massive transposonmobilization that may be useful in the annotation of new genome sequences andidentification of regulatory mechanisms.& 2005 Elsevier GmbH. All rights reserved.

5 Elsevier GmbH. All rights rese

n fragment length poly-A; GA, gibberellic acid19606; fax: 923 219609.(E. Cervantes).

Introduction

Germination is the process by which a seedinitiates growth after a period of quiescence. Itrequires seed imbibition, and in a strict sense, is

rved.

ARTICLE IN PRESS

J.G. de Diego et al.2

defined as the process leading to emergence of theradicle through the testa, a tissue of maternalorigin that surrounds the embryo (Bewley, 1997;Koornneef et al., 2002). Germination is thusfinished once the radicle has emerged.

Imbibition, i.e. water uptake by the seeds, isaccompanied by cell expansion, cell wall synthesisand activation of metabolism. Accumulating evidenceindicates that, in general, cell division occursfollowing germination (de Castro et al., 2000; Barrocoet al., 2005). The increase in cell growth that isrequired for germination is due to cell elongation.

In a very short time interval, a limited number ofcells elongate and go through differentiationprocesses based in rapid metabolic changes pre-ceding cell division. The process of germination isunder the control of environmental and hormonalfactors thus making the system appropriate for thestudy of plant development and the cellularresponses to these factors.

In Arabidopsis, as in many other plant species,germination is under the control of phytochrome. Theseeds require the presence of light to germinate, andafter activation by red light, phytochrome regulatesthe production of Gibberellic acid (GA) biosyntheticenzymes as well as of other genes, including transcrip-tion factors (Yamaguchi et al., 1998; Kamiya andGarcıa-Martınez, 1999; Tepperman et al., 2001).Nevertheless, the classical view that explains seedgermination as a consecutive series of causes andeffects ending in the activation of genes that encodeproteins necessary for radicle elongation may, in thelight of recent results, be modified for a moreintegrative view that takes into consideration globalgenome dynamics (Lippman et al., 2004; Nakabayashiet al., 2005)

GA is needed for Arabidopsis seed germination: ga1mutants, unable to synthesize GA, do not completegermination unless GA is added exogenously or thetesta broken or removed artificially (Silverstone etal., 1997). GA is required to overcome ABA induceddormancy and seed coat resistance to germination(Debeaujon and Koornneef, 2000). Processes underthe control of GA during germination of Arabidopsisthaliana seeds have been recently analyzed bymicroarray techniques. Upregulated genes includedthose whose products are involved in cell elongationand division, transcriptional regulation and enzymesresponsible for hormone biosynthesis (Ogawa et al.,2003). The variations in gene expression in themicroarray study are limited to the genes includedin the array. An alternative approach, such ascomplementary DNA (cDNA)-amplification fragmentlength polymorphism (AFLP), may be helpful touncover new transcripts whose products are involvedin germination.

Proteins involved in the repression of germina-tion are negatively regulated by GA and includemembers of the GRAS multiprotein family oftranscriptional regulators involved in cell differ-entiation. In Arabidopsis, the DELLA subfamily ofGRAS regulatory genes consists of GAI (GibberellicAcid Insensitive), RGA (Regulator of gibberellinresponse), RGA-LIKE1 (RGL1), RGL2, and RGL3. Allcontain DELLA protein motifs. Among them, RGL1and RGL2 have been demonstrated to be involved ingermination (Peng and Harberd, 2002). Mutants inRGL1 present reduced sensitivity to paclobutrazol,an inhibitor of GA biosynthesis (Wen and Chang,2002) and mutations in RGL2, restore germinationto germination-deficient mutants or wild-typeseeds treated with inhibitors (Lee et al., 2002).Thus, an important effect of GA in the induction ofgermination may be the reduction of the levels ofthese inhibitors. GA action is regulated by theubiquitin–proteasome pathway. The genes Sleepy1(SLY1) and Sneezy (SNE) encode F-box proteins of aSkp1-cullin-F-box (SCF), E3 ubiquitin ligase com-plex that positively regulates GA signaling troughtargeted proteolysis of the DELLA proteins usingubiquitin mediated destruction (McGinnis et al.,2003; Strader et al., 2004).

The mechanism by which DELLA proteins exerttheir inhibitory effect is still unclear. For RGA, itsnuclear localization is reduced by GA in coordina-tion with auxin and ethylene, leading to rootelongation (Achard et al., 2003; Fu and Harberd,2003) but the nuclear targets of RGA have not yetbeen identified.

It has been shown that many non-coding RNAsmay be expressed in a variety of tissues anddevelopmental conditions. Among them, somemay be precursors of microRNAs, whose roles indifferentiation processes have been recently de-monstrated both in plant and in animal systems(Reinhart et al., 2002; Houbaviy et al., 2003). Thiswork reports on the identification of sequencesregulated by GA early during imbibition. The cDNA-AFLP approach allows the identification of ex-pressed sequences independent of their codingcapacity. The analysis of their regulation in diversemutant backgrounds and during the course ofgermination may provide insights into this process.

Materials and methods

Plant material

A. thaliana cv. Wassilewskija, as well as itsderived mutant abc33 (ga1 allele described by

ARTICLE IN PRESS

Analysis of seed germination in Arabidopsis thaliana 3

Dubreucq et al., 1996) were obtained from the seedstocks of INRA Versailles. Seeds were sown in 1%agar plates with a light/dark cycle (16/8) at 25/20 1C.

Seeds of mutants ga1-1 in the ecotype Landsbergerecta were obtained from the Nottingham Arabi-dopsis stock centre. Lines eto1-1 (Chae et al.,2003) and etr1-1 (Chang et al., 1993), in theecotype Columbia, were a gift of Dr. Marta Roldan.

cDNA-AFLP

Seeds of wild-type Wassilewskija and mutantabc33 (ga1) were sown in water–1% agar and storedin the cold (4 1C) for stratification for 3 days.Afterwards, the seeds were allowed to germinatein a plant growth room under constant light at 20 1C.Seeds were collected after 24h and RNA wasextracted by the hot phenol method. Poly A+ RNAwas isolated and reverse transcribed. For theremainder of the protocol, the procedure of Bachemet al. (1996) was followed with slight modificationsfrom Bove et al. (2005). Sequences from cDNAs fromboth wild type and ga1 were amplified using theoligonucleotides AseI-NN and TaqI-NN

Primers used were:

Figure 1. Fragment of a gel showing profiles of differ- entially expressed transcripts as result of the cDNA-AFLP Adapter Taq I-50 50-GACGATGAGTCCTGAC-30

experiment.

Adapter Taq I-30 50-CGGTCAGGACTCAT-30

Adapter Ase I-50

50-CTCGTAGACTGCGTACC-30

Adapter Ase I-30

50-TAGGTACGCAGTC-30

A total of 64 combinations corresponding to AseI-CN/TaqI-NN were analyzed.

As a result of the AFLP experiment, a totalof approximately 5200 bands were detected inacrylamide gels (ca. 120–130 bands for eachprimer combination; Fig. 1). A total of 201 bandswere selected on the basis of being differentiallydetected at 24 h of imbibition, either in thewild type or in the mutant seeds. After excisionfrom the gels, the bands were cloned inpGEM-TR (PROMEGA), and the nucleotide sequencesof the inserts determined. All the sequencessmaller than 40 bp or showing vector contaminationwere discarded. BlastN program was run with theindividual sequences obtained against the A.thaliana gene banks (http://www.arabidopsis.org/Blast/).

RNA extraction and northern blot

For northern blots, seeds were sown in 1% agarplates with a light/dark cycle (16/8) at 25/20 1C.RNA was extracted by the method of Vicient and

Delseny (1999), run in agarose/formaldehydegels and transferred to nylon membranes(Hybond N, Amersham). Hybridization was per-formed with cDNA probes synthesized using radi-olabeled 32P.

Results

Identification of known genes that areinduced during imbibition.

Thirty-five non-redundant sequences were iso-lated as being differentially expressed duringimbibition of wild type and ga1 mutant seeds.Among them, seven corresponded to protein-encoding genes annotated in the Arabidopsisgenome (Table 1).

Clone 11 encodes a fragment of a type IIperoxiredoxin (At1g65980). Clone 20 contains anucleotide sequence partially encoding the b-ziptranscription factor ATHB2 (At2g18160; GBF5),that has been shown to be regulated by light (Rooket al., 1998). Clone 44 corresponds to an unknownprotein, probably of mitochondrial localization

ARTICLE IN PRESS

Table 1. Clones containing fragments of known genes that are induced during imbibition

Lab code TAIR gene code P-value Gene product Induced in

11 At1g65980.1 1e�059 Type II peroxirredoxin ga120 At2g18160.1 1e�049 b-Zip transcription factor ga144 At5g42825.1 0.0 Mitochondrial protein wt54b At5g50720.1 2e�013 ABA-responsive protein (HVA22) wt73 At2g21580.1 4e�045 Ribosomal protein S25B wt86 At3g27060.1 2e�040 Ribonucleotide reductase small subunit ga193 Chloroplast 1e�009 Ribosomal protein ga1109 At3g27280.1 7.2e�08 Prohibitin wt

J.G. de Diego et al.4

(At5g4282). Clone 54b encodes a fragment of HVA22(At5g50720; Chen et al., 2002), a stress and ABA-induced protein with diverse homologies in eukar-yotes included in the family TB2/DP1 proteinsdeleted in severe forms of familial adenomatouspolyposis, an autosomal dominant oncologicalinherited disease.

A sequence corresponding to At2g21580, encod-ing ribosomal protein S25, was obtained in clone 73as differentially expressed in the wild-type seedsduring imbibition. Partly encoded in clone 86 isAt3g27060, the gene for the small subunit of theenzyme ribonucleotide reductase (Chaboute et al.,2000), an enzyme essential in the S phase duringthe cell cycle.

Clone 93, is similar to a chloroplast-encodedsequence, most probably corresponding to 16Sribosomal RNA.

Finally, clone 109.1 encodes a fragment of prohi-bitin (At3g27280), a protein located in the mitochon-drial membrane in yeast and mammalian cells with areported role in the regulation of cell cycle. Aprohibitin gene was previously described in Arabi-dopsis (Snedden and Fromm, 1997). Four of thesesequences were cloned as up-regulated in the wildtype (Prohibitin, 40S Ribosomal protein, ATHVA22Eand unknown protein) and four in the mutant(Peroxiredoxin, rps12, RNR small subunit, ATHB2).

Identification of Arabidopsis sequences notencoding known genes that are inducedduring imbibition

Among the sequences cloned, five correspondedto A. thaliana annotation units reported in thedatabases, in regions where no genes have beendescribed (Table 2).

Most of them represent sequences unique in theArabidopsis genome. For example, clone 46 wassimilar to a sequence in BAC clone T8O18, but notto any other sequences in the Arabidopsis data-base. Clone 83 was similar to a sequence in T9I1,but no other database sequences. Clone 95 con-

tained an insert of 117 base pairs identical to asequence in clone F13B15. Clone 119 was almostidentical to a sequence in F2J7. To date, no knowngene or transposon has been described in theseregions.

The nucleotide sequence in clone 97 is similarto multiple Arabidopsis genomic regionsincluding 32 sequences in different BAC clones,but gave no significant similarity with codingregions. This sequence contained similarity withATHILA4A_LTR (see below). Interestingly, all ofthese sequences were cloned as induced in thega1 mutant.

Sequences induced during imbibition notshowing similarity with describedArabidopsis sequences

Eight sequences presented significant similaritywith fish DNA in diverse regions (promoters, exons,etc.) of various genes. The sequences includedmicrosatellites and transposons (Table 3). Of them,seven were differentially induced in the wild typeand one in the ga1 mutant seedlings. Finally, 14sequences (seven induced in the wild type andseven in the mutant) did not present any significantsimilarity to any sequences in the public databases(Table 4).

Some of the sequences contain repeatedmotifs

The sequences included in Tables 2–4 were usedas queries in searches in the REPBASE database(Jurka, 2000). Repeated motifs were found in someof the sequences indicating their relatedness withtransposons and repeated DNA. Sequence 97 (Table2) was found to be almost identical with ATHI-LA4A_LTR, a long terminal repeat from the ATHILA4endogenous retrovirus from Arabidopsis. Sequence54a (Table 3) was similar to a LINE retroelement,sequence 10 to a microsatellite (Table 3) and

ARTICLE IN PRESS

Table 3. Clones containing sequences induced during imbibition, not similar to any Arabidopsis sequences

Lab code (length)-GenBank_Accn

NCBI code P-value Organism and gene product Inducedin

10 (242)-DN289326 gi|46251374|gb|AY543804.1| 1e�32 Salmo salar clone Alu255 microsatellitesequence

ga1

54a(243)-DN289329 gi|34106193|gb|AY227655.1| 9e�80 Salvelinus fontinalis myostatin b/misoform gene

wt

56 (208)-DN289330 gi|18250326|gb|AY071854.1| 1e�72 Oncorhynchus mykiss clone GC71 MHCclass I antigen (Onmy-U71) gene

wt

74 (347)-DN289331 gb|AF063216.1|AF063216 9e�13 Oncorhynchus keta insulin-like growthfactor I precursor (IGF-I.1) gene

wt

79 (289)-DN289332 gi|34850934|dbj|AB113381.1| 7e�16 Salmo salar SYN2 gene for synapsin2 wt140 (420)-DN289343 gi|1332397|gb|L48685.1|TAQTRAN e�140 Tanichthys albonubes Tc1-like transposon

DNA, transposase genewt

166 (242)-DN289346 gi|15824444|gb|AF304198.1|AF304198 4e�48 Salmo salar transferrin gene wt167 (320)-DN289347 gi|22204537|emb|AL732598.10| 7e�07 Zebrafish DNA sequence from clone

CH211-10E8wt

Some of them were re-amplified from genomic DNA (54a, 56); from RNA in RT-PCR (167) or used as probes in northern blots (54a).

Table 4. Summary of clones containing sequences induced during imbibition, not similar to any database sequences

Induced in Code (length)- GenBank_Accn

wt 24 (1 1 1)-DN289327

81 (237)-DN289333

100(129)-DN289336

101(90)-DN289337

112(151)-DN289338

113(115)-DN289339

114(94)-DN289340

ga1 43 (107)-DN289328

89 (75)-DN289334

90 (174)-DN289335

120(74)-DN289341

123(95)-DN289342

160(94)-DN289344

161(192)-DN289345

The clone number is followed by the corresponding length (number of bases). Some of them were re-amplified from genomic DNA (43,81 and 113); from RNA in RT-PCR (113).

Table 2. Clones containing Arabidopsis sequences, not encoding known genes, that are induced during imbibition

Lab code-(length)-GenBank_Accn Localisation in clone/chromosome P-value Induced in

46 (201)-DN289348 T8O18 (63086–63200) chrom 2 2e�048 ga183 (209)-DN289349 T9I1 (12708–12508) chrom 1 e�106 ga195 (117)-DN289350 F13B15 (50302–50418) chrom 2 1e�060 ga197 (43)-DN289351 Multiple clones 3e�12 ga1119 (86)-DN289352 F2J7 (79827–79735) chrom 1 4e�026 ga1

The Genebank accession numbers are included in the first column.

Analysis of seed germination in Arabidopsis thaliana 5

sequence 140 (Table 3) was similar to transposonTc1.

PCR amplification of sequences fromgenomic DNA of Arabidopsis

Primers derived from the flanking regions of thecDNA clones were used to amplify bands fromArabidopsis genomic DNA (Fig. 2). Sequencesamplified included those corresponding to Arabi-dopsis DNA in non-coding regions (95), as well assequences similar with fish sequences (54a, 56, 166and 167) and sequences not having similarities inthe public databases (24, 43, 81, 90, 100, 113 and

161), In many cases, i.e. with primers for amplifi-cation of clones 24, 43, 54, 56, 90, 100, 113 and166, bands larger than expected were amplified,probably due to the presence of introns. Bandsobtained in PCR amplification of the remainingclones (81, 95, 161 and 167) were of the expectedsize.

RT-PCR amplification of sequences fromgerminating Arabidopsis seedlings

Primers were used to amplify cDNA in RT-PCRamplification reactions. Figure 3 shows the result ofRT-PCR amplification of clones 95 (Table 2), 113

ARTICLE IN PRESS

Figure 2. (a) PCR with Arabidopsis genomic DNA showingamplified sequences (sequence 95 from Table 2, se-quences 54, 56, 166 and 167 from Table 3 and sequences24, 43, 81, 90, 100, 113 and 161 from Table 4). Arrowindicates 100 bp. (b) Oligonucleotides used in theamplification reactions and expected product sizes.

Figure 3. (a) RT-PCR. RNA was isolated from Arabidopsisseedlings at 24 h after imbibition of wild type andmutant, seeds. For sequence 167, DNA was also amplifiedin eto and etr mutants. Arrow indicates 100 bp. Oligonu-cleotides used in the amplification reactions and ex-pected product sizes (assuming no introns) were as in Fig.2b.

J.G. de Diego et al.6

(Table 4) and 167 (Table 3). Sequence 167 wasamplified from RNA of wild type and mutants ga1,eto1-1 and etr1-1 and re-cloned to verify theidentity of the band in the RT-PCR experiment withthe band in the cDNA-AFLP. In Fig. 3, two bandsappear in each lane. The amplified fragmentscorrespond to the upper bands in clones 95 and167, and the lower bands for clone 113. All areequal to or smaller than the size of the clones(Tables 2–4).

Figure 4. (a) Northern blot showing expression of three climbibition (6, 12 and 24 h) in the wild-type seeds. Genessequence (90) and AtHVA22a (54b) are shown to be induced dtwo clones in dry seeds (0 h) and at three time points during imseeds. Expression of genes encoded in clones 54a and 24 is awell as in mutants eto1-1, etr1-1, ga1, and an as yet uncharacgermination (named 36, unpublished). Clone 54a contains a sno similarities to any database sequences. These two transimbibition until 12 h and increase again at 24 h. Their exprincreased in eto1-1 mutant backgrounds.

Northern analysis of sequences inducedduring imbibition

Northern blots presented in Fig. 4 show expres-sion of representative clones. In Fig. 4a, genesencoding the small subunit of ribonucleotidereductase, an unknown sequence encodedin clone 90 and AtHVA22, are shown to be inducedduring imbibition. The three genes show a similarpattern in the first 24 h following imbibition,with a peak at 6 h of imbibition and a slightdecrease (12 h) followed by a second peak, moreintense at 24 h. In Fig. 4b, expression of genesencoded in clones 54a and 24 is analyzed duringseed imbibition in the wild-type seeds, as well as inmutants eto1-1, etr1-1, ga1, and an yet unchar-acterized T-DNA insertion mutant that presentsaccelerated germination (named 36, unpublished).Clone 54a contains a sequence homologous to fishDNA. Clone 24 has no similarities to any databasesequences. These two transcripts are alreadypresent in dry seeds, decrease upon imbibition

ones in dry seeds (0 h) and at three time points duringencoding ribonucleotide reductase (86.1), an unknownuring imbibition. (b) Northern blot showing expression ofbibition (6, 12 and 24 h) in wild type as well as in mutant

nalyzed during seed imbibition in the wild type seeds, asterized T-DNA insertion mutant that presents acceleratedequence homologous to fish DNA sequences. Clone 24 hascripts are already present in dry seeds, decrease uponession level is reduced in ga1 as well as in etr1-1 and

ARTICLE IN PRESS

Analysis of seed germination in Arabidopsis thaliana 7

until 12 h and increase again at 24 h. The transcriptlevels are higher in ethylene overproducing mu-tants eto1-1 and reduced in etr1-1, ga1, andmutant 36 backgrounds.

Discussion

A cDNA-AFLP experiment was designed to identi-fy sequences induced during germination of

ARTICLE IN PRESS

J.G. de Diego et al.8

A. thaliana seeds in the wild-type Wassilewskijaand abc33, a GA synthesis deficient mutant in thisecotype (Dubreuq et al., 1996). After 24 h ofimbibition, the seeds already germinated in thewild type but were quiescent in the mutant, thus,in principle, we were expecting to identify genesassociated with the functions of germination in thewild type and genes related with quiescence orgermination block in the mutant. This generalassumption is not supported by our results for tworeasons. First, because sequences corresponding tofunctions associated with germination were iso-lated from the mutant seeds. Thus, although ga1mutants don’t complete germination, they are ableto initiate some of the biochemical processes ofgermination such as nucleic acid synthesis, tran-scription factor regulation and antioxidant de-fense. Second, the majority of the sequencesisolated are non-coding or encode transposons,and many of them are not in the public Arabidopsisgenome databases. The fact that some of thesesequences correspond to Arabidopsis genomicsequences and are expressed during germinationsuggests that germination is a complex processwhose regulation involves multiple genomic se-quences that are regulated by GA by still unknownmechanisms.

A total of 35 sequences were cloned andanalyzed. Of these, eight correspond to protein-encoding genes and the remaining 27 are classifiedin three groups: non-encoding Arabidopsis se-quences (5), sequences with homology to fish DNA(8) and sequences with no homology in anydatabase (14). Thus, a majority of the sequencescloned does not present similarity with Arabidopsissequences in the public genomic databases. Ofthese, many presented similarities with transpo-sons and/or had features of repeated sequences.Expression of transposons and sequences with nodatabase matches is high during sperm cell devel-opment in maize, a process that, like seedgermination, involves cell elongation and resumedgrowth after a period of quiescence (Engel et al.,2003). Also, a large population of retrotransposonshas been described recently in oocytes, and theirrole in the regulation of gene expression duringearly stages of vertebrate development has beensuggested (Peaston et al., 2004).

Some of the sequences cloned correspond toprotein-encoding genes that are induced upon seedimbibition. Among these, GBF5 is a B-zip transcrip-tion factor that was previously reported to beinduced by light (Rook et al., 1998); the smallsubunit of ribonucleotide reductase is the regula-tory subunit of this enzyme that has an essentialrole for obtaining reduced nucleotides required for

DNA synthesis, thus the induction at the onset ofradicle protrusion in a tissue where DNA synthesismust precede cell division was expected (Barrocoet al., 2005). The mRNA levels for this gene in thewild type show a rhythmic regulation with peaks ofinduction after 24 h of imbibition (Fig. 4). The factthat both clones were from upregulated genes inthe ga1 mutant may indicate a different periodicityin the expression of these genes with delayed peaksfor both of them in the ga1 mutant seeds whenimbibed in constant light. It is important to remarkthat the conditions in the cDNA-AFLP experimentinvolved constant light and this may be thereason for observing, in the ga1 mutant back-ground, induction of genes such as GBF5 and thesmall subunit of ribonucleotide reductase. Inconstant light, the regulation may be different toa dark/light cycle and the expression of thesegenes at 24 h after imbibition may be reduced inthe wild type.

Two of the fragments identified correspond togenes related with mitochondrial biogenesis. Clone44 encodes a fragment of At2g26360, a geneencoding a protein with characteristics of mito-chondrial carriers (Palmieri, 1994); whereas clone109 contains a region of At3g27280, a geneencoding prohibitin, a mitochondrial membraneprotein that has been involved in the control of cellcycle progression in mammalian cells (Snedden andFromm, 1997). Although mitochondria are believedto be maintained through seed desiccation, a rapidincrease in respiration is detected upon imbibition(Attucci et al., 1991). The cell elongation anddifferentiation processes that occur during germi-nation may be accompanied by an activation of theprocesses of mitochondrial biogenesis (Logan et al.,2001).

A peak of ABA has been shown to be producedduring imbibition (Grappin et al., 2000); in agree-ment with this, one of the fragments isolated fromwild type correspond to a gene responsive to ABA,AtHVA22a. Wild-type seeds, capable of completinggermination apparently express this gene greaterthan ga1 mutants.

Interestingly, we find mRNAs encoding tworibosomal proteins that are induced upon imbibi-tion, and one is plastid-encoded. The induction ofmRNAs encoding ribosomal proteins has beenalready reported in seed germination (Toorop etal., 2003; Nakabayashi et al., 2005) and in earlyphases of plastid development (Harrak et al.,1995). Recently, the ribosome is considered to bea versatile enzyme in which the different subunitsmay have important roles in the control oftranslation (see for example Kean, 2003). Diverseribosomal subunits may direct the selective trans-

ARTICLE IN PRESS

Analysis of seed germination in Arabidopsis thaliana 9

lation of different mRNAs due to sequence com-plementarities.

It was surprising to isolate so many non-codingsequences in both genotypes. This may be a resultof the experiment being performed in constantlight, and deserves further investigation because itshows that seeds imbibed under these conditionsmay be a source of non-coding expressed RNAsequences or sequences located in the genomicregion of limited accessibility (centromeres).

The sequences described in Table 2 correspond toregions of BAC clones for which no gene ortranscription unit has as yet been annotated. Someare located close to protein coding genes, forexample sequence 95 which is at about 200 basepairs upstream of At2g25490.1 encoding an F-boxprotein. It may be interesting to investigate furtherif the sequence derives from coordinated transcrip-tion with neighbor genes. The fact that all fiveclones corresponding to non-coding Arabidopsissequences were induced and isolated in the ga1mutant may indicate an effect of GA in determiningchromatin status or the particular responsiveness ofcertain genomic regions to this phytohormone.

Of particular interest was that, among 35sequences analyzed, 22 had no significant similaritywith the Arabidopsis genome sequence in thecurrent databases (TAIR, EBI, etc.). Of them, 13sequences were amplified from Arabidopsis geno-mic DNA. Thus, most of the new sequences nothaving homologies with the Arabidopsis genomemust be indeed Arabidopsis genomic sequences andare probably located in centromeric regions,immersed in repeated regions of poor accessibilityduring the isolation of fragments for cloning in BAClibraries for genomic sequencing.

The results obtained may be biased as aconsequence of the conditions in the growth roomused to imbibe the seeds during the cDNA-AFLPexperiment (24 h, constant light). Genes andsequences induced under these conditions may bedifferent from those under other, more naturalcircumstances, i.e. with a dark/light cycle. It hasbeen shown that many genes follow a circadianrhythm upon seed imbibition under light/dark cycle(Zhong et al., 1998). Thus, we may have identifiedat 24 h of imbibition as up-regulated in the GAmutant, genes following a delayed induction. Thatis to say these genes’ expression apparently peaksat this time in the GA mutant due to decreasedexpression in the wt. Thus, their apparent increaseis only due to different periodicity, delayed in themutant when compared with the wild type.Germination under continuous illumination mayresult in additional effects, i.e. the high rate ofnon-coding and transposon sequences cloned may

be a consequence of this. This may explain the factthat a high proportion of the clones obtained arehomologous to fish transposons or contain entirelynew sequences in the sense that no homology isfound in the databases.

Results of PCR reactions with Arabidopsis genomicDNA, RT-PCR and northern blot show that many ofthese new sequences belong to the Arabidopsisgenome and that the expression of some of themduring germination is hormone-dependent. In factthe expression of clones 24 and 54a (Fig. 4b) isreduced in ga1-1 mutant seeds. Ethylene has beenimplicated in the modulation of responses to GA(Achard et al., 2003) and also in the control of seedgermination because etr 1-1 mutants, insensitive toethylene, germinate slower than the wild type(Bleecker et al., 1988). Thus, it is not surprising tofind that expression of clones 24 and 54a is reducedin both mutant backgrounds ga1-1 and etr1-1,whereas a slight increase is observed at 24h ofimbibition in the mutant eto1-1 background. Furtheranalysis of the expression of these sequences mayyield information on the interaction between thesephytohormones early in development.

Previous work using differential cDNA display tocompare transcripts induced or repressed by GA inGA deficient seeds of tomato (Cooley et al., 1999)allowed the identification of genes induced by GAduring germination. Our results with Arabidopsisseeds germinated in constant light have shown theimportance of cDNA-AFLP protocol for the identi-fication of new genomic sequences expressed inseed germination. Our results like Cooley’s et al.(1999) and Nakabayashii et al. (2005) suggestcomplex genomic dynamics during this process.The new sequences obtained may be useful for theanalysis of gene regulation in germination.

Acknowledgements

The cDNA-AFLP experiment was done with sup-port of a grant of Ministerio de Educacion y Culturain the program Acciones Integradas Hispano-Fran-cesas to Philippe Grappin and Emilio Cervantes(HF1997-0076). Juana G. de Diego and F. DavidRodrıguez thank to Consejerıa de Ciencia y Culturade Castilla y Leon for grant SA/007-03.

References

Achard P, Vriezen WH, Van Der Straeten D, Harberd NP.Ethylene regulates Arabidopsis development via themodulation of DELLA protein growth repressor func-tion. Plant Cell 2003;15(12):2816–25.

ARTICLE IN PRESS

J.G. de Diego et al.10

Attucci S, Carde JP, Raymond P, Saint-Ges V, Spiteri A,Pradet A. Oxidative phosphorilation by mitochondriaextracted from dry sunflower seeds. Plant Physiol1991;95:390–8.

Bachem CWB, van der Hoeven RS, de Bruijn SM,Vreugdenhil D, Zabeau M, Visser RGF. Visualization ofdifferential gene expression using a novel method ofRNA fingerprinting based on AFLP: analysis of geneexpression during potato tuber development. Plant J1996;9(5):745–53.

Barroco RM, van Poucke C, Bergevoet JHW, de Veylder L,Groot SPC, Inze D, et al. The role of the cell cyclemachinery in resumption of postembryonic develop-ment. Plant Physiol 2005;137:127–40.

Bewley JD. Seed germination and dormancy. The PlantCell 1997;9:1055–66.

Bleecker AB, Estelle MA, Somerville C, Kende H.Insensitivity to ethylene conferred by a dominantmutation in Arabidopsis thaliana. Science1988;241:1086–9.

Bove G, Lucas P, Godin B, Oge L, Jullien M, Grappin P.Gene expression analysis by cDNA-AFLP highlights a setof new signaling networks and translational controlduring seed dormancy breaking in Nicotiana plumba-ginifolia. Plant Mol Biol 2005;57(4):593–612.

de Castro RD, van Lammeren AAM, Groot SPC, Bino RJ,Hilhorst HWM. Cell division and subsequent radicleprotrusion in tomato seeds are inhibited by osmoticstress but DNA synthesis and formation of microtub-ular cytoskeleton are not. Plant Physiol2000;122:327–36.

Chaboute M-E, Clement B, Sekine M, Philips G, Chaubet-Gigot N. Cell cycle regulation of the tobacco ribonu-cleotide reductase small subunit gene is mediated byE2F-like elements. The Plant Cell 2000;12:1987–99.

Chae HS, Faure F, Kieber JJ. The eto1, eto2, and eto3mutations and cytokinin treatment increase ethylenebiosynthesis in Arabidopsis by increasing the stabilityof ACS protein. The Plant Cell 2003;15:545–64.

Chang C, Kwok SF, Bleecker AB, Meyerowitz EM.Arabidopsis ethylene-response gene ETR1: similarityof product to two-component regulators. Science1993;262:539–41.

Chen CN, Chu CC, Zentella R, Pan SM, Ho THD. AtHVA22gene family in Arabidopsis: phylogenetic relationship,ABA and stress regulation, and tissue-specific expres-sion. Plant Mol Biol 2002;49:633–44.

Cooley MB, Yang H, Dahal P, Mella RA, Downie B, HaighAM, et al. Expression of vacuolar H-ATPase in responseto gibberellin during tomato seed germination. PlantPhysiol 1999;121:1339–47.

Debeaujon I, Koornneef M. Gibberellin requirement forArabidopsis seed germination is determined both bythe testa characteristics and embryonic abscisic acid.Plant Physiol 2000;122:415–24.

Dubreucq B, Grappin P, Caboche M. A new method for theidentification and isolation of genes essential forArabidopsis thaliana seed germination. Mol Gen Genet1996;252:42–50.

Engel ML, Chaboud A, Dumas C, McCormick S. Sperm cellsof Zea mays have a complex complement of mRNAs.Plant J 2003;34(5):697–707.

Fu X, Harberd N. Auxin promotes Arabidopsis root growthby modulating gibberellin response. Nature2003;421:740–3.

Grappin P, Bouinot D, Sota B, Miginiac E, Jullien M.Control of seed dormancy in Nicotiana plumbaginifo-lia: post-imbibition abscisic acid synthesis imposesdormancy maintenance. Planta 2000;210:279–85.

Harrak T, Lagrange C, Bisanz-Seyer S, Lerbs-Mache,Mache R. The expression of nuclear genes encodingplastid ribosomal proteins precedes the expression ofchloroplast genes during early phases of chloroplastdevelopment. Plant Physiol 1995;108:685–92.

Houbaviy HB, Murray MF, Sharp PA. Embryonic stem cell-specific MicroRNAs. Dev Cell 2003;5:351–8.

Jurka J. Repbase update: a database and an electronicjournal of repetitive elements. Trends Genet 2000;9:418–20.

Kamiya Y, Garcıa-Martınez JL. Regulation of Gibberellinbiosynthesis by light. Curr Opin Plant Biol 1999;2:398–403.

Kean KM. The role of mRNA 50-noncoding and 30-endsequences on 40S ribosomal subunit recruitment, andhow RNA viruses successfully compete with cellularmRNAs to ensure their own protein synthesis. Biol Cell2003;95:129–39.

Koornneef M, Bentsink L, Hilhorst H. Seed dormancy andgermination. Curr. Opin. Plant Biol. 2002;5(1):33–6.

Lee S, Cheng H, King KE, Wang W, He Y, Hussain A, et al.Gibberellin regulates Arabidopsis seed germination viaRGL2, a GAI/RGA-like gene whose expression is up-regulated following imbibition. Genes Dev2002;16:646–58.

Lippman Z, Gendrel A-V, Black M, Vaughn MW, Dedhia N,McCombie WR, et al. Role of transposable elements inheterochromatin and epigenetic control. Nature2004;430:471–6.

Logan DC, Millar AH, Sweetlove LJ, Hill SA, Leaver CJ.Mitochondrial biogenesis during germination in maizeembryos. Plant Physiol 2001;125:662–72.

McGinnis KM, Thomas SG, Soule JD, Strader LC, Zale JM,Sun TP, et al. The Arabidopsis SLEEPY1 gene encodes aputative F-Box subunit of an SCF E3 ubiquitin ligase.Plant Cell 2003;15:1120–30.

Nakabayashi K, Okamoto M, Koshiba T, Kamiya Y,Nambara E. Genome-wide profiling of stored mRNAin Arabidopsis thaliana seed germination: epigeneticand genetic regulation of transcription in seed. Plant J2005;41:697–708.

Ogawa M, Hanada A, Yamauchi Y, Kuwahara A, Kamiya Y,Yamaguchi S. Gibberellin biosıntesis and responsesduring Arabidopsis seed germination. Plant Cell2003;15:1591–604.

Palmieri F. Mitochondrial carrier proteins. FEBS Lett1994;346:48–54.

Peaston AE, Evsikov AV, Graber JH, de Vries WN, HolbrookAE, Solter D, et al. Retrotransposons regulate host

ARTICLE IN PRESS

Analysis of seed germination in Arabidopsis thaliana 11

genes in mouse oocytes and preimplantation embryos.Dev Cell 2004;7:597–606.

Peng J, Harberd NP. The role of GA-mediated signalling inthe control of seed germination. Curr. Opin. Plant Biol2002;5:376–81.

Reinhart BJ, Weinstein EG, Rhoades MW, Bartel B, BartelDP. MicroRNAs in plants. Genes Dev 2002;16:1616–26.

Rook F, Weisbeek, Smeekens S. The light regulatedArabidopsis bZip transcription factor gene ATB2encodes a protein with an unusually long leucinezipper domain. Plant Mol Biol 1998;37:171–8.

Silverstone AL, Mak PYA, Casamitjana E, Sun T-P. The newRGA locus encode a negative regulator of gibberellinresponse in Arabidopsis thaliana. Genetics 1997;146:1087–99.

Snedden WA, Fromm H. Characterization of the planthomologue of prohibitin, a gene associated withantiproliferative capacity in mammalian cells. PlantMol Biol 1997;33:753–6.

Strader LC, Ritchie S, Soule JD, McGinnis KM, Steber CM.Recessive-interfering mutations in the gibberellinsignaling gene SLEEPY1 are rescued by overexpressionof its homologue, SNEEZY. Proc Natl Acad Sci USA2004;101(34):12771–6.

Tepperman JM, Zhu T, Chang H-S, Wang X, Quail PH.Multiple transcription-factor genes are early targets ofphytochrome A signaling. Proc Natl Acad Sci USA2001;98(16):9437–42.

Toorop PE, Groot SPC, Hilhorst HWM. Expression of aribosomal protein gene during germination of cabbage(Brassica oleracea f. oleracea) seeds. In: Nicolas G,Bradford KJ, Come D, Pritchard HW, editors. Biology ofseeds: recent research advances. CABI Publishing;2003. p. 191–7.

Vicient C, Delseny M. Isolation of total RNA fromArabidopsis thaliana seeds. Anal Biochem 1999;268:412–3.

Wen CK, Chang C. Arabidopsis RGL1 encodes a negativeregulator of gibberellin responses. Plant Cell2002;14:87–100.

Yamaguchi S, Smith MW, Brown RGS, Kamiya Y, Sun T-P.Phytochrome regulation and differential expression ofgibberellin 3B-hydroxylase genes in germinating Ara-bidopsis seed. Plant Cell 1998;10:2115–26.

Zhong HH, Painter JE, Salome PA, Straume M, RobertsonMcClung C. Imbibition, but not release from stratifica-tion, sets the circadian clock in Arabidopsis seedlings.Plant Cell 1998;10:2005–17.