ARTICLE IN PRESS

Journal of Plant Physiology 166 (2009) 310—323

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

Abbreviationhybridization;�CorrespondE-mail addr

1These autho

www.elsevier.de/jplph

Characterization of CsSEF1 gene encoding putativeCCCH-type zinc finger protein expressed duringcucumber somatic embryogenesis

Agnieszka Grabowskaa,1, Anita Wisniewskab,�,1, Norikazu Tagashirac,Stefan Malepszyd, Marcin Filipeckid

aDepartment of Biochemistry, Faculty of Agriculture and Biology, Warsaw University of Life Sciences, Nowoursynowska159, 02-776 Warsaw, PolandbDepartment of Plant Physiology, Faculty of Agriculture and Biology, Warsaw University of Life Sciences,Nowoursynowska 159, 02-776 Warsaw, PolandcFaculty of Human Life Science, Hiroshima Jogakuin University, 4-13-1 Ushita-Higashi, Higashiku, Hiroshima city,732-0063 JapandDepartment of Plant Genetics Breeding and Biotechnology, Faculty of Horticulture and Landscape Architecture,Warsaw University of Life Sciences, Nowoursynowska 159, 02-776 Warsaw, Poland

Received 20 February 2008; received in revised form 27 May 2008; accepted 2 June 2008

KEYWORDSCCCH zinc finger;Cotyledonprimordium;Cucumber;Embryogenesis;Somatic embryo

ee front matter & 2008jplph.2008.06.005

s: ECS, embryogenic cZE, zygotic embryogeneing author. Tel.: +48 22ess: anita_wisniewska@rs shared equal particip

SummarySomatic embryos obtained in vitro are a form of vegetative reproduction that can beused in artificial seed technology, as well as a model to study the principles of plantdevelopment. In order to isolate the genes involved in somatic embryogenesis of thecucumber (Cucumis sativus L.), we utilized the suppression subtractive hybridization(SSH). One of the obtained sequences was the CsSEF1 clone (Cucumis sativus SomaticEmbryogenesis Zinc Finger 1), with a level of expression that sharply increased withthe induction of embryogenesis. The full length cDNA of CsSEF1 encodes the putative307 amino acid long protein containing three zinc finger motifs, two with CCCH andone with the atypical CHCH pattern. The CsSEF1 protein shows significant similarityto other proteins from plants, in which the zinc fingers arrangement and patterns arevery similar. Transcripts of CsSEF1 were localized in the apical part of somaticembryos, starting as early as the polarity was visible and in later developmental

Elsevier GmbH. All rights reserved.

ell suspension; SE, somatic embryogenesis or somatic embryos; SSH, suppression subtractivesis or zygotic embryo.59 325 33; fax: +48 22 59 325 21.sggw.pl (A. Wisniewska).ation in this work.

ARTICLE IN PRESS

CsSEF1 gene encoding putative CCCH-type zinc finger protein 311

stages marking the cotyledon primordia and procambium tissues. As a result oftransferring an antisense fragment of CsSEF1 into Arabidopsis thaliana abnormalitiesin zygotic embryos and also in cotyledons and root development were observed.& 2008 Elsevier GmbH. All rights reserved.

Introduction

Somatic embryogenesis (SE), the formation ofsomatic embryos from cells other than the zygote,most frequently under in vitro culture conditions, isan important process for biotechnology allowing forthe effective clonal propagation of plants. SE canalso serve as a model for studying the functions ofgenes involved in zygotic embryogenesis, sincethere are generally many fundamental similaritiesin the course of both processes (Dodeman et al.,1997). In dicotyledonous plants, the developmentof the somatic embryo usually involves the globularand heart/torpedo stages, which resemble thedevelopmental stages of a zygotic embryo(Zimmerman, 1993; Mordhorst et al., 1997). Somaticembryos of the cucumber correspond to zygoticones up to the heart stage, whereas later stagescan differ, mainly in the structure of cotyledons(Tarkowska et al., 1994). Somatic embryogenesis asa model for studying genes involved development isof special importance in species other thanArabidopsis thaliana, due to lack of rich collectionsof embryonic mutants. Moreover, there are notmany species that follow a ‘‘typical’’ Arabidopsisthaliana embryo development model (Kaplan andCooke, 1997). Substantial differences can occureven during the first division of the zygote, which inmost plants is transverse, but there are species inwhich the first division of the zygote is longitudinalor oblique. A dormancy of seeds, which is tradi-tionally regarded as the end-point of embryogen-esis in some plants, may not occur at all. In grasses,the embryo in the dormant stage already has leafprimordia and adventitious roots, whereas inorchids the development of the embryo is arrestedat the globular stage (Kaplan and Cooke, 1997).

For cucumber, two strategies aimed at obtainingembryogenic tissue have been developed: one isbased on a synthetic auxin (2,4-dichlorophenoxya-cetic acid) (Wroblewski et al., 1995), and thesecond involves the presence of a cytokinin in themedium (benzylaminopurine) (Burza and Malepszy,1998). These methods maintain a cell suspension inthe undifferentiated state (an embryogenic cellsuspension, ECS) as a result of constant auxin orcytokinin pressure in the liquid medium. Transfer-ring the culture to a medium lacking growthregulator initiates organized cell divisions that lead

to the formation of somatic embryos. Thesestrategies for the somatic embryogenesis of cu-cumber have many advantages that favor molecularstudies. These include a relatively short time forinduction of ECS from primary explants, long-termmaintenance of embryogenic potential by the ECS(for over 2.5 years), and easy observation andaccess to the individual developmental stages ofthe embryo (Burza et al., 2006).

Only a few genes involved in somatic embryogen-esis in the cucumber have been described thus far:CUS1, Cs-XTH1 and Cs-XTH3 (Filipecki et al., 1997;Malinowski et al., 2004). The largest numberof genes participating in SE, over a dozen, hasbeen characterized in carrot (for instance EMB-1(Wurtele et al., 1993), CEM-6 (Sato et al., 1995),CHB (Kawahara et al., 1995; Hiwatashi and Fukuda,2000), DcSERK (Shah et al., 2001) or KDC2(Formentin et al., 2006)). For other species, thereare only single examples. Recently, it was con-firmed that the soybean orthologue of the Arabi-dopsis MADS-domain transcription regulator AGL15is able to increase somatic embryo production(Thakare et al., 2008) and PaVP1 transcriptionfactor can be regarded as a good marker of somaticembryogenesis in Norway spruce (Fischerova et al.,2008). In Medicago truncatula, MtSERF1 is stronglyexpressed in somatic embryos and its transcriptiondepends on ethylene biosynthesis and perception(Mantiri et al., 2008).

Over 50 genes encoding transcription factors,which take part in zygotic or somatic embryogen-esis, have been identified and characterized inArabidopsis. Among them there are genes involvedin the initiation and maintenance of embryodevelopment, such as LEC1 (Meinke, 1992), L1L(Kwong et al., 2003), LEC2 (Stone et al., 2001;Braybrook et al., 2006; Stone et al., 2008) and BBM(Boutilier et al., 2002). Their ectopic overexpres-sion enables somatic tissues to acquire embryonicfeatures inducing somatic embryogenesis on seed-lings or tissues in culture, whereas their mutantsshow premature germination and non-embryoniccharacters in the embryos (Harada, 2001).

Zinc finger proteins can act as DNA-bindingtranscription factors; however, some can bind toRNA. The CCCH zinc fingers recognize specificAU-rich sequences and interact with RNA (reviewedby Hall, 2005). For Arabidopsis and rice, the

ARTICLE IN PRESS

A. Grabowska et al.312

comprehensive computational analyses were doneidentifying 68 and 67 CCCH family genes, respec-tively (Wang et al., 2008). Expression studiesindicated that CCCH proteins exhibit a variety ofexpression patterns, suggesting diverse functions.Compared with other gene families in rice andArabidopsis thaliana, the CCCH gene family is oneof the largest families in plants.

In the present paper, we describe a new gene,CsSEF1 (Cucumis sativus Somatic EmbryogenesisZinc Finger 1), with expression sharply increasedwith the induction of somatic embryogenesis. Thelocation of CsSEF1 transcripts in somatic embryoswas determined by in situ hybridization. Thepossible effects of CsSEF1 overexpression andsilencing of its close homologues were shown intransgenic Arabidopsis plants.

Materials and methods

Plant materials

The embryogenic tissue was initiated from shoot tips ofcucumber plants (Cucumis sativus L. var. Borszczagowski,denominated here ‘‘line B’’). The ECS was generated onMS medium (Murashige and Skoog, 1962) modified byWroblewski et al. (1995). Induction of somatic embryos(SEs) from the ECS for the suppression subtractivehybridization (SSH) technique and for preparing a cDNAlibrary have been described previously by Linkiewiczet al. (2004). The seeds of the Arabidopsis thalianaecotype Landsberg erecta (Ler) and Columbia (Col-0)were obtained from the Nottingham Arabidopsis StockCentre. The seeds of Arabidopsis T-DNA tagged lines withinsertion in loci: At2g19810 (‘Arabidopsis thaliana zincfinger (CCCH-type) family protein gene’)-SALK_151571and in At5g07500 (‘Arabidopsis thaliana PEI1; nucleicacid binding/transcription factor gene’)-SALK_108149(Alonso et al., 2003) were provided by the Salk InstituteGenomic Analysis Laboratory (La Jolla, CA, USA).

RNA isolation from cucumber organs

Total RNA was isolated using Trizols reagent (Invitro-gen) according to the manufacturer’s instructions, fromleaf, root, female flower, fruit, somatic embryos col-lected after 7 and 14 d post-SE induction and mix in equalproportion (w/w), mature zygotic embryos collectedfrom 15-d-old fruits, and ECS. mRNA was isolated fromthe total RNA using Dynabeadss oligo-dT (Dynal) accord-ing to the manufacturer’s instructions.

SSH library construction and screening

The SSH technique (Diatchenko et al., 1999) wascarried out using a PCR-Select cDNA Subtractive Kitaccording to the manufacturer’s protocol (Clontech,

USA). The ‘driver’ cDNA was prepared from ECS beforethe induction of somatic embryogenesis and the ‘tester’cDNA was prepared from tissue 2, 6, 12, 24 h, 3, 7, 14 dafter the induction of embryogenesis. The subtractedPCR products generated by SSH were ligated to pBlue-script SK(+) vector (Stratagene), transformed into Es-cherichia coli (XL1-Blue) (Stratagene) and plated onLuria-Bertani (LB) plates containing X-Gal/IPTG andampicillin. The total 384 white colonies were selectedfor further analysis, preamplified, replicated in four 96-well multiplates and arrayed on Hybond N nylonmembranes (Amersham). To identify differentially ex-pressed genes, reverse RNA hybridization analysis wasperformed with cDNA probes labeled using High Prime(Roche) and a32P dCTP (Amersham). The ‘tester’ and‘driver’ cDNAs from subtracted and unsubtracted li-braries were used in separate labeling reactions (fourprobes) and for hybridization with arrayed bacterialcolonies according to Sambrook et al. (1989). Radioactiveimage analysis was performed using a Kodak K screen andMolecular Imager (BioRad).

cDNA library construction and screening

Poly(A)+ RNA (4 mg) extracted from 7-d-old cucumbersomatic embryos were used to construct a library usingthe Stratagene ZAP-cDNAs Synthesis Kit and ZAP-cDNAs

Gigapack III Gold Cloning Kit following the manufacturer’sinstructions. The initial complexity of the library was9.6� 105 pfu. The 105 recombinants from the amplifiedlibrary were used for screening plated at a density of10,000 pfu/plate with competent XL1-Blue MRF’ cells.The phagemid screening was performed according to themanufacturer’s instructions using Porablot NCL mem-brane discs (Macherey–Nagel) and radiolabeled SSHfragment as a probe. The excision of the pBluescriptcandidate phagemids from Uni-ZAP XR vector wasperformed according to the manufacturer’s instructions(Stratagene). The candidate phagemids were amplified inSOLR E. coli strain and analyzed by EcoRI/XhoI restrictionand sequencing.

DNA isolations and hybridization

Genomic DNA was extracted from cucumber leavesaccording to Michaels et al. (1994). For DNA hybridizationblot, 5mg of DNA (per sample) was digested andelectrophoresed for 15 h (1.2 V/cm) in a 0.8% agarosegel, and then transferred on the Hybond N+ membrane(Amersham Pharmacia Biotech) according to Koetsieret al. (1993). Hybridization and washing were performedaccording to the membrane manufacturer’s protocol.Radioactive image analysis was performed using a KodakK screen and Molecular Imager (BioRad). Full length cDNAof the gene CsSEF1 was used as a molecular probe for DNAhybridization. The PCR-amplified DNA fragments werepurified using a QIAEX kit (Qiagen) and subsequentlyradiolabeled as described above.

ARTICLE IN PRESS

CsSEF1 gene encoding putative CCCH-type zinc finger protein 313

Real-time RT-PCR for CsSEF1 gene

Real-time RT-PCR was performed with the LightCy-cler–RNA Amplification Kit SYBR Green I (Roche) accord-ing to manufacturer’s instructions (RT at 55 1C for 10min,95 for 30 s, 45 cycles of 95 1C for 0 s (slope 20 1C/s), 60 1Cfor 10 s, 72 1C for 30 s; product length 211 bp; primers:50-TCC ACC CAG ACC GATACC-30 and 50-GGA GAC AAA GGTGGC GAG-30). The transcript level of CsSEF1 wasestimated relative to cucumber 25S rRNA level (amplifi-cation conditions as above; product length 300 bp;primers: 50-CCA GGT CAG GCG GGA CTA C-30 and 50-CGCAAC GGG CTC TCT CAC C-30). The transcript levels arepresented as values relative to ECS ( ¼ 1). Threeindependent experiments were performed.

In situ hybridization

Tissue fixation, embedding, sectioning and in situhybridization with DIG-labeled sense and antisense RNAprobes were performed as described by Long et al. (1996)with modifications described by Malinowski et al. (2004).Templates for the synthesis of RNA probes were preparedby cloning the CsSEF1 PCR fragment (1140 bp; primers:50-GCA CTT CCC ACC TCA TCT TCA-30 and 50-ATT GTA ACTCAC TCA CTT GCC-30) into the pCRII-TOPO plasmid(Invitrogen). The protocol used was as follows: 30 cyclesof 94 1C for 30 s, 60 1C for 30 s, 72 1C for 1min. PCR wasperformed using a 20 mL reaction mixture that contained20 ng of plasmid DNA, 200 mM each of dNTP, 500 nM ofeach primer, 1� Taq buffer with KCl, 1.25mM MgCl2 and1U Taq DNA polymerase (Fermentas). The RNA probeswere synthesized using SP6 and T7 RNA polymerases(Roche).

Gene constructs and bacterial strains

Transformation of Arabidopsis plants was performedusing the Agrobacterium tumefaciens strain LBA4404 orEHA105 carrying pCAMBIA1380 vector (hygromycin resis-tance selection marker; Cambia, Canberra, Australia)modified by insertion of the 35S promoter at the EcoRI/BamHI restriction site (A. Smigocki, Molecular PlantPathology Laboratory, Agricultural Research Service,Beltsville, MD, USA) and appropriate gene fragments.For silencing, a 518 bp fragment of CsSEF1 cDNA clone,which was obtained from the SSH cDNA library, wasintroduced into the HindIII restriction site of pCAM-BIA1380+35S. The T-DNA containing an antisense orienta-tion of cDNA fragment was selected and designated as35S::CsSEF1as. For overexpression, the CsSEF1 codingsequence (1071 bp) was amplified with specific primerscontaining HindIII and PstI restriction sites (underlined)added to the 50 end: 50-CTA AGC TTA ATC AAT GCT CCTCAATCC CA-30 and 50-GAC TGC AGATTG TAA CTC ACT CACTTG CC-30, restricted and ligated to pCambia 1380+35Svector. The PCR protocol used was as described above.This T-DNA was designated as 35S::CsSEF1s. Constructsfor transformation were verified by sequencing of PCR-amplified fragments and ligation sites.

Growth conditions and transformation of Arabidopsisplants

Germination of the seeds was synchronized by atreatment at 4 1C for 48 h in the dark. Plants were grownto the flowering stage in a growth chamber (VersatileEnvironmental Test Chamber MLR-35OH, Sanyo) at 22 1Cwith 10 h light (250 mE/m2 s) during the vegetative stageand 14 h light during generative growth (Wilson, 2000).

Floral dip of plants was essentially as described byClough and Bent (1998). The plants designated as T0 weregrown to maturity and seeds were harvested.

Seeds from floral-dipped plants (T1 generation) weresurface sterilized and suspended in 0.1% sterile agarose andplated on selective medium (1/2 MS) (Murashige and Skoog,1962) containing 1% sucrose, 0.8% agar, 15mg/L hygro-mycin, 200mg/L cefotaxime or 300mg/L timentin. Thegrowth conditions were as described above. After 3 weeks,resistant T1 seedlings were transferred to a fresh selectionmedium with 1.5% agar. After 1 week, some of theseseedlings were put into the soil. The T2 seeds from each T1plant were harvested and screened for the hygromycinresistance phenotype and phenotypic changes separatelyfor each T1 progeny. Macroscopic observations of thehygromycin-resistant seedlings were made after 1, 3 and4 weeks. Observations of the mature plants of T1 and T2progeny were performed during and after flowering.

Analysis of Arabidopsis transformants and mutants

Genomic DNA from T1 and/or T2 transformants, mutantsand non-transformed wild-type plants (control) was ex-tracted and PCR was performed by REDExtract-N-AMP

TM

PlantPCR Kit, according to the manufacturer’s protocol (Sigma).

Antibiotic-resistant plants were screened by PCR for thepresence of the hygromycin phosphotransferase gene (hpt)(primers: 50-GGC GAG TAC TTC TAC ACA-30 and 50-GCG AAGAAT CTC GTG CTT-30, length product 887bp long) andCsSEF1 cDNA sequence (50-TCC ACC CAG ACC GAT ACC-30

and 50-GGA GAC AAA GGT GGC GAG-30, length product211bp long). The PCR protocol used was 30 cycles of 94 1Cfor 30 s, 62 1C (for hpt gene) or 60 1C (for CsSEF1 gene) for30 s, 72 1C for 1min (for hpt gene) or 30 s (for CsSEF1 gene).

The insertions in T-DNA tagged lines (loci At2g19810 orAt5g07500) and homozygous states were confirmed bythree-primer-PCR with LBb1 (left border primer ofpROK2; 50-GCG TGG ACC GCT TGC TGC AAC T-30) andappropriate two genomic primers: 50-TCG CCG TCG TGTTTG TTT CTT-30 and 50-CAT TGC GAC TTG CAT CTT ACATCA-30 for At2g19810; 50-TTC CTT CAC GTA AAC GCT GCC-30 and 50-ACA CAA GTT CCC GGC GTT ACA-30 forAt5g07500 locus. The protocol used was as follows: 34cycles of 94 1C for 15 s, 60 1C for 30 s and 72 1C for 1min.

Results

Characterization of CsSEF1 sequence

Differential screening of the SSH library involvedhybridization of arrayed bacterial colonies with

ARTICLE IN PRESS

A. Grabowska et al.314

probes based on subtracted and unsubtracted cDNAisolated from tissues before and after induction ofsomatic embryogenesis. The application of the SSHstrategy resulted in a normalized library and probesenriched in cDNAs characteristic for the stagespreceding (probe cDNA only) and following theinduction of somatic embryogenesis (probe andlibrary). After the library screening, 117 cloneswere selected, of which the majority correspondedto transcripts induced during somatic embryogen-esis (data not shown). The nucleotide sequences ofthe studied clones were compared with a proteinsequence database using the algorithm blastx(Altschul et al., 1990). One of the clones, the518 bp long CsSEF1, was chosen for further ana-lyses. The sequence of this clone showed significantsimilarity to proteins possessing zinc fingers withatypical motif CCCH.

The CsSEF1 fragment obtained from the SSHlibrary was used to isolate seven independentclones from the conventional cDNA library. Thesequence of the longest cDNA clone (1217 bp) wassubmitted to the EMBL database (accession numberAJ870303). The sequence of the clone contains an84 bp long 50-UTR region, a 924 bp open readingframe and a 209 bp long 30-UTR region. Theputative protein coded by CsSEF1 consists of 307amino acids. The amino acid sequence of CsSEF1revealed the presence of three zinc fingermotifs: the upper had the common patternC-X5-H-X4-C-X3-H and two rarely found ones,middle C-X7-C-X5-C-X3-H and lower C-X5-C-X4-C-X3-H (Figure 1).

The protein coded by CsSEF1 showed the greatestsimilarity (using the algorithm blastp) to theArabidopsis thaliana ‘zinc finger (CCCH-type) fa-mily protein’ (NP_194648, 56% within a 337 aminoacid overlap, e-value 2e-66), Arabidopsis thaliana‘zinc finger (CCCH-type) family protein’(NP_179571, 56% similarity within a 345 amino acidoverlap, e-value 1e�65) as well as to ‘ATCTH;Arabidopsis thaliana Cys3His zinc finger protein’(NP_180161, 60% similarity within a 318 amino acidoverlap, e-value 2e�62) and to the describedArabidopsis thaliana ‘PEI1; nucleic acid binding/transcription factor’ (NP_196367, 70% similaritywithin a 127 amino acid overlap, e-value 2e�38),which also contains CCCH motif. The multiplesequence alignment of CsSEF1 and four homologousproteins from Arabidopsis showed high similarity inthe domain containing three zinc finger motifs(Figure 2A). The compared amino acid sequencesfrom other species showed the highest conservationwithin the domain containing three zinc fingers(Figure 2B), while regions flanking this domaindisplayed high variability.

Genomic DNA gel blot analysis for CsSEF1gene

In order to determine the number of copies ofCsSEF1 in the cucumber genome, genomic DNA gelblot analysis was employed. The enzymes used in theanalysis were selected with respect to the presence(for EcoRI, Eco32I) or absence (HindIII, XbaI) ofappropriate restriction sites within the cDNA se-quence. The molecular probe for the CsSEF1 clonewas the entire cDNA sequence (1217bp). As a resultof hybridization for each restriction enzyme, a singlestrong signal was found, and in the case of enzymesrecognizing a site within the studied sequence, aweaker second band was observed. These observa-tions point to the existence of a single copy of CsSEF1gene in the cucumber genome (Figure 3).

Analysis of CsSEF1 transcript accumulation

The level of transcription of CsSEF1 was determinedwith the use of real-time semi-quantitative RT-PCR,and the results were standardized relative to control25S rRNA gene from cucumber. An increased accumu-lation of transcripts in somatic embryos (SEs), leavesand female flowers, compared to ECS, was observed.The level of transcript accumulation in zygoticembryos and root, compared to ECS, was lower andsimilar, respectively (Figure 4). Similar results wereobtained after RNA gel blot analysis (data not shown).

In situ hybridization was carried out using cRNAsense and antisense probes specific for CsSEF1 mRNAto examine the spatial pattern of CsSEF1 expressionin cucumber somatic embryos. Cucumber somaticembryos in different stages were transversely andlongitudinally sectioned. Signals were observed in theglobular, heart/torpedo and mature stages of em-bryos (Figure 5). The analysis demonstrated a highaccumulation of transcripts in the putative apicaldomain of the late globular and early heart embryos(Figure 5A, B) and then in the cotyledon primordia ofthe heart embryo and older ones (Figure 5D, F). Thetranscript localization appears to be invariable duringsomatic embryo development. The transverse sectionalso showed the presence of the CsSEF1 transcript inthe procambium tissues, beginning from the heartstage onwards (Figure 5C). In the case of matureembryos with developed three cotyledon primordial,transcript accumulation was observed in everyprimordium (Figure 5F).

Transformation of Arabidopsis plants

Arabidopsis plants were transformed with twogene expression cassettes. The T-DNA 35S::CsSEF1as

ARTICLE IN PRESS

Figure 1. Nucleotide and putative amino acid sequences of CsSEF1. Start codon ATG is marked by an ellipse. Therestriction sites of enzymes used in DNA hybridization are underlined (besides HindIII and XbaI, which do not possestheir sites in cDNA sequence); EcoRI: GAATTC; Eco32I: GATATC. Molecular probe for DNA gel blot was full length cDNA ofCsSEF1. In putative protein sequence the characteristic zinc finger motifs are marked by double underline. Conservativecysteines and histidines are bold. First and last nucleotides of the sequence used to gene silencing are in boxes. Thetermination codon is marked with an asterisk. (A)n–poly(A) sequence.

CsSEF1 gene encoding putative CCCH-type zinc finger protein 315

for silencing contained the CsSEF1 fragment(518 bp) obtained from the SSH library, and theT-DNA 35S::CsSEF1s for overexpression carried afull length clone of CsSEF1 (1217 bp). The transfor-mant seeds were plated on selective media. Forboth constructs the transformation efficiency wasca. 0.6%. For further analyses we selected 14 plantsinto which 35S::CsSEF1as had been introduced and13 with 35S::CsSEF1s.

Molecular analysis of Arabidopsistransformants and mutants

Stable integration of the transgenes was con-firmed by PCR. Analysis with the use of primersspecific for hpt gene demonstrated the presence ofthe expected product (910 bp) in all the transgenic

plants carrying the 35S::CsSEF1s or 35S::CsSEF1ascassettes (Figure 6A, B). The negative control wasDNA isolated from non-transformed plants, whereasthe positive control was plasmid DNA (pCAM-BIA1380+35S) in equimolar concentration. PCR ana-lysis was also performed to confirm the presence ofCsSEF1 cDNA. In all the transformants into which35S::CsSEF1s and 35S::CsSEF1as cassettes were in-troduced (except for two plants 5 and 9), a 211bplong product was detected, indicating the integrationof the T-DNA was correct for most of the plants(Figure 6C, D). These results were also confirmed bygenome DNA gel blot hybridization using a molecularprobe consisting of a fragment of the hygromycinphosphotransferase gene that was performed withselected transgenic lines (data not shown).

The insertions in two T-DNA tagged Arabidopsismutants (At2g19810 and At5g07500 loci) and

ARTICLE IN PRESS

Figure 2. (A) An alignment (ClustalW) of CsSEF1 (CAI30889.1) and the most closely related proteins and PEI1 fromArabidopsis thaliana (GenBank accession number-Arabidopsis locus ID: PEI1, NP_196367-At5g07500;NP_180161-At2g25900; NP_194648-At4g29190 and NP_179571-At2g19810). (B) An alignment of the domain containingthree zinc finger motifs (ClustalW) of CsSEF1 and CAN83255 (Vitis vinifera), BAF04190 (Oryza sativa), ABK92961(Populus trichocarpa), XP_001774473 (Physcomitrella patens subsp. patens), ABN08215 (Medicago truncatula)ABI30334 (Capsicum annuum). Identical, conserved and semi-conserved residues in all aligned sequences are indicatedby asterisks (*), colons (:) and dots (.), respectively. Dashes (–) signify gaps.

A. Grabowska et al.316

homozygous states were confirmed by three-primer-PCR, where two primers are complementary to thegenomic sequences flanking the insertion site (giving

900bp long product from the wild-type allele) andone primer is complementary to the T-DNA sequence(LBb1, 110bp from the left border). The 400bp long

ARTICLE IN PRESS

Figure 3. DNA hybridization of cucumber genomic DNA.5 mg of cucumber genomic DNA was digested withrestriction enzyme indicated on each line. The blot washybridized with the molecular probe which was fulllength cDNA of CsSEF1 gene. The estimated sizes for DNAbands are shown on right.

Figure 4. The expression of CsSEF1 gene in variousorgans. Transcript levels were analyzed by real-time RT-PCR. Transcript levels are given as relative values to CES(the value of 1), after being normalized to the 25Sribosomal RNA levels. Data are shown as the means withvariation bars (SE) from three independent reactions.ECS-embryogenic cell suspension, SEs–somatic embryos,ZEs–zygotic embryos. R–roots, Fl–female flowers,L–leaves.

CsSEF1 gene encoding putative CCCH-type zinc finger protein 317

PCR fragment in the case of insertion in At2g19810locus and 500bp long fragment in the case ofAt5g07500 locus confirmed the insertions of T-DNAin both mutants and also their homozygous state(Figure 6E). According to the Salk Institute GenomicAnalysis Laboratory, the provider of the insertionmutants and methods to verify them, the LBb1primer can sometimes produce an extra, non-specific

band that is present in our At5g07500 mutantanalysis (700 bp long fragment; Figure 6E).

Phenotypic analysis of Arabidopsistransformants and mutants

In the case of T1 hygromycin-resistant seedlingscontaining 35S::CsSEF1as cassette, three phenoty-pic classes could be distinguished. The firstclass consists of seedlings with a single cotyledonand reduced or normally developed root system(Figure 7A, C). The second phenotypic class containsseedlings that had developed two cotyledons and areduced root system (Figure 7B). The third classconsisted of seedlings with two cotyledons and anormally developed root system (Figure 7D). Thepercentage participation of individual classes wasas follows: first class 62.5%, second 9.2% and third28.3%. Only the seedlings from the third class wereable to continue normal development. The matureplants obtained from these seedlings showedphenotypic alterations that mainly concerned theappearance of the siliques. In addition to siliquesthat looked the same as in the control plants(Figure 7G), siliques broader on the top (Figure 7H, I),shorter and flattened (Figure 7I, J), and showingother severe deformations (Figure 7N), could bedistinguished. These changes occurred with variedintensity in all 14 analyzed plants. Seeds wereproperly developed only in normally appearing silique(Figure 7K). In the phenotypically altered silique,normally developed seeds were accompanied byseeds in which development was arrested at differentstages (Figure 7L, M). In extreme cases, no seeds atall were observed (Figure 7N).

In the transformation of wild-type Arabidopsisplants with a 35S::CsSEF1s cassette for overexpres-sion and both types of mutants, no macroscopicchanges at either the seedling stage or in themature plants were observed. These plants devel-oped normal cotyledons as well as rosettes,inflorescence shoots, siliques and produced seeds.

Discussion

Zinc finger proteins belong to the most abundantgroup of proteins in eukaryotes. They have verydifferent structures as well as functions, whichinclude DNA or RNA recognition, RNA packaging,transcriptional activation, protein folding and as-sembly, and lipid binding. Laity et al. (2001) defineda zinc finger as any small, functional, independentlyfolded domain that requires coordination of one ormore zinc ions to stabilize its structure. Many

ARTICLE IN PRESS

Figure 5. In situ detection of CsSEF1 mRNA in cucumber somatic embryos. The signal was observed in the presumptivecotyledon primordia tissues of the globular stage of embryos (A), heart/torpedo (B) and torpedo stages of embryos(D, F) as well in procambium cells, from heart/torpedo stage (B) to mature stage (D). The transverse section shows alsothe presence of CsSEF1 transcript in the procambium tissues (C). The transcript accumulation was observed in primordiaof embryos which generated three primordia (F). The scale bars represent 250 mm. Results come from representativeexperiments that were repeated at least five times with similar results. The in situ hybridization was carried out usingcRNA antisense (A–D, F) and sense (E) probes specific for CsSEF1 mRNA. The arrows indicate cotyledon primordia (B, F)and procambium tissues (C, D).

A. Grabowska et al.318

proteins containing the classical CCHH zinc fingerare transcription factors that function by recogni-tion of specific DNA sequences, whereas thosehaving the CCCH zinc finger motif often bind to RNAtargets (Carballo et al. 1998; Tenlen et al. 2006;Pagano et al. 2007).

CsSEF1 is one of the many genes with levels ofexpression that are increased with the induction ofsomatic embryogenesis. The full length of CsSEF1cDNA codes a protein with three conserved zincfinger motifs, one with pattern CHCH (upper motif,different from the classical CCHH) and two withpatterns CCCH (middle and lower motifs). Thesemotifs are not the most typical; however, they arewell conserved among proteins from differentspecies. One of these plant proteins is PEI1, whosegene is the only CsSEF1 homologue from Arabidop-sis with documented participation in embryogen-esis. PEI1 is involved in the development of zygoticembryos from the globular stage to the heart stage

(Li and Thomas, 1998). Among other plant proteinswith zinc finger motifs, only a few have beendescribed that are involved in the development ofembryos. One of these is RIE1, a novel RING-H2finger protein with visible effects on seed develop-ment and constitutive expression in Arabidopsis.Mutation of the RIE1 gene was lethal under normalgrowing conditions, and mutant embryos werearrested from the globular to late torpedo stage(Xu and Li, 2003). There is also an interestingexample of a protein that contains zinc fingermotifs similar to those of the atypical middle andlower motifs in CsSEF1–HUA1 in Arabidopsis. Thisprotein possesses six such motifs and belongs toa family of nine Arabidopsis genes containingmultiple, tandem, CCCH-type zinc finger motifs(Li et al., 2001). HUA1 likely participates in a newregulatory mechanism governing flower develop-ment and acts only in the third and fourth whorls ofthe flower. No evidence exists for CCCH-type zinc

ARTICLE IN PRESS

Figure 6. PCR analysis of transgenic and mutant Arabidopsis plants. DNA was isolated from transgenic plants with the35S::CsSEF1s (A, C) and with the 35S::CsSEF1as cassette (B, D). PCR amplification was done with primers for hpt gene(A, B) and for CsSEF1 gene (C, D). DNA was also isolated from plants of two mutant lines (E); one with the T-DNAinsertion in At2g19810 locus (lines: 1–10) and second in At5g07500 (lines 12–21). DNA isolated from wild type plants(E, lines 11, 22). PCR amplification for confirmation of homozygous state of mutants was done with gene specific primerpair for At2g19810 locus (lines: 1–10) and gene specific primer pair for At5g07500 (lines 12–21) and third primer specificfor T-DNA sequence, the same for both loci. CN-the negative control, DNA from non-transformed plants; CP-positivecontrol, plasmid DNA (pCAMBIA1380+35S). M1, M2–DNA molecular markers, #SMO331 and #SMO311 (MBI Fermentas),respectively.

CsSEF1 gene encoding putative CCCH-type zinc finger protein 319

fingers being involved in DNA binding, but it wasshown that this type of zinc finger binds RNA(namely AGAMOUS RNA fragments) or is associatedwith RNA metabolism (Cheng et al., 2003).

The main criterion of CsSEF1 isolation wasapproximately nine-fold increase of the transcriptaccumulation after induction of somatic embryo-genesis in cucumber ECS. Interestingly, levels ofaccumulation of CsSEF1 transcripts in ZEs (RNAisolated from mature stage ZEs) were at a lowerlevel compared to ECS. This probably reflects thedifferences between somatic and zygotic embry-ogenesis, beginning with the fact that the acquisi-tion of competence for embryogenesis is stimulatedby different signals (Dodeman et al., 1997). Theconditions in which the zygotic and somaticembryos develop are also quite different (embryosac or in vitro culture). Therefore, the genes

participating in both somatic and zygotic embry-ogenesis can be activated or inhibited in differentways. The observed difference in transcript accu-mulation between SEs and ZEs could also resultfrom the difference in developmental stagesbecause most of cucumber somatic embryos tookto RNA isolation represented late-heart and torpe-do stages, whereas ZEs had developed cotyledonsand came from developed seeds. Moreover, in thecucumber, in contrast to carrot or alfalfa, forexample, there are significant differences in theanatomical structure between both kinds of em-bryos starting from late-heart stage (Tarkowskaet al., 1994). An increased level of expression ofthe CsSEF1 gene was also found in the leaves andflowers. This may indicate that the gene alsoparticipates in later developmental stages afterembryogenesis. In fact, there are only a few

ARTICLE IN PRESS

Figure 7. Phenotypes of 7-d-old seedlings of Arabidopsis thaliana T1 plants transformed with silencing construct (A-E).Seedling with single cotyledon and inhibited root growth (A), two cotyledons and reduced root growth (B), singlecotyledon and reduced root growth (C), two cotyledons and normal root growth (D) can be distinguished and wild-type,control plant (E). Scale bar-3mm. Siliques collected from wild-type, control plant (F) and T1 CsSEF1 antisense plants(G–N). Wild-type, control plant (F), normal (G), broader (H, I), short and flattened (J) and strongly deformed siliques(N) were observed. In opened siliques, seeds arrested in different stages were found (K–M). The arrows indicate severalaborted seeds. Scale bars-1mm.

A. Grabowska et al.320

ARTICLE IN PRESS

CsSEF1 gene encoding putative CCCH-type zinc finger protein 321

described genes in which activity is limited to onlythe process of embryogenesis. These include, forexample, LEC1 in Arabidopsis thaliana (Meinke,1992; Harada, 2001) and SERK in carrot (Schmidtet al., 1997). In the case of SERK, however, eventhough its expression in carrot is limited only toembryogenic tissues, the expression of its ortholo-gues in other species (Arabidopsis thaliana andMedicago truncatula) was detected in additionaltissues (Hecht et al., 2001; Nolan et al., 2003).

The location of CsSEF1 transcripts in somaticembryos was limited to cotyledon/leaf primordiaand procambium. However, Li and Thomas (1998)observed the expression of gene PEI1 from theglobular to the late cotyledon stages equallydistributed throughout the zygotic embryo. Theexpression patterns of genes PEI1 and CsSEF1therefore do not overlap, and so their range ofactivities is most likely different.

One of the many methods for studying thefunction of a gene is comparative analysis of plantphenotypes obtained after introducing gene con-structs, resulting in the silencing of the gene or itsoverexpression. It is common knowledge thatindependent transformants carrying such con-structs can differ considerably with respect todegree of silencing or overexpression, and this candepend on both the number of copies of thetransgene, on its integration site and on manyother factors (Filipecki and Malepszy, 2006).

In order to carry out a preliminary analysis ofgene functions, constructs for gene silencing andoverexpression were introduced into the Arabidop-sis genome. It was assumed that the level ofsimilarity between cucumber gene and its homo-logues in Arabidopsis is sufficient for obtainingapproximate information. This strategy also al-lowed the avoidance of the arduous cucumbertransformation procedure, which is time consumingand induces a high degree of somaclonal variation.PCR analysis demonstrated the correctness of theintegration of the transgene in most transgenicArabidopsis plants obtained in this study.

In our gene silencing experiment, the lack ofcotyledon or deformation of siliques in Arabidopsisplants with 35S::CsSEF1as cassette indicates aber-rations in the expression of gene/genes involved inthe formation of appropriate primordia. Indeed, incucumber CsSEF1 transcripts are present in cotyle-don/leaf primordia and in female flower, where thefourth whorl is responsible for the gynoeciumformation. In addition to the deformation of siliqueas such, an inhibition of the development of seedsinside them at varied stages was observed. Li andThomas (1998), after the transformation of Arabi-dopsis plants with a coding sequence of PEI1 gene in

antisense orientation, observed aberrations in thedevelopment of embryos, indicating a key role ofthis protein in the development of the embryosfrom the globular stage to the heart stage.Additionally, Li and Thomas (1998) observed mor-phological changes in the seedlings. They obtainedseedlings with one, two and multiple leaf primor-dia, and some of these developed regular leaves.The authors suggest that PEI1 may also be requiredfor the normal development of cotyledons that areformed already at the heart stage as well as beinginvolved in the proper formation of shoot apicalmeristem.

Silencing of a gene with the use of an antisense isfrequently problematic. It is difficult to showwhether only one target gene is silenced. Thissituation resembles the recently discussed cases ofa non-specific effect (‘off-target effect’) that canbe induced by siRNA (Jackson and Linsley, 2004).Based on a comparison between the nucleotidesequence of CsSEF1 and the Arabidopsis genome,the introduction of 35S::CsSEF1as cassette couldresult in the silencing of the most similargenes: At2g25900 and/or At2g19810. Caution inthe interpretation of phenotypes obtained as aresult of RNAi experiments postulated by Jacksonand Linsley (2004) concerns, above all, multigenefamilies. Li and Thomas (1998) achieved silencingby introducing the entire PEI1-coding sequence inthe antisense orientation and observed arresteddevelopment of Arabidopsis thaliana embryosat the globular stage. In this case, the silencingof other homologues that could have yieldedthe observed effect cannot be excluded. Such anidea is confirmed by our preliminary macroscopicobservations of the homozygous insertion mutantpei1 (SALK_108149), in which insertion of T-DNAfollowing codon 57 leads to the formation of a 58amino acid oligopeptide with no zinc finger motifsand in which no developmental disorders weredetected (data not shown). The seeds obtainedfrom mentioned mutant germinated, giving rise tonormal seedlings. It seems that the lack offunctional protein PEI1 in the mutant has no lethaleffect on the development or that the function ofthe gene has been taken over by its homologue. Thesilenced phenotype described by Li and Thomas(1998) could have been enhanced by thenon-specific silencing of other close homologuesof the PEI1 gene.

In the case of our overexpression experiment,the lack of phenotypic changes is difficult toexplain. It might be that cucumber protein lacksthe specificity for interaction with other proteins ornucleic acids from Arabidopsis. The putativeCsSEF1 protein has motifs that are characteristic

ARTICLE IN PRESS

A. Grabowska et al.322

for RNA-binding proteins and it may thus beinvolved in the post-transcriptional regulation ofgene expression influencing mRNA stability. Suchspecific interactions with sequences within 30-UTRof target mRNA were shown for mammalian CCCHzinc finger proteins TTP (tristetraprolin) andCaenorhabditis elegans MEX-5 (Brewer et al.,2004; Pagano et al., 2007). The potential role ofCsSEF1 in binding the specific mRNA regionsrequires confirmation by experimental approachesin the future.

Acknowledgments

This work was supported by Polish Ministry ofScience and Higher Education (grant no. PBZ/KBN/029/P06/2000).

References

Alonso JM, Stepanova AN, Leisse TJ, Kim CJ, Chen H,Shinn P, et al. Genome-wide insertional mutagenesisof Arabidopsis thaliana. Science 2003;301:653–7.

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basiclocal alignment search tool. J Mol Biol 1990;215:403–10.

Boutilier K, Offringa R, Sharma VK, Kieft H, Ouellet T,Zhang L, et al. Ectopic expression of BABY BOOMtriggers a conversion from vegetative to embryonicgrowth. Plant Cell 2002;14:1737–49.

Braybrook SA, Stone SL, Park S, Bui AQ, Le BH, Fischer RL,et al. Genes directly regulated by LEAFY COTYLEDON2provide insight into the control of embryo maturationand somatic embryogenesis. Proc Natl Acad Sci2006;103:3468–73.

Brewer BY, Malicka J, Blackshear PJ, Wilson GM. RNAsequence elements required for high affinity bindingby the zinc finger domain of tristetraprolin. J BiolChem 2004;279:27870–7.

Burza W., Malepszy S., Cytokinin control of cucumber(Cucumis sativus L.) somatic embryogenesis. IX Inter-national congress on plant tissue and cell culture.Jerusalem, Israel, vol. 68, 1998, June 14–19.

Burza W, Zuzga S, Yin Z, Malepszy S. Cucumber (Cucumissativus L.). In: Wang K, editor. Methods in molecularbiology. Agrobacterium protocols. Totowa, NJ: Huma-na Press Inc.; 2006. p. 427–38.

Carballo E, Lai WS, Blackshear PJ. Feedback inhibition ofmacrophage tumor necrosis factor-a production bytristetraprolin. Science 1998;281:1001–5.

Cheng Y, Kato N, Wang W, Li J, Chen X. Two RNA bindingproteins, HEN4 and HUA1, act in the processing ofAGAMOUS pre-mRNA in Arabidopsis thaliana. Dev Cell2003;4:53–66.

Clough SJ, Bent AF. Floral dip: a simplified method forAgrobacterium-mediated transformation of Arabidop-sis thaliana. Plant J 1998;16:735–43.

Diatchenko L, Lukyanov S, Lau YF, Siebert PD. Suppres-sion subtractive hybridization: a versatile method foridentifying differentially expressed genes. MethodsEnzymol 1999;303:349–80.

Dodeman VL, Ducreux G, Kreis M. Zygotic embryogenesisversus somatic embryogenesis. J Exp Bot 1997;313:1493–509.

Filipecki M, Malepszy S. Unintended consequences ofplant transformation: a molecular insight. J ApplGenet 2006;47:277–86.

Filipecki MK, Sommer H, Malepszy S. The MADS-box geneCUS1 is expressed during cucumber somatic embry-ogenesis. Plant Sci 1997;125:63–74.

Fischerova L, Fischer L, Vondrakova Z, Vagner M.Expression of the gene encoding transcription factorPaVP1 differs in Picea abies embryogenic linesdepending on their ability to develop somaticembryos. Plant Cell Rep 2008;27:435–41.

Formentin E, Naso A, Varotto S, Picco C, Gambale F, LoSchiavo F. KDC2, a functional homomeric potassiumchannel expressed during carrot embryogenesis. FEBSLett 2006;580:5009–15.

Hall TM. Multiple modes of RNA recognition by zinc fingerproteins. Curr Opin Struct Biol 2005;15:367–73.

Harada JJ. Role of Arabidopsis LEAFY COTYLEDON genesin seed development. J. Plant Physiol 2001;158:405–9.

Hecht V, Vielle-Calzada JP, Hartog MV, Schmidt EDL,Boutilier K, Grossniklaus U, et al. The ArabidopsisSOMATIC EMBRYOGENESIS RECEPTOR KINASE 1 gene isexpressed in developing ovules and embryos andenhances embryogenic competence in culture. PlantPhysiol 2001;127:803–16.

Hiwatashi Y, Fukuda H. Tissue-specific localization ofmRNA for carrot homeobox genes, CHBs, in carrotsomatic embryos. Plant Cell Physiol 2000;41:639–43.

Jackson AL, Linsley PS. Noise amidst the silence: off-target effects of siRNAs? Trends Genetics 2004;20:521–4.

Kaplan DR, Cooke TJ. Fundamental concepts in theembryogenesis of dicotyledons: a morphological inter-pretation of embryo mutants. Plant Cell 1997;9:1903–19.

Kawahara R, Komamine A, Fukuda H. Isolation andcharacterization of homeobox-containing genes ofcarrot. Plant Mol Biol 1995;27:155–64.

Koetsier PA, Schorr J, Doerfler W. A rapid optimizedprotocol for downward alkaline Southern blotting ofDNA. Biotechniques 1993;15:260–2.

Kwong RW, Bui AQ, Lee H, Kwong LW, Fisher RL, GolbergRB, et al. LEAFY COTYLEDON1-LIKE defines a class ofregulators essential for embryo development. PlantCell 2003;15:5–18.

Laity JH, Lee BM, Wright PE. Zinc finger proteins: newinsights into structural and functional diversity. CurrOpin Struct Biol 2001;11:39–46.

ARTICLE IN PRESS

CsSEF1 gene encoding putative CCCH-type zinc finger protein 323

Li J, Jia D, Chen X. HUA1, a regulator of stamen andcarpel identities in Arabidopsis, codes for nuclear RNAbinding protein. Plant Cell 2001;13:2269–81.

Li Z, Thomas TL. PEI1, an embryo-specific zinc fingerprotein gene required for heart-stage embryo forma-tion in Arabidopsis. Plant Cell 1998;10:383–98.

Linkiewicz A, Filipecki M, Tomczak A, Grabowska A,Malepszy S. The cloning of sequences differentiallytranscribed during the induction of somatic embry-ogenesis in cucumber (Cucumis sativus L.). Cell MolBiol Lett 2004;9:795–804.

Long JA, Moan EI, Medford JI, Barton MK. A member ofthe KNOTTED class of homeodomain proteins encodedby the SHOOT MERISTEMLESS gene of Arabidopsisthaliana. Nature 1996;379:66–9.

Malinowski R, Filipecki M, Tagashira N, Wisniewska A, GajP, Plader W, et al. Xyloglucan endotransglucosylase/hydrolase genes in cucumber (Cucumis sativus) –

differential expression during somatic embryogenesis.Physiol Plant 2004;120:678–85.

Mantiri FR, Kurdyukov S, Lohar DP, Sharopova N, SaeedNA, Wang XD, et al. The transcription factor MtSERF1of the ERF subfamily identified by transcriptionalprofiling is required for somatic embryogenesis in-duced by auxin plus cytokinin in Medicago truncatula.Plant Physiol 2008;146:1622–36.

Meinke DW. A homeotic mutant of Arabidopsis thalianawith leafy cotyledons. Science 1992;258:1647–50.

Michaels SD, John MC, Amasino RM. Removal of poly-saccharides from plant DNA by ethanol precipitation.Biotechniques 1994;17:274–6.

Mordhorst AP, Toonen MAJ, de Vries SC. Plant embry-ogenesis. Critical Reviews in Plant Science 1997;16:535–76.

Murashige T, Skoog F. A revised medium for rapid growthand bioassays with tobacco tissue culture. PlantPhysiol 1962;15:473–97.

Nolan KE, Irwanto RR, Rose RJ. Auxin up-regulatesMtSERK1 expression in both Medicago truncatularoot-forming and embryonic cultures. Plant Physiol2003;133:218–30.

Pagano JM, Farley BM, McCoig LM, Ryder SP. Molecularbasis of RNA recognition by the embryonic polaritydeterminant MEX-5. J Biol Chem 2007;23:8883–94.

Sambrook J, Fritsch EF, Maniatis T. Molecular cloning, Alaboratory manual. Cold Spring Harbor: LaboratoryPress; 1989.

Sato S, Toya T, Kawahara R, Whittier RF, Fukuda H,Komamine A. Isolation of a carrot gene expressedspecifically during early-stage somatic embryogenesis.Plant Mol Biol 1995;28:39–46.

Schmidt EDL, Guzzo F, Toonen MAJ, de Vries SC. Aleucine-rich repeat containing receptor-like kinasemarks somatic plant cells competent to form embryos.Development 1997;124:2049–62.

Shah K, Schmidt ED, Vlak JM, de Vries SC. Expression ofthe Daucus carota somatic embryogenesis receptorkinase (DcSERK) protein in insect cells. Biochimie2001;83:415–21.

Stone SL, Kwong LW, Yee KM, Pelletier J, Lepiniec L,Fischer RL, et al. LEAFY COTYLEDON2 encodes a B3domain transcription factor that induces embryodevelopment. Proc Natl Acad Sci 2001;98:11806–11.

Stone SL, Braybrook SA, Paula SL, Kwong LW, Meuser J,Pelletier J, et al. Arabidopsis LEAFY COTYLEDON2induces maturation traits and auxin activity: Implica-tions for somatic embryogenesis. Proc Natl Acad Sci2008;105:3151–6.

Tarkowska JA, Brzostecka D, Burza W, Malepszy S.Cytohistological analysis of somatic embryogenesis incucumber (Cucumis sativus L.). I. Comparison of cellsuspension containing and lacking natural fluores-cence with in vivo developing embryos. Acta Soc BotPol 1994;63:259–63.

Tenlen JR, Schisa JA, Diede SJ Page BD. Reduced dosageof pos-1 suppresses mex mutants and reveals complexinteractions among CCCH zinc-finger proteins duringCaenorhabditis elegans embryogenesis. Genetics 2006;174:1933–45.

Thakare D, Tang W, Hill K, Perry SE. The MADS-domaintranscriptional regulator AGAMOUS-LIKE15 promotessomatic embryo development in Arabidopsis andsoybean. Plant Physiol 2008;146:1663–72.

Wang D, Guo Y, Wu C, Yang G, Li Y, Zheng C. Genome-wideanalysis of CCCH zinc finger family in Arabidopsis andrice. BMC Genomics 2008;9:44.

Wilson ZA. Arabidopsis, A practical approach. Oxford:Oxford University Press; 2000.

Wroblewski T, Filipecki M, Malepszy S. Factors influencingcucumber (Cucumis sativus L.) somatic embryogen-esis. I. The crucial role of pH and nitrogen insuspension culture. Acta Soc Bot Pol 1995;64:223–31.

Wurtele ES, Wang H, Durgerian S, Nikolau BJ, Ulrich TH.Characterization of a gene that is expressed early insomatic embryogenesis of Daucus carota. Plant Physiol1993;102:303–12.

Xu R, Li QQ. A RING-H2 zinc-finger protein gene RIE1 isessential for seed development in Arabidopsis. PlantMolecular Biology 2003;53:37–50.

Zimmerman JL. Somatic embryogenesis: a model forearly development in higher plants. Plant Cell 1993;5:1411–23.