Transcriptome analysis during somatic embryogenesisof the tropical monocot Elaeis guineensis: evidence for conservedgene functions in early development

Hsiang-Chun Lin Æ Fabienne Morcillo ÆStephane Dussert Æ Christine Tranchant-Dubreuil ÆJames W. Tregear Æ Timothy John Tranbarger

Received: 26 September 2008 / Accepted: 21 January 2009 / Published online: 7 February 2009! Springer Science+Business Media B.V. 2009

Abstract With the aim of understanding the molecularmechanisms underlying somatic embryogenesis (SE) in oil

palm, we examined transcriptome changes that occur

when embryogenic suspension cells are initiated to developsomatic embryos. Two reciprocal suppression subtractive

hybridization (SSH) libraries were constructed from oil

palm embryogenic cell suspensions: one in which embryodevelopment was blocked by the presence of the synthetic

auxin analogue 2,4-dichlorophenoxyacetic acid (2,4-D) in

the medium (proliferation library); and another in which

cells were stimulated to form embryos by the removal of2,4-D from the medium (initiation library). A total of 1867

Expressed Sequence Tags (ESTs) consisting of 1567

potential unigenes were assembled from the two libraries.Functional annotation indicated that 928 of the ESTs cor-

respond to proteins that have either no similarity to

sequences in public databases or are of unknown function.Gene Ontology (GO) terms assigned to the two EST pop-

ulations give clues to the underlying molecular functions,

biological processes and cellular components involvedin the initiation of embryo development. Macroarrays

were used for transcript profiling the ESTs during SE.

Hierarchical cluster analysis of differential transcriptaccumulation revealed 4 distinct profiles containing a total

of 192 statistically significant developmentally regulated

transcripts. Similarities and differences between the globalresults obtained with in vitro systems from dicots, mono-

cots and gymnosperms will be discussed.

Keywords Somatic embryogenesis ! Oil palm !Auxin ! Transcript profiling

Abbreviations2,4-D 2,4-Dichlorophenoxyacetic acid

ARF ADP-ribosylation factor

EST Expressed sequence tagGO Gene Ontology

GST Glutathione S-transferases

HCA Hierarchical cluster analysisPR Pathogenesis-related

STM-like Shootmeristemless-like

SEm Somatic embryoSE Somatic embryogenesis

SSH Suppression subtractive hybridization

Hsiang-Chun Lin and Fabienne Morcillo contributed equally to thiswork and a portion was originally presented at the 2006 Congress ofthe Federation of European Societies of Plant Biology Lyon, France.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s11103-009-9464-3) contains supplementarymaterial, which is available to authorized users.

H.-C. Lin ! J. W. Tregear ! T. J. Tranbarger (&)IRD, UMR DIAPC, IRD/CIRAD Palm Development Group,911 Avenue Agropolis, BP 64501, 34394 Montpellier Cedex 5,Francee-mail: timothy.tranbarger@ird.fr

F. MorcilloCIRAD, UMR DIAPC, IRD/CIRAD Palm Development Group,911 Avenue Agropolis, BP 64501,34394 Montpellier Cedex 5, France

S. Dussert ! C. Tranchant-DubreuilIRD, UMR DIAPC, 911 Avenue Agropolis, 34394 MontpellierCedex 5, France

Present Address:H.-C. LinETH Zurich, Institute of Plant Science, LFW D 36.1,Universitatstrasse 2, 8092 Zurich, Switzerland

123

Plant Mol Biol (2009) 70:173–192

DOI 10.1007/s11103-009-9464-3

Introduction

Oil palm (Elaeis guineensis Jacq) belongs to the Arecaceae

family and is a perennial monocotyledonous plant that

recently became the world’s number one source of ediblevegetable oil. While the worldwide demand for palm oil

increases, new elite genotypes from breeding programs

provide a yield increase of only 1% per year and take over15 years to develop (Corley and Tinker 2003). To decrease

the time necessary to disseminate elite genotypes obtained

in breeding programs, multiplication by in vitro vegetativepropagation may be used. Because of its biological char-

acteristics, including the existence of a single shoot apical

meristem, the only viable method to multiply oil palm is byin vitro somatic embryogenesis (SE). SE can be performed

either on solid or in liquid medium while the liquid meth-

odology has the most promise to provide a cost effectivemethod to supply elite oil palm ramets on a commercial

scale (Corley and Tinker 2003; Duval et al. 1995). The

procedure of liquid SE of oil palm can be divided into 5 keysteps: (1) dedifferentiation via callogenesis originating from

selected explant material, (2) proliferation of embryogenic

cells in liquid culture, (3) synchronized redifferentiation viathe initiation of embryogenesis, (4) first signs of polarized

growth observed after suspension cells are transferred to

solid medium and (5) embryo morphogenesis. One limita-tion in this procedure is the production of well-polarized

somatic embryos (SEms; Aberlenc-Bertossi et al. 1999). SE

can also be used as a model system to study the molecular,physiological and cytological mechanisms underlying

embryo development, particularly early regulatory and

morphological events that are too difficult to access withzygotic embryos (Zimmerman 1993). In addition, SE is an

excellent example of plant totipotency that allows the

regeneration of whole plants from embryogenic cells (Feheret al. 2003; Verdeil et al. 2007). While the molecular basis

of SE even in model plant species such as Arabidopsis andrice is not well understood, molecular studies on SE of key

agronomic and forestry species need to be developed for

biotechnological applications. Indeed, studies of SE havebeen performed for a range of economically important

species. Transcriptomic and proteomic approaches have

been used successfully to catalogue genes and proteins thatare expressed during SE in economically important dicot-

yledonous, monocotyledonous and gymnosperm species

such as chicory, cotton, rapeseed, soybean, Medicagotruncatula, maize, rice, wheat, potato, Norway Spruce,

White Spruce and loblolly pine (Cairney et al. 2006; Che

et al. 2006; Imin et al. 2005; Joosen et al. 2007; Legrandet al. 2007; Lippert et al. 2005; Sharma et al. 2008; Singla

et al. 2007; Stasolla et al. 2003, 2004; Su et al. 2007; Thi-

baud-Nissen et al. 2003; van Zyl et al. 2003; Zeng et al.2006). These studies provide clues to the physiological and

developmental processes underlying SE and form a basis to

compare the SE pathways from both model and non-modeleconomically important species (Cairney and Pullman

2007; Cairney et al. 2006). In all cases, SE is based on a

somatic cell’s capacity to initiate a new morphologicalpathway towards embryo development and depends on a

cell’s plasticity to re-program gene expression and cellular

processes that lead to SEm development (Feher et al. 2003).In a recent survey of 124 SE protocols, it was shown that

the synthetic auxin analogue 2,4-dichlorophenoxyaceticacid (2,4-D) is the most commonly used treatment to induce

somatic cells to acquire embryogenic capacity (Gaj 2004).

With some species such as oil palm, auxin must subse-quently be removed from the medium in order to initiate

embryogenic cells to redifferentiate and form an embryo

(Gray 2005). The possibility to control the initiation ofembryo development by removing auxin from the culture

medium provides a simple model system to study gene

expression and regulation during the early stages of SEmdevelopment. In the present study, a transcriptome

approach was undertaken to analyze the gene expression

changes that underlie the initiation and development ofoil palm SEms. First, a PCR-based cDNA subtraction

approach was used to obtain two subtracted cDNA libraries

and expressed sequence tags (ESTs) from the early stagesof SE. Secondly, the ESTs were assigned to functional

categories and Gene Ontology terms. Finally, by means of

cDNA macroarray based expression analyses, the quanti-fication of transcript profiles during SE was conducted to

identify groups of genes with coordinated expression.

Materials and methods

Plant material production, histology and RNA

preparation

An oil palm embryogenic suspension cell line (LMC 221)

was obtained using a protocol previously described (de

Touchet et al. 1991) and proliferated on medium contain-ing 3.4 lM 2,4-D without activated charcoal. To initiate

embryo development, 500 mg of suspension cells at the

end of the 30-day proliferation cycle were transferred into50 ml of the liquid medium without 2,4-D in 250 ml

Erlenmeyer flasks. The cell suspensions were kept at 27"Cin the culture room on an orbital shaker set at 90 rota-tions min-1 for 30 days. Cell culture samples were

collected at the end of the 30-day proliferation cycle (day

0) and at 1, 2, 4, 8 and 16 days after 2,4-D was removedfrom the medium. Two independent replicates of each time

point were collected. A sample of suspension cells cultured

in medium without 2,4-D for 30 days was sieved (mesh size1 mm) and 40 mg of the resulting suspension cells were

174 Plant Mol Biol (2009) 70:173–192

123

plated on solid agar plates containing the basal medium

supplemented with 30 g l-1 sucrose, 0.5 g l-1 casein and8 g l-1 agar without hormone to facilitate SEm develop-

ment. Twenty days after plating, individual developing

SEms were collected that are characteristically oblong inshape approximately 2–3 mm in diameter with a white

epidermis (Fig. 1d). Preparation of samples for histological

analysis was preformed as previously described (Buffard-Morel et al. 1992). Each section was double stained with

periodic acid-Schiff (PAS) and naphthol blue–black(NBB). For transcript profiling analysis during the SE time

course, total RNA was extracted from 0.5 g of each sample

with TRIReagent (Sigma).

Suppression subtractive hybridization cDNA library

construction

A PCR based suppression subtractive hybridization (SSH)

kit was used to construct two reciprocal cDNA libraries,each from 1 lg total RNA according to the manufacturer’s

instructions (Clontech, Palo Alto, CA, USA). The SSH

cDNA libraries originate from cell suspensions culturedunder the two following contrasting conditions; cells cul-

tured for 30 days in liquid proliferation medium containing

2,4-D (proliferation conditions) and cells cultured for16 days in liquid pretreatment medium in the absence of

2,4-D (initiation conditions). The proliferation conditions

correspond to the treatment during which cells are multi-plied prior to the initiation of SEm development while the

initiation treatment corresponds to 16 days of the pre-

treatment necessary to initiate SEm development. One SSHcDNA library (proliferation library) was constructed using

cDNA originating from cells cultivated under proliferation

conditions as tester cDNA while the cDNA populationoriginating from cells grown under initiation conditions

was used as the driver cDNA. A second SSH cDNA librarywas constructed using cDNA originating from cells

cultured under initiation conditions while the cDNA pop-

ulation originating from cells grown under proliferationconditions was used as the driver cDNA. The cDNAs

resulting from the two subtracted cDNA populations were

cloned into the pGEM-T Easy vector (Promega, MadisonWI, USA).

EST generation, analysis and annotation

The ESTs originating from the two SSH cDNA librarieswere generated using standard high throughput sequencing

by GATC Biotech AG, Germany. The DNA templates

were subjected to 50 single pass automated sequencingusing the ABI3730 (Perkin Elmer, Foster City, CA, USA)

Fig. 1 Histological analysis ofoil palm embryogenicsuspension cells cultured inproliferation medium with 2,4-Dfor 30 days (a) or in initiationmedium without 2,4-D for16 days (b) and a longitudinalsection through a developingindividualized SEm isolatedfrom initiated cell suspensionssieved and plated onto solidagar plates without 2,4-D for20 days (c). Sections weredouble stained with periodicacid-Schiff (PAS) and naphtholblue–black (NBB). After20 days of growth on plates,individual SEms arise andbecome visually distinguishableby their white oblong shapes(d). E, Epidermis; EC,Embryogenic Cell; FL,Fragmentation Lines; M,Meristematic tissue; PS,Provascular Strands; SEm,Somatic Embryos. Scale bars100 lm for a and b; 400 lm forc; 2 mm for d

Plant Mol Biol (2009) 70:173–192 175

123

with M13 Reverse primer. The sequences were then sub-

jected to an automated procedure to verify cleanse, storeand analyze sequences (Jouannic et al. 2005). The average

read length for the two EST populations was approximately

550 bp. The automated analyses allowed the identificationof potential unigenes (contigs plus singletons) through

simultaneous cluster analysis. Finally, to assign putative

functions to the ESTs, BLASTX (http://www.ncbi.nlm.nih.gov/BLAST/) was used to compare sequences with the

GenBank non-redundant protein sequence database aspreviously described (Altschul et al. 1997). EST sequences

were manually assigned to functional categories based on a

previous catalogue system (Bevan et al. 1998). In addition,Gene Ontology (GO, http://geneontology.org/) based

annotation was performed using the GOblet website (http://

goblet.molgen.mpg.de) to assign GO molecular function,biological process and cellular component terms for com-

parison between the two EST populations (Altschul et al.

1997; Groth et al. 2004). The unigene sequences wereanalyzed using BLASTX against a GO-based plant uniprot

database with an E-value cutoff of 10-10.

cDNA array: construction, hybridization and signal

normalization

The transcript profiling approach used in this study is based

on cDNA macroarray hybridization (Alba et al. 2004;

Freeman et al. 2000). ESTs from the two SSH librarieswere amplified by PCR (Polymerase Chain Reaction) to

spot onto nylon membranes. On a nylon membrane matrix

(HybondTM-N?, Amersham Biosciences), the amplifiedEST PCR products were arrayed by robotics and constitute

the probe. In the present study, the nested primers NP1 and

NP2R (Clontech) were used to amplify from the cDNAclones the PCR products that were printed as single spots

onto the nylon membrane. These probes were hybridized

with cDNA populations made from RNA samples isolatedfrom the SE time course described above. The hybridiza-

tion signals were detected, quantified, normalized and

compared to obtain evidence of differential transcriptaccumulation using ArrayVisionTM (GE Healthcare).

Radioisotope [33P] dCTP labeled cDNA populations were

generated from the two experimental total RNA timecourses (0 days, 1 day, 2 days, 4 days, 8 days, 16 days and

SEm) by reverse-transcription using the kit LabelStar

(Qiagen). 15 lg of denatured total RNA from each samplewere mixed with 5 ll of 109 buffer RT, 5 ll of dNTP Mix

(dATP, dGTP, dTTP [20 mM] and dCTP [0.2 mM]), 5 llof 33P-labeled dCTP [3,000 Ci/mmol], 5 ll of Oligo-dTprimer [20 lM], 0.5 ll of RNase inhibitor [40 units/ll],7 ll of RNase-free water and 2.5 ll of LabelStar ReverseTranscriptase in a final volume of 50 ll for the labelingreaction. This mixture was incubated at 37"C for 120 min.

The membranes were pre-hybridized with 15 ml of

hybridization solution (3.75 ml of 209 SSC, 1.5 ml of 509Denhardt’s solution, 0.45 ml of 20% SDS, 60 ll of dena-tured salmon sperm DNA and 9.24 ml of RNase-free

distilled water) at 65"C for 3 h. About 25 ll of denatured(95"C for 5 min) probe mixture were applied onto the pre-

hybridized membranes and hybridized overnight at 65"C.After hybridization, the membranes were washed twice in19 SSC and 0.1% SDS and twice in 0.1% SDS and 0.19

SSC at 65"C for 10 min. After washing, the wet mem-branes were covered with Saran film and exposed to Kodak

Storage Phosphor Screen (Amersham Biosciences) for

5 days at room temperature before development. Hybrid-ization was performed with four membranes for each

labeled cDNA population and membranes were used only

for a single hybridization, so that there were four replicatesof each cDNA spot per hybridization. There were four

hybridizations performed with each of the seven experi-

mental samples (0 days, 1 day, 2 days, 4 days, 8 days,16 days and SEm) in duplicate, which represents two

independent biological experiments for each time course.

ArrayVisionTM analysis and statistical analyses

Hybridization signals were quantified with ArrayVisionTM

software. Background pixel values (intensities) were

determined locally by measuring blank spots around sig-

nals. Background intensities for each spot were subtractedto give corrected spot intensities that were normalized by

dividing by the average corrected value of all the spots on

the membrane. For each of the four replicate membranes,the corrected values of each spot were adjusted so that the

value of the spot was considered zero if its intensity was

lower than background. The three closest normalizedvalues from four replicate spots in each hybridization

experiment from the different replicates were selected for

subsequent statistical analyses.Gene expression datasets obtained from ArrayVisionTM

normalization of macroarray hybridizations were exported

into Statistica 7.1 software (StatSoft). First, hierarchicalcluster analysis (HCA, Ward’s pair-wise grouping method,

Euclidian distance) was performed on the gene expression

dataset to determine similarities between the 7 time pointsstudied (0 days, 1 day, 2 days, 4 days, 8 days, 16 days and

SEm) within each experiment. The Mantel test was then

used to assess the correlation between the two inter-stagedistance matrices to determine the reproducibility between

the two independent experiments. Second, HCA was

performed on the gene expression dataset to cluster simi-larities in gene expression profiles. Finally, statistical

significance of apparent changes in gene expression at

selected contrasting stages of development was calculatedusing ANOVA (P-value\ 0.05) only for the transcripts

176 Plant Mol Biol (2009) 70:173–192

123

found to have the same profile from both biological

replications.

Results

Histological analysis of undifferentiated suspension

cells used for SSH library construction

Suspension cells used for the following studies were rou-tinely examined by histological analysis (Fig. 1a, b).

Observations indicated that cell suspensions grown in the

presence of 2,4-D (proliferation cells) or after its removaland grown in the absence of 2,4-D for two weeks (initiation

cells) were comprised mainly of groups of cell clusters that

consisted of at least three major cell types (Fig. 1a, b). Theembryogenic cell types were characterized by their small

size, large central nucleus and dense cytoplasm stained

blue by the NBB, small vacuole (unstained), and thin cellwalls stained red by PAS. Embryogenic cells were found in

irregularly concentric clusters surrounded by cells that

make up so-called fragmentation lines described previously(de Touchet et al. 1991). This second cell type is elongated

with a smaller non-centralized nucleus and a less dense

staining cytoplasm. A third cell type always foundperipheral to the cell clusters appears to be degenerating

and have very little cellular contents. There were no

apparent qualitative cytological differences between cellscultured with or without 2,4-D (Fig. 1). Only cell suspen-

sions grown in the absence of 2,4-D for at least 16 days

results in the development of SEms identified by theirdifferentiated epidermis and the presence of provascular

strands (Fig. 1c, d).

Assembly and functional annotation of ESTs derived

from oil palm suspension culture SSH cDNA libraries

Single pass sequencing and processing resulted in the

production of 1867 high quality ESTs, 918 and 949 from

the proliferation and initiation libraries respectively (Gen-bank accession numbers GH635901–GH637767). Intra-

library cluster analysis resulted in 814 (89%, proliferation

library) and 839 (88%, initiation library) assembled unig-enes, while inter-library cluster analysis resulted in the

assembly of 1567 (84%) unigenes consisting of 1410 sin-

gletons and 157 contigs of two or more ESTs. To gaininsight into the functional diversity and differences

between the two libraries, the two sets of ESTs were ana-

lyzed by BLASTX, which indicated that approximately35% of the ESTs shared no significant similarity with

sequences from public databases while 14.5% encoded

polypeptide sequences matching with proteins of unknownfunction from Arabidopsis, rice, maize, or other species

(Supplementary Fig. 1). Approximately 50% of the ESTs

from the two libraries were assigned to 13 primary func-tional categories as described previously (Bevan et al.

1998). From the two libraries combined, the functional

categories most represented were Metabolism (132 ESTs),Transcription and Post-transcription (127 ESTs), Protein

Destination and Storage (123 ESTs), and Signal Trans-

duction (102 ESTs). Since the BLASTX analysis of theESTs from these libraries was performed, there were

32,136 oil palm ESTs made publicly available to NCBI(Ho et al. 2007; Low et al. 2008). From the 1567 unigenes

reported here, 980 shared no sequence similarity (E-value

10-10) to the 36,308 oil palm ESTs currently available (01/09/2008) at NCBI (data not shown). There were only 215

matches that consisted of alignments of more than 90%

identity.

Assigning Gene Ontologies to the SSH derived ESTs

Gene Ontology (GO) is an annotation framework that uses

a controlled vocabulary to assign functional attributes to

genes and gene products from any organism (Consortium2008). GO consists of three controlled ontologies divided

into molecular function, biological process and cellular

component that can allow insight into the functional dif-ferences between cDNA populations. The GO based

Goblet web-service was used to assign GO terms to the

unigene sets obtained from the two libraries by a BLASTXcomparison to the plant UNIPROT database annotated with

GO (Groth et al. 2004). The total numbers of molecular

function, biological process and cellular component GOaccessions assigned were 969, 978 and 891 for the prolif-

eration library ESTs and 828, 782 and 761 for the initiation

library unigenes respectively (Fig. 2 and SupplementaryFigs. 2 and 3). While the majority of the GO accessions

from the three ontologies were shared between the two

libraries, the comparison revealed differences in molecularfunction, biological process and cellular component

accessions assigned between the two libraries. For exam-

ple, there were 8 GO molecular functions represented by 34ESTs found only in the initiation library while there were 9

GO molecular functions represented by 26 ESTs found

only in the proliferation library (Fig. 2). For the biologicalprocess, there were 8 unique GO accessions corresponding

to 20 ESTs in the initiation library and 7 unique biological

process GO accessions corresponding to 14 ESTs in theproliferation library (Supplementary Fig. 2). For the

cellular component, there were 16 GO terms (49 ESTs)

and 13 GO terms (31 ESTs) found in either the initiation orthe proliferation library respectively (Supplementary

Fig. 3). Five of the 34 ESTs (EG-S41_010_C11, EG-

S41_006_E07, EG-S41_009_A04, EG-S41_008_H05 andEG-S41_010_A08) were assigned two GO molecular

Plant Mol Biol (2009) 70:173–192 177

123

function terms reducing the unique EST total from

the initiation library to 29 (Table 1). Likewise, 2 of the

20 ESTs (EG-S41_006_B07, EG-S41_006_E10) wereassigned two or more biological process GO terms reduc-

ing the unique EST total from the initiation library to 14(Table 2). For the cellular component, there were 4 ESTs

(EG-S41_006_B07, EG-S41_005_D09, EG-S41_003_H11,

EG-S41_004_C06) with two or more GO terms reducingthe unique EST total from the initiation library to 44

(Table 3). Finally, in account of the redundancy between

the three analyses, GO annotation allowed the identifica-

tion of a total of 74 ESTs with GO molecular function,

biological process and cellular component terms unique tothe initiation library (Tables 1, 2 and 3).

Hierarchical cluster analysis identifies reproducible

coordinated expression profiles during SE

In order to assess the reproducibility of the global

expression analysis during SE, HCA of the seven

112223456

1127

842

205

14141778

10732

11271

1105

938

1165332

1575

901117

212

22129

213

421

12399

11537322795287

475

710

65332

1686

1141424

321

4425

261222445

14

0% 20% 40% 60% 80% 100%

pattern bindingtranscriptional repressor activitytranscriptional activator activity

transposase activitytwo-component response regulator activity

receptor signaling protein activitycarboxylic acid binding

selenium bindingtranscription cofactor activity

lyase activityion transporter activity

protein transporter activityorganic acid transporter activity

enzyme inhibitor activityreceptor activity

carbohydrate transporter activityelectron transporter activity

isomerase activitycofactor bindingprotein binding

ion bindingcarrier activity

transferase activityoxidoreductase activity

nucleotide bindingamine transporter activity

hydrolase activityATPase activity, coupled to movement of substances

helicase activitytetrapyrrole binding

lipid bindingamine binding

structural constituent of cell wallchromatin binding

translation factor activity, nucleic acid bindingcarbohydrate binding

structural constituent of cytoskeletonnucleic acid binding

ligase activitystructural constituent of ribosome

small protein activating enzyme activityperoxidase activity

channel or pore class transporter activityvitamin binding

intracellular transporter activitytwo-component sensor molecule activity

transcription factor activityphosphatase regulator activity

GTPase regulator activityoxygen transporter activity

transcription initiation factor activityenzyme regulator activitymicrotubule motor activity

oxygen bindingmetal cluster binding

Distribution of Molecular Function GO Terms between LibrariesProliferationInitiation

Fig. 2 A graphicalrepresentation of the GOMolecular Function annotations(level 3) unique to and sharedbetween the two SSH libraries.The numbers on each barindicate the total unigenesassigned to a corresponding GOaccession

178 Plant Mol Biol (2009) 70:173–192

123

Table 1 ESTs with GO molecular function terms found only in the initiation library

GO molecularfunction and accession

EST IDa Description (BLASTX)* Species Accession E-value

Metal clusterbinding

GO:0051540

EG-S41_001_E01 C2 domain-containing protein-like O. sativa dbj|BAD38184.1 5E-98cEG-S41_006_D08 Nuclear protein-like A. thaliana dbj|BAD94321.1 7E-18

EG-S41_002_C09 Photosystem I assembly protein ycf3 Lactuca sativa dbj|BAE47595.1 2E-32

EG-S41_004_H09 Translational activator A. thaliana gb|AAD38254.1 3E-95

EG-S41_006_E02 Gamma-soluble NSF attachment protein A. thaliana gb|AAF01285.1| 1E-51

EG-S41_009_E08 Beta-adaptin-like protein B A. thaliana gb|AAF61672.1 2E-92

EG-S41_008_D10 Unknown protein O. sativa gb|AAN05340.1| 9E-12

EG-S41_005_D11 Pumilio-family RNA binding repeat O. sativa gb|AAP53549.1| 4E-55

EG-S41_004_B08 NADH dehydrogenase O. sativa gb|AAP68893.1| 6E-63

EG-S41_004_H02 NADH dehydrogenase O. sativa gb|AAP68893.1| 6E-63

EG-S41_004_C01 Protein phosphatase 2Aregulatory subunit A

Zea mays gb|AAY24685.1 2E-90

EG-S41_006_E03 DNA-binding protein O. sativa gb|ABA96555.1 6E-54

EG-S41_003_A03 Hypothetical protein OsJ_031095 O. sativa gb|EAZ16886.1| 2E-50

EG-S41_010_B02 ATP binding/damaged DNA binding A. thaliana ref|NP_180698.1 3E-46

Oxygen binding

GO:0019825

EG-S41_006_F04 Importin 9 O. sativa dbj|BAD69163.1 1E-53

EG-S41_010_C11 Nonsymbiotic hemoglobin Lotus corniculatus dbj|BAE46739.1 2E-54

EG-S41_006_E07 Non-symbiotic hemoglobin class 1 Malus x domestica gb|AAP57676.1 2E-67

EG-S41_006_F01 Translational activator O. sativa gb|AAT77858.1 4E-48

EG-S41_007_D03 Photosystem I assembly protein Ycf3 Ranunculusmacranthus

gb|AAZ03964.1 5E-16

Microtubule motoractivity

GO:0003777

cEG-S41_009_H07 Kinesin-like protein NACK1 Nicotiana tabacum dbj|BAB86283.1 9E-62

EG-S41_002_D09 Kinesin-related protein O. sativa gb|AAT07647.1| 2E-14cEG-S41_006_A08 Kinesin-related protein O. sativa gb|AAT07647.1| 2E-14cEG-S41_006_H10 ATATM; motor A. thaliana ref|NP_188630.1 2E-31

Enzyme regulatoractivity

GO:0030234

EG-S41_009_A04 Guanine nucleotide-exchangeprotein GEP2

O. sativa dbj|BAC79659.1| 2E-46

bEG-S41_008_H05 LEUNIG O. sativa dbj|BAD53054.1 1E-41

EG-S41_006_G05 Cystatin Actinidia eriantha gb|AAR92225.1 4E-23

EG-S41_010_A08 Phosphosulfolactate synthase-relatedprotein

Lycopersiconesculentum

gb|AAV36520.1 7E-27

Transcription initiationfactor activity

GO:0016986

b,cEG-S41_003_H05 bHLH protein family-like O. sativa dbj|BAD09664.1| 6E-17b,cEG-S41_006_B07 DNA polymerase epsilon

catalytic subunitO. sativa dbj|BAD23092.1| 3E-78

Oxygen transporteractivity

GO:0005344

EG-S41_010_C11 Nonsymbiotic hemoglobin Lotus corniculatus dbj|BAE46739.1 2E-54

EG-S41_006_E07 Non-symbiotic hemoglobin class 1 Malus x domestica gb|AAP57676.1 2E-67

GTPase regulatoractivity

GO:0030695

EG-S41_009_A04 Guanine nucleotide-exchangeprotein GEP2

O. sativa dbj|BAC79659.1| 2E-46

EG-S41_010_A08 Phosphosulfolactatesynthase-related protein

Lycopersiconesculentum

gb|AAV36520.1 7E-27

Phosphatase regulatoractivity

GO:0019208

EG-S41_008_H05 LEUNIG O. sativa dbj|BAD53054.1 1E-41

a Bold-italic indicates ESTs that were assigned more than one GO termb EST also listed in GO Table 2c EST also listed in GO Table 3

* BLASTX based similarities found in the nonredundant public database at NCBI

Plant Mol Biol (2009) 70:173–192 179

123

Tab

le2

ESTswithGO

biological

process

term

sfoundonly

intheinitiationlibrary

GO

biological

process

andaccession

ESTID

aDescription(BLASTX)*

Species

Accession

E-value

Positiveregulationofbiological

process

GO:0048518

cEG-S41_002_F08

Alpha-tubulin

Prunu

sdu

lcis

emb|CAA47635.1

4E-64

cEG-S41_007_C09

TubulinAlpha-6Chain

A.thaliana

gb|AAL38293.1|

8E-71

cEG-S41_007_F12

Beta-tubulin

G.hirsutum

gb|AAL92118.1

3E-28

EG-S41_007_A08

Microtubule-associated

protein

O.sativa

gb|AAT58855.1|

1E-57

bEG-S41

_006

_B07

DNA

polymeraseepsiloncatalyticsubunit

O.sativa

dbj|B

AD23092.1|

3E-78

EG-S41

_006

_E10

Cullin

A.thaliana

gb|AAG52544.1

4E-30

EG-S41_006_C07

Trehalose-6-phosphatesynthase

Ginkgobiloba

gb|AAX16015.1

1E-36

Regulationofgeneexpression,

epigenetic

GO:0040029

EG-S41_002_G04

OsN

AC7protein

O.sativa

dbj|B

AD19765.1|

2E-18

cEG-S41_003_E04

Shootm

eristemless-like

Petun

iaxhybrida

gb|AAM47027.1

3E-73

bEG-S41_003_H05

bHLH

protein

family-like

O.sativa

dbj|B

AD09664.1|

6E-17

cEG-S41_005_G05

TranscriptionregulatorSir2-likeprotein

A.thaliana

dbj|B

AB09243.1|

1E-45

b,cEG-S41

_006

_B07

DNA

polymeraseepsiloncatalyticsubunit

O.sativa

dbj|B

AD23092.1|

3E-78

Organ

developmentGO:0048513

EG-S41_005_H11

Lysyl-tRNA

synthetase

N.taba

cum

emb|CAC12821.1

6E-34

EG-S41

_006

_E10

Cullin

A.thaliana

gb|AAG52544.1

4E-30

CelldifferentiationGO:0030154

EG-S41_004_D12

WD-40repeatprotein

O.sativa

dbj|B

AC84349.1

1E-21

bEG-S41_008_H05

LEUNIG

O.sativa

dbj|B

AD53054.1

1E-41

Morphogenesis

GO:0009653

EG-S41

_006

_E10

Cullin

A.thaliana

gb|AAG52544.4

4E-30

Embryonic

developmentGO:0009790

EG-S41

_006

_E10

Cullin

A.thaliana

gb|AAG52544.1

4E-30

Reproductivestructure

developmentGO:0048608

EG-S41

_006

_E10

Cullin

A.thaliana

gb|AAG52544.1

4E-30

RegulationofdevelopmentGO:0050793

EG-S41

_006

_E10

Cullin

A.thaliana

gb|AAG52544.1

4E-30

aBold-italicindicates

ESTsthat

wereassigned

more

than

oneGO

term

bESTalso

listed

inGO

Table

1cESTalso

listed

inGO

Table

3

*BLASTX

based

similaritiesfoundin

thenonredundantpublicdatabaseat

NCBI

180 Plant Mol Biol (2009) 70:173–192

123

Tab

le3

ESTswithGO

cellularcomponentterm

sfoundonly

intheinitiationlibrary

GO

molecularfunctionandaccession

ESTID

aDescription(BLASTX)*

Species

Accession

E-value

Signal

peptidasecomplexGO:0005787

EG-S41_003_H09

Signal

peptidaseprotein-likeprotein

Cucum

ismelo

gb|AAO45754.1|

4E-18

EG-S41_008_B09

Signal

peptidaseprotein-likeprotein

Cucum

ismelo

gb|AAO45754.1|

4E-18

EG-S41_007_D10

RNA-bindingprotein

Avena

sativa

gb|AAO89229.1|

2E-30

EG-S41_005_H07

Calnexin

Helianthu

stuberosus

emb|CAA84491.1|

9E-41

EG-S41_002_H02

Calcium

ionbinding

A.thaliana

ref|NP_568202.1|

8E-68

EG-S41_010_G03

Grp94

Xerop

hyta

viscosa

gb|AAN34791.1|

2E-95

EG-S41_008_F08

24kDaseed

maturationprotein

O.sativa

dbj|B

AB89453.1|

2E-62

EG-S41_005_C01

Calretulin

Nicotiana

taba

cum

emb|CAA59694.1|

4E-41

EG-S41_002_F05

UDP-galactose

transporter

MSS4

A.thaliana

gb|AAG10147.1|

3E-80

Eukaryotictranslationelongation

factor1complexGO:0005853

EG-S41_009_F01

Translationelongationfactor1-gam

ma

Prunu

savium

gb|AAG17901.1|

3E-52

EG-S41_001_F11

Ribosomal

protein

L10a

O.sativa

dbj|B

AD28853.1|

9E-95

EG-S41_001_C03

Ribosomal

protein

L10a

O.sativa

dbj|B

AD28853.1|

6E-92

EG-S41_008_H10

40Sribosomal

protein-likeprotein

Solanu

mtuberosum

gb|ABC01888.1

9E-24

EG-S41_002_A06

Translationelongationfactor1A-2

Gossypium

hirsutum

gb|ABA12218.1|

1E-107

EG-S41_008_E12

Translational

elongationfactor1subunitBbeta

Pisum

sativum

gb|AAR15081.1|

3E-69

Chromatin

remodeling

complexGO:0016585

bEG-S41_005_G05

TranscriptionregulatorSir2-likeprotein

A.thaliana

dbj|B

AB09243.1|

1E-45

bEG-S41_007_C09

TubulinAlpha-6Chain

A.thaliana

ref|NP_171974.1

8E-71

EG-S41_008_A12

DNA-dam

age-repair/tolerance

resistance

protein

DRT111

A.thaliana

gb|AAC13593.1

2E-38

EG-S41_005_G08

Ribosomal

protein

L20

Typha

latifolia

gb|AAZ03851.1|

7E-37

EG-S41_004_F02

HistoneH2A

O.sativa

ref|XP_478633.1|

2E-31

Nucleotide-excisionrepair

complexGO:0000109

b,cEG-S41

_006

_B07

DNA

polymeraseepsiloncatalyticsubunitprotein

isoform

bO.sativa

ref|XP_465943.1|

3E-78

EG-S41_006_B08

Dek1-calpain-likeprotein

O.sativa

gb|AAL38190.1|

1E-106

b,cEG-S41_003_H05

bHLH

protein

family-like

O.sativa

ref|XP_483085.1|

6E-17

EG-S41_003_E04

Shootm

eristemless-like

Petun

iaxhybrida

gb|AAM47027.1|

3E-73

Cytochromeb6fcomplexGO:0009512

EG-S41_006_C09

Cytochromeb6/fcomplexsubunitV;petE

O.sativa

gb|AAS46133.1|

3E-17

EG-S41_003_F08

Cytochromeb6/fcomplexsubunitV;petE

O.sativa

gb|AAS46133.1|

3E-17

EG-S41_003_H06

Cytochromeb6/fcomplexsubunitV;petE

O.sativa

gb|AAS46133.1|

3E-17

EG-S41_003_F02

60Sribosomal

protein

L24

O.sativa

dbj|B

AD82702.1|

2E-17

Proton-transportingATPsynthase

complexGO:0045259

EG-S41

_005

_D09

ATPsynthaseCF0subunitIII

Eucalyptusglob

ulus

gb|AAX21015.1|

1E-14

EG-S41

_003

_H11

Transcriptioncoactivator

A.thaliana

ref|NP_192120.1

1E-30

EG-S41

_004

_C06

ribosomal

protein

L2

Yucca

schidigera

gb|AAZ04905.1|

2E-60

Proton-transportingATPsynthase

complex,couplingfactorF

GO:0045263

EG-S41

_005

_D09

ATPsynthaseCF0subunitIII

Eucalyptusglob

ulus

gb|AAX21015.1|

1E-14

EG-S41

_003

_H11

Transcriptioncoactivator

A.thaliana

ref|NP_192120.1

1E-30

EG-S41

_004

_C06

Ribosomal

protein

L2

Yucca

schidigera

gb|AAZ04905.1|

2E-60

Plant Mol Biol (2009) 70:173–192 181

123

Tab

le3continued

GO

molecularfunctionandaccession

ESTID

aDescription(BLASTX)*

Species

Accession

E-value

Microtubule

associated

complexGO:0005875

bEG-S41_002_D09

Kinesin-related

protein

O.sativa

ref|XP_475205.1|

2E-14

bEG-S41_006_A08

Kinesin-related

protein

O.sativa

ref|XP_475205.1|

2E-14

bEG-S41_009_H07

Kinesin-likeprotein

NACK1

Nicotiana

taba

cum

dbj|B

AB86283.1|

9E-62

SignalosomecomplexGO:0008180

EG-S41_001_C06

26Sproteasomenon-A

TPaseregulatory

subunit3

O.sativa

dbj|B

AC79193.1|

5E-55

EG-S41_001_C08

Cyclin,N-terminal

domain

O.sativa

gb|ABA98639.1|

2E-47

EG-S41_001_C10

Ribosomal

protein

S7

Narthecium

ossifrag

umgb|AAN31973.1|

2E-53

bEG-S41_007_F12

Beta-tubulin

Gossypium

hirsutum

gb|AAL92118.1|

3E-28

bEG-S41_006_H10

ATATM;motor

A.thaliana

ref|NP_188630.1|

2E-31

Origin

recognitioncomplexGO:0000808

EG-S41_007_H11

Origin

recognitioncomplexsubunit6-likeprotein

O.sativa

ref|NP_917945.1|

4E-56

EG-S41_008_C01

Origin

recognitioncomplexsubunit3-likeprotein

Zea

mays

gb|AAL10454.1|

5E-95

1,3-beta-glucansynthasecomplexGO:0000148

EG-S41_010_C08

Beta-1,3-glucansynthase

Nicotiana

alata

gb|AAK49452.2|

1E-106

GlycinecleavagecomplexGO:0005960

EG-S41_003_A01

glycinedecarboxylase

complexH-protein

O.sativa

dbj|B

AD45416.1|

2E-58

RNA

polymerasecomplexGO:0030880

EG-S41_007_G01

DNA

binding/DNA-directedRNA

polymerase

A.thaliana

ref|NP_196500.1

1E-35

ReplisomeGO:0030894

EG-S41

_006

_B07

DNA

polymeraseepsiloncatalyticsubunitprotein

isoform

bO.sativa

ref|XP_465943.1|

3E-78

DNA

polymerasecomplexGO:0042575

EG-S41

_006

_B07

DNA

polymeraseepsiloncatalyticsubunitprotein

isoform

bO.sativa

ref|XP_465943.1|

3E-78

aBold-italicindicates

ESTsthat

wereassigned

more

than

oneGO

term

bESTalso

listed

inGO

Table

1cESTalso

listed

inGO

Table

2

*BLASTX

based

similaritiesfoundin

thenonredundantpublicdatabaseat

NCBI

182 Plant Mol Biol (2009) 70:173–192

123

development stages examined (day 0 and 1, 2, 4, 8 and

16 days after the removal of 2,4-D and SEms) was per-formed using normalized intensity values of transcript

accumulation from two distinct biological replicates. The

two resulting inter-stage distance matrixes were not sig-nificantly different (r = 0.756 and P = 0.031) based on the

Mantel test, demonstrating the reproducibility of the bio-

logical processes studied.HCA was then used to identify clusters of co-expressed

genes based on normalized intensity values obtained fromhybridizations with labeled cDNAs synthesized from the

7 time points and the two biological replicates. HCA

revealed the existence of two main clusters of genes termedGroup 1 and 2. Group 1 (total of 882 ESTs) exhibited

constitutively low expression (normalized intensity values

less than 1 at all the time points examined) throughout SEand was not included in subsequent analyses (Supplemental

Table 1). Group 2 was further divided into four major

co-expressed gene clusters termed Profiles 1–4, and arepresented as an average for each profile (dendrograms not

shown; Fig. 3). Each of these 4 clusters of co-expressed

genes retained could be further divided into 3–5 individualsub-groups (Supplemental Fig. 4; dendrograms not shown).

A careful examination of the sub-group contents revealed

that this second level of classification was based on dif-ferent intensity ranges for a given profile. The four profiles

contained a total of 227 transcripts that had similar accu-

mulation profiles in the two biological repetitions.

ANOVA analysis identifies transcripts differentially

accumulated at certain stages of SE

In order to identify transcripts with differential accumula-

tion at selected stages of development during SE, we usedANOVA analysis to verify the accumulation peaks that

characterized each profile identified from the HCAdescribed above. The peak transcript accumulation that

characterized each profile (e.g. Day 8 for profile 1) was

compared to a contrasting developmental stage byANOVA. Significant differences (P\ 0.05) were found for

192 of the 227 transcripts identified in the HCA analysis in

both biological replications (Supplementary Tables 2, 3, 4and 5). From the 192 transcripts, 91 of these shared simi-

larities to sequences in the public databases, while the

remainder shared no significant homology to knownsequences or shared similarity with unknown proteins

(Fig. 4). The largest proportion (23 from 38 ESTs) of ESTs

from profile 1 had no significant homology to knownsequences or shared similarity with unknown proteins. In

contrast, profile 1 had a larger proportion of ESTs involved

in cell structure (3 from 38 ESTs) and signal transduction(3 from 38 ESTs) than the other profiles (Fig. 4 and

0.0

0.2

0.4

0.6

0.8

1.0

Mea

n S

tand

ardi

zed

Exp

ress

ion

0.0

0.2

0.4

0.6

0.8

1.0

Profile 3 Profile 4

Profile 1 Profile 2T0 T1 T2 T4* T8* T16 SEm T0 T1 T2 T4 T8* T16* SEm

T0 T1* T2 T4 T8* T16 SEm T0 T1 T2 T4 T8* T16 SEm*

Fig. 3 Hierarchical clusteranalysis (Ward’s pair-wisegrouping method, Euclidiandistance) was performed on thegene expression dataset(normalized signal intensitiesfrom the two biologicalrepetitions) to clustersimilarities in gene expressionprofiles. Profiles depicted arebased on the mean standardizedexpression as follows: themeans of the signals for eachEST at each developmentalstage were standardized bydividing by the maximum signalof each EST. x-axis: T0, 0 Days(with 2,4-D); T1, 1 Day; T2, 2Days; T4, 4 Days; T8, 8 Days;T16, 16 Days after 2,4-Dremoval; SEm, SomaticEmbryos. y-axis: meanstandardized expression.*Developmental stages at whichANOVA was used to determinesignificance of differences oftranscript accumulation

Plant Mol Biol (2009) 70:173–192 183

123

Supplemental Table 6). Profile 2 contained the most ESTs

(total of 66) that detected reproducible accumulation pat-terns; the largest proportion of those ESTs was assigned to

the functional category related to disease and defense (25

from 66 ESTs, Fig. 4 and Supplemental Table 6). Themajority of these ESTs encoded either methallothionein-

like or peroxidase related proteins. The second largest

proportion had no significant homology to sequences in thepublic databases (18 from 66 ESTs). Profile 2 included the

only ESTs with energy related functions and also contained

5 ESTs with protein destination and storage functions, 4 ofwhich are related to the ubiquitin/26S proteosome pathway.

The largest proportion of ESTs (38 from 57 ESTs) in

profile 3 had no significant homology to sequences in thepublic databases (Fig. 4). Profile 3 had proportionally more

ESTs in the functional category disease and defense

(4 from 57 ESTs) and transcription and post-transcription(3 from 57 ESTs). Three of the 4 disease and defense

related ESTs encode proteins similar to glutathione

S-transferases (Supplemental Table 6). The majority of theESTs (16 from 31 ESTs) in Profile 4 had either no sig-

nificant homology to sequences in the public databases or

had similarities to unknown or unclassified proteins(Fig. 4). Profile 4 had proportionally more ESTs with

functions related to disease and defense (4 from 31 ESTs),

protein destination and storage (4 from 31 ESTs), tran-scription and post-transcription (2 from 31 ESTs) and cell

structure (2 from 31 ESTs).

Discussion

Differences in the developmental kinetics, sampling,

experimental design and culture conditions used for global

gene expression analyses make a straightforward compar-

ison with different in vitro systems difficult. Indeed, SEinduction and SEm development can occur within different

time frames and SEms can arise either directly after an

initial induction treatment, or indirectly via the formationof embryogenic callus or suspension cells (Gray 2005). For

example, gymnosperm indirect SE is similar to oil palm in

that auxin must be removed from the cell culture mediumin order for SEms to develop (Lippert et al. 2005; Stasolla

et al. 2003, 2004; van Zyl et al. 2003). However, in gym-

nosperms proembryogenic masses that develop aredistinguishable by cellular organization and cell number

prior to hormone removal (von Arnold et al. 2002). By

contrast, oil palm cells grown in the presence of 2,4-D donot have any distinguishable cellular organization and are

indistinguishable from cells grown without 2,4-D for

14 days; SEms become distinguishable after 30 days ofliquid culture plus approximately 20 days of growth on

solid media without 2,4-D (Fig 1). This contrasts to the leaf

base SE system of wheat in which a 24 h 2,4-D treatmentinitiates SE and SEms emerge directly from leaf explants

after 10 days of culture on 2,4-D free medium (Singla et al.

2007). Another example is the potato and Medicago trun-catula SE systems that pass via an indirect callogenesis

phase followed by a period of weeks before SEms arise

(Imin et al. 2005; Sharma et al. 2008). In addition, theM. truncatula system does not require the removal of auxin

from the culture medium for SEms to develop. Finally, the

microspore-derived SEm system of Brassica napus usesheat-shock as the stimulus to initiate SE (Joosen et al.

2007; Tsuwamoto et al. 2007). In this system, SEms are

derived directly from haploid microspores and the devel-opment of discernable SEms occurs within 10 days of

induction treatment. Despite the differences between the

3

3

1

2

1

1

2

2

2

21

38

1

2

1

2

1

3

1

4

5

25

3

18

66

1

1

2

1

2

3

2

4

3

38

57

1

1

2

1

2

4

4

6

10

31

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Transposons

Transporters

Energy

Intracellular Traffic

Secondary Metabolism

Signal Transduction

Cell Structure

Cell Growth/Division

Metabolism

Protein Synthesis

Transcription and Post-transcription

Protein Destination and Storage

Disease/Defense

Unknown or Unclassified Proteins

No Significant Homology to Public Databases

Total

Functional Distribution of ESTs in Expression Profiles

Profile 1 Profile 2 Profile 3 Profile 4

Fig. 4 The functionaldistribution of ESTs identifiedwithin the 4 expression profilesby clustering and ANOVAanalyses. The numbers indicatethe total ESTs found withineach category based on Bevanet al. (1998). The percentagesindicate the percentage of ESTsfound within each category

184 Plant Mol Biol (2009) 70:173–192

123

systems, whether monocot, dicot, gymnosperm, direct or

indirect SE, common genes and functions emerge thatappear to be independent of the in vitro system examined

and will be discussed below (Table 4).

Fluctuations in gene expression patterns underlie early

oil palm SE and SEm development

In the present study, we find four expression profiles that

reveal fluctuations in global transcript abundance duringdevelopment (Fig. 3). Similar expression fluctuations dur-

ing development have been observed in various in vitro

systems including soybean, maize, Picea abies and Piceaglauca (Che et al. 2006; Stasolla et al. 2003, 2004; Thi-

baud-Nissen et al. 2003; van Zyl et al. 2003). Furthermore,

a dynamic global expression pattern was associated withnormal SEm development in contrast to developmentally

arrested cell lines unable to form SEms (van Zyl et al.

2003). These results indicate that patterns of gene expres-sion during SE of widely different systems can be

characterized by dynamic changes during development and

not always simple increases or decreases in transcriptabundance. In addition, depending on the system exam-

ined, there may be developmental stages during which

global transcript accumulation change in erratic ways, forinstance, an increase followed by a decrease and again an

increase (Che et al. 2006; Stasolla et al. 2003; 2004; van

Zyl et al. 2003). In oil palm we observed that 8 days afterthe removal of 2,4-D is distinct from all other time points.

Indeed it appears that the most significant global changes in

transcript accumulation occur at this time point: a peak oftranscript accumulation characterized in profile 1, while

profiles 2, 3 and 4 had lower transcript accumulation at this

stage. Expression fluctuations may be related to the lengthof time required for SE to occur in the various systems

examined. For example, very little fluctuations are seen

during the rapid SE systems such as with B. napus, wheatand M. truncatula while with slower responding systems

such as the gymnosperms, soybean, maize and in the

present study with oil palm, fluctuations in gene expressionappear to be common (Che et al. 2006; Joosen et al. 2007;

Lippert et al. 2005; Singla et al. 2007; Stasolla et al. 2003;

2004; Thibaud-Nissen et al. 2003; van Zyl et al. 2003).Alternatively, fewer time points sampled in the faster SE

systems may not allow adequate sampling to detect whe-

ther fluctuations occur in these systems. We hypothesizethat in the oil palm SE system a first global rearrangement

of gene expression occurs within the first 8 days after the

removal of 2,4-D from the medium followed by secondwave of changes that lead to SEm development. One

explanation of this pattern of expression may be a residual

amount of 2,4-D may remain in the suspension cells upto 8 days until it is fully removed from the system. Indeed,

at 8 days we observed no accumulation of a transcript

encoding an AUX/IAA transcriptional regulator typicallyinduced by the presence of auxin (Abel et al. 1994). The

association of auxin related gene expression and other

cellular processes occurring during SE are discussed in thesections below.

Two recent articles presented EST data from oil palm

from different tissues including embryogenic and nonem-bryogenic calli, suspension cells, SEms and zygotic

embryos (Ho et al. 2007; Low et al. 2008). The ESTs fromthese studies were derived mostly from randomly selected

unsubtracted cDNAs that represent typically the most

abundant transcripts. By contrast, the present studyemployed the SSH methodology shown to facilitate the

enrichment of rare transcripts (Diatchenko et al. 1996).

Indeed, a total of 882 ESTs presented here corresponded totranscripts that produced constitutively low intensity values

during oil palm SE. In the previous work suspension cul-

tures were used to generate ESTs but the exact stage ofdevelopment or culture conditions were not indicated (Ho

et al. 2007). Likewise, embryogenic and nonembryogenic

calli samples and SEms were used to generate ESTs but theexact stages of development and culture conditions were

undisclosed (Low et al. 2008). Despite these limitations,

ESTs found in our study that were also observed in theserecent oil palm articles will be discussed in the sections

that follow.

Oxidative stress and redox homeostasis gene expression

during oil palm SE are common to diverse SE systems

One common example of genes that are activated and

differentially expressed independent of the SE system

examined are those encoding glutathione S-transferases(GSTs). This large family of enzymes can catalyze the

conjugation of glutathione to various electrophilic mole-

cules and have a wide range of proposed functionsincluding detoxifying oxidative-stress metabolites (Dixon

et al. 2002). GSTs are expressed during SE of many species

while their exact roles are still unknown (Che et al. 2006;Galland et al. 2001; Joosen et al. 2007; Low et al. 2008;

Malik et al. 2007; Thibaud-Nissen et al. 2003; Tsuwamoto

et al. 2007; Vrinten et al. 1999). GST expression is char-acteristic for the transition towards somatic embryo

development in the in vitro culture systems of soybean,

barley, chicory and B. napus and oil palm (Galland et al.2001; Joosen et al. 2007; Low et al. 2008; Malik et al.

2007; Thibaud-Nissen et al. 2003; Tsuwamoto et al. 2007;

Vrinten et al. 1999). Indeed in recent articles, GST wasproposed to function as a central regulator of wheat SE

associated with auxin related gene expression (Singla et al.

2007). In this previous article, they observed a coordinatedincrease in GST and an AUX/IAA transcriptional regulator

Plant Mol Biol (2009) 70:173–192 185

123

Tab

le4

PutativefunctionsandsimilaritiesofEST/transcripts

inprofilesexam

ined

inDiscussion

Profile

1

Functional

category

andESTID

Description(BLASTX)

Species

Accession

E-value

aNorm

alized

signal

intensities

D0

D1

D2

D4*

D8*

D16

SEm

Antioxidativenetwork

orredoxstatus

EG-S31_010_D01

GlutathioneS-transferase2

Pap

aver

somniferum

gb|AAF22518.1

9E-30

7.96

9.65

9.75

6.66

14.63

4.46

8.52

EG-S41_010_E05

Cytosolicascorbateperoxidase

Nicotiana

taba

cum

dbj|B

AA12918.1|

1E-09

10.79

9.14

6.44

4.12

21.37

3.51

9.10

EG-S31_005_D07

Thioredoxin

HPop

ulus

trem

ulaxPop

ulus

trem

uloides

gb|AAL99941.1

7E-18

6.83

9.13

8.2

7.1

12.9

9.26

15.42

Patho

genesis-related(PR)proteins

EG-S31_009_D02

ChainA,Endo-Beta-1,3-G

lucanase

MusaAcuminata

pdb|2CYG|

1E-50

12.07

8.14

6.81

4.26

20.51

2.99

8.11

Intracellulartraffic

andpo

larization

EG-S31_007_C01

Actin

depolymerizingfactor5

A.thaliana

pir||B

84543

1E-20

18.53

14.96

11.98

7.93

33.17

8.29

12.89

EG-S41_010_G01

Profilin2

Lilium

long

iflorum

gb|AAF08303.1|

1E-28

9.36

7.08

6.68

5.74

17.48

5.84

9.83

Profile

2

Functional

Category

andESTID

Description(BLASTX)

Species

Accession

E-value

aNorm

alized

signal

intensities

D0

D1

D2

D4

D8*

D16*

SEm

Antioxida

tive

networkor

redo

xstatus

EG-S31_009_G02

Cationic

peroxidase

Cicer

arietinu

mem

b|CAB71128.2|

7E-51

0.10

0.22

0.86

2.54

0.18

5.98

3.09

EG-S41_007_F10

Peroxidase

Gossypium

hirsutum

gb|AAA99868.1|

3E-38

0.01

0.13

0.92

2.92

0.28

8.90

3.61

Disease/defense

EG-S31_004_H04

Metallothionein-likeprotein

Typha

latifolia

gb|AAK28022.1|

1E-15

3.32

1.49

2.39

4.23

0.46

8.26

5.44

EG-S41_003_H07

Metallothionein-likeprotein

Typha

latifolia

gb|AAK28022.1|

3E-20

1.36

1.03

3.38

8.99

0.10

14.52

6.48

EG-S31_004_A03

Metallothionein-likeprotein

type2

Narcissus

pseudo

narciss

gb|AAL16908.1|

1E-18

1.39

1.31

3.22

9.25

0.09

15.75

7.05

Intracellulartraffic

andpo

larization

EG-S41_003_B11

Clathrinassembly

protein

A.thaliana

gb|AAL77661.1

2E-48

0.54

0.54

0.57

0.71

0.26

0.74

0.58

EG-S41_009_A10

Endosperm

C-24sterolmethyltransferase

Zea

mays

gb|AAB70886.1|

2E-72

1.05

0.81

1.90

5.76

0.27

6.89

2.89

EG-S31_005_G01

Suppressorofactin1

O.sativa

ref|XP_466191.1|

1E-47

0.44

0.56

1.15

2.81

0.05

4.86

1.95

EG-S31_008_B02

Rab-typesm

allGTP-bindingprotein

Cicer

arietinu

mdbj|B

AA76422.1|

1E-98

0.13

1.51

1.00

3.84

0.21

2.34

2.09

Ubiqu

itin/26proteasomepa

thway

EG-S31_006_G01

Proteasomesubunitbetatype5precursor(20Sproteasome)

Spinacia

oleracea

sp|O24361|PSB5_SPIO

L5E-76

1.09

0.95

1.73

5.61

0.16

7.12

4.24

EG-S31_003_G01

UPL5;ubiquitin-protein

ligase

A.thaliana

ref|NP_192994.1

9E-94

0.35

0.35

0.50

0.62

0.17

1.19

0.65

EG-S31_006_D09

Ubiquitin-conjugatingenzymefamilyprotein-like

Solanu

mtuberosum

gb|ABB29951.1

1E-62

0.30

1.72

1.47

2.75

0.1

3.18

0.98

EG-S31_007_A07

Ubiquitin-related

modifier-1

O.sativa

dbj|B

AC01162.1

1E-12

0.19

0.54

0.65

1.07

0.05

1.27

0.53

Auxin

respon

se

EG-S31_009_A08

AUX/IAA

protein

O.sativa

gb|ABA99793.1

4E-31

0.26

0.56

0.72

0.62

0.00

0.26

0.38

186 Plant Mol Biol (2009) 70:173–192

123

Tab

le4

continued

Profile

3

Functional

category

andESTID

Description(BLASTX)

Species

Accession

E-value

aNorm

alized

signal

intensities

D0

D1*

D2

D4

D8*

D16

SEm

Antioxida

tive

networkor

redo

xstatus

EG-S41_007_F04

glutathioneS-transferase2

Pap

aver

somniferum

gb|AAF22518.1|

4E-31

0.29

2.78

2.04

1.99

0.25

0.69

0.40

EG-S31_003_H04

GlutathioneS-transferase2

Pap

aver

somniferum

gb|AAF22518.1|

6E-23

0.73

7.83

6.54

7.95

0.19

1.41

0.42

Patho

genesis-related(PR)proteins

EG-S31_009_B04

PR-4bprotein

Nicotiana

taba

cum

emb|CAA42821.1|

1E-14

31.69

7.98

9.66

16.36

0.75

4.43

0.82

Intracellulartraffic

andpo

larization

EG-S41_006_H01

Form

in2

Hom

osapiens

ref|NP_064450.2|

8E-06

1.56

3.11

1.85

0.47

0.16

0.24

0.29

EG-S41_001_E06

Sperminesynthase

Taxod

ium

distichu

mdbj|B

AD02823.1|

1E-96

1.90

1.96

2.29

1.73

0.27

0.90

1.64

Profile

4

Functional

category

andESTID

Description(BLASTX)

Species

Accession

E-value

aNorm

alized

signal

intensities

D0

D1

D2

D4

D8*

D16

SEm*

Antioxida

tive

networkor

redo

xstatus

EG-S31_007_C11

Cytosolicascorbateperoxidase

Dimocarpu

slong

angb|AAW49512.1|

9E-72

0.94

1.88

2.90

4.25

0.04

4.70

6.18

EG-S41_001_B08

Phospholipid

hydroperoxide

glutathioneperoxidase

O.sativa

gb|AAX96834.1|

1E-46

0.79

1.05

1.02

1.74

0.00

0.69

2.14

Disease/defense

EG-S31_002_C07

DefensinEGAD1

Elaeisgu

ineensis

gb|AAN52490.1|

2E-21

5.15

0.29

0.59

1.65

0.31

2.97

5.98

EG-S41_006_A11

Gam

ma-thionin

Castaneasativa

gb|AAL15885.1|

3E-10

0.19

1.70

0.70

1.15

0.00

0.57

20.90

Intracellulartraffic

andpo

larization

EG-S31_005_A06

ADP-ribosylationfactor

Solanu

mtuberosum

gb|ABA40446.1|

6E-07

10.86

10.91

7.37

5.47

0.51

4.84

8.62

Ubiqu

itin/26proteasomepa

thway

EG-S41_001_D08

Ubiquitin

family

O.sativa

gb|ABA91448.1

3E-18

1.20

1.21

1.46

2.29

0.21

1.21

4.29

Forcomplete

ESTprofile

dataseeSupplementary

Table

6aOnly

thenorm

alized

signal

intensities

from

onebiological

repetitionarepresentedhere

*ANOVAanalysiswas

usedto

determinesignificance

ofdifferencesoftranscriptaccumulationatthesedevelopmentalstages

andonly

those

withP-values\0.05in

both

biologicalrepetitions

arelisted

here.

D0,Day

0;D1,Day

1;D2,Day

2;D4,Day

4;D8,Day

8;D16,Day

16;SEM,somatic

embryo

Plant Mol Biol (2009) 70:173–192 187

123

transcript abundance. Our results indicate at least two and

possibly up to 4 members of the GST gene family areexpressed early during the initiation of SEm development.

The GSTs found in profile 3 are least abundant at 8 days,

which corresponds to the lowest accumulation of a tran-script for an AUX/IAA transcriptional regulator. This

result is similar to what was seen during wheat SE where

GST and AUX/IAA transcriptional regulator transcriptaccumulation are coordinated (Singla et al. 2007). By

contrast, another GST transcript is observed in profile 1with a peak of transcript accumulation at 8 days. Different

GST transcripts with opposite expression profiles were also

observed during P. glauca SE (Stasolla et al. 2003; van Zylet al. 2003). In addition, one GST is expressed preferen-

tially in non-embryogenic callus of oil palm and chicory

(Legrand et al. 2007; Low et al. 2008). Our results suggestup to two different contexts for GST function during oil

palm SE; one associated with a reduction of auxin induced

gene expression that may correspond to the elimination of2,4-D (the GSTs in profile 3) from the suspension cells, and

another opposite to auxin induced gene expression prior to

the phase that leads to the development of the SEms (theGSTs in profile 1). In addition, two other antioxidative

related transcripts, a cytosolic ascorbate peroxidase and a

thioredoxin H, are also observed with the GST in profile 1.Both cytosolic ascorbate peroxidase and thioredoxin H

transcripts are expressed in M. truncatula SE during the

early stages of cellular differentiation and proliferation(Imin et al. 2005). Likewise, a thioredoxin H transcript is

also expressed during early cotton SE associated with the

embryogenic callus and preglobular embryo stages, and anascorbate peroxidase is expressed at an early stage of B.napus microspore-derived embryo development (Zeng

et al. 2006; Joosen et al. 2007). By contrast, both cytosolicascorbate peroxidase and thioredoxin H transcripts are

repressed in normal embryogenic lines and accumulate in

non-embryogenic lines of P. abies (van Zyl et al. 2003;Stasolla et al. 2004). The authors suggest the high amount

of oxidative stress related transcripts may be related to the

inhibition of embryo development in the non-embryogenicline. Finally, expression studies with two thioredoxin

related transcripts in wheat showed conflicting results but

in general an accumulation related to SEm development(Singla et al. 2007). In the present study, we find thiore-

doxin H with GST and cytosolic ascorbate peroxidase

transcripts clustered in profile 1 and present in SEms. Wealso observe transcripts for a glutamine peroxidase and a

cytosolic ascorbate peroxidase in profile 4 that are highly

expressed in SEms. In addition, ESTs that encode peroxi-dases were found in profile 2 with peak transcript

accumulation at 16 days after 2,4-D removal. Numerous

peroxidase ESTs were also isolated from oil palm cellsuspensions and associated with the developing SEm (Ho

et al. 2007; Low et al. 2008). Peroxidases can also function

in the antioxidant network and are associated with manyaspects of SE (Malinowski and Filipecki 2002). For

example, a secreted cationic peroxidase was shown to

restore carrot SE and a cell wall bound cationic peroxidasewas proposed to be involved in the activation of cell

division and differentiation during SE of asparagus (Cor-

dewener et al. 1991; Takeda et al. 2003). In the wheat SEsystem, peroxidase transcripts were expressed at different

stages of SE including SEm development (Singla et al.2007).

Overall, these results suggest dynamic changes in the

antioxidative network and redox related mechanisms occurduring oil palm SE and other SE systems examined.

Compared with other studies, including monocot, dicot,

gymnosperm, direct and indirect SE, different genes fromlarge gene families that encode enzymes involved in the

antioxidative response are expressed and may have dif-

ferent functional roles that are developmental stage and SEsystem dependent.

Disease and defense related genes in diverse SEsystems

A range of stress treatments can trigger SE in differentspecies (Feher et al. 2003; Ikeda-Iwai et al. 2003; Pasternak

et al. 2002). The differentially expressed transcripts with

possible roles related to disease or defense comprised thelargest group found within the 4 profiles, particularly in

profile 2. The most predominant transcripts in profile 2

encode putative metallothionein-like proteins and peroxi-dases. While metallothioneins are expressed during stages

of SE, their expression can also characterize non-embryo-

genic callus and their roles during SE remain unclear(Chatthai et al. 1997; Dong and Dunstan 1996; Legrand

et al. 2007; Low et al. 2008). In plants, metallothioneins are

suggested to be involved in plant metal homeostasis, par-ticularly with copper, and recently were shown to protect

plants from copper and cadmium toxicity (Cobbett and

Goldsbrough 2002; Guo et al. 2008). We also found apathogenesis-related (PR)-4b protein EST differentially

expressed in profile 3 along with the GSTs. PR proteins are

a heterogeneous group of proteins encoded by genesoriginally found to be expressed in response to pathogen

attack, but also appear to have roles in response to abiotic

stress, during development, and different classes accumu-late during SE (Helleboid et al. 2000a; b; Seo et al. 2008).

ESTs encoding other classes of putative PR proteins

including an endo-beta-1,3-glucanase were found in profile1, a gamma-thionin (also known as defensins) and the oil

palm defensin EGAD1 in profile 4. The EGAD1 transcript

was previously observed to accumulate in oil palmembryogenic callus and SEms (Low et al. 2008; Tregear

188 Plant Mol Biol (2009) 70:173–192

123

et al. 2002). Transcripts for a PR-4 protein, an endo-beta-

1,3-glucanase and methalothioneins were associated withnon-embryogenenic callus in oil palm (Low et al. 2008). In

contrast, our results indicate that different PR-like proteins

and methalothionein transcripts are expressed in embryo-genic cells and throughout oil palm SE and support the

hypothesis for functional roles during SE.

Auxin response, vesicular trafficking and polarization

related transcripts are differentially expressed duringSE

An EST in profile 2 encodes an AUX/IAA transcriptionalregulator whose transcript accumulation is lowest at 8 days

after 2,4-D removal. AUX/IAA proteins are a family of

transcriptional regulators induced by auxin and involved inthe negative regulation of the auxin response (Abel et al.

1994; Overvoorde et al. 2005). AUX/IAA proteins bind to

auxin response factor proteins and inhibit the activation ofauxin-induced gene expression. Auxin induced gene

expression is achieved by the degradation of the AUX/IAA

proteins through the ubiquitin/26S proteasome pathwayrelieving the negative regulation to liberate auxin response

factors that initiate transcription. From the GO annotation

of the ESTs, a homolog of CULLIN, a key component ofthe ubiquitin/26S proteosome pathway, was found only in

the initiation library which is consistent with its role as an

important post-translational regulator during embryogene-sis (Thomann et al. 2005a). Two CULLINS of A. thalianahave been shown to be essential for normal embryo

development (Thomann et al. 2005b). Also in profile 2, thepresence of transcripts encoding components of the 26S

proteasome complex suggest coordinated regulation of the

auxin response at both the transcriptional and post-trans-lational levels may occur after the removal of 2,4-D for

8 days. In addition, an EST in profile 2 encodes an endo-

sperm C-24 sterol methyltransferase, an enzyme involvedin sterol biosynthesis important during Arabidopsis zygoticembryogenesis (Schrick et al. 2002). The activity of this

enzyme functions to determine the sterol compositionalchanges required for cell polarity related to auxin efflux in

Arabidopsis (Willemsen et al. 2003). Another transcript

related to polarization events includes a spermine synthaseinvolved in polyamine biosynthesis found in profile 3. In

Arabidopsis, a mutation in the ACL5 gene, which is

expressed in provascular cells and encodes a sperminesynthase, results in an increase in organ vascularization

proposed to be due to changes in polar auxin transport

(Clay and Nelson 2005). Finally, transcripts encodingcytoskeleton and vesicular trafficking components with

similarity to actin depolymerizing factor (ADF), suppressor

of actin (SAC), profilin 2, formin, and a clathrin assemblyprotein, all cellular components related to polarized

development were identified (Chen et al. 2002; Vidali et al.

2007; Zhong et al. 2005; Zigmond 2004). In addition, atranscript for an ADP-ribosylation factor (ARF) is

observed in profile 4, with peak accumulation in SEms.

ARF are key regulators of intracellular vesicle transportand the Arabipopsis ARF1 was shown to be important in

epidermal cell polarity (Xu and Scheres 2005). ARF

activity is regulated by ARF GDP/GTP exchange factors(ARF-GEFs) and the GNOM ARF-GEF is required for the

recycling of auxin transport components during Arabid-opsis embryogenesis (Geldner et al. 2003). In the present

study, a transcript for a guanine nucleotide-exchange pro-

tein GEP2 was found only in the initiation library(Table 1). Overall, our results suggest transcriptional

changes related to auxin response and polarization path-

ways occur during oil palm SE and provide genecandidates to study the cellular events required for the

production of well-polarized SEms.

Conclusions and prospects

The present study is the first global transcript profiling

analysis of oil palm SE. While oil palm is not a model

species for studies on embryogenesis, the suspension cul-ture based SEm system is the most promising for the

multiplication of elite palms and understanding the

molecular basis of this process allow comparisons withmodel and non-model species including monocots, dicots

and gymnosperms. Our results indicate that reproducible

transcript accumulation occurs in a coordinated mannerduring the time points investigated and that the largest

contribution to differential expression involves disease and

defense related genes, specifically those related to oxidativestress. In addition, several genes with roles previously

associated with SE or zygotic embryogenesis of other spe-

cies are also activated during oil palm SE suggesting similarmechanisms underlying SE in this species. In particular,

transcripts encoding proteins with possible roles related to

auxin response and cellular polarization events during earlydevelopment were identified. These results form a basis for

future studies towards understanding the roles of these

genes in embryo development, and whether the mechanismsof SE and gene function are conserved compared with

model and other agronomically important species.

Acknowledgments We would like to thank Xavier Sabau for theexpertise in cDNA arraying at the Robotics and DNA SequencingPlatform CIRAD, Montpellier Languedoc-Roussillon Genopole (http://www.genopole-montpellier-lr.org), Thierry Beule for helpful techni-cal advice on macroarray methodology, and Ivanna Fuentes forexcellent technical assistance with handling the PCR amplification ofthe EST plasmid inserts. This work was financed by institutionalfunds from IRD and CIRAD.

Plant Mol Biol (2009) 70:173–192 189

123

References

Abel S, Oeller PW, Theologis A (1994) Early auxin-induced genesencode short-lived nuclear proteins. Proc Natl Acad Sci USA91:326–330. doi:10.1073/pnas.91.1.326

Aberlenc-Bertossi F, Noirot M, Duval Y (1999) BA enhances thegermination of oil palm somatic embryos derived from embryo-genic suspension cultures. Plant Cell Tissue Organ Cult 56:53–57. doi:10.1023/A:1006241215717

Alba R, Fei Z, Payton P, Liu Y, Moore SL, Debbie P, Cohn J,D’Ascenzo M, Gordon JS, Rose JK, Martin G, Tanksley SD,Bouzayen M, Jahn MM, Giovannoni J (2004) ESTs, cDNAmicroarrays, and gene expression profiling: tools for dissectingplant physiology and development. Plant J 39:697–714. doi:10.1111/j.1365-313X.2004.02178.x

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W,Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a newgeneration of protein database search programs. Nucleic AcidsRes 25:3389–3402. doi:10.1093/nar/25.17.3389

Bevan M, Bancroft I, Bent E, Love K, Goodman H, Dean C,Bergkamp R, Dirkse W, Van Staveren M, Stiekema W, Drost L,Ridley P, Hudson SA, Patel K, Murphy G, Piffanelli P, WedlerH, Wedler E, Wambutt R, Weitzenegger T, Pohl TM, Terryn N,Gielen J, Villarroel R, De Clerck R, Van Montagu M, LecharnyA, Auborg S, Gy I, Kreis M, Lao N, Kavanagh T, Hempel S,Kotter P, Entian KD, Rieger M, Schaeffer M, Funk B, Mueller-Auer S, Silvey M, James R, Montfort A, Pons A, PuigdomenechP, Douka A, Voukelatou E, Milioni D, Hatzopoulos P, PiravandiE, Obermaier B, Hilbert H, Dusterhoft A, Moores T, Jones JD,Eneva T, Palme K, Benes V, Rechman S, Ansorge W, Cooke R,Berger C, Delseny M, Voet M, Volckaert G, Mewes HW,Klosterman S, Schueller C, Chalwatzis N (1998) Analysis of1.9 Mb of contiguous sequence from chromosome 4 of Arabid-opsis thaliana. Nature 391:485–488. doi:10.1038/35140

Buffard-Morel J, Verdeil JL, Pannetier C (1992) Embryogenesesomatique du cocotier (Cocos nucifera L.) a partir de tissusfoliaires: etude histologique. Can J Bot 70:735–741

Cairney J, Pullman GS (2007) The cellular and molecular biology ofconifer embryogenesis. New Phytol 176:511–536. doi:10.1111/j.1469-8137.2007.02239.x

Cairney J, Zheng L, Cowels A, Hsiao J, Zismann V, Liu J, Ouyang S,Thibaud-Nissen F, Hamilton J, Childs K, Pullman GS, Zhang Y,Oh T, Buell CR (2006) Expressed sequence tags from loblollypine embryos reveal similarities with angiosperm embryogene-sis. Plant Mol Biol 62:485–501. doi:10.1007/s11103-006-9035-9

Chatthai M, Kaukinen KH, Tranbarger TJ, Gupta PK, Misra S (1997)The isolation of a novel metallothionein-related cDNA expressedin somatic and zygotic embryos of Douglas-fir: regulation byABA, osmoticum, and metal ions. Plant Mol Biol 34:243–254.doi:10.1023/A:1005839832096

Che P, Love TM, Frame BR, Wang K, Carriquiry AL, Howell SH(2006) Gene expression patterns during somatic embryo devel-opment and germination in maize Hi II callus cultures. Plant MolBiol 62:1–14. doi:10.1007/s11103-006-9013-2

Chen CY, Wong EI, Vidali L, Estavillo A, Hepler PK, Wu HM,Cheung AY (2002) The regulation of actin organization by actin-depolymerizing factor in elongating pollen tubes. Plant Cell14:2175–2190. doi:10.1105/tpc.003038

Clay NK, Nelson T (2005) Arabidopsis thickvein mutation affectsvein thickness and organ vascularization, and resides in aprovascular cell-specific spermine synthase involved in veindefinition and in polar auxin transport. Plant Physiol 138:767–777. doi:10.1104/pp.104.055756

Cobbett C, Goldsbrough P (2002) Phytochelatins and metallothione-ins: roles in heavy metal detoxification and homeostasis. Annu

Rev Plant Biol 53:159–182. doi:10.1146/annurev.arplant.53.100301.135154

Consortium GO (2008) The Gene Ontology project in 2008. NucleicAcids Res 36:D440–D444. doi:10.1093/nar/gkm883

Cordewener J, Booij H, Zandt H, Engelen F, Kammen A, Vries S(1991) Tunicamycin-inhibited carrot somatic embryogenesis canbe restored by secreted cationic peroxidase isoenzymes. Planta184:478–486. doi:10.1007/BF00197895

Corley RHV, Tinker PB (2003) The oil palm, 4th edn. BlackwellScience, Oxford

de Touchet B, Duval Y, Pannetier C (1991) Plant regeneration fromembryogenic suspension culture of oil palm (Elaeis guineensisJacq). Plant Cell Rep 10:529–532

Diatchenko L, Lau YF, Campbell AP, Chenchik A, Moqadam F,Huang B, Lukyanov S, Lukyanov K, Gurskaya N, Sverdlov ED,Siebert PD (1996) Suppression subtractive hybridization: amethod for generating differentially regulated or tissue-specificcDNA probes and libraries. Proc Natl Acad Sci USA 93:6025–6030. doi:10.1073/pnas.93.12.6025

Dixon DP, Lapthorn A, Edwards R (2002) Plant glutathionetransferases. Genome Biol 3:reviews3004.1–3004.10

Dong JZ, Dunstan DI (1996) Expression of abundant mRNAs duringsomatic embryogenesis of white spruce [Picea glauca (Moench)Voss]. Planta 199:459–466. doi:10.1007/BF00195740

Duval Y, Engelmann F, Durand-Gasselin T (1995) Somatic embryo-genesis in oil palm (Elaeis guineensis Jacq.). Springer, Berlin

Feher A, Pasternak TP, Dudits D (2003) Transition of somatic plantcells to an embryogenic state. Plant Cell Tissue Organ Cult74:201–228. doi:10.1023/A:1024033216561

Freeman WM, Robertson DJ, Vrana KE (2000) Fundamentals ofDNA hybridization arrays for gene expression analysis. Bio-techniques 29:1042–1046, 1048–1055

Gaj MD (2004) Factors influencing somatic embryogenesis inductionand plant regeneration with particular reference to Arabidopsisthaliana (L.) Heynh. Plant Growth Regul 43:27–47. doi:10.1023/B:GROW.0000038275.29262.fb

Galland R, Randoux B, Vasseur J, Hilbert J (2001) A glutathioneS-transferase cDNA identified by mRNA differential display isupregulated during somatic embryogenesis in Cichorium. Bio-chim Biophys Acta 1522:212–216

Geldner N, Anders N, Wolters H, Keicher J, Kornberger W, Muller P,Delbarre A, Ueda T, Nakano A, Jurgens G (2003) TheArabidopsis GNOM ARF-GEF mediates endosomal recycling,auxin transport, and auxin-dependent plant growth. Cell112:219–230. doi:10.1016/S0092-8674(03)00003-5

Gray DJ (2005) Propagation from nonmeristematic tissues: nonzyg-otic embryogenesis. CRC Press Inc., Boca Raton

Groth D, Lehrach H, Hennig S (2004) GOblet: a platform for GeneOntology annotation of anonymous sequence data. NucleicAcids Res 32:313–317. doi:10.1093/nar/gkh406

Guo WJ, Meetam M, Goldsbrough PB (2008) Examining the specificcontributions of individual Arabidopsis metallothioneins tocopper distribution and metal tolerance. Plant Physiol146:1697–1706. doi:10.1104/pp.108.115782

Helleboid S, Chapman A, Hendriks T, Inze D, Vasseur J, Hilbert JL(2000a) Cloning of ß-1, 3-glucanases expressed during Cicho-rium somatic embryogenesis. Plant Mol Biol 42:377–386. doi:10.1023/A:1006344024877

Helleboid S, Hendriks T, Bauw G, Inze D, Vasseur J, Hilbert JL(2000b) Three major somatic embryogenesis related proteins inCichorium identified as PR proteins. J Exp Bot 51:1189–1200.doi:10.1093/jexbot/51.348.1189

Ho CL, Kwan YY, Choi MC, Tee SS, Ng WH, Lim KA, Lee YP, OoiSE, Lee WW, Tee JM, Tan SH, Kulaveerasingam H, Alwee SS,Abdullah MO (2007) Analysis and functional annotation of

190 Plant Mol Biol (2009) 70:173–192

123

expressed sequence tags (ESTs) from multiple tissues of oil palm(Elaeis guineensis Jacq.). BMC Genomics 8:381. doi:10.1186/1471-2164-8-381

Ikeda-Iwai M, Umehara M, Satoh S, Kamada H (2003) Stress-inducedsomatic embryogenesis in vegetative tissues of Arabidopsisthaliana. Plant J 34:107–114. doi:10.1046/j.1365-313X.2003.01702.x

IminN,NizamidinM, Daniher D, NolanKE, Rose RJ, Rolfe BG (2005)Proteomic analysis of somatic embryogenesis in Medicagotruncatula. Explant cultures grown under 6-benzylaminopurineand 1-naphthaleneacetic acid treatments. Plant Physiol 137:1250–1260. doi:10.1104/pp.104.055277

Joosen R, Cordewener J, Supena ED, Vorst O, Lammers M,Maliepaard C, Zeilmaker T, Miki B, America T, Custers J,Boutilier K (2007) Combined transcriptome and proteomeanalysis identifies pathways and markers associated with theestablishment of rapeseed microspore-derived embryo develop-ment. Plant Physiol 144:155–172. doi:10.1104/pp.107.098723

Jouannic S, Argout X, Lechauve F, Fizames C, Borgel A, Morcillo F,Aberlenc-Bertossi F, Duval Y, Tregear J (2005) Analysis ofexpressed sequence tags from oil palm (Elaeis guineensis).FEBS Lett 579:2709–2714. doi:10.1016/j.febslet.2005.03.093

Legrand S, Hendriks T, Hilbert JL, Quillet MC (2007) Characteriza-tion of expressed sequence tags obtained by SSH during somaticembryogenesis in Cichorium intybus L. BMC Plant Biol 7:27.doi:10.1186/1471-2229-7-27

Lippert D, Zhuang J, Ralph S, Ellis DE, Gilbert M, Olafson R, RitlandK, Ellis B, Douglas CJ, Bohlmann J (2005) Proteome analysis ofearly somatic embryogenesis in Picea glauca. Proteomics5:461–473. doi:10.1002/pmic.200400986

Low ET, Alias H, Boon SH, Shariff EM, Tan CY, Ooi LC, Cheah SC,Raha AR, Wan KL, Singh R (2008) Oil palm (Elaeis guineensisJacq.) tissue culture ESTs: identifying genes associated withcallogenesis and embryogenesis. BMC Plant Biol 8:62. doi:10.1186/1471-2229-8-62

Malik MR, Wang F, Dirpaul JM, Zhou N, Polowick PL, Ferrie AM,Krochko JE (2007) Transcript profiling and identification ofmolecular markers for early microspore embryogenesis inBrassica napus. Plant Physiol 144:134–154. doi:10.1104/pp.106.092932

Malinowski R, Filipecki M (2002) The role of cell wall in plantembryogenesis. Cell Mol Biol Lett 7:1137–1151

Overvoorde PJ, Okushima Y, Alonso JM, Chan A, Chang C, Ecker JR,Hughes B, Liu A, Onodera C, Quach H, Smith A, Yu G, TheologisA (2005) Functional genomic analysis of the AUXIN/INDOLE-3-ACETIC ACID gene family members in Arabidopsis thaliana.Plant Cell 17:3282–3300. doi:10.1105/tpc.105.036723

Pasternak TP, Prinsen E, Ayaydin F, Miskolczi P, Potters G, Asard H,Van Onckelen HA, Dudits D, Feher A (2002) The Role of auxin,pH, and stress in the activation of embryogenic cell division inleaf protoplast-derived cells of alfalfa. Plant Physiol 129:1807–1819. doi:10.1104/pp.000810

Schrick K, Mayer U, Martin G, Bellini C, Kuhnt C, Schmidt J,Jurgens G (2002) Interactions between sterol biosynthesis genesin embryonic development of Arabidopsis. Plant J 31:61–73. doi:10.1046/j.1365-313X.2002.01333.x

Seo PJ, Lee AK, Xiang F, Park CM (2008) Molecular and functionalprofiling of Arabidopsis pathogenesis-related genes: insights intotheir roles in salt response of seed germination. Plant CellPhysiol 49:334–344. doi:10.1093/pcp/pcn011

Sharma SK, Millam S, Hedley PE, McNicol J, Bryan GJ (2008)Molecular regulation of somatic embryogenesis in potato: anauxin led perspective. Plant Mol Biol, 16 Jun 2008. Epub aheadof print

Singla B, Tyagi AK, Khurana JP, Khurana P (2007) Analysis ofexpression profile of selected genes expressed during auxin-

induced somatic embryogenesis in leaf base system of wheat(Triticum aestivum) and their possible interactions. Plant MolBiol 65:677–692. doi:10.1007/s11103-007-9234-z

Stasolla C, van Zyl L, Egertsdotter U, Craig D, Liu W, Sederoff RR(2003) The effects of polyethylene glycol on gene expression ofdeveloping white spruce somatic embryos. Plant Physiol 131:49–60. doi:10.1104/pp.015214

Stasolla C, Bozhkov PV, Chu TM, Van Zyl L, Egertsdotter U, SuarezMF, Craig D, Wolfinger RD, Von Arnold S, Sederoff RR (2004)Variation in transcript abundance during somatic embryogenesisin gymnosperms. Tree Physiol 24:1073–1085

Su N, He K, Jiao Y, Chen C, Zhou J, Li L, Bai S, Li X, Deng XW(2007) Distinct reorganization of the genome transcriptionassociates with organogenesis of somatic embryo, shoots, androots in rice. Plant Mol Biol 63:337–349. doi:10.1007/s11103-006-9092-0

Takeda H, Kotake T, Nakagawa N, Sakurai N, Nevins DJ (2003)Expression and function of cell wall-bound cationic peroxidasein asparagus somatic embryogenesis. Plant Physiol 131:1765–1774. doi:10.1104/pp.102.014654

Thibaud-Nissen F, Shealy RT, Khanna A, Vodkin LO (2003)Clustering of microarray data reveals transcript patterns associ-ated with somatic embryogenesis in soybean. Plant Physiol132:118–136. doi:10.1104/pp.103.019968

Thomann A, Dieterle M, Genschik P (2005a) Plant CULLIN-basedE3s: phytohormones come first. FEBS Lett 579:3239–3245. doi:10.1016/j.febslet.2005.02.068

Thomann A, Brukhin V, Dieterle M, Gheyeselinck J, Vantard M,Grossniklaus U, Genschik P (2005b) Arabidopsis CUL3A andCUL3B genes are essential for normal embryogenesis. Plant J43:437–448. doi:10.1111/j.1365-313X.2005.02467.x

Tregear JW, Morcillo F, Richaud F, Berger A, Singh R, Cheah SC,Hartmann C, Rival A, Duval Y (2002) Characterization of adefensin gene expressed in oil palm inflorescences: inductionduring tissue culture and possible association with epigeneticsomaclonal variation events. J Exp Bot 53:1387–1396. doi:10.1093/jexbot/53.373.1387

Tsuwamoto R, Fukuoka H, Takahata Y (2007) Identification andcharacterization of genes expressed in early embryogenesis frommicrospores of Brassica napus. Planta 225:641–652. doi:10.1007/s00425-006-0388-8

vanZyl L,BozhkovPV,ClaphamDH,SederoffRR, vonArnoldS (2003)Up, down and up again is a signature global gene expression patternat the beginning of gymnosperm embryogenesis. Gene ExprPatterns 3:83–91. doi:10.1016/S1567-133X(02)00068-6

Verdeil JL, Alemanno L, Niemenak N, Tranbarger TJ (2007)Pluripotent versus totipotent plant stem cells: dependence versusautonomy? Trends Plant Sci 12:245–252. doi:10.1016/j.tplants.2007.04.002

Vidali L, Augustine RC, Kleinman KP, Bezanilla M (2007) Profilin isessential for tip growth in the moss Physcomitrella patens. PlantCell 19:3705–3722. doi:10.1105/tpc.107.053413

von Arnold S, Sabala I, Bozhkov P, Dyachok J, Filonova L (2002)Developmental pathways of somatic embryogenesis. Plant CellTissue Organ Cult 69:233–249. doi:10.1023/A:1015673200621

Vrinten PL, Nakamura T, Kasha KJ (1999) Characterization ofcDNAs expressed in the early stages of microspore embryogen-esis in barley (Hordeum vulgare) L. Plant Mol Biol 41:455–463.doi:10.1023/A:1006383724443

Willemsen V, Friml J, Grebe M, van den Toorn A, Palme K, ScheresB (2003) Cell polarity and PIN protein positioning in Arabid-opsis require STEROL METHYLTRANSFERASE1 function.Plant Cell 15:612–625. doi:10.1105/tpc.008433

Xu J, Scheres B (2005) Dissection of Arabidopsis ADP-RIBOSYL-ATION FACTOR 1 function in epidermal cell polarity. PlantCell 17:525–536. doi:10.1105/tpc.104.028449

Plant Mol Biol (2009) 70:173–192 191

123

Zeng F, Zhang X, Zhu L, Tu L, Guo X, Nie Y (2006) Isolation andcharacterization of genes associated to cotton somatic embryo-genesis by suppression subtractive hybridization and macroarray.Plant Mol Biol 60:167–183. doi:10.1007/s11103-005-3381-x

Zhong R, Burk DH, Nairn CJ, Wood-Jones A, Morrison WHIII, YeZH (2005) Mutation of SAC1, an Arabidopsis SAC domain

phosphoinositide phosphatase, causes alterations in cell mor-phogenesis, cell wall synthesis, and actin organization. Plant Cell17:1449–1466. doi:10.1105/tpc.105.031377

Zimmerman JL (1993) Somatic embryogenesis: a model for earlydevelopment in higher plants. Plant Cell 5:1411–1423

192 Plant Mol Biol (2009) 70:173–192

123