Isolation of an embryogenic line from non-embryogenic Brassica napus cv. Westar through microspore embryogenesis

Abstract

Brassica napus cultivar Westar is non-embryogenic under all standard protocols for induction of microspore embryogenesis; however, the rare embryos produced in Westar microspore cultures, induced with added brassinosteroids, were found to develop into heritably stable embryogenic lines after chromosome doubling. One of the Westar-derived doubled haploid (DH) lines, DH-2, produced up to 30% the number of embryos as the highly embryogenic B. napus line, Topas DH4079. Expression analysis of marker genes for embryogenesis in Westar and the derived DH-2 line, using real-time reverse transcription-PCR, revealed that the timely expression of embryogenesis-related genes such as LEAFY COTYLEDON1 (LEC1), LEC2, ABSCISIC ACID INSENSITIVE3, and BABY BOOM1, and an accompanying down-regulation of pollen-related transcripts, were associated with commitment to embryo development in Brassica microspores. Microarray comparisons of 7 d cultures of Westar and Westar DH-2, using a B. napus seed-focused cDNA array (10 642 unigenes), identified highly expressed genes related to protein synthesis, translation, and response to stimulus (Gene Ontology) in the embryogenic DH-2 microspore-derived cell cultures. In contrast, transcripts for pollen-expressed genes were predominant in the recalcitrant Westar microspores. Besides being embryogenic, DH-2 plants showed alterations in morphology and architecture as compared with Westar, for example epinastic leaves, non-abscised petals, pale flower colour, and longer lateral branches. Auxin, cytokinin, and abscisic acid (ABA) profiles in young leaves, mature leaves, and inflorescences of Westar and DH-2 revealed no significant differences that could account for the alterations in embryogenic potential or phenotype. Various mechanisms accounting for the increased capacity for embryogenesis in Westar-derived DH lines are considered.

Introduction

Plant breeding is facilitated by the rapid development of doubled haploid (DH) plant lines, homozygous at all loci. This is most readily accomplished through microspore embryogenesis, where freshly isolated uninucleate or binucleate microspores (male gametophytes) are induced in culture to shift from a gametophytic or pollen pathway to embryo development. After chromosome doubling, the resultant plant lines are maintained as breeding stocks. Successful and efficient microspore-derived embryo production is genotype dependent, and much labour is consumed in developing and optimizing tissue culture conditions. Homozygous DH lines have been integrated into breeding programmes for superior varieties of canola (Brassica spp.), barley (Hordeum vulgare L.), maize (Zea mays), wheat (Triticum aestivum L.), rice (Oryza sativa L.), pepper (Capsicum annum), asparagus (Asparagus officinalis), and several grasses such as Lolium and Festuca spp. (Thomas et al., 2003; Kopecky et al., 2005; Forster et al., 2007). Some selected DH lines also have been used to produce commercial hybrids, or directly as commercial lines (Forster et al., 2007). In addition, DH lines have become an extremely important resource for chromosome mapping studies and also facilitate selection of recessive or polygenic traits (Forster and Thomas, 2005).

Brassica napus is a model system for studies of microspore embryogenesis; however, not all genotypes respond equally well to inducing culture conditions. Brassica napus cv. Topas embryogenic line DH4079 is one of the most responsive genotypes, and >10% of cultured microspores form embryos (Ferrie, 2003). Some other genotypes of B. napus, for example cv. Allons, Garrison, and Westar, are poorly embryogenic, and <0.5% of cultured microspores form embryos using the standard B. napus protocol for microspore embryogenesis (Ferrie et al., 2005). A poor embryogenic response limits the utility of desirable cultivars in breeding programs.

Isolated microspores of suitable cultivars of B. napus can be induced to form embryos in vitro with appropriate culture media and stress treatments, for example heat (32 °C) or osmotic stress (polyethylene glycol) (Ferrie, 2003; Ferrie and Keller, 2007). Several factors are known to be critical for the optimum response of cultured microspores, including donor plant growth conditions, culture conditions, media, genotype, and age of the donor plants. Quantitative trait locus (QTL) analysis in Brassica cultivars for regeneration potential from protoplast cultures has revealed a low number of genes involved in this process (Holme et al., 2004). Zhang and Takahata (2001) have reported that microspore embryogenic ability is controlled by two multiple gene loci in B. napus. Also, random amplified polymorphic DNA (RAPD) markers with additive effects have been linked to microspore embryogenic ability in Chinese cabbage and oilseed rape (Zhang et al., 2003).

Improvements in microspore embryogenesis for poorly embryogenic cultivars of B. napus (i.e. cv. Westar) have been reported following additions of 24-epibrassinolide (EBR) or brassinolide to the culture medium (Ferrie et al., 2005). In addition, there are numerous reports in the literature of positive changes in embryogenic response in plant species and cultivars due to alterations in either donor plant conditions, media composition, or culture conditions (Li and Devaux, 2001; Croser et al., 2006; Kim and Moon, 2007); however, there have been no reports, or molecular studies, of cases of induced, stable, and heritable improvements in embryogenic potential in DH plant lines resulting from such manipulation of previously non-embryogenic, parental material. In the case described here, the rare embryos produced in the Westar microspore cultures following a 3 d exposure to brassinosteroid-supplemented media developed into heritably stable embryogenic lines after chromosome doubling (DH-1, DH-2, DH-3 and DH-4). Embryo development, transformation efficiency, gene expression profiles, hormone concentrations, and phenotypic differences have been compared in the non-embryogenic B. napus cv. Westar parental line and a Westar-derived embryogenic line, DH-2. The molecular characterization utilized a set of well-established marker genes for embryogenesis (Malik et al., 2007) and a newly developed Brassica seed-focused cDNA array (10 642 unigenes; Xiang et al., 2008).

Materials and methods

Plant material

Plants of B. napus cv. Westar were grown in 15 cm pots in a growth cabinet with a 16 h/8 h day/night photoperiod, light intensity of 400 μmol m⁻² s⁻¹, and day/night temperatures of 20 °C/15 °C. Following flower bud formation, and in preparation for microspore culture, the day/night temperatures were lowered to 10 °C/5 °C. Microspore collections and cultures were initiated as described by Ferrie and Keller (1995). For comparison purposes, it should be noted that microspores collected from 100 buds were sufficient for ∼20 culture plates. Embryogenesis was induced in microspores isolated at the late-uninucleate to early-binucleate stage (Ferrie and Keller, 1995) using a defined medium containing 13% sucrose and heat stress at 32 °C for 3 d. EBR (OlChemIm) at 10⁻⁶ M was added to the medium to improve the response in the poorly embryogenic cultivar Westar (Ferrie et al., 2005), and several of the resulting microspore-derived embryos from those cultures were regenerated into plants. The plantlets were treated with 0.34% colchicine solution for chromosome doubling in order to ensure recovery of DH plants. The colchicine-treated plantlets were transferred to soil, grown under the conditions described above, and used for subsequent microspore cultures. To assess embryogenic potential, microspores isolated from the Westar-derived DH lines (DH-1, DH-2, DH-3 and DH-4) and Westar were cultured on standard NLN-13% sucrose medium without the addition of brassinosteroids. For in-depth studies, microspore cultures of the DH-2 line and Westar were observed at 1, 3, 5, 7, 14, and 21 d for embryo development. The 5 d and 7 d microspore cultures from both lines were collected by centrifugation and stained with 1% acetocarmine to observe divisions during early embryogenesis. Light microscopy images were captured on a Leica DMR microscope.

Transformation

Agrobacterium tumefaciens strain GV3101:pMP90, carrying binary vector pHS723 which includes a β-glucuronidase–neomycin phosphotransferase (GUS–NPT) translational fusion driven by an enhanced 35S promoter with a nopaline synthase (NOS) terminator (Datla et al., 1991; Nair et al., 2000), was used for transformation. Hypocotyl explants were prepared from 5-d-old seedlings of Westar and the Westar-derived DH lines (DH-1, DH-2, DH-3 and DH-4), pre-cultured, and inoculated with Agrobacterium according to DeBlock et al. (1989), with modifications by Zou et al. (1997). Selection was carried out on shoot induction medium with 20 mg l⁻¹ kanamycin. Green regenerants were tested for GUS activity by incubating leaf pieces in X-Gluc substrate (Jefferson et al., 1986).

RNA isolation

Total RNA from 0 h, 1, 3, 5, and 7 d cultured microspores was isolated using the RNeasy Midi kit (Qiagen), including on-column DNase digestion. For semi-quantitative RT-PCR, 3 μg of total RNA was used for first-strand cDNA synthesis with oligo(dT)₁₆ primers (DNA Sequencing Lab, NRC-Plant Biotechnology Institute) and SUPERSCRIPT II reverse transcriptase (Invitrogen) according to the manufacturer's protocol. Gene-specific primer pairs were designed using Primer 3 software (http://gene.pbi.nrc.ca/cgi-bin/primer/primer3_www.cgi) to obtain PCR products that were 350–550 bp in length. PCRs were one cycle at 95 °C for 5 min and 30 cycles at 95 °C for 30 s, 55 °C for 30 s, 72 °C for 45 s using 0.6 μl of template cDNA from the first-strand cDNA synthesis reaction.

Expression analyses by real-time RT-PCR

Total RNA (150 ng) from each tissue, developmental stage, or cultivar was used for one-step real-time reverse transcription-PCR (RT-PCR) analyses using the QuantiTect SYBR Green RT-PCR Kit (Qiagen Inc.) and gene-specific primers. Primer pairs were designed using the Primer Quest software (Integrated DNA Technologies) to give PCR products from 100 to 400 bp. Real-time RT-PCR was performed on an Mx3000P™ Real-time PCR system (Stratagene, La Jolla, CA, USA). Relative expression was calculated using the 2^−ΔΔCT method (Livak and Schmittgen, 2001) with 18S rRNA as the internal control gene. The cycling parameters were one cycle at 50 °C for 60 min (reverse transcription reaction), one cycle at 95 °C for 15 min, then 35 cycles of 95 °C for 30 s, 58 °C for 1 min, 72 °C for 45 s. Primer sequences were described in a previous publication (Malik et al., 2007).

Microarray analysis using a B. napus seed-focused cDNA array (Bn10K)

For microarray analyses, the 7 d microspore cultures of Westar and the Westar-derived DH-2 line were size-selected on a mesh screen (Sefar; pore size 35 μm) in order to collect the microspores and/or cell clusters that had enlarged, and to discard the smaller physiologically unresponsive microspores. A 10 μg aliquot of total RNA from each sample was labelled with Cy3 or Cy5 (Amersham) and hybridized to the Brassica seed-focused cDNA microarray (Bn10K; http://brassicagenomics.ca) (Xiang et al., 2008) using the Pronto!™ Plus indirect system according to the manufacturer's instructions (Corning). The slides were scanned using a ScanArray 4000 laser scanner at a resolution of 10 μm. The image analysis and signal quantification were done with QUANTARRAY (GSI Lumonics, Watertown, MA, USA). Data storage and preliminary data processing, including normalization, were done on BioArray Software Environment (BASE 1.0; Saal et al., 2002). Background-subtracted signals were used to identify differentially expressed genes with Significance Analysis of Microarrays (SAM 2.0; Tusher et al., 2001).

Hormone profiling

One gram fresh weight tissue samples from young unexpanded leaves, mature leaves, and inflorescences bearing buds, with 3–4 open flowers, of matched samples of Westar and DH-2 were collected, frozen in liquid nitrogen, and freeze-dried. Three biological replicates (50 mg dry weight) were used for hormone profiling of abscisic acid (ABA) and related metabolites, auxins, and cytokinins using the protocol described in Chiwocha et al. (2005).

Results

Embryogenic potential of the DH lines

The DH lines (DH-1, DH-2, DH-3 and DH-4) were developed from four embryos randomly selected from Westar microspore cultures that had been previously treated with EBR to improve embryogenesis (Ferrie et al., 2005). In the absence of an EBR treatment, embryo development from Westar-derived microspores under standard inducing conditions was extremely rare (Table 1). The selected embryos were regenerated to plantlets, grown to flowering, and microspore embryogenesis was re-initiated in the DH lines. Surprisingly, all four of the lines had greatly improved rates of microspore embryogenesis as compared with the non-embryogenic parental line Westar (Table 1). The embryogenic response was variable among the lines, but always was consistently better than the Westar parent. The most embryogenic line, DH-2, produced thousands of embryos under standard culture conditions where Westar might only produce one embryo, and in this respect it was 30% as productive as a previously characterized highly embryogenic line, B. napus DH4079 (Ferrie et al., 2005; Malik et al., 2007). The DH lines were selfed (to maintain the original homozygosity) and F₂ seed was collected separately for each line. Plants (from F₂ seed) of the most embryogenic Westar-derived line (DH-2; Table 1) were used as donor plants for further studies of microspore embryogenesis (F₃ material). The improved embryogenic potential of the DH-2 line remained stable through at least these two subsequent generations (F₂ and F₃). Molecular and phenotypic comparisons were made between the non-embryogenic parental line, Westar, and the embryogenic Westar-derived DH-2 line.

Table 1.

Embryogenic response and transformation potential in four independently selected Westar-derived lines

B. napus, CV and line	Embryogenic responsea	Transformation response
		No. of explants	Green shoots	No. of GUS positives	Transformation efficiency (%)
Westar	1±0	510	72	39	7.6
Westar DH-1	40±11.9	484	49	24	3.9
Westar DH-2	2235.3±1366.5	850	8	3	0.4
Westar DH-3	131.0±99.5	510	33	17	3.3
Westar DH-4	685.3±657.4	96	26	1	1.0

Embryogenesis was assessed as the average frequency of embryos per plate from three replicate experiments from microspore cultures of Brassica napus cv. Westar and Westar-derived DH (doubled haploid) lines. On average there are ∼1 000 000 microspores transferred to each plate. Transformation success was assessed by determining the number of independent GUS-positive shoots relative to the number of original Agrobacterium-treated explants.

^aMean ±SE.

At the time of isolation (0 h), and the beginning of culture, the microspores of both lines had a prominent nucleus and the cytoplasm was restricted to the periphery. In a previous study it was noted that enlargement of the microspore was a prominent feature of 3 d heat-stress-induced embryogenic microspores of B. napus Topas DH4079, and that cell and nuclear divisions were predominant in the 5 d and 7 d induced microspores, with only marginal increases in size occurring during this latter period (Malik et al., 2007). No differences were noted between the microspores of the Westar parental line and the Westar-derived DH-2 line during the first 3 d of culture (data not shown). Following a heat stress treatment for embryo induction (3 d at 32 °C), responding Westar and DH-2 microspores enlarged similarly, to more than double their original size, and the numbers of enlarged microspores in the two lines were equivalent (∼30% of cultured microspores); however, by 5 d the induced microspores of the DH-2 line had undergone a first symmetric mitotic division (Fig. 1). In contrast, the enlarged microspores of the Westar line did not divide, although they were swollen and stained bright red with 1% acetocarmine (Fig. 1). Thereafter, the induced microspores of the DH-2 line continued to undergo random cell divisions, and by 7 d the embryogenic microspores of the DH-2 line appeared as cell clusters (globular and pre-globular embryos) with remnants of the ruptured exine still remaining on the developing embryo (Fig. 1). There were very few dividing structures in 7 d Westar cultures, and some of those dividing structures already had started to degenerate (Fig. 1).

Fig. 1.

Microspore-derived embryo development in Brassica napus cv. Westar and the Westar-derived DH-2 line. (A) Acetocarmine-stained 5 d enlarged microspores. Arrowheads indicate divisions in the microspores of the DH-2 line. (B) Acetocarmine-stained 7 d enlarged microspores in Westar and a dividing pre-globular embryo in the DH-2 line. (C) Twenty-one day mid-maturation stage embryos in the DH-2 line; no embryos developed in this microspore culture plate of Westar. Black bars=10 μm, white bars=35 μm.

Although B. napus is a model system for microspore embryogenesis, there are very few lines that are both embryogenic and transformable, for example B. napus DH12075 (Li et al., 2003) and cv. Lisandra (Fukuoka et al., 1998; Zhang et al., 2003), and thus appropriate for both genetic and reverse genetics applications. Westar is easily transformable (Cardoza and Stewart, 2004) but not highly embryogenic (Table 1; Ferrie et al., 2005), while B. napus Topas DH4079 is highly embryogenic (Malik et al., 2007), but not transformable using standard techniques (J Hammerlindl, unpublished data). Therefore, it was of interest to examine transformability of the Westar-derived DH lines (Table 1). Some of the lines were easily transformed with Agrobacterium (e.g. DH-1 and DH-3); however, the most embryogenic line (DH-2) was the least transformable of the lines developed (Table 1). This line also showed restricted organogenesis (shoot regeneration) and, therefore, perhaps transformability was limited by and could be improved with some focus on regeneration and shoot production.

Expression of embryogenesis-related genes

Previously, several clusters of differentially expressed genes marking the developmental transitions from freshly isolated microspores (0 h) to committed 7 d embryogenic cell clusters (globular and pre-globular embryos), as well as a set of 16 unambiguous marker genes for the induction of microspore embryogenesis, were identified (Malik et al., 2007). These molecular marker genes were not expressed in microspores at the time of culture (0 h) in the highly embryogenic B. napus line Topas DH4079, nor in microspores cultured under non-inductive conditions (18 °C), and thus can be used both quantitatively and qualitatively to measure a cultivar's responsiveness to embryogenesis-inducing conditions (Malik et al., 2007). Based on real-time RT-PCR, the transcript abundance for the marker gene BnLEC1 was not significantly different between the two lines at 5 d and 7 d (Fig. 2), confirming the acquisition of a certain level of embryogenic competence in cv. Westar (Fig. 1). However, BnLEC1 expression in Westar was markedly delayed at 1 d and 3 d as compared with the DH-2 cultures (Fig. 2). Expression of the other marker genes, namely BnLEC2, BnABI3, BnBBM1, BnUP1, and BnWOX9, was uniformly much reduced at all time points in Westar cultures as compared with DH-2 cultures (Fig. 2), reflecting the differences in embryogenic potential between the two lines. There were no significant differences in the level of expression of BnSERK1 between Westar and DH-2 during the early stages of microspore culture (Fig. 2), despite the marked differences in embryogenic potential.

Fig. 2.

Real time RT-PCR analyses of embryo-specific marker genes (BnLEC1, BnLEC2, BnABI3, BnBBM1, BnUP1, BnWOX9, and BnWOX2) and BnSERK1 in microspore cultures of non-embryogenic B. napus cv. Westar and the embryogenic Westar-derived DH-2 line. Stages of microspore-derived embryo (MDE) development (0 h, 1, 3, 5, and 7 d) are indicated for each of the lines (Westar, DH-2). Expression was calculated according to the 2^-ΔΔCT method (Livak and Schmittgen, 2001). Relative expression was based on comparisons with transcript levels in 0 h microspores of cv. Westar with 18S rRNA as the internal control for normalization.

Expression of pollen-related genes

Putative pollen-expressed genes have been identified from cDNA libraries made from 3 d and 5 d embryogenic cultures and cDNA libraries representing in vitro pollen (Joosen et al., 2007; Malik et al., 2007; M R Malik et al., unpublished data). Subsequent RT-PCR analyses have confirmed a number of predominantly pollen-expressed genes, for example BnPK12, BnCDPK, BnLEA1, BnPK21, and BnUP5 (see Supplementary Table S1 available at JXB online). Expression analyses by real-time RT-PCR showed at least a 2-fold greater expression of these pollen-related genes in the recalcitrant Westar cultivar, particularly at the 3 d and 5 d stages of microspore culture, as compared with the embryogenic Westar-derived DH-2 line (Fig. 3). Although the expression of these pollen-related genes diminished in cultures of both lines by 7 d, expression was still higher in the Westar microspores (Fig. 3).

Fig. 3.

Real time RT-PCR analyses of pollen-specific genes BnPK12 (At3g18810), BnCDPK (At2g31500), BnLEA1 (At4g13230), BnPK21 (At2g24370), and BnUP5 in microspore cultures of non-embryogenic B. napus cv. Westar and the embryogenic Westar-derived DH-2 line (the closest Arabidopsis match for each B. napus pollen-specific gene is given in parentheses). Stages of microspore-derived embryo (MDE) development (0 h, 1, 3, 5, and 7 d) are indicated for each of the lines (Westar, DH-2). Expression was calculated according to the 2^−ΔΔCT method (Livak and Schmittgen, 2001). Relative expression was based on comparisons with transcript levels in 0 h microspores of cv. Westar with 18S rRNA as the internal control for normalization.

Transcript profiling using the Bn10K seed cDNA microarray

Microarray analyses of gene expression using the B. napus 10K seed cDNA array (Xiang et al., 2008) identified significant expression of 637 genes in 7 d enlarged microspores of Westar and 456 genes in 7 d enlarged and dividing microspores of the DH-2 line (both samples >35 μm) and, of these, 314 genes that were expressed in both 7 d Westar and DH-2 (signal intensity >500 in two or more replicates). Using the software program SAM (Tusher et al., 2001), 117 differentially expressed genes were identified that mark the developmental differences between Westar and Westar-derived DH-2 (77 genes up-regulated in 7 d enlarged and dividing DH-2 microspores, 40 genes up-regulated in 7 d enlarged Westar microspores) (Fig. 4). Many of the genes up-regulated in the 7 d enlarged and dividing DH-2 microspores were related to either protein biosynthesis, response to stimulus, or cellular transport (Table 2). Notable amongst these are genes encoding 40S (RPS2C, RPS15A, RPS9B, RPSaA, RPS9B, RPS17A) and 60S (RPL10aB, RPL8A, RPL21E, RPP1B, RPL23C, RPL28C, RPL30C, RPL36aA, RPL3A) ribosomal proteins, elongation factor 2, heat shock cognate 70 kDa protein1 (HSC70-1), lipid transfer proteins, and seed storage proteins (Table 2). No differences were detected on the microarray in the levels of expression of the transcription factor genes LEC1 and LEC2 between the two lines, although these differences were demonstrated by real-time RT-PCR (Fig. 2), perhaps because microarray expression levels per se were very low for these genes (data not shown). Nonetheless, some known targets of the LEC2 transcription factor were up-regulated and differentially expressed in the 7 d enlarged and dividing microspores of the DH-2 line, for example 2S seed storage protein 1 (At4g27140), oleosin (At4g25140), and cysteine proteinase (At3g54940) (Table 2; Braybrook et al., 2006).

Fig. 4.

Microarray analysis of differentially expressed genes between 7 d enlarged (dividing) embryogenic microspores of DH-2 and 7 d induced (non-dividing) microspores of the parental line, Westar. Labelled total RNA (10 μg) was used for hybridization to the Bn10K seed cDNA array. Signal intensities were normalized and gene lists extracted using SAM (minimum 1.5-fold change in expression). A total of 117 differentially expressed genes were identified: 77 genes up-regulated in DH-2 and 40 genes up-regulated in Westar (negative log2 values). Gene identifications are listed in Tables 2 and 3.

Table 2.

Genes up-regulated in 7 d microspore cultures of embryogenic Westar-derived DH-2 line

Accession no.	Genes up-regulated in Westar DH-2	Best match to Arabidopsis	E-value	Biological process	Broad functional category
EE541057	60S ribosomal protein L32 (RPL32A)	AT4G18100.1	4e-71	Ribosome biogenesis and assembly	Cell organization and biogenesis
CN735630	Histone H2B, putative	AT2G37470.1	1e-41	Chromosome organization and biogenesis	Cell organization and biogenesis
DY010080	60S ribosomal protein L5 (RPL5B)	AT5G39740.1	e-123	Ribosome biogenesis and assembly	Cell organization and biogenesis
EE542594	60S ribosomal protein L37a (RPL37aC)	AT3G60245.1	3e-49	Ribosome biogenesis and assembly	Cell organization and biogenesis
EE541600	MEE26 (maternal effect embryo arrest 26)	AT2G34870.1	4e-9	Embryonic development ending in seed dormancy	Developmental processes
DY009433	EMBRYO DEFECTIVE 2386; identical to 60S ribosomal protein L19-1 (RPL19A)	AT1G02780.1	1e-93	Embryonic development ending in seed dormancy; translation	Developmental processes; protein metabolism
CN734060	EMBRYO DEFECTIVE 2171; 60S ribosomal protein L23 (RPL23A)	AT3G04400.1	2e-76	Embryonic development ending in seed dormancy; translation	Developmental processes; protein metabolism
CN727564	EMBRYO DEFECTIVE 2171; 60S ribosomal protein L23 (RPL23A)	AT3G04400.1	4e-77	Embryonic development ending in seed dormancy	Developmental processes
EE542973	Histone H2A, putative	AT5G59870.1	8e-52	Nucleosome assembly	DNA or RNA metabolism
EE550299	NAD2B; encodes subunit of mitochondrial NAD(P)H dehydrogenase	ATMG01320.1	3e-23	Electron transport	Electron transport or energy pathways
EE550879	COB; mitochondrial apocytochrome b	ATMG00220.1	e-114	Aerobic respiration	Electron transport or energy pathways
DY007371	Mitochondrial NADH dehydrogenase 5	ATMG00665.1	7e-41	Electron transport	Electron transport or energy pathways
CN732202	LIPID TRANSFER PROTEIN 3	AT5G59320.1	1e-52	Response to abscisic acid stimulus	Other biological processes
EE548239	Fructose-bisphosphate aldolase, putative	AT3G52930.1	e-117	Pentose-phosphate shunt	Other cellular processes
CN737059	GLUTATHIONE S-TRANSFERASE 29, ATGSTU18	AT1G10360.1	4e-72	Toxin catabolic process	Other cellular processes
EE541185	MATK; encodes a maturase located in the trnK intron in the chloroplast genome	ATCG00040.1	8e-70	RNA splicing	Other cellular processes
CN727662	GAPC-2; glyceraldehyde-3-phosphate dehydrogenase	AT1G13440.1	4e-72	Gluconeogenesis	Other cellular processes
CN733306	THI1 (THIAZOLE REQUIRING)	AT5G54770.1	7e-96	Thiamine biosynthetic process	Other cellular processes
EE462458	ATP synthase beta chain 2	AT5G08690.1	1e-52	ATP biosynthetic process	Other cellular processes
CN726197	OLEO1 (OLEOSIN1)	AT4G25140.1	6e-62	Sequestering of lipid	Other metabolic processes
EE548291	TPI; ATCTIMC (CYTOSOLIC TRIOSEPHOSPHATE ISOMERASE	AT3G55440.1	1e-97	Metabolic process	Other metabolic processes
DY010101	Calmodulin binding/elongation factor 1-alpha/EF-1-alpha	AT5G60390.2	e-138	Translational elongation	Protein metabolism
No Acc. No.	Elongation factor 1-alpha/EF-1-alpha	AT1G07920		Translational elongation	Protein metabolism
EE551198	RPS3; ribosomal protein S3	ATMG00090.1	e-127	Translation	Protein metabolism
EE461044	RPS3; encodes a chloroplast ribosomal protein S3	ATCG00800.1	4e-68	Translation	Protein metabolism
EE439665	Elongation factor 1-alpha/EF-1-alpha	AT5G60390.1	7e-94	Translational elongation	Protein metabolism
EE542027	Elongation factor 1-alpha/EF-1-alpha	AT5G60390.1	e-133	Translational elongation	Protein metabolism
EE550930	CLPP1; encodes the only ClpP (caseinolytic protease) encoded within the plastid genome	ATCG00670.1	2e-38	Proteolysis	Protein metabolism
CN725880	Cysteine proteinase, putative	AT3G54940.3	3e-85	Proteolysis	Protein metabolism
CN730121	40S ribosomal protein S4 (RPS4B)	AT5G07090.1	e-114	Translation	Protein metabolism
EE543924	RPS7, RPS7.1; encodes a chloroplast ribosomal protein S7	ATCG00900.1	1e-82	Translation	Protein metabolism
EE542577	60S ribosomal protein L36 (RPL36B)	AT3G53740.4	7e-50	Translation	Protein metabolism
EE550742	RPS6 (RIBOSOMAL PROTEIN S6)	AT4G31700.1	4e-95	Translation	Protein metabolism
CN737473	RPS15 (RIBOSOMAL PROTEIN S15)	AT1G04270.1	3e-82	Translation	Protein metabolism
CX270671	RPL2.2; encodes a chloroplast ribosomal protein L2	ATCG01310.1	1e-67	Translation	Protein metabolism
CN726155	60S ribosomal protein L10A (RPL10aB)	AT2G27530.2	e-102	Translation	Protein metabolism
EE542556	60S ribosomal protein L4/L1 (RPL4A)	AT3G09630.1	5e-89	Translation	Protein metabolism
CN737233	60S ribosomal protein L29 (RPL29B)	AT3G06680.1	7e-31	Translation	Protein metabolism
EE551248	RPL14; encodes a chloroplast ribosomal protein L14	ATCG00780.1	4e-43	Translation	Protein metabolism
EE543194	CCB452; cytochrome c biogenesis orf452	ATMG00180.1	5e-61	Translation	Protein metabolism
DY009980	RPL2.2; encodes a chloroplast ribosomal protein L2	ATCG01310.1	e-102	Translation	Protein metabolism
CN731939	60S acidic ribosomal protein P1	AT4G00810.2	2e-29	Translational elongation	Protein metabolism
EE541984	RPS11-BETA (putative ribosomal protein S11-beta)	AT5G23740.1	6e-75	Translation	Protein metabolism
EE541830	60S ribosomal protein L10 (RPL10C)	AT1G66580.1	e-125	Translation	Protein metabolism
CN735269	Eukaryotic translation initiation factor 5A, putative/eIF-5A, putative	AT1G26630.1	3e-70	Translational initiation	Protein metabolism
EE569706	ATP1; ATPase subunit 1	ATMG01190.1	e-142	Response to oxidative stress	Response to stress
EE462567	CRT1 (CALRETICULIN 1); calcium ion binding	AT1G56340.2	e-132	Response to oxidative stress	Response to stress
DY013565	HSP70-1, HEAT SHOCK COGNATE 70 KDA PROTEIN 1	AT5G02500.1	e-116	Response to cold	Response to stress
CN730192	AT1G56075.1, LOS1 (low expression of osmotically responsive genes 1)	AT1G56070.1	7e-52	Response to cold	Response to stress
EE439511	AT1G56075.1, LOS1 (low expression of osmotically responsive genes 1)	AT1G56070.1	e-144	Response to cold	Response to stress
CN737456	60S ribosomal protein L26 (RPL26A)	AT3G49910.1	4e-58	Response to cold	Response to stress
EE541625	ATCYP1, ROC5 (ROTAMASE CYP 5)	AT4G34870.1	5e-68	Aignal transduction	Signal transduction
EE541128	Protease inhibitor/seed storage/lipid transfer protein (LTP) family protein	AT5G38195.1	2e-32	Lipid transport	Transport
EE541584	LTP12 (LIPID TRANSFER PROTEIN 12)	AT3G51590.1	3e-51	Lipid transport	Transport
DY003849	2S seed storage protein 1	AT4G27140.1	1e-46	Lipid transport	Transport
DY013108	LTP2 (LIPID TRANSFER PROTEIN 2).	AT2G38530.1	2e-45	Phospholipid transfer to membrane	Transport
CN735143	LTP5 (LIPID TRANSFER PROTEIN 5)	AT3G51600.1	2e-52	Lipid transport	Transport
EE543261	Protease inhibitor/seed storage/lipid transfer protein (LTP) family protein	AT1G55260.1	2e-73	Lipid transport	Transport
CN728941	AAC2 (ADP/ATP CARRIER 2)	AT5G13490.2	1e-67	Transport	Transport
EE550070	Unknown protein	AT1G49310.1	3e-23	Biological process unknown	Unknown biological processes
EE541862	Unknown protein	AT1G49310.1	9e-7	Biological process unknown	Unknown biological processes
EE543031	BURP domain-containing protein	AT1G49320.1	e-101	Biological process unknown	Unknown biological processes
EE548292	Glycine-rich protein	AT3G24250.1	9e-8	Biological process unknown	Unknown biological processes
CX271266	BURP domain-containing protein	AT1G49320.1	5e-21	Biological process unknown	Unknown biological processes
ES265407	Unknown protein	AT1G75870.1	4e-29	Biological process unknown	Unknown biological processes
EE549026	Glycine-rich protein	AT2G30560.1	2e-10	Biological process unknown	Unknown biological processes
EE542023	UInknown protein	AT3G06090.1	2e-7	Biological process unknown	Unknown biological processes
EE550161	Similar to unknown protein (Arabidopsis thaliana)	AT1G49290.1	3e-9	Biological process unknown	Unknown biological processes
EE548723	No hits found
EE569070	No hits found
EE569674	No hits found
CN735696	No hits found
CN730585	No hits found
EE544345	No hits found
CN726406	No hits found
DY011515	No hits found
EE569173	No hits found

Locus identifiers for the best gene match in Arabidopsis are based on BlastX against the TAIR7_pep database. Biological process is taken from the Gene Ontology (GO) annotation on TAIR for each locus identifier (http://www.arabidopsis.org/tools/bulk/go/index.jsp). GenBank accession numbers identify the longest EST sequence for each gene from the collection of ESTs (∼67 000 ESTs) examined to construct this cDNA array (see Xiang et al., 2008), and these are included in the GAL file as descriptors for each reporter (Brassica gene) on the microarray (http://www.brassicagenomics.ca/cdnaarray.html).

Genes listed in bold were identified previously, based on EST abundance, as up-regulated during early stages of microspore embryogenesis in B. napus Topas DH4079 (Malik et al., 2007).

Organellar-derived genes are underlined.

The genes that were up-regulated in the 7 d enlarged Westar microspores (down-regulated during embryogenesis in the DH-2 line) were mostly related to pollen function, for example late embryogenesis abundant domain-containing protein/LEA (At4g13560.1), pectinesterase family protein (At3g05610), polygalacturonase (At3g07840.1), and SNF7 family protein (At5g63880.1) (Table 3). The developmental expression profile of the closest match in Arabidopsis for each of the 40 up-regulated genes in Westar was determined using the electronic Fluorescent Protein (eFP) Browser (http://bbc.botany.utoronto.ca/) and Genevestigator Gene Atlas (Zimmermann et al., 2004) on the The Arabidopsis Information Resource database (http://www.arabidopsis.org). The results of this search showed that at least 19 of these genes are highly expressed in stamen and/or pollen (Table 3).

Table 3.

Genes up-regulated in 7 d microspore cultures of poorly embryogenic Westar

Accession no.	Genes up-regulated in Westar (non-embryogenic)	Best match to Arabidopsis	E-value	Biological process	Broad functional category
ES264806	Pectinesterase family protein	AT3G05610.1*	2e-58	Cell wall modification	Cell organization and biogenesis
ES264447	UNE15 (unfertilized embryo sac 15)	AT4G13560.1	7e-29	Double fertilization forming a zygote and endosperm	Developmental processes
ES264881	Similar to BCP1 (Brassica campestris pollen protein 1)	AT3G26110.1*	1e-15	Pollen tube growth	Developmental processes
CN728581	RIC5 (ROP-INTERACTIVE CRIB MOTIF-CONTAINING PROTEIN 5)	AT3G23380.1	4e-23	Pollen tube growth	Developmental processes
ES265290	UNE15 (unfertilized embryo sac 15)	AT4G13560.1*	5e-43	Double fertilization forming a zygote and endosperm	Other biological processes
CN728588	NHL repeat-containing protein	AT5G14890.1*	2e-84	Double fertilization forming a zygote and endosperm	Other biological processes
ES264208	Similar to unknown protein (Arabidopsis thaliana)	AT5G39870.1*	2e-31	Double fertilization forming a zygote and endosperm	Other biological processes
CN727705	Family II extracellular lipase	AT1G20130.1*	4e-75	Lipid metabolic process	Other metabolic processes
CN728416	ATBETAFRUCT4, VAC-INV	AT1G12240.1	e-107	Carbohydrate metabolic process	Other metabolic processes
ES264826	GDSL-motif lipase	AT5G42160.1	2e-27	Lipid metabolic process	Other metabolic processes
ES265298	Exopolygalacturonase	AT3G14040.1*	3e-43	Carbohydrate metabolic process	Other metabolic processes
ES264781	Exopolygalacturonase	AT3G14040.1*	e-101	Carbohydrate metabolic process	Other metabolic processes
ES264609	LCR1 (low-molecular-weight cysteine-rich 1)	AT5G48543.1	4e-9	Carbohydrate metabolic process	Other metabolic processes
CN728494	Polygalacturonase	AT3G07840.1*	e-110	Carbohydrate metabolic process	Other metabolic processes
CN727751	Polygalacturonase, putative	AT5G48140.1*	1e-76	Carbohydrate metabolic process	Other metabolic processes
CN727871	Protein kinase family protein	AT3G01085.1	8e-76	Protein amino acid phosphorylation	Protein metabolism
CN728491	DNAJ heat shock N-terminal domain-containing protein	AT3G04980.1	2e-53	Protein folding	Protein metabolism
CN728503	Protein kinase family protein	AT3G01085.1	2e-86	Protein amino acid phosphorylation	Protein metabolism
CN728527	BTB/POZ domain-containing protein	AT4G08455.1	3e-78	Transport	Transport
CN727745	LCR11 (low-molecular-weight cysteine-rich 11)	AT4G11485.1	4e-7	Transport	Transport
CN728424	GAMMA-TIP3/TIP1;3	AT4G01470.1*	e-117	Transport	Transport
CN727761	Amino acid permease	AT1G71680.1*	9e-87	Amino acid transport	Transport
CN728482	Encodes a maternally expressed gene (MEG) family protein	AT2G16535.1	4e-17	Amino acid transport	Transport
CN728496	SNF7 family protein	AT5G63880.1*	e-101	Protein transport	Transport
CN728126	Encodes a defensin-like (DEFL) family protein	AT4G10603.1	6e-8	Transport	Transport
CN728594	Phytochrome kinase substrate-related	AT1G18810.1	6e-12	Biological process unknown	Unknown biological processes
ES264229	Invertase/pectin methylesterase inhibitor	AT3G17220.1*	1e-63	Biological process unknown	Unknown biological processes
CN728235	Unknown protein	AT1G15415.1	7e-27	Biological process unknown	Inknown biological processes
CN728176	Similar to unknown protein (Arabidopsis thaliana)	AT3G28840.1*	4e-34	Biological process unknown	Inknown biological processes
CN728423	Similar to unknown protein (Arabidopsis thaliana)	AT3G28790.1*	2e-31	Biological process unknown	Inknown biological processes
CN728566	Similar to unknown protein (Arabidopsis thaliana)	AT3G28780.1*	2e-63	Biological process unknown	Unknown biological processes
CN728544	Similar to unknown protein (Arabidopsis thaliana)	AT3G28790.1*	1e-16	Biological process unknown	Unknown biological processes
CN727786	Similar to unknown protein (Arabidopsis thaliana)	AT3G28790.1*	2e-41	Biological process unknown	Unknown biological processes
ES264118	No hits found
ES265176	No hits found
CN728230	No hits found
ES264291	No hits found
CN728497	No hits found
ES264343	No hits found
CN728523	No hits found

Locus identifiers for the best gene match in Arabidopsis are based on BlastX against the TAIR7_pep database. Biological process is taken from the Gene Ontology (GO) annotation on TAIR for each locus identifier (http://www.arabidopsis.org/tools/bulk/go/index.jsp). GenBank accession numbers identify the longest EST sequence for each gene from the collection of ESTs (∼67 000 ESTs) examined to construct this cDNA array (see Xiang et al., 2008), and these are included in the GAL file as descriptors for each reporter (Brassica gene) on the microarray (http://www.brassicagenomics.ca/cdnaarray.html).

^*These genes are highly expressed in the pollen and/or stamen of Arabidopsis (electronic Fluorescent Protein Browser; http://bbc.botany.utoronto.ca/, Winter et al., 2007).

The data indicate that induced microspores of DH-2 show up-regulation of many genes related to protein biosynthesis, while the heat-treated non-dividing microspores of Westar instead express genes related to carbohydrate and cell wall metabolism, indicative of a continued pollen developmental pathway. Additionally, some of the differentially expressed genes in induced DH-2 microspores are organellar transcripts suggestive of increased growth and multiplication of the mitochondria and chloroplasts, while no organellar transcripts were included in the list of differentially expressed genes representing Westar (pollen-like) microspores (Tables 2 and 3).

Phenotypic differences between Westar and Westar-derived DH-2 plants

In addition to the differences in embryogenic potential between Westar and DH-2, there are some striking differences in general plant morphology between these two lines (Fig. 5). The leaves of the DH-2 line are epinastic compared with leaves of Westar (Fig. 5A), and lateral branches in DH-2 are longer and more advanced than those of Westar plants of the same age, giving DH-2 plants a much bushier appearance (Fig. 5B). The petals of DH-2 plants are a paler yellow than petals of Westar flowers, and two petals of the DH-2 flowers are curled, thus imparting an asymmetry to the flower (Fig. 5C, D). The petals of DH-2 do not abscise normally, and a majority of the mature and dried siliques present retained petals, now white in colour, at the time of harvesting (Fig 5B). In addition, the uppermost portions of the inflorescences of DH-2 are looser and more elongated than the normally compact inflorescences of Westar (Fig. 5C). These features impart a different architecture to the plants and inflorescences of Westar and DH-2 lines; however, there are no differences in seed set and pollen fertility between the two lines.

Fig. 5.

Differences in morphology and architecture between plants of B. napus cv. Westar and the Westar-derived DH-2 line. (A) Young plants. The arrows indicate differences in leaf expansion between the two lines. (B) Plants during flowering and silique development. (C) Inflorescences. (D) Flowers on the day of anthesis. The arrow indicates wrinkled petals of the DH-2 line.

Hormone profiles

It was considered that the morphological differences between the Westar and Westar-derived DH-2 plants may be related to hormone levels and, therefore, ABA, auxins, and cytokinins were profiled in young and mature leaves and inflorescences of the two lines. Determination of ABA and its metabolites showed that the highest concentrations were of dihydrophaseic acid, an oxidation product of ABA, in all three tissues examined, but especially in young leaves and inflorescences of both lines (Fig. 6). There were no notable differences in the amounts of ABA or any of its intermediates between Westar and DH-2 in the tissues examined (Fig. 6). Isopentenyladenosine (iPA) was the most abundant cytokinin detected in any of the tissues, and showed the greatest accumulation in inflorescences of Westar and DH-2 (Fig. 6). Trans-zeatin (t-Z), trans-zeatin riboside (t-ZR), cis-zeatin riboside (c-ZR), and isopentenyladenine (2iP) were also present in higher amounts in the inflorescences as compared with the leaves, but at much lower concentrations than iPA (Fig. 6). IAA (indole-3-acetic acid) was present in high amounts in the inflorescences, at much lower levels in young leaves, and the least amount was found in mature leaves (Fig. 6). The auxin peptide conjugates, IAA-Asp (indole-3-aspartate) and IAA-Glu (indole-3-glutamate), were detected at low levels in the leaves and the inflorescences (Fig. 6). In summary, no notable differences in concentrations of any of these phytohormones were found in Westar and DH-2 tissues sampled that would account for the phenotypic differences and altered embryogenic response (Fig. 6). Methods are being developed that require less tissue in order to permit routine hormone profiling of microspores, developing embryos, and microspore-derived embryos.

Fig. 6.

Hormone profiling of metabolites and related compounds of ABA, auxin, and cytokinin in young and mature leaves and inflorescences (buds with 2–3 flowers) of B. napus cv. Westar and Westar-derived DH-2. Histograms indicate mean values (±SE) for each of the measured compounds, in three replicate tissue samples, each harvested from a different set of plants.

Discussion

The rescue of stable DH embryogenic lines from the rare embryos obtained from brassinosteroid-induced microspore cultures of the recalcitrant B. napus cultivar, Westar, has been described. The cultured haploid microspores bear the effects of the previous meiotic recombination, and thus collectively represent a population of all possible combinations of shuffled gene alleles prescribed within each parental genotype. Four of the rare embryos that developed from brassinosteroid-induced microspore cultures of cv. Westar were selected randomly to grow into F₁ plants, and all of these showed greatly increased rates of microspore embryogenesis in culture (Table 1) and in the later generations, as compared with the originating Westar parent. Thus, there was a direct relationship between microspores that formed embryos in the initial EBR-induced cultures, and the ability to express this characteristic heritably in succeeding generations.

The Westar-derived DH lines might have resulted from related genetic (or perhaps non-genetic) heritable changes, such as common re-arrangements or phenotypic exposure of recessive gene alleles due to fixed homozygosity at all loci. Alternatively, gene mutations should be considered. Uppermost estimates of the spontaneous mutation rate in gene-coding non-neutral DNA in plants are at 0.1–0.9 mutations per haploid genome per generation (Johnston and Schoen, 1995; Drake et al., 1998; Schultz et al., 1999). Taking the average (0.5), one would expect that out of 1 000 000 microspores per plate, there could conceivably be a single nucleotide mutation in one gene (out of >50 000 genes) in up to half of the cultured microspores in each plate. The level could even be higher if the heat exposure induced mutations. Regardless of the mechanism, it seems reasonable to speculate that there might be a limited number of ways to overcome the block in embryogenesis genetically in cv. Westar. In addition, it seems likely that the newly acquired heritable embryogenic potential of the Westar-derived DH lines was fixed early during the first microspore culture of cv. Westar, and that it is this same capability for embryogenesis that facilitated the first rare embryogenic events in culture, and later underlies the heritable improvements in embryo production for each of the Westar-derived DH lines.

The mechanism(s) by which the applications of brassinosteroids increased rates of embryogenesis in the original microspore cultures are still unknown (Ferrie et al., 2005). Brassinosteroids have been shown to have several effects on tissue culture material, including stimulating cell elongation, cell division, ethylene production, adventitious tissue formation, and increased resistance to abiotic stress (Miyazawa et al., 2003; Mussig and Altmann, 2003; Hardtke et al., 2007; Kagale et al., 2007). Brassinosteroids have also been used to improve somatic embryogenesis in conifers (Pullman et al., 2003). Brassinosteroid additions were most effective in improving microspore embryogenesis in various Brassica species and cultivars when included in the initial media during the heat stress treatment, and were relatively ineffective when included in media after heat stress induction (A M R Ferrie et al., unpublished data). A lingering question is whether the brassinosteroid additions in some way caused, or affected, the heritable increase in embryogenic potential, or whether these compounds merely permitted the expression, or rescue, of those alterations.

Iterative modifications to tissue culture protocols are frequently used to improve embryogenic responses for microspores from poorly embryogenic or recalcitrant varieties, cultivars, or species; however, there have been relatively few investigations as to whether the resulting microspore-derived embryos (and perhaps newly represented genotypes) maintain the same capacity to form embryos in the next generation. In cases where microspore-derived embryos result from the acquisition of such a potential, this may be a significant way to obtain stable embryogenic lines from recalcitrant varieties. Previous reports on Medicago and interspecific crosses of Helianthus have shown that the regenerated explants from tissue culture acquired and retained the ability to respond to in vitro conditions and form somatic embryos in multiple subcultures (Nolan et al., 1989; Fambrini et al., 1997). In the case of Westar and the Westar-derived DH lines, 100% of the DH embryos tested (four of four) showed increased embryogenesis in the next generation as compared with Westar (Table 1).

Westar DH-2 also showed several striking differences in morphology and architecture as compared with the Westar parental cultivar. These included leaf form, axillary branching patterns, petal colour, flower symmetry, and petal abscision (Fig. 5). These abnormal phenotypes were not associated with altered concentrations of ABA, auxins, or cytokinins, or their related metabolites between Westar and DH-2 (Fig. 6). Another notable difference between the two lines was a persistent difficulty in the quantitative isolation of RNA from both microspores and early microspore-derived embryos of the DH-2 line as compared with cv. Westar and all other Brassica species and cultivars previously investigated (Malik et al., 2007). The RNA yields were up to five times less from microspores and microspore cultures of DH-2, especially from the 5 d and 7 d stages, as compared with the parent line (data not shown). There were no differences in the extractability of RNA from any other tissues of DH-2, and DNA isolation was normal in all tissues examined (data not shown).

Future studies will be necessary to determine whether any of these phenotypic characteristics between DH-2 and Westar reflect pleiotropic effects linked directly to the embryogenic potential, or instead result from additional linked genes and/or somaclonal variations induced by tissue culture. Earlier reports in B. napus have identified two multiple gene loci with direct effects on microspore embryogenic ability (Zhang and Takahata, 2001; Zhang et al., 2003). Loci with additive gene effects on in vitro regeneration systems have been reported for other species, Brassica oleracea (Holme et al., 2004), Arabidopsis thaliana (Schianteralli et al., 2001), Oryza sativa (Taguchi-Shiobara et al., 1997), Solanum lycopersicum (Koorneef et al., 1993), Zea mays (Armstrong et al., 1992), and Hordeum vulgare (Komatsuda et al., 1995). Genes with additive functions might account for changes in embryogenic potential as well as the accompanying phenotypic variability due to dosage effects. Similarly, chromosomal rearrangements occurring during meiosis, including translocation and homologous and homeologous recombination events, might alter gene and allelic complements in a qualitative or dosage-dependent manner, thus affecting downstream morphological and physiological outcomes. Chromosomal rearrangements during meiosis have been well documented in B. napus cultivars and lines, including cv. Westar (Osborn et al., 2003; Udall et al., 2005). Additionally, there are reports of other genetic- or epigenetically regulated variations caused by tissue culture (Kaeppler et al., 2000; Guo et al., 2007). Tissue culture-induced variations in barley, studied with the aid of amplified fragment length polymorphism (AFLP)-based approaches, for example, have been linked to the genotype of the donor plants, medium composition, and the length of time in culture (Bednarek et al., 2007, and references within). These phenotypic variations may be detrimental in micropropagation experiments, but can be exploited to good effect in other instances where they give rise to useful traits.

Microscopic observations have shown that microspores of Westar become enlarged and swollen during the first 3 d of heat stress treatment; however, only a few of the induced microspores subsequently undergo cell divisions, thus accounting for the poor embryogenic response in cv. Westar (Fig. 1). The observations suggest a block to further embryogenic development in enlarged microspores of Westar by the 5 d stage, because similar-appearing enlarged microspores of the DH-2 line show cell divisions at 5 d and continue on through embryo development (Fig. 1). Quantitative real-time RT-PCR data revealed expression of LEC1, LEC2, ABI3, BBM1, UP1, and WOX9 in both Westar and DH-2, but notably that there was restricted transcription of all of these genes in the Westar cultivar, with the exception of LEC1 (Fig. 2). Moreover, transcription of LEC1 in 1 d and 3 d Westar cultures was delayed/inhibited, as compared with DH-2 (Fig. 2). These results provide molecular confirmation that microspores of Westar have attained some level of embryogenic competence under inductive conditions; however, this was not sufficient to drive the Westar microspores through to commitment to embryogenesis. Additionally, in conjunction with the timely and optimal expression of embryogenesis-related genes in competent microspores, pollen-expressed genes are down-regulated in Westar-derived DH-2, but not in cv. Westar, during the progression through to embryogenesis (Fig. 3).

So far these experiments have not revealed the nature of the molecular impediments that prevent the progression to embryogenesis in the original Westar material. To et al. (2006) have shown that LEC1, LEC2, ABI3, and FUS3 are major interacting and redundant regulators of embryogenesis and seed development, and each of these regulators is involved in embryo-specific transcriptional cascades affecting embryo identity and storage product accumulation (Wang et al., 2007). Ectopic overexpression of LEC1, LEC2, or BBM1 is sufficient to induce embryogenesis in somatic tissues (Lotan et al., 1998; Stone et al., 2001; Boutilier et al., 2002), and recent studies have further demonstrated that LEC1, LEC2, and FUS3 are required for zygotic and somatic embryogenesis in Arabidopsis (Gaj et al., 2005). LEC1 and LEC2 up-regulate the expression of ABI3 and FUS3 (Kagaya et al., 2005; To et al., 2006; Wang et al., 2007), while the regulon for LEC2 includes storage protein genes (Braybrook et al., 2006). The factors initially responsible for the transcriptional activation of LEC1, LEC2, and ABI3 during heat stress-induced microspore embryogenesis in B. napus are not known, although there are some data to implicate alkalinization, calcium signalling, and GTPase regulation (Pauls et al., 2006; Chan and Pauls, 2007). More recently, a downstream involvement of auxin pools and carbohydrate metabolism in the embryogenic response, correlated with the expression of some of these transcription factors, has emerged (Casson and Lindsey, 2006). Nonetheless, there is still little information on the role of LEC1 and LEC2 (and BBM1) in potentiating an environment conducive to the induction of an embryogenic programme, or the upstream events triggering the initial transcription of these embryogenesis-related genes.

Transcript profiling using expressed sequence tag (EST) frequencies from cDNA libraries and microarray analyses have provided considerable gene expression data for various developmental stages of microspore embryogenesis in B. napus (Joosen et al., 2007; Malik et al., 2007; Tsuwamoto et al., 2007; Xiang et al., 2008). In the present case, microarray comparisons of Westar and Westar-derived DH-2 have provided an opportunity to examine contrasting embryogenic responses of genetically related material cultured under identical conditions. The lists of differentially expressed genes compiled for 7 d DH-2 and Westar material clearly indicate there has been a major shift in metabolism in the DH-2 tissues during commitment to embryogenesis, involving increased translation and protein biosynthesis (Tables 2 and 3). The genes identified through the microarray comparison of Westar and DH-2 include some of the genes previously identified by EST profiling (Malik et al., 2007) and/or microarray analyses of highly embryogenic B. napus Topas DH4079 (Xiang et al., 2008), as well as many genes previously not listed or annotated (see Tables 2 and 3). These latter include mitochondrial- and chloroplast-derived genes, annotated embryo-defective genes, and several genes implicated in calcium responses.

In contrast, genes up-regulated in Westar include many pollen-related genes, thus underscoring the poorly embryogenic characteristics of this material (Table 3). Transcriptomic and proteomic profiling data for developing and mature pollen of Arabidopsis have shown a functional skew towards cell wall metabolism, carbohydrate metabolism, and cell structure (see Grennan, 2007, and references within). Surprisingly, in those same studies, genes related to protein biosynthesis were not detected in the pollen transcriptome, although their products were found in the proteome (Honys and Twell, 2004; Holmes-Davis et al., 2005), thereby indicating that some pollen proteins were formed early during pollen development, possibly in the uninucleate microspores which are enriched in genes related to protein biosynthesis (Malik et al., 2007).

It is well established that plants possess more plasticity in their genomes than most animal cells, thus allowing reprogramming of differentiated cells and expression of totipotency/pluripotency (Gutierrez, 2005; Costa and Shaw, 2007). Little is known about the details of genomic reprogramming in plant cells, but protein complexes including the Polycomb group of proteins (PcG) may be involved in the maintenance of silenced states and cellular memory. It is possible that the ability to aquire embryogenic potential or a totipotent state depends on the capacity of plant cells to modify gene expression in response to some external cues. Kinases, for example SERK (Schmidt et al., 1997), may be involved in the upstream perception of external stimuli, and transcription factors such as LEC1, LEC2, and BBM1 may act downstream to confer embryogenic potential (Lotan et al., 1998; Stone et al., 2001; Boutilier et al., 2002); however, the details of how these genes are de-repressed and the nature of their interactions to permit the expression of embryogenic potential are still not clear. Recent research progress on the stem cell potential of somatic animal tissues has taken a giant leap forward, and surged ahead of some studies in plant systems, with recent discoveries and demonstrations that expression of cassettes of four genes can transform skin cells into embryonic stem cells (Takahashi et al., 2007; Yu et al., 2007). Advances in our understanding of the cellular conditions associated with embryogenic competence and improved protocols for inducing embryogenic potential in a wide selection of plant cells will require further detailed studies of cell biology, proteomics, and metabolomics, in addition to transcript profiling. The development of genetically related lines differing in embryogenic responses, for example Westar and Westar-derived DH-2, provides an ideal opportunity for in-depth molecular studies of embryogenic potential and embryo development in plants.

Supplementary data

Supplementary data are available at JXB online.

Table S1. Accession numbers for embryo- and pollen-specific marker genes.