Introduction

In higher plants, life cycle starts from the fertilization of an egg cell by a sperm to produce a zygotic embryo. After fertilization, a zygotic embryo develops into a mature embryo through a series of developmental processes referred to as embryogenesis. In addition to the zygotic embryogenesis pathway, embryos can also arise from somatic tissues without undergoing meiosis and fertilization. Asexual embryos can be produced by manipulating somatic or gametophytic cells from a variety of species 1, 2. Moreover, the process of somatic embryo development is similar to that of zygotic embryos which are usually embedded inside maternal tissues, making them inaccessible to analysis. Thus, somatic embryos provide an attractive model for studying embryogenesis.

In vitro induction of somatic embryo formation is initiated by inducing de-differentiation of somatic cells, usually requiring high concentrations of 2,4-dichlorophenoxyacetic acid (2,4-D). Subsequently, the de-differentiated cells or calli re-differentiate and develop as somatic embryos by an unknown mechanism. Whereas phytohormones are generally believed to activate a signaling pathway for the induction of embryogenic competence, a number of genetic loci have also been shown to play critical roles during somatic embryo development 2, 3, 4. Mutations in the Arabidopsis LEAFY COTYLEDON1 (LEC1) and LEC2 genes cause similar defects in embryo maturation and partially transform cotyledons into leaves 5, 6. LEC1 and LEC2 encode a CCAAT-binding factor-type transcription factor and a plant-specific B3 transcription factor, respectively 7, 8. Ectopic expression of these two embryo-specific genes causes the formation of somatic embryos on vegetative tissues 4, 7, 8. A recent study shows that the auxin biosynthetic genes YUC2 and YUC4 are rapidly induced by LEC2, and YUC4 is likely a direct target of LEC2 4. These observations establish a molecular link between the regulation of embryogenic competence and the auxin activity. On the other hand, whereas LEC1 and LEC2 are key regulators controlling embryonic identity, the Arabidopsis PICKLE (PKL) gene, encoding a CHD3 chromatin-remodeling factor, appears to play an opposite role. The pkl mutant displays some embryonic traits after germination, including a higher capacity of embryogenic competence, the production of somatic embryos from root explants, and an elevated LEC1 expression 9, 10, 11. Because CHD3 proteins are involved in transcriptional repression, PKL has been proposed to prevent re-expression of the embryonic developmental state during germination, likely through the regulation of gibberellin (GA) signaling 10, 11. A more recent study identified three functionally redundant VP1/ABI3-LIKE (VAL) genes encoding B3 domain-containing proteins 12. Whereas val1 shows a pkl-like phenotype, the val1 val2 double mutant produces embryo-like structures around root and apical meristem regions, which is associated with increased expression of LEC1, L1L, ABI3, and FUSCA3 (FUS3). It has been proposed that VAL factors may function as transcriptional repressors of the LEC1/B3 genes in conjunction with PKL or related factors 12.

Using reverse genetic approaches, several genes have been shown to have the capacity to induce the formation of embryonic calli or somatic embryos from vegetative tissues. The carrot SOMATIC EMBRYOGENESIS RECEPTOR-LIKE KINASE (SERK) gene was shown to mark vegetative-to-embryonic transition in carrot suspension cell cultures 13, and its Arabidopsis orthologous gene was found to enhance embryogenic competence in cell cultures as well 14. The BABY BOOM (BBM) gene, encoding an AP2/ERF-type transcription factor, was identified from Brassica napus and Arabidopsis thaliana. Ectopic expression of BBM results in the formation of somatic embryos and cotyledon-like structures on seedlings 15. Moreover, overexpression of the Arabidopsis AGAMOUS-Like 15 (AGL15) gene and its homologous gene in soybean, which encode MADS domain-containing proteins, enhances the formation of secondary embryos and somatic embryos 16, 17. The promotion of somatic embryo formation by AGL15 is attributed, at least in part, to a reduced GA level 18. This notion is consistent with the observation made in the pkl mutant that displays a GA-deficiency phenotype 9, 10, and the observation that FUS3, encoding a B3 transcription factor critical for embryogenesis, negatively regulates GA biosynthesis 19.

We have used a functional screening strategy to identify Arabidopsis gain-of-function mutants, plant growth activator (pga), by culturing T-DNA-transformed root explants on a medium containing auxin but without cytokinin. In this screen, a chemical-inducible activation tagging vector was used, and therefore the expression of the mutant phenotype was strictly dependent on the chemical inducer estradiol 20, 21. Mutants identified from this screen were presumed to conditionally overexpress key components that regulate somatic embryo formation or the cytokinin response 21, 22. The pga6 mutant was found to efficiently produce somatic embryos on seedlings and explants in the presence of the inducer. Molecular analysis demonstrated that PGA6 was identical to the homeobox gene WUSCHEL (WUS) 21, which regulates stem cell fate in shoot and floral meristems 23, 24, 25. Moreover, ectopic expression of WUS also induces shoot and floral stem cell activities in roots and stems, respectively 26, 27. These observations suggest that WUS/PGA6 plays a critical role in the maintenance of the identity of embryonic stem cells. Here, we report the characterization of the pga37 mutant, which displays a similar but distinctive phenotype compared to pga6. The PGA37 gene, encoding a MYB transcription factor, induces embryogenic competence in a WUS-independent manner.

Results

The pga37 mutant phenotype

In previous genetic screens for pga mutants, we have identified two groups of mutants that displayed the ability to form either somatic embryos or shoots 21, 22. For example, overexpression of the PGA6/WUS gene causes vegetative-to-embryonic transition 21. The PGA22 gene, encoding isopentenyl transferase 8 (IPT8), catalyzes the first rate-limiting step of cytokinin biosynthesis, and the pga22 gain-of-function mutant shows a cytokinin overproduction phenotype, including the regeneration of shoots in the absence of externally supplied cytokinin 22.

Here, we report the characterization of pga37, which showed a phenotype similar to that of pga6. When cultured on the screening medium (SCM; containing IAA, kanamycin, and the chemical inducer 17-β-estradiol but without cytokinins; see Materials and Methods) for 7-10 days, pga37 root explants produced green-yellowish embryonic calli (Figure 1A), which were not found on control medium without the chemical inducer estradiol (Figure 1B). Upon culturing for a longer period of 3-5 weeks, somatic embryos were generated from embryonic calli (Figure 1C and 1D). Scanning electron microscopic analysis revealed the presence of somatic embryos at a variety of developmental stages, including the globular, heart, and cotyledon stage (Figure 1E–1G). Similar to that observed in the pga6 mutant, somatic embryos with three cotyledons were occasionally observed (Figure 1H).

Figure 1
figure 1

The PGA37 gain-of-function mutant phenotype. (A) pga37 root explants cultured on SCM containing 10 μM estradiol for 7 days. Note: yellowish green embryonic calli are visible. (B) pga37 root explants cultured on SCM without estradiol for 7 days. (C) pga37 root explants cultured on SCM containing 10 μM estradiol for 5 weeks. Somatic embryos are formed on the explants. (D) An enlarged view of panel (C). (E-H) Scanning electron microscopy of somatic embryos at various developmental stages. Images were taken from pga37 root explants cultured on SCM for 3-5 weeks. (E) A somatic embryo at the globular stage. (F) A somatic embryo at the heart stage. (G) A somatic embryo at the cotyledon stage. (H) An abnormal somatic embryo with three cotyledons. (I) A PGA37 seedling germinated from a somatic embryo. (J) A soil-grown pga37 seedling derived from a somatic embryo. (K) Seven-day-old pga37 seedlings germinated and grown on a MS medium in the absence (left) or presence (right) of 10 μM estradiol. (L) An enlarged view of the boxed region in panel (K). (M) Three-week-old pga37 seedlings germinated and grown on a MS medium in the absence (left) or presence (right) of 10 μM estradiol. (N) An enlarged view of the boxed region in panel (M). (O and P) Fat Red 7B staining of 2-week-old pga37 seedlings germinated and grown on MS medium in the absence (O) or presence (P) of 10 μM estradiol. Bar, 0.5 mm (A-D), 50 μm (E), 100 μm (F-H), 2 mm (I and L), 5 mm (K, N-P), and 2 cm (J and M).

After transfer from SCM to MS medium 28 without the inducer estradiol, these somatic embryos were able to continue development into healthy seedlings, and plantlets were fertile and set seeds (Figure 1I and 1J). When seeds harvested from pga37 plantlets were germinated and grown in the absence of the inducer, pga37 plants were indistinguishable from wild-type (WT) plants. However, when germinated and grown in the presence of inducer, pga37 seedlings were smaller than WT seedlings, and produced dark green cotyledons and true leaves (Figure 1K and 1M). Moreover, pga37 roots and the junction region between the root and the hypocotyl turned light green (Figure 1K–1N) similar to that previously observed in pickle (pkl) mutant roots 9, 10. Notably, the pga37 mutation also caused the accumulation of storage compounds such as lipids, which were normally found in embryos and seeds, in roots and other vegetative tissues (Figure 1O and 1P).

Taken together, these observations suggest that the pga37 mutation is capable of promoting vegetative-to-embryonic transition, thus likely representing a key genetic locus regulating Arabidopsis embryogenesis.

Molecular characterization of the PGA37 gene

To determine the genetic basis of pga37 (in the Wassilewskija (WS) background), we crossed the mutant with WT (WS). All tested F1 progeny from the cross showed the mutant phenotype in an estradiol-dependent manner. In an F2 population obtained from self-pollinated F1 plants, the phenotype of forming somatic embryos on explants segregated in a 3:1 ratio (mutant:WT = 85:26). These results indicated that pga37 is a dominant mutation in a single nuclear gene, and its mutant phenotype was strictly dependent on the inducer. Because of the dependency on the inducer, the mutated gene should be tightly linked to a T-DNA insert carrying the OLexA-46 promoter/enhancer. We identified the Arabidopsis genomic sequences flanking the OLexA-46 promoter/enhancer by inverse PCR 29. A T-DNA was found to be inserted approximately 7.8 kilo base pairs (kb) upstream from the putative translation start codon of At3g27785, which encodes the transcription factor MYB118. The left border (LB) of the T-DNA insertion was adjacent to the 5′-end of MYB118, thus allowing the inducible OLexA-46 promoter/enhancer to activate the expression of the transcription factor gene (Figure 2A). Using an F2 population obtained from a cross between pga37 (WS) and WT (Columbia-0 or Col-0), the mutation was roughly mapped onto chromosome III, approximately 4.4 cM from the simple sequence length polymorphism maker ciw11 (45 F2 plants were tested). MYB118 is approximately 4.0 cM south of ciw11 (Figure 2A). These results suggest that the pga37 mutation was likely caused by the identified T-DNA insertion.

Figure 2
figure 2

Molecular characterization of PGA37. (A) A schematic map of the pga37 mutant genome (not in scale). The insertion position is indicated by an arrow that also shows the orientation of left border (LB) of the T-DNA. The insertion position of the right border (RB) is unclear, as shown by a question mark. The pga37 mutation was mapped to approximately 4.4 cM from the ciw11 marker. (B) Expression of PGA37 in the pga37 mutant. Two-week-old pga37 seedlings were treated with different concentrations (μM) of estradiol for 16 h as indicated on the top of the panel. Total RNA was analyzed by northern blot hybridization using a PGA37 cDNA fragment as a probe. WT: wild type (Wassilewskija or WS) seedlings treated with dimethyl sulfoxide (DMSO; a solvent for estradiol). (C) Dose-dependency of the pga37 mutant phenotype. Two-week-old pga37 seedlings germinated and grown on MS medium containing various concentrations of estradiol (μM) as indicated on the top of the panel. WT: DMSO-treated WS seedlings. (D) Expression of PGA37 in pER10-PGA37 transgenic plants. Two-week-old seedlings of WT, pga37 and pER10-PGA37 transgenic plants (PGA37t) were treated with (+) or without (−) 10 μM estradiol for 16 h, and total RNA was analyzed by northern blot hybridization as described in panel (B). (E) Formation of somatic embryos on root explants of a pER10-PGA37 transgenic line. The explants were cultured on SCM for 5 weeks. (F) Phenotype of pER10-PGA37 transgenic plants. Two-week-old pER10-PGA37 seedlings germinated and grown on MS medium containing various concentrations of estradiol (μM) as indicated on the top of the panel. WT: DMSO-treated WS seedlings. Note that the transgenic plants showed a stronger phenotype than pga37 (see panel (C) and text for details). (G) An enlarged view of a pER10-PGA37 transgenic seedling in panel (F) marked by an arrow (grown in the presence of 10 μM estradiol). No distinctive root and hypocotyl structures were found, and embryonic calli were formed on the plantlet. Bar, 1 cm (C and F) and 1 mm (E and G).

In this 7.8-kb region between the insertion site and MYB118, neither protein- nor putative small RNA-encoding genes could be annotated using various prediction programs or by a manual search. Northern blot analysis showed that MYB118 expression was inducible in the pga37 mutant by estradiol in a dose-dependent manner (Figure 2B); moreover, the severity of the mutant phenotype was correlated with the MYB118 expression level (Figure 2B and 2C). A MYB118 cDNA fragment spanning the entire coding sequences was cloned into the estradiol-inducible expression vector pER10 20 and the resulting construct pER10-MYB118 was transformed into WT plants (WS). Expression of the MYB118 transgene was highly inducible by estradiol (Figure 2D). Upon induction, the pga37 mutant phenotype was recapitulated in pER10-MYB118 transgenic plants, including the somatic embryo formation phenotype (Figure 2E) and retarded growth and development of the transgenic seedlings (Figure 2F). Notably, when germinated under higher concentrations of inducer, growth and development of pER10-MYB118 transgenic plants were completely arrested, and the treated plants produced callus-like tissues without any recognizable structures of roots and hypocotyls (Figure 2F and 2G). Note that multiple independent pER10-MYB118 transgenic lines were analyzed and similar results were obtained.

A stronger phenotype of the pER10-MYB118 transgenic plants compared to pga37 was most likely caused by higher transgene expression levels. In most transgenic lines, MYB118 expression was substantially higher than that in pga37, as shown in Figure 2D for a representative transgenic line. Moreover, because of the relative distal position of the T-DNA insertion from MYB118 in the pga37 genome (approximately 7.8 kb), the OLexA-46 sequence was unlikely to directly activate the MYB118 transcription as a promoter, but rather it might act as a transcriptional enhancer. Consistent with this notion, only a single transcript was detected in inducer-treated pga37 plants, which was different from that observed in pga6 and pga22. In pga6 and pga22, two species of transcripts were detected, of which the shorter species presumably originated from the native promoters whereas the longer one likely initiated from the OLexA-46 promoter 21, 22. Nevertheless, the data presented above confirmed that overexpression of MYB118 was sufficient to recapitulate the pga37 mutant phenotype.

A PGA37::GUS reporter gene is preferentially expressed in embryos and endosperm

A reverse transcription-PCR (RT-PCR) analysis detected PGA37 expression in siliques but not in any other tissues/organs (Figure 3A). To more precisely follow the PGA37 expression pattern, a PGA37 promoter::β-glucuronidase (PGA37::GUS) reporter gene was transformed into WT plants. A 2.5-kb Arabidopsis genomic DNA fragment, including the putative promoter sequence, 5′-untranslated region (UTR), and sequences encoding the first 12 amino acid residues of MYB118, was fused in frame to the GUS coding sequence. The PGA37::GUS expression was specifically detected in developing embryos, starting approximately at the early heart stage and throughout embryogenesis (Figure 3B–3E). GUS activity was restricted to endosperm and embryos, but not detectable in the seed coat (Figure 3B and 3C). Upon entering the cotyledon stage, GUS expression was also detected in embryos (Figure 3D and 3E), but became undetectable after germination. Consistent with the results of RT-PCR analysis (Figure 3A), the PGA37::GUS expression was not detected in any other tissues/organs at various developmental stages. The embryo- and endosperm-specific expression pattern of PGA37 suggests that it plays a role during embryogenesis.

Figure 3
figure 3

Expression pattern of PGA37. (A) Total RNA from various tissues/organs of WT plants was subjected to reverse transcription (RT)-PCR analysis. Two micrograms of RNA was used for RT using oligo-dT as a primer, and the PCR was cycled for 30 times. Primers used in PCR were located in exons 2 and 4 of PGA37, respectively (see Figure 6I for the positions and orientation of the primers). Actin7 (ACT7; At5g09810) was used as an internal control. Se: seedling; Ro: root; St: stem; RL: rosette leaf; CL: cauline leaf; FB: flower bud; Fl: flower; Si: silique. (B-E) Analysis of β-glucuronidase (GUS) activity in PGA37::GUS transgenic plants. At least five independent transgenic lines were analyzed and similar results were obtained. (B) Early heart stage; (C) late heart stage; (D) torpedo stage; (E) cotyledon stage. Bar = 100 μm.

Increased expression of embryonic genes in pga37

Data presented above show that PGA37/MYB118 is capable of inducing vegetative-to-embryonic transition, leading to the formation of somatic embryos. To better understand the molecular mechanism underpinning this transition, we investigated the expression of several embryo- or seed-specific genes in pga37. Root explants derived from WT and pga37 plants were cultured on SCM for different times. In pga37 explants, green-yellowish embryonic calli were usually formed around 7-10 days, and somatic embryos were produced around 20-30 days. Expression of LEC1 was detectable around day 1, whereas expression of LEC2, FUS3 and WUS became detectable in later stages (Figure 4A). A similar expression pattern was also observed in explants derived from pER10-PGA37 transgenic plants. To confirm this result, pER10-PGA37 transgenic plants germinated and grown on MS medium were treated with inducer for varying times, and LEC1 expression was analyzed by RT-PCR. LEC1 expression could be detected after 6 h of induction, and a higher expression level was found at later time points (Figure 4B). Under these assay conditions, expression of WUS, FUS3, and LEC2 was not detected (data not shown). These results suggest that PGA37 is able to enhance LEC1 expression in non-embryonic tissues.

Figure 4
figure 4

PGA37-regulated expression of LEC1. (A) RT-PCR analysis of PGA37, LEC1, LEC2, FUS3, and WUS expression in pga37 root explants. Root explants cultured on CIM medium (containing 2,4-D; see Materials and Methods) for 3 days were transferred onto SCM. Root explants collected at the time points indicated on the top of the panel were used for the preparation of total RNA. RT-PCR was carried out as described in Figure 3A. PCR was cycled 25 times for PGA37 and ACT7, and 30 times for other genes. (B) RT-PCR analysis of the LEC1 expression in pga37 seedlings. Two-week-old seedlings germinated and grown on MS medium were treated with 10 μM estradiol for different times (hours; as indicated on the top of the panel). Expression of LEC2 and WUS was not detected under the assay conditions. (C) Expression of PGA37 and LEC1 in WT and pkl mutant plants. Two-week-old seedlings were germinated and grown on MS medium with (+) or without (−) 20 nM paclobutrazol (PAC). PAC is an inhibitor of gibberellin biosynthesis, which increases the phenotype penetrance of pkl and val1 mutants 9, 12. Total RNA was used for RT-PCR analysis as described in Figure 3. RNA prepared from pga37 seedlings treated with estradiol was used as a positive control for the analysis of the PGA37 expression (not shown).

Because pga37 roots showed a phenotype similar to that of pkl, which produced green roots and had an elevated LEC1 expression level 9, 10, it is possible that MYB118 expression might be elevated in pkl as well. However, MYB118 expression was not detectable under various conditions in pkl, whereas LEC1 expression was readily detected as reported previously (Figure 4C) 10, suggesting that MYB118 is not directly regulated by PKL.

PGA37 acts independently from WUS

The pga37 mutant shares some similarities with the pga6 mutant with respect to their capacity to promote vegetative-to-embryonic transition. To investigate possible interactions between WUS/PGA6 and PGA37, we constructed a pga37 wus-1 double mutant by genetic crosses of these two mutants. wus-1 has been previously shown as a null mutation 24. In the double mutant, PGA37 expression was inducible by estradiol in the wus-1 loss-of-function mutant background. When germinated and grown in the presence of estradiol, pga37 wus-1 showed a pga37-like phenotype, with characteristics of dark-green leaves, light green roots and growth inhibition (Figure 5A). Notably, developmental defects in the shoot apical meristem in wus-1 could not be rescued by PGA37 overexpression (Figure 5B–5E). To test if somatic embryo formation promoted by PGA37 is dependent on WUS, pga37 wus-1 root explants were cultured on a callus induction medium (CIM) and then transferred onto SCM (see Materials and Methods). Root explants derived from both pga37 wus-1 and PGA37 plants were able to efficiently form somatic embryos (Figure 5F), suggesting that WUS is not required for PGA37 function in promoting somatic embryo formation.

Figure 5
figure 5

PGA37 functions independently from WUS. (A) Two-week-old seedlings germinated and grown on MS medium in the presence or the absence of 10 μM estradiol. Note that the pga37 wus-1 double mutant shows a phenotype similar to that of pga37 in the presence of estradiol. Double mutant plants were smaller than WT plants and produced green roots. From left to right: pga37 on MS medium, pga37 wus-1 on MS medium, pga37 germinated on estradiol-containing medium, and pga37 wus-1 germinated on estradiol-containing medium. (B-E) Enlarged views of 2-week-old seedlings under different conditions. (B) A pga37 seedling germinated and grown on MS medium. (C) A pga37 seedling germinated and grown on MS medium containing 10 μM estradiol. (D) A PGA37 wus-1 seedling germinated and grown on MS medium. (E) A pga37 wus-1 seedling germinated and grown on MS medium containing 10 μM estradiol. (F) Root explants derived from pga37 (left) and pga37 wus-1 double mutant (right) plants were cultured as described in panel (A) (pretreated with 2,4-D for 3 days). Bar, 2 cm (A) and 2 mm (others).

Functional characterization of MYB115, a homologous gene of PGA37

PGA37 encodes a R2R3-type MYB transcription factor (MYB118) with unknown function 30, 31. PGA37/MYB118 belongs to a small subgroup of six members, classified on the basis of the relatively conserved carboxyl terminal sequences outside the MYB domain 30, 32. In this family, MYB115 (At5g40360) is most closely related to MYB118 30. These two proteins share approximately 32% identity (45% similarity) in the 150-amino acid segment located at the carboxyl terminal half. To functionally characterize MYB115, an estradiol-inducible expression construct pER10-MYB115 was generated, which was stably transformed into WT plants. Expression of the MYB115 transgene in transgenic plants was highly inducible by estradiol (Figure 6A).

Figure 6
figure 6

MYB115 functions similarly as PGA37/MYB118. (A) Expression of MYB115 in pER10-MYB115 transgenic plants. Two-week-old pER10-MYB115 transgenic seedlings germinated and grown on MS medium were treated with (+) or without (−) 10 μM estradiol for 16 h. Total RNA used for RT-PCR analysis as described in Figure 3A. The PCR was cycled for 25 times. Two transgenic lines (indicated on the top of the panel) with different MYB115 expression levels are shown. (B) Two-week-old pER10-MYB115 transgenic seedlings germinated and grown on MS medium supplemented with (right in each panel) or without (left in each panel) 10 μM estradiol. Two transgenic lines (#5 and #6) are shown. (C) An enlarged view of a plantlet shown in panel (B) marked by an arrow. (D) Germinating somatic embryos from pER10-MYB115 root explants cultured on SCM for 4 weeks. (E-G) Scanning electron microscopy of somatic embryos derived from pER10-MYB115 root explants cultured on SCM for 4 weeks. (E) A somatic embryo at the globular stage. (F) A somatic embryo at the cotyledon stage. (G) A somatic embryo with fused cotyledons. (H) Expression pattern of MYB115 analyzed by RT-PCR. Se: seedling; Ro: root; St: stem; RL: rosette leaf; CL: cauline leaf; FB: flower bud; Fl: flower; Si: silique. (I) Schematic maps of the myb118 (top) and myb115 (bottom) mutant genomes. A T-DNA was inserted in exon 3 of PGA37/MYB118, and a T-DNA was inserted in intron 1 of MYB115. Filled boxed and lines denote exons and introns, respectively. Dashed lines indicate non-transcribed sequences. Arrows above exons show the positions and orientation of primers used for RT-PCR analyses shown in panel (J). (J) Expression of PGA37/MYB118 (top) and MYB115 (bottom) in the myb118 myb115 double mutant. Total RNA prepared from immature siliques of WT, single, and double mutant plants was used for RT-PCR analysis. DM: the myb118 myb115 double mutant. Bar, 0.5 cm (B), 0.5 mm (C and D), 50 μm (E), and 200 μm (F and G).

When germinated in the presence of inducer, growth and development of pER10-MYB115 transgenic plants were significantly retarded, and the phenotypic strength was dependent on the transgene expression level (Figure 6A–6C). Overexpression of MYB115 caused a complete arrest of plant growth and development and the transformation of the plantlet into callus-like structures (Figure 6C), a phenotype similar to that of pER10-MYB118 (see Figure 2G). In root explants derived from pER10-MYB115 transgenic plants, somatic embryos could be efficiently formed in an inducer-dependent manner (Figure 6D–6G). Similar to that observed in pga37 and pER10-MYB118 transgenic plants, expression of LEC1, but not of LEC2 and WUS, was increased in pER10-MYB115 transgenic plants treated with the inducer for 1-2 days (data not shown). The results suggest that the structurally similar MYB118/PGA37 and MYB115 genes have similar functions.

A RT-PCR analysis showed that MYB115 was expressed in flowers and siliques, but its expression was not detectable in other tissues (Figure 6H). The expression pattern is similar to that of PGA37/MYB118 but in a more extended domain (see Figure 3), suggesting that these two genes have similar but distinctive functions during plant growth and development. To further investigate the function of PGA37/MYB118 and MYB115 in plant growth and development, we identified two T-DNA insertion lines SALK_111812 and SALK_044168 33. These two lines contained T-DNA insertions in exon 3 of PGA37/MFYB118 and intron 1 of MYB115, respectively (Figure 6I). Expression of PGA37/MYB118 was not detected in siliques of myb118 and myb118 myb115 plants (Figure 6J), suggesting that the T-DNA insertion likely caused a null mutation. However, reduced expression of MYB115 was detected in myb115 and myb118 myb115 siliques (Figure 6J), which presumably resulted from cryptic splicing of myb115 pre-mRNAs. Homozygous myb118 and myb115 plants were phenotypically indistinguishable from WT plants. In addition, no abnormalities were found in the myb118 myb115 double mutant at different developmental stages. When germinated and grown in the presence of auxin, cytokinin, abscisic acid or sucrose, the myb118 myb115 double mutant showed a normal response to these compounds as WT.

Discussion

In this study, we show that overexpression of PGA37/MYB118 and MYB115, two homologous Arabidopsis genes, efficiently induces the formation of somatic embryos from root explants, resulting in elevated expression levels of the embryo-specific gene LEC1. Moreover, overexpression of PGA37/MYB118 and MYB115 also causes the accumulation of lipids in vegetative tissues, the formation of green roots, and the disruption of post-germination morphogenesis. These observations suggest that PGA37/MYB118 and MYB115 are able to promote reprogramming of cell fate, acting as positive regulators of vegetative-to-embryonic transition.

The pga37 mutant shows phenotypic similarities with pga6 and transgenic plants overexpressing BBM 15, 21. The most striking phenotype of these mutants and transgenic plants is their capacity to form somatic embryos from vegetative tissues. Several lines of evidence suggest that PGA37 may act independently from WUS/PGA6. First, WUS expression is restricted to the organizing center during embryogenesis 24. However, PGA37 is expressed in a more expanded domain, including both the embryo and the endosperm, similar to the expression pattern of BBM 15. Although a non-cell-autonomous regulatory mechanism could not be excluded, it is unlikely that WUS directly targets PGA37. Second, overexpression of PGA37 is unable to rescue the shoot apical meristem defects in wus-1, thus disfavoring a model that PGA37 acts downstream of WUS. Third, WUS represses 21, whereas PGA37 promotes, LEC1 expression in root explants or embryogenic cells. Moreover, in pga6 root explants, LEC1 expression was found to be induced around 14 days after cultured on SCM 21. By contrast, LEC1 expression is detectable in pER10-PGA37 seedlings treated with estradiol for 12 h. These observations lead to the speculation that LEC1 expression observed in pga6 may represent a characteristic of embryogenic calli or somatic embryos. A similar observation was recently made by finding that LEC2-induced expression of genes encoding the seed proteins oleosin and storage proteins is prior to that of LEC1 and FUS3 4. In contrast to that of WUS/PGA6 and LEC2, PGA37 may act upstream of LEC1 to regulate vegetative-to-embryonic transition. Therefore, these different regulatory modes do not support the hypothesis that WUS and PGA37 act in a linear pathway. Instead, different mechanisms may be deployed by WUS/PGA6 and PGA37 to regulate vegetative-to-embryonic transition.

Compared to the previously characterized genetic loci involved in the regulation of vegetative-to-embryonic transition such as WUS/PGA6 and BBM, PGA37 employs a unique regulatory mechanism. The PGA37 activity is independent of WUS in the regulation of vegetative-to-embryonic transition. Moreover, PGA37/MYB118 is capable of enhancing LEC1 expression, but does not appear to be regulated by PKL. Whereas PKL has been shown to repress LEC1 expression 10, the expression of PKL itself is, in turn, positively regulated by LEC1, suggesting the involvement of a feedback regulatory mechanism (J Mu and J Zuo, unpublished data). Similarly, expression of LEC1, but not of PGA37/MYB118 and MYB115, is remarkably elevated in the val1 val2 double mutant 12. Notably, overexpression of PGA37/MYB118 and MYB115 induces the expression of LEC1 but not LEC2, consistent with the observation made in previous studies that these two loci may act independently in the control of embryo development and seed maturation 34. Collectively, these results suggest that PGA37/MYB118 and MYB115 may regulate LEC1 expression in a PKL-independent manner. LEC1 is a key regulator of embryogenesis and seed maturation 5, 6, 7, 35. Overexpression of LEC1 is sufficient to induce vegetative-to-embryonic transition 7, indicating a critical role of LEC1 in the initiation and/or maintenance of embryonic cell identity. Therefore, the promotion of vegetative-to-embryonic transition by PGA37/MYB118 and MYB115 should be attributed, at least partly, to an elevated LEC1 expression level.

PGA37/MYB118 is preferentially expressed in embryos and endosperm, implying a possible role of the MYB transcription factor in embryogenesis and seed development. This notion is supported by the phenotype of the gain-of-function mutants and of transgenic plants overexpressing PGA37. However, neither myb118 nor the myb118 myb115 double mutant displays any developmental defects under our assay conditions. The lack of an apparent phenotype may be caused by functional redundancy of this small class of MYB transcription factors and, in particular, owing to the residual expression of MYB115 in the double mutant. Alternatively, genetic redundancy may be present between the MYB118/MYB115 family and other transcription factor genes. Despite the lack of an apparent phenotype in myb118 and myb118 myb115 mutants, overexpression of these two genes causes strong inhibitory effects on plant growth and development and promotes the formation of embryonic calli on seedlings, suggesting that these two functionally redundant genes likely play a role in embryo or seed development.

In summary, we have identified and characterized PGA37/MYB118 and MYB115 that are capable of inducing vegetative-to-embryonic transition and the formation of somatic embryos. Because PGA37/MYB118 acts independently from WUS, the finding presented in this study thus reveals a new regulatory mechanism of somatic embryogenesis in Arabidopsis, which should have substantial impacts on both plant biotechnology and our understanding of this fundamental question in plant biology.

Materials and Methods

Plant material, growth conditions, and plant transformation

The Arabidopsis thaliana ecotypes Col-0, WS, and Landsberg erecta (Ler) were used in this study. Screening of pga mutants has been described in detail previously 21, 22. The myb118 (SALK_111812; Col-0), myb115 (SALK_044168; Col-0), and wus-1 (CS15; Ler) mutants were obtained from the Arabidopsis Biological Resource Center (ABRC, Ohio State University, Columbus, OH). The pkl-1 mutant (Col-0) 9 was kindly provided by J Ogas.

Plants were grown at 22 °C under a 16-h-light/8-h-dark cycle in soil or on agar plates containing 1/2 MS medium (Sigma Hong Kong and China), 2% sucrose, and 0.8% agar supplemented with or without the inducer 17-β-estradiol. Treatment of plants with 17-β-estradiol was carried out as described previously 20. Transformation of root explants was carried out using Agrobacterium tumefaciens (the ABI strain) as described previously 21. Unless indicated otherwise, root explants were cultured on a CIM (1× B5 salts, 2% glucose, 0.5 g/L MES, 0.5 mg/L 2,4-D, 0.05 mg/L kinetin and 0.25% phytagel, pH 5.7) for 3 days, and subsequently transferred onto a SCM (1× MS salts, 1% sucrose, 0.5 g/L MES, 0.15 mg/L IAA, 10 μM 17-β-estradiol, and 0.25% phytagel, pH 5.7) for 3-5 weeks. In planta transformation was done by the flower-dip method 36 using the Agrobacterium GV3101 strain.

Staining of seedlings with Fat Red 7B was carried out according to 9. Scanning electron microscopy was performed as described previously 37.

Molecular manipulations

Molecular cloning was performed according to standard methods 38. Genomic sequences flanking the OLexA-46 promoter in the pga37 genome were identified by inverse PCR 29. A PGA37/MYB118 cDNA clone and a MYB115 (AT5g40360) genomic clone were obtained by PCR using PWO DNA polymerase (Roche Diagnostics Hong Kong and China) and appropriate primer pairs (all primers used in this study are listed in Supplementary information, Table S1). The PCR fragments were cloned into the AscI (PGA37/MYB118) or XhoI (MYB115) and SpeI sites of the estradiol-inducible vector pER10 20.

To generate the PGA37::GUS reporter construct, a 2.5-kb genomic fragment, which spanned nucleotides −2 535 to +36 with respect to the putative translation start codon, was amplified from Col-0 genomic DNA using KOD-Plus DNA polymerase (Toyobo, Osaka, Japan) and primers PGA37PF and PGA37PB. The PCR fragment, containing the coding sequence for the first 12 amino acid residues of PGA37, was in-frame fused to the GUS coding sequence in pBI101.1 using the XbaI and BamHI sites. All constructs were verified by extensive restriction digests and DNA sequencing.

Analysis of gene expression by RT-PCR was performed essentially as described previously 22.

Analysis of GUS activity

GUS staining was performed as described 39. Histochemical staining of siliques was carried out as described 40 with minor modifications. Briefly, siliques were slit twice longitudinally and incubated in 90% acetone for 20 min at room temperature. After extensive wash, the sample was incubated in a GUS assay buffer (50 mM phosphate buffer [pH 7.2], 0.1% Triton X-100, 2 mM X-Gluc, 2 mM potassium ferrocyanide and 2 mM potassium ferricyanide) for about 6 h at 37 °C. After staining, siliques were cleared in ethanol-acetic acid and then mounted in Hoyer's medium. Dissected ovules were observed and photographed under an Olympus BX51 microscope equipped with Nomarski optics and a cooled KX85 CCD camera.

( Supplementary Information is linked to the online version of the paper on the Cell Research website.)