The WRINKLED1 (WRI1) transcription factor has been shown to play a role of the utmost importance during oil accumulation in maturing seeds of Arabidopsis thaliana. However, little is known about the regulatory processes involved. In this paper, comprehensive functional analyses of three new mutants corresponding to null alleles of wri1 confirm that the induction of WRI1 is a prerequisite for fatty acid synthesis and is important for normal embryo development. The strong expression of WRI1 specifically detected at the onset of the maturation phase in oil-accumulating tissues of A. thaliana seeds is fully consistent with this function. Complementation experiments carried out with various seed-specific promoters emphasized the importance of a tight regulation of WRI1 expression for proper oil accumulation, raising the question of the factors controlling WRI1 transcription. Interestingly, molecular and genetic analyses using an inducible system demonstrated that WRI1 is a target of LEAFY COTYLEDON2 and is necessary for the regulatory action of LEC2 towards fatty acid metabolism. In addition to this, quantitative RT-PCR experiments suggested that several genes encoding enzymes of late glycolysis, the fatty acid synthesis pathway, and the biotin and lipoic acid biosynthetic pathways are targets of WRI1. Taken together, these results indicate new relationships in the regulatory model for the control of oil synthesis in maturing A. thaliana seeds. In addition, they exemplify how metabolic and developmental processes affecting the developing embryo can be coordinated at the molecular level.
Seed development is a complex process that requires the coordinated growth of three distinct tissues. Once double fertilization has occurred in the embryo sac, the development of zygotic tissues, namely the embryo and endosperm, is initiated. Several layers of maternally derived integuments constitute the seed coat. In Arabidopsis thaliana, seed formation can be divided into three main stages: pattern formation (early embryo morphogenesis) (Mayer and Jürgens, 1998), storage compound accumulation (maturation) (Baud et al., 2002), and acquisition of desiccation tolerance by the embryo which becomes metabolically quiescent (late maturation). During the maturation phase, which is characterized by a strong increase in seed dry weight (DW), the embryo accumulates high amounts of storage proteins and lipids that fuel postgerminative seedling establishment (Bewley and Black, 1994).
Triacylglycerols (TAGs), esters of glycerol and fatty acids, constitute the storage form of lipids in A. thaliana embryos. TAG biosynthesis, which relies on the use of sucrose imported from maternal tissues (Baud et al., 2005), has been relatively well characterized (Voelker and Kinney, 2001). Incoming sucrose is cleaved via two distinct pathways involving either invertases or sucrose synthases (SUS) to produce hexose phosphates. These products are then metabolized through the glycolysis and/or the oxidative pentose phosphate pathway (OPPP) to provide precursors for fatty acid production (Schwender et al., 2003). The conversion of sucrose into pyruvate through glycolysis results in the loss of one-third of the carbon as CO2. However, Rubisco has recently been shown to act without the Calvin cycle to increase the efficiency of carbon use in maturing embryos of Brassica napus (Ruuska et al., 2004; Schwender et al., 2004). De novo fatty acid synthesis occurs in the plastid and generates C16:0, C16:1 and C18:1, which can be later elongated and/or desaturated in the ER. The final step of this network involves the assembly of TAGs through the sequential acylation of a glycerol backbone by the enzymes of the Kennedy pathway (Miquel and Browse, 1997).
Although the overall biochemical pathways producing storage lipids have been extensively described, factors regulating fatty acid synthesis and controlling total oil content in oilseed crops are still poorly understood (Thelen and Ohlrogge, 2002). Likewise, questions remain regarding the interplay between this metabolism and the developmental progression of embryogenesis (Brocard-Gifford et al., 2003). Mutants affected in maturation processes therefore constitute precious tools to address these questions [for reviews, see Vicente-Carbajosa and Carbonero (2005); Wobus and Weber (1999)]. In A. thaliana, four loci have been identified, namely FUSCA3 (FUS3), ABSCISIC ACID INSENSITIVE3 (ABI3) and LEAFY COTYLEDON1 and 2 (LEC1 and LEC2), that control a wide range of seed-specific characters and play a role of the utmost importance in seed maturation. FUS3, ABI3 and LEC2 belong to the B3 family of plant transcription factors (Giraudat et al., 1992; Luerssen et al., 1998; Meinke et al., 1994; Stone et al., 2001), whereas LEC1 encodes an NFY-B factor (Lotan et al., 1998). The abi3, lec1, lec2 and fus3 mutants share common phenotypes such as reduced accumulation of storage compounds, and exhibit specific phenotypes such as the lack of chlorophyll degradation, anthocyanin accumulation, intolerance to desiccation, or defects in cotyledon identity (see Table 1 in To et al., 2006). In addition to this first group of master regulators that exhibit a broad effect on seed maturation, other genes have been isolated that have an impact on more specific aspects of the maturation process. Of these, WRINKLED1 (WRI1), which encodes a transcription regulator of the AP2/EREB family, is of particular interest. The corresponding wri1-1 mutant produces wrinkled seeds that are severely depleted in TAGs (Cernac and Benning, 2004). In this low-seed-oil mutant, carbohydrate metabolism is compromised (Baud and Graham, 2006; Focks and Benning, 1998), rendering maturing embryos unable to efficiently convert sucrose into TAGs. Microarray data obtained from wri1-1 seeds indicated several interesting changes in carbon metabolism in response to disruption of the WRI1 gene (Ruuska et al., 2002). These results suggest that WRI1 may participate in the developmental control of storage metabolism during seed filling, but the mechanism involved remains to be elucidated. Interestingly, recent studies have demonstrated that sucrose can induce WRI1 in leaves (Masaki et al., 2005), and that WRI1 affects ABA sensitivity and sugar sensing but not through ABI3 in young seedlings (Cernac et al., 2006). The occurrence of similar regulatory cross-talk between WRI1 and sucrose or ABA in maturing embryos has yet to be validated.
In order to gain further insights into the control of oil metabolism during seed maturation in A. thaliana, the regulation of WRI1 was further investigated. Comprehensive characterization of three new wri1 alleles confirmed that strong WRI1 expression is a prerequisite for oil accumulation. Analyses of the tissue-specific and temporal expression of WRI1 were consistent with these results. Interestingly, using various molecular and genetic tools, we demonstrated that expression of WRI1 is controlled by the master regulator LEC2. In addition, it is shown that WRI1 is necessary for the accumulation of lipids in LEC2-over-expressing leaves. Finally, new putative targets of WRI1 were identified that encode members of the lipid biosynthetic network. Taken together, these results provide further insights into the understanding of the regulatory model controlling oil synthesis in A. thaliana seeds.
Comprehensive characterization of wri1 null alleles
The wri1-1 mutant isolated and extensively studied by Focks and Benning (1998) has been shown to exhibit a point mutation in the first intron of the WRI1 gene (Figure 1a), resulting in a splicing defect and leading to the accumulation of aberrant WRI1 transcripts in maturing wri1-1 seeds (Cernac and Benning, 2004). To investigate the effects of wri1 knockout mutations on oil accumulation in seeds, three null alleles were isolated and finely characterized. wri1-3 (N585693) and wri1-4 (N508559) were identified in the SALK collection (Alonso et al., 2003) and exhibited T-DNA insertions in the 5th and 6th introns of WRI1, respectively (Figure 1a). In the wri1-5 line (EAA5), isolated from the Versailles collection (Bechtold et al., 1993), both WRI1 and the neighbouring APETALA3 gene were truncated as a consequence of insertion of the T-DNA. These three new mutant lines produced severely wrinkled seeds (Figure 1b) and were shown to be allelic to wri1-1 (see Experimental procedures). To check that the three new wri1 alleles were null, WRI1 transcript levels were analysed by quantitative RT-PCR in seeds or embryos harvested 10 days after anthesis (DAA) (Figure 1c). In wri1-3 and wri1-4 seeds, relative WRI1 transcript levels were severely decreased compared with the wild-type. Using various sets of primers, it was possible to show that the transcripts detected in these lines consisted of a mixture of chimeric WRI1–T-DNA transcripts, together with traces of wild-type-like WRI1 transcripts, indicating that T-DNAs could be spliced on certain occasions (data not shown). WRI1 transcripts were barely detectable in wri1-5 embryos, and it was possible to establish that the transcripts detected in this material were all chimeric WRI1–T-DNA transcripts (data not shown). Contrary to the previously described wri1-1 allele, that accumulated aberrant WRI1 transcripts, the three new wri1 alleles appeared to be depleted in WRI1 transcripts.
To further characterize these new alleles, cytological observations were carried out during early embryonic morphogenesis and early maturation. The structure of the three types of tissue comprising the seed, i.e. integuments, endosperm and embryo, appeared normal. However, a reproducible delay in embryo elongation was observed in wri1 mutants. wri1 embryos reached the early-torpedo stage at 8 DAA (Figure 1d), when wild-type embryos were already late-torpedo-shaped (Figure 1e). At 10 DAA, embryo curvature was not as pronounced in wri1 mutants as in the wild-type (Figure 1f,g). Finally, at 12 DAA, wri1 embryos were smaller than their wild-type counterparts (Figure 1h,i). Dry desiccated seeds were then assessed to evaluate the impact of these mutations on seed maturation. A significant reduction in mutant seed DW was noted, ranging from 25–40% depending on the allele considered (Figure 1c). Fatty acid concentrations were then determined (Figure 1c). A 45–55% loss in fatty acid content was measured in the mutant lines. This reduction in oil content was accompanied by a significant modification of seed fatty acid composition. The relative amounts of the end products of desaturation (C18:3; data not shown) and elongation (C22:1; Figure 1c) were considerably increased, whereas intermediate fatty acid species in the desaturation and elongation pathways (C18:1, C18:2 and C20:1) were proportionally decreased. Taken together, these results strengthen the previous data obtained with the wri1-1 allele and confirm the effects of loss of function of WRI1.
Tissue specificity and time-course analysis of WRI1 expression
To gain further insights into the spatio-temporal regulation of WRI1 expression, several complementary techniques were used. A quantitative RT-PCR approach firstly confirmed that WRI1 was detectable at low levels in vegetative organs and highly accumulated in siliques (data not shown). The relative accumulation profile of WRI1 was further investigated in a developmental series of siliques ranging from 4 to 18 DAA (Figure 2a). WRI1 relative transcript levels were low during early silique development. A sharp peak in expression was observed from 8 to 10 DAA, corresponding to the onset of seed maturation. Then mRNA levels gradually decreased throughout the maturation phase. An in situ hybridization experiment was carried out to determine the precise spatial localization of WRI1 transcripts (Figure 2b–d). In early maturing seeds containing early- or late-torpedo-stage embryos, a signal could be detected both in embryonic tissues and in the endosperm; WRI1 transcripts were not detected in either seed coats or silique walls.
The spatio-temporal activity of the WRI1 promoter was then investigated. A 1028 bp fragment (referred as to ProWRI1) was fused translationally to the uidA reporter gene. The chimeric construct was assayed for the resulting expression pattern in transgenic A. thaliana lines. Beta-glucuronidase (GUS) staining was first detected in the endosperm and in the embryo of early maturing seeds (Figure 2e). In torpedo-shaped embryos, staining was intense in the hypocotyl and in the outer parts of young cotyledons (Figure 2f,g). Staining then gradually spread throughout the expanding cotyledons (Figure 2h,i). Sections of torpedo-shaped embryos showed that both the epidermis and the inner cell layers of the embryo were stained (data not shown); this was fully consistent with the results of in situ hybridization experiments (see above).
Importance of the WRI1 expression pattern
To evaluate the importance of WRI1 expression, complementation experiments were carried out with WRI1 cDNA placed under the control of promoters from genes exhibiting distinct seed expression patterns. Homozygous wri1-1 and wri1-4 mutants were first transformed with WRI1 cDNA, the expression of which was driven by its own promoter. Five independent primary transformants were selected in each of the mutant backgrounds considered, and their progeny were subjected to fatty acid analyses. Comparison of fatty acid concentrations measured in F1 progeny with the fatty acid content of mutant and wild-type seeds allowed a percentage of phenotypic reversion to be calculated (see Experimental procedures). If fully complemented, segregating seeds of F1 progeny are expected to exhibit at least a 75% reversion of the mutant phenotype (one insertion locus). Full complementation of both wri1-1 and wri1-4 alleles was obtained with the WRI1 promoter (Figure 3). When using the TT8 promoter, which is highly induced in the integuments of early developing seeds and slightly expressed in the hypocotyls of mature embryos (Baudry et al., 2006), very poor complementation of the wri1 phenotype was observed. When using the promoters of genes specifically expressed in zygotic tissues of the seed just prior to (AtSUC5) (Baud et al., 2005) or at the onset of seed maturation (LEC2 and FUS3) (Luerssen et al., 1998; Stone et al., 2001), the complementation observed was only partial. These results suggested a cell-autonomous activity of WRI1 and outlined the importance of its specific expression pattern.
WRI1 is a target of LEC2 and is necessary for LEC2 function
The fine characterization of wri1 mutants as well as transgenic lines over-expressing WRI1 clearly showed that induction of WRI1 triggers fatty acid accumulation (Cernac and Benning, 2004; Focks and Benning, 1998; this paper). However, the regulatory mechanisms involved remained to be elucidated. Analysing seed transcriptomic data displayed on GENEVESTIGATOR (http://www.genevestigator.ethz.ch), we observed that the peak in WRI1 expression (see above) was highly correlated with induction of LEC2 during the time course of seed development. We thus decided to introduce the ProWRI1:uidA construct in a lec2 mutant background to test whether this mutation that is known to prevent seed maturation could affect the WRI1 expression pattern. At the torpedo stage, WRI1 expression appeared to be restricted to the hypocotyl of lec2 embryos (Figure 4a). Contrary to what was observed in the wild-type, WRI1 expression remained localized in this tissue, never spreading to the cotyledon-like structures developed by mutant embryos at further developmental stages (Figure 4b,c). As WRI1 is already highly expressed in maturing seeds of the wild-type, testing the induction of WRI1 transcription by LEC2 over-expression in this tissue was not convenient. Santos Mendoza et al. (2005), using a chimeric fusion of LEC2 to the rat glucocorticoid receptor (GR), recently demonstrated that ectopic activation of LEC2 was able to trigger oil accumulation in A. thaliana leaves. We consequently used this system to investigate the relative expression level of WRI1 in vegetative organs of transgenic plants expressing the Pro35S:LEC2:GR construct. Leaves of plantlets grown for 2 weeks in vitro and then transferred onto a medium containing dexamethasone (DEX; a synthetic glucocorticoid that activates GR) were subjected to quantitative PCR analyses. From 3 days after induction onwards, a regular increase in WRI1 mRNA level could be observed (Figure 5a), whereas no induction could be detected in control plants (data not shown). A maximum was reached after 6 days of induction, the level of which (5.7% of the level of EF1αA4, i.e. constitutive expression) corresponded to values usually detected in early maturing seeds of wild-type plants (Figure 2a). To determine whether WRI1 transcript accumulation was under the direct control of LEC2:GR, transgenic plantlets expressing the Pro35S:LEC2:GR construct were treated with a solution containing both DEX and cycloheximide (CHX; an inhibitor of protein synthesis). This treatment allows the expression of direct target genes but the indirect activation of genes requiring de novo protein synthesis is blocked (Baudry et al., 2006). After 6 h of DEX treatment, a significant accumulation of WRI1 mRNAs was observed (Figure 5b), confirming the results obtained with plants grown on a DEX-containing medium (Figure 5a). This induction was absent in the controls, i.e. wild-type plants and transgenic lines expressing the Pro35S:TT8:GR construct (Baudry et al., 2004). After 6 h of a CHX treatment, a strong induction of WRI1 expression was seen in the three genotypes. When treated with both DEX and CHX, a further induction was observed specifically in Pro35S:LEC2:GR transgenic plants, demonstrating in planta that WRI1 is a direct target of LEC2.
These results strongly suggested that the accumulation of lipids in leaves expressing LEC2 ectopically (Santos Mendoza et al., 2005) may be a consequence of WRI1 expression. To confirm this hypothesis, the LEC2:GR fusion was introduced in wri1 backgrounds and the ability of the transgenic plants to accumulate lipids was tested. Two-week-old seedlings were transferred and cultured for 13 days on DEX-containing medium; their leaves were then subjected to fatty acid analyses. Lines 25 and 31 (two independent Pro35S:LEC2:GR transgenic lines) exhibited oil accumulation in their rosette leaves. This was characterized by an increase in total fatty acid content compared with the wild-type and by the production of very long chain fatty acids (VLCFA) (Figure 5c). In wri1-1 and wri1-4 backgrounds, Pro35S:LEC2:GR expression had no significant effect on leaf fatty acid content, demonstrating that WRI1 is necessary for LEC2-induced oil accumulation. We finally generated a Pro35Sdual:WRI1 construct that was introduced into a lec2 background to test whether WRI1 ectopic expression would be sufficient to restore lipid accumulation. Although this construct had been shown to be functional (by complementing wri1 mutants; see Experimental procedures), it was unable to complement the defect in lipid accumulation in the lec2 mutant background. The plants obtained were stunted, partially sterile (phenotypes that were similar to the ones previously observed when introducing the Pro35Sdual:WRI1 construct into a Ws wild-type background), and produced seeds that still exhibited a lec2 phenotype (data not shown). These results demonstrate that WRI1 is necessary for the accumulation of lipids controlled by LEC2, but is not sufficient to induce lipid accumulation in seeds in the absence of LEC2.
Identification of putative new targets of WRI1
To gain further insights into the mechanism controlling fatty acid accumulation in seeds, we used a quantitative PCR approach to determine putative new targets of WRI1. The microarray data obtained from wri1-1 seeds (Ruuska et al., 2002) showed some interesting changes in the expression levels of genes involved in carbon metabolism. This is consistent with alterations observed in mutant seeds both at the enzymatic (Baud and Graham, 2006; Focks and Benning, 1998) and biochemical levels (Focks and Benning, 1998; this paper). The relative mRNA levels of genes encoding enzymes involved in carbohydrate and/or fatty acid metabolism were thus investigated in a set of plant tissues exhibiting a wide range of WRI1 transcript levels. Whereas WRI1 was severely down-regulated in wri1-4 seeds and wri1-5 embryos compared with their wild-type counterparts (Figure 1c), transgenic Pro35Sdual:WRI1 lines T63 and T81 (see Experimental procedures) over-expressed WRI1 in their vegetative parts. We assumed that the expression profiles of WRI1 targets would exhibit similar patterns.
A set of genes involved in sucrose cleavage/hexose phosphate production pathways were firstly considered. Neither the UGPase-encoding gene At5g17310 (Figure 6) nor HXK2 (At2g19860; data not shown) exhibited altered expression levels in response to the up- or down-regulation of WRI1 expression. Interestingly, AtSUS2 (At5g49190) was strongly repressed in wri1 seeds, although no AtSUS2 transcripts could be detected in the leaves of WRI1 over-expressers. We then tested genes encoding plastidic enzymes involved in late glycolysis or de novo fatty acid synthesis: pyruvate kinases (At5g52920 and At3g22960), pyruvate dehydrogenase (PDH) E1α subunit (At1g01090), acetyl CoA carboxylase (ACCase) BCCP2 subunit (At5g15530), an enoyl-ACP reductase (At2g05990) and a phosphoglycerate mutase (At1g22170; data not shown). The whole set of genes was significantly induced in the leaves of Pro35Sdual:WRI1 transgenic plants, reaching expression levels normally attained in maturing seeds of the wild-type. Consistent with this was the drastic repression of these genes observed in wri1 seeds, where they exhibited the basal levels usually found in wild-type leaves. These genes thus appeared to be putative targets of WRI1. LAS (LIPOIC ACID SYNTHASE, At5g08415) and BIO2 (At2g43360) encode enzymes catalysing the last steps of the lipoic acid and biotin biosynthetic pathways, respectively. These co-factors are known to be required by PDH and ACCase, respectively. Interestingly, both LAS and BIO2 expression levels correlated with WRI1 mRNA levels in the tissues analysed, just like the set of genes described above. Genes involved in fatty acid desaturation (FAD2, At3g12120; data not shown) or elongation (ACC1, At1g36160; FAE1, At4g34520) were also considered, but alterations in WRI1 mRNA content had no significant effect on the accumulation level of the corresponding mRNAs.
Factors controlling the overall amount of fatty acids stored in seeds and integrating this biochemical process into the complex framework of seed development are largely unknown. In this paper, we report new insights into the regulation of WRI1, a transcription factor involved in the control of carbon metabolism during seed maturation. Comprehensive characterization of three new wri1 alleles provides additional evidence that induction of WRI1 constitutes a prerequisite for oil accumulation in seeds. The detailed analysis of WRI1 expression pattern together with complementation experiments of wri1 mutants (using various promoters) emphasize the importance of the fine spatio-temporal regulation of WRI1 mRNA accumulation in seeds. Finally, unravelling the transcriptional control exerted by the master regulatory factor LEC2 over WRI1, together with the key role played by WRI1 in LEC2 function, and, additionally, identifying putative new targets of WRI1 amongst the fatty acid biosynthetic network, provide important new insights into the regulatory model for the control of oil synthesis in maturing A. thaliana seeds.
Induction of WRI1 constitutes a prerequisite for oil accumulation
Characterization of wri1-1 (Cernac and Benning, 2004; Focks and Benning, 1998) shed promising light on the understanding of oil synthesis in seeds. The mutant was shown to accumulate aberrant WRI1 transcripts (with intron 1 unspliced). However, as wri1-1 was recessive, it was very likely that this mutation actually corresponded to a loss of function. This point is convincingly addressed in this paper with the isolation of three new wri1 mutants. These null alleles have been fully characterized at the molecular, cytological and biochemical levels. Like wri1-1, they produce wrinkled seeds, the fatty acid content of which is drastically reduced (45–55% loss on a concentration basis). Together with the analysis of transgenic lines expressing WRI1 ectopically and over-accumulating fatty acids in vegetative organs (Cernac and Benning, 2004; this paper), the fine characterization of these null wri1 alleles confirms the results of the elegant feeding experiments performed by Focks and Benning (1998) that established that induction of WRI1 constitutes a prerequisite for oil accumulation. The marked defect in TAG accumulation observed in wri1 mutants has an impact on embryo development, the elongation of which appears to be delayed. Mutations affecting the ability to furnish sucrose to zygotic tissues (suc5 mutants) (Baud et al., 2005) or affecting the ability to use this precursor to synthesize fatty acids (wri1 mutants) have a similar negative impact on embryo elongation. These results suggest that developmental and metabolic aspects are tightly associated during early maturation.
Importance of the WRI1 expression pattern
Both non-quantitative and quantitative RT-PCR studies used in this study indicate that WRI1 mRNA accumulates to a high level in A. thaliana seeds, but is only slightly detectable in vegetative organs. This is consistent with the conclusions of Cernac and Benning (2004) who used a Northern blot strategy that demonstrated strong accumulation of WRI1 mRNA in seeds. Fine characterization of the WRI1 expression pattern based on the use of complementary approaches such as real-time PCR, in situ hybridization and promoter:GUS analyses demonstrated that WRI1 transcription is finely regulated both in time and space. The mRNA accumulation of this gene dramatically increases 7 or 8 DAA, at the onset of the maturation phase, before decreasing consistently during late maturation. It should be noted that the late increase in mRNA accumulation observed towards the end of storage compound accumulation by Northern blot (Cernac and Benning, 2004) was not detected by real-time PCR; complementary analyses are currently being undertaken to elucidate this discrepancy. Activation of the WRI1 promoter is concomitant with embryo elongation from torpedo to upturned-U stage, at a time when fatty acid accumulation is initiated in the seed. When analysing the tissue specificity of WRI1 expression, it appears that the gene is expressed in both the endosperm and the embryo, but not in maternal integuments. Interestingly, both zygotic tissues are known to be the site of TAG deposition in seeds (Penfield et al., 2004), thus confirming at the spatial level the correlation existing between WRI1 expression and the synthesis of storage oil. In addition to this, the complementation experiments carried out on wri1 mutants using various seed promoters confirmed the importance of the WRI1 expression pattern for proper oil accumulation in the seed. Ultimately, questions arise concerning the regulatory mechanisms involved and the identity of the regulator(s) exerting transcriptional control on WRI1.
WRI1 specifies the regulatory action of LEC2 towards fatty acid synthesis
Among the regulators identified to date that are known to exert control over seed maturation, LEC2 is of particular interest for those who are studying the accumulation of storage lipids in embryonic tissues (Santos Mendoza et al., 2005; Stone et al., 2001). Transcriptomic data indicate that induction of LEC2 correlates with the onset of oil accumulation in A. thaliana seeds. Ectopic expression of LEC2 can confer embryonic characteristics to transgenic seedlings (Stone et al., 2001), triggering TAG accumulation in developing leaves (Santos Mendoza et al., 2005). A recent report (Braybrook et al., 2006) described a set of genes directly controlled by LEC2, providing original insights into the transcriptional control of embryo maturation. Among the targets of LEC2, AGL15 and IAA30 may represent key factors mediating the regulatory action of LEC2 on morphogenetic processes affecting developing embryos. The direct control exerted by LEC2 over EEL, At2S1, At2S2, At2S3, At2S4 and 2SLike further elucidates the link between LEC2 and the accumulation of storage proteins (Kroj et al., 2003). Although several genes encoding lipid-body proteins have also been identified, there were no clues to explain/hypothesize how LEC2 could possibly impact on lipid synthesis. Here, we demonstrate that WRI1 is a direct target of LEC2 that specifies the regulatory action of the master regulator towards the fatty acid biosynthetic network, such that WRI1 is necessary for LEC2-induced oil accumulation.
Interestingly, WRI1 transcription is not completely abolished in a lec2 background, but is restricted to the hypocotyl of the embryo. Consistent with this result is the high abundance of lipid bodies observed in the hypocotyl of lec2 embryos (Meinke et al., 1994). To et al. (2006) have recently demonstrated that LEC2, FUS3 and ABI3 are involved in a local, highly redundant gene regulation network governing most seed maturation aspects. From this genetic framework, it is clear that other master regulators could trigger WRI1 expression in maturing embryos. This hypothesis is consistent with the up-regulation of WRI1 in the turnip mutant that ectopically expresses LEC1 but not LEC2 (Casson and Lindsey, 2006). Finally, Cernac et al. (2006) recently established that sugar sensing is affected in wri1-1 seedlings, while Masaki et al. (2005) demonstrated that sucrose is able to enhance WRI1 expression in leaves. Cernac et al. (2006) proposed that WRI1 might constitute a mechanistic component of the sugar response, participating in sugar signalling through the hexokinase pathway. It would be very interesting to test whether a similar regulation occurs in maturing embryos, and whether LEC1 participates in the mediation of this sucrose response (Casson and Lindsey, 2006). Finally, the induction of WRI1 observed on CHX (Figure 5b) suggests the existence of negative regulators of WRI1 with a rapid turnover in vegetative organs of the plant, adding a degree of complexity to this new and intricate regulatory scheme.
WRI1 has an impact on the transcription level of genes essential for fatty acid synthesis
To unravel how WRI1 specifies the regulatory action of LEC2 towards fatty acid synthesis, identification of WRI1 targets is necessary. Using a quantitative PCR approach and investigating the impact of alterations in WRI1 expression on the accumulation of mRNAs encoding enzymes of carbon metabolism, it was possible to isolate a set of genes the expression of which is perfectly correlated with the WRI1 mRNA level. Among these putative WRI1 targets, the first subset of genes encodes enzymes predicted to be plastid-targeted and to catalyse reactions of late glycolysis and the fatty acid synthesis network (Figure 7). These results not only confirm some of the transcriptomic data previously obtained on wri1-1 seeds (Ruuska et al., 2002), but also indicate new targets involved in carbon metabolism such as pyruvate kinase (At3g22960) or phosphoglycerate mutase (At1g22170). This quantitative PCR approach also allowed the isolation of putative targets of WRI1 controlling biosynthetic pathways producing co-factors essential for PDH and ACCase activities, namely lipoic acid and biotin. Enzymes catalysing the ultimate reactions of these two pathways, namely lipoic acid synthase (Gueguen et al., 2000) and biotin synthase (Patton et al., 1998;Picciocchi et al., 2003), are encoded by two genes, the mRNA levels of which are perfectly correlated with the WRI1 transcript level.
Analysing various plant tissues with a broad range of variations in WRI1 expression level appears to be necessary to correctly identify putative targets of this transcription factor. For instance, AtSUS2 was reported to be down-regulated in wri1-1 seeds (Ruuska et al., 2002), but this must illustrate an indirect adaptation of carbon metabolism to the wri1 mutation as the gene is not induced in lines over-expressing WRI1. Alternatively, another seed-specific factor cooperating with WRI1 may be required to trigger AtSUS2 mRNA accumulation. Interestingly, using ExpressionAngler (http://bbc.botany.utoronto.ca/ntools) for transcriptomic analyses, it appears that most of the genes encoding the set of enzymes necessary to metabolize 3-phosphoglycerate and ultimately synthesize fatty acids share a similar expression pattern, making them all good putative targets of WRI1 in oil-accumulating tissues. Most of the enzymes involved in de novo fatty acid synthesis seem to be controlled by WRI1, but the situation seems quite different for those involved in non-plastidial fatty acid modifications. Expression of elongases and desaturases is not controlled by WRI1. This is fully consistent with the fatty acid composition of tissues exhibiting altered WRI1 transcription levels. In wri1 seeds, the fatty acid content is very low due to a reduced carbon flow through the fatty acid synthesis network, but the end products of fatty acid desaturation and elongation are over-represented, as these two pathways are unaffected by wri1 mutations. In contrast, in the leaves of WRI1-over-expressing lines, induction of the fatty acid synthesis network leads to an increase in fatty acid concentration. However, highly desaturated and/or elongated species are not over-accumulated in our experiments (data not shown). Synthesis and elongation/desaturation of fatty acids thus appear to be regulated separately in oil-accumulating seeds of A. thaliana, and this provides a possible explanation for the lack of feedback regulation to adjust the degree of unsaturation and elongation of the fatty acid substituents of TAGs to the flow of carbon into this pathway (Baud et al., 2005; Focks and Benning, 1998).
To conclude, a model is proposed for the regulation of fatty acid synthesis in maturing seeds of A. thaliana (Figure 7). In this framework, WRI1 appears as a node, acting at the interplay between the master regulatory element LEC2 and fatty acid metabolism to trigger sustained oil production. This model provides insight into the complex coordination of metabolic and developmental processes affecting developing embryos. An in-depth functional study of the promoters of some putative targets of WRI1 will have to be performed to determine the cis-element bound by the transcription factor and thus the precise molecular mechanism involved in this regulation. These analyses would also pave the way for the isolation of new partners of WRI1 involved in the overall regulatory network.
Plant material and growth conditions
Arabidopsis thaliana seeds of the ecotypes Wassilewskija (Ws) and Columbia (Col) were obtained from the Station de Génétique et d’Amélioration des Plantes (INRA, Versailles, France). Seeds were surface-sterilized and germinated on MS medium (Murashige and Skoog, 1962). After cold treatment at 4°C for 48 h in the dark, plates were kept in a growth chamber (16 h photoperiod; 15°C night/20°C day). After 10 days, the plantlets were transferred to compost, grown in a greenhouse under similar conditions, and irrigated twice a week with mineral nutrient solution. Material used for RNA extraction was frozen in liquid nitrogen immediately after harvest and stored at –80°C prior to extraction. DW measurements of leaf samples were performed after drying at 50°C for 48 h. Weight determination of leaf and seed samples was performed on a M2P balance (Sartorius, http://www.sartorious.com).
Genetic and molecular characterization of wri1 alleles
Mutants exhibiting T-DNA insertions in the WRI1 gene were isolated by reverse genetics. Two flanking sequences tags (FSTs) corresponding to left T-DNA borders that were anchored to the genome sequence of A. thaliana in the WRI1 gene were isolated from the SALK transformant collection; a third FST anchored in the same WRI1 gene was identified in the Versailles transformant collection. Plant genomic DNA flanking the left T-DNA borders of the corresponding mutants were amplified by PCR and sequenced, confirming the FSTs identified. In the wri1-3 line (N585693), the T-DNA insertion was located in the 5th intron of the WRI1 gene and resulted in a 10 bp deletion of genomic DNA at the insertion site (Figure 1a). In the wri1-4 line (N508559), the T-DNA insertion occurred in the 6th intron. Although the right T-DNA border could not be identified in this line, we checked by PCR experiments that no major deletion had occurred at the insertion site. In the wri1-5 line (EAA5), the T-DNA insertion was located in the 5th intron of WRI1. The insertion event generated a 3.7 kb deletion of genomic DNA such that both WRI1 and At3g54340 (APETALA3) were truncated in this mutant line.
Reciprocal crosses between wri1-1, wri1-3 and wri1-4 mutants were realized, and no complementation of the wrinkled seed phenotype was observed in F1 progeny, demonstrating that the three mutant lines were allelic. Due to the large deletion in the APETALA3 gene exhibited by the wri1-5 mutant, homozygous lines were male-sterile (Bowman et al., 1989); the wri1-5 mutation was thus maintained in a heterozygous state. When analysing wri1-5/WRI1 plant progeny, 25% of the seeds observed (n = 1615) exhibited a wrinkled phenotype. This corresponds to the ratio expected in the case of a normal recessive Mendelian trait (χ2 = 0.21, not significant , α < 0.05). As the sterility of wri1-5 is due to a transmission defect on the male side, crosses could be carried out between homozygous wri1-5 lines used as the female parent and each of the three wri1 mutants previously mentioned. No complementation of the wrinkled seed phenotype could be observed in F1 progeny, demonstrating that wri1-5 was allelic to the other wri1 mutations.
Nucleic acid analyses
DNA extraction, amplification and sequencing were performed according to the methods described by Baud et al. (2005). To amplify the T-DNA left border in wri1-5, primers TailA (5′-ATTGCCTTTTCTTATCGACC-3′) and 5693LB1 (5′-GGAGGAAATGCAGAGAGTGAC-3′) were used. The opposite border was amplified with TailA and EaaBlow (5′-ATATGTGTTTCGCTGGTTGG-3′); this border appeared to be a second left border, indicating that a complex insertion event occurred in the eaa5 line. To amplify the T-DNA left border in wri1-3, primers SigLB1 (5′-CGGAACCACCATCAAACAG-3′) and 5693LB1 were used. The opposite border was amplified with SigLB1 and 5693RB1 (5′-ATCTTCCGTTGTGGTGATGC-3′); this border also appeared to be a second left border, indicating that a complex insertion event occurred in wri1-3 line. To amplify the wri1-4 left border, primers 8559LB2 (5′-TGGTCCTAAAGCCTCTTTGAAC-3′) and SigLB1 were used.
Constructs and transformant analysis
Construction of the ProWRI1:uidA transgene. The WRI1 promoter used corresponds to the region -1028 to -1 bp relative to the WRI1 translational start codon, and was amplified using the proofreading Pfu Ultra DNA polymerase (Stratagene; http://www.stratagene.com/) from Ws genomic DNA using 5′-attB1-TGTCTGCTTAAATCATGTG-3′ and 5′-attB2-TAAACTCTGAGAAAGTTTAG-3′; attB1 and attB2 refer to the corresponding Gateway recombination sequences. The PCR product was introduced by BP recombination into the pDONR207 entry vector (Invitrogen; http://www.invitrogen.com/) and transferred to the binary vector pBI101-R1R2-GUS (F. Divol, J.-C. Palauqui, LBC-IJPB, and B. Dubreucq, unpublished data) by a LR recombination reaction to obtain a transcriptional fusion between the WRI1 promoter and the uidA gene. The resulting binary vector was electroporated into the Agrobacterium tumefaciens C58C1GV3101 (pMP90) strain (Koncz and Schell, 1986) and used for agro-infiltration of A. thaliana inflorescences (Bechtold et al., 1993). Ten transformants were selected on MS medium containing kanamycin (50 mg l−1) and then transferred to soil for further characterization.
Construction of the Pro35Sdual:WRI1 transgene. WRI1 cDNA was amplified from a mixture of seed cDNAs (of the Ws ecotype) using 5′-attB1-TAATGAAGAAGCGCTTAACCACT-3′ and 5′-attB2-TCAGACCAAATAGTTACAAGA-3′, cloned in pDONR207, and transferred to the binary vector pMDC32 (Curtis and Grossniklaus, 2003) as previously described. This construct was first tested for its ability to complement the wri1-1 mutation. Six independent hygromycin-resistant wri1-1 plants carrying the Pro35Sdual:WRI1 construct were selected and their progeny subjected to fatty acid analyses. A significant restoration of both fatty acid concentration and composition was measured in all six lines, demonstrating that a functional WRI1 protein was synthesized. Wild-type plants of the Ws ecotype were then infiltrated with this construct. We generated and analysed transgenic plants carrying the Pro35Sdual:WRI1 construct in a Ws wild-type background to avoid any indirect affect of the wri1 mutation. Segregation analyses amongst the progeny allowed selection of four primary transformants with a single insertion locus, which were further characterized. Using a quantitative PCR strategy, the WRI1 expression level was quantified in the rosette leaves of homozygous transgenic lines. Expression levels ranging from 100 to 400% of EF1αA4 (EF) were detected, indicating that WRI1 was efficiently ectopically expressed in the vegetative organs of these plants. Lines T63 and T81, which exhibited the highest WRI1 mRNA levels, were chosen for further analyses. Increased fatty acid concentrations in the leaves of these transgenic lines confirmed that WRI1 was efficiently synthesized in these vegetative organs. It should be noted that these transgenic lines looked stunted and were partially sterile. Although the amount of seeds produced per plant was drastically reduced, seed DW and fatty acid content were unmodified.
Construction of pBIB-R1R2-WRI1. WRI1 cDNA was amplified using the proofreading Pfu Ultra DNA polymerase (Stratagene) from a mixture of seed cDNAs (of the Ws ecotype) with BamAtgWri (5′-CGCGGATCCATGAAGAAGCGC-3′) and SstIStopWri (5′-CGCGAGCTCTCAGACCAAATAG-3′), primers that had been modified by the respective additions of BamHI and SstI restriction sites at their 5′ ends. The PCR product was digested with BamHI/SstI and cloned into BamHI/SstI-digested pBluescript SK+ (Stratagene). A blunt R1R2 cassette (Invitrogen) was then cloned into the construct thus obtained, digested with HincII. The R1R2–WRI1 cassette was released from pBluescript SK+ using XhoI/SstI and cloned into SalI/SstI-digested pBIB-HYG (Becker, 1990).
Construction of ProWRI1:WRI1, ProLEC2:WRI1, ProFUS3:WRI1, ProTT8:WRI1 and ProAtSUC5:WRI1 transgenes. WRI1 (see above), AtSUC5 (Baud et al., 2005), TT8 (Baudry et al., 2006), FUS3 and LEC2 (1 kb promoters; M. Santos Mendoza, unpublished data) promoters were amplified from DNA, cloned into pDONR207 by a BP recombination reaction, and finally transferred to the binary vector pBIB-R1R2-WRI by an LR recombination reaction. Transformants were selected on MS medium containing hygromycin (50 mg l−1) and then transferred to soil. The fatty acid contents of their progeny (FAT1) were determined and compared with fatty acid concentrations in mutant (FAwri) and wild-type (FAWT) seeds from control plants grown in the same conditions. The percentage of phenotypic reversion (PRT1) was calculated as follows: PRT1 = (FAT1 - FAwri)/( FAWT - FAwri) x 100.
In situ hybridization experiments were carried out as described by Lefebvre et al. (2006). The WRI1 fragment used as a probe was amplified from a cDNA mixture of the Ws ecotype using primers DYP133-180U (5′-CCGACGCAGCTCTATCTACAGAGG-3′) and DYP133-E (5′-CTATAGGTGCCGAGGTACAAG-3′), and subsequently cloned into EcoRV-digested pBluescript SK+ (Stratagene). A single-strand RNA probe was synthesized after linearization of plasmid DNA carrying the WRI1 fragment by digestion using BamHI. Control experiments were carried out using the SHOOT MERISTELMLESS (STM) probe (Vernoux et al., 2000).
Microscopy and image analysis
Developing seeds or excised embryos were placed in a quick clearing solution of chloral hydrate for 3-5 h (Boisson et al., 2001), and then observed under differential interference contrast optics using a light microscope (Axioplan 2; Zeiss; http://www.zeiss.com/). Photographs were taken using a ProgresResC10Plus image capture system (Jenoptik, http://www.jenoptik.com). Images were then prepared with Adobe Photoshop (http://www.adobe.com). For histochemical detection of GUS activity, tissues were incubated in 0.1 M phosphate buffer, pH 7.2, containing 2 mM 5-bromo-3-indolyl-β-d-glucuronide (X-Gluc; Duchefa, http://www.duchefa.com), 0.1% v/v Triton X-100 in water, 2 mM each of potassium ferrocyanide and potassium ferricyanide, depending on the construct, and 10 mM Na2-EDTA. Vacuum was applied for 1 h before incubating for 3 h at 37°C in the dark.
For seed fatty acid analyses, pools of 20 dry seeds were ground in a glass reaction tube prior to addition of methanol/sulphuric acid (100:2.5, v/v). For leaf fatty acid analyses, frozen material was placed directly in methanol/sulphuric acid in a glass reaction tube without any grinding. Extraction and analyses of fatty acyl methyl esters by gas chromatography were performed as described previously (Baud et al., 2002).
RNA extractions and quantitative RT-PCR studies were performed as previously described (Baud et al., 2004). The results obtained were standardized to the level of expression of the constitutive EF1αA4 gene (Nesi et al., 2000), amplified using the EF1F and EF1R primers (Baud et al., 2003). The efficiencies of the various primer sets used were checked and were almost similar. Primer sequences were as follows: FAE10, 5′-CTCTACATCGTAACCCGACCCA-3′; FAE11, 5′-TCGAGCGAGGACGGATCAT-3′; EAR10, 5′-TGGGACTTGGGTTCCTGCAC-3′; EAR11, 5′-CGCTTATTCGTTTTCACATCTTCAGGC-3′; BCCP10, 5′-GACCCGGTGAACCCCCT-3′; BCCP11, 5′-GTCAACGCTGACTGGTTTTCCAT-3′; UGPase10, 5′-AGTCCATCCCGAGTATAGTTGAGC-3′; UGPase11, 5′-CGAGCACGGCATTGTCAGG-3′; BIO10, 5′-GGAACCGCACGTATTGTAATGCC-3′; BIO11, 5′-CGTCAAAATCATTGTTTGGTGTGGTTAAAAGC-3′; PDH10, 5′-ATGTGTGCTCAAATGTATTACCGAGGC-3′; PDH11, 5′-ACCTTTGCTGAGGGCATGG-3′; LAS10, 5′-TCCTCCTCCTGACCCAATGGAAC-3′; LAS11, 5′-GTCTCTTCATAGCTTTGACAGTCTGCG-3′; PKP10, 5′-AGTCACTATCGTCCTTCCG-3′; PKP11, 5′-CTGTACGATTGCTATTTCCTC-3′; primers used to amplify AtSUS2 were as described by Baud et al. (2003). The QuantiTect primer assay (200) QT00791707 (Qiagen; http://www.qiagen.com/) was used to amplify WRI1.
We are grateful to M. Miquel for helpful discussions and technical advice for fatty acid analyses. J. Kronenberger, K. Madiona and N. Berger are acknowledged for their precious assistance. We thank A. Cernac and C. Benning (MSU, East Lansing, MI, USA) for kindly providing us with wri1-1 seeds, and A. Baudry for the gift of the TT8 promoter. This project has received grants from the Génoplante II AF2001019031 and TRIL-033 (P-009 Arabido-seed) projects.