- © 2005 American Society of Plant Biologists
Abstract
We identified the Arabidopsis (Arabidopsis thaliana) tanmei/emb2757 (tan) mutation that causes defects in both embryo and seedling development. tan mutant embryos share many characteristics with the leafy cotyledon (lec) class of mutants in that they accumulate anthocyanin, are intolerant of desiccation, form trichomes on cotyledons, and have reduced accumulation of storage proteins and lipids. Thus, TAN functions both in the early and late phases of embryo development. Moreover, the TAN and LEC genes interact synergistically, suggesting that they do not act in series in the same genetic pathway but, rather, that they have overlapping roles during embryogenesis. tan mutants die as embryos, but immature mutant seeds can be germinated in culture. However, tan mutant seedlings are defective in shoot and root development, their hypocotyls fail to elongate in the dark, and they die as seedlings. We isolated the TAN gene and showed that the predicted polypeptide has seven WD repeat motifs, suggesting that TAN forms complexes with other proteins. Together, these results suggest that TAN interacts with other proteins to control many aspects of embryo development.
Seed development in angiosperms begins with double fertilization of the egg cell and the central cell of the female gametophyte that leads to the formation of the diploid zygote and the triploid endosperm mother cell, respectively. The endosperm provides nutrients for the developing embryo and persists in the mature seed either as a layer of cells surrounding the embryo or as a storage tissue that serves a nutritional role for germinating seedlings (Lopes and Larkins, 1993; Olsen, 2004). Development of the zygote into the mature embryo can be divided conceptually into two overlapping phases (West and Harada, 1993; Goldberg et al., 1994). During the early morphogenesis phase, the basic body plan of the plant is established (Laux and Jurgens, 1997; Jurgens, 2001; Berleth and Chatfield, 2002). The initial cell divisions of the zygote generate the embryo proper and suspensor. Embryonic polarity is fixed with specification of the shoot-root axis. Along this axis, pattern elements, including the cotyledons, shoot apical meristem (SAM), hypocotyl, root, and root apical meristem, are established. Perpendicular to this axis, radial elements produce the embryonic tissue systems: protoderm, ground tissue, and procambium. During the late-maturation phase, the embryo undergoes unique developmental and physiological changes, including accumulation of nutrient reserves, acquisition of desiccation tolerance and dormancy, and developmental arrest and metabolic inactivation of the embryo as it desiccates (Bewley, 1997; Harada, 1997). Once the desiccated seed is exposed to an environment favorable for germination, the quiescent embryo becomes active metabolically and morphogenetically as postembryonic development is initiated. Although many aspects of embryogenesis have been characterized extensively, the molecular mechanisms underlying these events are poorly understood.
Genetic analyses have identified Arabidopsis (Arabidopsis thaliana) genes that play essential roles during embryogenesis. Several mutations that affect the early stages of the morphogenesis phase correspond to genes involved in morphogenetic processes that occur during vegetative development. For example, SHOOTMERISTEMLESS and WUSCHEL are involved in establishing the SAM, and SHORT ROOT and SCARECROW are involved in root apical meristem formation (Dilaurenzio et al., 1996; Long et al., 1996; Mayer et al., 1998; Helariutta et al., 2000). A different set of mutations has been shown to affect the maturation phase. abi3, abi4, and abi5 are mutations that cause defects in storage protein accumulation and/or tolerance to desiccation (Giraudat et al., 1992; Finkelstein et al., 1998; Finkelstein and Lynch, 2000). Furthermore, a large set of embryo defective (emb) mutations have been identified that disrupt embryo development and appear to be required for basic cellular functions (McElver et al., 2001; Tzafrir et al., 2004).
A distinct class of genes identified by the leafy cotyledon (lec) mutations of Arabidopsis play seed-specific roles in both the morphogenesis and maturation phases of embryogenesis (for review, see Harada, 2001). The LEC1, LEC2, and FUS3 genes are required early in embryogenesis to maintain embryonic cell fate and to specify cotyledon identity (Meinke, 1992; Keith et al., 1994; Meinke et al., 1994; West et al., 1994; Lotan et al., 1998). Late in embryogenesis, these genes are required to initiate and/or maintain the maturation phase and to inhibit premature germination of developing embryos (Baumlein et al., 1994; Keith et al., 1994; Meinke et al., 1994; West et al., 1994; Nambara et al., 2000; Raz et al., 2001). Moreover, ectopic expression of LEC1 or LEC2 induces embryonic characteristics and somatic embryogenesis in vegetative cells (Lotan et al., 1998; Stone et al., 2001). The LEC1 gene encodes a novel HAP3 subunit of the CCAAT binding factor, whereas LEC2 and FUS3 encode B3 domain transcription factors (Lotan et al., 1998; Stone et al., 2001; Kwong et al., 2003). We hypothesize that the LEC proteins are transcriptional regulators that integrate the early-morphogenesis and the late-maturation phase of embryogenesis by establishing cellular environments that promote embryo development.
In this article, we describe Arabidopsis mutants designated tanmei (tan) that have defects in embryogenesis. Analysis of tan mutants indicates that the gene plays roles in both the morphogenesis and maturation phases of embryogenesis, a hypothesis supported by genetic data suggesting that TAN has overlapping functions with the LEC genes. There are also indications that TAN is required for other aspects of plant development. We isolated the TAN gene and showed that it encodes a WD repeat protein predicted to function in protein-protein interaction. The potential role of TAN in plant development is discussed.
RESULTS
tan Mutation Causes Defects in Embryogenesis
We screened a T3 population of Arabidopsis mutagenized with T-DNA and identified embryo-defective mutants (Yadegari et al., 1994). As shown in Figure 1A, one line segregated immature seeds that were pale yellow as compared with their siblings that resembled green, wild-type seeds. After drying, the mutant seeds were shriveled and had a purple coloration as compared to wild-type seeds (Fig. 1, B and C), and they did not germinate, suggesting that the mutation caused embryo lethality. Furthermore, the mutation displayed monogenic, recessive inheritance. Twenty-five percent of progeny from plants heterozygous for the mutation had a mutant phenotype [68/286; χ2(3:1) = 0.2; P > 0.5]. Given that the mutant displayed embryonic lethality, we designated the corresponding gene TANMEI (TAN) because it is short lived.
Embryonic defects of tan mutants. A, Developing seeds within a silique from a heterozygous tan mutant plant. Mutant seeds have a pale yellow color (arrows). B and C, Dried mutant and wild-type seeds, respectively. Mutant seeds are desiccation intolerant. D and E, Mature mutant and wild-type embryos, respectively. Mutant embryo accumulates anthocyanin and displays aberrant cell growth (arrowheads). F and G, Mutant embryos were cleared and photographed with Nomarski optics. Trichome initials and/or aberrant cell growth are observed at the protoderm of the adaxial cotyledon surface (arrow head). Bars = 50 μm.
Growth Defects of tan Mutant Embryos
The tan-1 mutation caused defects in embryo morphology. As shown in Figure 1, D to F, mutant embryos differed from wild type in that their cotyledons did not curl within the seed at maturity because mutant embryos were shorter than wild type. Their embryonic axes were of comparable size to those of wild type, indicating that cotyledon growth was inhibited in the mutant. Mutant cotyledons also had regions of ectopic cell growth and patches of anthocyanin accumulation predominantly at their tips (Fig. 1D). We compared the SAMs of wild-type and tan-1 mutant embryos. As shown in Figure 2, the SAM of tan-1 mutant embryos (Fig. 2, A and B) did not appear to have the cell layer organization of wild-type sibling embryos at the bent-cotyledon stage (Fig. 2D) or of wild-type embryos at a similar morphological stage (Fig. 2C). In addition, root apices of tan-1 mutant embryos were translucent, while those of the wild-type embryos were opaque (Fig. 1, D and E). Thus, the morphology of tan-1 mutant embryos differed significantly from wild type.
SAM organization in tan mutant embryos. A and B, Medial sections through the SAM of tan-1 mutant embryos. tan mutant embryos were isolated from siliques at a stage when sibling embryos with a wild-type phenotype were at the bent-cotyledon stage. C, Wild-type embryo at the linear cotyledon stage. D, Wild-type embryo at the bent-cotyledon stage. a, Axis; c, cotyledon. Bars = 20 μm.
tan Mutation Affects the Maturation Program
We asked whether tan-1 mutants died as embryos because they were not able to withstand desiccation that occurs at the end of seed development. tan-1 mutant embryo cells from dried seeds cultured on callus-inducing medium did not grow or divide, whereas wild-type embryo cells proliferated to produce calli (data not shown). This result, along with the shriveled appearance of dried seeds, suggested that the cells in mutant embryos were dead because of their intolerance to desiccation. Because wild-type embryos normally acquire tolerance to desiccation during the maturation phase of embryogenesis (Harada, 1997), this result suggested that the tan mutation causes defects in maturation processes.
To determine whether the tan mutation affected other aspects of the maturation phase, we examined embryos at the ultrastructural level. As shown in Figure 3A, wild-type embryo cells in cotyledons and embryonic axes were packed with storage protein and lipid bodies. By contrast, accumulation of these organelles was defective in tan mutant embryos, particularly in apical regions. Storage protein and lipid bodies were much less prevalent in cotyledons (Fig. 3, B and C) versus axes (Fig. 3D) of tan-1 mutants, although limited numbers of storage organelles were observed throughout the embryo. We conclude that the TAN gene is required for aspects of the maturation phase of embryogenesis.
tan embryo cells contain fewer lipid and protein bodies. A, Cells from a wild-type embryo. B to D, Mutant embryonic cells in the tip and the middle of cotyledons and the embryonic axis, respectively. LB, Lipid body; PB, protein body; P, plastid; V, vacuole; N, nucleus. Bar = 1 μm; bar in A applies to B to D.
tan Mutant Embryos Have Dormancy Defects
To obtain further insight into the role of the TAN gene, we asked whether mutant embryos could be rescued by germinating immature seeds in culture. Seventy to 90% of immature tan-1 mutant seeds at 10 to 14 d after flowering germinated in the light without stratification and produced seedlings, whereas only approximately 10% of wild-type immature seeds germinated. Thus, the tan mutation appears to affect dormancy, although mutant seeds were not viviparous in planta. We note that immature tan-1 mutant seeds did not germinate on medium containing 1 μm abscisic acid, suggesting that defects in dormancy did not result from abscisic acid insensitivity.
tan Mutants Display Defects in Seedling Morphology
Although immature tan-1 mutant embryos could be germinated before desiccation, mutant seedlings did not continue to develop and eventually died. Thus, tan-1 mutants exhibit both embryo lethality and seedling lethality. Analyses of mutant seedlings revealed striking defects in seedling morphology caused by the tan mutation. As shown in Figure 4A, tan-1 mutant seedlings had a few trichomes either without branches or with one branch on the adaxial surfaces of their cotyledons that were initiated during embryogenesis (Fig. 1G). Trichomes are normally present only on leaves, stems, and sepals and are rarely observed on cotyledons of wild-type Arabidopsis seedlings (Fig. 4C; Hulskamp et al., 1994; West et al., 1994). Furthermore, approximately 50% of tan-1 seedlings had cotyledons that were fused along one side of their margins (Fig. 4B), another defect that was likely to be initiated during embryogenesis (Aida et al., 1997). It was also apparent that the tan mutation affected other aspects of seedling development. For example, mutant seedlings had stunted roots and hypocotyls that failed to elongate (Fig. 4, A and B). Moreover, true leaves rarely emerged, suggesting that the SAM is defective. To determine whether tan-1 mutants were competent to form vegetative SAMs, we cultured immature tan-1 embryos on callus-inducing medium. Unlike wild-type embryos that formed calli and adventitious shoots when cultured (Fig. 4H), tan-1 mutant cells underwent limited proliferation and eventually stopped growing without producing shoots (Fig. 4G). We conclude that the tan mutation has pleiotropic effects on seedling development.
tan mutants are defective in vegetative development. A to F, Mutant and wild-type seedlings germinated from immature seeds on medium. A, Mutant seedlings grown in the light have trichomes on cotyledons (arrow heads). B, Many mutant seedlings have cotyledons fused along one margin. C, Wild-type seedling grown in the light. D and E, Mutant seedlings grown in the dark show etiolation defects. F, Wild-type seedling grown in the dark. G and H, Mutant and wild-type calli on shoot-inducing medium, respectively. Unlike wild-type callus, tan calli did not regenerate adventitious shoots. Bars = 500 μm.
tan mutants share a subset of characteristics exhibited by the photomorphogenic fusca/constitutive photomorphogenic/de-etiolated mutants. This class of mutants accumulates anthocyanin in embryos, exhibits seedling lethality, and fails to etiolate when grown in the dark (Castle and Meinke, 1994; Chory et al., 1996; Wei and Deng, 1996). To determine whether tan-1 mutants are defective in photomorphogenesis, we germinated immature tan-1 seeds in the dark. As shown in Figure 4, D and E, dark-grown tan seedlings had short hypocotyls, and one-half the seedlings partially expanded their cotyledons. By contrast, wild-type seedlings had fully elongated hypocotyls and closed cotyledons (Fig. 4F). However, dark-grown tan-1 seedlings did not exhibit other characteristics of photomorphogenic mutants such as development of leaves and expression of light-inducible genes encoding chlorophyll a/b-binding protein and D1 protein of PSII (data not shown), which are up-regulated in dark-grown photomorphogenic mutants. Consistent with these results, we used genetic complementation experiments to show that TAN is not allelic with the FUS/COP/DET genes, COP1/FUS1, COP8/FUS8, COP9/FUS7, COP10/FUS9, COP11/FUS6, COP15/FUS5, and DET1/FUS2. We conclude that TAN is not a typical FUS/COP/DET gene.
Interaction between TAN and LEC Genes
tan-1 mutant embryos share many, but not all, characteristics of embryos with mutations in the LEC genes, LEC1, LEC2, and FUS3. For example, both tan-1 and lec mutations cause desiccation intolerance, defects in storage protein and lipid accumulation, and trichome formation on cotyledons (Harada, 2001). However, the lec mutations do not cause seedling lethality. Genetic tests showed that TAN was not allelic with the LEC genes (data not shown). Therefore, we conducted digenic mutant analysis to investigate the relationship between TAN and the LEC genes.
TAN did not display epistasis with the LEC genes because, as shown in Figure 5, embryos with digenic mutations in TAN and any of the three LEC genes had characteristics that were not predicted by the single mutant phenotypes. For example, the vast majority of tan-1 lec1-2 digenic mutant embryos arrested at the transition or heart stage, as shown in Figure 5, A and B, although a smaller number attained a shape similar to a torpedo-stage embryo (Fig. 5C). By contrast, each single mutant arrested at a late embryonic stage (Figs. 1D and 5D). Thus, TAN and LEC1 did not display epistasis and did not interact additively in that the digenic mutant arrested at an earlier morphological stage than either monogenic mutant. tan-1 lec1-2 mutant seeds were desiccation intolerant and germinated poorly even when cultured on agar plates. The morphology of tan-1 lec2-1 digenic mutants differed from lec2-1 mutant embryos (Fig. 5H) and resembled those with tan-1 lec1-1 mutations in that most arrested primarily at the transition to heart stage (Fig. 5, E and F) and a few arrested at the torpedo stage (Fig. 5G). By contrast, we found that tan-1 fus3-3 digenic mutants exhibited variable phenotypes. Approximately one-half of digenic mutants resembled transition- and heart-stage embryos (Fig. 5, I and J), whereas the remainder appeared to be intermediate forms of both parents (Figs. 1D and 5, K and L). Seeds with intermediate-form embryos germinated in culture and produced seedlings with trichomes on cotyledons, but they displayed lethality like tan-1. Together, these results indicate that the TAN and LEC genes, particularly LEC1 and LEC2, interact synergistically. Thus, TAN and LEC genes may operate in closely related pathways during embryogenesis.
Lack of epistasis between TAN and LEAFY COTYLEDON genes. A to C, Terminal phenotypes of tan lec1 digenic mutant embryos. Digenic mutant embryos are shaped like globular stage, heart-stage, and torpedo-stage embryos, respectively. D, Terminal phenotype of lec1 monogenic mutant embryo. E to G, Terminal phenotype of tan lec2 digenic mutant embryos. H, Terminal phenotype of lec2 monogenic mutant embryo. I to K, Terminal phenotype of tan fus3 digenic mutant embryos. L, Terminal phenotype of fus3 embryo. Bars = 100 μm.
Identification of the TAN Gene
To determine its molecular function, we isolated the TAN gene. Genetic segregation and genomic DNA gel-blot analyses indicated that the tan-1 mutation is linked with a single T-DNA locus (data not shown). Genomic clones containing the T-DNA were isolated from a library of tan-1 DNA (see “Materials and Methods”). Plant DNA fragments flanking the T-DNA were used to isolate seven cDNA clones from a library of silique RNA sequences and the corresponding genomic clones containing wild-type genomic DNA. The cDNA clones represented two RNAs, 1.5 and 0.8 kb in size, that were encoded by different genes and present in wild-type siliques and seedlings. We used reverse transcription (RT)-PCR analysis to determine which cDNA clone corresponded to the TAN gene. As shown in Figure 6B, the 0.8-kb RNA was present in wild-type and tan-1 mutant seedlings, whereas the 1.5-kb RNA was detected only in wild-type seedlings, suggesting that TAN encodes the 1.5-kb RNA. Consistent with this conclusion, a single recessive mutation [72/290, χ2(3:1) = 0.004; P > 0.9], tan-2 (SALK_097510), was identified that is allelic to and that caused an identical mutant phenotype as tan-1. In tan-2, the left border of a T-DNA was present in the putative TAN gene, but we were unable to position the right border. Thus, the precise nature of the T-DNA insertion in tan-2 is not known.
Identification of the TAN gene. A, Diagrammatic representation of the TAN locus. Green boxes represent exons of the TAN gene encoding a 1.5-kb RNA, gray boxes represent introns or intergenic regions, blue boxes represent the adjacent gene encoding an 0.8-kb RNA. Location of the T-DNA in tan-1 is shown. B, RT-PCR analysis showed that the 0.8-kb RNA is present in both wild-type and tan mutant seedlings, whereas the 1.5-kb RNA is detected in wild-type, but not tan, mutants.
Because the longest TAN cDNA clone contained only a 759-bp insert, we performed RT-PCR to obtain 5′ RNA sequences and obtained two classes of cDNA sequences. One class, represented by the RACE14 clone, contained the remaining 625 bp of the cDNA sequence and produced a long open reading frame (ORF) in the full-length RNA. Sequence alignment of the full-length RNA and the gene revealed that the TAN gene consists of 13 exons (Fig. 6A). This analysis also showed that the T-DNA in tan-1 disrupts the twelfth intron and that a deletion of 5 bp at the intron-exon junction occurred 7 bp downstream of the insertion (data not shown). To determine whether this cDNA clone represents the TAN gene, we constructed a full-length TAN cDNA clone, fused it with approximately 3 kb of 5′ flanking sequence from the TAN gene, and transferred the chimeric gene into plants heterozygous for the tan-1 mutation. The transgene suppression experiments provided strong evidence that the RACE14 clone represents a TAN cDNA clone. Embryos with mutant and wild-type phenotypes segregated at ratios suggesting that the transgene suppressed the tan-1 mutation and was present at a single locus in one transgenic plant [20/258, χ2(15:1) = 0.99; P > 0.3] and at two loci in two transgenic plants, [2/261, χ2(63:1) = 1.08; P > 0.3] and [3/259, χ2(63:1) = 0.28; P > 0.5]. The segregation of progeny from one transgenic plant was distorted, suggesting transgene suppression of the mutation, but the ratio was not consistent with either one or two transgene loci [e.g. 8/254, χ2(15:1) = 4.17; P > 0.01]. We conclude that the cDNA clone represents the 1.5-kb RNA encoded by the TAN gene.
Other data suggest that alternative splicing of the TAN gene occurred, producing nonfunctional transcripts. We obtained another class of cDNA clone that had 5 bp missing at the beginning of the second exon of TAN that changed the reading frame and is expected to produce a truncated protein. We also found that two expressed sequence tags corresponding to TAN cDNA clones from the Columbia ecotype resulted from different splicing events as compared with the RACE14 version of the RNA (accession nos. AA713092 and AK118989). Both alterations caused premature termination of the ORF, suggesting that RNAs corresponding to the expressed sequence tag clones are not functional. Subsequent sequencing of a bacterial artificial chromosome clone, F27B13 (accession no. AL050352), as part of the Arabidopsis Genome Initiative showed that the TAN gene is located on the bottom arm of chromosome IV (Arabidopsis Genome Initiative, 2000). Based on the locus designation, At4g29860, we learned that TAN corresponds to two recently reported emb mutations, emb2757-1 and emb2757-2, described as part of a dataset of 250 EMB genes identified from T-DNA insertion lines (Tzafrir et al., 2004). Morphological mutant phenotypes attributed to these lines, such as seed and embryo color and embryo morphology (http://www.seedgenes.org), are consistent with those described for tan-1 and tan-2. Thus, TAN appears to be allelic with EMB2757.
The TAN cDNA (accession no. AB191306) contains an ORF encoding a protein of 386 residues. As shown in Figure 7, the deduced amino acid sequence of the TAN gene contains seven WD repeats (Neer et al., 1994). WD repeats comprise 44 to 60 amino acid residues that often contain specific subsequence motifs, including the Trp-Asp (WD) dipeptide for which the repeat was named (Smith et al., 1999). The seven predicted WD repeats in TAN conform well with this motif with the possible exception of the third repeat.
Amino acid sequence similarity between TAN and other WD repeat proteins. Comparison of the predicted amino acid sequence of TAN with WD protein homologs from mouse, human, Anopheles, Drosophila, and Schizosaccharomyces pombe. Identical amino acid residues are shown by black boxes. Overlined sequences show putative WD repeats. Numbers indicate amino acid positions in the TAN protein.
RNA gel-blot (data not shown) and quantitative RT-PCR analyses, summarized in Figure 8A, showed that TAN RNA is present in floral buds, leaves, stems, roots, and four sequential stages of siliques. Quantitative RT-PCR experiments showed that TAN RNA was present at its highest levels in siliques that contain developing seeds. In situ hybridization analysis showed that TAN RNA accumulates throughout the embryo and in endosperm (Fig. 8B). We conclude that the TAN gene encodes a WD repeat protein that functions during embryogenesis and other phases of the life cycle.
Detection of TAN RNA in organs and developing embryos. A, Detection of TAN RNA levels with quantitative RT-PCR analysis. TAN RNA was detected in total RNA from flower buds (Fl), rosette leaves (Le), stems (St), roots (Ro), siliques containing seeds with embryos at the zygote to early globular stage (S1), globular stage to heart stage (S2), torpedo stage to curled-cotyledon stage (S3), and mature green stage (S4). TAN RNA levels were normalized to that of the 60S ribosomal protein, L14. Bars indicate the average of two determinations. B to D, Detection of TAN RNA in developing embryos with in situ hybridization. Wild-type seed sections were hybridized with a TAN antisense probe. Control experiments with a sense probe indicated that the hybridization signals were specific to TAN RNA. B, Heart-stage embryo. C, Bent-cotyledon-stage embryo. D, Mature green-stage embryo. Bars = 100 μm.
DISCUSSION
TAN Is a WD Repeat Protein
We identified the TAN/EMB2757 gene whose deduced amino acid sequence (Fig. 7) shows that it is one of the approximately 237 potential WD repeat proteins in Arabidopsis (van Nocker and Ludwig, 2003). This class of protein is involved in mediating interactions between proteins. WD repeat proteins have been predicted to have between four and 16 copies of the repeat (Smith et al., 1999), although structural analyses of WD repeat proteins have shown that several proteins predicted to have fewer or more repeats actually possess seven (Voegtli et al., 2003; Madrona and Wilson, 2004). Similarly, TAN is predicted to possess seven WD repeats, suggesting that it is a functional WD repeat protein.
Identification of TAN as a likely WD repeat protein provides insight into its molecular function. WD repeat proteins participate in protein-protein interactions and have roles in a variety of physiological functions (Neer et al., 1994; Smith et al., 1999). Of proteins with a known function, TAN is similar to the β-subunit of G proteins (Jones, 2002). However, this similarity is based primarily on sequences within the WD repeat. Portions of sequential WD repeats interact to form propeller blades, and the propeller structure creates a surface with which other proteins can interact (Neer et al., 1994; Smith et al., 1999). Thus, regions between repeats determine the specificity of each WD repeat protein, and WD repeat proteins with common functions share sequence similarity in inter-repeat regions. No significant sequence similarity was observed between the inter-repeat regions of TAN with other proteins in the database (http://bmerc-www.bu.edu/wdrepeat).
TAN Functions Early and Late in Embryo Development
Phenotypic analyses suggest that the TAN gene plays roles during both the early-morphogenesis and late-maturation phases of embryogenesis. As summarized in Table I, several lines of evidence indicate that TAN is required for normal progression through the maturation phase. The acquisition of desiccation tolerance and the accumulation of storage reserves are hallmarks of this developmental phase. The tan mutation causes defects in both processes because mutant embryos are intolerant of desiccation and do not properly accumulate storage proteins and lipids (Fig. 3). Consistent with this observation, tan mutant embryos contain reduced levels of RNA encoding the small heat shock protein, HSP 17.4, a protein whose accumulation is correlated with an embryo's competence for desiccation tolerance (data not shown; Wehmeyer and Vierling, 2000). Another indication that TAN functions during the maturation phase is that immature tan mutant embryos germinate efficiently, indicating that dormancy is compromised by the tan mutation (Fig. 4). Together, these results suggest strongly that TAN is required for normal progression through the maturation phase.
Summary of defects in tan mutants
Table I also summarizes results suggesting a role for TAN in the morphogenesis phase of embryogenesis. In support of this idea, tan mutant embryos do not appear to form functional embryonic SAMs, as indicated by defects in SAM organization in mutant embryos (Fig. 2), fusion of mutant cotyledons along one margin (Fig. 4), and the inability of seedlings to form stems or leaves (Fig. 4). Root apical meristem also appears to be defective in tan mutant embryos and seedlings (Figs. 1 and 4).
Although tan mutant seedlings have an abnormal morphology (Fig. 4), it is not clear whether the defects result from direct effects of the mutation on seedling development or indirect effects of the mutation that occurred originally during embryogenesis. For example, because SAM initiation occurs early in embryogenesis, abnormalities in the seedling SAM may be attributed to embryonic defects (Barton and Poethig, 1993). Furthermore, similar to tan, dark-grown lec1 seedlings also have shorter hypocotyls as compared with dark-grown wild-type seedlings, yet LEC1 gene expression is restricted to embryogenesis (Lotan et al., 1998; Brocard-Gifford et al., 2003; K. Yamagishi, J. Danao, and J. J. Harada, unpublished data). Thus, effects on seedling morphology could be explained by indirect effects of the tan mutation on embryo development. On the other hand, tan mutant calli failed to form shoots (Fig. 4), suggesting that the vegetative SAM is defective. Moreover, the TAN gene appears to be expressed throughout plant development (Fig. 8), suggesting a potential postembryonic role for the gene.
Relationship between TAN and the LEC Genes
Mutations in the lec genes induce many defects in embryo development (Harada, 2001) shared with those caused by the tan mutation, suggesting that the genes may have overlapping roles in embryogenesis. For example, a defining characteristic of lec mutants is the presence of trichomes on the adaxial surfaces of their cotyledons. Trichomes are present on true leaves, sepals, and stems and are not normally found on Arabidopsis cotyledons (Hulskamp et al., 1994; West et al., 1994). The presence of trichomes on the cotyledons of lec mutants has been attributed to heterochrony (West et al., 1994). That is, the shift from the maturation to the postgerminative phase of development induced by the lec mutations causes a heteroblastic change in which cotyledons adopt a partial leaf fate. Thus, one explanation for trichome initiation on the surfaces of tan mutant cotyledons (Figs. 1 and 4) is that defects in the maturation phase induce postgerminative development prematurely. However, we cannot eliminate the possibility that other factors are responsible for trichome formation on tan cotyledons. For example, twin1 mutants with embryonic transformation of the suspensor (Vernon and Meinke, 1994) and transgenic plants overexpressing GLABROUS1 and AtMYB23, genes involved in trichome initiation, develop cotyledonary trichomes after germination (Larkins et al., 1994; Kirik et al., 2001). Furthermore, trichomes have also been detected on cotyledons of constitutive photomorphogenesis and dwarfism (cpd) mutants that have defects in brassinosteroid biosynthesis (Szekeres et al., 1996). In this context, we confirmed that tan mutant seedlings contained wild-type levels of major sterols and their defects were not rescued by the application of 24-epibrassinolide (K.Yamagishi and S.Fujioka, unpublished data).
Another indication that tan mutants prematurely enter the postgermination phase is the high germination frequency of cultured immature seeds, indicating competence to germinate. We have shown that isocitrate lyase and lipid transfer protein genes, normally expressed primarily during postgerminative growth, are activated prematurely in tan mutant embryos (data not shown). Thus, similar to the lec mutations, the tan mutation appears to induce at least some aspects of postgerminative development prematurely during seed maturation.
tan mutants are similar to lec2 and fus3 mutants in that they all display a gradient of defects in storage reserve accumulation, with defects in storage protein accumulation being severe in apical regions of cotyledons and mild or nonexistent in embryonic axes (Fig. 3; Meinke et al., 1994). However, tan and lec2 mutants differ in that lec2 mutant embryos exhibit a gradient of desiccation intolerance in apical regions (Meinke et al., 1994; Stone et al., 2001), whereas tan mutants do not. Rather, all cells of tan mutant embryos appear to be intolerant of desiccation because no cells in mature dried embryos are able to proliferate and form callus. One interpretation of these results is that storage reserve accumulation and desiccation tolerance are regulated independently in pathways downstream of TAN. Alternatively, TAN may directly regulate each pathway, but desiccation tolerance may be more sensitive to loss of TAN activity.
Genetic interactions between TAN and LEC genes provide compelling evidence for an overlap in their functions. tan lec1, tan lec2, and a subset of tan fus3 digenic mutants display a synergistic phenotype similar to those observed for lec1 lec2 digenic mutants (Fig. 5; Meinke et al., 1994; Lotan et al., 1998). LEC1 and LEC2 both act early in embryogenesis (Harada, 2001). The lec1 lec2 digenic mutant arrests at an earlier morphological stage of embryogenesis than either monogenic mutant, and this result has been interpreted to indicate that the two genes have overlapping functions early in embryogenesis. By analogy, the striking similarities of tan lec1, tan lec2, and lec1 lec2 digenic mutants suggest that TAN also functions early in embryogenesis and that it may have overlapping function with the LEC genes, particularly LEC1 and LEC2. We note that both the lec1-2 and lec2-1 alleles used for the digenic mutant analysis are null alleles (Lotan et al., 1998; Stone et al., 2001), whereas it is not known whether fus3-3 represents a null allele (Luerben et al., 1998).
In conclusion, we have identified the TAN/EMB2757 gene that is required for several aspects of plant development. TAN is required both early and late in embryo development and potentially during vegetative development. Moreover, TAN interacts genetically with LEC genes, suggesting an overlap in functions. The demonstration that TAN is expressed throughout plant development and encodes a putative WD repeat protein suggests that it accomplishes its function by interacting with other proteins possibly to form regulatory complexes.
MATERIALS AND METHODS
Plant Material
The tan-1 mutant (previously designated Line 24) was derived from a population of Arabidopsis (Arabidopsis thaliana ecotype Wassilewskija [Ws]) mutagenized with T-DNA insertions (Feldmann and Marks, 1987; Yadegari et al., 1994). tan-2 (stock no. SALK_097510, ecotype Columbia), and cop, det, and fus mutants were obtained from the Arabidopsis Biological Resource Center (Columbus, OH). The tan-1 mutant was used for experiments unless otherwise indicated. fus3-3 mutants and the lec2-1 mutant were provided by Peter McCourt (University of Toronto; Nambara et al., 1992; Keith et al., 1994) and by David Meinke (Oklahoma State University; Meinke et al., 1994), respectively. Plant growth, digenic mutant construction, and lec1-2 and lec2-1 mutants are as described previously (West et al., 1994; Lotan et al., 1998). Because lec mutants can be maintained as homozygotes, tan-1 lec digenic mutant embryos segregated at a 3:1 ratio. The genotype of each digenic mutant was verified by backcrosses with each single mutant.
Induction of Calli and Shoots
Embryos were dissected from imbibed or immature seeds and cultured for six weeks on callus-inducing medium and, subsequently, for five weeks on shoot-inducing medium according to Valvekens et al. (1988), except that the carbon source was 20 g/L Suc.
Isolation and Sequence Analysis of Genomic and cDNA Clones
A genomic DNA library from a heterozygous tan-1 mutant was constructed in the λGEM 11 vector (Promega). A wild-type Ws-0 genomic library was provided by Ken Feldmann (University of Arizona). A cDNA library was prepared according to the manufacturer's specifications in the λZAPII bacteriophage vector (Stratagene) from wild-type siliques containing heart- to young-torpedo-stage embryos.
Clones were isolated from the tan genomic library using probes for the right and left T-DNA borders. HindIII fragments of 4.1 and 5.1 kb containing the plant DNA/T-DNA junctions were isolated and cloned into the pBluescript KS(+) phagemid (Stratagene) to create pRL3 and pL15, respectively. The DNA fragments were used to isolate genomic clones from a wild-type Ws-0 library. A 3.2-kb HindIII fragment present in a genomic clone that hybridized with both probes was inserted into the pBluescript KS(+) phagemid and used to identify TAN cDNA clones from wild-type silique cDNA libraries. Because the cDNA clones were smaller than the corresponding RNA, the 5′-end of the cDNA was obtained using the CapFinder PCR cDNA library construction kit (CLONTECH). cDNA was synthesized from poly(A)+ RNA of 3-week-old plants according to the manufacturer's instructions and subjected to the nested PCR with a 5′ PCR primer (CLONTECH) and gene-specific primers. The amplified products were inserted into pCR2.1 plasmid (Invitrogen) and fused with the cDNA clone at an MfeI site to make full-length cDNAs. Database searches were performed at the National Center for Biotechnology Information by using the BLAST network service. Alignment of protein sequences was done using DNASIS (Hitachi Software Engineering Co.). Analysis of WD repeat motifs was performed at the BioMolecular Engineering Research Center Web site (http://bmerc-www.bu.edu/index.html).
Generation of Transgenic Plants
For genetic suppression of the tan mutation, the full-length TAN cDNA was fused with a 3-kb fragment of TAN promoter and the octopine synthase terminator in the plasmid, pART7, in the proper transcriptional orientation (Gleave, 1992). The fusion gene was transferred into the plant transformation vector, BJ49. Constructs were transferred into heterozygous tan-1 mutant plants using in planta transformation procedures with Agrobacterium tumefaciens strain GV3101 (Bechtold et al., 1993). Genotypes of transgenic plants were verified in DNA amplification experiments.
RNA Analysis
RNA was isolated as described previously (Stone et al., 2001). For quantitative RT-PCR, RNAs were treated with DNase, MseI, and DdeI to control for DNA contamination. RT reactions were carried out with 1 μg of total RNA in a 20-μL reaction volume. PCR reactions were done with the Bio-Rad iQ SYBR Green supermix (Bio-Rad Laboratories) according to specifications, except that 0.25 μm of each primer and 1 μL of a 20-fold dilution of RT products were used. The Bio-Rad iCycler iQ multicolor real-time detection system (Bio-Rad Laboratories) was used with the following reaction conditions: one repeat of 3 min at 95°C followed by 50 repeats of 30 s at 95°C, 30 s at 60°C, 30 s at 72°C, followed by a melt curve encompassing 80 steps of 0.5°C from 60°C to 100°C. Images were captured during the 72°C extension step and at each step in the melt curve using a FAM filter. The following primer sets were used: AT4G27090 Fwd 5′-GAGAGGATCCAGATGAACTTCAAGA-3′ and AT4G27090 Rev 5′-GCTCTTGCCTGACAACACCAGCTTT-3′; TanMei Fwd 5′-GACCTCCGCCTGATCCAGTG-3′ and TanMei Rev 5′-CCATTTGCTCTACTATGAGCC-3′. Primary data analysis was performed with Bio-Rad iCycler iQ software. The Bio-Rad gene expression macro version 1.1 software (Bio-Rad Laboratories; Livak and Schmittgen, 2001; Pfaffl, 2001; Vandesompele et al., 2002) was used to calculate relative RNA levels normalized to an internal control (AT4G27090: 60S ribosomal protein L14; Thellin et al., 1999). Results are reported relative to RNA level in the S1 sample.
In situ hybridization experiments were performed as described previously (Dietrich et al., 1989).
Electron Microscopy
Transmission electron microscopy was performed according to Meinke et al. (1994).
Sequence data from this article can be found in the EMBL/GenBank data libraries under accession number AB191306.
Acknowledgments
We thank Jay Danao and Minsung Kim for their help with experiments, Bart Janssen for vectors, the Arabidopsis Biological Stock Center for seeds, K.C. McFarland of the Plant Cell Biology Training Facility for help in generating figures, and Sandra Stone for comments about the manuscript.
Footnotes
-
Article, publication date, and citation information can be found at www.plantphysiol.org/cgi/doi/10.1104/pp.105.060467.
-
↵1 This work was supported in part by research fellowships from the Japan Society for the Promotion of Science and the Special Postdoctoral Researchers Program of RIKEN (to K.Y.) and by grants from the Department of Energy and the National Science Foundation (to J.J.H.).
-
↵2 Present address: Cell Dynamics Research Group, Research Institute for Cell Engineering, National Institute of Advanced Industrial Science and Technology (AIST), 1–8–31 Midorigaoka, Ikeda, Osaka 563–8577, Japan.
-
↵3 Present address: Department of Chemical Biological Sciences, Faculty of Science, Japan Women's University, 2–8–1, Mejirodai, Bunkyo-ku, Tokyo 112–8681, Japan.
-
↵4 Present address: Reeve-Irvine Research Center, University of California, Irvine, CA 92697.
- Received February 9, 2005.
- Revised May 17, 2005.
- Accepted May 31, 2005.
- Published August 19, 2005.