Search

We have previously described a homeotic leafy cotyledon (lec) mutant of Arabidopsis that exhibits striking defects in embryonic maturation and produces viviparous embryos with cotyledons that are partially transformed into leaves. In this study, we present further details on the developmental anatomy of mutant embryos, characterize their response to abscisic acid (ABA) in culture, describe other mutants with related phenotypes, and summarize studies with double mutants. Our results indicate that immature embryos precociously enter a germination pathway after the torpedo stage of development and then acquire characteristics normally restricted to vegetative parts of the plant. In contrast to other viviparous mutants of maize (vp1) and Arabidopsis (abi3) that produce ABA-insensitive embryos, immature lec embryos are sensitive to ABA in culture. ABA is therefore necessary but not sufficient for embryonic maturation in Arabidopsis. Three other mutants that produce trichomes on cotyledons following precocious germination in culture are described. One mutant is allelic to lec1, another is a fusca mutant (fus3), and the third defines a new locus (lec2). Mutant embryos differ in morphology, desiccation tolerance, pattern of anthocyanin accumulation, presence of storage materials, size and frequency of trichomes on cotyledons, and timing of precocious germination in culture. The leafy cotyledon phenotype has therefore allowed the identification of an important network of regulatory genes with overlapping functions during embryonic maturation in Arabidopsis.

Embryogenesis plays a central role in the life cycle of flowering plants. Ever since the process of double fertilization was documented at the turn of the century, plant embryologists have attempted to characterize the cellular and biochemical changes that occur within developing seeds (Maheshwari, 1950; Raghavan, 1976). For many years, the emphasis was on comparative morphology and the analysis of cell division patterns during early stages of embryo development (Johansen, 1950; Wardlaw, 1955). Attention then shifted to experimental studies of somatic embryogenesis, embryo culture, and haploid embryos derived from microspores (Raghavan, 1986). Manipulation of zygotic embryos was limited by the presence of surrounding maternal tissues. Electron microscopy provided additional details on cellular changes associated with embryogenesis but generally failed to identify the mechanisms responsible. Plant embryologists believed that genetic factors played an important role in morphogenesis, but mutants with altered patterns of embryo development were largely ignored (Meinke, 1986). Recent advances in molecular biology have led to a renewed interest in plant embryology and the underlying patterns of gene expression that occur throughout seed development (Goldberg et al., 1989). Many genes transcribed during embryonic maturation have now been examined at the molecular level. Genetics provides a complementary approach to the study of plant embryo development by allowing the identification of genes with essential functions during this critical stage of the life cycle (Meinke, 1986). The purpose of this review will be to explore the potential benefits and limitations of mutant analysis in relation to plant embryo development. Additional information can be obtained through recent reviews on plant development (Steeves and Sussex, 1989; Lyndon, 1990), experimental plant embryogenesis (Johri, 1984; Raghavan, 1986; Williams and Maheswaran, 1986), and plant developmental genetics (Sheridan, 1988; Meinke, 1991a, 1991b). Plant embryologists originally attempted to explain characteristic patterns of cell division observed early in development by establishing fundamental laws of embryogeny (Johansen, 1950). These laws of parsimony, numbers, origin, disposition, and destination were thought to reflect the fundamental organization of embryos from different species. This view has gradually been replaced by the realization that cell division patterns are determined more by genetic and biophysical factors than by laws of embryogeny. Examination of developmental pathways in different angiosperms has nevertheless revealed a number of common features. The most critical events appear to be the formation of apical meristems, the establishment of basic patterns of symmetry and cellular organization, and the transition from a heterotrophic zygote dependent upon nutrient reserves of surrounding maternal tissues to an autotrophic embryo capable of surviving desiccation and producing a viable seedling after germination. Early stages of plant embryo development are characterized by cell division and morphogenesis. This is followed by a period of cell specialization and embryonic maturation in preparation for dormancy and germination. Several features distinguish embryogenesis in angiosperms from related pathways in animal systems: (1) the process of double fertilization and subsequent interactions between embryo and endosperm tissues, (2) the apparently minor role played by maternal mRNAs stored in unfertilized eggs, (3) the absence of cellular migration during embryo development, (4) the absence of a germ line established early in development, (5) the activation of large numbers of zygotic genes during very early stages of embryogenesis, (6) the presence of a male gametophyte active in transcription that leads to elimination of many deleterious alleles before fertilization, (7) the establishment of relatively few specialized cell types, (8) the formation of apical meristems that ultimately produce the adult plant, and (9) the small size of the zygote and its location deep within maternal tissues. The zygote in flowering plants is also not unique; a wide range of somatic cells can be

Developmental arrest of the embryo proper in aborted seeds from mutant 50B, a recessive embryo-lethal mutant of Arabidopsis thaliana, was shown to be followed by abnormal growth of the suspensor. Each of the 12 aborted seeds examined in sectioned material contained an abnormally large suspensor and an embryo proper arrested at a preglobular stage of development. Analysis of serial sections revealed that mutant suspensors contained 15-150 cells whereas wildtype suspensors were composed of only six to eight cells. Development of the mutant endosperm continued to a late nuclear or early cellular stage even in the absence of further development of the embryo proper. These results suggest that the missing gene product in mutant 50B is required for development of the embryo proper but not for continued growth of the suspensor or endosperm tissue. The pattern of abnormal development observed in this mutant provides further evidence that continued growth of the suspensor during normal development is inhibited by the developing embryo proper and that the full developmental potential of cells in the suspensor is expressed only when this inhibitory effect is removed through a mutation or experimental treatment that is lethal only to cells of the embryo proper. THE ZYGOTE in flowering plants develops into an early embryo composed of a terminal embryo proper and an attached suspensor. The embryo proper differentiates into the mature embryo whereas the suspensor usually degenerates during later stages of development. Many different patterns of morphogenesis have been observed in the development of the suspensor (Johansen, 1950; Maheshwari, 1950; Wardlaw, 195 5). Most members of the Cruciferae have a filamentous suspensor composed of a single file of 5-10 cells. In other families, such as the Leguminosae, the suspensor may be reduced to a single cell or enlarged into a massive structure composed of more than 100 cells (Lersten, 1983). Individual cells of the suspensor may also become polyploid, polytene, or multinucleate, or may form haustoria that invade surrounding maternal tissues. The suspensor is therefore an integral part of the plant embryo, but its size and cellular organization are highly variable. The function of the suspensor was originally thought to be limited to the passive role of attaching the embryo proper to surrounding ' Received for publication 1 1 February 1985; revision accepted 14 June 1985. This research was supported in part by NSF Grant PCM76-17222 to I. M. Sussex and NSF Grant PCM8215667 to D. W. Meinke. Initial studies were performed while M. P. F. Marsden was a Visiting Fellow at Yale University in the laboratory of Dr. Ian Sussex. 2 To whom reprint requests and correspondence should be addressed. maternal tissues (Maheshwari, 1950). The results of more recent studies on the ultrastructure of the developing suspensor (Schulz and Jensen, 1969; Schnepf and Nagl, 1970; Newcomb and Fowke, 1974; Yeung and Clutter, 1979), the transport of nutrients to the developing embryo proper (Yeung,1980), the growth in vitro of immature embryos cultured with and without a suspensor (Corsi, 1972; Cionini et al., 1976; Yeung and Sussex, 1979), the extent of endoreduplication and RNA synthesis in the developing suspensor (Sussex et al., l 973; Clutteretal., 1974;Nagl, 1974),andtheanalysis of growth regulators isolated from the massive suspensors of Phaseolus coccineus (Alpi, Tognoni and D'Amato, 1975; Lorenzi et al., 1978; Ceccarelli, Lorenzi and Alpi, 1981) and Tropaeolum majus (Przybyllok and Nagl,1977) all suggest that the suspensor plays an important role in promoting and regulating development of the early embryo proper. Much less is known about how development of the suspensor is regulated, what causes the suspensor to stop growing early in development, and how the suspensor interacts with other parts of the developing seed. Developmental interactions between the embryo proper and suspensor have been particularly difficult to study because they occur early in development when the embryo is small and deeply embedded in maternal tissue. One approach that has provided clues to the regulation of suspensor development has been

Mature seeds of Glycine max (L.) Merr. contain two major storage proteins, a glycosylated 7S protein (conglycinin) and a non-glycosylated 11S protein (glycinin). Accumulation of these proteins and their mRNAs during seed development in cv. Provar was studied by SDS polyacrylamide gel electrophoresis and by "Northern" (DNA-RNA) hybridization. The 11S acidic and basic subunits and the 7S α' and α subunits began to accumulate 18-20 d after pollination, shortly after the termination of cell division in developing cotyledons, whereas the 7S β and 11S A-4 subunits were not detected until one to two weeks later, during the maturation phase of development. Messenger RNAs for 7S and 11S proteins were first detected 14-18 d after pollination, several days before the accumulation of storage proteins. Extracts from embryonic axes contained reduced levels of the 7S α subunit, very little 11S protein, no detectable 7S β or 11S A-4 subunits, and an additional 7S subunit not found in cotyledons. Soybean axes and cotyledons therefore differ in their synthesis of seed storage proteins.

Wild-type plants of Arabidopsis thaliana strain "Columbia" regenerated at a high frequency from immature cotyledons cultured on a shoot-inducing medium containing 1.0 mg/l 6-benzylaminopurine and 0.1 mg/l 1-naphthaleneacetic acid. Cotyledon segments expanded rapidly and produced numerous shoots after 2-3 weeks in culture. Regeneration occurred in the absence of the original shoot apex. Hypocotyl segments from immature embryos produced root hairs and callus in culture but only rarely developed shoots. Hygromycin, kanamycin and G-418 inhibited cotyledon expansion and shoot formation in culture. Vancomycin was much less toxic to cotyledon segments than either carbenicillin or cefotaxime. Immature cotyledons therefore yield numerous regenerated plants that may be useful in future transformation studies.

Analysis of the expression of genes encoding the beta-conglycinin seed storage proteins in soybean has been used to extend our understanding of developmental gene expression in plants. The alpha, alpha', and beta subunits of beta-conglycinin are encoded by a multigene family which is organ-specific in its expression. In this study we report the differentially programmed accumulation of the alpha, alpha', and beta subunits of beta-conglycinin. Multiple isomeric forms of each subunit are present in the dry seed, but the timing of their accumulation is unique for each subunit. The previously reported variation in amount of alpha' and alpha subunits in axis and cotyledons is also reflected in the amount of subunit specific mRNA which is present in each tissue. The beta subunit, previously undetected in soybean axes, is found to be synthesized but rapidly degraded. These differences in beta-conglycinin protein accumulation may be reflected by the morphological differences observed in protein bodies between these two tissues.

Arabidopsis thaliana (Cruciferae) has been chosen as a model system for mutant analysis of plant embryo development. The isolation and characterization of embryo-lethal mutants of Arabidopsis are simplified by the following genetic and developmental characteristics: (1) Aborted seeds can be easily recognized in both immature and mature fruits; (2) spontaneous or physiological embryo abortion is rare; (3) individual fruits contain 30–60 seeds and show clear segregation of normal and aborted seeds in plants heterozygous for embryo-lethal mutations; (4) mature plants contain hundreds of seeds at every stage of development; (5) the generation time of 5–6 weeks is one of the shortest among flowering plants; (6) the diploid chromosome number is 10, and a large number of morphological mutants are available for linkage studies; (7) plants naturally self-pollinate but can be experimentally cross-pollinated; (8) the pattern of normal embryo development is very similar to that of Capsella , a plant used extensively for descriptive and experimental studies; (9) embryo and ovule culture can be used to study the biochemistry of mutant embryos; and (10) the isolation of temperature-sensitive mutants is possible because mature plants are small and viable over a wide range of temperatures. This report describes normal development in Arabidopsis thaliana , strain “Columbia,” and outlines the system used to identify embryo-lethal mutants. A method of classifying embryo-lethal mutants as cellular, nutritional, or developmental lethals is presented, and the potential application of these mutants to the study of normal embryo development is discussed.

Lethal mutants have been used in a variety of animal systems to study the genetic control of morphogenesis and differentiation. Abnormal development has been shown in some cases to be caused by defects in basic cellular processes. We describe in this report an embryo-lethal mutant of Arabidopsis thaliana that can be rescued by the addition of biotin to arrested embryos cultured in vitro and to mutant plants grown in soil. Mutant plants rescued in culture produced phenotypically normal seeds when supplemented with biotin but became chlorotic and failed to produce fertile flowers in the absence of biotin. Arrested embryos were also rescued by desthiobiotin, the immediate precursor of biotin in bacteria. Langridge proposed 30 years ago (1958, Aust. J. Biol. Sci. 11, 58-68) that the scarcity of plant auxotrophs might be caused by lethality prior to germination. The bio1 mutant of Arabidopsis described in this report clearly demonstrates that some auxotrophs in higher plants are eliminated through embryonic lethality. Further analysis of this mutant should provide valuable information on the nature of plant auxotrophs, the biosynthesis and utilization of biotin in plants, and the underlying causes of developmental arrest in lethal mutants of Arabidopsis.

The genetic control of embryo development in higher plants has been approached in part through the isolation and characterization of embryo-lethal mutants (Meinke 1986). The most extensive studies have dealt with defective kernel mutants of corn (Sheridan and Neuffer 1982), embryo-lethal mutants of Arabidopsis (Muller 1963; Meinke 1985), and variant cell lines of carrot unable to complete somatic embryogenesis in vitro (Breton and Sung 1982). Arabidopsis thaliana (Cruciferae) has been used as a model plant system for various studies in developmental and molecular genetics because it produces many seeds per plant and has a short generation time, low chromosome number, well-characterized mutants (Redei 1975), an established linkage map (Koornneef et al. 1983), and an unusually small genome with little repetitive DNA (Leutwiler et al. 1984). Recessive embryolethal mutants of Arabidopsis isolated following EMS seed mutagenesis have been shown previously to differ with respect to the stage of developmental arrest, the color of arrested embryos and aborted seeds, the percentage and distribution of aborted seeds in heterozygous siliques, the extent of abnormal embryo development, the response of mutant embryos in culture, the development of homozygous mutant plants, the formation of protein and lipid bodies, and the accumulation of seed storage proteins (Meinke and Sussex 1979a,b; Meinke 1982, 1985; Meinke et al. 1985; Marsden and Meinke 1985).

Seventeen embryo-lethal mutants ofArabidopsis thaliana with lethal phases ranging from the globular to mature cotyledon stages of development were analyzed by culturing arrested embryos on nutrient media designed to promote either callus formation or the completion of embryo development and the recovery of homozygous mutant plants. Enriched media supplemented with vitamins, amino acids, and nucleosides were used to identify potential auxotrophic mutants. Wild-type embryos produced extensive callus on basal and enriched media supplemented with 2,4-D and kinetin. Numerous roots developed when wildtype callus was grown in the presence of NAA and kinetin. Mutant embryos arrested prior to the heart stage of development formed only a slight amount of callus on basal and enriched media. Arrested embryos from mutants 122G-E and 112A-2A reached a later stage of development and gave the most interesting responses in culture. 122G-E mutant embryos failed to grow on basal media but produced extensive callus and homozygous mutant plants on enriched media. The specific nutrient required for growth of this mutant remains to be determined. Arrested embryos from mutant 112A-2A developed into abnormal plants without roots when placed in culture. Mutant callus also failed to form roots on a variety of root-inducing media. Expression of this mutant gene therefore disrupts development of the root apical meristem during both embryogenesis in vivo and organogenesis in vitro.

Embryo-lethal mutants of Arabidopsis thaliana were isolated by treating mature seeds with an aqueous solution of ethyl methanesulfonate (EMS), screening the resulting M-1 plants for siliques containing 25% aborted seeds following self-pollination, and verifying the presence of induced mutations in subsequent generations. Thirty-two recessive lethals with a Mendelian pattern of inheritance were examined in detail. Developmental arrest of mutant embryos ranged from the zygotic stage of embryogenesis in mutant 53D-4A to the linear and curled cotyledon stages of development in mutants 112A-2A and 130B-A-2. These lethal phases did not change significantly when plants were grown at 18 °C rather than at 24 °C. Differences between mutant lines were found in the color of arrested embryos and aborted seeds, the percentage and distribution of aborted seeds in heterozygous siliques, the size of arrested embryos, and the extent of abnormal development. Unusual mutant phenotypes included the presence of unusually large suspensors, distorted and fused cotyledons, reduced hypocotyls, and arrested embryos without distinct cotyledons or hypocotyl tissue. The isolation of eight new mutants with a non-random distribution of aborted seeds in heterozygous siliques provides further evidence that many of the genes that control early stages of embryogenesis in plants are also expressed prior to fertilization.

Normal and aborted seeds from two recessive embryo-lethal mutants (79A and 124D) of Arabidopsis thaliana were shown to be distributed nonrandomly along the length of heterozygous siliques. Significantly more than half of the aborted seeds in these two mutants were located in the top half of the silique, in the region closest to the stigma surface. Segregation ratios (percent aborted seeds) were unusually low at the base of the silique, and slightly higher than expected at the tip. In contrast, aborted seeds from four other embryo-lethal mutants (87A, 123B, 50B, and 71E) were distributed randomly along the length of the silique. These results suggest that the mutant genes in 79A and 124D are expressed during both the gametophytic (n) and sporophytic (2n) phases of development. These two mutants provide further evidence for the hypothesis that many genes expressed prior to fertilization also perform a critical function during growth and development of the sporophyte. Embryo-lethal mutants of Arabidopsis may therefore be useful in future studies of gametophytic gene expression and the regulation of pollen-tube growth in higher plants.

Mature seeds of Arabidopsis thaliana strain “Columbia” were soaked for 7.5 hr in an aqueous solution of the chemical mutagen ethyl methanesulfonate (0.05, 0.10, or 0.50%, v/v). Embryo-lethal mutants were identified in the resulting M-1 chimeral plants by screening the first five siliques of each plant and noting the frequency of aborted seeds. Three hundred sixty seeds were treated at each mutagen dose; the frequency of embryo-lethal mutants ranged from 1–3% of the M-1 plants grown from seeds exposed to 0.05% EMS, to 20–30% of the M-1 plants at the highest mutagen dose. Six embryo-lethal mutants identified through screening of M-1 plants were chosen for detailed studies in subsequent generations. All six mutants segregate as nonallelic, Mendelian recessive lethals, and are maintained as heterozygotes since homozygotes die as embryos. Fruits of heterozygous plants contain 25% aborted seeds and 75% phenotypically normal seeds ( 2 3 heterozygotes and 1 3 wild type). Segregation ratios are not temperature sensitive; the same frequency of aborted seeds is found in plants grown at 18, 25, and 32°C. Embryo arrest and eventual lethality in each mutant occur at a characteristic stage of early embryo development: globular-heart, globular, early globular, or preglobular. Arrested embryos from five of the six mutants resemble normal embryos at early stages of development. Developmental arrest of the embryo proper in the remaining mutant is followed by abnormal growth of the suspensor, an embryonic structure that attaches the embryo proper to the maternal tissue.

Arrested embryos from lethal (emb) mutants of Arabidopsis thaliana were rescued on a nutrient medium designed to promote plant regeneration from immature wild-type cotyledons. The best response was observed with mutant embryos arrested at the heart to cotyledon stages of development. Embryos arrested at a globular stage produced callus but failed to turn green or form normal shoots in culture. Many of the mutant plants produced in culture were unusually pale with abnormal leaves, rosettes, and patterns of reproductive development. Other plants were phenotypically normal except for the presence of siliques containing 100% aborted seeds following self-pollination. These results demonstrate that genes with essential functions during plant embryo development differ in their pattern of expression at later stages of the life cycle. Most of the 15 genes examined in this study were essential for embryogenesis but were required again for subsequent stages of development. Only EMB24 appeared to be limited in function to embryo development. These differences in the response of mutant embryos in culture may facilitate the classification of embryonic lethals and the identification of genes with developmental rather than housekeeping functions.

The major storage proteins isolated from wild-type seeds of Arabidopsis thaliana (L.) Heynh., strain “Columbia”, were studied by sucrose gradient centrifugation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Both the hypocotyl and cotyledons of mature embryos contained abundant 12 S (cruciferin) and 2 S (arabin) proteins that appeared similar in size and subunit composition to the cruciferin (12 S) and napin (1.7 S) seed-storage proteins of Brassica napus. The 12 S protein from Arabidopsis was resolved by SDS-PAGE into two groups of subunits with approximate relative molecular weights of 22–23 kDa (kilodalton) and 30–34 kDa. These polypeptides accumulated late in embryo development, disappeared early in germination, and were not detected in other vegetative or reproductive tissues. Accumulation of the 12 S proteins in aborted seeds from nine embryo-lethal mutants with different patterns of abnormal development was studied to determine the extent of cellular differentiation in arrested embryos from each mutant line. Abundant 12 S proteins were found in arrested embryos from two mutants with late lethal phases, but not in seven other mutants with lethal phases ranging from the globular to the cotyledon stages of embryo development. These results indicate that the accumulation of seed-storage proteins in wild-type embryos of Arabidopsis is closely tied to morphogenetic changes that occur during embryo development. Embryo-lethal mutants may therefore be useful in future studies on the developmental regulation of storage-protein synthesis.