History of Research in Goldberg Lab

From Cot Curves to Genomes to Fields of Crops

7/17/14

1. Overview
2. In the Beginning
3. Marching on to Specific Genes
4. Using a Spectrum of Approaches to Dissect Plant Gene Function
5. From Test Tube to Farm
6. Using Lasers and Genetics to Study Anther Development
7. Understanding "How to Make a Seed"
8. Back to the Future -- Using Genetics and Genomics to Dissect Embryo Development
9. Grant Support and Special Recognition

Overview [back to top]

My laboratory has been investigating the molecular processes controlling the development of specialized cells in higher plants. The major questions that I have addressed in my research are (1) how are genes organized in the genome, (2) what are the mechanisms that control the regulation of plant gene expression, (3) what are the sequences that program plant gene expression during development, (4) what are the genes that control the differentiation of specific plant cell types, and (5) what events cause an undifferentiated cell to take on a specialized state. The focus of my research has been on cells that are involved in plant reproductive processes (e.g., embryonic cells, cells of the male reproductive organ). I have used a variety of approaches and model plants to answer these questions -- always using the plant and approach best-suited to answer a specific question. Approaches include the use of molecular biology, cell biology, and genetics. Historically, I have used a variety of different plant systems, because there is not an ideal plant to investigate all questions of plant development and molecular biology and because approaches, techniques, and knowledge have changed over the years enabling a model plant, such as Arabidopsis, to take "center stage" at the present time. Plants that I have utilized for my experiments include (1) tobacco for molecular studies on plant organ systems and seeds, and as a transgenic plant to dissect gene control elements, (2) soybean, an economically-important crop, for investigating seed and embryo development, (3) canola, an economically-important crop, used as a model to genetically engineer for male fertility control, (4) Arabidopsis as a model plant for the genetic dissection of male reproduction and seed development, and, most recently, (5) Scarlet Runner Bean because it has "giant" embryos that are well-suited for using genomics to investigate the earliest stages of plant development. The plants used in my laboratory at the present time are the Scarlet Runner Bean, Arabidopsis, and tobacco.

In the Beginning [back to top]

At the beginning of my academic career (1973-1983) my laboratory investigated how DNA sequences are organized in plant genomes. These were the first "genomics" experiments carried out with plants and demonstrated that plant genomes have repetitive and single copy sequences that are organized in a pattern similar to that in animal genomes -- that is, plant genomes are complex, contain a large amount of genetic information, and contain many different types of interspersed repetitive sequence families (Zimmerman and Goldberg, Chromosoma 59, 227-252, 1977; Goldberg, Biochemical Genetics 16, 45-68, 1978). The organization of plant genomes turned out to be very different from that of fungal genomes which were also investigated in my laboratory. These genomes are much less complex and do not have large amounts of repetitive DNA sequences (Hudspeth et al., PNAS 74, 4332-4336, 1977). Thus, from the point of view of genome organization, experiments in my laboratory demonstrated that higher plants more closely resemble animals such as humans and mice.

During this period, my laboratory also carried out the first gene expression experiments to determine the extent to which genes are regulated during the entire plant life cycle and we made an estimate of the number of genes required to program all of plant development (Goldberg et al., Cell 14, 123-131, 1978 [this was one of the first plant papers published by the journal Cell]; Kamalay and Goldberg, Cell 19, 935-946, 1980; Goldberg et al., Developmental Biology 83, 201-217, 1981; Kamalay and Goldberg, PNAS 81, 2801-2805, 1984). These experiments demonstrated for the first time (1) how many genes are expressed in each vegetative and reproductive organ of a plant, (2) how many genes are active during different periods of embryo development, (3) that each organ system expresses a unique set of genes, (4) that plants have a complex set of nuclear RNA sequences similar to that found in animals (we now know that this additional nuclear RNA complexity is primarily due to spliced introns), (5) that tens of thousands of genes (~60,000) are required to program the entire life cycle of a higher plant, and (6) that transcriptional and post-transcriptional processes are important in regulating plant gene expression. Gene expression studies carried out with a representative of the fungal kingdom produced a very different picture -- that is, fungi do not have a complex set of nuclear RNAs (we now know this is due to a relative absence of introns in fungal genes) and that they require about a third of the genetic information to program their life cycle (Timberlake et al., Cell 10, 623-632, 1977).

The gene expression experiments carried out in my laboratory during this period provided the first insights into the complex processes regulating gene activity during the plant life cycle. The questions addressed during this period of my career were similar to those being addressed by the much more sophisticated sequencing and genomics tools of the present day -- however, the general answers have stood the test of time and provided the foundation for many of the experiments with specific genes once the gene cloning era began (Goldberg, Plant Physiology 125, 4-8, 2000).

Marching On to Specific Genes [back to top]

When it became possible to clone specific plant mRNAs and genes (~1976), my laboratory began to investigate the regulation of soybean seed protein genes during plant embryo development because they encoded highly abundant plant mRNAs and proteins (e.g., storage proteins, Kunitz trypsin inhibitor, lectin) and they are expressed primarily within seeds. During this period (1978-1990) seed protein genes served as a model system in my laboratory for investigating how sets of genes are "coordinately regulated" during plant development.

My laboratory was one of the first to generate plant genome and cDNA libraries and investigate the organization and expression of specific plant genes and genomic regions (Goldberg et al., Developmental Biology 83, 218-231, 1981; Fischer and Goldberg, Cell 29, 651-660, 1982; Goldberg et al., Cell 33, 465-475, 1983; Goldberg et al., Cell, 59, 149-160, 1989; Harada et al., Plant Cell 1, 415-425, 1989; Nielsen et al., Plant Cell 1, 313-328, 1989). These experiments demonstrated that (1) plant genes contain introns, (2) differentially regulated genes are tightly clustered among each other (leading to the hypothesis in the pre-transgenic-plant era that each gene contained its own cis-control element), and (3) on average, plant genes are spaced approximately 6-7 kb apart in complex plant genomes such as soybean -- a value not too different from that found today by genome sequencing projects!

During this period (1980-1990) my laboratory used a variety of approaches to attempt to unravel the mechanisms responsible for the "embryo specificity" of seed protein genes. That is, how are sets of genes activated and repressed during specific periods of embryo development? "Run-off" transcription assays demonstrated that seed protein genes are regulated primarily at the transcriptional level, but that post-transcriptional processes are also important in modulating mRNA levels (Walling et al., PNAS 83, 2123-2127, 1986). To identify the transcription factors that play a role in regulating seed protein gene transcription, my laboratory carried out one of the first DNA-binding protein experiments with plant genes and identified "nuclear factors" (which we now know are HMG proteins) that bind to the seed protein gene promoters (Jofuku et al., Nature 328, 734-737, 1987). My laboratory also carried out one of the first in situ hybridization experiments done with plants and demonstrated that seed protein mRNAs accumulate in a specific pattern during embryo development and are localized in specific embryo cells and regions (Barker et al., PNAS 85, 458-462; Perez-Grau and Goldberg, Plant Cell, 1, 1095-1109). In fact, my laboratory developed one of the first in situ hybridization protocols for use with plant cells which is still in general use today (Cox and Goldberg, In Plant Molecular Biology: A Practical Approach, IRL Press, pgs. 1-34, 1988).

We also used mutants defective in the expression of specific seed protein genes to dissect the mechanisms of seed protein gene expression. These experiments uncovered a large insertion element in the seed lectin gene and demonstrated that this insertion element prevented the transcription of this gene during embryo development (Goldberg et al., Cell 33, 465-475, 1983). In addition, mutants defective in Kunitz trypsin inhibitor and glycinin gene expression were shown to be due to a frameshift and an inversion, respectively (Jofuku et al., Plant Cell 1, 427-435, 1989; Cho et al., Plant Cell 1, 339-350, 1989). Finally, when the era of plant transformation arrived (1983), my laboratory carried out some of the first experiments to show that individual plant genes are regulated correctly in transgenic plants (Okamuro et al, PNAS 83, 8250-8255, 1986; Barker at al., PNAS 85, 458-462; Perez-Garau and Goldberg, Plant Cell 1, 1095-1109, 1989) -- something that we take for granted today! These experiments also demonstrated that large blocks of genes can be transferred form one plant to another and maintain their developmental-specific expression patterns (Okamuro et al, PNAS 83, 8250-8255) -- supporting our earlier hypothesis from the pre-transgenic-plant era that each plant gene has its own cis-element that is required to program expression during development.

Collectively the experiments carried out during this period of my career provided a conceptual understanding for how specific plant genes are organized and regulated during seed development and provided the foundation for many of the genomics experiments that we are carrying out today.

From Test Tube to Farm [back to top]

There are many aspects of my research on plant genes that have been important, exciting, and were "firsts" in their day. However, I will outline briefly what I consider to be one of the most exciting, lucky, challenging, and important research periods in my career (1987-1996) -- research that not only has been shown to improve crop production, but has also influenced me considerably on how carry out research to this day. That is, by establishing a team of scientists with complementary expertise that collectively do the work and collectively share the rewards -- a "laboratory without walls." I have been fortunate to have been able to translate basic research on plant gene regulation carried out in my laboratory to the improvement of major crop plants -- that is, go from the "test-tube to the farm."

"In the beginning" my laboratory demonstrated that the anther, or male reproductive organ of a plant, expresses a unique set of thousands of genes and that these genes appeared to be under strict transcriptional control (Kamalay and Goldberg, Cell 19, 935-946, 1980; Kamalay and Goldberg, PNAS 81, 2801-2805). In an extensive series of experiments to uncover DNA sequence elements responsible for anther-specific transcription, tobacco anther cDNA libraries were constructed and a large set of anther-specific cDNAs were identified. Using in situ localization procedures developed in my laboratory (Cox and Goldberg, In Plant Molecular Biology: A Practical Approach, IRL Press, pgs. 1-34, 1988; Goldberg, Science 240, 1460-1467, 1988), many of these cDNAs were shown to encode mRNAs that are localized specifically in the tapetal cell layer that surrounds the pollen chamber (Goldberg, Science 240, 1460-1467, 1988; Koltunow et al., Plant Cell 2, 1201-1224, 1990). Transcription experiments and studies with transgenic plants demonstrated that the tapetal-specific genes are under strict transcriptional control (as predicted from our earlier-era studies) and that for one tapetal-specific gene, designated as TA29, a region of ~120 bases 5' to the gene is sufficient and necessary to program tapetal-specific transcription (Koltunow et al., Plant Cell 2, 1201-1224, 1990). While these experiments were being carried out, I began a collaboration in 1987 with a biotechnology company, Plant Genetic Systems (PGS) of Gent, Belgium, in order to genetically engineer for male fertility control in major crop plants. That is, utilize information obtained in my laboratory on the regulation of anther-specific genes and gene regulation (specifically the TA29 gene) to develop a novel system for the generation of hybrid crops.

One of the major factors contributing to increases in crop productivity during the 20th century has been the breeding of hybrid varieties in major crops such as corn. Crosses between inbred plant lines give rise to hybrid progeny that often exhibit hybrid vigor. These hybrids are heartier, more resistant to disease, less susceptible to environmental stress, and generate higher yields than their inbred parents. For example, since the development of hybrid corn in the 1920's, the yield of corn has increased almost 300% in the United States alone! The reproductive biology of higher plants makes it difficult to perform directed crosses between inbred lines on a scale large enough for hybrid crop production. Because male and female reproductive organs in plants are present within the same flower (and plant), a high level of self-pollination and fertilization can occur within each inbred line. To prevent self-fertilization in the field, the male reproductive organ must be severed mechanically from one of the varieties in order to guarantee that the two varieties will cross and produce hybrid offspring. This is very expensive. For example, mechanical emasculation for hybrid corn costs farmers over $250,000,000 per year in the United States alone! In addition, at the time our research efforts began commercial hybrids did not exist for many major crops such as canola (oilseed rape), rice, wheat, and soybean --all major sources of food and fiber world-wide. The male reproductive organs (and flowers) of these crops are too small to be removed mechanically on a commercial scale. The production of hybrids in these crops by a practical and economically reasonable method would allow yield increases (and increase in food production) similar to what was obtained with corn in the century.

Our strategy was to use information obtained in my laboratory on the tapetal-specific transcription of the TA29 gene (Koltunow et al., Plant Cell 2, 1201-1224, 1990) to engineer a male-sterility gene, or a "genetic laser," that would selectively destroy tapetal cells which are necessary for pollen production. The resulting male-sterile plants could then be used as one of the parental varieties (the female line) in a hybrid breeding program. This would allow hybrid crops, such as corn, to be produced more efficiently and economically --- eliminating the need to hand emasculate one of the parental lines which, as I pointed out, is labor-intensive, expensive, and impractical for many crops. This strategy would also allow us to introduce "genetic lasers" into crops that have not been amenable to hybrid seed production strategies, such as oilseed rape and rice, and would permit the generation of new, higher-yielding hybrid varieties in a wide range of important crop plants.

To generate a dominant male-sterility gene ("genetic laser"), we fused the TA29 gene tapetal-specific transcriptional control region (Koltunow et al., Plant Cell 2, 1201-1224, 1990) with the barnase gene of a bacteria, Bacillus amyloliquifaciens. This bacteria uses barnase as a "weapon" to fight off predators. The barnase gene produces a cytotoxin (RNase) that degrades RNA and destroys the cells in which it is present. Genetically-engineered tobacco and canola plants containing the cytotoxic TA29/barnase gene were shown to be 100% male sterile as a result of selective tapetal cell ablation during a specific time interval of anther development (Mariani et al., Nature 347, 737-741, 1990). These results demonstrated, for the first time, that it is possible to generate genetically engineered male-sterile plants that develop normally but fail to produce pollen, and that the tobacco tapetal-specific promoter retains its specificity in a major crop plant such as canola.

A restorer system is required for hybrid seed production in many crops to either maintain parental male-sterile lines, restore fertility to male-sterile hybrids in order to generate fruits and seeds, or both. To generate a dominant male-fertility restorer gene, we fused the TA29 gene promoter with the barstar gene from Bacillus amyloliquifaciens. Barstar is a RNase- inhibitor that is specific for the barnase ribonuclease. Barstar/barnase complexes are highly stable and lead to the inactivation of barnase cytotoxic activity. Bacillus bacteria use barstar as a "fail-safe" system -- that is, to ensure that the Bacillus is not killed by the cytotoxic barnase protein before it is secreted from the cell. Introduction of the chimeric TA29/barstar gene into male-sterile canola plants containing the TA29/barnase gene lead to the formation of tapetal-cell-specific barnase/barstar complexes and restores fertility to genetically engineered male-sterile plants (Mariani et al., Nature 357, 384-387, 1992). These results demonstrated that it is possible to use the TA29/barnase male-sterility gene and the TA29/barstar male-fertility restorer gene to genetically engineer for male fertility control in crop plants, and that the TA29/barnase and TA29/barstar gene system provides a novel new breeding strategy for hybrid seed production. Simply put, the male fertility control system permits male-sterile plants to be generated and then restored to full male fertility. That is, it provides a new approach for hybrid crop production based on genetic engineering that is inexpensive and can be applied universally to all crop plants.

The tobacco TA29 promoter has now been shown to retain its tapetal cell specificity in a variety of monocot and dicot plants. Recent experiments with the TA29/barnase gene have resulted in production of hybrid canola and corn plants. These results indicate that the TA29/barnase and TA29/barstar gene system works in widely divergent plants and should be highly effective in hybrid seed production programs for a wide-range of monocot and dicot crop plants. Field trials carried out from 1992-1995 demonstrated that the male fertility control system works under actual field conditions to produce hybrid canola plants. And these plants produced a 25% increase in seeds compared to non-hybrid varieties! Regulatory approval was obtained to commercially produce the first genetically-engineered, hybrid canola plants in the world, and, in 1996, hybrid seeds produced by the TA29/barnase/barstar system were released commercially to farmers throughout Canada and the United States.

Currently, there are several million acres of hybrid canola plants in Canada containing the TA29 gene promoter. Thus, a high-risk project started in 1987 to genetically engineer for male sterility control in higher plants using knowledge gained from basic research on gene expression during anther development led to a novel strategy for hybrid seed production that works in major crop plants. And which has generated a significant increase in seed yield for one major crop, canola, which is used for vegetable oil production. At the present time the male fertility control system is being transferred to wheat, vegetables, and other crops, and field trials are beginning that will lead, in the future, to the commercial release of genetically engineered hybrid corn seed. I am very proud of the male-fertility control system because research in my laboratory to investigate the regulation of plant gene expression has made a significant impact on the "real-life" efforts of farmers in the field.

Using Lasers and Genetics to Study Anther Development [back to top]

The identification of genes transcribed in different anther cells and tissues (Koltunow et al., Plant Cell 2, 1201-1224, 1990) and the success that we had in using barnase to ablate tapetal cells (Mariani et al., Nature 347, 737-741, 1990; Mariani et al., Nature 357, 384-387, 1992) opened up the possibility that the barnase/barstar system could be used to investigate signaling processes during anther development. At this time, it also became possible to use Arabidopsis to isolate genes important in plant processes by identifying T-DNA tagged mutants. During this period (1990-2000), an extensive series of molecular and genetic experiments were initiated in my laboratory to understand the processes by which cells involved in anther dehiscence (i.e., release of pollen grains at flower opening) differentiated during anther development. Our major question was to determine the extent to which signaling processes played a role in the differentiation of anther cell and tissue types.

We developed a novel cell ablation approach using the barnase/barstar system in tobacco that demonstrated that the stomium (the anther wall cells that break and allow for pollen release) functions autonomously in anther development -- that is, ablating these cells has no effect on the differentiation of adjacent cells or the anther dehiscence process, except that pollen grains are not released (Goldberg et al., Phil. Trans. Royal Soc., 350, 5-17, 1995; Beals and Goldberg, Plant Cell 9, 1527-1545, 1997). The novel aspect of these experiments is that a chimeric barnase gene using one promoter was used to ablate the stomium, while a chimeric barstar gene using a different promoter was used to "protect" surrounding tissues and cells from the cytotoxic effects of barnase. This approach enables promoters with multiple, overlapping cell specifities to be used in targeted cell ablation experiments and eliminates the need for absolute promoter cell specificity (Beals and Goldberg, Plant Cell 9, 1527-1545, 1997).

A complementary series of genetic experiments were carried out that identified many male-sterile Arabidopsis mutants that were generated by T-DNA mutagenesis (Sanders et al., Sex Plant Reproduction 11, 1-27, 1999. Several of these male-sterile mutants had defects in the anther dehiscence process. One of these T-DNA tagged male-sterile mutants, designated as delayed dehiscence1, was shown to have a defect in the timing of pollen release -- that is, pollen develop normally but are not released until after the stigma, or pollen-receptive tissue of the female reproductive organ, is no longer able to accept pollen grains (Sanders et al., Sex Plant Reproduction 11, 1-27, 1999). The T-DNA tagged gene was shown to encode a key enzyme of the jasmonic acid biosynthesis pathway and application of exogenous jasmonic acid to floral buds was able to "rescue" the delayed dehiscence1 phenotype and restore male fertility (Sanders et al., Plant Cell 12, 1041-1062, 2000). These experiments demonstrated for the first time that jasmonic acid signaling plays a major role in release of pollen grains from the anther at flower opening.

For me, these experiments "completed the circle" and connected my initial experiments on the regulation of gene expression in the anther to the regulation of specific anther genes to the targeted ablation of specific anther cell types to the development of a new system to produce hybrid crops to the demonstration that an important plant hormone, jasmonic acid, is required for the release of pollen grains from a specific cell type (the stomium) during anther development -- that is, 20 years of attempting to unravel the processes regulating the activity of plant genes and the differentiation of plant cell types!

Understanding "How to Make a Seed" [back to top]

As pointed out above, two of the major questions that I have been addressing in my career is how genes are regulated during plant embryo development, and how differentiated organs and cell types of the embryo form following fertilization (e.g., Goldberg et al., Developmental Biology, 83, 201-217. 1981; Goldberg et al., Developmental Biology 83, 218-231, 1981; Fischer and Goldberg, Cell 29, 651-660, 1982; Goldberg et al., Cell 33, 465-475, 1983; Okamuro et al., PNAS, 83, 8240-8244, 1986, Walling et al., PNAS 83, 2123-2127; Jofuku et al., Nature, 328, 734-737; Barker et al., PNAS 85, 458-462, 1988; Jofuku and Goldberg, Plant Cell 1. 1079-1109, 1989; Goldberg et al., Cell 59, 149-160, 1989; Harada et al., Plant Cell 1, 415-425, 1989; Jofuku et al., Plant Cell 1, 427-435, 1989; Nielsen et al., 1989, Plant Cell 1, 313-328; Okamuro and Goldberg; Plant Cell 4, 1141-1146, 1992; Goldberg et al., Science 266, 605-614, 1994). Following my collaboration with PGS to engineer for male fertility control in crop plants (Mariani et al., Nature 347, 737-741, 1990; Mariani et al., Nature 357, 384-387, 1992), I became convinced that large team efforts involving several laboratories could be very effective in solving a "big problem" of plant biology -- for example, what are all the genes and processes required "to make a seed." In 1990, I initiated a project, designated as the "Embryo 21st Century Project", to dissect the mechanisms controlling seed and embryo development using a combination of genetics and molecular biology approaches. This project linked together faculty and students at several UC campuses including UCLA (my lab), UC Davis (John Harada's lab), UC Santa Cruz (lab of Jack Okamuro and Diane Jofuku), and UC Berkeley (Bob Fischer's lab). The idea was that (1) all of the labs would participate collectively in generating the ideas and approaches, (2) all the labs would collectively carry out the experiments, and (3) all of the labs would collectively share in the "rewards"; that is, publish the results together. What we created in 1990 was a "virtual laboratory or a "laboratory without walls." to dissect the processes controlling seed and embryo development. The Embryo 21st Century Project re-directed the work in my lab and the other participating labs to a focus primarily on Arabidopsis as a model system to study seed development.

The "Embryo 21st Century" Project has been enormously successful. It pre-staged the multi-campus research efforts that have now become "the norm" for large genome projects and it demonstrated how a "virtual lab" can be created in which all of the faculty, students, postdocs, and staff work collectively towards a common goal. In fact, it has provided an alternative model to the "old" lab by lab "pigeon-hole" approach for doing research in an academic setting. The "Embryo 21st Century" Project evolved formally into the University of California "Seed Institute" in 1998 when a partnership was created between the University and Ceres, Inc. (see Service Statement).

A large number of significant research results have come out of the "21st Century Embryo Project" and "Seed Institute" on the mechanisms of embryo and seed development. Most of these projects had their origin in several large genetic screens to identify mutants (and their corresponding genes) that had defects in embryo development (Yadegari et al., Plant Cell 6, 1713-1729, 1994; West et al., Plant Cell 6, 1731-1745, 1994).

· Early, undifferentiated plant embryos are divided into domains or territories in which distinct transcriptional events occur. These territories pre-stage the morphogenetic events that occur in these territories later in embryo development (Goldberg et al., Science 266, 604-614, 1994.)

· Plant embryos can undergo cell differentiation in the absence of morphogenesis; that is plant cells differentiate according to a strict temporal sequence regardless of whether embryonic organs form (Goldberg et al., Science 266, 604-614, 1994; Yadegari et al., Plant Cell 6, 1713-1729, 1994).

· The LEC1 gene (West et al., Plant Cell 6, 1731-1745, 1994; Lotan et al., Cell 93, 1195-1203, 1998) and the LEC2 gene (Stone et al., PNAS, In Press, 2001) encode different transcription factors and are central regulators of embryo development. The LEC1 gene encodes a HAP3-like, CAAT-box binding protein while the LEC2 gene encodes a B3 domain transcription factor. Each of these genes can induce embryo formation in non-embryonic cells when ectopically expressed -- for example, embryos on leaves (Lotan et al., Cell 93, 1195-1203, 1998; Stone et al., PNAS, In Press, 2001) -- indicating that each gene is capable of initiating a series of events leading to the formation of a mature embryo. LEC1 and LEC2 are among the first plant genes that have been shown to be central regulators of embryo formation, and both of these genes provide entry into the regulatory circuits leading to embryo development.

· The LEC1 and LEC2 genes are both activators and repressors of plant gene expression. An extensive series of recent RNA profiling experiments using Affymetrix microchips containing 8200 Arabidopsis sequences has indicated that both the LEC1 and LEC2 regulate, either directly or indirectly, the expression of hundreds of genes (Bui et al., unpublished results). These experiments were carried out in the UCLA Microchip Core Facility that I helped set up using Seed Institute funds (see Service Statement).

· Polycomb-group genes, like MEDEA (Kiyosue et al., PNAS, 96, 4186-4191) and FIE (Ohad et al., Plant Cell 11, 407-416, 1999) are important regulators of seed development and play a major role in the development of the endosperm. Mutations in these genes lead to the development of endosperm, fruit, and seed-like structures in the absence of fertilization (Kiyosue et al., PNAS, 96, 4186-4191; Ohad et al., Plant Cell 11, 407-416, 1999). The isolation of genes that can induce both embryo formation (LEC1, LEC2) and seed and fruit formation (FIE, MEDEA) in the absence of fertilization opens up the possibility of engineering for apomixis in crop plants; that is allowing for seed and embryo formation in the absence of fertilization. This would be of enormous value to agriculture because it would enable a hybrid to be made once using normal breeding methods and then the resulting hybrid vigor could be "fixed" forever using a genetically engineered apomixis system based on the genes identified in the "Embryo 21st Century Project" and the "Seed Institute." This would complement and greatly extend the use of the barnase/barstar system of male fertility control for establishing new hybrid crops. And it would mean, for the first time, that farmers could collect and "plant back" seeds from hybrid crops -- that is, the hybrids would not have to be made on a yearly basis.

· Imprinting of specific plant genes (e.g., MEDEA, FIE) plays a major role in endosperm and seed development (Kinoshita et al., Plant Cell, 11, 1945-1952, 1999; Yadegari et al., Plant Cell 12, 2367-2381, 2000).

· The chloroplast plays a significant role in embryo development (Apuya et al., Plant Physiology 126, 717-730; 2001). Mutations in the SCHLEPPERLESS gene, which encodes a 60a-chaperonin, cause defects in embryo and seed development (Apuya et al., Plant Physiology 126, 717-730; 2001). In addition, mutations in (1) the RASPBERRY2 gene (Yadegari et al., Plant Cell 6, 1713-1729, 1994), which encodes a nuclear-encoded chloroplast ribosomal protein (Yadegari, Ph.D. Thesis, UCLA, 1996), and (2) the RASPBERRY3 gene, which encodes a nuclear-encoded chloroplast protein similar to prokaryotic "cell-cycle" control protein (Apuya et al., unpublished results) block embryo morphogenesis -- that is, embryo cells differentiate in the absence of organ (cotyledon and axis) formation (Goldberg et al., Science 266, 605-614, 1994; Yadegari et al., Plant Cell, 6, 1713-1729, 1994). schlepperless, raspberry2, and raspberry3 mutant Arabidopsis plants have defective chloroplasts -- suggesting that defects in normal chloroplast processes (e.g., fatty acid biosynthesis) might indirectly affect embryo morphogenesis by preventing signaling molecules required for embryonic organ formation to be synthesized.


Back to the Future -- Using Genetics and Genomics to Dissect Embryo Development [back to top]

The Arabidopsis genetic screens yielded a wealth of information about genes that control embryo development. However, they have their limitations because of (1) redundancy of genes in the Arabidopsis genome and (2) the difficulty of distinguishing between mutations that affect "housekeeping-type" genes and those that affect genes that control embryo development (i.e., they both lead to similar phenotypes). To uncover "all of the genes" required to make an embryo, I initiated a genomics project within the "Seed Institute" to utilize a novel plant, the Scarlet Runner Bean, which has giant embryos (Weterings et al., Plant Cell, In Press, 2001). This is the central focus of research being carried out in my laboratory at the present time and most of the work described below is in progress (and summarized in the posters attached to the collection of reprints in the promotion dossier).

Scarlet Runner Bean embryos are ~100-times larger than those in Arabidopsis or tobacco plants. To uncover genes that play a major role in embryo development, my laboratory constructed cDNA libraries of mRNAs from two embryonic regions, the embryo proper and the suspensor, only 7-8 cell divisions after fertilization. The embryo proper leads to the formation of the mature embryo, while the suspensor is a terminally differentiated embryonic region that is required to nurture and sustain embryo-proper development during the early period of embryo development. Our rationale is that the Scarlet Runner Bean can serve as a "genomics engine" to uncover the genes and proteins required for the earliest stages of embryo development.

Suspensor and embryo-proper ESTs were sequenced using the high throughput sequencing capability of the UCLA DNA Sequencing Core Facility that I helped establish using Seed Institute funds (see Service Statement). At the present time, we have sequenced approximately 4,000 suspensor ESTs and 3500 embryo-proper ESTs and have provided the first profile of the spectrum of genes that are active in different regions of a plant embryo shortly after fertilization. These ESTs include a large number that encode transcription factors and signaling molecules -- including many that have not been observed previously in plants (e.g., an ortholog of neuralized, a Drosophila protein involved in nerve cell differentiation that operates via the notch signaling pathway). In addition, many suspensor ESTs encode enzymes required for the synthesis of gibberellic acid (GA), an important plant hormone. Using in situ hybridization, we have shown that the suspensor is the major site of GA biosynthesis in the embryo and that GA is probably exported to the embryo proper where it plays a major role in embryo proper growth and development.

An extensive series of experiments were carried out with two suspensor ESTs/cDNAs, designated G564 and C541, that encode proteins of unknown function (Weterings et al., Plant Cell, In Press, 2001). These experiments demonstrated that:

· G564 and C541 mRNAs accumulate asymmetrically in the basal (bottom) region of the three-celled embryo two divisions after fertilization.

· The asymmetric accumulation of G564 mRNA early in embryo development is due to transcriptional events-- that is, the G564 gene promoter is activated only in the basal embryo region after fertilization. It is possible that many of the transcription factors identified in the suspensor EST sequencing project interact with the G564 promoter and are responsible for this asymmetric transcription events.

· The apical and basal cells of a two-cell plant embryo are specified at the molecular level shortly after fertilization to develop into the embryo proper and the suspensor, respectively. It is possible, that many of the transcription factors and signaling proteins identified in the EST sequencing project play a role in these specification events.

Two strategies are being used to determine the function of transcription factors and signaling proteins uncovered in the EST sequencing project play in embryo development. First, orthologs of several Scarlet Runner Bean transcription factors and signaling proteins have been identified in Arabidopsis and these genes are in the process of being knocked-out using reverse genetics. Second, in collaboration with Roger Pennell (Ceres, Inc.) procedures are being developed for transforming and regenerating the Scarlet Runner Bean. Recently (9/7/01), transgenic Scarlet Runner Bean plants have been produced for the first time. If confirmed, these results open the exciting possibility of using anti-sense, RNA silencing (RNAi), and over-expression experiments in Scarlet Runner Bean directly to determine what role the transcription factors and signaling proteins uncovered in the early embryo EST sequencing project play in embryo development. In my opinion, this will transform the Scarlet Runner Bean into the most powerful system for dissecting the molecular mechanisms that control plant embryo development. And it will show the importance of choosing the correct plant to answer a specific question -- that is, one with giant embryos that can be isolated and manipulated to uncover molecules and events that control the earliest stages of plant development.

Finally, the experiments being carried out with the Scarlet Runner Bean bring research in my lab "back to the future" by making a direct connection between the gene expression experiments that I carried out 20 years ago with plant embryos ( Goldberg et al., Developmental Biology, 83, 201-217. 1981; Goldberg et al., Developmental Biology 83, 218-231, 1981) with powerful 21st century genomics approaches that should, with some luck, provide a conceptual understanding of the precise mechanisms and genes that are required to "make an embryo." I am very proud of the efforts carried out by all the participants in the "21st Century Embryo Project" and "Seed Institute" over the past decade because they have shown that there is a different way to carry out academic research and they have yielded significant new insights into the process of seed development.

Grant Support and Special Recognition [back to top]

My laboratory has been supported continuously since I began my academic career as an Assistant Professor over 40 years ago in 1973. I received my first grant from the National Science Foundation in 1974 to investigate the organization and expression of plant genomes. I have been awarded grants from the National Science Foundation, the United States Department of Agriculture, the United States Department of Energy, and several biotechnology companies. I was able to obtain $5,750,000 from Ceres, Inc. to create and support the "Seed Institute" for a five-year period (1998-2003).

Since arriving at UCLA in 1976 I have presented over 400 invited lectures on my research to various scientific forums all over the world. My research has been recognized by several awards. I was honored to have been the first Plant Molecular Biologist to receive the Tsune Kosuge Memorial Lecture Award from the UC Davis and Calgene, Inc. (1993). I was awarded the National Order of Scientific Merit, Grà Cruz from the President of Brazil in 1998 for my research contributions to Plant Molecular Biology and Agriculture. In addition, I was awarded the Gold Shield Prize for Excellence in Research and Undergraduate Teaching from the UCLA Academic Senate and Gold Shield Alumni (1998), was named UCLA Faculty Research Lecturer (1999), and was listed as making one of the "top 15 discoveries in UCLA history" (genetic engineering for male fertility control in crop plants). Finally, I was honored to be elected to the National Academy of Sciences in 2001.