Molecular Genetic Analysis of Senescence and the Reversal of Senescence in Soybean Cotyledons [Glycine max (L.) Bass]

Whole plant senescence is a programmed, endogenously controlled process in which cellular components are degraded and the resultant metabolites sequestered to other plant organs. All cells, with the possible exception of the meristem progenitors, go through cell division, expansion, elaboration of secondary structures to the primary cell wall, quiescence, and finally death. This senescence process is thought to be partially catalyzed by an increase in translation of degradative proteins coupled to a decrease in translation of vital proteins (Skadsen and Cherry, 1983). Inducers of senescence include wounding, fruit ripening, changes in hormone levels (auxin, cytokinins + auxin, Ca2+ + cytokinins, and ethylene), various stresses (anaerobiosis, heat, cold, UV light, Cd2+, and Li+), and pathogen attack (Theologis, 1992). The regulatory mechanisms that facilitate entry of a cell, tissue, or whole plant into this final developmental stage are largely unknown; however, hormonal controls are likely to be involved. A better understanding of the genetic regulation of senescence will be important for studies relating to : I) normal cellular differentiation, maintenance, and turnover; 2) plant defense mechanisms and the hypersensitive response; and 3) embryogenesis and development. In addition, economic benefits could result from reduction in post-harvest and post-production losses as well as improvement of crop yields. Evidence indicates that soybean cotyledon senescence and induced "rejuvenation" (i.e.,. the reversal of the senescence process) are associated with changes in gene expression (Marek and Stewart 1992; Skadsen and Cherry 1983). During germination the cotyledons emerge, become green, and develop leaf-like characteristics including photosynthetic and storage mobilizing capability. During normal development, the cotyledons senesce after expansion of the first and second trifoliates. It has been shown (Krul, 1974) that this process can be reversed by the removal of the epicotyl up to the "point of no return" (PONR), a developmental stage at which up to 90% of nucleic acids and 80% of proteins are lost from the senescing cotyledon. The objectives of this research project were to identify genes that show differential expression during senescence and rejuvenation in the soybean cotyledon system and to isolate clones of genes that are specifically up-regulated in expression during these processes. To accomplish these objectives, Northern blots were used to assay soybean cotyledon mRNA populations during various stages of senescence and rejuvenation using a collection of gene clones known or suspected to be involved in senescence and rejuvenation. These gene clones included: I) a tomato ethylene forming enzyme (EFE) and several other ethylene up-regulated genes; 2) from carnation, ACC synthase, glutathione stransferase and other unidentified senescence associated genes; 3) from soybean, three lipoxygenases, two vegetative storage proteins, rubisco large and small subunits and an iron superoxide dismutase; 4) from cucumber, malate synthase and isocitrate lyase; and 5) two Arabidopsis genes, Sag2 (cysteine protease) and Sag4 (currently unidentified). In addition, cDNA libraries were constructed from senescent, nonsenescent, and rejuvenated soybean cotyledons. Finally, a cDNA library made from senescing soybean cotyledons was screened for a Sag2 homologue. These studies revealed up-regulation in expression during cotyledon senescence for rubisco large and small subunits, two soybean vegetative storage proteins, and an Arabidopsis cysteine protease. A soybean homologue to the Arabidopsis cysteine protease (Sag2) was cloned from senescing soybean cotyledons and restriction analysis revealed that it contained two Pvull sites.


LIST OF TABLES
MANUSCRIPT I Table 1.
Gene    Senescence is a highly coordinated degenerative process that is characterized by a series of biochemical and physiological changes including increase in respiration, changes in hydrolytic enzyme activity, peroxidation of membrane lipids, and in some cases an increase in ethylene biosynthesis . This process represents the final developmental stage of the organism, tissue, or cell from the mature fully expanded state until death (Smart et. al., 1995). Cells remain viable and show tight metabolic regulation throughout the senescence process (Thomas, 1992;Matile, 1992;Smart, 1994).
Senescence is a metabolically active phase that requires transcription of new genes as well as down regulation of other genes (Skadsen and Cherry 1983).
Senescence includes, but is not limited to death of the whole plant. In some perennial species, senescence occurs in the portions of the plant that are above ground, but not in the bulb or the root. Deciduous woody plants lose their leaves but maintain the root system and all other above ground portions. A less dramatic example of senescence is the gradual loss of leaves from the base of the plant upward that is typical of many species.
Senescence at the cellular level is part of an important developmental process as exemplified during fertilization, when the synergid that resides adjacent to the newly fertilized egg deteriorates as does the pollen tube that delivers the sperm cells to the unfertilized egg.

Historical Perspective
The idea that a process that leads to the death of a plant or its pa11s is under endogenous control is a relatively new concept. Molish in his classic work "The Longevity of Plants" (1938) was the first to clearly recognize that death might be actively induced by endogenous factors that are a natural part of an organism's development . Since this time the concept of senescence as an important developmental process has been gaining favor. More recent evidence indicates that senescence is an important developmental process in the life cycle of the plant .

Senescence versus Aging
Senescence represents an active endogenously controlled process that leads to death of the plant, tissue, or cell. This contrasts strongly with the concept of aging. Aging encompasses a wide array of passive rather than active processes that appear to be unregulated and driven primarily by exogenous factors (Leopold, 1978). Senescence is a developmental process that always leads to death, whereas aging does not necessarily result in death.

2.) Regulation of Senescence by Plant Hormones
Ethylene Ethylene 1s one of the simplest organic molecules with biological activity (Theologis, 1992). It has been implicated to play a major role as a fruit ripening hormone (Abeles, 1973). Ethylene is required for activation and maintenance of several genes involved in the process of fruit ripening and flower petal senescence Holdsworth et al. , 1987a;Holdsworth et al. , 1987b;Sato and Theologis, 1992;Theologis, 1992;Woodson, 1988;Yenofsky et al. 1988). Increases in respiration, inhibition of growth, modification of pigments, loss of geotropic sensitivity (Burg, 1962) and changes in gene expression have been observed in ethylene treated fruit Holdsworth et al.,1987a;Holdsworth et al. , 1987b). The biosynthetic pathway for ethylene production has been elucidated by Yang and his associates. The rate limiting step is catalyzed by 1aminocyclopropane-1-carboxylate synthase ( ACC synthase) and the final step (conversion of ACC to ethylene) is catalyzed by ACC oxidase (Sato and Theologis. , 1989). Genes for both ACC synthase and ACC oxidase have been cloned (Slater et al. , 1985;Hamilton et al. , 1991 ;Spanu et al. , 1991).

Cytokinins
Cytokinins represent a class of plant hormones that play an important role in the regulation of plant senescence. Structurally, cytokinins are adenine derivatives with various N6-substituants . Unlike ethylene, the biosynthetic pathway for cytokinin production has not been elucidated . Although a lack of information exists regarding how cytokinins are synthesized, a great deal of information is known about the physiological effects and sites of production Short et al. , 1972). Generally cytokinins are associated with senescence deferring properties. Specifically, cytokinins are associated with maintenance of chlorophyll, protein, and RNA levels (Richmond and Lang, 1957). Molish (1938) has shown that root formation greatly increases the life of excised leaves . Kulaveva (1962) reported that cytokinin-like senescence retarding factors were produced by roots. Today it is widely accepted that roots are the major site of cytokinin production and that cytokinins are plant hormones that act to delay senescence.

Other Hormones
Other plant hormones have been implicated in the process of senescence, but few generalizations can be made about their effect on senescence. This is largely due to the fact that the other hormones have vastly different effects when applied to different tissues or on the same tissues at different stages of development. An example of the difficulty in assigning discrete roles to many hormones involved in senescence is illustrated by the naming of abscisic acid by its identification as an abscission-inducing factor; however, this is not true in many species (Klee and Ramano, 1994). In the case of auxin there are reports that applications of a-naphthalene acetic acid retards the senescence of Rhoeo and Mesembryanthemum leaf segments (Sacher, 1959). Auxin is also known to stimulate production of ethylene in concentrations only slightly above physiological levels and ethylene promotes senescence in some species (Smart, 1994).

3.) Model Systems for Use in the Molecular Genetic Analysis of Senescence
Many model systems have been used for molecular genetic analysis of senescence in leaves, fruits, and flower petals Lawton, 1989;Skadsen and Cherry, 1983;Sort and Torrey, 1972). Up-regulation, down-regulation and homeostasis of particular gene expression patterns have been reported in all these systems . Ethylene is an apparent trigger for mRNAs involved in senescence in carnation petals and various fruits Marek and Stewart, 1992;Theologis, 1992;Yenofsky, 1988). The function, if any, of any down-regulated genes during senescence remains unclear.

Fruit Ripening
Ethylene has been implicated as a maJor fruit ripening hormone. Increases in respiration, inhibition of growth, modification of pigments, and loss of geotropic sensitivity (Burg, 1962) and changes in gene expression have reported for ethylene treated fruit Holdsworth, et al. , 1987a;Holdsworth, et al. , 1987b ).
Ethylene is required for activation and maintenance of several genes involved in the process of fruit ripening . Several up-regulated mRNAs have been identified that encode cell-wall softening enzymes, including polygalacturonase from tomato (Slater, 1985) and cellulase from avocado (Christofferson,l 984). The rate-limiting enzyme in ethylene biosynthesis is 1 aminocyclopropane-1carboxylic acid (ACC) synthase, and the final step for the conversion of ACC to ethylene is catalyzed by ACC oxidase.

Flower Petals
Carnation petals have been used as a model system for the study of senescence Lawton, 1989;Peterman and Siedow, 1985;Woodson, 1988), which is associated with changes in cellular · biochemistry and physiology. These changes have also been shown to be regulated in part by ethylene (Lawton, et al. , 1989;Short, et al, 1988;. cDNA clones of ACC synthase, ACC oxidase, and glutathione-stransferase have been isolated from senescing carnation petals and should be useful in elucidating the regulatory mechanism of ethylene action (Mayer, et al, 1991 ;Peterman and Siedow, 1985;.

Leaves
Leaf senescence represents a programmed, highly coordinated degenerative process that results in the breakdown of cellular components and sequestering of the resultant metabolites to other plant organs. Recent evidence indicates that the ethylene signal in leaves is transduced via protein phosphorylation events and is mediated by specific kinases and phosphatases, however, such evidence does not exist for fruit .
Ultrastructural analysis reveals that the degreening of mesophyll cells associated with a loss of thylakoid structures and increases in plastoglobuli are characteristics of the senescence syndrome in leaves (Thomson and Plat-Aloia, 1987). Coupled to the disruption of the thylakoid membrane is a decrease in photosynthesis due to a decrease in ribulose-l,5,-bisphosphate carboxylase/oxygenase and noncyclic electron transport (Bate, et al. , l 990;Start et al. , 1986;Thomson and Plat-Aloia, 1987).
Similarities between leaf senescence and fiuit ripening activities have been reported, including breakdown of the photosynthetic apparatus, degradation of starch and chlorophyll, stimulation of respiration, in some cases a burst of ethylene production (Aharoni and Liberman, 1979a;Aharoni and Liberman, 1979b;Thimann, 1980;Stoddart, 1980), and expression of at least some related genes .
Although leaf senescence is similar to fiuit ripening, unique aspects have been identified (Aharoni and Liberman, 1979a;Aharoni and Liberman, 1979b;. In soybean, the signals that are produced by developing seeds appear to induce monocarpic whole plant senescence (whole plant senescence that results after a single reproductive phase) (Nooden and Leopold, 1988), however, in Arabidopsis, somatic tissue longevity does not seem to influenced by reproductive development . Leaf senescence is the least well characterized of all senescence processes. Traditionally, studies of leaf senescence relied on exogenous application of hormones or hormone-like substances. Although these studies have provided information about the effects of various hormones, they have provided little information about the hormonal regulation or mechanism of action.
We intend to use the soybean cotyledon model to examine leaf senescence. In the soybean cotyledon system, the organs being studied remain attached thoughout the senescence process. Detached leaves and leaf disks have been used for the study of leaf senescence . Although these systems allow for precise control of experimental parameters of the tissue being studied, care should be taken when attempting to extrapolate the data obtained from an artificial system back to intact tissues.
Because senescence is essentially a process of breakdown and remobilization, disruption of source/ sink relationships has been shown to disrupt the normal developmental profile of senescence and therefore may change patterns of gene expression .

Soybean Cotyledons as a Model System for Investigations of Senescence and the Reversal of Senescence
It has been demonstrated that senescence in soybean cotyledons is reversible up to a critical point by epicotyl removal . This "point of no return" (PONR) is defined as the latest developmental stage at which decapitation results in I 00% regeneration of cotyledons (14 days). Cotyledons that regenerate are termed rejuvenated.
Decapitation later than the PONR (i.e. , 18, 19 days) results in a lower frequency of cotyledon survival. Rejuvenated cotyledons restore DNA, total nucleic acid, and soluble protein to 5 day levels after losses of 25%, 90%, and 80%, respectively. Epicotyl removal results in a burst of nucleic acid synthesis, followed by synthesis of chlorophyll and protein.
Evidence exists that senescence and rejuvenation (the reversal of senescence) are associated with changes in gene expression and several unique mRNA transcripts have been observed (Benhard and Matile, 1994;Skadsen and Cherry, 1983;Smart, et. al. , 1995), although these products have not been fully characterized. The nature of the signals and mechanism by which they modulate these changes are largely unknown.
The soybean cotyledon system should prove to be a valuable tool for investigations of gene expression and regulation during senescence. The reversibility of this process offers a unique opportunity for our investigations in that many genes that appear to be upregulated during senescence should be down-regulated during rejuvenation. In addition, soybean cotyledons are excellent tissues for senescence studies because they senesce within 24 days of planting . This eliminates the need for the more conventional practice of detachment, which is also used to induce senescence (Theologis, 1992) but may induce non-senscence processes such as a wound response.

c. Rationale and Significance
Senescence represents a unique phase of plant development eventually leading to death of the plant. Currently, the regulatory mechanisms that facilitate entry into this process are unknown. A better understanding of senescence should provide workers with the information necessary to manipulate this process to avoid current problems with postharvest and post-production losses as well as to improve crop yields. The physiological changes that occur during whole plant and cellular senescence have been well characterized , but only recently has information become available concerning the molecular changes associated with senescing fiuit, flower petals, and leaves Holdsworth, et al. , l 987a;Holdsworth, et al. , 1987b;Lawton, et al. , 1989;Petermann and Siedow, 1985;Sato and Theologis, 1992;Van Der Straeten, et al. , 1990;. Genes that exhibit up-regulation during senescence have been isolated and characterized Holdsworth, et al. , 1987a;Holdsworth, et al. , 1987b;Lawton, et al. , 1989;Petermann and Siedow, 1985;Sato and Theologis, 1992;Van Der Straeten, et al. , 1990;, and all evidence implicates hormonal regulation of these genes. The process of leaf senescence is not well characterized at the molecular level. We believe that the soybean cotyledon system can be exploited to achieve a better understanding of the genetic control over this unique developmental stage. In addition, this system can be used to investigate the process of rejuvenation, or reversal of senescence, which cannot be addressed in other model systems. Experiments have been designed to isolate specific clones that are upregulated during both senescence and rejuvenation. We feel that these clones will ultimately be useful in identifying and characterizing the regulatory mechanisms that induces this gene expression though the analysis of the promotor regions of genomic copies of the genes.
To this end, Northern blots have been used to monitor mRNA fluctuations during various stages of senescence and rejuvenation. A collection of both homologous and heterologous clones that are known or suspected to be involved in the two developmental phases has been amassed and has been used as gene-specific probes. These experiments have resulted in developmental profiles of expression of particular genes during senescence and rejuvenation in the soybean cotyledon system. This data has provided a detailed characterization of our system and in turn has validated it as a unique model system for the study of leaf senescence. A cDNA clone, pSIC3e, homologous to the showed strong differential expression in the soybean cotyledon system. Sag2 showed a strong increase in expression up to the PONR, a decrease to presenescent levels on day 14+2, and increase to senescent levels on day 14+6. CARACC3 showed at least five distinct bands that increased to the PONR and decreased to pre-senescent levels on 14+6.
rbcL showed a strong decrease in expression up to PONR and an increase to presenescent on 14+6. pVSP 25 and pVSP 27 both showed a strong decrease to the 8 days, a very strong increase on day 12, followed by a decrease on PONR, and another increase on day 14 +6. These clones will be valuable in the identification of homologous clones that will be used to study their regulation.

INTRODUCTION
Plant leaf senescence involves a series of biochemical and physiological changes that comprise the final stage of development from the mature fully expanded state until death (Smart, et al. 1995). It is a process by which proteins, nucleic acids, and membranes are broken down and the resultant metabolites exported to the developing seeds, storage tissues or other non-senescing tissues. Visible signs of leaf senescence include a preferential degradation of chlorophyll compared to other pigments, which results in a yellowing of the leaf tissue (Matile et al. 1989). During leaf senescence, the photosynthetic apparatus is dismantled and the nutrients are exported to other tissues. Leaf senescence can be induced by a variety of factors, including extremes in temperature, mineral deficiencies, drought, pathogen attack, and shading (Nooden and Leopold, 1988).
Molecular studies indicate that leaf senescence is associated with changes in gene expression Lawton et al. 1989;Mayer et al. 1991 ;Woodson, 1987;Woodson, 1988).
Experimental evidence indicates that subsets of genes are up-regulated while others appear to be down-regulated (Woodson, 1987;. mRNAs for glutathione s-transferase, ACC synthase, cellulase, lipoxygenase, and polygalacturonase have been shown to increase, whereas mRNA for ribulose-1 ,5-bisphosphate carboxylase/oxygenase activase has been shown to decrease during senescence and/or ripening Meyer, et.al. 1991 ;. Genes that show differential patterns of expression during leaf or petal senescence have been cloned from a variety of species including maize, Brassica, Arabidopsis, soybean, and carnation Smart et al. 1995; It has been shown that senescence in soybean cotyledons is reversible, up to a critical point, by epicotyl removal . This "point of no return" (PONR) is defined as the latest developmental stage at which decapitation results in 100% regeneration (regreening) of cotyledons (14 days). Cotyledons that regenerate are termed "rejuvenated". Decapitation later than the PONR (i.e., 18, 19 days) results in a lower frequency of cotyledon survival. Rejuvenated cotyledons restore DNA, total nucleic acid, and soluble protein to 5 day levels after losses of 25%, 90%, and 80%, respectively (Krul,l974). Epicotyl removal results in a burst of nucleic acid synthesis, followed by synthesis of chlorophyll and protein.
Soybean cotyledons are well suited for the molecular genetic study of senescence.
Relatively large quantities of tissue from all developmental stages can be grown and harvested in 20 days, which makes correlative physiological and biochemical studies possible that are not feasible with other model systems such as Arabidopsis. Because the plants are not fully developed and are not large, only a small growth chamber is needed to carry out an experiment. In addition, the tissue being analyzed remains attached to the plant so natural source/sink relationships remain undisturbed during the senescence process. Finally, the reversal of senescence (i .e. rejuvenation) can be analyzed, which is not possible in other model systems.
Although much information is known about gross changes in nucleic acids, proteins, and gene expression during senescence, relatively little is known about the molecular mechanisms that trigger the senescence syndrome or how it is regulated.
Furthermore, virtually no molecular information (either gross or specific changes) is available for rejuvenated tissues. We investigated specific changes in gene expression during senescence and rejuvenation in soybean cotyledons as a first step towards gleaning more detailed information about the molecular events that catalyze these important developmental phases. These studies have resulted in developmental profiles of expression for specific genes during senescence and rejuvenation. These profiles will be useful for determining precise stages of senescence and rejuvenation by providing investigators with genes that can be used as molecular tags. In addition, these genes are candidates for future investigations ofregulatory regions (e.g., promotor regions and consensus sequences) that may be important for recognition by specific endogenous signals that induce the initial events of senescence and rejuvenation. Total and mRNA samples were re-suspended in diethylpolycarbonate (DEPC) treated water and stored at -70°C.

MATERIALS AND METHODS
Northern Blot Analysis-Northern analysis   coli strain XLl-Blue (Stratagene Inc., La Jolla, CA) using a standard heat shock protocol (Sambrook, et al, 1988) or electroporation (Gene Pulser II, Biorad Corporation) into E.coli strain XLl-Blue. Transformed cells were grown in large or small scale and plasmid DNA was purified using standard protocols (Sambrook J, et. al. 1988). CsCI gradients were used to purify plasmid DNA that was prepared on a large scale. Purified plasmid samples were resuspended in sterile distilled water, quantitated on a fluorometer, and stored at -20°C. Plasmid identity was verified by restriction digest analysis. Plasmid DNA was digested using an appropriate restriction enzyme to release the insert, run on a 1 % TAE-LMP agarose gel and the insert band was excised and purified from the LMP agarose using GELase™Agarose Gel-Digestion Preparation (Epicentre Technologies, Madison, WI) according to the manufacturer's recommendations. Purified insert DNAs were resuspended in sterile distilled water, quantitated on a fluorometer, and used as probes for Northern analysis. Table 1.
Cloned genes that were used as gene-specific probes against Northern blots of total RNA from different developmental stages of soybean cotyledons. "Plasmid" indicates the cloning vector for the insert. "Organism" indicates the type of organism the clone was recovered from . "Source" indicates the original source of the clones. All clones were recovered from cDNA libraries. "Function" is listed if known. "Cross-reacts" indicates whether the clone showed any hybridization when used as a probe against soybean cotyledon Northern blots. "Diff. Expressed" indicates whether the clone showed differential expression when used as a gene specific clone against the soybean cotyledon Northern blots. or rejuvenation were used as gene specific probes to assay changes in gene expression in the soybean cotyledon system (Table 1 ). Of the 17 clones used as probes on Northern blots, 6 showed some type of differential gene expression. Three of these clones displayed an expression pattern consistent with our predicted developmental pattern based on the known function of each gene-specific probe (Table 1 ).
CARACC3 , which encodes a carnation ACC synthase, was used as a probe on the Northern blots (Figure 1). At least five bands of varying molecular weight hybridized with this probe and all showed an increase in abundance from 5 to 14 days and a subsequent decrease through 14+6 days. rbcL served as a control that indicated a increase or decrease in chloroplast populations. rbcL encodes the large subunit of soybean rubisco 1,5 bisphosphate oxygenase, carboxylase. This clone hybridized to two bands ( Figure 2). The higher molecular weight band was strongest on 5 day, disappeared completely on 8, 12, and 14 day, then reappeared with increasing intensity in 14+2 and 14+4 day, and was barely visible in 14+6 day samples. The smaller band hybridized most strongly on 5 day followed by decreasing intensity through 14 day at which point it was not visible. A reappearance of the band was observed with increasing intensity in 14+2 and 14+4 day RNA samples followed by a virtual disappearance again in the 14+6 day sample. Sag2 is a senescence-related clone from Arabidopsis thaliana that encodes a cysteine protease.  . When used as a probe, it hybridized with increasing intensity from 5 to 14 day with approximately equivalent hybridization in 12 and 14 day samples hybridization levels decreased to 5 day levels in 14+2 day samples and then increased again in 14+4 and 14+6 day samples (Figure 3). VSP 25 and VSP 27 have been shown to encode vegetative storage proteins (Staswick, 1988). When used as gene-specific probes against Northern blots made from the developing soybean cotyledon, both VSP's were down-regulated in 5-8 day post germinated cotyledons, increased in 12 day cotyledons and decreased again in senescing cotyledons. Upon decapitation they were again upregulated (Figures 4 and 5). Ubiquitin is a protein involved in normal turnover of cellular proteins. When pUBS6, a soybean ubiquitin clone, was used as a gene specific probe against Northern blots made from the developing cotyledon system it hybridized with increasing intensity to a maximum value at 14 days, then decreased to presenescent levels on 14+6 days ( Figure 6). pSIC3e and pSIClOc are homologs of Sag2 that were cloned from senescing soybean cotyledons (see Manuscript II). As expected, they both exhibited similar patterns of expression to Sag2 (Figure 7).  is a major part of the senescence process (Smart et al. , 1995). However, during the later stages of rejuvenation a second increase in Sag2 mRNA was seen. This observation may be explained by the fact that rejuvenation in the soybean cotyledon system was defined morphologically and by the color of the cotyledons. It seems reasonable that proteins that are not essential to the metabolism of the cotyledon would be degraded before the chloroplasts and associated energy generating components because senescence requires energy to proceed. Alternatively, the cotyledons from 14+4 and 14+6 day stages could be far beyond the rejuvenated stage and could be actively senescing again. The former hypothesis is supported by the results of the Northern blot which was probed with rbcL.
rbcL encodes the rubisco 1,5 bis-phosphate oxygenase/carboxylase large subunit, which is a major enzyme involved in photosynthesis. The pattern of expression rbcL is consistent with the morphological observations that were used to discern various stages of senescence and rejuvenation in that we observed a decline in rbcL mRNA during senescence and an increase during rejuvenation.
Northern blots probed with CARACC3 resulted in at least five hybridizing bands that increased during senescence and decreased during rejuvenation. The major pathway for ethylene biosynthesis in plants is the Yang cycle (Yang and Hoffman, 1984).
CARACC3 encodes a carnation ACC synthase. This enzyme catalyzes the rate limiting step in ethylene biosynthesis and is known to have multiple isoforms (Wang and Woodson,19 91). Ethylene is a major hormone involved in fruit ripening (Theologis, 1992). Fruit ri . perung and senescence are closely related processes and some genes have been found that show sequence similarity ; therefore, it is not unexpected that ACC synthase appears to be differentially regulated during senescence and rejuvenation in the soybean cotyledon. The appearance of multiple bands suggests that ACC synthase is an important and highly regulated enzyme during senescence and rejuvenation in soybean cotyledons, and therefore ethylene may play a key role in the regulation of these processes. Experiments to discern the role of ethylene on senescence and rejuvenation in soybean cotyledons could be carried out using ethylene biosynthesis mutants. Northern blots could be used to evaluate differential gene expression of ACC synthase and other key enzymes involved in ethylene biosynthesis.
The VSP (vegetative storage proteins) proteins have been the subject of much investigation (Staswick, 1988;Staswick, 1989;Staswick, 1990). These vegetative storage proteins appear to be important as temporary storage molecules for nitrogen (amino acids) in vegetative tissues (leaves). They are regulated by multiple stimuli (Staswick, 1989;Staswick, 1990), but most notable is their up-regulation in depodded soybean plants which exhibit delayed senescence. Northern blots probed with p VSP 25 and p VSP 27 showed a transient increase just prior to senescence that was initially puzzling. After repeated analysis of several Northern blots the possibility of unequal loading was eliminated. Work with soybean leaves has shown similar transient up-regulation just prior to senescence (Staswick, 1988). This unique pattern of expression suggests a high level of regulation in the cotyledons of soybean.
pUBS6 is a soybean ubiquitin. Ubiquitin is involved in protein turnover. When pUBS6 was used as a gene specific probe against the soybean cotyledon system we observed an increase in intensity up to 14 days, followed by a return to presenescent levels on 14+6 days ( Figure 6). These observations agree with our predicted pattern of expression because senescence is a process that involves turnover of nutrients and subsequent export to other organs and/or tissues.

INTRODUCTION
Leaf senescence is an important developmental phase in which cells undergo distinct metabolic and structural changes prior to cell death (Nooden and Leopold, 1988).
It is a process by which proteins, nucleic acids, and membranes are broken down and the resultant metabolites shipped to the developing seeds or storage organs. Visible signs of leaf senescence include loss of chlorophyll and a degeneration of the photosynthetic apparatus, which results in a yellowing of the leaf tissue. Leaf senescence can be induced by a variety of factors including extremes in temperature, mineral deficiencies, drought, pathogen attack, and shading (Nooden and Leopold, 1988).
Molecular studies indicate that leaf senescence is associated with changes in gene expression Woodson, 1988;. Experimental evidence indicates that subsets of genes are upregulated while others appear to be down regulated (Woodson, 1987;. mRNAs for glutathione s-transferase, ACC synthase, cellulase, lipoxygenase, and polygalacturonase have been shown to increase, whereas mRNAs for ribulose-1 ,5-bisphosphate carboxylase/oxygenase activase has been shown to increase during senescence and/or ripening Meyer et.al. 1991 ;.
Although much information is known about gross changes in nucleic acids, proteins, and gene expression during senescence, relatively little is known about the molecular mechanisms that trigger the senescence syndrome or how it is regulated.
Furthermore, virtually no molecular information (either gross or specific changes) is available for rejuvenated tissues. We investigated specific changes in gene expression during senescence and rejuvenation in soybean cotyledons as a first step towards gleaning more detailed information about the molecular events that catalyze these important developmental phases. We have identified a senescence-associated Arabidopsis Sag2 gene that shows differential expression during senescence and rejuvenation m soybean cotyledons and have cloned the soybean homologue. This clone is a candidate for future investigations of important regulatory regions (i .e., consensus sequences) that are involved in the coordination of senescence and rejuvenation. coli strain XL 1-Blue using a standard heat shock protocol (Sambrook J, et. al. 1988) or electroporation (Biorad Corporation) into E.coli strain XLl-Blue. Transformed cells were grown large scale and small scale and plasmid DNA was purified using standard protocols (Sambrook J, et. al. 1988). CsCl gradients were used to purify plasmid DNA that was prepared large scale. Purified plasmid samples were re-suspended in sterile distilled water, quantitated on a tlourometer, and stored at -20°C. Plasmid identity was verified by restriction digest analysis. Plasmid DNA was digested using an appropriate restriction enzyme to release the insert, run on a 1 % T AE-Ll\.1P agarose gel and the insert band was excised and purified from the Ll\.1P agarose using GELase™ Agarose Gel-Digestion Products, Boston MA) by capillary diffusion in 10 X SSC. The membrane was fixed by UV crosslinking and probed with Sag2 that was purified and radiolabelled as described above. Ten samples containing at least one large molecular weight band that also hybridized strongly were selected for a tertiary screen.

Tertiary Screen
The second half of the elutions from each of the ten positive sample were plated onto fresh NZY plates at a lower titer and five plaques from each elution for a total of 50 were selected at random and eluted in 50 µI SM buffer overnight at 4°C and PCR amplified as described above. Single bands were observed in most samples when a 20 µl aliquot of each reaction was electrophoresed on a 1.0% T AE agarose gel at 60 volts for 2.5 hours.
Clones were selected for a quaternary screen based on a comparison of the size of the band on the tertiary screen blot to the size of the putative positive insert on the secondary screen.

Quaternary Screen
Single putative positive clones were excised and eluted in 50ul SM buffer overnight at 4°C and PCR amplified as described above. A 20 µl aliquot was run an a 1. 0% T AE agarose gel, transferred to membrane and hybridized to Sag2 as described above. Samples that hybridized strongly were selected for confirmation on Northern blots.

Northern Confirmation
Total RNA from 5, 8, 12, 14, 14+2, 14+4, and 14+6 day old cotyledons was extracted as described above. Northern analysis ) was carried out using 1 . 5 % agarose-formaldehyde gels. 20 µg of each sample were transferred to positively charged nylon Restriction products were run on a 1.0% T AE agarose gel, stained with eithidium bromide and visualized on a UV trans illurninator.

cDNA screening
From a total of 100,000 primary screen plaques, thirty showed strong hybridization when probed with Sag2 and were selected for a secondary screen (results not shown). For a secondary screen these plaques were eluted and half of each elution was amplified by PCR and separated on a 1.0% agarose gel where multiple bands (0 to 3 bands) were visible (an average of 1.7 bands per core) (Figures la, and le). Upon transfer to a positively charged nylon membrane and probing with Sag2, ten of the thirty samples were seen to contain high molecular weight bands that hybridized strongly; these were selected for a tertiary screen (Figures lb and ld). The ten positive samples from the secondary screen were replated at a low titer and five random isolates were selected from each sample. From a total of fifty plaques that were amplified by PCR, five random isolates from each of the ten samples from the secondary screen, twenty contained single bands that were the same molecular weight as the corresponding bands that hybridized strongly in the secondary screen (Figures 2a and 2b) and were chosen for a quaternary screen.
Eleven of these twenty hybridized strongly to Sag2 (Figures 3a and 3b).

Northern blot analysis
Two clones, pSIC 3e and pSICIOc, were selected for confirmation by Northern analysis.
Both clones showed similar patterns of expression compared to the original Northern blot probed with Sag2. pSIC3e and pSIC 1 Oc were not expressed on day 5, showed a strong increase in expression to a maximum level on day 14, followed by a dramatic decrease to five day levels on day 14+2 and another increase to day 14+6 (Figures 4a-c).

Restriction analysis
When restriction analysis was carried on pSIC3e and pSIC l Oc, both showed identical restriction patterns to Sag2 . Of the eight restriction enzymes used, only Pvull cut all three clones, which contained two restriction sites (results not shown). pSIC 3e was estimated to be I .4 kb and pSIC 1 Oc was estimated to be I . I kb (Figures 5a and 5b ).  from Arabidopsis thaliana. is a senescence-related clone that codes for a cysteine protease . Cysteine proteases have been implicated in all senescing model systems. Sag2 has been shown to be differentially regulated in our soybean cotyledon system. The gene is preferentially up-regulated in expression during cotyledon senescence ( Figure 4). Availability of Sag2 homologous gene sequences from soybean will allow for comparisons between Arabidopsis and soybean and better assess the function of this gene during senescence. The cDNA clones, pSIC3e and pSICIOc, are currently being sequenced and used to recover an intact genomic clone from soybean for future studies on the nature of the structure of the gene and its' associated regulatory sequences. An additional benefit of the soybean cotyledon system is that the process of rejuvenation or the reversal of senescence can be investigated. This process is interesting because at the molecular level it does not represent a simple reversal of senescence, but instead a unique developmental process. We have shown that Sag2 is up-regulated during senescence as well as during rejuvenation.
Sag2 and its' soybean homologues are rapidly down-regulated upon decapitation and then gradually accumulate again (Figure 4). Because this gene is so highly regulated it should be an interesting and important gene to study. Further experiments have been completed that show that Sag2 behaves in soybean leaf tissues as in cotyledons (results not shown). Patterns of expression can now be assessed in systems where the normal senescence program has been modified. Male sterile lines (ms 1 and ms 2 ) and depodded wild type soybean that do not exhibit senescence or abscission would provide two interesting systems. In addition, evergreen mutants (d 1 d 2 , G, and Gd 1 d 2 ) that do not senesce, but abscise normally would also be candidates to assay the expression of Sag2 and its' soybean homologues. These experiments should provide insight into the regulation of senescence. In addition, comparison of the evergreen mutant patterns of expression against the patterns of expression of depodded or male sterile lines should help define the regulatory differences between senescence and abscission.

GENERAL DISCUSSION
Senescence is an important biological process that is characterized by a sequence of biochemical and physiological changes that comprise the final stage of development of a plant. Plant cells undergoing senescence remain viable and metabolically active (Thomas and Stoddart, 1980). Evidence exists to suggest that these changes are in part preprogrammed and under genetic control, however little is known about the molecular mechanisms that control this process. A better understanding of this process will be important for studies relating to: 1) normal cellular differentiation, maintenance, and turnover; 2) plant defense mechanism and the hypersensitive response; and 3) embryogenesis and development. In addition, economic benefits could result from a better understanding of senescence through reduction of post harvest and post production losses as well as an improvement of crop yields.
Many models have been proposed to explain the role of the factors that may initiate senescence. It is well known that many different stimuli or "factors" can trigger the senescence process; it is thought that all these share a common molecular pathway (Smart 1994). These "factors" can be separated into two broad categories; namely, internal factors and external factors. Internal factors include hormones or growth regulators, reproduction, and shading. External factors include temperature, water, pathogens, and minerals. Internal and external factors are thought to share some common control steps in the initiation of the senescence program. The control steps then affect four broad categories of genes that include; 1) catabolic enzymes (pSIC clones are included in this group); 2) stress related proteins; 3) Nutrient remobilization (VSP25 and VSP27 are in this group); and 4) genes involved in abscission. By identifying genes that are up-regulated in expression during senescence, one can then use them as probes to elucidate the signals that regulated the senescence program.
Soybean has been used as a model system to study senescence (Chang et. al. 1979;Huber and Newman 1975;Marek and Stewart 1992;Peterman and Siedow 1985;Skadsen and Cherry 1983).
Soybean cotyledons proceed though a process of senescence approximately 14 days after germination. This process has been shown to be reversible by decapitation , therefore this system offers the unique ability to study the reversal of senescence or "rejuvenation" which is not possible in other model systems such as Arabidopsis, maize, carnation petals and others. In addition, soybeans are an agriculturally important crop. Furthermore, many soybean lines that have altered patterns of senescence are available. Isogenic evergreen mutants Harosoy with different combinations of single, double, and triple mutants involving the G, di> and d 2 genes are available. Also male sterile lines, ms 1 and ms 2 are available that do not senescence or have leaves that abscise. Finally, depodded wild type soybeans exhibit no senescence or abscission. The use of these mutant and wild type lines allow the processes of senescence and abscission to be examined and studied at the molecular level. The pSIC clones would be candidates for Northern analysis different stages of development during a normal senescence time frame in the lines.
We have used a collection of previously cloned genes to assay their expression during senescence and "rejuvenation" in soybean cotyledons. Different classes of genes were identified as being differentially regulated, these include: I) catabolic genes such as cysteine protease and ubiquitin ; 2) nutrient remobilizing genes i.e., vegetative storage proteins and ; 3) A regulatory gene ACC synthase. These genes serve as molecular markers and can be used to more precisely identify initial stages of senescence and "rejuvenation". It is at the early stages the we expect the regulatory events that control these processes to occur.
In addition, we have isolated a soybean homologue of Sag2. We can now begin to characterize the soybean gene structure and regulatory sequences. Ultimately it will be interesting to compare these results with the original Sag2 gene.
Future work to investigate the expression of the SIC clones at the whole plant level have begun. Experiments have been carried out that show that the pSIC clones appear to be up-regulated in senescence soybean leaf tissue (results not shown). This data suggests that conclusions that are drawn from the soybean cotyledon system may be applicable to leaf and possibly whole plant senescence. Currently work in has begun to sequence the pSIC clones and recover full length genomic copies. This will provide valuable information about regulatory sequences.