The Adaptive Reasons for and The Physiological
Causes of Senescence in Annual Plants
For annual plants, yearly senescence can be explained as a way of increasing the number of progeny surviving to reproductive age and/or a way of increasing the amount of turnover and thus the rate of evolution. The first explanation is probably more important than the second. Higher numbers of viable progeny are produced by a redistribution of nutrients from the plant to the seed during senescence and several less important reasons. On the level of plant physiology, with many or all species, evidence and theory point to senescence as a two step process. The first step involves only a breakdown of RUBISCO protein and the export of resulting amino acids to the seeds. The second phase is a general breakdown of the cells. On the level of the cell, general cell breakdown is possibly produced by a decline in Calmodulin activity, caused by a decline in polyamine levels, which is in turn caused by a decline in IAA and Cytokinin in the senescing cells.
When we think about death in living things we automatically think about death in human beings, which typically follows a period of progressive aging. However this appears not to be the way annual plants (plants that usually live only one year) die. Whole populations of these plants will die en masse at a set time of year, within the span of a few days, before bad weather has hit (Leopold 1975). Additionally, although aging would seem to be occurring, these species can easily be made to live much longer than they normally would by keeping them in a moonflower inducing photoperiod. As we shall see later this and many other pieces of evidence point to the idea that annual plants kill themselves.
Two broad questions will be taken up in this paper: Why annual plants kill themselves and how they do so. The question of why annual plants kill themselves can be split into two parts: the long run advantages (greater population growth in certain niches than perennial competitors or a higher rate of evolution than is achieved by the perennial strategy) and the short run advantages (e.g., the generation of larger numbers of seeds by the routing of nutrients from the adult plant to the seeds). These are evolutionary and ecological issues.
The question of how annual plants kill themselves (i.e., what goes on physiologically during senescence) can also be split into two parts; the mechanisms of senescence on the whole plant level and the mechanisms of senescence at the cell level. These are respectively the issues of whole plant physiology and cell physiology.
In exploring the topic of senescence, we shall find that the above questions are interrelated, the answers to some questions suggesting answers to others. Many of the points in the paper will be illustrated by experiments done on soybeans, and a model of soybean senescence will be presented toward the end of the paper. However, many of the conclusions reached in this model are applicable to other unrelated annuals. The paper will end with an outline of a series of experiments designed to find the solution to the plant physiology question of what causes the decline in photosynthesis during senescence.
Long Run Advantages of the
There appear to be two reasons why an evolving plant species might "choose" to be annual over perennial. One is that the annual strategy allows the plant a higher population growth in some niches than the perennial strategy. Most evolutionary competition models hypothesize that, when there are several species in a common niche, the species which has a greater rate of increase will "push" the other species into a new niche, or into extinction. The other possible reason is that to a plant with low allelic variation annual senescence may be a way of increasing overall genetic variation. This is because having the parent plant die each year prevents the same genotype from producing offspring year after year.
Cole (1954) gave an erroneous perspective on the question of population growth rate of annuals and perennials when he said, "For an annual species, the absolute gain in intrinsic population growth that can be achieved by changing to the perennial reproductive habit, would exactly be equivalent to adding one more individual to the litter size."
In mathematical terms this is:
N(T + l) = BaCN(T)
Equation Number 1. for annuals where:
N(T) = the size of the population at time T
Ba = the number of seeds produced by a mature plant
C = the percentage chance of a seed reaching maturity
N(T - 1) = (Ba + l)CN(T) = BaCN(T) + CN(T)
Equation Number 2. for perennials
This sort of analysis leads one to question why the annual strategy should be advantageous at all because (Ba + l)CN(T) will always be greater than BaCN(T). Hence the rate of population growth should always be higher for the perennial and the perennials should win out in all niches they share with annuals. This analysis, if it were true, would mean there should be many more perennial than annual species and that perennial species would be evolutionarily more advanced. Clearly neither of these conclusions is true. There are large numbers of annual species, and many or most them seem to be evolutionarily derived from perennial species.
It appears then that the annual strategy is adaptively advantageous in many circumstances. Some insight as to where the Cole formulas are wrong can be gained by looking at experiments measuring energy allocation in the life histories of annual and perennial plants. Harper and Ogden (1970), and Brock (1980) and many others have found that annual plants devote more or much more of their energy to reproduction than do perennials. This allocation of energy means that annuals probably, on average, produce more seeds. Charnov and Schaffer (1973) rewrote the Cole equation with these points in mind:
N(T + l) = BaCN(T)
Equation Number 3. remains the same for annuals
N(T + l) = BpCN(T) + PN(T)
Equation Number 4. for perennials where:
P = the chance of the perennial overwintering
If a perennial and an annual species are to remain in competition, then their rates of increase should equal each other or:
BaC = BpC + P
Equation Number 5. which can be rewritten as:
Ba = Bp - I - P/C
Equation Number 6.
Equation 6 shows that annuals must produce more seeds than perennials in order to compete. That is they have to make up for the P/C term found in the perennial equation.
The Charnov-Schaffer formulas lead to an understanding of the conditions under which each type of plant should predominate. The formulas predict that annual plants should win out when BaOBpC+P, whereas perennials should win out when BpC+P>BaC. In many environments, P is very small due to the harshness of winter conditions (e.g. tundra). In these settings it is fruitless for a plant to put energy into P and take it away from reproduction, because there is such a small chance that the plant will survive to the next year no matter how much energy is put into survival. Hence annuals should predominate in this sort of environment and in any environment where it is difficult to survive the whole year (e.g. the desert and the forest floor). On the other hand, if a plant is in an environment where it is hard for seedlings to take root and grow (C is small) whereas it is relatively easy for a mature plant to continue living (P is large), perennials should predominate (Charnov and Schaffer, 1973). An example of the latter sort of environment is the deciduous forest.
As a footnote, the Charnov-Schaffer equations probably should be further refined. Thinking logically, the more energy put into reproduction, the greater is the chance of reproductive success. However this higher reproductive success can be due to more seeds being produced or to the seeds being hardier (i.e. because of more nutrient storage in the seeds). That is, the entire term BaC should be larger than BpC when a species "decides" to use the annual strategy. It would not matter whether it was the B or C term or both which would be larger. This suggests that a fully adequate set of equations should read:
N(T + l) = BaCaN(T)
Equation Number 7. for annuals
N(T + l) = BpCpN(T) + PN(T)
Equation Number 8. for perennials
A different reason for the long run advantage of annual senescence might be that it produces a faster rate of turnover and thus a more rapid rate of evolution (Leopold, 1975). Because so many new plants are produced each year, annuals would be at an advantage over perennials when a new niche opened, because there would be more offspring to produce a mutant that could colonize the new niche.
If we look at the evolution of species from this viewpoint, in a complete turnaround from the Cole equations, annuals would seem to be at the greater advantage. Hence, turning the tables, one might wonder why there are so many perennial species. One possible reason for the success of perennials despite this disadvantage is that too high an evolutionary change rate produced by a high turnover can be harmful to a species. High turnover can destroy well adapted genotypes.
This suggests that annuals may actually try to keep their genetic variation, at each allele, down to a minimum (if it is possible to control the amount of allelic variation), because the variations get amplified so much with high turnover. Conversely, perennials may overcome the disadvantages of low turnover, by increasing the genetic variation at each allele (at the cost of even worse problems in maintaining successful genotypic coherence).
Mathematically, if the rates of evolution are to be equal and annuals and perennials are to compete equally, then:
Ea = GaBaCa = Rate of Evolution of Annuals
Equation Number 9. where Ga = Genetic Variation Coefficient
Ep = 6pBpCp = Rate of Evolution of Perennials
Equation Number 10.
Ea = Ep
Equation Number 11.
if they are to compete in evolution. Since BaCa is greater than BpCp, then Gp must be greater than 6a to make up the difference and thus satisfy Equation 11.
In a review article by Hamrick, Linhart, and Mitton (1979), it is reported that researchers have found single plant genetic variation as measured by the polymorphic index to be highest in woody perennials, next highest in annuals, still lower in herbaceous perennials and lowest in biennials. (The difference between the herbaceous perennials and the annuals was small and statistically insignificant.) In two other measures of gene variation (the percent alleles polymorphic, and the mean alleles per locus), annuals did have a statistically significant higher variation than did the herbaceous perennials, but they were still lower than the woody perennials. Also mentioned in the article, however, was the fact that longer generation times have, in general, been correlated with higher genetic variation.
The fact that biennials have the lowest variation in all three measurements of genetic variation mentioned above, is strange in the perspective of the genetic analysis just outlined. Biennials have turnover rates one half the size of annuals; hence, they should have twice the genetic variation. Clearly the reason for choosing the biennial strategy is not that it gives a higher evolutionary rate.
Overall the data only weakly support the notion that plants "choose" the annual strategy over the perennial because annual reproduction produces higher rates of evolution. However this interpretation that plants choose the annual strategy to get high turnover may be valid if we consider that the data deals with all annuals and perennials. Some annuals and perennials probably evolved their respective strategies because of the higher population growth these strategies afforded (as discussed above). Perhaps it is only in environments where neither the annual nor the perennial strategy is particularly favored in terms of growth rate that genetic variation per plant is clearly lower in annuals than perennials.
Moving on, we turn now to the next part of our question about why annuals senesce. The Charnov-Schaffer equations tell us that annual plants have a higher chance of reproductive success (BaCa) than do perennials in certain environments. This "tells us that the death of the annuals somehow increases the number or percentage chance of survival of the seeds. How this reproductive success is brought about is the question of the short run advantage of senescence. (For the sake of simplicity and space, the question of how the long run advantage of high evolutionary rate is brought about, will not be discussed.)
Advantages to Senescence
There appear to be four ways annual plants can improve their reproductive success by dying -i.e.. four physiological ecology reasons for the development of the annual strategy; a. The plants die in order that their biological nutrients and minerals may be transported to the seed, to enable its development (Sinclair and Dewitt, 1975). b. The plants die so that the seeds can get the biological nutrients more rapidly than would otherwise occur, so as to avoid frost (implied by Leopold, 1975). c. The plants die in order to leave space and fertilize the soil for the next generation (Pruitt, 1983). d. The plants die in order to insulate the seeds (with the dead leaves and stalks) against the ravages of winter (Pruitt, 1983).
For non-nitrogen fixing annuals such as wheat and oats, reason a seems to be particularly important. Seeds contain a very high amount of nitrogen. Non-legumes must absorb all of their nitrogen from the soil. Hence nitrogen is strictly limited and must be given up by the leaves if a large number of viable seeds are to be produced. This seems particularly true in light of some calculations made by Sheehy (1983) which show that legumes like soybeans could just, barely fill their seeds if they did not senesce and redistribute their nutrients. These calculations suggest that non-legumous annuals cannot absorb enough nitrogen during the seed development stage to fill the seeds and also keep the leaves alive.
The redistribution of nitrogen from mature leaves "to developing seeds involves (as we shall see later) a breakdown of proteins in adult leaves and the transport of amino acids to the seeds. (In fact there have been some researchers that have found that the level of protease activity is directly related to the protein content of the seed C. Perez, et al, 1973; Rao and Croy, 1971, 19723) It must be pointed out, in this context that killing the leaf and breaking down the protein involves a tradeoff for the seed. This is because nearly 507. of the protein in the leaf is made up of RUBISCO (Wittenbach, 1982). RUBISCO is Ribulose-Bisphosphate Carboxylase a very important protein of photosynthesis. This is often the first protein to decline during senescence (Wittenbach, 1982). Thus senescence in its early stages involves the stopping of photosynthesis. The seeds must then rely for pod-filling on the stored starch of the leaves, roots, and stems, which may be in short supply.
Soybeans and all legumes are different from the plants just discussed in that they manufacture their own nitrates and nitrites. Nevertheless, during senescence in soybeans, biological nutrients and minerals are redistributed to the seeds (Nooden, 1980). Thus either explanation a or b is an important part of the reason for senescence in soybeans. Such redistributions do not seem to be an absolute necessity, as is probably true in the non-nitrogen fixing species. As mentioned, calculations by Sheehy (1983) show that soybeans (and other legumes) are capable of filling their pods without senescence of the leaves, roots, and stems. Abu-Shakra et al (1978) in fact found five soybean plants from a hybrid cross which did not senesce during pod fill. The number and weight of the seeds from these plants were not diminished from levels usually expected in soybeans; in fact, they were somewhat larger than usual.
These findings suggest that explanation a is not correct for legumes. Hence we turn to explanation b, that senescence ensures more rapid transport of nutrients and minerals to the seeds than would otherwise occur. This points directly to a mechanism of frost avoidance. If the seeds were still in development (because of a longer development time produced by not senescing) and frost hit they would be killed.
The other two physiological ecology explanations for senescence apply to soybeans and many other annuals. Soybean seeds are contained in heavy pods that fall straight to the ground; hence the seeds if they germinate, grow right up where the parent plant was. If the parent plant did not die at all and lived to the next generation, the pods would have less of a chance for survival because of the lack of space and nutrients. Also the seeds would not be protected from the winter cold by the dead parent plant above it.
In contrast to soybeans, many other annuals such as wheat and oats have light, wind carried seeds. The local effects the parent plants have on fertilizing the soil and providing space, and the insulation provided by the dead plant, cannot 'be important here. Explanations c and d can only affect the survivability of seeds brought into the area of the plant from distances by the wind. Thus they would benefit non-progeny and often non-closely related plants. Unless we invoke the discredited notion of group selection, explanations c and d cannot be an important force in the adaptation of these kinds of plants.
Thus plants that have windblown seeds, and seeds which are consistently carried away from the immediate site of the parent plant (e.g. by herbivores) must senesce only because of reasons a or b. Those annuals whose seeds fall to the ground can senesce for any or all four reasons.
Questions of how senescence is brought about and what occurs during the process of senescence are questions of plant physiology and cell physiology, to which we now turn. Again, for the sake of simplicity and space, only considerations of how reasons a and b (the death and redistribution of minerals) are physiologically brought about, will be examined.
There appear to be three ways plants can senesce; a. They die simply because of aging; that is, accumulated entropy causes increased sensitivity to environmental stress, bringing on death. B. Simple nutrient withdrawal by the seeds from the leaves eventually brings on starvation or some other nutrient deficiency, causing death (Molisch, 1938). That is, the seeds take what nutrients they need by diffusion from the adjacent cells and this sets up a source sink diffusion gradient where the nutrients pass from the leaves to the seeds, with no active breakdown of the leaves. A hormonal or other kind of signal is released by the developing flower or seed, or the photo-induced leaves. This causes the plant to begin degrading itself and transporting the resulting nutrients out of the cells into the developing seed (Leopold, 1959).
Simple aging might be the reason for senescence in annuals. Yet this seems unlikely from an evolutionary viewpoint, because such a process would leave nutrients in the leaves, denying them to the next generation. If the annual is going to die anyway, why shouldn't it produce more or better seeds by the nutrient redistribution process. There may be some rare cases, where continued near peak function of the whole plant is necessary for the filling of seed pods. If starch is unobtainable from the breakdown of cellulose and other structures in the leaves, then the continued synthesis of starch may be necessary. Additionally if some mineral or biological molecule is needed strictly for seed development, then the whole plant might have to stay alive to absorb or synthesize the nutrient at the right time. However this is unlikely because there is no reason why the plants could not absorb or synthesize the special nutrient or nutrients early in life and then store it or them until the time of seed development.
Even before it was known that many annual species redistribute their nutrients during senescence, the notion that aging causes death was called into question. This was because, as mentioned in the first paragraph, many annual plants were noticed to die en masse within a span of a few days at the same time each year and before any bad weather had hit (Nooden, 1980). Death by aging always produces a bell shaped death curve, with a few dying young and a few dying old and most dying at an intermediate age.
Experiments by Molisch (1938), where he depodded plants from a number of species, confirmed that death is not caused by aging and suggested explanation b (senescence is caused by nutrient withdrawal). Molisch found that depodding treatments greatly delayed senescence. This suggested to him that the pods were killing the leaves by draining needed nutrients from the seeds. Careful work by Leopold (1959) confirmed these results in soybeans and showed that depodding delayed senescence in spinach as well, though "to much lesser extent. Once spinach plants had bolted, senescence was unstoppable.
Leopold first reasoned that the tremendous nutrient strain of bolting and not the comparatively small nutrient strain of the flower and fruit development, caused senescence. He tested this by treating non-photoinduced spinach plants with gibberel1 ins, which causes bolting but not flowering in plants. The results were that the Gibberellin treated plants lived almost as long as those that had not been flower induced, and much longer than those that had been flower induced. Hence, he noted, he was wrong (Leopold, 1961).
These results suggest that spinach undergoes senescence by scheme c. Mere nutrient drain is not enough to cause senescence. Additionally, since depodding only slightly delayed senescence, the hormone is probably made in the leaves in response to the flower inducing photoperiod. If the hormone was made in the flowers or pods, defloration or depodding would greatly delay senescence (Leopold, 1961).
Nooden and his coworker Lindoo greatly extended Leopold's work on soybean senescence. They first confirmed that depodding greatly delays leaf death in soybean (1978). Next they found that the cause of senescence can be traced not to the pod as a whole but to the seeds in the pod (1978). When the seeds were removed and the pods left on, senescence was greatly delayed.
Further, they found (1982) that if they prevented nutrient drain from the soybean leaf by heat killing a small cross section of the petiolar phloem while keeping the xylem intact, the plants still senesced. Again here also (with the dead phloem) the depodded plants had a delayed senescence. These experiments rule out explanation b as the cause of senescence in soybeans since nutrient drain was prevented and senescence still occurred. It does strongly suggest however that a hormone produced in the pods travels up the xylem and causes the leaf to senesce (Nooden, 1980).
However a variation to explanation b was implied by the very important findings of Richmond and Lang (1957), that Cytokinins delay senescence in detached plant leaves. It was suggested by Nooden (1978) that perhaps what happens in senescence in soybeans is that the developing pods drain off Cytokinins and minerals from the xylem stream, and these not coming to the leaf either signals or causes senescence. To test this Nooden (1978) tried leaving pods on the bottom half of a plant and leaves on the top and compared this to the reverse condition. He reasoned that if senescence was caused by Cytokinin and minerals draining into the pods, then when only the pods existed near the source of Cytokinin and minerals (which is the roots), senescence should be accelerated in the leaves above when compared to the other condition. He actually found the reverse; senescence was faster when the leaves were near the source of Cytokinin and minerals. The leaves had to be receiving as much of these substances as they needed, yet they were still senescing.
If we look closely, these findings disagree with Nooden's (1982) finding that the senescence signal travels in the xylem. The xylem only flows up in a plant, yet the signal was traveling down in the condition where the leaves were on the bottom. This author believes a possible way of rectifying these results is to conclude that the signal is IAA (Auxin) or a close "chemical cousin" of IAA. IAA has a special system of transport apart from the xylem and phloem, which makes it travel in only one direction: down from the apical meristem where Auxin is made. The signal, whether it was IAA or IAAX (the hypothetical chemical cousin of IAA), would use this system of transport. Although IAA normally only travels in live cells, the problem of how IAA could get around the block of dead cells in the 1982 experiment is solved when we realize that the chemiosmotic transport system is very imprecise and "sloppy" (see Goldsmith, 1974). when IAA is transported from one cell to another, it is simply kicked out the bottom of one cell to be taken up "at leisure" by the next cell. The IAA is free to diffuse anywhere in the space between the cells (though it has something of a tendency to be taken up the next cell below it that has Auxin carriers.)
The following four pieces of evidence support the idea that the signal is a chemical cousin of IAA. First, Lindoo and Nooden (1978) found that, of all the hormones and hormone combinations applied to deseeded soybean pods (including ABA), only IAA at 10 M speeded senescence. This is a very high concentration of IAA never actually found in cells. However it is known that close chemical analogues of a hormone will have the same effect as the hormone, but at a much higher concentration (see Abeles, 1973). Hence the strong concentration of IAA may have mimicked the effect of IAAX in this experiment. Secondly, and probably most importantly, TIBA was found by Dying (1982) to inhibit senescence in soybeans, flax, oats and wheat. TIBA is a known inhibitor of auxin transport, which accords with the author's idea that senescence is inhibited if IAAX is stopped in its transport. Thirdly, Atsumi and Hayashi (1979) found a 300X increase in the amount of IAA in senescing pea and French bean leaves. Their findings also implied a 3000 increase in Indole Pyruvic Acid, a close chemical cousin and precursor of IAA. Perhaps Indole Pyruvic Acid is IAAX. Fourthly, there is the experiment by Nooden et al (1979) which showed that senescence in detached leaves could only be delayed with the addition of synthetic Cytokinin (BA) and synthetic auxin (NAA). This coupled with the work of Tao (1983), who showed that Zeatin but not BA was broken down in senescing leaves seems to suggest that IAA is broken down in senescing soybeans, and thus cannot be the senescence signal. In these four findings, only a chemical derivative of IAA seems to be supported as being the senescence signal.
Some further aspects of how the senescence signal works are suggested by some of the article just listed and several other articles. Lindoo and Nooden (1978) for instance, made the important discovery that Cytokinin levels drop dramatically in soybean leaves just prior to senescence. Additionally, as mentioned, Nooden et al (1979) found that spraying the synthetic Cytokinin BA and the synthetic auxin NAA were the only way to prevent senescence in attached soybean leaves. Researchers up to that time had not found that exogenous application of Cytokinin to attached leaves delayed senescence (although they had found it in detached leaves.)
Insight into why synthetic and only synthetic auxin and Cytokinin could prevent leaf senescence in soybeans, was provided by the work of Tao (1983), mentioned before. He found that the naturally occurring Cytokinin Zeatin, was rapidly catabolized in senescing plant cells whereas BA was not. Thus it appears that in soybean leaves, endogenous IAA and Cytokinin are broken down and that this causes senescence. It follows that the exogenous application of synthetic analogues would prevent senescence, since these analogues are not recognizable by the hormone breakdown enzymes. Additionally a good case appears to exist for the idea that IAAX is the hormone which activates those enzymes that destroy IAA and Cytokinin.
The work of Mondal et al (1978), Franceschi et al (1983a,b), and especially Wittenbach (1982, 1983a,b) has recently questioned some of the assumptions of Nooden and other researchers on senescence. This work has given us a new picture of senescence in soybeans, in which senescence is still caused by hormones but is a two-step process instead of a one-step process assumed earlier. Mondal et al (1978) and Wittenbach (1982, 1983a) found that a decline in photosynthesis occurred in soybean leaves during the normal podfill period, whether the plant was podded or depodded! In fact both researchers found that depodding somewhat increased the rate of photosynthesis decline, although it still greatly delayed leaf yellowing and death. Nooden, Thimann and almost all workers in plant senescence had been determining the extent of senescence, by visually examining the degree of chlorophyll loss (leaf yellowing). Chlorophyll loss gives us an idea of the extent of actual cell death, but does not indicate the state of functional senescence in the leaf. From this point on "functional senescence" will refer to the decline in photosynthesis, whereas "final senescence" will refer to actual leaf and cell death.
Other important findings of Wittenbach (1983a) are that RUBISCO and stomate opening markedly decline during functional senescence of both podded and depodded plants. The decline in RUBISCO appears to occur prior to the beginning of the drop in photosynthesis and is believed by Wittenbach (1983a) to be the cause of the photosynthesis drop in functional senescence. Stomate closing on the other hand is not well correlated with the photosynthesis drop and seems to occur after it. Wittenbach (1983a) concluded that senescence in soybeans is a. two-step process where the first step is characterized by decreasing photosynthesis (functional senescence) and the second step by general cell breakdown, visually noticeable by a decline in chlorophyll.
Additional findings by Franceschi et al (1983a,b) and Wittenbach (1983b) bring to light some further plant physiological aspects of soybean senescence. Summarizing quickly, they found that a special group of cells exist in soybean leaves between the spongy mesophyll and the vascular bundles. These cells sequester in their vacuoles storage proteins that are made first from the breakdown of RUBISCO and then from the products of final senescence. At the beginning of podfill, these storage proteins appear to be broken down and exported, as amino acids in the phloem, to the developing seeds. (The name of the special cells is the paraveinal mesophyll.)
Getting back to the two-step senescence process, it is possible that it may not exist for some species. However, there would seem to be real theoretical advantages for all plants in having a two-part senescence. When a cell is dead, it can obviously no longer actively transport. Hence, if leaves were killed early on in the senescence process, and broken down by proteases, it might be much more difficult for the seed to get the nutrients it needs from the cells. RUBISCO is known to make up as much as 50 of all leaf proteins (Wittenbach, 1982). If this protein is simply broken up at first, the cells can live on stored starch and transport the amino acids from this protein. (A two-step process probably also contributes to the transport of certain minerals, which move only very slowly by passive uptake by the seeds, and thus would get trapped in the cells if the leaves died.) Once the RUBISCO source is exhausted however, then the plant has to start breaking down other enzymes which are more important for keeping the cells alive from minute to minute. This is when final senescence and cell breakdown would occur.
As we saw with evidence that the spinach senescence signal is produced in the leaves instead of the seeds, not all plants senesce in exactly the same way as soybeans. Nevertheless, there is evidence that a two-step senescence process exists in a plant totally unrelated to soybeans. This is Wittenbach's (1978) finding that wheat has two clear distinct phases of protein breakdown during senescence. Further parallels to the soybean pattern of senescence are suggested by the fact that the total protein content of the leaves does not go down during the first phase of wheat senescence. This implies that, as in soybeans, storage proteins increase during the decrease in RUBISCO, balancing out the loss of RUBISCO in the leaf. If all of this is true, the parallel between soybeans and wheat would have to have been independently evolved, because these two species are so distantly related. It seems reasonable to expect that there will be major differences between the two species in other mechanisms of senescence.
Looking back at the three original explanations of how senescence takes place physiologically, theories a and b are definitely ruled out for soybeans, and probably for oats, wheat, spinach and many other plants. With any plant species, two simple experiments could be done to rule out explanations a and b. First keep the plants in non-flower inducing photoperiods. If they live much longer than controls exposed to flower inducing photoperiods then this species is not dying because of aging. Secondly, take senescing plants of this species and spray on the plant a solution of Hoagland's nutrient solution mixed with Tween 80, which is a wetting agent. If the plant senesces as normal, than it is not dying from simple nutrient withdrawal-nutrient starvation. The plant in this case is probably dying from a hormone induced senescence. (If anybody is wondering, the last idea is not original with the author.)
As has been mentioned, it is probable that many or most annuals senesce by a process of hormone signals which cause the cells to begin degrading themselves. It is also probable that most annuals go through a two-phase sequence consisting of functional senescence, which involves the breakdown of RUBISCO and the export of resulting amino acids to the developing seeds, followed by final senescence, which involves the general breakdown of the cells in the main plant. Again because of space limitations and the need for simplicity (and because of the fact that so little work has been done on functional senescence), this section will concentrate mostly on final senescence. Some of the mechanisms examined will be applicable to both phases of senescence.
A number of cell physiological changes (listed below) are known to occur in all types of plants during final senescence. As we shall see, they can all be fitted into a model which holds that a decline in polyamines produced by a decline in Cytokinins causes final cell senescence. The cell biological changes during final senescence are: a. Large decreases in chlorophyll (Leopold 1959 I: soybeans and spinach!; Krizek, 1966 Cocklebur:!, etc.). A large decline in proteins and DNA (Hardwick and Wool house,1967 C Perilla frutescens D; Hurst and Gahan, 1975 C Lycopesicon esculentum]; Osborne 1962 Ccocklebur3; Smille and Krotkov1961 Cpea3). c. A large increase in the activities of hydrolytic enzymes- lipases, DNases, RNases, proteases and chlorophyllases (Chatter jee et al, 1976 Crice and Chibnall, 1954 Common bean; Chin and Beevers, 1970 [nasturtium]; Choe and Thimann,1975 Coats3). d. A drop in the level of polyamines, Spermine and Spermidine in the cells (Kaur-Sawhney, 1982 Coats 807: decrease); Palavan and Galston, 1982 Coats?]; Altman and Bachrach, 1981 Cdwarf bean 55 decrease, but no decrease in tobacco.) e. A large increase in respiration (Hardwick et al, 1968 CPerilla frutescensJ; Thimann et al, 1974 Coats, but may not occur in all species3; Nooden, 1980). f. Mitochondria remains intact and functioning until late stages of senescence (Hardwick and Wool house, 1967 CPerilla frutescensj; James, 1953).
The most important of these characteristics is the rise in hydrolytic enzyme activity (point c). This rise probably causes the decline in protein, chlorophyll and DNA. Protein and DNA are two of the most important elements of cell function. Their disappearance is probably the most direct antecedent of death. Bull how is this increase in hydrolytic enzyme activity brought about?
A first clue (not realized until many years later) came to light when Martin and Thimann (1972), found that Arginine, a precursor of the polyamines Spermine (spm) and Spermidine (spd) inhibited senescence in detached leaves. This led to the work of Galston and Kaur-Sawhney (1978) who found that spm and spd even more greatly delay senescence in detached leaves. The latter authors were then able to explain why polyamines inhibit detached leaf senescence-they found that DNase and protease activity were greatly inhibited by spm and spd. As mentioned before and developed below, there are real problems in trying to determine the pattern of whole plant senescence by examining senescence in detached young leaves. However the finding that polyamine levels decline in attached leaves before and during final senescence (see the references above) strongly supports the results on detached leaves.
Another intriguing finding by Suresh (1978 Coats?3) suggests where polyamines may fit in the causal chain entailed in final senescence. He found that the application of Cytokinin, Auxins or gibberellins all produced a many-fold rise in the levels of spm and spd. These substance are all known to prevent, to delay or to help prevent senescence. The findings suggest that intercellular levels of spm and spd might be under the control of these hormones. The research on all polyamines mentioned above, along with some other research mentioned later, led Galston (1982) to conclude that spm and spd are some sort of second messengers for Zeatin (a Cytokinin), IAA and GA (Gibberellic acid) in the final senescence process. The idea is that the normal levels of Zeatin, IAA, and GA keep up high levels :of spm and spd in the plant cells. These polyamines are then ^responsible for delaying final senescence, by inhibiting the synthesis or the activity of already synthesized hydrolytic proteins (the "ases".) When Zeatin and IAA (and probably GA too in this scheme) are broken down during final senescence (remember Tao, 1983), polyamine levels decrease and the hydrolytic enzyme activity is released from inhibition, causing the death of the plant.
Accepting the above scheme as the mode of final senescence in plants, we can go on to the next link in this chain of causation. This next link is the question of how polyamines actually inhibit hydrolytic enzyme activity. Much work has been done on polyamine activity in plant and animal cells. Although experiments done on animal cells are most often not applicable to plant cells, the work on animal polyamines may be an exception. Polyamines have been shown to be as important in animal cell functioning as in plant cell functioning. For instance, mutant animal cells that could not make polyamines were shown to be unable to move to the Gl phase of cell replication (Rupniak and Paul 1981.)
There are three possible ways in which polyamines can have their effect 0" hydrolytic enzyme activity; by binding to and thus directly inhibiting the hydrolytic enzymes, by binding to and thus stabilizing the substrates (membranes, proteins, etc.) from breakup, and by binding to an intermediate protein that inhibits all hydrolytic enzymes.
Thai polyamines inhibit hydrolytic enzymes by binding to these enzymes is suggested by the experiments of Kaur-Sawhney et al- (1978). These authors found that the normal rise in RNase activity in oat protoplasts during their senescence was inhibited by polyamines. More importantly, this inhibition occurred even after all spm was removed from the RNase suspension solution by exhaustive dialysis before the addition of F|NA. This suggests that inhibitory effect of spm and spd may be due to direct binding to RNase. In 1982, Kaur-Sawhney et al. found by the same process that polyamines bind to proteases and not to proteins, inhibiting their breakdown. Additionally Hasnain (1980) found that DNase was inhibited by polyamines, with some indication of attachment to the DNases as well as the DNA.
That the studies just mentioned may not be fully valid is suggested by x-ray crystallography that spm and spd have a penchant for binding to DNA and stabilizing it (Liquori et al, 1967; Suwalski, 1969; Zhurkin, 1980). This is consistent with the findings of Mager (1959) that spm and spd bind to membranes and proteins, stabilizing them. If the main action of polyamines is to bind to macromolecules, then it is understandable that hydrolytic enzymes would have difficulty breaking up these molecules, since the molecules are stabilized and "stitched together" by polyamines.
An interesting experiment by Chen (1983) also casts doubt on the conclusion that polyamines bind to hydrolytic enzymes and supports our third account of polyamine action. Chen gave radioactive spm an opportunity to bind to all the proteins found in rat (Morris 3924A) hepatoma cells and found that they bound to only one protein, 18kd in size. It is reasonable to assume that the available proteins included hydrolytic enzymes. Hence, this finding suggests that polyamines do not nonspecifically bind to hydrolytic enzymes and indeed that they do not nonspecifical1y bind to proteins in general. To be sure that this finding can be generalized to plants, this study should be replicated on oat cells from senescing plants. If polyamines directly bind to hydrolytic enzymes, then we should find a binding of polyamines to several proteins of different weights. However if we find that spm and spd bind only to one protein (as in Chen's study), then we would have to conclude that polyamines have their inhibitory effect through the action of a single protein rather than through direct action on all of three "ases."
If the latter conclusion were sustained, it would be necessary to explain the Kaur-Sawhney et al (1978, 1982) findings which suggest that spm and spd directly inhibit pure RNase and protease. A possible explanation is that, at least in the 1982 protease experiment, a crude (the supernatant from a high seed spin of a general cell preparation) rather than a pure enzyme preparation was used. Hence spm and spd could have been binding to one protein in the general protein preparation, which was then inhibiting the protease.
If Chen's findings can be shown to apply to senescing plants, it will be interesting to speculate about the nature of the single protein through which polyamines have their inhibitory effect on senescence. The author wishes to nominate the protein calmodulin for this role. Calmodulin is a calcium binding protein which is known to activate certain key cell proteins such as adenylate cyclase, NAD kinase, protein kinase, Ca pump of the plasma membrane, and phosphodiesterases (Rasmussen 1982.) Plant calmodulin is known to be so similar to animal calmodulin, that calmodulin isolated from zucchini has been found to stimulate calf brain C-amp-ase to the same extent as calmodulin isolated from the calf brain itself (Marme, 1982).
There are many pieces of evidence that suggest to this author that polyamines have their inhibitory effect through the protein calmodulin. Like calmodulin, polyamines have also been shown to activate the Ca plasma membrane pump (in response to testosterone, Koenig et al, 1983). It is also known that polyamines stimulate a calcium dependent phosphatidylinosi1-phosphodiesterase in rats (Eichberg et al, 1981), resembling calmodulin in this way as wel1.Polyamines have additionally been shown to activate protein kinase (see Morishita et al, 1983). This activation of protein kinase is increased by the exogenous addition of calmodulin and occurs in the absence of calcium (Criss et al, 1983)' This all seems to show that calmodulin and polyamines activate the same key regulator proteins and in fact work in such a synergistic way (Criss et al, 1983) as to suggest that polyamines bind to calmodulin and have some or all of their effects through catmodulin.
The case for the role of calmodulin is strengthened by indirect evidence that calmodulin inhibits senescence. This is suggested by the finding that calcium defers senescence in corn and Rumex (Pooviah and Leopold 1973). This bears on the action of calmodulin, since calcium has many or most of its effects on cells through the binding to and stimulation of calmodulin (see Marme, 1982). Thus the senescence may be inhibited by the stimulation of calmodulin. (This hypothesis should be examined with calmodulin inhibitors.) Other evidence for the role of calmodulin can be seen in the finding that calcium and calmodulin inhibit the rise of lipoxygenase (another hydrolytic enzyme which breaks up membranes and is closely associated with senescence), in senescing plant tissue (Lesham, 1982). (It should be remembered that polyamines "stabilize" the membranes of oat protoplasts from senescence according to Galston et al C19783 and that others have claimed to find that polyamines stabilize membranes in general CMager 19593.) It is thus the opinion of this author that polyamines do not "stabilize" membranes through the action of binding to these membranes, but by binding to calmodulin which is then stimulated to inhibit 1ipoxygenase, which would otherwise break down these membranes. This also suggests that the inhibition of all the hydrolytic enzymes (DNase, RNase, protease, chlorophyllase, and lipoxygenase) by spm and spd and the inhibition of final senescence by these polyamines is caused by their binding to and activation of calmodulin, which in turn inhibits these enzymes.
One final piece of evidence for the special role of calmodulin is that the molecular weight of calmodulin is known to be between 15kd and 19kd (Rasmussen, 1982). The protein which Chen (1983) found binding to spm had a molecular weight of 18kd, which is within this range.
Going back to the six physiological processes that occur at the cell level during senescence, we still have not accounted for the mitochondrial behaviors (points e and f). Tetley and Thimann (1974) concluded that the climacteric rise in respiration (point e) was due to the uncoupling of respiration from phosphorylation caused by hydrolytic enzymes. They reached this conclusion because they found they could produce similar rises in non-senescing leaves, by treating them with the classic respiration-phosphorylation uncoupler, DNP. Additionally, when they added DNP at the peak of the climacteric respiration, only a small rise in respiration was noticed. Nevertheless Malik and Thimann (1980 Coats:!) found that the climacteric rise in respiration was paralleled by a rise in ATP levels, so they concluded that the previous conclusion was erroneous.
This fact along with the fact that mitochondria appear to remain structurally intact and functioning (whereas the chloroplasts are broken up early on see Nooden, 19803) makes some theoretical sense in terms of the model we are developing. Energy may be needed for the active transport of large amounts of amino acids, minerals and other nutrients out of the dying cell. This could be examined experimentally, by treating senescing plant leaves with DNP to see if this slows down the nutrient translocation to the seeds.
A Model for
With all these ideas and pieces of information, we can now construct a model for senescence in soybeans. This model is not applicable in all its features to other plants, since annuals with different mechanisms may well have evolved at different times in response to the availability of receptive niches. It is useful nevertheless to have a model for senescence in soybeans, because the sequences of events in other senescing plants can be compared and contrasted to those predicted by this model, to bring the method of senescence in these plants into high relief.
The model is as follows:
A, In the niche where soybeans developed, BaCa was greater than BpCp + P (using the author's version of the Charnov-Schaffer equations) so these plants became annuals.
B. Soybeans cause themselves to senesce primarily in order to fill their seedpods rapidly and efficiently so as to avoid the frost. Secondary reasons might include protecting the seeds from winter with crop residues and fertilizing the ground around the seed.
C. Soybeans begin to functionally senesce by the selective activation of a chloroplast-located protease specific for RUBISCO. This protease is activated by a hormone manufactured in either:
1. the photo-induced leaves or
2. the developing flowers.
D. The resultant amino acids from the RUBISCO breakdown are transported to the paraveinal mesophyll and stored in the vacuoles in the form of three storage proteins of 27kd, 29kd and 52kd respectively. This takes place beginning with anthesis (full bloom).
E. When the pods begin forming, the storage proteins are broken down and transported to the seeds as amino acids.
F. At a point between early and mid-podfill, another signal is released by and only by the developing seeds. This travels by live cells (maybe the IAA transporting cells) and reaches the leaves. This signal may be a chemical cousin of IAA.
G. This hormone, upon arriving at the leaf cells, signals them to activate enzymes which break down cytokinin and IAA.
H. In the absence of cytokinin and IAA, polyamine levels decline dramatically.
I. In the absence of polyamines (spd and spm), calmodulin activity declines dramatically.
J. This frees already existing hydrolytic enzymes to begin breaking down the cells. Also, important enzymes of normal cell functioning such as many protein kinases, are prevented from working by the absence of calmodulin.
K. Mitochondria somehow stay intact and functioning late into senescence, fueling the active transport of nutrients out of the dying cell.
Experiments for Understanding
The problem of what causes functional senescence in soybean leaves (the decline in RUBISCO and photosynthesis) has received little attention so far, because its distinctiveness from chlorophyll loss and leaf death was not clear until the work of Wittenbach
(1982,1983a,b) and Mondal et al (1978). The proposed experiments are:
1. The basic question is, does the decline in RUBISCO and photosynthesis occur because of simple leaf aging or because of other processes. This question has essentially already been answered in the paper of Nooden (1977), when noninduced (kept in a nonflower inducing photoperiod) soybean plants were found to live over a year and grow 23 feet high. However, the rate of photosynthesis and F<UBISCO levels should also be measured in such non induced plants to completely rule out aging as a cause of functional senescence.
2. If functional senescence is not simple aging, then it may have to: do with the nutrient drain of the flower. To examine this possibility, flower-induced plants should be debudded and the photosynthesis and RUBISCO levels measured. If these level do not decline, then indeed functional senescence may have to do with ? nutrient drain (the nutrient drain of the developing flower may be a signal for or a cause of chloroplast breakdown). On the other hand a lack of decline in RUBISCO and photosynthesis in the debudded soybeans may be due to the existence of a hormone produced in the flower bud or flower itself, which normally causes functional senescence in flower-induced plants.
3. To distinguish between a nutrient drain signal and a flower derived hormone signal a study similar to Nooden's (1982) phloem destruction experiment could be done. In this case again a small cross section of the petiole would be steam killed. This would prevent nutrients from draining to the developing flower but would not prevent a hormone from the flower from traveling to the leaves. If functional senescence still occurred with this treatment, then functional senescence is caused by a hormone produced by the flower. If functional senescence did not occur with this treatment (combined, remember, with the finding of no functional senescence after debudding), then functional senescence could be assumed to be caused by the signal of a nutrient drain coming from the flowers.
4. Backing up to experiment number 2, if debudding of flower-induced plants did not stop functional senescence, then the senescence signal could probably be assumed to derive from the photoinduced leaves themselves. To determine whether the signal is derived from the cytoplasm or the chloroplast, the chloroplasts could be removed and subjected to experimentation. (It has been found that chloroplasts can continue to do photosynthesis outside of cells.)
If chloroplasts irradiated with the flower-inducing photoperiod continue to photosynthesize and to maintain a high level of RUBISCO, then the functional senescence signal is cytoplasmically derived. If; a decline in photosynthesis and RUBISCO does occur, then the senescence signal is endogenous to the chloroplast.
There are three areas that appear worthy of a little further discussion, the implications for agriculture of the knowledge gained from the study of plant senescence, the implications of this knowledge for a general understanding of the role of polyamines in plant biology, and finally the implications of the annual strategy for the broader issues of mortality and immortality. Improving Agriculture
There are two ways in which senescence research could help improve crop yields: by showing how to delay senescence in legumes and by showing how to increase the efficiency of hydrolytic enzyme activity during final senescence in nonlegumes.
With respect to the delay of senescence, it should be noted that soybeans (and most likely other legumes) senesce quite early, Presumably because they developed in a cold climate where frost comes quickly in the fall. This timetable has no function in latitudes where frost never hits, or hits long after soybeans have senesced. If soybeans could be prevented from functionally and finally senescing then (according to calculations by Sheehy C19833) the seeds still could be filled and protein and nutrients left in the leaves of the plant. This would be very valuable because, as it stands now, farmers must snake a choice each year between using legumes like soybeans as a crop or as nitrogen-fixing, fertilizer-producing green manure. : Delaying senescence would allow farmers to get a crop and fertilize the soil in the same season. Now farmers must plow their entire crop, beans and all, under the soil if they want to fertilize the soil. With senescence delayed, the beans could be harvested and then the leftover leaves, stems and roots could be plowed under. The material plowed under would have roughly the same amount of protein as the entire normal senescing plant (beans and all) since, in the non-senescing variety, no nutrient redistribution to the seeds would occur. This non-senescing variety could greatly reduce the need for fertilizers and crop rotation when growing legumes.
This prevention of senescence could be carried out in two ways. The first way is to try to find or induce mutants that do not functionally or finally senesce. In fact Abu-Shakra et al (1978) have found such a non-senescing variety of soybeans (as previously mentioned.) As predicted, this variety has a higher final amount of leaf protein but a normal seed yield compared to the common senescing varieties. This variety, if propagated, would be immediately useful to farmers The other way of preventing functional and final senescence would be to discover the hormones (the author really believes that it is two hormones that are the causes of these processes) that cause senescence and then to neutralize them in some way, e.g., if the final senescence hormone works by breaking down CK and IAA then exogenously apply these hormones to the plants.
The other lead for improving yields comes from the studies by Rao and Croy (1971. 1972) and Perez et al (1973), which showed that higher protease activity in senescing leaves led to higher grain yields. If it could be understood exactly how proteases are put into action during senescence, their efficiency could be increased, with resulting higher yields.
The Role of Polyamines
The polyamines Spermine and Spermidine appear to be general stimulators of growth and activity in the cell, through the action of calmodulin. Also through calmodulin activity, polyamines appear to be inhibitors of self destruction of the cell. Polyamines seem to be a kind of cellular glue which can be and is turned off selectively (in older leaves during progressive senescence) when needed. The action of polyamines might be one mechanism that plants (and animals too) have for keeping their constituent cells under control. The cells are in a sense always poised near the point of death, and all the plant has to do is stop supplying cytokinin, and the polyamine levels plummet, setting self-destruction mechanisms into action. This may be one of the methods plants use to keep their cells from using the plant as a giant agar plate to proliferate on. Conversely loss of control over the synthesis of polyamines may be one of the causes of tumor formation in plants and animals. The Relationship of Annual Senescence to Aging Patterns of Death
Annual plants appear to be different in their method of death from most of the rest of higher life forms. Aging appears to contribute very little to death in these plants. Annual senescence in plants is a genetically programmed short circuiting of life in order to further the progeny. Extrapolating this point to other organisms, it seems possible that death by aging may also be genetically programmed. In other words, it can be speculated that immortality would exist if it were not evolutionarily more adaptive to die.
However, there seems to be little adaptive significance to aging itself. In human beings for instance if death is adaptive, why don't we all die after a certain age. Graying hair, weaker organ function, and a mind that progressively fails over a period of many years, do not appear to be adaptive.
Looking at things from another perspective there are immortal organisms of sorts. These are the asexually reproducing single cell organisms. Their immortality, however, comes at the cost of a loss of individuality at each division. All higher organisms, to maintain their high levels of order and organization (their "individuality"), may need to reproduce by starting over again from one cell since high amounts of order (the highly evolved species such as man) can probably only be created from simpler structures. In other words, ontogeny must recapitulate phylogeny in order to produce the highly ordered and evolved individual. Yet because of this high ordering we must die. Lewis Thomas has said that death is the price we pay for individuality. Perhaps then annual plants are simply using an inevitable process to their own advantage, by causing and controlling an earlier death than would normally occur.