|Home | About | Journals | Submit | Contact Us | Français|
Male plants of spinach (Spinacea oleracea L.) senesce following flowering. It has been suggested that nutrient drain by male flowers is insufficient to trigger senescence. The partitioning of radiolabelled photosynthate between vegetative and reproductive tissue was compared in male (staminate) versus female (pistillate) plants. After the start of flowering staminate plants senesce 3 weeks earlier than pistillate plants. Soon after the start of flowering, staminate plants allocated several times as much photosynthate to flowering structures as did pistillate plants. The buds of staminate flowers with developing pollen had the greatest draw of photosynthate. When the staminate plants begin to show senescence 68% of fixed C was allocated to the staminate reproductive structures. In the pistillate plants, export to the developing fruits and young flowers remained near 10% until mid-reproductive development, when it increased to 40%, declining to 27% as the plants started to senesce. These differences were also present on a sink-mass corrected basis. Flowers on staminate spinach plants develop faster than pistillate flowers and have a greater draw of photosynthate than do pistillate flowers and fruits, although for a shorter period. Pistillate plants also produce more leaf area within the inflorescence to sustain the developing fruits. The 14C in the staminate flowers declined due to respiration, especially during pollen maturation; no such loss occurred in pistillate reproductive structures. The partitioning to the reproductive structures correlates with the greater production of floral versus vegetative tissue in staminate plants and their more rapid senescence. As at senescence the leaves still had adequate carbohydrate, the resources are clearly phloem-transported compounds other than carbohydrates. The extent of the resource redistribution to reproductive structures and away from the development of new vegetative sinks, starting very early in the reproductive phase, is sufficient to account for the triggering of senescence in the rest of the plant.
Many plants possess the characteristic of living more or less indefinitely until some agent or stress intervenes. However, some plants have evolved the monocarpic growth habit in which the entire plant undergoes a programmed cycle of organ degradation after reproduction, followed by the death of the plant. Two hypotheses have been advanced to explain this correlative phenomenon: (i) that allocation of nutrients to reproductive tissue results in the degradation of the vegetative structures or (ii) that a factor, probably hormonal, is produced by fruits and is then translocated to the leaves and stem apex, triggering their senescence. Literature supporting and challenging these proposals has been reviewed in Kelly and Davies (1988b) and Sklensky and Davies (1993). In either case, expression of senescence-promoting and senescence-retarding genes interacts with the plant's environmental and internal conditions to regulate the active senescence process (Noodén et al., 1997).
Molisch (1938) first proposed that a drain of nutrients to the fruits was the mechanism for the induction of whole plant senescence. An objection to this hypothesis emerged from work involving dioecious spinach plants (Leopold et al., 1959; Janick and Leopold, 1961), which normally undergo senescence following flowering similar to any other monocarpic species. However, as the removal of the very small pollen-producing flowers, which were assumed to be negligible as a nutrient sink, resulted in a delay of senescence it was suggested that these flowers must act in some other fashion to induce senescence. One such possibility could be by the production of a senescence factor (Kelly and Davies, 1988b), but experiments in peas failed to provide any support for the idea of a senescence factor (Hamilton and Davies, 1988a, b), supplying instead evidence that senescence is regulated by a shift in nutrient allocation from the vegetative to the reproductive tissue that occurs very early in the reproductive phase (Kelly and Davies, 1986, 1988a, b).
The explanation for whole plant senescence that best fits the evidence currently available is that a physiological transition is initiated at flowering and results in a resource allocation that is detrimental to the maintenance of vegetative tissues in comparison to the uninduced state. Many studies have demonstrated that alterations in the source–sink relations of vegetative and reproductive tissues can affect the course of senescence. Senescence can be prevented in pea genotypes with a slower rate of reproductive growth that enables continued resource allocation to vegetative growth (Kelly and Davies, 1986, 1988a). The endogenous auxin and gibberellin content of floral and vegetative tissue within the apical buds of these peas correlated with this resource allocation (Zhu and Davies, 1997): a higher gibberellin content of the apical vegetative tissues within the apical bud was associated with vigorous vegetative growth, slower floral development, and continued growth, whereas the greater rate of floral bud growth, which precedes senescence, was associated with a higher indoleacetic acid content in the floral buds. Senescence of soybean was also delayed by the mechanical prevention of pod expansion (Crafts-Brandner and Egli, 1987; Miceli et al., 1995). Such physical restrictions to soybean seed growth did not alter the time of initiation of senescence but decreased the rate of leaf senescence as judged by changes in photosynthesis (Crafts-Brandner and Egli, 1987), indicating that senescence was initiated at the same time regardless of sink size but that the rate of senescence was modulated by sink size. Restricting soybean pod growth did not affect total plant dry matter and N accumulations during seed-filling because there were proportional increases in the partitioning of assimilates into stems and leaves (Miceli et al., 1995). Similarly, pea plants bearing the alleles ar and n, which have smaller seeds and lower total seed yield, showed a resurgence of growth after apical growth had initially stopped, which is when normal plants undergo senescence. Only after further growth did the plants undergo full senescence (Murfet, 1985). The recessive alleles ar and n were proposed to impose a lower metabolic drain per reproductive node as a consequence of their restrictive effects on hilum anatomy and pod morphology, respectively, leading to a reduction in sink capacity. As a consequence the developing seed crop fails to cause plant senescence and death at the usual developmental time. Similarly, in maize, crop manipulations leading to kernel set restrictions enhanced post-flowering assimilate availability and reduced leaf senescence (Borras et al., 2003), whereas the acceleration of senescence by the imposition of water stress increased the rate of grain-filling in wheat and rice (Yang et al., 2001a, b; 2003).
The initiation of the senescence programme appears to occur early in the flowering period, and, in most cases, senescence can only be delayed, not prevented, by surgery or hormone application. Sterile and other mutants of Arabidopsis with delayed senescence did not exhibit prolonged leaf life, but did show an extended production of leaves and flowering stalks (Noodén and Penney, 2001). Following soybean pod removal the leaves did not show the dramatic visual yellowing associated with senescence, but function, as measured in terms of the rate of photosynthesis or the activity of ribulose bisphosphate carboxylase, was inhibited (Mondal et al., 1978; Wittenbach, 1982, 1983a, b). Delayed senescence in ‘stay-green’ genotypes of sorghum did appear to result in prolonged photosynthetic capability, but this seemed to be most closely correlated with greater nitrogen assimilation during the grain-filling period (Borrell et al., 2001).
Sinks have profound effects on the photosynthesis in source tissue (Paul and Foyer, 2001). Partitioning enzymes in sink tissue, such as sucrose phosphate synthase, showed both short- and long-term regulation by sugar concentrations (Huber and Huber, 1996). In addition to affecting partitioning enzymes, sugars also play a key role in the transcription regulation of such key photosynthetic genes as rbcS, cab, and atpD (nuclear genes encoding the small subunit of Rubisco, the chlorophyll a/b binding protein, and the D subunit of a thylakoid ATPase, respectively) (Krapp and Stitt, 1995). The interrelatedness of these facets of carbohydrate metabolism suggests that measurement of translocation between source and sink under conditions of varying senescence would provide useful information on the events leading to senescence.
In light of the clear results in peas that senescence is regulated by a shift in nutrient resource allocation early in reproductive development (Kelly and Davies, 1986, 1988a, b), the regulation of monocarpic senescence in spinach has been re-examined. The prior assumption was that the utilization of carbohydrate resources by staminate spinach flowers is negligible and would have no effect on senescence regulated by nutrient ‘exhaustion’. This has led to an investigation of photosynthate allocation in both male (staminate) and female (pistillate) spinach plants as an indication of overall phloem-derived resource allocation in order to determine whether the nutrient demand by staminate spinach flowers is insignificant, as has been previously assumed. It is demonstrated below that staminate flowers have a nutrient demand exceeding the rate of import of pistillate flowers, and thus could be a determining factor in bringing about monocarpic, whole-plant senescence, even in male plants.
Below are the abbreviated Materials and methods. Complete details can be found in the Supplementary material at JXB online.
Comparison among three different cultivars of spinach (Spinacia oleracea L.: Packer, Asgrow XPH 1450 and Asgrow XPH 1510) led to the selection of Asgrow XPH 1510, due to its retention of a vegetative state under short days and its tendency to flower rapidly under long-day conditions. Seeds of cv. Asgrow XPH 1510 (Asgrow Seed Company, Gonzales, CA), were sown in flats of peat–vermiculite mix under greenhouse conditions and grown in an 18 h photoperiod (long-day inductive conditions), with daylight supplemented by overhead lights. Seedlings emerged from the soil in 7–10 d and were subsequently transplanted to 12.5 cm pots. The first flower buds were just visible about 18 d later, designated ‘onset of flowering’. Thereafter, treatments and sampling were undertaken weekly. Sampling of staminate plants ceased after the fifth week from the start of flowering because, at that time, they senesced, whereas that of pistillate plants continued through the seventh week.
Plants were moved to growth chambers, with 18 h day lengths of 175 μE m−2 s−1, 19 °C day and 17 °C night. Three to four hours after the start of the light period, individual plants were exposed to 14CO2 for 1 h. A thick, transparent polyethylene bag (10×20 cm or 10×40 cm), containing a 3 ml plastic vial, was placed over a single leaf or over the entire young inflorescence region, respectively. Unless otherwise noted, the labelled leaf was located immediately below the inflorescence region, designated L–1. Leaves within the flower-bearing portion are indicated by positive L numbers, with the lowest leaf that subtends an axillary raceme being referred to as L1. The bags were sealed around the petioles with adhesive paper tape and petroleum jelly, whereas Cling florists’ adhesive sealed the inflorescence regions in an air-tight manner. A syringe was used to inject 1 μCi of NaH14CO3 in 20 μl of water into the vial, followed by 200 μl of 0.6 M HCl to release the 14CO2. The sealed bag contained approximately 1.5 μmol CO2, including 0.0182 μmol 14CO2. The injection hole was sealed with Scotch transparent adhesive tape. The plant was left to assimilate the 14CO2 for 1 h followed by a variable chase period (usually 3 h).
To determine the utilization of carbon over a longer term, both staminate and pistillate plants in the third week of floral development were left for chase periods of 1, 2, 3, or 4 d following labelling of the leaf immediately below the inflorescence region (L–1) for 1 h.
Following the post-labelling chase period, plants were dissected, weighed, frozen on dry ice, oven-dried for 2 d at 60 °C, reweighed, and combusted in a sample oxidizer. The resultant 14CO2 was captured by an absorbent solution, and this solution counted in a scintillation counter with quench correction. Recoveries were 90% or better, as measured by the combustion of standards. Because of variation in the total amount fixed and exported, the measure of the carbon demand of a group of structures was the percentage of the total export that was measured in that structure; this provides an indication of the relative importance of a given structure as a sink for the particular leaf labelled. Total export was calculated by a summation of the total recovered radioactivity from all harvested plant parts, except the labelled leaf. A mass-independent measure of sink strength, designated ‘relative specific activity’ (RSA), was calculated by dividing the dpm g−1 dry weight of an individual organ by the average dpm g−1 dry weight for the plant. At least three plants of each type were tested in each experiment and each experiment was repeated at least twice, usually several times.
Photosynthesis and respiration rates were measured using a Li-Cor 6200 (Lincoln, NE) infrared gas analyser. Photosynthesis was recorded using four leaves individually, from L–1 to L3. Respiration of the inflorescence was measured by CO2 evolution in darkness: 3.5 cm regions of the inflorescence, with the leaves removed, were enclosed in the cuvette, with sealing around the stem. This was repeated following removal of the flowers and fruits.
A dioecious species, the flowering regions of staminate and pistillate spinach plants appear quite different from one another (Fig. 1), but prior to the appearance of the inflorescence the young leaves enclose the apical bud and it is impossible visually to distinguish staminate from pistillate plants.
Under long-day inductive conditions, inflorescences develop in the axils of the upper leaves, and the internodes elongate. The staminate plants develop small racemes of flowers consisting of little more than clusters of anthers surrounded by green calyxes in the axil of a leaf. Larger axillary inflorescences produce two tiers of flower clusters, separated by a slender stalk, with the clusters not visibly subtended by a leaf. In these male-gamete-producing inflorescences, the shape of the leaves changes gradually from the base of the inflorescence, where the leaves resemble the sagittate leaves below, to the apex, becoming extremely small and ovoid (see Supplementary Fig. S1 at JXB online). The pistillate plants produce tiny green sessile flowers, with only the stigmata visible, and the leaves remain sagittate in shape. The upper leaves decline in size compared with the lower ones (see Supplementary Fig. S1 at JXB online), but less so than in staminate plants, remaining considerably larger (Fig. 1). The first flowering nodes are usually located 7–10 nodes above the cotyledons. At the onset of flowering in pistillate plants stigmata become visible as the stem elongates to separate young leaves spatially from those still surrounding the apical bud. Staminate plants, in contrast, produce a convex inflorescence bud no longer enclosed by the young leaves.
In the first week following the onset of flowering, both staminate and pistillate plants produced approximately 10 flowering nodes in 10 cm of inflorescence (Fig. 1; see Supplementary Fig. S2 and Table S1 at JXB online). Stigmata protruded from the axils of leaves in pistillate plants, but the anthers remained enclosed in the calyxes of staminate plants. By the end of the second week after the start of flowering some anthesis and pollen shedding had begun in staminate plants, the inflorescences of both types having attained about 25 cm in length, with up to 30 flowering nodes. Minute fruits began to appear in the lower leaf axils of the inflorescences of the pistillate plants. In the third week of flowering, plants of both sexes had 50 or more flowering nodes. The staminate plants shed copious amounts of pollen and pistillate plants had several large pointed fruits in the axils of lower leaves, although, near the apex, fruits remained small or indistinguishable from the flowers. At this stage, in many plants of both sexes, age-related senescence of leaves below the inflorescence may have begun. By week 4, the staminate plants showed yellowing nearly up to the level of the inflorescence, and the apical bud had converted into flower buds. The pistillate plants either continued production of flowering nodes or ceased node initiation in week 4, with development of the fruits occurring in either case. The staminate plants exhibited signs of the late, degradative stages of senescence in week 5, with overall yellowing, and sampling of these plants ceased before this age. Pistillate plants continued green, with the fruits enlarging, until the seventh week after the onset of flowering, although the lowest leaves often yellowed and withered with age. The overall senescence stage of the pistillate plants in week 8 corresponded to a similar stage of staminate plants in week 5.
Throughout the reproductive period, the leaf area in the region below the flowering area remained relatively constant (Fig. 2A). The large lower leaves in pollen-producing plants at weeks 4 and 5 resulted from the fact that the sturdier plants were more likely to last to that age. The reduction in leaf area in the older pistillate plants was due to the withering of the oldest leaves.
Leaves inside the inflorescence were larger in the pistillate plants than in the staminate plants (Fig. 2B). The two types were similar in the first and second weeks of floral development, but the photosynthetic area in the females increased more rapidly than the males. The highest leaf area average for the staminate plants was 277 cm2 during the fourth week of reproductive development; the pistillate plants averaged 52% larger at that time (421 cm2), and at their maximum in week 6 were 2.2 times larger (615 cm2).
Leaves of plants of similar size and age were treated with 14CO2 for 1 h at 1 week intervals. At the earliest flowering stage, 22% of the photosynthate from the leaf subtending the staminate inflorescence (L–1) was partitioned into the flowers (Fig. 3A). By comparison, in the pistillate plants, 8.6% went to the flowers, indicating that staminate plants allocated 2.5 times as much of the current photosynthate export to flowering structures than did pistillate plants of the same age (Fig. 3A). The similar relative specific activity (RSA), which takes into account sink mass, of staminate and pistillate flowers demonstrates that they have a similar sink strength, about twice the average sink strength of all tissues (Fig. 3B). By contrast, the leaves within the inflorescences of both types of plant (most of which were still expanding) drew about 11% of the export at this early stage (Fig. 3A), although their RSA was below average (Fig. 3B).
At 1–3 weeks following the onset of flowering, when staminate flowers were developing and undergoing anthesis, the differences between the partitioning patterns of staminate and pistillate plants had become increasingly large (Fig. 4A). Reproductive organs in pollen-producing plants received 4–6.5 times more photosynthate than in pistillate individuals [40% versus 9% (week 1) to 6% (week 3), respectively; Fig. 4A]. The mass of both types of flowers also rose rapidly (see Supplementary Table S1 at JXB online). The RSA for the pistillate reproductive structures continued to be similar to that of the staminate flowers through the first week (Fig. 4B), but fell to less than half of the comparable value in the second week and dropped even lower in later weeks.
The bulk of the current carbon fixed by the leaf just below the inflorescence was distributed to the tissue below the inflorescence region, being 63% of the exported radioactivity in the earliest stage of flowering in pistillate plants (Fig. 3A) and more than 50% for pollen-producing plants. However, the proportions become more disparate at later stages. At the time when partitioning to the shoot below the inflorescence declined rapidly for staminate plants (values range from 17–31% after the initial stage) (Fig. 4C), distribution to the comparable region in seed-bearing individuals stayed between 1.3 and 3 times higher, only falling to the 30% level well after the staminate plants had senesced. When the allocation to the lower region of the shoot was considered on a mass-corrected basis, the contrast remained: for staminate plants RSAs were between 0.4 and 1 during the period of flowering (Fig. 4D), whereas the RSAs in pistillate plants were larger at every stage, reaching a value of over 5 in the fifth week.
The percentage exported to the apical bud, which included both leaf and flower primordia, ranged from a fraction of 1% to 3.5% (Fig. 4E), but was, after the first week, higher in the staminate plants than the pistillate plants. The young leaves of the pistillate plants were large enough, relative to the flowers, to enclose the bud, while those of the staminate plants were not. Because of the small size of this structure, even the very small amount allocated resulted in a high RSA, but the values in staminate and pistillate plants were comparable at most stages. However, by week 7, after the staminate plants have senesced, the RSA for the apical buds of the pistillate plants exceeded 12 (Fig. 4F). At this stage, the bud included several large fruits.
By the fourth week of flowering, the staminate plants had begun to show the yellowing that is characteristic of whole plant senescence, and the export to the reproductive structures reached 68% (Fig. 4A). The pistillate plants were still robust, although individual older leaves had begun to senesce, and export from the leaf just below the inflorescence to the developing fruits and young flowers remained near 10% (Fig. 4A). Vegetative structures continued to receive the larger proportion of the exported photosynthate in the pistillate plants, including 51% to the lower vegetative tissue (Fig. 4C). However, in weeks 5, 6, and 7, the allocation to the fruits increased (Fig. 4A), while that to the lower portion of the plant declined. In week 7, the pistillate plants showed signs of overall senescence, and by week 8, most of the pistillate plants were dead.
Three weeks after the start of flowering, a full day of partitioning after 14CO2 labelling of L–1 for 1 h produced a similar photosynthate allocation pattern as did a 3 h partitioning period in both staminate and pistillate plants. The staminate flowers retained 47% of the radioactive carbon, while pistillate reproductive structures had 9% (Fig. 5A). However, whereas the amount in the pistillate flowers remained constant over the next 3 d, the proportion in the pollen-producing flowers had declined appreciably by day 2 and thereafter remained stable over the next 2 d. Whereas less than 9% of exported carbon went to the leaves within the staminate inflorescence, inflorescence leaves in the pistillate plants received 20–33% (Fig. 5B), with no consistent changes with the time. The amount of carbon located in inflorescence stem tissue was greater in the pistillate plants (22%) than the staminate plants for the first day (14%), but the difference disappeared in subsequent days (Fig. 5C). Even at the end of the four days, levels of total recovered radioactivity were similar to values for the 3 h partitioning period, suggesting that any respired carbon was immediately re-fixed.
The RSA of the staminate flowers still showed a substantial drop over the long chase period (Fig. 6). The RSAs of the pistillate flowers and fruits remained similar during partitioning over 8 d, with values of less than 1, indicating a below-average sink activity drawing from the leaf immediately below the inflorescence.
To determine the stage of maximum photosynthate import into the staminate flowers, the flowers of staminate plants in the second week of floral development were each separated into four categories following labelling of the leaf just below the inflorescence for 1 h and a chase period of 3 h. The four categories were: bud, being those flowers and flower parts still held within the axils; extruded, which had extruded anthers; anthesis, where anthers had begun to open; and old, where the anthers were withered and had largely shed their pollen. The flowers in the bud stage received the largest amount of radioactivity, 35%, compared with less than 5% at other stages (Fig. 7A), but they also had the largest mass. The buds were the only stage containing developing pollen, as determined by microscopic examination; in all the other stages, all of the pollen was mature. RSA differences between the staminate floral stages were not significant (Fig. 7B). The RSA for the empty anthers was just below 1, indicating less than average but not a cessation of import, despite the completion of development. However, as the flowers of the older stages were few in number, their total drain was relatively small.
Because of the differing contributions of individual leaves to the reproductive structures, an examination of partitioning of labelled photosynthate from all the tissues within the inflorescence region was examined. Using longer plastic bags, the entire inflorescence region of young plants was exposed to 14CO2 for 1 h, followed by a 4 h chase period. The greater allocation to reproductive structures in the young staminate plants compared with the predominant allocation to vegetative structures in the young pistillate plants, which live longer, was also evident for photosynthate from within the inflorescence (Fig. 8). This difference showed clearly at week 1, when the staminate flowers received 2.5 times more photosynthate than pistillate flowers, but disappeared by week 2. This contrasts with the situation when a leaf below the inflorescence was labelled, when the difference persisted throughout the life of the staminate plants.
Photosynthesis rates in the leaves of the staminate plants were constant during the first 3 weeks of flowering (Fig. 9A). The photosynthetic rates of the leaves declined substantially in week 4 and even further, essentially to zero, in week 5. The pistillate plants, on the other hand, showed a slight reduction in activity from week 1 to week 5, but thereafter displayed noticeable loss of function in weeks 6, 7, and 8 (Fig. 9B).
As the different regions of the inflorescence were at different stages of development, respiration was measured in consecutive 3.5 cm segments, numbering 1 at the apex down to the mature regions of the inflorescence, following the removal of the leaves from the inflorescence. Plants from the second week of flowering were measured at four consecutive segments, with the number of measurements increasing with the larger plants. The staminate plants, stripped of their leaves, had 3–3.5-fold greater rates of respiration at week 2 than did the pistillate plants near the inflorescence apices (Fig. 10A). This difference still existed, but with less magnitude, further down the axis in young plants. At this time, some anthers had begun to shed pollen, although very few in the first two segments. In older plants, at week 4, flowers existed in a mixture of pre-and post-anthesis stages: the majority of the staminate flowers had undergone anthesis while the region nearest the tip showed the fewest open anthers. At this time the respiration rates for the pistillate plants had risen to levels comparable with that of the staminate plants (Fig. 10B). The upper two segments of the pistillate inflorescence had very small flower clusters until week 4, at which time some of those in segment 2 had full-sized fruits.
As the pistillate flowers were initially extremely small, the degree to which the initial disparity in activity results from size was determined based on a calculation of the respiratory rate per gram of flower or fruit. To determine the respiration of the reproductive structures, carbon dioxide evolution of the stem regions was measured with both leaves and reproductive structures removed. This value was subtracted from the rate of respiration of the stem with flowers and fruits, and the resulting value divided by the fresh weight of the flowers removed from the segment, producing an estimate of the respiratory activity of the reproductive tissue on a tissue fresh weight basis. In the apical region, all of the tissues were very active in respiration, with the youngest plants showing the highest level of activity on a per gram basis (Fig. 10C). Below this region, the respiration of the staminate flowers declined, but still tended to be more than in the pistillate flowers. The older fruits showed a reduced respiration on a per gram basis compared with the young pistillate flowers (Fig. 10D).
The concentration of total non-structural carbohydrates (glucose+fructose+sucrose+starch) in the senescing staminate flowers at week 5 was as high as the highest level in pistillate fruits, which occurred close to fruit maturity at week 7 (Fig. 11A). These non-structural carbohydrates showed a continuous increase in leaf L3 of the staminate plants until their senescence (Fig. 11B), while in pistillate plants a very low level was maintained until the fifth week, when the amount rose, peaking 1 week later.
How senescence is triggered remains an enigma (Wingler et al., 2009), and even more so in monocarpic plants (Lacerenza et al., 2010), but here some light has been shed on the degree to which the diversion of nutrient flow from vegetative organs to reproductive organs is involved in the process. In this study, the examination of carbon partitioning throughout the flowering phase of spinach has demonstrated the importance of the reproductive structures of both staminate and pistillate plants to the pattern of resource allocation. The argument against nutrient-drain to the reproductive tissues triggering the senescence of spinach plants, namely that staminate flowers represent an insignificant photosynthate drain on the plant, is clearly refuted. Pollen-producing flowers demand, from an early age, much of the photosynthate of the plant. In fact, at many stages during the reproductive period, leaves on staminate spinach plants export larger quantities of photosynthate to staminate flowers than is provided to the pistillate flowers from a comparable leaf on a pistillate plant. The senescence-delaying removal of flowers from staminate plants (Leopold et al., 1959) could, in fact, affect the nutritional status of the plant in a similar manner to removing fruits from pistillate plants, thus influencing the senescence of the entire plant. The different time-course of senescence in staminate and pistillate plants, and differences in the size of inflorescence leaves, indicate that the rate of reproductive development and the amount of source tissue may also factor in the regulation of senescence of the plants. It is therefore possible to account for the senescence of staminate spinach plants merely on the basis of resource diversion, without invoking any fruit-derived hormonal senescence factor. This is not to say that photosynthate per se is the responsible factor, but rather that the diversion of any required components, such as nitrogenous materials or hormones that are moving with the photosynthate in the phloem, may also have a role in the regulation of senescence.
At the very early stages of reproduction, more fixed carbon was apportioned to the pollen-producing flowers than to the tiny flowers of the pistillate plants. The larger mass of the staminate racemes only partially explains the greater draw on assimilate at this stage. This higher resource allocation to flowers continued throughout the lives of the staminate plants. From an early stage, the fruit-producing plants designated a higher proportion of their carbon intake to the maintenance and production of vegetative tissue to support their reproductive stage than did the shorter-lived staminate plants. While the combined area of the leaves below the inflorescence is similar for each type of plant, the leaf area within the inflorescence becomes substantially larger in the seed-producing plants, providing them with a larger amount of photosynthetically-active tissue in close proximity to the developing sinks. A lower allocation of photosynthate to the stems and leaves developing in the staminate inflorescence results in a reduced growth of these organs, leading to a smaller total of source tissue. The leaves within the pistillate inflorescence stay green and retain photosynthetic activity, as the fruits develop, for weeks after the staminate plants have senesced, and thus contribute to the longer lifespan of the pistillate plants.
Two weeks after the start of flowering the terminal bud of the staminate plants is tightly compacted and received a higher proportion of assimilate than at other times. At this time flowers close to the apical region are undergoing anthesis, with those in the bud nearly ready to do so. The apical buds of the pistillate plants include larger developing leaves that enclose their flowers than do the pollen-producing plants with their exposed flower buds.
The contrast in the senescence programmes of staminate and pistillate plants lies primarily in the timing. Late in development, after the age at which staminate plants have senesced, the fruits began to draw higher quantities of photosynthate (and any other phloem-transported resources) from the lower leaves. This distinction in timing reflects the different roles of the gametophytes produced on each plant, namely the early development of pollen on the staminate plants versus the later development of fruits and seeds. As the fruits develop, the lower leaves allotted a very high proportion of their current photoassimilate to the developing fruits, though this amount fell somewhat towards the very end. By this time, large younger leaves within the fruit-bearing inflorescence have developed, and served as source tissue for the fruits after the oldest leaves have senesced. However, the leaf just below the inflorescence continued to designate a fairly large proportion of the resources to vegetative tissue production and maintenance, demonstrating the importance of the position of the source in consideration of the allocation of assimilates.
Numerous differences exist between staminate and pistillate plants in the long-term fate of the carbohydrate resources in vegetative and reproductive tissues. These differences are explicable as a consequence of their different functions and developmental timing leading to different rates and intensities of energy use.
The pollen-producing flowers had a relatively high degree of metabolic activity, losing substantial quantities of pre-fixed carbon. This rapid rate of respiration is responsible for at least part of the increased greater import of photosynthate into the pollen-producing flowers. By contrast, the pistillate flowers and fruits showed essentially constant levels of radiolabel.
In the shorter lifespan of the staminate plants, the allocation of resources to the flowers producing and shedding pollen indicates the time of greatest metabolic activity in these organs. The staminate flower buds in the leaf axils drew by far the highest proportion of label. Flowers at later stages, in which the anthers are extruded from the calyxes, following dehiscence at anthesis, received much smaller portions of assimilate. However, the allocation of radioactivity to different stages was similar on a mass-normalized basis, suggesting that the respiration of pollen and anthers continued beyond the earliest stages of flower development.
While staminate plants achieve, at most, a ratio of 2:1 for the contribution of photosynthetic tissue within and below the inflorescence respectively, pistillate plants more commonly have ratios of 4–5:1, possessing far larger, young, active leaves in the inflorescence region by the third week of flowering. The younger leaves were the most important as source tissue whereas the lower leaves of both staminate and pistillate plants exhibited an age-related decline in photosynthesis. However, even these older leaves maintained some function for a longer period of time in the pistillate plants. Comparable rates of carbon dioxide uptake per unit area were seen in both sexes for both the leaf just below the inflorescence and each of the three lower leaves of the flowering region, until the leaves of the staminate plants cease to photosynthesize in week 5. The larger size of these younger leaves in the pistillate plants indicates that more assimilate is being produced than in the staminate plants.
Young pollen-producing inflorescences are relatively large and respired at a far greater rate than did their much smaller pistillate counterparts. Even on a weight basis, the young staminate flowers respired somewhat more rapidly than did the flowers or fruits of the pistillate plants. Mid-age inflorescences were more comparable to one another, with the staminate flowers, which are mostly at, near, or post-anthesis, respiring at a much slower rate. As the fruits enlarged, their respiration increased slightly, but the per gram rate of respiration declined as the fresh weight rose. Staminate plants make an early, highly energetic contribution to reproduction, with no necessity to prolong function by producing large source leaves later in development. Pistillate plants maintain function longer, requiring an investment in source tissue that will continue to produce assimilate throughout the development of embryos and fruits.
An increase in the carbohydrate concentration in the fruits or pollen-producing flowers up to a peak probably reflects the maturation of those sinks. Starch concentration continued to increase in anthers in the last week of staminate flower development, after pollen development has ceased and pollen been shed from most of the anthers, and little purpose remains for carbohydrate allocation to the staminate flowers. Clearly once the shift in resources has occurred, the allocation pattern continues despite a lack of obvious function for the photosynthate at this stage.
The carbohydrate content of leaves directly contradicts any suggestion that the leaves senesce because of carbohydrate starvation. However, the pistillate plants maintained low total carbohydrate concentrations for a leaf in the inflorescence through the fourth week, whereas in the fifth and sixth weeks the carbohydrate levels showed a dramatic increase, at the time of the drop in photosynthesis for the leaves on the pistillate plants. Similarly, the same leaf of the staminate plants showed a gradual but substantial increase in total non-structural carbohydrate throughout the reproductive period, until it senesced in week 5. This evidence strongly suggests that lack of total carbohydrate does not cause leaf senescence.
To the extent to which fruits serve as an explanation for the senescence of pistillate plants because of the distribution of resources to these organs, the different timing, the greater respiration rate of pollen-producing flowers, and contrasting inflorescence morphology provides ample cause for the senescence of staminate plants.
The work of Leopold et al. (1959) that examined the effect of flower removal on the senescence of staminate spinach plants has been consistently misinterpreted as a counter-example to the nutrient drain hypothesis. By showing the different morphology, timing of senescence, and high rate of respiratory activity of staminate spinach plants, the current studies demonstrate that excision of pollen-producing flowers results in the loss of a very significant sink for photosynthetic carbon, and thus a sink for any other phloem-transported resource such as nitrogen or hormonal compound. In unpollinated plants senescence did proceed even though it was delayed (Leopold et al., 1959). The removal of unpollinated pistillate flowers also delayed senescence and the removal of flowers at a younger stage resulted in a greater delay of senescence. Our results show that the young staminate flowers draw large allocations of photoassimilate. By contrast, young pistillate flowers are a relatively small sink, but, because of their size, they are active sinks on a per gram basis. The lack of the fruits as a large sink in the unpollinated or de-flowered plants examined by Leopold et al. (1959) may have caused an alteration in the pattern of resource allocation, but the early shift to support the reproductive process was apparently sufficient to lead to eventual senescence.
G2 peas allocate less photosynthate to their vegetative buds in long days, when they senesce after flowering, than in short-days when the plants continue to flower without senescing, showing the importance of resource partitioning in the mediation of senescence phenomena (Kelly and Davies, 1988a). The flowers and pods of the pre-senescent long-day plants develop far more rapidly, correlating with the greater resource allocation to reproduction, well before senescence symptoms are visible (Kelly and Davies, 1986), also illustrating the early initiation of the regulation senescence.
During senescence it has been shown that phloem-transported compounds are diverted from vegetative to reproductive sinks. This indicates that there are possible changes in these respective sinks, and also in source leaves as they transfer from being a source of photosynthate to one of remobilized compounds. Extracellular invertases are important for apoplastic phloem unloading and are key enzymes in determining sink strength (Roitsch et al., 2003). The sink activity for pollen development is because of an anther-specific extracellular invertase activity that supplies the developing microspores with carbohydrates. Engelke et al. (2010) were able to cause male sterility through the antisense repression of the anther-specific cell wall invertase, or interference with this invertase activity by the expression of a proteinaceous inhibitor. They then achieved the restoration of fertility by replacing the down-regulated endogenous plant invertase with a localized yeast invertase.
While allocation of carbohydrate resources is undoubtedly influential in the process of whole plant senescence, other changes in the metabolism of the plant must be directing that allocation. Different structures of spinach receive different amounts of photosynthate, and the amount does not depend on size alone, as shown by the analysis of partitioning data taking into account the weight of the sink tissue. The distribution of assimilate must be directed by mediating factors. Strong candidates for such mediating factors are the known plant hormones which regulate the sink capacity of the various tissues (Zhu and Davies, 1997), although other factors are possible. An essential link between phytohormone action and sink strength is that numerous hormones affect the expression of extracellular invertase (Roitsch et al., 2003). Another mediating factor could be a signalling role of sugars (Koch, 1996). Interactions between sugar and hormone signalling also play a role in the induction of senescence especially in response to stress (Wingler and Roitsch, 2008).
In G2 peas, there is a strong correlation between hormone contents and the allocation of fixed carbon. The vegetative tissues of apical buds of long-day-grown plants that are destined to senesce experience a drop in concentrations of gibberellins, while the same tissues of short-day-grown plants maintain high levels of the hormone and remain vigorous. By contrast, IAA is higher in the rapidly-growing flower buds of long-day plants than in the slower-growing short-day flower buds (Zhu and Davies, 1997). This correlates strongly with the partitioning to these structures (Kelly and Davies, 1988a), suggesting that the combination of IAA and GA direct the allocation of phloem-transported resources, resulting in differential growth and maintenance.
There is now considerable evidence that senescence may be induced by carbohydrate accumulation and not by starvation (Levey and Wingler, 2005; van Doorn, 2008), and indeed that agrees with our findings in spinach that senescing leaves had higher carbohydrate content than prior to senescence. Sugar-induced senescence of source leaves may be a signal of low nitrogen availability (Wingler et al., 2006). Paul and Foyer (2001) suggested that the carbon:nitrogen and hormonal balances of the plant regulate photosynthesis, the development of leaves, and the whole plant nitrogen distribution, leading to the sink regulation of photosynthesis as well as senescence. Glucose has been shown to cause the induction of the senescence-specific gene SAG12 by over 900-fold, and two MYB transcription factor genes induced by glucose, in turn, induced genes for nitrogen remobilization (Pourtau et al., 2006). The influence of the reproductive sink on the induction of monocarpic senescence has been suggested to be due to its ability to stimulate the nitrogen mobilization process from the source tissue through high levels of reactive oxygen species (ROS) in the source tissue leading to protein breakdown prior to mobilization (Srivalli and Khanna-Chopra, 2004). Several proteases are induced by high carbohydrate in barley leaves at the same time that senescence is accelerated (Parrott et al., 2005, 2007). The enzyme pyruvate,orthophosphate dikinase (PPDK), which is up-regulated in naturally senescing leaves, functions in a pathway that generates the phloem transported amino acid glutamine. In Arabidopsis, overexpression of PPDK during senescence can significantly accelerate nitrogen remobilization from leaves, and thereby increase rosette growth rate and the weight and nitrogen content of seeds (Taylor et al., 2010). Thus in the reallocation of phloem-transported materials towards reproductive development in both male and female spinach plants we may be dealing with the supply of nitrogen, which is clearly needed for sustained growth, although our work reported here did not investigate whether nitrogen reallocation is occurring along with photosynthate.
It is suggested that the initiation of senescence associated with a reallocation of resources to reproductive growth in monocarpic plants is under genetic control, and occurs coincident with the initiation of flowering. The transfer of a single chromosome from a perennial relative was able to confer a polycarpic growth habit to monocarpic wheat, leading to a second phase of tiller initiation after the initial flowering and fruiting was complete (Lammer et al., 2004), so that the genetic regulation of senescence most probably relies on only a few loci. Floral initiation and leaf senescence of Arabidopsis accessions are linked (Levey and Wingler, 2005), so it is possible that photoperiod controls leaf senescence through its effect on floral initiation. A QTL for whole rosette senescence in Arabidopsis co-localized with FRI, a major determinant of flowering, and interacted epistatically with a QTL where the floral repressor FLC localizes (Wingler et al., 2010). Vernalization accelerated senescence in late-flowering lines with functional FRI and FLC alleles, and rapid rosette senescence on a glucose-containing medium was correlated with early flowering and high sugar content. Not surprisingly a correlation was found between the expression of flowering- and senescence-associated genes. An additional QTL was linked to nitrogen-use efficiency. A grain protein content (GPC) locus in barley strongly influences the timing of post-anthesis flag leaf senescence, with high-GPC germplasm not only senescing early but also showing an accelerated pre-anthesis plant development (Lacerenza et al., 2010). In Arabidopsis the developing reproductive structures appeared to cause the death of the plant by preventing the regeneration of leaves and the development of additional reproductive structures (Noodén and Penney, 2001). All this is in agreement with our results here and with the previous ideas expressed by Kelly and Davies (1988b) that senescence is triggered by a physiological change very early in the reproductive period. Lacerenza et al. (2010) suggest that one of these GPC genes may be a functional homologue of Arabidopsis glycine-rich RNA-binding protein 7, which has previously been implicated in the promotion of flowering. We may, therefore, be on the verge of a more detailed analysis of the interactions between the physiological and molecular networks controlling monocarpic senescence.
Our results clearly show an early reallocation of phloem-transported fixed carbon to reproductive development, so the nutrient-diversion hypothesis can account for the induction of senescence in vegetative tissues. However, the crucial compound is clearly not carbohydrate but a part of a global shift in the hormonal and/or nutrient balance, probably including nitrogen, resulting from flowering. This would then lead to changes in gene expression associated with the cessation of growth and the development of the senescence syndrome in the vegetative tissues.
The observed diminution of the leaves in the inflorescence of spinach (which occurs more rapidly in the staminate plants), and in the apical senescence in peas (Kelly and Davies, 1988a; Zhu and Davies, 1997), can be explained by this shift. The changes in allocation, including the proximity of the apical meristem to floral sinks, which may equal or exceed the apical meristem in sink strength, must affect the meristem itself. The meristem may then decline in size (and thus produce smaller leaves) and eventually often either senesces or converts to a flower primordium. The loss of the apical bud leads to a number of physiological changes, and the inevitable senescence of the whole plant, due to the inability to produce new organs. When the balance of carbohydrates is again altered by cessation of development in the floral sinks, the resultant feed-back inhibition could cause not only a repression of photosynthesis, but leaf senescence. Thus, it is not the drain to the reproductive sinks per se, but the permanent diversion away from the development of further new vegetative sinks that may be responsible for some of the observed phenomena in whole-plant senescence. As noted by Leopold et al. (1959), once flowering is initiated, even if flowers are removed or if pistillate flowers remain unpollinated, senescence will surely follow.
Supplementary data can be found at JXB online.
Supplementary material. A full version of the Materials and methods.
Supplementary Table S1. Descriptions and average dry weight of flowering structures for plants of the given flowering stage in weeks following the onset of flowering ±SE.
Supplementary Fig. S1. Shapes and sizes of leaves from staminate and pistillate spinach plants in ascending order: leaf L1, L11, L21, L31, L41, L51, numbering from the base of the inflorescence.
Supplementary Fig. S2. Photographs of staminate (A–E) and pistillate (F–J) spinach plants at various stages of development.
This work was supported, in part, by Hatch funds administered by the New York State College of Agriculture and Life Sciences at Cornell University. DES was supported, in part, by a National Science Foundation Fellowship. We thank Jeffrey Melkonian, David Wolfe, Marvin Pritts, Tim Setter, and Brian Flanagan for assistance and the use of their equipment, and the Asgrow Seed Company for donation of the spinach seed.