Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Extracellular matrix wikipedia , lookup
Cell membrane wikipedia , lookup
Cell culture wikipedia , lookup
Cytokinesis wikipedia , lookup
Cellular differentiation wikipedia , lookup
Endomembrane system wikipedia , lookup
Programmed cell death wikipedia , lookup
Signal transduction wikipedia , lookup
Journal of Experimental Botany, Vol. 59, No. 3, pp. 453–480, 2008 doi:10.1093/jxb/erm356 FOCUS REVIEW PAPER Physiology and molecular biology of petal senescence* Wouter G. van Doorn†,‡ and Ernst J. Woltering Wageningen University and Research Centre, PO Box 17, 6700 AA Wageningen, The Netherlands Present address: Mann Laboratory, Plant Science Department, University of California, Davis, One Shields Avenue, Davis, CA 95616, USA y Received 7 September 2007; Revised 16 December 2007; Accepted 17 December 2007 Abstract Petal senescence is reviewed, with the main emphasis on gene expression in relation to physiological functions. Autophagy seems to be the major mechanism for large-scale degradation of macromolecules, but it is still unclear if it contributes to cell death. Depending on the species, petal senescence is controlled by ethylene or is independent of this hormone. EIN3-like (EIL) transcription factors are crucial in ethylene-regulated senescence. The presence of adequate sugar levels in the cell delays senescence and prevents an increase in the levels of EIL mRNA and the subsequent up-regulation of numerous senescence-associated genes. A range of other transcription factors and regulators are differentially expressed in ethylene-sensitive and ethyleneinsensitive petal senescence. Ethylene-independent senescence is often delayed by cytokinins, but it is still unknown whether these are natural regulators. A role for caspase-like enzymes or metacaspases has as yet not been established in petal senescence, and a role for proteins released by organelles such as the mitochondrion has not been shown. The synthesis of sugars, amino acids, and fatty acids, and the degradation of nucleic acids, proteins, lipids, fatty acids, and cell wall components are discussed. It is claimed that there is not enough experimental support for the widely held view that a gradual increase in cell leakiness, resulting from gradual plasma membrane degradation, is an important event in petal senescence. Rather, rupture of the vacuolar membrane and subsequent rapid, complete degradation of the plasma membrane seems to occur. This review recommends that more detailed analysis be carried out at the level of cells and organelles rather than at that of whole petals. Key words: Autophagy, cell death, cell wall, gene expression, hormones, lipids, nucleic acids, pathogens, proteins, reactive oxygen species, regulation, remobilization, senescence, transcriptions factors, ultrastucture. Introduction This review is intended to give an update on the literature on petal senescence. It takes information from a number of recent papers on changes in gene expression, and relates it to the events that one might anticipate, in particular those that have been suggested by previous physiological and biochemical studies. This review will first discuss a model of petal senescence, then consider the underlying data regarding ultrastructure, the remobilization of macromolecules, transport of mobile compounds out of the petal, and the role of autophagy in cellular degradation. This review also pays attention to some issues of contention such as the localization of membrane degradation and the possibility that reactive oxygen species (ROS) contribute to cellular death. There has been some discussion in the literature about the distinction between senescence and programmed cell death (PCD). At least three main categories of PCD might be distinguished in plants: developmental, pathogen induced, and abiotic stress induced. Petal senescence is considered to be a subset of developmental PCD, and in the case of petal senescence the term ‘senescence’ is considered to be a synonym of ‘PCD’. The term ‘senescence’ will here be used, except when emphasis is placed on the actual death of cells. Most flowers have a whorl of green sepals and another whorl of variously coloured petals. In contrast, flowers from genera such as Alstroemeria, Hemerocallis, and Iris have two whorls of variously coloured flower leaves, called tepals. For the sake of easy reading, the tepals of these flowers will here * This paper is dedicated to the memory of our colleague Abraham Halevy, who passed away in 2006. Abe kindly suggested one of us (WvD) that we should produce the present review. z To whom correspondence should be addressed. E-mail: [email protected] ª The Author [2008]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved. For permissions, please e-mail: [email protected] 454 van Doorn and Woltering also be called petals, although sometimes reference will be made to the whorl from which they originated. Relationship between petal senescence, abscission, and ethylene Depending on the species, petal senescence is visibly shown by wilting or withering. Petal wilting is due to loss of turgor. Withering is a colour change and slow dehydration. There seems to be no difference in the mechanisms underlying wilting and withering. In most species with flowers showing visible senescence symptoms, the petals will abscise, whereby the time between visible symptoms and abscission is quite variable. In some species, for example carnation, abscission occurs much later than visible senescence. In other flowers, for example Tulipa and Alstroemeria, visible senescence and abscission occur at about the same time. The onset of senescence occurs long before the time of wilting and abscission of Alstroemeria petals (Breeze et al., 2004). Petals of many other species abscise when still fully turgid. It depends on the species whether senescence processes in these abscising petals have started or not (Yamada et al., 2007b). Another category consists of flowers with petals that show wilting because of water stress in the entire cut flowering stem. In many rose cultivars, for example, the petals of uncut flowers abscise without senescence symptoms. In contrast, the petals of these flowers often wilt due to an occlusion in the xylem, which blocks the flow of water to the petals. Such petal wilting is not due to senescence and will not be discussed here. In one group of flowers, petal senescence is highly ethylene sensitive, and regulated by ethylene, whilst in another group it is ethylene insensitive, and not regulated by this molecule. This difference was found to correlate with plant families or subfamilies (Woltering and van Doorn, 1988; van Doorn, 2001). The role of ethylene in hastening petal senescence seems to have evolved naturally as a mechanism to allow pollination-induced changes. Rapid petal wilting (but also petal abscission and changes in petal colour, insofar as advanced by ethylene) seem to direct pollinating animals to virgin flowers, thus increasing pollinator efficiency (van Doorn, 1997). Species often used in studies on petal senescence include carnation (Dianthus caryophyllus), Petunia hybrida, and orchids (ethylene-sensitive senescence), as well as Mirabilis jalapa, Hemerocallis hybrids (closely related to H. fulva), and Iris3hollandica (ethyleneinsensitive senescence). A model of petal senescence Based on our interpretation of the literature, the following picture of petal senescence is constructed. Numerous genes become (transiently) up-regulated during senescence, whilst many other genes are down-regulated. Most of the changes in gene expression are related to remobilization of macromolecules, and transport of the mobile compounds out of the petal. A considerable part of differential gene expression is related to plant defence. No genes have as yet been identified that are specific for cellular death. Degradation of macromolecules in the senescent cell is mainly due to autophagic processes in the vacuole, in addition to protein degradation in mitochondria, nuclei, and the cytoplasm, fatty acid breakdown in peroxisomes, and degradation of nucleic acids in the nuclei. It is as yet unknown how the onset of senescence is regulated. This is true for the onset of the increase in ethylene production in systems that are controlled by ethylene as well as for the regulation in species where senescence is not under the control of ethylene. Although many transcription factors and signal transduction factors are differentially expressed during petal senescence, as yet the functional role is known of only a few (e.g. EIL3 and Aux/IAA). It is also not clear how the petal cells die. Death is preceded by the disappearance of most of the cellular contents, and might therefore be due to ongoing remobilization, but might also occur independently of the remobilization process. Several factors that control programmed death of animal cells, such as caspases, proteins released by mitochondria, and proteins that control the release of compounds through membranes, have no apparent counterpart in plant cells, indicating that the regulation of cell death during petal senescence is quite different from the regulation of cell death in animal cells. Tonoplast rupture (or lack of tonoplast semi-permeability) is widely held to be the immediate cause of developmental cell death in plants, but the data that support this contention in petals are still weak. The mechanism that accounts for the rupture is completely unknown. The ovary is central in the regulation of ethylene-sensitive senescence, at least in carnation flowers. It mediates the hastening effect of pollination on senescence. After pollination the ovary produces ethylene which induces a rapid increase in petal ethylene production. In the absence of pollination, the ovary is also causal in petal senescence, although the mechanism by which ethylene is then produced in the ovary is not clear. This picture shows the many lacunae in our present understanding. Some testable hypotheses are put forward (in the Discussion section) regarding these gaps: (i) the regulation of the onset of senescence; (ii) the determination of the species-dependent petal life span; (iii) the mechanism whereby the rise in ethylene synthesis is produced; (iv) the role of hormones in the regulation of the onset of ethyleneinsensitive senescence; (iv) the relationship between remobilization and collapse of the tonoplast; and (vi) the mechanism of tonoplast collapse. Petal senescence 455 The data underlying the contentions of the above model of petal senescence will be discussed in the remainder of this review. Morphology and ultrastructure of dying petal cells One of the earliest changes in cell ultrastructure, at least during senescence of Iris petals, is the closure of the plasmodesmata. Plasmodesmata, if open, allow transfer of relatively small molecules, such as sugars, hormones, and RNA molecules, between neighbouring cells. If the plasmodesmata are closed, this transport is halted. In the outer petals of Iris the plasmodesmata of mesophyll cells close 2 d before those of the epidermis cells. The mesophyll cells also die 2 d earlier than the epidermal cells. Compared with epidermis cells, the mesophyll cells also exhibited all other ultrastructural changes, indicative of processes leading to cell death, ;2 d earlier (van Doorn et al., 2003). Although these data are suggestive of a role for plasmodesmata in the early regulation of senescence, the data thus far only show a correlation. Invaginations have been noted at the tonoplast of Ipomoea (¼ Convulvulus) tricolor petal cells; numerous vesicles were found in the vacuole. These vesicles are probably due to microautophagy and/or macroautophagy. The vacuoles also contained structures reminiscent of mitochondria. The vacuole therefore seems a main site of membrane and organelle degradation (Winkenbach, 1970b; Matile and Winkenbach 1971; Phillips and Kende, 1980; Smith et al., 1992). Petal cells in Ipomoea (Winkenbach, 1970a), carnation (Smith et al., 1992), Hemerocallis (Stead and van Doorn, 1994), and Iris (van Doorn et al., 2003) exhibit a decrease in the number of small vacuoles, and an increase in vacuolar size. The ongoing increase in vacuole volume in Iris was accompanied by the loss of a considerable part of the cytoplasm and with the disappearance of most organelles. In Iris, one of the earliest ultrastructural senescence symptoms (after closure of the plasmodesmata) was the loss of most of the endoplasmic reticulum (ER) and attached ribosomes. Several other organelles, such as Golgi bodies, then also become less frequent. Many mitochondria became degraded, but some remained until vacuolar collapse. Similar changes have been reported in carnation (Smith et al., 1992). The nucleus has been found to remain until a late stage of senescence, but shows several ultrastructural changes. In tobacco petals the nuclei showed nuclear blebbing, similar to that observed during apoptosis in animal cells (SerafiniFracassini et al., 2002). Nuclei in young cells have a rather even distribution of DNA (chromatin). During cell senescence, the chromatin clumps into patches that are found either throughout the nucleus or mainly at the nuclear periphery. In several species the nucleus showed brighter fluorescence after staining the DNA with a fluoroprobe, indicating DNA condensation. This is often accompanied by a decrease in nuclear diameter (Yamada et al., 2006a, b). In the petals of Ipomoea, a doubling of the number of DNA masses (‘nuclei’) per petal was found, together with a continuing decrease in the diameter of these masses (Yamada et al., 2006a). As the DNA masses contained an outer membrane, the results indicate nuclear fragmentation. Evidence for nuclear fragmentation was also found in the petals of Petunia hybrida and Argyranthemum frutescens, but not in petals of Antirrhinum majus (Yamada et al., 2006b). These data, although still limited, indicate nuclear fragmentation in the senescence of petals both of some ethylene-sensitive plants (Ipomoea; Petunia) and of an ethylene-insensitive plant (Argyranthemum). Other organelles also disappear and the volume of cytoplasm becomes very low, leaving only a small strip next to the cell wall. At least some mitochondria, and in carnation some plastids, stayed until a late phase of senescence (Winkenbach, 1970b; Smith et al., 1992; van Doorn et al., 2003). When the cytoplasm of the cell is almost or completely empty, electron microscopy (EM) work shows that the vacuolar membrane (tonoplast) collapses and then disappears. This occurs earlier than the collapse and disappearance of the plasma membrane (Winkenbach, 1970b; Smith et al., 1992; van Doorn et al., 2003). Although it cannot be excluded that this is an artefact of EM preparation, the earlier collapse of the tonoplast might be a real effect, similar to that in other plant cells, for example during the PCD of tracheal elements (Obara et al., 2001), and in several other examples of PCD/ senescence (Jones, 2001; Jones and Dangl, 2004). It will here be assumed that the earlier tonoplast collapse in petal cells is real, and the direct cause of cellular death, in line with the analysis of such effects in several other systems. In several species tested, the petal mesophyll cells died considerably earlier than the epidermis cells. This was observed in the dicotyledonous Ipomoea (Winkenbach, 1970b), and in the monocotyledonous Iris reichenbachii (Bancher, 1941), Iris3hollandica (van Doorn et al., 2003), Hemerocallis (Stead and van Doorn, 1994), Alstroemeria (Wagstaff et al., 2003), and Sandersonia (O’Donoghue et al., 2002). In contrast, no clear evidence for earlier death in the mesophyll, compared with the epidermis, was found in Gypsophila (Hoeberichts et al., 2005). In some species, cell death starts at the apical part and gradually moves towards the petal base (Iris, van Doorn et al., 2003; Alstroemeria, Wagstaff et al., 2003), but again no evidence for such a difference was found in Gypsophila (Hoeberichts et al., 2005) and it was also apparently not observed in carnation (Smith et al., 1992). Visible petal wilting coincided with collapse of the epidermis cells (carnation, Smith et al., 1992; Iris, van Doorn et al., 2003; Gypsophila, Hoeberichts et al., 2005). 456 van Doorn and Woltering Collectively, the data indicate that petal senescence is accompanied by large-scale autophagy. A considerable part of the cytoplasm is lost, and most organelles disappear, including the ER and ribosomes. Some mitochondria and the nucleus remain until late in the process but the nucleus usually changes in morphology relatively early during senescence. The evidence also indicates that the tonoplast collapses earlier than the plasma membrane, and that the collapse of the tonoplast is the actual moment of cellular death. The ultrastructure of petal senescence is similar in petals where the process is regulated by ethylene and in petals where it is not. Is petal death due to autophagy? The literature indicates an upsurge of autophagy both during starvation and during senescence. This follows from ultrastructural evidence, biochemical data, and gene expression data (Thompson and Vierstra, 2005). Autophagy during senescence seems to be a main process responsible for cell dismantling and remobilization of compounds that can be used in other parts of the plants. Autophagy may not necessarily be required for cell death. In PCD in animals, a few examples are now known where death was apparently due to autophagy (Verheye et al., 2007), but such evidence has not as yet been reported for plants. In Arabidopsis, a number of autophagy genes (ATG) have been knocked out, which resulted in hastened leaf yellowing (Hanaoka et al., 2002; Yoshimoto et al., 2004; Thompson et al., 2005; Xiong et al., 2005). This might be interpreted to indicate that death is not due to autophagy, but is rather delayed by autophagy. However, such an interpretation might be too hasty. The experiments in which autophagy genes were silenced showed pleiotropic effects, such as a small root system. Autophagy occurs at all stages of cell development. It is a mechanism to dispose of organelles that are no longer functional. The lack of normal baseline autophagy can lead to stress, perhaps shown as precocious leaf yellowing. Similarly, the presence of many dysfunctional mitochondria resulted in hastened petal senescence, probably because of stress (Chen et al., 2005). Since the root system of the transgenic Arabidopsis plants was smaller, it is possible that nutrient uptake was lower, or that the root produced fewer cytokinins and/or gibberellins. The experiments on ATG silencing, as carried out thus far, thus provide no answer to the question of what the relationship is between the upsurge in autophagy and leaf yellowing. Only experiments in which autophagy becomes attenuated during senescence, for example by silencing under the control of a senescence-specific promoter, can provide the necessary information. Autophagy, as discussed above, is a combination of at least micro- and macroautophagy (Thompson and Vierstra, 2005). Microautophagy entails invagination of the vacuolar membrane, transferring parts of the cyto- plasm, including organelles, to the vacuole. Macroautophagy is due to double membrane structures formed inside the cytosol, sequestering parts of the cytosol, including organelles. Autophagosomes merge with the vacuole. Prior to delivery of their contents to the vacuole, autophagosomes are also considered to merge with vesicles containing hydrolytic enzymes, thus the onset of degradation can already take place inside the autophagosomes (Inoue and Morioyasu, 2006). In addition to micro/ macroautophagy, the cell death programme induces the collapse of the tonoplast. Hydrolytic enzymes in the vacuole are released into the cell and will rapidly degrade most of what is left by that time, including the plasma membrane. After the collapse of the tonoplast, the cell can no longer regulate its degradation. Tonoplast collapse, therefore, seems to be the moment of cell death. The term mega-autophagy has been suggested for tonoplast collapse and its consequences (van Doorn and Woltering, 2005). Micro- and macroautophagy might progress, theoretically, to a stage whereby essential components disappear. Tonoplast collapse might be the result of such ongoing micro-/macroautophagy, although this event might also be regulated independently of autophagy. Even if tonoplast collapse is not due to micro/macroautophagy, the two processes must be closely co-ordinated because remobilization and nutrient transport out of the petals will be severely compromised if death occurs too early. Regulation of senescence at the transcriptional, translational, and post-translational level Senescence is regulated by levels of mRNA and proteins, and by activation of proteins. mRNA levels can be regulated by short interference RNA (siRNA) and microRNA (miRNA). A limiting factor can also be mRNA transport out of the nucleus. Proteins can be modified after translation. The eukaryotic translation initiation factor 5A (eIF-5A) is an example of post-translational modification. eIF-5A is apparently involved in translocation of specific mRNAs from the nucleus to the cytoplasm. During carnation petal senescence, the RNA abundance of eIF5A increased, as well as the mRNA abundance of a gene involved in post-translational modification of eIF-5A, deoxyhypusine synthase. eIF-5A is unique among translation initiation factors because it contains hypusine [N(4-amino-2-hydroxybutyl) lysine], which is formed post-translationally via the transfer and hydroxylation of a butylamino group to a specific lysine residue. These events activate eIF-5A. Deoxyhypusine synthase mediates the first step in the synthesis of hypusine in eIF-5A (Hopkins et al., 2007). Antisense suppression of deoxyhypusine synthase delayed senescence in Arabidopsis leaves (Wang et al., 2003) and in carnation petals (Hopkins et al., 2007). These results are reminiscent of the delay of petal senescence by inhibitors of translation such Petal senescence 457 as cycloheximide and narciclasine (van Doorn et al., 2004), showing the importance of de novo protein synthesis in senescence. This example illustrates the potential complexity of transcriptional, post-transcriptional, translational, and post-translational control of senescence. Post-translational modifications also occur, for example, in signal transduction (protein phosphorylation by kinases), during autocatalytic activation of cysteine proteases in the vacuole, and prior to protein degradation in proteasomes. Targeting for breakdown occurs through the transfer of several ubiquitin moieties to a protein, by a ubiquitin ligase complex. Senescence, therefore, can be regulated at the level of protein synthesis, protein activation, and protein degradation. Transduction of the death signal: comparison with animal cells In animal cells, several elements of the cell death signal transduction routes have been characterized. The various elements can be grouped, roughly, into caspase signal transduction, and the release of signal chemicals from mitochondria, the ER, and the lysosome (Cheng et al., 2006; Hail et al., 2006). Caspases are cysteine endoproteases that cleave at aspartate residues. Although several caspases are part of the signal transduction pathway, some are executioner proteins, responsible for the degradation of cellular substrates. Thus far, no caspase homologues have been found in plant genomic databases, but some plant caspaselike proteins such as the vacuolar processing enzyme (VPE) have been found. Although there is limited overall sequence identity between VPE and caspase-1, they share several structural properties. The activity of isolated VPEs was blocked by caspase-1 inhibitors (Rojo et al., 2003). Arabidopsis has four VPE genes (a to d). A VPEc was found to localize to vesicles that merged with vacuoles. In the vacuole, this VPEc was autocatalytically converted into the active form, which regulated maturation/activation of vacuolar proteases. A VPEc mutant showed 20-fold increased levels of cystatin, a cystein protease inhibitor. Homologues of cystatin suppress hypersensitive cell death in plant and animal systems (Rojo et al., 2003, 2004). Additionally, in a VPEd-deficient mutant, cell death of the two layers of the seed coat was delayed (Nakaune et al., 2005). The mRNA abundance of two VPE genes increased during petal senescence in Narcissus. One of the two was previously described as an asparaginyl endopeptidase, also known as legumain (Hunter et al., 2004a). A gene encoding a VPE was also up-regulated during carnation petal senescence (Hoeberichts et al., 2007). In addition to caspase-like enzymes, a class of cysteine proteases in plants are called metacaspases (HaraNishimura et al., 2005). A metacaspase was required for programmed death in suspensor cells (Bozhkov et al., 2004), and a plant metacaspase has been reported to induce cell death in yeast (Watanabe and Lam, 2005). Michelle Jones (personal communication, 2006) found an increase in mRNA abundance of a metacaspase in Petunia petals. However, no effect was observed in petal senescence of a Petunia line harbouring a metacaspase antisense gene. Another type of caspase-like protease is localized to the plant cell wall. Its activity was inhibited by chemicals that block either serine proteases or cysteine proteases. To distinguish this proteinase from caspases, which are not inhibited by serine protease inhibitors, it was called saspase. A saspase has been implicated in cell death of plants treated with the fungal toxin victorin (Coffeen and Wolpert, 2004). Saspase activity has apparently not been reported in relation to plant senescence. In animal cells, signals for cell death can be released from various organelles, such as the ER, the lysosome (vacuole), and mitochondria. It is still unclear whether release of factors from these organelles is causal in plant cell death. Release of cytochrome c from mitochondria, which is important in many examples of animal PCD, was not found during senescence in Petunia petals (Xu and Hanson, 2000). In animal cells, several other signal factors are released from mitochondria, but their possible role has not yet been reported in petals. In animal cells, Bax is one of the pro-apoptotic proteins of the Bcl-2 family. These proteins insert into membranes. This results in the release of cell death signal molecules. Bax-inhibitor-1 is also implicated in the regulation of animal cell death. A homologue of Bax-inhibitor-1 is present in plants. As no homologues of Bcl-2 proteins have been found in plants, the role of the plant Baxinhibitor-1 homologue is unclear. In animal cells, the mRNA abundance of a gene called DAD1 (defender against apoptotic cell death) substantially decreased prior to death. Overexpression of the gene delayed the death of these cells. The function of DAD1 is unclear; it has been suggested that the cell dies by lack of DAD1 activity because DAD1 is a household enzyme involved in glycosylation of proteins at ER membranes (van der Kop et al., 2003). DAD1 was also shown to interact with the Bcl-2 family proteins in animal cells (Makishima et al., 2000), which might suggest a regulatory role in cell death. In petals, thus far, only correlative evidence is available between senescence symptoms and DAD1 expression. The decrease in DAD1 abundance occurred relatively early during senescence in some species, but very late in other species (Table 1). The role of DAD1 in senescence, if any, is therefore as yet unclear. Free calcium ions may be a factor that is common to the regulation of cell death in plants and animals. In animals, a death signal from outside the cell can be transduced through an increase in inositol-1,4,5-triphosphate (IP3) levels. This molecule can induce an increase in 458 van Doorn and Woltering Table 1. Decrease of DAD1 transcript abundance during petal senescence Species Comments Reference Alstroemeria pelegrina Dianthus caryophyllus Gladiolus3 grandiflora Iris3hollandica Pisum sativum Decrease relatively early Decrease very late Wagstaff et al. (2003) van der Kop et al. (2003) Decrease occurs late Yamada et al. (2004) Decrease very late Treatments hastening or delaying senescence also hastened or delayed decrease in mRNA abundance, respectively van der Kop et al. (2003) Orzaez and Granell (1997) intracellular Ca2+ concentration. Cell death in pollen tubes depended on an increase in calcium ion levels, which resulted in a decrease of actin polymerization. Inducing a major decrease of actin polymerization, similar to the one occurring prior to normal death, was adequate to induce death. It was concluded that calcium-induced actin depolymerization is an early regulator of death in these cells (Thomas et al., 2006). A role for calcium ions in petal senescence has thus far not been established. A possible connection with actin might follow from work on senescence in Ipomoea petals, where the mRNA abundance of a gene encoding ataxin-2 transiently increased (Yamada et al., 2007a). In yeast, increased mRNA abundance of ataxin-2 caused premature cell death due to defects in actin filament formation (Satterfield et al., 2002). During senescence in carnation petals, an early rise occurred in transcript levels of a putative CBL-interacting protein kinase (CIPK), a regulator of phosphorylation cascades that transmit calcium signals. This rise was advanced by ethylene and delayed by silver thiosulphate. It was hypothesized, therefore, that a calcium signal is involved in carnation petal senescence, and that this signal is regulated by endogenous ethylene (Hoeberichts et al., 2007). In Iris petals, the transcript abundance of a gene encoding a cyclic nucleotide-gated ion channel (CNGC) increased (van Doorn et al., 2003). Some CNGCs are involved in calcium signal transduction (Arazi et al., 2000) and in increasing the internal calcium ion concentration (Köhler et al., 2001). In Arabidopsis, a GNGC has been suggested to be involved in developmental cell death (Köhler et al., 2001) and was required for ozone-induced cell death (Overmyer et al., 2005). The Iris transcript shows very high homology with this gene. During Iris petal senescence an increase was seen in the transcript abundance of a gene encoding a casein kinase (i.e. CK-type protein kinase), most of which are apparently signal transduction factors (van Doorn et al., 2003). Some CK proteins have been suggested to be regulators of the cell survival pathway (Ahmed et al., 2002). Taken together, it has presently not been established that petal senescence is regulated by any of the signal transduction mechanisms that are known from animal PCD. Although some of the genes found in petals might be involved in the regulation of senescence, their function is as yet unknown. Transcription factors and transcriptional regulators Transcription factors are usually classified according to specific functional amino acid sequences (domains). More than one type of such domains can be present in a protein. Table 2 gives an overview of transcription factors that have hitherto been found to be differentially expressed during petal senescence. It is to be expected that several more transcription factors will be discovered that are associated with petal senescence, similar to the long list of known transcription factors related to leaf senescence (Guo et al., 2004). With the exception of EIL, AP2, Aux/IAAs, NAC, knotted1, and GCIP, little is known about the role, in senescence, of the transcription factors shown in Table 2. EIN3 and EIN3-like (EIL) transcription factors are crucial in the ethylene signal transduction chain. Thus far, four EIL genes have been isolated in carnation petals. When carnation flowers were treated with ethylene, the petal mRNA abundance of Dc-EIL1-2 and Dc-EIL4 slightly decreased, whereas the mRNA abundance of Dc-EIL3 drastically increased. Ethylene promoted and STS delayed the increase in Dc-EIL3 mRNA, and had no effect on Dc-EIL1. These data indicate that Dc-EIL3 is an important regulator of gene expression during carnation petal senescence. It might act as a master switch of ethyleneinduced gene expression (Iordachescu and Verlinden, 2005; Hoeberichts et al. 2007). Targets of EIN/EIL transcription factors include genes with an ethylene-responsive element (ERE) in their promoter. Amongst these is a gene encoding the ethyleneresponsive transcription factor 2 (ERF2), which belong to the AP2 transcription factor family. In senescing daffodil petals, a gene (DAFSAG9) encoding an AP2 protein with very high homology to ERF2 was highly up-regulated (Hunter et al., 2002). Aux/IAA genes are transcriptional suppressors, also often considered transcription factors. Aux/IAA is central in the regulation of auxin responses. In the absence of auxin, Aux/IAA inhibits all auxin-responsive genes. When auxin is present, it binds to its receptor, located on an F-protein in a nuclear-localized ubiquitin ligase. This ligase then targets Aux/IAA for degradation in the proteasome. This results in the expression of auxinresponsive genes (Bishopp et al., 2006). The protein sequence of the NAC in carnation petals was highly homologous to that of NAC2 of Arabidopsis. NAC Petal senescence 459 Table 2. Differential expression, during petal senescence, of genes putatively encoding transcription factors or transcriptional regulators bHLH, basic helix–loop–helix; bZip, basic-leucine zipper; GCIP, Grap2- and cylin-interacting protein; HD, homeodomain; HLH-Zip, helix–loop– helix leucine zipper Factor Transcription factors related to ethylene signalling Dc-EIL1 Dc-EIL1-2 Dc-EIL3 Dc-EIL4 AP2/ERF Transcription factors and transcriptional regulators not directly related to ethylene signalling Aux/IAA bHLH/MYC bZIP C2H2 HLH-Zip HD HD-Zip HMG box JUMONJI MADS MYB NAC a Species Comments Reference Dianthus caryophyllus Dianthus caryophyllus Dianthus caryophyllus Dianthus caryophyllus Narcissus pseudonarcissus Waki et al. (2001) Iordachescu and Verlinden (2005) Iordachescu and Verlinden (2005) Iordachescu and Verlinden (2005) Hunter et al. (2002) Dianthus caryophyllus Dianthus caryophyllus Dianthus caryophyllus Mirabilis jalapa Alstroemeria pelegrina Mirabilis jalapa Iris3hollandica Dianthus caryophyllus Mirabilis jalapa Alstroemeria pelegrina Mirabilis jalapa Iris3hollandica Dianthus caryophyllus Mirabilis jalapa Dianthus caryophyllus Mirabilis jalapa Dianthus caryophyllus Narcissus pseudonarcissus Hoeberichts et al. (2007) Hoeberichts et al. (2007) Hoeberichts et al. (2007) Xu et al. (2007) Breeze et al. (2004) Xu et al. (2007) van Doorn et al. (2003) Hoeberichts et al. (2007) Xu et al. (2007) Breeze et al. (2004)a Xu et al. (2007) van Doorn et al. (2003) Hoeberichts et al. (2007) Xu et al. (2007) Hoeberichts et al. (2007) Xu et al. (2007) Hoeberichts et al. (2007) Hunter et al. (2002) GCIP Knotted-1 type Data from Supplement 3, for function see Martı́nez-Garcı́a and Quail (1999). factors (>100 in Arabidopsis) have diverse functions, such as organ separation, but a role for a NAC in senescence has not yet been described. AtNAC2 mRNA abundance was increased after treatment with the ethylene precursor 1aminocyclopropane-1-carboxylic-acid (ACC), by abscisic acid (ABA), and by an auxin (He et al., 2005). Knotted1 is a homeobox gene found in apical meristems. Overexpression of this gene yielded phenotypes similar to those in which the cytokinin gene ipt is overexpressed. Knotted1 mRNA was not found in tobacco leaves undergoing senescence, but when knotted1 was overexpressed, driven by the promoter of a senescencespecific cysteine protease (SAG12), it delayed leaf yellowing. This was associated with a 15-fold increase in cytokinin levels (Ori et al., 1999). These results suggest that the knotted1 found in carnation petals (Hoeberichts et al., 2007) is related to cytokinin metabolism. In senescing Iris petals, an up-regulated gene encoded a plant homologue of the animal Grap2- and cylininteracting protein (GCIP; van Doorn et al., 2003). GCIP is a helix–loop–helix leucine zipper family transcription factor that inhibits E2F1-mediated transcriptional activity The transcription factor E2F1 is a member of a family involved in cell cycle progression, differentiation, and cell death (Chang et al., 2000; Ma et al., 2007). In addition to the transcription factors mentioned (Table 2), WRKY factors might also be regulators of petal senescence. The mRNA abundance of a gene encoding a RING domain ankyrin repeat protein, probably (part of) a ubiquitin ligase, drastically increased during senescence in Mirabilis petals. Silencing of the gene in Petunia petals delayed petal wilting by a few days. The promoter of this gene had putative binding sites for bZip, HD-Zip, Myb, MADS, and also for WRKY (Xu et al., 2008). The data show that many transcriptional regulators become differentially expressed during petal senescence. However, these factors have as yet not been shown, for example by gene silencing or overexpression, to function as regulators of petal senescence. Ethylene receptors and signal transduction factors Receptors of the ethylene signal In Arabidopsis, ethylene is perceived by a family of five receptors (ETR1, ETR2, ERS1, ERS2, and EIN4; Mason and Schaller, 2005). The expression of genes encoding receptor proteins has been reported in petals of only a few species. In carnation, for instance, considerable amounts 460 van Doorn and Woltering of ETR1 and ERS2 mRNA were present in petals at the time of harvest. Exogenously applied ethylene hastened senescence but did not clearly change the mRNA abundance of ETR1 or ERS2 (Shibuya et al., 2002). Interestingly, two homologues of Arabidopsis ERS1 were isolated from the petals of Gladiolus which show ethylene-insensitive senescence. The mRNA abundance of one of these genes changed during petal senescence, whereas the other was highly and constitutively expressed (Arora et al., 2006). Overexpression of mutated ethylene receptor genes in petals showed the importance of these receptors in senescence (Table 3). The data show that senescence is delayed when a single class of ethylene receptors is expressed at low levels or is absent. CTR1 and EIN2 Signal transduction is usually taken to comprise the pathway from a receptor to a change in the levels of transcription factors. Part of the signal transduction pathway of ethylene has by now been elucidated (Bishopp et al., 2006; Etheridge et al., 2006). The pathway is relatively complex due to the presence of negative regulators. First, the ethylene receptors are negative regulators because ethylene binding stops their action. In the absence of ethylene, the receptors activate a downstream CTR1 protein. Secondly, active CTR1 is a negative regulator as it inhibits EIN2 and thereby prevents ethylene’s effects. In the presence of ethylene the receptors do not activate CTR1, which thus cannot block EIN2. This results in ethylene-related gene expression. The importance of EIN2 in petal senescence was shown in knockout experiments in which reduced EIN2 mRNA abundance produced a longer life in (attached) Petunia flowers (Shibuya et al., 2004). Active EIN2 regulates the levels of EIN3 and EIL transcription factors, by inhibiting their degradation in proteasomes (Bishopp et al., 2006). Ethylene regulation of petal senescence: beyond EIL Induction of the rise in ethylene production The onset of visible wilting in ethylene-sensitive petal senescence is accompanied by a sudden, transient increase Table 3. Delay of petal senescence in transgenic plants with mutated ethylene receptor genes Inserted gene Species Reference Mutated Mutated Mutated Mutated Petunia hybrida Dianthus caryophyllus Nemesia strumona Petunia hybrida Shaw et al. (2002) Bovy et al. (1999) Cui et al. (2004) Wilkinson et al. (1997); Jones et al. (2005) ERS ETR1 ETR1 ETR1 of respiration, associated with a transient upsurge in ethylene production. Pollination hastens these events. A signal related to pollination (reviewed elsewhere: O’Neill, 1997) causes the surge in ethylene production. A short treatment with ethylene can replace pollination (Nichols, 1966). It is not known what induces the rise in ethylene production in unpollinated flowers. The ethylene synthesis pathway starts with Sadenosylmethionine (SAM), which is converted to ACC, the direct precursor of ethylene. Genes in this pathway, such as SAM synthase, ACC synthase, and ACC oxidase, are up-regulated during ethylene-sensitive petal senescence (Jones, 2004; Hoeberichts et al., 2007). Preventing the conversion of ACC to ethylene, by decreasing the level of ACC oxidase mRNA through antisense techniques, prolonged the life of cut carnation flowers (Savin et al., 1995; Kosugi et al., 2002) and increased the life span of plant-attached flowers in Torenia fournieri (Aida et al., 1998). Extended flower longevity has also been reported in transgenic carnation flowers after decreasing the abundance of ACC synthase mRNA, using cosuppression (Michael et al., 1993). ten Have and Woltering (1997) excised petals and other organs of unpollinated carnation flowers at various times during vase life. The expression of genes encoding ACC synthase and ACC oxidase occurred first in the ovary, then in the stigma plus style (which cannot easily be separated in this species), and only last in the petals. This indicated (but did not show) that the ovary is the trigger of petal senescence. The mobile compound that brought the signal from the ovary to the petals was apparently ethylene (ten Have and Woltering, 1997). Strong evidence for an important role for the gynoecium (comprising stigma, style, and ovary) in carnation petal senescence was obtained after finding a method for careful gynoecium removal. In previous experiments, the gynoecium had been cut off. This apparently induced enough wounding ethylene to confound the gynoecium effect, and hasten senescence. Removal of the gynoecium by hand, in contrast, delayed senescence. In the absence of the gynoecium, petal ethylene did not reach normal high levels. Application of ACC, the ethylene precursor, in the vase water induced substantial ethylene production and early petal wilting in flowers from which the gynoecia had not been removed, but the treatment had little effect on ethylene production and did not induce petal wilting in flowers without gynoecia. This shows that ethylene generated in the gynoecium is the trigger for the onset of the increase in ethylene production in carnation petals (Shibuya et al., 2000; Nukui et al., 2004). It has been suggested that the rise in ethylene production in unpollinated flowers with ethylene-sensitive petal senescence occurs because the cells become more sensitive to their basal ethylene production, which in turn might relate to a decrease in cytokinin activity in the Petal senescence 461 petals. In a line of transgenic Petunia flowers, petal life was extended by overexpression of a gene encoding isopentenyltransferase (ipt), which results in high levels of several cytokinins. The mRNA abundance of a cysteine protease gene, indicative of ongoing senescence, remained very low for a considerable period of time when detached ipt-overexpressing flowers were treated with ethylene (Chang et al., 2003). These data strongly suggest that cytokinins regulate petal life span in vivo. In carnation, a relationship was found between petal cytokinin levels and the time to senescence, when comparing the shorter-lived detached flowers with flowers that remained uncut (van Staden and Dimalla, 1980). However, cytokinin levels in petals of unpollinated carnation flowers did correlate well with the increase in ethylene sensitivity, occurring at later stages of flower development (van Staden and Dimalla, 1980; van Staden et al., 1987). Cytokinins in petals therefore seem not to account fully for the rise in petal ethylene production. As the gynoecium has been shown to regulate senescence of carnation petals, the cytokinin levels in the gynoecium might be more important than the levels in petals. A detailed analysis of cytokinin activity in the gynoecium has apparently not been published. It is concluded that it is not yet known what signal starts the rise in ethylene production in unpollinated petals with ethylene-sensitive senescence. The following mechanism is here tentatively suggested: according to a default pathway the ovary will undergo senescence, after a species-specific period of time, if no pollen becomes deposited on the stigma. In the absence of pollination, ovary senescence induces petal senescence. If pollination occurs, ovary senescence is prevented. The pollination signal nonetheless induces a rapid rise in ovary ethylene levels, which induces petal senescence. The question as to what regulates the initial rise in ethylene production, in flowers that are not pollinated or treated with exogenous ethylene, is thereby not answered, but only shifts from the petals to the ovary. Carnation cultivars with a long vase life: relationship to ethylene Some carnation cultivars have an unusually long vase life. The long life in some of these cultivars was related to ethylene insensitivity. In other cultivars it was due to absence of ethylene production, combined with normal sensitivity to ethylene. Examples of the first type are a group of related cultivars (Chinera, Epomeo, Ginevra) derived from a long-lived non-commercial breeding line. The vase life of these flowers was highly correlated with their degree of ethylene insensitivity (van Doorn et al., 1993). During the vase life of Chinera flowers, the increase in the activities of ACC synthase and ACC oxidase, and the increase in the rate of ethylene production, occurred later than in normal, relatively short-lived, cultivars. Ethylene treatment stimulated the activity of the ethylene biosynthetic pathway but stimulation was less than in a relatively short-lived cultivar (Wu et al., 1991a, b; Woltering et al., 1993). The low ethylene sensitivity in these cultivars may relate to their abnormal ovary. In Chinera flowers, the ovary contained style-like and leaf-like structures rather than the normal reproductive organs. This may explain why the ovary does not produce the ethylene required to induce petal senescence (EJW, unpublished results). Examples of the second group are the cultivars Sandra (Wu et al., 1991a, b), Sandrosa (which is very similar to Sandra; Mayak and Tirosh, 1993), and White Candle (Nukui et al., 2004). Compared with cultivars with a shorter vase life, White Candle flowers showed lower expression of ACC synthase genes in the gynoecium, and repressed gynoecium ethylene production. The low ethylene production in the gynoecium was the apparent cause of the lack of transcription of ACC synthase genes in the petals and thus of low ethylene synthesis in the petals (Nukui et al., 2004). The data in this section support the idea that petal senescence in carnation requires a normal ovary, which determines both ethylene sensitivity of petal senescence and ethylene production in the petals. Hormones other than ethylene, in petals with ethylene-sensitive senescence As described above, cytokinins have been implicated in the onset of the rise in ethylene production, in unpollinated, ethylene-sensitive petal senescence. The importance of cytokinin levels also follows from other observations. Treatment with cytokinins delayed carnation petal senescence. It delayed the rise in ethylene production, possibly through a decrease in ethylene sensitivity (Eisinger, 1977; van Staden, 1995). Cytokinin levels can be regulated by synthesis, inactivation, and degradation. During Petunia petal senescence, cytokinins were inactivated by O-glucosylation. Cytokinin levels were also decreased by degradation, mainly through the activity of cytokinin oxidase/dehydrogenase. Feeding carnation flowers with 6-methylpurine, an inhibitor of cytokinin oxidase/dehydrogenase, resulted in a much longer petal life span (Taverner et al., 2000). The increase in mRNA abundance of two genes encoding cytokinin oxidase/dehydrogenase, during carnation petal senescence (Hoeberichts et al., 2007), might also indicate accelerated cytokinin breakdown. The role of ABA during ethylene-regulated petal senescence is still unclear. In cut carnations, a large increase in ABA levels was observed in the gynoecium, and a small increase in petals, prior to or concomitant with the ethylene upsurge (Nowak and Veen, 1982; Eze et al., 1986; Hanley and Bramlage, 1989; Onoue et al., 2000). The increase in ABA levels was completely prevented in carnations pre-treated with silver thiosulphate (Nowak and 462 van Doorn and Woltering Veen, 1982), showing that ethylene perception was required. Application of ABA to carnation flowers advanced the large rise in ethylene production and hastened petal wilting (Mayak and Dilley, 1976). If the gynoecia were removed, exogenous ABA no longer induced ethylene production and early petal wilting. It was therefore concluded that exogenous ABA acts through induction of ethylene synthesis in the gynoecium (Shibuya et al., 2000; Nukui et al., 2004). Auxins can reduce ethylene sensitivity in some tissues such as abscission zones, but in other tissues they can induce ethylene production. Application of IAA hastened the rise in ethylene production and petal wilting in cut carnation flowers (van Staden, 1995). The IAA effect was absent if the gynoecium was removed (Shibuya et al., 2000). A synthetic auxin (2,4-dichlorophenoxyacetic acid; 2,4-D) induced the expression of ACC synthase genes in the styles, ovaries, and petals (Jones and Woodson, 1999). In carnation petals, a transient increase was observed in the mRNA abundance of an Aux/IAA gene (Hoeberichts et al., 2007). The central role of the encoded Aux/IAA proteins in auxin responses has been described above. Application of GA3 to cut carnation flowers delayed the rise in ethylene production and postponed petal wilting (Saks and van Staden, 1993). Data on GA levels in carnation flowers have thus far apparently not been reported. Treatment with methyl-jasmonate hastened senescence in Petunia hybrida, Dendrobium, and Phalaenopsis (Porat et al., 1993, 1995). However, Xu et al. (2006) found only an earlier colour change and no promotion of petal wilting after treatment of Petunia inflata with methyl-jasmonate. It is concluded that there is ample evidence in ethylenesensitive petal senescence for a role for cytokinin in modulating responsiveness to ethylene. ABA and auxin also affect ethylene production, but their role during senescence is still unclear. Hormonal control of petal senescence that is insensitive to ethylene Remarkably little is known about the early regulation of ethylene-insensitive petal senescence. The required signal might be endogenous, from the petal cells themselves, and may not require hormones as an intermediate signal. Treatment with benzyladenine, a cytokinin, delayed petal senescence by a few days in Iris flowers (Wang and Baker, 1979; van der Kop et al., 2003), but data on cytokinin levels during senescence are lacking. In contrast to the effect of cytokinins, ABA treatment of Hemerocallis (Panavas et al., 1998b) or isolated Iris petals (Pak and van Doorn, unpublished results) hastened the time to visible senescence. ABA levels in petals of cut daylily flowers increased before flower opening and continued to increase during vase life (Panavas et al., 1998b). During petal senescence of non-cut and cut daffodil flowers, the petal ABA content also increased. However, the rise in ABA level occurred after an increase in the mRNA abundance of senescence-associated genes, indicating that ABA was not an early regulator. Attempts to extend floral longevity by using putative inhibitors of ABA biosynthesis were unsuccessful (Hunter et al., 2004b). The petal life span of cut Iris flowers was increased by about half a day (controls had a life of 4 d) after a pulse treatment with GA3 (Harkema and van Doorn, unpublished results). The life of cut daffodil flowers was also extended when they were held in vase solutions containing GA3 (Hunter et al., 2004b). Taken together, there are as yet no convincing data showing that any hormone controls the onset of ethyleneinsensitive petal senescence. Treatment with cytokinins caused a considerably delay of senescence, but its functional role has not been established. Regulation of senescence by sugars Effects of exogenous sugars, relationship with internal carbohydrate levels Sugars in the vase solution extend the life of many cut flowers. Sugars might serve to improve water relations (as they increase the levels of osmotic solutes), and provide the energy for maintaining cellular homeostasis, but other effects, such as on ethylene sensitivity (further discussed below), may be even more important. An apparent paradox in carbohydrate relations of cut flowers has still not been solved. In some species, such as carnation, the petal sugar levels remain rather high, and thus do not seem to become limiting. In apparent contradiction, sugar feeding extended the life of these petals. There may be several reasons for this paradox. One is the inaccurate separation of tissues and cells at various stages of senescence, which may result in overlooking a temporary decrease in sugar levels as a trigger of senescence. Often whole petals or petal parts are sampled, but the mesophyll cells often show earlier senescence than the epidermis cells, and the phloem cells, where sugars accumulate during senescence, show still later senescence. Another reason for the paradox may be inadequate separation of organelles in one type of cell. The onset of senescence may relate to low carbohydrate levels in one cellular compartment, such as the cytosol, while there is no such low level in another compartment, for example the vacuole. Inclusion of other sugars such as trehalose, mannitol, and inositol in the vase water also delayed senescence in some species. Examples are tulips (Iwaya-Inoue and Nonami, 2003) and Gladiolus (Otsubo and Iwaya-Inoue, 2000; Yamada et al., 2003; Arora and Singh, 2006). These sugars may delay senescence in a way similar to sucrose or glucose. Petal senescence 463 Competition for carbohydrates In cut lily inflorescences, the life span of individual flowers was shorter if they were left on the inflorescence rather than severed shortly after opening. The presence of flower buds and opening flowers apparently induced earlier senescence in the older flowers. A sharp decrease was observed in total carbohydrate content (starch, glucose, fructose, sucrose, and glycerol glucosides) in the petals of flowers that remained attached to the inflorescence. This was in contrast to a much smaller decrease in these compounds in the petals of detached individual flowers (van der Meulen-Muisers et al., 1995; van der Meulen-Muisers, 2000). Using three lily cultivars with a different petal life span, the time to visible withering of non-cut flowers coincided with the time until carbohydrate content had been reduced to about half. This indicates a correlation between the onset of senescence and lack of carbohydrates, in lily petals. The senescence of these petals is ethylene insensitive (van der Meulen-Muisers, 2000; van der Meulen-Muisers et al., 2001). These data are in contrast to the situation in carnation, where sugar levels did not decrease prior to visible senescence (van Doorn, 2004). Effect of sugars on ethylene sensitivity Exogenous sugars usually considerably delay visible senescence in flowers with ethylene-sensitive petal senescence, whereas sugars have only a small effect in flowers with ethylene-insensitive petal senescence (van Doorn and Stead, 1994; van Doorn, 2004). Flowers in which sugars have a large effect also generally show much earlier visible senescence when detached from the plant and placed in water, compared with flowers where sugars have only a small effect. This suggests that sugars overcome some senescence-enhancing effect of detachment, in ethylene-sensitive senescence. In carnation, feeding with exogenous sugars decreased the effect of exogenous ethylene (Mayak and Dilley, 1976). Part of such an effect might be due to the suppression of ethylene synthesis by the petals, following sugar treatment. In carnation flowers, sucrose (and an adequate antimicrobial compound) delayed the increase in mRNA abundance of almost all senescence-associated genes, very similarly to the effect of STS. At least part of the sugar effect could be explained by the abundance of EIL3 mRNA (Hoeberichts et al., 2007) and by lower levels of EIL3 protein. The presence of high sugar levels was observed to promote proteasomal degradation of EIN3 (Yanagisawa et al., 2003). As described above, EIL3 is a key transcription factor that translates the ethylene signal into the expression of numerous ethylene-responsive genes. How the sugar signal is translated into an antiethylene signal is not known. Possible sites of interaction are EIN2 and other regulators of proteasomal degradation. Sugars can also delay senescence in ethylene-insensitive petals. In Sandersonia, the effect of exogenous sugars on senescence was accompanied by a delay in the expression of genes involved in fatty acid and protein remobilization (Eason et al., 2002). Down-regulation of the tricarboxylic acid cycle? In senescing Alstroemeria petals, Breeze et al. (2004) found a steady, 7-fold decrease of the transcript abundance of 2-oxoglutarate dehydrogenase (i.e. aketoglutarate dehydrogenase), an enzyme involved in the production of succinyl-CoA in the tricarboxylic acid (TCA) cycle (Brugière et al., 2004). This decrease in transcript abundance might be associated with a decrease in enzyme activity, and thus with down-regulation of the TCA cycle. If true, this would be indicative of a major shift from anabolism to catabolism. Nonetheless, specific compounds such as sucrose and some amino acids (see below) are synthesized until a late stage of senescence. Remobilization, transport of mineral ions In cut flowers of which the petals show wilting or withering as a final symptom of senescence, macromolecules become degraded to mobile compounds. These are transported, through the phloem, out of the petal. For example, Hew et al. (1989) noted movement of phosphorus and soluble amino nitrogen out of the petals of Arachnis orchids. Macromolecules that are degraded include starch, lipids, proteins, nucleic acids, and, often, the molecules that form the cell wall. Carbohydrates are mainly transported in the phloem as sucrose, whilst hexoses are also found. Nitrogen is often mainly transported as glutamine and asparagine, and phosphorus is transferred as the inorganic anion (Eason et al., 2000). The phloem also carries cations such as potassium, calcium, and magnesium. These compounds can become loaded to the phloem during the remobilization phase of the living cells, but also after vacuolar collapse. By the time of visible senescence, the content of soluble phosphates in Ipomoea petals was reduced by 75% (Matile and Winkenbach, 1971). Inorganic phosphate seems to be mainly derived from degradation of proteins, nucleic acids, and phospholipids. In the vacuole, phosphates are remobilized by enzymes such as acid phosphatase and phosphodiesterase (Matile, 1997). How phosphates are remobilized from nucleic acids in the nucleus is unknown, but acid phosphatases are also reportedly present in nuclei. Inorganic cations also are removed from petals during senescence. About a quarter of the K+ content, ;12% of the Mg2+ content, but not more than 6% of the Ca2+ 464 van Doorn and Woltering content was withdrawn from cotton (Gossypium) petals during senescence (Phillis and Mason, 1936). In Ipomoea petals, ;60% of the K+ content, about half of the Mg2+ content, and about a quarter of the Ca2+ content was transferred to other plant parts (Winkenbach, 1970a). In Petunia, a large decrease was found in the content of potassium, but no significant changes were detected in the contents of calcium, magnesium, and micronutrients (Verlinden 2003). These data indicate that the pattern of mineral ion salvage depends on the species. Many enzymes contain metals, such as copper and zinc, which are probably released during protein degradation. Metallothioneins are known to bind copper and zinc ions (Hall, 2002). The increase in the expression of genes encoding metallothioneins during petal senescence (Hunter et al., 2002; Breeze et al., 2004) might therefore relate to metal transport, but a number of alternative roles for metallothioneins, for example the detoxification of ROS (Kiningham and Kasarskis, 1998), cannot be excluded. The two functions might also be connected. Intracellular free copper, for example, has to be strictly limited due to its toxic effects, in particular the generation of ROS. The expression of genes encoding metallothioneins and other proteins involved in heavy metal homeostasis is tightly regulated. It prevents toxic levels of ionic copper (Balamurugan and Schaffner, 2006). Several genes encoding proteins that transport inorganic cations or anions through membranes become differentially expressed during petal senescence (Table 4). Syntaxins are required for the fusion of transport vesicles to target membranes (Sanderfoot et al., 2001). ABC proteins are found in the tonoplast. The proteins carry Table 4. Differential expression, during petal senescence, of genes putatively involved in membrane transport Species Alstroemeria pelegrina Dianthus caryophyllus Iris3hollandica Narcissus pseudonarcissus a Gene Reference Potassium transporter ATP-binding cassette (ABC) transporter Sugar transporter Syntaxina Breeze et al. (2004) Amino acid permease Copper transporter Golgi nucleotide sugar transporter MATE efflux transporter Metal nicotianamide transporter Monosaccharide transporter Na+/myo-inositol transporter Oligopeptide transporter Cyclic nucleotide-gated ion channel Hoeberichts et al. (2007) Nitrate transporter Hunter et al. (2002) Data from Supplement 3. van Doorn et al. (2003) glutathione-S-conjugated molecules as well as other compounds (Lu et al., 1998). Conjugation with glutathione, most probably carried out by a glutathione-Stransferase (GST), is important for movement of various compounds over membranes. The transcript abundance of GST increased during carnation petal senescence (Itzhaki et al., 1994; Hoeberichts et al., 2007). However, GST can also be involved in detoxification, for example of peroxidized fatty acids (see below). The syntaxin that was up-regulated in Alstroemeria shows high homology with Arabidopsis SYP81, which is used as a marker of the ER (Uemura et al., 2004). CNGCs conduct K+ and Ca2+ ions (Köhler et al., 2001). Sugar synthesis and transport During petal senescence in Alstroemeria (Supplement 3 of Breeze et al., 2004) and Iris (van Doorn et al., 2003), the transcript abundance of a gene encoding a triose phosphate isomerase and that of genes encoding sucrose synthase increased. In Alstroemeria, the transcripts of a gene encoding a cell wall invertase also became more abundant. These data indicate an increase in sucrose synthesis. Nonetheless, in Iris, some sucrose synthase genes became down-regulated, which may indicate a shift of sucrose synthesis to other parts of the cells or tissues. Petal senescence is accompanied by a high sugar concentration in the phloem. In Ipomoea and Hemerocallis petals, the concentration of sugars in the phloem exudate increased by a factor of 15 and 14, respectively, during the period between the closed floral bud and the beginning of petal wilting (Wiemken et al., 1976; Bieleski and Reid, 1992). Sucrose is the main sugar in the phloem of senescent Ipomoea (Wiemken et al., 1976) and Hemerocallis (Bieleski, 1995) petals. In order to reach the phloem, the sugars must be transferred, at some point, through a membrane. Several genes encoding sugar transporters were up-regulated during Alstroemeria and carnation petal senescence (Table 4). Sugar transport from Ipomoea petal cells to the phloem was correlated with inactivation of an invertase, probably localized to the cell walls. It was suggested that this decrease was instrumental in preventing the cleavage of apoplastic sucrose, released from the cells during remobilization (Winkenbach, 1970a). Later work confirmed the increase of an invertase inhibitor protein in several flowers undergoing senescence, such as carnation, Alstroemeria, Dahlia, Gladiolus, and Petunia (Halaba and Rudnicki, 1989). As there are also invertases, as well as invertase inhibitors, in the vacuole (Greiner et al., 2000), it is not yet clear which invertases decrease in activity during senescence and in which compartments the invertase inhibitors become active. Additionally, invertase activity can be reduced because of invertase protein degradation. During leaf senescence, a vacuolar invertase was degraded Petal senescence 465 by proteases, which required the presence of a VPEc (Rojo et al., 2003). The data in this section show that sugar metabolism is active in senescent cells as many carbon skeletons that are remobilized from macromolecules are transported out of the petal, mainly as sucrose. Nucleic acid synthesis and breakdown Numerous mRNAs are produced during senescence. Such production requires intact mechanisms of both translation and post-translational RNA modification. It also requires an intact nucleus. RNA is often grouped into five classes: mRNA, tRNA, rRNA, siRNA, and miRNAs. Whether DNA and the RNAs other than mRNA are (still) produced during senescence is apparently not known. RNA and DNA are both degraded during senescence. In petals, DNA and RNA are present in the nucleus, mitochondria, chloroplasts (if present), and plastids. Degradation of nucleic acids in these organelles can take place before vacuolar collapse, by local enzymes. Breakdown of nucleic acids can also take place after vacuolar collapse, as vacuoles of cells undergoing senescence contain several DNases and RNases (Matile, 1997). No data have as yet been reported to indicate that nucleic acids are transported to the vacuole and are degraded there. RNA degradation RNA can be degraded by RNases and by some DNases. Winkenbach (1970a) found that RNA levels had already decreased by the time of onset of Ipomoea flower opening. RNA levels then further decreased as the flower opened, accompanied by a dramatic upsurge of RNase activity (Winkenbach, 1970b). RNase activity in Hemerocallis petals also increased during senescence (Panavas et al., 2000). In Petunia inflata, the initial decrease in mRNA levels was due to the increased activity of constitutive RNases. Novel RNase activities appeared rather late (Xu and Hanson, 2000). Overexpression of a specific RNase inhibitor prevented senescence in an animal system, indicating that RNA degradation is an important regulator (Castelli et al., 1998). Whether this also applies to plants remains to be evaluated. DNA degradation Degradation of DNA can be assessed in situ using the TUNEL (TdT-mediated dUTP-biotin nick end labelling) method, which labels DNA breaks in individual nuclei. The TUNEL method preferentially labels double-stranded breaks (Gravrieli et al., 1992). TUNEL-positive nuclei were found in petals of Pisum (Orzaez and Granell, 1997), Petunia (Langston et al., 2005), and Gypsophila. In Gypsophila, TUNEL-positive nuclei appeared in all cells, well before visible senescence symptoms and well before the increase in ethylene production (Hoeberichts et al., 2004). Nuclear DNA is often degraded in two phases. During the first stage, breakdown occurs into large fragments (50 000–200 000 bp). The second stage consists of breakdown into multiples of ;180 bp. When the DNA at this second stage is placed on a gel, the 180 bp multiple fragments often show a pattern reminiscent of a ladder; hence the name DNA laddering (Nagata et al., 2003). Accumulation of DNA fragments with multiples of 180 bp was found during the petal senescence of some species, but it was not observed in several other species (Table 5), indicating that it is not a good indicator of senescence/PCD in petals. In Ipomoea, the onset of the decrease in DNA content was accompanied by a large increase in DNase activity (Winkenbach, 1970a, b). In petals of Hemerocallis (Panavas et al., 1999), narcissus (Hunter et al., 2004a), and Petunia (Langston et al., 2005), the mRNA abundance of DNase (endonuclease) genes increased during senescence. In conclusion, RNA and DNA breakdown is often related to an increase in the expression of some genes encoding DNases, and is in some species accompanied by a dramatic increase in RNase and DNase activities. In some species, DNA degradation resulted in accumulation of oligosomal 180 bp fragments, but in many other species such accumulation was not observed. Amino acid synthesis, protein synthesis, and protein degradation Amino acid synthesis During senescence, amino acids are transferred to the phloem. In Ipomoea petals, Wiemken et al. (1976) noted a 4-fold increase of amino acids in the phloem exudate, during the period from flower opening to petal wilting. The Ipomoea exudate mainly contained glutamine, asparagine, and lysine, but also, at low concentrations, c-aminobutyric acid and serine, and a range of other amino acids. In Hemerocallis, glutamine and hydroxyproline Table 5. Observation, on agarose gels, of clear DNA laddering or no clear laddering, during petal senescence Species DNA laddering Reference Gladiolus hybrid Pisum sativum Alstroemeria pelegrina Antirrhinum majus Argyranthemum frutescens Ipomoea nil Iris3hollandica Petunia hybrida + + – – – – – – Yamada et al. (2003) Orzaez and Granell (1997) Wagstaff et al. (2003) Yamada et al. (2006b) Yamada et al. (2006b) Yamada et al. (2006a) van der Kop et al. (2003) Yamada et al. (2006b) 466 van Doorn and Woltering were the main transport amino acids (Bieleski, 1995), and in Sandersonia petals asparagine and glutamine were most prominent (Eason et al., 2002). The mRNAs of several genes encoding amino acid synthetases were differentially expressed (Table 6). The data support the idea of increased synthesis of at least glutamine, asparagine, and cysteine. Table 6 also shows increased transcript abundance of several other genes involved in amino acid metabolism. The isovaleryl-CoA oxidase, for example, is involved in valine and leucine catabolism (Däschner et al., 2001), and serine hydroxymethyltransferase converts two glycine molecules to one serine (Oliver, 1994). In conclusion, the data in this section show considerable synthesis of specific amino acids in cells undergoing senescence, and accumulation of these amino acids in the phloem. Protein synthesis Treatment with compounds that inhibit translation or transcription delayed the visible symptoms of petal senescence, showing that protein synthesis is required (Xu et al., 2007). Down-regulation of eIF-5A delayed carnation senescence (Hopkins et al., 2007), which leads to the same conclusion. In petals of Iris (van Doorn et al., 2003) and carnation (Hoeberichts et al., 2007), the mRNA abundance of a gene encoding a plant homologue of the animal translationally controlled tumour protein (TCTP) increased. In animal cells, the encoded protein has at least two functions. It blocks protein synthesis by interacting with a eukaryotic elongation factor (eEF). It also binds to pro-apoptotic members of the Bcl-2 family, and thus induces death (Rinnerthaler et al., 2006). It is not clear what the function of the protein is in plant cells. It is discussed here under protein synthesis, as the hypothesis may be put forward that in plant cells it also interacts with an eEF. Several genes related to protein synthesis are differentially expressed during petal senescence (Table 6). Both in daffodil (Hunter et al., 2002) and in Alstroemeria (Supplement 3 of Breeze et al., 2004), a DEAD/DEAH domain helicase was up-regulated. These genes have high homology to eIF-4A. Protein degradation Due to various rates of degradation, proteins have a specific half-life. Protein breakdown occurs in the proteasomes, vacuoles, mitochondria, the nucleus, and plastids. Bulk degradation mainly occurs in vacuoles. Prior to visible senescence symptoms, protein levels in petals decrease, sometimes drastically as in Ipomoea (Winkenbach, 1970a, b) and Petunia (Jones et al., 2005). A decrease in overall protein levels can be due to a decrease in synthesis as well as an increase in degradation. In Ipomoea, no upsurge of in vitro protease activity was detected during petal senescence (Winkenbach, 1970a, b), but in other species, such as Hemerocallis (Stephenson and Rubinstein, 1998), Iris (Pak and van Doorn, 2005), and Petunia (Jones et al., 2005), a massive increase in protease activity was found. Table 6. Differential expression, during petal senescence, of genes putatively involved in amino acid metabolism and protein synthesis Gene Amino acid metabolism Asparagine synthase Cysteine synthase Glutamine synthase Isovaleryl-CoA dehydrogenase Serine hydroxymethyl transferaseb Tryptophan synthase Tyrosine aminotransferase Protein synthesis Ribosomal protein L2 Ribosomal protein L23 Ribosomal protein L32 Ribosomal protein S7 Ribosomal protein S12 Ribosomal structural RNAc Translation initiation factor 4Ad Translation initiation factor 5A a b c d Species Comment Reference Sandersonia aurantiaca Alstroemeria pelegrina Alstroemeria pelegrina Sandersonia aurantiaca Dianthus caryophyllus Iris3hollandica Alstroemeria pelegrina Dianthus caryophyllus Up-regulated Up-regulated Up-regulated Highly expressed Up-regulated Down-regulated Down-regulated Up-regulated Eason et al. (2000) Breeze et al. (2004)a Breeze et al. (2004) Eason et al. (2000) Hoeberichts et al. (2007) van Doorn et al. (2003) Breeze et al. (2004) Hoeberichts et al. (2007) Alstroemeria pelegrina Alstroemeria pelegrina Dianthus caryophyllus Alstroemeria pelegrina Alstroemeria pelegrina Iris3hollandica Alstroemeria pelegrina Narcissus pseudonarcissus Narcissus pseudonarcissus Up-regulated Up-regulated Up-regulated Up-regulated Up-regulated Down-regulated Up-regulated Up-regulated Downregulated Breeze et al. (2004)a Breeze et al. (2004) Hoeberichts et al. (2007) Breeze et al. (2004) Breeze et al. (2004) van Doorn et al. (2003) Breeze et al. (2004)a Hunter et al. (2002) Hunter et al. (2002) Data from Supplement 3. Accession CF227016, previously described to be homologous to a glycine hydroxymethyltransferase. Accession CF226940, previous described as senescence-associated protein. These proteins were described as DEAD/DEAH box helicases. Petal senescence 467 Protein degradation in proteasomes Proteasomes are involved in degradation of specific proteins. They degrade misfolded proteins and can determine the level of normally folded proteins. As discussed above, proteasomes are important regulators of ethylene action, as they regulate the level of EIN3/EIL transcription factors. The level of other important transcriptional regulators, such as those involved in auxin and gibberellin responses, are also determined by proteasomal action (Bishopp et al., 2006). The proteasome system was apparently up-regulated during petal senescence in daylily (Courtney et al., 1994; Müller et al., 2004), daffodil (Hunter et al., 2002), and carnation. In carnation, the mRNA abundance increased of three genes encoding subunits of the 19S regulatory particle, one of the two large complexes of the 26S proteasome (Hoeberichts et al., 2007). Feeding isolated Iris petals with Z-Leu-LeuNva-H, an inhibitor of proteasome activity, produced a small but significant delay in the time to visible senescence (Pak and van Doorn, 2005), indicating that proteasome action is limiting senescence. Several monomers of ubiquitin, a 76 amino acid polypeptide, become attached to proteins targeted for degradation in the proteasome. Ubiquitin attachment involves three enzymes referred to as E1, E2, and E3. The cascade starts with E1 (ubiquitin-activating enzyme) which binds to ubiquitin and transfers it to E2 (ubiquitin-conjugating enzyme). Ubiquitin ligase (E3) binds to E2 and to the target protein, and transfers the ubiquitin moiety from E2 to the target protein. E3 can be monomeric or form a protein complex. Some types of multimeric E3 complexes contain an F-protein, which determines which specific protein will be ubiquitinated. During senescence, the mRNA abundance of several genes related to ubiquitination increased (Table 7). These results strongly indicate that petal senescence is associated with enhanced activity of the ubiquitin– proteasome route of protein degradation. The delay of visible senescence symptoms after chemical inhibition of proteasomes in Iris and after silencing a RING domain E3 protein in Petunia (Xu et al., 2008) even indicate that proteasome action is required for senescence. Non-proteasomal protein breakdown Proteasome-independent protease activity markedly increases prior to visible senescence (Stephenson and Rubinstein, 1998; Pak and van Doorn, 2005). Nonproteasome proteases are often divided into exo- and endoproteases, referring to the position in the target protein where the cleavage takes place. Endoproteases include cysteine-, serine-, aspartic-, and metalloproteases, named after the amino acid residues or the metals that are required for the cleavage reaction. The activities of these protease classes have been assessed by using inhibitors and in vitro measurements whereby a common protein is Table 7. Increase in the expression of genes putatively involved in protein ubiquitination (leading to proteasomal protein degradation), and genes encoding proteins in the 19S subunit of the proteasome, during petal senescence Gene group Species Reference Breeze et al. (2004) Hunter et al. (2002) E3 HECT domain Alstroemeria pelegrina Narcissus pseudonarcissus Ipomoea nil E3 U-box (arm) Dianthus caryophyllus E3 RING domain Dianthus caryophyllus F-box Mirabilis jalapa Petunia hybrida Dianthus caryophyllus Proteasomal subunits 19S particle Dianthus caryophyllus Ubiquitination E2 Yamada et al. (2007a) Hoeberichts et al. (2007) Hoeberichts et al. (2007) Xu et al. (2008) Xu et al. (2008) Hoeberichts et al. (2007) Hoeberichts et al. (2007) used as a substrate for all proteases extracted from a tissue. However, most protease inhibitors are not specific. Only E-64—and probably antipain—are specific inhibitors, both affecting cysteine proteases. Using E-64 in vitro, total protease activity was reduced by about half in petals of Hemerocallis (Stephenson and Rubinstein, 1998) and Iris (Pak and van Doorn, 2003) and by >90% in Petunia (Jones et al., 2005). Antipain reduced protease activity of Sandersonia petals by ;30% (Eason et al., 2002). Feeding Iris petals with a membrane-permeable form of E-64 also indicated that about half of the peak protease activity was due to cysteine proteases (Pak and van Doorn, 2003). The results of the other (non-specific) inhibitors might suggest a small amount of metalloprotease activity at the peak of protease activity in Hemerocallis, Iris, and Petunia, considerable serine protease activity in Hemerocallis and Iris, but almost no activity in Petunia, and little or no activity of aspartic proteases in these species (Stephenson and Rubinstein, 1998; Pak and van Doorn, 2003; Jones et al., 2005), but these data need verification by means other than inhibitors. Treatment of Iris flowers, at the onset of flower life, with AEBSF and DFP prevented the normal visible senescence symptoms. These compounds are general serine protease inhibitors, but AEBSF is also a general NADPH-oxidase inhibitor. Other effects of both compounds cannot be excluded. AEBSF and DFP delayed the onset of bulk protease activity during senescence, thus preventeing both cysteine and non-cysteine protease activity (Pak and van Doorn, 2005). Interestingly, AEBSF has also been reported to inhibit apoptotic PCD in animals (Solovyan and Keski-Oja, 2006). AEBSF, which completely suppressed cysteine protease activity in Iris, delayed senescence, but feeding Iris petals 468 van Doorn and Woltering with the membrane-permeable form of E-64, which reduced peak protease activity to about half, had no effect on senescence (Pak and van Doorn, 2003). The cysteine protease inhibitor 2,2#-dipyridyl delayed the time to wilting in Sandersonia petals and prevented the senescenceassociated rise in endoprotease activity (Eason et al., 2002). This effect might relate to cysteine protease inhibition, but side effects of 2,2#-dipyridyl cannot be excluded. Table 8 summarizes the present data on the increase in the expression of genes encoding non-proteasomal proteases, most of which are probably vacuole localized. Genes encoding metalloproteases have not yet been reported. Some of the encoded cysteine proteases contain the C-terminal KDEL domain, which is an ER retention motif (Eason et al., 2002; Wagstaff et al., 2002). KDEL domain cysteine proteases are generally found in vesicles that bud from the ER and merge with the vacuole. One type of vesicles containing KDEL cysteine proteases have been suggested not to merge with the vacuole but remain until they are destroyed after vacuolar collapse (Greenwood et al., 2005). These vesicles, called ricinosomes, were reportedly present in petals of Hemerocallis undergoing senescence (Schmid et al., 1999). As many as nine cysteine proteases have been isolated from senescing Petunia petals. In a transgenic Petunia line that was ethylene insensitive, the time to petal wilting was delayed by ;8 d. In this line, the increase in Table 8. Increase, during petal senescence, in the expression of genes putatively involved in non-proteasomal protein degradation Gene group Species Proteases in vacuoles and vesicles Aspartic proteases Alstroemeria pelegrina Hemerocallis hybrid Cysteine proteases Alstroemeria pelegrina Dianthus caryophyllus Hemerocallis hybrid Iris3hollandica Serine proteases Proteases in mitochondria and plastids Clp protease Reference Breeze et al. (2004) Panavas et al. (1999) Wagstaff et al. (2002) Hoeberichts et al. (2007) Valpuesta et al. (1995) van Doorn et al. (2003) Hunter et al. (2002) Narcissus pseudonarcissus Petunia hybrida Jones et al. (1995) Sandersonia aurantiaca Eason et al. (2002) Narcissus Hunter et al. (2002) pseudonarcissus Alstroemeria pelegrina Iris3hollandica expression of four cysteine protease genes occurred later, corresponding to the delay of petal wilting (Jones et al., 2005). In carnation petals, a gene encoding a cysteine protease inhibitor, abundant at the time of flower opening, became gradually down-regulated. Its mRNA had disappeared by the time of the increase in cysteine protease expression and petal wilting (Sugawara et al., 2002). This cysteine protease inhibitor showed high homology with cystatin. Hoeberichts et al. (2007) found a transient increase in the expression of another cystatin gene in carnation. Since in other systems cystatins suppressed PCD (Rojo et al., 2003, 2004), these vacuolar proteins might be regulators of petal senescence. Chloroplasts, non-chloroplast plastids, and mitochondria usually contain several families of ATP-dependent proteases (called Clp, Lon, and FtsH), and one family of ATP-independent proteases (called DegP). The Clp, Lon, and DegP families are serine proteases, and the FtsH family are metalloproteases (Sinvany-Villalobo et al., 2004). The transcript abundance of Clp proteases increased prior to visible senescence (Table 8). Plant nuclei also contain non-proteasomal proteases, for example a serine protease (Domı́nguez and Cejudo, 2006), which might be involved in degradation of nuclear proteins during senescence. In conclusion, petal senescence is accompanied by bulk non-proteasomal protein breakdown, mainly in vacuoles but possibly also in mitochondria, plastids, and nuclei. Many genes encoding proteases have been identified. Protease activity in the vacuole seems to be regulated by enzyme activation and by enzyme inhibition, in which VPEc seems to be involved. The level of cystatins, cysteine protease inhibitors, in vacuoles might be a regulator of petal senescence. Breeze et al. (2004) van Doorn et al. (2003) Lipid synthesis and breakdown Lipid synthesis Bieleski and Reid (1992) found evidence for rapid cessation of overall phospholipid synthesis during Hemerocallis petal senescence. Nonetheless, several genes that are apparently involved in lipid synthesis become up-regulated during petal senescence (Table 9). b-Ketoacyl-CoA synthase is the first enzyme of fatty acid elongation, whilst acyl thioesterase is involved in the last step. Acyl-CoA reductase may be associated with the increase in transcript abundance of a gene encoding CER1, and be involved in steryl/wax ester synthesis in membranes (Madey et al., 2001). Fatty acid hydroxylases are involved in various pathways, including those producing isoprenoids, carotenoids, and oxylipins. The fatty acid hydrolases CYP86, CYP94, and CYP96 are classified in the CYP86 clan, involved in fatty acid metabolism (Nelson et al., 2004). Petal senescence 469 Table 9. Differential expression, during petal senescence, of genes putatively involved in fatty acid synthesis Gene Species Comment Reference Acyl carrier protein (ACP) thioesterase ACP b-ketoacyl-CoA synthase Dianthus caryophyllus Hemerocallis hybrid Iris3hollandica Alstroemeria pelegrina Ipomoea nil Alstroemeria pelegrina Up-regulated Up-regulated Down-regulated Down-regulated Up-regulated –b Hoeberichts et al. (2007) Pananas et al. (1999) van Doorn et al. (2003) Breeze et al. (2004) Yamada et al. (2007a) Breeze et al. (2004) Acyl-CoA reductase Fatty acid hydroxylase (CYP96/CYP86)a Fatty acid hydroxylase (CYP94/CYP96) a b Accession AB267814. Previously described (more generally) as a cytochrome P450 protein. Two genes were up-regulated, whilst one was down-regulated. The encoded proteins were previously described as cytochrome P450 containing. The data indicate that mass lipid degradation during senescence is accompanied by the synthesis of some fatty acids and by some compounds, such as wax esters, derived from fatty acids. Table 10. Up-regulation, during petal senescence, of genes putatively involved in degradation of phospholipids and fatty acids Fatty acid degradation Phospholipid degradation Allene oxide synthase During senescence, degradation of free fatty acids mainly occurs by b-oxidation (Pistelli et al., 1992). This pathway, which repeatedly removes C2 units from the fatty acid chain, is present in peroxisomes. Several genes encoding enzymes that degrade fatty acids were reportedly upregulated during petal senescence (Table 10). Acyl-CoA oxidase isozymes catalyse the first step in peroxisomal fatty acid b-oxidation. The transcript abundance of a gene encoding an acyl-CoA synthase increased during senescence. The mRNA abundance of an acyl-CoA dehydrogenase, also part of the b-oxidation chain, also increased. Thiolases are often localized to peroxisomes, and involved in b-oxidation, but some isozymes are localized to mitochondria where they are also involved in fatty acid degradation. The end-product of b-oxidation (acetyl-CoA) is fed into the glyoxylate cycle, also localized to the peroxisomes. Isocitrate lyase and malate synthase are key enzymes in this cycle. Both genes were up-regulated during senescence in cucumber petals (Graham et al., 1994; McLaughlin and Smith, 1994). The glyoxylate pathway produces organic acids which can be converted to sugars (gluconeogenesis) in the cytosol (Chen et al., 2000; Cornah et al., 2004) or be used for the synthesis of amino acids. Phospholipid degradation and membrane leakiness All enzymes required for phospholipid degradation are apparently always present in the membranes, even in those of young cells, and are involved in continuous phospholipid turnover (Fobel et al., 1987; Brown et al., 1990). During senescence, a massive loss of phospholipids is often found, concomitant with an increase in phospholipid degradation products (Suttle and Kende, 1980; Thompson et al., 1998; Hopkins et al., 2007). Isolated plasma membranes from Petunia petals showed an increase in the levels of free polar lipid head groups, phosphatidic acid, and phosphatidylinositol monophos- Gene GDSL lipase/ hydroxylase Lipolytic acyl hydrolase Unspecific lipases Fatty acid degradation Acyl-CoA dehydrogenase Acyl-CoA oxidase Acyl-CoA synthasea Isocitrate lyase 3-Ketoacyl-CoA thiolase Malate synthase Species Reference Hemerocallis hybrid Panavas et al. (1999) Petunia hybrida Iris3hollandica Xu et al. (2006) van Doorn et al. (2003) Dianthus caryophyllus Thompson et al. (2000) Iris3hollandica Dianthus caryophyllus van Doorn et al. (2003) Hoeberichts et al. (2007) Iris3hollandica van Doorn et al. (2003) Dianthus caryophyllus Phalaenopsis orchids Iris3hollandica Hoeberichts et al. (2007) Do and Huang (1997) van Doorn et al. (2003) Cucumis sativa Dianthus caryophyllus Graham et al. (1994) Hoeberichts et al. (2007) Cucumis sativa Graham et al. (1994) a Some other contigs with high homology to acyl-CoA synthase were down-regulated during Iris petal senescence. phate (Borochov et al., 1994). This is indicative of phospholipid breakdown. There is considerable debate about the main sites of phospholipid degradation. The issue is to what degree the increase in phospholipid degradation in membranes during senescence occurs inside the autophagosomes and vacuoles and to what degree in the membrane before being sequestered in these organelles. All sides of the debate accept that some membrane degradation occurs inside chloroplasts and plastids (Hopkins et al., 2007). Many apparently intact membranes are observed inside vacuoles, suggesting that at least part of phospholipid degradation occurs there. In yeast and animal cells, there is compelling evidence for vacuolar/lysosomal degradation of phospholipids, but such firm evidence is still lacking in the case of plants (Inoue and Moriyasu, 2006). 470 van Doorn and Woltering At least one of the enzymes involved in mass phospholipid degradation, lipolytic acyl hydrolase, seemed to be mainly present in the vacuole and, under experimental conditions, able to degrade all phospholipids of the cell in a very short period. This enzyme might be involved in vacuolar degradation of phospholipids during senescence, and will probably rapidly degrade any membranes left by the time of vacuolar collapse (Travnicek et al., 1999). An argument against the role of autophagy in phospholipid degradation might be found in the finding that treatment with 3-methyladenine (3-MA), an inhibitor of autophagy, did not affect the degradation (Inoue and Moryasu, 2006). However, this inference might be false since 3-MA not only inhibits autophagy, but also has several side effects (Xue et al., 1999). A second issue of contention is the idea that the plasma membrane becomes permeable prior to vacuolar collapse, due to phospholipid degradation inside the membrane (Hopkins et al., 2007). Ions—and in some coloured petals also anthocyanins—are increasingly released by petals, during senescence (Suttle and Kende, 1980; Celikel and van Doorn, 1995). The release of ions into the apoplast might, theoretically, be passive (as a result of tonoplast rupture or as a result of increased plasma membrane leakiness), or active, regulated as part of remobilization and export through the phloem. These possibilities have not been experimentally disentangled, thus far. In contrast to ions, the release of (vacuole-located) anthocyanins from epidermis cells is most probably not active, and thus must be a result of loss of semipermeability in both the tonoplast and the plasma membrane. The amount of anthocyanin leakage from petals seems to be a reflection of the number of dead cells. Hopkins et al. (2007) defended the idea that the plasma membrane loses its semi-permeability before the cell dies. They referred to the work of Eze et al. (1986), who noted increased release of ions from carnation flowers, cut at the base of all petals and placed in water. In these experiments, however, the ion release was not shown to be due to membrane leakage. It might have been due to active efflux of ions into the apoplast. These tests also involved flowers under severe water stress, indicated by the considerable loss of fresh weight of the flower stems. The experiment needs to be repeated with attached flowers, and has to separate active and passive transport of ions into the apoplast. The data presently available, therefore, provide no decisive argument in favour of the hypothesis that the plasma membrane loses its semi-permeability before the cell dies. A change from the fluid to gel phase in membranes, and an increase in membrane fluidity (microviscosity) have been suggested to be due to enzyme-induced degradation processes inside the membranes. These changes were hypothesized to be the cause of a gradual increase in membrane leakage in senescent petals (Borochov and Woodson, 1989). However, as noted, increased leakage of otherwise intact membranes has not been demonstrated in such petals. Freeze-fracture EM of senescing carnation petals indicated the presence of gel phase lipid in the plasma membrane, ER, and tonoplast (Hopkins et al., 2007), but it is not clear how much this occurred above control levels and at what stage of senescence it took place (before or after vacuolar collapse). The data seem to be inadequate evidence for the idea of a gradual increase in membrane leakiness, during senescence. Biochemistry of phospholipid degradation Membranes usually consist of phospholipids, proteins, and sterols. Phospholipid breakdown during petal senescence has been studied mainly by using reconstituted membranes (microsomes) from the whole tissue. Such an approach is only a first approximation. The microsomes are a mixture of all membranes, which in one cell can be at various stages of degradation. It is also a mixture of the membranes of a large number of cells that may be at various stages of senescence. The work on microsomes, therefore, will not be considered here in detail. Only a few studies used specific membranes. The plasma membrane of Petunia petals, for example, showed a decrease in the phospholipid and protein levels during senescence (Borochov et al., 1994), but it is not clear if this was due to vacuolar collapse in some petal cells, or if it occurred prior to this collapse. The main phospholipid-degrading enzymes are (i) phospholipase D (PLD); (ii) phosphatidic acid phosphatase, also called phospholipase C (PLC); (iii) lipolytic acyl hydrolase (PLA1 and PLA2); and (iv) lipoxygenase. Phospholipid breakdown is apparently usually initiated by PLD, which removes the head groups, producing phosphatidic acid. Phosphatidic acid phosphatase then removes the phosphate moiety, and thus converts phosphatidic acid to diacylglycerol. Lipolytic acid hydrolases and other lipases remove the fatty acids from the glycerol backbone, or from phospholipids. As described, fatty acids are converted—mainly by b-oxidation—to smaller molecules. PLD activity, measured in vitro, decreased rapidly during senescence of Tradescantia petals (Suttle and Kende, 1980). In vivo membrane-bound PLD activity increased at least 2-fold early on during senescence in the tip of Iris petals (van Doorn et al., 2003). A purported inhibitor of PLD, lysophosphatidyl-ethanolamine, was found to delay the onset of visible petal senescence in Antirrhinum (Kaur and Palta, 1997). If this inhibitor is PLD specific, the result suggests that PLD activity is important in determining the time to the visible senescence symptoms. However, confirmation of this effect in other species has not been reported. In vitro lipolytic acyl hydrolase activity has thus far been shown only in Petal senescence 471 Tradescantia petals. It did not change during senescence (Suttle and Kende, 1980). Lipoxygenase oxidizes double bonds in fatty acids. Its activity in Alstroemeria petals decreased almost linearly from the time of harvest until the senescence symptoms (Leverentz et al., 2002). A small increase in lipoxygenase activity was found in Hemerocallis, but only by the time of visible senescence (Panavas and Rubinstein, 1988). No increase in lipoxygenase activity was observed during petal senescence in Phalaenopsis and Dendrobium orchids (Porat et al., 1995). Lipoxygenases in membranes produce hydroperoxides, which are the substrate of allene oxidases (Xu et al., 2006). As inhibitors of lipoxygenase had no effect on the time to visible senescence in Phalaenopsis and Dendrobium orchids (Porat et al., 1995), it is questionable whether lipoxygenases are important in determining the time to visible effects. In these experiments, it is not clear whether the in vitro measurements reflect in vivo changes. If they do reflect real changes, the results may still conceal what happens in specific cellular compartments. The present data regarding the expression of genes putatively involved in phospholipid degradation are summarized in Table 10. Genes encoding PLD or PLC have thus far not been isolated from petals undergoing senescence. Except for one gene, the functional significance of the phospholipid- and fatty acid-degrading enzymes has as yet not been reported. Decreased expression in Arabidopsis of a lipolytic acyl hydrolase found to be up-regulated by the time of visible senescence in carnation petals delayed leaf yellowing (Thompson et al., 2000). The gene might therefore also be important in the control of petal senescence. It is concluded that membrane degradation may be a central step in the programme that leads to cell death, as collapse of the tonoplast seems to execute the death sentence. A gradual loss in membrane permeability, rather than membrane collapse, has thus far not been unequivocally shown, for any membrane that remains functional until the end of senescence, such as the plasma membrane and the tonoplast. Already at a relatively early phase of senescence, membranes are degraded en masse, but it is not clear where this occurs: in the membranes in situ, prior to their uptake in autophagosomes and delivery in the vacuole, or after this uptake. Do reactive oxygen species contribute to petal cell death? ROS include both free radicals and hydrogen peroxide. They are produced during normal metabolism, in many organelles. There is good evidence that ROS contribute to the PCD that is induced by plant pathogens (Apel and Hirt, 2004; Van Breusegem and Dat, 2005). In contrast, there is as yet no evidence that an increase in ROS is part of the causal chain leading to developmental senescence, either as an early regulator of the onset of senescence or, because of the destructive nature of ROS, as a direct cause of cellular death. ROS might well be increasingly produced during senescence, but most of the ROS produced might be rendered innocuous. For example, ROS are probably formed when lipoxygenase peroxidizes double bonds in membrane C18–C22 fatty acids. The resulting fatty acid hydroperoxides are free radicals, able to react with other fatty acids in a non-enzymatic chain reaction. In Alstroemeria, fatty acid hydroperoxides accumulated in the outer whorl of petals, whereas no accumulation was found in the inner whorl. As these parts exhibited visible senescence symptoms at the same time, the fatty acid hydroperoxide levels were apparently of little importance for the production of the visible senescence symptoms (Leverentz et al., 2002). Evidence for polyunsaturated fatty acid oxidation also follows from accumulation of some of its end-products: malondialdehyde and other thiobarbituric acid (TBA)reactive compounds. In the inner Alstroemeria tepals (homologous to petals), the levels of TBA-reactive compounds remained almost constant, but in the outer whorl tepals (sepals) the levels peaked at an early stage and then declined throughout senescence. It was concluded that lipid peroxidation was ;2–3 times higher in the outer whorl compared with the inner one, but that these differences were irrelevant to the onset of senescence (Leverentz et al., 2002). It has been concluded that the activity of the main free radical-scavenging enzymes, during leaf senescence, such as superoxide dismutase (SOD), catalase, and ascorbate peroxidase (APX), tend to increase. This might indicate that an increase in free radical production is counteracted, at least in part, by an increase in their degradation (Buchanan-Wollaston, 1997). In petals, the activity of anti-oxidative enzymes such as SOD, catalase, APX, and glutathione reductase (GR) have thus far been assessed only crudely, that is only in vitro and only in whole petals (Droillard et al., 1987; Bartoli et al., 1995; Panavas and Rubinstein, 1998; Bailly et al., 2001). These data are of limited value, as they might not represent in vivo activity. Moreover, different isoforms of these enzymes are present in various cell compartments. These have to be separated before the role of anti-oxidative enzymes can be properly evaluated. Levels of non-enzymatic antioxidants, such as ascorbate and a-tocopherol, have also been determined only in whole petals (Bartoli et al., 1997; Panavas and Rubinstein, 1998), but must also be determined in each compartment separately. Nonetheless, application of anti-oxidants can be used as a method of preliminary assessment of the role of free radicals in senescence. In Dianthus (Paulin et al., 1986; 472 van Doorn and Woltering Droillard et al., 1987) and Hemerocallis (Panavas and Rubinstein, 1998), anti-oxidants delayed the time to visible senescence, but in Iris such compounds were ineffective (Celikel and van Doorn, 1995; Bailly et al., 2001). Taking the data on ROS together, it is concluded that in some species the end-products of lipid peroxidation accumulate during petal senescence, indicating that ROS are involved in membrane degradation. This process may be localized in a specific organelle (such as the vacuole) where it can be part of ongoing remobilization and not be harmful to the cell. In other organelles, the production of ROS may similarly increase. However, there is at present no conclusive evidence that ROS are either an early signal or a direct cause of cellular death during petal senescence. Nevertheless, it cannot at present be excluded that ROS contribute to cell death. The disruption of the tonoplast can, theoretically, come about by processes that involve ROS, causing membrane destabilization. How does tonoplast collapse come about? It is still an enigma how the rupture of the tonoplast occurs. Several hypotheses have been suggested. One is a mechanism whereby phospholipid synthesis is turned off. During senescence in daylily petals, the rate of phospholipid synthesis became lower than the rate of degradation (Bieleski and Reid, 1992). Tonoplast rupture may therefore be the result of a change in the equilibrium between phospholipid breakdown and synthesis. Since many lipid-degrading enzymes seem to be localized to the vacuole, it is remarkable that the tonoplast is not degraded very early. It has been suggested that the access of the degrading enzyme to the tonoplast is prevented by a coating of carbohydrates. Membrane proteins can be glycolysated at the vacuolar side of the membrane, with the carbohydrate parts sticking out, preventing access of hydrolases (Travnicek et al., 1999). This would require, however, a large amount of membrane proteins or very large glycosylation moieties covering the inner side of the membrane. Other hypotheses include the idea of membrane degradation in situ. Lipid de-esterification might result in the accumulation of free fatty acids, which can destabilize and perturb the membrane due a detergent-like action (Hopkins et al., 2007). Alternatively, real detergent-like molecules might accumulate in the vacuole and then insert in the tonoplast. Furthermore, as knockout of a VPEc delayed tonoplast rupture and cell death, induced by a virus (Hara-Nishimura et al., 2005), the activation of proteases, or other enzymes, by VPEs might be important as a regulator of tonoplast rupture. The data in this section show that if tonoplast rupture is the direct cause of cell death during petal senescence, it is as yet completely unknown how it comes about. Several testable hypotheses can be put forward. It is suggested here that future work on petal senescence requires considerably more focus on the processes related to the maintenance and collapse of tonoplast structure. Degradation of cell wall compounds Some types of PCD do not result in much cell wall degradation, for example PCD that leads to the formation of xylem vessels and tracheids. In petals, the degree of cell wall degradation seems to be dependent on the species. In Sandersonia, for example, relatively little wall degradation takes place. The desiccated flower, after senescence, still has the same bell shape as the nonsenescent flower. In other species, for example Ipomoea and Iris, considerable parts of the walls of mesophyll cells become degraded. During petal senescence, cell wall components are broken down both before and after vacuolar collapse. Ultrastructural work in Iris showed that the walls of mesophyll cells became inflated and much less electron dense, indicating degradation. This started prior to vacuolar collapse (van Doorn et al., 2003). In Ipomoea, the content of cellulose and hemicellulose even decreased from the moment of flower opening (Winkenbach, 1970a; Wiemken-Gehrig et al., 1974). Visible senescence symptoms in carnation petals coincided with a lower cell wall dry weight, due largely to a loss of neutral sugars, primarily galactose (45%) and arabinose (23%). Soluble pectin levels in the walls decreased. The amount of high molecular weight hemicelluloses decreased, associated with a decrease in xylose and galactose content (de Vetten and Huber, 1990). In Sandersonia, in contrast, levels of total pectin, cellulose, and high molecular weight hemicelluloses increased only after the first visible senescence symptoms, and despite the previous death of numerous mesophyll cells (O’Donoghue et al., 2002, 2005). Senescence was associated with an increase in the activities of b-galactosidase (Ipomoea, Winkenbach, 1970a; Wiemken-Gehrig et al., 1974; Sandersonia, O’Donoghue et al., 2002, 2005; Hemerocallis, Panavas et al., 1998a), a- and b-glucosidase (Ipomoea, Winkenbach, 1970a; Wiemken-Gehrig et al., 1974), and polygalacturonase (Hemerocallis, Panavas et al., 1998a). In carnation petals undergoing senescence, polygalacturonase activity was not detected (de Vetten and Huber, 1990). In Sandersonia, the activity of xyloglucan endotransglycolase/hydrolose (XET/XTH) was high and declined only by the time of flower wilting (O’Donoghue et al., 2002, 2005). An increase has been found in the transcript abundance of several genes that encode enzymes putatively involved in cell wall degradation (Table 11). These data tend to confirm the biochemical data on enzyme activity. Petal senescence 473 Table 11. Increase in transcript abundance, during petal senescence, of genes encoding enzymes that are putatively involved in cell wall degradation Table 12. Increase in transcript abundance, during petal senescence, of genes encoding enzymes that are putatively involved in defence Gene Species Reference Gene Species Reference b-Galactosidase Iris3hollandica Sandersonia aurantiaca Dianthus caryophyllus Dianthus caryophyllus Alstroemeria pelegrina van Doorn et al. (2003) O’Donoghue et al. (2005) Hoeberichts et al. (2007) Hoeberichts et al. (2007) Breeze et al. (2004)b Avr9/Cf-9 rapidly elicited protein 146 Chitinase IIa Dianthus caryophyllus Class III peroxidase Dianthus caryophyllus Dianthus caryophyllus Iris3hollandica Iris3hollandica Alstroemeria pelegrina Hoeberichts et al. (2007) van Doorn et al. (2003) van Doorn et al. (2003) Breeze et al. (2004)b Heat shock potein 201 Dianthus caryophyllus Polygalacturonase inhibitor-like leucine rich repeat protein Ribosome-inactivating proteins Salicylic acid-binding protein 2 Thaumatin/PR5 protein Dianthus caryophyllus Hoeberichts et al. (2007) Breeze et al. (2004)a Hoeberichts et al. (2007) Hoeberichts et al. (2007) Hoeberichts et al. (2007) Tropinone reductases Iris3hollandica b-Glucosidase b-Xylosidasesa Endoxyloglucan transferase Expansin Pectinesterase XET Xylose isomerase a b Previously described as glycosyl hydrolase family 3 proteins. Data from Supplement 3. The data discussed in this section indicate that cell wall degradation is part of petal senescence. In most species at least part of the cell walls becomes remobilized. a Expression of genes putatively involved in defence Several putative pathogen defence genes were upregulated during petal senescence (Table 12). Tropinone reductases are involved in tropane-alkaloid synthesis, which is stimulated by insect herbivory. They defend against this herbivory. In Iris, many genes encoding ribosome-inactivating proteins were up-regulated (van Doorn et al., 2003). It is not clear if these proteins function as antimicrobial agents or are involved in inhibiting ribosomes in the cells undergoing senescence. The transcript abundance of a putative class III peroxidase (cell wall peroxidase) decreased in Iris petals (van Doorn et al., 2003), but in carnation petals several genes were up-regulated (Hoeberichts et al., 2007). The enzymes may be involved in ROS-mediated defence against pathogens. Plant defence genes seem not to be causally related to remobilization or cell death. They might nonetheless be an integral part of senescence, as numerous cells collapse. Water, containing nutrients such as sugars and amino acids, is then released into the cell walls. The senescent petal might thus be an ideal substrate for micro-organisms, which from there can enter the remainder of the plant. Conclusions The present review points to a number of weaknesses in the present work on petal senescence. Analysis thus far often occurs at the level of whole petals, which always comprise cells that are at various stages of senescence. In the petals of all species, the cells around the phloem remain for a considerably longer time than the other cells. In several species the mesophyll cells die considerably earlier than the epidermis cells. Moreover, visible wilting Alstroemeria pelegrina Iris3hollandica Dianthus caryophyllus Dianthus caryophyllus van Doorn et al. (2003) Hoeberichts et al. (2007) Hoeberichts et al. (2007) van Doorn et al. (2003) Data from Supplement 3. (which tends to reflect the collapse of the epidermis cells) and other symptoms of senescence often start at the petal tips and progress toward the base. The level of precision will probably gain considerably by analysis of cells that are at the same stage of senescence. Moreover, detailed analysis has to include the level of organelles. Topics such as membrane composition, carbohydrate distribution, and ROS clearly need to be investigated at the organelle level, rather than at the level of whole cells. During the last decade, some progress has been made in understanding the physiological and molecular background of senescence in general. Ethylene receptors have been discovered in other plants, and several receptor genes have been isolated from petals. In addition, proteins involved in the ethylene signal transduction pathway have been discovered, and similar proteins have been found in senescing petals. One of the findings in (carnation) petals is the apparent large effect of EIL3. Numerous genes were found to be differentially expressed during petal senescence, and the putative functions of many of the sequenced genes have been suggested. The main processes suggested by these changes are remobilization and transport of mobile compounds out of the petal. The work suggests the importance of proteasomal protein degradation, and the apparent role of sugars in preventing the up-regulation of numerous genes that are controlled by ethylene. In addition, the gene expression studies revealed upregulation of several factors associated with plant defence. However, thus far, the work on gene expression has answered questions neither regarding the onset of the senescence program, nor about the way the cells die. Although several signal transduction and transcription 474 van Doorn and Woltering factors have been found, their role is as yet unknown. Functional analysis is likely to reveal some of the regulatory networks controlling the onset and the orderly progress of senescence. A main discovery has been the requirement of the ovary for carnation petal senescence. This requirement might also be true for other species with ethylene-sensitive petal senescence, at least those where the pollination signal has to pass the ovary before it can reach the petals. There remain several major gaps in our present understanding. These will be discussed here in the temporal order of the senescence process, from its initiation to the execution of cellular death. Some testable hypotheses will also be put forward. Some of the main questions that are still outstanding are as follows. flowers, this rise in ethylene production also occurs. It is hypothesized that it is due to a decrease in cytokinin levels, localized, at least in a species like carnation, in the ovary. If so, the question changes to what induces the decrease in cytokinin levels. (iv) What hormonal factors, if any, determine the time to visible ethylene-independent senescence? It is hypothesized here that this type of senescence, although it can be affected by hormones, is not naturally regulated by any hormone, but by gene expression in the petal cells. It is further hypothesized that ethylene-insensitive senescence is the basal pathway, and ethylene-sensitive senescence an additional level of control allowing pollination to advance the time to senescence. (v) (i) What determines the onset of the senescence process? In several species tested, changes in cellular ultrastructure, typical for senescence, occur prior to or during flower opening (e.g. Alstroemeria; Wagstaff et al., 2003; Iris, van Doorn et al., 2003). It is hypothesized here that, as in other developmental changes, a regulatory network determines the change from the non-senescent to the senescent stage. It is further hypothesized that the initiation of these changes is due to a time-controlled expression of genes within each cell. As the cells are at this stage connected by plasmodesmata, whereby the mesophyll cells and the epidermis cells belong to different multicellular domains (van Doorn et al., 2003), the earlier onset of senescence in mesophyll cells, found in several species, can be explained by a difference in the timing of the senescence signal in different domains. (ii) (iii) What determines the life span of petals? The time from flower opening to petal senescence (in pollinated as well as unpollinated flowers) is, in each species, highly constant, which indicates tight temporal control. It is hypothesized that this control is due to differences in the timing of the normal sequence of senescence events, and thus to differences in the changes in gene expression. In several species a circadian clock is involved in the timing of flower opening (van Doorn and van Meeteren, 2004). It might also be involved in the regulation of the timing of senescence events. What induces the rise in ethylene production in pollinated and unpollinated ethylene-sensitive flowers? It is clear that the pollination signal (which by itself is still not fully known; O’Neill, 1997) promotes the expression of genes in the ethylene synthesis pathway, thereby producing the rise in ethylene production in petals. In unpollinated Is cell death due to remobilization? Death at the end of the senescence process might be due to ongoing remobilization or might occur independently of remobilization. The data on cell death in Zinnia tracheal elements indicate tonoplast rupture long before most of the cell has been remobilized (Obara et al., 2001). This suggests that tonoplast rupture is not due to a degree of remobilization whereby the cells die of lack of energy and structure. It is hypothesized here that tonoplast rupture is the immediate cause of cell death during petal senescence, and that it occurs independently of remobilization. (vi) What, if tonoplast rupture is the immediate cause of death, induces it? Theoretically, numerous processes might be responsible. One is the activation of membrane-bound phospholipid-degrading enzymes; another is the disappearance of a mechanism that protects the tonoplast from being attacked by lipases in the vacuole. Focus on the role of the tonoplast in cell death, and the cause of its rupture, is one of the several challenges for future research on petal senescence. References Ahmed K, Gerber DA, Cochet C. 2002. Joining the cell survival squad: an emerging role for protein kinase CK2. Trends in Cell Biology 12, 226–230. Aida R, Yoshida T, Ichimura K, Goto R, Shibata M. 1998. Extension of flower longevity in transgenic Torenia plants incorporating ACC oxidase transgene. Plant Science 138, 91–101. Apel K, Hirt H. 2004. Reactive oxygen species: metabolism, oxidative stress, and signal transduction. Annual Review of Plant Biology 55, 373–399. Arazi T, Kaplan B, Fromm H. 2000. A high-affinity calmodulinbinding site in a tobacco plasma-membrane channel protein coincides with a characteristic element of cyclic nucleotidebinding domains. Plant Molecular Biology 42, 591–601. Arora A, Singh VP. 2006. Polyols regulate the flower senescence by delaying programmed cell death in Gladiolus. Journal of Plant Biochemistry and Biotechnology 15, 139–142. Petal senescence 475 Arora A, Watanabe S, Ma B, Takada K, Ezura H. 2006. A novel ethylene receptor homolog gene isolated from ethyleneinsensitive flowers of gladiolus (Gladiolus grandiflora hort. Biochemical and Biophysical Research Communications 351, 739–744. Bailly C, Corbineau F, van Doorn WG. 2001. Free radical scavenging and senescence in Iris tepals. Plant Physiology and Biochemistry 39, 649–656. Balamurugan K, Schaffner W. 2006. Copper homeostasis in eukaryotes: teetering on a tightrope. Biochimica et Biophysica Acta 1763, 737–746. Bancher E. 1941. Zellphysiologische Beobachtungen an Iris reichenbachii während des Abblühens. Österreichischen Botanischen Zeitschrift 90, 97–106. Bartoli CG, Simontacchi M, Guiamet JJ, Montaldi E, Puntarolo S. 1995. Antioxidant enzymes and lipid peroxidation during aging of Chrysanthemum morifolium Ram petals. Plant Science 104, 161–168. Bartoli CG, Simontacchi M, Montaldi E, Puntarolo S. 1997. Oxidants and antioxidants during aging of chrysanthemum petals. Plant Science 129, 157–165. Bieleski RL. 1995. Onset of phloem export from senescent petals of daylily. Plant Physiology 109, 557–565. Bieleski RL, Reid MS. 1992. Physiological changes accompanying senescence in the ephemeral daylily flower. Plant Physiology 98, 1042–1049. Bishopp A, Mähönen AP, Helariutta Y. 2006. Signs of change: hormone receptors that regulate plant development. Development 133, 1857–1869. Borochov A, Cho MH, Boss WF. 1994. Plasma membrane lipid metabolism of Petunia petals during senescence. Physiologia Plantarum 90, 279–284. Borochov A, Woodson WR. 1989. Physiology and biochemistry of flower petal senescence. Horticultural Reviews 11, 15–43. Bozhkov PV, Filonova LH, Suarez MF, Helmersson A, Smertenko AP, Zhivotovsky B, von Arnold S. 2004. VEIDase is a principal caspase-like activity involved in plant programmed cell death and essential for embryonic pattern formation. Cell Death and Differentiation 11, 175–182. Bovy AG, Angenent GC, Dons HJM, van Altvorst AC. 1999. Heterologous expression of the Arabidopsis etr1-1 allele inhibits the senescence of carnation flowers. Molecular Breeding 5, 301–308. Breeze E, Wagstaff C, Harrison E, Bramke I, Rogers H, Stead A, Thomas B, Buchanan-Wollaston V. 2004. Gene expression patterns to define stages of post-harvest senescence in Alstroemeria petals. Plant Biotechnology Journal 2, 155–168. Brown JH, Palyath G, Thompson JE. 1990. Influence of acyl length composition on the degradation of phosphatidylcholine by phospholipase D in carnation microsomal membranes. Journal of Experimental Botany 41, 979–986. Brugière S, Kowalski S, Ferro M, et al. 2004. The hydrophobic proteome of mitochondrial membranes from Arabidopsis cell suspensions. Phytochemistry 65, 1693–1707. Buchanan-Wollaston V. 1997. The molecular biology of leaf senescence. Journal of Experimental Botany 48, 181–199. Castelli J, Wood KA, Youle RJ. 1998. The 2–5A system in viral infection and apoptosis. Biomedicine and Pharmacotherapy 52, 386–390. Celikel F, van Doorn WG. 1995. Solute leakage, lipid peroxidation, and protein degradation during the senescence of Iris tepals. Physiologia Plantarum 94, 515–521. Chang H, Jones ML, Banowetz GM, Clark DG. 2003. Overproduction of cytokinins in petunia flowers transformed with PSAG12-IPT delays corolla senescence and decreases sensitivity to ethylene. Plant Physiology 132, 2174–2183. Chang MS, Chang CL, Huang CJ, Yang YC. 2000. p29, a novel GCIP-interacting protein, localizes in the nucleus. Biochemical and Biophysical Research Communications 279, 732–737. Chen ZH, Walker RP, Acheson RM, Tecsi LI, Wingler A, Lea PJ, Leegood RC. 2000. Are isocitrate lyase and phosphoenolpyruvate carboxykinase involved in gluconeogenesis during senescence of barley leaves and cucumber cotyledons? Plant and Cell Physiology 41, 960–967. Cheng WC, Berman SB, Ivanovska I, Jonas EA, Lee SJ, Chen Y, Kaczmarek LK, Pineda F, Hardwick JM. 2006. Mitochondrial factors with dual roles in death and survival. Oncogene 25, 4697–4705. Coffeen WC, Wolpert TJ. 2004. Purification and characterization of serine proteases that exhibit caspase-like activity and are associated with programmed cell death in Avena sativa. The Plant Cell 16, 857–873. Cornah JE, Germain V, Ward JL, Beale MH, Smith SM. 2004. Lipid utilization, gluconeogenesis, and seedling growth in Arabidopsis mutants lacking the glyoxylate cycle enzyme malate synthase. Journal of Biological Chemistry 279, 42916–42923. Courtney SE, Rider CC, Stead AD. 1994. Changes in protein ubiquitination and the expression of ubiquitin-encoding transcripts in daylily petals during floral development and senescence. Physiologia Plantarum 91, 196–204. Cui ML, Takada K, Ezura H. 2004. Overexpression of a mutated melon ethylene receptor gene Cm-ETR1/H69A confers reduced ethylene sensitivity in a heterologous plant, Nemesia strumosa. Plant Science 167, 253–258. Däschner K, Couée I, Binder S. 2001. The mitochondrial isovaleryl-coenzyme A dehydrogenase of Arabidopsis oxidizes intermediates of leucine and valine catabolism. Plant Physiology 126, 601–612. de Vetten NC, Huber DJ. 1990. Cell wall changes during the expansion and senescence of carnation (Dianthus caryophyllus) petals. Physiologia Plantarum 78, 447–454. Do YY, Huang PL. 1997. Gene structure of PACO1, a petal senescence-related gene from Phalaenopsis encoding peroxisomal acyl-CoA oxidase homolog. Biochemistry and Molecular Biology International 41, 609–618. Domı́nguez F, Cejudo FJ. 2006. Identification of a nuclearlocalized nuclease from wheat cells undergoing programmed cell death that is able to trigger DNA fragmentation and apoptotic morphology on nuclei from human cells. Biochemical Journal 397, 529–536. Droillard MJ, Paulin A, Massot JC. 1987. Free radical production, catalase and superoxide dismutase activities and membrane integrity during senescence of petals of cut carnations (Dianthus caryophyllus). Physiologia Plantarum 71, 197–202. Eason JR, Johnston JW, de Vre L, Sinclair BK, King GA. 2000. Amino acid metabolism in senescing Sandersonia aurantiaca flowers: cloning and characterization of asparagine synthetase and glutamine synthetase cDNAs. Australian Journal of Plant Physiology 27, 389–396. Eason JR, Ryan DJ, Pinkney TT, O’Donoghu EM. 2002. Programmed cell death during flower senescence: isolation and characterization of cysteine proteinases from Sandersonia aurantiaca. Functional Plant Biology 29, 1055–1064. Eisinger W. 1977. Role of cytokinins in carnation flower senescence. Plant Physiology 59, 707–709. Etheridge N, Hall BP, Schaller GE. 2006. Progress report: ethylene signalling and response. Planta 223, 387–391. Eze JMO, Mayak S, Thompson JE, Dumbroff EB. 1986. Senescence of cut carnation flowers: temporal and physiological relationships among water status, ethylene, abscisic acid and membrane permeability. Physiologia Plantarum 68, 323–328. 476 van Doorn and Woltering Fobel M, Lynch DV, Thompson JE. 1987. Membrane deterioration in senescing carnation flowers. Plant Physiology 85, 204– 211. Gavrieli Y, ShermanY Ben-Sasson SA. 1992. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. Journal of Cell Biology 119, 493–501. Graham IA, Denby KJ, Leaver CJ. 1994. Carbon catabolite repression regulates glyoxylate cycle gene expression in cucumber. The Plant Cell 6, 761–772. Greenwood JS, Helm M, Gietl C. 2005. Ricinosomes and endosperm transfer cell structure in programmed cell death of the nucellus during Ricinus seed development. Proceedings of the National Academy of Sciences, USA 102, 2238–2243. Greiner S, Koster U, Lauer K, Rosenkranz H, Vogel R, Rausch T. 2000. Plant invertase inhibitors: expression in cell culture and during plant development. Australian Journal of Plant Physiology 27, 807–814. Guo Y, Cai Z, Gan S. 2004. Transcriptome of Arabidopsis leaf senescence. Plant, Cell and Environment 27, 521–549. Hail N, Carter BZ, Konopleva M, Andreeff M. 2006. Apoptosis effector mechanisms: a requiem performed in different keys. Apoptosis 11, 889–904. Halaba J, Rudnicki RM. 1989. Invertase inhibitor in wilting flower petals. Scientia Horticulturae 40, 83–90. Hall J. 2002. Cellular mechanisms for heavy metal detoxification and tolerance. Journal of Experimental Botany 53, 1–11. Hanley KM, Bramlage WJ. 1989. Endogenous levels of abscisic acid in aging carnation flower parts. Journal of Plant Growth Regulation 8, 225–236. Hanaoka H, Noda T, Shirano Y, Kato T, Hayashi H, Shibata D, Tabata S, Ohsumi Y. 2002. Leaf PCD and starvation-induced chlorosis are accelerated by the disruption of an Arabidopsis autophagy gene. Plant Physiology 129, 1181–1193. Hara-Nishimura I, Hatsugai N, Nakaune S, Kuroyanagi M, Nishimura M. 2005. Vacuolar processing enzyme: an executor of plant cell death. Current Opinion in Plant Biology 8, 404– 408. He XJ, Mu RL, Cao WH, Zhang ZG, Zhang JS, Chen SY. 2005. AtNAC2, a transcription factor downstream of ethylene and auxin signaling pathways, is involved in salt stress response and lateral root development. The Plant Journal 44, 903–916. Hew CS, Tan SC, Chin TY, Ong TK. 1989. Influence of ethylene on enzyme activities and mobilisation of materials in pollinated Arachnis orchid flowers. Journal of Plant Growth and Regulation 8, 121–130. Hoeberichts FA, de Jong AJ, Woltering EJ. 2005. Apoptotic-like cell death marks the early stages of gypsophila (Gypsophila paniculata) petal senescence. Postharvest Biology and Technology 35, 229–236. Hoeberichts FA, van Doorn WG, Vorst O, Hall RD, van Wordragen MF. 2007. Sucrose prevents upregulation of senescence-associated genes in carnation petals. Journal of Experimental Botany 58, 2873–2885. Hopkins M, Taylor C, Liu Z, Ma F, McNamara L, Wang TW, Thompson JE. 2007. Regulation and execution of molecular disassembley and catabolism during senescence. New Phytologist 175, 201–214. Hunter DA, Ferrante A, Vernieri P, Reid MS. 2004b. Role of abscisic acid in perianth senescence of daffodil (Narcissus pseudonarcissus ‘Dutch Master’). Physiologia Plantarum 121, 313–321. Hunter DA, Steele BC, Reid MS. 2002. Identification of genes associated with perianth senescence in daffodil (Narcissus pseudonarcissus L. ‘Dutch Master’). Plant Science 163, 13–21. Hunter DA, Yi MF, Xu XJ, Reid MS. 2004a. Role of ethylene in perianth senescence of daffodil (Narcissus pseudonarcissus L. ‘Dutch Master’). Postharvest Biology and Technology 32, 269–280. Ionoue Y, Moriyasu Y. 2006. Autophagy is not a main contributor to the degradation of phospholipids in tobacco cells cultured under sucrose starvation conditions. Plant and Cell Physiology 47, 471–480. Iordachescu M, Verlinden S. 2005. Transcriptional regulation of three EIN3-like genes of carnation (Dianthus caryophyllus L. cv. Improved White Sim) during flower development and upon wounding, pollination, and ethylene exposure. Journal of Experimental Botany 56, 2011–2018. Itzhaki H, Maxson JM, Woodson WR. 1994. An ethyleneresponsive enhancer element is involved in the senescence-related expression of the carnation glutathione-S-transferase (GST1) gene. Proceedings of the National Academy of Sciences, USA 91, 8925–8929. Iwaya-Inoue M, Nonami H. 2003. Effects of trehalose on flower senescence from the view point of physical states of water. Environment Control in Biology 41, 3–15. Jones AM. 2001. Programmed cell death in development and defense. Plant Physiology 125, 94–97. Jones AM, Dangl JL. 1996. Logjam at the Styx: programmed cell death in plants. Trends in Plant Science 1, 114–119. Jones ML. 2004. Changes in gene expression during senescence. In: Nooden LD, ed. Plant cell death processes. Amsterdam: Elsevier, 51–71. Jones ML, Chaffin GS, Eason JR, Clark DG. 2005. Ethylenesensitivity regulates proteolytic activity and cysteine protease gene expression in petunia corollas. Journal of Experimental Botany 56, 2733–2744. Jones ML, Larsen PB, Woodson WR. 1995. Ethylene-regulated expression of a carnation cysteine proteinase during flower petal senescence. Plant Molecular Biology 28, 505–512. Jones ML, Woodson WR. 1999. Differential expression of three members of the 1-aminocyclopropane-1-carboxylate synthase gene family in carnation. Plant Physiology 119, 755–764. Kaur N, Palta JP. 1997. Postharvest dip in lysophosphatidylethanolamine, a natural phospholipid, may prolong vase-life of snapdragon flowers. HortScience 32, 888–890. Kiningham K, Kasarskis E. 1998. Antioxidant function of metallothioneins. Journal of Trace Elements in Experimental Medicine 11, 219–226. Köhler C, Merkle T, Roby D, Neuhaus G. 2001. Developmentally regulated expression of a cyclic nucleotide-gated ion channel from Arabidopsis indicates its involvement in programmed cell death. Planta 213, 327–332. Kosugi Y, Waki K, Iwazaki Y, Tsuruno N, Mochizuki A, Yoshioka T, Hashiba T, Satoh S. 2002. Senescence and gene expression of transgenic non-ethylene-producing carnation flowers. Journal of the Japanese Society for Horticultural Science 71, 638–642. Langston BJ, Bai S, Jones ML. 2005. Increases in DNA fragmentation and induction of a senescence-specific nuclease are delayed during corolla senescence in ethylene-insensitive (etr1-1) transgenic petunias. Journal of Experimental Botany 56, 15–23. Leverentz MK, Wagstaff C, Rogers HJ, Stead AD, Chanasut U, Silkowski H, Thomas B, Weichert H, Feussner I, Griffiths G. 2002. Characterisation of a novel lipoxygenase-independent senescence mechanism in Alstroemeria floral tissue. Plant Physiology 130, 273–283. Lu YP, Li ZS, Drozdowicz YM, Hörtensteiner S, Martinoia E, Rea PA. 1998. AtMRP2, an Arabidopsis ATP binding cassette transporter able to transport glutathione s-congugates and chlorophyll catabolites: functional comparisons with AtMRP1. The Plant Cell 10, 267–282. Petal senescence 477 Ma W, Stafford LJ, Li D, Luo J, Li X, Ning G, Liu M. 2007. GCIP/CCNDBP1, a helix–loop–helix protein, suppresses tumorigenesis. Journal of Cellular Biochemistry 100, 1376–1386. Madey E, Nowack LM, Su L, Hong Y, Hudak KA, Thompson JE. 2001. Characterization of plasma membrane domains enriched in lipid metabolites. Journal of Experimental Botany 52, 669–679. Makishima T, Yoshimi M, Komiyama S, Hara N, Nishimoto T. 2000. A subunit of the mammalian oligosaccharyltransferase, DAD1, interacts with Mcl-1, a member of the Bcl-2 protein family. Journal of Biochemistry 128, 399–405. Martı́nez-Garcı́a JF, Quail PH. 1999. The HMG-I/Y protein PF1 stimulates binding of the transcriptional activator GT-2 to the PHYA gene promoter. The Plant Journal 18, 173–183. Mason MG, Schaller GE. 2005. Histidine kinase activity and the regulation of ethylene signal transduction. Canadian Journal of Botany 83, 563–570. Matile P, Winkenbach F. 1971. Function of lysosomes and lysosomal enzymes in the senescing corolla of the morning glory (Ipomoea purpurea). Journal of Experimental Botany 22, 759–771. Mayak S, Dilley D. 1976. Effect of sucrose on response of cut carnation flowers to kinetin, ethylene and abscisic acid. Journal of the American Society of Horticultural Science 101, 583–585. Mayak S, Tirosh T. 1993. Unusual ethylene-related behavior in senescing flowers of the carnation Sandrosa. Physiologia Plantarum 88, 420–426. McLaughlin JC, Smith SM. 1994. Metabolic regulation of glyoxylate-cycle enzyme synthesis in detached cucumber cotyledons and protoplasts. Planta 195, 22–28. Michael M, Savin KW, Baudinette SC, et al. 1993. Cloning of ethylene biosynthetic genes involved in petal senescence of carnation and petunia, and their antisense expression in transgenic plants. In: Pech JC, Latché A, Balagué C, eds. Cellular and molecular aspects of the plant hormone ethylene. Dordrecht, The Netherlands: Kluwer, 298–303. Müller F, Ádori C, Sass M. 2004. Autophagic and apoptotic features during programmed cell death in the fat body of the tobacco hornworm (Manduca sexta). European Journal of Cell Biology 83, 67–78. Nagata S, Nagase H, Kawane K, Mukae N, Fukuyama H. 2003. Degradation of chromosomal DNA during apoptosis. Cell Death and Differentiation 10, 108–116. Nakaune S, Yamada K, Kondo M, Kato T, Tabata S, Nishimura M, Hara-Nishimura I. 2005. A vacuolar processing enzyme, delta VPE, is involved in seed coat formation at the early stage of seed development. The Plant Cell 17, 876–887. Nelson DR, Schuler MA, Paquette SM, Werck-Reichhart D, Bak S. 2004. Comparative genomics of rice and Arabidopsis. Analysis of 727 cytochrome P450 genes and pseudogenes from a monocot and a dicot. Plant Physiology 135, 756–772. Nichols R. 1966. Ethylene production during senescence of cut flowers. Journal of Horticultural Science 41, 279–290. Nowak J, Veen H. 1982. Effects of silver thiosulfate on abscisic acid content in cut carnations as related to flower senescence. Journal of Plant Growth Regulation 1, 153–159. Nukui H, Kudo S, Yamashita A, Satoh S. 2004. Repressed ethylene production in the gynoecium of long-lasting flowers of the carnation ‘White Candle’: role of the gynoecium in carnation flower senescence. Journal of Experimental Botany 55, 641–650. Obara K, Kuriyama H, Fukuda H. 2001. Direct evidence of active and rapid nuclear degradation triggered by vacuole rupture during programmed cell death in Zinnia. Plant Physiology 125, 615–626. O’Donoghue E, Eason JR, Somerfield SD, Ryan DA. 2005. Galactosidases in opening, senescing and water-stressed Sandersonia aurantiaca flowers. Functional Plant Biology 32, 911–922. O’Donoghue E, Somerfield SD, Heyes JA. 2002. Organisation of cell walls in Sandersonia aurantiaca floral tissue. Journal of Experimental Botany 53, 513–523. Oliver DJ. 1994. The glycine decarboxylase complex from plant mitochondria. Annual Review of Plant Physiology and Plant Molecular Biology 45, 323–337. O’Neill SD. 1997. Pollination regulation of flowers development. Annual Review of Plant Physiology and Plant Molecular Biology 48, 547–574. Onoue T, Mikami M, Yoshioka T, Hashiba T, Satoh S. 2000. Characteristics of the inhibitory action of 1,1-dimethyl-4(phenylsulfonyl)semicarbazide (DPSS) on ethylene production in carnation (Dianthus caryophyllus L.) flowers. Plant Growth Regulation 30, 201–207. Ori N, Juarez MT, Jackson D, Yamaguchi J, Banowetz GM, Hake GM. 1999. Leaf senescence is delayed in tobacco plants expressing the maize homeobox gene knotted1 under the control of a senescence-acitivated promoter. The Plant Cell 11, 1073–1080. Orzaez D, Granell A. 1997. The plant homologue of the defender against apoptotic death gene is down-regulated during senescence of flower petals. FEBS Letters 404, 275–278. Otsubo M, Iwaya-Inoue M. 2000. Trehalose delays senescence in cut gladiolus spikes. HortScience 35, 1107–1110. Overmyer K, Brosche M, Pellinen R, Kuittinen T, Tuominen H, Ahlfors R, Keinanen M, Saarma M, Scheel D, Kangasjarvi J. 2005. Ozone-induced programmed cell death in the Arabidopsis radical-induced cell death1 mutant. Plant-Physiology 137, 1092– 1104. Pak C, van Doorn WG. 2005. Delay of Iris flower senescence by protease inhibitors. New Phytologist 165, 473–480. Panavas T, LeVangie R, Mistler J, Reid PD, Rubinstein B. 2000. Activities of nucleases in senescing daylily petals. Plant Physiology and Biochemistry 38, 837–843. Panavas T, Pikula A, Reid PD, Rubinstein B, Walker EL. 1999. Identification of senescence-associated genes from daylily petals. Plant Molecular Biology 40, 237–248. Panavas T, Reid PD, Rubinstein B. 1998a. Programmed cell death of daylily petals: activities of wall-based enzymes and effects of heat shock. Plant Physiology and Biochemistry 36, 379–388. Panavas T, Rubinstein B. 1998. Oxidative events during programmed cell death of daylily (Hemerocallis hybrid) petals. Plant Science 133, 125–138. Panavas T, Walker EL, Rubinstein B. 1998b. Possible involvement of abscisic acid in senescence of daylily petals. Journal of Experimental Botany 49, 1987–1997. Paulin A, Droillard MJ, Bureau JM. 1986. Effect of a free radical scavenger, 3,4,5-trichloro-phenol, on ethylene production and on changes in lipids and membrane integrity during senescence of petals of cut carnations (Dianthus caryophyllus). Physiologia Plantarum 67, 465–471. Phillips HL, Kende H. 1980. Structural changes in flowers of Ipomoea tricolor during flower opening and closing. Protoplasma 102, 199–215. Phillis E, Mason TG. 1936. Further studies on the transport in the cotton plant. VI. Interchange between the tissues of the corolla. American Journal of Botany 50, 679–697. Pistelli L, Perata P, Alpi A. 1992. Effect of leaf senescence on glyoxylate cycle enzyme activities. Australian Journal of Plant Physiology 19, 723–729. Porat R, Borochov A, Halevy AH. 1993. Enhancement of petunia and dendrobium flower senescence by jasmonic acid methyl ester is via the promotion of ethylene production. Plant Growth Regulation 13, 297–301. Porat R, Reiss N, Atzorn R, Halevy AH, Borochov A. 1995. Examination of the possible involvement of lipoxygenase and 478 van Doorn and Woltering jasmonates in pollination-induced senescence of Phalaenopsis and Dendrobium orchid flowers. Physiologia Plantarum 94, 205–210. Rinnerthaler M, Jarolim S, Palle E, et al. 2006. MMI1 (YKL056c, TMA19), the yeast orthologue of the translationally controlled tumor protein (TCTP) has apoptotic functions and interacts with both microtubules and mitochondria. Biochimica et Biophysica Acta 1757, 631–638. Rojo E, Zouhar J, Carter C, Kovaleva V, Raikhel NV. 2003. A unique mechanism for protein processing and degradation in Arabidopsis thaliana. Proceedings of the National Academy of Sciences, USA 100, 7389–7394. Rojo E, Martin R, Carter C, et al. 2004. VPEc exhibits a caspase-like activity that contributes to defence against pathogens. Current Biology 14, 1897–1906. Sanderfoot AA, Pilgrim M, Adam L, Raikhel NV. 2001. Disruption of individual members of Arabidopsis syntaxin gene families indicates each has essential functions. The Plant Cell 13, 659–666. Saks Y, van Staden J. 1993. Evidence for the involvement of gibberellins in developmental phenomena associated with carnation flower senescence. Plant Growth Regulation 12, 105–110. Satterfield TF, Jackson SM, Pallanck LJ. 2002. A Drosophila homolog of the polyglutamate disease gene SCA2 is a dosage dependent regulator of actin filament formation. Genetics 162, 1687–1702. Savin KW, Baudinette SC, Graham MW, Michael MZ, Nugent GD, Lu CY, Chandler SF, Cornish EC. 1995. Antisense ACC oxidase RNA delays carnation petal senescence. Hortscience 30, 970–972. Schmid M, Simpson D, Gietl C. 1999. Programmed cell death in castor bean endosperm is associated with the accumulation and release of cysteine endopeptidase from ricinosomes. Proceedings of the National Academy of Sciences, USA 96, 14159–14164. Serafini-Fracassini D, Del Duca S, Monti F, Poli F, Sacchetti G, Bregoli AM, Biondi S, Della Mea M. 2002. Transglutaminase activity during senescence and programmed cell death in the corolla of tobacco (Nicotiana tabacum) flowers. Cell Death and Differentiation 9, 309–321. Shaw JF, Chen HH, Tsai MF, Kuo C, Huang LC. 2002. Extended flower longevity of Petunia hybrida plants transformed with boers, a mutated ERS gene of Brassica oleracea. Molecular Breeding 9, 211–216. Shibuya K, Barry KG, Ciardi JA, Loucas HM, Underwood BA, Nourizadeh S, Ecker JR, Klee HJ, Clark DG. 2004. The central role of PhEIN2 in ethylene responses throughout plant development in petunia. Plant Physiology 136, 2900–2912. Shibuya K, Nagata M, Tanikawa N, Yoshioka T, Hashiba T, Satoh S. 2002. Comparison of mRNA levels of three ethylene receptors in senescing flowers of carnation (Dianthus caryophyllus L.). Journal of Experimental Botany 53, 399–406. Shibuya K, Yoshioka T, Hashiba T, Satoh S. 2000. Role of the gynoecium in natural senescence of carnation (Dianthus caryophyllus L.) flowers. Journal of Experimental Botany 51, 2067–2073. Sinvany-Villalobo G, Davydov O, Ben-Ari G, Zaltsman A, Raskind A, Adam Z. 2004. Expression in multigene families. Analysis of chloroplast and mitochondrial proteases. Plant Physiology 135, 1336–1345. Smith MT, Saks Y, van Staden J. 1992. Ultrastructural changes in the petals of senescing flowers of Dianthus caryophyllus L. Annals of Botany 69, 277–285. Solovyan VT, Keski-Oja J. 2006. Proteolytic activation of latent TGF-beta precedes caspase-3 activation and enhances apoptotic death of lung epithelial cells. Journal of Cellular Physiology 207, 445–453. Stead AD, van Doorn WG. 1994. Strategies of flower senescence— a review. In: Scott RJ, Stead AD, eds. Molecular and cellular aspects of plant reproduction. Cambridge: Cambridge University Press, 215–238. Stephenson P, Rubinstein B. 1998. Characterization of proteolytic activity during senescence in daylilies. Physiologia Plantarum 104, 463–473. Sugawara H, Shibuya K, Yoshioka T, Hashiba T, Satoh S. 2002. Is a cysteine proteinase inhibitor involved in the regulation of petal wilting in senescing carnation (Dianthus caryophyllus L.) flowers? Journal of Experimental Botany 53, 407–413. Suttle JC, Kende H. 1980. Ethylene action and loss of membrane intregrity during petal senescence in Tradescantia. Plant Physiology 65, 1067–1072. Taverner EA, Letham DS, Wang J, Cornish E. 2000. Inhibition of carnation petal inrolling by growth retardants and cytokinins. Australian Journal of Plant Physiology 27, 357–362. ten Have A, Woltering EJ. 1997. Ethylene biosynthetic genes are differentially expressed during carnation (Dianthus caryophyllus L.) flower senescence. Plant Molecular Biology 34, 89–97. Thomas SG, Huang S, Li S, Staiger CJ, Franklin-Tong VE. 2006. Actin depolymerisation is sufficient to induce programmed cell death in self-incompatible pollen. Journal of Cell Biology 174, 221–229. Thompson AR, Doelling JH, Suttangkakul A, Vierstra RD. 2005. Autophagic nutrient recycling in Arabidopsis directed by the ATG8 and ATG12 conjugation pathways. Plant Physiology 138, 2097–2110. Thompson AR, Vierstra RD. 2005. Autophagic recycling: lessons from yeast help define the process in plants. Current Opinion in Plant Biology 8, 165–173. Thompson JE, Froese CD, Madey E, Smith MD, Hong Y. 1998. Lipid metabolism during plant senescence. Progress in Lipid Research 37, 119–141. Thompson JE, Taylor C, Wang TW. 2000. Altered membrane lipase expression delays leaf senescence. Biochemical Society Transactions 28, 775–777. Travnicek I, Rodoni S, Schellenberg M, Matile P. 1999. A thiobased phospholipase assay employed for the study of lipolytic acyl hydrolase in the ephemeral corolla of Ipomoea tricolor. Journal of Plant Physiology 155, 220–225. Uemura T, Ueda T, Ohniwa RL, Nakano A, Takeyasu K, Sato MH. 2004. Systematic analysis of SNARE molecules in Arabidopsis: dissection of the post-Golgi network in plant cells. Cell Structure and Function 29, 49–65. Valpuesta V, Lange NE, Guerrero C, Reid MS. 1995. Upregulation of a cysteine protease accompanies the ethyleneinsensitive senescence of daylily (Hemerocallis) flowers. Plant Molecular Biology 28, 575–582. Van Breusegem F, Dat JF. 2005. Reactive oxygen species in plant cell death. Plant Physiology 141, 384–390. van der Kop DAM, Ruys G, Dees D, van der Schoot C, de Boer AD, van Doorn WG. 2003. Expression of defender against apoptotic death (DAD-1) in Iris and Dianthus petals. Physiologia Plantarum 117, 256–263. van der Meulen-Muisers JJM . 2000. Genetic and physiological aspects of postharvest flower longevity in Asiatic hybrid lilies (Lilium L.). PhD Thesis, Wageningen University, Wageningen. The Netherlands.. van der Meulen-Muisers JJM, van Oeveren JC, Meijkamp BB, Derks FHM. 1995. Effect of floral bud reduction on individual flower longevity in Asiatic hybrid lilies. Acta Horticulturae 405, 46–57. van der Meulen-Muisers JJM, van Oeveren JC, van der Plas LHW, van Tuyl JM. 2001. Postharvest flower development Petal senescence 479 in Asiatic hybrid lilies as related to tepal carbohydrate status. Postharvest Biology and Technology 21, 201–211. van Doorn WG. 1997. Effects of pollination on floral attraction and longevity. Journal of Experimental Botany 48, 1615–1622. van Doorn WG. 2001. Categories of petal senescence and abscission: a re-evaluation. Annals of Botany 87, 447–456. van Doorn WG. 2004. Is petal senescence due to sugar starvation? Plant Physiology 134, 35–42. van Doorn WG, Balk PA, van Houwelingen AM, Hoeberichts FA, Hall RD, Vorst O, van der Schoot C, van Wordragen MF. 2003. Gene expression during anthesis and senescence in Iris flowers. Plant Molecular Biology 53, 845–863. van Doorn WG, Sinz A, Tomassen MM. 2004. Daffodil flowers delay senescence in cut Iris flowers. Phytochemistry 65, 571–577. van Doorn WG, Stead AD. 1994. The physiology of petal senescence which is not initiated by ethylene. In: Scott RJ, Stead AD, eds. Molecular and cellular aspects of plant reproduction. Cambridge: Cambridge University Press, 239–254. van Doorn WG, Woltering EJ. 2004. Senescence and programmed cell death: substance or semantics? Journal of Experimental Botany 55, 2147–2153. van Doorn WG, Woltering EJ. 2005. Many ways to exit? Cell death categories in plants. Trends in Plant Science 10, 117–122. van Doorn WG, Woltering EJ, Reid MS, Wu MJ. 1993. Reduced sensitivity to ethylene in a group of related carnation cultivars. In: Pech JC, Latché A, Balagué C, eds. Cellular and molecular aspects of the plant hormone ethylene. Dordrecht, The Netherlands: Kluwer, 188–194. van Staden J. 1995. Hormonal control of carnation flower senescence. Acta Horticulturae 405, 232–239. van Staden J, Dimalla GG. 1980. The effect of silver thiosulphate preservative on the physiology of cut carnations. II. Influence on endogenous cytokinin. Zeitschrift für Pflanzenphysiologie 99, 19–26. van Staden J, Featonby-Smith BC, Mayak S, Spiegelstein H, Halevy AH. 1987. Cytokinins in carnation flowers. II. Relationship between endogenous ethylene and cytokinin levels. Plant Growth Regulation 5, 75–82. Verheye S, Martinet W, Kockx MM, Knaapen MWM, Salu K, Timmermans JP, Ellis JT, Kilpatrick DL, De Meyer GRY. 2007. Selective clearance of macrophages in atheroslerotic plaques by autophagy. Journal of the American College of Cardiology 49, 707–715. Verlinden S. 2003. Changes in mineral nutrient concentrations in Petunia corollas during development and senescence. Hort Science 38, 71–74. Wagstaff C, Leverentz MK, Griffiths G, Thomas B, Chanasut U, Stead AD, Rogers HJ. 2002. Cysteine protease gene expression and proteolytic activity during senescence of Alstroemeria petals. Journal of Experimental Botany 53, 233–240. Wagstaff C, Malcom P, Arfan R, Leverentz MK, Griffith G, Thomas B, Stead AD, Rogers H. 2003. Programmed cell death (PCD) processes begin extremely early in Alstroemeria petal senescence. New Phytologist 160, 49–59. Waki K, Shibuya K, Yoshioka T, Hashiba T, Satoh S. 2001. Cloning of a cDNA encoding EIN3-like protein (DC-EIL1) and decrease in its mRNA level during senescence in carnation flower tissues. Journal of Experimental Botany 52, 377–379. Wang CY, Baker JE. 1979. Vase life of cut flowers with rhizobitoxine analogs, sodium benzoate and isopentenyl adenosine. HortScience 14, 59–60. Wang TW, Zhang CG, Wu W, Taylor C, Thompson JE. 2003. Pleiotropic effects of suppressing deoxyhypusine synthase expression in Arabidopsis thaliana. Plant Molecular Biology 52, 1223–1235. Watanabe N, Lam E. 2005. Two Arabidopsis metacaspases AtMCP1b and AtMCP2b are arginine/lysine-specific cysteine proteases and activate apoptosis-like cell death in yeast. Journal of Biological Chemistry 28, 14691–14699. Wiemken-Gehrig V, Wiemken A, Matile P. 1974. Mobilisation der Zellwandstoffen in der welkenden Blüte von Ipomoea tricolor (Cav.). Planta 115, 297–307. Wiemken V, Wiemken A, Matile P. 1976. Physiologie der Blüten von Ipomoea tricolor (Cav.): Untersuchungen an abgeschnittenen Blüten und Gewinning eines Phloemexsudates. Biochemie und Physiologie der Pflanzen 169, 363–376. Wilkinson JQ, Lanahan MB, Clark DG, Bleecker AB, Chang C, Meyerowitz EM, Klee HJ. 1997. A dominant mutant receptor from Arabidopsis confers ethylene insensitivity in heterologous plants. Nature Biotechnology 15, 444–447. Winkenbach F. 1970a. Zum Stoffwechsel der aufblühenden und welkenden Korolle der Prunkwinde Ipomoea purpurea. I. Beziehungen zwischen Gestaltwande l, Stofftransport, Atmung und Invertaseaktivität. Berichte der schweizerischen botanischen Gesellschaft 80, 374–390. Winkenbach F. 1970b. Zum Stoffwechsel der aufblühenden und welkenden Korolle der Prunkwinde Ipomoea purpurea. II. Funktion und de novo Synthese lysosomaler Enzyme beim Welken. Berichte der schweizerischen botanischen Gesellschaft 80, 391–406. Woltering EJ, Somhorst D, de Beer CA. 1993. Roles of ethylene production and sensitivity in senescence of carnation flower (Dianthus caryophyllus) cultivars White Sim, Chinera and Epomeo. Journal of Plant Physiology 141, 329–335. Woltering EJ, van Doorn WG. 1988. Role of ethylene in senescence of petals—morphological and taxonomical relationships. Journal of Experimental Botany 39, 1605–1616. Wu MJ, van Doorn WG, Reid MS. 1991a. Variation in the senescence of carnation (Dianthus caryophyllus L.) cultivars I. Comparison of flower lif e, respiration and ethylene biosynthesis. Scientia Horticulturae 48, 99–108. Wu MJ, Zacarias L, Reid MS. 1991b. Variation in the senescence of carnation (Dianthus caryophyllus L.) cultivars II. Comparison of sensitivity to exogenous ethylene and of ethylene binding. Scientia Horticulturae 48, 109–116. Xiong Y, Contento AL, Bassham DC. 2005. AtATG18a is required for the formation of autophagosomes during nutrient stress and senescence in Arabidopsis thaliana. The Plant Journal 42, 535–546. Xu X, Gookin T, Jiang C, Reid MS. 2007. Genes associated with opening and senescence of the ephemeral flowers of Mirabilis jalapa. Journal of Experimental Botany 58, 2193–2201. Xu X, Jiang C, Reid MS. 2008. Functional analysis of a putative ubiquitin ligase that is highly expressed during flower senescence. Journal of Experimental Botany. (in press). Xu Y, Hanson MR. 2000. Programmed cell death during pollination-induced petal senescence in Petunia. Plant Physiology 122, 1323–1333. Xu Y, Ishida H, Reisen D, Hanson MR. 2006. Upregulation of a tonoplast-localized cytochrome P450 during petal senescence in Petunia inflata. BMC Plant Biology 6, 8. doi: 10.1186/1471-2229-6-8. Xue L, Fletcher GC, Tolkovsky AM. 1999. Autophagy is activated by apoptotic signalling in sympathetic neurons: an alternative mechanism of death execution. Molecular and Cellular Neuroscience 14, 180–198. Yamada T, Ichimura K, Kanekatsu M, van Doorn WG. 2007a. Gene expression in opening and senescing petals of morning glory (Ipomoea nil) flowers. Plant Cell Reports 26, 823–835. Yamada T, Ichimura K, van Doorn WG. 2006b. DNA degradation and nuclear degeneration during programmed cell death in 480 van Doorn and Woltering petals of Antirrhinum, Argyranthemum, and Petunia. Journal of Experimental Botany 57, 3543–3552. Yamada T, Ichimura K, van Doorn WG. 2007b. Relationship between petal abscission and programmed cell death in Prunus yedoensis and Delphinium belladonna. Planta 226, 1195–1205. Yamada T, Takatsu Y, Kasumi M, Ichimura K, van Doorn WG. 2006a. Nuclear fragmentation and DNA degradation during programmed cell death in petals of morning glory (Ipomoea nil). Planta 224, 1279–1290. Yamada T, Takatsu Y, Kasumi M, Marubashi W, Ichimura K. 2004. A homolog of the defender against apoptotic death gene (DAD1) in senescing gladiolus petals is down-regulated prior to the onset of programmed cell death. Journal of Plant Physiology 161, 1281–1283. Yamada T, Takatsu Y, Manabe T, Kasumi M, Marubashi W. 2003. Suppressive effect of trehalose on apoptotic cell death leading to petal senescence in ethylene-insensitive flowers of gladiolus. Plant Science 4, 213–221. Yanagisawa S, Yoo SD, Sheen J. 2003. Differential regulation of EIN3 stability by glucose and ethylene signalling in plants. Nature 425, 521–525. Yoshimoto K, Hanaoka H, Sato S, Kato T, Tabata S, Noda T, Ohsumi Y. 2004. Processing of ATG8s, ubiquitin-like proteins, and their deconjugation by ATG4s are essential for plant autophagy. The Plant Cell 16, 2967–2983.