Author for correspondence: Carol Wagstaff Tel: +44 (0) 1784 443761 Fax: +44 (0) 1784 470756 Email: firstname.lastname@example.org
• In the Liliaceous species Alstroemeria, petal senescence is characterized by wilting and inrolling, terminating in abscission 8–10 d after flower opening.
• In many species, flower development and senescence involves programmed cell death (PCD). PCD in Alstroemeria petals was investigated by light (LM) and transmission electron microscopy (TEM) (to study nuclear degradation and cellular integrity), DNA laddering and the expression programme of the DAD-1 gene.
• TEM showed nuclear and cellular degradation commenced before the flowers were fully open and that epidermal cells remained intact whilst the mesophyll cells degenerated completely. DNA laddering increased throughout petal development. Expression of the ALSDAD-1 partial cDNA was shown to be downregulated after flower opening.
• We conclude that some PCD processes are started extremely early and proceed throughout flower opening and senescence, whereas others occur more rapidly between stages 4–6 (i.e. postanthesis). The spatial distribution of PCD across the petals is discussed. Several molecular and physiological markers of PCD are present during Alstroemeria petal senescence.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The concept of apoptosis (or cell death) originates from the Greek term describing the falling of leaves and petals during natural senescence. Senescence can be thought of as the terminal phase in the development of leaves and flowers, and thereby involves processes such as protein remobilization and protoplasmic elimination, which are characteristic of apoptosis and PCD. Modern research often uses apoptosis and PCD interchangeably, even in animal systems. However, caution needs to be exercised with the use of these terms. PCD has been defined in animal systems (Ellis et al., 1991) as a type of cell death that is a normal part of an organism's life cycle, is initiated by specific physiological signals, and requires de novo gene transcription. Apoptosis was originally used to define particular ultrastructural and biochemical characteristics of cells undergoing PCD, including chromatin condensation, membrane blebbing and DNA laddering (Kerr et al., 1972). However, the notion that all PCD manifests apoptotic features has been challenged (Schwartz et al., 1993); some animal cells appear to undergo physiologically programmed PCD without any of the apoptotic hallmarks. In plants, PCD has been recognized as part of a number of developmental processes including anther development, abscission of plant organs, megasporogenesis in angiosperms, sex determination, hypersensitive responses, tracheary element formation, destruction of the suspensor, and formation of aerenchyma (Jones & Dangl, 1996; Ranganath & Nagashree, 2000). In some of these cell types apoptotic characteristics can be detected, in others however, not all of the recognized signs of apoptosis are observed, leading some to question the similarities between senescence and PCD (Thomas et al., 2003).
The structural changes that occur during PCD can be followed by electron microscopy. In the early 1970s Matile & Winkenbach (1971) suggested that in Ipomoea petals the vacuole acts as an autophage. Invaginations of the tonoplast surround parts of the cytoplasm that are ultimately pinched off, resulting in lysosomal-like compartments within the vacuole. The membrane surrounding these vesicles decays and the cytoplasmic material then appears to be degraded by vacuolar enzymes. Such cytoplasmic degradation is concomitant with the visible collapse of the corolla but structural changes reminiscent of cell death have been described even before flower opening in Ipomoea (Phillips & Kende, 1980). In Sandersonia there is an increase in the air spaces between the mesophyll cells; this may be the result of tepal expansion without cell division or cell expansion. The mesophyll cells degrade further as the flower senesces such that the only identifiable cells other than epidermal cells are associated with vascular traces (O’Donoghue et al., 2002).
One of the easiest PCD markers to follow at a subcellular level is that of DNA laddering, which in animal cells is the result of the activation of a DNA ladder nuclease or caspase-activated DNase (CAD). However, several authors have questioned the correlation between laddering and apoptosis (Hengartner, 2000), and it has been shown that DNA fragmentation is not an absolute requirement for PCD (Vaux & Korsmeyer, 1999). Evidence for PCD in plants has been based extensively on the presence of DNA laddering, which has been observed in numerous systems. These include carpel cells of Pisum sativum (Orzáez & Granell, 1997), cowpea cells infected by fungi (Ryerson & Heath, 1996), during somatic embryogenesis of Norway spruce (Havel & Durzan, 1996), in developing anthers (Wang et al., 1999), in senescing tomato cells (Wang et al., 1996) and in senescing Petunia petals (Xu & Hanson, 2000).
Despite some similarities between animal and plant PCD at the cellular level, very few homologues of animal PCD-related genes have been identified in plants. One of the few is DAD-1 (defender against apoptotic death) that has been detected in Arabidopsis leaves (Gallois et al., 1997), citrus leaves (Moriguchi et al., 2000) and pea petals where, in all three, levels decline during senescence (Orzáez & Granell, 1997). A direct role in PCD was ascribed to this gene following the isolation of a hamster cell line in which a single amino acid change in the DAD-1 protein caused the cells to become apoptotic (Nakashima et al., 1993). In mammals DAD-1 encodes a subunit of oligosaccharyltransferase, an enzyme involved in N-linked glycosylation (Kelleher & Gilmore, 1997). Yeast and mammals with a disrupted DAD-1 gene show premature entry into PCD and express abnormal N-glycosylated proteins, although this has not yet been proven in plant systems. It has thus been suggested that DAD-1 may in fact not be a part of the PCD machinery, but rather an essential gene whose disruption triggers cells to enter PCD. Whether the plant DAD-1 protein performs the same function is not known. This gene, however, remains a useful marker as its expression is closely correlated with entry into cell death in several plant systems studied, although not in all (Danon et al., 2000).
The triggering and progression of PCD in plants can be usefully studied using petal senescence as a model system as this is a carefully programmed process; within a given species, it is possible to predict exactly when a bud will open and how rapidly the petals will senesce (Molisch, 1938). However, both the trigger and the co-ordination vary in different plant species. In one group of species the production of a burst of ethylene co-ordinates petal senescence, itself sometimes triggered by pollination (Stead, 1992). However, in another group, ethylene does not appear to co-ordinate senescence, even though it is almost always associated with the final event of petal abscission (Stead & van Doorn, 1994). Thus in these species, although a programme clearly exists, both the nature of the trigger and the co-ordination of cellular events is as yet unclear. Alstroemeria belongs to this latter group. Recent studies (Leverentz et al., 2002) indicate that unlike ethylene-insensitive species such as daylily and gladioli (Woltering & van Doorn, 1988; Peary & Prince, 1990; Rubinstein, 2000), lipoxygenase (responsible for oxidation of polyunsaturated fatty acids containing a 1,4 pentadiene moiety) does not appear to play a major role in bringing about petal senescence in Alstroemeria. Petal senescence in Alstroemeria is, however, associated with an increase in proteolytic activity and a dramatic depletion of complex lipids, suggesting active remobilization of nutrients from the floral tissues (Leverentz et al., 2002; Wagstaff et al., 2001, 2002).
In this paper cytological and molecular markers are used to chart the progression of PCD in Alstroemeria petals and thus this paper comprises the first integrated study of plant PCD using molecular, structural and physiological information. These results and observations are presented and used, together with information on this system from previous work (Leverentz et al., 2002; Wagstaff et al., 2002), to present a model for Alstroemeria senescence in which a coincidence between changes in membrane integrity and PCD markers is revealed.
Materials and Methods
Seven stages (S = stage) of Alstroemeria peruviensis var. Samora petal development and senescence were used as described in Table 1 and Wagstaff et al. (2001, 2002). Flowers were removed from the plant 2 d before flower and transported back to the laboratory dry. Individual cymes were then removed from each inflorescence and placed into dH2O. Petals from each stage were used for RNA extraction.
Table 1. Definition of floral stages during flower opening and senescence of Alstroemeria
Day relative to flower opening
Description of floral features
Outer sepals pigmented. Tips of sepals loosening
Sepals reflexed. No anthesis
Upper three anthers bent upwards and anthesed
Lower three anthers bent upwards and anthesed
Separation and reflexing of stigmatic lobes
Discolouration of petals. Translucence around margin of sepals. Reproductive organs lying on lower petal
Abscission of petals and sepals when lightly tapped
The top third of each petal and a quadrant from the top of each sepal (from several replicate flowers) was fixed in 3 : 1 ethanol : acetic acid and stored at 4°C. Fixed samples were rinsed in water then hydrolysed with 5 m HCl for 30 min. Following a wash in distilled water for 30 min on ice, samples were Feulgen stained for 1.5 h then rinsed in 45% acetic acid. Cells and nuclei were observed under a light microscope (Olympus BH-2) and images captured using a Fujitsu HC 300Z digital camera. Cell and nuclear areas were measured using SigmaScan (Jandel Scientific, San Rafael, CA, USA). Ten cells nearest the outer margins were examined and in each case cell size and their associated nuclei were measured for each developmental stage.
Tissue pieces (approximately 1 × 1 mm2) from either the edge or the middle of the petal lamina (avoiding the vascular tissue) were fixed in 3% glutaraldehyde: 4% formaldehyde in 0.1 m PIPES buffer (pH 7.6) for 2 h. Post fixation was performed in 1% aqueous osmium tetroxide for 2 h. The material was then dehydrated in a graded series of ethanol and embedded in Spurr resin (at 60°C).
Ultrathin (50–100 nm) cross sections of the petal material were obtained using a glass knife. Sections were collected on standard 200 mesh copper grids and positively stained firstly in uranyl acetate (alc) for 20 min followed by lead citrate solution for 5 min. Micrographs were taken with an H600 Hitachi TEM microscope at 75 kV.
Genomic DNA was extracted from 2 g (f. wt) of petals at each stage of development, essentially according to Doyle & Doyle (1987). EDTA was present in the extraction buffer at a final concentration of 25 mm. DNA (30 µg) was run on a 1.5% agarose gel and transferred to nylon membrane by capillary blotting. Pre-hybridization and hybridization were performed at 60°C in a solution containing 5× Denharts, 6× SSC, 0.1% SDS, 5% PEG, 0.1% tetrasodium pyrophosphate and 100 µg l−1 denatured herring sperm DNA. Random primed probes were prepared as described in Feinberg & Vogelstein (1983) using 5 µg genomic DNA digested with Sau3A.
Cloning of DAD-1 partial cDNA from Alstroemeria petals
RNA extraction and cDNA synthesis were performed as described in Wagstaff et al. (2002). Degenerate primers to DAD-1 (DAD1F: GGGTCRTTYCCHTTYAAC and DAD1R: CAYAGRACGAAAWCWGCAAA) were designed from a comparison of conserved regions of DAD-1 plant genes. Partial cDNAs were amplified from Alstroemeria cDNA using 0.625 units of Qiagen Taq polymerase, Qiagen buffer, 125 ng of cDNA from petals of flowers at stage 1, 1.5 mm MgCl2, 0.2 mm dNTPs and 1 µg of each PCR primer. Reactions were cycled in a Hybaid OMNE machine for 35 cycles of 94°C 1 min, 56°C 1 min, and 72°C 1 min. PCR products were cloned into pGEMEasy T (Promega) and sequenced using an ABI377 automated sequencer.
Specific Alstroemeria PCR primers were designed from the partial ALSDAD-1 cDNA clone (DADAF: GGGTCGTTTCCATTCAAC and DADAR: CATAGGACGAAATCTGCAA). An initial PCR reaction was conducted using the Alstroemeria specific primers on 125 ng of the cloned partial cDNA to estimate the optimal cycle number for exponential amplification. PCR conditions were as described above, except that an annealing temperature of 54°C was used. Tubes were removed from the thermocycler at 16, 18, 20, 22, 24, 26, 28 and 30 cycles and the products were analysed on a 1.5% agarose gel. A total of 28 cycles produced a barely visible band and was subsequently used for semi-quantitative RT-PCR in which 125 ng of cDNA from the seven defined stages of Alstroemeria petals was used as a template under the conditions as described earlier in this section. To test for contamination of the cDNA with genomic DNA, primers spanning a conserved intron in the β-tubulin gene were designed by comparison of available monocot β-tubulin sequences (TUBGENF: GAATGCHGAYGAGTGYATG and TUBGENR: CGGCGCRAABCCSACCAT). Using these primers with Alstroemeria genomic DNA template yields a PCR product of approximately 450 bp compared to the 231 bp from cDNA.
Quantification of expression of ALSDAD-1
Quantity One image analysis (Bio-Rad, Hemel Hempstead, UK) was used to quantify the signal from exposure of the radioactive blot to phosphorimager film. The signal was normalized to the data obtained previously of ubiquitin expression from the same batch of cDNA (Wagstaff et al., 2002).
Cell size and cell : nuclear size ratios in petals and sepals
Measurements of cell width and length (Fig. 1a) showed that the change in cell area was due almost entirely to elongation of the cells in one axis. Measurements of cell area showed a gradual increase in size from stage 1 to stage 7 in both sepals and petals (Fig. 1b) with an overall increase of 2.8-fold in petals and 2-fold in sepals. Conversely, nuclear area decreased with developmental stage with an overall decrease of 1.7-fold for petals and 1.8-fold in sepals. Thus the ratio of cell area to nuclear area showed a marked negative correlation with increasing petal age.
Transmission electron microscopy of epidermal cells from Alstroemeria petals showed increased nuclear shrinkage with increasing age of the petals (Fig. 2a–e). The reduction in nuclear size is noticeable from just before flower opening (< S1; Fig. 2a) and continues until the oldest stage (S6; Fig. 2e) by which time the nucleus appears to be the only remaining recognizable structure within the cell. The other noticeable change is associated with the cell wall. In young buds the wall is smooth (Fig. 2a) but when fully open the wall is ridged due to outgrowths of the wall (Fig. 2b), there is little further increase in the cell wall thickness, even in the oldest flowers (Fig. 2e). When the flower has just opened (S2) the epidermal cells contained a large vacuole with only a peripheral cytoplasm (Fig. 3) and the nucleus was surrounded by cytoplasm (Fig. 2b). Both the adaxial and abaxial petal surfaces had a ridged cellulose cell wall (Figs 2b and 3) that was absent in the petals of young flower buds (Fig. 2a). Even as the flower opened the integrity of the mesophyll cells at the petal and sepal margins appeared limited since, unlike the epidermal cells, the cytoplasm appeared disrupted with only the small plastids being identifiable (Fig. 3). Furthermore, much cellular material appeared to have accumulated in the intercellular spaces suggesting that the mesophyll cells had become disrupted. The cell walls of both adaxial and abaxial surfaces showed the characteristic ridges seen in many petals. Little or no content remained in the mesophyll cells 4 d after opening, and in places the cell wall appeared to be broken (Fig. 4). Although the epidermal cells remained intact their contents appeared to be reduced with only a very thin cytoplasmic layer in which numerous small lipid-containing plastids were visible.
By stage 5 (6 d after flower opening) many mesophyll cells and some epidermal cells had completely collapsed and lacked any cytoplasm (Fig. 5). This process appeared to continue with one epidermal surface collapsing whist the other remained intact (Fig. 6; Stage 6, 8 d after flower opening) but by this time there was little evidence of the previous mesophyll cell layer other than a series of collapsed cell walls.
Clear evidence of DNA laddering was obtained, even by visualization with ethidium bromide, but was clearer after the Southern blotting (Fig. 7). Laddering was present to some extent from 2 d post-flower opening (stage 3), but increased markedly during the final stages of senescence.
Isolation and characterization of a partial Alstroemeria DAD-1 cDNA and its expression during petal senescence Degenerate primers were designed by comparison to plant DAD-1 genes in the databases and used to isolate a partial DAD-1 cDNA from Alstroemeria petal tissue (ALSDAD-1). Comparison of the sequence to DAD-1 genes from other species confirms the putative identity of this cDNA (the nucleotide sequence reported here will appear in the EMBL, GenBank and DDBJ Nucleotide Sequence Databases under Accession number AJ514409). All the plant DAD-1 genes show very close homology at the amino acid level.
Semi quantitative RT-PCR was used to investigate the expression programme of ALSDAD-1 in petals. Early in petal development there was significant expression that then declined soon after flower opening (Fig. 8a). RT-PCR using β-tubulin primers spanning an intron only showed amplification of the correct sized product from the cDNA template (Fig. 8b), with no product of the size expected from the genomic DNA template. Normalization of the ALSDAD-1 gene expression relative to that of ubiquitin showed that the maximum expression of the ALSDAD-1 cDNA is more than 7-fold that of the minimum (Fig. 8c).
Two of the hallmarks of PCD in animal systems are DNA laddering and nuclear shrinkage (Kerr et al., 1972). Both of these processes are evident in Alstroemeria petals. By light microscopy nuclear area declined by nearly 50% of the maximum nuclear area seen at stage 1, however, the decline was continuous suggesting that the maximum nuclear area might be before flower opening. This was confirmed using slightly younger flowers and TEM. Both techniques revealed that the most rapid decline in nuclear area was between stages 4 and 5, i.e. concomitant with the first externally observed signs of petal deterioration. The degeneration of the nucleus occurs against a background of gradually increasing cell size, indeed cell size increased throughout the life of the petal, a situation that has been reported for other flowers (Phillips & Kende, 1980).
In Alstroemeria many indicators of PCD appear to start very early, indicating that some cell death is occurring from the earliest stages of flower development. This includes the gradual increase in the expression of a cysteine protease (Wagstaff et al., 2002) starting from the earliest tissues examined (stage 1) and the sharp decline in total LOX activity and lipid content (Leverentz et al., 2002), again starting early in floral development (stage 1 and stage 2, respectively). It also supports the structural data since the reduction in nuclear size occurs from a very early stage of flower development (before Stage 1). However, in addition to an early start of some PCD-associated processes, another feature of petal senescence in this species is that several of these processes appear to accelerate at the time at which the first visible signs of senescence are detectable. Total protease activity (Wagstaff et al., 2002), electrolyte leakage (Leverentz et al., 2002) and DNA laddering all rise sharply around stages 4–5. DAD-1 expression, used as another marker of PCD in this system, also declines 3-fold between stages 4 and 5. Although DNA laddering has not always been found in systems otherwise showing signs of PCD (Buckner et al., 2000; Herbert et al., 2001), it has been detected in many other plant PCD systems (Jones & Dangl, 1996; De Jong et al., 2000; Ranganath & Nagashree, 2000). A fall in DAD-1 expression has been associated with petal senescence in peas (Orzáez & Granell, 1997), leaf senescence in citrus (Moriguchi et al., 2000) and with silique maturation in Arabidopsis (Gallois et al., 1997). In the majority of theses cases the association of DAD-1 expression with senescence has been correlative, rather than causative, although a plant DAD-1 gene was shown to rescue a hamster cell line lacking the mammalian equivalent from entering into cell death (Gallois et al., 1997). In Alstroemeria petals, expression was very similar to that in pea petals, being high in young petals and declining in older petals. Put together, these data suggest to us that some key events in Alstroemeria petal senescence are occurring at around stage 4–5.
An important aspect of petal senescence is the spatial distribution of cell death across the petals. Inrolling and wilting in Alstroemeria clearly starts in the petal margins, and further studies will need to address whether cell death starts in the margins of the petals, and moves inwards as senescence progresses. Ultrastructural evidence to date suggests that this is likely to be the case with the cells surrounding the vascular tissue being the very last to degrade (data not shown). The very marked loss of contents and collapse of the mesophyll layers of Alstroemeria petals occurs against a background of marked epidermal cell expansion. Expansion occurs in just one plane when the petals are viewed from the surface, although evidence from the ultramicrographs suggests that there may be some loss of depth in the epidermal cells during the latter stages of senescence. Previous studies of this system (Wagstaff et al., 2001) have shown that f. wt declines from S4; the collapse of the mesophyll cells at this stage may contribute to the f. wt loss. Thus some cells within the petal remain fully functional and gene transcription and translation are occurring until the latest stages of petal senescence. Although we do not know from which cells transcription and translation occurs, it shows that at least some cells are active (Wagstaff et al., 2002), even when other cells are at an advanced state of senescence. This is in accordance with studies of the ultrastructure and water relations of senescing Iris petals (Celikel & van Doorn, 1995; Bailly et al., 2001) and collapsing cells of senescing Sandersonia petals (O’Donoghue et al., 2002) and serves to illustrate the importance of examining tissues on a cell by cell basis, as well as looking at the more global picture. The increases seen in DNA laddering and the decline in ALSDAD-1 gene expression are a reflection of the average state of the whole petal, whereas individual cells, possibly even neighbouring cells, may be in different physiological states.
Bringing together the biochemical, ultrastructural and molecular data for this system (from the data presented in this paper and previous publications) suggests a pattern of sharp acceleration of cell death against a background of increasing nutrient degradation (Fig. 9). Some processes appear to occur gradually, e.g. nuclear condensation and lipid peroxidation, whereas other events, e.g. upregulation of protease activity, loss of membrane integrity as measured by increased conductivity, and DNA laddering are temporally more precise in their occurrence and coincide with a specific stage in the senescence programme when the first visible signs of petal deterioration become apparent. It is as yet unknown in species such as Alstroemeria, where ethylene does not seem to act as a co-ordinator, what the trigger for petal senescence might be, or how its progress might be regulated. We propose a model for Alstroemeria petal senescence in which the trigger is in fact a threshold effect of one or more of the gradual biochemical processes, e.g. nuclear shrinkage or lipid breakdown. This then initiates the rapid onset of more catastrophic processes, e.g. loss of membrane integrity, DNA laddering, mesophyll collapse resulting in PCD.
This study has illustrated how early some PCD processes start during senescence. Structural degradation occurs to both cell membranes and nuclei before the flower is fully open. The implication of this is that biochemical pathways must be actively degrading these cell components in order to bring about these structural changes, and this has indeed been found for many of the processes discussed in this paper. Bringing together this large body of data in one system allows us to reveal a novel pattern of floral senescence. Thus PCD during petal senescence is not simply a gradual running down of resources, and therefore ‘death by starvation’ for the cell. A number of catastrophic events occur (DNA laddering, electrolyte leakage, increased protease activity) later in senescence that may be triggered by a threshold effect arising from the more gradual biochemical processes. Further studies using microarrays are in progress to sample gene expression across the transcriptome. This broader approach will help to identify some of the early changes in gene expression associated with senescence in this system and therefore point to the biochemical pathways that are of the earliest significance in senescence of Alstroemeria petals.
The authors would like to thank Gareth Lewis for sequencing and Lyndon Tuck for plant maintenance. The data for epidermal cell and nuclear area was obtained in the laboratory of Dr Dennis Francis (Cardiff) and we are grateful to him and Mike O’Reilly for advising Arfhan Rafiq. Electron microscopy was performed in the EM Unit at Royal Holloway and we are indebted to the staff there for their help. The authors also wish to thank Oak Tree Nurseries, Egham for provision of the bulk of the floral material. The work was funded by (MAFF) DEFRA.