Cell death processes that remove unwanted, infected, or damaged cells have evolved in eukaryotic organisms [1]-[3]. Traditionally, active cell death regimes have been denoted as programmed cell death (PCD), while cell death that occurs more passively has been called necrosis; however, recent studies call into question the validity of this dichotomy as there is evidence which suggests that necrosis is an active process as well [4]. PCD can be either developmentally regulated or environmentally induced [5], although significant overlap exists in the mechanisms [6]. Cell death traits differ among taxonomic groups and even histological origins within a species [3],[7]. It is for this reason that efforts have been employed to create cell death classification systems, which have been primarily based on cellular morphology, and more recently, biochemical and molecular data.

In animals, there are three distinct cell death morphology types: apoptosis, autophagic cell death, and necrosis [8]. First coined by Kerr et al. [9], apoptotic morphology is characterized by a reduction of cellular volume, chromatin condensation, nuclear fragmentation, conservation of organelle ultrastructure, retention of plasma membrane (PM) integrity until an advanced stage of the death process, and subsequent formation of apoptotic bodies (Reviewed by Kroemer et al. [8]). Cells that undergo apoptosis do not cause an inflammatory reaction and are engulfed by phagocytes. Autophagic cell death is characterized by a substantial increase in autophagy prior to death [8]. Autophagic cell death in animals typically consists of an increase in autophagosomes (double membrane vesicles), which later fuse with lysosomes, and in contrast to apoptosis, there is no chromatin condensation [10]. In animals there exist three types of autophagy: microautophagy, macroautophagy, and chaperone-mediated autophagy, while in plants there are additional forms of autophagy including but not necessarily limited to mega-autophagy, involving collapse of the tonoplast, and internal degradation of chloroplasts [11]. Although autophagy, or “self-eating”, is typically a pro-survival or reparatory mechanism, it has, in certain instances, been seen to promote cell death such as in Drosophila salivary glands during metamorphosis (as reviewed by Green [12]). Necrosis is typically associated with cell death induced by intense stressors, and has traditionally been seen as a more passive process. Necrotic morphology has been characterized by an increase in cellular volume, organelle swelling, early PM rupture, and subsequent spilling of intracellular components [10].

Currently, there is a marked lack of consensus over the classification of different plant PCD types. In the year 2000, Fukuda placed plant PCD into three categories based on cytological features including: apoptotic-like cell death, leaf senescence, and PCD where the vacuole plays a central role [7]. According to Fukuda, the morphological hallmark for apoptotic-like cell death is a retraction of the PM from the cell wall and cytoplasmic condensation [7]. Van Doorn and Woltering in 2005 stated that no plant examples conformed to the characteristics of true apoptosis [13]. They suggest that several PCD examples appeared to be autophagic, while many other PCD types fit into neither category [13]. Reape and McCabe in 2008, and furthermore in 2013, built on the apoptotic-like cell death classification [14],[15]. They discuss that despite true apoptosis not being present in plants, a number of similarities exist, specifically concerning PM retraction, which could be evolutionarily conserved [15]. Van Doorn et al., (2011) suggest there are two forms of plant PCD: vacuolar cell death and necrotic cell death, and that any use of the term apoptosis, or any derivative thereof when discussing plant PCD is a misapplication [16]. According to these authors, vacuolar cell death consists of degradation of the cell by both autophagy-like processes and the release of hydrolases immediately after tonoplast rupture [16]. Additionally, necrotic cell death is assumed to be a type of plant PCD due to the recent reports of internal signalling pathways during necrosis in animal models [16]. Alternatively, van Doorn (2011) later argued that since the vacuole is involved in almost all plant PCD types (including those not falling under the definition of vacuolar cell death), that plant PCD categories should be based on the rupture of the tonoplast in relation to cytoplasmic clearing [17]. Therefore, van Doorn [17] proposed two new categories: autolytic PCD, where rapid cytoplasmic clearing occurs post tonoplast collapse, and non-autolytic PCD, where despite the rupture of the tonoplast, no rapid cytoplasmic clearing occurs. Despite almost 15 years of attempts, well defined, workable definitions for plant PCD types based on morphology are still being developed.

Aponogeton madagascariensis, also known as lace plant, is a freshwater monocot endemic to the streams of Madagascar. Lace plant leaves possess a perforated lamina and are anchored to the corm by petioles with a sheathing base [18],[19]. This unique perforated leaf morphology has lent to its cultivation by aquarium enthusiasts for over 100 years [18]. Lace plant leaf perforations form via developmentally regulated PCD [20]. The lace plant is an excellent model organism for studying developmentally regulated PCD because of cell death predictability in window stage leaves (Figure 1A) between longitudinal and transverse veins (Figure 1B). Additionally, the thin, almost transparent leaves facilitate observation via live cell microscopy [5],[20]-[22]. Within an areole, a gradient of PCD can be seen during the window stage of development, which consists of three stages: Non (NPCD; Figure 1C), early (EPCD; Figure 1D) and late PCD (LPCD; Figure 1E) [24].

PCD initiates in the centre of the areole, which is nearly void of pigment as LPCD cells have lost their anthocyanin and the majority of their chlorophyll content (Figure 1E). In LPCD cells, the vacuole swells, displacing the nucleus and cytoplasmic components towards the PM. Subsequent rupture of the tonoplast and release of hydrolytic enzymes occurs, also known as mega-autophagy. Those cells that have lost their anthocyanin, yet still have chlorophyll, are EPCD cells (Figure 1D). Interestingly, the 4-5 cell layers adjacent to the veins do not undergo PCD during leaf perforation are in the NPCD stage (Figure 1C; [23]). Unlike cells which undergo PCD, NPCD cells retain their anthocyanin throughout perforation formation (Figure 1C). After the loss of anthocyanin, changes in the chloroplasts occur, resulting in a reduction of chlorophyll. Concurrently, there is an increase in transvacuolar strands (TVS), vesicles, vacuolar aggregates, and perinuclear accumulation of both chloroplasts and mitochondria [22].

The objective of our study is to elucidate cell death morphology of different degrees and types of stressors to contribute to the formation of a plant cell death classification system. Cell death due to extreme stressors typically causes a necrotic morphology. Conversely, cell death due to mild stressors occurs more gradually and usually displays morphology typical of developmental PCD. We assume this as a hypothesis for the present work. This study utilized the unique lace plant model system to compare various environmentally induced cell death morphologies in NPCD cells to the typical developmentally regulated PCD morphology described during perforation formation using live cell imaging techniques. Recognizing how cell death morphology and the vacuole and PM in particular vary in response to different modes of induction within a single system will provide a better understanding of the intracellular dynamics of cell death.

Heat shock treatments have been found to alter cell metabolism, disrupt mitochondria, and result in an increase in ROS [23],[25]. Depending on the severity, heat shock will result in cell death. In the 55°C treatment, anthocyanin disappearance was apparent immediately after treatment. It is suspected by the authors, that a spike in ROS may have played a role in anthocyanin disappearance, although anthocyanin is also a heat sensitive pigment [26]. In addition to anthocyanin disappearance, there was a marked increase in the number of vesicles within the cell. The characteristic increase in vesicles, the appearance of organelles in the vacuole, and an increased volume of the central vacuole, up until tonoplast collapse, provides evidence for both macro-, and mega-autophagy. The retraction of the PM from the cell wall after tonoplast collapse resembles the PM retraction observed during lace plant developmental PCD, however the ultrastructure of the PM at this point was not investigated. Although, it should be noted that the cell corpse in developmental PCD exhibits a more condensed morphology in comparison. Similarly, PM shrinkage has been shown in heat shock experiments using lace plant protoplasts [23]. Surprisingly, the protoplasts were less susceptible to heat shock at 55°C than the in situ cells in our experiment, with the protoplasts undergoing PCD after 20 min while the in situ cells underwent PCD after 10 min. The severity of the 65°C treatment resulted in cell death before completion of the treatment. The 65°C cell death morphology appeared remarkably different compared to the 55°C treatment, lacking PM retraction and with a loss of chlorophyll from the chloroplasts. The textured appearance along the periphery of the cell is believed to be the remains of cellular debris. Membranes within the cell are not believed to have retained their integrity. The subsequent morphology of the 65°C treatment is characteristic of what is commonly considered necrotic cell death [15],[16].

In the 100 mM NaCl treatment, there was a dramatic slowing in cytoplasmic streaming. Sodium chloride stress has been implicated in an increase in cytoplasmic Ca²⁺, which can arrest cytoplasmic streaming, by Na⁺ displacing Ca²⁺ from the PM, and from liberating Ca²⁺ from internal stores [27]. However, there is little research that assesses the effects of salinity on cytoplasmic streaming [28]. In the 100 mM and 400 mM NaCl treatments, the chloroplasts took on a wrinkled appearance. This wrinkled effect on chloroplast ultrastructure has similarly been observed in TEM images of tomato cells grown in a medium containing 100 mM NaCl [29]. Chloroplasts appeared swollen in the 400 mM NaCl treatment, but this effect was not observed in the 2 M NaCl experiments. In potato cultivars, electron microscopy showed that although the structural integrity of cells appeared intact, the chloroplasts appeared swollen when plants were irrigated with 100 mM, and 200 mM NaCl solution, respectively [30]. Swollen chloroplasts have also been seen in wheat and sweet potato leaves under salt stress [30],[31].

Interestingly, the vacuole appeared to increase in size in the 100 mM NaCl treatment, which occurs similarly in LPCD cells during leaf perforation developmental PCD. This rapid increase in vacuole size, in response to saline conditions, has been demonstrated in suspension-cultured of mangrove cells and barley root meristematic cells [32]. Na⁺ accumulation in the central vacuole and subsequent increase in vacuolar volumes has been shown to be an active process and is believed to be one strategy employed by the cell in response to salt stress [32]. Vesicle formation occurred in the 400 mM and 2 M NaCl treatments, suggesting an increase in macro-autophagy, perhaps to recycle damaged intracellular components [33]. High salt solutions have been shown to elicit autophagy in Arabidopsis thaliana by up-regulating autophagy related genes [33].The initial retraction of the PM from the cell wall in the 400 mM and 2 M NaCl treatments, in contrast to the late-stage PM retraction seen in developmental PCD, is plasmolysis and is due to changes in osmotic pressure. The filamentous structures between the PM and the cell wall observed during plasmolysis are speculated to be hechtian strands ([34]; Additional files 6, and 7). Interestingly, there were many more of these strands in the 400 mM NaCal compared to the 2 M NaCl treatment. At the final stages of death there was a contrast between these two treatments; cell treated with 400 mM NaCl exhibited tonoplast rupture and no PM retraction, whereas the 2 M NaCl treatment had a significant PM retraction. The authors speculate that the numerous strands connecting to the cell wall in the 400 mM NaCl treatment group played a role in negating the PM retraction which occurred after treatment at the higher concentration.

The most striking characteristic of the 12 M HCl treatment is the surprising dramatic retraction of the PM from the cell wall. The PM retraction in the 12 M HCl treatment resembled PM retraction seen at the end of developmentally regulated PCD in lace plant perforation formation. Although this retraction appears to be morphologically similar, this cell death lies in contrast to PCD in perforation formation, which typically takes several days. The cell death process in the 12 M HCl treatment was rapid and although the ultrastructural changes to the PM are unknown, the authors suspect this is a passive process. In the 3 mM HCl treatment, cytoplasmic streaming slowed, which may have been a result of either a change in cytoplasmic pH or an increase in cytosolic Ca²⁺. The vacuole swelled extensively, which may have been a cellular response to extracellular acidity by increasing the volume of the vacuole, as it is generally more acidic than the cytoplasm under normal conditions. The swelling of the vacuole in this treatment resembled swelling in LPCD cells during developmental PCD. The observation of vesicles was similar to the salt stress treatments, and may be indicative of an increase in macro-autophagy. Cell death occurred with the permeabilization of the tonoplast. In the 30 mM NaOH treatment, there was a considerable increase in vesicles compared to other treatments, which may also indicate an increase in macro-autophagy. The change in colour of the vacuole from pink to blue/green immediately before cell death suggests that the vacuolar pH was dramatically raised to near alkalinity as anthocyanin’s visible colour shifts.

Comparison between induced cell death and its developmental counterpart revealed that there are several common characteristics, including cessation of cytoplasmic streaming and tonoplast collapse (Table 1). Vacuolar dynamics appear to be consistent among the developmental and induced cell death videos (Table 1), and occupying the majority of a plant cell, it is likely to make a substantial contribution to cell death processes. Perinuclear chloroplast formations only occurred during developmental cell death. Likewise, the 12 M HCl treatment was the only cell death without anthocyanin disappearance, which is likely due to the response of the pigment to the low pH of the solution. In the NaOH treatments, nuclear condensation was not observed, in contrast to all other cell death types. Vesicle formation was a common characteristic of all cell death types except those that resulted in very rapid cell death, such as the 12 M HCL and 1 M NaOH treatments. PM retraction was seen in perforation formation, the 55°C, 2 M NaCl and 12 M HCl treatment groups. Interestingly, in all cases where cell death was observed, the vacuole played a central role, specifically seen with tonoplast collapse occurring in all cell death types. The results of this comparative study are summarized in Table 1.

In animal cells, there exists a morphologic classification system of cell death types, with three categories: apoptosis, autophagic cell death, and necrosis. Among the most apopotic-like characteristic seen in the lace plant is retraction of the PM due to a reduction in cell volume observed during leaf perforation developmental PCD. A similar morphology can be seen in the 55°C treatment, 2 M NaCl and the 12 M HCl treatment. Regarding autophagic cell death, an increase in vacuolar swelling and vesicle formation was observed in heat, salt, and most pH treatments. Notably, an increase in vesicles was not observed in the most extreme pH treatments (12 M HCl, 1 M NaOH). In most cell death morphologies induced by less severe stressors, such as in 30 mM NaOH treatment in leaf sheath tissue, whole organelles encapsulated by a vesicle were seen to fuse with the vacuole prior to tonoplast collapse. Recently, a dual role of autophagy as either an initiator of PCD during the HR (hypersensitive response), or a downstream executioner during developmental PCD has been proposed by Minina et al. [36]. The current authors believe that the examples of autophagy shown here are acting downstream, perhaps through the activation of some components similar to the lace plant leaf perforation developmental pathway. Interestingly, the high number of spherical opaque bodies which formed in the 2 M NaCl treatment fused with others that were in close proximity, and either disappeared or shrunk before PM retraction and cell death, but more research is needed to determine their composition and function.

Necrotic features such as early rupture of the PM, were typically seen in the most extreme treatments. A reduction of cellular volume, along with the active retraction of the PM is typically associated with a slower, more internally regulated form of cell death; however, PM retraction was seen in the most extreme acid treatment (12 M HCl) and took place within minutes. While the PM retractions observed in the 55°C and 2 M NaCl treatments were morphologically similar to the 12 M HCl treatment, the timeframe for cell death to occur was much longer in comparison. Considering the relatively slow timeframe for cell death from the 55°C and 2 M NaCl treatments (5.57 ± 0.21 h and 4.25 ± 0.38 h, respectively), the authors hypothesize that the PM retraction is an active process, whereas the 12 M HCl treatment represents a necrotic collapse. Further investigation is required, however, to determine whether or not the various induced cell death morphologies shown here are forms of PCD.

In 2000, Fukuda proposed the existence of three PCD categories in plants: apoptotic-like, leaf senescence, and one in which the vacuole plays a central role [7]. Since then, there have been several proposed classification systems for plant PCD, but currently none are unanimously accepted. Proposed classifications have often centered on characteristics of a particular organelle, notably the vacuole, or PM. The PM is commonly used due to its conspicuous appearance when retracted from the cell wall. This retraction, along with a reduction of cellular volume are characteristics present in apoptosis but absent in necrosis in animal models. Emphasis on the vacuole is likely due to its expansive nature in plant cells, often occupying up to 90 % of cellular volume which is unlike any autolytic organelle found in animal cells. Also, the vacuole is known for its autolytic properties, participating in cellular processes such as autophagy, which are associated to cell death events. Macro-autophagy occurs when cellular components are sequestered to the vacuole via double membrane vesicles. Vacuolar swelling followed by tonoplast collapse, known as mega-autophagy, is common in plant cell death processes, as was observed in this study. It is very possible that future cell death classifications will center on the role of the vacuole in plant cell death, as has been seen in previous classification proposals. Our data indicates that individual stressors typically result in different cell death morphologies amongst differing stressor intensities, despite being within a single system. Other researchers are encouraged to consider the means by which induced PCD studies are carried out. Although the experiments in this paper used isolated stressors, it has been found that more damage occurs to plants when multiple stressors occur simultaneously [35], and therefore it may be considered that treatments should replicate combined stressors that would occur naturally.

Lace plant (A. madagascariensis) is an ideal model organism for studying both developmental PCD and induced cell death, specifically due to its thin, near-transparent leaves which are suitable for live-cell microscopy. Using live-cell imaging, similarities and dissimilarities were demarcated among the induced treatments and developmental lace plant PCD. Cell death was successfully induced with physical and chemical stressors including heat shock, NaCl, HCl and NaOH. Although there were significant differences among the treatments, the vacuole played a central role as mega-autophagy was present in all forms of cell death that were observed. This study illuminates the variability of cell death morphology using different stressors and the authors believe that morphological data, though important should, be coupled with biochemical and molecular data before it can significantly contribute toward the formation of a cell death classification system in plants.