Keywords

3.1 Introduction

The transition of plant life from water onto land required the development of systems to ensure the hydration of all the plant organs exposed to the dry atmosphere as well as the strengthening of plant stature against gravity. Xylem (Greek xylon for wood) is the tissue which ensures both water and mineral transport from the roots up to the leaves and mechanically reinforces the axis of plant organs. Xylem is composed of several cell types including (1) tracheary elements (TEs), the conducting and support cells; (2) fibres, the non-conducting mechanical support cells; and (3) an associated xylem parenchyma. To allow TEs to conduct the hydro-mineral sap, these cells are hollowed by programmed cell death (PCD) to turn them into functional conducting corpses. The central importance of PCD to establish the hydro-mineral sap conduits becomes visible throughout the evolutionary history of TEs in plants. A subset of mosses or bryophytes, which are the most primitive living land plants, developed simplified conduits which relied on cell death to form a hollow cell delimited only by its primary cell walls to transport water and minerals [1, 2]. The earliest vascular land plants or tracheophytes, such as Aglaophyton and Horneophyton of the Devonian and Silurian periods, similarly had only primitive conduits formed by cell death and delimited only by the primary cell walls and no secondary cell walls [3]. PCD thus represents the first process acquired by higher land plants to establish a functional hydro-mineral sap conduit. The present-day TEs possess additional structural properties and morphological features which enable a highly efficient sap conduction [2, 4, 5]. These properties include (1) reinforcements of the TE sides with secondary cell walls to maintain the TE cell lumen open, (2) modification of the primary cell wall in between secondary thickenings to enable lateral distribution of the sap [6, 7] and (3) formation of thinned and/or perforated ends to allow the sap to pass from one TE to the next [8–10]. Once fully formed, TEs assemble end to end and/or laterally to establish a continuous vascular network throughout the plant body.

The study of TE formation in whole plants is challenging for several reasons. Very few TEs differentiate at a given time and even fewer are at the same differentiation stage. Another difficulty is that genetic or pharmacological modifications of TE formation often lead to drastic pleiotropic effects such as reduced growth, higher susceptibility to infection and problems in sap conduction [11–13]. In order to circumvent the deleterious effects caused by any modifications of TE formation and to obtain a large quantity of TEs at the same differentiation stage, simplified in vitro systems using suspension cell cultures have been established. The most common systems include cell suspensions from freshly isolated mesophyll cells of Zinnia elegans [14] or cell lines derived from roots of Arabidopsis thaliana [15–17]. Such in vitro suspension cell cultures enable TEs to differentiate under strict hormonal control, semi-synchronously, quickly and in large quantities. TEs formed in vitro share the same morphological, biochemical and genetic characteristics as TEs in whole plants [14, 15, 18–20]. Pharmacological and genetic modifications of TEs in these in vitro simplified systems are restricted to the single-cell level without interfering with the sap conduction or the overall development and physiology of the plant [21, 22].

Ménard and Pesquet [7] recently issued a comprehensive review on the different aspects of TE differentiation and function. The present review focuses on the cellular processes associated with PCD in TEs, their coordination with the other events occurring during TE differentiation as well as the signalling processes controlling TE PCD in the simplified in vitro TE differentiation systems in both Z. elegans and A. thaliana.

3.2 Cell Biology of TE Death

Programmed cell death during plant development is generally associated with the removal of unwanted cells. During TE formation, however, PCD is the main functionalising event of the differentiation programme as it empties the cell protoplast to convert the cell lumen into the future sap conduit. The maturation of TEs includes differentiation events occurring before TE cell death (pre-mortem), such as the cellulose and xylan deposition in the secondary cell wall thickenings [23], as well as events occurring after cell death (post-mortem), such as the lignification of secondary cell walls of TEs [22] and the clearing of the TE protoplast remains [24]. Altogether the succession of these events results in hollow TE corpses with reinforced and water-proofed side walls suitable for sap conduction. PCD is therefore essential as it initiates the “physiological life” of TEs to serve as sap conducting elements. The occurrence of TE PCD needs to be tightly coordinated with the full completion of pre-mortem processes and the initiation of post-mortem events. The major breakthrough in determining the chronological sequence of TE differentiation has relied on the combined use of real-time live-cell imaging and in vitro TE differentiating cell suspension cultures.

Contrary to observations made on fixed samples or in tissues with cells at different developmental stages, real-time live-cell imaging enables direct monitoring of the chronological progression of TE differentiation at the single cell level as well as to estimate the time length of each sequential event involved. Fukuda and Komamine, who had optimised the in vitro TE differentiation system in Zinnia and explored its potencies, produced the first real-time live-cell observations of TE differentiation [25]. In this study, TE secondary cell wall deposition was estimated to occur as quickly as 6 h in some cells. A pioneering study using “intermediate” (up to several hours) real-time live-cell imaging was performed later by Groover et al. [24] with bright-field microscopy to analyse the spatio-temporal relationship between secondary cell wall formation and TE PCD in Zinnia differentiating TEs. This breakthrough study identified the vacuole rupture as the central event triggering TE cell death. It also showed that (1) TE cell death occurs only once secondary cell wall synthesis is completed; (2) TE death is triggered by an implosion of the tonoplast; (3) the residual TE protoplast is rapidly removed post-mortem to clear the lumen of the TEs, within a few hours following death; and that (4) organelles like chloroplasts can remain visible in TEs even long after the vacuole rupture.

3.2.1 From Life to Death by TE Vacuole Rupture

TE PCD is triggered by the inward collapse and breakdown of the tonoplast. This rupture releases the vacuolar content into the TE cytoplasm which then initiates the post-mortem degradation of all cellular components (Fig. 3.1 and Supplementary Movie 3.1; [24]). This rapid inward rupture of the vacuole occurring during TE cell death is illustrated using short-term real-time imaging in Arabidopsis TE differentiating cell cultures in Supplementary Movie 3.1. Within 50 min, TE cells sequentially progress from an intact cell structure animated with cytoplasmic streaming (Fig. 3.1a, b) to a fragmented vacuole with an inward collapsed tonoplast in a slightly plasmolysed cell (Fig. 3.1c) and finally to a shrinking vestigial protoplast with no distinct inner organisation (Fig. 3.1d–f). The change in TE tonoplast integrity results from evenly distributed ruptures all around the vacuole (as indicated by the dotted lines in Fig. 3.1g, h). Differences in the vacuole uploading of fluorescein in dying TEs in Zinnia show that a change in membrane permeability precedes the vacuole implosion [26]. This tonoplast permeability change was estimated to last around 3 min using live-cell imaging in differentiating Zinnia cells [24] and could be mimicked using probenecid which is an inhibitor of organic anion transport [26]. Moreover, pharmacological inhibition of protein translation using cycloheximide in differentiating Zinnia TE cultures inhibited both TE cell death and tonoplast permeability changes, suggesting that the timing of vacuole rupture is actively controlled by the production of specific proteins [26]. However, no gene/protein candidate(s) has yet been identified or demonstrated to be responsible of TE tonoplast permeability change and/or rupture.

Fig. 3.1
figure 1figure 1

The vacuole rupture is the terminal event of TE programmed cell death observed using real-time live-cell imaging in differentiating in vitro cells of Arabidopsis. (a–f) Photo montage of time points before, during and after the TE vacuole rupture using real-time live-cell imaging (full movie provided in Supplementary Movie 3.1), bars = 4 μm. Note that TE rapidly shifts from an intact structure (a–b) to tonoplast implosion (c), followed by a gradual post-mortem degradation of the protoplast (d–f). Close-up of tonoplast structure in TEs, highlighted by a dotted white line, prior to vacuole rupture (g) and after its rupture (h), bars = 4 μm. Note that the tonoplast evenly breaks all around the vacuole in (h)

Full size image

3.2.2 Pre-mortem Cytological Changes Occurring Before TE Vacuole Rupture

The main pre-mortem cytological process is the deposition of the patterned secondary cell walls composed of cellulose and hemicelluloses. Long-term (2–3 days) real-time live-cell imaging using fluorescence microscopy revealed that Arabidopsis TEs deposit cellulose and hemicellulose into the secondary cell walls during a period of 10–16 h, after which the TEs remain alive for another 2–6 h and finally commit cell death within less than 10 min [17, 27]. The PCD of TEs appears to occur in both the Zinnia and the Arabidopsis TE differentiation systems only once the cellulose and xylan deposition are completed [17, 24, 26]. One key element for the proper formation and organisation of TE secondary cell walls is microtubules which specifically restrict and delimit the sites of secondary wall deposition [28, 29]. In both Zinnia and Arabidopsis TE differentiating cell cultures, a tight cortical microtubule network builds up while the secondary cell wall deposition progresses and then disassembles once TE cell wall synthesis is completed [30]. The mechanism controlling microtubule disassembly in TEs once the secondary cell wall deposition is completed is not yet fully understood. A Ca2+ influx, which is known to destabilise microtubules in plant and animal systems, was specifically associated with the triggering of Zinnia TE cell death [31, 32]. The subcellular localisation of Ca2+ accumulation in Zinnia TEs using chlortetracycline revealed that membrane-associated Ca2+ accumulation shifted from a broad cytoplasmic distribution to vesicular structures in between secondary thickenings as TE secondary cell wall was being deposited [33]. Treatments of Zinnia cell culture with CaCl2 and calcium ionophore A23187 induced premature cell death of TEs whereas incubation with Ca2+ chelators reduced cell death [31, 32]. Although changes in microtubule organisation have been shown to trigger PCD in animal systems [34], treatments with microtubule-stabilising or microtubule-destabilising drugs did not affect TE cell death in Zinnia or Arabidopsis TE suspension cell cultures [16, 35, 36].

During the active deposition phase of TE secondary cell walls, the TE cytoplasm accumulates small vacuolar structures in the cytoplasmic spaces in between secondary thickenings (Fig. 3.2a; [24, 37]). These specific structures have been suggested to be autophagosomal/lysosomal structures [38, 39]. Autophagosomes are specific subcellular organelles which are responsible of the degradation of damaged and/or superfluous cytoplasmic components [40]. Although the 13 classical autophagy-associated genes are not upregulated during TE formation in cell culture [23], autophagy has been associated with TE differentiation based on the activation state of the small GTP-binding protein RabG3b in Arabidopsis and poplar calli [38, 39]. Constitutive activation of RabG3b stimulated both TE formation and autophagosome formation whereas its inactivation or silencing led to the opposite effect [38]. These authors also proposed that autophagy positively regulates TE PCD, but the exact role of autophagy in this process remains to be functionally demonstrated. Several other organelles are also implicated in the control of TE PCD. Before the vacuolar rupture, the mitochondria lose their defined internal structure as well as the integrity of their external membrane (Fig. 3.2b; [41]). These changes were shown to be associated with the depolarisation of the mitochondrial membranes and the release of cytochrome c in Zinnia TE cell cultures [41]. Pharmacological agents (betulinic acid, cyclosporin A) were used by Yu et al. [41] to further show that TE PCD correlates positively with changes in the mitochondrial membrane potential but not with the cytochrome c release from the mitochondria. Cytochrome c release, contrary to the animal apoptotic cell death, is therefore not causally related to TE PCD. The combined observations of the pre-mortem cytological changes occurring during TE formation are schematised in Fig. 3.3a–c.

Fig. 3.2
figure 2figure 2

Transmission electron microscopy images illustrating the cytological changes prior to vacuole rupture in Arabidopsis TE differentiating cell cultures. (a) Accumulation of small vacuolar structures in the cytoplasm between secondary cell wall thickenings similar to what was reported by Groover et al. [24] in isolated Zinnia TEs. (b) Disorganised mitochondria similar to what was reported by Yu et al. [41] in isolated Zinnia TEs. Bars = 500 nm

Full size image
Fig. 3.3
figure 3figure 3

Schematic representation of the pre- and post-mortem events during TE PCD. (a–c) Pre-mortem events include: (a) active secondary cell wall formation delimited by a scaffold of microtubules. This stage is characterised by intact mitochondria and the accumulation of small vacuolar structures in the cytoplasm between secondary thickenings. (b) Completed secondary cell wall deposition still delimited by microtubules. This stage is characterised by intact mitochondria and the accumulation of small vacuolar structures. Note that hydrolytic enzymes (MC9, XCPs, LX and BFN1) accumulate in concert with secondary cell wall formation (shown by increasing font size). (c) TE prior death with altered tonoplast permeability, rupture of mitochondria external membrane Fig. 3.3 (continued) and their release of cytochrome c (cyt. c). Progression from this stage onwards is regulated by calcium intake, tonoplast permeability (affected by probenecid treatment) and ethylene. (d–f) Post-mortem events include: (d) the inward collapse of the tonoplast which alters the cell pH. Note that BFN1 is localised in ER-derived structures which surround the nucleus. (e) Complete degradation of the nucleus and disorganisation of the ER in the retracting protoplast. Note that secondary cell wall thickenings are starting to lignify. (f) TE with residual vestigial protoplast and lignified secondary cell wall ready for sap conduction

Full size image

3.2.3 Post-mortem Cytological Changes Occurring After TE Vacuole Rupture

Live-cell imaging in combination with fluorescence microscopy has been essential in defining the post-mortem events occurring after TE vacuole rupture. Short-term (up to 30 min) real-time live-cell imaging combined with fluorescent staining of the nuclei in Zinnia TEs revealed that the nucleus is the first organelle to disappear within 10 min after the tonoplast rupture [42]. The 10 min timing of nuclear degradation after the vacuole burst was confirmed using stable transgenic cell lines constitutively expressing translational fusion of histone2A and green fluorescent protein (GFP) with real-time live-cell imaging in Arabidopsis differentiating TEs (Fig. 3.4a–d and Supplementary Movie 3.2; [17, 26]). Supplementary Movie 3.2 illustrates this rapid nuclear degradation occurring during TE cell death using short-term real-time imaging in Arabidopsis TE differentiating cell cultures [17]. Within 15 min, TE development progresses from an intact cell structure with a nucleus (Fig. 3.4a) to a protoplast with no nucleus (Fig. 3.4b) which gradually retracts and degrades itself (Fig. 3.4c, d). This nuclear DNA fragmentation was further investigated using the terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) assay in whole plant and in differentiating cell cultures. Multiple studies confirmed that the fragmentation of TE nuclei occurs once the SCW has been deposited [22, 24, 43]. Figure 3.4f, g and Supplementary Movie 3.3 illustrate nuclei labelled in Zinnia TEs using the TUNEL assay. The degradation profiles of TE genomic DNA do not however exhibit a clear fragmentation through internucleosomal cleavage as observed in other types of PCD but rather a random degradation pattern [24, 44]. Chromatin condensation, which is another indicator of PCD, does not seem to occur regularly either [24], but it is clear that the nucleus can have a lobed structure before the vacuolar bursting [45]. In TEs, nuclear degradation is a consequence of PCD rather than the triggering mechanism like in many other plant and animal systems [46]. The active removal of the TE residual protoplast occurs post-mortem in the hours following the vacuole rupture (Figs. 3.1c–f and 3.2b–d; [17, 24, 41, 47]). The other cytoplasmic organelles are not degraded as quickly as nuclei, for instance, the endoplasmic reticulum (ER) is maintained intact up to 40 min after TE vacuole rupture [48]. The degradation of the other organelles include the disorganisation and rupture of mitochondria [41], the swelling of the Golgi apparatus [24, 37] and the breakdown of the ER into inflated vesicular structures [37, 49]. A schematic representation of the sequential cytological post-mortem changes occurring during TE formation is represented in Fig. 3.3d–f.

Fig. 3.4
figure 4figure 4

Post-mortem degradation of nuclei in Arabidopsis and Zinnia TE differentiating cells. (a–d) Real-time live-cell imaging of nucleus degradation, labelled by histone2a-GFP [17], in differentiating in vitro TEs of Arabidopsis; bars = 6 μm. Note that TE rapidly shifts from an intact structure with a nucleus (a) to a protoplast without nucleus (b), followed by a gradual post-mortem degradation of the protoplast (c–d). (e–h) TUNEL assay performed on Zinnia TEs observed using confocal microscopy with TE secondary cell wall stained by calcofluor (e), FITC-labelled genomic DNA breaks (f), chlorophyll autofluorescence (g) as well as merged images of E to G (H); bars = 7 μm. Arrow indicates TUNEL-positive nucleus in Zinnia TE

Full size image

3.3 Genes/Enzymes Associated with TE Post-mortem Protoplast Degradation

The optimal vascular function of TEs requires a complete removal of the protoplast once the cell has died. This process represents one of the post-mortem events occurring during TE formation to ensure that no remnants of the protoplast interfere with or occlude the sap flow and/or the lateral distribution. The complete clearing of TEs lasts from less than 10 min to more than 1 h depending on the in vitro TE experimental system used [17, 26, 42, 24]. The clearing process is achieved by a set of hydrolases including among others the A. thaliana xylem cysteine proteases 1 and 2 (XCP1 and XCP2), bifunctional nuclease 1/endonuclease 1 (BFN1/ENDO1) and metacaspase 9 (MC9). These genes/proteins are misleadingly referred to as “TE PCD markers” even though they affect the speed of TE clearing rather than TE cell death per se [47, 50–52]: they should therefore be called “TE autolysis markers”. Interestingly, all these genes are co-regulated with secondary cell wall synthesis genes such as cellulose synthases in both Zinnia and Arabidopsis (Fig. 3.5). They are also transcriptionally controlled by the master transcription factors vascular-related NAC-domain 6 and 7 which trigger the expression of the secondary cell wall synthesis genes as well (VND6/At5g62380 [53] and VND7/At1g71930 [54]).

Fig. 3.5
figure 5figure 5

Microarray expression profiles of genes associated with TE post-mortem autolysis compared to secondary cellulose synthase genes. In silico gene expression is presented from microarray experiments performed on TE differentiating cells of Arabidopsis thaliana (a, c), from Kubo et al. [15], and Zinnia elegans (b, d), from Demura et al. [18]. Secondary cellulose synthases are presented in (a, b) and autolytic genes in (c, d). Homologous autolytic genes include RNS3, BFN1, XCP1 and XCP2 in Arabidopsis and RBN1, ZEN1 and ZCP4 in Zinnia

Full size image

3.3.1 TE Nuclear Degradation and BFN1/ENDO1

The main enzyme responsible for the degradation of TE nucleus is BFN1/ENDO1 in Arabidopsis or its homolog with 54 % similarity, the endonuclease 1 (ZEN1) in Zinnia [55]. The expression of these nuclease genes is co-regulated with secondary cell wall cellulose synthase genes in TE differentiating cell suspension cultures of Zinnia [18, 20] and Arabidopsis [15, 17] (Fig. 3.5). BFN1/ENDO1 (AT1G11190) and ZEN1 (AB003131) are S1-like nucleases capable of degrading RNA as well as single- and double-stranded DNA [56]. The optimal nucleolytic activity of BFN1/ENDO1 and ZEN1 is dependent on the presence of divalent cations such as Ca2+, Mn2+, Mg2+ and Zn2+ [56]. Although believed to fulfil the same role in both species, BFN1/ENDO1 and ZEN1 exhibit distinct biochemical features: BFN1/ENDO1 is activated in neutral pH ~7 by Ca2+ and Mn2+ but inhibited by Zn2+ [56], whereas ZEN1 is activated in acidic pH ~5 by Zn2+ and is insensitive to Ca2+ and Mg2+ (Fig. 3.6b; [57, 58]). BFN1/ENDO1 is also expressed in cells undergoing other types of cell death processes such as leaf senescence, fruit abscission and the autolytic activity in the lateral root cap cells [59–61]. Both BFN1/ENDO1 and ZEN1 present a peptide signal and no C-terminal ER-retention sequence suggesting that these nucleases are located beyond the ER in the endomembrane secretory system. BFN1 is more precisely localised in specific filamentous compartments derived from the ER (Fig. 3.3a–d; [62]). These specific compartments have a neutral pH and do not colocalise with acidic vesicles labelled with LysoTracker (Fig. 3.3a–d; [62]). As PCD progresses, BFN1 is relocalised to ER filaments surrounding the nucleus where it is believed to contribute to the degradation of the nuclear DNA (Fig. 3.3a–d; [62]). The functional analysis of ZEN1 in differentiating TE cell cultures of Zinnia confirmed the role of ZEN1 in controlling the rate of post-mortem nuclear degradation but is not implicated in the vacuole rupture itself [58].

Fig. 3.6
figure 6figure 6

Enzymatic activity gel characterising the autolytic apparatus (nucleases and proteases) present in differentiating Zinnia TEs prior to vacuole rupture. TE autolytic hydrolases accumulate during the secondary cell wall deposition in Zinnia cell culture which appear between 60 and 72 h of culture [25]. Autolytic activities of total soluble proteins at these time points (presented by Coomassie staining) in (a) reveal that both nuclease activities (cf. [58]) in (b) and protease activities (cf. [71]) in (c) are present. Arrow heads indicate different nucleolytic and proteolytic activities. Note that treatment with STS does not affect the autolytic enzymes present in TEs

Full size image

3.3.2 RNA Degradation by Ribonucleases

The ribonucleases—ribonuclease 1 in Zinnia (Rbn1, AAC49325) and its 46 % homologous ribonuclease 3 in Arabidopsis (RNS3, AT1G26820)—have been associated with TE post-mortem autolysis on the basis of gene expression profiles and enzymatic activity during TE formation [15, 18, 63]. The expression of these ribonuclease genes is also co-regulated with secondary wall cellulose synthase genes in cell suspension culture of Zinnia [20, 64] and Arabidopsis [15, 17] (Fig. 3.5). These TE-specific ribonucleases present a peptide signal and no C-terminal ER-retention sequence suggesting that these enzymes are localised in the endomembrane secretory system (Fig. 3.3). Both Rbn1 and RNS3 are T2/S-like endoribonucleases capable of fully hydrolyzing RNAs and are associated with several types of cell death and senescence processes based on their expression patterns [65, 66]. RNS3 presents a very narrow pH optimum around 6 [67]. The tomato homolog of RNS3, the ribonuclease LX, was shown to accumulate in the ER of maturing TEs [66] and to participate in leaf senescence, but no apparent function has been yet defined in xylem differentiation [68].

3.3.3 Proteolysis of the TE Cytoplasmic Content

One of the first evidence of the post-mortem degradation of TE protoplasts is the reduction of cytoplasmic density observed by electron microscopy, suggesting that the cytoplasmic content is rapidly being hydrolysed [24, 47]. Both intracellular and extracellular activities of cysteine and serine proteases have been detected during TE formation in Zinnia cell cultures [53, 69–71] and in Arabidopsis xylem [72]. The cysteine proteases associated with TE autolysis have an optimal acid pH ~4.5 [70–73], which is obtained by the release of the acidic contents of the vacuole into the cytoplasm once the tonoplast ruptures [48]. These proteases include cysteine proteases such as XCP1 (AT4G35350) and XCP2 (AT1G20850) in Arabidopsis [72] or ZcP4 (AB091070) and p48h-17 (U19267) in Zinnia [18, 64, 70], putative serine proteases such as Xylem Serine Protease 1 (XSP1–AT4G00230) in Arabidopsis [72] or SP (AB091070) in Zinnia [64, 70, 71] and a phytepsin aspartate protease in Hordeum vulgare (X56136) [74]. XCP1 and XCP2 are specifically expressed in TEs [51] and co-regulated with secondary cell wall synthesis genes in both Arabidopsis [15, 23, 75] and Zinnia TE cell cultures [18, 20] (Fig. 3.5). XCP1 and XCP2 both present an N-terminal peptide signal targeting them to the secretory pathway. Constitutive 35S-driven over-expression in undifferentiated cells results in vacuolar accumulation of XCP1 [51], while immunolocalisation of XCP1 and XCP2 in Arabidopsis TEs shows first a localisation of these proteins close to ER followed by an accumulation in the central vacuole as TE PCD progresses (Fig. 3.3a–e; [50]). Interestingly, a homologous cysteine protease in Trifolium repens was shown to accumulate only in the ER and not in the vacuole [76]. Both genetic over-expression and knockout analyses clearly demonstrated that the XCP proteins are not involved in triggering TE PCD but are implicated in the post-mortem clearing of TEs [50, 51]. However, it is also clear that additional proteases participate in the post-mortem TE clearing process as the single and double loss-of-function mutants xcp1 and xcp2 in Arabidopsis only show a slight delay [50]. The asparaginyl-specific cysteine protease vacuolar processing enzymes (VPEs) which are associated with other types of developmental PCDs [77] could be candidates. Indeed, α- and γ-VPEs (AT2G25940 and AT4G32940) are expressed during the active deposition of TE secondary cell walls together with cellulose synthase genes in Arabidopsis TE differentiating cell cultures and in whole plants [15, 23, 78] as well as during poplar xylem formation [79]. Treatments of Zinnia TEs with caspase-1 inhibitor Ac-YVAD-CMK—considered to target VPEs in plants—reduced TE differentiation [80]. A quadruple VPE knockout mutant of Arabidopsis did not seem to have any defects in vegetative growth [81], which makes it unlikely that VPEs play a central role in TE autolysis. Other potential protease candidates include the ER-resident cysteine endopeptidases CEP1, CEP2 and CEP3 that are highly similar to XCP1/2 and are all expressed in the plant vascular tissues [82]. It has been recently shown that CEP1 (AT5G50260) is located in the ER during tapetum development and is then relocated during the later stages of tapetum development to the vacuole [83]. These authors speculated that the vacuolar localisation of CEPs activated the enzymes, due to the acidic content of the vacuole, thus accelerating the speed of autolysis. Although single CEP mutants show no obvious xylem defects [83, 84], investigations in double and triple mutants need to be performed to clarify the intervention of CEPs during TE autolysis.

3.3.4 Proteolysis by Cytoplasmic MC9

Caspases are a family of cysteine proteases responsible of both early signalling and execution of animal cell PCD. Caspase-like activities have frequently been observed in various plant tissues and even in differentiating TE cell cultures of Zinnia [80]. Treatments with caspase inhibitor Z-Asp-CH2-DCB in Zinnia differentiating TEs supported the functional role of caspase-like activities in post-mortem degradation of TEs [80]. Plants do not have canonical caspases per se but structurally related proteases named metacaspases (MC) which have distinct enzymatic activities [85]. In Arabidopsis, these form a small family of nine members [47], but only MC9 (AT5G04200) is specifically expressed in TEs. MC9 is also co-regulated with the TE secondary cell wall cellulose synthase genes in both Arabidopsis and poplar (Fig. 3.5; [15, 75, 79]). MC9 has a pH optimum of 5.5 [85] and presents a cytoplasmic localisation [47, 85]. Functional analyses showed that, just like XCP1 and 2, MC9 is involved in the TE content clearing process [47]. It is also plausible that MC9 is involved in processing other cysteine proteases in plant PCD [47], hence operating in analogy to the caspase cascades during animal PCD.

3.4 TE Cell Death Signalling

In order to coordinate the execution of cell death with the secondary cell wall formation, environmental cues and the post-mortem processes, a tight signalling mechanism of TE PCD in both time and space is paramount. Several molecular signals have been associated with TE cell death like Ca2+ intake (Fig. 3.5c; [32]) and nitric oxide release [86]. However, in both cases the associated drug treatments used (scavenger for NO [86] and chelators, ionophores for Ca2+ [32]) did not uncouple TE differentiation from TE cell death. The role of reactive oxygen species (ROS), classically associated with PCD, is not clear either. ROS have been shown to accumulate during TE differentiation but appears to be associated with post-mortem lignification rather than cell death itself (Fig. 3.7; [22, 24, 87]). An interesting signalling mechanism was proposed by Groover and Jones [32] to coordinate PCD with the completion of TE secondary cell wall deposition through the accumulation of an extracellular serine protease. While TEs develop their secondary cell walls, this serine protease hydrolyses the extracellular matrix which, together with an influx of calcium, can trigger the vacuolar collapse causing the cell death. To prove the significance of this mechanism, further studies are needed, but it is a fascinating idea that the timing of PCD could be determined by the accumulation levels of molecule(s) reflecting the thickness of the secondary cell wall.

Fig. 3.7
figure 7figure 7

Confocal laser scanning microscopy projections of reactive oxygen species (ROS) accumulation in differentiating TEs of Arabidopsis stained by dichloro-dihydro-fluorescein diacetate (DCFH-DA; green) and counterstained with LysoTracker Red (Fuschia, acidotropic dye also colouring plant cell walls) at 4 days (a), 5 days (b), 6 days (c) and 8 days (d) of culture. ROS signal (green) is concentrated in spherical structures at early time points (4 and 5 days) and then becomes more homogenous and intense at later time points (6 and 8 days). Bars = 10 μm

Full size image

3.4.1 Ethylene : Dual Function to Induce and Kill TEs

Ethylene, a gaseous hormone, represents a molecular signal associated with many types of cell death and senescence processes in plants [88]. Ethylene is produced in a triphasic manner during TE formation in Zinnia cell cultures with characteristic peaks of accumulation during the initiation, secondary cell wall formation and after TE death—generally referred to as phase II, III and IV of Zinnia TE formation (Fig. 3.8a; [89]). Ethylene is first required to trigger TE differentiation in the Zinnia differentiation system as treatments with ethylene synthesis inhibitor aminoethoxyvinylglycine (AVG) abolish TE formation [89]. Ethylene is then required to trigger TE PCD in the Zinnia TEs as treatment with ethylene perception inhibitor silver thiosulfate (STS) prevents TE post-mortem processes (inhibition of lignification and protoplast autolysis) without interfering with the pre-mortem secondary cell wall deposition (Fig. 3.8b, c; [22]). The role of ethylene produced by living parenchyma cells after TE cell death remains unclear; but its role supposedly regulates the coordination between TEs in formation, differentiated TEs and their associated parenchyma cells [89].

Fig. 3.8
figure 8figure 8

Ethylene production and role during TE formation in Zinnia differentiating cell cultures. (a) Measurements of ethylene production in the presence of silver thiosulfate (STS) during TE formation according to Pesquet and Tuominen [89]. Note that ethylene accumulation in Zinnia TE differentiating cells is under a positive feedback control. (b) Effect of the delayed addition of STS during TE cell differentiation. (c) Effect of delayed removal by washing of STS during TE differentiation. Note that both TE differentiation efficiency and TE lignification are presented (three replicates of ~300 cells); bars represent standard deviation. TE differentiation is subdivided into four successive stages including dedifferentiation (I), initiation (II), secondary cell wall formation (III) and after TE death (IV) [89]

Full size image

The delayed application (Fig. 3.8b) and removal (Fig. 3.8c) of STS during TE differentiation in Zinnia clearly confirms the dual role of ethylene in (1) increasing the TE differentiation efficiency in phase II and (2) controlling the TE post-mortem lignification in phase IV. Moreover, the STS treatment extends the pre-mortem phase III showing a longer gene expression of both secondary cell wall synthesis genes and autolytic hydrolases [22]. The STS treatment in Zinnia TEs presents no positive TUNEL labelling (Supplementary Movie 3.3; [22]), further confirming that TE post-mortem processes have not been triggered. However, both nucleolytic and proteolytic activities associated with TE post-mortem autolysis can be detected (Fig. 3.6), thereby confirming that ethylene is implicated in the signalling of TE cell death rather than TE autolysis. Genetic confirmation of ethylene intervention during TE PCD is difficult to achieve in whole plants: gain- and loss-of-function mutants affected in ethylene production or signalling do not show visible defects in TE cell death but only in TE precursor cell division [90, 91]. However, ethylene is produced in late maturing xylem [92] which supports the role of ethylene in the PCD of xylem cells in whole plants. It is nevertheless possible that the xylem tissue in the whole plant may somehow compensate defective ethylene signalling by activating an alternative pathway—which could not be possible in single-cell in vitro cultures.

3.5 Conclusion

The cell death of TEs is an extraordinary phenomenon as it is one of the few cell death processes in plants that initiates the functional “life” of the cells by enabling TEs to transport the hydro-mineral sap. The use of in vitro TE differentiating cell suspension cultures and real-time live cell imaging has provided breakthroughs in understanding the cell biology and physiology of TE cell death. Such discoveries include the elucidation of the central role of vacuole in triggering cell death, the timing and morphological characteristic of the cellular changes as well as the molecular and hormonal control of TE PCD. Great advances have also been made in understanding the TE clearing mechanism in which hydrolytic enzymes synthesised pre-mortem are activated rapidly post-mortem. A safety system, based on specific pH requirement of the enzymes, subcellular compartmentalisation and possibly proteolytic activation, prevents premature activation of the hydrolytic programme. Integrity of the vacuole is central in controlling this safety mechanism even though the importance of other organelles has to be emphasised as well. Increasing evidence supports the localisation of crucial hydrolytic enzymes in various ER-derived compartments rather than in the vacuole but also a translocation of enzymes between these various organelles. Therefore, we need to revisit the role of the vacuole as the main storage organelle for the post-mortem hydrolytic enzymes and also to increase our understanding of the function of the other components of the endomembrane system in the regulation of TE autolysis.

Another challenge for the future is to define the identity of the initial signals triggering TE PCD. It is widely accepted that cell death is carried out by the vacuolar rupture, yet important signalling events and crucial cellular changes which precede the tonoplast implosion are likely to trigger the TE cell death. Studies on other developmental PCD in plants suggest that processes such as ethylene production, cytoskeletal modifications, pH changes, a burst of ROS and autophagy occur before vacuolar rupture and contribute to the outcome of the cell death. These processes need to be characterised in detail during TE cell death. The in vitro TE differentiation cell culture systems represent a perfect tool to unravel these lines of investigations.