•Set out here is the first generic account of the cytological effects of dehydration and rehydration and exogenous abscisic acid on moss protonemata.
•Protonemal cells were subjected to slow and fast drying regimes, with and without prior exposure to abscisic acid. The cytological changes associated with de- and rehydration were analysed by light, fluorescence and transmission electron microscopy, together with pharmacological studies.
•Protonemata survive slow but not fast drying, unless pretreated with abscisic acid. Dehydration elicits profound cytological changes, namely vacuolar fragmentation, reorganization of the endomembrane domains, changes in the thickness of the cell wall and in the morphology of plastids and mitochondria, and the controlled dismantling of the cytoskeleton; these dynamic events are prevented by fast drying. In control cells, abscisic acid elicits changes that partially mimic those associated with slow drying, including controlled disassembly of cytoskeletal elements, thus enabling protonemal cells to survive normally lethal rates of water loss.
•Our demonstration that moss protonemata are an ideal system for visualizing and manipulating the cytological events associated with vegetative desiccation tolerance in land plants now opens up the way for genomic dissection of the underlying mechanisms.
In vascular plants, desiccation tolerance (DT), the ability to dry to equilibrium with air without dying, is generally restricted to seeds, pollen and spores, and is rare in vegetative tissues (vegetative DT). However, many bryophytes (mosses, liverworts and hornworts) are able to survive desiccation during their vegetative growth period (Proctor & Pence, 2002; Proctor & Tuba, 2002). As the closest living relatives of the first land plants (Renzaglia et al., 2007), bryophytes are highly suitable models for understanding the evolution of vegetative DT in green plants and for unravelling the major biological processes involved (Proctor et al., 2007a). Mosses, in particular, are now prime candidates for studies on the biological significance of stress-related genes (Cuming et al., 2007), following the development of homologous recombination techniques in the model moss Physcomitrella patens (Reski, 1999) and the sequencing of its genome (Rensing et al., 2008).
Although numerous physiological and cytological investigations have provided important insights into the underlying mechanisms of DT in adult moss tissues (Proctor et al., 2007b), to date, only a handful of studies have focused on the desiccation biology of the juvenile stage of moss development, the protonemata; this is surprising, given that protonemata have been used extensively as model systems to unravel the cellular basis for plant morphogenesis (Cove et al., 2006 and references therein).
Werner et al. (1991) have shown that the protonemata of Funaria hygrometrica withstand desiccation and recover only if the rate of drying is slow, and that, during the drying process, endogenous levels of abscisic acid (ABA) increase six-fold. When pretreated with exogenous ABA, protonemata are able to survive a much faster rate of water loss. More recently, Oldenhof et al. (2006) have demonstrated that, after exposure to ABA, Ph. patens, a drought- but not desiccation-tolerant species (Frank et al., 2005), is able to survive almost complete desiccation as long as the rate of water loss is slow. ABA has also been shown to increase the DT of adult tissues in mosses (Beckett et al., 2000; Beckett, 2001) and liverworts (Pence et al., 2005), and the freezing tolerance of moss protonemata (Nagao et al., 2005; Oldenhof et al., 2006; Rowntree et al., 2007).
Recent investigations into the consequences of drying and rewetting in bryophyte vegetative tissues have shown that the shutting down and resumption of metabolic activities associated with de- and rehydration are accompanied by cytological changes that are equally profound (Pressel et al., 2006; Proctor et al., 2007b). The major changes associated with drying are essentially the same across a range of tissues, and in both mosses (Pressel et al., 2006; Proctor et al., 2007b) and liverworts (Pressel et al., 2009). These are, by and large, comparable with those described in vascular plants (Quartacci et al., 1997; Dalla Vecchia et al., 1998; Navari-Izzo et al., 2000), and include vacuolar fragmentation, changes in the internal structure of plastids and mitochondria, and chromatin condensation in the nuclei whilst membrane integrity is retained. A study by Pressel et al. (2006) of the cytological consequences of drying and rewetting in the food-conducting cells (FCCs) of the moss Polytrichum formosum demonstrated, for the first time, a key role of the microtubular cytoskeleton in the desiccation biology of moss vegetative cells; highly ordered de- and repolymerization of microtubules (MTs) during a drying–rewetting cycle is a prerequisite for the survival of FCCs. The same is also true for the meristematic cells of Po.formosum (Pressel et al., 2006) and those of the liverwort Southbya nigrella (Pressel et al., 2009). Changes in MT dynamics have also been described in higher plants in response to cold and freezing stress (Pressel et al., 2006 and references therein), and Wang et al. (2007) have shown recently that salt tolerance in Arabidopsis requires the reorganization of the cortical MTs.
With their extensive MT network, comprising both cortical and endoplasmic arrays, which can be visualized directly, different from that of FCCs located deep within moss stems (Pressel et al., 2006), moss protonemata seem to be ideally suited for the investigation of the role of the cytoskeleton in desiccation biology. Numerous studies, in particular those combining drug treatments with fluorescence procedures, have demonstrated key roles for both the actin microfilaments and MTs in protonemal morphogenesis (Doonan & Duckett, 1988; Wacker et al., 1988; Doonan, 1991; Finka et al., 2007). However, none have addressed the possible relationships between the cytoskeleton and desiccation biology.
Thus, the main goals of the work described here are to document the cytological changes associated with drying and rewetting cycles in the protonemata in a range of mosses with different degrees of DT, subjected to different drying regimes with or without prior exposure to exogenous ABA, and to investigate further the possible effects of de- and rehydration on the cytoskeleton by a combination of pharmacological and fluorescence studies. As far as we are aware, this is the first study to address directly the cytological effects of de- and rehydration in this system.
Materials and Methods
Plant material and culture conditions
Protonemata of Physcomitrella patens (Hedw.) Bruch & Schimp. [Aphanoregma patens (Hedw.) Lindb.] (drought tolerant) collected from Bough Beech Reservoir, Sussex (UK), Funaria hygrometrica Hedw. (moderately DT; Werner et al., 1991) from Ashdown Forest, Sussex, Tortula muralis Hedw. from Queen Mary University of London campus and Campylopus introflexus (Hedw.) Brid. from Thursley Common, Sussex (the last two species both highly DT) were cultured axenically from spores on Parker medium solidified with 1% Phytagel and overlain with cellophane discs, as described previously (Pressel et al., 2008). Protonemal colonies, after 2–3 wk growth, were used for the desiccation experiments. Alternatively, 1- and 2-wk-old colonies were transferred on their cellophane overlays to medium containing 10 μm ABA and maintained for 7 d before experimentation. Syntrichia ruralis (Hedw.) F. Weber & D. Mohr (Tortula ruralis Hedw.), the moss used extensively in physiological and cytological studies of desiccation biology (Platt et al., 1994; Tuba et al., 1996; Oliver et al., 2000; Proctor & Smirnoff, 2000), was not included in this study because, when grown in culture, it produces only a limited protonemal system comprising thick-walled main axes packed with lipid and very little chloronema, and it is difficult to process for electron microscopy (Pressel, 2007).
Drying conditions and survival following rehydration
For the slow drying protocol, cellophane discs with fully hydrated protonemal colonies were placed on a single sheet of filter paper imbibed with 250 μl of distilled water in closed but not sealed Petri dishes. Placing protonemal colonies in empty Petri dishes (Werner et al., 1991) killed the protonemata of Funaria, whereas, in C. introflexus and T. muralis, only cells in the centre of protonemal colonies survived. In fact, using their protocol, we were unable to replicate the drying rate reported by Werner et al. (1991), 50% relative water loss within 4 h, and obtained a relative water loss in excess of 80% in 4 h.
For the fast drying protocol, protonemata were placed directly in a lamina flow cabinet. Independent of the drying regime, protonemata were considered to be desiccated when they had lost at least 90% of their water content. The dry weights were determined after the oven drying of protonemata at 100°C for 24 h. At least 10 replicates were performed for each species.
Protonemata were rehydrated by soaking in distilled water until the cells appeared fully turgid, and were then transferred to control medium and assessed for survival at 1, 2, 3 and 5-d intervals. Dead cells lacked organized contents, were devoid of plastid pigmentation and were frequently collapsed and therefore easily distinguishable from living cells. Possible longer term effects of the drying treatments on the subsequent growth and development of the protonemata were monitored for 2 wk after the experiments.
One- and 2-wk-old protonemal colonies were transferred to medium containing either 10 μm taxol (an MT-stabilizing drug) or 10 μm oryzalin (an MT-depolymerizing drug) for 18 h, washed with distilled water, slowly-desiccated, and their cytology and survival monitored following rehydration (in the absence of the drugs). Alternatively, slowly-dehydrated protonemata were rehydrated in the presence of either 10 μm oryzalin or 2 μm cytochalasin D (an inhibitor of actin polymerization) for 18 h, and their survival monitored after removal of the drug and transfer to standard medium. Controls were kept or rehydrated in 1% (v/v) dimethylsulphoxide, used as a solvent for the preparation of stock solutions of the inhibitors.
Light microscopy Control protonemata, ABA-treated protonemata and protonemata at 5 min, 30 min, 2, 4 and 24 h after rehydration in water or 18 h after rehydration in inhibitors were mounted in dH2O for inspection under a Leica DM RXA2 microscope (Leica, Southwold, England) equipped with differential interference contrast optics. Desiccated protonemal cells, both slow and fast dried, were mounted in immersion oil to prevent any rehydration from taking place (Tucker et al., 1975).
Transmission electron microscopy The protocol for chemical fixation and embedding of protonemata follows that of Proctor et al. (2007b). Cryofixation of protonemal cells was carried out in a Leica EM PACT high-pressure freezer. Thereafter, the frozen samples were transferred to a Leica EM AFS automatic freeze-substitution system. Samples were then kept at −90°C in precooled freeze-substitution media consisting of either (a) 1% osmium tetroxide (w/v) in anhydrous acetone for 3 d, or (b) 0.1% tannic acid (w/v) in anhydrous acetone for 24 h, followed by 2% anhydrous glutaraldehyde (v/v), 2% uranyl acetate (w/v) and 2% osmium tetroxide (w/v) in dry acetone for 48 h. Subsequently, samples were warmed in a stepwise manner (−60°C for 18 h, −40°C for 12 h, −20°C for 12 h and 0°C for 8 h), washed with acetone at room temperature and embedded in Spurr’s resin. Both protocols gave comparable results.
Thin sections, cut with a diamond knife, were sequentially stained with 5% (v/v) methanolic uranyl acetate for 15 min and lead citrate for 10 min, and examined with a Jeol 1200 EX2 electron microscope (Jeol, Welwyn Garden City, England); 0.5-μm-thick sections, stained with 1% toluidine blue, were photographed with a Leica DM RXA2 microscope equipped with differential interference contrast optics.
Indirect immunofluorescence of MTs For indirect immunofluorescence of MTs, the protocol by Goode (1992) was followed with slight modifications. Control, slowly-dried and rapidly dried protonemata, and slowly-dried protonemata rehydrated in oryzalin, were fixed in 8% paraformaldehyde, with 5% dimethylsulphoxide in MT-stabilizing buffer (MTSB), pH 6.9, for 40 min, washed in MTSB and stuck to polylysine-coated coverslips. Wall digestion was achieved by a 30-min treatment with 5% driselase (Sigma, Sigma-Aldrich, York, England) in MTSB. Cells were extracted with 1.5% Nonidet NP-40 with 1% dimethylsulphoxide in MTSB, washed, and incubated with monoclonal rat anti-tubulin antibody YOL 1/34 (Sera Labs, Haywards Heath, West Sussex, England) diluted at 1/10 in 0.1 m phosphate-buffered saline, pH 7.2, for 40 min at room temperature. Protonemata were washed in MTSB and incubated with the secondary antibody, fluorescein isothiocyanate-conjugated anti-rat (Sigma), for 40 min at room temperature, washed again in phosphate-buffered saline and mounted in Citifluor anti-fade glycerol solution (Agar Scientific, Stansted, Essex, England). For controls, either the primary or the secondary antibodies were omitted. Preparations were viewed under a Leica DM RXA2 microscope with epifluorescence optics.
Rhodamine–phalloidin staining of F-actin Rhodamine–phalloidin staining of F-actin followed the protocol by Goode (1992). Briefly, protonemata were incubated for 20 min at room temperature in lysis buffer comprising 100 mm piperazine-N,N′-bis(2-ethanesulphonic acid), 0.1% Triton X-100 (Sigma), 1 mm MgSO4, 5 mm ethyleneglycol-bis(β-aminoethylether)-N,N′-tetraacetic acid, 3 mm dithiothreitol, 50 μg ml−1 leupeptin (Sigma), 0.3 mm phenylmethylsulphonofluoride (Sigma), 5 μg ml−1 4′,6-diamidino-2-phenylindole and 0.1%n-propylgallate (Sigma), pH 7.0, and then mounted on a slide with 50/50 lysis buffer/rhodamine–phalloidin (at 0.1 μm in phosphate-buffered saline) (Molecular Probes, Manchester, England) and viewed immediately under a Leica DM RXA2 microscope fitted with epifluorescence optics.
Protonemal survival through drying–rewetting cycles
Protonemata of all four species lost in excess of 90% of their initial water content after 2–3 d of slow drying and within 30 min to 1 h of fast drying (data not shown). No significant differences were observed in the rate of water loss between the species; similarly, pretreatment with ABA did not alter significantly the rate of water loss for any of the drying regimes, as reported previously (Werner et al., 1991; Pence et al., 2005). For the slow and fast drying experiments in the presence or absence of ABA, our results for Physcomitrella and Funaria essentially confirmed those of earlier studies (Werner et al., 1991; Oldenhof et al., 2006): the entire protonema of Physcomitrella was killed by slow drying, whereas that of Funaria survived, except for some caulonemal tip cells and the two to five cells behind these. ABA rendered Ph. patens able to withstand slow but not fast drying, whilst exposure to ABA increased the DT of Funaria, with protonemal cells surviving normally lethal rates of water loss, except for some tip cells and side branches. The protonemata of C. introflexus and T. muralis behaved essentially as those of F. hygrometrica, although damage to peripheral cells was less pronounced in these two species.
Exposure to oryzalin and taxol greatly decreased the DT of protonemata in all species, with very few cells recovering from desiccation. Slow-dried cells rehydrated in oryzalin also failed to recover, and rehydration in cytochalasin had no effect on the subsequent survival of protonemata on drug-free medium.
Because Physcomitrella survives desiccation only after exposure to ABA, which, in turn, results in the differentiation of its protonemal cells into brood cells or brachycytes (Oldenhof et al., 2006), detailed transmission electron microscopy investigations were carried out in the remaining three species. The cytological responses of brood cells, desiccation-tolerant propagules (Goode et al., 1993), to drying and rewetting have been reported and discussed elsewhere (Rowntree et al., 2007).
The results for all three species are considered together, except when those for a particular species were clearly distinct from the others. The main results are summarized in Table 1. Pressel et al. (2008) provide a generic account of protonemal ultrastructure for comparison with the present data.
Table 1. The cytological effects of de- and rehydration and of exogenous abscisic acid on moss protonemata
Fully hydrated controls
Rehydration (5 min)
30 min–1 h
Abscisic acid-treated controls
ER, endoplasmic reticulum; MT, microtubule.
Large central vacuole
Numerous, small with electron-dense contents
As in desiccated state
Numerous, small with electron-dense contents
Heterogeneous vesicle population, associated with MTs, longitudinally- aligned
Numerous, small vesicles, some with electron-dense content, scattered
As in desiccated state together with multivesicular bodies
Heterogeneous vesicle population – becoming associated with MTs
As in controls – vesicles associated with MTs
Numerous, small vesicles, some with electron-dense contents
Numerous convoluted sheets, often fenestrated or swollen Smooth
Scattered, fenestrated sheets Rough and smooth
Labyrinthine often swollen tubular network with electron-opaque contents
As in controls
Figs 1a,b are light micrographs of slowly-desiccated protonemata mounted in oil. The original cylindrical filaments have become completely flattened, as is evident from the profiles at right angles to each other (Fig. 1a), and the volume of the cells has been reduced by c. 90%– a finding closely in line with the water loss data. The plastid-containing cytoplasm is aggregated against the end walls and along the edges of the flattened central region of the cells (Fig. 1a,b). The plastids are either rounded or have a highly folded appearance (Fig. 1b).
Transmission electron microscopy results show that, in both chloronemal and caulonemal cells, the plastids usually lack starch grains. Their membrane systems are intact; however, the thylakoids are often appressed to form large, curved grana with few stroma lamellae (Fig. 1c,d,f). Closely associated with the thylakoids are numerous, enlarged plastoglobules, between 100 and 200 nm in diameter (vs 50 nm in control plastids) (Fig. 1c,d). The plastids range between 2.5 and 6 μm in length, as opposed to up to 20 μm in control caulonema cells (Pressel et al., 2008) (Fig. 1c,d,f). The mitochondria are similarly spherical or ovoid, and contain thin parallel-sided cristae in an electron-transparent matrix (Figs 1c,f,2b). The nuclei are rounded or ovoid with irregular contours, and contain partially condensed chromatin; their envelopes are studded with closely packed pores (Fig. 1d,e).
The cytoplasm of chloronemal cells is packed with numerous small vacuoles with conspicuous electron-dense inclusions (Fig. 1c,d). In caulonema cells [vacuoles are lost during caulonemal differentiation to be replaced by endoplasmic reticulum (ER)-derived vesicles –Pressel et al., 2008], small vesicles, some with electron-dense contents, are a common occurrence (not illustrated). Free ribosomes form dense aggregates (Fig. 1e) and the ER consists of smooth sheets; these are either convoluted and scattered amongst the other organelles (Fig. 1f), or arranged parallel to and near the longitudinal walls (Fig. 1g). Golgi bodies are rare and comprise aggregates of small vesicles (not illustrated).
In both cell types, the cell wall is thickened and stratified (Fig. 1c) and, particularly in caulonema cells, large areas of the cell wall–plasma membrane interface are highly convoluted (Fig. 2a) and associated with numerous membranous tubules, 30–50 nm in diameter (Fig. 2b,c), which extend into the cytoplasm, both parallel and perpendicular to the long axes of the cells (Fig. 2b). Cortical and endoplasmic MTs, the latter forming conspicuous longitudinal arrays in hydrated caulonemal cells (Pressel et al., 2008), are completely absent.
Transmission electron microscopy of freeze-substituted protonemal cells (Fig. 2d–h) confirms the results obtained with standard chemical fixation described previously. Plastids and mitochondria are highly shrunken, the former with tightly appressed, intact thylakoids and prominent plastoglobules, 100–200 nm in diameter (Fig. 2d,e), and the latter with thin, tubular cristae (Fig. 2g,h). Chromatin condensation is evident in the nuclei (Fig. 2f). The main differences are the highly irregular and convoluted contours of the organelles (more so than in the chemically fixed material), especially of the nuclei (Fig. 2f,g) and microbodies (Fig. 2g,h), and the fact that the plastids are almost devoid of stroma (Fig. 2d,e). The membranous tubules and ER domains described earlier are very difficult to discern as they appear electron-transparent against an electron-dense background (Fig. 2h).
Rehydration of slowly-desiccated protonemata
Five minutes after rewetting, the cells regain their original dimensions and the plastids are scattered throughout their expanded lumen (Fig. 3a–c), often clustered into small groups (Fig. 3c). Ultrastructurally, organelles are as in the dry state (Fig. 4a,d,e,g), except for changes in the organization of the endomembrane domains and a marked increase in the size of the plastids. The membranous tubules associated, in the dry state, with the cell wall–plasma membrane interface have disappeared, and now irregular multivesicular bodies (Fig. 3h) and extensive and highly convoluted sheets of smooth ER (Fig. 3c,g) are present throughout the cytoplasm, especially in caulonema cells. The cell wall–plasma membrane interface remains highly convoluted with numerous small vesicles, < 300 nm in diameter, in close proximity (Fig. 3i). The plastids are up to 8 and 15 μm in length in chloronema (Fig. 3d–f) and caulonema (Fig. 3h) cells, respectively, and have highly irregular outlines, with large portions of their peripheral stroma thylakoid-free (Fig. 3f–h). Common to both desiccated and rehydrating cells is a close association between the plastids and numerous microbodies (Fig. 3f).
From 30 min to 1 h of rehydration, the sheets of smooth ER in caulonemal cells change from randomly oriented to longitudinally-aligned (Fig. 4a), and the ribosomes become organized into polysomes, either free in the cytoplasm or attached to swollen ER profiles (Fig. 4e). It is within the first hour following rehydration that the nuclei and mitochondria regain their control morphologies; the former become oval to elongate in shape (Fig. 4d), depending on the cell type, and contain dispersed chromatin; in the latter, cristae are saccate and the matrix is electron-opaque (Fig. 4e). However, the plastids are increasingly pleomorphic (Fig. 4b–d) and have an extensive stroma, with large areas completely devoid of thylakoids. These are no longer appressed, but now form discrete grana connected by stroma lamellae (Fig. 4c,d). Microbodies remain abundant, especially in close proximity to the plastids (as in Fig. 3f). A major event 30 min after rehydration is the reappearance of MTs; numerous short MTs are visible in association with the nuclear envelope (Fig. 4e) and with the plasma membrane (Fig. 4f).
Two hours after rehydration the cellular organization of caulonema cells is essentially back to that in the controls. The cellular polarity and the extensive network of endoplasmic MTs, typical of the hydrated caulonemal cells, are fully re-established; the MTs are longitudinally-aligned and closely associated with the organelles and vesicles (Fig. 4i), as described previously (Pressel et al., 2008). In chloronemata the vacuoles, albeit fewer and larger than in the desiccated state, remain numerous, with some appearing to fuse (Fig. 4g). A complete return to the highly vacuolated state, typical of the controls, takes longer, usually 3–4 h. The recovery of plastid morphologies is also a more protracted process. After 2 h from rehydration, although most of the organelles have recovered their control morphologies, plastids exhibit numerous stroma-filled extensions and invaginations of their envelopes (Fig. 4h), and regain their control morphology only after 4 h. After 4 h, both caulonema and chloronema cells are indistinguishable from the controls (not illustrated).
Fast drying and rehydration
Fast-desiccated protonemal cells mounted in immersion oil and viewed in the light microscope are highly shrunken, like their slowly-dried counterparts. However, their plastids remain somewhat scattered throughout the cells rather than crowded against the end walls (Fig. 5a). Ultrastructurally, fast-dried cells are markedly different from slowly-dried ones. Their walls have not increased in thickness (Fig. 5b,f) and the plastids, although shrunken (up to 6 μm in length), have swollen thylakoids and numerous starch grains (Fig. 5b,d). The mitochondria are irregularly shaped and contain swollen saccate cristae (Fig. 5c,d). Short, randomly oriented MT profiles are scattered throughout the cytoplasm and are not associated with the organelles (Fig. 5c). Although chromatin condensation is evident in the nuclei, these have highly irregular contours (Fig. 5b,c). Following rehydration, evidence of widespread cellular breakdown is readily apparent after only 1–2 h. The plasma membrane and nuclear envelope are ruptured, and degenerating plastids and mitochondria, with no discernible internal structure, are scattered throughout the electron-transparent cell lumina (Fig. 5e).
Cells rehydrated in 1% dimethylsulphoxide are indistinguishable from those rehydrated in distilled water (not illustrated). In cells rehydrated in oryzalin, MTs are absent, except for occasional, randomly oriented short profiles (not illustrated). Cellular organization appears to be severely disrupted and organelle polarity and alignment in caulonemata are absent (Fig. 5f,g). The cell wall–plasma membrane interface remains highly convoluted, as in the dried state (not illustrated); the nuclei are rounded with irregular contours (Fig. 5f); the plastids are clumped together and are highly pleomorphic with long, thin protrusions of their envelope and large plastoglobules (Fig. 5h); numerous small vacuoles/vesicles are present throughout the cytoplasm, especially in caulonemal cells (Fig. 5f). Dehydrated specimens treated with oryzalin during rehydration fail to recover when transferred to oryzalin-free medium, whereas fully hydrated specimens treated with oryzalin make a full recovery when the drug is removed (Pressel et al., 2008).
The parallel experiments with cytochalasin D produce much more subtle cytological changes. Longitudinal organelle alignment and polarity are retained, although the latter is frequently reversed relative to the controls (not illustrated). The plastids are packed with starch and, in particular, those of chloronemal cells are often much larger than in the controls (Fig. 5j). The caulonemal cytoplasm contains numerous large vesicles with convoluted membrane profiles around their periphery (Fig. 5i), never observed in the controls. Unlike specimens rehydrated in oryzalin, most of the cytochalasin-treated specimens recover and resume normal growth on drug-free medium.
Exogenous ABA elicits pronounced changes in the fine structure of protonemata. Some of these changes closely mimic those elicited by slow drying: chloronemal cells are packed with small vacuoles with electron-dense deposits (Fig. 6a,b), the cell wall increases in thickness and, in caulonemata, MT arrays typical of the controls are rarely observed (Fig. 6c). In some cells, longitudinal organelle alignment is lost (Fig. 6c). Starch deposits in the plastids are less abundant than in the controls in F. hygrometrica (Fig. 6b), but appear to be unaffected in C. introflexus (Fig. 6a). ABA-treated slow-dried cells are essentially indistinguishable from their nontreated counterparts (not illustrated), but ABA treatment changes dramatically the responses of cells to the fast rate of water loss. In ABA-treated fast-dried cells, MTs are absent and the endomembrane system is reorganized into single ER profiles and numerous membranous tubules (Fig. 6e,g). Abundant small vacuoles, with electron-dense contents, are also present (Fig. 6f). The overall appearance of the plastids, mitochondria and the nuclei is essentially the same as in slowly-dried cells (Fig. 6d–f), except that some plastids are somewhat elongate and contain starch grains (Fig. 6e,f), whereas the thylakoids remain organized into discrete grana and stroma lamellae (Fig. 6e) that trace an undulating course. Cellular integrity is retained also following rewetting, and 5–30 min after rehydration membranes remain intact (Fig. 6i–h).
Indirect immunofluorescence studies with α-tubulin antibody confirm the presence of net axial arrays of MTs in control cells, as described previously (Doonan, 1991; Goode et al., 1993) (Fig. 7a–c). In controls, in which either the primary or secondary antibodies were omitted, no staining is observed (not illustrated). In ABA-treated cells, the distribution of MT arrays lacks the net longitudinal alignment typical of controls; the MTs appear to be somewhat scattered or arranged around the plastids (Fig. 7d).
Indirect immunofluorescence of slow- and fast-dried protonema cells confirms the transmission electron microscopy observations. Following slow desiccation (or pretreatment with oryzalin), MTs are absent, as evidenced by a diffuse fluorescence or the presence of small tubulin aggregates (Fig. 7e–g). In fast-dried cells, fragmented MTs, with no apparent orientation, are clearly visible (Fig. 7h–j). Cells pretreated with ABA before fast drying are indistinguishable from their slow-dried, untreated counterparts (not illustrated).
In fully hydrated cells, rhodamine–phalloidin staining of F-actin reveals an intricate network of actin filaments, as reported previously (Goode et al., 1993; Finka et al., 2007) (Fig. 7k). After slow drying, only granular or diffuse staining is observed (Fig. 7l). Similar results are obtained for cells rehydrated in cytochalasin D (not illustrated), whereas, in fast-dried cells, short, randomly distributed actin microfilament fragments are stained (Fig. 7m,n).
This study demonstrates that the moss protonema is an ideal system to observe and manipulate cytology in relation to the desiccation of plant vegetative tissues. With two types of filament, chloronemata and caulonemata, the former cytologically similar to typical photosynthetic cells, the latter with cytology mirroring that of transport cells, that is a prominent endoplasmic microtubular network and extensive endomembrane domains (Pressel et al., 2008), protonemata offer a unique opportunity to gather a wealth of information on the cytological effects of drying and rewetting, unparalleled by any other plant tissue.
Like adult gametophytic tissues, protonemata tolerate losses in excess of 90% of their original water content and promptly recover on rewetting. However, different from leaf tissues of highly DT species, which survive almost independently of the rate of water loss, for example T. ruralis (Oliver & Bewley, 1997), protonemata usually tolerate drying as long as the rate of water loss is slow. Their tolerance to drying is greatly increased by exposure to ABA; when pretreated with ABA, cells are able to withstand normally lethal rates of drying, a result in line with previous studies (Werner et al., 1991; Oldenhof et al., 2006; Rowntree et al., 2007). As reported previously (Oldenhof et al., 2006), the protonema of Physcomitrella does not tolerate full desiccation, unless pretreated with ABA; although it appears to possess the genetic equipment found in DT plants (Cuming et al., 2007), Physcomitrella is drought- but not desiccation-tolerant (Frank et al., 2005).
These results are very much in line with those on adult moss and higher plant vegetative tissues. The close link between survival and slow drying rates in protonemata is clearly indicative of an inducible, protection-based desiccation strategy (Oliver et al., 1998, 2000; Alpert & Oliver, 2002; Proctor & Pence, 2002). Although, in dehydrated cells that recovered, there are no signs of damage, the broken membranes and disrupted cytoskeleton in fast-desiccated specimens clearly explain why these are damaged irreparably. Our finding that the cytology of fast-dried cells pretreated with ABA is essentially the same as that of slow-dried ones, except for the presence of starch grains in the plastids, further corroborates previous observations of a highly protective function of ABA during severe dehydration in moss protonemata (Werner et al., 1991; Oldenhof et al., 2006; Rowntree et al., 2007) and moss adult tissues (Beckett et al., 2000; Beckett, 2001).
The present study provides new insights into the underlying mechanisms of DT in plant vegetative tissues and, in particular, the crucial roles for the cytoskeleton in the ability to withstand near-complete dehydration, as well as novel aspects of the function of ABA in desiccation biology. Our results not only confirm a highly protective role of ABA on membranes (Beckett, 2001; Oldenhof et al., 2006), but also reveal that the pronounced increase in DT elicited by ABA may well depend on ABA-induced changes in cytoskeletal elements that allow for their controlled disassembly even during fast drying.
Our comparison of living specimens mounted in oil and freeze-substituted samples with cells processed by conventional fixation shows that, although in the latter, some level of rehydration must occur during fixation (e.g. the organelle-containing cytoplasm is no longer crowded at the end walls and there is a small increase in the volume of the plastid stroma), the shapes and disposition of the thylakoids and of the mitochondrial cristae in freeze-substituted and chemically fixed specimens are the same. Thus, conventional fixation does not produce any significant artefacts. In some instances, for example the visualization of ER domains and membranous tubules, it may even be preferable, given the tendency of freeze-substitution procedures to wash out membrane lipids, thus leaving white residues, as discussed by Wesley-Smith (2001).
The highly shrunken appearance of slowly-dried protonemata mounted in immersion oil, and resembling that of desiccated moss leaf cells (Tucker et al., 1975; Proctor et al., 2007b), is an inherent consequence of the withdrawal of water from cells (cytorrhesis) (Oparka, 1994), which also mirrors the cell shapes of dry DT angiosperms (Gaff et al., 1976; Hallam & Luff, 1980). Transmission electron microscopy observations of desiccated cells clearly show that the integrity of membranes is retained in the dry state, except in fast-dried specimens. The profoundly different appearances of cells subjected to slow and fast drying regimes immediately signal that major cytological changes during dehydration are a prerequisite for survival. Some of these changes, for example vacuolar fragmentation, changes in the thickness and substructure of the cell wall, and starch fluctuations in the plastids, can also be elicited in hydrated cells by exposure to ABA (Nagao et al., 2005; Rowntree et al., 2007), thus explaining why, once treated, protonemal cells are able to withstand fast drying.
In chloronemata, these changes and subsequent recovery closely resemble those in leaf cells of mosses (Tucker et al., 1975; Proctor et al., 2007b), liverworts (Pressel et al., 2009), pteridophytes (Platt et al., 1997) and angiosperms (Gaff et al., 1976; Gaff, 1997; Dalla Vecchia et al., 1998); In caulonemata, instead, there are remarkable similarities to FCCs (Pressel et al., 2006). Although some of the ultrastructural changes elicited by dehydration can simply be ascribed to the general withdrawal of water from the cells and are, indeed, also observed in fast-desiccated cells that do not recover following rehydration, for example close packing of free ribosomes, reduction in the volume of plastids and mitochondria, and chromatin condensation in the nuclei, others are clearly indicative of more active processes.
A more subtle change in the plastids during dehydration is an increase in the size and frequency of the plastoglobuli. Once considered to have little function other than lipid storage (Greenwood et al., 1963), plastoglobuli have now been implicated not only in the upregulation of plastid lipid metabolism during senescence, but also in response to oxidative stress on the photosynthetic apparatus of vascular plants (Austin et al., 2006 and references therein), drought (Chen et al., 1998; Rey et al., 2000), nitrogen starvation (Bondada & Syvertsen, 2003) and hypersalinity (Locy et al., 1996), amongst other environmental stresses, all causing an increase in their size and number. Various enzymes, including tocopherol cyclase, a key enzyme in tocopherol synthesis, have been identified as components of plastoglobuli. The idea of Vidi et al. (2006) that the final destination of tocopherol is the thylakoid membrane, where it most probably prevents oxidative degradation of fatty acids by reactive oxygen species, provides a plausible biochemical explanation for an active role for the enlarged and numerous protonemal plastoglobuli during oxidative stress that is generally associated with desiccation (Seel et al., 1992; Proctor & Tuba, 2002).
Although mitochondria and nuclei return to their normal state very rapidly, that is within the first hour from rewetting, plastids undergo a more protracted series of shape changes over several hours, although, internally, thylakoid disposition returns to normal more quickly. Stroma-filled extensions of the envelope become apparent after 30 min from rewetting and are particularly prominent after 2 h, especially in chloronemal plastids. These outgrowths are reminiscent of the stroma-filled tubular extensions of the plastid envelope membrane (stromules) described in the tissues of higher plants (Kwok & Hanson, 2004 for a review). Plastid protrusions and/or stromules have been reported in response to abiotic stresses, for example salt stress in the grass Deschanmpsia antartica (Gielwanowska et al., 2005) and high temperatures in Arabidopsis (Holzinger et al., 2007).
Although their functions remain to be elucidated, it has been argued that stromules may allow interorganellar molecular exchange by providing direct connections between the plastids, and enhance intraorganellar exchange by increasing the plastid surface area to volume ratios. Thus, their formation could denote an increase in metabolic activity, allowing cells to adapt to stress conditions (Köhler & Hanson, 2000). That the numerous extensions in protonemal plastids reported here could reflect a marked increase in the metabolic activity of these organelles is in line with physiological studies of recovery in bryophyte cells. Although respiration returns to normal rates within 30–60 min of rehydration, in parallel with the recovery rate of mitochondrial cytology (Tuba et al., 1996; Proctor & Pence, 2002), complete recovery of photosynthesis takes much longer (Proctor et al., 2007b). Similar physiological studies of photosynthetic recovery in protonemal cells are now needed to confirm this assumption, whereas further studies on the dynamics of plastid envelopes in rehydrating protonemal cells may be a key to understanding the functional enigma posed by stromules.
The cell wall and endomembrane domains
The substantial ultrastructural changes in both the wall and endomembrane domains, observed in dried and rehydrating protonemal cells, parallel, not surprisingly given their similar cytologies (Pressel et al., 2008), those reported previously in FCCs (Pressel et al., 2006), namely reorganization of the ER, the appearance of membranous tubules, the laying down of additional wall layers and the formation of a highly convoluted cell wall–plasma membrane interface. As in FCCs (S. Pressel & J. G. Duckett, pers. obs.), these dynamic events are prevented by fast drying, except in cells pretreated with ABA. In higher plants, major changes in wall chemistry and enzymatic activities are associated with both drought and desiccation. Xyloglucan polymer rearrangement is strongly implicated in the former and arabinose polymers in the latter (Moore et al., 2008). Not only do the altered matrix polysaccharides change wall properties during dehydration, but it has also been argued that soluble carbohydrates might be released during drying (Vicréet al., 2004). Thus, the highly convoluted cell wall–plasma membrane interface reported in protonemal cells not only may function in preventing their separation during shrinkage (Pressel et al., 2006), but may also be symptomatic of a high metabolic exchange between the wall and the cytoplasm during drying. Biochemical and immunocytochemical studies are now needed to elucidate exactly which components of both FCCs and protonemal walls change during de- and rehydration, and how these relate to the highly convoluted cell wall–plasma membrane interface and the highly dynamic vesicle populations associated with this.
One further indicator of possible changes in carbohydrate composition during desiccation is the disappearance of starch, a feature in common with DT angiosperms (Gaff, 1997). Exposure to ABA has also been shown to increase the soluble carbohydrate pool in a number of bryophytes (Nagao et al., 2005; Pence et al., 2005; Oldenhof et al., 2006). Nagao et al. (2005) showed that ABA treatment caused a marked reduction in the amount of starch in Ph. patens protonemal plastids, and argued that hydrolysis of starch was necessary to increase the free carbohydrate pool and thus protect the cells from the rigours of freezing. Our findings of an ABA-induced reduction in the starch grains in F. hygrometrica, but not in C. introflexus, suggest that fluctuations in carbohydrates in response to ABA might depend on the constitutive levels of free sugars in a given species. In the highly DT moss, T. ruralis, sucrose levels are constitutively high (Smirnoff, 1992) and do not increase during drying or after exposure to ABA (Bewley et al., 1978).
The importance of cytoskeletal dynamics in DT
Our observations demonstrate that, as in FCCs (Pressel et al., 2006), the controlled disassembly of MTs during drying, and their subsequent reassembly following rewetting, are a prerequisite for the survival of protonemal cells. When the proper dismantling of the MT network is prevented, either by fast drying (which results in irreversibly disrupted MTs) or pharmacologically by exposure to taxol, cells inevitably fail to recover. Similarly, when MT reassembly is inhibited by rehydration in the presence of oryzalin, cells do not regain their control organization and do not survive. Our findings that the removal of MTs with oryzalin before drying also prevents cell survival indicate that the controlled disassembly of MTs during drying is not only necessary, but may have a key role in the desiccation biology of protonemata. We envisage that MTs may have a sensory function in the response of protonemal cells to water stress, similar to that suggested for the MTs of wheat cells under cold stress (Abdrakhamanova et al., 2003).
Our finding that ABA affects the organization of MTs in fully hydrated cells is in line with previous reports on ABA-induced changes in MT orientation in higher plants (Shibaoka, 1994; Jiang et al., 1996) and during the differentiation of protonemal filaments into brood cells (Goode et al., 1993). That ABA-treated cells successfully depolymerize their MTs, even when fast dried, indicates that ABA may be the signal responsible for MT disassembly in response to water stress.
The possibility that drying and/or ABA elicit MT disassembly by altering calcium fluxes, possibly via the ER (whose function in the regulation of cytosolic calcium concentration is well established; Jones, 1985), is the focus of our current research.
This study also shows that, similar to MTs, actin microfilaments depolymerize during drying and their disassembly requires a slow rate of water loss. However, although MT reassembly is required for survival following rewetting, as shown by the oryzalin experiment, microfilament repolymerization seems to play a much less critical part in recovery as cells survive rehydration in cytochalasin. Nevertheless, possible roles for actin in the desiccation biology of protonemal cells cannot be excluded and deserve further scrutiny.
The present demonstration that moss protonemata are an ideal experimental system for desiccation studies has opened up the door to major new advances in functional genomics. Particularly inviting will be screening for mutants differing in their DT and sensitivity to ABA, whose phenotypes manifest changes in organelle disposition, the cell wall, plastid morphology and both the actin and microtubular cytoskeletons.
S.P. gratefully acknowledges the financial support of a Leverhulme Trust Early Career Fellowship. Cryofixation and freeze-substitution protocols were carried out by Dr T. Brain, CUI, King’s College, London, UK.