1. Morphological adaptation, changes of volume, tissue protection
Morphological adaptations are important to DT plants both mechanically, and in protection against the deleterious effects of excessive irradiance (Demmig-Adams & Adams, 1992; Björkman & Demmig-Adams, 1995). DT bryophytes mostly have rather thick cell walls and often look remarkably different wet and dry, exposing much less leaf area in the dry state (Fig. 3c–d). The dry leaves often show regular patterns of folding and shrinkage (Proctor, 1979a). Leaves of vascular resurrection plants typically roll rather than becoming flaccid in response to water loss, protecting the more delicate desiccated cell walls and giving the plants mechanical strength; the surface exposed in the dry state is often heavily pigmented, hairy or scaly (e.g. Ceterach, Fig. 3a–b). The lack of obvious wilting is due to the even shrinking of the mesophyll cells and intercellular spaces, compared with the collapse of mesophyll cells and enlargement of intercellular spaces that occurs in sensitive species (Lebkuecher & Eickmeier, 1991). Leaf area can decrease by as much as 85% in Craterostigma as the leaves contract and curl during desiccation. The leaf cells shrink, preserving the contact between plasmalemma and cell wall during desiccation, with accordion-like folding of the cell walls (Sherwin & Farrant, 1996; Hartung et al., 1998). Leaf folding can occur along or perpendicularly to the leaf axis. Ferns and some DT grasses curl rather than folding. These movements are caused by differential imbibition, rather than by osmotic effects (Gaff, 1989). The remarkable longitudinal zigzag folding of Nanuza plicata (Velloziaceae) leaves is a result of the alternating reinforced and nonreinforced bands in the epidermis (Rosetto & Dolder, 1996). The similar contraction of mesophyll between parallel veins of Xerophyta scabrida (Velloziaceae) during dehydration results in a decreasing specific leaf area, which in turn minimizes water loss (Tuba et al., 1996). Borya nitida has small, needle-like leaves with cutinised outer epidermal walls and stomata confined to longitudinal grooves (Gaff & Churchill, 1976). Similar morphology is often seen in plants tolerant of drought but not of desiccation. Other features traditionally seen as xeromorphic, like leathery or pubescent leaves, sclerenchyma, hairs and scales, also occur among DT ‘resurrection’ plants, but not consistently. They should be interpreted with the same caution as the ‘xeromorphic’ features of peat-bog Ericaceae, which a century ago were explained as adaptations to ‘physiological drought’. It is not possible to be sure from its anatomy whether a plant is desiccation tolerant or not.
Figure 3. (a) The desiccation tolerant (DT) fern Ceterach officinarum fully hydrated and metabolically active, and (b) dry during summer, with the leaves tightly rolled and the densely scaly undersides exposed. Chudleigh, Devon, UK. (c) Tortula (Syntrichia) intermedia, a DT moss (above) fully hydrated, and (below) dry. Chudleigh, Devon, UK. (d) The DT moss Leptodon smithii; (above) hydrated and fully expanded, and (below) dry, with the shoot systems tightly rolled in characteristic balls recalling the growth habit of Selaginella lepidophylla. Locronan, Brittany, France. (e) Xerophyta retinervis, a tall (c. 2 m) member of the Velloziaceae, seen here fully hydrated with green leaves, in open seasonally dry forest, Itala Game Reserve, KwaZulu-Natal, South Africa. (f) Xerophyta villosa, a lower-growing species, on sun-exposed rock outcrops, here hydrated and green. Itala Game Reserve, KwaZulu-Natal, South Africa. (g) The DT shrublet Myrothamnus flabellifolia, hydrated, with fully expanded leaves. Louwsburg, KwaZulu-Natal, South Africa. (For photographs of Craterostigma in hydrated and desiccated states see Bernacchia et al. (1996); Bohnert (2000); Scott (2000)).
Download figure to PowerPoint
2. Questions of scale
Appropriate adaptation depends greatly on scale (Niklas, 1994). Area increases as the square, and volume (and mass) as the cube, of linear dimensions. Surface tension operates on linear liquid–surface contacts and is therefore most effective and important at small scales. It is a trivial force for large vertebrates but life-or-death for insects; it plays little part in the rise of water up tall trees but is vitally important for lichens and bryophytes. Conversely, effects of gravity are trivial at small scales, but paramount at the scale of large animals or trees. Fluid boundary-layer thickness varies as the square root of the linear dimensions of an obstruction in the flow. To a first approximation the thickness of the atmospheric boundary layer may be thought of as an invariant part of the environment as perceived by an individual plant.
Scale thus has profound effects on mechanical and physiological constraints and adaptation in plants. The small size of bryophytes gives them larger cross-section to mass ratios, reducing the need for specialised supporting and conductive tissues; a bryophyte is similar in scale to an individual leaf or root of a vascular plant. Small vascular plants seldom show much secondary thickening; wood is a structural material evoked by the mechanical needs of trees. Large size both permits and demands a specialised internal conducting system. It is among the largest erect-growing mosses (e.g. Polytrichaceae, Dawsoniaceae, Mniaceae) that the nearest approach to a seed-plant vascular system and seed-plant rigidity are seen. The small size of bryophytes means that surface tension is a powerful force determining the distribution of water around their shoots, and that the shoots and leaves tend to lie within the laminar atmospheric boundary layer. Mass-transfer rates in their immediate vicinity are therefore dominated by (slow) molecular diffusion rather than (fast) turbulent mixing. Consequently, bryophytes do not in general need to package their photosynthetic systems within a ventilated epidermis, although Polytrichaceae, Dawsoniaceae and Marchantiales demonstrate that they have the evolutionary potential to produce functional equivalents of vascular-plant leaves. Individual bryophyte leaves are not comparable physiologically with vascular-plant leaves. A typical bryophyte colony is a photosynthetic system on a scale intermediate between a vascular-plant leafy canopy and a vascular-plant mesophyll. At low windspeeds, a higher-plant leaf as a gas-exchange system may be compared with the surface of a smooth bryophyte mat or cushion, in which the surfaces of the individual leaves are functionally a scaled-up mesophyll. Leaf canopies of bryophytes tend to have very high ‘LAIs’ by vascular-plant standards; a few estimates by one of us gave figures of 6 in Tortula (Syntrichia) intermedia, 18 in Mnium hornum and 20–25 in Scleropodium (Pseudoscleropodium) purum. The higher of these figures are in the same range as mesophyll/leaf-area quotients of vascular plants (Nobel, 1974, 1977). At higher windspeeds the bryophyte colony has no close analogue in terms of familiar vascular-plant physiology, but is readily comprehensible in terms of the physics of boundary layers, gaseous diffusion and heat balance. Evolution of much of the detailed morphology of bryophyte shoot systems has probably been driven by selection pressures balancing movement and storage of water against free gas exchange to the leaf surfaces (Dilks & Proctor, 1979; Proctor, 1979a). Carbon isotope discrimination measurements of bryophytes (Rundel et al., 1979; Teeri, 1981; Proctor et al., 1992) generally give δ13C figures in the same range as C3 vascular plants, which suggests that the relation between diffusive and ‘carboxylation’ resistances is similar in the two groups. This is consistent with the finding of Martin & Adamson (2001) that on a per-chlorophyll basis photosynthetic rates of bryophytes and vascular plants are similar.
Raven (1999) has emphasised that there is a lower size limit for homoihydric vascular plants. In fact, the vascular pattern of adaptation probably ceases to be optimal at a considerably larger scale than the limit he suggests, because the vascular pattern of adaptation only becomes seriously competitive for a plant large enough to tap reserves of water at significant depth in the soil, and tall enough to create its own individual boundary layer above ground.
3. Water-content components, water storage and water movement in vascular and nonvascular plants
The water associated with a plant may be divided into several components. Water inside the plasmalemma is symplast water. Water in the cell walls and intercellular spaces is often referred to as apoplast water, but this is less satisfactorily defined, and various components may be excluded from it. In vascular plants it is physically continuous with the xylem sap, which often may be better considered separately. Again, for some purposes, intercellular water may be recognised as a separate component, distinct from (and much more mobile than) the water in the cell walls (Beckett, 1995, 1997). Bryophytes generally lack intercellular spaces, but often carry large amounts of (physiologically important) external capillary water (Buch, 1945, 1947; Dilks & Proctor, 1979; Proctor et al., 1998). Water may be stored in various parts of the soil–plant system, buffering the plant against rapid changes in availability of water in the environment. Vascular plants typically rely almost entirely on the store of water in the soil. Halophytes and CAM succulents store large amounts of water in the symplast, and symplastic water storage is important in underground storage organs of geophytes, sometimes providing most or all of the water for development of inflorescences during the dry season. The structure of many bryophytes favours short-term storage of large quantities of external capillary water. Many algae and cyanobacteria have thick gelatinous cell walls or mucilaginous envelopes. Lichen thalli can store substantial amounts of water, generally in intercellular or apoplast locations (including the thick mucilaginous cell walls of Collema and other lichens with Nostoc as photobiont). In all of these nonvascular plants, much or most of the extracellular water can be lost before water potential falls sufficiently to affect metabolism. Although conventionally regarded as ‘poikilohydric’, when they are hydrated they probably experience less variation in cell water content or water potential than most ‘homoihydric’ vascular plants.
Water movement in a plant may take place entirely through the symplast, passing from cell to cell through the plasmadesmata, or water may pass through the plasmalemma of one cell, across the intervening cell wall, and enter through the plasmalemma of the other. These have generally been taken to be relatively high-resistance pathways. The apoplast pathway through the ‘free space’ of the cell walls and intercellular spaces commonly shows a lower resistance to water movement, and is generally assumed to be the main pathway within the root cortex and the mesophyll of vascular-plant leaves, though this assumption has been questioned recently (Steudle & Peterson, 1998). The xylem of vascular plants provides a pathway of very much lower resistance than either. Although ‘intracellular’ in origin, the lumina of the mature xylem vessels and tracheids are in effect a major extension of the apoplast pathway with which they are in contact. The hydroids of some large mosses (e.g. Polytrichaceae, Dawsoniaceae, Mniaceae) (Hébant, 1977) are in effect a parallel evolutionary development of different ontogeny serving a similar function (Ligrone et al., 2000). But typically, water movement in bryophytes is mostly external (Mägdefrau, 1935; Buch, 1945, 1947), and many species show elegant external capillary conducting structures (Dilks & Proctor, 1979; Proctor, 1979a). External capillary water movement is also important in the monocotyledon Xerophyta and other Velloziaceae, especially during the initial stages of remoistening.
In general, the internal symplastic and apoplastic pathways can support substantial differences in water potential. Despite recent controversy (Zimmerman et al., 1993; Canny, 1995; Milburn, 1996; Tyree, 1997), there can be little doubt that the internal xylem pathway of vascular plants, sealed from the exterior by the surrounding tissues, can also support substantial negative water potentials – though the water relations of trees are probably less simple, and local negative tensions in the xylem may not be as large, as has been commonly envisaged. The external conducting systems of bryophytes (and Xerophyta) are dependent on the maintenance of high water potentials (> c. −0.05 MPa) in their immediate surroundings – which must include the cells with which they are in contact.
4. What are the fundamental requirements for desiccation tolerance? Why are some species more tolerant than others?
The minimum requirements for survival of desiccation must include preservation intact of the genetic material, the mechanism for protein synthesis, and some essential spatial relationships of structures within the cell. Considerations of energetics and synthesis capacity in relation to observed rates of recovery suggest that retention intact of the major structural proteins and enzyme systems should be added to that minimal list.
The adaptations involved in desiccation tolerance are of two kinds, those essential to tolerance (without which tolerance would not be conceivable), and those that are consequences or corollaries of it. In the first group are factors that preserve the integrity of membranes and macromolecules in the dry state, and maintain essential spatial relationships within the cytoplasm; this must include ability of cell walls to shrink or fold without strain as the cytoplasm loses volume on drying. Control must also be maintained over the relative rates and integration of metabolic processes during drying and remoistening; vitrification of the cell contents as water is lost may be important in achieving all of these needs (Crowe et al., 1998; Buitink, 2000; Buitink et al., 2002). The second group includes larger-scale structural adaptations, and, for example, metabolic protection against oxidative damage (Smirnoff, 1993; Foyer et al., 1994; Alscher et al., 1997) – important to all plants, but a hazard often assumed as likely to bear more heavily on DT plants during drying and recovery.
High concentrations of disaccharide sugars are a common characteristic of DT organisms. Trehalose fills this role in organisms ranging from bacteria and fungi to nematodes, brine shrimps and the clubmoss Selaginella lepidophylla Adams et al. (1990). Yeast (Saccharomyces cerevisiae) produces little trehalose during rapid growth and is then relatively intolerant of drying, but trehalose accumulates and dehydration tolerance increases in the stationary phase, or following heat shock. In plants, sucrose is generally the dominant disaccharide that accumulates in DT cells. Disaccharides stabilise phospholipid bilayers by hydrogen bonding to the polar head groups, maintaining the spacing between them and preventing damaging phase transitions. Trehalose has also been shown to be effective in stabilising labile proteins during drying (Crowe et al., 1992), and sucrose has similar effects (Schwab & Gaff, 1990; Bustos & Romo, 1996; Suzuki et al., 1997). By contrast with the protective effects of the disaccharides, reducing sugars show a browning reaction with dry proteins, which leads to denaturation (Wettlaufer & Leopold, 1991). The possible stabilising role of linear polyols (Crowe et al., 1998; Buitink et al., 2002), which occur widely in DT (and non DT) liverworts, algae and lichens needs further investigation.
Sucrose levels show marked increases in the course of dehydration in most DT vascular plants (Albini et al., 1994; Bianchi et al., 1991, 1993; Müller et al., 1997; Ghasempour et al., 1998). With occasional exceptions, reducing sugars decrease or remain at low levels on drying (Koster & Leopold, 1988; Bianchi et al., 1991, 1993). Trehalose occurs widely in DT plants, but usually at low concentrations, and Ghasempour et al. (1998) found no consistent pattern with drying; trehalose seems to be completely absent from such DT species as Boea hygroscopica and Xerophyta villosa (Bianchi et al., 1991; Ghasempour et al., 1998). The carbon source for sucrose synthesis probably varies from species to species. In Craterostigma plantagineum the main source is 2-octulose stored in the leaves (Bianchi et al., 1991; Norwood et al., 2000). Possible sources in other species include starch (or other carbohydrate) in the leaves or other parts of the plant, or photosynthesis during the drying period, but few critical measurements have been made (Scott, 2000). In bryophytes, the available evidence indicates that sucrose concentrations remain constantly high, with no increase on drying, but reducing sugars remain at (or decline to) very low levels as the plant dries (Smirnoff, 1992).
It is generally agreed that disaccharides (and other sugars) have an essential role in desiccation tolerance, but that other factors must be important too. The most discussed of these is the part played by proteins and protein synthesis. Two questions may be asked. Are particular protein molecules an essential part of the equipment of desiccation-tolerant cells?; and, where and when is protein synthesis important in reinstating or repairing cell components damaged or inactivated during dehydration and re-wetting (Osborne et al., 2002; Walters et al., 2002)?
At least a partial answer to the first question may be drawn from analogy with seeds, where late-embryogenesis abundant (LEA) proteins play a major part in establishing the desiccation tolerance of the mature embryo. LEA proteins (including dehydrins (Close, 1996)) are hydrophilic and resistant to denaturation, and fulfil a role in membrane and protein protection parallel with that of the disaccharides. They are also important along with the sugars in maintaining vitrification of the cell contents; indeed the physical properties of intracellular glasses appear to be determined primarily by their protein rather than their sugar component (Buitink et al., 2000, 2002).
In answer to the second question, it is evident that a wide range of response exists between poikilochlorophyllous DT vascular plants such as Xerophyta humilis in which recovery entails extensive protein synthesis (mostly translation of existing transcripts) for reinstatement of the photosynthetic system (Dace et al., 1998), and DT bryophytes of exposed sites such as Racomitrium lanuginosum or Tortula (Syntrichia) ruralis in which recovery is essentially complete within an hour or two, even in the presence of protein-synthesis inhibitors (Proctor & Smirnoff, 2000; Proctor, 2000, 2001). Some (probably many) vascular DT plants fall between these extremes. Thus Craterostigma wilmsii requires protein synthesis for full recovery of photosynthetic function if dried fast (< 24 h), but not if it dries slowly over several days (Cooper, 2001). In all DT plants there is of course likely to be an ongoing need for some protein synthesis to balance protein degradation when partially hydrated (Gaff, 1989). Characteristic changes in gene expression during desiccation and rehydration have been reported from vascular resurrection plants including the fern Polypodium virginianum (Reynolds & Bewley, 1993a), the dicotyledon Craterostigma plantagineum (Ingram & Bartels, 1996), the monocotyledon (grass) Sporobolus stapfianus (Gaff et al., 1997; Blomstedt et al., 1998), and in the DT moss Tortula (Syntrichia) ruralis (Oliver, 1996; Oliver & Bewley, 1997; Oliver et al., 1998; Wood et al., 1999; Oliver et al., 2000a). It has been possible to point to a few homologies with proteins of known function in other organisms (e.g. genes in Craterostigma and Tortula which appear homologous with LEA proteins of seeds, and various metabolic enzymes), but the function of most of these genes is unknown and relating them to processes involved in desiccation tolerance remains a research challenge for the future.
It is widely considered that cell damage and metabolic disruption during drying and rehydration must exacerbate release of active oxygen species, and that high antioxidant activity to deal with it should be an essential part of the adaptation of DT plants (Stewart, 1990; Smirnoff, 1993; Navari-Izzo et al., 1997). Protection against active oxygen species produced during normal metabolism is essential to all plants (Foyer et al., 1994; Alscher et al., 1997), and up-regulation of active oxygen-scavenging enzymes in response to drought is well documented in nonDT plants. However, when responses to desiccation are compared, both DT and sensitive species show stimulation of antioxidant activity at one or another point in the drying–rehydration cycle but it generally appears that oxidative damage in the DT species is well controlled whereas the sensitive species show clear signs of oxidative damage (Dhindsa & Matowe, 1981; Seel et al., 1992a,b). Dhindsa & Matowe found activities of both superoxide dismutase (SOD) and catalase higher in the DT moss Tortula ruralis than in the sensitive Cratoneuron filicinum by a factor of around four. Seel et al. (1992b) found SOD and catalase activity substantially higher in (DT) Tortula ruraliformis than in (sensitive) Dicranella palustris, but activities of peroxidase and ascorbate peroxidase (AP) were somewhat greater in the sensitive than in the DT species; they suggested that the role of active oxygen-processing enzymes may be less important than that of antioxidants (tocopherols and glutathione) in determining desiccation tolerance. Sgherri et al. (1994a) found the activity of glutathione reductase (GR) and dehydroxyascorbate reductase (DHAR) approximately doubled, and that of AP halved, in dried leaves of the DT grass Sporobolus stapfianus, with a return towards normal levels after 24 h rehydration. The proportion of oxidized ascorbate roughly doubled on drying, and doubled again on rehydration, but there was little change in the ratio of oxidized to reduced glutathione. In the dicotyledon Boea hygroscopica (Sgherri et al., 1994b) they found total ascorbate doubled and glutathione increased 50-fold on drying; activity of GR and DHAR fell markedly on drying but AP was little changed. All the variables returned to normal levels after 24 h rehydration, in both slowly dried plants which recovered, and rapidly dried plants which did not. In Craterostigma wilmsii, Myrothamnus flabellifolia and Xerophyta viscosaSherwin & Farrant (1998) and Farrant (2000) found activity of all three enzymes, AP GR and SOD increased during at least some phase of the drying–rehydration cycle, but the detailed pattern of response varied greatly between species. Thus there is plenty of evidence of responses of antioxidants and antioxidant enzyme systems to drying and rehydration, but it is hard to discern any consistent general pattern. Lebkuecher & Eickmeier (1991, 1993) and Muslin & Homann (1992) showed that leaf curling in dry Selaginella lepidophylla and Polypodium polypodioides brings clear benefits in limiting photodamage during rehydration in bright light, but their results do not suggest any fundamental differences from nonDT plants. Over all, the evidence suggests that DT plants generally deal with the potential hazards of oxidative damage during the drying–re-wetting cycle by anticipating the problem at source rather than by invoking extravagantly high activity of antioxidant enzymes or antioxidants after the event. Protective mechanisms include leaf and stem curling, heavy anthocyanin pigmentation (Sherwin & Farrant, 1998; Farrant, 2000), progressive reduction of molecular mobility and controlled down-regulation of metabolism on drying (Hoekstra et al., 2001), and high levels of zeaxanthin-mediated photo-protection (Demmig-Adams & Adams, 1992; Eickmeier et al., 1993; Björkman & Demmig-Adams, 1995) manifested in chlorophyll-fluorescence measurements as high levels of (quickly relaxing) NPQ (Csintalan et al., 1999; Marschall & Proctor, 1999; M.C.F. Proctor, unpublished). A fuller discussion of active oxygen and antioxidant systems in relation to desiccation tolerance is given by Smirnoff (1993), and further references will be found in Walters et al. (2002).
5. Homoichlorophylly and poikilochlorophylly
Vascular DT plants fall into two groups depending on the degree to which they retain their chlorophyll when dry. Homoiochlorophyllous (HDT) species retain their photosynthetic apparatus and chlorophylls in a readily recoverable form. Poikilochlorophyllous (PDT) species dismantle their photosynthetic apparatus and lose all of their chlorophyll during drying; these must be resynthesised following rehydration (Tuba et al., 1994; Sherwin & Farrant, 1996; Tuba et al., 1998). The phenomenon of loss of chlorophyll during desiccation was first described by Vassiljev (1931) in Carex physoides from central Asia; the concept was reintroduced and the terms homoichlorophylly and poikilochlorophylly coined by Hambler (1961), and it was regarded as a special case in certain DT monocotyledonous plants (Hambler, 1961; Gaff & Hallam, 1974; Gaff, 1977, 1989; Bewley, 1979; Hetherington & Smillie, 1982a,b).
The HDT strategy is based on the preservation of the integrity of the photosynthetic apparatus by protective mechanisms considered in earlier sections of this review. The PDT strategy evolved in plants which are anatomically complex and which include the biggest in size of all DT species, and it can be seen as the evolutionarily youngest strategy. It is based on the dismantling of internal chloroplast structure by an ordered deconstruction process during drying, and its resynthesis upon rehydration by an ordered reconstruction process. These processes can thus be thought of as not only being superimposed on an existing cellular protection mechanism of vegetative desiccation tolerance (Oliver et al., 2000) but as a distinct new DT strategy (Tuba et al., 1994; Tuba et al., 1998). The selective advantage of poikilochlorophylly, in minimising photo-oxidative damage and not having to maintain an intact photosynthetic system through long inactive periods of desiccation, presumably outweighs the disadvantage of slow recovery and the energy costs of reconstruction.
Taxonomically the PDT plants appear to be restricted to the monocots (Gaff, 1977, 1989; Bewley & Krochko, 1982). Poikilochlorophylly is currently known in eight genera of four families (Cyperaceae, Liliaceae (Anthericaceae), Poaceae and Velloziaceae). Most occupy the almost soil-less rocky outcrops known as inselbergs, in strongly seasonal subtropical climates (Porembski & Barthlott, 2000); the best studied physiologically are the African Xerophyta scabrida, X. viscosa and X. humilis and the Australian Borya nitida (Gaff & Churchill, 1976; Hetherington & Smillie, 1982a,b; Hetherington et al., 1982a,b; Gaff & Loveys, 1984; Tuba et al., 1993a,b, 1994, 1996; Sherwin & Farrant, 1996; Dace et al., 1998; Farrant, 2000; Cooper, 2001).
HDT plants preserve much or all of their chlorophyll through a drying–re-wetting cycle. The fern Pellaea calomelanos retains chlorophyll and chloroplasts with discernible grana when dry (Gaff & Hallam, 1974). DT dicotyledons, although all retaining chlorophyll, vary in the details of their behaviour. In the southeast-European HDT Gesneriaceae Haberlea rhodopensis and Ramonda serbica, Markovska et al. (1994) found no significant changes in chlorophyll content on drying and re-wetting, but Drazic et al. (1999) found that Ramonda nathaliae lost 20% of its chlorophyll when desiccated in the glasshouse and 70% in natural habitats; Myrothamnus flabellifolia preserves its thylakoids but loses half its chlorophyll on drying (Farrant et al., 1999). The extent of chlorophyll loss in dicotyledons thus varies from species to species, and may be influenced by environmental factors. By contrast, Xerophyta scabrida, a characteristic PDT plant, loses all its chlorophylls and dismantles apparently all of the internal structures of the chloroplast during desiccation (Tuba et al., 1993a,b).
Desiccation-induced breakdown of the photosynthetic apparatus in PDT plants is different from the processes involved in leaf senescence. Indeed Gaff (1986) found that senescence reduced desiccation tolerance. The dismantling of the photosynthetic apparatus can be seen as a strictly organized protective mechanism, rather than ‘damage’ to be repaired after rehydration. If the PDT Borya nitida is desiccated rapidly, it does not have the time to break down chlorophyll and loses viability (Gaff & Churchill, 1976) and the same is true of Xerophyta humilis (Cooper, 2001). Chlorophyll breakdown in Borya nitida does not appear to be controlled by abscisic acid (ABA) (Gaff & Loveys, 1984), and Xerophyta scabrida preserves most of its chlorophyll when desiccated in the dark, so most of the loss seems to be a result of photooxidation under natural circumstances (Tuba et al., 1997). This suggests that PDT plants in general probably do not decompose their chlorophyll enzymatically, but rather do not invest in preserving it through the dry state. However, dismantling of the thylakoid membranes is strictly organized and leads to the formation of nearly isodiametric ‘desiccoplasts’, which, unlike chromoplasts, are able to regreen and photosynthesise after rehydration (Tuba et al., 1993b). Desiccoplasts contain granular stroma, a couple of translucent plastoglobuli possibly containing lipoquinones and neutral lipids. The thylakoid material is not arranged in plastoglobuli but can be found as osmiophilic lipid material stretched in the place of the former thylakoids (Tuba et al., 1993b). In X. villosa not only chloroplast internal membranes but even most of the mitochondrial cristae disappear on dehydration and the remaining ones appear to decompose within 30 min after rewetting. This is mirrored by a loss in insoluble or structural proteins (almost 50%), which is much less marked (generally c. 10%) in HDT plants (Gaff & Hallam, 1974).
On re-wetting X. scabrida, reconstruction of thylakoids has already begun by the time the tissue has reached full saturation. Chlorophyll synthesis starts 8–10 h after rehydration (Tuba et al., 1993a). In this process, osmiophil lipid material is used up. Primary thylakoids with a small stacking ratio make up an intermediate step in resynthesis. Regeneration of the photosynthetic apparatus is complete within 72 h after rewetting. The restitution of cristae in mitochondria was faster than that of chloroplast internal membranes. Net CO2 assimilation was first measurable after 24 h rehydration. At this point, chlorophyll content was just 35% of that in fully active control plants (Tuba et al., 1993b).
CO2 gas-exchange of desiccating Xerophyta scabrida leaves was studied by Tuba et al. (1996, 1997). Photosynthetic activity declined steeply to zero through the first 4 d of drying. By this time the leaves had lost approx. 60% of their water content. The fall in net CO2 assimilation was caused mainly by the rapid stomatal closure, but it was the decrease in chlorophyll content (as shown by the fluorescence signal) that finally brought photosynthesis to a halt. Dark respiration continued until the 14th day of dehydration, at 24% of the original water content. The declines in water content and respiration were closely correlated. The prolonged respiration during desiccation may cover the energy demand of controlled disassembly of the internal membrane structures in PDT plants (Tuba et al., 1997). In rehydrating Xerophyta scabrida leaves, respiratory processes are fully operational before full turgor is achieved (Tuba et al., 1994); fast recovery of respiration is important as the only ATP-source during the primary phase of rehydration (Gaff, 1989). A high initial respiration rate on re-moistening, so called ‘resaturation respiration’ (Smith & Molesworth, 1973), is seen in lichens, bryophytes and HDT vascular plants, but this elevated rate persists much longer in PDT species – 30 h in the case of X. scabrida (Tuba et al., 1994).
The HDT and PDT strategies solve the same ecological problem, but cover a broad temporal range of adaptation. The HDT pteridophytes and angiosperms are generally adapted to longer drying–wetting cycles than bryophytes and lichens, but to more rapid alternations of wet and dry periods than the PDT monocot species (Schwab et al., 1989; Ingram & Bartels, 1996; Sherwin & Farrant, 1996; Tuba et al., 1998), though some can survive dry for very long periods of time. The PDT strategy has evolved in habitats where the plants remain in the desiccated state for 5–8(−10) months. Under these conditions it is evidently more advantageous to dismantle the whole photosynthetic apparatus and reconstitute it after rehydration. Of course there is variation within each category and the categories overlap in their ecological adaptation, and two or more may coexist in one habitat, for example on inselbergs (Ibisch et al., 1995). Both ends of this ecological spectrum have particular points of interest. There is probably a trade-off between the ‘cost’ of protection and repair to the photosynthetic apparatus if this is kept in a quickly recoverable state through prolonged periods of desiccation, and the ‘cost’ of reconstituting the photosynthetic apparatus de novo.
6. Constitutive and induced tolerance
In the bryophytes and lichens of habitats that frequently dry out, desiccation tolerance appears to be essentially constitutive and influenced to only a limited extent by previous desiccation history or rate of drying. Recovery takes place quickly, and depends essentially on reactivation of components conserved undamaged through the drying–re-wetting cycle. By contrast, in DT angiosperms tolerance is generally induced in the course of slow drying (Gaff, 1980, 1989, 1997). These plants may tolerate severe and prolonged desiccation under appropriate conditions, but show little or no tolerance if they are dried rapidly. This antithesis may at least in part reflect different evolutionary origins of desiccation tolerance, primitive in bryophytes (and lichens), secondary and polyphyletic in vascular plants (Oliver et al., 2000a), where it has probably often evolved as a fall-back in plants already more or less tolerant of drought stress. Slow drying is a natural consequence of the vascular-plant habit, and in some cases elements of the induction process are obvious, as in the conversion of 2-octulose to sucrose in Craterostigma, and the controlled dismantling of the photosynthetic system in Borya and Xerophyta. However, the antithesis is not clear-cut. Leaf cells of mosses in exposed sunny situations switch from full turgor to air dryness with a few minutes, but many forest bryophytes dry much more slowly, and a degree of drought hardening is readily demonstrated (Höfler, 1946; Abel, 1956; Proctor, 1972; Dilks & Proctor, 1976). Also, desiccation tolerance appears to be essentially constitutive in various small saxicolous ferns (Hymenophyllym tunbrigense, H. wilsonii, H. sanguinolentum, Asplenium trichomanes, A. ruta-muraria) and Selaginella species (S. cf. underwoodii) which can dry out and recover as quickly as many bryophytes (Kappen, 1964; M.C.F. Proctor, unpublished). These species dry out quickly once water becomes limiting, and show no conventional adaptations to conserve water within the plant, a habit probably related to the high temperatures reached in their saxicolous habitats in the absence of evaporative cooling. By contrast the taller (and often epiphytic) polypody ferns (Polypodium vulgare agg.) show strong stomatal control over water loss and dry out slowly (Lange et al., 1971; M.C.F. Proctor, unpublished).
It is generally accepted that ABA produced in roots under water stress induces various drought responses in vascular plant shoots (Schulze, 1986; Zeevaart & Creelman, 1988). Molecular-biological evidence (Ingram & Bartels, 1996) has tended to confirm earlier speculation that ABA may have a general role in evoking desiccation tolerance in DT plants. However physiological evidence for this is equivocal. Chamaegigas intrepidus and Craterostigma lanceolatum increased their ABA content 20–30-fold on drying (Schiller et al., 1997). On the other hand, ABA concentration did not increase in drying leaves of Polypodium virginianum, but exogenous ABA enhanced survival of rapid drying (Reynolds & Bewley, 1993b). Gaff & Loveys (1984) found a substantial rise in ABA levels (along with increased desiccation tolerance) in detached leaves of both (homoichlorophyllous) Myrothamnus flabellifolia and (poikilochlorophyllous) Borya nitida equilibrated at 96% rh; cell survival was significantly promoted by exogenous ABA without concurrent water stress in B. nitida, and slightly so in M. flabellifolia. In the grass Sporobolus stapfianus isolated leaves became DT only if detached below 61% relative water content (RWC), but ABA peaked in shoots at 15% RWC, and originated in the leaves rather than the roots (Gaff & Loveys, 1992). In the same species, Ghasempour et al. (1998) found desiccation tolerance was promoted only marginally by exogenous ABA, whereas brassinolide and methyljasmonic acid had much larger effects. In the DT thalloid liverwort Exormotheca holstii, high levels of ABA went with high desiccation tolerance under field conditions, ABA level and desiccation both declining under well-watered cultivation, but desiccation tolerance could be restored by exogenous ABA (Hellwege et al., 1994). Tolerance of water stress was increased significantly by ABA treatment even in the aquatic thalloid liverwort Riccia fluitans (Hellwege et al., 1996). Similar effects of ABA in promoting desiccation tolerance have been found in protonema of the moss Funaria hygrometrica (Werner et al., 1991; Bopp & Werner, 1993) and leafy shoots of Atrichum androgynum (Beckett, 1999). In the highly DT moss Tortula ruralis ABA is undetectable (Oliver et al., 1998). A systematic search for endogenous ABA in bryophytes is much needed.
7. Recovery; re-establishment of water-relations and metabolism
Highly DT bryophytes growing in exposed situations typically re-moisten quickly and diffusely. The thin leaves of bryophytes (and lichen thalli and algal mats) equilibrate rapidly with the humidity of the surrounding air, and can regain turgor from overnight dewfall even if no obvious deposition of liquid water has taken place (Lange et al., 1968, 1990, 1991, 1994; Csintalan et al., 2000). Liquid water from rain or cloudwater deposition is imbibed by the leaves directly as it spreads by capillarity over the surface of the shoots. In species with well-developed internal conduction and water-repellent leaf surfaces, such as many large Polytrichaceae and Mniaceae, imbibition is slower and may take an hour or two to complete even when liquid water is present. The water-conducting hydroids of mosses collapse without embolising on drying (J.G. Duckett, pers. comm.), so they are able to refill immediately through their lateral walls as the stem re-imbibes water.
Vascular resurrection plants in general require liquid water for rehydration (Gaff, 1977), and the continuity of water within the xylem must be restored as the tissues rehydrate if the leaves are to remain turgid for more than a few hours after rain. Embolism in xylem vessels has been assumed to hinder resaturation of the photosynthetic tissues in some species (Sherwin & Farrant, 1996). The fact that resurrection plants are generally low herbs or shrubs (Gaff, 1977; Kappen & Valladares, 1999) is consistent with the idea that restitution of xylem transport is a real challenge for these plants, and probably imposes a practical limit on their height. If normal function of the vascular system is to be re-established, it is vital that the vascular bundles remain essentially undamaged in the desiccated state. Rosetto & Dolder (1996) found that vascular bundles of the desiccation tolerant Nanuza plicata showed no significant changes due to dehydration. By contrast, drought stress severely impaired vascular bundles in nonDT barley leaves (Pearce & Beckett, 1987), which would preclude the reestablishment of water flow even if all the other factors necessary for rehydration were present. In Myrothamnus flabellifolia, a DT shrub with a height of at most a few decimetres (Figs 3g and 4a), Sherwin et al. (1998) considered capillary rise in the narrow reticulate xylem vessels sufficient to achieve continuity of water column, but Schneider et al. (2000) and Wagner et al. (2000) presented evidence that root pressure is implicated in the initial filling of a proportion of xylem elements, water then moving radially in a complex manner into the adjacent xylem and other tissues. In Pellaea calomelanos, Gaff & Hallam (1974) found capillary rise unable to transport water even into the lowest leaves, and concluded that root pressure must also play a part. In Borya nitida, the leaves on the upper half of the shoots had still not rehydrated from the soil after 6 d (Gaff & Churchill, 1976), so in this case rehydration through the leaf cuticle seems also to be essential. Xerophyta scabrida roots die back when desiccated, therefore rehydration can only occur via the leaves. In this species leaf turgor is regained within 6 h and maximum water content after 12 h in immersed leaves (Tuba et al., 1994). After re-establishment of metabolism in the leaves, adventitious roots develop and ensure continuous water supply (Tuba et al., 1993a). The relative importance of the water received by the roots and water absorbed through leaves varies from species to species (Gaff, 1977).
Figure 4. (a) Habitat of Myrothamnus flabellifolia, Cheilanthes sp. and Selaginella dregei over flat-bedded sandstone slabs near Louwsburg, KwaZulu-Natal, South Africa. (b) Sandy dry grassland near Fülöpháza, Hungary; the desiccation tolerant (DT) moss Tortula ruralis carpets much of the ground under the sparse grass cover. (c) Arnavatnsheii, C. Iceland: the DT moss Racomitrium lanuginosum dominant on stony slopes in foreground (with yellowish DT lichens, Alectoria sp.), and giving its grey-green colour to much of the rest of the landscape. (d) Lichen field dominated by the orange DT lichen Teloschistes capensis, near Cape Cross, Namibia. (e) Quercus robur wood under high rainfall, Black Tor Copse, Devon, UK; DT mosses and lichens cover almost every available surface. (f) Cool-temperate rainforest, Newell Creek, SW Tasmania; dense DT bryophyte cover on trunks, branches and fallen logs.
Download figure to PowerPoint
In desiccation-tolerant mosses that have been air dry for a few days, respiration recommences almost instantaneously on remoistening and net photosynthetic carbon fixation typically reaches two-thirds or more of normal levels within the first few minutes. The initial respiratory carbon loss is typically made good within 30 min or so (Tuba et al., 1996; Proctor & Pence, 2002) Chlorophyll-fluorescence measurements show that initial recovery of the photosystems is extraordinarily rapid, and independent of protein synthesis (Csintalan et al., 1999; Proctor & Smirnoff, 2000). Complete return to unstressed levels of fluorescence parameters may take a number of hours, but is unaffected by protein-synthesis inhibitors in the dark. There are indications that some cytoplasmic protein synthesis following remoistening is needed for return to predesiccation levels of CO2 fixation (Proctor & Smirnoff, 2000), but this may be ‘maintenance’ repair of ongoing photo-damage during recovery in light, rather than the repair of damage from desiccation. Over all, the results suggest that recovery of respiration and photosynthesis in these plants is largely a matter of reassembly and reactivation of components which have survived drying and re-wetting essentially intact. A substantial ‘repair’ element in the recovery of DT bryophytes was suggested by Bewley (1979) and Bewley & Krochko (1982) primarily in relation to restoration of membrane integrity. Oliver & Bewley (1984, 1997) postulated that desiccation tolerance in ‘fully DT plants’ (primarily bryophytes) is essentially ‘repair based’, while that in ‘modified DT plants’ (vascular plants) is based on inducible ‘protection’ systems set in place in the course of slow drying. This division broadly reflects the phylogenetic distribution of DT plants (Oliver et al., 2000a), but its value depends heavily on how ‘repair’ is defined – and it underplays not only the wide range of behaviour within both bryophytes and DT vascular plants and the speedier recovery of many bryophytes, but also the features in common between the two groups. Leakage of membranes for a short time after re-wetting occurs in all desiccation-tolerant organisms (Crowe et al., 1992), and ‘repair’ processes in the limited sense originally envisaged by Bewley must be common to all DT plants. Indeed the greatest commitment to repair is seen in the poikilochlorophyllous vascular plants where much of the photosynthetic system must be rebuilt each time the plant dries out. Much further research is needed on the distribution and mechanisms of inducible desiccation tolerance in both bryophytes and vascular plants. Oliver and colleagues have reported the synthesis of many rehydration-specific proteins in Tortula ruralis (Oliver, 1991; Oliver et al., 1998; Wood et al., 1999; Oliver et al., 2000b), and this too is an area in which we may expect interesting developments in the future.