Desiccation Damage

As water is removed from cells, the physical and physiological properties of the cells change. These changes, often characterized by a reduced cell size or lack of integrated metabolism, do not in themselves imply damage. They may be purely consequences of water removal and may be completely reversible once water is added back to the system. Therefore, damage from desiccation is not indicated by differences between the hydrated and dry state, but rather by the resumption of normal activity upon rehydration.

The number of different stresses that can be associated with removal of water from cells can be attributed to the multiple roles that water plays in supporting life. Water plays a structural role: at the cellular scale, water, fills spaces and provides turgor, while, at the molecular scale, water provides hydrophilic and hydrophobic associations and controls intermolecular distances that determine the conformation of proteins, polar lipids and the partitioning of molecules within organelles. With water present, reactive surfaces of metals or molecules are not as exposed, and this limits reactivity among molecules. Water also plays a role in controlling metabolism, as it is a reactant and product of many reactions. As a dilutant, water affects the chemical potential of other molecules, potentially shifting the likelihood of reactions. Water also provides the fluid matrix that allows diffusion of substances to reactive sites. Changes in water concentration affect viscosity of the matrix and the overall mobility of dissolved or suspended molecules. The drier the medium becomes, the more viscous it becomes, until it is essentially a solid matrix, trapping molecules (Slade and Levine, 1991; Williams et al., 1993; Leopold et al., 1994; Buitink et al., 1998b; Wolfe and Bryant, 1999). As one would expect from all the roles of water, there will be a number of strains that the tissues undergo when water is removed.

9.3.1. Mechanical strains and structural damage

9.3.1.1. Cellular and subcellular scales

The first sign of desiccation/drought stress is the loss of turgor pressure. This occurs at water potentials of about -1 to -2 MPa, coinciding with the water potential range designated as 'permanent wilting point' for non-transpiring vegetative tissue (Levitt, 1980). At lower water potentials, cells lose water and shrink (Meryman, 1974, Steponkus, 1979; Levitt, 1980; Steponkus and Lynch, 1989; Steponkus et al., 1995). Osmotic adjustments, which lessen the water potential difference between cells and the environment and augment the amount of dry matter in cells, can prevent water loss and cell contraction at water potentials between -1 and -2.5 MPa (Levitt, 1980; Jones and Gorham, 1983). Osmotic adjustments are fairly ineffective at reducing strains when cells are exposed to lower water potentials (Wolfe and Bryant, 1999). In slow-freezing experiments, believed to mimic dehydration stress, protoplasts can undergo reversible contraction-expansion cycles, or 'osmotic excursions', when slowly cooled and warmed from > 0°C (^w = -0.5 MPa) to temperatures of -2 to -5°C (-2.5 ^ ^ -6 MPa) (Meryman, 1974; Steponkus, 1979; Steponkus and Lynch, 1989). A 60-80% reduction in cell volume occurs when the water potential of cells decreases from about -0.5 MPa to about -4.5 to -6 MPa (-4 to -5°C) (Meryman, 1974; Steponkus, 1979; Steponkus and Lynch, 1989). Similar contraction was calculated for immature embryo cells in which 88% of the cell volume was occupied by water (Fig. 9.1). However, cells filled with dry matter reserves (mature embryos in Fig. 9.1) do not contract as much as highly vac-uolated cells (immature embryos in Fig. 9.1). For a similar reduction in water potential to -5 MPa, the cells of fully mature bean axes contract only by about 18%, and complete desiccation only causes a 24% reduction in volume in these cells (Fig. 9.1). When cells that have not been acclimatized to the water stress shrink by 50-80%, they burst when returned to the original water potential. This observation led to the concept of 'minimum critical volume' (Meryman, 1974), which describes the limits to which a cell can contract in a reversible osmotic excursion. As seen for mature embryos (Fig. 9.1), this strain of cell contraction can be avoided by accumulating dry matter.

Differences in the degree to which cell walls contract compared with protoplasm may cause mechanical stress and damage to the plasmalemma or plant cells during dehydration. The tight attachment of the plasmalemma to the cell wall is believed to create tension to the membrane in shrinking cells (e.g. Murai and Yoshida, 1998b), which is most profound at the cell wall-plasmalemmma attachments near the plasmodesmata (Iljin, 1957; Bewley and Krochko, 1982). Plasmolysis, where the plasma membrane separates from the cell wall, appears to mitigate damage to whole cells during severe water stress (Murai and Yoshida, 1998b), and there is some evidence to suggest that cells in desiccation-tolerant seeds are slightly plasmolysed (Perner, 1965). Observations of plasmolysis may be an artefact of the aqueous fixatives used to study dry organisms (Opik, 1985; Platt et al., 1997; Wesley-Smith, 2001). In studies using anhydrous chemical fixation (Opik, 1985) or freeze substitution (Wesley-Smith, 2001, Wesley-Smith et al., 2001), the plasma membrane remained closely appressed to the cell walls, and both the cell wall and the plasmalemma became highly convoluted during desiccation of tolerant cells. Opik (1985) demonstrated that the plasmalemma separated from the cell wall during rehydration as a result of differential swelling or weakening of the cell wall-plasmalemma association caused by detergents such as dimethylsulphoxide. The mechanical properties of the cell wall, including its elasticity, ability to fold and associations with plasmodesmata, influence the degree of plasma membrane disruption consequent upon contraction or expansion (Webb and Arnott, 1982; Opik, 1985; Murai and Yoshida, 1998b; Vicre et al., 1999).

Cell membranes must fold or vesiculate to accommodate the volume changes during cell contractions. Conservation of membrane surface area during contraction is critical for successful rehydration. If the surface area of the plasmalemma is reduced too much, the cell bursts upon rehydration, suggesting that there is a critical minimum surface area, rather than a critical minimum volume, to which cells can survive (Steponkus, 1979; Steponkus and Lynch, 1989; Steponkus et al., 1995). Protoplasts from cells that are not acclimatized to the cold contract through invaginations of the plasma membrane, which eventually form endocytotic vesicles that cannot be reincorporated into the plasmalemma upon warming (Steponkus and Lynch, 1989; Steponkus et al., 1995). The plasma membrane of protoplasts from cells more tolerant of water stress (i.e. acclimatized by low temperatures) contracts through exocytotic extrusions which remain continuous with the plasma membrane and help to conserve the membrane surface area (Steponkus and Lynch, 1989; Steponkus et al., 1995). High phospho-lipid:sterol ratios and high amounts of diunsaturated fatty acids in the plasmalemma appear to facilitate exocytotic folding in shrinking protoplasts and greater elasticity of the expanding membranes (Steponkus and Lynch, 1989; Steponkus et al., 1995). Protoplasts with these properties tend to survive to lower water potentials (Steponkus et al., 1995).

The mechanism by which membrane surface area is conserved in intact cells is largely unknown. There are some studies of the effect of dehydration on cell volume and membrane configuration in cells from plant embryos, but these are often confounded by problems associated with using aqueous fixatives (Platt et al., 1997; Wesley-Smith et al, 2001). In addition, the studies often use mature embryos (recalcitrant or orthodox) where > 50% of the cell volume is occupied by dry matter (e.g. Farrant et al., 1997). These cells will not experience the same degree of shrinkage as highly vacuolated cells (Fig. 9.1), and so the need for conserving membrane surface area is not as critical. Circumventing the problem of cell shrinkage may explain why most orthodox and recalcitrant embryos (except for A. marina and other recalcitrant seeds with highly vacuolated cells (Farrant et al., 1992, 1993)) are fairly tolerant of water stress, surviving to water potentials of —12 MPa or less, compared with the benchmark of —5 MPa described above for protoplasts from non-acclimatized cells. None the less, drying results in some degree of cell contraction, which is mostly completed when the water potential of the cell is reduced to —12 MPa (Fig. 9.1). In cells that survive water potentials of about —12 MPa but not lower, both endo- and exocytotic vesicles have been observed (P. Berjak and N.W. Pammenter, unpublished date). These observations are not reported in extremely dried embryos, perhaps because of technical problems of fixation. In severely dried cells of fully desiccation-tolerant seeds, the plasmalemma stays intact and closely attached to the cell wall as this folds, suggesting that the membrane surface area remains relatively constant during drying even though the cell volume is diminished (Opik, 1985). Some membrane constituents may be removed during cell contraction as evidenced by whorls of membrane close to the plasmalemma in seed cells (Webster and Leopold, 1977; Opik, 1985; Wesley-Smith et al., 2001) and circular membrane structures and plas-toglobuli within chloroplasts in sections of leaf tissue from desiccation-tolerant angiosperms (Farrant et al., 1999; Farrant 2000; Mundree et al., 2000). These membrane bodies have been proposed to provide additional membrane reserves upon rehydration (Webster and Leopold, 1977; Farrant et al., 1999; Mundree et al., 2000), although mechanisms by which they would be reinserted are not clear and their very presence may be artefacts of aqueous fixation. Alternatively, these membrane abnormalities may arise from other organelles, such as endoplasmic reticula, and may participate in autophagy or vacuole formation (Wesley-Smith et al., 2001). The shapes of nuclei, mitochondria and plastids in dried cells of desiccation-tolerant seeds are irregular and convoluted, suggesting that the surface area of the membranes of these organelles are also conserved simply by folding (Opik, 1985).

The membranes of cell vacuoles are likely to experience tensions similar to those described for protoplast membranes during osmotic excursions, and so are prone to rupture, with lethal consequences, following exposure to water potentials of -2.5 to -5 MPa (Murai and Yoshida, 1998a). Highly vacuolated cells of immature seeds (Berjak et al., 1984, 1994; Farrant et al., 1989, 1997) and desiccation-sensitive vegetative tissue (Farrant and Sherwin, 1997; Farrant, 2000) are particularly sensitive to tonoplast dissolution. Replacing the water in vacuoles with solid material reduces the degree to which vacuoles must contract, thereby lessening the tension on tonoplast membranes during drying. Dry matter reserves naturally accumulate during embryogenesis in orthodox and some recalcitrant seeds, and may explain the progressive tolerance to low water contents in developing embryos (Vertucci and Farrant, 1995; Farrant et al., 1997; Farrant and Walters, 1998). There is also accumulation of dry matter in vacuoles of vegetative tissues in many of the desiccation-tolerant angiosperm species during acclimatization to water stress (Farrant, 2000).

Water loss results in a general contraction of cell volume. The plasmalemmae of plant cells can be damaged if they are sheared from cell walls, which contract less than protoplasm, or if contraction results in an irreversible loss of membrane surface area. In addition to protection by filling cells with dry matter (described above), the consequences of volume changes can also be lessened by initial high surface area-to-volume ratios of cells and vacuoles (Iljin, 1957; Bewley, 1979) and may explain why cells from non-vascular plants, which usually have small vacuoles and lack plasmo-desmata, do not appear to suffer physical damage upon contraction (reviewed by Bewley and Krochko, 1982). Damaging effects of cell contraction are usually manifested during rehydration, suggesting that the stress and damage are not direct effects of desiccation, but rather indications of rehydration stress and mechanical failure.

A dismantling of mitochondria and chloroplasts is associated with severe water stress. Mitochondria observed in mature orthodox seeds lack defined cristae (Bergtrom et al., 1982; Thomson and Platt-Aloia, 1984; Farrant et al., 1997) and mitochondrial proteins are easily extractable from dried pollen (Hoekstra and van Roekel, 1983). Conversely, mitochondria from immature embryos and recalcitrant seeds are more defined, and the greater differentiation has been linked to greater sensitivity to desiccation (Farrant et al., 1997). Chloroplast structure also degrades during water stress. Dried leaves of the desiccation-tolerant grasses Borya nitida and Xerophyta humilis become yellow, concurrent with the loss of grana stacks in the chloroplasts (Gaff and Hallam, 1974; Farrant, 2000). Using fluorescence-induction kinetics to study partial processes of photosynthesis, researchers found a decrease in the efficiency of photosystem II at water potentials between -3 and -4 MPa (Wiltens et al., 1978; Hetherington, et al., 1982b; Vertucci et al., 1985; Sherwin and Farrant, 1998; Tuba et al., 1998; Csintalan et al., 1999) or during acclimatization to winter (Öquist and Strand, 1986). This decline could be a consequence of photochemical damage, but is more likely to be a reflection of protective dismantling of photosystem II (Demmig-Adams and Adams, 1992; Farrant, 2000). Indeed, the dismantling of the photosynthetic apparatus during drying of B. nitida and X. humilis is required for survival: plants dried too rapidly stay green and do not recover (Gaff and Hallam, 1974; Farrant et al., 1999).

Slight water stress (-1 > >-3 MPa) enhances the protein synthesis that is believed to be important for conferring tolerance (Ried and Walker-Simmons, 1993; Vertucci and Farrant, 1995; Ingram and Bartels, 1996; Oliver et al., 1998; Mundree et al., 2000; Whittaker et al., 2001). Further drying reduces the rate of protein synthesis in both tolerant and sensitive cells (Bewley and Krochko, 1982; Salmen Espindola et al., 1994; Ingram and Bartels, 1996; Oliver et al., 1998; Mundree et al., 2000; Whittaker et al., 2001), perhaps because of a dismantling of endoplasmic reticulum, dictyosomes and polysomes (Webster and

Leopold, 1977; Thomson and Platt-Aloia, 1984; Farrant et al., 1997; Wesley-Smith et al., 2001).

Indirect evidence from recalcitrant seeds suggests that, during dehydration, the cytoskeleton is disrupted at fairly high water potentials (-3.8 MPa for Trichilia dregeana and -3.5 MPa for Quercus robur) leading to an abnormal distribution of organelles within cells (Berjak et al., 1999; Mycock et al., 2000). Although it is tacitly assumed that cytoskeletal disassembly must occur during dehydration in desiccation-tolerant seeds (and vegetative tissues), it is its failure to reconstitute that characterizes this aspect of dehydration-related injury in recalcitrant material (Mycock et al., 2000).

There is clearly a general trend towards contraction or disassembly of cellular machinery during water stress to about -5 MPa. In most desiccation experiments, plant materials are stressed further and cell survival is assayed by whether or not organelles reassemble upon rehydration. In desiccation-sensitive cells that do not rupture, the protein-synthesizing machinery does not recover, nor do mitochondria and chloroplasts resume normal function; organelles become irregularly shaped and disorganized (reviewed by Bewley and Krochko, 1982; Farrant et al., 1989; Berjak et al., 1990; Mycock et al., 2000). The contraction and dismantling of organelles described above are clearly signs of water stress, but it is unclear whether these changes are symptoms of damage occurring at - 5 MPa, or means of protection when water stress intensifies. It is also unclear whether the failure to reconstitute organelles indicates a primary site of damage or a general debilitation when cells die. These cause and effect arguments have led researchers to study the primary effects of dehydration on the structure of macromol-ecules.

9.3.1.2. Molecular scale

Removing water from cells pushes cellular constituents together, causing them to interact in ways that might not otherwise occur. A consequence of these molecular aggregations is an increased ordering of molecular structures, and it may seem ironic that primary lesions during drying are directly attributed to order rather than to loss of it. Drying-induced compaction of molecules requires greater packing efficiency, resulting in localized enrichments of similar-type molecules in a process known as demixing (Lis et al., 1982; Bryant and Wolfe, 1989; Rand and Parsegian, 1989; Bryant et al., 1992). Molecules remix upon rehydration, but the reactions that occurred in the desiccated state may have irreversible consequences.

Intermolecular associations of polar lipids are intrinsically linked to the water content of the medium. Under aqueous conditions, polar lipids spontaneously align to form micelles or bilayer structures depending on the polar head group of the lipid. Acyl chains within bilayers are more-or-less mobile, giving considerable fluidity to the structure and allowing proteins and other constituents to be inserted. Drying brings membrane bilayers into close proximity and causes membrane constituents to segregate laterally into different domains enriched with particular lipid classes or proteins (Lis et al., 1982; Bryant and Wolfe, 1989; Rand and Parsegian, 1989; Bryant et al., 1992; Crowe and Crowe, 1992; Steponkus et al., 1995; Hoekstra and Golovina, 1999) (Fig. 9.2). The closer packing between membranes and among membrane constituents results in greater rigidity of the fatty acid domain within the bilayer. There are two mechanisms, based on either intra- or interlamel-lar events, used to explain why fatty acid domains become more rigid. If water molecules are removed from between adjacent polar head groups, the associated fatty acids compress because of the increased strength of van der Waals attractions (Crowe et al., 1990; Crowe and Crowe, 1992; Hoekstra and Golovina, 1999). Alternatively, as different bilayers come into close apposition, strong repulsive hydration forces keep them separate, but create isotropic tensions that lead to lateral compression within the acyl domain (Lis et al., 1982; Wolfe, 1987;

Rand and Parsegian, 1989; Bryant and Wolfe, 1992; Wolfe and Bryant, 1999). Increased rigidity eventually leads to phase transitions within the membrane from a fluid to a gel state (Ladbrooke and Chapman, 1969; Cullis and de Kruijff, 1979). While these phase transitions are completely reversible, they are believed to interfere with the semi-permeable properties of membranes. Permanent damage comes from the exclusion of proteins from parts of the bilayer (Rand and Parsegian, 1989; Bryant and Wolfe, 1992; Crowe and Crowe, 1992; Hoekstra and Golovina, 1999) (Fig. 9.2). Transient damage occurs upon rehydration: the rush of water on to an inelastic membrane may cause it to rupture (Murphy and Noland, 1982; Steponkus et al., 1995; Hoekstra et al., 1999) or imperfect packing among different domains may cause leakage of cellular constituents (Crowe and Crowe, 1992; Hoekstra et al., 1999).

The close approach of membrane systems and the lateral demixing of membrane components can lead to an even greater threat to membranes than lamellar fluid-to-gel transitions. Membranes can fuse together, causing the complete loss of compartmentation within the cell (Crowe et al., 1986; Crowe and Crowe, 1992; Steponkus et al., 1995) (Fig. 9.2). Fusion is known to occur among liposomes and native membrane fractions, although the mechanism that causes polar lipids to cross over to a different bilayer is unclear. In principle, the hydration characteristics of individual lipids and lipids in a mixture, the intrinsic curvature of different head groups, the water content and the temperature allow the formation of inverted micelles within closely appressed bilayers (Cullis and de Kruijff, 1979; Crowe et al., 1986; Steponkus et al., 1995) (Fig. 9.2). In domains enriched with non-bilayer-forming lipids such as phosphatidylethanolamine-diglycerides or monogalactosyl-diglycerides, the polar head groups coalesce into rings and the acyl chains extend radially outwards in what is known as a hexagonal phase (Cullis and de Kruijff, 1979; Siegel et al., 1994; Steponkus et al., 1995). Fusion via hexagonal-phase changes is rare in native mem branes, but has been demonstrated in cells from non-acclimatized leaves that were lethally cooled to -5°C (oat, = -6 MPa) or -10°C (rye, = -12 MPa) (Steponkus et al., 1995) and more frequently in animal cells (Cullis and de Kruijff, 1979; Crowe and Crowe, 1992). Evidence of cell fusion, but not via hexagonal-phase changes, is common in desiccation-damaged cells, protoplasts and liposomes (e.g. Crowe et al., 1986; Steponkus et al., 1995). In oat and rye leaves acclimatized to cold (but clearly not fully desiccation-tolerant), fusion of plasmalemma and endomembrane systems is suggested at temperatures between -10 and -40°C (-12 > > -48 MPa), depending on the level of cold tolerance achieved (Steponkus et al., 1995). Upon rehydration, improperly fused membranes produce vesicles that exclude cell constituents or are combinations of different membrane systems (e.g. the plasmalemma fuses with chloroplast outer membrane or with endoplasmic reticulum) (Fig. 9.2). Because the osmotic balance inside and outside the cells has been completely disrupted, vesicles produced from membrane fusions are identified by their inability to expand during rehydration (Steponkus et al., 1995).

Most of our understanding of how polar lipids behave in water-stressed situations comes from model studies of liposomes with known composition. In these systems, phase transitions are usually studied, even though they may only be harbingers of real damage. Phase transitions of prepared membrane systems occur at a range of water contents and temperatures depending on the saturation of the acyl chains and the presence of non-phospholipids (e.g. Ladbrooke and Chapman, 1969; Cullis and de Kruijff, 1979; Crowe et al., 1989; Steponkus et al., 1995). A water potential of about -12 MPa is often cited as critical. It has been suggested that structural water needed for the proper spacing of polar head groups is removed at ^ -12 MPa (Ladbrooke and Chapman, 1969; Crowe et al., 1990). Also, at = -12 MPa, large, potentially deforming hydration forces result from the close approach of molecules (Wolfe, 1987).

Hydrated membrane systems

Plasma membrane

Plasma membrane

Endomembrane

Plasma membrane

Fig. 9.2. Schematic drawing of the effect of dehydration on cellular membranes. Different membrane systems may become closely appressed, leading to demixing of lipids and proteins and the loss of proteins from parts of the bilayer. Closely appressed membranes may then form non-bilayer structures that lead to fusion between different membrane systems. Upon rehydration, cellular contents leak out and fused membrane particles do not swell (i.e. they are 'osmotically unresponsive'). (Adapted from Steponkus et al. (1993), with permission.)

Plasma membrane

Plasma membrane

Non-bilayer phase formation and membrane fusion

Endomembrane

Plasma membrane

Rehydrated membranes

Leakage of œH T^JF Endomembrane contents

Fig. 9.2. Schematic drawing of the effect of dehydration on cellular membranes. Different membrane systems may become closely appressed, leading to demixing of lipids and proteins and the loss of proteins from parts of the bilayer. Closely appressed membranes may then form non-bilayer structures that lead to fusion between different membrane systems. Upon rehydration, cellular contents leak out and fused membrane particles do not swell (i.e. they are 'osmotically unresponsive'). (Adapted from Steponkus et al. (1993), with permission.)

There is little information for comparing membrane phase behaviour among orthodox and recalcitrant embryos, maturing embryos as they become more tolerant of desiccation, or leaves from desiccation-tolerant angiosperms as they adjust to low water potentials. Changes in bilayer spac-ings or lamellar fluid-to-gel transitions have been detected in both desiccation-tolerant and sensitive plant cells during dehydration, with little difference in behaviour detected with degree of tolerance (McKersie and Stinson, 1980; Seewaldt et al., 1981; Priestley and de Kruijff, 1982; Singh et al., 1984; Kerhoas et al., 1987; Crowe et al., 1989; Hoekstra et al., 1991, 1992; Sun et al., 1994; Hoekstra and Golovina, 1999). In tolerant soybean cotyledons, a gel-like transition occurred when seeds were dried to less than 0.2 g H2O g-1 dry mass (Seewaldt et al., 1981), a water content that corresponds to a water potential of about -12 MPa (e.g. Vertucci and Roos, 1990) (Fig. 9.1). Water potentials between -10 and -15 MPa also mark the survival limit for recalcitrant seeds (described above). A membrane-mediated mechanism is often invoked to explain damage in desiccation-sensitive embryos and pollen because the membrane integrity of these cells appears to be compromised upon rehydration (McKersie and Stinson, 1980; Senaratna and McKersie, 1983; Vertucci and Leopold, 1987; Berjak et al., 1992, 1993; Poulsen and Eriksen, 1992; Sun and Leopold, 1993; Sun et al., 1994; Wolkers et al., 1998a).

The different views of dehydration stress (i.e. removal of structural water versus enhancement of hydration forces) have promoted different ideas for the mechanisms of protection. According to the 'Water Replacement Hypothesis', if structural water is removed, small hydrophilic molecules such as sugars must be inserted between polar lipid head groups to maintain proper intermolecular spacings and membrane integrity (Clegg, 1986; Crowe et al., 1990; Crowe and Crowe, 1992). An alternative, but not mutually exclusive, model suggests that high concentrations of compatible solutes can help resist water loss between molecular surfaces, relieving the size of hydration forces (Wolfe and Bryant, 1999; Koster et al., 2000; Bryant et al., 2001). As dehydration proceeds, the concentration within the interfaces increases, with a concomitant increase in viscosity (Fig. 9.2). The high viscosity of these interfacial solutions provides mechanical resistance to the further compression of macromolecules (Wolfe and Bryant, 1999; Koster et al., 2000; Bryant et al., 2001). In both protective models, the goal is to keep molecules separated so that harmful interactions are prevented. Sugars accomplish this capably in model membrane systems (Crowe et al., 1986, 1989, 1990; Crowe and Crowe, 1992; Wolfe and Bryant, 1999; Koster et al., 2000; Bryant et al., 2001). However, the presence of adequate quantities of sugars in cells and the vitrification of cellular constituents do not appear to prevent polar lipid phase changes in desiccation-tolerant cells (Seewaldt et al., 1981; Priestley and de Kruijff, 1982; Crowe et al., 1989; Hoekstra et al., 1989, 1992, 1999; Leopold et al., 1994) or damage in desiccation-sensitive cells (Berjak et al., 1992, 1993; Sun and Leopold, 1993; Still et al., 1994; Sun et al., 1994; Vertucci and Farrant, 1995; Vertucci et al., 1995; Farrant and Walters, 1998; Wolkers et al., 1998a; Hoekstra and Golovina, 1999; see Chapter 10). Changing the composition of membranes (reviewed by Steponkus et al., 1995) and reducing their surface area by dismantling endomembrane systems (described above) may be the important tools for maintaining compartmentation in drying cells.

Structural changes of proteins with hydration have received wide attention in the literature. Early work using a variety of proteins showed that protein structure was conserved during drying to extremely low levels (Schneider and Schneider, 1972; Kuntz and Kauzmann, 1974; Ruegg and Hani, 1975; Ruegg et al., 1975; Fujita and Noda, 1978; Careri et al., 1980; Takahashi et al., 1980; Jaenicke, 1981; Rupley et al., 1983). In parallel studies, it was demonstrated that some proteins even maintained functional activity (albeit at low levels)

when dry (Acker, 1969; Potthast, 1978; Labuza, 1980; Rupley et al., 1983). Secondary structure of cytoplasmic proteins (extracted from desiccation-tolerant pollen) was conserved upon drying in the absence of protectant sugars, demonstrating innate stability perhaps because of the high degree of a-helical structures (Wolkers and Hoekstra, 1995). The reversibility of sorption-desorption isotherms of numerous proteins supported the idea that con-formational changes of proteins during hydration were slight and reversible, making proteins an ideal model for studying hydration properties of biological materials (Bull, 1944; D'Arcy and Watt, 1970). Slight, reversible changes in protein structure, particularly secondary structure, have been attributed to volumetric changes from the loss of water rather than to changes in the native structure of proteins. These changes occur at fairly low moisture levels (between 0.2 and 0.1 g H2O g-1 dry mass or -70 to -200 MPa) (Ruegg and Hani, 1975; Griebenow and Klibanov, 1995). Drying, in fact, stabilizes protein structures, making them particularly resistant to ageing (Franks et al, 1991; Costantino et al., 1998) and heat denaturation (e.g. Echigo et al., 1966; Ruegg et al., 1975; Fujita and Noda, 1978; Takahashi et al., 1980; Jaenicke, 1981; Leopold and Vertucci, 1986; Wolkers and Hoekstra, 1997). The extreme stability of protein structure with low hydration may be attributed to stronger intramolecular associations compared with the situation of polar lipids. Such interactions would reduce the need for hydrogen bonding with water to maintain structural integrity (obviating the need for water replacement by sugars as suggested by Crowe and co-workers (e.g. Crowe and Crowe, 1992)) and/or provide mechanical strength that resists deformation when molecules are compressed (obviating the need for mechanical barriers to compression as suggested by Wolfe and co-workers (e.g. Wolfe and Bryant, 1999)).

The conformations of some proteins and polypeptides are irreversibly damaged by drying or freeze-drying in the absence of protectants (Hanafusa, 1969; Carpenter et al., 1987, 1990; Franks et al., 1991; Prestrelski et al., 1993). Enzymes such as lactate dehydrogenase and polypeptides such as poly-L-lysine are particularly labile (Prestrelski et al., 1993), and damage is exacerbated if molecules are freeze-dried rather than air-dried (Franks et al., 1991). Rate of drying also has a large effect on the conservation of protein structure, with greater preservation achieved by rapid drying conditions (Wolkers et al., 1998a,b). Often, desiccation-labile proteins are used to study the effects of protectants (Carpenter et al., 1987, 1990; Prestrelski et al., 1993). Clearly, these studies are essential to the pharmaceutical industry, but similar mechanisms of protection must not be presumed to apply in vivo in dehydrating plants. A tremendous amount of work has demonstrated that proteins are rather robust; thus, a need for protection must be demonstrated before a protective mechanism is implied. Studies must show that desiccation-labile enzymes exist in vivo, that they are not produced de novo during rehydration and that they are irreversibly damaged in desiccation-sensitive cells.

The structure and activity of proteins are compromised if they are stored under extremely dry conditions of approximately 0.1 g H2O g-1 dry matter or about -200 MPa or less (Kuntz and Kauzmann, 1974; Luscher-Mattli and Ruegg, 1982; Sanches et al., 1986; Labrude et al., 1987). Substantial deterioration of the lattice of protein crystals was attributed to the refolding of polypep-tide chains to increase packing efficiency (Kuntz and Kauzmann, 1974; Luscher-Mattli and Ruegg, 1982). Other studies have shown that severe drying exposes haem groups on proteins, promoting free radical production (Sanches et al., 1986; Labrude et al., 1987). At such low water contents, proton exchanges among charged amino acids could be measured, suggesting that these sites were exposed (Careri et al., 1980; Rupley et al., 1983). Deterioration at similar water potentials and in similar time frames is observed in stored seeds and pollen (e.g. Vertucci and Leopold, 1987; Vertucci and Roos, 1990; Buitink et al., 1996). Although these organisms survive the initial stress of complete water removal, they age progressively more rapidly when stored at

^ -220 MPa (< 20% RH). Perhaps mechanisms suggested to cause damage in proteins at low water contents (e.g. exposure of reactive sites on the proteins, increased relaxation of molecular structures as they fill voids left by water, or relaxation of the glassy matrix that embeds the proteins) are responsible for the deterioration of stored seeds and pollen. Protein structure is stable in seeds stored at about 30% RH (Golovina et al. 1997), but stability of protein structures in seeds stored at lower humidities has not been documented. Increased ageing rates of seeds and pollen stored below a critical water content have also been attributed to reduced viscosity of the aqueous medium in cells that are almost completely dry (Buitink et al., 1998b).

Upon dehydration, the same destabilizing forces that perturb lipid and same protein structures may also affect nucleic acid structure (Rau et al., 1984). DNA is a particularly stable molecule (Wayne et al., 1999) which maintains its structure in the absence of water and reversibly unfolds at high temperatures (Bonner and Klibanov, 2000). The intermolecular distances of dehydrating DNA strands are comparable to those of condensed DNA in hydrated nuclei (Rau et al., 1984), suggesting that DNA structures are resistant to perturbations resulting from dense packing. When DNA is replicated and so is decondensed during germination, the cells concomitantly become susceptible to desiccation injury (Deltour and Jacqmard, 1974; Crevecoeur et al., 1988) and rapidly dividing cells during embryogenesis also appear to be sensitive to desiccation (Myers et al., 1992). Desiccation did not affect the structure of condensed or decondensed chromatin in desiccation-tolerant or sensitive maize embryos, respectively (Leprince et al., 1995a). However, in those studies, the chelation of Ca2+ (and other divalent cations) by the ethylenediamine tetra-acetic acid (EDTA) present in the medium used for chromatin spreading, may have relaxed previously condensed chromatin, possibly accounting for the reportedly similar results from desiccation-tolerant and sensitive material (Pammenter and Berjak, 1999) (see also Chapter 12).

When unprotected cells are dried, organelles and macromolecules experience mechanical or structural damage. This type of desiccation damage is termed sensu stricto because the primary stress is water removal (Pammenter and Berjak, 1999; Walters et al., 2001). Membrane structures appear more prone to desiccation damage sensu stricto than do proteins or DNA, perhaps because of the intense hydrogen bonding within proteins and nucleic acid structures. Protection from damage often lies in the ability of the structure or the surrounding medium to offer mechanical resistance to the stress or accommodate the stress through enhanced elasticity.

9.3.2. Metabolically derived damage

Loss of turgor precipitates a number of changes in metabolic pathways of plant cells. Assimilation of CO2 (if the tissue is photosynthetic) and growth are impaired. Often protein synthesis is temporarily stimulated during mild water stress (reviewed by Farrant et al., 1989; Ingram and Bartels, 1996; Oliver et al., 1998), with a switch in metabolism believed to lead to the production of putative protection mechanisms (reviewed by Vertucci and Farrant, 1995; Ingram and Bartels, 1996; Oliver et al, 1998; Chapters 1, 5 and 11). Observations of increased polysomes and rough endoplasmic reticulum in slightly water-stressed recalcitrant embryos suggest that certain (possibly similar) metabolic pathways may also be induced in seeds that do not acquire full tolerance of desiccation (Berjak et al., 1984; Farrant et al., 1989; Pammenter et al., 1998). These changes in metabolism do not indicate that cells have already experienced damage; when briefly stressed, most organisms resume normal metabolism once the water stress is relieved. However, prolonged mild stress (which could be considered akin to drought) is deleterious to both vegetative and embryonic tissues. Many recalcitrant seeds lose viability if maintained for long periods at constant high water contents (e.g. Chin and Roberts, 1980; Berjak et al., 1989; Pammenter et al., 1994; Walters et al., 2001), and similar damage is observed in orthodox seeds (Walters et al., 2001). The loss of viability has been associated with the continuation of metabolism (including cell division) (Farrant et al., 1989), which will ultimately lead to a greater demand for water to maintain high water potentials (Berjak et al., 1989; Pammenter et al., 1994).

Metabolism slows at water potentials less than about -2 MPa, but not all reactions are affected by dehydration in the same way. Protein synthesis slows down at relatively high water potentials (reviewed by Bewley and Krochko, 1982; Clegg, 1986; Salmen Espindola et al., 1994; Ingram and Bartels, 1996; Mundree et al., 2000; Whittaker et al., 2001), while respiration continues to much lower levels (Vertucci and Leopold, 1984; Vertucci and Roos, 1990; Salmen Espindola et al., 1994; Leprince and Hoekstra, 1998; Leprince et al., 1999; Farrant, 2000; Walters et al., 2001). Various reactions within photosyn-thetic (Wiltens et al., 1978; Hetherington et al., 1982b; Vertucci et al., 1985; Vertucci and Leopold, 1986; Farrant, 2000) and respiratory (Vertucci and Leopold, 1986; Leprince and Hoekstra, 1998; Leprince et al., 2000) pathways respond differently to low water contents. The differing responses to water stress among and within metabolic pathways can lead to imbalances in metabolism. Metabolic imbalances may be confounded by the respiration of fungi that occurs at water potentials as low as -20 MPa in orthodox and recalcitrant seed tissues (Mycock and Berjak, 1990; Goodman, 1994; Calistru et al., 2000). Damage by metabolic stress is most pronounced in cells at water potentials between -2 and — 5 MPa with a diminishing effect as cells are dried to —12 MPa (Leprince et al., 2000; Walters et al., 2001). Both desiccation-sensitive and -tolerant organisms are damaged when stored at intermediate water potentials, though the time-dependency of the damage varies considerably among species and tissues (Walters et al., 2001).

A by-product of continued respiration and light harvesting when other metabolic processes are shut off is the accumulation of high-energy intermediates that leak out of mitochondria and plastids and form reactive oxygen species (ROS) and free radicals (Puntarulo et al., 1991; Dean et al., 1993; Hendry, 1993; Leprince et al., 1993, 1994, 1995b; Smirnoff, 1993; Foyer et al., 1994; Halliwell and Gutteridge, 1999). Reactive oxygen species and free radicals react with proteins, lipids and nucleic acids, causing permanent damage to enzymes (Wolff et al., 1986; Dean et al., 1993; Halliwell and Gutteridge, 1999), membranes (Senaratna and McKersie, 1983, 1986; Chan, 1987; McKersie et al., 1988, 1989; Finch-Savage et al., 1996; Halliwell and Gutteridge, 1999; Leprince et al., 2000) and chromosomes (Dizdaroglu, 1994). Peroxidation of lipids decreases the fluidity within membranes (McKersie et al., 1988, 1989), interfering with their selective permeability upon rehydration (as described above). Upon dehydration, high levels of free radicals have been detected in desiccation-sensitive embryos (Senaratna and McKersie, 1983, 1986; McKersie et al., 1988; Hendry et al., 1992; Leprince et al., 1993, 1994, 1995b, 1999, 2000; reviewed by Vertucci and Farrant, 1995; Pammenter and Berjak, 1999). The origin and sequence of events following the appearance of these toxic compounds are still unclear. They may be produced by the water-stressed cell (Leprince et al., 1994, 1995b, 1999, 2000; Leprince and Hoekstra, 1998) or as a result of the associated fungi (Goodman, 1994; Finch-Savage, 1999), and they may precede (or precipitate) damage (Finch-Savage et al., 1996; Leprince et al., 2000) or arise after the cell has already died (Finch-Savage, 1999).

There are several ways that cells can protect themselves from metabolic imbalance and ROS-mediated damage. At higher moisture levels, free-radical-scavenging enzymes efficiently detoxify ROS (Bewley, 1979; Dhindsa, 1987; Hendry, 1993; Smirnoff, 1993; Foyer et al., 1994; Kranner and Grill, 1997; Sherwin and Farrant, 1998; Pammenter and Berjak, 1999; Farrant,

2000). These enzymes appear ineffective at low water contents, and tocopherol and ascorbic acid may be more effective (reviewed by McKersie et al., 1988; Pammenter and Berjak, 1999). Amphipathic molecules such as tocopherol can partition between aqueous and lipid domains according to the water content of the cell and polarity of the molecule (Golovina et al., 1998). A controlled shutdown of metabolism upon drying may also mitigate the consequences of unbalanced metabolism (as reviewed by Leprince et al., 1993; Vertucci and Farrant, 1995; Hand and Hardewig, 1996; Pammenter and Berjak, 1999). Cells with more organelles and greater definition of organelle structure appear to be more sensitive to desiccation (Bewley, 1979; Hetherington, 1982a; Gaff, 1989; Berjak et al., 1990; Farrant et al., 1997; Farrant and Walters, 1998; Farrant, 2000), either because there are more membrane structures to protect (described above) or because the higher metabolism leads to greater ROS production. Conditions that reduce metabolism such as low temperature (Leprince et al., 1995b) or highly complex substrates (Leprince et al., 1990) also tend to reduce sensitivity to desiccation. Desiccation-sensitive cells respire at comparatively greater rates than tolerant cells at the same water content (Leprince et al., 1999; Walters et al., 2001), which may reflect properties of the mitochondria themselves or of the cellular matrix. It has been suggested that changes in viscosity with dehydration are not as marked in desiccation-sensitive cells, and so metabolism is not as restricted (Leprince and Hoekstra, 1998). It has also been suggested that the packing of macromolecules during dehydration of desiccation-sensitive cells is not as dense (Wolkers et al., 1998a,c), and this might facilitate the diffusion of oxygen through the cell matrix.

0 0

Post a comment