Biophysical Limitation of Cell Elongation in Cereal Leaves

Abstract

Grass leaves grow from the base. Unlike those of dicotyledonous plants, cells of grass leaves expand enclosed by sheaths of older leaves, where there is little or no transpiration, and go through developmental stages in a strictly linear arrangement. The environmental or developmental factor that limits leaf cell expansion must do so through biophysical means at the cellular level: wall‐yielding, water uptake and solute supply are all candidates. This Botanical Briefing looks at the possibility that tissue hydraulic conductance limits cell expansion and leaf growth. A model is presented that relates pathways of water movement in the elongation zone of grass leaves to driving forces for water movement and to anatomical features. The bundle sheath is considered as a crucial control point. The relative importance of these pathways for the regulation of leaf growth and for the partitioning of water between expansion and transpiration is discussed.

INTRODUCTION

Plants grow in size through the expansion of individual cells. Water uptake into expanding cells is osmotically driven and requires a yielding wall. The driving force is a gradient in water potential ‘ψ’ between the cell exterior (less negative ψ) and interior. So long as water and solutes are sufficiently available, and the wall yields to the existing turgor pressure, biophysical requirements are met for cell expansion. A cell bathed in a solution with nutrients will not encounter any shortages in water and solutes. The only biophysical growth limitation that such a cell can experience is a poorly extensible or non‐yielding wall. However, the situation can be very different for a cell embedded in a tissue and separated by many cell layers from the nearest xylem or phloem elements. Now, the rate of solute or water supply to the cell can become limiting, and growth may be limited by tissue properties, rather than by properties of the enlarging cell (Fig. 1).

Fig. 1. Elongating cells embedded in a tissue can be located considerable distances and many cell layers away from the nearest water source. The diagram shows a cross‐sectional view of the area next to a major vascular bundle in the leaf elongation zone of a grass, barley. Water and solutes exiting the protoxylem (PX) or metaxylem (MX) have to pass several layers of cells before reaching peripherally located tissue (here, mesophyll, MS). Epidermal cells are located even further away (compare Figs 3 and 4). During its passage, water has to cross two bundle sheaths, the mestome sheath (MSH) and the parenchymatous bundle sheath (PBS). The walls of the mestome sheath may be suberized, as indicated by the bold line. Water can move along an apoplastic path or along a combined symplastic/transcellular path. For simplicity, each path is shown only for one direction of water flow.

Many studies have explained differences in plant growth rates through changes in cell wall mechanical properties—yield threshold and extensibility (e.g. Cramer and Bowman, 1991; for reviews, see Barlow, 1986; Cosgrove, 1993, 1998; Hsiao and Xu, 2000). However, some researchers, in particular Boyer and colleagues (Boyer, 1974; Boyer et al., 1985; Nonami et al., 1997), have pointed out that growth can be limited by tissue hydraulic properties. Yet other studies have indicated that the supply or availability of solutes to expanding cells is co‐limiting growth, particularly during water (Frensch, 1997; Hsiao et al., 1998) or salt (Fricke and Peters, 2002) stress. The vast majority of studies have been carried out on giant algal cells, hypocotyl tissue, roots or dicotyledonous leaves; few studies have focused on cereal leaves.

During the past 5 years the pressure‐probe technique (Steudle, 1993) and picolitre osmometry (Malone and Tomos, 1992) have been used to study the biophysical limitation of cell expansion in cereal leaves (Fricke et al., 1997; Fricke and Flowers, 1998; Martre et al., 1999). The results suggest that growth is (co‐)limited by the rate of water or solute supply to expanding cells. However, reaching this conclusion is far from simple. The short half‐time of water exchange of plants cells (Steudle, 1993) argues against a limitation of growth by tissue hydraulic properties. In addition, some form of destructive preparation has to take place to gain access to the leaf growth zone of grasses. The preparative method affects leaf elongation velocity and may lead to artefacts, particularly concerning cell turgor.

THE GROWING GRASS LEAF

The shoot apical meristem of grasses is located at the base of leaves. Cells are produced in two intercalary meristems. The first subdivides very early into two meristems that give rise to the leaf blade and leaf sheath, respectively. The second intercalary meristem gives rise to the internode. During the early stages of plant development, internode elongation is suppressed and only leaves expand. The primordia of subsequent leaves are therefore in close proximity and leaves emerge from the subtending sheath of older leaves, except leaf one, which emerges from the coleoptile (Jewiss, 1966; Langer, 1979).

Blade elongation is achieved through the expansion and displacement of cells in a zone that stretches to about 20–60 mm from the point of leaf insertion. Depending on the species and leaf number, elongation velocity ranges from about 1·0 to 3 mm h^–1, but can exceed 4·0 mm h^–1 (maize, Acevedo et al., 1971). The elongation zone is enclosed by sheaths of older leaves. Therefore, expanding cells are never exposed to a dry atmosphere, and water and solute supply to cells located peripherally from xylem cells are unlikely to be driven by transpiration. In that respect, and in the spatial arrangement of sequential developmental stages, grass leaves are fundamentally different from leaves of dicotyledonous plants.

THE GRASS LEAF ELONGATION ZONE

Along the basal leaf growth zone, cells expand in all directions, not just longitudinally. There is increasing evidence that lateral expansion continues beyond the point at which elongation ceases (MacAdam and Nelson, 1987; Fricke and Flowers, 1998), and both processes might be under independent control as suggested for roots (Liang et al., 1997). A perfect example is growth under reduced irradiance: leaves become longer but narrower. Compara tively little is known about the processes controlling or contributing to lateral cell expansion (MacAdam and Nelson, 2002), partly because it is easier to quantify elongation growth and partly because of the larger contribution of elongation to leaf area increase.

Growth rate can be expressed on an absolute and on a relative basis. The latter has the advantage that it is independent of the size of the element considered and that it relates future growth to existing resources. Figure 2 shows the spatial pattern of relative elemental growth rate (REGR) along the growth zone of leaf three of barley. The profile of REGR of plants grown under ‘control’ conditions is bell‐shaped and typical of grass leaves. Maximum REGRs along the growth zone of grasses typically range from about 8 to 18 % h^–1. This is small in comparison with root cells, which can elongate at up to about 50 % h^–1 (Pritchard, 1994), but compensated for by an elongation zone that is three to eight times longer in leaves than in roots.

Fig. 2. Profile of relative elemental growth rate (REGR) along the growth zone of leaf three of barley. Plants were grown under control conditions, subjected to 120 mm NaCl (Fricke and Peters, 2002) or grown under source‐reduced conditions (Fricke, 2002). The latter was achieved by removing the blades of older leaves at the time leaf three emerged from encircling sheaths. Profiles of REGR were determined by pin‐pricking and corrected for the reduction in leaf elongation velocity due to pricking.

Environmental changes that reduce the velocity of leaf elongation affect the spatial distribution of REGR in various ways. They may shorten the elongation zone, reduce maximum REGR, cause a proximal shift of high REGR at reduced or unchanged elongation zone length, or result in a general decrease in REGR (e.g. Bernstein et al., 1993; Ben‐Haj‐Salah and Tardieu, 1995; Fricke et al., 1997; Hu et al., 2000; Fricke, 2002; Fricke and Peters, 2002). Figure 2 shows two responses for the third leaf of barley. Salt‐treated plants had reduced REGR along the entire zone. In contrast, source‐reduced plants maintained high REGR selectively in the proximal part of the elongation zone; values were even shifted towards the base. Such a response enables a plant to keep the duration of elongation short, despite general reduction in leaf elongation velocity. This is because cells are displaced along the elongation zone through the cumulative elongation of cells at more proximal positions. If source‐reduced plants had maintained high REGR in the distal, rather than proximal, portion of the growth zone, elongation of leaf three would have taken weeks, not days. The question is therefore not so much why REGR was maintained in the proximal portion, but how it was maintained.

DRIVING FORCES AND PATHWAYS OF WATER MOVEMENT

The driving force for water movement can change with environmental conditions and with location in the plant. This has implications for the pathway of water movement and for the speed of water movement. The apoplastic path (the path along cell walls) has potentially the least resistance to water flow since it involves crossing of no or only a few membranes. However, apoplastic barriers, such as the endodermis in the root, substantially increase the hydraulic resistance, and the cross‐sectional area of the wall path is small compared with that of the cell‐to‐cell path (transcellular, crossing membranes and wall space, and symplastic, moving through plasmodesmata). Still, studies on roots show that the hydraulic conductance of the apoplastic path can be 100 times larger than the combined symplastic/transcellular conductance (Steudle and Peterson, 1998).

The nature of driving forces has no effect on water flows (hydraulic conductance) in individual cells (Tyerman and Steudle, 1982). However, in multi‐layered tissues, the situation is different. Steudle and co‐workers (reviewed in Steudle, 2000) observed that the radial hydraulic conductance of roots can be one to three orders of magnitude larger under transpiring than under non‐transpiring conditions. According to the composite model of water transport in the root, during transpiration water moves (hydrostatic gradient) along the low‐resistance apoplastic path, and during non‐transpiring conditions (osmotic gradients) from cell to cell along the high‐resistance symplastic and transcellular path (Steudle, 2000). Similarly, Westgate and Steudle (1985) observed for the midrib tissue of maize leaves that water flows (hydraulic conductance) in the presence of hydrostatic gradients exceeded by far flows during sorption kinetics (osmotic and matric forces). Boyer (1974) concluded that in sunflower leaves water moved during transpiration (hydrostatic gradient) along the apoplast, and during growth (osmotic gradient) along the symplastic/transcellular (‘protoplasmic’) path. The distinction between different pathways of water movement does not reflect a distinction between various sources of water at the cell level: water equilibrates readily across membranes of higher plant cells (Steudle, 1993), and protoplasts are in local ψ equilibrium with the apoplast.

Water leaving the xylem can reach peripheral growing cells along the symplastic/transcellular path when driven by osmosis; movement along the apoplast requires appropriate hydrostatic gradients. Outwardly directed gradients of positive pressure exist during periods of positive xylem pressure, for example, at night or at 100 % relative humidity, while gradients of negative pressure (tension) exist only during growth. The magnitude of the latter gradients depends on how much growth‐associated water potentials lead to apoplastic tension (Boyer, 2001) or, alternatively, to lowering of apoplastic solute potential (Cosgrove and Cleland, 1983). During periods of transpirational water loss from the emerged portion of a growing blade of a grass, xylem tension draws water along the apoplast the opposite way (in the growth zone), from peripheral, enclosed cells to xylem. Whether transpiration actually removes water along the apoplast from growing tissues and slows down expansive growth depends on the relative magnitude of growth‐induced wall tension and xylem tension (Tang and Boyer, 2002), on the conductivity of the path, and on counteracting osmotic forces by growing cells.

It seems that grass plants can regulate, independent of the overall water status, the partitioning of water between leaf growth and transpiration. For example, in barley, growth‐induced water uptake in the elongating leaf accounts for only about 1 % of the amount of water lost through transpiration from the emerged portion of the blade, and this figure is little affected by salinity or source‐reduction (Fricke, 2002; W. Fricke, unpubl. res.). In maize, about 2 % of transpiration water is used for leaf growth (Tang and Boyer, 2002).

LIMITATION OF GROWTH BY MECHANICAL AND BY HYDRAULIC CONSTRAINTS

Cell and tissue expansion can be limited by mechanical or hydraulic constraints, or both. In addition, continued solute transport is needed to compensate for growth‐associated dilution of cell osmotic pressure, π_C (MPa). For a growing tissue, this can be expressed by

R = [(ϕL)/(ϕ + L)] (ψ₀ + π_C – Y)(1)

where R denotes relative growth rate [= relative rate in volume expansion, ΔV/(VΔt); this approximates REGR along the elongation zone of grass leaves since longitudinal expansion by far exceeds lateral expansion], ψ₀ (MPa) denotes the water potential of the water source (here: xylem), ϕ (MPa^–1 s^–1) denotes the volumetric extensibility and Y (MPa), the yield threshold of the wall (Boyer et al., 1985). The term L (MPa^–1 s^–1) denotes the average hydraulic conductance of the tissue. It differs from the usually defined conductivity L_p (m MPa^–1 s^–1) in that it incorporates the volume and area geometries of the cell or tissue (Cosgrove, 1981).

The coefficient (ϕL)/(ϕ + L) determines the rate of tissue and cell enlargement at a given ψ₀. When ϕ >> L, the coefficient approximates L and expansion is limited by hydraulic properties; when L >> ϕ, the coefficient approximates ϕ and expansion is limited by mechanical properties; however, when L and ϕ are comparable in size, growth is co‐limited by hydraulic and by mechanical properties. The latter has been observed for growing stem tissue of soybeans and the elongation zone of grass leaves (Table 1). Both ϕ and L are average values for the growing tissues, and it is possible that overall limitation of organ growth is due to extreme values in only a few cells or one specialized tissue. For example, in growing dicotyledonous stems, pith or epidermis may exert mechanical growth control (for a critical discussion, see Peters and Tomos, 1996), while cells with a low diffusivity for water (i.e. long diffusion times per distance travel), next to vascular bundles, may exert hydraulic limitation (Nonami et al., 1997; Tang and Boyer, 2002). In the growing grass leaf, with a much looser mechanical association between mesophyll and epidermis, mechanical limitation of extension could be due particularly to stretching of protoxylem elements, while suberization of the bundle sheath (see below) or layers of small cells with low diffusivity for water (Tang and Boyer, 2002) may limit growth hydraulically.

Table 1.

Values of volumetric extensibility, ϕ, and hydraulic conductance, L, in growing stem and leaf tissues

Plant tissue	ϕ × 104 (MPa–1 s–1)	L × 104 (MPa–1 s–1)	Treatment	Reference
Maize, leaf	3·3	0·7	Control and short‐term exposure to NaCl	Cramer and Bowman (1991)
Begonia, leaf	0·33		Control	Serpe and Matthews (1992)
Maize, leaf	0·72–2·31		Control and osmotic stress	Hsiao et al. (1998)
Barley, leaf	1·17–1·89	0·94–1·44	Control and source reduction	Fricke (2002)
Soybean, stem	0·95	0·78	Control	Boyer et al. (1985)

The hydraulic conductance, L, also incorporates the volume and area of the tissue.

Methods included a range of techniques: cell‐pressure probe, pressure clamp, picolitre osmometry, extensibility measurement and guillotine thermocouple psychrometry.

There is evidence for both a mechanical and a hydraulic limitation of growth. Numerous studies have shown that growth rates of tissues change during stress, yet (cell) turgor remains unaffected (e.g. Termaat et al., 1985; Palmer et al., 1996; Fricke et al., 1997; Thompson et al., 1997). This has led to the conclusion—not necessarily by the authors—that growing tissues adjust to environmental changes through changes in wall properties (ϕ, Y; for discussion, see Hsiao et al., 1998). The precise molecular basis of ϕ and Y is not known, but there is increasing evidence that the wall has chemo‐rheological properties (Cosgrove, 1993) and that ϕ and Y are under enzymatic, protein‐biochemical or chemical control. Expansins, xyloglucan‐endotransglycosylase (XET), peroxidase, ascorbate and, most recently, yieldins have been associated with wall loosening and hardening and with changes in yield threshold (Thompson et al., 1997; Cosgrove, 1998; de Souza and MacAdam, 1998; Fry, 1998; Okamoto‐Nakazato et al., 2000); however, evidence is conflicting (Palmer and Davies, 1996; Schünmann et al., 1997; Peters et al., 1999).

Other studies have concluded that tissue hydraulic properties limit or co‐limit growth. This conclusion has been based on analyses of water potential gradients (Δψ) between water source (xylem) and expanding cells (Boyer et al., 1985). The magnitude of Δψ indicates whether growth is (co‐)limited by tissue hydraulic properties or not, since, in analogy to Ohm’s law, relative growth rate equals the product of L and Δψ. A value for Δψ of more than about 0·05 MPa is generally considered indicative of significant hydraulic limitation. In higher plants, significant growth‐associated water potential gradients have been reported for hypocotyl tissue (Nonami et al., 1997), and for leaves of dicotyledonous plants (Boyer, 1974) and grasses (Fricke et al., 1997; Fricke and Flowers, 1998; Martre et al., 1999, 2001; Tang and Boyer, 2002). Other studies have not found significant gradients (pea stem, Malone and Tomos, 1992; Rhicinus hypocotyl, Meshcheryakov et al., 1992) or have found only small gradients (Cosgrove and Steudle, 1981). The main argument against the existence of significant Δψ is that the experimental procedure to determine cell ψ causes artefacts and that plant cells have short half‐times of water exchange.

Philip (1958) and Molz and Boyer (1974) provided a theoretical basis for the prediction of Δψ in growing tissues. Nonami et al. (1997) used this approach to confirm measured, radial Δψ in soybean hypocotyls. The authors showed that Δψ was largely due to low water diffusivity in small cells surrounding the xylem. Fricke and Flowers (1998), studying the growth zone of barley leaves, also predicted values of Δψ that matched measured ones, although the authors had to make assumptions concerning cell elastic modulus and half‐times of water exchange of cells deeper in the tissue. Martre et al. (2001) computed water potentials in tall fescue leaves. The authors predicted values of Δψ between the xylem and expanding cells that were similar to those obtained using the cell pressure‐probe technique and picolitre osmometer (Martre et al., 1999).

MEASUREMENT OF TURGOR: ARTEFACTS OR ‘TRUE’ VALUES?

A major objection to the existence of significant growth‐associated Δψ in grasses is the experimental approach used to obtain ψ. Water potential can be determined at a tissue level in the growing and non‐growing part of the leaf, using isopiestic psychrometry (e.g. Tang and Boyer, 2002), or water potential can be determined for the water source (leaf xylem) and for peripheral expanding cells. In studies on Δψ in grass leaves, xylem ψ has been deduced from ψ in epidermal cells of the exposed blade (Fricke et al., 1997). This provides a most‐negative estimate for xylem ψ. In addition, xylem ψ may be significantly less negative in the elongation zone compared with that in the mature zone (Martre et al., 2000). If anything, estimation of xylem ψ from ψ in the emerged epidermis should lead to an under‐, not overestimation of Δψ between the xylem and epidermal cells in the elongation zone.

Direct determination of ψ in individual cells in situ is not possible; ψ has to be calculated from data on cell turgor (cell pressure probe) and cell osmolality (picolitre osmometry). Whether osmolality of extracted cell sap reflects the actual osmotic force in plantae cannot be said with certainty. A detailed study of reflection coefficients of main osmolytes in leaf cells is missing, but it is generally assumed that reflection coefficients are close to 1 (for a study on isolated epidermis and non‐electrolytes, see Tyerman and Steudle, 1982). For the analysis of growth, the cell probed for turgor must be located within the growth zone and should, ideally, expand during analysis. The latter may not be the case. When the tip of the cell pressure probe is inserted into a cell, subcellular compartmentation can be lost, and elongation reduced or arrested. Fortunately, this does not matter. Osmolality of neighbouring grass (barley) leaf epidermal cells is almost identical (Fricke, 1997), and the 2–4 min it takes to record turgor in a cell is too short to cause significant changes in solute content. The half‐time of water exchange of epidermal cells is in the second to sub‐second range (Fricke, 1997). Hence, the probed cell and (growing) neighbouring cells are in ψ‐equilibrium. Since osmolality is the same, turgor must be the same too. The main problem for determination of cell ψ is that some form of destructive plant preparation has to take place to access the leaf elongation zone and this affects leaf elongation velocity and, potentially, cell turgor.

There exists only a handful of studies in which cell turgor has been measured in the elongation zone of grass leaves (Table 2). It is striking that despite a large variation in residual elongation velocity and a wide range of elongation velocities of stress treatments, turgor was very similar (Table 2; elongation velocities ranged from <0·1 to almost 2·0 mm h^–1); turgor averaged 0·53 ± 0·04 MPa, including epidermal and mesophyll cells. It appears that cell turgor in the elongation zone of grass leaves falls within a very narrow range. Furthermore, two pressure‐probe studies on growing leaves of dicotyledonous plants yielded similar turgor: 0·48 ± 0·01 and 0·52 ± 0·01 MPa in sunflower leaves grown at high and low N, respectively (Palmer et al., 1996), and 0·53 MPa in leaves of Begonia argenteo gutatta (Serpe and Matthews, 1992).

Table 2.

Cell turgor in the elongation zone of cereal leaves, and residual leaf elongation velocity following plant preparation for turgor analysis

Plant and leaf number	Preparation for turgor analysis	% Residual elongation velocity	Turgor (MPa)	Experimental treatment	Reference
Barley, L3	Partial removal of sheaths	∼50	0·48 ± 0·05	Control and two levels of N‐limitation	Fricke et al. (1997)
			0·53 ± 0·07
			0·53 ± 0·07
Barley, L3	Window‐cut	42–46	0·56–0·57	Control and source‐reduction	Fricke (2000)
			0·52–0·58
English ryegrass, L4	Window‐cut	>80	0·50 ± 0·02	Control and changing temperature	Thomas et al. (1989)
Tall fescue	Window‐cut	>80	0·53 ± 0·01	Control	Martre et al. (1999)
Wheat, L1	Window‐cut	NI	∼0·45–0·50	Control and following addition of 25 mm or 150 mm NaCl	Arif and Tomos (1993)
Maize, L3	NI	∼60–70†	∼0·50–0·60	Control and following addition of polyethylene glycol	Thompson et al. (1997)
Barley, L1	Window‐cut	>80	0·63–0·68	Control and change in temperature	Pollock et al. (1990)
Barley, L3	Peeling back of older sheaths	NI†	0·54 ± 0·03	Control	Thiel et al. (1988)
Barley, L3	Window‐cut	46–53	0·47–0·51	Control and growth at 75 mm or 120 mm NaCl	Fricke and Peters, (2002)

Cell turgor was measured using the cell pressure probe. The elongation zone of cereal leaves is enclosed by the coleoptile or by sheath(s) of older leaves. Some form of destructive plant preparation has to take place to gain access to cells. Two methods are employed: complete removal of older leaves or partial removal. The latter consists mostly of cutting a small window into older sheath(s) (‘Window‐cut’). Regardless of the approach, leaf elongation velocity is reduced.

^† Thompson, pers. comm.

NI, not indicated in reference.

Using storage tissue of red beet (Beta vulgaris L.), Tomos et al. (1984) concluded that excision of tissue caused turgor to decrease substantially due to solutes leaking from damaged cells into the apoplast. The same could occur in the grass leaf elongation zone when it is made accessible for turgor analyses; however, this is not the case (Fricke and Peters, 2002).

VASCULATURE OF THE GROWING GRASS LEAF

The bulk of axial water movement through the elongation zone occurs through xylem elements in the midrib (MR) and large lateral veins (LV). For example, in barley, there are six to eight large lateral veins, three to four on either axial half of the blade (Fig. 3A; see also Dannenhoffer and Evert, 1994). Several studies suggest that most of the axial water transport through the elongation zone occurs through protoxylem vessels (PX, Fig. 3B and C; Dannenhoffer and Evert, 1994; Fricke and Flowers, 1998; Martre et al., 2001; Tang and Boyer, 2002). The much larger and faster‐conducting metaxylem vessels only mature at and beyond the distal end of the elongation zone (MX, Fig. 3D). The switch from a low‐conducting protoxylem path (elongation zone) to a high‐conducting metaxylem path (maturation zone) may explain the ten‐fold lower axial hydraulic conductance in the elongation zone than in the mature blade (Martre et al., 2000).

Fig. 3. Cross‐sectional view of the third leaf of barley at 10–12 mm (A and B) and 26–30 mm (C) from the point of leaf insertion, and halfway along the emerged part of the blade (D). At 26–30 mm from the point of leaf insertion, cells expand at near maximum relative rates, and growth‐associated water potential gradients are largest. The photographs show autofluorescence of lignified xylem vessels or vein tissue (blue) and mesophyll (red). Pictures were taken with an Axioscope (Zeiss, Germany: excitation filter, G 365; chromatic beam splitter, FT 395; barrier filter, LP 420). Note, the change in autofluorescence of metaxylem vessels between B/C and D. Cross‐sections of the entire leaf (A) and of larger leaf sections (B and C) show that not every vein and surrounding tissue is supplied with water by its ‘own’ vessels. Instead, almost 100 % of water is supplied by six large lateral veins (LV) and the midrib (MR). Along the elongation zone, water is conducted within protoxylem vessels (PX); the larger metaxylem vessels (MX) are lignified and thought to be fully functional only beyond the elongation zone (D). Protoxylem vessels are often separated by several cell layers from the mestome sheath (MSH) and parenchymatous bundle sheath (PBS). In contrast, metaxylem vessels border always directly at the mestome sheath. PH, Phloem; ABEP, abaxial epidermis; ADEP, adaxial epidermis. Bar as shown in part A represents 200 µm (A), 50 µm (B and C) or 25 µm (D).

The number of small cells, with a supposedly low diffusivity for water (Tang and Boyer, 2002), that water in the protoxylem (but not in the metaxylem) has to pass before it leaves the bundle is considerable, as is the distance that the water has to travel to outlying cells, particularly abaxial epidermal cells (Fig. 3A–C).

When the entire leaf is still enclosed by subtending sheaths and is only about 4 mm long, protophloem development precedes protoxylem development (Dannenhoffer and Evert, 1994). It is possible that the very young grass leaf uses predominantly organic solutes for osmolality generation, and that a significant proportion of water is provided through the phloem.

THE BUNDLE SHEATH: CONTROL POINT OF WATER (AND SOLUTE) SUPPLY TO GROWING LEAF CELLS?

The above considerations suggest that growth‐associated Δψ and limitation of elongation by tissue hydraulic conductance are partly due to low diffusivity for water of cells surrounding the protoxylem (see Tang and Boyer, 2002). Another possibility is that bundle sheath cells constitute a barrier to apoplastic water movement, and that water passage through bundle sheath membranes is controlled by the properties of aquaporins. Such a function of the bundle sheath is supported by studies that show suberization (Evert et al., 1996; see also Fig. 4) and high abundance of aquaporins (Frangne et al., 2001) in bundle sheath cells. In addition, the magnitude of growth‐associated Δψ is similar between xylem and mesophyll (Martre et al., 1999) and between xylem and epidermis (Fricke et al., 1997). This suggests that the main hydraulic constriction lies at the bundle sheath–mesophyll boundary, between bundle sheaths (parenchymatous and mestome sheath) or within the vascular bundle. Martre et al. (2001) studied the hydraulic architecture of tillers of tall fescue and predicted up to 0·3 MPa gradients in ψ between xylem vessels and adjacent mesophyll tissue in the transpiring, mature portion of the blade.

Fig. 4. Cross‐sectional view of the MR tissue of the mature blade of leaf three of barley, following staining with berberine hemisulfate (Brundrett et al., 1988). Leaves were viewed under fluorescence, with the same filter setting as in Fig. 3, and counter‐stained with aniline blue (A) or not counter‐stained (B). Following counter‐staining with aniline blue, berberine hemisulfate stains lignified walls bright yellow, Casparian bands intense yellow‐white and suberin blue white or blue (Brundrett et al., 1988). Note the bright fluorescence of metaxylem vessels, and the fluorescence of radial walls of the mestome sheath (MSH) and of the border region separating xylem and phloem. Fluorescence at the leaf surface points to guard cells or sclerenchymateous tissue. Red fluorescence in B originates from chlorophyll (mesophyll). Bars = 25 µm (A) and 50 µm (B). PBS, Parenchymatous bundle sheath.

Data on growth‐associated water flows, on Δψ and on bundle sheath dimensions can be used to estimate the hydraulic conductivity of bundle sheath cells that is required to sustain these flows (for equations, see Westgate and Steudle, 1985). Seven bundles supply the bulk of water to the growth zone of barley leaves (Fig. 3). On average, each bundle sheath has an outer circumference of approx. 400 µm, about 200 µm bordering the xylem part of bundles. Within a 10 mm long tissue section, this amounts to a bundle sheath membrane surface of about 14 mm² bordering the xylem. Between 20–30 mm from the point of leaf insertion, REGR is maximal and averages 8·66 % h^–1; the water content of this tissue segment is 6·4 mg (Fricke, 2002) and growth‐associated water flow is therefore 0·554 µl h^–1. If we assume that the bundle sheath membrane constitutes the main hydraulic resistance for peripheral water movement, the driving force for water movement across the bundle sheath amounts to a Δψ of 0·1–0·3 MPa (Fricke et al., 1997; Fricke and Peters, 2002). This gives a hydraulic conductivity for bundle sheath cells of 3·6 × 10^–8 to 11·2 × 10^–8 m s^–1 MPa^–1, a value well within the range of previously published conductivities of higher plant cells (e.g. Tyerman and Steudle, 1982; Steudle and Peterson, 1998) and considerably lower than conductivities of peripheral MR tissue cells (Westgate and Steudle, 1985).

The location of the hydraulic constriction within the bundle sheath may vary among grass genera. In those grasses with one bundle sheath (e.g. Zea), the sheath is parenchymatous and contains chloroplasts. The sheath can be suberized. In those grasses with two bundle sheaths (e.g. Triticum, Hordeum and Avena), the outer sheath is also parenchymatous and contains chloroplast, but it is not suberized. The inner sheath, termed the sclerenchymatous bundle or mestome sheath, is colourless and contains thickened walls and extensive suberin lamella (O’Brien and Kuo, 1975; Hattersley and Browning, 1981). Suberiz ation in the mestome sheath is particularly prominent in large LVs, those veins that transport the bulk of water through the growth zone (Dannenhoffer et al., 1990; see also Fig. 4).

THE SOLUTE ASPECT: CONSTANT OSMOTIC ADJUSTMENT!

There are a few studies on grass leaves, and some on roots, where osmolality in the elongation zone has been determined at the cellular level (Pritchard, 1994, and references therein; Fricke et al., 1997; Martre et al., 1999; Fricke, 2002; Fricke and Peters, 2002). These studies show that cell osmolality and turgor change little along the elongation zone. The implication is that growing cells constantly deposit solutes to maintain osmolality and that relative rates of volume expansion are matched by similar relative rates of solute deposition. The reason why osmolality does not change along the elongation zone, but instead remains constant with time and development is not known. But, if we assume that wall properties (ϕ, Y) change little along the elongation zone, and that the constancy in turgor indicates that turgor is close to the wall yield threshold, osmolality must be maintained to maintain the driving force for water uptake (Δψ) and the mechanical force for cell expansion (turgor). In this respect, growing cells exhibit a special form of osmotic adjustment, and they are constantly achieving (growth) at minimal effort (solute transport).

Since expansion growth of cells must be matched by adequate rates of solute supply or solute uptake, either rate can become growth‐limiting. This seems to be the case for barley exposed to high external NaCl (Fricke and Peters, 2002). High salt overloads the capacity of epidermal cells to maintain osmolality during growth dilution and to adjust osmotically to the large decrease in external ψ. Reduction in leaf‐ and cell‐elongation rate at high external NaCl is therefore not so much a detrimental effect on the plant, but a mechanism through which the plant guarantees sustained water uptake and growth (see also Munns et al., 1982). Similarly, Frensch (1997) concluded that growth in osmotically stressed maize roots is limited by solute supply. It appears that limitation of growth by solute supply occurs either in plants exposed to large decreases in external ψ, or in tissues that depend on a high rate of radial transport of phloem‐borne solutes (discussed in Cosgrove, 1993).

BIOPHYSICAL LIMITATION OF LEAF CELL EXPANSION: A MODEL

The above considerations can be summarized in the following model (Fig. 5A–F). Substantial water potential gradients exist between xylem and peripheral mesophyll and epidermal cells in the growth zone of grass leaves (Fig. 5A–C, E and F). By analogy to the composite model of water transport in the root (Steudle, 2000), water moves along two main pathways, depending on the driving force, and encounters major apoplastic barriers. Growth‐associated water uptake is mainly osmotically driven. Apoplastic tension induced by lowering of cell ψ (Boyer, 2001) may also pull water peripherally, but this depends on how much lowered cell ψ translates into lowered apoplastic solute potential.

Fig. 5. Model for water transport between xylem and peripheral cells in the elongation zone of grass leaves. Various scenarios are considered (A–F). Tissue (cross‐sectional view) is divided into tissue located inside vascular bundles (blue; xylem, phloem, parenchyma), bundle sheaths (purple; mestome sheath and parenchymatous bundle sheath), mesophyll (green) and epidermis (yellow). Water moves along a combined transcellular/symplastic path, i.e. crossing protoplasts (black arrows; the length of arrows in A–F indicates whether a particular pathway applies only for passage of the bundle sheaths or for the entire path between vein tissue and epidermis), or along a predominantly apoplastic path (white arrows). Proposed gradients in water potential (ψ) are also shown. A, Water reaches peripheral cells along a symplastic/transcellular path, and a large drop in water potential occurs at the bundle sheath (lowest hydraulic conductance), due to suberization of the mestome sheath and a low hydraulic conductivity (aquaporin regulation) of cells. For the same reason, water moving towards the xylem (transpiration tension) has to cross the bundle sheath along the symplastic/transcellular path. Growth‐associated Δψ does not translate into apoplast tension, hence there is no outward‐directed hydraulic force. B, Same pathways of water movement as in A, except that layers of small parenchymatous cells, which are located within the bundle, represent the main hydraulic barrier (see drop in ψ; compare Tang and Boyer, 2002). C, As in A, except that water moves peripherally along the symplastic/transcellular path only at the bundle sheath region (as indicated by short black arrow) due to apoplastic barriers. Before and beyond the bundle sheath, water moves apoplastically (for simplicity not shown), either driven hydraulically by growth‐induced apoplast tension (outward) or by xylem tension (inward). D, This is the only scenario with insignificant growth‐associated Δψ. This could be due to overestimation of osmotic forces based on in vitro analyses of osmolality—if cells have reflection coefficients for main osmolytes that are <<1·0—or due to consistent underestimation of turgor pressure in plants prepared for turgor analyses (see Table 2) regardless of residual elongation velocity. Water moves in either direction mainly along the low‐resistance apoplast path, and the bundle sheath represents neither an apoplastic nor an osmotic barrier for water movement. E, As in A, except that the bundle sheath does not represent an apoplastic barrier for water movement. Despite this, water moves peripherally along the symplastic/transcellular path since growth‐associated ψ does not cause apoplast tension (but apoplast solute potential which cannot drive water movement). Water potential drops at the bundle sheath due to low conductivity of cells, most likely because of low aquaporin activity (this is also implicit in A). Inward flow of water is driven by transpiration tension. This scenario implies that during periods of positive xylem pressure (no transpiration) water can move peripherally along the apoplast and that ψ gradients may be insignificant (not indicated in figure). F, As in E, except that low‐conducting cells in the xylem parenchyma represent the major hydraulic constriction.

According to the model, the bundle sheath of grass leaves fulfils a role similar to that of the endodermis in roots, both in terms of water and in terms of solute transport, except that the direction of flows is outwards, not inwards. For water transport, aquaporin function will be crucial. Aquaporin activity is responsive to hormones (abscissic acid, reviewed in Steudle, 2000), activity is fine‐tuned by phosphorylation/de‐phosphorylation (Kjellbom et al., 1999) and expression varies diurnally (Clarkson et al., 2000). Therefore, aquaporins show all the prerequisites that are needed in the growing leaf to respond to environmental, diurnal and developmental changes in water demand. Future studies into the expression and phosphorylation of aquaporins, combined with determination of hydraulic conductivity in bundle sheath cells (cell pressure probe) and chemical analyses of wall composition (e.g. suberin, reviewed in Steudle, 2000) will show whether bundle sheath aquaporins really control water supply to growing leaf cells, whether hydraulic limitation resides within the xylem parenchyma as suggested for maize (Tang and Boyer, 2002), or whether growth‐associated Δψ are overestimated due to solute reflection coefficients <<1 (i.e. actual osmotic force << force suggested by in vitro osmolality; Fig. 5D).

Water circulates along two pathways between xylem and peripheral cells. Changes in the driving forces (Δψ; xylem ψ and cell ψ) and hydraulic conductance (suberization, aquaporins, plasmodesmata) control the partitioning of water between growth and transpiration.

By analogy to Ohm’s Law, relative growth rate equals the product of L and Δψ. This implies that cells can grow at unchanged rates despite generating smaller Δψ and accumulating fewer solutes, simply by increasing tissue hydraulic conductance. Therefore, localized changes in hydraulic conductance of bundle sheath cells will affect the growth‐associated solute demand of the entire leaf. For example, increase in L enables leaves to elongate at substantial velocities when solutes or energy are scarce (Fricke, 2002), or when cells have to adjust osmotically to considerable decreases in external ψ, such as during salt or water stress (Hsiao and Xu, 2000; Steudle, 2000).

The above model is testable: it does not require major changes in wall properties, and it provides mechanisms through which leaf cell elongation can be actively regulated in the short‐ (aquaporin activity, plasmodesmata) and long‐term (aquaporin abundance, solute accumulation, suberization). Regulation will be subject to hormonal control. Judging from studies on the REGR profile of dwarf‐ and slender mutants, it appears that gibberellins are required to achieve high relative elongation rates in the distal part of the elongation zone (Schünmann et al., 1997).

ACKNOWLEDGEMENTS

The author would like to thank Paisley University for financial support and Winfried S. Peters (Frankfurt University) for discussions concerning REGR profiles. In a short review such as this, it has not been possible to cite all notable contributions to the subject of grass leaf growth or plant water relationships.

Received: 18 October 2001; Returned for Revision: 12 January 2002; Accepted: 17 May 2002