Plant Physiol. (1985) 78, 183-191 0032-0889/85/78/0183/09/$01.00/0 Water Transport in the Midrib Tissue of Maize Leaves' DIRECT MEASUREMENT OF THE PROPAGATION OF CHANGES IN CELL TURGOR ACROSS A PLANT TISSUE Received for publication September 5, 1984 and in revised form January 4, 1985 MARK E. WESTGATE2 AND ERNST STEUDLE* Arbeitsgruppe Membranforschung am Institutffur Medizin, Kernforschungsanlage Jiulich, Postfach 1913, D-5J 70 Julich, Federal Republic ofGermany

ABSTRACT cellular (vacuolar) and symplasmic components. The porous nature of the wall material and the relative impermeability of Water movement across plant tissues occurs along two paths: from cell membranes to water has led to the conclusion that water cell-to-cell and in the apoplasm. We examined the contribution of these transport in plant tissues occurs primarily in the apoplasm. two paths to the kinetics of water transport across the parenchymatous However, the rapid rates of water exchange exhibited by cells of midrib tissue of the maize (Zea mays L.) leaf. Water relations parameters higher plants (17) suggests that substantial cell-to-cell transport (hydraulic conductivity, Lpr, cell elastic coefficient, e; half-time of water of water could occur under some conditions. In this paper, we exchange for individual cells, T,,2) of individual parenchyma cells deter- investigated the kinetics of water transport in the midrib tissue mined with the pressure probe varied in different regions of the midrib. of the maize leaf at the individual cell and whole tissue levels in In the adaxial region, Lp = (03 ± 0.3).10' centimeters per second per order to determine the contribution of apoplasmic and cell-to- bar, e = 103 - 72 bar, and T,,2 = 7.9 ± 4.8 seconds (n = seven cells); cell paths to water movement. whereas, in the abaxial region, Lp = (2.5 - 0.9). 10i centimeters per The controversy concerning which is the dominant pathway second per bar, e = 41 ± 9 bar, and T112 = 1.3 ± 0.5 seconds (n = 7). is very old. Pfeffer (10) was perhaps the first to suggest that water This zonal variation in Lp, e, and T1/2 indicates that tissue inhomogeneities was transmitted across the tissue osmotically and exist for these parameters and could have an effect on the kinetics of that cells were water transport across the tissue. involved. Sachs (12), on the other hand, proposed that water movement in tissues (and even over longer distances) was me- The diffusivity of the tissue to water (D,) obtained from the sorption diated the kinetics of rehydrating tissue was D, = (1.1 ± 0.4).10' square centi- by imbibition of wall material via capillary forces meters per second = The of ('imbibition theory'). Although Pfeffer's point ofview was widely (n 6). diffusivity the cell-to-cell path (Dj) accepted, experiments showing that fluorescent or electron-dense calculated from pressure probe data ranged from De = 04 10' square material introduced into the centimeters per second in the adaxial region to = 6.1 10-' transpiration stream could move De square quickly within the apoplasm suggested that this compartment centimeters per second in the abaxial region of the tissue. D, = De suggests substantial cell-to-cell transport of water occurred during rehy- represented the main path for water transport in plant tissues dration. However, the tissue diffusivity calculated from the kinetics of (18, 19). However, movement of water from cell to cell may pressure-propagation across the tissue (D,') was D,' = (33.1 ± 8.0). 101 have gone undetected in such experiments due to the low perme- square centimeters per second (n = 8) and more than 1 order ofmagnitude ability ofcell membranes to dyes or electron-dense materials and larger than D,. Also, the hydraulic conductance of the midrib tissue (Lp. the preferential binding of dyes to the cell wall. per square centimeter of surface) estimated from pressure-induced flows Philip ( 11) was the first to formulate a general theory for water across several parenchyma cell layers was Lp., = (8.9 ± 5.6). 101 movement in plant tissues. His analysis assumed that water centimeters per second per bar (n = 5) and much larger than Lp. moved from cell to cell and a possible apoplasmic component These results indicate that the preferential path for water transport was not considered. Philip's theory was extended by Molz and across the midrib tissue depends on the nature of the driving forces Ikenberry (9) who incorporated both cell-to-cell and apoplasmic present within the tissue. Under osmotic conditions, the cell-to-cell path pathways. They assumed that each cell within the tissue was in dominates, whereas under hydrostatic conditions water moves primarily local water-flux equilibrium with its immediate surroundings. in the apoplasm. With this assumption, they quantified the contribution of the two pathways in terms ofthe hydraulic conductances and storage capacities for water. Using the experimental data available, they concluded that both pathways could contribute to a similar extent to the overall transport of water across the tissue. The actual contribution ofthe cell-to-cell path to the transport of water in a plant tissue can be determined directly if the water It is generally thought that there are two possible pathways for exchange characteristics ofthe individual cells along the path are water movement in plant tissues: within the apoplasm and from known. With the introduction of the pressure probe technique cell to cell. The former pathway is composed primarily of a cell (5, 17, 22, 23), it is now possible to measure the water relations wall component, whereas the latter is comprised of both trans- parameters of individual plant cells. For most higher plant cells, the rate of water exchange with the local environment is fast (on 'Supported by a grant from the Deutsche Forschungsgemeinschaft, the order of seconds) and the hydraulic conductivity of the cell Zi 99/8. membrane is relatively high (17). Comparison of the half-times 2 Present address: United States Department of Agriculture, Agricul- for osmotic shrinking and swelling ofwhole tissues with the rates tural Research Service, North Central Soil Conservation Research Lab- of water exchange for individual cells indicates that a large oratory, North Iowa Avenue, Morris, MN 56267. portion of the water could flow along a cell-to-cell path during 183 184 WESTGATE AND STEUDLE Plant Physiol. Vol. 78, 1985 tissue hydration and dehydration (1, 13, 15). However, Steudle vessels at the periphery. Therefore, the kinetics ofwater transport and Boyer (13) found that hydrostatic pressure gradients applied across the tissue could be measured with the pressure probe and across the cortex of soybean hypocotyls caused large water flows analyzed using standard methods for radial diffusion in a solid along the apoplasmic path. They concluded that the hydraulic cylinder. conductance of a tissue (i.e. the dominant path for water flow) The mechanism(s) controlling water transport in this tissue could depend on the nature of the driving force applied across were investigated using a combination of four techniques. (a) the tissue and that under 'osmotic conditions' the apoplasm was The Julich pressure probe was used to measure the water ex- rather ineffective in water transport. These results were different change characteristics of individual parenchyma cells within the from those obtained for barley roots (14) where it was shown midrib. (b) Water uptake (sorption) of dehydrated tissue via the that the hydraulic conductance ofthe root tissue was independent xylem gave estimates ofthe kinetical parameters governing tissue of the nature of the driving forces applied. Water movement rehydration. (c) A pressure-perfusion technique (13) was used to across the roots was predominately along a cell-to-cell path. measure the tissue hydraulic conductivity in the presence of a In this study we investigated the problem ofwater transport in stationary hydrostatic pressure gradient. (d) In addition, a pres- plant tissues further using the midrib tissue of the maize leaf. sure-propagation technique was developed to follow the propa- This tissue was chosen because it has a fairly regular shape (half gation ofchanges in cell turgor across the tissue. The combination of a cylinder), it contains large cells up to 12 cell layers distant of these techniques provides the basis for the study of water from the nearest xylem vessel, and it is supplied with water by transport in tissues under physically well-defined conditions. MATERIALS AND METHODS Maize plants (Zea mays L., cv B73xMo17) were grown from seeds in a controlled environment chamber (day/night temper- atures, 30/20 ± 20C; RH, 90/45 ± 5%; photoperiod, 14 h; light intensity, approximately 350 4mol.m-2.s-' PAR at the level of the leaves). Seeds were sown in soil mix (Einheitserde, Type ED 73; Gebr. Patzer KG, Sinntal, FRG) and plants were well watered throughout development. All experiments were conducted using the mature blade ofleaf 3 from plants 13 to 17 d old. The leaf blades were about 25 cm long and 1.5 cm wide. All measurements were made on cells in the midrib region 6 to 7 cm from the ligule. At this point, the midrib was about 2 mm in diameter and the shape could be approximated by half a cylinder (Fig. 1). The midrib cylinders were supplied with water by vessels at the cylinder surface (five A main vascular bundles containing vessels approximately 30 to 40 ,gm in diameter which carried most of the water and 12 to 14 minor bundles with much smaller diameters). The cylinder radius, R3, was equivalent to 8 to 12 cell layers. The bulk of the midrib tissue consisted of rather large parenchyma cells which increased in diameter from the cylinder axis towards the periph- 3 Abbreviations: R (cm), radius of the midrib cylinder; ac, and a, (cm2), cross-sectional area of the cell-to-cell path and cell wall path; A I:T jii (cm2), cell surface area; Ct C(A (cm3n-bar-'), per-cell water storage \, \ capacity of the cell-to-cell path and cell wall path; C., (cm3.bar-'), water storage capacity of the xylem; d (cm), tissue thickness; D, (cm2 s%'), diffusivity of the tissue for water; D, (cm2 - s-'), diffusivity of the cell-to- cell path for water; Fu, fractional uptake of water during sorption; Jvm and Jv,(cm3.s'), stationary water flow across the midrib tissue and the xylem; k,, rate constant for water uptake during sorption; I and 4, (cm), length of an individual cell and midrib pore path; Lp (cmrns-'-bar-'), hydraulic conductivity of cell membranes; Lp,. (cm2-s-' bar'), hy- draulic conductivity of the cell wall path; Lpm and Lpp (cm*s'-bar-'), FIG. 1. A, Cross-section of the midrib of a mature maize leaf. The hydraulic conductance of the midrib tissue and midrib pore path; P and section was taken from leaf No. 3 (from the bottom of the plant) P, (bar), hydrostatic pressure in an individual cell and the xylem; r, rp, approximately 7 cm above the ligule 14 d after planting. The midrib and r,(cm), radius ofa cylindrical cell, an intercellular pore, and a xylem tissue is composed primarily of large vacuolated parenchymatous cells vessel; I/RX (cm3.s'- bar '), hydraulic conductance ofthe midrib xylem which are supplied with water from xylem vessels at the periphery. The vessels; 11/2 (s), half-time for tissue responses to a change in xylem water geometry of the midrib can be approximated by half a cylinder (R, potential; T,,2 (s), half-time for water exchange of an individual cell; tax radius). R ranged from 0.09 to 0.10 cm (8 to 12 cell layers). B, Schematic (s), time constant for a change in Px; V and V.x (cm3), volume of an representation of the midrib tissue in A showing three zones (I-III) for individual cell and the xylem; Ax (cm), cell diameter, X (cm), distance which the water relations parameters (P, (, T,,2, and Lp) were determined. ofa cell from the peripheral xylem vessels; ye and ay,, mean fraction of Each zone has approximately three to four cell layers in the radial the cross-sectional area available for cell-to-cell and apoplasmic water direction. Cell dimensions: zone I, 2r = 64 20 Am (n = 376), 1 = 129 transport; e and E. (bar), elastic coefficient of an individual cell and the ± 12 Am, V= 0.42 x 106 cm3; zone II, 2r= 97 ± 25 Mm (n = 629), 1= xylem; q (bar-s), viscosity of water; ir' and ir. (bar), osmotic pressure of 150 ± 16 Am, V= 1.1Ox 10-6 cm3; zone III, 2r= 112 ± 27 Am (n = cell sap and xylem sap; p, ratio ofwater transported in apoplasm to water 480), 1 = 160 ± 17 Am, V= 1.58 x 10-5 cm3. Mean values ± SD; n = transported from cell-to-cell; ai, reflection coefficient of the cell wall number of cells measured. material; iI' (bar), water potential oftissue. WATER TRANSPORT IN PLANT TISSUES 185 ery (Fig. IA). Measurements from several leaves indicated that there was a good linear correlation (correlation coefficient, 0.92) between microscope (A) cell diameter (2r) and cell length (1). Since cell diameters were somewhat variable, the midrib was divided into three zones (Fig. Cell sap 1 B, I-III) and average cell diameters were determined for each licone oil zone. The relationship between 2r and / was used to determine I meniscus from measurements of the cell diameter. Average cell volumes leaf tissue clamp (V) and cell surface areas (A) for the parenchyma cells located a in each of the three zones were calculated from mean 2r and I vent to Xprobe values assuming the cells to be cylinders. The average volumes ranged from 0.42 to 1.58. 10-6 cm3 and average surface areas ranges from 3.7 to 7.6. 10-8 cm2. The cell wall thickness, esti- mated from cross-sections, was 0.3 to 0.9 ,um. Intercellular air spaces ranged from 0.5 to 9 ,um in diameter. Cell Water Relations. Cell water relations parameters (cell turgor pressure, P; cell elastic coefficient, E; half-time of water exchange for individual cells, T,2; and the hydraulic conductiv- ity, Lp) were determined using the pressure probe as described to infusion' -pressure transducer for 13-16, Corrections for the efflux of pump measuring pressure previously (5, 20). rapid applied to leaf xylem water were not necessary for e determinations since the meniscus closed water reservoir in the tip of the pressure probe could be moved with sufficient FIG. 2. Experimental arrangement used to monitor the response of speed (15). The water relations parameters were related to indi- cell turgor (P) to changes in pressure in the maize leaf. vidual cells within the midrib tissue (i.e. to cells with a certain xylem (Ps) The volume and leafblade was excised from the plant at the ligule and recut under water. surface area) by measuring the depth of the tip of A 1-cm ofthe was the pressure probe under the epidermis and classifying each cell region midrib exposed by cutting away the blade. The into zone I, II, or III (Fig. I B). cut end ofthe leafwas mounted in a pressure chamber and a water-tight Sorption Kinetics. Sorption kinetics of isolated midribs were seal was formed around the leafwith silicone rubber (see "Materials and measured to estimate the tissue diffusivity for water (D,) which Methods"). The seal ensured that water was supplied to the leaf only via comprises both the apoplasmic and the cell-to-cell transport the exposed vessels in the midrib. A perfusion pump and vacuum jet paths for water movement (9, 13). Midrib sections (about 6 cm line were connected in parallel to the reservoir supply line so that in length) were isolated from the leaves using a razor blade. They hydrostatic pressures greater than atmospheric (Ps > 0) and less than were dried in air to about 85% of their original weight, coated atmospheric (P, <0) could be applied and changed instantaneously. The with vaseline and Saran paper (13), and recut under water to a pressure in the chamber was monitored with a pressure transducer. A final length of about 3 cm. This initiated rehydration via water- small collector was included in the vacuum line to accommodate changes filled vessels. The tissue was weighed after certain time intervals in volume ifair bubbles formed in the reservoir at P, <0. A conventional to determine the fractional uptake of water, Fui (Fu = water pressure probe was used to measure cell turgor pressure (P). The leafwas uptake at time t/total water uptake). D, was estimated from the held firmly in place by the silicone seal in the pressure chamber and a rate constant for water uptake (k,) which is given by the slope of retaining clamp at the point of measurement. For clarity, the upper part In (l-Fu) verslus time (6): ofthe pressure chamber is not shown. measured by the probe after a step change in P. should directly = In (2).R- k, (1) reflect the kinetics of the propagation of water potential (or (2.405)2 turgor or cell volume) across the tissue. k, was evaluated from the linear part of the curve where the The distance of the cell from the peripheral vessels (X) could sorption kinetics were exponential to a good approximation. be varied in order to test the proposed type of kinetics (e.g. for This evaluation of D, assumes a homogeneous tissue with uni- diffusion; t'12, X2). For the specific case of a solid cylinder, the form cells. Because of the inhomogeneities of the midrib tissue, relationship between the half-time for a change in water potential this procedure only gives a rough estimate of D,. (or turgor or cell volume), expressed by the dimensionless quan- Propagation of Water Potential through the Tissue after a tity DI'.1/2 *R-2, and the relative distance from the cylinder Change in Xylem Water Potential (Pressure-Propagation Tech- surface, given by the dimensionless parameterX/R, is well known nique). A new method was developed to follow the kinetics of (e.g. Fig. lB in Ref. 11). Therefore, if t ]/2 and the position of the cell turgor changes in response to a change in xylem hydrostatic cell within the cylinder (X/R) are known, it is possible to estimate pressure (P,). In this pressure-propagation technique, an excised D,' from a standard curve (1 1). leaf was clamped via a water-tight seal to a closed water reservoir The pressure-propagation experiments required a water-tight filled with distilled and degassed water (Fig. 2). Using an infusion seal around the leaf that did not restrict the flow of water in the pump (Unita, Braun/Melsungen, FRG) or a vacuum jet, the xylem from the cut end of the midrib to the point of cell turgor hydrostatic pressure in the reservoir could be varied between measurement. The seals were prepared from a silicone paste several bars above atmospheric to near vacuum (about -I bar (Xantopren-plus from Bayer, Leverkusen, FRG). This liquid water potential). The hydraulic conductances of the clamp (i.e. silicone material could be hardened with an activator while in of the vessels within the leafclamped by the water-tight seal) and contact with the leaf forming a seal without damaging the tissue. the vessels in the leaf up to the point of cell turgor measurement Even after 4 h in contact were much larger than that of the tissue outside the vessels (see with the material, no visible damage to the half-times for )4 the epidermal cells was evident. The seal was formed around the "Results"). Thus, changes in cell turgor (t /2 leaf and between the leaf and a perspex clamp (Fig. 2). The leaf was arranged within the seal so that only the cut end of the 4 tl/2 denotes the half-time for the propagation of a change in water midrib was exposed to the water reservoir of the clamp. The cut potential (or turgor or cell volume) across the tissue following a change end of the midrib was in contact with water during all manipu- in PA. A primed symbol is used to distinguish this value from that lations to prevent air from entering the xylem. obtained in sorption kinetics (ti/2) which may reflect a different process. To estimate the hydraulic conductance ofthe xylem vessels in 186 WESTGATE AND STEUDLE Plant Physiol. Vol. 78, 1985 osmotic environment at all times. However, it is remarkable that crosnape the T1n values were shorter and the Lp values were higher (by 1 order ofmagnitude) for the cells in zones II and III than for the cells in zone I. This is the first time that such large differences in Lp and T,,2 have been observed over such short distances in the same tissue. In the leaf of Oxalis carnosa (16), the epidermal bladder cells on the adaxial and abaxial epidermis showed sig- nificant differences in Lp and especially in . These differences as well as those found between epidermal and cortical cells in growing pea epicotyl (1) also occurred over short distances but may have reflected differences between tissues. The differences in T1/2, Lp, and e within the midrib parenchyma tissue indicate 'sitcone seaing that tissue inhomogeneities for these parameters exist and could material have an effect on the kinetics ofwater transport across the tissue. It should be pointed out that the values for individual arrangement for measuring pressure-induced T,12 FIG. 3. Experimental cells are measured accurately with the pressure probe. However, stationary water flow across the midrib parenchyma tissue. A Lp and e are derived values which depend on the accurate section of the leaf blade was excised from the leaf near the ligule and measurement of cell volume and surface area. Since an average mounted in the pressure clamp as in Figure 2. Except for a small (2 volume (V) and surface area (A) were used in the calculation of 0.4cm) region ofthe adaxial surface, the entire leaf section was encased and Lp, these values may be reliable only within a factor of 2. in a silicone sealing material. A region of exposed midrib tissue was For an extensive discussion of errors involved in measurements carefully removed forming a trough. Water flowing across the remaining ofe and Lp, see Steudle et al. (15) and Tomos et al. (20). midrib tissue was collected in the trough at P> Sorption Experiments. To obtain an estimate of the total from the trough at P, < 0. The rate of water flow was determined from diffusivity of the leaf midrib tissue for water (D,), sorption the movement ofthe air/water meniscus in a calibrated capillary attached experiments were performed on isolated sections of the midrib to a small water reservoir above the trough. in a fashion similar to that described by Steudle and Boyer (13). according the clamped area and of the vessels in the leaf, the leaf was cut D,, calculated from the rate of rehydration to equation after the probe measurements at the position where the probe 1, is determined by the water relations parameters and the (9): had been inserted. This caused a large drop in the pressure geometry of the cells and apoplasm gradient across the leaf. A second cut of the leaf right at the seal caused a further drop in pressure indicating that the clamp D = Ax(Lp.-ac. + Lp.ac. Ax/2) ccc + (2) resistance was small. The hydraulic conductance of the vessels CC. (AJvxIAPx) was determined by collecting the water flowing out where Lp. (cm2-s'- barC') denotes the hydraulic conductivity of the vessels at stationary flows (Jv,) and stationary pressures of the cell wall material; Lp (cm-s'-bar') is the hydraulic (Ps) applied to the clamp. conductivity of the cell membranes; a. and a, (cm2) are the Pressure-Perfusion Experiments. The hydraulic conductance cross-sectional areas of the cell wall and cell-to-cell path, respec- of the midrib parenchyma tissue (Lpm) was measured directly tively (acc >> a,,). Ax represents the cell diameter (here equal to using a pressure-perfusion technique. A small region of the 2r); and C. andCcc (cm3 bar') are the per-cell water storage adaxial surface of the midrib was removed forming a V-shaped capacities of the cell wall and cell-to-cell paths, respectively. As channel approximately 16 mm long and 420zm wide at the seen in equation 2, D, reflects the rate of water transport in two surface and approximately 500 um deep. One end of the leaf parallel pathways (cell-to-cell and apoplasm). The value ofD, for section (about 6 cm long) was attached to the leaf clamp (Fig. 2) the midrib tissue of the maize leaf obtained from sorption and water was forced across this modified midrib tissue (approx- kinetics was D, = (1.1 ± 0.4)10' cm2r-s' (TableI). This is imately three to six cell layers) under positive or negative hydro- similar to D, for soybean hypocotyls (8, 13) and pea epicotyls (1) static pressures. The conductance to flow (Lpm) was measured calculated using similar methods. as (AJvm/APm). The rate of water flow was monitored in a small The diffusivity of the cell-to-cell path alone (Dc) can be cal- calibrated capillary which was attached to a collection chamber culated provided the half-times for water exchange(T,/2) and the situated above the midrib (Fig. 3). All other paths for water flow cell diameters (2r) for cells along the path are known. Assuming were sealed by coating the remainder of the leaf section with the a tissue of uniform cells, Dc is given by(1 1): silicone sealing material. Therefore, water forced into or out of the xylem vessels should have moved across the parenchyma Dc=2 LpAx(f+wi) (3) cells in the midrib and into or out of the collection vessel. 2 RESULTS where a is a shape factor which can be taken as unity to a good approximation. Since T,2 = V/A -ln(2)fLp(e+ri)(I 1, 22) and Cell Water Relations Parmeters. Table I summarizes the V/A = r/2 for a cylindrical cell withI >> 2r, equation 3 can be water relations parameters for the parenchyma and epidermal rewritten as: cells in the three zones ofmidrib tissue. Cell turgor (P) measured

uniform throughout ln(2) r2 directly with the pressure probe, was fairly 2c T1 (4) the tissue ranging from 6.8 to 7.9 bar. These values agree well with those obtained for mature maize leaves using the thermo- DC couple psychrometer (P = 6.5-8 bar [7,211). The latter method If D,,1 it can be concluded that the cell-to-cell path dominates transport across the tissue (13, averages a large number of cells and yields an estimate of P water 15). Average values ofDc vP from + The cell elastic modulus (e), the half-times for water for the three zones of midrib tissue ranged 0.4 to 6.1106 wr. (Table This indicates that there was a exchange of individual cells(T,,2), and the cell hydraulic con- cm2.sS' II). substantial transport of water in the tissue during in ductivity (Lp) were similar to those reported for other higher cell-to-cell rehydration experiments. plant cells (17, 23). The shortT,,2 and high Lp values indicate the sorption of Water a that the cells should be in near equilibrium with their local Propagation Potental through the Tissue after WATER TRANSPORT IN PLANT TISSUES 187 Table I. Water Relations Parameters (P, E, T,/2, Lp) ofIndividual Parenchyma and Epidermal Cellsfrom the Midrib ofa Mature Maize LeafNo. 3 Obtained Using the Pressure Probe Data are mean ± SD (n = number of cells) for cells measured in the three regions defined in Figure 1.

Cell Cell Cell Elastic Pressure Cell Half- Cell CellType Volume Turgor Modulus Range Time Hydraulic (VX109) (P) for (T12 Conductivity cm3 bar bar bar s cm-s'.bar-' Parenchyma Zone I (n = 7) 422 6.9 ± 0.7 103 ± 72 3.4_P<9.7 7.9 ± 4.8 0.3 ± 0.3 Zone II (n = 14) 1101 7.5 ± 1.0 67 ± 60 3.0'P_10.4 1.5 ± 0.4 2.0 ± 1.2 Zone III (n = 7) 1580 7.9 ± 1.0 41 ± 9 4.3'P_10.4 1.3 ± 0.5 2.5 ± 0.9 Epidermis Zone I (n = 10) 154 ± 75 6.8 ± 0.7 48 ± 24 3.9'P=10.1 4.6 ± 4.8 0.4 ± 0.3 Zone II (n = 15) 247 ± 85 7.5 ± 1.4 34 ± 16 2.6_P'12.2 1.7 ± 0.8 1.4 ± 0.7 Zone III (n= 5) 190± 138 7.6± 1.7 30±28 4.3'P'-11.4 1.5±0.3 1.5±0.8

Table II. Diffusivity ofMidrib Parenchyma Tissue Calculatedfrom the Kinetics ofthe Tissue Response to a Change in Xylem Pressure (D,'), the Kinetics ofTissue Rehydration (D,), and the Kinetics ofCell Turgor Pressure Relaxations (D,) The relative position of the cell within the half-cylinder of parenchymatous tissue (X/R) was determined from the depth at which the probe was inserted below the epidermis (R-X) assuming an average R = 0.0938 cm. D,' was calculated from D,' * tl/2*R-2 values obtained from Figure lb in Philip (1958) for an ideal cylinder of tissue. D, values were calculated from the rates of rehydration of the midrib tissue measured in sorption experiments (see "Materials and Methods"; t,/2 = 1895 ± 1128 s [range, 1119-4158 s]) using equation 1. D, values were calculated from average T/2 values and cell dimensions for cells in zones I to III (see Fig. I and Table I) using equation 4. CellNo. X/R t1'/2 D,'x 106 D,x 106 D,x 106 s cm2-s' cm2.s-' cm2.s' 71 0.230 11 33.6 99 0.310 24 27.1 Mean: Zone I: 1.1 ± 0.4 (6) 0.4 70 0.358 15 48.1 98 0.362 32 23.1 Range: Zone II: 0.4-1.5 5.4 101 0.586 37 35.7 Zone III: 6.1 88 0.614 54 25.7 97 0.776 50 33.1 100 0.839 44 38.4 Mean ofD,' = 33.1 ± 8.0 (8)

Change in Xylem Water Potential. The measurement of the the xylem could be changed stepwise and the response of the diffusivity for the entire midrib tissue based on sorption kinetics turgor of the cells outside the xylem monitored simultaneously. was a 'cumulative' method of determining D, and assumed that This approach assumes that the change in cell water potential in the tissue was uniform ( 11). Measurements ofthe water relations response to a change in xylem water potential will be expressed parameters of individual cells within the midrib suggested that primarily as a change in cell turgor. this assumption was questionable (Table I). Therefore, it was Since desirable to measure D, in an independent and more direct way. dP 6e This was done by following the changes in cell turgor of individ- (5) ual cells within the midrib tissue in response to changes in xylem d4 E +lri pressure (PA) at the periphery of the midrib. Conceptually, this and e >> Wri (Table I), this assumption holds to a good approxi- type of measurement is similar to the osmotic experiments mation. described by Cosgrove and Steudle (1) for the growing pea The experimental arrangement is shown in Figure 2. It was epicotyl, but is not complicated by solute diffusion, effects of essential that the half-time of the cell turgor response (t l/2 ) was unstirred layers, or by the growth of the tissue. controlled by the hydraulic conductance ofthe tissue and not by Experiments were designed so that the hydrostatic pressure of that of the xylem or clamp. Figure 4 shows that the hydraulic 188 WESTGATE AND STEUDLE Plant Physiol. Vol. 78, 1985 Table III. Calculated Time Constants, t,, for Loading the Midrib Xylem Vessels with Water t, was estimated from the hydraulic resistance of the xylem (R.) and the xylem water capacitance (VI/(Et + 7r). Rx was obtained by measuring stationary flows, Jvx, across the leaf from the cut end within the clamp (cf Fig. 2) to (a) the point at which turgor was measured and (b) the point at which the leaf protrudes from the clamp. VA was estimated from the total cross-sectional area of the xylem vessels within the five major .4 bundles of the midrib (Fig. 1) for each leaf. E, + ir. was assumed to be 100 bar. I0 0 Volume of M' midri E Conductance of u Leaf Midrib Vessels MVdsbtsx ¢x ao cm3 .s' bar' cm3 ms 11 I a 1.48 6.40 43 b 10.0 2.32 2 2 a 1.11 5.09 46 b 10.3 2.04 2 CE 3a 1.90 5.92 31 -0 21.0 b 27.3 2.41 1 a)0 4 a 2.5 5.90 24 b 25.0 2.37 1 5 a 2.48 7.90 32 b 18.9 2.80 2

I I Stationary Pressure at the Clamp (bbar) 112- . N * *,,.l. . FIG. 4. Dependence of stationary water flow (JvW in the midrib vessels of an excised maize leaf on the hydrostatic pressure applied at the I10 clamp (Fig. 2). The leaf was cut at two positions to open the vessels. (0,

U, A, V, *), Measurements in which the vessels were opened at the point P0p where turgor pressures were measured (about 7 cm from the cut end, Fig. 2). (0, 0, A, V, 0), Measurements in which the vessels were opened PE right at the clamp (about 2 cm from the cut end). Data shown were obtained from five different leaves. Note that the hydraulic resistance of the alone is much smaller than the total hydraulic resistance of I- clamp a the xylem system. _ 0 40 P 0 *'0l 1 30 -,0 0- 20 40 60 'O 10 20 30 40 conductance of the xylem (AJv./AP.) was much higher across 11 B than across the entire leaf xylem up t~--l1s v=18s f;2=18sB the clamp (open symbols) 10 to the of turgor measurement (closed symbols). This indi- point 0~ cates that sealing the leaf to the perspex clamp did not limit the 9 I flow of water in the vessels. Furthermore, the hydraulic con- a.. 8 Cell TuWgor Presure ductance of the xylem up to the point of turgor measurement 0 also was low. The apparent decrease in conductance is due, in 11 part, to a decrease in the diameter of the xylem vessels toward the leaf apex. The hydraulic conductance of the major xylem in the midrib, estimated by r,x4 according to Hagen- 0 vessels Pressure Poiseuille's law (2r. = vessel diameter), decreased by about 30% between the clamp and the point of turgor measurement. The -1 axial conductance of the vessels per cm length ranged from 0.8 to 2.0 x 10-3 cm4's-'.bar-' which is in good agreement with literature values (2, 3). Using the measured xylem conductance, a time constant (t.,) for loading the vessels with water can be 0 120 240 360 480 600 720 840 960 1080 estimated provided that the water capacity of the xylem (Cx = Time, t(s) V,/[e, + lx,]) is known since t1, = Rx. C. Assuming eA + wr, = 100 bar, t, = 24 to 46 ms (Table III). This calculation of t, FIG. 5. A, Traces ofAP/A Vfor estimation of e and pressure relaxation represents an upper limit because E, + 7r. = 100 bar would be an curves for an individual epidermal cell on the adaxial surface of the extremely low value considering the structure of the xylem and midrib (zone III). B, Time course of the change in cell turgor pressure to an instantaneous in pressure (Ps) for estimates of E) in the literature (4). Therefore, the time required (P) in response change xylem for the propagation of a pressure pulse within the xylem should the same cell as in A. The cell was approximately three to four cell layers nearest Note that the half-time for water be negligible compared with t l- or T,/ provided that there are distant from the xylem. exchange for the tissue no air bubbles in the system. (T1/2) ofthis cell is much shorter than the halftime response Figure 5 shows the response ofcell turgor to a change in xylem (t1',/2). WATER TRANSPORT IN PLANT TISSUES 189 pressure for an epidermal cell located near the edge ofthe midrib tissue (zone III) approximately three to four cell layers distant from the nearest xylem vessel. t',I ranged from 15 to 18 s and was similar for P- > 0 and P. < 0 (Fig. SB). It is evident that t'112 reflects the tissue response to APJ since t,/ >> T112 for this cell (Fig. 5A). If the change in water potential of the xylem is propagated across the tissue according to a diffusion type of kinetics, then the response time of individual cells at different positions within the tissue should vary in direct proportion to the square of the distance of the cell from the peripheral vessels. That is, the overall diffusivity of the tissue to water calculated from the cell response times (D,') should be similar regardless of the position of the cell provided that the tissue is uniform. To test this, the tl/2 for cells at various positions within the midrib tissue were measured. This test was performed only for cells in zones II and III because these regions were fairly homogeneous (Fig. 1 and Table I) and x2 should be valid. tl/2 L- L- Figure 6 (A-C) shows that, indeed, t /2 values are larger for a o cells farther from the xylem. t 1/2 increased from 1 I to 50 s as the m position of the cell in the tissue (X/R) increased from 0.23 to a.0 x 0.78. The longer half-times required to reach a new steady state for cells further from the xylem was characteristic of the tissue a- and not the individual cells since T,2 values for these cells ranged I from 1.0 to 2.0 s (Fig. 7, A-C) which was much shorter than 0 t'l2 in each case. Results for a number of cells within the midrib a~. are a- presented in Table II. The tissue diffusivities (D,') calculated L- from these l,,2 values ranged from D,' = 23.1 to 48.1. 10-6 cm2* 0 E s-'. Thus D,' for zones II and III was fairly uniform indicating that the propagation of the signal from the xylem across the im x midrib tissue followed a diffusion type of kinetics. However, D,' L. was I order of magnitude larger than D. or D, (Table II) suggest- I-@ ing that water transport occurred via a different mechanism (or along a different path) during the sorption and pressure-propa- gation experiments. Pressure-Perfusion Experiments and Midrib Hydraulic Con- ductance. The discrepancy between D,' and D, indicated that the hydraulic conductance of the tissue was considerably higher when a hydrostatic pressure gradient was directly applied to (and may have been present in) the tissue. Therefore, an experiment was designed to estimate the hydraulic conductance ofthe midrib tissue directly in the presence of a hydrostatic pressure gradient. Stationary water flows were induced across a defined region of 0 120 240 360 480 600 720 80 960 midrib tissue using the experimental arrangement shown in Time. t (s) Figure 3. The pressure-induced flows were quite large (range, 2 FIG. 6. Response of cell turgor pressure (P) to an instantaneous to -4. 10-5 cm3-s-') even for small pressure gradients (range, change in xylem pressure (Ps) for three parenchyma cells in zones II and +0.1 to -0.3 bar). The resulting hydraulic conductance was Lp,, III of the midrib. Cells in A to C were located at increasing distances = (8.9 ± 5.6). 10-5 cm - s-' -bar-' (n = 5) when the conductance from the xylem vessels. The relative position of each cell within the was expressed per cm2 of cut area in the channel (Fig. 3). midrib (X/R; X = distance from the vessel, R = radius of midrib) was Although there was considerable variability in Lp,l, it is evident determined from the depth of the tip of the probe which was inserted that Lpm is larger than the individual cell Lp. from the adaxial side of the leafblade. A, X/R = 0.230; B, X/R = 0.362; C, X/R = 0.776. Note that the response time (t,12) increased with DISCUSSION increasing distance from the xylem. Several techniques were combined to elucidate the mecha- of water moving along a cell-to-cell path. This is in agreement nism(s) of water transport across the parenchymatous tissue in with earlier studies comparing the diffusivity of the cell-to-cell the midrib of the maize leaf. A new technique (turgor pressure path with that of cell-to-cell and apoplasmic paths combined (1, propagation) was developed to study the propagation of changes 13, 15). These estimates of D, and D. assume that the tissue is in cell turgor pressure across this tissue. The results show that composed of uniform cells having similar water exchange char- the propagation of turgor (or water potential or volume) across acteristics. However, we observed a zonal variation in T112, Lp, the midrib follows a diffusion type of kinetics. These direct and E (as well as 2r) within the midrib tissue. The parenchyma measurements at the cellular level support the general theoretical cells near the adaxial surface of the midrib had a longer T112, a predictions for the kinetics of water transport in plant tissues (9, lower Lp, a larger E, and a smaller diameter on average than cells 11). near the abaxial surface close to the peripheral vascular bundles The diffusivity of the midrib parenchyma tissue estimated (Table I). The fact that D, was smaller than D, for zones II and from sorption kinetics (D,) was similar to that estimated from III but larger than D, for zone I (Table II) probably reflects the the water relations parameters of individual cells (D.). This effect of the cells in zone I on the overall D, for the tissue. These indicates that during rehydration there was a substantial amount results indicate that tissue inhomogeneities for the water relations 190 WESTGATE AND STEUDLE Plant Physiol. Vol. 78, 1985

I I I %-"" I I I transport must be dominating during pressure-driven flow. Steu- 11- A]- dle and Boyer (13) suggested that under the conditions of the ap Ul ns A 10 'sec:I? Tui2=12s T',2=t2s sorption experiments, the gradient in water potential within the apoplasm was mainly osmotic due to the solutes dissolved in the 9 A wall water. They argued, however, that a gradient in osmotic pressure within the wall would be ineffective in driving water 8 transport because of the low reflection coefficient of the wall O7 1I v material (u<,.< 1). Using the approach of Molz and Ikenberry (9), and assuming that only osmotic gradients are present and CC,. is constant, D, is given by (13):

.8 Ax 6. Tu.1.0sTi,2=2.Os AX(a,.-Lp,:w-ac:, + - Lp-aj 0Ub5 D,= + C.(6) A ~ ~ c in analogy to equation 2. The ratio (p) between the amount of water transported in the apoplasm and the amount transported from cell-to-cell is then given by: a,= 2,,.* Lp. a,. LO6 (7) Ax/2*Lpa,& I ' ' v v . I

A 40 Thus, if a¢c << 1, the wall component becomes diminishingly small (13). Ti,2=1.~sTTme,=t.6se1,Tt(1ss The hypothesis that the hydraulic conductance of the tissue (i.e. the dominant path for water transport) depends on the physical nature of the driving force in the tissue also involves matric forces. At low ip, water will be drawn from the surface layers and largest pores of the cell wall. It could be argued that this would result in a lower Lp,. and consequently a lower D, in the dehydrated tissue than in the fully hydrated state. If so, one FIG. 7. Pressure relaxation curves for the three parenchyma cells in would expect a nonlinear diffusion kinetic (an increase in D,) Figure 6. Note that the half-times for water exchange (T112) for all three during rehydration. However, this was not observed in the sorp- cells were much shorter than the half-time for the tissue response (11/2) tion experiments. At high p,,., as in the pressure-propagation (at shown in Figure 6. P, > 0) and pressure-perfusion experiments, matric forces in the wall would be negligible. This is supported by the observation of parameters T112, Lp, and exist and that they can have an effect Steudle and Boyer (13) that infiltration of hypocotyl segments of on the kinetics of water transport across the tissue. soybean seedlings with water did not affect Lp,. Therefore, while The diffusivity of the midrib tissue estimated from pressure the development of matric potentials and changes in Lp,,. cannot propagation measurements (D,') was more than order of mag- be excluded, these results support the conclusion that water nitude larger than D,. In addition, the hydraulic conductance of transport in the apoplasm driven by gradients in matric potential the midrib tissue (in the radial direction) estimated from pres- is small compared to that driven osmotically along a cell-to-cell sure-perfusion measurements (Lpm) was much larger than Lp. path during rehydration. These results are incompatible with the cell-to-cell transport of Water transport in the apoplasm may occur along two parallel water and suggest that, under the conditions of the pressure paths, i.e. a component within the wall material and a component propagation and pressure perfusion experiments, water move- within or along the interconnected intercullar air spaces. If we ment occurred primarily in the apoplasm. In both cases, a assume that all the water moves in the walls, we can calculate an gradient in hydrostatic pressure was imposed across the tissue. Lp,CH for the wall material since (13): The gradient was transient during pressure-propagation but was stationary during perfusion. This is in contrast to the sorption + Lp LPM YC*Lp. 'y,.- (8) experiments in which a gradient in osmotic or matric potential d 2d/Ax likely developed in the apoplasm of the dehydrated tissue. In all experiments, the driving force for water transport is the gradient and -yc denote the mean fractions of the cross-sectional area in total water potential imposed across the tissue. This gradient available for apoplasmic and cell-to-cell transport, respectively must be similar in magnitude and direction in the apoplasm and ('yc + yc,, = 1; yec,,. << 'y,). d is the tissue thickness and Ax is the from cell-to-cell. However, the flow of water in the apoplasm cell diameter. For typical experimental values (Lpm = 5.5. 10- apparently varied with experimental conditions. This suggests cm s I-bari', y... = 0.025, d = 350 ,um, Ax = 10 ,um, and Lp that not all components of the total water potential are effective = 2.2. 10- cm-s'-bar'), Lp,.. = 1.2. 10-4 cm2. s'- bar7'. This in moving water through the apoplasm and that the dominant value is 3 to 4 orders of magnitude larger than literature data for pathway for water transport depends on the physical nature Lpcw (10-8 to 10-7 cm2-s'Ibar' [9, 13]) which suggests that (hydrostatic, matric, or osmotic) of the driving force. water transport within or along the intercellular air spaces also In part, these results are similar to those of Steudle and Boyer occurred in the presence of hydrostatic pressure gradient. This ( 13) for the radial water transport in growing segments of soybean has been suggested for the soybean hypocotyl, where Lp,4. was hypocotyls. They found that D, was similar to D, and concluded estimated to be Lp., = 8. 10-' cm2.s ' *bar-' (13). that cell-to-cell transport was dominating during hydration. Air spaces 0.5 to 9 Mm in diameter were observed in freehand However, the radial conductance of the tissue to water, Lp, cross-sections of midrib tissue. The number of pores per cm2 in (which is equivalent to our Lpm and measured using the pressure- the radial direction estimated from longitudinal sections was N perfusion technique) was much too high to be interpreted by the = 2.5. 104. If we assume an average minimum pore diameter of same mechanism. Therefore, they concluded that apoplasmic 2rp = 1 utm and a pore length of lp = 700 Mm which is twice as WATER TRANSPORT IN PLANT TISSUES 191 long as the tissue thickness because of tortuosity, the hydraulic LITERATURE CITED conductance of the pore path (Lp,,) can be calculated provided 1. COSGROVE DJ, E STEUDLE 1981 Water relations of growing pea epicotyl all other components are negligible. From Hagen-Poiseuille's law segments. Planta 153: 343-350 we have: 2. GIORDANO R, A SALLEO, S SALLEO, F WANDERLINGH 1978 Flow in xylem vessels and Poiseuille's law. Can J Bot 56: 333-338 N. w 3. GREACEN FL, P PONSANA, KP BARLEY 1976 Resistance to flow in the roots of Lp,, = 8 (9) cereals. In OL Lange, L Kappen, ED Schulze, eds, Water and Plant Life, 8 11ep Ecological Studies, Vol 19. Springer-Verlag, Heidelberg, pp 86-100 where n is the viscosity ofwater (n = 8.95 x 10-9 bar s at 25C). 4. HAMMEL HT 1967 Freezing ofxylem sap without cavitation. Plant Physiol 42: cm 55-66 Using the values given above, Lp,) 1 110-4 s-' *bar-' which 5. HUSKEN D, E STEUDLE, U ZIMMERMAN 1978 Pressure probe technique for is within the range ofexperimental values for Lp,E. This compar- measuring water relations of cells in higher plants. Plant Physiol 61: 158- ison suggests that the high Lp,, values measured during pressure- 163 perfusion may be due in part to transport of water in the 6. JOsT W 1960 Diffusion in Solids, Liquids, Gases. Academic Press, New York intercellular space. However, such a conclusion should be taken 7. MICHELENA VA, JS BOYER 1982 Complete turgor maintenance at low water potentials in the elongating region of maize leaves. Plant Physiol 69: 1145- cautiously since minimum pore diameters and pore continuity 1149 (in the radial direction) could not be determined. In the intact 8. MOLZ FJ, JS BOYER 1978 Growth induced water potentials in plant cells and aerated tissue, water may flow as a film on the wet wall surfaces tissues. Plant Physiol 62: 423-429 around the air spaces. This type of flow also would have a fairly 9. MOLZ FJ, E IKENBERRY 1974 Water transport through plant cells and cell low resistance and may be pressure dependent (13). If so, this walls: Theoretical development. Soil Sci Soc Am Proc 38: 699-704 10. PFEFFER W 1897 Pflanzenphysiologie, Vol 1. Verlag Engelmann, Leipzig may explain the finding (for both the maize leafand the soybean 11. PHILIP JR 1958 Osmosis and diffusion in tissues: Half-times and internal hypocotyl) that flooding ofintercellular air spaces did not change gradients. Plant Physiol 33: 275-278 Lp,, significantly. 12. SACHS J 1865 Handbuch der Experimentalphysiologie der Pflanzen. Verlag It is interesting that there was no difference in the radial Engelmann, Leipzig hydraulic conductance ofbarley roots (Lpr per cm2 of outer root 13. STEUDLE E, JS BOYER 1984 Hydraulic resistance to radial water flow in growing hypocotyl of soybeans measured by a new pressure perfusion technique. surface) in the presence of a hydrostatic or osmotic gradient as Planta. In press determined using a special 'root pressure probe' (14). Lpr of the 14. STEUDLE E, WD JESCHKE 1983 Water transport in barley roots. Planta 158: barley root was about 10-7 cms-'bar-' and by 1 order of 237-248 magnitude smaller than the cell Lp. This indicated that a sub- 15. STEUDLE E, JAC SMITH, U LUTTGE 1980 Water relation parameters ofindivid- stantial cell-to-cell transport of water occurred in the root cortex ual mesophyll cells of the CAM plant Kalanchoe daigremontiana. Plant Physiol 66: 1155-1163 regardless of the nature of the driving force and suggests that 16. STEUDLE E, H ZIEGLER, U ZIMMERMANN 1983 Water relation ofthe epidermal either the wall Lp(., is very low or that the Casparian strip blocks bladder cells of Oxalis carnosa Molina. Planta 159: 38-45 the system of the interconnected intercellular air spaces in the 17. STEUDLE E, U ZIMMERMANN 1984 Water relations of plant cells: Further barley root. development ofthe pressure probe and oftechniques for measuring pressure- The possibility that the preferred path for water flow across dependent transport. In WJ Cram, K Janacek, R Rybova, K Sigler, eds, Membrane Transport in Plants. Academia, Prague, pp 73-82 plant tissues could depend on the nature of the driving forces 18. STRUGGER S 1938 Die lumineszenz-mikroskopische Analyse des Transpira- present has general consequences for the water relations of plant tionsstromes in Parenchymen. Flora (Jena) 133: 56-68 tissues and may be important for the control of water flows 19. TANTON TW, SH CROWDY 1972 Water pathways in higher plants. J Exp Bot within the plant (e.g. during water uptake into roots, radial water 23: 600-625 transport in stems, and also for the water transport across the 20. TOMOs AD, E STEUDLE, U ZIMMERMANN, ED SCHULZE 1981 Water relations of leaf epidermal cells of Tradescantia virginiana. Plant Physiol 68: 1135- different leaf tissues during transpiration). More direct evidence, 1143 however, is necessary to support this hypothesis. These experi- 21. WESTGATE ME, JS BOYER 1984 Transpiration- and growth-induced water ments should incorporate measurements with osmotic gradients potentials in maize. Plant Physiol 74: 882-889 set up across tissues and more quantitative measurements to 22. ZIMMERMANN U, E STEUDLE 1978 Physical aspects of water relations of plant evaluate the effects of intercellular air spaces on water transport. cells. Adv Bot Res 6: 45-117 23. ZIMMERMANN U, E STEUDLE 1980 Fundamental water relations parameters. Acknowledgments-The authors are grateful to Dr. J. Wieneke, Institut fur In RM Spanswick, WJ Lucas, J Dainty, eds. Plant Membrane Transport: Radiogranomie, KFA Julich, for suggesting the use of the liquid silicone paste as a Current Conceptual Issues. Elsevier/North-Holland Biomedical Press, Am- sealing material and to H. Jackel for growing the plants. sterdam, pp 113-127