Planta
Planta (1989) 177:281-295
9 Springer-Verlag 1989
Osmotic responses of maize roots Water and solute relations Ernst Steudle and Jiirgen Frensch Lehrstuhl f/Jr Pflanzen6kologie, Universit/it Bayreuth, Universit/itsstrasse 30, D-8580 Bayreuth, Federal Republic of Germany
Abstract. Water and solute relations of excised seminal roots of young maize (Zea mays L) plants, have been measured using the root pressure probe. U p o n addition of osmotic solutes to the root medium, biphasic root pressure relaxations were obtained as theoretically expected. The relaxations yielded the hydraulic conductivity (Lpr), the permeability coefficient (Pj , and the reflection coefficient (a~r) of the root. Values of Lpr in these experiments were by nearly an order of magnitude smaller than Lpr values obtained from experiments where hydrostatic pressure gradients were used to induce water flows. The value of P~r was determined for nine different osmotica (electrolytes and nonelectrolytes) which resulted in rather variable values (0.1.10 . 8 - 1.7.10 .8 m.s-1). The reflection coefficient as~ of the same solutes ranged between 0.3 and 0.6, i.e. a~r was low even for solutes for which cell membranes exhibit a ~s ~ 1. Deviations from the theoretically expected biphasic responses occurred which may have reflected changes of either P~r or of active pumping induced by the osmotic change. The absolute values of Lp~, P~r, and a~ have been critically examined for an underestimation by unstirred layer effects. The data indicate a considerable apoplasmic component for radial movement of water in the presence of hydrostatic Abbreviations and symbols: A t = r o o t surface area; Cm (Cx) =solute concentration in the meduum (xylem); D s (D e) =diffusion coefficient in the stele (cortex); Jvr=radial water transport across the root; kwr (k~)=rate constant of water (solute exchange between root xylem and medium; Lp. = root hydraulic conductivity; P~= root pressure; P~ = permeability coefficient of root for solute " s " ; T~/2 (T~S/2)=half-time of water (solute) exchange between root xylem and medium; Vs(V~) = volume of system (xylem); c5~ (~) = thickness of unstirred layer in the stele (cortex); ~s (e~)= elastic modulus of measuring system (xylem); a~,=reflection coefficient of root for solute "'s"; =flow constriction factor; superscripts " e n " and " e x " denote flows from the medium into the root xylem or from the xylem into the medium, respectively
gradients and also some solute flow by-passing root protoplasts. In the presence of osmotic gradients, however, there was a substantial cell-to-cell transport of water. Cutting experiments demonstrated that the hydraulic resistance for the longitudinal movement of water was much smaller than for radial transport except for the apical ends of the segments (length = 5 to 20 mm). The differences in Lpr as well as the low asr values sugge, st that the simple osmometer model of the root with a single osmotic barrier exhibiting nearly semipermeable properties should be extended for a composite membrane model with hydraulic and osmotic barriers arranged in series and in parallel.
Key words: Hydraulic conductivity - Reflection coefficient - R o o t permeability - R o o t pressure Water relations - Zea (water relations)
Introduction The hydraulic resistance of the root is an important factor in the water supply of the shoot a n d in the acclimation of plants to conditions of water shortage. Next to the stomata, roots usually exhibit the largest resistance to water flow in the soil-plantatmosphere continuum and, like the stomata, roots show some variability of this resistance. The reasons for the variability are not completely understood, but seem to be related to a coupling between water and solute uptake into the root Which may be described by the formalism of irreversible thermodynamics (Dalton et al. 1975; Fiscus 19175). This approach suggests that the root (analogous to the cell) may be treated as a two-compartment system with the xylem as the internal and the s0il solution as the external compartment, separatediby a membrane-like structure which is usually identified by
282
the endodermis. In fact, there is a large body of evidence that the osmotic barrier in the root has membrane-like properties such as selectivity which is a prerequisite for a proper functioning of the root. However, the detailed mechanisms for the uptake of solutes and water as well as the interactions between water and solutes in the root are still under debate. The development of root pressure under conditions of low transpiration provides the best evidence that roots (analogous to cells) behave like osmometers in the sense of the simple two-compartment model. Although solutes are pumped actively into the xylem and there is a passive leak across the endodermis, roots are usually looked at as rather perfect osmometers with reflection coefficients (a~r) close to unity, which would indicate a nearly semipermeable membrane (Pitman 1982; Weatherley 1982; Kramer 1983; Dainty 1985; Fiscus 1986; Passioura 1988). However, there is also evidence that this picture may not be completely correct and that the simple, equivalent membrane, model of the root has to be extended. This evidence comes from the fact that there are apoplasmic by-passes in the root for water and solutes and that different root zones could contribute differently to the overall transport (Sanderson 1983; Hanson et al. 1985; Yeo et al. 1987). The situation may, thus, be complicated if there is more than one osmotic barrier arranged in series or in parallel, such as the epidermis/hypodermis complex and the endodermis (Perumalla and Peterson 1986; Clarkson et al. 1987), or the different pathways (apoplasmic, symplasmic, and transcellular path) or root zones. This complex composite structure of the root may cause deviations from the linear relations between flows and forces usually assumed, as has been shown by the basic treatment of membrane transport processes using the principles of irreversible thermodynamics (Kedem and Katchalsky 1963a, b; see House 1974; Zimmermann and Steudle 1978). Recent measurements of water and solute flows using the root pressure probe technique also indicate that the traditional view of roots as perfect osmometers should be extended to allow for the complications mentioned above (Steudle and Jeschke 1983; Steudle 1987a, b; Steudle et al. 1987; Steudle and Brinckmann 1989). In these experiments, the water and solute permeability as well as reflection coefficients could be obtained from measurements of the root pressure in excised roots of barley, maize, and bean. In maize roots, a large apoplasmic flow of water was indicated when hydrostatic gradients were applied in contrast to osmotic gradients (Steudle et al.
E. Steudle and J. Frensch: Osmotic responses of maize roots
1987). The finding was different from results obtained for barley and bean where flows were independent of the nature of the driving force, and water predominantly moved from cell to cell (Steudle and Jeschke 1983; Steudle and Brinckmann 1989). Moreover, for bean roots the solute permeability was rather low as compared with maize. These results indicated differences in the transport mechanisms for water and solutes depending on the species and the type of experiment used. For all species, however, the root reflection coefficients (asr) were substantially smaller than unity even for solutes for which cell membranes exhibit a as ~ 1. In this paper these experiments have been further extended to obtain more evidence for the composite membrane model of the root. Since unstirred layer effects could have artificially affected the absolute values of transport coefficients and, for example, could have yielded reflection coefficients that were too low (Dainty 1985), these effects have been thoroughly investigated. In order to vary water flows, pressure relaxations as well as pressure clamp experiments have been performed (Wendler and Zimmermann 1982). Furthermore, in root cutting experiments the contribution of the longitudinal (axial) to the overall hydraulic resistance has been estimated using the root pressure probe.
theory
Calculation of transport coefficients (Lpr, ~r, P~r) from root pressure probe experiments. In an excised root attached to the root pressure probe the accumulation of nutrients in the root xylem provides the driving force for water uptake and for the development of a root pressure (Fig. 2). However, there are two important differences between this system and an ideal osmometer. Firstly, the excised root is not impermeable to solutes and (at z e r o water flow) maintains a steady state (stationary root presure) which is dependent on metabolic energy (pump/leak system) rather than a thermodynamic equilibrium as in the case of an osmometer. The second difference is that the "osmotic barrier" in the root is complex and does, perhaps, not only comprise the endodermis. For water and solutes, there are different permeation pathways in series (rhizodermis, cortex, stele) and in parallel (apoplasmic, symplasmic, and transcellular path) which may be used. Furthermore, it has to be considered that during root development there may be regions of different permeability in the root and, thus, parameters such as the hydraulic conductivity (Lpr), the solute permeability coefficient (PJ , and the re-
E. Steudle and J. Frensch: Osmotic responses of maize roots
283
flection coefficient (a~r) of a root or root segment reflect average properties which, for example, could depend on the growing conditions. This has to be kept in mind when the root is treated as a simple osmometer consisting of two well-stirred compartments (medium and xylem) separated by a homogeneous membrane. In spite of these difficulties, the root pressure probe can be used to perform hydrostatic and osmotic experiments on excised roots which are analogous (though not identical) to those performed on plant cells using the cell pressure probe (Hiisken et al. 1978; Steudle et al. 1983; Steudle and Zimmermann 1984). A rigorous treatment of the processes based on a two-compartment model has been given elsewhere (Steudle and Jeschke 1983; Steudle et al. 1987). Here, briefly summarized, are the basic equations which have been used to calculate transport coefficients. In the experiments, either the root pressure or the osmotic pressure of the medium were changed and the re-establishment of a new stationary root pressure was measured. From the root-pressure relaxations obtained, the hydraulic conductivity of the root (Lpr) can be evaluated. Experiments in which water is taken up by the roots ("endosmotic experiments") have to be treated in a somewhat different way from those in which water flow is from the xylem into the medium ("exosmotic experiments"). It has been readily shown in the aforementioned papers that in the presence of impermeable solutes the rate constant of water exchange between xylem and medium during relaxations (kwr) will be given by: APr
as
(1)
where Ar=surface area of the root; es=elastic modulus of the measuring system; Vso=volume of the measuring system. A P~ and A Vs are the changes in root pressure (P~) and in the volume of the measuring system, respectively. The term f corrects for concentration changes in the xylem. Strictly, f should be zero for exosmotic and f > 0 for endosmotic experiments under ideal contitions, i.e. with well-stirred xylem solution. However, in the real situation the concentration in the xylem may be polarized and should vary along the root, and this (besides other factors) could result in deviations which could be accounted for by using the empirical parameter f which can be determined experimentally in the hydrostatic experiments (for details see Steudle et al. 1987). In the presence of permeating solutes, biphasic relaxations similar to those which have been measured with individual plant
cells (Tyerman and Steudle 1982; Steudle and Tyerman 1983) are obtained in osmotic experiments. The pressure-time (volume-time) courses in these experiments will be given by: Vs- Vso
P~- Pro
Vso
~S
Lpr" Ar"O-st"RTCms =
Vso(kwr~ksr)
[exp(-/%~. t)
- exp ( - k~r"t)],
(2)
where Vso and Pro are the volume of the measuring system and the stationary root pressure at the beginning of an experiment, respectively. Cr.s =concentration of solute "s" in the medium; ksr = r a t e constant for solute uptake across the root cylinder; R=universal gas constant; T=absolute temperature. Equation (2) describes a biphasic process. Upon the addition of the osmotic solute to the medium, root pressure first decreases because of the reduction of the water potential of the medium ("water phase"). However, when a mimimum in root pressure (Prr,in) has been reached, ~ increases again due to the uptake of solutes, which reduces the water potential of the xylem, and water then follows ("solute phase"). This process should proceed until the original root pressure (Pro) is reattained and Cms = Cxs. On the other hand, when a root has taken up a permeating solute and is then treated with the original medium, a biphasic response should occur with a root-pressure maximum. The second phase of the biphasic relaxations will be governed by the passive uptake of solutes into the root xylem, and the rate constant (k~r) of this process will be: ksr
Ar'Psr
Vx
(3).
Thus, by measuring ksr and determining A r a s well as the volume of the xylem (Vx), the permeability coefficient of the root to a given solute (P~r) can be evaluated. Reflection coefficients of roots (%) can be evaluated at J v r = 0 from the maximum changes in root pressure (Pro-- Prmi,(max)) and the corresponding changes of external osmotic pressure (rcs=RT. Cms). Setting dVs/dt, dP~/dt,-O one obtains from Eq. (2): asr =
Pro - Prmin
An~
ax+RT. Cxo am
exp (ksr.
train),
(4)
where ex is the elastic modulus of the xylem and Cxo the stationary solute concentration in the xylem at the beginning of the experiment, tmin denotes
284
E. Steudle and J. Frensch: Osmotic responses of maize roots
the time necessary to reach the minimum. The factor @x+ RT. Cxo)/exon the right side of eqn. (4) corrects for the shrinking and swelling of the xylem during the relaxations (which is usually negligible), and the factor exp (k~.tm~n) for the uptake of permeating solutes during the water phase.
Effects of ion-pumping. It has to be emphazised that eqns. (1) to (4) strictly hold only for passive water and solute movements and for the simple model outlined above. Complications may occur if the rate of active transport (ion-pumping) is changing (e.g., by osmoregulatory responses of the root). In this case, the approach may be still used provided that the changes in P~ due to these changes are sufficiently slow (i.e., if their half-times are much larger than those of the water and passive solute flows). For water, this may generally be true, but for slowly permeating solutes, half-times could become comparable. In this case, P~r (ksr) may be estimated for the maximum slope of the solute phase, if there is a sufficient lag time for the onset of active transport processes. For this time period, it can be shown that the maximum slope of the solute phase will be obtained at t=2"tmi n and that the solute flow at that time will be given by (Steudle unpublished):
dPr~_ __O.sr.RT. Jsr ~ = (Pro- P~mi,)exp ( -- k~. train) 9ksr.
(5)
As in eqn. (4), the exponential term in eqn. (5) denotes the correction for solute flow occurring between minimum and maximum slope. Equation (5) is valid for exosmotic experiments, but should also hold for endosmotic experiments, if Prmi. is replaced by P~max.
Effects of unstirred layers. Unstirred layer effects at the osmotic barrier(s) in the root could be important and could result in an underestimation of Lpr, P~r, and as~ (Dainty 1985; Steudle et al. 1987). Two types of effects have to be considered, the sweepaway effect and purely diffusional effects. In the sweep-away effect, the water flow across the osmotic barrier changes the concentration right at the barrier (Cb~) in that solutes are swept away on one side and are concentrated on the other by the convective flow of water. The effect is opposed by a back-diffusion of solutes. For a steady water flow (Jvr), the relative change in concentration on each side of the membrane (barrier) will be given by (Dainty 1963):
Cbs = exp ( - Jvr" 6/D~), Cms
(6)
where 6 is the thickness of the unstirred layer and D~ the diffusion coefficient of the solutes. Usually, there is some uncertainty in estimating 6 and Ds in tissues. In the experiments in this paper, the unstirred layer on the outer side of the endodermis could, in principle, be as thick as the entire cortex and D~ would be reduced as compared with the free solution. Furthermore it is possible that water flow could be more or less confined to certain pathways or root zones (e.g., to the apoplast or to the absorption zone) and, hence, the local Jw could be substantially inreased by flow constriction. In this case, Jw/Cb has to be used in eqn. (6) instead of Jvr where ~b is the flow-constriction factor (Barry and Diamond 1984). Purely diffusional unstirred-tayer effects ("gradient-dissipation effects"; Barry and Diamond 1984) arise, whenever the diffusional resistances of unstirred layers next to a membrane or osmotic barrier become comparable to that of the barrier itself. If we assume for the root that the endodermis is the only barrier and that D~ will be the same in the cortex and in the stele, we get (considering the cylindrical geometry of the root) in the presence of stelar and cortical unstirred layers (see Appendix A): 1
-
1
E
R
4- D ~ l n a ,
(7)
where 1/P~ ea~ is the measured and l/P~r the true permeation resistance per unit area. R = root radius; E = radius of endodermis; (E-a)= thickness of unstirred layer in the stele. Equation (7) assumes that the entire cortex is functioning as an external unstirred layer. Using reasonable stimates of D~ and a as well as measured values of R and E it is, thus, possible to estimate the influence of gradient-dissipation effects on measurements of P~r, although (different from the techniques used in this paper) eqn. (7) again refers to steady-state conditions. Both gradient-dissipation and sweep-away effects could be also important for the evaluation of Crsr, although the latter effects should be small in the experiments because O-~r was estimated at Jw = 0. In the absence of sweep-away effects and of frictional interactions between solute and Water flow, the approach of Hingson and Diamond (1972) may be used to correct the measured o-~ for unstirred layers. For the cylindrical geometry of the root we get (Appendix A):
E. Steudle and J. Frensch: Osmotic responses of maize roots
1/P r + E/Ds ln(R/a)
285
(8)
Material and methods Plant material. Maize caryopses (Zea mays L. cv. B73 x Mo17; Siidwestdeutsche Saatzucht, Rastatt, FRG) were sterilized by treating them with 1% NaOCt solution and were germinated for 2-3 d at 25~ in the dark on wet filter paper. By that time, the main roots of the seedlings were 30-50 mm long and could be transferred to hydroculture for 3-9 d (1/4 Johnson solution; for composition, see Steudle et al. t987). The plants used in the experiments developed homorrhizic root systems from which the end segments of seminal roots were excised at a length of 38 110 mm (root diameter: 0.7-1.1 ram). The cross-sectional area of the total xylem (Fig. 1) was about 12% of the root cross-section (Steudle et al. 1987). However, only the protoxylem and the early metaxylem vessels should have been mature and functioning in the roots used and not the
Fig. 1. Cross section of the seminal root of a 6-d-old maize seedling. The section was taken at a distance of about 80 mm behind the root apex. Areas of late (l) and early (e) metaxylem were estimated to be 9 and 3% of the total area of the cross section. Note that the endodermis is interrupted by a secondary root initial (r,i.)
big elements of late metaxylem (St. Aubin et al. 1986) and, therefore, only this effective part of the xylem, which was 3% of the cross section of the root, has been used to estimate /~ in the calculation of Psr [Eq. (3)]. Measurement of root pressure and of water and solute relations of roots. Root pressures (Pr), hydraulic conductivity per m 2 of outer root surface area (Lp~), permeability (P~r), and reflection (asr) coefficients were measured as described previously (Steudle and Jeschke 1983; Steudle et al. 1987). Briefly, the root segments were tightly connected to the root pressure probe (Fig. 2) via seals which had been prepared from liquid silicone material (Xantopren plus from Bayer, Leverkusen, FRG) in order to adapt the inner diameter of the seal to individual roots. The length of the seals was 8 mm to provide a sufficient sealing area around the root and, at the same time, avoid large local compressive forces which might have constrained xylem vessels. The proper functioning of the seals was checked after each experiment by cutting the root right at the seal (see below). It was ensured that the half-times of pressure relaxations decreased drastically after the cut. This finding proved that the hydraulic resistance of the sealing area was negligible as compared with that of radial water movement across the root cylinder. In hydrostatic experiments, water flows were induced by moving the meniscus either forward to increase the pressure in the system (exosmotic water flow) or backward to reverse the flow (endosmotic water flow). In the case of the osmotic experiments, the original nutrient solution was rapidly exchanged for media containing test solutions at known concentrations which were determined cryoscopically (Osmomat 030 from Gonotec, Berlin, West Germany). Solutions Were circulat-
286
E. Steudle and J. Frensch: Osmotic responses of maize roots Microscope
,
Pressure horn bee 9 , ,
I~IeTQ L
/
I
I
Pressure ~((/
~
fronsducer
/
End segment of reQt
...........;;~:~:"':~::':":'~::":"'::;':. . . . . . . . . .
screw
L"T" Perfusion chQmber
Sitic0ne seQts
Fig. 2. Root pressure probe for measuring root pressures (P~) and water- and solute-relations parameters of roots. The excised root was tightly connected to a pressure chamber with the aid of a silicone seal so that the root pressure could be built up in the system. Water flows across the root could be induced by either changing the pressure in the pressure chamber with the aid of the metal rod ("hydrostatic experiments") or by exchanging the medium in the perfusion chamber for solutions of different osmotic pressure (" osmotic experiments"; see Figs. 3 and 5). During the experiments, a meniscus, which served as a reference point, was formed within a capillary between the silicone oil in the pressure chamber and the water upon the excised root. With the aid of the meniscus, defined volume changes could be produced in order to determine the elastic extensibility of the measuring system. During the experiments, the tube in which the root segment was fixed was perfused with aerated solution in order to minimize unstirred layer effects
ed along the root segments which were fixed in a narrow glass tube sitting in a bigger one, as shown in Fig. 2. Hydrostatic pressure relaxations and the water phases of biphasic root pressure relaxations in the presence of permeating solutes (see below) yielded k,~, or T,~/2 from which Lp~ was calculated using eqn. (1) of the Theory section. The elastic extensibility of the measuring system (A Vs/AP~)was determined by moving the metal rod in the root pressure probe instantaneously and recording the corresponding A P~. A Vs was calculated from the shift of the oil-water meniscus in the measuring capillary (observed with a stereomicroscope) and the diameter of the capillary ( = 360 gm). The geometric root surface area (At) was estimated from the root length and diameter and Vx from the cross-sectional area of the early metaxylem and protoxylem (see above). In some hydrostatic experiments, Lpr was determined in a constant flow mode, i.e., ,/,~, was kept constant by clamping P~ to a pressure higher than Pro and measuring the resulting water flow from the movement of the meniscus in the measuring capillary (Steudle and Jeschke 1983). In the corresponding osmotic experiments which resembled the pressure-clamp technique used for plant cells (Wendler and Zimmermann 1982), Lpr was calculated from the efflux of water from the root in the presence of an applied osmotic gradient while P, was kept at P~o. In the calculation of Lp r it was assumed that, during the initial phase, the amount of solutes taken up by the root was negligible. In biphasic relaxations where the final (steady) root pressure P,o=~P~o, this deviation could have been due to changes in active transport or to other processes (see Theory section and Discussion). In these cases, the maximum initial solute flow
~l/
g~ ~, S0tution 0utter
(and k~r) was estimated from the maximum slope of the solute phase (eqn. (4)). In all cases, however, where P,e~,P,o, k~,(P~r) values were obtained by measuring the rate constant of the entire solute phase (eqn (2)), i.e., active components were neglected. Reflection coefficients were calculated from the maximum changes in root pressure (Pro--Prmin(max)) and from the changes in the external osmotic pressure according to eqn. (4).
Cutting experiments. Information about the development of the root xylem was obtained from cutting experiments in which roots attached to the root pressure probe were excised from the apex. Sections of length of 5-10 mm were successively cut with a razor blade until a drop in P, indicated that mature xylem was hit. Hydrostatic pressure relaxations performed occasionally between cuts indicated changes in the hydraulic resistance. Since the time constant of the relaxations was given by the product of hydraulic resistance and water capacitance of the system, a change in T/I/2 directly reflected changes in the hydraulic resistance, because the hydraulic capacitance (A Vs/APr) remained constant.
Results Figures 3 and 4 indicate that the roots, basically, responded as would be expected for osmometers in the presence of permeating or non-permeating solutes. In both the hydrostatic and osmotic type of experiments the initial water flows during relaxations (i.e., (dP~/dt)t=o)were linearly related to the driving forces applied across the root. Furthermore, in the osmotic experiments the maximum changes in root pressure were proportional to the changes in the osmotic pressure of the medium (Arc~ eqn. (4)), whereas kw~(T/'~/z) remained the same (see, for example, Fig. 3 D, E). The responses of the roots to permeating solutes were biphasic as predicted by eqn. (2). In the hydrostatic relaxations it was also found that P~e>Pro during exosmotic flow and P~ePre; mannitol) were observed, whereas in others P~o
Planta (1989) 177:281-295
9 Springer-Verlag 1989
Osmotic responses of maize roots Water and solute relations Ernst Steudle and Jiirgen Frensch Lehrstuhl f/Jr Pflanzen6kologie, Universit/it Bayreuth, Universit/itsstrasse 30, D-8580 Bayreuth, Federal Republic of Germany
Abstract. Water and solute relations of excised seminal roots of young maize (Zea mays L) plants, have been measured using the root pressure probe. U p o n addition of osmotic solutes to the root medium, biphasic root pressure relaxations were obtained as theoretically expected. The relaxations yielded the hydraulic conductivity (Lpr), the permeability coefficient (Pj , and the reflection coefficient (a~r) of the root. Values of Lpr in these experiments were by nearly an order of magnitude smaller than Lpr values obtained from experiments where hydrostatic pressure gradients were used to induce water flows. The value of P~r was determined for nine different osmotica (electrolytes and nonelectrolytes) which resulted in rather variable values (0.1.10 . 8 - 1.7.10 .8 m.s-1). The reflection coefficient as~ of the same solutes ranged between 0.3 and 0.6, i.e. a~r was low even for solutes for which cell membranes exhibit a ~s ~ 1. Deviations from the theoretically expected biphasic responses occurred which may have reflected changes of either P~r or of active pumping induced by the osmotic change. The absolute values of Lp~, P~r, and a~ have been critically examined for an underestimation by unstirred layer effects. The data indicate a considerable apoplasmic component for radial movement of water in the presence of hydrostatic Abbreviations and symbols: A t = r o o t surface area; Cm (Cx) =solute concentration in the meduum (xylem); D s (D e) =diffusion coefficient in the stele (cortex); Jvr=radial water transport across the root; kwr (k~)=rate constant of water (solute exchange between root xylem and medium; Lp. = root hydraulic conductivity; P~= root pressure; P~ = permeability coefficient of root for solute " s " ; T~/2 (T~S/2)=half-time of water (solute) exchange between root xylem and medium; Vs(V~) = volume of system (xylem); c5~ (~) = thickness of unstirred layer in the stele (cortex); ~s (e~)= elastic modulus of measuring system (xylem); a~,=reflection coefficient of root for solute "'s"; =flow constriction factor; superscripts " e n " and " e x " denote flows from the medium into the root xylem or from the xylem into the medium, respectively
gradients and also some solute flow by-passing root protoplasts. In the presence of osmotic gradients, however, there was a substantial cell-to-cell transport of water. Cutting experiments demonstrated that the hydraulic resistance for the longitudinal movement of water was much smaller than for radial transport except for the apical ends of the segments (length = 5 to 20 mm). The differences in Lpr as well as the low asr values sugge, st that the simple osmometer model of the root with a single osmotic barrier exhibiting nearly semipermeable properties should be extended for a composite membrane model with hydraulic and osmotic barriers arranged in series and in parallel.
Key words: Hydraulic conductivity - Reflection coefficient - R o o t permeability - R o o t pressure Water relations - Zea (water relations)
Introduction The hydraulic resistance of the root is an important factor in the water supply of the shoot a n d in the acclimation of plants to conditions of water shortage. Next to the stomata, roots usually exhibit the largest resistance to water flow in the soil-plantatmosphere continuum and, like the stomata, roots show some variability of this resistance. The reasons for the variability are not completely understood, but seem to be related to a coupling between water and solute uptake into the root Which may be described by the formalism of irreversible thermodynamics (Dalton et al. 1975; Fiscus 19175). This approach suggests that the root (analogous to the cell) may be treated as a two-compartment system with the xylem as the internal and the s0il solution as the external compartment, separatediby a membrane-like structure which is usually identified by
282
the endodermis. In fact, there is a large body of evidence that the osmotic barrier in the root has membrane-like properties such as selectivity which is a prerequisite for a proper functioning of the root. However, the detailed mechanisms for the uptake of solutes and water as well as the interactions between water and solutes in the root are still under debate. The development of root pressure under conditions of low transpiration provides the best evidence that roots (analogous to cells) behave like osmometers in the sense of the simple two-compartment model. Although solutes are pumped actively into the xylem and there is a passive leak across the endodermis, roots are usually looked at as rather perfect osmometers with reflection coefficients (a~r) close to unity, which would indicate a nearly semipermeable membrane (Pitman 1982; Weatherley 1982; Kramer 1983; Dainty 1985; Fiscus 1986; Passioura 1988). However, there is also evidence that this picture may not be completely correct and that the simple, equivalent membrane, model of the root has to be extended. This evidence comes from the fact that there are apoplasmic by-passes in the root for water and solutes and that different root zones could contribute differently to the overall transport (Sanderson 1983; Hanson et al. 1985; Yeo et al. 1987). The situation may, thus, be complicated if there is more than one osmotic barrier arranged in series or in parallel, such as the epidermis/hypodermis complex and the endodermis (Perumalla and Peterson 1986; Clarkson et al. 1987), or the different pathways (apoplasmic, symplasmic, and transcellular path) or root zones. This complex composite structure of the root may cause deviations from the linear relations between flows and forces usually assumed, as has been shown by the basic treatment of membrane transport processes using the principles of irreversible thermodynamics (Kedem and Katchalsky 1963a, b; see House 1974; Zimmermann and Steudle 1978). Recent measurements of water and solute flows using the root pressure probe technique also indicate that the traditional view of roots as perfect osmometers should be extended to allow for the complications mentioned above (Steudle and Jeschke 1983; Steudle 1987a, b; Steudle et al. 1987; Steudle and Brinckmann 1989). In these experiments, the water and solute permeability as well as reflection coefficients could be obtained from measurements of the root pressure in excised roots of barley, maize, and bean. In maize roots, a large apoplasmic flow of water was indicated when hydrostatic gradients were applied in contrast to osmotic gradients (Steudle et al.
E. Steudle and J. Frensch: Osmotic responses of maize roots
1987). The finding was different from results obtained for barley and bean where flows were independent of the nature of the driving force, and water predominantly moved from cell to cell (Steudle and Jeschke 1983; Steudle and Brinckmann 1989). Moreover, for bean roots the solute permeability was rather low as compared with maize. These results indicated differences in the transport mechanisms for water and solutes depending on the species and the type of experiment used. For all species, however, the root reflection coefficients (asr) were substantially smaller than unity even for solutes for which cell membranes exhibit a as ~ 1. In this paper these experiments have been further extended to obtain more evidence for the composite membrane model of the root. Since unstirred layer effects could have artificially affected the absolute values of transport coefficients and, for example, could have yielded reflection coefficients that were too low (Dainty 1985), these effects have been thoroughly investigated. In order to vary water flows, pressure relaxations as well as pressure clamp experiments have been performed (Wendler and Zimmermann 1982). Furthermore, in root cutting experiments the contribution of the longitudinal (axial) to the overall hydraulic resistance has been estimated using the root pressure probe.
theory
Calculation of transport coefficients (Lpr, ~r, P~r) from root pressure probe experiments. In an excised root attached to the root pressure probe the accumulation of nutrients in the root xylem provides the driving force for water uptake and for the development of a root pressure (Fig. 2). However, there are two important differences between this system and an ideal osmometer. Firstly, the excised root is not impermeable to solutes and (at z e r o water flow) maintains a steady state (stationary root presure) which is dependent on metabolic energy (pump/leak system) rather than a thermodynamic equilibrium as in the case of an osmometer. The second difference is that the "osmotic barrier" in the root is complex and does, perhaps, not only comprise the endodermis. For water and solutes, there are different permeation pathways in series (rhizodermis, cortex, stele) and in parallel (apoplasmic, symplasmic, and transcellular path) which may be used. Furthermore, it has to be considered that during root development there may be regions of different permeability in the root and, thus, parameters such as the hydraulic conductivity (Lpr), the solute permeability coefficient (PJ , and the re-
E. Steudle and J. Frensch: Osmotic responses of maize roots
283
flection coefficient (a~r) of a root or root segment reflect average properties which, for example, could depend on the growing conditions. This has to be kept in mind when the root is treated as a simple osmometer consisting of two well-stirred compartments (medium and xylem) separated by a homogeneous membrane. In spite of these difficulties, the root pressure probe can be used to perform hydrostatic and osmotic experiments on excised roots which are analogous (though not identical) to those performed on plant cells using the cell pressure probe (Hiisken et al. 1978; Steudle et al. 1983; Steudle and Zimmermann 1984). A rigorous treatment of the processes based on a two-compartment model has been given elsewhere (Steudle and Jeschke 1983; Steudle et al. 1987). Here, briefly summarized, are the basic equations which have been used to calculate transport coefficients. In the experiments, either the root pressure or the osmotic pressure of the medium were changed and the re-establishment of a new stationary root pressure was measured. From the root-pressure relaxations obtained, the hydraulic conductivity of the root (Lpr) can be evaluated. Experiments in which water is taken up by the roots ("endosmotic experiments") have to be treated in a somewhat different way from those in which water flow is from the xylem into the medium ("exosmotic experiments"). It has been readily shown in the aforementioned papers that in the presence of impermeable solutes the rate constant of water exchange between xylem and medium during relaxations (kwr) will be given by: APr
as
(1)
where Ar=surface area of the root; es=elastic modulus of the measuring system; Vso=volume of the measuring system. A P~ and A Vs are the changes in root pressure (P~) and in the volume of the measuring system, respectively. The term f corrects for concentration changes in the xylem. Strictly, f should be zero for exosmotic and f > 0 for endosmotic experiments under ideal contitions, i.e. with well-stirred xylem solution. However, in the real situation the concentration in the xylem may be polarized and should vary along the root, and this (besides other factors) could result in deviations which could be accounted for by using the empirical parameter f which can be determined experimentally in the hydrostatic experiments (for details see Steudle et al. 1987). In the presence of permeating solutes, biphasic relaxations similar to those which have been measured with individual plant
cells (Tyerman and Steudle 1982; Steudle and Tyerman 1983) are obtained in osmotic experiments. The pressure-time (volume-time) courses in these experiments will be given by: Vs- Vso
P~- Pro
Vso
~S
Lpr" Ar"O-st"RTCms =
Vso(kwr~ksr)
[exp(-/%~. t)
- exp ( - k~r"t)],
(2)
where Vso and Pro are the volume of the measuring system and the stationary root pressure at the beginning of an experiment, respectively. Cr.s =concentration of solute "s" in the medium; ksr = r a t e constant for solute uptake across the root cylinder; R=universal gas constant; T=absolute temperature. Equation (2) describes a biphasic process. Upon the addition of the osmotic solute to the medium, root pressure first decreases because of the reduction of the water potential of the medium ("water phase"). However, when a mimimum in root pressure (Prr,in) has been reached, ~ increases again due to the uptake of solutes, which reduces the water potential of the xylem, and water then follows ("solute phase"). This process should proceed until the original root pressure (Pro) is reattained and Cms = Cxs. On the other hand, when a root has taken up a permeating solute and is then treated with the original medium, a biphasic response should occur with a root-pressure maximum. The second phase of the biphasic relaxations will be governed by the passive uptake of solutes into the root xylem, and the rate constant (k~r) of this process will be: ksr
Ar'Psr
Vx
(3).
Thus, by measuring ksr and determining A r a s well as the volume of the xylem (Vx), the permeability coefficient of the root to a given solute (P~r) can be evaluated. Reflection coefficients of roots (%) can be evaluated at J v r = 0 from the maximum changes in root pressure (Pro-- Prmi,(max)) and the corresponding changes of external osmotic pressure (rcs=RT. Cms). Setting dVs/dt, dP~/dt,-O one obtains from Eq. (2): asr =
Pro - Prmin
An~
ax+RT. Cxo am
exp (ksr.
train),
(4)
where ex is the elastic modulus of the xylem and Cxo the stationary solute concentration in the xylem at the beginning of the experiment, tmin denotes
284
E. Steudle and J. Frensch: Osmotic responses of maize roots
the time necessary to reach the minimum. The factor @x+ RT. Cxo)/exon the right side of eqn. (4) corrects for the shrinking and swelling of the xylem during the relaxations (which is usually negligible), and the factor exp (k~.tm~n) for the uptake of permeating solutes during the water phase.
Effects of ion-pumping. It has to be emphazised that eqns. (1) to (4) strictly hold only for passive water and solute movements and for the simple model outlined above. Complications may occur if the rate of active transport (ion-pumping) is changing (e.g., by osmoregulatory responses of the root). In this case, the approach may be still used provided that the changes in P~ due to these changes are sufficiently slow (i.e., if their half-times are much larger than those of the water and passive solute flows). For water, this may generally be true, but for slowly permeating solutes, half-times could become comparable. In this case, P~r (ksr) may be estimated for the maximum slope of the solute phase, if there is a sufficient lag time for the onset of active transport processes. For this time period, it can be shown that the maximum slope of the solute phase will be obtained at t=2"tmi n and that the solute flow at that time will be given by (Steudle unpublished):
dPr~_ __O.sr.RT. Jsr ~ = (Pro- P~mi,)exp ( -- k~. train) 9ksr.
(5)
As in eqn. (4), the exponential term in eqn. (5) denotes the correction for solute flow occurring between minimum and maximum slope. Equation (5) is valid for exosmotic experiments, but should also hold for endosmotic experiments, if Prmi. is replaced by P~max.
Effects of unstirred layers. Unstirred layer effects at the osmotic barrier(s) in the root could be important and could result in an underestimation of Lpr, P~r, and as~ (Dainty 1985; Steudle et al. 1987). Two types of effects have to be considered, the sweepaway effect and purely diffusional effects. In the sweep-away effect, the water flow across the osmotic barrier changes the concentration right at the barrier (Cb~) in that solutes are swept away on one side and are concentrated on the other by the convective flow of water. The effect is opposed by a back-diffusion of solutes. For a steady water flow (Jvr), the relative change in concentration on each side of the membrane (barrier) will be given by (Dainty 1963):
Cbs = exp ( - Jvr" 6/D~), Cms
(6)
where 6 is the thickness of the unstirred layer and D~ the diffusion coefficient of the solutes. Usually, there is some uncertainty in estimating 6 and Ds in tissues. In the experiments in this paper, the unstirred layer on the outer side of the endodermis could, in principle, be as thick as the entire cortex and D~ would be reduced as compared with the free solution. Furthermore it is possible that water flow could be more or less confined to certain pathways or root zones (e.g., to the apoplast or to the absorption zone) and, hence, the local Jw could be substantially inreased by flow constriction. In this case, Jw/Cb has to be used in eqn. (6) instead of Jvr where ~b is the flow-constriction factor (Barry and Diamond 1984). Purely diffusional unstirred-tayer effects ("gradient-dissipation effects"; Barry and Diamond 1984) arise, whenever the diffusional resistances of unstirred layers next to a membrane or osmotic barrier become comparable to that of the barrier itself. If we assume for the root that the endodermis is the only barrier and that D~ will be the same in the cortex and in the stele, we get (considering the cylindrical geometry of the root) in the presence of stelar and cortical unstirred layers (see Appendix A): 1
-
1
E
R
4- D ~ l n a ,
(7)
where 1/P~ ea~ is the measured and l/P~r the true permeation resistance per unit area. R = root radius; E = radius of endodermis; (E-a)= thickness of unstirred layer in the stele. Equation (7) assumes that the entire cortex is functioning as an external unstirred layer. Using reasonable stimates of D~ and a as well as measured values of R and E it is, thus, possible to estimate the influence of gradient-dissipation effects on measurements of P~r, although (different from the techniques used in this paper) eqn. (7) again refers to steady-state conditions. Both gradient-dissipation and sweep-away effects could be also important for the evaluation of Crsr, although the latter effects should be small in the experiments because O-~r was estimated at Jw = 0. In the absence of sweep-away effects and of frictional interactions between solute and Water flow, the approach of Hingson and Diamond (1972) may be used to correct the measured o-~ for unstirred layers. For the cylindrical geometry of the root we get (Appendix A):
E. Steudle and J. Frensch: Osmotic responses of maize roots
1/P r + E/Ds ln(R/a)
285
(8)
Material and methods Plant material. Maize caryopses (Zea mays L. cv. B73 x Mo17; Siidwestdeutsche Saatzucht, Rastatt, FRG) were sterilized by treating them with 1% NaOCt solution and were germinated for 2-3 d at 25~ in the dark on wet filter paper. By that time, the main roots of the seedlings were 30-50 mm long and could be transferred to hydroculture for 3-9 d (1/4 Johnson solution; for composition, see Steudle et al. t987). The plants used in the experiments developed homorrhizic root systems from which the end segments of seminal roots were excised at a length of 38 110 mm (root diameter: 0.7-1.1 ram). The cross-sectional area of the total xylem (Fig. 1) was about 12% of the root cross-section (Steudle et al. 1987). However, only the protoxylem and the early metaxylem vessels should have been mature and functioning in the roots used and not the
Fig. 1. Cross section of the seminal root of a 6-d-old maize seedling. The section was taken at a distance of about 80 mm behind the root apex. Areas of late (l) and early (e) metaxylem were estimated to be 9 and 3% of the total area of the cross section. Note that the endodermis is interrupted by a secondary root initial (r,i.)
big elements of late metaxylem (St. Aubin et al. 1986) and, therefore, only this effective part of the xylem, which was 3% of the cross section of the root, has been used to estimate /~ in the calculation of Psr [Eq. (3)]. Measurement of root pressure and of water and solute relations of roots. Root pressures (Pr), hydraulic conductivity per m 2 of outer root surface area (Lp~), permeability (P~r), and reflection (asr) coefficients were measured as described previously (Steudle and Jeschke 1983; Steudle et al. 1987). Briefly, the root segments were tightly connected to the root pressure probe (Fig. 2) via seals which had been prepared from liquid silicone material (Xantopren plus from Bayer, Leverkusen, FRG) in order to adapt the inner diameter of the seal to individual roots. The length of the seals was 8 mm to provide a sufficient sealing area around the root and, at the same time, avoid large local compressive forces which might have constrained xylem vessels. The proper functioning of the seals was checked after each experiment by cutting the root right at the seal (see below). It was ensured that the half-times of pressure relaxations decreased drastically after the cut. This finding proved that the hydraulic resistance of the sealing area was negligible as compared with that of radial water movement across the root cylinder. In hydrostatic experiments, water flows were induced by moving the meniscus either forward to increase the pressure in the system (exosmotic water flow) or backward to reverse the flow (endosmotic water flow). In the case of the osmotic experiments, the original nutrient solution was rapidly exchanged for media containing test solutions at known concentrations which were determined cryoscopically (Osmomat 030 from Gonotec, Berlin, West Germany). Solutions Were circulat-
286
E. Steudle and J. Frensch: Osmotic responses of maize roots Microscope
,
Pressure horn bee 9 , ,
I~IeTQ L
/
I
I
Pressure ~((/
~
fronsducer
/
End segment of reQt
...........;;~:~:"':~::':":'~::":"'::;':. . . . . . . . . .
screw
L"T" Perfusion chQmber
Sitic0ne seQts
Fig. 2. Root pressure probe for measuring root pressures (P~) and water- and solute-relations parameters of roots. The excised root was tightly connected to a pressure chamber with the aid of a silicone seal so that the root pressure could be built up in the system. Water flows across the root could be induced by either changing the pressure in the pressure chamber with the aid of the metal rod ("hydrostatic experiments") or by exchanging the medium in the perfusion chamber for solutions of different osmotic pressure (" osmotic experiments"; see Figs. 3 and 5). During the experiments, a meniscus, which served as a reference point, was formed within a capillary between the silicone oil in the pressure chamber and the water upon the excised root. With the aid of the meniscus, defined volume changes could be produced in order to determine the elastic extensibility of the measuring system. During the experiments, the tube in which the root segment was fixed was perfused with aerated solution in order to minimize unstirred layer effects
ed along the root segments which were fixed in a narrow glass tube sitting in a bigger one, as shown in Fig. 2. Hydrostatic pressure relaxations and the water phases of biphasic root pressure relaxations in the presence of permeating solutes (see below) yielded k,~, or T,~/2 from which Lp~ was calculated using eqn. (1) of the Theory section. The elastic extensibility of the measuring system (A Vs/AP~)was determined by moving the metal rod in the root pressure probe instantaneously and recording the corresponding A P~. A Vs was calculated from the shift of the oil-water meniscus in the measuring capillary (observed with a stereomicroscope) and the diameter of the capillary ( = 360 gm). The geometric root surface area (At) was estimated from the root length and diameter and Vx from the cross-sectional area of the early metaxylem and protoxylem (see above). In some hydrostatic experiments, Lpr was determined in a constant flow mode, i.e., ,/,~, was kept constant by clamping P~ to a pressure higher than Pro and measuring the resulting water flow from the movement of the meniscus in the measuring capillary (Steudle and Jeschke 1983). In the corresponding osmotic experiments which resembled the pressure-clamp technique used for plant cells (Wendler and Zimmermann 1982), Lpr was calculated from the efflux of water from the root in the presence of an applied osmotic gradient while P, was kept at P~o. In the calculation of Lp r it was assumed that, during the initial phase, the amount of solutes taken up by the root was negligible. In biphasic relaxations where the final (steady) root pressure P,o=~P~o, this deviation could have been due to changes in active transport or to other processes (see Theory section and Discussion). In these cases, the maximum initial solute flow
~l/
g~ ~, S0tution 0utter
(and k~r) was estimated from the maximum slope of the solute phase (eqn. (4)). In all cases, however, where P,e~,P,o, k~,(P~r) values were obtained by measuring the rate constant of the entire solute phase (eqn (2)), i.e., active components were neglected. Reflection coefficients were calculated from the maximum changes in root pressure (Pro--Prmin(max)) and from the changes in the external osmotic pressure according to eqn. (4).
Cutting experiments. Information about the development of the root xylem was obtained from cutting experiments in which roots attached to the root pressure probe were excised from the apex. Sections of length of 5-10 mm were successively cut with a razor blade until a drop in P, indicated that mature xylem was hit. Hydrostatic pressure relaxations performed occasionally between cuts indicated changes in the hydraulic resistance. Since the time constant of the relaxations was given by the product of hydraulic resistance and water capacitance of the system, a change in T/I/2 directly reflected changes in the hydraulic resistance, because the hydraulic capacitance (A Vs/APr) remained constant.
Results Figures 3 and 4 indicate that the roots, basically, responded as would be expected for osmometers in the presence of permeating or non-permeating solutes. In both the hydrostatic and osmotic type of experiments the initial water flows during relaxations (i.e., (dP~/dt)t=o)were linearly related to the driving forces applied across the root. Furthermore, in the osmotic experiments the maximum changes in root pressure were proportional to the changes in the osmotic pressure of the medium (Arc~ eqn. (4)), whereas kw~(T/'~/z) remained the same (see, for example, Fig. 3 D, E). The responses of the roots to permeating solutes were biphasic as predicted by eqn. (2). In the hydrostatic relaxations it was also found that P~e>Pro during exosmotic flow and P~ePre; mannitol) were observed, whereas in others P~o
