Growth of Arabidopsis thaliana seedlings under water deficit studied by control of water potential in nutrient‐agar media

Abstract

We have characterized the growth responses of Arabidopsis thaliana seedlings to water deficit. To manipulate the water potential, we developed a method whereby the nutrient‐agar medium could be supplemented with polyethylene glycol (PEG 8000); PEG was introduced into gelled media by diffusion, which produced media with water potential as low as −1.6 MPa. For dark‐grown plants, hypocotyl growth had a hyperbolic dependence on water potential, and was virtually stopped by −1 MPa. In contrast, primary root elongation was stimulated by moderate deficit and even at −1.6 MPa was not significantly less than the control. That these results did not depend on a direct effect of PEG was attested by obtaining indistinguishable results when a dialysis membrane impermeable to PEG was placed between the medium and the seedlings. For light‐grown seedlings, moderate deficit also stimulated primary root elongation and severe deficit reduced elongation only partially. These changes in elongation were paralleled by changes in root system dry weight. At moderate deficit, lateral root elongation and initiation were unaffected and at higher stress levels both were inhibited. Primary root diameter increased steadily with time in well‐watered controls and under water deficit increased transiently before stabilizing at a diameter that was inversely proportional to the deficit. Along with stimulated primary root elongation, moderate water deficit also stimulated the rate of cell production. Thus, A. thaliana responds to water deficit vigorously, which enhances its use as a model to uncover mechanisms underlying plant responses to water deficit.

Introduction

Plants live in a variable environment to which they acclimate their growth and development. Plant growth is often limited by water deficiency; typically, however, roots are affected less than shoots. In fact, even under mild water deficit, shoots may stop growing completely while roots continue to grow. Continued root growth allows the plant to plumb the soil for water and can be especially important for seedling establishment.

Although the responses of plants to water deficit have been studied in many species, we sought to use Arabidopsis thaliana to take advantage of the potent molecular and genetic tools available for this species. In addition, roots of A. thaliana seedlings have a well‐defined anatomy (Dolan et al., 1993) and cells in the epidermis and cortex are visible in living roots using Nomarski microscopy, which is helpful when looking at growth responses at the cellular level. Furthermore, Beemster and Baskin developed a method for A. thaliana in which the expansion and division parameters of the root meristem can be quantified with high spatial and temporal resolution (Beemster and Baskin, 1998).

Despite the advantages of A. thaliana, to our knowledge only one previous study has used this species to investigate growth at low water potential. Vartanian et al. examined ‘drought rhizogenesis’, the production of short, tuber‐like lateral roots in response to soil drying (Vartanian et al., 1994). The authors also presented data on total root system biomass; shoot growth was not assayed. Unfortunately, comparative data for well‐watered controls were not included, so how water deficit affected root growth could not be assessed.

The limited use of A. thaliana for studies of growth responses to water deficit has been, perhaps, because its small size compounds the difficulty of controlling plant water status reproducibly and quantitatively. In addition, because the roots are extremely thin, it is difficult to image them for growth measurement in soil or vermiculite. Therefore, the objective of this study was to develop a method to grow A. thaliana seedlings at constant and defined water potentials on an agar‐solidified nutrient medium. Nutrient‐agar media are often used in studies of A. thaliana, and hence a large amount of physiological data have been obtained for this condition. By placing the agar medium vertically, the root grows on the surface; this minimizes the possibility of oxygen limitation, which may occur in solution culture when water potential is manipulated with high‐molecular‐weight solutes (Verslues et al., 1998). An agar medium also facilitates the method developed previously (Beemster and Baskin, 1998), which requires marking and imaging the roots on the surface of a growth substrate.

The difficulty encountered here was introducing a high‐molecular‐weight solute into the medium. Whereas low‐molecular‐weight compounds are readily incorporated into an agar medium at even rather high concentrations, larger compounds interfere with the entanglement of the polysaccharide chains and prevent the medium from gelling. Although some authors have decreased water potential with low‐molecular‐weight solutes, this approach suffers because small molecules may be taken up into cells, particularly in long‐term experiments, and because they penetrate the cell wall readily, thus removing water only from the cell. In drying soil, water is lost from the cell wall as well as from the cell, and this condition is reproduced using high‐molecular‐weight solutes. We describe here how a high‐molecular‐weight solute (polyetheylene glycol 8000) was successfully introduced into agar media; and, using this system, we characterized several growth responses of A. thaliana seedlings to water deficit.

Materials and methods

Plant growth

Seeds of Arabidopsis thaliana L. (Heynh), ecotype Columbia, were surface‐sterilized with 15% household bleach and grown vertically on nutrient‐agar media in 10 cm Petri plates. The medium comprised: 1.2% agar, 0.5% sucrose, 4 mM KNO₃, 1 mM Ca(NO₃)₂, 0.3 mM MgSO₄, 2 mM KH₂PO₄, 89 μM iron citrate, 46.3 μM H₃BO₃, 9.1 μM MnCl₂, 0.77 μM ZnSO₄, 0.31 μM CuSO₄, and 0.11 μM MoO₃. To minimize evaporation but still permit gas exchange, plates were wrapped with one layer of bandage tape (Micropore, 3M Company, St Paul, MN).

For light‐grown plants, seeds were stored at 4 °C and on day zero, seeds on plates were placed in a growth chamber under constant conditions (19 °C, 135 μmol m⁻² s⁻¹), as described previously (Baskin and Wilson, 1997). Six days after plating, similar seedlings (roots 1.5–2.0 cm long, cotyledons expanded) were transferred to plates with lowered water potential or control plates with the same water potential. The plates were kept under a clear, acrylic cover to reduce evaporation. For each experiment, three plates with six to seven seedlings each were prepared. For dark‐grown plants, seeds on plates were put in a light‐tight box, held at 4 °C for 5 d to break dormancy, and then transferred to a 20 °C growth chamber. The time of transfer to 20 °C defined day zero. After 6 d, seedlings with a root length of 4–7 mm and a hypocotyl length of 3–12 mm were transferred to plates with lowered water potentials or to the same water potential (controls). During transplantation, plants were exposed to the fluorescent lighting in the laboratory (approximately 3 μmol m⁻² s⁻¹) for about 10 min. The plates were returned to the light‐tight box and put back in the chamber. For each experiment, two plates with 7–11 plants each were prepared per treatment.

For experiments with membranes, seeds were sown on thoroughly cleaned and autoclave‐sterilized dialysis tubing (molecular weight cutoff 3500 Da; SpectraPor 3, Fisher Scientific, Pittsburgh, PA) that was cut open and spread as a single layer on top of the agar medium. Seedlings that were outside the desired size range were removed from the membrane, and the remaining seedlings were transplanted by transferring the membrane.

Water deficit treatment

Water potential was lowered by the addition of various amounts of polyethylene glycol (PEG) (molecular weight 8000; Sigma, St Louis, MO) to the growth medium. Approximately 20 ml of PEG solution was poured on top of an equal volume of solidified nutrient‐agar in a Petri plate, and after 24 h, the solution on top of the plate was poured off and the plate used for experiments. During the 24 h period, the PEG diffused into the agar medium, thus lowering its water potential. That the 24 h period was long enough to reach an approximate equilibrium was verified by measurements of water potential of the top and bottom surfaces of the agar medium. The PEG solution had the same nutrient composition as the growth medium; nutrient solution without PEG was poured on the plates for the ‘well‐watered’ control medium, which gave a water potential for the agar medium of approximately −0.10 MPa. Water potentials were measured using isopiestic thermocouple psychrometry (Boyer and Knipling, 1965) at the end of each experiment, and the data are reported as the mean of the measured potentials.

PEG changes when autoclaved: the water potential of a −1.6 MPa PEG solution decreased by 20% after autoclaving, while the water potential of a comparable sodium chloride solution decreased by only 4%. Therefore, the PEG‐nutrient solution was sterilized by passing it through a filter (0.22 μm). In some experiments, solid non‐sterile PEG was mixed with sterile nutrient solution without leading to detectable contamination.

Growth measurement

Root and hypocotyl elongation were measured by scoring the back of the plastic plate with a razor blade every 24 h, starting immediately after transplanting. For scoring plates in the experiments on dark‐grown plants, plates were illuminated for less than 1 min by light from a flashlight filtered through green acrylic (0.2 μmol m⁻² s⁻¹). At the end of treatment, the plates were photocopied at 1.4 times enlargement and elongation was measured with a digitizing tablet or a ruler as the distance along the root or hypocotyl between each mark. For light‐grown plants, root and shoot dry weight, number of laterals emerged, length of the laterals, and primary root diameter were also measured at the end of the 5 d treatment. For measuring dry weight, shoots and roots were pooled per plate, dried for 8 h at 60 °C and weighed on a microbalance (C‐31, Cahn Instruments, Cerritos, CA). The number of laterals that were visible through a dissecting microscope was counted. The length of the laterals was measured on the photocopies of the plates with a digitizing tablet. Root diameter was measured at each score mark by viewing the roots on the agar plates through a compound microscope at low magnification. Diameters were measured with a video digitizer (Image 1/AT, Universal Imaging, West Chester, PA).

To determine cell flux on days 2, 3, and 4 after transplanting, at the end of the experiment the lengths of 20 cortical cells were measured in parts of the primary root that passed out of the elongation zone on those days. Beemster and Baskin determined that in well‐watered A. thaliana roots, cells resided in the zone of rapid elongation for 6–8 h (Beemster and Baskin, 1998). To ensure that the measured cells had matured on the day in question, cell lengths were measured only in the region of the root that had grown during the latter half of that day. For each day, the elongation rate of a root was divided by the relevant average cell length of the same root, and this ratio was averaged over the sample. Four to six roots were used on each of the three plates per treatment.

Results

To study the effect of low water potential on the growth of A. thaliana seedlings, seeds were germinated and plants grown for 6 d under well‐watered conditions (i.e. nutrient‐agar medium, water potential approximately −0.10 MPa) and then transplanted onto plates with lowered water potentials. To prepare medium with lowered water potential, high‐molecular‐weight polyethylene glycol (PEG 8000) was used. The PEG could not be mixed with molten agar because at higher concentrations of PEG, agar does not solidify. Therefore, as described in Materials and methods, the water potential was lowered by allowing the PEG to diffuse into already solid agar media. In this way, water potentials in the agar medium as low as −1.6 MPa were attained reproducibly, and remained approximately constant over at least 5 d in the growth chamber.

We studied growth responses that were elicited by water deficit in both light‐grown and dark‐grown A. thaliana seedlings because each is known to respond distinctly to various stimuli. Results for dark‐grown seedlings over a 3 d treatment period are shown in Fig. 1A. Hypocotyl elongation had an approximately hyperbolic response to water deficit, decreasing to 60% of the control at a water potential of −0.2 MPa and to near zero below −0.9 MPa. In contrast, primary root elongation was stimulated by lowered water potential between −0.2 and −0.9 MPa, and even at −1.6 MPa was not lower than the well‐watered control. Results for light‐grown roots over a 5 d treatment period are shown in Fig. 1B (hypocotyl elongation was already too low by the time of transplanting to be measurable on the photocopied images of the plates). Primary root elongation, similarly to dark‐grown seedlings, was stimulated by moderate water deficit. Root elongation was reduced by water potentials below −0.5 MPa, although at the lowest water potential (−1.2 MPa), elongation was still 50% of the control.

To determine whether there was a direct effect of PEG on growth, plants were grown on a dialysis membrane spread on the agar. The molecular weight of the PEG (8000 Da) was larger than the cut‐off of the membrane (3500 Da). The membrane did not affect the total elongation over 3 d of hypocotyls or roots (Fig. 1A) or the kinetics of the response (data not shown), indicating that PEG did not affect growth directly.

Because root elongation in the light was more vigorous than in the dark, further investigation used light‐grown plants. The dry weights of the whole plant and the shoot decreased with increasing stress (Fig. 2). The dry weight of the whole root system paralleled the changes in primary root elongation, being increased at moderate deficit and decreased at the most severe deficit. Increased root‐system dry weight was not explained by stimulated initiation or elongation of lateral roots, because there was no significant difference in the number or length of lateral roots of seedlings grown under moderate stress compared to well‐watered conditions (Table 1). Severe stress reduced both the number and the length of the laterals.

To gain further insight into the response of primary root elongation to water deficit, the kinetics of the response during the 5 d exposure were studied (Fig. 3). For the control, elongation rate accelerated steadily throughout the experiment, whereas under water deficit, rates leveled off and approached steady state. At high stress levels (−0.8 and −1.2 MPa), elongation rates throughout the experiment were lower than the control; however, under moderate stress (−0.23 and −0.51 MPa), elongation rates during the first 3–4 d of treatment were substantially greater than the control.

Water deficit is known to cause maize primary roots to become thinner (Sharp et al., 1988; Liang et al., 1997). Thinning is believed to be adaptive so that, under limited water supply, roots can concentrate their use of resources to maintain elongation (Sharp et al., 1990). Therefore, we determined whether water deficit also caused root thinning in A. thaliana. We expected that, at least at moderate stress levels where elongation rate was stimulated, root diameter would decrease for the sake of conserving water. Primary root diameter was measured for each day at the end of the 5 d water‐stress treatment. For the control, diameter increased steadily during the experiment (Fig. 4). Under water stress, root diameter increased over the first day of the experiment, and for all deficits except the most severe, the roots actually became thicker than the control. However, thickening was not sustained, and except for the −0.23 MPa treatment in which diameter continued to increase, by the end of the experiment root diameter had stabilized at a level that was roughly inversely proportional to the stress level and less than that of the well‐watered roots.

The total increase of volume in the primary root was calculated from volume increases per day, assuming the shape of the root was a cylinder. Compared to controls, roots at moderate water deficit added significantly more volume, and only at the highest stress was water uptake reduced (Table 2). At moderate deficit, volume increased because of the transient stimulation of elongation and radial expansion; by the end of the experiment, the rate of volume increase for all deficits was less than the control.

In well‐watered primary roots, the length of mature cortical cells increased slightly (Table 3). Moderate water deficit increased cell length modestly whereas severe stress had little effect. For the three treatments, however, the changes in cell length, if any, were smaller than the concomitant changes in elongation rate. This approximate constancy suggests that mature cell length may be regulated to fall within a preferred range.

To study how water deficit affected cell production, we measured cell flux at the end of the elongation zone. When a root elongates at steady state, cell production rate for a file of cells equals the rate at which cells in that file exit the elongation zone, i.e. the cell flux, and is calculated as the ratio of elongation rate to mature cell length. However, when root elongation rate and mature cell size increase over time, this cell flux underestimates cell production rate. In work on roots grown on 3% sucrose, in which elongation rate and mature cortical cell length increase over time similarly to the well‐watered roots studied here, cell production rate was quantified kinematically, accounting for non‐steady state behaviour (Beemster and Baskin, 1998). Cell production rates thus obtained exceeded the cell flux calculated as the ratio of elongation rate to mature cell length by approximately 10% only (Beemster and Baskin, unpublished data). Because the kinematic analysis is tedious, cell production rate was estimated here as the ratio of root elongation rate to mature cell length without correcting for non‐steady state behaviour.

In well‐watered roots, the flux of cortical cells increased over time in parallel with elongation rate (Table 3), similar to the increase previously reported for roots grown on medium containing 3% sucrose (Beemster and Baskin, 1998). At the severest water deficit assayed (−1.28 MPa), cell flux was approximately 50% of the control; however, at moderate deficit (−0.22 MPa), cell flux was stimulated significantly during the second day of the response. Note that under moderate deficit, elongation rate and cell length increased more than they did under well‐watered conditions; therefore, accounting for non‐steady‐state behaviour would result in calculating cell production to have been increased by moderate deficit to an even greater extent.

Discussion

As with many plants, the shoot of A. thaliana seedlings was acutely sensitive to water deficit, growing slowly at moderate stress and not at all at severe stress. The response of the primary root was opposite in that moderate deficit stimulated elongation. Even at more negative water potentials, primary roots still grew at appreciable rates. Increased primary root elongation rate was correlated with an increased cell production rate. Water stress also prevented the steady increase in root diameter that occurred in the control so that by the end of the experiment, stressed roots were thinner than controls. These results suggest that the species A. thaliana is relevant for studies of plant response to water deficit not only because of its well‐developed molecular genetics, but also because the species acclimates vigorously to water deficit.

The use of PEG in agar to limit water availability

The addition of PEG to agar‐solidified media provided a growth substrate with consistent and reproducible water potential. However, the use of PEG might be problematic. In solutions it increases viscosity, and in solution‐culture even moderate concentrations of PEG 8000 have been shown to hinder oxygen diffusion to the root (Verslues et al., 1998). A. thaliana roots grown on the surface of an agar medium often have a film of solution around the root, which could become an increasing barrier to diffusion with increasing concentrations of PEG. However, under severe stress (<−0.8 MPa) few if any roots had a visible film of solution around them, and under moderate stress, roots grew faster than control roots, not slower as would be expected for oxygen‐deficient roots. Moreover, when plants were grown on a membrane impermeable to PEG 8000, any water film around the root would have been free of PEG; for all PEG concentrations no differences in root or shoot growth rates were found with or without the membrane. Therefore, the effects reported here most likely result from the changed water status of the medium rather than from some other consequence of PEG addition.

Solid growth media supplemented with PEG are also frequently used for tissue culture experiments (Tschaplinski et al., 1995; Capuana and Debergh, 1997; Linossier et al., 1997). Not only is the response of cells and explants to water deficit interesting intrinsically, lowering the water potential of the culture medium often enhances regeneration and hence is important practically. However, the deficits imposed in tissue‐culture studies have been limited by the inability of agar to gel when containing more than about 15% PEG. Although gellan gum (phytagel) does gel with higher concentrations of PEG, in our hands, gellan gum‐media led to irreproducible root elongation rates among controls and, in stress treatments, to effects on root elongation rate as a function of the age of the medium. These effects might be related in some way to the fact that gellan gum‐media increase in gel strength over time (Klimaszewska and Smith, 1997). Interestingly, elongation was affected to the same extent when a dialysis membrane was interposed between the gellan gum‐medium and the seedlings, which appears to rule out a direct effect of any gel property on the roots. In the light of these effects on elongation, gellan gum‐media seem to be inadvisable for physiological experiments. Therefore, the diffusion‐based method used here for adding PEG to agar media used here may be useful, not only for studies on the minuscule seedlings of A. thaliana but also for studies of any plant in culture.

Sucrose supplement

Often, nutrient‐agar media for A. thaliana contain up to 3% sucrose because this promotes plant growth (Baskin and Wilson, 1997). It is assumed that growth is improved because sucrose is a source of carbohydrate; however, the addition of 3% sucrose to a medium would lower its water potential by about 0.2 MPa. It was found that lowering the water potential of the growth medium by this amount with PEG improved root growth. Therefore, the enhancement of root growth by sucrose might occur because of an osmotic rather than a nutritional effect.

Initially, we did experiments in a medium that lacked sucrose completely; germination and growth were irregular to such an extent that it was difficult to select enough uniform seedlings for transplanting. Therefore, 0.5% sucrose was added to the media, which lowered the water potential by a small amount only. However, sucrose, often compared to a hormone, is considered to play a prominent role in regulating plant metabolism (Koch, 1996), and in this context even 0.5% could have profound physiological consequences, particularly for dark‐grown plants, which presumably have less endogenous sucrose. By the same token, the macronutrients in the agar medium also influence seedling growth and may have affected the behaviour of the plant in response to water deficit. The connections between responses to levels of various nutrients and water are knotted and untangling them is an active area of research (McDonald and Davies, 1996). With the method demonstrated here, such studies can be extended to A. thaliana; the availability of mutants with defined alterations in specific components of these responses may support particularly incisive experiments.

Root growth of A. thaliana under water deficit

Water deficit increased the dry weight of the A. thaliana root system, except at the highest deficit tested. Water deficit nearly always increases the ratio of root to shoot dry weight, and there are many examples where water deficit increases the absolute dry weight of the root system (Sharp and Davies, 1979; Meyer and Boyer, 1981). In this respect, A. thaliana responds typically to water deficit, which validates its use as a model system. Not typical is the strikingly enhanced primary root elongation. Most studies have found that water deficit inhibits primary root elongation rate (Gingrich and Russell, 1956; Mirreh and Ketcheson, 1973; Sharp and Davies, 1989; Materechera et al., 1992; Spollen et al., 1993), although stimulation has been shown for Pinus pinaster (Nguyen and Lamant, 1989; Triboulot et al., 1995) and Glycine max (Creelman et al., 1990). The experiments on pine lowered water potentials with PEG whereas those on soybean used vermiculite culture and withheld water; therefore, enhanced primary root growth at moderate stress is not unique to agar or to the use of PEG as an osmoticum.

However, stimulated root elongation is not necessarily unusual in the context of the whole root system. Water deficit often stimulates root system growth, as assessed by the increase of length density of the whole or partial root system (Read and Bartlett, 1972; Klepper et al., 1973; Sharma and Ghildyal, 1977; Jupp and Newman, 1987; Schmidhalter et al., 1998). Increased root system length requires that either root elongation rate or proliferation increases. In many of the above studies, water deficit did increase root proliferation, but also appeared to stimulate the elongation rate of at least some roots within the system, although elongation rate was not measured directly. Under water stress, the stimulated primary root elongation of A. thaliana seedlings thus resembles the enhanced root system growth of the older plants of other species, perhaps reflecting an accelerated development associated with A. thaliana's short life cycle.

In addition to stimulating elongation of the A. thaliana primary root, moderate water deficit also stimulated the rate of cell production. To our knowledge, this is the first report of stimulated rates of cell production under water deficit. In species where water deficit inhibited primary root elongation rate, cell production rate was similarly inhibited (Fraser et al., 1990; Dubrovsky et al., 1998). In A. thaliana at both moderate stress where root elongation rate increased, as well as at severe stress where elongation rate decreased, the change in elongation rate was paralleled closely by the change in cell production rate. This suggests that the root's elongation rate at given levels of water deficit is determined principally by the supply of cells to the zone of rapid elongation, as recently argued to account for the steady increase in elongation rate of well‐watered A. thaliana roots (Beemster and Baskin, 1998). In contrast, in pine seedling primary roots responding to water deficit, cell production rate stayed constant despite increases or decreases in root elongation rate at specific stress levels (Triboulot et al., 1995). Apparently, regulating the supply of cells plays a pivotal role in executing the response to water deficit in the primary roots of some species but not in others; it would be interesting to determine the effect of water stress on cell production in the root systems of older plants.

Root diameter

For many years, root systems under water deficit have been observed to produce thin roots, although such observations have rarely distinguished between new roots being initiated in narrower diameter classes and individual roots lessening their radial expansion. However, several studies have measured the diameter of the primary root of water‐stressed plants, and consistently reported that it was thinner than that of the well‐watered control (Taylor and Ratliff, 1969; Eavis, 1972; Read and Bartlett, 1972; Sharp et al., 1988). Thus, the thin primary roots of A. thaliana under water deficit resemble the response seen in other plants.

Despite the presumed importance of limiting radial expansion to conserve water, few reports have examined how roots become thinner at low water potential. In the primary root of maize, individual roots became thinner with time under stress, and the rates of radial expansion were not correlated with rates of elongation (Liang et al., 1997). In A. thaliana, however, individual roots did not become thinner; instead, water deficits below −0.23 MPa prevented the steady increase in diameter that occurred under well‐watered conditions. In fact, the diameter of the root appeared to be closely correlated with elongation rate, so much so that, except under severe stress, the root simultaneously elongated faster and grew fatter than the control, failing to minimize water uptake. Root diameter and elongation rate have been correlated in other species for various circumstances (Wilcox, 1972; Hackett and Rose, 1972; Thaler and Pagès, 1996). Thus, in A. thaliana, root diameter and elongation rate may be reduced under water deficit by a common mechanism, whereas in maize, the thinning of roots at low water potential appears to reflect the action of a specific mechanism for limiting radial expansion.

That A. thaliana roots did not thin over time may be a consequence of the small absolute diameter of the root, which is scarcely more than 100 μm, roughly a tenth of the maize root. First, in the A. thaliana root meristem, where final root diameter is attained in this species, cells may already be close to the minimal cytoplasmic volume needed to function. Second, decreased root diameter may be significant not only because it reduces the quantity of water that must be absorbed to support elongation but also because it increases the surface to volume ratio, thus facilitating water uptake. Given that the surface to volume ratio of the thread‐like roots of A. thaliana is already high, there is relatively less to gain by reducing root diameter further.

We thank Steve Wells for help with constructing thermocouples used in measuring water potential, and Mike Keller for help with statistics. This work was supported in part by grant No. IBN 9817132 from the US National Science Foundation (to TIB), and by award No. 95‐37100‐1601 from the US Department of Agriculture (to WGS and RES). This is contribution No. 13 006 from the Missouri Agricultural Experiment Station journal series.

References