Hydraulic Conductance and Mercury-Sensitive Water Transport for Roots of Opuntia acanthocarpa in Relation to Soil Drying and Rewetting

Abstract

Drought-induced changes in root hydraulic conductance (L_P) and mercury-sensitive water transport were examined for distal (immature) and mid-root (mature) regions of Opuntia acanthocarpa. During 45 d of soil drying, L_P decreased by about 67% for distal and mid-root regions. After 8 d in rewetted soil, L_P recovered to 60% of its initial value for both regions. Axial xylem hydraulic conductivity was only a minor limiter of L_P. Under wet conditions, HgCl₂ (50 μm), which is known to block membrane water-transport channels (aquaporins), decreased L_P and the radial hydraulic conductance for the stele (L_{R, S}) of the distal root region by 32% and 41%, respectively; both L_P and L_{R, S} recovered fully after transfer to 2-mercaptoethanol (10 mm). In contrast, HgCl₂ did not inhibit L_P of the mid-root region under wet conditions, although it reduced L_{R, S} by 41%. Under dry conditions, neither L_P nor L_{R, S} of the two root regions was inhibited by HgCl₂. After 8 d of rewetting, HgCl₂ decreased L_P and L_{R, S} of the distal region by 23% and 32%, respectively, but L_P and L_{R, S} of the mid-root region were unaltered. Changes in putative aquaporin activity accounted for about 38% of the reduction in L_P in drying soil and for 61% of its recovery for the distal region 8 d after rewetting. In the stele, changes in aquaporin activity accounted for about 74% of the variable L_{R, S} during drought and after rewetting. Thus, aquaporins are important for regulating water movement for roots of O. acanthocarpa.

When the soil is wet, the root system is the primary limitation for plant water uptake (Nobel and Cui, 1992; Sperry et al., 1998). The root hydraulic conductance based on the root surface area (also referred to as the root hydraulic conductivity [L_P]) thus has a major influence on the shoot water status, and, in turn, on plant growth and development (Frensch, 1997). Water moves from the surface of a root to the root xylem through a series of tissues, each with a hydraulic conductance that can change with root development (Melchior and Steudle, 1993) and with the availability of soil moisture (North and Nobel, 1996). L_P for most species is limited by the radial hydraulic conductance (L_R) of the tissues outside the xylem (North and Nobel, 1991; Melchior and Steudle, 1993). Knowledge of L_P and L_R not only improves understanding of individual root functioning (Steudle, 2000) but also is important for modeling water uptake by an entire root system (Doussan et al., 1998; Sperry et al., 1998).

Many studies have addressed the regulation of transpiration by stomata in relation to environmental factors (Schulze, 1994; Jones, 1998). Less is known about possible mechanisms regulating root water uptake. In many species, root L_P considerably decreases as the soil dries (Cruz et al., 1992; North and Nobel, 1996; Lo Gullo et al., 1998). Such decreases are associated with substantial anatomical modifications, such as the development of Casparian bands and suberin lamellae in the exodermis and the endodermis (Enstone and Peterson, 1998; North and Nobel, 2000). The roots of most dicotyledons exhibit secondary growth and produce a suberized periderm outside the stele, which also restricts water uptake (North and Nobel, 1996). The axial hydraulic conductivity of the root may also decrease during drought due to xylem cavitation (Linton and Nobel, 1999). Under natural conditions when the soil dries gradually, such modifications can regulate water flow from the root surface to the root xylem and can limit a backflow of water from the root xylem to the drying soil (Taleisnik et al., 1999). After substantial soil water loss, the soil hydraulic conductivity decreases markedly and root shrinkage can occur, which increases the resistance (reciprocal of conductance) at the soil-root interface (Nobel and Cui, 1992).

Suberization and lignification of cell walls affect water movement through the root apoplast, yet modifications in the cell-to-cell pathway may allow more flexible control of root water transport (Steudle, 2000). Likely candidates for such control are the proteinaceous membrane water-transport channels (aquaporins) that are present in both the plasma membrane and the tonoplast for a wide range of plant tissues (Maurel, 1997). In roots of maize (Zea mays), tonoplast aquaporin ZmTIP1 mRNA occurs in all tissues of the apical growth zone (Chaumont et al., 1998), whereas in mature regions of the roots, it apparently is limited to the endodermis and xylem parenchyma cells (Barrieu et al., 1998). Similar patterns of aquaporin localization occur in roots of sunflower (Helianthus annuus; Sarda et al., 1999), the common ice plant (Mesembryanthemum crystallinum; Kirch et al., 2000), and tobacco (Nicotiana tabacum; Otto and Kaldenhoff, 2000). The root/shoot ratio and the hydraulic conductivity of protoplasts for antisense versus control plants of Arabidopsis suggest that, in addition to cytosolic osmoregulation, aquaporins are important for the bulk flow of water in a plant (Kaldenhoff et al., 1998). Some aquaporin genes are up-regulated during drought and salinity stress (Guerrero et al., 1990; Yamaguchi-Sinosaki et al., 1992; Yamada et al., 1997), whereas others are down-regulated (Yamada et al., 1995). For sunflower, exposure of part of the root system to air induces a complex regulation of the expression of closely related δ-TIP genes (Sarda et al., 1999).

Sulfhydryl reagents, such as HgCl₂, inhibit water flow via most (Maurel, 1997) but not all aquaporins (Daniels et al., 1994); subsequent use of 2-mercaptoethanol to reverse this inhibition facilitates the study of aquaporin-mediated water uptake by whole roots (Maggio and Joly, 1995; Wan and Zwiazek, 1999; Barrowclough et al., 2000). Such studies on whole root systems or root regions indicate that aquaporins can account for 60% to 80% of the root L_P. HgCl₂ (100 μm) decreases L_P for the root system of Populus tremuloides in 1 h without reducing the respiration rate, suggesting that the inhibition of root water uptake is not due to metabolic inhibition (Wan and Zwiazek, 1999). In contrast, HgCl₂ rapidly depolarizes the plasma membrane (half-maximal depolarization at 8 μm) for cortical root cells of bread wheat (Triticum aestivum) and has other effects in addition to the direct inhibition of water channel activity (Zhang and Tyerman, 1999). Patterns of aquaporin transcription and translation follow the diurnal rhythm for L_P of roots of Lotus japonicus (Henzler et al., 1999). Also, the water transport activity of aquaporins in roots is affected by salinity (Carvajal et al., 1999, 2000), nutrient deprivation (Carvajal et al., 1996), and drought (North and Nobel, 2000). L_P for the distal region of young roots of the desert monocotyledon Agave deserti is reduced 60% by HgCl₂ (50 μm) under wet soil conditions, whereas HgCl₂ has no effect after 45 d in drying soil (North and Nobel, 2000), suggesting that reduced aquaporin activity could help limit root water loss to a dry soil. The resumption of aquaporin function after soil rewetting could allow renewed root water uptake.

Changes in L_P in response to soil drying and rewetting were determined for Opuntia acanthocarpa, a dicotyledon in the Cactaceae that is sympatric with A. deserti in the northwestern Sonoran Desert. Anatomical changes and the involvement of mercury-sensitive water transport in L_P regulation were examined. Three hypotheses were tested: (a) Soil moisture affects aquaporin-mediated water uptake, which could contribute to the variable root resistance to water flow that is reported for several species during soil drying and subsequent rewetting; (b) aquaporin activity is greater in the immature distal region (where an endodermis is present) than in the mature mid-root region (where several periderm layers are present); and (c) aquaporin activity is localized primarily in the stele, as suggested by molecular marker studies on other species.

RESULTS

Root Morphology, Anatomy, and Changes in Radial Pathway

The diameters for distal and mid-root regions of O. acanthocarpa in wet soil were not significantly different (P = 0.54; Table I). After 45 d in drying soil, the diameter decreased by 21% (P = 0.03) for the distal root region but was unchanged (P = 0.77) for the mid-root region. For both regions, root diameter did not change (P = 0.86, n = 9) after 8 d of rewetting. After 4 d of rewetting, one or two new lateral roots arose 1 to 2 cm proximal to the dead tip. These new laterals had a diameter comparable to that of the distal region under wet conditions and were about 4 cm in length after 8 d of rewetting. No new lateral roots were produced during the 8 d of rewetting along the mid-root region.

Table I.

Diameter, no. of cell layers in the cortex and the periderm, and no. of lignified layers in the periderm for roots of Opuntia acanthocarpa in wet soil (Ψ_soil > −0.25 MPa) or after 45 d in drying soil (Ψ_soil < −10 MPa)

Region	Soil	Root Diameter	No. of Cell Layers
			Cortex	Periderm	Lignified layers in periderm
		mm
Distal	Wet	1.71 ± 0.10a	9.4 ± 0.4a	0.2 ± 0.2a	0
	Dry	1.36 ± 0.09b	6.9 ± 0.5b	6.1 ± 0.5b	0
Midroot	Wet	1.80 ± 0.11a	7.2 ± 1.0ab	6.2 ± 0.4b	0.23 ± 0.16a
	Dry	1.66 ± 0.06a	7.0 ± 0.4b	7.8 ± 0.7b	2.00 ± 0.31b

Measurements were made at 40 mm (the distal root region) or 160 mm (the midroot region) from the root tip. Data are means ± 1 se (n = 7–19 roots from five–six plants).

^a  

^b Different letters within a column indicate a significant difference (P < 0.05). 

Under wet soil conditions, roots of O. acanthocarpa had an epidermis, a parenchymatous cortex, and an endodermis beginning at about 10 mm back from the root tip (Fig. 1A). Periderm began to develop at about 70 mm from the tip (Fig. 1B). The number of cortical cell layers was not significantly different (P = 0.10) for distal and mid-root regions (Table I). The epidermis for both distal and mid-root regions collapsed under drying conditions, and intercellular spaces or fissures were enlarged in the cortex due to cell shrinkage, which often pulled the cortex away from the periderm in the mid-root region. After 45 d of soil drying, the number of layers of cortical cells decreased by 27% (P = 0.99; Table I) for the distal region but was unchanged for the mid-root region (P = 0.10).

Figure 1.

Photomicrographs of cross sections of main roots of O. acanthocarpa in wet soil (A and B) and after 45 d in drying soil (C and D). Sections were made at 40 mm (A and C, distal root region) and 160 mm (B and D, mid-root region) from the root tip. All cross sections are stained with toluidine blue O. Single arrows indicate endodermis; double arrows indicate mucilage cells. e, Epidermis; c, cortex; p, periderm, s, stele. Scale bars = 100 μm.

Under wet soil conditions, at 160 mm from the tip (middle of the mid-root region) the periderm consisted of about six layers of cells with suberized tangential and radial walls (Fig. 1B; Table I). After 45 d in drying soil, a periderm with two to three layers of lightly suberized cells began to develop at about 15 mm from the tip; at 40 mm (middle of the distal root region) the periderm consisted of about six layers of cells with suberized tangential and radial walls (Table I, Fig. 1C). The number of periderm layers for the mid-root region did not change (P = 0.33) during soil drying, but two to three layers had thick tangential and radial cell walls that stained more strongly for lignin than for suberin (Table I), whereas the outer layers had thin cell walls that stained strongly for suberin but not for lignin (Fig. 1D). In drying soil, large mucilage cells associated with the secondary xylem were present in both distal and mid-root regions (Fig. 1D). Under all soil conditions, a suberized exodermis was not apparent in either the distal or the mid-root region. Anatomical changes in response to 8 d of rewetting were slight for both distal and mid-root regions.

L_P and Effect of HgCl₂

The volume flux density (J_V) into the distal root region of O. acanthocarpa immersed in water or in solutions containing HgCl₂ (50 μm) or 2-mercaptoethanol (10 mm) was linear with the applied pressure difference from 10 to 40 kPa (Fig. 2; r² = 0.99 for all curves). Also, J_V in the absence of an applied pressure was near zero. Similar linear relationships were obtained for the mid-root region. The slopes of the relationships equal L_P.

Figure 2.

Relationship between J_V and applied pressure difference for the distal root region of O. acanthocarpa in wet soil. J_V was measured sequentially in water, HgCl₂ (50 μm), and 2-mercaptoethanol (10 mm). Data are means ± 1 se for n = 6 roots from six plants.

Under wet conditions, L_P was 32% lower (P < 0.05) for the distal root region (including the tip) than for the mid-root region (Fig. 3). During 45 d of soil drying, L_P of the distal and the mid-root region decreased by 60% and 73%, respectively, and then did not differ for the two regions (P = 0.90). In soil rewetted for 1 d, L_P of both distal and mid-root regions did not change significantly (P = 0.99). After 8 d of rewetting, L_P of both distal and mid-root regions was restored to 60% of its initial value (Fig. 3). The tips of the main roots became necrotic after 45 d of soil drying. Excision of the new distal root growth after 4 d of rewetting had no effect on L_P of rewetted distal root regions, presumably because of the immaturity of the xylem vessels.

Figure 3.

L_P for the distal (A) and the mid-root (B) region for roots of O. acanthocarpa in wet soil and during soil drying and rewetting. Water was withheld from −45 to 0 d, when Ψ_soil was >−0.25 and <−10 MPa, respectively. After rewetting, Ψ_soil was maintained above −0.1 MPa. L_P was measured for root regions in water and then in 50 μm HgCl₂. Data are means ± 1 se for n = 5 to 7 roots from four to six plants. The asterisk below a pair of points indicates a statistically significant difference due to the HgCl₂ treatment (paired t test, P < 0.05).

Under wet conditions, L_P of the distal root region decreased by 32% ± 5% (P < 0.01) when transferred to 50 μm HgCl₂ (Fig. 3A); subsequent transfer to 10 mm 2-mercaptoethanol permitted recovery to 101% ± 7% of the initial value. Under drought conditions, HgCl₂ did not significantly change L_P of the distal root region (P = 0.69), but HgCl₂ decreased L_P of the distal root region by 21% ± 4% after 1 d of rewetting (P < 0.01) and by 22% ± 2% after 8 d of rewetting (P < 0.01). The inhibition of L_P of rewetted distal root regions by HgCl₂ was not dependent on new apical growth. Under all soil moisture conditions, L_P for the mid-root root region did not change significantly (P > 0.20) after transferring to 50 μm HgCl₂ (Fig. 3B).

Root Axial Hydraulic Conductivity (K_h)

Under wet, drying, and rewetted conditions, K_h was similar (P > 0.09) for distal and mid-root regions (Fig. 4). During 45 d of soil drying, K_h of both distal and mid-root regions decreased (P < 0.05) to 10% of its value under wet conditions. Rewetting for 8 d caused K_h to increase to 41% and 55% (not significantly different, P = 0.24) of its initial value for the distal and the mid-root region, respectively.

Figure 4.

K_h for the distal and the mid-root region for roots of O. acanthocarpa in wet soil and during soil drying and rewetting (conditions as for Fig. 1). Data are means ± 1 se for n = 5 to 12 roots from four to six plants.

Root Radial Hydraulic Conductances and Effect of HgCl₂

Under wet conditions, L_R was 57% lower (P < 0.001) for the distal than for the mid-root region (Fig. 5). As for Figure 2, very little flow occurred in the absence of an applied pressure difference for segments from which tissues external to the stele had been removed. L_R of the epidermis/cortex/periderm (L_{R, E/C/P}) was then similar (P = 0.52) to that of the stele (L_{R, S}) for the distal region, but L_{R, E/C/P} was 2-fold higher (P < 0.001) than L_{R, S} for the mid-root region. Also under wet conditions, L_{R, E/C/P} was one-quarter as high (P < 0.001) for the distal as for the mid-root region, whereas L_{R, S} was not significantly different (P = 0.16) for the two regions. During 45 d in drying soil, L_R decreased 81% for both distal and mid-root regions (Fig. 5; P < 0.001). L_{R, E/C/P} and L_{R, S} then decreased by 86% and 48%, respectively, for the distal root region and by 90% and 66%, respectively, for the mid-root region.

Figure 5.

L_R for intact root regions, epidermis/cortex/periderm (E/C/P), and stele for distal (A) and mid-root (B) regions for roots of O. acanthocarpa under wet conditions (Ψ_soil > −0.25 MPa), after 45 d in drying soil (Ψ_soil < −10 MPa), and after 8 d of rewetting (Ψ_soil > −0.1 MPa). For the stele, L_P was measured in water and then in HgCl₂ (50 μm) for distal and mid-root regions and also in 2-mercaptoethanol (ME, 10 mm) for the distal region. After determining K_h, L_R for intact root regions and the stele was then calculated using Equation 3 and for the epidermis/cortex/periderm tissues using Equation 4. Data are means ± 1 se (n = no. of roots from three–six plants).

When the epidermis/cortex and periderm were removed sequentially, the L_Rs indicated that the periderm accounted for only 12% (n = 2) of the resistance external to the stele for the mid-root region under wet conditions. Under dry conditions, the periderm accounted for 41% (n = 6) and 77% (n = 2) of the resistance external to the stele for the distal and the mid-root region, respectively.

For the distal root region after 8 d of rewetting following 45 d of drying, L_R increased to 81% of its value under wet conditions, reflecting a 3-fold increase in L_{R, E/C/P} and a doubling of L_{R, S}, leading to values that were not significantly different (P > 0.80) from those under wet conditions. For the mid-root region at 8 d of rewetting, L_R increased by 42% (P < 0.05) but remained only 57% (P < 0.001) of the value under wet conditions (Fig. 5). L_{R, E/C/P} then increased to only 32% of its value under wet conditions, whereas L_{R, S} was unchanged (P = 0.91).

Under wet conditions, when the stele of both distal and mid-root regions was placed in 50 μm HgCl₂, L_{R, S} decreased by 41% (Fig. 5; P < 0.02). As for L_P, the inhibition by HgCl₂ on L_{R, S} for the distal region was fully reversed (P = 0.34) by 10 mm 2-mercaptoethanol (Fig. 5A). In contrast, after 45 d in drying soil, L_{R, S} of both distal (P = 0.84) and mid-root (P = 0.13) regions was not affected by 50 μm HgCl₂. After rewetting for 8 d, L_{R, S} of the distal region decreased significantly (by 35%, P < 0.01) in response to HgCl₂, but no inhibition occurred for the mid-root region (Fig. 5B).

DISCUSSION

For root regions of O. acanthocarpa immersed in water or solutions containing HgCl₂ or 2-mercaptoethanol, J_V was proportional to the applied pressure difference. To be specific, the relationship was linear and virtually no flux occurred in the absence of an applied pressure difference. The y-intercept equals L_Pς(Π_x - Π₀), where ς is the root reflection coefficient and Π_x and Π₀ are the osmotic pressures at the root xylem and the root surface, respectively (Fiscus, 1975; Lo Gullo et al., 1998). Thus, the osmotic pressure difference between the root surface and the root xylem was negligible, and/or the root reflection coefficient was low, and/or L_P of the pathway used by the water when the driving force is an osmotic pressure difference is very low. This implies that the effective driving force for water flow through the root segments of O. acanthocarpa was the applied pressure difference.

Although a decline in L_P or L_R following immersion of a root segment in HgCl₂ is taken to indicate involvement of aquaporins in radial water flow, such treatments can also lead to artifacts due to toxicity and non-selectivity of the inhibitor (Zhang and Tyerman, 1999; Barrowclough et al., 2000). For both the distal root region and the dissected stele of O. acanthocarpa, inhibition by 50 μm HgCl₂ was fully reversed by subsequent treatment with 10 mm 2-mercaptoethanol, suggesting that a toxic reaction was unlikely, as is also observed for roots of Allium cepa (Barrowclough et al., 2000) and A. deserti (North and Nobel, 2000). The 32% reduction of L_P by 50 μm HgCl₂ for roots of O. acanthocarpa under wet conditions is smaller than inhibitions using that concentration for A. cepa (57%–84% inhibition; Barrowclough et al., 2000), A. deserti (60%; North and Nobel, 2000), Capsicum annuum (66%; Carvajal et al., 1999), Cucumis melo (80%; Carvajal et al., 2000), barley (Hordeum vulgare; 90%; Tazawa et al., 1997), P. tremuloides (47%; Wan and Zwiazek, 1999), or bread wheat (66%; Carvajal et al., 1996). In any case, not all aquaporins are sensitive to mercury (Daniels et al., 1994; Otto and Kaldenhoff, 2000), and the penetration of HgCl₂ may by blocked by suberized cell layers (Barrowclough et al., 2000), such as the endodermis or periderm.

Under wet conditions, HgCl₂ similarly inhibited L_P for the distal root region and the stele of O. acanthocarpa, indicating that aquaporins were not located solely in the tissues external to the stele (namely the epidermis, the cortex, and/or the periderm), but were located either in the stele alone or in both the stele and the external tissues. Considering measured cortical cell membrane hydraulic conductances ranging from 0.5 to 9 × 10⁻⁷ m s⁻¹ MPa⁻¹ (Steudle, 1992) and a cortex with 9.4 cell layers arranged in series, L_R for the cortex of O. acanthocarpa could range from 0.3 to 5 × 10⁻⁸ m s⁻¹ MPa⁻¹, i.e. only 0.7% to 12% of the L_{R, E/C/P} of the distal root region. Hence, when water flow is induced in roots of O. acanthocarpa by a hydrostatic pressure difference, most flow in the cortex appears to be through the apoplastic pathway rather than a cell-to-cell pathway, so aquaporins would have little influence on L_{R, E/C/P}. Moreover, under wet conditions, the distal root region of O. acanthocarpa lacked a periderm, and the endodermis was discontinuous and only slightly lignified and suberized. Therefore, these tissues also would not greatly restrict the flow of either water or HgCl₂ to the stele, suggesting that the HgCl₂-inhibited water transport, most likely through aquaporins, was located primarily in the cells of the stele. This is consistent with the lack of correlation between the variations of the hydraulic conductivity of cortical cells and the root L_P or the level of aquaporins over a 24-h period for L. japonicus (Henzler et al., 1999).

For the mid-root region under wet conditions, HgCl₂ did not affect L_P but decreased L_{R, S} by 41%. Considering the high L_R for tissues external to the stele in the mid-root region, most of the water flow in these tissues was apoplastic, as for the distal region. The lack of reduction in L_P for intact mid-root segments presumably reflects the failure of HgCl₂ to cross the periderm, as occurs for the suberized exodermis of A. cepa (Barrowclough et al., 2000), and does not indicate that aquaporins are inactive in water uptake or absent at mid-root. Under wet conditions, water crosses the periderm; thus, aquaporin activity in the stele in both distal and mid-root segments may well influence conductance. The apparent restriction of aquaporin activity to the stele provides functional support for molecular evidence that aquaporin mRNA is primarily located in the parenchyma cells associated with the xylem and the phloem in mature regions of roots for several species, as revealed by in situ hybridization (Barrieu et al., 1998; Sarda et al., 1999; Kirch et al., 2000; Otto and Kaldenhoff, 2000).

During soil drying, L_P decreased 60% for the distal region, and HgCl₂ then had no effect on L_P or L_{R, S}. Thus, aquaporins may not have been active under dry conditions. The contribution of aquaporin closure to the overall reduction in L_R under dry conditions can be calculated from the ratio:

1

where L_P^Wet is L_P for root regions under wet conditions, L_P^Dry is L_P for root regions under drying conditions, and L_P^HgCl₂ is L_P for regions under wet conditions but measured in HgCl₂. The value of the ratio is 0.52, indicating that closure of aquaporins accounted for 48% of the decrease in L_P for the distal region for O. acanthocarpa. Performing a similar calculation with L_{R, S} indicates that aquaporin closure accounted for 83% of the decrease in L_{R, S} for the distal region. After 8 d of rewetting, L_P increased to 60% of its initial value, and HgCl₂ then inhibited L_{R, S} of the distal region by 35%. Reopening of aquaporin channels could account for 63% of the partial recovery of L_P in the distal region and 95% of the recovery of L_R,S. Hence, aquaporins appeared to contribute substantially to changes in L_P with soil water availability. Soil drying caused L_P for the mid-root region to decrease by 73%, a greater decrease than for the distal region, perhaps due to the increased suberization and lignification of the periderm during drying (North and Nobel, 1992; Enstone and Peterson, 1998). Under dry conditions, HgCl₂ had no effect on L_P or L_{R, S} for the mid-root region, as for the distal region. Closure of aquaporin channels under drying conditions could account for 37% of the reduction in L_P and 64% of the reduction in L_{R, S} for the mid-root region. After 8 d of rewetting, L_P increased to 60% of its initial value, as for the distal region, but L_{R, S} for the stele of the mid-root region did not increase after rewetting, nor was it affected by HgCl₂, indicating a continued absence of aquaporin activity. This could be related to the greater maturity of the tissues and the associated lower metabolic activity in the mid-root compared to the distal region (Palta and Nobel, 1989).

Under adverse conditions, the closure of root aquaporins may help to limit water loss when the soil water potential is lower than that of the root (Steudle, 2000). Under conditions of moderate water stress, molecular work suggests that aquaporins are often up-regulated (Yamaguchi-Sinosaki et al., 1992; Yamada et al., 1997; Sarda et al., 1999; Kirch et al., 2000). The exposure of roots of sunflower to a mild drought, through partial exposure of the root system to air, shows a complex response by related δ-TIP genes encoding aquaporins located in the parenchyma cells surrounding the phloem. Some genes are repressed, some are unaffected, and others are overexpressed—overall, the level of δ-TIP expression increases (Sarda et al., 1999). In contrast, aquaporin activity in the roots of the desert succulents A. deserti (North and Nobel, 2000) and O. acanthocarpa is eliminated or is no longer susceptible to inhibition by HgCl₂ under dry conditions. It is interesting that HgCl₂ (50 μm) has an effect similar to that of 50 mm NaCl on L_P for roots of C. annuum, and the inhibition by HgCl₂ is partially reversed by dithiothreitol (5 mm). However, in plants treated with NaCl, only a slight effect of HgCl₂ is observed, suggesting that either the reduction of the activity or the abundance of mercury-sensitive water channels is the primary cause of L_p reduction following a short-term salinity stress (Carvajal et al., 1999), similar to results for C. melo (Carvajal et al., 2000; Martinez-Ballesta et al., 2000). Also, the exposure of the common ice plant, a facultative halophyte, to salt shock (400 mm NaCl) reduces the levels of the plasma membrane aquaporins MIP-A and MIP-C mRNA in roots, but the levels recover within 2 d to the prestress level after full turgor is restored (Yamada et al., 1995). The duration and the severity of drought imposed, in contrast to the milder water deficits in other experiments, as well as the use of a xerophytic (or halophytic) species may explain the different results. Moreover, posttranslational and posttranscriptional modification can also regulate aquaporin water transport activity (Maurel, 1997). For example, in leaves of spinach (Spinacia oleracea), phosphorylation of the plasma membrane aquaporin PM28A decreases with decreasing apoplastic water potential (Johansson et al., 1996, 1998).

Under wet conditions, L_{R, S} of both distal and mid-root regions of O. acanthocarpa was not significantly different. In the mid-root region, L_{R, S} was only 30% of L_{R, E/C/P}, indicating that the stele predominated in limiting radial water movement. In the distal region, L_{R, S} and L_{R, E/C/P} were similar, as is the case for roots of Opuntia ficus-indica (North and Nobel, 1996), reflecting the tight cell packing in the cortex in the distal root region for both species. Also, in the mid-root region of both species, L_{R, E/C/P} was higher than in the distal root region due to the death and subsequent loss of resistance of the epidermal and cortical cells during root development. Although the periderm was more highly developed in the mid-root than in the distal region for O. acanthocarpa, the water permeability of its suberized and lignified cell walls was relatively high under wet conditions, as is the case for the roots of several other species (Vogt et al., 1983; North and Nobel, 1996).

The decrease in L_R for roots of O. acanthocarpa during soil drying resulted not only from the closure of aquaporins but also from an increase in the suberization of peridermal cells and the decrease of their hydraulic conductance upon drying (Vogt et al., 1983; North and Nobel, 1996). In the distal root region during soil drying, the increase in L_R due to the collapse of the cortex was outweighed by the reduction in L_R due to the six-layer periderm, which was not present under wet conditions. For the mid-root region, L_R of the periderm decreased by 97% during 45 d in drying soil, which would help prevent the cambial cells of the stele from dehydrating. After 8 d of rewetting, L_{R, E/C/P} recovered to 67% and 32% of its initial value for the distal and the mid-root region, respectively. Such tissues were then the main limitation for radial water movement for the distal segment. In contrast, for the mid-root region, the aquaporin activity did not recover after rewetting, and the stele was then the main radial resistance.

Values of K_h for O. acanthocarpa agree well with those previously reported for O. ficus-indica and Ferocactus acanthodes (North and Nobel, 1992), although K_h for the latter two species is restored almost to its initial value after 7 d of rewetting. In all cases, the decrease in K_h during drying probably resulted from xylem embolism, which can also help limit plant water loss by preventing backflow from the succulent shoot through the roots to a drying soil (North and Nobel, 1992; Linton and Nobel, 1999). Under wet conditions, L_R for roots of O. acanthocarpa was not significantly different from L_P, and K_h limited L_P by only 10% to 15%. Under wet conditions, K_h also does not significantly limit L_P for roots of A. deserti (North and Nobel, 1991) and maize (Melchior and Steudle, 1993), whereas for other species, including O. ficus-indica and F. acanthodes (North and Nobel, 1992), K_h can limit L_P, even in moist soil. Under dry conditions, K_h limited L_P by about 16%. After 8 d of rewetting, K_h recovered only partially, and K_h then limited L_P by 9%. Although K_h and L_P recovered to approximately the same extent after rewetting, K_h was then less of a limiting factor than under wet conditions due to an irreversible decrease in the root radius during drying, leading to a corresponding increase in the L_R (Landsberg and Fowkes, 1978). In any case, these results should be taken with caution, as K_h was probably underestimated due to blockage of the vessels at the cut end by mucilage within the stele, as indicated by the rapid decrease in the axial water flux after excising root segments at mid-length.

In conclusion, the first and the third hypotheses framed in the introduction were generally supported: Aquaporin activity, as inferred from HgCl₂ experiments, was affected by drought conditions and was greatest in the stele in both distal and mid-root regions, accounting for 32% of L_P under wet conditions. HgCl₂ decreased L_P for both distal and mid-root regions under wet conditions and for the distal region after rewetting, but had no effect on L_P of either the intact or the dissected root regions under drying conditions. The second hypothesis was partially supported, as aquaporin activity after rewetting was restricted to the distal root region. Such variations in aquaporin activity with root development and water availability may regulate L_P during drought and may help O. acanthocarpa take up water under conditions of heterogeneous soil moisture that prevail in its native habitat.

MATERIALS AND METHODS

Plant Material and Culture Conditions

Thirty-two plants of Opuntia acanthocarpa var. ganderi C.B. Wolf (Cactaceae) about 30 cm tall were collected from Agave Hill in the University of California Philip L. Boyd Deep Canyon Desert Research Center (field site at 33°38′N, 116°24′W, 820-m elevation) in the northwestern Sonoran Desert. They were grown in a greenhouse at the University of California (Los Angeles) in 36-cm-long × 28-cm-wide × 12-cm-deep plastic tubs containing a 1:1 mixture of washed quartz sand:soil from Agave Hill. Plants received a mean total daily photosynthetic photon flux of 38 mol m⁻² d⁻¹ (80% of ambient solar radiation), with daily maximum/minimum air temperatures averaging 28°C/16°C and a day/night relative humidity averaging 40%/70%. The soil water potential (Ψ_soil, MPa) in the rooting zone was maintained above −0.25 MPa by biweekly watering with 0.1-strength Hoagland's solution number 1 supplemented with micronutrients before water was withheld. Plants were maintained for at least 30 d in the greenhouse before soil drying commenced. The water content of the soil was determined by weighing 14 to 18 g of soil before and after drying for 48 h in a forced-draft oven at 105°C, and the soil water potential in the rooting zone was calculated using a moisture release curve for Agave Hill soil (Young and Nobel, 1986).

At 45 d of soil drying, five plants were watered so that Ψ_soil was increased to −0.1 MPa within 1 d and was maintained at that value by watering on alternate days. The five plants were in a 1:1:2 mixture of washed quartz sand:soil from Agave Hill:vermiculite, the latter to facilitate the excavation of individual roots after 1, 4, and 8 d of rewetting. New roots arising from the stem or from old woody roots were 250 to 300 mm long and averaged 1.8 mm in diameter after 30 d in wet soil. Two regions of such new main roots were examined: distal, from the tip to 80 mm back; and mid-root, from 120 to 200 mm back from the tip.

Anatomy

To investigate anatomical features, root segments were sectioned with a razor blade and stained with 0.05% (w/w) toluidine blue O in water for 30 s. Sections were mounted in water and examined with a BH2 microscope (Olympus, Lake Success, NY) at a magnification of 100× to 440×. Sections to be examined for lignin were stained with 0.5% (w/w) phloroglucinol in water followed by 20% (v/v) HCl (Jensen, 1962) and mounted in water. Suberin lamellae were stained with 0.1% (w/w) Sudan red 7B in 70% (v/v) ethanol. The sections were mounted in 75% (v/v) glycerin and examined under bright field. Suberin lamella appeared red, and Casparian bands were not stained. Suberin and lignin were also located in untreated sections by their autofluorescence under violet and UV light (Peterson et al., 1981).

Root Hydraulic Conductances (L_P, K_h, L_R, L_R,E/C/P, and L_R,S)

Root hydraulic conductance based on the root surface area (L_P, m s⁻¹ MPa⁻¹, also referred to as root hydraulic conductivity (Nobel et al., 1990; Henzler et al., 1999), was measured on individual distal and mid-root regions that were 60 to 80 mm long (Nobel et al., 1990). Root segments about 150 mm in length were gently excavated with a fine spatula and jets of water and rapidly trimmed by 50 mm under distilled water with a razor blade, leaving a segment 100 mm long. All tissues external to the stele were removed from a 15-mm length at the proximal end of the root segments, which prevented flow of water axially through or around cortical cells. The exposed stele was then trimmed 5 mm under water and immediately inserted into a 10-mm-long Tygon tubing (i.d. 0.9, 1.1, or 1.6 mm, depending on stele diameter) affixed to a glass capillary (i.d. 0.8, 1.0, or 1.6 mm) that was half filled with distilled water. A watertight seal between the stele and the tubing was achieved by inserting the tubing through a silicone gasket in a brass compression fitting (McCown and Wall, 1979). The junction between the tubing and the stele as well as the distal cut end of the mid-root region were sealed with hydrophilic vinyl polysiloxane (Reprosil, Dentsply International, Milford, DE) and coated with acrylic copolymer (clear nail protector). The same type of seal was applied to the bases of excised lateral roots, when present, which were cut at about 5 mm from the main root. The root segment was then suspended in 200 mL of distilled water.

Water flow through the root segment was induced by applying a partial vacuum, adjusted with a needle valve and monitored with a PS309 digital manometer (Validyne Engineering, Northridge, CA), to the open end of the capillary. The flow rate (Q_V, m³ s⁻¹) was determined by monitoring the movement of the meniscus in the capillary with a traveling microscope capable of resolving 0.01 mm. Pressure was first decreased to −40 kPa; after the flow rate was stabilized, usually within 20 min, the vacuum pressure was successively increased to −30, −20, and −10 kPa, and the flow rate was recorded at each pressure after it stabilized in less than 10 min. L_P was calculated as the slope of the relationship between the volumetric flux density (flow rate per unit root surface area; J_V, m s⁻¹) and the applied pressure difference. The root surface area was calculated from root length and mean diameter.

To measure K_h (m⁴ s⁻¹ MPa⁻¹), root segments were trimmed under water at about 20 mm from the proximal seal. About 1 mm at the cut end of the segment was immersed in distilled water. Q_V was used to calculate K_h:

2

where the pressure difference ΔP (10 kPa) was applied along the length Δx (m) of the root segment. Measurements of Q_V were made during the first 10 s after cutting, because Q_V decreased by about 30% at 60 s.

The radial hydraulic conducance (L_R, m s⁻¹ MPa⁻¹, calculated based on the outer root surface area) at any point along a root equals J_V at the root surface divided by the difference in water potential from the root surface to the root xylem. L_R averaged over the entire root segment was calculated by incorporating measured values of L_P and K_h together with the length (l, m) and the radius (r_root, m) of the root segment into a model of Landsberg and Fowkes (1978) based on leaky cable theory:

3

where α (m⁻¹) equals (2πr_rootL_R/K_h)^1/2_, which represents the length along the root xylem across which the pressure halves (Landsberg and Fowkes, 1978). L_R was initially set equal to L_P and gradually increased to solve Equation 3 by iteration.

Assessment of Aquaporins

After measurement of L_P in water, the root segments were transferred to 50 μm HgCl₂ for 10 min at a vacuum pressure of −30 kPa. Then the pressure was decreased to −40 kPa and L_P was measured as in water. The root segments were then briefly rinsed in water, immersed in 10 mm 2-mercaptoethanol under the same conditions (10 min, −30 kPa), and L_P was measured again. To check for artifacts due to repeated measurements of L_P on the same root segments, comparable segments were repeatedly measured in water with the same protocol; L_P varied less than 3% among three such measurements.

L_R of Concentric Root Tissues

L_R of the tissues external to the stele and that of the stele was measured by sequentially removing tissue layers (North and Nobel, 1996). After L_P was measured for an intact root segment, the epidermis and the cortex plus the endodermis for the distal root region or plus the periderm for the mid-root region were stripped from the stele using fine forceps under a stereomicroscope. L_P was then measured on the stele in water and in 50 μm HgCl₂, as described above. After L_P was measured for an intact root segment and its stele, K_h was measured and L_R was calculated using Equation 3 and the surface area of the root segment. Radial conductances are in series and are based on the outer surface area of the intact root segment in all cases (note that the same Q_V occurs across each of the concentric tissue layers). Thus, the reciprocal of L_R for an intact root segment equals the sum of the reciprocal of L_R for the epidermis/cortex/periderm (L_{R, E/C/P}) and the stele (L_{R, S}), so:

4

Because conductance of each layer was based on the outer surface area of the intact root segment, comparisons could be made between root segments of different lengths and diameters. In some cases, the epidermis/cortex and the periderm were removed sequentially so that the L_R of the periderm could be calculated using a relation analogous to Equation 4.

Statistics

All statistical analyses were done using SigmaStat 2.0 (SPSS, Chicago). Differences in L_P due to watering treatments were analyzed using one-way ANOVA (α = 0.05) followed by a Tukey's test, after verifying that the treatment effects were normally distributed with equal variance. Differences in L_P due to HgCl₂ treatment were analyzed using paired t tests, after verifying that the treatment effects were normally distributed with equal variance. Differences in L_R for the different tissue layers, as well as the effects of the watering treatments on L_R, were analyzed on log-transformed data using one-way ANOVA (α = 0.05) followed by a Tukey's test. Differences in the number of cell layers were analyzed using a Kruskal-Wallis test, followed by pair-wise testing. Data are presented as means ±1 se (n = no. of measurements).