A quick determination of root resistance to water transport in paddy rice

Abstract Hydraulic resistance in plants is one of the most important factors responsible for changes in leaf water potential that is an indicator of plant water stress. Although the hydraulic resistance to passive water transport (Rpa) is a robust index in paddy rice (Oryza sativa), measurement is both time-consuming and labour-intensive. Here, we describe on a quick method to measure hydraulic resistance to osmotic water transport (Ros) by measuring the xylem sap exudation rate and osmotic water potential. In a greenhouse experiment, Ros responded significantly to soil temperature, but under field conditions soil temperature varied considerably less than air temperature. In the field experiment, Ros of six rice cultivars at two growth stages was strongly positively correlated with Rpa. We conclude that measuring Ros could be used to evaluate root water transport capacity in paddy rice under conditions with adequate soil water.


Introduction
Water stress is an important issue in crop production because it decreases yield dramatically. In rice (Oryza sativa L.), even when soil water is adequate, water stress is observed as a mid-day and early afternoon depression of the rate of photosynthesis (Hirasawa & Hsiao, 1999;Ishihara & Saitoh, 1987). This phenomenon is caused by the closure of stomata in response to a reduction in the leaf water potential when transpiration of the leaf increases under a high vapour-pressure deficit (Hirasawa et al., 1989. To avoid this stomatal closure, plants must absorb enough water from their roots to meet the demand created by their leaves ; indeed, a rice cultivar with low hydraulic resistance tends to maintain higher stomatal conductance than a cultivar with high hydraulic resistance (Jiang et al., 1988).
In addition, Taylaran et al. (2011) showed that root water transport capacity affects both daytime and morning rates of photosynthesis under a low vapour-pressure deficit in paddy rice. A strong relationship between the rate of photosynthesis and hydraulic conductance, which is the inverse of hydraulic resistance, has been found in other plant species (Brodribb et al., 2007;Meinzer & Grantz, 1990). Hence, one important way to enhance photosynthesis would be to mitigate the mid-day depression by increasing the plant's water transport capacity.
A water transport capacity in plants can be measured in terms of the resistance to water transport by analogy with Ohm's Law (Steudle & Peterson, 1998). Hirasawa and Ishihara (1991) developed a procedure to calculate the hydraulic resistance to passive water transport from the roots to the leaves (R pa ) in paddy rice under conditions with sufficient soil moisture. This R pa is calculated from the leaf transpiration rate and the difference in water potential between the soil and the leaf. These authors also suggested that differences in R pa among plants could be used as an index of their root water transport capacity, since root hydraulic resistance often changes as plants age and in response to changes in growing conditions, whereas shoot hydraulic resistance remains unchanged (Hirasawa & Ishihara, 1991;. Although R pa has served as a robust index of root water transport capacity, measuring R pa remains time-consuming and labour-intensive, precluding widespread application of this method in field studies, since R pa must be determined for many plants within a short time period. A quick determination of root hydraulic resistance is therefore desirable, as it could improve and promote studies of crop physiology seedlings were transplanted on 17 May. The plant density was 22.2 hills m −2 (at a spacing of 30 cm × 15 cm), with three plants per hill. As a basal dressing, manure was applied at 20 t per ha, and inorganic fertilizer was applied at 50 kg N ha −1 , 136 kg P 2 O 5 ha −1 , and 72 kg K 2 O ha −1 . One-third of the total N was applied as ammonium sulphate; another onethird was applied as elution-controlled urea Chisso Asahi Fertilizer,Tokyo,Japan), and the final one-third was applied as elution-controlled urea (LPS-100; Chisso Asahi Fertilizer). No top dressing was applied. The experimental plots (each 3 m 2 ) were arranged in a randomized complete block design with six replicates. The plants were measured at the full heading stage (1 to 2 weeks after flowering) and at late maturity (4 weeks after flowering).

Hydraulic resistance to osmotic water transport (R os )
The R os (×10 3 MPa s g −1 stem −1 ) was calculated using the method of Miyamoto et al. (2001) as, where σ is the root-reflection coefficient for solutes in the xylem, Ψ soil os (MPa) is the osmotic potential of the soil solution immediately outside the roots, Ψ xylem sap (MPa) is the osmotic potential of the xylem sap, and E (g s −1 ) is the exudation rate per stem. The osmotic potentials of the paddy water (-0.012 MPa, on average) and of the surface water in the pots (-0.017 MPa, on average) were used as Ψ soil os . For σ, we used a value of 0.4, following Miyamoto et al. (2001). Stems were cut at 15 cm above the ground, and their exudates were collected in pre-weighed absorbent cotton placed on each cut surface. To prevent loss of water by evaporation, the cotton was covered with a plastic bag and shaded with aluminium foil. The collection of stem exudates began at 08:30, when E typically reaches its daily maximum (Hirasawa et al., 1983). E shows a time-dependent decrease after cutting off the stems (Morita & Abe, 2002). Since E didn't change within 4 h after cutting and then decreased in our preliminary experiment, we collected the exudates for 3 h. To determine E, we used the increase in weight of the cotton divided by the collection time. After weighing, the collected exudates were quickly frozen at -80 °C and stored until the osmotic potential could be measured. After thawing the samples overnight at 4 °C, we measured the osmotic potentials of the stem exudates and those of the pot and paddy water with a freezing-point osmometer (OM802; Vogel, Giessen, Germany), and calculated R os of the potted plants and the field plants. For the field-grown plants, we selected warm and sunny days for sampling. Six replicates were used and the values from two or three hills were averaged for each replicate.
and support breeding to select plants with superior water transport. Such a method would be particularly suitable for making comparisons among different environments, analysing developmental profiles, evaluating natural genetic variation in water relations, screening of mutants, and analysing quantitative trait loci. The hydraulic resistance to osmotic water transport (R os ) offers an alternative index for evaluating root water uptake capacity. It is well known that the xylem sap rises spontaneously from the cut surface of a plant stem (Schurr, 1998). This can be explained as follows: active xylem loading increases the concentration of ions inside the xylem relative to the surrounding medium (the apoplast of the stele), thereby generating an osmotic gradient that attracts water into the xylem vessels and creating a force that pushes the xylem sap towards the cut stem surface (Schurr, 1998). Therefore, by measuring the gradient of osmotic potential between the xylem sap and the solution around the roots, and combining this with measurements of the exudation rate, it becomes possible to calculate R os (by analogy with Ohm's Law). In this way, several studies have evaluated the hydraulic conductance to osmotic water transport in maize (Zimmermann & Steudle, 1998) and rice (Miyamoto et al., 2001;Sakurai-Ishikawa et al., 2011). However, these studies were conducted under laboratory conditions, and this method has not yet been applied to rice grown in the fields. To determine the method would also be suitable under filed conditions, we examined whether R os could serve as an alternative to R pa to determine the water uptake capacity of paddy rice.

Plant materials
In our first experiment, we grew rice plants, ('Sasanishiki'), in a greenhouse under natural sunlight. Seeds were sown on 14 April 2007. The seedlings were transplanted on 8 May 2007 into 1/5000a Wagner pots filled with a 1:1 (v/v) mixture of paddy soil (an alluvial clay loam) and upland soil (a diluvial volcanic ash), at 10 plants per pot. Within each pot, the plants were arranged in a circle around the periphery (Satake et al., 1969). A compound of NPK fertilizer (1.0 g each of N, P 2 O 5 , and K 2 O per pot) was applied at planting; additional N fertilizer was applied as ammonium sulphate at 0.3 g per pot at the tillering stage and then again at 1.0 g per pot at the booting stage. The pots were watered daily to maintain the submerged growing conditions.
In our second experiment, we grew rice plants in a paddy field (an alluvial clay loam) at the University farm of Tokyo University of Agriculture and Technology (35°40'N, 139°28'E). We used six rice cultivars: 'Sasanishiki, ' 'Koshihikari, ' 'Nipponbare, ' 'Akenohoshi, ' 'Takanari, ' and 'Habataki. ' Seeds were sown on 26 April 2007 and the To evaluate the temperature dependency of E, Ψ xylem sap , and R os , we manipulated the soil temperature by putting the pots in a water bath 12 h before the measurements. The water temperature was controlled within a range from 15 to 40 °C, with a heater and a cooler that were regulated and monitored by a thermocouple thermometer. The exudate was collected only from the main stems. During collection, we covered each plant with a polyethylene bag to limit the effects of transpiration from the remaining plant parts. Three plants grown in a single pot were measured at the full heading stage and an average value was calculated.

Hydraulic resistance to passive water transport (R pa )
The R pa (×10 6 MPa s m −1 ) was calculated using the method of Hirasawa and Ishihara (1991) as where Ψ soil (MPa) is the water potential of the soil immediately outside the roots, Ψ leaf (MPa) is the water potential of a single leaf, and T (mmol m −2 s −1 ) is the steady-state transpiration rate per unit leaf area. Since the plants were grown under submerged conditions, so Ψ soil was much higher than Ψ leaf , and therefore Ψ soil could be regarded as negligible (assigned a value of zero). The T of an intact leaf was measured in an air-sealed acrylic assimilation chamber (Tsunoda, 1974) under natural sunlight. Air, with a dew point controlled at 10 (±0.1) °C, was pumped into the chamber at a rate of 6.67 × 10 −5 m −3 s −1 . The humidity of the air pumped into and out of the chamber was measured with a dew point hygrometer (Model 660; EG & G Inc., Waltham, MA, USA). Once T reached a constant value, the leaf water potential was measured in a pressure chamber (Model 3005; Soil Moisture Equipment Inc., Santa Barbara, CA, USA). Six replicates were used and the values of four to six leaves per replicate were averaged. The measurements were conducted from 09:00 to 15:00 during the period of 13 August to 20 September 2007.

Statistical analysis
Differences and interactions between the treatment effects on the E, Ψ xylem sap , and R os response variables were tested separately using two-way analysis of variance (ANOVA) In each ANOVA, the two treatment factors were rice cultivar (six cultivars) and plant stage of development (full heading vs. late maturity). All tests were performed in JMP software v. 12 (SAS Institute, Cary, NC, USA).

Results and discussion
E in rice is strongly affected by soil temperature (Yamaguchi et al., 1995). However, the responses of Ψ xylem sap and R os to soil temperature have not been determined before now. As we expected, the value of E increased with increasing soil temperature (Figure 1(A)), but only in the range of 15-30 °C; E peaked at about 30 °C and then decreased as temperature increased to 40 °C. In contrast, Ψ xylem sap varied little over the range of soil temperatures examined (Figure 1(B)). As a result, R os decreased from 20 to 33 °C and slightly increased in the range of 33 to 40 °C (Figure 1(C)). These results indicate that R os can respond to changes in soil temperature and that the response is likely driven by changes in E.
Along with air temperature, we monitored the rice paddy field's actual soil temperature at a depth of 15 cm from 19 August to 28 September in 2007 (Figure 2). The average air temperature during the period when most of the sampling occurred (i.e. from 09:00 to 12:00) varied greatly (by 13 °C), from 17 to 30 °C, whereas soil temperature varied less (by 4.9 °C, from 20.5 to 25.4 °C). The specific (A) (B) (C) Figure 1. dependence on soil temperature of (a) the exudation rate from the cut surface of stem the (E), (B) the water potential of the xylem sap (Ψ xylem sap ), and (c) the hydraulic resistance to osmotic water transport (R os ). the plants were grown in a greenhouse in 1/5000a Wagner pots; each symbol represents the average of three plants grown in a pot.
We compared E, Ψ xylem sap , and R os of the six cultivars at full heading and late maturity (Table 1). There were significant between-cultivar differences in E at both stages. 'Akenohoshi' had the highest E at both stages. E was always higher at full heading than at late maturity; significant differences were detected between stages, and there was a significant cultivar × stage interaction. There were also significant differences in Ψ xylem sap among the cultivars at both stages, although the range of values was smaller than that for E. Ψ xylem sap was always higher (less negative) at full heading than at late maturity, and significant differences were detected at both stages; the cultivar × stage interaction was also significant. There were significant differences in R os among the cultivars at both stages. 'Akenohoshi' had the lowest R os among the cultivars at both stages owing to its higher E at full heading and its higher E and higher (less negative) Ψ xylem sap at late maturity than the other cultivars. R os was always lower at full heading than at late maturity, and the difference between stages and the cultivar × stage interaction were also significant. These results are consistent with those of Jiang et al. (1988), who showed that 'Akenohishi' maintained a lower R pa than 'Nipponbare' throughout the ripening stage, and that R pa of both cultivars gradually declined as the plants aged, suggesting that R os may reliably reflect variable differences in root water transport capacity among rice cultivars and between growth stages.
We also measured the R pa of these rice plants to examine the relationship between R os and R pa across the six cultivars and two growth stages (Figure 3). We found that R os provides a good estimate of R pa (R 2 = 0.91, n = 11). Taken together, our results suggest that R os can be as reliable as R pa as an index of water uptake capacity in paddy rice.
heat of water is 4.18 J g −1 K −1 , which is four times that of air, so the soil (which is in contact with the water) will change temperature more slowly. Furthermore, canopy vegetation cover can slow the transfer of heat between the soil and the air (Takami et al., 1989). Both factors might explain the smaller changes in soil temperatures. Moreover, we selected warm and sunny days for sampling in the paddy field; excluding the data for one cool day, soil temperature ranged from 24.6 to 26.6 °C. The effect of this variation on R os can be ignored, since the changes in R os over this range of soil temperature were much smaller than the betweenplant variation at a given soil temperature (Figure 1(C)). Figure 2. the relationship between soil and air temperatures in the rice paddy field from 19 august to 28 September 2007. the soil temperature was measured 15 cm below the soil surface. the air temperature was measured at 30 cm above the ground in the same field. each symbol is the average of recordings from 09:00 to 12:00, which corresponds to the time when stem exudates were sampled. the dashed line indicates a 1:1 bivariate relationship (y = x). Table 1. the exudation rate from the cut stem surface (E), water potential of the xylem sap (Ψ xylem sap ), and the hydraulic resistance to osmotic water transport (R os ) in six rice cultivars grown in a paddy field.
Values are means ± Se (n = 6). nd, not determined. Significance: *p < 0.05; **p < 0.01; ***p < 0.001. data were not obtained for 'takanari' at late maturity. the anoVa for cultivar at the full heading stage was performed for six cultivars, whereas the anoVa for cultivar at late maturity, stage, and cultivar × stage was performed with only five cultivars, excluding 'takanari' . cultivar Full heading *** *** ** late maturity *** * ** Stage *** *** *** cultivar × Stage * *** ** highly impeded in rice and that a large portion of the water moves via the cell-to-cell pathway. If so, this might explain the close relationship we observed between R pa and R os . Although we were unable to directly compare R pa and R os in the same units (because R pa should be expressed per unit leaf area), we can still use R os to compare root water uptake capacity among cultivars and growth stages. The measurement of R pa is time-consuming and labour-intensive. For example, it typically takes 5-10 min per leaf to obtain a stable transpiration rate after inserting a leaf into the chamber. The measurement of Ψ leaf then takes an additional 5 min. To do this work, it requires two persons: one to measure the transpiration rate and another to measure Ψ leaf . This method also requires a humidity control system, a plastic chamber, a dew-point hygrometer for measuring T, and a pressure chamber to measure Ψ leaf (Hirasawa & Ishihara, 1991). In contrast, measuring R os takes a single person only 2 min, to set the absorbent cotton upon the cut stem surface. After the exudate has been collected, the cotton can be stored in a freezer before measurements of its weight increase (i.e. the amount of exudate absorbed) and Ψ xylem sap . The osmotic method demonstrated in this paper therefore offers the large advantages of simplicity and ease of measurement, thus enabling plant researchers to process more samples per unit of sampling time than they could with the passive method.

Cultivar
In this paper, we have shown how a quick method to measure the hydraulic resistance to root water transport (an index of root water uptake capacity), could be used under conditions with adequate soil water by simply evaluating the exudation rate and the difference in osmotic potential between the soil and the exudates. The proposed method offers a promising approach for evaluating the root water uptake capacity of different rice cultivars, investigating lines grown in several areas under a wide range of environmental conditions, and supporting genetic studies designed to identifying quantitative trait loci and genes associated with differences among rice cultivars in their root water transport capacity.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported in part by the Japan Science and Technology Agency, Precursory Research for Embryonic Science and Technology (to S. A.), by the Ministry of Agriculture, Forestry and Fisheries of Japan (Genomics-based Technology for Agricultural Innovation, RBS-2006, to T. H.), and by the Global Innovation Research Organization in Tokyo University of Agriculture and Technology (to S. A. and T. H.).
Passive water transport occurs in the xylem between roots and leaves along water potential gradients generated by transpiration (Zimmermann et al., 1995). Water transport is thought to occur via an apoplastic bypass, especially during transpiration (Steudle & Peterson, 1998), whereas osmotic water transport occurs via active xylem loading (Schurr, 1998), as described in the Introduction. Osmotic water transport is dominant under low light or dark conditions, which are characterised by low transpiration rates, during which time water flows mainly through cell-to-cell pathways (Steudle & Peterson, 1998). The mechanisms underpinning these two modes of water transport are different, and the amount of water transported by osmotic transport is considerably less (by a factor of 10) than that transported by passive transport (Steudle & Peterson, 1998).
Given these differences, why did we find such a close relationship between R pa and R os (Figure 3)? The axial hydraulic conductance of roots is much larger than their radial conductance (Frensch & Steudle, 1989). Thus, the hydraulic conductance of whole root system is determined by the product of root surface area and root hydraulic conductance per unit surface area, which is generally referred to as root hydraulic conductivity (L pr ). However, the root surface area of the neighbouring plants used in the measurements of R pa and R os should be very similar. During the radial passage of water, potential barriers include several root structures: the exodermis, sclerenchyma, aerenchyma, and endodermis. Miyamoto et al. (2001) suggested that the root endodermis is the main barrier to water transport in rice. Furthermore, they found that the values of L pr measured osmotically and hydrostatically were similar in rice, but differed significantly in maize (Zea mays). This suggests that the apoplastic pathway in passive water transport is Figure 3. the relationship between hydraulic resistance to osmotic water transport (R os ) and hydraulic resistance to passive water transport (R pa ) for rice plants of six cultivars grown in a paddy field. the filled and open symbols indicate the measurements at full heading and late maturity, respectively. the dashed curves represent the 95% confidence interval. Values are means ± Se for n = 6 replicates per symbol.