Springer 2006

Plant and Soil (2006) 280:65–76 DOI 10.1007/s11104-005-2508-9

Mechanical impedance and nutrient acquisition in rice Donna A. Brown1,3, Lawrence J. Clark2,4, Jonathan R. Howarth1, Saroj Parmar1 & Malcolm J. Hawkesford1 1

Crop Performance and Improvement Division, Rothamsted Research, AL5 2JQ, Harpenden, Hertfordshire, UK. 2Agriculture and Environment Division, Rothamsted Research, AL5 2JQ, Harpenden, Hertfordshire, UK. 3 Present address: The Physiological Society, P.O. Box 11319, London, WC1X 8WQ, UK. 4Corresponding author* Received 22 December 2004. Accepted in revised form 26 August 2005

Key words: gene expression, mechanical impedance, rice

Abstract The aim of this work was to examine the effect of mechanical impedance on nutrient acquisition and gene expression in rice (Oryza sativa L.). Roots were mechanically impeded in a sand-core apparatus to vary impedance independently of aeration and water status. The effect of impedance on plant growth, anion concentration and expression of genes for anion transporters was compared for six varieties with differences in root penetration ability. Impedance decreased shoot growth more than root growth in all varieties, resulting in increased root/shoot ratios. Impedance substantially increased shoot tissue nitrate concentration in all varieties but only caused a small increase in shoot sulphate and phosphate concentrations. High impedance increased expression of the sulphate transporter OsST1 in five varieties, which was associated with decreased sulphate concentration in root tissues. In contrast, impedance decreased expression of the phosphate transporter OsPT2 expression in all varieties, which was associated with decreased phosphate concentration in root tissues. Localisation of expression of the sulphate transporter by in situ hybridisation indicated high levels of expression in lateral bud primordia. It was suggested that the decreased root phosphate concentrations of impeded roots were caused by low phosphate transporter gene expression, while the increase in sulphate transporter gene expression was due to a derepression mechanism of control.

Introduction Roots experience mechanical impedance due to the force required to displace soil particles as they elongate (Bengough and Mullins, 1990). As soil strength increases, root elongation rate decreases due to the increasing resistance of soil particles to displacement, although compensatory growth may occur in weaker horizons. Strong soil can be a serious agricultural problem, as the ability of the root system to access water and nutrients from the deeper soil layers is restricted (Barraclough and Weir, 1988). * E-mail: [email protected]

In lowland rice (Oryza sativa L.) fields, strong soil in the form of a hardpan is very common and restricts rooting at depth, so restricting nutrient uptake (Kundu et al., 1996) and making rainfed crops more vulnerable to drought (Wade et al., 1999). While it has long been recognised that deep rooting enhances drought resistance in upland rice (Yoshida and Hasegawa, 1982), the challenges of breeding root systems that will confer drought resistance in the rainfed lowlands have only recently been recognised (Wade et al., 1999). There is evidence that rice cultivars differ in their ability to penetrate hardpans in the field (Samson et al., 2002) and in laboratory screens (Clark et al., 2002; Yu et al., 1995). A wax layer

66 laboratory screen can reveal very large differences in root penetration ability, although these differences vary with screening conditions and assessment criteria (Clark et al., 2000). An interesting feature of the behaviour of different rice varieties is that despite large differences in their ability to penetrate strong wax layers, there is little difference in maximum rooting depth in uniformly strong sand (Table 1). However, those varieties that had greater root diameter when exposed to uniform mechanical impedance had better penetration of wax layers. Mechanical impedance increased the diameter of rice roots, as has been observed in other species (Atwell, 1990; Materechera et al., 1991). There is evidence that mechanical impedance causes changes in gene expression in maize roots (Huang et al., 1998), although it is not yet clear how these changes may be responsible for responses of plants to impedance. A major role of roots is in mineral nutrient acquisition and there is evidence that impedance may affect nutrient concentrations in roots (Atwell, 1990). We therefore investigated the effect of mechanical impedance on expression of genes encoding phosphate and sulphate ion transporters in rice roots, together with the parallel changes in tissue anion concentrations. Six varieties that differ in their ability to penetrate wax layers in the laboratory and hardpans in the field were compared. Compacted soil was not used to impose mechanical impedance due to the interactions of Table 1. Effect of mechanical impedance on root growth in six rice cultivars Variety

Number of root axes penetrating

Longest nodal root axis in strong sand (non-flooded)

80% wax 60% wax Length Diameter layer layer (mm) (mm) (non-flooded) (flooded) Azucena Bala IR20 CT9993 KDML 105 IR58821

5.4 1.2 – – – –

5.2 2.5 2.1 3.8 1.1 5.8

258 298 262 320 256 272

1.09 0.64 0.68 0.84 0.56 0.94

Data are taken from experiments reported in Clark et al. (2000), Price et al. (2000) and Clark et al. (2002).

impedance with water and aeration. Instead, a sand-core system was used to vary mechanical impedance independently of aeration and water status.

Materials and methods Strategy The impact of mechanical impedance on the anion concentration of root and shoot tissues of six varieties was measured, together with an analysis of expression of genes of two transporters involved in anion acquisition. The varieties in Table 1 were studied, as their responses to mechanical impedance in controlled environments are known. In addition, Samson et al. (2002) reported the rooting behaviours of IR20, CT9993, KDML 105 and IR58821 in rainfed lowland fields and found that IR58821 has relatively good hardpan penetration. Azucena and Bala are parents of a recombinant inbred line mapping population that has been used to study the genetics of root traits in rice, including wax layer penetration (Price et al., 2000). Here, the effect of mechanical impedance on gene expression was studied in sand cores, rather than wax layers, in order to impede the whole root system and obtain sufficient plant material for analysis. Plant material and growth conditions Experiments were carried out in controlled environments using a 14 h day-length with day/night temperatures of 30 and 20 C, respectively, a relative humidity of 60% (day) and 70% (night). The photosynthetic photon flux density was 300 lmol m)2 s)1 by fluorescent tubes, supplemented by tungsten bulbs. Seeds of rice were set to germinate between two sheets of wet filter paper in Petri dishes wrapped in aluminium foil to exclude light. After 4 days, the radicle was about 10–20 mm long and the seedlings were then grown in the sand-core screen under the above environmental conditions. A hole for the radicle was made in the sand core using forceps and the seedlings transplanted with the seed just below the surface of the sand. Two seedlings were planted per core to give sufficient plant material for analysis.

67 Sand-core screen The screen was developed from the sand-core method of Materechera et al. (1991), as described by Clark et al. (2002). The apparatus allows mechanical impedance to be varied independently of aeration and water status of the sand (RH 65 grade silica sand, WBB Minerals, Sandbach, UK, with a particle size distribution by percentage of mass retained of 710 lm, 0.1%; 500 lm, 0.7%; 355 lm, 7.3%; 250 lm, 28.3%; 180 lm, 41.5%; 125 lm, 19.2%; 90 lm, 2.4%; 63 lm, 0.4%; <63 lm, 0.1%). The sand is allowed to pack under its own weight in an excess of nutrient solution, which is then allowed to drain to a water table 0.3 m below the top of the sand core. When a weight is placed on the sand surface, a confining pressure is imposed. This increases the strength of the medium as the resistance of sand grains to displacement is increased, but there is negligible compaction of the sand, as the grains of sand are already well consolidated. This approach therefore contrasts with approaches where moist sand is compressed to different bulk densities, such as that of Croser et al. (1999). Each sand-core apparatus consisted of a tank containing six tubes in a 3  2 arrangement (arrangement of one tube is shown in Figure 1). The sand above the nutrient solution level was kept watered by capillary action throughout the experiment. The nutrient solution composition was 1.5 mM Ca(NO3)2, 0.15 mM CaH4(PO4)2, 1.0 mM KCl, 0.3 mM MgSO4, with the following micronutrients: 50 lM B, 50 lM Fe, 10 lM Mn, 1 lM Zn, 1 lM Cu and 0.5 lM Mo. This provided an approximately 10-fold excess of nutrients as there were 15 L of nutrient solution available per tube. The tubes were lined with polytetrafluoroethylene to decrease friction between the sand and the tube and so increase the depth to which the effect of the weight was transmitted. In contrast to a wax layer screen, roots appear to encounter relatively uniform mechanical impedance with depth. Preliminary experiments with cv. IR36 showed that the high impedance treatment decreased the elongation of the seminal root axis by 44% compared to the low impedance control, regardless of whether the control roots had reached depths of 100, 200 or 300 mm.

The effect of two levels of impedance was tested: ‘impeded’ and ‘control’. Immediately after planting, a steel 200 N weight was placed on the plastic disc for the impeded treatments (giving a pressure of 11.6 kPa on the sand), while the controls used a ‘mock weight’ made of plastic foam to simulate the physical environment around the shoot caused by the weight. The control and impeded treatments were applied to six varieties [Azucena, Bala, IR20, CT9993-5-10-1-M (CT9993), Khao Dawk Ma Li 105 (KDML 105) and IR58821-23-B-1-2-1 (IR58821)] as a 2  6 factorial. The nutrient solution in the tanks was changed 3 weeks after planting. Plants were harvested 5 weeks after planting, with the aid of a root harvesting apparatus (Clark et al., 2000). The lengths of the longest root axis of each tube were measured at harvest, together with root and shoot fresh weights. Plants were frozen in liquid nitrogen within 5 min of the start of the root washing procedure, and subsequently kept at )80 C. In the first experiment, a randomised complete block design was used, with three replicates. Two tanks were treated as a block. In the second experiment, a completely randomised design was used. Different numbers of replicate cores (at least two) were used for each treatment combination so that there would be sufficient plant material for analysis from all the treatments. The first experiment was used to obtain plant material for northern analysis and the second for in situ hybridisation and anion determination. Nitrate, phosphate and sulphate Anions were measured by extracting approximately 20 mg of freeze-dried, homogenised plant material in 1 ml of deionised water at 80 C for 30 min, after which the extract was filtered through a 0.45 lm filter. Anion concentrations in the extracts were determined by ion chromatography (Dionex 2000i/sp) using an AS9SC separation column fitted with an AS9G guard column (Dionex, Sunnyvale, CA, USA). The eluent solution consisted of 1.8 mM Na2CO3, 1.7 mM NaHCO3. Northern analysis OsST1 and OsPT2 were obtained from Dr Frank Smith (CSIRO–Brisbane). Primers were designed

68 to the coding regions to generate fragments of approximately 300–500 bp. The primers employed were: OsST1F, 5¢-GTACAAGGACCAG CCGATG-3¢; OsST1R, 5¢-GCCTTTAAGCTGC TGAAGGGC-3¢; OsPT2F, 5¢-GGTCATGTAC

GCCTTCACC-3¢; OsPT2R, 5¢-GAAGAACCC GATGACGTTG-3¢. PCR products were separated by agarose gel electrophoresis and purified using the QIAquick Gel Extraction kit (Qiagen, Valencia, CA, USA). Purified DNA was cloned

Figure 1. Sand-core apparatus showing arrangement of one tube. Redrawn from Clark et al. (2002).

69 into pGEM-T Easy vector (Promega, Madison, WI, USA) and sequenced on an ABI PRISM 310 genetic analyser using an ABI PRISM BigDye Termination kit (Applied Biosystems, Foster City, CA, USA). RNA was extracted by the method of Verwoerd et al. (1989). Ten micrograms of total RNA was fractioned on a 1% agarose/formaldehyde gel and blotted onto Hybond-N+ nylon membrane (Amersham, Buckinghamshire, UK) as described in Sambrook et al. (1989). Membranes were probed overnight at 65 C using specific OsST1 or OsPT2 gene fragments radiolabelled with [a-32P]dCTP (Amersham) using Prime-AGene oligonucleotide labelling system (Promega). Equal loading of RNA in each lane was verified by probing with a constitutively expressed 18S rDNA fragment. Membranes were washed to high stringency in 1 SSC, 0.1% SDS at 65 C before visualising on Kodak Biomax MS (Kodak, Rochester, NY, USA) autoradiography film. In situ hybridisation cDNA probes for in situ hybridisation were labelled using a DIG In Vitro Transcription kit (Roche, Basel, Switzerland). A 1-kb OsST1 cDNA fragment was subcloned into pGEM-T Easy Vector system (Promega) and amplified by PCR using M13 reverse and forward primers. One microgram of the PCR product was used as template to synthesise DIG labelled RNA using T7 RNA or T3 RNA polymerase. Root sections were fixed in 4% formaldehyde, embedded in paraffin wax and processed and hybridised under the conditions described by Gotor et al. (1997). Colorimetric alkaline phosphatase detection of hybridisation was carried out as described in the DIG non-radioactive detection kit protocol (Roche). Slides were developed in the dark at room temperature for 3–7 days. Statistical analyses Plant growth data and anion concentration data were subjected to residual maximum likelihood (REML) variance components analysis (Payne, 2000). This approach was preferred to analysis of variance because of the combination of the factorial design and the different numbers of replicates

in experiment two. In all analyses, the fixed model was variety*impedance. For anion analyses from experiment two, the random model was tube. Growth data from both experiments were analysed together, with an experiment/block/tube random model. The significance of main effects and interactions was estimated from the Wald statistics. Predicted means from REML analysis are shown, together with standard error of differences (SED).

Results The impedance treatment had a substantial effect on plant growth in all varieties (Figure 2). High impedance decreased maximum root length by about 30% (Figure 2a), root fresh weight by about 60% (Figure 2b) and shoot fresh weight by 75% (Figure 2c). As a consequence, the root to shoot ratio increased from 1.10 to 1.75, with SED 0.065 (there was no significant effect of variety). It was evident that high impedance affected plant growth similarly in the six varieties. RNA was extracted from all root samples and expression of genes encoding two anion transporters was determined by northern analysis in the six varieties (Figure 3). The group 1 sulphate transporter OsST1 (Hawkesford, 2003), most likely responsible for initial uptake into the root symplasm, showed greater expression in impeded tissue than in the control in all varieties except IR58821. The expression was most highly induced in Azucena, Bala and IR20. Decreased expression of a phosphate transporter, OsPT2, was observed in response to impedance in all varieties. Sulphate, phosphate and nitrate concentrations were analysed in root and shoot tissues of the six varieties for both control and impeded treatments. Impedance increased shoot concentrations of these ions, with the greatest increase seen for nitrate (Figure 4). Variety by impedance interactions were not significant for phosphate and sulphate (P > 0.4), although there was evidence for an interaction for nitrate (P = 0.06). The effect of impedance on root nitrate concentrations was not significant, although different varieties had different nitrate concentrations (Figure 5a). All impeded roots had lower phosphate and sulphate concentrations than the

70

Figure 2. Impact of mechanical impedance on growth of rice. (a) Maximum root length, (b) root fresh weight and (c) shoot fresh weight for control plants (open bars) and impeded plants (hatched bars). Data shown are means from both experiments. The bar is the SED for comparison of interaction means from REML analysis.

71

Figure 3. Northern blots showing expression of sulphate transporter (ST: OsST1: AF493792) and a phosphate transporter (PT: OsPT2: AF493789) in roots of the six rice varieties grown under control and impeded conditions. Approximately 10 lg RNA was loaded per track. An approximately 1 kb 18S rDNA fragment isolated from wheat by RT-PCR using sequence was hybridised to the filter to verify equal loading of the samples (18S).

controls (Figure 5b, c), although effects of variety or interactions were not significant (P > 0.1). In situ hybridisation using OsST1 demonstrated localisation of expression of the specific gene in root tissues (Figure 6). Expression in all sections was similar for both control and impeded root sections. Expression was observed in the exodermis and cells in ribbons of cells of the cortex, where aerenchyma were present (Figure 6a, b). No difference in aerenchyma formation between control and impeded roots was observed. Most noticeable was strong expression in the tissues where lateral roots were initiating (Figure 6c, d).

Discussion The use of the sand-core system, rather than compacted soil, was intended to allow mechanical impedance to be varied independently of aeration, water status and root–soil contact. It is possible that the high impedance treatment may have indirectly affected aeration, to the extent that the root systems were shallower when impeded and located further above the water table. The water release characteristics of this sand are such that air-filled porosity is <1% at 0.10 m above the water table, 5% at 0.20 m above the water table and 18% at 0.30 m above the water table (W. R. Whalley, unpublished results). However, it seems unlikely that this effect could have explained differences between varieties in their

responses to impedance, as the roots of different varieties grew to broadly similar depths. Any differences in matric potential between control and impeded plants would have been negligible, as the maximum height above the water table corresponds to )3 kPa. Root and shoot growths of the six rice varieties responded similarly to high impedance. This is consistent with previous data from sand cores, where root systems were challenged with high impedance (Clark et al., 2002). However, comparisons between these varieties can show large differences in response to high impedance when this occurs as a sudden increase in impedance, such as that provided by a wax layer (Table 1) or a hardpan (Samson et al., 2002). It appears that better ability to penetrate a wax layer or a hardpan does not necessarily mean greater elongation through strong sand. Although varietal differences in expression of phosphate and sulphate transporters were seen in this study, these cannot be related simply to growth responses to impedance. Both sulphate and phosphate transporters are encoded by large gene families with functional specialisation of the subtype (Hawkesford, 2003; Mudge et al., 2002). Northern analysis of a representative gene for each family and encoding a transporter involved in initial root uptake demonstrated distinct patterns of expression in response to impedance. In all varieties, high impedance decreased expression of the phosphate transporter, possibly because the reduced growth resulted in a decreased demand for phosphate. In contrast, high impedance increased abundance of transcripts of the sulphate transporter in five of the varieties. The increased sulphate transporter expression seen under impeded conditions may be the result of higher expression in individual cells or may be a consequence of altered root morphology. Increased abundance of lateral root initiation in impeded conditions has been reported (Atwell, 1988; Iijima and Kono, 1991). High levels of expression of the sulphate transporter were seen in initiating laterals (Figure 6c, d). Whilst there was little observable difference in levels of expression between control and impeded roots as observed by in situ analysis, this was not the case with the northern analysis; generally northern analysis is a better quantitative method for comparing multiple treatments. Increased sulphate

72

Figure 4. Anion concentration of shoot tissue of rice. (a) Nitrate, (b) phosphate and (c) sulphate concentration of shoot tissue for control plants (open bars) and impeded plants (hatched bars). The bar is the SED for comparison of interaction means from REML analysis.

73

Figure 5. Anion concentration of root tissue of rice. (a) Nitrate, (b) phosphate and (c) sulphate ion concentration of root tissue for control plants (open bars) and impeded plants (hatched bars). The bar is the SED for comparison of interaction means from REML analysis.

74

Figure 6. In situ hybridisation of sulphate transporter OsST1 in cv. Bala. (a) OsST1 hybridisation to a section within 2 mm of the apex of an impeded root, showing exodermal and endodermal expression; (b) control sense probe hybridisation against impeded root section; (c) impeded and (d) unimpeded root tissue, showing high expression in developing lateral primordial. Sections in (b), (c) and (d) were taken from mature root >20 mm from apex. Solid bars = 0.1 mm.

transporter expression is usually the result of a derepression mechanism of control (Smith et al., 1997) when demand for S exceeds supply. To ascertain the nutrient status of the plants and the impact of modified transporter gene expression, anion concentrations of roots and shoots were determined. Notably, high impedance decreased both sulphate and phosphate concentrations of root tissues. These are conditions where a derepression of expression would be expected, as seen for the sulphate transporter. Although variety by impedance interactions in root sulphate concentration were not significant, it is interesting that the largest decreases were seen in Azucena, Bala and IR20, the varieties in which greatest derepression of the sulphate transporter expression occurred. This suggests that a decreased S-nutritional status was affecting expression of this gene. This is in contrast to the data of Atwell (1990), who reported enhanced sulphate ion concentration in impeded wheat roots.

The decreased root phosphate concentration in response to high impedance may have been a consequence of the lower phosphate transporter expression under impeded conditions. It is unlikely that differences in expression of the phosphate transporter were due to derepression of expression in the controls due to demand exceeding supply, as root phosphate concentrations were greater in the controls than in the impeded plants, and in any case, nutrients were provided in excess. High impedance increased shoot anion concentrations, especially for nitrate. The most likely explanation is that the decreased growth caused an excess accumulation of nitrate. The decreased shoot growth itself was presumably caused by root to shoot signalling of mechanical impedance (Masle and Passioura, 1987) or decreased root volume (Ternesi et al., 1994). It is also possible that the accumulation of nitrate may have triggered the increased sulphate transporter

75 expression, as crosstalk between the pathways influencing gene expression is likely (Hawkesford and Wray, 2000). The analysis of the transporter gene expression and the ion concentration of the plant tissues indicated the substantial impact of mechanical impedance on nutrient acquisition. This has implications for crop nutrition, crop quality and tissue diagnosis of nutrient status as crops with impeded roots may have imbalances in shoot nutrients. Clearly identification of impedance conditions, in for example compacted soils, is a priority and identification of shoot-specific, impedance-regulated gene expression is one approach that would facilitate diagnosis of root-impeded crops. Conclusions As well as the expected decrease in root and shoot growth, mechanical impedance also led to changes in shoot and root anion concentrations. High impedance increased phosphate, sulphate and nitrate concentrations of shoot tissue but decreased phosphate and sulphate concentrations of root tissue. In root tissue, high impedance decreased expression of the phosphate transporter OsPT2 but increased expression of the sulphate transporter OsST1. Acknowledgements Rothamsted Research is grant-aided by the Biotechnology and Biological Sciences Research Council. This project was supported by a BBSRC ROPA award (9912405). We thank Dr W. R. Whalley, Rothamsted Research, for sharing data on air-filled porosity of the sand, and Laura Hopkins and J.J. Zhou for assistance with tissue harvesting and preparation. Rice seeds were a gift of the International Rice Research Institute, Philippines.

References Atwell B J 1988 Physiological responses of lupin roots to soil compaction. Plant Soil 111, 277–281. Atwell B J 1990 The effect of soil compaction on wheat during early tillering. II. Concentrations of cell constituents. New Phytol. 115, 37–41.

Barraclough P B and Weir A H 1988 Effects of a compacted subsoil layer on root and shoot growth, water use and nutrient uptake of winter wheat. J. Agr. Sci., Cambridge 110, 207–216. Bengough A G and Mullins C E 1990 Mechanical impedance to root growth: A review of experimental techniques and root growth responses. J. Soil Sci. 41, 341–358. Clark L J, Aphale´ S L and Barraclough P B 2000 Screening the ability of rice roots to overcome the mechanical impedance of wax layers: Importance of test conditions and measurement criteria. Plant Soil 219, 187–196. Clark L J, Cope R E, Whalley W R, Barraclough P B and Wade L J 2002 Root penetration of strong soil in rainfed lowland rice: Comparison of laboratory screens with field performance. Field Crops Res. 76, 189–198. Croser C, Bengough A G and Pritchard J 1999 The effect of mechanical impedance on root growth in pea (Pisum sativum). I. Rates of cell flux, mitosis, and strain during recovery. Physiol. Plant. 107, 277–286. Gotor C, Cejudo F J, Barroso C and Vega J M 1997 Tissuespecific expression of ATCYS-3A, a gene encoding the cytosolic isoform of O-acetylserine(thiol)lyase in Arabidopsis. Plant J. 11, 347–352. Hawkesford M J 2003 Transporter gene families in plants: The sulphate transporter gene family – redundancy or specialization? Physiol. Plant. 117, 155–165. Hawkesford M J and Wray J L 2000 Molecular genetics of sulphate assimilation. Adv. Bot. Res. 33, 159–223. Huang Y, Jordan W R, Wing R A and Morgan P W 1998 Gene expression induced by physical impedance in maize roots. Plant Mol. Biol. 37, 921–930. Iijima M and Kono Y 1991 Interspecific differences of the root system structures of four cereal species as affected by soil compaction. Jpn. J. Crop Sci. 60, 130–138. Kundu D K, Ladha J K and Lapitan de Guzman E 1996 Tillage depth influence on soil nitrogen distribution and availability in a rice lowland. Soil Sci. Soc. Am. J. 60, 1153–1159. Masle J and Passioura J B 1987 The effect of soil strength on the growth of young wheat plants. Aust. J. Plant Physiol. 14, 643–656. Materechera S A, Dexter A R and Alston A M 1991 Penetration of very strong soils by seedling roots of different plant species. Plant Soil 135, 31–41. Mudge S R, Rae A L, Diatloff E and Smith F W 2002 Expression analysis suggests novel roles for members of the Pht1 family of phosphate transporters in Arabidopsis. Plant J. 31, 341–353. Payne R W 2000 A Guide to GenStat Release 8. Part 2 – Statistics, VSN International, Oxford, 997 pp. Price A H, Steele K A, Moore B J, Barraclough P B and Clark L J 2000 A combined RFLP and AFLP linkage map of upland rice (Oryza sativa L.) used to identify QTLs for rootpenetration ability. Theor. Appl. Genet. 100, 49–56. Sambrook J, Fritsch E F and Maniatis T 1989 Molecular Cloning: A Laboratory Manual. 2nd edition Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Samson B K, Hasan M and Wade L J 2002 Penetration of hardpans by rice lines in the rainfed lowlands. Field Crops Res. 76, 175–188. Smith F W, Hawkesford M J, Ealing P M, Clarkson D T, Vanden Berg P J, Belcher A R and Warrilow A G S 1997 Regulation of expression of a cDNA from barley roots encoding a high affinity sulphate transporter. Plant J. 12, 875–884.

76 Ternesi M, Andrade A P, Jorrin J and Benlloch M 1994 Root– shoot signalling in sunflower plants with confined root systems. Plant Soil 166, 31–36. Verwoerd T C, Dekker B M M and Hoekema A 1989 A smallscale procedure for the rapid isolation of plant RNAs. Nucleic Acids Res. 17, 2362. Wade L J, Fukai S, Samson B K, Ali A and Mazid M A 1999 Rainfed lowland rice: Physical environment and cultivar requirements. Field Crops Res. 64, 3–12.

Yoshida S and Hasegawa S 1982 The rice root system: Its development and function. In Drought Resistance in Crops with Emphasis on Rice. pp. 97–114. IRRI, Los Ban˜os, Philippines. Yu L, Ray J D, O’Toole J C and Nguyen H T 1995 Use of wax– petrolatum layers for screening rice root penetration. Crop Sci. 35, 684–687. Section editor: R.W. Bell