Drought resistance and recovery in mature Bituminaria bituminosa var. albomarginata

Annals of Applied Biology ISSN 0003-4746 RESEARCH ARTICLE Drought resistance and recovery in mature Bituminaria bituminosa var. albomarginata K. Foster1,2,3 , H. Lambers1 , D. Real1,2,3 , P. Ramankutty1,2 , G.R. Cawthray1 & M.H. Ryan1,2 1 School of Plant Biology and Institute of Agriculture, The University of Western Australia, Perth, Australia 2 Future Farm Industries Cooperative Research Centre, The University of Western Australia, Perth, Australia 3 Department of Agriculture and Food Western Australia, South Perth, Australia Keywords Alfalfa; climate change; osmotic adjustment; paraheliotropism; plant adaptation; Psoralea bituminosa; rehydration; water-use efficiency. Correspondence K. Foster, School of Plant Biology and Institute of Agriculture, The University of Western Australia, 35 Stirling Highway, Crawley (Perth) WA 6009, Australia. Email: kevin.foster@agric.wa.gov.au Received: 30 March 2014; revised version accepted: 10 September 2014. doi:10.1111/aab.12171 Abstract Few studies have investigated the response of perennial legumes to drought stress (DS) and their ability, following rewatering, to regrow and restore photosynthetic activity. We examined these responses for two genotypes of drought-tolerant tedera (Bituminaria bituminosa var. albomarginata) and one genotype of lucerne (Medicago sativa). Plants were grown outdoors in 1-m deep PVC pots with a reconstructed field soil profile, regularly watered for 8 months (winter to mid-summer), and then moved to a glasshouse where either watering was maintained or drought was imposed for up to 47 days, before rewatering for 28 days. Drought stress greatly decreased shoot dry matter (DM) production in both species. Lucerne plants showed severe leaf desiccation after 21 days of withholding water. Relative leaf water content (RWC = 42%) and midday leaf water potential (LWP = −6.5 MPa) decreased in tedera in response to DS, whereas leaf angle (85∘ ) and lateral root DM both increased. Proline and pinitol accumulated in tedera leaves during DS, and their concentration declined after rewatering. Nine days after rewatering, previously drought-stressed tedera had similar RWC and LWP to well-watered control plants. In tedera and lucerne, 28 days after rewatering, photosynthesis and stomatal conductance were greater than in the well-watered controls. The lateral root DM for one tedera genotype decreased during the recovery phase but for lucerne, the lateral root DM did not change during either the drought or the recovery phases. Overall, the root systems in tedera showed greater plasticity in response to DS and rewatering than in lucerne. In conclusion, tedera and lucerne showed different physiological and morphological strategies to survive and recover from DS. Proline and soluble sugars may act as a carbon source for regrowth in tedera during recovery. In comparison with lucerne, tedera’s more rapid recovery after rewatering should contribute to a greater aboveground DM yield under alternating dry and wet periods. Tedera genotypes are highly heterogeneous and selecting genotypes with enhanced concentrations of pinitol and proline could be a valuable tool to improve plant performance during DS and recovery. Introduction There is a need in Australia and elsewhere to develop pasture plants and grazing systems to address the issue of climate change (Cullen et al., 2009). It is critical to understand the potential role perennial plants can play Ann Appl Biol (2014) © 2014 Association of Applied Biologists in a changing climate, and develop systems that can deliver sustainability and, importantly, profitability for the livestock industries (Nie, 2011). However, with the fluctuations in rainfall patterns predicted from climate change, there may also be an increase in the occurrence of episodic drought (IPCC, 2007) where plants will be 1 K. Foster et al. Drought resistance and recovery in tedera repeatedly exposed to drought in the field. While drought resistance and recuperative potential are known to vary within several perennial grass species (Su et al., 2008), the combined responses to water stress and rewatering among perennial legumes is relatively unknown (Filippou et al., 2011). Lucerne (Medicago sativa L.) is a widely grown perennial legume in southern Australia and is considered to have strong drought avoidance due to its deep rooting system (Li et al., 2010); it survives prolonged drought by limiting aboveground growth and often sheds its leaves (Loo et al., 2006). While lucerne does have the ability to respond to summer/autumn rainfall (Moore et al., 2006), the uncertainty of summer rainfall events makes it difficult to predict lucerne growth in any season and to match growth with the demand of livestock (Revell et al., 2012). However, there remains a strong demand for new summer-active perennials and the need to maintain nitrogen fixation has focused attention on the inclusion of perennial legumes rather than perennial grasses (Dear et al., 2003). Tedera (Bituminaria bituminosa (L.) C.H. Stirton var. albomarginata) is a promising fodder species and has recently been introduced into Australia for evaluation as a pasture legume species in Mediterranean climates (Real et al., 2009). Tedera is native to Lanzarote, Canary Islands, Spain, where it is found at low altitude and in areas with an annual rainfall of 300 mm, a Mediterranean rainfall pattern, a 5–6 month dry season (Real et al., 2009), and high amounts of sunshine and high temperatures which when combined with strong winds, result in very high evaporation rates (Díaz, 2004). Tedera is a drought-resistant perennial legume (Pang et al., 2011; Suriyagoda et al., 2013) and compared to lucerne, is physiologically better adapted to water deficit, retaining more of its leaves and having a higher leaf water-use efficiency in summer (Foster et al., 2013). However, the drought-resistance mechanisms enabling tedera to grow in the harsh habitat of the Canary Islands are not known, nor has the recuperative potential of this promising fodder species, once drought stress (DS) is relieved, been investigated. Species differences in drought resistance and the degree of recovery from drought have been associated with various physiological, morphological and biochemical factors including changes in root distribution, root diameter, osmotic adjustment (OA) and accumulation of sugars and organic solutes (DaCosta & Huang, 2006). However, evergreens like tedera that endure severe water stress over summer may also rely on physiological changes that are more permanent, which may prohibit rapid recovery from DS (Mittler et al., 2001). Drought stress can also result in permanent dysfunction of some 2 vessels via tyloses or resins that are released in response to DS (Chen et al., 2010). Rapid and complete photosynthetic recovery following rewatering is likely the key to prevent significant declines in crop yield following episodic drought (Chaves et al., 2009). There is anecdotal information from the field that the photosynthetic potential of tedera might be enhanced following rewatering after drought, and so help compensate for the loss of dry matter (DM) production due to drought. A better understanding of the mechanisms of both drought resistance and recovery following rewatering in tedera will allow for specific traits to be targeted in selection/breeding programmes prior to the release of the first tedera cultivar and enhance the potential for the adoption of this species into low-medium-rainfall pasture systems. We conducted a glasshouse experiment to test our hypotheses: (1) the accumulation of compatible sugars and proline will increase in tedera during DS and then decrease during recovery; (2) the area-based rate of photosynthesis of tedera is enhanced following rewatering after DS; (3) tedera will exhibit superior recuperative potential compared with lucerne after DS and rewatering; and (4) the roots of tedera will show greater plasticity in response to DS and rewatering than those of lucerne. Materials and methods This experiment comprised three drought treatments applied to mature plants of lucerne and two genotypes of tedera. The two tedera genotypes differed in growth habit (tedera 4 is an erect genotype and tedera 6 is a semi-prostrate genotype) and were the most drought resistant of seven genotypes in field experiments at Newdegate (Latitude: 33 06.56 S Longitude: 118 50.15 E), Western Australia (K. Foster, personal observation). The legume species used for comparison was lucerne (SARDI TEN) which, where suited, it is highly productive, relatively drought resistant and persistent, and it is the standard by which other perennial legumes are often compared (Dear et al., 2003; Pang et al., 2011; Suriyagoda et al., 2013). Soil was obtained from a field site at Newdegate in three layers (0–40, 40–80, 80–100 cm) at the beginning of autumn in April 2009 from a dry profile. This soil represents a major soil type of agricultural importance in south-western Australia (Moore et al., 1998). Soil was mixed completely within each layer, passed through a 5-mm sieve and stored dry. This site has a texture-contrast soil with coarse sand over an alkaline clay loam at 80–100 cm. Bulked analyses were performed on these samples of soil by CSBP FutureFarm analytical laboratories (Bibra Lake, Western Australia). In the 0–40 cm Ann Appl Biol (2014) © 2014 Association of Applied Biologists K. Foster et al. layer, pH (CaCl2 ) was 5.5 and the soil chemistry (mg kg−1 dry soil) was 27 for bicarbonate-extractable phosphorus (P; Colwell, 1963), 34 for mineral nitrogen (N), 47 for bicarbonate-extractable potassium (K) and 15.1 for sulphur (S). In the 40–80 cm soil layer, pH (CaCl2 ) was 6.2 and the soil contained 19 mg kg−1 bicarbonate-extractable P, 7 mineral N, 22 bicarbonate-extractable K and 1.7 S. In the 80–100 cm soil layer, pH (CaCl2 ) was 8.0 and the soil contained 5 mg kg−1 bicarbonate-extractable P, 9 mineral N, 85 bicarbonate-extractable K and 5.6 S. The three soil layers that were collected from the field site were later reconstructed in polyvinyl chloride (PVC) cylindrical pots (15 cm diameter, 100 cm high), which were closed at the bottom with a solid plastic cap with two holes to allow free drainage of water. Each pot was cut lengthwise up to 80 cm on both sides and taped with waterproof 50 mm wide tape. Filter paper (Whatman No.5) was placed over the holes in the bottom of the pot, and pots were filled with 29.1 kg of dry soil. The soil was packed in 20-cm layers to a bulk density similar to that at the field site (approximately 1.40 g cm−3 ). There was no significant difference among species in the pot weight before or after water was added (data not shown). Lucerne seeds were scarified with fine sandpaper (P240 grit) to overcome seed-coat impermeability. Seeds of tedera were scarified by cutting the outer seed coat using a surgical scalpel blade. On 28 May 2009, seeds were pre-germinated in 90-mm Petri dishes with wet filter paper (Whatman No.1). Seedlings were then transplanted into plastic pots (Track PK48 clear) with a commercial potting mix [Waldecks premium potting mix, pH (CaCl2 ) 6] and transferred to a controlled-temperature glasshouse at 20/12∘ C (day/night) for 4 weeks. Rhizobia were applied at 0.50 g per tray to the soil surface and then tap water was applied [lucerne, AL Group (strain RRI128) from Nodulaid at Becker Underwood; tedera, WSM4083 from the Centre for Rhizobium Studies, Murdoch University]. All pots were watered to pot capacity (approximately 11% v/v) and allowed to drain for 48 h. One seedling was transplanted into each pot on 29 June 2009 (early winter). Plants then grew in the open air and were watered each week if necessary. Plants were fertilised in October and November with the equivalent of 150 kg ha−1 superphosphate and potash at a 3:1 ratio. Plants were not cut. Hence, plants were grown under close to the usual environmental conditions for an establishing perennial legume pasture for 8 months from winter to mid-summer. All pots were moved on 12 January 2010 (midsummer) into a naturally lit glasshouse (to avoid summer rainfall events) at the Department of Agriculture and Food Western Australia (DAFWA) with temperature set at 30∘ C/15∘ C (day/night), which is the average January temperature for Newdegate, in the southern wheatbelt Ann Appl Biol (2014) © 2014 Association of Applied Biologists Drought resistance and recovery in tedera of Western Australia (B.O.M., 2008). Plants were allowed to acclimate in the glasshouse for 14 days before imposing the treatment, and during this time they were rewatered to pot capacity on 12, 19 and 26 January. Both tedera genotypes and lucerne plants were flowering and setting seeds while plants were outside. After the last watering, drought was imposed in the DS treatments, that is, no more watering occurred, while the well-watered (WW) control continued to be watered. The imposition of drought in mid-summer mimics what may be expected to occur in the field. The experiment was a randomised complete block design of genotype (tedera 4, tedera 6, lucerne) by drought (DS, WW) by harvest (Treatments 1, 2 and 3). Plants in treatment 1 were harvested after 33 days of drought (Fig. 1, Harvest 1). Plants in treatment 2 were harvested after 47 days of drought (Harvest 2). Days 0–47 are therefore referred to as the ‘drought phase’. Plants in treatment 3 were rewatered back to pot capacity after 47 days of drought, and rewatered every seven days for the following 28 days before being harvested at day 75 (Harvest 3). Days 48–75 are therefore referred to as the ‘recovery phase’. Well-watered control plants were watered from large plastic containers adjusted to glasshouse temperature (i.e. cold tap water was not added to plants) to pot capacity once a week and harvested at the same time as the plants in the DS treatment (i.e. days 33, 47 and 75). There were four replicates. Plants in treatment 1 were Figure 1 Experimental design (one of four replicates shown). Plants were either well watered throughout the 75 days of the experiment or experienced drought for up to 47 days, before being rewatered for 28 days. Plants in harvest 1 (H1) were harvested after 33 days of drought; plants in harvest 2 (H2) were harvested after 47 days of drought and plants in harvest 3 (H3) were harvested after 47 days of drought followed by a recovery phase of 28 days of weekly rewatering. 3 K. Foster et al. Drought resistance and recovery in tedera harvested on 27 February and those in treatment 2 on 12 March 2010. All plants in the DS treatment 3 and their WW controls were rewatered back to pot capacity on 12 March, and harvested on 12 April 2010. For leaf angle, water potential (pre-dawn and midday), leaf RWC, net photosynthesis, stomatal conductance, OA and compatible solutes and sugars, readings were taken at the start of the treatments (i.e. day 1 of the drought phase). Subsequent readings, unless stated otherwise, were taken at regular intervals for treatment 3 plants only. Measurements Plant water use To determine plant water use, all pots were weighed on day 1 (26 January), and every 7 days thereafter, with a specifically designed pot crane fitted with an electronic balance. Photosynthesis and instantaneous leaf water-use efficiency Net photosynthesis (A) and stomatal conductance (gs ) were measured on youngest fully expanded leaves between 09:00 h and 12:00 h using a LI-COR 6400 portable gas exchange system (LI-COR, Lincoln, NE, USA) with a red/blue LED light source, at a photosynthetically active radiation (PAR) of 1500 μmol quanta m−2 s−1 , a CO2 concentration of 380 μmol mol−1 and a leaf chamber air flow rate of 200 μmol s−1 , with block temperature set at 30∘ C. The humidity of the air coming into the leaf chamber was kept the same as that in the glasshouse. Plants were measured on days 1, 7, 16, 31 and 47 of the drought phase, and then on days 21 and 28 of the recovery phase. Leaflets of the trifoliates were separated and placed in the standard 6 cm2 cuvette; care was taken to keep the leaf angle close to the in situ orientation. Leaves that were folded or rolled due to DS were gently unfolded before measuring. After measurement, each leaf was imaged and the leaf area was determined; this was used to calculate A and gs . Intrinsic leaf water-use efficiency (WUEL ) was calculated as A/gs (Ahmadi and Siosemardeh, 2005). Some readings for stomatal conductance, and therefore for WUE, occurred on plants at permanent wilting point (PWP) and these readings were therefore zero. Relative leaf water content, pre-dawn and midday leaf water potential Relative leaf water content (RWC) of fully expanded leaves was calculated as RWC (%) = (FW – DW)/(SFW – DW) × 100, where FW is fresh weight, DW is dry weight and SFW is saturated fresh weight. Leaves were removed 4 (days 1, 7, 14, 30 and 47 of the drought phase and days 9, 21 and 28 of the recovery phase) between 12:00 h and 14:00 h and placed in zip-lock plastic bags and then on ice. Fresh weight was recorded, then the leaves were immersed in a 90-mm Petri dish filled with deionised water for 24 h at room temperature (20–25∘ C) before reweighing to attain saturated fresh weight (Turner, 1981). Turgid leaf weight was measured and leaves were oven-dried at 80∘ C for 48 h to determine dry weight. Pre-dawn leaf water potential (LWPP ) (03:00–05:00 h) and midday leaf water potential (LWPM ) (12:00–14:00 h) were measured in a pressure chamber (Scholander Model 1002, PMS Instruments, Corvallis, OR, USA) on petioles of young fully expanded leaves. Pre-dawn LWP was measured for all plants on day one and treatment 3 plants were measured on days 7, 16 and 23 of the drought phase. Subsequent LWP readings were taken at midday on days 30 and 47 of the drought phase, and days 9, 21 and 28 of the recovery phase. The pressure chamber used allowed measurements to −7 MPa. Photosynthetically active radiation Photosynthetically active radiation was measured at the level of the terminal trifoliate leaflet of the 2nd or 3rd uppermost leaves (two leaves per plant) from treatment 3 on 24 February at 12:00–15:00 h (leaf elevation angles were at maximum for tedera genotypes) with clear skies and 36∘ C outside using a LI-192 Underwater Quantum Sensor [Model UWQ 7534, (LI-COR) calibration date 25 March 2008] set for reading in air with a LI-1400 data logger unit (LI-COR). Care was taken to keep the sensor head parallel to the leaf in its normal orientation. Osmotic adjustment Leaf samples were removed at noon (days 1, 7, 31 and 47 of the drought phase, and days 21 and 28 of the recovery phase) and immediately placed in a water-tight vial, snap frozen in liquid N2 and stored in a −80∘ C freezer. Samples were later thawed, sap expressed using a leaf press, and sap osmolality measured using a freezing point osmometer, which was calibrated against 50 and 850 mOsm kg−1 standard solutions (Fiske Associates, Norwood, MA, USA). Osmotic potential (OP, MPa) of samples was then calculated from osmolality OP as 2.447 × osmolality/1000. Data on RWC were used to convert the OP at the given water content to that at 100% RWC. Osmotic adjustment was calculated as the difference in OP at full turgor (OP 100; i.e. 100% leaf RWC) between DS and WW plants, according to the method of Ludlow et al. (1983). Ann Appl Biol (2014) © 2014 Association of Applied Biologists K. Foster et al. Compatible solutes and soluble sugars Leaf samples within treatment 3 were removed on three occasions; pre-drought, day 47 of the drought phase and day 28 of the recovery phase. On day 47, lucerne leaves were desiccated and could not be sampled. Leaf samples were immediately placed in a water-tight vial, snap frozen in liquid N2 and stored in a freezer at −20∘ C. Samples were placed in a lyophiliser for 72 h (Labconco Freeze Dry System, Model LyphLock 12, Kansas City, MO, USA) directly from the −20∘ C freezer; samples did not thaw before lyophilisation. Leaves were ground (60 s) in a modified coffee grinder (Sunbeam Autogrinder Model EM0415, China) to a fine powder, collected into a 2 mL Eppendorf tube and stored at −20∘ C. Ethanol extraction of compatible solutes, soluble sugars and sugar alcohols from plant tissues Samples were brought to room temperature and 20 mg were placed into a 1.5 mL screw-top Eppendorf tube. Samples were extracted twice with 1 mL of 80% (v/v) ethanol (99.5% analytical grade) by incubating at 80∘ C in a water bath for 20 min. Tubes were centrifuged (Model: Biofuge 13, Heraeus Instruments, Thermo Electron Corporation, Langenselbold, Germany) for 20 min at 13 793 g. The supernatants were combined in a 2 mL vial and the extract stored at −20∘ C. The pellet was dried at 60∘ C for 24 h and the dry weight measured to calculate the ethanol-insoluble weight. The initial HPLC analysis of compatible solutes (glycinebetaine, proline, proline betaine and hydroxyproline), soluble sugars (fructose, glucose and sucrose) and sugar alcohols (pinitol, sorbitol and mannitol) was adapted from Slimestad & Vågen (2006). The HPLC system (Waters, Milford, MA, USA) consisted of a 600E pump, 717 plus autosampler and a 996 photo-diode array detector (PDA). As detection of fructose, glucose and sucrose with the PDA at 195 nm is very insensitive, an Alltech (Deerfield, IL, USA) evaporative light-scattering detector (ELSD) was used to improve sensitivity by a minimum of 100-fold. Separation was achieved at 22 ± 1.0∘ C on a Prevail ES Carbohydrate column (250 × 4.6 mm i.d. with 5 μm packing; Alltech) using a gradient elution profile of acetonitrile (Eluent A) and water (Eluent B) at 1 mL min−1 . Samples in the autoinjector were held at 10∘ C, the ELSD drift tube was held at 85∘ C and eluent nebulisation with high purity nitrogen gas was at a flow rate of 2.6 L min−1 . Quantification based on peak area for compatible solutes used the PDA, while peak area from the ELSD was used for soluble sugars and sugar alcohols. Calibration curves were generated from peak area versus the mass of standard analytes injected, with a linear Ann Appl Biol (2014) © 2014 Association of Applied Biologists Drought resistance and recovery in tedera relationship for the PDA output and a power relationship for the ELSD output. A standard was analysed every 10 samples to check for instrument/detector drift. Retention times of standards were used to identify analytes in the sample extracts, with the PDA spectral data and peak purity used to confirm compatible solutes. Typical sample injections were 20 μL and runtime was 20 min per sample, with EMPOWER™ 2 software (Waters) used for data acquisition and processing. For the tedera samples under the HPLC conditions used and described above, pinitol in the extracts co-eluted with fructose. For pinitol and fructose analyses, 16 samples were re-analysed for the genotype tedera 4 only (due to time constraints and cost), pre-drought, day 47 of the drought phase and day 28 of the recovery phase (four replicates each from the DS treatment). In addition, four replicates from the WW control on day 47 of the drought phase were also analysed for pinitol. To separate and quantify fructose and pinitol, a Sugar-Pak (Waters) column (300 × 6.5 mm i.d.) was held at 90 ± 0.5∘ C and separation achieved using a mobile phase of 2.5 mg L−1 Ca-EDTA at 0.6 mL min−1 (Naidu, 1998). Detection and quantification of pinitol were undertaken with the PDA as this offered good sensitivity, as well as peak spectral and purity comparisons with the standards. Not detected (ND) denotes failure to detect any quantity above the detection limit, where the detection limits were: proline (10 μmol g−1 DW), proline betaine (10 μmol g−1 DW), fructose (11 mg g−1 DW), glucose (16 mg g−1 DW), sucrose (3 mg g−1 DW) and pinitol (25 mg g−1 DW). Shoot and root growth characters At each of the three harvests (days 33 and 47 of the drought phase and day 28 of the recovery phase) plants were cut at the soil surface and shoots dried at 80∘ C for 48 h and weighed. Plant components were separated into leaf, stem and reproductive components (seed and flowers), and weighed. Leaf mass ratio (LMR) was calculated as the proportion of the total harvested biomass that was leaf biomass. The weight of desiccated leaves that had dropped from the plant was not quantified. At each of the three harvests, pots were cut open; roots were placed on a 2-mm sieve, soil was gently washed away from the roots, and then immediately frozen with dry ice. Roots of considerable length were found coiled in the bottom 10 cm of the pot, but these were not harvested separately. Nodules were visible on lateral roots and the taproots of all genotypes at all three harvests in the DS treatments and WW controls. Roots (crowns, taproot and laterals) were then oven-dried at 80∘ C for 72 h and DM recorded. The root to shoot ratio (r : s ratio) was calculated as total root DM/total shoot DM. 5 Drought resistance and recovery in tedera K. Foster et al. Individual leaf angle and plant height Date + Genotype. Rewatering. Date. Leaf water potential had the following fixed components: Constant + Genotype + Harvest + Genotype. Harvest + Harvest.Drought + Genotype.Harvest. Drought + Date + Genotype.Date + Date. Harvest + Genotype.Date. Harvest + Date. Harvest. Drought + Genotype. Date.Harvest.Drought. [Correction added on 6 November 2014, after first online publication: Several changes were made to the terms in the formulae in this paragraph.]. All models assume that the error terms are normally and independently distributed with zero mean and constant variance. Where necessary, the data were transformed before analysis using the natural logarithmic transformation to ensure that the model assumptions were met. The data were analysed by analysis of variance (ANOVA) or by repeated measures analysis of variance using GENSTAT RELEASE 12.1 [(PC/Windows XP) Copyright 2009, Lawes Agricultural Trust (Rothamsted Experimental Station)]. Various plots (scatter plots, histograms and normal probability plots) of the residuals were examined to ensure that the model assumptions were met. Standard Error of the Differences (SED) are given to allow for means comparisons (McNicol, 2013). Leaf angle inclination was measured from a horizontal line from the base of the primary pulvinus of the terminal leaflet (days 1, 7, 18, 25, 37 and 47 of the drought phase, and days 6, 21 and 28 of the recovery phase) using a transparent plastic protractor with a swing arm blade in the centre of the base (Fiskars, Madison, WI, USA). Leaf angles above horizontal (0∘ ) were represented as positive values; leaf angles below horizontal were assigned negative values. Plant height was measured from the soil surface of the pot at day 30 of drought and again 28 days after rewatering. Statistical analysis A preliminary analysis was done using the data at the start of the treatments (26 January 2010 – day 1) to check for the presence of initial variation among the treatments. If variation was present (i.e. for photosynthetic rates, relative leaf water content), these initial pre-drought readings were used as covariates. Scatter plots of the response variables against the covariates were examined to ensure that analysis of covariance assumptions were met. A linear mixed model was fitted to all response variates because both fixed and random effects were present. The random components were common to all models and were Block + Block.Genotype. Block was considered random because the environmental conditions (Blocks) in which the plants were grown were used to represent all possible field environments in which these plants can be grown. Similarly, the genotypes grown in these environments (Block.Genotype) were considered representative of the growth of the entire genotype in any environment, and not just the specific genotypes examined in this experiment. For the photosynthetic active radiation, the fixed components were: Constant + Genotype + Drought + Genotype. Drought. Responses for destructive harvest measurements had the following fixed components: Constant + Genotype + Harvest + Genotype. Harvest + Harvest.Drought + Genotype. Harvest.Drought. Plant height and OA were analysed using repeated measures analysis of variance. Responses for plant height and other non-destructive measures (excluding OA and LWP) had the following fixed components: Constant + Genotype + Drought + Genotype.Drought + Rewatering + Genotype.Rewatering + Drought.Rewatering + Rewatering.Date + Genotype.Drought.Rewatering + Genotype.Rewatering.Date + Drought.Rewatering.Date + Genotype.Drought.Rewatering.Date. Osmotic adjustment had the following fixed components: Constant + Genotype + Rewatering + Genotype. Rewatering + Rewatering. 6 Results Soil water On day 1, soil water content was around 11% (v/v) in all pots (data not shown). Soil water content then continuously declined in all DS pots. By day 16 of the drought phase, soil water content was reduced to 4% (v/v) for all DS pots. By day 21, lucerne was approaching PWP; thereafter, aboveground productivity stopped, with plants neither growing nor using water. However, tedera plants did not reach PWP and continued to function during this period. The soil water content slowly declined for tedera genotypes, resulting in the lowest soil water content [approximately 2% (v/v)] at day 47 of DS. Note that during the drought and recovery phases, soil water content in all WW control pots was returned to pot capacity every 7 days [approximately 11% (v/v)]. Photosynthesis and stomatal conductance For all genotypes, photosynthetic rates (A) in the DS treatment and WW controls at day 7 were very similar at approximately 13–15 μmol CO2 m−2 s−1 (Fig. 2). At day 16, A in the DS treatment was reduced for all species and by day 31 tedera had reached a low 4–5 μmol CO2 m−2 s−1 . At day 47, A for tedera in the DS treatment decreased to 5% of the WW controls (approximately 0.5 μmol CO2 m−2 s−1 ). At days 31 and 47, lucerne plants in the DS treatment were at PWP and hence A and gs were zero. Ann Appl Biol (2014) © 2014 Association of Applied Biologists K. Foster et al. Drought resistance and recovery in tedera had dropped sharply in the DS treatment by day 31, to approximately 15% of the WW controls. Both tedera genotypes in the DS treatment maintained this level of low gs from day 31 to 47 of the drought phase. The gs of all genotypes increased markedly in the recovery phase and was higher than that of the WW controls. For tedera in the DS treatment at day 21 of the recovery phase, gs was more than double that in the WW control and the value for tedera 4 was higher than that of lucerne. However, by day 28 of recovery, the gs of all genotypes was similar. Intrinsic leaf water-use efficiency In the drought phase, the WUEL for all genotypes in the DS treatment and WW control was similar from day 7 to 16 (Fig. 3b). The WUEL of tedera genotypes in the DS treatment increased sharply at day 31 compared with that in the WW control plants, with WUEL of tedera 4 higher (50%) than that of tedera 6. At the same time, the values of gs decreased rapidly in tedera. However, between day 31 and day 47, WUEL for both tedera genotypes decreased, particularly for tedera 4, which had a lower WUEL than its WW control. At day 28, the WUEL of all the genotypes was similar in the DS treatment and WW control. The WUEL for lucerne in the DS treatment after rewatering was similar to that of the WW controls, although values for the latter decreased at day 21 of the recovery phase. The WUEL of the WW controls varied slightly over time. Figure 2 Photosynthesis for two genotypes of tedera and one genotype of lucerne (n = 4, SED = 0.9455). There was a three-way interaction among treatment, rewatering and day (P < 0.001), but no significant effect of genotype. Lucerne plants at days 31 and 47 of the drought phase were at permanent wilting point and hence photosynthesis was zero. In the WW control, A increased in tedera from day 7 and reached a peak at day 31 (16–18 μmol CO2 m−2 s−1 ) and declined by day 47. On day 1 of the recovery phase, all plants in the DS treatment and WW controls were rewatered back to pot capacity. At day 21 and 28 of the recovery phase, A in the DS treatment was higher for all genotypes, up to 50% for tedera 4, than that in the WW controls. By day 47, A in the WW controls had declined in tedera from day 31 by approximately 50%, whereas for lucerne, A declined during the recovery phase by approximately 30%. The gs of lucerne and tedera 6 in the WW control and DS treatment at day seven of the drought phase were similar; however, gs of tedera 4 was higher in the DS treatment (Fig. 3a). By day 16 of the drought phase, gs for all genotypes in the DS treatment had generally decreased compared to their WW controls. The gs of tedera Ann Appl Biol (2014) © 2014 Association of Applied Biologists Relative leaf water content Relative leaf water content remained around 80% in the WW controls throughout the drought and recovery phases (Fig. 4). In the DS treatment, RWC declined for tedera 6 to a low of 42% by day 47 of the drought phase in green leaf tissues. By day 9 after rewatering, previously drought-stressed tedera had a similar RWC as the WW controls, whereas for lucerne, RWC was still lower than that in the WW controls. Pre-dawn and midday leaf water potential The effect of the DS treatment on LWP (pre-dawn and midday) differed between lucerne and tedera (data not shown). At day 7 of DS, there was no difference in LWPp among the treatments or genotypes. A decrease in the soil water content [4% (v/v)] in all pots in the DS treatment at day 16 resulted in a large decrease in the LWPP in lucerne to −4.2 MPa. At day 23 for lucerne, the LWPP was −4.5 MPa, and plants were close to PWP, whereas for tedera, LWPp was less negative at approximately −2 MPa. At day 30 of the drought phase, LWPM for tedera in the DS 7 K. Foster et al. Drought resistance and recovery in tedera Figure 3 (A) Stomatal conductance (n = 4; SED = 0.0390) and (B) intrinsic leaf water-use efficiency (n = 4, SED = 7.46) for two genotypes of tedera and one genotype of lucerne. For (A) and (B) there was a four-way interaction among genotype, treatment, rewatering and day (P < 0.001). Lucerne plants at days 31 and 47 of the drought phase were at permanent wilting point and hence gs and WUEL were zero. treatment was −4 MPa, whereas xylem sap could not be obtained from lucerne plants at −7 MPa. At day 47 of the drought phase, the LWPM for tedera in the DS treatment had declined to below −6.5 MPa. In the WW control, the LWPM for tedera ranged from −1.2 to −1.7 MPa, while in lucerne it declined to −2.5 MPa during the drought phase. The severity of the DS, measured in terms of LWP (both pre-dawn and midday), did not differ between the two tedera genotypes. The LWPM in the DS treatment was three to four times more negative than that of the WW controls; however, neither tedera genotype had reached PWP by day 47 of DS. By day 9 of the recovery phase, LWPM for tedera in the DS treatment had returned to the values of the WW controls. However, for lucerne, LWPM in the DS treatment 8 was higher (−1.6 MPa), that is, less negative, than that in the WW controls (−2 MPa). However, by day 28, LWPM in the DS treatment and WW control for lucerne did not differ. Other plant growth responses Leaf angle and photosynthetic active radiation At day 7 of the drought phase, there was no difference in leaf angle (≤5∘ ) among the genotypes or between the DS treatment and WW control (Fig. 5). In the DS treatment, by day 18 the leaf angles of all genotypes increased greatly and were 80∘ –85∘ . At day 25, the majority of the lucerne leaves were drooping towards the ground (i.e. −80∘ , below horizontal). The leaf angles remained at 70∘ –85∘ Ann Appl Biol (2014) © 2014 Association of Applied Biologists K. Foster et al. Figure 4 Relative leaf water content for two genotypes of tedera and one genotype of lucerne (n = 4, SED = 0.033). There was a three-way interaction among treatment, rewatering and day (P < 0.001), but no significant effect of genotype. for tedera 6 during the entire drought phase, whereas for tedera 4 the leaves declined to 45∘ –50∘ by day 47. By day 6 of the recovery phase the leaf angle for tedera in the previously drought-stressed plants was approximately 5∘ , less than in the WW controls, all of which had increased their leaf angles just prior to the next rewatering. By day 21 of the recovery phase, the leaf angles of all genotypes in both treatments did not differ (≤5∘ ). For tedera 4, the increase in leaf angle in the DS treatment reduced the upper leaf area exposed to vertical radiation when compared to the WW controls. At day 30, this reduced the average PAR on upper leaf surfaces of tedera 4 to approximately 850 ± 55 μmol photons m−2 s−1 (mean ± SE) compared to approximately 1860 ± 36 μmol photons m−2 s−1 in the WW control (Table S1, Supporting Ann Appl Biol (2014) © 2014 Association of Applied Biologists Drought resistance and recovery in tedera Figure 5 Mean leaf angle for two genotypes of tedera and one genotype of lucerne (n = 4, SED = 4.19). There was a four-way interaction (P < 0.001) of genotype, treatment, rewatering and day. Note that 0∘ is horizontal. Information), a reduction of approximately 55% compared to the WW control plants. Plant height and pubescence There was a two-way interaction of time by genotype (P < 0.05) for plant height (Table S2). The plant height at day 30 of drought phase differed among the genotypes in the DS treatment with lucerne plants taller (560 mm) than tedera 4 and 6 (266 and 208 mm, respectively). By day 21 after rewatering, the height of the lucerne plants in the DS treatment had increased (702 mm), but there was no change for tedera. Increased pubescence was observed on stems and new leaves of both tedera genotypes at day 30 of the drought phase in the DS treatment when compared to the WW control. 9 K. Foster et al. Drought resistance and recovery in tedera Table 1 Shoot DM and root DM for two genotypes of tedera and one genotype of lucernea Drought phase Recovery phase Day 33 Total shoot DM (g) Tedera 4 Tedera 6 Lucerne SED Total root DM (g) Tedera 4 Tedera 6 Lucerne SED Day 47 Day 28 WW DS WW DS WW DS 26.6 20.5 20.3 2.7 9.2 11.2 12.2 30.4 27.9 22.1 9.2 10.3 13.2 46.4 37.8 32.8 17.2 15.8 20.1 19.3 18.8 15.3 1.3 11.6 11.7 13.1 18.1 21.0 21.2 14.7 13.1 12.9 23.9 24.8 28.4 12.7 11.6 15.3 DM = dry matter; DS = drought-stressed; WW = well-watered controls. was a three-way interaction of genotype, harvest and treatment for shoot DM (P < 0.001) and root DM (P < 0.05). a There Plant growth Total shoot dry matter and leaf to total mass ratio The shoot DM of all genotypes in the DS treatment during the drought phase was greatly reduced compared to that of the WW control (Table 1). For tedera genotypes, this reduction in shoot DM was between 63% and 70% at day 47. For lucerne, this reduction was only 40%; while all leaves were desiccated, they had not been shed, as is often seen in the field. In the recovery phase in the DS treatment, shoot DM for tedera 4 increased by 86%, whereas that of both tedera 6 and lucerne increased by approximately 50%. However, the total shoot DM for all genotypes was still less (38–60%) than that of their equivalent WW controls. By day 28 of the recovery phase the canopy structure in tedera was similar to that of the WW controls, whereas the canopy of lucerne plants in the DS treatment contained both dried and green shoots. Consequently, the leaf to total plant mass ratio (LMR) for tedera genotypes at the end of the recovery phase in the DS treatment was higher (0.48–0.52) than for lucerne plants (0.32). Root dry matter production The root DM for all the genotypes in the DS treatment was reduced compared to that of the WW control plants (Table 1) at all three harvests. At harvest one, the reduction in root DM was 37–40% for tedera and 14% for lucerne. For tedera genotypes, the root DM in the WW control plants did not change during the drought phase. The total root DM increased at harvest two in the DS treatment for tedera 4 only, however the lateral roots continued to grow for both tedera genotypes (data not shown). At harvest three in the rewatered DS treatment, the total root DM for tedera and lucerne did not change during 10 the recovery phase although the lateral roots for tedera decreased but this was significant for tedera 4 only. For lucerne, the lateral root DM did not change during either the drought or the recovery phases. Root to shoot ratio The root to shoot ratios (r : s ratio) for tedera 4 and lucerne in the DS treatment at day 33 (Harvest 1) of the drought phase were higher than those of their WW controls (Fig. 6). However, at day 47 of the drought phase, the r : s ratio was higher for tedera only in the DS treatment, while there was no change for lucerne. At day 28 of the recovery phase, the r:s ratio of tedera in the rewatered DS treatment was reduced, and all genotypes were now similar to their respective WW controls. The r : s ratio changed little for the genotypes in WW controls over the three harvests. Osmotic adjustment At day 7 of the drought phase, there was no evidence of OA in either lucerne or tedera in the DS treatment (data not shown). However, by day 30 the OA was −0.12 MPa and −0.20 MPa for tedera 4 and tedera 6, respectively. By day 47, OA increased to −0.5 MPa for both tedera genotypes. Osmotic adjustment for lucerne at day 30 and day 47 was not measured, as plants were at PWP. In the DS treatment at the end of the recovery phase, OA for tedera 4 and 6 had declined to −0.13 MPa and −0.12 MPa, respectively, and for lucerne OA was −0.04 MPa. Compatible solutes and sugars Proline was not detected in either species or either treatment at day 1 (Table S1), but accumulated to high Ann Appl Biol (2014) © 2014 Association of Applied Biologists K. Foster et al. Drought resistance and recovery in tedera Figure 7 Pinitol and fructose concentrations in leaf tissue of tedera 4 at days 1 and 47 of the drought phase and day 28 of the recovery phase (mean ± SE, n = 4). Figure 6 Root : shoot ratio for two tedera genotypes and one genotype of lucerne at day 33 and day 47 of the drought phase and day 28 of the recovery phase (n = 4, SED = 0.12). There was a three-way interaction of genotype, treatment and harvest (P < 0.001). concentrations (191–205 μmol g−1 DW) in tedera in the DS treatment at the end of the drought phase (day 47). At the end of the recovery phase in the DS treatment (day 28), proline was again not detected in tedera or lucerne. Proline betaine was present in the leaves of lucerne in both treatments at day 1 (28–35 μmol g−1 DW); it is the main betaine in lucerne (Trinchant et al., 2004). However, at the end of the drought phase in the DS treatment, lucerne was at PWP and leaves could not be measured. In the recovery phase, the proline betaine concentration for lucerne in the DS treatment was again similar to that of the WW controls (58–67 μmol g−1 DW), but still higher than pre-drought levels. Concentrations of glucose (36–40 mg g−1 DW) and fructose (80–95 mg g−1 DW) which co-eluted with pinitol for tedera at day 47 in the DS treatment were more Ann Appl Biol (2014) © 2014 Association of Applied Biologists than three times greater than at day 1 of the drought phase. At the end of the recovery phase (day 28), fructose and glucose concentrations for tedera 4 and tedera 6 in the DS treatment had declined (48–57 mg g−1 DW and 16–25 mg g−1 DW, respectively) and were similar to those in the WW control. The fructose concentrations in lucerne in the DS and WW plants were similar at the end of the recovery phase (45–48 mg g−1 DW), but glucose concentrations were lower than pre-drought concentrations. Sucrose was either not detected or at low concentrations, with little change observed over time. Pinitol concentrations in the DS treatment at the end of the drought phase for tedera 4 (Fig. 7) were 68% higher than those in the WW plants (3-fold higher than pre-drought values). In the recovery phase, tedera 4 maintained enhanced pinitol concentrations, above pre-drought concentrations, but the concentrations were similar to those in the WW plants at day 47 of the drought phase. Discussion Key findings from this study are: (a) concentrations of proline and pinitol in tedera substantially increased in response to water deficits; (b) gs and A for both species in the DS treatment at days 21 and 28 of the recovery phase exceeded those of the WW control plants; (c) tedera 4 showed more root plasticity in response to DS and rewatering than did tedera 6 or lucerne; (d) OA plays an important role in drought tolerance in tedera; and (e) tedera had a more rapid recovery after rewatering than lucerne. Overall, tedera and lucerne showed different physiological and morphological strategies to survive and recover from DS. 11 K. Foster et al. Drought resistance and recovery in tedera Physiological and morphological responses to drought stress In tedera, the decreasing RWC was also associated with a decrease in A and gs , approximately in parallel. However, the decrease in A in both species is unlikely to arise from stomatal closure alone. Stomatal and non-stomatal limitations can both contribute to the reduction of photosynthesis under severe drought conditions (Flexas et al., 2004). The photosynthetic apparatus in tedera, although severely curtailed by low gs , appears highly resistant to DS, in contrast to that of lucerne (where leaves were brown). Even under extreme DS at day 47, the leaves of tedera were still pliable and green, and gas exchange continued, albeit at a low rate and very low LWPM . In tedera, reduced stomatal conductance was one of the primary mechanisms for acclimating to water stress. Stomatal closure occurred when significant reductions were observed in RWC and LWPM in the DS treatment, and it is likely the main cause of reduced photosynthesis in tedera under DS. In contrast, lucerne leaves did not show a reduced gs , by day 31 of DS; instead leaves wilted and died which reduced transpiration. Tedera’s WUEL increased as water supply declined in the DS treatment, but only until day 31, before declining at day 47 when water deficit was severe. Elevated WUE values are commonly observed in water-stressed plants (Pou et al., 2008), and WUE is an important index of a plant’s acclimation to an arid environment (Xing and Wu, 2012). However, severe water stress may have resulted in a decrease in the activities of photosynthetic enzymes, resulting in a lower WUEL (Zhao et al., 2004); a decrease in WUE has also been reported in both winter wheat (Shangguan et al., 2000) and spring wheat (El Hafid et al., 1998) subjected to severe stress. Plant growth responses to drought stress Total shoot and root dry matter production The long-term DS drastically reduced aboveground plant production in both tedera and lucerne. The reduction in leaf shoot DM under DS for tedera (63–69%) was similar to that reported for other legumes (Pandey et al., 1984). Nonetheless, any green leaf retention under DS is a highly desirable trait in a perennial pasture species over summer, as leaf biomass is likely to make up most of the nutritional intake for livestock. In contrast to tedera, lucerne plants sacrificed green leaves and became dormant under declining moisture conditions. Overall, long-term severe DS dramatically reduced total root DM in both species. However, for tedera 4, total root DM increased between Harvest 1 and 2; even though taproot DM did not change, the lateral roots continued to 12 grow. Lateral root growth in wheat (Triticum aestivum L.) is also promoted by water deficit (Ito et al., 2006). Likewise, Pang et al. (2011) reported an increase in the proportion of roots in the topsoil for tedera in response to DS; the mechanisms underlying the increase may include OA (Saab et al., 1992). For tedera, enhanced allocation of assimilates to the lateral roots likely provides a store of assimilates, available after rewatering. Consistent with this interpretation for tedera 4, the lateral root DM decreased after rewatering although total root DM did not change, whereas lateral root DM for lucerne did not change during the drought or recovery phase. Other aboveground plant growth responses Morphological responses to DS in tedera were striking. The extent of the leaf movements increased as the RWC and water potential decreased, thereby reducing light interception. Therefore, steep leaf angles at midday, with the consequent reduced radiation interception, would constitute a crucial adaptation in tedera to surviving in arid environments due to reduced leaf temperature and transpiration. Droughted plants were also increasingly pubescent on stems and new leaves, which would partly protect them from excessive light injury and decrease transpirational water loss (Pfeiffer et al., 2003). Root to shoot ratio The increase in r : s ratio under DS reflects the adaptive growth balance of the leaf canopy and root system. Many crop plants divert assimilates to root growth under water stress, resulting in a high r:s ratio (Whitfield et al., 1986), which is one of the mechanisms involved in the acclimation of plants to drought (Turner, 1997). Indeed, the increase in r : s ratio in tedera 4 at day 47 of DS was due to root biomass accumulating and not just a decrease in aboveground biomass. This suggests a stronger acclimation in tedera 4 to prolonged DS than in tedera 6 or lucerne. This should be advantageous because it enhances soil resource acquisition and suggests a close relationship with drought resistance in this species. Response of compatible solutes and sugars to drought stress Our observations are consistent with our first hypothesis that DS induces the accumulation of compatible osmolytes in leaf tissue in tedera, at least at the reproductive stage. The accumulation of substantial concentrations of proline in tedera at the same time as significant changes occurred in RWC and LWPM suggests it has a role in OA in leaf tissue during the later phase of water stress, as has been reported for cowpea [Vigna unguiculata (L.) Ann Appl Biol (2014) © 2014 Association of Applied Biologists K. Foster et al. Walp.] (Oliveira Neto et al., 2009). Osmotic adjustment has been found in tedera genotypes previously (Pang et al., 2011); our results also suggest it plays an important role in drought tolerance, at least in the leaves. In tedera, OA may also allow stomata to remain at least partially open, and CO2 assimilation to continue, at LWPs that could otherwise be inhibitory and thereby enable plants to extract more water from the soil and adjust their water potential. High concentrations of proline in tedera during drought may also have protected the photosynthetic system from permanent damage (Lawlor, 2001). Osmotic adjustment in tedera likely helps assists to maintain leaf metabolism and root growth at very low LWP by maintaining the turgor pressure in the cells. This is the first report of pinitol in tedera in response to water deficit and it suggests this sugar alcohol contributes to drought tolerance in this species, at least in the specific genotype tested. Consistent with our results, Ford (1984) concluded that pinitol accumulation may indicate the ability of a legume to tolerate low LWP. With its slow turnover relative to sugars (Paul & Cockburn, 1989), and limited reactivity, pinitol is a good candidate for an osmoprotectant. Pinitol concentrations also increased in the control plants at day 47. Likewise, Streeter et al. (2001) found pinitol accumulated in the leaves of WW soybean [Glycine max (L). Merr.] plants. These authors suggested there is some advantage for plants that are more likely to experience DS to accumulate pinitol before the onset of stress and that pinitol has a key role in stress tolerance of soybean. Silvente et al. (2012) also showed that a drought-tolerant soybean genotype had higher amounts of pinitol, even under WW conditions, than a drought-sensitive genotype. Drought resistance and recovery in tedera Feller (2007) also reported a completely restored photosynthetic rate 4 weeks after rewatering, but in contrast to our results, gs remained at a lower level. The resumption of CO2 assimilation of water-stressed tedera plants after rewatering indicates that the basic mechanisms of photosynthetic photochemistry and biochemistry were not permanently damaged by water deficit (Cornic, 2000). Our second hypothesis, that the area-based photosynthetic rate of tedera is enhanced during the rewatering phase after drought was also supported. After rewatering, A and gs were greater in the DS plants than in the WW controls, indicating an overcompensation of gas exchange. However, the decline in aboveground DM in tedera due to long-term severe DS was not completely replaced by enhanced growth following rewatering. Similar results were found for soybean (Bunce, 1977; Wang et al., 2006; Lobato et al., 2008) and for some grasses (Xu et al., 2010). On rewatering, tedera was able to rapidly restore LWP and unfold green leaves, and this likely played a critical role in the resumption of plant growth, as previously found in some grasses (Munne-Bosch & Alegre, 2004). Although the recovery phase was limited to 28 days, during this period, tedera did exhibit superior recuperative potential to lucerne, confirming our third hypothesis. In contrast to tedera, lucerne plants in the DS treatment had brown desiccated leaves and dry stems upon rewatering. For lucerne to recover from DS, new shoots must be initiated following rewatering. Consequently, lucerne had fewer green leaves and a lower LMR, and its aboveground growth rate was slower than that of tedera over subsequent weeks. The new shoot growth resulted in an increase in plant height after rewatering in lucerne; however, there was no change for tedera. Physiological traits associated with post-drought recovery Root growth responses associated with post-drought recovery Tedera and lucerne quickly returned to pre-drought levels of physiological activity once DS had been relieved. For tedera, rewatering following the DS treatment resulted in a rapid reduction in the leaf folding angle to almost horizontal by day 6. By day 9 after rewatering, both leaf RWC and LWPM in the DS treatment and WW control plants were similar, indicating the quick reversibility of the responses to the DS in tedera. The reason for the full recovery of leaf water relations in tedera indicates that, after rewatering, embolised xylem vessels were likely rapidly refilled (Holbrook et al., 2001). Despite more than 16 days of almost full stomatal closure in both tedera genotypes in the DS treatment (days 31–47), severe DS did not appear to impair vascular capacity for water transport, as there was complete recovery of photosynthesis and gs after rewatering. Gallé & Ann Appl Biol (2014) © 2014 Association of Applied Biologists Root DM of tedera and lucerne did not change following rewatering of the DS plants, however, lateral root DM of tedera 4 decreased and perhaps thereby supplied energy and nutrients for the resumption of new shoot growth in tedera. The increase in root DM for tedera and lucerne in the WW controls during the recovery phase may be a response to the reduction in photoperiod in early autumn (April), with more assimilate investment towards perennial organs like taproots for survival. The root system of tedera showed greater phenotypic plasticity in response to DS and rewatering than that of lucerne which supports our fourth hypothesis. High levels of root plasticity to DS and rewatering in tedera may be a key factor in determining plant persistence and productivity under climate change. 13 K. Foster et al. Drought resistance and recovery in tedera Changes in compatible solutes and sugars associated with post-drought recovery Rewatering of the DS plants resulted in proline no longer being detected in leaves of tedera at day 28. Proline has been reported to act as a source of energy, carbon and nitrogen (Van Heerden & Krüger, 2002; Szabados & Savoure, 2010). This suggests an important role for proline in plant recovery from severe DS in tedera, and is consistent with our first hypothesis. Our results also suggest an important role for pinitol in plant recovery from long-term DS in tedera 4. The pinitol concentration decreased substantially in leaves of tedera 4 following rewatering, but was still higher than pre-drought values, reflecting a slower turnover that likely contributed to metabolism for growth of new shoots. Soluble sugars may also act as a carbon source for regrowth in tedera during recovery, and glucose and fructose in tedera decreased by up to 55% after rewatering. Osmotic adjustment in tedera also declined at the end of the recovery phase and osmotically adjusted leaves likely metabolised much of these solutes. Osmotic adjustment is considered one of the critical processes in plant acclimation to drought, because it sustains tissue metabolic activity and also enables regrowth upon rewatering (Morgan, 1984). Concluding comments The maintenance of plant functions at low LWP in tedera and their more rapid recovery after rewatering should contribute to greater aboveground DM yield under intermittent drought periods than possible for lucerne recovering from dormancy. The ability of tedera to rapidly regrow following a period of DS depends on the presence of the residual photosynthetic leaf area. Selecting tedera genotypes that maximise the number of leaves that survive the dry season will likely further enhance this recovery and persistence. Following rewatering, tedera can rapidly adjust from steep leaf angles to the horizontal, to again obtain maximum photosynthesis. This would also be a key factor in facilitating a rapid growth response to rewatering in tedera. This study contributes to our understanding of stress biology of plants through the identification of key compounds in drought resistance and recovery in perennial legumes. The capacity to synthesise and accumulate pinitol could also be an important adaptive feature in other tedera genotypes. Selecting tedera genotypes with enhanced concentrations of pinitol during DS could be a valuable tool to improve plant performance under, as well as recovery from, DS. The identification of perennial plants with resistance to DS and, especially, the physiological properties utilised 14 by perennial legumes to cope with DS and maximise response to rewatering is of vital importance in the face of predicted climate change. Tedera has the potential to provide out-of-season forage across the low-to-medium rainfall zones in southern Mediterranean Australia and, critically, to quickly respond to rewatering events, to help balance the feed supply across the entire season. However, appropriate grazing-management strategies for tedera over autumn and summer are yet to be defined. Acknowledgements This research was supported by Department of Agriculture and Food Western Australia (DAFWA) and the Future Farm Industries Cooperative Research Centre (FFI CRC). We thank Michael Renton for his input into the experimental design and statistical approach and also Daniel Kidd, Dion Nicol, Neil Turner, Kevin Goss, Clinton Revell and Mark Sweetingham for their ongoing support and advice. References Ahmadi A., Siosemardeh A. (2005) Investigation on the physiological basis of grain yield and drought resistance in wheat: leaf photosynthetic rate, stomatal conductance and non stomatal limitation. International Journal of Agriculture and Biology, 7, 807–811. B.O.M. (2008) (Bureau of Meteorology) Climate statistics for Australian locations. Summary Statistics Newdegate. URL http://www.bom.gov.au/climate/averages/tables/cw_0106 92.shtml [accessed on 11 November 2011] Bunce J.A. (1977) Leaf elongation in relation to leaf water potential in soybean. Journal of Experimental Botany, 28, 156–161. Chaves M.M., Flexas J., Pinheiro C. (2009) Photosynthesis under drought and salt stress: regulation mechanisms from whole plant to cell. Annals of Botany, 103, 551–560. Chen J., Zhang Q., Li X., Cao K. (2010) Gas exchange and hydraulics in seedlings of Hevea brasiliensis during water stress and recovery. Tree Physiology, 30, 876–885. Colwell J.D. (1963) The estimation of the phosphorus fertilizer requirements of wheat in southern New South Wales by soil analysis. Australian Journal of Experimental Agriculture and Animal Husbandry, 3, 190–198. Cornic G. (2000) Drought stress inhibits photosynthesis by decreasing stomatal aperture–not by affecting ATP synthesis. Trends in Plant Science, 5, 187–188. Cullen B.R., Johnson I.R., Eckard R.J., Lodge G.M., Walker R.G., Rawnsley R.P., McCaskill M.R. (2009) Climate change effects on pasture systems in south-eastern Australia. Crop & Pasture Science, 60, 933–942. DaCosta M., Huang B. (2006) Deficit irrigation effects on water use characteristics of bentgrass species. Crop Science, 46, 1779–1786. Ann Appl Biol (2014) © 2014 Association of Applied Biologists K. Foster et al. Dear B.S., Moore G.A., Hughes S.J. (2003) Adaptation and potential contribution of temperate perennial legumes to the southern Australian wheatbelt: a review. Australian Journal of Experimental Agriculture, 43, 1–18. Díaz F. (2004) Sistemas agrícolas tradicionales de las zonas áridas de las Islas Canarias. Tesis Doctoral. Universidad de La Laguna. Flexas J., Bota J., Loreto F., Cornic G., Sharkey T.D. (2004) Diffusive and metabolic limitations to photosynthesis under drought and salinity in C3 plants. Plant Biology, 6, 269–279. Ford C.W. (1984) Accumulation of low molecular weight solutes in water-stressed tropical legumes. Phytochemistry, 23, 1007–1015. Filippou P., Antoniou C., Fotopoulos V. (2011) Effect of drought and rewatering on the cellular status and antioxidant response of Medicago truncatula plants. Plant Signaling & Behaviour, 6, 270–277. Foster K., Ryan M.H., Real D., Ramankutty P., Lambers H. (2013) Seasonal and diurnal variation in the stomatal conductance and paraheliotropism of tedera (Bituminaria bituminosa var. albomarginata) in the field. Functional Plant Biology, 40, 719–729. Gallé A., Feller U. (2007) Changes of photosynthetic traits in beech saplings (Fagus sylvatica) under severe drought stress and during recovery. Physiologia Plantarum, 131, 412–421. El Hafid R., Smith D.H., Karrou M., Samir K. (1998) Physiological responses of spring durum wheat cultivars to early-season drought in a Mediterranean environment. Annals of Botany, 81, 363–370. Holbrook N.M., Ahrens E.T., Burns M.J., Zwieniecki M.A. (2001) In vivo observation of cavitation and embolism repair using magnetic resonance imaging. Plant Physiology, 126, 27–31. IPCC. (2007) Climate Change 2007: impacts, adaptation and vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Eds M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C.E. Hanson. Cambridge, UK: Cambridge University Press. Ito K., Tanakamaru K., Morita S., Abe J., Inanaga S. (2006) Lateral root development, including responses to soil drying, of maize (Zea mays) and wheat (Triticum aestivum) seminal roots. Physiologia Plantarum, 127, 260–267. Lawlor D.W. (2001) Photosynthesis. 3rd edition. Oxford: Bios Scientific Publishers. Li W.R., Zhang S.Q., Ding S.Y., Shan L. (2010) Root morphological variation and water use in alfalfa under drought stress. Acta Ecologica Sinica, 30, 5140–5150. Lobato A.K.S., Oliveira Neto C.F., Santos Filho B.G., Cruz F.J.R., Neves H.K.B., Lopes M.J.S. (2008) Physiological and biochemical behaviour in soybean (Glycine max cv. Sambaiba) plants under water deficit. Australian Journal of Crop Science, 2, 25–32. Ann Appl Biol (2014) © 2014 Association of Applied Biologists Drought resistance and recovery in tedera Loo C., Dolling P.J., Mokhtari S. (2006) Lucerne. In Perennial pastures for Western Australia, Bulletin 4690, pp. 59–75. Eds G. Moore, P. Sanford and T. Wiley. Western Australia, Australia: Department of Agriculture and Food. Ludlow M.M., Chu A.C.P., Clements R.J., Kerslake R.G. (1983) Adaptation of species of Centrosema to water stress. Australian Journal of Plant Physiology, 10, 119–130. McNicol J. (2013) Significance testing after analysis of variance. Annals of Applied Biology, 162, 288–289. Mittler R., Merquiol E., Hallak-Herr E., Rachmilevitch S., Kaplan A., Cohen M. (2001) Living under a “dormant” canopy: a molecular acclimation mechanism of the desert plant Retama raetam. Plant Journal, 25, 407–416. Moore G., Dolling P., Porter B., Leonard L. (1998) Soil acidity. In Soil Guide. A Handbook for Understanding and Managing Agricultural Soils, pp. 127–140. Ed. G. Moore. Western Australia, Australia: Department of Food and Agriculture. Moore G., Sanford P., Wiley T. (2006) Perennial pastures for Western Australia. Bulletin 4690. Perth, Australia: Department of Food and Agriculture. Morgan J.M. (1984) Osmoregulation and water stress in higher plants. Annual Review of Plant Physiology, 35, 299–319. Munne-Bosch S., Alegre L. (2004) Die and let live: leaf senescence contributes to plant survival under drought stress. Functional Plant Biology, 31, 203–216. Naidu B.P. (1998) Separation of sugars, polyols, proline analogues, and betaines in stressed plant extracts by high performance liquid chromatography and quantification by ultra violet detection. Australian Journal of Plant Physiology, 25, 793–800. Nie Z. (2011) Use of perennial grass in grazing systems of Southern Australia to adapt to a changing climate. URL http://www. intechopen.com/books/climate-change-research-and-tech nology-for-adaptation-and-mitigation/use-of-perennialgrass-in-grazing-systems-of-southern-australia-to-adaptto-a-changing-climate. Oliveira Neto C.F., Lobato A.K.S., Costa R.C.L., Maia W.J.M.S., Santos Filho B.G., Alves G.A.R., Brinez B., Neves H.K.B., Santos Lopes M.J., Cruz F.J.R. (2009) Nitrogen compounds and enzyme activities in sorghum induced to water deficit during three stages. Plant, Soil and Environment, 55, 238–244. Pandey R.K., Herrera W.A.T., Villepas A.N., Pendelton J.W. (1984) Drought response of grain legumes under irrigation gradient. III. Plant Growth. Agronomy Journal, 76, 557–560. Pang J., Yang J., Ward P., Siddique K.H.M., Lambers H., Tibbett M., Ryan M. (2011) Contrasting responses to drought stress in herbaceous perennial legumes. Plant and Soil, 348, 299–314. Paul M.J., Cockburn W. (1989) Pinitol, a compatible solute in Mesembryanthemum crystallinum L. Journal of Experimental Botany, 40, 1093–1098. 15 K. Foster et al. Drought resistance and recovery in tedera Pfeiffer T.W., Peyyala R., Ren Q., Ghabrial A.S. (2003) Increased soybean pubescence density: yield and soybean mosaic virus resistance effects. Crop Science, 43, 2071–2076. Pou A., Flexas J., Alsina M.M., Bota J., Carambula C., de Herralde F., Galmes J., Lovisolo C., Jimenez M., Ribas-Carbo M., Rusjan D., Secchi F., Tomas M., Zsofi Z., Medrano H. (2008) Adjustments of water use efficiency by stomatal regulation during drought and recovery in the drought-adapted Vitis hybrid Richter-110 (V. berlandieri × V. rupestris). Physiologia Plantarum, 134, 313–323. Real D., Correal E., Méndez P., Santos A., Rios S., Sternberg Z.M., Dini-Papanastasi O., Pecetti L., Tava A. (2009) Grassland species profiles. Bituminaria bituminosa C.H. Stirton (synonym: Psoralea bituminosa L.). Rome, Italy: Food and Agriculture Organization. URL http://www.fao.org/ ag/AGP/agpc/doc/Gbase/new_species/tedera/bitbit.htm. Revell C.K., Ewing M.A., Nutt B.J. (2012) Breeding and farming system opportunities for pasture legumes facing increasing climate variability in the south-west of Western Australia. Crop & Pasture Science, 63, 840–847. Saab I.N., Sharp R.E., Pritchard J. (1992) Effect of inhibition of abscisic acid accumulation on the spatial distribution of elongation in the primary root and mesocotyl of maize at low water potentials. Plant Physiology, 99, 26–33. Shangguan Z.P., Shao M.A., Dyckmans J. (2000) Nitrogen nutrition and water stress effects on leaf photosynthetic gas exchange and water use efficiency in winter wheat. Environmental and Experimental Botany, 44, 141–149. Silvente S., Sobolev A.P., Lara M. (2012) Metabolite adjustments in drought tolerant and sensitive soybean genotypes in response to water stress. PLoS ONE, 7, 1–11. DOI:10.1371/journal.pone.0038554. Slimestad R., Vågen I.M. (2006) Thermal stability of glucose and other sugar aldoses in normal phase high performance liquid chromatography. Journal of Chromatography A, 1118, 281–284. Streeter J.G., Lohnes D.G., Fioritto R.J. (2001) Pattern of pinitol accumulation in soybean plants and relationships to drought tolerance. Plant, Cell and Environment, 24, 429–438. Su K., Bremer D.J., Keeley S.J., Fry J.D. (2008) Rooting characteristics and canopy responses to drought of turf-grasses including hybrid bluegrasses. Agronomy Journal, 100, 949–956. Suriyagoda L.D.B., Real D., Renton M., Lambers H., Ryan M.H. (2013) Establishment, survival, and herbage production of novel, summer-active perennial pasture legumes in the low-rainfall cropping zone of Western Australia as affected by plant density and cutting frequency. Crop & Pasture Science, 64, 71–85. Szabados L., Savoure A. (2010) Proline: a multifunctional amino acid. Trends in Plant Science, 15, 89–97. Trinchant J.C., Boscari A., Spennato G., Van de Sype G., Le Rudulier D. (2004) Proline betaine accumulation and metabolism in alfalfa plants under sodium chloride stress. 16 Exploring its compartmentalization in nodules. Plant Physiology, 135, 1583–1594. Turner N.C. (1981) Techniques and experimental approaches for the measurement of plant water status. Plant and Soil, 58, 339–366. Turner N.C. (1997) Further progress in crop water relations. Advances in Agronomy, 58, 293–338. Van Heerden P.D., Krüger G.H.J. (2002) Separately and simultaneously induced dark chilling and drought stress effects on photosynthesis, proline accumulation and antioxidant metabolism in soybean. Journal of Plant Physiology, 159, 1077–1086. Wang L., Zhang T., Ding S. (2006) Effect of drought and rewatering on photosynthetic physioecological characteristics of soybean. Acta Ecologica Sinica, 26, 2073–2078. Whitfield D.M., Wright G.C., Gyles O.A., Taylor A.J. (1986) Growth of lucerne (Medicago sativa L.) in response to frequency of irrigation and gypsum application on a heavy clay soil. Irrigation Science, 7, 37–52. Xing D., Wu Y. (2012) Photosynthetic response of three climber plant species to osmotic stress induced by polyethylene glycol (PEG) 6000. Acta Physiologiae Plantarum, 34, 1659–1668. Xu Z., Zhou G., Shimizu H. (2010) Plant responses to drought and rewatering. Plant Signaling & Behaviour, 5, 649–654. Zhao C.X., Deng X.P., Zhang S.Q., Ye Q., Steudle E., Shan L. (2004) Advances in the studies on water uptake by plant roots. Journal of Integrative Plant Biology, 46, 505–514. Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Photosynthetically active radiation (PAR) (μmol photons m−2 s−1 ) for two genotypes of tedera and one genotype of lucerne in the drought-stressed treatment and the well-watered control on day 30 in the drought phase at 12:00–15:00 h (average of two leaves per plant) (mean ± SE, n = 4). Table S2. Plant height (mm) for two genotypes of tedera and one genotype of lucerne in the drought-stressed treatment at day 30 in the drought phase and day 28 in the recovery phrase (n = 4, SED = 58). There was a two-way interaction of time by genotype (P < 0.05). Table S3. Proline, proline betaine, fructose/pinitol and glucose concentrations in leaf tissue for two genotypes of tedera and one genotype of lucerne at day 1 and day 47 of the drought phase, and day 28 of the recovery phase in the drought-stressed treatment (DS) and the well-watered control (WW) (mean ± SE, n = 4). PWP = permanent wilting point, * = not detected, that is failure to detect any quantity above the detection limit: proline (10 μmol g−1 DW), proline betaine (10 μmol g−1 DW), fructose (11 mg g−1 DW), glucose (16 mg g−1 DW) and pinitol (25 mg g−1 DW). Ann Appl Biol (2014) © 2014 Association of Applied Biologists