Oecologia DOI 10.1007/s00442-014-3033-4

Physiological ecology - Original research

Convergence of a specialized root trait in plants from nutrient‑impoverished soils: phosphorus‑acquisition strategy in a nonmycorrhizal cactus A. Abrahão · H. Lambers · A. C. H. F. Sawaya · P. Mazzafera · R. S. Oliveira 

Received: 8 March 2014 / Accepted: 30 July 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  In old, phosphorus (P)-impoverished habitats, root specializations such as cluster roots efficiently mobilize and acquire P by releasing large amounts of carboxylates in the rhizosphere. These specialized roots are rarely mycorrhizal. We investigated whether Discocactus placen‑ tiformis (Cactaceae), a common species in nutrient-poor campos rupestres over white sands, operates in the same way as other root specializations. Discocactus placenti‑ formis showed no mycorrhizal colonization, but exhibited a sand-binding root specialization with rhizosheath formation. We first provide circumstantial evidence for carboxylate exudation in field material, based on its very high shoot manganese (Mn) concentrations, and then firm evidence, based on exudate analysis. We identified predominantly oxalic acid, but also malic, citric, lactic, succinic, fumaric, and malonic acids. When grown in nutrient solution with P concentrations ranging from 0 to 100 μM, we observed an increase in total carboxylate exudation with decreasing P supply, showing that P deficiency stimulated carboxylate release. Additionally, we tested P solubilization by citric, malic and oxalic acids, and found that they solubilized P from the strongly P-sorbing soil in its native habitat, when the acids were added in combination and in relatively low

Communicated by Jason P. Kaye. A. Abrahão · A. C. H. F. Sawaya · P. Mazzafera · R. S. Oliveira (*)  Departamento de Biologia Vegetal, Institute of Biology, University of Campinas, UNICAMP, P.O. Box: 6109, Campinas, SP 13083‑970, Brazil e-mail: [email protected]; [email protected] H. Lambers · R. S. Oliveira  School of Plant Biology, University of Western Australia, Perth, Australia

concentrations. We conclude that the sand-binding root specialization in this nonmycorrhizal cactus functions similar to that of cluster roots, which efficiently enhance P acquisition in other habitats with very low P availability. Keywords Carboxylates · Cerrado · Manganese · Root exudates · Root hairs

Introduction Nutrient availability is one of the main environmental filters influencing plant species distribution (Lambers et al. 2008a). In old landscapes, plant productivity tends to be phosphorus (P)-limited, rather than nitrogen (N)-limited, because P is originally present in the soil parent material and then lost due to erosion, leaching and occlusion, while N is virtually absent from the parent material and gradually accumulates through biological fixation and atmospheric deposition (Walker and Syers 1976; Turner and Condron 2013). Hopper (2009) introduced the concept of OCBILs (Old, Climatically-Buffered, Infertile Landscapes), based on three floristic provinces: the fynbos vegetation in South Africa, the South West Australian Floristic Region (SWAFR), and the Venezuelan Pantepuis, but surmised that other parts of South America could be classified as OCBILs. OCBILs are among the most nutrient-impoverished environments in the world (Lambers et al. 2010). The plant species diversity in OCBILs reflects diversity of plant nutritional strategies (Lambers et al. 2010; Olde Venterink 2011). A number of plant responses and adaptations have been described for P-impoverished habitats (Lambers et al. 2008a, b, 2010), including root proliferation in shallow soil horizons (Lynch and Brown 2001) and increased abundance and length of root hairs (Bates and Lynch 2001).

13

Oecologia

The presence of abundant root hairs allows the formation of rhizosheaths in sand-binding roots. Rhizosheaths are sheaths of soil particles strongly adhered to the root surface formed through the exudation of mucilage (McCully 1999), with increasing evidence for a role in nutrient acquisition (Smith et al. 2011; Hayes et al. 2014). A relatively large fraction of plant species in severely P-impoverished environments are nonmycorrhizal and exhibit alternative root specializations (Miller 2005), such as cluster roots (Neumann and Martinoia 2002; Shane et al. 2005) and dauciform roots (Lamont 1974; Shane et al. 2005). These root specializations are recognized for their ability to acquire sparingly soluble forms of P (Skene 1998; Shane et al. 2005), mainly by releasing a large amount of root exudates, such as organic anions and phosphatases (Raghothama 1999) in a small volume of rhizosphere soil (Dinkelaker et al. 1995). These nonmycorrhizal specializations are a mining strategy to access poorly available nutrients, as opposed to the scavenging mycorrhizal strategy (Lambers et al. 2008a, b). Organic anions (or carboxylates) are root exudates involved in nutrient acquisition of the mining strategy (Hinsinger 2001; Lambers et al. 2006). Some species increase carboxylate exudation in response to P, K or Fe deficiency (Hoffland et al. 1992; Jones 1998; Playsted et al. 2006). Mobilization of P by organic anions can occur by direct ligand exchange, where the organic anion replaces P on ligand-exchange surfaces, or by ligand-promoted mineral dissolution, by ligand adsorption followed by subsequent detachment of the metal–ligand complex (Johnson and Loeppert 2006; Lambers et al. 2008a). The release of carboxylates occurs predominantly at the root tips. The release of carboxylates mobilizes micronutrients from the soil (Lambers et al. 2006). In cluster-rooted plants, as a result of rapid rates of carboxylate release, soil manganese (Mn) is mobilized (Grierson and Attiwill 1989; Dinkelaker et al. 1995) and accumulates in leaves (Gardner and Boundy 1983; Shane et al. 2005; Muler et al. 2014). This suggests that plant Mn concentrations may be used as a proxy for carboxylate release in P-poor soils (Hayes et al. 2014). In Brazil, extensive areas of severely nutrient-impoverished soils are found in the cerrado, a global biodiversity hotspot (Myers et al. 2000), particularly across the campos rupestres at the Espinhaço range. These areas are known for their plant species diversity (Alves and Kolbek 1994), but with the exception of two recent studies focused on carnivorous plants (Pereira et al. 2012; Nishi et al. 2013), remarkably little is known about the nutritional strategies of campos rupestres species. Here, we assess the mycorrhizal status, morphology, anatomy, and analyze the functioning of a sand-binding root specialization in Discocactus placentiformis (Lehmann) K. Schumann 1894 (Cactaceae), a common species in campos rupestres. We also investigate

13

the ability of organic acids to solubilize P from the soil in its natural habitat. Based on observations on severely P-impoverished habitats in Australia (Lambers et al. 2006) and South Africa (Lamont 1982), we hypothesized that D. placentiformis roots would be nonmycorrhizal, and have alternative root adaptations. These alternative strategies were expected to be similar in function to cluster roots and dauciform roots (Shane et al. 2006; Playsted et al. 2006), and the roots were hypothesized to exhibit increased carboxylate exudation under P deficiency. We expected that D. placentiformis presented high concentrations of shoot Mn, as a result of a P-mobilizing carboxylate-releasing strategy.

Materials and methods Study site and species selection Soils of campos rupestres are severely nutrient-impoverished (Benites et al. 2003; de Campos 2012; Cândido 2012) with high porosity and permeability. The soils in these environments are rather heterogeneous, ranging from thin soil layers between rock outcrops to deep sandy depositions in valleys and depressions. Soil resin P at the collection site is less than 1 mg kg−1 (CB-1 in De Campos 2012). Discocactus placentiformis (Fig. 1a) is a globose cactus species that occurs abundantly only on sandstone rocks, quartz sand, and gravel in these cerrado campos rupestres (Machado et al. 2013). Most subpopulations are small and isolated from each other in the states of Bahia and Minas Gerais of Brazil (Martin and James 2009), exploited for the production of cactus candy, when located close to towns and villages, and threatened by habitat loss due to Eucaplyptus plantations (Machado et al. 2013). We observed in the field that the roots of D. placentiformis strongly adhered sand and were morphologically similar to sand-binding roots described for Anarthriaceae and Cyperaceae (Smith et al. 2011; Shane et al. 2011). We collected 25 young plants and transplanted them into PVC tubes (2.4 l; 10 cm diameter and 30 cm high) filled with soil from their native habitat in a sandy shrubland in Parque Estadual da Serra do Cabral, in the municipality of Joaquim Felício, Minas Gerais, south-eastern Brazil (17°42′S, 44°11′W, 1030 m altitude) in January 2010 and 2011. We transported the plants to the greenhouse of the Laboratory of Plant Functional Ecology, University of Campinas, where they were allowed to acclimate for 6 months until the beginning of the experiment. Root morphology, anatomy and mycorrhizal colonization We collected the whole root system and fine, ramified roots (<5 mm) from six individuals during the growing season

Oecologia

Fig. 1  Discocactus placentiformis’ root specialization. a D. pla‑ centiformis in its native habitat. b General view of the root system. c Washed root system of D. placentiformis (p) with primary growth root structure and covered with thin root hairs (s) with secondary growth structure without root hairs. d Root hairs from the root system of D. placentiformis in structure. e Cross-section of the D. pla‑ centiformis root system with primary growth structure. Root hairs (h)

are extensions of epidermal cells, being characterized as unicellular trichomes. f Cross-section of the D. placentiformis root system with secondary growth structure. Note the absence of root hairs. g D. pla‑ centiformis roots in agar plates with bromocresol purple that switches from purple to yellow at pH < 5. Scale bars (b) 4 cm, (c) 0.5 cm, (d) 500 μm, (e) 100 μm, (f) 50 μm

(wet summer). The roots were preserved in 70 % ethanol (v/v) (Koske and Gemma 1989) and their percentage of arbuscular mycorrhizal colonization was determined at the Laboratory of Soil Microbiology, ”Luiz de Queiroz” College of Agriculture (ESALQ), Piracicaba-SP, following the protocols of Giovanetti and Mosse (1980). Roots with primary and secondary growth were sampled for anatomical procedures. Root fragments were cut and fixed in FAA 50 (Johansen 1940). Roots were dehydrated in a graded ethanol series and embedded in Historesin®. Serial sections with rotatory microtome were stained in toluidine blue, and observed under an Olympus BX51 microscope (New York, USA) equipped with an Olympus® DP71 camera (Tokyo, Japan). We measured the root hairs using the camera software. For morphological descriptions, the plants were removed from the soil and their roots gently washed to

remove adhering sand. Individual roots were photographed with a digital camera mounted on a stereoscopic microscope (Leica® MSV266) using Leica Application Suite (LAS) v.3.8.0 (Heerbrugg, Switzerland) for focus stacking. Root exudation Acid exudation To check if the roots of D. placentiformis were capable of releasing any acid exudates, prior to the hydroponic experiment, we removed four plants from the pots, gently washed their roots with distilled water and placed young root fragments onto agar plates with 1 % (w/v) bromocresol purple, a pH indicator (Neumann et al. 2000). This pH indicator switches from purple to yellow at pH values below 5. We

13

Oecologia

chose young, but fully developed root sections, as they are usually the most active in exudate release (Watt and Evans 1999). We also collected six additional individuals in the field for Mn analysis. Plants were dried until constant weight, ground, digested with nitric and perchloric acids and analyzed by atomic absorption spectrophotometry (Perkin-Elmer, MA, USA) at the ESALQ. Phosphorus‑concentration effects on carboxylate exudation To determine if D. placentiformis was capable of regulating P uptake depending on P availability, an experiment was carried out by growing plants in nutrient solutions at a range of P concentrations (0, 10, 50, and 100 μM P, n = 5, 10, 5, and 5, respectively) for 8 months. Each plant was cultivated in a plastic bag inside a PVC pot filled with 2 l of nutrient solution containing (in μM): 400 NO3−, 200 Ca2+, 200 K+, 154 SO42−, 54 Mg2+, 20 Cl−, 2,0 FeEDTA, 0.24 Mn2+, 0.10 Zn2+, 0.02 Cu2+, 2.4 H3BO3, and 0.3 Mo4+ prepared in deionized water (pH = 5.8). PVC pots were maintained inside a polystyrene box to provide thermal insulation. The polystyrene box was covered with a black polypropylene perforated lid to allow roots to enter the solution. Shoots were supported in the center of a foam disc that made a light-tight seal. Each plant was provided with an aeration tube with an aquarium stone at the end of the tube to facilitate air dissolution in the nutrient solution. Nutrient solutions were replaced three times per week. Plants in the greenhouse were subjected to natural daily and seasonal cycles of solar radiation, temperature was not controlled in the greenhouse and varied from 8 to 36 °C, and relative humidity varied from 10 to 100 %. Exudate collection and analyses After detecting acid exudates by the bromocresol purple assay, we wanted to identify the root organic acid exudates. Three 1–2 cm sections of fresh root tips per plant were collected and gently shaken in 0.3 or 0.6 ml of 0.1 % (v/v) Mili-Q water-solved formic acid (pH = 5.8) to avoid microbial consumption of the organic acids. We did not use calcium chloride because previous experiments showed that it interfered in ionization of organic acids on UPLCMS analyses. After 5 min, the root tips were removed and the samples were frozen at −20 °C until analysis. Each root tip was weighed in order to express acid concentrations per unit root weight. Working standards were citric, malic, oxalic, succinic, lactic, maleic, tartaric, DL-isocitric, malonic, and fumaric acid, which are known to be released from roots and may be involved in P mobilization (Roelofs et al. 2001). Samples were analyzed using an ultra-high performance liquid chromatography system with a triple quadrupole

13

mass spectrometer (UPLC-MS) and an electrospray ionization source, Acquity UPLC-TQD (Waters, Milford, MA, USA). The compounds were separated with a Waters Acquity BEH C18 column (2.1 × 50 mm) with 1.7 μm particle size at a temperature of 30 °C. The mobile phase consisted of Mili-Q water containing 0.1 % (v/v) of formic acid (phase A), and methanol (phase B) at a flow-rate of 0.2 ml min−1 and the injected volume was 10 μl. Elution was carried out using a gradient starting with 99 % A, maintained until 2.5 min, then ramping up to 50 % A by 3 min, maintaining this concentration for 4 min, then returning to the initial conditions and re-equilibrating the system until 6 min. All data were acquired and processed with MassLynxV4 software (Waters) MS detection in the negative ion mode and Selected Ion Mode was used for the specific detection of analytes. Source temperature was 150 °C, and desolvation temperature was 350 °C. Capillary voltage was −2.8 kV and cone −25 V. Phosphate‑sorption curves To measure native soil P-sorption capacity (S  =  S0  +  S′), we produced phosphate-sorption curves. Five soil replicates (0–30 cm) from campos rupestres of Serra do Cabral were collected, air-dried, 2 mm-sieved and stored at 4 °C until required. Initial P concentration (S0, in mg kg−1 soil) was measured using Olsen’s and Colwell’s method for extraction with a soil: solution ratio of 1:100 (Olsen et al. 1954; Colwell 1963) and introduced in the sorption equation as suggested by Barrow (2000). An aliquot of 6 ml of acidified soil extract was collected and the concentration of phosphate was measured with the malachite green method (Motomizu et al. 1983). Sorption curves were prepared by adding 5 g of soil mixed with 100 ml of background solution (0.02 M KCl) containing P added as KH2PO4 in concentrations of 0, 2, 4, 8, 20 and 25 mg P l−1. Samples were shaken at room temperature for 16 h at 100 rpm. The concentration of phosphate was measured with the malachite green method (Motomizu et al. 1983). Sorbed P (S′, in mg kg−1 soil) was calculated as the difference between P added (in mg l−1) and P observed in solution (C, in mg l−1). Both Freundlich (1) and Langmuir’s (2) equations were fitted to the adsorption curves as in Sharpley et al. (2008) and Graetz and Nair (2009), as follows:

S = aC b

(1)

1 C C = + S KSmax Smax

(2)

The coefficients of the Freundlich Eq. (1) were obtained by using the linear form (log S = log a + b log C), where S is P-sorption capacity (S  =  S0  +  S′), S0 is initial soil P,

Oecologia

Effect of organic acids on P sorption P solubilization by carboxylic acids by the three major carboxylic acids released by D. placentiformis was measured: citric (CIT), malic (MAL), oxalic (OXA) acids, and the three of them together (CMO). One gram of steamsterilized soil was placed in 15-ml centrifuge tubes and mixed with 5 mM KCl background solution with carboxylate acids added alone or in mixtures. The samples were shaken for 20 min at 260 rpm and centrifuged at 11,000g for 20 min. The supernatant was collected with a syringe and filtered through a 45-µm syringe-filter, and P was analyzed as described above. All treatments were carried out in three to five replicates. Statistical analyses Statistical analyses were performed using R (R Core Team 2012) and Statistica 8.0 (StatSoft 2007). Comparisons of acid concentrations within plants, regardless of the treatment, were made using Friedman ANOVA (nonparametric two-way ANOVA) (Kendall 1970). Comparisons of fractions of total exudation of each acid were performed with the Kruskal–Wallis test (nonparametric one-way ANOVA). Comparisons of total organic acid exudation between treatments were done by model selection with linear mixed models. The linear mixed model approach allowed including hierarchy in the factors, as we had several measures of root exudation per plant. The linear mixed-effect models were analyzed with the lme4 package (Bates et al. 2012). Model selection was performed by evaluating likelihood using a χ 2 distribution, and included two different models. In the first model, P treatment was used as fixed effect, and in the second model, no fixed effect was used. Random effects in both models included plants and repeated measures within plants. Values with Cook’s distance greater than one were considered outliers and removed from the analyses using the influence ME package to calculate Cook’s distance (Nieuwenhuis et al. 2012). Post-hoc comparisons were performed using Tukey contrasts with the lsmeans package (Lenth 2014). Fumaric and malonic acids were found in only a few samples, and therefore a linear regression was used, avoiding outlier exclusion to reduce the loss of degrees of freedom. The effect of organic acids on P solubilization was verified with simple linear regressions.

50

Acid exudation (µmol g−1 fwt min−1)

as measured by Colwell’s method, S′ is calculated sorbed P (both in mg P kg−1 soil), C is P concentration observed in solution (C, in mg l−1), and a and b are coefficients. In the Langmuir Eq. (2), Smax is maximum P sorption capacity, calculated as the reciprocal of the slope of the C/S vs. C plot; and K, is the binding energy, calculated as the slope/ intercept of the same plot.

Malonic Fumaric Succinic Lactic Citric Malic Oxalic

40

30

20

10

0

0

10

50

100

−1

P supply (µmol l )

Fig. 2  Median exudation of carboxylates from root tips of Discocac‑ tus placentiformis grown at a range of P concentrations in nutrient solution. Oxalic red, malic blue, citric green, lactic purple, succinic orange, fumaric yellow and malonic brown (n  = 5 plants per treatment, 3 roots tips per plant)

Results Morphology, anatomy and mycorrhizal colonization The root system of D. placentiformis consisted of a main branched tap root with several branched lateral roots (Fig.  1b). In the field, a small cactus (around 8 cm diameter) had roots up to 40–50 cm long. During primary root growth, these lateral roots were white and showed a high density of root hairs along their entire axis. During secondary growth, root hairs disappeared and roots turned to yellow or brown (Fig. 1c). Anatomical sections showed that root hairs from roots with 200 μm diameter were more than 1 mm long (Fig. 1d, e) and were lost during secondary growth (Fig. 1f). No mycorrhizal colonization was observed on any individual of this species. Root exudates Using bromocresol purple, we observed that D. placenti‑ formis roots does exude acids, as shown by the yellow aura around the roots (Fig. 1g). In UPLC-MS analyses, both the quantity and composition of carboxylates released from D. placentiformis roots varied according to P supply (Fig. 2). We identified predominantly oxalic, but also malic, citric, lactic, succinic, fumaric, and malonic acid in decreasing order of concentration (Friedman ANOVA χ 2 = 112.1429,

13

200 50

100

Sorbed P (mg kg −1 soil)

150

400

0

Total exudation (µmol g−1 fwt min−1)

Oecologia

0

20

40

60

80

300

200

100

100

−1

P supply (µmol l )

Phosphate‑sorption equations Campos rupestres soils presented native P (S0, Colwell P) of 6.5 ± 0.8 mg P kg−1 soil (mean ± standard deviation).

13

0.5

1.0

1.5

2.0 −1

Solution P concentration (mg l )

4.0

4.5

5.0

Fig. 4  Sorption behavior of phosphate (P) on native sandy soils from Serra do Cabral, Minas Gerais, Brazil, fitted with the Langmuir (solid line) and Freundlich (dashed line) equations (n = 5 soil replicates). P added ranged from 0 to 25 mg l−1

3.5

p < 0.0001). Maleic and isocitric acids were observed in trace amounts only and were not included in the analyses. When we analyzed total exudation with the linear mixed models, the model that included P treatment as a fixed factor fitted better (AIC = 802.33) than the model without the P treatment. (AIC = 804.58) ( χ 2  = 4.2512; p  = 0.0392; Fig.  3). The best model showed that total exudation (sum of all acids) declined significantly with increasing P supply (y = −0.29x + 42.97, Correlation of fixed effects = −0.72). Although total exudation seemed to peak in the first level of P addition, post hoc pairwise comparisons using Tukey contrasts did not show differences between each pair of treatments. A high variability in total carboxylate concentrations was observed in all treatments, but only one value was considered an outlier (Cook’s distance greater than 1) and removed from the analyses (Fig. 3). When we analyzed the exudation of each acid separately with linear mixed-model approach we observed that oxalic acid exudation decreased with increasing P concentration (y  = 30.837–0.185x, χ 2  = 4.461; p  = 0.035), but P treatment did not affect exudation of citric, succinic, lactic, malic, fumaric, or malonic acids. Shoot Mn concentrations in D. placentiformis were remarkably high, ranging from 451 to 1156 mg kg−1 dry weight (DW) (960  ± 358 mg kg−1 DW, mean ± standard deviation), whereas the plant community presented foliar Mn concentrations of 127 ± 90 mg kg−1 (DW) (Oliveira et al., unpublished data).

0.0

Solubilized phosphate (mg kg−1)

Fig. 3  Total exudation from Discocactus placentiformis root tips grown at a range of concentrations of phosphorus in nutrient solution (each circle represents a root tip, n = 5 plants per treatment, 3 roots tips per plant, χ 2 = 4.2512; p = 0.0392); outliers were excluded)

0

0

5

10

15

20 −1

Organic acid concentration (µmol l )

Fig. 5  Soil phosphorus (P) concentration in native sandy soils from Serra do Cabral, Minas Gerais, Brazil, as affected by the addition of organic acids (citric, malic and oxalic acids), n = 3–5 replicates

An average of 93 % of the added P was sorbed by the soil in all treatments, and a maximum of P sorption (Smax) of 762 mg P per kg−1 of soil was calculated by Langmuir’s model. Experimental data conformed well to both Freundlich’s (r2 = 0.957, p < 0.001; Fig. 4) and Langmuir’s equation (r2  = 0.9328, p < 0.001; Fig. 4), but Langmuir’s model fitted better (smaller residual sum of squares, RSS = 5.55 10−7) than Freundlich’s model (RSS = 0.11).

Oecologia

Phosphate solubilization by organic acids In the treatment with the three acids together (CMO), between 0 and 20 μM, we observed an increase in soluble P concentration from 3.4 to 4.2 mg P kg−1 of soil (Fig. 5; r2 = 0.29; p = 0.025). We found no effect when we added single carboxylic acids alone, and no further increase in P solubilization when mixtures of CMO were added in excess of 20 μM, as expected because the negative effect of pH outweighed the positive effect of carboxylic anions.

Discussion Similar to what is known for dominant plant species on severely P-impoverished soils in Australia and South Africa (Lamont 1982), D. placentiformis was found to be nonmycorrhizal. It presented high Mn in its shoots, and a specialized root morphology, which was associated with the release of carboxylates, especially when plants were subjected to P deficiency. Root morphology and anatomy Roots of D. placentiformis are structurally different from cluster roots (Lamont 1982), since they do not form abundant determinate rootlets (Shishkova et al. 2008). However, they are functionally similar to cluster roots (Shane et al. 2006), in terms of their abundance of fine root hairs and the release of organic anions that enhance P uptake from P-impoverished soils. The morphology of the root system of D. placentiformis is similar to that of early stages of sand-binding roots of Lyginia barbata (Anarthriaceae) (Lamont 1982; Shane et al. 2005). The main difference between both root systems is that, in later developmental stages, during secondary growth, D. placentiformis loses its root hairs, while root hairs of sand-binding roots of L. barbata become lignified and are persistent (Shane et al. 2011). Sand-binding activity in D. placentiformis is similar to that of grasses (McCully 1999) and other cacti (North and Nobel 1992) that form rhizosheaths. Rhizosheath formation in grasses and cacti is restricted to young roots and develops until the disintegration of the epidermis and the loss of root hairs. Root hairs are a common root structure, but rhizosheath formation by abundant long root hairs is restricted to a few plant groups (Smith et al. 2011). The layer of soil tightly bound to the roots may promote water acquisition (North and Nobel 1992) as well as nutrient acquisition even in locally dry conditions (Nambiar 1976). In P-poor environments, root-hair abundance and length provide a competitive advantage to wild-type plants of Arabidopsis thaliana when compared to a root-hairless mutant (Bates and Lynch 2001). Root hairs also increase

surface area and expand the volume of soil from which nutrients can be absorbed, increasing the P-depletion zone beyond that of roots without root hairs (Peterson and Farquhar 1996). The long root hairs (>0.5 cm) of D. placenti‑ formis confer intimate contact with soil particles promoted by the rhizosheath and an extra surface area for P acquisition, where P is a limiting resource. Root carboxylate exudation Discocactus placentiformis presented very high shoot Mn, providing an indication of rhizosphere carboxylate release (Jauregui and Reisenauer 1982; Gardner and Boundy 1983; Grierson and Attiwill 1989; Shane et al. 2005). The bromocresol purple experiment confirmed that the roots of D. placentiformis were capable of releasing acids in the rhizosphere, and the UPLC-MS analysis confirmed that they also released carboxylates. Overall carboxylate exudation was increased under P deficiency, as it is in Lupinus albus (Gardner et al. 1983; Keerthisinghe et al. 1998; Shane et al. 2003a, b) and species within Cyperaceae, Proteaceae, and Brassicaceae (Dinkelaker et al. 1995; Playsted et al. 2006; Chen et al. 2013). Roelofs et al. (2001) mentioned that exudation rates in Proteaceae species are faster than those previously presented. Here, we present exudation rates one order of magnitude greater than those reported by Roelofs et al. 2001. In the absence of root structures like cluster roots, which allow a local build-up of exudates, more carboxylates would need to be released to have the same effect. When comparing L. albus and L. angustifolius, for example, the latter does not form cluster roots, but releases more carboxylates than L. albus (Pearse et al. 2006). Similarly, to be effective, D. placentiformis might have to release more carboxylates than cluster-rooted species. Oxalate was the carboxylate released in greatest quantities by D. placentiformis, while dauciform roots of Cyperaceae and cluster roots of Proteaceae release mainly citrate under P deficiency (Shane et al. 2003a, b; Playsted et al. 2006), although oxalate is also released (Hoffland et al. 2006). The ability to mobilize P increases with the number of carboxyl groups. Citrate (a trivalent anion), malate, and oxalate (both divalent anions) have multiple carboxyl groups, and a great affinity for Fe3+ and Al3+, and replace phosphorus previously complexed by metals, increasing P availability (Oburger et al. 2009). Therefore, oxalate mobilizes P from soil particles less efficiently than citrate (Bolan et al. 1994). However, the release of oxalate by D. placentiformis might be more efficient, because synthesis of the two-carbon oxalate requires less carbon than that of the six-carbon citrate; moreover, it is not a metabolic intermediate like citrate and malate (Dong et al. 2004). For example, Oryza sativa and Glycine max also release mainly oxalate under P deficiency (Dong

13

Oecologia

et al. 2004; Hoffland et al. 2006), albeit in much smaller quantities. Finally, up-regulation of the rate of exudation at low P supply, as observed in D. placentiformis increases P mobilization, even if exuded in relatively low continuous amounts (Fox and Comerford 1992). The release of oxalate by root tips can also be related to stress caused by Al (Zheng et al. 1998). Oxalate chelates toxic metals, thus diminishing their inhibitory effects (Ma et al. 1998). Cerrado soils have high plantavailable Al concentrations (Viani et al. 2011, 2014), and this is also true for the soils inhabited by the present cactus (Cândido 2012). We suggest that oxalate, in addition to mobilizing P, is also important for Al detoxification in D. placentiformis. Phosphate‑sorption equations Campos rupestres soils presented low available P, between 5.7 and 7.5 mg P kg−1 soil (NaHCO3 extraction), or less than 1 mg P kg−1 soil (resin P) (de Campos 2012), similar to OCBILs (Witkowski and Mitchell 1987; Laliberté et al. 2012). We observed high sorption of P by soils, despite the sandy nature of campos rupestres soils (95 % of sand). The maximum P-sorption capacity is closely related to soil texture. Soils with higher proportions of silt or clay (lower proportions of sand) present higher availability of P-binding sites, increasing sorption capacity (Ullah et al. 1983), while soils with higher proportions of sand-sized particles are only poorly capable of retaining phosphorus. The P concentrations added ranged from 1 to 25 mg l−1, while the observed concentrations only reached 2 mg l−1. A surprisingly high maximum sorption capacity (Smax) of 762 mg P per kg−1 of soil was calculated using Langmuir’s model. Further mineralogical analyses are needed to assess the nature of these high P-sorbing particles in campos rupestes despite the high proportion of sand. Organic matter decreases P-sorption capacity, as it competes with P for binding sites (Afif et al. 1995). The low organic matter content (12 g kg−1 soil) of campos rupestres soils (de Campos 2012) additionally contributes to its high maximum sorption capacity. This indicates that in addition to the low P status of campos rupestres soils, most of the P is sorbed to the soils. This characteristic of the soil requires specialized root functioning for P acquisition.

decreasing P solubility (Palomo et al. 2006). The range of organic acid concentrations that solubilized P was the same range that D. placentiformis released. Our results indicate that the release of carboxylic acids in combination, and at low concentrations was capable of efficient solubilization of P from the campos rupestres soils. Concluding remarks We found that a nonmycorrhizal cactus functioned successfully on severely P-impoverished soils, based on its high relative abundance (Machado et al. 2013). It occurred on soils with relatively high P-buffering capacity in cerrado campos rupestres. The binding of sand grains by lateral roots indicates the release of root exudates, usually involved in mineral nutrition (Nagarajah et al. 1970). In our study, we observed that the exudation of carboxylates by the root tips of D. placentiformis decreased at high P supply, a similar response to that found for cluster roots and dauciform roots in other families. However, their sand-binding activity and rhizosheath formation suggests that other exudates, such as mucilages, may have been released as well. The intimate contact between roots and soil particles found in species forming cluster roots, dauciform roots or rhizosheaths, as in D. placentiformis is part of their P-mining strategy (Lambers et al. 2008a, b). We provide evidence that the occurrence of the mining strategy can be expanded to South American P-impoverished habitats such as campos rupestres, and support the inclusion of these habitats as OCBILs. Taken together, these results indicate a functional convergence of nutrient-acquisition strategies for plants inhabiting severely P-impoverished habitats. Acknowledgments  This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, grant CNPq 474670/2008-2), Fundação de Apoio à Pesquisa do Estado de São Paulo (FAPESP, grant 10/17204-0) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes). We thank H.G. Cândido, C. Pereira and A.V. Scatigna for the support in collecting the plants in the field. Additionally, we thank M.C. Campos, S.M. Carmello-Guerreiro, N. Urquiza, J. Tamashiro, J.C. Galvão, L. Pereira, D.C. da Silva and D.P. de Souza for the methodological support and F.S.C. Takahashi for the help with statistical analyses. We also thank P.B. Costa, S.M.S. Costa, L. Franci and A. Dias, for their assistance. Finally, we thank the anonymous reviewers for their useful reviews of this manuscript.

P solubilization by organic acids We observed P solubilization by organic acids when added in combination at low concentrations (up to 20 μM of citric, malic and oxalic acids). At higher concentrations, the solubilizing effect of the organic anions would have been offset by the negative effect of protons that were added with the carboxylic acids, which further acidified soil solution,

13

References Afif E, Barrón V, Torrent J (1995) Organic matter delays but does not prevent phosphate sorption by cerrado soils from Brazil. Soil Sci 159:207–211 Alves RJV, Kolbek J (1994) Plant species endemism in savanna vegetation on table mountains (Campo Rupestre) in Brazil. Vegetation 113:125–139. doi:10.1007/BF00044230

Oecologia Barrow NJ (2000) Towards a single-point method for measuring phosphate sorption by soils. Aust J Soil Res 38:1099–1113. doi:10.1071/SR99135 Bates TR, Lynch JP (2001) Root hairs confer a competitive advantage under low phosphorus availability. Plant Soil 236:243–250. doi:1 0.1023/A:1012791706800 Bates D, Maechler M, Bolker B (2012) lme4: linear mixed-effects models using S4 classes. R Package version 0.99875-6 Bolan NS, Naidu R, Mahimairaja S, Baskaran S (1994) Influence of low-molecular-weight organic acids on the solubilization of phosphates. Biol Fertil Soils 18:311–319. doi:10.1007/ BF00570634 Brazilian woodland savannah and seasonally dry forest species. Persp Plant Ecol Evol Syst 16(2): 64–74. doi:10.1016/j.ppees.2014.02.001 Cândido HG (2012) Estratégias de aquisição de nutrientes e estequiometria ecológica em comunidades de campos rupestres. Brasil. Master’s Dissertation, Universidade Estadual de Campinas, Campinas, Brazil Chen SL, Yang LT, Lin ZH, Tang N (2013) Roles of organic acid metabolism in plant tolerance to phosphorus-deficiency. In: Lüttge U, Beyschlag W, Rancis D, Cushman J (eds) Progress in Botany 74. Springer, New York, pp 213–237 Colwell JD (1963) The estimation of the phosphorus fertilizer requirements of wheat in southern New South Wales by soil analysis. Aust J Exp Agric Anim Husbandry 3:190–197. doi:10.1071/ EA9630190 Benites MV, Caiafa AN, de Mendonça ES, Schaefer CE, Ker JC (2003) Solos e vegetação nos complexos rupestres de altitude da Mantiqueira e do Espinhaço. Flor Amb 10:76–85 De Campos MCR (2012) Phosphorus-acquisition and phosphorusconservation mechanisms of plants native to south-western Australia or to Brazilian rupestrian fields. PhD thesis, University of Western Australia, Perth Dinkelaker B, Hengeler C, Marschner H (1995) Distribution and function of proteoid roots and other root clusters. Bot Acta 108:183–200 Dong D, Peng X, Yan X (2004) Organic acid exudation induced by phosphorus deficiency and/or aluminium toxicity in two contrasting soybean genotypes. Physiol Plant 122:190–199. doi:10.1111/j.1399-3054.2004.00373.x Fox TR, Comerford NB (1992) Influence of oxalate loading on phosphorus and aluminum solubility in spodosols. Soil Sci Soc Am J 56:290–294. doi:10.2136/sssaj1992.03615995005600010046x Gardner WK, Boundy KA (1983) The acquisition of phosphorus by Lupinus albus L. IV. The effect of interplanting wheat and white lupin on the growth and mineral composition of the two species. Plant Soil 70:391–402. doi:10.1007/BF02374894 Gardner WK, Barber DA, Parbery DG (1983) The acquisition of phosphorus by Lupinus albus L. III. The probable mechanism by which phosphorus movement in the soil/root interface is enhanced. Plant Soil 70:107–124. doi:10.1007/BF02374754 Giovanetti M, Mosse B (1980) An evaluation of techniques for measuring vesicular arbuscular mycorrhizal infection in roots. New Phytol 84:489–500. doi:10.1111/j.1469-8137.1980.tb04556.x Graetz DA, Nair VD (2009) Phosphorus sorption isotherm determination. In: Kovar JL, Pierzynski GM (eds) Methods of phosphorus analysis for soils, sediments, residuals, and waters. Virginia Tech University, Virginia, pp 33–37 Grierson PF, Attiwill PM (1989) Chemical characteristics of the proteoid root mat of Banksia integrifolia L. Aust J Bot 37:137–143. doi:10.1071/BT9890137 Hayes P, Turner BL, Lambers H, Laliberté E (2014) Foliar nutrient concentrations and resorption efficiency in plants of contrasting nutrient-acquisition strategies along a 2-million-year dune chronosequence. J Ecol 102:396–410 Hinsinger P (2001) Bioavailability of soil inorganic P in the rhizosphere as affected by root-induced chemical changes: a review. Plant Soil 237:173–195. doi:10.1023/A:1013351617532

Hoffland E, van den Boogaard R, Nelemans J, Findenegg G (1992) Biosynthesis and root exudation of citric and malic acids in phosphate-starved rape plants. New Phytol 122:675–680. doi:10.1111/j.1469-8137.1992.tb00096.x Hoffland E, Wei C, Wissuwa M (2006) Organic anion exudation by lowland rice (Oryza sativa L.) at zinc and phosphorus deficiency. Plant Soil 283:155–162. doi:10.1007/s11104-005-3937-1 Hopper SD (2009) OCBIL theory: towards an integrated understanding of the evolution, ecology and conservation of biodiversity on old, climatically buffered, infertile landscapes. Plant Soil 322:49– 86. doi:10.1007/s11104-009-0068-0 Jauregui MA, Reisenauer HM (1982) Dissolution of oxides of manganese and iron by root exudate components. Soil Sci Soc Am J 46:314–317. doi:10.2136/sssaj1982.0361599500460002 0020x Johansen DA (1940) Plant microtechnique. McGraw-Hill, New York Johnson SE, Loeppert RH (2006) Role of organic acids in phosphate mobilization from iron oxide. Soil Sci Soc Am J 70:222–234. doi :10.2136/sssaj2005.0012 Jones DL (1998) Organic acids in the rhizosphere – a critical review. Plant Soil 205:25–44. doi:10.1023/A:1004356007312 Keerthisinghe G, Hocking PJ, Ryan PR, Delhaize E (1998) Effect of phosphorus supply on the formation and function of proteoid roots of white lupin (Lupinus albus L.). Plant Cell Environ 21:467–478. doi:10.1046/j.1365-3040.1998.00300.x Kendall MG (1970) Rank correlation methods, 4th edn. Griffin, London Koske RE, Gemma JN (1989) A modified procedure for staining roots to detect VA mycorrhizas. Mycol Res 92:486–488. doi:10.1016/ S0953-7562(89)80195-9 Laliberté E, Turner BL, Costes T, Pearse SJ, Wyrwoll K, Zemunik G, Lambers H (2012) Experimental assessment of nutrient limitation along a 2-million-year dune chronosequence in the southwestern Australia biodiversity hotspot. J Ecol 100:631–642. doi:10.1111/j.1365-2745.2012.01962.x Lambers H, Shane MW, Cramer MD, Pearse SJ, Veneklaas EJ (2006) Root structure and functioning for efficient acquisition of phosphorus: matching morphological and physiological traits. Ann Bot 98:693–713. doi:10.1093/aob/mcl114 Lambers H, Chapin FS, Pons TL (2008a) Plant physiological ecology (2nd edn). Springer, New York Lambers H, Raven JA, Shaver GR, Smith SE (2008b) Plant nutrientacquisition strategies change with soil age. Trends Ecol Evol 23:95–103. doi:10.1016/j.tree.2007.10.008 Lambers H, Brundrett MC, Raven JA, Hopper SD (2010) Plant mineral nutrition in ancient landscapes: high plant species diversity on infertile soils is linked to functional diversity for nutritional strategies. Plant Soil 334:11–31. doi:10.1007/s11104-010-0444-9 Lamont B (1974) The biology of dauciform roots in the sedge Cyathochaete avenacea. New Phytol 73:985–996. doi:10.1111/j.1469-8137.1974.tb01327.x Lamont B (1982) Mechanisms for enhancing nutrient uptake in plants, with particular reference to mediterranean South Africa and Western Australia. Bot Rev 48:597–689. doi:10.1007 /BF02860714 Lenth RV (2014) lsmeans: least-Squares Means. R Package version 2:05 Lynch JP, Brown KM (2001) Topsoil foraging—an architectural adaptation of plants to low phosphorus availability. Plant Soil 225:225–237. doi:10.1023/A:1013324727040 Ma JF, Hiradate S, Matsumoto H (1998) High aluminum resistance in buckwheat. II Oxalic acid detoxifies aluminum internally. Plant Physiol 117:753–759. doi:10.1104/pp.117.3.753 Machado M, Braun P, Taylor NP, Zappi D (2013) Discocactus placen‑ tiformis. IUCN 2013 IUCN Red List Threatened Species Version 2013.2

13

Oecologia Martin LA, James G (2009) Unusual habitats, unusual plants. Cactus Succul J 81:106–112. doi:10.2985/015.081.0303 McCully ME (1999) Roots in soil: unearthing the complexities of roots and their rhizospheres. Annu Rev Plant Physiol Plant Mol Biol 50:695–718. doi:10.1146/annurev.arplant.50.1.695 Miller RM (2005) The nonmycorrhizal root: a strategy for survival in nutrient-impoverished soils. New Phytol 165:655–658. doi:10.1111/j.1469-8137.2005.01331.x Motomizu S, Wakimoto T, Tôei K (1983) Spectrophotometric determination of phosphate in river waters with molybdate and malachite green. Analyst 108:361–367. doi:10.1039/AN9830800361 Muler AL, Oliveira RS, Lambers H, Veneklaas EJ (2014) Does cluster-root activity benefit nutrient uptake and growth of co-existing species? Oecologia 174:23–31. doi:10.1007/s00442-013-2747-z Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J (2000) Biodiversity hotspots for conservation priorities. Nature 403:853–858. doi:10.1038/35002501 Nagarajah S, Posner AM, Quirk JP (1970) Competitive adsorption of phosphate with polygalacturonate and other organic anions on kaolinite and oxide surfaces. Nature 228:83–85. doi:10.1038/228083a0 Nambiar EKS (1976) Uptake of Zn65 from dry soil by plants. Plant Soil 44:267–271. doi:10.1007/BF00016978 Neumann G, Martinoia E (2002) Cluster roots: an underground adaptation for survival in extreme environments. Trends Plant Sci 7:162–167. doi:10.1016/S1360-1385(02)02241-0 Neumann G, Massonneau A, Langlade N, Dinkelaker B, Hengeler C, Römheld V, Martinoia E (2000) Physiological aspects of cluster root function and development in phosphorus-deficient white lupin (Lupinus albus L.). Ann Bot 85:909–919. doi:10.1006/a nbo.2000.1135 Nieuwenhuis R, te Grotenhuis M, Pelzer B (2012) Influence.ME: tools for detecting influential data in mixed effects models. R J 4:38–47 Nishi AH, Vasconcellos-Neto J, Romero GQ (2013) The role of multiple partners in a digestive mutualism with a protocarnivorous plant. Ann Bot 111:143–150. doi:10.1093/aob/mcs242 North GB, Nobel PS (1992) Drought-induced changes in hydraulic conductivity and structure in roots of Ferocactus acan‑ thodes and Opuntia ficus-indica. New Phytol 120:9–19. doi:10.1111/j.1469-8137.1992.tb01053.x Oburger E, Kirk GJD, Wenzel WW, Puschenreiter M, Jones DL (2009) Interactive effects of organic acids in the rhizosphere. Soil Biol Biochem 41:449–457. doi:10.1016/j.soilbio.2008.10.034 Olde Venterink H (2011) Does phosphorus limitation promote species-rich plant communities? Plant Soil 345:1–9. doi:10.1007/ s11104-011-0796-9 Olsen S, Cole CV, Watanabe FS, Dean LA (1954) Estimation of available phosphorus in soils by extraction with sodium bicarbonate. USDA Circular, issue 939, 1–19 Palomo L, Claassen N, Jones DL (2006) Differential mobilization of P in the maize rhizosphere by citric acid and potassium citrate. Soil Biol Biochem 38:683–692. doi:10.1016/j.soilbio.2005.06.019 Pearse SJ, Veneklaas EJ, Cawthray GR, Bolland MDA, Lambers H (2006) Carboxylate release of wheat, canola and 11 grain legume species as affected by phosphorus status. Plant Soil 288:127–139. doi:10.1007/s11104-006-9099-y Pereira CG, Almenara DP, Winter CE, Fritsch PW, Lambers H, Oliveira RS (2012) Underground leaves of Philcoxia trap and digest nematodes. Proc Natl Acad Sci USA 109:1154–1158. doi: 10.1073/pnas.1114199109 Peterson RL, Farquhar ML (1996) Root hairs: specialized tubular cells extending root surfaces. Bot Rev 62:1–40. doi:10.1007 /BF02868919 Playsted CWS, Johnston ME, Ramage CM, Edwards DG, Cawthray GR, Lambers H (2006) Functional significance of dauciform

13

roots: exudation of carboxylates and acid phosphatase under phosphorus deficiency in Caustis blakei (Cyperaceae). New Phytol 170:491–500. doi:10.1111/j.1469-8137.2006.01697.x R Core Team (2012) R: a language and environment for statistical computing Raghothama KG (1999) Phosphate acquisition. Annu Rev Plant Physiol Plant Mol Biol 50:665–693. doi:10.1146/annurev.arplant.50. 1.665 Roelofs RFR, Rengel Z, Cawthray GR, Dixon KW, Lambers H (2001) Exudation of carboxylates in Australian Proteaceae: chemical composition. Plant Cell Environ 24:891–903. doi:10.1046/j.1365-3040.2001.00741.x Shane MW, de Vos M, de Roock S, Cawthray GR, Lambers H (2003a) Effects of external phosphorus supply on internal phosphorus concentration and the initiation, growth and exudation of cluster roots in Hakea prostrata R. Br. Plant Soil 248:209–219. doi:10.1 023/A:1022320416038 Shane MW, de Vos M, de Roock S, Lambers H (2003b) Shoot P status regulates cluster-root growth and citrate exudation in Lupi‑ nus albus grown with a divided root system. Plant Cell Environ 26:265–273. doi:10.1046/j.1365-3040.2003.00957.x Shane MW, Dixon KW, Lambers H (2005) The occurrence of dauciform roots amongst western Australian reeds, rushes and sedges, and the impact of phosphorus supply on dauciform-root development in Schoenus unispiculatus (Cyperaceae). New Phytol 165:887–898. doi:10.1111/j.1469-8137.2005.01283.x Shane MW, Cawthray GR, Cramer MD, Kuo J, Lambers H (2006) Specialized “dauciform” roots of Cyperaceae are structurally distinct, but functionally analogous with “cluster” roots. Plant Cell Environ 29:1989–1999. doi:10.1111/j.1365-3040.2006.01574.x Shane MW, McCully ME, Canny MJ, Pate JS, Lambers H (2011) Development and persistence of sandsheaths of Lyginia barbata (Restionaceae): relation to root structural development and longevity. Ann Bot 108:1307–1322. doi:10.1093/aob/mcr244 Sharpley AN, Kleinman PJA, Weld JL (2008) Environmental soil phosphorus indices. In: Carter MR, Gregorich EG (eds) Soil sampling and methods of analysis. Taylor & Francis, Boca Raton, pp 141–159 Shishkova S, Rost TL, Dubrovsky JG (2008) Determinate root growth and meristem maintenance in angiosperms. Ann Bot 101:319– 340. doi:10.1093/aob/mcm251 Skene KR (1998) Cluster roots: some ecological considerations. J Ecol 86:1060–1064. doi:10.1046/j.1365-2745.1998.00326.x Smith RJ, Hopper SD, Shane MW (2011) Sand-binding roots in Haemodoraceae: global survey and morphology in a phylogenetic context. Plant Soil 348:453–470. doi:10.1007/ s11104-011-0874-z StatSoft I (2007) STATISTICA (data analysis software system) Turner BL, Condron LM (2013) Pedogenesis, nutrient dynamics, and ecosystem development: the legacy of T.W. Walker and J.K. Syers. Plant Soil 367:1–10. doi:10.1007/s11104-013-1750-9 Ullah MH, Jabbar A, Khan MA (1983) The influence of soil pH and texture on the adsorption of phosphorus by soils. Pakistan J Agric Res 4:41–46 Viani RAG, Rodrigues RR, Dawson TE, Oliveira RS (2011) Savanna soil fertility limits growth but not survival of tropical forest tree seedlings. Plant Soil 349(12):341–353. doi:10.1007/ s11104-011-0879-7 Viani RAG, Rodrigues RR, Dawson TE, Lambers H, Oliveira RS (2014). Soil pHaccounts for differences in species distribution and leaf nutrient concentrations of Brazilian woodland savannah and Seasonally Dry forest species. Perspect Plant EcolEvol Syst 16:64–74 Walker TW, Syers JK (1976) The fate of phosphorus during pedogenesis. Geoderma 15:1–19. doi:10.1016/0016-7061(76)90066-5 Watt M, Evans JR (1999) Linking development and determinacy with organic acid efflux from proteoid roots of white lupin grown with

Oecologia low phosphorus and ambient or elevated atmospheric CO2 concentration. Plant Physiol 120:705–716. doi:10.1104/pp.120.3.705 Witkowski ETF, Mitchell DT (1987) Variations in soil phosphorus in the fynbos biome, South Africa. J Ecol 75:1159–1171

Zheng SJ, Ma JF, Matsumoto H (1998) High aluminum resistance in buckwheat. I. Al-induced specific secretion of oxalic acid from root tips. Plant Physiol 117:745–751. doi:10.1104/pp.117.3.745

13