SEARCH

SEARCH BY CITATION

Keywords:

  • Banksia;
  • carboxylates;
  • citrate;
  • endemism;
  • malate;
  • phosphorus;
  • Proteaceae;
  • species distribution

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References
  • • 
    Banksia species (Proteaceae) occur on some of the most phosphorus (P)-impoverished soils in the world. We hypothesized that plasticity in the exudation of P-mobilizing carboxylates would be greater in widespread than in rare Banksia species.
  • • 
    Glasshouse experiments were conducted to identify and quantify carboxylate exudation in three widespread and six narrowly distributed Banksia species.
  • • 
    High concentrations of carboxylates (predominantly malate, citrate, aconitate, oxalate) were measured in the rhizosphere of all nine species of Banksia on six different soils, but widespread species did not have greater plasticity in the composition of exuded carboxylates.
  • • 
    Based on the evidence in the present study, rarity in Banksia cannot be explained by limited phenotypic adjustment of carboxylate exudation.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

On a global scale, ecosystems with greatest terrestrial plant diversity occur in the tropics (tropical rainforests in South America and Asia) and in Mediterranean regions (Cowling & Lamont, 1998; Myers et al., 2000). Most of these species-rich communities occur on severely nutrient-impoverished soils (Beadle, 1966; Pate & Dell, 1984; Specht & Specht, 1999). Around 75–80% of the Western Australian ‘kwongan’ (scrub heath) species are endemic to the region (Hopper & Gioia, 2004). This region exhibits limited topographical and rainfall variation (Hopper & Gioia, 2004), and the large diversity of the region is strongly influenced by variation in edaphic factors, particularly phosphorus (P) availability (Beadle, 1966). For example, the extreme P sensitivity of Hakea prostrata (Proteaceae) excludes this species from habitats with a slightly higher P supply (Shane et al., 2004a,b), whereas the paraphyletic Grevillea crithmifolia is unaffected by higher P availability and can inhabit slightly more fertile sites (Shane & Lambers, 2006). Similarly, the ironstone endemics, Hakea oldfieldii and Hakea tuberculata (Proteaceae), are edaphically restricted to their habitat, showing rooting patterns that allow superior performance in their own habitat, while excluding them from others (Poot & Lambers, 2003). Hakea verrucosa, the only known mycorrhizal Hakea species, is also edaphically restricted, namely to ultramafic soils in the south-west of Australia (Boulet & Lambers, 2005).

Plant life in the severely nutrient-impoverished systems in Western Australia is associated with a plethora of adaptive mechanisms to acquire and use scarcely available P and micronutrients (Beadle, 1966; Pate & Dell, 1984; Handreck, 1997). The predominantly nonmycorrhizal Proteaceae have ‘proteoid’ or ‘cluster’ roots (Purnell, 1960; Shane & Lambers, 2005). There is a substantial amount of literature on the morphology of the cluster roots of Proteaceae (reviewed in Lamont, 2003), and, more recently, our understanding of their functioning is becoming firmly established (Shane & Lambers, 2005; Lambers et al., 2006). Cluster roots release exudates (carboxylates, protons, phenolics, phosphatases) which enhance the availability of complexed metals and P for the plant, both when P occurs in organic and in inorganic forms (Neumann & Martinoia, 2002; Lambers et al., 2006). Compared with mycorrhizal species without root clusters, nonmycorrhizal plants with root clusters release far more carboxylates (Grierson, 1992; Keerthisinghe et al., 1998; Kamh et al., 1999; Roelofs et al., 2001), and these carboxylates are crucially important in accessing P in the world's most P-impoverished soils (Lambers et al., 2006). What is not known is whether variation in carboxylate composition, such as has been described for calcicole vs calcifuge species (Ström et al., 1994), also offers an explanation for the distribution pattern of kwongan species.

The exceptional success of Proteaceae in the kwongan flora is most probably associated with their highly effective mechanisms for P acquisition (Lambers et al., 2006). Banksia species exhibit a remarkable level of endemism in Western Australia: six species are intrinsically rare and limited to < 100 ha; examples are Banksia brownii, Banksia verticillata and Banksia goodii (George, 1981). More than 10 species are highly regionalized or edaphically restricted endemics; among these are Banksia burdettii, Banksia chamaephyton, Banksia hookeriana, Banksia lanata, Banksia laricina and Banksia scabrella (George, 1981; Lamont & van Leeuwen, 1988). The exact edaphic factors that restrict the distribution of the narrowly distributed Banksia species are not known. Therefore, these patterns of endemism provide ideal opportunities for experimental testing of the role of exuded carboxylates in the distribution of Banksia species. Given over 50 million years in which present forms of Banksia have existed, fine-tuning of composition and rate of carboxylate exudation to different soil types may have isolated populations of Banksia species, as suggested for calcicole and calcifuge species (Ström et al., 1994). Since many of the rare species of this genus have edaphically restricted distributions (George, 1981), we hypothesize that a species’ limited ability to alter its rate or composition of exuded carboxylates restricts efficient acquisition of P to certain soils.

Exudation of carboxylates by plants differs according to plant species, age and physiology. Plants that naturally occur on acid soils (calcifuge species) release more acetic acid, while those growing on calcareous soils (calcicole species) release more oxalic and citric acids (Ström et al., 1994; Zhang et al., 1997). Plants may also alter their carboxylate exudation in response to soil properties: for example, the proportion of malate to citrate exuded by chickpea and white lupin was correlated with soil pH (Veneklaas et al., 2003). We hypothesize that Banksia species, which differ more subtly in their soil preferences, release different types and amounts of carboxylates, dependent on soil type. Different carboxylate exudation patterns will have adaptive value if the effectiveness of different carboxylates varies with soil properties, in particular the extent and mechanism of P sorption. Furthermore, different carboxylates vary in their effectiveness in releasing P from different sources (Randall et al., 2001). Species with a wide distribution with respect to soil type may have genetically distinct ecotypes on different soil types. Alternatively, such widespread species may be more phenotypically plastic in their exudate composition, as found for B. grandis (Lambers et al., 2002). Species that are restricted to one soil type would lack both significant genetic differentiation and phenotypic plasticity. If exudation patterns are correlated with edaphic preferences, and lead to enhanced P acquisition and growth, this would provide a likely explanation for edaphic endemism and range-restriction in Banksia species in south-west Australian kwongan. In restricted species, carboxylate composition is expected to be unchanging across soil types, and (above- and below-ground) growth is expected to be highest on the parent soil. In widespread and edaphic-generalist species, carboxylate composition of the rhizosphere would vary according to whichever composition was optimal for that soil, and growth would be invariant across soil types. To investigate this, we measured the growth, P content and rhizosphere carboxylates of three widespread and six narrowly distributed Banksia species grown in a range of soil types, including one from their natural habitat.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Seed collection and plant growth

Banksia fruiting cones were collected from the northern sandplains between Moore River National Park (115°50′ E, 31°50′ S) and Geraldton, Western Australia (114°36′ E, 28°49′ S). Cones of B. burdettii E.G. Baker, B. chamaephyton A.S. George, B. hookeriana Meissner, B. lanata A.S. George, B. laricina C. Gardner, and B. scabrella A.S. George were collected in March–June 2003. In addition, three widely distributed, co-occurring species, B. attenuata R. Brown, B. menziesii R. Brown and B. prionotes Lindley, were collected. B. hookeriana is closely related to B. prionotes, B. burdettii is closely related to B. menziesii, and B. chamaephyton, B. scabrella, B. lanata and B. laricina are distantly related to other species (Mast & Givnish, 2002). The release of Banksia seeds is usually triggered by fire. Cones were burnt for 3–5 min using kerosene, soaked for 1 h in water, dried for 7 d, and the seeds were removed. Five seeds were germinated in pots lined with perforated plastic bags with 3 kg of soil, and seedlings were thinned to one per pot after 2 wk. The plants were cultivated in a cooled glasshouse at the University of Western Australia, with light shade provided during the summer months, and watered with deionized water. Plants were grown in each of six soils, which were selected as representing the typical soils in which each of the six rare species naturally grew (Table 1). The widespread species B. attenuata occurred naturally on all soils, while B. menziesii occurred on soils 1, 5 and 6 and B. prionotes occurred on soils 2 and 6. Soil analyses included pH, plant-available P (Colwell method; Rayment & Higginson, 1992), organic carbon (Walkley and Black method; Rayment & Higginson, 1992) and oxalate-extractable iron (Fe) and aluminium (Al) (Rayment & Higginson, 1992). Phosphorus-retention index (PRI), a measure of phosphate sorption by soils, was calculated according to Allen et al. (2001), as follows: a solution containing 10 mg P l−1 in 0.02 m KCl was mixed with the soil for 16 h, and the concentration of P was measured. The sorbed P was calculated and the PRI was calculated as the sorbed P divided by the P concentration (Allen et al., 2001).

Table 1.  Characteristics of the six soils used in the experiments
SiteSpeciespH (H2O)Organic C (mg g1)Total P (µg g1)PRI (ml g1)Fe (µg g1)Al (µg g1)
  1. Soils are listed in increasing phosphorus-retention index (PRI); a higher value indicates stronger sorption, but all values indicate low sorption capacity. C, carbon; P, phosphorus; iron (Fe) and aluminium (Al) are oxalate-extractable amounts.

1 Banksia laricina 5.211.6140.5 54 43
2 Banksia burdettii 5.4 5.4170.5 47 46
3 Banksia chamaephyton 5.4 8.0120.6 38 40
4 Banksia scabrella 5.5 4.6170.6 42 43
5 Banksia lanata 5.3 7.0401.4 91 61
6 Banksia hookeriana 5.7 7.7234.7103215

Plant analysis

Plants were grown for 35 wk, and harvested during winter when cluster roots were active. Shoots were separated into stems and leaves, and fresh mass and leaf area were measured using an Epson 1680 scanner and the Winrhizo (Regent Instruments, Quebec, Canada) root scanner program. The bags lining the pots were cut open, and the cluster roots (conglomerates of dense fine roots and soil) were removed from the bulk soil. Fresh, mature cluster roots were identified, and soil was washed off the remaining roots, which were then separated into cluster roots and noncluster roots, and their fresh mass measured. Dry mass of shoot parts and roots was determined after drying for 48 h at 80°C. Subsamples of seeds, shoots and roots were finely ground, digested using nitric/perchloric acid digestion and analysed for P, sulphur (S), calcium (Ca), potassium (K), magnesium (Mg), iron (Fe), boron (B), molybdenum (Mo), manganese (Mn), zinc (Zn), nickel (Ni), sodium (Na) and Al using a Varian Vista Axial Inductively Coupled Plasma Atomic Emission Spectrometer (Varian, Palo Alto, CA, USA).

Fresh mature clusters were sampled at the stage of development in which carboxylates are released in an exudative burst (Watt & Evans, 1999; Shane et al., 2004b). Root clusters were gently shaken to remove excess soil. The soil that remained attached to cluster roots was defined as rhizosphere soil. Clusters were transferred to a container with a known amount of 0.2 mm CaCl2 to ensure membrane integrity (Veneklaas et al., 2003), and gently shaken to extract the root exudates. The pH of the sample was measured, and a subsample of this rhizosphere extract was taken and filtered through a 0.22 µm syringe filter into a 1 ml high-pressure liquid chromatography (HPLC) vial. The HPLC samples were acidified with orthophosphoric acid, and frozen at −20°C, until analysis. Working standards of malate, malonate, lactate, acetate, maleate, citrate, succinate, cis-aconitate, trans-aconitate acid and oxalate (ICN Biomedicals Inc., Aurora, OH, USA) were used to identify carboxylates on an Altima C-18 reverse-phase column (250 × 4.6 mm; Alltech, Deerfield IL, USA) (Cawthray, 2003). Oxalate was measured using a Prevail organic acid column (250 × 4.6 mm; Alltech) run in 0.1% (w : v) trifluoroacetic acid (Sigma-Aldrich, Castle Hill, Australia), pH 1.9. Adsorption isotherms for malate, citrate and aconitate were determined for soils 1 and 6, using the methods of Jones & Brassington (1998). Although adsorption of carboxylates differed among soils, negligible adsorption (< 0.2 mmol kg−1) of carboxylates occurred over a range of equilibrium concentrations (0–20 mm).

Statistical analyses

Analyses of variance were performed using Genstat edition 7 (VSN International Ltd, Rothamsted, UK). Multiple comparisons were made using Tukey's HSD test. Log transformation was used, when necessary, to conform data to assumptions of homogeneity and normality.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

The growth of the nine Banksia species varied significantly among species and soil type (Fig. 1). All Banksia species produced large cluster root mass ratios (the ratio between cluster-root mass and total plant biomass) regardless of treatment (Fig. 1). Values ranged from 0.32 to 0.75, with a mean of 0.60. Cluster roots are, however, short-lived (approx. 3 wk in Hakea prostrata; Shane et al., 2004b), and many of the harvested clusters were senesced. The shoot mass of B. attenuata and of B. chamaephyton did not vary significantly according to soil type, but the shoot mass of most species varied marginally (e.g. in B. menziesii and B. scabrella) or to a large extent (e.g. in B. prionotes and B. laricina) irrespective of a species’ rarity (Fig. 1). The shoot mass of B. laricina in soil 6 was only 19% of the shoot mass of B. laricina grown in the soil on which it grows naturally (soil 1). No species had the greatest shoot mass when grown in its ‘native’ soil.

image

Figure 1. Leaf, stem, noncluster- and cluster-root mass of nine Banksia species grown in six soils. The soil in which each species is grown is indicated, and the soil from the natural habitat of the species is designated the home soil (H), except for B. attenuata, which occurred at all sites. Details of the soils used are shown in Table 1. Within each species, bars that do not share the same letter differ significantly in leaf mass (n = 6). Significant differences within species are indicated as: ns, no significant difference; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Download figure to PowerPoint

Rhizosphere carboxylate concentrations from fresh, mature root clusters ranged from 2.5 to 101 µmol g−1 root FM, and varied considerably according to species and soils (Fig. 2a). Banksia laricina showed the highest concentrations of rhizosphere carboxylates and B. attenuata the lowest. Citrate was the most abundant carboxylate (0.3–54 µmol g−1 root FM), and oxalate, aconitate and malate were commonly observed in rhizosphere samples; fumarate was found at very low concentrations (0–80 nmol g−1 root FM, median 2.5 nmol g−1 root FM) in the rhizospheres of all Banksia species (not shown in Fig. 2a). There was no pattern in either carboxylate composition or concentrations that was clearly associated with either plant species or soil characteristics (Fig. 2b). Carboxylate concentrations and compositions were equally as variable in rare banksias and widespread species when grown in a range of soils. Rhizosphere pH was typically reduced by 0.8 units compared with the pH in bulk soil, and this was not dependent upon species, soil type or rhizosphere carboxylate concentrations (data not shown).

image

Figure 2. (a) Rhizosphere carboxylate concentrations and (b) mean composition from root clusters of nine Banksia species grown in six soils (n = 6). The soil in which each species is grown is indicated, and the soil from the natural habitat of the species is designated the home soil (H), except for B. attenuata, which occurred at all sites. Details of the soils used are given in Table 1. Fumarate was observed in the rhizospheres of all Banksia species at very low concentrations (0–80 nmol g−1 root FM, median 2.5 nmol g−1 root FM) and is not shown in the graph. Composition of carboxylates in the rhizosphere of Banksia species is the fraction of total carboxylates on a mol g−1 root FM basis. Within each species, bars that do not share the same letter differ significantly in total carboxylate concentrations. Significant differences within species are indicated as: ns, no significant difference; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Download figure to PowerPoint

Total plant P content varied depending upon species and soil treatments (Fig. 3). The P content of the seedlings was strongly influenced by P content of the sown seeds – 12–70% of total seedling P at harvest was estimated to have been derived from seed P (Fig. 3). The potential contribution of seed P to seedling P was strongly correlated to seed size (r2 = 0.73).

image

Figure 3. Total plant phosphorus (P) content of Banksia seedlings compared with initial seed phosphorus contents for nine Banksia species grown in six soil types (± SE, n = 4). The soil in which each species is grown is indicated, and the soil from the natural habitat of the species is designated the home soil, ‘H’, except for B. attenuata, which occurred at all sites. Details of the soils used are shown in Table 1. For each species, bars that do not share the same letter differ significantly in total plant P. Significant differences within species are indicated as: ns, no significant difference; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Experimental testing of nine Banksia species growing in a range of soils showed that, contrary to our hypothesis, the three broadly distributed species did not have greater plasticity in carboxylate-exudation patterns compared with the six narrowly distributed species. In previous studies, phenotypic plasticity has been clearly demonstrated in terms of leaf morphological changes in response to, for example, light (Bradshaw, 1965; Sultan, 2000). Root morphological plasticity has also been assessed, primarily in response to nutrient supply and flooding (Hutchings & De Kroon, 1994; Bell & Sultan, 1999; Hodge, 2004; Kembel & Cahill, 2005), but few, if any, studies have assessed root physiological plasticity among a range of rare and widespread plant species. Rare species have previously been shown to have lower reproductive output than congeneric widespread species (Murray et al., 2002; Lavergne et al., 2004), but other generic differences relating to root traits have seldom been reported for rare and common species (Poot & Lambers, 2003).

Phenotypic plasticity in carboxylate exudation in Banksia species

We hypothesized that specialization of the P-acquisition mode explains edaphic endemism and range-restriction in Banksia species in the south-western Australian kwongan. However, contrary to our hypothesis, the widespread species, B. attenuata, B. menziesii and B. prionotes did not express greater plasticity in their rhizosphere carboxylate composition compared with narrowly endemic species in the soils used. Therefore, the present data do not support our hypothesis. Plasticity of carboxylate exudation has previously been observed in the widespread species B. grandis, which had a different carboxylate composition according to whether aluminium phosphate or iron phosphate was supplied (Lambers et al., 2002). In that study, plants grown with Fe–P exuded monocarboxylates in addition to the tricarboxylates and dicarboxylates exuded when plants were grown with Al–P. The authors argued that the carboxylate composition reflected the plants’ ability to perceive a difference in the chemistry of the rhizosphere. We found no indication that phenotypic plasticity in carboxylate exudation was greater in widespread species compared with that of narrowly distributed species. Therefore, we reject our hypothesis that plasticity in carboxylate composition accounts for differences in distribution patterns among widespread and narrowly distributed Banksia species.

Five tricarboxylates and dicarboxylates were identified in the rhizospheres of Banksia root clusters, with citrate in the greatest abundance. Citrate and malate are the most commonly reported exudates in crop species (Gardner et al., 1982; Hoffland et al., 1989), while oxalate has also been observed, for example in the rhizospheres of buckwheat and rice (Kirk et al., 1999; Zheng et al., 2005). The tricarboxylic acid aconitate has been observed in exudates from many Australian Proteaceae (Grierson, 1992; Roelofs et al., 2001).

Carboxylates play a pivotal role in the mobilization of P, particularly from P that is sorbed to hydrous oxides of Al, Fe or Ca (Randall et al., 2001). The function of carboxylates is to complex metals, thus displacing P from charged surfaces. The formation constants of carboxylates with these cations indicate that desorption of P is typically greatest for tricarboxylic acids compared with dicarboxylic acids, which, in turn, have greater desorption capacity than monocarboxylic acids (Ryan et al., 2001). This desorption capacity allows nonmycorrhizal Proteaceae to inhabit the most P-impoverished soils, where mycorrhizal species that lack this capacity show lower abundance (Lambers et al., 2006). When widespread Banksia species were grown on soils collected at sites where rare species occurred, the widespread species typically did not exude the same composition of carboxylates that rare species did. Thus, the present results do not provide evidence for the hypothesis that a particular carboxylate composition is optional for P mobilization on a given soil. All soils showed low P sorption and carboxylate sorption was negligible. The reason why Banksia species exude a range of carboxylates, rather than only citrate, which is very effective at mobilizing P, is not clear. The composition of exuded carboxylates can vary with root-cluster development (e.g. in H. prostrata aconitate concentrations decreased with age, and citrate concentrations increased; Shane et al., 2004b). Further work is needed to understand the benefits, if any, of exuding a range of carboxylates that vary in their capacity to solubilize P.

While P-deficient plants tend to exude greater amounts of carboxylates than P-adequate plants (Hoffland et al., 1989; Shane et al., 2003), we found no correlations between the concentration of rhizosphere carboxylates (or individual carboxylates) and leaf P concentrations. Neither total concentrations of rhizosphere carboxylates nor individual carboxylate concentrations varied in any predictable fashion with plant P status. Given the measured leaf P concentrations (0.14–0.32 mg P g−1 DW), the most likely explanation for the lack of a relationship between leaf [P] and carboxylates is that none of the plants grown in our soils experienced a leaf [P] sufficiently high to inhibit root-cluster formation or reduce exudation rate (Randall et al., 2001).

Growth and P acquisition of Banksia species

Although soils used in these experiments differed little in measured characteristics, they often had a pronounced effect upon the growth of Banksia species. The widespread B. attenuata grew moderately well in all soil types, but the equally widespread species B. prionotes grew poorly in four soils. Conversely, the narrowly distributed species B. lanata and B. scabrella had similar growth in all soils, and none of the narrowly distributed species exhibited the best growth in their native soil (Fig. 1). The variation in growth in these soils indicates that subtle differences exist among the soils that have not been identified by us, but that are important for plant growth.

The accumulation of P by Banksia species was equally variable among widespread and narrowly endemic species. Of the widespread species, only B. attenuata did not differ in total seedling P content when grown in a range of soils, but B. chamaephyton, B. hookeriana, B. lanata and B. scabrella all accumulated similar amounts of P when grown in a range of soils. Plant P analyses indicated that seedlings contained a significant amount of putative seed P, particularly in species with large seeds (Fig. 3). Studies by Stock et al. (1990) and Milberg & Lamont (1997) have suggested that the large seeds of the Proteaceae allow the growth of seedlings to proceed in nutrient-impoverished environments which is particularly important in the absence of functional mycorrhizal associations (Pattinson & McGee, 2004).

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Widespread Banksia species did not show a greater ability to adjust rhizosphere carboxylate exudation in response to growth in different soils than narrowly distributed species. In addition, widespread species did not alter the composition of carboxylates to match those of the rare species on their native soil. Thus, there is no evidence to suggest that a particular carboxylate composition was optimal for a given soil. Widespread and rare species had equally variable growth in a range of soils. Neither growth nor P accumulation by narrowly distributed banksias was optimized on their native soil. Based on the evidence in the present study, rarity in Banksia cannot be explained by limited phenotypic adjustment of carboxylate exudation.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

This research was funded by the Australian Research Council. Banksia seeds were collected under the appropriate collection permit from the Department of Conservation and Land Management, Western Australia. Greg Cawthray provided excellent assistance with field sampling and carboxylate analyses. Ben Croxford, Chris Szota and Rob Creasy are thanked for assistance with field sampling.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References