INTRODUCTION
Understanding strategies of nutrient allocation and their underlying mechanisms in plants adapted to phosphorus (P)-impoverished soils is an important topic in plant physiological ecology (Lambers et al. , 2006, Veneklaas et al. , 2012). Phosphorus-impoverished soils limit the growth and yield of crops, pastures and forests throughout the world (Conroy et al. , 1990, Fujita et al. , 2003, Herbert & Fownes, 1995, Seneweera & Conroy, 1997, Thomas et al. , 2006). Moreover, a recent meta-analysis showed that P limitation of above-ground plant production is pervasive in natural terrestrial ecosystems (Hou et al. , 2020). Low soil P availability is widespread in Australia (Kooyman et al. , 2016, Viscarra Rossel & Bui, 2016), and plants generally respond to low soil P availability by having a low foliar P concentration (Epstein & Bloom, 2005).
Foliar P concentration is the sum of the concentrations of several major P fractions in leaf cells including inorganic P (Pi) and various P-containing organic compounds (i.e. nucleic acids, phospholipids and small phosphate esters (Veneklaas et al. , 2012). Therefore, variation in foliar P concentration among plant species or due to environmental conditions will reflect differences in both the concentrations of the foliar P fractions and the relative proportions among these fractions. The allocation of P to foliar fractions is likely related to life-history strategy, because these fractions are functionally related to growth, reproduction, and stress tolerance. Shifting P-allocation patterns in leaves is an important mechanism for plants to acclimate to low soil P availability (Hidaka & Kitayama, 2011, Yan et al. , 2019). If strong P limitation occurs, plants shift the allocation of P among foliar P fractions, and this might increase plant fitness under the prevailing conditions (Hidaka & Kitayama, 2011).
Chapin III & Kedrowski (1983) investigated foliar P fractions of four Alaskan tree species and found that nucleic acid P was the largest pool throughout the growing season, and that there was no difference in the proportion of foliar P concentrations between different forest types. However, these species demonstrated a relatively high foliar P concentration (1.3–2.1 mg g-1 dry mass (DM)), so no adaptation of the tree species to P limitation can be expected. Hidaka & Kitayama (2009) found that plants growing on P-impoverished tropical soils increased both leaf mass per area (LMA) and photosynthetic P-use efficiency (PPUE) compared with plants on P-richer soil. These authors suggested that a greater proportion of cellular P may be allocated to metabolic P, rather than to structural P to maintain high PPUE. Yanet al. (2019) investigated foliar P fractions of three species along a two-million-year chronosequence with a strong gradient of available P in south-western Australia, and found that their P allocation pattern was associated with their distribution along the chronosequence, and concluded that the differences are likely adaptive. How plants allocate P among foliar P fractions and exhibit adaptive strategies to efficiently use P in two species in the same genus with contrasting life-history strategies in extremely P-impoverished ecosystems with a Mediterranean climate remains unclear.
The relationship between growth rate and P investment, and the rapidly-emerging field of ecological stoichiometry have shown that species with fast growth rates have low N:P ratios (Reef et al. , 2010). This pattern has been explained by the Growth Rate Hypothesis (GRH), which proposes that fast growth rates are associated with a proportionally greater requirement for P than for N, because organisms must allocate a disproportionately greater proportion of P to P-rich ribosomal RNA (rRNA) to meet the protein synthesis demands needed to support the rapid growth rates (Elser & Hamilton, 2007, Elser et al. , 1996, Sterner & Elser, 2002). Nucleic acids have an N:P stoichiometry of 4:1 (Reef et al. , 2010), and are a major fraction of organic P, with RNA by far the largest proportion (Geider & La Roche, 2002). Within the RNA pool, rRNA is the largest P fraction.
Tree species in ancient landscapes have experienced long-term low soil P status; thus, they likely possess adaptations to P limitation. Non-mycorrhizal Proteaceae are an important component of the vegetation on severely P-impoverished soils in south-western Australia (Hayeset al. , 2014, Lambers et al. , 2013, Pate et al. , 2001). Species in this family typically form cluster roots that effectively mine soil P by releasing large amounts of low-molecular-weight carboxylates to desorb P from soil particles (Shane & Lambers, 2005). It is striking that mature leaves of Proteaceae species from south-western Australia exhibit relatively fast rates of area-based photosynthesis, despite having extremely low leaf P concentrations (Denton et al. , 2007, Lambers et al. , 2012, Sulpice et al. , 2014), while leaves of P-starved crop plants tend to have slow rates of photosynthesis per unit leaf area (Brooks et al. , 1988, Fredeenet al. , 1990, Rao et al. , 1989). Consequently, some of these Proteaceae exhibit a very high photosynthetic P-use efficiency (PPUE, Denton et al. , 2007, Lambers et al. , 2010, Sulpiceet al. , 2014). This high PPUE in Proteaceae from severely P-impoverished habitats is brought about mainly by low foliar rRNA concentrations (Sulpice et al. , 2014) and extensive replacement of phospholipids with galactolipids and sulfolipids during leaf development (Lambers et al. , 2012).
The slow-growing resprouter Banksia attenuata and the faster-growing seeder B. sessilis (Pate et al. , 1991) both produce compound cluster roots (Shane & Lambers, 2005), but have different life histories (Shi et al. , 2020). Banksia sessilis is a short-lived obligate seeder that occurs on shallow sand over laterite or limestone (Hayes et al. , 2019, Pate & Bell, 1999) and allocates more biomass to cluster roots than B. attenuata , which invests more in deep roots (Shi et al. , 2020). This strategy enhances P mobilisation from laterite or limestone by releasing more carboxylates and/or exuding these at a faster rate thanB. attenuata (Shi et al. , 2020). In contrast to B. sessilis , B. attenuata is restricted to deep sand (FloraBase, http://florabase.dpaw.wa.gov.au/) and does not grow fast and complete its life cycle quickly (Bowen & Pate, 2017, Knox & Clarke, 2005, Pateet al. , 1990). McArthur & Wilson (1967) coined the termsr strategy and K strategy to describe selection for rapid population growth in uncrowded populations and selection for competitive ability in crowded populations, respectively. Over time, the meaning of these terms has broadened (Parry, 1981), and according to the broader context, B. sessilis is an r strategist, while B. attenuata is a K strategist. We do not know the physiological pattern of allocating P among foliar P fractions that allows species to exhibit a particular life-history strategy and efficient use of P in contrasting low-P environments. Therefore, we aimed to compare P-allocation patterns in these two Banksia species with contrasting life history. Thus, we measured leaf P and N concentrations, LMA, and concentrations and proportions of P in foliar P-containing fractions in B. attenuata and B. sessilis grown with different soil P availability.
We hypothesised that:
1) With decreasing soil P availability, the foliar total P concentrations of both B. attenuata and B. sessilis would decrease.
2) Banksia sessilis , which exhibits a more opportunisticr -life strategy than B. attenuata , would have a higher foliar NTotal : PTotal ratio and invest more P in nucleic acids than B. attenuata when grown on the same substrate.