P for Two, Sharing a Scarce Resource: Soil Phosphorus Acquisition in the Rhizosphere of Intercropped Species

Over the last 50 years, the use of nitrogen (N) and phosphorus (P) fertilizers has increased at a faster rate than global food production, resulting in substantial decreases of N and P efficiency in agriculture (Fig. 1). Thus, considerable amounts of N and P fertilizers have been wasted in agroecosystems, and the alteration of N and P biogeochemical cycles is among the most visible impacts of human activities on ecosystem services, far exceeding the hypothesized “planetary boundaries” for N and approaching those for P (Rockström et al., 2009). Projections for the future suggest that a substantial further increase in N and P fertilizer use will occur to cope with increasing food demand (Tilman et al., 2001; Vance et al., 2003). Tilman et al. (2001) predicted that global consumption of fertilizer P will increase from 34.3 Tg year−1 in 2000 to a mean projected value of 83.7 Tg year−1 in 2050 (56.2–118 Tg year−1 depending on the calculation scenario). Such increases in fertilizer consumption will further threaten the global N and P cycles. Bouwman et al. (2009) estimated that in 2000, the total input of fertilizer P in croplands amounted to 21 Tg year−1, of which 9 Tg was accumulated in soils and 1 Tg was lost to erosion. Their projections for 2050 for various scenarios of future agricultural development amounted to 29 to 46 Tg year−1 total P fertilizer input and 10 to 23 and 3 to 5 Tg year−1 accumulated in soil and lost to erosion, respectively. This suggests that even in the most optimistic scenario, the contribution of P fertilization of cropland to P movement by erosion will triple over the period 2000 to 2050. While further intensification of agroecosystems is clearly needed to cover the growing food demand over the next decades, we cannot afford to accept the “business as usual” scenario that relies on ever-increasing agricultural inputs and the resulting waste of nutrients in some regions of the world (Vitousek et al., 2009). Relative increase in world annual production of cereals, and global annual consumption of fertilizer N and fertilizer P over the period from 1961 to 2008. The 100-basis in 1961 corresponds to 876.9 Tg cereals year–1, 11.6 Tg fertilizer N year–1, and 4.8 Tg fertilizer P year–1. In 2008, these values were 2,520.7 Tg cereals year–1, 101.6 Tg fertilizer N year–1, and 17.1 Tg fertilizer P year–1. (Data are from FAOSTAT [http://faostat.fao.org/], accessed February 20, 2011). Contrary to N, P is reasonably abundant in the Earth’s crust (1.2 g kg–1 on average) and thus in soils, where it primarily occurs as inorganic P in apatite minerals derived from the bedrock. However, with soil formation and weathering, total P content decreases over time and organic P content builds up at the expense of inorganic P, as shown in soil chronosequences (Richardson et al., 2004). At early stages of development (in young soils), therefore, terrestrial ecosystems are primarily N limited, while at later stages of development (in older soils), they become P limited (Vitousek and Farrington, 1997). Thus, soil P scarcity is especially critical in the tropics, where deeply weathered soils dominate. In addition, unlike N (especially nitrate), phosphate ions are poorly mobile and present at low concentration in soil solution, due to strong and multiple interactions with soil constituents (Hinsinger, 2001). These comprise adsorption onto soil minerals (metal oxides and clay minerals), precipitation as P minerals (predominantly apatite-like minerals), and immobilization as various organic P compounds (soil organic matter and phytate, which is the P storage form in seeds). P limitation is thus widespread, estimated to affect about 5.7 billion ha worldwide (Gaume, 2000). While, in spite of their considerable energy cost, the reservoir of atmospheric N2 used for manufacturing N fertilizers is much larger than required, the situation for P is rather different. World reserves of P ores are indeed finite, and the exact time when their consumption will peak is a matter for debate. High-grade phosphate rocks are definitely expected to be exhausted within the next decades (Cordell et al., 2009), which calls into question the sustainability of current P fertilizer use in developed and emerging countries. Increasing P efficiency in crops without further increasing P inputs requires better exploration and exploitation of soil resources in agroecosystems. To achieve this, we must breed more P-efficient crop genotypes that will make better use of belowground (root architecture and rhizosphere-related) traits (Vance et al., 2003; Lynch, 2007). Another promising option for achieving ecological intensification of agroecosystems (Cassman, 1999) is to make better use of plant diversity, especially niche complementarity and facilitation occurring in the rhizospheres of intercropped species (Zhang et al., 2010). Enhanced productivity of multispecies agroecosystems (intercropping) compared with that of monospecific agroecosystems (each of the component species being grown alone) may be explained by two major processes that result in improved resource use: complementarity and facilitation (Fridley, 2001). Experimentally, these processes can be difficult to tease apart (Loreau and Hector, 2001). Complementarity may be defined as a decrease in interspecific competition and competitive exclusion through resource partitioning between intercropped species (Fig. 2). Species may use a given resource differently in time, in space, and in forms (Fridley, 2001). A well-known example is the complementarity of N use between cereals and N2-fixing legumes, where both species compete for the same pool of soil N, while only the legume can substantially access the additional pool of atmospheric N2 through symbiotic fixation. Facilitation occurs when one species enhances the growth or survival of another (Callaway, 1995). This can occur through (1) direct positive mechanisms, such as favorable alteration of light, temperature, soil moisture, soil nutrients, etc., and (2) indirect mechanisms, such as beneficial changes in soil mycorrhizal or microbial communities. Hereafter, we use facilitation to mean positive interactions by which a species can modify the biotic/abiotic environment of its roots (rhizosphere), ultimately benefitting the intercropped species by increasing nutrient availability (Callaway, 2007). Direct, root-mediated processes altering P availability in the rhizosphere and indirect, microbially mediated processes will be addressed. Competition (A), complementarity (resource use partitioning; B), and facilitation (C) between two intercropped species. Pools represent different forms of a single resource (e.g. nutrient). Solid arrows indicate uptake of the resource by the intercropped species, while dotted arrows indicate mechanisms by which species B can alter resource availability, increasing the size of the available pool at the expense of the unavailable pool (indicated by the curved arrow), thereby improving uptake of the resource by the intercropped species A (facilitation). Such positive interactions are particularly valuable when resources are limited, as occurs in low-input agroecosystems. For example, beneficial effects of intercropping have been observed at lower rates of P fertilizer application but were no longer significant at higher rates (Li et al., 2007). In ecology, the “stress gradient hypothesis” proposes that positive interactions (facilitation) increase in importance and intensity with increasing environmental stress (Brooker et al., 2005). For both natural and managed ecosystems, nutrient use in multispecies stands has been mainly studied for N and especially for cereal/legume intercropping systems. Few studies have focused on cereal-legume interactions with regard to soil P. Recently, this research field has attracted new interest with the reported evidence of enhanced P acquisition for cereals intercropped with legumes (Li et al., 2007). Given the low mobility of phosphate ions in soils, the volume and geometry of the rhizosphere largely determines the pool of P readily accessible to plants. Spatial complementarity can occur (1) when the two species have contrasting root architecture, exploring different soil horizons, and/or (2) because of the plasticity of root systems, combined with possible avoidance strategies (Hauggaard-Nielsen and Jensen, 2005; Li et al., 2006; de Kroon, 2007). In both cases, intercropping may ultimately result in better exploration of the whole soil volume compared with monocropping. The role of root distribution in resource use in intercropping systems has been mostly documented for N. But the conclusion that cereal has better N uptake efficiency due to faster root growth may also apply to P. The discovery that plants can adjust root location depending on nutrient availability and the presence of neighboring plants within a single species (Gersani et al., 2001; Cahill et al., 2010) raises new questions about how soil P is shared in multispecies stands and how “root decisions” (Hodge, 2009) occur for intercropped species. In addition, intercropped species may exhibit contrasting phenologies (Rose et al., 2007) and/or growth periods (e.g. different sowing dates), which may result in differential P requirements over time (Li et al., 1999, 2007). Contrasted sowing/harvest dates may also reduce competition and increase P availability by mineralization of crop residue, which enhances P acquisition of the intercropping system. Soil P occurs as various pools that require different biochemical or chemical reactions to release phosphate ions that are readily taken up by roots. Complementarity can thus occur for two intercropped species tapping into two distinct pools of soil P resources (e.g. inorganic and organic; Li et al., 2008). Intercropped species may also have access to different fractions of each of these pools (Turner, 2008). Cu et al. (2005) observed that wheat (Triticum aestivum) and intercropped white lupin (Lupinus albus) depleted two distinct inorganic P fractions. In Li et al. (2003a), chickpea (Cicer arietinum) mobilized soil organic P and left more inorganic P available to the intercropped wheat. Li et al. (2008) tested this hypothesis on a larger number of inorganic and organic P fractions for durum wheat (Triticum durum) and common bean (Phaseolus vulgaris). Almost all values of soil P pools in the rhizosphere of the two intercropped species were intermediate between those in the monocropped cereal or legume. As for Cu et al. (2005), the rhizospheres of the intercropped species were not dissociated to measure the changes in P fractions separately for each of the intercropped species. Facilitation of P uptake is defined as the positive interactions that result from the ability of one species to increase soil P availability to the benefit of the intercropped species (Callaway, 2007). Thus, for example, the size of the pool of available P is increased at the expense of the pool that is unavailable to species A (Figs. 2C and 3) as a consequence of exudation by species B. This section will focus on how plant roots are able to directly change rhizosphere P availability through either P uptake and exudation of P-mobilizing compounds or as a consequence of interactions with the uptake of other nutrients (Hinsinger, 2001; Vance et al., 2003; Devau et al., 2010). Root exudates play a major role in P bioavailability via several mechanisms: protons/hydroxyls and carboxylates solubilize inorganic P, while root-borne phosphatases hydrolyze organic P (Hinsinger, 2001; Vance et al., 2003). Most cereal/legume intercropping studies implicitly assume that the cereal shall benefit from the legume species (one-way because legumes are to larger amounts of et al., et al., carboxylates and Vance et al., 2003; et al., and phosphatases et al., in their But cereals can also change rhizosphere P availability, and one the facilitation of legume P acquisition by P-efficient cereals or even facilitation in intercropped cereal/legume systems. and microbially mediated positive interactions (facilitation) altering P availability in the rhizosphere of two intercropped species. arrows indicate how species B can P that is not available to species either directly or via soil such as and mycorrhizal or These processes result in increases in the size of the available P pool at the expense of the unavailable pool (indicated by the curved Solid arrows indicate P uptake by the two species from the available P N uptake and the of N used are to play a role in uptake and changes in rhizosphere and P availability et al., 2003). The intercropped cereal will benefit from due to N2 in soils, where such may increase P availability through of P minerals (Hinsinger, 2001). have shown a lower rhizosphere in legumes than cereals, while rhizosphere of the intercropped cereal/legume is intermediate (Li et al., Cu et al., 2005). However, has a with plant Li et al. (2008) reported a significant benefit when durum wheat was intercropped with common to durum wheat grown but they not the with the observed changes of rhizosphere and P These that release rate by the intercropped legume was than that when the legume was grown This as a result of legume N2 due to lower rhizosphere because of competition for soil N with the intercropped as documented (Hauggaard-Nielsen et al., 2009). when legumes to more on soil N when it is readily available due to the of concentration on N2 fixation. This where competition for one resource may facilitation for the acquisition of another resource only but also release and increase in rhizosphere can substantially increase soil P This occurs as a consequence of enhanced of phosphate ions and changes in of oxides and clay minerals to which phosphate ions are (Hinsinger, 2001). Devau et al. both and the substantial contribution of of the durum wheat rhizosphere to observed increases in P These were documented in a soil but occur for a of soil and that P availability is primarily by These suggest that cereals, which exhibit improved N when intercropped and may P acquisition by the intercropped legume. and other organic by roots can compete with phosphate ions for adsorption on thereby P and enhanced P availability (Hinsinger, 2001). exudation rates are and with plant species and and Vance et al., being much in cereals than in legumes such as chickpea and white lupin and et al., These species can amounts of and which are some of the most P-mobilizing carboxylates (Hinsinger, 2001). Thus, it is expected that legumes P acquisition of the intercropped legume species not carboxylates at rates than most one (Li et al., 2010) focused on and in the rhizosphere of intercropped and white lupin or bean for but no on plant or P The was and additional is needed to the of this in intercropping systems. which have the role of P-mobilizing have mainly focused on (Li et al., et al., et al., phosphatases may soil P availability through organic P. Li et al. reported increased P acquisition from organic in wheat intercropped with chickpea compared with wheat grown with a its roots from those of no such when the of P was inorganic which suggests that chickpea P acquisition by the intercropped wheat as a consequence of its ability to hydrolyze the organic P by Li et al. observed both in and plants that higher was observed for chickpea than for not affect the of either species, and P acquisition were The enhanced and P acquisition of the cereal not be to increased in intercropping compared with grown because it also when only inorganic P was et al. a but with another different species and growth intercropping was observed when P was in organic form but P and P availability all increased when inorganic P was between rhizosphere processes such as and P facilitation have not been clearly is to that a of in the rhizosphere is not directly the consequence of the of root-borne but rather from microbial activities of the on intercropping systems has to the of the due to To a this also to rhizosphere and organic of the rhizosphere processes occur over rather While protons/hydroxyls can several from root carboxylates and most are expected to be onto soil and thus poorly mobile in the over than 1 or et al., 2009). Thus, in the of root or mycorrhizal root is for the of positive interactions that the release of P-mobilizing Thus, in to niche of roots of the two intercropped species is needed for P facilitation to This is by root either in the field or in (Fig. which have shown that better growth and/or P acquisition is when is no between the roots of intercropped species. et al. (2008) that where the of is as be expected to occur in most soils, only the neighboring roots of the species benefit from the increased P availability due to by roots of the species. This not only from the mobility of P-mobilizing compounds by such as but especially carboxylates or is also for the phosphate ions to be able to the roots of the species. The for between roots may be by root and more mycorrhizal which can access to P resources from roots of the intercropped species. of either uptake or release of P-mobilizing compounds be as the of a root can its This has been especially documented for the release of and carboxylates in roots of white lupin and bean intercropping in two field with a root and between the species a that root but and no between the root systems of and bean from Li et al. with from the of of the B and and P uptake (C) of and bean from Li et al. with from For each species, different indicate significant among the exudation of P-mobilizing other processes may also changes in rhizosphere P availability as a consequence of interactions with the uptake of other nutrients, as for example N uptake and its on rhizosphere The role of uptake has been by Devau et al. P availability in the rhizosphere of durum wheat not be explained by P uptake combined with rhizosphere due to However, when for Devau et al. a between and rhizosphere P with P via its on especially for clay minerals at thereby the adsorption of phosphate ions and P The uptake of by roots thus the of increasing P Given that uptake can substantially between species, with legumes up more than cereals, legumes may thereby P acquisition of the intercropped we have evidence to the facilitation of P acquisition in between increased P acquisition and rhizosphere processes altering P availability are in with the role of N facilitation in cereal/legume a studies have to measure both rhizosphere processes and changes in P availability et al., et al., Li et al., or P acquisition in intercropping systems. In addition, rhizosphere processes may with soil and plant species (Hinsinger, 2001; et al., 2009), and P availability in the rhizosphere of intercropped species has been mostly studied in These interactions to be in a of soil soils et al., Li et al., 2010). For the hypothesis of facilitation of cereal P acquisition due to rhizosphere has only been tested in soils (Li et al., Cu et al., 2005). also be tested in soils, where more at increasing P direct chemical processes in the facilitation can also occur as a consequence of in the microbial or (Fig. species a on rhizosphere microbial due to in amounts and of root exudates and et al., et al., 2010). For in in the rhizosphere of white lupin were to while in were to the presence of and et al., In addition, roots a of and that microbial and are in and et al., 2003). can thus with plant species et al., 2006; et al., and/or plant interactions et al., 2004). In to the of is increasing evidence that soil chemical such as have a major on the of soil microbial et al., et al., 2010). To this apply to rhizosphere changes question et al., 2009), which be in cereal/legume intercropping systems. studies have shown significant changes in microbial in the rhizosphere of intercropped species compared with those of crops et al., et al., Li et al., 2010). of roots of the intercropped species can result in a common microbial which be a of the of each of the two species et al., However, in the rhizosphere of a of organic compounds be expected are in root between the intercropped species. In that the microbial of intercropped species not be a of the of each of the two species. effects of intercropping on microbial are et al. of bean intercropping on the of The and growth may also among in microbial are not as as changes in microbial and which can plant growth by altering root growth via production or nutrient availability through mineralization et al., et al., et al., 2011). et al. observed a between microbial P and However, is a between microbial and because of these are by a of which form et al., The consequence of such is that changes in the of a single species have on a given et al., 2003). However, even for with low of and their was not by intercropping et al., 2011). are that changes in microbial can be with changes in microbial activities et al., In to the of microbial on P acquisition of intercropped species, of be such as those in the of inorganic or organic P pools in The to increase the availability of inorganic P is soil as it in the production of organic and which are traits among rhizosphere (Richardson et al., 2009). also use phytate, pool of soil organic P et al., through the production of et al., 2008). These that the of from the rhizosphere of species was that for a legume lupin their was on these suggest that intercropping species such as cereals and legumes may different resulting in to in the To this it will be to to microbial from or to and and have been far only for one the et al., to only in et al., 2007). Another indirect in the rhizospheres of intercropped species is enhanced nutrient mineralization due to the The is defined as the change in soil organic matter resulting from the of organic matter and 2008). Thus, it can occur in the rhizosphere via root and 2009). et al. that use the energy from this to in to release organic N when inorganic N is P limitation has been to a but it may be in ecosystems that are primarily P limited, such as in the A positive of to the of organic N and P and may ultimately plant growth et al., 2000). of the is not accessible to all and is in to with those have not been clearly but a are to this et al., et al., 2011). As a the is on both the microbial and the of the organic be used by that their energy to growth rather than or compounds are more to some than et al., 2011). The and the of may thus the and the of the rhizosphere microbial and intensity (2009) indeed a larger in the rhizosphere of than in that of wheat. The has been in the of cereal/legume but of one of the intercropped species that release P in benefit the growth of the plant species via their The of common mycorrhizal the roots of two plants may also increase P availability to intercropped species. of nutrients between the intercropped species via has been documented for N but not for P. Li et al. (2009) that was of N to the intercropped species when the was the cereal with no significant role of In the of N was much increased by the development of when the species was the legume bean which up to of the N taken up by evidence for improved growth and in cereal/legume intercropping systems for ecological intensification of agroecosystems. of of nutrient efficiency in such systems has N, especially with regard to legume N2 fixation. on P are and of these between complementarity and facilitation or evidence that a rhizosphere the improved P efficiency when cereals and legumes are research is needed to positive interactions occur in intercropped cereals and legumes, which be to P acquisition efficiency for both low input and agroecosystems (Zhang et al., 2010). to be for such Given the of the mechanisms, also be further developed and