The mixotrophic nature of photosynthetic plants

Additional keywords: endocytosis, mixotrophy, organic nutrients, plant nutrition, root hairs.

Terrestrial primary production and the resulting vast phytomass is fuelled by the photosynthetic power of the plant shoot, complemented by the root’s acquisition of water and nutrients from soil. Although most terrestrial organisms live predominantly either above- or belowground, plants, with the notable exception of epiphytes, span both. Green plants are considered autotrophs that are capable of converting inorganic material into organic matter using the electromagnetic energy of light. The atmosphere provides the substrates for photosynthesis, whereas all other essential elements are predominantly derived from soil (Marschner 1995). A plant is truly autotrophic when its roots acquire only inorganic nutrients, including nitrate, ammonium, base cations, ortho- and polyphosphates and sulfate. Autotrophy has been shown in hydroponic experiments on the essentiality of particular elements in which plants complete their life cycle in the absence of exogenous organic compounds (Arnon and Stout 1939; Marschner 1995; Epstein and Bloom 2005). However, roots in soil are exposed to numerous organic compounds, and much of the soil’s nitrogen (N) and phosphorus (P) is organic. Soil N occurs predominantly as amino acid-based molecules (Schulten and Schnitzer 1997), organic P, for example, as phytate and nucleotides, and organic sulfur as a variety of organosulfur compounds (Dyer and Wrenshall 1941; Schulte and Kelling 1999). In addition, soils contain a large range of carbon compounds such as organic acids and sugars, including those released by roots, and there are bidirectional fluxes (soil–roots) of organic matter (Jones et al. 2009).

There is broad agreement that plants acquire N and P in inorganic form. Inorganic N is relatively mobile in soil, although orthophosphate has significant restrictions on its mobility, but has greater mobility than phytate or polyphosphate. The acquisition of inorganic ions and molecules by roots involves plasma membrane located transporters that catalyse selective uptake, and accumulation, of inorganic N and P. It is generally assumed that plants are net exporters of organic compounds into soil and importers of inorganic ions and molecules, but recently the role of organic compounds as plant nutrients is being re-evaluated (reviewed by Paungfoo-Lonhienne et al. 2012).

It therefore seems timely to explore mixotrophy – the combined nutritional modes of autotrophy and heterotrophy – in ecological and mechanistic contexts. We focus here on N and P because they occur as organic and inorganic compounds in soils, and their supply is a major determinant of plant growth in natural ecosystems and bioproduction systems (reviewed by Lambers et al. 2008; Peltzer et al. 2010; Ma et al. 2012). We briefly outline nutrient acquisition strategies to highlight root specialisations, and discuss mixotrophy in algae and in plants. The last section discusses how larger organic molecules may enter roots, and the function of organic compounds as signalling molecules that increase root branching and the presence of root tips and associated root hair zones.

The four main specialisations for nutrient acquisition are root exudates and their extreme occurrence in the functioning of cluster roots, and symbioses with mycorrhizal fungi and N₂ fixing bacteria (Fig. 1). Although exudates universally contribute to root function, their amount and composition varies among species and growth situations, the other specialisations are species specific. The mycorrhizal symbioses are most common and formed by most of plant species with a range of fungal partners, different extent of mycorrhizal colonisation, and varying benefits derived from the symbiosis. The dependence of plants on mycorrhizal fungi for accessing nutrients from organic matter and soil minerals is well established (Smith and Read 2008). Mycorrhizas are not the focus here, although we discuss mycoheterotrophy in the context of green orchids.

Fig. 1.  Overview of plant strategies for nutrient acquisition. Four acknowledged strategies are (i) root exudates, (ii) mycorrhizal associations, (iii) N₂ fixing symbioses in leguminous and actinorhizal species, (iv) cluster roots.

An estimated 10% of plant taxa have N₂ fixing symbioses (Franche et al. 2009), providing an alternative N source to soil N at the cost of photosynthates required to operate N₂ fixation by the bacterial symbionts and, especially, the production and maintenance of the specialised structures (nodules, rhizothamnia and coralloid toots) housing the symbiont. Far fewer plant species have N₂ fixing symbioses than mycorrhizas, although most ecosystems contain N₂ fixing plants. N₂ fixing symbioses are beyond the scope of this review (see Sprent 2001).

Several thousand plant species do not form mycorrhizal symbioses, including all species of the Proteaceae, many sedges, and species in several other plant families (Tester et al. 1987; Lambers et al. 2008), including the model plant Arabidopsis thaliana and other Brassicaceae. However, the recently discovered symbiotic associations with endophytic dark septate fungi are quite common (Caldwell et al. 2000), and occur in Arabidopsis thaliana (Mandyam et al. 2013). Most Proteaceae, many sedges, some N₂ fixing species (perhaps most actinorhizal species and some legumes) and other species produce cluster roots in response to low nutrient availability. Cluster roots are considered particularly beneficial in P-depauperate soils (reviewed by Dinkelaker et al. 1995; Neumann and Martinoia 2002; Lambers et al. 2006). Cluster roots impact a small volume of soil by exuding protons, organic acids, secondary compounds and enzymes (Dinkelaker et al. 1995; Shane and Lambers 2005; Paungfoo-Lonhienne et al. 2009), and, less well characterised, through interaction with microbes (Marschner et al. 2002; Weisskopf et al. 2011).

Ecosystems have been classified along a gradient of ‘organic-to-inorganic nutrients dominated’, with the former requiring greater specialisation by the plant to access nutrients than the latter. In slow-mineralisation organic-nutrient rich ecosystems, such as heathlands and boreal forests, specialised ecto- and ericoid mycorrhizal fungi depolymerise organic matter via extracellular enzymes, and supply nutrients to plants (Read and Perez-Moreno 2003; Lambers et al. 2008). In rapid-mineralisation ecosystems such as grasslands, plants are have mostly arbuscular mycorrhizas and are thought to acquire inorganic nutrients that are derived from breakdown of organic matter by soil microbes (Read and Perez-Moreno 2003).

In light of all root specialisations, it remains uncertain what proportion of N and P is acquired by roots as inorganic or organic compounds. The difficulties associated with studying rhizosphere processes, including chemical characterisation of nutrients and quantification of their fluxes, have limited progress in understanding the importance of organic N and P for plant nutrition. However, roots take up a variety of organic compounds (Näsholm et al. 2009; Paungfoo-Lonhienne et al. 2012). An important consideration for the discussion on mixotrophy is that irrespective of root specialisation, a significant proportion of the root surface is free of fungi or other microbes (Bais et al. 2006).

Root exudates affect the availability of nutrients directly or indirectly (reviewed by Walker et al. 2003; Bais et al. 2006; Jones et al. 2009). Exudates are complex chemical mixtures of low and high molecular weight compounds, including organic acids, phytosiderophores (only in the Poaceae), sugars, vitamins, amino acids, purines, nucleosides, inorganic ions, dissolved gases, high molecular weight carbohydrates (mucilage), enzymes and root border cells (reviewed by Dakora and Phillips 2002). Exudates are communication agents and facilitate nutrient acquisition through various mechanisms (see reviews by Walker et al. 2003; Bais et al. 2006). The mechanisms that plants use to interpret the chemical signals in the rhizosphere that originate from soil organisms and other plants, remain largely unknown (Bais et al. 2006). This has relevance to mixotrophy, and is discussed below.

Root-derived enzymes released into the apoplast and rhizosphere may contribute to mixotrophy if they generate organic compounds that are subsequently acquired by roots. Root exoenzymes have been studied for their role in catalysing the breakdown of large organic molecules as well as other functions. High-molecular mass inorganic compounds (e.g. poly-phosphates) and organic compounds (e.g. protein, phytate, DNA) can be degraded by root-derived phosphatases, proteases, phytases and DNAses (reviewed by Paungfoo-Lonhienne et al. 2012). Phosphatases are exuded, for example, from cluster roots of Proteaceae and Fabaceae (Dinkelaker et al. 1995; Yamamura et al. 2002; Wasaki et al. 2003). In legumes, root phosphatase activity is higher than in other forbs regardless of the presence of nodules (Venterink 2011), and soil associated with N₂ fixing plants has higher extracellular phosphatase activity than soil without N₂ fixers (Houlton et al. 2008). Root-derived proteases have been described in Hakea (Proteaceae), carnivorous plants, Arabidopsis and other species including crops (Adlassnig et al. 2012). Proteases are present in the root apoplast (Tornero et al. 1996; Hamilton et al. 2003), and the presence of protein degradation products in the apoplast of inner root cortex suggests that proteases are active in the apoplast and at the root surface (Paungfoo-Lonhienne et al. 2008).

From an evolutionary viewpoint, the release of enzymes into the rhizosphere is advantageous only if the nutrient gain outweighs the nutrient loss in the exudates, or alternatively, if breakdown products have important signalling roles (Paungfoo-Lonhienne et al. 2008). The presence of enzymes in the apoplast may catalyse the degradation of larger molecules that enter roots via transpiration-driven movement of the soil solution and diffusion, but this is not well characterised. Overall, there has been comparatively little research on root exoenzymes for plant nutrition, as it was generally assumed that microbes drive the breakdown of soil organic compounds. The rhizosphere was treated as a ‘black box’ and researchers were satisfied to consider microbial activity in general terms (Silberbush 2013). Recent advances in molecular methods characterising microbial communities in soil, rhizosphere and roots promise to advance knowledge of rhizosphere dynamics (Bulgarelli et al. 2012; Lundberg et al. 2012).

The positive relationship between the release of enzymes and access to organic nutrients that has been observed in plants tested in controlled conditions has not always been confirmed in soil. For example, wheat genotypes with greater exudation of phosphatase acquired more P from organic P when grown axenically, but not when grown in soil (George et al. 2008). Similarly, transgenically-enhanced root exudation of phytase improved plant growth with phytate in controlled conditions, but to a lesser extent in soil (reviewed by Richardson et al. 2009). Root-derived exoenzymes may have numerous roles, although how much they contribute to the breakdown of organic compounds in the apoplast relative to the rhizosphere, and subsequent uptake of the resulting compounds, is unknown.

Mixotrophy, as phagomixotrophy, is a common and biogeochemically important strategy in eukaryotic phytoplankton and a range of planktonic and benthic protists and metazoans which have acquired photosynthesis by endosymbiosis of cyanobacterial or microalgal cells or through kleptoplasty (Raven 1997; Stoecker 1998; Raven et al. 2009; Esteban et al. 2010; Hartmann et al. 2012; Flynn et al. 2013). In most planktonic habitats the strategy predicted to maximise the abundance of mixotrophs is a predominantly photosynthetic mode of energy gain in combination with consuming bacteria and other microbes which increases acquisition of essential elements (Crane and Grover 2010), thereby outcompeting planktonic phototrophs and phagotrophs (Tittel et al. 2003; Hartmann et al. 2012).

Is mixotrophy an ancestral state in algae? Endocytosis, itself a synapomorphy of eukaryotes, was how all extant algae and hence embryophytic (‘higher’ plants), regardless of whether or not they exhibit mixotrophy, obtained their plastids by primary, secondary or tertiary endosymbiosis followed by genetic integration (Raven 1997; Raven et al. 2009). It is likely that phagomixotrophy was retained in some clades after the genetic integration of plastids (Raven et al. 2009). Phagotrophy also underlies kleptoplasty, where plastids from an algal food item are retained in a functional form in an otherwise heterotrophic cell (Raven et al. 2009; Flynn et al. 2013). Such kleptoplastids are capable of long-term functionality if the nucleus from the plastid source is retaining a functional state with the plastids which occurs in the ciliate Myrionecta rubra (Johnson et al. 2007). However, in some cases, plastid longevity occurs without retention of food alga nuclei or prior horizontal gene transfer from the algal nucleus to the grazer nucleus (Wagele et al. 2011; Pillet and Pawlowski 2013).

In addition to phagomixotrophy, most algae are also capable of osmomixotrophy, i.e. uptake of individual soluble organic molecules by transporters in the plasmalemma (Raven et al. 2009; Flynn et al. 2013) and perhaps, of soluble macromolecules by fluid-phase endocyotosis, so-called pinocytosis. The significance of osmomixotrophy in phytoplankton under natural conditions is unclear; it is possible that the net flux of low molecular mass organic solutes is outwards under many conditions (see Flynn et al. 2013). Osmomixotrophy also occurs in photosynthetic bacteria such as cyanobacteria, as well as in proteobacterial anoxygenic photosynthetic bacteria. Particularly widespread in the surface ocean are anoxygenic aerobic photosynthetic proteobacteria, and the otherwise heterotrophic proteobacteria with proteorhodopsin (Zubkov 2009; Raven and Donnelly 2013). Here the photochemical reactions do not power autotrophic CO₂ assimilation, but spare the use of respiratory processes in energising solute transport and ADP phosphorylation, and so may legitimately be considered mixotrophs since the cells use two sources of energy, i.e. electromagnetic radiation and dissolved organic compounds (Zubkov 2009; Raven and Donnelly 2013).

Osmomixotrophy as a source of organic carbon permitting growth (in the dark) or increasing growth (in the light) is not universal in autotrophs. Obligate photoautotrophy, i.e. the inability to grow other than with light as the energy source and inorganic nutrients supplying essential elements, has been reported for several algae (including cyanobacteria) (Raven 2012). As with algae that cannot be grown in laboratory culture under any conditions so far used, so with obligate photoautotrophic algae: perhaps the next culture attempt with different conditions will allow growth (for so far unculturable organisms) or growth with a contribution from external organic carbon (for obligate photoautotrophs). However, for the moment at least there are obligately photolithotrophic algae (Raven 2012). It is important to note that the obligate photoautotrophy does not preclude the assimilation of exogenous organic compounds, just their involvement in growth (Raven 2012). A final point about obligate photoautotrophy is that technical complications mean that it is not readily tested for embryophytic, and especially vascular plants.

An endocytosis-like uptake of protein by the planktonic heterotrophic planctomycete bacterium Gemmata obscuriglobus was recently reported (Lonhienne et al. 2010). This process is probably not involved in mixotrophy in anoxygenic photosynthetic bacteria, since the planctomycetes have a uniquely complex cell structure for a bacterium, and no planctomycete is known to have photosynthesis (Fuerst et al. 1993). However, a planctomycete contains proteorhodopsin (McCarren and DeLong 2007), so the endocytosis-like means of protein uptake could be involved in this variant on mixotrophy. In summary, the categorisation of autotrophic phytoplankton and heterotrophic microzooplankton does not account for the dual mode of nutrition and warrants revision of the dichotomous concept (Flynn et al. 2013).

A related process, which overlaps with mixotrophy phylogenetically and functionally, is auxotrophy – the inability of some algae to make all enzyme cofactors known as vitamin in metazoan nutrition. The most widespread vitamin requirement is that of cobalamin (vitamin B₁₂), with 155 of the 306 algae examined needing an external supply of vitamin B₁₂ (Croft et al. 2006). Where the vitamin B₁₂ auxotrophs are also phagomixotrophs the prey is presumably a source of the vitamin, bearing in mind that only archaea and bacteria are able to synthesise vitamin B₁₂. Auxotrophic algae that are not phagomixotrophs must obtain the vitamin from associated bacteria or archaea or from the growth medium. Vitamin B₁₂ dependence comes about from loss of an alternative, vitamin B₁₂-independent, pathway of methionine synthesis (Helliwell et al. 2011). For other vitamins, auxotrophy comes about from loss of the capacity to synthesise the vitamin. Since embryophytic plants have no known requirement for exogenous vitamins, the assumption is that they evolved from a charophycean green alga with no requirements for exogenous vitamins. This is the most economical hypothesis, but it does not rule out the alternative of an auxotrophic algal ancestor with subsequent horizontal gene transfer that overcame the need for external vitamins.

Mixotrophy has been a recent focus of phytoplankton research (e.g. Bronk et al. 2007; Gómez-Baena et al. 2008; Zubkov et al. 2008; Poretsky et al. 2010; Flynn et al. 2013), and photosynthetic plants are also investigated for their ability to satisfy energy and nutritional needs with organic carbon (C) compounds. An obvious parallel to algal phagomixotrophy in green plants is the carnivorous mode of nutrition (Fig. 2; Raven et al. 2009). We do not consider carnivorous angiosperms in detail here, but note that they are photosynthetic (Juniper et al. 1989; Raven et al. 2009; Król et al. 2012), can take up organic C in amino acids and peptides via transporters (Schulze et al. 1999) and via endocytosis (Adlassnig et al. 2012), but have a significant C cost in trap construction and maintenance and in enzyme secretion (Ellison and Adamec 2011; Sirová et al. 2011).

Fig. 2.  Acknowledged mixotrophic strategies of green plants. Animal prey, fungi or host plants provide organic nutrients to carnivorous, partially myco-heterotrophic and hemiparasitic photosynthesing plants (photograph courtesy of Jeremiah Harris, Leilani Nepenthes, Colorado Springs CO (Dionaea muscipula), Eleanor Saulys and the Connecticut Botanical Society (Corallorhiza trifida), and John Barkla, NZ Plant Conservation Network (Korthalsella lindsayi) and reproduced with permission.

Another variant on mixotrophy is the mycoheterotrophy in photosynthetically competent plants. It was originally thought that mycoheterotrophy only occurs in specialist achlorophyllous plants that obtain all their C and N from mycorrhizal fungi (Leake 2004; Leake and Cameron 2010). More recently it was discovered that some green orchids are mixotrophs, gaining organic C and N via a combination of mycoheterotrophy and photosynthesis (Fig. 2; reviewed by Selosse and Roy 2009). Substantial fungus-mediated mixotrophy occurs in the photosynthesising ericaceous tribe Pyroleae (Tedersoo et al. 2007) and the orchid Cephalanthera damasonium (Julou et al. 2005) with an estimated 10–67% of carbon obtained from fungal partners. Selosse and Roy (2009) draw analogies between mixotrophy of plants and algae, and suggest that partial mycoheterotrophy may be more common in light-limited environments than currently acknowledged. The authors also highlight the dearth of knowledge of cellular mechanisms enabling the transfer of nutrients between fungi and partially mycoheterotrophic plants.

Analogous to mycoheterotrophy by photosynthetically competent plants, but without a fungal link, is hemiparasitism (Fig. 2; Marshall and Ehleringer 1990). Here a photosynthetically competent vascular (usually angiosperm) plant obtains all water and nutrient elements which cannot be supplied by its own photosynthesis from another (usually angiosperm) vascular plant (Raven 1983). Hemiparasites are generally linked exclusively to the host xylem and acquire N mostly as organic N as all root-acquired, including organic N, ammonium and N₂ fixed by root nodules or rhizothamnia, move in the xylem as organic N, although only nitrate and most P move as inorganic constituents (Raven 1983). Acquisition of organic C and N by hemiparasites from the xylem stream of the host involves the parallel acquisition of two or more atoms of organic C per atom N when dicarboxylic amino acid and their amides are the form in which N is moved, and one C per N (with not all C recoverable as organic C), when ureides such as allantoin are the main N transport molecules.

Mixotrophy is deemed the exception in terrestrial plants, excepting carnivorous, partial mycoheterotrophic or hemi-parasitic plants, but a recent study using ¹³C-labelled litter showed that Quercus petraea trees in north-eastern France are mixotrophic in spring (Bréda et al. 2013).

In recent decades it has been ‘rediscovered’ that nitrate and ammonium are not the sole N sources for plants; early plant nutrition researchers considered a surprising variety of organic compounds as nutrient sources (reviewed by Paungfoo-Lonhienne et al. 2012). Amino acids are present in soil, and plants use or even prefer amino acids over other N forms (Chapin et al. 1993; Chapin 1995; Näsholm et al. 1998; Schimel and Bennett 2004), and amino acid transporters catalysing the uptake of amino acids into roots have been identified (Hirner et al. 2006; Lee et al. 2007; Svennerstam et al. 2007, 2011). The difficulties associated with quantifying N conversions in soils, detecting simultaneous uptake and release of N by soil and symbiotic organisms and roots, and identifying the chemical form in which N is acquired by plants have prevented unequivocal assessment of the contribution of organic N to the plants’ N uptake (reviewed by Näsholm et al. 2009). Ecological and physiological studies have evaluated amino acids for their role as organic N and C sources, but debate continues as to whether these compounds are generated at sufficient rates in soil to be significant N sources, and whether plants can compete with microbes for their uptake. Larger organic N compounds in soils include peptides and proteins. Membrane transporters catalysing the uptake of di- and tri-peptides into roots have now been identified (Komarova et al. 2008). It has been suggested that use of peptides as N and C source contributes to the success of the endemic angiosperm Deschampsia antarctica with increasing temperatures on the Antarctic Peninsula (Hill et al. 2011). We found that green fluorescent protein enters root hairs and supplements inorganic N in low-N supplied Arabidopsis but have not identified the mechanisms that permit such uptake (Paungfoo-Lonhienne et al. 2008).

Experimental and technical innovations are advancing knowledge on organic nutrients and include microdialysis as tool to quantify N flux via passive diffusion (Inselsbacher and Näsholm 2012). Soil microdialysis showed that amino acids account for 80% of the soil N supply in undisturbed boreal forest soil, whereas conventional extraction techniques detected inorganic N as the dominant soil N form. Microdialysis and analysis of a larger range of soil organic compounds have great potential to improve knowledge of N and other nutrients (Inselsbacher et al. 2011; Warren 2013), as widely used soil extraction methods were established to identify fertiliser needs in agriculture focussed on inorganic nutrients. Promising techniques include high resolution research techniques of soil, roots and soil microorganisms including confocal and nano-SIMS imaging with fluorescent and stable isotope labelled N compounds (Paungfoo-Lonhienne et al. 2008; Clode et al. 2009; Whiteside et al. 2009; Kilburn et al. 2010).

Similar questions as for N arise about the role of organic P for plant nutrition. Inorganic P (orthophosphate, P_i) is considered the main source for plants and is acquired via membrane transporters (Mudge et al. 2002; Smith et al. 2003). Organic P forms are generally considered unavailable for plant uptake, but have been more recently investigated as sources for plants. As discussed above, exoenzymes cleaving phosphate esters have been described in several plants species, and the ability of axenically cultivated Arabidopsis and wheat to grow with nucleic acid and other organic P forms as sole P sources was thought to be due to exuded DNAases, phytases and phosphatases (Chen et al. 2000; Richardson et al. 2000). We showed that organic P consisting of 25-nucleotide DNA, that was protected from enzymatic degradation, entered root cells (Fig. 3; Paungfoo-Lonhienne et al. 2010a), indicating that organic P compounds can enter roots. So far there have been no reports that angiosperms can use the uptake system of some cyanobacteria and certain other bacteria which acquire and then cleave organic phosphonates (Dyhrman et al. 2006).

Fig. 3.  Proposed scheme of plant nutrient sources in which soil microbes and root-derived enzymes contribute to depolymerisation and mineralisation of organic matter, and plants acquire not only inorganic nutrients directly and organic nutrients indirectly via mycorrhizal symbionts, but also take up a suite of organic compounds and microbes (modified from Paungfoo-Lonhienne et al. 2012).

Adding to the uncertainties of the role of organic nutrients is the scant mechanistic understanding of how large organic compounds enter roots. The membrane-spanning proteins that catalyse selective passage of ions or molecules across the plasma membrane cannot transport large macromolecules such as proteins, polynucleotides or polysaccharides. Eukaryotic cells ingest macromolecules via endocytosis by progressively enclosing the substance in a section of the plasma membrane, followed by invagination and pinching off to form an intracellular vesicle containing the ingested substance (Alberts et al. 1989). Endocytosis retrieves membrane material and associated cargo from the plasma membrane for internal utilisation or destruction, and for processing and recycling to the plasma membrane (Robinson 2005). Endocytosis is common to all eukaryotes and enabled by cytoskeleton-forming proteins, but was thought not to occur in prokaryotes due to missing cellular components and thick and rigid cell walls. However, uptake of proteins by bacterium Gemmata obscuriglobus (Planctomycetes) occurs as an energy-dependent process analogous to eukaryotic endocytosis (Lonhienne et al. 2010), and some bacteria incorporate other bacteria as endosymbionts (McCutcheon and von Dohlen 2011).

Endocytosis serves multiple purposes in plants including membrane recycling, protein transport, cell-to-cell communication, cell signalling and cell wall morphogenesis (Holstein 2002; Šamaj et al. 2004; Šamaj et al. 2005; Geldner and Jurgens 2006). The role of endocytosis for nutrient uptake remains a controversial topic. In the 1960s pinocytosis, a type of endocytosis that involves invagination of the plasma membrane around a fluid phase and subsequent formation of a small versicle within the cytoplasm, was considered by some researchers as a main route of entry of nutrients into root cells (Jensen and McLaren 1960; Bradfute et al. 1964; Bradfute and McLaren 1964; reviewed by Cram 1980). Cram (1980) critically reviewed this hypothesis, pointing out major problems such as the required implausibly high density of selective binding sites in the membrane and the thermodynamic and mechanistic difficulties of creating a fluid-filled spaces by invagination of the plasmalemma in a turgid plant cell. The points raised by Cram (1980) were reiterated and developed by Raven (1987) and Robinson et al. (2008). However, FM dyes, the most reliable endocytosis tracer (Emans et al. 2002; Bolte et al. 2004), have unequivocally demonstrated that plant cells perform endocytosis (Emans et al. 2002) and contain endosomes (Geldner 2004; Voigt et al. 2005). Membrane internalisation in plant cells occurs principally via clathrin-coated vesicles similar to most animal cells (Pérez-Gómez and Moore 2007). Even though the analysis of the Arabidopsis genome has revealed considerable conservation of components of the endocytotic machinery among eukaryotes, endocytosis research in plants has lagged behind that of animals. A proposed reason has been the lack of obvious candidates for endocytotic cargo molecules (Robinson et al. 2008).

There is evidence for a vesicle-mediated transport system between vacuole and plasma membrane (Echeverria 2000). Internalised lucifer yellow, an endocytosis marker, occurs in transition zone cells of the inner cortex of intact maize root apices indicating that large molecules are transported across cortex cells (Baluska et al. 2004, who acknowledged that results from the use of lucifer yellow can be compromised by the occurrence of carrier-mediated movement of this tracer across plant membranes). That larger molecules enter roots via endocytosis has long been proposed as exogenously supplied macromolecules such as DNA were detected in roots (reviewed by Paungfoo-Lonhienne et al. 2012). Sucrose and glucose enter cells by membrane transporters (Sauer 2007) and are also incorporated into intact walled plant cells via endocyctosis (Etxeberria et al. 2005a, 2005b). Endocytosis and an additional unknown mechanism have been proposed to facilitate entry of peptides into walled plant cells (Eggenberger et al. 2011).

It was assumed that carnivorous plants first digest organic compounds in animal-catching organs such as pitchers into inorganic and smaller organic compounds (amino acids, peptides), before transport into plant cells by membrane transporters (Schulze et al. 1999). Recently, endocytosis was identified as a mechanism facilitating entry of proteins into tissues of carnivorous plants. Proteins are absorbed by endosomes which fuse with primary lysosomes which results in larger compartments where degradation takes place (Adlassnig et al. 2012). Endocytosis also plays an integral role in the endocytic uptake of Rhizobium bacteria in legume roots, involving internalisation into plant cells (reviewed by Brewin 2004).

Microbes occur in much higher density in the rhizosphere than bulk soil and are attracted by root-derived compounds and debris, and it is conceivable that plants acquire and digest microbes as sources of nutrients. We tested this possibility and presented the first evidence that root cells incorporate non-pathogenic and non-symbiotic microbes as nutrient sources (Fig. 3; Paungfoo-Lonhienne et al. 2010b). As microbes compete with plants for soil nutrients, uptake and digestion of microbes could be an adaptation for securing nutrients (Paungfoo-Lonhienne et al. 2010b).

It is possible that the effects of external organic compounds on root function are related to mixotrophy. For example, organic compounds that trigger increased root branching by increasing the number of root tips and extent of root hair production, may stimulate processes that rely on the presence of these structures especially nutrient uptake. Indeed, the observed effects of organic polymers on root morphogenesis include enhanced root branching and root hair growth (Fig. 4) and are congruent with the notion that organic polymers facilitate root proliferation in sites that are rich in organic matter (Paungfoo-Lonhienne et al. 2008; Paungfoo-Lonhienne et al. 2010a).

Fig. 4.  Addition of protein to nutrient-replete axenic growth medium enhances the thickness of lateral roots and length of root hairs. Wild-type Arabidopsis plants (Columbia 0) were grown for 17 days on 0.5× MS medium (Murashige and Skoog 1962) with or without protein added (330 μg bovine serum albumen mL^–1 growth medium; TGA Lonhienne, D Rentsch, T Näsholm, S Schmidt, C Paungfoo-Lonhienne, unpubl. data).

An extension of the root system allows exploration of soil for the continued absorption of inorganic and organic nutrients. The effect of inorganic nutrients on root morphology depends on the nutrient solute and its concentration. For example, supply of ammonium stimulates lateral root initiation and higher-order lateral root branching, while nitrate stimulates lateral root elongation (Forde and Walch-Liu 2009; Lima et al. 2010). Although there is a vast literature on the quantitative effects of inorganic nutrients on root growth and morphology (reviewed by Hermans et al. 2006), much less is known about the effects of organic forms on these characteristics. Amino acids and peptides alter root morphology and biomass, root length, thickness, surface area and root hairs (Walch-Liu et al. 2006a, 2006b; Cambui et al. 2011; Soper et al. 2011), providing indirect evidence that root structure and uptake of these compounds may be linked. Extension of the root system is particularly important for the acquisition of less mobile and less soluble nutrients such as most P-compounds (Bates and Lynch 1996). Root hairs have a large effect on the uptake of P-compounds as evidenced by morphologically different root hairs of crop species and varieties and their P-compound acquisition. Length and surface area of root hairs were significantly correlated with the depletion of P-compounds in the rhizosphere in low P growth condition (Bates and Lynch 1996; Gahoonia and Nielsen 1996; Gahoonia et al. 1997; Ma et al. 2001). Long root hairs are considered a beneficial trait for sustaining crop yields in low P soil (Gahoonia and Nielsen 2004). Although root hairs are under genetic control, their growth is modulated by environmental conditions. Increased root hair length under P-limitation is an adaptation that aids P-compound acquisition. Root colonisation with arbuscular mycorrhizal fungi is more pronounced in species and genotypes that have few or short root hairs (Smith and Read 2008). Root hair length increased in the presence of DNA in plants grown in P-replete medium (Paungfoo-Lonhienne et al. 2010a), pointing to a role of organic compounds as signalling agents. Similarly, phytate as a P source in axenic growth conditions increased root growth but dramatically decreased shoot growth (Richardson et al. 2000). The responsiveness of plants to organic compounds can be interpreted as a further indicator for their involvement in plant nutrient relations.

Root hairs are specialised epidermal cells and the outermost interface between roots and soil. Root hairs can constitute up to 70% of the root surface. They are tip-growing extensions originating from root epidermal cells, so-called trichoblasts, which play an important role in water and nutrient uptake (reviewed by Datta et al. 2011). Root hairs are the major site of inorganic P uptake (Pi, phosphate) (Gahoonia and Nielsen 1998) containing the high affinity phosphate transporters responsible for P_i uptake (Mudge et al. 2002; Schunmann et al. 2004), and transporters for ammonium and nitrate (Lauter et al. 1996).

Tip growth in plant cells is an actin-based process that requires targeted exocytosis and compensatory endocytosis to occur at the growth cone and involves polarised membrane trafficking and the presence of endosomal compartments at the tip of root hairs (Ovečka et al. 2005; Voigt et al. 2005). Endocytosis occurs at the tip of growing root hairs and subsequent trafficking of the incorporated membrane. Mercado-Blanco and Prieto (2012) speculated that this process facilitates passive entry of bacteria into root hairs. Root hairs of N₂ fixing species facilitate the endophytic colonisation of roots, and recent studies indicate that root hairs of non-N₂ fixing angiosperm incorporate microbes (Fig. 5d; Prieto et al. 2011; Mercado-Blanco and Prieto 2012). Our research indicates that uptake of protein, DNA and microbes appears to be restricted to root-hair-producing trichoblasts (Paungfoo-Lonhienne et al. 2008, 2010a, 2010b) (Fig. 5a–c). These observations support the argument that root hairs are one of the entry routes of large molecule into root cells.

Fig. 5.  Confocal laser scanning microscopy images of roots incubated with fluorescent protein, DNA and microbes. (a) Green fluorescent protein with Arabidopsis roots (Paungfoo-Lonhienne et al. 2008) (the copyright notice), (b) fluorescent-labelled DNA with Arabidopsis roots (Paungfoo-Lonhienne et al. 2010a) (www.plantphysiol.org, Copyright American Society of Plant Biologists, accessed 9 January 2013) and (c) yeast cells expressing GFP in roots of tomato (Paungfoo-Lonhienne et al. 2010b) (the copyright notice). (d) EGFP-tagged Pseudomonas putida PICP2 colonised olive root hairs (Mercado-Blanco and Prieto 2012) (reprinted from Plant and Soil, Vol. 361, p. 303, 2012 with permission of Springer).

Mixotrophy is currently considered an exception in higher plants, and restricted to carnivorous, hemi-parasitic and partially hetero-mycotrophic species. However, similar to other mixotrophic organisms, notably phytoplankton, plants seem to be able to complement photosynthetic energy gain with essential elements from organic compounds and microbes. Although the contribution of organic carbon is likely to be less important for plants that are not light-limited, organic N and P compounds could supplement essential elements acquired as inorganic compounds. Knowledge of whether mixotrophy is important for plants only in certain ecosystems or whether it is a more general feature of plants is lacking. Similar to the uncertainty surrounding organic C fluxes in the rhizosphere (Jones et al. 2009), we have little quantitative information on how much organic versus inorganic N and P compounds contribute to plant nutrition. In semi-controlled growth conditions, several studies have shown a preference of plants for organic N monomers over inorganic N (e.g. Stoelken et al. 2010). Similarly, in highly artificial axenic growth conditions, Arabidopsis satisfied ~10% of N demand with protein (Paungfoo-Lonhienne et al. 2008). A knowledge gap remains in our understanding of how microbes, especially procaryotes, act in the rhizosphere, apoplast and as root endophytes. The rapid advances in microbial research will shed light into these questions, and more powerful analyses techniques will decipher the chemical nature and fluxes of organic compounds at much finer resolution than what has been possible in the past.

Approaches that span from molecular to ecological techniques, the use of mutants, sophisticated microscopy and modelling can be used to address the questions raised here. We argue that a new framework that considers plants as mixotrophs has numerous benefits including advancing the sustainable use of soils and efficient nutrient supply to plants with nutrients contained in organic materials. This is pertinent as the use of fertilisers is accelerating, much of which is applied as inorganic substances (or in the case of urea is rapidly converted to inorganics), and nutrient inefficiencies characterise many modern crop systems (Tilman et al. 2002; Davidson et al. 2012). Illustrating the dramatic discrepancy in fertiliser use is food production in China that increased 3.4-fold from 1961 to 2009, but was accompanied by 37- and 91-fold increases in N and P fertiliser application (Zhang et al. 2013). The resultant pollution is a problem of global significance (Gruber and Galloway 2008), motivating research and development on improving nutrient supply, uptake and assimilation by plants.