Plant Nutrition: Root Transporters on the Move

The dynamics of nutrient transporters and channels emerges as a critical and highly regulated cellular process controlling nutrient uptake and distribution in plant roots.

Abstract

Nutrient and water uptake from the soil is essential for plant growth and development. In the root, absorption and radial transport of nutrients and water toward the vascular tissues is achieved by a battery of specialized transporters and channels. Modulating the amount and the localization of these membrane transport proteins appears as a way to drive their activity and is essential to maintain nutrient homeostasis in plants. This control first involves the delivery of newly synthesized proteins to the plasma membrane by establishing check points along the secretory pathway, especially during the export from the endoplasmic reticulum. Plasma membrane-localized transport proteins are internalized through endocytosis followed by recycling to the cell surface or targeting to the vacuole for degradation, hence constituting another layer of control. These intricate mechanisms are often regulated by nutrient availability, stresses, and endogenous cues, allowing plants to rapidly adjust to their environment and adapt their development.

Plants take up nutrients and water from the soil and transport them to the leaves to support photosynthesis and plant growth. However, most soils around the world do not provide optimal conditions for plant colonization. Consequently, plants have evolved sophisticated mechanisms to adjust to deficiency or excess of nutrients and water supply. Membrane transport proteins, including channels and transporters, play crucial roles in the uptake of nutrients and water from the soil and in their radial transport to the root vasculature. Newly synthesized membrane transport proteins have to be properly targeted to a defined compartment, usually the plasma membrane, to efficiently ensure their function. The trafficking of membrane transport proteins along the secretory pathway is tightly controlled and involves the recognition of exit signals by gatekeeper protein complexes. After reaching the plasma membrane, membrane transport proteins can be endocytosed and subsequently recycled to the cell surface or targeted to the vacuole for degradation. Because the subcellular localization of proteins directly influences their activity, modulating the localization of membrane transport proteins constitutes a powerful way to control nutrient and water uptake in plants. This review discusses the fundamental mechanisms at stake in membrane protein secretion and endocytosis, with a specific focus on membrane transport proteins, and how endogenous and exogenous cues affect their dynamics to integrate uptake of nutrients and water to plant growth conditions.

GENERAL PRINCIPLES OF MEMBRANE PROTEIN TRAFFICKING THROUGH THE SECRETORY AND ENDOCYTIC PATHWAYS

Transport between the Endoplasmic Reticulum and the Plasma Membrane

The endoplasmic reticulum (ER) is the departure point of the secretory pathway where synthesis, folding, disulfide bond formation, and oligomerization of the proteins take place. Misfolded proteins are selected by the endoplasmic reticulum-associated degradation (ERAD) machinery and retrotranslocated into the cytosol for ubiquitin/proteasome-mediated degradation (Guerra and Callis, 2012). On the other hand, functional proteins are exported from the ER using vesicles coated with coat protein complex II (COPII) that is composed of three cytosolic components: the GTPase Sar-1, Sec23/Sec24, and Sec13/Sec31 heteromers (Barlowe et al., 1994; Aridor et al., 2001). Recruitment of cargo proteins in COPII vesicles is mediated by Sec24. Thus far, multiple independent cargo binding sites recognizing diverse ER sorting signals, such as diacidic motifs, have been identified on Sec24 (Miller et al., 2003; Mossessova et al., 2003). Diacidic motifs correspond to the (D/E)x(D/E) sequence, where x represents any amino acid, and are functionally conserved from virus to animals (Nishimura and Balch, 1997; Zuzarte et al., 2007). Although the role of the COPII complex was mostly investigated in yeast (Saccharomyces cerevisiae) and mammals, several studies highlighted that the COPII machinery is conserved in plants and is responsible for protein export from the ER (Mikosch et al., 2006; Takagi et al., 2013).

In parallel to the anterograde route mediated by COPII, a retrograde pathway is operated by vesicles coated with COPI, which allow the continual recycling of proteins and lipids from the Golgi to the ER in order to maintain an equilibrium with COPII transport. Although COPI-mediated transport was mostly studied in yeast and mammals, the presence of homologs of COPI proteins in plants suggested that the function of this complex was conserved and COPI-containing vesicles were identified (Contreras et al., 2000; Pimpl et al., 2000).

The Golgi apparatus has a major role in sorting proteins toward other cellular compartments including the plasma membrane and the vacuole. In plants, the Golgi is involved in the synthesis and the assembly of complex polysaccharides of the cell wall and in the production of glycolipids for the plasma membrane and the tonoplast. The Golgi is composed of individual cisternae and is subdivided into three parts, the cis-, medial-, and trans-Golgi, which make this organelle polarized (Hwang and Robinson, 2009). Proteins coming from the ER arrive first in the cis-Golgi and then pass sequentially in the two other compartments, finally reaching the trans-Golgi network (TGN) that in plants exists as an independent organelle from the Golgi (Brandizzi and Barlowe, 2013). Vesicles coming from the TGN lastly fuse with the plasma membrane to deliver the cargo proteins.

The Endocytic Pathway from the Cell Surface to the Vacuole

Plasma membrane proteins can be internalized using clathrin-dependent or clathrin-independent pathways. Clathrin-mediated endocytosis (CME) initiates at specific foci of the plasma membrane called clathrin-coated pits by the recruitment of the heterotetrameric adaptor protein2 complex and the hexameric clathrin complex (Chen et al., 2011a; Baisa et al., 2013). Along with the formation of the clathrin cage, clathrin-coated pits mature into clathrin-coated vesicles that will be released from the plasma membrane. In plants, key components of the CME machinery are conserved, and our knowledge of CME mechanisms increased considerably in recent years (Chen et al., 2011a; Baisa et al., 2013). A crucial step in CME is the selection of cargo proteins in clathrin-coated vesicles via the recognition of specific sorting signals. Subunits of the Adaptor Protein2 complex select cargo proteins by recognizing di-Leu motifs [DE]xxx[LIM] (x is any amino acid) and Tyr-based motifs YxxΦ (Φ is a bulky hydrophobic amino acid; Traub, 2009). Ubiquitination, a posttranslational modification in which the 76-amino acid polypeptide ubiquitin is attached onto a Lys residue of a protein, is also recognized by the CME machinery and plays a crucial role in cargo internalization from the plasma membrane (Lauwers et al., 2010). Importantly, CME is not the only way membrane proteins are internalized. The Arabidopsis (Arabidopsis thaliana) membrane microdomain-associated protein Flotillin1 was recently associated with clathrin-independent endocytosis (Li et al., 2012), as previously demonstrated in mammals (Hansen and Nichols, 2009).

After internalization, membrane proteins are targeted to early endosomes (EEs). In plants, EEs have been demonstrated to coincide with the TGN in the early endosome/trans-Golgi network (EE/TGN; Dettmer et al., 2006) and may contain specialized subdomains with secretory or endocytic functions (Contento and Bassham, 2012). From the EE/TGN, membrane proteins can be recycled to the plasma membrane, which involves vesicle budding regulators named ADP ribosylation factor-guanine nucleotide exchange factor (ARF-GEF) proteins (e.g. GNOM) that are sensitive to the toxin brefeldin A (BFA; Geldner et al., 2003). By inhibiting ARF-GEF activity, BFA triggers the accumulation of endocytosed protein in large bodies in plant roots, making this drug a very interesting tool to study endocytosis and recycling. Alternatively, endocytosed transporters are targeted to late endosomes named multivesicular bodies (MVBs) that constitute an intermediate compartment before the vacuole where protein degradation occurs. Ubiquitination is known to play a critical role in MVB sorting (MacGurn et al., 2012). The endosomal sorting complex for transport (ESCRT), composed of four subcomplexes (ESCRT-0 to ESCRT-III), captures ubiquitinated cargos in the endosome membrane and allows their sorting in intraluminal vesicles of the MVB and subsequent vacuolar/lysosomal targeting (Henne et al., 2011). Most of the ESCRT proteins identified in yeast and mammals are conserved in plants, except a canonical ESCRT-0 subcomplex whose role is to initially recognize and concentrate ubiquitinated cargos (Leung et al., 2008). However, Arabidopsis proteins related to Target of Myb1 were recently proposed to compensate for the absence of ESCRT-0 by functioning as early gating factors for recognition and sorting of ubiquitinated cargos (Korbei et al., 2013). Interestingly, deubiquitination of cargos by the action of deubiquitinating enzymes was proposed to promote the recycling of ubiquitinated substrates at an early stage of the endosomal sorting process, and hence to rescue these cargos from degradation (Bomberger et al., 2009). At late steps of endocytosis, cargo proteins can be retrieved to earlier endocytic compartments by the retromer complex (for review, see Seaman, 2012). In mammals, the retromer is composed of two subcomplexes: the core retromer constituted by Vacuolar Protein Sorting (VPS)26, VPS29, and VPS35 and a dimer of Sorting Nexin (SNX; Seaman, 2012). In Arabidopsis, all members of the retromer are conserved; however, recent data suggest that the core retromer and the SNX dimer probably behave as independent units (Jaillais et al., 2007; Pourcher et al., 2010; Zelazny et al., 2013). Even if the subcellular localization of the plant retromer is still a matter of debate, this complex was shown to primarily localize in Ara7/RabF2b-positive MVBs (Jaillais et al., 2007).

SECRETION OF PLANT MEMBRANE TRANSPORT PROTEINS

Thus far, the mechanisms by which plant membrane transport proteins travel along the secretory pathway have been poorly studied. In this section, we present the knowledge gained over the past few years on membrane transport protein exit from the ER and targeting to the plasma membrane.

Export from the ER

The role of diacidic motifs in the ER export of membrane transport proteins in plants is well illustrated by the water channels aquaporins. Some Plasma membrane Intrinsic Protein2 (PIP2) proteins from Arabidopsis and maize (Zea mays) carry in their N-terminal part a DxE motif that has been demonstrated by mutation analyses to be essential for ER exit (Fig. 1; Zelazny et al., 2009; Sorieul et al., 2011). Extensive work in yeast and mammals demonstrated that diacidic motifs are recognized by the Sec24 protein from the COPII complex (Miller et al., 2003; Mossessova et al., 2003). This recognition seems to be conserved in plants because the Arabidopsis potassium channel KAT1 is recruited into COPII vesicles via binding of a diacidic motif to Sec24 (Sieben et al., 2008). Interestingly, in Arabidopsis roots, overexpression of AtPIP2;1 mutated in the diacidic export motif induced a strong reduction in the root hydraulic conductivity compared with wild-type plants (Sorieul et al., 2011). Because aquaporins form tetramers, it was proposed that the mutated AtPIP2;1 interacts with endogenous PIPs and induces ER retention of the complex, leading to a decrease in PIP levels at the cell surface. Heteromerization between PIP1 and PIP2 proteins, which constitute two groups of PIPs with different properties, appears as a way to control their ER export. When expressed alone in maize cells, ZmPIP1 proteins are retained in the ER, whereas ZmPIP2 proteins are targeted to the plasma membrane (Zelazny et al., 2007). However, upon heteromerization with ZmPIP2s, ZmPIP1s are efficiently addressed to the plasma membrane, suggesting that ZmPIP2s provide export signals that are sufficient to overcome the ER retention capacity of ZmPIP1s.

Figure 1.

Diagram illustrating the constitutive dynamics of PIP2;1 in Arabidopsis root cells. CIE, Clathrin-independent endocytosis; PM, plasma membrane.

Posttranslational modifications, such as ubiquitination, constitute important signals regulating the fate of membrane proteins. In Arabidopsis, overexpression of the ER resident E3 ubiquitin ligase Ring membrane-anchor1 H1 (Rma1H1) decreases PIP2;1 levels and inhibits its trafficking from the ER to the plasma membrane, concomitantly conferring enhanced tolerance to drought stress (Lee et al., 2009). Rma1H1 was shown to ubiquitinate PIP2;1, likely leading to its degradation by the proteasome. Because Rma1H1 expression is induced by stresses such as dehydration (Lee et al., 2009), the enhanced degradation of PIP2;1 at the ER level might result from an ERAD- degradation process of PIP2;1. Phosphorylation has been also demonstrated to control ER export of membrane transport proteins. Mutation of a C-terminal Ser to the phosphomimic Asp in the Arabidopsis Phosphate Transporter1;1 (PHT1;1) leads to ER retention, suggesting that phosphorylation prevents PHT1;1 exit from the ER in root cells (Fig. 2A; Bayle et al., 2011). Interestingly, phosphorylation plays an opposite effect in the trafficking of some Arabidopsis aquaporin. PIP2;1 is phosphorylated on Ser283 and mutation of this residue to Ala triggers a strong retention of PIP2;1 in the ER (Prak et al., 2008).

Figure 2.

Diagram illustrating the substrate-regulated trafficking pathways of PHT1 (A), IRT1 (B), and BOR1 (C) in Arabidopsis root cells. The putative secretory pathways of IRT1 (B) and BOR1 (C) are indicated by dashed oval structures. Putative pathways are indicated by dashed arrows. Note that in A, ubiquitination of PHT1 in the ER requires the ubiquitin-conjugating E2 phosphate2 enzyme (PHO2) but also an unknown E3 ubiquitin ligase. B, Boron; Co, cobalt; Mn, manganese; Zn, zinc; PM, plasma membrane.

In yeast, the first step in COPII vesicle formation required the Sec12 protein, a guanine nucleotide-exchange factor (Barlowe and Schekman, 1993). Interestingly, mutation of the Arabidopsis ER-localized and Sec12-related Phosphate Transporter Traffic Facilitator1 (PHF1) impairs ER exit of PHT1;1 and inorganic phosphate (Pi) uptake, without modifying the localization of other plasma membrane proteins. This argues for a specific action of PHF1 on Pi transporters (González et al., 2005). Interestingly, the importance of PHF1 in the ER exit of PHT1 proteins has also been demonstrated in rice (Oryza sativa), suggesting a conserved mechanism in plants (Chen et al., 2011b).

From the Golgi Apparatus to the Plasma Membrane

The mechanisms involved in the trafficking of membrane transport proteins between the Golgi apparatus and the plasma membrane are still poorly understood in plants. En route to the cell surface, Arabidopsis PHT1;1 and maize PIP2;5 have been demonstrated to partially colocalize with Golgi markers (Bayle et al., 2011; Hachez et al., 2013). The post-Golgi trafficking of ZmPIP2;5 was recently demonstrated to be regulated by the plasma membrane-localized soluble N-ethylmaleimide-sensitive-factor attachment protein receptor SYP121 known to regulate vesicular fusion (Besserer et al., 2012). SYP121 and ZmPIP2;5 physically interact and expression of the dominant-negative fragment of SYP121-Sp2 decreases the delivery of ZmPIP2;5 to the plasma membrane. As a result, ZmPIP2;5 water channel activity is negatively regulated by SYP121-Sp2 when expressed in protoplasts (Besserer et al., 2012). In addition, SYP121 regulates the delivery of plant potassium channels (Sutter et al., 2006). A fluorescent imaging-based genetic screen recently identified BFA-visualized exocytic trafficking defective5 (bex5) as a novel dominant mutant defective in the exocytosis of PIPs. BEX5 encodes the EE/TGN-localized RabA1b protein that plays a role in protein trafficking between the EE/TGN and the plasma membrane in Arabidopsis roots, probably by regulating vesicle formation (Feraru et al., 2012).

Although most of membrane transport proteins pass though the Golgi apparatus on their way to the plasma membrane, some are residents of the Golgi such as the Arabidopsis Metal Tolerance Protein11 (MTP11) manganese transporter (Peiter et al., 2007). MTP11 was proposed to allow the sequestration of manganese excess into vesicles that traffic to the plasma membrane to release manganese from the cell via exocytosis. Interestingly, the Arabidopsis Ca²⁺/Mn²⁺ pump endomembrane-type CA-ATPase3 has also been proposed to localize in the Golgi and to possibly transport calcium and manganese into this organelle (Mills et al., 2008). However, the localization of endomembrane-type CA-ATPase3 remains unclear because this protein was subsequently shown to localize in Ara7/RabF2b-positive late endosomes, where it might play a role in manganese detoxification (Li et al., 2008).

ENDOCYTOSIS AND DEGRADATION OF MEMBRANE TRANSPORT PROTEINS

Mechanisms of Internalization and Recycling from the EE/TGN

CME is the major mechanism of endocytosis in plants (Dhonukshe et al., 2007). Tyrphostin A23 (TyrA23), which impairs the recognition of YxxΦ motifs by the CME machinery, inhibits the endocytosis of the iron transporter IRT1, the ammonium transporter AMT1;3, and the aquaporin PIP2;1, indicating that these membrane transport proteins undergo CME (Dhonukshe et al., 2007; Barberon et al., 2011; Li et al., 2011; Wang et al., 2013). CME of PIP2;1 and IRT1 was confirmed by the use of a dominant-negative form of clathrin (Dhonukshe et al., 2007; Barberon et al., 2014). In root epidermal cells, IRT1 localizes in the EE/TGN but rapidly cycles with the cell surface (Fig. 2B), as revealed by TyrA23 treatment (Barberon et al., 2011). IRT1 transports not only iron, but also highly reactive manganese, zinc, and cobalt ions (Vert et al., 2002). Therefore, limiting the plasma membrane pool of IRT1 by an internalization/recycling process appears as a protective mechanism to limit the absorption of potentially toxic substrates (Zelazny et al., 2011).

Ubiquitination emerges as an essential signal driving transporter endocytosis (Lauwers et al., 2010). In plants, the dynamics of IRT1 between the cell surface and the EE/TGN is dependent on its multi-monoubiquitination on cytosol-exposed Lys residues (Barberon et al., 2011), a process likely mediated by the E3 ubiquitin ligase IRT1 DEGRADATION FACTOR1 (Shin et al., 2013). The ubiquitination-defective IRT1_K154RK179R mutant version accumulates at the plasma membrane (Barberon et al., 2011) and recent evidence point to a defect in internalization rather than to a mis-sorting in the MVB and subsequent recycling to the plasma membrane (M. Barberon and G. Vert, unpublished data). Transgenic plants expressing nonubiquitinatable IRT1_K154RK179R show severe growth defects due to uncontrolled metal uptake (Barberon et al., 2011, 2014). Ubiquitination of phosphate transporters has also been proposed to control their internalization from the plasma membrane. The E3 ubiquitin ligase NITROGEN LIMITATION ADAPTATION (NLA) is localized at the plasma membrane and was demonstrated to mediate the ubiquitination of PHT1;1, likely inducing its CME (Lin et al., 2013). Mutation of NLA causes high Pi accumulation in plants due to increased PHT1 transporter levels, highlighting the importance of ubiquitin-mediated endocytosis in the regulation of phosphate uptake.

Thus far, plant membrane transport proteins have been mainly demonstrated to be internalized using CME; however, the importance of the clathrin-independent pathways is only emerging. Single-molecule analysis indeed demonstrated that PIP2;1 is internalized using both a clathrin-dependent and a membrane raft-associated pathway involving Flotillin1 (Li et al., 2011).

Recycling of membrane transport proteins from the EE/TGN to the cell surface appears as a widespread phenomenon in plants, as illustrated by the accumulation in BFA bodies of transporters such as PHT1;1, IRT1, and the boron transporters BOR1 and BOR2, or the water channel PIP2;1 (Takano et al., 2005; Dhonukshe et al., 2007; Barberon et al., 2011; Bayle et al., 2011; Miwa et al., 2013). However, the identity of the ARF-GEFs required for the recycling of these membrane transport proteins remains an open question.

Sorting at the MVB: Vacuolar Targeting or Recycling

Ubiquitination constitutes an essential signal to sort endocytosed membrane proteins in the MVB and to trigger vacuolar targeting. The boron transporter BOR1 is mono- or diubiquitinated, likely on Lys590, and mutation of this residue to Ala blocks the vacuolar degradation of BOR1 (Kasai et al., 2011). Pharmacological approaches showed that ubiquitination is not involved in BOR1 internalization from the plasma membrane but is rather essential for sorting in the MVB (Fig. 2C; Kasai et al., 2011). How BOR1 is selected in the MVB remains to be determined; however, this mechanism probably involves the ESCRT complex (Henne et al., 2011). Similarly to BOR1, the iron transporter IRT1 is targeted to the vacuole for degradation (Barberon et al., 2011). Monoubiquitination on K154 and K179 likely controls the IRT1 internalization step; however, ubiquitination of other Lys residues or polyubiquitin chain extensions may control the sorting in MVBs of IRT1, en route to the vacuole.

In Arabidopsis, the putative retromer component SNX1 has been implicated in IRT1 recycling (Ivanov et al., 2014). Mutation of SNX1 leads to reduced iron import efficiency in roots and correlates with an enhanced degradation of IRT1. In addition, when IRT1 and SNX1 were transiently expressed in tobacco (Nicotiana tabacum) leaves, IRT1 and SNX1 partially colocalized in endosomes. SNX1 therefore appears to drive the recycling of internalized IRT1 to prevent its premature degradation (Ivanov et al., 2014). Whether other transporters found in SNX1-positive endosomes, such as PHT1;1 (Bayle et al., 2011), are recycled after retromer recognition remains an open question.

REGULATION OF MEMBRANE TRANSPORT PROTEIN TRAFFICKING BY ENVIRONMENTAL AND DEVELOPMENTAL CUES

Nutrients are essential for plant growth and development but are also toxic when present in excess. Therefore, the regulation of the dynamics of nutrient transporters by their substrates constitutes an elegant way to rapidly adjust to nutrient availability. As mentioned above, phosphorylation of PHT1;1 on a specific Ser residue prevents exit from the ER (Bayle et al., 2011). Interestingly, phosphoproteomic analysis revealed that Ser phosphorylations at the C terminus of PHT1;1 were less frequent when plants were grown under phosphate starvation. The authors proposed that phosphorylation of Ser residues in PHT1;1 impairs the recognition of a proximal ER export motif, thus preventing PHT1;1 exit from the ER when the internal level of phosphate is high (Bayle et al., 2011). Under Pi-sufficient conditions, ubiquitination of PHT1s at the ER and post-ER levels also appears as a way to regulate Pi uptake (Huang et al., 2013). In the presence of Pi, the ubiquitin-conjugating E2 enzyme PHO2 accumulates and participates in PHT1 ubiquitination, and consequently enhances the degradation of these transporters (Huang et al., 2013). However, whether PHT1s are degraded by an ERAD process involving the proteasome or in the vacuole remains unclear. Indeed, contradictory results have been obtained concerning the sensitivity of PHT1 degradation to MG132, a well-known inhibitor of the proteasome (Huang et al., 2013; Park et al., 2014). In addition to controlling PHT1 exit from the ER, phosphate nutrition modulates PHT1 endocytosis. Under Pi-sufficient conditions, PHT1;1 undergoes ubiquitin-dependent and NLA-mediated endocytosis and degradation in the vacuole (Lin et al., 2013). Interestingly, NLA expression itself is posttranscriptionally repressed by microRNA827 that is induced by Pi deprivation, adding another level of regulation by phosphate nutrition.

Boron-induced endocytosis of Arabidopsis BOR1 also represents a good example of substrate-induced endocytosis. Under boron limitation, BOR1 is localized at the plasma membrane, where it allows boron radial transport to vascular tissues. Upon higher boron availability, BOR1 is rapidly endocytosed and targeted to the vacuole for degradation (Takano et al., 2005). The Tyr-based motifs located in the large loop of BOR1 are not required for internalization from the plasma membrane under high boron, but rather allow the recruitment of BOR1 toward an EE/TGN subcompartment that becomes/fuses with the MVBs (Takano et al., 2010). In addition, boron excess induces BOR1 ubiquitination, which in turn accelerates MVB sorting and vacuolar targeting (Kasai et al., 2011).

The endocytosis of the broad spectrum metal transporter IRT1 is not regulated by the availability of its primary substrate iron (Barberon et al., 2011). However, the non-iron metal substrates of IRT1 (zinc, manganese, and cobalt) were very recently shown to regulate IRT1 dynamics between the EE/TGN and the cell surface (Barberon et al., 2014). Upon depletion of zinc, manganese, and cobalt, IRT1 relocalizes from the EE/TGN to the outer polar plasma membrane domain facing the rhizosphere. Interestingly, the nonubiquitinable IRT1_K154RK179R was found at the plasma membrane in the presence of metals, providing evidence that the response to secondary substrates of IRT1 may be mediated by ubiquitination (Barberon et al., 2014). In Arabidopsis root cells, the early effect of ammonium excess on the dynamics of the ammonium transporter AMT1;3 at the plasma membrane was recently investigated by using single-particle approaches (Wang et al., 2013). Upon ammonium excess, AMT1;3 proteins are amassed into clusters before being internalized mainly by CME. Whether this cluster formation of transporters in response to nutrient excess represents a general mechanism remains to be clarified in the future.

The relocalization of membrane transport proteins in response to environmental changes allows plants to adjust growth and development in their ever-changing environment. In Arabidopsis, salt stress triggers a strong inhibition of the root hydraulic conductivity and concomitantly induces the accumulation of PIP2;1 in intracellular vesicular structures in root cells (Boursiac et al., 2005, 2008). Single-molecule analysis demonstrated that under salt stress, the internalization of PIP2;1 from the plasma membrane involving a clathrin-independent membrane raft-associated pathway was enhanced (Li et al., 2011). Fluorescence recovery after photobleaching approaches revealed that salt treatment also enhanced the cycling of PIP2;1 in Arabidopsis roots (Luu et al., 2012). The relocalization mechanism of plant aquaporins is not restricted to PIPs, because Tonoplast Intrinsic Proteins1;1 were relocalized into putative intravacuolar invaginations in response to sodium chloride treatment (Boursiac et al., 2005).

Internalization is also controlled by endogenous cues and serves developmental programs and adaptation to environmental conditions. Notably, CME is regulated by several plant hormones, likely reflecting extensive cross talk between different pathways involved in sorting decisions. Salicylic acid and auxin were found to repress the endocytosis of different cell surface proteins including the aquaporin PIP2;1 (Paciorek et al., 2005; Du et al., 2013).

POLARIZATION OF MEMBRANE TRANSPORT PROTEINS IN PLANT ROOTS

To properly ensure their physiological function in root cells, some membrane transport proteins must be polarly localized. Several nutrient transporters have been shown to display a lateral polarity in root cells, with a specific enrichment at the outer or inner plasma membrane domains. In Arabidopsis, boron uptake from the soil is performed by the boric acid channel Nodulin26-like Intrinsic Protein5;1 (NIP5;1) that is localized at the outer domain of plasma membrane in root epidermal cells (Takano et al., 2010). By contrast, the borate exporter BOR1 localizes to the inner domain of root endodermis cells but also to other cell types (Takano et al., 2010). The opposite polarity of NIP5;1 and BOR1 is thought to facilitate the transcellular transport of boron from the rhizosphere to the vascular tissue of the root. Similarly, the boric acid channel OsNIP3;1 and OsBOR1, a close ortholog of AtBOR1, might allow boron transport across the root in rice (Fuji et al., 2009). Interestingly, Arabidopsis BOR4 is a borate exporter that is found at the outer domain of the plasma membrane in root epidermal cells and that mediates the extrusion of boron from the root, likely to avoid boron toxicity (Miwa et al., 2007). In rice, cell polarity is also essential for silicon transport across the root. This mechanism implies the Low silicon rice1 silicon influx channel and the Low silicon rice2 silicon exporter that are localized to the outer plasma membrane domain and the inner plasma membrane domain of the same root cells, respectively (Ma et al., 2006, 2007). In the absence of its secondary substrates, the iron transporter IRT1 is localized at the outer polar domain of the plasma membrane of root epidermal cells (Barberon et al., 2014). This polar localization of IRT1 is likely critical for proper radial transport of iron and metals in the root; however whether a polarly localized metal transporter is found at the inner plasma membrane domain of root epidermal cells is still unknown.

The mechanisms driving the polar localization of membrane proteins have been mostly described for PIN-FORMED (PIN) proteins (for review, see Dettmer and Friml, 2011; Luschnig and Vert, 2014). However, nutrient transporters now emerge as very interesting models to study cell polarity in plants, and its functional outcome. Mutation of Tyr-based motifs in BOR1 leads to loss of polarity, showing that these motifs are important to maintain BOR1 polarity (Takano et al., 2010). Similar to PIN proteins that are secreted nonpolarly and then targeted to polar domains by endocytosis and recycling (Dhonukshe et al., 2008), the authors proposed that BOR1 polarity would be, at least in part, determined by a recycling-based mechanism requiring the Tyr-based motifs. Moreover, the lateral diffusion of BOR1 in the plasma membrane was sufficiently slow to allow polar localization by such a process (Takano et al., 2010). In contrast with BOR1, IRT1 polarity is not established by a recycling process involving CME because IRT1 polar localization is unchanged in presence of a dominant-negative form of clathrin (Barberon et al., 2014). Moreover, IRT1 polarity is independent of ubiquitination because the nonubiquitinable IRT1_K154RK179R form is still localized to the outer plasma membrane domain. Interestingly, IRT1 interacts with the endosome-recruited FYVE1 protein that controls its localization and polarity (Barberon et al., 2014). Plants overexpressing FYVE1 accumulate IRT1 at the cell surface in a nonpolar fashion. Surprisingly, such plants show hypersensitivity to low iron and decreased metal uptake. This contrasts greatly with the scenario in which the loss of IRT1 ubiquitination triggers constitutive outer plasma membrane domain localization and correlates with metal overaccumulation (Barberon et al., 2011). This demonstrates that polarization of IRT1 is critical for proper radial transport of iron in roots and suggests that iron exits the root epidermis using efflux transporters rather than plasmodesmata. This is consistent with old observations that root epidermal cells are symplasmically isolated (Duckett et al., 1994). In contrast with BOR1 and IRT1 that require recycling mechanisms to be polarly localized, BOR4 polarity was proposed to be directly achieved by a polar secretion (Langowski et al., 2010). Similarly, Medicago truncatula phosphate transporter MtPT4, which is essential for symbiotic phosphate transport and for maintenance of the symbiosis, is polarly targeted by a mechanism that involves transient changes in the secretory system of the colonized root cells (Pumplin et al., 2012).

CONCLUSION

The mechanisms of plant plasma membrane protein trafficking are heavily influenced by studies on the auxin efflux carrier PINs. Nutrient transporters and water channels recently emerged as very interesting models to further study plant membrane protein dynamics. Altogether, they offer the possibility to grasp the mechanisms driving plant membrane protein trafficking and how developmental and environmental cues affect their subcellular dynamics to adjust to ever-changing growth conditions. A better understanding of the regulatory mechanisms controlling nutrient transporters and channels may provide biotechnological targets for crop improvement to rationalize fertilizer usage and to achieve better use of arable land. The knowledge gained over the past decade on plant membrane transport proteins brought to light the similarities and differences in membrane protein sorting strategies and therefore offered an evolutionary perspective on membrane protein trafficking. Plant membrane protein trafficking, which greatly suffered over the years from excessive homology-based comparisons with yeast and mammalian cells, can now be considered as an exciting field of research by itself.

Glossary

ER

endoplasmic reticulum

ERAD

endoplasmic reticulum-associated degradation

COPII

coat protein complex II

COPI

coat protein complex I

TGN

trans-Golgi network

CME

clathrin-mediated endocytosis

EE

early endosome

EE/TGN

early endosome/trans-Golgi network

BFA

brefeldin A

MVB

multivesicular body

ESCRT

endosomal sorting complex for transport

Pi

inorganic phosphate