Substrate (un)specificity of Arabidopsis NRT1/PTR FAMILY (NPF) proteins

Abstract

The conventional approach to categorizing transporters has been to class them according to their sequence homology, defining a ‘family’ (or a ‘superfamily’ if they are numerous), and according to their substrate specificity or selectivity. This general view is still relevant for some transporters, but it is being increasingly challenged. Here, we take the NRT1/PTR FAMILY (NPF) as one such example. NPF members do indeed display sequence and structural homologies with peptide transporter (PTR) proteins involved in the uptake of di- and tri-peptides in most other organisms. And in plants they were initially characterized as nitrate or peptide transporters. However, in recent years several other substrates have been identified, namely nitrite, chloride, glucosinolates, auxin (IAA), abscisic acid (ABA), jasmonates (JAs), and gibberellins (GAs). Some of these transporters are even capable of transporting more than one different substrate (e.g. nitrate/auxin, nitrate/ABA, nitrate/glucosinolates, or GA/JA). In this review, we give an overview of the substrate-specificity of the Arabidopsis NPF.

Introduction

The 53 NPF (NRT1/PTR FAMILY) members in Arabidopsis belong to the large peptide transporter (PTR) family (Léran et al., 2014). PTR transporters are thought to be present in all organisms (Daniel, 2004). In animals, they belong to the SLC15 family (Hediger et al., 2004), and are di-/tri-peptide transporters that are able to transport peptide-like drugs such as beta-lactam antibiotics (Meredith, 2009). The two most studied members, PepT1 (SCLC15A1) and PepT2 (SCLC15A2), have been cloned by functional expression in Xenopus oocytes. They constitute the major route for absorption of dietary nitrogen (Daniel et al., 2006). PTRs have 12 transmembrane segments, and structure–function studies have allowed the mapping of the amino acids important for the function of the transporters (Meredith and Price, 2006). Crystal structures of two bacterial PTRs (PepT_St and PepT_So) have been published (Newstead et al., 2011; Solcan et al., 2012; Doki et al., 2013). Recently, the crystal structure of a plant NPF, AtNPF6.3/NRT1.1/CHL1, has been determined by two independent research groups (Parker and Newstead, 2014; Sun et al., 2014; Tsay, 2014; Sun and Zheng, 2015), demonstrating that the overall folding of this plant protein is reminiscent of the other bacterial PTRs (Tsay, 2014; Sun and Zheng, 2015).

At the functional level, since the characterization of the first plant member NPF6.3/NRT1.1/CHL1 by Yi-Fang Tsay in 1993 (Tsay et al., 1993), a very active research community has resulted in the identification of a substrate for 32 Arabidopsis NPF members. In contrast to their animal and bacterial counterparts, the plant NPF proteins can transport a huge variety of substrates, including dipeptides, nitrate, nitrite, chloride, glucosinolates, and amino-acids, as well as several plant hormones including auxin (IAA), abscisic acid (ABA), jasmonates (JAs), and/or gibberellins (GAs).

Because of the diversity of substrates discovered in recent years, we believe this is a good time to summarise and describe the different functional studies that have aimed to identify the physiological substrates of NPF in Arabidopsis (Table 1).

Different systems have been used to identify NPF substrates

The two systems of choice that have been extensively used to characterize NPF members are Xenopus oocytes and the yeast Saccharomyces cerevisiae. Functional characterization in Xenopus oocytes has been done using electrophysiological approaches (two-electrode voltage clamp) or substrate quantification using labeled or non-labeled molecules. Three kinds of labeled substrates have been used: radioactive isotopes (³H, ¹⁴C), non-radioactive isotopes (²H, ¹³C, ¹⁵N), or fluorescent-tagged molecules. In yeast, the first approach used was functional complementation of mutant yeast deficient for peptide transporters (Steiner et al., 1994). Accumulation studies have also been performed in this system (e.g. Rentsch et al., 1995). Another elegant and very powerful indirect approach has been developed by the group of Mitsunori Seo using reporters to provide evidence of hormone transport (Kanno et al., 2012; Chiba et al., 2015). In order to identify new ABA transporters, they developed a screen based on a yeast strain expressing the first steps of the ABA signal transduction pathway split as a two-hybrid system: a PYR/PYL/RCAR ABA-receptor is fused to GAL4-DNA-binding domain and a type 2C protein phosphatase is fused to the GAL4 activation domain. Expression of a functional transporter of ABA allows the entry of external ABA that promotes the binding of PYR–ABA to the PP2C and restores the GAL4 transcription factor. Initially designed to identify ABA transporters (Kanno et al., 2012), this method was adapted to screen for GA and JA-Ile (jasmonoyl-isoleucine, the bioactive form of JA) transporters (Chiba et al., 2015; David et al., 2016). The same Japanese group has also performed accumulation studies of hormones in insect cell lines. Finally, two transporters were characterized using plasma membrane proteoliposomes from Arabidopsis or from NPF2.3-enriched Lactococcus lacti membranes (Segonzac et al., 2007; Taochy et al., 2015).

Nitrate transport by 17 NPF members

More than 20 years ago, molecular characterization of the chl1-5 mutant was the starting point for the functional characterization of the NPF. Indeed, Yi-Fang Tsay (Tsay et al., 1993) used Xenopus oocyte to elegantly demonstrate that NPF6.3/NRT1.1/CHL1 is a bona fide nitrate transporter (see below). In the following years, the Tsay laboratory was the leading group involved in the characterization of more than half of the NPF nitrate transporters. The nitrate transporter capacity of different NPF members has been studied in Xenopus oocytes and, so far, 17 have been demonstrated to be able to transport nitrate. The physiological role of these transporters has already been described (Krapp et al., 2014; Krapp, 2015; Noguero and Lacombe, 2016; O’Brien et al., 2016).

Most of the NPF nitrate transporters have been studied using two-electrode voltage clamp of Xenopus oocytes, namely: NPF1.1 and 1.2 (Hsu and Tsay, 2013), 2.9 (Wang and Tsay, 2011), 2.10 and 2.11 (Nour-Eldin et al., 2012), 2.12 (Almagro et al., 2008), 2.13 (Fan et al., 2009), 3.1 (Pike et al., 2014), 4.6 (Huang et al., 1999), 6.2 (Chiu et al., 2004), 6.3 (Tsay et al., 1993), 7.2 (Li et al., 2010), and 7.3 (Lin et al., 2008). All these transporters display the same behavior: in acidic external conditions, addition of NO₃^– induces a negative current at negative membrane potentials. This behavior is proof of an influx of positive charges, demonstrating that these transporters are symporters, the positive charge being H⁺. In these conditions, the stoichiometry has not been precisely determined but more H⁺ than NO₃^– should be transported to explain the current properties.

Another widely used technique allowing both influx and efflux characterization employs labeling with the ¹⁵N stable isotope of nitrogen in NO₃^– accumulation studies; although several radioactive N isotopes exist their half-lives are very short (the longest one, ¹³N, is about 10 min) and so they are very difficult to manage. In most biological tissues, ¹⁵N is present at 0.37% (the rest being ¹⁴N). The ¹⁵N can be easily quantified using mass spectrometric analysis.

Using one of the two techniques described above, low-affinity transporters have been characterized with a K_m in the millimolar range between 2–12 mM.

As mentioned previously, the first functionally characterized NPF member was NPF6.3/NRT1.1/CHL1 (Tsay et al., 1993). At the functional level, it is one of the most intriguing ones. Expression studies in Xenopus oocytes demonstrated that it behaves as a NO₃^– transporter activated by external H⁺. Later, it was proved that it functions as a dual-affinity transporter (Liu et al., 1999; Liu and Tsay, 2003) with two K_m, one in the millimolar range (around 5 mM) and one close to 50 µM, the transition between high- and low-affinity mode being controlled by phosphorylation of the T101 residue by CIPK23 (Liu and Tsay, 2003; Ho et al., 2009). CIPK23 is activated by CBL9 (Ho et al., 2009) and CBL1 (Léran et al., 2015). CIPK23/CBL is able to inhibit low-affinity influx of NO₃^– through NPF6.3, as well as efflux (Ho et al., 2009; Léran et al., 2013, 2015). At high external NO₃^– concentration, the negative effect of CIPK23 can be counteracted by the protein phosphatase PP2C, ABI2 (Leran et al., 2015).

Within the 17 NPF nitrate transporters, 12 are influx transporters, two are efflux transporters (NPF2.3 and NPF2.7), and three are influx/efflux transporters (NPF6.2, NPF6.3, and NPF7.3). Bidirectionality should be tested for the ‘influx’ and ‘efflux’ transporters. Several NPF NO₃^– transporters are multisubstrate transporters, being also able to transport another substrate such as glucosinolates, ABA, auxin, or nitrite.

Nitrite transport by one NPF member

Nitrite is the first metabolite of nitrate assimilation produced by the nitrate reductase in the cytosol and is transported into chloroplasts to prevent toxic accumulation. A study of a NPF3.1 knock-out has suggested a role in nitrite transport (Sugiura et al., 2007). NPF3.1 expression has been observed in smaller veins in vascular tissues and a role in nitrite transport has been demonstrated by expression in Xenopus oocytes and by two-electrode voltage clamp (Pike et al., 2014). In this system, NPF3.1, the only Arabidopsis member of the NPF3 subfamily, behaves as a low-affinity nitrite (and nitrate) transporter (K_m>5 mM). This property is shared with VvNPF3.2 from Vitis vinifera (Pike et al., 2014).

Chloride transport by one NPF member

NPF2.4 is a plasmamembrane-localized protein (Li et al., 2016). Its expression in the root stele and the phenotype of NPF2.4 overexpression lines and of artificial microRNA plants designed to knock-down NPF2.4 expression demonstrate its role in the root-to-shoot transfer of chloride (Li et al., 2016). This role is further supported by expression in the Xenopus oocyte combined with two-electrode voltage clamping and accumulation of ³⁶Cl^–. In NPF2.4-expressing oocytes, a hyperpolarization-induced external chloride-induced and Na/K-induced chloride efflux is recorded; whereas accumulation studies demonstrate the ability of NPF2.4 to perform chloride influx (Li et al., 2016). These different properties need to be studied in more detail to strengthen the determination of the chloride selectivity of this NPF member. A role of NPF2.4 in nitrate efflux can not currently be excluded.

Transport of peptides and amino acids

Di- and tri-peptides are the major nitrogen source for a lot of organisms. In bacteria and mammals, the uptake of these short peptides is performed by homologs of NPF members (see Introduction). Shortly after the identification of NPF6.3 as a nitrate transporter, NPF8.3 was identified as a di-peptide transporter (Rentsch et al., 1995). Using complementation of yeast strains deficient for peptide uptake, four di- and tri-peptide transporters from the NPF family were identified: NPF8.1, NPF8.2, NPF8.3, and NPF5.2. In-depth electrophysiological characterization of three of these transporters has been performed in Xenopus oocytes using two-electrode voltage clamp, for NPF8.1, 8.2 and 8.3 (Chiang et al., 2004; Hammes et al., 2010). They are H⁺-coupled, voltage-dependent acid-activated transporters.

NPF8.3, previously known as PTR2 or NTR1, was initially described as a low-affinity histidine transporter (Frommer et al., 1994), but this property was later determined to be a side activity of the transporter (Rentsch et al., 1995).

Glucosinolate transport by four NPF members

In Arabidopsis, glucosinolate derivatives are a group of nitrogen- and sulfur-rich secondary metabolites that are well known as major plant defense compounds against herbivores. They also have a wide range of biological functions and properties that make them useful for humans, including antipathogenic and carcinopreventive properties, and they are also used as flavor components. They accumulate in all parts of the plant, and long-distance transport from senescing leaves to flowers, seeds, and fruit has been well described (Jørgensen et al., 2015a).

A screen for uptake of 4-methylthiobutyl glucosinolate (4MTB) in Xenopus oocytes led to the identification of the first glucosinolate transporter, NPF2.10/GTR1. Following on from this, other members of NPF subfamily 2 were tested individually in Xenopus oocytes. NPF2.11, and to a lesser extent NPF2.9, NPF2.13, and NPF2.14 (but it is not yet certain that the npf2.14 gene is expressed in Arabidopsis Col-0), are also able to accumulate 4MTB (Nour-Eldin et al., 2012). NPF2.10 and NPF2.11 have been studied in detail in Xenopus oocytes using two-electrode voltage clamp. Both transporters are high-affinity H⁺/glucosinolates symporters that are also able to transport 8-methylthiooctyl glucosinolates, 8MTO (Andersen et al., 2013).

Recently, the highly conserved E₁X₁X₂E₂K motif shared by POT/PTR family members across kingdoms has been demonstrated to be essential for accumulation of glucosinolates by NPF2.11 into Xenopus oocytes (Jørgensen et al., 2015b).

Auxin

NPF6.3 expressed in Xenopus oocytes is able to transport IAA (Krouk et al., 2010) and 2,4-D (Bouguyon et al., 2015). This substrate is also transported by NPF6.3 expressed in other systems, namely in yeast and in BY-2 cells (Krouk et al., 2010; Bouguyon et al., 2015). Interaction between both substrates (nitrate and auxin) has been studied and explains at least part of the role of NPF6.3 as a nitrate sensor involved in nitrate-dependent lateral root developement (Krouk et al., 2010, 2011; Bouguyon et al., 2015; Krouk, 2016). Whereas IAA transport is inhibited at external nitrate concentration >0.5 mM, nitrate transport is not sensitive to external IAA up to 5 microM (Krouk et al., 2010).

Abscisic acid

Abscisic acid (ABA) is involved in a wide range of physiological and developmental processes including seed dormancy, germination, cell division, and elongation, as well as adaptation to biotic and abiotic stresses such as drought, salinity, and cold (Cutler et al., 2010).

While evidence for long-distance transport of ABA has been demonstrated for many years, the molecular basis of this transport was only determined in 2010 with the identification of the first ABA transporters belonging to the ABC transporter family (Kang et al., 2010; Kuromori et al., 2010; Boursiac et al., 2013). Shortly after this, using an ABA-dependent two-hybrid screening system (see Introduction), Kanno et al. (2012) identified four NPF members displaying ABA transport activity: NPF4.6/AIT1/NRT1.2, NPF4.5/AIT2, NPF4.1/AIT3, and NPF4.2/AIT4. NPF4.6 and NPF4.1 were the more effective in inducing β-gal activity. Their ABA transporting activities were therefore tested and confirmed by ABA accumulation studies in yeast cells (Kanno et al., 2012). NPF4.6 is expressed around vascular tissues, and mutants defective in AtNPF4.6 have lower surface temperatures due to more open stomata than the wild-type, indicating that the protein functions as an ABA transporter in vivo (Kanno et al., 2012). Since NPF4.6 is also a nitrate transporter (Huang et al., 1999), the effect of nitrate on ABA accumulation has been tested (Kanno et al., 2013) but interaction between the two substrates has not been demonstrated.

Using the same ABA-dependent two-hybrid system and screening 45 out of the 53 Arabidopsis NPF members, Chiba et al. (2015) confirmed that NPF4.6, NPF4.1, and NPF4.5 – but also additional NPF members such as NPF1.1, NPF2.5, NPF5.1, NPF5.2, NPF5.3, NPF5.7, and NPF8.2 – are able to transport ABA. Recently, Tal et al. (2016) have demonstrated the ability of NPF3.1-expressing oocytes to accumulate ABA.

Gibberellins

Gibberellins (GAs) are a family of plant hormones involved in many developmental processes such as root and shoot elongation, dormancy, and fruit and seed development. In Arabidopsis, among the large family of GA compounds, the major bioactive members are GA₁ and GA₄, and recent studies have shown that the predominant form subject to long-distance transport is GA₁₂, an inactive precursor of GA₁ and GA₄ (Regnault et al., 2015).

First evidence of GA transport activity by NPF members came with the characterization of the ABA transporter NPF4.1, which has also the ability to transport GA₃in vitro in yeast (Kanno et al., 2012). Then, using a modified yeast two-hybrid system with GA-specific receptor complexes to detect GA₁, GA₃, or GA₄ transport activity, Chiba et al. (2015) identified 18 among 45 Arabidopsis NPF members that were able to transport GA. Twelve NPF members transport GA₁, GA₃, and GA₄; three members transport only GA₁ and GA₃; and three NPF transport only GA₁ and GA₄ (Table 1).

Uptake experiments in Xenopus oocytes have shown that NPF2.10 is able to transport GA₃ but that no significant import of GA₁, GA₄, GA₉, or GA₂₀ could be detected. In addition, application of exogenous GA₃ (but also GA₁ and GA₄) fully restores the reduced-fertility phenotype due to impaired stamen development of a NPF2.10 knock-out mutant, suggesting that NPF2.10 may play a crucial role by providing GA required for stamen development (Saito et al., 2015).

Recently, in a screen for Arabidopsis mutants deficient in GA accumulation (using Fluorescein-tagged GA molecules, or GA-Fl), Tal et al. (2016) identified a mutant lacking NPF3.1 that accumulates very low levels of GA₃-Fl and GA₄-Fl. Uptake experiments in Xenopus confirmed that NPF3.1-expressing oocytes accumulate GA₃ and GA₄, but also GA₁, GA₈, and GA₂₀. The plasmamembrane localization of NPF3.1 together with its ability to transport GA and the phenotype of the mutant demonstrate that NPF3.1 is a GA importer in the elongating endodermal cells of the root (Tal et al., 2016). David et al. (2016) confirmed in yeast that NPF3.1 is able to transport GA₁, GA₃, and GA₄, but also the inactive GAs, GA₈ and GA₁₉, and they demonstrated that NPF3.1 has a role in GA transport in planta under low-nitrate conditions.

A GA₁₂ transporter has not yet been characterized.

Jasmonates

Jasmonates are membrane lipid-derived hormones that have many functions, from regulation of plant defenses and responses to environmental stresses, to contribution to plant growth and developmental processes. Jasmonates include jasmonic acid (JA) and its derivatives methyl-jasmonate (MeJA) and jasmonoyl-isoleucine (JA-Ile), with JA being the prohormone with low activity, MeJA being the volatile methyl ester, and JA-Ile being the bioactive form of JA (Gfeller et al., 2010).

In a search for new regulators of JA signalling, Saito et al. (2015) identified NPF2.10, a gene encoding a high-affinity H⁺/glucosinolate transporter that was tightly co-expressed with JA biosynthetic genes. Import experiments in Xenopus oocytes have shown that in addition to GA and glucosinolate, NPF2.10 is also able to transport the bioactive JA-Ile. Furthermore, the npf2.10 KO mutant displays altered responses to exogenous JA treatment, and co-treatment with GA and JA is more effective to restore seed production of this mutant compared to treatment with one hormone alone (Saito et al., 2015).

Using their modified yeast two-hybrid system with JA-Ile-specific receptor complexes, Chiba et al. (2015) also identified NPF members able to transport the bioactive JA-Ile in yeast, namely: NPF1.1, NPF1.2, NPF4.1, NPF5.7, NPF8.1, and NPF8.2, and to a lesser extent NPF2.4, NPF2.6, NPF2.7, NPF2.10, NPF2.13, NPF3.1, and NPF5.1. However, the functions of these proteins in JA/JA-Ile transport in vivo remain unknown.

Conclusions

Whilst the data summarized in Table 1 should give us a clearer view of the NPF substrates, what becomes evident is the multi-substrate nature of most of the NPF members. Competitive interactions between different substrates, such as nutrients and hormones, should now be studied in detail to strengthen the hypothesis that NPF members are the basis of the integration of environmental and physiological information linked to the relative availability of the different nutrients.

Although a substrate is known for 32 of the Arabidopsis NPF members (Table 1), the detailed transport mechanisms are not well described for most of them. Exhaustive electrophysiological experiments, accumulation assays, and structure–function studies need to be performed in order to decipher the transport pathways of the different substrates. Structural data should help in understanding the molecular basis of the selectivity of these transporters, leading to an explanation of how a transporter can transport structurally different substrates (e.g. nitrate/ABA or nitrate/glucosinolates) while still being selective.

All the transporters characterized and described in this review are, or are thought to be, plasmamembrane-localized, but at least two of the Arabidopsis NPF members with unknown substrates are tonoplast-localized (NPF8.4 and 8.5; Weichert et al., 2012). The molecular determinants of this subcellular localization have been studied (Komarova et al., 2012). Tonoplast-localized low-affinity nitrate transporter are present in other plants (Hu et al., 2016).

Finally, it should be noted that ABC transporters were thought to be the only family of what we may term ‘portnawak’ transporters, by which we mean a family of transporters with a very broad substrate specificity or without selectivity. However, it is now clear that this behavior is shared with the large majority of transporter families, the latest one to be identified being the SWEET transporters that were believed to be sugar-selective transporters but which were recently demonstrated also to be gibberellin transporters (Kanno et al., 2016).

Note added in proof

Payne et al. (2017) very recently demonstrate that a Catharanthus roseus NPF (CrNPF2.9) is tonoplast localized and transport strictosidine, a central monoterpene indole alkaloid, from the vacuole to the cytosol.

Acknowledgements

CCF was the recipient of a fellowship from Agence Nationale de la Recherche (ANR-11-PDOC-020-01-NITRONET). Work in the BL laboratory is financially supported by the ANR (ANR-14-CE34-0007-01-HONIT) and the Région Languedoc-Roussillon (Chercheur d’Avenir).

References

Author notes