Plants absorb nitrogen (N) mostly in its two major inorganic forms, ammonium and nitrate ions. These ions are eventually assimilated into glutamine, then the N is transferred through transamination reactions to all other organic N molecules including other amino acids. In addition to the critical role as the central metabolites, amino acids serve as the N carrier between different organelles, tissues and organs. Distribution of assimilated N from the source (typically photosynthetic leaves) to sink tissues is mainly achieved through translocation of amino acids. To meet such needs, plant genomes encode numerous amino acid transporters (∼100 identified members of transporter families in Arabidopsis, Tegeder and Hammes, 2018), which function in transporting amino acids across the plasma membrane, as well as across organelle membranes.

Given the central role of amino acids in cellular metabolism and N balance between organs, it is not surprising that multiple mechanisms seem to exist to monitor the levels of amino acids. For example, TARGET OF RAPAMYCIN Complexes (TORCs) are cytosolic kinases that are activated by branched chain amino acids and are conserved among fungi, metazoans and plants (Nakajo et al., 2004; Fumarola et al., 2005; Kingsbury et al., 2015; Shimobayashi and Hall, 2016; Cao et al., 2019; Schaufelberger et al., 2019). Another cytosolic kinase, GENERAL CONTROL NON-REPRESSIBLE 2 (GCN2), is important for sensing N deficiency and is also conserved in fungi, metazoans, and plants. GCN2 does not sense amino acids per se, however it is activated when bound to de-acetylated tRNAs, which are more abundant when the cell is deprived of amino acids. Activated GCN2 represses protein synthesis by deactivating EUKARYOTIC ELONGATION INITIATION FACTOR 2α (eIF2α) through phosphorylation (Chantranupong et al., 2015). Therefore, TORC and GCN2 monitor the cytosolic amino acid sufficiency and deficiency, respectively. Additionally, GLUTAMATE RECEPTOR-LIKE (GLR) proteins, which structurally resemble the neuronal glutamate receptors from metazoans, seem to be involved in the perception of extracellular amino acids (Qi et al., 2006; Stephens et al., 2008; Vincill et al., 2012; Tapken et al., 2013; Kong et al., 2016; Goto et al., 2020). In addition to the proteogenic amino acids sensed by the above pathways, specialized non-proteogenic amino acids, such as β-aminobutyric acid (Luna et al., 2014) and pipecolic acid (PIP) and its derivative N-hydroxypipecolic acid (NHP) (Huang et al., 2020), function as defense signaling molecules, both of which activate the SA-dependent defense pathway.

The existence of amino acid sensing mechanisms in plants suggests that alteration of amino acid levels via altered amino acid metabolism or transport might trigger pleiotropic responses. Indeed, several independent studies in which the transport of amino acids has been enhanced or altered seem to confirm this possibility. A study using knockout lines for the amino acid importer LYSINE HISTIDINE TRANSPORTER 1 (LHT1) revealed that amino acid contents within the tissue and the extracellular fluid were both altered, and these mutants showed a constitutive stress response mediated by salicylic acid (SA) (Hirner et al., 2006; Liu et al., 2010). Similarly, overexpression of CATIONIC AMINO ACID TRANSPORTER 1 (CAT1) caused a stunted phenotype, associated with increased SA levels, resistance against Pseudomonas synringae, and the upregulation of genes associated with the development of systemic acquired resistance (pathogenesis- related genes or PR) (Yang et al., 2014). Similar observations were reported upon over-expression of GLUTAMINE DUMPER 1 (GDU1), a single-transmembrane protein which promotes amino acid export through an unknown mechanism (Pilot et al., 2004; Pratelli et al., 2010). Similar to lht1 knockouts and CAT1 overexpressor mutants, SA levels in GDU1 overexpressing plants were elevated (Liu et al., 2010).

Recent discovery of UMAMITs which are bidirectional facilitators of amino acid transport, offer the possibility to interrogate the response of plants to increased amino acid export directly. The goal of this study was to investigate the effects of overexpressing UMAMIT genes that were shown to promote amino acid transport in plants (UMAMIT14, 18, 24, and 25, Ladwig et al., 2012; Müller et al., 2015; Besnard et al., 2016, 2018) on plant development, growth, and reproduction to further examine the link between amino acid transport and stress responses. The results presented here show strong evidence that amino acid export activity positively correlates with stress phenotypes and pathogen resistance, most likely due to the establishment of a constitutive SA-mediated stress response.

In the present study, it has been shown that over-expressing four independent UMAMITs that function as amino acid exporters causes a stunted growth phenotype (Figures 1, 2 and Supplementary Figure 1). Further studies of two such over-expressor lines, 35S:UMAMIT14-4 and 35S:UMAMIT14-6, showed altered physiological traits such as a decrease in biomass and seed yield (Table 1). UMAMIT14 over-expression also led to accumulation of SA and upregulation of a marker gene responding to the SA-dependent signaling pathway, PR1 (Malamy et al., 1990; Van Loon and Van Strien, 1999). ALD1 expression was also increased, indicating that the PIP/NHP-mediated pathway was also activated. Corroborating the activation of SA and PIP/NHP pathways, increased resistance to an SA-responsive pathogen was also observed (Figures 4, 5). Thus, a variety of elevated immune response hallmarks were observed.

The results from our study are in agreement with previous reports in which the mis-regulation of amino acid transport led to stress responses (Pilot et al., 2004; Hirner et al., 2006; Yang et al., 2014; Liu et al., 2016). Additionally, a forward genetics screen revealed that the RESISTANCE TO PHYTOPHTHORA PARASITICA 1 (RTP1) mutant, which shows an elevated SA-mediated stress response, is caused by a defect in the UMAMIT36 gene, although amino acid transport activity for UMAMIT36 has yet to be established (Pan et al., 2016). It is worthwhile noting that such defense-inducing phenotypes are likely associated with both increased (in case of CAT1 over-expression and RTP1 loss-of-function) and decreased (LHT1 knock-out, GDU1-D, UMAMIT14, UMAMIT 18, UMAMIT 24, UMAMIT 25 over-expression) amino acid retention in the cytosol. In cases where amino acid profiles of the mutants were investigated, there was no clear common trend for the contents of specific amino acids being altered (Pilot et al., 2004; Hirner et al., 2006). Examination of amino acid profiles in 35S-UMAMIT14-4 line corroborates with these results, showing some amino acid composition changes but with no general trend common with lht1 or GDU1-D mutants (Supplementary Figure 3). This is similar to the results from multiple studies reporting the amino acid contents in various mutants of amino acid biosynthesis genes (Van Damme et al., 2009; Hwang et al., 2011; Stuttmann et al., 2011; Alvarez et al., 2012). While all mutants showed enhanced defenses responses, they all displayed different amino acid accumulations and profiles, without any clear correlation between single amino acid contents or phenotypes.

The exact mechanism through which the SA-mediated defense pathway is triggered in these amino acid transporter or metabolic enzyme mutants is unknown. One potential pathway is via pipecolic acid (PIP) and its derivative, N-hydroxypipecolic acid (NHP). PIP and NHP are generated in the tissue attacked by biotrophic pathogens, and NHP functions as a long-distance signal to induce systemic resistance in distal leaves (Bernsdorff et al., 2016; Chen et al., 2018). PIP is generated from Lys; therefore a change in Lys metabolism might trigger an overproduction of PIP/NHP. In addition, PIP is also an analog of the proteogenic amino acid Pro. Currently there is no evidence showing that any of these amino acid transporters transport PIP/NHP at a physiologically relevant concentration. If they do function as PIP/NHP transporters, at least some of the phenotypes might be explained by mis-localization of these signaling molecules. Another potential pathway could be via GCN2; amino acid deficiency caused by transporter misregulation could increase the concentration of free tRNA, which is perceived by GCN2 to induce stress responses (Li et al., 2013). Additionally, the change in the concentration of extracellular amino acids might trigger the responses of GLRs (reviewed in Forde and Roberts, 2014). Deciphering the exact mechanism will be complex, because these pathways are interconnected. For example, PIP/NHP induces the production of SA (Navarova et al., 2012), which, in turn, induces GCN2-dependent elF2 phosphorylation (Lageix et al., 2008). Although the downstream components of GLRs are not well known, AtGLR3;3 seems to be required for defense responses against biotrophic pathogens, which is dependent on the SA pathway (Manzoor et al., 2013). Genetic experiments conferring altered transporter activities, defects in PIP/NHP synthesis, and/or defects in GCN2- or GLR-mediated signaling could help to elucidate the canonical components required for the immune responses triggered by misregulation of amino acid transporters.

SA is also known to interact with other plant hormones. For example, there is a well-characterized reciprocal antagonism between SA and JA pathways (Ikemura et al., 1990). SA also suppresses the function of auxin by downregulating the auxin receptor, TIR1 (Wang et al., 2007). MYB96, a transcription factor downstream of ABA signaling is also known to induce SA response, and activation tagging line myb97-1 shows a dwarf phenotype which is dependent of SA production (Seo and Park, 2010). Therefore, it is possible that other hormones are involved in triggering the stress response observed in UMAMIT14 overexpressing plants. In our current study, however, we did not observe any consistent increase or decrease of other plant hormones (IAA, ABA, and JA) in the UMAMIT14 overexpression lines. Therefore, it is likely that the phenotypic responses we observe is primarily due to the activation of SA pathway.

While constitutive activation of the SA pathway via loss or gain of function of amino acid transport function to combat pathogens is unlikely to be an agronomically useful strategy due to the loss of biomass and seed yield, these characterized UMAMITs provide tools to study the link between amino acid levels and plant immune responses. For example, an inducible expression system of UMAMITs may be useful in identifying the early events triggered by enhanced amino acid export through RNAseq studies. Amino acids serve as the main N form acquired from the host in biotrophic pathogens, many of which induce accumulation of a specific set of amino acids at the infected site (Mckee et al., 1972; Kumar and Prasad, 1992; Horst et al., 2010a,b; Ward et al., 2010). Therefore, the pathway triggered by the overexpression of UMAMITs could represent the endogenous monitoring system for pathogen-induced alteration of amino acid metabolism. Positive regulation of SA pathway via amino acid export could interact with other SA-dependent defense mechanism activated in parallel, such as PAMP- and effector-triggered immunity (Pieterse et al., 2012). Indeed, a recent study revealed that amino acid transporter expression profiles in Arabidopsis leaves are different between compatible (i.e., the pathogen can establish an infection) and incompatible (i.e., the pathogen is incapable of infection) interactions, suggesting that amino acid transporter reprogramming is required to establish the infection (Sonawala et al., 2018). The significant question is whether the amino acid transporters are induced by the pathogens to support their growth, similar to the case of sugar exporters SWEET11 and 12 (Gebauer et al., 2017), or whether the plant actively regulates amino acid transporter activity to counter infection, as suggested by differential regulation of local amino acid transporters upon perception of pathogen-associated molecular patterns (PAMP) (Anderson et al., 2014) and systemic alteration of amino acid transporter activities in distal leaves (Schwachtje et al., 2018). Careful examination using genetics, physiology and molecular biology will be necessary to parse out such differences.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

JB, SO, JM, and GP designed the experiments. JB, US, BM, SF, and EC conducted experiments. JB, SO, SF, EC, and GP contributed in writing the manuscript. SO and GP supervised the project. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.