A single amino acid transporter controls the uptake of priming-inducing beta-amino acids and the associated tradeoff between induced resistance and plant growth

Abstract

Selected β-amino acids, such as β-aminobutyric acid (BABA) and R-β-homoserine (RBH), can prime plants for resistance against a broad spectrum of diseases. Here, we describe a genome-wide screen of fully annotated Arabidopsis thaliana T-DNA insertion lines for impaired in RBH-induced immunity (iri) mutants against the downy mildew pathogen Hyaloperonospora arabidopsidis, yielding 104 lines that were partially affected and four lines that were completely impaired in RBH-induced resistance (IR). We confirmed the iri1-1 mutant phenotype with an independent T-DNA insertion line in the same gene, encoding the high-affinity amino acid transporter LYSINE HISTIDINE TRANSPORTER 1 (LHT1). Uptake experiments with yeast cells expressing LHT1 and mass spectrometry-based quantification of RBH and BABA in leaves of lht1 mutant and LHT1 overexpression lines revealed that LHT1 acts as the main transporter for cellular uptake and systemic distribution of RBH and BABA. Subsequent characterization of lht1 mutant and LHT1 overexpression lines for IR and growth responses revealed that the levels of LHT1-mediated uptake determine the tradeoff between IR and plant growth by RBH and BABA.

Introduction

The innate immune system enables plants to perceive and react to attacks by pathogens and herbivores. The basal component of this regulatory system is under the control of pattern recognition receptors that perceive molecular nonself-patterns from the attacker or damaged-self patterns that form during an attack (Choi and Klessig, 2016). Following recognition of these alarm signals, a signaling network is initiated that orchestrates the induction of cellular defense mechanisms, including reactive oxygen species (ROS), callose-rich cell wall depositions, and the induction of defense-related genes (Chisholm et al., 2006; Bigeard et al., 2015). Besides this pattern-triggered immunity (PTI), innate immunity can be triggered by susceptibility-inducing pathogen effectors. If the challenged plant expresses a resistance (R) gene that can recognize the activity of such a pathogen effector, the innate immune response is referred to as effector-triggered immunity (ETI; Cui et al., 2015). In addition to innate immunity, plants can acquire long-lasting resistance, which develops after recovery from biotic stress. This induced resistance (IR) is typically based on the priming of the innate immune system, which mediates a faster and/or stronger induction of inducible defenses upon secondary attack (Wilkinson et al., 2019; De Kesel et al., 2021). Moreover, IR can be triggered by root colonization of selected plant-beneficial microbes or treatment with specific chemical agents, such as microbe-associated molecular patterns, volatile organic compounds, and nonproteinogenic β-amino acids (Mauch-Mani et al., 2017; De Kesel et al., 2021).

β-amino butyric acid-IR (BABA-IR) has emerged as a popular model system to study the molecular mechanisms controlling immune priming in plants. BABA-IR has been reported in more than 40 plant species against different types of pathogens (Cohen, 1994; Cohen et al., 2016). In Arabidopsis (Arabidopsis thaliana), BABA primes both salicylic acid (SA)-dependent and independent defense mechanisms and protects plants against biotrophic, hemibiotrophic, and necrotrophic pathogens (Zimmerli et al., 2000; Ton et al., 2005; Schwarzenbacher et al., 2020). Recent evidence suggests that BABA accumulates during exposure to biotic and abiotic stress (Thevenet et al., 2017), which provides biological relevance and supports previous evidence that an aspartyl tRNA aspartase, IMPAIRED IN BABA-INDUCED DISEASE IMMUNITY 1 (IBI1), acts as a plant receptor for BABA (Luna et al., 2014). BABA was also suggested to act as a microbial rhizosphere signal, based on the finding that induced systemic resistance upon root colonization by Pseudomonas simiae WCS417 is blocked in the ibi1-1 mutant (Luna et al., 2014). Despite the apparently high efficiency by which plant roots are capable of taking up BABA from the soil (Zimmerli et al., 2000; Ton et al., 2005), a cellular transporter for this well-known priming agent has not been identified.

Although BABA-IR is effective against a broad spectrum of plant diseases, high doses of BABA results in major growth reduction (Wu et al., 2010; Luna et al., 2014). This undesirable side effect is partly caused by disruptive binding of R-BABA to the aspartic acid-binding pocket of the IBI1 enzyme, causing the accumulation of uncharged tRNA^Asp and GENERAL CONTROL NON-DEREPRESSIBLE 2-dependent inhibition of translation (Luna et al., 2014; Buswell et al., 2018). To search for less phytotoxic IR analogs of BABA, we previously screened a small library of structurally related β-amino acids for IR activity and phytotoxicity in Arabidopsis. This screen resulted in the identification of R-β-homoserine (RBH), which induces resistance in Arabidopsis and tomato (Solanum lycopersicum L.) cultivar Micro-Tom) against biotrophic and necrotrophic pathogens without growth reduction (Buswell et al., 2018). A recent study comparing four IR agents for their effectiveness in strawberry (Fragaria × ananassa) against Botrytis cinerea also identified RBH as the most effective IR agent without negative effects on plant growth (Badmi et al., 2019). Like BABA, RBH primes defense activity of callose-rich papillae, which in Arabidopsis are formed at relatively early stages of infection by the biotrophic oomycete Hyaloperonospora arabidopsidis (Hpa). Interestingly, despite its structural similarity to BABA, RBH does not require the IBI1 receptor to induce resistance in Arabidopsis (Buswell et al., 2018). Furthermore, unlike BABA, RBH does not prime SA-dependent induction of gene expression but primes camalexin production upon infection by Hpa and the expression of jasmonic acid-dependent defense genes after infection by the necrotrophic fungus Plectosphaerella cucumerina (Zimmerli et al., 2000; Ton et al., 2005; Buswell et al., 2018). Hence, RBH-IR is controlled by partially distinct pathways relative to BABA-IR. Importantly, the molecular mechanisms responsible for the uptake and perception of RBH are unknown.

In this study, we conducted a genome-wide screen of Arabidopsis T-DNA insertion mutants for impaired in RBH-induced immunity (iri) phenotypes against Hpa, yielding 104 and 4 lines that are partially and completely impaired in RBH-IR, respectively. Of the latter, we characterized the iri1 mutant, which is affected in the high-affinity amino acid transporter LYSINE HISTIDINE TRANSPORTER 1 (LHT1). We provide evidence that the level of LHT1-mediated uptake determines the balance between IR responses and plant tolerance to RBH and BABA. Furthermore, mass spectrometry analysis of leaves from RBH- and BABA-treated wild-type, lht1 mutant and LHT1-overexpressing plants revealed that LHT1 is critical for the uptake and systemic distribution of both RBH and BABA, while uptake experiments with LHT1-expressing yeast cells demonstrated that LHT1 acts as a high-affinity transporter of BABA and RBH. In support of other studies that have linked LHT1 to plant–microbe interactions and plant immunity, we conclude that LHT1 acts as a master regulator of the tradeoff between growth and IR by priming-inducing β-amino acids.

Results

Genome-wide screen for iri mutants

To search for new regulatory genes of RBH-IR, we screened 23,547 T-DNA insertion lines from the SALK and SAIL collections (Alonso et al., 2003) for an iri phenotype against Hpa. This set of T-DNA insertion lines covers >90% of all annotated protein-coding genes in the Arabidopsis genome. In contrast to conventional ethyl methanesulfonate-based mutant screens, which rely on the selection of mutant phenotypes in individual plants, the collection of fully annotated homozygous T-DNA insertion mutants allowed us to screen five genetically identical seedlings per line for quantification of the iri mutant phenotype, including partial loss of RBH-IR. To reduce false positives, we performed the screen in three successive stages. In the first stage, we screened seedlings in 400-well trays, in which the soil was soaked to saturation with RBH to a final soil concentration of ∼0.5 mM, followed by inoculation with Hpa conidiospores and scoring for visual sporulation at 5- to 7-day postinoculation (dpi; Figure 1A). Each tray yielded ∼1–2 lines displaying sporulation for at least two seedlings/well by 7 dpi; these lines were selected and rescreened during stage 2, using the same 400-well tray selection system. Stage 2 yielded 427 putative iri mutant lines (Figure 1A). These putative iri mutant lines were taken forward for final validation in stage 3, which was based on categorical scoring of Hpa colonization in trypan-blue-stained leaves from control- and RBH-treated plants (0.5 mM) of each candidate line (Figure 1A). To validate the statistical robustness of this screening stage, we conducted a pilot experiment that compared Hpa colonization between 40 pots of Columbia-0 (Col-0) seedlings pretreated with either water or RBH (0.5 mM). Categorical scoring of trypan blue-stained leaves confirmed statistically uniform distributions of Hpa colonization within each treatment (Supplemental Figure S1A). Of the 427 putative iri lines from stage 2, we confirmed 104 lines as having partially impaired RBH-IR in stage 3, as evidenced by statistically enhanced levels of Hpa colonization in RBH-treated mutant plants compared to RBH-treated wild-type plants, while still showing a statistically significant reduction in Hpa colonization by RBH treatment compared to the water controls (Figure 1A; Supplemental Figure S1B; Supplemental Data Set 1). An additional four lines, named iri1-1 to iri4-1, showed a full impairment of RBH-IR, as indicated by statistically identical levels of Hpa colonization between RBH- and water-treated plants within each line (Figure 1A; Supplemental Figure S1B; Supplemental Data Set 1).

Identification of IRI1/LHT1 as a critical regulator of RBH-IR against Hpa

Since SALK/SAIL lines can carry multiple T-DNA insertions and/or T-DNA-induced mutations (Alonso and Ecker, 2006), it is possible that the iri mutant phenotypes are caused by mutations in genes other than those identified and annotated by PCR border recovery analysis. To address this possibility, we quantified RBH-IR in independent T-DNA insertion lines in the annotated genes for each of the four complete iri lines (Figure 1, B and C; Supplemental Figure S2, A and B). Since RBH-IR against Hpa in Arabidopsis is associated with greater effectiveness of callose-rich papillae (Buswell et al., 2018), we quantified the effectiveness of callose-mediated cell wall defense at 3 dpi, as detailed previously (Schwarzenbacher et al., 2020). All original iri lines consistently lacked RBH-IR and concomitantly failed to augment callose-mediated defense upon RBH treatment (Figure 1D; Supplemental Figure S2C), confirming the importance of this postinvasive defense barrier in RBH-IR against Hpa. However, independent T-DNA insertions in the annotated genes inactivated in iri2-1, iri3-1 or iri4-1 did not affect RBH-IR and showed wild-type levels of callose-mediated defense against Hpa (Supplemental Figure S2C), indicating that their iri phenotypes are caused by T-DNA-induced mutations in other genes. In contrast, an independent T-DNA insertion mutant (iri1-2) in the annotated gene disrupted in iri1-1 displayed a complete iri phenotype (Figure 1, B and C) and was concomitantly impaired in RBH-induced priming of callose defense (Figure 1D). The iri1-1 and iri1-2 mutants carry a T-DNA insertion in the fifth intron and the second intron of LYSINE HISTIDINE TRANSPORTER1 (LHT1; At5g40780; Figure 1B; Supplemental Figure S3, A and B), respectively. LHT1 encodes a high-affinity amino acid transporter for acidic and neutral amino acids in roots and mesophyll cells (Chen and Bush, 1997; Hirner et al., 2006; Svennerstam et al., 2007). We will therefore refer to IRI1 as LHT1 thereafter.

LHT1 controls RBH uptake from the soil

Since LHT1 was characterized as an amino acid transporter (Chen and Bush, 1997), we hypothesized that the lack of RBH-IR in lht1 mutants (lht1-5, for iri1-1; and lht1-4, for iri1-2) might be caused by impaired RBH uptake from the soil. To test this hypothesis, we saturated the soil with increasing RBH concentrations and determined RBH concentrations in the leaves of Col-0, iri1-1 and a previously characterized LHT1 overexpression line (Hirner et al., 2006; 35Spro:LHT1), which shows a 27-fold higher LHT1 expression level than Col-0 plants under our experimental conditions (Supplemental Figure S3C). At 2 days after soil treatment, we harvested replicate leaf tissues for RBH quantification by hydrophilic interaction liquid chromatography coupled to quadrupole time-of-flight mass spectrometry (HILIC-Q-TOF; Figure 2A) or challenged the leaves with Hpa to quantify RBH-IR (Figure 2B). The three genotypes differed statistically in their RBH shoot concentrations after soil treatment with increasing RBH concentrations, as evidenced by a highly statically significant interaction between soil treatment and genotype (two-way ANOVA; P < 0.001; Figure 2A). For both Col-0 and 35Spro:LHT1, RBH shoot accumulation showed a dose-dependent rise with increasing RBH concentrations in the soil. The 35Spro:LHT1 seedlings accumulated statistically higher RBH concentrations in their shoots than Col-0 after saturating the soil to a final concentration of 0.15 mM or 0.5 mM RBH, whereas RBH concentrations in the shoot of lht1-5 were hardly detectable by HILIC-Q-TOF and failed to show a dose-dependent increase with RBH soil treatment (Figure 2A). The observed variation in RBH shoot concentrations correlated with RBH-IR intensity against Hpa (Figure 2B); while RBH failed to induce statistically significant levels of resistance in lht1-5 at all concentration tested, 35Spro:LHT1 plants showed increased levels of RBH-IR compared to Col-0 at all RBH concentrations tested. Notably, the relatively low concentration of 0.05 mM RBH failed to protect Col-0 against Hpa seedlings, whereas the same RBH concentration induced a statistically significant reduction in Hpa colonization in 35Spro:LHT1 (Figure 2B). Thus, RBH uptake from the soil by LHT1 increases by overexpression of LHT1, which in turn boosts RBH-IR against Hpa.

Tolerance to RBH depends on LHT and not on catabolism

In contrast to BABA, RBH induces resistance in Arabidopsis without concomitant growth inhibition (Buswell et al., 2018). To examine whether LHT1 controls tolerance to RBH, we quantified seedling growth of Col-0, lht1-5, and 35Spro:LHT1 on Murashige and Skoog (MS) agar medium. To strengthen the evidence that RBH-induced phytotoxicity in 35Spro:LHT1 depends on LHT1 uptake, we conducted this experiment in the presence of increasing concentrations of L-Ala, a high-affinity substrate of LHT1 (Hirner et al., 2006), expecting that if tolerance is controlled by LHT1-dependent uptake, the L-Ala in the medium would outcompete RBH for uptake and antagonize RBH-induced phytotoxicity. Indeed, while green leaf areas (GLAs) of Col-0 and lht1-5 were unaffected by increasing concentrations of RBH after 1 week of growth, growth of the 35Spro:LHT1 overexpression line showed a dose-dependent repression with increasing RBH concentrations, which was antagonized by L-Ala in a dose-dependent manner (Figure 3). Together with our earlier finding that RBH uptake increased in the 35Spro:LHT1 line (Figure 2A), these results indicate that natural tolerance of Arabidopsis to RBH (Buswell et al., 2018) is determined by RBH uptake capacity of LHT1.

To exclude a role for catabolism in RBH tolerance, we repeated the experiment on MS medium without inorganic nitrogen (N_inorg; ${\text{NO}}_3^{–}$ and ${\text{NH}}_4^{+}$⁠), supplemented with increasing concentrations of RBH and L-Ala. Importantly, Arabidopsis failed to grow on agar medium without N_inorg (Supplemental Figure S4), and increasing RBH concentrations in the growth medium failed to rescue growth. Hence, Arabidopsis cannot metabolize RBH as an N source, which rules out metabolic breakdown (catabolism) as a mechanism of RBH tolerance. In contrast, increasing L-Ala concentrations added to the agar medium rescued seedling growth of all genotypes, albeit to varying degrees. While 35Spro:LHT1 seedlings showed the strongest growth response to increasing L-Ala concentrations, Col-0 displayed an intermediate growth response, followed by a relatively weak growth response in lht1-5 (Supplemental Figure S4), thus confirming the contribution of LHT1 to L-Ala uptake (Hirner et al., 2006; Svennerstam et al., 2007, 2011). Notably, increasing RBH concentrations in the presence of L-Ala caused a dose-dependent growth reduction in 35Spro:LHT1 seedlings but not in Col-0 or lht1-5 (Supplemental Figure S4), which supports our conclusion that increased RBH uptake through LHT1 overexpression renders Arabidopsis sensitive to RBH-induced stress due to accumulation of phytotoxic RBH concentrations that cannot be catabolized. Hence, tolerance of Arabidopsis to RBH is controlled by LHT1-dependent uptake of RBH, rather than catabolism of RBH.

LHT1 also controls BABA uptake, BABA-IR, and BABA tolerance

Given the published broad substrate range of the LHT1 transporter for acidic and neutral amino acids (Hirner et al., 2006; Svennerstam et al., 2007), we examined whether LHT1 also plays a role in the uptake of BABA. To this end, we harvested replicate shoot tissues of Col-0 and lht1-5 seedlings to quantify in planta concentrations of BABA at 2 days after saturating the soil with increasing concentrations of the chemical (0, 0.025, 0.05, 0.15, and 0.5 mM), using HILIC-Q-TOF (Figure 4A). While saturating the soil on which Col-0 seedlings grew with increasing BABA concentrations resulted in a dose-dependent increase of BABA concentrations in the shoot (Figure 4A), a similar treatment of the lht1-5 mutant failed to increase shoot BABA concentrations (Figure 4A), indicating that BABA uptake is dependent on LHT1. To corroborate this, we saturated the soil of Col-0, lht1-5 and 35Spro:LHT1 seedlings with increasing BABA concentrations and scored BABA-IR against Hpa (Figure 4B). As reported previously, BABA was more efficient than RBH in protecting Col-0 against Hpa (Buswell et al., 2018), already reducing Hpa colonization at 0.025-mM BABA and reaching maximum levels of resistance at concentrations of 0.05 mM and higher (Figure 4B). The 35Spro:LHT1 line showed even higher levels of resistance at 0.025-mM BABA compared to Col-0, indicating that these seedlings are sensitized to respond to BABA. In contrast, the lht1-5 mutant was severely compromised in its effectiveness of BABA-IR, and only displayed weak levels of IR at soil BABA concentrations of 0.25 and 0.5 mM (Figure 4B). Thus, like RBH-IR, BABA-IR depends on a functional LHT1 transporter and is enhanced by overexpression of LHT1.

To determine whether LHT1 also controls BABA-induced phytotoxicity, we quantified the growth of Col-0, lht1-5 and 35Spro:LHT1 seedlings growing on MS agar plates supplemented with phytotoxic concentrations of BABA. As shown in Figure 5, GLA values of Col-0 after 1 week of growth declined with increasing BABA concentrations. This BABA-induced stress increased dramatically in 35Spro:LHT1 seedlings and decreased in lht1-5 seedlings (Figure 5). The fact that lht1-5 seedlings still showed growth repression at higher BABA concentrations suggests that additional mechanisms contribute to BABA-induced phytotoxicity. To compare the severity of RBH- and BABA-induced phytotoxicity, we cultivated Col-0, lht1-5 and 35Spro:LHT1 seedlings on MS agar plates containing the same doses of RBH or BABA (0.25 mM, 0.5 mM, 1 mM or 2.5 mM). Of the three genotypes tested, only 35Spro:LHT1 seedlings were affected in growth by both chemicals at concentrations of 0.25 mM and above (Supplemental Figure S5A), with BABA causing more severe growth repression than RBH (Supplemental Figure S5B). Quantification of GLAs of 35Spro:LHT1 across all inhibitor concentrations confirmed that BABA is more potent in repressing growth than RBH (Supplemental Figure S5B). Collectively, our results indicate that LHT1 is the dominant transporter for BABA uptake from the soil, controlling both BABA-IR and BABA-induced stress.

LHT1 transports both RBH and BABA

Having established that LHT1 is responsible for the uptake of RBH and BABA, we next examined the kinetics by which LHT1 transports these β-amino acids. To this end, we heterologously expressed the Arabidopsis LHT1 coding sequence in the yeast (Saccharomyces cerevisiae) 22Δ10α strain, which lacks 10 amino acid transporter genes and is completely deficient in the uptake of amino acids (Besnard et al., 2016). In contrast to empty vector (EV)-transformed 22Δ10α cells, the LHT1-expressing 22Δ10α strain was capable of growing on agar plates containing 1 mM L-Ala as the only nitrogen (N) source (Figure 6A), while supplementing liquid growth medium without inorganic (NH₄)₂SO₄ with increasing L-Ala concentrations steadily improved planktonic growth by LHT1-expressing 22Δ10α cells (Figure 6B). Increasing RBH and BABA concentrations in liquid growth medium with 1 mM L-Ala repressed growth by LHT1-expressing 22Δ10α cells completely (Supplemental Figure S6, A and B, respectively), despite the fact that both chemicals only marginally repressed 22Δ10α growth in liquid medium with 10 mM (NH₄)₂SO₄ as an N source (Supplemental Figure S7). These results not only show that yeast fails to metabolize RBH and BABA, but they also suggest that increasing RBH and BABA concentrations outcompete L-Ala for cellular uptake.

To study the kinetics of RBH and BABA uptake, we carried out experiments with ¹⁴C-labeled L-Ala in the absence and presence of RBH or BABA. To this end, we incubated EV- and LHT1-expressing 22Δ10α cells for 2, 5, and 10 min in buffer containing 50 µM or 500-µM L-Ala with a fixed amount of ¹⁴C-L-Ala for incubation, after which we quantified cellular L-Ala uptake by ¹⁴C scintillation. In contrast to EV-transformed cells, LHT1-expressing cells showed a linear uptake for L-Ala over time (Supplemental Figure S8), confirming the functionality of the transporter in yeast. To determine whether RBH and BABA competitively inhibit the LHT1 transporter for L-Ala uptake, we incubated LHT1-expressing cells for 5 min in buffer containing increasing concentrations L-Ala and a fixed amount of ¹⁴C-L-Ala in the presence or absence of 500-µM RBH or 500-µM BABA (Figure 6, C and D). Plotting the uptake velocity (V_uptake; fmol L-Ala/cell) against L-Ala concentration revealed a dose-dependent increase until saturation (V_max; Figure 6, C and D). Based on these data, we calculated that LHT1 has a K_m value of 9.4 µM for L-Ala-uptake, which is in line with previously reported K_m values for acidic and neutral amino acids (Hirner et al., 2006). Although V_uptake in the presence of either 500-μM RBH or 500-μM BABA decreased across a lower range L-Ala concentration, it still reached similar V_max values at higher L-Ala concentrations, indicating that RBH and BABA are competitive inhibitors of L-Ala uptake by LHT1. To calculate the inhibition constants (Ki) of RBH and BABA, we conducted further uptake experiments in the presence of multiple inhibitor concentrations (0, 250, and 1,000 μM RBH/BABA) and increasing L-Ala concentrations. Dixon plots of the inverse uptake velocity (1/V_uptake) against inhibitor concentration (Cornish-Bowden, 1974; Yoshino and Murakami, 2009) were created to determine Ki values at the intersecting lines of the different L-Ala concentrations (1, 5, 25, 50, and 250 µM; Figure 6, E and F). Predicted intersects were called at modeled RBH/BABA concentrations that had the smallest 1/V_uptake range between the various L-Ala concentrations (Supplemental Figure S9), revealing a Ki of 87.9 µM for RBH and a Ki of 68.9 µM for BABA (Figure 6, E and F). Hence, LHT1 is a transporter of both β-amino acids and shows a higher affinity for BABA than for RBH.

Discussion

Using annotated T-DNA insertion lines for a genome-saturating mutant screen

We used a genome-covering collection of Arabidopsis T-DNA insertion lines in a forward mutant screen for regulatory genes of IR. The availability of homozygous T-DNA insertions with high genomic coverage (Alonso et al., 2003) facilitates a near genome-saturating screen. The use of this resource has several benefits compared to conventional mutant screens. First, the availability of T-DNA flanking sequences mapped to the Arabidopsis genome allows for immediate identification of gene candidates without having to commit to a time-consuming generation of mapping populations and linkage analysis. Second, the collection of homozygous mutant lines enables the screening of small populations that all carry the same mutant allele, which facilitates the identification of partial (leaky) mutant phenotypes, as illustrated by the identification of 104 iri lines that are partially affected in RBH-IR (Figure 1A; Supplemental Figure S1; Supplemental Data Set 1). This relatively high number of partial iri mutants supports the notion that IR is a highly quantitative form of resistance, relying on the additive contribution of multiple genes (Ton et al., 2006; Ahmad et al., 2010; Wilkinson et al., 2019). Thus, the within-genotype replication of this screen enables selection for genes that make a quantitative contribution to complex multigenic traits. A disadvantage of using annotated T-DNA insertion lines in a forward mutant screen is that a single T-DNA insertion line can carry multiple mutations (O'Malley et al., 2015). These mutations are not necessarily covered by the annotated T-DNA flanking sequences, since they can be caused by truncated T-DNA elements or mis-repairs of integration sites from abortive T-DNA integrations (leaving mutational footprints; Gelvin, 2021). Indeed, several other studies have reported that mutant phenotypes in this collection of T-DNA insertion lines do not always co-segregate with the annotated T-DNA insertion (De Muyt et al., 2009; Dobritsa et al., 2011; Wilson-Sánchez et al., 2014). To account for this issue, we validated the mutant phenotypes of the four complete iri mutants in independent T-DNA insertion lines of their disrupted annotated genes for both RBH-IR and augmented cell wall defense against Hpa (Figure 1, C and D; Supplemental Figure S2). Even though the iri phenotypes of the four original mutant lines were robust and reproducible (Figure 1, C and D; Supplemental Figure S2), only the phenotype of the lht1-5 (iri1-1) mutant could be confirmed in an independent T-DNA insertion line in the annotated disrupted gene. Identifying the causal mutation in the other three iri lines will require thermal asymmetric interlaced PCR (TAIL-PCR) to identify flanking sequences of alternative T-DNA insertions and/or conventional linkage analysis in segregating mapping populations.

The role of LHT1 in plant–biotic interactions

IRI1 encodes the broad-range amino acid transporter LHT1. Cellular transporters play important roles in the control of plant–pathogen interactions by facilitating pathogen feeding (Elashry et al., 2013; Marella et al., 2013), secretion of antibiotic compounds (Lu et al., 2015; Khare et al., 2017), transporting defense plant hormones (Serrano et al., 2013), or contributing to plant defense responses (Liu et al., 2010; Yang et al., 2014). Furthermore, the LHT1 ortholog LjLHT1.2 in birdsfoot trefoil (Lotus japonicus) is transcriptionally induced by arbuscular mycorrhizal fungi (AMF; Guether et al., 2011), suggesting that it facilitates AMF-dependent uptake of organic nitrogen. Given the role of LHT1 in IR, it is tempting to speculate that LHT1 also plays a role in mycorrhiza-IR (Cameron et al., 2013). In Arabidopsis, LHT1 has been implicated in the direct regulation of SA-dependent disease resistance. Liu et al. (2010) reported that lht1 mutant lines had increased basal resistance against the hemibiotrophic bacterium Pseudomonas syringae pv. tomato, the hemibiotrophic fungus Colletotrichum higginsianum, and the biotrophic fungus Erysiphe cichoracearum. The study furthermore provided evidence that LHT1 controls plant immunity by cellular uptake of L-glutamine (L-Gln), which is a precursor of the redox-buffering compound glutathione. Liu et al. (2010) proposed that the lower L-Gln uptake capacity in lht1 mutants suppresses cellular redox buffering capacity, thereby enabling augmented elicitation of ROS- and SA-dependent defenses upon pathogen attack. Our experiments did not reveal statistically significant differences in basal defense against the biotrophic oomycete Hpa between wild-type and lht1 mutant plants (Figures 1 and 2), in contrast to the results shown by Liu et al. (2010). This discrepancy may be explained by the fact that we used relatively young plants (2- to 3-week-old seedlings), which do not express SA-dependent age-related resistance (Kus et al., 2002). Indeed, other studies have reported that lht1 seedlings display normal growth phenotypes without the enhanced SA levels observed in older plants (Liu et al., 2010; Zhang et al., 2022). Accordingly, it is possible that glutamine-dependent redox regulation contributes to age-related resistance in older plants. Since LHT1 expression is lower in seedlings (Hirner et al., 2006), it is also possible that other amino transporters contribute to the cellular delivery of glutamine in these younger seedlings, such as AMINO ACID PERMEASE 1 (AAP1; Boorer et al., 1996) or CATIONIC AMINO ACID TRANSPORTER 8 (Yang et al., 2010). Interestingly, in contrast to the negative role of LHT1 in innate immunity reported by Liu et al. (2010), a recent study by Yoo et al. (2020) revealed that LHT1 contributes positively to ETI-related resistance in Arabidopsis against P. syringae pv. maculicola carrying the avirulence gene AvrRpt2. Moreover, Zhang et al. (2022) showed that LHT1 is the dominant transporter responsible for increased amino acid uptake during early PTI against pathogenic P. syringae, when it has a positive contribution to resistance by restricting bacterial colonization. Hence, LHT1 has been reported to have both positive and negative roles in innate plant resistance. It should be noted, however, that the immune-related function of LHT1 described in our study is related to IR by priming-inducing β-amino acids, rather than innate resistance.

The role of LHT1 in β-amino acid-IR

Our results have shown that LHT1 is the dominant transporter for cellular uptake of RBH and BABA from the soil (Figures 2 and 4). LHT1 localizes to the cell membrane (Hirner et al., 2006), which enables cellular import of RBH and BABA from the apoplast. LHT1 is expressed in root tips, lateral roots and mature leaves (Hirner et al., 2006), enabling cellular uptake of RBH and BABA in both roots and leaves. Since LHT1 is not expressed in the leaf vein, we propose that the activity of RBH and BABA in leaves is preceded by long-distance transport via the xylem and apoplastic distribution in the leaves. While BABA was applied exogenously in our experiments, recent studies have reported that biotic and abiotic stresses can elicit low concentrations of endogenous BABA in Arabidopsis (Thevenet et al., 2017; Balmer et al., 2019). Under these conditions, BABA only accumulates in locally stressed tissues and not systemically in nonstressed tissues (Balmer et al., 2019), indicating that stress-induced accumulation of BABA does not contribute to systemic defense signaling. Although the biosynthesis pathway of stress-induced BABA remains unknown, it seems plausible that this local biosynthesis occurs inside the cell. The Ki values of RBH (87.9 µM) and BABA (68.9 µM) indicate that these β-amino acids have marginally lower affinities for LHT1 than endogenous alpha-amino acids (Hirner et al., 2006). Since alpha-amino acids typically reach apoplastic concentrations between 1 and 10 µM (Zhang et al., 2022), it would be difficult for BABA to compete with these substrates. Moreover, Hpa-induced BABA concentrations do not exceed 25 ng g⁻¹ fresh weight (242.7 nM; Thevenet et al., 2017), which seems too low to be a competitive substrate for LHT1. Hence, cellular uptake of BABA by LHT1 does not appear to play a major role in Hpa-induced BABA accumulation, which would also explain why the lht1 mutant and 35Spro:LHT1 overexpression lines were not majorly affected in basal resistance to Hpa (Figure 2). Nevertheless, we cannot exclude that Hpa locally induces much higher BABA concentrations in the cells directly interacting with the parasite, and that LHT1 plays a role in countering diffusion of this intracellular BABA into the apoplast. In this context, it is interesting to note that Hpa infection induces LHT1 expression (Sonawala et al., 2018; Supplemental Figure S10), which could play a role in upholding defense-inducing intracellular concentrations of BABA in Hpa-challenged cells and would also explain why stress-induced BABA is not distributed systemically (Balmer et al., 2019).

While our results provide strong evidence that LHT1 is the dominant transporter for the uptake of RBH and BABA (Figures 2–6), they do not necessarily mean that the contribution of LHT1 to RBH-IR or BABA-IR solely depends on its uptake activity. For instance, while treatment with 0.05mM RBH resulted in similar foliar concentrations in both 35Spro:LHT1 and wild-type plants (Figure 2A), this relatively low RBH concentration only triggered a significant IR response in 35Spro:LHT1 plants and not in wild-type plants. This uncoupling of RBH concentration from IR suggests that the function of LHT1 in RBH-IR may involve an additional defense signaling activity that becomes active at low RBH concentrations. Such a transporter-receptor co-functionality (transceptor activity) has been reported for NITRATE TRANSPORTER 1.1 (NRT1.1) for nitrate uptake and signaling. Replacing Pro-492 with Leu-492 in NRT1.1 disabled the nitrate transport activity of this protein but not its ability to induce NRT2.1 expression (Ho et al., 2009), which is a nitrate-responsive gene that has concomitantly been linked to the regulation of disease resistance (Camanes et al., 2012). Although no amino acid transporters have been reported with receptor co-functionality (Dinkeloo et al., 2018), it is tempting to speculate that LHT1 might act as a transceptor of β-amino acids. Site-directed mutagenesis of LHT1 and testing whether its RBH and BABA transport activity can be uncoupled from its role in RBH-IR and BABA-IR would be required to test this hypothesis.

Since the lht1 mutant still displayed residual levels of BABA-IR and BABA-induced stress after treatment with high BABA doses (Figures 4, B and 5, B), we cannot exclude the possibility that other amino acid transporters have a minor contribution to BABA uptake. A recent study reported that LHT2 has a similar substrate specificity as LHT1, including several D-amino acids and 1-aminocyclopropane-1-carboxylate (ACC; Choi et al., 2019), and could thus have a complementary contribution to BABA uptake.

RBH and BABA compete with proteinogenic amino acids for uptake by LHT1

We used LHT1-expressing yeast cells to assess competitive inhibition of L-Ala uptake by RBH and BABA. Our uptake essays revealed a Km for LHT1 of 9.4 µM for L-Ala (Figure 6C), which supports previously reported Km values of LHT1 for proteinogenic amino acids (Hirner et al., 2006). Furthermore, the inhibitory kinetics of RBH or BABA on L-Ala uptake confirmed competitive inhibition, as evidenced by the fact that L-Ala uptake in the presence of RBH or BABA still reached maximum velocities at higher L-Ala concentrations (Figure 6, C and D). Of the two β-amino acids, BABA had a lower Ki than RBH (68.9 versus 87.9 µM), suggesting that LHT1 has a higher affinity for BABA than RBH (Figure 6, E and F). This difference in affinity is consistent with our observation that BABA has a stronger inhibitory effect on growth of 35Spro:LHT1 than RBH (Supplemental Figure S5). Since the affinity of LHT1 has been reported to be similar or higher for a range of acidic and neutral amino acids, including L-Gln (Hirner et al., 2006; Svennerstam et al., 2007), our results also explain previous findings by Wu et al. (2010), who showed that BABA-induced phytotoxicity in Arabidopsis can be alleviated by co-application with L-Gln.

LHT1: not just a transporter for proteinogenic amino acids

Although LHT1 was initially identified as a transporter for proteinogenic amino acids (Chen and Bush, 1997), subsequent studies have shown that it transports a much wider range for nonproteinogenic amino acids, such as the ethylene precursor ACC (Shin et al., 2015) and xenobiotic amino acid conjugates (Chen et al., 2018; Jiang et al., 2018). Consistent with this broad-spectrum uptake activity, we showed that LHT1 is the main transporter of the β-amino acids RBH and BABA. Of particular interest is the regulatory function of LHT1 in the tradeoff between β-amino acid-IR and plant growth. For BABA, overexpression of LHT1 in Arabidopsis increased BABA-IR at the relatively low concentration of 0.025-mM BABA (Figure 4) but it also dramatically increased plant sensitivity to BABA-induced growth repression (Figure 5; Supplemental Figure S5). However, RBH elicited high levels of IR in wild-type plants at soil concentrations of 0.15-mM RBH and above (Figure 2B) but did not repress growth across all concentrations tested (Figure 3), supporting our earlier conclusion that RBH induces disease resistance without costs on plant growth (Buswell et al., 2018). Interestingly, 35Spro:LHT1 overexpression plants increased the level of IR at relatively low RBH concentrations (Figure 2B), but also repressed growth in a dose-dependent manner (Figure 3; Supplemental Figure S5). Direct comparison of RBH- and BABA-induced growth repression in 35Spro:LHT1 plants confirmed that BABA is more active than RBH (Supplemental Figure S5B), which is also apparent from the IR response (Figures 2, B and 4, B). It is worth noting that the molecular mechanisms of RBH-induced stress remain unclear, and its lower toxicity in plants might come from a combination of uptake and intracellular modes of action.

The observed tradeoffs between β-amino acid-IR and plant growth reveal two important conclusions. First, like BABA, RBH can repress plant growth, but this phytotoxicity depends on LHT1-dependent uptake capacity rather RBH catabolism. Second, our results show that the tradeoff between β-amino acid-IR and growth can be optimized in favor of the IR response by manipulating the LHT1 gene. This conclusion holds major translational value for breeding programs aiming to exploit BABA-IR in vegetable crops that are protected by BABA but also suffer from BABA-induced phytotoxicity (Cohen et al., 2016; Yassin et al., 2021).

Materials and methods

Biological material

All Arabidopsis (A. thaliana) genotypes were in accession Col-0. The iri1-1 mutant (lht1-5l) and iri1-2 mutant (lht1-4) were described previously by Svennerstam et al. (2007) and Liu et al. (2010); the 35Spro: LHT1 overexpression lines were described by Hirner et al. (2006). The iri mutant screen was performed with fully annotated T-DNA insertion lines from the SALK and SAIL collections (Alonso et al., 2003) and purchased from the Nottingham Arabidopsis Stock Centre (sets N27941, N27951, N27942, N27943, N27944, and N27945). The annotated T-DNA insertions in iri1-1 (SALK_115555), iri1-2 (SALK_036871), iri2-1 (SALK_204380), SAIL_902_B08, iri3-1 (SALK_118654), SALK_078838, iri4-1 (SALK_076708), and SALK_046376 were confirmed by PCR before further testing (Supplemental Table S1), as described below. Hyaloperonospora arabidopsidis strain WACO9 was maintained in its asexual cycle by alternate conidiospore inoculations of Col-0 and hypersusceptible Wassilewskija (Ws) NahG plants.

Plant growth conditions

For soil-based IR experiments, seeds were sown in a 2:1 (v/v) Scott’s Levington M3 compost/sand mixture and stratified for 2–4 days in the dark at 4°C. Plants were subsequently cultivated under short-day conditions (8-h light (Sylvania GroLux T8 36W or Valoya NS1 LED); 150-µmol photons m⁻² s⁻¹; 21°C; and 16-h dark; 18°C) with a ∼60% relative humidity (RH). Plants for seed propagation were grown in long-day growth conditions (16-h light (Sylvania GroLux T8 36W); 150-µmol photons m⁻² s⁻¹; 21°C; and 8-h dark; 18°C) with ∼60% RH. For plate assays, seeds were surface sterilized (vapor-phase sterilization method) prior to sowing on half-strength MS medium (pH = 5.7 and 1% sucrose), solidified with 1.5% (w/v).

Mutant screen

Approximately 10–15 seeds for each seed line were sown in individual wells of 400-well trays (Teku JP 3050/230 H). Each tray was filled with ∼2.4 L of compost/sand mixture. After sowing, stratification of seeds and seed germination, seedlings were thinned to five seedlings/well. Two-week-old seedlings were treated with RBH by watering each tray with 1.5 L of 2× concentrated RBH solution (1 mM), which was left overnight to saturate the soil. Excess RBH solution (∼300 mL) was removed the next morning, resulting in a final soil concentration of ∼0.5-mM RBH. Challenge inoculation was performed at 2 days after RBH treatment by spraying seedlings with a suspension of Hpa conidiospores (10⁵ spores mL⁻¹). Trays were sealed with clingfilm after inoculation to maintain 100% RH and promote infection. To verify RBH-IR, each tray contained three randomly distributed wells with Col-0 seedlings. Furthermore, to verify favorable conditions for Hpa disease, three additional wells with Col-0 seedlings were cut out from each tray and left outside during RBH-uptake to prevent RBH-IR prior to inoculation. At 5–7 dpi, trays were visually inspected for Hpa sporulation when sporulation on Col-0 seedlings in the untreated wells of the tray became apparent. Lines developing sporulation within 7 dpi were scored as stage 1 impaired in RBH-induced-immunity (S1 iri) lines, while nongerminated lines were scored as stage 1 nongerminated (S1 ug). All S1 iri and S1 ug lines were pooled for the stage 2 screen in 400-well trays, as described above. S1 iri lines allowing visible sporulation in two screens time were scored as Stage 2 iri (S2 iri). S1 ug lines that germinated upon rescreening and showed sporulation were re-tested for S2 iri phenotypes. Of the 26,631 T-DNA insertion lines, 23,547 lines germinated and could be screened for iri mutant phenotypes. The 427 putative iri1 lines selected after stage 2 were pooled for seed bulking and validated by controlled IR assays in stage 3 (S3) of the screen, as described below.

IR assays

Two-week-old seedlings were grown in 60-mL pots, after which the soil was saturated with water (R)-β-homoserine (Sigma-Aldrich, St. Louis, MO, USA; #03694), or R/S-BABA (Sigma-Aldrich, #A44207) to the indicated concentrations, as described previously (Buswell et al., 2018). Two days after chemical treatment, seedlings were spray inoculated with a suspension of Hpa conidiospores (10⁵ spores mL⁻¹) and maintained in 100% RH to promote infection. Leaves were collected at 6–7 dpi for trypan blue staining for microscopy scoring of Hpa colonization by categorizing them into four classes, ranging from healthy leaves (I) to heavily colonized leaves (IV), as described in detail by Schwarzenbacher et al. (2020). To investigate augmented induction of cell wall defense by chemical priming treatment, leaves were harvested at 3 dpi for aniline blue/calcofluor staining and analysis by epifluorescence microscopy (Leica DM6B; light source: CoolLED pE-2; 365-nm excitation filter, L 425-nm emission filter, 400-nm dichroic filter). For each genotype/treatment combination, germinated coniodiospores on 10 leaves from independent seedlings were scored either as arrested (spores or germ tubes fully encased in callose), or nonarrested by callose depositions (no callose or lateral callose deposition along the germ tube/hyphae), as detailed by Schwarzenbacher et al. (2020). Statistical differences in Hpa colonization or callose defense were analyzed by pairwise Fisher's exact tests, using R software (version 3.5.1; Supplemental Data Set 2). For multiple comparisons, an additional Bonferroni multiple correction was applied, using the R package “fifer” (fifer_1.1.tar.gz; Supplemental Data Set 2).

Plant growth assays

Surface-sterilized seeds were sown onto half-strength MS agar plates and cultivated for 2 weeks under standard plant growth conditions, as indicated above. Photographs were taken after 1 and 2 weeks of growth with a Nikon D5300 digital camera. GLAs were quantified from digital photographs of 1- or 2-week-old seedlings, using Fiji/ImageJ software (Rueden et al., 2017). Statistical differences in the natural logarithm of (1+GLA) were analyzed by two-way analysis of variation (ANOVA), using R software (version 3.5.1; Supplemental Data Set 2).

Genotyping verification by PCR and gene expression analysis by RT–qPCR

Genomic T-DNA insertions of all iri1, iri2, iri3, and iri4 lines were confirmed by PCR using LP + RP and LBb1.3/LB3 + RP primers (Supplemental Table S1) To quantify LHT1 expression levels by reverse transcription–quantitative PCR (RT–qPCR), shoot tissues from five 2-week-old seedlings were collected and combined as one biological replicate. A total of five replicates were collected at the same time and snap-frozen in liquid nitrogen and homogenized. Total RNA was extracted using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany; cat. no. 74904) and first-strand cDNA was synthesized from 800 ng total RNA using a Maxima First Strand cDNA Synthesis Kit (Thermo Fisher, Waltham, MA, USA; cat. no. K1641). The cDNA was diluted 20 times in nuclease-free water before qPCR. All qPCR reactions were performed with 2-µL diluted cDNA and primer concentrations at a final concentration of 250 nM in a Rotor-Gene Q real-time PCR cycler (Qiagen; Q-Rex version 1.0), using a Rotor-Gene SYBR Green PCR Kit (Qiagen; cat. no. 204074). The qPCR amplification of LHT1 was performed with gene-specific primers (FP: ATCTCCGGCGTTTCTCTTGCTG, RP: GCCCATGCGATTGTTGAGTAGCTG) and normalized to the transcript levels of two housekeeping genes (At1g13440 [GLYCERALDEHYDE-3-PHOSPHATE DEHYDROGENASE C2, GAPC2], and At2g28390 [MONENSIN SENSITIVITY 1, MON1]), as detailed previously (Schwarzenbacher et al., 2020).

Quantification of in planta RBH and BABA concentrations by hydrophilic interaction liquid chromatography coupled to quadrupole time-of-flight mass spectrometry

Shoot tissues were collected at 2 days after soil-drenching and divided into four replicate tubes per treatment (five plants per tube, from separate trays), frozen at −80°C, freeze-dried and weighed. Dry tissue was crushed and extracted into 1 mL of cold extraction buffer (methanol:water:formic acid, 10:89.99:0.01, v/v/v). Extracts were centrifuged at 16,000 g for 5 min at 4°C, after which each supernatant was divided between three aliquots. RBH and BABA standards were prepared as individual standards from 0.1 to 100 μM. Separation was performed with a Waters Acquity HILIC BEH C18 analytical column, 1.7-mm particle size, 2.1 × 50 mm. The mobile phase was 20-mM ammonium formate with 0.1% (v/v) formic acid (A) and acetonitrile with 0.1% (v/v) formic acid (B). The gradient started at 99% (v/v) A and reached 65% (v/v) A in 4 min. The gradient changed to 1% (v/v) A up to 6 min and was held there for 1.5 min and then returned to initial conditions. The solvent flow rate was 0.3-mL min⁻¹, with an injection volume of 4 µL. Mass spectra were recorded in positive electro-spray ionization mode, using a Waters UPLC system interfaced to a Waters quadrupole time-of-flight mass spectrometer (Q-TOF; G2Si Synapt). Nitrogen was used as the drying and nebulizing gas. Desolvation gas flow was adjusted to ∼150 L h⁻¹ and the cone gas flow was set to 20 L h⁻¹ with a cone voltage of 5 V and a capillary voltage of 2.5 kV. The nitrogen desolvation temperature was 280°C and the source temperature was 100°C. The instrument was calibrated in 20–1,200 m/z range with a sodium formate solution. Leucine enkephalin (Sigma-Aldrich, St Louis, MO, USA) in methanol: water (50:50, v/v) with 0.1% (v/v) formic acid was simultaneously introduced into the quadrupole time-of-flight (qTOF) instrument via the lock-spray needle for recalibrating the m/z axis. Quantification of amino acids in tissues was based on the standard curves, using MassLynx version 4.1 software (Waters, Elstree UK). Amino acids identities were confirmed by co-elution of product fragment ions with parent ions and matching peak retention times to individual amino acid standards. Statistical differences in RBH and BABA between genotypes and soil-drench treatments were tested by two-way ANOVA followed by Welch t tests to test cross-genotype differences at each RBH/BABA concentration, using R software (version 3.5.1; Supplemental Data Set 2).

Yeast transformation

The LHT1 (At5g40780) coding sequence with stop codon was amplified from wild-type Col-0 cDNA with Phusion High-Fidelity DNA Polymerase (New England Biolabs, Ipswich, MA, USA; #M0530L) and cloned into the pENTR plasmid (Invitrogen). LHT1 was then subcloned into pDR196 (Meyer et al., 2006) by restriction (EcoRI and XhoI) and ligation (T4 DNA ligase). EV- and LHT1-harboring plasmids were confirmed by Sanger sequencing and introduced into competent cells of the 22Δ10α strain (Besnard et al., 2016), using heat shock transformation (Gietz and Schiestl, 2007).

Yeast growth assays

To assess the growth of LHT1- and EV-transformed 22Δ10α yeast strains, cells were first cultivated in liquid Yeast Nitrogen Base (YNB) medium (Alfa Aesar, #H26271, without amino acids and ammonium sulfate) supplemented with 10 mM ammonium sulfate at 30°C and 220 rpm for 2 days. Cells were washed by centrifugation at room temperature (3,000g; 5 min) and resuspended in distilled water to an optical density at 600 nm (OD₆₀₀) of 0.3–0.5. To assess whether yeast can metabolize RBH and BABA, 5 µL of the cell suspension was added to 2-mL Yeast Nitrogen Base and increasing concentrations of RBH or BABA (0.2–5 mM). To assess toxicity of RBH and BABA, 5 μL of the suspension was added to 2-mL YNB medium with 10-mM ammonium sulfate and increasing concentrations of RBH or BABA (0.2–5 mM). To assess competition between L-Ala and RBH or BABA, 5 µL of the suspension was added to 2-mL YNB medium supplemented with 1-mM L-Ala and increasing concentrations of RBH or BABA (0.2–5 mM). Cultures were incubated at 30°C with 220 rpm shaking for 3 days, after which the OD₅₉₅ was determined in a plate reader (FLUOstar OPTIMA; BMG LABTECH; Germany).

Assessment of uptake and inhibition kinetics of LHT1 in yeast

Transformed 22Δ10α cells were grown in YNB medium supplemented with 10-mM (NH₄)₂SO₄ at 30°C with shaking at 220 rpm for 2 days. Yeast cells were collected by centrifugation at room temperature (3,000 g; 5 min), washed in distilled water, and resuspended in ice-cold washing buffer (0.6-M sorbitol, 50-mM sodium phosphate, pH 4.5) to OD₆₀₀ of 5. Before the uptake assay, cells were energized by adding 1-M glucose (final concentration 50 mM) to the growth medium for 10 min. To assess time-dependent uptake of L-[¹⁴C] Ala in EV- and LHT1-transformed cells (Supplemental Figure S8), 1.5 mL of the energized cell culture was added to 1.5-mL uptake buffer, containing 50-nCi L-[¹⁴C]Ala (158-mCi mmol⁻¹; Perkin Elmer; NEC856) with unlabeled L-Ala (50 or 500 µM). After 2, 5, or 10 min of incubation in a thermomixer (Grant bio ES-20; Grant Instruments; UK; 30°C, 220 rpm), the cell suspensions were mixed with 2-mL ice-cold water and kept on ice to inhibit L-Ala uptake. Cells were then centrifuged (3,000 g; 5 min; 4°C) and washed 4 times with 2-mL ice-cold water, after which pellets were stored at −20°C for quantification of radioactivity the following day. To determine uptake and inhibition kinetics (Figure 6, C and D), LHT1-transformed cells were incubated in the same uptake medium, containing 50-nCi L-[¹⁴C] Ala with increasing concentrations (1–3,000 µM) of unlabeled L-Ala and/or 500-µM inhibitory RBH or BABA. After 5 min of incubation, cells were washed, collected, and stored as described above. To assess radioactivity, frozen pellets were resuspended in 750-μL distilled water, from which 200 μL was loaded onto Combusto-Pads (Perkin Elmer, Waltham, MA, USA; part number 5067034) and combusted in a sample oxidizer (Model 307 Sample Oxidizer; Perkin Elmer; USA). Trapped ¹⁴CO₂ was quantified by liquid scintillation counting (Tri-Carb 3100TR; Perkin Elmer; USA). L-Ala uptake velocities over the 5-min time window (V_uptake) were expressed as fmol L-Ala/cell and plotted against the L-Ala concentration, using the R package drc (Ritz et al., 2015) to determine the kinetics of L-Ala uptake in the absence and presence of RBH or BABA.

To estimate inhibition constants (Ki) of RBH and BABA (Figure 6, E and F), L-Ala uptake velocities were determined in the presence of 0, 250, and 1,000-µM RBH or BABA, using a medium containing increasing concentrations of L-Ala (1, 5, 25, 50, and 250 µM) with a fixed quantity of 50-nCi L-[¹⁴C]Ala. Dixon plots were created by plotting inverse L-Ala uptake velocities (1/V_uptake) against inhibitor concentration (RBH or BABA), after which five linear models for each L-Ala concentration were generated using the lm function (R base). Exact Ki values of RBH and BABA were determined by modeling 1,200 1/V_uptake values in the range between −200 and 1,000 µM of the inhibitor concentration using the predict() function (R base), after which Ki values were selected by calculating the inhibitor concentration yielding the minimum range in 1/V_uptake.

Accession numbers

LHT1 (IRI1): At5g40780.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Validation of putative iri mutants at stage 3 of the mutant screen.

Supplemental Figure S2. Characterization of RBH-IR in SALK/SAIL lines carrying independent T-DNA insertions in the same genes that are annotated to carry T-DNA insertions in iri2-1, iri3-1, and iri4-1.

Supplemental Figure S3. Genetic characterization of two independent lht1 mutant lines and the LHT1 overexpression line.

Supplemental Figure S4. Transgenic overexpression of LHT1 improves Arabidopsis growth on medium with L-alanine as the only N source, which is antagonized by co-application of RBH.

Supplemental Figure S5. Comparison of growth repression by low concentrations of BABA and RBH.

Supplemental Figure S6. RBH and BABA compete with L-alanine for LHT1 uptake and inhibit yeast growth.

Supplemental Figure S7. RBH and BABA have minimal effects on yeast growth but cannot be used as N source by yeast.

Supplemental Figure S8. Transformation of the yeast 22Δ10α mutant with LHT1 rescues uptake of L-[¹⁴C] alanine.

Supplemental Figure S9. Modeling exact inhibitor constants (Ki) of RBH and BABA.

Supplemental Figure S10. Induction of LHT1 expression by Hpa.

Supplemental Table S1. Primers used for characterization of T-DNA insertion lines.

Supplemental Data Set 1. Annotated genomic T-DNA insertions of the 108 confirmed iri lines, RBH-IR phenotypes, and expression profiles of the associated T-DNA-tagged genes.

Supplemental Data Set 2. Details of statistical tests and results used in the manuscript.

Acknowledgments

We thank Dr. Henrik Svennerstam for providing the seeds of the 35Spro:LHT1 line, Professor Guillaume Pilot for providing the 22Δ10α yeast line, Professor Stephen Rolfe and Dr. Pedro Rocha for advice on the enzyme kinetic experiment. We thank Dr. Karin Posthuma (Enza Zaden) for advice and support throughout the project. We gratefully acknowledge PhD student support from The De Laszlo Foundation.

Funding

This work was supported by a grant from the European Research Council (ERC; no. 309944 “Prime-A-Plant”) to J.T., a Research Leadership Award from the Leverhulme Trust (no. RL-2012-042) to J.T., a Biotechnology and Biological Sciences Research Council (BBSRC) Industry Parnership Award grant to J.T. (BB/P006698/1) and Supplementary grant from Enza Zaden to J.T., and a ERC-Proof-of-Concept grant to JT (no. 824985 “ChemPrime”). K.F. is supported by an ERC-Consolidator Grant (MYCOREV-865225).

Conflict of interest statement. None declared.

J.T. conceived the research. C.N.T., W.B., P.Z., R.S., and J.T. designed the experiments. C.N.T., W.B., P.Z., H.W., I.J., and K.F. conducted the experiments. C.N.T., W.B., P.Z., and J.T. analyzed the data. C.N.T., W.B., and J.T. wrote the paper.

The author(s) responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://dbpia.nl.go.kr/plcell) is: Jurriaan Ton (j.ton@sheffield.ac.uk).

References

Author notes