Auxin and Tryptophan Homeostasis Are Facilitated by the ISS1/VAS1 Aromatic Aminotransferase in Arabidopsis

Abstract

Indole-3-acetic acid (IAA) plays a critical role in regulating numerous aspects of plant growth and development. While there is much genetic support for tryptophan-dependent (Trp-D) IAA synthesis pathways, there is little genetic evidence for tryptophan-independent (Trp-I) IAA synthesis pathways. Using Arabidopsis, we identified two mutant alleles of ISS1 (Indole Severe Sensitive) that display indole-dependent IAA overproduction phenotypes including leaf epinasty and adventitious rooting. Stable isotope labeling showed that iss1, but not WT, uses primarily Trp-I IAA synthesis when grown on indole-supplemented medium. In contrast, both iss1 and WT use primarily Trp-D IAA synthesis when grown on unsupplemented medium. iss1 seedlings produce 8-fold higher levels of IAA when grown on indole and surprisingly have a 174-fold increase in Trp. These findings indicate that the iss1 mutant’s increase in Trp-I IAA synthesis is due to a loss of Trp catabolism. ISS1 was identified as At1g80360, a predicted aromatic aminotransferase, and in vitro and in vivo analysis confirmed this activity. At1g80360 was previously shown to primarily carry out the conversion of indole-3-pyruvic acid to Trp as an IAA homeostatic mechanism in young seedlings. Our results suggest that in addition to this activity, in more mature plants ISS1 has a role in Trp catabolism and possibly in the metabolism of other aromatic amino acids. We postulate that this loss of Trp catabolism impacts the use of Trp-D and/or Trp-I IAA synthesis pathways.

IN plants, the aromatic amino acids tryptophan (Trp), tyrosine (Tyr), and phenylalanine (Phe) are used for the synthesis of proteins and as precursors to a variety of specialized metabolites. While most secondary metabolites help protect the plant against abiotic and biotic stress (Tzin and Galili 2010), some such as Trp-derived indole-3-acetic acid (IAA) are essential growth regulators (Woodward and Bartel 2005). IAA, the primary auxin in plants, functions in establishing cell polarity during embryogenesis (Weijers et al. 2005; Kleine-Vehn et al. 2008), determination of leaf patterning (Bainbridge et al. 2008), and initiation of lateral roots and shoots (Celenza et al. 1995; Peret et al. 2009).

There are two general routes proposed for IAA biosynthesis in plants: from a Trp-dependent (Trp-D) pathway or through an indolic precursor of Trp in a Trp-independent (Trp-I) pathway (Woodward and Bartel 2005; Tivendale et al. 2014) (Figure 1). Within the Trp-D IAA biosynthetic pathway, there are three established pathways in plants that lead to IAA production: (i) the indole-3-pyruvic acid (IPA) pathway (ii) the indole-3-acetaldoxime (IAOx) pathway, and (iii) the indole-3-acetamide (IAM) pathway (Figure 1) (Ljung 2013; Zhao 2014).

Figure 1.

IAA biosynthetic pathway in Arabidopsis. (A) Simplified Trp biosynthetic pathway. IGP, indole-3-glycerolphosphate; TSA1, tryptophan synthase alpha; TSB1, tryptophan synthase beta. (B) IPA/YUC Branch; TAA1, Tryptophan Aminotransferase of Arabidopsis-1 (taa1* denotes mutant alleles named sav3, wei8, tir2, and ckrc1); TAR1, TAA1-Related-1; TAR1, TAA1-Related-2; YUC1-YUC11, YUCCA1-YUCCA11 (yuc* denotes multiple yuc mutants yuc1, yuc1-D, yuc2, yuc3, yuc4, yuc5, yuc6, yuc6-1D, yuc6-2D, yuc7, yuc8, yuc9, yuc10); IPA, indole-3-pyruvic acid. (C) IAOx branch. CYP79B2, cytochrome P450 (79B2); CYP79B3, cytochrome P450 (79B3); IAOx, indole-3-acetaldoxime; IG, indole-3-glucosinolate; CAM, camalexin; IAN, indole-3-acetonitrile; NIT1-3, nitrilase1-3; IAM, indole-3-acetamide; AMI1, indole-3-acetamide hydrolase-1. (D) Trp-independent (Trp-I) branch with IGP or indole as the starting substrate. Dashed lines indicate the gene or enzyme has not been identified.

In the IPA pathway, the TAA1 family of Trp aminotransferases converts Trp into IPA (Stepanova et al. 2008; Tao et al. 2008; Yamada et al. 2009; Zhou et al. 2011) followed by the direct conversion to IAA by the YUCCA (YUC) family of flavin monooxygenase-like (FMO) enzymes (Mashiguchi et al. 2011; Won et al. 2011; Zhao 2012). Both TAA1 and YUC gene families are highly conserved across the plant kingdom and TAA1 and YUC genes are coexpressed in a spatial and temporal manner (Stepanova et al. 2011) consistent with TAA1 and YUC being the primary route for IAA biosynthesis.

In Arabidopsis, IAOx is produced by CYP79B2 and CYP79B3 and is an intermediate in the synthesis of IAA and two classes of defense compounds, indole-glucosinolates (IGs) and camalexin (Hull et al. 2000; Zhao et al. 2002; Glawischnig et al. 2004; Sugawara et al. 2009). Disruption of the metabolic flux from IAOx to IGs in the sur1-1 (Mikkelsen et al. 2004) or sur2-1 (Morant et al. 2010) mutant causes an increase in conversion of IAOx to IAA, supporting this pathway as a route to IAA. However, IAOx is likely a minor contributor to overall IAA production under most conditions because the double cyp79B2 cyp79B3 mutant produces no IAOx and appears wild type (WT) (Zhao et al. 2002; Sugawara et al. 2009). In addition, IAOx production appears to be limited to the Brassicaceae family (Sugawara et al. 2009; Nonhebel et al. 2011).

For the IAM pathway, IAM is produced from Trp, perhaps by a monooxygenase, similar to what is used by bacteria (Woodward and Bartel 2005; Ljung 2013), although in Arabidopsis the IAOx pathway may also produce IAM (Sugawara et al. 2009). In Arabidopsis IAM can be converted to IAA in vitro by IAM hydrolase encoded by AMI1 (Pollmann et al. 2003). While the IAM pathway has been proposed to exist throughout the plant kingdom (Mano et al. 2010), its importance in overall IAA metabolism remains to be characterized genetically.

Evidence of the Trp-I IAA synthesis pathway comes primarily from stable isotope labeling studies of Trp auxotrophs in maize (Wright et al. 1991) and Arabidopsis (Normanly et al. 1993), IAA homeostasis mutants (Quint et al. 2009) and in wild-type Lemna gibba under environmental perturbations (Rapparini et al. 2002). Indole or indole-3-glycerol phosphate, precursors of Trp, have been proposed as the branch point to Trp-I IAA biosynthesis (Ouyang et al. 2000) and recent work suggests that a cytosolic tryptophan synthase α-subunit called indole synthase (INS) participates in Trp-I IAA synthesis in embryogenesis in Arabidopsis (Wang et al. 2015). Indirect evidence comes from analysis of the Arabidopsis alf3-1 mutant, which produces lateral roots that arrest growth and die after the initial stages of lateral root formation (Celenza et al. 1995). However, the alf3-1 phenotype is rescued by medium supplemented with indole but not with Trp (Celenza et al. 1995), suggesting that the rescue is via Trp-I IAA biosynthesis.

In this work we describe the identification of indole severe sensitive 1 (iss1), a mutant that displays indole-dependent adventitious roots, leaf narrowing, and longer petioles due to elevated Trp-I IAA synthesis. Indole-grown iss1 mutants have elevated IAA and Trp levels as well as alterations in phenylpropanoid metabolism. ISS1 encodes an aromatic aminotransferase (AroAT); an allele of ISS1, named vas1, was identified previously in a screen for suppressors of taa1 and was found to carry out the conversion of IPA to Trp using methionine as an amino donor (Zheng et al. 2013). The results presented herein provide evidence that ISS1/VAS1 plays a role in Trp catabolism in addition to its role in modulating IPA levels and can function as a more general AroAT.

Materials and Methods

Plant materials and growth conditions

All plants listed are in the Columbia (Col-0) background except iss1-2, which is in the Wassilewskija (Ws) background. iss1-1 was identified by screening pools of random T-DNA insertion lines in the Col-0 background [CS76502, CS76504, CS76506, and CS76508; Arabidopsis Biological Resource Center (ABRC) (http://abrc.osu.edu/)] grown on plant nutrient medium containing 0.5% (w/v) sucrose (PNS) and solidified with 0.6% agar (Haughn and Somerville 1986) containing 80 μM indole (PNS + indole). iss1-2 was identified by screening EMS-mutagenized Wassilewskija (Ws) seeds on PNS + indole. The wei8-4 alleles of TAA1 (SALK_022743C) and trp2-1 (CS8327) were obtained from the ABRC.

For positional cloning, the homozygous iss1-2 mutant in the Ws background was crossed to WT (Col-0) and the F₁ progeny were allowed to self-fertilize. The F₂ progeny were grown on PNS + indole and ∼1100 plants displaying the iss1 phenotype were selected as a mapping population. Microsatellite markers were used to isolate the iss1 locus to a region on chromosome 1 between BAC clones F18B13 and F5I6 using primers listed in Supporting Information, Table S1.

To generate mutant combinations, strains were crossed and the F₁ progeny were allowed to self-fertilize. Putative double or triple mutants were identified in subsequent generations by PCR genotyping (Table S2).

35S::ISS1 constructs were generated using the Univector Plasmid-Fusion System (UPS) (Liu et al. 1998) using the pKYLX-myc9-loxP host vector (Guo and Ecker 2003). For plant transformation, the floral dip method was used (Clough and Bent 1998) and three homozygous 35S::ISS1 lines were selected by kanamycin resistance.

Surface sterilized seeds were grown on solidified PNS medium. Unless otherwise noted, PNS + indole contained 80 μM indole; Trp, 5-methyl-tryptophan (5MT), or p-fluorophenylalanine (PFP) supplements are as indicated. Plants were grown at 22° under constant light using a yellow low pass filter (Stasinopoulos and Hangarter 1990) at a light intensity of 20–40 μE m⁻²⋅s⁻². For UV-B treatment, plants were incubated at room temperature for 24 hr with a UV-B light intensity of 0.8–1.1 μE m⁻²⋅s⁻². Root growth measurements were performed using ImageJ (http://imagej.nih.gov/ij/) and are reported as the mean root length ± SE with analysis by a two-tailed Student’s t-test.

Quantification of IAA, Trp, Phe, and Tyr and secondary metabolites

Unless otherwise noted, all metabolite measurements are reported as the average of three samples ± SE and analyzed statistically as indicated in the legends for Table 1 and Figure 4, Figure 6, and Figure 9. Solid phase extraction of IAA was performed on 30–75 mg of plant tissue according to Barkawi et al. (2010). Methylated IAA samples were subjected to gas chromatography-selected ion monitoring-mass spectrometry (GC-SIM-MS) analysis as described by Barkawi et al. (2010). A total of 100 pg/μl of authentic methyl-IAA was used to determine the retention time (RT). The molecular and quinolinium ions of endogenous IAA and the [¹³C₆]-IAA internal standard (Cambridge Isotope Laboratories, Tewksbury, MA), are listed in Table S3. Levels of free IAA were quantified using the quinolinium ion as described by Barkawi et al. (2010).

Table 1. iss1-1 mutants have an indole-dependent increase in free IAA levels.

Genotype/treatment	IAA ng/g FW	±SE	Fold change
WT/PNS	18.9	1.7	—
WT/PNS + indole	31.3	6.9	1.6
iss1-1/PNS	24.7	3.8	1.3
iss1-1/PNS + indole	157.1	5.1	8.3*

Shown are free IAA levels extracted from 21-day-old WT and iss1-1 plants grown on PNS or PNS + indole. Fold change is given relative to WT grown on PNS. *P < 0.0001 using a two-tailed Student’s t-test. Other fold changes are not significant.

Figure 4.

iss1-1 has an indole-dependent increase in Trp-independent IAA biosynthesis. (A) The ratio of [¹⁵N] ANA to [¹³C₁₁ ¹⁵N₂] Trp incorporation into Trp (white bars) and IAA (black bars) is shown. Only the ratio of incorporation into IAA for indole-grown iss1-1 seedlings is significantly different from other samples (*P < 0.01 two-tailed Student’s t-test). Details of the labeling procedure are described in Materials and Methods. (B) For the samples used in A, the percentage IAA made through a Trp-I pathway was determined by calculating the increased enrichment of [¹⁵N] into IAA relative to Trp as described in Materials and Methods. Only indole-grown iss1-1 seedlings showed a significant difference in Trp-I IAA synthesis (*P < 0.01 two-tailed Student’s t-test).

Figure 6.

The iss1-1 mutant has altered Trp metabolism. Relative levels of Trp (A), Phe (B), and Tyr (C) are presented normalized to WT plants grown on PNS. Samples were extracted from whole 15-day-old Arabidopsis seedlings grown on PNS (white bars) or PNS + indole (black bars) and metabolites were detected by GC-MS following derivitization as described in Materials and Methods. The Trp levels in WT grown on PNS + indole are significantly different from WT plants grown on PNS (P < 0.005 using a two-tailed Student’s t-test). Trp levels in iss1-1 grown on PNS + indole are significantly different than other samples (P < 0.0001 using two-tailed Student’s t-test). Both WT and iss1-1 grown on PNS + indole had higher levels of Phe (P < 0.0005 two-way ANOVA) and Tyr (P < 0.005 two-way ANOVA) compared to WT grown on PNS.

Figure 9.

The iss1 mutant has altered phenylpropanoid metabolism. Relative levels of coniferin (A) and flavonoids (B) are presented normalized to WT plants grown on PNS. Samples were extracted from whole 15-day-old Arabidopsis seedlings grown on PNS (white bars), PNS + indole (black bars), or PNS following 24-hr UV-B exposure (gray bars) and detected as described in Materials and Methods. For coniferin, upper brackets indicate a significant decrease in coniferin in iss1-1 compared to WT (***P < 0.0005 two-way ANOVA). Lower bracket indicates a significant increase in coniferin levels in iss1-1 when grown on indole compared to PNS (*P < 0.005 two-tailed Student’s t-test). For flavonoids, a significant increase was found after UV-B treatment for both WT and iss1-1 (P < 0.0001 two-way ANOVA).

Extraction and analysis of Trp, Phe, and Tyr were performed according to Chen et al. (2010) using methyl chloroformate derivitization. For Trp and Phe quantification, known amounts of [²H₅]-Trp, and [²H₅]-Phe (Cambridge Isotope Laboratories) were included as internal standards. Tyr was quantified using a standard curve of unlabeled Tyr. The molecular and major fragment ions of Trp, [²H₅]-Trp, Phe, [²H₅]-Phe, and Tyr are listed in Table S3.

Glucosinolates were isolated from methanol extracts using anion exchange (DEAE Sephadex A25) and converted to desulfoglucosinolates by adding aryl sulfatase as described (Brown et al. 2003). HPLC of desulfoglucosinolates was carried out using a Waters 2795 HPLC equipped with a Waters 2996 photodiode array detector and a Phenomenex Luna 5 micrometer 250 × 4.6 mm C18 column. Desulfoglucosinolates were confirmed by comparing the RT and UV spectra to purified standards and quantified at 229 nm relative to an external sinigrin standard (Brown et al. 2003).

Phenylpropanoids were quantified from methanol extracts using a Waters 2795 HPLC equipped with a Waters 2996 photodiode array detector and a Phenomenex Luna 5 micrometer 250 × 4.6 mm C18 column. The mobile phase was a gradient of water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B) as follows: 95% A; 1–25 min, linear gradient to 40% A; 25–28 min, linear gradient to 100% B; 28–33 min, 100% B; 33–37 min linear gradient to 95% A; 37–42 min 95% A. Detection of separated compounds was monitored using full spectrum detection from 200–400 nm. Compounds were identified by comparison of RT and UV spectra compared to previously published results (Hemm et al. 2004; Kerhoas et al. 2006) and quantified using chromatographs extracted at 260 nm relative to WT control. Peak areas were normalized to wet tissue weight. For UV-B inducible flavonoid quantification, the peak areas of the two most abundant UV-B responsive flavonoids were summed.

The mass of coniferin was confirmed by HPLC combined with high-resolution mass spectrometry (HRMS) obtained on a Waters Q-ToF (hybrid quadrupole/time-of-flight) API US system (Waters, Milford, MA) by electrospray (ESI) in the [positive] mode. Mass correction was done by an external reference using a Waters Lockspray accessory. HPLC was as for phenylpropanoid quantification. The MS settings were: capillary voltage = 3 kV, cone voltage = 35, source temperature = 120°, and dissolvation temperature = 350°.

Dual stable isotope labeling

Plants were grown on PNS + indole plates until the iss1-1 phenotype was observed, typically 20 days after germination. A total of 4–6 WT plants or 10–12 mutant plants were transferred to a sterile six-well tissue culture plate. Labeling was initiated by adding 3 ml of liquid PNS containing 10 μM [¹⁵N] ANA (Cambridge Isotope Laboratories) and 10 μM [¹³C₁₁ ¹⁵N₂] TRP (Cambridge Isotope Laboratories), and returning the plates to the growth chamber for 12 hr. (Because [¹⁵N] ANA and [¹³C₁₁ ¹⁵N₂] TRP were dissolved in 2-propanol, a mock labeling of 2-propanol alone was done to determine that the addition of the solvent did not perturb IAA synthesis.) Tissue was then transferred to a mesh screen that was stretched over a beaker and rinsed with sterile deionized H2O to remove excess label and media. The tissue was briefly patted dry with paper towels and then weighed, frozen on dry ice, and stored at −80°. IAA and Trp extraction and quantification were performed as described above with 5 ng [¹³C₆]-IAA and 1 μg of [²H₅]-Trp added as internal standards. The quinolinium ion of each isotopomer is listed in Table S3 and was used for isotope enrichment analysis. The percent incorporation of each isotopomer for Trp and each isotopomer for IAA was calculated from the total amount of Trp and IAA, respectively (using the internal standards [¹³C₆] IAA and [²H₅]Trp).

Using the percent incorporation for each isotopomer, the percent IAA_Trp-D and IAA_Trp-I were calculated using the following formulas as previously described (Liu et al. 2012): IAA_Trp-D = [¹³C₁₀ ¹⁵N]IAA + [¹⁵N]IAA_Trp-D and IAA_Trp-I = [¹⁵N]IAA_Trp-I. Since [¹³C₁₁ ¹⁵N₂]Trp can only be converted to [¹³C₁₀ ¹⁵N]IAA via the Trp-D pathway, any [¹³C₁₀ ¹⁵N]IAA detected is from the Trp-D pathway. Additionally, any [¹⁵N]Trp converted to [¹⁵N]IAA from the Trp-D pathway should be proportional to the conversion of [¹³C₁₁ ¹⁵N₂]Trp to [¹³C₁₀ ¹⁵N]IAA. Therefore, the ratio of Trp-dependent [¹⁵N]IAA to [¹³C₁₀¹⁵N]IAA is equal to the ratio of [¹⁵N]Trp to [¹³C₁₁ ¹⁵N₂]Trp. Or [¹⁵N]IAA_Trp-D = [¹³C₁₀ ¹⁵N]IAA x [¹⁵N]Trp/[¹³C₁₁ ¹⁵N₂]Trp. Once the amount of [¹⁵N]IAA_Trp-D was calculated, the amount of [¹⁵N]-IAA_Trp-I can be calculated by subtracting [¹⁵N]IAA_Trp-D from [¹⁵N]IAA_total or [¹⁵N]IAA_Trp-I = [¹⁵N]IAA_total − [¹⁵N]IAA_Trp-D.

Heterologous expression constructs

For the expression of Arabidopsis cDNAs in yeast and plants, the Univector Plasmid-Fusion System (UPS) was used (Liu et al. 1998). UPS-compatible Arabidopsis cDNAs and expression vectors (pHOST) are listed in Table S4. The iss1-2 mutation was introduced into ISS1 cDNA carried on an Escherichia coli plasmid by using PCR and Gibson assembly according to supplier’s instructions (New England Biolabs, Ipswich, MA).

Complementation of the yeast aro8 aro9 Phe and Tyr auxotrophy

Yeast strains 11965 and 14569 in the S288C background (GE Dharmacon, Lafayette, CO) were used to generate a haploid aro8::HygR aro9::G418R double mutant (MPY1; Table S5) for complementation assays. ISS1, iss1-2, TAA1, and MEE17 cDNAs recombined into pHY326-loxH were transformed into MPY1 and grown on yeast minimal (SD) medium (Ausubel et al. 1994–1998), with or without 50 mg/liter Phe and 30 mg/l Tyr, and supplemented with either 2% glucose or 2% galactose. Complementation was assayed following incubation at 30° for 48 hr.

For pseudohyphal growth assays, diploid yeast strains PY652 and PY653 (Σ1278B background) were obtained from Reeta Prusty-Rao (Worcester Polytechnic Institute, Worcester, MA) and strain MPY5 was derived from PY653 (Table S5). To induce pseudohyphal growth, yeast was grown on solid synthetic low ammonium dextrose (SLAD) medium with or without Phe and Tyr supplemented with either 0.5% glucose or 0.5% galactose. Pseudohyphal growth was observed following incubation at 30° for 72–96 hr.

HPLC-based aromatic aminotransferase enzyme assay

ISS1 and iss1-2 cDNAs were subcloned into the IPTG-inducible pMAL-c4X MBP expression vector (New England Biolabs). MBP fusion protein was purified according to the supplier’s instructions (New England Biolabs). Purified protein was quantified by Bradford assay. The presence of recombinant protein was confirmed by SDS-PAGE and native PAGE analysis. An empty pMAL-c4X vector was used as a control.

AroAT activity was measured using a method adapted from Stepanova et al. (2008) as follows. The reaction buffer consisted of 0.1 M borate pH 6.5, 4 mM EDTA, 0.2 mM pyridoxal 5′-phosphate (PLP), 5 mM amino donor, and 0.5 mM amino acceptor. The reaction was initiated by adding 5–60 μg/200 μL purified MBP-tagged protein. A zero-time point was taken immediately after adding the protein. The reaction was incubated at 37° for 30, 60, and 120 min. At each time point, 200 μl was removed from the reaction volume and added to a 1.5-ml microcentrifuge tube containing 200 μl methanol to stop the reaction. The stopped reaction was briefly vortexed and centrifuged for 30 sec. After centrifugation, 100 μl was analyzed for production of the appropriate aromatic amino acid and α-keto acid using a Waters 2795 HPLC equipped with a Waters 2996 photodiode array detector and a Phenomenex Luna 5 micrometer 250 × 4.6 mm C18 column. The mobile phase was a gradient of water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B) as follows: 90% A; 1–5 min; 5–10 min, linear gradient to 85% A; 10–20 min, linear gradient to 83% A; 20–30 min, linear gradient to 75% A; 30–40 min, linear gradient to 65% A; 40–50 min, linear gradient to 55% A; 55–58 min, linear gradient to 45% A; 58–60 min, linear gradient to 10% A; 61–66 min 90% A. Detection of separated compounds was monitored at 254 nm and 378 nm. Products were compared to the RT of authentic standards for Trp, Phe, Tyr, indole-3-pyruvate, phenylpyruvate, and 4-hydroxyphenylpyruvate.

Data availability

Strains and plasmids are available upon request. Table S1 lists oligonucleotide sequences used in the map-based cloning of ISS1. Table S2 lists oligonucleotide sequences used for plant genotyping. Table S3 lists molecular and fragment ions of derivitized IAA and aromatic amino acids. Table S4 lists plasmids used. Table S5 lists Saccharomyces cerevisiae stains used. Table S6 shows indole glucosinolate quantification. Table S7 shows specific activities for purified MBP-ISS1. Figure S1 shows the strategy used for map-based cloning of ISS1. Figure S2 shows rescue of iss1-1 by overexpression of the ISS1 cDNA. Figure S3 shows the amino acid sequence alignment of ISS1 with other fold type-I aminotransferases.

Results

The iss1-1 mutant displays an indole-dependent high auxin phenotype

Because the branch point between Trp-I IAA biosynthesis and Trp-D IAA biosynthesis is proposed to be an indolic precursor to Trp (Woodward and Bartel 2005), we reasoned that mutants displaying a high-IAA phenotype when supplemented with indole, but not Trp, would have an increase in Trp-I IAA biosynthesis. We screened through a publicly available single T-DNA insertion collection from which one seedling, indole severe sensitive (iss1-1) developed a high IAA phenotype that was observable after 2–3 weeks growth on PNS medium supplemented with 80 μM indole (PNS + indole). Compared to the wild-type Columbia plants, the iss1-1 mutant displayed increased root growth inhibition by indole (Figure 2, A and B) and phenotypes consistent with increased auxin, including increased adventitious root formation, narrow leaves, and elongated petioles (Figure 3). iss1-1 appeared normal when grown on PNS medium or PNS supplemented with 80 μM Trp (data not shown). After backcrossing to Col-0, the indole sensitivity segregated as a single recessive Mendelian trait. The iss1-2 allele was recovered from an EMS-mutagenized M2 population in the Ws accession, had a similar phenotype to iss1-1, and failed to complement iss1-1.

Figure 2.

The iss1-1 mutant displays altered root growth sensitivity to indole, 5-methyl tryptophan, and p-fluorophenylalanine. The 15-day-old average root length normalized to WT is shown for WT (white bars) and iss1-1 (black bars) plants grown on unsupplemented PNS (A) or PNS + indole (B), PNS + 1 μM 5MT (C) or PNS 15 μM PFP (D). iss1-1 root growth is significantly different from WT for B and C (**P < 0.001 two-tailed Student’s t-test) and D (*P < 0.005 two-tailed Student’s t-test).

Figure 3.

The iss1-1 mutant displays an indole-dependent high IAA phenotype. Depicted are representative 18-day-old WT and iss1-1 mutant plants grown on PNS (A, B, G, and H), or PNS containing 40 μM indole (C, D, I, and J), or 80 μM indole (E, F, K, and L). Shown are dark field images of the seedling (A, C, E, G, I, and K) with an arrowhead indicating the root/hypocotyl junction and bright field images of the hypocotyl (B, D, F, H, J, and L).

The iss1-1 mutant has an indole-dependent increase in IAA due to elevated Trp-I IAA biosynthesis

To determine if the high auxin phenotype observed in iss1-1 is indeed caused by an increase in IAA levels, we quantified IAA using GC-SIM-MS. On PNS medium, IAA levels were similar between iss1-1 and WT (Table 1). However, on PNS + indole medium, iss1-1 accumulated approximately eightfold greater IAA than WT, consistent with the high IAA phenotype (Table 1).

Because the iss1-1 mutant displayed phenotypes consistent with elevated levels of IAA when grown on indole but not when grown on Trp, we hypothesized that the increase in IAA was due to an increase in Trp-I IAA biosynthesis relative to Trp-D IAA biosynthesis. To distinguish between Trp-D and Trp-I IAA biosynthesis, we fed 20-day-old iss1-1 and WT plants grown on either PNS or PNS + indole media the stable isotopes [¹⁵N] anthranilate (ANA) and [¹³C₁₁ ¹⁵N₂] Trp for 12 h. The rationale for dual labeling is that any [¹³C₁₀ ¹⁵N] detected in IAA results only from the conversion of [¹³C₁₁¹⁵N₂] Trp via a Trp-D IAA pathway. In contrast, [¹⁵N] IAA derives from either [¹⁵N] ANA first being converted into [¹⁵N] Trp and then into [¹⁵N] IAA via a Trp-D IAA pathway or by the conversion of [¹⁵N] ANA into [¹⁵N] IAA via a Trp-I IAA pathway (and thus bypassing Trp). Therefore by quantifying the relative conversion of [¹³C₁₁¹⁵N₂] Trp to [¹³C₁₀ ¹⁵N] IAA and comparing this to the relative amounts of [¹⁵N] Trp and [¹⁵N] IAA, the percentage of labeled IAA coming from the Trp-D IAA and Trp-I IAA can be determined (Quint et al. 2009; Liu et al. 2012). In addition, if IAA is derived only from a Trp-D pathway, the proportion of IAA (the product) labeled with [¹⁵N], cannot be greater than the proportion of Trp (the precursor) labeled with [¹⁵N]. If the proportion of IAA labeled with [¹⁵N] is greater than the proportion of Trp labeled with [¹⁵N] then Trp cannot be the sole precursor for the [¹⁵N]-labeled IAA.

We first examined relative isotopic enrichment of the Trp pool and found that both iss1-1 and WT grown on PNS or PNS + indole media had similar ratios of [¹⁵N] Trp to [¹³C₁₁ ¹⁵N₂] Trp, indicating that the uptake of [¹⁵N] ANA relative to [¹³C₁₁ ¹⁵N₂] Trp is not changed by genotype or growth condition (Figure 4A). Thus the relative access of [¹⁵N] ANA and [¹³C₁₁ ¹⁵N₂] Trp to the total Trp pool was not altered by the genotype or growth condition. We also examined relative isotopic enrichment of the IAA pool and in contrast to what we found for the Trp pool, there was a significant increase in the ratio of [¹⁵N] IAA to [¹³C₁₀ ¹⁵N] IAA for iss1-1 plants grown on PNS + indole medium compared to iss1-1 grown on PNS medium or WT grown in either condition (Figure 4A).

By calculating the increased enrichment of [¹⁵N] into IAA relative to Trp, the amount of IAA derived from Trp-I IAA synthesis can be calculated. On PNS, Trp-I IAA synthesis accounted for 38% of the IAA produced in WT and 15% of the IAA produced in iss1-1 (Figure 4B). On PNS + indole medium, the amount of IAA produced via Trp-I IAA synthesis decreased to 18% for WT seedlings, while Trp-I IAA synthesis increased to 75% in indole-grown iss1-1 (Figure 4B), suggesting that the high levels of IAA in indole-grown iss1-1 derive primarily from the Trp-I IAA pathway. In wild-type plants, the addition of indole did not cause a significant increase in Trp-I IAA and instead showed a trend toward a decrease in Trp-I IAA, suggesting that in WT plants added indole is metabolized to Trp.

In addition to being able to measure the relative isotopic enrichment of Trp and IAA, the dual labeling experiment also allowed us to quantify the IAA and Trp pools of WT and iss1-1 grown with and without an indole supplement. As we found previously, iss1-1 plants showed an ∼10-fold increase in IAA when grown on indole, whereas WT plants showed no change when grown with and without an indole (Table 2). We found that growth on PNS + indole increased unlabeled Trp severalfold for WT plants as might be expected. Surprisingly, as discussed further below, indole-grown iss1-1 plants showed a >100-fold increase in Trp (Table 3); nonetheless as noted above, the ratio of [¹⁵N] Trp to [¹³C₁₁ ¹⁵N₂] Trp (white bars in Figure 4) did not change under this condition. Thus, we did not observe a dilution effect from the large amounts of unlabeled Trp in iss1-1 grown in the presence of indole.

Table 2. Stable isotope incorporation into IAA.

Genotype/treatment	IAA (unlabeled) (ng/g)	[15N] IAA (ng/g)	[13C1015N] IAA (ng/g)
WT/PNS	17.0 ± 4.9	3.2 ± 0.2	1.1 ± 0.1
WT/PNS + indole	14.5 ± 3.8	2.0 ± 0.2	1.1 ± 0.1
iss1-1/PNS	17.2 ± 3.4	7.7 ± 1.3	4.5 ± 1.4
iss1-1/PNS + indole	162.2 ± 21.5	22.5 ± 2.3	2.8 ± 0.6

Shown are average amounts ±SD for unlabeled, [¹⁵N] IAA, and [¹³C₁₀¹⁵N] IAA found in the seedlings used in Figure 4.

Table 3. Stable isotope incorporation into Trp.

Genotype/treatment	Trp unlabeled (μg/g)	[15N] Trp (μg/g)	[13C1115N2] Trp (μg/g)
WT/PNS	8.8 ± 0.1	16.8 ± 4.0	10.6 ± 4.0
WT/PNS + indole	34.5 ± 6.3	9.8 ± 0.8	6.3 ± 1.0
iss1-1/PNS	5.9 ± 1.2	14.8 ± 7.1	8.4 ± 4.6
iss1-1/PNS + indole	1071.4. ± 38.5	41.8 ± 0.2	25.5 ± 0.3

Shown are average amounts ±SD for unlabeled, [¹⁵N] Trp, and [¹³C₁₁¹⁵N₂] Trp found in the seedlings used in Figure 4.

We also examined IAA synthesis pathway use genetically. Both the IPA and IAOx Trp-D IAA pathways can be reduced or eliminated, respectively, by mutations in key steps in each pathway (Zhao et al. 2002; Stepanova et al. 2008; Tao et al. 2008). Thus we asked if the taa1 mutation or the cyp79B2 cyp79B3 double mutant could suppress iss1-1 indole sensitivity. As seen in Figure 5, the iss1-1 taa1 double mutant and the iss1-1 cyp79B2 cyp79B3 triple mutant remained sensitive to indole.

Figure 5.

Mutations in Trp-dependent IAA synthesis pathways do not suppress the iss1-1 indole sensitivity. Shown are representative 18-day-old plants of the genotypes shown, grown on PNS (top row) or PNS + indole (bottom row).

The iss1-1 mutant has altered Trp metabolism

The observation from the dual-labeling experiment that iss1-1 accumulated Trp when grown on PNS + indole led us to quantify Trp levels as well as the other aromatic amino acids. Similar to our previous results, iss1-1 grown on PNS + indole medium accumulated 174-fold higher levels of Trp, suggesting an overall disruption of Trp metabolism (Figure 6). In contrast, Phe and Tyr levels were similar between WT and iss1-1 plants when grown on PNS or PNS + indole media. Independent of the genotype, we found that growth on PNS + indole relative to PNS medium caused a twofold increase in Phe and a fourfold increase in Tyr (Figure 6).

Because the iss1-1 mutant accumulated Trp significantly when grown on PNS + indole medium, we predicted that Trp catabolism is altered in the iss1-1 mutant. Therefore we compared iss1-1 growth to WT in the presence of the toxic Trp analog 5MT. Changes in 5MT sensitivity reflect changes in Trp metabolism, such that mutants with decreased Trp catabolism show increased 5MT sensitivity compared to WT (Zhao et al. 2002; Celenza et al. 2005; Tao et al. 2008). The iss1-1 mutant had an increased sensitivity to 5MT compared to WT consistent with a decrease in Trp catabolism (Figure 2C).

The trp2 mutant partially suppresses the iss1-1 phenotype

We hypothesized that iss1-1 has altered Trp metabolism, as evidenced by both elevated IAA and Trp levels when iss1-1 was grown in the presence of indole and its increased sensitivity to 5MT. Changes in both IAA and Trp suggest a model in which the iss1-1 mutation primarily causes a decrease in Trp catabolism and this buildup of Trp leads to increased usage of a Trp-I IAA pathway. Alternatively iss1-1 simply could directly deregulate Trp-I IAA synthesis; however if this was the case, it is not obvious why Trp levels would increase in addition to IAA. To distinguish between these models, we asked if Trp synthesis was required for the indole-dependent iss1-1 high IAA phenotype and used double mutant analysis to examine the interaction between iss1-1 and the Trp synthase-β (TSB1) mutant trp2-1. The trp2-1 mutant is unable to convert indole to Trp. trp2-1 accumulates indole (Last et al. 1991) and exhibits elevated Trp-I IAA biosynthesis (Normanly et al. 1993). If increased Trp-I IAA synthesis in indole-grown iss1-1 plants was the result of a buildup of Trp, then by blocking the conversion of indole to Trp (by way of the trp2-1 mutation), the iss1-1 mutant phenotype should be suppressed by trp2-1. On the other hand, if the iss1-1 mutant directly increases Trp-I synthesis, the trp2-1 mutation should enhance the iss1-1 phenotype when grown on indole and perhaps show elevated IAA phenotypes even in the absence of indole. We found that trp2-1 (Figure 7) suppressed the indole sensitivity phenotype in iss1-1. In particular, unlike the single iss1-1 mutant, iss1-1 trp2-1 plants had normal roots and leaves when grown on PNS + indole medium, suggesting that the ability to convert added indole to Trp by TSB1 is needed for the iss1-1 high IAA phenotype. Consistent with this observation, we found that growth of iss1-1 trp2-1 on PNS + indole + Trp medium restored the high IAA phenotype.

Figure 7.

Analysis of the iss1-1 trp2-1 double mutant phenotype. Shown are representative 21-day-old WT, iss1-1, iss1-1 trp2-1, and trp2-1 plants grown on PNS (top row), or PNS containing 80 μM Trp (second row), 80 μM indole (third row), or 80 μM Trp, 80 μM indole (bottom row).

To confirm that elevated Trp levels are important for the iss1-1 phenotype, we compared the severity of the iss1-1 phenotype of plants grown on PNS media with either added indole, Trp, or indole together with Trp. iss1-1 plants grown on PNS + indole + Trp medium had a much more pronounced high-IAA phenotype compared to iss1-1 plants grown on PNS + indole medium, consistent with elevated Trp levels contributing to the iss1-1 high IAA phenotype (Figure 7).

iss1-1 makes normal levels of indolic glucosinolates

The increased sensitivity of iss1-1 to 5MT as well as the indole-dependent increase in Trp could be explained by reduction in catabolism of Trp into IGs. IGs are an important sink for Trp in Arabidopsis, and mutants with reduced IGs display increased sensitivity to 5MT (Celenza et al. 2005; Bender and Celenza 2009). The iss1-1 mutant had similar IG levels to WT whether grown on PNS or PNS + indole media, indicating that conversion of Trp to IGs is not altered in the iss1-1 mutant. Interestingly both iss1-1 and WT showed a 1.5-fold increase in total IG levels when grown on PNS + indole medium (Table S6).

ISS1 encodes an aminotransferase

Using map-based cloning (Lukowitz et al. 2000) the ISS1 gene was identified as At1g80360 (Figure S1). Comparison between the iss1-1 and Col-0 At1g80360 genomic sequences revealed a complex mutation beginning at base 335 of the coding sequence. This mutation consists of a 20-base deletion and a 51-base insertion of T-DNA that would result in a severely truncated protein when expressed (Figure S1). Sequencing of the iss1-2 allele revealed a single point mutation that changes arginine 362 to a Trp residue in the predicted protein (Figure S1). Introduction of an expression construct with the At1g80360 cDNA driven by the CaMV 35S promoter rescued the iss1-1 mutant from the indole-dependent high IAA phenotype, confirming the identity of ISS1 (Figure S2).

At1g80360 encodes a predicted fold-type I aminotransferase (AT), that was recently identified as VAS1, an AT capable of converting IPA to Trp (Zheng et al. 2013). ATs require PLP as a cofactor and catalyze the reversible reaction in which an amino group is transferred from an amino acid donor to a 2-oxo acid acceptor (Jensen and Gu 1996). BLAST analysis revealed that ISS1 is conserved across plant species; however, none of these presumed orthologs have a characterized enzyme activity. In the Arabidopsis genome, there are no ISS1 paralogs; however, the proteins most similar to ISS1 are the bifunctional aspartate/prephenate AT encoded by MEE17 (24% amino acid identity; Figure S3) and the Tyr AT TAT3 (23% amino acid sequence identity; Figure S3). ISS1 shares only 12% amino acid sequence identity with TAA1, another Arabidopsis Trp AT (Figure S3).

Heterologous expression of ISS1 rescues the yeast Phe and Tyr auxotroph mutant aro8 aro9

Because the ISS1 protein is most similar to enzymes involved in aromatic amino acid biosynthesis and iss1-1 mutants have altered Trp metabolism, we hypothesized that ISS1 functions as a general AroAT. To test this hypothesis, we asked if heterologous expression of ISS1 in aro8 aro9 mutant yeast could rescue the Phe and Tyr auxotrophy. In yeast, the AroATs encoded by ARO8 and ARO9 convert phenylpyruvate (PPY) to Phe and 4-hydroxyphenylpyruvate (4-HY) to Tyr (Iraqui et al. 1998; Urrestarazu et al. 1998). As seen in Figure 8A, galactose-induced ISS1 expression was able to rescue the aro8 aro9 Phe and Tyr auxotropy. The iss1-2 mutant cDNA did not rescue the aro8 aro9 mutant, indicating that the point mutant can no longer function as an AroAT (Figure 8A). TAA1 weakly rescued the aro8 aro9 under galactose induction, consistent with its in vitro activity primarily using Trp as a substrate (Tao et al. 2008). In addition, MEE17 failed to rescue the yeast mutant (Figure 8A), consistent with its reported activity with prephenate, but not with Phe or Tyr as substrates (Maeda et al. 2010).

Figure 8.

Heterologous expression of ISS1 rescues the aro8 aro9 mutant yeast. (A) pHY326-loxH (vector) or pHY326-loxH expressing ISS1, iss1-2, TAA1, or MEE17 were transformed into aro8 aro9 double mutant haploid yeast (MPY1) in the S288C background and grown on SC medium –Phe and –Tyr containing glucose (top panels) or galactose (bottom panels) for 48 hr at 30°. White bar, 5 mm. (B) The same plasmids used in A were transformed into aro8 aro9 double mutant diploid yeast (MPY5) in the Σ1278b background and grown on SLAD medium with either glucose (not shown) or galactose for 72 hr at 30°. WT is strain PY652. Black bar, 20 μm.

Many species of fungi, including certain backgrounds of Saccharomyces cerevisiae, are able to transition from a vegetative state to filamentous state under conditions of low nitrogen and high cell density (Gimeno et al. 1992). High cell density and low nitrogen availability cause the increased production of tryptophol (Trp-OH) and phenylethanol (Phe-OH) by deamination of Trp and Phe, respectively, which in turn causes the transition from vegetative growth to filamentous growth in diploid yeast (Chen and Fink 2006). The aro8 aro9 double mutant fails to produce Trp-OH or Phe-OH, and therefore is unable to transition to filamentous growth (Chen and Fink 2006). To test if ISS1 also functions catabolically as an AroAT, we grew aro8 aro9 yeast expressing ISS1 on low nitrogen medium (SLAD) with galactose as carbon source and observed a weak rescue of the filamentous growth phenotype (Figure 8B), indicating that ISS1 is able to participate both in the biosynthesis of Phe and Tyr and in the catabolism of Phe and/or Trp. aro8 aro9 yeast expressing iss1-2, TAA1, or MEE17 failed to develop filaments (Figure 8B).

Recombinant ISS1 protein has aromatic aminotransferase activity

Mutant alleles of ISS1 called vas1 (reversal of sav3 phenotype 1) have been reported previously (Zheng et al. 2013). These authors determined that VAS1 uses IPA as the amino acceptor to make Trp as a counterbalance to the conversion of Trp to IPA by TAA1. Because the iss1-1 mutant phenotype suggested a role in Trp catabolism and heterologous expression of ISS1 demonstrated a role in Phe/Tyr metabolism, we further characterized AroAT activity of heterologously expressed ISS1. The ISS1 wild-type cDNA and the iss1-2 mutant allele were cloned into the pMAL-c4X vector to express proteins with N-terminal maltose binding protein (MBP) tags. To confirm the fusion proteins were still active as AroATs, the fusion proteins were transformed into the DL39 E. coli mutant. DL39 has mutations in tyrB (Tyr AT), aspC (aspartate AT), and ilvE (branched chain AT) and is auxotrophic for Tyr, Phe, aspartic acid, leucine, isoleucine, and valine (Riewe et al. 2012). Expression of MBP-ISS1, but not MBP-iss1-2 rescued the DL39 Phe and Tyr auxotrophy, confirming that the MBP tag was not hindering the activity of the recombinant enzymes and also confirming that ISS1 has AroAT activity (data not shown).

Following purification using the MBP tag, we confirmed MBP-ISS1 was able to use Trp, Phe, and Tyr as substrates in vitro, consistent with ISS1 being an AroAT. In addition, the reverse reactions were also catalyzed by MBP-ISS1 (Table S7), indicating that ISS1 is an AroAT.

The iss1-1 mutant has an altered phenylpropanoid profile

Our analyses of the iss1-1 mutant indicate a role for ISS1 in Trp metabolism. However, based on the yeast heterologous expression experiments and the in vitro assays, ISS1 is a broader AroAT with activities not limited to only Trp metabolism. To determine if iss1-1 mutant plants also display altered Phe metabolism, we tested iss1-1 for sensitivity to the toxic Phe analog PFP. Like 5MT, PFP is incorporated into proteins instead of Phe and feedback inhibits chorismate mutase (CM) (Palmer and Widholm 1975). We found that iss1-1 plants were more resistant to PFP compared to WT (Figure 2D). Resistance to PFP could be explained by relaxed feedback inhibition or increased Phe turnover.

The fact that Phe levels are not elevated in the iss1-1 mutant relative to WT (Figure 6B) suggests that feedback inhibition is not dramatically altered in iss1-1 mutant plants. However, if Phe turnover was disrupted in iss1-1 mutant plants, then production of Phe metabolites such as phenylpropanoids might be altered. iss1-1 plants grown on PNS medium had a significant decrease in coniferin, a monolignol glucoside, while flavonoid levels were similar to WT (Figure 9). Interestingly, in iss1-1 plants grown on PNS + indole medium, coniferin levels were significantly higher compared to growth on PNS medium, but these levels were still were significantly lower than WT plants when grown on PNS + indole medium.

Exposure to UV-B activates the production of certain phenylpropanoids, including some flavonoids (Ryan et al. 2001; Tohge et al. 2011). To determine whether the iss1-1 mutant is affected in UV-B-inducible flavonoid production, we measured phenylpropanoid levels after 24 hr exposure to UV-B (Figure 9). After UV-B induction, several flavonoids increased ∼12-fold in both WT and iss1-1 plants while coniferin levels were unaffected by UV-B.

Discussion

The iss1 mutant reveals increased use of Trp-independent IAA synthesis

Using a screen designed to find mutants with elevated indole-dependent IAA synthesis, two alleles of ISS1 were identified that displayed a high IAA phenotype when grown on indole but not when grown on Trp. The iss1-1 mutation is a complex insertion/deletion and iss1-2 changes an amino acid conserved in other aminotransferases, thus we suspect that both alleles are null or severe hypomorphs. When grown on indole (but not Trp), both iss1 mutant alleles display narrowed leaves and increased lateral and adventitious root growth, phenotypes consistent with elevated IAA levels.

In support of the indole-dependent growth phenotype, iss1-1 seedlings grown on indole showed an eightfold increase in IAA levels compared to WT grown under the same conditions; however, IAA levels were similar to WT levels when grown on unsupplemented medium. We used dual labeling with [¹⁵N] ANA (an upstream precursor of Trp) and [¹³C₁₁ ¹⁵N₂] Trp to determine that these indole-dependent elevated IAA levels were due to an increase in Trp-I IAA biosynthesis. Our data demonstrate that on unsupplemented medium, the majority of labeled IAA in both iss1-1 and WT is from a Trp-D IAA pathway, consistent with what has been reported previously for wild-type Arabidopsis seedlings (Quint et al. 2009; Yu 2014). This finding suggests that Trp-D IAA biosynthesis is not obviously disrupted in iss1-1 plants under normal lab growth conditions and is consistent with our finding that IAA levels are similar for WT and iss1-1 grown without an indole supplement. In contrast, indole grown iss1-1 seedlings had a higher percentage of [¹⁵N] IAA compared to [¹⁵N] Trp, consistent with increased use of Trp-I IAA synthesis, while WT plants grown on indole used primarily Trp-D IAA synthesis. These labeling results strongly suggest that the iss1-1 mutant has an indole-dependent increase in Trp-I IAA synthesis. Both WT and iss1-1 grown with or without an indole supplement showed nearly identical ratios of [¹⁵N] Trp to [¹³C₁₁ ¹⁵N₂] Trp incorporation into the Trp pool. This finding indicates that uptake of [¹⁵N] ANA is not perturbed in the iss1-1 mutant and is consistent with indole-grown iss1-1 having increased Trp-I IAA synthesis. If the increased ratio of [¹⁵N] IAA to [¹³C₁₀ ¹⁵N] IAA observed in the iss1-1 mutant was due to altered uptake of [¹⁵N] ANA in the iss1-1 mutant, the [¹⁵N] Trp to [¹³C₁₁ ¹⁵N₂] Trp ratio would also have been elevated in the mutant.

Consistent with this model, we found that the iss1-1 taa1 double mutant and the iss1-1 cyp79B2 cyp79B3 triple mutant showed the same degree of indole sensitivity as the single iss1-1 mutant, suggesting that the IPA and IAOx Trp-D IAA synthesis pathways are not responsible for the increased IAA synthesis in indole-grown iss1-1 plants.

ISS1 functions in Trp catabolism

Surprisingly, iss1-1 seedlings also have a 174-fold increase in Trp when grown on indole, indicating that the increased usage of Trp-I IAA synthesis may be due indirectly to a loss of Trp catabolism. Consistent with decreased Trp catabolism, iss1 has increased sensitivity to 5MT. In addition, we found that the indole-sensitive phenotype depends on having a functional Trp synthase-β. The trp2-1 mutation in TSB1 suppresses the iss1-1 indole-dependent growth phenotype consistent with a buildup of Trp being required for increased usage of Trp-I IAA synthesis. This suppression also indicates that the increased Trp-I IAA synthesis is not due simply to a buildup of indole and also suggests that Trp-I IAA synthesis branches off from IGP and not indole. Consistent with high Trp levels being responsible for the iss1 phenotype, growth of the iss1-1 trp2-1 double mutant on a combination of 80 μM Trp and 80 μM indole restored the iss1-1 mutant phenotype (Figure 7). In addition, single iss1-1 mutant seedlings grown under this condition showed a more severe phenotype than iss1-1 mutants grown on 80 μM indole alone (Figure 7). iss1-1 plants grown on 80 μM Trp appeared similar to WT grown in the same condition (Figure 7).

We hypothesize that ISS1’s catabolic function plays a homeostatic role that would be required when rapid shifts in metabolic flux would occur in nonlaboratory conditions. The Trp catabolism pathway that is affected in iss1 mutants is undetermined at this time. In yeast and bacteria, Trp-derived IPA is converted to Trp-OH via the Ehrlich pathway (Hazelwood et al. 2008). While Trp-OH has been detected in cucumber seedlings (Rayle and Purves 1967) more recent studies with peas revealed very low levels of Trp-OH (Quittenden et al. 2009).

Where does ISS1 fit into IAA metabolism?

Mutant alleles of ISS1, called vas1, were identified as suppressors of the taa1 mutant phenotype: failure of hypocotyl elongation in response to shade (Zheng et al. 2013). Under normal growth conditions, vas1 mutants showed subtle increases in hypocotyl and petiole lengths and a significant increase in IAA and IPA levels, suggesting that VAS1 has altered IPA/YUC-dependent IAA metabolism (Zheng et al. 2013). Consistent with a role for VAS1 as a suppressor of TAA activity, the triple vas1-2 sav3-1 tar2-1 mutant had normal hypocotyl length and is fertile (Zheng et al. 2013). Additionally, vas1 mutant plants displayed a fivefold increase in the ethylene precursor ACC, pointing to a link between IAA synthesis and ethylene metabolism (Zheng et al. 2013).

Because the vas1 phenotype is suggestive of a role in the IPA/YUC pathway, recombinant VAS1 enzyme characterization focused on using Trp as an amino donor or IPA as an amino acceptor. Using IPA as the amino acceptor, VAS1 had the highest catalytic (Kcat/Km) activity with methionine as the amino donor followed by Phe as the amino donor (Zheng et al. 2013). Based on the vas1 phenotype and the enzyme kinetics of the recombinant VAS1 protein, it was proposed that VAS1 normally functions as a counterbalance to TAA1-dependent production of IPA while also impacting ethylene synthesis (Zheng et al. 2013). In this view, VAS1 preferentially catalyzes the reverse reaction converting excess IPA to Trp, thus reducing the levels of free IPA.

Our results are distinct from those reported for VAS1 and suggest that this gene has an additional role in Trp (and possibly Phe and/or Tyr) metabolism. We note that the indole-sensitivity phenotype of the iss1-1 mutant is strongest in plants >18 days old, while the vas1 phenotype was characterized in 7- to 9-day-old seedlings. In addition, the aerial portions of 7-day-old vas1 seedlings were found to have a twofold increase in IAA (Zheng et al. 2013), whereas we found that three-week old iss1 whole seedlings had no significant change in IAA levels unless supplemented with indole. We also note that while the expression of VAS1 overlaps with the meristem-specific expression of TAA1 and YUC genes, gene expression databases show that ISS1/VAS1 is expressed at similar levels in meristematic and vegetative tissues, whereas TAA1 and YUC genes are expressed at much lower levels in vegetative tissues compared to meristematic tissues (www.weigelworld.org/resources/microarray/AtGenExpress/). We propose that the differing roles for ISS1/VAS1 in Trp metabolism depend on the tissue type and/or developmental timing (Figure 10). In meristematic tissues, where the expression of TAA1 (Stepanova et al. 2008; Yamada et al. 2009) and YUC genes (Cheng et al. 2007) is highest, ISS1/VAS1 modulates levels of IPA by converting excess IPA into Trp (Zheng et al. 2013). In vegetative tissues, where IAA biosynthetic genes are not highly expressed, ISS1/VAS1 functions in Trp degradation, by converting Trp into IPA, leading into an uncharacterized Trp degradation pathway. Because aminotransferases catalyze reversible reactions, ISS1/VAS1 would be expected to carry out the synthesis or degradation of Trp, depending on the relative concentration of Trp and IPA. Expression of ISS1/VAS1 in different metabolic contexts is consistent with the different roles we propose.

Figure 10.

Model for ISS1/VAS1’s role in Trp homeostasis. In differentiated tissues where IAA synthesis is low, the ISS1/VAS1 aromatic aminotransferase catabolizes excess Trp (red arrow). ISS1/VAS1 functions anabolically in dividing tissues where flux from Trp into auxin (and IPA) is high (green arrow).

Based on its role in modulating Trp and IAA levels, we do not think that ISS1/VAS1 is directly part of a Trp-I IAA biosynthesis pathway; rather we propose that the loss of ISS1/VAS1 activity in Trp homeostasis increases Trp-I IAA synthesis. A number of reports indicate that Trp-I and Trp-D IAA synthesis are under exquisite developmental and environmental regulation (Michalczuk et al. 1992; Ljung et al. 2001; Epstein et al. 2002; Rapparini et al. 2002; Ribnicky et al. 2002; Yamada et al. 2009); perhaps one way of regulating Trp-I IAA synthesis is by modulating Trp metabolic flux.

ISS1 is a member of a third class of plant aromatic aminotransferases

We found that heterologous expression of either ISS1 or TAA1 rescued the Phe and Tyr auxotrophies of yeast and E. coli mutants defective in AroAT activity. However, only ISS1 expression was able to rescue the filamentous growth phenotype of the diploid aro8 aro9 yeast mutant, indicating that ISS1 can carry out both aromatic amino acid catabolism and biosynthesis in vivo. In agreement with these in vivo findings, ISS1 has in vitro AT activity using Phe, Tyr, or Trp as the amino donor or amino acceptor.

Two distinct families of ATs are known to participate in plant aromatic amino acid metabolism: the TAA family of Trp ATs, and Tyr ATs (TATs). TAA1, and the related TAR proteins are >50% identical, while TAA1 and ISS1 are only 12% identical, consistent with ISS1 not being in the TAA1 family. The seven Arabidopsis TATs share between 38 and 51% sequence identity with each other (Prabhu and Hudson 2010) while TATs and ISS1 share only 23% sequence identity. Two TATs were able to rescue the DL39 E. coli Phe and Tyr auxotrophy (Prabhu and Hudson 2010; Riewe et al. 2012) and were found to have AroAT activity in vitro (Prabhu and Hudson 2010; Riewe et al. 2012). Additionally, the tat5 mutant has an 11-fold increase in Tyr levels and a decrease in Tyr secondary metabolites, while showing no changes in Phe or Trp metabolism (Riewe et al. 2012). Based on the phenotype of the iss1 mutant and limited similarity of ISS1 to TAT proteins, we suggest that ISS1 is a member of a plant AroAT family distinct from those previously described. An open question is whether ISS1 contributes to Phe and/or Tyr biosynthesis. While plants primarily convert prephenate to Phe/Tyr by transamination followed by dehydration/decarboxylation (Tzin and Galili 2010), plants possess enzyme activities capable of carrying out the dehydration/decarboxylation of prephenate first, followed by transamination to Phe or Tyr (Tzin et al. 2009; Yoo et al. 2013). The ability of ISS1 to rescue the yeast aro8 aro9 mutant combined with its in vitro activities is consistent with ISS1 also serving a role in Phe/Tyr metabolism.

iss1 mutants reveal an interaction between Trp and Phe metabolism

Given that several recent reports have demonstrated a high degree of coordination between Trp and Phe metabolism (Yamada et al. 2008; Tzin et al. 2009; Huang et al. 2010), it was not surprising that iss1 plants have altered Phe metabolism as exhibited by reduced coniferin production and altered sensitivity to PFP. In plants, the majority of Phe is used for the production of the two classes of phenylpropanoids, monolignols and flavonoids (Huang et al. 2010). On unsupplemented medium, iss1-1 plants displayed a decrease in coniferin while flavonoid levels were unchanged. This finding suggests that the iss1-1 mutation disrupts regulation of specific steps downstream of the general phenylpropanoid pathway. Environmental stresses such as UV-B regulate the phenylpropanoid pathway as exemplified by increased production of certain flavonoids (Ryan et al. 2001; Tohge et al. 2011). We observed similar fold increases in flavonoid levels between iss1-1 plants and WT plants grown on unsupplemented medium following 24-hr UV-B exposure, suggesting that iss1-1 affects constitutive phenylpropanoid production rather than inducible phenylpropanoid synthesis.

The fact that iss1-1 is resistant to PFP suggests an alteration in Phe metabolism. In plants, resistance to the toxic Phe analog PFP is associated with a loss of feedback inhibition (Palmer and Widholm 1975) or an increased demand for Phe (Berlin and Widholm 1977). We found that Phe levels in the iss1-1 mutant are similar to WT, suggesting that feedback inhibition is not changed. In addition, the decrease in phenylpropanoid content in iss1-1 would suggest decreased Phe turnover. Nonetheless, because the iss1 mutant’s effects on Phe metabolism are modest, these phenotypes may be due to direct disruption of a Phe metabolic pathway that ISS1 has a minor role in or to indirect effects caused by dysregulation of Trp metabolism.

Conclusions

From this work and the recent characterization of vas1, ISS1/VAS1 is an AroAT that has multiple roles in Trp metabolism. ISS1/VAS1 functions in meristematic tissue to modulate IAA levels by conversion of IPA to Trp, while in vegetative tissues it functions in Trp catabolism. iss1 mutant plants have an indole-dependent increase in Trp-I IAA synthesis that results in a significant increase in IAA. Additionally, the fact that high levels of Trp appear to be required for the redirection of indole into Trp-I IAA synthesis points to an important role for Trp catabolism in IAA homeostasis. That iss1 has altered Trp as well as Phe metabolism supports the idea of coordinated flux of metabolites through the Trp and Phe biosynthetic pathways. ISS1/VAS1-related proteins appear across the plant kingdom and form a distinct clade within the α I family of PLP-dependent enzymes. This suggests a conserved function of ISS1/VAS1-related proteins in plant metabolism, and future investigations of ISS1/VAS1 will be useful for revealing interactions between primary and secondary aromatic amino acid metabolism.