Protocol: analytical methods for visualizing the indolic precursor network leading to auxin biosynthesis

Background The plant hormone auxin plays a central role in regulation of plant growth and response to environmental stimuli. Multiple pathways have been proposed for biosynthesis of indole-3-acetic acid (IAA), the primary auxin in a number of plant species. However, utilization of these different pathways under various environmental conditions and developmental time points remains largely unknown. Results Monitoring incorporation of stable isotopes from labeled precursors into proposed intermediates provides a method to trace pathway utilization and characterize new biosynthetic routes to auxin. These techniques can be aided by addition of chemical inhibitors to target specific steps or entire pathways of auxin synthesis. Conclusions Here we describe techniques for pathway analysis in Arabidopsis thaliana seedlings using multiple stable isotope-labeled precursors and chemical inhibitors coupled with highly sensitive liquid chromatography-mass spectrometry (LC–MS) methods. These methods should prove to be useful to researchers studying routes of IAA biosynthesis in vivo in a variety of plant tissues. Supplementary Information The online version contains supplementary material available at 10.1186/s13007-021-00763-0.


Background
Plant life is characterized by strictly regulated developmental events that achieve optimum growth and reproduction. This is accomplished through an extremely complex hormonal signaling network in which the plant growth hormone auxin plays a central and defining role. To this end, auxin helps regulate almost all aspects of plant growth and development including embryogenesis, tissue architecture and tropic responses [1]. Maintenance of auxin homeostasis involves multiple pathways for the biosynthesis of indole-3-acetic acid (IAA), the principal auxin in plants, and several regulatory pathways as well as subsequent catabolic events. These additional input/ output processes include conjugation and hydrolysis of sugar and cyclitol conjugates, amino acid, peptide and protein conjugates, formation and β-oxidation of indole-3-butyric acid as well as deactivation by ring oxidation of IAA and its amino acid conjugates [2,3]. Nevertheless, how much IAA is made and accumulates remains the critical regulatory event in many aspects of plant development [4].
Although several biosynthetic pathways for the bioactive auxin IAA have been proposed, many of them have not been well defined and flux information is largely lacking (Fig. 1). The predominant biosynthetic route to IAA in Arabidopsis thaliana is widely believed to be through the YUCCA pathway, in which the amino acid tryptophan (Trp) is converted to indole-3-pyruvic acid (IPyA), which is then converted to IAA by YUCCA flavin monooxygenase enzymes [5]. Species-specific evidence for the synthesis of IAA from Trp through indole-3-acetaldoxime (IAOx), which is converted to indole-3-acetamide (IAM) and sometimes an indole-3-acetonitrile (IAN) intermediate has been shown in Arabidopsis [6,7]. Other potential intermediates of IAA synthesis downstream of Trp have been proposed, such as indole-3-acetaldehyde (IAAld) [8][9][10] and tryptamine (TAM) [11], though their places within the web of auxin biosynthesis have not been well detailed. A Trp-independent route has also been proposed based on tryptophan synthase mutants, metabolic flux analysis and in vitro analyses, in which indole or another upstream compound serves as the IAA precursor [1,[12][13][14]; however, unbound chemical intermediates, if they are involved in this pathway, have as yet not been identified [15]. The purpose of this protocol is to describe improved techniques for characterization of the auxin metabolic network utilizing recently discovered chemical inhibitors and technical advances in mass spectrometry (Fig. 2). These tools will allow researchers to characterize auxin biosynthesis during specific developmental events or environmental responses.
Metabolic inhibitor approaches are complementary to genetic and biochemical studies and are particularly useful in studying IAA biosynthesis. While auxin biosynthesis mutants may have severe developmental defects that alter growth and confound comparisons to wild type plants [16], biosynthetic reactions can be turned off at specific developmental time points with chemical inhibitors. Additionally, genetic redundancy can be overcome by inhibiting an entire enzyme family with a single chemical treatment [17]. Such is the case with inhibitors targeting both steps in the YUCCA pathway. The YUCCA enzymes are encoded by multiple genes in Arabidopsis thaliana and mutations in small sets of these genes encoding the flavin monooxygenase proteins results in significant morphological defects [18]. A number of chemical inhibitors have been developed to inhibit the YUCCA pathway of auxin biosynthesis (Table 1), providing valuable tools to study the function of this pathway in different plant tissues and environmental conditions. Similarly, TAA1/TAR/ISS1/VAS1 (Tryptophan Aminotransferase of Arabidopsis 1/ Tryptophan Aminotransferase Related/ Indole Severe Sensitive 1 and reversal of sav3 phenotype 1) form a set of enzymes with overlapping biochemical functions that catalyze the penultimate step in the IPyA pathway [19]. Alternative aromatic amino acid substrates, such as L-kynurenine, can act as competitive inhibitors of tryptophan aminotransferase and a series of potent inhibitors have been developed to pyridoxal phosphate-dependent enzymes with enhanced specificity to TAA1 and related enzymes ("pyruvamines"; see Table 1) [20]. Major pathways for IAA biosynthesis. Solid arrows refer to pathways with enzymes identified in at least one species, and dashed arrows to undefined ones. AMI1, indole-3-acetamide hydrolase-1; ANT, anthranilate; CHA, chorismic acid; IAAld, indole-3-acetaldehyde; CYP79B2/3, cytochrome P450 (79B2/3); IAM, indole-3-acetamide; IAN, indole-3-acetonitrile; IAOx, indole-3-acetaldoxime; IGP, indole-3-glycerol phosphate; INS, indole synthase; IPyA, indole-3-pyruvic acid; ISS1, Indole Severe Sensitive 1; NIT, nitrilase; Ser, serine; TAA1, tryptophan aminotransferase of Arabidopsis 1; TAR, tryptophan aminotransferase-related; TAM, tryptamine; Trp, tryptophan; TSA, tryptophan synthase α; TSB, tryptophan synthase β; YUCCA, Arabidopsis flavin monooxygenase The issues of redundancy with tryptophan synthase (TS) are a bit different. Arabidopsis and maize have two copies of the genes that encode each of the two proteins that form the αββα heterodimeric complex that catalyzes the formation of tryptophan from indole-glycerolphosphate and serine in the plastids. In addition, maize has genes BX1 and IGL for TSα-like cytosolic enzymes that serve as sources of free indole [21]. Arabidopsis also has a cytosolic TSα-like enzyme encoded by the indole synthase (INS) gene [22]. TS is, however, a well-researched and highly conserved bi-enzyme complex [23] such that inhibitors are available ( Table 1) that target specifically TSα, TSβ as well as the 25-Å long tunnel to the β-subunit where indole diffuses in order to participate in the TSβ pyridoxal 5'-phosphatemediated β-addition reaction with serine. Determining the possibility of a tryptophan-independent pathway is largely dependent on having Trp auxotroph mutants, which are difficult to obtain due to redundancy of Trp synthase genes and the fact that mutations in both copies of TSβ are seedling lethal [12,13,24]. The protocols described here largely overcome these issues by Indoleacrylic acid trans-indole-3-acrylic acid Trp synthase β and α Allosteric inhibitor [56,57] (1-Fluorovinyl)glycine α-(1′-fluoro)vinyl glycine Trp synthase β PLP-enzyme mechanism-based inhibitor [58] Arylsulfide phosphonates Trp synthase inter-subunit interface Allosteric inhibitor [62] Aryl sulfonamides [F9]; N-(4'-Trifluoromethoxy benzenesulfonyl)-2-aminoethyl Phosphate Trp synthase β α-Site allosteric ligand [63] Benzamide N-(4-Carbamoyl benzyl)-5-(3chloro phenyl)-1,2-oxazole-3-carboxamide Trp synthase α α-Site ligand [64] employing chemical inhibitors, and can complement genetic studies. Mass spectrometry (MS) has historically been and continues to be an important technique in deciphering routes of auxin biosynthesis, enabling accurate quantitation of IAA and its precursors, identification of intermediates, and tracking of isotopic labels through distinct pathways. Quantitative methods for IAA and precursor analysis by MS have been invaluable tools in elucidating auxin biosynthesis pathways and have continuously evolved over time with advances in analytical sensitivity and resolution [4,[25][26][27][28][29]. Stable isotope tracing experiments also lend insight into auxin biosynthesis when plant tissue is supplied with one or more labeled precursors, such as indole and/or anthranilate [30][31][32], and label incorporation into suspected downstream intermediates is monitored to determine whether synthesis from the labeled precursor has occurred. This approach can also provide information regarding direction of flow and flux through different steps [6]. Additionally, labeled precursors that are unique to one pathway in particular can be applied to measure contributions of a specific pathway to the IAA pool [5,33].

Results and discussion
In this paper, we describe methods utilizing metabolic inhibitors coupled with a modified approach of isotope dilution/tracing and using liquid chromatography-high resolution-mass spectrometry (LC-HR-MS) for qualitative and quantitative analysis of a comprehensive set of IAA precursors and IAA itself to characterize auxin biosynthesis in Arabidopsis (see Additional file 1). A distinct advantage of this method is its ability to resolve potential precursor compounds by chromatographic retention, absolute mass and by elemental composition, enabling complex mixtures of different stable isotopes (for example, multiple labeled compounds with 13 C and 15 N can be resolved) to be used in the experimental procedures (see Additional file 2). Readers may also consult a complementary paper that was published while this manuscript was in preparation [34]. Growing seedlings on fully 15 N-labeled media as described here enables accurate quantitation of biosynthetic intermediates by reverse isotope dilution, using unlabeled internal standards which are typically more readily available than isotopically labeled standards [35]. The addition of one or more 15 N atoms at a mass addition of 0.9970 can be resolved from the more abundant natural occurrence of 13 C, which is 0.0034 heavier than 12 C, which improves the utility of this approach when using high resolution analysis. Seedlings are first germinated on nylon mesh and are easily transferred onto media containing chemical treatments at the desired developmental time point.
Next, stable isotope-labeled precursor compounds are fed to the plant. Labeled serine is used as a tracer for Trpdependent biosynthesis specifically [33], while labeled indole and anthranilate can feed into both Trp-dependent and Trp-independent pathways [19,31,32] (Fig. 1). The techniques described here offer several advantages over previously described methods in their ease of preparation, high level of sensitivity, capacity for monitoring many compounds at once (see Additional file 2), and the ability of high resolution analysis to distinguish between different 'heavy' atoms, as might be required with [ 13 C 1 ] IAA and [ 15 N 1 ]IAA labeling products. As shown in Additional file 1, the use of multiple labels makes it easy to see that the addition of the tryptophan monooxygenase inhibitor YDF increases the incorporation of labeled indole into IAA but decreases labeling from labeled anthranilate and to a lesser degree from labeled tryptophan. Furthermore, this IAA labeling pattern for labeled indole and anthranilate is not reflected in any of the proposed intermediates following YDF treatment.
We also describe a technique for identifying novel intermediates based on the characteristic quinolinium ion produced from MS fragmentation of 3-substituted indolic compounds. This method involves using a series of injections of the same sample with increasingly narrow mass ranges, similar to the methods utilized by Yu et al. [36] and Tang et al. [37] where they targeted and identified novel indolic compounds. By monitoring exact masses of [ 13 C 8 , 15 N 1 ]-and [ 15 N 1 ]quinolinium ions after treatment with [ 13 C 8 , 15 N 1 ]-and [ 15 N 1 ]indole, this method can identify unknown compounds synthesized downstream from indole. A similar approach would likely be applicable in investigations of other classes of compounds that form characteristic signature ions. High resolution accurate mass analysis significantly reduces factors such as false negative molecular ions, low abundance ions, multiple isomers, and matrix effects, which otherwise would make it difficult to confirm possible compound identities.

Growing, labeling, and collecting plant material
Wild-type Columbia-0 ecotype Arabidopsis thaliana seeds or specific metabolic mutant lines need to be surface sterilized sodium hypochlorite then imbibed for 5-10 days at 4 °C to promote uniform germination. Typically seeds would be sown in a single row onto 20 μm nylon mesh covering the agar growth medium.

IAA extraction
Homogenized samples are incubated on ice for 50 min to allow isotopic standard equilibration with the endogenous IAA. They are then diluted tenfold with water such that ion exchange will be effective, centrifuged to remove solid materials, and loaded onto two consecutive SPE micro spin column (TopTips) steps, first ion exchange on an amino phase and then on an epoxide support.
• Bondesil-NH 2 resin (Agilent, 12,213,020) suspended in water, 1:4 w:v Homogenized samples are incubated on ice for 50 minutes to allow isotopic standard equilibration with the endogenous compounds, diluted 10-fold with water to allow proper interaction with the solid phase, centrifuged to remove solids, and loaded onto a SPE micro spin column (TopTips) containing hydrophilic-lipophilic balanced (HLB) resin conditioned with acetonitrile followed by 20% acetonitrile in water. After loading, the spin columns are washed with 5% acetonitrile and compounds are eluted with 80% acetonitrile.

Indole extraction
Indole is a very lipophilic and somewhat volatile compound that cannot be purified using the techniques used for the other compounds. Thus, its purification involves a simple solvent partitioning. It was important to select an apolar solvent with a boiling point below the melting point of indole. We found pentane to be well-suited as its boiling point is 36.0 °C, well below the indole melting point of 52.5 °C.

LC-MS analysis
UPLC utilizes a column with an end-capped octadecylsilane fully porous 1.8 µm silica resin with high carbon loading (20%) in order to obtain highest sensitivity for indolic compounds (see Additional file 2).

Growing seedlings with inhibitor and stable isotope precursor treatments
Seedlings are grown in vitro on mesh squares, allowing them to be easily transferred to chemical inhibitor treatments at the desired timepoints. A liquid solution containing stable isotope-labeled precursors is then supplied to seedlings, and synthesis of isotopically labeled IAA and intermediates can be identified and distinguished by LC-HR-MS.
1. In a laminar flow hood, moisten sterile nylon mesh squares with sterile water and use forceps to place squares flat on germination media (see Note 7 and Table 2) in square Petri dishes. 2. Clean Arabidopsis seeds by shaking in 20% bleach solution for 5 min and rinsing 4 times with sterile water. 3. Sow seeds approximately 0.5 cm apart in single row on mesh. 4. Store plates at 4̊ C in the dark for 3-7 days to stratify seeds. Remove plates from cold and place vertically in growth conditions. 5. Transfer seedlings onto inhibitor media (Table 1) to begin auxin biosynthesis inhibition treatment (see Note 8). In a laminar flow hood, use forceps to gently lift mesh with seedlings from germination plates and lay flat onto plates containing inhibitor media. Cover plates and place vertically under growth conditions. 6. Begin isotopic labeling treatments by flooding plates with 3 mL of labeling solution (

Proposed IAA biosynthesis pathway intermediates: Anthranilate, Ser, IPyA, IAAld, IAOx, IAN, IAM
Samples are prepared for analysis of biosynthesis intermediates by SPE using an HLB resin. SPE is an effective sample preparation technique for these compounds because it provides a high level of recovery and is relatively easy to use with large sample sets. IAA can also be extracted using the following method, but with some loss of sensitivity compared to methods described in the previous section.

Unknown indolic compounds (double indole labeling samples)
An unbiased extraction method is used for discovery of unknown compounds synthesized from indole.
27. Transfer supernatant to a clean tube and centrifuge again at 25,000 g for 10 min. at 4 °C to remove all debris.

LC-MS analysis
Samples are analyzed using LC-HRAM-MS to chromatographically separate components of chemical matrix and obtain high resolution m/z data. Specific LC-MS methods are tailored for different sample types and analysis objectives (see method details in "Materials" section).
28. Carefully transfer each sample to a 50 μL glass insert so that no air pockets remain at the bottom of the insert. Assemble insert into autosampler vial with cap. 29. Inject 5-10 μL of sample for LC-MS analysis using methods described in the LC-MS analysis subsection of the Materials section.

Data analysis
IAA analysis Extracted ion chromatograms (EICs) of labeled and unlabeled quinolinium ions generated by fragmentation of labeled internal standard and unlabeled endogenous IAA are viewed (see Additional file 2). Narrow mass ranges are used to filter out background noise.
30. Under the "Ranges" tab in "Chromatogram Ranges" in Xcalibur, set the chromatogram viewing options to display two mass ranges: 130.0641-130.0661 (corresponding to unlabeled quinolinium ion), and 136.0843-136.0863 ([ 13 C 6 ] quinolinium produced from [ 13 C 6 ]IAA internal standard). Under "Display" tab, check "Peak Area. " Use "peak selection" tool to select and calculate area of peaks corresponding to unlabeled IAA and the internal standard. Endogenous IAA levels can be calculated using isotope dilution [25,39].
Targeted IAA precursor analysis Peak areas from EICs of multiple compounds are determined using a script. Mass ranges surrounding the exact masses of ions produced from the compounds of interest, as well as their labeled forms synthesized from the supplied labeled precursors, are kept within a narrow window to exclude background noise.

31.
Raw data files are converted to mzXML format using the msconvert tool from the ProteoWizzard software [40] prior to input into R. Quantitative data for each indolic compound is extracted using the Metabolite-Turnover script developed in the Hegeman lab (https:// github. com/ Hegem anLab/ Metab olite-Turno ver, [41]). In this script, the ProteinTurnover [42] and the XCMS package [43] are employed to extract EICs for each isotopmer of IAA and intermediates. This quantification approach using linear regression [44] is preferred over that using peak area [39] when the MS data has high background noise due to low analyte abundance. 32. Exact masses for isotopomers of interest are calculated using the University of Wisconsin-Madison Biological Magnetic Resonance Data Bank exact mass calculator (http:// www. bmrb. wisc. edu/ metab olomi cs/ mol_ mass. php). Isotopomers of proposed IAA biosynthetic intermediates derived from several isotopic labeling strategies are listed in Table 3.
(See Note 12) 33. In the data output csv files, the slope of each linear regression line represents the ratio of the respective isotopic trace to its monoisotopmer. This ratio is used to calculate the relative abundance of labeled compounds, allowing us to track label incorporation from upstream precursors into IAA intermediates through multiple pathways.
Double indole labeling data analysis Supplying plants two differentially labeled form of indole provides a way to identify indole-derived compounds, as downstream intermediates will incorporate both labels. These samples are analyzed in a series of LC-MS/MS injections, initially scanning broadly for formation of labeled quinolinium ions, and then narrowing in on precise ions in subsequent injections until a molecular ion can be identified and fragmented to provide further structural information.    12 We recommend using a mass range window of the calculated m/z value ± 0.003.