Analytical methods for stable isotope labeling to elucidate rapid auxin kinetics in Arabidopsis thaliana

The phytohormone auxin plays a critical role in plant growth and development. Despite significant progress in elucidating metabolic pathways of the primary bioactive auxin, indole-3-acetic acid (IAA), over the past few decades, key components such as intermediates and enzymes have not been fully characterized, and the dynamic regulation of IAA metabolism in response to environmental signals has not been completely revealed. In this study, we established a protocol employing a highly sensitive liquid chromatography-mass spectrometry (LC-MS) instrumentation and a rapid stable isotope labeling approach. We treated Arabidopsis seedlings with two stable isotope labeled precursors ([13C6]anthranilate and [13C8, 15N1]indole) and monitored the label incorporation into proposed indolic compounds involved in IAA biosynthetic pathways. This Stable Isotope Labeled Kinetics (SILK) method allowed us to trace the turnover rates of IAA pathway precursors and product concurrently with a time scale of seconds to minutes. By measuring the entire pathways over time and using different isotopic tracer techniques, we demonstrated that these methods offer more detailed information about this complex interacting network of IAA biosynthesis, and should prove to be useful for studying auxin metabolic network in vivo in a variety of plant tissues and under different environmental conditions.


Introduction
Plant growth control is regulated by changes in the levels of active hormones, their transport to and away from the site of action and the sensitivity of plant tissues to changes in these concentrations [1].In particular, auxin levels in plants appear to be controlled by several different metabolic pathways.This combination of metabolic regulatory mechanisms has been termed "auxin homeostasis" to highlight the expected high degree of control exerted by plants to maintain optimal growth rates.Auxin homeostatic regulation involves an interactive network of redundant pathways, the complexity of which has been shown by the application of a combination of molecular, genetic, and analytical approaches [2][3][4][5].
A major limitation of many studies of phytohormone biosynthesis is the lack of dynamic information to allow interpretation of data in the context of metabolic fluxes.Metabolic analysis may demonstrate an increased abundance of a hormone or its metabolite under specific conditions.Often such studies cannot determine whether this is the result of increased flux from a biosynthetic enzyme, decreased flux by a catabolic enzyme, changes in formation or hydrolysis of conjugating moieties, or alteration in transport of the metabolite [6].Furthermore, phytohormone metabolic pathways are largely interconnected, and a number of interacting metabolic processes maintain metabolite levels.These cannot be adequately resolved by measurement of steady-state metabolite concentrations alone.
These considerations are of concern because the vast bulk of studies investigating phytohormone metabolism have approached physiological questions by measuring hormone levels and/or metabolite levels to infer a metabolic origin, fate, or regulation, with only a few notable exceptions.Although many of these studies have been informative, they employ limited static "windows" that are often extrapolated to predict time-dependent metabolic processes.The full limitations of such approaches are difficult to predict with any certitude.Similarly, approaches that seek to measure reactions where a supplied metabolic precursor is applied and its conversions to a product of interest, typically using an isotopic label, provide different information but also are subject to some inherent limitations.These can be compounded by measurements that are limited in number of time points.Also, often measurements are made after protracted labeling periods.These times can range from somewhat less than an hour to several hours [7][8][9][10][11][12][13]. In other cases, labeling times can continue for a day to even weeks [14][15][16].Clearly, temporal understanding can be a concern given that indole-3-acetic acid (IAA) turnover have been measured or modeled to be in the range of a few minutes to 1-10 hours [17][18][19].
To overcome some of the possible limitations of prior efforts, we developed a method for the targeted metabolomic analysis of the indole auxin biosynthetic network [20].We have now further adapted this method by first analyzing when seedlings of Arabidopsis make the transition from heterotrophic to autotrophic auxin biosynthesis, and then by using seedlings of Arabidopsis to analyze the incorporation at stable isotopic labels into newly synthesized IAA as well as potential biochemical precursors at very short time intervals with time-dependent sampling.By these small molecule implementations of Stable Isotope Labeled Kinetics (SILK) methods, we demonstrate that measuring entire pathways over time and using different isotopic tracer precursor techniques offer improved and more detailed information about this complex interacting network of metabolic reactions.

Harvest, homogenization and extraction
These steps were performed as detailed below in two separate options.In each case, plant samples were thoroughly homogenized (with added internal standard, if desired, for quantitative analysis) and equilibrated for a short time before centrifugation to remove large particles.
Option 1a: Absolute quantification of unlabeled IAA and [ 15 N 1 ]IAA.[ 13 C 6 ]IAA (CAS 100849-36-3; IAA (phenyl-13 C₆), 99atom%, Cambridge Isotope Laboratories, CLM-1896) was used as internal standard to find the change from the heterotrophic IAA utilization from stored reserves and precursors to autotrophic de novo IAA production.5-20 mg of seedlings were collected into a 1.5 mL microcentrifuge tube (Fisherbrand, 05-408-129) every day at the same time from day 1 to day 14 after the transfer to the growth chamber.Microcentrifuge tubes were weighed before and after sample collection using an analytical balance to record sample fresh weight.Immediately after weighing, tubes were submerged in liquid nitrogen to flash freeze and placed on dry ice.Samples were stored at -80˚C until extraction.20 μL of homogenization buffer per 10 mg tissue and 2-3 stainless steel beads (1.6 mm diameter, Next Advance, SSB16) were added to each sample with 10 ng of [ 13 C 6 ]IAA mixed per 1 mL homogenization buffer (65% isopropanol, 35% 0.2 M imidazole (pH 7.0)).Samples were homogenized in Geno/Grinder (SPEX SamplePrep) for 4 minutes at 1750 RPM, followed by incubating on ice for approximately 1 hour.90 μL of water were added to each homogenized sample per 10 μL homogenization buffer.After mixing, samples were centrifuged at 25,000 g for 10 min.at 4˚C.
IAA biosynthesis intermediates analysis.A script was employed to determine peak areas from EICs of multiple compounds.Narrow mass ranges centered around the exact masses of ions generated by the compounds of interest, along with their labeled counterparts synthesized from the supplied labeled precursors, were defined to minimize background noise.Before inputting into R, raw data files were converted to mzXML format using the msConvert tool from the ProteoWizard software [31].Quantitative data for each indolic compound of interest was extracted using the Metabolite-Turnover script developed in the Hegeman lab.(https:// github.com/HegemanLab/Metabolite-Turnover,[32]).Within this script, the ProteinTurnover [33] and the XCMS package [34] are utilized to extract EICs for every isotopomer of IAA and intermediates.This quantification approach employing linear regression [35] is preferred over peak area-based quantification [36] when the MS data exhibits high background noise due to low analyte abundance.Exact masses for isotopomers of proposed IAA biosynthetic intermediates were calculated using the University of Wisconsin-Madison Biological Magnetic Resonance Data Bank exact mass calculator (http://www.bmrb.wisc.edu/metabolomics/mol_mass.php).These isotopomers of interest, derived from various isotopic labeling strategies, are listed in Tillmann et al. [20].In the.CSV data output files, the slope of each linear regression line denotes the ratio between the corresponding isotopic trace and its monoisotopomer.This ratio was used to calculate the relative abundance of labeled compounds, enabling us to monitor the label incorporation from upstream precursors into IAA intermediates through multiple pathways.

Results and discussion
An important advantage of SILK is that it can be used to assess not only the status on on-going metabolic activity but also the effect of various metabolic modulators on a target metabolic network [37].The quantification of compounds as well as the rate of labeling has several advantages over static measurements of concentration [6].For example, both metabolic activity and pathway disruptions are not adequately visualized by simply quantifying the concentrations before and after treatments.SILK, however, quantifies the metabolism of target compounds and can, for example, reflect which side of the biosynthesis/catabolism process has been affected and may reveal metabolism changes even if overall metabolite concentrations are essentially unchanged.In addition, SILK can be adapted, as employed in this method, to measure labeling patterns very rapidly in complex networks and determine changes in labeling as altered by metabolic disruptions over short time intervals [38].
For the analysis of IAA biosynthesis in a seedling it is better defined after the plant has become fully autotrophic for its auxin needs; this limits inputs from large stores of storage forms [39].There are only a few studies in Arabidopsis that have accessed this critical transition period, although a few reports exist employing deuterium oxide with plants with larger seeds such a maize [40][41][42][43] and bean [44].These prior studies utilized deuterium oxide primarily because of its rapid uptake into cellular compartments and because the low-resolution mass spectrometers available at the time were better suited to analyze multiple labels to avoid interference with naturally occurring heavy atoms, primarily endogenous 13 C, that gave IAA approximately a 11.1% m+1/z increase in the isotopic envelope.The growth of plants on deuterium oxide however was shown to be inhibited [44,45] thus making evaluation of temporal changes difficult to reconcile.In addition, deuterium exchange processes [27,40,44] complicates isotopic analysis.Modern high resolution mass spectrometers, however, can easily distinguish between a mass increase of 1.00335 for 13 C, 1.00628 for 2 H and a mass increase of 0.99703 for 15 N, allowing the structural nitrogen atom to serve as a monitor of new synthesis.Methods for growth with 15 N have been developed [20].Structural atoms such as 15 N, 18 O and 13 C avoid many of the problems of exchangeable deuterium, as high concentrations are not toxic, nor do they alter growth rates, and they are not generally subject to rearrangements [46] during mass spectrometry processes.One disadvantage to using inorganic 15 N to measure de novo synthesis of a nitrogen-containing molecule is the recycling of nitrogen that occurs [47], however that change in the baseline of enrichment becomes important only where absolute rates of synthesis are being measured.As shown in Fig 2, low levels of 14 N become consistent at day 8, remains low, and "new IAA" with the [ 15 N]-label becomes the predominate form.Our labeling experiments were thus done with 12-day-old seedlings so they are well beyond the "auxin autotropic" stage based on [ 15 N]-incorporation.
Previous studies using isotopic labeling measured label incorporation after hours (3 or 6 hrs, [10]; 6, 10, 21 hrs, [48]) or days (3 or 6 days, [44]; 19 days, [14]) which has the potential to complicate data interpretation for a complex pathway where rates of reactions are potentially much faster than the analysis times [18].The goals of this current research were to develop a procedure where labeling times could be reduced, and the kinetics of labeling could be measured over relatively short time intervals.
By labeling plants with [ 13 C 6 ]anthranilate, the appearance of [ 13 C 6 ]IAA could be detected after 1 hour of incubation (Fig 3), with little or no change in the [ 15 N 1 ]IAA pool except for a decline immediately upon transfer to the label.After 2 hours of incubation, labeling from [ 13 C 6 ]anthranilate into IAA was reduced by treatment with the monooxygenase inhibitor YDF (Fig 3).Labeling with [ 13 C 8 , 15 N 1 ]indole was more rapid than labeling with [ 13 C 6 ]anthranilate, suggesting perhaps more rapid uptake as might be expected for a lipid soluble compound.Interestingly, YDF had little effect on the incorporation of label into [ 13 C 8 , 15 N 1 ]IAA (Fig 4).This is similar to what we previously reported after a longer-term incubation [49] but without time resolved data.The effect of IPyA pathway inhibition by both the tryptophan amino transferase inhibitor PVM2153 and the monooxygenase inhibitor YDF was more pronounced than YDF alone, with little incorporation of 13 C into IAA from [ 13 C 6 ]anthranilate labeling even after 4 hours with both inhibitors (Fig 5).With [ 13 C 8 , 15 N 1 ]indole and both inhibitors, labeling of IAA was reduced more than with YDF alone, but still not as pronounced as with [ 13 C 6 ] anthranilate (Fig 6).
Inhibition of the biosynthesis of tryptophan as well as reactions related to tryptophan is complex in plants in part because indolic compounds and derivatives are often accumulated as part of the biotic resistance system [50].In addition to two gene copies encoding most of the enzymes involved in the conversion of anthranilate to tryptophan, including tryptophan synthase (TS) α and β subunits, Arabidopsis has the TSα-like indole synthase (INS), a type2 AtTSβ protein, and an undefined AtTSβtype1 [16,[51][52][53][54].A number of different inhibitors are known for the canonical TSαββα heterotetramer [20], however, such metabolic modulators have not been explored as to how they alter the activities of INS, nor standalone type1 and type2 AtTSβ proteins.Nevertheless, compounds that inhibit the activity of these early steps related to IAA metabolism have promise for further analysis of the precursors involved in the pathways.In a proof-of-concept study we used a well-established inhibitor of TS, an arylsulfide phosphonate (I26; [4-[(2-aminophenyl) sulfanyl]butyl] phosphonic acid) that was originally designed as a transition state analog targeting TSα.TS inhibitors are, however, well known to exert pleiotropic effects across the TSαββα heterotetramer due to complex allosteric interactions and little is known about their effects on the TSα-like INS enzyme, AtTSβtype2 or ).This result is not expected based on the first reports on I26 and its predicted mechanism.Nevertheless because, as discussed in Michalska et al. [55], TSαβ is allosterically regulated by the physical switching of the α-and β-subunits between an open low activity confirmation and a closed high activity conformation, interaction between TSαβ can be complex.In their open conformations, active sites are freely accessible to supplied substrates while in their closed states, sites are inaccessible to external substrates.While this switching normally prevents the escape of the intermediate, indole, produced by the α subunit, "locking" the TSα in the closed confirmation could block supplied labeled indole from entry.Such an inhibitory mechanism was first postulated based on the protein structures in the presence of aryl sulfonamides, also proposed as TSα inhibitors [56,57].These structures showed the α and β subunits in closed conformations with blocked access into the α and β sites in the presence of inhibitors.Interestingly, the inhibition pattern  15 N 1 ]indole labeling of IAA when the predominate Trp pathway to IAA [49], the IPyA pathway, is blocked.However, there are several possible explanations for the data.First, TS could be involved in both Trp-dependent and Trp-independent IAA biosynthesis, although data from Trp auxotrophic mutants would suggest otherwise (reviewed in [58]).Indole-3-glycerol-phosphate (IGP) could be the precursor to IAA.Its formation by a reverse activity of any TSα-like enzyme would be expected to be inhibited by I26 as described for the forward reaction.Finally, an enzyme activity that forms a product important for IAA biosynthesis from indole could be a secondary target for inhibition by this allosteric mimic.
Stable isotope labeling followed by high resolution LC-MS/MS analysis is a powerful technique that can be used effectively to investigate labeling kinetics between multiple experimental groups and in tissues exposed to effectors of metabolic pathway activities.Determining relative changes to specific steps over short time intervals in a less well characterized metabolic pathway can aid in the understanding of how specific steps respond to disruptions, especially in steps catalyzed by multiple enzymes with similar and overlapping functions.These technologies can be easily adapted to measure processes in particular cellular fractions to determine in real-time cellular compartment dynamics.Similarly, stable isotopes allow the methods to be extended to be used in dynamic pulse-chase experimental designs followed by LC-MS/MS in order to investigate individual compound turnover.