Conversion of Endogenous Indole-3-Butyric Acid to Indole-3-Acetic Acid Drives Cell Expansion in Arabidopsis Seedlings

Genetic evidence in Arabidopsis ( Arabidopsis thaliana ) suggests that the auxin precursor indole-3-butyric acid (IBA) is converted into active indole-3-acetic acid (IAA) by peroxisomal b -oxidation; however, direct evidence that Arabidopsis converts IBA to IAA is lacking, and the role of IBA-derived IAA is not well understood. In this work, we directly demonstrated that Arabidopsis seedlings convert IBA to IAA. Moreover, we found that several IBA-resistant, IAA-sensitive mutants were deﬁcient in IBA-to-IAA conversion, including the indole - 3 - butyric acid response1 ( ibr1 ) ibr3 ibr10 triple mutant, which is defective in three enzymes likely to be directly involved in peroxisomal IBA b -oxidation. In addition to IBA-to-IAA conversion defects, the ibr1 ibr3 ibr10 triple mutant displayed shorter root hairs and smaller cotyledons than wild type; these cell expansion defects are suggestive of low IAA levels in certain tissues. Consistent with this possibility, we could rescue the ibr1 ibr3 ibr10 short-root-hair phenotype with exogenous auxin. A triple mutant defective in hydrolysis of IAA-amino acid conjugates, a second class of IAA precursor, displayed reduced hypocotyl elongation but normal cotyledon size and only slightly reduced root hair lengths. Our data suggest that IBA b -oxidation and IAA-amino acid conjugate hydrolysis provide auxin for partially distinct developmental processes and that IBA-derived IAA plays a major role in driving root hair and cotyledon cell expansion during seedling development.

The auxin indole-3-acetic acid (IAA) controls both cell division and cell expansion and thereby orchestrates many developmental events and environmental responses. For example, auxin regulates lateral root initiation, root and stem elongation, and leaf expansion (for review, see Davies, 2004). Normal plant morphogenesis and environmental responses require modulation of auxin levels by controlling biosynthesis, regulating transport, and managing storage forms (for review, see Woodward and Bartel, 2005a). In some storage forms, the carboxyl group of IAA is conjugated to amino acids or peptides or to sugars, and free IAA can be released by hydrolases when needed (Bartel et al., 2001;Woodward and Bartel, 2005a). A second potential auxin storage form is the side chain-lengthened compound indole-3-butyric acid (IBA), which can be synthesized from IAA (Epstein and Ludwig-Mü ller, 1993) and is suggested to be shortened into IAA by peroxisomal b-oxidation (Bartel et al., 2001;Woodward and Bartel, 2005a).
Unlike the simple one-step release of free IAA from amino acid conjugates, release of IAA from IBA is suggested to require a multistep process (Zolman et al., 2007(Zolman et al., , 2008. Conversion of IBA to IAA has been demonstrated in a variety of plants (Fawcett et al., 1960; for review, see Epstein and Ludwig-Mü ller, 1993) and may involve b-oxidation of the four-carbon carboxyl side chain of IBA to the two-carbon side chain of IAA (Fawcett et al., 1960;Zolman et al., 2000Zolman et al., , 2007.
Mutation of genes encoding the apparent b-oxidation enzymes INDOLE-3-BUTYRIC ACID RESPONSE1 (IBR1), IBR3, or IBR10 results in IBA resistance, but does not alter IAA response or confer a dependence on exogenous carbon sources for growth following germination (Zolman et al., 2000(Zolman et al., , 2007(Zolman et al., , 2008, consistent with the possibility that these enzymes function in IBA b-oxidation but not fatty acid b-oxidation. Both conjugate hydrolysis and IBA b-oxidation appear to be compartmentalized. The IAA-amino acid hydrolases are predicted to be endoplasmic reticulum localized (Bartel and Fink, 1995;Davies et al., 1999) and enzymes required for IBA responses, including IBR1, IBR3, and IBR10, are peroxisomal (Zolman et al., 2007(Zolman et al., , 2008. Moreover, many peroxisome biogenesis mutants, such as peroxin5 (pex5) and pex7, are resistant to exogenous IBA, but remain IAA sensitive (Zolman et al., 2000;Woodward and Bartel, 2005b).
Although the contributions of auxin transport to environmental and developmental auxin responses are well documented (for review, see Petrášek and Friml, 2009), the roles of various IAA precursors in these processes are less well understood. Expansion of root epidermal cells to control root architecture is an auxin-regulated process in which these roles can be dissected. Root epidermal cells provide soil contact and differentiate into files of either nonhair cells (atrichoblasts) or hair cells (trichoblasts). Root hairs emerge from trichoblasts as tube-shaped outgrowths that increase the root surface area, thus aiding in water and nutrient uptake (for review, see Grierson and Schiefelbein, 2002). Root hair length is determined by the duration of root hair tip growth, which is highly sensitive to auxin levels (for review, see Grierson and Schiefelbein, 2002). Mutants defective in the ABCG36/ PDR8/PEN3 ABC transporter display lengthened root hairs and hyperaccumulate [ 3 H]IBA, but not [ 3 H]IAA, in root tip auxin transport assays (Strader and Bartel, 2009), suggesting that ABCG36 functions as an IBA effluxer and that IBA promotes root hair elongation. The related ABCG37/PDR9 transporter also can efflux IBA (Strader et al., 2008b;Rů žička et al., 2010) and may have some functional overlap with ABCG36 (Rů žička et al., 2010). In addition to lengthened root hairs, abcg36/pdr8/pen3 mutants display enlarged cotyledons, a second high-auxin phenotype. Both of these developmental phenotypes are suppressed by the mildly peroxisome-defective mutant pex5-1 (Strader and Bartel, 2009), suggesting that IBA contributes to cell expansion by serving as a precursor to IAA, which directly drives the increased cell expansion that underlies these phenotypes. However, whether IBAderived IAA contributes to cell expansion events during development of wild-type plants is not known.
Here, we directly demonstrate that peroxisomedefective mutants are defective in the conversion of IBA to IAA, consistent with previous reports that these genes are necessary for full response to applied IBA. We found that a mutant defective in three suggested IBAto-IAA conversion enzymes displays low-auxin pheno-types, including decreased root hair expansion and decreased cotyledon size. We further found that these mutants suppress the long-root-hair and enlarged cotyledon phenotypes of an abcg36/pdr8 mutant, suggesting that endogenous IBA-derived IAA drives root hair and cotyledon expansion in wild-type seedlings.
Because the tested IBA-resistant, IAA-sensitive mutants produced low levels of [ 13 C 8 -15 N 1 ]IAA when supplied with [ 13 C 8 -15 N 1 ]IBA (Fig. 1C), we examined endogenous IAA and IBA levels in wild-type and ibr1 ibr3 ibr10 seedlings. We found that free IAA and IBA levels in ibr1 ibr3 ibr10 seedlings were not notably different from levels found in wild type (Supplemental Fig. S1). Because measuring auxin levels in whole seedlings may mask cell-or tissue-specific variations in auxin levels, we examined ibr1 ibr3 ibr10 for auxinrelated phenotypes.

Mutants Deficient in IBA-to-IAA Conversion Fail to Expand Root Hairs
We previously found that abcg36/pdr8/pen3 mutants display defects in IBA efflux and mild developmental phenotypes, including root hairs that were longer than those of wild type ( Fig. 2B; Strader and Bartel, 2009), which suggest increased auxin levels in these cells. We therefore examined mutants with decreased IBA-to-IAA conversion for the opposite phenotype. Many of these mutants are defective not only in IBA-to-IAA conversion ( Fig. 1C) and IBA responsiveness ( Fig. 1A), but also fatty acid b-oxidation, which results in dependence on an external fixed carbon supply such as Suc for growth following germination (Fig. 1B). To avoid complications that might result from the limited fixed carbon available to mutants with defects in fatty acid utilization, we examined root hair lengths of IBA response mutants defective in enzymes suggested to be dedicated to IBA-to-IAA conversion: IBR1, IBR3, and IBR10 (Zolman et al., 2007(Zolman et al., , 2008. These mutants, unlike many other peroxisome-defective mutants, do not require exogenous carbon sources to fuel early seedling growth ( Fig. 1B; Zolman et al., 2000Zolman et al., , 2007Zolman et al., , 2008, suggesting that fatty acid b-oxidation is functioning normally. Unlike the longer root hairs of pen3 mutants ( ) single mutants each displayed shorter root hairs than wild type ( Fig. 2A). Similar to findings in IBA-responsive root elongation assays (Zolman et al., 2008), the ibr1 ibr3 ibr10 triple mutant had shorter root hairs than any of the ibr single mutants (Fig. 2F).
Growing seedlings on the auxin efflux inhibitor 1-naphthylphthalamic acid (NPA) resulted in longer root hairs in wild type (Fig. 3, B and E), ibr1 ibr3 ibr10 (Fig. 3, B and H), and pen3 (Fig. 3, B and K). However, unlike IAA and NAA, NPA-treated ibr1 ibr3 ibr10 root hairs remained shorter than NPA-treated wild-type root hairs (Fig. 3, B, E, and H; Supplemental Fig. S2). Similarly, ibr1 ibr3 ibr10 root hairs treated with the auxin efflux inhibitor 2,3,5-triiodobenzoic acid (TIBA) were longer than untreated root hairs but remained shorter than TIBA-treated wild-type root hairs (Supplemental Fig. S2). The ability of natural and synthetic auxins to fully rescue ibr1 ibr3 ibr10 root hair elongation defects and the ability of auxin transport inhibitors to partially rescue ibr1 ibr3 ibr10 root hair elongation defects are consistent with the possibility that endogenous IBA is an important IAA source during normal root hair elongation.

Mutants Deficient in IBA-to-IAA Conversion Have Smaller Cotyledons
Hypocotyl elongation and cotyledon cell expansion also depend on auxin. The triple hydrolase mutant, which has decreased free IAA levels in whole seedlings (Rampey et al., 2004), displays reduced hypocotyl lengths when grown in the light at 22°C but not at 28°C (Fig. 6A; Rampey et al., 2004), suggesting that conjugates provide some of auxin that drives hypo-cotyl elongation at normal growth temperatures. When we examined ibr1 ibr3 ibr10 for a similar phenotype, we found that 8-d-old ibr1 ibr3 ibr10 seedlings displayed wild-type hypocotyl lengths at both temperatures (Fig. 6A), suggesting that conjugate-derived IAA has a larger role than IBA-derived IAA in driving hypocotyl elongation in the light and that neither precursor contributes to the increased IAA that drives hypocotyl elongation at high temperature Zhao et al., 2002).
We previously found that abcg36/pdr8/pen3 mutants, in addition to displaying elongated root hairs, display enlarged cotyledons ( Fig. 6D; Strader and Bartel, 2009), suggesting that auxin levels are increased not only in root epidermal cells, but also in cotyledon cells. This increase in cotyledon size probably results from increased cotyledon cell expansion, Figure 3. Auxin application restores root hair elongation to ibr mutants. A, Photographs of 5-d-old Col-0 and ibr1-2 ibr3-1 ibr10-1 seedling roots vertically grown under yellow-filtered light at 22°C on medium supplemented with ethanol (mock) or 100 nM IAA. Scale bar = 1 mm. B, Mean root hair lengths (+SE) of 5-d-old Col-0, ibr1-2 ibr3-1 ibr10-1, and pen3-4 seedlings vertically grown under yellow-filtered light at 22°C on medium supplemented with ethanol (mock), 100 nM IAA, 100 nM NAA, or 10 mM NPA. Root hair lengths were measured using NIH Image software (n = 500 total root hairs from at least 12 seedlings). C to K, Histograms of root hair lengths from section B of Col-0 (C-E), ibr1-2 ibr3-1 ibr10-1 (F-H), and pen3-4 (I-K) grown on IAA (C, F, and I), NAA (D, G, and J), or NPA (E, H, and K). Histograms for mock-treated seedlings are shown in every section for comparison. Wt because after germination, Arabidopsis cotyledons grow by cell expansion without cell division (Mansfield and Briarty, 1996). We examined cotyledons of 7-d-old ibr1 ibr3 ibr10 mutants and ilr1 iar3 ill2 mutants and found that ibr1 ibr3 ibr10 displayed smaller cotyledons than wild type, whereas ilr1 iar3 ill2 cotyledons resembled wild-type cotyledons (Fig. 6, B and C). We further found that the pen3 ibr1 ibr3 ibr10 quadruple mutant exhibited cotyledon sizes similar to the ibr1 ibr3 ibr10 triple mutant parent (Fig. 6D). The cotyledons of wild type, pen3, ibr1 ibr3 ibr10, and pen3 ibr1 ibr3 ibr10 appeared to be similarly sized soon after emergence from the seed coat, but pen3 mutant cotyledons expanded more rapidly and ibr1 ibr3 ibr10 and pen3 ibr1 ibr3 ibr10 mutant cotyledons expanded more slowly than those of the wild type (Fig. 6D), suggesting that the small cotyledons of seedlings with impaired IBA-to-IAA conversion result from a cell expansion defect rather than delayed germination.

DISCUSSION
The active auxin pool is tightly controlled through regulation of IAA biosynthesis, regulation of IAA transport, and use of storage forms. One means of controlling free IAA levels is through the formation and hydrolysis of IAA conjugates, wherein IAA is conjugated to amino acids, peptides, or sugars, to be released by hydrolases when free IAA is needed (Bartel et al., 2001;Woodward and Bartel, 2005a). A second potential auxin storage form is the chainlengthened compound IBA, which requires peroxisomal b-oxidation to IAA for auxin activity (Bartel et al., 2001;Woodward and Bartel, 2005a).
Because hydrolysis of IAA-amino acid conjugates also can contribute to the active auxin pool, we examined root hairs in the triple auxin-conjugate hydrolase mutant ilr1-1 iar3-2 ill2-1 (Rampey et al., 2004). Lightgrown ilr1 iar3 ill2 seedlings have slightly decreased free IAA levels, a longer primary root, and a shorter hypocotyl than wild type (Rampey et al., 2004). We found that ilr1 iar3 ill2 had an approximately 30% reduction in root hair length, compared to the approximately 55% reduction in root hair length in the ibr1 ibr3 ibr10 triple mutant, suggesting that IBA-derived IAA may play a more substantial role in driving root hair expansion than conjugate-derived IAA. The reduced hypocotyl elongation phenotype of light-grown ilr1 iar3 ill2 triple mutant seedlings ( Fig. 6A; Rampey et al., 2004) was not observed in ibr1 ibr3 ibr10 (Fig.  6A), suggesting that IBA-derived IAA plays a negligible role in hypocotyl elongation. Conversely, ilr1 iar3 ill2 cotyledon expansion was not impaired, suggesting that conjugate-derived IAA plays a minimal role in this growth process. However, the hydrolase and ibr mutations are present in different Arabidopsis accessions, and a direct comparison of the importance of conjugates versus IBA in providing IAA for root hair elongation, hypocotyl elongation, and cotyledon expansion would be facilitated by the development of a triple hydrolase mutant in the Columbia (Col) accession.
Active auxin pools are controlled by a variety of processes, including regulation of biosynthesis, transport, and storage (for review, see Woodward and Bartel, 2005a). Our data suggest that IBA-derived IAA drives cell expansion in specific tissues, including root hair and cotyledon cells, indicating that IBA is an important IAA precursor in young Arabidopsis seedlings. The reverse reaction, the formation of IBA from IAA, is catalyzed by an IBA synthetase (for review, see Ludwig-Mü ller 2000) that is induced by drought and abscisic acid in maize (Zea mays) seedlings (Ludwig-Mü ller et al., 1995). It will be interesting to learn whether and how IBA-to-IAA conversion is regulated and whether IBA b-oxidation contributes to active IAA in other tissue types and developmental stages.
To examine Suc-dependent hypocotyl elongation in the dark, seeds were plated on PN or PNS, incubated under white light for 1 d, and then incubated in the dark for an additional 4 d, after which hypocotyl lengths were measured. To examine hypocotyl elongation in the light, seedlings were grown under continuous illumination through yellow long-pass filters for 8 d on PNS at either 22°C or 28°C, and hypocotyl lengths were measured.
To examine auxin-responsive root elongation, seedlings were grown under yellow long-pass filters to slow indolic compound breakdown (Stasinopoulos and Hangarter, 1990) for 8 d on PNS supplemented with the indicated auxin, and the lengths of primary roots were measured.
To examine root hairs, seedlings were vertically grown for 5 d at 22°C under continuous white light (unless otherwise noted), the 4-mm root sections adjacent to the root-shoot junction were imaged using a dissecting microscope, and root hair lengths were measured using National Institutes of Health (NIH) Image software.
To examine cotyledon expansion, cotyledons of seedlings grown under continuous white light at 22°C were removed, mounted, and imaged through a dissecting microscope. Cotyledon areas were measured using NIH Image software.
Synthesis of [ 13 C 8 -15 N 1 ]IBA Synthesis of [ 13 C 8 -15 N 1 ]IBA was modified from previously described methods (Cohen and Schulze, 1981;Sutter and Cohen, 1992). A total of 0.08 mol of g-butyrolactone, 4.26 3 10 24 moles (0.05 g) of [U-13 C 8 , 98 atom%+; 15 N, 96 atom% to 99 atom%]indole (Cambridge Isotope Laboratories Inc.), and 0.08 mol of NaOH were mixed in a 23-mL Teflon insert fitted into a screw-top reaction bomb (Parr Instruments). This mixture was incubated in a heating mantle, temperature was increased 2°C/min, and then the mixture was held at 220°C for 24 h. The reaction was stopped by the addition of 50 mL water, and the mixture stirred in an Erlenmeyer flask until the sample was dissolved. The reaction mixture was extracted with chloroform twice to remove any unreacted indole, and the aqueous layer brought to pH 2.5 with HCl. [ 13 C 8 -15 N 1 ]IBA was then partitioned 33 into ethyl actetate from the aqueous layer.
Feeding Assays with [ 13 C 8 -15 N 1 ]IBA Eight-day-old seedlings (50-130 mg fresh weight) were incubated in 250 mL uptake buffer (10 mM MES, 10 mM Suc, and 0.5 mM CaSO 4 , pH 5.6) supplemented with 10 mM [ 13 C 8 -15 N 1 ]IBA. Seedlings were incubated for 1 h at room temperature, rinsed four times in uptake buffer, placed in a 1.5-mL tube, frozen in liquid nitrogen, and stored at 280°C prior to analysis.

Quantification of Auxins
For free [ 13 C 8 -15 N 1 ]IAA analysis, 150 mL of homogenization buffer (65% isopropanol, 35% 0.2 M imidazole buffer, pH 7.0) containing 7.5 ng [ 13 C 6 ]IAA internal standard (99 atom%, Cambridge Isotope Laboratories; Cohen et al., 1986) were added to each tissue sample from the [ 13 C 8 -15 N 1 ]IBA feeding assay. For free IAA and IBA analysis, 150 mL of homogenization buffer containing 10 ng [ 13 C 6 ]IAA and 10 ng [ 13 C 8 -15 N 1 ]IBA were added to each 50 mg fresh weight tissue sample (5-d-old seedlings grown under white light on PNS). Two 3-mm tungsten carbide beads (Qiagen) were added to each tube and samples were homogenized with a Mixer-Mill (Qiagen) for 5 min at 25 Hz, then incubated on ice for 1 h to allow the internal standards to equilibrate with the endogenous auxins in the extracts.
After equilibration, samples were centrifuged for 5 min at 10,000g and 100 mL of supernatant were collected and placed into a deep 96-well plate (Continental Lab Products). IAA extraction was performed according to Barkawi et al. (2008), with the IAA eluted in 600 mL MeOH and transferred to 1.5 mL screw-capped glass vials. For methylation, 900 mL ethereal diazomethane (Cohen, 1984) was added to each sample, and tubes were capped and incubated for 5 min at room temperature, then dried under N 2 in a 55°C sand bath and resuspended in 30 mL ethyl acetate. Samples were analyzed by GC-MS-SIM using an Agilent 6890 GC/5973 MS (Agilent Technologies) run in electron impact ionization mode at 70 eV and equipped with a fused silica capillary column (HP-5MS, 30 m 3 0.25-mm i.d., 0.25 mm film). The injector temperature was 280°C and the GC oven temperature was programmed to ramp from 70°C to 280°C at 10°C/min. Helium was used as the carrier gas at a flow rate of 1 mL/min. IAA and IBA levels were calculated by monitoring ions at mass-to-charge ratio (m/z) 130 and 189 for endogenous IAA, m/z 139 and 198 for [ 13 C 8 -15 N 1 ] IAA formed from added [ 13 C 8 -15 N 1 ]IBA, m/z 136 and 195 for the [ 13 C 6 ]IAA standard, m/z 130 and 217 for endogenous IBA, and m/z 139 and 226 for the [ 13 C 8 -15 N 1 ]IBA standard. Quantities were calculated by using standard isotope dilution equations (Cohen et al., 1986).

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Auxin levels in whole ibr1 ibr3 ibr10 seedlings resemble those in wild type.
Supplemental Figure S2. Auxin transport inhibitors only partially restore root hair elongation to ibr1 ibr3 ibr10.