Biochemical analyses of indole-3-acetaldoxime-dependent auxin biosynthesis in Arabidopsis

Abstract

Auxins are hormones that regulate many aspects of plant growth and development. The main plant auxin is indole-3-acetic acid (IAA), whose biosynthetic pathway is not fully understood. Indole-3-acetaldoxime (IAOx) has been proposed to be a key intermediate in the synthesis of IAA and several other indolic compounds. Genetic studies of IAA biosynthesis in Arabidopsis have suggested that 2 distinct pathways involving the CYP79B or YUCCA (YUC) genes may contribute to IAOx synthesis and that several pathways are also involved in the conversion of IAOx to IAA. Here we report the biochemical dissection of IAOx biosynthesis and metabolism in plants by analyzing IAA biosynthesis intermediates. We demonstrated that the majority of IAOx is produced by CYP79B genes in Arabidopsis because IAOx production was abolished in CYP79B-deficient mutants. IAOx was not detected from rice, maize, and tobacco, which do not have apparent CYP79B orthologues. IAOx levels were not significantly altered in the yuc1 yuc2 yuc4 yuc6 quadruple mutants, suggesting that the YUC gene family probably does not contribute to IAOx synthesis. We determined the pathway for conversion of IAOx to IAA by identifying 2 likely intermediates, indole-3-acetamide (IAM) and indole-3-acetonitrile (IAN), in Arabidopsis. When ¹³C₆-labeled IAOx was fed to CYP79B-deficient mutants, ¹³C₆ atoms were efficiently incorporated to IAM, IAN, and IAA. This biochemical evidence indicates that IAOx-dependent IAA biosynthesis, which involves IAM and IAN as intermediates, is not a common but a species-specific pathway in plants; thus IAA biosynthesis may differ among plant species.

Plants produce small molecule hormones for intracellular and intercellular signal transduction in response to developmental and environmental cues. Auxins are 1 type of plant hormones that are involved in many developmental processes, including cell division, cell differentiation, phototropism, root gravitropism, apical dominance, and vascular differentiation (1). Although the structure of the predominant natural auxin, indole-3-acetic acid (IAA), has been known since the 1930s, elucidation of auxin biosynthesis has remained challenging probably because of the occurrence of multiple biosynthetic pathways (2–5). Both tryptophan (TRP)-dependent and independent IAA biosynthetic pathways have been proposed (2). Recently identified auxin biosynthetic genes all belong to the TRP-dependent pathway (3–5).

As shown in Fig. 1A, there are 4 proposed pathways for the biosynthesis of IAA from TRP in plants: (i) the CYP79B pathway, (ii) the YUCCA (YUC) pathway, (iii) the indole-3-pyruvic acid (IPA) pathway, and (iv) the indole-3-acetamide (IAM) pathway (5). In Arabidopsis, indole-3-acetaldoxime (IAOx) is known to be produced from the CYP79B pathway by the cytochrome P450 monooxygenases, CYP79B2 and CYP79B3 (6, 7). It has been suggested that IAOx is a common intermediate for the synthesis of IAA, 3-indolylmethyl-glucosinolate (IG), and camalexin (CL) in Arabidopsis (8–11). IG is a metabolite produced to deter herbivores (12, 13), and CL is a phytoalexin produced in Arabidopsis in response to pathogen infection (11, 14). The physiological importance of IAOx-dependent IAA biosynthesis has been demonstrated by analysis of cyp79b2 cyp79b3 double mutants, which have shorter hypocotyls and decreased IAA levels when grown at high temperatures (8). Disruption of IG biosynthesis resulted in auxin-overproduction phenotypes because of the presumed diversion of IAOx to IAA synthesis. Mutations in SUPERROOT1 (SUR1) and SUPERROOT2 (SUR2), which encode 2 enzymes in IG biosynthesis, cause the production of massive adventitious roots from hypocotyls (9, 10). However, the exact mechanisms by which IAOx is converted to IAA are not clear. Both indole-3-acetonitrile (IAN) and indole-3-acetaldehyde (IAAld) are suggested as potential intermediates (9, 15), but the biosynthetic routes and the genes involved have not been identified. Moreover, IAN has also been proposed as an intermediate in IAA biosynthesis via IG metabolism (5). Because the production of IAN from IG occurs in response to tissue damage, it is not clear whether IG metabolism contributes to IAA biosynthesis under normal growth conditions.

Fig. 1.

Proposed IAA biosynthetic pathway in plants. (A) Previously proposed IAA biosynthetic pathway. The YUC and CYP79B pathways that lead to IAOx synthesis are illustrated by blue and red arrows, respectively. (B) The IAA biosynthetic pathway proposed in this study. Each number indicates a proposed pathway for biosynthesis of IAA from TRP. The bold arrows indicate novel pathways converting IAOx to IAA. Dotted square indicates the IAOx metabolic pathway in Arabidopsis. The IAOx-dependent IAA biosynthetic pathway is shown in the gray squares. Cloned genes that potentially encode enzymes for IAA, CL, and IG biosyntheses are shown in italics.

IAA biosynthesis via the YUC pathway is essential for many developmental processes including embryogenesis, seedling growth, vascular patterning, and flower development (16, 17). Overexpression of YUC genes leads to elongated hypocotyls and epinastic cotyledons, and inactivation of YUC genes causes severe developmental phenotypes that can be rescued by in situ production of IAA (17). YUC was proposed to catalyze the conversion of tryptamine (TAM) to N-hydroxy-tryptamine (HTAM), a rate-limiting step in a TRP-dependent IAA biosynthesis pathway (18). The roles of YUC genes in IAA biosynthesis and plant development have been extended to rice, maize, and petunia (19–21). Unlike the YUCs that exist throughout the plant kingdom, the CYP79B family has so far only been identified in Arabidopsis, Brassica napus, and Sinapis alba (22), suggesting that the YUC pathway is likely a universally used auxin biosynthesis pathway whereas IAA synthesis by the CYP79B family may have more specific roles in crucifers.

Genetic studies of IAA biosynthesis in Arabidopsis have suggested that the CYP79B and YUC pathways may converge at IAOx (5, 8, 10). It has also been proposed that IAOx is converted to IAA via several possible routes as shown in Fig. 1A (5). To test these ideas, we analyzed the relative contribution of the 2 pathways to the production of IAOx in Arabidopsis and other plants. We also analyzed the possible intermediates in the pathway converting IAOx to IAA. In this article, we show that the majority of IAOx is produced by the CYP79B family in Arabidopsis. The IAOx levels in the cyp79b2 cyp79b3 double mutants, and in rice, maize, and tobacco that do not have apparent CYP79B orthologues were below our detection limits. The YUC pathway probably does not contribute to IAOx synthesis, because IAOx levels were not significantly altered in the yuc quadruple mutants. We found that IAM and IAN are intermediates in IAOx-mediated IAA biosynthesis in Arabidopsis. This biochemical evidence indicates that IAOx-dependent IAA biosynthesis is not a common but a species-specific pathway in plants.

Results

The Majority of IAOx Is Produced by CYP79B2 and CYP79B3 in Arabidopsis.

We analyzed the IAOx levels in the cyp79b2 cyp79b3 double mutants to determine the contribution of CYP79B pathway to IAOx production in Arabidopsis (Fig. 1A). We used D₅-IAOx as an internal standard and purified IAOx from plants using HPLC and solid phase extraction columns. From WT seedlings, we detected MS/MS ions ([M + H − 17]⁺) corresponding to trans and cis forms of endogenous IAOx at m/z 158.1 and those of D₅-IAOx at m/z 163.1; see supporting information (SI) Figs. S1A and S1B. The identity of IAOx was confirmed by in vivo feeding of a [¹³C₈, ¹⁵N]indole to Arabidopsis where the MS/MS ions for endogenous IAOx showed an increase of 9 mass units (Fig. S2). As shown in Table 1, the level of IAOx in WT plants was 1.7 ± 0.1 ng/gram fresh weight (gfw) (n = 2) by liquid chromatography-electrospray ionization-mass/mass spectrometry (LC-ESI-MS/MS) analysis. The level of IAOx was increased in sur1–1 seedlings (2.5-fold) as compared to that in WT seedlings (Table 1). We measured IAOx levels in 3 T-DNA insertion lines of cyp79b2 cyp79b3 (Fig. S3). All of the 3 cyp79b2 cyp79b3 null lines did not show a visible phenotype on MS agar media under our growth conditions (Fig. S3), although cyp79b2 cyp79b3 previously showed subtle growth phenotypes in soil (8). Endogenous IAOx was not detected from these mutants (Table 1 and Figs. S1C and Figs. S1D). These results suggested that IAOx is mainly produced by CYP79B2 and CYP79B3 in Arabidopsis (Fig. 1B).

Table 1.

The levels of IAA and its precursors in Arabidopsis, rice, maize, and tobacco

Plants	Intermediates (ng/gfw)
	IAOx	IAN	IAM	IAA
A. thaliana	1.7 ± 0.1*	9,720 ± 2,120	9.9 ± 2.1	11.2 ± 2.9
cyp79b2–1 cyp79b3–1	ND*	ND	0.5 ± 0.1	11.4 ± 1.7
cyp79b2–2 cyp79b3–2	ND*	ND	0.6 ± 0.1	11.5 ± 1.2
cyp79b2–1 cyp79b3–3	ND*	ND	0.4 ± 0.1	10.8 ± 3.4
sur1–1	4.2 ± 1.1	9,230 ± 2,130	337 ± 4	235 ± 9
Rice	ND*	ND	3.2 ± 0.9	23.2 ± 5.7
Maize	ND*	ND	11.2 ± 0.1*	93.1 ± 22.8
Tobacco	ND*	ND	1.0 ± 0.5	11.3 ± 5.2

Seedlings of Arabidopsis (2-week-old), coleoptiles of rice (2-day-old), and maize (3-day-old), and apexes of tobacco (2-month-old) were used for quantifying the levels of IAA and its precursors. Data are means ± SD (n = 3). *Quantification was performed in two independent experiments (data are means ± SD, n = 2). ND, not detected.

To investigate the contribution of the YUC genes to IAOx production in Arabidopsis, we analyzed the IAOx levels in yuc1 yuc2 yuc4 yuc6 quadruple mutants. The aerial parts of 4-week-old yuc1 yuc2 yuc4 yuc6 seedlings that showed growth defects in soil were used for IAOx analysis (Fig. S4). The IAOx level was not significantly altered in yuc1 yuc2 yuc4 yuc6 seedlings (2.9 ± 0.4 ng/gfw, n = 3) as compared to that in the WT plants (2.9 ± 0.7 ng/gfw, n = 3), suggesting that IAOx is not mainly produced through the YUC pathway in Arabidopsis. To further exclude a possibility that IAOx is involved in the YUC pathway, we fed ¹⁵N₂-TAM to WT seedlings and analyzed ¹⁵N₂-incorporation to IAOx. Arabidopsis WT plants produce TAM via the TRP metabolic pathway (Fig. S2) and possessed 209 ± 15 pg/gfw (n = 4) of this compound. However, ¹⁵N-incorporation into IAOx was not observed (< 1%) even though seedlings were cultured in MS liquid media containing ¹⁵N₂-TAM (100 μM) for 10 days, while 21.5 ± 3.5% (n = 2) of total IAA was ¹⁵N-labeled in the same plants. All of these results suggest that the YUC pathway may not produce IAOx in Arabidopsis (Fig. 1B).

We also analyzed the levels of IAOx in rice, maize, and tobacco, which are noncruciferous plants. As they do not have apparent orthologues of CYP79B genes, these plants can also be used to determine the contribution of CYP79B genes to IAOx production. As shown in Table 1, we detected IAA, but not IAOx from all of these plants, indicating that IAOx is most likely not a common but a species-specific metabolite of plants such as Arabidopsis (Fig. 1B).

IAN, IG, and CL Are Each Produced Independently from IAOx in Arabidopsis.

IAN has been proposed as a common metabolite of both the CYP79B and YUC pathways as shown in Fig. 1A (3). We quantified IAN levels in 3 independent cyp79b2 cyp79b3 mutants by GC-MS to clarify the contribution of the CYP79B pathway to IAN production. As shown in Table 1, IAN was not detected in any of the cyp79b2 cyp79b3 mutant seedlings in contrast to a previous report (8). We found that treatment with IAOx at 3 μM and 10 μM for 24 h can restore the levels of IAN in cyp79b2 cyp79b3 mutants dose dependently to 8% and 23% of those in WT plants, respectively (Table 2). Moreover, endogenous IAN was completely ¹³C-labeled by feeding ¹³C₆-IAOx to cyp79b2 cyp79b3 mutants under the same conditions (Fig. 2A). In addition, we fed ¹⁵N₂-TAM to WT seedlings and analyzed ¹⁵N₂-incorporation to IAN to elucidate the contribution of YUC genes to IAN synthesis. As in the case of IAOx, we could not observe the enrichment of ¹⁵N atoms in IAN (<1%). These results indicate that the majority of IAN in Arabidopsis is derived from the CYP79B pathway but not from the YUC pathway (Fig. 1B).

Table 2.

The levels of IAA and its precursors in IAOx-treated cyp79b2 cyp79b3 and WT seedlings

	IAOx (μM)	Intermediates (ng/gfw)
		IAN	IAM	IAA
cyp79b2–2 cyp79b3–3	3	779	4.0	21.1
	10	2,280	16.0	23.5
WT	3	8,630	12.8	35.9
	10	9,660	20.1	54.8

Seedlings of cyp79b2–2 cyp79b3–2 and WT (10-day-old) were transferred to MS agar media IAOx containing, where they were incubated for 24 h (at 21°C) aseptically, before being subjected to analysis.

Fig. 2.

In vivo feeding of ¹³C₆-IAOx to cyp79b2 cyp79b3 and WT seedlings. (A) Incorporation of ¹³C₆ into metabolites in cyp79b2–2 cyp79b3–2 seedlings fed with 3 μM and 10 μM of ¹³C₆-IAOx. (B) Incorporation of ¹³C₆ into metabolites in WT seedlings fed with 10 μM of ¹³C₆-IAOx. Ten-day-old seedlings were fed with ¹³C₆-IAOx for 24 h at 21 °C. Data are means ± SD, n = 2.

IG metabolism has been proposed as a major pathway for IAN synthesis (Fig. 1A), while myrosinase (MYR) generates IAN from IG in response to tissue damage (5, 12, 13). To investigate this notion, we analyzed IAN levels in sur1–1 null mutants, which do not produce IG (10). We found that IAN levels were not significantly reduced in sur1–1 mutants in comparison to those in the WT plants (Table 1), suggesting that IG metabolism is not a major pathway for IAN synthesis under normal growth conditions. To test if increasing the metabolic flux from IAOx to IAA can affect the level of IAN in sur1–1 mutants, we applied IAOx to WT plants and quantified IAN levels. As shown in Table 2, application of IAOx increased IAA levels, but did not significantly alter the level of IAN in WT plants. Arabidopsis also produces IAN from IAOx by CYP71A13 for the synthesis of CL in response to pathogen attack (14). Consistent with this idea, IAN levels were not decreased in cyp71a13 knockout mutants (16.7 ± 2.2 μg/gfw, n = 3) without pathogen attack. These results suggest that IAN, IG, and CL are independently produced from IAOx in Arabidopsis under normal growth conditions (Fig. 1B).

We also analyzed the levels of IAN in rice, maize, and tobacco. As shown in Table 1, we detected IAA, but not IAN from all 3 plant species we investigated, suggesting that IAN is not a common intermediate in IAA biosynthesis. However, as IAN was previously detected in maize coleoptiles (23) and cyp79b2 cyp79b3 double mutants (8), we cannot completely exclude the possibility that IAN may be a common intermediate in IAA biosynthesis. IAN may be detected if we increase the amount of plant material used for IAN analysis, use plants grown under different conditions, or use previously reported methods (23, 24).

IAM Is Involved in IAA Biosynthesis via IAOx.

Arabidopsis nitrilases (NIT1, NIT2, and NIT3) can convert IAN to IAA and IAM in vitro (25). Since our results indicated that IAN is mainly produced from IAOx, we hypothesized that IAM is a possible intermediate in the pathway converting IAOx to IAA (Fig. 1B). If that is the case, we expect IAM levels to be reduced in cyp79b2 cyp79b3 and increased in sur1–1 mutants. We detected a MS/MS ion ([M + H − 45]⁺) for endogenous IAM at m/z 130.1 from WT seedlings along with a MS/MS ion for ¹³C₆-IAM at m/z 136.1 as an internal control peak (Fig. S5). As shown in Table 1, the IAM level in WT plants was 9.9 ± 2.1 ng/gfw (n = 3). We found that IAM levels were drastically decreased to approximately 0.4–0.6 ng/gfw in all cyp79b2 cyp79b3 mutants, but substantially increased (more than 34-fold) in sur1–1 seedlings. These results suggest that the majority of IAM in Arabidopsis is produced from IAOx.

Synthesis of IAM from IAOx was directly demonstrated by in vivo labeling experiments. First, we examined the restoration of IAM levels in cyp79b2 cyp79b3 mutants by application of IAOx. As shown in Table 2, the IAM level was moderately restored in cyp79b2 cyp79b3 seedlings (40.4%) by applying IAOx (3 μM) for 24 h. We found that IAM was ¹³C₆-labeled (83.3%, n = 2) when ¹³C₆-IAOx was fed to cyp79b2 cyp79b3 seedlings under the same conditions (Fig. 2A). IAM was even more efficiently ¹³C₆-labeled (97.5%, n = 2) when cyp79b2 cyp79b3 seedlings were fed with 10 μM of ¹³C₆-IAOx (Fig. 2A). IAA was moderately labeled from IAOx in these conditions (Fig. 2A). These results provided further evidence indicating that IAM is mainly produced from IAOx in Arabidopsis (Fig. 1B).

To determine if IAN is a main precursor of IAM, we performed an in vivo labeling experiment with WT plants. Since the IAN levels are nearly 3 orders of magnitude higher than those of IAM in WT plants (Table 1), we speculated that ¹³C-enrichment of IAN and IAM in ¹³C₆-IAOx-fed plants would be different if these intermediates are produced independently from IAOx. We found that the ¹³C-enrichment of IAM and IAA was approximately 6-fold higher than that of IAN in the WT plants fed with ¹³C₆-IAOx (Fig. 2B). The greater ¹³C-enrichment of IAM relative to IAN suggests that IAN and IAM are produced independently from IAOx in Arabidopsis (Fig. 1B).

It was previously demonstrated that cyp79b2 cyp79b3 double mutants show growth defects and reduced IAA levels as compared to WT plants at 26 °C (8). We therefore examined the phenotypic complementation of cyp79b2 cyp79b3 mutants by applying IAM. As shown in Fig. 3 A and B, cyp79b2 cyp79b3 mutants displayed the growth defects at 26 °C under our growth conditions. We found that the phenotype of cyp79b2 cyp79b3 mutants can be restored to WT levels by treatment with 10 μM of IAM for 7 days at 26 °C (Fig. 3 C and D). The level of IAA in cyp79b2 cyp79b3 mutants was 20% lower than that in WT plants at 26 °C, but it was also 30% increased by application of 10 μM IAM. IAM rescue of cyp79b2 cyp79b3 mutants depends on IAA production because auxin resistance 1–3 (axr1–3), in which the auxin signal pathway is disrupted, can suppress IAM-dependent hypocotyl elongation (Fig. S6). In contrast, application of IAA or IAOx to cyp79b2 cyp79b3 mutants resulted in the inhibition of root growth and did not complement the phenotype under the same conditions (data not shown), presumably because of the difference in diffusion and/or metabolic rate between IAM and these compounds in plants as suggested previously (26). Together, all of these results indicated that IAM is an intermediate of IAOx-dependent IAA biosynthesis in Arabidopsis (Fig. 1B).

Fig. 3.

Effect of IAM on the growth of cyp79b2 cyp79b3 at high temperatures. (A–C) Seven-day-old seedlings of WT, cyp79b2–2 cyp79b3–2, and 10 μM IAM-fed cyp79b2–2 cyp79b3–2. Seedlings were germinated and grown at 26 °C. (D) Fresh weight of WT and cyp79b2–2 cyp79b3–2 seedlings grown in the presence of 0, 1, 2, 5, and 10 μM of IAM for 7 days. Data are means ± SD, n = 16. (Scale bar, 2 cm.)

We analyzed IAM levels in rice, maize, and tobacco by LC-ESI-MS/MS. As shown in Table 1, we have detected IAM from all of these plants. Because these plants do not have apparent CYP79B orthologues and are devoid of IAOx, our data suggest that IAM is also produced in the IAOx-independent pathway, but the mechanisms are still unknown (Fig. 1B).

Simultaneous Application of IAM and IAN to WT Plants Phenocopies sur1 Mutants.

Although IAM and IAN are identified as metabolites of IAOx, it remains unclear how IAA is overproduced in sur1 mutants. Because IAM levels are greatly increased in sur1–1 mutants (Table 1), we investigated if WT plants show sur1-like phenotype by application of IAM. Application of IAM (60 μM) triggered elongation of hypocotyls and petioles of WT plants, but did not cause formation of massive adventitious roots like that observed in sur1–1 (Fig. 4 A–C). In contrast, application of 30 μM IAN to WT plants results in the massive adventitious roots and root growth inhibition (Fig. 4D) as reported previously (27). We found that simultaneous application of IAM and IAN to WT plants phenocopied sur1–1 mutants (Fig. 4E), suggesting that both IAM and IAN may be overproduced in sur1–1. Since the level of IAN was not increased in sur1–1 and IAOx-treated WT plants (Tables 1 and 2), excessive IAN may be promptly converted to IAM and IAA in Arabidopsis. To test this idea, we applied D₂-IAN (30 μM) to WT plants and analyzed D₂-enrichment in IAM and IAA by LC-ESI-MS/MS. We found that 69.9% and 34.8% of total IAM and IAA, respectively, were labeled with D₂ atoms. These results indicate that IAA is likely overproduced in sur1 mutant by conversion of IAOx to both IAN and IAM (Fig. 1B).

Fig. 4.

Effect of IAM and IAN on the growth of Arabidopsis seedlings. (A–E) Two-week-old seedlings of WT, sur1–1, WT treated with 60 μM IAM, WT treated with 30 μM IAN, and WT treated with 60 μM IAM and 30 μM IAN. Arrows indicate outgrowth of numerous adventitious roots. (Scale bar, 3 mm.)

Discussion

In this work we demonstrated that the majority of IAOx is produced from the CYP79B pathway because cyp79b2 cyp79b3 mutants and noncruciferous plants such as rice, maize, and tobacco did not produce IAOx. We also showed that the YUC pathway probably does not contribute to IAOx synthesis, because the IAOx levels were not significantly altered in yuc quadruple mutants and ¹⁵N₂-incorporation to IAOx was not observed after feeding ¹⁵N₂-TAM to Arabidopsis. Furthermore, we showed that IAOx is converted into IAN and IAM, the 2 likely intermediates in a pathway that converts IAOx into IAA.

Biosynthesis of IAOx.

Genetic studies have identified several genes in IAA biosynthesis, but it has been difficult to determine whether both the CYP79B and YUC pathways contribute to IAOx synthesis. Analysis of IAOx levels in CYP79B-deficient and YUC-deficient mutants is critical for biochemical dissection of these pathways. It has been assumed that oximes are unstable compounds that do not accumulate in the cell, and the low K_s and K_m of SUR2 for IAOx would prevent accumulation of IAOx in plants (9). The analysis of IAOx in Chinese cabbage was previously performed using GC-MS and proved difficult, because the full-mass spectrum of endogenous IAOx was not completely identical with that of authentic IAOx because of the presence of contaminant ions (28). Here we demonstrated that IAOx is detectable using LC-ESI-MS/MS coupled with a rapid IAOx purification from plants under neutral conditions. The analysis of IAOx in WT, cyp79b2 cyp79b3 and yuc quadruple mutants of Arabidopsis indicate that the majority of IAOx is produced by CYP79B enzymes. Our results from in vivo feeding experiments suggest that TAM, a proposed substrate of YUC, may not be a precursor of IAOx in Arabidopsis.

It is well established that IAOx can be produced from TRP by CYP79B2 and CYP79B3, but it was not clear whether other pathways also contribute to the production of IAOx. We demonstrated that IAOx was not detectable in rice, maize, and tobacco that presumably do not have CYP79B genes. Similar to CYP79B2, Arabidopsis CYP79A2 catalyzes the conversion of L-phenylalanine to phenylacetaldoxime (29). Rice, which does not have the CYP79B family, possesses the CYP79A family (30). Although CYP79A2 catalyzes a similar enzyme reaction, our results suggested that the CYP79A family do not contribute to the production of IAOx in Arabidopsis and rice. These results suggest that IAOx is produced from the CYP79B pathway but not from other pathways in the model plants we investigated.

Conversion of IAOx to IAA.

Genetic evidence demonstrated that IAOx can be converted to IAA, but the exact mechanisms are not known. We have shown that IAOx is converted to IAN in Arabidopsis plants (Fig. 2). It is known that IAN has IAA-like effects (27), and presumably IAN is converted to IAA in vivo. We also discovered that IAOx is converted to IAM (Fig. 2), which is a well-known IAA biosynthesis intermediate. It appears that IAM and IAN are independently produced from IAOx, because ¹³C-enrichment of IAN from ¹³C₆-IAOx was much lower than that of IAM in WT plants (Fig. 2B). It is known from animal systems that oximes can be converted to amides by a Beckmann-rearrangement-type enzymatic reaction requiring NADPH (31). Therefore, a similar enzymatic reaction might be involved in the formation of IAM from IAOx in Arabidopsis. However, we do not exclude the possibility that IAN is a direct precursor of IAM in Arabidopsis (Fig. S7), because it has previously been proposed that the metabolic pool of IAN relevant for IAA biosynthesis is separated from the larger second pool by strict compartmentation or as a consequence of metabolite channeling (32). This was suggested by the fact that IAN was labeled from D₅-TRP in WT plants, but its isotopic abundance was lower than that of IAA. It will be important to elucidate whether the multiple knockout mutants of nitrilases (NIT1, NIT2, and NIT3) show reduced IAM levels under normal conditions, and growth defects under high temperature conditions like those of the cyp79b2 cyp79b3 mutants.

The Role of IAOx in IAA Biosynthesis.

It is evident that IAA can be made from several TRP-dependent pathways (Fig. 1) (2–5). IAOx was proposed as a key intermediate in IAA biosynthesis on the basis of several observations. (i) Overexpression of CYP79B2 leads to high auxin phenotypes such as elongated hypocotyls similar to known auxin overproduction lines such as yuc1D and iaaM overexpression lines (8, 33). (ii) Inactivation of SUR1 or SUR2 blocks IG biosynthesis, and leads to auxin overproduction phenotypes because of the diversion of IAOx to IAA synthesis (9, 10). (iii) Overexpression of SUR2 results in a bushy phenotype resembling multiple YUC knockout mutants (9, 17). It was unclear whether IAOx-mediated IAA biosynthesis plays an essential role in plant growth and development for 2 reasons. First, inactivation of CYP79B2 and CYP79B3 caused subtle developmental defects and only a slight decrease in IAA levels (8). cyp79b2 cyp79b3 double mutants did not show a visible phenotype or a decrease in IAA levels under our standard growth conditions (Table 1 and Fig. S3). Furthermore, CYP79B genes appear not to exist universally throughout the plant kingdom. Our biochemical analysis of IAOx levels in cyc79b2 cyp79b3 indicated that IAOx-mediated IAA biosynthesis is probably not responsible for the production of the bulk of IAA in Arabidopsis. The CYP79B pathway does play a role in IAA production when Arabidopsis plants are grown at higher temperatures (8). In plants grown at 26 °C, we observed apparent growth defects and decreased IAA levels in cyp79b2 cyp79b3 mutants in comparison to WT plants, and the phenotype and IAA levels could be rescued by treatment with IAM (Fig. 3).

Genetic studies on the YUC pathway and the IPA pathway suggest that the 2 pathways play a more prominent role in plant growth and development than IAOx-mediated IAA biosynthesis. The YUC genes have been shown to be essential for embryogenesis, seedling growth, vascular patterning, and flower development (16, 17). The IPA pathway also plays a key role in embryogenesis and shade avoidance (34, 35). Unlike the CYP79B genes, the YUC and TAA1 gene families appear to be widely distributed in plants, suggesting that the YUC and IPA pathways are universally conserved for IAA biosynthesis. Our data suggest that IAOx-dependent IAA biosynthesis is probably limited to Arabidopsis and other crucifers and plays a more restricted role in plant growth and development. Further analysis of IAOx, as well as CYP79B, in other plant species may allow us to understand the evolution of IAOx-dependent IAA biosynthesis and whether CYP79B is the only IAOx-producing pathway.

Materials and Methods

Plant Materials and Growth Conditions.

All mutants are in the Arabidopsis thaliana Col-0 background. sur1–1, cyp79b2–1, and cyp79b2–2 mutant seeds were obtained from the Arabidopsis Biological Resource Center. cyp79b3–2 and cyp79b3–3 mutant seeds were from the Nottingham Arabidopsis Stock Centre. cyp79b2 cyp79b3 double mutants were generated by crossing the corresponding single mutants as described in SI Text Materials and Methods. Plants were grown at 21 or 26 °C under continuous light, 30–50 μmol of photons per m² per second, with cool-white illumination. Plants were germinated and grown in Murashige-Skoog (MS) agar media (pH 5.7) supplemented with thiamin hydrochloride (3 μg/mL), nicotinic acid (5 μg/mL), pyridoxin hydrochloride (0.5 μg/mL), myo-inositol (100 μg/mL) and 1% (wt/vol) sucrose.

Rice (Oryza sativa) seeds were germinated in H₂O, and grown at 30 °C on agar media in the dark. Two days after germination, coleoptiles were excised from the seedlings with a razor blade and kept at −80 °C until use. Coleoptile tips of maize (Zea mays) were prepared as previously reported (36). WT tobacco (Nicotiana tabacum) plants were grown on soil in a growth room at 25 °C with 16 h light for 2 months. Shoot apexes were excised from the adult plants with a razor blade and kept at −80 °C until use.

Chemical Synthesis, LC-ESI-MS/MS, GC-MS and Labeling Experiments.

IAOx, D₅-IAOx, ¹³C₆-IAOx, ¹³C₆-IAM, D₂-IAN, and ¹⁵N₂-TAM were synthesized as described in SI Text Materials and Methods. LC-ESI-MS/MS and GC-MS analyses of IAA and IAA intermediates and in vivo labeling experiments were performed as described in SI Text Materials and Methods.