See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/51822022

The Arabidopsis YUCCA1 Flavin Monooxygenase

functions in the Indole-3-Pyruvic acid branch of
Auxin Biosynthesis

ARTICLE in THE PLANT CELL · NOVEMBER 2011

Impact Factor: 9.58 · DOI: 10.1105/tpc.111.088047 · Source: PubMed

CITATIONS

79

8 AUTHORS, INCLUDING:

Ondrej Novak Wenrong He

Palacký University of Olomouc Peking University
148 PUBLICATIONS 2,583 CITATIONS 7 PUBLICATIONS 304 CITATIONS

SEE PROFILE SEE PROFILE

Hongwei Guo Jose Alonso

Peking University North Carolina State University
47 PUBLICATIONS 3,168 CITATIONS 101 PUBLICATIONS 16,748 CITATIONS

SEE PROFILE SEE PROFILE

Available from: Ondrej Novak

Retrieved on: 25 August 2015
The Plant Cell, Vol. 23: 3961–3973, November 2011, www.plantcell.org ã 2011 American Society of Plant Biologists. All rights reserved.

The Arabidopsis YUCCA1 Flavin Monooxygenase Functions in

the Indole-3-Pyruvic Acid Branch of Auxin Biosynthesis W

Anna N. Stepanova,a Jeonga Yun,a Linda M. Robles,a Ondrej Novak,b,c Wenrong He,d Hongwei Guo,d Karin Ljung,b
and Jose M. Alonsoa,1
a Department of Genetics, North Carolina State University, Raleigh, North Carolina 27695
b Umeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences,
SE–901 83 Umea, Sweden
c Laboratory of Growth Regulators, Faculty of Science, Palacký University and Institute of Experimental Botany Academy of

Sciences of the Czech Republic, CZ–783 71 Olomouc, Czech Republic

d State Key Laboratory of Protein and Plant Gene Research, College of Life Sciences, Peking University, Beijing 100871, China

The effects of auxins on plant growth and development have been known for more than 100 years, yet our understanding of
how plants synthesize this essential plant hormone is still fragmentary at best. Gene loss- and gain-of-function studies have
conclusively implicated three gene families, CYTOCHROME P450 79B2/B3 (CYP79B2/B3), YUCCA (YUC), and TRYPTOPHAN
AMINOTRANSFERASE OF ARABIDOPSIS1/TRYPTOPHAN AMINOTRANSFERASE-RELATED (TAA1/TAR), in the production
of this hormone in the reference plant Arabidopsis thaliana. Each of these three gene families is believed to represent
independent routes of auxin biosynthesis. Using a combination of pharmacological, genetic, and biochemical approaches,
we examined the possible relationships between the auxin biosynthetic pathways defined by these three gene families. Our
findings clearly indicate that TAA1/TARs and YUCs function in a common linear biosynthetic pathway that is genetically
distinct from the CYP79B2/B3 route. In the redefined TAA1-YUC auxin biosynthetic pathway, TAA1/TARs are required for the
production of indole-3-pyruvic acid (IPyA) from Trp, whereas YUCs are likely to function downstream. These results,
together with the extensive genetic analysis of four pyruvate decarboxylases, the putative downstream components of the
TAA1 pathway, strongly suggest that the enzymatic reactions involved in indole-3-acetic acid (IAA) production via IPyA are
different than those previously postulated, and a new and testable model for how IAA is produced in plants is needed.

INTRODUCTION
Since the discovery of the auxin indole-3-acetic acid (IAA) in the
1930s (Thimann and Koepfli, 1935), the regulatory roles and
molecular mechanisms of action of this key plant hormone have
become the research focus for many plant biologists. Over the
years, numerous reports of the critical involvement of this hor-
mone in nearly every aspect of a plant’s life have flooded
scientific literature (reviewed in Bartel, 1997; Woodward and
Bartel, 2005; Vanneste and Friml, 2009). Despite this massive
research effort, our current understanding of how this hormone is
made in plants remains, at best, fragmentary (Normanly, 2010;
Zhao, 2010). Based primarily on biochemical experimental
approaches, two general pathways for IAA production have
been proposed, the Trp-dependent and the Trp-independent
(Normanly et al., 1993). Very little is known about enzymatic
reactions and metabolic intermediates implicated in the Trp-
independent pathway (Östin et al., 1999), and none of the genes
involved have been characterized. It is, however, clear that this
1 Address correspondence to jmalonso@ncsu.edu.
The authors responsible for distribution of materials integral to the
findings presented in this article in accordance with the policy described
in the Instructions for Authors (www.plantcell.org) are: Anna N. Stepanova
(atstepan@ncsu.edu) and Jose M. Alonso (jmalonso@ncsu.edu).
W
Online version contains Web-only data.
www.plantcell.org/cgi/doi/10.1105/tpc.111.088047
pool of auxin is derived from the Trp precursors indole and/or
indole-3-glycerol phosphate (Ouyang et al., 2000). The situation
is somewhat less dire for the Trp-dependent pathway, where four
biosynthetic routes with candidate intermediates and potential
enzymatic activities (responsible for the conversion of these
intermediates into IAA) have been postulated (reviewed in
Normanly, 2010; Zhao, 2010) (Figure 1). Each of these four
putative biosynthetic routes is defined by the first metabolic
intermediate generated from Trp, namely: i) indole-3-acetaldox-
ime (IAOx), ii) indole-3-pyruvic acid (IPyA), iii) indole-3-acetamide
(IAM), or iv) tryptamine (TAM). Of these, the IAOx route is believed
to be largely restricted to the Brassica genus. In Arabidopsis
thaliana, IAOx is produced from Trp by a duo of highly related
cytochrome P450s, CYP79B2 and CYP79B3 (Zhao et al., 2002;
Sugawara et al., 2009). The bulk of IAOx produced is typically
used in the synthesis of defense compounds, such as indole
glucosinolates and camalexins (Glawischnig et al., 2004; Bender
and Celenza, 2009). Under standard growth conditions, IAOx can
also be converted to IAA via at least two intermediates, IAM and
indole-3-acetonitrile (IAN) (Sugawara et al., 2009); however, this
route is not believed to be a major source of IAA during normal
plant development. On the other hand, high auxin levels can be
reached when the conversion of IAOx into indole glucosinates is
blocked by means of genetic mutations, such as superroot1
(sur1), sur2, or UDP-glucosyltransferase 74b1 (ugt74b1) (Bak
et al., 2001; Grubb et al., 2004; Mikkelsen et al., 2004). In these
3962 The Plant Cell
Figure 1. Schematic Representation of the Auxin Biosynthesis Pathways in Arabidopsis.
Classical view (A) and updated view (B) of the pathway modified to reflect the findings of this study. Solid arrows indicate single steps, whereas dashed
arrows imply possible multiple steps. Arrows, gene names, and intermediates shown in gray versus black indicate hypothetical versus well-
characterized steps, enzymatic activities, and metabolites, respectively. AMI, INDOLE-3-ACETAMIDE HYDROLASE; HTAM, N-hydroxyl TAM; IAAld,
indole-3-acetaldehyde; IGs, indole glucosinolates; NIT, NITRILASE.
mutants, large quantities of IAA can be produced via IAOx, but in
this case, the potential metabolic steps between the precursor
IAOx and the final product IAA have not been determined.
Although several of the metabolic intermediates of the IAOx
biosynthetic branch (such as IAM or IAN) are shared with the
proposed IAM and TAM routes, the possible metabolic connec-
tions between these pathways are unknown.
Unlike the IAOx pathway, which is limited almost exclusively to
Brassica spp., the IPyA, IAM, and TAM routes of Trp-dependent
auxin biosynthesis are believed to be common to all plants,
based on the ubiquitous distribution of at least some of the genes
implicated in these pathways (De Smet et al., 2011). For example,
WEI8/TAA1, which encodes a Trp aminotransferase responsible
for the conversion of Trp into IPyA, has homologous sequences
across a wide array of evolutionarily diverse plant species,
ranging from mosses (such as Physcomitrella) to monocots
(such as maize [Zea mays] and rice [Oryza sativa]) and dicots
(such as Arabidopsis) (Yamada et al., 2009). In this IPyA pathway,
only the first biosynthetic step, the conversion of Trp into IPyA,
is well characterized at both biochemical and genetic levels
(Stepanova et al., 2008; Tao et al., 2008; Yamada et al., 2009). By
contrast, the two commonly postulated downstream steps, that
is, the conversion of IPyA into indole-3-acetaldehyde and then
into IAA by the consecutive action of a putative indole pyruvate
decarboxylase (PDC) and an indole aldehyde oxidase (AAO),
have very little experimental support in plants. Based on se-
quence similarities with the bacterial indole PDCs, in Arabidop-
sis, six putative PDCs can be grouped into two small families of
four (PDC1, PDC2, PDC3, and PDC4) and two genes (PDC-Like2
[PDL2] and PDL1) each (see Supplemental Figure 1 online). One
of the two genes of the latter family, PDL2/ALS/AHAS/CSR1/
IMR1/TZP5, has been functionally characterized and shown to
be involved in Val and Ile biosynthesis (Smith et al., 1989).
Similarly, two of the four members of the PDC family, PDC1 and
PDC2, have been implicated in ethanol fermentation (Ismond
et al., 2003; Kürsteiner et al., 2003). There are also four predicted
AAOs in Arabidopsis, AAO1, AAO2, AAO3, and AAO4. Although
initial experiments showed that AAO1 is highly expressed in the
sur2 auxin-overproducing mutant (Seo et al., 1998), no additional
evidence for a role of this gene in IAA biosynthesis was found.
Moreover, several of the members of this small gene family have
been shown to function in the biosynthesis of another plant
hormone, abscisic acid, raising additional doubts about the role
of AAOs in auxin production (Seo et al., 2000). Thus, in the
absence of direct in planta evidence for the role of the predicted
PDCs and AAOs in the IPyA pathway, the postulated current
models should be considered working hypotheses.
Another proposed route for IAA production in plants is that
defined by IAM. Although this pathway was originally believed to
be present only in auxin-synthesizing bacteria, the identification
of IAM in several plant species (Sugawara et al., 2009), as well as
the characterization of tobacco [Nicotiana tabacum] and Arabi-
dopsis IAM hydrolase enzymes capable of catalyzing the con-
version of IAM into IAA, have opened the possibility of this
pathway’s existence in plants (Pollmann et al., 2003; Nemoto
et al., 2009). Nevertheless, to date there is no compelling
evidence for a physiological role of this putative route of auxin
biosynthesis in plants under normal growth conditions.
Finally, perhaps the most puzzling of all routes of auxin
production is that defined by the YUC family of flavin monoox-
ygenases, which in Arabidopsis is represented by 11 genes. This
pathway is frequently referred to as the TAM route, because it
was believed to start by the conversion of Trp into TAM (Zhao
et al., 2001). TAM was then believed to be converted into
N-hydroxyl TAM by the action of YUCs and then to IAOx (Zhao
et al., 2001), an intermediate also shared by the CYP79B2/B3
branch of the auxin biosynthetic pathway. Recent findings have
substantially weakened this model. Sugawara et al. (2009)
showed that most, if not all, detectable IAOx in Arabidopsis is
likely produced via CYP79B2/B3, suggesting that the YUC

pathway does not converge on IAOx as previously believed.
In addition, serious doubts about the involvement of YUC1 in
the production of N-hydroxyl TAM as well as about the nature
of the in vivo substrate of this enzyme have recently been
raised (Tivendale et al., 2010). Thus, YUCs, despite their un-
disputed central role in auxin production, have yet to be posi-
tioned in a defined Trp-dependent route of auxin biosynthesis.
Based on the phenotypic similarities of many loss-of-function
mutants of the YUC and TAA1 gene families, as well as the
overlapping expression patterns of the corresponding genes, it
was suggested that YUCs might function together with TAA1 in
the IPyA route of IAA production (Strader and Bartel, 2008).
Further support for this possibility was obtained by the less-than-
additive phenotypes of the maize mutants vanishing tassel2 (vt2)
and sparse inflorescence1 (spi1) (Phillips et al., 2011), defective in
the likely orthologs of Arabidopsis TAA1 and YUC, respectively.
On the other hand, the observation that YUC1 overexpression
lines accumulate high levels of IAN, an IAA intermediate that has
been closely associated with the IAOx route of auxin production
(Nonhebel et al., 2011) but has no known link to the IPyA branch,
may argue against the idea of a shared TAA1-YUC pathway.
Despite the continuous improvements in the analytical tools
suitable for quantification of IAA and IAA intermediates and the
recent advances in the genetic characterization of several com-
ponents of the auxin biosynthetic pathway, our view of how IAA is
produced in plants resembles that of a puzzle made of a few
unconnected pieces. In this study, to clarify the possible rela-
tionship between the different postulated routes of IAA bio-
synthesis in Arabidopsis, we took a combination of genetic,
biochemical, and pharmacological approaches. We present
evidence that conclusively shows the coinvolvement of YUCs
and TAA1/TARs in the same route of auxin production. These
findings, together with the lack of auxin-related phenotypes in
multiple pdc mutant combinations, provide an excellent starting
point for the construction of a new working model of the IPyA
branch of auxin biosynthesis.
RESULTS
The CYP79B2/B3 Function Is Not Required for
YUC1-Mediated Auxin Production
To examine the relationship between the different auxin biosyn-
thetic routes genetically defined in Arabidopsis, we first consid-
ered the possible points of action for the YUC1 monooxygenase.
Theoretically, the YUC activity may be required in the CYP79B2/
B3 route, in the TAA1/TAR route, or in a branch of auxin bio-
synthesis that is independent of either of these two gene families.
To discriminate between these three alternatives, we first exam-
ined the possibility that YUCs and CYP79B2/B3 function in the
same pathway. A priori, this seemed a very unlikely scenario,
because the CYP79B2/B3 pathway is known to operate mainly in
Brassica spp. (Bender and Celenza, 2009; Sugawara et al., 2009),
whereas auxin biosynthesis via YUCs has also been demonstrated
in species beyond this genus (Tobeña-Santamaria et al., 2002;
Gallavotti et al., 2008). Nevertheless, the relationship between
these two pathways may not be as clear as one may have
Rethinking Auxin Biosynthesis via IPyA 3963
anticipated. The detailed quantification of several IAA interme-
diates clearly showed that IAN accumulates in Arabidopsis
plants overexpressing YUC1 (Nonhebel et al., 2011). This result
is quite surprising, because Sugawara et al. (2009) have found
that most (if not all) of the IAN found in Arabidopsis is produced
via CYP79B2/B3. Together, these findings suggest that the
levels of IAN, an intermediate in the CYP79B2/B3 pathway, are
directly or indirectly affected by YUC activity. Thus, to experi-
mentally address the possibility of some type of interdepen-
dence between YUCs and CYP79B2/B3, we took two reciprocal
genetic approaches. First, we crossed a well-characterized
YUC1-overexpressing (ox) line (35S:YUC1) (Zhao et al., 2001)
into the cyp79b2/b3 double T-DNA knockout (KO) mutant.
Unfortunately, this cross proved to be not informative, because
we observed silencing of the YUC1ox transgene, possibly
caused by the presence of multiple copies of the 35S promoter
in the cyp79b2/b3 T-DNA mutant plants. To bypass this problem,
we used a transgenic construct, TAA1:YUC1, designed to confer
the expression pattern of TAA1 to the YUC1 cDNA (see
Methods). As shown in Figure 2, the expression of this construct
in Columbia ecotype (Col-0) plants resulted in the characteristic
high-auxin phenotypes (typically observed in the YUC1 over-
expression lines) at all developmental stages examined, includ-
ing dark-grown seedlings (Figure 2A), light-grown seedlings
(Figure 2B), and adult plants (Figure 2C). Of note, the same
high-auxin phenotypes were observed when this TAA1:YUC1
construct was introgressed into the double KO cyp79b2/b3
mutant plants (Figure 2). These results suggest that CYP79B2/B3
activity is not required for auxin production via YUC1.
To test further the relationship between the CYP79B2/B3 and
YUC gene families, we examined the effects of activating both
the CYP79B2/B3 and the YUC1 pathways simultaneously. To
activate the CYP79B2/B3 route, we used a sur2 loss-of-function
mutation. As shown in Figure 2, the high-auxin phenotypes of
sur2 are completely suppressed by the cyp79b2/b3 mutations,
indicating that the sur2 phenotype is the result of the CYP79B2/
B3 pathway hyperactivation. If CYP79B2/B3 and YUCs function
in parallel or independent pathways, the simultaneous activation
of both routes is expected to result in additive phenotypes. The
morphological defects of the sur2 YUCox line are in fact fully
consistent with the additive model (see Supplemental Figure 2
online), providing additional support for CYP79B2/B3 and YUCs
working in parallel/independent pathways.
Taken together, the aforementioned results clearly indicate
that YUC1 does not function in the CYP79B2/B3 branch of auxin
biosynthesis. Therefore, the previously reported accumulation of
IAN in the YUC1ox plants (Nonhebel et al., 2011) is likely a
secondary effect of altering the normal IAA homeostasis rather
than a direct consequence of increasing the metabolic flow
through the YUC pathway.
Simultaneous Knockdown of Several TAA1/TAR and YUC
Family Members Has Less-Than-Additive Effects
Based on the similarities of the loss-of-function phenotypes of
TAA1/TARs and of multiple YUCs, it has been suggested that the
YUC and TAA1/TAR gene families may function in the same
branch of auxin biosynthesis rather than in the two independent
3964 The Plant Cell
Figure 2. The cyp79b2/b3 Double Mutant Can Suppress the High-Auxin
Phenotypes of sur2, but Not That of TAA1:YUC1.
Col, Col TAA1:YUC1, sur2, cyp79b2/b3, TAA1:YUC1 cyp79b2/b3, and
sur2 cyp79b2/b3 were grown in plates for 3 d in the dark (A), 3 d in the dark
followed by 4 d in constant light (B), or for 3 weeks in soil under a 16-h light/
8-h dark cycle (C). Representative plants of each genotype are shown.
pathways proposed by the current models (Strader and Bartel,
2008). A much stronger argument in favor of such possibility was
recently obtained upon the analysis of double loss-of-function
mutants in the maize orthologs of YUC1 (spi1) and TAA1 (vt2),
which were found to display a less-than-additive phenotype
(Phillips et al., 2011). Thus, for our first approach to test the
interaction between these two gene families in Arabidopsis, we
used a similar experimental strategy. One important difference,
however, is that in contrast with the situation in maize, where spi1
and vt2 single mutants show dramatic developmental defects,
the high degree of functional redundancy among TAA1/TARs as
well as YUCs in Arabidopsis made the generation of higher-order
mutant combinations necessary. Starting with a wei8 tar2 double
mutant, where the IPyA branch of IAA biosynthesis is critically
impaired (Stepanova et al., 2008), higher-order mutants with up
to four additional yuc KOs were generated. The expectation was
that if TAA1/TARs and YUCs work in the same pathway, knock-
ing out one or more YUC genes in the wei8-1 tar2-1 genetic
background would have little, if any, effect. On the other hand, if
TAA1/TARs and YUCs work in parallel pathways, even a minor
reduction in the YUC pathway activity would produce strong
effects in a plant with a compromised TAA1/TAR pathway. As
shown in Figure 3, loss of up to four YUC genes (YUC1, YUC2,
YUC4, and YUC6) in the wei8-1 tar2-1 double mutant back-
ground had no additional phenotypic effects beyond those of the
corresponding parental lines. It is important to point out that,
although the contribution of these four YUC genes to the overall
auxin production may not be very large in 3-d-old seedlings
(Figure 3A), in older plants the role of YUC1, YUC2, YUC4, and
YUC6 in synthesizing auxin is significant, as demonstrated both
by the phenotypic analysis (Figure 3B) and the quantification of
IAA and IAA metabolites (see below) in the yuc1/2/4/6 quadruple
mutant. Thus, in agreement with the similar findings in maize
(Phillips et al., 2011), Arabidopsis loss-of-function studies sup-
port (but do not necessarily prove) the idea that TAA1/TARs and
YUCs work in the same biosynthetic pathway.
Pharmacological Inhibition of TAA1/TAR Activity
Suppresses the High-Auxin Phenotypes of
YUC1 Overexpression
To test further the hypothesis of a linear TAA1-YUC1 pathway
model, we decided to take advantage of kynurenine (Kyn), a
potent inhibitor of in vivo TAA1/TAR activity (He et al., 2011). The
rationale for these experiments was that if TAA1/TARs and YUCs
work together to produce IAA, blocking the in vivo activity of
TAA1/TARs with Kyn would suppress the high-auxin phenotypes
of YUC1ox plants. To achieve high levels of YUC1 activity in
plants, we used both the YUC1ox (35S:YUC1) and the TAA1:
YUC1 constructs. In contrast with what we observed in the
cyp79b2/b3 T-DNA KO lines, no silencing of the YUC1ox con-
struct was observed in the Col-0 wild-type background. As
shown in Figure 4, both YUC1 constructs result in plants with the
characteristic high-auxin phenotypes both in dark-grown (Figure
4A) and light-grown (Figure 4B) plants. It is important to point out
that the plants expressing the YUC1 cDNA under the control of
the TAA1 regulatory sequences displayed a much stronger root
phenotype, characterized by an extremely short root length both

Figure 3. Simultaneous Loss of Function of Multiple YUC and TAA1/TAR
Family Members Results in Less-Than-Additive Phenotypic Defects in
Quadruple, Quintuple, and Sextuple Knockout Mutants.
Col, yuc1/2/4/6, wei8-1 tar2-1, wei8-1 tar2-1 yuc1/6, wei8-1 tar2-1 yuc1/
2/6, wei8-1 tar2-1 yuc1/4/6, and wei8-1 tar2-1 yuc1/2/4/6 plants were
grown for 3 d in the dark on AT plates supplemented with 10 mM ACC (A)
or for 3 d in the dark on ACC-supplemented plates followed by 15 d in
constant light on unsupplemented AT plates (B). The genotype of every
plant shown was determined by PCR as described in Methods.
in light- and dark-grown seedlings (Figure 4). On the other hand,
the YUC1ox dark-grown seedlings displayed a hookless pheno-
type that was not observed in the TAA1:YUC1 plants. Remark-
ably, Kyn treatment blocked all of the high-auxin phenotypes of
both types of YUC1ox lines, namely the hookless phenotype of
the dark-grown YUC1ox seedlings, the short root of both
YUC1ox and TAA1:YUC1 in dark- and light-grown seedlings,
as well as the dark-green epinastic cotyledons in the light-grown
plants (Figure 4). Different concentrations of Kyn were required to
Rethinking Auxin Biosynthesis via IPyA 3965
suppress these different phenotypes. For example, the short root
was completely suppressed at Kyn concentrations as low as
1 mM, whereas much higher concentrations were required to
suppress the epinastic cotyledons. Interestingly, at high Kyn
concentrations, both the wild-type and TAA1:YUC1 dark-grown
seedlings become hookless, whereas the hook curvature was
partially restored in the YUC1ox plants. As a control for these
experiments, we used sur2, an auxin-overproducing mutant that
(as was shown in the previous section) works in a parallel and
independent route of auxin production. As shown in Figure 4, Kyn
had little effect on the high-auxin defects of this mutant at the
concentrations tested. Taken together, the results presented
Figure 4. Kyn, a Chemical Inhibitor of the TAA1/TAR Trp-Aminotrans-
ferase Activity, Blocks the High-Auxin Phenotypes of YUC1ox and TAA1:
YUC1 Lines.
Col, Col YUC1ox, Col TAA1:YUC1, and sur2 plants were grown in plates
for 3 d in the dark (A) or for 3 d in the dark followed by 4 d in constant light
(B) in the presence of 0, 1, 5, 10, 25, 50, or 100 mM Kyn. Representative
plants are shown.
3966 The Plant Cell
indicate that the high-auxin phenotypes of both YUC1ox and
TAA1:YUC1 require the function of TAA1/TARs and, therefore,
support the linear action model for TAA1/TARs and YUCs.
Gene Loss- and Gain-of-Function Studies Suggest That
YUC1 and TAA1/TARs Function in the Same
Biosynthetic Pathway
To test further the linear model hypothesis for the action of TAA1/
TARs and YUCs, we took a genetic approach. We reasoned that
if the proposed linear model was correct, then the TAA1/TAR and
YUC activities should coexist in the same cells and at the same
time. Although previous studies based on gene expression
suggest a good deal of overlap between these two gene families
(Cheng et al., 2006; Cheng et al., 2007; Stepanova et al., 2008;
Tao et al., 2008), the presence of 11 YUCs and three TAA1/TARs
makes the direct approach to examine the coexpression of these
genes’ products challenging and possibly inconclusive. To by-
pass this potential problem, we decided to generate transgenic
plants in which a YUC1 cDNA is driven by the regulatory
sequences of TAA1 (i.e., the TAA1:YUC1 construct), thereby
ensuring that the cells expressing TAA1 also express YUC1. The
expectation is that this construct would result in high levels of
auxin only if the cells that normally express TAA1 also had all of
the enzymes required to produce IAA via the YUC1 pathway. By
contrast, no additional auxin would be produced if, in the cells
expressing TAA1 (and, therefore, the TAA1:YUC1 construct),
one or more of the components of the YUC pathway were
missing. As shown in Figure 5, Col-0 plants that express YUC1
cDNA under the control of the regulatory sequences of TAA1
show high-auxin phenotypes very similar to those of YUC1ox
plants at all developmental stages examined (Figures 2, 4, and 5)
(Zhao et al., 2001). The shorter roots and normal hooks observed
in the TAA1:YUC1 dark-grown seedlings (as compared with
those of the YUC1ox transgenics) can be explained by the
specific and highly localized expression pattern of TAA1 in the
root meristem and the apical hook (Stepanova et al., 2008).
Importantly, these results suggest that TAA1/TARs and YUCs
coexist in the same cells and at the same time, which is a
prerequisite for the linear model being tested herein.
Next, we examined whether or not the high-auxin phenotypes
observed in the TAA1:YUC1 plants require functional TAA1 and
TARs. To this end, we introgressed the TAA1:YUC1 transgene
into the following mutant backgrounds: wei8-2, wei8-2 tar1-1,
wei8-2 tar2-1, wei8-2 tar1-1 tar2-1, and yuc1/2/4/6. Interestingly,
the single wei8-2 mutant could fully suppress the short root
phenotype of TAA1:YUC1 (Figures 5A and 5B), whereas the
epinastic cotyledons of light-grown seedlings and the general
morphology of the TAA1:YUC1 adult plants were suppressed
only in the wei8-2 tar2-1 double and wei8-2 tar1-1 tar2-1 triple
mutant backgrounds (Figure 5). Although these results strongly
support the idea that functional TAA1 and TARs are required for
the auxin biosynthetic activity of TAA1:YUC1, it is formally
possible that TAA1:YUC1 can confer a high auxin biosynthetic
activity in the wei8-2 tar2-1 mutants, but this additional auxin is
not sufficient to reach a certain threshold required for rescuing
the low-auxin phenotypes of wei8-2 tar2-1. To rule out this
possibility, we used two different approaches. First, we intro-
gressed the TAA1:YUC1 transgene into the quadruple mutant
yuc1/2/4/6, a yuc mutant combination with auxin-deficient phe-
notypes very similar to those of the wei8 tar2 doubles (Figure 5C;
see Supplemental Figure 3 online) and, as shown below, with
similar levels of free IAA. As an additional control, we also
generated double wei8-2 tar2-1 mutants harboring a TAA1:iaaM
transgene, in which the expression of the bacterial auxin bio-
synthetic gene iaaM is under the control of the same TAA1
regulatory sequences that drive the expression of YUC1 in the
TAA1:YUC1 construct. Importantly, the TAA1:YUC1 transgene
was able to revert the auxin-deficient phenotypes of the qua-
druple yuc mutant, resulting in plants phenotypically similar
to Col TAA1:YUC1 (Figure 5; see Supplemental Figure 3 online).
Furthermore, the TAA1:iaaM construct effectively comple-
mented the wei8-2 tar2-1 auxin defects (see Supplemental
Figure 4 online). These results provide strong genetic support to
the idea that auxin production via YUC1 requires the activity of
the TAA1/TAR gene family or, in other words, that TAA1/TARs
and YUCs function in the same genetic pathway. In addition, the
different degrees by which the wei8-2 single mutant suppresses
TAA1:YUC1 in the various tissues examined could be indicative
of the different levels of functional overlap between TAA1 and
other family members in these tissues. Thus, for example, if most
of the cells that express TAA1 in roots do not express any of the
other family members, it would be expected that the expression
of YUC1 in these TAA1-specific tissues will render no additional
IAA in the wei8 mutant background. On the other hand, if the cells
that express TAA1 in the cotyledons also express TARs, the
expression of YUC1 in these cells would still produce significant
amounts of IAA even in the wei8-2 mutant.
Quantification of Free IAA and Several IAA Metabolites
Indicates That the Function of TAA1 and TARs Is Required for
IAA Production via YUCs
Taken together, the pharmacological and genetic data pre-
sented above strongly support the idea that TAA1/TARs and
YUCs function in a common branch of the auxin biosynthetic
pathway. These types of experiments, although highly infor-
mative, have an important caveat. They rely on the use of
morphological changes as an indirect readout of the IAA bio-
synthetic activity (the primary effect of the imposed genetic or
pharmacological conditions). Furthermore, the approaches de-
scribed thus far cannot clarify the relative position of TAA1/TARs
and YUCs in the biosynthetic pathway. Therefore, we examined
the levels of free IAA and IAA metabolites in several TAA1/TAR
and YUC mutants with or without the TAA1:YUC1 transgene.
In agreement with previous studies (Stepanova et al., 2008;
Tao et al., 2008), significantly lower levels of free IAA were de-
tected in the single wei8-2 as well as in the double wei8-2 tar2-1
mutants (Figure 6A). Importantly, the levels of free IAA in the
quadruple yuc1/2/4/6 were also much lower compared with
those in the wild-type plants. In fact, the IAA levels in the yuc
quadruple mutant were nearly as low as in wei8-2 tar2-1 (Figure
6A). It is important to point out that these auxin measurements
were done using 2-week-old plants. At this developmental stage,
both wei8-2 tar2-1 and yuc1/2/4/6 show clear auxin-deficient
phenotypes. We also measured free IAA in 3-d-old dark-grown

Figure 5. Functional TAA1 and TARs Are Required for YUC1-Mediated
Auxin Biosynthesis.
Col, wei8-2, wei8-2 tar1-1, wei8-2 tar2-1, wei8-2 tar1-1 tar2-1, and yuc1/
Rethinking Auxin Biosynthesis via IPyA 3967
seedlings, but in that case, no reduction in the levels of free IAA
was detected in the quadruple yuc1/2/4/6 mutant, consistent
with the lack of obvious auxin-deficient phenotypes at this
developmental stage (Figure 5A). In addition to free IAA, we
also measured the most abundant IAA conjugates and catabo-
lites in Arabidopsis, namely 2-oxindole-3-acetic acid (oxIAA),
indole-3-acetyl Asp (IAAsp) and indole-3-acetyl Glu (IAGlu)
(Östin et al., 1998; Kowalczyk and Sandberg, 2001). All three
IAA metabolites were reduced in the auxin-depleted genotypes
wei8-2, wei8-2 tar2-1, and yuc1/2/4/6 (Figures 6B to 6D), as
would be expected of plants impaired in the production of IAA.
Once we had established that the levels of free IAA and IAA
conjugates were similarly reduced in both the wei8-2 tar2-1
double mutant and in the yuc1/2/4/6 quadruple mutant, we
examined the effects of expressing the TAA1:YUC1 transgene in
different genetic backgrounds. First, we observed a small in-
crease in the levels of free IAA in Col-0 plants expressing TAA1:
YUC1 (Figure 6A). Although this increase was not significant, we
found a fivefold to 10-fold increase in the levels of IAA metab-
olites in this line (Figures 6B to 6D). These results clearly indicate
that TAA1:YUC1 enhances the IAA biosynthetic activity in the
otherwise wild-type control plants and that IAA catabolism and
conjugation are induced to maintain IAA homeostasis. These
effects of TAA1:YUC1 on the levels of IAA and IAA metabolites
were somewhat reduced in the wei8-2 mutant background and
were nearly completely obliterated in the wei8-2 tar2-1 double
mutants (Figure 6). By contrast, the quadruple yuc mutant had
only negligible effects on the enhanced auxin biosynthetic ac-
tivity conferred by the TAA1:YUC1 transgene (Figure 6). The
simplest interpretation of the overall results is that the auxin
biosynthetic activity of YUCs requires functional TAA1/TARs.
Therefore, our findings firmly position the two gene families in the
same branch of the auxin biosynthesis pathway. These results,
however, do not provide conclusive information on the relative
position of TAA1/TARs with respect to YUCs.
Gene Loss- and Gain-of-Function Analyses Suggest That
None of the Annotated Arabidopsis PDCs Are Involved in
Auxin Biosynthesis
The results described in previous sections show that YUC1, a
flavin monooxygenase, functions downstream of TAA1/TARs in
the conversion of IPyA into IAA. To narrow down the number of
putative YUC1 substrates, we examined whether or not any of the
four predicted Arabidopsis PDCs were involved in auxin produc-
tion. Again, we took a genetic approach and generated loss- and
gain-of-function alleles for the four predicted Arabidopsis PDCs.
We have obtained homozygous T-DNA mutants in all four PDC
genes, PDC1, PDC2, PDC3, and PDC4. T-DNA insertion sites were
identified by sequencing of the PCR-amplified T-DNA/genomic
DNA junction fragments. pdc1-10, pdc2-10, pdc3-10, and pdc4-11
harbor T-DNAs in exons, whereas pdc4-10 has a T-DNA insertion
2/4/6 with or without the TAA1:YUC1 transgene were grown on plates for
3 d in the dark (A), on plates for 3 d in the dark followed by 4 d in constant
light (B), or in soil for 4 weeks under a 16-h light/8-h dark cycle (C).
Representative plants are shown.
3968 The Plant Cell

in the intron (see Supplemental Figures 1A to 1C and Supplemental
Data Set 1 online). Phenotypic analysis of dark- and light-grown
seedlings in plates and of soil-grown adult plants failed to uncover
any auxin-related phenotypes. Similarly, the very sensitive root
elongation assay in response to ethylene (which can pick up even
very mild auxin deficiency) was unsuccessful at detecting any
phenotypic abnormalities. To deal with the possible functional re-
dundancy problem, we generated and characterized several dou-
ble and triple pdc mutant combinations. Again, no auxin-related
phenotypes were observed (see Supplemental Figure 1D online).
To test further the potential role of these genes in auxin biosynthe-
sis, we constructed four different pdc triple mutant combinations in
the sensitized wei8-1 mutant background, but again failed to
observe any auxin-related defects beyond those of wei8 alone (see
Supplemental Figure 1D online). Finally, we also took a comple-
mentary approach and overexpressed all four Arabidopsis PDCs in
the wild-type plants using the strong constitutive 35S promoter.
Again, we were unable to detect any auxin-related phenotypes in
any of these lines (see Supplemental Figure 1D online).
Consistent with our findings, PDC1 and PDC2 were previously
implicated in ethanolic fermentation: A KO allele of PDC1 was
shown to be hypersensitive to anoxia (Kürsteiner et al., 2003),
whereas overexpression of either PDC1 or PDC2 resulted in
increased resistance to low oxygen levels (Ismond et al., 2003).
These data, together with the lack of auxin-related defects in
multiple pdc mutants and the inability of any of the PDCs to use
IPyA as a substrate in vitro (Ye and Cohen, 2009), suggest that
the PDC gene family may not be involved in auxin biosynthesis in
plants. Therefore, a completely revised working model of how
IPyA is converted into IAA is required.
DISCUSSION
To date, none of the four predicted routes of auxin biosynthesis in
plants have been experimentally resolved in full. The current
models of the auxin biosynthetic pathway, although obviously
very useful, should be considered as “working hypotheses”

Figure 6. Quantification of the Levels of IAA and IAA Catabolites/Conjugates.

Levels of free IAA (A), oxIAA (B), IAAsp (C), and IAGlu (D) were quantified in 2-week-old plants of Col, wei8-2, wei8-2 tar2-1, and yuc1/2/4/6 with and
without the TAA1:YUC1 transgene. The bars represent average levels (6 SD) of three independent biological replicates. The tables below the bar graphs
indicate P-values of the genotype comparisons in an ANOVA analysis. *, **, and *** correspond to P-values of 0.05 > p > 0.01, 0.01 > p > 0.001, and p <
0.001, respectively. F.W., fresh weight.

rather than well established and solidly grounded facts. Indeed,
recent studies have raised some legitimate concerns about the
validity of several aspects of these models, primarily regarding
the position of the YUC genes in auxin biosynthesis. This work
was aimed at specifically addressing this uncertainty and pro-
viding a solid foundation for the future elucidation of the
complete IPyA route of IAA biosynthesis. Our genetic, pharma-
cological, and biochemical data provide compelling evidence for
YUCs and TAA1/TARs working in the same Trp-dependent
branch of the auxin biosynthetic pathway. Although previous
reports have suggested this possibility, to our knowledge, no
direct proof had been presented until now. Based on the fact that
Trp is the substrate for TAA1 and that YUC1 is unlikely to be
involved in Trp production, the simplest model would position
YUCs downstream of TAA1/TARs. On the other hand, serious
doubts about the role of PDCs downstream of the IPyA route in
the IAA biosynthetic pathway have been raised herein and in Ye
and Cohen (2009). Thus, in contrast with auxin-synthesizing
bacteria, plants may produce IAA from IPyA using a different
enzymatic mechanism that involves the catalytic activity of
YUCs. In fact, based on examples from the literature (Lockridge
et al., 1972), we speculate that flavin monooxygenases, like
YUCs, may catalyze the conversion of IPyA into IAA.
In addition to shedding new light on the relationship between
the different branches of auxin biosynthesis, our results also
raise several obvious questions that should become the subject
of future investigations. Perhaps the most important of these
issues is how IPyA, a labile molecule (Badenoch-Jones et al.,
1984; Tam and Normanly, 1998), is converted into IAA via one or
more enzymatic reactions that comprise (possibly, among
others) the monooxygenase activity of the YUC enzyme. Al-
though at this point we do not have experimental evidence to
show how this is accomplished in plants, we can make several
observations that may guide us in this future endeavor. For
example, the known chemical instability of IPyA, the high affinity
of Trp-aminotransferases for IPyA, and the cytoplasmic locali-
zation of both TAA1 and YUCs may be indicative of the existence
of a multiprotein functional complex implicated in the catalysis of
two or more sequential enzymatic reactions, including those of
TAA1 and YUCs. Although merely speculative at this point for the
TAA1-YUC case, this idea of “metabolic channeling” is by no
means new, and there are many examples, including the Trp
biosynthetic pathway itself, where the product of one enzyme is
directly fed to the next enzyme in the pathway (Miles, 2001).
Another interesting observation came from the comparison of
the morphological phenotypes and IAA levels across the different
mutants examined. Thus, for example, it was surprising to see
that the strong high-auxin phenotypes of the TAA1:YUC1 lines
did not correspond to a statistically significant increase in the
levels of free IAA. By contrast, a similar degree of reduction in the
levels of free IAA in the 2-week-old double wei8-2 tar2-1 and
quadruple yuc1/2/4/6 mutants resulted in phenotypic alterations
of different severity (although at later developmental stages, the
two genotypes become nearly indistinguishable). In theory, these
results could be indicative, as has been suggested before for the
YUC genes (Tobeña-Santamaria et al., 2002), of these genes
being involved in the production of a bioactive compound other
than IAA. We, however, disfavor this idea for several reasons.
Rethinking Auxin Biosynthesis via IPyA 3969
First, both exogenous auxin treatment (see Supplemental Figure
5 online) (Stepanova et al., 2008) and stimulation of endogenous
auxin production by means of sur2 (Stepanova et al., 2008) or
iaaM (Cheng et al., 2006; Tao et al., 2008) complement most of
the phenotypic defects of the loss-of-function mutants. Further-
more, the dramatic increase in the levels of the major IAA
catabolite oxIAA and of the IAA conjugates IAAsp and IAGlu in
the TAA1:YUC1 lines suggests that the apparent lack of corre-
lation between the IAA levels and the phenotypic alterations
observed is caused by a very effective homeostasis mechanism
that regulates the levels of free IAA.
Finally, our results also show that the expression of YUC1
under either a constitutive promoter or the regulatory sequences
of TAA1 results in very similar but not identical high-auxin
phenotypes. This suggests that nearly every cell that is capable
of producing IAA via YUC also expresses TAA1, and the differ-
ence between the constitutive YUC1ox (35S:YUC1) and the
TAA1:YUC1-overexpressing lines is caused by the few cells or
developmental stages in which either TAR1 or TAR2 are ex-
pressed but TAA1 is not. This hypothesis is also consistent with
the lack of phenotypes in the tar1 or tar2 single or tar1 tar2 double
mutants and with the fact that the phenotypic consequences of
the loss of function of these two genes can be observed only in
the absence of functional TAA1 (Stepanova et al., 2008). An
alternative explanation would be that the YUC1ox and the TAA1:
YUC1-overexpressing lines produce high levels of auxin in very
different places, but the auxin transport and catabolism machin-
eries mask these differences. In either case, however, the few but
clear phenotypic differences observed between the YUC1ox and
the TAA1:YUC1 lines (i.e., the shorter root and normal apical
hook curvature of the TAA1:YUC1 lines compared with the
YUC1ox) indicate that not only the overall levels but also the
patterns of auxin biosynthesis play important roles, at least in
processes such as root elongation and differential cell growth in
the apical hook. These findings provide further evidence for the
key role of local auxin biosynthesis in plant development.
Taken together, the results reported here represent one addi-
tional step toward the construction of a revised model of the IPyA
branch of the auxin biosynthetic pathway. The future research in
this area is likely to focus on determining how IPyA is converted to
IAA in plants and on establishing the metabolic intermediates
involved and the enzymes responsible for each step. These studies
should clarify the biochemical role of YUCs in auxin production.
Having a complete view of the major TAA1/YUC-mediated route of
auxin biosynthesis will greatly advance our understanding of the
role of auxin biosynthesis in plant growth and development and will
open new avenues to manipulating the levels and distribution of
this essential plant hormone for the benefit of mankind.
METHODS
Strains, Constructs, Plant Transformation, and Genotyping
All Arabidopsis thaliana mutant strains and transgenic lines reported herein
are in Col background. wei8-1, wei8-2, tar1-1, tar2-1, tar2-2, and tar2-3
were described in Stepanova et al. (2008). sur2 was reported in Stepanova
et al. (2005). The cyp79b2/b3 double mutant (Zhao et al., 2002) was
provided by Dr. J. Celenza. The yuc1/2/4/6 quadruple mutant (Cheng et al.,
3970 The Plant Cell
2006) and the YUC1ox line (Zhao et al., 2001) were a gift from Dr. Y. Zhao.
pdc1-10, pdc2-10, pdc3-10, pdc4-10, and pdc4-11 (see Supplemental
Figure 1C online for the origin of these lines) were obtained from ABRC
(Alonso et al., 2003). Genotype of the TAA1 locus in various wei8-2–
containing mutant combinations was determined by sequencing.
Extraction of plant genomic DNA for genotyping purposes was per-
formed in cetyltrimethylammonium bromide (CTAB) buffer (1.4 M NaCl,
20 mM EDTA, 100 mM Tris-HCl, pH 8, 3% CTAB [Calbiochem]) either
individually or in a 96-well format. For individual sample extraction, fresh
or frozen tissues (typically one average-size adult leaf or five to 20
seedlings) were ground dry in microfuge tubes for 6 to 15 s in the
presence of ;100 mL of 1-mm glass beads in a Silamat S5 shaker (Ivoclar
Vivadent), followed by 6 to 15 s with 250 mL of CTAB buffer. For high-
throughput extraction, tissues were collected into the individual wells of
96-well plates, frozen at 2808C, ground dry in a Qiagen shaker with 4.5-
mm steel zinc-plated beads, and resuspended in 250 mL of CTAB buffer.
Samples were incubated at 658C for 30 min, extracted once with 1 volume
of chloroform, and precipitated with 1 volume of isopropyl alcohol. DNA-
containing pellets were washed once with 70% ethanol, were air-dried,
and were resuspended in 100 to 300 mL deionized water.
Genotyping of DNA samples extracted from various T-DNA mutants
and/or their derivatives was performed by PCR. The origin of all mutant
alleles used in this study and their respective genotyping primers are
listed in Supplemental Tables 1 and 2 online. The typical 10-mL PCR
reaction contained 1 mL of genomic DNA, 1 mL of 103 PCR buffer, 0.25 mL
of 2 mM deoxynucleotides triphosphate, 0.25 mL of homemade Taq
polymerase, 0.25 mL of 20 mM forward primer, 0.25 mL of 20 mM reverse
primer, and 7 mL of H2O. A typical PCR program was (30 s at 948C, 30 s at
568C, 1 to 3 min at 728C [1 min per kb]) times 40 cycles. PCR products
were separated on 13 Tris-acetate-EDTA 1 to 2% agarose gels, and the
gel images were photographed and scored.
The YUC1 cDNA and iaaM genomic clone were provided by Drs. Y.
Zhao and G. Jurgens, respectively. TAA1:YUC1 and TAA1:iaaM con-
structs were generated by bacterial recombineering (Zhou et al., 2011) in
the backbone of a transformation-ready BAC clone JAtY73L21. The TAA1
gene (nucleotides 221 to 2256) was replaced by the open reading frame of
YUC1 or iaaM using GalK selection/counterselection (Warming et al.,
2005), and the recombineered BAC clones were transformed into the
wei8-2 (for TAA1:YUC1) or wei8-2 tar2-1/+ (for TAA1:iaaM) background
by a modified Agrobacterium-mediated “flower-dip” method (Zhou et al.,
2011). Selection of the T1 generation was performed on AT plates (13
Murashige and Skoog salts, 1% Suc, pH 6.0, with 1 M KOH, 0.7% Bacto
Agar) supplemented with 20 mg/mL phosphinothricin (Basta). For TAA1:
YUC1, one representative line (of more than 12 that displayed charac-
teristic high-auxin phenotypes, including epinastic cotyledons and
leaves) was crossed to Col, cyp79b2/b3, wei8-2 tar2-1, wei8-2 tar1-1
tar2-2, and yuc1/2/4/6, and the progeny of the crosses was genotyped in
F2, F3, and/or F4. sur2 cyp79b2/b3 triple mutant was generated by
crossing sur2 with the cyp79b2/b3 double mutant. Various pdc mutant
combinations (with and without wei8-1) were generated by sequential
crossing of single and then higher-order mutants and by genotyping of
the progeny in F2 and F3 generations of each cross. PDC overexpression
constructs were generated by Gateway technology (Invitrogen) in the
pGWB2 vector (Nakagawa et al., 2007). A PDC3 cDNA clone in the
pENTR223 vector was obtained from ABRC (G22953) and was subcloned
into pGWB2 by LR Clonase. Genomic DNA fragments of PDC1, PDC2,
and PDC4 (from the start codons to the ends of the open reading frames
minus stop codons) were PCR-amplified by Pfx50 polymerase using BAC
DNA minipreps (clones JAtY57K14, JAtY66I17, and JAtY54M22, respec-
tively) as templates. PCR fragments were subcloned into pENTR/D-Topo,
sequence-verified, and then transferred to pGWB2 by LR recombination.
All four 35S:PDC constructs were introduced into Col plants by the
Agrobacterium-mediated transformation using the flower-dip method
(Clough and Bent, 1998). Transformants (20 or more per gene) were
selected on AT plates supplemented with 100 mg/mL kanamycin and
analyzed in T1 and T2.
Multiple wei8-1 tar2-1 yuc combinations were obtained by crossing
wei8-1 tar2-1/+ tar1-1 with the yuc1/+ yuc2 yuc4 yuc6/+ mutant (because
homozygous versions of these mutant combinations are sterile/lethal).
Ten out of 60 F1 plants tested had inherited the mutant versions of tar2-1,
yuc1, and yuc6. A total of 248 individual F2 plants from the progeny of one
of these 10 plants (number 79) were genotyped for all seven segregating
loci. A single F2 plant, number 79-114, with the genotype wei8-1 t2-1/+
tar1-1/+ yuc1 yuc2/+ yuc4/+ yuc6/+ was selected for further analysis,
because it had undergone recombination between the closely linked tar2
and yuc1 loci. Then, 47 individual F3 progeny of number 79-114 were
genotyped for tar2-1, yuc2, yuc4, and yuc6 mutations. Two of these F3
progeny, numbers 79-114-44 and 79-114-62, were shown to be geno-
typically wei8-1 t2-1/+ yuc1 yuc2/+ yuc4/+ yuc6. F4 progeny of these two
lines were planted on AT plates supplemented with 10 mM 1-amino-
cyclopropane-1-carboxylate (ACC), phenotypically preselected in the
triple response assay for the characteristic wei8-1 tar2-1 double mutant-
like defects (lack of apical hook, long agravitropic or missing root, and/or
missing hypocotyls, and/or monopteros-like defects) in 3-d-old etiolated
seedlings, and photographed individually. A total of 164 preselected F4s
were grown for an additional 15 d in constant light, were photographed,
and were then harvested for DNA extraction and genotyping. Three out of
164 F4s were found to be homozygous sextuple wei8-1 t2-1 yuc1 yuc2
yuc4 yuc6 mutants, which corresponds to ;1/55 instead of the expected
1/16 sextuple mutants among the tar2-1 preselected plants.
Physiological Assays
Seeds were surface-sterilized with 50% bleach (spiked with approxi-
mately five drops of Triton-X100 per liter to avoid seed clumping) for ;10
min, washed three or more times with sterile diH2O, resuspended in
melted precooled 0.7% low-melting-point agarose in water, and planted
on the surface of AT plates (see above) supplemented or not with ACC or
Kyn. Plates with seeds were kept at 48C for 2 to 4 d, light-treated for 1 to
2 h to equalize germination, wrapped in aluminum foil, and placed
horizontally in a 228C dark incubator for 3 d. Phenotypes of etiolated
seedlings were scored, and their photographs were taken at 72 h of age.
For the light assay, plates with dark-grown 3-d-old seedlings were
transferred to constant light and were incubated for an additional 4 to
15 d, at which point pictures were taken, phenotypes were scored, and/or
tissues were collected for IAA level quantification. For soil-grown plants,
seeds were germinated on plates and then transplanted to soil (50:50 mix
of Fafard 4P [Fafard] and Sunshine Professional Growing Mix [Sun Gro
Horticulture]) and grown under a 16-h light/8-h dark cycle.
For the triple response assay and Kyn treatment, seeds were germi-
nated for 3 d in the dark on horizontal AT plates supplemented with 10 mM
ACC or with the indicated concentrations of Kyn. 50 mM ACC (Calbio-
chem) stocks were made in diH2O, whereas stocks of 50 mM Kyn (Sigma-
Aldrich) were prepared in 0.1 N HCl in water.
Quantification of Free IAA and IAA Metabolites
Col, Col TAA1:YUC1 (hemizygous), wei8-2, wei8-2 TAA1:YUC1 (homo-
zygous), wei8-2 tar2-1/+, wei8-2 tar2-1/+ TAA1:YUC1 (homozygous), and
yuc1/2/4/6 TAA1:YUC1 (hemizygous) seedlings were grown on AT plates
in the dark for 3 d (the indicated genotypes correspond to those of the
parental plants). Seedlings from homozygous lines were picked ran-
domly, and those from the segregating lines were selected based on their
phenotype (see below), transferred to fresh AT plates, and grown for
another 11 d in constant light. At 2 weeks of age, all tissues (three

independent biological replicates per genotype, with each replicate
originating from a different seed batch) were collected, weighed, and
frozen in liquid nitrogen. Phenotype selection of the segregating lines was
performed in 3-d-old etiolated seedlings in the following way. Col TAA1:
YUC1 lines were selected for two copies of the TAA1:YUC1 transgene
based on the very short root phenotype (Col plants hemizygous for this
transgene show intermediate root length). wei8-2 tar2-1 and wei8-2
tar2-1 TAA1:YUC1 plants were selected based on their highly agravi-
tropic root and open apical hooks, which are characteristic of the wei8
tar2 double mutants. yuc1/2/4/6 and yuc1/2/4/6 TAA1:YUC1 (homozy-
gous for the transgene) were selected from the yuc1/2/4/6 TAA1:YUC1
hemizygous populations based on their long versus very short root
phenotype, respectively (hemizygous seedlings show an intermediate
root phenotype and were discarded). Correlations between the root
length and the copy number of the TAA1:YUC1 transgene were ob-
served in multiple transgenic lines (in Col and yuc1/2/4/6 background)
across several generations. Genotype of the preselected wei8-2 tar2-1
homozygotes was confirmed by the characteristic phenotypes of
2-week-old plants and/or by PCR.
For quantification of IAA and its metabolites, whole plants were
collected and purified in triplicates (20 mg fresh weight/sample). Extrac-
tion and purification were done according to Edlund et al. (1995) with
minor modifications. The frozen samples were homogenized, extracted,
and purified with the addition of the following [13C6]-labeled internal
standards: [indole-13C6]-IAA (10 pmol), [indole-13C6]-oxIAA (10 pmol),
[indole-13C6]-IAAsp (5 pmol), and [indole-13C6]-IAGlu (5 pmol) (Kowalczyk
and Sandberg, 2001). All samples were analyzed by liquid chromatogra-
phy-multiple reaction monitoring-mass spectrometry. The eluates were
evaporated to dryness and dissolved in 20 mL of mobile phase before
mass analysis using a 1290 Infinity LC system and a 6460 Triple Quad LC/
MS system (Agilent Technologies). A total of 10 mL of each sample was
injected onto a reversed-phase column (Kinetex C18 100A, 50 3 2.1 mm,
1.7 mm; Phenomenex) and eluted with 10-min gradient (0 to 10 min, 10/90
to 50/50% A/B; flow rate, 0.2 mL min21; column temperature, 308C,
where A was 0.1% acetic acid in methanol and B was 0.1% acetic acid in
water) (see Supplemental Figure 6 online). At the end of the gradient, the
column was washed and equilibrated to initial conditions for 6 min.
Quantification was obtained by monitoring the precursor ([M+H]+) and the
appropriate product ions. The multiple reaction monitoring transitions
182.2>136.1, 198.2>152.1, 297.2>136.1, and 311.1>136.1 were used for
labeled IAA, oxIAA, IAAsp, and IAGlu, respectively, and 176.2>130.1,
192.2>146.1, 291.2>130.1, and 305.1>130.1 were used for unlabeled
IAA, oxIAA, IAAsp, and IAGlu, respectively. Tandem mass spectrometry
conditions were as follows: drying/sheath gas temperature, 250/4008C;
drying/sheath gas flow, 10/12 L min21; nebulizer pressure, 60 psig;
capillary voltage, 2250 V; fragmentor, 65 to 85 V; collision energy, 9 to 22
V. The limits of detection (signal-to-noise ratio, 1:3) were close to 1 pg for
IAA and its metabolites. The linear range was established to be 0.01 to
100 ng per injection with a correlation coefficient of 0.9987 to 0.9998.
Chromatograms were analyzed using the MassHunter software (Agilent
Technologies), and the compounds were quantified by standard isotope
dilution analysis (Rittenberg and Foster, 1940).
Accession Numbers
Sequence data from this article can be found in the Arabidopsis Genome
Initiative or GenBank/EMBL databases under the following accession
numbers: At4g33070 (PDC1), At5g54960 (PDC2), At5g01330 (PDC3),
At5g01320 (PDC4), At3g48560 (PDL1), At5g17380 (PDL2), At5g20960
(AAO1), At3g43600 (AAO2), At2g27150 (AAO3), At1g04580 (AAO4),
At1g70560 (TAA1), At1g23320 (TAR1), At4g24670 (TAR2), At4g39950
(CYP79B2), At2g22330 (CYP79B3), At4g31500 (SUR2), At4g32540
(YUC1), At4g13260 (YUC2), At5g11320 (YUC4), and At5g25620 (YUC6).
Rethinking Auxin Biosynthesis via IPyA 3971
Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure 1. Loss- and Gain-of-Function Analysis of the
Putative PDC Gene Family in Arabidopsis.
Supplemental Figure 2. Simultaneous Activation of the CYP79B2/B3
and YUC Pathways in the sur2 YUC1ox Lines Leads to Additive
Phenotypic Defects.
Supplemental Figure 3. TAA1:YUC1 Can Suppress the Flower
Defects of yuc1/2/4/6, but Not of wei8-2 tar2-1.
Supplemental Figure 4. Functional TAA1 and TARs Are Not Required
for iaaM-Mediated Auxin Biosynthesis.
Supplemental Figure 5. Exogenous IAA Can Restore Some but Not
All of the wei8 tar2 Mutant Phenotypes.
Supplemental Figure 6. Representative Chromatograms of Free IAA
and IAA Catabolites/Conjugates.
Supplemental Table 1. Primers Used to Genotype the Different
Mutants Used in This Study
Supplemental Table 2. Sequences of the Primers Used in This Study
Supplemental Data Set 1. Text File of the Alignment Used for the
Phylogenetic Analysis in Supplemental Figure 1A.
ACKNOWLEDGMENTS
We thank members of the Alonso-Stepanova lab for helpful comments
and stimulating discussions. This work was supported by National
Science Foundation Grants DBI0820755 and MCB0923727 to J.M.A.
and A.N.S., and grants to K.L. and O.N. from the Swedish Governmental
Agency for Innovation Systems and the Swedish Research Council
(Vetenskapsrådet). O.N. was also supported by the Ministry of Educa-
tion, Youth and Sports of the Czech Republic (MSM6198959216).
AUTHOR CONTRIBUTIONS
A.N.S. and J.M.A. designed experiments and performed research, ana-
lyzed data, and wrote the article. J.Y. and L.M.R. performed research.
O.N. and K.L. performed research, analyzed data, and contributed new
analytic tools. W.H. and H.G. contributed information about a new
chemical tool.
Received June 8, 2011; revised October 19, 2011; accepted October 26,
2011; published November 22, 2011.
REFERENCES
Alonso, J.M., et al. (2003). Genome-wide insertional mutagenesis of
Arabidopsis thaliana. Science 301: 653–657.
Badenoch-Jones, J., Summons, R.E., Rolfe, B.G., and Lethan, D.S.
(1984). Phytohormones, Rhizobium mutants, and nodulation in le-
gumes. IV. Auxin metabolism in pea root nodules. J. Plant Growth
Regul. 3: 23–39.
Bak, S., Tax, F.E., Feldmann, K.A., Galbraith, D.W., and Feyereisen,
R. (2001). CYP83B1, a cytochrome P450 at the metabolic branch
point in auxin and indole glucosinolate biosynthesis in Arabidopsis.
Plant Cell 13: 101–111.
Bartel, B. (1997). Auxin biosynthesis. Annu. Rev. Plant Physiol. Plant
Mol. Biol. 48: 51–66.
3972 The Plant Cell
Bender, J., and Celenza, J.L. (2009). Indolic glucosinolates at the
crossroads of tryptophan metabolism. Phytochem. Rev. 8: 25–37.
Cheng, Y., Dai, X., and Zhao, Y. (2006). Auxin biosynthesis by the
YUCCA flavin monooxygenases controls the formation of floral organs
and vascular tissues in Arabidopsis. Genes Dev. 20: 1790–1799.
Cheng, Y., Dai, X., and Zhao, Y. (2007). Auxin synthesized by the
YUCCA flavin monooxygenases is essential for embryogenesis and
leaf formation in Arabidopsis. Plant Cell 19: 2430–2439.
Clough, S.J., and Bent, A.F. (1998). Floral dip: a simplified method for
Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant
J. 16: 735–743.
De Smet, I., et al. (2011). Unraveling the evolution of auxin signaling.
Plant Physiol. 155: 209–221.
Edlund, A., Eklöf, S., Sundberg, B., Moritz, T., and Sandberg, G.
(1995). A microscale technique for gas chromatography-mass spec-
trometry measurements of picogram amounts of indole-3-acetic acid
in plant tissues. Plant Physiol. 108: 1043–1047.
Gallavotti, A., Barazesh, S., Malcomber, S., Hall, D., Jackson, D.,
Schmidt, R.J., and McSteen, P. (2008). sparse inflorescence1 en-
codes a monocot-specific YUCCA-like gene required for vegetative
and reproductive development in maize. Proc. Natl. Acad. Sci. USA
105: 15196–15201.
Glawischnig, E., Hansen, B.G., Olsen, C.E., and Halkier, B.A. (2004).
Camalexin is synthesized from indole-3-acetaldoxime, a key branch-
ing point between primary and secondary metabolism in Arabidopsis.
Proc. Natl. Acad. Sci. USA 101: 8245–8250.
Grubb, C.D., Zipp, B.J., Ludwig-Müller, J., Masuno, M.N., Molinski,
T.F., and Abel, S. (2004). Arabidopsis glucosyltransferase UGT74B1
functions in glucosinolate biosynthesis and auxin homeostasis. Plant
J. 40: 893–908.
He, W., et al. (2011). A small-molecule screen identifies L-kynurenine as
a competitive inhibitor of TAA1/TAR activity in ethylene-directed auxin
biosynthesis and root growth in Arabidopsis. Plant Cell 23: 3944–
3960.
Ismond, K.P., Dolferus, R., de Pauw, M., Dennis, E.S., and Good,
A.G. (2003). Enhanced low oxygen survival in Arabidopsis through
increased metabolic flux in the fermentative pathway. Plant Physiol.
132: 1292–1302.
Kowalczyk, M., and Sandberg, G. (2001). Quantitative analysis of
indole-3-acetic acid metabolites in Arabidopsis. Plant Physiol. 127:
1845–1853.
Kürsteiner, O., Dupuis, I., and Kuhlemeier, C. (2003). The pyruvate
decarboxylase1 gene of Arabidopsis is required during anoxia but not
other environmental stresses. Plant Physiol. 132: 968–978.
Lockridge, O., Massey, V., and Sullivan, P.A. (1972). Mechanism of
action of the flavoenzyme lactate oxidase. J. Biol. Chem. 247: 8097–
8106.
Mikkelsen, M.D., Naur, P., and Halkier, B.A. (2004). Arabidopsis
mutants in the C-S lyase of glucosinolate biosynthesis establish a
critical role for indole-3-acetaldoxime in auxin homeostasis. Plant J.
37: 770–777.
Miles, E.W. (2001). Tryptophan synthase: A multienzyme complex with
an intramolecular tunnel. Chem. Rec. 1: 140–151.
Nakagawa, T., Kurose, T., Hino, T., Tanaka, K., Kawamukai, M.,
Niwa, Y., Toyooka, K., Matsuoka, K., Jinbo, T., and Kimura, T.
(2007). Development of series of gateway binary vectors, pGWBs, for
realizing efficient construction of fusion genes for plant transforma-
tion. J. Biosci. Bioeng. 104: 34–41.
Nemoto, K., Hara, M., Suzuki, M., Seki, H., Muranaka, T., and Mano,
Y. (2009). The NtAMI1 gene functions in cell division of tobacco BY-2
cells in the presence of indole-3-acetamide. FEBS Lett. 583: 487–492.
Nonhebel, H., Yuan, Y., Al-Amier, H., Pieck, M., Akor, E., Ahamed,
A., Cohen, J.D., Celenza, J.L., and Normanly, J. (2011). Redirection
of tryptophan metabolism in tobacco by ectopic expression of an
Arabidopsis indolic glucosinolate biosynthetic gene. Phytochemistry
72: 37–48.
Normanly J. (2010). Approaching cellular and molecular resolution of
auxin biosynthesis and metabolism. Cold Spring Harb. Perspect. Biol.
(online), doi/10.1101/cshperspect.a001594.
Normanly, J., Cohen, J.D., and Fink, G.R. (1993). Arabidopsis thaliana
auxotrophs reveal a tryptophan-independent biosynthetic pathway for
indole-3-acetic acid. Proc. Natl. Acad. Sci. USA 90: 10355–10359.
Östin, A., Ilić, N., and Cohen, J.D. (1999). An in vitro system from maize
seedlings for tryptophan-independent indole-3-acetic acid biosynthe-
sis. Plant Physiol. 119: 173–178.
Östin, A., Kowalyczk, M., Bhalerao, R.P., and Sandberg, G. (1998).
Metabolism of indole-3-acetic acid in Arabidopsis. Plant Physiol. 118:
285–296.
Ouyang, J., Shao, X., and Li, J. (2000). Indole-3-glycerol phosphate, a
branchpoint of indole-3-acetic acid biosynthesis from the tryptophan
biosynthetic pathway in Arabidopsis thaliana. Plant J. 24: 327–333.
Phillips, K.A., Skirpan, A.L., Liu, X., Christensen, A., Slewinski, T.L.,
Hudson, C., Barazesh, S., Cohen, J.D., Malcomber, S., and
McSteen, P. (2011). vanishing tassel2 encodes a grass-specific
tryptophan aminotransferase required for vegetative and reproductive
development in maize. Plant Cell 23: 550–566.
Pollmann, S., Neu, D., and Weiler, E.W. (2003). Molecular cloning and
characterization of an amidase from Arabidopsis thaliana capable of
converting indole-3-acetamide into the plant growth hormone, indole-
3-acetic acid. Phytochemistry 62: 293–300.
Rittenberg, D., and Foster, G.L. (1940). A new procedure for quanti-
tative analysis by isotope dilution, with application to the determina-
tion of amino acids and fatty acids. J. Biol. Chem. 133: 727–744.
Seo, M., Akaba, S., Oritani, T., Delarue, M., Bellini, C., Caboche, M.,
and Koshiba, T. (1998). Higher activity of an aldehyde oxidase in the
auxin-overproducing superroot1 mutant of Arabidopsis thaliana. Plant
Physiol. 116: 687–693.
Seo, M., Peeters, A.J., Koiwai, H., Oritani, T., Marion-Poll, A.,
Zeevaart, J.A., Koornneef, M., Kamiya, Y., and Koshiba, T.
(2000). The Arabidopsis aldehyde oxidase 3 (AAO3) gene product
catalyzes the final step in abscisic acid biosynthesis in leaves. Proc.
Natl. Acad. Sci. USA 97: 12908–12913.
Smith, J.K., Schloss, J.V., and Mazur, B.J. (1989). Functional expres-
sion of plant acetolactate synthase genes in Escherichia coli. Proc.
Natl. Acad. Sci. USA 86: 4179–4183.
Stepanova, A.N., Hoyt, J.M., Hamilton, A.A., and Alonso, J.M. (2005).
A link between ethylene and auxin uncovered by the characterization
of two root-specific ethylene-insensitive mutants in Arabidopsis. Plant
Cell 17: 2230–2242.
Stepanova, A.N., Robertson-Hoyt, J., Yun, J., Benavente, L.M., Xie,
D.Y., Dolezal, K., Schlereth, A., Jürgens, G., and Alonso, J.M.
(2008). TAA1-mediated auxin biosynthesis is essential for hormone
crosstalk and plant development. Cell 133: 177–191.
Strader, L.C., and Bartel, B. (2008). A new path to auxin. Nat. Chem.
Biol. 4: 337–339.
Sugawara, S., Hishiyama, S., Jikumaru, Y., Hanada, A., Nishimura,
T., Koshiba, T., Zhao, Y., Kamiya, Y., and Kasahara, H. (2009).
Biochemical analyses of indole-3-acetaldoxime-dependent auxin bi-
osynthesis in Arabidopsis. Proc. Natl. Acad. Sci. USA 106: 5430–
5435.
Tam, Y.Y., and Normanly, J. (1998). Determination of indole-3-pyruvic
acid levels in Arabidopsis thaliana by gas chromatography-selected
ion monitoring-mass spectrometry. J. Chromatogr. A 800: 101–108.
Tao, Y., et al. (2008). Rapid synthesis of auxin via a new tryptophan-
dependent pathway is required for shade avoidance in plants. Cell
133: 164–176.

Thimann, K.V., and Koepfli, J.B. (1935). Identity of the growth-pro-
moting and root-forming substances of plants. Nature 135: 101–102.
Tivendale, N.D., Davies, N.W., Molesworth, P.P., Davidson, S.E.,
Smith, J.A., Lowe, E.K., Reid, J.B., and Ross, J.J. (2010). Reas-
sessing the role of N-hydroxytryptamine in auxin biosynthesis. Plant
Physiol. 154: 1957–1965.
Tobeña-Santamaria, R., Bliek, M., Ljung, K., Sandberg, G., Mol, J.N.,
Souer, E., and Koes, R. (2002). FLOOZY of petunia is a flavin mono-
oxygenase-like protein required for the specification of leaf and flower
architecture. Genes Dev. 16: 753–763.
Vanneste, S., and Friml, J. (2009). Auxin: A trigger for change in plant
development. Cell 136: 1005–1016.
Warming, S., Costantino, N., Court, D.L., Jenkins, N.A., and Copeland,
N.G. (2005). Simple and highly efficient BAC recombineering using galK
selection. Nucleic Acids Res. 33: e36.
Woodward, A.W., and Bartel, B. (2005). Auxin: Regulation, action, and
interaction. Ann. Bot. (Lond.) 95: 707–735.
Yamada, M., Greenham, K., Prigge, M.J., Jensen, P.J., and Estelle,
M. (2009). The TRANSPORT INHIBITOR RESPONSE2 gene is re-
Rethinking Auxin Biosynthesis via IPyA 3973
quired for auxin synthesis and diverse aspects of plant development.
Plant Physiol. 151: 168–179.
Ye, S., and Cohen, J.D. (2009). Reevaluation of the role of the indole
pyruvate decarboxylase (IPyA) pathway in auxin biosynthesis in
plants. Poster 34, Hormone Biology. In Plant Biology 2009, July 18–
July 22, 2009, Honolulu, HI. American Society of Plant Biologists,
Rockville, MD. Abstract no. P34020.
Zhao, Y. (2010). Auxin biosynthesis and its role in plant development.
Annu. Rev. Plant Biol. 61: 49–64.
Zhao, Y., Christensen, S.K., Fankhauser, C., Cashman, J.R., Cohen,
J.D., Weigel, D., and Chory, J. (2001). A role for flavin monooxyge-
nase-like enzymes in auxin biosynthesis. Science 291: 306–309.
Zhao, Y., Hull, A.K., Gupta, N.R., Goss, K.A., Alonso, J., Ecker, J.R.,
Normanly, J., Chory, J., and Celenza, J.L. (2002). Trp-dependent
auxin biosynthesis in Arabidopsis: involvement of cytochrome P450s
CYP79B2 and CYP79B3. Genes Dev. 16: 3100–3112.
Zhou, R., Benavente, L.M., Stepanova, A.N., and Alonso, J.M. (2011).
A recombineering-based gene tagging system for Arabidopsis. Plant
J. 66: 712–723.