Sesquiterpene-like inhibitors of a 9-cis-epoxycarotenoid dioxygenase regulating abscisic acid biosynthesis in higher plants

Article abstract of DOI:10.1016/j.bmc.2009.01.076

Abscisic acid (ABA) is a carotenoid-derived plant hormone known to regulate critical functions in growth, development and responses to environmental stress. The key enzyme which carries out the first committed step in ABA biosynthesis is the carotenoid cl

Full text of DOI:10.1016/j.bmc.2009.01.076

Bioorganic & Medicinal Chemistry 17 (2009) 2902–2912
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Sesquiterpene-like inhibitors of a 9-cis-epoxycarotenoid dioxygenase
regulating abscisic acid biosynthesis in higher plants
Jason Boyd, Yuanzhu Gai, Ken M. Nelson, Erica Lukiwski, James Talbot , Mary K. Loewen,
Stacey Owen, L. Irina Zaharia, Adrian J. Cutler, Suzanne R. Abrams ,à, Michele C. Loewen à
*
Plant Biotechnology Institute, National Research Council of Canada, 110 Gymnasium Place, Saskatoon, SK, Canada S7N 0W9
a r t i c l e i n f o
a b s t r a c t
Article history:
Abscisic acid (ABA) is a carotenoid-derived plant hormone known to regulate critical functions in growth,
development and responses to environmental stress. The key enzyme which carries out the first commit-
ted step in ABA biosynthesis is the carotenoid cleavage 9-cis-epoxycarotenoid dioxygenase (NCED). We
have developed a series of sulfur and nitrogen-containing compounds as potential ABA biosynthesis
inhibitors of the NCED, based on modification of the sesquiterpenoid segment of the 9-cis-xanthophyll
substrates and product. In in vitro assays, three sesquiterpene-like carotenoid cleavage dioxygenase
(SLCCD) inhibitor compounds 13, 17 and 18 were found to act as inhibitors of Arabidopsis thaliana NCED
Received 21 December 2008
Revised 30 January 2009
Accepted 31 January 2009
Available online 8 February 2009
Keywords:
Abscisic acid biosynthesis
ABA biosynthesis inhibitor
9-cis-Epoxycarotenoid dioxygenase
Carotenoid metabolism
Xanthoxin analog
3 (AtNCED3) with Ki0s of 93, 57 and 87
lM, respectively. Computational docking to a model of AtNCED3
supports a mechanism of inhibition through coordination of the heteroatom with the non-heme iron in
the enzyme active site. In pilot studies, pretreatment of osmotically stressed Arabidopsis plants with com-
pound 13 resulted lower levels of ABA and catabolite accumulation compared to levels in mannitol-
stressed plant controls. This same inhibitor moderated known ABA-induced gene regulation effects
and was only weakly active in inhibition of seed germination. Interestingly, all three inhibitors led to
moderation of the stress-induced transcription of AtNCED3 itself, which could further contribute to low-
ering ABA biosynthesis in planta. Overall, these sesquiterpenoid-like inhibitors present new tools for con-
trolling and investigating ABA biosynthesis and regulation.
ABA analog
Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
1. Introduction
of ABA occurs principally through oxidation of one of the methyl
groups of the ring (80-carbon atom, using convention for ABA num-
Abscisic acid (1, ABA) is a plant hormone involved in the regu-
lation of important developmental functions including seed matu-
ration, desiccation tolerance and dormancy, as well as adaptation
to environmental stress through stomatal closure and modification
of gene expression.1–3 The biosynthesis of ABA 1 begins with iso-
pentenyl diphosphate which enters the mevalonic acid-indepen-
bering) mediated by members of a class of P450 monooxygenase
enzymes, CYP 707A.5 The catabolite phaseic acid (6, PA) which oc-
curs as the result of reversible cyclization of 80-hydroxyABA, is re-
duced by an unknown reductase to afford dihydrophaseic acid (7,
DPA). ABA can also be metabolized to the glucose conjugate 8.3
First identified in maize (VP14), NCEDs have also been found in
a variety of other species including Arabidopsis thaliana (AtNCED3),
bean (PvNCED1), tomato (LeNCED1), avocado (PaNCED1 and
PaNCED3) and cowpea (VuNCED1).6–11 AtNCED3 is a member of
the carotenoid cleavage enzyme family of A. thaliana, which con-
sists of nine enzymes.12 In general, the family is characterized by
dent 2-C-methyl-D-erythritol-4-phosphate pathway producing
plastidic isoprenoids, including carotenoids.4 Enzymatic cleavage
of C40-carotenoid cis-xanthophylls (neoxanthin 2 and violaxanthin
3) at the 110–120 double bond by a 9-cis-epoxycarotenoid dioxy-
genase (NCED) produces C15 (xanthoxin 4) and C25 metabolites
and represents the first committed step in ABA biosynthesis
(Fig. 1). Xanthoxin 4 is subsequently converted by an alcohol dehy-
drogenase (ABA2) into abscisyl aldehyde 5, which is oxidized to
ABA 1 by an abscisic aldehyde oxidase (AAO3).3 The catabolism
a plastid-targeting transit peptide, an amphipathic a-helix domain
and a catalytic domain which contains four conserved histidine
residues responsible for non-heme iron coordination. AtNCED3 is
found in both the stroma and bound to the thylakoid membrane,
accounts for NCED activity in roots, contributes to NCED activity
in developing seeds and is the major stress-induced NCED in leaves
of A. thaliana.12 Recently, immunohistochemical analysis revealed
that the AtNCED3 protein is detected exclusively in the vascular
parenchyma cells of water-stressed plants.13
* Corresponding author. Tel.: +1 306 975 5569; fax: +1 306 975 4839.
E-mail address: sue.abrams@nrc-cnrc.gc.ca (S.R. Abrams).
Current address: Department of Biochemistry, University of Saskatchewan, 107
Wiggins Road, Saskatoon, SK, Canada S7N 0W5.
à
These authors contributed equally to this work.
0968-0896/$ - see front matter Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2009.01.076

J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
2903
Figure 1. ABA biosynthesis and catabolism pathway of higher plants from the committed step of C40-carotenoid cleavage of either 9-cis-neoxanthin 2 or 9-cis-violaxanthin 3
by AtNCED3.
Due to ABA’s important role in plant physiology, significant ef-
fort has been expended on investigating functional aspects of ABA
1 biosynthesis, regulation and action. ABA-deficient mutants are
powerful tools for elucidating ABA’s role in planta, as are chemical
inhibitors of ABA 1 biosynthesis which have broad applicability to
many plant species. General carotenoid biosynthesis inhibitors
such as fluridone, a potent broad spectrum herbicide that inhibit
phytoene desaturase in the carotenoid biosynthesis pathway, have
been used to inhibit ABA 1 biosynthesis.14,15 While fluridone does
inhibit ABA 1 biosynthesis, a corresponding general repression of
the carotenoid biosynthesis pathway limits its application for bio-
chemical investigations including those of carotenoid cleavage en-
zymes and products. To address this problem, Abamine
compounds 9 and 10 were developed as inhibitors of NCED’s, based
on early observations that a number of inhibitors of soybean
lipoxygenase were effective in reducing ABA accumulation in
stressed soybean cell cultures and seedlings.16 One of the active
compounds, nordihydroguaiaretic acid, served as the starting
structure for generation of analogs with improved NCED inhibitory
activity, leading to development of the tertiary amines Abamine (9,
ABM) and Abamine SG (10, ABM-SG) (Fig. 2).17,18 Arabidopsis plants
treated with ABM 9 showed a significant decrease in drought toler-
ance and under simulated osmotic stress ABM 9 inhibited stomatal
closure in spinach leaves. The latter effect was counteracted by co-
application of ABA 1. ABM-SG 10 strongly inhibited the expression
of ABA-responsive and catabolic genes in plants under osmotic
stress. Finally, both ABM 9 and ABM-SG 10 reduced ABA metabolite
accumulation by 35% and 77%, respectively and were shown to act
as competitive inhibitors of the cowpea NCED enzyme, with K0is of
18.5 lM and 38.8 lM, respectively.

2904
J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
Figure 2. Structures of AtNCED3 SLCCD inhibitors.
In this report we describe the design, synthesis and character-
by oxidation of allylic alcohol 22 with MnO2, followed by imine
ization of a novel set of sesquiterpene-like carotenoid cleavage
dioxygenase (SLCCD) inhibitors 11–18. The current compounds
were designed starting with the sesquiterpenoid subunit of the
substrate and product of the NCED enzyme. Of eight initial inhibi-
tors designed and tested (Fig. 2), three were found to inhibit re-
combinant AtNCED3 activity strongly. These have been fully
characterized in vitro, with kinetic inhibition constants comparing
favorably to those of the ABM-type compounds. Computational
docking of the inhibitors correlated with these findings and sup-
ported the proposed functional mechanism. In preliminary
in vitro analyses, one inhibitor in particular, SLCCD inhibitor com-
pound 13 was found to moderate ABA-responsive genes and ABA
metabolism. Interestingly, all three inhibitors reduced expression
of AtNCED3, presenting a second mechanism for inhibition of
ABA 1 biosynthesis by the molecules.
formation using phenyl amine and then reduction to the amine.
2.2. In vitro assays and kinetic analyses
Recombinant AtNCED3 including a C-terminally located gluta-
thione-S transferase fusion tag was expressed in Escherichia coli
and purified by affinity chromatography. In vitro assays demon-
strated the functionality of the recombinant purified enzyme prod-
uct. Sample HPLC profiles (Supplementary data Fig. 1) show
cleavage of the 90-cis-neoxanthin 2 substrate (tR 14.2 min. with
three maxima at 415, 438 and 467 nm) producing the expected
C25-allenic apo-aldehyde cleavage product (tR 11.6 min. with a
maxima of 423 nm). Further kinetic analysis fitted by non-linear
regression analysis defined a Km of 24
lates well with the Km0s of 27
M and 49.0
ously for VP14 and VuNCED1.18,24
l
M (Fig. 4). This value corre-
l
l
M determined previ-
Using this recombinant enzyme and assay system, the eight po-
tential inhibitor compounds were tested for their relative ability to
inhibit AtNCED3 activity at 1 mM concentration (Fig. 5). Com-
pounds 12, 17 and 18 completely inhibited AtNCED3 activity at
1 mM, while 13 inhibited AtNCED3 activity by 75%. Compound
12 is one of the stereoisomers of racemic 13. The latter being easier
to synthesize (and thus of higher potential practical application), it
was decided to move forward with compounds 13, 17 and 18 for
detailed in vitro and in vitro testing. Dixon plots indicated that
compounds 13, 17 and 18 competitively inhibit recombinant AtN-
CED3 with Ki0s comparable or better than those observed for ABM
and ABM-SG (Table 1 and Supplementary data Fig. 2).
2. Results
2.1. Design and synthesis of the SLCCD inhibitors
The present compounds were designed to incorporate the 9-cis
double bond geometry of the substrates and product of AtNCED3
as well as a heteroatom at carbon 12 (carotenoid numbering) of
the inhibitor molecules. All of the SLCCD inhibitors 11–18 were
synthesized from 4-oxoisophorone 19 (Fig. 3). Bakers0 yeast reduc-
tion of 19 afforded (ꢀ)-(R)-2,2,6-trimethylcyclohexa-1,4-dione19
which was converted into chiral nonracemic allylic alcohols 20,
21, 23 and 24.20 Racemic allylic alcohol 22 was prepared in a sim-
ilar manner, except that reduction of 19 was accomplished using
zinc in acetic acid.21 The terminal allylic alcohols were then con-
verted to the corresponding ethyl sulfides by reaction with ethyl
disulfide in the presence of tributylphosphine.22 Inhibitor 16 was
obtained by reacting 2-thiopheneacetyl chloride and allylic alcohol
22 (protected as the neopentylglycol ketal). The xanthoxin-like
allylic alcohol 22 was prepared through a Sonogashira coupling
between the terminal acetylene in 2123 and (Z)-3-iodobut-2-en-
1-ol. Alcohol 22 was then converted to the phenyl sulfide 13 with
54% yield. The nitrogen-containing inhibitor 18 was synthesized
2.3. Homology modeling and SLCCD inhibitor docking
Recently a crystal structure was determined for Synechocystis
apocarotenoid-15,150-oxygenase (ACO), a fungal homologue of the
NCEDs.25 AtNCED3 shares 25% identity and 45% similarity with
ACO at the amino acid level. Homology modeling using the Swiss-
Model servers generated a hypothetical protein structure of AtN-
CED3 which maintained the octahedral coordination of the four
active site histidines at 2.14, 2.05, 2.16 and 2.31 Å from the iron atom
for H164, H211, H276 and H450, respectively.26 Structural differ-

J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
2905
Figure 3. Synthesis of AtNCED3 SLCCD inhibitors. (a) See Ref. 20; (b) n-Bu3P, (C2H5S)2; (c) TBAF, THF; (d) see Ref. 22; (e) (Z)-3-iodobut-2-en-1-ol, (Ph3P)4Pd, CuI, (i-Pr)2NH; (f)
n-Bu3P, (C6H5S)2; (g) MnO2; (h) C6H5NH2, ; (i) NaBH4.
D
Figure 4. Kinetic analysis of recombinant purified AtNCED3 activity. Michaelis–
Menton plot for cleavage of 9-cis-neoxanthin 2 by recombinant AtNCED3 indicating
a Km of 24 lM.
ences between the AtNCED3 model and ACO were limited to small
surface exposed loops related to a few minor alignment gaps. As con-
trols to test the AtNCED3 model, 90-cis-neoxanthin 2, the substrate of
ACO (all-trans-(3R)-hydroxy-80-apo-b-carotenol (3-ON)), and
Figure 5. Relative inhibition of recombinant AtNCED3 activity by various SLCCD
compounds at 1 mM concentration.

2906
J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
Table 1
binding pocket. The nitrogen of 18 docked 2.67 Å away from the
iron atom. The sulfur atoms of 12 and 17 docked 2.84 and 2.65 Å
away from the iron atom, respectively. Other SLCCD inhibitor mol-
ecules that performed poorly in the in vitro trials generally were
not targeted to the catalytic site of the binding pocket, or in some
instances were not targeted to the binding pocket at all during
docking.
Kinetic parameters for AtNCED3 cleavage of the substrate 9-cis-neoxanthin 2, and
inhibitory effects of SLCCD inhibitor compounds
Compound
Ki (
lM)
Km (lM)
9-cis-Neoxanthin
13
17
18
10 (ABM-SG)
9 (ABM)
—
24
—
—
—
—
—
93
57
87
86
2.4. Effect of SLCCD inhibitors on ABA accumulation under
osmotic stress
132
A. thaliana plants were treated with either ABM 9 or the inhib-
itor compounds 13, 17 and 18, to evaluate their ability to reduce
ABA biosynthesis induced by an osmotic stress. Essentially plants
were treated inhibitor compound for 2 h followed by mannitol
stress-treatment in the presence of the same compounds. Mannitol
stress has been shown to result in loss of turgor with a correspond-
ing increase in ABA levels through the induction of AtNCED3 in A.
thaliana.7,27 As expected, mannitol treatment alone resulted in an
elevation of the levels of ABA and catabolites peaking 24 h after
the imposition of treatment, compared to the levels in non-trea-
ted/non-stressed plants (Fig. 7). Accumulation of ABA and catabo-
lites dropped off by 48 h as described previously.28 Treatment with
compound 13 for 2 h prior to and then during mannitol stress-
treatment resulted in levels of ABA and catabolites remaining com-
parable to those of the non-treated/non-stressed control plants in
the first 6 h. By 24 h, the total levels of ABA and catabolites in
the compound 13 treated plants increased to only 8211 pmol/g,
significantly below those of the mannitol-stressed only plants
(15,147 pmol/g). Similar treatment of plants with ABM 9 resulted
in higher levels of ABA and catabolites at the first time point, with
levels remaining constant (and higher than those for treatment
with compound 13) over the remaining time course of the experi-
ment. The remaining two inhibitors, 17 and 18, were less effective
than 13 in reducing the effect of the osmotic stress on ABA and
catabolite pools. Interestingly, the overall effect observed for com-
pound 13 is not represented in individual plots of ABA levels (or
any one other catabolite) alone (Supplementary data Fig. 4). It is
only when total accumulation of ABA and its catabolites are con-
sidered that the overall effect becomes evident.
xanthoxin 4 structures were docked (Fig. 6A and Supplementary
data Fig. 3A and B). The 3-ON molecule docked to the AtNCED3 mod-
el with a similar orientation as observed in the ACO crystal with its b-
ionone ring oriented towards the tunnel entrance but shifted in to-
ward the catalytic site by 5 Å. This positions the C12 and C13 bond
within 3.95 Å of the iron atom and 2.10 Å of a coordinated active site
water molecule. Docking of 9-cis-neoxanthin 2 resulted in the epox-
ide ring entering the proteinchannel first, yieldinga final orientation
with the C11–C12 bond 4.4 Å away from and directly over the iron
atom and 2.3 Å away from the active site water molecule. The xanth-
oxin molecule docked in the opposite orientation from the 9-cis-
neoxanthin substrate, with its epoxide ring towards the tunnel en-
trance and its C10 carbon atom 3.6 Å and 1.9 Å from the iron atom
and water molecule, respectively.
Docking results correlated well with the in vitro enzyme assay
data. Structures representing 12 (the more active stereoisomer of
the racemic compound 13), 17 and 18 (Fig. 6B and C and Supple-
mentary data Fig. 3C, respectively) all docked in the same orienta-
tion as xanthoxin, in close proximity to the iron atom in the
Figure 7. Total ABA metabolite levels in mannitol stressed Arabidopsis thaliana
plants treated with SLCCD inhibitors. Plants were treated with 33 lM inhibitors
Figure 6. Computational docking of compounds to the AtNCED3 homology model.
Docking was performed using the Autodock v3.1 software.38 Conserved histidine
residues are shown coordinating the iron (orange). The active site water (light blue)
is shown in relation to the docked molecules. Molecules include (A) 9-cis-
neoxanthin 2, (B) compound 17 and (C) compound 12. Sulfur heteroatoms in the
two SLCCD compounds are highlighted in yellow.
(compounds 13 (inverted closed triangles), 17 (closed circles), 18 (open circles) and
ABM 9 (open triangles)) for 2 h prior to being stressed with mannitol. Plants were
harvested 6 h after the inhibitor treatment and compared to plants that were
mannitol stressed only (open squares), or non-treated/non-stressed plants (closed
squares). Metabolites quantified and summed at each time point include abscisic
acid-glucose ester, dihydrophaseic acid, phaseic acid and abscisic acid.

J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
2907
2.5. Effect of SLCCD inhibitors on A. thaliana seed germination
Seed germination assays were performed for compounds 13, 17
and 18 to assess the ABA-like character of the inhibitors.29 Inhibi-
tors had relatively little effect on seed germination at low concen-
tration compared to non-inhibitor treated and ABA treated controls
(Fig. 8). At increasing concentrations (0.33 lM) the inhibitors did
lead to reductions of seed germination by approximately 15%, com-
pared to 47% for the (+)-ABA 1. Both compounds 13 and 17 reduced
seed germination by 26% at 1 lM while 18 showed a more pro-
nounced effect with a 50% reduction compared to 61% for (+)-
ABA 1. At the highest concentration tested, compound 13 still only
had a modest impact on seed germination at 38% reduction, while
compounds 17 and 18 showed 51% and 71% reductions, respec-
tively, compared to 96% for (+)-ABA 1.
2.6. Effect of SLCCD inhibitors on target gene transcript levels
under osmotic stress
In light of the observed effectiveness of compound 13 in moder-
ating ABA and catabolite levels in vitro and its limited effect on
seed germination, it was targeted for further evaluation. Specifi-
cally, quantitative reverse-transcription PCR was used to assess
inhibitor induced changes in gene transcript levels in mannitol
stressed plants. The gene targets chosen for this purpose were AtN-
CED3, the ABA and drought inducible Rd29B and the ABA (induc-
ible) catabolic genes CYP707A1 and CYP707A3. Transcript levels
were normalized against UBQ10 mRNA levels.5,30,31 Mannitol treat-
ment led to the induction of expression of all four target genes
within 4 h of the stress-treatment (Fig. 9). Subsequently the man-
nitol-induced gene transcription levels decrease back to non-trea-
ted/non-stressed levels by 24 h post-treatment and remained low
through 48 h (Supplementary data Figs. 5 and 6). In general, pre-
treatment with compound 13 at both 10 and 33 lM concentrations
prior to mannitol-stress led to reductions in the accumulation of
mRNA transcript levels at 6 h post-compound treatment for
Rd29B, CYP707A1 and CYP707A3 compared to the mannitol-
stressed control (Fig. 9A). The inhibition of mannitol-induced
Rd29b transcription by compound 13 (about 90%) is especially
striking and is consistent with the mannitol effect on Rd29b being
primarily mediated by ABA. This result indicates the potential of
this inhibitor for dissecting the role of ABA in physiological and
developmental processes. As observed in mannitol stressed only
Figure 9. Target gene transcript levels in mannitol stressed Arabidopsis thaliana
plants treated with SLCCD inhibitors. (A) Effect of compound 13. Plants were treated
with 10 or 33 lM compound 13 for 2 h prior to being stressed with mannitol. Plants
were harvested 6 h after the SLCCD compound treatment and compared to
mannitol stressed only plants or non-treated/non-stressed plants. Target genes
included Rd29B, CYP707A1 and CYP707A3. Transcript levels were normalized against
the UBQ10 gene. (B) Effects of SLCCD compounds on AtNCED3 expression.
Experiments were carried out as described for A, but with each of SLCCD
compounds 13, 17 and 18.
plants, transcript levels in compound 13 pre-treated plants de-
creased back to non-treated/non-stressed levels by 24 h and re-
mained low through 48 h (Supplementary data Fig. 5). In
addition to this, compound 13 was also found to decrease the rel-
ative expression levels of AtNCED3 in mannitol stressed plants
(Fig. 9B). While the former results emphasize the lack of ABA-like
character for compound 13, the moderation of AtNCED3 transcrip-
tion represents a useful inhibitor-dependent side-effect that likely
further contributes to lowering ABA levels in planta. Testing of
compounds 17 and 18 demonstrated similar, although not as pro-
nounced effects on AtNCED3 expression (Fig. 9B, Supplementary
data Fig. 6).
Figure 8. Germination of Arabidopsis thaliana plants in the presence of each of
SLCCD inhibitor compounds 18 (inverted closed triangles), 13 (closed circles), 17
(open circles), (+)-ABA 1 (open triangles), and germination of plants on media
without added compounds (closed square).

2908
J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
structures to enter the cavity. Therefore in contrast to ACO, AtN-
3. Discussion
CED3 likely determines substrate positioning based on where the
molecule interacts with the internal terminus of the binding pock-
et. Docking to the AtNCED3 model highlights that this is likely the
case, as 3-ON was oriented with its C13–C14 bond over the iron,
and the b-ionone ring pulled inside the tunnel entrance. Docking
of 90-cis-neoxanthin 2 resulted in the epoxide ring being buried
in the AtNCED3 catalytic pocket. This positioned the C11–C12 bond
over the iron atom in a suitable position for catalytic cleavage at
the expected location. These results emphasize the validity and po-
tential utility of the AtNCED3 model.
The xanthoxin 4 molecule docked with its epoxide ring in the
opposite orientation (similar to 3-ON) to that of the 90-cis-neoxan-
thin 2. While this likely does not represent its native orientation
following cleavage of the 9-cis-neoxanthin 2 substrate, it empha-
sizes the accommodating size of the AtNCED3 entrance tunnel
and that the preferred orientation of single ring containing mole-
cules is with the ring pointing toward the entrance. Docking results
for the SLCCD inhibitors seem to follow this preference with the
hydroxylated rings preferentially pointing toward the entrance.
In the ACO crystal structure a coordinated water molecule occu-
pies the fifth ligand position within the iron octahedral coordina-
tion structure. The water molecule, theorized to be an oxygen
donor and required for catalytic activity, is located 3.2 Å from the
C15 of the substrate and 2.07 Å from the non-heme iron atom.25
Each of the three active SLCCD inhibitors docked with their hetero-
atoms (nitrogen or sulfur) within 2.8 Å of the iron atom such that
they would be sufficient to occupy the coordinate space of the
water molecule in the ACO structure and stop catalysis.
3.1. Design/synthesis and inhibitory activities of SLCCD
inhibitors
The design of inhibitors described herein focuses on specific
interaction with the non-heme iron atom within AtNCED3, a defin-
itive motif of carotenoid cleavage enzymes. It was envisioned that
a molecule maintaining characteristics of the native enzyme sub-
strate 9-cis-neoxanthin 2 or xanthoxin 4 product, but presenting
a nitrogen or sulfur heteroatom might specifically occupy the ac-
tive site of the enzyme with the heteroatom interacting with the
non-heme iron, resulting in inactivation of the enzyme. Similar
concepts have been applied to inhibitors of other dioxygenase
enzymes.32,33
In earlier ABA structure activity studies, analogs with the side
chain having a triple bond conjugated to a cis double bond were
found to be highly active and were also readily synthesized.20
Therefore the enyne feature was incorporated into the design of
the present set of eight potential ABA biosynthesis inhibitors. The
epoxy alcohol analogs 17 and 18, which most closely resemble
the substrate and product of the NCED, strongly inhibited the
NCED enzyme activity in vitro, and demonstrated higher inhibitory
function than ABM 9 in this assay. However, in the experiment
simulating drought stress, 17 and 18 were relatively weak inhibi-
tors of ABA biosynthesis. As well, the aniline derivative 18 had a
fairly pronounced (and undesirable) ABA-like effect on seed germi-
nation, with the thiophenyl analog 17 demonstrating a moderate
effect. In addition, these compounds were prepared from the epoxy
alkyne 25, a compound that was not readily accessible, so these
structural types were not pursued.
3.3. In vivo effects of SLCCD inhibitors
Compounds with a tertiary alcohol at the junction of ring and
side chain and either ketone or alcohol at C-4 were also envisioned
to be possible inhibitors, as the general shape of the molecule and
oxygen atom would be maintained. The keto allylic alcohol precur-
sors 20–22 were more conveniently prepared, affording both race-
mic and enantiomerically pure compounds. This was desirable as
we had found earlier that the individual enantiomers 20 and 21
of the allylic alcohol 22 had different properties as competitive
inhibitors of ABA perception.34 The analog 20 competitively
blocked ABA perception, while its enantiomer was a weak ABA
agonist. On observing significant NCED inhibition with the racemic
compound 13, comparable with that of ABM 9 and ABM-SG 10, we
anticipated similar differences might be found in the present case,
and the thioethyl derivatives of compounds 20 and 21 were syn-
thesized and tested. Again, the stereochemistry of the analogs
had an effect. Compound 12 inhibited the enzyme as strongly as
the more xanthoxin-like compounds, while the other enantiomer
11 had reduced activity in the in vitro enzyme assay.
The basic premise of this work lies in the design of inhibitors
that bind to and inactivate the NCED enzyme responsible for the
first committed step in ABA 1 biosynthesis. In a recent study on ef-
fects of drought stress on signaling and gene expression in Arabid-
opsis, it had been shown that the levels of ABA and its catabolites
phaseic acid 6, dihydrophaseic acid 7 and ABA glucose ester 8 were
all found to increase on imposition of the stress.28 In the present
study to compare the effects of potential inhibitors on ABA biosyn-
thesis capacity, an osmotic stress treatment of Arabidopsis plants
was substituted for the drought stress. ABA biosynthetic inhibitors
were designed and tested and in the case of compound 13 were
shown to significantly reduce the accumulation of ABA 1 and the
catabolites 6–8 in plants subjected to osmotic stress. While the
rationale for inhibitor design was based on maintaining structural
characteristics similar to the enzymes substrate and products to
maximize specificity, this also meant that the inhibitors share
structural characteristics with ABA 1 itself. Obviously an inhibitor
of ABA 1 biosynthesis should not mediate ABA signaling.
Toward assessing the ABA-like character of the inhibitors their
ability to mediate known ABA 1 effects at the levels of seed germi-
nation and gene regulation were determined. In general, the SLCCD
inhibitors were found to be weaker germination inhibitors than
(+)-ABA 1, with compound 13 having 60–70% less effect. Interest-
ingly, low concentrations of compounds 13 and 17 had slight pro-
motion effects on seed germination. As well, treatment of
mannitol-stressed plants with compound 13 led to a reduction of
transcript levels for three genes known to be (+)-ABA 1 induc-
ible.5,30 The reduction of transcription mediated by this inhibitor
is in agreement with previous observations made for alternate
inhibitors and likely results from the reduction of endogenous
ABA 1 levels.17 Overall, these results emphasize that SLCCD inhib-
itor 13 does not generally simulate ABA-inducible responses and
thus does not maintain ABA-like characteristics.
Two diasteromeric hydroxy compounds 14 and 15 were synthe-
sized to explore the effect of changing the oxidation level of the C-4
or position of the oxygen atom. In the in vitro enzyme assay, the hy-
droxyl compounds did not afford greater activity, and were not pur-
sued. Compound 16was incorporated into the setof test molecules to
determine if positioning the sulfur atom further from the cyclohexa-
none ring would have an effect on activity compared to that of 13.
3.2. Computational analysis of SLCCD inhibitor-enzyme
complexes
In the ACO structure the binding pocket entrance is proposed to
act as a bottleneck, arresting movement of 3-ON to the interior and
positioning the C15–C150 bond over the iron molecule in a trans
conformation.25 In contrast, AtNCED3 must accept substrate mole-
cules with rings at both extremities, and thus it would be expected
that the binding pocket entrance be sufficiently large to allow ring

J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
2909
Finally, these pilot in vitro studies demonstrate that mannitol
stress leads to induction of AtNCED3 gene expression as reported
previously.7 While stress-induced, it is not clear whether AtNCED3
is specifically ABA-inducible. But from the results reported here, it
is clear that application of the SLCCD inhibitors significantly re-
duces AtNCED3 mRNA levels under stress conditions, which would
further contribute to reducing ABA 1 biosynthesis in planta. While
this characteristic was not specifically sought in designing the
inhibitors, in terms of the overall objective of inhibiting ABA 1 bio-
synthesis, a reduction in the primary biosynthetic enzyme is a very
useful side-effect.
The relatively poor effects of inhibitors 17 and 18 in planta were
surprising considering their effectiveness in vitro and docking re-
sults in silico. This lack of efficacy in vitro could be due to many fac-
tors, including stability of the different compounds in the plant and
the presence of the hydrophobic aromatic rings in both 17 and 18
structures, possibly reducing their permeability through the roots
and transport to the site of action. This lack of efficacy in moderat-
ing ABA levels in vitro could be due to many factors, including sta-
bility of the different compounds in the plant and the presence of
the hydrophobic aromatic rings in both 17 and 18 structures, possi-
bly reducing their permeability through the roots and transport to
the site of action. The discrepancy between in vitro and in vitro re-
sults is consistent also in the AtNCED3 expression profiling where
13 led to the highest reduction of stress-induced gene expression.
(2 mL) was stirred at room temperature for 6 h. Work up as de-
scribed above, followed by purification by FCC (ethyl acetate/hex-
ane, 15:85 v/v) to afford 12 (22 mg, 63%). ½a D25
ꢁ
+15 (c 1.0, CHCl3).
The spectral characterization data was identical to enantiomer 11.
5.1.3. (4S,5R/4R,5S)-(30Z)-4-(50-(Ethylthio)-30-methylpent-30-en-
10-ynyl)-4-hydroxy-3,3,5-trimethylcyclohexanone (13)
A solution of allylic alcohol 22, protected as the neopentylglycol
ketal20, (34 mg, 0.1 mmol), ethyl disulfide (34
lL, 0.27 mmol) and
n-Bu3P (62
lL, 0.25 mmol) in CH2Cl2 was stirred at room tempera-
ture for 4.5 h. Work up as described above, followed by purification
by FCC (ethyl acetate/hexane, 10:90 v/v) to afford the sulfide
(22.1 mg, 58%). 1H NMR (CDCl3) d: 5.67 (1H, ddq, 1.5, 7.75,
7.75 Hz, @CH), 3.54 (2H, d, 11.25 Hz, OCH2), 3.36 (2H, ddd, 1.75,
11.25, 13.25 Hz, OCH2), 3.31 (2H, dd, 0.75, 7.75 Hz, SCH2), 2.48
(2H, q, 7.5 Hz, SCH2CH3), 2.24 (1H, dd, 3.25, 14.25 Hz, H-2), 2.18
(1H, m, H-5), 1.96 (1H, dt, 3.25, 13.5, H-6), 1.87 (3H, d, 1.0 Hz,
CH3), 1.57 (1H, dd, 13.5, 13.5 Hz, H-6), 1.46 (1H, d, 14.25 Hz, H-
2), 1.22 (3H, t, 7.5 Hz, CH3), 1.12 (3H, s, CH3), 1.09 (3H, s, CH3),
1.04 (3H, d, 7.5 Hz, CH3), 1.04 (3H, s, CH3), 0.82 (3H, s, CH3). To a
solution of the ketal protected sulfide (160 mg, 0.4 mmol) in ace-
tone (5 mL) was added 2 N HCl (8 drops). The mixture was stirred
at room temperature for 40 min. After evaporation of acetone,
ether was added and washed with satd NaHCO3, dried and concen-
trated to give a residue which was purified by FCC (ethyl acetate/
hexane 20:80 v/v) to provide 13 (100 mg, 80%). The spectral char-
acterization data was identical to pure enantiomer 11.
4. Conclusions
5.1.4. (1S,4R,6R)-(30Z)-1-(50-(Ethylthio)-30-methylpent-30-en-10-
ynyl)-2,2,6-trimethylcyclohexane-1,4-diol (14)
While in vitro studies identified SLCCD compound 17 as the
most promising candidate inhibitor, hormone profiling data con-
vincingly demonstrated that SLCCD 13, a more easily synthesized
racemic compound, best met the objective of reducing the total
ABA metabolite levels in planta. Overall, these sesquiterpenoid-like
inhibitors present new tools for controlling and investigating ABA
biosynthesis, regulation and effects.
A solution of allylic alcohols 23 and 2420 (200 mg, 0.55 mmol),
(C2H5S)2 (102 lL, 0.83 mmol) and n-Bu3P (203 lL, 0.83 mmol) in
CH2Cl2 (5 mL) was stirred at room temperature for 6 h. Work up
as described above, followed by purification by FCC (ethyl ace-
tate/hexane, 5:95 v/v) to provide 25 (41 mg, 17%), 26 (18.4 mg,
8%) and recovery of the unreacted starting material (70 mg, 35%).
For 25: 1H NMR (CDCl3) d: 5.67 (1H, dt, 1.5, 7.75 Hz, @CH), 3.92
(1H, m, H-4), 3.32 (2H, dd, 0.75, 7.75 Hz, CH2S), 2.49 (2H, q,
7.25 Hz, SCH2CH3), 2.32 (1H, m, H-6), 1.87 (3H, d, 1.0 Hz, CH3),
1.76 (1H, br s, OH), 1.62 (1H, dd, 3.5, 14.25 Hz, H-3), 1.57 (2H, m,
H-5), 1.49 (1H, d, 14.25 Hz, H-3), 1.23 (3H, t, 7.25 Hz, SCH2CH3),
1.20 (3H, s, CH3), 1.06 (3H, s, CH3), 1.04 (3H, d, 6.5 Hz, CH3), 0.86
(9H, s, SiCMe3), 0 (6H, s, SiMe2). To a solution of 25 (41 mg,
0.1 mmol) in THF (1.5 mL) was added TBAF (1 M solution in THF,
0.5 mL, 0.5 mmol). The reaction mixture was stirred at room tem-
perature for 1 day and diluted with ether. The mixture was washed
with water (10 mL ꢂ 3), dried, concentrated and fractionated by
FCC (10% ethyl acetate/hexane, 10:90 v/v increased to 35:65 v/v)
5. Experimental
5.1. Synthesis of ABA analogue inhibitors
5.1.1. (4S,5R)-(30Z)-4-(50-(Ethylthio)-30-methylpent-30-en-10-
ynyl)-4-hydroxy-3,3,5-trimethylcyclohexanone (11)
A solution of alcohol 2020 (25 mg, 0.1 mmol), ethyl disulfide
(25 lL, 0.2 mmol) and n-Bu3P (49 lL, 0.2 mmol) in CH2Cl2
(1.5 mL) was stirred at room temperature for 4.5 h. Ethanol
(1 mL) was added to the reaction and the resulting mixture was
stirred for 20 min. Ethanol was removed by evaporation and
CH2Cl2 (15 mL) was added. The organic phase was washed with
0.5 N NaOH and brine successively, dried and concentrated to give
a residue which was purified by FCC (ethyl acetate/hexane, 15:85
v/v) to provide 11 (19.2 mg, 62%) and recover 20 (4 mg, 19%).
to provide 14 (21.3 mg, 71%). ½a D25
ꢀ13 (c 0.94, CH2Cl2); IR (KBr):
ꢁ
3332, 2976, 2879, 1450 cmꢀ1 1H NMR (CDCl3) d: 5.67 (1H, dt,
;
1.5, 7.75 Hz, @CH), 4.0 (1H, m, H-4), 3.30 (2H, dd, 1.0, 7.75 Hz,
CH2S), 2.47 (2H, q, 7.5 Hz, SC2H5), 2.32 (1H, m, H-6), 1.85 (3H, s,
CH3), 1.63 (4H, m, H-5 and H-3), 1.19 (3H, t, 7.5 Hz, SC2H5), 1.19,
(3H, s, CH3), 1.08 (3H, s, CH3), 1.03 (3H, d, 6.5 Hz, CH3); 13C NMR
(CDCl3) d: 133.2, 120.0, 94.1, 85.7, 79.1, 66.8, 44.5, 40.1, 38.8,
31.9, 31.6, 27.5, 25.2, 23.3, 23.1, 16.1, 14.9; HRMS CI+ NH3 m/z calcd
for C17H32NO2S: 314.2154, found: 314.2162.
½
a 2D5
ꢁ
ꢀ16 (c 0.48, CHCl3); IR (KBr): 3463, 2975, 2872, 1688 cmꢀ1
;
1H NMR (CDCl3) d: 5.76 (1H, dt, 1.25, 7.75 Hz, @CH), 3.31 (2H, d,
8.75 Hz, CH2S), 2.65 (1H, d, 14.25 Hz, H-2), 2.48 (2H, q, 7.5 Hz,
SCH2CH3), 2.29 (3H, m, H-5 and H-6), 2.08 (1H, d, 14.25 Hz, H-2),
1.89 (3H, s, CH3), 1.22 (3H, t, 7.5 Hz, SCH2CH3), 1.20 (3H, s, CH3),
1.14 (3H, s, CH3), 0.97 (3H, s, CH3); 13C NMR (CDCl3) d: 209.2,
134.2, 119.5, 92.6, 86.8, 77.4, 52.9, 47.0, 42.2, 37.4, 31.6, 25.9,
25.4, 23.2, 20.8, 16.6, 14.9; HRMS EI+ m/z calcd for C17H26O2S:
294.1654, found: 294.1655.
5.1.5. (1R,4R,6R)-(30Z)-1-(50-(Ethylthio)-30-methylpent-30-en-10-
ynyl)-2,2,6-trimethylcyclohexane-1,4-diol (15)
To a solution of 2620 (18.4 mg, 0.045 mmol) in THF (1.2 mL) was
added TBAF (1.0 M solution in THF, 0.13 mL, 0.13 mmol). The reac-
tion was stirred at room temperature for 1 day. Work up as for 14
to provide product 15 (8.5 mg, 64%) and recovered starting mate-
5.1.2. (4R,5S)-(30Z)-4-(50-(Ethylthio)-30-methylpent-30-en-10-
ynyl)-4-hydroxy-3,3,5-trimethylcyclohexanone (12)
A solution of alcohol 2120 (28 mg, 0.11 mmol), diethyl sulfide
rial (4.5 mg, 24%). ½a D25
ꢁ
+10 (c 0.25, CH2Cl2); IR (KBr): 3388, 2968,
2922, 1458 cmꢀ1 1H NMR (CDCl3) d: 5.71 (1H, dt, 1.5, 7.75 Hz,
;
(28 lL, 0.22 mmol) and n-Bu3P (55 lL, 0.22 mmol) in CH2Cl2

2910
J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
@CH), 3.87 (1H, m, H-4), 3.33 (2H, dd, 1.0, 7.75 Hz, CH2S), 2.51 (2H,
q, 7.7 Hz, SC2H5), 2.00 (1H, m, H-6), 1.88 (3H, d, 1.5 Hz, CH3), 1.67
(1H, ddd, 2.5, 4.5, 12.75 Hz, H-5), 1.57 (1H, dd, 11.5, 12.5 Hz, H-5),
1.35 (1H, dd, 12.5, 24.25 Hz, H-3), 1.23 (3H, t, 7.25 Hz, CH3), 1.13
(3H, s, CH3), 1.07 (3H, d, 6.5 Hz, CH3), 1.02 (3H, s, CH3); 13C NMR
(CDCl3) d: 133.4, 119.9, 93.9, 86.3, 78.3, 66.2, 46.8, 41.7, 39.9,
35.7, 31.6, 27.0, 25.3, 23.2, 20.8, 16.5, 14.9; HRMS EI+ m/z calcd
for C17H28O2S: 296.1810, found: 296.1822.
5.1.9. (1R,3S,6R)-(30Z)-1,5,5-Trimethyl-6-(30-methyl-50-
(phenylamino)-pent-30-en-10-ynyl)-7-oxabicyclo[4.1.0]heptan-
3-ol (18)
A solution of aldehyde 29 (16 mg, 0.065 mmol) and aniline
(10 lL, 0.11 mmol) in ethanol (1.5 mL) was refluxed for 30 min.
The reaction mixture was cooled to room temperature and then
NaBH4 (7.4 mg, 0.2 mmol) was added. The resulting mixture was
stirred at room temperature for 15 min and water (3 mL) with gla-
cial acetic acid (1 drop) was added. The ethanol was evaporated off
and water phase was extracted with ether, dried, concentrated and
fractionated by FCC (ethyl acetate/hexane, 35:65 v/v) to provide
5.1.6. (1R,3S,6R)-(30Z)-6-(50-Hydroxy-30-methylpent-30-en-10-
ynyl)-1,5,5-trimethyl-7-oxa-bicyclo[4.1.0]heptan-3-ol (28)
A mixture of compound 2720 (18 mg, 0.1 mmol), (Z)-3-iodobut-
2-en-1-ol (30 mg, 0.15 mmol), CuI (15 mg, 0.08 mmol) and
(Ph3P)4Pd (23 mg, 0.02 mmol) in (i-Pr)2NH (0.3 mL) was stirred at
room temperature for 17 h. Satd NH4Cl solution was added to
quench the reaction. The mixture was extracted with ether, dried,
concentrated and fractioned by FCC (ethyl acetate/hexane, 60:40 v/
product 18 (17 mg, 81%). ½a D25
ꢀ13 (c 1.4, CHCl3); IR (KBr): 3410,
ꢁ
2960, 1602, 1504 cmꢀ1 1H NMR (C6D6) d: 7.15 (2H, m, C6H5),
;
6.73 (1H, dd, 7.25, 7.25 Hz, C6H5), 6.52 (2H, dd, 1.0, 8.5 Hz, C6H5),
5.47 (1H, ddq, 1.5, 6.5, 6.5 Hz, @CH), 3.80 (2H, m, CH2NH), 3.62
(1H, m, H-3), 2.09 (1H, ddd, 1.5, 5.0, 14.5 Hz, H-2), 1.67 (3H, d,
1.25, CH3), 1.44 (3H, s, CH3), 1.40 (2H, m, H-2 and H-4), 1.31 (3H,
s, CH3), 1.23 (3H, s, CH3), 1.05 (1H, dd, 9.75, 13.0 Hz, H-4); 13C
NMR (C6D6) d: 148.4, 136.5 129.5, 119.8, 117.7, 113.2, 92.9, 84.5,
66.6, 63.6, 45.7, 44.1, 40.0, 34.4, 30.1, 26.2, 22.9, 22.0; HRMS
TOF+ m/z calcd for C21H28NO2: 326.2114, found: 326.2123.
v) to provide compound 28 (18.1 mg, 72%). ½a D25
ꢀ8.0 (c 1.2, CHCl3);
ꢁ
IR (KBr): 3333, 2959, 2923 cmꢀ1; 1H NMR (CDCl3) d: 5.85 (1H, ddq,
1.0, 6.75, 6.75 Hz, @CH), 4.26 (2H, d, 6.75 Hz, @CHCH2), 3.79 (1H,
m, H-3), 2.32 (1H, ddd, 1.75, 5, 14.25 Hz, H-2), 1.84 (3H, d, 1 Hz,
CH3), 1.74 (1H, br s, OH), 1.61 (1H, dd, 8.75, 14.25 Hz, H-2), 1.57
(1H, m, H-4), 1.47 (3H, s,CH3), 1.22 (3H, s, CH3), 1.19 (1H, dd,
10.5, 13.0 Hz, H-2), 1.08 (3H, s, CH3); 13C NMR (C6D6) d: 137.8,
119.1, 92.4, 84.3, 66.6, 63.8, 63.6, 61.5, 45.7, 40.0, 34.4, 30.0,
26.2, 22.9, 22.0; HRMS CI+ m/z calcd for C15H23O3: 251.1647, found:
251.1646.
5.1.10. (2Z,4E)-5-(10-Hydroxy-20,20,60-trimethyl-40-
oxocyclohexyl)-3-methylpenta-2,4-dienyl 2-(thiophen-200-yl)
acetate (16)
To a solution of the allylic alcohol, racemic 20 from Ref. 20,
(34 mg, 0.1 mmol), Et3N (42
lL, 0.3 mmol) in CH2Cl2 (1.5 mL) was
added 2-thiopheneacetyl chloride (18
l
L, 0.15 mmol). The reaction
5.1.7. (1R,3S,6R)-(30Z) -1,5,5-Trimethyl-6-(30-methyl-50-
(phenylthio)-pent-30-en-10-ynyl)-7-oxa-bicyclo[4.1.0]heptan-3-
ol (17)
A solution of alcohol 28 (56.6 mg, 0.23 mmol), phenyl disulfide
(98.9 mg, 0.45 mmol) and n-Bu3P (112 lL, 0.45 mmol) in dry
mixture was stirred at room temperature for 4 h and diluted with
CH2Cl2. The organic phase was washed with satd NaHCO3, dried,
concentrated and fractionated by PTLC (ethyl acetate/hexane,
20:80 v/v) to the ketal protected thiophene ester (13 mg, 28%).
1H NMR (CDCl3) d: 7.19 (1H, d, 1.25 Hz, SCH), 6.93 (2H, m, thio-
phene CH@CH), 6.67 (1H, d, 15.5 Hz, CH@CH), 5.98 (1H, d,
15.5 Hz, CH@CH), 5.47 (1H, t, 7.0 Hz, @CHCH2O), 4.80 (2H, d,
7.0 Hz, CH2O), 3.82 (2H, s, COCH2), 3.58 (2H, dd, 5, 10.25 Hz,
OCH2), 3.41 (2H, dd, 5, 10.25 Hz, OCH2), 2.30 (1H, dd, 2.75,
14.5 Hz, H-3), 2.17 (1H, m, H-60), 1.98 (1H, dd, 3.25, 14.25 Hz, H-
50), 1.86 (3H, s, CH3), 1.40 (1H, d, 14.0 Hz, H-30), 1.35 (1H, d,
14.0 Hz, H-30), 1.12 (3H, s, CH3), 1.06 (3H, s, CH3), 0.85 (3H, s,
CH3), 0.78 (3H, s, CH3), 0.77 (3H, d, 8.0 Hz, CH3). To a solution of
the ketal protected thiophene ester (13 mg, 0.028 mmol) in ace-
tone (1.5 mL) was added 2 N HCl (2 drops). The mixture was stirred
at room temp. for 1 h. After removing acetone, ether was added
and washed with satd NaHCO3, dried and concentrated to give a
residue which was purified by FCC (ethyl acetate/hexane, 20:80
CH2Cl2 (3 mL) was stirred at room temperature for 3 h. Ethanol
(1 mL) was added to the reaction and stirred for 30 min. Ethanol
was evaporated off and more CH2Cl2 added. The organic phase
was washed with 0.5 N NaOH, followed by water and then dried,
concentrated, and fractionated by FCC (ethyl acetate/hexane,
30:70 v/v) to give product 17 (42 mg, 54%). ½a D25
ꢁ
ꢀ16 (c 0.84,
CHCl3); IR (KBr): 3438, 2961, 2924, 1583 cmꢀ1
;
1H NMR (CDCl3)
d: 7.31 (2H, m, C6H5), 7.23 (2H, m, C6H5), 7.14 (1H, m, C6H5), 5.73
(1H, ddq, 1.0, 7.5, 7.5 Hz, @CH), 3.82 (1H, m, H-3), 3.71 (2H, dd,
0.75, 7.5 Hz, CH2S), 2.34 (1H, ddd, 1.75, 5.0, 14.5 Hz, H-2), 1.81
(3H, d, 1.0 Hz, CH3), 1.63 (1H, dd, 8.5, 14.5 Hz, H-2), 1.58 (1H, m,
H-4), 1.47 (3H, s, CH3), 1.23 (3H, s, CH3), 1.21 (1H, m, H-4), 1.10
(3H, s, CH3); 13C NMR (CDCl3) d: 135.9, 132.9, 129.3, 128.9, 126.0,
120.7, 91.7, 84.2, 67.1, 63.8, 63.7, 45.8, 39.8, 34.4, 33.9, 29.9,
25.7, 22.9, 21.7; HRMS EI+ m/z calcd for C21H26O2S: 342.1654,
found: 342.1659.
v/v) to provide 16 (8 mg, 75%). IR (KBr): 3517, 2959, 1714 cmꢀ1
;
1H NMR (CDCl3) d: 7.19 (1H, d, 1.5 Hz, SCH), 6.93 (2H, m, thiophene
CH@CH), 6.80 (1H, d, 15.5 Hz, CH@CH), 6.12 (1H, d, 15.5 Hz,
CH@CH), 5.54 (1H, t, 7.0 Hz, @CHCH2O), 4.81(2H, d, 7.0 Hz,
CH2O), 3.82 (2H, s, COCH2), 2.46 (1H, d, 15.0 Hz, H-30), 2.30 (2H,
m, H-50 and H-60) 2.13 (2H, m, H-50 and H-30), 1.90 (3H, s, CH3),
1.02 (3H, s, CH3), 0.90 (3H, s, CH3), 0.85 (3H, d, 6.5 Hz, CH3); 13C
NMR (C6D6) d: 209.3, 170.4, 136.7, 135.6, 135.0, 129.9, 128.2,
126.8, 125.1, 123.1, 78.1, 61.0, 52.9, 47.1, 41.6, 37.4, 35.4, 25.2,
22.8, 20.9, 15.9; HRMS EI+ m/z calcd for C21H28O4S: 376.1708,
found: 376.1720.
5.1.8. (10R,40S,60R)-(2Z)-5-(40-Hydroxy-20,20,60-trimethyl-70-oxa-
bicyclo[4.1.0]heptan-10-yl)-3-methylpent-2-en-4-ynal (29)
A mixture of alcohol 28 (89 mg, 0.36 mmol) and MnO2 (774 mg,
8.9 mmol) in petroleum ether (10 mL) and ethyl acetate (5 mL) was
stirred at room temperature for 4 h. The reaction mixture was fil-
tered through a pad of Celite 545Ò and washed with ethyl acetate.
The combined filtrates and washings were concentrated and puri-
fied by FCC (ethyl acetate/hexane, 30:70 v/v) to afford aldehyde 29
(73.3 mg, 83%). ½a D25
ꢁ
ꢀ11 (c 3.0, CHCl3); IR (KBr): 3456, 2918, 1601,
5.2. AtNCED In vitro assay substrate preparation
1593 cmꢀ1 1H NMR (C6D6) d: 10.27 (1H, d, 8.0 Hz, CHO), 5.88 1H,
;
dd, 0.75, 8.0 Hz, @CH), 3.55 (1H, m, H-4), 2.03 (1H, ddd, 1.0, 5.0,
15.5 Hz, H-5), 1.46 (3H, d, 0.75 Hz CH3), 1.36 (2H, m, H-3 and H-
5), 1.27 (3H, s, CH3), 1.13 (3H, s, CH3), 1.11 (3H, s, CH3), 0.99 (1H,
dd, 10.0, 13.0 Hz, H-3); 13C NMR (C6D6) d: 190.8, 140.1, 136.3,
98.4, 82.5, 67.0, 63.4, 45.4, 39.8, 34.3, 29.6, 26.1, 24.1, 21.8; HRMS
CI+ m/z calcd for C15H21O3: 249.1491, found: 249.1489.
Fresh spinach was macerated under liquid nitrogen and ex-
tracted five times with three volumes of methanol/0.1% KOH. Sam-
ples were dried using a roto-evaporator, resuspended in acetone
and then chilled on ice for 1 h. The solvent was subsequently trans-
ferred to a new flask, roto-evaporated and resuspended in acetoni-
trile/acetone (1:1) mixture. The mixture was applied to a gravity

J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
2911
flow column containing C-18 silica gel (Sigma) equilibrated in 65%
acetonitrile/35% water (solvent C). The column was washed with
49% acetone (solvent D)/51% solvent A and 20 mL of 55% solvent
D/45% solvent A while collecting 5 mL fractions. Fractions contain-
tural template.25 To model AtNCED3, amino acid alignments were
made between ACO, AtNCED3 and VP14. AtNCED3 shares 25% and
45% amino acid identity and similarity with ACO, and 64% and 76%,
respectively with VP14.37 Highly conserved amino acids including
H183, H238, H304 and H484 forming the octahedral coordination
of the non-heme iron required for catalysis of the dioxygenase
reaction were used to aid in development of a suitable alignment
and ultimately build the homology model. Homology modeling
jobs were submitted to the Swiss-Model servers using the DEEPVIEW
program as an interface.26 Each generation of the AtNCED3 homol-
ogy model was energy minimized within DEEPVIEW using 1000 steps
of steepest descent followed by 1000 steps of conjugate gradient
minimization until the RMS gradient of the potential energy was
less than 0.01 kJ.
ing neoxanthin were pooled, dried, and resuspended in 100
methanol. The pooled mixture was separated using an Agilent
1100 series HPLC and a Supelcosil LC-18 (25 cm ꢂ 10 mm, 5 m)
lL of
l
(Supelco) column equilibrated with solvent A. The HPLC method
consisted of a linear gradient over 30 min from 100% solvent A to
100% solvent D with a flow rate of 4 mL/min at 22 °C and moni-
tored with a PDA detector at 436 nm. The neoxanthin fractions
were collected, dried and resuspended in ethanol. Neoxanthin
was quantified by determining its OD439 using a PerkinElmer
Lambda 35 UV–vis spectrometer and applying its extinction coeffi-
cient of 2243 (A1%1 cm).30
5.5. In silico docking of AtNCED3 active site SLCCD inhibitor
5.3. Recombinant AtNCED3expression, purification and in vitro
assays
interactions
Inhibitor structures were created using CS ChemOffice v9 (Cam-
bridgeSoft). In silico docking of inhibitor structures to the AtNCED3
homology model were performed using AutoDock v3.1 on a Silicon
Graphics Octane2 Workstation.38 Inhibitor structures were docked
within a grid box encompassing the entire catalytic pocket of AtN-
CED3 corresponding to 80 ꢂ 36 ꢂ 30 points using a spacing of
0.375 Å between grid points. The docking parameters consisted
of 20 Lamarckian Genetic Algorithm runs using a population size
of 100 individuals and 1,000,000 energy evaluations. Final docked
structures having orientations less than or equal to 0.5 Å root mean
square deviation were clustered.
AtNCED3 was over-expressed using the pRL296 expression vec-
tor (a gift from M. Cygler, BRI, Montreal) in E. coli (BL21)DE3 cells
as a glutathione-S-transferase fusion protein and affinity purified
using glutathione sepharose 4 fast flow resin (GE Healthcare) as
described previously.35 Essentially, cells were grown to an OD600
of 0.45 at 37 °C and 200 rpm shaking. The culture was induced
with 1 mM isopropyl-b- -thiogalactoside for 16 h at 15 °C and
D
200 rpm shaking. The cells were pelleted and resuspended in
50 mM Tris–HCl (pH 8.0) 1 mM DTT and 0.5% protease inhibitor
cocktail set III (CalBiochem). Cells were lysed using a french press
at 20,000 psi and affinity purified as per manufacturer0s instruc-
tions (GE Healthcare). Protein concentration was determined by
the method of Bradford.36
5.6. In vivo application of SLCCD inhibitors to A. thaliana Col-0
Enzymatic assays contained 100 mM bis–tris (pH 6.7), 5
FeSO4, 10 mM ascorbate, 0.05% Triton X-100, catalase (1 mg/mL),
neoxanthin and inhibitor to a total volume of 5 L of ethanol and
g AtNCED3 to a total assay volume of 100 L. Assays were incu-
bated at 22 °C for 20 min. The assays were stopped with the addi-
tion of 50 L of 25% Triton X-100 and extracted with 150 L of
ethyl acetate. All procedures were performed under red-light to
minimize photo-induced damage to assay components and prod-
ucts.6 Fine chemicals and solvents were purchased from Sigma–Al-
l
M
For each condition to be tested, three hundred wild-type A. tha-
liana Col-0 seeds (LEHLE) were sterilized, stratified and sewn onto
200 mL of Sunshine Mix #3 (Sun Gro) potting material in an
8 ꢂ 8 ꢂ 4 cm pot. Plants were watered continuously with 25 g/
100 mL of 20–20–20 (PlantProd) fertilizer and grown at 22 °C with
a 16 h photoperiod for 22 days. Plants were pre-treated with
50 mL/pot of Buffer A (10 mM HEPES pH 6.5) (Sigma) 10 or
l
8
l
l
l
l
33
50 mL/pot of Buffer A containing 0.4 M mannitol (Sigma) 10 or
33 M test compound. Non-treated/non-stressed control plants
lM test compound for 2 h. Plants were then soaked with
drich. 75
series HPLC machine equipped with
(3.3 cm ꢂ 4.6 mm, 3 m) (Supelco) column pre-equilibrated with
lL of the assay extract was injected into an Agilent 1100
l
a
Supelcosil LC-18
were simply soaked in Buffer A at the designated time points. Aer-
ial plant tissue was harvested after 6, 12 and 48 h from the time of
initial inhibitor treatment and flash frozen in liquid nitrogen. Half
of the tissue samples were lyophilized for metabolite profiling and
the other half taken for quantitative reverse-transcription poly-
merase chain reaction (qRT-PCR) analysis. While analytical exper-
iments were carried out on pooled samples (ꢃ300 plants/
condition), experiments were not replicated due to limitations in
the availability of inhibitor compounds for treatments at high
concentrations.
l
15% acetonitrile (solvent B)/85% water (solvent A). Solvent B in-
creased to 35% over 10 min, followed by a linear gradient of 65%
solvent B to 100% solvent D over 10 min. Solvent D was maintained
at 100% for 2 min and then the column was returned to 15% solvent
B for 5 min. The flow rate was maintained at 1.5 mL/min. and mon-
itored with a photodiode array (PDA) detector at 436 and 262 nm.
Preliminary in vitro screening experiments of all selected po-
tential inhibitors were not reproduced due to limitations in the
availability of some of the compounds. Instead the top three per-
forming inhibitors were selected for subsequent kinetic evaluation.
Evaluation of recombinant AtNCED3 kinetic parameters for Km was
accomplished using Michaelis–Menten equation plotted with Enz-
Fitter v2.0.18.0 (Biosoft). The Ki for inhibitors was determined
using a Dixon plot and concentration ranges of 250, 200, 150,
5.7. Metabolite profiling of A. thaliana hormone levels
Freeze-dried tissue was homogenized using a multi-tube ball
mill (Mini-BeadBeater-96, Biospec Products Inc., Bartlesville, Okla-
homa, USA) and 50 mg of each sample was weighed out into indi-
100, 50 and 0
l
M inhibitor in the presence of either 55, 30 or
vidual Falcon tubes. To each sample, 100 lL of a cocktail of internal
10
l
M 9-cis-neoxanthin 2.5
standards comprised of (ꢀ)-5,80,80,80-d4-ABA, (ꢀ)-70,70,70-d3-PA,
(ꢀ)-5,80,80,80-d4-70OH ABA, (ꢀ)-70,70,70-d3-DPA and (+)-4,5,80,80,80-
5.4. Homology modeling of AtNCED3
d5-ABAGE, each at a concentration of 0.2 ng/lL and dissolved in
a mixture of water/acetonitrile (1:1, v/v) with 0.5% glacial acetic
acid, was added. Further, 3 mL of isopropanol/water/glacial acetic
acid (80:19:1, v/v/v) extraction solvent was added, and samples
were placed in the fridge (4 °C, in the dark) on an orbital shaker
A homology model of AtNCED3 was built using the X-ray crystal
structure of Synechocystis sp. PCC 6803 ACO (pdb code: 2biw; avail-
able at the RCSB Protein Data Bank) at 2.39 Å resolution as a struc-

2912
J. Boyd et al. / Bioorg. Med. Chem. 17 (2009) 2902–2912
at ꢃ350 rpm. After 18–24 h, the samples were centrifuged at
Acknowledgments
4.4 krpm for 10 min, the supernatant was transferred to a dispos-
able culture tube, and a second portion of 500 lL extraction sol-
This work was supported by the National Research Council of
Canada. We would like to thank Tadao Asami for generously sup-
plying us with quantities of ABM and ABM-SG used as controls in
our study. This manuscript represents NRCC# 50133.
vent mixture was added to wash the pellet. After vortexing and
centrifuging again at 4.4 krpm for 10 min, each wash was com-
bined with its appropriate supernatant. The organic extract was
dried under reduced pressure, then re-dissolved in 100
lL metha-
Supplementary data
nol/glacial acetic acid (99:1, v/v) followed by 900 L of aqueous 1%
l
glacial acetic acid. This mixture was extracted with 2 mL hexane,
and then the aqueous layer was dried down under reduced pres-
sure. The sample was further reconstituted in 2 mL aqueous 1% gla-
cial acetic acid and loaded onto an Oasis MCX SPE cartridge (3 cc,
Waters Corporation, Mississauga, Ontario, Canada). After a wash
with 3 mL aqueous 1% glacial acetic acid, samples were eluted with
1 mL methanol/glacial acetic acid (99:1, v/v) and then dried down
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2009.01.076.
References and notes
1. Zeevaart, J. A.; Creelman, R. A. Annu. Rev. Plant Physiol. Plant Mol. Biol. 1988, 39,
439.
2. McCarty, D. R. Annu. Rev. Plant Physiol. Plant Mol. Biol. 1995, 46, 71.
3. Nambara, E.; Marion-Poll, A. Annu. Rev. Plant Biol. 2005, 56, 165.
4. Milborrow, B. V. J. Exp. Bot. 2001, 52, 1145.
5. Kushiro, T.; Okamoto, M.; Nakabayashi, K.; Yamagishi, K.; Kitamura, S.; Asami,
T.; Hirai, N.; Koshiba, T.; Kamiya, Y.; Nambara, E. EMBO J. 2004, 23, 1647.
6. Schwartz, S. H.; Tan, B. C.; Gage, D. A.; Zeevaart, J. A. D.; McCarty, D. R. Science
1997, 276, 1872.
7. Iuchi, S.; Kobayashi, M.; Taji, T.; Naramoto, M.; Seki, M.; Kato, T.; Tabata, S.;
Kakubari, Y.; Yamaguchi-Shinozaki, K.; Shinozaki, K. Plant J. 2001, 27, 325.
8. Qin, X.; Zeevaart, J. A. D. PNAS 1999, 96, 15354.
9. Burbidge, A.; Grieve, T. M.; Jackson, A.; Thompson, A.; McCarty, D. R.; Taylor, I.
B. Plant J. 1999, 17, 427.
10. Chernys, J. T.; Zeevaart, J. A. D. Plant Physiol. 2000, 124, 343.
11. Iuchi, S.; Kobayashi, M.; Yamaguchi-Shinozaki, K.; Shinozaki, K. Plant Physiol.
2000, 123, 553.
under reduced pressure. The extract was re-dissolved in 100
lL
methanol/glacial acetic acid (99:1, v/v) followed by 900 L of aque-
l
ous 1% glacial acetic acid. This mixture was further cleaned on an
Oasis HLB SPE cartridge (1 cc, Waters Corporation, Mississauga,
Ontario, Canada). After a wash with 1 mL aqueous 1% glacial acetic
acid, the fraction containing ABA and ABA metabolites was eluted
with 1 mL acetonitrile/water/glacial acetic acid (30:69:1, v/v/v)
and then was evaporated to dryness. The final residue was dis-
solved in 200
0.1% glacial acetic acid and 100 pg/
l
L of acetonitrile/water (15:85, v/v) containing
l
L ( )-30,50,50,70,70,70-d6-ABA
as a recovery standard. Finally, the sample was subjected to LC–
ES–MS/MS analysis and quantification, as described in Owen and
Abrams, 2008.39
12. Tan, B.-C.; Joseph, L. M.; Deng, W.-T.; Liu, L.; Li, Q.-B.; Cline, K.; McCarty, D. R.
Plant J. 2003, 35, 44.
13. Endo, A.; Sawada, Y.; Takahashi, H.; Okamoto, M.; Ikegami, K.; Koiwai, H.; Seo,
M.; Toyomasu, T.; Mitsuhashi, W.; Shinozaki, K.; Nakazono, M.; Kamiya, Y.;
Koshiba, T.; Nambara, E. Plant Physiol. 2008, 12. pp. 108.116632.
14. Nagamune, K.; Hicks, L. M.; Fux, B.; Brossier, F.; Chini, E. N.; Sibley, L. D. Nature
2008, 451, 207.
15. Toh, S.; Imamura, A.; Watanabe, A.; Nakabayashi, K.; Okamoto, M.; Jikumaru, Y.;
Hanada, A.; Aso, Y.; Ishiyama, K.; Tamura, N.; Iuchi, S.; Kobayashi, M.; Yamaguchi,
S.; Kamiya, Y.; Nambara, E.; Kawakami, N. Plant Physiol. 2008, 146, 1368.
16. Creelman, R. A.; Bell, E.; Mullet, J. E. Plant Physiol. 1992, 99, 1258.
17. Kitahata, N.; Han, S.-Y.; Noji, N.; Saito, T.; Kobayashi, M.; Nakano, T.; Kuchitsu,
K.; Shinozaki, K.; Yoshida, S.; Matsumoto, S. Bioorg. Med. Chem. 2006, 14, 5555.
18. Han, S.-Y.; Kitahata, N.; Sekimata, K.; Saito, T.; Kobayashi, M.; Nakashima, K.;
Yamaguchi-Shinozaki, K.; Shinozaki, K.; Yoshida, S.; Asami, T. Plant Physiol.
2004, 135, 1574.
19. Baumeler, A.; Brade, W.; Haag, A.; Eugster, C. H. Helv. Chim. Acta 1990, 73, 700.
20. Lamb, N.; Abrams, S. R. Can. J. Chem. 1990, 68, 1151.
21. Isler, O.; Lindlar, H.; Montavon, M.; Rüegg, R.; Saucy, G.; Zeller, P. Helv. Chim.
Acta 1956, 39, 2041.
22. Nakagawa, I.; Hata, T. Tetrahedron Lett. 1975, 17, 1409.
23. Furuichi, N.; Hara, H.; Osaki, T.; Mori, H.; Katsumura, S. Angew. Chem., Int. Ed.
2002, 41, 1023.
5.8. Seed germination assay
A. thaliana Col-0 seeds were sterilized by washing them with
10% sodium hypochlorite and 20% sodium dodecyl sulfate (Sigma)
for 5 min and then rinsing four times with sterile water. Seeds
were moist chilled for 4 days and then plated on germination med-
ium (0.41% MS salts, 1% sucrose, 0.05% MES and 0.1% Gamborg’s
vitamins, pH 5.7, 0.7% agar) (Sigma) containing either 0.1, 0.33,
1.0 or 3.33 lM of inhibitor or (+)-ABA 1. As a control, seeds were
sewn and germinated on media only without inhibitors or (+)-
ABA 1. Germination was recorded over seven days and indexes cal-
culated as described previously.40 Experiments were not replicated
due to limitations in the availability of inhibitor compounds for
treatments at high concentrations.
24. Schwartz, S. H.; Tan, B. C.; McCarty, D. R.; Welch, W.; Zeevaart, J. A. D. Biochim.
Biophys. Acta (BBA)—General Subj. 2003, 1619, 9.
5.9. Quantitative reverse-transcription PCR (qRT-PCR)
25. Kloer, D. P.; Ruch, S.; Al-Babili, S.; Beyer, P.; Schulz, G. E. Science 2005, 308, 267.
26. Schwede, T.; Kopp, J.; Guex, N.; Peitsch, M. C. Nucleic Acids Res. 2003, 31, 3381.
27. Creelman, R. A.; Zeevaart, J. A. D. Plant Physiol. 1985, 77, 25.
28. Huang, D.; Wu, W.; Abrams, S. R.; Cutler, A. J. J. Exp. Bot. 2008, 59, 2991.
29. Cutler, A. J.; Rose, P. A.; Squires, T. M.; Loewen, M. K.; Shaw, A. C.; Quail, J. W.;
Krochko, J. E.; Abrams, S. R. Biochemistry 2000, 39, 13614.
30. Yamaguchi-Shinozaki, K.; Shinozaki, K. Plant Physiol. 1993, 101, 1119.
31. Norris, S. R.; Meyer, S. E.; Callis, J. Plant Mol. Biol. 1993, 21, 895.
32. Han, S.-y.; Inoue, H.; Terada, T.; Kamoda, S.; Saburi, Y.; Sekimata, K.; Saito, T.;
Kobayashi, M.; Shinozaki, K.; Yoshida, S.; Asami, T. Bioorg. Med. Chem. Lett.
2002, 12, 1139.
33. Abe, M.; Matsuki, H.; Domae, M.; Kuwata, H.; Kudo, I.; Nakanishi, Y.; Hara, N.;
Mitsuyama, T.; Furukawa, T. Am. J. Respir. Cell Mol. Biol. 1996, 15, 565.
34. Wilen, R. W.; Hays, D. B.; Mandel, R. M.; Abrams, S. R.; Moloney, M. M. Plant
Physiol. 1993, 101, 469.
35. Guo, S.; Boyd, J.; Sammynaiken, R.; Loewen, M. C. Biochemistry and Cell Biology
2008, 86, 262.
Frozen plant material (250 mg) was ground under liquid nitrogen
and extracted for mRNA as suggested by the manufacturer (PolyA-
Tract System 1000, Promega). The resulting mRNA was quantified
and checked for quality using a Nano-Drop ND-1000 Spectropho-
tometer. QuantiTect Reverse Transcription Kit (Qiagen) was used
to produce cDNA as directed by the manufacturer from 20 ng of
starting mRNA. Quantitative PCR was performed on 1 lL of cDNA
product using a Bio-Rad iCycler and the QuantiTect SYBR Green
PCR Kit (Qiagen) coupled with QuantiTect Primer Assays (Qiagen)
for the gene targets; AtNCED3 (NM_112304), Rd29B
(NM_124609), CYP707A1 (NM_118043), CYP707A3 (NM_123902)
and UBQ10 (NM_178968). The pre-validated primer sets are as fol-
lows indicated by the GeneGlobe (https://www1.qiagen.com/Gene-
Globe/default.aspx) product name and (catalogue number):
At_NCED3_1_SG (QT00769573), At_RD29B_1_SG (QT00840399),
At_CYP707A1_1_SG (QT00808339), At_CYP707A3_1_SG (QT007
39242), At_UBQ10_va.1_SG (QT01123745). Relative changes in
transcript level were normalized using UBQ10 and quantified as pre-
viously described.41
36. Bradford, M. M. Anal. Biochem. 1976, 72, 248.
37. Nicolas Guex, M. C. P. Electrophoresis 1997, 18, 2714.
38. Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.; Belew, R. K.;
Olson, A. J. J. Comput. Chem. 1998, 19, 1639.
39. Owen, S. J.; Abrams, S. R. In Plant Hormones: Methods and Protocols, 2nd Ed.,
Cutler, S., Bonetta, D., Eds.; Humana Press, A Part of Springer Science + Business
Media, 2009, pp. 39–51.
40. Walker-Simmons, M. K. Plant, Cell Environ. 1988, 11, 769.
41. Livak, K. J.; Schmittgen, T. D. Methods 2001, 25, 402.