Chemical screening and development of novel gibberellin mimics

Article abstract of DOI:10.1016/j.bmcl.2017.07.012

Gibberellin (GA) plays versatile roles in the regulation of plant growth and development and therefore is widely used as a regulator in agriculture. We performed a chemical library screening and identified a chemical, named 67D, as a stimulator of seed germination that was suppressed by paclobutrazol (PAC), a GA biosynthesis inhibitor. In vitro binding assays indicated that 67D binds to the GID1 receptor. Further studies on the structure–activity relationship identified a chemical, named chemical 6, that strongly promoted seed germination suppressed by PAC. Chemical 6 was further confirmed to promote the degradation of RGA (for repressor of ga1-3), a DELLA protein, and suppress the expression levels of GA3ox1 in the same manner as GA does. 67D and its analogs are supposed to be agonists of GID1 and are expected to be utilized in agriculture and basic research as an alternative to GA.

Full text of DOI:10.1016/j.bmcl.2017.07.012

Accepted Manuscript
Chemical screening and development of novel gibberellin mimics
Kai Jiang, Hiroaki Shimotakahara, Ming Luo, Masato Otani, Hidemitsu
Nakamura, Said Salama Moselhy, Khalid Omer Abualnaja, Abdulrahman
Labeed Al-Malki, Taha Abduallah Kumosani, Nobutaka Kitahata, Takeshi
Nakano, Masatoshi Nakajima, Tadao Asami
PII:
S0960-894X(17)30708-4
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.07.012
BMCL 25125
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
6 June 2017
29 June 2017
3 July 2017
Please cite this article as: Jiang, K., Shimotakahara, H., Luo, M., Otani, M., Nakamura, H., Moselhy, S.S., Abualnaja,
K.O., Al-Malki, A.L., Kumosani, T.A., Kitahata, N., Nakano, T., Nakajima, M., Asami, T., Chemical screening and
development of novel gibberellin mimics, Bioorganic & Medicinal Chemistry Letters (2017), doi: http://dx.doi.org/
10.1016/j.bmcl.2017.07.012
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Chemical screening and development of novel gibberellin mimics
Kai Jianga,i, Hiroaki Shimotakaharaa,i, Ming Luob,i, Masato Otania, Hidemitsu Nakamuraa, Said Salama
Moselhyc,d,e,f, Khalid Omer Abualnajac,d, Abdulrahman Labeed Al-Malkic,d,e, Taha Abduallah
Kumosanic,e,g, Nobutaka Kitahataa, Takeshi Nakanoh, Masatoshi Nakajimaa, Tadao Asamia,c,d
*
aDepartment of Applied Biological Chemistry, University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo
113-8657
bKey Laboratory of Biotechnology and Crop Quality Improvement of Ministry of Agriculture,
Biotechnology Research Center, Southwest University, Chongqing 400716, China
cBiochemistry Department, Faculty of Science, King Abdulaziz University, Jeddah, Saudi Arabia
dBioactive Natural Products Research Group, King Abdulaziz University, Jeddah, Saudi Arabia
eExperimental Biochemistry Unit, King Fahd Medical Research Center, King Abdulaziz University,
Jeddah, Saudi Arabia
fBiochemistry Department, Faculty of Science, Ain Shams University, Cairo, Egypt
gProduction of Bioproducts for Industrial Applications Research Group, King Abdulaziz University,
Jeddah, Saudi Arabia
hGene Discovery Research Group, Center for Sustainable Resource Science, RIKEN, Wako, Saitama
351-0198, Japan.
iThese authors contributed equally to this work.
*Corresponding author. Tel.: +81 3 5841 5192; Fax: +81 3 5841 8025.
E-mail address: asami@mail.ecc.u-tokyo.ac.jp (T. Asami).

Abstract
Gibberellin (GA) plays versatile roles in the regulation of plant growth and development and therefore
is widely used as a regulator in agriculture. We performed a chemical library screening and identified a
chemical, named 67D, as a stimulator of seed germination that was suppressed by paclobutrazol (PAC),
a GA biosynthesis inhibitor. In vitro binding assays indicated that 67D binds to the GID1 receptor.
Further studies on the structure–activity relationship identified a chemical, named chemical 6, that
strongly promoted seed germination suppressed by PAC. Chemical 6 was further confirmed to promote
the degradation of RGA (for repressor of ga1-3), a DELLA protein, and suppress the expression levels
of GA20ox2 and GA3ox1 in the same manner as GA does. 67D and its analogs are supposed to be
agonists of GID1 and are expected to be utilized in agriculture and basic research as an alternative to
GA.
Keywords: chemical screening, gibberellin mimic, plant regulator

Gibberellin (GA), a tetracyclic diterpenoid carboxylic acid, is an important plant hormone that
regulates seed germination, stem elongation, leaf expansion and flower development.1 In higher plants,
GA is synthesized from geranylgeranyl diphosphate (GGDP) through a series of enzymes, including
ent-copalyl diphosphate synthase (CPS) and ent-kaurene synthase, which convert GGDP to ent-
kaurene; ent-kaurene oxidase and ent-kaurenoic acid oxidase, which oxidize ent-kaurene to GA12; and
GA20-oxidase (GA20ox) and GA3-oxidase (GA3ox), which catalyze the conversion of GA12 to
bioactive GAs, such as GA1 or GA4.1 The GA signaling pathway has been established since the GA
receptor GID1 was first identified in rice and Arabidopsis.2 GID1 proteins are mainly present in the
nucleus but are also located in the cytoplasm.2,3 Upon binding to bioactive GA, GID1 undergoes
conformational changes in its lid region and binds to a DELLA protein, which is a repressor protein in
GA signaling. After the formation of a GID1-GA-DELLA complex, DELLA is polyubiquitinated by an
SCF (SKP1, CULLIN, F-BOX) E3 ubiquitin-ligase complex and subsequently degraded by the 26S
proteasome, which then induces GA responses, such as the expression of GA biosynthesis enzymes
(GA20ox2 and GA3ox1) via feedback regulation pathways to maintain GA homeostasis.4 Although
more than a hundred GAs have been identified, only a few of them, such as GA1, GA3, GA4 and GA7,
display bioactivity. GA3 can be produced by microbes with a relatively low cost of production
compared with that of other bioactive GAs. However, the bioactivity spectrum of GA3 is narrow. For
example, GA4 but not GA3 is assumed to increase the fruit size of ‘Kousui’ Japanese pear.5 For better
use of the GA signal in both agriculture and basic research, GA mimics that can be artificially
synthesized at a relatively low production cost are in urgent need. A series of N-substituted
phthalimides have been reported to display GA activities.6 The chemical AC94377 was determined to
share the same signaling pathway as GA, even though it has a completely different structure as GA,7

which indicates that synthetic chemicals that mimic GA can be developed.
In this study, to discover novel GA mimics, we performed a chemical library screening using the
model plant Arabidopsis. GA plays a vital role in seed germination of Arabidopsis. The seeds fail to
germinate under treatment with paclobutrazol (PAC), a GA biosynthesis inhibitor that inhibits the
oxidation of ent-kaurene to ent-kaurenoic acid.8 Using 96-well plates, we screened compounds from
Riken Natural Products Depository (NPDePO) that can rescue seed germination inhibited by PAC (Fig.
1a). A chemical ((S)-3-phenyl-2-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido) propanoic
acid) named 67D was confirmed to rescue seed germination inhibited by PAC in a concentration-
dependent manner (Fig. 1b). Prohexadione-calcium (PHX) functions on different targets of PAC in GA
biosynthesis and inhibits the conversion of GA12 to bioactive GAs. Thus, PHX inhibits the seed
germination of Arabidopsis. Seeds co-treated with 67D showed resistance to PHX regarding seed
germination (Supplementary Fig. 1). ga1-3, a mutant that lacks the function to convert GGDP to
copalyl pyrophosphate, cannot germinate by itself due to defects in GA biosynthesis.9 Treatment with
67D rescued the seed germination of ga1-3 (Fig. 1c). These results indicate that 67D is a GA mimic
but not a GA biosynthesis regulator in promoting the seed germination of Arabidopsis.
To investigate whether 67D shares the same signaling pathway as GA, we first performed a GID1-
binding assay in accordance with previous reports.7,10 In brief, recombinant GID1 protein, which was
produced by Rosetta (DE3), was first incubated with radiolabeled GA (hot GA) that can bind to GID1.
If the test compound such as GA4 competes with hot GA to bind to the GID1 receptor, the radioactivity
of GID1 fraction will be reduced in a competent compound-concentration-dependent manner. The
results in Figure 2 show that 67D reduced the radioactivity of GID1b in the same manner as GA4 did,
which indicates that 67D can bind to the GID1 receptor. The automated docking of 67D to

Figure 1. Chemical screening and evaluation of the GA-like activity of 67D. (a) Illustration of
chemical screening in a 96-well plate. Thirty millimolar chemicals from the library were added to 1/2
Murashige and Skoog medium (MS) containing 3 µM PAC. The chemicals were ultimately diluted 100
times to a concentration of 30 µM. Wild-type (Col-0) seeds were germinated on 1/2 MS medium
containing dimethyl sulfoxide (DMSO) as a mock treatment, PAC and GA as positive controls, or PAC
and chemicals from the library for screening. The germinated seeds after incubation for 7 days in
darkness were counted and the ratios determined. Chemicals that promote seed germination were
considered candidates. (b) 67D rescued the seed germination of Col-0 that was suppressed by PAC.
Col-0 seeds were germinated on 1/2 MS medium containing 3 µM 67D and increasing concentrations
of PAC, as indicated. Seed germination rate was determined after incubation for 7 days in darkness.
Error bars indicate standard error (n = 3). (c) 67D promoted the seed germination of ga1-3, a GA
biosynthesis mutant. ga1-3 seeds were germinated on 1/2 MS media containing DMSO as a mock
treatment, 10 nM GA4 as a positive control, or 1 µM or 10 µM 67D. The germination rate was
determined after incubation for 10 days in darkness. Afterward, the germinated seeds were counted and
the ratios determined. The mock sample was treated with DMSO. Twenty-five seeds of each sample
were used in each experiment. Error bars indicate standard error (n = 3). **P < 0.01 (two-tailed
Student’s ttest). The mock control and each compound treatment (GA4 or 67D) were compared.

Figure 2. GID1 binding assay. Recombinant GID1b protein was produced by the Escherichia coli
Rosetta (DE3) Novagen strain. GID1b was incubated with 850,000 dpm/µL [1,2,16,17-3H4]-16,17-
dihydro-GA (hot GA); unlabeled GA4 or 67D was then added. Due to competition with hot GA, the
amount of radiation in the GID1b fraction was reduced by incubation with GA4. Incubation with 67D
reduced the amount of radiation in the GID1b fraction in a concentration-dependent manner. Error bars
indicate standard error (n = 3). **P < 0.01 (two-tailed Student’s ttest). The mock control and each
compound treatment (GA4 or 67D) were compared.

GID1a (PDB: 2ZSI) indicated that the anthracene moiety of 67D occupies the space of deep pocket of
GID1a. The remaining moiety of 67D resides in the same position as GA4, which is supposed to be
important for its GA mimic activities (Supplementary Fig. 2).
To develop chemicals with stronger GA-like activities than 67D has, we designed and synthesized
11 analogs of 67D based on the docking results (Fig. 3 and Supplementary note), and we then
evaluated the analogs using both in vitro and in vivo experiments. In vitro binding affinity assays
showed that changes in the benzyl group of 67D affect its affinity for GID1 (Fig. 4a). The replacement
of the benzyl group with hydrogen (chemical 2) or a methyl group (chemical 3) impaired their binding
activities for GID1. Chemicals 4, 5, 6, 7, 8, 9 and 11, which have relatively longer carbon chains than
do chemicals 2 and 3, can bind to GID1 and show the same or even stronger binding activity to GID1
(Fig. 4a). Esterification at the carboxylic acid group (chemical 12) completely impaired binding
activity to GID1, which indicates the importance of the carboxylic acid for binding to GID1.
Compared with chemical 8, which binds to GID1, chemical 10, which has a modified carbon chain due
to the introduction of a methylthiol, lost binding activity to GID1. Furthermore, we evaluated these
synthesized chemicals by examining their effects on germination rates under treatment with PAC. In
accordance with the results of the in vitro binding assays, chemicals 4, 5, 6, 7 and 8 showed greater
effects than did 67D in an in vivo experimental system. The compounds (chemicals 2, 3 and 12) that
showed no in vitro binding activity to GID1 also displayed no GA-like activities in this experimental
system (Fig. 4b). However, chemicals 4, 9 and 11 did not mimic the activity of 67D in promoting seed
germination, despite displaying a similar affinity as 67D for GID1. In fact, upon binding to GID1, GA
induces conformational changes in GID1 and promotes the interaction of GID1 with DELLA, both of
which are important for activating GA signals. Thus, it is possible that chemicals 4, 9 and 11 can bind

Figure 3. Synthesis of 67D and its derivatives. (a) Scheme of the synthesis of 67D and its derivatives
via two-step methods. (b) Chemical structures of 67D (1) and its derivatives synthesized in this study.

Figure 4. Evaluation of 67D and its derivatives via in vitro and in vivo experiments. (a) In vitro GID1
3
binding assay. Recombinant GID1b that was pre-incubated with H-GA (hot GA) was further co-
incubated with GA4 or 0.3 mM test chemicals (67D and chemicals 2–12). The relative amount of
radiation was detected. Error bars indicate standard error (n = 3). *P < 0.05, **P < 0.01, and ***P <
0.001 (two-tailed Student’s ttest). The mock control and each compound treatment were compared. (b)
Rescue of seed germination rate that was suppressed by PAC. Col-0 seeds were germinated on 1/2 MS
containing 1 µM PAC and chemicals at a concentration of 10 µM as indicated. Seed germination rate
was determined after incubation for 7 days in darkness. Error bars indicate standard error (n = 3). *P <
0.05, **P < 0.01, and ***P < 0.001 (two-tailed Student’s ttest). The mock control and each compound
treatment were compared.

to GID1 but do not induce proper conformational changes of GID1. Surprisingly, chemical 10 did not
show binding activity to GID1 but did promote seed germination. In Arabidopsis, there are three GA
receptors: GID1a, GID1b and GID1c.3 Here, we performed a binding activity assay using one of the
three receptors (GID1b). Griffiths et al. investigated the absolute transcript levels of GID1a, GID1b
and GID1c.11 They found that in 24 h imbibed seeds, GID1a shows the highest transcript level, GID1b
shows relative low transcript level and GID1c shows the lowest transcript level. These data suggests
the redundancy among three GID1s in regulating the seed germination. In a similar manner as the
GID1 agonist AC94377, we supposed chemical 10 might have selectivity for other GID1s rather than
GID1b in promoting seed germination. These results suggest that binding to the GID1 receptor is
necessary but not adequate for 67D analogs to mimic GA activity in vivo. Finally, chemical 6, which
has the same affinity as 67D for GID1 and the highest in vivo activity among 67D and its analogs, was
determined to be the best candidate in this study. Chemical 6 was then chosen to further investigate its
mode of action. To determine whether chemical 6 triggers the degradation of DELLA, a pRGA:GFP-
RGA transgenic line was used. In this line, RGA, which is an Arabidopsis DELLA protein, is fused
with green fluorescence protein (GFP) as a reporter, which allows the detection of RGA degradation by
observing the fluorescence of GFP.12 The roots of pRGA:GFP-RGA were soaked in 10 µM GA4
solution for 4 h or 30 µM 67D solution for 7 h and were observed using a confocal laser scanning
microscope (Zeiss LSM700, , Carl Zeiss Inc., Thornwood, NJ). The results in Figure 5a show that
treatment with 67D induces the disappearance of fluorescence in the roots in the same manner as GA
did, which indicates 67D promotes the degradation of the DELLA protein. After the degradation of the
DELLA protein, GA responses that were suppressed by DELLA are induced. For example, the
transcript levels of the GA biosynthesis genes AtGA20ox2 (At5g51810) and AtGA3ox1 (At1g5550) are

Figure 5. GA-like activities of chemical 6. (a) Chemical 6 triggered the degradation of GFP-RGA.
pRGA:GFP-RGA seedlings 7 days after germination were soaked with DMSO as a mock treatment,
GA4 as a positive control, or chemical 6 for several hours. Green fluorescence in primary roots was
observed using confocal scanning microscopy. Bar = 50 µm. (b) Transcript levels of the GA-
responsive gene GA3ox1 in response to chemical treatments. Ten-day-old Col-0 seedlings were treated
with DMSO as a mock treatment, 1 µM GA4 as a positive control, or chemical 6 for 3 h. RNAs were
extracted and converted to cDNA using a total RNA extraction mini kit (RBC Bioscience, Taipei,
Taiwan) and the PrimeScript RT Reagent Kit (Takara Bio, Inc., Shiga, Japan). The transcript levels
of GA3ox1 were normalized to those of Actin 7 (Act7). Primer information is available in
Supplementary Figure 2. Error bars indicate standard error (n = 3). *P < 0.05 (two-tailed Student’s
ttest). The mock control and each compound treatment were compared.

down-regulated by GA in a feedback manner.4 We then investigated the effects of 67D on the gene
expression levels of AtGA20ox2 and AtGA3ox1. Total RNAs were extracted from Arabidopsis
seedlings incubated for 3 h after transplantation onto 1/2 Murashige and Skoog (MS) medium that
contained 67D (1 μM and 10 μM), and the expression analysis was performed using Real-Time RT-
PCR and specific primers of AtGA20ox2 and AtGA3ox1 (Supplementary Fig. 3).13,14 The results
showed that 67D suppresses the transcript levels of these genes in the same manner as did GA (Fig.
5b).
At the present, various GA biosynthesis inhibitors such as PAC and prohexadione have been
reported. However, reports on GA mimics, which are also useful regulators, are rare. The commercial
chemical AC94377 displays GA-like activities.6 A recent report confirmed that AC94377 is an agonist
of GID1 that shares a signaling pathway with GA.7 AC94377 also has selectivity specific to the GID1
receptor; therefore, AC94377 is expected to have use as a specific plant growth regulator and as a
prospective synthetic GA mimic. In this study, we performed a chemical screening to identify novel
GA mimics by focusing on seed germination. By screening 9,600 compounds, we identified a chemical,
67D, that rescued both seed germination suppressed by PAC and the seed germination of ga1-3, which
indicates that 67D is a functional GA mimic. Subsequent structure–activity relationship studies suggest
that the carboxylic acid group is essential for the GA-like activity of 67D. In contrast, the benzyl group
can be replaced with a carbon chain without impairing binding activity to GID1, even increasing it.
Finally, chemical 6 was a better candidate by showing higher activity than 67D did in promoting seed
germination suppressed by PAC. Further studies indicated that chemical 6 triggered DELLA
degradation and down-regulated the expression levels of GA20ox2 and GA3ox1 in the same manner as
did GA, indicating that chemical 6 is an agonist of GID1 that shares the same signaling pathway as GA.

As a potential GID1 agonist, 67D and its derivatives could be leading chemicals for the development
of GA mimics. With a simple two-step synthesis procedure and low production cost, 67D and its
derivatives are expected to be applied to both basic research and agriculture as growth regulators.

Acknowledgments
This study was funded by a grant from the Core Research for Evolutional Science and Technology
(CREST) Program of Japan Science and Technology Agency (JST), and by a grant from the Deanship
of Scientific Research (DSR) at King Abdulaziz University, Jeddah, Saudi Arabia, under Grant No. 1-
130-36-HiCi. The authors thank the DSR for technical and financial support.

References
1.
Yamaguchi S. Gibberellin metabolism and its regulation. Annu Rev Plant Biol. 2008;59:225–
251.
2.
(a) Ueguchi-Tanaka M, Ashikari M, Nakajima M, Itoh H, Katoh E, Kobayashi M, Chow TY,
Hsing YI, Kitano H, Yamaguchi I, Matsuoka M. GIBBERELLIN INSENSITIVE DWARF1
encodes a soluble receptor for gibberellin. Nature. 2005;437:693–698.
(b) Nakajima M, Shimada A, Takashi Y, Kim Y-C, Park S-H, Ueguchi-Tanaka M, Suzuki H,
Katoh E, Iuchi S, Kobayashi M, Maeda T, Matsuoka M, Yamaguchi I. Identification and
characterization of Arabidopsis gibberellin receptors. Plant J. 2006;46:880–889.
Hauvermale AL, Ariizumi T, Steber CM. Gibberellin signaling: a theme and variations on
DELLA repression. Plant Physiol. 2012;160:83–92.
3.
4.
Phillips AL, Ward DA, Uknes S, Appleford NE, Lange T, Huttly AK, Gaskin P, Graebe JE,
Hedden P. Isolation and expression of three gibberellin 20-oxidase cDNA clones from
Arabidopsis. Plant Physiol. 1995;108:1049–1057.
5.
6.
Ito A, Sakamoto D, Itai A, Nishijima T, Oyama-Okubo N, Nakamura Y, Moriguchi T,
Nakajima I. Effects of GA3+4 and GA4+7 Application Either Alone or Combined with
Prohexadione-Ca on Fruit Development of Japanese Pear ‘Kosui’. Hortic J. 2015;85:201–208.
(a) Gott KA, Thomas TH. Comparative effects of gibberellins and an N-substituted
phthalimide on seed germination and extension growth of celery (Apium graveolens L.). Plant
Growth Regul. 1986;4:273–279.
(b) Lal M, Jaiswal VS. Modification of flower sex and acid phosphatase activity by
phthalimides in female plants of Morus nigra L. Plant Growth Regul. 1988;7:29–37.

(c) Rodaway SJ, Gates DW, Brindle C. Plant Growth Regul. 1991;10:243–259.
(d) Thomas TH. Control of early seedling growth in varietal lines of hexaploid wheat
(Triticum aestivum), durum wheat (Triticum durum), and barley (Hordeum vulgare) in
response to the phthalimide growth regulant, AC 94,377. Sci Hortic (Amsterdam).
1984;23:113–117.
(e) Chandler PM, Harding CA, Ashton AR, Mulcair MD, Dixon NE, Mander LN.
Characterization of gibberellin receptor mutants of barley (Hordeum vulgare L.). Mol Plant
2008;1:285–294.
7.
8.
9.
Jiang K, Otani M, Shimotakahara H, Yoon J-M, Park S-H, Miyaji T, Nakano T, Nakamura H,
Nakajima M, Asami T. Substituted pthalimide AC94377 is a selective agonist of the gibberellin
receptor GID1. Plant Physiol. 2017;173:825-835
Hedden P, Graebe JE. Inhibition of gibberellin biosynthesis by paclobutrazol in cell-free
homogenates of Cucurbita maxima endosperm and Malus pumila embryos. J Plant Growth
Regul. 1985;4:111–122.
Sun TP, Kamiya Y. The Arabidopsis GA1 locus encodes the cyclase ent-kaurene synthetase A
of gibberellin biosynthesis. Plant Cell 1994;6:1509–1518.
10. Nakajima M, Takita K, Wada H, Mihara K, Hasegawa M, Yamaguchi I, Murofushi, N. Partial
purification and characterization of a gibberellin-binding protein from seedlings of Azukia
angularis. Biochem Biophys Res Commun. 1997;241:782–786.
11. Griffiths J, Murase K, Rieu I, Zentella R, Zhang Z-L, Powers SJ, Gong F, Phillips AL, Hedden
P, Sun T, Thomas SG. Genetic Characterization and Functional Analysis of the GID1
Gibberellin Receptors in Arabidopsis. Plant Cell 2006; 18:3399–3414.

12. Dill A, Jung HS, Sun TP. The DELLA motif is essential for gibberellin-induced degradation of
RGA. Proc Natl Acad Sci U.S.A. 2001;98:14162–14167.
13. Ogawa M, Hanada A. Gibberellin biosynthesis and response during Arabidopsis seed
germination. Plant Cell. 2003;15:1591–1604.
14. Goda H, Sasaki E, Akiyama K, Maruyama-Nakashita A, Nakabayashi K, Li W, Ogawa M,
Yamauchi Y, Preston J, Aoki K, Kiba T, Takatsuto S, Fujioka S, Asami T, Nakano T, Kato H,
Mizuno T, Sakakibara H, Yamaguchi S, Nambara E, Kamiya Y, Takahashi H, Hirai MY,
Sakurai T, Shinozaki K, Saito K, Yoshida S, Shimada Y. The AtGenExpress hormone and
chemical treatment data set: experimental design, data evaluation, model data analysis and data
access. Plant J. 2008;55:526–542.

Supplementary data
Supplementary Figure 1. 67D rescued the seed germination of Col-0 that was suppressed by PHX.
Col-0 seeds were germinated on 1/2 MS medium containing DMSO as a mock treatment, 5 µM PHX,
or 5 µM PHX plus 10 µM 67D. The seed germination rate was determined after incubation for 7 days
in darkness. Error bars indicate standard error (n = 3).

Supplementary Figure 2. Automated docking of 67D to GID1a. The docking simulation was performed
with suflex-dock (Biopharmics LLC) and visualized with UCSF CHIMERA. PDB data of GID1a was
downloaded from RCSB Protein Data Bank under PDB code 2ZSI. GID1a and DELLA domain of
RGA were displayed in gray. 67D (cyan) superposes with GA4 (purple) in the pocket of GID1a (gray).

Supplementary Figure 3. Primers used for qPCR analysis.

Supplementary note
Anthracene-9,10-endo-α,β-succinic (anhydride). 7.0 g anthracene and 19.0 g maleic acid were mixed in 140 ml
toluene with reflux for 4 hours. The chemical was purified by recrystallization with ethyl acetate, and collected by
1
suction filtration. HNMR (CDCl3, 500MHz): δ3.52-3.53 (dd, 2H, J1=2.0 Hz, J2=1.4 Hz), δ4.82-4.83 (d, 2H, J=1.6
Hz), δ7.19-7.40 (m, 8H, Ar-H)
1 (67D), 2(S)-3-Phenyl-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido) propanoic acid. One gram of
(0.36×10-2 mol) anthracene-9,10-endo-α,β-succinic(anhydride) and l-phenylalanine 0.66 g (0.40×10-2 mol) were
mixed in 20 ml acetate acid with reflux for 12 h. After reaction, the mixture was concentrated by vacuum and mixed
with 0.6 N HCl. Suction filtration was carried out with distilled water. Precipitate was dried under vacuum and
dissolved in acetate acid. Chemical was purified by recrystallization with distilled water. 1H NMR(CDCl3,500MHz) δ
2.50-2.54(dd,1H,J1=14.5 Hz,J2=8.5 Hz), δ 3.03-3.06(dd,1H,J1=9.0 Hz,J2=3.5 Hz), δ 3.15-3.18(dd,1H,J1=8.5
Hz,J2=4.0 Hz), δ 3.20-3.24(dd,1H,J1=11.5 Hz,J2=6.5 Hz), δ 4.60-4.63(dd,1H,J1=8.5 Hz,J2=7.0 Hz),δ 4.72(d,1H,J=3.5
Hz), 4.78-4.79(d,2H,J=3.5 Hz),δ 7.02-7.37(m,13H,Ar−H)
2, 2-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido)acetic acid. The chemical was synthesized with the
same procedure as described in 1. 1H NMR(CDCl3,500MHz) δ 1.25(s,2H), δ 3.32(s,1H,), δ 3.84(s,1H), δ 4.80(s,2H),δ
−
7.02-7.37(m,8H,Ar H)
3, 2(S)-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido)propanoic acid. The chemical was synthesized
with the same procedure as described in 1. 1H NMR(CDCl3,500MHz) δ 0.78-0.79(d,3H,J=7.0 Hz), δ 3.20-
3.22(dd,1H,J1=8.5 Hz,J2=3.5 Hz), δ 3.27-3.29(dd,1H,J1=8.5 Hz,J2=3.5 Hz),δ 4.39-4.43(q,1H,J=7.5 Hz), δ 4.78-
−
4.80(dd,2H,J1=4.0 Hz,J2=4.0 Hz),δ 7.11-7.39(m,8H,Ar H)
4, 2(S)-3-methyl-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido)butyric acid. The chemical was
1
synthesized with the same procedure as described in 1. H NMR(CDCl3,500MHz) δ 0.22-0.23(d,3H,J=7.0 Hz), δ

0.74-0.76(d,3H,J=6.5 Hz),
δ
2.23-2.27(q,1H,J=7.0 Hz),δ 3.25-3.27(dd,1H,J1=9.0 Hz,J2=2.5),
δ
3.29-
3.31(dd,1H,J1=8.5 Hz,J2=3.5 Hz), δ 3.54-3.54(d,1H,J=0.5 Hz), δ 4.13-4.14(d,1H,J=8.5 Hz), δ 4.83(s,2H),δ 7.10-
−
7.40(m,8H,Ar H)
5, 2(S)-4-methyl-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido) penthanoic acid. The chemical was
1
synthesized with the same procedure as described in 1. H NMR(CDCl3,500MHz) δ 0.57-0.59(d,3H,J=6.5 Hz), δ
0.61-0.62(d,3H,J=6.9 Hz), δ 1.49-1.55(m,2H), δ 3.21-3.24(dd,1H,J1=8.8 Hz,J2=3.4 Hz), δ 3.28-3.30(dd,1H,J1=8.8
Hz,J2=3.4 Hz), δ 4.45-4.48(dd,1H,J1=10.1 Hz,J2=5.6 Hz), δ 4.78(d,2H,J=3.4),δ 7.10-7.37(m,8H,Ar−H)
6, 2(S)-3-methyl-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido) penthanoic acid. The chemical was
1
synthesized with the same procedure as described in 1. H NMR(CDCl3,500MHz) δ 0.23-0.26(m,2H), δ 0.56-
0.59(t,3H,J=6.8 Hz), δ 0.72-0.73(d,3H,J=6.7 Hz),δ 1.96-1.97(m,1H), δ 3.23-3.25(dd,1H,J1=8.9 Hz,J2=3.2 Hz), δ 3.27-
3.30(dd,1H,J1=9.0 Hz,J2=3.2 Hz), δ 4.17-4.19(d,1H,J=8.4 Hz), δ 4.78(s,2H),δ 7.10-7.37(m,8H,Ar−H)
HRMS (ESI+) m/z calcd. for C24H22NO4 (M+H) 388.1554, found 388.1550
7, 2-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido)penthanoic acid. The chemical was synthesized with
1
the same procedure as described in 1. H NMR(CDCl3,500MHz) δ 0.18-0.24(m,1H), δ 0.51-0.57(m,1H), δ 0.61-
0.64(t,3H,J=7.4 Hz), δ 1.43-1.48(m,1H), δ 1.70-1.75(m,1H),δ 3.22-3.24(dd,1H, J1=8.7 Hz,J2=3.4 Hz), δ 3.28-
3.31(dd,1H,J1=8.8 Hz,J2=3.4 Hz), δ 4.34-4.37(dd,1H,J1=10.4 Hz,J2=4.8 Hz), δ 4.77-4.79(dd,2H, J1=6.0 Hz,J2=3.4
Hz),δ 7.10-7.40(m,8H,Ar−H)
8, 2-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido)hexanoic acid. The chemical was synthesized with
1
the same procedure as described in 1. H NMR(CDCl3,500MHz) δ 0.27-0.32(m,1H), δ 0.57-0.63(m,1H), δ 0.75-
0.78(t,3H,J=7.2 Hz), δ 1.01-1.10(m,2H), δ 1.35-1.43(m,1H), δ 1.74-1.81(m,1H),δ 3.21-3.24(dd,1H, J1=8.7 Hz,J2=3.4
Hz), δ 3.27-3.30(dd,1H,J1=8.7 Hz,J2=3.4 Hz), δ 4.30-4.33(dd,1H,J1=9.9 Hz,J2=5.0 Hz), δ 4.77-4.79(dd,2H, J1=6.4
Hz,J2=3.4 Hz),δ 7.10-7.40(m,8H,Ar−H)

9, 2-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido)octanoic acid. The chemical was synthesized with the
same procedure as described in 1. 1H NMR(CDCl3,500MHz) δ 0.31-0.39(m,1H), δ 0.61-0.69(m,1H), δ 0.86-
0.89(t,3H,J=7.4 Hz), δ 0.98-1.08(m,2H), δ 1.08-1.18(m,2H), δ 1.20-1.27(m,2H), δ 1.33-1.40(m,1H), δ 1.73-
1.80(m,1H),δ 3.22-3.24(dd,1H, J1=8.7 Hz,J2=3.2 Hz), δ 3.27-3.30(dd,1H,J1=8.5 Hz,J2=3.4 Hz), δ 4.30-
4.33(dd,1H,J1=9.7 Hz,J2=5.0 Hz), δ 4.77-4.79(dd,2H, J1=5.8 Hz,J2=3.4 Hz),δ 7.09-7.39(m,8H,Ar−H)
10, 2(S)-4-methylthio-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido) butanoic acid. The chemical was
synthesized with the same procedure as described in 1. 1H NMR(CDCl3,500MHz) δ 1.46-1.49(distorted dd,1H,J=7.4
Hz), δ 1.62-1.66(t,2H,J=6.0 Hz), δ 1.95(s,3H),δ 2.05-2.09(distorted dd,1H),δ 3.22-3.24(dd,1H, J1=9.0 Hz,J2=3.5 Hz),
δ 3.30-3.32(dd,1H,J1=8.8 Hz,J2=3.5 Hz), δ 4.62-4.64(dd,1H,J1=8.7 Hz,J2=4.8 Hz), δ 4.78-4.79(dd,2H, J1=8.0
Hz,J2=3.5 Hz),δ 7.14-7.37(m,8H,Ar−H)
11, (S)-4-phenyl-2-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido)-butanoic acid. The chemical was
synthesized with the same procedure as described in 1. NMR(CDCl3, 500MHz): δ 1.84-1.87(dd, 1H, CHCH2), δ 1.88-
1.92(m, 2H, CCH2), δ 2.06-2.09(dd, 1H, CHCH2), δ 3.14-3.17(dd, 1H, 11,12-H), δ 3.19-3.21(dd, 1H, 11,12-H), δ
−
4.35-4.38(dd, 1H,CHCOOH), δ 4.77-4.78(d,1H,9,10-H), 4.78-4.79(d,1H,9,10 H), δ 7.02-7.37(m,13H,ArH)
12, 2(S)-3-Phenyl-(9,10-dihydro-9,10-ethanoanthracene-11,12-dicarboximido) propanoic acid methyl ester. 1.70 g
67D was mixed with 50 ml methanol and cooled down. The mixture was added with 1.5 ml thionyl chloride drop by
drop. After reaction for 24 h at room temperature, the mixture was dried by vacuum and dissolved in acetate acid. The
chemical was recrystallization by adding distilled water. 1H NMR(CDCl3,500MHz) δ 2.41-2.46(dd,1H,J1=14.5
Hz,J2=8.5 Hz), δ 2.99-3.01(dd,1H,J1=8.5 Hz,J2=3.5 Hz), δ 3.19-3.23(dd,1H,J1=14.5 Hz,J2=7.0 Hz), δ 3.47(s,3H)δ
4.53-4.56(dd,1H,J1=8.0 Hz,J2=7.0 Hz),
δ
4.70-4.71(d,1H,J=3.5 Hz),4.77-4.78(d,1H,J=3.0 Hz),δ 7.01-
7.36(m,13H,Ar−H)
The synthetic yields of all the above chemicals were over 70%.