Biological Activity of Brassinosteroids – Direct Comparison of Known and New Analogs in planta

Article abstract of DOI:10.1002/hlca.201600305

A systematic investigation of structural modifications of brassinosteroids is presented. We describe in detail their synthetic preparation, which includes significant improvements of previously reported protocols as well as access to new analogs with functional modifications of the steroid skeleton and of the C(17)-attached side chain. We report the biological potency of the prepared brassinosteroid analogs as plant hormones, which were carefully established in the French bean second internode elongation assay and discuss our observations in light of the recently reported structural data detailing the molecular interactions between brassinolide in the trimeric complex with the protein receptors kinases BRASSINOSTEROID INSENSITIVE 1 (BRI1) and SOMATIC EMBRYOGENESIS RECEPTOR KINASE 1 (SERK1). In a further part of this work we describe the preparation of H₂O-soluble pro-forms of 24-epicastasterone and we discuss their physical properties, hydrolytic stabilities and biological activity.

Full text of DOI:10.1002/hlca.201600305

Received Date: 22-Sep-2016
Revised Date: 24-Oct-2016
Accepted Date: 26-Oct-2016
Article Type: Full Paper
Biological activity of Brassinosteroids - direct comparison of known and new analogs
in planta
Sebastian Wendeborn,*1 Mathilde Lachia,1 Pierre M. J. Jung,1 Joerg Leipner,1 David
Brocklehurst, 2 Alain De Mesmaeker,1 Regis Mondiere1
1 Syngenta Crop Protection AG, Chemical Research, Schaffhauserstrasse 101, CH-4332
Stein, Switzerland
2 Syngenta Jealotts Hill Int. Research Center, Bracknell, Berkshire, RG42 6EY, United
Kingdom
* corresponding author: email address: sebastian.wendeborn@syngenta.com
Abstract:
A systematic investigation of structural modifications of brassinosteroids is presented. We
describe in detail their synthetic preparation, which includes significant improvements of
previously reported protocols as well as access to new analogs with functional modifications
of the steroid skeleton and of the C(17)-attached side chain. We report the biological
potency of the prepared brassinosteroid analogs as plant hormones, which were carefully
established in the French bean second internode elongation assay and discuss our
observations in light of the recently reported structural data detailing the molecular
interactions between brassinolide in the trimeric complex with the protein receptors kinases
BRASSINOSTEROID INSENSITIVE
1
(BRI1) and SOMATIC EMBRYOGENESIS
RECEPTOR KINASE 1 (SERK1). In a further part of this work we describe the preparation of
water-soluble pro-forms of 24-epicastasterone and we discuss their physical properties,
hydrolytic stabilities and biological activity.
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/hlca.201600305
This article is protected by copyright. All rights reserved.

OH
OH
OH
R5
R5
O
NH2
O
OH
OH
OH
HO
H
H
O
R1
R2
H
R1
R2
H
O
H
H
H
H
H
HO
R4
R4
R3
O
O
R3
O
Structure Activity Relationships of brassinosteroid analogs and preparation of pro-forms with improved properties
Keywords:
brassinosteroids; castasterone; water-soluble pro-forms; plant hormones; plant growth
promotion; French bean second internode bioassay
1. Introduction
Plant hormones play a crucial role in orchestrating plant development, ranging from seed-
germination,1 flowering,2 fruit development and fruit ripening3 to processes which terminate
the seasonal plant life cycle4 (senescence). The plant senses and translates environmental
signals, such as light intensity and wavelength, temperature, soil quality (such as humidity,
nutritional status and mechanical resistance) into plant hormone signaling, which allow the
plant to react and adapt to environmental opportunities and challenges (nutrients, water,
light, biotic and abiotic stresses) thereby optimizing growth under the given environmental
conditions. The detailed understanding of plant hormone signaling molecules is not only of
fundamental but also of commercial interest, as it opens opportunities to influence plant
development in modern farming, which may result in more robust and higher yielding crops.
In the past, for example, breeding of crop varieties, in particular wheat, with low gibberellic
acid content was a key contribution to the ‘green revolution’: A reduction of the concentration
of this growth promoting plant hormone results in plants with shorter stems, allowing an
increased proportion of the plants’ assimilates/biomass to be utilized for seed production. At
the same time, shorter plants are less prone to lodging therefore reducing the risk of crop
loss, a further significant benefit.5 Likewise, targeted and specific plant hormone signaling
modifications may open opportunities to help crop-plants to cope with environmental, biotic
and abiotic stresses, and to maintain and increase crop productivity. Therefore, both
academic and industrial laboratories continue to conduct intensive research programs in the
area of plant hormone research.6
This article is protected by copyright. All rights reserved.

The brassinosteroid plant hormones are a class of relatively small molecules with
pronounced effects on plant development, positively influencing plant growth and organ
development (Figure 1). Brassinosteroids promote plant growth through cell-elongation,7 and
play a crucial role in the cell cycle.8 Moreover, numerous studies have demonstrated that
brassinosteroids are involved in the mitigation of abiotic stresses, such as cold, drought and
heat.9
OH
24'
24
OH
OH
OH
22
21
18
27
OH
20
23
25
26
OH
OH
12
11
OH
16
H
17
14
13
H
19
H
HO
HO
H
1
4
HO
HO
15
H
2
HO
HO
HO
HO
H
8
9
10
7
O
5
O
6
n
H
O
O
H
n
3
n
O
H
H
O
O
O
n = 1
n = 0
28-homobrassinosteroid
28-homocastasterone
brassinolide 1
castasterone 2
24-epibrassinosteroid 3
24-epicastasterone 4
Figure 1. Representative brassinosteroids and carbon numbering
For example, rice plants genetically modified to over-express sterol C-22 hydroxylase, a key
enzyme in the biosynthetic pathway to brassinosteroids, and consequently with an increased
endogenous concentration of brassinosteroids, were more heat tolerant, produced more
tillers, and more and heavier seeds.10 On the other hand, treatment of wheat seeds with 28-
homo brassinolide resulted in higher harvested grain yields under both non- and drought-
stressed conditions.11 Numerous other examples are reported in the literature,
demonstrating the beneficial effects of brassinosteroids on crop yield under optimal and sub-
optimal growing conditions.12 Brassinosteroids are highly potent molecules, exhibiting their
biological effects at very low concentration, and therefore some commercial applications in
agriculture are possible already today, despite their relatively high production costs.13,14
Originally isolated from rape pollen, brassinolide (1) is a plant steroid produced from
campesterol through several oxidative steps.15 Recently the detailed molecular interactions
of brassinolide with its target, the ectodomain of the plasma-membrane bound leucine-rich
repeat receptor kinase BRASSINOSTEROID INSENSITIVE 1 (BRI1), was elucidated by x-
ray crystallography.16,17 The brassinosteroid-BRI1 complex triggers a signaling cascade:18
this complex allows an interaction with the ectodomain of another membrane bound kinase
SOMATIC EMBRYOGENESIS RECEPTOR KINASE 1 (SERK1) and both kinases, BRI1 and
SERK1, undergo cross phosphorylation of their cytoplasmic subunits, thereby transferring
the hormonal signal into the cell. Through the involvement of other proteins in the signaling
cascade, eventually the activation of transcription factors is achieved and specific genes in
This article is protected by copyright. All rights reserved.

the cell nucleus are activated for transcription. Thanks to the elucidation by x-ray
crystallography of the BRI1 – brassinolide – SERK1 complex, detailed structural knowledge
of the brassinosteroid binding is now available and this should facilitate the understanding
and the design of potent brassinosteroid analogs and mimics.
Over the last decades, several research groups have prepared and studied synthetic
analogs of brassinosteroids with the aim of understanding the structure-activity-relationship
(SAR) of this fascinating plant hormone and optimizing their biological performance for
applications in agriculture.19,20 However, while there is a wealth of information available, it is
very difficult to compare the numerous analogs reported by different research groups, as
frequently different testing conditions were used to evaluate their biological potency.
In this publication we report the preparation and biological evaluation of several new
brassinosteroid analogs and compare them to a number of previously reported analogs,
many of which we have re-prepared and tested side by side in our biological screening for
potency. We discuss our findings in the context of the recently published crystallographic
data.
2. Relative biological potency of brassinosteroid derivatives
We evaluated the biological activity of brassinosteroid analogs using the French bean
second internode elongation assay (Figure 2),15 one of the most frequently used bioassays
for the detection of brassinosteroids biological activity, beside the also frequently used rice
lamina inclination test.21 Owing to the bell-shaped rate dependence of the elongation
response typical for plant-hormone dose-responses,22 the application amount at which the
largest increase of second internode elongation occurred was determined and is reported
accordingly in the tables in the following sections as optimal amount. Details of the French
bean second internode elongation assay are described in the experimental section. As an
additional test we evaluated the use of the Arabidopsis dwarf mutant det2-123 which is
defective at an early stage in brassinosteroid biosynthesis24 and has a short hypocotyl
phenotype which is rescued by the exogenous application of brassinolide.25 Brassinosteroid
derivatives were thus assessed for biological activity based on their ability to induce
hypocotyl elongation in det2-1 seedlings and simultaneously for their inhibitory effect on root
growth.26 The data obtained with the Arabidopsis dwarf mutant det2-1 appeared to be less
accurate and less sensitive; we therefore report the corresponding results in the
supplementary material and focus the discussion on the results obtained with the French
bean second internode elongation assay.
This article is protected by copyright. All rights reserved.

Figure 2: Example of the French bean second internode elongation assay. Plants were applied only with solvent
(A) or with 5 µg plant-1 of 24-epi-brassinosteroid 3. The point of application is marked with a blue arrow. The
second internodes are marked with red lines. As can be seen, 24-epi brassionsteroid 3 induces a significant
elongation of the second internode (for experimental details see supplementary material section 3.1)
In most studies performed with brassinosteroids, six types of analogues have been used
(Figure 1), which differ in the alkyl substitution of the side chain and the oxidation state of the
B-ring: (1) brassinolide, (2) 24-epibrassinolide and (3) 28-homobrassinolide, the latter two
being synthetically accessible in five to seven steps from the steroid precursors ergosterol
and stigmasterol, respectively. Frequently, the castasterone analogues (4) castasterone, (5)
24-epicastasterone and (6) 28-homocastasterone, which contain a cylohexanone B-ring
instead of the 7-membered ring lactone typical for the brassinosteroids are also used to
evaluate functional modifications of the steroid skeleton.27,28 This is justified, as castasterone
retains a large portion of the biological activity of brassinolide, as demonstrated by the
similar high activity of 24-epicastasterone and 24-epibrassinolide in the French bean second
internode elongation assay (Table 1). Therefore, many of the analogs we discuss here are
analogs of 24-epicastasterone.
This article is protected by copyright. All rights reserved.

2.1 Biological activity of reference compounds
OH
OH
OH
OH
OH
OH
OH
OH
H
H
H
HO
HO
H
HO
HO
HO
HO
HO
HO
H
H
H
H
H
O
H
H
O
H
O
O
H
H
H
H
O
O
brassinolide 1
24-epibrassinolide 3
castasterone 2
24-epicastasterone 4
Table 1. Biological activity of reference brassinosteroids in the French bean second internode elongation
bioassays. Compounds were tested at 0.005, 0.05, 0.5, 5 and 50 µg plant-1 and compared to mock treated
controls (=100%). Values are means ± SD of four replications from one experiment were compounds were
evaluated side by side.
French bean second internode bioassay
Compoun
Optimal amount (µg plant-1)
Relative activity (%)
d
1
3
2
4
0.05
0.5
5
235 ± 33
210 ± 22
246 ± 57
218 ± 51
5
To establish the relative activity of 24-epicastasterone (4) and castasterone (2), two
important reference compounds in our study, we compared them side by side with
brassinolide (1) and 24-epibrassinolide (3) in our French bean second internode assay. As
reported in Table 1, the biological activity of 24-epicastasterone (4) and castasterone (2) are
very similar at identical application rates. By contrast, 24-epibrassinolide (3) retains only
about 10% of the biological activity of brassinolide (1), which is approximately 100 times
more active than castasterone.
This article is protected by copyright. All rights reserved.

2.2 Biological activity of A- and B-ring modified brassinosteroids
2.2.1 Modifications of the A-ring
OH
24
OH
R4
H
R3
R1
2
3
H
O
H
R2
H
Table 2. Biological activity of A- and B-ring modified 24-epicastasterone in the French bean second internode
elongation bioassays. Compounds were tested at 0.5, 5 and 50 µg plant-1 and compared to mock treated controls
(=100%). Values are means ± SD of four replications from different experiments. n.t.: not tested; NA not active
French bean second
internode bioassay
Optimal
amount
Relative
activity
(%)
Compound
R1
R2
R3
R4
Reference
(µg plant-1)
29
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
OH
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
OH
5
n.t.
146 ± 6
NA
30,31,32,33
OH
H
6
50
-
OMe
H
7
this study
this study
this study
OCH2C(O)OH
H
8
-
NA
OCH2C(O)NMe2
H
H
9
-
NA
OCH2C(O)NH2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
50
50
-
115 ± 25
142 ± 57
NA
this study
34
F
H
H
H
H
H
H
H
H
H
H
H
H
H
31,33,34,35,36
37
OH
OMe
-
NA
OCH2CH2OMe
OCH2CH2OH
OCH2C(O)OH
OCH2C(O)OMe
OCH2C(O)OEt
OCH2C(O)NMe2
OCH2C(O)NHMe
OCH2C(O)NH2
H
-
NA
this study
this study
this study
this study
this study
this study
this study
-
NA
-
NA
-
NA
-
NA
-
NA
-
NA
-
NA
50
220 ± 63
n.t.
this study
31
H
24-Epicastasterone analogs with no functionalization of the A-ring are inactive: 5 (R1, R2,
R3, R4 = H) showed no activity in our det-2 Arabidopsis hypocotyl elongation assay
(supplementary data Table S1) indicative of the importance of the hydroxyl functionalities of
This article is protected by copyright. All rights reserved.

the A-ring. This is in agreement with very weak activity observed for the corresponding 28-
homocastasterone analog (24, Table 3) in the rice laminar inclination test.38 However, upon
the introduction of the C(3)-hydroxyl group in the natural α-configuration (6) onto the A-ring,
significant activity is observed in the French bean second internode elongation assay. This
is in agreement with the previously reported activity in the rice laminar inclination test, were
this compound exhibits approximately 50% of the activity of 24-epicastasterone.30 The
corresponding fluoro-analog 11, showed almost identical activity. This is in good agreement
with a previously reported moderate activity for this compound.34 The related analog (25,
Table 3) in the 28-homo series was reported to exhibit excellent activity in the rice laminar
inclination test, while the isomeric β-C(3)-F analog (26) was significantly less active.39,40
Methylation of the C(3)-OH group compromised biological activity: the α-C(3)-OMe analog 7
was inactive in the French bean second internode elongation assay. Other, larger α-isomeric
substituents in this position also led to compounds 8, 9 and 10 with no or very little activity.
OH
OH
R4
R3
H
O
H
R1
R2
Table 3. Structures of 28-homocastasterone analogs
R1
R2
R3
R4
ref.
compound
24
25
26
H
F
H
H
F
H
H
H
H
H
H
38
39
40
H
In contrast to 6 (= 24-epityphasterol) with hydroxylation of the A-ring only at C(3) in the α-
configuration, the analog 12 (= 24-epiteasterone), possessing a β-C(3)-OH was lacking any
biological activity in our biological assessment. Previous studies in the rice laminar assay
reported that teasterone (possessing a 24-ent methyl group) retains about 10-20% of the
activity of castasterone,41 and 6 (24-epityphasterol) retains 60% and 87% of the activity of
24-epicastasterone and 24-epibrassinolide respectively. 30,32 The lack of activity of 12 in our
This article is protected by copyright. All rights reserved.

assay could be rationalized by a possibly higher sensitivity of the rice laminar assay for
those analogs compared to French bean second internode elongation assay used in our
study, possibly also facilitated by a more efficient biochemical conversion of teasterone to
typhasterol in rice, as both are consecutive intermediates in the biosynthesis of brassinolide.
Due to the facile synthetic accessibility of β-C(3)-OR analogs, we prepared a number of
analogs, including 13, which was previously reported to be quite active in the rice laminar
inclination test at 10 and 100 ng plant-1.37 However, these demonstrated no activity in our
assays. 14 and 15 were equally inactive as well as analogs 16, 17, 18, 19, 20 and 21 which
showed no ability to promote internode elongation in our French bean second internode
elongation assay.
Very little is known about castasterone analogs which are deoxygenated in the C(3) position.
In the biosynthetic pathway, oxygenation of the C(3) position precedes oxygenation of the
C(2) position, and therefore C(2)-OH/C(3)H2 analogs 22 and 23 are not intermediates in the
biosynthesis of castasterone.41,42 The previously prepared 23 had approximately half the
activity of 24-epicastasterone in the rice laminar inclination test.31 The recently reported
structural data detailing the molecular interactions between brassinolide in the trimeric
complex with BRI1 and SERK1 suggests that the C(2)-OH group of brassinolide is engaged
in a tight hydrogen-bonding network involving both the proton and the oxygen lone pair of
the C(2)-OH with SERK1’s His62 (pdb 4LSX)17 (or His61(pdb 4M7E)) imidazole and
backbone NH (Figure 3). C(3)-OH is involved in hydrogen bonds with the of Val63 and
Tyr642 (pdb: 4LSX) (or Val62 and Tyr642 in 4M7E). A recent computational analysis
discusses the possible contributions of those hydrogen bonds in detail.43
This article is protected by copyright. All rights reserved.

Figure 3. Brassinolide in its complex with BRI1 and SERK1 (pdb code: 4LSX). Hydrogen bonds discussed in the
text are highlighted
As x-ray crystallographic data indicates that the α-C(2)OH of brassinolide is involved in tight
binding with SERK1, we decided to prepare the previously unknown analog 22 which is
hydroxylated with the natural α-configuration at the C(2) position and deoxygenated in the
C(3) position. Indeed, 22 showed very strong internode elongation (Table 2 and Table 4) at
50 µg plant-1 in the French bean second internode assay, similar to the activity observed for
24-epicastasterone 4 at 0.5 µg plant-1 (Table 1). In a side by side comparison 22 and 24-epi
brassinolide (3) showed biological activity already at 0.5 µg plant-1 which was, however,
statistically significantly higher at this amount for 24-epi brassinolide than for compound 22
(Table 4). By contrast, compound 6, deoxygenated at C(2), induced second internode
elongation only at 5 µg plant-1 or higher. Taken together, this confirms the high relevance of
both, the C(2)-OH and the C(3)-OH, but points to a higher relevance of the α-C(2)-OH in the
overall contribution to the biological activity of brassinosteroids. The tight interactions
observed in the crystallographic data for the BRI1-brassinolide-SERK1 complex of this
position are in agreement with this observation.
This article is protected by copyright. All rights reserved.

Table 4. Biological activity of C(2) and C(3) deoxygenated 24-epicastasterone analogues (6, 22) and 24-
epibrassinolide (3) in the French bean second internode elongation bioassays. Compounds were tested at 0.5, 5
and 50 µg plant-1 and compared to mock treated controls (=100%). Values are means ± SD of four replications
from one experiment.
Compound
Application amount (µg plant-1)
0.5
5
50
3
204 ± 41
158 ± 35
NA
170 ± 11
167 ± 37
142 ± 6
167 ± 8
22
6
220 ± 63
146 ± 6
Among the four hydroxyl groups present in brassinolide, the equatorial α-C(2)-OH is the
most accessible and therefore likely to be glycosylated. While important for activity, the α-
C(2)-OH is therefore also a significant soft spot of brassinolide for secondary metabolism.
Consequently, we have investigated protected pro-forms of brassinosteroids (vide infra).
2.2.2 C(5)-fluorinated analogs
Previously, the C(5)-fluorinated 28-homocastasterone (30, Figure 4) has been prepared and
studied for its biological effects in the rice lamina inclination test, which suggested weaker
activity compared to the non-fluorinated 28-homocastaterone. However, fluorination at C(5)
(31, Table 3) is reported to increase the biological effect in the rice lamina inclination test in
the 28-homotyphasterol-series which is deoxygenated in the C(2) position.44
This article is protected by copyright. All rights reserved.

OH
OH
OH
OH
OH
OH
HO
HO
H
O
H
H
O
H
H
O
H
HO
HO
F
F
F
27
28
29
OH
OH
R
H
O
H
HO
F
30 R = OH C(5)-fluorinated 28-homocastasterone
31 R = H C(5)-fluorinated 28-homotyphasterol
Figure 4. C(5)-fluorinated brassinosteroid analogs
2
OH
O
OH
5
H
3
O
O
F
H
Figure 5. Proposed hydrogen bonding in C(5) fluorinated 24-epicastasterone between fluorine and C(3)-OH
C(5)-Fluorination of castasterone is expected to increase the electrophilicity of the
neighbouring carbonyl group. In addition, and as suggested by Ramírez and coworkers44
and studied experimentally in a recent comparative study of fluorinated and non-fluorinated
conformationally locked cyclohexanols by Graton et al., a hydrogen-bond between the 1,3-
diaxial C(5)-F and the C(3)-OH can be formed (Figure 5) and this could result in a decrease
of the hydrogen bond donating capacity of the α-C(3)-OH group towards an external H-bond
acceptor by approximately 1.2 kcal mol-1.45 Notably, in the complex with BRI1–SERK1, the
C(3)-OH of brassinolide is engaged through a hydrogen bond with the backbone-C=O of
Val63, which should be weakened as a result of C(5)-fluorination. The lone pair of C(3)-OH
appears to be hydrogen-bonding to Tyr462, an interaction which could also suffer from the
intramolecular hydrogen bond interaction between the C(3)-OH and a fluorine atom at C(5)
which would place the C(3)OH unfavorably opposite the Tyr462-OH. Based on those
considerations we expected a reduction of biological activity upon C(5)-fluorination of 24-
This article is protected by copyright. All rights reserved.

epicastasterone analogs, even though a reduction of solvation as a result of fluorination may
counterbalance such negative effects. In order to validate this hypothesis, three analogs
were prepared and tested (Figure 4). Fluorination of 24-epicastasterone at C(5) significantly
reduces its activity in the French bean second internode elongation assay (Table 5). While
24-epicastasterone had the highest activity at 5 µg plant-1 with an effect of 218% of control,
this value was reduced to 154% for 27 at 10 times higher treatment rate (50 µg plant-1). In
agreement with results discussed above, C(2)-deoxygenation (28) and C(3)-OH inversion
(29) have significant negative effects on biological potency of such C(5)-fluorinated analogs.
We therefore conclude that C(5)-fluorination in the 24-epicastasterone and in the 24-
epityphasterol series significantly reduces biological activity in the French bean second
internode elongation assay. This is in agreement with the proposed interactions of C(3)-OH
in the interaction of brassinolide in its complex with BRI1 and SERK1, as discussed above.
Table 5. Biological activity of C(5)-fluorinated 24-epicastasterone 27, - typhasterol 28, and – teasterone 29 in the
French bean second internode elongation bioassays. Compounds were tested at 0.5, 5 and 50 µg plant-1 and
compared to mock treated controls (=100%). Values are means ± SD of four replications from one experiment.
NA: not active.
French bean second internode bioassay
Optimal amount
(µg plant-1)
Relative activity
(%)
Compound
27
28
29
50
-
154 ± 9
NA
-
NA
2.2.3 C(7) Methylation of the B-ring
All biologically active brassinosteroids possess a carbonyl group at C(6). This carbonyl is
flanked by a C(7)H2- (as in castasterone) or a OC(7)H2-group (as in brassinolide). To
investigate the relevance of this CH2-group we prepared C(7) methylated analogs of 24-
epicastasterone (4) and 24-epibrassinosteroid (3). Methylation of castasterone at C(7)
significantly reduces the biological signal in our biological assays. The observed internode
elongation was below the significance threshold. Only 32 showed weak activity in the
Arabidopsis det-2 hypocotyl elongation assay (Table S1, supplementary material), the
corresponding -isomer 33 lacks activity in both assays. Those results are in agreement with
the structural data of brassinolide in complex with BRI1 and SERK1 (Figure 3): the C(7)-
methylene group of brassinolide is in close contact with Tyr462 (CH- interaction). A similar
This article is protected by copyright. All rights reserved.

contact is expected for castasterone. Methylation of C(7) therefore most likely results in a
steric clash with Tyr462.
OH
OH
OH
OH
HO
HO
HO
HO
H
H
H
O
H
Me
O
O
-Me: 32
-Me: 33
34
Table 6. Biological activity of C(7) methylated 24-epicastasterone (32, 33) and brassinosteroid-analog 34 in the
French bean second internode elongation bioassays. Compounds were tested at 0.5, 5 and 50 µg plant-1 and
compared to mock treated controls (=100%). Values are means ± SD of four replications from one experiment.
NA: not active.
French bean second internode bioassay
Optimal amount
(µg plant-1)
Relative activity
(%)
Compound
34
32
33
4
-
-
NA
NA
-
NA
5
165 ± 13
2.3 Side chain modifications
The 9-carbon containing side chain of the brassinosteroids consists of a syn-diol and a
terminal terpene unit. Previously reported analogs indicate that some variations in the side
chain can improve the biological activity. Thus, 24-epibrassinolide (3) and 28-
homobrassinolide (Figure 1) have similar activity although generally a little bit lower than the
natural brassinolides. Metabolic deactivation of brassinosteroids is known to limit their
biological performance, and hydroxylation of the isopropyl unit as well as glycosylation of the
different hydroxyl groups present in the A-ring and the side chain have been reported.46
Consequently, different research groups have tried to improve the biological activity of the
brassinosteroids by blocking some of the metabolic pathways. However, methylation of
either of the hydroxy groups of the side chain reduced the biological activity of brassinolide
even though the C(22)-OMe brassinolide was more potent than C(23)-OMe brassinolide.47
Takatsuto and co-workers were more successful with the C22-C23-epoxybrassinosteroid
This article is protected by copyright. All rights reserved.

2,3-diacetates (TS303) which was further tested under field conditions and found to be
superior to brassinolides and 24-epibrassinosteroid.48
O
O
N
N
HO
HO
HO
HO
H
H
H
H
H
H
H
O
O
H
O
O
35
36
Table 7. Biological activity of side chain modifications in the French bean second internode elongation bioassays.
Compounds were tested at 0.05, 0.5 and 5 µg plant-1 and compared to mock treated controls (=100%). Values
are means ± SD of four replications from one experiment. NA: not active.
French bean second internode bioassay
Optimal amount
(µg plant-1)
Relative activity
(%)
Compound
35
36
-
NA
5
133 ± 22
In an independent study, an isoxazoline ring was introduced to replace the C(22),C(23)-diol.
Although close analogues of 35 and 36 were reported to display good biological activity,49 35
showed no statistically significant and 36 only a modest activity in the French bean second
internode elongation assay when prepared and tested in our laboratories, the “natural”
stereochemistry at C(22) being preferred.
The recently published structure of the BRI1-brassinolide complex and the BRI1-SERK3-
brassinolide complex helped rationalize these observations and showed that C(22)-OH and
C(23)-OH are involved in crucial interactions with Ser 647 and Tyr 597 as hydrogen bond
acceptor (see Figure 6).17 While the nitrogen and oxygen of the isoxazoline ring of 35 and 36
provides options for such hydrogen bonds, a change in the conformation of the side-chain
and suboptimal trajectories of the hydrogen bonds which result of this modification may
nevertheless result in lower receptor affinity. More recently, molecular modeling studies
highlighted the importance of the hydrogen bond between C(23)-O of brassinolide and the
backbone NH of Ser 647, confirming the higher sensitivity of C(23)-OH to methylation.43
This article is protected by copyright. All rights reserved.

Figure 6. Brassinolide binding mode with BRI1-SERK3 complex with key interactions of 22-OH, 23-OH and alkyl
side chain (pdb code: 4LSX).
Further investigations were carried out for the replacement of the alkyl side chain to improve
the activity of brassinolides. Pharis and co-workers studied the replacement of the terminal
isopropyl group by cyclopropyl and cyclobutyl and reported a 5-7 fold increase of activity in
the rice laminar assay.50 Independently, the replacement of the alkyl chain by an aromatic
ring was also proposed.51 Thus, 23-phenyl brassinosteroid was reported to be equally potent
to 24-epibrassinolide. An analysis of the receptor binding site revealed mostly lipophilic
interactions with Ile 563, Met 657 and Trp 564 and probably some free space available for
side chain modification (Figure 6), which is in agreement with the good activity observed for
such modified brassinosteroid analogs.
To further explore the potential of side-chain derivatives, we decided to prepare additional
analogs of C(23)-arylated brassinosteroids, in the hope that variation of the aryl-group and
its substituents would result in high receptor affinity and metabolically more robust
derivatives.
We prepared and tested a set of analogues with aryl and heteroaryl group at C(23) with
different sizes and electronic properties. The cyclopropyl-moiety was presumably too small
This article is protected by copyright. All rights reserved.

to allow sufficient interactions in the binding pocket of BRI1 and therefore 37 gave no
biological signal in the Arabidopsis det-2 assay and French bean second internode
elongation assay. However, five membered heteroaryl-rings were large enough and such
analogs were able to trigger hypocotyl elongation in the Arabidopsis det-2 assay, and,
gratifyingly, also resulted in significant activity in the French bean second internode
elongation assay. Thiophene 40 was among the most active 5-ring analogs resulting in
second internode length of 187% of control at 5 µg plant-1, whereas the corresponding more
polar oxazole 38 showed only moderate activity in the Arabidopsis det-2 assay and no
significant activity in the French bean second internode elongation assay, probably due to
unfavorable interactions with the lipophilic residues on this part of the brassinosteroids
receptor; the chlorinated thiazole 39, however, showed moderate activity in the French bean
second internode assay. Aryl rings were also potent, tolerating substitutions with halogen (45
and 46), with activities comparable to the effects induced by 24-epicastasterone (Table 1). In
order to increase the solubility and bioavailability of these analogues, we introduced a
carboxylic acid substituent on the phenyl ring, but 41 and 42 were inactive in our assays.
The pyridine ring in 43 was better tolerated with similar activity to the corresponding phenyl
analogue 44. Overall, these results confirmed earlier observations that apolar groups are
preferred. The introduction of the larger naphthyl group resulted in reduced biological
activity, as it most likely exceeded the upper limit of the binding pocket size: 48 still retained
some activity in Arabidopsis det-2 assay but performed only weakly in French bean second
internode elongation assay. As expected from the excellent activity of 46 containing a para-
chloro substituted phenyl ring at C(23), the corresponding brassinosteroid analog 47
possessing a 7-membered ring lactone, was the most potent compound we identified, with
an activity approaching the potency of the natural brassinolide.
These results demonstrate that different aromatic groups can replace the branched alkyl
chain normally attached at C(23) in the brassinosteroids. Most potent analogs are obtained
with lipophilic groups, enabling a biological response as strong as the natural products.
Smaller substituents at C(23) or the introduction of polar functionalities, however, result in
significantly less potent analogs. Further studies will be required to demonstrate increased
resistance of those analogs towards oxidative metabolism. However, the replacement of the
sensitive isopropyl side group in the side chain by an aromatic ring should reduce the rate of
oxidation by cytochrome p450 enzymes.
This article is protected by copyright. All rights reserved.

Cl
OH
22 OH
23 R
OH
HO
HO
HO
HO
HO
H
H
H
H
O
O
O
47
37-46, 48
Table 8. Biological activity of C(23)-aryl brassinosteroid analogs in the French bean second internode elongation
bioassays. Compounds were tested at 0.5, 5 and 50 µg plant-1 and compared to mock treated controls (=100%).
Values are means ± SD of four replications from one experiment. NA: not active.
French bean second internode
bioassay
Optimal amount
(µg plant-1)
Relative activity
(%)
Compound
R
Reference
this study
37
-
-
NA
NA
38
this study
39
40
41
50
5
158 ± 30
187 ± 18
NA
this study
this study
this study
-
42
43
44
45
-
NA
this study
this study
51
5
162 ± 29
205 ± 54
172 ± 50
50
50
this study
This article is protected by copyright. All rights reserved.

46
47
50
5
187 ± 41
192 ± 53
this study
this study
48
50
124 ± 12
this study
2.4 C(2) Acylated pro-forms of 24-epicastasterone with improved chemical properties
To enhance the biological availability of exogenously applied molecules for in planta effects,
the preparation of a pro-form, in the case of pesticides a ‘pro-cide’, is a frequently applied
strategy. In many cases, a rather polar cidal molecule is protected with a lipophilic and
hydrolytically labile moiety to increase the logP of the resulting pro-cide; upon a foliar
application such pro-cides have enhanced up-take through the waxy cuticle and upper
epidermis of the leaf before the active parent cidal principle is released and distributed within
the plant. A typical example for such a pro-cide is the herbicide pinoxaden.52
The very limited transport of brassinolide and closely related analogs in plants is explained
by their chemical properties, in particular a relatively high logP and very low water
solubility.53,54 We therefore aimed to improve the physical chemical properties of 24-
epicastasterone (log P = 4.01, solubility (H2O, pH = 7.2) = 0.37 ppm) through preparation of
appropriate pro-forms. Several reports, although not in the context of brassinosteroid
analogs, support the concept of pro-forms of biologically active molecules to increase water
solubility of otherwise insoluble organic molecules by a factor of up to 100, mostly with
amino acid derivatives.55 A pro-form may also protect a molecule from undesired metabolism
as for example glycosylation; the extent of glycosylation of the four hydroxyl group present in
brassinolide and castasterone can vary from plant species to plant species.56 Chemically,
the equatorial -C(2)-OH is the sterically most accessible among the four secondary
hydroxyl functions present in brassinolide and castasterone (Figure 7).57,36 Accordingly, we
prepared a number of new C(2)O-acylated 24-epicastasterone analogs with the aim to
enhance water solubility and transport compared to the rather insoluble and lipophilic parent
steroid, while at the same time protecting the C(2)- and possibly also the C(3)-hydroxyl
group against undesired glycosidation (Figure 7). In addition, variation of the substitution of
the side chain would also allow fine-tuning of the hydrolytic stability of the ester functionality.
This article is protected by copyright. All rights reserved.

tune solubility
O
FG
OH
OH
O
R
HO
OH
OH
OH
O
OH
O
24-epi Castasterone
tune stability
log P = 4.28
H2O solubility : 0.37ppm
Figure 7. Rational for preparation of 24-epicastasterone pro-forms
24-Epicastasterone analogs with acyl groups containing amino functions, carboxylates or a
combination of the two, as well as analogs with neutral substituents were prepared. To
evaluate the potential for hydrolytic release of the parent 24-epicastasterone (4), all analogs
were evaluated for their hydrolytic stability at pH 5.5, 7 and 8. Compared to the parent 24-
epicastasterone, C(2)-O-esters containing protonable amine-functions (49, 50, 56) increased
water solubility by a factor of as much as 140 to 375. C(2)-O-esters with a carboxylic acid in
the side-chain (52, 53, 54) increased the water solubility at pH 7.2 dramatically by a factor of
1400-25000 as extrapolated from the water-solubility measurements of the non-ionized
species at pH 2.2. The zwitterionic analog 55 containing basic amino and acidic functions in
the side-chain also improved water solubility significantly by a factor of 1600 to 602 ppm. As
expected, the water solubility of the neutral analog 51 was not improved.
Interestingly, analogs containing C(2)-O-esters with a carboxylic acid in the side-chain (52,
53, 54) were stable to hydrolysis, allthough migration of the acyl group into the C(3) position
was observed. By contrast, the C(2)-O-esters with basic amino functions (49, 50) were
significantly less stable and acyl-migration into the C(3) position and hydrolysis was
observed at pH 7 and pH 8 with T ranging from 1 to 8 hours. Most likely the neighboring
½
protonated amine increases the electrophilicity of the carbonyl by hydrogen-bonding and
inductive effects making it more susceptible to hydrolysis. This effect appears to be
‘neutralized’ by the carboxylate in the zwitterionic analog 55, which did not hydrolyse at pH 5
to 8. The neutral analog 51 proved to be hydrolytically stable and no acyl-migration into the
C(3) was observed.
This article is protected by copyright. All rights reserved.

Among the 24-epicastasterone pro-forms prepared, 49 showed strong activity in the French
bean second internode elongation assay (Table 9). This analog had only limited, non-
enzymatic hydrolytic stability at pH 7 and therefore its potency in the French bean second
internode elongation assay came with no surprise. Interestingly, 54, which was hydrolytically
stable at pH 5, 7 and 8, induced significant elongation of the second internode (Table 9),
even though less than the parent 24-epicastasterone. Analog 55, with a zwitterionic aspartic
acid residue attached to the C(2) position, was equally resistant to hydrolysis and exhibited
similar biological activity compared to 54, which lacked the primary amino group. Even
though we have no experimental data about the in vitro activity of 54 and 55, and therefore
cannot rule out intrinsic activity, we believe that this is unlikely as the α-C(2)-OH appears so
crucial for interacting with SERK1, as discussed above; more likely the esters present in 54
and 55 are hydrolysed in-planta by an esterase, thereby releasing the active 24-
epicastasterone.
We demonstrated that a ‘pro-form’ approach can be used to significantly improve water
solubility of the parent 24-epicastasterone. For example, 55 had a water solubility of 602
ppm while the parent 24-epicastasterone hardly dissolved in water (0.37 ppm). As
demonstrated, the exact nature of the pro-form moiety can be used to fine-tune the
hydrolytic stability and therefore the release of the brassinosteroid analog upon uptake into
the plant. Further investigations will be required to understand the fate, distribution and
behavior of such compounds in planta as well as their activity on the receptors. Indeed,
brassinosteroids functionalized at either one or both alcohols in the C(2) and C(3) position
may bind to the brassinosteroid receptor BRI1 but interfere with further complexation to
SERK1, to which C(2)-OH and C(3)-OH form key hydrogen bonds, as outlined above.
Therefore, the discussed compounds could act as brassinosteroid antagonists in their pro-
form, as recently reported for the brassinolide-2,3-acetonide.58
This article is protected by copyright. All rights reserved.

OH
OH
OH
OH
HO
RO
RO
HO
H
O
H
H
O
H
Table 9. Chemical properties, hydrolytic stability and biological activity of C(2)-O esterified 24-epicastasterone
analogs and biological activity in the French bean second internode elongation bioassays. Compounds were
tested at 0.5, 5 and 50 µg plant-1 and compared to mock treated controls (=100%). Values are means ± SD of
four replications from different experiments. NA: not active.
French bean second
internode bioassay
Hydrolytic stability (h) at
H2O
solubility
(ppm / at
pH)
Calc.
Solubility
at 7.2
Optimal
amount
Relative
activity
(%)
Compound
R
LogP
pH 5
pH 7
pH 8
(µg plant-1)
(ppm)
BL (1)
3.26
4.01
3.24 / 7.3
0.05
235 ± 33
24-epi CA (4)
0.37 / 7.2 c)
139 / 7.2 b)
5
218 ± 51
178 ± 68
49
stable a)
stable
1.9
1.2
50
O
H3N+
Cl-
50
51
52
53
54
55
80.6 / 7.23
0.74 / 7.2
1.7
<1.0
5
-
147 ± 48
NA
O
H3N+
O
Cl-
O
stable
stable
stable a)
stable d)
stable a)
stable a)
stable
stable a)
stable d)
stable
stable a)
O
N
H
O
7.08 / 2.12
13.18 / 2.15
9558
5141
stable a)
stable d)
stable
-
NA
O
O
O
HO
HO
O
-
NA
O
50
50
143 ± 16
168 ± 63
HO
O
NH2
602 / 7.3
53.7 / 7.2
stable a)
O
HO
stable d)
stable d)
stable d)
50
123 ± 6
O
56
N
a) partial migration of R onto C(3)-O observed; b) partial migration of R onto C(3)-O during solubility
measurement occurred, the value reported is the sum of the solubility of both isomers; c) three isomers
(diastereomers and partial migration of R onto C(3)-O observed; d) poor MS sensitivity, no quantification
possible, but no hydrolysis observed
3. Androstan and Diosgenin analogs
This article is protected by copyright. All rights reserved.

Androstan analogues,35,59,60 and spirostanic analogues,61,62 have been reported to possess
brassinosteroid type activity (Figure 8). Based on the considerations discussed above, in
particular the relevance of the brassinolides α-C(2)-OH, the β-C(3)-OH and the tight
interactions of the side chain attached to C(17) in the binding cavity of the BRI1 receptor,
this came as a surprise to us and we therefore decided to prepare several analogs to
evaluate their biological activity in our assays. In the androstan series, compound 58,
fluorinated at C(5),45,63 was reported to possess significant activity in the French bean
second internode elongation assay. By contrast, in our hands this compound as well as its β-
C(3)-OH isomer 57 exhibited no activity in the same assay. The spirostanic analogues of
64
natural steroids, in particularly laxogenin, but also BB-6 and BB-16,
additionally
hydroxylated at C(2) or C(5) were reported to induce growth-promoting effect in radish.62,65 In
our hands, the two isomeric spirostanic analogs 59 and 60 fluorinated at C(5) were inactive
in our French bean second internode elongation assay. In conclusion we were not able to
reproduce the reported growth promoting effects of androstan and spirostanic analogues.
O
OH
O
H
H
H
R2
R1
R2
R1
H
O
F
F
O
57 R1 = H, R2 = OH
58 R1 = OH, R2 = H
59 R1 = H, R2 = OH
60 R1 = OH, R2 = H
O
O
O
O
O
O
H
H
H
HO
HO
H
O
H
H
O
H
H
O
H
HO
HO
H
H
HO
BB-6
Laxogenin
BB-16
Figure 8. Androstan and Spirostanic derivatives
Table 10. Biological activity of androstan analogs in the French bean second internode elongation bioassays.
Compounds were tested at 0.5, 5 and 50 µg plant-1 along with 24-epibrassinolide (3) and compared to mock
treated controls (=100%). Values are means ± SD of four replications from one experiment. NA: not active.
French bean second internode bioassay
This article is protected by copyright. All rights reserved.

Optimal amount
(µg plant-1)
Relative activity
(%)
Compound R1
R2
Reference
57
58
59
60
3
OH
H
H
OH
OH
H
-
-
NA
NA
This work and 34,60
This work and 34
This work
H
-
NA
OH
-
NA
This work
5
171 ± 19
This work
4. Synthesis of brassionsteroids and their analogs
4.1 Synthesis of 24-epicastasterone and 24-epibrassinosteroid
Scheme 1 summarizes the synthesis of 24-epicastasterone and 24-epibrassinosteroid. We
followed the sequence starting from ergosterol as originally reported by Thompson and later
improved by McMorris, who introduced the one-pot transformation of the cyclopropyl-
intermediate 64 to unsaturated 65 and the more selective dihydroxylation of the sidechain
applying Sharpless enantioslective dihydroxylation methodology.27,66,67
This article is protected by copyright. All rights reserved.

c) 3% RuO2. (H2O)2
1.5 eq. NaBrO3
a) MsCl, pyridine, r.t
b) KHCO3, acetone, H2O
H
H
H
H
H
H
H
44%
40-60%
H
O
H
HO
63
OH
+
62
ergosterol, 61
d) 6 eq. Li/NH3/THF-78°C
H
then
87%
10 eq. 4 methyl pentenone
then NH4Cl(aq)
H
sideproduct
f) tBuOH/H2O
K2CO3, K3Fe(CN)6
OH
DHQD
MeSO2NH2
cat OsO4
40-62%
H
e) PyrHCl, LiBr, DMA
160°C, 6h
H
H
OH
H
H
H
H
H
HO
HO
78%
H
O
H
H
O
H
O
H
64
65
H
4
g) CF3CO2OH
CHCl3
50%
OH
OH
H
H
HO
HO
H
H
3
O
H
O
Scheme 1. Synthesis of 24-epicastasterone and 24-epibrassinosteroid
Further improvements of the Thompson sequence addressed the reduction of the ,-
unsaturated ketone 63 using solvated electrons. Sodium dithionite (Na2S2O4) was reported
by Cheprakov as an alternative to Li/NH3, however attempts to reduce 63 with this reagent
were unsuccessful in our hands.68 Unfortunately, the reported Li/NH3 reduction of 63
provided the desired product only in varying yields and substantial amounts of over-reduced
by-products resulting from reductive cleavage of the cyclopropyl-ring and further reduction of
the ketone to the alcohol were observed. Low yields and difficulties in separating 66 from
desired 64 required further optimization. Attempts to avoid this issue by reducing the excess
of lithium resulted in incomplete reactions.
We reasoned that Li/NH3 reduction led to double charged anion Int 2, which should not
accept further electrons from the reaction medium (Scheme 2). However, when quenching
the reaction mixture, the slow addition of aqueous NH4Cl might result in a faster protonation
of Int 2 relatively to the quenching of the excess of electrons and 64 would then be produced
in a strongly reducing environment, allowing further reaction to the observed side-products
during the quenching procedure. Consequently, we proposed a sequential work-up: first
quenching of the electrons and then protonation of Int-2.
This article is protected by copyright. All rights reserved.

R
H
R
H
R
H
R
H
R
H
e-
H+
e-
e-
H
H
.
-
H
O
H
H
H
H
O
H
H
O
H
H
O
H
.
O
63
Int 1
Int 2
64
Int 3
R =
R
H
R
R
H
e-
+
H
2 H ,
e-
+
2 H ,
H
H
H
H
H
H
O
H
H
O
H
.
OH
67
66
Int 4
Scheme 2. Proposed reaction sequence leading to over reduction of 64.
Gratifyingly, quenching the reaction with single electron acceptor 4-methyl pentenone prior
to protonation with NH4Cl led to consistently high yields of 64 and avoided the formation of
over-reduced side products. Quenching of the reaction with benzene (a weaker electron
acceptor)69 prior to protonation, led to the exclusive formation of 66 (Table 11).
Table 11. Yield of products obtained from the reaction of 63 with Li/NH3/THF in dependence of reaction
quenching conditions
Observed reaction products
quenching conditions
64
66
67
Aqueous NH4Cl (aq.)
variable
14%
64%
0%
24-34%
20 eq. PhH, then NH4Cl (aq.)
10 eq. 4-methyl pentenone, then NH4Cl (aq.) 73-78%
4.2 Synthesis of C(7) methylated 24-epicastasterone and 24-epibrassinosteroids
The intermediate anion generated by reduction from 63 with solvated electrons could be also
quenched with an excess of MeI to produce the cyclopropyl derivative 68. No over-reduced
material was obtained in this reaction as methyl iodide also efficiently quenches the excess
of electrons (Scheme 3). After isomerization of the cyclopropyl group, the resulting isomers
69 and 70 were further converted and then separated to the corresponding 24-
epicastasterone and brassinosteroid derivatives 32, 33, 34. (Scheme 3).
This article is protected by copyright. All rights reserved.

PyrHC, LiBr, DMA
160°C, 6h
H
6 eq. Li/NH3/THF-78°C, 5min
H
H
H
H
H
H
H
O
H
quench:
8 eq. MeI
NH4Cl(aq)
H
O
H
H
O
H
H
O
H
(92%)
63
68
70 (25%)
69 (65%)
Sharpless asymmetric dihydroxylation
OH
OH
OH
OH
OH
H
H
OH
H
H
H
H
O
HO
HO
HO
HO
HO
HO
H
H
H
O
H
H
H
34 (28%)
O
O
32 (19%)
33 (73%)
TFAA, H2O2
CHCl3
Scheme 3. Synthesis of C(7) methylated 24-epicastasterone and 24-epibrassinosteroids
4.3 Synthetic modifications of the A-Ring
In order to scope the relative relevance of the C(2) and C(3) hydroxyl groups as well as their
stereochemical requirement for biological activity, mono-hydroxylated analogs were
prepared. 65 was treated with 9-BBN followed by oxidative work-up. However, competitive
reduction of the ketone was observed. The more hindered thexylborane allowed access of
the C(2) and C(3) mono-hydroxylated analogs 73 and 74 with -configuration, as
hydroboration took place exclusively from the less hindered -face, opposite to the C(10)
methyl group (Scheme 4). Sharpless asymmetric dihydroxylation of the side chain produced
the corresponding 24-epityphasterol 6 and the so far not reported C(3)-deoxy 24-
epicastaterone 22.
This article is protected by copyright. All rights reserved.

a) 3 eq. 9BBN
H
H
H
H
H
H
H
O
H
H
H
H
H
65
HO
72 (66%)
71 (13%)
H
H
H
OH
OH
BH
b)
2 eq.
H
H
H
H
HO
H
H
H
O
H
74 (13%)
73 (29%)
HO
H
H
O
c ) Sharpless asymmetric dihydroxylation
OH
OH
OH
H
OH
H
H
H
HO
H
H
H
O
H
HO
H
H
O
6
22
Scheme 4. Preparation of 24-epityphasterol 6 and C(2)-deoxy 24-epicastaterone 22.
This article is protected by copyright. All rights reserved.

H
a) AcOH, cat H2SO4, 95%
b) NaOH, MeOH, 60%
H
H
H
H
O
H
H
O
H
HO
75
64
H
e) DEAD, Ph3P, HCOOH, 66-95%
f) K2CO3, MeOH; 69%
d) DAST, Et3N HF, 5°C, 37%
c) Sharpless asymmetric dihydroxylation
OH
H
O
H
H
H
H
OH
H
H
H
H
O
H
H
O
H
37%
24%
77
F
HO
H
O
H
H
H
76
HO
H
12 (37%)
c) Sharpless asymmetric dihydroxylation
OH
OH
OH
H
OH
H
H
H
H
H
H
H
11 (50%)
6 (48%)
F
HO
H
H
O
O
Scheme 5. Preparation of 24-epiteasterone 12, its epimer 6, and its fluoro analog 11.
The-C(3)-OH 75 was conveniently prepared from cyclopropyl intermediate 64, with
catalytic H2SO4 in AcOH followed by hydrolysis of the produced acetate, as described
previously (Scheme 5).34,39 Asymmetric dihydroxylation of 75 provides 24-epi teasterone
(12). DAST-fluorination and Mitsunobu reaction, respectively, provide the corresponding
previously described inverted -configurated C(3)-fluoride34 and hydroxyl derivatives 76 and
77, which were further elaborated by asymmetric dihydroxylation reactions to provide 11 and
6.
The acid-catalyzed ring-opening of 64 with alcohols allowed the preparation of several other
C(3)-substituted analogs, as outlined in Scheme 6 and Scheme 7.
OH
OH
H
b) Sharpless
asymmetric
dihydroxylation
H
H
a) ROCH2CH2OH,
cat H2SO4
H
H
H
H
O
H
H
O
H
H
O
H
75- 89%
35%
O
O
H
H
64
RO
RO
14 R = Me
15 R = H
78 R = H
79 R = Me
Scheme 6. Preparation of C(3)-substituted analogs 14 and 15
This article is protected by copyright. All rights reserved.

As 75 (Scheme 7) was unstabe to basic conditions, alkylation of the C(3)-OH group was
accomplished through carbene insertion under acidic conditions to provide, upon
dihydroxylation of the side-chain, C(2)-OMe analogs 7 and 13. Likewise, Rh-catalysed
insertion of diazoethylacetate provided access to the corresponding α- and β-alkoxylated
ethyl acetates 80, 81, which were further derivatized to 8, 9, 10, 16, 17, 19, 21, 84, 85, 86
and 87 using classical chemical transformations.
OH
b) Sharpless asymmetric
dihydroxylation
OH
H
H
H
a) TMSCHN2, HBF4
H
H
H
H
O
H
H
O
H
H
O
H
O
HO
O
H
H
H
-OH 75
-OH 77
 OMe 87 (70%)
-OMe 88 (41%)
 OMe 13 (39%)
-OMe 7 (43%)
f) N2CHCO2Et, Rh(AcO)2
OH
c) 7N NH3, MeOH, 55°C, 48h
b) Sharpless asymmetric
dihydroxylation
H
H
OH
H
H
38%
H
O
H
O
H
O
H
O
O
O
H
-C(2)-O: 80 (81%)
-C(2)-O: 81 (70%)
O
H
d) NaOH, EtOH, THF
O
-C(2)-O: 82 (35%)
-C(2)-O: 18 (38%)
d) NaOH, EtOH, THF
OH
OH
H
H
H
H
H
O
H
R
O
H
O
H
HO
H
O
-C(2)-O: 83 (86%)
O
H
-C(2)-O; R = H2N: 10 (75%)
O
e) TMSCHN2
MeOH, 45%
-C(2)-O; R = OH: 16 (78%)
-C(2)-O; R = OMe: 17
-C(2)-O; R = OH: 8 (76%)
g) Me2C=C(NMe2)Cl, CH2Cl2
HNMe2 or NH3 or MeNH2
b) Sharpless asymmetric
dihydroxylation
-C(2)-O; R = Me2N: 19 (32%)
-C(2)-O; R = Me2N: 9 (33%)
-C(2)-O; R = H2N: 21 (57%)
H
H
-C(2)-O; R = Me2N: 84 (61%)
-C(2)-O; R = Me2N: 85 (crude)
-C(2)-O; R = H2N: 86 (39%)
-C(2)-O; R = MeHN: 87 (77%)
H
O
H
R
O
H
O
Scheme 7. Preparation of additional C(3)-substituted analogs
This article is protected by copyright. All rights reserved.