Enzymatic synthesis and bioactivity of estradiol derivative conjugates with different amino acids

Article abstract of DOI:10.1016/j.tet.2005.02.090

A series of N-protected amino acid-estradiol derivative conjugates have been synthesized by coupling of 17β-aminoestra-1,3,5 (10)-trien-3-ol (1) or 17β-hydrazonoestra-1,3,5 (10)-trien-3-ol (2) with different amino acids via the catalysis of subtilisin Carlsberg in organic solvent. Various factors, including the structure of amino acid residue, different N-protecting groups of amino acids, different esters of carboxyl group and water content of the reaction media that influence the efficiency of enzymatic reactions were systematically studied. In vitro biological activity studies revealed that the binding interactions between estradiol derivative conjugates and estrogen receptor can be affected by the properties of the conjugated amino acid, but the effects of the change in binding properties did not result in changes in biological activities in both MCF-7 and HeLa cell lines.

Full text of DOI:10.1016/j.tet.2005.02.090

Tetrahedron 61 (2005) 5933–5941
Enzymatic synthesis and bioactivity of estradiol derivative
conjugates with different amino acids
Ai-xin Yan,a,† Robbie Y. K. Chan,b Wai-Sum Lau,b Kin-sing Lee,b Man-Sau Wong,b
Guo-wen Xing,a Gui-ling Tiana and Yun-hua Yea,
*
aThe Key Laboratory of Bioorganic Chemistry and Molecular Engineering, Ministry of Education,
College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China
bOpen Laboratory of Chirotechnology and Department of Applied Biology and Chemical Technology,
The Hong Kong Polytechnic University, Kowloon, Hong Kong, China
Received 15 November 2004; revised 24 February 2005; accepted 28 February 2005
Abstract—A series of N-protected amino acid-estradiol derivative conjugates have been synthesized by coupling of 17b-aminoestra-1,3,5
(10)-trien-3-ol (1) or 17b-hydrazonoestra-1,3,5 (10)-trien-3-ol (2) with different amino acids via the catalysis of subtilisin Carlsberg in
organic solvent. Various factors, including the structure of amino acid residue, different N-protecting groups of amino acids, different esters
of carboxyl group and water content of the reaction media that influence the efficiency of enzymatic reactions were systematically studied. In
vitro biological activity studies revealed that the binding interactions between estradiol derivative conjugates and estrogen receptor can be
affected by the properties of the conjugated amino acid, but the effects of the change in binding properties did not result in changes in
biological activities in both MCF-7 and HeLa cell lines.
q 2005 Elsevier Ltd. All rights reserved.
1. Introduction
important class of molecules for drug development. A wide
variety of steroid derivatives conjugated with amino acids or
peptides have been prepared with the purposes to enhance
the oral antiarrhythmic activity,3 to help delivering the pro-
drugs to specific target tissues,4 as well as to achieve the
‘permissive action’.5
Amino acid-steroidal conjugates include such a class of
compounds in, which the amino acids are linked with
steroids through amide or ester bonds. This class of
conjugates covers a spectrum of important molecules in
nature, such as bufetoxin, a 3-arginyl derived steroid
product isolated from the Chinese hoptoad; cholyl glycine
and cholyl taurine, which exist in the bile of animals and
contains a glycyl or a taurinyl group at 17 position of
steroid; polymastiamide A,1 a tyrosine conjugated steroid
analog isolated from the Norwegian marine sponge
polymastia boletiformis as well as a starfish steroid in
which a 24-carboxylic acid functionality is linked via an
amide bond to D-cysteinolic acid.2 These compounds have
been found to play a diversity of critical roles in a large
number of organisms. Introduction of the amino acid or
peptide to the steroid backbone offers a combination of a
hydrophilic functional moiety as well as a hydrophobic
carrier in a same molecule and therefore represents as an
Previous studies on steroidal derivatives have indicated that
an unblocked 3-oxygenated group that should maintain as a
carbonyl or a hydroxyl group is required to exhibit high
binding affinities with their receptors.6,7 As a consequence,
we focused our interests on the modification of the other
functional groups on steroid matrix rather than the 3
position. In the previous paper, a series of 17b-amino or
17b-hydrazono steroidal derivatives have been synthesized
using a variety of chemical coupling reagents and the
relative binding affinities of the deprotected derivatives to
the estrogen receptor were examined by competitive
radioligand binding assay.8 Recently, we focus our attention
on enzymatic reactions in organic media owing to its
noticeable advantages over the chemical method, such as
enzymatic coupling usually can be performed under mild
conditions, displays high regio- and stereoselectivity,
requires minimal protection and is generally racemization
free. Moreover, most organic compounds are soluble in
organic solvents and enzymes are stable in organic media,9
Keywords: Conjugate; Enzymatic synthesis; Chemoselectivity; Steroids;
Estrogen receptor.
*
Corresponding author. Tel.: C86 10 62756354; fax: C86 10 62751708;
e-mail: yhye@pku.edu.cn
† Present address: Department of Biochemistry and Cell Biology, SUNY at
Stony Brook, NewYork 11794-5215, USA.
0040–4020/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2005.02.090

5934
A. Yan et al. / Tetrahedron 61 (2005) 5933–5941
therefore, the formation of peptide bond is favorable over
the hydrolysis of the product in organic media.10–12 A large
number of bioactive peptides have been synthesized via the
catalysis of different proteases in organic media and the
factors that influence the enzymatic peptide bond formation
have been investigated by us and others.12–15 However, at
present, proteases are limited in the synthetic application in
terms of their substrate specificity. In order to extend the
synthetic applications of proteases in organic media, a
number of approaches such as medium-engineering, sub-
strate engineering, enzyme engineering have been
developed.16–19 In an earlier study, we have extended the
nucleophilic substrates of proteases to amino alcohols and a
series of N-protected peptide alcohols have been generated
with the catalysis of subtilisin Carlsberg or a-chymo-
trypsin.15 In this paper, we present our exploration of
utilizing two steroid derivatives, 17 b-aminoestra-1, 3, 5
(10)-trien-3-ol 1 or 17 b-hydrazonoestra-1, 3, 5 (10)-trien-3-
ol 2, as the nucleophilic substrates of subtilisin Carlsberg to
synthesize a series of N-protected amino acid-estradiol
conjugates in an organic solvent. Various factors, which
influence the catalytic efficiency were studied and the
optimum reaction conditions were systematically investi-
gated. Our results demonstrate that steroid derivatives 1 and
2 can be applied as nucleophilic substrates for subtilisin in
organic media under certain reaction condition. The
synthetic strategy was briefly reported previously by us.20
Scheme 1. Schematic representative of the enzymatic coupling reaction
between a variety of N-protected amino acid esters and 1 or 2 carried out in
DMF in the presence of certain amount of essential water. (a), (b) Subtilisin
Carlsberg, DMF, 8% water (v/v); PZZ, Boc or Fmoc; RZMe, Et, tBu, Ph,
Bzl, Cam, or CH2CF3; AAZAla, Leu, Phe, Asp(OBzl), Asp, or Lys(Z).
reactions (P, AA, R etc.) were also shown in this scheme and
they were examined subsequently.
2.1. N-Protected amino acid trifluoroethyl ester is the
best acyl donor among various esters
Since the kinetic approach requires the use of amino acid
esters as the acyl donor substrates, we first analyzed the
influence of different esters in the acyl donor component for
their feasibility to be coupled with substrate 1 or 2 under the
catalysis of subtilisin Carlsberg. As shown in Table 1,
among the various esters we analyzed, the trifluoroethyl
ester of Z-protected alanine acts as the best acyl donor. The
t-butyl ester, phenyl ester, benzyl ester and Cam(Carbox-
amidomethyl)ester also displays activation effects, yet not
as good as trifluoroethyl ester, whereas methyl or ethyl ester
as acyl donor substrate does not yield any detectable
product. This result was expected because a strong electron
withdrawing group, trifluoroethyl, causes electron
deficiency on the carbonyl group. As a result, the free
energy required to form the acyl donor-enzyme inter-
mediate, which is believed as the prerequisite for the
subsequent aminolysis, is diminished so that the enzymatic
coupling becomes achievable.24 The IR spectrum of the
carbonyl group in the trifluoroethyl esters, which shows a
shift to high move number (1771 cmK1) compared with the
Incorporation of amino acids with 17b-amino or 17b-
hydrazono estra-1, 3, 5 (10)-trien-3-ol (1 and 2) resulted in a
series of molecules, which is analogous to estradiol, an
endogenous estrogen hormone. Previous efforts to develop
novel anti-estrogenic agents have resulted in several
estradiol derivatives, which contain the modifications on
7a or 11b position, such as ICI 164387, ICI 182780, RU
39411, RU 5866821 whereas no systematic studies on the 17
b-derivatives were examined so far. Thus, it is interesting to
measure the effects of the synthesized compounds on the
estrogen receptor (ER) positive cells with regard to their
agonistic or antagonistic activities. The MCF-7 and HeLa
cell lines22a,b were chosen in this paper to analyze the
estrogen like activities of these synthetic analogues as well
as to determine their abilities to activate ER dependent gene
transcription. In addition, the binding affinities of selected
synthesized compounds towards estrogen receptor were
determined by competitive receptor binding assay.
carbonyl group in an unactivated ester (1723–1747 cmK1
)
Table 1. Coupling of compound 1 or 2 with different Z-Ala-OR catalyzed
by subtilisin Carlsberg in DMF
2. Results and discussion
Acyl donor
Nucleophile
Product
Yield (%)
Z-Ala-OMe (6a)
Z-Ala-OEt (6b)
1
1
1
1
1
1
1
2
2
2
2
2
2
2
1a
1a
1a
1a
1a
1a
1a
2a
2a
2a
2a
2a
2a
2a
0
0
In order to generate a nucleophile moiety at the 17 position
of steroid backbone, we started with estrone and derived its
17-carbonyl group to amine (compound 1) or hydrazono
(compound 2) as described previously.8 Because these two
compounds are not the natural neucleophilic substrates of
proteases, subtilisin Carlsberg, which has been shown to
display flexible specificity for its P01 residue was utilized
and the kinetic approach, which is usually faster than
thermodynamic one10 was employed in our study. The
reaction was carried out in DMF in the presence of certain
amount of essential water23 as demonstrated in Scheme 1.
Various factors, which affect the efficiency of enzymatic
Z-Ala-OBut (6c)
Z-Ala-OPh (6d)
Z-Ala-OBzl (6e)
Z-Ala-OCam (6f)
Z-Ala-OCH2CF3 (6g)
Z-Ala-OMe (6a)
Z-Ala-OEt (6b)
Trace
18.7
Trace
32.2
35.0
0
0
Z-Ala-OBut (6c)
Z-Ala-OPh (6d)
Z-Ala-OBzl (6e)
Z-Ala-OCam (6f)
Z-Ala-OCH2CF3 (6g)
22.3
21.5
Trace
36.9
56.0

A. Yan et al. / Tetrahedron 61 (2005) 5933–5941
Table 2. Effect of different protecting groups on the yields of coupling products
5935
Entry
Acyl donor
Nucleophile
Product
Yield (%)
1
2
3
4
5
6
BocAlaOCH2CF3 (5b)
Z-AlaOCH2CF3 (6g)
FmocAlaOCH2CF3 (7a)
BocAlaOCH2CF3 (5b)
Z-AlaOCH2CF3 (6g)
FmocAlaOCH2CF3 (7a)
1
1
1
2
2
2
1b
1a
1c
2b
2a
2c
77.0
35.0
0
87.0
56.0
40.0
also supports the proposal. However, it is noticeable that
such activation effect of trifluoroethyl esters is not sufficient
to provide the free energy required to form the amide bond
without the catalysis of subtilisin. For an instance, we
carried out the control reaction, that is, the reaction between
compound 2 and Boc-Ala-OCH2CH3 (5b) without enzyme
added, the result was negative and we did not get the desired
product (2b). Our previous studies demonstrated that the
methyl or ethyl esters of corresponding amino acids or
peptides acted as attainable substrates for proteases
catalyzed native peptide bond formation in organic
media,14,25 but they are not acceptable in the current study
when non-natural substrates of proteases act as acyl
acceptors. This result implied that under special reaction
conditions, subtilisin could accommodate steroid derivative
as its acyl acceptor substrate to catalyze the non-peptidyl
amide bond formation in organic solvent, however, to
compensate for the unfavorable effects of the weak
nucleophilic substrates, strong electron withdrawing ester
is required in acyl donor substrate.
substrates. As demonstrated both in Tables 1 and 2, 2
behaves as a preferable nucleophile than 1 in the enzymatic
reactions we studied. This result is consistent with the
proposal that despite that reaction rate is determined mainly
by the specificity of the enzyme toward the acyl donor, a
specific binding of the nucleophile to the S0 subsite of the
protease is crucial for high yields.12 The preference of the
compound 2 as an acyl acceptor probably is the conse-
quence of both electronic and stereo effect of hydrozono
group. The electron accumulation on the nucleophilic
moiety in 2 due to the conjugating effect between two
nitrogen atoms and double bond of N]C in hydrazono
group obviously favors nucleophilic attack. On the other
hand, the nucleophilic moiety of amino group in 2 is not
directly attached to the bulk steroid backbone thereby is
more accessible to the acyl donor-enzyme complex in the
process of aminolysis compared with the amino group in 1.
2.4. The optimum reaction condition
It was believed that although most of the proteases are stable
in non-aqueous media, certain amount of water (essential
water) is required to maintain their catalytic conformation.9
On the other hand, however, excess of water in the system
will cause the undesired hydrolysis of the product. Hence,
there must be an optimum water content range within which
high yield of the coupling product could be achieved. We
subsequently examined the optimum water content of the
reaction using the condensation of Boc-Ala-OCH2CF3 (5b)
and 1 as model reaction and found that the optimum water
content is 8% (V/V) (Fig. 1). As expected, without water or
low water content in the system led to a complete depletion
of the catalytic activity of subtilisin whereas when the water
content of the system is higher than 8%, the yield of the
coupling product declined again due to the side effect of
2.2. Boc- is more favorable than Z- or Fmoc- moiety as
the N-protecting group in the acyl donor substrate in the
subtilisin catalyzed non-typical peptide bond formation
To avoid the unexpected coupling between amine and
carboxyl groups in a same molecule, which results in
oligomerization, the amine group in the amino acid ester
usually is blocked by selective protecting groups during the
coupling procedure. We then investigated the effect of
diverse N-protecting groups on the enzymatic reactions. As
revealed in Table 2, Boc is more favorable as the amino
protecting group for the acyl donor than Z or Fmoc both in
the presence of compound 1 and 2 as an acyl acceptor. This
observation is consistent with the suggestion that the S2
binding pocket of subtilisin prefers to a small, hydrophobic
moiety.26 Notably, under the circumstance when a poor
hydrophilic component 1 (see discussion below) acted as an
acyl acceptor and an unfavorable N-protecting group was
placed in the acyl donor (Fmoc), the capability of subtilisin
to catalyze the aminolysis was depleted or if any, was
undetectable (see entry 3 in Table 2).
2.3. Compound 2 is a better acyl acceptor than
Compound 1 in the subtilisin catalyzed non-peptidyl
amide bond formation
It was traditionally believed that proteases act on the amide
bond formed by natural amino acids, hydrolysis in
physiological condition and aminolysis in organic solvent.
However, we have shown that under particular condition,
subtilisin is able to accommodate non amino acid
component, steroid derivatives, 1 or 2 as its acyl acceptor
Figure 1. Effect of water content on the enzymatic reaction. Using the
coupling between 1 (0.2 mmol) and 5b (0.2 mmol) as the model reaction,
0.0, 0.06, 0.12, 0.18, 0.24, 0.30 and 0.36 mL phosphate buffer (KH2PO4-
wNa2HPO4, 50 mmol/L, pH 7.73) were added to certain amount of
anhydrous DMF to the total volume of 3 mL. The reaction was carried out
at 40 8C oil bath for 48 h and the product was purified by column
chromatography or preparative TLC.

5936
A. Yan et al. / Tetrahedron 61 (2005) 5933–5941
hydrolysis (Fig. 1). We also examined the effect of molar
ratio between two reactive components on the yield of
coupling product using the same model reaction and found
that the molar ratio of acyl donor versus acyl acceptor of 2:1
resulted in highest yield (Table 3).
residue was selected to be condensed with the nucleophiles
indicating that the regioselectivity of subtilisin was main-
tained under the conditions we studied. In the case of using
Boc-Lys(Z)-OCH2CF3 as an acyl donor, which bears a bulk
moiety in its P1 position, no coupling product was detected
even in the presence of a favorable acyl acceptor substrate 2.
Therefore, we speculate that the S1 binding pocket of
subtilisin favors a small, acidic moiety rather than a bulk,
basic one under the reaction conditions studied.
Table 3. Effect of molar ratio of acyl donor versus acyl acceptor on the
yield of the coupling product
Boc-Ala-OCH2CF3:
compound 1
1:2
1:1
2:1
2.6. In vitro estrogen like activity of the synthetic amino
acid-steroid conjugates
Product
Yield (%)
1b
83.4
1b
77.0
1b
88.0
Since the amino acid-steroid conjugates we synthesized
are analogs of estradiol, an endogenous hormone, we
removed the artificial N-protecting groups on amino acid
moieties (Scheme 2) and measured their abilities to
stimulate cell proliferation in ER positive human breast
cancer (MCF-7) cell line. As revealed in Figure 2, 10K9 M
of all synthetic conjugates, except for the lysine
conjugates (3h and 4h), significantly increased [3H]-
methylthymidine incorporation in MCF-7 cells (p!0.05).
To determine whether the up-regulation of thymidine
incorporation by the analogues is mediated via the ER,
cells were incubated with different analogues in the
presence or absence of the estrogen antagonist ICI
182780. ICI 182780 (10K6 M) completely blocked the
increase in [3H]-methylthymidine incorporation in MCF-7
by all synthetic conjugates and 17-b-estradiol, suggesting
that their effects on MCF-7 cell proliferation are mediated
through ER. Several conjugates with amino acids of
different hydrophilicity (Leu, Phe, Lys) were chosen to
test for their abilities to induce ER dependent activation
of luciferase gene transcription in HeLa cells. Figure 3
indicated that all chosen synthetic 17b-derivatives of
estradiol increased ERE-dependent luciferase activities
and that these activities can be blocked by co-treatment
with ICI 182780. However, it appears that the hydro-
philicity of amino acid in these conjugates did not affect
2.5. Characterization of the S1 binding pocket of
subtilisin under the reaction conditions studied
To study whether the non-native acyl acceptor affect the
substrate selectivity and regioselectivity of subtilisin in its
S1 position in the reactions studied, we synthesized a series
of amino acid trifluoroethyl esters, which bear a variety of
different side chains as the acyl donors to be reacted with 1
or 2. As listed in Table 4, among the various acyl donor
substrates, the one contains alanine residue, which bears a
small, alkyl side chain in the P1 position resulted in highest
yield (entries 1 and 7). Longer alkyl residue of the leucyl
side chain (entries 2 and 8) or bulk aromatic residue of the
phenylalaninyl side chain (entries 3 and 9) in this position
led to poor yield of the corresponding product. Particularly,
when the side chain of aspartic acid was blocked by benzyl
group, subtilisin was incapable to accept it as an acyl donor
substrate. However, when the side chain was deblocked, the
corresponding a-carboxyl active ester becomes to be
recognized by the enzyme. A moderate coupling yield was
obtained when 1 acts as the acyl donor and a significant
improvement of the yield was achieved when 2 acts as the
acyl donor, which is in agreement with the proposal that 2
acted as a better acyl acceptor than 1. It is noteworthy that
only a-carboxyl group rather than b-carboxyl group in Asp
Table 4. Coupling of various of acyl donors bearing different amino acid residues at P1 position with 1 or 2 under the catalysis of subtilisin Carlsberg
Entry
Acyl donor
Acyl acceptor
Product
Yield (%)
1
2
3
4
5
6
7
8
Boc-Ala-OCH2CF3 (5b)
Boc-Leu-OCH2CF3 (5d)
Boc-Phe-OCH2CF3 (5e)
Boc-Asp(OBzl)-OCH2CF3 (5f–1)
Boc-Asp-OCH2CF3 (5f)
Boc-Lys(Z)-OCH2CF3 (5h)
Boc-Ala-OCH2CF3 (5b)
Boc-Leu-OCH2CF3 (5d)
Boc-Phe-OCH2CF3 (5e)
Boc-Asp(OBzl)-OCH2CF3 (5f–1)
Boc-Asp-OCH2CF3 (5f)
Boc-Lys(Z)-OCH2CF3 (5h)
1
1
1
1
1
1
2
2
2
2
2
2
1b
1d
1e
—
1f
—
2b
2d
2e
—
2f
88.0
32.3
22.3
—
40.0
—
89.0
49.0
32.6
—
80.0
—
9
10
11
12
—
Scheme 2. Structures of a series of amino acid-steroid conjugates for bioactivity assay.

A. Yan et al. / Tetrahedron 61 (2005) 5933–5941
5937
Figure 2. HeLa cell-based ER-luciferase reporter gene system was treated with 10K9 M estradiol and different estradiol analogues with or without 10K6
M
pure estrogen-receptor antagonist. Arbitrary amount of light emitted was measured after substrate luciferin was added. Data represented meanGSEM, for nZ
3. Mean difference was compared by one-way ANOVA and followed by Tukey’s Post Test. The effect of each individual E2 and its analogues treatment was
significantly blocked by ICI 182780 (#p!0.001); *p!0.001 versus control.
their abilities to stimulate ERE-dependent luciferase
activities.
concentration of estradiol or its derivative conjugates
displaced the binding of radiolabel estradiol. The calculated
EC50 for 17-estradiol, 3g, 3d and 3e are 4 nM, 5.8 nM,
1.7 mM and 13.9 nM, respectively. The order of affinity for
ER is: E2O3gO3e[3d. These results are in agreement
with our previous studies8 that E2 has a much stronger
affinity toward ER than 3d does. These results also indicated
that estradiol derivative conjugated with Leu has a much
lower affinity for ER than conjugates with Gly or Phe. Thus,
the degree of bulkiness of the amino acid in estradiol
2.7. Difference in binding affinities of selected synthetic
amino acid-steroid conjugates towards estrogen
receptor
We then determined the binding affinities of several
conjugates towards human estrogen receptor alpha using
competitive binding assay. As shown in Figure 4, increasing
Figure 3. HeLa cell-based ER-luciferase reporter gene system was treated with 10K9 M estradiol and different estradiol analogues with or without 10K6
M
pure estrogen-receptor antagonist. Arbitrary amount of light emitted was measured after substrate luciferin was added. Data represented meanGSEM, for nZ
3. Mean difference was compared by one-way ANOVA and followed by Tukey’s Post Test. The effect of each individual E2 and its analogues treatment was
significantly blocked by ICI 182780 (#p!0.001); *p!0.001 versus control.

5938
A. Yan et al. / Tetrahedron 61 (2005) 5933–5941
application in routine organic synthesis. This problem
may be solved by modern molecular biology tools. Our
results provide helpful information for expanding the
proteases catalyzed amide bond formation to routine
organic synthesis in the future. The in vitro bioactivity
assay indicates that all of the synthesized steroid-amino acid
conjugates possess ER-dependent biological activities in
human MCF-7 cell lines. The competitive binding assay
indicated that the binding affinities of the conjugates for ER
were indeed influenced by the structure of the conjugate
amino acids. However, such hydrophobicity differences did
not alter the abilities of these conjugates to induce ERE-
dependent gene transcription in HeLa cells.
Figure 4. Competitive ER binding assay. Receptor binding assays were
carried out as described in methodology. The percent [3H]-estradiol bound
was calculated based on the equation: 100%(TB-NSB/TB). The data are the
mean of triplicate determinations, and the experiment was repeated twice.
The displacement of [3H]-estradiol increases as concentration of non-
radiolabeled estradiol or conjugated estradiol derivative analog increases.
4. Experimental
4.1. General
Subtilisin Carlsberg, (EC 3.4.21.14, specific activity
20.3 U/mg powder) was purchased from Shanghai Dong-
Feng Biochemical Technology Company. DCC, HOBt,
DMAP and trifluoroethanol were supplied by Aldrich
Chemical Company. All the amino acids were obtained
from Nova Biochemical Company. The estrone used in this
paper, was a generous gift from Prof. Guang-Dian Han in
Institute of Materia Medica, Chinese Academy of Medical
Sciences. Melting points were recorded by Yanaco micro
melting point apparatus given without correction;
Molecular masses were determined with VG-ZAB-HS,
derivative conjugates might affect the binding interactions
with ER.
Taken together, these results seem to indicate that even
though these compounds have different affinities for ER,
they have similar abilities to activate human breast cancer
cell proliferation as well as ERE-dependent transcription of
luciferase gene in HeLa cells. One explanation for this could
be the high sensitivity of both MCF-7 cell as well as
luciferase gene construct in HeLa cell to activated estrogen
receptor. In other word, even though their abilities to bind
ER are weaker than E2, all the compounds demonstrate
sufficient binding to ER for subsequent activation of MCF-7
cell growth and ERE-dependent gene transcription.
1
Bruker APEXeII; H was recorded with Bruker ARX400
(200 and 400 MHz); Elemental analysis were performed
with Carlo Erba 1106, Hezaeus CHN-Rapid, ST-02 or
Elementar Vario EL (Germany); Optical rotation was
analyzed by Perkin–Elmer 241 MC or 341 LC. Standard
abbreviations for amino acids and protecting groups are
according to the suggestion of IUPAC-IUB joint Commis-
sion on Biochemistry Nomenclature, Eur. J. Biochem. 1984,
138, 937.
3. Conclusion
4.2. Synthesis of 17b-aminoestra-1,3,5(10)-trien-3-ol (1)
and 17b-hydrazonoestra-1,3,5(10)-trien-3-ol (2)
In this paper, we have successfully extended the application
of subtilisin Carlsberg to the catalysis of non-peptidyl amide
bond formation. 17 b-aminoestra-1,3,5 (10)-trien-3-ol 1 or
17 b-hydrazonoestra-1, 3, 5 (10)-trien-3-ol 2 are compatible
nucleophilic substrates of subtilisin at P01 position in DMF
with 8% water (V/V). N-protected amino acid trifluoroethyl
ester is the best acyl donor among various esters, and Boc- is
more favorable than Z- or Fmoc- moiety as the N-protecting
group in the acyl donor substrate. Compound 2 is a better
acyl acceptor than Compound 1 in the subtilisin catalyzed
non-peptidyl amide bond formation. Probing the selectivity
of the S1 binding pocket of subtilisin indicates that this
subsite prefers to a small, acidic moiety rather than a bulk,
basic one under the reaction conditions in this study. Most
importantly, when Boc-Asp-OCH2CF3 (5f) as an acyl donor
in which the side chain b-carboxyl group was not protected,
only a-carboxyl group rather than b-carboxy group was
selected to be reacted with the nucleophiles indicating that
the high regioselectivity of the enzymatic synthesis in
organic media was maintained. However, the substrate
selectivity of the enzyme on its S1 position has led to
moderate yields in the cases of those amino acids, which
bear bulk side chains, partially limited its extensive
These two starting materials were prepared virtually as
described in literature.8 17b-aminoestra-1, 3, 5 (10)-trien-3-
ol (1), as white crystals, 52.0%, mp 235–237 8C; 17b-
hydrazonoestra-1, 3, 5 (10)-trien-3-ol (2). As white needle
crystals, 81.0%, mp 252–254 8C.
4.3. General procedure of synthesis of N-protected
amino acid trifluoroethyl esters27
DCC (5 mmol) and HOBt (2.5 mmol) were added to an ice
cooled solution of the N-protected amino acid (5 mmol) and
trifluoroethanol (5 mmol) in anhydrous DMF or CH2Cl2
(10 mL) followed by addition of DMAP (0.5 mmol) and
Et3N drop by drop till the pH becomes 8–9. The mixture was
stirred overnight at room temperature and the DMF or
CH2Cl2 was removed in vacuo. The residue was dissolved in
EtOAc (100 mL); washed with 5% citric acid, 5% NaHCO3
and saturated NaCl; and dried over MgSO4. The solvent was
removed and the residue was recrystallized from
EtOAc/hexane.

A. Yan et al. / Tetrahedron 61 (2005) 5933–5941
5939
4.3.1. Z-Ala-OMe (6a), Z-Ala-OEt (6b). These two
compounds were synthesized as described in literature.28
Z-Ala-OMe (6a). As white crystals, yield 70.0%, mp
44–46 8C, [a]2D2ZK34.5 (cZ0.5 in MeOH). [Lit.28, mp
46–47 8C, [a]2D0ZK36.5 (cZ1.0 in MeOH)]; Z-Ala-OEt
(6b). As light yellow crystals, yield 83.0%, mp 25–27 8C,
[a]2D2ZK30.2 (cZ1.0 in MeOH); IR nmax (cmK1) 3343,
3034, 1724, 1455, 1530. [Lit.28, oil, [a]D20ZK38.4 (cZ1.0
in MeOH)].
Calcd for C18H22O6NF3%: C, 53.33; H, 5.47; N, 3.46.
Found: C, 53.15; H, 5.64; N, 3.37. dH (CDCl3, 200 MHz);
1.45 (s, 9H, 3CH3), 2.80–3.21 (m, 2H, CH2), 4.39–4.71 (m,
3H, CH, CH2O), 5.13 (s, 2H, CH2O), 5.48 (d, JZ8.2 Hz,
1H, CONH), 7.26–7.39 (m, 5H, Ar-H); IR nmax (cmK1
)
3360, 3050, 1700, 1780, 1450, 1542.
4.3.12. Boc-Lys(Z)OCH2CF3 (5h). As white crystals, yield
85.0%, mp 85–86 8C, [a]2D2ZK13.5 (cZ1.0 in MeOH); MS
(FAB) m/zZ463.4 [MCH]C. Anal. Calcd for
C21H29O6N2F3%: C, 54.54; H, 6.32; N, 6.06. Found: C,
54.82; H, 6.36; N, 5.93. dH (CDCl3, 200 MHz); 1.43 (s, 9H,
3CH3), 1.50–1.79 (m, 6H, 3CH2), 3.15–3.21 (m, 2H, CH2),
4.34–4.69 (m, 3H, CH, CH2O), 4.68–4.81 (m, 2H, 2CONH),
4.3.2. Z-Ala-OBut (6c). As colorless oil, yield 71.0%,
[a]2D2ZK33.2 (cZ1.2 in MeOH); IR nmax (cmK1) 3341,
3066, 3034, 1722, 1455, 1527. [Lit.29, colorless oil, 58%,
[a]2D0ZK22.9 (cZ1.2 in EtOH).
5.10 (s, 2H, CH2O), 7.26–7.45 (m, 5H, Ar-H); IR nmax (cmK1
)
4.3.3. Z-Ala-OPh (6d). As white crystals, yield 42.0%, mp
94.5–95.5 8C, [a]D22ZK18.9 (cZ2.1 in CHCl3); IR nmax
(cmK1) 3356, 3062, 1690, 1700, 1771, 1485, 1527. [Lit.30,
mp 94–95 8C, [a]D25ZK18.8 (cZ2.0 in CHCl3)].
3360, 3050, 1695, 1778, 1450, 1550.
4.3.13. Boc-Asp-OCH2CF3 (5f). The standard hydrogen-
ization procedure was followed to generate 5f from
compound 5f–1. As white crystals, yield 84.0%, mp 120–
121 8C, [a]2D2ZK19.7 (cZ1.0 in EtOH); MS (FAB) m/zZ
316.2 [MCH]C.
4.3.4. Z-Ala-OBzl (6e). As white crystals, yield 83.0%, mp
)
34–36 8C, [a]D22ZK28.5 (cZ1.0 in MeOH); IR nmax (cmK1
3339, 3064, 1692, 1718, 1747, 1454, 1527. [Lit.31, mp
38–39 8C, [a]D25ZK31.4 (cZ1.0 in MeOH); IR nmax (cmK1
3338, 1747, 1688].
)
4.4. General procedure of enzymatic condensation
between various N-protected amino acid trifluoroethyl
esters and 1 or 2
4.3.5. Z-Ala-OCam (6f). As white crystals, yield 84.0%,
mp 65–68 8C, [a]D22ZK15.0 (cZ2.0 in DMF); IR nmax
(cmK1) 3351, 3202, 3065, 3034, 1691, 1737, 1751, 1771,
1455, 1536. [Lit.32, 80%, mp 60–64 8C, [a]2D2ZK13.0 (cZ
2.1 in DMF)].
To a mixture of P-AA-OCH2CF3 (0.2 mmol) and compound
1 or 2 (0.2 mmol) in anhydrous DMF (2.7 mL), 0.3 mL
phosphate buffer (KH2PO4wNa2HPO4, 50 mmol/L,
pH 7.73) and subtilisin Carlsberg (4 mg) were added. The
mixture was stirred 48 h at 40 8C oil bath and DMF was
subsequently removed in vacuo. The residue was dissolved
in EtOAc (60 mL), washed with 5% citric acid, 5%
NaHCO3 and saturated NaCl and dried over MgSO4. The
solvent was removed and the residue was recrystallized
from appropriate solvents or purified by column chroma-
tography or preparative TLC.
4.3.6. Z-Ala-OCH2CF3 (6g). As white crystals, yield
63.9%, mp 43–46 8C, [a]2D2ZK23.9 (cZ1.0 in MeOH);
MS (EI) m/zZ305.2[M]C. Anal. Calcd for C13H14NO4F3%:
C, 51.15; H, 4.62; N, 4.59. Found %: C, 51.55; H, 4.65; N,
4.72; IR nmax (cmK1) 3333, 3038, 1689, 1756, 1774, 1454,
1536.
4.3.7. Boc-Ala-OCH2CF3 (5b).27 As white crystals, yield
69.0%, mp 106–108 8C, [a]2D2ZK17.5 (cZ0.5 in MeOH).
4.4.1. 17b-(N-t-Butyloxycarbonyl-L-alanyl)aminoestra-1,
3, 5 (10)-trien-3-ol (1b). As white powder, mp 129–
131 8C, [a]1D9ZK31.0 (cZ0.5 in EtOH); MS (FAB) m/zZ
443.5 [MCH]C. Anal. Calcd for C26H38O4N2%: C, 70.56;
H, 8.65; N, 6.33. Found: C, 70.76; H, 8.86; N, 6.12.
4.3.8. Fmoc-Ala-OCH2CF3 (7a). As white crystals, yield
87.0%, mp 120–121 8C, [a]2D2ZK21.1 (cZ0.5 in MeOH).
Anal. Calcd for C20H18NO4F3%: C, 61.07; H, 4.61; N, 3.56.
Found: C, 61.24; H, 4.52; N, 3.37.
4.4.2. 17b-(N-t-Butyloxycarbonyl-L-alanyl)hydrazono-
estra-1, 3, 5 (10)-trien-3-ol (2b). As white powder, mp
174–176 8C, [a]1D9ZC120.5 (cZ0.4 in DMF); MS (FAB)
m/zZ456.6[MCH]C. Anal. Calcd for C26H37O4N3%: C,
68.54; H, 8.19; N, 9.22. Found: C, 68.71; H, 8.38; N, 8.91.
4.3.9. Boc-Leu-OCH2CF3 (5d). As colorless oil, yield
75.0%; MS (FAB) m/zZ314.3 [MCH]C, [a]D22ZK25.4
(cZ0.9 in DMF).33
4.4.3. 17b-(N-Benzyloxycarbonyl-L-alanyl)aminoestra-1,
3, 5 (10)-trien-3-ol (1a). As white powder,8 mp 108–
110 8C, [a]1D9ZK16.7 (cZ0.5 in EtOH); MS (FAB) m/zZ
477.6 [MCH]C.
4.3.10. Boc-Phe-OCH2CF3 (5e). As white crystals, yield
87.0%, mp 75–77 8C, [a]2D2ZK11.8 (cZ1.1 in MeOH); MS
(FAB) m/zZ348.4[MCH]C. Anal. Calcd for
C16H20O4NF3%: C, 55.33; H, 5.80; N, 4.03. Found: C,
55.69; H, 5.85; N, 3.94. dH (CDCl3, 200 MHz); 1.41 (s, 9H,
3CH3), 3.10 (m, 2H, CH2), 4.28–4.80 (m 3H, CH, CH2O),
4.91 (d, JZ8.0 Hz, 1H, CONH), 7.07–7.41 (m, 5H, Ar-H);
IR nmax (cmK1) 3350, 3055, 1692, 1768, 1452, 1540.
4.4.4. 17b-(N-Benzyloxycarbonyl-L-alanyl)hydrazono-
estra-1, 3, 5 (10)-trien-3-ol (2a). As white powder,8 mp
173–176 8C, [a]1D9ZC56.8 (cZ1.0 in DMF); MS (FAB)
m/zZ490.6 [MCH]C.
4.3.11. Boc-Asp(OBzl)-OCH2CF3 (5f–1). As white
crystals, yield 90.0%, mp 59–61 8C, [a]2D2ZK14.3 (cZ
1.0 in MeOH); MS (FAB) m/zZ406.2 [MCH]C. Anal.
4.4.5. 17b-(N-9-Fluorenylmethyloxycarbonyl-L-alanyl)-
hydrazonoestra-1, 3, 5 (10)-trien-3-ol (2c). As white

5940
A. Yan et al. / Tetrahedron 61 (2005) 5933–5941
powder, mp 147–150 8C, [a]1D9ZC95.2 (cZ0.3 in EtOH);
MS (FAB) m/zZ578.6 [MCH]C. Anal. Calcd for
C36H39O4N3$1/2H2O %: C, 73.70; H, 6.87; N, 7.16.
Found: C, 74.10; H, 6.93; N, 6.78.
bovine serum (FBS) for five days before the cells were
seeded in a 96-well microtitre plate at a density of 2!104
cells per well. Cells were then treated with 10K9 M estrogen
or its analogues amino acid estrogens with or without
estrogen receptor antagonist ICI 182, 780 at 10K6 M
(Tocris, UK). Drug treatments continued for three days
with the medium changed everyday.
4.4.6. 17b-(N-t-Butyloxycarbonyl-L-leucinyl)aminoestra-
1, 3, 5 (10)-trien-3-ol (1d). As white powder,8 mp
221–223 8C, [a]2D0ZK32.6 (cZ0.2 in EtOH); MS (FAB)
m/zZ485.6 [MCH]C.
4.6.2. [3H]-Methylthymidine incorporation assay. After
drug treatment, cells were washed with serum-free medium
once before adding serum-free medium containing
1 mCi/mL [3H]-methylthymidine (Amersham Pharmacia
Biotech, UK). Cells were returned to incubator for 2 h
before they were chilled at 4 8C for 15 min. Cells were then
washed with ice-cold phosphate-buffered saline (PBS)
twice. 200 mL ice-cold 0.75 M trichloroacetic acid (TCA)
per well was added and incubated at 4 8C further for 15 min.
TCA was then removed and 200 mL fresh TCA per well was
added. Plates were baked at 75 8C for 2 h. TCA from each
well was transferred into scintillation vial and its radioac-
tivity was measured by a scintillation counter.22a
4.4.7. 17b-(N-t-Butyloxycarbonyl-L-leucinyl)hydrazono-
estra-1, 3, 5 (10)-trien-3-ol (2d). As white powder,8 mp
129–131 8C, [a]2D0ZC57.8 (cZ0.8 in DMF); MS (FAB)
m/zZ498.5 [MCH]C.
4.4.8. 17b-(N-t-Butyloxycarbonyl-L-phenylalanyl)-
aminoestra-1, 3, 5 (10)-trien-3-ol (1e). As white powder,
mp 128–130 8C, [a]2D0ZC21.1 (cZ0.4 in EtOH); MS
(FAB) m/zZ519.6 [MCH]C. Anal. Calcd for
C32H42O4N2$H2O %: C, 71.61; H, 8.26; N, 5.22. Found:
C, 71.40; H, 8.02; N 5.00.
4.4.9. 17b-(N-t-Butyloxycarbonyl-L-phenylalanyl)-
hydrazonoestra-1, 3, 5 (10)-trien-3-ol (2e). As white
powder, mp 128–130 8C, [a]2D0ZC76.2 (cZ0.9 in EtOH);
MS (FAB) m/zZ532.6 [MCH]C. Anal. Calcd for
C32H41O4N3: C, 72.29; H, 7.77; N, 7.90. Found: C, 72.24;
H, 7.73; N, 7.88.
4.6.3. Luciferase assay. HeLa cell based human estrogen-
receptor (ER)-luciferase reporter cell line (a gift from Dr. H.
Gronemeyer, Strasbourg, France)22b was used to determine
the degree of activation of ER induced by estrogen and
peptidyl estrogen. Cells were cultured in phenol-red free
DMEM with 5% charcoal-stripped FBS for five days before
they were seeded at a density of 3!105 cells per well in a
96-well plate. Different drug treatments, as indicated in the
result section, were used to treat cells for 24 h before they
were washed with PBS and lysed in lysis buffer (Promega).
Cells were frozen at K80 8C and were thawed to room
temperature before the substrate luciferin (Promega) was
added to the cell lysate. The amount of light produced, as a
direct correlation of the degree of ER activation, was
measured by TR717 microplate luminometer (Applied
Biosystems, USA).
4.4.10. 17b-(N-t-Butyloxycarbonyl-L-aspartyl)amino-
estra-1, 3, 5 (10)-trien-3-ol (1f). As white powder, mp
162–164 8C, [a]2D0ZK8.9 (cZ0.9 in EtOH); MS (FAB)
m/zZ487.6 [MCH]C. Anal. Calcd for C27H38O6N2$H2O
%: C, 64.27; H, 7.99; N, 5.55. Found: C, 64.57; H, 7.77; N
5.32.
4.4.11. 17b-(N-t-Butyloxycarbonyl-L-aspartyl)hydrazono-
estra-1, 3, 5 (10)-trien-3-ol (2f). As white powder, mp
205–208 8C (dec), [a]2D0ZC74.7 (cZ1.0 in DMF); MS
(FAB) m/zZ500.6 [MCH]C. Anal. Calcd for
C27H37O6N3%: C, 64.91; H, 7.46; N, 8.41. Found: C, 65.13;
H, 7.69; N, 8.44.
4.6.4. ER binding assay. Escherichia. coli TOPP-3 cells
(Stratagene, La Jolla, CA) harboring the desired expression
plasmid pAF9His were grown in LB medium and induced
by IPTG.34 The cell lysate was used for the binding assay.
Total [3H]-estradiol binding (TB) to the hERa was
determined at different concentrations of non-radiolabeled
ligand in the presence of 10 nM [3H]-estradiol. Radioactive
counts were also measured in the presence of a 2000-fold
molar excess of non-radiolabeled estradiol attempting
displacement of [3H]-estradiol from hERa (NSB).
4.5. General procedure of deprotection of Boc by 50%
trifluoroacetic acid
The Boc-protected amino acid-steroidal conjugates
(1 mmol) were dissolved in 6 mL DCM: TFA (1:1), several
drops of anisole were added to quench the cations generated.
The mixture was stirred at room temperature and the
completion of the reaction was detected by TLC. The DCM
and TFA were removed in vacuo followed by the addition of
anhydrous ether to obtain precipitates of products.
4.6.5. Statistical analysis. Data are reported as the meanG
standard error of mean (SEM). Significance of differences
between group means was determined by analysis of
variance (ANOVA) when more than two groups were
compared. Group means differing by P values of 0.05 and
less are considered statistically significant.
4.6. In vitro binding affinity assay
4.6.1. Culture of estrogen receptor positive and negative
human breast cancer cell lines (MCF7 and MDA-MB-
231). MCF7 and MDA-MB-231 cells were routinely
cultured in Dulbecco’s Modified Eagle Medium (DMEM)
with 5% fetal bovine serum (FBS) at a confluency below
70.22 Culture medium was shifted to phenol-red free
DMEM supplemented with 1% charcoal-stripped fetal
Acknowledgements
The authors thank the Hong Kong Polytechnic University
for the financial support. We are grateful of Dr. Yun-sheng

A. Yan et al. / Tetrahedron 61 (2005) 5933–5941
5941
19. Liu, P.; Tian, G. L.; Lo, W. H.; Lee, K. S.; Ye, Y. H. Prep.
Biochem. Biotechnol. 2002, 32, 29–37.
Huang for the synthesis of the precursors of 4g, 3g, 3h and
4h by chemical method.
20. Yan, A. X.; Xing, G. W.; Ye, Y. H.; Tian, G. L.; Wong, M. S.;
Lee, K. S. Tetrahedron Lett. 2000, 41, 5379–5381.
21. William, H. C.; Jordan, V. C. In Molecular, Cellular, and
Systemic Mechanisms of Antiestrogen Action; Vedeckis,
W. V., Ed.; Hormones and Cancer; Birkhauser: Boston,
1996; pp 307–355.
References and notes
1. Kong, F.; Andersen, R. J. J. Org. Chem. 1993, 58, 6924–6927.
2. De Riccardis, F.; Minal, L.; Riccio, R. Tetrahedron Lett. 1993,
34, 4381–4384.
22. (a) Chan, R. Y. K.; Dong, Al.; Guo, D.; Wong, M. S. Regul.
J. Clin. Endo. Metab. 2002, 87, 3691–3695. (b) Chen, W. F.;
Huang, M. H.; Tsang, C. H.; Yang, M. S.; Wong, M. S.
Biochim. Biophys. Acta 2003, 1638, 187–196.
23. Zaks, A.; Klibanov, A. M. J. Biol. Chem. 1988, 263,
8017–8021.
3. Mokotoff, M.; Zhao, M.; Richard, J.; Marshall, R. J.; Winslow,
E.; Wong, L. K.; Liao, Q.-J. Steroids 1990, 55, 399–404.
4. Lapatsanls, L. J. Chem. Eng. Data 1980, 25, 287–289.
5. Wang, C.; Zhao, M.; Yang, J.; Peng, S. Q. Steroids 2001, 66,
811–815.
24. Miyazawa, T.; Tanaka, K.; Ensatsu, E.; Yanagihara, R.;
Yamada, T. Tetrahedron Lett. 1998, 39, 997–1000.
25. Xing, G. W.; Liu, D. J.; Ye, Y. H.; Ma, J. M. Tetrahedron Lett.
1999, 40, 1971–1974.
6. Tang, Z. M.; Chen, M. Y.; Zhang, Q. Acta Pharmacologica
Sinica 1983, 4, 186–191.
7. Jordan, V. C. Pharmacol. Rev. 1984, 36, 245–276.
8. Ye, Y.-H.; Huang, Y.-S.; Wang, Z.-Q.; Chen, S.-M.; Tian, Y.
Steroids 1993, 58, 35–39.
26. Matsubara, H.; Hagihara, B.; Nakai, M.; Komaki, T.;
Yonetani, T.; Okunuki, K. J. Biochem. 1958, 45, 251–258.
27. Liu, P.; Yan, A. X.; Ye, Y. H. Huaxue Tongbao 2001, No. 12,
777–780.
9. Zaks, A.; Klibanov, A. M. Science 1984, 224, 1249–1251.
10. Wong, C.-H.; Wang, K. T. Experitia 1991, 47, 1123–1129.
11. Yan, A. X.; Tian, G. L.; Ye, Y. H. Chin. J. Chem. Prog. 2001,
13, 203–208. (Chem. Abstr. 2001, 134, 237763).
12. Bordusa, F. Chem. Rev. 2002, 102, 4817–4867.
13. Xing, G. W.; Tian, G. L.; Ye, Y. H. J. Peptide Res. 1998, 52,
300–304.
28. Yamada, T.; Isono, N.; Inui, A.; Miyazawa, T.; Kuwata, S.;
Watanabe, H. Bull. Chem. Soc. Jpn. 1978, 51, 1897–1898.
29. Strazzolini, P.; Scuccato, M.; Giumanini, A. G. Tetrahedron
2000, 56, 3625–3633.
30. Barton, M. A.; Singh, B.; Fraga, E. J. Am. Chem. Soc. 1977,
99, 8491–8498.
14. Ye, Y. H.; Tian, G. L.; Xing, G. W.; Dai, D. C.; Chen, G.; Li,
C. X. Tetrahedron 1998, 54, 12585–12596.
31. Miyazawa, T.; Tanaka, K.; Ensatsu, E.; Yanagihara, R.;
Yamada, T. J. Chem. Soc., Perkin Trans. 1 2001, 87–94.
32. Martinex, J.; Laur, J.; Castro, B. Tetrahedron Lett. 1983, 24,
5219–5222.
15. Liu, P.; Tian, G. L.; Lee, K. S.; Wong, M. S.; Ye, Y. H.
Tetrahedron Lett. 2002, 43, 2423–2425.
16. Fernandez, M. M.; Margot, A. O.; Falender, C. A.; Blanch,
H. W.; Clark, D. S. Enzyme Microb. Technol. 1995, 17, 964.
17. Khumtaveeporn, K.; Ullmann, A.; Matsumoto, K.; Davis,
B. G.; Jones, J. B. Tetrahedron: Asymmetry 2001, 12,
249–261.
33. Park, O. J.; Jeon, G. J.; Yang, J. W. Enzyme Microb. Technol.
1999, 25, 455–462.
34. Nygaard, F. B.; Harlow, K. W. Protein Expr. Purif. 2001, 21,
500–509.
18. Witte, K.; Seitz, O.; Wong, C.-H. J. Am. Chem. Soc. 1998, 120,
1979–1982.