Design, synthesis and evaluation of antiproliferative activity of fluorinated betulinic acid

Article abstract of DOI:10.1016/j.bmc.2019.05.016

Betulinic acid (BA), a pentacyclic triterpenoid, exhibits broad spectrum antiproliferative activity, but generally with only modest potency. To improve BA's pharmacological properties, fluorine was introduced as a single atom at C-2, creating two diastereomers, or in a trifluoromethyl group at C-3. We evaluated the impact of these groups on antiproliferative activity against five human tumor cell lines. A racemic 2-F-BA (compound 6) showed significantly improved antiproliferative activity, while each diastereomer exhibited similar effects. We also demonstrated that 2-F-BA is a topoisomerase (Topo) I and IIα dual inhibitor in cell-based and cell-free assays. A hypothetical mode of binding to the Topo I-DNA suggested a difference between the hydrogen bonding of BA and 2-F-BA to DNA, which may account for the difference in bioactivity against Topo I.

Full text of DOI:10.1016/j.bmc.2019.05.016

Accepted Manuscript
Design, Synthesis and Evaluation of Antiproliferative Activity of Fluorinated
Betulinic Acid
Jizhen Li, Ling-Chu Chang, Kan-Yen Hsieh, Pei-Ling Hsu, Stephen J. Capuzzi,
Ying-Chao Zhang, Kang-Po Li, Susan L. Morris-Natschke, Masuo Goto, Kuo-
Hsiung Lee
PII:
S0968-0896(19)30523-1
https://doi.org/10.1016/j.bmc.2019.05.016
BMC 14903
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
1 April 2019
8 May 2019
10 May 2019
Please cite this article as: Li, J., Chang, L-C., Hsieh, K-Y., Hsu, P-L., Capuzzi, S.J., Zhang, Y-C., Li, K-P., Morris-
Natschke, S.L., Goto, M., Lee, K-H., Design, Synthesis and Evaluation of Antiproliferative Activity of Fluorinated
Betulinic Acid, Bioorganic & Medicinal Chemistry (2019), doi: https://doi.org/10.1016/j.bmc.2019.05.016
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Design, Synthesis and Evaluation of Antiproliferative Activity of Fluorinated Betulinic
Acid
a,b
c
b
b
c
Jizhen Li *, Ling-Chu Chang , Kan-Yen Hsieh , Pei-Ling Hsu , Stephen J. Capuzzi ,
a
b
b
b
Ying-Chao Zhang , Kang-Po Li , Susan L. Morris-Natschke , Masuo Goto *, Kuo-Hsiung
b,d
Lee *
a
Department of Organic Chemistry, College of Chemistry, Jilin University, 2519 Jiefang Road,
Changchun, 130023, China
b
Natural Products Research Laboratories, UNC Eshelman School of Pharmacy, University of
North Carolina, Chapel Hill, North Carolina, 27599, United States
c
Laboratory for Molecular Modeling, Division of Chemical Biology and Medicinal Chemistry,
UNC Eshelman School of Pharmacy, University of North Carolina, Chapel Hill, NC, 27599,
United States
d
Chinese Medicine Research and Development Center, China Medical University and Hospital,
Taichung, Taiwan
*
Corresponding authors. (J. Li) E-mail: ljz@jlu.edu.cn; (M. Goto) E-mail: goto@med.unc.edu;
(K. H. Lee) Tel. +1-919-962-0066, E-mail: khlee@unc.edu.
1

ABSTRACT
Betulinic acid (BA), a pentacyclic triterpenoid, exhibits broad spectrum antiproliferative
activity, but generally with only modest potency. To improve BA’s pharmacological properties,
fluorine was introduced as a single atom at C-2, creating two diastereomers, or in a trifluoromethyl
group at C-3. We evaluated the impact of these groups on antiproliferative activity against five
human tumor cell lines. A racemic 2-F-BA (compound 6) showed significantly improved
antiproliferative activity, while each diastereomer exhibited similar effects. We also demonstrated
that 2-F-BA is a topoisomerase (Topo) I and IIα dual inhibitor in cell-based and cell-free assays.
A hypothetical mode of binding to the Topo I-DNA suggested a difference between the hydrogen
bonding of BA and 2-F-BA to DNA, which may account for the difference in bioactivity against
Topo I.
Keywords:
Betulinic acid
Fluorine introduction
Topoisomerase
Antiproliferative activity
2

1
. Introduction
Betulinic acid (1, BA, 3β-hydroxy-lup-20(29)-en-28-oic acid, Figure 1) is a naturally
occurring lupane triterpenoid with anticancer, anti-HIV, antibacterial, antimalarial, and other
biological activities.1-3 In the development of new pharmacological candidates sourced from
natural products, much attention has been given to the anti-HIV and anticancer properties of 1.4, 5
For example, bevirimat (DSB, 3-O-(3′,3′-dimethylsuccinyl)-betulinic acid) was developed through
side chain modification at the C-3 position of 1. Bevirimat demonstrated good anti-HIV activity,
acted as a maturation inhibitor without showing cytotoxicity6-11 and reached phase II anti-AIDS
clinical trials. Moreover, further modifications of bevirimat at C-28 resulted in better anti-HIV
1
2
effects. With regard to its antiproliferative activity, compound 1 not only showed broad spectrum
cytotoxicity, but also good selectivity against melanoma cells and malignant tumors via induction
4
of apoptosis. Moreover, although 1 exhibits only modest cytotoxicity against tumor cells, it has a
good safety profile in normal cell lines. Indeed, it is safe for use even at a dose of 500 mg/kg body
1
3
weight. Even though various derivatives of 1, modified mainly at C-3 and C-28, have been
prepared by chemical methods and many of them have exhibited increased cytotoxicity,14-27 an
effective antitumor candidate has not been achieved.
Based on 1’s chemical structure, the pentacyclic triterpenoid scaffold is likely a major
contributor to the pharmacological properties. However, the hydrophobicity of the mainly carbon
and hydrogen combinations in the pentacyclic rings leads to poor water solubility, modest efficacy,
and diverse pharmacological and physicochemical characteristics that hamper the development of
2
8
1
-derivatives for anticancer clinical use. According to reported structure-activity relationship
(SAR) results, a hydrogen donor group at either C-3 or C-28 can significantly improve inhibition
3

of cell proliferation.29 Retention of a polar substituent at C-3 was also essential for the
pharmacological activities of pentacyclic triterpenes.30
We expect that the addition of simple side chains to produce branched 1-derivatives will not
produce an effective anticancer candidate of this compound type. However, one possible solution
to obtain an effective lead compound is to introduce a distinctive polar group into the skeleton.
Fluorine has strong electronegativity, small atomic size, and low polarization of the C-F bond, and
can mimic either a hydrogen or hydroxyl group.31-34 Roughly, 20% of the currently marketed
3
5
pharmaceuticals contain at least one fluorine atom. Incorporation of fluorine into a molecule is
3
6
an important medicinal chemistry approach in drug discovery, because of the atom’s beneficial
effects on conformation, pKa, intrinsic potency, membrane permeability, metabolic pathways, and
pharmacokinetic properties. Thus, we postulate that the introduction of fluorine into the skeleton
of 1 could be useful in modulating the biological activity or pharmaceutical profile.
From earlier results, the C-2 position, as well as the neighboring C-3 position, can play key
3
7
roles in modulating the biological activity. Indeed, derivatives of bevirimat with or without a 2-
3
8
F substitution exhibited large differences in anti-HIV activity. During other research studies, 2,2-
difluoro derivatives of dihydrobetulinic acid with increased cytotoxicity were synthesized.39, 40
Herein, we report the introduction of fluorine on the skeleton of 1 (Figure 1) at positions C-2 (as
F) and C-3 (as CF ) and evaluation of the effect on antiproliferative activity and interaction with
3
topoisomerases (Topos) I and IIα.
2
. Results and discussion
2
.1. Chemistry
Compound 1 (BA) and methyl ester 3 were prepared from betulonic acid 2 following reported
4
1
methods. Fluorine was then introduced at C-2 as shown in Scheme 1. The methyl ester 3 was
4

treated with lithium diisopropylamide (LDA) and N-fluorobenzenesulfonimide (NFSI) at -78 C,
resulting in the addition of one fluorine atom at C-2 to afford 4 in 43% yield via an electrophilic
substitution reaction. Subsequently, 2-F-substituted BA analogues 58 were prepared according
to established synthetic methods. The reduction of the carbonyl to a hydroxyl group was
accomplished easily by using NaBH to provide 5 in 81% yield. Hydrolysis of 5 with LiI and DMF
4
at 140 C gave compound 6 (2-F-BA) with a free C-28-CO H in 58% yield. Esterification of the
2
C-3 hydroxyl group with 2,2-dimethylsuccinic anhydride provided 7 and 8 in 54% and 50% yields
from 5 and 6, respectively.
For all C-2 fluoro-substituted analogues 48, the ratio of 2α-F to 2β-F was 7:3. The correct
orientations were confirmed by single crystal X-ray of compound 5 as shown in Figure 2.
Compound 6 is denoted as 2-F-BA in this paper.
Scheme 2 illustrates the synthesis of compounds with a CF group at C-3. When compound
3
+
-
3
was treated with CF Si(CH ) in the presence of Bu N F , followed by acidification with HCl
3 3 3 4
solution, a CF functional group was inserted directly at C-3 to provide 9 in 83% yield. Then, the
3
desired CF -substituted 10 was obtained in 38% yield by hydrolytic removal of the methyl ester
3
with LiI in DMF.
In drug development, it is necessary to determine the effect of chemical configuration on
biological activity. Accordingly, we prepared not only compound 6 (2-F-BA), but also its two α
and β diastereomers (2α-F-BA and 2β-F-BA) via the synthetic routes shown in Scheme 3. Benzyl
ester 11 was prepared from betulonic acid 2 and 4-chlorobenzyl bromide in 76% yield. Then,
fluorine was introduced at C-2 using NFSI reagent in 62% yield. The two α and β diastereomers,
1
3a and 13b, were separated by HPLC equipped with a preparative C18 column. The pre-
separation ratio of 13a and 13b was 3:2. Finally, the target compounds 6a and 6b were obtained
5

in 43% yield from 13a and 13b, respectively, via reduction and deprotection. In addition, the CF3
substituted compound 12 was also prepared in 83% yield by the treatment of intermediate benzyl
ester 11 with CF Si(CH ) . Briefly, the two key functional groups, F and CF , were successfully
3
3 3
3
introduced at C-2 and C-3 of the BA skeleton, providing 13 target compounds, 412, 13a, 13b,
6
a, 6b, for evaluation.
H
H
H
H
CO CH3
H
CO H
H
CO H
2
2
2
a
b
H
c
H
H
9
0%
75%
O
O
HO
H
H
H
3
2
1 (BA)
43%
H
H
H
H
CO CH3
H
H
2
CO CH3
CO H
2
2
F
F
F
b
d
H
H
H
8
1%
O
HO
58%
HO
H
H
H
4
5
6 (2-F-BA) (: =
:3)
7
e 54%
e 50%
H
H
H
H
CO CH3
CO H
2
2
F
F
O
O
H
H
H
HO C
O
HO C
O
2
2
H
7
8
Scheme 1. Synthetic route for fluorine introduction at C-2.
a) CH I, K CO , DMF, 40 C; (b) NaBH , EtOH/THF, rt; (c) NFSI, LDA, THF, -78 Crt; (d)
(
3
2
3
4
LiI, DMF, 140 C; (e) 2,2-dimethylsuccinic anhydride, DMAP, pyridine, 120 C.
6

H
H
H
H
H
H
CO CH3
CO CH3
CO2H
2
2
a
F C
b
F3C
3
H
H
H
O
83%
38%
H
HO
H
HO
H
3
9
10
Scheme 2. Synthetic route for CF group introduction at C-3.
3
+
-
(
a) CF Si(CH ) , Bu N F , THF, rt; then 1.0 M HCl; (b) LiI, DMF, 140 C.
3
3 3
4
H
H
H
H
CO H
H
CO CH Ph(4-Cl)
2 2
2
H
CO CH Ph(4-Cl)
2 2
H
a
H
F3C
HO
b
H
O
O
H
yield 76%
H
yield 83%
H
2
11
c yield 62%
12
H
H
H
H
CO CH Ph(4-Cl)
2 2
CO CH Ph(4-Cl)
F
2
2
F
H
H
d
+
O
O
H
H
1
3a
13b
yield 43%
yield 43%
d
H
H
H
CO H
H
CO H
2
2
F
F
H
H
H
HO
HO
H
6
a
6b
Scheme 3. Synthetic route to α and β diastereomers of 6 (2-F-BA)
7

+
-
(
a) 4-chlorobenzyl bromide, K CO , DMF, rt; (b) CF Si(CH ) , Bu N F , THF, 0 Crt; then 1.0
2
3
3
3 3
4
M HCl; (c) i) NFSI, LDA, THF, -78 Crt; ii) HPLC (C18 column, CH CN/H O); (d) i) NaBH ,
3
2
4
EtOH/THF, rt; ii) Pd/C, H , MeOH/THF, 40 C.
2
2
.2. Antiproliferative activity of fluorinated derivatives against human tumor cell line panel
Initially, the newly synthesized fluorinated compounds (410 and 12) together with lead
compound 1 were evaluated for antiproliferative activity using five human tumor cell lines, A549,
MDA-MB-231, MCF-7, KB, and P-glycoprotein (P-gp) overexpressing multidrug-resistant
(MDR) KB subline KB-VIN (Table 1). As seen from the results in Table 1, lead compound 1
showed only weak antiproliferative activity, while simple fluorination at C-2 (R = F) (compound
1
6
) led to significantly improved potency and perhaps greater affinity with the cellular target.
Furthermore, the pure diastereomers 6a (2α-F-BA) and 6b (2β-F-BA) showed similar
antiproliferative profiles against the tested tumor cell lines, indicating that the effect of fluorine is
diastereomer-independent. In contrast, compound 8 with fluorination at C-2 (R = F) but also
1
esterification at C-3 (R = 3,3-dimethylsuccinate) was inactive, indicating that a free hydroxyl
3
group at C-3 is required for the antiproliferative activity. Similarly, esterification of the carboxylic
acid (C-28) was detrimental, compare 6 (R = CO H) with 5 (R = CO Me). In fact, none of the
4
2
4
2
compounds with a C-28 ester (4, 5, 7, 9, 1113a,b) exhibited significant potency (IC >10 µM)
5
0
against any tested cell line. Interestingly, compound 10 with CF at C-3 (R = CF ) also exhibited
3
2
3
improved antiproliferative activity against all cell lines, except MDA-MB-231, compared with 1.
The lower potency against this cell line suggests that the cellular target expressed in MDA-MB-
2
31 may be slightly resistant to 10. These SAR studies indicate that two specific oxygenated
groups at C-3 (R = OH) and C-28 (R = CO H) are critical for the antiproliferative activity of 1
3
4
2
and its analogs. Fluorine at C-2 (R = F) (compound 6) significantly stimulated the antiproliferative
1
8

activity against all human tumor cell lines tested, including triple-negative breast cancer (TNBC)
and MDR subline.
Based on the screening results, compound 6 showed significant potency against both
hormone-dependent (MCF-7) and triple-negative (MDA-MB-231) breast cancer cell lines, and
thus, was selected for further evaluation against normal cells. Non-tumorigenic mammary gland
epithelial cells (MCF-10A) were treated with 6 for 72 h at various concentrations (Fig. 3A). MCF-
1
0A was well tolerant of 6 at a dose of 10 µM, the compound’s IC against MDA-MB-231 and
50
MCF-7. Surprisingly, 80% of MCF-10A cells survived at 30 µM, while more than 80% of MCF-
cells were killed. The effect of 6 on the colony formation of MCF-7 was also evaluated, and the
7
results demonstrated that 6 inhibited colony formation more efficiently than 1 (Fig. 3B, 3C). These
analyses indicate 6 probably targets the same protein as 1, but with different affinity. Thus,
fluorination of 1 increases the antiproliferative activity. Thus, compound 6 is useful not only for
use in mechanism of action studies to more clearly demonstrate the antiproliferative activity of 1,
but also for further development as an anticancer agent.
2
.3. Mechanism of action studies with 6
Initially, we confirmed that compound 6 induced apoptosis (Fig. 4A, 4B). Dose- and time-
dependent apoptotic changes were clearly seen in MCF-7 cells treated with 6, while such effects
were limited in cells treated with 40 µM of 1. Apoptosis was mediated by caspase-7, -9, and PARP
pathways, as confirmed by Western blot analysis (Fig. 4C). These observations indicate that
induction of apoptosis, not necrosis, is the final stage of the antiproliferative effects of 6.
We performed a gene expression profile as a primary step to explore the target of 6. The
alternation of gene expression caused by 6 was analyzed by cDNA microarray followed by
profiling compared with vehicle and 1 (Fig. 5). MCF-7 cells were treated with ether 1 or 6 for 24
9

h at 40 µM, and DMSO was used as a vehicle control. The cDNA was synthesized from total RNA
to probe a microarray chip. Heat map representation showed significant differences between cells
treated with each compound (Fig. 5A). Duplicate experiments (labeled as 1 and 2 for each
treatment with each compound) confirmed that the microarray results were reproducible. The
expression pattern of 1-treated cells was compared with that of DMSO, and the numbers of altered
genes are shown in Fig. 5B. Treatment with 6 dramatically increased the number of altered genes
(Fig. 5B). Cells treated with 6 showed 1119 up-regulated and 984 down-regulated genes, while 1-
treated cells showed 221 up- regulated and 408 down-regulated genes (Fig. 5B). These results
indicated that gene expression in MCF-7 was altered dramatically by treatment with 6. Gene
ontology enrichment analysis of the top ten biological processes suggested that compound 6 altered
the gene expression related to DNA replication and DNA repair (Fig. 5C). These analyses agreed
with the gene ontology enrichment analysis of the top ten molecular functions (Fig. 5D). Genes
encoding proteins related to double-stranded DNA binding, mismatch repair, DNA binding, and
DNA insertion or deletion binding were enriched. Thus, we postulated that bioactivity of 6 is
highly related to the DNA duplication (S) phase. The data were deposited in the Gene Expression
Omnibus database (Accession No: GSE90104).
2
.4. Identification of 6 as a DNA topoisomerase inhibitor
Because compound 1 shows modest cytotoxicity against tumor cell lines, numerous
synthetically modified related compounds have also been evaluated for antiproliferative activity
2
7
and mechanism of action (MOA). However, when the synthetic modifications lead to improved
bioactivity, the next question is whether the detected bioactivity is identical to the bioactivity
shown by unmodified 1. Indeed, modified 1-derivatives with aromatic-sulfonyl substitutions at the
C-3 (R ) and C-28 (R ) positions clearly showed inhibitory activity toward both DNA Topos I and
3
4
1
0

3
IIα, while the effects of 1 against both enzymes were unclear in a cell-free assay. Our newly
synthesized potent compound 6 and 1 differ only by the substitution of fluorine for one proton.
Thus, we initially assumed that 6 should show the same biological actions, but with improved
affinity to the cellular target/receptor of 1.
3
Based on our microarray analysis combined with previously reported results, compound 6 is
highly likely to target DNA Topos. Accordingly, we evaluated the ability of 6 to relax supercoiled
DNA (Fig. 6A) or kinetoplast DNA (kDNA) (Fig. 6B) in MCF-7 nuclei. Nuclear extracts were
prepared from MCF-7 after treatment with compounds subjected to the DNA relaxation assay. The
supercoiled DNA was relaxed and nicked by nuclear extract prepared from vehicle (DMSO) and
from 40 µM 1-treated cells, demonstrating that the enzyme, i.e., Topo I, was actively catalyzing
supercoiled DNA toward relaxation/nicking in the nucleus. In contrast, this catalytic activity was
inhibited by treatment with 6 at 40 µM, as well as camptothecin (CPT), a known Topo I inhibitor.
These results implied that 6 efficiently blocked catalytic enzyme(s), such as Topo I, toward
relaxation/nicking of supercoiled DNA in the cell nucleus. Similar effects were observed using
kDNA.
Subsequently, we evaluated 6 as a topoisomerase(s) inhibitor using purified recombinant
human Topos I (Fig. 6C) and II (Fig. 6D) in a cell-free assay. Supercoiled DNA was incubated
with Topo I in the presence of different doses of compound, DMSO (vehicle) or 1 µM CPT
(positive control) (Fig. 6C). Under this condition, Topo I catalyzed DNA relaxation/nicking was
observed in the presence of 1 at 10 µM, while catalytic activity was clearly inhibited at 30 µM.
When 6 was added, DNA relaxation and nicking were repressed in a dose response manner. In
fact, Topo I was partially inhibited by 6 at 3 µM comparable to CPT at 1 µM, and significant
inhibition was seen by 6 at 10 µM comparable to 1 at 30 µM. We also investigated the effect of 6
1
1

on the catalytic activity of Topo IIα (Fig. 6D). In this cell-free assay, DNA nicking was not
prevented by 1, even at 100 µM. In contrast, compound 6 clearly caused dose-responsive inhibition
from 30 to 100 µM comparable to the effect of VP-16 (etoposide) at 100 µM. From all assay
results, compound 6 directly inhibited catalytic activities of Topo I and IIα at 10 µM and 30 µM,
respectively, while 1 inhibited only Topo I at 30 µM. These observations demonstrate that 6 is a
dual inhibitor of Topos I and IIα. Thus, the effects against Topos were significantly improved by
simply substituting fluorine for a proton at C-2.
As previously reported, a docking model of 1 with Topo I showed that the C-28 carboxyl
group plays a critical binding role via H-bonding, while the C-3 hydroxyl group cannot interact
directly with amino acids on the enzyme. Other studies showed that, when aromatic-sulfonyl
moieties were introduced at C-3, C-28, or both positions, the association of the 1-derivative with
Topo I was stabilized by charge-transfer interactions of the π-aromatic cloud, forming H-bond
3
connections with the positively charged pocket on Topo I. We further found that the
antiproliferative activity of 6 was dramatically reduced when the 28-carboxylic acid (R = CO H)
4
2
was converted to a methyl ester (5, R = CO Me). Based on these findings, the C-3 hydroxyl and
4
2
C-28 carboxylic acid groups are critical to stabilize the association with Topos, and fluorine at C-
2
dramatically activates the involvement.
We have analyzed docking modes of multiple stereoisomers of 1 and 6 into the crystal
structure of human DNA Topo I (PBD: 1K4T). The top two best docked poses for 1 and 6 are
shown in Figure 7. In this model, the hydroxyl group at C-3 of 1 acts as a hydrogen bond donor
that interacts with the carboxylic acid moiety of Asp533 in Topo I. On the other hand, the hydroxyl
group at C-3 of 6 is in the opposite plane relative to 1 and makes a hydrogen bond with guanosine
1
2 of the DNA. This difference in hydrogen bonding may account for difference in Topo I activity.
1
2

Our hypothetical mode of binding of 6 suggested that the compound may inhibit Topo I when the
enzyme is associated with DNA. Therefore, 6 can inhibit Topo I activity more efficiently than 1.
Potential design and synthesis of 6 as a Topo I inhibitor are ongoing based on this docking model.
Studies have shown that the substitution of fluorine for proton protects a compound from
4
1
detoxification, such as P450-catalyzed oxidation in tumor cells. In our cell-free Topo assays, a
significant difference was clearly seen in the inhibitory effects between 1 and 6 on catalyzing
supercoiled DNA, strongly suggesting that fluorine at C-2 increased or stabilized the compound’s
association with Topos. However, this effect may not be caused by protection from detoxification
by P450.
Although two diastereomers (α:β = 7:3) form compound 6, both pure diastereomers (6a, 6b)
exhibited similar antiproliferative activity. Thus, we would conclude that the bioactivity against
Topos is diastereomer-independent.
3
. Conclusion
We synthesized several analogues of 1 with F or CF groups at C-2 or C-3, respectively, and
3
evaluated the new compounds for antiproliferative activity against five human tumor cell lines,
including a P-gp over expressing MDR subline. Compounds 6 (2-F) and 10 (3-CF ) showed
3
increased antiproliferative potency relative to 1. Because replacement of proton with fluorine at
C-2 created a new chiral center, the absolute chemical structure of 6 was determined by X-ray
crystallography of precursor 5 as a racemic compound (2α:β = 7:3). Its separate diastereomers (6a,
6
b) demonstrated similar antiproliferative profiles. Compound 6 was also demonstrated to be a
dual Topo I and IIα inhibitor. Docking studies showed that a hydrogen bond is formed between
the hydroxyl group at C-3 of 6 and guanosine 12 of DNA with Topo I.
1
3

4
. Experimental section
4
.1. Chemistry
All reagents were purchased from commercial sources and used without further purification,
1
13
19
unless otherwise indicated. All solvents were dried and newly distilled. The H, C and F NMR
spectra were obtained on a Bruker AVANCE-400 NMR spectrometer recorded at 400, 100 and
3
76 MHz, respectively, in CDCl with TMS as internal standard. All shifts were given in ppm. All
3
coupling constants (J values) were reported in Hertz (Hz). Column chromatography was performed
on silica gel 100200 or 200300 mesh. MS (ESI) data were measured on an Agilent
1
290/microTOF QII series LCMS/MS system. Single crystal X-ray diffraction data were collected
using a Bruker-AXS SMART APEX2 CCD diffractometer (Mo Kα, λ = 0.71073 Å ). The
characteristic data of compound 1 and methyl ester 3 were identical with those reported .42, 43
4
4
.2. Synthesis
.2.1. Methyl ester of 2-fluoro-betulonic acid 4.
Compound 3 (1.01 g, 2.15 mmol) was dissolved in anhydrous THF (10 mL) under nitrogen
protection and cooled to -78 °C with a dry ice-acetone solution. LDA [LiN(i-Pr) , 1.0 M/L) (2.15
2
mL, 2.15 mmol) was added slowly to the stirred reaction solution. After 1 h, NFSI (N-
fluorobenzene sulfonamide) (0.68 g, 2.15 mmol) dissolved in 23 mL THF was added dropwise
to the reaction mixture. After 1 h, the reaction was allowed to continue for another 3 h at rt. The
solvent was then removed under reduced pressure, and the residue was separated directly by silica
gel column chromatography (hexane/EtOAc as eluent) to give product 4 (0.52 g). White solid,
1
yield: 43%. H NMR (CDCl , 400 MHz, ppm) δ: 5.21-5.36 (ddd, 1H, J = 48Hz, J = 12Hz, J =
3
1
2
3
8
Hz, CH-F, in C2), 4.74 (s, 1H, C=CH, in C29), 4.61 (s, 1H, C=CH, in C29), 3.67 (s, 3H, CO CH ),
2 3
1
4

2
2
0
1
4
3
.97-3.01 (m, 1H, in C19), 2.45-2.47 (m, 1H, in C18), 2.22-2.25 (m, 2H, in C22), 1.88-1.91 (m,
H, in C1), 1.68 (s, 3H, CH , in C30), 1.14 (s, 3H, CH ), 1.07 (s, 3H, CH ), 0.97 (s, 3H, CH ),
3
3
3
3
1
3
.95 (s, 3H, CH ), 0.92 (s, 3H, CH ). C NMR (100 MHz, CDCl ) δ 210.31 (d, J = 12.4 Hz),
3
3
3
76.56 (s), 150.31 (s), 109.77 (s), 89.19 (d, J = 185.4 Hz), 56.78 (s), 56.46 (s), 51.29 (s), 50.25 (s),
9.37 (s), 48.62 (s), 47.11 (d, J = 16.6 Hz), 46.96 (s), 42.54 (s), 40.82 (s), 38.30 (d, J = 9.4 Hz),
8.12 (s), 36.91 (s), 33.89 (s), 32.09 (s), 30.54 (s), 29.58 (s), 25.25 (s), 24.74 (d, J = 1.4 Hz), 21.24
1
9
(
s), 21.19 (s), 19.31 (s), 19.06 (s), 16.88 (s), 16.09 (s), 14.60 (s). F NMR (CDCl , 376 MHz, ppm)
3
+
δ: -195.61. ESI-HRMS (m/z): calcd for C H FO [M+H] : 487.3509, found 487.3567.
3
1
47
3
4
.2.2. Methyl ester of 2-fluoro-betulinic acid 5.
Compound 4 (188 mg, 0.39 mmol) was dissolved in EtOH/THF (25 mL, 2:3 v/v) at rt. NaBH4
150 mg, 3.9 mmol) was added to the solution, which was stirred for 24 h. Then 3 N HCl was
(
added dropwise to a solution pH of 5~7. The organic solvent was removed under reduced pressure.
The mixture was extracted with EtOAc, washed with brine, and dried over anhydrous Na SO .
2
4
After concentration in vacuo, the residue was purified by silica gel column chromatography
1
(
hexane/EtOAc as eluent) to give product 5 (144 mg). White solid, yield: 81%. H NMR (CDCl ,
3
4
00 MHz, ppm) δ: 4.73 (s, 1H, C=CH, in C29), 4.61 (s, 1H, C=CH, in C29), 4.45-4.62 (m, 1H, J1
48Hz, CH-F, in C2), 3.67 (s, 3H, CO CH ), 3.20-3.26 (m, 1H, CH-OH, in C3), 2.92-2.96 (m,
=
2
3
1
H, in C19), 2.14-2.23 (m, 3H, in C18 and C22), 1.87-1.90 (m, 2H, in C1), 1.67 (s, 3H, CH , in
3
C30), 1.04 (s, 3H, CH ), 0.95 (s, 3H, CH ), 0.90 (s, 3H, CH ), 0.88 (s, 3H, CH ), 0.81 (s, 3H, CH ).
3
3
3
3
3
1
3
C NMR (100 MHz, CDCl ) δ 176.60 (s), 150.42 (s), 109.67 (s), 93.25 (d, J = 166.2 Hz), 81.16
3
(d, J = 15.7 Hz), 56.52 (s), 55.27 (s), 51.27 (s), 50.49 (s), 49.44 (s), 46.97 (s), 44.04 (d, J = 15.4
Hz), 42.47 (s), 40.76 (s), 39.65 (d, J = 7.5 Hz), 39.00 (d, J = 10.6 Hz), 38.17 (s), 36.94 (s), 34.13
s), 32.15 (s), 30.59 (s), 29.62 (s), 28.46 (s), 25.36 (s), 21.05 (s), 19.36 (s), 18.13 (s), 17.28 (s),
(
1
5

1
9
1
6.60 (s), 15.97 (s), 14.68 (s). F NMR (CDCl , 376 MHz, ppm) δ: -191.24. ESI-HRMS (m/z):
3
+
calcd for C H FO [M+H] : 489.3666, found 489.3710.
3
1
49
3
4
.2.3. 2-Fluoro-betulinic acid 6 (2-F-BA).
Compound 5 (0.12 g, 0.24 mmol) was dissolved in anhydrous DMF (8 mL), and LiI (0.33 g,
2
.4 mmol) was slowly added to the solution. The glass bottle containing the reaction mixture was
shaken by hand for 25 min. Then the reaction mixture was stirred at 140 °C for 48 h under argon
protection. The reaction mixture was poured into water, extracted with EtOAc, washed with brine,
dried over anhydrous Na SO , and concentrated in vacuo. The target product 6 (66 mg) was
2
4
isolated by silica gel column chromatography (hexane/EtOAc as eluent). White solid, yield: 58%.
1
H NMR (CDCl , 400 MHz, ppm) δ: 4.73 (s, 1H, C=CH, in C29), 4.61 (s, 1H, C=CH, in C29),
3
4
2
.49-4.62 (m, 1H, CH-F, in C2), 3.20-3.26 (q, 1H, J = 12 Hz, J = 4.4 Hz, CH-OH, in C3), 2.96-
1 2
.99 (m, 1H, in C19), 2.25-2.28 (m, 1H, in C18), 2.14-2.18 (m, 2H, in C22), 1.95-1.98 (m, 2H, in
C1), 1.68 (s, 3H, CH , in C30), 1.04 (s, 3H, CH ), 0.97 (s, 3H, CH ) , 0.92 (s, 3H, CH ), 0.88 (s,
3
3
3
3
1
3
3
9
4
3
H, CH ), 0.80 (s, 3H, CH ). C NMR (100 MHz, CDCl ) δ 181.04 (s), 150.24 (s), 109.81 (s),
3 3 3
3.23 (d, J = 166.4 Hz), 81.18 (d, J = 15.9 Hz), 56.31 (s), 55.26 (s), 50.44 (s), 49.24 (s), 46.90 (s),
4.02 (d, J = 16.1 Hz), 42.52 (s), 40.78 (s), 39.65 (d, J = 7.4 Hz), 39.00 (d, J = 10.6 Hz), 38.32 (s),
7.02 (s), 34.13 (s), 32.13 (s), 30.54 (s), 29.64 (s), 28.47 (s), 25.34 (s), 21.01 (s), 19.35 (s), 18.11
1
9
(
s), 17.28 (s), 16.59 (s), 16.04 (s), 14.67 (s). F NMR (CDCl , 376 MHz, ppm) δ: -191.22. ESI-
3
+
HRMS (m/z): calcd for C H FO [M+H] : 475.3509, found 475.3573.
3
0
47
3
4
.2.4. Methyl ester of 2-fluoro-3β-O-(3′,3′-dimethylsuccinyl)-betulinic acid 7.
Compound 5 (16 mg, 0.032 mmol), 2,2-dimethyl succinic anhydride (41 mg, 0.32 mmol),
DMAP (8 mg, 0.064 mmol), and pyridine (1.5 mL) were placed in a 10 mL glass tube and sealed.
1
6

The mixture was stirred at 120 C for about 12 h. The reaction mixture was then transferred into a
0 mL flask. Pyridine was removed under reduced pressure. HCl (3N, 10 mL) was added, the
5
mixture was extracted three times with EtOAc, and the organic layer washed with 3N HCl for a
second time. Then the organic layer was washed with brine, dried over anhydrous Na SO , and
2
4
concentrated in vacuo. The residue was purified by silica gel column chromatography
1
(
hexane/EtOAc as eluent) to provide target product 7 (11 mg). White solid, yield: 54%. H NMR
(
CDCl , 400 MHz, ppm) δ: 4.76 (d, 1H, J = 12Hz, CH-OCO), in C3), 4.73 (s, 1H, C=CH, in C29),
3
4
1
1
0
.60 (s, 1H, C=CH, in C29), 4.51-4.67 (m, 1H, CH-F, in C2), 3.66 (s, 3H, CO CH ), 2.92-2.98 (m,
2 3
H, in C19), 2.62-2.76 (2d, J = 16 Hz, CO-CH -C(CH ) ), 2.17-2.23 (m, 3H, in C18 and C22),
2
3 2
.87-1.90 (m, 2H, in C1), 1.67 (s, 3H, CH , in C30), 1.32 (s, 6H, 2CH , CO H-C(CH ) -CH -),
3
3
2
3 2
2
1
3
.90 (s, 3H, CH ), 0.88 (s, 3H, CH ), 0.86 (s, 3H, CH ), 0.83 (s, 3H, CH ), 0.81 (s, 3H, CH ). C
3
3
3
3
3
NMR (150 MHz, CDCl ) δ 176.50 (s), 176.23 (s), 170.83 (s), 150.27 (s), 109.60 (s), 89.47 (d, J =
3
1
4
3
73.2 Hz), 81.47 (d, J = 15.1 Hz), 56.38 (s), 54.89 (s), 51.13 (s), 50.28 (s), 49.28 (s), 46.85 (s),
4.40 (d, J = 17.8 Hz), 42.32 (s), 40.60 (s), 39.42 (d, J = 6.9 Hz), 38.55 (d, J = 10.2 Hz), 38.01 (s),
6.80 (s), 33.89 (s), 31.98 (s), 30.42 (s), 29.47 (s), 28.13 (s), 25.33 (s), 25.26 (s), 25.16 (s), 24.92
1
9
(
s), 20.93 (s), 19.17 (s), 18.02 (s), 17.39 (s), 17.14 (s), 15.81 (s), 14.51 (s). F NMR (CDCl 376
3,
+
MHz, ppm) δ: -190.57. ESI-HRMS (m/z): calcd for C H FO [M+H] : 617.4217, found 617.4169.
3
7
57
6
4
.2.5. 2-Fluoro-3β-O-(3′,3′-dimethylsuccinyl)-betulinic acid 8.
Compound 6 (20 mg, 0.0421 mmol) was converted to product 8 (12.5 mg) using the same
1
procedure as for the synthesis of 7 described above. White solid, yield: 50%. H NMR (CDCl ,
3
4
4
00 MHz, ppm) δ: 5.17-5.30 (d, 1H, J = 8Hz, CH-OCO, in C3), 4.81-4.95 (m, 1H, CH-F, in C2),
.74 (s, 1H, C=CH, in C29), 4.62 (s, 1H, C=CH, in C29), 2.92-2.98 (m, 1H, in C19), 2.67-2.77
(
m, CO-CH -C(CH ) ), 2.19-2.29 (m, 3H, in C18 and C22), 1.95-1.97 (m, 2H, in C1), 1.70 (s, 3H,
2 3 2
1
7

CH , in C30), 1.35 (s, 6H, 2CH , CO H-C(CH ) -CH -), 1.09 (s, 3H, CH ), 0.93 (s, 3H, CH ), 0.91
3
3
2
3 2
2
3
3
1
3
(
s, 3H, CH ), 0.89 (s, 3H, CH ), 0.88 (s, 3H, CH ). C NMR (150 MHz, CDCl ) δ 177.19 (s),
3 3 3 3
1
76.73 (s), 170.57 (s), 150.10 (s), 109.70 (s), 87.69 (d, J = 176.3 Hz), 77.26 (d, J = 15.9 Hz), 56.13
(
s), 54.50 (s), 50.11 (s), 49.07 (s), 46.71 (s), 44.11 (d, J = 17.8 Hz), 42.46 (s), 40.77 (s), 38.79 (d,
J = 10.3 Hz), 38.33 (d, J = 5.1 Hz), 38.05 (s), 36.91 (s), 33.90 (s), 31.97 (s), 30.37 (s), 29.53 (s),
7.52 (s), 25.59 (s), 25.35 (s), 25.19 (s), 24.78 (s), 21.36 (s), 20.71 (s), 19.21 (s), 17.65 (s), 16.99
2
1
9
(
s), 15.95 (s), 14.62 (s). F NMR (CDCl , 376 MHz, ppm) δ: -191.50. ESI-HRMS (m/z): calcd for
3
+
C H FO [M+H] : 603.3983, found 603.4034.
3
6
55
6
4
.2.6. Methyl ester of 3-trifluoromethyl-betulinic acid 9.
CF Si(CH ) (200 mg, 0.43 mmol) was added to the ester 3 (200 mg, 1.40 mmol) in anhydrous
3
3 3
+
-
THF (8.0 mL), and the mixture was cooled to 0 C in an ice bath. Bu N F (1M/THF, 0.56 mL,
4
0
.56 mmol) was added slowly to the stirring solution over 20 min. The reaction mixture continued
to stir for 24 h at rt. After removing the organic solvent under vacuum, HCl (1.0 N, 30 mL) was
added to the residue, and 2.0 mL THF was added. The mixture was kept for an additional 15 h at
rt. After extraction with EtOAc, the solution was washed with brine and dried over anhydrous
Na SO . The desired product (192 mg) was obtained by silica gel column chromatography
2
4
1
(
hexane/EtOAc as eluent). White solid, yield: 83%. H NMR (CDCl , 400 MHz, ppm) δ: 4.72 (s,
3
1
H, C=CH, in C29), 4.59 (s, 1H, C=CH, in C29), 3.66 (s, 3H, CO CH ), 2.92-2.97 (m, 1H, in
2 3
C19), 2.15-2.24 (m, 3H, in C18 and C22), 1.88-1.91 (m, 2H, in C1), 1.67 (s, 3H, CH , in C30),
3
1
3
1
.02 (s, 3H, CH ), 0.95 (s, 3H, CH ), 0.92 (s, 3H, CH ), 0.90 (s, 3H, CH ), 0.88 (s, 3H, CH ). C
3 3 3 3 3
NMR (100 MHz, CDCl ) δ 176.67 (s), 150.49 (s), 127.33 (q, J = 290.1 Hz), 109.60 (s), 77.87 (q,
3
J = 24.4 Hz), 56.56 (s), 51.53 (s), 51.26 (s), 50.63 (s), 49.46 (s), 46.98 (s), 42.39 (s), 40.86 (s),
4
0.66 (s), 38.24 (s), 36.95 (s), 36.90 (s), 35.61 (s), 34.47 (s), 32.17 (s), 30.60 (s), 29.71 (s), 26.64
1
8

1
9
(
s), 25.42 (s), 24.16 (s), 20.81 (s), 20.22 (s), 19.34 (s), 18.78 (s), 16.20 (s), 16.06 (s), 14.76 (s). F
+
NMR (CDCl , 376 MHz, ppm) δ: -68.59. ESI-HRMS (m/z): calcd for C H F O [M+H] :
3
32 49
3
3
5
39.3634, found 539.3669.
4
.2.7. 3-Trifluoromethyl-betulinic acid 10.
Compound 9 (22.7 mg, 0.0421 mmol) was converted to product 10 (8.3 mg) using the same
1
procedure for the synthesis of compound 6 described above. White solid, yield: 38%. H NMR
(
CDCl , 400 MHz, ppm) δ: 4.76 (s, 1H, C=CH, in C29), 4.63 (s, 1H, C=CH, in C29), 2.96-3.01
3
(m, 1H, in C19), 2.70 (s, 1H, OH), 2.20-2.24 (m, 3H, in C18 and C22), 1.95-1.97 (m, 2H, in C1),
1
.71 (s, 3H, CH , in C30), 1.03 (s, 3H, CH ), 1.00 (s, 3H, CH ), 0.97 (s, 3H, CH ), 0.93 (s, 3H,
3 3 3 3
1
3
CH ), 0.92 (s, 3H, CH ). C NMR (100 MHz, CDCl ) δ 181.97 (s), 150.31 (s), 127.12 (q, J =
3
3
3
2
4
98.9 Hz), 109.73 (s), 77.90 (q, J = 25.4 Hz), 56.38 (s), 50.58 (s), 49.24 (s), 46.91 (s), 42.43 (s),
0.86 (s), 40.67 (s), 38.38 (s), 37.05 (s), 36.91 (s), 35.59 (s), 34.47 (s), 32.14 (s), 30.55 (s), 29.71
(s), 26.63 (s), 25.40 (s), 25.31 (s), 24.17 (s), 20.77 (s), 20.20 (s), 19.33 (s), 18.77 (s), 16.20 (s),
1
9
1
6.10 (s), 14.75 (s). F NMR (CDCl , 376 MHz, ppm) δ: -68.58. ESI-HRMS (m/z): calcd for
3
+
C H F O [M+H] : 525.3477, found 525.3518.
3
1
47
3
3
4
.2.8. 4-Chlorobenzyl ester of betulonic acid 11.
To a suspension of betulonic acid (0.94 g, 2.07 mmol) and K CO (1.43 g, 10.35 mmol) in
2
3
amhydrous DMF (10 mL) was added 4-chloro-benzyl bromide (0.85 g, 4.14 mmol) at rt. The
reaction mixture was stirred for 24 h. Then the reaction mixture was poured into water and
extracted with CH Cl for three times. The organic layer was washed with brine, dried over
2
2
anhydrous Na SO , and concentrated under reduced pressure. The residue was purified by silica
2
4
gel column (petroleum ether/EtOAc as eluent) to give product 11 (0.90 g). White solid, yield: 76%.
1
9

1
H NMR (400 MHz, CDCl ) δ: 7.38-7.31 (m, 4H, Ph), 5.11 (2d, J = 12Hz, 2H, CH Ph), 4.75 (s,
3
2
1
H, C=CH, in C29), 4.63 (s, 1H, C=CH, in C29), 3.06-2.99 (m, 1H, in C19), 2.52-2.35 (m, 2H, in
C2), 1.70 (s, 3H, CH ), 1.08 (s, 3H, CH ), 1.04 (s, 3H, CH ), 0.97 (s, 3H, CH ), 0.93 (s, 3H, CH ),
3
3
3
3
3
1
3
0
1
4
.80 (s, 3H, CH ). C NMR (100 MHz, CDCl ) δ: 213.18 (s), 176.02 (s), 150.34 (s), 134.17 (s),
3 3
33.25 (s), 129.88 (s), 128.30 (s), 109.73 (s), 62.64 (s), 56.53 (s), 53.27 (s), 51.53 (s), 50.60 (s),
9.35 (s), 42.40 (s), 37.01 (s), 36.88 (s), 35.60 (s), 30.57 (s), 29.70 (s), 26.64 (s), 26.36 (s), 25.39
(s), 24.17 (s), 22.83 (s), 22.62 (s), 22.35 (s), 20.81 (s), 20.60 (s), 20.24 (s), 19.29 (s), 18.75 (s),
+
1
5
6.21 (s), 16.18 (s), 14.74 (s). ESI-HRMS (m/z): calcd for C H ClO [M+H] : 579.3599, found
3
7
52
3
79.3604.
.2.9. 4-Chlorobenzyl ester of 3-trifluoromethyl-betulinic acid 12.
4
Compound 12 (242 mg) was synthesized from ester 11 ( (260 mg, 0.45 mmol) following the
1
same synthesis procedure of 9 as described above. White solid, yield: 83%. H NMR (400 MHz,
CDCl ) δ: 7.37-7.31 (m, 4H, Ph), 5.14-5.06 (2d, J = 12Hz, 2H, CH Ph), 4.74 (s, 1H, C=CH, in
3
2
C29), 4.62 (s, 1H, C=CH, in C29), 3.05-2.98 (m, 1H, in C19), 2.29-2.26 (m, 1H, in C2), 2.20-2.11
(
m, 1H, in C2), 1.69 (s, 3H, CH ), 1.02 (s, 3H, CH ), 0.96 (s, 3H, CH ), 0.93 (s, 3H, CH ), 0.89 (s,
3 3 3 3
1
3
3
1
5
H, CH ), 0.77 (s, 3H, CH ). C NMR (100 MHz, CDCl ) δ: 176.00 (s), 150.36 (s), 135.48 (s),
3 3 3
34.74 (s), 134.18 (s), 129.99 (s), 128.80 (s), 109.68 (s), 92.53 (s), 65.39 (s), 57.95 (s), 56.48 (s),
4.92 (s), 53.92 (s), 52.79 (s), 49.87 (s), 49.29 (s), 47.30 (s), 42.44 (s), 40.63 (s), 39.59 (s), 38.30
(s), 36.88 (s), 35.28 (s), 34.11 (s), 32.05 (s), 30.56 (s), 29.62 (s), 26.60 (s), 26.45 (s), 25.49 (s),
1
9
2
1.01 (s), 19.60 (s), 19.31 (s), 14.58 (s), 13.68 (s). F NMR (376 MHz, CDCl ) δ: -68.60. ESI-
3
+
HRMS (m/z): calcd for C H ClF O [M+H] : 649.3557, found: 649.3549.
3
8
53
3
3
4
.2.10. 4-Chlorobenzyl ester of 2α-fluorobetulonic acid 13a and 2β-fluorobetulonic acid 13b.
2
0

Compound 11 (0.515 g, 0.89 mmol) was dissolved in anhydrous THF (10 mL) under nitrogen
protection and cooled to -78 °C with a dry ice-ethyl alcohol solution. Then the mixture was stirred
for 30 minutes at -78 °C. LDA [LiN(i-Pr) , 1.0 M) (1.07 mL, 1.07 mmol) was added slowly to the
2
stirred reaction solution. After 1 h, NFSI (0.336 g, 1.068 mmol) dissolved in 23 mL THF was
injected dropwise into the reaction mixture. After 1 h, the reaction was allowed to continue for
another 12 h at rt. The solvent was then removed under reduced pressure, and the residue was
separated directly by silica gel column chromatography (petroleum ether/EtOAc as eluent) to give
diastereomers 13a and 13b (total 0.33 g), which were separated by preparation HPLC (system:
SHIMADZU, Model: LC-20AP, column: Shim-pack GIST, dimensions: 20×150mm, material:
C18 5μm, solvents: acetonitrile:H O = 95:5). According to HPLC analysis (C18), the ratio of 13a
2
1
and 13b is 3:2. White solid, total yield: 62%. 13a: H NMR (400 MHz, CDCl ) δ: 7.38-7.32 (m,
3
4
H, Ph), 5.38-5.21 (ddd, 1H, J =48Hz, J =12Hz, J =8Hz, CH-F, in C2), 5.11 (2d, J =12 Hz, 2H,
1 2 3
CH Ph), 4.75 (s, 1H, C=CH, in C29), 4.63 (s, 1H, C=CH, in C29), 3.05-2.98 (m, 1H, in C19), 1.69
2
(
s, 3H, CH ), 1.15 (s, 3H, CH ), 1.14 (s, 3H, CH ), 1.09 (s, 3H, CH ), 0.95 (s, 3H, CH ), 0.79 (s,
3 3 3 3 3
1
3
3
H, CH ). C NMR (100 MHz, CDCl ) δ: 210.32 (d, J = 13.4 Hz), 175.61 (s), 150.18 (s), 134.96
3 3
(
s), 134.00 (s), 129.82 (s), 128.68 (s), 109.87 (s), 89.17 (d, J = 190.4 Hz), 64.92 (s), 56.76 (s),
6.40 (s), 50.20 (s), 49.31 (s), 48.62 (s), 47.17 (s), 47.00 (s), 46.93 (s), 42.52 (s), 40.75 (s), 38.31
s), 38.21 (s), 36.84 (s), 33.85 (s), 32.01 (s), 30.48 (s), 29.44 (s), 25.23 (s), 24.74 (s), 21.19 (s),
5
(
1
9
1
9.30 (s), 19.02 (s), 16.88 (s), 15.87 (s), 14.56 (s). F NMR (376 MHz, CDCl ) δ: -195.62. ESI-
3
+
1
HRMS (m/z): calcd for C H ClFNaO [M+Na] ：619.3325, found: 619.3313. 13b: H NMR
3
7
50
3
(
400 MHz, CDCl ) δ: 7.34-7.29 (m, 4H, Ph), 5.45-5.28 (ddd, 1H, J =52Hz, J =12Hz, J =8Hz, 1H,
3 1 2 3
CH-F, in C2), 5.08 (2d, J = 12 Hz, 2H, CH Ph), 4.73 (s, 1H, C=CH, in C29), 4.61 (s, 1H, C=CH,
2
in C29), 3.03-2.98 (m, 1H, in C19), 1.69 (s, 3H, CH ), 1.12 (s, 3H, CH ), 1.10 (s, 3H, CH ), 0.98
3
3
3
2
1

1
3
(
s, 3H, CH ), 0.78 (s, 3H, CH ), 0.73 (s, 3H, CH ). C NMR (100 MHz, CDCl ) δ: 213.37 (d, J =
3 3 3 3
1
2.5 Hz), 175.59 (s), 150.21 (s), 134.94 (s), 134.00 (s), 129.78 (s), 128.68 (s), 109.81 (s), 88.27
(
d, J = 190.8 Hz), 64.92 (s), 56.46 (s), 51.96 (s), 49.78 (s), 49.25 (s), 48.40 (d, J = 18.0 Hz), 46.85
s), 46.15 (s), 42.41 (s), 40.37 (s), 38.40 (s), 37.34 (s), 37.28 (s), 36.83 (s), 32.74 (s), 31.92 (s),
(
3
0.50 (s), 29.46 (s), 29.32 (s), 25.52 (s), 22.04 (s), 19.86 (s), 19.38 (s), 18.30 (s), 15.04 (s), 14.56
1
9
(
s). F NMR (376 MHz, CDCl ) δ: -191.10. ESI-HRMS (m/z): calcd for C H ClFNaO
3 37 50 3
+
[
M+Na] : 619.3325, found: 619.3313.
4
.2.11. 2-α-Fluoro-betulinic acid 6a (2-α-F-BA) and 2-β-fluoro-betulinic acid 6b (2-β-F-BA).
Compound 13a or 13b (120 mg, 0.20 mmol) was dissolved in EtOH/THF (25 mL,
1
0mL/15mL) at rt. NaBH (77 mg, 2.0 mmol) was added to the solution, which was stirred for 24
4
h. Then 3.0 N HCl was added dropwise to the solution until its pH data 5~6. The organic solvent
was removed under reduced pressure. The mixture was extracted with CH Cl , washed with brine,
2
2
dried over anhydrous MgSO . After concentration in vacuo, the residue was purified by silica gel
4
column chromatography (petroleum ether/EtOAc as eluent) to give white intermediate (76 mg).
To a solution of the intermediate (70 mg, 0.117 mmol) in THF/CH OH (10 mL/ mL) was added
3
1
0% Pd on carbon (14 mg) at room temperature. The mixture was subjected to 1 atm of H and
2
was heated at 40 C, then the reaction was monitored by TLC. After most of substrate was
consumed, the reaction mixture was filtered and the filtrate was concentrated in vacuo. Then the
solution was washed with brine, dried over anhydrous MgSO , and concentrated under reduced
4
pressure. The residue was purified by silica gel column chromatography (petroleum ether/EtOAc
as eluent) to give product 6a or 6b (34 mg) as a white solid. White solid, yield: 43% for 6a or 6b
1
respectively. 6a: H NMR (400 MHz, CDCl ) δ: 4.76 (s, 1H, C=CH, in C29), 4.64 (s, 1H, C=CH,
3
in C29), 4.57-4.50 (m, 1H, CH-F, in C2), 3.32-3.24 (m, 1H, CH-OH, in C3), 3.03 (m, 1H, in C19),
2
2

1
.71 (s, 3H, CH ), 1.08 (s, 3H, CH ), 1.00 (s, 3H, CH ), 0.95 (s, 3H, CH ), 0.92 (s, 3H, CH ), 0.83
3 3 3 3 3
1
3
(
s, 3H, CH ). C NMR (100 MHz, CDCl ) δ: 182.71 (s), 150.39 (s), 109.77, (s), 93.85 (d, J =
3
3
1
4
2
72.6 Hz), 77.86 (d, J = 18.0 Hz), 56.87 (s), 55.20 (s), 50.58 (s), 48.77 (s), 44.20 (s), 42.71 (s),
0.79 (s), 38.29 (d, J = 14.0 Hz), 37.44 (s), 36.62 (s), 34.22 (s), 32.08 (s), 29.78 (s), 29.60 (s),
9.23 (s), 26.84 (s), 23.03 (s), 22.78 (s), 21.04 (s), 18.00 (s), 16.83 (s), 16.55 (s), 16.07 (s), 14.71
1
9
(
s), 14.60 (s). F NMR (376 MHz, CDCl ) δ: -191.23. ESI-HRMS(m/z): calcd for C H FO
3 30 48 3
+
1
[
M+H] : 475.3582, found: 475.3580. 6b: H NMR (400 MHz, CDCl ) δ: 4.76 (s, 1H, C=CH, in
3
C29), 4.64 (s, 1H, C=CH, in C29), 4.57-4.50 (m, 1H, CH-F, in C2), 3.31-3.24 (m, 1H, CH-OH, in
C3), 3.03-2.99 (m, 1H, in C19), 1.71 (s, 3H, CH ), 1.08 (s, 3H, CH ), 1.00 (s, 3H, CH ), 0.95 (s,
3
3
3
1
3
3
1
H, CH ), 0.92 (s, 3H, CH ), 0.83 (s, 3H, CH ). C NMR (100 MHz, CDCl ) δ: 182.67 (s),
3 3 3 3
50.36 (s), 109.74, (s), 93.82 (d, J = 172.3 Hz), 77.82 (d, J = 18.0 Hz), 56.83 (s), 55.16 (s), 50.55
(s), 48.73 (s), 44.17 (s), 42.68 (s), 40.75 (s), 38.27 (d, J = 13.9 Hz), 37.41 (s), 36.58 (s), 34.19 (s),
3
2.05 (s), 29.74 (s), 29.57 (s), 29.20 (s), 26.80 (s), 23.00 (s), 22.74 (s), 21.01 (s), 17.97 (s), 16.80
1
9
(
s), 16.52 (s), 16.03 (s),14.67 (s), 14.56 (s). F NMR (376 MHz, CDCl )δ : -195.15. ESI-
3
+
HRMS(m/z): calcd for C H FO [M+H] : 475.3582, found: 475.3581.
3
0
48
3
4
.3. Cell lines and cell proliferation assay
Cell lines and antiproliferative activity assay protocols were described previously. In brief,
4
4
five human tumor cell lines and one normal mammary gland epithelial cell line (MCF-10A) were
used as follows: A-549 (lung carcinoma), KB (originally isolated from epidermoid carcinoma of
the nasopharynx), KB-VIN (vincristine-resistant P-gp overexpressing KB subline), and two breast
cancer cell lines MCF-7 (estrogen and progesterone receptor-positive and HER2 negative luminal
type) as well as MDA-MB-231 (triple negative basal type). These cell lines were obtained from
American Type Culture Collection (ATCC) (Manassas, VA, USA) or from the Lineberger
2
3

Comprehensive Cancer Center (UNC-CH). KB-VIN was a generous gift of Professor Y.-C. Cheng
Yale University). All cell lines were maintained and assayed in RPMI-1640 medium containing
mM L-glutamine and 25 mM HEPES (Corning), supplemented with 10% heat-inactivated fetal
(
2
bovine serum (Corning), 0.25 µg/mL amphotericin B, 100 µg/mL streptomycin, and 100 IU/mL
penicillin (Corning) in a humidified atmosphere in air containing 5% CO . Compounds were
2
prepared at 10 mM in DMSO and diluted with culture medium with the final DMSO concentration
3
≤
0.4% (v/v), a concentration with no effect on cell growth. The 411 × 10 cells/well were cultured
for 72 h with various concentrations of test compounds in a 96-well plate at 37 °C. The IC (µM)
5
0
values were determined from triplicated assays with duplicated samples using sulforhodamine B
assay. Cell viability assay shown in Figure 3 was determined by the 3-(4,5-dimethylthiazol-2-yl)-
2
,5-diphenyl-2H-tetrazolium bromide (MTT) method. In brief, cell viability was evaluated by the
reduction of MTT. Cells were cultured in 96-well plates and treated with the compound. After
treatment, MTT solution (1 mg/mL) was added to each well and plates were incubated for another
2
h. The resulting blue formazan product was dissolved in DMSO and detected
spectrophotometrically by absorbance at 570 (for MTT) and 630 (for background).
4
.4. Colony formation assay
Human breast cancer cell line MCF-7 and immortalized mammary epithelial cell line MCF-
0A were obtained from ATCC. MCF-7 cells were cultured in DMEM/F12 medium supplemented
1
with 10% (v/v) FBS, 100 units/ml penicillin, and 100 g/mL streptomycin. MCF-10A cells were
cultured in DMEM/F12 medium supplemented with 5% (v/v) horse serum, 0.5 g/ml
hydrocortisone, 10 g/mL insulin, 20 ng/mL epidermal growth factor, 0.1 g/mL cholera
enterotoxin, 2 mM L-glutamine, 100 units/mL penicillin, and 100 g/mL streptomycin. Cells were
maintained in a humidified incubator containing 5% CO . After exposure to compounds, cells were
2
2
4

collected and washed extensively with PBS. One thousand cells per well were seeded onto 6-well
plates and maintained in a 37 C, 5% CO incubator. Three weeks later, colonies were fixed with
2
formaldehyde (3.7%, v/v) and stained with crystal violet (0.5%, w/v), then counted.
4
.5. Apoptosis assay
Annexin V-FITC/propidium iodide double staining kit (BD Biosciences) was used according
to the manufacturer’s manual. MCF-7 cells were harvested after 24, 48, or 72 h treatment with
compounds and washed with ice-cold PBS. Cells were labeled with annexin V-FITC and
propidium iodide for 15 min at rt and analyzed by FACSCalibur flow cytometer (Becton
Dickinson).
4
.6. Western blotting
Harvested cells in PBS containing proteinase inhibitors (1 mM PMSF, 10 µg/mL each of
leupep0tin, aprotinin, and pepstatin A) and phosphatase inhibitors (1 mM Na VO and 1 mM NaF)
3
4
were sonicated. Whole cell lysates were separated by SDS-PAGE and transferred to Immobilon P
membrane (EMD Millipore). Protein-blotted membrane was incubated with primary antibody
followed by labeling with horseradish peroxidase (HRP)-conjugated secondary antibody (EMD
Millipore). HRP was detected by chemiluminescent substrate kit (EMD Millipore) and visualized
by the Luminescent image analyzer LAS4000 (Fuji Photo Film Co., Japan). Primary antibodies to
caspase-7, caspase-9, and PARP were purchased from Cell Signaling, and β-actin was from EMD
Millipore.
4
.7. Microarray
Gene expression profiling by microarray was performed at Phalanx Biotech (Hsinchu,
Taiwan) using Human Whole Genome OneArray (HOA6.2). The data were deposited in the Gene
Expression Omnibus database (Accession number: GSE90104,
2
5

https://www.ncbi.nlm.nih.gov/geo/). Total RNA was extracted using TRIzol method (Invitrogen,
Carlsbad, CA, USA). Cy5-labeled cDNA was prepared from 5 µg of total RNA subjected to
hybridization to a HOA 6.2 array. The fold changes of genes were calculated by Rosetta Resolver
7
.2 with error model adjusted by Amersham Pairwise Ratio Builder for signal comparison of
sample.
4
.8. Topoisomerase activity assay
Topo I and IIα activity tests were performed using the assay kits from TopoGen Inc. (Buena
Vista, CO, USA) according to the manufacturer’s instructions. For cell-based assay, nuclear
extracts from compound-treated MCF-7 cells were incubated with supercoiled DNA for 1 h or
kinetoplast DNA (kDNA) for 15 min at 37 C for Topo I or Topo IIα, respectively. For a cell-free
assay, compounds were added to the reaction mixture for a recombinant human Topo I with
supercoiled DNA or a recombinant human Topo IIα with kDNA. Reaction mixtures were separated
by 1% agarose gel in TAE buffer followed by staining with ethidium bromide and photographed
using Gel Doc XR (Bio-Rad). CPT (1 µM) or VP-16 (100 µM) was used as a control for Topo I
or Topo IIα, respectively. DMSO was used as a vehicle control.
4
.9. Molecular Docking
The lowest energy configurations at pH 7 were generated for BA and 2-F-BA with the Ligand
Preparation application (Schrödinger Release 2016-4: LigPrep, Schrödinger, LLC, New York,
NY, 2016). Next, the crystal structure of Human DNA Topoisomerase I bound to topotecan and
4
5
in a covalent complex with a 22 base pair DNA duplex (PBD: 1K4T) was downloaded from the
Protein Databank (http://www.rcsb.org/pdb/home/home.do)46 and prepared for docking as
follows. Briefly, bond orders were assigned, hydrogen atoms were added, selenomethionine
residues were converted to methionines, and missing side chains were filled with the Prime
2
6

application (Schrödinger Release 2016-4: Prime, Schrödinger, LLC, New York, NY, 2016). The
protein grid for docking was set by selecting the center of the co-crystal ligand (topotecan) in the
PDB structure. The Glide application was used for docking of all compounds into this complex
(Schrödinger Release 2016-4: Glide, Schrödinger, LLC, New York, NY, 2016). Flexible ligand
sampling was allowed with the XP (extra precision) scoring function.47
ASSOCIATED CONTENT
Supporting Information Available: High resolution mass spectral data for targets, HPLC
analysis of the target compounds, single crystal X-ray structure of compound 5, synthetic
1
1
13
1
13
procedures as well as H-NMR of compound 3, H-, C-NMR of compound 11, and H-, C-,
19F-NMR data of compounds 4-10, 12, 13a, 13b, 6a and 6b.
Note
The authors declare no competing financial interest.
Acknowledgments
This work was mainly supported by NIH grant from the National Cancer Institute (CA177584
awarded to K.H.L). This work was also supported in part by a grant from the IBM Junior Faculty
Development Award, and the Eshelman Institute for Innovation, Chapel Hill, North Carolina
(
awarded to M.G). Financial support from Department of Science & Technology of Jilin Province
Grant No. 20170414052GH for J.L) and National Natural Science Foundation of China (NSFC
(
No. 51573069 for J.L) were also gratefully acknowledged.
2
7

Abbreviations used
betulinic acid (BA), camptothecin (CPT), multidrug-resistant (MDR), structure-activity
relationship (SAR), triple-negative breast cancer (TNBC), topoisomerase (Topo)
30
20
29
1
9
H
H
2
8
1
H
CO H
2
H
CO2H
F
2
F
H
H
3
HO
HO
H
H
1
(Betulinic acid, BA)
F-substituted positions of BA skeleton
Figure 1. BA and F-substituted positions.
2
8

Figure 2. Single crystal X-ray of compound 5. The ratio of F in the C-2 α:β position was 7:3.
2
9