Synthesis and anti–tumour, immunomodulating activity of diosgenin and tigogenin conjugates

Article abstract of DOI:10.1016/j.jsbmb.2019.105573

A series of novel diosgenin (DSG) and tigogenin (TGG) derivatives with diosgenin or tigogenin steroid aglycons linked to levulinic and 3,4–dihydroxycinnamic acids, dipeptides and various amino acids by an ester bond at the C3–oxygen atom of the steroid skeleton has been synthesized. Diosgenyl esters have been prepared by an esterification reaction (DCC/DMAP) of diosgenin with the corresponding acids. All analogues have been evaluated in vitro for their antiproliferative profile against cancer cell lines (MCF–7, MDA–MB–231, PC–3) and human umbilical vein endothelial cells (HUVEC). Analogue2c (l–serine derivative of TGG), the best representative of the series showed IC₅₀ of 1.5 μM (MCF–7), and induced apoptosis in MCF–7 by activating caspase–3/7. The immunomodulatory properties of six synthesized analogues have been determined by examining their effects on the expression of cytokine genes essential for the functioning of the human immune system (IL–1, IL–4, IL–10, IL–12 and TNF–α). Biological evaluation has revealed that new compounds 4c and 16a do not induce the expression of pro–inflammatory cytokines in THP–1 cells after the lipopolysaccharide (LPS) stimulation. They also stimulate the expression of anti–inflammatory IL–10 that acts stronger than diosgenin itself. An in silico ADME properties(absorption, distribution, metabolism, excretion) study was also performed to predict the pharmacokinetic profile of the synthesized compounds. To shed light on the molecular interactions between the synthesized compounds and the glucocorticoid receptor and the estrogen receptor, 2c, 4c and 16a compounds were docked into the active binding sites of these receptors. The in silico and in vitro data suggested that this new group of compounds might be considered as a promising scaffold for further modification of more potent and selective anticancer and immunomodulatory agents.

Full text of DOI:10.1016/j.jsbmb.2019.105573

Journal Pre-proof
Synthesis and anti–tumour, immunomodulating activity of diosgenin and
tigogenin conjugates
´
´
O. Michalak, P. Krzeczynski, M. Cieslak, P. Cmoch, M. Cybulski, K.
´
´
´
Krolewska-Golinska, J. Kaz´mierczak-Baranska, B. Trzaskowski, K.
Ostrowska
PII:
S0960-0760(19)30352-8
DOI:
https://doi.org/10.1016/j.jsbmb.2019.105573
SBMB 105573
Reference:
To appear in:
Journal of Steroid Biochemistry and Molecular Biology
Received Date:
Revised Date:
Accepted Date:
13 June 2019
20 December 2019
23 December 2019
´
´
Please cite this article as: Michalak O, Krzeczynski P, Cieslak M, Cmoch P, Cybulski M,
´
´
´
Krolewska-Golinska K, Kaz´mierczak-Baranska J, Trzaskowski B, Ostrowska K, Synthesis and
anti–tumour, immunomodulating activity of diosgenin and tigogenin conjugates, Journal of
Steroid Biochemistry and Molecular Biology (2020),
doi: https://doi.org/10.1016/j.jsbmb.2019.105573
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. 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.
© 2019 Published by Elsevier.

Synthesis and anti–tumour, immunomodulating activity of diosgenin
and tigogenin conjugates
O. Michalaka*, P. Krzeczyńskia, M. Cieślakb, P. Cmochc, M. Cybulskia, K. Królewska–Golińskab, J.
Kaźmierczak–Barańskab, B. Trzaskowskid, K. Ostrowskae
a Łukasiewicz Research Network–Pharmaceutical Research Institute, 8 Rydygiera Str., 01–793 Warsaw, Poland
b Department of Bioorganic Chemistry, Centre of Molecular and Macromolecular Studies, Polish Academy of Sciences,
112 Sienkiewicza Str., 90–363 Łódź, Poland
c Institute of Organic Chemistry, Polish Academy of Sciences, 44/52 Kasprzaka Str., 01–224, Warsaw, Poland
d
Chemical and Biological Systems Simulation Lab, Center of New Technologies, University of Warsaw, 2C Banacha Str.,
02–097 Warsaw, Poland
e Department of Organic Chemistry, Faculty of Pharmacy, Medical University of Warsaw, 1 Banacha Str., 02–097 Warsaw,
Poland
*Correspondence: Łukasiewicz Research Network–Pharmaceutical Research Institute, 8 Rydygiera Str.,
01–793 Warsaw, Poland; o.michalak@ifarm.eu; Tel. +48 22 456 3925
Graphical abstract
Highlights
1


A set of diosgenin and tigogenin derivatives substituted with various amino acids,
dipeptides or levulinic and 3,4–dihydroxycinnamic acid were synthesized
Analogue 2c (L–serine in position 3 of tigogenin) showed the highest activity against the breast
cancer cell line (MCF–7) and its affinity profile to the active site of the estrogen receptor (ER)
was confirmed by molecular docking



A diosgenin derivative with caffeic acid 16a and the analogue of tigogenin with glutamic acid
4c


exhibited preferred immunomodulatory effects
A strong binding interaction of compounds 4c and 16a (Ki=0.23 pM and Ki=1.14 pM) with
the active site of the glucocorticoid receptor (GR) was estimated using molecular docking
Abstract:
A series of novel diosgenin (DSG) and tigogenin (TGG) derivatives with diosgenin or tigogenin
steroid aglycons linked to levulinic and 3,4–dihydroxycinnamic acids, dipeptides and various amino
acids by an ester bond at the C3–oxygen atom of the steroid skeleton has been synthesized. Diosgenyl
esters have been prepared by an esterification reaction (DCC/DMAP) of diosgenin with the
corresponding acids.
All analogues have been evaluated in vitro for their antiproliferative profile against cancer cell
lines (MCF–7, MDA–MB–231, PC–3) and human umbilical vein endothelial cells (HUVEC). Analogue
2c (L–serine derivative of TGG), the best representative of the series showed IC50 of 1.5 µM (MCF–7),
and induced apoptosis in MCF–7 by activating caspase–3/7.
The immunomodulatory properties of six synthesized analogues have been determined by
examining their effects on the expression of cytokine genes essential for the functioning of the human
immune system (IL–1, IL–4, IL–10, IL–12 and TNF–α). Biological evaluation has revealed that new
compounds 4c and 16a do not induce the expression of pro–inflammatory cytokines in THP–1 cells after
the lipopolysaccharide (LPS) stimulation. They also stimulate the expression of anti–inflammatory IL–
10 that acts stronger than diosgenin itself.
An in silico ADME (properties, absorption, distribution, metabolism, excretion) study was also
performed to predict the pharmacokinetic profile of the synthesized compounds. To shed light on the
molecular interactions between the synthesized compounds and the glucocorticoid receptor and the
estrogen receptor, 2c, 4c and 16a compounds were docked into the active binding sites of these receptors.
The in silico and in vitro data suggested that this new group of compounds might be considered
as a promising scaffold for further modification of more potent and selective anticancer and
immunomodulatory agents.
Keywords
Diosgenin, Tigogenin, Inflammation, Cancer, Anti–tumor, Docking
1. Introduction
Due to their interesting biological properties, natural compounds have recently attracted much
attention of both academics and pharmaceutical industry. This also applies to steroidal saponins which
are not only basic precursors for the synthesis of the sex hormones and corticosteroids in the
pharmaceutical industry, but also display various types of pharmacological activity, such as
antibacterial, anti–inflammatory, anticancer and hypocholesterolemic [1].They are therefore expected
to become key building blocks of new drugs with different signal transduction properties.
2

Diosgenin and tigogenin belong to the group of steroidal saponins. Diosgenin is present mainly
in the Dioscoreaceae plant family, but also in some species of Solanaceae and Fabaceae families in
combination with various sugars in the glycosides form, such as dioscin and gracilin. Tigogenin is
isolated from Yucca (Agavaceae) [1], but its triglycoside is isolated from the genus Hosta and
Tribulusplants [2] (Fig.1).
Fig.1. Diosgenin (DSG), tigogenin (TGG) and their natural glycosides:
dioscin (DSC), gracilin (GRL) and tigogenin natural triglicoside (TNT)
Diosgenin and diosgenyl glycosides display, among others, antiviral, anti–inflammatory,
antidiabetic and anticancer activity [3], while tigogenin triglycoside has been widely used in China as a
traditional medicine against various diseases [4]. In vitro assays have revealed that tigogenin has a very
weak anti–proliferative activity [5], but tigogenin triglycoside and its three derivatives bearing different
carbohydrate moieties exhibit a potent cytostatic activity against human promyelocytic leukemia HL–
60 cells and human epithelial cervical cancer cells [2]. Guo–Long L. et al. have synthesized a series of
tigogenin neoglycosides and their in vitro antitumor activity against five human cancer cell lines has
been evaluated. Some tigogenin neoglycosides have displayed enhanced antitumor activity against one
or more human cancer cell lines [6].
Diosgenin’s anticancer potential has been extensively studied both in vitro and in vivo [7]. The
in vitro studies have been performed on various human cancer cell lines to confirm the influence of
diosgenin on various molecular targets critical for carcinogenesis. It has been shown that diosgenin
inhibited the proliferation of HT–29 human colon cancer cells and induced apoptosis partly by
modulating the expression of bcl–2 and caspase–3 in vitro [8]. In the HCT–116 human colon cancer
cells diosgenin has modulated the expression of 3–hydroxy–3–methylglutaryl–CoA reductase (HMG–
CoA), the enzyme that inhibits the biosynthesis of cholesterol [9]. Diosgenin has been reported to induce
selective apoptosis in AU565 human breast adenocarcinoma cells through PARP cleavage [10]. It has
been demonstrated that diosgenin displays a significant anticancer activity against both the estrogen
receptor positive (ER+) and estrogen receptor negative (ER−) breast cancer cell lines by inhibiting
prosurvival signaling pathways Akt, Raf/MEK, NF–κB and inducing apoptosis [11]. The compound
also inhibits the proliferation of PC–3 human prostate cancer cells due to the reduction of the activity as
well as the mRNA expression of MMP–2 (matrix metalloproteinase) and MMP–9 [12]. It has been
shown that diosgenin suppresses the proliferation of hepatocellular carcinoma (HCC) at the G1 phase
of the cell cycle and inhibits both constitutive and inducible activation of STAT3 with no effect on
STAT5 [13]. Diosgenin also causes the G2/M cell cycle arrest and apoptosis in human leukemia K562
cells [14].
3

In addition, it has been shown that both diosgenin and tigogenin inhibit growth and induce
apoptosis of synovial cells in vitro (human cells of the synovial membrane) which play a significant role
in rheumatoid arthritis [15]. It has also been reported that diosgenin exhibits an anti–inflammatory
activity by downregulating the TNF–α–induced expression of ICAM–1 via the inhibition of NF–κB p65
activation [16]. Both specific and non–specific cellular immune responses might be supported by
diosgenin action which indirectly exerts an anti–tumour effect by activating the immune system [17].
M. Singh et al. have synthesized several analogues of diosgenin by modifying the spiroketal
ring. They have demonstrated that diosgenin analogues inhibit the production of pro–inflammatory
cytokines (TNF–α, IL–6 and IL–1β) both under in vitro and in vivo conditions [18].
Romero–Hernández et al. have proved that newly obtained thioxo and selenoxo diosgenin
derivatives behave as strong antiproliferative agents against human tumor cells, with a higher potency
than that of diosgenin itself [19]. Kvasnica et al. [20] have studied the anticancer activity of the platinum
(II) compounds conjugated with L–histidine and L–methionine esters with various steroids, including
diosgenin. Some compounds have displayed a significant cytotoxic activity towards acute lymphoblastic
leukaemia CEM cell lines with the IC50 values in the 14–25 μM range.
B. Huang et al. have synthesized several diosgenyl esters of primary amino acids and their N–
salicylamides [21]. The compounds have been MTT–tested to evaluate their cytotoxic activity against
the cell lines of human breast carcinoma MDA–MB–231, mouse colon cancer C26 and human
hepatocellular cancer Hep G2. The diosgenin analogue substituted with the 6–aminohexanoic acid
moiety at position C–3 of the steroid core structure (aminocaproic acid analogue) proves to be more
effective than aspirin or diosgenin alone against all three tested cancer cell lines. The IC50 values range
from 4.7 μM against the C26 cells to 14.6 μM against HepG2 cells. Moreover, in the anti–inflammatory
activity test several salicylate conjugates have significantly inhibited ear swelling caused by xylene,
displaying comparable or stronger anti–inflammatory activity than that of diosgenin and aspirin,
whereas among other esters only the aminocaproic acid analog has exhibited similar characteristics to
that of the reference compounds.
While studying the influence of the structure on the anticancer activity and anticoagulant
properties, the same authors [22] have obtained three series of diosgenyl esters with different aglycone
skeleton structures. Compounds with the side chains including 4–8 carbon atoms displayed a much
stronger inhibitory activity towards cancer cell lines (C26 (colon), B16 (melanoma), HepG2 (liver),
A549 (lung), MDA–MB–231 (breast)) than diosgenin itself or its analogues with shorter or longer
carbon chains. Among the studied compounds, the one with the 6–aminocaproic moiety including six
carbon atoms was more active than other analogues, especially against C26 cell lines (IC50 5.5 μM).
R. Ur Masood et al. have synthesized novel triazoles of diosgenin and have evaluated their anti–
proliferative activity against several human cancer cell lines: breast (HBL–100), lung (A–549), colon
(HT–29 and HCT–116). Five analogues have been identified as potent antiproliferative agents against
all the tested cancer cells [23].
Several novel analogues of diosgenin at C7 position have been prepared by Hamid et al. Six of
the analogues exhibit anticancer activity against a panel of human cancer cell lines with IC50 ranging
from 12 to 35 µM. Additionally, these analogues inhibit lipopolysaccharideinduced pro–inflammatory
cytokines (TNF–α and IL–6) [24].
The above mentioned studies demonstrate that diosgenin analogues substituted at position C–3
of the steroid core structure possess a significant antitumor activity. Their immunomodulating profiles
are poorly studied, which we have considered worth investigating. Thus, it was assumed that the
evaluation of new substituents at position C3 may be of interest from the point of view of optimizing
diosgenin analogues as potential therapeutic agents (anticancer or immunomodulatory). For in vitro
biological evaluation we have selected the following cancer cell lines: PC3, MCF–7, MDA–MB–231
4

[25] due to the fact that prostate cancer is the second most frequently diagnosed cancer in men (15% of
all male cancers) and breast cancer among women (25% of all female cancers).
This work reports the synthesis of diosgenin derivatives with new structural fragments at
position C3, such as: levulinic and 3,4–dihydroxycinnamic acids, dipeptides and various amino acids.
Steroid aglycons have been connected to new substituents by an ester bond at the C3–oxygen atom of
their skeleton. The hydrogenation of the double bond at the C5–C6 position of new diosgenin amino
acid derivatives has led to new tigogenin analogues. All compounds have been evaluated for their
antiproliferative activity in vitro against HUVEC and selected cancer cell lines. Particular derivatives
have been examined for their influence on the inflammatory response.
2. Materials and methods
2.1. Abbreviations
AcOEt — ethyl acetate; Boc — tert–butyloxycarbonyl group; DCC— N,N′–dicyclo-
hexylocarbodiimide; DCU — Dicyclohexylurea; DCM — dichloromethane; DIPEA — N,N–
diisopropylethylamine; DMAP — 4–dimethylaminopyridine; HCl — hydrochloric acid; HOBt — N–
hydroxybenzotriazole monohydrate; MS — mass spectrometry; NMR — nuclear magnetic resonance;
TBTU — O–(1H–benzotriazol–1–yl)–1,1,3,3–tetramethyluronium tetrafluoroborate; TEA —
triethylamine; THF — tetrahydrofuran; TMS — trimethylsilyl; caff — caffeic acid; LPS —
lipopolysaccharide; PC3 — prostate human cancer cell lines; MCF–7, MDA–MB–231— breast cancer
cell lines; HCC— hepatocellular carcinoma; C26— colon cancer cell lines; HepG2— human liver
cancer cell line; K562—human leukemia cell line; B16 — melanoma murine cell lines; HBL–100 —
breast cancer cell lines; A–549 — lung cancer cell lines; HT–29 and HCT–116 — colon cancer cell
lines; s, d, t, mov — singlet, doublet, triplet and overlapping multiplet/s in NMR spectra.
2.2. General procedures
Diosgenin was procured from KOCH–LIGHT LABORATORIES, UK. Other materials, solvents
and reagents were of commercial origin and used without additional operations. Reactions were
monitored on the silica gel TLC plates 60 F254 (Merck, Darmstadt, Germany). The visualization was
performed by UV light (254 and/or 365nm) then with CeMo stain and subsequent charring [26]. Melting
points were determined by the Melting Point System (Mettler Toledo MP70). Solvents were evaporated
under reduced pressure at 40°C on the Büchi Rotavapor. Flash column chromatography was performed
on silica gels (200–300 mesh).
The 1H NMR and 13C NMR spectra were recorded in CDCl3 (protected diosgenin derivatives 2–18,
16a, 2b and 4b) and CD3OD (2a–3a, 5a–14, 2c and 4c) solutions with the Varian–NMR–vnmrs600
spectrometer (at 298 K) equipped with a 600 MHz PFG Auto XID (1H/15N–31P 5 mm) indirect
probehead. Standard experimental conditions and standard Varian programs (ChemPack 4.1) were used.
To identify the structures of all isolated products correctly, a careful analysis of the results of 1D and
1
2D NMR experiments was performed. The 1D and 2D measurements covered: H selective NOESY,
2D: COSY, NOESY, 1H–13C gradient selected (g–HSQC) and (g–HMBC) optimized for 1J(C–H) = 150
n
1
13
Hz and J(C–H) = 8 Hz, respectively. The H and C NMR chemical shifts are given relative to the
TMS signal at δ = 0.0 ppm. The concentration of the solutions used for the measurements was about
10–20 mg of the compounds in 0.6 cm3 of the solvent.
The mass spectra were recorded on the MaldiSYNAPT G2–S HDMS (Waters) Spectrometer via
electrospray ionization (ESI–MS). High–resolution mass spectrometry (HRMS) measurements were
performed using the Synapt G2-Si mass spectrometer (Waters) equipped with an ESI source and a
quadrupole–Time–of–flight mass analyzer. The results of the measurements were processed using the
MassLynx 4.1 software (Waters).
5

2.2. Chemical synthesis
The studied compounds were synthesized starting from diosgenin (1), as shown below (Scheme 1).
Scheme 1. i) DCC, DMAP, DCM, r.t; ii) TBTU, HOBt, DIPEA, DCM, r.t.; 2–16 yield 60–98%;
iii) HCl/AcOEt, r.t.; 2a–14a yield 63–93%; iv) 1,3–dimethylbarbituric acid, Ph3P, Pd(OAc)2,
EtOH, 75°C; 16a yield 98%; v) H2, Pd/C, AcOEt, r.t.; 2a, 4b yields 79%, 94%; vi) HCl/AcOEt,
r.t; 2c, 4c yields 65%, 70%; vii) Ac2O, DMAP, TEA, DCM, r.t.; yield 83%; viii) NaBH3CN,
Ac2O, r.t., yield 98%.
2.2.1.General procedure for the synthesis of compounds 2–16
2.2.1.1. Synthesis of analogue 2 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–O–benzyl–
L–serinate
DMAP (40 mg, 0.327 mmol) and DCC (280 mg, 1.357 mmol) were added to the solution of
Boc–L–Ser(Bzl)–OH (400 mg, 1.354 mmol) in anhydrous CH2Cl2 (10 mL) and the mixture was stirred
for 10 min at room temperature. Then diosgenin (406 mg, 0.979 mmol) was added and the reaction
mixture was stirred at room temperature for 24 h (TLC control). After the reaction had been completed,
the DCU precipitate was filtered off. The filtrate was washed successively with the NaHCO3 aq. solution
and NaCl aq. solution. The extract was dried over anhydrous MgSO4, filtered and evaporated to dryness.
The crude product was purified through a silica gel column eluted with ethyl acetate–hexane. Yield 93%,
white solid, m.p. 161.6°C; [M+H]+ calcd for C42H62NO7: 692.4526, found: 692.4531, [M+Na]+ calcd
for C42H61NO7Na: 714.4346, found: 714.4352.
1H NMR (CDCl3) δ [ppm]: L–serine part: 7.33 and 7.28 (5H, 2×m, phenyl), 5.39 (d, J = 8.8 Hz,
NH), 4.56 and 4.47 (2H, 2×J = 12.2 Hz, both H4’ protons), 4.39 (1H, m, H2’), 3.87 and 3.68 (2H, 2×dd,
J = 2.9 and 9.2 Hz/3.0 and 9.2 Hz, both H3’ protons), 1.45 (9H, s, 3×CH3 of t–Bu),
Diosgenin (DSG) part: 5.36 (1H, m, H5), 4.67 (1H, m, H3), 4.41 (1H, m, H16), 3.47 (1H, m, one of H27
protons), 3.38 (1H, t, J = 11.0 Hz, one of H27 protons), 2.27 (2H, m, both H4), 2.01 and 1.56 (2H, mov,
both H7), 1.99 and 1.29 (2H, mov, both H15), 1.88 and 1.58 (2H, mov, both H2), 1.86 and 1.13 (2H, mov,
both H1), 1.87 (1H, mov, H20), 1.78 (1H, mov, H17), 1.74 and 1.19 (2H, mov, both H12), 1.69 and 1.59
(2H, mov, both H23), 1.63 (1H, mov, H8), 1.62 and 1.45 (2H, mov, both H24), 1.62 (1H, mov, H25), 1.51
and 1.46 (2H, mov, both H11), 1.12 (1H, mov, H14), 1.02 (3H, s, CH3–19), 0.98 (1H, m, H9), 0.97 (3H,
d, J = 7.0 Hz, CH3–21), 0.79 (6H, s and d (both overlapping), 2×CH3 group: CH3–18 and CH3–26),
13C NMR δ [ppm]: L–serine part: 170.0 (C1’), 155.5 (NHCO), [137.6, 128.4, 127.8, 127.6 –
phenyl ring], 79.8 (OC(CH3)3), 73.3 (C4’), 70.2 (C3’), 54.2 (C2’), 28.3 (OC(CH3)3),
6

DSG part: 139.5 (C5), 122.5 (C6), 109.3 (C22), 80.9 (C16), 75.2 (C3), 66.9 (C27), 62.1 (C17), 56.4
(C14), 49.9 (C9), 41.6 (C20), 40.3 (C13), 39.7 (C12), 37.9 (C4), 36.9 (C1), 36.7 (C10), 32.1 (C7), 31.9
(C15), 31.4 (2C, C8 and C23), 30.3 (C25), 28.8 (C24), 27.7 (C2), 20.8 (C11), 19.3 (C19), 17.1 (C26),
16.3 (C18), 14.5 (C21).
Analogue 3 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–isoleucinate:
yield 98%, white solid, m.p. 181.7°C; [M+H]+ calcd for C38H62NO6: 628.4577, found: 628.4582;
[M+Na]+ calcd for C38H61NO6Na: 650.4397, found: 650.4406.
1H NMR δ [ppm]: L–isoleucine part: 5.04 (1H, d, J = 8.7 Hz, NH), 4.22 (1H, dd, J = 4.5, 8.7
Hz, H2’), 1.85 (1H, mov, H3’), 1.44 (10H, s, one of H4’ protons and nine protons of t–Bu), 1.18 (1H,
mov, one of H4’ protons), 0.93 (6H, 2×d, J = 7.0 Hz, CH3 at C3’ and CH3–5’) + DSG part,
13C NMR δ [ppm]: L–isoleucine part: 171.6 (C1’), 155.5 (NHCO), 79.6 (OC(CH3)3), 57.9 (C2’),
38.2 (C3’), 28.3 (OC(CH3)3), 25.1 (C4’), 15.5 (CH3 at C3’) and 11.7 (CH3–5’) + DSG part.
Analogue 4 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–5–benzyl–L–glutamate:
yield 94%, white solid, m.p. 154.7°C; [M+H]+ calcd for C44H64NO8: 734.4632, found: 734.4633,
[M+Na]+ calcd for C44H63NO8Na: 756.4451, found: 756.4454.
1H NMR δ [ppm]: L–glutamic acid part: 7.38–7.31 (5H, m, phenyl), 5.12 (2H, s, H6’), 5.09 (1H,
d, J = 8.6 Hz, NH), 4.29 (1H, m, H2’), 2.47 and 2.44 (2H, 2 overlapping multiplets, both H4’), 2.20 and
1.95 (2H, m and part of overlapping multiplets, both H3’), 1.44 (9H, s, protons of t–Bu) + DSG part,
13C NMR δ [ppm]: L–glutamic acid part: 172.6 (C5’), 171.6 (C1’), 155.4 (NHCO), [135.8,
128.6, 128.3 128.2 – phenyl ring], 79.9 (OC(CH3)3), 66.5 (C6’), 53.0 (C2’), 30.3 (C4’), 28.3 (OC(CH3)3),
27.9 (C3’) + DSG part.
Analogue 5 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–glycinate:
yield 90%, white solid, m.p. 137.4°C; [M+H]+ calcd for C34H54NO6: 572.3951, found: 572.3961,
[M+Na]+ calcd for C34H53NO6Na: 594.3771, found: 594.3784.
1H NMR δ [ppm]: glycine: 5.00 (1H, br), 3.88 (2H, m), 1.45 (9H, t–Bu) + DSG part,
13C NMR δ [ppm]: glycine: 169.7 (C1’), 155.7 (NHCO), 79.9 (OC(CH3)3), 42.6 (C2’), 28.3
(OC(CH3)3) + DSG part.
Analogue 6 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–valinate:
yield 98%, white solid, m.p. 137.5°C; [M+H]+ calcd for C37H60NO6: 614.4421, found: 614.4420,
[M+Na]+ calcd for C37H59NO6Na: 636.4240, found: 636.4240.
1H NMR δ [ppm]: L–valine: 5.02 (1H, d, J = 8.8 Hz), 4.17 (1H, dd, J = 4.6, 8.8 Hz, H2’), 1.45
(9H, s, t–Bu), 2.12 (1H, mov, H3’), 0.95 and 0.88 (both CH3 groups at C3’) + DSG part,
13C NMR δ [ppm]: L–valine part: 171.7 (C1’), 155.7 (NHCO), 79.9 (OC(CH3)3), 58.5 (C2’),
31.4 (C3’), 28.3 (OC(CH3)3), 18.9 and 17.5 (both CH3 groups at C3’) + DSG part.
Analogue 7 — (25R)–spirost–5–en–3β–yl Nα,Nτ–bis(t–butoxycarbonyl)–L–histidinate:
yield 94%, white solid, m.p. 163.2°C; [M+H]+ calcd for C43H66N3O8: 752.4850, found:
752.4860.
1H NMR δ [ppm]: L–histidine part: 8.00 (1H, s, H6’), 7.14 (1H, s, H8’), 5.67 (1H, d, J = 8.4 Hz,
NH), 4.51 (1H, m, H2’), 3.03 (2H, m, H3’), 1.60 (9H, s, t–Bu at N7 of imidazole), 1.44 (9H, s, t–Bu at
NH at C2’) + DSG part,
13C NMR δ [ppm]: 171.1 (C1’), 155.5 (NHCO), 146.8 (NHCO – imidazole), 138.7 (C4’ –
imidazole), 136.8 (C6’ – imidazole), 114.6 (C8’ – imidazole), 85.5 (OC(CH3)3 – imidazole), 79.9
(OC(CH3)3), 53.3 (C2’), 30.3 (C3’), 28.3 (OC(CH3)3), 27.9 (OC(CH3)3 – imidazole) + DSG part.
7

Analogue 8 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–methioninate:
yield 98%, white solid, m.p. 121.9°C; [M+H]+ calcd for C37H60NO6S: 646.4141, found:
646.4149, [M+Na]+ calcd for C37H59NO6SNa: 668.3961, found: 668.3972.
1H NMR δ [ppm]: L–methionine part: 5.12 (1H, d, J = 8.0 Hz), 4.36 (1H, m), 2.53 (2H, m), 2.33
(2H, m), 2.11 (3H, s, S–CH3), 1.44 (9H, s, protons of t–Bu) + DSG part,
13C NMR δ [ppm]: L–methionine part :171.6 (C1’), 155.3 (NHCO), 79.9 (OC(CH3)3), 32.4 and
30.0 (C3’ and C4’), 28.3 (OC(CH3)3), 15.5 (CH3–S) + DSG part.
Analogue 11 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–alaninate:
yield 95%, white solid, m.p. 221.7°C; [M+H]+ calcd for C35H56NO6: 586.4108, found: 586.4105.
1H NMR δ [ppm]: L–alanine part: 5.06 (1H, d, J = 6.6 Hz, NH), 4.26 (1H, m, H2’), 1.45 (9H, s,
t–Bu), 1.37 (3H, d, J = 7.2 Hz, CH3–3’) + DSG part,
13C NMR δ [ppm]: L–alanine part: 172.8 (C1’), 155.1 (NHCO), 79.7 (OC(CH3)3), 49.3 (C2’),
28.3 (OC(CH3)3), 18.8 (CH3–3’) + DSG part.
Analogue 12 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–phenylalaninate:
yield 78%, white solid, m.p. 143.5°C; [M+H]+ calcd for C41H60NO6: 662.4421, found: 662.4422.
1H NMR δ [ppm]: L–phenylalanine part: 7.29, 7.24, 7.15 (5H, m, Ph), 4.53 (1H, m, H2’), 3.07
(2H, m, H3’), 1.43 (9H, s, t–Bu) + DSG part,
13C NMR δ [ppm]: L–phenylalanine part: 171.2 (C1’), 155.1 (NHCO), 136.1, 129.4, 128.4,
126.9 (Ph), 79.8 (OC(CH3)3), 54.5 (C2’), 38.4 (C3’), 28.3 (OC(CH3)3) + DSG part.
Analogue 13 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–leucinate:
yield 78%, white solid, m.p. 150.9°C; [M+H]+ calcd for C38H62NO6: 628.4577, found: 628.4581,
[M+Na]+ calcd for C38H61NO6Na: 650.4397, found: 650.4408.
1H NMR δ [ppm]: L–leucine part: 4.89 (1H, d, J = 8.4 Hz, NH), 4.26 (1H, m, H2’), 1.71 (1H,
m, H4’), 1.59 and 1.48 (both H3’), 1.44 (9H, s, t–Bu), 0.95 and 0.94 (2×3H, 2×CH3 at C4’) + DSG part,
13C NMR δ [ppm]: L–leucine part: 172.9 (C1’), 155.4 (NHCO), 79.8 (OC(CH3)3), 52.2 (C2’),
41.9 (C3’), 28.3 (OC(CH3)3), 24.8 (C4’), 22.8 and 22.0 (both CH3 groups at C4’) + DSG part.
Analogue 14 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–prolinate:
yield 94%, white solid, m.p. 185.4°C; [M+H]+ calcd for C37H58NO6: 612.4264, found: 612.4261,
[M+Na]+ calcd for C37H57NO6Na: 634.4084, found: 634.4081.
NMR: mixture of two diasteroisomers 1) and 2) in the proportion 1.0 : 0.6.
1H NMR δ [ppm]: L–proline part:
1) 4.19 (1H, m, H2’), 3.56 and 3.45 (2H, 2×mov, both H5’), 2.22 and 1.96 (2H, 2×mov, both
H3’), 1.93 and 1.85 (2H, 2×mov, both H4’), 1.42 (9H, s, t–Bu) + DSG part,
2) 4.28 (1H, m, H2’), 3.51 and 3.38 (2H, 2×mov, both H5’), 2.18 and 1.94 (2H, 2×mov, both
H3’), 1.94 and 1.87 (2H, 2×mov, both H4’), 1.46 (9H, s, t–Bu) + DSG part,
13C NMR δ [ppm]: L–proline part:
1) 172.6 (C1’), 145.9 (NHCO), 79.8 (OC(CH3)3), 59.2 (C2’), 46.3 (C5’), 31.0 (C3’), 23.5
(C4’), 28.3 (OC(CH3)3) + DSG part,
2) 172.3 (C1’), 145.3 (NHCO), 79.6 (OC(CH3)3), 59.0 (C2’), 46.5 (C5’), 30.0 (C3’), 24.2
(C4’), 28.4 (OC(CH3)3) + DSG part.
Analogue 15 — (25R)–spirost–5–en–3β–yl levulinate:
yield 60%, white solid, m.p. 139.7°C; [M+Na]+ calcd for C32H48O5Na: 535.3399, found:
535.3395.
8

1H NMR δ [ppm]: levulinic acid part: 2.74 (2H, t, H2’), 2.56 (2H, t, H3’), 2.20 (3H, s, CH3–5’)
+ DSG part,
13C NMR δ [ppm]: levulinic acid part: 206.7 (C4’), 172.1 (C1’), 29.9 and 28.3 (C2’ and C3’) +
DSG part.
Analogue 16 — (25R)–spirost–5–en–3β–yl (E)–3–(3,4–dialliloxyphenyl) acrylate:
yield 87%, white solid, m.p. 159.5°C; [M+H]+ calcd for C42H57O6: 657.4155, found: 657.4156,
[M+Na]+ calcd for C42H56O6Na: 679.3975, found: 679.3975.
1H NMR δ [ppm]: (E)–3–(3,4–dialliloxyphenyl)acrylic acid part: 7.58 (1H, d, J = 15.9 Hz,
CH=CH trans), 7.07 (2H, m, two phenyl protons), 6.87 (1H, d, J = 8.5 Hz, one phenyl proton), 6.26
(1H, d, J = 15.9 Hz, CH=CH trans), 6.12 – 6.04 (2H, 2×m, both CH protons of CH2=CH– groups), 5.45
and 5.41 (4H, 2×m, both CH2 protons of CH2=CH– groups), 4.64 (4H, m, both CH2 protons of =CH–
CH2– groups) + DSG part,
13C NMR δ [ppm]: (E)–3–(3,4–dialliloxyphenyl)acrylic acid part: 166.6, 150.5, 148.5, 144.3,
133.1, 132.9, 127.7, 122.4, 117.9, 117.8, 116.4, 113.4, 112.6, 69.9, 69.7 + DSG part.
2.2.1.2. Synthesis of analogue 10 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–isoleucyl–
L–prolinate
DIPEA (92 µL, 0.531 mM) was added to the solution of Boc–L–Ile–OH (41 mg, 0.177 mmol),
TBTU (114 mg, 0.354 mmol) and HOBt (48 mg, 0,354 mmol) in DCM (5 mL) and the mixture was
stirred for 15 min at room temperature. Then Boc–Pro–DSG 14 (150 mg, 0.274 mM) was added and the
reaction mixture was stirred at room temperature for 24 h (TLC monitoring). After the reaction had been
completed, DCM (15 mL) was added. The organic layer was washed successively with the aq. NaHCO3
solution and then aq. NaCl solution. The extract was dried over anhydrous MgSO4, filtered and
evaporated to dryness. The crude product was purified through a silica gel column eluted with ethyl
acetate–hexane.
Analogue 10 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–isoleucyl–L–prolinate:
yield 94%, oil; [M+H]+ calcd for C43H69N2O7: 725.5105, found: 725.5118.
1H NMR δ [ppm]: L–isoleucine–L–proline part: 5.17 (1H, d, J = 9.4 Hz, NH at C7’), 4.47 (1H,
m, H2’), 4.29 (1H, m, H7’), 3.82 (1H, m, one of H5’) 3.66 (1H, m, one of H5’), 2.22 and 1.96 (2H,
2×mov, both H3’), 2.04 and 1.97 (2H, 2×mov, both H4’), 1.76 (1H, m, H8’), 1.60 and 1.14 (2H, both
H9’), 1.42 (9H, s, t–Bu), 1.014 (3H, m, CH3–at C8’), 0.90 (3H, m, CH3–10’) + DSG part,
13C NMR δ [ppm]: L–isoleucine–L–proline part: 171.4 (C1’), 171.3 (C6’), 155.8 (NHCO), 79.4
(OC(CH3)3), 59.2 (C2’), 56.2 (C7’), 47.3 (C5’), 37.9 (C8’), 29.2 (C3’), 28.3 (OC(CH3)3), 24.9 (C4’),
24.2 (C9’), 15.3 (CH3 at C8’), 11.2 (CH3–10’) + DSG part.
Analogue 9 — (25R)–spirost–5–en–3β–yl N–(t–butoxycarbonyl)–L–leucyl–L–valinate:
yield 97%, oil; [M+H]+ calcd for C43H71N2O7: 727.5261, found: 727.5277, [M+Na]+ calcd for
C43H70N2O7Na: 749.5081, found: 749.5099.
1H NMR δ [ppm]: L–leucine–L–valine part: 6.51 (1H, d, J = 8.8 Hz, NH, N3), 4.88 (1H, d, J =
7.3 Hz, NH at C5’), 4.49 (1H, dd, J = 4.8, 8.9 Hz, H2’), 4.12 (1H, m, H5’), 2.18 (1H, m, (CH3)2CH at
C2’), 1.68 and 1.48 (both H6’), 1.60 (H7’), 1.44 (9H, s, t–Bu), 0.95 and 0.91 (12H, 4×CH3, both CH3
groups at C7’and both CH3 groups (CH3)2CH at C2’) + DSG part,
13C NMR δ [ppm]: L–leucine–L–valine part: 172.3, 171.0 (C1’ and C4’), 155.7 (NHCO), 80.0
(OC(CH3)3), 57.0 (C2’),53.1 (C5’), 40.9 (C6’), 31.4 ((CH3)2CH at C2’), 24.8 (C7’), 22.8 and 22.2 (both
CH3 groups at C7’), 19.0 and 17.6 (both CH3 groups of (CH3)2CH at C2’) + DSG part.
9

2.2.2 General procedure for the synthesis of compounds 2a–3a, 5a–14a, 16a
2.2.2.1. Synthesis of (25R)–spirost–5–en–3β–yl O–benzyl–L–serinate hydrochloride.
Product 2 (112 mg, 0.162 mmol) was treated with 2.9M HCl/AcOEt (5.8 mM, 2 mL) and stirred
for 24 h (TLC monitoring). The precipitated salt 2a was collected by filtration and dried under vacuum
at room temperature.
Analogue 2a — (25R)–spirost–5–en–3β–yl O–benzyl–L–serinate hydrochloride:
yield 65%, m.p. 189.2°C; HRMS (ESI) m/z: [M+H]+ calcd for C37H54NO5: 592.4002, found:
592.4008.
1H NMR δ [ppm] (CD3OD): O–benzyl–L–serine: 7.35–7.30 (5H, phenyl), 4.66 (2H, d, J = 12.0
Hz, one of H5’ protons), 4.53 (1H, d, J = 12.0 Hz, one of H5’ protons), 4.25 (1H, dd, J = 3.1 and 3.8
Hz, H2’), 3.94 (1H, dd, J = 4.0 and 10.5 Hz, one of H3’ protons), 3.83 (1H, dd, J = 3.0 and 10.5 Hz, one
of H3’ protons) + DSG part: 5.40 (1H, m, H6), 4.66 (1H, m, H3), 4.40 (1H, m, H16), 3.45 (1H, m, one
of H27), 3.32 (mov with CH3 from CD3OD, one of H27), 2.28 (2H, m, both H4), 2.04 and 1.60 (both
H7), 1.99 and 1.30 (both H15), 1.89 (H20), 1.93 and 1.15 (both H1), 1.90 and 1.66 (both H2), 1.78 and
1.22 (both H12), 1.75 (H17), 1.70 and 1.56 (both H23), 1.68 (H8), 1.62 and 1.42 (both H24), 1.60 (H25),
1.57 and 1.52 (both H11), 1.16 (H14), 1.06 (3H, s, CH3–19), 0.97 (3H, d, J = 6.9 Hz, CH3–21), 0.82
(3H, s, CH3–18), 0.79 (3H, d, J = 6.5 Hz, CH3–26),
13C NMR δ [ppm]: O–benzyl–L–serine: 168.0 (C1’), [138.4, 129.6, 129.3, 129.2 – phenyl], 74.5
(C5’), 68.1 (C3’), 54.5 (C2’) + DSG part: 140.5 (C5), 124.0 (C6), 110.6 (C22), 82.2 (C16), 77.9 (C3),
67.9 (C27), 63.8 (C17), 57.7 (C14), 51.5 (C9), 42.9 (C20), 41.4 (C13), 40.8 (C12), 38.9 (C4), 38.0 (C1),
37.9 (C10), 33.1 (C7), 32.7 (C15), 32.4 (C23), 31.4 (C25), 29.9 (C24), 28.5 (C2), 21.9 (C11), 19.7 (CH3–
19), 17.5 (CH3–26), 16.8 (CH3–18), 14.9 (CH3–21).
Analogue 3a — (25R)–spirost–5–en–3β–yl–L–isoleucinate hydrochloride:
yield 73%, m.p. 276.6°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C33H54NO4: 528.4053,
found: 528.4047.
1H NMR δ [ppm] (CD3OD): isoleucine part: 3.95 (1H, d, J = 4.0 Hz, H2’), 1.98 (1H, m, H3’),
1.56 and 1.38 (2H, both H4’), 1.04 (3H, d, J = 7.0 Hz, CH3 at C3’), 1.01 (3H, t, J = 7.0 Hz, CH3–5’) +
DSG part,
13C NMR δ [ppm]: isoleucine: part: 169.3 (C1’), 58.3 (C2’), 37.9 (C3’), 26.8 (C4’), 14.9 (CH3
at C3’), 12.0 (CH3–5’) + DSG part.
Analogue 5a — (25R)–spirost–5–en–3β–yl–glycinate hydrochloride:
yield 65%; m.p. 242.7°C (dec.); HRMS (ESI) m/z: [M+Na]+ calcd for C29H45NO4Na: 494.3246,
found: 494.3246.
1H NMR δ [ppm] (CD3OD): glycine part: 3.81 (2H, s, H2’) + DSG part,
13C NMR δ [ppm]: glycine part:168.0 (C1’), 41.2 (C2’) + DSG part.
Analogue 6a — (25R)–spirost–5–en–3β–yl–L–valinate hydrochloride:
yield 80%; m.p. 284.3°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C32H52NO4: 514.3896,
found: 514.3889.
1H NMR δ [ppm]: valine part: 3.88 (1H, d, J = 4.5 Hz, H2’), 2.29 (1H, m, H3’), 1.09 and 1.08
(6H, 2×dov, both CH3 groups at C3’) + DSG part,
13C NMR δ [ppm]: valine part: 169.4 (C1’), 59.4 (C2’), 31.0(C3’), 18.4 and 18.2 (both CH3
groups at C3’) + DSG part.
Analogue 7a — (24R)–spirost–5–en–3β–yl–L–histidinate hydrochloride:
10

yield 87%; m.p. 227.5°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C33H50N3O4: 552.3801,
found: 552.3799
1H NMR δ [ppm] (CD3OD): histidine part: 8.96 (1H, s, H6’ – proton between both nitrogens in
diazole ring), 7.56 (1H, s, H8’), 4.43 (1H, mov, H2’), 3.50 – 3.36 (2H, 2×mov, both H3’) + DSG part,
13C NMR δ [ppm]: histidine part: 168.4 (C1’), 135.9, 128.6, 119.7 – diazol ring, 53.0 (C2’), 26.7
(C3’) + DSG part.
Analogue 8a — (25R)–spirost–5–en–3β–yl–L–methioninate hydrochloride:
yield 77%; m.p. 253.3°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C32H52NO4S: 546.3617,
found: 546.3620.
1H NMR δ [ppm] (CD3OD): methionine part: 4.17 (1H, t, J = 6.4 Hz, H2’), 2.66 (2H, m, both
H4’), 2.23 (1H, m, one of H3’) and 2.14 (1H, m, one of H3’), 2.13 (3H, s, CH3 at S) + DSG part,
13C NMR δ [ppm]: methionine part: 169.8 (C1’), 52.9 (C2’), 30.8 (C3’), 30.1 (C4’) 15.0 (CH3
at S) + DSG part.
Analogue 9a — (25R)–spirost–5–en–3β–yl–L–leucyl–L–valinate hydrochloride:
yield 63%; m.p. 228.4°C (dec.); HRMS (ESI) m/z: [M–H] calcd for C38H62N2O5Cl: 661.4347,
found: 661.4343.
1H NMR δ [ppm] (CD3OD): L–leucine–L–valine part: 4.31 (1H, d, J = 5.9 Hz, H2’), 2.19 (1H,
m, CH–(CH3)2 at C2’), 1.002 and 1.006 (both CH3 groups of CH–(CH3)2 at C2’), 3.98 (1H, dd, J= 5.6
and 8.4 Hz, H5’), 1.75 (2H, one of H6’and H7’), 1.67 (1H, one of H6’), 1.022 and 1.008 (both CH3
groups at C7’) + DSG part,
13C NMR δ [ppm] (CD3OD): L–leucine–L–valine part: 171.8 (C1’), 171.0 (C4’), 59.8 (C2’), 52.8
(C5’), 41.9 (C6’), 25.3 (C7’), 31.5 (CH–(CH3)2 at C2’), 23.2 and 22.2 (both CH3 groups at C7’), 19.4
and 18.6 (both CH3 groups of CH–(CH3)2 at C2’) + DSG part.
Analogue 10a — (25R)–spirost–5–en–3β–yl–L–isoleucyl–L–prolinate hydrochloride:
yield 70%; m.p. 210.3°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C38H61N2O5: 625.4580,
found: 625.4568.
1H NMR δ [ppm] (CD3OD): 4.48 (1H, m, H2’), 4.10 (1H, d, J = 5.6 Hz, H7’), 3.76 (1H, m, one
of H5’), 3.63 (1H, m, one of H5’), 2.32 and 1.97 (both H3’), 2.07 and 2.01 (both H4’ and H8’), 1.66 and
1.25 (both H9’), 1.14 (CH3 at C8’), 1.00 (CH3–10’) + DSG part,
13C NMR δ [ppm] (CD3OD): 172.5 (C1’), 168.6 (C6’), 61.0 (C2’), 57.7 (C7’), 48.9 (C5’), 37.6
(C8’), 30.2 (C3’), 26.2 (C4’), 24.7 (C9’), 15.1 (CH3 at C8’), 11.7 (CH3–10’) + DSG part.
Analogue 11a — (25R)–spirost–5–en–3β–yl–L–alaninate hydrochloride:
yield 92%; m.p. 275.4°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C30H48NO4: 486.3583,
found: 486.3581.
1H NMR δ [ppm] (CD3OD): L–alanine part: 4.04 (1H, q, J = 7.2 Hz, H2’), 1.52 (3H, d, J = 7.2
Hz, CH3–3’) + DSG part,
13C NMR δ [ppm] (CD3OD): L–alanine part: 170.5 (C1’), 49.9 (C2’), 16.3 (CH3–3’) + DSG part.
Analogue 12a — (25R)–spirost–5–en–3β–yl–L–phenylalaninate hydrochloride:
yield 93%; m.p. 253.0°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C36H52NO4: 562.3896,
found: 562.3892.
1H NMR δ [ppm] (CD3OD): L–phenylalanine part: 7.33 – 7.28 (5H, m, phenyl), 4.27 (1H, t, J =
7.1 Hz, H2’), 3.21 (2H, m, both H3’) + DSG part.
13C NMR δ [ppm] (CD3OD): L–phenylalanine part: 169.5 (C1’), 135.4, 130.6, 130.1, 128.9 –
phenyl, 55.2 (C2’), 37.7 (C3’) + DSG part.
11

Analogue 13a — (25R)–spirost–5–en–3β–yl–L–leucinate hydrochloride:
yield 84%; m.p. 263.8°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C33H54NO4: 528.4053,
found: 528.4047.
1H NMR δ [ppm] (CD3OD): L–leucine part: 3.98 (1H, m, H2’), 1.79 (1H, mov, H4’), 1.80 and
1.68 (2H, 2×mov, both H3’), 1.02 and 1.00 (2×3H, both CH3 groups at C4’) + DSG part.
13C NMR δ [ppm] (CD3OD): L–leucine part: 170.4 (C1’), 52.6 (C2’), 40.9 (C3’), 25.7 (C4’),
22.6 and 22.4 (both CH3 groups at C4’) + DSG part.
Analogue 14a — (25R)–spirost–5–en–3β–yl–L–prolinate hydrochloride:
yield 74%; m.p. 249.8°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C32H50NO4: 512.3740,
found: 512.3733.
1H NMR δ [ppm] (CD3OD): L–proline: 4.41 (1H, m, H2’), 3.38 (2H, mov, H5’), 2.08 (2H, m,
H4’), 2.43 and 2.12 (2H, 2×mov, H3’) + DSG part.
13C NMR δ [ppm] (CD3OD): L–proline: 169.6 (C1’), 60.8 (C2’), 47.2 (C5’), 29.5 (C3’), 24.5
(C4’) + DSG part.
2.2.2.2 (25R)–spirost–5–en–3β–yl (E)–3–(3,4–dihydroxyphenyl)acrylate
1,3–dimethylbarbituric acid (1.5 g, 9.608 mM), Ph3P (296 mg, 1.129 mM) and Pd(OAc)2 (299
mg, 1.333 mM) were added to the suspension of DSG–O–caff–All (16) (700 mg, 1.066 mM) in EtOH
(53 mL). The reaction mixture was heated to 75°C for 3 hours. After the reaction had been completed
(TLC control), the mixture was filtered through a cellite pad and the filtrate was evaporated under
reduced pressure. DCM was added and washed with 5% aq. NaHCO3 and water. The extract was dried
over anhydrous MgSO4, filtered and evaporated to dryness. The crude product was purified through a
silica gel column eluted with DCM/MeOH (95:5) (v/v) to give a white solid.
Analogue 16a — (25R)–spirost–5–en–3β–yl (E)–3–(3,4–dihydroxyphenyl)acrylate:
yield 98%, m.p. 218.4°C (dec.); HRMS (ESI) m/z: [M–H] calcd for C36H47O6: 575.3373, found:
575.3369.
1H NMR δ [ppm]: (E)–3–(3,4–dihydroxyphenyl)acrylic acid part: 7.57 (1H, d, J = 15.9 Hz,
H2’), 7.11 (1H, d, J = 1.8 Hz, H5’), 6.99 (1H, dd, J = 1.8 and 8.4 Hz, H9’), 6.87 (1H, d, J = 8.4 Hz,
H8’), 6.56 (1H, br, OH), 6.43 (1H, br, OH), 6.24 (1H, d, J = 15.9 Hz, H3’) + DSG part,
13C NMR δ [ppm]: (E)–3–(3,4–dihydroxyphenyl)acrylic acid part: 167.4 (C1’), (146.6, 145.0,
144.1, 127.4, 122.3, 115.8, 115.4, 114.3 – phenyl and double bond carbons) + DSG part.
2.2.3.Synthesis of compounds 2b and 4b
(25R)–5α–spirostan–3β–yl N–(t–butoxycarbonyl)–L–serinate 2b:
A solution of 2 (290 mg, 0.419 mmol) in 20 ml AcOEt was hydrogenated in the presence of
10% palladium on charcoal (70 mg) for 6 hrs. After filtration through a cellite pad, the solution was
evaporated. The product was obtained as a white solid, yield 199 mg (79%).
Analogue 2b — (25R)–5α–spirostan–3β–yl N–(t–butoxycarbonyl)–L–serinate:
yield 79%, m.p. 214.9°C; [M+H]+ calcd for C35H58NO7: 604.4213, found: 604.4218, [M+Na]+
calcd for C35H57NO7Na: 626.4033, found: 626.4041.
1H NMR δ [ppm]: L–serine part: 5.46 (1H, d, J = 6.4 Hz, NH), 4.29 (1H, m, H1’), 3.89 (2H, m,
both H3’), 1.43 (9H, s, t–Bu); TGG part: 4.75 (1H, m, H3), 4.36 (1H, m, H16), 3.44 (1H, ddd, J = 2.0,
4.1, 10.9 Hz, one of H27), 3.36 (1H, t, J = 10.9 Hz, one of H27), 1.83 (H20), 1.73 (H17), 1.59 (H25),
1.51 (H8), 1.14 (H5), 1.07 (H14), 0.64 (H9), 1.67 and 1.10 (both H12), 1.71 and 1.00 (both H1), 1.59
and 1.36 (both H4), 1.65 and 0.87 (both H7), 1.95 and 1.22 (both H15), 1.64 and 1.56 (both H23), 1.59
12

and 1.42 (both H24), 1.25 (both H6), 1.81 and 1.51 (both H2), 1.47 and 1.26 (both H11), 0.93 (3H, d, J
= 7.0 Hz, CH3–21), 0.81 (3H, s, CH3–19), 0.76 (3H, d, J = 6.4 Hz, CH3–26), 0.73 (3H, s, CH3–18),
13C NMR δ [ppm]: L–serine part: 170.2 (C1’), 155.8 (OCOC(CH3)3), 80.2 (OCOC(CH3)3), 55.9
(C2’), 63.8 (C3’), 28.3 (OCOC(CH3)3); TGG part: 109.2 (C22), 80.8 (C16), 75.4 (C3), 66.8 (C27), 62.1
(C17), 56.2 (C14), 54.1 (C9), 44.6 (C5), 41.6 (C20), 40.5 (C13), 40.0 (C12), 36.6 (C1), 35.5 (C10), 35.0
(C8), 33.8 (C4), 32.1 (C7), 31.7 (C15), 31.3 (C23), 30.3 (C25), 28.8 (C24), 28.4 (C6), 27.3 (C2), 21.0
(C11), 17.1 (CH3–26), 16.4 (CH3–18), 14.5 (CH3–21), 12.2 (CH3–19).
Analogue 4b — (25R)–5α–spirostan–3β–yl N–(t–butoxycarbonyl)–L–glutamate:
yield 94%, m.p. 183.2°C; [M+H]+ calcd for C37H60NO8: 646.4319, found: 646.4323, [M+Na]+
calcd for C37H59NO8Na: 668.4138, found: 668.4143.
1H NMR δ [ppm]: L–glutamic acid part: 5.17 (1H, d, J = 8.0 Hz, NH), 4.28 (1H, m, H2’), 2.44
(2H, m, both H4’), 2.17 and 1.92 (2H, 2×mov, both H3’), 1.43 (9H, s, t–Bu); TGG part: 4.74 (1H, m,
H3), 4.39 (1H, m, H16), 3.46 (1H, ddd, J = 2.0, 4.1, 10.9 Hz, one of H27), 3.36 (1H, t, J = 10.9 Hz, one
of H27), 1.85 (H20), 1.75 (H17), 1.61 (H25), 1.52 (H8), 1.15 (H5), 1.09 (H14), 0.65 (H9), 1.69 and 1.12
(both H12), 1.72 and 1.00 (both H1), 1.57 and 1.37 (both H4), 1.66 and 0.89 (both H7), 1.97 and 1.24
(both H15), 1.66 and 1.58 (both H23), 1.60 and 1.43 (both H24), 1.27 (both H6), 1.80 and 1.52 (both
H2), 1.48 and 1.28 (both H11), 0.95 (3H, d, J = 7.0 Hz, CH3–21), 0.83 (3H, s, CH3–19), 0.78 (3H, d, J
= 6.4 Hz, CH3–26), 0.75 (3H, s, CH3–18),
13C NMR δ [ppm]: L–glutamic acid part: 177.2 (CO–5’), 171.6 (CO–1’), 155.6 (OCOC(CH3)3),
80.2 (OCOC(CH3)3), 52.9 (C2’), 30.1 (C4’) 28.7 (C3’), 28.3 (OCOC(CH3)3); TGG part: 109.3 (C22),
80.8 (C16), 75.2 (C3), 66.9 (C27), 62.2 (C17), 56.2 (C14), 54.2 (C9), 44.6 (C5), 41.6 (C20), 40.6 (C13),
40.0 (C12), 36.6 (C1), 35.6 (C10), 35.1 (C8), 33.8 (C4), 32.1 (C7), 31.8 (C15), 30.3 (C25), 28.8 (C24),
28.5 (C6), 27.4 (C2), 21.0 (C11), 17.1 (CH3–26), 16.5 (CH3–18), 14.5 (CH3–21), 12.3 (CH3–19).
2.2.4.Synthesis of compounds 2c and 4c
Synthesis of (25R)–5α–spirostan–3β–yl L–serinate hydrochloride 2c:
Product 2b (96 mg, 0.159 mmol) was treated with 2.9M HCl/AcOEt (8.7 mM, 3 mL) and stirred
for 24 h (TLC monitoring). The precipitated salt was collected by filtration and dried under vacuum at
room temperature.
Analogue 2c— (25R)–5α–spirostan–3β–yl L–serinate hydrochloride:
yield 98%, m.p. 241.5°C (dec.); HRMS (ESI) m/z: [M+H]+ calcd for C30H50NO5: 504.3689,
found: 504.3688.
1H NMR δ [ppm] (CD3OD): L–serine part: 4.06 (1H, m, H2’), 3.97 (1H, dd, J = 4.8 and 11.7
Hz, one of H2’), 3.92 (1H, dd, J = 3.5 and 11.7 Hz, one of H2’), TGG part: 4.84 (H3), 4.38 (H16), 3.45
and 3.32 (2H, 2×m, both H27), 1.90 (H20), 1.74 (H17), 1.61 (H8), 1.60 (H25), 1.23 (H5), 1.16 (H14),
0.74 (H9), 1.98 and 1.26 (both H15), 1.90 and 1.61 (both H2), 1.80 and 1.08 (both H1), 1.73 and 0.97
(both H7), 1.66 and 1.48 (both H4), 1.74 and 1.17 (both H12), 1.70 and 1.56 (both H23), 1.63 and 1.42
(both H24), 1.55 and 1.36 (both H11), 1.33 (both H6), 0.96 (CH3–21), 0.90 (CH3–19), 0.80 (CH3–18),
0.79 (CH3–26), 0.74 (H9),
13C NMR δ [ppm] (CD3OD): L–serine part: 168.4 (C1’), 60.7 (C3’), 56.1 (C2’), TGG part: 110.6
(C22), 82.2 (C16), 77.7 (C3), 67.9 (C27), 63.8 (C17), 57.5 (C14), 55.6 (C9), 45.9 (C5), 42.9 (C20), 41.7
(C13), 41.1 (C12), 37.8 (C1), 36.7 (C10), 36.5 (C8), 34.8 (C4), 33.3 (C7), 32.6 (C15), 32.4 (C23), 31.4
(C25), 29.9 (C24), 29.6 (C6), 28.4 (C2), 22.1 (C11), 17.5 (CH3–26), 16.9 (CH3–18), 14.9 (CH3–21),
12.6 (CH3–19).
Analogue 4c— (25R)–5α–spirostan–3β–yl L–glutamate hydrochloride:
13

yield 65%, m.p. 212°C; HRMS (ESI) m/z: [M+H]+ calcd for C32H52NO6: 546.3795, found:
546.3791.
1H NMR δ [ppm] (CD3OD): L–glutamic acid part: 4.06 (1H, t, J = 6.8 Hz, H2’), 2.15 (2H, m,
both H3’), 2.53 (2H, m, both H4’) + TGG part,
13C NMR δ [ppm] (CD3OD): L–glutamic acid part: 175.4 (C5’), 169.7 (C1’), 53.4 (C2’), 30.3
(C4’), 26.7 (C3’) + TGG part.
2.2.5.Synthesis of compounds 17 and 18
2.2.5.1. Synthesis of (25R)–spirost–5–en–3β–yl acetate 17
Acetic anhydride (0.7 ml, 7.42 mmol) was added to the solution of diosgenin (2 g, 4.82 mmol),
DMAP (0.6 g, 4.91 mmol) and TEA (6.4 ml, 45.98 mmol) in DCM (100 ml) at room temperature. The
solution was stirred at room temperature for 24 h. After the reaction had been completed (TLC control),
the organic layer was washed successively with water (90 ml) and a saturated aq. NaHCO3 solution (80
ml). The extract was dried over anhydrous MgSO4, filtered and evaporated to dryness. The crude product
was crystallized from ethyl acetate to afford a white solid.
Analogue 17 — (25R)–spirost–5–en–3β–yl acetate:
yield 83%, m.p. 204.7°C; [M+H]+ calcd for C29H45O4: 457.3318, found: 457.3323.
1H NMR δ [ppm]: 2.03 (3H, s, CH3 of CH3CO at C3) + DSG part: 5.38 (1H, m, H6), 4.60 (1H,
m, H3), 4.41 (1H, m, H16), 3.47 (1H, ddd, J = 2.3, 4.3 and 10.9 Hz, one of H27), 3.38 (1H, t, J = 11.0
Hz, one of H27), 2.32 (2H, mov, both H4), 2.00 and 1.56 (mov,both H7), 1.98 and 1.29 (mov,both H15),
1.87 (1H, mov, H20), 1.86 and 1.58 (mov, both H2), 1.86 and 1.13 (mov,both H1), 1.78 (1H, mov, H17),
1.74 and 1.19 (mov, both H12), 1.68 and 1.59 (mov, both H23), 1.63 (2H, mov, H8 and H25), 1.62 and
1.45 (mov, both H24), 1.51 and 1.46 (mov, both H11), 1.12 (1H, mov, H14), 1.04 (3H, s, CH3–19), 0.98
(1H, mov, H9), 0.97 (3H, d, J = 6.8 Hz, CH3–21), 0.79 (6H, 2×dov CH3–25 and CH3–18),
13C NMR δ [ppm]: 170.5 (C1’), 21.4 (CH3 of CH3CO at C3) + DSG part: 139.7 (C5), 122.3
(C6), 109.3 (C22), 80.8 (C16), 73.9 (C3), 66.8 (C27), 62.1 (C17), 56.4 (C14), 49.9 (C9), 41.6 (C20),
40.2 (C13), 39.7 (C12), 38.1 (C4), 36.9 (C1), 36.7 (C10), 32.0 (C7), 31.8 (C15), 31.4 (C23 and C8),
30.3 (C25), 28.8 (C24), 27.7 (C2), 20.8 (C11), 19.3 (C19), 17.1 (C26), 16.3 (C18), 14.5 (C21).
2.2.5.2.Synthesis of (25R)–furost–5–en–3β–acetoxy–26–ol 18
DSG–3–acetate (1 g, 2.19 mmol) was dissolved in acetic anhydride (70 ml) at room temperature,
and then sodium cyanoborohydride (1 g, 15.91 mmol) was added in portions over a period of 30 min.
This mixture was stirred for 3 h at room temperature (TLC control). Then, the reaction was poured into
ice–cold water (450 ml), extracted with AcOEt (3×70 ml), and washed with water. The extract was dried
over anhydrous MgSO4, filtered and evaporated to dryness. The crude product was purified through a
silica gel column with hexane–ethyl acetate.
Analogue 18 — (25R)–furost–5–en–3β–acetoxy–26–ol:
yield 98%, m.p. 107.4°C; HRMS (ESI) m/z: [M+Na]+ calcd for C29H46O4Na: 481.3294, found:
481.3296.
1H NMR δ [ppm]: 2.03 (3H, s, CH3CO– at C3) + DSG part: 5.37 (1H, m, H6), 4.60 (1H, m,
H3), 4.31 (1H, m, H16), 3.50 and 3.44 (2H, 2×dd, J = 6.2, 10.7 Hz, both H27), 3.33 (1H, m, H22), 2.32
(2H, m, both H4), , 2.00 and 1.31 (2×mov, both H15), 1.99 and 1.54 (2×mov, both H7), 1.86 and 1.59
(2×mov, both H2), 1.86 and 1.14 (2×mov, both H1), 1.75 (m, H20), 1.72 and 1.13 (2×mov, both H12),
1.67 (mov, H25), 1.62 (mov, H8), 1.60 (mov, H17), 1.59 (mov, both H23), 1.51 and 1.45 (2×mov, both H11),
1.47 and 1.35 (2×mov, both H24), 1.09 (1H, mov, H14), 1.03 (3H, s, CH3–19), 1.00 (3H, d, J = 6.7 Hz,
CH3–21), 0.95 (1H, mov, H9), 0.91 (3H, d, J = 6.8 Hz, CH3–27), 0.81 (3H, s, CH3–18),
14

13C NMR δ [ppm]: 170.5 (CO at C3), 21.4 (CH3CO at C3) + DSG part: 139.7 (C5), 122.4 (C6),
90.3 (C22), 83.2 (C16), 73.9 (C3), 68.1 (C27), 65.1 (C17), 56.9 (C14), 50.0 (C9), 40.7 (C13), 39.4
(C12), 38.1 (C4), 37.9 (C20), 37.0 (C1), 36.7 (C10), 35.7 (C25), 32.2 (C15), 32.0 (C7), 31.5 (C8), 30.4
(C23), 30.1 (C24), 27.7 (C2), 20.6 (C11), 19.3 (CH3–19), 18.9 (CH3–21), 16.6 (CH3–26), 16.4 (CH3–
18).
2.3. Biological studies
2.3.1. Cells and a cytotoxicity assay
Human umbilical vein endothelial cells (HUVEC, purchased from Life Technologies, Carlsbad,
CA, USA) were cultured in the Medium 200 with a low serum growth supplement (Life Technologies
Carlsbad, CA, USA) according to the manufacturer’s instructions. The MCF–7 cells (mammary gland,
breast; derived from the metastatic site) were cultivated in the Eagle's Minimum Essential Medium
supplemented with 1% non–essential amino acids. PC–3 cells (prostate; derived from the metastatic site:
bone) were cultured in the RPMI 1640 medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA).
MDA–MB–231 cells (mammary gland) were cultured in the DMEM (Gibco, Thermo Fisher Scientific,
Waltham, MA, USA medium). THP–1 cells (human monocytic leukemia) were grown in RPMI 1640
supplemented with 0.05 mM β–mercaptoethanol (Sigma, St. Louis, MO, USA). All cell culture media
were supplemented with a 10% fetal bovine serum FBS (Gibco, Thermo Fisher Scientific, Waltham,
MA, USA) and antibiotics. The cells were grown at 37oC in a 5% CO2 atmosphere.
For the cytotoxicity studies 10×103 cells/well of HUVEC or 7×103 cells/well of cancer cells
were seeded on a 96–well plate (Nunc, Roskilde, Denmark). 24 hours later the cells were exposed to the
test compounds for additional 48 hours. Stock solutions of the test compounds were freshly prepared in
ethanol. The final concentrations of the compounds tested in the cell cultures were: 1×10-1 mM, 4×10-2
mM, 1.6×10-2 mM, 6.5×10-3 mM, 2.5×10-3 mM and 1×10-3 mM. The concentration of ethanol in the cell
culture medium was 1%.
The IC50 values were calculated from dose–response curves and used as a measure of cellular
sensitivity to a given treatment.
The cytotoxicity of all compounds was determined by the MTT [3–(4,5–dimethylthiazol–2–yl)–
2,5–diphenyltetrazolium bromide; Sigma, St. Louis, MO] assay as described [27]. Briefly, after 24 or
48 hours of incubation with drugs, the cells were treated with the MTT reagent and the incubation was
continued for 2 hours. MTT–formazan crystals were dissolved in 20% SDS and 50% DMF at pH 4.7
and the absorbance was read at 570 and 650 nm on the ELISA–PLATE READER (FLUOstar Omega).
As control (100% viability), we used cells grown in the presence of a vehicle (1% Ethanol) only.
2.3.2. Caspase-3/7 enzymatic activity assay
MCF–7 cells were cultured in the Eagle's Minimum Essential Medium supplemented with
antibiotics and a 10% fetal bovine serum in a 5% CO2 atmosphere at 37°C. 20×103 cells were seeded in
each well on a 96–well plate (Nunc, Roskilde, Denmark). After 24 hours the cells were exposed to the
test compound 2c at the concentration of 2×IC50 and 5×IC50 for another 18 hours. The cells were also
exposed to 1% ethanol (vehicle control) or 1μM staurosporine (a strong inducer of cell apoptosis, Sigma,
St. Louis, MO). The induction of cell apoptosis was analyzed by the Apo–ONE® Homogeneous
Caspase–3/7 Assay (Promega, Madison, WI, USA). After 18 hours of incubation with the test
compounds, the cells were treated with the caspase–3/7 reagent according to the manufacturer’s
instructions and incubated for additional 1.5 hour at room temperature. The fluorescence in each well
was measured in triplicate (excitation at 485 nm, emission measured at 520 nm) using a microplate
reader FLUOStar Omega (BMG–Labtech, Germany). For the normalization of data, the level of caspase
activation in the control cells (exposed to 1% ethanol) was taken as 1.0.
15

2.3.3.Cytokine mRNA quantification by real–time RT–PCR
A real–time RT–PCR method was used to measure the mRNA level of a given cytokine (IL–1,
IL–4, IL–10, IL–12, TNF–α) in THP–1 cells (Institute of Medical Biology of Polish Academy of
Sciences, Łódź, Poland) exposed to diosgenin or its derivatives. To obtain RNA for the gene expression
analysis, THP–1 cells were seeded on a 6–well plate (Nunc) in the amount of 1×106 cells/well and
exposed to the test compounds at the concentration of 5 µM for another 6 hours. The control cells were
exposed to 1% ethanol or 5 μg/ml LPS (Sigma, St. Louis, MO, USA). The total RNA pool was isolated
from the cell lysates using a TriPure Isolation Reagent (Roche, Basel, Switzerland) according to the
manufacturer’s instruction. The RNA purity and integrity were checked spectrophotometrically with a
NanoDrop ND–1000 spectrophotometer (Thermo Fisher Scientific).
Reverse transcription and PCR amplification reactions were performed in one step using the
LightCycler® 1.0 Instrument (Roche) and LightCycler RNA Amplification Kit SYBR Green I (Roche).
Each sample contained 250 ng of total RNA, 2 mM MgCl2, 2 µl of 5x SYBR Green, 0.25 µM of forward
and reverse primers, and 0.2 µl of the enzyme mix (total sample volume of 10 µl). The qRT–PCR
reactions were optimized for each studied gene and performed according to a general protocol: a reverse
transcription reaction (RT) for 10 minutes at 55°C, and a denaturation step for 30 seconds at 95°C. Three
steps of the PCR quantification reaction included: I – denaturation (0 seconds at 95°C), II – annealing
(10 seconds at 61°C), III – product extension (8 seconds at 72°C), 45 cycles in total. Subsequent melting
experiments were performed by quick denaturation at 95oC, annealing over 10 seconds at 65°C and
heating up to 95°C at 0.1 C/s. The IL–1, IL–4, IL–10, IL–12, and the TNF–α mRNA levels were
normalized against the reference GAPDH mRNA. Changes in the mRNA expression invoked by the
tested compounds were calculated using the ΔΔCt method (calibrator–RNA isolated from THP–1 cells
exposed to 1% ethanol, mean values ± standard deviation from three experiments are given). Primer
sequences for the real–time RT–PCR reactions were taken from literature data [28].
Table 1. Primers used for the amplification of cytokine mRNA (IL–1, IL–4, IL–10, IL–12, TNF–α) and
GAPDH in real–time RT–PCR reaction.
mRNA
IL–1
Sequence
Fwd 5’–CCTGTCCTGCGTGTTGAAAGA–3’
Rev
Fwd 5’–AACAGCCTCACAGAGCAGAAGAC–3’
Rev 5’–GCCCTGCAGAAGGTTTCCTT–3’
Fwd 5’–GCTGGAGGACTTTAAGGGTTACCT–3’
Rev 5’–CTTGATGTCTGGGTCTTGGTTCT–3’
Fwd 5’–CATGGTGGATGCCGTTCA–3’
Rev 5’–ACCTCCACCTGCCGAGAAT–3’
Fwd 5’–CCCCAGGGACCTCTCTCTAATC–3’
Rev 5’–GGTTTGCTACAACATGGGCTACA–3’
Fwd 5’–CATCATCTCTGCCCCCTCTC–3'
Rev 5'–CTGTTGAAACCATAGCACCT–3'
5’–GGGAACTGGGCAGACTCAAA–3’
IL–4
IL–10
IL–12
TNF–α
GADPH
16

2.4. Statistical analysis
The results were subjected to a nonparametric analysis of the ANOVA variance followed by the
Tukey post–hoc test. The analyzes were performed using the GraphPad_Prism 8 program (GraphPad
Software Inc., USA). As a borderline level of significance, p <0.05 was accepted.
2.5. Computational methods
In the computational part of this work we first built the models of all studied molecules manually
and then the LigProp 3.3 software (Schrodinger Inc.) was used to prepare all atom 3D structures.
Afterwards, we evaluated all ADME properties using the QikProp 4.6 software (Schrodinger Inc.) with
default options. The most important selected ADME properties have been presented in Table 4 and
analyzed in section 3.4.
2.6. Molecular docking
For the molecular docking part we used the crystal structure (PDB code: 2YAT [29]) of the estrogen
receptor–ligand binding domain complex and the crystal structure (PDB code: 4UDD [30]) of the GR.
The ligand and water molecules were removed, which was followed by the addition of hydrogen atoms
with AutoDockTools 4 [31]. Atomic interaction energy grids were calculated using probes
corresponding to each atomic type found in the 2c molecule at the 0.375 Å grid resolution. In all docking
experiments all ligands were treated in a fully flexible manner with the Gasteiger partial charges added
by AutoDockTools 4. For each ligand–receptor pair we used a two–step docking protocol to assess the
most likely binding site of the ligands in the receptors. In the first step we used a 116×126×126 (2YAT)
or 126×126×126 (4UDD) Å cubic boxes that spanned over all the receptors and as amino acids were
treated as rigid. We used Autodock 4.2 [31] with the Genetic Lamarckian Algorithm and standard
options, but including 200 dockings per compound and 5,000,000 energy evaluations per docking [32].
Such an approach allowed us to identify three potential binding sites within the estrogen receptor with
relatively low estimates of 2c binding affinities (<220 nM) and three binding sites within the GR with
low estimates of 4c and 16a binding affinities (<70 nM). In the second step a more accurate docking to
these potential sites was performed using the same docking protocol. However, the selected amino acids
were treated as flexible (see Table 2) and docked to smaller, better defined sites.
Table 2. List of flexible residues
Potential binding site List of flexible residues
2YAT site 1
2YAT site 2
2YAT site 3
4UDD site 1
4UDD site 2
4UDD site 3
GLU353, ARG394, PHE404, GLU423, PHE425, HIS524
GLU423, ASP426, ARG434, HIS513
SER381, THR460, ARG515, ASN519, ARG548
MET560, LEU563, ASN564, GLN570, ARG611, TYR735, THR739, ILE747,
GLU540, GLN570, TRP610, ARG611, ARG614, GLN615
ASN619, TYR640, LYS644, HIS645, TYR648, GLU727, GLU730
17

3. Results and discussion
3.1. Chemistry
In the present study a series of diosgenyl derivatives, including ten amino acid diosgenyl esters
(2a–8a and 11a–14a), two dipeptide derivatives (9a and 10a), two esters of caffeic and levulinic acid
(16a and 15) have been synthesized.
The amino acid derivatives 2–8 and 11–14 were prepared by the esterification reaction involving
N,N′–dicyclohexylocarbodiimide (DCC) and 4–dimethylaminopyridine (DMAP), corresponding N–t–
butoxycarbonyl amino acids, and diosgenin in dry dichloromethane at room temperature. After the
reaction had been completed, the post–reaction mixture was filtered and water was added to the residue.
The organic layer was separated and washed successively with NaHCO3 aq. solution, then NaCl aq.
solution. The extract was dried over anhydrous MgSO4, filtered and evaporated to dryness. Products 2–
8 and 11–14 were then purified by flash chromatography (ethyl acetate/hexane mixture). The yield of
2–8 and 11–14 intermediates ranged from 78 to 98%. In the next step Boc groups were removed by
2.9M HCl gas in ethyl acetate. The deprotection gave appropriate hydrochlorides 2a–8a and 11a–14a
as final products with good yields.
Dipeptide derivatives 9 and 10 were synthesized using a “step by step” method. (25R)–spirost–5–
en–3β–yl L–valinate hydrochloride (2a) was coupled with Boc–L–Leu using TBTU as the coupling
activator in the presence of HOBt and DIPEA in DCM [33]. Likewise, (25R)–spirost–5–en–3β–yl N–
(t–butoxycarbonyl)–L–prolinate was coupled with Boc–L–Ile by the TBTU method in DCM. Crude
products 9 and 10 were then separated by extraction and purified by flash chromatography (ethyl
acetate–hexane mixture). The pure fraction was evaporated, allowing to obtain 9 and 10. The yields
were 97% and 94%, respectively. Then, Boc removal with hydrogen chloride in ethyl acetate gave
hydrochlorides 9a and 10a.
Tigogenin derivatives 2c and 4c with a single bond in the C5–C6 position were obtained in the
course of reducing the N–protected compounds 2 and 4, with a simultaneous removal of the benzyl
groups. The reduction was carried out under hydrogenation conditions with Pd/C as a catalyst in ethyl
acetate. After the reaction had been completed, the mixture was filtered through a cellite pad, the filtrate
was evaporated and products 2b and 4b were obtained with good yields. In the next step Boc groups
were removed by hydrogen chloride in ethyl acetate to give hydrochlorides 2c and 4c with the yields of
65% and 70%.
A diosgenin derivative with caffeic acid 16 was obtained as a result of the condensation reaction of
the allyl caffeic acid derivative with diosgenin (DCC, DMAP, THF) with the subsequent removal of the
allyl groups (All) by 1,3–dimethylbarbituric acid in the presence of triphenylphosphine and
palladium(II) acetate at 75°C. After the reaction had been completed, the mixture was filtered through
a cellite pad and the filtrate was evaporated. The crude product was then separated by extraction and
purified by flash chromatography with a mixture of dichloromethane and methanol to give product 16a
(yield 98%).
Levulinic acid diosgenyl ester 15 was obtained analogously as 16, i.e. as a result of the esterification
involving DCC, DMAP, levulinic acid and diosgenin in dry tetrahydrofuran at room temperature. The
crude product was isolated by extraction and then purified by flash chromatography with a mixture of
ethyl acetate and hexane to obtain 15 with the 60% yield.
Compound 17 was obtained by the acetylation of diosgenin with acetic anhydride in the presence
of DMAP and triethylamine in dichloromethane at room temperature. The spectral data of acetate 17
was as reported earlier [34]. The F–ring of the spiroketal bond was opened under mild reductive cleavage
conditions using sodium cyanoborohydride in acetic acid at room temperature [35]. As a result of the
extraction and purification by flash chromatography we obtained product 18 with a good yield (98%).
Structures of all diosgenin analogues have been confirmed by spectral techniques.
18

3.1.1 NMR
Taking into account predictable differences in the structures of compounds 4/4b and 17/18, we
performed a detailed 1H and 13C NMR comparative analysis for these pairs of analogues.
Although hydrogenation of the C5–C6 bond in 4 can theoretically lead to two different
enantiomeric products originating from the attack of hydrogen on the C5 carbon from both sides of the
molecule, the known examples of the diosgenin double bound hydrogenations indicated the
stereoselectivity of the process [36]. In order to confirm that the spatial position of the methyl group at
C10 (CH3–19) limits the substitution of hydrogen from the above, we studied the hydrogenation product
of 4. Similarly to literature data, we isolated one product which was identified as 4b tigogenin derivative
with H3 and H5 protons in the cis–axial positions and the methyl group at C10 in the trans–axial position
to proton H5. This fact was fully confirmed by the NOE effect observed in the 2D NOESY experiment
where the cross–peak at δ = 4.74/1.15 ppm corresponding to H3/H5 protons was evident. Further
meaningful effects proving the H3/H5 cis relation were noticed for the following cross–peaks: δ =
0.65/1.15 ppm (H9/H5), δ = 0.65/1.00 (H9/one of H1 protons), δ = 4.74/1.57 ppm (H3/one of H4
protons), δ = 4.74/1.80 (H3/one of H2 protons) and δ = 4.74/1.00 (H3/one of H1 protons). The formation
of tigogenin derivative 4b from diosgenin 4 was strongly related to a few significant effects reflected in
1
the changes of the H/13C NMR chemical shifts. First of all, the change of a C5–C6 double bond to a
single one caused significant alternations of 13C NMR, as well as 1H NMR chemical shifts. For C5 and
C6 nuclei a shielding increase of ca. 95.0 ppm was recorded. Other differences were observed for C4,
C8, C9, C10 and C19 nuclei. In all these cases medium shielding effects: 4.0 ppm (C4), 1.0 ppm (C10),
7.0 ppm (CH3–19) and deshielding effects: 3.5 ppm (C8) and 4.3 ppm (C9) were calculated. The
1
hydrogenation of the C5–C6 double bond was also related to considerable changes of the H NMR
chemical shifts but this time only shielding effects were noticed. The biggest one was observed for H6
proton (ca. 4.1 ppm), whereas smaller ones were for both H4 (ca. 0.7 – 1.0 ppm), both H7 protons (ca.
0.4 – 0.7 ppm), H8 (ca. 0.1 ppm), H9 (ca. 0.3 ppm) and protons of the methyl group at C10 (CH3–19)
(ca. 0.2 ppm).
1
The opening of the diosgenin ring F led to several changes of the H/13C NMR chemical shifts
13
related to this structural modification. A relatively strong C shielding increase by ca. 19.0 ppm was
observed only in the case of C22 nucleus, whereas for other carbon nuclei in the ring opening area the
opposite effects of 13C shielding were noticed: C16 (ca. 2.5 ppm), C17 (ca. 3.0 ppm), CH3–19 (ca. 4.4
ppm), C27 (ca. 1.2 ppm) and C25 (ca. 5.5 ppm). For the respective protons these changes were smaller
and less consistent.
3.2.Biological studies
3.2.1. Anticancer activity
All synthesized compounds have been screened for their cytotoxic activity against human cancer cell
lines, i.e. breast (MCF–7 and MDA–MB–231), and prostate (PC–3). Normal human umbilical vein
endothelial cells (HUVEC) have also been used. The cytotoxicity was measured with an MTT assay and
the cells treated with ethanol (1%) served as control (100% viability).
Table 3. The IC50 values (µM) of the compounds after a 48 hours incubation with the cells. Mean values
with standard deviations are shown.
Compound MCF–7a
PC–3a
MDA–MB–231a
HUVEC
> 100
> 100
> 100
—
TGG
1 (DSG)
21.0 ± 3.6
10.0 ± 2.9
20.0 ± 3.9
10.0 ± 2.5
19

1.4 ± 1.3
8.5 ± 2.6
70.0 ± 7.4
8.5 ± 2.5
7.5 ± 3.6
5.0 ± 2.3
7.5 ± 2.9
3.4 ± 2.9
8.5 ± 2.4
4.0 ± 2.7
50.0 ± 2.8
41.0 ± 4.4
—
10.0 ± 2.3
10.0 ± 3.3
> 100
10.5 ± 4.3
17.0 ± 3.5
> 100
4.5 ± 1.7
7.0 ± 2.2
> 100
2c
3a
4c
16.0 ± 2.7
12.0 ± 2.4
11.0 ± 2.3
20.0 ± 5.6
7.5 ± 2.1
9.2 ± 3.0
7.5 ± 3.8
60.0 ± 4.8
100.0 ± 3.8
—
30.0 ± 3.6
10.5 ± 2.7
11.0 ± 3.4
11.0 ± 4.1
6.0 ± 2.0
13.0 ± 3.3
8.5 ± 3.0
71.0 ± 6.2
97.0 ± 5.0
—
10.5 ± 3.8
9.5 ± 2.7
5.0 ± 1.9
10.0 ± 3.6
3.0 ± 2.1
8.5 ± 2.4
3.7 ± 2.7
60.0 ± 7.0
51.0 ± 3.1
—
5a
6a
7a
8a
9a
2a
10a
18
15
16ab
11a
12a
13a
14a
7.5 ± 4.0
22.0 ± 3.4
7.5 ± 4.5
5.4 ± 3.7
19.0 ± 2.5
> 100
23.0 ± 5.2
> 100
8.5 ± 1.6
11.0 ± 3.3
7.0 ± 3.2
7.5 ± 2.2
15.0 ± 3.9
10.0 ± 1.9
17.0 ± 3.7
13.0 ± 3.5
aMCF–7: mammary gland carcinoma; MDA–MB–231: breast cancer; PC–3: prostate cancer
bresult cannot be interpreted — compound interferes with the component of the MTT test
The results of the studies on the cytotoxicity of diosgenyl derivatives are summarized in Table 3.
Diosgenin and tigogenin were used as reference compounds. Most of the obtained derivatives display a
cytotoxic activity higher or comparable to that of diosgenin (IC50: 10 – 21µM). Interestingly, tigogenin
(TGG) did not show any toxicity in the tested concentration range towards the cancer cells.
Derivative 2c (L–serine in position 3 of TGG) shows the highest activity towards MCF–7 cells
(IC50: 1.4 µM), while its activity towards PC–3 and MDA–MB–231 cells is lower and the IC50 value
equals 10 and 10.5 µM, respectively. Furthermore, compounds 7a and 14a (L–histidine and L–proline in
position 3 of DSG) are also active in the tested cell line systems with the highest activity against the
MCF–7 cell line. The IC50 value towards MCF–7 for the derivatives containing the above mentioned
amino acids is 5 µM (for the reference compound 21 µM). The activity towards the PC–3 and MDA–
MB–231 cells is lower and the IC50 value is within the range of 10 to 13 µM. Dipeptide diosgenin
derivatives 9a and 10a have also been characterized by a high cytotoxic activity towards MCF–7 cancer
cells, i.e. 3.4 µM and 4 µM, respectively, while their activity towards the PC–3 and MDA–MB–231
cells is also lower (IC50 from 6 to 8.5 µM). On the other hand, in the MCF–7 the IC50 values for
derivatives 3a, 5a, 6a, 8a, 2a, 11a, 13a containing amino acids (L–isoleucine, glycine, L–valine, L–
methionine, O–benzyl–L–serine, L–alanine and L–leucine) in position 3 of DSG are by 7 µM to 8.5 µM
lower than the IC50 values for two remaining cancer cell lines. Compounds 4c, 12a and 15 substituted
with L–glutamine, L–phenylalanine or the levulinic acid display a low cytotoxic activity towards cancer
cells.
The tested derivatives display a high antiproliferative activity towards cancer cells. The IC50 values
show that the MCF–7 breast cancer cells are slightly more sensitive to the studied compounds than the
MDA–MB–231 and PC–3 cells. At the same time, these compounds show a similar or slightly higher
20

cytotoxicity towards normal cells (HUVEC). An initial biological study has proved that the type of a
substituent in position 3 of the DSG and TGG derivatives significantly affects their activity in vitro. The
compounds containing amino acids such as: L–serine, L–histidine and L–proline as well as dipeptides (–
L–leucyl–L–valine and –L–isoleucyl–L–proline) have proved to be the most active.
3.2.2. Activation of apoptosis in MCF–7 cells
Caspases are a family of endoproteases that provide critical links in cell regulatory networks
controlling inflammation and cell death. Caspases involved in the apoptosis have been subclassified by
their mechanism of action and are either initiator caspases (caspase–8 and –9) or executioner caspases
(caspase–3, –6, and –7) [37]. Caspase–3 and caspase–7 (caspase 3/7) are strongly activated during the
apoptosis, irrespective of the specific death–initiating stimulus, and coordinate the demolition phase of
the apoptosis by cleaving a diverse array of protein substrates [38]. Thus, caspases 3/7 serve as markers
of the apoptosis.
It has been shown previously that in some cancer cells diosgenin and its analogues activate caspases
leading to apoptosis [39]. Having demonstrated that derivative 2c (L–serine in position 3 of TGG) is
highly toxic towards breast cancer cells (especially MCF–7) we examined its proapoptotic activity.
MCF–7 cells were incubated with 2c for 18 hours and subsequently the activity of caspases 3 and 7 was
measured. Staurosporine which inhibits tumor cell growth by inducing cell death via intrinsic apoptotic
pathways [40] was used as the positive control.
Figure 2. Activity of caspase 3/7 in MCF–7 cells treated with the test compound 2c (2,8 µM and 7
µM) or staurosporine for 18 hours. The apoptosis was determined by the Apo–ONE® Homogeneous
Caspase–3/7 Assay (Promega, Madison, WI, USA). Abbreviations: MCF–7 — untreated cells;
Ethanol — MCF–7 cells treated with 1% ethanol. The caspase activation level in the cells exposed to
1% ethanol was normalized to 1.0. Means ± SD are shown. ANOVA, post hoc Tuckey: *p < 0.05, **p
< 0.01, ***p < 0.001, ****p<0.0001 vs. Ethanol (1%).
Data presented in Figure 2 clearly demonstrate that derivative 2c induces apoptosis in MCF–7 cells in a
dose–dependent manner. The incubation of MCF–7 cells with 2c at 2.8 μM (2×IC50) and 7 μM (5×IC50)
led to 1.86–fold and 3.37–fold increase in the activity of caspase 3/7, respectively.
These results suggest that the cytotoxic activity of compound 2c towards MCF–7 cancer cells is likely
due to the induction of apoptosis.
21

3.2.3. Immunomodulatory activity
Six compounds have been selected for the study of their immunomodulatory properties. They
differed in the steroid aglycon part (DSG, TGG) and in the substituents in position 3 of the steroid
skeleton. Their immunomodulatory properties have been determined by examining the effect on the
expression of several cytokine genes (IL–1, IL–4, IL–10, IL–12, TNF–α) essential for the correct
functioning of the human immune system.
The studies have been performed by real time RT–PCR to measure the relative mRNAs levels in
the THP–1 cells after a 6 hours incubation with LPS (5 µg/ml) or the tested compounds (5 µM). Under
these experimental conditions the viability of the THP–1 cells has exceeded 90% (data not shown).
Levels of the analyzed mRNAs in THP–1 cells treated with 1% ethanol have been used as control (K).
LPS (a strong activator of the cytokine expression and secretion) has been used as the reference
compound (positive control).
Out of five tested cytokines, three: IL–1, TNF–α and IL–12 have a pro–inflammatory activity, while
IL–10 is an anti–inflammatory cytokine. Depending on the biological context, IL–4 can have either a
pro– or anti–inflammatory effect.
Pro–inflammatory IL–1, TNF–α and IL–12 are the mediators in various acute and chronic
inflammation linked diseases [41]. The immunomodulatory IL–10 plays a crucial role in the
maintenance of homeostasis, particularly in controlling immunopathology of infectious, allergic and
autoimmune diseases [42]. IL–4 is a cytokine that regulates multiple biological functions. It can regulate
proliferation, apoptosis and plays a critical role in the regulation of immune responses [43].
The results shown in Figure 3 reveal that the effect of compounds 2c, 4c, 9a, 15, 16a, 18 on the
expression of pro–inflammatory IL–1 is similar to that of diosgenin. Interestingly, analogues 4c and 16a
show a statistically significant inhibition of the expression of TNF–α when compared to diosgenin. On
the other hand, 9a stimulates the expression of TNF–α with a statistical significance vs. K. Comparison
of TGG and 4c (a tigogenin derivative) reveals that both compounds have a similar effect on the
expression of the TNF–α, IL–1 and IL–12 genes in the THP–1 cells.
We have also evaluated the effect of the synthesized diosgenin and tigogenin analogues on the
expression of anti–inflammatory IL–10. Our studies indicate that analogues 4c, 9a, 16a cause a
significant induction of the IL–10 expression. This effect is statistically significant vs. both K (untreated
cells) and diosgenin or tigogenin treated cells. Most of the tested compounds also strongly upregulates
the expression of IL–4 in comparison to the untreated cells (K), except for 2c and 9a which do not have
a significant effect on the expression of this cytokine.
From among diosgenin analogues tested, compounds 4c and 16a exhibit the desired
immunomodulatory effects. They do not induce the expression of pro–inflammatory cytokines, while
stimulating the expression of anti–inflammatory IL–10 in the THP–1 monocytes. Moreover, their
immunomodulatory activity is higher than that of diosgenin or tigogenin.
22

Figure 3. Relative expression of the cytokine genes in the THP–1 cells after incubation with the tested
compounds in relation to control (K). Means ± SD are shown. ANOVA, post hoc Tuckey: *p < 0.05,
**p < 0.01, ***p < 0.001, ****p<0.0001 vs. Ethanol (1%).
3.4. Computational methods
Table 4.Predicted ADMET properties calculated for the tested compounds
Compd MWa Dipoleb Volumec SASAd donorHBe accptHBf logPg #metabh Pi Ro5j Ro3k
TGG 416.6 2.05
DSG 1 414.63 1.94
1324.16
1317.60
681.95
685.17
1.00
1.00
1.00
1.00
1.00
2.00
1.00
1.00
2.00
1.00
2.25
3.20
3.20
6.70
5.70
5.00
7.00
5.00
5.00
7.00
5.50
6.25
6.01
6.00
8.73
6.41
7.94
6.52
6.36
7.61
6.58
8.01
6.03
1
3
5
4
3
4
3
3
4
3
5
4180
4172
3414
1116
3826
73
1
1
2
2
2
2
1
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
2a
2c
3a
4c
5a
6a
7a
8a
9a
591.83 5.75
501.71 6.43
527.79 5.43
543.74 7.68
471.68 6.03
513.76 5.54
551.77 11.26
545.82 7.30
626.92 7.56
1956.67 11077.78
1620.07
1749.95
1739.09
1547.09
1699.41
1766.49
1785.13
1847.83
893.33
954.16
955.76
858.38
929.95
966.60
983.09
920.80
1605
3826
764
3038
63
23

10a
11a
12a
13a
14a
15
624.90 11.35
485.71 5.72
561.80 5.82
527.79 5.88
511.74 6.23
512.73 2.71
576.77 3.99
456.66 6.37
1922.44 1006.99
2.00
1.00
1.00
1.00
1.00
0.00
2.00
0.00
8.50
5.00
5.00
5.00
5.00
5.50
5.00
3.50
6.06
6.86
8.58
7.83
6.56
6.46
7.54
6.41
6
3
4
3
4
4
4
2
2
1
2
2
2
2
2
1
1
1
1
1
1
1
1
1
631
2517
3727
3323
3274
1297
2548
1602.34
1820.29
1740.23
1541.81
1602.20
1794.87
1412.44
885.47
993.01
938.66
782.77
829.08
932.28
728.03
16a
18
462
b
c
d
aMW – molecular weight (Da); dipole – dipole moment (D); volume – total molecular volume (Å3); SASA –
e
solvent accessible surface (Å2); donorHB – estimated number of hydrogen bonds that would be donated by the
solute to water molecules in an aqueous solution; faccptHB – estimated number of hydrogen bonds that would be
g
accepted by the solute from water molecules in an aqueous solution; logP – octanol/water partition coefficient;
h#metab – number of likely metabolic reactions; P – apparent Caco–2 permeability (nm/sec); Ro5 – number of
i
j
violations of Lipinski’s rule of five; kRo3 – number of violations of Jorgensen’s rule of three.
Ten predicted ADME properties have been calculated for all analyzed compounds. Principal
descriptors are given in Table 4. Lipinski's rule describing the essential physicochemical characteristics
of well–known medicines with good oral bioavailability determines that the mass of a potential drug
substance should not exceed 500 Da [44]. Molecular weights of the tested molecules fall within the
range 414.627 – 626.918 which constitutes the limits currently suggested for multi–target drugs (130.0
– 725.0). For all compounds the computed dipole moment and the total solvent–accessible volume
values are within the limits traditionally prescribed for those parameters. Total solvent accessible surface
area (SASA) values have been predicted as 1077.783 Å2 for 2a and 1006.987 Å2 for 10a, exceeding the
traditionally cutoff value of 1000.000, but only to a small extent. For other compounds they fall within
the limit of 300.0 – 1000. The estimated number of hydrogen bonds that would be donated by the solute
to the water molecules in an aqueous solution (donorHB) and the estimated number of hydrogen bonds
that would be accepted by the solute from the water molecules (accptHB), same as the number of likely
metabolic reactions (#metab), are in accordance with Lipinski's rule. The logP values fall within the
range 6.000 – 8.729 for the majority of compounds, which is not in an agreement with the values for the
well absorbed compounds (recommended values for log P are –2.0 – 6.5), and indicates even higher
lipophilicities.
The aim of the ADME screening was to eliminate poor lead candidates from the family of
diosgenin and tigogenin derivatives at this early stage of our studies. It has bee demonstrated that the
compounds tested did not meet the requirements of one (TGG, DSG, 5a, 11a, 18) or two (2a, 2c, 3a,
4c, 6a–10a, 12a–14a, 15) requirements of Lipinski’s rule of five and one requirement of Jorgensen’s
rule of three (taking into account aqueous solubility of the compound, Caco–2 cell permeability and
number of primary metabolites). These values indicate the number of potential ADME risk factors that
the compounds might possess in terms of their suitability for an orally active drug. There is, however, a
number of drugs that fall outside of Lipinski’s rule of five, particularly among the natural products and
their derivatives. In some cases, they do not meet the molecular weight and number of hydrogen bond
acceptors criteria [45]. In view of these findings the ADME profile for our diosgenin and tigogenin
derivatives shows that they constitute a promising scaffold for further modification of more potent and
selective agents.
24

3.5. Molecular interaction study of DSG and TGG analogues through docking
The aim of the molecular interaction study has been to explore the molecular interaction of the
DSG and TGG analogues with the estrogen receptor (ER) and glucocorticoid receptor (GR). The
estrogen receptor is an important therapeutic target in the clinical treatment of breast cancer [46], while
the GR is expressed in almost every cell in the body and regulates genes that control, inter alia, the
immune response [47]. Assuming that the structures of diosgenin and tigogenin are similar to that of
glucocorticoids (GCs) which bind to the GR and are important chemicals widely used in the therapy of
inflammatory diseases, we have hypothesized that 4c and 16a might function by affecting the GRs
involved in the anti–inflammatory pathways.
3.5.1. Molecular docking of 2c to the estrogen receptor
Molecular docking of 2c to three different potential binding sites of the estrogen receptor has
revealed that two of these sites show relatively low affinities (low nM for site 3 and high nM for site 2)
with respect to site 1 (Ki = 3.73 pm) which was the binding site of the co–crystalized estradiol derived
metal chelate. We can see that the predicted binding pose of 2c is very similar to the binding pose of the
estradiol derivative, since both ligands occupy the same space and interact with the same residues, apart
from HIS524. 2c is particularly predicted to make two strong hydrogen bonds between the terminal –
OH group and GLU353/ARG394, just like the estradiol derivative from the crystal structure. The
obtained affinity value of Ki = 3.73 pM (Table 5) pm is better than the values obtained for 1 (Ki = 39.65
pm) and TGG (Ki = 26.30 pm) suggesting a more suitable fit to the binding site of the estrogen receptor
than for these two natural compounds, though the binding pose is very similar in all three cases. Indeed,
while the predicted binding pose of all three compounds is comparable, either 1 or TGG can only make
one hydrogen bond to GLU353 which stabilizes these systems in the binding site, but to a lesser degree
than 2c. For the sake of comparison we have also performed molecular docking of diethylstilbestrol, a
known nonsteroidal estrogen receptor agonist used in the past for a variety of medical conditions [48].
The obtained pose was virtually identical with the diethylstilbestrol pose from the crystal structure of its
complex with the estrogen receptor [49], with one of the hydroxyl groups interacting with both GLU353
and ARG394. The estimated Ki for this pose is 26.88 nM, which is four orders of magnitude worse than
for 2c.
Table 5. Estimated computational Ki values from the molecular docking.
compound
estrogen receptor Ki
glucocorticoid receptor Ki
TGG
26.30 pM
39.65 pM
3.73 pM
—
9.17 pM
11.89 pM
—
1
2c
4c
0.23 pM
1.14 pM
—
16a
—
diethylstilbestrol
mifepristone
26.88 nM
—
17.15 pM
25

Figure 4. Comparison of the a) calculated pose of 2c and b) crystal structure pose of the estradiol
derivative in the estrogen receptor binding site.
3.5.2. Molecular docking of 4c and 16a to the glucocorticoid receptor
Molecular docking of 4c and 16a to three different potential binding sites of the GR has revealed that
all three sites show relatively high free energies of binding and the Ki values in the low nM range. Site
3, located at the outer surface of the GR is, however, an unlikely candidate for a binding site. We have
therefore focused our attention on the first two sites. All four best found poses of 4c and 16a in these
two sites are stabilized mostly by hydrophobic/van der Waals interactions with a small number of
hydrogen bonds. The best pose of 16a is stabilized in the binding pocket (corresponding to the binding
pocket of the desisobutyrylciclesonide from the complex crystal structure) by three hydrogen bonds to
ASN564, ARG611 and THR739 with the binding affinity of 1.14 pM (Table 5). Interestingly, 4c in the
same binding site also displays a low Ki of 5.6 pM and a very similar pose stabilized by hydrogen bonds
to ARG611 and THR739 as well as to GLN570. The poses of 4c and 16a in the second site are, on the
other hand, quite different. Here, 16a with the binding affinity of 1.8 pM is also stabilized by a hydrogen
bond to ARG611 and partially occupies the binding pocket of desisobutyrylciclesonide, while 4c is
shifted away from site 1 and is in different orientation than 16a, stabilized by a hydrogen bond to
ARG614 (binding affinity of 0.2 pM). 1 and TGG are predicted to bind preferentially to site 1 with
lower binding affinities than both 4c and 16a (Ki = 9.68 pM in the case 1 and Ki = 11.88 for TGG). It is
interesting to note, however, that TGG is predicted to bind in a pose similar to 4c with the lone hydroxyl
group pointing towards ARG611, while in the case of 1 the hydroxyl group is pointing towards TYR735.
In this case we also decided to compare these results with the binding affinity of a known drug
26

interacting with the glucocorticoid receptor. We have selected mifepristone which is used for medical
abortion [50]. In molecular docking we were able to find mifepristone pose very similar to the one from
the crystal structure [51] with the estimated binding affinity of 521.9 pM, two–three orders of magnitude
worse than for 4c and 16a.
Figure 5. Comparison of the calculated pose of a) 4c and b) 16a (both in site 1) and c) 4c and d) 16a
(site 2) in the GR binding site.
The obtained affinity values for all investigated compounds (in the fM–pM range) seem
unrealistically high, but given the expected accuracy of the docking algorithm of around 1–2 kcal/mol
we can expect affinities in the low nM values for all the tested systems. Unfortunately, while the protocol
used in this investigation is generally accurate in predicting ligand poses within the binding sites of
receptors, it may be often inaccurate when it comes to estimating binding free energies; it has been
shown that the latter values may err by a few orders of magnitude [52,53]. Therefore, we should treat
high computational affinity values only as indicators of possible high–affinity compounds, but such
results need to be later confirmed by experimental estimates. At the same time, with relatively high
confidence we can treat the obtained ligand poses as accurate.
4. Conclusion
In conclusion, we have synthesized a set of diosgenin and tigogenin derivatives substituted with
various amino acids, a dipeptide or levulinic and 3,4–dihydroxycinnamic acid at C–3 position that have
been further evaluated for their antiproliferative and anti–inflammatory activity. Diosgenin and
tigogenin analogues display a high antiproliferative activity towards breast (MCF–7 and MDA–MB–
27

231) and prostate (PC–3) human cancer cell. The serine derivative of tigogenin 2c shows the highest
cytotoxic activity towards MCF–7 cells and induces apoptosis. Possible mechanism of the
antiproliferative action of compound 2c is through the caspase dependent apoptosis pathway.
Diosgenin and tigogenin conjugates also show immunomodulatory properties. Structure–activity
relationship studies indicate that compounds 4c and 16a in most cases exhibit desired effects. They do
not induce the gene expression of the pro–inflammatory cytokines (i.e. IL–12, IL–1, TNF–α) but
strongly stimulate the expression of anti–inflammatory IL–10 and IL–4.
Ten predicted ADME properties have been calculated for all the obtained compounds. The
compounds tested have not met the requirements of one (1, 5a, 11a, 18) or two (2a, 2c, 3a, 4c, 6a–10a,
12a–14a, 15) regulations of Lipinski’s rule of five and one requirement of Jorgensen’s rule of three
(taking into account only such parameters as molecular weight and the number of hydrogen bond donors
and acceptors). The synthesized compounds exhibit reasonable ADME properties, which may be an
indication of their good oral drug–like behavior.
A molecular docking study has revealed a high binding affinity of diosgenin and tigogenin
analogues to the active site of the glucocorticoid receptor (GR) and the estrogen receptor (ER). Our
findings have shown that analogue 2c interacts with amino acids GLU353/ARG394 at the binding site
of ER with inhibition constant Ki = 3.73 pM, This value is four orders of magnitude better than for
reference diethylstilbestrol. Compounds 4c and 16a also revealed strong binding interaction with the
active site GR with the inhibition constant Ki = 0.23 pM and Ki = 1.14 pM respectively. The estimated
binding affinity (Ki) of 4c and 16a is two–three orders of magnitude better than for the known drug
mifepristone. The molecular docking results indicated that 2c, 4c and 16a derivatives exhibited a higher
affinity for the ER and the GR active site than the known drug.
Promising in vivo anti–tumor and immunomodulating activity, molecular docking, and drug–
likeness properties of compounds 2c, 4c, 16a indicate that these compounds can serve as potential lead
compounds for further development and investigation of novel anticancer and immunomodulatory
agents.
Conflict of interest
There is no conflict of interest.
Acknowledgements
This work was supported under the framework of the statutory project funded by the Polish Ministry of
Science and Higher Education.
References
[1]a) L. Fernández–Cabezón, B. Galán, J.L. García New Insights on Steroid Biotechnology, Front.
Microbiol. 9 (2018) 958. https://doi.org/10.3389/fmicb.2018.00958. b) A. Parus, Postępy Fitoterapii,
Pharmacological activities of saponins, 3 (2013) 200–204.
[2] G. Gu, L. An, M. Fang, Z. Guo, Efficient one–pot synthesis of tigogenin saponins and their
antitumor activities, Carbohydr. Res. 383 (2014) 21–26. https://doi.org/10.1016/j.carres.2013.10.015.
28

[3] a) Z. Wang , Y. Cheng , N. Wang , D. M. Wang , Y. W. Li , F. Han , J. G. Shen , D. P. Yang , X.
Y. Guan, J. P. Chen, Dioscin induces cancer cell apoptosis through elevated oxidative stress mediated
by downregulation of peroxiredoxins, Cancer Biol. Ther. 13 (2012) 138–147.
https://doi.org/10.4161/cbt.13.3.18693. b) K. Patel, M. Gadewar, V. Tahilyani, D. K. Patel, A review
on pharmacological and analytical aspects of diosgenin: a concise report, Nat. Prod. Bioprospect. 2
(2012) 46–52. https://doi.org/10.1007/s13659–012–0014–3. c) J. Raju, R. Mehta, Cancer
chemopreventive and therapeutic effects of diosgenin, a food saponin, Nutr. Cancer 61 (2009) 27–35.
https://doi.org/10.1080/01635580802357352.
[4] F. Wang, J. Tang, Z. Wu, Flora of China; Science Press, 1980.
[5] C. Corbiere, B. Liagre, A. Bianchi, K. Bordji, M. Dauca, P. Netter, J. L. Beneytout, Different
contribution of apoptosis to the antiproliferative effects of diosgenin and other plant steroids,
hecogenin and tigogenin, on human 1547 osteosarcoma cells, Int. J. Oncol. 22 (2003) 899–905.
https://doi.org/10.3892/ijo.22.4.899.
[6] G. L. Li, H. J. Xu, S. H. Xu, W. W. Wang, B. Y. Yu, J. Zhang, Synthesis of tigogenin MeON–
Neoglycosides and their antitumor activity, Fitoterapia 125 (2018) 33–40.
https://doi.org/10.1016/j.fitote.2017.12.014.
[7] a)J. Raju, J. M. Patlolla, M. V. Swamy, C. V. Rao, Diosgenin, a steroid saponin of
Trigonellafoenum graecum (Fenugreek), inhibits azoxymethane–induced aberrant crypt foci formation
in F344 rats and induces apoptosis in HT–29 human colon cancer cells, Cancer Epidemiol. Biomarkers
Prev. 13 (2004) 1392–1398; b) V. S. Malisetty, J. M. R. Patlolla, J. Raju, L. A. Marcus, C. L. Choi, C.
V. Rao, Chemoprevention of colon cancerby diosgenin, a steroidal saponin constituent of fenugreek,
Proc. Am. Assoc. Cancer Res. 46 (2005) 2473; c) T. Yamada, M. Hoshino, T. Hayakawa, H. Ohhara,
H. Yamada, T. Nakazawa, T. Inagaki, M. Iida, T. Ogasawara, A. Uchida, C. Hasegawa, G. Murasaki,
M. Miyaji, A. Hirata, T. Takeuchi, Dietary diosgenin attenuates subacute intestinal inflammation
associated with indomethacin in rats, Am. J. Physiol. 273 (2009) G355–G364.
https://doi.org/10.1152/ajpgi.1997.273.2.G355.
[8] C. Lepage, B. Liagre, J. Cook–Moreau, A. Pinon, J. L. Beneytout, Cyclooxygenase–2 and 5–
lipoxygenase pathways in diosgenin–induced apoptosis in HT–29 and HCT–116 colon cancer cells,
Int. J. Oncol. 36 (2010) 1183–1191. https://doi.org/10.3892/ijo_00000601.
[9] J. Raju, R. P. Bird, Diosgenin, a naturally occurring furostanol saponin suppresses 3–hydroxy–3–
methylglutaryl CoA reductase expression and induces apoptosis in HCT–116 human colon carcinoma
cells, Cancer Lett. 255 (2007) 194–204. https://doi.org/10.1016/j.canlet.2007.04.011.
[10] Ch. T. Chiang, T .D. Way, S. J. Tsai, J. K. Lin, Diosgenin, a naturally occurring steroid,
suppresses fatty acid synthase expression in HER2–overexpressing breast cancer cells through
modulating Akt, mTOR and JNK phosphorylation, FEBS Lett. 581 (2007) 5735–5742.
https://doi.org/10.1016/j.febslet.2007.11.021.
[11] a)S. Srinivasan, S. Koduru, R. Kumar, G. Venguswamy, N. Kyprianou, C. Damodaran, Diosgenin
targets Akt-mediated prosurvival signaling in human breast cancer cells, Int. J. Cancer 125 (2009)
961–967. https://doi.org/10.1002/ijc.24419. b)S. Sowmyalakshmi, R. Ranga, C.G Gairola, D. Chendil,
Effect of diosgenin (Fenugreek) on breast cancer cells. Proc. Am. Assoc. Cancer Res. 46 (2005) 1382.
[12] P. S. Chen, Y. W. Shih, H. Ch. Huang, H. W. Cheng, Diosgenin, a steroidal saponin, inhibits
migration and invasion of human prostate cancer PC–3 cells by reducing matrix metalloproteinases
expression, PLoS ONE 6 (2011) e20164. https://doi.org/10.1371/journal.pone.0020164.
[13] F. Li, P. P. Fernandez, P. Rajendran, K. M. Hui, G. Sethi, Diosgenin, a steroidal saponin, inhibits
STAT3 signaling pathway leading to suppression of proliferation and chemosensitization of human
hepatocellular carcinoma cells, Cancer Lett. 292 (2010) 197–207.
https://doi.org/10.1016/j.canlet.2009.12.003.
[14] M. J. Liu, Z. Wang, Y. Ju, R. N. S. Wong, Q. Y. Wu, Diosgenin induces cell cycle arrest and
apoptosis in human leukemia K562 cells with the disruption of Ca2+ homeostasis, Cancer Chemother.
Pharmacol. 55 (2005) 79-90. https://doi.org/10.1007/s00280-004-0849-3.
[15] a) B. Liagre, P. Vergne–Salle, C. Corbiere, J. L.Charissoux, J. L. Beneytout, Diosgenin, a plant
steroid, induces apoptosis in human rheumatoid arthritis synoviocytes with cyclooxygenase–2
overexpression, Arthritis Res. Ther. 6 (2004) R373.https://doi.org/10.1186/ar1199.b) J. Stefanowicz-
Hajduk, J. R. Ochocka, Steroidal saponins – occurrence, characteristic and application in therapeutics,
Postępy Fitoterapii 1 (2006) 36;c) B. Liagre, P. Vergne–Salle, D. Y. Leger, J. L. Beneytout, Inhibition
29