Synthesis and antimicrobial activity of (?)-cleistenolide and analogues

Article abstract of DOI:10.1016/j.bioorg.2020.104491

Using the “chiral pool” approach, two modified total syntheses of the biologically active δ-lactone cleistenolide (1) have been achieved starting from D-glucose. These approaches also enabled the preparation of novel analogues and derivatives of natural product 1. The applied strategy for the synthesis of 1 involves: the initial degradation of the chiral precursor for a single C-atom, C₂-fragment chain extension using Z-selective Wittig reaction, and the final δ-lactonization. All tested cleistenolide analogues displayed antimicrobial activity against a panel of nine microbial strains, most of them superseding the activity of cleistenolide itself, and, in some cases, coming close in value to the observed minimal inhibitory concentrations of chloramphenicol. Increased lipophilicity of the derivatives and the non-sterically congested conjugated lactone moiety were a prerequisite for analogues with high inhibitory activity against S. aureus and, in general, Gram-positive bacteria.

Full text of DOI:10.1016/j.bioorg.2020.104491

Journal Pre-proofs
Synthesis and antimicrobial activity of (−)-cleistenolide and analogues
Goran Benedeković, Mirjana Popsavin, Niko S. Radulović, Zorica
Stojanović-Radić, Sándor Farkas, Jovana Francuz, Velimir Popsavin
PII:
S0045-2068(20)31789-2
DOI:
https://doi.org/10.1016/j.bioorg.2020.104491
YBIOO 104491
Reference:
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
20 May 2020
8 November 2020
16 November 2020
Please cite this article as: G. Benedeković, M. Popsavin, N.S. Radulović, Z. Stojanović-Radić, S. Farkas, J.
Francuz, V. Popsavin, Synthesis and antimicrobial activity of (−)-cleistenolide and analogues, Bioorganic
Chemistry (2020), doi: https://doi.org/10.1016/j.bioorg.2020.104491
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.
© 2020 Elsevier Inc. All rights reserved.

Graphical Abstract
Synthesis and antimicrobial activity of (−)-cleistenolide and analogues
Goran Benedeković a, Mirjana Popsavin a, Niko S. Radulović b, Zorica Stojanović-Radić c, Sándor
Farkas a, Jovana Francuz a, and Velimir Popsavin a,d,*
a Department of Chemistry, Biochemistry and Environmental Protection, Faculty of Sciences,
University of Novi Sad, Trg Dositeja Obradovića 3, Novi Sad, Serbia
b Department of Chemistry, Faculty of Sciences and Mathematics, University of Niš, Višegradska
33, Niš, 18000, Serbia
c Department of Biology and Ecology, Faculty of Sciences and Mathematics, University of Niš,
Višegradska 33, Niš, 18000, Serbia
d Serbian Academy of Sciences and Arts, Knez Mihajlova 35, 11000 Belgrade, Serbia
* Corresponding author. E-mail address: velimir.popsavin@dh.uns.ac.rs (V. Popsavin).
1

Synthesis and antimicrobial activity of (−)-cleistenolide and analogues
Goran Benedeković a, Mirjana Popsavin a, Niko S. Radulović b, Zorica Stojanović-Radić c, Sándor
Farkas a, Jovana Francuz a, and Velimir Popsavin a,d,*
a Department of Chemistry, Biochemistry and Environmental Protection, Faculty of Sciences,
University of Novi Sad, Trg Dositeja Obradovića 3, Novi Sad, Serbia
b Department of Chemistry, Faculty of Sciences and Mathematics, University of Niš, Višegradska
33, Niš, 18000, Serbia
c Department of Biology and Ecology, Faculty of Sciences and Mathematics, University of Niš,
Višegradska 33, Niš, 18000, Serbia
d Serbian Academy of Sciences and Arts, Knez Mihajlova 35, 11000 Belgrade, Serbia
Abstract
Using the "chiral pool" approach, two modified total syntheses of the biologically active δ-lactone
cleistenolide (1) have been achieved starting from D-glucose. These approaches also enabled the
preparation of novel analogues and derivatives of natural product 1. The applied strategy for the
synthesis of 1 involves: the initial degradation of the chiral precursor for a single C-atom, C2-
fragment chain extension using Z-selective Wittig reaction, and the final δ-lactonization. All
tested cleistenolide analogues displayed antimicrobial activity against a panel of nine microbial
strains, most of them superseding the activity of cleistenolide itself, and, in some cases, coming
close in value to the observed minimal inhibitory concentrations of chloramphenicol. Increased
lipophilicity of the derivatives and the non-sterically congested conjugated lactone moiety were a
prerequisite for analogues with high inhibitory activity against S. aureus and, in general, Gram-
positive bacteria.
Keywords: cleistenolide; synthesis of analogues; pyranones; antimicrobial lactones; hierarchical
clustering
* Corresponding author. E-mail address: velimir.popsavin@dh.uns.ac.rs (V. Popsavin).
2

1. Introduction
(−)-Cleistenolide (1, Figure 1), is a naturally occurring α,β-unsaturated δ-lactone that was
isolated from Cleistochlamys kirkii Oliver, and exhibited in vitro antibacterial activity against
Staphylococcus aureus and Bacillus anthracis and antifungal activity against Candida albicans [1].
A recent study additionally revealed that (−)-cleistenolide (1) exhibits antibacterial activity against
some Gram-positive strains [2]. Ten total syntheses of the natural product 1 have been completed so
far [3], using different methods such as ring closing metathesis [3–8], Z-selective olefination [9–11]
or Yamaguchi lactonization [12] as the key steps. The most efficient route (49% yield over eight
steps) is achieved by Du, Linhardt et al. [12] starting from D-arabinose. The second synthesis by
efficiency was realized starting from D-glucose, in six steps, and in an overall yield of 42.5% [9].
Despite the existence of these numerous synthetic pathways to 1, there is still a demand to
develop alternative synthetic procedures, especially those that would enable the preparation of
novel cleistenolide analogues.
Herein, we present two alternative approaches to the total syntheses of (−)-cleistenolide (1),
and its new analogues 3–10, as well as their preliminary antimicrobial screening.
O
O
O
O
R
O
R
O
O
O
O
O
R1
R
O
RO
O
RO
OR1 OH
OR1 OR2
OAc OAc
OAc OAc
5 R = R1 = R2 = Bn
6 R = Bz, R1 = Ac, R2 = Bn
7 R = R1 = R2 = Bz
1 R = R1 = H (Cleistenolide)
2 R = H, R1 = OMe
3 R = Cl, R1 = H
9 R = R1 = H
10 R = Bz, R1 = Ac
4
8 R = R1 = R2 = 3-MeOBz
Fig. 1. Chemical structures of (−)-cleistenolide (1) and the corresponding analogues (2–10).
Compounds 3 and 2 are direct isosteres of the lead 1, while the analogue 4 represents a new
hybrid of cinnamic acid and the natural product 1. Cinnamate hybrids are a unique family of
compounds with multiple biological activities [14], including antimicrobial activity [15]. Finally,
we have designed four highly hydrophobic (5–8) and two hydrophilic derivatives (9 and 10) having
an α,β-unsaturated δ-lactone scaffold. The rationale for obtaining this type of compound is to
evaluate the effect of their lipophilic/hydrophilic properties on antimicrobial activity.
Part of this work was recently published in the form of a preliminary communication,
describing that the natural product 1 and some of its analogues strongly inhibit the growth of
selected tumour cells [13]. Considering the problem of bacteraemia and invasive candidiasis in
3

cancer patients, an important task of this work was to prepare and test the antimicrobial activity of 1
and related compounds for a possible use in multifunctional therapy [16].
2. Results and discussion
2.1. Chemistry
The synthesis of 1 and 9 is summarized in Scheme 1. The sequence started from commercially
available diacetone D-glucose 11.
RO
RO
BnO
BnO
BnO
BnO
O
O
O
O
OH
d
a
OH
c
O
O
O
O
O
O
OH
HO
RO
BnO
BnO
12 R = H
13 R = Bn
11
b
14
15
e
h
O
O
O
RO2C
OH
g
f
O
O
O
O
O
BnO
HO
BnO
OBn OBn
OAc OAc
OH OH
OBn OBn
16 R = Et
16a R = Me
9
()-Cleistenolide (1)
5
Scheme 1. Reagents and conditions: (a) 60% aq AcOH, 24 h, rt; (b) NaH, DMF, 0 °C (1 h) → rt (2 h), 90% from 11; (c) 50% aq
TFA, 18 h, rt, 83%; (d) H5IO6, H2O, EtOAc, 2.5 h, rt; (e) Ph3P=CH-CO2Et, MeOH, 20 h, rt, 86% from 14; (f) TsOH, CH2Cl2, 12 h,
rt, 92%; (g) BzBr, FeCl3, rt, 2 h, then AcBr, rt, 20 h, 74%; (h) FeCl3, CH2Cl2, 20.5 h, rt, 92%.
Hydrolytic removal of the terminal isopropylidene protection produced monoacetonide 12
which was not purified, but was subsequently treated with benzyl bromide in dry DMF, in the
presence of sodium hydride, to give the known [17,18] tri-O-benzyl derivative 13 in 90% overall
yield (from two steps). Hydrolytic removal of the isopropylidene protecting group in 13 gave the
corresponding lactol 14 as only reaction product. Although lactol 14 is most likely a mixture of α-
1
13
and β-anomer, their mutual ratio cannot be determined due to the complexity of both H and C
NMR spectra. The purity of product 14 was confirmed by HRMS data [m/z 451.2112 (M++H), calcd
for C24H27O8: 451.2115]. Oxidative cleavage of 14 with periodic acid, followed by Z-selective
Wittig olefination [19,20] of the resulting aldehyde 15 with stabilized C2-ylide (Ph3P=CHCO2Et),
afforded the unsaturated ester 16 (72% from the last three steps). The unsaturated ester 16 cyclised
to 5 after treatment with toluene-4-sulfonic acid in anhydrous dichloromethane. The desired lactone
5 was thus obtained in a yield of 92%. Treatment of 5 with BzBr/FeCl3 reagent system resulted in
the selective introduction of the benzoyl group into the primary position. The subsequent in situ
4

acetylating of the resulting diol (not shown in the Scheme 1) yielded the natural product 1 in 74%
yield. This modified synthesis of 1 was accomplished in seven synthetic steps, in a total yield of
44%. In terms of yield, this synthesis is slightly less efficient than the synthesis reported by Du,
Linhardt et all. [11] (49% from eight steps). However, our synthesis is for one step shorter. In order
to optimize the initial stages of the synthesis, sodium periodate was used for diol cleavage, the
Wittig reaction was performed using Ph3P=CH-CO2Me, and the initial five-step sequence was
carried out without isolation or purification of intermediates 12, 13, 14 and 15. This provided the
key intermediate 16a in a total yield of 75% from five synthetic steps.
This approach to the synthesis of (−)-cleistenolide (1) is somewhat similar to the Yadav's route
that proceeded in six steps from D-glucose diacetonide, in 42.5% overall yield [9]. Our approach
(Scheme 1) initially provided natural product 1 in almost equal total yield (44% from the seven
synthetic steps). However, the optimization of some key steps, as mentioned above increased the
yield of target 1 to 65.6% (from six steps).
Finally, to obtain the triol 9, which is a possible divergent intermediate for the synthesis of
various new (−)-cleistenolide analogues and derivatives, FeCl3-mediated global deprotection of
benzyl groups in 5 was performed to obtain the desired triol 9 in 92% yield.
An alternative synthesis of cleistenolide (1) is shown in Scheme 2. The 3-O-benzyl-1,2:5,6-di-
O-isopropylidene-α-D-glucofuranose (17), which is readily available from D-glucose [21], was used
as a convenient starting compound in this approach.
BzO
AcO
BzO
AcO
O
RO
HO
O
O
O
O
OH
a
c
d
O
O
O
O
O
O
O
BnO
OH
BnO
BnO
BnO
18 R = H
19 R = Bz
17
21
20
b
e
O
O
RO2C
OH
BzO
AcO
OH
g
i
f
O
O
O
BzO
O
BzO
O
BnO
OAc OBn
OAc OAc
OAc OR
6 R = Bn
10 R = H
23 R = Et
24 R = Me
()-Cleistenolide (1)
h
22
Scheme 2. Reagents and conditions: (a) 60% aq AcOH, rt, 24 h; (b) BzCl, DMAP, CH2Cl2, rt, 1 h; (c) AcCl, DMAP, rt, 1 h, 69%
from 17; (d) 90% aq TFA, CH2Cl2, 0 °C (0.5 h) → rt (2 h); (e) H5IO6, EtOAc, H2O, 0 °C (0.5 h) → rt (3.5 h); (f) Ph3P=CH-CO2Et,
MeOH, rt, 20 h; (g) TsOH, CH2Cl2, rt, 20 h, 38% from 20; (h) TiCl4, CH2Cl2, rt, 24 h, 70%; (i) Ac2O, TsOH, rt, 4 h, 95%.
5

Selective hydrolytic removal of the exocyclic isopropylidene protection in 17 gave the
corresponding diol 18, which was not purified but allowed to react in crude form with BzCl in
anhydrous CH2Cl2, in the presence of DMAP. Only the primary OH group is benzoylated under
these reaction conditions (~1 eq of BzCl, rt, 1 h). Subsequent in situ acetylating of the resulting
secondary alcohol 19 gave the known [22] fully protected derivative 20 in a total yield of 69%
(from 3 steps). Compound 20 was further treated with 9:1 aqueous trifluoroacetic acid in CH2Cl2, at
0 °C for 0.5 h, then at room temperature for 2 h, to give the corresponding lactol 21. The product is
a mixture of the corresponding α- and β-anomers (β/α = 2:1) as determined by the integration of
anomeric proton signals in the 1H NMR spectrum [δ 5.23 (s, 1 H, H-1β), 5.54 (d, 0.5 H, J1,2 = 4.0
Hz, H-1α)]. Oxidative cleavage of 21 with periodic acid in an ethyl acetate and water gave the
aldehyde 22. Due to its instability the aldehyde 22 was subjected directly to a C2 homologation
using Z-selective olefination conditions [19] to afford the corresponding Z-enoates as a 1:1 mixture
of ethyl (23) and methyl ester (24). The methyl ester 24 is obviously the product of trans-
esterification of 23 with methanol. The similar reaction was not observed in the Wittig reaction of
15 under the same conditions (Scheme 1). The mixture of esters was not separated, but was further
converted to the pyranone 6, after treatment with toluene-4-sulfonic acid in anhydrous
dichloromethane. Lactone 6 was obtained in 38% overall yield (four steps). Deprotection of the
benzyl group in 6 was achieved with titanium(IV) chloride in anhydrous dichloromethane, to afford
alcohol 10 in 70% yield. Acetylation of 10 (Ac2O, TsOH, rt, 4 h) gave the natural product 1 in 95%
yield. Although the overall yield of this synthetic pathway is lower (17.4% from 9 steps) than the
previous one, it has an important advantage because it provides access to a divergent intermediate
(10) for the synthesis of various novel C-4 substituted analogues.
The melting point and spectral data of thus obtained synthetic sample 1 were in excellent
1
agreement with data reported in the literature [3–13]. H and 13C NMR spectra (see the
Supplementary material) confirmed the purity of the synthesized natural product. However, the
value of optical rotation was different. We observed the values of [α]D = −212.2 (c 0.5, CHCl3), and
[α]D = −212.4 (c 0.5, CH2Cl2). Lit [1] recorded for the isolated natural product: [α]D = −63.5 (c 0.7,
CHCl3); the closest literature [8] value for the synthetic sample: [α]D = −164.6 (c 0.5, CH2Cl2). At
the moment, we have no explanation for this discrepancies.
The conversion of triol 9 to cleistenolide analogues 2–4, and related 5,6-dihydro-2H-pyran-2-
one derivatives 7 and 8 is shown in Table 1.
6

Table 1. Preparation of (−)-cleistenolide analogues 2–4, 7 and 8.
Product
(yield)
Entry
Reagents and conditions
Ar
R
1
2
3
4
5
(a) 3-methoxybenzoyl chloride, Py/CH2Cl2, rt, 4 h; (b) Ac2O, rt, 20 h
(a) 2,4,6-trichlorobenzoyl chloride, Py/CH2Cl2, rt, 4 h; (b) Ac2O, rt, 20 h
(a) cinnamoyl chloride, Py/CH2Cl2, rt, 2 h; (b) Ac2O, rt, 18 h
benzoyl chloride, Py/CH2Cl2, rt, 3.5 h
3-methoxyphenyl
2,4,6-trichlorophenyl
styryl
phenyl
3-methoxyphenyl
Ac
Ac
Ac
Bz
2 (67%)
3 (85%)
4 (47%)
7 (86%)
8 (88%)
3-methoxybenzoyl chloride, Py/CH2Cl2, rt, 3.5 h
3-methoxy-Bz
The first three analogues 2–4 were prepared through a two-step sequence consisting of the
initial selective arylation of 9 with one molar equivalent of ArCOCl, followed by acetylating of the
intermediately formed diol (entries 1–3). The cleistenolide isosteres 2 and 3 were obtained as the
only reaction products in good yields (entries 1 and 2), while the cinnamic acid hybrid (4) was
obtained in a slightly lower yield (entry 3). Treatment of 9 with six molar equivalents of benzoyl
chloride, or 3-methoxybenzoyl chloride gave excellent yields of the corresponding ester derivatives
7 (entry 4), or 8 (entry 5).
All these analogues (Table 1), along with the lead 1 and 5,6-dihydro-2H-pyran-2-one
derivatives (5, 6, 9 and 10), used as intermediates in the synthetic pathways described (Schemes 1
and 2), were evaluated for their antimicrobial activity.
2.2. Antimicrobial activities
Cleistenolide and the mentioned analogues were evaluated for their antimicrobial activity
against a panel of ATCC microbial strains, four Gram-positive and four Gram-negative bacteria,
and one yeast. The tested compounds were shown to possess inhibitory action (MIC) in the range
0.04–6.00 μM/ml (Table 2) and exhibited a generally wide spectrum of activity by affecting the
Table 2. Antimicrobial activity (MIC in µmol/ml) of cleistenolide and its analogues.
Strain/compound
1
2
3
4
5
6
7
9
8
10
CHL*
NYS*
Gram-positive bacteria
S. aureus
0.75
0.75
0.37
1.50
1.50
0.75
0.75
1.50
0.37
3.00
0.18
0.18
0.37
0.75
0.09
0.75
1.50
0.75
6.00
6.00
0.37
1.50
0.09
0.04
0.18
1.50
0.09
3.00
6.00
>6.00
6.00
0.37
3.00
0.75
6.00
0.75
0.75
0.18
0.75
19.4
4.82
2.41
9.65
NT
NT
NT
NT
S. epidermidis
K. rhizophila
B. cereus
6.00
Gram-negative bacteria
E. coli
6.00
6.00
3.00
6.00
6.00
6.00
3.00
6.00
6.00
6.00
3.00
6.00
6.00
3.00
1.50
6.00
6.00
3.00
6.00
6.00
1.50
6.00
1.50
3.00
3.00
6.00
6.00
6.00
6.00
6.00
6.00
6.00
>6.00
3.00
3.00
6.00
3.00
6.00
1.50
6.00
4.82
38.68
38.68
9.65
NT
NT
NT
NT
P. aeruginosa
A. baumannii
S. enteritidis
Yeast
C. albicans
3.00
1.50
3.00
1.50
6.00
0.37
6.00
6.00
>6.00
6.00
NT
2.52
CHL – chloramphenicol; NYS – nystatin; * – results given in nM/ml, NT – not tested.
7

growth of all tested strains. The only exceptions were 8 and 9 which failed to be active at the tested
concentration range against two and one strains, respectively.
The highest activity was observed in the case of compound 6 against Gram-positive B. cereus,
where the growth inhibition was achieved at the concentration of 40 nM/ml, which was only ca. 4
times lower compared to chloramphenicol. Considering the overall susceptibility of the strains, the
lowest average MIC of 1.2 μM/ml was determined for S. aureus. Among the tested compounds, 4
and 6 exhibited the highest antimicrobial effect with average MIC values of 2.21 and 1.59 μM/ml,
respectively. On the other hand, the lowest antimicrobial action was noted in the case of compounds
5 and 9, which had average MIC values of 4.58 and 6.00 μM/ml, respectively. Among the tested
compounds, it appears that the Gram-positive strains showed a significantly higher susceptibility to
the action of all tested compounds. On the other hand, the yeast, C. albicans, seemed to be affected
in a similar way to that observed for Gram-negative strains. Among the tested strains, the Gram-
negative strains S. enteritidis and P. aeruginosa demonstrated the highest resistance.
To further obtain insights into the antimicrobial nature of cleistenolide and its congeners, a
hierarchical clustering analysis of the activity data was undertaken. Minimal inhibitory
concentrations were utilized in the statistical treatment and the results are presented in Figure 2 as
the heatmap and two dendrograms.
Fig. 2. Heatmap representing color-coded MIC values of the tested cleistenolide and its analogues with row and column dendrograms
obtained by a hierarchical clustering analysis (Euclidian distance and average linkage clustering method.
Clearly visible is the mentioned higher activity of almost all cleistenolide analogues toward
Gram-positive bacteria (a cluster of three bacteria S. aureus, S. epidermidis, and K. rhizophila), as
well as a lower susceptibility of the majority of Gram-negative strains (the cluster composed of P.
8

aeruginosa, E. coli, and S. enteritidis). Anticandidal activity was intermediate, as that observed for
one Gram-positive (B. cereus) and Gram-negative (A. baumannii). This pronounced, alas
unquantified [1], activity against S. aureus was noted initially when cleistenolide was originally
isolated. The seminal work [1] states, again without MIC values, that cleistenolide in addition to
antistaphylococcal activity was active against B. anthracis and C. albicans.
Based on the clustering of the tested derivatives, it is evident data the most hydrophilic tested
compound (9) was the least active one, while still preserving some unselective activity in the
highest tested concentration. The sterically congested (second cluster) per-O-benzyl ether derivative
(5) and the electron-rich trimethoxybenzoate (8) appear to have a somewhat decreased antimicrobial
effect compared to the mostly acetyl derivatives (the remaining compounds forming the third
cluster). It follows that the likely pharmacophore is the Michael acceptor, α,β-unsaturated lactone,
as all of the derivatives showed a certain degree of activity; however, two factors influence the
expression of this activity: lipophilicity and steric hindrance, the first being positive, and the second
one negatively influencing the electrophilicity of the conjugated system. The lack of
lipopolysaccharides in Gram-positive bacteria, while these are a major component of the outer
membrane of Gram-negative bacteria, contributing greatly to the structural integrity of the bacteria,
appears to correlate well with the observed higher susceptibilities of Gram-positive bacteria; it
suggests the cell membrane as a possible cellular target of cleistenolide and hence the increased
lipophilicity as a structural requirement of higher active analogues. It seems that the activities could
be fine-tuned by altering the identities of the esterifying aromatic acids furthest to the
pharmacophore and by minding the steric requirements of the remaining, preferentially lipophilic,
more proximal O-substituents.
3. Conclusions
Two independent synthetic sequences have been developed for the total synthesis of the natural
biologically active lactone cleistenolide (1), starting from commercially available D-glucose
derivatives. The first route (Scheme 1) represents a modified procedure recently published as a
preliminary communication [4]. After optimizing the reaction conditions of several key synthetic
steps, the target 1 was obtained in an overall yield of 43.7% from seven steps. Obviously, the
modified procedure is significantly more efficient than that described in the preliminary
communication [4] which gave the target 1 in an overall yield of 33% (from seven steps). However,
after additional optimizations of the initial stages of the synthesis, the natural product 1 was
obtained in a total yield of 65.6% over 6 steps. This is the most efficient total synthesis of (−)-
9

cleistenolide so far. The second synthetic pathway (Scheme 2) gave the natural product 1 in a lower
overall yield (17.4% from 9 steps). Nevertheless, this approach has the important advantage of
providing access to an important intermediate (compound 10), suitable for the preparation of a
variety of novel C-4 substituted analogues. By combining both approaches mentioned above,
several novel analogues and derivatives of 1 were synthesized and their antimicrobial activity was
evaluated. The structural variations in the tested cleistenolide derivatives all resulted in compounds
that possess a wide range of antimicrobial activities, and that display, in some cases, MIC values
close in to that of commercial antibiotic chloramphenicol. Increased lipophilicity of the derivatives
and the non-sterically congested conjugated lactone moiety were a prerequisite for analogues with
high inhibitory activity against S. aureus and, in general, Gram-positive bacteria.
4. Experimental section
4.1.1. General experimental procedures
For general experimental procedures and for the preparation of initial intermediate 17, see the
Supplementary data.
4.1.2. 3,5,6-Tri-O-benzyl-1,2-O-isopropylidene-α-D-glucofuranose (13)
A solution of 11 (2.45 g, 9.41 mmol) in 60% aq AcOH was stirred at room temperature for 24 h
and then evaporated by co-distillation with toluene. The remaining crude 12 was dried in high
vacuum over night. To a cooled (0 °C) and stirred solution of crude 12 in anhydrous DMF (70 mL)
was slowly added 95% NaH (2.0 g, 83.33 mmol) in portions during 0.5 h. When the evolution of
hydrogen was ceased, BnBr (2.0 mL, 16.82 mmol) was added and the resulting mixture was stirred
at 0 °C for 0.5 h. The cooling bath was removed and the stirring was continued at room temperature
for 2 h. The excess of NaH was destroyed by treatment with absolute MeOH (10 mL) and the
volatiles were evaporated. The residue was distributed between CH2Cl2 (40 mL) and H2O (300 mL),
and extracted successively with CH2Cl2 (3 × 40 mL) and EtOAc (30 mL). Organic phases were
combined, dried and evaporated. The remaining crude mixture was purified on a column of flash
23
silica (3:1 light petroleum/Et2O) to give pure 13 (4.179 g, 90%) as a colourless glassy solid, [α]D
=
−45.2 (c 1.0, CHCl3), lit. [17] [α]D = −33.0 (c 9.3, CHCl3), lit. [18] [α]D = −61.8 (c 0.3, CHCl3), Rf =
0.42 (7:3 light petroleum/Et2O). IR (film): νmax 1605, 1586, 1078, 1027. 1H NMR (400 MHz,
CDCl3): δ 1.39 and 1.58 (2 × s, 3 H each, Me2C), 3.79 (dd, 1 H, J5,6a = 5.8, J6a,6b = 10.6 Hz, H-6a),
4.01 (dd, 1 H, J5,6b = 1.7, J6a,6b=10.6 Hz, H-6b), 4.17 (ddd, 1 H, J5,6b = 1.7, J5,6a = 5.8, J4,5 = 9.3 Hz,
H-5), 4.22 (d, 1 H, J3,4 = 2.9 Hz, H-3), 4.42 (dd, 1 H, J3,4 = 2.9, J4,5 = 9.3 Hz, H-4), 4.54–4.94 (m, 7
10

H, H-2 and 3 × CH2Ph), 6.00 (d, 1 H, J1,2 = 3.6 Hz, H-1), 7.28–7.56 (m, 15 H, 3 × Ph). HRMS
(ESI): m/z 513.2239 (M++Na), calcd for C30H34NaO6: 513.2248.
4.1.3. Ethyl (2Z)-4,6,7-tri-O-benzyl-2,3-dideoxy-D-arabino-hept-2-enoate (16)
A solution of 13 (2.16 g, 0.28 mmol) in 50% aq TFA (30 mL) was stirred at room temperature
for 18 h. The solution was evaporated by co-distillation with toluene to remove the traces of acid
and water. The residue was purified by flash column chromatography (3:1 Et2O/light petroleum) to
afford pure 14 as a colourless glassy solid, Rf = 0.45 (4:1 Et2O/light petroleum). IR (film): νmax
3405, 1605, 1586, 1061, 1028 cm−1. 1H NMR (250 MHz, CDCl3): All signals are heavily
overlapped. 13C NMR (62.5 MHz, CDCl3): δ 70.4, 70.8, 71.8, 72.3, 72.5, 72.6, 73.4, 73.5, 73.8,
73.9, 76.1, 76.7, 77.7, 80.0, 82.2, 83.4, 96.9, 103.4, 127.4, 127.5, 127.6, 127.65, 128.1, 128.2,
128.26, 128.3, 128.6, 137.8, 138.2, 138.5. HRMS (ESI): m/z 451.2112 (M++H), calcd for C24H27O8:
451.2115. To a stirred solution of purified 14 (1.6 g, 3.55 mmol) in EtOAc (20 mL) were
successively added H2O (20 mL) and H5IO6 (1.57 g, 6.89 mmol) and the resulting solution was
stirred at room temperature for 2.5 h. The mixture was poured to 10% aq NaCl (50 mL) and
extracted with EtOAc (3 × 30 mL). The combined organic layers were dried and evaporated and the
residue was dried in high vacuum for 12 h. The resulting crude 15 was dissolved in MeOH (125
mL) and treated with 2-(triphenylphosphoranylidene)-acetic acid ethyl ester (3.6 g, 10.33 mmol).
The mixture was stirred at room temperature for 20 h and then evaporated. The residue was purified
on a column of flash silica (3:2 light petroleum/Et2O) to give pure 16 (1.5 g, 72% from 13) as a
colourless glassy solid, [α]D23 = −10.4 (c 1.0, CHCl3), Rf = 0.33 (3:2 light petroleum/Et2O). IR
(film): νmax 3476, 1717, 1650, 1605, 1586, 1028. 1H NMR (400 MHz, CDCl3): δ 1.23 (t, 3 H, J =
7.2 Hz, CH2CH3), 2.51 (bs, 1 H, OH), 4.19 (q, 2 H, CH2CH3), 3.70–3.93 (m, 4 H, H-7a, H-7b, H-6
and H-5), 4.40, 4.50, 4.59, 4.74 (4 × d, 4 H, Jgem = 11.6 Hz, 2 × CH2Ph), 4.59 (s, 2 H, CH2Ph at C-
7), 5.45 (ddd, J3,4 = 8.8, J2,4 = 0.8, J4,5 = 2.5 Hz, H-4), 6.02 (d, 1 H, J2,3 = 11.8 Hz, H-2), 6.40 (dd, 1
H, J2,3 = 11.8, J3,4 = 9.0 Hz, H-3), 7.24–7.39 (m, 15 H, 3 × Ph). The sample is contaminated with
4% of E-isomer. 13C NMR (100 MHz, CDCl3): 14.1 (CO2CH2CH3), 60.3 (CO2CH2CH3), 70.7 (C-
7), 71.2, 72.3, 73.4 (3 × CH2Ph), 73.6 (C-4), 74.4 (C-5), 77.8 (C-6) 122.8 (C-2), 127.4, 127.5,
127.59, 127.6, 127.7, 127.8, 127.83, 127.9, 128.0, 128.2, 128.26, 128.3, 128.33, 128.4, 137.9,
138.2, 138.5 (Ph), 146.9 (C-3), 165.6 (CO2CH2CH3). HRMS (ESI): m/z 513.2240 (M++Na), calcd
for C30H34NaO6: 513.2253.
4.1.4. Methyl (2Z)-4,6,7-tri-O-benzyl-2,3-dideoxy-D-arabino-hept-2-enoate (16a)
11

A solution of 11 (3 g, 11.53 mmol) in 60% aq AcOH (45 mL) was stirred at room temperature
for 24 h, then evaporated by co-distillation with toluene to remove the traces of acid and water. To a
stirred and cooled (0 °C) solution of 12 in anhydrous DMF (75 mL) was added NaH (1.66 g, 69.16
mmol) and after cessation of intensive H2 evolution, BnBr (4.8 mL, 40.34 mmol) was slowly added
over 10 minutes. The mixture was stirred at 0 °C for 0.5 h, the cooling bath was then removed and
stirring at room temperature was continued for additional 2 h. Absolute MeOH (12 mL) was slowly
added to the reaction solution. After cessation of intensive evolution of hydrogen, the mixture was
evaporated. The residue was partitioned between CH2Cl2 (40 mL) and H2O (250 mL), the organic
layer was separated and extracted with CH2Cl2 (3 × 40 mL). The organic phases were combined,
washed with 10% aq NaCl (3 × 150 mL), dried and evaporated. The residue (crude 13) was treated
with 50% aq TFA (60 mL) while vigorous stirring at room temperature for 18 h. The solution was
evaporated with toluene to completely remove the acid and water and the resulting residue was
dried in high vacuum overnight. To a solution of crude 13 in absolute MeOH (500 mL), was added
NaIO4 (3.68 g, 17.29 mmol) and the mixture was stirred at room temperature for 2 h. The mixture
was cooled to 0 °C and treated with Ph3P=CHCO2Me (7.5 g, 22.46 mmol). After 1 h, the cooling
vessel was removed and the reaction continued at room temperature for 3 h. A new portion of
Ph3P=CHCO2Me (4.05 g, 12.12 mmol) was then added and stirring continued for additional 20 h.
The reaction mixture was evaporated and the residue was purified twice by means of flash column
chromatography (1:1 light petroleum/Et2O). Pure product 16a (4.1 g, 75% from 11) was obtained in
the form of colourless syrup, [α]D = −19.1 (c 1.0, CHCl3). Rf = 0.32 (3:2 light petroleum/Et2O). The
IR, NMR, and HRMS of the thus obtained 16a were fully consistent with the values previously
reported by us [4].
4.1.5. 4,6,7-Tri-O-benzyl-2,3-dideoxy-D-arabino-hept-2-eno-1,5-lactone (5)
Procedure A. A solution of 16 (1.55 g, 3.16 mmol) and TsOH·H2O (0.015 g, 0.08 mmol) in
CH2Cl2 (120 mL) stirred at room temperature for 12 h and then evaporated. The residue was
purified on a column of flash silica (1:1 light petroleum/Et2O) to afford pure 5 (1.25 g, 92%) as a
colourless solid. Crystallization from light petroleum/Et2O gave colourless crystals, mp 64 °C,
[α]D23 = −194.6 (c 0.5, CHCl3), Rf = 0.30 (1:1 light petroleum/Et2O). IR, NMR (1H and 13C) and
HRMS data were in full agreement with previously reported values [4].
Procedure B. Compound 16a was converted to lactone 5 (in a yield of 94%) according to the
reported procedure [4].
12

4.1.6. 2,3-Dideoxy-D-arabino-hept-2-eno-1,5-lactone (9)
A solution of 5 (0.08 g, 0.18 mmol) and FeCl3 (0.058 g, 0.36 mmol) in anhydrous CH2Cl2 (8
mL) was stirred at room temperature for 0.5 h. An additional amount of FeCl3 (0.015 g, 0.09 mmol)
was added, the mixture was stirred at room temperature for 20 h and then evaporated. The residue
was digested with dry CH3CN (2 × 4 mL), organic phase was filtered and evaporated. The residue
was purified on two columns of flash silica (9:1 EtOAc/MeOH) to afford pure 9 (0.029 g, 92%) as a
colourless glassy solid. Recrystallization from 2-propanol gave a white powder, mp 104–106 °C,
[α]D23 = −162.6 (c 0.5, MeOH), Rf = 0.33 (9:1 EtOAc/MeOH). IR (KBr): νmax 3419, 3336, 1719,
1627. 1H NMR (400 MHz, D2O): δ 3.81 (dd, 1 H, J7a,7b = 12.3, J6,7a = 2.7 Hz, H-7a), 3.92 (dd, 1 H,
J6,7b = 5.2, J7a,7b = 12.3 Hz, H-7b), 4.08 (ddd, 1 H, J6,7a=2.7, J6,7b = 5.2, J5,6 = 8.1 Hz, H-6), 6.21 (d,
1 H, J2,3 = 9.7 Hz, H-2), 7.25 (dd, 1 H, J2,3 = 9.7, J3,4 = 6.1 Hz, H-3). 13C NMR (100 MHz, D2O): δ
61.5 (C-4), 62.8 (C-7), 71.2 (C-6), 81.9 (C-5), 124.5 (C-2), 148.8 (C-3), 169.2 (C-1). HRMS (ESI):
m/z 213.0152 (M++K), calcd for C7H10KO5: 213.0160.
4.1.7. 5-O-Acetyl-6-O-benzoyl-3-O-benzyl-1,2-O-isopropylidene-α-D-glucofuranose (20)
A solution of 17 (2.9 g, 8.28 mmol) in 60% aq AcOH (60 mL) was vigorously stirred at room
temperature for 24 h. The mixture was evaporated by co-distillation with toluene to remove traces
of acid and water, and the resulting residue was dried overnight under high vacuum. To a stirred
solution of crude 18 and BzCl (0.96 mL, 8.28 mmol) in anhydrous CH2Cl2 (75 mL) was added
DMAP (3.033 g, 24.83 mmol) and the mixture was stirred at room temperature for 1 h. Acetyl
chloride (0.88 mL, 12.42 mmol) and DMAP (2.6 g, 21.28 mmol) were finally added to the mixture
and the stirring at room temperature was continued for the next 1 h. Reaction mixture was poured in
10% aq NaCl (300 mL) and extracted with CH2Cl2 (3 × 40 mL). Combined organic phases were
washed with 10% aq NaCl (200 mL), dried and evaporated and the resulting residue was purified on
two columns of flash silica (3:7 Et2O/light petroleum). Pure product 20 (2.6 g, 69%) was isolated as
a colourless glassy solid, which crystallizes from Et2O/hexane as colourless needles, mp 96–97 °C,
lit. [12] mp 95–96 °C, [α]D= −64.2 (CHCl3, c 0.5), lit. [12] [α]D= −67.1 (CHCl3, c 1.22), Rf = 0.25
(3:1 light petroleum/Et2O). IR (film): νmax 1748, 1724, 1602, 1585, 1165. 1H NMR (250 MHz,
CDCl3): δ 1.35 and 1.51 (2 × s, 3 H each, Me2C), 1.91 (s, 3 H, MeCO), 3.99 (d, 1 H, J3,4 = 3.1 Hz,
H-3), 4.34–4.52 (m, 3 H, CH2Ph, H-4, H-6a), 4.66 (d, 1 H, J1,2 = 3.7 Hz, H-2), 4.66 (d, 1 H, Jgem
=
11.7 Hz, CH2Ph), 4.92 (dd, 1 H, J5,6b = 2.1, J6a,6b = 12.3 Hz, H-6b), 5.50 (ddd, 1 H, J5,6b = 2.1, J =
5.0, J = 8.5 Hz, H-5), 5.96 (d, 1 H, J1,2 = 3.7 Hz, H-1), 7.28–8.18 (m, 10 H, Ph). 13C NMR (62.5
MHz, CDCl3): δ 20.8 (MeCO), 26.2 and 26.8 (Me2C), 63.7 (C-6), 68.5 (C-5), 71.9 (CH2Ph), 77.8
13

(C-4), 80.4 (C-3), 81.6 (C-2), 105.2 (C-1), 111.9 (Me2C), 128.0, 128.2, 128.3, 128.6, 129.6, 130.0,
132.9, 136.8 (Ph), 166.1 (COPh), 169.5 (MeCO). HRMS (ESI): m/z 495.1410 (M++K), calcd for
C25H28KO8: 495.1416.
4.1.8. 6-O-Acetyl-7-O-benzoyl-4-O-benzyl-2,3-dideoxy-D-arabino-hept-2-eno-1,5-lactone (6)
To a solution of 20 (1.25 g, 2.74 mmol) in CH2Cl2 (12.5 mL) which was vigorously stirred at 0
°C was added 90% aq TFA (25 mL). After 0.5 h the cooling was stopped and the reaction continued
at room temperature for additional 2 h. The solution was evaporated by co-distillation with toluene
to remove the traces of acid and water. A small amount of the residue was purified (for
characterization purposes only) by flash column chromatography (7:3 Et2O/light petroleum).
Colourless solid, mp 130–132 °C (from CH2Cl2/hexane), [α]D = +62.4 (CHCl3, c 0.5), Rf = 0.31 (7:3
Et2O/light petroleum). Anomeric ratio: β/α = 2:1 (from 1H NMR). 1H NMR (400 MHz, CDCl3): δ
1.95 (s, 3 H, MeCO), 3.50 (bs, 2 H, 2 × OH), 3.96–5.52 (overlapped H-2, H-3, H-4, H-5, 2 × H-6
and CH2Ph), 5.23 (s, 1 H, H-1β), 5.54 (d, 1 H, J1,2=4.0 Hz, H-1α), 7.29−8.07 (m, 10 H, Ph). 13C
NMR (100 MHz, CDCl3): δ 21.23 (MeCO), 64.2 (C-6β), 64.3 (C-6α), 69.5, 69.8, 72.3, 73.0, 74.8,
77.0, 77.3, 79.2, 81.6, 82.8, 97.1 (C-1α), 104.0 (C-1β), 128.3, 128.4, 128.7, 128.72, 128.8, 129.0,
130.0, 130.1, 133.4, 136.6, 137.4 (Ph), 166.9 (COPh), 170.3 (MeCO). HRMS (ESI): m/z 434.1802
(M++NH4), calcd for C18H19Cl3NO8: 434.1809. To a cooled (0 °C) solution of crude intermediate 21
in EtOAc (15 mL), were successively added H2O (15 mL) and H5IO6 (1.25 g, 5.49 mmol). After 0.5
h the cooling bath was removed and the reaction continued at room temperature for additional 3.5 h.
The mixture was poured to 10% aq NaCl (40 mL) and extracted with EtOAc (3 × 25 mL). The
combined organic phases were washed with 10% aq NaCl (30 mL), then were dried and evaporated,
and the resulting residue was dried under high vacuum for 2 h. To the solution of crude 22 in dry
MeOH (100 mL) was added ECMP (1.15 g, 3.3 mmol) and the resulting mixture was stirred at
room temperature for 20 h. The solution is evaporated and the resulting residue was purified by
flash chromatography (3:2→1:1 light petroleum/Et2O). An inseparable 1:1 mixture (1H NMR) of
ethyl and methyl esters (23 and 24) was obtained (0.6 g), Rf = 0.47 (1:1 light petroleum/Et2O). The
mixture was used in the next step without separation. A solution of purified mixture of 23 and 24
(0.315 g, 0.69 mmol) and TsOH·H2O (0.006 g, 0.03 mmol) in dry CH2Cl2 (31 mL) was stirred at
room temperature for 20 h. the reaction solution is evaporated, the resulting residue was purified by
flash chromatography (2:3 light petroleum/Et2O). Pure 6 (0.225 g, 79%, 38% from 20) as a syrup
that crystallized from a mixture of Et2O/traces CH2Cl2/hexane as a colourless solid, mp 100 °C, [α]D
= −222.2 (CHCl3, c 0.5), Rf = 0.30 (3:2 light petroleum/Et2O). IR (KBr): νmax 1741, 1728, 1602,
1585, 1177. 1H NMR (400 MHz, CDCl3): δ 2.08 (s, 3 H, MeCO), 4.05 (dd, 1 H, J3,4 = 5.6, J4,5 = 2.7
14

Hz, H-4), 4.51 and 4.61 (2 × d, 2 H, Jgem = 11.8 Hz, CH2Ph), 4.54 (dd, 1 H, J6,7a = 4.2, J7a,7b = 12.6
Hz, H-7a), 4.66 (dd, 1 H, J4,5 = 2.6, J5,6 = 8.8 Hz, H-5), 5.03 (dd, 1 H, J7a,7b = 12.5, J6,7a = 2.3 Hz,
H-7b), 5.62 (ddd, 1 H, J6,7b = 2.3, J6,7a = 4.2, J5,6=8.8 Hz, H-6), 6.24 (d, 1 H, J2,3 = 9.8 Hz, H-2),
7.03 (dd, 1 H, J2,3 = 9.8, J3,4=5.6 Hz, H-3), 7.28–8.07 ( m, 10 H, Ph). 13C NMR (100 MHz, CDCl3):
δ 20.8 (MeCO), 61.8 (C-7), 64.0 (C-4), 68.5 (C-6), 71.2 (CH2Ph), 76.9 (C-5), 124.4 (C-2), 128.2,
128.3, 128.34, 128.6, 129.5, 129.6, 133.1, 136.6 (Ph), 142.4 (C-3), 161.6 (C-1), 166.0 (COPh),
169.3 (MeCO). HRMS (ESI): m/z 449.0990 (M++K), calcd for C23H22KO7: 449.0997.
4.1.9. 6-O-Acetyl-7-O-benzoyl-2,3-dideoxy-D-arabino-hept-2-eno-1,5-lactone (10)
To a stirred solution of 6 (0.32 g, 0.78 mmol) in CH2Cl2 (18 mL) was added 0.66% solution of
TiCl4 (2.61 mL, 1.56 mmol) in dry CH2Cl2. After 24 h, the reaction mixture was poured into 10%
aq NaHCO3 (60 mL) and the resulting suspension extracted with CH2Cl2 (4 × 20 mL). Organic
phases were washed with 10% aq NaCl (50 mL), dried and evaporated, and the resulting residue
was purified by flash chromatography (Et2O). Pure 10 (0.176 g, 70%) was obtained, as a white
solid, mp 118–120 °C. Recrystallization from CH2Cl2/hexane yielded a white powder, mp 124 °C,
[α]D23 = −32.0 (c 0.5, CHCl3), lit. [6] [α]D = −29.5 (c 0.45, CHCl3), Rf = 0.30 (Et2O). IR (KBr): νmax
3420, 1743. 1H NMR (400 MHz, CDCl3): δ 2.12, (s, 3 H, MeCO), 3.69 (br s, 1 H, OH), 4.25 (t, 1 H,
J = 5.0 Hz, H-4), 4.52–4.63 (m, 2 H, H-5 and H-7a), 4.95 (dd, 1 H, J7a,7b = 12.5, J6,7b=2.0 Hz, H-
7b), 5.52 (ddd, 1 H, J6,7b = 2.2, J6,7a = 4.9, J5,6 = 9.5 Hz, H-6), 6.15 (d, 1 H, J2,3 = 9.7 Hz, H-2), 7.04
(dd, 1 H, J2,3 = 9.7, J3,4=3.6 Hz, H-3), 7.34–8.08 ( m, 5 H, Ph). 13C NMR (100 MHz, CDCl3): δ 20.9
(MeCO), 59.3 (C-4), 62.5 (C-7), 69.1 (C-6), 77.6 (C-5), 123.2 (C-2), 128.5, 129.6, 129.7, 133.3
(Ph), 143.7 (C-3), 162.4 (C-1), 166.3 (COPh), 170.8 (MeCO). HRMS (ESI): m/z 343.0787
(M++Na), calcd for C16H16NaO7: 343.0788.
4.1.10. (−)-Cleistenolide (1)
Procedure A. To a stirred suspension of 10 (0.176 g, 0.55 mmol) in Ac2O (5 mL, 10.6 mmol)
was added TsOH·H2O (0.005 g, 0.026 mmol). The mixture was stirred at room temperature for 4 h
and then evaporated by co-distillation with toluene to remove the traces of Ac2O. The residue was
purified by flash column chromatography on silica gel (1:2 light petroleum/Et2O) to yield pure 1
(0.19 g, 95%) in the form of a white solid, mp 135 °C, lit. [8] mp 133–134 °C, [α]D = −212.2 (c 0.5,
CHCl3), or [α]D = −212.4 (c 0.5, CH2Cl2), lit. [8] [α]D = −164.6 (c 0.48, CH2Cl2). Recrystallization
from Et2O/hexane gave white needles, mp 138 °C, Rf = 0.36 (3:7 light petroleum/Et2O).
15

Procedure B. To a stirred solution of 5 (0.05 g, 0.11 mmol) and BzBr (0.017 mL, 0.14 mmol)
in dry CH2Cl2 (5 mL) was added FeCl3 (0.005 g, 0.03 mmol) and the mixture was stirred at room
temperature for 2 h. After that, to the mixture was added AcBr (0.16 mL, 2.2 mmol) and the
reaction was continued at room temperature for additional 20 h. Reaction mixture was diluted with
CH2Cl2 (10 mL), poured in 10% aq NaHCO3 (60 mL) and extracted with CH2Cl2 (2 × 15 mL). The
organic phases were combined, washed with 10% aq NaCl (30 mL), dried and evaporated, and the
resulting residue was purified by flash column chromatography (2:1 light petroleum/Et2O) to afford
pure 1 (0.030 g, 74%) as a colourless solid, mp 135 °C, lit. [8] mp 133–134 °C. Recrystallization
from Et2O/hexane gave colourless needles, mp 138 °C, [α]D = −212.2 (c 0.5, CHCl3), lit. [8] [α]D =
−164.6 (c 0.5, CH2Cl2), Rf = 0.36 (7:3 Et2O/light petroleum). IR (KBr): νmax 1752, 1732, 1601,
1183. 1H NMR (400 MHz, CDCl3): δ 2.06 and 2.11 (2 × s, 3 H each, 2 × MeCO), 4.54 (dd, 1 H,
J6,7a = 4.5, J7a,7b = 12.5 Hz, H-7a), 4.82 (dd, 1 H, J5,6 = 9.6, J4,5 = 2.7 Hz, H-5), 4.94 (dd, 1 H, J6,7b
2.4, J7a,7b = 12.5 Hz, H-7b), 5.44 (dd, 1 H, J3,4 = 6.1, J4,5 = 2.7 Hz, H-4), 5.53 (ddd, 1 H, J5,6 = 9.6,
J6,7a = 4.5, J6,7b = 2.4 Hz, H-6), 6.30 (d, 1 H, J2,3 = 9.7 Hz, H-2), 7.01 (dd, 1 H, J2,3 = 9.7, J3,4 = 6.1
=
Hz, H-3), 7.42–8.08 (m, 5 H, Ph). 13C NMR (100 MHz, CDCl3): δ 20.5 and 20.7 (2 × MeCO), 59.8
(C-4), 62.0 (C-7), 67.7 (C-6), 75.5 (C-5), 125.4 (C-2), 128.5, 129.6, 129.7, 133.3 (Ph), 139.7 (C-3),
161.1 (C-1), 166.0 (COPh), 169.5 and 169.9 (2 × MeCO). HRMS (ESI): m/z 385.0897 (M++Na),
calcd for C18H18NaO8: 385.0894.
4.1.11. 4,6-Di-O-acetyl-7-O-(3-methoxybenzoyl)-2,3-dideoxy-D-arabino-hept-2-eno-1,5-lactone (2)
To a stirred solution of 9 (0.1 g, 0.57 mmol) in a mixture of anhydrous Py (2 mL) and CH2Cl2
(2 mL) was first added 3-methoxybenzoyl chloride (0.08 mL, 0.58 mmol) and the mixture was
stirred at room temperature for 4 h. Acetic anhydride (4 mL) was added to the solution and the
stirring at room temperature was continued for the next 20 h. The mixture was evaporated by co-
distillation with toluene to remove traces of Py and Ac2O. The residue was suspended in water (70
mL) and extracted with CH2Cl2 (3 × 20 mL). The combined organic solutions were dried and
evaporated, and the remaining crude mixture was purified by flash column chromatography (9:1
toluene/Me2CO) to give pure 2 (0.152 g, 67%) as a colourless glassy solid. Recrystallization from
2-propanol gave a white powder, mp 85 °C, [α]D23 = −187.8 (c 0.5, CHCl3), Rf = 0.30 (9:1
toluene/Me2CO). IR (KBr): νmax 1750, 1734, 1636, 1604, 1588, 1182, 1153. 1H NMR (400 MHz,
CDCl3): δ 2.05 and 2.10 (2 × s, 3 H each, 2 × MeCO), 3.86 (s, 3 H, CH3O), 4.52 (dd, 1 H, J7a,7b
=
12.5, J6,7a = 4.6 Hz, H-7a), 4.80 (dd, 1 H, J4,5 = 2.5, J5,6 = 9.6 Hz, H-5), 4.92 (dd, 1 H, J7a,7b = 12.5,
J6,7b = 2.1 Hz, H-7b), 5.43 (dd, 1 H, J4,5 = 2.6, J3,4 = 6.0 Hz, H-4), 5.52 (ddd, 1 H, J6,7b = 2.1, J6,7a
=
4.6, J5,6 = 9.6 Hz, H-6), 6.29 (d, 1 H, J2,3 = 9.7 Hz, H-2), 7.00 (dd, 1 H, J2,3 = 9.7, J3,4 = 6.1 Hz, H-
16

3), 7.09–7.64 (m, 4 H, MeOPh). 13C NMR (100 MHz, CDCl3): δ 20.5 and 20.7 (2 × MeCO), 55.5
(OMe), 59.7 (C-4), 62.2 (C-7), 67.6 (C-6), 75.5 (C-5), 125.3 (C-2), 114.4, 119.6, 122.0, 129.5,
130.9, 159.6 (MeOPh), 139.7 (C-3), 161.1 (C-1), 165.9 (ArCO), 169.5, 169.9 (2 × MeCO). HRMS
(ESI): m/z 431.0732 (M++K), calcd for C19H20KO9: 431.0739.
4.1.12. 4,6-Di-O-acetyl-7-O-(2,4,6-trichlorobenzoyl)-2,3-dideoxy-D-arabino-hept-2-eno-1,5-lactone
(3)
To a stirred solution of 9 (0.05 g, 0.29 mmol) in a mixture of anhydrous Py (1 mL) and CH2Cl2
(1 mL) was first added 2,4,6-trichlorobenzoyl chloride (0.04 mL, 0.26 mmol) and the mixture was
stirred at room temperature for 4 h. Acetic anhydride (2 mL) was added to the solution and the
stirring at room temperature was continued for the next 20 h. The mixture was evaporated by co-
distillation with toluene and the residue was purified on a column of flash silica (1:1 Et2O/light
petroleum) to give pure 3 (0.114 g, 85%) as a colourless solid. Recrystallization from
CH2Cl2/hexane gave a white powder, mp 134 °C, [α]D23 = −193.6 (CHCl3, c 0.5), Rf = 0.30 (1:1
Et2O/light petroleum). IR (KBr): νmax 1747, 1631, 1580, 1548, 1193, 1161. 1H NMR (400 MHz,
CDCl3): δ 2.07 and 2.09 (2 × s, 3 H each, 2 × MeCO), 4.61 (dd, 1 H, J7a,7b = 12.5, J6,7a = 2.5 Hz, H-
7a), 4.80 (dd, 1 H, J4,5 = 1.7, J5,6 = 9.5 Hz, H-5), 5.00 (bd, 1 H, J7a,7b = 12.5 Hz, H-7b), 5.39 (dd, 1
H, J4,5 = 1.8, J3,4 = 6.0 Hz, H-4), 5.44 (bd, 1 H, J5,6 = 9.5 Hz, H-6), 6.27 (d, 1 H, J2,3 = 9.7 Hz, H-2),
6.99 (dd, 1 H, J2,3 = 9.6, J3,4 = 6.1 Hz, H-3), 8.36 (s, 2 H, 2,4,6-trichlorophenyl). 13C NMR (100
MHz, CDCl3): δ 20.4 and 20.7 (2 × MeCO), 59.7 (C-4), 62.8 (C-7), 67.4 (C-6), 74.9 (C-5), 125.3
(C-2), 128.1, 131.6, 132.6, 136.5 (2,4,6-trichlorophenyl), 139.7 (C-3), 161.0 (C-1), 163.5 (ArCO),
169.4 and 169.9 (2 × MeCO). HRMS (ESI): m/z 482.0169 (M++NH4), calcd for C18H19Cl3NO8:
482.0171.
4.1.13. 4,6-Di-O-acetyl-7-O-cinnamoyl-2,3-dideoxy-D-arabino-hept-2-eno-1,5-lactone (4)
To a stirred solution of 9 (0.08 g, 0.46 mmol) in a mixture of anhydrous Py (2.5 mL) and
CH2Cl2 (2.5 mL) was first added trans-cinnamoyl chloride (0.095 g, 0.57 mmol) and the mixture
was stirred at room temperature for 2 h. Acetic anhydride (5 mL) was added to the mixture and the
stirring at room temperature was continued for the next 18 h. The solution was evaporated by co-
distillation with toluene. The residue was suspended in H2O (80 mL) and extracted with CH2Cl2 (3
× 20 mL). The combined organic phases were washed with 10% aq NaCl (50 mL) dried and
evaporated. Oily residue was purified by flash column chromatography (3:2 Et2O/light petroleum)
to afford pure 4 (0.084 g, 47%) as a colourless solid. Recrystallization from Et2O/light petroleum
17

gave a white powder, mp 110 °C, [α]D23 = −212.8 (CHCl3, c 0.5), Rf = 0.34 (3:2 Et2O/light
petroleum). IR (KBr): νmax 1750, 1736, 1642, 1578, 1180. 1H NMR (400 MHz, CDCl3): δ 2.08 and
2.10 (2 × s, 3 H each, 2 × MeCO), 4.44 (dd, 1 H, J7a,7b = 12.6, J6,7a = 3.9 Hz, H-7a), 4.78–4.87 (m, 2
H, H-5 and H-7b), 5.39–5.49 (m, 2 H, H-4 and H-6), 6.30 (d, 1 H, J2,3 = 9.7 Hz, H-2), 6.46 (d, 1 H,
J2′,3′ = 16.0 Hz, H-2′), 7.02 (dd, 1 H, J2,3 = 9.6, J3,4 = 6.1 Hz, H-3), 7.35–7.61 (m, 5 H, Ph), 7.73 (d,
1 H, J2′,3′ = 16.0 Hz, H-3′). 13C NMR (100 MHz, CDCl3): δ 20.5 and 20.7 (2 × MeCO), 59.8 (C-4),
61.5 (C-7), 67.7 (C-6), 75.3 (C-5), 117.2 (C-2′), 125.4 (C-2), 128.2, 128.9, 130.6, 134.2 (Ph), 139.7
(C-3), 145.8 (C-3′), 161.2 (C-1), 166.1 (C-1′), 169.5 and 169.9 (2 × MeCO). HRMS (ESI): m/z
411.1045 (M++Na), calcd for C20H20NaO8: 411.1050.
4.1.14. 4,6,7-Tri-O-benzoyl-2,3-dideoxy-D-arabino-hept-2-eno-1,5-lactone (7)
To a stirred solution of 9 (0.04 g, 0.23 mmol) in a mixture of anhydrous Py (1.5 mL) and
CH2Cl2 (1.5 mL) was added slowly benzoyl chloride (0.16 mL, 1.38 mmol) in portions. The
mixture was stirred at room temperature for 3.5 h, and the mixture was evaporated by co-distillation
with toluene to remove traces of Py. The residue was dissolved in CH2Cl2, absorbed on flash silica
and purified by flash column chromatography (11:9 Et2O/light petroleum). Pure product 7 (0.096 g,
86%) was obtained as a turbid oil, which crystallizes from CH2Cl2/hexane as a white powder, mp
120 °C, [α]D23 = −276.8 (CHCl3, c 1), Rf = 0.35 (3:7 EtOAc/light petroleum), or Rf = 0.39 (3:2
Et2O/light petroleum). IR (film): νmax 1726, 1601, 1585, 1177. 1H NMR (400 MHz, CDCl3): δ 4.78
(dd, 1 H, J6,7a = 4.6, J7a,7b = 12.5 Hz, H-7a), 5.08 (dd, 1 H, J6,7b = 2.2, J7a,7b = 12.6 Hz, H-7b), 5.12
(dd, J4,5 = 2.4, J5,6 = 9.2 Hz, H-5), 5.66 (dd, 1 H, J3,4 = 6.0, J4,5 = 2.5 Hz, H-4), 5.93 (m, 1 H, H-6),
6.35 (d, 1 H, J2,3 = 9.7 Hz, H-2), 7.23 (dd, 1 H, J2,3 = 9.7, J3,4 = 6.0 Hz, H-3), 7.35–8.04 (m, 15 H,
Ph). 13C NMR (100 MHz, CDCl3): 61.0 (C-4), 62.2 ( C-7), 68.5 (C-6), 75.9 (C-5), 125.3 (C-2),
128.36, 128.4, 128.5, 128.8, 129.5, 129.6, 129.8, 133.2, 133.5, 133.8 (Ph), 140.2 (C-3), 161.2 (C-1),
164.8, 165.2, 165.9 (3 × PhC=O). HRMS (ESI): m/z 525.0948 (M++K), calcd for C28H22KO8:
525.0946.
4.1.15. 4,6,7-Tri-O-(3-methoxybenzoyl)-2,3-dideoxy-D-arabino-hept-2-eno-1,5-lactone (8)
To a stirred solution of 9 (0.04 g, 0.23 mmol) in a mixture of anhydrous Py (1.5 mL) and
CH2Cl2 (1.5 mL) was slowly added 3-methoxybenzoyl chloride (0.2 mL, 1.38 mmol). The mixture
was stirred at room temperature for 3.5 h, and the mixture was evaporated by co-distillation with
toluene to remove traces of Py. The residue was purified on two columns of flash silica (2:1
Et2O/light petroleum). Pure product 8 (0.117 g, 88%) was obtained in the form of turbid syrup,
18

[α]D23 = −242.4 (CHCl3, c 1), Rf = 0.33 (17:3 toluene/EtOAc) or Rf = 0.31 (7:3 Et2O/light
petroleum). IR (film): νmax 1731, 1601, 1489, 1182. 1H NMR (400 MHz, CDCl3): δ 4.77 (dd, 1 H,
J6,7a = 4.5, J7a,7b = 12.7 Hz, H-7a), 5.03 (dd, 1 H, J6,7b = 1.7, J7a,7b = 12.7 Hz, H-7b), 3.78, 3.81 and
3.86 (3 × s, 9 H, 3 × OMe), 5.14 (dd, J4,5 = 2.4, J5,6 = 9.1 Hz, H-5), 5.63 (dd, 1 H, J3,4 = 6.0, J4,5
2.2 Hz, H-4), 5.91 (m, 1 H, H-6), 6.35 (d, 1 H, J2,3 = 9.5 Hz, H-2), 7.04–8.04 (m, 13 H, H-3 and
=
Ph). 13C NMR (100 MHz, CDCl3): 55.38, 55.4 and 55.5 (3 × OMe), 61.2 (C-4), 62.4 (C-7), 68.6 (C-
6), 75.9 (C-5), 125.3 (C-2), 114.2, 114.2 and 114.25 (3 × C-2′), 119.7, 120.0, 120.4 (3 × C-4′),
122.0, 122.3 (C-6′), 129.5, 129.6 (C-5′), 140.3 (C-3), 159.5, 159.57, 159.6 (C-3′) 161.2 (C-1),
165.8, 165.2, 164.7 (3 × C=O). HRMS (ESI): m/z 615.1268 (M++K), calcd for C31H28KO11:
615.1263.
4.2. Antimicrobial activity
4.2.1. Test microorganisms
The compounds were tested against the panel of microbial strains belonging to the American
Type Culture Collection reference strains, including the following: Gram-positive bacteria
(Staphylococcus aureus ATCC 6538, Staphylococcus epidermidis ATCC 12228, Bacillus cereus
ATCC 11778, and Kocuria rhizophila ATCC 9431), Gram-negative bacteria (Pseudomonas
aeruginosa ATCC 9027, Escherichia coli ATCC 8739, Salmonella enteritidis ATCC 13076, and
Acinetobacter baumannii ATCC 19606) and yeast Candida albicans (ATCC 10231). Bacterial
strains were maintained on Nutrient Agar (NA) at 37 °C, while the yeast strain was grown on
Sabouraud Dextrose Agar (SDA) at 30 °C at the Microbiology Laboratory (Department of Biology,
Faculty of Sciences and Mathematics, University of Niš).
4.2.2. Screening of antimicrobial activity (microdilution method)
Antimicrobial activity was evaluated using a broth microdilution method in microtiter plates as
described earlier [23]. Briefly, cell suspensions standardized to 0.5 McFarland (DEN-1, Biosan)
turbidity were made using an overnight culture (18 h) of the test microorganisms. Stock solutions of
the test compounds were made in pure DMSO and diluted with appropriate sterile broth (Sabouraud
Dextrose or Mueller Hinton broth). The highest final concentration (10%, v/v) of DMSO needed to
achieve complete dissolution of the compounds was confirmed not to affect either bacterial or
fungal growth. After making doubling dilutions of the test substances in the concentration range
0.002–6.00 µmol/ml, the inoculum was added to all wells and the plates were incubated at 37 °C
during 24 h in the case of bacteria or at 30 °C during 48 h for the yeast. Chloramphenicol and
19

nystatin served as positive controls. The bacterial growth was determined by adding 20 μl of 0.5%
(w/w) triphenyltetrazolium chloride (TTC) aqueous solution. MIC was defined as the lowest
concentration of the test compound that inhibited visible growth (red-coloured pellet on the bottom
of the wells after the addition of TTC). All experiments were done in triplicate and repeated two
times.
4.3. Statistical treatment
Hierarchical clustering and heatmap preparation were performed using the freeware Morpheus
[24]. The proximity between two objects in the dendrograms was measured by the Euclidean
distance. The average-linkage method was applied as the aggregation criterion. The number of
object classes (groups of observations) was chosen based on the increase of within-group and
between-group dissimilarities. The method was applied utilizing original variables (MIC values).
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.
Acknowledgement
The authors acknowledge financial support of the Ministry of Education, Science and
Technological Development of the Republic of Serbia (Grants No. 172006, contract No. 451-03-
68/2020-14/200125; and 172061, contract No. 451-03-68/2020-14/200124). This work has received
funding from the Serbian Academy of Sciences and Arts under strategic projects programme (grant
agreement No 01-2019-F65), as well as by a research project (No. F-130) from the same institution.
Appendix A. Supplementary material
Supplementary data associated with this article can be found in online version.
References and notes
[1] S. Samwel, S. J. M. Mdachi, M. H. H. Nkunya, B. N. Irungu, M. J. Moshi, B. Moulton, B. S.
Luisi, Cleistenolide and cleistodienol: Novel bioactive constituents of Cleistochlamys kirkii, Nat.
Prod. Commun. 2 (2007) 737–741.
20

[2] F. Pereira, A. M. Madureira, S. Sancha, S. Mulhovo, X. Luo, A. Duarte, M.-J. U. Ferreira,
Cleistochlamys kirkii chemical constituents: Antibacterial activity and synergistic effects against
resistant Staphylococcus aureus strains, J. Ethnopharmacol. 178 (2016) 180–187.
[3] P. Karier, G. C. Catrinescu, N. Diercxsens, K. Robeyns, M. L. Singleton, I. E. Markó, Total
synthesis of (−)-cleistenolide and formal synthesis of herbarumin I via a diastereoselective
modulable allylation, Tetrahedron 74 (2018) 7242–7251.
[4] T. V. Kumar, K. S. Babu, J. M. Rao, A simple and efficient stereoselective synthesis of (−)-
cleistenolide, Tetrahedron Lett. 53 (2012) 1823–1825.
[5] D. C. Babu, K. Ashalatha, C. B. Rao, J. P. S. Jondoss, Y. Venkateswarlu, Total synthesis of (−)-
cleistenolide, Helv. Chim. Acta 94 (2011) 2215–2220.
[6] D. C. Babu, J. J. P. Selavam, D. K. Reddy, V. Shekhar, Y. Venkateswarlu, Total synthesis of
(−)-cleistenolide, Tetrahedron 67 (2011) 3815–3819.
[7] P. Ramesh, H. M. Meshram, Total synthesis of (−)-cleistenolide, Tetrahedron Lett. 52 (2011)
2443–2445.
[8] B. Schmidt, O. Kunz, A. Biernat, Total synthesis of (−)-cleistenolide, J. Org. Chem. 75 (2010)
2389–2394.
[9] A. B. Reddy, B. K. Swamy, J. S. Yadav, A concise total synthesis of cleistenolide, Tetrahedron:
Asymmetry 27 (2016) 788–790.
[10] R. S. Ghogare, S. B. Wadavrao, A. V. Narsaiah, Enantioselective construction of 6-substituted-
α,β-unsaturated-δ-lactone: total synthesis of anti-bacterial agent (−)-cleistenolide, Tetrahedron Lett.
54 (2013) 5674–5676.
[11] B. V. S. Reddy, B. P. Reddy, T. Pandurangam, J. S. Yadav, The stereoselective total synthesis
of (−)-cleistenolide, Tetrahedron Lett. 52 (2011) 2306–2308.
[12] C. Cai, J. Liu, Y. Du, R. J. Linhardt, Stereoselective total synthesis of (−)-cleistenolide, J. Org.
Chem. 75 (2011) 5754–5756.
[13] G. Benedeković, I. Kovačević, M. Popsavin, J. Francuz, V. Kojić, G. Bogdanović, V.
Popsavin, New antitumour agents with α,β-unsaturated δ-lactone scaffold: Synthesis and
antiproliferative activity of (−)-cleistenolide and analogues, Bioorg. Med. Chem. Lett. 26 (2016)
3318–3321.
21

[14] E. Pontiki, A. Peperidou, I. Fotopoulos, D. Hadjipavlou-Litina, Cinnamate hybrids: A unique
family of compounds with multiple biological activities, Curr. Pharm. Biotechnol. 19 (2018) 1019–
1048.
[15] M. Sova, Antioxidant and antimicrobial activities of cinnamic acid derivatives, Mini-Rev.
Med. Chem. 12 (2012) 749–767.
[16] V. W. C. Soo, B. W. Kwan, H. Quezada, I. Castillo-Juárez, B. Pérez-Eretza, S. J. García-
Contreras, M. Martínez-Vázquez, T. K. Wood, R. García-Contreras, Repurposing of anticancer
drugs for the treatment of bacterial infections, Curr. Topics Med. Chem. 17 (2017) 1157–1176.
[17] Y. Du, F. Kong, Synthesis and glycosidic reaction of 1,2-anhydromanno-, lyxo-, gluco-, and
xylofuranose perbenzyl ethers, J. Carbohydr. Chem. 15 (1996) 797–817.
[18] H. Hori, Y. Nishida, H. Ohrui, H. Meguro, Regioselective de-O-benzylation with Lewis acids,
J. Org. Chem. 54 (1989) 1346–1353.
[19] J. E. Harvey, S. A. Raw, R. J. K. Taylor, A versatile and stereocontrolled route to pyranose and
furanose C-glycosides, Org. Lett. 6 (2004) 2611– 2614.
[20] S. Valverde, M. Martin-Lomas, B. Herradon, S. Garcia-Ochoa, The reaction of carbohydrate-
derived alkoxyaldehydes with methoxycarbonylmethylenetriphenylphosphorane: stereoselective
synthesis of β-unsaturated esters, Tetrahedron 43 (1987) 1895–1901.
[21] For an efficient procedure for the preparation of 17 see the Supplementary data.
[22] A. S. Meyer, T. Reichstein, L-Idose aus D-glucose, sowie ein neuer weg zur L-idomethylose,
Helv. Chim. Acta. 29 (1946) 152–162.
[23] N. Radulović, Z. Stojanović-Radić, P. Stojanović, N. Stojanović, V. Dekić, B. Dekić, A small
library of 4-(alkylamino)-3-nitrocoumarin derivatives with potent antimicrobial activity against
gastrointestinal pathogens, J. Serb. Chem. Soc. 80 (2015) 315–327.
[24] Morpheus, https://software.broadinstitute.org/morpheus
22

TITLE:
Synthesis and antimicrobial activity of (−)-cleistenolide and analogues
AUTHORS:
Goran Benedeković, Mirjana Popsavin, Niko S. Radulović, Zorica Stojanović-Radić, Sándor
Farkas, Jovana Francuz, Velimir Popsavin
Highlights



New synthesis of (−)-cleistenolide was achieved using the "chiral pool" approach
The same approach was used to prepare several novel (−)-cleistenolide analogues
Most analogues showed antimicrobial activity against some bacterial strains
23