Semisynthetic latrunculin derivatives as inhibitors of metastatic breast cancer: Biological evaluations, preliminary structure-activity relationship and molecular modeling studies

Article abstract of DOI:10.1002/cmdc.200900430

The microfilament cytoskeleton protein actin plays an important role in cell biology and affects cytokinesis, morphogenesis, and cell migration. These functions usually fail and become abnormal in cancer cells. The marine-derived macrolides latrunculins A and B, from the Red Sea sponge Negombata magnifica, are known to reversibly bind actin monomers, forming 1:1 stoichiometric complexes with G-actin, disrupting its polymerization. To identify novel therapeutic agents for effective treatment of metastatic breast cancer, several semisynthetic derivatives of latrunculin A with diverse steric, electrostatic, and hydrogen bond donor and acceptor properties were rationally prepared. Analogues were designed to modulate the binding affinity toward G-actin. Examples of these reactions are esterification, acetylation, and N-alkylation. Semisynthetic latrunculins were then tested for their ability to inhibit pyrene-conjugated actin polymerization, and subsequently assayed for their antiproliferative and anti-invasive properties against MCF7 and MDA-MB-231 cells using MTT and invasion assays, respectively.

Full text of DOI:10.1002/cmdc.200900430

MED
DOI: 10.1002/cmdc.200900430
Semisynthetic Latrunculin Derivatives as Inhibitors of
Metastatic Breast Cancer: Biological Evaluations,
Preliminary Structure–Activity Relationship and Molecular
Modeling Studies
Mohammad A. Khanfar,[a] Diaa T. A. Youssef,[b] and Khalid A. El Sayed*[a]
The microfilament cytoskeleton protein actin plays an impor-
tant role in cell biology and affects cytokinesis, morphogenesis,
and cell migration. These functions usually fail and become ab-
normal in cancer cells. The marine-derived macrolides latruncu-
lins A and B, from the Red Sea sponge Negombata magnifica,
are known to reversibly bind actin monomers, forming 1:1 stoi-
chiometric complexes with G-actin, disrupting its polymeri-
zation. To identify novel therapeutic agents for effective treat-
ment of metastatic breast cancer, several semisynthetic deriva-
tives of latrunculin A with diverse steric, electrostatic, and hy-
drogen bond donor and acceptor properties were rationally
prepared. Analogues were designed to modulate the binding
affinity toward G-actin. Examples of these reactions are esterifi-
cation, acetylation, and N-alkylation. Semisynthetic latrunculins
were then tested for their ability to inhibit pyrene-conjugated
actin polymerization, and subsequently assayed for their anti-
proliferative and anti-invasive properties against MCF7 and
MDA-MB-231 cells using MTT and invasion assays, respectively.
Introduction
Actin is a cytoskeleton protein that forms versatile dynamic
polymers, which can define cell polarity, organize cytoplasmic
organelles, control cell shape, promote stable cell–cell and
cell–matrix adhesions, and generate protrusive forces required
for migration.[1] These functions usually fail and become abnor-
mal in cancer cells.[1]
understanding of the structure–activity relationship (SAR) of
these macrolides is fairly limited. Previous attempts at SAR
studies were focused on the macrocyclic ring and semiquanti-
tative testing.[15,16] An X-ray crystallographic structure of latrun-
culin A bound to actin monomers reveals that both the thiazo-
lidinone and THP moieties are the key pharmacophores that
direct the main binding and orientation of latrunculin A inside
the ATP binding cleft.[12,13] Meanwhile, the macrocyclic ring is
involved in hydrophobic interactions, and the hydrocarbon
chain from C5 to C7 is exposed to the solvent, forming few
contacts with actin and therefore playing a limited role in
binding.[12] The detailed binding mode of 1 to G-actin was
identified as having the following hydrogen bonding interac-
tions: C1 carbonyl oxygen atom via water to Glu214 carboxyl
group, C17 lactol hydroxy group to Arg210 NH (major bind-
ing), C17 pyran oxygen atom to Tyr69 hydroxy group, thiazoli-
dinone NH to Asp157 carboxyl group, and thiazolidinone C20
carbonyl oxygen atom to Thr186 hydroxy group.[12,13] The only
hydrogen bond donor (HBD) group was found to be the thia-
zolidinone NH, whereas the rest of binding functions act as hy-
drogen bond acceptors (HBA).[12,13] Therefore, efforts focused
The marine-derived macrolides latrunculins A and B were
first isolated by Kashman and co-workers from the Red Sea
sponge Negombata magnifica.[2] Latrunculin A (1) was the first
marine macrolide known to contain a 16-membered ring and
the unique 2-thiazolidinone moiety connected by a tetrahydro-
pyran (THP) ring.[2,3] Latrunculins A and B and their derivatives
show antiangiogenic, antiproliferative, antimicrobial, and anti-
metastatic activities.[4,5] The most important biological effect of
latrunculins is their ability to disrupt microfilament organiza-
tion and to inhibit microfilament-mediated processes without
affecting the organization of the microtubule system.[5–13] La-
trunculin A reversibly binds cytoskeletal actin monomers, form-
ing 1:1 stoichiometric complexes with G-actin, thereby disrupt-
ing its polymerization.[5–13] It has striking selectivity, rapid onset
of action, and remarkable potency that exceed those of cyto-
chalasin D by 1–2-fold.[9] Latrunculin A was also reported to de-
crease intraocular pressure and increase outflow facility with-
out corneal effects in monkeys.[7,8] Latrunculin A also shows an-
tiviral and antibacterial activities, inhibits the stress-activated
MAP kinase (SAPK) pathway,[11] and suppresses hypoxia-in-
duced factor (HIF-1) activation in breast cancer cells.[14] There-
fore, latrunculin A is considered a potential lead, appropriate
for further optimization studies.
[a] M. A. Khanfar, Prof. K. A. El Sayed
Department of Basic Pharmaceutical Sciences, College of Pharmacy
University of Louisiana at Monroe, Monroe, LA 71201 (USA)
Fax: (+1)318-342-1737
E-mail: elsayed@ulm.edu
[b] Prof. D. T. A. Youssef
Department of Pharmacognosy, Faculty of Pharmacy
Suez Canal University, Ismailia 41522 (Egypt)
Despite the remarkable physiological properties of latruncu-
lins and their widespread use as biochemical tools, the present
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.200900430.
274
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 274 – 285

Latrunculin Derivatives
on the optimization of the lactol hydroxy group and thiazolidi-
none NH of latrunculins should reflect better SAR insight.
Several C17 lactol hydroxy- and/or thiazolidinone NH-substi-
tuted latrunculin A derivatives 2–14 were produced by semi-
synthetic procedures (Scheme 1 and Table 1). Various aliphatic
and aromatic substituents were used. Steric limitation, electro-
static, HBD and HBA properties were varied at these positions,
and the biological activities of the resulting derivatives were
tested. Because actin is essential for cellular proliferation and
migration, these activities were quantified by cell proliferation
(MTT) and cell invasion assays using MCF7 and MDA-MB-231
breast cancer cell lines, respectively.
Table 1. Structure of latrunculin A derivatives.
Compound
R1
R2
1
2
3
H
CH3
a
H
H
H
4
5
6
7
8
9
10
11
12
13
14
b
H
CH3
CH2CH3
c
CH2CH2CH2OH
d
b
e
Results and Discussion
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
Chemistry
Latrunculin A (1) was isolated from the Red Sea marine sponge
Negombata magnifica and identified by detailed 1D and 2D
NMR studies along with comparison with published data.[2] Re-
action of 1 with methanol or phenylethanol in the presence of
boron trifluoride diethyl etherate yielded the respective known
17-methoxy (compound 2)[17] and new phenylethoxy (com-
CH2CH3
d
CH2OH
H
H
pound 3) derivatives of
1
(Scheme 1). The configuration
at C17 was maintained after
acetalization of the C17 hydroxy
group. This was confirmed by
molecular modeling and NMR
spectroscopy. The distance be-
tween proton H-18 and the C17
OCH3 group in 2 was calculated
by SYBYL to be 2.403 ꢁ, which
is in the range of NOE coupling.
Proton H-18 and the C17 OCH3
group showed a strong dipole–
dipole coupling in NOESY ex-
periments, confirming a similar
b orientation.
The
13C NMR
chemical shift of C17 in 2 (dC:
99.9 ppm) was almost identical
to the reported value of the
same carbon atom in 17b-me-
thoxylatrunculin
A
(dC: 99.8
ppm).[17] Optical rotation values
for compound 2 and 17b-me-
thoxylatrunculin A were nearly
identical.[17]
Therefore,
the
b configuration at C17 was
maintained in these semisyn-
thetic latrunculin A derivatives.
High-resolution mass spec-
trometry (HRMS), 1H NMR, and
13C NMR data for compound 3
indicated the presence of the
17-O-phenylethyl side chain.
The downfield shifts of the C17
Scheme 1. Semisynthetic transformations of latrunculin A (1). Reagents and conditions: a) ROH, Et2O·BF3, room
temperature; b) 1. NaH, 08C, THF, 1 h, 2. RX, room temperature; c) AcOH/H2O/THF (3:1:1), 608C; d) benzoic anhy-
dride, DMAP, CHCl3 (anhyd), room temperature, 12 h; e) CH2O(aq) (35%), EtOH, 24 h, 608C.
ChemMedChem 2010, 5, 274 – 285
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
275

K. A. El Sayed et al.
MED
and H-18 signals in 3 (d: +3.2 and 0.20 ppm, respectively) rela-
tive to that of the starting material 1 suggested possible ether-
ification at C17.[2] The doublet of triplet of oxygenated methyl-
ene H2-1’ (dH: 3.81, 3.51 ppm) showed COSY coupling with the
High-resolution electrospray ionization mass spectrometry
(HRESIMS) data for compound 9 suggest the molecular formu-
la C30H39NO5S, 12 degrees of unsaturation, and the N-benzyla-
tion of 2. The downfield benzylic nitrogenate methylene pro-
3
3
benzylic H2-2’ triplet (dH: 2.89 ppm). Protons H2-1’ showed J
tons H2-1’ (dH: 5.11 and 4.35 ppm) showed a J HMBC correla-
HMBC correlations with C17 and the aromatic quaternary
carbon atom C3’ (dC: 138.3 ppm). Protons H2-2’ showed 3J
HMBC correlations with the aromatic methine carbon atoms
C4’/C8’ (dC: 128.9 ppm). Protons H-4’/H-8’ showed COSY cou-
plings with protons H-5’/H-7’ and 3J HMBC correlations with
C6’ (dC: 126.7 ppm).
tion with C20 carbonyl (dC: 170.2 ppm), connecting this moiety
to the thiazolidinone ring. Protons H2-1’ also showed a 3J
HMBC correlation with the symmetric aromatic methine car-
bons C3’/C7’ (dC: 128.4 ppm). Protons H-3’/H-7’ showed a 3J
HMBC correlation with the aromatic methine carbon C5’ (dC:
127.7 ppm) and COSY correlation with protons H-4’/H-6’ (dH:
7.34 ppm). The latter protons also show COSY correlation with
Compound 4 was prepared by treatment of 1 with benzoic
anhydride in chloroform in the presence of 4-dimethylamino-
3
H-5’ (dH: 7.31 ppm) and a J HMBC correlation with the quater-
1
pyridine (DMAP) as a catalyst (Scheme 1). Analysis of H and
nary aromatic carbon C2’ (dC: 132.5 ppm).
13C NMR data indicate benzoylation at C17. The HRMS data for
4 suggest the molecular formula C29H35NO6S. The aromatic
double doublet H-3’/H-7’ (dH: 7.69 ppm) showed COSY cou-
The H and 13C NMR data for 10 were quite similar to those
1
of compound 9, with an N-benzoyl functionality instead of N-
benzyl in 9. The aromatic methine protons H-3’/H-7’ (dH:
7.72 ppm) showed a 3J HMBC correlation with the carbonyl
carbon C1’ (dC: 169.7 ppm). Notably, benzoylation of the lactol
hydroxy group at C17 in 4 and the thiazolidinone NH in 10 re-
sulted in a significant downfield shift of proton H-18 (d: +1.22
and +1.49 ppm, respectively) relative to that of 1, possibly
due to the anisotropic effect of the benzene ring and the car-
bonyl group.
3
pling with protons H-4’/H-6’ (dH: 7.42 ppm) and J HMBC corre-
lations with the carbonyl carbon atom C1’ (dC: 169.2 ppm). Pro-
tons H-4’/H-6’, in turn, showed COSY coupling with H-5’ (dH:
3
7.53 ppm) and J HMBC correlations with the aromatic quater-
nary carbon C2’ (dC: 133.4 ppm). The downfield shift of H-18
(dH: 5.07, >+1.00 ppm) relative to that of the starting material
1 is possibly due to the anisotropic effect of the newly intro-
duced C17-O-benzoyl functionality.
Reaction of 2 with p-methoxyphenylacetyl chloride afforded
11, which was highly unstable in the reaction mixture; rapid
degradation of the formed amide bond was observed. There-
fore, the reaction time was shortened to five minutes after the
addition of reagent. The HRESIMS data for compound 11 sug-
gest the molecular formula C32H41NO7S. The 1H and 13C NMR
data for 11 suggest successful N-p-methoxyphenylacetylation.
N-Substitution of 2 was carried out as shown in Scheme 1
using sodium hydride in anhydrous tetrahydrofuran at 08C
and the corresponding alkyl and aryl halides to afford 5–11.[17]
The identity of 5 was confirmed by NMR spectroscopic analy-
sis, which showed replacement of the thiazolidinone NH
proton (dH: 5.80 ppm) with an N-methyl singlet (dH: 2.97 ppm
and dC: 36.7 ppm); this is also supported by comparison with
published data.[17] Similarly, the N-ethyl functionality in 6 was
evident based on the COSY coupling between the downfield
nitrogenated methylene H2-1’ (dH: 3.71 and 3.35 ppm) with the
methyl triplet H3-2’ (dH: 1.17 ppm). Protons H2-1’ also showed
3J HMBC correlations with the carbonyl C20, connecting the
new ethyl group with the thiazolidinone ring.
3
The methoxy singlet H3-9’ (dH: 3.77 ppm) showed a J HMBC
correlation with the quaternary aromatic oxygenated carbon
atom C6’ (dC: 159.8 ppm). The aromatic protons H-4’/H-8’ (dH:
3
7.21 ppm) showed J HMBC correlations with C6’ and the ben-
zylic methylene C2’ (dC: 42.3 ppm). They also show COSY cou-
pling with protons H-5’/H-7’ (dH: 6.81 ppm). The proton singlet
2
H2-2’ (dH: 3.79 ppm) showed J HMBC correlations with the car-
The HRMS data for compound 7 suggest an additional
bonyl carbon C1’ (dC: 175.7 ppm) and the quaternary aromatic
carbon C3’ (dC: 126.3 ppm). Proton H-18 (dH: 5.30 ppm) showed
a J HMBC correlation with C1’ carbonyl, connecting the new p-
1
degree of unsaturation. Analysis of H and 13C NMR data fur-
3
ther confirmed the new N-cyclopentyl moiety. The methine
3
quintet H-1’ (dH: 3.88 ppm) showed J HMBC correlations with
methoxyphenylacetyl moiety with the thiazolidinone ring.
To explore the importance of N-substitution on the pharma-
cological effect of the unsubstituted C17 lactol group, com-
pounds 6 and 9 were demethylated by heating with aqueous
acetic acid to afford 12 and 13, respectively, with a free C17
lactol hydroxy functionality (Scheme 1).[17] Compounds 12 and
the thiazolidinone carbonyl carbon atom (dC: 171.5 ppm) and
the methylene carbon atoms C3’/C4’ (dC: 24.5 ppm). Protons
H2-3’/4’ showed COSY coupling with both H2-2’/5’ protons,
which, in turn, showed COSY coupling with H-1’.
1
Analysis of H and 13C NMR data for compound 8 suggested
1
N-hydroxypropylation of the starting material 2. The new
13 showed H and 13C NMR data identical to those of 6 and 9
3
methylene singlet H2-1’ (dH: 3.55 ppm) showed a J HMBC cor-
with the replacement of C17 methoxy with a hydroxy group.
The D2O-exchangeable broad proton singlets at d: 4.01 and
3.91 ppm were assigned as the new C17 hydroxy signals in 12
and 13, respectively.
N-Hydroxymethyllatrunculin A (14) was prepared as previ-
ously reported to study the effect of extending the location of
the only HBD in 1 (the thiazolidinone NH) by a methylene unit
relation with C29 carbonyl (dC: 170.2 ppm), connecting this
moiety to the thiazolidinone ring. Protons H2-1’ also showed a
3J HMBC correlation with the oxygenated methylene carbon
atom C3’ (dC: 59.2 ppm). The two chemically unequivalent H2-
3’ protons (dH: 3.76 and 3.58 ppm) showed COSY coupling
with H2-2’ (dH: 1.78 ppm), confirming the introduction of the
new N-hydroxypropyl side chain.
and its replacement with
(Scheme 1).[17]
a
primary alcohol group
276
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 274 – 285

Latrunculin Derivatives
Biological activity
The cytoskeleton is composed of three distinct elements: actin
microfilaments,
microtubules,
and
intermediate
fila-
ments.[1,5,9–11] Microtubules form a polarized network, enabling
the movement of organelles and proteins throughout the
cell.[5,10] Intermediate filaments are rigid components that
maintain the overall cell shape.[5,10] The actin cytoskeleton and
other proteins involved in its regulation and function consti-
tute more than 25% of the total cellular protein content.[1,5,9–11]
Cytoskeletal elements coordinate and regulate cellular motility,
adhesion, division, and exo- and endocytosis.[5,9–11] Actin fila-
ments were found to be disrupted in malignant transformed
cells.[1,5,9–11] Alteration of actin polymerization or its remodeling
is important in regulating the morphology and phenotypic
events of a malignant cell.[1,5,9–11] Actin alterations are progres-
sive, and distinct actin remodeling profiles are correlated with
various stages of cancer development and progression.[1,5,9–11]
Compound 1 reversibly binds cytoskeletal actin monomers,
forming a 1:1 complex with G-actin, thereby disrupting its
polymerization.[5–13] Compound 1 has shown antiangiogenic,
antiproliferative, antimicrobial, and antimigratory activities; it
has been directly correlated with the amelioration of diseases
in which the stability of cytoskeletal actin is intricately associat-
ed with the pathology.[1–6,10,14]
^
Figure 1. Time course for actin polymerization in the absence ( ) and pres-
An actin polymerization kit was used to assess the direct
actin binding activity of each analogue at two concentrations:
100 nm and 1 mm. This kit is based on the enhanced fluores-
cence of pyrene-conjugated actin that occurs during polymeri-
zation. The increase in fluorescence that results when pyrene–
G-actin monomers assemble to form pyrene–F-actin can be
used to follow polymerization as a function of time.[18,19]
Table 2 lists the activities as the percentage of actin polymeri-
zation inhibition relative to the negative vehicle (DMSO) con-
trol along with calculated IC50 values. Figure 1 shows the time-
dependent inhibition profiles of actin polymerization for com-
pounds 1 and 3. Compounds 3 and 14 were profoundly more
~
&
ence of a) 1 and b) 3 at doses of 0.1 ( ) and 1.0 mm ( ). Polymerization was
monitored by an increase in fluorescence intensity.
active than the parent latrunculin A (1) in the inhibition of
actin polymerization at both doses, whereas compounds 4, 9,
12, and 13 were nearly equipotent to 1. This specific binding
assay proves that the antiproliferative and anti-invasive activi-
ties of latrunculin A derivatives are due to the disruption of
actin polymerization, except in the case of compound 11.
The antiproliferative and anti-invasive activities of 1–14 were
quantified using MTT and cell invasion assays, respectively. La-
trunculin A (1) was used as a positive control in all biological
assays because it is an established actin polymerization inhibi-
tor and has documented cytotoxic, antiproliferative, and anti-
invasive activities.[1,4–6,9–16] The 3-[4,5-dimethylthiazol-2-yl]-2,5-
diphenyltetrazolium bromide (MTT) assay detects living cells in
a quantitative colorimetric fashion, as it exploits the cells’ abili-
ty to reduce MTT into an insoluble purple formazan dye.[20,21]
This method is routinely used to assess a given compound’s
cytotoxicity and proliferation inhibitory activity. The antiproli-
ferative activities of 1–14 were measured against two human
mammary gland adenocarcinoma cell lines, MCF7 and MDA-
MB-231, at five different concentrations: 0.1, 0.5, 1, 10, and
50 mm. The calculated IC50 values are listed in Table 3. While
many of the prepared semisynthetic latrunculins retained activ-
ity, some of them were more active than the parent natural
product 1. Compounds 3 and 14 were more potent than 1
against both cell lines, with respective 4.8- and 4.0-fold increas-
es in activity against MCF7, and notable cytotoxic effects at
doses greater than 1 mm. Although MDA-MB-231 cells were
more resistant than MCF7 against most derivatives, compound
Table 2. Pyrene actin polymerization assay of latrunculin A derivatives.
Actin polymerization inhibition [%]
Compound
100 nm
1 mm
IC50 [nm][a]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
30
6
31
23
11
0
20
0
33
17
0
73
21
94
67
19
18
49
21
67
53
13
70
64
90
284Æ72
9233Æ825
184Æ48
466Æ51
>10000
>10000
1049Æ271
8548Æ574
333Æ76
834Æ109
>10000
32
24
35
283Æ51
445Æ79
210Æ54
[a] IC50 values were determined from nonlinear dose–response curves
using GraphPad Prism; values are the mean ÆSEM; each experiment was
conducted in triplicate.
ChemMedChem 2010, 5, 274 – 285
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
277

K. A. El Sayed et al.
MED
of latrunculin A derivatives with their actin-disrupting effect, 1–
14 were tested at nontoxic concentrations. With the exception
of compound 3, the MTT assay results showed no or insignifi-
cant toxicity at concentrations of ꢀ1.0 mm for latrunculin deriv-
atives over 72 h. Therefore, all compounds were tested at 0.1,
0.5, and 1.0 mm. Latrunculin 3 killed 17% of MDA-MB-231 cells
at 1.0 mm, while the same concentration of 3 inhibited MDA-
MB-231 cell invasion by 95%. This shows a clear distinction be-
tween the cytotoxic and anti-invasive activity levels. Therefore,
the anti-invasive activities of 1–14 are attributed to the inhibi-
tion of actin polymerization and are not due to cell death.
Derivatives 3, 11, 13, and 14 were 3.3-, 4.6-, 3.1-, and 2.1-
fold more active than latrunculin A (1), respectively, at 0.5 mm.
17-O-Phenylethyllatrunculin A (3) was the most potent, with a
concentration of 100 nm required to inhibit MDA-MB-231 cell
invasion versus a dose of 1 mm for 1. Upon the addition of 1,
3, and 14 to MDA-MB-231 cells in the upper assay chamber,
rapid changes in cell morphology were observed, and actin fil-
aments were disrupted, as characterized by remarkable cell de-
formity (data not shown). The actin cytoskeleton is dynamically
remodeled during cell migration, and this reorganization pro-
duces the forces necessary for cell motility.[5,10,11] This suggests
that the anti-invasive activities of 1, 3, and 14 are mediated
through direct disruption of actin cytoskeleton remodeling. In-
terestingly, the second most active derivative, N-p-methoxy-
phenylacetyl-15-O-methyllatrunculin A (11) showed potent
anti-invasive activity at 100 nm (Figure 2). This activity was
nearly fourfold more active than that of 1 at the same concen-
tration without antiproliferative, cytotoxic, or cell-shape-modi-
fying activities. Furthermore, because it shows weak actin poly-
merization inhibition (Table 2), compound 11 has different tar-
get(s) other than the microfilament actin. Whereas compounds
4 and 12 are equipotent to compound 1 at concentrations of
1 and 0.5 mm, the rest were less active.
Table 3. Antiproliferative activities of latrunculin A derivatives.
IC50 [mm][a]
Compound
MCF7
MDA-MB-231
1
2
3
4
5
6
7
8
9
10
11
12
13
14
0.48Æ0.04
10.19Æ1.72
0.10Æ0.07
0.52Æ0.14
13.07Æ3.69
20.64Æ2.77
5.30Æ1.32
19.20Æ3.91
0.78Æ0.24
0.55Æ0.091
15.4Æ1.21
0.95Æ0.28
0.63Æ0.02
0.12Æ0.01
4.19Æ1.49
23.42Æ0.85
2.71Æ0.52
5.05Æ0.36
19.93Æ3.27
>50.0
5.07Æ0.72
>50.0
6.82Æ1.41
4.48Æ1.79
26.17Æ3.91
7.19Æ1.01
2.72Æ0.39
2.99Æ0.15
[a] IC50 values were determined from nonlinear dose–response curves
using GraphPad Prism; values are the mean ÆSEM; each experiment was
conducted in triplicate.
14 retained potent activity against both cell lines. Although
derivative 10 was less active than 1 in the invasion assay, this
compound, along with 4 and 9, was nearly equipotent to the
parent natural product 1 in the proliferation assay.
The anti-invasive activities of 1–14 were measured using a
96-well basement membrane extract (BME) cell invasion assay
against the highly metastatic MDA-MB-231 cells.[22,23] This assay
employs a simplified Boyden chamber design with a polyethy-
lene terephthalate membrane (8 mm pore size). Detection of
cell invasion is quantified with calcein acetomethylester
(AM).[22,23] Cells internalize calcein AM, and intracellular esteras-
es cleave the AM moiety to generate free calcein. Free calcein
fluoresces brightly, and this fluorescence is used to quantify
the number of cells that have invaded across the BME.[22,23] The
anti-invasive activities of 1–14 at three different concentrations
are shown in Figure 2. To correlate the anti-invasive activities
It has been reported that latrunculin A binds to actin mono-
mers in the presence of ATP and acts by interfering with con-
Figure 2. Anti-invasive activities of latrunculin derivatives 1–14 against MDA-MB-231 cells using a Cultrexꢂ assay kit. Error bars indicate the SEM of n=3 ex-
periments per concentration.
278
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 274 – 285

Latrunculin Derivatives
formational changes that are necessary for polymerization.[12,13]
Occupation of the latrunculin binding site clamps ATP in its
buried site by preventing movement between subdomains II
and IV.[12,13] Therefore, it is less likely that latrunculin A or its de-
rivatives can act as ATP-competitive inhibitors.
Preliminary structure–activity relationships
X-ray crystallographic data revealed the binding site of latrun-
culin A is located above the actin nucleotide binding site in a
cleft between subdomains II and IV.[12,13] The thiazolidinone
and THP ring systems are the primary pharmacophores that
direct the orientation and binding affinity, while the macrocy-
clic ring is mainly involved in hydrophobic interactions, with
no contact with actin at the C5–C7 chain.[12,13] Furthermore,
the bare macrocyclic system does not induce polymerization
of F-actin at 10 mm concentration.[16] Therefore, the two acces-
sible functional groups, C17 hydroxy and the thiazolidinone ni-
trogen atom, were used to investigate the steric limitations,
electronic properties, and HBD and HBA characteristics at
these positions to probe the chemical tolerance surrounding
the binding-determinant amino acids Tyr69, Asp157, Thr186,
Glu214, and Arg210.
Figure 3. MOLCAD visualization of the electrostatic surface potential of the
actin binding site with compound 9 docked using SurFlex-Dock program.
Some non-interacting amino acids were omitted for clarity in viewing the
important internal residues.
of Arg210 and possibly Arg206, while the acetyl moiety in 11
is too long to allow such interactions, given the added steric
effect of the methoxy group. Because this interaction is absent
in N-methyl and N-ethyl substituents of 2, derivatives 5 and 6
were drastically less active both in the antiproliferation and
anti-invasion assays. SurFlex-Dock software uses Hammerhead
screens for the binding of flexible ligands to a protein binding
site.[24,25] Ligand fragments are generated by breaking rotatable
bonds and are then aligned onto the protomol probes. The
highest-scoring fragmented poses are retained and proceed
until the ligand is complete.[24,25] This docking procedure
allows the N-benzyl substitution of 9 to be accommodated be-
tween the macrocyclic ring and Arg210 by rotating the thiazo-
lidinone moiety outward by 2.6 ꢁ and reallocating the macro-
cycle ring outward by 1.7 ꢁ. Superimposition of the docked
pose of 9 with the crystallographic structure of 1 (see Figure 6
below) shows that this shift in the thiazolidinone ring position
does not disrupt the hydrogen bonding interaction with
Thr186, but instead places the aromatic ring at an optimum
distance near the guanidine moiety of Arg210 away from the
interior compact area (Figure 3 and Figure 6).
Previously reported latrunculin derivatives were used as the
basis by which to design new derivatives to get a better un-
derstanding of the SAR.[4,14–16] Our earlier results revealed the
importance of both the thiazolidinone and THP ring systems in
actin binding.[4,14] However, a clear SAR was not revealed. For
example, it was shown that aliphatic substitution at the C17
lactol hydroxy group decreases activity,[4] unlike some aromatic
substitutions, which increase activity.[14] N-Acetylation of the
thiazolidinone nitrogen atom does not disrupt actin binding,
although it has a strong interaction with Asp157 and is consid-
ered the only HBD pharmacophore in the molecule.[4] There-
fore, the chemically accessible C17 lactol hydroxy and thiazoli-
dinone NH groups were selected to probe the chemical space
around these two critical pharmacophores.
Although the internal binding site of latrunculin A is com-
pact, aromatic or aliphatic substitutions with defined steric and
electronic properties will not change the binding orientation
or the important interactions, as described below by simulated
docking for each active derivative. Despite the fact that the
C17 lactol hydroxy group acts as an HBA and not an HBD, the
anti-invasive and antiproliferative activity of methylated deriva-
tive 2 was extensively suppressed despite the fact that the me-
thoxy oxygen atom can still act as an HBA (Figure 2, Table 3).
This result is consistent with those previously reported for 15-
O-methyllatrunculin B.[4] In contrast, aromatic substitution of
the thiazolidinone nitrogen atom restores actin binding activi-
ty, as is the case for compounds 9 and 10, but not for 11, in
which the extra steric effect of the C9’ methoxy group weak-
ens ligand binding. This outcome can be explained by
MOLCAD visualization of the docked pose of 9 in the actin
binding site (Figure 3). The benzyl moiety fills the pocket
formed by Gly156, Gly182, and Arg210, and establishes a
strong charge-transfer interaction between the p electron aro-
matic cloud of 9 and the positively charged guanidine group
Although analogue 11 was dramatically less active than 1 in
the antiproliferation assay, it was quite potent in attenuating
the metastasis of MDA-MB-231. This surprising observation can
be explained by the potential activity of 11 on protein target(s)
other than actin, as previously observed with other latrunculin
derivatives.[14,15] It is evident that certain chemical changes in
latrunculin A can ultimately change its molecular target.
N-Substitution of 1 with ethyl and benzyl groups preserves
the potency of 1 with IC50 values of 0.63 and 0.95 mm for 13
and 12, respectively, against MCF7 cells, whereas MDA-MB-231
cells are more resistant, with IC50 values of 2.72 for 13 and
7.19 mm for 12. This result is consistent with the previously
tested N-acetyllatrunculin B, which has potent antimigratory
activity against murine brain metastatic melanoma cells
(B16B15b).[4] On the other hand, the N-hydroxypropyl analogue
8
has significantly decreased activity for two reasons:
1) blocking the C17 hydroxy group with no aromatic replace-
ChemMedChem 2010, 5, 274 – 285
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
279

K. A. El Sayed et al.
MED
ment effect as discussed above, and 2) the propyl chain is too
long to allow hydrogen bonding interactions between the hy-
droxy group and the carboxylate of Asp157.
The most active derivative in actin binding, compound 3,
was docked into the active site of actin using SurFlex-Dock to
understand its significant potency. The docked pose (Figure 4)
Figure 5. Hydrogen bonding interaction of docked compound 14 with the
carboxylate moiety of Asp157 at optimal distance calculated by SYBYL 8.0.
Asp157 due to an optimal distance of 2.01 ꢁ between the two
functionalities (Figure 5), whereas this distance is 2.87 ꢁ in the
case of compound 1 as calculated by SYBYL 8.0 (Figure 6). Fur-
thermore, the primary hydroxy group forms a more stable hy-
drogen bond than the amide nitrogen atom of the thiazolidi-
none moiety.
Figure 4. Detailed view of the docked structure of 3 with the corresponding
interacting amino acids of the actin binding site.
shows optimum electrostatic interactions between the p elec-
tron cloud of the aromatic ring and the positively charged
guanidine moiety of Arg210, and maintenance of the strong
hydrogen bonding interactions with Tyr69, Asp157, Thr186,
and Glu214. The calculated distance (using SYBYL 8.0) between
the aromatic ring and the guanidine moiety is 3.32 ꢁ, which is
ideal for electrostatic interactions. This bulky phenylethyl
moiety is accommodated between the macrocyclic ring, which
is shifted outward and exposed more to the solvent (Figures 4
and 6), and Arg210 without inappropriate crash (see below) or
conformational change in actin. In a previous study, 17-O-phe-
nylcarbamoyllatrunculin A was observed to be fivefold more
potent than 1 in invasion assays against PC-3M-CT+ cells.[14]
The calculated length of the phenylcarbamoyl (CONHC6H5)
moiety, from carbonyl carbon atom to the end of the aromatic
ring, is 6.33 ꢁ, which is very close to that determined for the
phenylethyl moiety of 3 (6.12 ꢁ). Therefore, a distance of 6.10–
6.35 ꢁ between aromatic substituents and the C17 hydroxy
group can significantly improve the actin binding affinity of la-
trunculin A derivatives. It also appears that the C17 carbamate
HBA/HBD feature does not play a direct role in the ATP binding
site of actin.
Figure 6. Superimposition of the crystallographic structures of 1 (blue) with
the docked pose structures of 3 (yellow), 9 (green), and 14 (red).
One of the scoring functions that SurFlex-Dock calculates is
crash. Crash is the degree of inappropriate penetration by the
ligand into the protein and interpenetration (self-clash) be-
tween ligand atoms that are separated by rotatable bonds.
Crash scores close to zero are favorable. Negative numbers in-
dicate penetration.[24] All crash values were in the range of
À0.11 to À0.87, which indicate appropriate penetration with
no ligand–protein crash.
In the proliferation assay, latrunculin derivative 14 shows im-
proved activity over the parent latrunculin A (1) by 4- and 1.4-
fold against MCF7 and MDA-MB-231 cells, respectively, and by
2.7-fold against MDA-MB-231 cells in the invasion assay at
1 mm. This pronounced activity of 14 can be partly explained
by improved direct actin binding via the more accessible HBD
N-hydroxyethyl group relative to the thiazolidinone NH in 1
(Figure 5). The proton of the primary hydroxy group in 14 can
form a strong hydrogen bond with the carboxylate moiety of
Conclusions
Actin is involved in numerous normal cellular activities and
also plays a role in various pathological conditions. Owing to
the high potency and cytotoxicity of latrunculin A (1), this
280
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 274 – 285

Latrunculin Derivatives
MgSO4. The residue after solvent evaporation was subjected to
column chromatography (silica gel 60).[17]
study is aimed at describing the structural features that can be
altered to modulate the activity of 1.
Latrunculin A is a potent and selective inhibitor of actin
polymerization. To probe the chemical space around the bind-
ing-determinant moieties—the C17 hydroxy group and thiazo-
lidinone nitrogen atom—they were substituted by various aro-
matic and aliphatic groups. It has been shown that the C17/
C15 lactol hydroxy group is essential for actin binding affinity
and therefore noticeably affects the antiproliferative and anti-
invasive activities. Alkyl substitutions at the C17 hydroxy group
remarkably diminish biological activity; this is in contrast to ar-
omatic substitutions at the thiazolidinone nitrogen, which re-
store activity with notable increases in potency. The SAR study
of latrunculins provides insight for optimizing their activities
toward actin polymerization.
General procedure B
Alkylation of the thiazolidinone nitrogen atom: A solution of 2
in dry THF was gradually added to a suspension of NaH (60% dis-
persion in mineral oil) in dry THF at 08C. The mixture was then
stirred for 1 h at 08C. Alkyl halide was added, and the mixture was
stirred at room temperature until 2 was completely depleted. Et2O
and H2O were then added, the two layers were separated, and the
organic layer was dried over Na2SO4. Evaporation of the solvents
afforded a residue which was subjected to column chromatogra-
phy (silica gel 60).[17]
An optimum balance between moderate inhibition of actin
polymerization and low cytotoxicity can improve the therapeu-
tic index of semisynthetically modified latrunculins for future
use in controlling cancer cell invasion and migration, decrease
the intraocular pressure in glaucoma, and inhibit the formation
of tau hyperphosphorylated structures in Alzheimer’s disease.
Indeed, meeting such a balance could improve the treatment
of other diseases in which the stability of the actin cytoskele-
ton is intricately connected to the given pathology.
General procedure C
Demethylation of 17-O-methyllatrunculin A derivatives: A solu-
tion of 17-O-methyllatrunculin A analogue in AcOH/H2O/THF (3:1:1)
was stirred and heated at 608C. The reaction was monitored by
TLC until complete depletion of the starting material (~0.5–1 h).
The mixture was then cooled to room temperature, neutralized
with an aqueous solution of 10% NaHCO3, and Et2O was then
added. The upper organic layer was then washed with H2O, dried
over anhydrous Na2SO4, and evaporated under reduced pressure.
The residue was subjected to column chromatography (silica
gel 60).[17]
Experimental Section
17-O-Methyllatrunculin A (2): Compound 2 was prepared accord-
ing to procedure A from 1 (200 mg, 0.72 mmol), MeOH (5 mL), and
Et2O·BF3 (0.22 mL, 1.77 mmol). Elution with n-hexane/EtOAc (8:2)
afforded 2 (144 mg, 72.0%).[17]
General experimental procedures
Measurements of optical rotation were carried out on a Rudolph
Research Analytical Autopol III polarimeter. IR spectra were record-
ed on a Varian 800 FTIR spectrophotometer. H and 13C NMR spec-
17-O-Phenylethyllatrunculin A (3): Compound 3 was synthesized
according to general procedure A from 1 (15 mg, 0.036 mmol),
phenylethanol (1 mL, 7.131 mmol), and Et2O·BF3 (15 mL,
0.117 mmol). Elution with n-hexane/EtOAc (8:2) afforded 3 (5.5 mg,
36.7%): colorless oil, [a]2D5 =+34.5 (c=0.18 in CHCl3); 1H NMR
(400 MHz, CDCl3): d=5.41 (1H, brs, H-2), 2.70 (1H, m, H-4a), 3.57
(1H, dt, J=11.9, 3.5 Hz, H-4b), 2.33 (1H, m, H-5), 5.87 (1H, dt, J=
15 and 4.8 Hz, H-6), 6.43 (H1, dd, J=15 and 10.6 Hz, H-7), 6.08 (1H,
dd, J=10.8 and 10.6 Hz, H-8), 5.01 (1H, dd, J=10.6 and 10.4 Hz, H-
9), 2.81 (1H, m, H-10), 1.11 (1H, m, H-11a), 1.74 (1H, m, H-11b), 1.46
(2H, m, H-12), 4.24 (1H, m, H-13), 1.42 (1H, m, H-14a), 1.66 (1H, m,
H-14b), 5.17 (1H, brt, J=2.9 Hz, H-15), 1.89 (1H, dd, J=15.4 and
4.4 Hz, H-16a), 2.15 (1H, m, H-16b), 4.05 (1H, dd, J=7.3 and 7.3 Hz,
H-18), 3.07 (1H, dd, J=11.4 and 6.6 Hz, H-19a), 3.13 (1H, dd, J=
11.4 and 6.6 Hz, H-19b), 1.91 (3H, s, H3-21), 0.98 (3H, d, J=6.2 Hz,
H3-22), 3.51 (1H, dt, J=8.8 and 7.0 Hz, H-1’a), 3.81 (1H, dt, J=8.8
and 7.0 Hz, H-1’b), 2.89 (2H, t, J=7.3 Hz, H-2’), 7.17 (1H, dd, J=8.4
and 1.8 Hz, H-4’), 7.29 (1H, m, H-5’), 7.22 (1H, brd, J=7.0 Hz, H-6’),
7.29 (1H, m, H-7’), 7.17 (1H, dd, J=8.4 and 1.8 Hz, H-8’), 5.68 ppm
(1H, s, NH); 13C NMR (75 MHz, CDCl3): d=166.4 (qC, C1), 118.2 (CH,
C2), 158.1 (qC, C3), 30.7 (CH2, C4), 31.6 (CH2, C5), 132.3 (CH, C6),
124.3 (CH, C7), 127.4 (CH, C8), 135.8 (CH, C9), 29.8 (CH, C10), 32.2
(CH2, C11), 31.2 (CH2, C12), 63.5 (CH, C13), 35.3 (CH2, C14), 67.0 (CH,
C15), 29.8 (CH2, C16), 100.1 (qC, C17), 57.1 (CH, C18), 27.9 (CH2,
C19), 175.2 (qC, C20), 21.8 (CH3, C21), 25.3 (CH3, C22), 61.6 (CH2,
C1’), 36.9 (CH2, C2’), 138.3 (qC, C3’), 128.9 (CH, C4’), 128.6 (CH, C5’),
1
tra were recorded in CDCl3, with (CH3)4Si as an internal standard,
on a JEOL Eclipse NMR spectrometer operating at 400 MHz for
1H NMR and 100 MHz for 13C NMR. HREIMS experiments were con-
ducted at the University of Michigan on a Micromass LCT spec-
trometer. Analytical HPLC analyses were performed on a Dionexꢂ
Summit II system using a Phenomenex Luna 250ꢃ4.6 mm re-
versed-phase C18 column, and isocratic elution (100% MeOH) with
UV detection set at l 235 nm to verify the purity of each latruncu-
lin. Purities of >98% were established for latrunculins 1–14 except
analogue 9, which showed 91% purity. Thin-layer chromatographic
analysis was carried out on precoated silica gel 60 F254 500 mm TLC
plates (EMD Chemicals), using the developing systems n-hexane/
EtOAc (1:1) or CHCl3/MeOH (9:1). For column chromatography,
silica gel 60 (EMD Chemicals, 63–200 mm), fine silica gel 60 (EM Sci-
ence, <63 mm), and C18 silica gel (Bakerbond, Octadecyl 40 mm)
were used. For Sephadex LH-20 column chromatography, n-
hexane/CHCl3 (1:3), CHCl3, and CHCl3/MeOH (9:1) solvent systems
were used.
Chemical syntheses
General procedure A
Acetalization of the C17 hydroxy group of latrunculin A:
Et2O·BF3 was added to a solution of 1 in ROH. The mixture was
stirred for 12 h at room temperature and was then neutralized
with an aqueous solution of 10% NaHCO3. The solvent was evapo-
rated, and the residue was extracted with CHCl3 and dried over
126.7 (CH, C6’), 128.6 (CH, C7’), 128.9 ppm (CH, 8’); IR (neat): n˜max
=
;
3528, 3416, 2927, 2855, 1687, 1455, 1132, 1090, 1020, 1984 cmÀ1
HRMS-ESI: m/z [M+Na]+ calcd for C30H39NO5SNa: 548.2447, found:
548.2463.
ChemMedChem 2010, 5, 274 – 285
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
281

K. A. El Sayed et al.
MED
17-O-Benzoyllatrunculin A (4): A solution of 1 (20 mg, 0.047) in
anhydrous CHCl3 (3 mL) was stirred with benzoic anhydride (11 mg,
0.050 mmol) and a catalytic amount of DMAP for 12 h at room
temperature. The reaction mixture was neutralized with a solution
of NaHCO3. The organic layer was washed with H2O (2ꢃ5 mL),
dried over anhydrous Na2SO4, and evaporated under reduced pres-
sure. The reaction residue was subjected to chromatography (silica
gel 60) using isocratic n-hexane/EtOAc (9:1) to afford 4 (4.7 mg,
23.5%): colorless oil, [a]2D5 =+47.1 (c=0.25 in CHCl3); 1H NMR
(400 MHz, CDCl3): d=5.62 (1H, brs, H-2), 2.36 (1H, dt, J=13.2 and
7.3 Hz, H-4a), 3.02 (1H, dt, J=13.2 and 7.32 Hz, H-4b), 2.22 (2H, dt,
J=10.8 and 6.2 Hz, H-5), 5.67 (1H, dt, J=15.4 and 6.2 Hz, H-6), 6.20
(1H, dd, J=15.4 and 10.6 Hz, H-7), 5.94 (1H, dd, J=11.0 and
11.0 Hz, H-8), 4.89 (1H, dd, J=10.6 and 10.6 Hz, H-9), 2.50 (1H, m,
H-10), 1.01 (1H, m, H-11a), 1.70 (1H, m, H-11b), 1.48 (2H, m, H-12),
3.92 (1H, m, H-13), 1.01 (1H, m, H-14a), 1.69 (1H, m, H-14b), 5.19
(1H, t, J=4.0 Hz, H-15), 1.63 (1H, m, H-16a), 1.81 (1H, m, H-16b),
5.07 (1H, dd, J=7.0 and 7.0 Hz, H-18), 3.43 (1H, dd, J=11.0 and
6.6 Hz, H-19a), 3.53 (1H, dd, J=11.0 and 6.6 Hz, H-19b), 1.88 (3H, s,
H3-21), 0.66 (3H, d, J=6.2 Hz, H3-22), 7.69 (1H, dd, J=8.0 and
1.5 Hz, H-3’), 7.42 (1H, dd, J=8.0 and 7.4 Hz, H-4’), 7.53 (1H, dd,
J=7.4 and 1.5 Hz, H-5’), 7.42 (1H, dd, J=8.0 and 7.4 Hz, H-6’), 7.69
(1H, dd, J=8.0 and 1.5 Hz, H-7’), 5.61 ppm (1H, s, NH); 13C NMR
(75 MHz, CDCl3): d=166.6 (qC, C1), 118.3 (CH, C2), 157.1 (qC, C3),
33.2 (CH2, C4), 30.7 (CH2, C5), 132.5 (CH, C6), 126.2 (CH, C7), 127.8
(CH, C8), 135.9 (CH, C9), 29.3 (CH, C10), 31.5 (CH2, C11), 30.9 (CH2,
C12), 64.3 (CH, C13), 35.0 (CH2, C14), 71.0 (CH, C15), 35.7 (CH2, C16),
97.7 (qC, C17), 61.2 (CH, C18), 29.7 (CH2, C19), 172.1 (qC, C20), 24.8
(CH3, C21), 21.5 (CH3, C22), 169.2 (qC, C1’), 133.4 (qC, C2’), 129.4
(CH, C3’), 128.2 (CH, C4’), 133.1 (CH, C5’), 128.2 (CH, C6’),
129.4 ppm (CH, C7’); IR (neat): n˜max =3694, 3522, 3202, 2928, 2956,
1690, 1602, 1451, 1378, 1281, 1153, 1118, 971 cmÀ1; HRMS-ESI: m/z
[M+Na]+ calcd for C29H35NO6SNa: 548.2083, found: 548.2087.
2924, 1681, 1604, 1282, 1163 cmÀ1; HRMS-ESI: m/z [M+Na]+ calcd
for C25H37NO5SNa: 486.2290, found: 486.2281.
17-O-Methyl-N-cyclopentanelatrunculin A (7): Compound 7 was
prepared according to general procedure B starting with 2 (10 mg,
0.022 mmol) in THF (1 mL), NaH (2 mg, 0.041 mmol) in THF (1 mL)
and iodocyclopentane (0.1 mL, 0.867 mmol). Elution with n-
hexane/EtOAc (8:2) afforded 7 (7.6 mg, 70.6%): colorless oil, [a]D25
=
1
+53.6 (c=0.63 in CHCl3); H NMR (400 MHz, CDCl3): d=5.64 (1H,
brs, H-2), 1.41 (2H, m, H-4), 2.28 (2H, m, H-5), 5.80 (1H, dt, J=15.4
and 3.6 Hz, H-6), 6.37 (1H, dd, J=15.4 and 11.0 Hz, H-7), 6.04 (1H,
dd, J=11.0 and 10.6 Hz, H-8), 4.99 (1H, dd, J=10.6 and 10.6 Hz, H-
9), 2.83 (1H, m, H-10), 1.13 (1H, m, H-11a), 1.50 (1H, m, H-11b),
0.91 (1H, m, H-12a), 1.61 (1H, m, H-12b), 4.16 (1H, m, H-13), 1.41
(1H, m, H-14a), 1.61 (1H, m, H-14b), 5.16 (1H, brt, J=3.0 Hz, H-15),
1.75 (1H, m, H-16a), 1.86 (1H, m, H-16b), 3.96 (1H, dd, J=9.5 Hz,
H-18), 3.42 (1H, m, H-19a), 3.47 (1H, dd, J=9.9 and 6.6 Hz, H-19b),
1.91 (3H, s, H3-21), 1.02 (1H, d, J=6.2 Hz, H3-22), 3.88 (1H, t, J=
8.8 Hz, H-1’), 2.14 (2H, m, H-2’), 1.81 (2H, m, H-3’), 2.05 (2H, m, H-
4’), 1.81 (2H, m, H-5’), 3.28 ppm (3H, s, O-CH3); 13C NMR (75 MHz,
CDCl3): d=166.5 (qC, C1), 118.5 (CH, C2), 157.7 (qC, C3), 31.3 (CH2,
C4), 28.6 (CH2, C5), 132.4 (CH, C6), 125.0 (CH, C7), 127.8 (CH, C8),
135.6 (CH, C9), 29.8 (CH, C10), 30.8 (CH2, C11), 29.5 (CH2, C12), 63.9
(CH, C13), 35.0 (CH2, C14), 66.7 (CH, C15), 32.4 (CH2, C16), 101.6
(qC, C17), 63.1 (CH, C18), 26.2 (CH2, C19), 171.7 (qC, C20), 25.1 (CH3,
C21), 21.8 (CH3, C22), 60.7 (CH, C1’), 29.9 (CH2, C2’), 24.5 (CH2, C3’),
24.5 (CH2, C4’), 29.9 (CH2, C5’), 47.5 ppm (CH3, O-CH3); IR (neat):
n˜max =3696, 3672, 2928, 2855, 1685, 1602, 1290 cmÀ1; HRMS-ESI:
m/z [M+Na]+ calcd for C28H41NO5SNa: 526.2603, found: 526.2603.
17-O-Methyl-N-(3’-hydroxypropyl)latrunculin A (8): Compound 8
was prepared according to general procedure B from 2 (10 mg,
0.022 mmol) in THF (1 mL), NaH (2 mg, 0.041 mmol) in THF (1 mL),
and 3-iodo-1-propanol (0.1 mL, 1.043 mmol). Elution with n-
hexane/EtOAc (8:2) afforded 8 (4.6 mg, 40.6%): colorless oil, [a]D25
=
17-O-Methyl-N-methyllatrunculin A (5): Compound 5 was pre-
pared according to general procedure B using 2 as starting materi-
al (10 mg, 0.022 mmol) in THF (1 mL), NaH (2 mg, 0.041 mmol) in
THF (1 mL), and MeI (0.1 mL, 1.600 mmol). Elution with n-hexane/
EtOAc (8:2) afforded 5 (7.1 mg, 70.1%).[17]
1
+53.6 (c=0.83 in CHCl3); H NMR (400 MHz, CDCl3): d=5.64 (1H, s,
H-2), 2.24 (2H, m, H-4), 2.25 (2H, m, H-5), 5.79 (1H, dt, J=5.4 and
5.5 Hz, H-6), 6.39 (1H, dd, J=5.0 and 11.0 Hz, H-7), 6.04 (1H, dd,
J=10.7 and 10.6 Hz, H-8), 5.01 (1H, dd, J=10.6 and 10.6 Hz, H-9),
2.84 (1H, m, H-10), 1.09 (1H, m, H-11a), 1.65 (1H, m, H-11b), 1.43
(2H, m, H-12), 4.17 (1H, m, H-13), 1.46 (1H, m, H-14a), 1.77 (1H, m,
H-14b), 5.15 (1H, brt, J=3.0 Hz, H-15), 1.81 (1H, m, H-16a), 2.01
(1H, m, H-16b), 4.01 (1H, dd, J=9.9 and 4.0 Hz, H-18), 3.27 (1H, dd
J=12.8 and 4.5 Hz, H-19a), 3.37 (1H, m, H-19b), 1.91 (3H, s, H3-21),
1.01 (3H, d, J=6.6 Hz, H3-22), 3.55 (2H, m, H-1’), 1.78 (2H, m, H-2’),
3.58 (1H, m, H-3’a), 3.76 (1H, dt, J=8.1 and 5.8 Hz, H-3’b), 3.31
(3H, s, O-CH3), 5.64 ppm (s, OH); 13C NMR (75 MHz, CDCl3): d=
166.3,(qC, C1), 118.3 (CH, C2), 154.2 (qC, C3), 30.2 (CH2, C4), 29.8
(CH2, C5), 132.3 (CH, C6), 124.8 (CH, C7), 127.9 (CH, C8), 153.6 (CH,
C9), 29.7 (CH, C10), 31.4 (CH2, C11), 31.2 (CH2, C12), 63.3 (CH, C13),
35.2 (CH2, C14), 66.5 (CH, C15), 32.4 (CH2, C16), 101.0 (qC, C17),
61.4 (CH, C18), 29.4 (CH2, C19), 170.2 (qC, C20), 25.2 (CH3, C21),
21.8 (CH3, C22), 41.8 (CH2, C1’), 30.8 (CH2, C2’), 59.2 (CH2, C3’),
47.5 ppm (CH3, O-CH3); IR (neat): n˜max =3694, 3601, 3505, 2929,
2857, 2360, 1671, 1602, 1456, 1291, 1127, 1089 cmÀ1; HRMS-ESI:
m/z [M+Na]+ calcd for C26H39NO6SNa: 516.2396, found: 516.2388.
17-O-Methyl-N-ethyllatrunculin A (6): Compound 6 was prepared
according to general procedure B using 2 as starting material
(10 mg, 0.022 mmol) in THF (1 mL), NaH (2 mg, 0.041 mmol) in THF
(1 mL), and EtI (0.1 mL, 1.147 mmol). Elution with n-hexane/EtOAc
(8:2) afforded 6 (7.6 mg, 70.6%): colorless oil, [a]2D5 =+33.7; (c=1.3
in CHCl3); 1H NMR (400 MHz, CDCl3): d=5.63 (1H, brs, H-2), 2.61
(1H, m, H-4a), 2.89 (1H, m, H-4b), 2.25 (2H, m, H-5), 5.78 (1H, dt,
J=15.4 and 4.0 Hz, H-6, 6.37 (1H, dd, J=15.4 and 11.0 Hz, H-7),
6.04 (1H, dd, J=10.9 and 10.6 Hz, H-8), 4.99 (1H, dd, J=10.6 and
10.6 Hz, H-9), 2.84 (1H, m, H-10), 1.11 (1H, m, H-11a), 1.76 (1H, m,
H-11b), 1.46 (2H, m, H-12), 4.16 (1H, m, H-13), 1.07 (1H, m, H-14a),
1.72 (1H, m, H-14b), 5.14 (1H, brt, J=4.0 Hz, H-15), 1.79 (dd, J=
15.4 and 4.4 Hz, H-16a), 2.13 (1H, m, H-16b), 4.03 (1H, dd, J=9.9
and 3.0 Hz, H-18), 3.25 (1H, dd, J=11.7 and 3.0 Hz, H-19a), 3.35
(1H, m, H-19b), 1.91 (3H, brs, H3-21), 1.02 (3H, d, J=6.24 Hz, H3-
22), 3.35 (1H, dq, J=17.4, 7.0 Hz, H-1’a), 3.71 (1H, dq, J=17.4 and
7.0 Hz, H-1’b), 1.17 (3H, t, J=7.0 Hz, H-2’), 3.33 ppm (3H, s, O-CH3);
13C NMR (75 MHz, CDCl3): d=167.3 (qC, C1), 118.5 (CH, C2), 158.5
(qC, C3), 32.5 (CH2, C4), 30.7 (CH2, C5), 132.3 (CH, C6), 125.0 (CH,
C7), 127.8 (CH, C8), 135.6 (CH, C9), 29.4 (CH, C10), 31.4 (CH2, C11),
31.3 (CH2, C12), 63.2 (CH, C13), 35.1 (CH2, C14), 66.7 (CH, C15), 35.2
(CH2, C16), 101.2 (qC, C17), 60.6 (CH, C18), 25.5 (CH2, C19), 173.6
(qC, C20), 25.1 (CH3, C21), 21.8 (CH3, C22), 40.4 (CH2, C1’), 29.7
(CH3, C2’), 47.6 ppm (CH3, O-CH3); IR (neat): n˜max =3693, 3601, 2957,
17-O-Methyl-N-benzyllatrunculin A (9): Compound 9 was pre-
pared according to general procedure B using 2 as starting sub-
strate (15 mg, 0.033 mmol) in THF (1 mL), NaH (2 mg, 0.041 mmol)
in THF (1 mL), and benzyl chloride (0.1 mL, 0.866 mmol). Elution
with n-hexane/EtOAc (9:1) afforded 9 (11.3 mg, 75.3%): colorless
1
oil, [a]2D5 =+28.5 (c=0.48 in CHCl3); H NMR (400 MHz, CDCl3): d=
5.63 (1H, brs, H-2), 1.91 (1H, m, H-4a), 3.45 (1H, dt, J=13.6 and
11.0 Hz, H-4b), 2.26 (2H, m, H-5), 5.82 (1H, dt, J=13.0 and 5.1 Hz,
282
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 274 – 285

Latrunculin Derivatives
H-6), 6.35 (1H, dd, J=15.4 and 11.0 Hz, H-7), 6.04 (1H, dd, J=10.8
and 10.6 Hz, H-8), 5.01 (1H, dd, J=10.6 and 11.4 Hz, H-9), 2.80 (1H,
m, H-10), 0.89 (1H, m, H-11a), 1.67 (1H, m, H-11b), 1.42 (1H, m, H-
12a), 1.61 (1H, m, H-12b), 4.19 (1H, m, H-13), 1.38 (1H, m, H-14a),
1.88 (1H, m, H-14b), 5.15 (1H, brt, J=3.0 Hz, H-15), 1.51 (1H, m, H-
16a), 1.95 (1H, m, H-16b), 3.82 (1H, dd, J=9.2 and 2.9 Hz, H-18),
3.23 (1H, dd, J=15.0 and 2.6 Hz, H19a), 3.28 (1H, dd, J=15.0 and
3.0 Hz, H-19b), 1.90 (3H, s, H3-21), 1.01 (3H, d, J=6.2 Hz, H3-22),
4.35 (1H, d, J=14.6 Hz, H-1’a), 5.11 (1H, d, J=14.6 Hz, H-1’b), 7.23
(1H, dd, J=8.4 and 1.4 Hz, H-3’), 7.34, (1H, dd, J=7.3 and 1.4 Hz,
H-4’), 7.31 (1H, dd, J=8.4 and 7.3 Hz, H-5’), 7.34 (1H, dd, J=7.3
and 1.4 Hz, H-6’), 7.23 (1H, dd, J=8.4 and 1.4 Hz, H-7’), 3.17 ppm
(3H, s, O-CH3); 13C NMR (75 MHz, CDCl3): d=166.5 (qC, C1), 118.4
(CH, C2), 157.8 (qC, C3), 31.4 (CH2, C4), 29.8 (CH2, C5), 132.4 (CH,
C6), 124.9 (CH, C7), 127.9 (CH, C8), 135.6 (CH, C9), 29.7 (CH, C10),
30.7 (CH2, C11), 30.4 (CH2, C12), 63.4 (CH, C13), 35.2 (CH2, C14), 66.9
(CH, C15), 32.5 (CH2, C16), 101.7 (qC, C17), 59.7 (CH, C18), 25.5
(CH2, C19), 173.4 (qC, C20), 25.2 (CH3, C21), 21.8 (CH3, C22), 48.3
(CH2, C1’), 132.5 (qC, C2’), 128.4 (CH, C3’), 128.8 (CH, C4’), 127.7
(CH, C5’), 128.8 (CH, C6’), 128.4 (CH, C7’), 47.6 ppm (CH3, O-CH3); IR
(neat): n˜max =3523, 3371, 2897–2816, 1731, 1744, 1544, 1344,
4.97 (1H, dd, J=10.6 and 10.4 Hz, H-9), 2.76 (1H, m, H-10), 1.59
(2H, m, H-11), 1.11 (1H, m, H-12a), 1.83 (1H,m, H-12b), 4.01 (1H, m,
H-13), 1.47 (1H, m, H-14a), 1.88 (1H, m, H-14b), 5.03 (1H, brt, J=
3.1 Hz, H-15), 2.22 (2H, m, H-16), 5.30 (1H, d, J=9.0 Hz, H-18), 3.39
(1H, dd, J=12.1 and 4.0 Hz, H-19a), 3.45 (1H, dd, J=12.3 and
9.1 Hz, H-19b), 1.91 (3H, s, H3-21), 0.98 (3H, d, J=6.2 Hz, H3-22),
3.79 (2H, s, H-2’), 7.21 (1H, d, J=8.8 Hz, H-4’), 6.81 (1H, d, J=
8.8 Hz, H-5’), 6.81 (1H, d, J=8.8 Hz, H-7’), 7.21 (1H, d, J=8.8 Hz, H-
8’), 3.77 (3H, s, H-9’), 3.30 ppm (3H, s, O-CH3); 13C NMR (75 MHz,
CDCl3): d=166.3 (qC, C1), 118.4, (CH, C2), 158.9 (qC, C3), 32.3 (CH2,
C4), 30.9 (CH2, C5), 131.8 (CH, C6), 125.1 (CH, C7), 127.4 (CH, C8),
136.4 (CH, C9), 29.8 (CH, C10), 31.9 (CH2, C11), 31.7 (CH2, C12), 63.0
(CH, C13), 34.7 (CH2, C14), 66.9 (CH, C15), 33.1 (CH2, C16), 101.2
(qC, C17), 57.6 (CH, C18), 25.3 (CH2, C19), 171.4 (qC, C20), 25.2 (CH3,
C21), 21.8 (CH3, C22), 175.7 (qC, C1’), 42.3 (CH2, C2’), 126.3 (qC, C3’),
130.8 (CH, C4’), 114.0 (CH, C5’), 159.8 (qC, C6’), 114.0 (CH, C7’),
130.8 (CH, C8’), 55.4 (CH3, C9’), 48.2 ppm (CH3, O-CH3); IR (neat):
n˜max =3693, 3601, 2928, 2855, 2360, 1694, 1602, 1291, 1090 cmÀ1
;
HRMS-ESI: m/z [M+Na]+ calcd for C32H41NO7SNa: 606.2501, found:
606.2500.
1354 cmÀ1
;
HRMS-ESI: m/z [M+Na]+ calcd for C30H39NO5SNa:
N-Ethyllatrunculin A (12): Compound 12 was prepared according
to general procedure C starting with 6 (5 mg, 0.011 mmol) in acidic
solution (1 mL): 3 h at 608C. Elution with n-hexane/EtOAc (9:1) af-
forded 12 (1.3 mg, 26.0%): colorless oil, [a]2D5 =+63.6 (c=0.11 in
548.2447, found: 548.2443.
17-O-Methyl-N-benzoyllatrunculin A (10): Compound 10 was pre-
pared according to general procedure B using 2 as starting sub-
strate (10 mg, 0.022 mmol) in THF (1 mL), NaH (2 mg, 0.041 mmol)
in THF (1 mL), and benzoyl chloride (0.1 mL, 0.866 mmol). Elution
with n-hexane/EtOAc (8:2) afforded 10 (5.2 mg, 52.0%): colorless
1
CHCl3); H NMR (400 MHz, CDCl3): d=5.69 (1H, brs, H-2), 2.64 (1H,
m, H-4a), 2.76 (1H, m, H-4b), 2.28 (2H, m, H-5), 5.73 (1H, dt, J=
15.4 and 4.0 Hz, H-6), 6.41 (1H, dd, J=15.0 and 11.0 Hz, H-7), 5.97
(1H, dd, J=10.8 and 11.0 Hz, H-8), 5.01 (1H, dd, J=10.4 and
10.0 Hz, H-9), 2.86 (1H, m, H-10), 1.11 (1H, m, H-11a), 1.76 (1H, m,
H-1b), 1.46 (2H, m, H-12), 4.25 (1H, m, H-13), 1.45 (1H, m, H-14a),
1.80 (1H, m, H-14b), 5.44 (1H, brt, J=3.3 Hz, H-15), 1.79 (1H, dd,
J=15.4 and 4.4 Hz, H-16a), 2.14 (1H, m, H-16b), 3.83 (1H, dd, J=
11.7 and 2.2 Hz, H-18), 3.35 (1H, dd, J=12.1 and 2.2 Hz, H-19a),
3.49 (1H, dd, J=12.1 and 2.6 Hz, H-19b), 1.93 (3H, s, H3-21), 1.01
(3H, d, J=6.6 Hz, H3-22), 3.45 (1H, dq, J=17.1 and 7.0 Hz, H-1’a),
3.76 (dq, J=17.1 and 7.0 Hz, H-1’b), 1.18 (3H, t, J=7.0 Hz, H-2’),
4.01 ppm (1H, s, OH); 13C NMR (75 MHz, CDCl3): d=167.4 (qC, C1),
118.0 (CH, C2), 158.7 (qC, C3), 32.7 (CH2, C4), 30.3 (CH2, C5), 132.7
(CH, C6), 125.0 (CH, C7), 127.8 (CH, C8), 135.5 (CH, C9), 29.3 (CH,
C10), 31.4 (CH2, C11), 31.3 (CH2, C12), 63.6 (CH, C13), 35.2 (CH2,
C14), 66.6 (CH, C15), 35.4 (CH2, C16), 101.0 (qC, C17), 60.5 (CH,
C18), 25.4 (CH2, C19), 173.5 (qC, C20), 25.2 (CH3, C21), 21.7 (CH3,
C22), 40.6 (CH2, C1’), 29.9 ppm (CH3, C2’); IR (neat): n˜max =3698,
3606, 3021–2856, 2360, 1660, 1602, 1289, 1089 cmÀ1; HRMS-ESI:
m/z [M+Na]+ calcd for C24H35NO5SNa: 472.2134, found: 472.2133.
1
oil, [a]2D5 =+42.7 (c=0.15 in CHCl3); H NMR (400 MHz, CDCl3): d=
5.65 (1H, brs, H-2), 2.26 (1H, dt, J=13.2 and 7.3 Hz, H-4a), 3.48
(1H, dt, J=13.3 and 11.0 Hz, H-4b), 2.25 (2H, m, H-5), 5.76 (1H, dt,
J=14.6 and 6.2 Hz, H-6), 6.34 (1H, dd, J=14.6 and 11.7 Hz, H-7),
5.99 (1H, dd, J=11.0 and 10.8 Hz, H-8), 4.89 (1H, dd, J=11.0 and
10.8 Hz, H-9), 2.71 (1H, m, H-10), 0.77 (1H, m, H-11a), 1.13 (1H, m,
H-11b), 1.60 (2H, m, H-12), 4.05 (2H, m, H-13), 1.33 (1H, m, H-14a),
1.63 (1H, m, H-14b), 5.25 (1H, brt, J=3.3 Hz, H-15), 2.20 (1H, m, H-
16a), 3.42 (1H, m, H-16b), 5.34 (1H, dd, J=9.9 and 1.4 Hz, H-18),
3.28 (1H, dd, J=11.7 and 1.4 Hz, H-19a), 3.60 (1H, dd, J=11.7 and
9.9 Hz, H-19b), 1.92 (3H, s, H3-21), 0.95 (3H, d, J=6.6 Hz, H3-22),
7.72 (1H, dd, J=8.4 and 1.5 Hz, H-3’), 7.42 (1H, brd, J=7.3 Hz, H-
4’), 7.54 (1H, dd, J=7.3 and 1.5 Hz, H-5’), 7.42 (1H, brd, J=7.3 Hz,
H-6’), 7.72 (1H, dd, J=8.4 and 1.5 Hz, H-7’), 3.35 ppm (3H, s, O-
CH3); 13C NMR (75 MHz, CDCl3): d=166.5 (qC, C1), 118.3 (CH, C2),
157.9 (qC, C3), 32.1 (CH2, C4), 30.1 (CH2, C5), 131.9 (CH, C6), 125.2
(CH, C7), 127.3 (CH, C8), 136.2 (CH, C9), 29.8 (CH, C10), 30.8 (CH2,
C11), 30.7 (CH2, C12), 63.2 (CH2, C13), 35.1 (CH2, C14), 66.7 (CH,
C15), 32.2 (CH2, C16), 100.0 (CH, C17), 58.2 (CH, C18), 28.9 (CH2,
C19), 171.9 (qC, C20), 25.2 (CH3, C21), 21.8 (CH3, C22), 169.7 (qC,
C1’), 133.8 (qC, C2’), 129.9 (CH, C3’), 133.2 (CH, C4’), 128.1 (CH, C5’),
133.2 (CH, C6’), 129.9 (CH, C7’), 48.2 ppm (CH3, O-CH3); IR (neat):
n˜max =3693, 3531, 2928, 2856, 2356, 1684, 1453, 1351, 1278,
N-Benzyllatrunculin A (13): Compound 12 was prepared accord-
ing to general procedure C from 9 (9 mg, 0.0176 mmol) in acidic
solution (1 mL): 5 h at 608C. Elution with n-hexane/EtOAc (9:1) af-
forded 12 (1.8 mg, 20.0%): colorless oil, [a]2D5 =+63.6 (c=0.17 in
1
CHCl3); H NMR (400 MHz, CDCl3): d=5.68 (1H, brs, H-2), 1.91 (1H,
1149 cmÀ1
;
HRMS-ESI: m/z [M+Na]+ calcd for C30H37NO6SNa:
m, H-4a), 3.38 (1H, dt, J=13.6 and 11.0 Hz, H-4b), 2.25 (2H, m, H-
5), 5.75 (1H, dt, J=13.0 and 5.1 Hz, H-6), 6.36 (1H, dd, J=15.4 and
11.0 Hz, H-7), 5.97 (1H, dd, J=10.8 and 10.6 Hz, H-8), 5.01 (1H, dd,
J=11.4 and 10.6 Hz, H-9), 2.90 (1H, m, H-10), 0.89 (1H, m, H-11a),
1.75 (1H, m, H-11b), 1.44 (1H, m, H-12a), 1.67 (1H, m, H-12b), 4.27
(1H, m, H-13), 1.36 (1H, m, H-14a), 1.89 (1H, m, H-14b), 5.42 (1H,
brt, J=3.0 Hz, H-15), 1.48 (1H, m, H-16a), 2.02 (1H, m, H-16b), 3.88
(1H, dd, J=9.2 and 2.9 Hz, H-18), 3.33 (1H, dd, J=11.0 and 3.0 Hz,
H-19a), 3.66 (1H, dd, J=11.0 and 2.6 Hz, H-19b), 1.90 (3H, s, H3-21),
1.01 (3H, d, J=6.2 Hz, H3-22), 4.43 (1H, d, J=15.0 Hz, H-1’a), 5.15
(1H, d, J=15.0 Hz, H-1’b), 7.26 (1H, m, H-3’), 7.34 (1H, dd, J=7.3
and 1.4 Hz, H-4’), 7.31 (1H, m, H-5’), 7.34 (1H, dd, J=7.3 and
562.2239, found: 562.2234.
17-O-Methyl-N-(7’-methoxyphenylacetyl)latrunculin A (11): Com-
pound 11 was prepared according to general procedure B from 2
(10 mg, 0.022 mmol) in THF (1 mL), NaH (2 mg, 0.041 mmol) in THF
(1 mL), and p-methoxyphenylacetyl chloride (0.1 mL, 0.636 mmol).
Elution with n-hexane/EtOAc (8:2) afforded 11 (2.2 mg, 22.0%): col-
1
orless oil, [a]2D5 =+63.6 (c=0.1 in CHCl3); H NMR (400 MHz, CDCl3):
d=5.61 (1H, brs, H-2), 1.91 (1H, m, H-4a), 2.13 (1H, m, H-4b), 2.25
(2H, m, H-5), 5.77 (1H, dt, J=15.4 and 5.8 Hz, H-6), 5.34 (1H, dd,
J=15.7 and 10.3 Hz, H-7), 6.00 (1H, dd, J=10.6 and 10.6 Hz, H-8),
ChemMedChem 2010, 5, 274 – 285
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
283

K. A. El Sayed et al.
MED
1.4 Hz, H-6’), 7.26 (1H, m, H-7’), 3.91 ppm (1H, s, OH); 13C NMR
(75 MHz, CDCl3): d=166.7 (qC, C1), 117.5 (CH, C2), 158.4 (qC, C3),
32.1 (CH2, C4), 30.5 (CH2, C5), 132.0 (CH, C6), 125.0 (CH, C7), 127.7
(CH, C8), 136.4 (CH, C9), 29.3 (CH, C10), 31.7 (CH2, C11), 30.4 (CH2,
C12), 63.7 (CH, C13), 35.0 (CH2, C14), 66.9 (CH, C15), 32.7 (CH2, C16),
101.1 (qC, C17), 62.6 (CH, C18), 27.1 (CH2, C19), 173.8 (qC, C20),
24.6 (CH3, C21), 21.7 (CH3, C22), 48.3 (CH2, C1’), 136.4 (qC, C2’),
128.2 (CH, C 3’), 128.7 (CH, C4’), 127.4 (CH, C5’), 128.7 (CH, C6’),
128.2 ppm (CH, C7’); IR (neat): n˜max =3693, 3567, 2996–2855, 2360,
2337, 1663, 1456, 1280, 1091 cmÀ1; HRMS-ESI: m/z [M+Na]+ calcd
for C29H37NO5SNa: 534.2290, found: 534.2304.
Preparation of various dilutions of latrunculin A derivatives
for cell culture assays
A stock solution of each latrunculin A derivative was prepared by
dissolving the compound in DMSO at a concentration of 50 mm
for MTT assays and 1 mm for Cultrex assays. About 1 mL of the
former solutions was transferred to 999 mL of serum-free medium
to obtain 50 mm and 1 mm dilutions (0.001% DMSO) for MTT and
Cultrex assays, respectively. Serial dilutions were then conducted
to obtain the desired concentrations.
MTT proliferation assays
N-Hydroxymethyllatrunculin A (14): A solution of 1 (10 mg) in
EtOH (3 mL) was treated with an aqueous solution of 35% CH2O
(3 mL) and stirred for 24 h at 608C.[17] Brine solution (5 mL) was
then added, and the mixture was extracted with CHCl3 (2ꢃ5 mL).
The residue was subjected to chromatography over silica gel 60
using n-hexane/EtOAc (8:2) as a solvent system to afford 14
The growth of MCF7 and MDA-MB-231 cancer cell lines was mea-
sured by using an MTT kit (TACSTM, Trevigen Inc.).[20,21] Cells in expo-
nential growth were plated in a 96-well plate at a density of 8ꢃ103
cells per well, and allowed to attach for 24 h. Complete growth
medium was then replaced by 100 mL RPMI serum-free medium
(GIBCO-Invitrogen, NY, USA) containing various doses (50, 10, 1,
0.5, and 0.1 mm) of the specific tested compound, and culturing
was continued at 378C under 5% CO2 for 72 h. The cells were then
treated with MTT solution (10 mL per well) at 378C for 4 h. The
color reaction was stopped by the addition of solubilization/stop
solution (100 mL per well), and incubation at 378C was continued
to ensure complete dissolution of the formazan product. Absorb-
ance of the samples was determined at l 570 nm with an ELISA
plate reader (BioTek, VT, USA). The number of cells per well was cal-
culated against a standard curve prepared by plating various con-
centrations of cells, as determined by hemocytometer, at the start
of each experiment. The IC50 value for each compound was calcu-
lated by nonlinear regression (curve fit) of log(concentration)
versus the number of cells per well, implemented with GraphPad
Prism version 5.0 (GraphPad Software, La Jolla, CA, USA).
1
(2.8 mg, 28%): colorless oil, [a]2D5 =+70 (c=2.2 in CHCl3); H NMR
(400 MHz, CDCl3): d=5.70 (1H, brs, H-2), 2.97 (1H, m, H-4), 2.03
(1H, m, H-5), 5.75 (1H, dt, J=15 and 5.6 Hz, H-6), 6.40 (H1, dd, J=
15.5 and 11 Hz, H-7), 5.99 (1H, dd, J=10.8 and 10.6 Hz, H-8), 5.01
(1H, dd, J=10.6 and 10.4 Hz, H-9), 2.71 (1H, m, H-10), 1.05 (1H, m,
H-11a), 1.75 (1H, m, H-11b), 1.42 (2H, m, H-12), 4.38 (1H, m, H-13),
1.48 (1H, m, H-14a), 1.72 (1H, m, H-14b), 5.42 (1H, brs, H-15), 1.82
(1H, m, H-16a), 2.05 (1H, m, H-16b), 3.88 (1H, dd, J=7.3 and
7.3 Hz, H-18), 3.33 (1H, dd, J=12 and 9.5 Hz, H-19a), 3.51 (1H, dd,
J=12 and 2.6 Hz, H-19b), 1.93 (3H, s, H3-21), 1.00 (3H, d, J=6.5 Hz,
H3-22), 4.43 (1H, s, C17-OH), 4.65 (1H, dd, J=11.0 and 5.0 Hz, H-
1’a), 5.19 (1H, dd, J=11.0 and 9.0 Hz, H-1’b), 3.84 ppm (1H, dd, J=
9.0 and 5.0 Hz, CH2OH); 13C NMR (75 MHz, CDCl3): d=165.5 (qC,
C1), 117.5 (CH, C2), 158.2 (qC, C3), 32.4 (CH2, C4), 30.5 (CH2, C5),
132.0 (CH, C6), 126.1 (CH, C7), 127.3 (CH, C8), 136.3 (CH, C9), 29.2
(CH, C10), 31.8 (CH2, C11), 31.2 (CH2, C12), 62.8 (CH, C13), 35.0 (CH2,
C14), 67.9 (CH, C15), 32.1 (CH2, C16), 98.5 (qC, C17), 66.5 (CH, C18),
28.7 (CH2, C19), 175.5 (qC, C20), 21.6 (CH3, C21), 24.5 (CH3, C22),
68.9 ppm (CH2, C1’); IR (neat): n˜max =3400 (br), 2900, 1670.
Cultrexꢀ cell invasion assays
Anti-invasive activities were measured using the Cultrexꢂ cell inva-
sion assay (Trevigen) as previously described.[22,23] About 50 mL of
basement membrane extract (BME) coat was added per well. After
incubation for 4 h at 378C under 5% CO2, 50000 MDA-MB-231
cells per 50 mL in serum-free RPMI medium were added per well to
the top chamber containing the tested compound at the desired
concentration (0.1, 0.5, 1 mm). About 150 mL of RPMI medium was
added to the lower chamber containing 10% FBS and penicillin/
streptomycin with the use of fibronectin (1 mLmLÀ1) and N-formyl-
Met-Leu-Phe (10 nm) as chemoattractants. Cells were allowed to
migrate to the lower chamber at 378C under 5% CO2. After 24 h,
the top and bottom chambers were aspirated and washed with
washing buffer supplied with the kit. About 100 mL of cell dissocia-
tion dilution/calcein-AM solution was added to the bottom cham-
ber and incubated at 378C under 5% CO2 for 1 h. The cells inter-
nalize calcein-AM, and the intracellular esterases cleave the AM
moiety to generate free calcein. Fluorescence of the samples was
determined at lex 485 and lem 520 nm, with an ELISA plate reader
(BioTek, VT, USA). The number of cells that invaded through the
BME coat was calculated by a standard curve.
Biological procedures
Pyrene actin polymerization assays
Actin polymerization assays were carried out as per manufacturer
protocols (Cytoskeleton; Denver, CO, USA). Briefly, 5 mm final con-
centration of monomeric actin (1:10 pyrene labeled) was incubated
on ice for 10 min with the indicated concentrations of latrunculin A
derivatives. Samples were then equilibrated for 10 min in an ELISA
plate reader (BioTek, VT, USA), after which polymerization was in-
duced by the addition of KCl, MgCl2, and ATP. Compounds 1–14
were tested at 0.1, 0.5, 1.0, and 10 mm, and IC50 values for each
compound were calculated using GraphPad Prism version 5.0.
Cell culture
Molecular modeling
Breast cancer cell lines MCF7 and MDA-MB-231 were purchased
from ATCC (Manassas, VA, USA). The cell lines were grown in 10%
fetal bovine serum (FBS) and RPMI 1640 (GIBCO-Invitrogen, NY,
USA) supplemented with glutamine (2 mmolLÀ1), penicillin G
(100 mgmLÀ1), and streptomycin (100 mgmLÀ1) at 378C under 5%
CO2.
Three-dimensional structure building and all modeling were per-
formed with the SYBYL program package,[26] version 8.0, installed
on Dell desktop workstations equipped with dual 2.0 GHz Intelꢂ
Xeonꢂ processors running the Red Hat Enterprise Linux (version 5)
operating system. The X-ray crystal structure of 1 is available,[12]
284
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 274 – 285

Latrunculin Derivatives
[7] M. Okka, B. Tian, P. L. Kaufman, Arch. Ophthalmol. 2004, 122, 1482–
1488.
[8] P. L. Kaufman, B. Geiger, US Patent 772,412, 2002.
[9] I. Spector, N. R. Shochet, D. Blasberger, Y. Kashman, Cell Motil. Cytoskele-
ton 1989, 13, 127–144.
[10] C. Hayot, S. Farinelle, R. De Decker, C. Decaestecker, F. Darro, R. Kiss, M.
van Damme, Int. J. Oncol. 2002, 21, 417–425.
[11] Y. Gashet, S. Tournier, J. A. B. Millar, J. S. A. Hyams, Nature 2001, 412,
352–355.
which provided a template for modeling other latrunculins. To
maintain the active conformation, the actin-bound crystallographic
structure of 1 was used as template to sketch other derivatives
using SYBYL 8.0. Conformations of each compound were generat-
ed using ConfortTM conformational analysis. Energy minimizations
were performed using the Tripos force field with a distance-depen-
dent dielectric and the Powell conjugate gradient algorithm with a
convergence criterion of 0.01 kcalmolÀ1 AÀ1 [27]
Partial atomic charg-
.
es were calculated using the semiempirical program MOPAC 6.0
and applying the AM1 method.[28]
[12] W. M. Morton, K. R. Ayscough, P. J. McLaughlin, Nat. Cell Biol. 2000, 2,
376–378.
[13] M. R. Bubb, L. Govindasamy, E. G. Yarmola, S. M. Vorobiev, S. C. Almo, T.
Somasundaram, M. S. Chapman, M. Agbandje-McKenna, R. McKenna, J.
Biol. Chem. 2002, 277, 20999–21006.
Molecular docking
[14] K. A. El Sayed, M. A. Khanfar, H. M. Shallal, A. Muralidharan, B. Awate,
D. T. A. Youssef, Y. Liu, Y. D. Zhou, D. G. Nagle, S. Shah, J. Nat. Prod.
2008, 71, 396–402.
[15] T. Amagata, T. A. Johnson, R. H. Cichewicz, K. Tenney, S. L. Mooberry, J.
Media, M. Edelstein, F. A. Valeriote, P. Crews, J. Med. Chem. 2008, 51,
7234–7242.
[16] A. Fꢄrstner, D. Kirk, M. Fenster, C. Aꢅssa, D. De Souza, C. Nevado, T.
Tuttle, W. Thiel, O. Mꢄller, Chem. Eur. J. 2007, 13, 135–149.
[17] D. Blasberger, S. Carmely, M. Cojocaru, I. Spector, N. R. Shochet, Y. Kash-
man, Liebigs Ann. Chem. 1989, 12, 1171–1188.
The SurFlex-Dock program version 2.0 interfaced with SYBYL 8.0
was used to dock the compounds into the active site of actin.[29,30]
SurFlex-Dock employs an idealized active site ligand (protomol) as
a target to generate putative poses of molecules or molecular frag-
ments.[24,25] These putative poses were scored using the Hammer-
head scoring function.[24,25] The program was used to dock the
training set molecules into the active site of G-actin. The 3D struc-
ture was taken from the RCSB Protein Data Bank (PDB ID: 1ESV).[12]
[18] J. A. Cooper, T. D. Pollard, Methods Enzymol. 1982, 85, 182–210.
[19] L. Zhai, P. Zhao, A. Panebra, A. L. Guerrerio, S. Khurana, J. Biol. Chem.
2001, 276, 36163–36167.
Acknowledgements
[20] T. Mosmann, J. Immunol. Methods 1983, 65, 55–63.
[21] M. C. Alley, D. A. Scudiero, A. Monks, M. L. Hursey, M. J. Czerwinski, D. L.
Fine, B. J. Abbott, J. G. Mayo, R. H. Shoemaker, M. R. Boyd, Cancer Res.
1988, 48, 589–601.
[22] K. P. Tamilarasan, G. K. Kolluru, M. Rajaram, M. Indhumathy, R. Saranya,
S. Chatterjee, BMC Cell Biol. 2006, 7, 17–30.
This work was supported by the NIH grant number P20RR16456
from the INBRE Program of the National Center for Research Re-
sources.
[23] P. R. Borghesani, J. M. Peyrin, R. Klein, J. Rubin, A. R. Carter, P. M.
Schwartz, A. Luster, G. Corfas, R. A. Segal, Development 2002, 6, 1435–
1442.
Keywords: actin · antitumor agents · cancer · latrunculin A ·
structure–activity relationships
[24] W. Welch, J. Ruppert, A. N. Jain, Chem. Biol. 1996, 3, 449–462.
[25] J. Ruppert, W. Welch, A. N. Jain, Protein Sci. 1997, 6, 524–533.
[26] SYBYL Molecular Modeling Software, version 8.0, 2007, Tripos L.P.,
St.Louis, MO (USA): http://www.tripos.com.
[27] M. Clark, R. D. Cramer III, N. van Opdenbosch, Comput. Chem. 1989, 10,
982–1012.
[28] J. J. Stewart, J. Comput. Aided Mol. Des. 1990, 4, 1–105.
[29] E. Kellenberger, J. Rodrigo, P. Muller, D. Rognan, Proteins Struct. Funct.
Bioinf. 2004, 57, 225–242.
[1] a) D. J. Newman, G. M. Crag, S. Holbeck, E. A. Sausville, Curr. Cancer
Drug Targets 2002, 2, 279–308; b) J. R. Peterson, T. J. Mitchison, Chem.
Biol. 2002, 9, 1275–1285; c) G. Fenteany, S. Zhu, Curr. Topics Med. Chem.
2003, 3, 593–616.
[2] a) Y. Kashman, A. Groweiss, U. Shmueli, Tetrahedron Lett. 1980, 21,
3629—3632; b) A. Groweiss, U. Shmueli, Y. Kashman, J. Org. Chem.
1983, 48, 3512–3516.
[3] Y. Kashman, A. Groweiss, R. Lidor, D. Blasberger, S. Carmely, Tetrahedron
1985, 41, 1905–1914.
[4] K. A. El Sayed, D. T. A. Youssef, D. Marchetti, J. Nat. Prod. 2006, 69, 219–
223.
[30] A. N. Jain, J. Med. Chem. 2003, 46, 499–511.
[5] C. Hayot, O. Debeir, P. van Ham, M. van Damme, R. Kiss, C. Decaestecker,
Toxicol. Appl. Pharmacol. 2006, 211, 30–40.
Received: October 15, 2009
Revised: November 17, 2009
[6] M. Coue, S. L. Brenner, I. Spector, E. D. Korn, FEBS Lett. 1987, 213, 316–
318.
Published online on December 30, 2009
ChemMedChem 2010, 5, 274 – 285
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
285