Synthesis and biological activity of analogues of ptilomycalin A

Article abstract of DOI:10.1016/S0040-4039(01)00400-2

Benzo-fused model compounds 21a and 21b, resembling in structure the marine metabolite ptilomycalin A, were prepared and were shown to display significant activity against a series of cancer cell lines and to also possess a significant activity against th

Full text of DOI:10.1016/S0040-4039(01)00400-2

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 3377–3381
Synthesis and biological activity of analogues of ptilomycalin A
Gregory P. Black,a Simon J. Coles,e Amnon Hizi,c Andrew G. Howard-Jones,a
Michael B. Hursthouse,e Alan T. McGown,d Shoshana Loya,c Christopher G. Moore,a
Patrick J. Murphy,a,* Nigel K. Smithd and Nigel D. A. Walsheb
aDepartment of Chemistry, University of Wales, Bangor, Gwynedd LL57 2UW, UK
bPfizer Limited, Central Research, Sandwich, Kent CT13 9NJ, UK
cDepartment of Cell Biology and Histology, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel
dChristie CRC Research Centre, Paterson Institute for Cancer Research, Christie NHS Trust, Wilmslow Road,
Manchester M20 4BX, UK
eDepartment of Chemistry, University of Southampton, Highfield, Southampton SO17 1BJ, UK
Received 4 January 2001; revised 20 February 2001; accepted 8 March 2001
Abstract—Benzo-fused model compounds 21a and 21b, resembling in structure the marine metabolite ptilomycalin A, were
prepared and were shown to display significant activity against a series of cancer cell lines and to also possess a significant activity
against the DNA polymerase activity of the reverse transcriptase of human immunodeficiency virus type 1 (HIV-1 RT). © 2001
Elsevier Science Ltd. All rights reserved.
Ptilomycalin A 1 and related natural products have
been the subject of considerable synthetic interest owing
to their high levels of diverse biological activity. This
includes cytotoxicity towards several cancer cell lines as
well as antifungal and antiviral activities.1 Despite this
range of activity, only one report of the synthesis of a
reported, by Hart and Grillot, which detailed the prepa-
ration of the racemic bicyclic guanidine 22 (Scheme 1).
As can be seen, ptilomycalin A can be considered as
four separate components, a guanidine-containing poly-
cyclic core, an ester linkage to an v-hydroxyacid spacer
unit and a terminal spermidine residue. In common
structural analogue of ptilomycalin
A has been
Scheme 1.
* Corresponding author.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00400-2

3378
G. P. Black et al. / Tetrahedron Letters 42 (2001) 3377–3381
Scheme 2. (a) (i) Guanidine, DMF, 3 h; (ii) MeOH, HCl, 0°C to rt, 24 h; (iii) NaBF4 (satd, aq.), (25–35%). n=Variable, X=CH2,
O.
with this, the analogue 2 was designed to mimic the
natural material 1 with one exception in that it was
based upon a much simpler bicyclic guanidine core.2
Unfortunately, it was reported that 2 was inherently
unstable and over a period of a few weeks completely
decomposed via an unidentified process which pre-
vented the biological evaluation of this model com-
pound. The structure of the decomposition products
were not fully determined, although NMR spectro-
scopic analysis revealed that the signal due to the
methylene group adjacent to the ester oxygen was no
longer present. This observation led the authors to
speculate that the role of the spiro N,O-acetal groups in
the natural product 1 is to sterically protect the ester
linkage from hydrolysis or ammonolysis.2 As part of
our general synthetic programme in this area of chem-
istry, we were interested in preparing analogues of 1 to
evaluate the structural features that are necessary to
impart biological activity. The preliminary results of
this work are outlined in this communication.
formed as a single diastereoisomer and X-ray analysis†
of 9 confirmed the relative stereochemistry to be identi-
cal to that found in ptilomycalin A. Encouragingly, the
fluoroborate anion was found to interact with the
guanidinium ion in a bidentate ligating mode similar to
that observed with carboxylates and phosphates (Fig. 1:
The data for the N-H···F hydrogen bonds, H···F, N···F
,
distances and N-H···F angles are 2.19/2.06 A, 2.99/2.92
4
,
A and 155.72/174.57°, respectively). The relative stere-
ochemistry of the molecule is a key factor in the
synthesis of analogues of 1 as it is known that disrup-
tion of the all cis arrangement found in ptilomycalin A
has a considerable effect on biological activity1 (Scheme
3).
After several unsuccessful attempts at derivatising the
aromatic ring of 9 by electrophilic substitution, we
decided to adopt a more direct approach by preparing
4-substituted phthaldehydes which could be modified to
the required model compounds. One reported5 prepara-
tion of 4-substituted phthaldehydes is via ozonolysis of
2-substituted naphthalenes and hence we considered
naphthalene 13 to be a suitable precursor. Isomerisa-
tion of the alkyne in commercially available 7-hexade-
cyn-1-ol 10 to the terminus by means of the ‘Zipper’
reaction,6 followed by silyl protection of the alcohol
group gave 11 in good overall yield. This was then
converted into the corresponding vinyl boronic acid
using catecholborane and coupled to 2-bromo-6-
methoxynaphthalene under Suzuki conditions to give
alkene 12 in 78% yield. Hydrogenation of the alkene
group under standard conditions gave the required
substrate 13 in 91% yield. The key reaction of 13 with
Our strategy was to follow the basic design of the
natural molecule but to replace the potentially labile
ester linkage with a more stable function. It was envis-
aged that model compounds of general structure 3,
possessing a benzo-fuzed pentacyclic guanidine group
could be prepared in a similar manner to substances
reported in our model studies on the synthesis of
ptilomycalin A. These studies detailed the preparation
of pentacycles such as 5 from the bis-enone 4 by
sequential addition of guanidine, deprotection and
spirocyclisation3 (Scheme 2).
It was envisaged that the linker group in 3 would be
based on either an ether linkage (X=O) or an alkylated
benzene (X=CH2). The preparation of 3 in a conver-
gent manner would offer the possibility of varying the
chain length between the guanidine and spermidine
moieties, thereby enabling an investigation of the role
of this parameter in the biological activity of ptilomy-
calin A and related metabolites. We firstly needed to
determine if it was indeed possible to prepare the
required benzo-fuzed guanidine pentacycle and thus
reacted phthaldehyde 6 with the previously reported
phosphorane 73 leading to the bis-enone 8 in 85% yield.
Reaction of this with guanidine under our standard
conditions routinely produced the desired pentacyclic
guanidine 9 in 30–40% overall yield as a crystalline
compound. From NMR analysis of the crude reaction
mixtures it was also apparent that this material was
†
Crystal data for C21H28N3O2·BF4·CHCl3: Mr=500.45, monoclinic,
,
a=18.303(3), b=16.421(2), c=17.961(4) A, i=119.30(6)°, U=
3
4707.6(11) A , space group C2/c, Z=8, Dc=1.412 g cm−3, F(000)=
,
2084, (Mo Ka)=0.275 mm−1. Data were collected at 150(2) K, for
a crystal of dimensions 0.23×0.22×0.17 mm, on a FAST TV Area
detector diffractometer following previously described proce-
dures.10a A total of 9061 data were recorded (index ranges=
−19h19, −15k17, −19l19) and merged to give 3049 unique reflec-
tions (Rint=0.0723). The structure was solved via direct methods
(SHELX-S)10b and then refined by full matrix least-squares on all
F2 data (SHELX-93).10c The final R, Rw indices [I>2(I)] were
o
0.0797, 0.1935 and 0.1123, 0.2055 for all data, respectively, with 367
parameters. Full details of the data collection, structure refinement,
atomic coordinates, bond lengths and angles, and thermal parame-
ters have been deposited at the Cambridge Crystallographic Data
Centre (Deposition Number=148301).

G. P. Black et al. / Tetrahedron Letters 42 (2001) 3377–3381
3379
Figure 1.
ozone under the conditions described by Pappas et al.5
did indeed lead to selective ozonolysis of the methoxy-
substituted ring, however, the yields for this process
were not as high as those reported for the simpler
substrates described in the original work. Typical yields
of phthaldehyde 14 were of the order of 20–25% which
could be improved by stopping the reaction at ca. 50%
completion and recycling the recovered starting mate-
rial, eventually leading to an overall yield of 40%
(Scheme 4).
We also investigated a simpler approach to the prepara-
tion of a 4-substituted phthaldehyde and prepared 17
via a convenient three-step route which involved alkyla-
tion of diol 15 using 2-bromomethylnaphthalene, fol-
lowed by silyl protection of the remaining alcohol
leading to 16 in high yield. Ozonolysis of 16 led to the
dialdehyde 17, but again the yields for this step were
poor (20–30%), however, the ease of preparation of the
precursors enabled the synthesis of gram quantities for
further study (Scheme 5).
Scheme 3. (a) CH2Cl2, reflux, 6 h, (85%); (b) (i) DMF,
guanidine, 0–20°C, 5 h; (ii) MeOH, HCl, 0°C, 1 h, then 20°C,
15 h; (iii) NaBF4 (satd, aq.); (iv) trituration and crystallisa-
tion, (30–40%).
Scheme 4. (a) Li, 1,3-diaminopropane, 70°C, KOt-Bu, Et2O, (72%); (b) Imid, TBSCl, DMF, 2.5 h, (98%); (c) (i) catecholborane,
reflux, 3 h, rt, 16 h; (ii) PhH, Pd (PPh3)4, 2-bromo-6-methoxynaphthalene, 2.5 h, then NaOEt/EtOH (2 M), reflux, 2 h; (iii) NaOH
(3 M), rt, 16 h, (78%); (d) EtOAc, Pd/C, H2, 10 min, (91%); (e) CH2Cl2, O3, −78°C then PPh3, (40%).

3380
G. P. Black et al. / Tetrahedron Letters 42 (2001) 3377–3381
Scheme 5. (a) NaH, THF, 2-bromomethylnaphthalene, (64%); (b) Imid, TBSCl, DMF, (93%); (c) CH2Cl2, O3, −78°C then PPh3,
(20–30%).
With the two aldehydes 14 and 17 in hand we proceeded
to the synthesis of the analogues by reacting them with
the phosphorane 7 under standard conditions to yield the
enones 18a and 18b in 70 and 76% yield, respectively.
Addition of guanidine was accomplished as before to give
19a in 25% yield and 19b in 28% yield, which are typical
yields for this reaction. Analysis of the proton and
carbon NMR spectra for these substances indicated that
they gave near identical signals to compound 9 and had
also been formed as single diastereomers. Oxidation of
the terminal hydroxyl groups was accomplished using
PDC/DMF7 in 54 and 85% yields, respectively, and
coupling of the carboxylic acids with bis-Boc protected
spermidine8 using EDCl proceeded in 88 and 87% yield
for the two substrates. Finally, acidic deprotection of the
Boc groups gave the required analogues 21a and 21b in
quantitative yields (Scheme 6).
The compounds 21a and 21b were tested against four
cancer cell lines and also for their capacity to inhibit the
enzyme reverse transcriptase of human immunodefi-
ciency virus type 1 (HIV-1 RT)9 (Table 1). Of the two
analogues, compound 21a showed by far the best activity
against all the cell lines studied, demonstrating an IC50
activity at less than 1 mg ml−1 in each case, and also
displayed 64% inhibition of HIV-1 RT. Compound 21b
also showed activity against the same cell lines but to a
lesser extent than 21a, together with 55% inhibition of
HIV-1 RT. Comparison of these results with those
obtained from ptilomycalin A indicated that the activity
of 21a was comparable in magnitude to the natural
material for all the cell lines studied and that both 21a
and 21b gave near identical levels of HIV-1 RT inhibition
to 1. The levels of activity for the simple benzo ana-
logue 9 were also investigated and this proved to be
Scheme 6. (a) 3 equiv. 7, CH2Cl2, 24–48 h, (70, 76%); (b) (i) guanidine, DMF, 7 h, 0°C to rt; (ii) MeOH, HCl, 0°C to rt, 24 h;
(iii) NaBF4 (satd, aq.), (25, 28%); (c) PDC, DMF, (54, 85%); (d) bis-(N-1, N-8)-Boc-spermidine, EDCl, HOBT; (88, 87%); (e) for
21a: CHCl3, CF3CO2H; for 21b: HCl (3N, aq.); (quant.).
Table 1.
Compound
K562a
A2780a
H-460a
P388a
HIV-1 RTb (%)
Ptilomycalin A 1
0.35
0.52
7.28
0.27
0.92
11.02
22.06
0.35
0.52
5.87
0.11c
0.69
4.25
9.11
60
64
55
0
21a
21b
9
24.93
21.36
a Cytotoxic activity (IC50/mg ml−1); K562: human chronic myelogenous leukaemia; A2780: human ovarian carcinoma; H-460: human large cell
carcinoma; lung. High DT-diaphorase; P388: mouse, lymphoid neoplasm.
b Percentage of inhibition of DNA polymerase activity of HIV-1 reverse transcriptase in the presence of final concentrations of 10 mM inhibitor.
c Data obtained from Refs. 1a and 1b.

G. P. Black et al. / Tetrahedron Letters 42 (2001) 3377–3381
3381
the least active compound of the three, which indicates
References
that the presence of a spacer chain and spermidine
residue are essential for the compounds to be able to
demonstrate significant biological activity. It is also
apparent that inclusion of the ester linkage is not
necessary for activity and that the obvious changes in
the overall geometric arrangement of the guanidine and
spermidine residues has not affected the levels of activ-
ity to a great extent.
1. (a) Kashman, Y.; Hirsh, S.; McConnell, O. J.; Ohtani, I.;
Kusumi, T.; Kakisawa, H. J. J. Am. Chem. Soc. 1989,
111, 8925–8926; (b) Ohtani, I.; Kusumi, T.; Kakisawa,
H.; Kashman, Y.; Hirsh, S. J. Am. Chem. Soc. 1992, 114,
8472–8479; (c) For a review, see: Heys, L.; Moore, C. G.;
Murphy, P. J. Chem. Soc. Rev. 2000, 29, 57–67.
2. (a) Grillot, A.-L.; Hart, D. J. Tetrahedron 1995, 51,
11377–11392; (b) Grillot, A.-L.; Hart, D. J. Heterocycles
1994, 39, 435–438.
3. Murphy, P. J.; Williams, H. L.; Hibbs, D. E.; Hurst-
house, M. B.; Malik, K. M. A. Tetrahedron 1996, 52,
8315–8332.
4. Murphy, P. J.; Williams, H. L.; Hibbs, D. E.; Hurst-
house, M. B.; Malik, K. M. A. J. Chem. Soc., Chem.
Commun. 1996, 445–447.
In conclusion, we have successfully prepared two novel
benzo-fused pentacyclic analogues of ptilomycalin and
have shown that both compounds display significant
activity against a series of cell lines and also HIV-1 RT
inhibitory activity. Drawbacks in the synthesis of these
compounds were the poor yields observed for the
ozonolysis of the naphthalene intermediates leading to
the 4-substituted phthaldehydes. In spite of this we are
currently investigating alternative methods for the syn-
thesis of these analogues and intend to use this work as
a platform to prepare a series of compounds in which
variations in length of the alkyl spacer are introduced
to ascertain any effects this might have on biological
activity.
5. Pappas, J. J.; Keaveney, W. P.; Berger, M.; Rush, R. V.
J. Org. Chem. 1968, 33, 787–792.
6. Brown, C. A.; Yamashita, A. J. Am. Chem. Soc. 1975, 97,
891–892.
7. Corey, E. J.; Schmidt, G. Tetrahedron Lett. 1979, 399–
402.
8. Cohen, G. M.; Cullis, P. M.; Hartley, J. A.; Mather, A.;
Symons, M. C. R.; Wheelhouse, R. T. J. Chem. Soc.,
Chem. Commun. 1992, 298–300.
Acknowledgements
9. The assay used is as described in: Hizi, A.; Tal, R.;
Shaharabany, M.; Loya, S. J. Biol. Chem. 1991, 266,
6230–6239.
We are grateful to Professor Yoel Kashman for the gift
of a sample of ptilomycalin A and thanks are also given
to the EPSRC and Pfizer Central Research for a CASE
studentship to G.P.B., to the EPSRC for a quota award
to C.G.M. and to U.C.N.W. and the TFW scheme for
funding to A.G.H.-J. The support of the EPSRC Mass
Spectrometry centre at Swansea is also acknowledged.
10. (a) Drake, S. R.; Hursthouse, M. B.; Malik, K. M. A.;
Miller, S. A. S. Inorg. Chem. 1993, 32, 4653–4657; (b)
Sheldrick, G. M. Acta Crystallogr. 1990, A46, 467–473;
(c) Drake, S. R.; Hursthouse, M. B.; Malik, K. M. A.;
Miller, S. A. S. Inorg. Chem. 1993, 32, 4653–4657; (d)
Sheldrick, G. M. University of Gottingen: Germany,
1993, unpublished work.
.
.