Synthesis and biological evaluation of didemnin photoaffinity analogues

Article abstract of DOI:10.1016/S0960-894X(01)00339-0

The synthesis of four benzophenone-containing analogues of the antiproliferative natural product didemnin B is presented. In vitro protein biosynthesis inhibition potency and antitumor activity were evaluated. The results indicate that all four analogues are biologically active and could serve as photoaffinity reagents for the study of receptor-binding interactions of didemnins. These analogues could also be useful in studying antitumor effects of didemnins.

Full text of DOI:10.1016/S0960-894X(01)00339-0

Bioorganic & MedicinalChemistry Letters 11 (2001) 1871–1874
Synthesis and Biological Evaluation of Didemnin Photoaffinity
Analogues
Matthew D. Vera,a Amy J. Pfizenmayer,a Xiaobin Ding,a Deepika Ahuja,b
Peter L. Toogoodb and Madeleine M. Joulliea,*
aDepartment of Chemistry, University of Pennsylvania, Philadelphia, PA 19104-6323, USA
bDepartment of Chemistry, University of Michigan, Ann Arbor, MI 48109-1055, USA
Received 21 March 2001; accepted 2 May 2001
Abstract—The synthesis of four benzophenone-containing analogues of the antiproliferative natural product didemnin B is pre-
sented. In vitro protein biosynthesis inhibition potency and antitumor activity were evaluated. The results indicate that all four
analogues are biologically active and could serve as photoaffinity reagents for the study of receptor-binding interactions of
didemnins. These analogues could also be useful in studying antitumor effects of didemnins. # 2001 Elsevier Science Ltd. All rights
reserved.
The didemnin class of marine depsipeptides (1 and 2,
Fig. 1) has attracted considerable interest resulting from
between didemnin B and its two binding proteins is not
understood at any level of structural detail.
the wide spectrum of biological activities exhibited by
1À3
didemnin B, the initiaellad congener.
Although
In order to address this aspect of didemnin biochem-
istry, we now report the synthesis and biological eval-
uation of didemnin photoaffinity analogues. Before
attempting to use any analogues for photoaffinity label-
ing, we decided to address two separate questions. First,
we wished to establish whether the presence of a pho-
tophore in the side chain would be tolerated. Secondly,
we wished to determine whether there is an optimal
length of the linker between the photophore and the
didemnin nucleus. Analogues 3–6 (Fig. 1) were designed
to explore both issues.
didemnins have shown antiviral, immunosuppressive,
and protein biosynthesis inhibition effects, the anti-
proliferative and cytotoxic actions have led to clinical
trials of some congeners against cancer.4 The clinical
efficacy of didemnins remains uncertain. Moreover, lit-
tle is understood about the molecular basis of their
biological activities.
Two didemnin binding proteins, eukaryotic elongation
factor 1a (EF-1a)5 and palmitoyl protein thioesterase 1
(PPT1)6 have been identified. While the mechanistic
relevance of the interaction with PPT1 is unclear,7
binding of didemnin B to EF-1a explains the protein
biosynthesis inhibition effects.8 Using a tritiated didem-
nin B radioligand, we recently reported evidence show-
ing that a ternary complex consisting of ribosome,
didemnin B, and EF-1a serves to inhibit the EF-2-medi-
ated translocation step of eukaryotic peptide elongation
in vitro.9,10 Although mechanistically interesting, the
inhibition of protein biosynthesis by didemnins is not
Photoaffinity analogue candidates 3–6 are formally
didemnin A derivatives containing the 4-benzoylbenzoic
acid moiety. We chose this type of benzophenone deri-
vative as the photoreactive group because of its multiple
advantages in photoaffinity labeling.12,13 Benzophe-
nones are easily manipulated in ambient light, but are
photoactivated at about 350 nm, a wavelength which
will not damage the target protein. The activated
diradical of benzophenone tends to avoid solvent mole-
cules and nucleophiles, cross-linking preferentially at
unreactive carbon centers. Because the photoexcitation
of benzophenones is reversible, a given molecule of the
analogue which is not proximal to the ligand binding
site can undergo several excitation–relaxation cycles
without suffering nonspecific side reactions. Thus,
sufficient to explain the cellular apoptosis induced by
11
these naturalproducts.
Moreover, the interaction
*Corresponding author. Tel.: +1-215-898-3158; fax: +1-215-898-
5129; e-mail: mjoullie@sas.upenn.edu
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00339-0

1872
M. D. Vera et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1871–1874
benzophenones tend to cross-link with greater site-spe-
cificity than photophores such as arylazides, which
undergo an irreversible photolysis reaction.
among side-chain variants.7 The consequences of benzo-
phenone incorporation for antitumor activity, however,
were more difficult to predict.
The side-chain portion of didemnins is easily modified
by chemicalsynthesis. Based upon the SAR profiel for
protein biosynthesis inhibition by didemnins,3,8 EF-1a
was anticipated to be tolerant of a structural change
such as a benzophenone moiety in the side chain. Like-
wise, PPT1 has been shown to be rather undiscriminating
Following our approach to the synthesis of didemnin
congeners and analogues,14,15 the side-chain portions of
3–6 were synthesized and then coupled to an advanced
macrolide intermediate as a final step. Scheme 1 shows
the synthesis of the d-leucine derivative which is com-
mon to all four analogues. d-Leucine was converted to
its benzylcarbamate ( 7) and then dialkylated with
dimethyl sulfate. Catalytic hydrogenolysis of the carba-
mate followed by trapping with HCl provided amine
hydrochloride salt 9. As shown in Scheme 2, 9 was
coupled to 4-benzoylbenzoic acid using BOP-Cl, a
reagent which is generally useful in coupling to N-
methylamines. Ester saponification deilvered side chain
12.
Benzophenone derivative 10 (Scheme 3) was elongated
by BOP-mediated coupling with a glycine methyl ester
linker. Saponification of ester 13 followed by coupling
with amine salt 9 and saponification afforded side chain
15.
Scheme 2. (a) BOP-Cl, NMM, DMF, 0 ꢀC, 30 min, then amine
hydrochloride 9, 58%; (b) LiOH, H2O, THF 85%.
Figure 1. Structures of didemnins A and B and photoaffinity ana-
logues 3–6.
Scheme 1. (a) Benzylcholroformate, NaHCO
,
H2O, 87%; (b)
Scheme 3. (a) BOP, NMM, DMF, glycine methyl ester hydrochloride
salt, 82%; (b) LiOH, H2O, THF, MeOH, 83%; (c) BOP-Cl, NMM,
DMF, amine salt 9, 30%; (d) LiOH, H2O, THF, MeOH, 93%.
3
Me2SO4, powdered KOH, THF, Bu4N+HSO4 (cat.) 86%; (c) H2, 10%
Pd/C, EtOAc/MeOH (1:1); (d) HCl/MeOH, 96% (two steps).

M. D. Vera et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1871–1874
1873
The side chains leading to analogues 5 and 6 were ela-
borated using the strategy described above, incorporat-
ing either one or two 6-aminohexanoic acid moieties as
linkers. The four benzophenone-containing side-chain
acids were then coupled to the previously reported15,16
macrocycle salt (16) using the BOP coupling reagent to
obtain analogues 3–6 in 68–81% yield (Scheme 4, Table 1).
(GI50) shows that all four analogues retain potency
comparable to didemnin B. In contrast to the result for
in vitro protein biosynthesis, there does not appear to
be any clear-cut relationship between linker length and
activity. However, totalgrowth inhibition (TGI) by the
four analogues requires significantly higher concentra-
tions than are required by didemnin B. Interestingly, all
four of the analogues exhibit three to 10-fold lower
toxicity as measured by the LC50.
In order to evaluate their suitability for photoaffinity
labeling of the EF-1a binding protein, analogues 3–6
were assayed for protein biosynthesis inhibition potency
(Table 1) in a cell-free translation system as described
previously.8,9 In this assay, didemnin B (2) exhibited an
IC50 of 3 mM. The results in Table 1 show that protein
biosynthesis inhibition potency remains intact in spite of
the incorporation of the large benzophenone moiety.
The length of the linker seems to exert a marginal effect,
with the two shortest side chains (3 and 4) being nearly
equipotent with didemnin B. Increasing side-chain
length correlates with modest reductions in inhibitor
potency. These results are consistent with others3,8
showing that protein biosynthesis inhibition is broadly
tolerant of side-chain modification.
In conclusion, benzophenone-containing analogues of
didemnins are accessible by total synthesis. Biological
evaluation of the analogues indicates that incorporation
of benzophenone in the side-chain peptide could be a fea-
sible strategy for photoaffinity labeling of the molecular
targets of didemnins.
Acknowledgements
The authors thank the Developmental Therapeutics
Program of the NationalCancer Institute for evaul ating
compounds 3–6 in the NCI-60 tumor cell screen. Addi-
tionally, we gratefully acknowledge the NIH (CA-
40081) and NSF (CHE-01449) for financialsupport of
this work.
Table 2 shows preliminary results from the NCI-60
tumor cell screen.17 The data shown indicate that the
benzophenone photoaffinity analogues may also be
usefulin studying antitumor effects of didemnins.
Examination of the 50% growth inhibition parameter
References and Notes
1. Li, W.-R., Joullie, M. M. In Studies in Natural Products
Chemistry, Vol. 10, Stereoselective Synthesis (Part F); Atta-ur-
Rahman, Ed.; Elsevier: Amsterdam, 1992; p 241–302.
2. Rinehart, K. L., Jr.; Gloer, J. B.; Hughes, R. G., Jr.; Renis,
H. E.; McGovren, J. P.; Swynenberg, E. B.; Stringfellow,
D. A.; Kuentzel, S. L.; Li, L. H. Science 1981, 212, 933.
3. Sakai, R.; Rinehart, K. L.; Kishore, V.; Kundu, B.; Fair-
cloth, G.; Gloer, J. B.; Carney, J. R.; Namikoshi, M.; Sun, F.;
Hughes, R. G., Jr.; Gravalos, G.; de Quesada, T. G.; Wilson,
G. R.; Heid, R. M. J. Med. Chem. 1996, 39, 2819.
4. Rinehart, K. L. Med. Res. Rev. 2000, 20, 1.
5. Crews, C. M.; Collins, J. L.; Lane, W. S.; Snapper, M. L.;
Schreiber, S. L. J. Biol. Chem. 1994, 269, 15411.
6. Crews, C. M.; Lane, W. S.; Schreiber, S. L. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 4316.
Scheme 4. (a) BOP, NMM, CH2Cl2.
Table 1. Coupling yields and protein biosynthesis inhibition activity
for compounds 2–6
Compound
Coupling yield (%)
IC50 (mM)
Ref
2
3
4
5
6
3.0
4.0
4.5
24
68
78
74
81
18a
18b
18c
18d
7. Meng, L.; Sin, N.; Crews, C. M. Biochemistry 1998, 37,
10488.
8. SirDeshpande, B. V.; Toogood, P. L. Biochemistry 1995,
34, 9177.
19
9. Ahuja, D.; Geiger, A.; Ramanjulu, J. M.; Vera, M. D.;
SirDeshpande, B. V.; Pfizenmayer, A.; Abazecd, M.; Krosky,
D. J.; Beidler, D.; Joullie, M. M.; Toogood, P. L. J. Med.
Chem. 2000, 43, 4212.
Table 2. Mean-graph results from the NCI-60 tumor cell screen
10. Ahuja, D.; Vera, M. D.; SirDeshpande, B. V.; Morimoto,
H.; Williams, P. G.; Joullie, M. M.; Toogood, P. L. Biochem-
istry 2000, 39, 4339.
11. Beidler, D. R.; Ahuja, D.; Wicha, M. S.; Toogood, P. L.
Biochem. Pharmacol. 1999, 58, 1067.
12. Dorman, G.; Prestwich, G. D. Biochemistry 1994, 33,
5661.
13. Fleming, S. A. Tetrahedron 1995, 51, 12479.
Compound
GI50 (nM)
TGI (mM)
LC50 (mM)
2
3
4
5
6
13
3.0
13
4.3
17
0.066
0.35
2.0
21
5.0
3.8
15
23
19
48

1874
M. D. Vera et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1871–1874
14. Li, W.-R.; Ewing, W. R.; Harris, B. D.; Joullie, M. M. J.
Am. Chem. Soc. 1990, 112, 7659.
15. Mayer, S. C.; Ramanjulu, J.; Vera, M. D.; Pfizenmayer,
A.; Joullie, M. M. J. Org. Chem. 1994, 59, 5192.
16. Mayer, S. C.; Carroll, P. J.; Joullie, M. M. Acta Cryst.
1995, C51, 1609.
17. Boyd, M. R.; Paull, K. D. Drug Dev. Res. 1995, 34, 91.
18. Compounds 3–6 gave satisfactory 1H and 13C NMR
spectra. (a) Compound 3: Rf 0.41 (5% MeOH/CHCl3); [a]D25
À43.8ꢀ (c 0.445, CHCl3); HRMS m/z calcd for
C63H86N6O14+H (M+H) 1151.6280, found 1151.6261. (b)
Compound 4: Rf 0.29 (5% MeOH/CHCl3); [a]2D5 À33.4ꢀ (c
0.445, CHCl3); HRMS m/z calcd for C65H89N7O15+H
(M+H) 1208.6494, found 1208.6479. (c) Compound 5: Rf 0.25
(5% MeOH/CHCl3); [a]2D5 À60.9ꢀ (c 0.525, CHCl3); HRMS
m/z calcd for C69H97N7O15+H (M+H) 1264.7120, found
1264.7153. (d) Compound 6: Rf 0.21 (5% MeOH/CHCl3);
[a]2D5 À46.4ꢀ (c 0.835, CHCl3); HRMS m/z calcd for
C75H108N8O16+Na (M+Na) 1399.7780, found 1399.7775.