Observations on the reactivity of thiyl radicals derived from 3,6-epidithiodiketopiperazine-2,5-diones and related congeners

Article abstract of DOI:10.1016/j.bmcl.2005.03.022

A range of thiyl radicals derived from the reduced form of epidithiodiketopiperazines (ETPs) act as polarity reversal catalysts for the hydrosilylation of an enol lactone but not for H-atom abstraction from a model ribose ester.

Full text of DOI:10.1016/j.bmcl.2005.03.022

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2239–2242
Observations on the reactivity of thiyl radicals derived from
3,6-epidithiodiketopiperazine-2,5-diones and related congeners
S. T. Hilton,a W. B. Motherwell,a,* P. Potier,b C. Pradeta and D. L. Selwoodc
aChemistry Department, University College London, Christopher Ingold Laboratories, 20 Gordon Street, London WC1H OAJ, UK
bInstitut de Chimie des Substances Naturelles, CNRS, 91198 Gif-sur-Yvette, France
cThe Wolfson Institute for Biomedical Research, University College London, The Cruciform Building, Gower Street,
London WC1E 6BT, UK
Received 10 November 2004; revised 22 January 2005; accepted 7 March 2005
Abstract—A range of thiyl radicals derived fromthe reduced formof epidithiodiketopiperazines (ETPs) act as polarity reversal
catalysts for the hydrosilylation of an enol lactone but not for H-atomabstraction froma model ribose ester.
Ó 2005 Elsevier Ltd. All rights reserved.
The 3,6-epidithiopiperazine-2,5-dione (ETP) family of
natural products, formally derived by the incorporation
of a disulfide bridge into a cyclic dipeptide to give the
core structure (1), exhibit a wealth of structural diversity
and a fascinating range of biological activities (Fig. 1).1
The fungal metabolite gliotoxin (2), for example, has
been shown to cleave DNA in vitro and exhibits signi-
ficant antitumour, antiviral, antibacterial and im-
munosuppressive properties.1,2 Although the exact
mechanism by which this class of natural products oper-
ates has yet to be fully established, the bridgehead disul-
fide moiety, which exists in an atypical conformation
with significant torsional strain,3 is known to be essen-
tial for biological activity. Elegant studies by Waring
and Chai have confirmed the unusual reactivity of these
disulfides towards one electron reduction overall and
also established the equilibriumconstant for the gluta-
thione–gliotoxin pair formed by thiol disulfide inter-
change.1,4,5 In consequence, two mechanisms by which
gliotoxin induces apoptosis have thus far been pro-
posed, with both redox cycling to produce active oxygen
species, and protein interactions via thiol–disulfide inter-
change implicated.1,4
In light of the above studies on the nature and chemical
reactivity of the disulfide bridge, we therefore elected to
examine the possibility outlined in Scheme 1, for DNA
strand cleavage under anaerobic conditions such as the
oxygen deficient environment present in many types of
Glutathione
X = H
O
O
R1
XS
R2
R1
N
S
S
N
3
RSH
X = SR
N
N
O
R2
SH
R4
R3
R1
R2
+
O
O
O
O
+1e M SET
N
S
N
1
S
S
S
2+
H
N
eg X= M
OH
N
O
O
Base
R1
S
P O
Base
P O
H
XS
R2
O
OH
O
Strand
Cleavage
N
1
2
N
O
P
Figure 1.
O
P
O
P
O
Keywords: DNA cleavage; Thiyl radicals; Epidithiodiketopiperazines;
Polarity reversal catalysis; Mechanism.
6
4
5
*
Scheme 1. Proposed mechanism for DNA strand cleavage under
anaerobic conditions.
Corresponding author. Tel.: +44 020 7679 7533; fax: +44 020 7679
7524; e-mail: w.b.motherwell@ucl.ac.uk
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.03.022

2240
S. T. Hilton et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2239–2242
tumour cell.6 We reasoned that if the thiyl radical (4) de-
rived fromETP ( 1) via one of the indicated ring opening
sequences, were sufficiently electrophilic in character to
abstract the 40-hydrogen atomfromthe ribose ring ( 5),
then strand cleavage of DNA could be initiated via elim-
ination of phosphate from( 6) and subsequent nucleo-
philic capture of the resultant radical cation as
established in a beautiful series of mechanistic studies
both by Giese and co-workers7 and Crich and Mao.8
O
O
O
R2
Br
R2
Br
a
N
N
N
N
R1
R1
O
9 R1 = R2 = CH2Ph
10 R1 = Ph, R2 = CH3
O
11, 12
b
O
HS
R2
AcS
R1
R2
c
N
N
N
N
R1
SH
SAc
Although the reactivity of thiyl radicals of type (4) pos-
sessing such adjacent functionality was unknown, we
were particularly encouraged by an elegant study by
Roberts who demonstrated that sufficiently electrophilic
thiols such as triphenylsilanethiol and 1-thio-b-D-gluco-
pyranose tetraacetate could act as polarity reversal cata-
lysts to promote the radical chain racemisation of (R)-
tetrahydrofurfuryl acetate (7) via the planar radical (8)
(Scheme 2).9 The substrate tetrahydrofurfuryl acetate
can of course be considered as a simple model for the
ribose ring of DNA in this sequence. In a conceptually
similar vein, von Sonntag had also demonstrated that
thiyl radicals derived from1,4-dithiothreitol are capable
of effecting epimerisation of cis 2,5-dimethyltetra-
hydrofuran.10
O
15, 16
O
13, 14
Scheme 3. Reagents and conditions: (a) NBS, (PhCO)2O2, CCl4,
reflux; (b) KSAc, CH2Cl2, 0 °C—(13) 53% from 9; (14) 44% from 10;
(c) HCl, EtOH, reflux—(15) 68% from 13; (16) 94% from 14.
Ph3SiH
Ph3Si
O
O
O
O
O
HS
Ph
N
Ph
SH
17
18 70%
N
O
(15)
Scheme 4. Regents and conditions: Thiol (15) (2.5 mol %), Ph3SiH,
TBHN (tert-butyl hyponitrite) (10 mol %), dioxane 60 °C, 3 h, 70%.
The organic soluble cis dithiols (15 and 16) (Scheme 3)
required as thiyl radical precursors were accordingly
prepared fromthe corresponding diketopiperazines ( 9
and 10) using the classical methodology developed by
Trown and Kishi involving sequential radical bro-
mination with N-bromosuccinimide, followed by thioace-
tate displacement and acid catalysed thiol ester
hydrolysis.11,12
The successful obtention of the desired hydrosilylated
product, which requires the derived thiyl radical to ab-
stract a hydrogen atomfromtriphenylsilane as a key
propagation step was very encouraging and provided
strong presumptive evidence for the required electro-
philic character of the intermediate thiyl radicals derived
fromthe piperazine-2,5-dione 3,6-dithiol framework.
With the necessary cis dithiols in hand, we then elected
in the first instance to establish the probable intermedi-
acy of the derived thiyl radical through its involvement
once again, as a polarity reversal catalyst for the free
radical addition of triphenylsilane to the enol lactone
(17) (Scheme 4).9
In order to explore the potential reactivity of such thiyl
0
radicals for hydrogen atomabstraction fromthe 4 -posi-
tion of DNA under anaerobic conditions, we then direc-
ted our attention to the model study outlined in Scheme
5, which is of course based on the seminal work of Cai
and Roberts (vide supra Scheme 2).9 The para-chloro-
benzoate ester of (R)-tetrahydrofurfuryl alcohol (19),
readily prepared in essentially quantitative yield (94%
and 99% ee) by reaction of the alcohol with the acid
chloride in the presence of triethylamine, was selected
since the two enantiomers were readily separated on chi-
ral HPLC using a ChiralcelÓ OD column. Following the
protocol developed by Cai and Roberts,9 appropriate
control experiments using 5 mol % of both triphenyl-
silanethiol and ethylthioglycolate (vide infra) as the
H
O
AcO
XS
XSH
H
O
AcO
XSH
O
(R)-7
AcO
XS
(R)-7
8
AcO
O
H
(S)-7
Scheme 2.
O
O
H
O
H
O
O
O
O
Ar
O
HS
R1
R2
N
Cl
N
19
19
SH
15 R1 = R2 = CH2Ph 94% ee
16 R1 = Ph, R2 = CH3 91% ee
Scheme 5. Reagents and conditions: Thiol (2.5 mol %), TBHN (10 mol %), PhH, 60 °C, 3 h.

S. T. Hilton et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2239–2242
2241
catalysts and tert-butylhyponitrite (10 mol %) as initia-
O
O
a,b,c
d,e,f
tor in benzene at 60 °C for 3 h led to essentially complete
racemisation for the former (1.5% ee) and partial race-
misation (65% ee) for the latter, thereby confirming
the validity of the para-chlorobenzoate ester as a model
substrate.
Cl
HS
Cl
NEt2
24
27
O
O
H
H
Ph
N
Ph
N
OMe
OMe
O
O
SH
In the event however, extremely limited racemisation
was observed when either of the two organic soluble
cis dithiols (15 or 16) was employed as the racemisation
catalyst. Whilst it could be argued for the case of the
dibenzyl derivative (15) that intramolecular hydrogen
abstraction fromthe benzylic position by the thiyl radi-
cal might well be a competitive process, this is certainly
not feasible for (16). Since a possible explanation for the
lack of reactivity might involve thiyl radical stabilisation
via intramolecular hydrogen bonding as implied in Fig-
ure 2, we therefore decided to focus on the more imme-
diate electronic environment surrounding the thiol
through progressive study of a range of monothiol
derivatives.
28
26
Scheme 6. Reagents and conditions: (a) HNEt2, NEt3, CH2Cl2, 0 °C,
99%; (b) KSAc, CH2Cl2, 0 °C, 96%; (c) HCl, MeOH, 43%; (d) tBuOCl,
MeOH, NaOMe, 64%; (e) BF3ÆOEt2, CH2Cl2, AcSH, 59%; (f) NaOMe,
MeOH, 83%.
The results for a series of experiments using thiols (21–
26) as polarity reversal catalysts for racemisation of
the chiral ester (19) are shown in Table 1 and reveal that,
save for the case of the ester (25) where racemisation had
occurred to a significant extent (65%) all other mercap-
tans were virtually without effect. Fromthese observa-
tions it is clear that the presence of the second thiol
moiety in cis dithiols (15 and 16) did not have a detri-
mental effect on reactivity. The requirements for efficient
hydrogen atomabstraction from( 19) by a thiyl radical
XSÅ are clearly related to the electron withdrawing nat-
ure of X, probably because the S–H bond is stronger,
and the thiyl radical more electrophilic. The most strik-
ing observation in this respect can be seen through a
comparison of the simple acyclic mercapto ester (25)
with the corresponding diethyl amide (24) in which the
reduced electrophilic character of the amide carbonyl
group relative to the ester has effectively led to an inert
system. Given that the electron withdrawing nature of
the hemithioacetal unit in carbohydrate thiols such as
2,3,4,6-tetra-O-acetyl-1-b-D-glucopyranose contributes
to their efficiency as polarity reversal catalysts,9 it was
anticipated that the further inclusion of the additional
amidic nitrogen atom on the a-carbon of the mercaptan
would have contributed to increased reactivity in the
three a-mercapto piperazinediones (21, 22 and 23) and
in the acyclic model (26) when compared to the diethyl
amide (24). In the event however, even with incorpora-
tion of this additional electron withdrawing amide unit
a sufficiently reactive catalytic entity has not been pro-
duced. It should be noted however that 26, unlike 21–
23, is prone to elimination of hydrogen sulfide and hence
loss of potential activity.
Those compounds selected are shown in Figure 3, and
range fromthe three a-mercaptodiketopiperazines (21,
22 and 23) in which the basic heterocyclic core has been
retained to the acyclic thiols (24, 25 and 26) whose elec-
trophilic character varies as a function of the electron
withdrawing substituents attached to the carbon atom
bearing the thiol.
The synthesis of the monothiodiketopiperazines (21, 22
and 23) was carried out in a similar manner to that illus-
trated for the dithiodiketopiperazines (15 and 16) save
that only 1 equiv of N-bromosuccinimide was used in
the bromination step. The synthesis of the two mercap-
tans (24 and 26) is outlined below (Scheme 6).13,14
H
S
S
O
R
N
N
R
O
20
Figure 2. Possible hydrogen bonding stabilisation of the thiyl radical
(20).
O
O
O
O
Table 1. Reaction of the monothiols (Fig. 3) in the racemisation
studies
N
N
Ph
N
N
Ph
N
N
SH
SH
SH
SH
SH
Entry
Thiol
Ee (%)
O
O
O
1
2
3
4
5
6
21
22
23
24
25
26
96
98
98
93
65
95
21
22
23
NH
Ph
O
N
SH
MeO
O
24
O
O
25
26
Reagents and conditions: Thiol (5 mol %), TBHN (10 mol %), PhH,
60 °C, 3 h.
Figure 3.

2242
S. T. Hilton et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2239–2242
Recent studies by Waring and co-workers15 have sug-
gested that intracellular levels of gliotoxin exist almost
exclusively in the reduced formand that when glutathi-
one levels fall following induction of apoptotic cell
death, the oxidised disulfide is then released. In this
manner, the toxin can act further on neighbouring cells
in a pseudocatalytic way. The above studies have dem-
onstrated however, that under the anaerobic conditions
found in many types of tumour cell, the thiyl radicals
derived fromthe reduced formof epidithiodiketo-
piperazines are not sufficiently reactive per se in
intermolecular hydrogen atomabstraction fromthe
model ribose ester (19). However, as noted by Robins
in a model study of ribonucleotide reductase,16 signifi-
cant rate enhancements are observed in the intramolecu-
lar mode, even when the thiyl radical does not possess
additional electron withdrawing groups. For the ETP
family of natural products, the possibility therefore ex-
ists that one of the sulfur atoms could be involved either
in covalent modification through thiol disulfide inter-
change whilst the second thiol can then be available
for radical generation.
Med. Res. Rev. 1988, 8, 499–524; (d) Johnson, J. R.;
Bruce, F. W.; Dutcher, J. D. J. Am. Chem. Soc. 1943, 65,
2005–2009.
2. Eichner, R. D.; Waring, P.; Geue, A. M.; Braithwaite, A.
W.; Mullbacher, A. J. Biol. Chem. 1988, 263, 3772–3777.
¨
3. (a) Fridrichsons, J.; McL. Mathieson, A. Acta Crystallogr.
1967, 23, 439–448; (b) Fridrichsons, J.; McL. Mathieson,
A. Acta Crystallogr. 1965, 18, 1043–1052.
4. Chai, C. L. L.; Heath, G. A.; Huleatt, P. B.; OꢀShea, G. A.
J. Chem. Soc., Perkin Trans. 2 1999, 389–391.
5. Bernado, P. H.; Chai, C. L. L.; Deeble, G. J.; Liu, X.;
Waring, P. Bioorg. Med. Chem. Lett. 2001, 11, 483–485.
6. Rockwell, S. Oncol. Res. 1997, 9, 383–390.
7. (a) Glatthar, R.; Spichty, M.; Gugger, A.; Batra, R.;
Damm, W.; Mohr, M.; Zipse, H.; Giese, B. Tetrahedron
2000, 56, 4117–4128; (b) Giese, B.; Dussy, A.; Meggers, E.;
Petretta, M.; Schwitter, U. J. Am. Chem. Soc. 1997, 119,
11130–11131.
8. Crich, D.; Mao, X. J. Am. Chem. Soc. 1997, 119, 249–250.
9. (a) Cai, Y.; Roberts, B. P. Chem. Commun. 1998, 1145–
1146; (b) Haque, M. B.; Roberts, B. P.; Tocher, D. A.
J. Chem. Soc., Perkin Trans. 1 1998, 2881–2889; (c)
Roberts, B. P. Chem. Soc. Rev. 1999, 28, 25–35; (d) Cai,
Y.; Roberts, B. P.; Tocher, D. A. J. Chem. Soc., Perkin
Trans. 1 2002, 1376–1386.
10. Akhlad, M. S.; Schuchmann, H.; Von Sonntag, C. Int. J.
Radiat. Biol. 1987, 51, 91–102.
11. Trown, P. W. Biochem. Biophys. Res. Commun. 1968, 33,
402–407.
12. (a) Fukuyama, T.; Kishi, Y. J. Am. Chem. Soc. 1976, 98,
6723–6724; (b) Fukuyama, T.; Nakatsuka, S.; Kishi, Y.
Tetrahedron 1981, 37, 2045–2078.
13. Gebbink, R. J. M. K.; Klink, S. I.; Feiters, M. C.; Nolte,
R. J. M. Eur. J. Inorg. Chem. 2000, 253–264.
14. Miknis, G. F.; Williams, R. M. J. Am. Chem. Soc. 1993,
115, 536–547.
Acknowledgements
We are grateful to the BBSRC (Grant no. B16703) for
the provision of a postdoctoral fellowship (to S.T.H)
for this work, as well as to Professor Brian Roberts
for helpful discussions.
References and notes
15. Bernado, P. H.; Brasch, N.; Chai, C. L. L.; Waring, P.
J. Biol. Chem. 2003, 278, 46549–46555.
16. Robins, M. J.; Ewing, G. J. J. Am. Chem. Soc. 1999, 121,
5823–5824.
1. (a) Waring, P.; Beaver, J. Gen. Pharmacol. 1996, 27, 1311–
1316; (b) Chai, C. L. L.; Waring, P. Redox Rep. 2000, 5,
257–264; (c) Waring, P.; Eichner, R. D.; Mullbacher, A.
¨