The synthesis and conformational aspects of a novel dioxopiperazine - A possible β-turn constraint

Article abstract of DOI:10.1039/a904943f

The synthesis and conformational aspects of a dioxopiperazine was discussed. The cis configuration of the dioxopiperazine was rationalized using nuclear magnetic resonance (NMR) spectroscopy. Computational energy calculations were performed to explain the reluctance to cyclise of N-terminal partially protected dipeptides containing α, α-amino groups.

Full text of DOI:10.1039/a904943f

View Article Online / Journal Homepage / Table of Contents for this issue
The synthesis and conformational aspects of a novel
dioxopiperazine—a possible ꢀ-turn constraint
John S. Davies,*a Malgosia Stelmach-Diddams,a Regis Fromentin,a Alun Howellsa and
Ron Cottonb
1
a Department of Chemistry, University of Wales, Swansea, Singleton Park, Swansea,
UK SA2 8PP
b Zeneca Pharmaceuticals, Alderley Park, Macclesfield, Cheshire, UK SK10 4TG
Received (in Cambridge, UK) 21st June 1999, Accepted 1st November 1999
A synthetic route to a novel α,α-diamino β-keto ethoxycarbonyl-containing dioxopiperazine, capable of mimicking
a β-turn, is reported. The cis configuration of the dioxopiperazine has been rationalised by NMR spectroscopy,
while computational energy calculations have been used to explain the reluctance to cyclise of N-terminal partially
protected dipeptides containing α,α-amino groups.
Introduction
The biological activity of peptides and proteins is often
enshrined in a small loop of amino acid sequence which has
been generally designated a β-turn.1 This β-turn motif has
received a great deal of attention as the conformational basis
for designing pharmaceutically desirable peptidomimetics for
application as peptide receptor agonists and antagonists.2 The
constraining of peptides by cyclisation into cyclic penta-3 and
hexapeptides4 offers a means of maximising the β-turn con-
formations, although even at this level of constraint there are
many energy minimised conformations present. An alternative
means of constraining peptides for the development of peptido-
mimetics is to mimic a β-turn using a rigid building block,
which can influence the conformation of biologically active
motifs.5 Amongst the best known examples, (S)- and (R)-
Gly[ANC-2] 1,6 BTD 2,7 and (S,S)-spiro-Pro-Leu 38 have been
covered very quickly that we had focused on a rather demand-
ing combination of circumstances in that it contains β-keto and
α,α-diamino components. Precursors of the former would be
liable to facile decarboxylation while the α,α-diamino unit
would have to be orthogonally protected. The eventual route
that proved successful is outlined in Scheme 1. Whilst diester 7
known to stabilise β-turns in linear or cyclic peptides. However
when these are incorporated5 into cyclic pentapeptides, only in
the case of spiro-insert 3 was there evidence that the insert
appeared at the β-turn. In cyclohexapeptides the BTD insert 2
provided a stable and typical βIIЈ/βII arrangement. Our inter-
est in cyclopentapeptide analogues,9 led us to the molecular
modelling (A. Slater using ICI Viking programme) of the
minimum constraint requirement for a β-turn, which would
fit tightly at the turn without causing steric hindrance, and
preserve the polar backbone at the β-turn. The result turned
out to be a reverse cis-amide link in the form of a novel dioxo-
piperazine ring as portrayed in structure 4. This we saw as a
prototype for β-turn stabilisation, and as the nucleus of a tem-
plate to induce β-sheet structures, a topic which has seen a great
deal of activity of late.10 In order to produce a unit compatible
with contemporary protocols for peptide synthesis the synthon
5 became our target.
Scheme 1 Reagents and conditions: (i) DCC–HOBT; (ii) 1 eq. base; (iii)
DCC–HONSu; (iv) H2/Pd/C.
was commercially available, the diprotected diamino acid 6
proved more difficult to obtain.
gem-Diamino alkyl units have been explored for many years
in the synthesis of retro inverso peptide analogues,11 but only
two of the published methods proved fruitful in our hands. The
synthetic route based on the work of Bock et. al.12 (Scheme 2)
and synthesis via our adaptation of a procedure by Waki et al.13
(Scheme 3) were the only synthetic sequences to give significant
Results and discussion
In attempting the synthesis of dioxopiperazine 5, it was dis-
J. Chem. Soc., Perkin Trans. 1, 2000, 239–243
This journal is © The Royal Society of Chemistry 2000
239

View Article Online
Scheme 2 Reagents and conditions: (i) Me2CHSH–AcOH–H2SO4; (ii)
Boc-NH2–HgCl2.
Fig. 1
Scheme 3 Reagents and conditions: (i) DCC–HONSu–NH3; (ii) TIB,
pyridine, t-BuOH; (iii) base.
yields of 6. The coupling of 6 and 7 was routine and best
accomplished by DCC–HOBT to give dipeptide 8, which was
then subjected to various conversions in order to synthesise
dioxopiperazine 5. The literature is full of facile syntheses of
dioxopiperazines from dipeptide esters, many being formed far
too readily14 as side products. Yet, when the Z-amino protect-
ing group was removed from 8 by catalytic hydrogenation over
palladium/charcoal, and the free amino diester was subjected to
typical conditions for cyclisation to dioxopiperazine (refluxing
in ethanol for periods up to 96 hours, or hydrogenation at
elevated pressures), there was no evidence that 5 had been
formed. For a similar amino dipeptide diethyl ester, but without
an α,α-diamino combination these conditions sufficed.15 To test
that the cyclisation was not impeded by the bulkiness of the
Boc-amino group, the alternative route was also tested, by
removing the Boc group in 8 with trifluoroacetic acid. After
neutralisation with DIEA (diisopropylethylamine), stirring at
room temperature for 24 h, followed by refluxing in ethanol
overnight, no tendency to cyclise was detected.
Eventually, a successful synthetic route was achieved by care-
fully hydrolysing 8 to a mono ester 9, followed by the activation
of the carboxy group as its N-hydroxysuccinimide ester 10.
Removal of the Z-group protection in 10, using H2/Pd/C in the
presence of base, led to the spontaneous cyclisation to the
required synthon 5. A side reaction product (in 13% yield) dur-
ing a cyclisation attempt was identified as the decarboxylated
analogue 11. So, although the tendency for the dioxopiperazine
Fig. 2
between the RR or SS forms) was characterised as depicted in
Fig. 1. The characterisation was made on the basis of the
following NMR data and arguments. The dioxopiperazine 5
1
showed the following signals in its H NMR spectrum. The
numbering refers to atoms listed in Fig. 1: δH (250 MHz; d6-
DMSO; Me4Si) 1.2 (3H, t, CH2CH3), 1.40 [9H, s, C(CH3)3], 4.2
(2H, q, CH2CH3), 4.65 [1H, s, CH(4)], 5.22 [1H, d, J 11, CH(2)],
7.25 [1H, d, J 11, NH(5)], 8.5 [1H, s, NH(1)] and 8.80 [1H, s,
NH(3)]. The boat form was assumed as very few chair forms
of dioxopiperazines have been found to exist in the solution
phase.16 In CDCl3 solution at room temperature, two separate
signals for both H(2) and H(4) were seen, giving evidence of
two conformers, since on heating to 60 ЊC, the separate signals
for both H(2) and H(4) almost coalesced. Both signals under-
went simultaneous upfield shifts, indicative of increased shield-
ing, due to ring inversion. This increased shielding can occur in
two situations: (a) if the ring is in the chair conformation and
the substituents are trans; (b) the conformation is in the boat
form and the substituents are cis. However, further evidence
for the boat conformation comes from the H-coupling con-
stants in CDCl3 solution: δ H(4) 4.70, JH(3)–H(4) = 4 Hz, H(2)
5.52, JH(1)–H(2) = 5.6 Hz, H(5) 6.08, JH(5)–H(2) = 8 Hz.
If our configurational deductions for the cis-dioxopiperazine
5 are correct, then this cis-form has preferentially formed from
a dipeptide which contains four diastereoisomers. There is pre-
cedence in the literature17 where self-condensation of -amino
acid esters preferentially formed cis-dioxopiperazines. This was
attributed to the rates of cyclisation of two kinds of precursor
diastereoisomeric dipeptide esters, and showed that pre-cis-
peptide esters cyclised faster than pre-trans in methanol solu-
tion. Rationalising our own results in this way we would have
to argue that in the representations in Fig. 2 the pre-cis is
more energetically favoured than the pre-trans form. So this
selectivity, and why the α,α-diamino precursor dipeptide was
so reluctant to cyclise, led us to our computational energy cal-
culations using a model system designed to compare the N-
terminal α,α-diamino system with a conventional sterically
hindered N-terminal valyl unit. A C-terminal seryl residue was
used, as this could be later used in an alternative synthesis of
analogues of 5. The HOBT active ester was used for com-
parison with a C-terminal alkyl ester.
and some of the linear mono ester precursors to decarboxylate
was troublesome, the biggest surprise was the lack of cyclis-
ation between the released α-amino group with the esters at the
C-terminal positions. This could be due either to the lack of
nucleophilicity of the amino group or the lack of reactivity of
β-keto esters. The latter seemed unlikely as there was prece-
dence in the literature,15 supporting their adequacy as dioxo-
piperazine precursors. We therefore initiated empirical energy
calculations as described later in this paper to try and clarify the
position.
It was very important to the applicability of the proposed
insert 5 as a constraining unit, that the side chain protected
amino and carboxy groups were pointing in the same direc-
tion—i.e. in a cis disubstituted ring. The synthesis of 5 had
been carried out from a dipeptide precursor 10 which was
racemic at both chiral centres. So four diastereoisomeric
dioxopiperazines [RR and SS (cis-forms) or the RS and SR
(trans-forms)] could emanate from such a system. All the
dioxopiperazine material isolated from the synthesis of 5 was
therefore analysed by NMR spectroscopy, and only one form
which we believe is the cis-form (we cannot differentiate
Computational energy calculations
Molecular orbital calculations were carried out on empirical
structures for molecules, using the AM118 method of the
MOPAC 93 program.19 The preliminary calculations of heat of
formation and formulae for the representative molecules calcu-
lated, appear in Fig. 3. All calculations were carried out assum-
ing the gas phase, and these heat of formation figures were
inserted into eqn. (1) to calculate the nett energy change due to
240
J. Chem. Soc., Perkin Trans. 1, 2000, 239–243

View Article Online
Table 1 Nett energy change due to cyclisation
Conformation
after
cyclisation
Calculated atomic charges
Energy/
Chirality
kcal molϪ1
1
2
3
4
5
6
A
B








trans
cis
trans
cis
trans
cis
trans
cis
Ϫ0.980
Ϫ0.343
Ϫ0.344
Ϫ0.338
Ϫ0.345
Ϫ0.361
Ϫ0.363
Ϫ0.362
Ϫ0.367
Ϫ0.030
Ϫ0.031
Ϫ0.033
Ϫ0.030
0.107
0.105
0.109
0.259
0.262
0.274
0.264
0.310
0.307
0.308
0.311
0.289
0.272
0.291
0.288
0.281
0.277
0.292
0.290
Ϫ0.347
Ϫ0.347
Ϫ0.292
Ϫ0.295
Ϫ0.366
Ϫ0.365
Ϫ0.297
Ϫ0.294
Ϫ0.276
Ϫ0.286
Ϫ0.174
Ϫ0.171
Ϫ0.254
Ϫ0.253
Ϫ0.166
Ϫ0.179
0.270
Ϫ8.890
Ϫ7.500
3.780
4.170
Ϫ7.110
Ϫ6.570
C
D
0.104
explain the reduced reactivity of the α-amino group. At the
present time the model is not refined enough to distinguish for
us the configurational selectivity which was experienced in the
synthesis, but this might be more of a kinetic rather than a
thermodynamic effect.
To further check whether the difficulties in cyclisation were
due to the Boc-amino group substituent, we attempted to syn-
thesise cyclo-(Val-Ser)20 and cyclo-(Boc-amino α-Gly-Ser),
under similar conditions. Under conditions20 of refluxing in
methanol over 5 days only the former dioxopiperazine could be
produced.
Experimental
1H NMR spectra were recorded, either at 250 MHz on a Bruker
WM 250 spectrometer, or at 400 MHz on a Bruker AC-400 FT
spectrometer, using tetramethylsilane as internal standard.
Chemical shifts are given in ppm; J values are given in Hz.
Abbreviations used, s = singlet, d = doublet, t = triplet, q =
quartet, m = multiplet. Mass spectra were determined both at
the EPSRC Centre, Swansea, and at Zeneca Pharmaceuticals,
on a VG 12 - 253 for low resolution EI/CI, and for high reso-
lution (or FAB) a VG Autospec mass spectrometer using Csϩ
ions for bombardment on samples dissolved either in a thio-
glycerol or NOBA (3-nitrobenzyl alcohol). The numbers in
brackets after m/z values represent the intensity of the peak
relative to the base peak. CHN Analyses were carried out on
either a Carlo Erba 1106 or a Perkin Elmer 2400 CHN instru-
ment. Mp’s were determined on a Gallenkamp hot stage appar-
atus and are uncorrected.
Heat of
formation
of
Heat of
formation
of DKP’s
etc./kcal
molϪ1
Chirality
of
dipeptide
dipeptides/
kcal molϪ1
A
B








Ϫ192.66
Ϫ193.54
Ϫ21.69
Ϫ22.71
Ϫ266.82
Ϫ266.77
92.87
CH3OH
HOBt
DKP trans (isopropyl)
DKP cis (isopropyl)
Ϫ57.05
106.01
Ϫ136.59
Ϫ136.22
C
D
DKP trans (Boc-amino) Ϫ205.99
DKP cis (Boc-amino)
Ϫ205.55
92.97
Column chromatography was routinely carried out on silica
gel 60 (Merck) (in a few cases neutral silica gel was used) which
was prepared as follows: silica gel 60 (Merck) (355 g) was
stirred in 0.1 M sodium hydroxide solution for 30 min. After
removal of base the silica was re-dispersed in buffer solution
(pH 7, 1 L), followed by filtration and washing with water (3 L),
and finally with pure methanol (1 L). After filtration, the silica
was partially air dried, before being dried in an air oven at
120 ЊC for 16 h.
Fig. 3
Nett energy (kcal molϪ1) =
(Hdioxopiperazine ϩ HX–OH) Ϫ (Hdipeptide
)
(1)
cyclisation, where H = Heat of formation, X = CH3 or benzo-
triazole. The results appear in Table 1. If the energy value is
positive, the reaction is deemed to be thermodynamically
impossible.
Synthesis of 2-(N-benzyloxycarbonylamino)-2-(N-tert-butoxy-
The results show that the use of an active ester (reactions A
and C compared with B and D) is very important in nett energy
changes, although it is not manifested in a great increase in the
electrophilicity of the ester carbonyl carbon (4), but electron
densities at oxygens (5) and (6) do show change. The result
confirms the difficulties we have had in cyclisation without the
use of an active ester. The effect of the Boc-amino group com-
pared to the isopropyl group at carbon (2) has increased the
electrophilicity of carbons (2) and (3), so the increase in the
nucleophilicity on the N(1) α-amino group is therefore a little
surprising. However, if the difference in charge between atoms
(1) and (2) are calculated the average difference in the atoms
bearing the isopropyl group is Ϫ0.373, but is changed to an
average value of Ϫ0.257 for the two examples (C and D)
bearing the Boc-amino substituent. This difference could well
carbonylamino)ethanoic acid 6
Method A (based on the method of Bock et al.12). A suspen-
sion of N-benzyloxycarbonyl-2-hydroxyglycine21 (29.0 g, 0.129
mol) and propane-2-thiol (39.3 g, 0.516 mol) in glacial acetic
acid (100 mL) was cooled in an ice bath under stirring, followed
by the addition of conc. sulfuric acid (15 mL). After stirring the
mixture at ambient temperature for 2 days, it was poured into
ice and the aqueous solution was extracted with ethyl acetate.
Extraction of the organic phase with 5% sodium hydrogen
carbonate, followed by acidification of the aqueous layer and
re-extraction of the product into ethyl acetate, gave on work up
and trituration with petroleum ether (40–60 ЊC), N-benzyloxy-
carbonyl-2-isopropylthioglycine (28.5 g, 78%), as a white
crystalline solid, mp 82–84 ЊC (lit.12 82–84 ЊC); δH (250 MHz;
J. Chem. Soc., Perkin Trans. 1, 2000, 239–243
241

View Article Online
d6-DMSO) 1.19 and 1.23 [6 H, 2d, J 7.0, (CH3)2CH], 3.09
[1 H, m, (CH3)2CH], 5.05 [1 H, d, NHCH, (partially obscured
by a singlet at 5.05)], 5.05 (2 H, s, C6H5CH2), 7.35 (5 H, s,
C6H5CH2), 7.35 (1 H, d, J 8.5, NHCH); m/z CI (NH3): 284 (7%)
(M ϩ H)ϩ, 210 (4) [M ϩ H Ϫ (CH3)2CS]ϩ, 193(2) (M Ϫ C7H7)ϩ,
152(46) (NH3COOCH2Ph)ϩ.
mixture of ethanol–water, homogenised by vigorous stirring or
using an ultrasonic bath, was hydrolysed by adding an equimo-
lar amount of 1 M potassium hydroxide, over 30 min–1 h, and
by stirring at ambient temperature for 3 h. Work up procedures
as described under method A gave 4.2 g (95%) of acid 6, with
analytical data identical to the product of method A.
N-Benzyloxycarbonyl-2-isopropylthioglycine (4.1 g, 14.5
mmol), tert-butylcarbamate (7.3 g, 62.3 mmol) and mercury()
chloride (5.9 g, 21.7 mmol) in tetrahydrofuran (75 mL) were
heated at 60 ЊC under nitrogen for 20 h. The cooled mixture was
diluted with diethyl ether (150 mL) and filtered. The filtrate was
further diluted with diethyl ether (200 mL), and the organic
phase extracted with 10% sodium hydroxide solution. Work up
after acidification of the alkaline aqueous layer, gave a colour-
less gum which on trituration with diethyl ether, and recrystal-
lisation from the same solvent gave 2-Boc-amino-2-Z-amino-
ethanoic acid as a white crystalline compound (1.17 g, 25%),
mp 158–159 ЊC (decomp.) (lit.12 158–159 ЊC decomp.) (Found:
C, 55.7; H, 6.2; N, 8.6. C15H20N2O6 requires: C, 55.55; H, 6.2; N,
8.6%); δH (100 MHz; d6-DMSO) 1.43 [9H, s, (CH3)3C], 5.08
(2H, s, C6H5CH2), 5.35 (1H, t, J 8, NHCH), 7.3 (1H, br d,
partially obscured, NHCH), 7.38 (5H, s, C6H5CH2), 7.8 (1H, br
d, J 8, NHCH); m/z (FAB) 347 (16.5%) (M ϩ Na)ϩ, 325 (7.6)
(M ϩ H)ϩ, 269 (35) (M ϩ H Ϫ C4H8)ϩ, 152 (9.1), 91 (100)
(C7H7)ϩ. A major by-product in the mother liquor was identi-
fied as ZNHCH(OH)COOH.
Diethyl [2-(benzyloxycarbonylamino)-2-(tert-butoxycarbonyl-
amino)ethanoylamino]malonate, 8
Acid 6 (4.2 g, 13.0 mmol) and diethyl aminomalonate 7 (4.1 g,
19.4 mmol) were dissolved in a minimum volume of dry DMF,
and cooled in an ice-bath. HOBT (2.56 g, 19 mmol) was added
followed by dropwise addition of triethylamine until the pH of
the mixture reached 7–7.5 (monitored with moist pH paper).
DCC (2.81 g, 13.65 mmol) was added, and soon precipitation of
dicyclohexylurea occurred. After 12 h stirring at approx 0 ЊC,
the precipitate was removed by filtration, and the filtrate rotary
evaporated. The gummy residue was dissolved in a minimum
volume of hot ethyl acetate, and left to cool to precipitate more
of the urea. After further filtration, the filtrate was diluted with
more ethyl acetate and washed with 10% citric acid solution and
sodium bicarbonate as described earlier. Two neutral com-
ponents detected by TLC, in the dried ethyl acetate layer, were
purified by silica gel chromatography. Fraction eluted in 96:4
(v:v) chloroform–methanol was crystallised from ethyl acetate–
light petroleum as dipeptide diester 8 (9.39 g, 75%) mp 114 ЊC,
δH (100 MHz; d6-DMSO) 1.21 (6 H, t, J 7, CH3CH2), 1.40 [9 H,
s, (CH3)3C], 4.19 (4 H, q, J 7, CH3CH2), 5.03 (1 H, d, J 7,
NHCH), 5.09 (2 H, s, C6H5CH2), 5.48 [1 H, t, J 8.4, (-NH)2-
CH], 7.16 and 7.60 [2 H, br d, J 8.4, (-NH)2CH], 8.65 (1 H, d,
J 7, NHCH); m/z (FAB) 504 (1.5%) (M ϩ Na)ϩ, 482 (2.5)
(M ϩ H)ϩ, 426 (6) (482 Ϫ C4H8)ϩ, 365 (7) (482 Ϫ BocNH2)ϩ,
321 (14) (365 Ϫ CO2)ϩ, 231 (37) (321 ϩ H Ϫ C7H7)ϩ, 91 (100)
(C7H7)ϩ. A by-product (0.5 g, 10%) eluted from the column with
92:8 (v:v) chloroform–methanol was identified as ZNHCH-
(OH)CONHCH(CO2Et)2.
Method B (adapted from the work of M. Waki et al.13). N-
Benzyloxycarbonyl-2-amino-monoethyl malonate22 (10.0 g, 35
mmol), N-hydroxysuccinimide (6.13 g, 53.2 mmol) in dry tetra-
hydrofuran was cooled to 0 ЊC. DCC (8.06 g, 39.05 mmol) was
added, and the reaction mixture was stirred vigorously at 0 ЊC
for 1 h, when dry tetrahydrofuran (100 mL) saturated with gas-
eous ammonia was added via a syringe needle through a sep-
tum. After one more hour of stirring at 0 ЊC, the precipitated
dicyclohexylurea and hydroxysuccinimide were removed by
filtration and the filtrate, after rotary evaporation, was re-
dissolved in hot ethyl acetate (100 mL). After extraction of the
ethyl acetate layer with 10% citric acid followed by sodium
bicarbonate solution and water, the dried solvent layer on
rotary evaporation began to precipitate out N-benzyloxy-
carbonyl-2-carboxamidoglycine (9.45 g, 95%), mp 112–114 ЊC
(from hot isopropanol) (Found: C, 55.8; H, 5.7; N, 9.85.
C13H16N2O5 requires: C, 55.7; H, 5.75; N, 10.0%); δH (60 MHz;
d6-DMSO) 1.07 (3 H, t, J 7, CH3CH2), 3.95 (2 H, q, J 7,
CH3CH2), 4.63 (1 H, d, J 8, NHCH), 4.87 (2 H, s, C6H5CH2),
7.13 (5H, s, C6H5CH2), 7.20–7.70 (3 H, m, NHCH and
CONH2); m/z (EI) 280 (2.4%) Mϩ, 263 (0.5) (M Ϫ NH3)ϩ, 237
(2.4) (M Ϫ CONH)ϩ, 91 (100) (C7H7)ϩ.
Monoethyl 2-[2-(benzyloxycarbonylamino)-2-(tert-butoxy-
carbonylamino)ethanoylamino]malonate, 9
Dipeptide diester 8 (4.6 g, 9.55 mmol) was hydrolysed with 1 M
potassium hydroxide under the general conditions described for
making 6 in method B above, giving monoethyl dipeptide ester 9
(4.24 g, 98%), as a white crystalline compound, mp 99 ЊC (from
ethanol–light petroleum) (Found: C, 51.1; H, 6.0; N, 8.8.
C20H27N3O9ؒH2O requires: C, 51.1; H, 6.2; N, 8.9%); δH (250
MHz; d6-DMSO) 1.19 (3 H, t, J 7, CH3CH2), 1.39 [9 H, s,
(CH3)3C], 4.15 (2 H, q, J 7, CH3CH2), 4.88 (1 H, d, J 6.7,
NHCH), 5.06 (2 H, s, C6H5CH2), 5.46 [1 H, t, J 8.1, (-NH)2CH],
7.32 and 7.70 [2 H, 2br s, (-NH)2CH], 8.40 (1 H, br s, NHCH);
m/z (FAB): 929 (1.2%) (2 × M ϩ Na)ϩ, 476 (12.4) (M ϩ Na)ϩ,
454 (8.6) (M ϩ H)ϩ, 398 (18.1) (454 Ϫ C4H8)ϩ, 337 (7.7)
(454 Ϫ BocNH2)ϩ, 203 (38.5) (337 ϩ H Ϫ C7H7OCO)ϩ, 91
(100) (C7H7)ϩ.
N-Benzyloxycarbonyl-2-carboxamidoglycine (5.0 g, 17.8
mmol) and TIB [I,I,-bis(trifluoro-acetoxy)iodo]benzene (11.48
g, 26.7 mmol), in dry 2-methylpropan-2-ol (150 mL) were
refluxed for 15 min, and then pyridine (4.22 g, 53.4 mmol) was
added, the reaction being complete in 1 h. After removal of
solvent by rotary evaporation, the residue was re-dissolved in
chloroform, and the insoluble pyridinium trifluoroacetate was
removed by filtration. Work up of the organic layer using the
usual acid–base extractions described before, yielded on rotary
evaporation, a colourless oil (5.21 g, 83%), purified by column
chromatography with chloroform as eluent, to give ethyl 2-
N-benzyloxycarbonylamino-2-tert-butyloxycarbonylaminoethan-
oate, mp 122 ЊC (chloroform–light petroleum) (Found: C, 58.1;
H, 7.0; N, 7.9. C17H24N2O6 requires: C, 57.9; H, 6.9; N, 7.95%);
δH (250 MHz; d6-DMSO) 1.17 (3 H, t, J 7, CH3CH2), 1.39 [9 H,
s, (CH3)3C], 4.11 (2 H, q, CH3CH2), 5.06 (2 H, s, C6H5CH2),
5.32 [1 H, t, J 7.5, (-NH)2CH], 7.36 (5 H, s, C6H5CH2), 7.60
(1 H, d, J 7, NHCH), 8.02 (1 H, d, J 7.7, CHNH); m/z (CI)
353 (10%) (M ϩ H)ϩ; 297 (57) (M ϩ H Ϫ C4H8)ϩ, 253 (61)
Ethyl 5-(N-tert-butoxycarbonylamino)-3,6-dioxopiperazine-2-
carboxylate 5, via ethyl succinimido-2-[2-(benzyloxycarbonyl-
amino)-2-(tert-butoxycarbonylamino)ethanoylamino]malonate
10
Monoethyl dipeptide ester 9 (1.0 g, 2.2 mmol) and N-hydroxy-
succinimide (0.38 g, 3.3 mmol) (previously dried over KOH at
80 ЊC for 10 h), were dissolved in dry DMF (20 mL) and cooled
in an ice bath with stirring. DCC (0.477 g, 2.31 mmol) was
added and within 5 min dicyclohexylurea started precipitating,
and the reaction was deemed complete within 30 min. The pre-
cipitate was removed by filtration, the DMF removed by rotary
evaporation. (In trial runs the reaction was worked up at this
stage by re-dissolution of the residue in ice-cold ethyl acetate.
Any further precipitated urea was removed before subjecting a
(297 Ϫ CO )ϩ, 236 (10) (253 Ϫ NH )ϩ, 102 (100) (NH ᎐CH-
᎐
2
3
2
CO2Et)ϩ. This ethyl ester (4.65 g, 13.2 mmol) in a 7:3 (v:v)
242
J. Chem. Soc., Perkin Trans. 1, 2000, 239–243

View Article Online
concentrated solution to chromatography on silica gel. The
main fractions containing N-hydroxysuccinimido ester 10 (0.72
g, 60%), were eluted with a solvent mixture of chloroform–(4–
10%) methanol. However over 30% of the yield was lost by
reaction of the active ester with methanol at the chromato-
graphy stage, so in subsequent reactions the active ester was not
isolated.) In the ‘non-isolated’ cases the DMF solution of the
active ester 10, was diluted to 200 mL, and 20% Pd/C (100 mg)
added. Hydrogen gas was purged through under stirring for
24 h, and the catalyst removed by repeated filtration through
thick filter paper. The filtrate was rotary evaporated at <35 ЊC,
and the gummy residue re-dissolved in ethyl acetate for sub-
sequent extraction of acid and base material. The neutral frac-
tion remaining appeared to contain a number of components
(by TLC), which were separated by column chromatography on
silica gel. Dioxopiperazine 5 was eluted in the later fractions
(97:3, v:v chloroform–methanol), which after evaporation gave
a colourless oil (0.26 g, 40%), which crystallised on trituration
with ether. Mp 188–190 ЊC (from chloroform) (Found: C, 46.4;
H, 6.1; N, 13.0. C12H19N3O6ؒ¹ H2O requires: C, 46.45; H, 6.5;
N, 13.54%); δH (400 MHz; d6-DMSO) details appear in the dis-
cussion; δH (400 MHz, CDCl3; ambient temp; atoms as num-
bered in Fig. 1) 1.3 (3 H, m, CH3CH2), 1.48 [9 H, s, (CH3)3C],
4.38 (2 H, m CH3CH2), 4.69 and 4.70 (1H, 2d, J3,4 3.5, J3,4 4.0,
H(4) conformers), 5.52 and 5.55 [1H, d and dd, J1,2 or J2,5 5.6,
J5,2 8, J1,2 3 H(2) conformers], 5.77 and 6.08 [1 H, br s and d, J5,2
8, H(5)], 6.49 and 6.52 [1 H, br s, H(3)], 6.58 and 5.68 [1 H, br s,
H(1)]. At 60 ЊC in CDCl3, H(4) gave a broad singlet at 4.66, and
H(2) a broad singlet at 5.52 and doublet at 5.50.
ethyl acetate–light petroleum) gave a white product, mp 149–
151 ЊC (62% yield); δH (400 MHz d6-DMSO) 1.4 [9 H, s,
(CH3)3C], 3.6–3.8 (5 H, s and d, OCH3 and CH2OH), 4.3–4.4
(1 H, br s, CHCH2OH), 5.05 (2 H, s, CH2O), 5.3–5.4 (1 H, t,
NHCHNH), 7.0 (1 H, br s, CHCONH), 7.3–7.4 (5 H, m,
C6H5), 7.5–7.6 (1 H, br s, OCONH), 7.9 (1 H, br s, OCONH);
δC (d6-DMSO) 28.1 [C(CH3)], 51.8 and 54.9 (2 × CH, and
OCH3), 61.1 (CH2OH), 65.7 (CH2O), [(CH3)3C], 128 and 136.6
(C6H5), 167.68 and 170.3 (4 × CO). Product of hydrolysis of
this ester had m/z (electrospray): 434 (M ϩ Na)ϩ. C18H25N3O8ؒ
Na requires 434.
Attempted cyclisation. When the above dipeptide ester was
refluxed after hydrogenation under exactly the same conditions
as for the valyl analogues, no dioxopiperazine formed.
Acknowledgements
We thank the EPSRC for a CASE Award with Zeneca Pharm-
aceuticals for M. S.-D., and a Quota Studentship for A. H., and
the ERASMUS scheme for a studentship for R. F.
¯
²
References
1 G. Müller, Angew. Chem., Int. Ed. Engl., 1996, 35, 2767; G. D. Rose,
L. M. Gierasch and J. A. Smith, in Advances in Protein Chemistry,
eds. C. B. Anfinsen, J. T. Edsall and F. M. Richards, Academic Press.
Florida, 1985, vol. 37, pp. 1–109.
2 R. Hirschmann, in Peptides 1996, eds. R. Ramage and R. Epton,
Mayflower Scientific Ltd., Kingswinford, UK, 1998, p. 3; J. B. Ball
and P. F. Alewood, J. Mol. Recognit., 1990, 3, 55; G. Holzemann,
Kontakte (Darmstadt), 1991, pp. 3 and 55; M. Kahn, Synlett, 1993,
821; A. Giannis and T. Kolter, Angew. Chem., Int. Ed. Engl., 1993,
32, 1244; R. M. Liskamp, J. Recl. Trav. Chim. Pays-Bas, 1994, 113,
1; S. Hanessian, G. McNaughton-Smith, H.-G. Lombart and W. D.
Lubell, Tetrahedron, 1997, 53, 12789.
A by-product (13% yield) of some of the attempts to syn-
thesise
5 was identified as 3-tert-butyloxycarbonylamino-
piperazin-2,5-dione (11), a white crystalline solid, mp 93–101 ЊC
(lowered by traces of N-hydroxysuccinimide); δH (400 MHz;
d6-DMSO) 1.39 [9 H, s, (CH3)3C], 3.6 (1 H, dd, CH-H), 3.89
(1 H, d, CH-H), 4.92 (1 H, dd, CH-NHBoc), 7.96 (1 H, d, NH),
8.12 (1 H, s, NH), 8.53 (1 H, s, NH); 2D COSY spectrum
showed CH at 4.92 coupled to both NH ’s at δ 7.96 and 8.53, the
CH at 3.60 coupled to both NH at 8.12 and CH at δ 3.89. 13C
DEPT 135 confirmed a CH2 signal at δ 44.26. m/z (CI): 246
(M ϩ NH4)ϩ.
3 R. Haubner, D. Finsinger and H. Kessler, Angew. Chem., Int. Ed.
Engl., 1997, 36, 1374.
4 K. D. Kopple, J. W. Bean, K. K. Bhandry, J. Briand, C. A.
D’Ambrosia and C. E. Peishoff, Biopolymers, 1993, 33, 1093.
5 R. Haubner, W. Schmitt, G. Hölzemann, S. L. Goodman, A.
Jonczyk and H. Kessler, J. Am. Chem. Soc., 1996, 118, 7881.
6 R. M. Friedinger, D. S. Perlow and D. F. Veber, J. Org. Chem., 1982,
47, 104.
7 U. Nagai and K. Sato, Tetrahedron Lett., 1985, 26, 647.
8 M. G. Hinds, N. G. Richards and J. A. Robinson, J. Chem. Soc.,
Chem. Commun., 1988, 1447.
9 J. S. Davies, J. Howe, J. Jayatilake and A. M. Riley, Lett. Pept. Sci.,
1997, 4, 441; J. S. Davies, C. Enjalbal, C. J. Wise, S. E. Webb and
G. E. Jones, J. Chem. Soc., Perkin Trans. 1, 1994, 2011.
10 J. S. Nowick, E. M. Smith and M. Pairish, Chem. Soc. Rev., 1996,
401.
Cyclo(Val-L-Ser)20 and Cyclo(Val-D-Ser)20
Comparison of the effect of isopropyl vis à vis Boc-amino side
chains on the cyclisation yields in dioxopiperazine formation
was checked under the known conditions20 that had proved
successful for the isopropyl case: Z-Val-Ser-OMe22 or Z-Val--
Ser-OMe20 (0.103 g, 0.3 mmol) were independently suspended
in absolute methanol (10 mL) with 10% Pd/C (10 mg), and
stirred under hydrogen for 15 h. After removal of catalyst and
evaporation of solvent, the crude residues were suspended in
absolute methanol (10 mL) and refluxed for 110 h. Removal of
solvent and work up gave the cyclo(Val-Ser), mp 246–247 ЊC
(lit.,20 245–249 ЊC), and cyclo(Val-D-Ser), mp 224–225 ЊC (lit.,20
230–232 ЊC), with NMR data identical to published data. For
the analogue with the Boc-amino side chain:
11 M. Chorev and M. Goodman, Int. J. Pept. Protein Res., 1983, 21,
258.
12 M. G. Bock, R. M. Dipardo and R. M. Friedinger, J. Org. Chem.,
1986, 51, 3718.
13 M. Waki, Y. Kitajima and N. Izumiya, Synthesis, 1981, 266.
14 P. M. Fischer, M. Solbakken and K. Undheim, Tetrahedron, 1994,
50, 2277.
15 H. E. Zaugg. M. Freifelder, H. J. Glenn, B. W. Horrom, G. R. Stone
and M. R. Vernsten, J. Am. Chem. Soc., 1956, 78, 2626.
16 M. J. O. Anteunis, Bull. Soc. Chim. Belg., 1978, 87, 627.
17 H. Naraoka and K. Harada., J. Chem. Soc., Perkin Trans. 1, 1986,
1557.
18 M. J. S. Dewar, E. G. Zoebish, E. F. Healy and J. J. P. Stewart, J. Am.
Chem. Soc., 1985, 107, 3902.
19 MOPAC 93, J. J. P. Stewart and Fujitsu Ltd., Tokyo, Japan copyright
(Fujitsu Ltd., 1993), obtained from QCPE Department of
Chemistry, Indiana University, Bloomington, Indiana 47 405, USA.
20 M. Falorni, M. Satta, S. Conti and G. Giacomelli, Tetrahedron:
Asymmetry, 1993, 4, 2389.
21 U. Zoller and D. Ben Ishai, Tetrahedron, 1975, 31, 863.
22 T. Iwasaki, H. Horikawa, K. Matsumoto and M Miyoshi, J. Org.
Chem., 1977, 42, 2419.
(2-N-Benzyloxycarbonylamino-2-N-tert-butoxycarbonyl)eth-
anoyl-L-serine methyl ester. 2-Boc-amino-2-Z-aminoethanoic
acid 6 (1.04 g, 3.22 mmol) and HOBT (0.493 g, 3.22 mmol) in
DMF (20 mL) were cooled in an ice bath, before being treated
with -serine methyl ester hydrochloride (0.499 g, 3.22 mmol),
triethylamine (0.9 mL, 6.44 mmol) and DCC (0.663 g, 3.22
mmol), in that order, and the total mixture was stirred at room
temperature for 24 h. After evaporation of DMF, dicyclo-
hexylurea and HOBT precipitated out of cold acetone and
ethyl acetate respectively. Combined filtrates in ethyl acetate
were washed successively with 10% citric acid, 0.1 M NaHCO3
and water, and after drying, the solvent was evaporated to a
crude residue. Chromatography on silica gel (with gradient of
Paper a904943f
J. Chem. Soc., Perkin Trans. 1, 2000, 239–243
243