Synthesis and solution conformation of β-hairpin mimetics utilizing a template derived from (2S,3R,4R)-diaminoproline

Article abstract of DOI:10.1002/(SICI)1522-2675(20000216)83:2<444::AID-HLCA444>3.0.CO;2-R

A novel template was synthesized for stabilizing β-hairpin conformations in cyclic peptide mimetics. The template is a diketopiperazine derived formally from L-aspartic acid and (2S,3R,4R)-diaminoproline, the latter being available by an efficient synthet

Full text of DOI:10.1002/(SICI)1522-2675(20000216)83:2<444::AID-HLCA444>3.0.CO;2-R

444
Helvetica Chimica Acta ± Vol. 83 (2000)
Synthesis and Solution Conformation of b-Hairpin Mimetics Utilizing a
Template Derived from (2S,3R,4R)-Diaminoproline
by Marc E. Pfeifer, Kerstin Moehle, Anthony Linden, and John A. Robinson*
Institute of Organic Chemistry, University of Zurich, Winterthurerstrasse 190, 8057 Zürich
A novel template was synthesized for stabilizing b-hairpin conformations in cyclic peptide mimetics. The
template is a diketopiperazine derived formally from l-aspartic acid and (2S,3R,4R)-diaminoproline, the latter
being available by an efficient synthetic route from vitamin C. The template was incorporated by solid-phase
peptide synthesis into a cyclic loop mimetic containing the sequence (-Ala-Asn-Pro-Asn-Ala-Ala-template-).
This mimetic was shown by NMR to adopt a stable b-hairpin conformation in (D6)DMSO solution. The
template may prove to be generally useful for creating small-molecule mimetics of hairpin loops on proteins of
diverse function.
1. Introduction. ± Conformationally defined synthetic molecules that mimic surface
epitopes on proteins are potentially valuable reagents in biology, as well as in drug and
vaccine design [1]. Short linear peptides are inherently flexible molecules, especially in
aqueous solution, and so are often poor mimics of the secondary structures (turns, a-
helices, b-strands) found on the surfaces of folded proteins. To circumvent this folding
problem, much attention has been focused on the design of templates that constrain
peptide chains into biologically relevant secondary structures [2]. We report here a
novel bicyclic template 1, comprising a diketopiperazine derived from l-aspartic acid
and (2S,3R,4R)-diaminoproline, which in the context of a cyclic peptide mimetic can
stabilize b-hairpin conformations.
The design of 1 stems from our earlier studies with the related template 2 [3][4],
which was used to generate the loop mimetic 4 containing the motif Asn-Pro-Asn-Ala
(NPNA). The biological interest in the NPNA motif stems from its occurrence in a
tandemly repeated form (as NPNA)ꢀ37) in the circumsporozoite (CS) protein of the
malaria parasite Plasmodium falciparum, where it is believed to prefer a b-turn, or
turn-like conformation [5][6]. Linear synthetic peptides containing tandemly repeated
NPNA motifs were evaluated in the late 1980s as potential malaria vaccine candidates
[7]; however, their efficacy was found to be low [8]. Such linear peptides in H2O are
conformationally flexible [5], whereas in the intact CS protein, the NPNA repeats
presumably adopt a folded conformation. The structure of the native CS protein is
unknown. NMR Studies showed that the NPNA motif in 4 adopts a stable b-turn
conformation, and immunological studies showed that 4 can elicit anti-sporozoite
antibodies in mice [3]. In the solution structure of 4, the amido N-atom in the 4-
aminoproline moiety of the template prefers a pseudo-equatorial position (an Eg ring
pucker), as shown in Fig. 1, which leads preferentially to a bulged loop conformation in
aqueous solution, rather than to a stable b-hairpin geometry.

Helvetica Chimica Acta ± Vol. 83 (2000)
445
Since b-hairpins frequently play important roles in protein-protein recognition, and
templates that specifically stabilize this secondary structure could be generally useful in
peptide-mimetic design, we set out to modify 2 by introducing an additional axial amino
group at the 3-position of the proline moiety, as in 1, which could then be used to more
accurately anchor a peptide loop in a b-hairpin geometry. We describe below the
synthesis of this new template, and of the mimetic 3 containing the ANPNAA loop.
Conformational studies with NMR and MD support the conclusion that 3 adopts a
stable b-hairpin geometry, rather than the bulged loop conformation deduced earlier
for 4, as shown in Fig. 1. Some aspects of this work were described earlier in a
communication [9].
2. Results and Discussion. ± 2.1. Synthesis of the Template. A convenient gram-scale
synthesis of (2S,3R,4R)-diaminoproline was established by exploiting a known route to

446
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 1. Average solution structures of 3 (this work) and 4 (see [3]) deduced by NMR and simulated annealing
(see text). All H-atoms apart from amide NHs are omitted for clarity. N-Atoms blue, O-atoms red, C-atoms
grey, and amide H-atoms white.
the bicyclic b-lactam 5 from vitamin C [10 ± 12], which has been implemented already
on a multi-kilogram scale in a commercial synthesis of b-lactamase inhibitors. As shown
in Scheme 1, 5 was converted to 6 by a Mitsunobu reaction [13] and subsequently into 9
via 7 and 8 by exchange of protecting groups. The b-lactam ring was then opened to
yield, after esterification, the orthogonally protected (2S,3R,4R)-diaminoproline
derivative 10. Coupling to Z-Asp(OtBu)-OH, cyclization to afford the diketopiper-
azine, and further manipulation of the protecting groups gave 1 in good overall yield.
The intermediate 6 was also converted to the orthogonally protected 3,4-
diaminoproline derivative 12 (Scheme 2). However, hydrolysis of the b-lactam in 11
(steps iii) and iv)) led partially to opening of the phthalimide protecting group, thereby
reducing the yield and complicating the purification of 12. The structure of 12 could be
confirmed, however, by X-ray crystallography. In the crystal structure (Fig. 2), the
pyrrolidine ring has a C(b)-endo envelope conformation (c1 ˆ 38.98), i.e. C(b) is
above the plane defined by the other four ring atoms (bE). This places the carbamate
N
C(b) N-atom in axial position (as desired for 3) and the g-amido substituent in an
equatorial position. The carbamate NH group forms an intramolecular H-bond with
the ester carbonyl O-atom (H-bond length 2.44 Š and donor-H-acceptor angle 1178).
The intermediate 12 could then be converted to the diketopiperazine 13, as shown
in Scheme 2. It is potentially a useful feature of the template that the g-amino
substituent in the proline moiety can be functionalized with a variety of acyl groups
(benzoyl in 1, acetyl in 13). For the present purpose, however, only the derivative 1 was
carried forward for further studies.

Helvetica Chimica Acta ± Vol. 83 (2000)
447
1
Scheme 1
)
DMB ˆ 2,4-dimethoxybenzyl, Z ˆ PhCH2OCO, Phth ˆ phthalimido, Bz ˆ PhCO, Boc ˆ tBuOCO, Fmoc ˆ (9H-
fluoren-9-ylmethoxy)carbonyl. i) PhthH, Ph3P, THF, DEAD; 61%. ii) MeNHNH2, DMF, 808. iii) Bz2O, Et3N,
CH2Cl2; 66% (2 steps). iv) K2S2O8, Na2HPO4, MeCN/H2O, 788; 77%. v) Boc2O, Et3N, DMAP, CH2Cl2; 60%. vi)
Na2CO3, THF/H2O. vii) CH2N2, Et2O; 98% (2 steps). viii) H2, Pd/C, DMF; 96%. ix) Z-Asp(OtBu)-OH, HATU,
i
HOAt, Pr2EtN, CH2Cl2; 80%. x) H2, Pd/C, DMF; 100%. xi) CF3COOH, CH2Cl2, 89%. xii) Fmoc-ONSu,
iPr2EtN, CH2Cl2; 60%.
1
Scheme 2
)
DMB ˆ 2,4-dimethoxybenzyl, Z ˆ PhCH2OCO, Phth ˆ phthalimido, Boc ˆ tBuOCO. i) K2S2O8, Na2HPO4,
MeCN/H2O, 788. ii) Boc2O, Et3N, DMAP, CH2Cl2; 65% (2 steps). iii) Na2CO3, THF/H2O. iv) CH2N2, Et2O;
62% (2 steps). v) H2, Pd/C; 100%. vi) Z-Asp(OtBu)-OH, HATU, HOAt, iPr2EtN, CH2Cl2; 80%. vii) H2, Pd/C;
95%. viii) MeNHNH2, 808, EtOH. ix) Ac2O; 70% (2 steps).
1
)
Abbreviations: DEAD, diethyl diazenedicarboxylate; DMAP ˆ 4-(dimethylamino)pyridine; DMF, dime-
thylformamide; HBTU, O-(1H-benzotriazol-1-yl)-N,N,N',N'-tetramethyluronium hexafluorophosphate;
HATU, O-(7-aza-1H-benzotriazol-1-yl)-N,N,N',N'-tetramethyluronium hexafluorophosphate; HOBt, 1-
hydroxy-1H-benzotriazole; HOAt, 1-hydroxy-7-aza-1H-benzotriazole; CIP, 2-chloro-1,3-dimethylimid-
azolidinium hexafluorophosphate; Mtt, 4-methyltrityl.

448
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 2. X-Ray crystal structure of 12
2.2. Synthesis of the Loop Mimetic. To evaluate the hairpin-inducing properties of
the template, the target 3 containing the sequence -Ala-Asn-Pro-Asn-Ala-Ala- was
chosen, so that a direct comparison of its preferred conformation could be made with
that deduced for 4 in earlier studies [3][4]. The template 1 could be incorporated into
the cyclic peptide 3 by standard solid-phase methods and Fmoc ((9H-fluoren-9-
ylmethoxy)carbonyl) chemistry [14]. For example, 1 was coupled to Tentagel-S-AC
resin, and the peptide chain was elaborated to afford H-Ala-Asn(Mtt)-Pro-Asn(Mtt)-
Ala-Ala-template-resin1). After cleavage from the resin with 1% CF3COOH in
CH2Cl2, the linear precursor was cyclized in solution with HATU/HOAt1) in DMF,
and all side-chain protecting groups were then removed with CF3COOH/CH2Cl2 35 :60

Helvetica Chimica Acta ± Vol. 83 (2000)
449
i
and Pr3SiH (5% v/v). After purification by HPLC, the cyclic peptide 3 was obtained
from 1 in an unoptimized 11% yield.
2.3. Conformational Studies. The conformation of 3 was studied in DMSO solution
3
rather than H2O, due to the limited solubility of 3 in the latter solvent. J Coupling
constants, amide temperature coefficients, and amide proton H/D exchange rates, as
well as NOEs were measured by NMR spectroscopy. From distance restraints derived
from NOE intensities, average structures were then calculated by dynamic simulated
annealing. A typical solution structure was also used as a starting point for MD
simulations in explicit DMSO solvent. The results of these experiments and simulations
are presented below.
The 1H-NMR spectrum of 3 shows a single major species (> 98%) on the chemical-
shift time scale in (D6)DMSO solution, whereas 4 in D2O showed ca. 10% of a minor
conformer due to cis/trans isomerism at the Asn-Pro peptide bond. The spectrum was
assigned by standard methods [15] at 305 K, a temperature at which the amide-proton
resonances are optimally resolved (Table 1). The chemical shifts and line-widths were
essentially invariant over the concentration range 21 ± 0.6 mm, indicating an absence of
significant aggregation under these conditions. The amide-proton temperature
coefficients, relative H/D-exchange rates, and key NOE connectivities measured for
3 are summarized in Fig. 3, and coupling constants are collected in Table 2.
H/D-Exchange. The NH C(b) amide proton of the diaminoproline moiety of the
template in 3 exchanges very slowly with deuterium in (D6)DMSO/(D4)MeOH 9 :1,
with a half-life of several days, whereas under the same conditions, the half-lives for
Ala1 and Ala6 NHs are around 30 min, and the aspartate NH of the template and the
side-chain amide protons of Asn2 and Asn4 exchange completely within a few minutes.
By contrast, with mimetic 4, the NH C(g) amide proton of the template exchanged
relatively quickly [3], in accord with the absence of a stable intramolecular H-bond to
this NH group in 4. A further interesting comparison arises with the disulfide-bridged
analogue 14, also studied in earlier work [3]. In this molecule, the disulfide bridge fixes
the loop backbone into a stable b-hairpin conformation. The corresponding NH C(g)
Table 1. 1H-NMR (600 MHz) Chemical Shifts for 3 at 305 K in (D6)DMSO
Residueb)
Chemical shift [ppm]a)
NH
H
C(a)
CH2(b)c) or Me(b)
Others
Ala1
8.55
8.50
±
7.95
7.26
8.48
8.24
4.74
4.84
4.01
4.52
4.37
4.77
4.60
1.22
Asn2
2.83, 2.79
2.17, 1.56
2.48, 2.32
1.08
1.30
4.69
7.74 (NH(d),(E); 7.02 (NH(d),(Z))
1.92, 1.80 (CH2(g)); 3.80, 3.42 (CH2(d))
7.13 (NH(d),(E)); 6.77 (NH(d),(Z))
Pro3
Asn4
Ala5
Ala6
d
Pro(NH2)72
)
4.78 (CH(g)); 3.54, 3.94b) (CH2(d))
8.22 (HNCOPh); 7.86, 7.49, 7.42 (arom. H)
Asp8
8.33
4.18
2.77, 2.98c)
a) Chemical shifts are measured relative to internal SiMe4.
b) Pro(NH2)7 and Asp8 are the 3,4-diaminoproline and aspartatic acid moieties of the template.
c) Stereospecific assignments are in italics, in the order pro-R, pro-S.
d
7
) The Pro(NH2)2 NH and HNCOPh refer to NH C(b) and NH C(g), respectively.

450
Helvetica Chimica Acta ± Vol. 83 (2000)
Table 2. Coupling Constantsa) [Hz] for 3
Residueb)
3J(a,NH)
3J(a,b)
Others
Ala1
7.5
7.6
±
9.4
7.0
9.1
8.9d)
6.8
Asn2
5.0, 4.1
10.0, 7.4c)
10.3, 3.9c)
6.8
6.9
3.5
2J(b,b') ˆ 16.1
n.d.
Pro3
Asn4
2J(b,b') ˆ 14.6
Ala5
Ala6
7
3
Pro(NH2)2
3J(NH C(g)H) ˆ 7.0, J(b,g) ˆ 4.3,
3J(g,dpro-S ) ˆ 10.1, J(g,dpro-R ) ˆ 9.6, J(d,d') ˆ 11.6,
3
2
5J(a,a') ˆ 2.0
2J(b,b') ˆ 16.9
Asp8
< 2.0
4.1, 2.6e)
a
1
) Measured from 1D H-NMR spectra and/or E.COSY experiments.
b) Pro(NH2)27 and Asp8 are the 3,4-diaminoproline and aspartic acid moieties of the template.
c) Stereospecific assignments not available.
d
3
) Refers to the J(NH C(b),H) coupling.
e) Given in the order pro-R, pro-S.
a) The values in brackets refer to the HN-C(g) amide, the others to the HN-C(b)
amide.
b) Pro(NH2)72 and Asp8 are the 3,4-diaminoproline and aspartic acid moieties of
the template.
Fig. 3. Summary of NOEs, H/D exchange rates, and amide temperature coefficients measured for 3 by NMR. The
temperature coefficients ( Dd/T) in ppb/K were determined over the range 295 ± 320 K in (D6)DMSO;
relative H/D exchange rates of peptide NH protons were determined by monitoring residual peak intensities
after dissolution in (D6)DMSO ‡ 10% CD3OD. The half-lives (H/D t1/2 in min) were derived by fitting residual
peak intensities to an exponential function.
amide proton of 14 again exchanges very slowly, mirroring the behaviour of the
NH C(b) amide proton of 3. These data support the conclusions of structure
calculations (see also below) that both 3 and 14 adopt similar b-hairpin conformations,
with the NH C(b) amide proton in 3 and the NH C(g) amide proton in 14 involved
in intramolecular H-bonding across the hairpin.
Coupling Constants. The side chains of Asn2 and Asn4 in 3 have preferred c1 values,
since non-averaged 3J(a,b) values are seen for Asn2 (5.0 and 4.1 Hz) and for Asn4 (3.9
and 10.3 Hz) (Table 2). A similar pattern was observed earlier for 14 (i.e., Asn2, 3J(a,b)
3.9 and 5.4 Hz; Asn4, 3J(a,b) 5.4 and 10.1 Hz), whereas in 4, high and low 3J(a,b) values

Helvetica Chimica Acta ± Vol. 83 (2000)
451
3
were found for the methylene protons of both Asn2 and Asn4 [3]. The J(a,NH) for
Ala1, Asn2, and Ala5 of 3 are lower than expected for an ideal b-hairpin, although the
Ala6 and Asn4 3J(a,NH) values are close to those expected (i.e., >9.0 Hz) for a b-
strand and a type-I b-turn, respectively.
The 3J(g,d) coupling constants within the proline moiety of the template in 3 both
have large values (9.6 and 10.1 Hz), as expected for a preferred bE envelope
conformation, with the b-amido group axial and the g-benzamido substituent
equatorial. This conformation gives rise to small and large dihedral angles for
H
C(g) C(d) Hpro-R and H C(g) C(d) Hpro-S, respectively, and hence to large
3J(g,d) values for both couplings. In comparison, the 3J(g,d) values in 4 (5.7 and 8.9 Hz)
were close to the averaged values expected from rapid flipping, whereas in 14 two small
values (<2.0 and 4.2 Hz) were found [3], as expected for the axial position of the g-
amido N-atom. The two small 3J(a,b) values for the aspartate moiety of the template in
3 indicate a preferred c1 torsion of around ‡ 608, thus placing the C(g) carbonyl group
above the diketopiperazine ring, as found earlier for 4 and 14 [3]. Finally, the long-
range 5J(a,a') coupling (Table 2) is well precedented and indicates a significant
puckering of the diketopiperazine ring [16][17].
NOEs and Structure Calculations. A summary of key NOEs observed in NOESY
spectra of 3 in (D6)DMSO is shown in Fig. 3. A cross-strand H C(a)/H C(a) NOE
between Ala1 and Ala6 could not be observed due to the small chemical-shift difference
between these H C(a) resonances (Table 1). However, several other NOEs were
observed that strongly indicate a stable b-hairpin conformation, in particular between
Asn2NHi/Ala5NHi‡3 and Ala1H C(ai)/Ala5NHi‡4. NOESY Experiments also revealed
for Ala1/Asn2 as well as for Ala5/Ala6 strong H C(ai)/NHi‡1 NOEs, but an absence of
NHi/NHi‡1 NOEs, as would be expected in the proposed b-strands. A b-turn in the

452
Helvetica Chimica Acta ± Vol. 83 (2000)
NPNA motif was indicated by a relatively strong Asn4NHi/Ala5NHi‡1 NOE, as well as
by NOEs between Asn2H C(b) and Ala5NH (Fig. 3).
Since the NMR data provide evidence for a preferred conformation, average
solution structures were determined by dynamic simulated annealing (SA), using
distance restraints derived from NOE build-up curves. From 30 structures generated in
this way, the energies of 15 were within 20 kcal/mol of the global energy minimum and
showed no distance restraint violations >0.3 Š. Of this set, 8 structures were within
10 kcal/mol of the energy minimum, and all possessed essentially the same backbone b-
hairpin conformation. The remaining seven possessed a distorted hairpin, and were not
considered further because the NH C(b) was directed out towards solvent rather than
participating in intramolecular H-bonding, as expected on the grounds of the very slow
H/D exchange rate of this amide proton (Fig. 3).
The final 8 structures superimposed over the backbone C-, C(a)-, and N-atoms in
residues Ala1±Ala6 with an average pairwise r.m.s.d. of only 0.11 Š (Fig. 4). Three
cross-strand H-bonds were observed, one involving the template Asp8C(g)O/
7
Pro(NH2)2 NH C(b) groups, as well as backbone Asn2NH/Ala5CO and Asn2CO/
Ala5NH. The three amide NH groups involved in these H-bonds are also those that
have the slowest H/D exchange rates (Fig. 3), whereas the Ala1 and Ala6 peptide NHs
point outwards from the hairpin, are not involved in cross-strand H-bonding, and have
Fig. 4. Superimposition over the backbone N-, C(a)-, and C-atoms of the final solution structures for 3
determined by simulated annealing (see text)

Helvetica Chimica Acta ± Vol. 83 (2000)
453
relatively fast H/D exchange rates. The Asn4 peptide NH forms H-bonds with the side-
chain carbonyl O-atom of Asn2. The Asn2-Pro3-Asn4-Ala5 motif is locked into close to a
type-I b-turn conformation. The Asn2 side-chain conformation in the minimum energy
structure (Fig. 1) also has both H C(a) C(b) H torsion angles at ca. 608, which is
consistent with the two small 3J(a,b) values (Table 2) found for Asn2. With this c1 angle,
the Asn2 side-chain amide CO group can H-bond with either the Asn4 or Ala5 peptide
NH groups (Fig. 1).
The average solution structure deduced for 3 was compared with that deduced for
the disulfide-bridged loop mimetic 14 in earlier work. As shown in Fig. 5, the backbone
conformations of the two mimetics are essentially identical.
Fig. 5. Superimposition over the backbone N-, C(a)-, and C-atoms of the solution structures for 3 and 14 (see
[3]) determined by simulated annealing (see text)
MD Simulations. Simulations of 3 were carried out in explicit DMSO solvent at
300 K to explore how the structural properties inferred from NOEs might be influenced
by motional averaging. The minimum-energy structure found for 3 by SA was used as a
starting point for 2-ns MD simulations, both with and without time-averaged distance
restraints (TA-DR) [18][19]. The results of both simulations were similar, suggesting
that the restraints are well satisfied by structures at or near a local or global energy
minimum in the force field used.

454
Helvetica Chimica Acta ± Vol. 83 (2000)
In the 2-ns simulation with TA-DR, only two of 39 average restraint violations were
within the range 0.5 ± 1.0 Š, the others were <0.5 Š. The average 3J coupling constants
3
showed average violations for several backbone J(a,NH) values of ca. 1.6 Hz. The
reasons for the discrepancies are unknown, but conceivably may reflect an insuffi-
ciently long simulation time or an inadequate representation of the mimetic in the force
field. The Pro3 and Asn4 f/y angles remain for most of the simulation with TA-DR
within Æ308 of the values expected for a bI-turn conformation (Fig. 6). And three well-
populated cross-strand backbone-backbone H-bonds are found in the restrained and
unrestrained trajectories (Fig. 7). The population of all three simultaneously would
correspond to a class 2 :2 b-hairpin, the most abundant b-hairpin type found in high-
resolution protein crystal structures [20]. However, the most frequently found turn
types in 2 :2 b-hairpins in proteins are the type-I' and -II', and less so the type-I turn
[21][22]. Here, type-I' and -II' turns are not possible due to the presence of proline
(f ꢀ 608) in the i ‡ 1 position. The Asn2CO/Ala5NH H-bond is only poorly populated
in the unrestrained simulation. Instead, the side-chain CO of Asn2 frequently acts as an
H-bond acceptor for the main-chain NHs of Asn4 and Ala5 (Figs. 1 and 7). There is
again a good correlation between the involvement of amide NHs in intramolecular H-
bonding, indicated experimentally by relatively slow H/D exchange rates and the
population of H-bonds during the MD simulation (Figs. 3 and 7). On the other hand,
the significantly slower rate of H/D exchange seen for the template NH C(b), in
comparison to all others (Fig. 3), is not so well reflected in the different H-bonding
populations in the restrained simulation (Fig. 7).
Fig. 6. Variations in f and y torsion angles for Pro3 and Asn4 during the MD simulation with TA-DR
Large variations of the ring pucker in the diaminoproline moiety of the template in
3 were not observed during either simulation, unlike in simulations reported earlier
with 4 [3]. The c1 torsion angle (N C(a) C(b) C(g)) remains close to 358
b
throughout both simulations (Fig. 8,a), with the E ring pucker and the NH C(b)
substituent in an axial position and the peptide loop in a b-hairpin conformation. For
comparison, an unrestrained MD simulation was also performed under the same
g
conditions but starting with an energy-minimized structure adopting an inverted E

Helvetica Chimica Acta ± Vol. 83 (2000)
455
Fig. 7. Population of H-bonds during the MD simulation with TA-DR. The H-bonds are shown by broken lines
and the percent population over the entire 2-ns simulation with TA-DR (in bold) and without restraints (in
italics) are given.
conformation (c1 ꢀ 208; C(g)-endo) within the diaminoproline moiety of the template
(i.e., with the NH C(b) substituent in an equatorial position (Fig. 8,b)) and a bulged-
loop backbone conformation. This structure was derived by computer modelling, using
the SA structure deduced for 4 in earlier work, and exchanging templates. About 200 ps
into the simulation of this model system, the pyrrolidine ring of the template flipped
back to the bE conformation (c1 ꢀ 358; C(b)-endo; Fig. 8,b), and the loop backbone
into a 2 :2 b-hairpin geometry, as seen in the simulations starting from the NMR
solution structure of 3.
In summary, the simulations indicate only a limited motional averaging within the
backbone of the mimetic 3, and reinforce the conclusion that the molecule adopts a
well-defined b-hairpin conformation.
3. Conclusions. ± An efficient synthetic route to the novel template 1 has been
developed. The template can be incorporated into a linear peptide-mimetic precursor
using solid-phase Fmoc chemistry, with subsequent solution-phase macrocyclization to
afford mimetic 3. NMR Spectra and MD simulations of 3 indicate a well defined b-
hairpin conformation in DMSO solution. The deduced average solution structures
(Fig. 1) are well satisfied by the NOEs and amide H/D-exchange rates, and comprise a
well-populated class 2 :2 b-hairpin conformation with a type-I b-turn in the NPNA
motif at the tip of the hairpin loop. It seems that the template may be useful for

456
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 8. a) c1 Torsion angle in the template (N C(a) C(b) C(g)) monitored during the course of an MD
simulation with TA-DR, starting from the solution SA structure, and b) as in a) but starting from a structure with
g
the template in a E (c1 ꢀ 208) conformation
inducing b-hairpin conformations in six (or more) residue loops of diverse sequence,
although no systematic investigation of the dependence of loop conformation on
peptide sequence has so far been carried out. We have shown in other studies [23] that a
template comprising a heterochiral diproline unit (d-Pro-l-Pro) is very effective in
stabilizing b-hairpin conformations in octapeptide loops. Given the wide occurrence of
b-hairpins in proteins of diverse function, both 1 and the d-Pro-l-Pro unit may prove to
be generally useful in hairpin-mimetic design.
The authors thank the Swiss National Science Foundation for financial support and Dr. Pflieger, F.
Hoffmann-La Roche Ltd., Basel, for a generous gift of compound 5.

Helvetica Chimica Acta ± Vol. 83 (2000)
457
Experimental Part
General. See [24].
Benzyl (1S,4R,5R)-6-(2,4-Dimethoxybenzyl)-4-(1,3-dihydro-1,3-dioxo-2H-isoindol-2-yl)-7-oxo-2,6-diaza-
bicyclo[3.2.0]heptane-2-carboxylate (6). To a suspension of alcohol 5 (20 g, 48.5 mmol), triphenylphosphine
(31.8 g, 121.2 mmol), and phthalimide (21.4 g, 145.4 mmol) in anh. THF (500 ml), diethyl diazenedicarboxylate
(DEAD; 9.1 ml, 58.1 mmol) was slowly added at 308 under Ar. After stirring at 308 for 3 h, the soln. was
stirred at r.t. for 16 h. The soln. was washed with sat. aq. NaHCO3 soln. and brine, dried (MgSO4), and
evaporated and the product purified by FC (hexane/AcOEt 4 :6 ! 3 :7): 6 (16.0 g, 61%). TLC (hexane/AcOEt
4 :6): Rf 0.44. White solid. M.p. 56 ± 588. [a]D20
ˆ
55.1 (c ˆ 0.490, AcOEt). IR (KBr): 2940w, 2832w, 1765s,
1712vs, 1611m, 1589m, 1506m. 1H-NMR (300 MHz, (D6)acetone): 7.89 ± 7.80 (m, 4 H); 7.44 ± 7.28 (m, 5 H); 6.81
(d, J ˆ 8.3, 1 H); 6.22 (d, J ˆ 2.4, 1 H); 6.19 (dd, J ˆ 8.3, 2.4, 1 H); 5.24 (br. d, J ˆ 3.7, 1 H); 5.17 (s, 2 H); 4.55 ±
4.36 (m, 4 H); 4.30 (d, J ˆ 15.0, 1 H); 3.98 (d, J ˆ 15.0, 1 H); 3.64 (s, 3 H); 3.50 (s, 3 H). 13C-NMR (75 MHz,
(D6)acetone): 169.0 (s); 166.0 (s); 161.6 (s); 159.0 (s); 154.5 (s); 138.0 (s); 135.2 (d); 132.8 (s); 131.3 (d); 129.4
(d); 128.8 (d); 128.7 (d); 124.2 (d); 117.2 (s); 105.5 (d); 99.3 (d); 67.75 (d); 67.71 (t); 59.3 (d); 58.3 (d); 55.7 (q);
‡
‡
55.5 (q); 43.7 (t); 40.6 (t). ESI-MS: 564.2 ([M ‡ Na] ), 542.2 ([M ‡ H] ).
Benzyl (1S,4R,5R)-4-(Benzoylamino)-6-(2,4-dimethoxybenzyl)-7-oxo-2,6-diazabicyclo[3.2.0]heptane-2-
carboxylate (7). To a soln. of 6 (8 g, 14.8 mmol) in DMF (250 ml), methylhydrazine (3.9 ml, 5 equiv.,
74.1 mmol was added). After stirring at 808 for 1.5 h, DMF was evaporated, and CH2Cl2 (250 ml), Et3N (2.1 ml,
15.1 mmol), and benzoic anhydride (3.5 g, 15.5 mmol) were added. After stirring overnight, the suspension was
washed with 10% aq. citric acid soln. and brine, dried (MgSO4), and evaporated. Purification by FC (hexane/
AcOEt 3 :7; Rf 0.66) gave 7 (5.0 g, 66% over 2 steps). White solid. M.p. 142 ± 1438. TLC (hexane/AcOEt 2 :8):
Rf 0.76. [a]2D0
ˆ
43.4 (c ˆ 0.53, AcOEt). IR (KBr): 3285s, 3054w, 3020w, 3000w, 3020w, 2930m, 2835w, 1755vs,
1729vs, 1712vs, 1660s, 1629vs, 1585s, 1535vs, 1506s. 1H-NMR (300 MHz, (D6)acetone): 7.91 ± 7.24 (m, 11 H); 6.99
(d, J ˆ 8.2, 1 H); 6.45 (d, J ˆ 2.3, 1 H); 6.41 (dd, J ˆ 8.2, 2.3, 1 H); 5.17 (br. s, 3 H); 4.61 ± 4.51 (m, 1 H); 4.54
(d, J ˆ 15.1, 1 H); 4.39 ± 4.19 (m, 2 H); 4.17 (d, J ˆ 15.1, 1 H); 3.76 (s, 3 H); 3.61 (s, 3 H); 3.55 (dd, J ˆ 10.8, 10.2,
1 H). 13C-NMR (75 MHz, (D6)acetone): 167.7 (s); 166.1 (s); 161.9 (s); 159.3 (s); 154.5 (s); 138.0 (s); 135.1 (s);
132.4 (d); 131.4 (d); 129.4 (d); 129.3 (d); 128.8 (d); 128.4 (d); 117.2 (s); 105.5 (d); 99.2 (d); 68.5 (d); 67.6 (t); 58.6
‡
‡
(d); 55.8 (q); 55.7 (q); 51.1 (d); 47.5 (t); 41.3 (t). ESI-MS: 538.3 ([M ‡ Na] ), 516.3 ([M ‡ H] ). Anal. calc. for
C29H29N3O6 (515.56): C 67.6, H 5.7, N 8.2; found: C 67.5, H 5.7, N 8.2.
Benzyl (1S,4R,5R)-4-(Benzoylamino)-7-oxo-2,6-diazabicyclo[3.2.0]heptane-2-carboxylate (8). To a reflux-
ing soln. of 7 (4.4 g, 8.5 mmol) in MeCN (75 ml) and H2O (10 ml), a soln. of K2S2O8 (9.2 g, 4 equiv., 34.0 mmol)
and Na2HPO4 ´ 2H2O (12.2 g, 8 equiv., 68.5 mmol) in H2O (50 ml) was added in 4 portions under vigorous
stirring every 5 min. Upon conversion (TLC control, 55 min), the orange soln. was immediately cooled and the
two-phase system made slightly alkaline by adding crystalline NaHCO3. The aq. phase was extracted with
AcOEt, the org. phase washed successively with sat. aq. NaHCO3 soln. and brine, dried (MgSO4), and
evaporated, and the residue purified by FC (hexane/AcOEt 1:1): 8 (2.4 g, 77%). Off-white solid. M.p. 92 ± 948.
TLC (hexane/AcOEt 1:4): Rf 0.37. [a]D20
ˆ
4.9 (c ˆ 0.51, AcOEt). IR (KBr): 3295m (br.), 3060w, 2950w,
1769vs, 1722s, 1710s, 1696s, 1680s, 1665s, 1645s, 1602m, 1579m, 1534s. 1H-NMR (300 MHz, (D6)acetone): 7.92 ±
7.28 (m, 10 H); 5.26 (br. s, 1 H); 5.15 (br. s, 2 H); 4.54 ± 4.48 (m, 2 H); 4.47 (br. dd, 1 H); 3.42 (dd, J ˆ 10.8, 10.2,
1 H). 13C-NMR (75 MHz, (D6)acetone): 167.9 (s); 166.3 (s); 154.4 (s); 138.0 (s); 135.3 (s); 132.4 (d); 129.4 (d);
129.2 (d); 128.8 (d); 128.4 (d); 69.8 (d); 67.6 (t); 56.6, 54.7 (d); 51.4 (d); 50.7 (d); 47.7 (t). ESI-MS: 388.2 ([M ‡
‡
Na] ).
2-Benzyl 6-(tert-Butyl) (1S,4R,5R)-4-(Benzoylamino)-7-oxo-2,6-diazabicyclo[3.2.0]heptane-2,6-dicarboxy-
late (9). To 8 (1.46 g, 4.0 mmol), Et3N (558 ml, 4.0 mmol), and 4-(dimethylamino)pyridine (DMAP; 0.49 g,
4.0 mmol) in anh. CH2Cl2 (50 ml), (Boc)2O (1.74 g, 8.0 mmol) was added. After stirring for 1 h, the org. phase
was washed with 1% aq. citric acid soln. and brine, dried (MgSO4), and evaporated, and the residue purified by
FC (hexane/AcOEt 3 :2): 9 (1.12 g, 60%). Yellowish solid. TLC (hexane/AcOEt 1:1): Rf 0.5. M.p. >658 (dec.).
[a]2D1
ˆ
49.6 (c ˆ 0.482, AcOEt). IR (KBr): 3342s (br.), 3060m, 3030w, 2978m, 2930m, 1812vs, 1727vs, 1712vs,
1692vs, 1665vs, 1602vs, 1578s, 1550vs, 1536vs. 1H-NMR (300 MHz, (D6)acetone): 8.08 (br. d, J ˆ 6.4, 1 H); 7.95
(d, J ˆ 7.3, 2 H); 7.57 ± 7.29 (m, 8 H); 5.31 (d, J ˆ 4.8, 1 H); 5.12 (s, 2 H); 4.87 (dd, J ˆ 5.0, 4.8, 1 H); 4.68
(dddd, J ˆ 11.0, 10.7, 6.4, 5.0, 1 H); 4.34 (br. m, 1 H); 3.36 (dd, J ˆ 11.0, 10.7, 1 H); 1.29 (s, 9 H). NOE: 5.31
(H C(1)) ! 4.87 (H C(5)); 4.87 (H C(5)) ! 5.31 (H C(1)); 4.68 (H C(4)) ! 4.87 (H C(5)); 4.34
(Hpro-R C(3)) ! 4.68 (H C(4)); 3.36 (Hpro-S C(3)) ! 4.34 (Hpro-R C(3)). 13C-NMR (75 MHz, (D6)acetone):
167.4 (s); 163.9 (s); 154.4 (s); 154.1 (s); 149.6 (s); 137.7 (s); 135.1 (s); 132.4 (d); 129.3 (d); 129.2 (d); 128.9 (d);

458
Helvetica Chimica Acta ± Vol. 83 (2000)
128.3 (d); 84.0 (s); 68.0 (d); 67.9 (t); 59.9 (d); 58.9 (d); 51.2 (d); 50.5 (d); 48.1 (t); 28.1 (q). ESI-MS: 488.2
‡
‡
‡
([M ‡ Na] ), 410.1 ([M ‡ H tBu] ), 388.1 ([M ‡ Na tBu CO2] ).
Methyl (2S,3R,4R)-4-(Benzoylamino)-1-[(benzyloxy)carbonyl]-3-{[(tert-butoxy)carbonyl]amino}prolinate
(10). To a soln. of 9 (1.2 g, 2.6 mmol) in THF (65 ml) and H2O (20 ml), crystalline Na2CO3 (0.68 g, 6.4 mmol)
was added portionwise. When the turbid soln. had cleared (a few ml of H2O may be added), THF was
evaporated and the aq. soln. acidified to pH 4.5 with dil. aq. AcOH soln. The white suspension was extracted
with AcOEt and the org. phase washed with brine, dried (Na2SO4), and evaporated. The crude product (free
‡
‡
‡
carboxylic acid; ESI-MS: 989.6 ([2M ‡ Na] ), 788.4 ([2(M tBu CO2) ‡ Na] ), 506.2 ([M ‡ Na] ), 484.2
([M ‡ H] ), 428.1 ([M ‡ H tBu] ), 384.0 ([M ‡ H tBu CO2] )) in AcOEt (20 ml) was titrated with CH2N2
in Et2O. The org. phase was evaporated: 10 (1.26 g, 98%). White solid. TLC (hexane/AcOEt 1:1): Rf 0.55. M.p.
68 ± 708. [a]2D0 ˆ ‡11.1 (c ˆ 0.5, AcOEt). IR (KBr): 3349s (br.), 3065m, 3033m, 2979s, 2899m, 1715vs, 1668vs,
1603m, 1581s, 1536vs. 1H-NMR (300 MHz, (D6)acetone; (E)/(Z)-isomers): 7.86 (br. d, 3 H); 7.58 ± 7.31
(m, 8 H); 6.31 (2 br. d, 1 H); 5.19 ± 4.94, 4.81 ± 4.66 (2m, 5 H); 3.94 ± 3.66 (m, 5 H); 1.36, 1.35 (2s, 9 H). 13C-NMR
(75 MHz, (D6)acetone; (E)/(Z)-isomers): 173.6, 172.9 (2s); 167.2 (s); 156.4 (s); 154.8, 154.4 (2s); 137.7, 137.6
(2s); 135.3 (s); 132.4 (d); 129.3 (d); 128.9 (d); 128.7 (d); 128.6 (d); 128.2 (d); 79.8 (s); 67.7 (t); 62.4 (d); 62.0 (d);
‡
‡
‡
‡
55.2 (d); 54.3 (d); 53.0 (q); 52.3 (t); 52.1 (t); 51.6 (d); 50.7 (d); 28.5 (q). ESI-MS: 520.4 ([M ‡ Na] ), 498.3
([M ‡ H] ), 442.3 ([M ‡ H tBu] ), 398.2 ([M ‡ H tBu CO2] ).
‡
‡
‡
Methyl (2S,3R,4R)-4-(Benzoylamino)-3-{[(tert-butoxy)carbonyl]amino}prolinate. A soln. of 10 (1.18 g,
2.37 mmol) in DMF (80 ml) was stirred with 10% Pd/C (0.2 g) under H2 for 1.5 d. DMF was evaporated, the
residue suspended in AcOEt, the mixture filtered through Celite, and the filtrate evaporated: white solid (0.83 g,
96%). TLC (CH2Cl2/MeOH 95 :5): Rf 0.32. M.p. 73 ± 748. [a]D20
ˆ
4.4 (c ˆ 0.39, AcOEt). IR (KBr): 3365m
(br.), 2978w, 2931w, 1745m, 1730m, 1710s, 1694s, 1680s, 1670s, 1645s, 1604s, 1579m, 1540s, 1490m. 1H-NMR
(600 MHz, (D6)acetone): 7.88 (d, J ˆ 7.1, 2 H); 7.77 (br. d, J ˆ 6.8, 1 H); 7.56 ± 7.43 (m, 3 H); 5.92 (br. d, J ˆ 7.8,
1 H); 4.80 (dddd, J ˆ 7.1, 6.8, 6.1, 5.7, 1 H); 4.56 (ddd, J ˆ 7.8, 6.4, 6.1, 1 H); 4.08 (d, J ˆ 6.4, 1 H); 3.69 (s, 3 H);
3.30 (dd, J ˆ 10.6, 7.1, 1 H); 3.13 (dd, J ˆ 10.6, 5.7, 1 H); 1.33 (s, 9 H). 13C-NMR (150 MHz, (D6)acetone): 174.0
(s); 167.0 (s); 156.1 (s); 135.4 (s); 132.1 (d); 129.0 (d); 128.1 (d); 79.0 (s); 63.6 (d); 55.5 (d); 52.6 (d); 52.3 (q);
‡
‡
‡
50.2 (t); 28.4 (q). ESI-MS: 386.3 ([M ‡ Na] ), 364.3 ([M ‡ H] ), 308.3 ([M ‡ H tBu] ).
Methyl {N2-[(Benzyloxy)carbonyl]-O4-(tert-butyl)-l-aspart-1-yl}-(2S,3R,4R)-4-(benzoylamino)-3-{[(tert-
butoxy)carbonyl]amino}prolinate. To the foregoing product (0.79 g, 2.17 mmol), Z-Asp(OtBu)-OH (0.84 g,
i
2.6 mmol), HATU (0.99 g, 2.6 mmol), and HOAt (0.3 g, 2.2 mmol) in anh. CH2Cl2 (25 ml), Pr2EtN (1.1 ml,
6.43 mmol) was added, and the soln. was stirred at r.t. for 24 h1). After washing with sat. aq. NaHCO3 soln. and
brine, the org. phase was dried (MgSO4) and evaporated and the residue purified by FC (hexane/AcOEt 2 :3):
white solid (1.16 g, 80%). TLC (hexane/AcOEt 1:1): Rf 0.56. M.p. 97 ± 988. [a]D20
ˆ
21.2 (c ˆ 0.48, AcOEt). IR
(KBr): 3320m (br.), 3060w, 2975m, 2930m, 1739vs, 1728vs, 1714vs, 1660vs, 1650vs, 1602m, 1580m, 1534vs, 1503s,
1486s. 1H-NMR (300 MHz, (D6)acetone): 7.96 (d, J ˆ 9.1, 1 H); 7.84 (d, J ˆ 7.3, 2 H); 7.52 ± 7.39 (m, 3 H); 7.32 ±
7.25 (m, 5 H); 6.73 (d, J ˆ 8.6, 1 H); 6.18 (br. d, 1 H); 5.03 ± 4.98 (s ‡ m, 3 H); 4.78 ± 4.59 (m, 3 H); 4.05 ± 3.96
(m, 2 H); 3.68 (s, 3 H); 2.73 (dd, J ˆ 16.3, 7.9, 1 H); 2.50 (dd, J ˆ 16.3, 6.2, 1 H); 1.32, 1.31 (2s, 18 H). 13C-NMR
(75 MHz, (D6)acetone): 172.6 (s); 170.9 (s); 170.2 (s); 167.2 (s); 156.6 (s); 156.3 (s); 137.9 (s); 135.4 (s); 132.4
(d); 129.3 (d); 128.8 (d); 128.5 (d); 128.3 (d); 81.3 (s); 79.8 (s); 67.2 (t); 62.1 (d); 53.5 (d); 52.9 (q); 52.7 (t); 52.0
‡
‡
(d); 50.0 (d); 38.7 (t); 28.5 (q); 28.2 (q). ESI-MS: 691.5 ([M ‡ Na] ), 669.4 ([M ‡ H] ), 613.4 ([M ‡ H
tBu] ), 557.4 ([M ‡ H 2tBu] ); 513.3 ([M ‡ H 2tBu CO2] ).
‡
‡
‡
tert-Butyl (3S,7R,8R,8aS)-7-(Benzoylamino)-8-{[(tert-butoxy)carbonyl]amino}octahydro-1,4-dioxopyrro-
lo[1,2-a]pyrazine-3-acetate. The foregoing product (1.14 g, 1.7 mmol) in DMF (100 ml) was stirred with 10% Pd/
C (0.2 g) under H2 for 5 d. DMF was evaporated, the residue suspended in CH2Cl2, the mixture filtered through
Celite, and the filtrate evaporated: white solid (0.86 g, 100%). TLC (CH2Cl2/MeOH 9 :1): Rf 0.59. M.p. 144 ±
1468. [a]2D0
ˆ
18.8 (c ˆ 0.48, AcOEt). IR (KBr): 3350s (br.), 2980m, 2930m, 1783m, 1770m, 1712vs, 1693vs,
1682vs, 1668vs, 1660vs, 1645vs, 1580m, 1533vs, 1487vs. 1H-NMR (600 MHz, (D6)acetone): 7.85 (br. d, J ˆ 7.2,
3 H); 7.51 (t, J ˆ 7.2, 1 H); 7.42 (t, J ˆ 7.2, 2 H); 7.31 (br. s, 1 H); 6.41 (d, J ˆ 7.7, 1 H); 4.82 (m, 1 H); 4.67
(m, 2 H); 4.36 (dd, J ˆ 4.5, 4.1, 1 H); 3.85 (dd, J ˆ 11.8, 9.0, 1 H); 3.61 (dd, J ˆ 11.8, 10.6, 1 H); 3.00 (dd, J ˆ 17.3,
4.1, 1 H); 2.82 (dd, J ˆ 17.3, 4.5, 1 H); 1.45 (s, 9 H); 1.34 (s, 9 H). 13C-NMR (75 MHz, (D6)acetone): 171.5 (s);
167.7 (s); 165.0 (s); 164.5 (s); 157.4 (s); 135.5 (s); 132.1 (d); 129.1 (d); 128.2 (d); 82.2 (s); 79.6 (s); 61.8 (d); 55.0
‡
‡
(d); 52.7 (d); 51.9 (d); 48.0 (t); 38.0 (t); 28.4 (q); 28.3 (q). ESI-MS: 525.4 ([M ‡ Na] ), 503.3 ([M ‡ H] ), 447.3
([M ‡ H tBu] ), 403.3 ([M ‡ H tBu CO2] ). Anal. calc. for C25H34N4O7 (502.56): C 59.7, H 6.8, N 11.2;
‡
‡
found: C 59.4, H 6.9, N 11.0.
[(3S,7R,8R,8aS)-7-(Benzoylamino)-3-(carboxymethyl)octahydro-1,4-dioxopyrrolo[1,2-a]pyrazin-1-yl]am-
monium Trifluoroacetate. To the foregoing product (0.74 g, 1.47 mmol) in anh. CH2Cl2 (40 ml), CF3COOH

Helvetica Chimica Acta ± Vol. 83 (2000)
459
(9 ml, 0.12 mol) was added at 08, and the soln. was stirred for 5 h at r.t. After evaporation, the residue was
redissolved in MeCN/H2O and lyophilized: white solid (0.6 g, 89%). M.p. 148 ± 1508. [a]2D0 ˆ ‡45.6 (c ˆ 0.29,
acetone). IR (KBr): 3700 ± 2500m (br.), 1730s, 1712w, 1682vs, 1651vs, 1581m, 1536s, 1515s, 1490m. 1H-NMR
(600 MHz, (D6)acetone): 7.95 (d, J ˆ 7.3, 2 H); 7.55 (t, J ˆ 7.3, 1 H); 7.45 (t, J ˆ 7.3, 2 H); 5.35 (dd, J ˆ 5.2, 3.3,
1 H); 5.28 (ddd, J ˆ 10.5, 8.8, 5.2, 1 H); 5.09 (dd, J ˆ 3.3, <2.0, 1 H); 4.54 (br. dd, J ˆ 7.4, 3.8, 1 H); 4.19 (dd, J ˆ
11.9, 8.8, 1 H); 3.95 (dd, J ˆ 11.9, 10.5, 1 H); 3.15 (dd, J ˆ 17.6, 3.8, 1 H); 2.85 (dd, J ˆ 17.6, 7.4, 1 H). NOE: 5.35
(H C(1)) ! 5.09 (H C(8a)); 5.28 (H C(2)) ! 3.95 (Hpro-R C(3)); 5.09 (H C(8a)) ! 5.35 (H C(1)); 4.19
(Hpro-S C(3))! 3.95 (Hpro-R C(3)); 3.95 (Hpro-R C(3))! 5.28 (H C(2), 4.19 (Hpro-S C(3)). 13C-NMR (75 MHz,
(D6)acetone): 173.1 (s); 168.2 (s); 165.1 (s); 164.6 (s); 161.1 (q); 134.0 (s); 132.6 (d); 129.2 (d); 128.3 (d); 117.1
‡
‡
(q); 61.9 (d); 61.4 (d); 52.7 (d); 50.4 (d); 47.3 (t); 36.5 (t). ESI-MS: 369.2 ([M ‡ Na] ), 347.2 ([M ‡ H] ).
(3S,7R,8R,8aS)-7-(Benzoylamino)-8-{[(9H-fluoren-9-ylmethoxy)carbonyl]amino}octahydro-1,4-dioxopyr-
rolo[1,2-a]pyrazine-3-acetic Acid (1). To a suspension of the foregoing product (0.58 g, 1.26 mmol) in anh.
i
CH2Cl2 (45 ml), and Pr2EtN (1.28 ml, 6 equiv., 7.48 mmol), Fmoc-ONSu (1.27 g, 3 equiv., 3.76 mmol) was
added, and the soln. was stirred at r.t. for 24 h. The org. phase was washed with 10% aq. citric acid soln., dried
(MgSO4), and evaporated. TLC (CH2Cl2/MeOH 9 :1, ‡ 0.1% CF3COOH): Rf 0.49. White solid (0.43 g, 60%).
M.p. >1868 (dec.). [a]D20
ˆ
4.6 (c ˆ 0.53, acetone). IR (KBr): 3327m (br.), 3065w, 2953w, 2896w, 1694vs,
1665vs, 1603w, 1579w, 1534s. 1H-NMR (600 MHz, (D6)acetone): 7.83 (d, J ˆ 7.5, 2 H); 7.82 (d, J ˆ 6.8, 2 H); 7.67,
7.65 (2d, J ˆ 7.5, 2 H); 7.42 (t, J ˆ 7.3, 1 H); 7.38 (t, J ˆ 7.5, 2 H); 7.36 (t, J ˆ 7.5, 2 H); 7.30 (t, J ˆ 7.5, 2 H); 7.27
(t, J ˆ 7.5, 2 H); 7.23 (2t, J ˆ 7.5, 2 H); 5.01 (ddd(d), J ˆ 10.1, 9.4, 4.1, 1 H); 4.87 (t, J ˆ 4.1, 1 H); 4.76 (d(d), J ˆ
4.1, <2.0, 1 H); 4.49 (br. dd, J ˆ 5.3, 5.1, 1 H); 4.23 (dd, J ˆ 9.5, 7.6, 1 H); 4.15 (t, J ˆ 7.6, 1 H); 4.09 (dd, J ˆ 9.5,
7.6, 1 H); 3.88 (dd, J ˆ 11.4, 9.4, 1 H); 3.75 (dd, J ˆ 11.4, 10.1, 1 H); 3.08 (dd, J ˆ 17.5, 5.3, 1 H); 3.05 (dd, J ˆ 17.5,
5.1, 1 H). 13C-NMR (150 MHz, (D6)acetone): 172.9 (s); 167.8 (s); 165.3 (s); 165.1 (s); 157.6 (s); 145.1 (s); 145.0
(s); 142.0 (s); 135.5 (s); 132.1 (d); 129.1 (d); 128.50 (d); 128.49 (d); 128.2 (d); 128.1 (d); 126.6 (d); 126.4 (d);
‡
120.7 (d); 67.7 (t); 62.3 (d); 55.6 (d); 52.5 (d); 51.0 (d); 48.0 (d); 47.8 (t); 36.6 (t). ESI-MS: 591.2 ([M ‡ Na] ),
‡
569.3 ([M ‡ H] ).
Benzyl
(1S,4R,5R)-4-(1,3-Dihydro-1,3-dioxo-2H-isoindol-2-yl)-7-oxo-2,6-diazabicyclo[3.2.0]heptane-2-
carboxylate. To a refluxing soln. (788) of 6 (12.0 g, 22.2 mmol) in MeCN (220 ml) and H2O (10 ml), a soln. of
K2S2O8 (48.0 g, 8 equiv., 0.18 mol) and Na2HPO4 ´ 2 H2O (63.1 g, 16 equiv., 0.36 mol) in H2O (60 ml), was added
in portions with vigorous stirring. Upon conversion (TLC control, 30 min), the orange soln. was immediately
cooled and the two-phase system made slightly alkaline by adding crystalline NaHCO3. The aq. phase was
extracted with AcOEt, the org. phase washed with sat. aq. NaHCO3 soln. and brine, dried (MgSO4), and
evaporated, and the residue purified by FC (hexane/AcOEt 1:4): off-white solid (7.1 g, 82%). TLC (hexane/
AcOEt 1:1): Rf 0.64. M.p. > 508 (dec.). [a]2D1 ˆ ‡6.1 (c ˆ 0.43, AcOEt). IR (KBr): 3300m, 3060w, 3028w, 2944w,
1
1780vs, 1760vs, 1722vs, 1714vs, 1696vs, 1600vs, 1497m. H-NMR (300 MHz, (D6)acetone): 7.87 (s, 4 H); 7.46 ±
7.30 (m, 6 H); 5.31 (br. s, 1 H); 5.19 (s, 2 H); 4.75 ± 4.27 (3m, 4 H). 13C-NMR (75 MHz, (D6)acetone): 169.2 (s);
165.4 (s); 154.4 (s); 137.8 (s); 135.2 (d); 133.1 (s); 129.7 (d); 129.2 (d); 128.5 (d); 123.9 (d); 68.9 (d); 67.6 (t); 55.9
‡
‡
(br. d); 54.5 (br. d); 53.4 (br. d); 52.9 (br. d); 44.4 (br. t). ESI-MS: 414.1 ([M ‡ Na] ), 392.4 ([M ‡ H] ).
2-Benzyl 6-(tert-Butyl) (1S,4R,5R)-4-(1,3-Dihydro-1,3-dioxo-2H-isoindol-2-yl)-7-oxo-2,6-diazabicyclo-
[3.2.0]heptane-2,6-dicarboxylate (11). To the foregoing product (6.2 g, 15.8 mmol) in anh. CH2Cl2 (120 ml)
and Et3N (2.2 ml, 15.8 mmol), (Boc)2O (6.9 g, 31.6 mmol) and DMAP (1.93 g, 15.8 mmol) were added. After
stirring for 1 h, the org. phase was washed with 1% aq. citric acid soln., sat. NaHCO3 soln., and brine and
evaporated. The residue was purified by FC (hexane/AcOEt 1 :1): yellowish solid (5.72 g, 74%). TLC (hexane/
AcOEt 1:1): Rf 0.63. M.p. 92 ± 938. [a]D21
ˆ
107.5 (c ˆ 0.43, AcOEt). IR (KBr): 2974w, 2936w, 1816vs, 1778s,
1725vs, 1711vs, 1696vs. 1H-NMR (300 MHz, (D6)acetone): 7.88 (s, 4 H); 7.45 ± 7.31 (m, 5 H); 5.41 (br. d, J ˆ 4.9,
1 H); 5.22 ± 5.17 (m, 2 H); 4.96 (ddd, J ˆ 10.8, 8.5, 5.5, 1 H); 4.84 (br. dd, J ˆ 5.5, 4.9, 1 H); 4.66 (t, J ˆ 10.8, 1 H);
4.23 (m, 1 H); 1.21 (s, 9 H). NOE: 5.41 (H C(1)) ! 4.84 (H C(5)); 5.20 (PhCH2) ! 7.43 (Ho, Z); 4.96
(H C(4)) ! 4.84 (H C(5)), 4.66 (Hpro-S C(3)); 4.84 (H C(5)) ! 5.41 (H C(1)), 4.96 (H C(4)); 4.66
(Hpro-s C(3)) ! 4.23 (Hpro-R C(3)); 4.23 (Hpro-R C(3)) ! 4.96 (H C(4)), 4.66 (Hpro-S C(3)). 13C-NMR
(75 MHz, (D6)acetone): 168.9 (s); 163.0 (s); 154.3 (s); 148.6 (s); 137.7 (s); 135.3 (d); 132.9 (s); 129.8 (d); 129.3
(d); 128.6 (d); 124.0 (d); 83.5 (s); 67.1 (d); 66.9 (t); 59.0 (d); 58.1 (d); 51.2 (d); 50.7 (d); 43.5 (t); 27.8 (q). ESI-
‡
‡
‡
MS: 513.9 ([M ‡ Na] ), 457.9 ([M ‡ Na tBu] ), 414.0 ([M ‡ Na tBu CO2] ).
Methyl (2S,3R,4R)-1-[(Benzyloxy)carbonyl]-3-{[(tert-butoxy)carbonyl]amino}-4-(1,3-dioxo-1,3-dihydro-
2H-isoindol-2-yl)prolinate (12). As described for 10, with 11 (3.2 g, 6.5 mmol), THF (120 ml), H2O (40 ml),
and Na2CO3 (0.83 mg, 7.8 mmol) (on workup acidification to pH 4.0, and drying with MgSO4): carboxylic acid.
‡
‡
‡
ES-MS: 564.3 ([M ‡ MeOH ‡ Na] ), 542.3 ([M ‡ MeOH ‡ H] ), 486.1 ([M ‡ MeOH ‡ H tBu] ), 442.1
‡
([M ‡ MeOH ‡ H tBu CO2] ). The corresponding ester was purified by FC (hexane/AcOEt 2 :3). TLC

460
Helvetica Chimica Acta ± Vol. 83 (2000)
Rf 0.82. M.p. 72 ± 748. [a]2D1 ˆ ‡11.7 (c ˆ 0.44, AcOEt). IR (KBr): 3420w (br.), 2960w, 2920w, 1775w, 1752w,
1724s, 1715s, 1610w, 1510w. 1H-NMR (300 MHz, (D6)acetone; (E)/(Z)-isomers): 7.88 (s, 4 H); 7.44 ± 7.32
(m, 5 H); 5.59 (2 br. d, 1 H); 5.21 ± 4.82 (m, 5 H); 4.61 (m, 1 H); 4.06 (m, 1 H); 3.72 (s, 3 H); 3.62 (s, 3 H); 1.15
(s, 9 H). 13C-NMR (75 MHz, (D6)acetone; (E)/(Z)-isomers): 170.4 (s); 169.9 (s); 169.2 (s); 155.7 (s); 155.3 (s);
154.7 (s); 137.8 (s); 137.7 (s); 135.4 (d); 132.7 (s); 129.3 (d); 128.7 (d); 128.6 (d); 128.5 (d); 124.0 (d); 79.9 (s); 67.6
(t); 62.6 (d); 62.3 (d); 54.9 (d); 53.9 (d); 52.4 (q); 52.2 (d); 51.7 (d); 44.6 (t); 44.2 (t); 28.2 (q); 28.0 (q). ESI-MS:
‡
‡
‡
546.3 ([M ‡ Na] ), 468.3 ([M ‡ H tBu] ), 424.2 ([M ‡ H tBu CO2] ).
Methyl (2S,3R,4R)-3-{[(tert-Butoxy)carbonyl]amino}-4-(1,3-dihydro-1,3-dioxo-2H-isoindol-2-yl)prolinate.
The foregoing product (0.52 g, 1.0 mmol) was stirred in DMF (5 ml) with 10% Pd/C under H2 for 3 d. DMF was
evaporated, the residue suspended in AcOEt and filtered through Celite, and the filtrate evaporated: white solid
(0.38 g, 100%). TLC (CH2Cl2/MeOH 95 :5): Rf 0.44. M.p. 68 ± 718. [a]2D0 ˆ ‡30.4 (c ˆ 0.39, AcOEt). IR (KBr):
3600 ± 3200w (br.), 2974m, 1780m, 1757s, 1725vs, 1715vs, 1693s, 1680s, 1632w, 1566w, 1550w, 1530m, 1519m.
1H-NMR (300 MHz, (D6)acetone): 7.91 (s, 4 H); 6.53 (d, J ˆ 9.3, 1 H); 5.31 (m, 1 H); 5.12 (d, J ˆ 6.6, 1 H); 5.06
(m, 1 H); 4.76 (dd, J ˆ 12.1, 9.0, 1 H); 3.91 (dd, J ˆ 12.1, 9.3, 1 H); 3.82 (s, 3 H); 1.18 (s, 9 H). 13C-NMR
(75 MHz, (D6)acetone): 171.6 (s); 169.3 (s); 155.6 (s); 135.2 (d); 132.7 (s); 123.8 (d); 79.3 (s); 64.3 (d); 56.5 (d);
‡
‡
‡
54.4 (d); 52.1 (q); 46.9 (t); 28.0 (q). ESI-MS: 412.2 ([M ‡ Na] ), 390.1 ([M ‡ H] ), 356.0 ([M ‡ Na tBu] ),
‡
334.1 ([M ‡ H tBu] ).
Methyl N2-[(Benzyloxy)carbonyl]-l-aspart-1-yl}-(2S,3R,4R)-3-{[(tert-butoxy)carbonyl]amino}-4-(1,3-di-
hydro-1,3-dioxo-2H-isoindol-2-yl)prolinate. To the foregoing product (0.69 g, 1.77 mmol), Z-Asp(OtBu)-OH
(0.68 g, 2.1 mmol), HATU (0.8 g, 2.1 mmol), and HOAt (0.24 g, 1.76 mmol) in anh. CH2Cl2 (20 ml), Pr2EtN
i
(0.9 ml, 5.26 mmol) was added, and the soln. was stirred at r.t. for 24 h1). After washing with 10% aq. citric acid
soln., sat. aq. NaHCO3 soln., and brine, the org. phase was dried (MgSO4) and evaporated. The residue was
purified by FC (hexane/AcOEt 1:1): white solid (0.78 g, 63%). TLC (hexane/AcOEt 1:1): Rf 0.48. M.p. 78 ±
808. [a]2D0 ˆ ‡9.4 (c ˆ 0.47, AcOEt). IR (KBr): 3440 ± 3200w (br.), 2979m, 2932w, 1780m, 1732vs, 1715vs, 1668s,
1661s, 1652s, 1515s, 1505s. 1H-NMR (300 MHz, (D6)acetone): 7.88 (m, 4 H); 7.39 ± 7.27 (m, 5 H); 6.74 (d, J ˆ 8.4,
1 H); 5.59 (br. d, 1 H); 5.10 ± 4.88 (m, 7 H); 4.39 (m, 1 H); 3.68 (s, 3 H); 2.83 (dd, J ˆ 16.1, 6.0, 1 H); 2.57
(dd, J ˆ 16.1, 7.7, 1 H); 1.47 (s, 9 H); 1.14 (s, 9 H). 13C-NMR (75 MHz, (D6)acetone): 170.8 (s); 169.7 (s); 169.03
(s); 169.02 (s); 157.3 (s); 156.2 (s); 138.1 (s); 135.3 (d); 132.6 (s); 129.1 (d); 128.6 (d); 123.9 (d); 81.3 (s); 79.8 (s);
66.8 (t); 62.1 (d); 53.0 (d); 52.7 (d); 52.2 (q); 50.4 (d); 44.2 (t); 38.6 (t); 28.10 (q); 28.06 (q). ESI-MS: 717.5
‡
‡
‡
‡
([M ‡ Na] ), 695.5 ([M ‡ H] ), 661.4 ([M ‡ Na tBu] ), 639.3 ([M ‡ H tBu] ), 605.4 ([M ‡ Na
2
tBu]
‡
t
‡
t
‡
), 583.4 ([M ‡ H 2 Bu] ), 539.2 ([M ‡ H 2 Bu CO2] ). Anal. calc. for C35H42N4O11 (694.75): C 60.5,
H 6.1, N 8.1; found: C 60.3, H 6.2, N 7.9.
tert-Butyl
(3S,7R,8R,8aS)-8-{[(tert-Butoxy)carbonyl]amino}-7-(1,3-dihydro-1,3-dioxo-2H-isoindol-2-
yl)octahydro-1,4-dioxopyrrolo[1,2-a]pyrazine-3-acetate. The foregoing dipeptide (0.31 g, 0.45 mmol) in DMF
(15 ml) was stirred with 10% Pd/C (30 mg) under H2 for 5 d. DMF was evaporated, the residue suspended in
CH2Cl2, the mixture filtered through Celite, and the filtrate evaporated: white solid (0.24 g, 95%). TLC (CH2Cl2/
MeOH 95 :5): Rf 0.33. M.p. 1918 (dec.). [a]2D0 ˆ ‡21.8 (c ˆ 0.23, AcOEt). IR (KBr): 3700 ± 3200m (br.), 2977m,
2929w, 1779m, 1720vs, 1695vs, 1684vs, 1669vs, 1526s. 1H-NMR (600 MHz, CDCl3): 7.83 (m, 2 H); 7.71 (m, 2 H);
6.72 (br. s, 1 H); 6.20 (d, J ˆ 8.8, 1 H); 5.05 (m, 1 H); 4.93 (dd, J ˆ 12.2, 7.8, 1 H); 4.85 (m, 1 H); 4.51 (d, J ˆ 4.5,
1 H); 4.37 (br. dd, 1 H); 3.64 (dd, J ˆ 12.2, 9.6, 1 H); 3.22 (dd, J ˆ 17.2, 6.4, 1 H); 3.18 (dd, J ˆ 17.2, 3.4, 1 H); 1.49
(s, 9 H); 1.13 (s, 9 H). 13C-NMR (75 MHz, CDCl3): 170.5 (s); 168.3 (s); 164.1 (s); 155.9 (s); 134.4 (d); 131.9 (s);
123.7 (d); 82.8 (s); 79.7 (s); 60.2 (d); 54.8 (d); 52.3 (d); 51.1 (d); 43.4 (t); 37.8 (t); 28.3 (q); 28.2 (q). ES-MS: 551.4
‡
‡
‡
‡
t
([M‡Na] ), 529.3 ([M‡H] ), 473.1 ([M‡H tBu] ), 429.1 ([M‡H tBu CO2] ), 373.2 ([M‡H 2 Bu
‡
CO2] ).
tert-Butyl (3S,7R,8R,8aS)-7-Acetamido-8-{[(tert-butoxy)carbonyl]amino}octahydro-1,4-dioxopyrrolo[1,2-
a]-pyrazine-3-acetate (13). A soln. of the foregoing product (50 mg, 0.095 mmol) and MeNHNH2 ´ H2O (24 ml,
3.6 equiv., 0.34 mmol) in EtOH (2 ml) was stirred at 808 for 1 h. The solvent was evaporated, and iPr2EtN (16 ml,
0.095 mmol), CH2Cl2 (2 ml), and Ac2O (18 ml, 2 equiv., 0.19 mmol) were added. After stirring for 15 min, the
suspension was filtered, CH2Cl2 evaporated, and the residue purified by HPLC (gradient 5 ± 95% MeCN/H2O
(‡0.1% CF3COOH) over 15 min, then 95% MeCN/H2O (‡0.1% CF3COOH) for 5 min); tR 11.9 min): 13
1
(29 mg, 70% over 2 steps). White solid. H-NMR (300 MHz, (D6)acetone): 4.69 (m, 1 H); 4.58 (m, 2 H); 4.34
(dd, 1 H); 3.67 (dd, J ˆ 11.8, 9.0, 1 H); 3.38 (dd, J ˆ 11.8, 10.2, 1 H); 2.97 (dd, J ˆ 17.4, 4.0, 1 H); 2.82 (dd, J ˆ
‡
‡
17.4, 4.4, 1 H); 1.87 (s, 3 H); 1.46 (s, 9 H); 1.40 (s, 9 H). ESI-MS: 903.5 ([2M ‡ Na] ), 463.3 ([M ‡ Na] ), 441.3
‡
‡
‡
([M ‡ H] ), 341.1 ([M ‡ H tBu CO2] ), 285.2 ([M ‡ H 2 tBu CO2] ).
Cyclo{l-alanyl-l-alanyl-[(3S,7R,8R,8aS)-7-(benzoylamino)-8-aminooctahydro-1,4-dioxopyrrolo[1,2-a]pyra-
zine-3-acetyl]-l-alanyl-l-asparaginyl-l-prolyl-l-asparaginyl} (Cyclo(-Ala-Ala-temp-Ala-Asn-Pro-Asn; 3). Fmoc-

Helvetica Chimica Acta ± Vol. 83 (2000)
461
Fig. 9. Molecular topology building block of the template used for simulations of 3 with GROMOS96 and the
43A1 force field. a) Atom names X, atom numbering nX, integer atom code Xn; b) bond-type codes (thin),
bond-angle-type codes (italics).
temp-OH 1 was coupled to TentagelTM-S-AC resin (1 equiv.) (Rapp Polymere, Tübingen; 0.27 mmol/g) using
CIP1) for activation in pyridine/CH2Cl2 1:1. The solid-phase peptide synthesis was performed by chain
elongation with Fmoc-Ala-OH, Fmoc-Ala-OH, Fmoc-Asn(Mtt)-OH, Fmoc-Pro-OH, Fmoc-Asn(Mtt)-OH,
Fmoc-Ala-OH (0.36 mmol each) using HOBt/HBTU for activation1). Upon completion of chain assembly, the
N-terminal Fmoc protecting group was removed with 20% piperidine/DMF, and the resin was washed with
MeOH and CH2Cl2. Cleavage of the peptide from the resin was performed with 1% CF3COOH/CH2Cl2 (30 ml)
for 15 min, repeated four times. Each filtrate was neutralized with pyridine (3 ml in total) in CH2Cl2 (10 ml in
total), and the combined org. phase was evaporated to give linear, protected peptide (20.2 mg, 67%) after
purification by HPLC (20% MeCN/H2O (‡0.1% CF3COOH) for 4 min, gradient 20±70% within 0.5 min, 70±95%
‡
within 10.5 min, and 95% for 5 min). ESI-MS: 1397.98 ([M‡ Na] ), 710.57 ([M‡ H‡ Na]2‡), 699.70 ([M‡ 2H]2‡).
For cyclization, the linear peptide (19 mg, 12.6 mmol) in 1% iPr2EtN/DMF was stirred with HOAt/HATU1)
(2 equiv.) overnight at r.t. Evaporation of DMF and purification by HPLC (20% MeCN/H2O (‡0.1%
CF3COOH) for 4 min, gradient 20 ± 70% within 0.5 min, 70 ± 95% within 10.5 min, and 95% for 5 min) afforded
‡
the side-chain-protected cyclic peptide (7 mg, 40%). ESI-MS: 1379.76 ([M ‡ H] ).
The protected cyclic peptide (6 mg) was treated with CF3COOH/CH2Cl2/iPr3SiH 35 :60 :5 (8.4 ml) at r.t. for
1 h. Evaporation, precipitation with iPr2O, and purification by HPLC (2% MeCN/H2O (‡0.1% CF3COOH) for

462
Helvetica Chimica Acta ± Vol. 83 (2000)
Table 3. Codes for Improper Dihedrals and Proper Torsional Dihedrals Used for the GROMOS96 Molecular
Topology Building Block File of the Template in 3
Improper dihedrals
Atom sequence
Proper torsional dihedrals
Atom sequence
Type
code
Type
code
i
j
k
l
i
j
k
l
1
3
4
5
7
9
10
12
9
1
4
20
4
9
10
9
3
22
3
7
5
2
1
5
6
8
7
1
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
1
1
2
1
1
2
1
1
3
4
5
3
4
20
21
21
22
23
25
25
27
27
1
1
3
4
5
1
3
4
5
7
3
4
5
7
9
10
21
22
23
1
25
27
29
21
31
33
22
4
19
17
19
4
12
12
10
16
14
18
18
14
16
10
12
16
22
3
7
4
9
5
16
14
18
18
16
14
11
13
15
17
19
29
21
24
26
28
30
32
20
21
22
3
23
25
27
29
28
31
21
17
19
19
17
20
4
19
20
17
20
4
9
20
21
22
22
23
25
27
27
28
29
12
10
14
16
9
9
10
12
10
12
14
16
18
21
22
23
25
27
29
31
9
18
18
14
20
23
22
27
25
27
28
25
23
29
21
33
5 min, gradient 2 ± 10% within 0.5 min, 10 ± 40% within 39.5 min, 40 ± 95% within 0.5 min, and 95% for 5 min)
1
‡
gave 3 (1.8 mg, 40%). H-NMR (600 MHz, (D6)DMSO, 305 K): Tables 1 and 2. ESI-MS: 867.68 ([M ‡ H] ).
Amino acid analysis (molar ratio): Ala 3.00, Asx 3.27, Pro 0.90 (modified proline derivative from template not
determined).
NMR Experiments, Structure Calculations, and Simulations. 1D and 2D 1H-NMR spectra: 600 MHz;
Bruker-AMX600 spectrometer, typical peptide concentration 10 mg/ml in (D6)DMSO; analysis of 2D spectra
by ꢀFelixꢁ software (MSI, San Diego).
To derive NOE distance restraints, it was assumed that the initial rate approximation is valid and that each
peptide rotates as a single isotropic rotor. The NOEs were determined from NOESY spectra measured at 305 K
with mixing times of 40, 80, 120, and 250 ms, with 2048 Â 256 data points, and zero-filling to 4096 Â 2048.
Transformation was performed with a sine-bell function. Cross-peak volumes were determined by integration,
and build-up curves were checked to ensure a smooth exponential increase in peak intensity for all NOEs used
in deriving distance restraints. The relative cross-peak volumes were assumed to be proportional to r 6, and were
used to derive distance restraints for simulated annealing (SA) calculations, performed with the methods
described in detail elsewhere [25].
For MD simulations with and without TA-DR, the GROMOS96 suite of programs was used with the 43A1
force field [26]. The dynamics of 3 in DMSO was studied at 300 K at 1 atm pressure and with periodic boundary
conditions. The temp. was maintained by weak coupling (tT ˆ 0.1 ps) to a temp. bath. The system contained the
mimic and 314 DMSO molecules in a truncated octahedral box. The upper distance restraints were the exact
values obtained from NOE build-up curves, where necessary with pseudoatom corrections, a memory decay
time tdr ˆ 50 ps, and a force constant Kdr ˆ 1000 kJ mol 1 nm 2. A molecular building block for the template was
constructed using the existing GROMOS96 atom-type and force-field parameters. The parameters used are
shown in Fig. 9 and Table 3. The SHAKE algorithm was used to maintain bond lengths with a relative precision

Helvetica Chimica Acta ± Vol. 83 (2000)
463
Table 4. Crystallographic Data for 12
Crystallized from
AcOEt
Empirical formula
Formula weight [g mol
Crystal colour, habit
Crystal dimensions [mm]
Temperature [K]
C27H29N3O8 ´ 0.5 H2O
1
]
532.55
colorless, prism
0.20 Â 0.22 Â 0.50
173 (1)
monoclinic
C2
Crystal system
Space group
Z
4
Reflections for cell determination
25
2q range for cell determination [8]
Unit-cell parameters a [Š]
22 ± 38
27.275 (7)
6.017 (6)
17.068 (5)
110.83 (2)
2618 (2)
1.351
b [Š]
c [Š]
b [8]
V [Š3]
3
Dx [g cm
]
1
m(MoKa) [mm
2q(max) [8]
]
0.101
55
Total reflections measured
Symmetry-independent reflections
Reflections used (I > 2s(I))
Parameters refined
3366
3295
2303
348
Final R
0.0591
0.0522
2.169
wR
Goodness of fit s
7
Secondary extinction coefficient
4(5) ´ 10
Final Dmax/s
D1 (max; min) [e Š
0.0001
0.26; 0.25
3
]
of 10 4, and the integrator time step was 0.002 ps. Nonbonded interactions evaluated at every step were within a
short range cut-off of 9 Š. For long-range interactions, calculated every 5 steps, the cut-off was 15 Š. Structures
were saved for analysis every 100 steps (0.2 ps). After short simulations to relax the solute and solvent, the
simulations with and without TA-DR were each run for 2 ns.
Crystal-Structure Determination for 122). All measurements were conducted at low temp. on a Rigaku-
AFC5R diffractometer using graphite-monochromated MoKa radiation (l 0.71069 Š) and a 12-kW rotating
anode generator. The intensities were collected using w/2q scans, and three standard reflections, which were
measured after every 150 reflections, remained stable throughout the data collection. The intensities were
corrected for Lorentz and polarization effects, but not for absorption. The structure was solved by direct
methods using SHELXS86 [27] which revealed the positions of all non-H-atoms of the prolinate molecule 12
plus a H2O molecule which sits on a two-fold axis. The non-H-atoms were refined anisotropically. The unique H-
atom of the H2O molecule was located in a difference electron density map, and its position was held fixed. All
other H-atoms were fixed in geometrically calculated positions (d(C H) ˆ 0.95 Š). All H-atoms were assigned
fixed isotropic displacement parameters with a value equal to 1.2Ueq of the atom to which each was bonded. A
correction for secondary extinction was applied. Refinement was carried out on F using full-matrix least-squares
procedures which minimized the function Sw(j Fo j j Fc j )2, where 1/w ˆ s2(Fo) ‡ (0.005Fo)2. All calculations
were performed using the TEXSAN crystallographic software package [28]. The data collection and refinement
2
)
Crystallographic data (excluding structure factors) for the structure of 12 have been deposited with the
Cambridge Crystallographic Data Centre as supplementary publication No. CCDC-136003. Copies of the
data can be obtained, free of charge, on application to the CCDC, 12 Union Road, Cambridge CB21EZ,
UK (fax: ‡44-(0)-1223-336033; email: deposit@ccdc.cam.ac.uk).

464
Helvetica Chimica Acta ± Vol. 83 (2000)
parameters are listed in Table 4, and a view of the molecule, produced with ORTEPII [29], is shown in Fig. 2.
The enantiomorph was assigned from the known absolute configuration of the starting material. The H2O
molecule forms H-bonds with the (benzyloxy)carbonyl O-atom of each of two symmetry-related prolinate
molecules.
REFERENCES
[1] D. P. Fairlie, M. L. West, A. K. Wong, Curr. Med. Chem. 1998, 5, 29.
[2] D. Obrecht, M. Altorfer, J. A. Robinson, Adv. Med. Chem. 1999, 4, 1.
[3] C. Bisang, L. Jiang, E. Freund, F. Emery, C. Bauch, H. Matile, G. Pluschke, J. A. Robinson, J. Am. Chem.
Soc. 1998, 120, 7439.
[4] F. Emery, C. Bisang, M. Favre, L. Jiang, J. A. Robinson, Chem. Commun. 1996, 2155.
[5] H. J. Dyson, A. C. Satterthwait, R. A. Lerner, P. E. Wright, Biochemistry 1990, 29, 7828.
[6] C. Bisang, C. Weber, J. Inglis, C. A. Schiffer, W. F. van Gunsteren, I. Jelesarov, H. R. Bosshard, J. A.
Robinson, J. Am. Chem. Soc. 1995, 117, 7904.
[7] D. A. Herrington, D. F. Clyde, G. Losonsky, M. Cortesia, J. R. Murphy, J. Davis, S. Baqar, A. M. Felix, E. P.
Heimer, D. Gillessen, E. Nardin, R. S. Nussenzweig, V. Nussenzweig, M. R. Hollingdale, M. M. Levine,
Nature (London) 1987, 328, 257.
[8] V. Nussenzweig, R. S. Nussenzweig, Adv. Immunol. 1989, 45, 283.
[9] M. E. Pfeifer, J. A. Robinson, Chem. Commun. 1998, 1977.
[10] I. Heinze-Krauss, P. Angehrn, R. L. Charnas, K. Gubernator, E.-M. Gutknecht, C. Hubschwerlen, M.
Kania, C. Oefner, M. G. P. Page, S. Sogabe, J.-L. Specklin, F. Winkler, J. Med. Chem. 1998, 41, 3961.
[11] C. Hubschwerlen, G. Schmid, Helv. Chim. Acta 1983, 66, 2206.
[12] C. Hubschwerlen, Synthesis 1986, 961.
[13] O. Mitsunobu, M. Wada, T. Sano, J. Am. Chem. Soc. 1972, 94, 679.
[14] E. Atherton, R. C. Sheppard, ꢀSolid Phase Peptide Synthesis ± A Practical Approachꢁ, IRL Press, Oxford,
1989.
[15] K. Wüthrich, ꢀNMR of Proteins and Nucleic Acidsꢁ, Wiley-Interscience, New York, 1986.
[16] D. B. Davies, M. A. Khaled, J. Chem. Soc., Perkin Trans. 2 1976, 187.
[17] D. B. Davies, M. A. Khaled, J. Chem. Soc., Perkin Trans. 2 1976, 1238.
[18] A. P. Nanzer, W. F. van Gunsteren, A. E. Torda, J. Biomol. NMR 1995, 6, 313.
[19] A. E. Torda, R. M. Scheek, W. F. van Gunsteren, Chem. Phys. Lett. 1989, 157, 289.
[20] B. L. Sibanda, T. L. Blundell, J. M. Thornton, J. Mol. Biol. 1989, 206, 759.
[21] B. L. Sibanda, J. M. Thornton, J. Mol. Biol. 1993, 229, 428.
[22] C. Mattos, G. A. Petsko, M. Karplus, J. Mol. Biol. 1994, 238, 733.
[23] M. Favre, K. Moehle, L. Jiang, B. Pfeiffer, J. A. Robinson, J. Am. Chem. Soc. 1999, 121, 2679.
[24] D. Gramberg, C. Weber, R. Beeli, J. Inglis, C. Bruns, J. A. Robinson, Helv. Chim. Acta 1995, 78, 1588.
[25] C. Bisang, C. Weber, J. A. Robinson, Helv. Chim. Acta 1996, 79, 1825.
[26] W. F. van Gunsteren, S. R. Billeter, A. A. Eising, P. H. Hünenberger, P. Krüger, A. E. Mark, W. R. P. Scott,
I. G. Tironi, ꢀBiomolecular Simulation: The GROMOS96 Manual and User Guideꢁ, Hochschulverlag AG
an der ETH Zürich, Zürich, 1996.
[27] G. M. Sheldrick, ꢀSHELXS86ꢁ, Acta Crystallogr., Sect. A 1990, 46, 467.
[28] ꢀTEXSAN: Single Crystal Structure Analysis Software, Version 5.0ꢁ, Molecular Structure Corporation,
The Woodlands, Texas, 1989.
[29] C. K. Johnson, ꢀORTEPII. Report ORNL-5138ꢁ, Oak Ridge National Laboratory, Oak Ridge, Tennessee,
1976.
Received October 21, 1999