Diversity-oriented synthesis of a drug-like system displaying the distinctive N→C=O interaction

Article abstract of DOI:10.1021/jo800719j

(Chemical Equation Presented) This study describes the syntheses and characterization of two hydrazino ureas. These fold into a six-membered ring by virtue of the infrequently observed ^δ+N→C=O ^δ- interaction when solvated by polar pr

Full text of DOI:10.1021/jo800719j

Diversity-Oriented Synthesis of a Drug-Like System Displaying the
Distinctive NfCdO Interaction
Michael Waibel and Jens Hasserodt*
Laboratoire de Chimie, UMR CNRS 5182, UniVersite´ de Lyon-ENS, 46 alle´e d’Italie, 69364 Lyon cedex
07, France
jens.hasserodt@ens-lyon.fr
ReceiVed April 1, 2008
This study describes the syntheses and characterization of two hydrazino ureas. These fold into a six-
membered ring by virtue of the infrequently observed δ+NfCdOδ- interaction when solvated by polar
protic media. The highly polar functional group resulting from this interaction is hypothesized to especially
reproduce electronic but also steric features of the transition states of peptide hydrolysis. The urea moiety
constitutes an additional key element of modern HIV-1 protease (HIV PR) inhibitors and is meant to
interact with the enzyme flaps. We have developed an efficient, convergent synthetic route to enantiopure
compounds that uses CDI to couple two independent building blocks, one derived from amino acids and
the other one from easily accessible hydrazines. It is thus amenable to rapid generation of diversity in
order to screen for novel HIV PR inhibitors. A complete study using one- and two-dimensional NMR as
well as UV spectroscopy confirmed the sole existence of the cyclic constitution of target compounds 7
and 8 in methanol. Total reversal to the linear aldehydic form is observed upon passage to apolar media.
Introduction
An alternative approach has been the use of heterocyclic
nonpeptidic templates, such as a cyclic urea (1)9 or a 5,6-
dihydro-4-hydroxy-2-pyrone (2).10,11 While the hydroxyl group(s)
interact(s) with the catalytic aspartyl diad of HIV PR, the
carbonyl function on the opposite side engages in hydrogen
bonding with two protein flaps of the enzyme. Incorporation of
heterocycles displaying such features into an inhibitor can yield
higher affinities due to a favorable entropic effect in that (a)
the carbonyl function replaces a key enzyme-bound water
molecule that is usually attached to the flaps,9,12 and (b) fewer
degrees of freedom have to be frozen during complexation, due
to the elevated rigidity of the core.9,13 Repression of confor-
mational freedom is also known to lead to better oral bioavail-
The design of transition state (TS) analogs of enzymatic
reactions has proved to be a powerful strategy for the develop-
ment of potent enzyme inhibitors on the one hand as well as
for the creation of artificial catalysts on the other.1–3 Inhibitors
of aspartic proteases (PR) have great importance for the
treatment of a variety of diseases including malaria, cancer,
hypertension, AIDS, and Alzheimer’s disease.4,5A common
approach in designing inhibitors for HIV-1 protease (HIV PR)
has been the incorporation of a TS isostere into a peptidomimetic
that reflects the linear shape of the biologically occurring peptide
substrate. Several isosteric moieties have been explored, for
example, hydroxyethylamine, hydroxyethylene, statine, phos-
phinate, reduced amide, R-ketoamide, and more recently a
silicon-based unit.6–8
(8) Ghosh, A. K.; Chapsal, B. D.; Weber, I. T.; Mitsuya, H. Acc. Chem.
Res. 2008, 41, 78–86.
(9) DeLucca, G. V.; EricksonViitanen, S.; Lam, P. Y. S. Drug DiscoVery
Today 1997, 2, 6–18.
(1) Hollfelder, F.; Kirby, A. J.; Tawfik, D. S. Nature 1996, 383, 60–63.
(2) Mader, M. M.; Bartlett, P. A. Chem. ReV. 1997, 97, 1281–1301.
(3) Wolfenden, R. Bioorg. Med. Chem. 1999, 7, 647–652.
(10) Hagen, S.; Prasad, J. V. N. V.; Tait, B. D. AdV. Med. Chem. 2000, 5,
159–195.
(11) Thaisrivongs, S.; Strohbach, J. W. Biopolymers 1999, 51, 51–58.
(12) Dunitz, J. D. Science 1994, 264, 670–670.
(4) Abbenante, G.; Fairlie, D. P. Med. Chem. 2005, 1, 71–104.
(5) De Clercq, E. Nat. ReV. Drug DiscoV. 2007, 6, 1001–1018.
(6) Brik, A.; Wong, C. H. Org. Biomol. Chem. 2003, 1, 5–14.
(7) Bursavich, M. G.; Rich, D. H. J. Med. Chem. 2002, 45, 541–558.
(13) Dunn, B. M. Chem. ReV. 2002, 102, 4431–4458.
(14) Veber, D. F.; Johnson, S. R.; Cheng, H. Y.; Smith, B. R.; Ward, K. W.;
Kopple, K. D. J. Med. Chem. 2002, 45, 2615–2623.
10.1021/jo800719j CCC: $40.75
Published on Web 07/16/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 6119–6126 6119

Waibel and Hasserodt
ability.14 Generally, the search for potent enzyme inhibitors as
well as the design of artifical enzymes has greatly benefited
from two complementary strategies, one based on structural
considerations and the other on the generation and screening
of combinatorial libraries.15,16
We have been interested in exploring the moiety resulting
from a tertiary amine-carbonyl interaction (δ+NfCdOδ-, see
examples 3,17 4,18 and 519) for its capacity to reproduce steric
and electronic properties of the transition states of peptide
hydrolysis (for comparison see the second TS (6) of the preferred
dipeptide sequence of HIV PR20,21).
hybridization between sp2 and sp3; (c) a C-O bond order
between 1 and 2; and (d) partial charges on the oxygen and the
nitrogen, resulting in an enhanced dipole moment.25 This
strongly polarized unit may thus form particularly strong
hydrogen bonds and electrostatic interactions with the catalytic
diad of the two aspartate residues. Our first-generation prototype
5 showed the δ+NfCdOδ- interaction to a degree of 70% at
20 °C in MeOH.19 The resulting non-zero fraction of the
(unfolded) aldehydic form not only complicates analysis of the
system but also rules out any exploration as a drug candidate,
since aldehydes are justly regarded as too reactive toward
nitrogen nucleophiles present in living organisms.
In this study it was our objective to identify a core unit that
would show total δ+NfCdOδ- interaction, while offering a
nonreactive carbonyl group on the opposite side of the molecule
to satisfy the hydrogen-bonding partners on the enzyme flaps.
This core unit should be of reduced molecular complexity and
should be easily assembled to allow for the subsequent hunt
for a promising lead structure by a combined computational
and combinatorial tactic.
Historically, the δ+NfCdOδ- interaction was observed in a
rare class of alkaloids as a transannular contact across medium-
sized rings.22,23 It has been described as a through-space
homoconjugation where electron density is shifted from the n
nitrogen orbital to the carbonyl π* orbital,24 giving rise to an
enhanced dipole moment and a hypsochromic shift of the
carbonyl UV absorption.25,26 Studies on natural and synthetic
examples have shown that the interaction is favored in polar
protic media and when incorporated into preorganized, often
multicyclic frameworks.17,19,27 Recently, the δ+NfCdOδ-
moiety has featured prominently in the total synthesis of the
marine natural product (-)-sarain A (4).18
Results and Discussion
Compound Design and Synthetic Strategy. A urea moiety
not only proved to interact well with the flaps but offers the
advantage to largely rigidify a cyclic system. Grafting an
aldehyde group and a tertiary amine moiety on each nitrogen,
respectively, the preorganizing effect of a Z/Z-configured urea
should greatly favor the δ+NfCdOδ- interaction. Also, use of
a tertiary hydrazine as the nitrogen donor was expected to favor
the interaction due to its enhanced nucleophilicity and could
be introduced by creating a urea hydrazide. We quickly realized
that such features would not allow us to assemble a five-
membered heterocycle resulting from the δ+NfCdOδ- interac-
tion, but a six-membered one. Yet δ+NfCdOδ- interactions
have never been observed in six-membered rings outside the
context of multicyclic systems. It remained to be seen whether
such a compound shows a strong propensity to fold.
From a synthetic point of view, the urea moiety may
ultimately be very useful for the creation of a combinatorial
library, since rapid, convergent assembly from three building
blocks can be envisaged (Scheme 1). We chose to connect two
types of building blocks via the urea carbonyl group in which
one contains the (protected) carbonyl acceptor and the other
one the tertiary-nitrogen donor in the form of a hydrazine.
Numerous substituants (R1-R6) may be introduced. However,
at least two limitations need to be put forward here: (a) not
unexpectedly, we have found that if R6 represents a hydrogen
With respect to the strategy of transition-state mimicry, we
find the following of its characteristics particularly attractive:
(a) an N-C bond order between 0 and 1; (b) a carbon
(15) Maly, D. J.; Huang, L.; Ellman, J. A. ChemBioChem 2002, 3, 16–37.
(16) Reymond, J. L. Angew. Chem., Int. Ed. 2004, 43, 5577–5579.
(17) McCrindle, R.; McAlees, A. J. J. Chem. Soc., Chem. Commun. 1983,
61–62.
(18) Becker, M. H.; Chua, P.; Downham, R.; Douglas, C. J.; Garg, N. K.;
Hiebert, S.; Jaroch, S.; Matsuoka, R. T.; Middleton, J. A.; Ng, F. W.; Overman,
L. E. J. Am. Chem. Soc. 2007, 129, 11987–12002.
(19) Gautier, A.; Pitrat, D.; Hasserodt, J. Bioorg. Med. Chem. 2006, 14, 3835–
3847.
(20) Okimoto, N.; Tsukui, T.; Hata, M.; Hoshino, T.; Tsuda, M. J. Am. Chem.
Soc. 1999, 121, 7349–7354.
(21) Piana, S.; Bucher, D.; Carloni, P.; Rothlisberger, U. J. Phys. Chem. B
2004, 108, 11139–11149.
(22) Kermack, W. O.; Robinson, R. J. Chem. Soc. 1922, 121, 427–441.
(23) Leonard, N. J. Rec. Chem. Prog. 1956, 17, 243–257.
(24) Rademacher, P. Chem. Soc. ReV. 1995, 24, 143–150.
(25) Leonard, N. J.; Morrow, D. F.; Rogers, M. T. J. Am. Chem. Soc. 1957,
79, 5476–5479.
(26) Leonard, N. J.; Oki, M. J. Am. Chem. Soc. 1955, 77, 6239–6240.
(27) (a) Davies, J. E.; Kirby, A. J.; Komarov, I. V. HelV. Chim. Acta 2003,
86, 1222–1233. (b) Kirby, A. J.; Komarov, I. V.; Bilenko, V. A.; Davies, J. E.;
Rawson, J. M. Chem. Commun. 2002, 2106–2107. (c) Carroll, J. D.; Jones, P. R.;
Ball, R. G. J. Org. Chem. 1991, 56, 4208–4213. (d) Pilme, J.; Berthoumieux,
H.; Robert, V.; Fleurat-Lessard, P. Chem.-Eur. J. 2007, 13, 5388–5393. (e)
Bell, M. R.; Archer, S. J. Am. Chem. Soc. 1960, 82, 151–155.
6120 J. Org. Chem. Vol. 73, No. 16, 2008

DistinctiVe δ+NfCdOδ- Interaction
SCHEME 1. Retrosynthetic Assembly of the Target and
Opportunities for Generating Diversity
commonly reduced with NaBH4 or NaCNBH3, their application
to conjugated hydrazones has not been described. The common
reagent for this task is reported to be borane,33 which is also
known for its capacity to reduce amides.34 Since we intended
to apply the same reduction step to the synthesis of the second
target compound of this study, bearing an amide moiety (8,
Scheme 4), we searched for milder conditions to prevent amide
reduction. Catalytic hydrogenation at room temperature and
atmospheric pressure gave low yields. LC-MS monitoring of
the reaction indicated that the desired product was formed within
a few hours, but during the same time it underwent side reactions
to a significant degree. Despite its literature absence, we then
attempted the application of NaCNBH3 under neutral conditions
and also in the presence of stoichiometric amounts of acetic
acid, to no avail. On the other hand, using a large excess of
acetic acid at room temperature produced the desired hydrazine
14 in only 30 min.
The reaction of primary hydrazines with carbamoylimida-
zoles, such as 12, has been reported, but secondary hydrazine
14 did not react with compound 12, even when refluxing the
mixture.30 The imidazole unit was thus activated with excess
MeI to produce a highly reactive imidazolium intermediate.31
Subsequently, treatment of activated 12 with 14 gave hydrazino
urea 15 in a convenient overnight reaction. The three-step
reaction sequence from 11 to 15 (reaction of 11 with CDI,
activation with MeI, and reaction with 14) effectively couples
the amino acid with the hydrazine building block and may be
regarded as central to the synthesis of the target system. It offers
an overall yield of 71%, while requiring just one purification
step by column chromatography for 15. Removal of the dimethyl
acetal protecting group was achieved by exposure to in situ
generated TMSI,35 furnishing desired target compound 7 in
excellent yield. The E/Z-configuration of 7 as depicted in
Scheme 2 is the only one observed (see section on configura-
tional analysis). All reaction conditions applied in the synthesis
of 7 are also included in that of 8 (see below). Since no
diastereomer formation (racemization) was observed in the
preparation of 8, we suggest that racemization also does not
occur in the assembly of 7.
Synthesis of Prolyl Hydrazino Urea 8. In order to demon-
strate the potential for structural diversification, we established
the synthesis of a second target compound (8, Scheme 4) where
the hydrazine unit is derived from an amino acid (proline). By
synthesizing 8, we would simultaneously be able to prove that
ring sizes other than that of a piperidyl ring give equally reliable
folding and that sterically more demanding substitution patterns
on this ring do not impair the δ+NfCdOδ- interaction. We
would also show that amino acid derivatives are a suitable
starting material to establish elevated molecular complexity in
our system. The preparation of 8 necessitated the elaboration
of a synthetic sequence toward the second hydrazine building
block. The common way of making hydrazino acids from amino
acids is to reduce nitrosated amino acids by use of Zn/HOAc.36
Even though this procedure has been for years the only reported
method for this task, no explanation has been given for the low
yields that are generally disclosed. Nevertheless, we embarked
on its application for its straightforward character. Nitroso
atom, a spontaneous cyclization reaction takes place, furnishing
a five-membered ring that displays no δ+NfCdOδ- interaction;
we thus concluded that R6 has to be a substituent other than
hydrogen; (b) for the time being, we only succeeded in our
synthetic efforts if R3 ) H, most likely for steric reasons.
The aldehyde building block can be conveniently generated
from an amino acid precursor, thus allowing for an enormous
choice of substitution and stereoconfiguration patterns in view
of the market size of this commodity. We initially chose
phenylalanine as this precursor. The hydrazine building block
may be generated (a) from secondary amines of any kind28 with
subsequent reductive alkylation of the newly introduced primary
nitrogen, or (b) by consecutive alkylation of commercially
available hydrazines,29 once again demonstrating the immense
potential of this scheme for the generation of diversity. The
use of carbonyl di-imidazole (CDI) as the connecting reagent
appeared ideal for two reasons: (a) the desired urea function is
formed without further manipulations; (b) two independent
nucleophiles can be introduced consecutively, since the product
of the initial step (e.g., 12, Scheme 2) is much less reactive
than CDI itself and will not react with a second amino
nucleophile.30 The remaining imidazole unit can however be
methylated to produce the corresponding, highly reactive
imidazolium salt that was shown to react with various nucleo-
philes under mild conditions.31 In the following, we describe
the synthesis and the configurational and constitutional analysis
of two independent manifestations (7, Scheme 2, and 8, Scheme
4) of the generic structure in Scheme 1.
Synthesis of Piperidyl Hydrazino Urea 7. In the synthesis
of the phenylalanine part of 7 (Scheme 2), aldehyde 9 was
prepared as previously described from commercially available
Cbz-protected phenylalaninol.32 The following protection step
by acid-catalyzed acetalization furnished compound 10 in a
convenient overnight reaction at room temperature. Removal
of the Cbz group using straightforward catalytic hydrogenation
quantitatively furnished amine 11. Reaction with CDI yielded
carbamoylimidazole 12 as the only product in only 15 min.
Compound 12 is not stable on silica, but the conversion is clean
and quantitative, making purification unnecessary. Intermediate
12 can thus be produced in a very efficient four-step synthesis
from commercial sources.
Benzyl hydrazine 14 was obtained by reacting commercially
available hydrazine 13 with benzaldehyde, followed by in situ
reduction of the intermediate hydrazone. Although imines are
(28) Vidal, J.; Damestoy, S.; Guy, L.; Hannachi, J. C.; Aubry, A.; Collet, A.
Chem.-Eur. J. 1997, 3, 1691–1709.
(29) Smith, R. F.; Brophy, K. A.; Rodriguez, G.; Dennis, L. A.; Ryan, W. J.
Synth. Commun. 1990, 20, 183–188.
(30) Kurz, T.; Widyan, K. Tetrahedron Lett. 2004, 45, 7049–7051.
(31) Grzyb, J. A.; Shen, M.; Yoshina-Ishii, C.; Chi, W.; Brown, R. S.; Batey,
R. A. Tetrahedron 2005, 61, 7153–7175.
(33) Perdicchia, D.; Licandro, E.; Maiorana, S.; Baldoli, C.; Giannini, C.
Tetrahedron 2003, 59, 7733–7742.
(32) Lam, P. Y. S.; Ru, Y.; Jadhav, P. K.; Aldrich, P. E.; DeLucca, G. V.;
Eyermann, C. J.; Chang, C. H.; Emmett, G.; Holler, E. R.; Daneker, W. F.; Li,
L. Z.; Confalone, P. N.; McHugh, R. J.; Han, Q.; Li, R. H.; Markwalder, J. A.;
Seitz, S. P.; Sharpe, T. R.; Bacheler, L. T.; Rayner, M. M.; Klabe, R. M.; Shum,
L. Y.; Winslow, D. L.; Kornhauser, D. M.; Jackson, D. A.; EricksonViitanen,
S.; Hodge, C. N. J. Med. Chem. 1996, 39, 3514–3525.
(34) Brown, H. C.; Krishnamurthy, S. Tetrahedron 1979, 35, 567–607.
(35) Jung, M. E.; Andrus, W. A.; Ornstein, P. L. Tetrahedron Lett. 1977,
4175–4178.
(36) Klosterman, H. J.; Lamoureux, G. L.; Parsons, J. L. Biochemistry 1967,
6, 170–177.
J. Org. Chem. Vol. 73, No. 16, 2008 6121

Waibel and Hasserodt
SCHEME 2. Preparation of Target Compound 7
SCHEME 3. Classical Approach in the Synthesis of
Hydrazino Proline Derivatives
14. The overall yield of deprotection, hydrazone formation, and
reduction was 93%, while purification by column chromatog-
raphy was only necessary following the last step. The prepara-
tion of urea 22 was carried out in a manner identical to that of
15 by reacting hydrazine 21 with methylated 12. Deprotection
of the dimethyl acetal using NaI and TMSCl gave the second
target compound 8 in excellent yield. Since diastereomer
formation was not observed during this reaction sequence, we
conclude that the assembly of 8 occurred without racemization.
As in the synthesis of 7, most steps are high-yielding, do not
require purification by column chromatography, and except for
the amination of proline, are performed at room temperature.
Configurational and Constitutional Analysis. The cycliza-
tion tendency of compounds 7 and 8 was studied by NMR and
UV spectroscopy. The 1H NMR spectrum of 7 in CDCl3 showed
only one set of signals, including the one of the aldehydic proton
(Figure 1), thus indicating that only one of the four possible
proline 1636 was converted into its methyl ester 16a, and
subsequently the two served as model compounds to test the
Zn/HOAc reduction (see Scheme 3). In both cases, LC-MS
analysis of samples directly taken from the crude reaction
mixture showed clean conversion to 17 and 17a (see S39-S43
in Supporting Information), however low yields were obtained
after reaction workup. The generally described workup proce-
dure consists of saturating the reaction solution with H2S, which
results in the precipitation of large amounts of colloidal zinc
sulfide. The resulting slurry was filtered, but only traces of
compounds 17 and 17a were obtained. In light of the LC-MS
data, we suggest that the inherent problem of this procedure is
the encapsulation of the target compound by the considerable
amounts of colloidal zinc sulfide. We do not rule out that a
method might eventually be found to overcome this problem.
In order to obtain good yields of the hydrazino acid building
block, we turned to a rather recent innovation. The oxaziridine
reagent 18 (Scheme 4) effectively transfers an NH-Boc moiety
to an amino acid.28,37 Treatment of (S)-proline with freshly
prepared 18 gave thus N-Boc-protected hydrazino proline 19
in good yield. In order to prepare hydrazino dipeptide 20,
compound 19 was reacted with H-Val-NHtBu using EDC and
HOBt as coupling reagents.38 The Boc group was then removed
in a matter of 2 h by use of TFA before applying the reductive
amination step toward 21, as worked out in the preparation of
1
configurational isomers is present. The corresponding H-1H
NOESY spectrum revealed contacts of the urea proton (position
3 of E-7 in Scheme 5, 6.85 ppm) with the piperidine protons
(position 8, 2.64-2.62 ppm and 2.48-2.36 ppm; see S28 in
Supporting Information), while no interaction with the benzylic
protons (position 6, 4.51 ppm) was observed. These data suggest
that 7 adopts the E/Z configuration as shown in Scheme 5. The
13C NMR spectrum of 7 in CDCl3 revealed only one set of
signals with the characteristic signal of the aldehyde carbon at
200.7 ppm
By contrast, in MeOH-d4 this signal disappeared in favor of
two signals further upfield. At 100.5 and 100.1 ppm, they are
plainly centered in the zone between the olefinic/aromatic and
the aliphatic region, suggesting a partial change in hybridization
from sp2 to sp3. It may be assumed that these signals correspond
to the newly formed pseudotetrahedral carbon, resulting from
the δ+NfCdOδ- interaction (cyclized Z-7 in Scheme 5). Their
splitting can be explained with the fact that a new stereogenic
center has been formed without any particular stereoselectivity
and that two diasteromers were generated as it was previously
observed.27c This is supported by the splitting of most of the
other carbon signals in the spectrum. These observations are
1
(37) Vidal, J.; Hannachi, J. C.; Hourdin, G.; Mulatier, J. C.; Collet, A.
Tetrahedron Lett. 1998, 39, 8845–8848.
reflected in the H NMR spectrum taken in MeOH-d4, where
the signal corresponding to the aldehyde proton has disappeared.
(38) Li, W. H.; Hanau, C. E.; d’Avignon, A.; Moeller, K. D. J. Org. Chem.
1995, 60, 8155–8170.
1
A H-13C HMBQ spectrum revealed that the signal of the
6122 J. Org. Chem. Vol. 73, No. 16, 2008

DistinctiVe δ+NfCdOδ- Interaction
SCHEME 4. Preparation of Target Compound 8
methine proton of the newly formed pseudotetrahedral center
(position 1, Scheme 5) is located right underneath the signal of
the benzylic protons (position 6, δ 4.61-4.57). Futher evidence
for the δ+NfCdOδ- interaction in MeOH-d4 could be obtained
from a 1H-13C HMBC spectrum (see S29 and S30 in Support-
ing Information for the HMBQ and HMBC spectra). The
determined to be approximately 13 kcal/mol39 and thus sign-
ficantly lower than that of carboxamide bonds (around 20
kcal/mol).40,41 Still, a barrier of 13 kcal/mol explains the relative
slowness of the process observed here at room temperature.
Compound 8 showed a behavior identical to that of 7 when
being examined by 1H and 13C NMR. The open aldehydic form
was the only one present in CDCl3, whereas time-resolved
folding occurred in MeOH-d4 to give cyclized Z-8 in the form
of two diastereomers (Scheme 5). Interestingly, when samples
of 7 or 8 in MeOH-d4 are evaporated to dryness and redissolved
in CDCl3, only the linear aldehydic form is detected, thus
demonstrating that this unique cyclization process is completely
reversible.
3
spectrum showed J coupling between the methine proton and
the carbon atoms adjacent to the piperidine nitrogen (positions
8, δ 55.41, 55.38, 54.92, and 54.86 ppm). To our knowledge,
this is the first time that the δ+NfCdOδ-interaction has been
characterized by 2D NMR spectroscopy and that actual coupling
through this weak bond has been demonstrated.
Interestingly, upon dissolution in MeOH-d4, the “folding” of
7 can be monitored by NMR. Figure 1 demonstrates that about
40% is cyclized after 10 min at room temperature, 70%
conversion is reached after 30 min, and cyclization is complete
in 3 h. These kinetics might be explained with the existence of
a configurational equilibrium present in 7 (Scheme 5). In the
absence of a polar protic medium, the thermodynamic form is
E-7, where the planar secondary amide of the urea moiety is
E-configured. In MeOH-d4, this configurational equilibrium is
shifted completely to the Z form by the coupled constitutional
equilibrium in favor of the cyclized Z form. The activation
barrier of the E/Z equilibrium of regular ureas has been
The above observation that the δ+NfCdOδ- interaction is
especially favored in polar protic media is in complete agreement
with literature data.17,19,27 It is however the first time that such
an interaction is observed in the formation of a six-membered
ring outside the context of multicyclic or macrocyclic systems
that benefit from elevated levels of preorganization. Up until
now, linear systems were only known to fold to five-membered
rings.17,27c This leads us to conclude that there are no principal
obstacles to the formation of a six-membered ring brought about
by an δ+NfCdOδ- interaction, aside from an unfavorable
entropic term. Our system benefits from a urea moiety that, once
Z-configured, endows the molecule with a considerable degree
of preorganization thanks to its coplanarity.
UV spectroscopy is an alternative means to detect the
δ+NfCdOδ- interaction.19,26 It is characterized by a band that
is significantly more intense and found at a shorter wavelength
than the one of the aldehyde function. This band can be ascribed
to the transition between the bonding and antibonding molecular
orbitals newly formed by the through-space interaction of the
n nitrogen orbital and the π orbitals of the carbonyl function
(Figure 2).24 Figure 3 shows two diagrams, each depicting two
UV spectra taken from compounds 7 and 8 in CHCl3 and
MeOH, respectively. In CHCl3, the spectra have their maximum
at 241 nm, corresponding to the absorption of the aldehyde
function. The extinction coefficients were calculated to be ꢀ )
877 L mol-1 cm-1 for 7, and ꢀ ) 733 L mol-1 cm-1 for 8. In
MeOH, the spectra show the intense δ+NfCdOδ- band at 208
nm, while no aldehyde absorption can be observed. The
(39) Zhao, Y. M.; Raymond, M. K.; Tsai, H.; Roberts, J. D. J. Phys. Chem.
1993, 97, 2910–2913.
(40) Drakenberg, T.; Dahlqvist, K.-I.; Forsen, S. J. Phys. Chem. 1972, 76,
2178–2183.
FIGURE 1. 1H spectra (500 MHz) of 7 in CDCl3, and at certain time
intervals beginning with initial dissolution in MeOH-d4. (9) E-7
(position 1 proton), (b) cyclized Z-7 (position 2 proton).
(41) Scherer, G.; Kramer, M. L.; Schutkowski, M.; Reimer, U.; Fischer, G.
J. Am. Chem. Soc. 1998, 120, 5568–5574.
J. Org. Chem. Vol. 73, No. 16, 2008 6123

Waibel and Hasserodt
SCHEME 5. Equilibria between the E Configuration and the Putative Z Configuration (Not Observed) of the Open Form, and
the Cyclic Constitution Z-7
extinction coefficients corresponding to this band were deter-
mined to be ꢀ ) 21200 and 18082 L mol-1 cm-1 for 7 and 8,
respectively. These results are in accord with previous reports19,26
and support the observations of the NMR investigations of the
present study.
ties yet to be explored. The rigidity provided by the urea core
as well as the nucleophilicity of the hydrazine nitrogen make
the cyclic form the only constitutional isomer present in
methanol. This may likely be extended to aqueous media but
remains to be verified by investigating derivatives with increased
water solubility. Importantly, the absence of any aldehydic form
is essential to any exploration of the δ+NfCdOδ- moiety as a
possible pharmacophore. Both target molecules revert com-
pletely and rapidly to the linear, aldehydic form when redis-
solved in chloroform, thus making this system an attractive
starting point for the development of a molecular switch.
The synthetic pathway elaborated here allows for fast and
economical generation of diversity using commercially available
enantiopure amino acid derivatives and secondary amines as
building blocks. All reactions but one were carried out at room
temperature, are high-yielding, and in most cases do not require
purification by column chromatography. These assets make the
pathway particularly attractive for a combinatorial approach to
the discovery of novel HIV PR inhibitors, or of antagonists to
any other drug target for that matter. In silico screening by QM/
MM methods of selected candidates based on our system in
complex with HIV PR is currently in progress. The results will
be applied to the generation of promising lead structures in due
course.
FIGURE 2. Energy levels of the δ+NfCdOδ- molecular orbitals,
constructed from the nitrogen n orbital and the carbonyl π orbitals;
and observed transitions.
Experimental Section
(S)-(1-Benzyl-2,2-dimethoxy-ethyl)-carbamic Acid Benzyl
Ester (10). Aldehyde 9 (625 mg, 2.21 mmol) was dissolved in
anhydrous MeOH (20 mL), and p-toluenesulfonic acid monohydrate
(40 mg, 0.21 mmol) was added. The reaction was stirred overnight,
and then most of the solvent was removed under vacuum. NaHCO3
(saturated aqueous solution, 20 mL) was added, and the mixture
was extracted with CH2Cl2 (3 × 30 mL). The combined organic
extracts were dried over Na2SO4, and the solvent was evaporated
under vacuum. Purification by flash chromatography (33% EtOAc/
cyclohexane) gave 10 as a white solid (625 mg, 86%). Mp 75 °C;
1H NMR (500 MHz, CDCl3) δ 7.35-7.19 (m, 10H), 5.07 (d, J )
12.4 Hz, 1H), 5.01 (d, J ) 12.4 Hz, 1H), 4.94 (d, J ) 9.2 Hz, 1H),
4.17 (d, J ) 3.3 Hz, 1H), 4.11-4.09 (m, 1H), 3.43 (s, 3H), 3.39
(s, 3H), 2.93 (dd, J ) 13.8, 6.1 Hz, 1H), 2.79 (dd, J ) 13.8, 8.0
Hz, 1H); 13C NMR (50 MHz, CDCl3) δ 155.9, 137.7, 136.5, 129.2,
128.3, 127.8, 127.7, 126.2, 104.6, 66.4, 55.6, 55.4, 53.4, 35.8;
HRMS (ESI) calcd for C19H23NO4Na [M + Na]+ 352.1525, found
352.1526.
FIGURE 3. UV spectra of compounds 7 and 8 in MeOH (dotted line)
and in CHCl3 (continuous line).
(S)-Imidazole-1-carboxylic Acid (1-Benzyl-2,2-dimethoxy-
ethyl)-amide (12). CDI (457 mg, 2.82 mmol) was suspended in
anhydrous THF (3 mL), and a solution of amine 11 (500 mg, 2.56
mmol) in THF (2 mL) was added dropwise at room temperature.
The reaction was stirred for 30 min, and then the solvent was
evaporated. The residue was redissolved in CH2Cl2 (10 mL) and
washed with water (2 × 5 mL). The organic phase was dried over
Conclusion
We have integrated the unusual δ+NfCdOδ- moiety into a
drug-like molecule (low molecular weight, rigid, and largely
inert) that displays key elements present in well-known HIV
PR inhibitors, while offering unique hydrogen bonding proper-
6124 J. Org. Chem. Vol. 73, No. 16, 2008

DistinctiVe δ+NfCdOδ- Interaction
Na2SO4, and the solvent was evaporated under vacuum to give pure
12 as colorless oil (814 mg, quant yield). H NMR (500 MHz,
(NNCH2pip), 42.3 (NCH2Ph), 35.3 (CH2Ph), 26.3 (CH2pip), 26.2
(CH2pip), 23.1 (CH2pip); 1H NMR (500 MHz, MeOH-d4), (1:1
mixture of two diasteromers) δ 7.33-7.10 (m, 18H; Ar), 6.94-6.88
(m, 2H; Ar), 4.61-4.57 (m, 4H; NCH2Ph, δ+NfCHdOδ-),
4.49-4.45 (m, 2H; NCH2Ph), 4.14-4.05 (m, 2H; NHCH), 3.12-3.05
(m, 2H; CH2Ph), 2.81-2.72 (m, 4H; CH2Ph, NNCH2pip), 2.64-2.58
(m, 2H; NNCH2pip), 2.41-2.37 (m, 4H; NNCH2pip), 1.67-1.60 (m,
10H; CH2pip), 1.11-1.04 (m, 2H; CH2pip); 13C NMR (127 MHz,
MeOH-d4) δ 160.9 (NCON), 160.8 (NCON), 142.4 (Ar), 140.9 (Ar),
131.39 (Ar), 131.37 (Ar), 130.2 (Ar), 130.0 (Ar), 129.03 (Ar),
129.02 (Ar), 128.4 (Ar), 128.1 (Ar), 100.5 (δ+NfCHdOδ-), 100.1
1
CDCl3) δ 8.04 (s, 1H), 7.33-7.22 (m, 6H), 7.07 (s, 1H), 5.90 (d,
J ) 8.5 Hz, 1H), 4.42-4.37 (m, 1H), 4.26 (d, J ) 2.7 Hz, 1H),
3.49 (s, 3H), 3.39 (s, 3H), 3.02-2.94 (m, 2H); 13C NMR (50 MHz,
CDCl3) δ 148.6, 137.0, 135.9, 130.5, 129.2, 128.7, 126.8, 115.8,
104.1, 56.0, 55.7, 53.5, 36.0; HRMS (ESI) calcd for C15H20N3O3
[M + H]+ 290.1505, found 290.1505.
Benzyl-piperidin-1-yl-amine (14). Benzaldehyde (2.44 g, 23.02
mmol) and hydrazine 13 (2.00 g, 20.00 mmol) were stirred at room
temperature in anhydrous MeOH (150 mL) overnight. Acetic acid
(30 mL, 525 mmol) and NaCNBH3 (6.30 g, 100 mmol) were added,
and stirring was continued for 2 h. Most of the solvent was removed
under vacuum. Then a pH value around 8 was adjusted by adding
NaHCO3 (saturated aqueous solution, 40 mL) and NaOH (aqueous
solution, 10 mL), and the mixture was extracted with CH2Cl2 (3 ×
40 mL). The combined organic extracts were dried over Na2SO4,
and the solvent was evaporated under vacuum. After purification
by flash chromatography (50% EtOAc/cyclohexane), 14 was
(
δ+NfCHdOδ-), 57.5 (NHCH), 57.4 (NHCH), 55.41 (NNCH2pip),
55.38 (NNCH2pip), 54.92 (NNCH2pip), 54.86 (NNCH2pip), 43.9
(NCH2Ph), 37.44 (CH2Ph), 37.36 (CH2Ph), 28.36 (CH2pip), 28.34
(CH2pip), 28.30 (CH2pip), 28.24 (CH2pip), 25.1 (CH2pip); MS (ESI)
m/z 366.2 ([M + H]+, 100). HRMS (ESI) calcd for C22H28N3O2
[M + H]+ 366.2182, found 366.2187. Anal. Calcd for C22H27N3O2:
C, 72.30; H, 7.45; N, 11.50. Found: C, 72.19; H, 7.48; N 11.46.
(S,S)-[2-(1-tert-Butylcarbamoyl-2-methyl-propylcarbamoyl)-
pyrrolidin-1-yl]-carbamic Acid tert-Butyl Ester (20). Compound
19 (1.94 g, 8.43 mmol) was dissolved in CH2Cl2 (60 mL), and
EDC (2.26 g, 11.77 mmol) and then HOBt (12.65 mL of a 1 N
solution in N-methylmorpholine, 12.65 mmol) were added. After
5 min, a suspension of HValNHtBu·HCl (1.75 g, 8.43 mmol) and
triethylamine (1.17 mL, 8.43 mmol) in CH2Cl2 (20 mL) was added,
and the reaction was stirred overnight. NaHCO3 (saturated aqueous
solution, 80 mL) was added, and the mixture was extracted with
CH2Cl2 (3 × 60 mL). The combined organic extracts were dried
over Na2SO4, and the solvent was evaporated under vacuum.
Purification by flash chromatography (50% EtOAc/CH2Cl2) gave
20 as colorless oil (3.24 g, quant yield). 1H NMR (500 MHz, CDCl3)
δ 8.28 (br s, 1H), 6.01 (br s, 1H), 5.82 (br s, 1H), 4.05-4.02 (m,
1H), 3.56 (dd, J ) 10.2, 4.8 Hz, 1H), 3.41-3.38 (m, 1H), 2.87
(dd, J ) 16.4, 9.4 Hz, 1H), 2.35-2.27 (m, 2H), 2.01-1.85 (m,
2H), 1.82-1.72 (m, 1H), 1.45 (s, 9H), 1.32 (s, 9H), 0.95 (d, J )
6.7 Hz, 6H); 13C NMR (50 MHz, CDCl3) δ 174.0, 170.2, 155.6,
80.1, 68.2, 59.3, 55.7, 50.9, 29.8, 29.5, 28.5, 28.2, 23.1, 19.4, 17.6;
HRMS (ESI) calcd for C19H36N4O4Na [M + Na]+ 407.2634, found
407.2633.
1
obtained as a colorless oil (3.44 g, 91%). H NMR (500 MHz,
CDCl3) δ 7.36-7.23 (m, 5H), 3.97 (s, 2H), 2.67 (br s, 4H),
1.66-1.62 (m, 4H), 1.57-1.52 (m, 2H); 13C NMR (50 MHz,
CDCl3) δ 138.9, 128.3, 128.0, 126.7, 57.3, 52.6, 25.8, 23.7; HRMS
(CI) calcd for C12H19N2 [M + 1]+ 191.1548, found 191.1549.
(S)-1-Benzyl-3-(1-benzyl-2,2-dimethoxy-ethyl)-1-piperidin-1-
yl-urea (15). Compound 12 (257 mg, 0.89 mmol) was dissolved
in anhydrous acetonitrile (1.6 mL), and iodomethane (215 µL, 3.45
mmol) was added. The reaction was stirred overnight at room
temperature, then the solvent was evaporated, and the yellow oil
was dried under vacuum. The residue was redissolved in anhydrous
CH2Cl2 (9 mL), and benzylhydrazine 14 (169 mg, 0.89 mmol) and
triethylamine (123 µL, 0.89 mmol) were added. After stirring
overnight at room temperature, the reaction was quenched by adding
NaHCO3 (saturated aqueous solution, 10 mL), and the mixture was
extracted with CH2Cl2 (3 × 10 mL). The combined organic extracts
were dried over Na2SO4, and the solvent was removed under
vacuum. After purification by flash chromatography (33% EtOAc/
cyclohexane), urea 15 was obtained as a colorless oil (260 mg,
1
71%). H NMR (500 MHz, CDCl3) δ 7.29-7.13 (m, 10H), 6.61
(d, J ) 9.0 Hz, 1H), 4.60 (d, J ) 16.3 Hz, 1H), 4.45 (d, J ) 16.3
Hz, 1H), 4.31-4.27 (m, 2H), 3.473 (s, 3H), 3.468 (s, 3H), 3.06
(dd, J ) 13.9, 4.5 Hz, 1H), 2.78 (dd, J ) 13.9, 8.9 Hz, 1H),
2.71-2.69 (m, 1H), 2.52-2.48 (m, 1H), 2.42-2.40 (m, 1H),
2.36-2.32 (m, 1H), 1.65-1.42 (m, 5H), 1.03-0.95 (m, 1H); 13C
NMR (50 MHz, CDCl3) δ 157.8, 140.5, 138.6, 129.4, 128.2, 128.1,
127.2, 126.4, 126.0, 105.9, 56.3, 55.0, 53.4, 53.1, 52.1, 42.0, 35.3,
26.4, 23.2; HRMS (ESI) calcd for C24H33N3O3Na [M + Na]+
434.2420, found 434.2421.
(S,S)-1-Benzylamino-pyrrolidine-2-carboxylic Acid (1-tert-
Butylcarbamoyl-2-methyl-propyl)-amide (21). Compound 20
(1.48 g, 3.85 mmol) was dissolved in TFA (80 mL) and stirred for
2 h at room temperature. The solvent was evaporated, and the
yellow oil was dried under vacuum. The residue was redissolved
in MeOH (30 mL), and according to the preparation of 14, reacted
with benzaldehyde (471 mg, 4.44 mmol), followed by treatment
with NaCNBH3 (1.22 g, 19.30 mmol) and acetic acid (6 mL, 105
mmol). After reaction workup (see synthesis of compound 14), and
purification by flash chromatography (5% MeOH/CH2Cl2), 21 was
obtained as a white solid (1.34 g, 93%). Mp 82 °C; 1H NMR (500
MHz, CDCl3) δ 7.68 (d, J ) 9.0 Hz, 1H), 7.35-7.22 (m, 5H),
5.93 (br s, 1H), 4.06-3.98 (m, 3H), 3.55-3.52 (m, 1H), 3.24 (dd,
J ) 9.6, 5.8 Hz, 1H), 2.36 (dd, J ) 16.5, 8.2 Hz, 1H), 2.27-2.13
(m, 2H), 1.89-1.75 (m, 3H), 1.31 (s, 9H), 0.95 (d, J ) 6.8 Hz,
3H), 0.88 (d, J ) 6.8 Hz, 3H); 13C NMR (50 MHz, CDCl3) δ 173.7,
170.2, 138.2, 128.7, 128.2, 127.1, 69.6, 58.3, 56.1, 54.8, 51.0, 30.8,
28.4, 28.1, 22.2, 19.2, 18.0; HRMS (ESI) calcd for C21H35N4O2
[M + H]+ 375.2760, found 375.2761.
(S)-1-Benzyl-3-(1-benzyl-2-oxo-ethyl)-1-piperidin-1-yl-urea
(7). To a solution of compound 15 (120 mg, 0.29 mmol) in
anhydrous acetonitrile (6 mL) were added NaI (110 mg, 0.73 mmol)
and TMSCl (93 µL, 0.73 mmol). The reaction was stirred at room
temperature for 1 h and was then quenched by adding NaHCO3
(saturated aqueous solution, 5 mL). The mixture was extracted with
CH2Cl2 (3 × 10 mL), and the combined organic extracts were
washed with Na2S2O3 (saturated aqueous solution, 10 mL) and then
with water (10 mL). The organic phase was dried over Na2SO4,
and the solvent was evaporated under vacuum. After purification
by flash chromatography (50% EtOAc/cyclohexane), 7 was obtained
as a colorless oil (97 mg, 92%). [R]25 -30.6 (c 1.086, CHCl3);
D
Acknowledgment. This article is dedicated to the memory
of Nelson Leonard who kindly communicated with us in
October, 2000 during the initial stages of this project. We are
indebted to Maurice Santelli, Marseille, for suggesting the
hydrazine moiety as a potential δ+NfCdOδ- partner. Special
thanks go to Laure Guy and Jean-Christophe Mulatier, Lyon,
for their advice in the preparation of hydrazinoproline via
oxaziridines. We thank Bernard Fenet and Denis Bouchu, both
of Lyon, for the preparation of 2D NMR spectra and HRMS
1H NMR (500 MHz, CDCl3) δ 9.66 (s, 1H; CHO), 7.25-7.12 (m,
10H; Ar), 6.85 (d, J ) 6.9 Hz, 1H; NH), 4.51 (s, 2H; NCH2Ph),
4.50-4.46 (m, 1H; NHCH), 3.12 (dd, J ) 14.0, 6.0 Hz, 1H;
CH2Ph), 3.05 (dd, J ) 14.0, 7.4 Hz, 1H; CH2Ph), 2.64-2.62 (m,
1H; NNCH2piperidine), 2.48-2.36 (m, 3H; NNCH2pip), 1.54-1.50 (m,
3H; CH2pip), 1.36-1.22 (m, 2H; CH2pip), 0.95-0.87 (m, 1H; CH2pip);
13C NMR (127 MHz, CDCl3) δ 200.7 (CHO), 158.0 (NCON), 140.0
(Ar), 136.4 (Ar), 129.3 (Ar), 128.6 (Ar), 128.3 (Ar), 127.4 (Ar),
126.9 (Ar), 126.7 (Ar), 60.2 (NHCH), 53.5 (NNCH2pip), 53.4
J. Org. Chem. Vol. 73, No. 16, 2008 6125

Waibel and Hasserodt
1
spectra, respectively. This work was supported by the Rhone-
Alpes Region through research cluster 5, and by the French
Research Ministry and the CNRS. M.W. acknowledges a
European Doctoral Fellowship from the Research Ministry.
and 13C NMR spectra of compounds 7-22. H NMR spectral
monitoring of the cyclization process of 7 in MeOH-d4. 1H-1H
NOESY, H-13C HMBQ, and H-13C HMBC spectra of 7.
LC-MS analyses (chromatograms, mass spectra) of the conver-
sions of 16 to 17 and 16a to 17a. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
1
Supporting Information Available: General experimental
procedures. Experimental protocols and characterization data
for compounds 8, 9, 11, 16, 16a, 17, 17a, 19, and 22. 1H NMR
JO800719J
6126 J. Org. Chem. Vol. 73, No. 16, 2008