Characterization of the two fundamental conformations of benzoylureas and elucidation of the factors that facilitate their conformational interchange

pubs.acs.org/joc

Characterization of the Two Fundamental Conformations of

Benzoylureas and Elucidation of the Factors That Facilitate

Their Conformational Interchange

Guillaume Lessene,^†Brian J. Smith,^†Robert W. Gable,^‡and

Jonathan B. Baell*^,†

^†The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville,

Victoria 3052, Australia., and ^‡School of Chemistry, The University of Melbourne,

Victoria 3010, Australia

jbaell@wehi.edu.au

Received April 28, 2009

The ability of the benzoylurea core to mimic R helices relies on its ability to form an intramolecular

hydrogen bond. The conformational behavior of benzoylureas is investigated in depth in this study

via the use of NMR, IR, X-ray, and computational analysis. The results show that the closed

conformation maintained by an intramolecular hydrogen bond is favored in most of the cases

studied except when steric and electronic effects combined with a solvent possessing a high

hydrogen bond accepting ability, such as DMSO, are involved. The study highlights the propensity

for benzoylureas to switch conformation depending on the environment of the molecule for a

particular set of substituents. We anticipate that our summary of the phenomenon of internal

hydrogen bonding and its analysis may further serve as a useful reference source for future workers

in this area.

Introduction

protein-protein interactions that comprise a helical epitope

in the binding interface.^7,8We identified the benzoylurea

moiety to be particularly suitable for this task as it contains

heteroatoms that introduce a certain degree of amphi-

philicity within the scaffold without being excessively hydro-

phobic.⁹Furthermore, we envisaged this system to be

synthetically versatile such that, when adequately functio-

nalized, it could act as a R-helical protein mimetic. In

particular, we were interested in the topographical similarity

between the benzoylphenylurea and the terphenyl system,

already shown by Hamilton and co-workers to be able to

The discovery of organic scaffolds that mimic targeted

binding epitopes is directly relevant to the development of

small molecule inhibitors of protein-protein interactions.

We have previously reported on our successful efforts in the

design of small molecule structural and functional mimetics

of pharmacologically active toxins.^1-6Our current interest is

the design of R-helical mimetics as inhibitors of those

(1) Baell, J. B.; Duggan, P. J.; Forsyth, S. A.; Lewis, R. J.; Lok, Y. P.;

Schroeder, C. I. Bioorg. Med. Chem. 2004, 12, 4025–4037.

(2) Baell, J. B.; Duggan, P. J.; Lok, Y. P. Aust. J. Chem. 2004, 57, 179–185.

(3) Baell, J. B.; Forsyth, S. A.; Gable, R. W.; Norton, R. S.; Mulder, R. J.

J. Comput.-Aided Mol. Des. 2001, 15, 1119–1136.

(4) Baell, J. B.; Harvey, A. J.; Norton, R. S. J. Comput.-Aided Mol. Des.

2002, 16, 245–262.

(5) Harvey, A. J.; Gable, R. W.; Baell, J. B. Bioorg. Med. Chem. Lett.

2005, 15, 3193–3196.

(7) Baell, J. B.; Huang, D. C. Biochem. Pharmacol. 2002, 64, 851–863.

(8) Lessene, G.; Czabotar, P. E.; Colman, P. M. Nat. Rev. Drug Discovery

2008, 7, 989–1000.

(9) Moreover, benzoylureas have a good track record in marketed

pharmaceuticals. For example, see: Ware, G.; Whitcare, D. The Pesticide

Book, 6th ed.; Meister Media Worldwide, Willoughby, OH, 2004.Okada, H.;

Kato, M.; Koyanagi, T.; Mizuno, K. Chem. Pharm. Bull. 1999, 47, 430–433.

Li, J.-N.; Song, D.-Q.; Lin, Y.-H.; Hu, Q.-Y.; Yin, L.; Bekesi, G.; Holland, J.

F.; Jiang, J.-D. Biochem. Pharmacol. 2003, 65, 1691–1699.

(6) Norton, R. S.; Baell, J. B.; Angus, J. A. In Calcium channel pharma-

cology; McDonough, S. I., Ed.; Kluwer/Academic/Plenum Press: New York,

2003; pp 143-179.

DOI: 10.1021/jo900871a

r

Published on Web 08/07/2009

J. Org. Chem. 2009, 74, 6511–6525 6511

2009 American Chemical Society

Lessene et al.

JOCArticle

CHART 1. Classification of Benzoylurea Derivatives Consid-

ered in the Present Study

FIGURE 1. Comparison of (a) the benzoylphenylurea core and

(b) Hamilton’s terphenyl system.

FIGURE 2. Two different benzoylurea conformations represented

in the Cambridge Crystallographic Database.

or even local energy minimum.¹⁵Since appropriate scaffold

conformation is clearly critical to mimetic design, we under-

took a conformational analysis of model compounds under

a variety of experimental conditions and support our find-

ings with selected use of high-level density functional theory

computational calculations.

geometrically and functionally mimic biologically active

R-helical binding epitopes (Figure 1).^10-13

The pivotal feature of the benzoylureas is the central

acylurea group that is required to exist in the intended

“closed” conformation necessary to project functional

groups appropriately (A, Figure 2), forming a pseudo-6-

membered ring with an intramolecular hydrogen bond and a

central Z amide bond (E and Z geometry will refer to the

position of the R¹and R²groups in relation to their

neighboring carbonyl group). Preliminary molecular me-

chanics modeling used in our design process confirmed this

hypothesis. However, alternative conformations can be pos-

tulated: formally there exist eight possible conformations

through rotation about three bonds.

A search of the Cambridge Crystallographic Database

(CCD) confirmed that many benzoylureas exist in this

pseudo-6-membered ring conformation. However, there

are some benzoylureas for which the C¹-N¹bond adopts a

Z conformation¹⁴as shown by Bin Figure 2. This conformation

is not predicted by molecular mechanics to occupy a global

We describe here for the first time why benzoylureas can

adopt two distinct conformations;the “closed” form A or

the “twisted” form B. We show that the preferred conforma-

tion adopted for a given benzoylurea depends on the nature

of substituent groups and can be logically explained by a

subtle interplay of electrostatic and steric effects.¹⁶We

further elucidate why benzoylureas are better described

structurally as bis-amides rather than as acylureas; confor-

mationally, they behave as two distinct amide units without

any urea character whatsoever. This structural characteristic

enables strengthening of an internal hydrogen bond that

stabilizes the closed form, which is otherwise inherently

sterically disfavored. This hydrogen bond is not a reso-

nance-assisted hydrogen bond as defined by Gilli et al., but

despite this, is shown by gas-phase calculations to remain

very strong.^17-19

(10) Orner, B. P.; Ernst, J. T.; Hamilton, A. D. J. Am. Chem. Soc. 2001,

123, 5382–5383.

Results and Discussion

(11) Kutzki, O.; Park, H. S.; Ernst, J. T.; Orner, B. P.; Yin, H.; Hamilton,

A. D. J. Am. Chem. Soc. 2002, 124, 11838–11839.

(12) Yin, H.; Lee, G.-i.; Sedey, K. A.; Kutzki, O.; Park, H. S.; Orner, B. P.;

Ernst, J. T.; Wang, H.-G.; Sebti, S. M.; Hamilton, A. D. J. Am. Chem. Soc.

2005, 127, 10191–10196.

(13) This work as well as applications to generate biologically active, Bcl-

x_L-binding, R helical mimetics was presented in part at the 20th Royal

Australian Chemical institute Organic Chemistry Conference, Cairns,

Australia, 2004, abstract P(RACIOC) 40 and at Connect 2005, Sydney,

Australia, 2005, Abstracts 248 and 364. The worldwide patent disclosing

these structures was published in 2006 (Lessene G.; Baell, J. B. R-Helical

Mimetics, PCT/AU2005/000968). This system has apparently been “redis-

covered” in the interim and published ( Rodriguez, J. M.; Hamilton, A. D.

Angew. Chem., Int. Ed. 2007, 46, 8614–8617. Rodriguez, J. M.; Ross, N. T.;

Katt, W. P.; Dhar, D.; Lee, G.-i.; Hamilton, A. D. ChemMedChem 2009, 4,

649–656).

Compounds Studied. Compounds of interest in this study

are shown in Chart 1, divided into four classes characterized

by the nature of the substituents R¹and R². A nonbranched

R¹substituent is defined as any alkyl group that comprises a

methylene moiety R to the N¹nitrogen atom, such as methyl

or ethyl, whereas we define this substituent as “branched”

when the carbon R to the nitrogen is branched, such as for

R¹= isopropyl, sec-butyl, tert-butyl, etc. Thus, we define

(16) Kohmoto et al. have reported the closed form of a benzoylureas

(Kohmoto, S.; Iwasaki, N.; Fukui, D.; Nishio, T.; Iida, I.; Kishikawa, K.;

Yamamoto, M.; Yamada, K. J. Chem. Soc., Perkin Trans. 2 1996, 985-988)

and analyzed a twisted form of an alkylacylurea ( Kohmoto, S.; Kasimura,

H.; Nishio, T.; Iida, I.; Kishikawa, K.; Yamamoto, M.; Yamada, K. J.

Chem. Soc., Perkin Trans. 2 1994, 1565-1568) but have not extended this to

an understanding of conversion between these two forms.

(17) Gilli, G.; Gilli, P. J. Mol. Struct. 2000, 552, 1–15.

(18) Gilli, P.; Bertolasi, V.; Pretto, L.; Lycka, A.; Gilli, G. J. Am. Chem.

Soc. 2002, 124, 13554–13567.

(19) Gilli, P.; Bertolasi, V.; Ferretti, V.; Gilli, G. J. Am. Chem. Soc. 2000,

122, 10405–10417.

(14) Of 25 structures in version 5.30 of CCD, three were in this alternative

conformation.

(15) Maximin2 using the Tripos Forcefield as implemented in Sybyl7.01

(Tripos Associates, MO); a third, extended (EEZ) form has been identified as

a putative biologically active conformation using the Catalyst/HIPHOP

software program but in light of the likely generic inability of molecular

mechanics to properly analyse benzoylurea conformation, this finding is

questionable (see: Kurogi, Y.; Miyata, K.; Okamura, T.; Hashimoto, K.;

Tsutsumi, K.; Nasu, M.; Moriyasu, M. J. Med. Chem. 2001, 44, 2304-2307).

6512 J. Org. Chem. Vol. 74, No. 17, 2009

Lessene et al.

JOCArticle

SCHEME 1. Retrosynthesis of Benzoylureas

SCHEME 2. Benzoylurea Preparation

We applied this approach to the benzoylureas and were pleased

to see that this procedure furnished all the desired compounds

in high yields and with great ease of purification (Scheme 2).

Spectroscopic Study. A clear feature that distinguishes

between the closed and twisted benzoylurea conforma-

tions in Figure 2 is the respective presence or absence

of an intramolecular hydrogen bond. IR and ¹H NMR

spectroscopy are classical techniques used to analyze

hydrogen-bonding behavior.²⁸IR spectroscopy results

are relatively straightforward to interpret. Hydrogen

bonding can be intermolecular, involving the solvent or

another molecule of solute, or can be intramolecular.

Assignment of an intramolecular hydrogen bond by IR

spectroscopy, therefore, is clearer in a non-hydrogen-

bonding solvent (or very weak hydrogen-bonding solvent),

such as chloroform, with dilute conditions (to reduce the

potential for dimer formation of the solute), and for solutes

with only a limited number of amide NH groups (to

minimize band overlap).^28-31The time scale in IR spec-

troscopy is fast compared with conformational changes

and allows the observation of the two characteristic bands

of N-H bond vibration. In non-hydrogen-bonding sol-

vents, a broadband at longer wavelengths (usually around

3200-3340 cm^-1) corresponds to a hydrogen-bonded NH,

while a sharp band at shorter wavelengths (usually around

3400-3450 cm^-1) corresponds to a solvent-exposed NH.

Caution must be used in the use of IR spectra to estimate

conformer ratios since a hydrogen-bonded conformation

might have a distinctly different absorption coefficient

from a non-hydrogen-bonded conformation. In this study,

we used IR spectroscopy to qualitatively assess the pre-

sence of hydrogen-bonded versus non-hydrogen-bonded

conformations to support quantitative interpretation of

data obtained via ¹H NMR spectroscopy.

four distinct classes of compounds: class I compounds (1)

possess a nonbranched N¹substituent (ethyl) and a phenyl

ring on N²; class II compounds (2a-c) possess a non-

branched N¹substituent (ethyl) and a varying alkyl R²

group; class III compounds (3) possess a branched N¹

substituent (isopropyl) and a phenyl ring on N²; class IV

compounds (4a-c) possess a branched N¹substituent

(isopropyl) and an alkyl R²group.²⁰

Synthesis of Benzoylureas. Although a large number of

benzoylureas are known, the chemistry surrounding this

group of compounds is rather disparate, especially when R¹

is not a hydrogen atom. Methods using carbodiimides^21,22or

ureas²³are known and furnish “symmetrical” benzoylureas

(R¹= R²). Examples where R¹and R²are different alkyl

groups have been presented in an Eli Lilly patent and in a

study of semisynthetic β-lactam antibiotics.^24,25The key

feature in these two preparations is the formation of an

acylcarbamoyl chloride from reaction of a chlorocarbonyl

donor with an activated secondary amide. Subsequent reac-

tion with amines furnishes differentially substituted benzoy-

lureas (Scheme 1). In the Eli Lilly report,²⁴the amide was

activated by generation of the anion via the use of butyl-

lithium, whereas Ohi et al.²⁵achieved activation through

silylation with trimethylsilyl choride (TMSCl).

We attempted the butyllithium-mediated method first, but

this was not successful with amides bearing a branched

R¹substituent and so we turned to the silylation-based

approach. We noted that a milder silylation-based route to

the synthesis of acylcarbamoyl chlorides than that used by

Ohi et al. has also been reported, using trimethylsilyl triflate

(TMSOTf) as the silylating agent and phosgene as the

chlorocarbonyl donor.^26,27This synthetic route has not been

applied to the synthesis of benzoylureas, but we reasoned

that it should be possible to do so by reacting the acylcarbamoyl

chloride intermediate with either an aniline using triethyla-

mine as an acid scavenger or with 2 equiv of an alkylamine.

NMR has advantages over IR as a spectroscopic techni-

que since data from experiments in polar solvents as well as

nonpolar solvents can be readily and meaningfully inter-

preted.^32-34If the benzoylurea interconverts rapidly on an

NMR time scale between two different conformations,

A and B, and if the limiting chemical shift values of a given

proton signal of each form are known, the equilibrium

(20) All classes are drawn as extended, conformationally “neutral” forms,

indicative of neither the closed nor twisted conformations.

(21) Kishikawa, K.; Yamamoto, M.; Kohmoto, S.; Yamada, K. J. Org.

Chem. 1989, 54, 2428–2432.

(22) Lee, W. W.; Martinez, A. P.; Goodman, L. J. Med. Chem. 1974, 17,

326–330.

(23) Smith, T. D.; Jones, P. G.; Schmutzler, R. Z. Naturforsch. B 1992, 47,

526–532.

(28) Aaron, H. S. In Topics in Stereochemistry; Allinger, N. L., Eliel, E.

L., Eds.; John Wiley & Sons: New York, 1979; Vol. 11, pp 1-52.

(29) Gellman, S. H.; Dado, G. P.; Liang, G. B.; Adams, B. R. J. Am.

Chem. Soc. 1991, 113, 1164–1173.

(24) Cooper, R. D. G.; Herron, D. K. Eli Lilly Co.: France, 1975.

(25) Ohi, N.; Aoki, B.; Shinozaki, T.; Moro, K.; Kuroki, T.; Noto, T.;

Nehashi, T.; Matsumoto, M.; Okazaki, H.; Matsunaga, I. Chem. Pharm.

Bull. 1987, 35, 1903–1909.

(30) Stevens, E. S.; Sugawara, N.; Bonora, G. M.; Toniolo, C. J. Am.

Chem. Soc. 1980, 102, 7048–7050.

(31) Analogous arguments apply to OH, SH, and amino NH groups, but

our focus here is the amide NH.

(26) Horner, J. H.; Musa, O. M.; Bouvier, A.; Newcomb, M. J. Am.

Chem. Soc. 1998, 120, 7738–7748.

(27) Esker, J. L.; Newcomb, M. Tetrahedron Lett. 1992, 33, 5913–5916.

(32) Baxter, N. J.; Williamson, M. P. J. Biomol. NMR 1997, 9, 359–369.

(33) Cierpicki, T.; Otlewski, J. J. Biomol. NMR 2001, 21, 249–261.

(34) Kessler, H. Angew. Chem., Int. Ed. 1982, 21, 512–523.

J. Org. Chem. Vol. 74, No. 17, 2009 6513

Lessene et al.

JOCArticle

constant can be retrieved providing the ratio of the two

conformers (eq 1)³⁵

TABLE 1. Interpretation of NMR Spectroscopic Data in Terms of

Conformational Behavior (These Rules Are a Synthesis of Findings from

Refs 29 and 30 in Chloroform and Ref 34 in DMSO)^a

δ_A-δ_obs

AaB K_eq

¼

ð1Þ

δ_obs-δ_B

where δ_Ais the chemical shift of a given proton signal in

conformation A, δ_Bis the chemical shift of a given proton

signal in conformation B, and δ_obsis the observed chemical

shift.

Equation 1 is most useful when δ_Aand δ_Bare well

separated and is highly suited for situations where one

conformation (A) is hydrogen bonded and the other (B) is

not (since protons usually undergo a significant downfield

shift upon hydrogen bonding).

The temperature dependence of an NH chemical shift has

also been used as a tool to study hydrogen bonding.³⁶Early

experiments involved conformationally rigid cyclic peptides

in strong hydrogen-bonding solvents where amide NH pro-

tons were either locked in an internal hydrogen bond or

hydrogen-bonded to solvent.³⁷Whereas the former gave rise

to a small negative temperature dependence of the chemical

shift (NTDCS), the latter gave rise to large NTDCS values.

Hence, a widely used interpretation for polar solvents such

as DMSO-d₆states that an NH NTDCS value of greater than

-3 ppb/K is diagnostic of an internally hydrogen-bonded

NH proton, whereas a value of less than -4 ppb/K infers a

solvent-exposed NH proton.^34,38While this interpretation

has found to have no predictability in the case of less

conformationally rigid peptides,³⁹we consider it applicable

in the context of small molecules.⁴⁰

In nonpolar solvents such as chloroform, the situation is

more complex, and equivocal results may be obtained,³⁹

especially if conformational equilibria are involved. In semi-

nal papers by Gellman et al., a series of di- and triamides were

used as tools in IR and NMR studies to develop an under-

standing of how NTDCS values in nonpolar solvents can be

applied to the interpretation of conformationally heteroge-

neous systems.^29,41-44The conclusions for studies under-

taken in CDCl₃and CD₂Cl₂are summarized in Table 1.

Briefly, a small NTDCS of the NH proton (>-5 ppb/K) can

indicate one of three possibilities: (i) an equilibrium between

^aThis summary relates mainly to a two-state system, which applies to

our case. Other more complex behaviors can be observed with molecules

inherently more flexible or having more than one hydrogen bond donor or

acceptor. For example, a small positive NTCDS has been observed for a

non-hydrogen-bonded NH becoming increasingly hydrogen bonded in an

enthalpically less favored conformation with increasing temperature.

From our reading it is clear that these rules are still not fully worked out

and our interpretations are only intended to be guides.^{42 b}Amounts of

water can influence the results of the NMR analysis.^29,54Dry CDCl₃was

c

used for this study. To put this into context, it has been stated in one

particular case that a large enthalpic difference of around 6 kJ/mol in favor

of A over B could correspond to a NTDCS of -10 ppb/K. The proportion

of B in this case increases slightly with increasing temperature.^{29 d}The

degree of enthalpic difference has been related to the how close the

geometry of the hydrogen bond in A is to the optimum linear geometry,

but this can be readily counteracted by steric strain, for example.⁴⁵

an internally hydrogen-bonded conformation and non-hy-

drogen-bonded conformation where the two conformations

are of similar enthalpy; (ii) a proton mostly engaged in an

intramolecular hydrogen bond whose conformational envir-

onment is invariant over the course of the experiment; or

(iii) a proton mostly not engaged in an intramolecular

hydrogen bond whose conformational environment does

not change over the course of the experiment. On the other

hand, a large NH proton NTDCS (<-6 ppb/K) implies that

the hydrogen-bonded and the solvent exposed states are

in equilibrium but that neither of these states is largely

preferred and A is enthalpically favored. The complexity of

the interpretation of NMR data underscores the importance

for dual consideration of the IR data to unequivocally

address the presence of one or more conformers when a

solvent such as chloroform is involved. For our study in

chloroform, we therefore chose to combine the two experi-

ments and used the rules presented in Table 1 to interpret the

data obtained.

(35) Nowick, J. S.; Holmes, D. L.; Mackin, G.; Noronha, G.; Shaka, A.

J.; Smith, E. M. J. Am. Chem. Soc. 1996, 118, 2764–2765.

(36) A slow rate of deuterium exchange is also often used to indicate

internal hydrogen bonding. In peptides and proteins, a slow rate of exchange

indicates a buried proton, which may or may not be intramolecularly

hydrogen bonded (but would usually be expected to be so), or a solvent-

exposed proton that is intramolecularly hydrogen bonded. A faster rate of

exchange indicates a solvent-exposed and non-hydrogen-bonded proton. In

principle, these experiments could be informative for small molecules where

solvent exposure is not an issue, and whether an NH proton is hydrogen-

bonded or not can be judged by the rate of deuterium exchange. However,

these experiments can be tedious to perform and results hard to interpret

where rapid equilibria between hydrogen-bonded and non-hydrogen-

bonded conformations on an NMR time scale occur, either for a peptide

or small molecule.

(37) Llinas, M.; Klein, M. P. J. Am. Chem. Soc. 1975, 97, 4731–4737.

(38) Martinezmartinez, F. J.; Arizacastolo, A.; Tlahuext, H.; Tlahuextl,

M.; Contreras, R. J. Chem. Soc., Perkin Trans. 2 1993, 1481–1485.

(39) Andersen, N. H.; Neidigh, J. W.; Harris, S. M.; Lee, G. M.; Liu, Z.

H.; Tong, H. J. Am. Chem. Soc. 1997, 119, 8547–8561.

(40) This supposition is supported by the fact that the data we obtain for

benzoylureas are consistent with interpretation v and vi in Table 1.

(41) Dado, G. P.; Gellman, S. H. J. Am. Chem. Soc. 1992, 114, 3138–3139.

(42) Dado, G. P.; Gellman, S. H. J. Am. Chem. Soc. 1993, 115, 4228–4245.

(43) Dado, G. P.; Gellman, S. H. J. Am. Chem. Soc. 1994, 116, 1054–1062.

(44) Gellman, S. H.; Adams, B. R.; Dado, G. P. J. Am. Chem. Soc. 1990,

112, 460–461.

NTDCS values can also be useful to study the thermo-

dynamics of systems undergoing rapid conformational equi-

librium on an NMR time scale. Provided that the NTDCS

6514 J. Org. Chem. Vol. 74, No. 17, 2009

Lessene et al.

JOCArticle

TABLE 2. Spectroscopic Data (See the Experimental Section for

Details)

b

c

CDCl₃

DMSO-d₆

compd^a

δ(NH)^d

NTDCS^e

δ(NH)^d

NTDCS^e

1

11.33

8.97

8.83

8.90

10.30

7.76

6.70

6.19

7.08

5.42

5.27

5.29

-4.2

-3.9

-5.0

-5.4

-9.1

-8.8

-3.9

-1.0

-2.5

-3.9

-5.8

-4.6

10.04

8.11

7.93

7.84

9.93

7.82

7.75

7.58

9.91

7.73

7.67

7.41

-1.3

-2.3

-2.0

-1.6

-4.5

-4.7

-4.6

-5.2

-4.8

-5.4

-5.7

2a

2b

2c

3

4a

4b

4c

5a

5b

5c

5d

^aCompounds 5a-d are model compounds that cannotform an internal

hydrogen bond.^50,515a is acetanilide, 5b is N-methylacetamide, 5c is

N-isopropylacetamide, and 5d is N-tert-butylacetamide. ^bSample concen-

tration: 3.0 mM. ^cSample concentration: 10.0 mM. ^dppm. ^eppb/K.

On the other hand, an exposed amide NH proton in CDCl₃

undergoes a significant downfield shift when the solvent is

changed to DMSO-d₆due to the strong hydrogen-bond-

accepting ability of the latter solvent (Figure 3b).⁴⁸A third

situation arises when the conformation differs in the two

solvents, in particular where there is a loss of an intramole-

cularly NH hydrogen-bonded conformation (Figure 3c).⁴⁹

In such a situation, the otherwise large upfield chemical shift

of an NH proton that would occur on the loss of an

intramolecular hydrogen bond would be greatly reduced

through intermolecular hydrogen bonding of this proton

with DMSO-d₆. Clearly, the observation of an upfield shift

of an amide NH proton resonance in changing from an

environment of CDCl₃to DMSO-d₆does not in itself

distinguish between cases (a) and (c) but does confirm the

presence of an intramolecular hydrogen bond in the former

solvent.

We have obtained IR data for the NH protons in the

benzoylureas in chloroform and analyzed the chemical shift

of the NH proton in each compound and its temperature

dependence, both in CDCl₃and DMSO-d₆. We will show

how this set of spectral data can be used to define the

conformational behavior of this series of compounds, which

elucidates a fundamental conformational property common

to all benzoylureas. The results of our spectroscopic study

are presented in Table 2 for the NMR data, while the IR

spectra are shown in Figure 4.

FIGURE 3. Expected changes in an NH chemical shift (shown here

using a benzoylurea system) in going from CDCl₃to DMSO-d₆and

when conformational interchange between intramolecularly hydro-

gen-bonded and nonintramolecularly hydrogen-bonded conforma-

tions can occur.

for model compounds representing the fully hydrogen-

bonded state and the fully non-hydrogen-bonded state are

known,⁴⁶it has been shown that in nonpolar solvents,

thermodynamic data can be obtained (eq 2) using K_eq

obtained at each temperature via eq 1.^43,47,45

ΔH° 1 ΔS^°

R T

ð2Þ

ln K_eq¼ -

þ

R

Hence, while interpretation of NTDCS values in nonpolar

solvents such as chloroform is much more complicated than

was initially appreciated, these experiments can provide

useful qualitative and quantitative information when

coupled with another technique such as IR spectroscopy.

Comparison of the values of chemical shift for a given

NH proton in CDCl₃or DMSO-d₆can also provide

useful conformational information.^37,47We illustrate this

in Figure 3 using a benzoylurea system rather than a generic

structure. Llinas and Klein have shown that if an amide NH

proton is intramolecularly hydrogen-bonded and its amide

carbonyl is solvent exposed, its chemical shift is further

upfield in DMSO-d₆compared with CDCl₃(Figure 3a).³⁷

The rationale for this observation is that chloroform can

donate a weak hydrogen bond to the carbonyl oxygen atom.

Class I. The IR spectrum of compound 1 showed a

diagnostic broad absorption band between 3200 and 3350

cm^-1, indicative of NH internal hydrogen bonding at this

temperature. This hydrogen-bonded NH proton conse-

quently resonates very far downfield at 11.33 ppm in the

NMR spectrum, compared with 7.08 ppm for acetanilide

(Table 2), a relative downfield shift of 4.25 ppm. In DMSO-

d₆, the NMR signal remained far downfield (10.04 ppm) but

upfield compared with the value in CDCl₃. The small

NTDCS of -1.3 ppb/K and the upfield shift from CDCl₃

(45) Gung, B. W.; Zhu, Z. H.; Zou, D.; Everingham, B.; Oyeamalu, A.;

Crist, R. M.; Baudlier, J. J. Org. Chem. 1998, 63, 5750–5761.

(46) Selection of model compounds that accurately represent the two

extreme states is not necessarily trivial (see, for example, ref 41).

(47) Gellman, S. H.; Dado, G. P. Tetrahedron Lett. 1991, 32, 7377–7380.

(48) Where a compound has an amide group with an exposed NH and an

exposed carbonyl, a downfield NH chemical shift is still observed in DMSO-

d₆relative to CDCl₃because even though both solvents induce a relative

downfield shift, the influence that DMSO-d₆has by hydrogen bonding to the

NH proton is much greater than the influence that CDCl₃has by protonating

the carbonyl group.

(49) This is not considered unusual since DMSO is a strong hydrogen

bond acceptor and could compete with an intramolecular hydrogen bond in a

nonpolar solvent. In principle, if the intramolecular hydrogen bond in CDCl₃

is very weak through participation by a poor hydrogen bond acceptor, this

situation could give rise to a downfield shift when the solvent is changed to

DMSO.

J. Org. Chem. Vol. 74, No. 17, 2009 6515

Lessene et al.

JOCArticle

FIGURE 4. Infrared spectra for compounds 1, 2a-c, 3, and 4a-c. All samples at 3.0 mM in CHCl₃. Spectra shown in absorbance mode.

to DMSO-d₆are diagnostic of a hydrogen-bonded NH

group in CDCl₃(Figure 3a). The thermal stability of the

hydrogen-bonded form in the polar, strong hydrogen bond-

accepting environment of DMSO would be expected to be

even greater in the nonpolar environment of CDCl₃. This

suggests that the CDCl₃NTDCS value of -4.2 ppb/K

equates to interpretation (ii) in Table 1 and that 1 exists in

the hydrogen-bonded form in this solvent at all temperatures

studied.

In summary, we conclude that class I compounds are likely

to adopt a hydrogen-bonded conformation in both CDCl₃

and DMSO-d₆. This conformation remains favored during

the variable-temperature experiment.⁵²

Class II. The IR spectra for compounds 2a-c showed

a classic broad NH absorption band between 3200 and

3400 cm^-1, indicative of an intramolecular hydrogen bond.

The NH resonances in the NMR spectra at 8.97, 8.83, and

8.90 ppm for 2a, 2b, and 2c, respectively, in CDCl₃are also

diagnostic of a hydrogen-bonded conformation, being sub-

stantially downfield relative to their respective non-hydro-

gen-bonded model compounds 5b, 5c, and 5d (5.42, 5.27, and

5.29 ppm, respectively), so that the relative downfield shifts

for this NH proton in 2a, 2b, and 2c relative to 5b, 5c, and

5d were 3.55, 3.55, and 3.61 ppm, respectively.⁵³These shifts

are not quite as great as the downfield shift of 4.25 ppm for

class I compounds, which is consistent with the greater

hydrogen bond strength through resonance stabilization

that an anilide NH proton should be able to provide over

an alkylamide NH proton. In DMSO-d₆, the small NTDCS

(50) Hence, 5a is a model for noninternally hydrogen-bonded conforma-

tions of 1 and 3, 5b is a model for noninternally hydrogen-bonded conforma-

tions of 2a and 4a, 5c is a model for noninternally hydrogen-bonded

conformations of 2b and 4b, and 5d is a model for noninternally hydrogen-

bonded conformations of 2c and 4c.

(51) Several different model systems for internally hydrogen-bonded and

noninternally hydrogen-bonded conformations of benzoylureas were inves-

tigated before finally choosing the most representative models used herein.

(52) We have made several other class I compounds and have found that

all behave similarly (data not shown).

(53) Both 2c and 5d exhibit a slight downfield shift of the NH proton

resonance relative to their respective counterparts 2b and 5c. We propose that

this could be due to the effect of the tert-butyl group in stabilizing carbonyl

protonation by CDCl₃. It is possible that the slightly greater relative down-

field shift in 2c is due to steric compression by the tert-butyl group pushing

the NH proton closer to the C¹carbonyl.

6516 J. Org. Chem. Vol. 74, No. 17, 2009

Lessene et al.

JOCArticle

values of -2.3, -2.0, and -1.6 ppb/K, respectively, clearly

indicate a hydrogen-bonded conformation, which is stable

over the variable temperature NMR experiment, a situation

that is likely to be even more favored in CDCl₃. We conclude,

therefore, that the CDCl₃NTDCS values for the NH proton

of -3.9, -5.0, and -5.4 ppb/K, respectively, align with

interpretation (ii) in Table 1, even though the last value lies

just outside the conventional range for this interpretation.⁵⁴

All compounds exhibited an upfield shift in the resonance

of the NH proton in changing solvents from CDCl₃to

DMSO-d₆, consistent with the situation in Figure 3a.

In summary, we conclude that class II compounds

adopt a hydrogen-bonded conformation in both CDCl₃

and DMSO-d₆. This conformation remains favored during

the variable-temperature experiment.

further upfield than those in the corresponding class II

compounds 2a-c, but not shifted upfield as much as those

in the corresponding respective model compounds 5b-d.

Using these NMR chemical shift data, we estimated the

respective proportion of the closed form to be 66%, 40%,

and 25%, respectively. The hydrogen-bonded form is, there-

fore, preferred in CDCl₃only for 4a where R²is methyl,

and changing to isopropyl and then tert-butyl gives rise to

an increasing amount of the non-hydrogen-bonded form.

The CDCl₃NTDCS data is as follows: compound 4a dis-

played a large NTDCS (-8.8 ppb/K), 4b a moderate NTDCS

(-3.9 ppb/K), and 4c a very small (-1 ppb/K) NTDCS.

Considering the IR spectra, these values are consistent with

interpretations iv, i, and iii, respectively, in Table 1.

Class IV compounds can adopt a non-hydrogen-bonded

conformation in CDCl₃, and it is expected that a strong

hydrogen-bond-disrupting solvent such as DMSO would

completely destabilize the hydrogen-bonded form, analo-

gous with the behavior observed for 3. This is confirmed

by the large DMSO-d₆NH NTDCS values of -4.7, -4.6,

and -5.2 ppb/K, respectively, for 4a, 4b, and 4c. The NH

resonance in all class IV compounds is shifted downfield in

DMSO-d₆compared with CDCl₃, markedly so for 4c, less so

for 4b, and only marginally for 4a. This is consistent with

situations b and c in Figure 3, where in the case of 4a, the

situation depicted in Figure 3b is minor but in Figure 3c is

major (as 4a is mostly internally hydrogen-bonded in

chloroform). In contrast, for 4c, the situation depicted in

Figure 3b is major and in Figure 3c is minor (as 4c is mostly

not internally hydrogen bonded in chloroform), and so there

is a more marked downfield NH chemical shift observed in

DMSO-d₆relative to CDCl₃. The situation for compound 4b

lies somewhere between those for 4a and 4c.

Class III. Compound 3 is closely related to 1 (class I) but

exhibits a distinctly different conformational behavior. At

room temperature, the IR spectrum of compound 3 shows a

strong, broad absorption band between 3200 and 3350 cm^-1

,

indicative of hydrogen bonding, and of the closed form.

However, there is a small sharp NH band between 3400 and

3450 cm^-1indicating the presence of a non-hydrogen-bonded

conformation. In chloroform, the NH ¹H NMR signal is far

downfield at 10.30 ppm but is still upfield from the correspond-

ing signal in 1, consistent with a mixture dominated by the

hydrogen-bonded form but containing a significant amount of

the non-hydrogen-bonded form. Using 1 and acetanilide as

models for hydrogen-bonded and non-hydrogen-bonded con-

formations, respectively, we can deduce from the chemical

shift value that the conformer population of 3 comprises 24%

of the non-hydrogen-bonded conformation and 76% of the

hydrogen-bonded conformation. The large NTDCS of -9.1

ppb/K is supportive of situation (iv) in Table 1.

The upfield shift observed for the chemical shift in DMSO-d₆

at 300 K compared to CDCl₃is also indicative of a mixture

dominated by an internally hydrogen-bonded conformation in

CDCl₃(Figure 3a). In the more polar solvent, compound 3

adopts a non-hydrogen-bonded conformation: the DMSO-d₆

NTDCS value of -4.5 ppb/K is diagnostic of a solvent-exposed

proton. As expected, the NH chemical shift of 9.93 ppm is very

close to the value for acetanilide in the same conditions (9.91

ppm). This downfield value is testament to the strong interac-

tion between the NH proton and DMSO for both compounds

and is only marginally upfield from the NH chemical shift of

the hydrogen-bonded form of 3 in CDCl₃. In summary, class

III compounds prefer to adopt a hydrogen-bonded form at

room temperature in chloroform; however, they are readily

perturbed toward a non-hydrogen-bonded form in polar hy-

drogen-bond-accepting solvents.⁵⁵

Class IV. The IR spectrum of compounds in this class

revealed an intriguing but consistent trend in conformational

behavior. As the N²substituent becomes progressively bulk-

ier, from NHMe (4a) to NH-i-Pr (4b) to NH-t-Bu (4c), the

intensity of the IR absorption band corresponding to a

hydrogen-bonded NH becomes progressively smaller and

that for a non-hydrogen-bonded NH becomes progressively

larger to become the major feature for 4c. Consistent with

this trend, the chemical shift of the respective NH protons are

In conclusion, most of the benzoylureas preferentially

adopt hydrogen-bonded conformations in chloroform with

the exception of class IV compounds 4b and 4c, for which the

equilibrium is shifted toward the non-hydrogen-bonded

conformer at room temperature. In DMSO, compounds

from classes I and II retain a stable hydrogen-bonded con-

formation, but class III and IV compounds adopt the non-

hydrogen-bonded conformation.

Conformation in the Solid State. The inherent good crystal-

linity that we observed for class IV compounds naturally lent

them to X-ray crystallographic analysis. Shown in Figure 5a

is the solid-state structure of 4a, which crystallized from

acetonitrile in the twisted conformation. The spectroscopic

data for 4a had indicated that while it preferred to exist as the

closed form in CDCl₃, the polar environment of DMSO

stabilized the twisted form, and it appears the crystal-pack-

ing stabilization has done so similarly. In the crystal lattice,

this compound formed loose dimers with favorable electro-

static interactions between the N²-H of one molecule with

the C¹dO of another and the N²-H of the second molecule

with the C¹dO of the first. The lack of propensity for this

compound to form tight dimers, even when crystallized,

was illustrated by the fact that the distance between the

respective N²atom with its intermolecular O¹counterpart

was long, 3.16 A.⁵⁶Similar remarks apply to 4c, which also

˚

crystallized from acetonirile in the twisted conformation, as

shown in Figure 5c.

(54) Adrian, J. C.; Wilcox, C. S. J. Am. Chem. Soc. 1991, 113, 678–680.

(55) We have synthesized other class III compounds, and all behave

similarly (data not shown).

(56) Steiner, T. Angew. Chem., Int. Ed. 2002, 41, 48–76.

J. Org. Chem. Vol. 74, No. 17, 2009 6517

Lessene et al.

JOCArticle

FIGURE 5. Molecular structure and crystal structure of (a) 4a; (b) compound 6 from Ref 16; (c) 4c; (d) 7 from Ref 23; and (e) 8.

Focusing on 4a, the phenyl ring is significantly out-

of-plane with the adjoining carbonyl group by 35.2°. The

C¹-N¹amide bond is in a Z conformation causing R¹to be

trans to the benzoyl ring but eclipsed with the C¹carbonyl

group. Conversely, the N¹-C²amide bond is neither E nor

Z but twisted so that R¹is not planar with the C²-N²amide

group. The net effect is to cause the plane of the C²(dO)-N²

amide to face toward the plane of the benzoyl ring. The

amide plane can face either way, and indeed, the “acylurea

twist” in 4c is the mirror image of the twist in 4a, with

a dihedral angle for C¹-N¹-C²dO²of 121.4° in 4a and

-130.5° in 4c. The C²-N²amide is in a Z conformation

according to our definition in Figure 2. For comparison, we

have included in Figure 5b the published crystal structure of

compound 6, one of only three other representatives of class IV

benzoylureas in CCD. The overall orientation of the acylurea

scaffold is very similar in all these compounds, and the twisted

conformation in 6 is essentially identical to that in 4a.

In stark contrast, those benzoylureas in this study that

preferred the closed conformation did not yield so readily to

X-ray crystallographic analysis. We attributed this to the

decreased ability of the closed conformation to form favor-

able intermolecular electrostatic interactions when R¹¼ H

compared with the twisted form. This is supported by the

observation that the crystal structures of benzoylureas that

are unsubstituted at N¹, and are consequently able to form

favorable electrostatic intermolecular interactions in the

solid state, are well represented in the CCD, numbering 21.

However, there is only one example of a benzoylurea with an

unbranched N¹alkyl group, this being the class II bis-

benzoylurea 7 with an N¹-Me substituent.^57,23As shown in

Figure 5d, 7 adopts a closed conformation, as was expected

on the basis of the solution conformational behavior that we

observed for class II compounds. Our attempts to crystallize

compounds in the closed conformation with an alkyl

R¹group were successful with only one compound made in

the course of our research, that being 8 (Figure 5e). This is,

therefore, the first report of a class III benzoylurea crystal

structure. This particular example shows that the six-mem-

bered ring maintained by the hydrogen bond is coplanar with

the anilide phenyl ring.

In summary, the X-ray crystallographic results support

the notion that the hydrogen-bonded and non-hydrogen-

bonded conformation populations observed by NMR and

IR spectroscopy represent the closed and twisted conforma-

tions, respectively, observed in the solid-state structures of

benzoylureas.

Our experimental data show that benzoylureas can adopt

two distinct conformations: a closed form stabilized by an

intramolecular hydrogen bond and a twisted form that does

not contain an intramolecular hydrogen bond. We have

shown that conformational preference is determined by the

substituents, or indeed both forms can exist in solution,

(57) There are no examples of class III benzoylureas in the Cambridge

Crystallographic Database.

6518 J. Org. Chem. Vol. 74, No. 17, 2009

Lessene et al.

JOCArticle

FIGURE 6. Factors influencing the conformation of benzoylureas.

rapidly interconverting on an NMR time scale. Such

dynamic conformational behavior was not expected for this

hindered and partially conjugated system. There are several

features of these compounds that influence their confor-

mational preference as shown schematically in Figure 6.

Fundamentally, it is a matter of whether the strength of the

intramolecular hydrogen bond stabilizing the closed form,

which can be influenced by the nature of the R²substituent,

can compensate for the opposing hindrance in the closed

form between the benzamide ring and R¹and between

R¹and the C²carbonyl group, the magnitude of which

clearly depends on the steric bulk of the R¹substituent. An

additional challenge to the intramolecular hydrogen bond is

whether it can compete in the environment of a strong

hydrogen-bond-accepting solvent such as DMSO. These

factors will now be discussed in more detail.

FIGURE 7. IR spectrum of compound 8.

TABLE 3. Chemical Shift and TDC for Compounds 8

Influence of Substituent R¹. In order to alleviate steric

hindrance induced in the closed form, the benzoyl phenyl

ring can rotate relative to the plane of the carbonyl group

C¹dO¹. Alternatively, conversion to the Z conformation

around the C¹-N¹bond and rotation about the N¹-C²

bond leads to a twisted conformation, as shown by the

crystal structures of compounds 4a and 4c (Figure 5). This

ameliorates the hindered environment of R¹by replacing

the cis C²dO²carbonyl group in the closed form with an out-

of-plane N²H group in the twisted form. Thus, R¹has only

one steric partner in the twisted form (the C¹carbonyl group) but

two in the closed form (the C²carbonyl group and the C¹phenyl

ring). This explains why 3 converts to the twisted conformation

more readily than 1. In both compounds, we expect the C¹

phenyl ring to be rotated significantly out-of-plane relative to the

C¹carbonyl group so that steric hindrance is minimized in the

closed form.⁵⁸We propose that this is sufficient to maintain

the closed conformation for 1 in both CDCl₃and DMSO-d₆but

not for 3 with its bulkier R¹group.

a

b

CDCl₃

DMSO-d₆

compd

δ(NH)^c

NTDCS^d

-3.0

δ(NH)^c

NTDCS^d

8 (closed)

8 (twisted)

11.59

10.53

9.44

∼ 0

-6.0

^aSample concentration: 3.0 mM. ^bSample concentration: 10.0 mM.

^cppm. ^dppb/K.

to reduce steric hindrance, rather than through a sterically

less demanding R¹group such as in 1, and that this subtle

effect could be sufficient to stabilize the closed conformation

such that 8 would mirror the conformational behavior of 1.

Indeed, unlike 3, the IR spectrum of 8 in chloroform

exhibited only a hydrogen-bonded NH absorption band

(Figure 7) and, as shown in Figure 5e, 8 crystallized in the

closed conformation, despite R¹being a bulky isopropyl

group.⁶⁰

In CDCl₃the exclusively hydrogen-bonded NH signal in 8

(Table 3) is displaced far downfield at 11.59 ppm, and the

small NTDCS value of -3.0 ppb/K therefore corresponds to

interpretation ii in Table 1. Also, unlike 3, 8 maintained some

In order to test this hypothesis, we prepared compound 8

(Table 3), in which the 2,6-dimethyl substitution forces the

plane of the phenyl ring to be roughly perpendicular to the

plane of the carbonyl group.⁵⁹Hence, the local steric inter-

action between the C¹phenyl ring and R¹in the closed form

should be decreased to the greatest possible extent. We

anticipated that this could provide an alternative approach

(58) Tertiary benzamides readily rotate about the phenyl-carbonyl bond

to accommodate highly distorted but less sterically hindered structures. A

survey of 36 tertiary benzamides in the Cambridge Crystallographic Data-

base revealed deviation from planarity of 35-85°, with a Ph-CONR²

torsion angle of around 75° predominating.

(60) It cannot be ruled out that the closed conformation in 8 is not

relatively stabilized but that the twisted conformation is relatively destabi-

lized. However, scrutiny of the twisted form of these compounds reveals no

reason why 2,6-dimethyl substitution should have such an effect.

(59) Bowles, P.; Clayden, J.; Helliwell, M.; McCarthy, C.; Tomkinson,

M.; Westlund, N. J. Chem. Soc., Perkin Trans. 1 1997, 2607–2616.

J. Org. Chem. Vol. 74, No. 17, 2009 6519

Lessene et al.

JOCArticle

FIGURE 8. Comparison between (a) benzanilide and (b) benzoyl-

urea.

closed conformation even in DMSO, and this could be

observed as a separate NH signal downfield at 10.53 ppm

in addition to an upfield signal at 9.44 ppm that corre-

sponded to the twisted conformation.⁶¹This allowed us to

measure unequivocally the NH NTDCS values in DMSO-d₆

for the closed and twisted forms. As listed in Table 3, these were

∼0 and -6 ppb/K, respectively, confirming that in DMSO-d₆

an NTDCS above -3 ppb/K is diagnostic of an internally

bonded NH proton while a value of less than -4 ppb/K is

diagnostic of an NH proton hydrogen-bonded to solvent.

Thus, the greater propensity that we observe for benzoy-

lureas with a branched R¹group to adopt the twisted

conformation seems to be driven principally by steric effects

and explains why we observe that the closed form of class III

and IV compounds (R¹=branched) is destabilized relative to

the closed form of class I and II compounds (R¹= Et),

respectively.

In the appropriate environment, an amide group can be

considered a phenyl ring isostere due to its planarity and

partial resonance.^62-64Replacement of the C²(dO)N²R²

amide group in benzoylureas with a phenyl ring provides a

benzanilide that can be used as model systems for the

benzoylureas in which the R¹steric effects are maintained

but the ability to stabilize a closed-like conformation with a

hydrogen bond is removed, as shown in Figure 8.

FIGURE 9. Structure comparison between benzanilides 9 (Ref 65)

and benzoylurea 4c.

arrangement to the C¹(dO)-N¹amide group for both

compounds.

It is clear from the comparison with the benzanilides that

the closed-type conformation in both systems is inherently

sterically demanding.⁶⁶This is exacerbated by the steric bulk

of R¹to the point where the twisted conformation, which

alleviates some of this steric hindrance, becomes favored.

Twisting occurs more readily for benzanilides than benzoy-

lureas because the latter can stabilize the closed conforma-

tion through formation of an intramolecular hydrogen

bond.

Influence of Substituent R². Class IV compounds clearly

show that the nature of R²influences conformational pre-

ference between the closed and twisted forms. In the series

4a-c (R²= Me, i-Pr, t-Bu), the closed form becomes

progressively less stable than the twisted form, comprising

66%, 40%, and 25%, respectively, of the conformer popula-

tion in CDCl₃at room temperature. For these compounds,

the R²substituent becomes progressively bulkier and more

electron-rich while R¹remains invariant. The R¹group in

this class is i-Pr, which we have shown introduces steric

hindrance between itself and both the benzoyl phenyl ring

and the C²dO carbonyl group. Thus, there is a propensity

for these compounds to adopt the twisted form when the

intramolecular hydrogen bond is no longer sufficient to

stabilize the closed form. It would be expected that R²is

decreasingly able to stabilize the internal NH bond in the

series R²=Me, i-Pr, t-Bu based on the increasing electron-

rich character.

Indeed, while benzanilides favor the “closed” conforma-

tion (a) in Figure 8 when R¹=H, they switch to a twisted-like

conformation even when R¹= Me with the amide group

presenting an E conformation as observed for the twisted

benzoylurea conformation.^62,65This is clearly shown in

Figure 9, where the anilide ring for benzanilide 9 and

benzoylurea 4c assumes the same almost perpendicular

(61) Separation of the NH signals for the two conformations was an

unanticipated but useful consequence of the 2,6-dimethyl substitution slow-

ing down the dynamics of conformational interchange. We attribute this to

steric hindrance, though a similar phenomenon in certain 2,6-disubstituted

tertiary benzamides has been related to an electronic effect on amide

resonance. See: Lewin, A. H.; Frucht, M. Tetrahedron Lett. 1970, 11,

1079–1082.

While R²can affect the electronic nature of the hydrogen

bond and, therefore, its conformation, this does not discount

at this point a steric contribution of R²to the conformational

preference. Shown in Table 4 are the bond angles at the

N²amide nitrogen atom in selected benzoylureas in the solid

state; this angle reflects the steric environment of the amide

group since its magnitude affects the distance of R²from the

eclipsed carbonyl group.

(62) Stewart, W. E.; Siddall, T. H. Chem. Rev. 1970, 70, 517.

(63) Curran, D. P.; Hale, G. R.; Geib, S. J.; Balog, A.; Cass, Q. B.;

Degani, A. L. G.; Hernandes, M. Z.; Freitas, L. C. G. Tetrahedron:

Asymmetry 1997, 8, 3955–3975.

(64) Ahmed, A.; Bragg, R. A.; Clayden, J.; Lai, L. W.; McCarthy, C.;

Pink, J. H.; Westlund, N.; Yasin, S. A. Tetrahedron 1998, 54, 13277–13294.

(65) Itai, A.; Toriumi, Y.; Saito, S.; Kagechika, H.; Shudo, K. J. Am.

Chem. Soc. 1992, 114, 10649–10650.

(66) While π-π stacking appears as though it could contribute to the

stability of the twisted conformation of benzanilides (aryl-aryl stacking) and

benzoylureas (aryl-amide stacking) as they are depicted in Figure 8, this is

either not the case or is insignificant, since N-methylalkylanilides also adopt

the twisted conformation (see ref 63). Interestingly, this type of conforma-

tional interchange where a twisted form alleviates steric hindrance appears to

be a general phenomenon as analogous behavior is observed in N-alkyl-N⁰-

phenylureas (see: Nowick, J. S.; Abdi, M.; Bellamo, K. A.; Love, J. A.;

Martinez, E. J.; Noronha, G.; Smith, E. M.; Ziller, J. W. J. Am. Chem. Soc.

1995, 117, 89–99) and has been used as a deliberate design strategy to occupy

large pockets in protein-protein interactions. See: Park, C.-M.; Oie, T.;

Petros, A. M.; Zhang, H.; Nimmer, P. M.; Hensry, R. F.; Elmore, S. W. J.

Am. Chem. Soc. 2006, 128, 16206–16212.

Steric hindrance in an increasingly bulky R²group in the

series 4a (R²=Me), 4c (R²=t-Bu), 8 (R²=Ph) is marked by

an increase in this bond angle. However, in compounds 4a

and 7, where R²is small, both twisted and closed conforma-

tions, respectively, are observed. Moreover, in compounds

4c and 8, in which R²is bulky, both twisted and closed

conformations, respectively, are also observed. Thus, the

size of R²does not appear to influence conformational

6520 J. Org. Chem. Vol. 74, No. 17, 2009

Lessene et al.

JOCArticle

TABLE 4. Bond Angle (deg) around the N²Amide Nitrogen Atom in

Selected Benzoylureas, Derived from Crystal Structure Data

TABLE 5. DFT Calculations of Gas-Phase Energies of Model Acyl-

ureas MeC(dO)NMeC(dO)NHR in Twisted Conformation Compared

with the Closed Conformation^a

R

ΔH (kJ mol^-1

)

Me

i-Pr

t-Bu

Ph

31.0

29.0

29.1

30.6

compd (conformation)

(C²-N²-R²)

d(N O)^a

R²

3 3 3

4a (twisted)

4c (twisted)

7 (closed)

8 (closed)

a

121.2

124.7

121.2

126

n.a.

2.64

2.58

Me

tBu

Me

Ph

^aB3LYP/6-311þG(2df,p)//B3LYP/6-31G(d) calculations, reported

as the amount by which the enthalpy of the twisted form is higher than

that of the more favored closed form. Computational details are

provided in the Supporting Information.

˚

(N-O) distance measured in A. n.a.: this distance was not measured

in the twisted form.

Third, the N²-O¹distance in a class II benzoylurea such as

7 is 2.64 A and in a class I benzoylurea such as 8 is 2.58 A. A

˚

value of between 2.48 and 2.65 A has been used to indicate

preference. Compound 3 is a useful probe here because R¹, as

in 4a-c, is an isopropyl group, but R²is a phenyl group. The

in-plane R²phenyl group in 3 would be expected to exert

more steric hindrance than the R²tert-butyl group in 4c, and

yet, the NMR data indicate that while 4c in CDCl₃is

predominantly in the twisted conformation, 3 is mainly

closed. Indeed, the proportion of 3 in the closed conforma-

tion (76%) is even greater than that in 4a, where R²is a small

methyl group. These data are not in accordance with any

steric influence of R²on conformational preference. Rather,

they are completely in accordance with the trend of increas-

ing hydrogen-bond donor abilities expected along the series

R²=t-Bu, i-Pr, Me, Ph.

the presence of a very strong hydrogen bond of around

.

30-40 kJ mol^{-1 19,68}

In order to estimate the strength of the hydrogen bond, we

applied density functional theory (DFT) calculations to

model acylureas in either the closed or twisted conforma-

tions.⁶⁹We reasoned that the use of simple methyl groups for

model acyl urea system MeC(dO)NMeC(dO)NHR would

minimize the influence of sterics and give a reasonable

estimation of hydrogen bond strength in the gas phase. As

shown in Table 5, these calculations confirm that in the gas

phase, the hydrogen bond strength in acyl ureas is indeed

estimated to be very strong, at approximately 30 kJ mol^-1

.

In summary, the bulkiness of the R²group does not affect

the conformational preference because the steric effects are

similar in both the closed and twisted forms of benzoylureas.

It is the electronic effect that the R²group has on increasing

the hydrogen bond strength in the closed form that stabilizes

this conformation. Such stabilization is, however, readily

overwhelmed by the nature of the R¹group because the

closed conformation is inherently sterically more demand-

ing. Thus, the hydrogen bond strength in 3 is intrinsically

stronger than in class II compounds, and yet it is the latter

class with R¹=ethyl that maintains the closed conformation

in both chloroform and DMSO, whereas 3 already exhibits

some twisted conformation in chloroform and adopts an

entirely twisted conformation in DMSO.

In carbon tetrachloride, a maximum enthalpic gain ob-

tained from the formation of a geometrically optimal

amide-amide intramolecular hydrogen bond has been re-

ported at around 17 kJ mol .

^{-1 29}When this geometry is not

optimal this enthalpic gain is readily decreased and not able

to compensate for small gauche destabilizing interactions. In

our case, we have shown that R¹induces significant steric

constraints on the molecular scaffold. In addition, the ben-

zoylurea core does not place the hydrogen bond partners in

an optimal orientation. Despite this, our calculations suggest

a solvation effect of approximately 10 kJ mol^-1in transfer-

ring from the gas phase to chloroform (see the Supporting

Information). This places the enthalpic gain from hydrogen-

bond formation in benzoylureas greater than that for two

amides in the optimum, linear geometry.

Hydrogen Bond Strength of the Benzoylurea Core in the

Closed Conformation. A number of observations suggested

to us that the strength of the intramolecular hydrogen bond

in the closed form of benzoylureas could be relatively

strong.⁶⁷First, some of the compounds such as 1 and 2a-c

maintain an intramolecular hydrogen bond even in DMSO-

d₆, despite the fact that the NH is not buried but rather is

readily accessible to solvent. Second, the downfield shift of

the intramolecularly hydrogen-bonded NH proton in the

NMR spectrum of several benzoylureas is very significant

relative to model compounds for the non-hydrogen-bonded

NH proton. In the case of 8, the NH proton in the NMR

spectrum resonates in CDCl₃at 11.59 ppm and the downfield

shift relative to acetanilide is 4.51 ppm in the same solvent.

It can be observed that the enthalpy decreases slightly on

going from R=methyl, to R=isopropyl and tert-butyl but

increases again for R=Ph. While these differences are small

and should not be overinterpreted, this trend supports our

earlier proposal that electron-rich R²groups destabilize the

closed conformation in class IV compounds because of

hydrogen-bond weakening and that the steric bulk R²does

not influence conformational preference. Consideration of

the nature of the electronics of the benzoylurea core in the

closed conformation can shed light on those factors that

contribute to hydrogen bond strength.

Resonance Forms of the Benzoylurea Core in the Closed

Conformation. Hydrogen bonds are strengthened by reso-

nance forms that decrease the difference in electronegativity

between the hydrogen-bond donor (NH) and hydrogen-bond

(67) The geometry of the intramolecular hydrogen bond in the closed

conformation of the benzoylurea core is not optimal (NH O nonlinear,

3 3 3

see: Peters, D.; Peters, J. J. Mol. Struct. 1980, 68, 255-270) and comprises a

donor and acceptor of different pK_a’s (carbonyl CdO and amide NH). It is

clear from these two factors that we are not in presence of the strongest type

of hydrogen bond, which is linear and comprises a donor and acceptor of

equivalent pK_a’s (e.g., FHF). The terms “strong” or “weak” are employed as

relative characterizations and not in their absolute meaning.

(68) Note that it has been stated that firm conclusions on hydrogen bond

strength cannot be drawn from spectral and crystallographic data. See:

Guthrie, J. P. Chem. Biol. 1996, 3, 163–170. Perrin, C. L.; Nielson, J. B.

Annu. Rev. Phys. Chem. 1997, 48, 511–544.

(69) See the Supporting Information.

J. Org. Chem. Vol. 74, No. 17, 2009 6521

Lessene et al.

JOCArticle

torsion angle of 121-137° that would presumably discou-

rage substantial resonance between the amide groups along

the N¹-C²bond.⁷¹In these two structures, the amide groups

behave as if they are structurally independent of each other.

Therefore, it is quite remarkable then, that these respective

bonds lengths in 7 and 8 are revealed in Table 6 to be

essentially identical to those in 4a and 4c. Thus, even in the

closed conformation, the acylurea group behaves like two

independent amide groups, with substantial resonance with-

in each amide group as represented by resonance structure

B in Figure 10, but not between the two amide groups, as

represented by resonance structure C and which would result

in a shorter C¹-N²bond length. That is, resonance form

C appears not to contribute to an overall acylurea resonance,

and these compounds behave structurally like “bis-amides”

without any urea character.⁷²This was quite unexpected as

the planarity of the closed form should allow for maximum

conjugation between the two amide groups.⁷³

FIGURE 10. Representations of valence bond resonance forms

B and C of the benzoylurea core A and tautomeric interconversion

involving D.

acceptor (CdO). Using simple valence-bond structures of the

benzoylurea system, partial resonance can be achieved in two

different ways. As shown in Figure 10, the benzoylurea core

A could plausibly comprise resonance forms B and C. Here,

resonance form B and tautomer D would confer on each

amide group within the core a similarly resonance-shortened

C¹-N¹or C²-N²bond. Conversely, resonance form C would

lead to a shortening of the N¹-C²bond.

Conversely, the ¹³C NMR data reveal that the C²carbon

in all benzoylureas studied here, irrespective of conforma-

tion, experiences an electronic environment of a typical urea,

shifted upfield to 151-155 ppm (see the Supporting Infor-

mation) relative to a typical amide of around 170 ppm. It is

instructive that the C²carbonyl carbon atom ultimately

behaves electronically like a urea carbonyl carbon atom,

being flanked by two nitrogen atoms; this is despite the fact

that the electron sharing between the carbonyl group and the

flanking nitrogen atoms, as witnessed by the differences in

the respective N-carbonyl bond lengths, is distinctly differ-

ent from that in a typical urea, where mutual sharing

occurs.⁷⁴This categorically demonstrates that the chemical

shift of a given ¹³C carbon atom cannot be used to infer the

degree of partial double bond character with flanking atoms.

Since RAHB requires all bonds in the pseudo-six-mem-

bered ring to participate in the π-conjugated system, the

absence of C as a significant resonance form precludes any

RAHB in the benzoylureas, despite their topological simi-

larity with systems in which this effect is observed. The

resonant spacers in Gilli’s studies strictly comprise double

bonds (see C⁰) and therefore favor strong resonance to occur.

Gilli and co-workers have shown that the intramolecular

hydrogen bond strength in certain systems is similar to the

benzoylureas that also contain an intramolecular hydrogen

bond and that form a pseudo-six-membered ring and can be

enhanced by the intervention of resonance between the

acceptor and the donor.^18,19They devised a nomenclature

to define the nature of these systems, which feature sequences

such as H-N-R_ndO where “R_n(n=1, 3, 5, etc.) is a resonant

spacer of n atoms forming a chain of alternating single and

double bonds”. For example, intramolecular resonance-as-

sisted hydrogen bonding (RAHB) in β-enaminone C⁰in

Figure 11, for which n = 3 and hence the appellation

R₃-RAHB (intra), helps to stabilize sharing of the intramo-

lecular hydrogen-bonded proton between the NH and car-

bonyl oxygen atom by way of tautomer D⁰. This leads to a

˚

small d(N-O) value of 2.661 A, which has been associated

for this system with a gas-phase hydrogen-bonded energy

(E_HB) of 29 kJ mol^{-1 19}

.

By way of contrast, the nonresonant

.

hydrogen bond in E⁰was estimated to be 12 kJ mol^-1

(71) The extent of torsional distortion alone, in the absence of bond

length data, has been shown to be an unreliable gauge of the extent of

conjugation (see, for example: Chattopadhyay, D.; Mazumdar, S. K. Z.

Kristallogr. 1986, 177, 103-106).

By analogy with the β-enaminone system, the interconver-

sion between C and D in Figure 11 would suggest that

benzoylureas could in principle exhibit RAHB.

Bond lengths give insight into the electronics underpin-

ning the presence or absence of resonance stabilization and

thus can help determine the relative distribution of the

individual resonance forms. Shown in Table 6 are selected

bond length values for compounds 4a, 4c, 7, and 8.

(72) This difference is starkly demonstrated by reference to C-N bond

lengths in the CCD. There is only one entry with an R factor of <0.05 for an

N-alkyl-N-phenyl-N’-phenylurea. Here, the N-carbonyl and N’-carbonyl

˚

bond lengths are both 1.370 A, in stark contrast with the values of 1.43

˚

and 1.33 A, respectively, for the benzoylureas in the current study. This may

be a general phenomenon for acylureas, as the more common unsubstituted

benzoylureas in the CCD (i.e., with R¹=H) exhibit N¹-C²bond lengths of

1.41 A, only marginally shorter than those of a single bond. We are not aware

˚

of this phenomenon having previously been reported.

(73) The two amide groups are by no means equivalent as the C¹-N¹

distance is consistently longer than the C²-N²distance in both closed and

twisted forms. This could perhaps be attributable to the electronegativity of

the C²carbonyl attached to N¹that might discourage substantial N¹-C¹

resonance. In any case, neither amide group is “typical”: the average HN-

carbonyl bond length in benzanilides and alkylacylanilides for entries in the

The twisted forms of 4a and 4c serve as useful controls

where intramolecular hydrogen bonding is removed as a

factor that could influence acylurea electronics. In these two

control systems, the C¹-N¹and C²-N²bond lengths are

˚

1.36/1.38 A and 1.32/1.33 A, respectively. Both are substan-

˚

1

2

CCD is the same, at 1.347 A, and this is not affected by whether the NH is

intramolecularly hydrogen bonded or not. Only 2/40 (5%) of representative

˚

tially shorter than the N -C bond lengths of 1.43 A, which

corresponds to the length of a pure single C(sp2)-N(sp2)

bond.⁷⁰This is not unexpected considering the N¹-C²

˚

entries in the CCD reveal NH-carbonyl bond lengths shorter than 1.335 A

1

2

˚

and none are longer than 1.375 A. Thus, the C -N and C -N distances are

remarkably longer and shorter, respectively, for benzoylureas in the current

study than in a comparable amide bond. A full investigation into why this is

so is beyond the scope of this paper.

(70) Gilli, G.; Bertolasi, V.; Bellucci, F.; Ferretti, V. J. Am. Chem. Soc.

1986, 108, 2420–2424.

(74) Bharatam, P. V.; Moudgil, R.; Kaur, D. J. Phys. Chem. A 2003, 107,

1627–1634.

6522 J. Org. Chem. Vol. 74, No. 17, 2009

Lessene et al.

JOCArticle

FIGURE 11. Example in a β-enaminone of R₃-RAHB, as defined by Gilli et al.^18,19

˚

TABLE 6. Bond Distances (A) in Compounds 4a,c, 7, and 8

compd (conformation)

d(C¹-O¹)

d(C¹-N¹)

d(N¹-C²)

d(C²-O²)

d(C²-N²)

d(O¹N²)

torsion angle C¹-N¹-C²-O²(deg)

3 3 3

4a (twisted)

4c (twisted)

7 (closed)

1.22

1.23

1.36

1.38

1.37

1.43

1.44

1.22

1.21

1.22

1.21

1.32

1.33

1.35

N/A

2.64

2.58

121.4

-130.6

179.3

8 (closed)

-172

TABLE 7. Conformationally Constrained Compounds

FIGURE 12. Comparison of δ(NH) for compound 10 from ref 29

and compound 2a.

b

c

CDCl₃

DMSO-d₆

compd^a

δ(NH)^d

NTDCS^e

δ(NH)^d

NTDCS^e

It appears that in the benzoylurea system the two amides, by

being decoupled electronically, occupy a low energy state

that also allows for a sufficiently strong hydrogen bond to

stabilize the closed form in certain cases.⁷⁵The hydrogen

bond does not itself influence the resonance, as the amide

bond lengths in Table 6 are the same for closed and twisted

forms.

11a

11b

12

9.36

11.77

11.69

-1.9

-2.2

9.19

11.33

11.0

-1.8

-2.1

^aThese compounds showed a characteristic hydrogen-bonded NH

b

broad IR band (see the Supporting Information). Sample concentra-

tion: 3.0 mM. ^cSample concentration: 10.0 mM. ^dppm. ^eppb/K.

Effect of Conformational Rigidity on the N²H Chemical

Shift in the Closed Formation. The hydrogen-bonded NH

proton in benzoylureas resonates significantly further down-

field than even some closely related intramolecularly hydro-

gen-bonded bis amides such as 10 (Figure 12). The chemical

shift for the NH proton in this compound has been reported

by Gellmann et al. to be 7.7 ppm (in CD₂Cl₂), which is

1.3 ppm further upfield than the corresponding NH signal in

2a.²⁹

TABLE 8. van’t Hoff Analysis Performed on Compounds 3 and 4a-c^a

Parameter

3

4a

4b

4c

ΔG

-2.85

-5.32

-8.3

-1.70

-4.89

-10.7

1.01

1.47

1.52

2.74

4.39

5.52

The intrinsic hydrogen bond strength in 2a should be

similar to that in 10, in which amide resonance forms could

similarly contribute. The downfield chemical shift of the NH

proton in 2a relative to 10 is consistent with a reduced

conformational flexibility in 2a, caused by the amide C¹-N¹

bond linkage, increasing the probability of the closed form

over noninternally hydrogen-bonded conformations relative

to 10.^56,76However, the N¹-C²bond in 2a is essentially single

in character and probably does not confer any extra rigidity

relative to the corresponding bond in 10. The phenyl ring of

the benzoyl group does not appear to influence the NH

chemical shift, which in alkanoylureas is reported to resonate

between 8.75 and 9.32 ppm in CDCl₃.^77,78

ΔH°^b

ΔS°^c

^aCompounds 1, 2a, 2b, and c were used as references for the hydrogen-

bonded conformation of 3, 4a, 4b, and 4c, respectively, and compounds

5a, 5b, 5c, and 5d were used as references for the non-hydrogen-bonded

conformations of 3, 4a, 4b, and 4c, respectively, in these calculations.

van’t Hoff analysis was performed using eqs 1) and 2) at 298.15 K. ^bIn kJ

mol^-1. ^cIn J mol^-1K^-1

.

In compounds 11a and 11b shown in Table 7, we covalently

constrained a benzoylurea to further increase its rigidity to

investigate how this may affect the NH chemical shift. This

modification directly decreases torsional flexibility about

the phenyl-carbonyl, C¹-N¹, and N¹-CH₂bonds and by

pinning back the N¹substituent also indirectly stabilizes the

N¹-C²bond by reducing hindrance between the N¹alkyl

group and the C²carbonyl group. As shown in Table 7, the

NH signals in the NMR spectra of 11a and 11b resonate far

downfield, more so in both CDCl₃and DMSO than their

respective less constrained counterparts 2a and 1 in Table 2.

(75) Interestingly, benzoylureas do not appear in Gilli’s survey of crystal-

lographic data, although papers by Smith (see ref 23) and Kohmoto (see ref

16) containing the X-ray structure of compound 10 were published before

Gilli’s study.

(76) Bilton, C.; Allen, F. H.; Shields, G. P.; Howard, J. A. K. Acta

Crystallogr. B 2000, 56, 849–856.

(77) Lin, W. O.; Altoe, A. P. Monatsh. Chem. 1982, 113, 101–109.

(78) Yogo, M.; Hirota, K.; Senda, S. Chem. Pharm. Bull. 1982, 30, 1333–

1337.

J. Org. Chem. Vol. 74, No. 17, 2009 6523

Lessene et al.

JOCArticle

FIGURE 13. Substituent influence on benzoylurea conformation in CDCl₃and DMSO-d₆.

We also prepared 12 as a “rigidified” class I compound,

on the basis of our findings in our investigation on the

conformational influence of R¹that the 2,6 dimethyl sub-

stitution in compound 8 helped to lock class III compounds

into the closed conformation. As shown in Table 7, the

NH NMR data for compound 12 are very similar to those

for compound 11b, despite the fact that the C¹phenyl rings

are expected to adopt perpendicular orientations. Results

from 11b and 12 suggest that the torsional aspect of the

C¹-phenyl ring has little influence on the electronics of

benzoylureas.⁷⁹Rather, the further downfield NH chemical

shift in 11a, 11b, and 12 relative to their less constrained

counterparts 2a, 1, and 1, respectively, is due to increased

rigidity leading to a higher probability of the intramolecular

hydrogen-bonded conformation. However, this effect on the

NH chemical shift within the acylurea series is less marked

than the one observed when the bis-amide core 10 (Figure 12)

is compared with the next closest analogue acylurea 2a

(Figure 10).

What is more interesting is the fact that the equilibrium

entropy favors the twisted conformation for 3 and 4a, but

the closed form for 4b and 4c; this suggests that the bulky R¹

and R²groups in 4b and 4c limit the conformational freedom

of these molecules.

Summary. It is now possible to draw a more complete

picture of the conformational behavior of the benzoylurea

scaffold. The steric interaction between the substituent

R¹and the phenyl ring is the key factor influencing the

conformation. Most of the compounds adopt a closed con-

formation in chloroform at room temperature. In this con-

formation, the intramolecular hydrogen bond overcomes

inherent steric hindrance arising from interactions between

the benzoyl phenyl ring and the C²dO amide carbonyl group

and R¹. Thus, the closed conformation of benzoylureas

becomes relatively less stable when R¹substituents are

branched and hence bulkier, and if the internal hydrogen

bond is weakened due to an electron-rich R²substituent. The

extreme case is represented by 4c, which exists predomi-

nantly in the twisted form, even in CDCl₃. A more polar

solvent such as DMSO accentuates this trend, and for those

compounds with nonfavorable steric interactions and weak-

er internal hydrogen bonding, the switch to the twisted

conformation occurs more readily because of the competi-

tive hydrogen-bond-acceptor ability of this solvent. Fig-

ure 13 summarizes the expected conformational behavior

of benzoylureas at 300 K in the two solvents studied.

Conclusion. De novo interactive design, as for many other

drug discovery strategies, relies on the innovative use of

suitable scaffolds. For our purposes, we have chosen ben-

zoylureas for their ability to form a central six-membered

ring induced by an intramolecular hydrogen bond maintain-

ing the molecule in the correct conformation required for

mimicry of the R-helix. We have shown that both steric and

electronic effects, as well as the relative strength of an

internal hydrogen bond, are critical factors influencing the

conformation of this scaffold in a given environment. We

have also shown that this simple molecular system has

interesting conformational properties and that the appella-

tion “benzoylurea” is rather misleading as the structure

behaves essentially as two sequential amide groups rather

than an acylated urea; a more appropriate appellation would

Thermodynamic Analysis of Benzoylurea Conformational

Interchange. To obtain further insight into the thermody-

namics of benzoylurea conformational change, we used the

NMR data to derive energetic information on compounds 3

and 4a-c in chloroform using a van’t Hoff analysis

(Table 8).

The enthalpy values shown in Table 8 forcompounds 3 and

4a show that for these compounds the hydrogen bond is able

to overcome steric repulsion in the closed state and further

stabilize this conformation by -5.32 and -4.89 kJ mol^-1

,

respectively. The data in Table 8 confirm that when R²is an

isopropyl (4b) or tert-butyl (4c) group, the equilibrium

enthalpy favors the twisted form due to increasing steric

repulsion in the closed form. The values obtained are in

agreement with our interpretation of the conformational

behavior of class IV compounds and especially of compound

4b (small NTDCS, compound equilibrating between two

conformers of similar enthalpy, i in Table 1), for which we

obtain the lowest ΔH° values in the series (1.47 kJ mol^-1).

(79) The influence of increased rigidity on the NH chemical shift is why 1,

for example, is a better model than constrained compounds 12 or 11b for the

closed conformation of 3 in our analysis of conformational equilibria in 3

using the NH chemical shift value.

6524 J. Org. Chem. Vol. 74, No. 17, 2009

Lessene et al.

JOCArticle

therefore be “bisamide”. These compounds may, therefore,

comprise useful probes for investigations of peptide hydro-

gen bonds as they represent a system where the two amide

groups are linked by the shortest possible distance, and so

entropic factors, which can be a major consideration in

biological systems, are minimized.

warmed to room temperature over 4 h. The reaction was then

concentrated in vacuo. The residue was dissolved in 1 mL of dry

acetonitrile and added at 0 °C to a mixture of triethylamine

(224 μL, 1.61 mmol) aniline hydrochloride (104.2 mg, 0.8 mmol)

in 1 mL of dry acetonitrile. The reaction was then stirred at room

temperature for 1 h after which time the reaction was complete.

The reaction was then concentrated. The residue was diluted with

dichloromethane and treated with water. The aqueous phase was

extracted three times with dichloromethane. The combined organ-

ic phases were washed three times with HCl 1 N and three times

with saturated NaHCO₃, water, and brine. Purification by column

chromatography with silica gel (CH₂Cl₂/petroleum ether 5:95 to

In the course of this study we have been also able to con-

firm conclusions from NMR and IR previously presented by

Gellman and Stevens.^29,30Specifically, the (negative) tempera-

ture dependence of an amide NH chemical shift for a small

molecule in CDCl₃can furnish useful conformational informa-

tion when accompanied by IR data, but not on its own. The

results shown by us and others indicate that the NTDCS in

chloroform may be highly influenced by the internal structure

of the molecule and cannot necessarily be anticipated. Con-

versely, in DMSO the situation is greatly simplified, and there is

a marked difference between compounds that are intramole-

cularly hydrogen bonded with small NH NTDCS values

of >-3 ppb/K to those where the NH is not intramolecularly

hydrogen bonded and where the NH NTCDS values are

relatively larger and all <-4 ppb/K.

Lastly we wish to return to our original problem, which was

to know whether or not designed benzoylureas would adopt the

desired closed conformation. We now know that when R¹and

R²are suitably chosen, the molecules do indeed adopt the

closed conformation with an internal hydrogen bond as de-

sired, in both chloroform (hydrophobic environment) and

DMSO (polar environment). We are now in a position to

rationally functionalize benzoylureas to be R-helix mimetics

and will report on this in due course.⁸⁰

1

60:40) afforded 1 as a colorless film (127 mg, 71%): H NMR

(CDCl₃, ppm) 11.35 (br s, 1H), 7.59 (d, ³J=7.53 Hz, 2H), 7.55-

7.46 (m, 5H), 7.37 (t, ³J=7.47 Hz, 2H), 7.14 (t, ³J=7.38 Hz, 1H),

3.85 (q, ³J=6.99 Hz, 2H), 1.19 (t, ³J=6.96 Hz, 3H); ¹³C NMR

(CDCl₃, ppm) 175.8, 152.3, 138.1, 136.7, 130.8, 129.3, 129.1, 126.1,

124.5, 120.9, 42.6, 15.3; MS (ESI) 269.1 (M þ H^þ); HRMS (ESI)

C₁₆H₁₆N₂O₂calcd M^þ268.1212, found 268.1216.

Variable-Temperature NMR. All samples for variable-tem-

perature NMR in CDCl₃were dried in vacuo over P₂O₅. CDCl₃

was stored on activated sieves and K₂CO₃. NMR samples were

prepared in a glovebag with a two-dilution process to obtain a 3

mM solution. For each temperature, samples were equilibrated,

and acquisition required 64 scans. The reading chemical shift for

NH protons was obtained by applying line broadening (lb=1) to

the original spectrum. Spectra were acquired at 300, 305, 310,

315, and 320 K. A remaining peak of water was observed on

all spectra but fell within the range of previously published

studies.²⁹

All samples for variable-temperature NMR in DMSO-

d₆were dried in vacuo over P₂O₅. NMR samples were prepared

without particular precautions to obtain a 10 mM solution. For

each temperature, samples were equilibrated, and acquisition

required 16 scans. Spectra were acquired at 300, 305, 310, 315,

and 320 K.

Experimental Section

General Experimental Methods. Analytical thin-layer chro-

matography (TLC) was performed on silica gel 60F254 alumi-

num-backed plates. Flash chromatography was performed on

silica 60 using ACS analytical grade solvents. All reactions were

performed in flame-dried Schlenk flasks. Dry diethyl ether, dry

triethylamine, and dry acetonitrile were used as received.

(CAUTION! Phosgene solution must be handled with utmost

care in a well-vented fume-cupboard.) N-Isopropylacetamide

was prepared according to previously reported methods.

General Method of Formation of the Benzoylurea: 1-Benzoyl-

1-ethyl-3-phenylurea (1). N-Ethylbenzamide (100 mg, 0.67 mmol)

was suspended in 1.2 mL dry diethyl ether in an oven-dried Schlenk

flask. Triethylamine (103 μL, 0.74 mmol) was added followed by

trimethylsilyl trifluoromethanesulfonate (133 μL, 0.74 mmol). The

reaction was stirred at room temperature for 16 h. Stirring was

stopped to allow the orange oil formed during this first step to settle

at the bottom of the flask. The oil was removed via a syringe.

Phosgene solution (670 μL of a 20% solution in toluene,

CAUTION!) was added dropwise to the clear solution at 0 °C

with stirring (133 μL, 0.74 mmol). The reaction was then slowly

Graphs for all NMR variable-temperature experiments are

available in the Supporting Information section.

Infrared Spectroscopy. All samples for infrared spectroscopy

in CHCl₃were dried in vacuo over P₂O₅. CHCl₃was washed

with water, dried over K₂CO₃, stirred over CaCl₂, and distilled

˚

over fresh CaCl₂. It was then stored over 4 A sieves and used

quickly afterward. Samples were prepared in a glovebag with a

two-dilution process to obtain a 3 mM solution. Spectra were

obtained in absorbance mode with a spectral window 3000-

3600 cm^-1and 128 scans. Background subtraction was applied.

Acknowledgment. We thank the Leukemia and Lymphoma

Society (SCOR 7015-02) for funding the overall medicinal

chemistry program within which this study was conducted.

This work was supported by NHMRC IRISS Grant No.

361646 and a Victorian State Government OIS grant.

Supporting Information Available: Detailed procedures for

preparation of compounds 2a-c, 3, 4a-c, 8, 11a,b, and 12.

Results of NMR temperature experiments, CIF files for new

structures presented, and computational details for Table 5.

This material is available free of charge via the Internet at

http://pubs.acs.org.

(80) We have successfully applied the foundation work here to the

development of biologically active helical mimetics that bind to Bcl-x_Land

have disclosed this to researchers in a variety of formats (see ref 13).

J. Org. Chem. Vol. 74, No. 17, 2009 6525

Article Doi

Characterization of the two fundamental conformations of benzoylureas and elucidation of the factors that facilitate their conformational interchange

DOI: 10.1021/jo900871a

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jo900871a

Full text of DOI:10.1021/jo900871a

Products guided by the article

R&D Labs maybe for 1185291-03-5

Relevant to this article

Hot Product