Sequence-selective peptide recognition with designed modules

FULL PAPER

DOI: 10.1002/ejoc.200500393

Sequence-Selective Peptide Recognition with Designed Modules

Mark Wehner,^[a]Dominic Janssen,^[a]Gerhard Schäfer,^[a]and Thomas Schrader*^[a]

Keywords: Amino acids / Molecular recognition / Peptides / Rational design / Receptors / Sequence selectivity / Noncoval-

ent interactions

A concept for the rational design of sequence-selective pep-

tide receptors has been extended: in addition to recognition

modules for polar, aromatic and basic amino acids, the series

has now been completed with new receptor units for apolar

and acidic amino acids. The underlying strategy uses the

intermolecular β-sheet stabilization of a dipeptide as a pre-

requisite to bind its N-terminal amino acid side chain

through a strategically placed recognition tip at the end of a

U-turn protruding from the receptor moiety. Thus, a diami-

nopyrazole has been covalently attached to Kemp’s triacid by

way of a cyclic imide, while a meta-substituted aniline was

coupled as an amide to the pendant third carboxylate arm,

bringing the two aromatic units into a sub-van der Waals dis-

tance in a tight conformational lock. NMR titrations, Karplus

analyses and Monte-Carlo simulations demonstrate the ef-

fective sequence-selective recognition of alanine-containing

dipeptides. No example of such a rationally designed set of

peptide receptors had existed previously.

Germany, 2006)

neral concept introduced by the Still group starts from a

rigid scaffold bearing two parallel peptide strands, the se-

quence of which is varied statistically.^[3]From the resulting

combinatorial receptor library, the best binder for a given

small peptide is identified by an efficient high-throughput

screening method. Alternatively, the binding profile for a

given host can be established by varying the sequence of its

guest, which is often fluorescence-labelled. In recent years

the groups of Wennemers,^[4]Kilburn^[5]and Schmuck^[6]have

developed their own optimized systems used for, for exam-

ple, recognition of the Alzheimer’s peptide, the d-Ala-d-Ala

sequence and many more related small peptides. Very few

attempts to create sequence-selective peptide receptors by

rational design have been made. Schneider et al. (Figure 1)

recently reported that a zwitterionic peptide will be espe-

cially tightly bound if its N terminus is inserted into a pen-

dant crown ether while the C terminus is drawn into the

centre of the positive charges of a triply pyridinium-substi-

tuted porphyrin.^[7]π-Stacking interactions make these re-

ceptors selective for aromatic residues; with an additional

Zn²⁺central ion, histidine-containing peptides are bound

preferentially (Mizutani).^[8]An elegant extension is the con-

struction of mixed receptor lipid monolayers at the air/

water interface, introduced by Kunitake for distinction be-

tween dipeptides with very similar amino acid composi-

tions.^[9]

Introduction

Although considerable progress has been achieved in

designing artificial receptors for small biomolecules, includ-

ing most amino acids, the task of sequence-specific amino

acid recognition in peptides still poses a severe problem: an

efficient host for oligopeptides must be able to distinguish

the consecutive order of side chains. Human efforts to

mimic nature’s potent peptide hosts (mainly high molecular

weight proteins) with artificial structures are hampered by

the complexity of three-dimensional protein structures and

our lack of knowledge for prediction of protein folding. In

the early stages (and well beyond), success came through

serendipity: Rebek created dimeric host molecules based on

Kemp’s triacid, featuring a chiral cleft adorned with conver-

gent imide and amide groups, their mutual arrangement en-

forcing a steric constraint on sterically demanding dipep-

tides, which were only bound if the second amino acid had

a sterically undemanding side chain.^[1]Early work by Still

produced macrocyclic, tricyclic and even polycyclic receptor

structures, with unsymmetric arrangements of hydrogen

bond donors and acceptors with remarkable selectivities.

These artificial receptor molecules were developed for apo-

lar solvents, and bound to di- and tripeptides in an impres-

sively, albeit unpredictably, sequence- or stereoselective

fashion.^[2]

Subsequent approaches used a more efficient avenue

through the employment of combinatorial methods. The ge-

Prominent examples of the development of specific re-

ceptor units for small peptides are also found in nature:

vancomycin provides a binding pocket as an apolar envi-

ronment for very strong hydrogen bonds directed towards

the backbone of its target peptide. The typical d-Ala-d-Ala

sequence essential for most bacterial cell walls is thus ster-

[a] Philipps-Universität Marburg, Fachbereich Chemie,

Hans-Meerwein-Straße, 35032 Marburg, Germany

E-mail: schradet@staff.uni-marburg.de

Supporting information for this article is available on the

WWW under http://ww w .eurjoc.org or from the author.

138

Eur. J. Org. Chem. 2006, 138–153

Sequence-Selective Peptide Recognition with Designed Modules

FULL PAPER

ton,^[28]Bartlett^[29]), albeit in organic solvents. Aminopyra-

zoles, diaminoquinolones and azacyclohexenones (Figure 2)

are thus all able to form the maximum number of possible

hydrogen bonds to the top and bottom faces of an extended

peptide strand backbone.

Figure 1. Rational design of sequence-selective peptide receptors.

Left: Schneider’s porphyrin-based tripeptide receptor fills its crown

ether with Gly-Gly-Phe’s N-terminal ammonium group, while the

C-terminal Phe residue π-stacks with the porphyrin. Right: Mizut-

ani’s receptor uses a coordinative bond with the central Zn²⁺ion

for superior affinity towards histidine-containing peptides, rein-

forced by ion pairing with additional carboxyalkyloxy groups at

the periphery.

eoselectively recognized under physiological conditions.^[10]

The RGD–integrine interaction is responsible for a variety

of biological events governing the correct molecular func-

tion of proteins such as fibrinogen^[11](blood coagulation),

fibronectine^[12,13](cell matrix binding), osteopontine^[14,15]

(bone formation) or certain growth factors^[16](cell differen-

tiation and angiogenesis). All these proteins carry the criti-

Figure 2. Intermolecular β-sheet recognition by aminopyrazoles

and diaminoquinolones requires perfect three-dimensional comple-

mentarity in their D-A-D arrangements with exact alternating dis-

tances of 2.7 and 3.7 Å.

Starting from the efficient conformational lock repre-

cal RGD sequence on a solvent-exposed loop; mutagenesis sented by aminopyrazole β-sheet templates we recently put

studies showed that this is indeed the primary binding site forward a new concept for the rational design of sequence-

for interaction with their biological counterparts (e.g., cell selective peptide receptors, using the intermolecular β-sheet

surface-bound receptors such as the integrines).^[17–20]Many stabilization of a dipeptide as a prerequisite to bind one

proteases bind their peptidic substrates in a shallow groove, of its amino acid side chains through a strategically placed

where it is locked in an extended β-sheet-like conformation recognition tip at the end of a U-turn protruding from the

that produces a horizontal presentation of the guest’s amino receptor moiety.^[30]To the best of our knowledge, no exam-

acid side chains favourable for lateral recognition. β-Sheet ple of such a rationally designed set of peptide receptors

mimics make use of this principle: they replace the flexible had existed previously. The U-turn was constructed from

peptide backbone with heterocyclic recognition elements the very well known, highly preoriented skeleton of Kemp’s

possessing a much higher degree of rigidity, while simulta- triacid.^[31]Several elegant applications in the area of peptide

neously retaining the hydrogen bond donor and acceptor recognition and stereocontrolled photochemistry have used

pattern of a β-sheet.^[21]A logical extension affords the first this remarkable platform,^[32]which allows the insertion of

soluble β-sheet models, in which combinations of these arti- an amino-substituted heterocycle between two of its car-

ficial (Nowick^[22]) and natural (Kemp,^[23]Feigel,^[24]Kelly,^[25]boxylic acid functionalities, whilst still leaving a third bind-

Gellman^[26]) peptide strands are attached to common scaf- ing site for an interchangeable recognition tip (Figure 3).

folds and interact along their inner faces like zippers to This sets the stage for the synthetic procedure: condensa-

form two- or three-stranded parallel or antiparallel β- tion of the protected aminopyrazole unit with Kemp’s tri-

sheets. Very recently the first intermolecular versions of acid through the imide derivative must be followed by at-

these β-sheet models were created (Schrader,^[27]Hamil- tachment of the additional binding site, preferentially again

Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

139

M. Wehner, D. Janssen, G. Schäfer, T. Schrader

FULL PAPER

by amide formation. This sounds simple, and indeed offers sition, and the crude product is finally stirred with silica gel

a very short, straightforward route to a number of closely for a convenient and quantitative Boc deprotection. To our

related recognition modules, but it also poses the nontrivial great pleasure, the purified products are extremely soluble

problem of creating a new Kemp’s acid derivative with two in solvents of low polarity; despite their trifluoroacetylated

parallel aromatic moieties at sub-van der Waals distances aminopyrazole components, known to self-associate

(2.6 Å and less; vide infra). The amide coupling will thus strongly through hydrogen bonds, these Kemp’s acid ad-

pass through a sterically extremely demanding tetrahedral ducts are very soluble in chloroform.

intermediate stage. After extensive screening of various

Monte-Carlo simulations carried out for various poten-

coupling reagents we found that PyClop activation of the tial host–guest complexes in chloroform suggested: a) for-

corresponding monoacid could be carried out effectively at mation of an intramolecular hydrogen bond between the

elevated temperatures and provided moderate yields of the pyrazole NH group and one of the imide carbonyls, which

desired sterically congested coupling products. Only one would produce a favourable conformational lock for the

other report relating to the synthesis of doubly aromatic aminopyrazole ring, and b) an optimal tip orientation over

substituted derivatives of Kemp’s triacid has been pub- the horizontal amino acid side chain, if the aniline deriva-

lished, and this group also pointed out the synthetic diffi- tive is meta-substituted. Indeed, in the imide intermediate,

culties of the critical coupling step.^[33]

a distinct NMR signal was observed for the pyrazole NH

proton, shifted downfield to 12.5 ppm ([D₆]DMSO), indi-

cating the postulated intramolecular hydrogen bond. By the

synthetic approach outlined above, m-ethylaniline,^[37]m-

aminoacetophenone, benzyl m-aminobenzoate^[38]and even

m-aminonitrobenzene could be coupled to the imide inter-

mediate 2. The benzyl ester was finally cleaved by hydro-

genolysis and the resulting free acid was converted into the

tetrabutylammonium benzoate 8 – the binding site for ba-

sic, cationic amino acid residues. m-Aminoacetophenone

derivative 4 constitutes a host molecule for OH- and SH-

containing amino acids, m-aminonitrobenzene 5 should

pick electron-rich aromatic amino acids, and m-ethylaniline

3 serves as a reference compound with a comparable sub-

stituent size, but suitable only for recognition of the peptide

backbone.

As already stated above, 3–9 are among the few Kemp’s

triacid derivatives containing two closely stacked aromatic

rings. Modelling experiments predict a high degree of bond-

angle strain and repulsive van der Waals interactions; in

fact, the two aromatic planes are not completely parallel,

but diverge at a dihedral angle of ca. 15°. From the unsym-

metric geometries of the two aromatic groups, the existence

of two conformers must be assumed, and these should be

Figure 3. Design of the new peptide recognition modules with

amino acid specificity. XH = representative amino acid side chain

functionality; Y = recognition tip of each module.

Results and Discussion

In order to retain the DAD hydrogen bond donor and distinguishable through their spectroscopic properties. Only

1

acceptor pattern of the heterocycle we chose 3,5-diami- single sets of signals appear, however, in both the H and

nopyrazole as core unit. After its selective monotrifluo- the ¹³C NMR spectra of 3–9. In addition, although the two

roacetylation with TFAA (1; chemoselectivity follows the aromatic planes should be in close proximity, NOESY spec-

HSAB principle), it was condensed with the simple anhy- tra obtained for 3 find only one cross-peak between them,

dride of Kemp’s triacid, which is accessible from the parent as opposed to numerous contacts between alkyl substitu-

molecule by sublimation (Figure 4).^[34]If the condensation ents of the cyclohexyl skeleton. Close inspection of the cal-

procedure is carried out in toluene under dehydrating con- culated H,H distances in the productive conformation con-

ditions, the analytically pure product precipitates almost firms that only one distance is calculated to be less than

quantitatively on cooling to room temperature. The two 4.5 Å, and this is precisely the one observed in the NOESY

subsequent steps are carried out in one-pot fashion and re- experiment (3.2 Å). From Monte-Carlo simulations and

quire exact control of reaction conditions: ring protection molecular dynamics calculations^[39]we obtained indications

of the deactivated pyrazole nucleus takes place at 90 5 °C, of greater flexibility of these aromatic moieties within the

whilst above this temperature the labile Boc group is cleaved Kemp’s triacid scaffold than would be expected from the

off.^[35]Careful removal of excess Boc₂O is followed by pre- Lewis structures. The best explanation seems to be a con-

activation of the remaining carboxylic acid with PyClop/ certed rotation of the whole trans-amide bond, carrying the

DIEA and subsequent addition of the aniline derivative.^[36]aniline derivative further away from the aminopyrazole and

Only on prolonged heating at 50 5 °C in chloroform also allowing the ring-flip necessary for a fast interconver-

(Ͼ12 h) does amide formation take place without decompo- sion of the two conformers on the NMR timescale.

140

www.eurjoc.org

Eur. J. Org. Chem. 2006, 138–153

Sequence-Selective Peptide Recognition with Designed Modules

FULL PAPER

Figure 4. Synthetic route to all new Kemp’s acid-based host compounds. a) TFAA. b) Sublimation. c) Toluene, 110 °C (–H₂O), Et₃N.

d) Boc₂O, 90 °C. e) m-Aminobenzene derivative, PyCloP, 50 °C, CHCl₃. f) Silica gel, (–Boc). a–f) 17–22% overall yield. R = m-C₂H₅(3),

–

+

m-(C=O)CH₃(4), m-NO₂(5), m-CO₂(6), p-CH₂–C₆H₅(7), m-CH₂–C₆H₅(8), p-CH₂–NH₃(9).

The first important question to answer is that concerning mined by Job plots.^[41]Preliminary investigations of the ste-

potential self-association of the new host compounds, a ric tolerance in the guest peptide structure were informative,

critical point with the simple monomeric and dimeric linear since it became clear that neither a bulky N-terminal acyl

acylated aminopyrazoles. In the concentration range be- group (such as Boc) nor two sterically demanding amino

tween 10^–2m and 10^–3m later used in the binding experi- acids (Ac-Val-Val-OMe) were tolerated in the complex. It

ments, no chemically induced shift of any receptor proton must be assumed that the sterically congested chemical en-

could be observed. Obviously, the sterically demanding vironment around the aminopyrazole renders the Kemp’s

framework of the Kemp’s triacid effectively prevents the ag- acid-based peptide hosts selective for relatively slim guests.

gregation of host molecules with one another.

This view is also supported by force-field calculations. We

The binding profiles of the different receptor modules thus replaced the C-terminal valine with alanine, and ob-

next had to be examined, the challenge being to create mod- served the restoration of the expected downfield shifts of

ules for each category among the five major classes of the top face NH (Δδ_max= 0.1–0.8 ppm). Accordingly, we

amino acids: polar, aromatic, basic, acidic and nonpolar. can now write the general structure of the preferred guests

Binding experiments were generally carried out through as Ac-NH-AA1-Ala-OMe. Such peptides present only one

NMR titrations,^[40]whereas stoichiometries were deter- side chain (the one at the N terminus) to the receptor. Since

Table 1. Overview of all association constants K_a[M^–1] for complex formations between hosts 3–9 and representative dipeptides, deter-

mined by NMR titrations in CDCl₃. All dipeptides were protected as N-acetyl methyl esters. Stoichiometries were always 1:1, except

when stated otherwise.

Host

Standard

Selective binding

Special

TriflAMP^[a]

Reference 3

Ala-Ala: 50^[b]

Orn-Ala: 280

Ala-Ala: 80

–

propylamine: 490

–

Phe-Ala: 40

Ala-Ala: 70

Ser-Ala: 900

Phe-Ala: 350

Orn-Ala: 2360

Val-Ala: 110

Val-Ala: 70

Glu-Ala: 1300

Phe-Ala: Յ 40

Ser-Val: no shifts

Ala-Phe: no shifts

Gly-Lys: 130

Ala-Val: 130

Leu-Val: no shifts

–

Val-Ala: 50

Val-Val: no shifts

–

Leu-Val: no shifts

–

XH-binder 4

Arene-binder 5

Cation-binder 6

Alkyl-binder 7

Alkyl-binder 8

Anion-binder 9

acetate: Ͻ 10

[a] TriflAMP = 3-trifluoroacetylamino-5-methylpyrazole. [b] Errors were determined as standard deviations from the nonlinear re-

gressions; they usually varied between 10% and 40% (exceptions: very weak binding).

Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

141

M. Wehner, D. Janssen, G. Schäfer, T. Schrader

FULL PAPER

alanine has a much higher propensity to adopt an α-helical matic in a parallel orientation at the ideal van der Waals

conformation than to form a β-sheet,^[42]the intended stabi- distance of 3.4 Å.

lization of the extended conformation in the guest peptide

had to be expected to become more difficult – lower K_a

values being the consequence. After a perfect 1:1 stoichiom-

etry had been established by means of a Job plot, nonlinear

regression analysis of the binding isotherms between refer-

ence compound 3 and alanine-containing apolar peptides

in CDCl₃confirmed this assumption: K_avalues remained

below 100 m^–1(Table 1). Moreover, identical values of

around 50 m^–1were obtained in titrations of these peptides

with the best monomeric aminopyrazole peptide binders

previously developed by our group.

Basic Amino Acids

A biomimetic approach towards the recognition of ionic

amino acid residues involves the coulombic attraction be-

tween those in the guest and oppositely charged residues in

the host compound. Basic amino acids are thus often

bound by aspartate or glutamate anions. A short basic

amino acid, ornithine, was chosen because molecular me-

chanics calculations in this case suggested a perfect align-

ment of the two ionic groups. In summary, a marked prefer-

ence of 9:1 was found for the selective recognition of Ac-

Orn-Ala-OMe by the corresponding aminopyrazole recep-

tor unit 6 (K_a= 2400 m^–1) over the closely related reference

Polar Amino Acids

A host molecule to serve this purpose must form strong host 3 (280 m^–1). A single ion pair-reinforced hydrogen

hydrogen bonds to the OH or SH groups in serine, threo- bond between the host module’s benzoate tip and the

nine and cysteine. As already described in our first report, guest’s ammonium cation was thus designed and was exper-

Ac-Ser-Ala-OMe furnished a high association constant of imentally found to increase the affinity towards basic pep-

ca. 900 m^–1when it was titrated with 4, as opposed to con- tide by roughly one order of magnitude.

trol experiments with the dialanine derivative or the refer-

In order to examine the sequence selectivity of this ionic

recognition process, we carried out a synthesis of the in-

verted sequence Ac-Ala-Orn-OMe, but failed in the last

step, which involved hydrogenolytic cleavage of the δ-Z

group: even under acidic conditions we could not avoid in-

tramolecular cyclization of ornithine’s free amino group

onto its C-terminal methyl ester to furnish the stable pre-

ferred piperidone. As an alternative we prepared the related

lysine derivative Ac-Gly-Lys-OMe. Its tosylate did not cy-

clize and could be titrated with host 6 to give a low associa-

tion constant of 130 m^–1. In spite of the small structural

differences this result clearly again indicates a strong prefer-

ence of host 6 for peptides with N-terminal basic residues.

ence compound 3 (K_a= 80 m^–1, Table 1).

Aromatic Amino Acids

π-Stacking interactions are often used in protein recogni-

tion events, especially within extended hydrophobic do-

mains. For the most efficient attraction of Phe, Tyr or Trp

residues a m-aminonitroarene was introduced in the host

molecule. Comparison of their associated EPS colour codes

demonstrates that the electron-rich aromatics among the

proteinogenic amino acids should all form favourable com-

plexes with the electron-deficient nitroarene.^[43]The nitro-

aniline-functionalized host 5 (Figure 5) showes an affinity

towards Ac-Phe-Ala-OMe (K_1:1= 350 m^–1) nine times

higher than toward the reference host 3 (K_aϽ 40 m^–1). In a

molecular modelling study of the complex the nitroarene is

Nonpolar Amino Acids

To achieve efficient recognition of nonpolar amino acids

reproducibly flipped out to receive the phenylalanine aro- is a difficult task, especially in apolar chloroform, in which

Figure 5. Phenylalanine recognition within Ac-Phe-Ala-OMe by nitroarene-modified aminopyrazole host 5. Left and middle: front view.

Right: side view. A: Kemp’s triacid with arene tip. B: aminopyrazole. C: aromatic guest dipeptide. Note the perfect π-stacking between

the nitroarene and the electron-rich Phe benzene ring.

142

www.eurjoc.org

Eur. J. Org. Chem. 2006, 138–153

Sequence-Selective Peptide Recognition with Designed Modules

FULL PAPER

hydrophobic interactions are intrinsically weak. Further- the coupling constants dropped markedly, suggesting that

more, most aliphatic amino acids carry bulky side chains, the two amino acids avoid a perfect β-sheet conformation

so a compromise must be found between steric repulsion to optimize their hydrogen bonding to the peptide back-

and maximum van der Waals contact. Since protein–protein bone. The effect is most pronounced in the case of the meta

3

complexes often use phenylalanine units to provide large derivative 8 [Val: J(H,H) from 8.8 to 7.8 Hz; Ala: J(H,H)

nonpolar π-faces, we chose the diphenylmethane unit as from 7.1 to 6.2 Hz]. It lowers the free energy gain on com-

such a compromise: this offers a large π-face, and simulta- plex formation and might thus explain the lower affinity of

neously, because of its CH₂hinge group, forms a shallow 8 towards aliphatic dipeptides. As a further improvement of

groove for maximum van der Waals contact. Modelling ex- this concept, we intend in the future to extend the di-

periments suggest that this binding site is easily accessible phenylmethane unit into a dihydromethanoanthracene moi-

for secondary alkyl groups, so that most apolar proteinog- ety, which could form a much more extended cleft for in-

enic amino acids should be able to insert their alkyl side creased hydrophobic contacts (Figure 6).

chains into the shallow groove of the corresponding recep-

tor moiety. According to the optimized procedure described

above, both the m- and the p-benzylaniline derivatives of

Kemp’s acid with a pendant aminopyrazole imide could be

prepared in a straightforward synthesis. Both new hosts

turned out still to be very soluble in chloroform. Binding

Acidic Amino Acids

The simplest case of such a receptor based on our

experiments with ethyl reference host 3 and the new di- Kemp’s triacid aminopyrazole module would be the re-

phenylmethane-modified 7 and 8 demonstrate the validity cognition of an aspartate or glutamate side chain by a per-

of both the assumptions detailed above: leucine-containing manently charged cationic binding site attached to the m-

dipeptides are rejected by both 3 and 7 (para derivative), aniline arm. Force field calculations and synthetic consider-

due to steric restrictions. However, valine-containing dipep- ations suggested aminomethylaniline as a good candidate

tides again show pronounced downfield shifts in their NH for glutamate recognition (Figure 7). Its synthesis was

protons, demonstrating that backbone recognition takes achieved through Z-protection of the more basic benzyl-

place. This time, comparison of the affinity towards Ac- amine, amide coupling with PyClop at elevated temperatures

Val-Ala-OMe with the reference host 3 (55 m^–1) reveals a and Boc deprotection on the column. Final removal of the

moderate, but distinct twofold increase in K_afor the p-di- Z group was effected by hydrogenolysis and subsequent

phenylmethane-modified host 7 (105 m^–1). The meta-substi- treatment with 1 equiv. of HCl. Binding experiments with

tuted host 8 binds the guest peptide somewhat less tightly Ac-Glu-Ala-OMe tetrabutylammonium salt resulted in

(70 m^–1). Although steric restrictions are far more impor- chemical shift changes exclusively at the top-face NH pro-

tant in these cases, surprisingly, no sequence selectivity is ton, whereas the bottom face NH remained untouched, in-

found this time: the inverted peptide sequence Ac-Ala-Val- dicating the desired 1:1 complex formation with concomi-

OMe binds equally well to 7 (134 m^–1). We attribute this tant backbone recognition. That strong ion pairs were

divergent behaviour to the facile rotation of the flexible di- formed simultaneously is well documented by the elevated

phenylmethane linker around its benzylic bond. Subsequent binding constant of 1300 m^–1. A final titration with tetrabu-

Karplus analysis also produced interesting results: from the tylammonium acetate did not furnish any appreciable

preoriented dipeptide (Ac-Val-Ala-OMe) to the complex chemical shift change.

Figure 6. Valine recognition within Ac-Val-Ala-OMe by diphenylmethane-modified aminopyrazole host 7. Left and middle: front view.

Right: side view. A: Kemp’s triacid with arene tip. B: aminopyrazole. C: aliphatic guest dipeptide. Note the close fit between valine’s

isopropyl group and the diphenylmethane “roof”.

Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

143

M. Wehner, D. Janssen, G. Schäfer, T. Schrader

FULL PAPER

Figure 7. Glutamate recognition by benzylammonium-modified aminopyrazole host 9.

Karplus Analysis

Direct information relating to the peptide conformation can

be drawn from Karplus analyses of the NH-a-CH coupling

constants, which correlate with the characteristic torsion

angle θ (Figure 8).^[44]For convenient comparison, we exam-

ined complexes of three hosts carrying additional binding

sites, each with their best peptide binding partners. If hosts

3–8 were added to the peptide solution, signals were con-

stantly sharpened and in most cases ³J values increased rel-

ative to the free peptides. Both the peptide’s top and bottom

side showed this effect. However, the experimentally ob-

Figure 8. Top: Karplus equation, modified for peptides. Bottom:

assignment of NH¹and NH²and corresponding torsion angles θ

and ψ in an N/C-protected dipeptide.

tained values did not reach the calculated coupling con-

stants derived from the calculated complex geometries

(MacroModel 7.0), see Table 2. This makes sense, because

at the relatively low concentrations used for the NMR ti-

tration experiments, only a certain fraction of all molecules

was fixed in the complex.

Conclusion

The host modules 7 and 9 for nonpolar and acidic amino

acids constitute exceptions to this rule: most probably for

In the experiments described above we have shown that

steric reasons, the peptide backbones avoided the adoption peptide receptors selective for some of the main classes of

of perfect β-sheet conformations in order to maximize hy- amino acids in a peptidic environment can be developed by

drogen bond interactions with the receptors, and thus pro- rational design. Further work towards improved efficiency

3

duced lower J values in the complexes as opposed to the and selectivity of all receptor modules is in progress in our

free peptides.

laboratory.

Table 2. Experimentally measured and calculated ³J coupling constants [Hz] of free dipeptides and their complexes with the corresponding

best host molecules discussed above (MacroModel 7.2, Amber*, CHCl₃, 1000 steps).

Host

Dipeptide

Experiment

free

Calculation

complex

H²

solution

H¹

H²

H¹

H²

4

5

6

7

9

Ac-Ser-Ala-OMe

6.3

7.3

5.3

8.8

7.0

6.3

7.6

6.0

7.1

6.8

6.6

7.0

8.6

6.4

7.6

8.3

6.4

6.7

5.0

8.7

5.4

6.5

8.0

9.0

9.7

6.6

6.7

5.0

6.7

Ac-Phe-Ala-OMe

Ac-Orn-Ala-OMe^.HOAc

Ac-Val-Ala-OMe

Ac-Glu-Ala-OMe^.LiOH

144

www.eurjoc.org

Eur. J. Org. Chem. 2006, 138–153

Sequence-Selective Peptide Recognition with Designed Modules

FULL PAPER

in a Dean–Stark trap, until no water was produced any longer. The

solvent was removed by distillation, until precipitation started at

the boiling point. After the system had been cooled to –24 °C the

white precipitate was filtered off and dried in vacuo (R_F= 0.45 in

CH₂Cl₂/methanol, 5:1). Yield 1.88 g (4.5 mmol, 94 %); m. p.

If the dipeptides carry their hydroxy, aromatic or cationic

amino acid residues at the C terminus, binding constants

drop drastically. Thus, the serine- or phenylalanine-contain-

ing dipeptides 11 and 14 are not complexed at all. Even a

highly flexible lysine residue at the C terminus is bound

with a K_avalue of only 130 m^–1as opposed to N-terminal

ornithine with 2400 m^–1. These examples show that a high

simple sequence selectivity is achieved with our new recep-

tor modules, which we attribute mainly to steric reasons. In

1

221 °C. H NMR (500 MHz, [D₆]DMSO, 25 °C): δ = 1.25 (s, 6 H,

1), 1.26 (s, 3 H, 2), 1.43 (d, ²J(H,H) = 13.9 Hz, 2 H, 3), 1.66 (d,

2

²J(H,H) = 12.9 Hz, 1 H, 4), 2.10 (d, J(H,H) = 12.9 Hz, 1 H, 5),

2

2.53 (d, J(H,H) = 13.9 Hz, 2 H, 6), 6.55 (br., 1 H, 7), 12.02 (br.,

1 H, 8), 12.48 (br., 1 H, 9) ppm. ¹³C NMR (126 MHz, [D₆]DMSO,

some cases, complex formation is so sensitive towards steric 25 °C): δ = 25.1 (a), 30.2 (b), 39.9 (c), 40.9 (d), 41.7 (e), 42.9 (f),

2

59.7 (g), 89.2 (h), 96,5 (i), 115.7 [q, J(C, F) = 287.0 Hz; j], 154.0

[q, ³J(C, F) = 36.9 Hz]; (k), 175.4 (l), 176.9 (m) ppm. MS (CI, NH₃,

200 °C): m/z = 417 [M + H]⁺. HRMS (EI): m/z, found 416.1304;

calcd. 416.1308. C₁₇H₁₉F₃N₄O₅: found C 49.04, H 4.60, N 13.46;

calcd. C 49.56, H 4.38, N 12.94.

factors that the same receptor module completely rejects

the same dipeptide with an inverted sequence. Modelling

studies indicate that only one mutual orientation of both

binding partners is productive, whilst the other one seems

to be sterically hindered.

General Procedure for the Amidation of 2 with Aniline Derivatives:

A suspension of 2 (250 mg, 0.6 mmol) and Boc₂O (10 mL,

46.7 mmol) was heated to 90 °C for 30 min, after which the starting

material had disappeared in the TLC. The excess of Boc₂O was

condensed off at a water bath temperature of Ͼ70 °C. The remain-

ing solid was dissolved in chloroform (40 mL) and treated with

diisopropylethylamine (0.84 mL, 4.8 mmol) and PyCloP (0.89 g,

2.1 mmol). Stirring was continued overnight. The solution was

warmed to 50 °C and a suspension of chloroform (25 mL) and the

corresponding aniline derivative (2.5 equiv., 1.5 mmol) was added

dropwise over 3 h. After the addition heating was continued for

another 7 hr, and the solution was then stirred at room temperature

overnight and at 50 °C for another 6 h. The solution was concen-

trated to dryness and the remaining yellow oil was suspended in a

little ethyl acetate/hexane (1:2). It was then placed on a chromatog-

raphy column filled with silica gel 60. After standing overnight, elu-

tion was carried out until no substance was eluted any longer; the

product was then eluted with ethyl acetate. The crude product was

chromatographed a second time over silica gel 60 with ethyl acetate/

hexane (2:1).

Outlook

In the future we intend to connect two or more peptide

receptor modules covalently for the five major classes of

amino acids and to achieve sequence-selective recognition

of larger peptides, hopefully also in more polar solutions.

Dimeric aminopyrazoles and aminopyrazole/peptide hybrid

receptor molecules have recently been shown to bind a

hexapeptidic key sequence taken from Aβ (1–40) – in water!

A detailed spectroscopic investigation indicates that an in-

creased number of hydrogen bonds plus hydrophobic stack-

ing interactions are responsible for this possible drastic sol-

vent change. Recognition of larger peptide sequences in

peptides and proteins might then become feasible with con-

structs of several peptide mimics attached to a central scaf-

fold, as introduced by Nowick’s group. We thus propose

a tridentate fork, made up of a central peptide chain and

peripheral aminopyrazole oligomers, as an inhibitor of

HIV-protease dimerization for a potential application in

AIDS therapy.

Experimental Section

Receptor Molecules

N-(3-Ethylphenyl)-1,5,7-trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoroace-

tylamino)-2H-pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carboxam-

ide (3): The preparation followed the general procedure outlined

above. R_F= 0.64 (ethyl acetate): Yield 0.1 g (0.1 mmol, 23%); m. p.

152 °C. ¹H NMR (300 MHz, [D₄]methanol, 28 °C): δ = 1.37 (t,

³J(H,H) = 7.6 Hz, 3 H, 1), 1.56 (s, 6 H, 2), 1.58 (s, 3 H, 3), 1.68

2

(d, J(H,H) = 14.3 Hz, 2 H, 4), 1.83 (d, J(H,H) = 12.9 Hz, 1 H,

2

3

5), 2.40 (d, J(H,H) = 12.9 Hz, 1 H, 6), 2.78 (q, J(H,H) = 7.6 Hz,

1,5,7-Trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoroacetylamino)-2H-pyr- 2 H, 7), 3.07 (d, ²J(H,H) = 14.3 Hz, 2 H, 8), 6.19 (br., 1 H, 9), 7.15

azol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carboxylic Acid (2): A sus-

pension of 1,3,5-trimethyl-1,3,5-cyclohexanetricarboxylic acid an-

hydride (1) (114 g, 4.8 mmol), (0.94 g, 4.8 mmol) and triethylamine

(0.1 mL, 0.7 mmol) were heated at reflux in dry toluene (150 mL)

(t, ³J(H,H) = 4.0 Hz, 1 H, 10), 7.32 (s, 1 H, 11), 7.37 (m, 2 H,

12) ppm. ¹³C NMR (50.4 MHz, [D₄]MeOH, 25 °C): δ = 15.9 (a),

26.0 (b), 29.2 (c), 32.2 (d), 41.4 (e), 43.5 (f), 45.6 (g), 45.7 (h), 110.4

(i), 122.7 (j), 126.4 (k), 129.5 (l), 136.3 (m), 146.1 (n), 176.0 (o),

Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

145

M. Wehner, D. Janssen, G. Schäfer, T. Schrader

FULL PAPER

178.3 (p) ppm. MS (FD): m/z = 519 [M]⁺(100%). C₂₅H₂₈F₃N₅O₄:

calcd. C 57.80, H 5.43, N 13.48; found C 57.87, H 5.82, N 13.75.

Benzyl 3-({1,5,7-Trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoroacetylami-

no)-2H-pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carbonyl}amino)-

benzoate (6a): The preparation followed the general procedure out-

lined above. R_F= 0.37 (ethyl acetate/hexane, 2:1); Yield 26 mg

N-(3-Acetylphenyl)-1,5,7-trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoro-

acetylamino)-2H-pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carbox-

amide (4): The preparation followed the general procedure outlined

above. R_F= 0.49 (ethyl acetate); Yield 54 mg (0.1 mmol, 18%);

1

(0.04 mmol, 7%); m. p. 159 °C. H NMR (300 MHz, [D₆]acetone,

28 °C): δ = 1.24 (s, 6 H, 1), 1.28 (s, 3 H, 2), 1.46 (d, ²J(H,H) =

14.2 Hz, 2 H, 3), 1.59 (d, ²J(H,H) = 13.3 Hz, 1 H, 4), 2.18 (d,

1

m. p. 185 °C. H NMR (500 MHz, [D₄]MeOH, 25 °C): δ = 1.32 (s,

2

²J(H,H) = 13.3 Hz, 1 H, 5), 2.82 (d, J(H,H) = 14.2 Hz, 2 H, 6),

2

6 H, 1), 1.35 (s, 3 H, 2), 1.49 (d, J(H,H) = 14.2 Hz, 2 H, 3), 1.64

3

6.29 (s, 1 H, 7), 7.33 (m, 6 H, 8), 7.69 (d, J(H,H) = 7.8 Hz, 1 H,

2

(d, J(H,H) = 13.6 Hz, 1 H, 4), 2.14 (d, J(H,H) = 13.6 Hz, 1 H,

3

9), 7.80 (d, J(H,H) = 8.3 Hz, 1 H, 10), 7.97 (s, 1 H, 11) ppm. ¹³C

2

5), 2.52 (s, 3 H, 6), 2.83 (d, J(H,H) = 14.2 Hz, 2 H, 7), 5.93 (br.,

NMR (50.4 MHz, [D₆]acetone, 25 °C): δ = 25.9 (a), 31.2 (b), 41.4

(c), 43.1 (d), 43.8 (e), 44.7 (f), 67.8 (g), 122.9 (h), 126.0 (i), 126.8

(j), 128.8 (k), 129.3 (l), 129.8 (m), 131.5 (n), 137.4 (o), 139.4 (p),

166.4 (q), 175.6 (r), 176.4 (s) ppm. MS (FD): m/z = 625 [M]⁺

(100%). HRMS (EI): m/z, found 625.2156; calcd. 625.2148.

1 H, 8), 7.34 (t, ³J(H,H) = 8.2 Hz, 1 H, 9), 7.51 (d, ³J(H,H) =

3

7.6 Hz, 1 H, 10), 7.67 (d, J(H,H) = 7.6 Hz, 1 H, 11), 7.99 (s, 1 H,

12) ppm. ¹³C NMR (50.4 MHz, [D₄]MeOH, 25 °C); δ = 25.9 (a),

26.8 (b), 31.4 (c), 42.0 (d), 43.7 (e), 44.2 (f), 44.9 (g), 123.7 (h),

125.8 (i), 128.5 (j), 130.0 (k), 138.8 (l), 139.6 (m), 176.5 (n), 178.3

(o), 200.6 (p) ppm. MS (FD) m/z = 533 [M]⁺(100%). HRMS (EI)

m/z, found 533.1876; calcd. 533.1886.

3-({1,5,7-Trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoroacetylamino)-2H-

pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carbonyl}amino)benzoic

Acid (6b): Compound 6a (50 mg, 0.08 mmol) was dissolved in

methanol (50 mL) and treated with Pd/C (55 mg). The flask was

connected to a hydrogen-filled balloon and the mixture was stirred

for 6 h. The reaction mixture was filtered through silica gel and the

filtrate was concentrated to dryness. The remaining solid was

washed twice with ethyl acetate and twice with dichloromethane.

1,5,7-Trimethyl-N-(3-nitrophenyl)-2,4-dioxo-3-[5-(2,2,2-trifluoroace-

tylamino)-2H-pyrazol-3-yl]-3-azabicyclo[3.3.1]nonan-7-carboxamide

(5): The preparation followed the general procedure outlined above;

R_F= 0.58 (ethyl acetate); Yield 85 mg (0.1 mmol, 18 %); m. p.

1

208 °C. H NMR (300 MHz, [D₄]MeOH, 28 °C): δ = 1.31 (s, 6 H,

1

Yield 16 mg (0.03 mmol, 72%); m. p. 230 °C. H NMR (300 MHz,

1), 1.34 (s, 3 H, 2), 1.47 (d, ²J(H,H) = 14.6 Hz, 2 H, 3), 1.61 (d,

[D₄]MeOH, 28 °C): δ = 1.31 (s, 6 H, 1), 1.34 (s, 3 H, 2), 1.44 (d,

2

²J(H,H) = 13.2 Hz, 1 H, 4), 2.14 (d, J(H,H) = 13.2 Hz, 1 H, 5),

2

²J(H,H) = 14.3 Hz, 2 H, 3), 1.59 (d, J(H,H) = 12.9 Hz, 1 H, 4),

2.84 (d, ²J(H,H) = 14.6 Hz, 2 H, 6), 5.93 (br., 1 H, 7), 7.42 (t,

2

2.15 (d, J(H,H) = 12.9 Hz, 1 H, 5), 2.84 (d, J(H,H) = 14.3 Hz, 2

3

³J(H,H) = 8.3 Hz, 1 H, 8), 7.66 (d, J(H,H) = 9.0 Hz, 1 H, 9), 7.88

3

H, 6), 5.93 (br., 1 H, 7), 7.28 (t, J(H,H) = 8.3 Hz, 1 H, 8), 7.57

3

(d, J(H,H) = 9.3 Hz, 1 H, 10), 8.37 (s, 1 H, 11) ppm. ¹³C NMR

(d, ³J(H,H) = 8.0 Hz, 1 H, 9), 7.68 (d, J(H,H) = 7.6 Hz, 1 H, 10),

3

(50.4 MHz, [D₄]MeOH, 25 °C): δ = 26.0 (a), 31.3 (b), 42.0 (c), 43.8

(d), 44.3 (e), 44.8 (f), 117.9 (g), 120.1 (h), 128.9 (i), 130.5 (j), 140.6

(k), 149.8 (l), 176.3 (m), 177.7 (n) ppm. MS (FD): m/z = 536 [M]⁺

(100%). HRMS (EI): m/z, found 536.1633; calcd. 533.1631.

7.87 (s, 1 H, 11) ppm. ¹³C NMR (50.4 MHz, [D₆]acetone, 25 °C):

δ = 26.0 (a), 31.5 (b), 41.9 (c), 44.1 (d), 45.0 (e), 124.5 (f), 126.8

(g), 126.9 (h), 129.5 (i), 129.5 (j), 139.2 (k), 173.9 (l), 176.3 (m),

178.1 (n) ppm. MS (ESI negative): m/z = 534 [M]⁺.

146

www.eurjoc.org

Eur. J. Org. Chem. 2006, 138–153

Sequence-Selective Peptide Recognition with Designed Modules

FULL PAPER

Tetrabutylammonium 3-({1,5,7-Trimethyl-2,4-dioxo-3-[5-(2,2,2-tri-

fluoroacetylamino)-2H-pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-

carbonyl}amino)benzoate (6c): Compound 6b (20 mg, 0.037 mmol)

was dissolved in methanol (10 mL) and treated with NBu₄OH (1 m,

37 μL), and the solution was concentrated to dryness. Yield

28.7 mg (0.037 mmol, 100%); m. p. Ͼ260 °C. ¹H NMR (300 MHz,

N-(3-Benzylphenyl)-1,5,7-trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoro-

acetylamino)-2H-pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carbox-

amide (8): The preparation followed the general procedure outlined

above except for the use of a less polar column chromatography

gradient, first with ethyl acetate/hexane, 1:2 and subsequently with

ethyl acetate/hexane, 3:2. Yield 16 mg (0.03 mmol, 13%); m. p.

3

[D₄]MeOH, 28 °C): δ = 0.99 (t, J(H,H) = 7.3 Hz, 12 H, 1), 1.30

(s, 6 H, 2), 1.35 (s, 3 H, 3), 1.38 (d, ²J(H,H) = 14.1 Hz, 2 H, 4),

1

2

224 °C. H NMR (300 MHz, CDCl₃): δ = 1.34 (s, 6 H, H-16, H-

1.38 (br., 8 H, 5), 1.59 (m, 9 H, 6), 2.16 (d, J(H,H) = 13.3 Hz, 1

2

17), 1.34 (s, 3 H, H-18), 1.46 (d, J = 14.6 Hz, 2 H, H-13, H-15),

H, 7), 2.83 (d, J(H,H) = 14.1 Hz, 2 H, 8), 3.18 (m, 8 H, 9), 5.87

2

3

(br., 1 H, 10), 7.25 (t, J(H,H) = 7.6 Hz, 1 H, 11), 7.64 (d, ³J(H,H)

1.63 (d, J = 12.4 Hz, 1 H, H-11), 2.18 (d, J = 12.4 Hz, 1 H, H-

2

3

11), 2.84 (d, J = 14.6 Hz, 2 H, H-13, H-15), 3.89 (s, 2 H, H-27),

= 7.6 Hz, 1 H, 12), 7.67 (d, J(H,H) = 10.3 Hz, 1 H, 10), 7.70 (s, 1

6.89 (b, 1 H, H-7), 7.16 (m, 9 H, H-aryl) ppm. ¹³C NMR

(400 MHz, CDCl₃): δ = 25.3 (C-16, C-17), 27.6 (C-18), 40.6 (C-10.

C-12), 41.4 (C-14), 42.8 (C-11), 44.2 (C-27), 44.8 (C-13, C-15), 95.7

(C-7), 120.2 (C-23), 120.5 (C-22), 123.1 (C-1), 125.9 (C-25), 126.0

(C-31), 128.2 (C-Ar), 128.7 (C-Ar), 128.9 (C-Ar), 137.7 (C-21),

142.1 (C-26, C-28), 174.5 (C-19), 175.3 (C-8, C-9) ppm. MS (ESI

positive): m/z = 604 [M + Na]⁺. HRMS for C₃₀H₃₀F₃N₅O₄Na⁺[M

+ Na]⁺(ESI positive): m/z, found 604.2148; calcd. 604.2155.

H, 11) ppm. ¹³C NMR (50.4 MHz, [D₄]MeOH, 25 °C): δ = 14.0

(a), 20.7 (b), 24.8 (c), 26.2 (d), 28.6 (e), 31.6 (f), 42.0 (g), 44.2 (h),

45.2 (i), 59.6 (j), 97.4 (k), 124.0 (l), 125.5 (m), 126.6 (n), 129.2 (o),

139.0 (p), 139.2 (q), 176.2 (r), 178.1 (s) ppm. MS (ESI negative):

m/z = 776 [M]⁺(18%), 534 [M – H]⁺, Carboxylic acid anion with-

out NBu₄]⁺(100%). HRMS for [M – NBu₄]⁺(ESI negative): m/z,

found 534.1595; calcd. 534.1600.

N-(4-Benzylphenyl)-1,5,7-trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoro-

acetylamino)-2H-pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carbox-

amide (7): The preparation followed the general procedure outlined

above except for the use of a less polar column chromatography

gradient, with ethyl acetate/hexane, 1:2, first and subsequently with

Benzyl [4-({1,5,7-Trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoroacetylami-

ethyl acetate/hexane, 3:2. Yield 19 mg (0.03 mmol, 13%); m. p. no)-2H-pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carbonyl}amino)-

1

207 °C; H NMR (300 MHz, CDCl₃): δ = 1.34 (s, 6 H, H-16, H-

benzyl]carbamate (9a): The preparation followed the general pro-

cedure outlined above except for the use of a different column

chromatography gradient, finally using dichloromethane/methanol,

40:1. Yield 9 mg (0.02 mmol, 10 %); m. p. 207 °C; ¹H NMR

(200 MHz, [D₄]MeOH): δ = 1.30–1.50 (m, 11 H), 2.23 (d, ²J =

2

17), 1.35 (s, 3 H, H-18), 1.50 (d, J = 14.4 Hz, 2 H, H-13, H-15),

2

1.65 (d, J = 13.4 Hz, 1 H, H-11), 2.22 (d, J = 13.4 Hz, 1 H, H-

2

11), 2.88 (d, J = 14.4 Hz, 2 H, H-13, H-15), 3.94 (s, 2 H, H-25),

7.12 (d, ³J = 8.3 Hz, 2 H, H-23), 7.21 (m, 3 H, H-27, H-28, H-29),

7.25 (m, 2 H, H-27, H-28, H-29), 7.27 (d, ³J = 8.3 Hz, 2 H, H-

22) ppm. ¹³C NMR (400 MHz, CDCl₃): δ = 22.8 (C-18), 25.6 (C-

16, C-17), 29.5 (C-14), 29.9 (C-10, C-12), 40.9 (C-13, C-15), 41.3

2

13.6 Hz, 1 H, H-11), 2.66 (d, J = 13.6 Hz, 1 H, H-11), 2.86 (d, J

= 13.9 Hz, 2 H, H-13, H-15), 4.23 (s, 2 H, H-25), 5.07 (s, 2 H, H-

3

28), 6.34 (s, 1 H, H-7), 6.88 (d, J = 8.3 Hz, 2 H, H-23), 7.14 (d,

(C-11), 43.2 (C-26), 95.9 (C-7), 123.4 (C-1), 126.1 (C-29), 128.5 (C- ³J = 8.0 Hz, 2 H, H-22), 7.33 (m, 5 H, H-30, H-31, H-32) ppm.

Ar), 129.1 (C-Ar), 129.6 (C-Ar), 133.8 (C-Ar), 139.6 (C-21, C-24), M S ( E S I p o s i t i v e ) : m / z = 6 5 5 [ M + H ] ⁺. H R M S fo r

141.2 (C-28), 175.1 (C-19), 175.7 (C-8, C-9) ppm. MS (ESI nega-

C₃₂H₃₃F₃N₆O₆H⁺[M + H]⁺(ESI positive, MeOH): m/z, found

655.2498; calcd. 655.2492.

tive): m/z = 582 [M – H]^–.

Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

147

M. Wehner, D. Janssen, G. Schäfer, T. Schrader

FULL PAPER

3

²J(H,H) = 4.3 Hz, 1 H, 4), 3.52 (s, 3 H, 5), 4.18 (ddd, J(H,H) =

3

7.3 Hz, 1 H, 6), 4.43 (m, 1 H, 7), 7.98 (d, J(H,H) = 8.6 Hz, 1 H,

8), 8.40 (d, J(H,H) = 7.0 Hz, 1 H, 9) ppm. ¹³C NMR (50.4 MHz,

3

[D₄]MeOH, 25 °C): δ = 17.4 (a), 22.4 (b), 39.0 (c), 52.8 (d), 55.8 +

55.9 (e,f), 127.7 (g), 129.4 (h), 130.3 (i), 138.5 (j), 173.1 + 173.6 (k,

l), 174.4 (m) ppm. MS (EI): m/z = 292 [M]⁺(15 %), 233 [M –

CH₃C(O)NH]⁺(25%), 162 (22%), 159 (38%), 131 [C(O)NH-Ala-

OMe]⁺(29 %), 120 (100 %). HRMS (EI): m/z, found 292.1422;

calcd. 292.1423. C₁₅H₂₀N₂O₄: calcd. C 61.63, H 6.90, N 9.58; found

C 61.53, H 7.39, N 10.14.

[4-({1,5,7-Trimethyl-2,4-dioxo-3-[5-(2,2,2-trifluoroacetylamino)-2H-

pyrazol-3-yl]-3-azabicyclo[3.3.1]nonane-7-carbonyl}amino)benzyl-

amine Hydrochloride (9): Compound 9a (13.5 mg, 2 mmol) was dis-

solved in methanol (10 mL), treated with Pd(OH)₂(20 % on C,

20 mg) and quantitatively hydrogenated under normal pressure for

2 d. After filtration and evaporation of the solvent the free amine

9 was obtained as a colourless oil.

The free amine (0.4 mg, 2 mmol) was subsequently dissolved in dry

methanol (10 mL) and cooled to 10 °C. Under pH control (glass

electrode) a solution of HCl (1 m) in diethyl ether was carefully

added dropwise, until a pH of 7.0–7.5 was reached. Stirring was

continued at ambient temperature for another half hour, and the

solvent was then concentrated to dryness and the resulting hydro-

chloride salt was dried in vacuo. Colourless oil, which solidifies

after several days. Yield 11 mg (99%).

N-Acetyl-L-alaninyl-L-phenylalanine Methyl Ester (11): N-Acetyl-l-

alanine (0.48 g, 3.7 mmol), l-phenyalanine methyl ester hydrochlo-

ride (0.78 g, 3.6 mmol) and N-methylmorpholine (2.5 mL,

23 mmol) were dissolved in dry dichloromethane (50 mL) and co-

oled down to –10 °C. T3P (2,6 mL, 16 mmol) was added slowly

and the reaction mixture was stirred for 3 h in the cold and after-

wards for 3 d at room temperature. The solution was concentrated

and the residue was subsequently washed with ethyl acetate and

HCl (1 n), satd. aqueous NaHCO₃, satd. aqueous NaCl and water.

The organic layer was dried with magnesium sulfate and filtered,

and the solvents were removed. The residue was chromatographed

over silica gel with elution with CH₂Cl₂/methanol (20:1); R_F= 0.23;

1

2

Compound 9: H NMR (300 MHz, [D₄]MeOH): δ = 1.44 (d, J =

14.4 Hz, 2 H, H-13,15), 1.50 (s, 6 H, H-16,17), 1.52 (s, 3 H, H-18),

2

2.06 (d, J = 11.0 Hz, 1 H, H-11), 2.28 (d, J = 11.0 Hz, 1 H, H-

11Ј), 2.84 (d, ²J = 14.4 Hz, 2 H, H-13Ј,15Ј), 3.73 (s, 2 H, H-25),

7.08 (d, ³J = 8.4 Hz, 2 H, H-23), 7.21 (d, ³J = 8.4 Hz, 2 H, H-

22). HRMS for C₂₄H₂₂D₅F₃N₆O₄⁺+ Li [M + Li]⁺(ESI positive,

1

Yield 0.65 g (2.2 mmol, 61%); m. p. 122 °C. H NMR (300 MHz,

3

+

[D₆]DMSO, 28 °C): δ = 0.94 (d, J(H,H) = 2.0 Hz, 3 H, 1), 1.60 (s,

CD₃OH): m/z, found 532.2580; calcd. 532.2515. C₂₄H₂₂D₅F₃N₆O₄

+ Na [M + Na]⁺: found 548.2324; calcd. 548.2252.

3 H, 2), 2.80 (m, 2 H, 3), 3.39 (s, 3 H, 4), 4.10 (m, 1 H, 5), 4.26

3

(m, 1 H, 6), 7.02 (m, 5 H, 7), 7.78 (d, J(H,H) = 8.0 Hz, 1 H, 8),

Peptides

8.07 (d, ³J(H,H) = 7.6 Hz, 1 H, 9) ppm. ¹³C NMR (50.4 MHz, [D₄]-

MeOH, 25 °C): δ = 18.1 (a), 22.6 (b), 50.2 (c), 52.8 (d), 54.9 + 55.2

(e, f), 128.0 (g), 129.6 (h), 130.4 (i), 138.0 (j), 173.0 + 173.3 + 175.0

(k, l, m) ppm. MS (FD): m/z = 292 [M]⁺. HRMS (EI): m/z, found

292.1421; calcd 292.1423.

N-Acetyl-L-phenylalaninyl-L-alanine Methyl Ester (10): l-Alanine

methyl ester hydrochloride (0.1 g, 0.7 mmol), N-acetyl-l-phenylala-

nine (0.2 g, 1 mmol) and N-methylmorpholine (0.55 mL, 5 mmol)

were dissolved in dry dichloromethane (50 mL) and cooled to

–10 °C. T3P (0.5 mL, 3.1 mmol) was then added slowly, and stir-

ring was continued for 3 h in the cold and afterwards for 3 d at

room temperature. The solution was concentrated and sub-

sequently washed with HCl (1 n), satd. aqueous NaHCO₃, satd.

aqueous NaCl and water. The organic layer was dried with magne-

sium sulfate, filtered and concentrated to dryness. R_F= 0.16 (ethyl

acetate); Yield 0.18 g (0.6 mmol, 88%); m. p. 149–155 °C. ¹H NMR

(300 MHz, [D₆]DMSO, 28 °C): δ = 1.20 (d, ³J(H,H) = 7.3 Hz, 3 H,

N-Acetyl-(O-tert-butyl)-L-serinyl-L-alanine Methyl Ester (12): N-

Acetyl-l-serine tert-butyl ether (0.61 g, 3 mmol), l-alanine methyl

ester hydrochloride (0.43 g, 3 mmol) and N-methylmorpholine

(1.8 mL, 16.4 mmol) were dissolved in dry dichloromethane

(50 mL) and cooled down to –10 °C. T3P (2.1 mL, 13 mmol) was

added slowly and stirring was continued for 3 h in the cold and

another 3 d at room temperature. The solution was concentrated

and washed with ethyl acetate and HCl (1 n), satd. aqueous

NaHCO₃, satd. aqueous NaCl and water. The organic layer was

dried with magnesium sulfate, filtered and concentrated to dryness.

The resulting crude product was NMR-pure and was used without

further purification. Yield 0.32 g (1.1 mmol, 37%); m. p. 108 °C.

2

1), 1.63 (s, 3 H, 2), 2.60 (dd, J(H,H) = 10.3 Hz, 1 H, 3), 2.89 (dd,

148

www.eurjoc.org

Eur. J. Org. Chem. 2006, 138–153

Sequence-Selective Peptide Recognition with Designed Modules

FULL PAPER

¹H NMR (300 MHz, [D₆]DMSO, 28 °C): δ = 1.17 (s, 9 H, 1), 1.34

and kept overnight at room temperature. The solution was concen-

trated to dryness and was subsequently chromatographed over sil-

(d, ³J(H,H) =Hz, 3 H, 2), 1.91 (s, 3 H, 3), 2.60 (dd, ²J(H,H) =

2

10.3 Hz, 1 H, 3), 2.89 (dd, J(H,H) = 4.3 Hz, 1 H, 4), 3.52 (s, 3 H, ica gel (ethyl acetate, R_F= 0.11). Yield 84 mg (0.3 mmol, 51%);

3

5), 4.18 (ddd, J(H,H) = 7.3 Hz, 1 H, 6), 4.43 (m, 1 H, 7), 7.98 (d,

m. p. 129 °C. ¹H NMR (300 MHz, [D₆]DMSO, 28 °C): δ = 0.86

3

³J(H,H) = 8.6 Hz, 1 H, 8), 8.40 (d, J(H,H) = 7.0 Hz, 1 H, 9).

(m, 6 H, 1), 1.85 (s, 3 H, 2), 2.03 (m, 1 H, 3), 3.62 (s, 3 H, 4), 4.20

3

(m, 3 H, 5 + 6), 4.72 (m, 1 H, 7), 8.22 (d, J(H,H) = 11.0 Hz, 1 H,

3

8), 8.31 (d, J(H,H) = 8.0 Hz, 1 H, 9) ppm. ¹³C NMR (50.4 MHz,

[D₄]MeOH, 25 °C): δ = 18.4 + 19.5 (a + b), 22.5 (c), 32.0 (d), 52.6

(e), 56.6 + 59.3 (f, g), 63.0 (h), 172.7 + 173.3 + 173.5 (i, j, k) ppm.

HRMS (ESI for C₁₁H₂₀O₅N₂Na): m/z, found 283.12586; calcd.

283.12699. C₁₁H₂₀O₅N₂: calcd. C 50.76, H 7.74, N 10.76; found C

50.47, H 7.68, N 10.22.

N-Acetyl-L-serinyl-L-alanine Methyl Ester (12a): Compound 12

(0.23 g, 0.8 mmol) was dissolved in formic acid (p. a., 40 mL) and

kept overnight at room temperature. The solution was concentrated

to dryness and the residue was purified by chromatography

(CH₂Cl₂/methanol, 20:1, R_F= 0.1). Yield: 0.12 g (0.5 mmol, 65%);

m. p. 135 °C. ¹H NMR (300 MHz, [D₆]DMSO, 28 °C): δ = 1.28 (d,

³J(H,H) =Hz, 3 H, 1), 1.86 (s, 3 H, 2), 3.54 (m, 2 H, 3), 3.62 (s, 3

3

H, 4), 4.30 (m, 1 H, 5), 7.91 (d, J(H,H) = 7.0 Hz, 1 H, 6), 8.28 (d,

³J(H,H) = 7.3 Hz, 1 H, 7) ppm. ¹³C NMR (50.4 MHz, [D₄]MeOH,

25 °C): δ = 17.5 (a), 22.6 (b), 52.8 (c), 56.6 + 56.7 (d, e), 63.2 (f),

172.3 + 173.5 (g, h), 174.6 (i) ppm. MS (FD): m/z = 233 [M +

H]⁺. HRMS (ESI, for C₉H₁₆O₅N₂Na): m/z, found 255.09478;

calcd. 255.09569. C₉H₁₆O₅N₂: calcd. C 46.55, H 6.94, N 12.06;

found C 45.59, H 6.48, N 12.00.

N-Acetyl-L-valinyl-L-serine Methyl Ester (14): N-Acetyl-l-valine

(0.67 g, 4.2 mmol), l-serine methyl ester hydrochloride (0.61 g,

3.9 mmol), PyCloP (2.1 g, 5 mmol) and DIEA (2.6 mL, 15 mmol)

were dissolved in dry dichloromethane (50 mL) and were stirred

for 2 days at room temperature. The solution was concentrated to

dryness and the residue was purified by chromatography over silica

gel (CH₂Cl₂/methanol, 20:1, R_F= 0.49). Yield 0.1 g (0.4 mmol,

9%); m. p. 212 °C. ¹H NMR (300 MHz, [D₆]DMSO, 28 °C): δ =

3

0.99 (dd, J(H,H) = 7.0 Hz, 6 H, 1), 2.00 (s, 3 H, 2), 2.09 (m, 1 H,

3), 3.76 (s, 3 H, 4), 3.81 (m, 3 H, 5), 4.43 (m, 2 H, 6 +7), 8.03 (d,

³J(H,H) = 9.0 Hz, 1 H, 8), 8.41 (d, ³J(H,H) = 7.3 Hz, 1 H, 9) ppm.

¹³C NMR (50.4 MHz, [D₄]MeOH, 25 °C): δ = 17.3 + 18.6 (a + b),

22.4 (c), 32.0 (d), 52.7 (e), 56.2 + 60.2 (f, g), 62.8 (h), 172.1 + 173.4

(i, j), 173.9 (k) ppm. MS (CI, NH₃, 200 °C): m/z = 261 [M + H]⁺

(59 %), 278 [M + NH₃]⁺(100 %). HRMS (ESI calcd. for

C

₁₁H₂₀O₅N₂Na): m/z, found 283.12599; calcd. 283.12699.

C₁₁H₂₀O₅N₂: calcd. C 50.76, H 7.74, N 10.76; found C 50.47, H

7.68, N 10.22.

N-Acetyl-O-tert-butyl-L-serinyl-L-valine Methyl Ester (13): N-Ace-

tyl-l-serine tert-butyl ether (0.51 g, 2.5 mmol), l-valine methyl ester

hydrochloride (0.39 g, 2.3 mmol), PyCloP (1.1 g, 2.6 mmol) and

DIEA (1.6 mL, 9 mmol) were stirred in dry dichloromethane

(50 mL) for 2 days at room temperature. The solution was concen-

trated to dryness and the residue was chromatographed over silica

gel (ethyl acetate, R_F= 0.44). The crude product was NMR spec-

troscopically pure and was used for the next step without further

1

purification. Yield 0.3 g (1.1 mmol, 48%); m. p. 107 °C. H NMR

(300 MHz, [D₆]DMSO, 28 °C): δ = 0.85 (m, 6 H, 1), 1.08 (s, 9 H,

2), 1.83 (s, 3 H, 3), 2.00 (m, 1 H, 4), 3.40 (m, 2 H, 5), 3.60 (s, 3 H,

5), 4.17 (dd, ³J(H,H) = 6.6 Hz, 1 H, 7), 4.41 (dd, ³J(H,H) = 8.0 Hz,

1 H, 8), 7.89 (d, ³J(H,H) = 8.3 Hz, 1 H, 9), 8.00 (d, ³J(H,H) =

8.6 Hz, 1 H, 10) ppm.

N-Acetyl-N^δ-benzyloxycarbonyl- -ornithine (15): N^δ-Benzyloxycar-

L

bonyl-l-ornithine (0.5 g, 1.9 mmol) was suspended in water

(100 mL) and treated with acetic acid anhydride (1 mL). The reac-

tion mixture was sonicated for 15 min, treated once more with ace-

tic acid anhydride (1 mL) and sonicated again. This procedure was

repeated six times. The solvent was evaporated and completely

eliminated by threefold azeotropic distillation with small amounts

N-Acetyl-

L-serinyl-L-valine Methyl Ester (13a): Compound 13

(0.2 g, 0.63 mmol) was dissolved in dry formic acid (p. a., 40 mL)

Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

149

M. Wehner, D. Janssen, G. Schäfer, T. Schrader

FULL PAPER

of toluene. The crude product was pure according to NMR and

was used for the next step without further purification. Yield 0.58 g

3.61 (s, 3 H, 6), 4.14 (br., 3 H, 7), 4.26 (m, 2 H, 8), 8.05 (d, ³J(H,H)

= 8.0 Hz, 1 H, 9), 8.47 (d, ³J(H,H) = 6.6 Hz, 1 H, 10) ppm. ¹³C

(1.9 mmol, 100%); m. p. 96 °C. ¹H NMR (300 MHz, [D₆]DMSO, NMR (50.4 MHz, [D₄]MeOH, 25 °C): δ = 17.3 (a), 22.5 (b), 24.1

3

28 °C): δ = 1.58–2.08 (m, 4 H, 1), 1.63 (s, 3 H, 2), 3.16 (q, J(H,H) + 25.0 (c, d), 30.2 (e), 40.2 (f), 52.8 (g), 53.7 + 53.9 (h, i), 173.3 +

= 6.3 Hz, 2 H, 3), 4.31 (m, 1 H, 4), 5.17 (s, 2 H, 5), 7.51 (m, 5 H,

6): δ = 8.27 (d, J(H,H) = 8.0 Hz, 1 H, 7), 12.48 (br., 1 H, 8) ppm.

173.7 + 174.5 + 174.6 (j, k, l, m). HRMS (ESI for C₁₁H₂₃O₄N₃

[M – Ac]⁺): m/z, found 260.16110; calcd. 260.16103. C₁₃H₂₅N₃O₆:

calcd. C 48.89, H 7.89, N 13.16; found C 48.80, H 7.96, N 12.64.

3

N-tert-Butoxycarbonyl-L-valinyl-L-valine Methyl Ester (17): N-tert-

Butoxycarbonyl-l-valine (0.3 g, 1.1 mmol), l-valine methyl ester

(0.23 g, 1.4 mmol), PyCloP (0.58 g, 1.4 mmol) and DIEA (0.9 mL,

5.2 mmol) were dissolved in dry dichloromethane (50 mL) and

stirred overnight. The solution was concentrated and the residue

was purified by chromatography over silica gel to give a colourless

solid (R_F= 0.50 with ethyl acetate/n-hexane, 1:2). Yield 0.23 g

(0.7 mmol, 65 %); m. p. 167 °C. ¹H NMR (300 MHz, CDCl₃,

25 °C): δ = 0.86 (m, 3 H, 1), 1.40 (s, 9 H, 2), 2.10 (m, 2 H, 3), 3.67

(s, 3 H, 4), 3.83 (ddd, ³J(H,H) = 8.9 Hz, 1 H, 5), 4.48 (ddd, ³J(H,H)

= 8.6 Hz, 1 H, 6), 4.95 (d, ³J(H,H) = 6.0 Hz, 1 H, 7), 6.25 (d,

³J(H,H) = 7.6 Hz, 1 H, 7) ppm. ¹³C NMR (50.4 MHz, [D₄]MeOH,

25 °C): δ = 17.7 + 17.8 + 18.8 + 19.2 (a, b, c, d), 28.2 (e), 30.5 +

31.2 (f, g), 52.0 (h), 57.0 (i, j), 155.7 (k), 171.5 + 172.0 (l, m) ppm.

MS (FD): m/z = 348 [M + NH₄]⁺(100%), 331 [M + H]⁺(90%).

C₁₆H₃₀N₂O₅: calcd. C 58.16, H 9.15, N 8.48; found C 57.94, H

9.19, N 8.49.

N-Acetyl-N^δ-benzyloxycarbonyl-

L-ornithinyl-L-alanine Methyl Ester

(15a): Compound 15 (1.36 g, 4.4 mmol), l-alanine methyl ester hy-

drochloride (0.58 g, 4.1 mmol), DIEA (4 mL, 23 mmol) and Py-

CloP (2.1 g, 5 mmol) were stirred in dry dichloromethane (50 mL)

for three days at room temperature. The solution was concentrated

and the residue was purified by chromatography over silica gel

(CH₂Cl₂/methanol, 20:1, R_F= 0.08). The remaining oil was recrys-

tallized from methanol/diethyl ether. Yield 1.21 g (3.1 mmol, 75%);

m. p. 176 °C. ¹H NMR (300 MHz, [D₆]DMSO, 28 °C): δ = 1.28 (d,

³J(H,H) = 7.3 Hz, 3 H, 1), 1.42–1.67 (m, 4 H, 2), 1.83 (s, 3 H, 3),

3

2.98 (q, J(H,H) = 6.0 Hz, 2 H, 4), 3.60 (s, 3 H, 5), 4.25 (m, 2 H,

3

6), 5.01 (s, 2 H, 7), 7.35 (m, 5 H, 8), 7.97 (d, J(H,H) = 8.0 Hz, 1

H, 9), 8.35 (d, ³J(H,H) = 7.0 Hz, 1 H, 10) ppm. ¹³C NMR

(50.4 MHz, [D₄]MeOH, 25 °C): δ = 17.4 (a), 22.5 (b), 27.4 (c), 30.6

(d), 41.3 (e), 52.9 (f), 54.2 + 56.0 (g, h), 70.9 (i), 128.9 + 129.1 +

129.6 (j, k, l), 138.5 (m), 159.1 (n), 173.3 + 174.2 + 174.6 (o, p,

q) ppm. MS (EI): m/z = 393 [M]⁺(5%), 291 [Ac-Orn(Z)]⁺(6%),

104 [Ala-OMe + H]⁺(16%), 91 [C₆H₅–CH₂]⁺(100%). HRMS (EI):

m/z, found 393.1900; calcd. 393.1894.

N-Acetyl-L-leucinyl-L-alanine Methyl Ester (18): Ac-Leu-OH

(173.2 mg, 1 mmol), O-(6-chloro-1H-benzotriazol-1-yl)-1,1,3,3-tet-

ramethyluronium hexafluorophosphate (HCTU, 414 mg,

1.0 mmol), 6-chloro-1-hydroxybenzotriazole (Cl-HOBt, 424 mg,

2.5 mmol) and 2,6-lutidine (0.35 mL, 3.0 mmol) were dissolved in

dry dichloromethane and DMF (3:1, 40 mL). The solution was co-

oled in an ice bath and treated with a solution of H-Ala-OMe hy-

drochloride (140 mg, 1.0 mmol) and 2,6-lutidine (0.12 mL,

1.0 mmol) in dichloromethane (1 mL). The reaction mixture was

stirred for 2 h at 0 °C. Subsequently, the solution was washed thrice

with satd. aqueous NaHCO₃, followed by HCl (1 m) and once with

satd. NaCl. The aqueous layer was extracted with dichloromethane.

After drying over Na₂SO₄, the combined organic phases were fi-

nally concentrated to dryness. Yield 189 mg (0.73 mmol, 73%);

N-Acetyl-L-ornithinyl-L-alanine Methyl Ester Hydroacetate (16):

Compound 15a (0.2 g, 0.5 mmol) was dissolved in methanol

(40 mL) and treated with acetic acid (10 mL) and Pd/C (100 mg).

Hydrogenolysis was carried out under hydrogen at normal pressure

overnight. The catalyst was filtered off over silica gel and the resi-

1

3

m. p. 114 °C. H NMR (200 MHz, [D₆]DMSO): δ = 0.84 (d, J =

due was concentrated to dryness. A colourless solid remained. 6.5 Hz, 3 H, H-12/13), 0.88 (d, ³J = 6.5 Hz, 3 H, H-12/13), 1.27 (d,

3

Yield 0.13 g (0.4 mmol, 81%); m. p. 40 °C. ¹H NMR (300 MHz, ³J = 7.5 Hz, 3 H, H-14), 1.42 (d, J = 6.0 Hz, 2 H, H-10), 1.57 (m,

3

[D₆]DMSO, 28 °C): δ = 1.28 (d, J(H,H) = 7.3 Hz, 3 H, 1), 1.45–

1 H, H-11), 1.81 (s, 3 H, H-1), 3.30 (s, 3 H, H-9), 4.23 (m, 1 H, H-

3

1.70 (m, 4 H, 2), 1.82 (s, 3 H, 3), 1.83 (s, 3 H, 4), 2.61 (m, 1 H, 5), 4), 4.33 (q, J = 6.3 Hz, 1 H, H-7), 7.98 (d, J = 8.5 Hz, 1 H, H-3/

150 Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

Sequence-Selective Peptide Recognition with Designed Modules

FULL PAPER

6), 8.39 (d, ³J = 7.0 Hz, 1 H, H-3/6) ppm. ¹³C NMR (200 MHz, (0.82 mmol, 82 %); m. p. 103 °C; ¹H NMR: (300 MHz, [D₆]-

3

[D₆]DMSO): δ = 16.7 (C-14), 21.7 (C-12/13), 22.4 (C-11), 22.9 (C- DMSO): δ = 1.28 (d, J = 7.3 Hz, 3 H, H-17), 1.39 (s, 9 H, H-14,

12/13), 24.0 (C-1), 41.0 (C-10), 47.4 (C-9), 50.5 (C-7), 51.7 (C-4),

168.9 (C-2), 172.1 (C-8), 172.8 (C-5) ppm. MS (ESI positive):

H-15, H-16), 1.73 (m, 2 H, H-10), 1.82 (s, 3 H, H-1), 2.23 (m, 2 H,

H-11), 3.61 (s, 3 H, H-9), 4.25 (m, 2 H, H-4, H-7), 8.01 (d, ³J =

m/z = 281 [M + Na]⁺. HRMS for C₁₂H₂₂N₂O₄Na⁺[M + Na⁺]⁺8.0 Hz, 1 H, H-3/6), 8.40 (d, ³J = 6.8 Hz, 1 H, H-3/6) ppm. ¹³C

(ESI positive): calcd. 281.1477; found 281.1479.

NMR: (300 MHz, DMSO): δ = 16.7 (C-17), 22.4 (C-1), 27.6 (C-

14, C-15, C-16), 27.7 (C-10), 31.1 (C-11), 47.5 (C-9), 51.2 (C-7),

51.7 (C-4), 79.6 (C-13), 169.0 (C-2), 171.1 (C-12), 171.6 (C-8), 172.8

(C-5) ppm. MS (ESI positive): m/z = 331 [M + H]⁺⁺

.

Acetyl-L-valinyl-L-alanine Methyl Ester (19): The preparation fol-

lowed the procedure outlined above, with Ac-Val-OH and H-Ala-

OMe hydrochloride. Yield 190 mg (0.78 mmol, 78%); m. p. 207 °C;

3

¹H NMR (200 MHz, [D₆]DMSO): δ = 0.83 (d, J = 7.0 Hz, 3 H,

H-11/12), 0.87 (d, ³J = 7.3 Hz, 3 H, H-11/12), 1.27 (d, ³J = 7.5 Hz,

3 H, H-13), 1.85 (s, 3 H, H-1), 1.91 (m, 1 H, H-10), 3.60 (s, 3 H,

N-Acetyl-L-glutamyl-L-alanine Methyl Ester (21a): Compound 21

3

H-9), 4.20 (m, 2 H, H-4, H-7), 7.88 (d, J = 9.5 Hz, 1 H, H-3/6),

(50 mg, 0.15 mmol) was stirred in a mixture of dichloromethane/

trifluoroacetic acid (DCM/TFA, 10 mL, 1:1) for 2 h with subse-

quent evaporation to dryness. Yield 41 mg (0.15 mmol, 99%); m. p.

64 °C; ¹H NMR: (300 MHz, [D₆]DMSO): δ = 1.28 (d, ³J = 7.3 Hz,

3 H, H-14), 1.76 (m, 2 H, H-10), 1.83 (s, 3 H, H-1), 2.25 (m, 2 H,

H-11), 3.61 (s, 3 H, H-9), 4.28 (m, 2 H, H-4, H-7), 8.02 (d, ³J =

3

8.42 (d, J = 6.3 Hz, 1 H, H-3/6) ppm. ¹³C NMR (300 MHz, [D₆]-

DMSO): δ = 16.7 (C11, C-12), 18.1 (C-13), 18.9 (C-1), 30.5 (C-10),

47.4 (C-9), 51.6 (C-7), 57.1 (C-4), 169.0 (C-2), 171.0 (C-8), 185.1

(C-5) ppm. MS (ESI positive): m/z = 267 [M + Na]⁺. HRMS for

C₁₁H₂₀N₂O₄Na⁺[M + Na]⁺(ESI positive): calcd. 267.1321; found

267.1324.

3

8.0 Hz, 1 H, H-3/6), 8.41 (d, J = 7.0 Hz, 1 H, H-3/6), 12.13 (b, 1

H, H-13) ppm. ¹³C NMR: (300 MHz, DMSO): δ = 16.6 (C-14),

22.3 (C-1), 27.5 (C-10), 29.9 (C-11), 47.4 (C-9), 51.3 (C-7), 51.7 (C-

4), 169.0 (C-2), 171.2 (C-8), 173.8 (C-5), 195.4 (C-12) ppm. MS

(ESI positive): m/z = 275 [M + Na]⁺. HRMS for C₁₁H₁₈N₂O₆H⁺

[M + Na]⁺(ESI positive): calcd. 275.1243; found 275.1233.

N-Acetyl-L-alaninyl-L-valine Methyl Ester (20): The preparation

followed the procedure outlined above, with Ac-Ala-OH and H-

Val-OMe hydrochloride. Yield 166 mg (0.68 mmol, 68%); m. p.

105 °C; ¹H NMR (200 MHz, [D₆]DMSO): δ = 0.84 (d, ³J = 3.5 Hz,

3

3 H, H-12/13), 0.89 (d, J = 3.3 Hz, 3 H, H-12/13), 1.16 (d, J =

7.0 Hz, 3 H, H-10), 1.81 (s, 3 H, H-1), 2.02 (m, 1 H, H-11), 3.62

(s, 3 H, H-9), 4.15 (dd, ³J = 6.3 Hz, 1 H, H-7), 4.38 (q, ³J = 7.3 Hz,

3

1 H, H-4), 8.03 (d, J = 7.7 Hz, 1 H, H-3/6), 8.09 (d, J = 8.3 Hz,

1 H, H-3/6) ppm. ¹³C NMR (300 MHz, [D₆]DMSO): δ = 18.0 (C-

12/13), 18.1 (C-12/13), 18.8 (C-10), 22.3 (C-1), 29.8 (C-11), 47.7 (C-

9), 51.5 (C-4), 57.2 (C-7), 168.9 (C-2), 171.8 (C-8), 172.7 (C-5) ppm.

M S (E SI po s i tive ) : m/z = 2 6 7 [ M + N a ]⁺. H RM S for

C₁₁H₂₀N₂O₄Na⁺[M + Na]⁺(ESI positive): calcd. 267.1321; found

267.1313.

Methyl Ester and Tetrabutylammonium Salt of N-Acetyl-

L-gluta-

myl- -alanine (21b): Compound 21a (20 mg, 0.07 mmol) was dis-

L

solved in methanol (5 mL). A solution of HONBu₄in methanol

(1 m , 0.07 mmol, 70 μL) was added dropwise. The solution was

concentrated to dryness and gave a viscous oil. Yield 36 mg

1

(0.07 mmol, 99%); H NMR: (300 MHz, [D₆]DMSO): δ = 0.92 (t,

³J = 7.3 Hz, 12 H, H-17), 1.29 (m, 11 H, H-13, H-16), 1.56 (m, 8

H, H-15), 1.68 (m, 2 H, H-10), 1.77 (s, 3 H, H-1), 1.99 (m, 2 H,

H-11), 3.17 (t, ³J = 8.3 Hz, 8 H, H-14), 3.59 (s, 3 H, H-9), 4.15 (m,

2 H, H-4, H-7), 9.09 (d, ³J = 6.6 Hz, 1 H, H-3/6), 9.29 (d, ³J =

6.8 Hz, 1 H, H-3/6) ppm. ¹³C NMR: (300 MHz, DMSO): δ = 13.3

(C-17), 16.8 (C-13), 19.1 (C-16), 22.5 (C-1), 23.0 (C-15), 28.9 (C-

10), 35.3 (C-11), 47.5 (C-9), 51.5 (C-7), 53.0 (C-4), 57.5 (C-14),

168.6 (C-2), 172.1 (C-8), 173.1 (C-5), 175.1 (C-12) ppm. MS (ESI

negative): m/z = 272 [M – NBu₄– H]^2–.

Host–Guest Titrations: Two parent solutions were prepared; one

contained the indicated amount of guest found in the tables, dis-

solved in CDCl₃(7 mL). This solution was evenly distributed

among 10 NMR tubes (0.7 mL per tube). For the host parent solu-

N-Acetyl-(O-tert-butyl)-L-glutamyl-L-alanine Methyl Ester (21):

The preparation followed the procedure outlined above, with Ac-

Glu(OtBu)-OH and H-Ala-OMe hydrochloride. Yield 270 mg

Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

151

M. Wehner, D. Janssen, G. Schäfer, T. Schrader

FULL PAPER

tion the corresponding amount of host was dissolved in the indi-

cated amount of solvent (in general 1.5 mL). Defined amounts of

the host parent solution were now added successively to the 10

samples (usually in steps of 40 μL up to 400 μL). Only sharp guest

signals observable across the whole range of concentrations were

selected for the evaluation. Association constants were determined

from the Δδ values by nonlinear regression with the PC program

SigmaPlot 2000. Volume changes were taken into account. The

complete set of data is contained in the tables in the Supporting

Information; for Supporting Information see also the footnote on

the first page of this article

from the nonlinear regressions, and indicates that saturation had

not yet been reached by the end of each titration.

[1] M. Famulok, K.-S. Jeong, G. Deslongchamps, J. Rebek Jr., An-

gew. Chem. 1991, 103, 880–882.

[2] S. S. Yoon, W. C. Still, J. Am. Chem. Soc. 1993, 115, 823; C.-

T. Chen, H. Wagner, W. C. Still, Science 1998, 279, 851–853.

[3] Review:W. C. Still, Acc. Chem. Res. 1996, 29, 155–163.

[4] H. Wennemers, M. Conza, M. Nold, P. Krattiger, Chem. Eur.

J. 2001, 7, 3342–3347.

[5] T. Braxmeier, M. Demarcus, T. Fessmann, S. McAteer, J. D.

Kilburn, Chem. Eur. J. 2001, 7, 1889–1898.

Job Plots: The indicated amounts of host and guest compound

were dissolved in chloroform. Both solutions were combined in de-

fined varying ratios, ranging from molar fractions of 0.1 to 0.9. All

selected signals could be assigned unambiguously throughout the

entire titration experiment. The solvent volumes and the complete

set of complexation-induced shifts are contained in the appropriate

tables in the Supporting Information (for details see the footnote

on the first page of this article).

[6] C. Schmuck, M. Heil, ChemBioChem 2003, 4, 1232–1238.

[7] M. Sirish, V. A. Chertkov, H.-J. Schneider, Chem. Eur. J. 2002,

8, 1181–1188.

[8] T. Mizutani, K. Wada, S. Kitagawa, J. Am. Chem. Soc. 1999,

121, 11425–11431.

[9] K. Ariga, A. Kamino, X. Cha, T. Kunitake, Langmuir 1999,

15, 3875–3885.

[10] D. H. Williams, B. Bardsley, Angew. Chem. Int. Ed. 1999, 38,

1172–1193.

[11] J. M. Gardner, R. O. Hynes, Cell 1985, 42, 439–448.

[12] M. D. Pierschbacher, E. Ruoslathi, J. Sundelin, P. Lind, P. A.

Peterson, J. Biol. Chem. 1982, 257, 9593–9595.

[13] A. Miyauchi, J. Alvarez, E. M. Greenfield, A. Teti, M. Grano,

S. Colucci, A. Zambonin-Zallone, F. P. Ross, S. L. Teitelbaum,

D. A. Cheresh, K. A. Hruska, J. Biol. Chem. 1991, 266, 369–

374.

[14] M. H. Helfrich, S. A. Nesbitt, E. L. Dorey, M. A. Horton, J.

Bone Miner. Res. 1992, 7, 335–343.

[15] M.A. Horton, G.A. Rodan, Integrins as therapeutic targets in

bone disease, Boca Raton: CRC Press, 1996.

[16] J. Folkman, M. Klagsbrun, Science 1987, 235, 442–447.

[17] E. Ruoslathi, M. D. Pierschbacher, Science 1987, 238, 491–497.

[18] S. M. Albelda, C. A. Buck, FASEB J. 1990, 4, 2868–2880.

[19] R. O. Hynes, A. D. Lander, Cell 1992, 68, 303–322.

[20] J. Travis, Science 1993, 260, 906–908.

[21] A. B. Smith III, R. Hirschmann, A. Pasternak, R. Akaishi,

M. C. Guzman, D. R. Jones, T. P. Keenan, P. A. Sprengeler,

P. L. Darke, E. A. Emini, M. K. Holloway, W. A. Schleif, J.

Med. Chem. 1994, 37, 215.

[22] a) Review: J. S. Nowick, Acc. Chem. Res. 1999, 32, 287–296; b)

J. H. Tsai, A. S. Waldman, J. S. Nowick, Bioorg. Med. Chem.

1999, 7, 29–38.

[23] D. S. Kemp, B. R. Bowen, C. C. Muendel, J. Org. Chem. 1990,

55, 4650.

[24] V. Brandmeier, W. H. B. Sauer, M. Feigel, Helv. Chim. Acta

1994, 77, 70.

[25] P. Chitnumsub, W. R. Fiori, H. A. Lashuel, H. Diaz, J. W.

Kelly, Bioorg. Med. Chem. 1999, 7, 39–59.

[26] H. L. Schenck, S. Gellman, J. Am. Chem. Soc. 1998, 120, 4869.

[27] T. Schrader, C. Kirsten, Chem. Commun. 1996, 2089.

[28] W. S. Weiner, A. D. Hamilton, Bioorg. Med. Chem. Lett. 1998,

8, 681–686.

[29] S. T. Phillips, M. Rezac, U. Abel, M. Kossenjans, P. A. Bartlett,

J. Am. Chem. Soc. 2002, 124, 58–66.

[30] T. Schrader, M. Wehner, Angew. Chem. 2002, 114, 1827–1831.

[31] J. Rebek Jr., Angew. Chem. Int. Ed. Engl. 1990, 29, 245–255.

[32] T. Bach, H. Bergmann, K. Harms, J. Am. Chem. Soc. 1999,

121, 10650–10651.

[33] S. Shinoda, H. Tsukube, Y. Nishimura, I. Yamazaki, A. Osuka,

J. Org. Chem. 1999, 64, 3757–3762.

[34] B. Askew, P. Ballester, C. Buhr, K. S. Jeong, S. Jones, K. Parris,

K. Williams, J. Rebek, Jr., J. Am. Chem. Soc. 1989, 111, 1082–

1090.

Self-Association Experiments: A self-association experiment with

serine binder 4 over a concentration range of 2×10^–2to 2 ×10^–3

m

produced no change in the chemical shift (less than 0.03 ppm) of

any NMR signal. Obviously the bulky substituted cyclohexane

moiety of Kemp’s triacid prevents any close approach of two host

molecules to interact through their aminopyrazole binding sites.

Considerable self-association might be expected for the ornithine-

containing dipeptide. Although a self-association experiment in

CDCl₃over the concentration range of 3 × 10^–2to 3 × 10^–3m pro-

duced a large upfield shift of the C-terminal NH (Δδ = 1.2 ppm),

the binding curve remained almost completely linear, indicating a

negligible extent of self-association. We explain this upfield shift in

terms of an intermolecular hydrogen bond between the appropriate

amide proton and the acetate counterion of another peptide mole-

cule.

Molecular Modelling Experiments: Force-field calculations were

performed on a Silicon Graphics O2 workstation with Macro-

Model 7.0. The MM3 and OPLS-AA force-fields were both used

for most calculations and produced similar results in chloroform.

Monte-Carlo conformational searches were conducted over 1000

steps.

NOESY Experiments: The serine binder 4 was dissolved at a con-

centration of 6.1 mgmL^–1= 11.8 mm in a mixture of CD₂Cl₂and

CD₃CN (2:1); most NMR signals were sharp including the amidic

NH protons of aminopyrazole and aminoacetophenone. A stan-

dard NOESY experiment was performed at 400 MHz and room

temperature. Only positive cross-peaks were produced. All ob-

served cross-peaks were reciprocal; they were strong in the Kemp’s

triacid part of the host molecule and weaker in the aromatic region.

However, a distinct reciprocal NOE was found between two specific

aromatic CH protons of the aminopyrazole and the aminoace-

tophenone moiety (see also Figure 6).

Karplus Analysis: ³J coupling constants between the NH proton

and the corresponding α-protons were measured by NMR in the

free and complexed states of three selected peptides: Ac-Ser-Ala-

OMe, Ac-Phe-Ala-OMe and Ac-Orn-Ala-OMe were each com-

plexed with the host molecule designed for the specific additional

noncovalent interaction. These values were compared with the cal-

3

culated J coupling constants obtained from Monte-Carlo simula-

[35] S. Dey, P. Garner, J. Org. Chem. 2000, 65, 7697–7699.

[36] a) T. Yvernault, G. Yvernault, J. C. Bollinger, C. R. Acad. Sci.

Paris 1978, 287, 519–521; b) J. Coste, E. Frerot, P. Jouin, J.

Org. Chem. 1994, 59, 2436–2446.

tions of the optimized complex geometries. In most cases, the ex-

perimental value found for the complex in solution was in-between

the experimentally measured value for the free peptide and the cal-

culated value for the pure complex. This correlates with the results

[37] J. L. Sargent, J. Org. Chem. 1954, 19, 599–607.

152

www.eurjoc.org

Eur. J. Org. Chem. 2006, 138–153

Sequence-Selective Peptide Recognition with Designed Modules

FULL PAPER

[38] S. Kubik, J. Am. Chem. Soc. 1999, 121, 5846–5855.

[39] MacroModel 7.0, Schrödinger Inc., 2000. Force-field: Amber*

and OPLS-AA, solvent: chloroform (Monte-Carlo: 2000 steps;

MD: 100 ps, 300 K, snapshots taken every 10 ps).

[40] a) H. J. Schneider, R. Kramer, S. Simova, U. Schneider, J. Am.

Chem. Soc. 1988, 110, 6442; b) C. S. Wilcox, in: Frontiers in

Supramolecular Chemistry (Ed.: H. J. Schneider), Verlag

Chemie, Weinheim, 1991, p. 123.

[41] a) P. Job, Compt. Rend. 1925, 180, 928; b) M. T. Blanda, J. H.

Horner, M. Newcomb, J. Org. Chem. 1989, 54, 4626.

[42] J. S. Nowick, S. Insaf, J. Am. Chem. Soc. 1997, 119, 10903;

the Ala-coil: K. M. Gernert, M. C. Surles, T. H. LaBean, J. S.

Richardson, D. C. Richardson, Protein Sci. 1994, 4, 2252–

2260.

[43] EPS calculations predicted a notable difference between the

electron-rich surfaces in Phe and Trp and the electron-poor

surface of the m-nitroaniline in host 5: F.-G. Klärner, J. Pan-

itzky, D. Preda, L. T. Scott, J. Mol. Model. 2000, 5, 318.

[44] a) M. Delepierre, C. M. Dobson, F. M. Poulsen, Biochemistry

1982, 21, 4756; b) V. F. Bystrov, Prog. Nucl. Magn. Reson. Spec-

trosc. 1976, 10, 41.

Received: June 1, 2005

Published Online: November 9, 2005

Eur. J. Org. Chem. 2006, 138–153

www.eurjoc.org

153

Article Doi

Sequence-selective peptide recognition with designed modules

DOI: 10.1002/ejoc.200500393

Source and publish data:

Authors:

Article abstract of DOI:10.1002/ejoc.200500393

Full text of DOI:10.1002/ejoc.200500393

Products guided by the article

R&D Labs maybe for 35515-72-1

Relevant to this article

Hot Product