Synthesis and conformation of dipeptide taste ligands containing homo-β-amino acid residues

Article abstract of DOI:10.1002/(SICI)1099-1395(199907)12:7<577::AID-POC159>3.0.CO;2-F

The synthesis and conformational properties of a series of dipeptide taste ligands, differing from the commercial sweetener aspartame by the presence of a methylene group between the C^α and the C′ carbon atoms (as in homo-β-residues) in either the L-Asp or the L-Phe residues, are described. Homo-β-residues such as homo-β-aspartic acid, homo-β-phenylglycine and homo-β-phenylalanine, obtained by homologation of the corresponding proteinogenic α-amino acids, have been used in the solution peptide synthesis of the following aspartame analogues in protected and unprotected forms: NH₂-homo-β-(L or D)-Asp-L-Phe-OMe, NH₂-L-Asp-homo-β-L-Phg-OMe and NH₂-L-Asp-homo-β-L-Phe-OMe. Lengthening of the peptide skeleton at the L-Asp site results in a drastic loss of sweetness with the production of tasteless compounds; on the other hand, lengthening of the skeleton at the C-terminal L-Phe site partially mantains the sweet taste in both NH₂-L-Asp-hamo-β-L-Phe-OMe and NH₂-L-Asp-homo-β-L-Phg-OMe. The solution conformation of the synthesized dipeptide taste ligands was investigated by NMR and circular dichroism techniques. The analysis of NMR data combined with restrained molecular dynamics calculations shows that all peptides are fairly flexible and they do not assume a preferred conformation in DMSO and methanol. The peptides containing homo-β-L-Phe and homo-β-L-Phg do adopt a discrete number of conformation among which mainly extended and 'L-shaped' conformation are represented The circular dichroism spectra are consistent with the NMR results, indicating a significant flexibility for these compounds. Copyright

Full text of DOI:10.1002/(SICI)1099-1395(199907)12:7<577::AID-POC159>3.0.CO;2-F

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 12, 577–587 (1999)
Synthesis and conformation of dipeptide taste ligands
containing homo-b-amino acid residues†
C. Isernia,1 E. Bucci,2 L. De Napoli,3 P. Di Lello,2 R. Iacovino,1 D. Montesarchio,3 G. Piccialli,3 F. Rossi,2
M. Saviano2 and E. Benedetti2*
1Dipartimento di Scienze Ambientali, Seconda UniversitaÁ di Napoli, Via Arena 22, I-81100 Caserta, Italy
2Centro di Studio di Biocristallogra®a, CNR and Dipartimento di Chimica, UniversitaÁ degli Studi di Napoli `Federico II,' Via Mezzocannone 4,
I-80134 Naples, Italy
3Dipartimento di Chimica Organica e Biologica, UniversitaÁ degli Studi di Napoli `Federico II,' Via Mezzocannone 16, I-80134 Naples, Italy
Received 8 October 1998; revised 2 December 1998; accepted 3 December 1998
ABSTRACT: The synthesis and conformational properties of a series of dipeptide taste ligands, differing from the
epoc
commercial sweetener aspartame by the presence of a methylene group between the Ca and the C' carbon atoms (as in
homo-b-residues) in either the L-Asp or the L-Phe residues, are described. Homo-b-residues such as homo-b-aspartic
acid, homo-b-phenylglycine and homo-b-phenylalanine, obtained by homologation of the corresponding proteino-
genic a-amino acids, have been used in the solution peptide synthesis of the following aspartame analogues in
protected and unprotected forms: NH2-homo-b-(L or D)-Asp-L-Phe-OMe, NH2-L-Asp-homo-b-L-Phg-OMe and NH2-
L-Asp-homo-b-L-Phe-OMe. Lengthening of the peptide skeleton at the L-Asp site results in a drastic loss of sweetness
with the production of tasteless compounds; on the other hand, lengthening of the skeleton at the C-terminal L-Phe site
partially mantains the sweet taste in both NH2-L-Asp-homo-b-L-Phe-OMe and NH2-L-Asp-homo-b-L-Phg-OMe.
The solution conformation of the synthesized dipeptide taste ligands was investigated by NMR and circular
dichroism techniques. The analysis of NMR data combined with restrained molecular dynamics calculations shows
that all peptides are fairly flexible and they do not assume a preferred conformation in DMSO and methanol. The
peptides containing homo-b-L-Phe and homo-b-L-Phg do adopt a discrete number of conformations among which
mainly extended and ‘L-shaped’ conformation are represented. The circular dichroism spectra are consistent with the
NMR results, indicating a significant flexibility for these compounds. Copyright 1999 John Wiley & Sons, Ltd.
KEYWORDS: synthesis; homo-b-amino acids; conformational studies; sweetner
Additional material for this paper is available from the epoc website at http://www.wiley.com/epoc
INTRODUCTION
preparation of homo-b-amino acids using a homologation
reaction sequence.5,6 In a recent paper,7 a new and
efficient procedure to homologate N-protected a-amino
acids, with the production of the corresponding chirally
pure homo-b-derivatives, was reported. According to this
method, we have prepared in enantiopure form N- and C-
protected homo-b-(S)-leucine derivatives8 and several
new homo-b-amino acids with aliphatic, aromatic and
functionalized side-chains in order to use them in the
synthesis of analogues of biologically active compounds.
Since the discovery9 of the potent sweet-tasting a-L-
aspartyl-L-phenylalanine methyl ester (aspartame), the
special interest in dipeptide sweeteners has increased and
a large number of analogues related to aspartame have
been synthesized.10 It has been demonstrated11 that the
sweet-tasting activity of the dipeptide is preserved if the
aspartyl moiety is restricted to L-Asp or to the shorter
homologue aminomalonic acid (Ama) in which the side-
chain methylene group is missing; elongation of the side-
chain as in the higher homologue NH2-L-Glu-L-Phe-OMe
led to bitter compounds.12 For a sweet-tasting ligand the
b-Amino acids, although less abundant in nature than
a-amino acids, have become increasingly important
during the last decade. They are found in a variety of
natural products,1,2 including peptides.3 According to
Spatola,4 b-amino acids and homo-b-amino acids
‡
‡
are H3NCH2CHRCOO and H3NCHRCH2COO ,
respectively. The growing interest in this new class of
residues rests both on their yet largely unexplored
conformational behaviour and on the possibility of
inserting them as convenient new molecular tools in
bioactive natural products in order to improve both the
resistance to biodegradation and the pharmacokinetic
properties.
a-Amino acids are the ideal starting material for the
*Correspondence to: E. Benedetti, Centro di Studio di Biocristallo-
grafia, CNR and Dipartimento di Chimica, Universita` degli Studi di
Napoli ‘Federico II,’ Via Mezzocannone 4, I-80134 Naples, Italy.
E-mail: benedetti@chemna.dichi.unina.it
†Dedicated to the memory of Professor Giacomino Randazzo.
Copyright 1999 John Wiley & Sons, Ltd.
CCC 0894–3230/99/070577–11 $17.50

578
C. ISERNIA ET AL.
‘bioactive conformation’ was proposed to be an ‘L-
shaped’ structure.11,12
the determination of the temperature coefficients were
recorded at 298–318 K in steps of 5 K.
The synthesis of several aspartame dipeptide analo-
gues whose backbone was elongated by one methylene
group was performed by Miyoshi et al.,13 who analysed
the steric effect in relation to the potency of sweetness.
The structure–taste relationships accumulated so far
prompted us to investigate the effect on taste of the
systematic insertion of a methylene group between the Ca
and C' carbon atoms of either the L-Asp or the L-Phe
residue in dipeptide analogues. For this purpose, using an
innovative synthetic procedure7,8 in the preparation of
homo-b-amino acids, we prepared a series of new taste
ligand dipeptide analogues.
In the present paper we report the synthesis and
solution characterization by NMR and circular dichroism
(CD) techniques of the following homo-b-residues
containing dipeptides: NH2-homo-b-(L or D)-Asp-L-
Phe-OMe (9a and 9b), NH2-L-Asp-homo-b-L-Phg-OMe
(14) and NH2-L-Asp-homo-b-L-Phe-OMe (19). Length-
ening of the peptide skeleton at the L-Asp site results in a
drastic loss of sweetness with the production of tasteless
compounds; on the other hand, lengthening of the
skeleton at the C-terminal L-Phe site partially mantains
the sweet taste in both NH2-L-Asp-homo-b-L-Phe-OMe
and NH2-L-Asp-homo-b-L-Phg-OMe.
Phase-sensitive double quantum filtered correlated
spectroscopy (DQFCOSY),14 homonuclear Hartmann–
Hahn (HOHAHA)15 and nuclear Overhauser effect
spectroscopy in the rotating frames (ROESY)16 were
performed according to the States–Haberkorn method
and applying standard phase cycling schemes.
These experiments were typically acquired with 256
increments and 2048 data points in t2. The FIDs were
multiplied by a square shifted sine-bell and zero filled to
1K in F1 prior to Fourier transformation. The mixing time
for TOCSY was 70 ms; a continuous spin lock was used
for the compensated17 ROESY spectrum (mixing time
150 ms). Cross-peak intensities were evaluated by
volume integration using the available Varian software.
Restrained molecular dynamics. Restrained molecular
dynamics (RMD) simulations were carried out on a
Silicon Graphics O2 workstation employing the
INSIGHT/DISCOVER program.18 The consistent
valence force field (CVFF) was utilized for all simula-
tions.19–21 The equations of motion were solved using the
so-called Leapfrog integration algorithm.22
Starting models of the N-protected (13, 18) and
deprotected dipeptides (14, 19) for the RMD were built
using standard parameters supplied by the INSIGHT
software package. One hundred steps of energy mini-
mizations to eliminate hot spots23 using the conjugate
gradient method18 were performed for the starting
structure. The energy-minimized structure was used as
the initial structure for the RMD in vacuo at 300 K. All
unambiguous NOE effects were used for the RMD
calculations. In each RMD simulation, performed with a
time step of 0.5 fs, the molecule was equilibrated for 50
ps. After this first step, an additional 100 ps of simulation
without rescaling was carried out since energy conserva-
tion was observed and the average temperature remained
essentially constant around the target value of 300 K.
During this step, the structures were extracted and
minimized every 1 ps.
EXPERIMENTAL
Materials and methods. a-Amino acids and the activat-
ing agent benzotriazol-1-yloxytrispyrrolidinophospho-
nium hexafluorophosphate (PyBop) were purchased
from Novabiochem (Switzerland). Solvents were of
reagent grade; the other organic reagents, commercially
available, were used without further purification. High-
performance liquid chromatographic (HPLC) analyses
were performed on a Beckman System Gold instrument,
equipped with a UV 166 module detector and a
Chromatopac R6A recorder. Purifications were carried
out on a Millipore-Waters Delta Prep 3000 HPLC
system.
The final structure was then checked for consistency
with all observable NOE. Coordinates and velocities for
the simulation were dumped to a disk every 10 steps
during the last 20 ps of the simulations. The dumped data
were used for statistical analysis.
NMR characterization. NMR measurements were carried
out on Varian Gemini 200, Bruker 250 and Varian
1
UNITY 400 MHz spectrometers. All H NMR spectra
were recorded at 298 K and referenced to DMSO-d6
(ꢀ = 2.5 ppm), to CD3OH (ꢀ = 3.33 ppm) and CDCl3
(ꢀ = 7.26 ppm). For structural characterization, NMR
samples were prepared by dissolving about 5 mg of
peptide in 0.75 ml of DMSO-d6 (99.96% 2H atom,
Cambridge Isotope Laboratories) or in 0.75 ml of
CD3OH (99% 2H atom, Cambridge Isotope Labora-
tories). Additional very dilute solutions were prepared by
dissolving 0.2 mg of peptides in 0.75 ml of both solvents.
One-dimensional experiments were typically acquired
using 32–64 scans with 32K data points. The spectra for
Circular dichroism characterization. CD measurements
were carried out on a Jasco J715 spectropolarimeter.
Peptides 9a, 14 and 19 were dissolved in water–
5
trifluoroethanol (1:1) at concentrations of 2.29 Â 10
,
5
5
3.05 Â 10 and 1.91 Â 10 M, respectively. A single
Jasco cylindrical cuvette with a pathlength of 1 cm was
used for all measurements. The spectra were recorded
1
using a scan speed of 5 nm min with a resolution of
0.2 nm, a sensitivity of 50 mdeg, a time constant of 16 s
and a spectral window from 260 to 190 nm.
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)

DIPEPTIDE TASTE LIGANDS CONTAINING HOMO-b-AMINO ACID RESIDUES
579
Scheme 1
Syntheses
HPLC and their structures confirmed by 1H and 13C NMR
and fast atom bombardment mass spectrometric (FAB-
MS) data.
The syntheses of the NH2-homo-b-phenylalanine methyl
ester 17 and NH2-homo-b-phenylglycine methyl ester 12
were carried out starting from Boc-L-Phe-OH and Cbz-L-
Phg-OH, following a slightly modified literature proce-
dure.7 In an analogous manner, NH2-homo-b-aspartic
acid has been prepared from Boc-L-Asp-(OBzl)-OH.
As depicted in Scheme 1, the starting material was
converted into the corresponding N-protected homo-b-
amino acid via the b-amino alcohol, the b-amino iodide
and the b-amino nitrile followed by hydrolysis of the last
compound under acidic conditions. The key step in the
preparation of the desired compounds was the conversion
of the N-alkoxycarbonyl-protected b-amino alcohol into
the corresponding iodide by a polymer-bound triaryl-
phosphine–I2 complex, under very mild, high-yielding
and non-racemizing conditions. This is a very convenient
procedure to homologate a-amino acids directly; the
alternative to this process involves an application of the
Arndt–Eistert reaction, under classical24 or modified25
conditions, the major drawback of which resides in the
use of the extremely hazardous diazomethane, which
strongly limits scale-up reactions. We successfully
applied the protocol described by Caputo et al.7 on a
large scale using inexpensive reagents to obtain the target
compounds in the N-protected form. The homo-b-amino
acid derivatives were then coupled with the a-amino acid
partner in the preparation of the desired dipeptide taste
ligands following standard protocols for peptide synth-
esis in solution (Schemes 1–3). After the usual work-up,
protected and deprotected dipeptides were purified by
Synthesis of the dipeptide NH2-homo-b-Asp-Phe-
OMe (9a and 9b). Synthesis of 3-(N-tert-butyloxycar-
bonylamino)-4-hydroxy benzylbutyrate (2). To a solu-
tion of 3.2 g of Boc-Asp(Bzl)-OH (1, 10 mmol) in THF
Scheme 2
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)

580
C. ISERNIA ET AL.
Scheme 3
(30 ml), 4-methylmorpholine (1.2 ml, 11 mmol) and
methyl chloroformate (0.86 ml, 11 mmol) were added at
15°C under stirring. After 15 min, 1 M NaBH4 aqueous
solution (15 ml) was added dropwise and the resulting
mixture was left at room temperature (r.t.) until the
disappearance of gas evolution, then concentrated under
reduced pressure, redissolved in AcOEt and washed
repeatedly with water. The organic phase, dried over
Na2SO4, was taken to dryness to give 2.60 g of pure 2
(8.4 mmol, 84% yield) as a foam: m.p. 59.30°C; Rf 0.85
in petroleum ether–AcOEt (1:9, v/v). 1H NMR
(200 MHz, CDCl3), ꢀ (ppm): 7.34 (s, 5H, phenyl
protons); 5.23 (d, 1H, NH); 5.12 (s, 2H, CH2Ph); 4.02
(m, 1H, a CH); 3.68 (d, 2H, CH2OH); 2.65 (d, 2H,
CH2COOBzl); 1.44 (s, 9H, 3 CH3 Boc group). 13C NMR
(62 MHz, CDCl3), ꢀ (ppm): 170.44 (COOBzl); 154.58
(OCONH); 135.09, 128.06, 127.92, 127.79 (phenyl ring
carbons); 79.70 [(CH3)3C]; 66.12 (CH2Ph); 64.43
(CH2OH); 47.71 (a CH); 36.10 (CH2COOBzl); 28.19 (3
CH3 Boc group).
[(CH3)3C]; 66.52 (CH2Ph); 47.55 (a CH); 38.61
(CH2COOBzl); 28.16 (3 CH3 Boc group); 11.04 (CH2I).
Synthesis of 3-(N-tert-butyloxycarbonylamino)-4-cyano
benzylbutyrate (4). A 1.26 g amount of 3 (3 mmol) was
dissolved in CH2Cl2 (100 ml) and the solution was
treated with 0.97 g of tetrabutylammonium cyanide
(6.1 mmol) at reflux for 4 h and then kept at r.t. overnight.
The reaction mixture, dried in vacuo, was purified by
flash chromatography, eluting the column with increasing
amounts of AcOEt in petroleum ether. The fractions
eluted with 20% AcOEt in petroleum ether gave 530 mg
of the pure title compound (1.7 mmol, 56% yield): Rf
0.15 in cyclohexane–AcOEt (8:2, v/v). 1H NMR
(400 MHz, CDCl3) ꢀ (ppm): 7.34 (s, 5H, phenyl protons);
5.18 (d, 1H, NH); 5.07 (s, 2H, CH2Ph); 4.12 (m, 1H, a
CH); 2.35 (m, 4H, overlapped signals, CH2CN and
CH2COOBzl); 1.40 (s, 9H, 3 CH3 Boc group). 13C NMR
(100 MHz, CDCl3) ꢀ (ppm): 169.84 (COOBzl); 154.45
(OCONH); 134.94, 128.34, 128.19, 127.97 (phenyl ring
carbons); 116.78 (CN); 79.98 [(CH3)3C]; 66.61 (CH2Ph);
44.13 (a CH); 37.13 (CH2COOBzl); 27.98 (3 CH3 Boc
group); 22.59 (CH2CN).
Synthesis of 3-(N-tert-butyloxycarbonylamino)-4-iodo
benzylbutyrate (3). A mixture of 4.0 g of polystyryldi-
phenylphosphine (12 mmol), swollen in anhydrous
CH2Cl2 (100 ml), and 2.8 g of I2 (11 mmol), was left at
r.t. for 15 min, then 0.80 g of imidazole (12.5 mmol) and
1.5 g of 2 (5 mmol) were successively added under
stirring at r.t. and the resulting mixture was refluxed for
1 h, then cooled, filtered on Celite and washed with
CH2Cl2 (100 ml). The organic phase, washed with 5 M
Na2S2O3 (50 ml) and successively with water until
neutralization, was dried over Na2SO4 and concentrated
under reduced pressure, affording 1.5 g of 3 (3.5 mmol,
70% yield); m.p.: 50.40°C (from CH2Cl2–n-hexane); Rf
0.9 in petroleum ether–AcOEt (1:9, v/v). 1H NMR
(200 MHz, CDCl3) ꢀ (ppm): 7.32 (s, 5H, phenyl protons);
5.20 (d, 1H, NH); 5.10 (s, 2H, CH2Ph); 3.91 (m, 1H; a
CH); 3.38 (d, 2H, CH2I); 2.70 (d, 2H, CH2COOBzl); 1.40
(s, 9H, 3 CH3 Boc group). 13C NMR (50 MHz, CDCl3), ꢀ
(ppm): 170.37 (COOBzl); 154.52 (OCONH); 135.27,
128.45, 128.23, 128.06 (phenyl ring carbons); 79.79
Synthesis of 3-(N-¯uorenylmethoxycarbonylamino)-4-
cyano benzylbutyrate (5). A 530 mg amount of 4
(1.7 mmol) was treated with 50% TFA in CH2Cl2
(10 ml) under stirring for 1 h at r.t. The reaction mixture
was concentrated under reduced pressure and then co-
evaporated several times with diethyl ether. The residue,
redissolved in THF (10 ml), was taken to pH 9.0 by
dropwise addition of 10% Na2CO3 aqueous solution and
successively treated with 540 mg of FmocOSu
(1.7 mmol). The reaction mixture was left at 0°C for
30 min and at r.t. for 1 h, then filtered and concentrated
under reduced pressure. The crude compound was
purified by flash chromatography, eluting the column
with increasing amounts of AcOEt in petroleum ether.
The fractions eluted with petroleum ether–AcOEt (8:2,
v/v), collected and taken to dryness afforded 310 mg of
pure 5 (0.7 mmol, 41% yield): Rf 0.25 in petroleum
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)

DIPEPTIDE TASTE LIGANDS CONTAINING HOMO-b-AMINO ACID RESIDUES
581
1
ether–AcOEt (8:2, v/v). H NMR (250 MHz, CDCl3), ꢀ
(ppm): 7.80–7.22 (complex system, 13H, aromatic Fmoc
and phenyl protons); 5.48 (d, 1H, NH); 5.12 (s, 2H,
CH2Ph); 4.47 (d, 2H, CH2 Fmoc); 4.35 (t, 1H, CH Fmoc);
4.20 (m, 1H, a CH); 2.75 (d, 4H, overlapped signals,
CH2CN and CH2COOBzl). 13C NMR (62 MHz, CDCl3)
ꢀ (ppm): 169.95 (COOBzl); 155.20 (OCONH); 143.45,
141.19, 134.93, 128.48, 128.25, 128.04, 127.94, 126.98,
124.80 and 119.91 (aromatic Fmoc and phenyl ring
carbons); 116.67 (CN); 66.97 (CH2 Fmoc); 66.90
(CH2Ph); 46.97 (CH Fmoc); 44.90 (a CH); 36.97
(CH2COOBzl); 22.61 (CH2CN). MS (70 eV, EI), m/z:
diastereomers 7a and 7b, which were successively
separated by HPLC on a Baker Chiralcel OD-R analytical
column (250 Â 0.46 mm i.d., 5 mm), eluted with a linear
gradient of CH3CN in H2O (from 5 to 100% CH3CN in
20 min, flow-rate 0.8 ml min 1, ꢁ = 254 nm).
7a: Rf 0.4 in cyclohexane–AcOEt (2:3, v/v); tR
1
20.2 min. H NMR 1D and 2D (400 MHz, CD3OH), ꢀ
(ppm): 8.33 (d, 1H, NH Phe residue); 7.83–7.18 (complex
system, 13H, aromatic Fmoc and phenyl protons); 7.01
(d, H, NH homo-b-Asp); 4.70 (m, 1H, a CH Phe); 4.25
(m, 1H, a CH homo-b-Asp); 4.39–4.20 (complex system,
3H, CH2 and CH Fmoc); 3.67 (s, 3H, OCH3); 3.16 and
2.96 (2 dd, 2H, b CH2 Phe); 2.58–2.40 (complex signals,
‡
440 (M ).
‡
4H, 2 CH2 homo-b-Asp). FAB-MS, m/z: 531 ([M ‡ H] ).
Synthesis of 3-(N-¯uorenylmethoxycarbonylamino)glu-
taric acid (6). A 250 mg amount of 5 (0.57 mmol),
dissolved in dioxane (5 ml), was treated with 37% HCl
(5 ml) and the resulting solution was left at reflux for 4 h,
then at r.t. overnight. The reaction mixture was extracted
with AcOEt and washed exhaustively with water. The
organic phase, dried over anhydrous Na2SO4 and
concentrated in vacuo, was purified by HPLC on a
preparative RP-18 column (Vydac C18, 25 Â 2.2 cm i.d.),
eluted with a linear gradient of CH3CN, containing 0.1%
TFA (solvent B), in H2O, containing 0.1% TFA (solvent
A) (from 20% to 80% B in 30 min, flow-rate 12 ml
min 1, ꢁ = 220 nm). The peak at retention time (tR)
15.2 min, taken to dryness, gave 150 mg of pure 6
(0.41 mmol, 71%): Rf 0.2 in petroleum ether–AcOEt (2:3,
7b: Rf 0.4 in cyclohexane–AcOEt (2:3, v/v); tR
1
21.2 min. H NMR 1D and 2D (400 MHz, CD3OH), ꢀ
(ppm): 8.37 (d, 1H, NH Phe residue); 7.83–7.13 (complex
system, 13H, aromatic Fmoc and phenyl protons); 7.06
(d, 1H, NH homo-b-Asp); 4.68 (m, 1H, a CH Phe); 4.25
(m, 1H, a CH homo-b-Asp); 4.39-4.19 (complex system,
3H, CH2 and CH Fmoc); 3.68 (s, 3H, OCH3); 3.14 and
2.96 (2 dd, 2H, b CH2 Phe); 2.56–2.42 (complex signals,
‡
4H, 2 CH2 homo-b-Asp). FAB-MS, m/z: 531 ([M ‡ H] ).
From the crude 7, the HPLC purification furnished also
a peak at tR 21.6 min, which was collected and identified
as the tripeptide 8 (5 mg, 0.0072 mmol, 12% yield): Rf
0.9 in cyclohexane–AcOEt (2:3, v/v). 1H NMR
(250 MHz, CDCl3), ꢀ (ppm): 8.02 (d, 1H, NH Phe
residue); 7.76–7.19 (complex system, 19H, aromatic
Fmoc and phenyl protons); 6.56 (d, 1H, NH homo-b-Asp
residue); 4.92 (m, 2H, 2 a CH Phe); 4.45–4.16 (m, 4H,
CH2 and CH Fmoc system and a CH homo-b-Asp); 3.96
and 3.92 (2 s, 3H each, 2 OCH3); 3.12–2.35 (complex
system, 6H, 2 CH2 homo-b-Asp and CH2 Phe). FAB-MS,
1
v/v). H NMR (200 MHz, CD3OD), ꢀ (ppm): 7.79–7.25
(complex system, 8H, aromatic Fmoc protons); 4.34–
4.18 (complex system, 4H, overlapped signals, CH2 and
CH Fmoc and a CH); 2.60 (d, 4H, J = 7.0 Hz, 2
CH2COOH). 13C NMR (50 MHz, CD3OD), ꢀ (ppm):
174.84 (2 COOH); 158.21 (OCONH); 145.57, 142.83,
129.04, 128.44, 126.55 and 121.19 (aromatic Fmoc
carbons); 68.09 (CH2 Fmoc); 48.67 (a CH); 46.93 (CH
Fmoc); 39.78 (2 CH2COOH).
‡
‡
m/z: 691 (M ); 692 ([M ‡ H] ). This compound was the
main reaction product when the coupling was performed
using stoichiometric amounts of the reagents.
Synthesis of Fmoc-homo-b-(L,D)-Asp-L-Phe-OMe (7a
and 7b). A 50 mg amount of 6 (0.14 mmol) was
dissolved in anhydrous THF (1 ml) and treated with
70 mg of PyBOP (0.14 mmol) under stirring at r.t. A
solution of 14.6 mg of L-phenylalanine methyl ester
chloridate (0.06 mmol) and DIEA (60 ml, 0.24 mmol) in
anhydrous THF (2 ml) was added to the reaction vessel in
1 h. The resulting mixture was left at r.t. for 2 h, then
concentrated under reduced pressure and purified by
HPLC on a preparative RP-18 column (Vydac C18,
25 Â 2.2 cm i.d.), eluted with a linear gradient of CH3CN
(0.1% TFA) and H2O (0.1% TFA) from 20% to 80% B in
60 min (flow-rate 12 ml min 1, ꢁ = 220 nm). The peak at
tR 17.8 min, taken to dryness, gave 10 mg of pure 7
(0.02 mmol, 31% yield); the peak at tR 15.3 min gave
15 mg of unreacted Fmoc-homo-b-Asp-OH (6).
Synthesis of homo-b-(L,D)-aspartyl-L-phenylalanine
methyl ester (9a and 9b). A 5 mg (0.01 mmol) amount
of 7a was dissolved in CH3CN (2 ml) and treated with
piperidine (2 ml) at r.t. After 1 h the formation of a
compound at Rf 0.15 in CHCl3–CH3OH (9:1, v/v) was
observed, with the concomitant disappearance of the
starting material. The reaction mixture was then con-
centrated under reduced pressure and purified by HPLC
on an analytical RP-18 column (Varian 90A,
250 Â 0.46 mm i.d., 5 mm) eluted with a linear gradient
of CH3CN in H2O (from 5 to 100% CH3CN in 25 min,
flow-rate 0.8 ml min 1, ꢁ = 254 nm). The peak at tR
11.43 min gave the desired compound 9a in almost
1
quantitative yield. H NMR (400 MHz, D2O), ꢀ (ppm):
7.45–7.29 (complex system, 5H, phenyl protons); 4.75
(m, 1H, a CH Phe); 3.79 (s, 3H, OCH3); 3.61 (t, 1H, a CH
homo-b-Asp); 3.30–3.03 (AB part of an ABX system,
2H, CH2 Phe); 2.56 (d, 2H, CH2 backbone homo-b-Asp);
1H NMR data for 7 revealed the presence of two
different sets of signals, in a 1:1 ratio, attributable to
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)

582
C. ISERNIA ET AL.
2.32 (d, 2H, CH2 side-chain homo-b-Asp). FAB-MS, m/z:
and 2.84 (2 dd, 2H, CH2 homo-b-Phg); 2.76 and 2.56 (2
dd, 2H, CH2 Asp); 1.43 (s, 9H, OtBu protons).
‡
309 ([M ‡ H] ).
The same procedure on compound 7b afforded 9b: tR
1
11.51 min. H NMR (400 MHz, D2O), ꢀ (ppm): 7.45–
Synthesis of NH2-L-aspartyl-homo-b-L-phenylglycine
methyl ester (14). A 126 mg amount of 13 (0.26 mmol),
dissolved in CH3OH (5 ml), was hydrogenated over
0.5 mg of 10% Pd on charcoal under stirring at r.t. After
2 h the catalyst was removed, the resulting solution
concentrated under reduced pressure and the residue
treated with TFA (10 ml) at r.t. under stirring for 1 h to
ensure the removal of the tert-butyl group. The mixture
was co-evaporated several times with diethyl ether, taken
to dryness and then lyophilized, affording 100 mg of pure
14 (0.25 mmol, 95% yield): Rf 0.15 in CHCl3–CH3OH
(95:5 v/v); tR 13.3 min on an RP-18 column (LiChrosorb,
LabService, 250 Â 4.6 mm i.d., 5 mm) eluted with a linear
gradient of CH3CN in H2O (from 20 to 80% CH3CN in
7.28 (complex system, 5H, phenyl protons); 4.85 (m, 1H,
a CH Phe); 3.78 (s, 3H, OCH3); 3.69 (m, 1H, a CH homo-
b-Asp); 3.39–3.02 (AB part of an ABX system, 2H, CH2
Phe); 2.54–2.68 (AB part of an ABX system, 2H, CH2
backbone homo-b-Asp); 2.40–2.28 (AB part of an ABX
system, 2H, CH2 side-chain homo-b-Asp). FAB-MS, m/z:
‡
309 ([M ‡ H] ).
Analogously to 7a and 7b, deprotection of tripeptide 8
(5 mg, 0.0072 mmol) was achieved by addition of
piperidine (2 ml) in CH3CN (2 ml) at r.t. After 1 h the
formation of a compound at Rf 0.35 in CHCl3–CH3OH
(9:1, v/v) was observed, with the concomitant disap-
pearance of the starting material. The reaction mixture
was then concentrated under reduced pressure and
purified by HPLC on a preparative RP-18 column (Vydac
C18, 25 Â 2.2 cm i.d.), eluted with a linear gradient of
CH3CN, containing 0.1% TFA, in H2O, containing 0.1%
TFA (from 20% to 80% B in 60 min, flow-rate 12 ml
min 1, ꢁ = 220 nm). The peak at tR 41.7 min gave the
1
30 min, flow-rate 1.0 ml min 1, ꢁ = 220 nm). H NMR
1D and 2D (400 MHz, CD3OH), ꢀ (ppm): 8.86 (d, 1H,
NH homo-b-Phg); 7.41–7.26 (complex system, 5H,
phenyl protons); 5.38 (apparent q, 1H, a CH homo-b-
Phg); 4.16 (apparent q, 1H, a CH Asp); 3.67 (s, 3H,
OCH3); 2.98 (dd, 1H, Hl of CH2 Asp); 2.88 (d, 2H, CH2
homo-b-Phg residue); 2.86 (dd, 1H, Hh of CH2 Asp).
1
desired compound 10 in almost quantitative yield. H
NMR (400 MHz, CD3OH), ꢀ (ppm): 8.44 and 8.41 (two
ds, 2H, 2 NH Phe residues); 7.96 (bs, 1H, NH homo-b-
Asp residue); 7.33–7.19 (complex system, 10H, phenyl
protons), 4.73 (m, 2H, a CH Phe); 3.74 (s, 6H, 2 OCH3);
3.72 (m, 1H, a CH homo-b-Asp); 3.18 and 2.94 (AB parts
of two ABX systems, 4H, 2 CH2 Phe); 2.60–2.44
(complex signals, 4H, side-chain and backbone CH2
Synthesis of the dipeptide NH2-L-Asp-homo-b-L-Phe-
OMe (19). Synthesis of homo-b-L-phenylalanine methyl
ester (17). The title compound7 was obtained in
quantitative yield by treating with an ethereal diazo-
methane solution homo-b-phenylalanine (16) dissolved
in CH3OH, which had previously been obtained in 77%
yield by acidic hydrolysis with 37% HCl of 15, prepared
according to a literature procedure.7
‡
homo-b-Asp). FAB-MS, m/z: 470 ([M ‡ H] ).
Synthesis of the dipeptide NH2-L-Asp-homo-b-L-
Phg-OMe (14). Synthesis of Cbz-L-Asp(OtBu)-homo-
b-L-Phg-OMe (13). The dipeptide ester 13 was prepared
from 1.0 g of Cbz-L-Asp(OtBu)-OH (11, 3.1 mmol) and
0.56 g of NH2-homo-b-L-Phg-OMe (12, 3.1 mmol),
prepared according to a literature procedure,7 in anhy-
drous THF (10 ml) in the presence of 1.6 g of PyBop
(3.1 mmol) and DIEA (3.0 ml, 12 mmol). After 3 h under
stirring at r.t., the reaction mixture was taken to dryness,
redissolved in AcOEt and washed several times with 10%
citric acid, saturated NaHCO3 and water. The organic
phase was dried over anhydrous Na2SO4 and concen-
trated to give 1.2 g of pure 13 (2.48 mmol, 80% yield): Rf
0.65 in CHCl3–CH3OH (95:5); tR 19.1 min on an RP-18
column (LiChrosorb, LabService, 250 Â 4.6 mm i.d.,
5 mm) eluted with a linear gradient of CH3CN in H2O
(from 20 to 80% CH3CN in 30 min, flow-rate 1.0 ml
Synthesis of Boc-L-Asp(OtBu)-homo-b-L-Phe-OMe (18).
A 147 mg amount (0.51 mmol) of Boc-L-Asp(OtBu)-OH
and 267 mg (0.51 mmol) of PyBop, dissolved in CH2Cl2
(5 ml), were treated with 117 mg (0.51 mmol) of 17 and
500 ml (2.0 mmol) of DIEA under stirring at r.t. After 2 h
the reaction mixture was concentrated under reduced
pressure and worked up as described for 13. The crude
mixture was purified by HPLC on a preparative RP-18
column (Vydac C18, 25 Â 2.2 cm i.d.), eluted with a
linear gradient of CH3CN, containing 0.1% TFA, in H2O,
containing 0.1% TFA (from 20% to 80% B in 30 min,
flow-rate 12 ml min 1, ꢁ = 220 nm). The peak at tR
29.2 min, taken to dryness, gave 108 mg (0.23 mmol,
47% yield) of pure 18; Rf 0.75 in CHCl3–CH3OH (98:2).
1H NMR 1D and 2D (400 MHz, CD3OH), ꢀ (ppm): 7.93
(d, 1H, NH homo-b-Phe); 7.33–7.17 (complex system,
5H, phenyl protons); 6.83 (d, 1H, NH Asp); 4.37 (m, 1H,
a CH Asp); 4.42 (m, 1H, a CH homo-b-Phe); 3.65 (s, 3H,
OCH3); 2.81 and 2.87 (2 dd, 2H, b CH2 side-chain Phe);
2.62 (dd, 1H, Hl of CH2 Asp); 2.55 and 2.47 (2 dd, 2H,
CH2 backbone homo-b-Phe); 2.42 (dd, 1H, Hh of CH2
Asp); 1.46 and 1.45 (two s, 18H, Boc and OtBu protons).
1
min 1, ꢁ = 220 nm). H NMR 1D and 2D (400 MHz,
CD3OH), ꢀ (ppm): 8.51 (d, 1H, NH homo-b-Phg); 7.40
(d, 1H, NH Asp); 7.38–7.24 (complex system, 10H,
aromatic phenyl protons); 5.34 (m, 1H, a CH homo-b-
Phg); 5.13 and 5.09 (AB part of an ABX system, 2H, CH2
Cbz); 4.54 (m, 1H, a CH Asp); 3.63 (s, 3H, OCH3); 2.89
‡
FAB-MS, m/z: 465 ([M ‡ H] ).
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)

DIPEPTIDE TASTE LIGANDS CONTAINING HOMO-b-AMINO ACID RESIDUES
583
Figure 1. Circular dichroism spectra of peptides 9a (^), 14 (ÐÐ) and 19 (-----). The aspartame CD spectrum is shown for
comparison (inset)
Synthesis of NH2-L-aspartyl-homo-b-L-phenylalanine
methyl ester (19). A 50 mg amount (0.107 mmol) of 18
was treated with 50% TFA in CH2Cl2 (5 ml) at r.t. under
stirring. After 2 h the solution was concentrated under
reduced pressure and then co-evaporated several times
with diethyl ether, giving 45 mg (0.102 mmol, 95% yield)
of pure 19, as the trifluoroacetate salt: Rf 0.1 in CHCl3–
CH3OH (9:1, v/v). 1H NMR 1D and 2D (400 MHz,
CD3OH), ꢀ (ppm): 8.37 (d, 1H, NH homo-b-Phe); 7.35–
7.20 (complex system, 5H, phenyl protons); 4.45 (m, 1H,
a CH homo-b-Phe); 4.05 (apparent q, 1H, a CH Asp);
3.66 (s, 3H, OCH3); 2.91 (dd, 1H, Hl of CH2 side-chain
Phe); 2.82–2.89 (partially overlapped signals, 2 dd, 2H,
Hh of CH2 side-chain Phe and Hl of CH2 Asp); 2.76 (dd,
1H, Hh of CH2 Asp); 2.60 and 2.47 (2 dd, 2H, CH2
formation was achieved by x-ray diffractometry, the CD
band between 210 and 230 nm was interpreted as a result
of the superposition of a band distinctive of the spectra of
homo-b-amino acids containing peptides and a positive
contribution of the aromatic phenyl side-chain.26
The band near 195 nm is a characteristic feature of the
spectra of peptides containing a phenyl residue in the
side-chain.26
The major differences in the spectra may be explained
as a consequence of the presence of the following two
alternative arrays of chromophores:
‡
backbone homo-b-Phe). FAB MS, m/z: 309 ([M ‡ H] ).
RESULTS AND DISCUSSION
Circular dichroism
In order to investigate the structure–activity relationships
(SAR) of the dipeptide taste ligands with respect to the
natural dipeptide aspartame, a solution conformational
analysis was carried out by means of CD and NMR
techniques. For CD experiments all peptides were
dissolved in a 1:1 (v/v) TFE–H2O mixture to minimize
differences in solubility between the protected and the
deprotected forms and to increase possible conforma-
tional differences.
The CD spectra of dipeptides 9a, 14 and 19 are shown
in Fig. 1. Gaussian deconvolution shows that all the
spectra may be regarded as linear combinations of a
medium-intensity CD band with a maximum at about
195 nm and a second positive band with a maximum
between 210 and 220 nm. By comparison with the CD
spectra of aliphatic homo-b-amino acids containing
peptides (unpublished data), whose unstructured con-
The intrinsic chromophores are the same, so that strong
changes in intensity between the corresponding CD
spectra should arise from a difference in the number of
conformers which contribute to the CD. A less intense
signal in the peptides containing the chromophore II
corresponds to a wider set of conformers, whose CD
contributions tend to cancel each other. This is not
unexpected, since the CH2 interposed between the phenyl
and the backbone a-carbon in chromophore II contributes
to an overall increased flexibility of the corresponding
side-chain.
Peptide 14, showing the more intense CD spectrum,
presents a type I chromophore. It almost mantains the
aspartame sweetness even if larger differences in its CD
spectrum are seen with respect to that of aspartame.
In contrast, all type II peptides share a CD spectrum
very reminescent of that of natural aspartame; however,
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)

584
C. ISERNIA ET AL.
Table 1. 1H chemical shifts (ppm) of the dipeptides in DMSO-d6 at 298 K
Compound
AA
NH
aH
bH
hbH
2.79
2.80
2.41
Others
13
Asp
Phg
7.53
8.46
4.38
5.20
2.60/2.42
CH2(Cbz) 5.01; 2,6H Cbz 7.33; 3,4,5H Cbz 7.30; Boc 1.37;
2,6H Phg 7.31; 3,4,5H Phg 7.30; OMe 3.54
14
18
19
Asp
Phg
8.13
8.95
4.05
5.23
2.84/2.72
OMe 3.57
Asp
Phe
6.93
7.80
4.21
4.21
2.45/2.31
2.77/2.70
Boc/OtBu 1.38
OMe 3.54; 2,4,6H 7.19; 3,5H 7.27
Asp
Phe
n.d.
8.46
3.91
4.26
2.70/2.62
2.81/2.76
2.41/2.51 OMe 3.56; 2,4,6H 7.22; 3,5H 7.31
3
3
Table 2. JNH-aCH (Hz), JaH-bCH (Hz) and temperature
coef®cients, Dꢀ/DT (ppb K 1), for the dipeptides in DMSO
and in CD3OH (in parentheses) at 298 K
Table 4. Side-chain populations (%) in DMSO and in CD3OH
(in parentheses)
Amino acid
Asp
13
14
18
19
Parameter
3JNH-aCH (Asp)
13
14
18
19
PI (g )
PII (t)
62
(55)
20
(26)
18
(19)
51
(56)
33
(14)
16
(30)
61
(56)
21
(24)
18
(20)
52
(57)
11
(15)
37
(28)
8.5
(8.0)
4.8
(5.4) (4.2)
9.4 8.2
(8.6) (8.7)
8.5 8.3
(8.3) (8.6)
—
(—)
—
(—)
7.4
(7.9) (7.6)
7.4
(6.1) (7.6)
5.4 1.5
( 8.0) (—) ( 8.4)
6.0 3.7 5.2
( 8.5) ( 6.0) ( 7.3)
Broad
(—)
6.2
8.6
(7.8)
4.9
(5.2)
9.3
(8.7)
8.5
—
(—)
3.8
(4.2)
8.4
JaCH-bCHpro-S (Asp)
3JaCH-bCHpro-R (Asp)
3JNH-aCH (Phe or Phg)
3JaCH-bCH (Phe)
PIII (g‡)
(8.8)
8.2
Phe
PI (g )
PII (t)
—
—
—
—
—
—
35
(41)
42
(43)
23
37
(37)
40
(46)
23
(8.6)
7.2
(8.2)
—
(—)
—
(—)
—
7.0pro-S
(7.3) (7.6)pro-S
3JaCH-b'CH (Phe)
6.4
(7.1) (6.6)pro-R
5.7
5.4pro-S
(5.5) (5.0)pro-S
7.1
8.0pro-R
(7.2) (8.4)pro-R
6.6pro-R
PIII (g‡)
(16)
(17)
3JaCH-hbCH (Phe or Phg)
3JaCH-hb'CH (Phe or Phg)
Dꢀ/DT (Asp)
hbCH2
PI (g )
PII (t)
49
(53)
25
(22)
26
—
6.3
—
(—)
4.1
PIII (g‡)
(25)
Dꢀ/DT (Phe or Phg)
( 5.5)
CD3OH solutions, showed ROESY spectra with very
little significant structural information (data not shown)
and revealed an enhanced flexibility in solution, probably
because of the extra CH2 group in the homo-b-Asp
residue. In addition, the observed NOEs cannot be
explained in terms of one preferential conformation, but
in terms of different conformations in rapid exchange.
The 1H chemical shift assignments in DMSO at
298 K for 13, 14, 18 and 19 are reported in Table 1.
replacement of the L-Asp residue with homo-b-Asp
causes a loss of taste.
NMR
The conformation of the dipeptides was investigated by
1H NMR spectroscopy in CD3OH and DMSO-d6
solutions at 298 K.
3
3
The JNH-a CH and JaH-bCH coupling constants and the
temperature coefficients are listed in Table 2. Tables 3
and 4 contain the calculated torsion angles f and the side-
Dipeptides 7a, 7b, 9a and 9b, studied by high-
resolution 1H NMR spectroscopy in DMSO-d6 and
3
Table 3. Torsion angles f in DMSO and in CD3OH (in parentheses) determined by JNH-aCH coupling
constants
Amino acid
Asp
13
14
18
19
154, 91, 40–80
( 157, 86, 37–83)
—
(—)
154, 91, 40–80
( 157, 84, 43, 80)
—
(—)
Phe or Phg
154, 91, 40–80
154, 89, 40–80
154, 91, 40–80
157, 86, 37–83
( 154, 89, 40–80) ( 154, 91, 40–80) ( 154, 91, 40–80) ( 157, 86, 37–83)
Copyright 1999 John Wiley & Sons, Ltd. J. Phys. Org. Chem. 12, 577–587 (1999)

DIPEPTIDE TASTE LIGANDS CONTAINING HOMO-b-AMINO ACID RESIDUES
585
Figure 2. Relevant NOE correlations (strong = solid line,
medium = dashed line, weak = dotted line) for Cbz-NH-L-
Asp(OtBu)-homo-b-L-Phg-OMe recorded in DMSO at 298 K
Figure 4. Relevant NOE correlations (strong = solid line,
medium = dashed line, weak = dotted line) for NH2-L-Asp-
homo-b-L-Phg-OMe recorded in DMSO at 298 K
3
chain populations. Chemical shifts and temperature
coefficients were shown to be independent on the peptide
concentration in the range 0.5–22 mM, suggesting that
there is no aggregation at the concentrations used for the
2D analysis.
The NOEs were obtained from the ROESY experi-
ments and were classified according to their relative
intensities as strong, medium and weak. The main NOE
correlations are summarized in Figs 2–5. As expected,
coupling constants, temperature coefficients and NOE
data, observed for all molecules, do not indicate the
existence of a single preferred conformation in solution.
For all compounds a strong inter-residue NOE was
observed between the a-proton of the Asp and the NH
proton of the following residue, indicating that the torsion
angle (Asp) is restricted to values between 60° and
180°.
chain of Asp residue was carried out using the Jab1
3
and Jabh experimental values and following the treat-
ment suggested by Jardetzky and Roberts;28 in each case
the g conformation is the preferred structure. The
populations of peptide 19 side-chains were calculated
assuming the Hbh as pro-R and the Hb1 as pro-S protons
in analogy with peptides 13, 14 and 18, since NOE
correlations NH–Hb1 and NH–Hbh could not be
evaluated. A similar treatment for the Phe side-chain in
18 indicates a preference for the trans orientation. Using
the same procedure we determined the homo-b—CH2
backbone populations for 19 and a preference for the g
orientation was found.
The torsion angles f for the Phe residues was
evaluated from the JNH-aCH coupling constants, using a
3
Karplus-type relationship reported by Bystrov;29 all
compounds have similar values which are consistent
with the medium NOEs NH Phe–Ha Phe observed.
Medium and weak NOE connectivities, reported in
Figs 2–5, were observed for all compounds; the C-
terminal part of the molecules appears less defined since
no correlation with the methyl ester group was detected.
An important difference between compounds 14 and 19
In 13, 14 and 18, the prochiralities of the two b-protons
of the Asp residue were assigned27 using the coupling
constants 3Jab and the NOE connectivities Asp NH–Asp
Hb1 and Asp NH–Asp Hbh. For each compound the
proton at higher field (Hbh) is pro-R and that at lower
field (Hb1) is pro-S. Analysis of populations for the side-
Figure 3. Relevant NOE correlations (strong = solid line,
medium = dashed line, weak = dotted line) for Boc-NH-L-
Asp(OtBu)-homo-b-L-Phe-OMe recorded in DMSO at 298 K
Figure 5. Relevant NOE correlations (strong = solid line,
medium = dashed line, weak = dotted line) for NH2-L-Asp-
homo-b-L-Phe-OMe recorded in DMSO at 298 K
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)

586
C. ISERNIA ET AL.
and compounds 13 and 18 was found. For the deprotected
dipeptides (14 and 19) the NOEs NH Phe–Hb1 Asp and
NH Phe–Hbh Asp have the same medium intensities,
whereas in the protected compounds the NH Phe–HbproR
Asp effect was found to be weaker than that for NH Phe–
HbproS Asp. An accurate analysis was conducted also in
CD3OH and the related NMR conformational parameters
are reported in Tables 2–4. The structural data obtained
are similar to those reported in DMSO.
shaped’ structure. The conformational parameters for the
homo-b-L-Phe residue in 19 shows a similar behaviour
with dihedral angles f, m and of 134°, 179° and 90°
for the extended conformation and 134°, 95° and
112° or 134°, 95° and 112° for the two equi-energetic
extended and ‘L-shaped’ conformations and preferred
trans conformation (w1) for the Phe residue side-chain for
both the extended and ‘L-shaped’ conformers.
CONCLUSION
Molecular dynamics
The conformational effect due to the insertion of a
methylene group in the backbone skeleton of the
dipeptide taste ligand aspartame has been investigated.
The solution conformation of the newly synthesized
dipeptide taste ligands was investigated by 1H NMR and
CD spectroscopy.
The NMR spectra were fully assigned by a combina-
tion of TOCSY, DQFCOSY and ROESY maps. Chemi-
cal shifts and temperature coefficients were shown to be
independent of the peptide concentration in the range 0.5
–22 mM. For the dipeptides NH2-homo-b-(L or D)Asp-L-
Phe-OMe (9a and 9b), the NMR data were poor with
little structural information, revealing an enhanced
flexibility in solution certainly due to the extra CH2
moiety of the homo-b-Asp residue. Thus only for the
dipeptides NH2-L-Asp-homo-b-L-Phg-OMe (14) and
NH2-L-Asp-homo-b-L-Phe-OMe (19) were computa-
tional analyses possible.
The CD spectra are consistent with the NMR results,
indicating great flexibility for peptides 9a, 9b and 19 and
a reduced flexibility for peptide 14. Moreover, the
populations of the different families of conformers are
the same on comparing analogues 9a, 9b and 19 with
aspartame, even if these new dipeptide taste ligand
compounds differ in their sweetness potency with respect
to aspartame.
The analysis of the conformational behaviour of the
synthesized dipeptides shows that all of the compounds
can adopt preferentially extended and ‘L-shaped’ con-
formations with different relative populations. Several
other conformations are also possible but with smaller
relative populations.
The relationship between conformation and sweet taste
underlines that the lengthening of the peptide skeleton at
the L-Asp site results in a loss of sweetness with the
production of tasteless compounds, thus confirming that
(i) the C-terminal end is the more important function in
imparting the sweet taste to the molecule and (ii) the
orientation of the amide group linking the Asp and Phe
residue of aspartame analogues plays a decisive role in
obtaining very sweet compounds. On the other hand,
lengthening of the skeleton at the C-terminal L-Phe site
maintains the sweet character in both NH2-L-Asp-homo-
b-L-Phe-OMe and NH2-L-Asp-homo-b-L-Phg-OMe. The
decrease in sweetness potency in these compounds with
All the structurally significant NMR parameters were
used for RMD simulations in vacuo for the various
compounds.† The conformational statistical analysis of
the RMD simulations at 300 K reveals that a preferred
conformation was found for 13. The average structure of
peptide 13 appears consistent with all the observable
NMR NOE effects. The conformational analysis under-
lines that the preferred conformation of these dipeptide
taste ligands can be described with an ‘L-shape.’ This
class of conformers also corresponds to a minimum
energy conformation. In particular, the aspartyl residue
has dihedral angles f and of 111.1° and 134.4°, and a
g
preferred side-chain conformation (w1) for the Asp
residue. Similar behaviour has also been observed in
crystal structures of aspartame and its analogues. The
analysis of the homo-b-L-Phg residue shows dihedral
angles f, m and of 97.5°, 71.2° and 86.4°,
respectively.
The conformation analysis of RMD simulation for 18
shows that it can adopt mainly extended and ‘L-shaped’
conformations, among others, with similar energy. The
‘L-shaped’ conformer presents dihedral angles of the
aspartyl residue of f, and w1 of 128°, 103° and 58°,
respectively, similar to those observed in several crystal
structures of dipeptide taste ligand analogues of aspar-
tame. The analysis of the homo-b-L-Phe residue shows
dihedral angles f, m and of 120.3°, 80.2° and 82.1°,
respectively.
The analysis of the RMD results for 14 and 19
indicates that they can adopt mainly extended and ‘L-
shaped’ conformations, among others, and that these
conformations have comparable energies. For both
compounds, the ‘L-shaped’ conformer presents average
values of dihedral angles and w1 (147° and 70°,
respectively) of the aspartyl residue similar to those
found in other dipeptide taste ligands. The analysis of the
homo-b-L-Phg residue in 14 shows dihedral angles f, m
and of 137°, 173° and 73° for the extended
conformation and 145°, 45° and 70° for the ‘L-
†The figures of the superimposition obtained for the minimized
structures of compounds 13, 14, 18 and 19 during the simulation are
deposited as Supplementary Material on the EPOC website
(http://www.wiley.com/epoc).
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)

DIPEPTIDE TASTE LIGANDS CONTAINING HOMO-b-AMINO ACID RESIDUES
587
8. E. Benedetti, R. Iacovino, C. Pedone, F. Rossi, M. Saviano, C.
Isernia, D. Montesarchio, L. de Napoli and G. Piccialli, Letters in
Peptide Science 4, 129–134 (1997).
9. R. H. Mazur, J. M. Schlatter and A. H. Goldkamp, J. Am. Chem.
Soc. 91, 2684–2691 (1969).
10. D. J. Ager, D. P. Pantaleone, S. A. Henderson, A. R. Katritzky, I.
Prakash and D. E. Walters, Angew. Chem. Int. Ed. Engl. 37, 1802–
1817 (1998).
11. R.-H. Mattern, Y. Amino, E. Benedetti and M. Goodman, J. Pept.
Res. 50, 286–299 (1997).
12. M. Goodman, T. Yamazaki, Y. F. Zhu, E. Benedetti and R. K.
Chadha, J. Am. Chem. Soc. 115, 428–432 (1993).
13. M. Miyoshi, K. Nunami, H. Sugano and T. Fuji, Bull. Chem. Soc.
Jpn. 51, 1433–1440 (1978).
respect to aspartame can be ascribed to an increase of the
chain flexibility due to the methylene insertion, where the
C-terminal ester group is almost comparable in size with
the side-chain aromatic ring. This would make more
likely the exchange in the position of either group in the
interaction with the receptor. In the light of these results,
it seems reasonable to confirm that the sweet power of
aspartame analogues can only be modulated by modify-
ing the C-terminal moiety of the molecule.
14. U. Piantini, O. W. Sørensen and R. R. Ernst, J. Am. Chem. Soc.
104, 6800–6801 (1982).
Acknowledgements
15. L. Braunschweiler and R. R. Ernst, J. Magn. Reson. 53, 521–528
(1983).
16. A. A. Bothner-By, R. L. Stephens, J. Lee, C. D. Warren and R. W.
Jeanloz, J. Am. Chem. Soc. 106, 811–813 (1984).
17. C. Griesinger and R. R. Ernst, J. Magn. Reson. 75, 261–271
(1987).
18. Insight/Discover Package. Biosym Technologies, San Diego, CA.
19. S. Lifson, A. T. Hagler and P. Dauber, J. Am. Chem. Soc. 101,
5111–5121 (1979).
The authors gratefully acknowledge Professor Murray
Goodman for his useful suggestions and for critical
reading of the manuscript. They thank Dr Gabriella de
Vita and Miss Rita Carolla for competent technical
assistance.
20. A. T. Hagler, S. Lifson and P. Dauber, J. Am. Chem. Soc. 101,
5122–5130 (1979).
21. A. T. Hagler, P. Dauber and S. Lifson, J. Am. Chem. Soc. 101,
5131–5141 (1979).
22. L. Verlet, Phys. Rev. 159, 98–103 (1967).
23. C. L. Brooks, III, B. M. Pettitt and M. Karplus, Proteins: a
Theoretical Perspective of Dynamics, Structure and Thermo-
dynamics. J Wiley, New York (1988).
24. K. Pucinska and B. Liberak, Tetrahedron 43, 3509–3517 (1987).
25. J. Podlech and D. Seebach, Liebigs Ann. Chem. 1217–1228 (1995).
26. S. Fermandjian, J. L. Morgat and P. Fromageot, Eur. J. Biochem.
24, 252–258 (1971).
27. G. Wagner, W. Braun, T. F. Havel, T. Schaumann, N. Go and K.
Wu¨thrich, J. Mol. Biol. 196, 611–639 (1987).
28. O. Jardetzky and G. C. K. Roberts, NMR in Molecular Biology.
Academic Press, New York (1981).
29. V. F. Bystrov, Prog. Nucl. Magn. Reson. Spectrosc. 10, 41–81
(1976).
REFERENCES
1. K. S. Chu, G. R. Negrete and J. P. Konopelski, J. Org. Chem. 56,
5196–5202 (1991).
2. H. Sone, T. Nemoto, H. Ishiwata, M. Ojika and K. Yamada,
Tetrahedron Lett. 34, 8449–8452 (1993).
3. H. Morita, S. Nagashima, K. Takeya, H. Itokawa and Y. Iitaka,
Tetrahedron 51, 1121–1132 (1995).
4. A. F. Spatola, in Chemistry and Biochemistry of Amino Acids,
Peptides and Proteins, edited by B. Weinstein, Vol. VII,
pp. 331–333. Marcel Dekker, New York (1983).
5. X. Daura, W. F. van Gusteren, D. Rigo, B. Jaun and D. Seebach,
Chem. Eur. J. 3, 1410–1417 (1997).
6. J. L. Matthews, M. Overhand, F. N. M. Kuhnle, P. E. Ciceri and D.
Seebach, Liebigs Ann. Chem. 1371–1378 (1997).
7. R. Caputo, E. Cassano, L. Longobardo and G. Palumbo,
Tetrahedron 51, 12337–12350 (1995).
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 577–587 (1999)