Conformation of different S-deoxo-Xaa3-amaninamide analogues in DMSO solution as determined by NMR spectroscopy. Strong CD effects induced by βI, βII conformational change

Article abstract of DOI:10.1021/ja9529706

The amatoxines form a class of naturally occurring bicyclic octapeptides responsible for the poisonous effect of Amanita mushrooms. For the investigation of structure-activity relationships many derivatives of these peptides have been synthesized and test

Full text of DOI:10.1021/ja9529706

4380
J. Am. Chem. Soc. 1996, 118, 4380-4387
Conformation of Different S-Deoxo-Xaa3-amaninamide
Analogues in DMSO Solution as Determined by NMR
Spectroscopy. Strong CD Effects Induced by âI, âII
Conformational Change
Wolfgang Schmitt,† Giancarlo Zanotti,‡ Theodor Wieland,§, and Horst Kessler*,†
Contribution from the Institut fu¨r Organische Chemie und Biochemie, Technische UniVersita¨t
Mu¨nchen, Germany, Centro di Chimica del Farmaco del CNR, Dipartimento di Studi
Farmaceutici, UniVersita` ‘La Sapienza’, Roma, Italy, and Max-Planck Institut fu¨r medizinische
Forschung, Heidelberg, Germany
ReceiVed August 28, 1995X
Abstract: The amatoxines form a class of naturally occurring bicyclic octapeptides responsible for the poisonous
effect of Amanita mushrooms. For the investigation of structure-activity relationships many derivatives of these
peptides have been synthesized and tested up to now. To examine the particular role of the configuration of the
amino acid in position 3, three new analogues of S-deoxo-amaninamide have been synthesized in which residue 3
is replaced by L-Ala, D-Ala, or Gly. CD spectra in methanol show large deviations which indicate conformational
differences or isomeric bridging during synthesis. To discriminate both possibilities a detailed NMR analysis was
performed. The conformation of these compounds has been determined in DMSO solution by highfield NMR
spectroscopy, distance geometry, and molecular dynamic calculations. The resulting structures are surprisingly similar
and closely related to the x-ray structure of â-amanitin: Conformational differences are only found in the turn region
of Hyp2-Xaa3. Hence, the strong differences of the CD spectra rise only from type I/type II variations of this â-turn.
Introduction
polymerase, DNA template, and the nascent mRNA is formed.7,8
After the first phosphodiester bond is built the translocation of
the polymerase along the DNA template, and therefore the
synthesis of mRNA and so protein synthesis is blocked.2,8
To date nine different naturally occurring peptides of the
amatoxin family have been isolated;1-3 their chemical structures
are shown in Table 1. They all exhibit the same bicyclic
framework which is important for their activity (see Figure 1).
To investigate structure-activity relationships on the inhibi-
tory action of these toxins more then 40 different analogues
have been synthesized.1-3,9 One result is the importance of the
chemical nature of the side chain in position 3.2,3,9,10 Therefore,
three new analogues of S-deoxo-amaninamide containing either
Ala, D-Ala, or Gly instead of residue 3 have been synthesized
to evaluate the influence of the chirality at this position.
The CD spectra of these three analogues differ dramatically
in the region of 210-260 nm (see Figure 2). The L-Ala3 analog
(1) shows negative values in this region, and the curve of the
D-Ala3 analog (2) is always positive, while the spectrum of the
Gly3 analog (3) lies in-between. This was interpreted as an
indication of conformational differences between these com-
pounds in solution. For this class of bicyclic peptides theoreti-
cally two different diastereomers are possible: a normal form
and the related iso-form (see Figure 3). The sulfide bridge can
be formed either above (normal) or below the cyclic peptide
backbone. In comparison with the CD spectra of S-deoxo-D-
Ile3-amaninamide (number 4 in ref 10), where both of these
isomers could be isolated, the spectrum of 1 resembles that of
the normal form of 4, while 2 shows a similar curvature as iso-
Most of the deadly mushroom intoxications are caused by
ingestion of Amanita phalloides or Amanita Virosa. These
mushrooms contain three different classes of toxins.1,2 Phal-
lotoxins (bicyclic heptapeptides) and virotoxins (monocylic
heptapeptides) are quick acting substances which cause death
of white mice in 2-4 h after parenteral administration.3,4 They
act on liver cells where they bind strongly to F-actin.2,5 This
results in a weakening of the plasma membrane and leads to
the appearance of vacuoles filled with blood. As a consequence
the liver swells and the amount of the circulating blood is thus
reduced by 60-65%, and the organism dies due to haemody-
namic shock. A third class of toxic peptides are the slowly
acting amatoxins (bicyclic octapeptides) which exhibit LD50
values up to ten times lower, depending on the animal species,
than the other two kinds of toxins mentioned above. The
amatoxins are the peptides responsible for the intoxicating effect
after ingestion of A. phalloides. They lead to death after 4-8
days.3 Amatoxins bind very selectively to the 140 kDa subunit
(SB3) of DNA-dependent RNA polymerase B in nuclei of all
eukariontic cells.6 A very stable ternary complex of toxin,
† Technische Universita¨t Mu¨nchen.
‡ Universita` ‘La Sapienza’.
§ Max-Planck Institut fu¨r medizinische Forschung.
Deceased November 24, 1995.
X Abstract published in AdVance ACS Abstracts, April 15, 1996.
(1) Wieland, Th. Peptides of the Poisonous Amanita Mushrooms;
Springer Verlag, New York, 1986.
(2) Wieland, Th. Int. J. Peptide Protein Res. 1983, 22, 257-276.
(3) Wieland, Th.; Faulstich, H. CRC Crit. ReV. Biochem. 1978, 5, 185-
260.
(7) Cochet-Meilhac, M.; Chambon, P. Biochim. Biophys. Acta 1974, 353,
160-184.
(4) Faulstich, H.; Baku, A.; Bodenmu¨ller, H.; Wieland, Th. Biochemistry
1980, 19, 3334-3343.
(8) Vaisius, A. C.; Wieland, Th. Biochemistry 1982, 21, 3097-3101.
(9) Wieland, Th.; Go¨tzendorfer, C.; Zanotti, G.; Vaisius, A. C. Eur. J.
Biochem. 1981, 117, 161-164.
(5) Faulstich, H.; Scha¨fer, A. J.; Weckauf, M. Hoppe-Seyler’s Z. Physiol.
Chem. 1977, 358, 181-184.
(10) Zanotti, G.; Petersen, G.; Wieland, Th. Int. J. Peptide Protein Res.
1992, 40, 551-558.
(6) Brodner, O. G.; Wieland, Th. Biochemistry 1976, 15, 3480-3484.
S0002-7863(95)02970-2 CCC: $12.00 © 1996 American Chemical Society

S-Deoxo-Xaa3-amaninamide Analogues
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4381
Table 1. Naturally Occurring and Synthetic Amatoxins
compound
R1
R2
R3
R4
X
R-amanitin
NH2
OH
NH2
OH
OH
OH
OH
OH
OH
OH
OH
OH
H
CH(CH3)CH(OH)CH2OH
CH(CH3)CH(OH)CH2OH
CH(CH3)CH(OH)CH3
CH(CH3)CH(OH)CH3
CH(CH3)CH(OH)CH2OH
CH(CH3)CH(OH)CH2OH
CH(CH3)CH2CH3
CH(CH3)CH2CH3
CH(CH3)CH2CH3
CH3 (l)
OH
OH
OH
OH
H
SO
SO
SO
SO
SO
SO
SO
SO
SO
S
â-amanitin
γ-amanitin
ꢀ-amanitin
amanin
amaninamide
amanullin
amanullic acid
proamanullin
OH
NH2
NH2
OH
NH2
NH2
NH2
NH2
H
OH
OH
OH
OH
OH
OH
S-deoxo-Ala3-amaninamide 1
S-deoxo-D-Ala3-amaninamide 2
S-deoxo-Gly3-amaninamide 3
OH
OH
OH
CH3 (d)
H
S
S
Figure 3. Schematic drawing of the two possible isomers for
amatoxins.
Figure 1. General structure of the amatoxin-peptides; the different
substituents are listed in Table 1.
DMSO, dimethyl sulfoxide; Hpi, L-3a-hydroxy-1,2,3,3a,8,8a-hexahy-
dropyrrolo-(2,3-b)-indole-2-carboxylic acid; HOBt, hydroxybenzotria-
zole; MA, mixed anhydride; NMM, N-methylmorpholine; TLC, thin
layer chromatography; TFA, trifluoroacetic acid; THF, tetrahydrofurane.
Ddz-Hyp-D-Ala-OtBu (2.1). To the solution of 1.28 g (5 mmol)
of Ddz-Hyp-OH in 50 mL of CH2Cl2 were added at -10 °C under
stirring 0.68 g (5 mmol) of iBCCl and 0.68 g (6.7 mmol) of NMM.
After 10 min the solution of 0.725 g (5 mmol) of H-D-Ala-OtBu in 20
mL of CH2Cl2 was added, and the reaction mixture was kept 3 h at
room temperature, washed with aqueous NaHCO3 solution, 0.5 M
KHSO4, and water, dried over Na2SO4, and evaporated to give 2.15 g
(89% yield) of pure title compound. Rf ) 0.7 in ethyl acetate.
Ddz-Asn-Hyp-D-Ala-OtBu (2.2). The solution of 2.15 g (4.48
mmol) of Ddz-Hyp-D-Ala-OtBu (2.1) in a mixture of 3.42 mL of TFA
and 150 mL of CH2Cl2 was stirred at room temperature for 20 min
then extracted with aqueous NaHCO3 and water and dried over Na2-
SO4. This solution was mixed with a solution of 1.58 g (4.48 mmol)
of Ddz-Asn-OH and 1.20 g (8.96 mmol) of HOBt in 20 mL of DMF,
cooled to 0 °C, and combined with a solution of 0.920 g (4.48 mmol)
of DCCI in 20 mL of DMF. After stirring for 1 h at 0 °C and overnight
at room temperature, the solvent was evaporated in vacuo, and the
residue was dissolved in ethyl acetate. The solution, cleared from
dicyclohexylurea and worked up as described in the previous paragraph,
gave a residue which was chromatographed on a Sephadex LH 20
column (5 cm × 250 cm) in methanol as eluent, yielding 0.82 g (30%)
of 2.2. Rf ) 0.25 in ethyl acetate.
Figure 2. CD spectra of the three analogues: S-deoxo-Ala3-amani-
namide (solid line), S-deoxo-D-Ala3-amaninamide (dotted line) and
S-deoxo-Gly3-amaninamide (dashed line).
4.10 Hence, one could draw the conclusion that 1 exhibits a
conformation analogous to the normal form and 2 exists in the
iso-form. This assumption should be proven by careful con-
formational analysis in solution using NMR spectroscopy.
Boc-Hpi-Gly-Ile-Gly-Cys(STrt)-Asn-Hyp-D-Ala-OtBu (2.3). The
solution of 0.82 g (1.38 mmol) of 2.2 in a mixture of 1.05 mL of TFA
and 40 mL of CH2Cl2 was stirred at room temperature for 20 min and
neutralized with NMM. This precooled solution was added after 8
min to a MA solution prepared at -10 °C from 1.23 g (1.38 mmol) of
Boc-Hpi-Gly-Ile-Gly-Cys(STrt)-OH in 30 mL of THF with 0.168 g
(1.38 mmol) of iBCCl and 0.168 g (1.66 mmol) of NMM. After 3 h
stirring at room temperature the reaction mixture was worked up as
described for analogous cases. Peptide 2.3 was obtained by chroma-
tography with methanol in a 2.5 cm × 250 cm column of Sephadex
LH 20. Fractions (25 mL) 35-40 afforded 0.98 g (57% yield) of title
compound. Rf ) 0.5 in ethyl acetate-methanol 9:1.
2-Mercapto-L-tryptophylglycyl-L-isoleucylglycyl-L-cysteinyl-L-as-
paraginyl-trans-hydroxy-L-prolyl-D-alanine Cyclic (1f5) Sulfide
(2.4). The solution of 0.98 g (0.78 mmol) of octapeptide 2.3 in 250
mL of TFA was kept at room temperature for 3 h and evaporated in
vacuo. The residue was washed thoroughly with diethyl ether and
Experimental Methods
Synthesis. The amino acids, other chemicals, and solvents were of
analytical grade. Analysis by TLC was performed on silica (Merck
60 F 254) using the solvent mixtures indicated for each substance.
Detection of the substances on TLC plates was feasible as dark spots
on illumination with UV. The inhibitory capacities were tested with
an enzyme preparation from embryos of Drosophila melanogaster in
the laboratory of E.K.F. Bautz, Heidelberg.
Abbreviations. iBCCl, isobutyloxycarbonylchloride; Boc, tert-
butyloxycarbonyl; DCCI, dicyclohexylcarbodiimide; Ddz, R,R-di-
methyl-3,5-dimethoxybenzyloxycarbonyl; DMF, dimethylformamide;

4382 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Schmitt et al.
chromatographed on a Sephadex LH 20 column (5 cm × 150 cm) in
aqueous 0.004 N NH4OH as eluent. Fractions (30 mL) 31-40 afforded
0.3 g (46% yield) of seco-thioether 2.4. Rf ) 0.3 in n-butanol-acetic
acid-water 4:1:1.
CD Spectra. Circular dichroism (CD) measurements were carried
out in methanolic solutions on a Jasco J-600 instrument with appropriate
quartz cell.
NMR Measurements. Proton and carbon spectra have been
recorded on a Bruker AMX-500 at 500 MHz. Data were processed
on a Bruker X32 workstation using the Uxnmr program. All spectra
were recorded at a temperature of 300 K, except for the temperature
coefficients, which were measured for the amide proton resonances by
variation of the temperature from 300 to 325 K. The sample
concentration of 1 was 20 mM in DMSO-d6, that of compound 2 was
15 mM, and that of 3 was 5 mM. All samples were vacuum-sealed
after three freeze-thaw-pump cycles. The spectra were calibrated
relative to DMSO-d6 (1H: 2.49 ppm, 13C: 39.5 ppm) as internal
standard.
The z-filtered TOCSY spectra11,12 for 1 and 2 were recorded with
4096 data points in t2 and 512 data points in t1 with a relaxation delay
of 1 s and a spinlock mixing time of 80 ms using a 10 kHz DIPSI
sequence (64 scans). The spectrum of 3 was recorded under the same
conditions, but with 430 t1 increments and 128 scans per increment.
For processing a π/2 shifted squared sine bell was used prior to Fourier
transformation in both dimensions.
P.E. COSY spectra13 were recorded with 8192 data points in t2 and
512 in t1 with a relaxation delay of 1 s using a 37° read pulse and 16,
23, and 64 scans for compound 1, 2, and 3, respectively. For the one-
dimensional reference spectra 16 384 data points were acquired using
the same number of scans as in the 2D spectra and a delay of 4.5 µs
before acquisition. A π/2 shifted squared sine bell apodization was
applied in both dimensions before phase-sensitive Fourier transformation
(TPPI).
For the carbon assignment HMQC spectra14-16 using a BIRD
sequence for 12CH presaturation17,18 and GARP-decoupling during
acquisition with 2048 data points in t2 and 512 (1 and 2) or 256 (3)
increments in t1 were recorded with a relaxation delay of 192 ms and
16 (1), 32 (2) or 96 (3) scans. The signal after the BIRD sequence
was minimized with a delay of 198 ms. Additionally HMQC-TOCSY
spectra19 using a 10 kHz MLEV-17 spinlock20 with a mixing time of
80 ms were acquired with 16 scans for compound 1 and 64 scans for
2. The first t1 increment used was calculated according to (4‚IN0-
180°[1H]-1.28‚90°[13C])/2 to obtain a phase correction of 180° (zero
order) and -360° (first order) in t1.21 The spectra were processed by
Fourier transformation after the application of a π/2 shifted squared
sine bell apodization.
Cyclic(L-asparaginyl-trans-hydroxy-L-prolyl-D-alanyl-L-trypto-
phylglycyl-L-isoleucylglycyl-L-cysteinyl) (4f8) Sulfide (2). The
solution of 0.3 g (0.42 mmol) of 2.4 in 80 mL of THF and 150 mL of
DMF was treated at -10 °C under stirring with 34 mg of TFA and 5
min later with 41 mg of iBBCI. After 10 min stirring at the same
temperature, the precooled solution of 100 mg of NMM in 240 mL of
THF and 450 mL of DMF was added. After standing 1 h at 0 °C and
overnight at room temperature, the solution was evaporated in vacuo,
and the residue was chromatographed on a silica gel column (2.5 cm
× 50 cm) in CHCl3-CH3OH-H2O (65:25:4) as eluent. The residue
obtained from fractions (15 mL) 25-35 was further purified by HPLC
in a CH3CN-H2O gradient yielding 80 mg (23%) of title compound.
Rf ) 0.8 in CHCl3-CH3OH-H2O 65:25:4.
Mass spectrum measurement (FAB) of (M + H)+ exhibited a mass
of 813 as expected.
Ddz-Hyp-Gly-OtBu (3.1). To the solution of 1.6 g (4 mmol) of
Ddz-Hyp-OH in 30 mL of CH2Cl2 were added at -10 °C under stirring
0.544 g (4 mmol) of iBCCl and 0.544 g (5.4 mmol) of NMM. After
10 min the solution of 0.52 g (4 mmol) of H-Gly-OtBu in 20 mL of
CH2Cl2 was added. After 3 h stirring at room temperature the reaction
mixture, worked up as usual, gave 1.6 g (85% yield) of 3.1. Rf ) 0.6
in ethyl acetate.
Ddz-Asn-Hyp-Gly-OtBu (3.2). The solution of 1.6 g (3.43 mmol)
of Ddz-Hyp-Gly-OtBu in a mixture of 2.6 mL of TFA and 100 mL of
CH2Cl2 was stirred at room temperature for 20 min, neutralized with
NMM, and mixed at 0 °C with a solution of 1.2 g (3.43 mmol) of
Ddz-Asn-OH, 0.926 g (6.86 mmol) HOBt and 0.7 g (3.43 mmol) DCCI
in 40 mL of DMF. After stirring 1 h at 0 °C and overnight at room
temperature, the reaction mixture, worked up as described in the
previous paragraphs, gave a residue which was chromatographed on a
Sephadex LH 20 column (5 cm × 250 cm) in methanol yielding 1 g
(50%) of 3.2. Rf ) 0.2 in ethyl acetate.
Boc-Hpi-Gly-Ile-Gly-Cys(STrt)-Asn-Hyp-Gly-OtBu (3.3). The
solution of 1 g (1.7 mmol) of 3.2 in mixture of 1.3 mL of TFA and 50
mL of CH2Cl2 was stirred for 20 min at room temperature, neutralized
with NMM, and added after 8 min to a MA solution prepared at -10
°C from 1.5 g (1.7 mmol) Boc-Hpi-Gly-Ile-Gly-Cys(STrt)-OH in 30
mL of THF with 0.23 g (1.7 mmol) of iBCCl and 0.23 g (2.3 mmol)
of NMM. After 3 h stirring at room temperature the reaction mixture,
worked up as usual, gave a residue which was chromatographed on a
Sephadex LH 20 column (5 cm × 250 cm) in methanol. Fractions
(30 mL) 98-105 afforded 0.93 g (47% yield) of title compound. Rf )
0.4 in ethyl acetate-methanol 9:1.
The HMBC spectra22 were recorded with 4096 data points in t2 and
384 points in t1 with a relaxation delay of 1 s and a 60 ms delay for
the evolution of the long-range coupling with 160 (1 and 2) and 224
scans (3). To increase the resolution in t1, the carbonyl resonances
were folded back.21 The first t1 increment was calculated according to
(4‚IN0-180°[1H]-1.28‚90°[13C])/2 for a proper phasing of folded and
nonfolded resonances.
HETLOC spectra23-25 for 1 and 2 were recorded with 192 respec-
tively 256 scans, 4096 points in t2 and 512 point in t1 using a relaxation
delay of 1 s and a delay of 198 ms for minimization of signals from
2-Mercapto-L-tryptophylglycyl-L-isoleucylglycyl-L-cysteinyl-L-as-
paraginyl-trans-hydroxy-L-prolylglycine Cyclic (1f5) Sulfide (3.4).
The solution of 0.93 g (0.75 mmol) of octapeptide 3.3 in 250 mL of
TFA was kept 3 h at room temperature, and the residue of the
evaporation was washed with diethyl ether and chromatographed on a
Sephadex LH 20 column (5 cm × 150 cm) in 0.004 N NH4OH.
Fractions (30 mL) 35-45 yielded 0.38 g (63%) of 3.4. Rf ) 0.25 in
n-butanol-acetic acid-water 4:1:1.
(11) Titman, J. J.; Neuhaus, D.; Keeler, J. J. Magn. Reson. 1989, 85,
111-131.
(12) Titman, J. J.; Keeler, J. J. Magn. Reson. 1990, 89, 640-646.
(13) Mueller, L. J. Magn. Reson. 1987, 72, 191-196.
(14) Mueller, L. J. Am. Chem. Soc. 1979, 101, 4481-4484.
(15) Bendall, M. R.; Pegg, D. T.; Doddrell, D. M. J. Magn. Reson. 1983,
52, 81-117.
Cyclic(L-asparaginyl-trans-hydroxy-L-prolylglycyl-L-tryptophylg-
lycyl-L-isoleucylglycyl-L-cysteinyl) (4f8) Sulfide (3). To the stirred
solution of 0.38 g (0.47 mmol) of 3.4 in 125 mL of THF and 250 mL
of DMF were added at -10 °C 54 mg of TFA and 5 min later 64 mg
of iBCCl. After 10 min stirring at -10 °C the precooled solution of
100 mg of NMM in 300 mL of THF and 600 mL of DMF was added.
After stirring 1 h at 0 °C and overnight at room temperature, the
reaction, worked up as previously described, gave a residue which was
chromatographed on a silica gel column (2.5 cm × 50 cm) in CHCl3-
CH3OH-H2O 65:25:4. The residue obtained from fractions (7 mL)
20-32 was further purified by HPLC in a CH3CN-H2O gradient
yielding 90 mg (24%) of title compound. Rf ) 0.7 in CHCl3-CH3-
OH-H2O 65:25:4.
(16) Bax, A.; Griffey, R. H.; Hawkins, B. L. J. Magn. Reson. 1983, 55,
301-315.
(17) Bax, A.; Subramanian, S. J. Magn. Reson. 1986, 67, 565-569.
(18) Byrd, R. A.; Summers, M. F.; Zon, G.; Spellmeyer Fouts, C.;
Marzilli, L. G. J. Am. Chem. Soc. 1986, 108, 504-505.
(19) Lerner, A.; Bax, A. J. Magn. Reson. 1986, 69, 375-380.
(20) Bax, A.; Davis, D. G. J. Magn. Reson. 1985, 65, 355-360.
(21) Schmieder, P.; Kessler, H. Magn. Reson. Chem. 1991, 29, 375-
380.
(22) Bax, A.; Summers, M. F. J. Am. Chem. Soc. 1986, 108, 2093-
2094.
(23) Kurz, M.; Schmieder, P.; Kessler, H. Angew. Chem., Int. Ed. Engl.
1991, 30, 1329-1330.
(24) Schmieder, P.; Kurz, M.; Kessler, H. J. Biomol. NMR 1991, 1, 403-
420.
Mass spectrum measurement (FAB) of (M+H)+ exhibited a mass
of 799 as expected.
(25) Wollborn, U.; Leibfritz, D. J. Magn. Reson. 1992, 98, 142-146.

S-Deoxo-Xaa3-amaninamide Analogues
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4383
Figure 4. Formation of the first ring of analogs 1-3 by the Savige-Fontana reaction and cyclization to yield the bicyclic octapeptides.
the BIRD sequence. The mixing time of the 10 kHz MLEV-17 spinlock
used was 40 ms. The complementary spectrum of 3 was recorded with
300 t1 increments and 656 scans.
next 5 ps at 200 K. The trajectory used for analysis was calculated
for 180 ps at 300 K, and a structure was stored every 1000 fs.
The test calculations were performed in vacuo with the Gromos
program using NOEs and coupling constants as restraints (kNOE ) 2000
kJ/(mol‚nm2), kJ ) 0.25 kJ/mol). The starting structures were manually
built using the Biopolymer module of the InsightII program (BIO-
SYM).34 These structures were optimized by molecular dynamic
calculations for 2 ps at 1000 K, 2 ps at 500 K, and finally for 5 ps at
300 K. The trajectories were then recorded for 50 ps at a temperature
of 300 K.
For the determination of proton-proton distances NOESY spectra26
with a mixing time of 200 ms, a relaxation delay of 1 s and 4096
respectively 512 data points in t2 and t1 have been recorded. In addition
NOESY and ROESY spectra with shorter mixing times were recorded
to ensure the lack of spin diffusion. The spectra were processed by
applying a π/2 shifted squared sinebell in both dimensions and zero
filling to 1024 complex points in t1 before phase-sensitive Fourier
transformation. The integral intensities were measured using the
integral subroutine within the Uxnmr software. The resulting crosspeak
volumes were transformed into distances using the isolated two-spin
approximation with reference to the Hyp2-Hâ,Hâ-crosspeak which was
set to a distance of 178 pm.
Results
The synthesis of the amanitin analogs followed the route
described in ref 10. As depicted in Figure 4 the respective
N-terminal l-Hpi-octapeptide is subjected to the indole-2-
thioether forming Savige-Fontana reaction, and the monocyclic
tryptophan octapeptide so obtained is converted to the bicyclic
amatoxin analog by internal peptide bond formation. The
synthesis of analog 1 has already been published (number 1i in
ref 35).
Structure Calculations. A Silicon Graphics Indy with R4400
processor was used for all calculations. The distance geometry
calculations were carried out using a modified version of the Disgeo
program.27,28 For each compound 100 structures were embedded into
the four-dimensional Cartesian space. These structures were minimized
for 400 steps with the steepest descent algorithm followed by 5 ps
distance driven dynamics (DDD)29,30 at 300 K using NOEs and chiral
volumes as restraints and 4 ps DDD with a weak coupling to a
temperature bath of 1 K. After projection into the three-dimensional
space 100 steps of steepest descent minimization using chiral volumes
as restraints were performed to relax steric strain. Finally a distance
and angle driven dynamic calculation (DADD)31 was carried out for 5
ps at a temperature of 500 K and subsequently with a weak temperature
coupling at 1 K using NOEs, 3J-coupling constants and chiral volumes
as restraints. For the following molecular dynamic simulations a
modified version of the Gromos software32 was used, to permit
additionally the inclusion of coupling constants as restraints. The
structure of each compound with the lowest total error resulting from
the DG calculations was chosen as a starting structure. It was placed
in a truncated octahedron (a ) b ) c ) 3.43415 (1), 3.42954 (2) and
3.28463 nm (3)) and soaked with 143 to 164 DMSO molecules. The
system was relaxed by using a maximum of 2000 steps with the steepest
descent algorithm and a force constant for the NOEs of 2000 kJ/(mol‚-
nm2). In the first step only the solvent was relaxed while the peptide
was held fixed. In the second step the whole system was relaxed. The
dynamic simulations were performed using the Shake algorithm33 with
a stepsize of 1 fs. The temperature relaxation time was 10 fs, the
nonbonded interactions were updated every 10 steps with a cutoff
distance of 0.8 nm. The upper and lower bounds for NOEs are set to
plus and minus 10% of the experimentally determined distances. Force
constants for NOEs and coupling constants were set to 2000 kJ/(mol‚-
nm2) and 0.25 kJ/mol, respectively. For equilibration the first 5 ps
were calculated with a strong temperature coupling at 100 K and the
The assignment of all proton and carbon shifts was performed
following a standard strategy previously described.36-39 Se-
quential assignment was carried out using heteronuclear long
range couplings from HMBC experiments. Proton chemical
shifts are listed in Table 2; carbon chemical shifts are available
as supporting information. For all of the three compounds only
one single conformation in DMSO solution was detected.
Hence, no cis-trans isomerization occurs under measurement
conditions. The proton chemical shifts of the three analogues
are quite similar and each exhibits good dispersion. Other than
the signals from residue 3 only few differences in proton shifts
can be detected. The resonance of Asn1-HN in 1 is shifted 0.2
ppm downfield with respect to 2 and 3. Also the â-protons of
Asn1 in compound 1 are shifted about 0.4 ppm downfield
relative to 2. These â-protons are separated by 0.3 ppm in 1
and 2, but they are degenerated in 3. A further downfield shift
of about 0.7 ppm relative to 2 and 3 is detected for the δ-amide
protons of Asn1 in compound 1. The separation of the â-protons
of Trp4 in 1 is smaller than in 2 and 3. These comparisons
give a first hint, that the structural difference between these three
analogues should occur in the region of residues 1-4.
3
The homonuclear JHN,HR coupling constants were in most
cases extracted from the 1D proton spectra recorded with 16 384
data points. Only those of Trp4-HN and Cys8-HN in compound
1, of Cys8-HN in 2 and of Gly3-HN as well as that of Gly7-HN
in 3 have been determined from the related P.E. COSY
(26) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J. Chem. Phys.
1979, 71, 4546-4553.
(27) Havel, T. F. Disgeo, Quantum Chemistry Exchange Program;
Exchange No. 507, Indiana University, 1986.
(34) InsightII; Biosym Technologies Inc.: San Diego, CA 1993.
(35) Zanotti, G.; Mo¨hringer, C.; Wieland, Th. Int. J. Peptide Protein
Res. 1987, 30, 450-459.
(36) Kessler, H.; Bermel, W.; Mu¨ller, A.; Pook, K.-H. In Peptides:
Analysis, Synthesis and Biology; Udenfried, S., Meienhofer, J., Hruby, V.,
Eds.; Academic Press: New York, 1985; Vol. 7, pp 437-473.
(37) Wu¨thrich, K. NMR of Proteins and Nucleic Acids; John Wiley &
Sons: New York, 1986.
(38) Kessler, H.; Seip, S. In Two-Dimensional NMR Spectroscopy:
Applications for Chemists and Biochemists; Croasmun, W. R., Carlson, R.
M. K., Eds.; VCH Publisher: New York, 1994; pp 619-654.
(39) Kessler, H.; Schmitt, W. in Encyclopedia of Nuclear Magnetic
Resonance; Grant, D. M., Harris, R. K., Eds.; John Wiley & Sons: New
York, 1996, pp 3527-3537.
(28) Crippen, G. M.; Havel, T. F. Distance Geometry and Molecular
Conformation; Research Studies Press LTD.: Taunton, Somerset, UK, 1988.
(29) Kaptein, R.; Boelens, R.; Scheek, R. M., van Gunsteren, W. F.
Biochemistry 1988, 27, 5389-5395.
(30) Kuntz, I. D.; Thomason, J. F.; Oshiro, C. M. Methods Enzymol.
1989, 177, 159-204.
(31) Mierke, D. F.; Geyer, A.; Kessler, H. Int. J. Peptide Protein Res.
1994, 44, 325-331.
(32) van Gunsteren, W. F.; Berendsen, H. J. C. Groningen Molecular
Simulation (Gromos) Library Manual; Biomos, B. V., Eds.; Nijenborgh
16, NL 9747 AG Groningen, 1987.
(33) van Gunsteren, W. F.; Berendsen, H. J. C. Mol. Phys. 1977, 34,
1311-1321.

4384 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Schmitt et al.
Table 3. Vicinal Coupling Constants Defining the Backbone
Table 2. Proton Chemical Shifts [ppm] for the Three
S-Deoxoamaninamide Analogues in DMSO-d6
φ-Angle as Determined in DMSO-d6 at 300 K in [Hz]
residue
Asn1
proton
1
2
3
residue
Asn1
1
2
3
HN
8.40
4.76
3.17
2.85
8.03
7.38
8.17
4.87
2.70
2.47
7.27
6.86
8.22
4.84
2.73
2.72
7.25
7.04
3JHN,HR ) 3.9
3JHN,HR ) 4.8
3JHN,HR ) 4.6
3JHN,C′ ) 3.1
3JHN,Câ ) 1.9
HR
3JHN,C′ ) 2.7
3JHN,C′ ) 1.8
Hâ1
Hâ2
HNδ1
HNδ2
Hyp2
Xaa3
3JHR,C′i-1 ) 4.3
3JHN,HR ) 8.9
3JHN,HR ) 8.6
3JHN,Câ ) 1.4
3JHR,C′i-1 ) 2.9
3JHN,HRproR ) 7.0
3JHN,HRproS ) 5.8
3JHRproS,C′i-1 ) 4.4
3JHN,Câ ) 1.0
Hyp2
HR
4.29
2.20
1.89
4.39
5.25
3.81
3.73
4.48
2.06
1.90
4.45
5.31
3.64
3.63
4.42
2.11
1.90
4.44
5.29
3.69
3.69
Hâ1
Hâ2
Hγ
Trp4
Gly5
3JHN,HR ) 9.1
3JHN,Câ ) 1.1
3JHR,C′i-1 ) 1.4
3JHN,HR ) 9.6
3JHN,HR ) 9.8
3JHN,Câ ) 1.2
3JHN,Câ ) 0.3
HOH
Hδ1
Hδ2
3JHN,HRproS ) 7.7
3JHN,HRproR ) 4.8
3JHN,C′ ) 3.1
3JHN,HRproS ) 6.8
3JHN,HRproR ) 2.5
3JHN,HRproS ) 6.5
3JHN,HRproR ) 2.7
3JHRproS,C′i-1 ) 5.4 3JHN,C′ ) 1.7
Xaa3
Trp4
HN
HR
Hâ
8.27
4.31
1.35
8.90
4.36
1.28
8.84
3JHRproS,C′i-1 ) 4.5 3JHRproR,C′i-1 ) 5.1 3JHRproR,C′i-1 ) 5.0
3.89 (HR1)
3.66 (HR2)
3JHRproR,C′i-1 ) 3.8
Ile6
3JHN,HR ) 5.2
3JHN,HR ) 5.7
3JHN,Câ ) 1.3
3JHN,HR ) 5.5
HN
7.83
4.93
3.28
3.01
11.24
7.54
7.24
7.00
7.10
8.10
4.83
3.71
2.83
11.21
7.51
7.25
7.02
7.11
8.03
4.92
3.57
2.90
11.22
7.51
7.25
7.02
7.11
HR
Hâ1
Hâ2
HNꢀ1
Hꢀ3
HZ2
HZ3
HCH
Gly7
3JHN,HRproS ) 7.8
3JHN,HRproR ) 4.7
3JHN,HRproS ) 6.9
3JHN,HRproS ) 7.4
3JHN,HRproR ) 4.9
3JHN,HRproR ) 4.7
3JHRproS,C′i-1 ) 2.9 3JHRproS,C′i-1 ) 3.3 3JHRproR,C′i-1 ) 5.8
3JHRproR,C′i-1 ) 5.7 3JHRproR,C′i-1 ) 5.6
Cys8
3JHN,HR ) 9.7
3JHN,Câ ) 0.3
3JHR,C′i-1 ) 4.1
3JHN,HR ) 9.2
3JHN,HR ) 9.4
3JHN,Câ ) 1.0
3JHN,C′ ) ∼0
3JHN,Câ ∼0
3JHN,C′ ) 1.0
3JHR,C′i-1 ) 4.5
Gly5
Ile6
HN
8.15
4.08
3.44
7.70
3.94
3.54
7.79
3.98
3.51
HR1
HR2
Table 4. Vicinal Coupling Constants Defining the ø-Angle as
Determined in DMSO-d6 at 300 K in [Hz]
HN
8.51
3.74
1.58
1.53
1.10
0.77
0.81
8.56
3.83
1.65
1.52
1.13
0.80
0.82
8.54
3.81
1.63
1.53
1.12
0.80
0.82
HR
Hâ
residue
Asn1
1
2
3
Hγ1,1
Hγ1,2
Hγ2(Me)
Hδ
3JHR,HâproR ) 4.7
3JHR,HâproS ) 4.2
3JC′,HâproR ) 7.5
3JC′,HâproS ∼0
3JHR,HâproR ) 6.3
3JHR,HâproS ) 4.1
3JC′,HâproR ) 4.5
3JC′,HâproS ) 5.4
Gly7
Cys8
HN
8.86
3.93
3.41
8.85
3.94
3.49
8.86
3.94
3.45
Hyp2
3JHR,HâproR ) 7.3
3JHR,HâproR ) 7.6
3JHR,HâproR ) 7.3
HR1
HR2
3JHR,HâproS ) 11.5 3JHR,HâproS ) 10.1 3JHR,HâproS ) 10.9
3JC′,HâproR ) ∼0
3JC′,HâproS ) 4.5
3JHâproR,Hγ ∼9
3JC′,HâproR ) ∼0
3JC′,HâproS ) 5.4
3JHâproR,Hγ ) 8.0
3JHâproS,Hγ ) 3.7
HN
7.80
4.64
3.18
2.81
7.54
4.62
3.33
2.96
7.60
4.63
3.28
2.93
HR
3JHâproS,Hγ ) 3.6
3JHγ,HδproS ) 3.2
3JHγ,HδproR ∼9
Hâ1
Hâ2
Trp4
3JHR,HâproS ) 13.0 3JHR,HâproS ) 12.7 3JHR,HâproS ) 13.0
crosspeak according to the method of Kim and Prestegard.40
3JHR,HâproR ) 6.4
3JC′,HâproS ) 3.0
3JC′,HâproR ) 9.0
3JHR,HâproR ) 5.4
3JC′,HâproR ) 8.5
3JHR,HâproR ) 5.7
3
The heteronuclear JHN,Câ couplings were measured from the
HETLOC experiments.24 Following the procedure proposed by
Titman and Keeler,11,12 the heteronuclear 3JHN,C′ and 3JHR,C′ i-1
are calculated via the traces of HMBC and z-filtered TOCSY
spectra. P.E. COSY crosspeaks have been used for the
Ile6
Cys8
3JHR,Hâ ) 9.0
3JHR,Hâ ) 9.5
3JHR,Hâ ) 8.3
3JHR,HâproR ) 9.4
3JHR,HâproS ) 3.0
3JC′,HâproS ) 5.3
3JHR,HâproR ) 10.8 3JHR,HâproR ) 8.6
3JHR,HâproS ) 3.3
3JC′,HâproR ) 4.8
3JC′,HâproS ) 7.2
3JHR,HâproS ) 3.7
3JC′,HâproR ) 4.5
3JC′,HâproS ) 6.5
measurement of JHR,Hâ coupling constants.13 For compound
3
1 42 coupling constants and for 2 39 coupling constants were
determined, whereas only 28 could be extracted for substance
3, due to the lower signal to noise ratio in the related
heteronuclear spectra (see Tables 3 and 4). With these values
the populations of the side-chain rotamers for Asn1 were
calculated using the Pachler equations.41 In 1 the rotamer with
ø1 ) +60° is predominant (66%), whereas in compound 2 the
population of the rotamers +60° and -60° are nearly equivalent.
Due to the degeneration of the â-protons in 3 no preferred
orientation of the side chain of Asn1 could be derived. From
the homonuclear and the heteronuclear coupling constants
diastereotopic assignment of Trp4 and Cys8 â-protons was
performed. In the course of the calculations the assignment of
the prochiral R-protons of glycin residues was also possible as
in every structure only one of the two R-protons fulfilled the
NOE distance constraints. Subsequent calculations were carried
out without pseudoatoms.
A comparison of the temperature coefficients shows major
differences among the examined analogues only for the amide
protons of residue 3 (see Table 5). This amide proton seems
to form an additional hydrogen bond in compound 1.42 The
determined NOE derived distances, together with the upper and
lower limits used for the calculations, are available as supporting
information. The NOESY spectra indicate that all peptide bonds
have trans configuration, as no strong HR,HR crosspeaks were
detected.
(40) Kim, Y.; Prestegard, J. H. J. Magn. Reson. 1989, 84, 9-13.
(41) Pachler, K. G. R. Spectrochim. Acta 1964, 20, 581-587.
(42) Llinas M.; Klein, M. P., J. Am. Chem. Soc. 1975, 97, 4731-4737.

S-Deoxo-Xaa3-amaninamide Analogues
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4385
Table 5. Temperature Coefficients Given in [ppb/K] of the Amide
Protons of All Three Analogues, Determined in the Range of 300
and 325 K
1
2
3
Asn1-HN
Xaa3-HN
Trp4-HN
Gly5-HN
Ile6-HN
-2.0
-1.5
-1.3
-2.0
-5.1
-4.3
-1.7
-1.5
-4.0
-2.5
-0.7
-4.8
-5.0
-1.0
-1.1
-2.8
-1.6
-0.7
-4.4
-4.5
-0.8
Gly7-HN
Cys8-HN
Table 6. Dihedral Angles of the Refined Structures of the
S-Deoxoamaninamides (Means of a 150 ps MD Trajectory)
residue
Asn1
dihedral
1
2
3
φ
-173.3
164.6
59.0
-121.1
-162.7
150.3
-107.9
138.5
-148.2
146.3
-57.6
98.6
ψ
ø1
ø2
Hyp2
φ
-66.7
12.0
-28.3
41.0
-38.9
22.4
-64.3
119.1
-27.9
39.1
-35.4
18.5
-65.1
109.2
-27.8
36.9
32.4
15.6
ψ
ø1
ø2
ø3
ø4
Xaa3
Trp4
φ
ψ
-122.3
104.5
-18.4
119.9
-30.6
7.0
φ
-143.5
-38.0
176.2
-91.0
-53.3
176.5
-85.9
-53.6
174.0
ψ
ø1
ø2
-127.9
-130.4
-115.4
Gly5
Ile6
φ
ψ
142.3
-176.7
160.6
-168.8
169.8
-175.0
φ
-73.0
117.4
-59.2
159.3
-73.0
134.9
-55.9
-56.3
-68.0
119.4
-53.9
157.3
ψ
ø1
ø2
Figure 5. Stereoscopic picture of the conformation of S-deoxo-Ala3-
amaninamide (A), of S-deoxo-D-Ala3-amaninamide (B) and of S-deoxo-
Gly3-amaninamide (C) in DMSO.
Gly7
Cys8
φ
ψ
98.8
9.2
59.3
33.8
65.2
16.1
φ
-130.4
-78.4
-174.2
178.2
-131.7
-79.1
167.3
-106.6
-91.5
164.8
ψ
ø1
ø2
The mean distance restraint violation is low (12 pm for 1
and 2, 14 pm for 3) and the resulting dihedral angles (see Table
6) fit to the experimentally measured coupling constants. In
all three structures every proton having a low negative temper-
ature coefficient is involved in a hydrogen bond. As the
information on hydrogen bond donors was not incorporated into
the calculations, the fulfillment of these data is an additional
proof for the correctness of the determined structures.
The structure of 1 shows a bent backbone and exhibits the
normal form. It has a âII-turn with Ile6 in the i+1 position
and on the opposite side a âI-turn with Hyp2 in the i+1 position.
The latter âI-turn is stabilized via an additional hydrogen bond
from the Ala3-HN to the carboxyl oxygen of the Asn1 side chain.
Therefore, the orientation of this side-chain is fixed to a ø1 angle
of +60°, which is in accordance with the analysis of the
sidechain rotamer populations. The conformation of the peptide
backbone is fixed by additional hydrogen bonds between Asn1-
HN/Gly5-C′ and Gly5-HN/Asn1-C′. Figure 6 shows the backbone
structure of 1 to illustrate the hydrogen bonding network.
The conformation of compound 1 as determined in solution
is nearly identical with the X-ray structure of â-amanitin.47 The
RMSD value for the superposition of all backbone atoms is only
22 pm (Figure 7). There are only minor differences in the side-
chain orientations, even the stabilization of the âI-turn is the
same in both structures.
-169.2
-178.2
All calculated distance limits were incorporated as restraints
for structure calculation. The coupling constants (for which
parameters for the Karplus curves exist) were used directly as
restraints, following the method introduced by Kim and Prest-
egard43 and improved by Mierke and Kessler.44-46 Distance
geometry calculations on all three analogues converged (RMSD
< 100 pm), and the conformations of the 20 structures with
the lowest total error were very similar. The RMSD value for
the backbone atoms of 1 was 28 pm, that for 2 was 52 pm, and
53 pm for 3. These results suggest that the structure can be
described by one predominant conformation for every analog
in solution. Hence, in each case the DG structure with the
lowest total error was taken as a starting structure for the
subsequent molecular dynamic simulation in solution using NOE
values and coupling constants as restraints. Only the last 150
ps of the resulting trajectories were analyzed, the first 30 ps
are neglected for equilibration of the structure at the simulation
temperature. The averaged and minimized structures for the
three compounds are shown in Figure 5.
(43) Kim, Y.; Prestegard, H. Proteins Struct. Funct. Genet. 1990, 8, 377-
385.
(44) Mierke, D. F.; Kessler, H. Biopolymers 1993, 33, 1003-1017.
(45) Mierke, D. F.; Kessler, H. Biopolymers 1992, 32, 1277-1282.
(46) Eberstadt, M.; Mierke, D. F.; Ko¨ck, M.; Kessler, H. HelV. Chim.
Acta 1992, 75, 2583-2592.
(47) Kostansek, E. C.; Lipscomb, W. N.; Yocum, R. R., Thiessen, W.
E. Biochemistry 1978, 17, 3790-3795.

4386 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Schmitt et al.
Figure 6. Stereoscopic picture of the backbone conformation of
S-deoxo-Ala3-amaninamide (3); hydrogen bonds are painted as dashed
lines.
Figure 8. Structures of the turns with Hyp2 in the i+1 position in all
three analogues.
explained by allylic 1,3-strain.48 Due to the repulsion of the
carbonyl oxygen of Hyp2 in compound 1 with the methyl group
of Ala3 an orientation of the amide plane with coplanar
arrangement of the oxygen and the R-proton is favored. As
the chirality at the R-carbon of residue 3 in compound 2 is
inverted an opposite orientation of the amide plain is found.
Compound 3 has no side chain in position 3 (Gly). Therefore
none of these two orientations of the amide plain is energetically
favored. Hence, we suggest that 3 interconverts rapidly on the
NMR time scale between the conformation of 1 (âI-turn) and
that of 2 (âII-turn).
Figure 7. Superposition of the structures of S-deoxo-Ala3-amaninamide
(black line) and â-amanitin (gray line).
The conformation of the analogues 2 and 3 (see Figure 5) is
quite similar to that of 1. The RMSD value for the superposition
of all backbone atoms is 22 pm. Both structures show the
normal form and exhibit also a bent backbone where both halves
are nearly orthogonal to each other. These structures consist
of two âII-turns, one with Ile6 in the i+1 and the other one
with Hyp2 in the i+1 position.
A comparison of the conformations of all three examined
analogues shows that the only difference is found in the structure
of the â-turn with Hyp2 in the i+1 position. While the L-Ala3-
analogue has a âI-turn, which is stabilized through an additional
hydrogen bond between Ala3-HN and the carboxyl oxygen of
the Asn1 side chain, the D-Ala3- and the Gly3-analogue both
show a âII-turn (see Figure 8).
The structures of 2 and 3 exhibit the same predominant
conformation in solution but not with the same stability. If one
compares the temperature coefficients of the amide proton of
residue 3; the value for compound 2 (-4.0 ppb/K) indicates a
solvent exposed orientation, whereas the value for 3 (-2.8 ppb/
K) lies in the border region where both, an internal and an
external orientation are possible and that for 1 (-1.5 ppb/K) is
typical for the internal orientation of the amide protons. The
Xaa3-HN/HR and the Xaa3-HN/Hyp2-HR distances show the same
pattern: the sequential NOE distances are 213 pm (2) and 228
pm (3), and the interresidual distances are 281 pm (2) and 364/
332 pm (3). Additionally the Xaa3-HN/Trp4-HN distance in 3
is 14 pm shorter than in 2. The homonuclear 3JHN,HR couplings
of residue 3 in the glycin analogue are 7.0 and 5.8 Hz. These
data, in particular the coupling constants, indicate that the Gly3-
analogue is more flexible than the D-Ala3-analogue. This is
shown by the more averaged values for compound 3 which are
not compatible with an absolutely fixed peptide plane between
Hyp2 and Gly3. These conformational differences can be
A simple comparison of distances within the normal and the
iso form to distinguish them is not unambiguous because of
conformational differences and easily mislead the interpretation.
Hence, detailed calculations taking into account the whole
network of distances and coupling constants were performed.
All three analogues show the structure of the normal form in
solution even though interchange to the iso-form was allowed
during the DG calculations. But the initial assumption, that 2
might exist in the iso form was disproven via additional test
calculations. As distance geometry calculations are independent
on the starting structure they should lead in every case to the
correct isomer. To test this a distance geometry calculation for
2 was performed using the same protocol but without any
restraints. The resulting structures show a ratio of 69:21
between the normal and the iso form. Hence, if the experimental
data could fit the iso form, the simulation should have lead to
this isomer. A further experiment was performed to exclude
the iso form via molecular dynamic simulations in vacuo. As
the two different isomers are not interconvertible modified
analogues of the three compounds were employed. The two
residues forming the transannular bridge were replaced by
alanine. For each of the three S-deoxo-amaninamides two
different starting structure were created; one with a backbone
conformation similar to the normal isomer and the other
analogous to the iso form. Apart from the distances to the side
chains of Trp4 and Cys8 all other restraints were used for these
calculations. Since both simulations converged for all three
compounds to the normal isomer, the existence of the iso form
in solution can be safely excluded.
The region of the CD spectra in-between a wavelength of
210 and 260 nm where deviations are detected is related to the
(48) Hoffmann, R. W. Angew. Chem., Int. Ed. Engl. 1992, 31, 1124-
1134.

S-Deoxo-Xaa3-amaninamide Analogues
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4387
carbonyls. The initial assumption that these differences may
arise from two isomeric forms can be excluded through the
performed test calculations. As the normal form of S-deoxo-
D-Ile3-amaninamide shows a CD spectrum similar to compound
1 (l-configuration) the influence of the chirality in this position
can also be denied. Therefore, the detected differences in the
CD spectra must arise from the discussed turn structures. The
âI-turn in compound 1 leads to negative, the âII-turn in 2 to
positive values. The Gly3-analogue shows a higher flexibility
than analogs 1 and 2. Hence, the resulting CD spectrum lies
in-between these of the other two as a result of conformational
averaging in the turn region with Hyp2 in the i+1 position.
Recent studies have shown that the most frequently occurring
types of â-turns (type I and type II) give significantly different
CD spectra: âI-turns generally give class C CD spectra and
âII-turns result in class B spectra.49,50 Our interpretation that
the differences between the CD spectra of the three analogs are
caused by variations of the â-turn structure fits very well to the
results given in ref 50. The red-shift of the maxima of about
20 nm compared to the spectra shown therein might be a result
of the different solvents. In ref 50 the authors state that an
increase in conformational flexibility leads to a decreased band
intensity, which corresponds very well to the effect found for
compound 3.
control without any addition. As the side chain in position 3
of the three analogues is very short compared to that of the
natural amatoxin-peptides, the detected rather small inhibitory
activities could be a result of a side chain which is too short.
Conclusion
The herein examined S-deoxo-amaninamide analogues all
appear in the normal form in solution. The backbone is bent
and both halves are nearly orthogonal to each other. Each
compound exhibits two â-turns with residues 2 and 6 in the
i+1 position. The conformational differences between the
analogues lie in the structure of the turn with Hyp2 in i+1. The
Ala3-analogue shows a âI-turn in this position, which is further
stabilized by an additional hydrogen bond to the Asn1 sidechain.
Its conformation is identical to that found for â-amanitin in the
crystal structure. Both of the other compounds show a âII-
turn. However, the Gly3 analogue seems to exhibit a higher
flexibility in this region.
Based on the results of conformational analysis we explain
the strong differences of the CD spectra in-between a wave-
length of 210 and 260 nm with the different types of turns with
Hyp2 in the i+1 position. The assumption that these differences
may arise from two isomeric forms can be excluded. The âI-
turn in compound 1 leads to negative, and the âII-turn in 2 to
positive values. As a result of conformational averaging the
Gly3-analogue shows a decreased band intensity.
Biological tests show that the inhibitory effect of the three
amatoxin analogs on RNA polymerase II of Drosophila mela-
nogaster embryos was more than 100 times smaller than that
of R-amatoxin. The L-Ala3-compound (1) showed Ki ) 10-5
(R-amanitin: Ki ) 10-8), comparable to that value obtained by
the same compound with RNA-polymerase of calf thymus (3
× 10-5). Surprisingly the D-Ala3-analogue (2) was about 5
times more inhibitory (Ki ) 5 × 10-6) than 1. This effect might
be caused by the methyl of D-Ala3 group in 2, which is more
exposed compared to 1. Glycine-analogue 3 exhibited no
inhibitory effect, in contrary it seemingly stimulated RNA
formation in 10-6 mol solution by 10-20% as compared to the
Acknowledgment. This contribution is dedicated to Ivar Ugi
on the occasion of his 65th birthday.
Supporting Information Available: Tables of carbon
chemical shifts, of the experimentally determined NOE values
and of the proton-proton distances resulting from the MD
calculations for all three compound are available (9 pages). This
material is contained in many libraries on microfiche, im-
mediately follows this article in the microfilm version of the
journal, can be ordered from the ACS, and can be downloaded
from the Internet; see any current masthead page for ordering
information and Internet access instructions.
(49) Hollo´si, M.; Majer, Z. S.; Ro´nai, A. Z.; Magyr, A.; Medzihradszky,
K.; Holly, S.; Perczel, A.; Fasman, G. D. Biopolymers 1994, 34, 177-185.
(50) Perczel, A.; Hollo´si, M.; Foxman, B. M.; Fasman, G. D. J. Am.
Chem. Soc. 1991, 113, 9772-9784.
JA9529706