close

Вход

Забыли?

вход по аккаунту

?

Dynamic Forcing a Method for Evaluating Activity and Selectivity Profiles of RGD (Arg-Gly-Asp) Peptides.

код для вставкиСкачать
[2] A. M. Caminade, J. P. Majoral, R. Mathieu, Chen?. Rev. 1991, 91, 575.
[3] H. G. Ang, B. 0. West, Ausl. J. Chem. 1967. 20. 1133; M. Baudler, F.
Salzer. J. Hahn. Z . Naturforsch. B 1982, 37, 1529: A. L. Rheingold, M. E.
Fountain, Organometullics 1984, 3, 1471; B. 0. West in Homoalomic
Rings, Chains and Macromolecules of Main-Group Elemenrs (Ed.: A. L.
Rheingold), Elsevier. Amsterdam. 1977. p. 409.
[4] H. G. Ang, J. S. Shannon, B. 0. West, Chem. Commzm. 1965, 10; M. A.
Bush, P. Woodward, J. Chem. Soc. A 1968. 1221.
[5] M. Baudler, J. Hahn, H. Dietsch. G. Furstenberg. 2. Nuturforsch. B 1976,
31. 1305.
[6] X-ray structure analysis: Stoe Stadi IV, MO,,. empirical absorption correction. data collection and refinement: wscan. profile analysis. 1: lattice
constants (200 K): a = h =1040.9(1), c = 2486.4(4) pm. Y = 0 = 90".
'; = 1 2 0 , V = 2333.0 x lo6 pm3; space group R3 (No. 148). 2 = 3 ,
ji(MoKJ = 8.8cm-I. 28,,,, = 54'; 1136 reflections, 1047 with I > 2u(l),
Ni,P,C anisotropic, H isotropic (84 parameters). R , = 0.026, R , = 0.033.
S = 1.28.-3:latticeconstants(200K):a= 1163.1(9),h= 1181.5(8),r=
1880.0(9)pm, a = 107.75(3), fi = 90.31(3). ;, = 118.66(4)'. V = 2 1 2 3 . 2 ~
lo6 pm3: space group P i (No. 2), 2 = 2, p(MoK,)= 23.3 cm-I,
20,,, = 54 , 9729 reflections refined. 7934 with I > 2 u ( 0 Ni,P,C.O anisotropic. C atoms of the C,H, groups isotropic, H of the CH, groups
refined as rigid groups (461 parameters). R , = 0.039, R , = 0.044.
S = 0.90.-Further details of the crystal structure investigations are available on request from the Fachinformationszentrum Karlsruhe,
Gesellschaft fur wissenschaftlich-technische Information mbH. D-W-7514
Eggenstein-Leopoldshafen 2 (FRG), on quoting the depository number
CSD-55935. the names of the authors. and the journal citation.
[7] D. Fenske, J. Ohmer, J. Hachgenei. K. Merzweiler. Angew. Chem. 1988,
100, 1300. Angew. Chem. lnt. Ed. Engl. 1988, 27. 1277.
[8] R. A. Jones, M. H. Seeberger, B. R. Whittlesey, J. Am. Chem. SOC.1985,
107. 6424.
[9] J. J. Daly, J1 Chem. Sac. 1965, 4789.
[lo] J. Hahn. M. Baudler, C. Kruger, Yi-Hung Tsay. 2. Naturforsch. B 1982.
37, 797.
[ l l ] L. D. Lower, L. F. Dahl, J. Am. Chem. Sor. 1976, 98, 5046.
[12] The calculations were carried out with a Workstation IBM 6000/320 and
the program system TURBOMOLE (R. Ahlrichs. M. Bdr, M. Haser. H.
Horn. C. Kolmel. Chem. PIzys. Lett. 1989. 162, 165). The following basis
sets were used: Ni, a Wachters basisoftype (14s,llp,6d)/[8~,7p,4d]extended by (2p.ld) ( A .J .H . Wachters, J. Chem. Phys. 1970. 52. 1034); P,
(1 1~.7p,ld)/[6~,4p,ld].
C(P-bound): (8s,4p)/[4s,2p], from: S. Huzinaga,
Approximate Atomic Wavefuncrions, Department of Chemistry Report,
University ofAlberta, Edmonton, Alberta, Canada, 1971; remaining C,H:
minimum basis (9s,3p)/[2s,lp], and (3s)/[ls], respectively. which were completely energy-optimized for neopentane.
[13] The reaction Ni(3d") + 4 -1 is 168 kJmol-' exothermic in the SCF approximation. MP2 calculations for the corresponding Me-compounds
gave a further stabilization of ca. 613 kJmol-' by correlation effects.
which should also apply in the case of the tBu compound. Since the
3F(3d84s2)ground state of Ni is 1.83 eV (176 kJmol-I) more stable than
the 'S(3d") state (A. A. Radzig, B. N. Smirnow: Reference Data on
Atoms, Molecules and Ions, Springer, Berlin 1985, p. 157) the stability of 1
can he estimated to a very rough approximation to be 600 kJmol-'.
[I41 R. S. Mulliken, J. Chem. Phys. 1955.23, 1833; C. Ehrhardt, R. Ahlrichs.
Theor. Lil. Chim. Acta 1985. 68. 231, and references cited therein.
(151 M. R. A. Blomberg, U. B. Brandemark, P. E. M. Siegbahn, J. Wennerberg, C. W. Bauschlicher. Jr. J. Am. Cliem. Soc. 1988, 110. 6650.
Dynamic Forcing, a Method for Evaluating
Activity and Selectivity Profiles of RGD
(Arg-Gly-Asp) Peptides
By Gerhard Miiller, Marion Gurrath, Horst Kessler,*
and Rupert Timpl
A detailed knowledge of the receptor-bound conformation of substrate molecules and of the complementary receptor surface is of primary importance for rational drug design.
[*I Prof. Dr. H. Kessler. DipLChem. G. Muller, DipLChem. M. Gurrath
Institut fur Organische Chemie der Technischen Universitlt Munchen
Lichtenbergstrasse 4, D-W-8046 Garching (FRG)
Dr. R. Timpl
Max-Planck-Institut f i r B~ochemie,Martinsried (FRG)
[**I This investigation was supported by the Fonds der Chemischen Industrie
(doctoral fellowship for GM), the Studienstiftung des deutschen Volkes
(doctoral fellowship for MG), and the Deutsche Forschungsgemeinschaft.
326
0 VCH Verlagsge.seflschaftmbH.
W-6940 Weinheim, 1992
Provided that the three-dimensional structure of a receptor
or of the receptor-substrate complex are unknown, investigations of conformationally restricted model peptides containing biologically relevant recognition sequences offer a
straightforward approach to characterize the requirements
for binding.[']
The tripeptide sequence RGD (Arg-Gly-Asp) was identified as a universal cell-recognition sequence within several
extracellular matrix proteins. These proteins bind to
transmembrane cell surface-receptors, the so-called integrins, and are involved in a variety of physiologically important processes like cell differentiation, platelet aggregation,
and tumor metastasis.[2.31
We have synthesized several cyclic RGD peptides to study
the conformational influence of the RGD sequence on the
specificity of different integrin-RGD peptide interactions.
In all peptides the same lead sequence was fixed in different
conformations by variation of the chirality of selected
residues. The peptides show a wide variety of biological activities in inhibition assays of tumor cell
Of primary interest are the cyclic pentapeptides cycle(-Arg-GlyAsp-D-Phe-Val-) = c(RGDFV) (1) and cycZo(-Arg-GlyAsp-Phe-D-Val-) = c(RGDF_V) (2). Compared to the prototype linear standard peptide GRGDS and the linear parent
sequence of 2, both peptides inhibit tumor cell adhesion,
mediated by the laminin fragment P1, with activities which
are two orders of magnitude higher (Fig. 1).
HBL-100 cell line
peptides
dEGDFV)
cWDFV)
c(RGpFV)
4RGD-W 1
0.1
0.1
c(RGDFV) 2
30
470
42
0.01
0.1
1.o
IC,
-
10
100
Fig. 1. IC,,
M]values of pentapeptides against laminin P1 (hatched) and
vitronectin (tilled) mediated tumor-cell adhesion. The D amino acid is printed
in italics and underlined.
The conformation of the RGD sequence within the cyclic
peptides is therefore closely related to the structure in the
receptor-bound state. The increased activity of the cyclic
peptide 2 compared to the linear precursor peptide RGDFV
is due to a smaller loss of conformational entropy of cyclic
compounds during binding.
Peptide 1 also inhibits vitronectin-mediated tumor-cell
adhesion in a submicromolar range, but 2 is less active there.
This demonstrates for the first time that enhanced selectivity
can be achieved by conformational restriction of RGD peptides. In the following discussion, this biological observation
will be correlated with structural features of the peptides by
using molecular dynamics (MD) techniques.
The analysis of the conformations of peptides 1 and 2
results in identical structures for the common PII'y turn motif. However, the position of the RGD sequences within the
underlying structural templates is different in these two compounds. In peptide 1 the RGD sequence forms a tight y turn
with Gly in the central position. The Arg and Asp side chains
0570-0833/92j0303-0326 3 3 .5 0 f .25/0
A n g w . Chem. Inr. Ed. Engl. 31 (1992) No. 3
are oriented almost parallel to one another on the same side
of the peptide ring. As expected the D residue (Phe) occupies
the i + 1 position of the PII’ turn. In contrast the essential
Arg and Asp side chains point in opposite directions in peptide 2, so that the guanidino and carboxy functionalities are
separated by a larger distance (Fig. 2).
To verify this assumed receptor-specific transition we employed the method of dynarnicforcing.t61Dynamic forcing is
a restrained molecular mechanics technique, in which a dihedral restraining is superimposed onto molecular dynamics
simulations by extending the force field with an additional
harmonic potential function that forces selected torsions into defined target values. For the simulation of 1 the mainchain dihedral angles of peptide 2 served as target values.
Peptide 2 was forced to change its conformation to a structure related to that of 1. By evaluating energetic parameters
and by scaling the restraining force constant, dynamic forcing gives an impression of possible transition pathways.
Peptide 1 changes its conformation to the desired target
structure of 2 during the initial phase of the 30ps simulation
(Fig. 3, center). The structural analysis of the trajectory
360
I
E
E
G
210
0
30
tlpsl-
Fig. 2. Stereoplot of the NMR-derived conformations of c(RGD_FV)1(above)
and c(RGDFV) 2 (below). Oxygen atoms are filled, nitrogen atoms are shaded,
and carbon atoms are shown as open spheres. Only hydrogen atoms of polar
atom groups are shown explicitly. Both conformations are oriented so that the
PI[’ turn is located in the upper part of the molecule. The hydrogen bonds are
indicated as dotted lines.
Based on the results of the inhibition experiments we propose the structure of 2 as the bioactive conformation for the
laminin PI receptor. The significantly different conformation of l is thought to be related to the receptor-bound
conformation for the vitronectin receptor. This is sufficient
to explain the interaction of the c(RGDFz)-laminin PI receptor and the c(RGDFV)-vitronectin receptor, but not the
nonselective binding of the more active peptide 1 to the
laminin P1 receptor.
Valuable information was obtained for solving this contradiction from careful examination of the MD simulations.
For 1 an unexpected hydrogen-bond pattern results from a
70ps trajectory of a M D simulation in a dimethyl sulfoxide
(DMSO) solvent
Peptide 1forms the predicted hydrogen bonds between Arg’NH and Asp3C0 and between
Asp3NH and Arg’CO, and surprisingly an additional pair
which is similar to that determined for peptide 2 (Table 1).
iL
290,
1
30
0
I
210 I
30
t[psl-
t Ips1
-
Fig. 3. Analysis of simulations with dynamic forcing for c(RGDFI/) 2 (increase of the force constant is marked by arrows) and c(RGD_FV) 1 of 0-3Ops
as well as for 1 of 30-6Ops. E in kcal mol-’.
shows stable intermediates as “snapshots” of the transition
(Fig. 4). The peptide reaches a PII’y, structure via the already
populated shifted PIy, conformation. The inverse y turn (y,)
then changes into a regular y turn, so that the PII’y target
conformation, identical to that determined for 2 (Fig. 2, bottom) is adopted. The interconversion of the yi into the y turn
is a sterically hindered process as indicated by the energy
barrier around 5-7ps (Fig. 3, center).
Table 1. Analysis of hydrogen-bond patterns over the last 70ps of the loops
M D simulations of c(RGDFV) 1 and of c(RGDF_V) 2. rOA= donor-acceptor
distance, OonA = donor-proton-acceptor angle.
1
2
Donor
Acceptor
r,, [pm] OD,,, [“I
Population [%] Turn
Arg’NH
Asp3NH
Gly’NH
Phe4NH
Gly’NH
Phe4NH
Asp3C0
Arg’CO
Phe4C0
Gly’CO
Phe4C0
Gly’CO
341
294
348
321
160
139
135
134
41
85
20
37
gr shift
7, shifi
313
283
161
142
98
95
PII’
i
PI1
This fact indicates a second accessible conformation for peptide 1, a finding which could not be extracted from any
experimental data.
Angea. Chem. Int.
Ed. Engl. 31 (1992) No. 3
C’ VCH
Fig. 4. Monitored intermediates of 1 with Stable turn structures during the
simulated conformational transition.
Peptide 2 is not able to adopt the target conformation
under the same simulation conditions, although the force
constant for the restraining term was increased. This is the
reason for the stepwise increase in the potential energy profile for 2 in Figure 3 (left). Only energetically unfavorable
conformations were observed.
We have extended the simulation of peptide 1 to 60ps,
changing the applied target torsion values after the first 30ps
Vedagsgesellschuft mbH, W-6940 Weinheim. 1992
0570-0833/92/0303-0327 $3.50+ ,2510
327
60
in such a way that the peptide is forced back into its initial,
NMR-derived conformation. Again the peptide changes to
the target structure in the initial phase of the simulation
(6ps) (Fig. 3, right). It is obvious that the NMR-derived
conformation is the energetically favored structure, but the
conformational transition induced by the laminin PI receptor is easy to perform.
The positioning of the essential RGD sequence in conformationally restricted molecules that are defined by a rational
design principle led to highly active and selective compounds. The combination of biological data and conformational characteristics allows the definition of conformationactivity relationships and furthermore implies a dynamic
model of receptor-substrate interactions. Hence, from these
computer simulation techniques it was possible to explain
the biological profile of RGD peptides.
From the recently published conformational analysis of
two native RGD-containing proteins, kistrin['] and echistatin,['] members of the disintegrin family, it was not possible to derive a bioactive RGD conformation, because the
RGD sequence is located in both proteins in exposed and
highly flexible loop regions. Our study of conformationactivity relationships with an underlying dynamic aspect
demonstrates the importance of taking into account conformational changes of substrate molecules during receptor
binding. The receptor exerts a conformational pressure on a
substrate, enabling different substrate conformations to be
induced, depending on the receptor environment.
Experimental Procedure
For structure refinement MD simulations were performed with the GROMOS
force field [9] and NMR-derived interproton distances as constraints. The initial structure was built up manually so that no secondary structure elements
were preformed. After energy minimization the structure was simulated by MD
for 2ps at 1000 K, 3ps at 500 K, and further 5ps a t 300 K in vacuum with a
distance-restraining force constant of 4000 kJmol-' nm-'. The obtained conformation served as the initial structure for a 15Ops solvent simulation in a
DMSO solvent box [5) with a box length of approximately 3.5 nm containing
about 150-180 DMSO molecules. After the first 70ps MD the distance restraining force constant was reduced from 1000 kJmol-'nm-* to
500 k J m ~ l - ' n m -forfurther30ps.
~
Thelast 50psofthe 150ps trajectory were
simulated without any restraint to make sure that the obtained conformation
after loops restrained M D was a low-energy structure.
For the simulation of the conformational transitions of 1 and 2 the CVFF
(Consistent Valence Force field) was employed which is implemented in the
DISCOVER (BIOSYM) program package [lo]. The force field was extended by
a harmonic potential function for the dihedral-angle restraints. These simulations were performed in vacuum over 30 ps. The dihedral-angle restraining
force constant was set for the first 7 . 5 at~ 1~kcalrad-2 and was increased in
7 . 5 ~ steps
s
to 2.5 kcairad-*. 5 kcalrad-'. and 10 kcalrad-*.
Received: October 15, 1991 [Z 4969 IE]
German version: Angew. Chem. 1992, 104. 341
CAS Registry numbers:
1, 137813-35-5; 2, 137894-01-0;GRGDS, 96426-21-0
[I] H. Kessler, Angen. Clirm. 1982. 94,509-520; Angev,. Chem. fnt. Ed. Engl.
1982, 21, 512-523.
[2] M. D. Pierschbacher, E. Ruoslahti, Nuture 1984, 309, 30-33.
(31 S. E. D'Souza, M. H. Ginsberg, E. F. Plow, Trends Biochem. Sci. 1991, 16,
246 - 250.
[4] M. Aumailley, M. Gurrath, G. Muller. J. Calvete, R. Timpl, H. Kessler,
FEBS Lert. 1991, 291, SO- 54.
151 D. F. Mierke. H. Kessler, .
I
Am. Chem. Sor., 1991, 113, 9466-9470.
[6] R. S. Struthers, G. Tanaka, S. C. Koerber, T. Solmajer, E. L. Baniak,
L. M. Gierasch, W. Vale, J. Rivier, A. T. Hagler, Proleins: Srrurt. Funcr.
Gene/. 1990, 8, 295-304.
171 M. Adler, R. A. Lazarus, M. S. Dennis, G. Wagner. Science 1991. 238,
491-497.
[8] V. Saudek, R. A. Atkinson. J. T. Pelton. Blochemistry 1991,30.7369-7372.
[9] W. F. van Gunsteren, H. J. C. Berendsen. Groningen Molecular Simulution
( C R O M O S ) Library Munuul Biomm, Groningen, 1987.
[lo] P. Dauber-Osguthorpe, V. A. Roberts, D. J. Osguthorpe, J. Wolff, M. Genest, A. T. Hagler. Pro/eins: Struct. Funcr. Genet. 1988, 4, 31-47.
328
51 VCH Verlugsgesellschufi mbH, W-6940 Weinheim,1992
Selective Inhibition of Trypanosomal
Triosephosphate Isomerase by a Thiopeptide **
By Horst Kessler,* Hans Matter, Armin Geyer,
Hans-Jiirgen Diehl, Matthias Kock, Guido Kurz,
Fred. R. Opperdoes, Mia Callens, and Rik K. Wierenga
Dedicated to Professor Gerhard Quinkert
on the occasion of his 65th birthday
Although much effort has been invested in the study of
sleeping sickness its control has not yet been achieved."] The
germs causing this tropical disease (Trypanosoma brucei brucei) obtain their energy exclusively from glycolysis. Thus, the
synthesis of selective inhibitors of glycosomal enzymes of the
pathogen is a promising approach.[21In our investigations of
the inhibition of trypanosomal triosephosphate isomerase
we have
(TIM), EC 5.3.1.1), a key enzyme in this
developed several cyclic hexapeptides which have shown remarkable effectiveness and selectivity.[41Several determinations of the crystal structure of TIM are known,r51but the
binding site of peptide inhibitors has not been localized until
now.
Here we report the improvement of biological activity of
a cyclic peptide, which was achieved by conformational
changes in the peptide backbone. All side-chain functionalities were unmodified. The selective introduction of a thiocarbonyl group into the cyclic hexapeptide cyclo(Gly'-Pro2Phe3-Val4-Phe5-Phe6) induces a drastic change in the
backbone
resulting in a remarkable enhancement in the selective inhibition of trypanosomal TIM.
The OjS exchange (thionylation) within conformationally restricted peptides can take place with high regioselectivitY,16.71
The introduction of a thiocarbonyl group as a modification of a peptide bond is an isosteric replacement of a carbonyl group. The local geometry in the vicinity of this fragment is scarcely influenced.[" However, by the selective
exchange of a CO-NH unit by its sulfur analogue the formation of transannular hydrogen bonds within the peptide can
be influenced. The C=S fragment exhibits weaker acceptor
properties, and due to its enhanced acidity the neighboring
amide proton is a better donor than the 0x0 analogue. In
addition, the larger van der Waals radius of the sulfur and
the longer C=S bond limit the {cp, $} conformational space
in a characteristic manner: The thiocarbonyl fragment can
be positioned at the i + 1 and i + 2 positions of a PI turn but
only at the i + 2 position of a PI1 turn. The unfavorable
steric interaction of the sulfur with the side chain of the
L-amino acid in the i + 2 position prohibits the adoption of
the PI1 turn."]
The thionylation of the peptide described here causes a
remarkable increase of its inhibitory effect on trypanosomal
TIM. Although the cyclic oxopeptide cyclo(Gly'-Proz-Phe3VaI4-Phe5-Phe6)has only weak biological activity (Table l),
the analogous thiopeptide cyclo(Gly1-Pro2-Phe3-Val4[*I Prof. Dr. H. Kessler. DipLChem. H. Matter, Dip].-Chem. A. Geyer.
DipLChem. H. J. Diehl, Dipl.-Chem. M. Kock, Dip!.-Chem. G. Kurz
Organisch-chemisches lnstitut der Technischen UniversitPt Munchen
Lichtenbergstrane 4, D-W-8046 Garching (FRG)
Prof. Dr. F. R. Opperdoes, Dr. M. Callens
International Institute of Cellular and Molecular Pathology
Avenue Hippocrate 75. 8.1200 Brussel (Belgium)
Dr. R. K. Wierenga
European Molecular Biology Laboratory
Meyerhofstraae 1, D-W-6900 Heidelberg (FRG)
[**I This work was supported by the Fonds der Chemischen Industrie, the
Deutsche Forschungsgemeinschaft, and a BRIDGE project of the EEC.
H. M. and M. K. thank the Fonds der Chemischen Industrie for a grant.
0570-0833/92j0303-03283 3.50+ ,2510
A n g e w Chen?. Int. Ed. Engl. 31 (199.2) No. 3
Документ
Категория
Без категории
Просмотров
0
Размер файла
485 Кб
Теги
profiler, forcing, selectivity, rgd, evaluation, method, activity, asp, dynamics, gly, peptide, arg
1/--страниц
Пожаловаться на содержимое документа