вход по аккаунту


Design of Novel Nonpeptidic Thrombin Inhibitors and Structure of a ThrombinЦInhibitor Complex.

код для вставкиСкачать
Design of Novel, Nonpeptidic
Thrombin Inhibitors and Structure
of a Thrombin-Inhibitor Complex**
Tyr 6 0 A
Ulrike Obst. Volker Gramlich, FranFois Diederich,*
Lilt7 Weber." and David W. Banner
i o PI-ofossor. A l i w t E~c~lieni~ioscr
011 //w oc.c.rr.vio/i of' his 7Ol/i hir.[h(lriJ,
I n the context of our work on host-guest interactions. we
have developed receptors for various substrates."] The
avnilabilit> of high resolution X-ray structures of enzymes now
ullows the reverse process of rational design of inhibitors with
the aid of molecular modeling. To this end, instead of using the
enzqme's substrate a s model, a small molecule, complementary
to the active site with respect to hydrogen bonding as well as
ionic and dispersive interactions, should be designed dc, 1 7 0 1 ~ .
EnLymes that are best suited for the rational design approach
are those with rigid. well-defined binding pockets in the active
site. Thrombin. 21 trypsin-like serine protease[21with a central
role i n the blood coagulation process.[31 meets these requirements. The natural substrate for thrombin is tibrinogen. which
is dissolved in the blood plasma. The enzyme selectively cleaves
the peptide bond next to an arginine residue to form fibrin.
which can polyinerize. Thrombin is also the main activator of
platelet aggregation. A selective inhibition of thrombin has potential therapeutic application in thrombotic diseases.
A number of X-ray structures of thrombin-inhibitor complexes were ;I\ ailable,[41which allowed the study of the important interactions crucial for inhibition. The active site of
thrombin features three binding pockets (see Fig. 1). The bottom of the recognition pocket S1 contains Asp189 for interactions with guanidinium or ammonium groups of arginine or
lysine residues. The large hydrophobic pocket (D-pocket, signifying ii pocket "distal" to the catalytic site'lb1) is preferentially
occupied by aretie rings, which can undergo C H . . TI arenearene inicractions with the indole unit of Trp215. Fibrinogen.
the natural substrate for thrombin, places the benzyl group o f a
phenylalanine residue in this pocket. An additional smaller hydrophobic pocket is positioned close to the catalytic site ("proximal": P-pockctl"hl). I n the thrombin-fibrinogen complex this
pocket is occupicd by ;I valine residue. It was our goal to develop
lic template to which three side chains could be
attached in order to reach each of the three binding pockets
(hydrophobic hide chains for the P and D pockets, and an aromatic unit with ii basic group for the recognition pocket S1 of
the e n q i n e ) . The template should be rigid to avoid hydrophobic
collapse of the side chains'b1and to avoid the loss of conformational degreeh 01' t'reedom on coinplexation to the enzyme. In
addition. the template should contain a hydrogen-bond acceptor (for exainplc. a carbonyl group) to form a hydrogen bond to
A s p 189
Fig I . Schematic representation of inhibitor 1 energy-niiiiimi/rd in the iiclive site
of Ili roin hi n .
the amide NH of Gly216, a feature present in all known
thrombin-inhibitor complexes. The synthesis should be simple
and short and allow variation of the side chains. With such a
synthesis, affinity and selectivity can be subsequently optimized
through combinatorial solid-phase synthesis. The design of the
target molecule was undertaken with the programs Insight I1
and Discover;['] the force field used was the CVFF. A result of
the modeling was molecule 1 (see Fig. 1). It has no structural
similarity to either the natural thrombin substrate, fibrinogen,
or other known synthetic thrombin inhibitors. This molecule
has a series of potential interactions with the enzyme (Fig. 1): I )
a salt bridge between the amidinium group of the inhibitor and
the carboxylate of Asp189 in the recognition pocket S1 of the
enzyme. 2) two hydrogen bonds of the NH of Gly216 and the
OH of Tyr60A to the carbonyl groups of the succinimide
residue, 3) edge-to-face interactions between the indole ring of
Trp215 and the benzyl group of the inhibitor in the D-pocket of
the enzyme, where the benzyl group experiences additional benefits from the neighboring hydrophobic residues Ilel74 and
Leu99 (not shown in Fig. I ) , and 4) contacts between two
methyl groups of the inhibitor and the hydrophobic P-pocket of
The synthesis of 1 (Scheme 1) began with a 1.3-dipolar cycloaddition in a three component reaction. An azomethine ylide
was generated in situ from 4-bromophenylalanine (2) and acetone (3) with decarboxylation and treated with N-benzylmaleimide (4). The pyrrolidine nitrogen atom of the cycloadduct (5) was methylated (-+6), and the bromo substituent
converted into a cyano group. The conversion of the nitrile 7
into amidine 1 was achieved with a Pinner reaction.["'
In the preparation of the bicyclic templates, i t is iilso possible
to make shorter basic benzamidinium needle^"'^'' lacking the
methylene group, by using 4-cyanobenzaldehyde in the 1,3dipolar cycloaddition (Scheme 2). Compounds l l a m are more
easily prepared than the analogous cycloadduct 5. In addition,
their synthesis allows a larger structural variation of the "upper" side chains, a significant feature with respect to the planned
Scheme 1. Synthesis of the designed thrombin inhibitor 1 . BS racemate. a) toluene.
reflux, 68 h. 2 5 % : b) HCHO. HC0,H. 100 C, 10 h. 81 %; c ) CuCN, D M F . reflux,
56 h , 6 3 % ; d ) H C I , MeOH. 4"C,48 h:ejNaHCO,; f)NH,CI,MeOH,'H,O.65"C.
3 h, 57%.
9a, g, i, I, m
I l a 71 % j 3 7 %
d 65 Yo k 27 /o'
g 61 Yo I 34%
conibinatorial synthesis. The structure of the major
diastereomer formed in the cycloaddition was determined by
X-ray crystallography of nitrile 12b. The product was formed
from an (inti-ylide through an endo transition state'''' and was
shown to have a different conformation in the crystal['21(Fig. 2)
than is required for the binding of the final product 14b to
thrombin. In the structure of 12b an intramolecular CH . ' . x
interaction and a different conformation of the pyrrolidine ring can be
observed. At any rate,
the structure of 12b in
the crystal does not
about the conformation
of amidine 14b either in
solution or in the active
site of thrombin.
The originally proposed compound 1 was
found to have a K, value
Fig 2. Crqstal structure of 12b.
for thrombin of 18 VM
as a racemic
(for comparison: benzamidine has a K, value of 300 pM[4c1).
Compounds 14a-m, lacking the CH, group in the benzamidinium needle, were significantly more active (Table 1). These cotnpounds are conformationally more restricted than 1 because of
the rigid framework and the attachment of the needle. and are
therefore excellent candidates on which to study structure-activity relationships. These compounds can all be expected to
follow the same binding mode.
i 32% m 17%
Table 1. Activities o f the thrombin inhibitors 14a-m and selectivities with respect
to complexation to trypsin.
83 %
f 77%
K, [pM]
Selectivity [a]
1 .I
I .5
1 .0
B ii
piperonyl [h]
piperonyl [b]
[a] k',(trypsinj;K, (thrombin). [b] Piperonyl =&
14a a7 yo
b 70 o/o
d 55%
e aa
f 60 Yo
g 70%
h 89%
i 86 %
j 40%
k 69 70
1 a3
m 35
Scheme 2. Synthesis of the henzdmldinium chlorides 14a-m as andlogs of 1 . a)
D M E 8O"C, 5 h ; b) HCHO, HC0,H. IOO-C, 10 h , c) Ac20. pyridine. 0 C to room
temperature. 2 h; d) HCI, MeOH. 4°C. 48 h; e) NH,. MeOH. 6 5 ' C . 3 h. For the
residues R ' - R 2 see Table 1. Racemic mixtures are formed at all stages
On the pyrrolidine nitrogen on the right in the depicted structural formula, three structural modifications were evaluated.
The compounds with a methylated nitrogen atom (l4b,e,h)
showed increased activity relative to the unmethylated ones
(14a,d,g), whereas a dramatic reduction of activity was observed with acetylated compounds (14c.f). A possible explanation is that acetylation leads to a conformational change in the
bicyclic template. Another possibility is that the pyrrolidine is
preferentially bound in the protonated state: with acetylation
protonation is no longer possible. The highest thrombin affinities (lowest Ki values) were achieved with the addition of a
five-mcmbered ring on the right side of the molecule (14i-k). A
six membered ring led to a reduction in affinity (141), probably
becausc this large hydrophobic moiety displaces a water molecule close t o the catalytic group Ser195. A thiazolidine ring
(14m) i s probably also too big. The activity is less influenced by
the substitileiits on the succinimide nitrogen atom for the
"aryl binding pocket" (D-pocket) . Butyl, benzyl, and cyclohexylmethyl substituents give rise to 110 large variations. The
highest activity was achieved with a benzodioxol-5-ylmethyl
( = piperonyl) substituent (14j, K , = 90 nM).
Thc binding mode of the currently most active compound 14j
was revealed by X-ray crystallography" 'I of its complex with
thrombin. The inhibitor was bound in the active site of
thrombin in the predicted fashion (Figs. 3 and 4). As expected,
based on thc rigidity of the molecule and the molecular modeling, only the (3aS,4R,8aS,8bR) enantiomer of the racemic mixture was found in the crystal structure. The benzamidinium side
chain binds well in the recognition pocket S1 of the enzyme.
Hydrogen bonds are formed to the carboxylate group of
Asp1 89, to the carbonyl group of Gly219, and to ;I water molecule present at the bottom of the pocket. One carbonyl group of
the inhibitor acts as a hydrogen bond acceptor for the NH of
Gly216, whereas the other is located in the ensyme's P-pocket.
The move of this carbonyl group from aqueous solution to the
small hydrophobic P-pocket of thrombin must cost desolvation
energy. This energy is not required for the binding of the inhibitor to trypsin, because it lacks the loop which in thrombin
forms the P-pocket. This could explain why this class of inhibitors is only slightly more selective (up to 20-fold) for
thrombin than for trypsin. Another hydrogen bond is formed
between TyrhOA and an oxygen of the piperonyl group. The
aromatic portion of the piperonyl group is located in the Dpocket of the enzyme and is involved in the evpected CH . . . n
interaction with Trp21.5.
This inhibitor is still not optimal, especially with respect to its
selectivity. Nevertheless, our results show that with a rational
approach it is possible to develop new lead structures for enzyme inhibition in a relatively short time.
The synthetic strategy with 1,3-dipolar cycloaddition provides a large potential for structural variation. Most a-amino
acids and several carbonyl compounds are feasible reagents to
generate the azomethine ylides." ' I Most electron-poor olefins
or acetylenes are conceivable dipolarophiles. Investigations involving the 1,3-dipolar cycloaddition with azomethine ylides on
a solid phase are in progress. We hope that these combinatorial
methods on a resin will lead to thrombin inhibitors with improved activity, selectivity, and good pharmacological properties.
E.vperinwntu1 Procedure
I l j : A mixture of proline (1.15 g, 10 mmol), 4-cyanohenraldehqde (1.31 g,
10 mmol). and N-piperonylmaleimide (2.31 g. 10 mmol) w a s heated i n DMF
(10 mL) for 5 h a t 80 C. The solvent was removed under high v:icuuin. and
the residue purified chromatographically on silica gel (hexane:AcOEt:NEt
49 S:49.5:1) Yield 3 7 % ; colorless needles (recrystallired from methanol), m.p.
176-179 C.
Asp 189
Fif. 3. Scheni.itic ireprescntation ofthe interactions of 14j v.ith thrombin according
to thr crv5t;il \ ~ i ~ i c t u r e .
14j. Dry HCI was passed through a solution of 1 Ij (831 nig, 2 mmol) in dry CHCI,
CH ,OH S: 1 (6 mL) a t 0 C for 10 min The rcaction mixture u'as kepl fur two days
:it 4 C . After addition of Et,O a precipitate was formed. which \\as dried and mixed
with a NaHCO, solution ( 5 %. 8 mL) and CHCI, (20 mL).Alicr ii quick extraction
the aqueous layer again cxtracted tu,ice w i t h CHCI,. The organic phase was
dried over Na,SO,. and the solvent was distilled off. The rc\idue was dissolved in
MeOH (7 mL) and NH,CI (I50 mg) in H Z O(1.5 mL) was addcd. The mixture was
stiri-ed ill 65 C for 3.5 h. After cooling. the solvent &as distilled off, and the residue
was yeparated chroniatofraphicully (RP18 silica
gel. gradient from H,O to M e O H ) . Yield 4 0 % ;
colorless powder. m.p 177 181 C. 1R (KBr):
t = 1 7 0 1 ( C = O ) . 1678 l ( ' = N ) c m - ' ; ' H N M R
(250 MHz. (CD,)$O): 6 = I 68 (ni. 2 H ) . 1.98
(m, 2H). 2 45 (in. 1 HI. 2 77 (rn. 1 H). 3.47 (d,
J =7.8 HL, 1 H ) . 3.56 (111. I H). 3.76 (t. .I =
8 . 3 H L . 1 H). 4.22 (d. J = X.5 Hz. 1 H ) . 4.33 (s,
7H). 6.02 ( 5 . 2 H ) . 6.67. 0.XX (AB. J =7.X Hr,
? H I . 6.70 (s, 1 H ) . 7.44. 7.71 ( A B , J = 8.2 Hz,
4 ~ ) 9.m
(s, 2 ~ ) Y. ?t; ( \ . ? H I : 1 % N~ M R
(62.5 MHL. (CD,),SO) n = 23 0. 29.0, 41.2.
48.6, 50 3. 67.4. 67.5. in1 (1. i o x . ~ .inx.7. i?1.2.
126.5. 121 5. 178.7. 129.7. 145.6. 146.6, 147.2.
165.4, 17S.4. 17X 1 ; ESl(eleclrospray)-MS: in;:
433.4 ([M + HI'. 100). \ntirfactory nnal. for
Received. ApiiI 12. 1995 [Z7X841E]
. l i m 1995. 107.
Gerinan version: A i i w ~ r C
1874 1x77
Fig 4. Stci-c[irepi-~scnt;Itionof [he difference electron denutieq of 14j (dashed linec), thrombin (thin lines). and a
model 0 1 the inhihitor (thick lines) [15]. The very small peaks represent water molecules or small changes of the
prolein comp;ircd to tlic (unrefined) model. CornDarative c a l c ~ l a t i ~showed
rhat the electron densities o f the
cwhonyl group\ mu\t he wcak
this resolution due to truncation of the Fourier series.
Keywords: azomethine ylides . cycloadditions . de novo design . enzyme inhibitors . thrombin
[I]a) B. K.Peterson. F Dicdericli. A t i p w . i % ~ i i 1994.
/Oh. 16XX 1690. Airgrit..
C/IOIII h i / . Ed. EiigI. 1994. 33. 1625-16X; b) V Alcarar. E Diedcrtch. i/>id.
1992. l0-f. 1503 1505 and 1992. 31. 1571-1523: c) F. Dicdericli. i : i ~ / o p / i ~ i t i ~ ~ s .
Thc Royal Society of Chemistry. Cainbridse. 1991
[2] I V. K u r i n o \ . R . W. Harrison. S r r ~ /Biol
1994. 1. 735 743.
[3] M . C.E. r a n Dan-Mierar. A. D. Muller in Blood ~ ' m i ~ i i / ~ i / (Eds.:
K F. A.
L\vanl. H C. Hrmker). Elsevier. Aiiisterdaiii, 1986. pp. 1 11.
[4] a ) W. Bode. 1. Maqi-, U . Bautiiann. R. Hubcr, S. R. Stone. .I Holsteengc.
The homogeneous catalytic hydrogenation of carbon dioxide
EMBO J 1989. H. 3467 3475; h) D. W H;uiiier. P. H;idvdry. J Biol C / i m
into formic acid has been the subject of recent experimental
1991. 266, 20085 -20093; c ) K Hilpert. J. Ackerm;uin. D. W Bmner. A Gnct.
Metathesis as a Critical Step for the Transition Metal Catalyzed Formation of Formic Acid
from C 0 2 and H,? An Ab Initio Investigation**
Franqois Hutschka, Alain Dedieu,*
K . Gubcrnator, P. Hadsiiry. L. Labier. K Muller. G. Schinid. T R. Tschopp,
H . v a n de Waterhccmd. J Mw/. Clici?~.1994. 37, 38X9--3901
[ 5 ] Chyinotrypsinogrn numbering [4a]
[6] ;I) D. H. Rich in Prrs/ic~i.tii,cc
111 , ~ ~ ~ ~ / ; ~ i i i ~ i / i ' / i ~ ~ i i i iB
. sTcat;~.
/ ~ t ~ (W
E dFuhi-er.
E. Kqburz. R.Giger). Verlag Helvetica Cliiinica Acta. Basel.VCH. Weinheim
1993, pp 15 2 5 : b) R . Hii-cchminn. K . C Nicokiou. S. P i e t ~ i n i c o .E. M
Lecihy, J Salvino. R. Arison. M . A. Ctchy. P. G Spoors. W. C Shakespeare.
P.A . Sprengeler. P. H;imle\. A. B. Snuth Ill.T. Reisinr, K . Raynor. L M:icchler. C. Donnldson, W Vale. R W.Freidinger. M . R Cascleri, C. D Stl.adcr. .I.
h i . Chriii SOC.1993, 11.5, 12550-12568: c ) P. Y. S . Lam. P. K . .[adha\,. C J.
Eycrtiimn. C. S . Hodge. Y. Ku. L. T Bacheler. J. L. Meek. M . J. Otto. M . M
Kii)ncr. Y N. Woiig, C -H. Chang. P. C. Weher, D. A. Jackx>n.T. K . Sharpc.
S. Ericksoii-Viit;incii. S<i i w c 1994. 3 3 . 3x0 ~ 3 8 4
[7] Insight 11. Version 2.3.0. Snn Diego. Bios)in Technologies. 1993: Discover.
Version 2.9.5. San Diego. Biosyni Technologies, 1994. Modeling procedure:
The proposed inhibitor was separately geometry optirnized a n d docked mto its
expected binding site. The coordinates o f t h e enLynie w
of the inhibitor inside the enzyme \\:is iniiiiiniied.
[XI a ) 0 . Tsuge. S. Kanemasa. M. Ohe. S. Ttikenaka, Bid/. i%eili. SO( Jpi. 1987.
6(J. 4079 4089: b) R.Grigg. S. Thianpatunagul, J. Clirrii. S t x . < ' / i t w i . i'omiiiiiii.
1984, 180 1x1. c ) R. Grigg. S. Sui-cndrnkumar. S. Tliianpataiiagul. D
Vipond. ihid 1987. 47 49.
191 All new compounds gave satisfactor) an;ilqtiwl roiilts.
[ l o ] G . Wagner. I. Wunderlich. P/iorii~a:ie1977. 32. 76 79.
[ I I ] Rcview oil azoinethiiie ylides: 0.Tsnge. S.K:inein;is;i. Arli.. f f r i i v i i ( r d . C / i ( w i .
1989. 4j.231 ~347.
[I?] 12b: C L 2 H S , N I O I .colorle~s plates. crystal diiiiciisionc 0.4 x 0.4 x 0 3 nmi:
,\I, = 359.43: trtclinic. space group Pi. u = 9.352(X). h = 10.004(7!.
~ ~ = I 0 . X X 2 ( 9 ) ~ , a = 9 5 . 5 5 ( h ) . / i = 9 3 3 9 ( 7 ) . ; . = 1 1 0 . 7 3 (1'=041
Sqntcx P3 V diff~ictonieter (Mo,, ~ i d i a t i o n .
Z = 7.
= 1.269 gcin
;= 0.71073 A. graphite iiioiiochroiiiiitor. T = 793 KI, 1885 I-etlecttons iiiciisured. 1767 indepeiidciit reilections. 1365 rellcctions with F > 4 On(F!. The
structure was ~ o l ~ ebyd direct methods [ I 31 and refined by full-matrix.
least-squares analysis (heavy cltoms anisotropic. H atoma fixed isotropically
H positions are based on stei-eochcmtc;il considerations). R = 0.0471. II R =
0.0422. Further details of the crystal structure investigationz may hr obtatnrd
from the Director of the Cambridge Crystallographic Data Centre. 12 U tiion
Road. GB-Cambridge CB2 1 EZ ( U K ) . o n quoting the fiill joiirn;il citation
[I31 Siemens SHELXTL PLUS (VMS), Siemens Analytical Instruments. M;idix~n.
USA. 1986.
[I41 The octi\ity of the throinbin inhibitors wasdrteriniiled ~ c c o r d i n gt o K Lotlenheig. J. A Hall. .I.W Fenton. G. M . .lackson. T/irord? Rr,\. 1982. 28, 313
(chi-oiiiogenic substrate S-2238).
11 51 The dava ( 7 5 % conipletc t o a resolution 01'3 10 A) were tncasured a n d \+torkcd
up a s described i n [4b,c]. For calculation of the differcuce elcctron densities
(Fig. 4) the coordinates for protein and water molecules o f h u m a n thrombiii
~ c r used
(I-elincdt o ii I-csolution of 2 6
see Fig. 4 in ref. [4c]). Llcspitc the
modest quality of the dat:i the binding mode of the inhibitor could be clcarly
idciitified. After refinement of protein and inhibi~orthe cry~t~tllographic
value was 12.9% for 7043 rcilcctions ~ i t hF > q with acceptable geoiiictrq
(st;indard dcvlation [or the boiid length0.012 A aiid for thc hond angles 2.0 I .
The atoins of the inhibitor h m c ;in a~ei-ageB value of 50
higher then the atoms o f t h e protein. Either not all inhibitor sites arc occupied
01- the inhibitor I S present i n several siintl;ir conformation\. h'e have reccntly
coiilirmed this result at higher resoIiition ( 1 . 9 A ) in ti different crysliil tqpe
(i.21, d e t a i l s 01' t h i s hctter dcl'tned s t r ~ i c t i ~will
r e be published clscwhcrc.
studies," 31 and hydridorhodium(1) complexes have been
shown to be very efficient catalysts.['. 'I It is generally believed
that the catalytic cycle starts with the CO, insertion into the
metal - hydrogen bond. a process that is well understood, experimentally and the~retically.[~.
'I The subsequent steps are more
speculative. A sequence involving Rh' and Rh"' intermediates
was put forward in the two most recent examples. In these
either [Rh(nbd)(PMe,Ph),]BF, (nbd = norbornadiene)['] or
[(Rh(cod)(p-H)),]and [{Rh(cod)(/i-CI],] (cod = 1,5-cyclooctadiene) in the presence of the bidendate phosphane ligand
Ph,P(CH,),PPh, were used a s the hydrogenation catalyst.']". ' 'I
I n the former case the catalytically active species is
believed to be the cationic dihydridorhodium(ll1) complex
[Rh(PMe,Ph),(H),(solvent)]+, and insertion of CO, has been
shown to yield a hydridoformatorhodium(II1) complex. Reductive elimination of formic acid and subsequent oxidative addition of H z are assumed to regenerate the dihydridorhodiuin(~~~)
cation. In the second, much more efficient system, thc neutral
species [ (Ph2P(CH2),PPh,)Rh(H)] is believed to carry the catalytic cycle by insertion of CO,. H z would then oxidatively add
to the formatorhodium(r) intermediate and reductive elimination of formic acid from a dihydridorhodium(ii1) intermediate
would lead back to the rhodium(1) compound (path A in Scheinc
1 ) . In thc course of our theoretical studies of CO, coordination
and reactivity.["I wc have started to investigate these catalytic
processes.['] The purpose of this preliminary communcation is
to propose on the basis of our calculations an alternative mechanism (summarized in path B in Schemc 1) in which the rhodium center retains the formal oxidation state I throughout the
reaction. Change of oxidation state is avoided by a ci metathesis
reaction between the H, moleculc and the formate complex.
Ab initio MO,'MP2 calculations were carried out on the cis[HRh(PH,),] + Hz+ CO, model system using the Gaussian 92
program system.[81Geometries were optimized by the gradient
technique at the MP2 level, with the LANLlDZ basis set.18,91
Single-point MP2 energy values, to be discussed here. were obtained with an all electron basis set and with polarization functions on the active hydrogen atoms and on the carbon and
oxygen atoms of CO, . [ I 3 ' We have checked that the use of the
LANLIDZ basis set was appropriate for the geometry opti-
Di- A. Dedieu, F Hutschka
I.;iboratoirc dc C'hiinie Qti;uitique. UPR 13') du CNRS
Universitb Louis Pasteur
4 rue Blaise Pasc;il. F-67000 Srra\bourg (France)
Telefax: In1 codc +xx 61 20 x5
Dr. W. Lcttner' ' I
Max Planck Gescll~chal't. Arheitsgruppe CO1-Chemic
a n dcr Uni\.ersit;it Jen;i (Geriii;iiiy)
N e w addl-es\
Mnr-Plaiick-lii\citut fiir Kohlenforschting. ivlulhciin' Rulir (Gcrin;iny)
Tlii.; woi-k \vas pnrtl! supported by a French Germiin PROCOPE contract
( n o 941 32). Calctilntioii~were carried out o n the IBM RS 6(100 workstations
0 1 our laborntor) and of tlic Institiit du Dr\,cloppenient et de Rcssourcea cn
Iiil'oi-tnntiquc Scientiliqiir (IDRIS, Orsay). ,ind o n the DEC 3001)'600S work\tations 01' the Cenlre Unilcrsit;iire R&pionaI dc Ressoiirces Informariqucs
(CUKRI, Strashourg). We thank the staff of these two ccuters for their cooper,itioii We;irc,ilso griitcl'ul to Ur. L. Pndel and Mi\ Fersingfoi- their help and
teclitiiciil assi\tance.
Без категории
Размер файла
478 Кб
structure, complex, design, inhibitors, nonpeptide, thrombinцinhibitor, novem, thrombin
Пожаловаться на содержимое документа