close

Вход

Забыли?

вход по аккаунту

?

Discovery of Low-Molecular-Weight Ligands for the AF6 PDZ Domain.

код для вставкиСкачать
Zuschriften
Protein–Protein Interactions
DOI: 10.1002/ange.200503965
Discovery of Low-Molecular-Weight Ligands for
the AF6 PDZ Domain**
Mangesh Joshi, Carolyn Vargas, Prisca Boisguerin,
Annette Diehl, Gerd Krause, Peter Schmieder,
Karin Moelling, Volker Hagen, Markus Schade,* and
Hartmut Oschkinat*
Despite their central role in most regulatory processes and
disease mechanisms, protein–protein interactions (PPIs)
remain largely unconquered ground for drug discovery and
chemical-tool generation. In many cases, these interactions
are mediated by protein interaction domains like Src homology 2 (SH2), Src homology 3 (SH3), WW domain, and
postsynaptic density/discs large/zona occludens-1 (PDZ).[1]
PDZ domains may be considered “drugable” because of a
shallow ridge on their surface which is, however, not a proper
cavity. Hence, they are good test cases for the development of
PPI inhibitors.
PDZ domains comprise 90 residues and occur more
than 450 times in approximately 250 human proteins. In most
cases, they recognize the C terminal four to seven residues of
membrane receptors and ion channels,[2–4] although they are
also known to bind to internal sequence motifs.[5] Consisting
of a six-stranded b-sheet flanked by two a helices, PDZ
[*] Dr. M. Schade
Combinature Biopharm AG
Robert-Roessle-Strasse 10, 13125 Berlin (Germany)
Fax: (+ 49) 30-94894-038
E-mail: schade@combinature.com
M. Joshi, C. Vargas, Dr. A. Diehl, Dr. G. Krause, Dr. P. Schmieder,
Dr. V. Hagen, Prof. Dr. H. Oschkinat
Leibniz-Institut f=r Molekulare Pharmakologie FMP
Robert-Roessle-Strasse 10, 13125 Berlin (Germany)
Fax: (+ 49) 30-94793-169
E-mail: oschkinat@fmp-berlin.de
Dr. K. Moelling
Institut f=r Medizinische Virologie
UniversitAt Z=rich
8028 Z=rich (Switzerland)
Dr. P. Boisguerin
Institut f=r Medizinische Immunologie
CharitD-UniversitAtsmedizin Berlin
Hessische Strasse 3–4, 10115 Berlin (Germany)
[**] This work was supported by grants from the European Community
(QLK3-2000-00924-SH2 Libraries) and the Deutsche Forschungsgemeinschaft (DFG) (OS 106/4-2). The authors thank Dietmar
Leitner for assistance in NMR spectroscopic measurements, Arvid
Soderhall and Jens Laettig for their help in molecular dynamics
(MD) simulations, Viviane Uryga-Polowy for the MS measurements,
and Daniel Geißler for helpful discussions. The structure of the
AF6 PDZ domain is deposited in the protein data bank (PDB) under
accession code 1XZ9 and the AF6 PDZ-5 f complex under accession
code 2EXG.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3874
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. 2006, 118, 3874 –3879
Angewandte
Chemie
domains bind cognate peptides in an extended conformation
whereby the terminal carboxy group is hydrogen bonded with
the backbone amides of the highly conserved GLGF loop.[6–8]
The protein AF6 (ALL-1 fusion partner on chromosome 6),
also known as s-afadin,[9] contains one type II PDZ domain
(abbreviated: AF6 PDZ), two N-terminal Ras-association
(RA) domains, one forkhead-association (FHA) domain, and
one dilute (DIL) domain. AF6 PDZ mediates interactions
with a subset of ephrine-receptor protein-tyrosine kinases,[10, 11] the poliovirus receptor-related protein PRR2/
nectin,[12] the junctional adhesion molecule (JAM),[13] and
the breakpoint-cluster-region protein (BCR).[14] It binds
target peptides with comparatively low affinities in the 20–
150 mm range.[4] When the C-terminal valine of full-length
BCR is mutated to Ala, binding to full-length AF6 is
abrogated in various cell-based assays.[14]
Herein, we report the identification of novel, lowmolecular-weight ligands for the AF6 PDZ domain by
NMR spectroscopy-based screening and chemical synthesis.
These compounds are active in competition assays and
represent building blocks for the design of tight-binding
competitors. Furthermore, we have determined the solution
structure of AF6 PDZ in complex with the ligand of highest
affinity. The synthetic ligand induces the formation of a
hydrophobic subpocket in AF6 PDZ that is absent in
published structures of both apo and peptide-bound PDZ
domains. This unexpected ligand–subpocket interaction redefines the proteinBs drugability and opens the door to smallmolecule modulators for the entire family of PDZ domains.
We selected NMR spectroscopy-based screening as the hit
discovery method of choice for the identification of weakly
binding small organic frameworks because of its high
sensitivity towards weakly binding ligands and unique structural support for a fragment-based lead-generation approach.
In the first round of screening, a library of 5000 synthetic
compounds was screened against uniformly 15N-labeled
protein by using 2D 1H-15N-heteronuclear single quantum
correlation (HSQC) experiments. Three chemically distinct
classes of binders were identified as indicated by the examples
in Figure 1 a. Comparison of the three compounds with the
structure of the conserved C-terminal valine of natural
peptide ligands (Figure 1 b) revealed a consistent pharmacophore pattern of one moiety showing at least one hydrogen
bond acceptor (red), a moiety showing one hydrogen bond
donor (yellow), and a hydrophobic core of variable size.
Compound 3 occurs in two tautomeric forms in aqueous
solution (Figure 1 a) and in organic solvents (Figure 1 c), as
observed by NMR spectroscopy.
In the second round of screening, initial structure–activity
relationships (SAR) were obtained by assaying analogues
within the remaining 15 000 compounds of our fragment
library and further commercially available compounds. Each
of the three hit classes showed consistent SAR (data not
shown). Hit compound 3 was selected for further optimization
because of its comparatively high binding affinity and
chemical tractability.
Derivatives of compound 3 were synthesized by condensation of 2-thioxo-4-thiazolidinone with a series of aldehydes
according to the procedures described earlier[15, 16] (FigAngew. Chem. 2006, 118, 3874 –3879
Figure 1. a) Novel ligands for the AF6 PDZ domain as identified by
NMR spectroscopy-based screening (H-bond acceptors, red; H-bond
donors, yellow). b) Structure of the conserved C-terminal valine of
natural peptide ligands. c) Synthesis of 2-thioxo-4-thiazolidinone derivatives (for structures of R1, see Table 1). Different tautomers owing to
different solvents are shown in a) and c).
ure 1 c). Some of the condensation products 4 (all Z isomers)
were reduced to yield the racemates 5, generating a more
flexible linkage between the two ring systems. Normal-phase
HPLC and a chiral column (Chiralpak IA, 5 mm, 250 E
4.6 mm2, Chiral Technologies Europe) were used to separate
the enantiomers of 5 f. Although this attempt was initially
successful, interconversion produced the racemate again.
Aldehydes were chosen to sample hydrophobic moieties of
variable sizes in the R1 position of 4 or 5, ranging from a small
isopropyl group to various substituted and unsubstituted
aromatic ring systems (Table 1).
To monitor the binding activity of the 4 and 5 compound
series initially, the chemical shift perturbation (CSP) of Leu25
HN was used. This residue is in the GLGF loop, in contact
with the five-membered ring of the ligands (see below), and
should hence be in a similar chemical environment in all
investigated complexes. In the 4 compound series, the
derivatives with a 3-thienyl (4 a), iPr (4 b), C6H5 (4 c), or 4MeC6H4 (4 d) substituent at the R1 position showed weak
chemical-shift changes for Leu 25 HN. Compounds 4 a and 4 c
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.de
3875
Zuschriften
Table 1: Structure-activity relationship of 4 and 5 compound series.
Compd Compd[a] R1
4a
4b
4c
4d
4e
4f
4g
4h
4i
4j
5a
–
5c
–
–
5f
5g
5h
–
–
CSP[b] of Leu 25
HN
4
5[a]
3-thienyl
0.20
iPr
0.27
C6H5
0.17
0.14
4-MeC6H4
3-MeC6H4 < 0.02
4-CF3C6H4
0.50
0.21
3-CF3C6H4
4-BrC6H4
0.24
3-indolyl
< 0.02
4-PhC6H4 < 0.02
4
0.17 460 32
–
–
0.21 680 58
–
–
–
–
0.88 348 30
< 0.02
–
–
–
–
–
–
–
Dissociation
constant [mm]
5[a]
–
–
576 40
–
–
100 12
–
–
–
–
[a] Analyzed as racemic mixtures: () compound not synthesized or value not
determined. [b] Chemical shift perturbation (CSP) in ppm units. CSPs
< 0.02 ppm are considered inactive.
showed dissociation constants of 460 32 and 580 62 mm,
respectively. Chemical-shift changes below 0.02 ppm were
observed for derivatives with 3-MeC6H4 (4 e), the bulky
indolyl group (4 i), and 4-PhC6H4 (4 j) as substituents and were
categorized as nonbinding. Interestingly, the 4-CF3C6H4 (4 f)-,
3-CF3C6H4 (4 g)-, and 4-BrC6H4 (4 h)-substituted derivatives
showed binding.
Significantly, improved binding to AF6 PDZ was
observed when compound 4 f was reduced to 5 f. The parasubstitution of the phenyl ring in 5 in the R1 position by a
trifluoromethyl (CF3) is important for binding. In the 5
compound series, however, the 3-CF3C6H4-substituted derivative (5 g) showed negligible CSPs. Taken together, a fivemembered ring system with hydrogen bond accepting
capacity is critical for activity, and 5 with 4-CF3C6H4
substitution in the R1 position (5 f) showed the tightest
binding of the compounds investigated. By using the racemic
mixture of 5 f, we determined a dissociation constant of
100 mm. Assuming that only one of the enantiomers binds
preferentially, its dissociation constant may be approximated
to 50 mm, which is in the same range as that of the natural
peptide ligands.
To demonstrate that 5 f binds competitively to the
AF6 PDZ domain, a series of 15N-filtered 1D spectra with
increasing amounts of 5 f were recorded on a sample
containing equimolar concentrations of U-15N-labeled
AF6 PDZ and the natural peptide ligand[10, 4] (NH2-IQSVEV-COOH). Figure 2 shows the result of the titration experiment, in which only signals from the 14N and carbon-bound
protons appear. The peptide signal at d = 8.1 ppm, which
vanishes in the presence of the protein, re-emerges with
increasing 5 f concentration. In addition, the signals at d =
8.3 ppm and d = 7.6 ppm revert back to the “unbound” state
(black arrows). The other signals of the peptide could not be
observed as well because of the overlap with the aromatic
protons of the protein or 5 f (gray arrows). This experiment
proves that the peptide ligand is ejected from the proteinBs
binding groove by 5 f.
To understand the binding mode of compounds 5 at
atomic resolution, we determined the solution structure of the
complex between AF6 PDZ and 5 f (see the Supporting
3876
www.angewandte.de
Figure 2. Competition experiment with the AF6 PDZ domain,
NH2-IQSVEV-COOH peptide ligand, and 5 f. 1) 100 mm peptide and
100 mm AF6 PDZ domain; 2–8) Spectra recorded at 100, 200, 400, 600,
800, 1000, and 1400 mm concentrations of 5 f, respectively; 9) 1H NMR
spectrum of the peptide (100 mm).
Information). The structure is well defined by 11.8 interresidual NOEs per amino acid. The ensemble of the 20
lowest-energy structures has a mean a-carbon root-meansquare deviation (rmsd) of 0.31 0.09 J for the well-structured regions (residues 11–28, 40–70, 77–95) and 1.21 0.23 J for all heavy atoms (Table 2). The overall structure
of the AF6 PDZ domain in complex with 5 f is similar to other
PDZ structures.[6, 8, 17] The protein has six b strands (bA to bF)
and two a helices (aA, aB) (see the Supporting Information).
The binding mode of 5 f complexed to AF6 PDZ is
defined by 11 intermolecular NOEs obtained from 2D 13C-F2filtered NOESY and 2D 13C-F2-filtered HMQC-NOESY
spectra. Since the signal of 5-H of 5 f overlaps with the
water signal, only NOEs to 8-H/12-H and 9-H/11-H, which
are degenerate in the bound state, were extracted from the
spectra. All 11 intermolecular NOEs occur between these
degenerate protons, 8-H/12-H and 9-H/11-H, and the sidechain protons of Met 23, Leu 25, Ile 27, Ala 80, Met 83, and
Thr 84. As the thiazolidinone ring did not show NOE
restraints, its location was determined by computational
docking of the whole molecule utilizing all observed NMR
spectroscopic data. By enforcing the 11 intermolecular NOE
restraints, we docked the R and S enantiomers of 5 f into the
lowest-energy AF6 PDZ structure in a constrained MD
simulation lasting for 2 ns using Amber 8.[20] The tautomeric
form of 3 (Figure 1) was chosen for the simulations as the
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. 2006, 118, 3874 –3879
Angewandte
Chemie
Table 2: Structural statistics of AF6 PDZ-5 f complex.
Restraints
total no. of experimental restraints
total no. of NOE restraints
intraresidue (i = j)
sequential (j ij j = 1)
medium-range (2 j ij j 5)
long-range (j ij j > 5)
no. of H-bond restraints
no. of dihedral angle restraints (TALOS)
average inter-residual NOEs per residue
no. of NOE violations > 0.3 N
no. of dihedral angle violations > 58
fy Space (residues)[a,b]
most favored regions [%]
additionally allowed regions [%]
generously allowed regions [%]
disallowed regions [%]
rmsd values[c,d]
Ca
heavy atoms
1518
1440
428
390
151
471
14
64
12
0
0
70.3
22.5
6.3
1.0
0.31 0.09
1.21 0.23
[a] Residues considered: 11–95. [b] From Procheck-NMR.[18] [c] Residues
considered: 11–17, 25–29, 40–46, 60–66, 69–72, and 77–95. [d] Calculated by using MOLMOL.[19]
experiments were done in aqueous solution. During docking,
the backbone of the protein was kept rigid and the side chains
of the residues were restrained by using the same distance
constraints as in the Cyana structure calculation. The total
energy obtained for the complex was significantly higher for
the S enantiomer than for the R enantiomer, suggesting that
the R enantiomer of 5 f is the active binder.
The final complex structure shows the 4-CF3C6H4 group of
5 f embedded in a deep hydrophobic pocket surrounded by
the side chains of residues Met 23, Leu 25, Ile 27, Ala 80,
Met 83, Val 90, and Leu 92 (Figure 3 b and c). N3 and O4 of 5 f
form H-bonds with the backbone HNs of residues Gly 24 and
Leu 25, respectively, which belong to the conserved GLGF
loop (GMGL in the case of AF6 PDZ). The SH group is
involved in hydrophobic interactions. It also explains the
lower affinity of compounds 5 a and 5 c as compared to 5 f
because of incomplete occupancy of the proteinBs hydrophobic pocket. The weak binding of 5 g can be attributed to
steric hindrance because the meta-substituted ring would
require a larger hydrophobic cavity for binding.
To better understand the binding mode of 5 f, we
compared the structures of the AF6 PDZ domain in the
ligand-free (see the Supporting Information) and ligandbound forms. Although the superposition of the structures
with the lowest Cyana target function show a good overall
agreement with an a-carbon rmsd of 1.25 J for the b strands,
three areas of significant deviation located in aA, aB, and bE
were observed. Upon binding of 5 f, the Ca atom of Gln 76 at
the beginning of aA was displaced by 3.3 J, which results in
the widening of the peptide-binding groove (Figure 4 a). In
addition, closer inspection of the superimposed structures
reveals significant side-chain rearrangements of residues
Met 23, Leu 25, and Met 83 and a backbone rearrangement
for the residue Ala 80 in the ligand-binding groove to
accommodate the bulky 4-CF3C6H4 group of 5 f. The side
Angew. Chem. 2006, 118, 3874 –3879
Figure 3. a) Surface representation of the AF6 PDZ domain without
ligand, 1XZ9. Surface coloring indicates hydrophobic (yellow) and
hydrophilic (green) areas. b) AF6 PDZ domain in complex with 5 f.
Surface coloring indicates hydrophobic (yellow) and hydrophilic
(green) areas. The hydrogen bonding interaction between the PDZ
domain and 5 f are shown as yellow-dotted lines. c) Schematic
representation of the AF6 PDZ-5 f interaction. Hydrogen bonds are
shown as green-dotted lines. Hydrophobic interactions are highlighted
by red line fences. The Figure was generated by using the program
Ligplot.[21]
chains of the other hydrophobic residues lining this pocket
(Ile 27, Val 90, and Leu 92) are virtually unchanged. As a
consequence, 5 f triggers an induced-fit rearrangement to
shape a new hydrophobic pocket without unbound AF6 PDZ.
To gain insight into the possible interaction between 5 f
and other classes of PDZ domains and to aid the structure-
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.de
3877
Zuschriften
Figure 4. a) Comparison of AF6 PDZ in ligand-free (blue) and ligand-bound (red) state
(stereoview). b) Comparison of AF6 PDZ ligand-bound (red) state and syntrophin PDZ domain
bound to nNOS b-finger, 1QAV, blue, n-NOS b-finger not shown (stereoview). c) Comparison
of AF6 PDZ ligand-bound (red) state and erbin PDZ domain bound to Erbb2 receptor peptide,
1 MFG, blue, peptide not shown (stereoview).
based extension of 5 f for developing specific ligands, we
compared the structure of AF6 PDZ in the ligand-bound
form to the crystal structures of the syntrophin (1QAV)[5] and
erbin (1MFG)[22] PDZ domains in complex with their
respective ligands. AF6 PDZ in the ligand-bound form superimposes rather well with the two other PDZ domains with
rmsd values of 1.7 and 1.5 J, respectively. In the structure of
the syntrophin PDZ domain, the residues homologous to
Met 23, Leu 25, Ile 27, Ala 80, Met 83, Thr 84, and Leu 92 adopt
similar positions (Figure 4 b). Only Leu 149 undergoes a slight
side-chain rearrangement to accommodate the 4-CF3C6H4
group of 5 f, suggesting that it may interact similarly with
5 f. Indeed, weaker binding of 5 f to the syntrophin PDZ
domain (Kd = 270 21 mm) was observed in an NMR spectroscopic screening experiment. In the syntrophin PDZ
domain, His 141 narrows the peptide-binding groove at the
homologous position of Gln 76 in AF6 PDZ, suggesting that
3878
www.angewandte.de
variation of 5 f specifically designed to
interact with either His 141 or Gln 76 may
be helpful in developing specific ligands
for either of the two domains.
In the structure of the erbin PDZ
domain, the bulky side chain of Phe 1293,
which is homologous to Leu 25 of our
AF6 PDZ construct, severely clashes with
the CF3 moiety of 5 f (Figure 4 c). As the
conformational freedom of the Phe 1293
side chain is limited, specificity may be
controlled by the CF3 group or related
moieties. Other residues including the
homologues of Ile 27 and Ala 80, which
play an important role in the recognition
of hydrophobic residues in the 2 position with respect to the C-terminal residue of cognate peptides, show similar
orientations, thus being less-promising
interaction points for the design of specific ligands.
By using NMR spectroscopic screening, we identified three chemically distinct classes of compounds binding to the
challenging PPI target AF6 PDZ. We
improved a 2-thioxo-4-thiazolidinone
derivative to become a 100-mm Kd ligand
with a molecular weight of 291 Da, and
determined the solution structure of the
R enantiomer complexed with AF6 PDZ.
The 3D structure reveals a new hydrophobic subpocket formed through
induced-fit binding of 5 f. This finding
redefines the drugability of PDZ domains
and discloses 5-aryl-2-thioxo-4-thiazolidinones and related frameworks as promising candidates for the development of
potent and selective small-molecule modulators of individual domains from the
large PDZ family.
Experimental Section
All NMR spectroscopic experiments were performed at 295 K on
Bruker DRX600 and DMX750 spectrometers in standard configuration by using triple-resonance probes equipped with self-shielded
gradient coils. Two samples of 1.5 mm U-15N, 13C-labeled AF6 PDZ
with 5 f (1:1 protein/ligand) at pH 6.5 in Na phosphate buffer solution
(20 mm) with NaCl (50 mm) and dimethyl sulfoxide (DMSO; 10 %
v/v) either in D2O (99.98 % v/v) or H2O/D2O (90:10 % v/v).
Backbone chemical-shift assignments for the complex were
achieved as described earlier.[4] Side-chain assignments were obtained
from 3D HCCH-COSY, HCCH-TOCSY (100 ms), and 13C-edited
NOESY (80 ms) experiments recorded on a sample in D2O (99.98 %
v/v); and from 15N-edited NOESY (80 ms) spectra recorded on a
sample in H2O/D2O (90:10 % v/v).[23]
Data were processed by using XWIN-NMR (version 1.3, Bruker
BioSpin GmbH, Rheinstetten, Germany) and analyzed by using
SPARKY.[24] Structures were calculated from the restraints listed in
Table 2 using CYANA.[25] The 20 lowest-energy structures without
distance violations greater than 0.3 J and no angle violations greater
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. 2006, 118, 3874 –3879
Angewandte
Chemie
than 58 were accepted into the final ensemble. A total of 64 dihedralangle restraints were obtained from backbone chemical shifts by using
TALOS.[26] Furthermore, 14 hydrogen-bond restraints were identified
from the NOE pattern in the 3D 13C-edited NOESY spectrum.
15
N-labeled AF6 PDZ at 50 mM concentration in a 20 mm Na
phosphate buffer solution containing NaCl (50 mm), and [D]6DMSO
(10 % v/v) at pH 7.0 was used for screening experiments. Ligand
binding was detected at 300 K by acquiring 1H-15N HSQC spectra in
the presence and absence of compounds. Compounds were initially
tested at 400 mm each in mixtures of 16 compounds, with subsequent
deconvolution to mixtures of 4 compounds at 400 mm each and then to
individual compounds. Spectra were acquired with 16 scans and 128
points in the indirect dimension on a Bruker DRX600 spectrometer
equipped with a cryoprobe. Chemical-shift mapping of the binding
site was achieved by comparing the shifts of protein alone to those of
the protein in presence of ligands. Chemical shifts were quantified by
using the formula Dd = [(DH)2 + (DN/5)2] = , in which Dd is the
weighted chemical-shift change, and DH and DN are the chemical
shift changes in the proton and the nitrogen dimensions, respectively.
Dissociation constants were obtained for selected compounds by
monitoring the chemical shift changes as a function of ligand
concentration. Data were fit by using a one-site binding model. A
nonlinear least-square optimization was performed by varying the
values of Kd and the chemical shift of the fully saturated protein.
1
2
Received: November 8, 2005
Revised: January 26, 2006
Published online: May 3, 2006
.
Keywords: inhibitors · ligand effects · NMR spectroscopy ·
PDZ domains · protein structures
[15] G. Bruno, L. Constantino, C. Curinga, R. Maccari, F. Manforte,
F. NicolQ, R. OttanR, M. G. Vigorita, Bioorg. Med. Chem. 2002,
10, 1077 – 1084.
[16] S. Kukolya, S. Draheim, B. Graves, D. Hunden, J. Pfeil, R.
Cooper, J. Ott, F. Counter, J. Med. Chem. 1985, 28, 1896 – 1903.
[17] N. L. Stricker, K. S. Christopherson, B. A. Yi, P. J. Schatz, R. W.
Raab, G. Dawes, D. E. Bassett Jr, D. S. Bredt, M. Li, Nat.
Biotechnol. 1997, 15, 336 – 342.
[18] R. A. Laskowski, J. A. Rullmann, M. W. MacArthur, R. Kaptein, J. M. Thornton, J. Biomol. NMR 1996, 8, 477 – 486.
[19] R. Koradi, M. Billeter, K. WSthrich, J. Mol. Graphics 1996, 14,
29 – 32.
[20] D. A. Case, T. A. Darden, T. E. Cheatham III, C. L. Simmerling,
J. Wang, R. E. Duke, R. Luo, K. M. Merz, B. Wang, D. A.
Pearlman, M. Crowley, S. Brozell, V. Tsui, H. Gohlke, J.
Mongan, V. Hornak, G. Cui, P. Beroza, C. Schafmeister, J. W.
Caldwell, W. S. Ross, P. A. Kollman, AMBER 8, University of
California, San Francisco, 2004.
[21] M. S. Wallace, R. A. Laskowski, J. M. Thornton, Protein Eng.
1995, 8, 127 – 134.
[22] G. Birrane, J. Chung, J. A. Ladias, J. Biol. Chem. 2003, 278,
1399 – 1402.
[23] V. Kanelis, J. D. Forman-Kay, L. E. Kay, IUBMB Life 2001, 52,
291 – 302.
[24] T. D. Goddard, D. G. Kneller, SPARKY 3, University of
California, San Francisco 2003.
[25] P. Guntert, C. Mumenthaler, K. WSthrich, J. Mol. Biol. 1997,
273, 283 – 298.
[26] G. Cornilescu, F. Delaglio, A. Bax, J. Biomol. NMR 1999, 13,
289 – 302.
[1] T. Pawson, P. Nash, Science 2003, 300, 445 – 452.
[2] Z. Songyang, A. S. Fanning, C. Fu, J. Xu, S. M. Marfatia, A. H.
Chishti, A. Crompton, A. C. Chan, J. M. Anderson, L. C.
Cantley, Science 1997, 275, 73 – 77.
[3] B. Z. Harris, W. A. Lim, J. Cell Sci. 2001, 114, 3219 – 3231.
[4] U. Wiedemann, P. Boisguerin, R. Leben, D. Leitner, G. Krause,
K. Moelling, R. Volkmer-Engert, H. Oschkinat, J. Mol. Biol.
2004, 343, 703 – 718.
[5] B. J. Hillier, K. S. Christopherson, K. E. Prehoda, D. S. Bredt,
W. A. Lim, Science 1999, 284, 812 – 815.
[6] D. A. Doyle, A. Lee, J. Lewis, E. Kim, M. Sheng, R. MacKinnon,
Cell 1996, 85, 1067 – 1076.
[7] M. Cabral, J. H. Petosa, M. J. Sutcliffe, S. Raza, O. Byron, F. Poy,
S. M. Marfatia, A. H. Chishti, R. C. Liddington, Nature 1996,
382, 649 – 652.
[8] J. Schultz, U. Hoffmuller, G. Krause, J. Ashurst, M. J. Macias, P.
Schmieder, J. Schneider-Mergener, H. Oschkinat, Nat. Struct.
Biol. 1998, 5, 19 – 24.
[9] R. Prasad, Y. Gu, H. Alder, T. Nakamura, O. Canaani, H. Saito,
K. Huebner, R. P. Gale, P. C. Nowell, K. Kuriyama, Cancer Res.
1993, 53, 5624 – 5628.
[10] B. Hock, B. BPhme, T. Karn, T. Yamamoto, K. Kaibuchi, U.
Holtrich, S. Holland, T. Pawson, H. Rubsamen-Waigmann, K.
Strebhardt, Proc. Natl. Acad. Sci. USA 1998, 95, 9779 – 9784.
[11] M. Buchert, S. Schneider, V. Meskenaite, M. T. Adams, E.
Canaani, T. Baechi, K. Moelling, C. M. Hovens, J. Cell Biol.
1999, 144, 361 – 371.
[12] K. Takahashi, H. Nakanishi, M. Miyahara, K. Mandai, K. Satoh,
A. Satoh, H. Nishioka, J. Aoki, A. Nomoto, A. Mizoguchi, Y.
Takai, J. Cell Biol. 1999, 145, 539 – 549.
[13] K. Ebnet, C. U. Schulz, M. K. Meyer Zu Brickwedde, G. G.
Pendl, D. Vestweber, J. Biol. Chem. 2000, 275, 27 979 – 27 988.
[14] G. Radziwill, R. A. Erdmann, U. Margelisch, K. Moelling, Mol.
Cell. Biol. 2003, 23, 4663 – 4672.
Angew. Chem. 2006, 118, 3874 –3879
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.de
3879
Документ
Категория
Без категории
Просмотров
1
Размер файла
517 Кб
Теги
discovery, af6, domain, pdz, molecular, low, weight, ligand
1/--страниц
Пожаловаться на содержимое документа