Binding, Domain Orientation, and Dynamics of the Lck SH3-SH2 Domain Pair
and Comparison with Other Src-Family Kinases
†
Gregor Hofmann,
‡
Kristian Schweimer,
§
Anke Kiessling,
§
Edith Hofinger,
§
Finn Bauer,
§,|
Silke Hoffmann,
§
Paul Ro¨sch,
§
Iain D. Campbell,
‡
Jo¨rn M. Werner,
‡,⊥
and Heinrich Sticht*
,|
Department of Biochemistry, UniVersity of Oxford, South Parks Road, Oxford OX1 3QU, U.K., Lehrstuhl fu¨r Biopolymere,
UniVersita¨t Bayreuth, D-95440 Bayreuth, Germany, and Institut fu¨r Biochemie, Emil-Fischer-Zentrum,
UniVersita¨t Erlangen-Nu¨rnberg, D-91054 Erlangen, Germany
ReceiVed May 3, 2005; ReVised Manuscript ReceiVed August 1, 2005
ABSTRACT: The catalytic activity of Src-family kinases is regulated by association with its SH3 and SH2
domains. Activation requires displacement of intermolecular contacts by SH3/SH2 binding ligands resulting
in dissociation of the SH3 and SH2 domains from the kinase domain. To understand the contribution of
the SH3-SH2 domain pair to this regulatory process, the binding of peptides derived from physiologically
relevant SH2 and SH3 interaction partners was studied for Lck and its relative Fyn by NMR spectroscopy.
In contrast to Fyn, activating ligands do not induce communication between SH2 and SH3 domains in
Lck. This can be attributed to the particular properties of the Lck SH3-SH2 linker which is shown to be
extremely flexible thus effectively decoupling the behavior of the SH3 and SH2 domains. Measurements
on the SH32 tandem from Lck further revealed a relative domain orientation that is distinctly different
from that found in the Lck SH32 crystal structure and in other Src kinases. These data suggest that flexibility
between SH2 and SH3 domains contributes to the adaptation of Src-family kinases to specific environments
and distinct functions.
Src family tyrosine kinases are implicated in a large
number of cellular processes including cell adhesion and
spreading, focal adhesion formation and disassembly, cell
migration, cell cycle progression, apoptosis, cell differentia-
tion, and gene transcription (1, 2). Mutations in these proteins
can also lead to cancer (3). All members of the family (Src,
Blk, Fgr, Fyn, Hck, Lck, Lyn, Yes, Yrk) have a common
molecular architecture that includes regulatory SH3
1
and SH2
domains and a kinase domain. In its inactive state, the kinase
is inhibited by binding of the SH2 domain to a conserved
tyrosine in the C-terminal regulatory tail and by binding of
the SH3 domain to the SH2-kinase linker region (4-6).
Dephosphorylation of this conserved tyrosine (7) or binding
of competing ligands to the SH2 or SH3 domains (8)
activates the kinase. Young et al. (9) suggested that coupling
between SH2 and SH3 is important in controlling activation,
because mutations of residues in the SH3-SH2 linker
activated c-Src. Molecular dynamics simulations also sug-
gested that, in the inactive state, the coupling between the
SH3-SH2 dynamics was tight, and that this no longer
applies when the regulatory tail is detached from the SH2
domain (9). Hence, they proposed that the SH3-SH2-kinase
assembly represented an “inducible snap lock”.
While the crystal structures of c-Src and Hck have
provided a wealth of structural information on the closed,
inactive state of Src-kinases, the open state is much less well
characterized, in terms of both structure and dynamics.
Crystal structures have been obtained for the SH3-SH2
domain pairs of Lck (10) and Fyn (11). These fragments
might be expected to behave in a similar way to the domains
when they are detached from the kinase in the activated state.
There has, therefore, been some previous interest in studying
the solution behavior of SH3-SH2 domain pairs including
those of Src (12), Abl (13), and Fyn (14). Comparison of
the domain orientation observed in the Fyn SH32 crystal
structure with the domain orientation of peptide bound Fyn
SH32 observed in solution showed a relatively modest
average domain reorientation along one rotation axis. This
is consistent with a situation that is intermediate between a
rigid and a completely decoupled interdomain interface (14).
Intriguingly, the orientation between the SH3 and SH2
domains in the crystal structure of Lck SH32 is very different
from that observed for other Src-family members. Here
solution NMR spectroscopy is used to determine the domain
orientation of Lck SH32 in solution and to characterize the
dynamics of this domain pair. Binding of peptides derived
from physiologically relevant ligands to both Fyn SH32 and
Lck SH32 is also investigated to provide further insight into
†
This work was supported by grants of the Deutsche Forschungs-
gemeinschaft (SFB466: A5, C11) to P.R. and H.S. and by the
Wellcome Trust to I.D.C. and J.M.W.
* To whom correspondence should be addressed. Tel:
+49 9131 8524614. Fax: +49 9131 8522485. E-mail:
H.Sticht@biochem.uni-erlangen.de.
‡
University of Oxford.
§
Universita¨t Bayreuth.
|
Universita¨t Erlangen-Nu¨rnberg.
⊥
Present address: School of Biological Sciences, University of
Southampton, Southampton SO16 7PX, U.K.
1
Abbreviations: HMTA, hamster polyomavirus middle T-antigen;
HSQC, heteronuclear single-quantum coherence; NMR, nuclear mag-
netic resonance; PAG, polyacrylamide gel; RDC, residual dipolar
coupling; RMSD, root-mean-square deviation; SH2, Src-homology
domain 2; SH3, Src-homology domain 3; SH32, SH3-SH2 domain
pair; Tip, tyrosine kinase interacting protein
13043Biochemistry 2005, 44, 13043-13050
10.1021/bi050814y CCC: $30.25 © 2005 American Chemical Society
Published on Web 09/10/2005
the contribution made by SH3-SH2 domain pairs to the
regulation of Src kinases.
MATERIALS AND METHODS
Protein Production and Peptide Ligands. Expression and
purification of the SH3-SH2 domain pairs of human Lck
(Lck SH32; residues 57-225) and Fyn (Fyn SH32; residues
81-247 with C239S, C240S, and C246S) were performed
as described previously (11, 14, 15). Minimal medium with
15
NH
4
Cl and if necessary D
2
O instead of H
2
O was used to
produce
15
N-labeled Lck SH32 and
15
N,
2
H-labeled Fyn SH32.
Purity of the sample was confirmed by SDS-polyacrylamide
gel electrophoresis, and concentration was assessed by the
optical density at 280 nm, using extinction coefficients
calculated from the amino acid sequence.
Peptide ligands for SH3 and SH2 binding experiments
were purchased from Biosyntan (Berlin, Germany) and
Coring (Gernsheim, Germany) and contained blocked end
groups. The Tip peptide comprises residues 167-199 of the
tyrosine-kinase interacting (Tip) protein from Herpesvirus
saimiri strain C488 and the phosphorylated HMTA peptide
the CQ(pY)EEIP sequence from the middle T antigen of
hamster polyomavirus. The P f C point mutation at the
N-terminal sequence position of the investigated peptide was
shown to have no effect on SH2 binding affinity (Hofinger
and Sticht, unpublished).
NMR Sample Preparation and Data Acquisition. Protein
solutions were prepared in the following buffer conditions:
100 mM K
2
HPO
4
/KH
2
PO
4
, pH 6.5, 20 mM NaCl for Lck
SH32 and 50 mM K
2
HPO
4
/KH
2
PO
4
, pH 7.0, 200 mM Na
2
-
SO
4
for Fyn SH32. For the peptide titrations, aliquots of
peptide of known concentration were first dried using a
Speed-Vac apparatus and successive aliquots were then
dissolved in the sample solution.
Polyacrylamide gels were prepared as described by Chou
et al. (16) and were squeezed into an open-ended NMR tube
by application of pressure, using a home-built squeezer. For
recording residual dipolar couplings in liquid crystalline
medium, a sample of 450 µM Lck SH32 in monododecyl-
pentaethyleneglycol-ether/hexanol (molar ratio 0.95; 3% (wt)
C12E5/H
2
O) was used that was prepared as described in
Ru¨ckert & Otting (17). For all systems, proper alignment
was checked by measuring the quadrupolar splitting of the
deuterium resonance of D
2
O. NMR experiments were
performed on spectrometers operating at
1
H frequencies of
500, 600, and 750 MHz. Unless otherwise stated, all spectra
were recorded at 25 °C. Backbone resonance assignments
of Lck SH32 and Fyn SH32 have been described previously
(11, 15).
Measurement and Interpretation of Relaxation Data. Two-
dimensional {
1
H}-
15
N heteronuclear NOE experiments and
a series of
1
H-
15
N correlation spectra for the determination
of
15
N T
1
and T
2
relaxation time constants were acquired
using previously described methods incorporating pulsed
field gradients for coherence pathway selection and water
suppression (18-20).
15
N-T
2
constants were measured using
a spin-echo sequence with a Carr-Purcell-Meiboom-Gill
(CPMG) delay of 419.4 µs. Dipolar and chemical shift
anisotropy (CSA) cross-correlation occurring during delay
periods were removed by applying
1
H 180° pulses. This was
performed once in the middle of the CPMG delay block for
15
N-T
2
experiments and every 5 ms in the delay period of
15
N-T
1
experiments (21, 22). The total sample heating for
different experiments was equalized by application of a
“heat” pulse train to keep the total number of
15
N 180° pulses
constant for all experiments. Each series of T
1
and T
2
measurements consisted of 8 autocorrelation spectra with
increasing
15
N relaxation time delays, chosen to sample
approximately the entirety of the observed intensity decays.
All spectra were processed with mild resolution enhancement
and linear prediction in the indirect dimension using Felix
2.3 (Biosym, San Diego, CA). After zero filling, the digital
resolution of the spectra was 6.1 Hz/point in the
1
H
dimension and 1.3 Hz/point in the
15
N dimension. Relaxation
parameters were estimated from two parameter exponential
fits to the intensity decays in the series of T
1
T
2
correlation
spectra. Errors were estimated using the baseline noise (23).
To test initially for the aggregation state of free Lck SH32,
experimental T
1
and T
2
relaxation times were compared with
calculated relaxation times as a function of isotropic cor-
relation time and order parameter (S
2
) using the Lipari and
Szabo model (24). For Lck SH32 correlation times and
diffusion tensors were derived using the T
1
and T
2
values of
residues that are part of secondary structure elements and
whose values were inside the S
2
) 1 envelope of the Lipari-
Szabo model. The principal axes, D
x
, D
y
, and D
z
,ofthe
diffusion tensor were determined by global least-squares fits
of the T
1
/T
2
ratios derived from the spectral density functions
of a diffusing particle to the experimental values (25-27).
Three models of increasing complexity were tested: a sphere
[D
x
) D
y
) D
z
, D ) (D
x
+ D
y
+ D
z
)/3], a symmetric top
[D
|
) D
z
, and D
⊥
) (D
x
+ D
y
)/2], and a fully asymmetric
tensor (D
x
* D
y
* D
z
)(28). The analysis was performed
with programs written in-house. The NH bond length of
backbone amides was assumed to be 1.02 Å, and the
chemical shift anisotropy (CSA) was assumed to be -170
ppm with the CSA tensor taken to be collinear with the
dipolar vector. Model comparisons were performed by using
the F-test and calculation of the probability Q to obtain this
F by chance (29). Generally, Q values below 5% were
assumed to justify the use of the more complex model.
Domain Orientation from Residual Dipolar Couplings.
Apparent J
NH
constants of
15
N-labeled Lck SH32 were
obtained using the in-phase, antiphase (IPAP) scheme (30).
Two sets of data were collected at temperatures of 25 °C
and 20 °C from an aligned gel sample and a sample
containing a liquid crystalline phase (as described above),
respectively. All spectra were processed with resolution
enhancement in both
15
N and
1
H dimensions and linear
prediction in the indirect dimension using Felix 2.3. After
zero filling the digital resolution of the spectra was 0.45 and
7.81 Hz for the
15
N and
1
H dimensions, respectively.
The residual dipolar
15
N-
1
H coupling (RDC) was calcu-
lated from the apparent J
NH
constants for spectra collected
for the aligned and isotropic media: RDC ) J
NH,aligned
-
J
NH,isotropic
. The error in the RDCs was estimated from the
ratios of the line widths and signal-to-noise values in each
spectrum using error propagation. Nonlinear optimization of
the fit of the experimental residual dipolar couplings to
calculated values, as a function of alignment tensor param-
eters, was performed using the program MODULE (31) and
routines written in-house. Only dipolar couplings from
residues in regular secondary structure elements were used
13044 Biochemistry, Vol. 44, No. 39, 2005 Hofmann et al.
for analysis. Best fits were determined by χ
2
minimization,
where χ
2
is defined as
σ
2
was taken to be the error in the measured RDCs, estimated
to be 2.0 Hz from the ratios of the line width and signal-
to-noise values in each spectrum using error propagation.
Ligand Binding Studies. The binding of Tip(167-199) and
HMTA to Lck SH32 and Fyn SH32 was monitored by
chemical shift perturbations in a series of
1
H-
15
N HSQC
experiments upon titration of unlabeled ligand to the
15
N
labeled SH32 domains. All titrations were performed to an
at least 4-fold excess of ligand. Eight spectra with Tip peptide
concentrations ranging from 0 to 1.72 mM were collected
for experiments on Fyn SH32, and the corresponding protein
concentration was 250 µM. Binding constants for Tip and
Fyn SH32 were estimated from chemical shift changes of
individual nuclei at different peptide concentrations by
nonlinear curve fitting (32) according to the equation
where ∆ is the chemical shift change, ∆
0
is the chemical
shift change at saturation, K
d
is the binding constant, and
[L] and [P] are the concentrations of peptide ligand and
protein, respectively. This formula assumes a simple [protein]
+ [ligand] a [complex] binding model. An estimate of the
binding constant K
d
for the entire complex was made as the
average of the measured K
d
values for individual nuclei. The
error in K
d
was estimated as the standard deviation of the
sampled K
d
values.
Since intermediate and slow exchange phenomena on the
NMR time scale were observed for several resonances in
the Tip(167-199)-Lck SH32 titration experiment, the
binding constant was determined by fluorescence measure-
ments in an identical fashion as described for Tip(168-187)
and Lck SH3 (33). The W170L point mutation present in
the Tip peptides used for fluorescence measurements was
previously shown to have no effect on the Tip-Lck affinity
(33). Since the concentration of Lck SH32 was always low
compared to the ligand concentration, the experimental data
were fitted to the equation F ) F
max
[L]/(K
d
+ [L]), where
[L] is the final ligand concentration at each measurement
point, F is the measured protein fluorescence intensity at
the particular peptide concentration, and F
max
is the observed
maximal fluorescence intensity of the protein when saturated
with the peptide. Nonlinear regression curve fitting was
carried out to fit the experimental data to the equation, with
F
max
and K
d
as fitted parameters. The change in protein
concentration that occurred as a result of peptide addition
was properly corrected.
RESULTS AND DISCUSSION
SH3 and SH2 Domain Communication by Ligand-Binding
to Lck SH32 and Fyn SH32. Two peptides that are specific
for SH3 and SH2 domain binding respectively were used to
investigate the effect on interdomain communication in Lck
SH32 and Fyn SH32.
The herpesviral tyrosine kinase interacting protein (Tip)
contains a proline-rich sequence [Tip(167-199)] that binds
to the SH3 domains of several Src-family kinases (33), and
this interaction was shown to be sufficient for a moderate
activation of Lck (34). Importantly, Fyn is not activated by
Tip (35) even though both Lck and Fyn are expressed in
T-lymphocytes.
HMTA is derived from the middle T-antigen of hamster
polyomavirus. It contains a phosphotyrosine site and is the
strongest Lck SH2 ligand known (K
d
∼ 140 nM (36-38)).
HMTA activates a range of Src-family kinases including both
Lck and Fyn (39-42).
Chemical shift changes induced by titrating a protein
solution with a ligand were used to identify local environ-
mental changes caused by complex formation. With respect
to domain-domain interactions, chemical shift changes in
the domain that is not involved in direct binding of the
respective ligand are of particular interest, because these
indicate possible propagation of structural changes from one
domain to the other. Such changes were indeed found for
the interaction of the SH32 domain pairs of Src (12) and
Fyn (14) with peptides derived from natural ligands. Com-
munication of this kind between domains was suggested to
be important in the activation of Src-kinases (9, 14) and may
be involved in kinase regulation.
Tip(167-199) binding to Lck SH32 and Fyn SH32 was
monitored by changes in
1
H and
15
N chemical shifts of the
proteins. The dissociation constant, K
d
, for the Tip(167-
199):Lck SH32 complex was determined to be 3.8 ( 0.2
µM using fluorescence spectroscopy. Both the chemical shift
changes and the K
d
are similar to the results for the
interaction of an almost identical Tip-peptide with the Lck
SH3 domain alone (33). No chemical shift changes could
be observed in the SH2 domain. Hence, Tip(167-199)
binding to the SH3 domain of Lck SH32 does not appear to
induce structural changes in the SH2 domain of this domain
pair (Figure 1a).
The dissociation constant of the Tip(167-199):Fyn SH32
complex was 64 ( 18 µM. This affinity is about 15 times
lower than for the Tip(167-199):Lck SH32 complex and
about 2-fold lower than for a focal adhesion kinase derived
SH3 ligand that activates Fyn (14). Only one noticeable
chemical shift change is found in the SH2 domain (residue
Leu163). Thus, generally no structural change seems to be
transmitted through the SH3-SH2 linker upon Tip(167-
199) binding (Figure 1a,b). Since such changes were typically
observed in experiments with ligands that do activate Fyn
(14), communication between domains may be a critical
feature of natural partners of Fyn. This feature is not shared
by Tip(167-199).
The interdomain communication in Lck SH32 was further
investigated by monitoring the chemical shift changes upon
titration of the SH2 specific ligand HMTA. This did not lead
to any chemical shift changes in the SH3 domain, suggesting
the absence of structural communication between domains
even for this very high affinity Lck SH2 ligand (Figure 1c).
In contrast, binding of HMTA to both Src SH32 (12) and
Fyn SH32 (14) did confer chemical shift changes in the
respective SH3 domains. This suggests that Lck differs from
other Src-kinases in that it has no facility for structural
communication through the SH3-SH2 linker, possibly
χ
2
)
∑
i)0
N
(RDC
experiment,i
- RDC
simulated,i
)
2
σ
2
∆ ) ∆
0
K
d
+ [L] + [P] -
x
(K
d
+ [L] + [P])
2
- 4[L][P]
2[P]
Lck SH3-SH2 Domain Orientation Biochemistry, Vol. 44, No. 39, 2005 13045
resulting in regulation pathways distinct from other members
of the family.
Domain Orientation of Lck SH32 in Solution. In aqueous
solution, through-space dipolar interactions are averaged to
zero by isotropic tumbling and are no longer observed.
However, it is possible to observe residual dipolar couplings
(RDCs) in a weakly aligned medium (43). A set of such
couplings provides orientational constraints that allow the
determination of a preferred orientation of molecules or
domains in the anisotropic medium. Comparing the alignment
tensors of individual domains in domain pairs provides a
sensitive measure of interdomain orientation (44).
To assess the SH3-SH2 domain orientation of Lck in
solution, we measured its
1
H
15
N RDCs (Figure 2) in an
aligned polyacrylamide gel medium (PAG). Alignment
tensors were then determined separately for the SH3 and SH2
domains using the published crystal structure (pdb code 1lck).
A good fit was obtained for the SH3 domain and a reasonable
one for the SH2 domain, showing that the structure of the
isolated domains is similar in solution and in the crystal. R
values calculated according to Clore et al. (45) were 0.23
for the SH3 domain and 0.35 for the SH2 domain. The worse
quality of the fit for the SH2 domain is likely to be caused
by structural differences induced by the presence of a
phoshpotyrosine peptide bound to the SH2 domain in the
crystal structure.
The alignment tensors corresponding to the SH3 and SH2
domains, however, have significantly different orientations
suggesting that the SH3-SH2 domain orientation in solution
is different from that found in the crystal state (Table 1).
We generated model solution structures by performing rigid
body rotations of the SH2 and SH3 domains that achieve
alignment of the individual alignment tensors. Residue E123
in the Lck SH32 linker was chosen as a hinge for these
FIGURE 1: (a) Interaction of Lck SH32 with the Tip(167-199) peptide that binds to the SH3 domain. Residues are colored in red, orange,
and yellow if the magnitude of the changes of the normalized chemical shift upon titration exceeds 0.12, 0.06, or 0.04 ppm, respectively.
Residues for which insignificant changes of the normalized chemical shifts (<0.04 ppm) were detected or for which no shift data could be
obtained are colored in gray and white, respectively. All normalized values of the chemical shifts given were calculated as ∆
norm
) [(∆
HN
)
2
+ (∆
N
/10)
2
)]
1/2
. (b) Ribbon diagram of the crystal structure of Fyn SH32 showing chemical shift changes on Tip(167-199) binding. Residues
are colored in red, orange, and yellow if the magnitude of the changes of the normalized chemical shift upon titration exceeds 0.20, 0.10,
or 0.04 ppm, respectively. Residues for which insignificant changes of the normalized chemical shifts (<0.04 ppm) were detected or for
which no shift data could be obtained are colored in gray and white, respectively. (c) Interaction of Lck SH32 with the HMTA peptide that
binds to the SH2 domain. Color coding as in panel a.
FIGURE 2: Equivalent sections of overlaid IPAP spectra of Lck SH32 in isotropic buffer solution (left) and in an aligned PAG medium
(right). Some HN couplings are indicated by arrows, and their magnitude is given in hertz.
13046 Biochemistry, Vol. 44, No. 39, 2005 Hofmann et al.
rotations because, in heteronuclear NOE, experiments re-
vealed an exceptionally high flexibility (see below). Align-
ment tensors are degenerate with respect to 180° rotations
around their axes, giving rise to four different degenerate
solutions. Two of these models were incompatible with the
connectivity of the molecule or resulted in prohibitive steric
clashes. To identify the correct model of the two remaining
models a second set of RDCs in a different alignment
medium was measured and subjected it to the same analysis.
Comparison of the pairwise RMSDs between all the resulting
models revealed that only two of the resulting orientations
were reasonably close (backbone RMSD 2.6 Å) while the
others were quite different (backbone RMSD > 5.0 Å). The
structure favored from the two alignment media was termed
Lck-RDC. In this model the interdomain orientations differ
considerably from the orientations present in the crystal
structures of either Lck SH32 or Fyn SH32, or indeed in the
inactive state of Src kinases (Figure 3; Table 2).
Thus, the alignment data allowed the determination of the
average relative SH3-SH2 domain orientation in solution,
but it cannot give unambiguous information whether the two
domains are mobile with respect to each other or not. The
axial and rhombic components of the alignment tensor of
the SH3 and SH2 domain are quite similar (Table 1), which
might indicate a rigid domain orientation, but this observation
alone is not sufficient to rule out interdomain dynamics. For
that reason we performed a comprehensive relaxation
analysis which is considered to be more sensitive for
elucidating dynamic behavior of molecules.
Interdomain Dynamics of Lck SH32. NMR relaxation of
nuclei in a protein in solution is mediated by rotational
diffusion. Analysis of transversal (T
1
) and longitudinal (T
2
)
relaxation times allows determination of characteristic tum-
bling times of the domains in the protein. Furthermore, the
diffusion tensors to the corresponding molecules or domains
can be determined reflecting the shape of the molecule. In
addition, the heteronuclear {
1
H}
15
N NOE is a relaxation
parameter that is particularly sensitive to local motion on a
subnanosecond time scale. The lower the value for the
heteronuclear {
1
H}
15
N NOE, the greater the local flexibility
of the protein.
We measured the {
1
H}
15
N heteronuclear NOE for Lck
SH32 to determine sites of enhanced local flexibility (Figure
4). The SH3-SH2 linker region turned out to be extremely
flexible. Residue E123 has a negative NOE value indicative
of a high degree of mobility. This flexible hinge region was
retained in Tip-bound Lck SH32 (data not shown). Hetero-
nuclear NOE data gathered for Fyn SH32 (11) also indicate
Table 1: Alignment Tensors of the SH3 and SH2 Domains of Lck
SH32
a
domain A
a
[10
-5
]
b
A
r
[10
-5
]
b
R [deg]
c
β [deg]
c
γ [deg]
c
χ
total
2
SH3 3.7 ( 0.4 1.5 ( 0.5 112 ( 14 59 ( 5 157 ( 5 6.19
SH2 3.8 ( 0.5 1.4 ( 0.4 149 ( 9 108 ( 3 141 ( 5 21.35
a
For the fit on the Lck SH32 crystal structure 19 RDCs were included
in the fit for the SH3 domain and 27 for the SH2 domain. Since the axes
of the Lck SH3 and SH2 alignment tensors are not collinear, domain
reorientation is necessary to obtain alignment and therefore no fit was
performed for the SH32 domain pair of the crystal structure.
b
A
a
and A
r
are the axial and rhombic components of the alignment tensor.
c
R, β, and
γ are Euler angles for the rotation of the alignment tensor into the molecular
frame.
FIGURE 3: Comparison of structure of Lck-RDC (green), that is
favored from combination of the data from two oriented media,
with the crystal structures of Lck SH32 (magenta), Fyn SH32 (blue),
and the Hck SH32 (yellow) domain pair as part of the inactive
kinase. All structures are overlaid on their SH2 domain.
Table 2: Summary of the Relative Domain Orientations
structure tilt angle
a
[deg] twist angle
a
[deg]
Hck SH32
b
13 306
Fyn SH32 38 334
Lck SH32 crystal structure 117 345
Lck SH32 RDC structure 94 302
a
Tilt and twist angles for the rotation relating inertia tensors corre-
sponding to fragments 106-109 and 171-176 in Hck and homologous
fragments in Fyn and Lck.
b
As part of the Hck crystal structure of the
inactive kinase.
FIGURE 4: The backbone
15
N{
1
H}-NOE of unbound Lck SH32 at
a
15
N Larmor frequency of 50 MHz. Secondary structure elements
are indicated by black bars. The negative value measured for E123
which is located in the SH3-SH2 linker region indicates increased
backbone flexibility at this sequence position.
Lck SH3-SH2 Domain Orientation Biochemistry, Vol. 44, No. 39, 2005 13047
![FIGURE 1: (a) Interaction of Lck SH32 with the Tip(167-199) peptide that binds to the SH3 domain. Residues are colored in red, orange, and yellow if the magnitude of the changes of the normalized chemical shift upon titration exceeds 0.12, 0.06, or 0.04 ppm, respectively. Residues for which insignificant changes of the normalized chemical shifts (<0.04 ppm) were detected or for which no shift data could be obtained are colored in gray and white, respectively. All normalized values of the chemical shifts given were calculated as∆norm ) [(∆HN)2 + (∆N/10)2)]1/2. (b) Ribbon diagram of the crystal structure of Fyn SH32 showing chemical shift changes on Tip(167-199) binding. Residues are colored in red, orange, and yellow if the magnitude of the changes of the normalized chemical shift upon titration exceeds 0.20, 0.10, or 0.04 ppm, respectively. Residues for which insignificant changes of the normalized chemical shifts (<0.04 ppm) were detected or for which no shift data could be obtained are colored in gray and white, respectively. (c) Interaction of Lck SH32 with the HMTA peptide that binds to the SH2 domain. Color coding as in panel a.](/figures/figure-1-a-interaction-of-lck-sh32-with-the-tip-167-199-dux63mhj.webp)
