Conformational Changes Upon Calcium Binding and Phosphorylation

Conformational
Calcium Binding and Phosphorylation
in a SyntheticChanges
FragmentUpon
of Calmodulin
in a Synthetic Fragment of Calmodulin
Luca Settimo,1 Serena Donnini,2 Andre H. Juffer,2 Robert W. Woody,3 Oriano Marin4
1

CRIBI Biotechnology Centre, University of Padova, via U.Bassi, 58/b, 35131 Padova, Italy

The Biocenter and the Department of Biochemistry, University of Oulu, P.O. Box 3000,
FIN-90014 University of Oulu, Finland
3

Department of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, CO 80523-1870

Department of Biological Chemistry, University of Padova, Viale G. Colombo 3, 35121 Padova, Italy

Received 5 June 2006; revised 8 October 2006; accepted 11 December 2006

Published online 15 December 2006 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/bip.20657

ABSTRACT:

modication is predicted to alter the secondary structure in

We have recently investigated by far-UV circular

its surrounding and also to hinder the physiological

dichroism (CD) the effects of Ca2 binding and the

bending of the central helix of CaM through an alteration

phosphorylation of Ser 81 for the synthetic peptide CaM

of the hydrogen bond network established by the side chain

[54106] encompassing the Ca2-binding loops II and III

of residue 81. Using quantum mechanical methods to

and the central a helix of calmodulin (CaM) (Arrigoni

predict the CD spectra for the frames obtained during the

et al., Biochemistry 2004, 43, 1278812798).Using

MD simulations, we are able to reproduce the relative

computational methods, we studied the changes in the

experimental intensities in the far-UV CD spectra for our

secondary structure implied by these spectra with the aim

peptides. Similar conformational changes that take place

to investigate the effect of Ca2 binding and the functional

role of the phosphorylation of Ser 81 in the action of the
full-length CaM. Ca2 binding induces the nucleation of

in CaM [54106] upon Ca2 binding and

phosphorylation may occur in the full-length CaM.
#

2006 Wiley Periodicals, Inc. Biopolymers (Pept Sci) 88:

helical structure by inducing side chain stacking of

373385, 2007.

hydrophobic residues. We further investigated the effect of

Keywords: calmodulin; calcium; phosphorylation;

Ca2 binding by using near-UV CD spectroscopy.

conformational changes; circular dichroism

Molecular dynamics simulations of different fragments

containing the central a-helix of CaM using various
experimentally determined structures of CaM with bound
Ca2 disclose the structural effects provided by the
phosphorylation of Ser 81. This post-translational
Correspondence to: Luca Settimo; e-mail: luca.settimo@poste.it
This article contains supplementary material available via the Internet at http://
www.interscience.wiley.com/jpages/0006-3525/suppmat.

C 2006
V

Wiley Periodicals, Inc.

PeptideScience Volume 88 / Number 3

This article was originally published online as an accepted

preprint. The Published Online date corresponds to the preprint
version. You can request a copy of the preprint by emailing the
Biopolymers editorial office at biopolymers@wiley.com

INTRODUCTION

almodulin (CaM) is a ubiquitous eukaryotic Ca2binding protein that binds and activates different targets1 and has a very important physiological role. Several structures of CaM have been reported using Xray crystallography and nuclear magnetic resonance
(NMR) (see e.g., reviews in Refs. 2 and 3).

373

374

Settimo et al.

These studies revealed that CaM has a dumbbell-like

shape with N- and C-terminal lobes separated by a central
helix. Calmodulin-like domains have been identied also in
other proteins.4 Two EF-hand Ca2-binding motifs are present in each lobe of CaM. When CaM binds Ca2, a conformational change occurs in each lobe of CaM with exposure
of hydrophobic residues that bind the target.3 All the structures reported for the full-length CaM can be classied in
one of the following three functional states: (i) Ca2-free
CaM, such as the structure determined using NMR5; (ii)
Ca2-bound CaM, as the structure published by X-ray crystallography by Chattopadhyaya et al.,6 which reported a
straight and continuous central a-helix of CaM, or by Fallon
and Quiocho,7 which, in contrast, reported the central ahelix of CaM bent in the middle; (iii) Ca2-bound CaM in
complex with a ligand or target, as the structure determined
by X-ray crystallography by Meador et al.8 or by NMR by
Elshorst et al.9 The central helix of CaM has a fundamental
role since it assures the correct positioning of the N- and Cterminal lobes in the binding and activation of the target,
and it is very exible in solution in both apo5,10 and Ca2bound form.9,11 The conformational changes provided by
Ca2 binding and the conformation and exibility of the
central helix of CaM have been widely studied and investigated using different methods such as: (i) NMR5,916; (ii)
small-angle X-ray scattering17,18; (iii) spin-label electron paramagnetic resonance19; (iv) optical spectroscopic techniques
such as far-UV CD,20 Fourier transform infrared spectroscopy,21,22 uorescence,23,24 optical rotatory dispersion measurements25; (v) structural studies on protein mutants at residues in the central helix2631; (vi) cross-linking.32
Several modeling studies and molecular dynamics (MD)
simulations have been performed on CaM and they have
been reviewed recently.3 In particular, MD simulations have
been performed on the full-length CaM (i) in the Ca2loaded form3339; (ii) in the Ca2-free form40; (iii) in the
Ca2-loaded form in complex with a target peptide.41 Also
the central helix of CaM (fragment 6592) has been analyzed
using MD.35 All these MD studies underscore the exibility
of the middle part of the central helix of CaM (corresponding approximately to the region encompassed by residues 7482), important for the correct functioning of CaM.
CaM is phosphorylated by different kinases as reviewed in
Ref. 42. In particular, the casein kinase 2 (CK2) is a serine/
threonine kinase that phosphorylates Thr 79, Ser 81, Thr 110,
and Thr 117 in the full-length CaM.4347 Thr 79 and Ser 81,
located in the middle of the central a helix of CaM, are the
most phosphorylated residues in the full-length protein.44,46,47
The interaction between CK2 and the central a helix of CaM
has been recently studied by molecular modeling.48

We have recently reported the synthesis, purication, and

characterization of the fragment encompassing residues 54
106 of human CaM, wild-type and phosphorylated in position 81.47 These peptides encompass the Ca2-binding loops
II and III and the central a-helix of the full-length CaM. We
will refer to these two peptides as CaM [54106] and CaM
[54106]pS81, respectively. In the same study, we investigated by far-UV CD the effect of Ca2 binding and phosphorylation of residue 81 on CaM [54106] (see Figure 6 in
Ref. 47). We also found that CaM [54106]pS81 could not be
phosphorylated at Thr 7947 despite the fact a phosphorylated
side-chain at n 2 is a strong positive determinant for CK2
phosphorylation of short peptides.49
It is generally known that phosphorylation of CaM
decreases the binding and activation of CaM-dependent targets.42 However, there are no studies in the literature that
investigate the structural consequences of the phosphorylation of CaM. As reported recently by Doig and coworkers,
the structural effects consequent to protein phosphorylation
are very variable (e.g. phosphorylation can stabilize or destabilize a helices) and highly dependent upon the position and
possible interactions that the phosphorylated residue can establish with the neighboring residues.50,51
In the present study, we (i) analyze spectroscopic data for
the fractions of secondary structure elements using several
methods; (ii) experimentally analyze the effect of Ca2 binding using near-UV CD; (iii) investigate conformational
changes induced by Ca2 binding and phosphorylation of
Ser 81 in our peptides[54106] using molecular modeling
and MD simulation; (iv) rationalize the experimentally
observed change in far-UV CD spectra by calculating theoretical CD spectra using structures from the MD trajectories.
Our models also explain why CK2 is not able to phosphorylate Thr 79 in CaM[54106]pS81.
This study aims to investigate the structuralfunctional
role correlated to Ca2 binding and phosphorylation in the
full-length CaM by using a fragment of CaM.

MATERIALS AND METHODS

Spectroscopic Measurements
Circular dichroism (CD) spectra were recorded at 258C at pH 7.5
on a Jasco (Tokyo, Japan) Model J-710 spectrometer equipped with
a thermostated cell-holder and a Neslab RTE-110 water circulating
bath. The instrument was calibrated with ()-10-camphorsulfonic
acid.52 The far-UV CD spectra of the peptides [54106]CaM and
[54106]pS81CaM in the absence or presence of Ca2, which are
analyzed with computational methods, have been already
reported.47

Biopolymers (Peptide Science) DOI 10.1002/bip

Conformational Changes in CaM

375

Table I MD Simulations Carried Out in This Study

Abbreviation

CaM[fragment]

1cll[54106]
1cll[54106]pS81
1cll[6493]
1cll[6493]pS81
1prw[6493]
1prw[6493]pS81
1cff_4[6493]
1cff_4[6493]pS81

Initial Coordinates (pdb code)

54106
54106
6493
6493
6493
6493
6493
6493

1cll (with 2 Ca2)

1cll (with 2 Ca2)
1cll
1cll
1prw
1prw
1cff_model#4
1cff_model#4

Length of the MD
100 ns
100 ns
20 ns
20 ns
20 ns
20 ns
20 ns
20 ns

The abbreviations used in the rst column are used through all the text of this manuscript. Refer to the
Methods section for further details.

Near-UV CD spectra were taken using the same buffer and temperature, in the presence and absence of 10 mM CaCl2. The results
were expressed as the mean residue ellipticity.53

Computational Analysis of the Far-UV CD Spectra

The secondary structure fractions of the peptides CaM[54106] and
CaM[54106]pS81 in the absence and presence of Ca2 were determined from CD spectra47 using the CDPro software package that
includes SELCON3, CDSStr, and CONTIN/LL.5458 The SP43
(IBasis 4) and SDP48 (IBasis 7) protein reference sets were used.
SP43 contains 43 proteins having known structures and CD spectra
ranging from 190 to 240 nm. SDP48 contains ve unfolded proteins
in addition to those in SP4357 and was used because it gives better
results with protein fragments and partially unfolded proteins.
These large protein reference sets were chosen since our CD
spectra lacked short-wavelength data (180190 nm). In addition,
K2D,59,60 implemented in DICHROWEB,61,62 and the method
described by Scholtz et al.63 were also used to predict secondary
structure fractions from the CD spectra.

Molecular Dynamics Simulations

Three different experimentally determined structures of Ca2-CaM
were downloaded from the Protein Data Bank (PDB)64: (i) the ensemble of 26 structures determined by NMR in solution by Elshorst
et al.9 (PDB code: 1cff); the crystal structures determined (ii) by
Chattopadhyaya et al.6 (PDB code: 1cll) and (iii) by Fallon and
Quiocho7 (PDB code: 1prw).
Eight MD simulations were carried out for different fragments
of these structures as described in Table I (abbreviations in Table I
will be used throughout the text of this report).
The addition of the phosphate groups for the simulations of the
fragments having the phosphoserine in position 81 (pS81 in Table I)
was done using SYBYL 7.0 (Tripos, St. Louis). Except for some preliminary calculations using GROMOS 96,65,66 all molecular dynamics (MD) simulations were performed with the Gromacs suite of
programs67,68 utilizing the OPLS force eld.69 Force-eld parameters for the phosphoserine were adapted from the phosphotyrosine
parameters used by Price and Jorgensen.70
The protein structures were placed in the center of a dodecahedral box, which was subsequently lled with TIP4P water molecules
(TIP4P transferable intermolecular potentials with four point

Biopolymers (Peptide Science) DOI 10.1002/bip

charges).71 The N-terminus and C-terminus were charged. The size

of the box was chosen so as to contain the protein and at least
1.2 nm of solvent on all sides. The number of water molecules was
roughly 15,000 for the fragments of CaM encompassing residues
54106 and 10,000 for the fragments of CaM encompassing residues
6493. Phosphoserine was considered in its completely dissociated
form (charge:
2) in agreement with the pKa reported for this
residue.72
The peptides 54106 were simulated with two Ca2 atoms taken
from the crystal structure (bound to the Ca2-binding loops II and
III), whereas the peptides encompassing residues 6493 were simulated without Ca2 since these fragments correspond to the central
a-helix of CaM and do not contain Ca2-binding sites. Sodium
counterions were added in each simulation to maintain the overall
charge of the system equal to zero.
Coulomb interactions were treated with fast particle mesh Ewald
(PME),73,74 using a grid spacing of 0.12 nm and cubic interpolation.
The cut-off distances for the Coulomb and Lennard-Jones interactions were 0.8 and 1.4 nm, respectively. Interactions between atoms
that were within 0.8 nm were evaluated every step, while interactions between atoms within the longer cutoff distance were evaluated every 3 steps. Constant pressure and temperature were maintained by a weak coupling of the system to an external bath at 1 bar
and 300 K using the Berendsen barostat and thermostat75 with coupling times of 1 and 0.1 ps, respectively. The bond distances and
bond angles of water were constrained using the SETTLE algorithm.76 All other bond distances were constrained with the LINCS
algorithm.77 Dummy nonpolar hydrogen atoms78 were employed.
The integration time step was 4 fs. Prior to the simulations, the
potential energy of each system was minimized using steepest
descent, followed by a 20-ps MD simulation with position restraints
on the protein to relax the water molecules. 100-ns simulations of
the peptides 1cll[54106]and 1cll[54106]pS81 and 20-ns simulations for the remaining fragments ([6493] of 1cll, 1prw and model
4 of 1cff) were performed (Table I).
The conformations generated in the MD simulations were analyzed using the analysis tools implemented in the Gromacs suite of
programs, such as the secondary structure predictions using DSSP.79

Theoretical Calculations of CD Spectra

The computation of CD spectra from the protein structure was performed using the matrix method80 as described in Sreerama and

376

Settimo et al.

Table II Fractions (Percentages) of Secondary Structure Elements From the Experimental CD Spectra for the Peptides CaM[54106]
and CaM[54106]pS81 in the Absence (apo) or Presence of Ca2 (the Spectra are Reported in Figure 6 in Ref. 47)
CaM[54106]apo

CONTIN/LL
SELCON3
CDSStr
K2D
Scholtz

CaM[54106]-Ca2

CaM[54106]pS81apo

CaM[54106]pS81-Ca2

Total
aa

Total
ba Remainingb

Total
aa

Total
ba Remainingb

Total
aa

Total
ba Remainingb

Total
aa

Total
ba Remainingb

12.1c
8.7d
9.4c
7.4d
6.0c
5.0d
8
9

30.2c
18.1d
29.4c
18.4b
33.0c
21.0d
41

27.1c
25.6d
26.3c
26.2b
30.0c
28.0d
25
25

17.9c
9.0d
18.9c
9.1b
19.0c
10.0d
15

10.0c
6.3d
10.1c
8.6b
5.0c
4.0d
7
7.1

32.6c
17.2d
28.6c
18.1b
32.0c
21.0d
49

21.8c
18.7d
17.3c
17.1
19.0c
18.0d
10
18

24.6c
12.9d
26.7c
18.5
25.0c
15.0d
34

57.7c
73.3d
59.5c
61.6b
60.0c
73.0d
50

55.1c
65.3d
55.0c
63.9b
51.0c
61.0d
60

57.5c
76.6d
55.6c
67.0b
62.0c
74.0d
44

53.5c
68.5d
53.1c
62.3b
56.0c
66.0d
56

Different methods were used as described in the methods section.

a
The total fractions of a-helix and b-structure in CONTIN/LL, SELCON3, and CDSStr implemented in CDPRO are reported, respectively, as the sum of
regular (ar) plus distorted (ad) a-helix and as the sum of regular (br) plus distorted (bd) b-structure.
b
For CONTIN/LL, SELCON3, and CDSStr the remaining corresponds to the sum of turns (T) and unordered (U) whereas for K2D it corresponds to
the prediction of random structure.
c
Analysis using SP43 protein reference set.
d
Analysis using SP48 protein reference set.

Woody.54 In particular, the average CD spectrum over the whole trajectory was calculated with the following steps: (i) generation of the
snapshots corresponding to the structures of the peptides being
simulated picked up every 100 ps for each trajectory (Table I), (ii)
calculation of the CD curve using theoretical quantum-mechanical
methods including aromatic residues54 for each of these frames, (iii)
average of the theoretical CD spectra (for all the frames of the whole
trajectory).

Figures
Figures were prepared using Molscript81 and Raster3D.82

RESULTS AND DISCUSSIONS

Analysis of the Far-UV CD Spectra
Table II reports the percentage of secondary structure elements calculated from the far-UV CD spectra of CaM[54
106] and CaM[54106]pS81 in the absence and presence of
Ca2 (see Figure 6 in Ref. 47) using ve different methods
described in the Methods section. We used more than one
method to predict the secondary structure fractions from
these experimental curves to make the results more robust.
CDPro analysis using the 43-protein basis set gives signicant
b-sheet contents, ranging from ca. 2050% (Table II). These
may result from the limited wavelength range. A basis set
that includes unfolded proteins (48-protein basis) has been
shown57 to give more accurate analysis of secondary structure when studying partially unfolded proteins and peptide

fragments. With this basis set, the b-sheet content is decreased by about 10% for each of the fragments (Table II).
As can be seen in Table II, the predictions from the CD
curves of the peptide CaM[54106] and CaM[54106]pS81
show an increase of the a-helical content in the presence
of Ca2 and a decrease in the a-helical content upon phosphorylation of Ser 81.
In the following sections, we investigate the structural
effects correlated to Ca2 binding and phosphorylation of
Ser 81, using mainly molecular modeling and MD simulations rationalizing the results with the experimental data.

Effect of Ca2 Binding

As can be seen in Table II, all the methods predict an increase
of the a-helical content of the peptide in the presence of
Ca2. For CaM[54106], the a-helical content increases by
1224% in the presence of Ca2. Surprisingly, the addition
of Mg2 instead of Ca2 does not change the CD spectrum
from the prole of apo-CaM[54106] (Figure 6A in Ref. 47).
Interestingly, the CD spectra for the full-length CaM in
the absence and presence of Ca2 and Mg2 are similar to
the curves for our peptide CaM[54106]. In fact, for the fulllength CaM, there is also an intensication of the dichroic
signal at 222 nm content upon Ca2 binding,20,24,83,84 and
there is no change in the ellipticity upon Mg2 binding.85
Ellipticities for CaM can vary with the ionic strength,24 leading to different values of ellipticities and a-helix fractions
reported in different studies (see e.g. Ref. 84). Brokx and
Vogel86 suggested that in CaM, as has been claimed with
Biopolymers (Peptide Science) DOI 10.1002/bip

Conformational Changes in CaM

FIGURE 1 Near-UV CD spectra of peptide CaM[54106] in 10 mM

MOPS, pH 7.5, 0.1M NaCl, 258C in the presence (dashed line) and
absence (continuous line) of 10 mM CaCl2.

Ca2 binding to troponin-C,87 a change in the alignment of

the a-helices with respect to one another could also be responsible for this increase in the CD signal, a mechanism
originally suggested by Manning.88 Even though NMR shows
that Ca2-CaM has a high degree of mobility near the middle of the central helix,11 apparently similar to what is seen
in the apo form,5 some studies clearly show that Ca2 binding rigidies and increases the a-helical content of the central
helix of CaM: (i) a peptide corresponding to the third Ca2binding loop of CaM binds a multivalent cation (lanthanide)
and provides a-helix nucleation,89 CaM[54106] contains
the second and third Ca2-binding loop of CaM at the N-terminus and C-terminus, respectively so a similar effect might
be invoked for both termini of our peptide, which could help
stabilize the central a-helix; (ii) a spin-label electron paramagnetic resonance study shows rigidication of the central
part7681 of CaM upon Ca2 binding19; (iii) Fourier transform infrared spectroscopy studies show an increased portion
of a-helix in the presence of Ca2 21,22; (iv) a study in which
optical rotatory dispersion and proteolysis with trypsin were
used shows an increase of the a-helix content upon Ca2
binding25; (v) uorescence studies23,24 show that the central
helix of CaM undergoes conformational changes upon Ca2
binding, and in particular Yao and coworkers report that the
central helix becomes more extended and rigid.23
To further investigate experimentally the effect of Ca2
binding, we measured the near-UV CD in the absence and
presence of Ca2 for CaM[54106] (Figure 1).
The near-UV CD spectra in the presence and absence of
Ca2 for CaM[54106] are very similar to the published
Biopolymers (Peptide Science) DOI 10.1002/bip

377

near-UV CD spectra for the full-length CaM84,85,90: the

stronger intensities in the near-UV CD spectrum in the presence of Ca2 show that the aromatic side chains in our peptide (Phe 65, 68, 89, 92, and Tyr 99) are in a more rigid structure. In addition, as mentioned earlier, it is interesting to see
that CaM[54106] can distinguish between Ca2 and Mg2.
Therefore, the radius of the metal is very important in this
conformational change, as already anticipated by Chao
et al.91 Ca2 has a larger radius than Mg2 (ionic radii are 0.99
and 0.65 A for Ca2 and Mg2, respectively92). Therefore for
Ca2, the charge is distributed in a larger volume and over a
larger surface in comparison to Mg2. As a consequence, a
larger number of residues (at least ve in each Ca2-binding
loop of CaM) can interact with Ca2, in comparison with the
smaller Mg2. Thus, the conformational change is expected to
take place when more residues are involved in the contact with
the metal ion.
The N-terminal half of the CaM[54106] fragment, from
residue 54 to 76 of the structure determined by Chattopadhyaya et al.6 is displayed in Figure 2. The formation of a
hydrophobic cluster when Ca2 is bound is evident. This
cluster, formed by Ile 63 and Val 55 of the Ca2-binding loop
II with Phe 68 and Met 72 in the N-terminal part of the central helix, nucleates the formation and stabilization of the
central a-helix because it leads to the stacking of hydrophobic residues Phe 68, Met 72, and Met 76 parallel to the axis
(Figure 2). The same effect can be described also in the C-terminal part of our peptides. The importance of helix-nucleation upon Ca2 binding was also reported by Siedlecka et al.
for a peptide encompassing the third Ca2-binding loop of
CaM.89
In addition, in our simulations for the peptide [54106]
in the presence of Ca2, these ions are bound to the Ca2binding loops II and III in our peptide (1cll[54106] and
1cll[54106]pS81, Table I) even after 100 ns, showing that
the Ca2 is stably bound.

Effect of Phosphorylation of Ser 81

The phosphorylation of Ser 81 has an important effect on
the a-helical content of Ca2-CaM[54106] (Table II): phosphorylation of Ser 81 decreases the helical content by 515%
in comparison to the wild-type peptide, according to the
analyses of the far-UV CD spectra (Table II).
Quadroni et al. have measured and compared the CD
spectra of Ca2-CaM and Ca2-phosphorylated CaM.93 The
phosphorylation described in Ref. 93 was obtained by treatment of CaM with CK2, a kinase that extensively phosphorylates Ser 81 in the central helix of CaM.44,46 Interestingly, the intensity of the CD signal at 222 nm is stronger for

378

Settimo et al.

FIGURE 2 Structure showing the nucleation of the a-helix provided by the binding of Ca2 in the N-terminal half of the fragment
54106 taken from the structure of CaM (PDB code: 1cll). Backbone is represented as tube colored in green; carbon atoms colored
in gray; Ca2 in the Ca2-binding loop II colored in orange. The
coordination of the Ca2 atom is displayed by black dashed lines.

the wild-type Ca2-CaM in comparison to Ca2-phosphorylated CaM (Figure 5 in Ref. 93), as we observed for
CaM[54106] and CaM[54106]pS81 (Figure 6B in our previous study47).
The lower intensity at 222 nm in the CD spectra seen for
Ca2-CaM[54106]pS81 in comparison to Ca2-CaM[54
106] is not correlated with weaker binding of Ca2 to the
phosphopeptide in comparison to the wild-type peptide, but

results from the phosphorylation of Ser 81 because we have

conrmed, using radiolabeling,94 that both CaM[54106]
and CaM[54106]pS81 bind Ca2 with similar afnity
(Figure 1 in the supplementary material section), in agreement with published calcium binding studies on the wildtype and phosphorylated CaM.93,95
To investigate the structural effects correlated to the phosphorylation of Ser 81, we performed MD simulations on
CaM fragments encompassing the central a-helix of CaM,
having serine or phosphoserine in position 81 (pSer 81). To
explore as much as possible all the possible states in which
the central helix can exist, we performed MD simulations
starting with different fragments taken from several experimentally determined structures of CaM (Table I; see Methods
section). We considered both crystal structures and also a
structure determined by NMR, since the middle part of the
central a helix of CaM is very exible in solution. We calculated the CD spectra for snapshots corresponding to the
structures of the peptides being simulated for each trajectory
using the matrix method80 and compared the calculated
spectra with the experimental curves reported previously
(Figure 6 in Ref. 47). The matrix method for predicting CD
spectra54 is a quantum mechanical method in which the perturbation matrix is diagonalized and the eigenfunctions are
used to calculate the electric and magnetic dipole transition
moments, and the eigenvalues give the transition energies.
Simulations Done With Fragments of a Crystal Structure of
CaM (PDB Code 1cll Starting Structures). Initially, we performed 60-ns simulations for the fragment [54106], wildtype and with phosphoserine (including Ca2) starting from
the structure solved by Chattopadhyaya et al.6 (PDB code:

FIGURE 3 Secondary structure analysis (by DSSP) for the simulation of 1cll[6493] (a) and
1cll[6493]pS81 (b) as a function of the simulation time.

Biopolymers (Peptide Science) DOI 10.1002/bip

Conformational Changes in CaM

379

FIGURE 4 Flexibility of the central a-helix of CaM. Change in the interhelical angle (angle
between the helical segment 6474 and 8393) as a function of the simulation time for 1cll[64
93](continuous line) and 1cll[6493]pS81 (dashed line).

1cll) using the GROMOS 96 force eld (43a1).65,66 We

observed the formation of p-helix during both simulations
(data not shown). It has been reported that some force-elds
(e.g. GROMOS 96) generate p-helix during the MD simulations, but this has been shown to be an artifact not supported
by experimental data.96,97 Upon changing the force-eld to
OPLS and rerunning the simulations (1cll[54106] ,1cll[54
106]pS81,1cll[6493] and 1cll[6493]pS81 in Table I), we
did not observe the formation of p-helix (Figure 3).
CD spectra averaged over the whole trajectory (see Methods section for details) of 1cll[6493] and 1cll[6493]pS81
are very similar to the CD spectra averaged over the whole
trajectory for 1cll[54106] and 1cll[54106]pS81, respectively. In addition, we observed that the removal of Ca2binding loops II and III do not affect the structure of the central a-helix during the MD simulation. Therefore, the CD
signal is mainly generated by the central a-helix of CaM and
not by the Ca2-binding loops, so we decided to simulate
only the central a-helix 6493 of CaM to study the local
effect of the phosphorylation of Ser 81.
The structure reported by Chattopadhyaya et al.6 has a
straight a-helix, which is unlikely to be observed in solution.11
However, there is some experimental evidence for a straight
central a-helix in at least a fraction of the molecules in solution.19 Interestingly, the central helix of 1cll[6493] bends in
its middle between 6 and 15 ns in the simulation (Figures 3a
Biopolymers (Peptide Science) DOI 10.1002/bip

and 4), in agreement with the NMR structure reported for the
full-length CaM.11 Surprisingly, this bending is not seen for
1cll[6493]pS81 (Figures 3b and 4), where a stable interaction
between the phosphate group of pSer 81 and the amino group
of Lys 77 rigidies the central a-helix of CaM, hindering the
bending. However, the calculated CD spectra over the whole
trajectory for the simulations of 1cll[6493] and 1cll[64
93]pS81 have similar ellipticity at 222 nm (Figure 2 in the supplementary material section) and therefore differ from the experimental data (Figure 6 in Ref. 47).
Simulations Done With Fragments of a Crystal Structure of
CaM (PDB Code 1prw Starting Structures). If a phosphate
group is added to the hydroxyl group of Ser 81 in the crystal
structure of CaM determined by Fallon and Quiocho,7 there
is a destabilizing interaction between the negative charges of
the phosphate group and the negative partial charge associated with the helix-dipole moment in the C-terminal of
the helix6476 (Figure 5). This destabilizes the helical portion
6476 in 1prw[6493]pS81, in agreement with the fact that
the phosphorylated peptide of CaM possesses less helical
content than the wild-type peptide.
The a-helix is exible in the middle during the simulation
(see the large changes in interhelical angle in Figure 6) in
agreement with the experimental evidence mentioned before
for the central a-helix of the full-length CaM. In 1prw[64

380

Settimo et al.

93]pS81, the central helix is interrupted by a region lacking

secondary structure (white strip in Figure 7). The calculated
CD curves (average over the whole trajectory) (Figure 8) are
in agreement with the experimental curves that show higher
ellipticity for wild-type than for the phosphorylated peptide.
In 1prw[6493], the hydroxyl group of Ser 81 donates a
hydrogen bond to the main-chain oxygen of Asp78 throughout
the trajectory as captured in the snapshot of the structure
seen at 10 ns (Figure 9a). In 1prw[6493]pS81, the hydroxyl
oxygen cannot donate a hydrogen bond to the main-chain
oxygen (Figure 9b). Instead, the phosphate group of phosphoserine 81 can interact with the sidechain of Lys 75, with
consequent rigidication of the structure (seen in Figure 6
as mentioned above) and stabilization of a loop (Figure 9b)
that is seen throughout the simulation (white strip in Figure 7).

FIGURE 5 The structure of 1prw[6493]pS81 at the beginning of

the simulation. The addition of a phosphate group to the fragment
1prw[6493] (coordinates from the structure having PDB code: 1prw),
gives an unfavorable interaction (as indicated by the asterix *) with
the C-terminal negative charge provided by the dipole moment of helix
[6476] (helix colored in blue). The helical dipole moments (arrows)
with their partial charges are indicated for the N-terminal helix (in
blue) and C-terminal helix (in red) of fragment [6493].

Simulations Done With Fragments of a Model From an Ensemble of CaM Structures Solved by NMR (PDB Code 1cff
Starting Structures). Elshorst et al.9 determined by NMR
the structure of a complex of Ca2-CaM and reported an ensemble of 26 models (PDB code 1cff).
We calculated the theoretical CD spectra for the fragment
54106 of each of these 26 models. The fragment [54106]
of model 4 from 1cff produces a CD spectrum very similar to
that of Ca2-CaM[54106] reported by us47 and has favor-

FIGURE 6 Flexibility of the central a-helix of CaM. Change in the interhelical angle (angle
between the helical segment 6474 and 8393) as a function of the simulation time for 1prw[64
93] (continuous line) and 1prw[6493]pS81 (dashed line).

Biopolymers (Peptide Science) DOI 10.1002/bip

Conformational Changes in CaM

381

FIGURE 7 Secondary structure analysis (by DSSP) for the simulation of 1prw[6493] (a) and
1prw[6493]pS81 (b) as a function of the simulation time.

able F and C angles in the Ramachandran plot (Figure 3 in

the supplementary material section).
The secondary structural elements of 1cff_4[6493] do
not change signicantly during the simulation, except for the
formation of a bend that is not persistent (Figure 10a). By
contrast 1cff_4[6493]pS81 forms a stable bend in the middle of the central a-helix (green strip in Figure 10b). This
bend is stabilized by an interaction between the side chain of
phosphoserine 81 and Arg 86 throughout the simulation.

In this case, the calculated CD curves over the whole trajectory (Figure 11) are in good agreement not only with the
intensities but also with the shapes of the experimental
curves (Figure 6 in Ref. 47).
In agreement with the results of our simulations, one possible reason for the shape of the experimental curves (more
intense peak at 205 nm in comparison to the peak at 222
nm) is that the main-chain oxygen atoms in the middle of
the central a-helix of CaM are accessible to the solvent and

FIGURE 8 Average of the theoretical curves for the whole trajectory of the peptide 1prw[6493]
(continuous line) and 1prw[6493]pS81 (dashed line).

Biopolymers (Peptide Science) DOI 10.1002/bip

382

Settimo et al.

establish hydrogen bonds to water molecules in addition to

main-chain i ? i 4 hydrogen-bonds.98,99

FIGURE 9 Frame-snapshots observed at 10 ns of simulation for

1prw[6493] (a) and 1prw[6493]pS81 (b). Key hydrogen bonds
established by the side-chain of residue 81 are shown.

Comparison of the Results From the MD Simulations. The

validity of our MD simulations was assessed by different
methods,100 e.g., by running simulations on different starting
conformations and for different simulation times, as can be
seen in Table I.
Both the simulations done on the fragments taken from
1prw and 1cff predicted correctly the stronger intensity at
222 nm of the dichroic signal for the wild-type peptide relative to the phosphopeptide. However, the averaged CD spectra calculated from the simulations of the fragments of 1cff
showed the best agreement. Even the shape of the curves,
with a more intense peak at 205 nm in comparison to the
peak at 222 nm, was predicted correctly. In addition, the percentages of the secondary structure elements seen during this
simulation (Figure 10) agree quite well with the data
reported in Table II. The improved results with this starting
structure may be due to the fact that the structure was determined in solution.
Conversely, the simulations of the peptide from the structure 1cll (having the continuous central a-helix) give results
that agree poorly with experiment. We have extended the simulation of 1cll[54106] to 100 ns and we observed results very
similar to those for the shorter fragment 1cll[6493] (e.g. even
at 100 ns 1cll[54106]pS81 had a straight central a-helix as in
the starting structure). However, this fact could be consistent
with the nding reported by Qin and Squier that there is a
population of Ca2-CaM having a straight central a-helix.19

FIGURE 10 Secondary structure analysis (by DSSP) for the simulation of 1cff_4[6493] (a) and
1cff_4[6493]pS81 (b) as a function of the simulation time.

Biopolymers (Peptide Science) DOI 10.1002/bip

Conformational Changes in CaM

383

FIGURE 11 Average of the theoretical curves for the whole trajectory of the peptide 1cff_4[64
93] (continuous line) and 1cff_4[6493]pS81 (dashed line).

In all the MD simulations of the phosphopeptide, we

observed the formation of an interaction between the side
chain of one lysine or arginine residue and the phosphate
group of phosphoserine 81. In general, this interaction rigidies the central a-helix of CaM, decreasing the exibility of
the central a-helix and most likely affecting the recognition
and activation to the CaM-dependent targets as seen experimentally.42 An additional experimental result supports this
model. We have reported47 that CaM[54106]pS81 cannot
be phosphorylated at Thr 79, despite the fact that a phosphorylated side-chain at n 2 is a strong positive determinant
for CK2 phosphorylation of short peptides.49 In our simulation, the negatively charged phosphate group of phosphoserine 81 interacts with a positively charged residue (Lys or
Arg), which would preclude the interaction of phosphoserine
81 with CK2. The phosphate group of pSer 81 accepts a
hydrogen bond from the Lys/Arg residue rather than interacting with the basic residues near the active site of CK2
where the central helix of CaM is predicted to bind.48 In
addition, the basic residues of CK2 that are thought to recognize the acidic residues in the consensus sequence of phosphorylation48 would interact unfavorably with the Lys/Arg
that forms a salt bridge with pSer 81. As a consequence, the
binding of CaM[54106]pS81 to CK2 would be impaired,
with a consequent decrease of the phosphorylation of Thr
79. We ran some MD simulations with fragments containing
pThr 79, and found also in this case a similar phenomenon
Biopolymers (Peptide Science) DOI 10.1002/bip

in which the phosphate group interacts with Lys or Arg residues. This agrees with the absence of peptides bis-phosphorylated in positions 79 and 81.47 The rigidifying stabilization of helical structure given by a salt bridge between a
phosphoserine and lysine residues has been reported in the
literature51; interestingly, also in our simulations done with
1cll fragments having pSer 81 or pThr 79, we observe similar
interactions in which the phosphorylated residue interacts
with the side chain of the lysine located at i-4 position, with
the formation of a salt bridge between Lys 75 and pThr 79 or
between Lys 77 and pSer 81.

CONCLUSIONS
The synthetic fragment encompassing the central helix of
CaM and the Ca2-binding loops II and III gives far-UV and
near-UV CD spectra similar to those of CaM when (i) Ca2
is bound and (ii) it is phosphorylated in position 81, a residue highly phosphorylated by CK2. We have explained with
modeling and MD simulations the structural reasons for
these effects. In particular, when Ca2 binds to loops II and
III, near-UV CD indicates that there is rigidication of the
structure, similar to that seen for the full-length CaM, and
we suggest that important interactions between hydrophobic
residues nucleate the formation of a-helix. Introduction of a
phosphate at Ser 81 affects the stabilization of a-helix by
inuencing the hydrogen-bond network established by the

384

Settimo et al.

side-chain of residue 81. Quantum-mechanical calculations

of the CD spectra on snapshots during the MD simulations
give results in agreement with the experimental intensities
that we reported for the peptide CaM[54106] and CaM[54
106]pS81. It is very important to run different MD simulations on different starting structures for CaM, since the central part of the central a-helix of CaM has a peculiar exibility in solution.
The conformational changes that take place in this peptide upon Ca2 binding and phosphorylation might also
occur in the full-length CaM. The advantage of using a relatively small fragment of CaM to perform these structural
investigations is that it is easier to study with molecular
modelling and MD simulations in comparison to the fulllength protein.
We thank Prof. Lorenzo A. Pinna for useful discussions about the
study and to Dr. Dan Price for providing the parameters for the
construction of the topology for phosphoserine for the simulation
with OPLS. Also thanks to Dr. Alice Glattli for additional discussions on the formation of p-helix using GROMOS 96 force eld.
Access to the software was provided by the Centre for Scientic
Computing (CSC, Espoo, Finland).

REFERENCES
1. Van Eldik, L.; Watterson, D. Calmodulin and Signal Transduction; Academic Press: San Diego, CA, 1998.
2. Finn, B. E.; Forsen, S. Structure 1995, 3, 711.
3. Yang, C.; Jas, G. S.; Kuczera, K. Biochim Biophys Acta 2004,
1697, 289300.
4. Weljie, A. M.; Clarke, T. E.; Juffer, A. H.; Harmon, A. C.; Vogel,
H. J. Proteins 2000, 39, 343357.
5. Kuboniwa, H.; Tjandra, N.; Grzesiek, S.; Ren, H.; Klee, C. B.;
Bax, A. Nat Struct Biol 1995, 2, 768776.
6. Chattopadhyaya, R.; Meador, W. E.; Means, A. R.; Quiocho, F. A.
J Mol Biol 1992, 228, 11771192.
7. Fallon, J. L.; Quiocho, F. A. Structure 2003, 11, 13031307.
8. Meador, W. E.; Means, A. R.; Quiocho, F. A. Science 1992, 257,
12511255.
9. Elshorst, B.; Hennig, M.; Forsterling, H.; Diener, A.; Maurer,
M.; Schulte, P.; Schwalbe, H.; Griesinger, C.; Krebs, J.; Schmid,
H.; Vorherr, T.; Carafoli, E. Biochemistry 1999, 38, 12320
12332.
10. Zhang, M.; Tanaka, T.; Ikura, M. Nat Struct Biol 1995, 2, 758
767.
11. Barbato, G.; Ikura, M.; Kay, L. E.; Pastor, R. W.; Bax, A. Biochemistry 1992, 31, 52695278.
12. Tjandra, N.; Kuboniwa, H.; Ren, H.; Bax, A. Eur J Biochem
1995, 230, 10141024.
13. Jaren, O. R.; Kranz, J. K.; Sorensen, B. R.; Wand, A. J.; Shea, M. A.
Biochemistry 2002, 41, 1415814166.
14. Ikura, M.; Spera, S.; Barbato, G.; Kay, L. E.; Krinks, M.; Bax, A.
Biochemistry 1991, 30, 92169228.
15. Brokx, R. D.; Scheek, R. M.; Weljie, A. M.; Vogel, H. J. J Struct
Biol 2004, 146, 272280.

16. Torok, K.; Lane, A. N.; Martin, S. R.; Janot, J. M.; Bayley, P. M.
Biochemistry 1992, 31, 34523462.
17. Heidorn, D. B.; Trewhella, J. Biochemistry 1988, 27, 909915.
18. Seaton, B. A.; Head, J. F.; Engelman, D. M.; Richards, F. M. Biochemistry 1985, 24, 67406743.
19. Qin, Z.; Squier, T. C. Biophys J 2001, 81, 29082918.
20. Bayley, P.; Martin, S.; Jones, G. FEBS Lett 1988, 238, 6166.
21. Trewhella, J.; Liddle, W. K.; Heidorn, D. B.; Strynadka, N. Biochemistry 1989, 28, 12941301.
22. Jackson, M.; Haris, P. I.; Chapman, D. Biochemistry 1991, 30,
96819686.
23. Yao, Y.; Schoneich, C.; Squier, T. C. Biochemistry 1994, 33,
77977810.
24. Sun, H.; Yin, D.; Squier, T. C. Biochemistry 1999, 38, 12266
12279.
25. Liu, Y. P.; Cheung, W. Y. J Biol Chem 1976, 251, 41934198.
26. Tabernero, L.; Taylor, D. A.; Chandross, R. J.; VanBerkum, M. F.;
Means, A. R.; Quiocho, F. A.; Sack, J. S. Structure 1997, 5,
613622.
27. Sorensen, B. R.; Shea, M. A. Biophys J 1996, 71, 34073420.
28. VanScyoc, W. S.; Sorensen, B. R.; Rusinova, E.; Laws, W. R.;
Ross, J. B.; Shea, M. A. Biophys J 2002, 83, 27672780.
29. Kataoka, M.; Head, J. F.; Persechini, A.; Kretsinger, R. H.;
Engelman, D. M. Biochemistry 1991, 30, 11881192.
30. Raghunathan, S.; Chandross, R. J.; Cheng, B. P.; Persechini, A.;
Sobottka, S. E.; Kretsinger, R. H. Proc Natl Acad Sci USA 1993,
90, 68696873.
31. Weber, P. C.; Lukas, T. J.; Craig, T. A.; Wilson, E.; King, M. M.;
Kwiatkowski, A. P.; Watterson, D. M. Proteins 1989, 6, 7085.
32. Persechini, A.; Kretsinger, R. H. J Biol Chem 1988, 263, 12175
12178.
33. Mehler, E. L.; Pascual-Ahuir, J. L.; Weinstein, H. Protein Eng
1991, 4, 625637.
34. Vorherr, T.; Kessler, O.; Mark, A.; Carafoli, E. Eur J Biochem
1992, 204, 931937.
35. van der Spoel, D.; de Groot, B. L.; Hayward, S.; Berendsen,
H. J.; Vogel, H. J. Protein Sci 1996, 5, 20442053.
36. Wriggers, W.; Mehler, E.; Pitici, F.; Weinstein, H.; Schulten, K.
Biophys J 1998, 74, 16221639.
37. Yang, C.; Jas, G. S.; Kuczera, K. J Biomol Struct Dyn 2001, 19,
247271.
38. Fiorin, G.; Biekofsky, R. R.; Pastore, A.; Carloni, P. Proteins
2005, 61, 829839.
39. Likic, V. A.; Gooley, P. R.; Speed, T. P.; Strehler, E. E. Protein Sci
2005, 14, 29552963.
40. Yang, C.; Kuczera, K. J Biomol Struct Dyn 2002, 19, 801819.
41. Yang, C.; Kuczera, K. J Biomol Struct Dyn 2002, 20, 179197.
42. Benaim, G.; Villalobo, A. Eur J Biochem 2002, 269, 36193631.
43. Nakajo, S.; Hayashi, K.; Daimatsu, T.; Tanaka, M.; Nakaya, K.;
Nakamura, Y. Biochem Int 1986, 13, 687693.
44. Meggio, F.; Brunati, A. M.; Pinna, L. A. FEBS Lett 1987, 215,
241246.
45. Nakajo, S.; Masuda, Y.; Nakaya, K.; Nakamura, Y. J Biochem
(Tokyo) 1988, 104, 946951.
46. Sacks, D. B.; Davis, H. W.; Crimmins, D. L.; McDonald, J. M.
Biochem J 1992, 286 (Part 1), 211216.
47. Arrigoni, G.; Marin, O.; Pagano, M. A.; Settimo, L.; Paolin, B.;
Meggio, F.; Pinna, L. A. Biochemistry 2004, 43, 1278812798.
48. Rekha, N.; Srinivasan, N. BMC Struct Biol 2003, 3, 4.

Biopolymers (Peptide Science) DOI 10.1002/bip

Conformational Changes in CaM

49. Perich, J. W.; Meggio, F.; Reynolds, E. C.; Marin, O.; Pinna, L.
A. Biochemistry 1992, 31, 58935897.
50. Andrew, C. D.; Warwicker, J.; Jones, G. R.; Doig, A. J. Biochemistry 2002, 41, 18971905.
51. Errington, N.; Doig, A. J. Biochemistry 2005, 44, 75537558.
52. Toumadje, A.; Alcorn, S. W.; Johnson, W. C., Jr. Anal Biochem
1992, 200, 321331.
53. Schmid, F. X. Spectral Methods of Characterizing Protein Conformation and Conformational Changes; IRL Press: Oxford,
UK, 1989.
54. Sreerama, N.; Woody, R. W. Methods Enzymol 2004, 383, 318
351.
55. Sreerama, N.; Venyaminov, S. Y.; Woody, R. W. Anal Biochem
2001, 299, 271274.
56. Sreerama, N.; Woody, R. W. Anal Biochem 2000, 287, 252260.
57. Sreerama, N.; Venyaminov, S. Y.; Woody, R. W. Anal Biochem
2000, 287, 243251.
58. Sreerama, N.; Venyaminov, S. Y.; Woody, R. W. Protein Sci
1999, 8, 370380.
59. Andrade, M. A.; Chacon, P.; Merelo, J. J.; Moran, F. Protein Eng
1993, 6, 383390.
60. Merelo, J. J.; Andrade, M. A.; Prieto, A.; Moran, F. Neurocomputing 1994, 6, 443454.
61. Lobley, A.; Wallace, B. A. Biophys J 2001, 80, 373a.
62. Lobley, A.; Whitmore, L.; Wallace, B. A. Bioinformatics 2002,
18, 211212.
63. Scholtz, J. M.; Qian, H.; York, E. J.; Stewart, J. M.; Baldwin, R. L.
Biopolymers 1991, 31, 14631470.
64. Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.;
Weissig, H.; Shindyalov, I. N.; Bourne, P. E. Nucleic Acids Res
2000, 28, 235242.
65. Tironi, I. G.; van Gunsteren, W. F. Mol Phys 1994, 83, 381403.
66. van Gunsteren, W. F.; Billeter, S. R.; Eising, A. A.; Hunenberger,
P. H.; Kruger, P.; Mark, A. E.; Scott, W. R. P.; Tironi, I. G. Biomolecular Simulation: The GROMOS96 Manual and User
Guide; Verlag der Fachvereine: Zurich, Switzerland, 1996.
67. Berendsen, H. J. C.; van der Spoel, D.; van Drunen, R. Comput
Phys Comm 1995, 91, 4356.
68. van der Spoel, D.; van Buuren, A. R.; Apol, E.; Meulenhoff,
P. J.; Tieleman, D.; Sijbers, A. L. T.; Hess, B.; Feenstra, K. A.;
Lindhal, E.; van Drunen, R.; Berendsen, H. J. C. Gromacs User
Manual, Version 3.0; University of Groningen: Groningen, The
Netherlands, 2001.
69. Jorgensen, W. L.; Tirado-Rives, J. J Am Chem Soc 1988, 110,
16571666.
70. Price, D. J.; Jorgensen, W. L. J Comput Aided Mol Des 2001,
15, 681695.
71. Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey,
R. W.; Klein, M. L. J Chem Phys 1983, 79, 926935.
72. Folsch, G.; Osterberg, R. J Biol Chem 1959, 234, 22982303.

Biopolymers (Peptide Science) DOI 10.1002/bip

385

73. Darden, T.; York, D.; Pedersen, L. J Chem Phys 1993, 98, 10089
10092.
74. Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.;
Pedersen, L. G. J Chem Phys 1995, 103, 85778593.
75. Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; Di
Nola, A.; Haak, J. R. J Chem Phys 1984, 81, 36843690.
76. Miyamoto, S.; Kollman, P. A. J Comput Chem 1992, 13, 952962.
77. Hess, B.; Bekker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M.
J Comput Chem 1997, 18, 14631472.
78. Feenstra, K. A.; Hess, B.; Berendsen, H. J. C. J Comput Chem
1999, 20, 786798.
79. Kabsch, W.; Sander, C. Biopolymers 1983, 22, 25772637.
80. Bayley, P. M.; Nielsen, E. B.; Schellman, J. A. J Phys Chem 1969,
73, 228243.
81. Kraulis, J. J Appl Crystallogr 1991, 24, 946950.
82. Merritt, E. A. a. B., David J. Methods Enzymol 1997, 277, 505
524.
83. Klee, C. B. Biochemistry 1977, 16, 10171024.
84. Martin, S. R.; Bayley, P. M. Biochem J 1986, 238, 485490.
85. Martin, S. R.; Masino, L.; Bayley, P. M. Protein Sci 2000, 9,
24772488.
86. Brokx, R. D.; Vogel, H. J. Protein Sci 2000, 9, 964975.
87. Gagne, S. M.; Tsuda, S.; Li, M. X.; Chandra, M.; Smillie, L. B.;
Sykes, B. D. Protein Sci 1994, 3, 19611974.
88. Manning, M. C. J Pharm Biomed Anal 1989, 7, 11031119.
89. Siedlecka, M.; Goch, G.; Ejchart, A.; Sticht, H.; Bierzyski, A.
Proc Natl Acad Sci USA 1999, 96, 903908.
90. Crouch, T. H.; Klee, C. B. Biochemistry 1980, 19, 36923698.
91. Chao, S. H.; Suzuki, Y.; Zysk, J. R.; Cheung, W. Y. Mol Pharmacol 1984, 26, 7582.
92. Pauling, L. The Nature of the Chemical Bond and the Structure
of Molecules and Crystals: An Introduction to Modern Structural Chemistry, 3rd ed.; Cornell University Press: Ithaca, NY,
1939; p 514.
93. Quadroni, M.; LHostis, E. L.; Corti, C.; Myagkikh, I.; Durussel,
I.; Cox, J.; James, P.; Carafoli, E. Biochemistry 1998, 37, 6523
6532.
94. Kawasaki, H.; Kasai, H.; Okuyama, T. Anal Biochem 1985, 148,
297302.
95. Aiuchi, T.; Hagiwara, T.; Omata, K.; Nakaya, K.; Nakamura, Y.
Biochem Int 1991, 23, 145149.
96. Feig, M.; MacKerell, A. D.; Brooks, C. L., III. J Phys Chem B
2003, 107, 28312836.
97. Freedberg, D. I.; Venable, R. M.; Rossi, A.; Bull, T. E.; Pastor,
R. W. J Am Chem Soc 2004, 126, 1047810484.
98. Blundell, T.; Barlow, D.; Borkakoti, N.; Thornton, J. Nature
1983, 306, 281.
99. Manning, M. C.; Woody, R. W. Biopolymers 1991, 31, 569586.
100. van Gunsteren, W. F.; Mark, A. E. J Chem Phys 1998, 108,
61096116.