Influence of modulated structural dynamics on the
kinetics of a-chymotrypsin catalysis
Insights through chemical glycosylation, molecular dynamics and
domain motion analysis
Ricardo J. Sola
´
and Kai Griebenow
Laboratory for Applied Biochemistry and Biotechnology, Department of Chemistry, University of Puerto Rico, Rı
´
o Piedras Campus, San Juan,
PR, USA
Unraveling the general mechanisms by which enzymes
catalyze chemical reactions is fundamental to the
understanding of biochemical processes. While the
chemical basis of enzyme catalysis is largely under-
stood the same cannot be said about the influence of
the intrinsic protein structural dynamics on enzyme
catalysis [1–4]. Although it has been known for dec-
ades that proteins are highly dynamic molecules which
undergo a variety of structural motions [5,6] only
recently has the relationship between protein structural
Keywords
a-chymotrypsin; enzyme catalysis;
glycosylation; molecular dynamics; serine
protease
Correspondence
K. Griebenow, Department of Chemistry,
University of Puerto Rico, Rı
´
o Piedras
Campus, Facundo Bueso Bldg Laboratory-

3
) without affecting sub-
strate binding (K
S
) at increasing glycosylation levels. Statistical correlation
analysis between the catalytic (DG
„
k
i
) and structurally dynamic (DG
HX
)
parameters determined revealed that the enzyme acylation and deacylation
steps are directly influenced by the changes in protein structural dynamics.
Molecular modelling of the a-CT glycoconjugates coupled with molecular
dynamics simulations and domain motion analysis employing the Gaussian
network model revealed structural insights into the relation between the
protein’s surface glycosylation, the resulting structural dynamic changes,
and the influence of these on the enzyme’s collective dynamics and catalytic
residues. The experimental and theoretical results presented here not only
provide fundamental insights concerning the influence of glycosylation on
the protein biophysical properties but also support the hypothesis that for
a-CT the global structural dynamics directly influence the kinetics of
enzyme catalysis via mechanochemical coupling between domain motions
and active site chemical groups.
Abbreviations
a-CT, a-chymotrypsin; exchange, kinetics (k
HX
); GNM, Gaussian network model; H ⁄ D, hydrogen ⁄ deuterium; MD, molecular dynamics; pNA,
p-nitroanilide; QM, quantum mechanics; Suc, N-succinyl; SBzl, thio-benzyl; SS-mLac, mono-(lactosylamido)-mono-(succinimidyl) suberate;

ate (EP
2
)
TET2
, and liberation of the reaction’s second
product with restoration of the original free enzyme.
From a structural perspective a-CT is composed of
two six-stranded b-barrel domains with the nature of
its collective structural dynamics being attributed to
interdomain hinge-bending motions [16,20,21]. Due to
the location of the active site residues at the interface
between these two structurally rigid b-sheet domains it
has been suggested that global structural flexibility
could directly influence their displacements, thus
impacting the reaction kinetics [16,21–23]. Theoretical
free-energy calculations of the catalytic cycle for struc-
turally related serine proteases (trypsin, elastase) have
also suggested the necessity of structural displacements
for the catalytic residues so that acylation and deacyla-
tion can take place [24–29]. Thus, both local active site
residues and global domain motions are thought to be
implicated in the catalytically relevant structural
dynamics of the enzyme.
The influence of structural dynamics on the
enzyme’s kinetics has also been suggested in previous
experimental works. From
1
H-NMR studies on the
His57–Asp102 low barrier hydrogen bond, Frey and
coworkers proposed the involvement of a conforma-

Due to the well documented effect of natural glycans
in modulating glycoprotein structural dynamics and
function [32–35], chemical glycosylation represents a
straightforward methodology to study the role of pro-
tein structural dynamics on enzyme catalysis [36].
Herein we designed a series of differentially glycosylat-
ed a-CT variants with sequentially reduced structural
dynamics through chemical glycosylation with mono-
functionally activated glycans of differing molecular
masses [36,37]. These were employed in this work to
address experimentally the questions of whether and
how the enzyme’s structural dynamics influence the
kinetics of a-CT catalysis. This was done by determin-
ing the changes in the global structural dynamics
(DG
HX
) [38] for the various chemically glycosylated
a-CT conjugates through amide hydrogen ⁄deuterium
(H ⁄ D) exchange kinetic (k
HX
) experiments and then
performing statistical correlation analysis with their
kinetic parameters (K
S
, k
2
, and k
3
) for the hydrolysis
of N-succinyl-Ala-Ala-Pro-Phe p-nitroanilide (Suc-Ala-

that could potentially alter the substrate binding
affinities of the conjugates and thus impact their cata-
lytic behavior. The chemistry used for chemical glyco-
sylation is based on the succinimidyl functionality
(Fig. 2) which allows coupling of the glycans to the
protein surface via the lysine e-amino groups at pH 9
and above (Table 1). The resulting conjugates are het-
erogeneous mixtures of noncrosslinked single protein
species characterized by a variable distribution of gly-
cans attached to the protein’s surface. Average glycan
molar contents for these a-CT glycoconjugates were
sequentially increased to levels of around 7–8 mol of
glycan per mol of protein. This is approximately 50–
60% of the total glycan content that can theoretically
be attached to a-CT by the chemistry employed
because the protein has 14 surface accessible lysine
residues. Previous structural characterizations revealed
that protein structural integrity was not adversely
impacted during the chemical glycosylation and that
the thermodynamic stability of the conjugates was
increased with increasing glycosylation [36,37].
Changes in a-CT’s structural dynamics upon
chemical glycosylation
Determination of H ⁄ D exchange kinetics represents
one of the principal techniques for the experimental
measurement of changes in protein structural dynamics
[9,34,36,38–46]. Due to the heterogeneous nature of
the glycoconjugates we chose to determine the global
amide H ⁄ D exchange rates by FTIR spectroscopy
[7,36,44,45]. These measurements thus represent the

ously described by us [36].
Additionally, molecular models of the Lac-a-CT
glycoconjugates (Fig. 4) were constructed based on the
lysine reactivity index presented in Table 1 (see below)
to provide a detailed picture of the possible changes in
structural dynamics upon chemical glycosylation. These
glycoconjugate structures were then subjected to
conformational energetic equilibration by molecular
dynamics (MD) simulation methods (Fig. S2). Models
Fig. 2. Succinimidyl activated lactose mole-
cule (SS-mLac) employed for the chemical
glycosylation of a-CT and for the molecular
modelling and molecular dynamics simula-
tions. The succinimidyl functionality serves
as leaving group during the glycosylation
reaction.
R. J. Sola
´
and K. Griebenow Structural dynamics and serine protease catalysis
FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS 5305
for the dextran modified protein could not be construc-
ted due to the technical limitations involved in model-
ling linear polymeric molecules of such large size
(> 300 A
˚
). While molecular modelling and MD simula-
tions have previously been employed with great success
to provide a deeper mechanistic understanding towards
the roles of glycans on glycoprotein and glycocon-
jugates structure, stability, dynamics, and function

results, other noncovalent interactions such as internal
hydrogen bonds could also contribute to the increase in
these parameters. Analysis of the changes in internal
hydrogen bond composition for the protein-glycan con-
jugates indicates that for all of the conjugates there was
also an increase in these internal hydrogen bonds
formed due to glycosylation (Table S2). However, they
are too small to sustain the observed changes in the
bond and angle parameters. These are most probably
increased due to the increased VDW interactions. The
changes in some of these parameters (e.g. reduced
dihedral and increased VDW energies) also suggest a
more rigid and compact protein structure for the glyco-
conjugates. This increase in rigidity due to glycosylation
can be also be appreciated from the decrease in the cal-
culated Debye–Waller temperature B-factors (Table 3,
[60]). This reduction in dynamics due to chemical glyco-
sylation does not appear to be caused by the modified
lysine residue charges as it has been well established that
natural glycosylation also reduces substantially the
dynamics of natural glycoproteins where the modifica-
tion occurs in noncharged residues [32–34]. However,
future experiments will be performed to investigate this.
The observed changes in the coulombic energy
parameter also highlight the large contribution that
the internal electrostatics have towards decreasing the
total energy of the conjugates, which agrees with the
hypothesis of global electrostatics being relevant to
Table 1. Reactivity order based on the calculated electrostatic pot-
entials (EP) for the N

)1
).
Structural dynamics and serine protease catalysis R. J. Sola
´
and K. Griebenow
5306 FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS
protein stability [61]. The decrease in structural
dynamics due to glycosylation could also be attributed
to the decrease in the coulombic energy parameter
because electrostatics are also known to influence
protein dynamics [62]. This decrease in the internal
electrostatic energy of the protein as a result of glyco-
sylation and its consequences on protein dynamics and
stability seems to be in agreement with the notion that
glycosylation perturbs the protein’s surrounding solva-
tion-shell [36]. This could lead to solvent dielectric
Table 2. Kinetic and thermodynamic parameters derived from amide H ⁄ D exchange rates for a-CT and for the various lactose-a-CT and
dextran-a-CT conjugates at pH 7.1, 25 °C.
Glycoconjugate A
1
b
k
HX,1
(min
)1
) A
2
b
k
HX,2

b
A
i
are the fractions of amide protons in the i
th
population that exchange with a
rate constant k
HX,i
.
c
Gibbs free-energy of microscopic unfolding per mol of peptide hydrogen for the fast exchanging amide protons [48].
Fig. 4. Representative a-CT and Lac-a-CT glycoconjugates structures after equilibration of conformational energetics by MD simulations with
YASARA
Dynamics
. Coloring scheme: domain 1 (blue), domain 2 (red), catalytic triad (yellow), and mLac glycans (grey). Structures were ren-
dered with
PYMOL [92].
R. J. Sola
´
and K. Griebenow Structural dynamics and serine protease catalysis
FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS 5307
shielding [63] thereby transforming the protein bio-
physical properties from being solvent slaved to non-
slaved [64,65]. We have analyzed the effect that
glycosylation has on the protein-solvent hydrogen
bonds and the solvent accessible surface areas for the
protein portion of the conjugates to provide evidence
for this concept within our system. While the total
number of hydrogen bonds and solvent accessible area
increases for the conjugates with increased glycosyla-

This allowed us to then examine the effects of chemical
glycosylation on the kinetics of enzymatic catalysis
from both an experimental and theoretical perspective.
Changes in the kinetics of a-CT catalysis upon
chemical glycosylation
The catalytic behavior of a-CT after chemical glyco-
sylation was determined from the hydrolysis of
Table 3. Global energetic parameters and Debye–Waller temperature factors calculated for the protein portion of a-CT and the various lac-
tose-a-CT conjugate structures modeled and submitted MD simulations at pH 7.1, 25 °C. Energy values in McalÆmol
)1
.
Glycoconjugate
Energy
Global B-FactorBond Angle Dihedral Planar VdW Coulomb Total
a-CT 11.8 5.6 13.5 0.084 15.4 ) 136.5 ) 90.1 14.1
Lac
1
-a-CT 11.1 5.7 13.1 0.083 15.6 ) 133.3 ) 87.7 13.8
Lac
3
-a-CT 15.1 7.0 13.0 0.079 21.1 ) 168.2 ) 111.9 13.3
Lac
5
-a-CT 16.5 7.4 12.9 0.077 23.4 ) 182.5 ) 122.2 12.8
Lac
7
-a-CT 20.2 8.9 12.9 0.082 30.1 ) 224.1 ) 151.9 12.4
Lac
14
-a-CT 23.5 10.1 12.7 0.086 35.9 ) 259.4 ) 177.1 11.0

demonstrated that native-like activity and dynamics
could be restored at increased temperature regimes for
these glycoconjugates [36]. Evaluation of the glycocon-
jugates surface potential reveals that the decreased
kinetics are also not due to a perturbation of the
enzyme’s active site groove electrostatics due to lysine
charge modification (Fig. S3).
Because for the substrate used the k
cat
and K
M
parameters are a combination of the reaction’s individ-
ual rate constants (K
S
, k
2
, and k
3
) we determined these
by kinetic chemical dissection with a thio-benzyl (SBzl)
functionalized substrate as previously described by
Stein and coworkers [17]. This experiments revealed
that both the kinetics of enzyme acylation (k
2
) and
deacylation (k
3
) are reduced by chemical glycosylation,
also as a function of the glycan molar content of the
conjugates (Table 4). In contrast, the substrate binding

? k
3
) this generalized assumption is not
always accurate for all substrates as previously pointed
out by Hedstrom [16]. This can be appreciated experi-
mentally in the already mentioned work by Stein [17],
where they measured the changes in k
S
, k
2
, and k
3
as a
function of pH and temperature for three different
sized amide substrates (Suc-F-pNA, Suc-AF-pNA, and
Suc-AAPF-pNA). While for the two smaller substrates
k
2
is generally smaller than k
3
, for the larger substrate
that we use in our study k
2
is equivalent to k
3
.
Correlation between the changes in a-CT’s global
structural dynamics and enzyme kinetics
Next we performed a statistical correlation analysis
(Fig. 6) between the structural dynamic (DG

]. k
2
¼ k
3
k
cat
⁄ (k
3
– k
cat
). k
3
is equal to k
cat
for the hydrolysis of Suc-AAPF-SBzl.
Glycoconjugate k
cat
(s
)1
) K
M
(mM) K
S
(mM) k
2
(s
)1
) k
3
(s

R ¼ 0.9245, P < 0.0001; DG
HX,1
⁄ DG
„
k
3
:R¼ 0.9370,
P < 0.0001). Interestingly, the reaction’s activation
energy for both steps increases linearly with a decrease
in the structural dynamics of the enzyme (DG
„
k
2
¼
1.06DG
HX,1
+ 9.94; DG
„
k
3
¼ 1.12DG
HX,1
+ 9.65).
This linear relation can be rationalized if one considers
that the enzyme’s dynamical free-energy can be trans-
ferred to the reaction’s activation energy by influencing
the transition-state activation energy (DG
„
¼ DG
TS

dues and that the protein conformational fluctuations
responsible for H ⁄ D exchange are not necessarily in
the same timescales as the protein motions of catalysis.
Nevertheless, the slope values for the linear correla-
tions obtained here [which are close to unity (m % 1)]
clearly support the notion that the dynamical energy
of the enzyme is transferred directly into catalysis. The
correlations thus provide direct experimental evidence
indicating that both acylation and deacylation rates
are influenced by the changes in an enzyme’s structural
dynamics. This observed similar response for k
2
and k
3
to the changes in the enzyme’s structural dynamics
could be attributed to the fact that the enzyme
employs similar structural and chemical mechanisms
for proton transfer during the acylation and deacyla-
tion steps but just in a reverse order [16]. These results
provide support to the kinetic mechanism previously
presented by Kawai et al. [19,31] in which a substrate-
induced conformational change occurs during the for-
mation of the first tetrahedral intermediate and during
the breakdown of the second tetrahedral intermediate.
Nonetheless, an observation that becomes clearly
evident from our results is that to some degree the
Fig. 6. Statistical correlation analysis (ANOVA) between the Gibbs
free-energy of microscopic unfolding per mol of peptide hydrogen
for the fast exchanging amide protons (DG
HX,1

(kcalÆmol
)1
) DG
„
k
3
(kcalÆmol
)1
)
Lac-a-CT
0.0 ± 0.0 15.65 ± 0.01 15.69 ± 0.01
1.8 ± 0.7 15.81 ± 0.03 15.81 ± 0.01
2.5 ± 0.4 15.84 ± 0.03 15.86 ± 0.01
3.8 ± 0.4 15.86 ± 0.02 15.94 ± 0.01
5.2 ± 0.3 15.87 ± 0.03 15.93 ± 0.02
7.4 ± 0.3 16.00 ± 0.03 16.07 ± 0.04
Dex-a-CT
0.0 ± 0.0 15.65 ± 0.01 15.69 ± 0.01
1.4 ± 0.1 15.68 ± 0.01 15.72 ± 0.02
2.5 ± 0.3 15.74 ± 0.01 15.77 ± 0.02
4.2 ± 0.1 15.78 ± 0.02 15.81 ± 0.03
6.7 ± 0.4 16.00 ± 0.04 16.04 ± 0.05
7.6 ± 0.1 16.15 ± 0.02 16.19 ± 0.01
Structural dynamics and serine protease catalysis R. J. Sola
´
and K. Griebenow
5310 FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS
catalytically relevant dynamics of a-CT appear to be
an intrinsic structural feature of the protein. This
notion is indirectly supported by more detailed NMR

the enzyme.
Structural insights into the mechanochemical
nature of a-CT catalysis
A more detailed analysis of the influence of chemical
glycosylation on the dynamics of a-CT from the theor-
etical simulations was additionally performed to gain a
deeper perspective into the mechanism of coupling
between the structural dynamic and functional proper-
ties of the enzyme. Although decreases in the dynamics
of catalytically important regions (e.g. catalytic triad,
S1 binding site, and L1 specificity site) can certainly be
observed from the analysis of the MD trajectories
(Fig. S4 and Table S4), these changes are not necessar-
ily relevant to the changes in catalysis as the timescales
that are accessible to MD simulation techniques are
computationally limited so that catalytically important
phenomena which occur on larger time scales (e.g. col-
lective domain motions) are not accurately sampled.
The Gaussian network model (GNM) was developed
to provide a simple and computationally inexpensive
yet accurate description of residue mobilities within
the collective vibrational modes of proteins and supra-
molecular structures [73,74]. Results from this type of
calculation have been found to be in excellent agree-
ment with X-ray crystallographic B-factors, H ⁄ D
exchange free energies of amide protons, and NMR-
relaxation order parameters [75,76]. Due to this GNM
has been extensively used to describe the influence of
collective structural motions on the functional proper-
ties of proteins. Because these calculations are tradi-

both domains are correlated and thus move in the
Fig. 7. Catalytic steps influenced by the
enzyme’s intrinsic structural dynamics.
R. J. Sola
´
and K. Griebenow Structural dynamics and serine protease catalysis
FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS 5311
same direction (squared patterns in the upper left and
in the lower right areas of plot) within the collective
vibrational modes. Additionally, significant cross-
correlations are observed between catalytically relevant
residues (Cys42–Ser195, His57–Asp102, His57–Ser195,
Gly140–Ser195, Cys182–Ser214) present in similar and
in separate domains. The motion of these collective
vibrational modes can be better appreciated from a
movie generated with the normal mode analysis morph
server at Yale University ()
(Video S1) [80].
When the GNM analysis was applied to the a-CT
glycoconjugates (Fig. 9), the relative mobilities of both
interdomain connecting loops (85–105, 160–175) were
largely reduced with an increase in the relative mobility
of the interdomain hinge residues (120–150) and the
C-terminal a-helix (230–245) (Fig. 10). While most of
the glycosylation sites occur in these interdomain con-
necting loop regions (Fig. 9) we can also see a reduc-
tion in the dynamics of regions far away from the
glycosylation sites. This is most probably due to the
very well known fact that the network of hydrogen
bonds within the protein’s interior can relay informa-

i
)
2
]
1)2
) in the
slowest two collective vibrational modes
versus residue index (A) and interresidue
cross-correlation map (B) for a-CT calculated
by GNM. Coloring scheme on scale: positive
correlations (red), negative correlations
(blue).
Structural dynamics and serine protease catalysis R. J. Sola
´
and K. Griebenow
5312 FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS
Because the cross-correlation maps for the glycosylated
conjugates were not significantly different from those
of the nonglycosylated protein (results not shown), this
result implies that glycosylation can reduce the kinetics
of the protein’s collective domain motions without
altering the shape of the collective vibrational mode,
the protein structure, or even the interresidue connec-
tions. More specifically the decrease in mobility of the
Asp102 loop (85–105) seems to be largely dependent
on the degree of glycosylation. Interestingly, the mobil-
ity of the His57 loop (55–65) and the calcium binding
loop (70–80) which are adjacent to the Asp102 loop
were also reduced (but to a lesser extent) as a function
of the glycosylation degree. The decrease in mobility

ics and active-site chemistry through domain motions
[17,18,21,78].
Conclusions
In this article, we further demonstrate the value of
chemical glycosylation as a tool for studying the effects
of modulated structural dynamics on the protein bio-
physical properties. We have also been able to provide
some novel insights concerning the mechanism by
which glycans modulate the fundamental biophysical
protein properties. By applying this methodology to
the study of the kinetics of catalysis by the serine pro-
tease a-CT we were able to statistically correlate the
structurally dynamic behavior of the enzyme with its
kinetics of catalysis. From a mechanistic perspective
we have provided evidence that supports a catalytic
mechanism for a-CT in which both the enzyme acyla-
tion (k
2
) and deacylation (k
3
) steps are influenced by
the enzyme’s intrinsic thermally activated structural
dynamics. Additionally, through the use molecular
Fig. 9. Changes in relative mobility (D[(DR
i
)
2
]
1)2
) in the slowest two

D exchange measurements
Amide H ⁄ D exchange FTIR spectra were recorded on a
Nicolet NEXUS 470 infrared spectrophotometer equipped
with a thermally controlled sample cell (Spectra-Tech Inc.,
Shelton, CT) using CaF
2
windows and 25 lm Teflon spac-
ers (Buck Scientific, East Norwalk, CT). Kinetic experi-
ments were designed and performed similarly to those
reported previously [7,36,42–46,69].
Processing of H ⁄ D exchange data
H ⁄ D exchange spectra were processed for quantitative ana-
lysis in the form of hydrogen exchange (HX) decay plots (X
versus time) [36,42,44,45]. In these plots the fraction of un-
exchanged peptide hydrogen atoms (X) was determined as:
X ¼
wðtÞÀwð1Þ
wð0ÞÀwð1Þ
where w(t) is the ratio of the amide II (1550 cm
)1
) and
amide I (1637.5 cm
)1
) absorbencies corrected with the base-
line absorbance (1789 cm
)1
) at time t, w(0) is the amide
II ⁄ amide I ratio of the undeuterated proteins and w [8] is
the amide II ⁄ amide I ratio for the fully deuterated proteins.
WðtÞ¼

Þt
þ A
3
where A
1
, A
2
, and A
3
are the fractions of the fast, slow and
stable amide protons and k
HX,1
and k
HX,2
are the apparent
exchange rate constants for the fast and slow amide pro-
tons. Results were interpreted thermodynamically under the
EX
2
exchange mechanism (pH 7.1) [41,43] where the Gibbs
free-energy of microscopic unfolding per mol of peptide
hydrogens for the fast exchanging amide protons is based
on the chemical exchange rate constant (k
0
) and the meas-
ured rate constant (k
HX,1
) [42,44,45]:
DG
HX;1

kinetic parameters, initial velocities were determined for
seven initial substrate concentrations in the range between
0.01 and 0.5 mm k
cat
and K
M
parameters were determined
from Eadie–Hofstee plot analysis. Determination of the indi-
vidual rate constants (K
S
, k
2
, and k
3
) was performed by
chemical kinetic dissection experiments employing as sub-
strate Suc-Ala-Ala-Pro-Phe-SBzl as described previously by
Stein and coworkers [17]. For these reactions the thiol
product derived from thioester hydrolysis was detected by
a coupled assay with 5,5¢-dithiobis(2-nitrobenzic acid)
(DTNB) (100 lL; [DTNB]
o
¼ 2mm) at 412 nm (e
412
¼
13 mm
)1
Æcm
)1
) [17]. The rest of the kinetic method was the

tion model [88]. Lysine reactivity order (nucleophilicity) for
the in silico creation of the various glycoconjugates was
determined by calculating the electrostatic potentials (EP)
for the N
e
of the lysine residues of a-CT at pH 9 were the
chemical glycosylation reaction takes places using yasara’s
cell neutralization and pKa prediction module (Table 1)
[89]. The parametrized SS-mLac molecule was then coupled
in silico to the various lysine residues of the MD optimized
protein structure using this reactivity order, yielding the
glycoconjugates: Lac1-a-CT, Lac3-a-CT, Lac5-a-CT, Lac7-
a-CT, Lac14-a-CT. The novel amide bonds of the resulting
glycoconjugates were also QM parametrized with yasara’s
AutoSMILES protocol.
The protein models were subjected to yasara’s MD pro-
tocol using the AMBER99 force field with the GLYCAM
force field parameters for carbohydrates [54,86]. Simulation
temperature was 25 ° C, density 0.997, and pH 7.1. Van der
Waals pairs cutoff distance was 7.86 A
˚
and particle mesh
Ewald (PME) long range electrostatics were employed [90].
Multiple timesteps of 1.5 fs for intramolecular and 3 fs for
intermolecular forces and periodic cell boundaries with a
simulation cell 20 A
˚
larger than the protein along each axis
were used. Filling of the simulation cell with water, predic-
tion of charged residues pKa’s, placement of counter ions,

)] within a cutoff distance (r
c
¼ 6.0 A
˚
). The dynamics of
the resulting network are then defined by the Ni · Nj Kir-
chhoff connectivity matrix of interresidue contacts (G)
where the off diagonal elements of G are defined as: G
ij
¼
)1 if the distance between residues i and j (R
ij
) is shorter
than r
c
, meaning that they interact and G
ij
¼ 0ifR
ij
is lar-
ger than r
c
and the residues do not interact. The statistical
thermodynamics of the network are then described by its
potential V ¼ (c ⁄ 2)(DR)G(DR)
T
; where DR is a vector, with
DR
i
representing the displacement of the i

ij
¼hDR
i
Á DR
j
i=½hDR
2
i
iÁhDR
2
j
i
1=2
¼½C
À1

ij
=ð½C
À1

ii
½C
À1

ij
Þ
1=2
so that positive values describe residue movement in the
same direction while negative values describe movement in
opposite directions. From this analysis the dynamics of the

Tc
À1
Þk
k
½u
k

ii
Statistical analysis
All statistical analyses were performed by one-way analysis
of variance (anova) with a P-value of < 0.05 considered
significant using sigmaplot 8.0 (SPSS UKD Ltd, Woking,
UK) statistical analysis module.
Acknowledgements
We thank Dr Elmar Krieger at YASARA Biosciences
for the helpful discussions on the modelling studies.
This work was supported by Grant P20 RR16439 from
the National Institutes of Health (NIH-COBRE II) to
KG. RJS was supported by fellowships from the
R. J. Sola
´
and K. Griebenow Structural dynamics and serine protease catalysis
FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS 5315
Alfred P. Sloan Foundation and University of Puerto
Rico Deanship of Graduate Studies and Research
(UPR-DEGI).
References
1 Bruice TC & Benkovic SJ (2000) Chemical basis for
enzyme catalysis. Biochemistry 39, 6267–6274.
2 Somogyi B, Welch GR & Damjanovich S (1984) The

Kern D (2005) Intrinsic dynamics of an enzyme under-
lies catalysis. Nature 438, 117–121.
12 Benkovic SJ & Hammes-Schiffer S (2003) A perspective
on enzyme catalysis. Science 301, 1196–1202.
13 Garcia-Viloca M, Gao J, Karplus M & Truhlar DG
(2004) How enzymes work: Analysis by modern rate
theory and computer simulations. Science 303, 186–195.
14 Hammes-Schiffer S (2002) Impact of enzyme motion on
activity. Biochemistry 41, 13335–13343.
15 Antoniou D, Caratzoulas S, Kalyanaraman C, Mincer
JS & Schwartz SD (2002) Barrier passage and protein
dynamics in enzymatically catalyzed reactions. Eur J
Biochem 269, 3103–3112.
16 Hedstrom L (2002) Serine protease mechanism and
specificity. Chem Rev 102, 4501–4524.
17 Case A & Stein RL (2003) Mechanistic origins of the
substrate selectivity of serine proteases. Biochemistry 42,
3335–3348.
18 Hengge AC & Stein RL (2004) Role of protein confor-
mational mobility in enzyme catalysis: acylation of
alpha-chymotrypsin by specific peptide substrates.
Biochemistry 43, 742–747.
19 Kawai Y, Matsuo T & Ohno A (2000) Kinetic study
on conformational effect in hydrolysis of
p-nitroanilides
catalyzed by a-chymotrypsin. J Chem Soc Perkin Trans
2, 887–891.
20 Gerstein M, Lesk AM & Chothia C (1994) Structural
mechanisms for domain movements in proteins.
Biochemistry 33, 6739–6749.

sis in the Catalytic Triad Residue of Serine Proteases:
Correlation between Structural Rearrangement and
Chemical Shifts in the Acylation Process. Biochemistry
45, 5413–5420.
30 Cassidy CS, Lin J & Frey PA (1997) A new concept for
the mechanism of action of chymotrypsin: the role of the
low barrier hydrogen bond. Biochemistry 36, 4576–4584.
31 Kawai Y, Matsuo T & Ohno A (1998) Conformational
effect (induced-fit) on catalytic activity of alpha-chymo-
trypsin. Bull Chem Soc Jpn 71, 2187–2196.
Structural dynamics and serine protease catalysis R. J. Sola
´
and K. Griebenow
5316 FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS
32 Joao HC, Scragg IG & Dwek RA (1992) Effects of gly-
cosylation on protein conformation and amide proton
exchange rates in RNase B. FEBS Lett 307, 343–346.
33 Joao HC & Dwek RA (1993) Effects of glycosylation
on protein structure and dynamics in ribonuclease B
and some of its individual glycoforms. Eur J Biochem
218, 239–244.
34 Rudd PM, Joao HC, Coghill E, Fiten P, Saunders MR,
Opdenakker G & Dwek RA (1994) Glycoforms modify
the dynamic stability and functional activity of an
enzyme. Biochemistry 33, 17–22.
35 Kohen A, Jonsson T & Klinman JP (1997) Effects of
protein glycosylation on catalysis: changes in hydrogen
tunneling and enthalpy of activation in the glucose
oxidase reaction. Biochemistry 36, 2603–2611.
36 Sola

ium exchange. Biochemistry 43, 3844–3852.
43 Hvidt A & Nielsen SO (1966) Hydrogen exchange in
proteins. Adv Protein Chem 21, 287–386.
44 Li J, Cheng X & Lee JC (2002) Structure and dynamics
of the modular halves of Escherichia coli cyclic AMP
receptor protein. Biochemistry 41, 14771–14778.
45 Tatulian SA, Cortes DM & Perozo E (1998) Structural
dynamics of the Streptomyces lividans K
+
channel
(SKC1): secondary structure characterization from
FTIR spectroscopy. FEBS Lett 423, 205–212.
46 Zavodszky P, Johansen JT & Hvidt A (1975) Hydro-
gen-exchange study of the conformational stability of
human carbonic-anhydrase B and its metallocomplexes.
Eur J Biochem 56, 67–72.
47 Pace CN (1990) Conformational stability of globular
proteins. Trends Biochem Sci 15, 14–17.
48 Privalov PL & Tsalkova TN (1979) Micro- and macro-
stabilities of globular proteins. Nature 280, 693–696.
49 Bohne A & von der Lieth CW (2002) Glycosylation of
proteins: a computer based method for the rapid
exploration of conformational space of N-glycans.
Pac Symp Biocomput 285–296.
50 Bosques CJ, Tschampel SM, Woods RJ & Imperiali B
(2004) Effects of glycosylation on peptide conformation:
a synergistic experimental and computational study.
J Am Chem Soc 126, 8421–8425.
51 Karplus M & Kuriyan J (2005) Molecular dynamics
and protein function. Proc Natl Acad Sci USA 102,

45–53.
60 Kundu S, Melton JS, Sorensen DC & Phillips GN Jr
(2002) Dynamics of proteins in crystals: comparison of
experiment with simple models. Biophys J 83, 723–732.
61 Strickler SS, Gribenko AV, Gribenko AV, Keiffer TR,
Tomlinson J, Reihle T, Loladze VV & Makhatadze GI
(2006) Protein stability and surface electrostatics: a
charged relationship. Biochemistry 45, 2761–2766.
R. J. Sola
´
and K. Griebenow Structural dynamics and serine protease catalysis
FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS 5317
62 Simonson T (2003) Electrostatics and dynamics of pro-
teins. Reports on Progress in Physics 66, 737–787.
63 Affleck R, Haynes CA & Clark DS (1992) Solvent
dielectric effects on protein dynamics. Proc Natl Acad
Sci USA 89, 5167–5170.
64 Fenimore PW, Frauenfelder H, McMahon BH & Parak
FG (2002) Slaving: solvent fluctuations dominate pro-
tein dynamics and functions. Proc Natl Acad Sci USA
99, 16047–16051.
65 Fenimore PW, Frauenfelder H, McMahon BH &
Young RD (2004) Bulk-solvent and hydration-shell fluc-
tuations, similar to alpha- and beta-fluctuations in
glasses, control protein motions and functions. Proc
Natl Acad Sci USA 101, 14408–14413.
66 Haouz A, Glandieres JM & Alpert B (2001) Involve-
ment of protein dynamics in enzyme stability. The case
of glucose oxidase. FEBS Lett 506, 216–220.
67 Tsai AM, Udovic TJ & Neumann DA (2001) The

of protein dynamics: comparison of theoretical results
for hen lysozyme with X-ray diffraction and NMR
relaxation data. Proteins 37, 654–667.
76 Bahar I, Wallqvist A, Covell DG & Jernigan RL (1998)
Correlation between native-state hydrogen exchange
and cooperative residue fluctuations from a simple
model. Biochemistry 37, 1067–1075.
77 Hilser VJ, Garcia-Moreno EB, Oas TG, Kapp G &
Whitten ST (2006) A Statistical Thermodynamic
Model of the Protein Ensemble. Chem Rev 106, 1545–
1558.
78 Yang LW & Bahar I (2005) Coupling between catalytic
site and collective dynamics: a requirement for
mechanochemical activity of enzymes. Structure 13,
893–904.
79 Bahar I, Atilgan AR, Demirel MC & Erman B (1998)
Vibrational dynamics of folded proteins: Significance of
slow and fast motions in relation to function and stabi-
lity. Phys Rev Lett 80, 2733–2736.
80 Flores S, Echols N, Milburn D, Hespenheide B, Keating
K, Lu J, Wells S, Yu EZ, Thorpe M & Gerstein M
(2006) The Database of Macromolecular Motions: new
features added at the decade mark. Nucleic Acids Res
34, D296–D301.
81 Gohlke H, Kuhn LA & Case DA (2004) Change in
protein flexibility upon complex formation: analysis of
Ras-Raf using molecular dynamics and a molecular
framework approach. Proteins 56, 322–337.
82 Adebodun F & Jordan F (1989) Multinuclear magnetic
resonance studies on the calcium (II) binding site in

& Pedersen LG (1995) A Smooth Particle Mesh Ewald
Method. J Chem Phys 103, 8577–8593.
91 Chennubhotla C, Rader AJ, Yang LW & Bahar I
(2005) Elastic network models for understanding
Structural dynamics and serine protease catalysis R. J. Sola
´
and K. Griebenow
5318 FEBS Journal 273 (2006) 5303–5319 ª 2006 The Authors Journal compilation ª 2006 FEBS
biomolecular machinery: from enzymes to supramolecu-
lar assemblies. Phys Biol 2, S173–S180.
92 DeLano WL (2002) The PyMOL molecular graphics
system. .
Supplementary material
The following material is available for this article
online:
Fig. S1. H⁄ D exchange kinetic decay plots for a-CT,
the a-CT-lactose conjugates and the a-CT-dextran con-
jugates.
Fig. S2. Total energy equilibration during the produc-
tion phase of MD simulations for the various
glycoconjugates.
Fig. S3. Particle mesh Ewald electrostatic potential cal-
culated for a-CT and for the Lac14-CT conjugate
(positive potential: blue surface; negative potential: red
surface).
Fig. S4. Changes in root-mean squared displacements
(nrmsd) (A
˚
) versus residue index for the various
Lac-a-CT glycoconjugates structures with respect to