Stabilising the DNA-binding domain of p53 by rational design of its hydrophobic core.

Abstract

The core domain of the tumour suppressor p53 is of inherently low thermodynamic stability and also low kinetic stability, which leads to rapid irreversible denaturation. Some oncogenic mutations of p53 act by just making the core domain thermosensitive, and so it is the target of novel anti-cancer drugs that bind to and stabilise the protein. Increasing the stability of the unstable core domain has also been crucial for biophysical and structural studies, in which a stabilised quadruple mutant (QM) is currently used. We generated an even more stabilised hexamutant (HM) by making two additional substitutions, Y236F and T253I, to the QM. The residues are found in the more stable paralogs p63 and p73 and stabilise the wild-type p53 core domain. We solved the structure of the HM core domain by X-ray crystallography at 1.75 A resolution. It has minimal structural changes from QM that affect the packing of hydrophobic core residues of the beta-sandwich. The full-length HM was also fully functional in DNA binding. HM was more stable than QM at 37 degrees C. Anomalies in biophysics and spectroscopy in urea-mediated denaturation curves of HM implied the accumulation of a folding intermediate, which may be related to those detected in kinetic experiments. The two additional mutations over-stabilise an unfolding intermediate. These results should be taken into consideration in drug design strategies for increasing the stability of temperature-sensitive mutants of p53.

Full text

Page 1

Stabilising the DNA-binding domain of p53 by rational
design of its hydrophobic core
Kian Hoe Khoo†, Andreas C. Joerger†, Stefan
M.V. Freund and Alan R. Fersht1
MRC Centre for Protein Engineering, Hills Road, Cambridge CB2 0QH, UK
1To whom correspondence should be addressed.
E-mail: arf25@cam.ac.uk
†These authors contributed equally.
The core domain of the tumour suppressor p53 is of
inherently low thermodynamic stability and also low
kinetic stability, which leads to rapid irreversible dena-
turation. Some oncogenic mutations of p53 act by just
making the core domain thermosensitive, and so it is the
target of novel anti-cancer drugs that bind to and stabil-
ise the protein. Increasing the stability of the unstable
core domain has also been crucial for biophysical and
structural studies, in which a stabilised quadruple mutant
(QM) is currently used. We generated an even more
stabilised hexamutant (HM) by making two additional
substitutions, Y236F and T253I, to the QM. The residues
are found in the more stable paralogs p63 and p73 and
stabilise the wild-type p53 core domain. We solved the
structure of the HM core domain by X-ray crystallogra-
phy at 1.75 A˚resolution. It has minimal structural
changes from QM that affect the packing of hydrophobic
core residues of the b-sandwich. The full-length HM was
also fully functional in DNA binding. HM was more
stable than QM at 378 8 8 8 8C. Anomalies in biophysics and
spectroscopy in urea-mediated denaturation curves of
HM implied the accumulation of a folding intermediate,
which may be related to those detected in kinetic exper-
iments. The two additional mutations over-stabilise an
unfolding intermediate. These results should be taken
into consideration in drug design strategies for increasing
the stability of temperature-sensitive mutants of p53.
Keywords: drug design/folding intermediate/p53/protein
stability/structure
Introduction
The p53 tumour suppressor protein is a sequence-specific
transcription factor, which is involved in the regulation of a
complex gene regulatory network. It is at the hub of different
signalling pathways that determine the fate of the cell, which
include cell cycle arrest and apoptosis (Vogelstein et al.,
2000). p53 has a complex domain structure that consists of a
folded core and a tetramerisation domain with natively
unfolded N and C termini (Joerger and Fersht, 2008). The
DNA-binding core domain (residues 94–312) determines the
overall stability of full-length p53. The core domain has a
low melting temperature that is only slightly higher than
body temperature (Bullock et al., 1997). Some 50% of all
human cancers have dysfunctional p53 because of mutation.
Most of these mutations are localised in the core domain
(Olivier et al., 2002; Joerger and Fersht, 2007). They either
remove essential DNA-contact residues on the protein
surface (contact mutants) or induce loss of thermodynamic
stability, which may have concomitant structural changes
(structural mutants). Some structural mutants are highly
destabilised by .3 kcal/mol and are largely unfolded at
body temperature (Bullock et al., 2000). Wild-type core
domain is also kinetically unstable at 378C because it
denatures irreversibly and aggregates with a half-life of
?9 min (Friedler et al., 2003). Destabilised mutants such as
V143A and R249S unfold even faster. The irreversible dena-
turation may proceed via folding intermediates (Butler and
Loh, 2005).
p53 is tightly regulated in vivo and accumulates upon cel-
lular stress (Kubbutat et al., 1997). It is possible that the low
thermodynamic and kinetic stability of p53 have evolved so
that the protein has a low spontaneous half-life in vivo. The
thermodynamic stability of the p53 core domain correlates
with expression levels in Escherichia coli (Mayer et al.,
2007). This fact suggests that increasing the thermodynamic
stability of p53 would increase the accumulation of the
protein in mammalian cells, thus enhancing its potency in
inducing apoptosis. Accordingly, stabilised p53 would be
advantageous for use in gene therapy. Various efforts have
been undertaken to stabilise the core domain of p53, includ-
ing in vitro evolution (Matsumura and Ellington, 1999) and
semi-rational design (Nikolova et al., 1998). The latter has
resulted in a stabilised quadruple mutant M133L/V203A/
N239Y/N268D of the core domain (QM or T-p53C, residues
94–312). QM is 2.5 kcal/mol more stable than wild-type p53
core domain, while retaining the overall structure of the wild-
type protein (Nikolova et al., 1998; Joerger et al., 2004).
Wild-type p53 is so unstable that it is difficult to use in bio-
physical and structural studies. QM has been used as a scaf-
fold for elucidating the structural effects of numerous
destabilising oncogenic mutations by X-ray crystallography
(Joerger et al., 2005, 2006) and for measuring DNA binding
(Ang et al., 2006). Recently, it has also been used for deter-
mining the quaternary structure of full-length p53 in a multi-
pronged approach combining nuclear magnetic resonance,
electron microscopy and small-angle X-ray scattering (Tidow
et al., 2007).
In this study, we added two additional substitutions, Y236F
and T253I, to the QM and generated an even more stable hex-
amutant (HM). These two mutations were found initially from
solving the solution structure of p53 and discovering that
Tyr236 and Thr253 do not have properly paired hydrogen
bonds (Canadillas et al., 2006). The more stable p63 and p73
paralogs have Phe and Ile at positions equivalent to residues
236 and 253 in p53, respectively, and substitution of these resi-
dues into wild-type p53 stabilises it (Canadillas et al., 2006).
Here, we report the crystal structure of the HM and the effects
of the additional two mutations on its biophysical properties.
# 2009 The Author(s).
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License
(http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction
in any medium, provided the original work is properly cited.
421
Protein Engineering, Design & Selection vol. 22 no. 7 pp. 421–430, 2009
Published online June 10, 2009 doi:10.1093/protein/gzp018

Page 2

Results
Structure of HM p53 core domain
We solved the crystal structure of HM p53 core domain at
1.75 A˚resolution to elucidate the effects of the mutations on
the hydrophobic core of the protein (Table I). The crystals of
HM used for the structure solution were isomorphous to
those reported for QM and contained two molecules in the
asymmetric unit. The structure of the p53 core domain con-
sists of a central b-sandwich that provides the basic scaffold
for the DNA-binding surface (Fig. 1a). The latter is formed
by a loop-sheet-helix motif and two large loops (L2 and L3)
held together by a zinc ion, which is coordinated by three
cysteines (Cys176, Cys238 and Cys242) and a histidine
(His179). Overall, the structures of QM and HM are virtually
identical, apart from, in the immediate environment of the
mutation site in the hydrophobic core of the central
b-sandwich (Figs 1b and 2a). The Ca-atoms of equivalent
chains can be superimposed with a root-mean-square devi-
ation of 0.16 A˚(chain A) and 0.18 A˚(chain B). Instead of
the buried hydrogen bond of the Tyr236/Thr253 pair in the
wild-type and QM, Phe236 and Ile253 form hydrophobic
interactions, inducing only minor structural changes in their
immediate environment (Fig. 2a). The Cd-atom of Ile253
packs against the aromatic ring of Phe236 and the hydro-
phobic side chains of Ala161, Ile195, Ile255 and Val272.
Val272 is the only contacting residue with a notable shift of
main-chain atoms (0.4 A˚). Moreover, the side chain of
Val272 has flipped as a result of mutation (1108 rotation of x
for chain A) to form favourable packing interactions with the
Ile253 side chain. Interestingly, we also observed a flip of
the side chain of Leu254 even though it does not directly
interact with the mutated side chains but points away from
the hydrophobic core.
The p63 residues equivalent to Phe236 and Ile253 in p53
(Phe267 and Ile284) also pack against each other, as observed
in the recently solved solution structure of the p63 core
domain (PDB entry 2RMN). There are, however, small differ-
ences in the way these two residues pack in p63 because of
amino acid substitutions in the immediate environment of
these residues (e.g. Val272 in the p53 mutant corresponds to
Ala303 in p63) and shifts of the protein backbone at the end
of b-strand S8 where Phe267 is located (Fig. 2b).
Fluorescence anisotropy
We measured the DNA-binding properties of HM to probe
whether the two additional mutations have an effect on the
functional properties of the protein. The dissociation constant
of 11.3+3.0 nM for full-length HM and p21 consensus
sequence, measured by fluorescence anisotropy (Fig. 3), was
within the experimental error of that published for full-length
QM (12.0 nM) (Weinberg et al., 2005).
Table I. Data collection and refinement statistics for HM p53 core domain
Data collection
Space group
Cell [a, b, c (A˚)]
Molecules/ASU
Resolution (A˚)a
Unique reflections
Completeness (%)a
Multiplicitya
Rmerge(%)a,b
,I/sI.a
Wilson B value (A˚2)
Refinement
Number of atoms
Proteinc
Water
Zinc
Rcryst(%)d
Rfree(%)d
Root-mean-square deviation bonds (A˚)
Root-mean-square deviation angles (8)
Mean B value (A˚2)
Ramachandran plot statisticse
Most favoured/additional allowed (%)
Generously allowed/disallowed (%)
P212121
65.33, 71.08, 104.92
2
65.4–1.75 (1.84–1.75)
48 639
97.6 (92.4)
6.4 (5.7)
7.8 (27.7)
17.7 (5.5)
15.6
3102
455
2
17.0
19.9
0.008
1.2
17.6
93.4/6.6
0/0
aValues in parentheses are for the highest resolution shell.
bRmerge=P(Ih,i2 ,Ih.)/PIh,i.
5% of the amplitudes chosen at random and not used in the refinement.
eCalculated with PROCHECK (Laskowski et al., 1993).
cNumber includes alternative conformations.
dRcrystand Rfree=PkFobsj 2 jFcalck/PjFobsj, where Rfreewas calculated over
Fig. 1. Crystal structure of HM p53 core domain. (a) Ribbon diagram of the structure of HM. The mutated residues are shown as van der Waals models.
(b) Stereo view of the Catrace of HM (green) superimposed onto QM (PDB entry 1UOL; magenta) (Joerger et al., 2004) and wild-type p53 core domain
(PDB code 2OCJ; grey) (Wang et al., 2007), showing that the overall structures are virtually identical. Mutation sites in QM are marked with magenta spheres,
the sites of the two additional mutations Y236F and T253I in HM with green spheres.
K.H.Khoo et al.
422

Page 3

Differential scanning calorimetry
p53 denatures irreversibly and forms aggregates on approach-
ing its melting temperature, Tm. Differential scanning calori-
metry (DSC) does not record a true Tm because of this
irreversible denaturation. However, very high scanning rates
give a good approximation of the Tm(Boeckler et al., 2008).
The double-mutant Y236F/T253I had a very similar apparent
Tm to wild-type p53 (45.1+0.18C versus 45.3+0.28C)
(Fig. 4). Addition of the two mutations, Y236F/T253I, to
Fig. 2. Structural details of the hydrophobic core of HM. (a) Stereo view of the hydrophobic core of HM (green) superimposed onto QM (light pink). The
mutation sites (Y236F and T253I) and selected neighbours are shown as stick models. (b) Stereo view of the hydrophobic core of HM p53 core domain
(green) superimposed onto the p63 core domain (PDB entry 2RMN; yellow). Selected residues in the hydrophobic core are shown as stick model. The first
number in the labels refers to the position of a particular residue in p53, and the second number refers to the position in p63.
Fig. 3. DNA binding of full-length HM. Full-length HM binds to p21
consensus DNA with a Kd=11.3+3.0 nM, as measured by fluorescence
anisotropy.
Fig. 4. Thermal stability of p53 core domain variants. Melting curves for
wild-type p53 (black), double mutant (red), QM (blue) and HM (green), as
measured by DSC.
Stabilising p53 core domain by rational design
423

Page 4

QM (Tm= 51.0+0.28C) did not lead to a significant increase
in Tmfor HM (51.3+0.28C).
Thermodynamics of urea denaturation
The reversible unfolding, and hence thermodynamic stability,
of HM and QM can be measured at lower temperatures by
equilibrium ureadenaturation
Denaturation of p53 core domain by urea was monitored by
changes in the tryptophan fluorescence at 352 nm at different
temperatures (Fig. 5). At 108C, HM unfolded at a higher
urea concentration, [urea]50%, of 5.08 M urea than the
3.97 M urea found for QM. At higher temperatures (208C,
308C and 378C), the [urea]50%of HM was also higher than
that for QM (Table II). Measurements at temperatures higher
than previously described (Bullock et al., 1997) were poss-
ible because QM and HM are kinetically more stable than
the wild-type p53 (unpublished data for QM, G. Jaggi and
A.R.F.). For experiments carried out at 108C and 208C,
samples were equilibrated overnight. For experiments at
higher temperatures, shorter equilibration times were used to
prevent aggregation, as the rate of aggregation of protein
increases with temperature (see Materials and methods).
The m-value of p53, which is the rate of change of free
energyof denaturationwith
(=@DGD–N/@[urea]),isaless
[urea]50%, which is usually very reproducible (Serrano et al.,
1992). The m-values of HM (Table II) were markedly lower
than those of QM at 108C (e.g. 2.16 versus 2.96 kcal/mol/M
at a protein concentration of 0.5 mM). The differences in
values between QM and HM decreased with increasing
temperature, and at 378C the m-values were similar.
(Bullocket al.,1997).
denaturant
accurate
concentration
measure than
Fluorescence spectra in urea solutions
The m-value is a measure of the change in solvent-accessible
surface area upon denaturation. Low values of m can result
from either incomplete unfolding or the accumulation of an
intermediate upon denaturation. We searched for evidence of
an intermediate by examining the fluorescence spectra during
denaturation as a function of urea concentration. Changes in
both the tryptophan peak at 352 nm and the tyrosine peak at
305 nm could be observed by monitoring the full fluor-
escence spectra of QM and HM during unfolding at different
urea concentrationsat 108C
Tryptophan fluorescence is highest in the denatured state.
The maximum tyrosine signal is found for the native state.
An isoemission point, characteristic of a two-state transition,
was observed for QM but not for HM during unfolding at
108C, 208C and 308C (Fig. 6). This observation implies that
HM does not undergo a two-state transition during unfolding
with urea but that there must be one or more intermediates.
At 378C, however, there was an isoemission point for both
HM and QM implying that denaturation was two-state for
both, consistent with their m-values being similar (Table II).
To check for any effects due to aggregation, we performed
the experiments at different protein concentrations of 0.25
and 1.0 mM at 108C (Fig. 7). The [urea]50%of QM and HM
stayed largely invariant at the protein concentrations tested
(Table II). Some concentration dependence was observed for
the m-values of HM but not for QM (Table II). The m-value
of HM increased to 2.68 kcal/mol/M when 1 mM protein was
used but was invariant at lower concentrations. The isoemis-
sion point during denaturation at 1.0 mM protein implied
with280 nmexcitation.
two-state denaturation for HM (Fig. 7). At 0.25 mM, no isoe-
mission point was observed for HM, suggesting that there
was a significant population of an intermediate. There was an
isoemission point for QM at protein concentrations of 0.25
and 1.0 mM (Fig. 7).
15N HSQC
There was a good overlay of peaks in the15N HSQC spectra
of the denatured state of QM and HM at 6.7 M urea with the
exception of a few peaks, which were likely to result from
the two mutations. Most of the peaks had
were within random coil values (data not shown). Both QM
and HM were thus mainly unfolded at high concentrations of
denaturant.
1H signals that
Discussion
In this study, we report the design of a stabilised core
domain mutant of p53 (HM). The crystal structure of HM
shows that the two newly introduced mutations, Y236F and
T253I, improve hydrophobic core packing, which in turn
leads to higher stability of the core domain. Residues in
functional regions are not affected by the mutations. These
structural observations are consistent with fluorescence aniso-
tropy data, showing that the mutations have no effect on the
intrinsic DNA-binding properties of the full-length protein.
The two mutations mimic the hydrophobic core of the p53
paralogs p63 and p73, which are thermodynamically more
stable than p53 (Klein et al., 2001; Patel et al., 2008).
Phylogenetic studies suggest that p63 and p73 more closely
resemble the common ancestral protein, whereas p53
evolved later (Kaghad et al., 1997; Yang et al., 1998).
Hence, it appears that Tyr236 and Thr253 have been
selected by nature to allow p53 to have a more unstable
hydrophobic core.
It has been suggested that the intrinsic instability of the
p53 core domain might be important for the function of the
protein (Canadillas et al., 2006). Since p53 is at the hub of
cellular DNA-damage response pathways, its activity has to
be finely balanced and quickly switched on and off in
response to incoming signals. Having an inherently unstable
core domain may be one way of achieving this by allowing
for rapid cycling between folded and unfolded state [in
addition to modulation by post-translational modifications,
such as ubiquitination followed by nuclear export and protea-
somal degradation (Toledo and Wahl, 2006)]. As a negative
trade-off, p53 is more susceptible to destabilising mutations.
If wild-type p53 core domain had a higher intrinsic stability,
then many structural p53 cancer mutations that reduce the
thermodynamic stability of the protein but do not perturb the
architecture of functional interfaces would result in more
benign phenotypes at physiological conditions (Joerger et al.,
2006).
Unlike p53, p63 and p73 are not mutated in human
cancer. Both p63 and p73 seem to have unique roles in
development, whereas p53 plays a more important role in
tumour suppression, despite the similarity in their ability to
transcribe cell-cycle arrest and apoptotic genes (Moll and
Slade, 2004). More recent studies, however, show that
depending on the cellular context, p63 and p73 can act as a
tumour suppressor and maintain the integrity of the genome
(Pietsch et al., 2008; Rosenbluth and Pietenpol, 2008). p63,
K.H.Khoo et al.
424

Page 5

Fig. 5. Equilibrium denaturation of p53 core domain mutants. Urea-induced unfolding curves at different temperature of 108C, 208C, 308C and 378C are
shown for QM (a–d) and HM (e–h), represented by the change in relative fluorescence versus concentration of denaturant. The measurements were carried out
in 25 mM sodium phosphate, pH 7.2, 150 mM KCl, 5 mM DTT, with 0.5 mM protein.
Stabilising p53 core domain by rational design
425

Page 6

for example, plays a unique role in protecting the female
germ line during meiotic arrest (Suh et al., 2006). It would
be interesting to see whether the differences in thermodyn-
amic stability of the different members of the p53 family are
a contributing factor to their different functional properties in
the cellular environment.
Equilibrium urea denaturation at 108C has been used for
the determination of the free energy of unfolding, DGD–N, of
the p53 core domain and has provided much of the frame-
work for the classification of the effects of oncogenic
mutations on stability (Bullock et al., 2000). We monitored
the transition of intrinsic tryptophan and tyrosine fluor-
escence upon urea denaturation, in both QM and HM.
We observed lower m-values for HM when compared with
QM, which can either suggest a more significant residual
structure in the denatured state (Dill and Shortle, 1991) or a
decrease in cooperativity of unfolding, such as in the case of
staphylococcal nuclease (Carra and Privalov, 1996) or
barnase (Sanz and Fersht, 1993). A cooperative transition for
urea unfolding has been previously demonstrated for wild-
type p53 core domain by the presence of an isoemission
point (Bullock et al., 2000). The absence of an isoemission
point in the tryptophan and tyrosine spectral transition of
HM at lower temperatures suggests a lack of cooperativity in
unfolding. QM, on the other hand, appears to undergo a
two-state transition from 108C to 378C with no significant
population of intermediate species and with the presence of a
clear isoemission point. The m-value of QM is close to that
of wild-type p53 at 108C (2.96 versus 3.26 kcal/mol/M), as
reported previously (Bullock et al., 2000).
Moreover, we showed by HSQC that the denatured states of
HM and QM at high urea concentration of 6.65 M urea were
unfolded to a similar extent with little residual structure. The
HSQC results show that both HM and QM have similar
denatured states, whereas the native states of QM and HM are
shown to be similar by X-ray crystallography. Since the
m-value of a protein is affected by the difference in
solvent-accessible area between the native and the denatured
state (Dill and Shortle, 1991), it is possible that a significant
population of a folding-intermediate state in HM during unfold-
ing might explain its lower m-value. This population of an
intermediate state accumulates at lower temperatures of 108C,
208C and 308C, leading to a non-two-state transition (Fig. 6).
However, this population of HM intermediate may be lower at
378C, leading to an apparent two-state transition, as seen from
the presence of an isoemission point. Such a disappearance of
an intermediate population due to higher temperatures has been
previously described for barnase (Oliveberg et al., 1995). We
also observed that this intermediate species was less populated
at a higher concentration of HM protein (Fig. 7), as seen from
the presence of an isoemission point at 1.0 mM protein.
The stabilisation of an intermediate due to the two
mutations might also explain why HM does not have a sig-
nificantly higher apparent Tm than QM, as measured by
DSC, because of faster aggregation at higher temperatures. It
could be that this intermediate is the one that has been pre-
viously observed by kinetic experiments (Butler and Loh,
2005).
Previous calculations of the free energy of unfolding,
DGD–N, of p53 mutants have made use of an apparent
two-state assumption. However, the use of a two-state model
can lead to the erroneous conclusion that HM is less stable
than QM at lower temperatures due to the lower m-values of
HM, even though the urea midpoint of unfolding, [urea]50%,
is consistently higher for HM than QM. Three-state analysis
has been used previously for the calculations of the free
energy of unfolding, DGD–N, for low m-value mutants in the
case of staphylococcal nuclease. This analysis helped to
resolve contradictory results between the stability of mutants
(Carra and Privalov, 1996). We calculated the free energy of
unfolding of HM and QM at 378C, as both proteins display
isoemission points and can be approximated to undergoing a
two-state transition at this temperature. HM is found to be
1.9 kcal/mol more stable than QM at 378C, by taking the
average m-value of both HM and QM at this temperature.
In summary, we have further stabilised the native state of
p53 by 1.9 kcal/mol at 378C, when compared with the pre-
viously described QM, by the addition of two mutations,
Y236F and T253I, which mimic the hydrophobic core of the
paralogs p63 and p73. We propose that the two mutations
also lead to the stabilisation of a folding intermediate.
Understanding the formation of p53 intermediates will be
important for elucidating its folding pathway and the mech-
anism of misfolding of p53 due to oncogenic mutations.
Materials and methods
Molecular cloning and protein purification
p53 core domain (residues 94–312) was purified as
described previously using a pRSETA-derived plasmid
(Joerger et al., 2004). Full-length p53 (residues 1–393) was
cloned intoa pET24a-HLTV
N-terminal fusion of 6?His/lipoamyl domain/TEV protease
plasmid containing an
Table II. Equilibrium denaturation of p53 core domain mutants
Protein concentration (mM)a
Temperature (8C)QMHM
m (kcal/mol/M) [urea]50%(M)m (kcal/mol/M)[urea]50%(M)
0.5
0.5
0.5
0.25
0.5
1.0
37
30
20
10
10
10
22.48+0.23
22.75+0.09
22.88+0.07
22.92+0.10
22.96+0.18
23.10+0.01
2.56+0.08
3.33+0.04
3.81+0.03
4.05+0.01
3.97+0.02
4.12+0.01
22.18+0.19
22.24+0.20
22.27+0.13
22.10+0.25
22.16+0.10
22.68+0.17
3.37+0.11
4.04+0.07
4.55+0.14
4.99+0.01
5.08+0.02
5.06+0.16
aMeasurements were carried out in 25 mM sodium phosphate, pH 7.2, 150 mM KCl, 5 mM DTT. All experiments were done in triplicate except for the ones at
0.25 and 1.0 mM protein, which were done in duplicate.
K.H.Khoo et al.
426

Page 7

Fig. 6. Temperature dependence of fluorescence spectra upon urea denaturation of p53 core domain mutants. Raw fluorescence spectra at different
temperatures of 108C, 208C, 308C and 378C with urea unfolding are shown for QM (a–d) and HM (e–h). The measurements were done in 25 mM sodium
phosphate, pH 7.2, 150 mM KCl, 5 mM DTT, with 0.5 mM protein.
Stabilising p53 core domain by rational design
427

Page 8

cleavage site. Additional mutations were introduced using
the QuickChange site-directed mutagenesis kit (Stratagene).
Protein expression and purification of the full-length p53
mutants followed published protocols (Tidow et al., 2007).
Crystallisation and structure solution
Crystals of HM p53 core domain were grown at 218C using
the sitting drop vapour diffusion technique. Three microlitres
of protein solution (6 mg/ml in 20 mM phosphate buffer pH
7.2, 150 mM NaCl, 4 mM DTT) were mixed with 3 ml reser-
voir buffer (100 mM Hepes pH 7.2, 19% PEG 4000) above a
reservoir solution of 400 ml. The crystals obtained belonged
to space group P212121 and were isomorphous to those
reported for QM (Joerger et al., 2004). They were flash
frozen in mother liquor containing 20% glycerol as cryo pro-
tectant. Data collection was carried out at beamline I03 of
the Diamond Light Source. Data processing was performed
usingMOSFLM (Leslie,1992) and
(Collaborative Computational Project, 1994). The structure
was solved by molecular replacement with CNS (Bru ¨nger
et al., 1998) using the structure of QM (PDB entry 1UOL) as
a search model. Subsequently, the structure was refined by
iterative cycles of refinement with CNS (Bru ¨nger et al.,
1998) and PHENIX (Adams et al., 2002), and manual model
building with MAIN (Turk, 1992). Structure validation was
performed using PROCHECK (Laskowski et al., 1993) and
MOLPROBITY (Davis et al., 2007). Structural figures were
the CCP4 suite
created using PYMOL (www.pymol.org). Data collection
and refinement statistics are summarised in Table I.
Fluorescence anisotropy
Binding of HM to fluorescein-labelled 30mer dsDNA con-
taining the 30p21 response element (50-TAGAGGAAG
AAGACTGGGCATGTCTGGGCA) was measured by fluor-
escence anisotropy. Measurements were recorded on a
PerkinElmerLifeSciences
trometer equipped with a Hamilton Microlab titrator and con-
trolled by laboratory software. The excitation and emission
wavelengths used were 480 and 530 nm, respectively, and
the slit widths for excitation and emission were 15 and
20 nm. The photomultiplier voltage used was 950 V with an
integration time of 5 s for each measurement. The initial con-
centration of p21 DNA was 10 nM. The experiments were
performed at 108C in 25 mM imidazole, pH 7.2, 213.4 mM
NaCl, 5 mM DTT, with a total ionic strength of 225 mM.
LS55LuminescenceSpec-
Differential scanning calorimetry
DSC experiments
VP-Capillary DSC instrument (Microcal, Amherst, MA,
USA) with an active cell volume of 125 ml. Temperatures
from 108C to 858C were scanned at a rate of 2508C/h.
Protein samples were buffer-exchanged into a buffer of
25 mM sodium phosphate, pH 7.2, 150 mM NaCl, 5 mM
DTT. This buffer was also used for baseline scans. About
30 mM protein was used. A pressure of 2.5 bars (nitrogen)
wereperformed usinga Microcal
Fig. 7. Concentration dependence of fluorescence spectra upon urea denaturation of p53 core domain mutants. The fluorescence spectra at different
concentrations of 0.25 and 1.0 mM protein with urea unfolding are shown for QM (a and b) and HM (c and d). The measurements were done in 25 mM
sodium phosphate, pH 7.2, 150 mM KCl, 5 mM DTT, at 108C.
K.H.Khoo et al.
428

Page 9

was applied to the cell. The data were analysed with
ORIGIN software (Microcal). Experiments were carried out
in triplicate.
Urea denaturation
Samples for urea denaturation experiments were prepared
using a Hamilton Microlab dispenser from stock solutions of
urea, buffer and protein to contain 1 mM protein in 25 mM
sodium phosphate buffer, pH 7.2, 150 mM KCl, 5 mM DTT
and increasing concentrations of urea. Samples were equili-
brated overnight for measurements done at 108C and 208C.
Samples were allowed to equilibrate for 4 h for measure-
ments at 308C. For the experiments at 378C, the samples
were kept at 308C and allowed to equilibrate in the cuvette at
378C for 9 min before measurements were taken. The intrin-
sic fluorescence spectra of p53, excited at 280 nm, were
recorded in the range of 300–400 nm on a PerkinElmer Life
Sciences LS50B spectro-fluorometer equipped with a Waters
2700 sample manager and controlled by laboratory software.
The data were analysed as previously described (Bullock
et al., 2000).
Nuclear magnetic resonance
15N HSQC spectra were acquired for the denatured states of
QM and HM at 208C. About 100 mM15N labelled QM and
HM core domain protein in 25 mM sodium phosphate buffer,
pH 7.2, 150 mM KCl, 5 mM DTT and 6.65 M urea
were used.
Curve fitting
The data from urea denaturation for QM and HM were fitted
to Eq. (1), which assumes a two-state model in which the
fluorescence of the folded and unfolded states is dependent
on denaturant concentration:
F¼ðaNþbN½D?ÞþðaDþ bD½D?Þexpfm½D? ? ½D?50%Þ=RTg
1 þ expfmð½D? ? ½D?50%=RTg
where F is the fluorescence at the given concentration of
denaturant, aNand aDare the intercepts and bNand bDare
the slopes of the baselines at low (N) and high (D) denatur-
ant concentrations, respectively. [D] is the concentration of
denaturant; [D]50%, the concentration of denaturant at which
half of the protein is denatured; m, the slope of the transition;
R, the gas constant and T the temperature in K. The data
were fitted to this equation by non-linear least-square analy-
sis using the general curve-fit option of the Kaleidagraph
program (Abelbeck Software, Reading, PA, USA), which
gives the calculated values for individual experimental
measurements of m and [D]50%together with their standard
errors.
The free energy of denaturation of proteins in the presence
of denaturant DGD–N
is, to a first approximation, linearly
related to the concentration of denaturant [Eq. (2)]:
D
DGD
D?N¼ DGH2O
D?N? m½D?
Protein data bank accession number
Coordinates and structure factors of HM p53 core domain
have been deposited in the Protein Data Base (PDB entry
2WGX).
Acknowledgements
We thank Drs Chris Johnson, Robert Sade, and Dmitry Veprintsev for
helpful discussions and technical advice, and Caroline Blair for her help in
molecular biology and protein purification. We also thank the staff at the
Diamond Light Source (beamline I03) for help with data collection. This
publication reflects the authors’ views and not necessarily those of the EC.
The Community is not liable for any use that may be made of the
information.
Funding
K.H.K. is supported by the Singapore Agency for Science,
Technology and Research (A*STAR). This research was sup-
ported by Cancer Research, UK, the Medical Research
Council, and by EC FP6 funding. Funding to pay the Open
Access publication charges for this article was provided by
the Medical Research Council.
References
Adams,P.D., Grosse-Kunstleve,R.W., Hung,L.W., Ioerger,T.R., McCoy,A.J.,
Moriarty,N.W., Read,R.J.,Sacchettini,J.C.,
Terwilliger,T.C. (2002) Acta Crystallogr. D, 58, 1948–1954.
Ang,H.C., Joerger,A.C., Mayer,S. and Fersht,A.R. (2006) J. Biol. Chem.,
281, 21934–21941.
Boeckler,F.M., Joerger,A.C., Jaggi,G., Rutherford,T.J., Veprintsev,D.B. and
Fersht,A.R. (2008) Proc. Natl Acad. Sci. USA, 105, 10360–10365.
Bru ¨nger,A.T., et al. (1998) Acta Crystallogr. D, 54, 905–921.
Bullock,A.N., et al. (1997) Proc. Natl Acad. Sci. USA, 94, 14338–14342.
Bullock,A.N., Henckel,J. and Fersht,A.R. (2000) Oncogene, 19, 1245–1256.
Butler,J.S. and Loh,S.N. (2005) J. Mol. Biol., 350, 906–918.
Canadillas,J.M., Tidow,H., Freund,S.M., Rutherford,T.J., Ang,H.C. and
Fersht,A.R. (2006) Proc. Natl Acad. Sci. USA, 103, 2109–2114.
Carra,J.H. and Privalov,P.L. (1996) FASEB J., 10, 67–74.
Collaborative Computational Project, N (1994) Acta Crystallogr. D, 50,
760–763.
Davis,I.W., et al. (2007) Nucleic Acids Res., 35, W375–W383.
Dill,K.A. and Shortle,D. (1991) Annu. Rev. Biochem., 60, 795–825.
Friedler,A., Veprintsev,D.B., Hansson,L.O. and Fersht,A.R. (2003) J. Biol.
Chem., 278, 24108–24112.
Joerger,A.C. and Fersht,A.R. (2007) Oncogene, 26, 2226–2242.
Joerger,A.C. and Fersht,A.R. (2008) Annu. Rev. Biochem., 77, 557–582.
Joerger,A.C., Allen,M.D. and Fersht,A.R. (2004) J. Biol. Chem., 279,
1291–1296.
Joerger,A.C., Ang,H.C., Veprintsev,D.B., Blair,C.M. and Fersht,A.R. (2005)
J. Biol. Chem., 280, 16030–16037.
Joerger,A.C., Ang,H.C. and Fersht,A.R. (2006) Proc. Natl Acad. Sci. USA,
103, 15056–15061.
Kaghad,M., et al. (1997) Cell, 90, 809–819.
Klein,C., Georges,G., Kunkele,K.P., Huber,R., Engh,R.A. and Hansen,S.
(2001) J. Biol. Chem., 276, 37390–37401.
Kubbutat,M.H., Jones,S.N. and Vousden,K.H. (1997) Nature, 387, 299–303.
Laskowski,R.A., MacArthur,M.W., Moss,D.S. and Thornton,J.M. (1993)
J. Appl. Crystallogr., 26, 283–291.
Leslie,A.G.W. (1992) Joint CCP4 and ESF-EACMB Newsletter on Protein
Crystallography, Vol. 26. Daresbury Laboratory, Warrington, UK.
Matsumura,I. and Ellington,A.D. (1999) Protein Sci., 8, 731–740.
Mayer,S., Rudiger,S., Ang,H.C., Joerger,A.C. and Fersht,A.R. (2007) J. Mol.
Biol., 372, 268–276.
Moll,U.M. and Slade,N. (2004) Mol. Cancer Res., 2, 371–386.
Nikolova,P.V., Henckel,J., Lane,D.P. and Fersht,A.R. (1998) Proc. Natl
Acad. Sci. USA, 95, 14675–14680.
Oliveberg,M., Tan,Y.J. and Fersht,A.R. (1995) Proc. Natl Acad. Sci. USA,
92, 8926–8929.
Olivier,M., Eeles,R., Hollstein,M., Khan,M.A., Harris,C.C. and Hainaut,P.
(2002) Hum. Mutat., 19, 607–614.
Patel,S., Bui,T.T., Drake,A.F., Fraternali,F. and Nikolova,P.V. (2008)
Biochemistry, 47, 3235–3244.
Pietsch,E.C.,Sykes,S.M.,McMahon,S.B.
Oncogene, 27, 6507–6521.
Rosenbluth,J.M. and Pietenpol,J.A. (2008) Genes Dev., 22, 2591–2595.
Sanz,J.M. and Fersht,A.R. (1993) Biochemistry, 32, 13584–13592.
Serrano,L., Kellis,J.T., Jr, Cann,P., Matouschek,A. and Fersht,A.R. (1992)
J. Mol. Biol., 224, 783–804.
Sauter,N.K. and
and Murphy,M.E.(2008)
Stabilising p53 core domain by rational design
429

Page 10

Suh,E.K., Yang,A., Kettenbach,A., Bamberger,C., Michaelis,A.H., Zhu,Z.,
Elvin,J.A., Bronson,R.T., Crum,C.P. and McKeon,F. (2006) Nature, 444,
624–628.
Tidow,H.,Melero,R.,Mylonas,E.,
Carazo,J.M., Svergun,D.I., Valle,M and Fersht,A.R. (2007) Proc. Natl
Acad. Sci. USA, 104, 12324–12329.
Toledo,F. and Wahl,G.M. (2006) Nat. Rev. Cancer, 6, 909–923.
Turk,D. (1992) PhD thesis, Technische Universita ¨t Mu ¨nchen, Germany.
Vogelstein,B., Lane,D. and Levine,A.J. (2000) Nature, 408, 307–310.
Wang,Y., Rosengarth,A. and Luecke,H. (2007) Acta Crystallogr. D, 63,
276–281.
Weinberg,R.L., Veprintsev,D.B., Bycroft,M. and Fersht,A.R. (2005) J. Mol.
Biol., 348, 589–596.
Yang,A., Kaghad,M.,Wang,Y.,Gillett,E.,
Andrews,N.C., Caput,D. and McKeon,F. (1998) Mol. Cell, 2, 305–316.
Freund,S.M., Grossmann,J.G.,
Fleming,M.D.,Dotsch,V.,
Received May 5, 2009; revised May 5, 2009;
accepted May 6, 2009
Edited by Valerie Daggett
K.H.Khoo et al.
430