Tài liệu Báo cáo khóa học: Mutations in the hydrophobic core and in the protein–RNA interface affect the packing and stability of icosahedral viruses doc

Mutations in the hydrophobic core and in the protein–RNA interface
affect the packing and stability of icosahedral viruses
Sheila M. B. Lima
1
, David S. Peabody
2
, Jerson L. Silva
1
and Andre
´
a C. de Oliveira
1
1
Departamento de Bioquı
´
mica Me
´
dica, Instituto de Cie
ˆ
ncias Biome
´
dicas and Centro Nacional de Ressona
ˆ
ncia Magne
´
tica Nuclear de
Macromole
´
culas, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil;
2
Department of Molecular Genetics and

Microbiology and Cancer Research and Treatment Center, University of New Mexico School of Medicine, Albuquerque, NM, USA
The information required for successful assembly of an
icosahedral virus is encoded in the native conformation of
the capsid protein and in its interaction with the nucleic acid.
Here we investigated how the packing and stability of virus
capsids are sensitive to single amino acid substitutions in the
coat protein. Tryptophan fluorescence, bis-8-anilinonaph-
thalene-1-sulfonate fluorescence, CD and light scattering
were employed to measure urea- and pressure-induced
effects on MS2 bacteriophage and temperature sensitive
mutants. M88V and T45S particles were less stable than the
wild-type forms and completely dissociated at 3.0 kbar of
pressure. M88V and T45S mutants also had lower stability
in the presence of urea. We propose that the lower stability of
M88Vparticlesisrelatedtoanincreaseinthecavityof
the hydrophobic core. Bis-8-anilinonaphthalene-1-sulfonate
fluorescence increased for the pressure-dissociated mutants
but not for the urea-denatured samples, indicating that the
final products were different. To verify reassembly of the
particles, gel filtration chromatography and infectivity
assays were performed. The phage titer was reduced dra-
matically when particles were treated with a high concen-
tration of urea. In contrast, the phage titer recovered after
high-pressure treatment. Thus, after pressure-induced dis-
sociation of the virus, information for correct reassembly
was preserved. In contrast to M88V and T45S, the D11N
mutant virus particle was more stable than the wild-type
virus, in spite of it also possessing a temperature sensitive
growth phenotype. Overall, our data show how point sub-
stitutions in the capsid protein, which affect either the

packing or the interaction at the protein–RNA interface,
result in changes in virus stability.
Keywords: hydrostatic pressure; MS2 bacteriophage; tem-
perature-sensitive mutants; urea; fluorescence.
The protein shells of viruses generally have several key
functions, including shielding of the nucleic acid, particle
maturation and conferring the ability to penetrate the host
cell and undergo disassembly. The coat proteins are usually
arranged in a shell with an icosahedral shape [1]. The
information required for successful assembly of a virus
particle is encoded in the native conformation of a capsid
protein subunit. Structural and thermodynamic approaches
have been employed to identify the general rules that govern
virus assembly [2–7].
The MS2 bacteriophage is an RNA virus of the
family Leviviridae, a group of single-stranded RNA
bacteriophages that infect F+ Escherichia coli cells. The
icosahedral shell of the MS2 virus particle has a
diameter of 260 A
˚
and is made up of 180 copies of
the coat protein subunit (M
r
13.7 · 10
3
)inaT¼3
surface lattice. Each virion also contains one copy of
the maturase protein, which is responsible for attach-
ment of the phage to E. coli F-pili. The coat protein
has two functions in the viral life cycle. First, it acts as

a translational repressor of the replicase gene. A coat
protein dimer binds specifically to an RNA stem–loop
structure (known as the translational operator) and
prevents initiation of replicase translation [8–10]. Second,
coat protein serves as the major virus structural protein,
forming the shell in which the RNA genome is
contained [11,12].
The tertiary structure and topology of the MS2 coat
protein is different from those of other simple icosahedral
viruses [13]. The main chain of the protein subunit folds
into a five stranded antiparallel b-sheet (strands bC–bG)
facing the interior of the phage particle, with an
N-terminal hairpin (strands bAandbB) and two a-helices
(aAandaB) shielding most of the upper surface of the
b-sheet from the environment. Upon dimerization, exten-
sive contacts are formed between the subunits so that
the b-sheet becomes extended to form a continuous
10-stranded sheet. The two polypeptide chains are so
intimately intertwined that it seems clear that the dimer
must be the basic unit of coat protein folding; each
subunit depends on the other for acquisition of its native
fold. The 3D structure of the MS2 bacteriophage has
been determined at 2.8 A
˚
(Fig. 1A) [14]. Even so, the
mechanism of assembly and nucleic acid recognition are
still far from completely understood.
Correspondence to A. C. de Oliveira, Avenida Bauhinia,
400 - CCS/ICB/Bl E, sl. 08, Cidade Universita
´

ria,
CEP 21941-590, Rio de Janeiro, RJ, Brazil.
Fax: + 55 21 2270 8647, Tel.: + 55 21 2562 6756,
E-mail:
Abbreviations: bis-ANS, bis-8-anilinonaphthalene-1-sulfonate; LB,
Luria–Bertani; p.f.u., plaque-forming units; ts, temperature sensitive;
WT, wild-type.
(Received 26 September 2003, accepted 7 November 2003)
Eur. J. Biochem. 271, 135–145 (2004) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03911.x
High pressure is an efficient tool for studies on the
folding of proteins [15–18] and on the assembly of
supramolecular structures, such as viruses [6,7,19–23]. In
general, it has been found that individual capsid proteins
(monomers or dimers) are generally much less stable to
pressure than the assembled icosahedral particles [21,22].
The isolated capsid and the assembly intermediates assume
different partially folded states in the assembly pathway
[7]. Hydrostatic pressure permits controlled perturbation of
the subunit interactions and is a powerful tool for using to
study cavities in proteins [24,25]. The coat protein of
bacteriophage MS2 contains two tryptophan residues, thus
permitting the use of intrinsic fluorescence as a probe of
structural changes. Here, we study the stability against
pressure and urea of the MS2 bacteriophage and three
temperature sensitive (ts) mutants (Fig. 1B). Urea and
hydrostatic pressure are utilized to promote capsid disso-
ciation and denaturation, where the conformational chan-
ges are analyzed by fluorescence spectra, light scattering,
CD, HPLC, infectivity assays and the bis-8-anilinonaph-
thalene-1-sulfonate (bis-ANS) binding assay. We find that

pressure promotes dissociation of the wild-type (WT)
bacteriophage and ts mutants. Two mutations that appear
to lead to the formation of cavities, one in the hydrophobic
core (M88V) and the other in the protein–RNA interaction
(T45S), decrease the stability of the capsid. Our findings
illuminate the role of packing in the icosahedral lattice of
the virus capsid.
Experimental procedures
Chemicals
All reagents were of analytical grade. Distilled water was
filtered and deionized through a Millipore water purification
system. The experiments were performed at 20 °Cin
standard buffer: 50 m
M
Tris, 150 m
M
NaCl, pH 7.5.
MS2 bacteriophage and mutant samples
After growth of E. coli C3000 in Luria–Bertani (LB)
medium to an attenuance (D), at 600 nm, of 1.2, the culture
was infected with MS2 and, after a further 5 h of
incubation, was treated with lysozyme [8]. The samples
were processed by pelleting the bacterial debris by centrif-
ugation (8000 r.p.m. for 10 min; RPR 9.2 rotor; Beckman)
at 4 °C. The supernatant was precipitated with ammonium
sulfate (330 gÆL
)1
). The phages were precipitated by
centrifugation (10000 r.p.m. for 45 min; RPR 12.2 rotor;
Beckman) at 4 °C. The precipitate was dissolved in standard

buffer and purified by high-speed centrifugation
(35 000 r.p.m. for 14 h; SW41 rotor; Beckman) in a sucrose
gradient (10–50%). The sample concentration utilized in all
the experiments was 50 lgÆmL
)1
, except for CD experi-
ments (where the sample concentration was 100 lgÆmL
)1
).
Virus concentrations were determined by the method of
Bradford [26] using lysozyme as a standard. They were
confirmed by measuring the absorbance at 280 nm.
Spectroscopic measurements under pressure
Two important parts form the high pressure system: the
pressure generator and the high-pressure cell [27]. Fluores-
cence spectra and light scattering measurements were
recorded on an ISSK2 spectrofluorometer (ISS Inc.,
Champaign, IL, USA). Fluorescence spectra were quanti-
fied by evaluating the spectral center of mass, <m>, as
follows:
Fig. 1. Structure of the whole capsid of bacteriophage MS2 and of the
coat protein dimer bound to RNA. (A) The MS2 bacteriophage capsid is
colored according to the asymmetric units. (B) The coat protein of
MS2 bacteriophage bound to RNA, showing the location of the amino
acids (in space fill display) substituted in the temperature sensitive (ts)
mutants. The two polypeptide chains of the dimer are shown as blue
and green ribbons and the RNA molecule is shown in red (space fill).
Met88 is represented in brown (at the dimer surface), Thr45 in yellow
(interacting with RNA) and Asp11 in red, which appears to form a salt
bridge to Lys113 (cyan) on the alpha-helix of the adjacent subunit of

the dimer (PDB file: 2MS2).
136 S. M. B. Lima et al. (Eur. J. Biochem. 271) Ó FEBS 2003
<m> ¼ Rmi:F i=RFi ð1Þ
where Fi stands for the fluorescence emitted at wavenumber
mi and the summation is carried out over the range of
appreciable values of F. The values of center of spectral
mass of tryptophan were converted into degree of denatur-
ation/dissociation (a
p
), according to the following equation:
a
p
¼ðm
p
À m
i
Þ=ðm
f
À m
i
Þð2Þ
where m
p
is the value at each pressure, m
i
is the value at 1.0
bar, and m
f
is the value at 3.4 kbar. The volume change DV
can be calculated from the following thermodynamic

relation [15,16]:
ln½a
n
p
=ð1 À a
p
Þ ¼ pDV=RT þ ln½k
atm
=n
n
C
ðnÀ1Þ
ð3Þ
where k
atm
is the denaturation/dissociation constant at
atmospheric pressure, p corresponds to a given pressure, R
is the gas constant, T is the absolute temperature, n is the
number of subunits, and C is the protein concentration.
Each experiment was performed at least three times with
different protein preparations.
Light scattering
Light scattering measurements were made in an ISSK2
spectrofluorometer. Scattered light was collected at an angle
of 90° of the incident light. The samples were excited on
320 nm and collected in the same wavelength. This wave-
length was chosen because protein and RNA do not absorb
at 320 nm.
Chemical denaturation
The samples were incubated with increasing concentrations

of urea (1–9
M
) and allowed to equilibrate for 30 min prior
to making measurements. The measurements were made in
the presence and absence of urea. Each experiment was
performed at least three times with different protein
preparations.
Size-exclusion chromatography
HPLC was carried out using a prepackaged SynChropak
GPC500 column (250 · 4.6 mm inner diameter; SynChro-
paK Inc., Linden, IN, USA). The system was equilibrated
in 50 m
M
Tris, 0.2
M
sodium acetate buffer containing
0.5 gÆL
)1
sodium azide (pH 7.0). A flow rate of 0.3 mLÆ
min
)1
was utilized. Sample elution was monitored by the
fluorescence at 330 nm (excitation at 280 nm) and the
absorbance at 260 nm. The equipment used was a Shim-
adzu model.
CD
Conformational changes in MS2 bacteriophage and ts
mutants treated with urea were analyzed. The MS2
bacteriophage and ts mutant samples were diluted to a
final concentration of 100 lgÆmL

)1
and the spectra were
obtained in 10 m
M
Tris, 30 m
M
NaCl (pH 7.5) buffer using
a 0.1 cm pathlength quartz cuvette. The spectropolarimeter
used was a Jasco J-715 1505 model.
Infectivity assays
An overnight culture of E. coli was diluted 1 : 20 (v/v) in
LB medium and cultured at 37 °C for 2 h in a rotary shaker.
Several phage dilutions, made in a standard buffer, were
plated in LB semisolid medium containing E. coli.The
plates were incubated overnight at 37 °C, after which the
MS2 and ts mutants were diluted and titered by quantifi-
cation of plaques resulting from phage-induced bacterial
lysis. The results are expressed as p.f.u. (plaque-forming
units) per mL.
Isolation of ts coat mutants
The MS2 coat protein gene was randomly mutagenized by
error-prone PCR [28] and introduced into pMS27 [29], a
plasmid from which infectious MS2 genomic RNA can be
produced by transcription from the T7 promoter. Transfec-
tion into strain CSH41(pAR1219) [30], which produces T7
RNA polymerase, and plating at 32 °C led to the production
of plaques, % 200 of which were picked to lawns of CSH41 on
duplicate plates. One plate was incubated at 32 °Candthe
other at 40 °C. After identification of mutants exhibiting a
growth defect at 40 °C, virus stocks were produced by

growth in LB medium at 32 °C and the viruses were purified,
by sedimentation to equilibrium, in CsCl density gradients.
Further characterization showed that the three mutants
analysed in this study exhibited modest reductions in plating
efficiencies (four to 20-fold reductions in plaque number) at
elevated temperature, but in each case the plaques produced
were dramatically smaller than at the permissive tempera-
ture, whereas WT plaque size was unaffected. The coat genes
of ts mutants were recovered by RT–PCR and cloned in
pCT119 [31] and their amino acid substitutions were
determined by DNA sequence analysis.
Results
Chemical stability of ts MS2 coat mutants
The coat protein of MS2 is the major structural protein of the
virus and acts as a translational repressor that inhibits
synthesis of the viral replicase late during infection (Fig. 1).
We isolated ts mutants of MS2, as described above in the
Experimental procedures. Each mutant particle exhibited a ts
growth phenotype, producing fewer and dramatically smal-
ler plaques at non-permissive temperatures. Further charac-
terization verified the ts character of the mutants. Each
exhibited a ts defect for repression of translation of a
replicase–b-galactosidase fusion protein from plasmid pRZ5
[31,32]. Furthermore, although WT coat protein was found
almost entirely in the soluble fraction of cell lysates at either
growth temperature, each of the ts mutant proteins was
found predominantly in the insoluble fraction when cultured
at the non-permissive temperature, but exhibited WT
solubilities at the permissive temperature. More than 12 ts
coat mutants were produced with these characteristics (D. S.

Peabody, unpublished results). We describe here some addi-
tional properties of three: D11N, T45S and M88V (Fig. 1B).
We used light scattering and intrinsic tryptophan fluor-
escence to monitor whole particle disassembly and subunit
denaturation, respectively. Intrinsic fluorescence of the coat
Ó FEBS 2003 Cavities and stability in MS2 capsid mutants (Eur. J. Biochem. 271) 137
protein in the absence of urea is blue shifted because the
tryptophan residues are buried in the hydrophobic interior
of the protein. As the protein unfolds, the tryptophan
residues become more exposed to the solvent and their
fluorescence maxima shift towards the red (Fig. 2). Figure 2
shows that the WT and mutant forms have different
susceptibilities at an intermediate urea concentration
(4.5
M
), but all fully dissociate and denature at a high urea
concentration (9.0
M
).
Trp32 clearly resides within the hydrophobic core of the
protein. The other, Trp82, is only partially solvent-exposed
[33]. Its environment is determined primarily by interactions
within the dimer, not by interactions between dimers. Thus,
tryptophan fluorescence should predominantly monitor
dimer denaturation rather than capsid dissociation. Mean-
while, light scattering measurements are sensitive to the size
of the particle and can be used to monitor capsid
dissociation. The WT virus and each mutant were subjected
to increasing concentrations of urea (1–9
M

). Figure 3A
shows that the curves relating to the spectral center of mass
and light scattering are practically superimposable for WT
MS2, indicating that the subunit dissociation and the
denaturation processes are coupled. The T45S and M88V
mutants were both significantly less stable than the WT
form (Figs 2 and 3B,C) and denatured at rather lower urea
concentrations (midpoints at % 3
M
for the mutants vs.
% 4.5
M
for the WT form). The dissociation and denatur-
ation processes for T45S and M88V were also coupled, as
shown by the overlapping of the fluorescence and light
scattering curves.
The large red-shift in the fluorescence emission spectra
of the WT form and T45S, M88V and D11N mutants
incubated with 9.0
M
urea demonstrated the complete
denaturation of coat protein in all cases. Surprisingly
however, D11N, despite its ts growth phenotype, was
more stable to urea treatment than the WT capsid (Figs 2
and 3B). Moreover, the light scattering data showed no
significant change during urea treatment of the D11N
mutant, suggesting that, even in the denatured state, the
characteristics of a large particle are retained. Apparently,
the subunits remain aggregated in a particle of approxi-
mately virus-size (Fig. 3C). Size-exclusion HPLC was also

utilized. Only for the D11N mutant was a small fraction
Fig. 2. Tryptophan fluorescence spectra of bacteriophage MS2. Spectra
of wild-type (WT) (circles), M88V (diamonds), D11N (squares) and
T45S (triangles) particles were recorded at atmospheric pressure in the
absence (filled symbols), or presence of 4.5
M
(unfilled symbols) or
9.0
M
(lines) urea. The excitation wavelength was 280 nm, and the
emission wavelength range was 300–420 nm. Standard buffer: 50 m
M
Tris/150 m
M
NaCl (pH 7.5). The sample concentration utilized was
50 lgÆmL
)1
.
Fig. 3. Dissociation and denaturation of wild-type (WT) and mutant MS2 particles. (A) Light scattering and spectral center of mass measurements of
WT MS2 as a function of urea concentration. To verify the dissociation and denaturation processes, we measured the light scattering of the particles
at 320 nm (d)andthespectralcenterofmassoftheparticles(s). (B) Urea-induced denaturation of MS2, as measured by tryptophan fluorescence
for: (d), WT MS2; (r), M88V; (j), D11N; and (m), T45S. (C) Urea-induced dissociation, as measured by light scattering for: (d), WT MS2; (r),
M88V; (j), D11N; and (m), T45S. For tryptophan fluorescence emission, the sample was excited at 280 nm and the emission was measured at 300–
420 nm. For the light scattering measurements, the sample was excited at 320 nm and the emission measured from 315 to 325 nm. Standard buffer:
50 m
M
Tris/150 m
M
NaCl (pH 7.5). Fluorescence data points are the average and standard deviation of three experiments (A and B) and light
scattering curves are representative of three measurements. The sample concentration utilized was 50 lgÆmL

)1
.
138 S. M. B. Lima et al. (Eur. J. Biochem. 271) Ó FEBS 2003
of the particle eluted in the same position as untreated
particles (Fig. 4). In contrast, the other mutants and the
WT capsid were irreversibly dissociated and denatured by
a high concentration of urea (Fig. 4). The monomers were
expected to elute close to the total volume of the column
(% 12 min). However, as previously found with other
dissociated capsid proteins, the lack of elution of the
proteins can be explained by non-specific binding of the
denatured coat proteins to the gel [34,35]. The elution of
some D11N as capsids after urea treatment is probably
the result of a small fraction that was not denatured by
urea.
To explore the biological activity of the viral particles, cell
infectivity assays were performed. Treatment of WT, M88V
and T45S particles with 4.5
M
urea resulted in drastic
reductions of phage titer (Table 1). The titer of WT MS2
was reduced by % 1000-fold. M88V and T45S were the most
susceptible, each showing a 1 · 10
6
-fold reduction in titer.
However, consistent with other measures of its stability, the
D11N mutant was relatively unaffected and showed only a
10-fold loss of titer.
Pressure stability of the MS2 bacteriophage
and ts mutants

The effects of high pressure on tryptophan fluorescence
emission spectra of the MS2 and ts mutants were also
investigated. Pressure produced complete dissociation of the
two mutants M88V and T45S. However, up to 3.4 kbar,
hydrostatic pressure was unable to promote complete
dissociation of WT and D11N particles, as measured by
fluorescence (Fig. 5A) and by light scattering (Fig. 5B). In
agreement with the urea studies described above, the D11N
mutant was more stable than the WT bacteriophage. All the
curves for pressure denaturation of WT and mutant
particles seem to have more than one transition, which
may indicate partially dissociated or denatured states.
However, because both fluorescence and light scattering
reveal the average properties, we cannot fully characterize
these potential intermediates.
The reversibility of the process was analyzed by deter-
mining the values of spectral center of mass (Fig. 5A,C),
which were measured after decompression and utilizing
HPLC (Fig. 4). Figure 5C shows that the spectra of the
sample subjected to compression and decompression are
similar to the non-treated sample, even in the case of the
M88V mutant. The elution of the samples after pressuriza-
tion in the same position as the native virus showed that the
particles were able to reassemble correctly, suggesting that
dissociation by pressure is at least partially reversible. To
further investigate the recovery, infectivity assays were
performed and when we used high pressure the phage titer
was similar to that of the control (Table 1). Thus, in spite of
the dissociation induced by pressure, the information for
correct reassembly seems to be largely preserved. In

principle, the recovery of the titer implies that the pressure
disassembly process is reversible, whereas the urea denatur-
ation process is not.
The changes in stability can be clearly evaluated by the
values of pressure and urea that lead to 50% denaturation
(p
½
and U
½
, respectively, shown in Table 2). Table 2 also
shows the apparent volume changes obtained by treating
Fig. 4. High-performance gel filtration chromatography of bacterio-
phage MS2 and the temperature sensitive mutants. Elution profiles of
(A) M88V mutant, (B) wild-type MS2 and (C) D11N mutant. The
unbroken line corresponds to the samples that were incubated with
9.0
M
urea, the dashed line corresponds to the samples without treat-
ment, and the dotted line represents the samples treated with high
pressure and returned to atmospheric pressure. The flow rate was
0.3 mLÆmin
)1
and the elution of the samples was monitored by tryp-
tophan fluorescence (excitation at 280 nm, emission at 330 nm).
Table 1. Effects of pressure and urea on bacteriophage MS2 and
mutants: infectivity assays (37 °C). ND, non-detected; p.f.u., plaque-
forming units.
Titer (p.f.u.ÆmL
)1
)

Control
3.4 kbar
of pressure
4.5
M
urea
9.0
M
urea
MS2 10
8
10
8
10
5
ND
T45S 10
8
10
8
10
2
ND
M88V 10
8
10
7
10
2
ND

D11N 10
8
10
8
10
7
10
2
Ó FEBS 2003 Cavities and stability in MS2 capsid mutants (Eur. J. Biochem. 271) 139
the data of Fig. 5 (Eqn 3). M88V and T45S presented much
larger changes in volume per unit of coat protein than the
WT capsid. A volume change for dissociation of D11N
particles could not be determined because of the lack of
change induced by pressure.
Changes in secondary structure upon dissociation
and denaturation
To further confirm the urea-induced changes in secondary
structure, we analyzed the UV CD spectra of MS2 and the ts
mutants in the presence and absence of 4.5
M
urea, the
concentration at which the greatest difference among the
various samples was observed (Fig. 6). CD spectra evidenced
a great loss of secondary structure for M88V and T45S in
urea. In the absence of urea, the signal was smaller for the
M88V mutant than for the other mutants and WT bacterio-
phage, although all experiments were carried out under
identical conditions and virus concentrations (100 lgÆmL
)1
).

However, the WT bacteriophage and the D11N mutant
showed little change in structure, confirming their higher
stabilities. At higher urea concentrations, both WT bacterio-
phage and the D11N mutant lost the ellipticity at 218 nm,
indicating complete denaturation (results not shown).
Bis-ANS binding assay of MS2 and ts mutants
The fluorophore, bis-ANS, binds non-covalently to non-
polar segments in proteins, especially those in proximity to
positive charges [36]. Its binding is accompanied with a large
increase in its fluorescence quantum yield and it has been
Fig. 5. Pressure stability of bacteriophage MS2 and the temperature
sensitive (ts) mutants. The effect of pressure on the samples was ana-
lyzed at room temperature. (A) The effect was measured by the tryp-
tophan fluorescence emission of the spectral center of mass. The
samples were excited at 280 nm and the emission was measured from
300 to 420 nm for: (d), WT MS2; (r), M88V; j), D11N; and (m),
T45S. (B) Light scattering measurements of MS2 and ts mutants under
pressure. (d) MS2 bacteriophage and the ts mutants (r)M88V,(j)
D11N, and (m) T45S. The excitation wavelength was 320 nm and the
emission wavelength range was 315–325 nm. The incubation time at
each pressure was 10 min. Other conditions were as described in the
legend to Fig. 2. The unfilled symbols correspond to the respective
values after pressure release. Fluorescence data points are the average
and SD of three experiments, and light scattering curves are the rep-
resentative of three measurements. (C) Fluorescence emission spectra
of M88V mutant particles before (unbroken line), under 3.4 kbar of
pressure (broken line, or after decompression (dotted lines).
Table 2. [U]
½
and p

½
values for wild-type (WT) bacteriophage MS2
and mutants. CM, spectral center of mass; LS, light scattering.
Substitution
Surface (S),
buried (B)
or RNA
interaction (I)
[U]
½
(
M
)
p
½
(kbar)
DV/n
a
(mLÆmol
)1
)
CM LS
WT 4.6 3.0 3.1 4.65
M88V B 3.2 1.6 1.4 19.81
T45S I 3.0 1.8 1.5 14.12
D11N S 5.8 >3.2 >3.2 –
a
The apparent volume change of association (DV) was determined
by replotting the data in Fig. 5 according to (Eqn 3) and normal-
ized by dividing by the number of subunits in the capsid (180)

DV/n.
140 S. M. B. Lima et al. (Eur. J. Biochem. 271) Ó FEBS 2003
used to probe protein structural changes [37,38]. At
atmospheric pressure and in the absence of urea, the MS2
bacteriophage and D11N mutant did not bind bis-ANS
(Fig. 7B), showing that these particles do not present
exposed hydrophobic segments. High urea concentrations
did not promote significant binding of bis-ANS to any of
the particles, with the exception of T45S, which bound a
small amount of bis-ANS, increasing, by twofold, the
emission of the probe (Fig. 7B). M88V and WT particles
treated with pressure did not show significant changes in
bis-ANS binding. However, when the T45S mutant was
denatured and dissociated by pressure, a sixfold increase in
the emission of the probe occurred (Fig. 7A), suggesting
that the conformation of the pressure-denatured state is
different from that of the urea-denatured state.
Discussion
In the last 20 years, the structure of many viruses has been
solved by X-ray crystallography [1]. However, despite
extensive knowledge of their structures, the mechanisms of
virus assembly and disassembly are still poorly understood.
In viruses, interactions within and between capsid subunits
must be strong enough to ensure virus particle stability and
protection of the genome, but weak enough to permit
uncoating or release of the genome upon interaction with
the cell. The coat protein of the bacteriophage MS2 has two
basic functions, (a) specific RNA binding for translational
repression and genome encapsidation and (b) formation of
the capsid structure. Structural and genetic analysis permit

the identification of amino acid residues involved in the
protein–RNA and protein–protein interactions that mediate
these functions [9,31,32,39–42]. Analysis of the effects of
substitutions of key amino acids should ultimately provide
information about the roles they play in capsid assembly
and disassembly. In this work, we studied the effects of
chemical (i.e. urea) and physical (i.e. high pressure) dena-
turing agents in the structure and stability of the MS2
bacteriophage, and of three ts mutants, with the aim to
eventually understand the assembly and disassembly pro-
cesses. Our results allow us to infer how the introduction of
a small cavity in the coat protein affects the whole stability
of the virus particle. However, local mutations also tend to
affect the global conformation of the protein. For the
pressure sensitivity, modification in the dynamics of the
protein, even with the average structure not affected, may
result in changes in pressure stability [43].
M88V showed a large decrease in stability when exposed
to conditions of high pressure and high concentrations of
urea. Its diminished stability can probably be explained by
the large potential of the substitution to create a cavity, as
well to sterically interfere with side-chain packing in the
protein’s interior. Mutations that create cavities in hydro-
phobic environments generally cause proteins to become
less stable [44,45]. Using the program
VMD
[46] and a probe
radius of 1.4 A
˚
, we analyzed the pdb coordinates of WT

bacteriophage [14] and the M88V mutant (by substitution
using the same program) for the existence of internal cavities
in this region in the structure. A significant cavity in the WT
phage structure in the neighboring Met88 residue was
identified. After substitution of this residue with valine, the
cavity volume increased to 43 A
˚
3
, reflecting a reduction in
the surface area of the residue (% 59 A
˚
2
)andinthe
interactions occurring there (Fig. 8). Met88 resides in the
Fig. 6. UV CD spectra of wild-type MS2
bacteriophage and the temperature sensitive (ts)
mutants. Conformational changes in the sec-
ondary structure of bacteriophage MS2 and ts
mutants were analyzed in the presence of
4.5
M
urea (hollow symbols). Filled symbols
correspond to the samples in the absence of
urea. Wavelength range: 300–210 nm. The
samples of MS2 bacteriophage and ts mutants
were diluted to a final concentration of
100 lgÆmL
)1
and the spectra were measured in
buffer (10 m

M
Tris/30 m
M
NaCl, pH 7.5)
using a 0.1 cm pathlength quartz cuvette. The
data are representative of three experiments.
Ó FEBS 2003 Cavities and stability in MS2 capsid mutants (Eur. J. Biochem. 271) 141
middle of the central b-strand of the coat protein dimer and
its side-chain projects into the protein’s interior (Figs 1B
and 8). The two Met88 residues are in close proximity to one
another at the dimer’s twofold symmetry axis and thus
interact with each other across the dimer interface. Each
also interacts with residues on the adjacent b-strand (F) and
with amino acids in the alpha-helical regions as they pass
over the b-sheet. Thus, the M88V mutant substitution
affects the dimer interface, explaining its increased sensitiv-
ity to pressure (Dp
½
¼ 1.4 kbar, Table 2). The increase
in volume of 43 A
˚
3
corresponds to 26 mLÆmol
)1
of coat
subunit, which, by itself, could explain the difference (of
% 15 mLÆmol
)1
) between the volumes measured for the
M88V mutant and the WT phage (Table 2). In addition to

the creation of a cavity, the substitution of Val for Met
might produce steric clashes as a result of introduction of
the beta branched side-chain where the unbranched Met
side-chain is ordinarily packed.
The physical basis for the reduced stability of the T45S
mutant may have a similar explanation. Although residue
45 resides on the surface of the protein and makes no
obvious stabilizing interactions with other amino acid side-
chains, this part of the protein interacts with genomic RNA
[33,41,47]. Such interactions probably contribute to the
stability of the virus particle (Fig. 1B). The crystal structure
of coat protein in complex with the translational operator
shows interaction between Thr45 and RNA [41]. The X-ray
structure of the virus particle itself shows significant electron
density in the vicinity of Thr45, indicating that many
individual subunits apparently contact RNA at this posi-
tion, albeit in a presumably non-specific manner [13,14].
Thus, it is clear that the T45S substitution destabilizes the
capsid because of a perturbation of the protein–RNA
interaction, as assessed by the several criteria reported here.
The high sensitivity of coat protein folding/stability to
Thr45 substitution was previously inferred from electroph-
oretic studies [41]. Nineteen substitutions were introduced in
position 45 and none of the amino acids was a completely
acceptable replacement for threonine. Every mutant showed
loss of translational repression, increased insolubility and/or
degradation, and failure to produce normal quantities of
virus-like particles. However, the T45S mutant was the most
affected in the translational repressor activity [41].
The T45S mutant was more sensitive to urea than to high

pressure. This sensitivity to urea treatment can be explained
by the involvement of other forces, such as hydrogen bonds,
in the interactions at the protein–RNA interface. X-ray
structural analysis suggests that the hydroxyl group of
Thr45 makes H-bonds with N6 and N7 of A-4 [47].
Moreover, the lack of binding of bis-ANS in this mutant,
following treatment with urea, suggests that the dissociation
and denaturation processes induced by urea and pressure
occur in a distinct way. More likely, urea provokes a general
unfolding, not leaving any hydrophobic cleft for bis-ANS
binding. High pressure is well known to denature proteins to
partially folded states [15,17,38,48–50]. The large increase in
bis-ANS binding when the T45S mutant was treated with
pressure can also be explained by the presence of a cavity in
the T45S mutant, which is released by dissociation, leading
to binding of the dye.
The changes promoted by pressure were reversible for
WT and all the ts mutants, whereas the effects of urea were
irreversible. Only the D11N mutant showed a residual titer
after treatment with 4.5
M
urea. The UV CD spectra, under
different conditions, provide further insights into the
stability of the WT and mutant capsids. The less stable
mutants presented higher positive ellipticity peaks in
the RNA region (270 nm), whereas the negative peak
Fig. 7. Binding of bis-8-anilinonaphthalene-1-sulfonate (bis-ANS) to
dissociated and denatured capsids. (A) bis-ANS binding to MS2 bac-
teriophage and mutants under pressure. Inset: fluorescence emission
spectra of bis-ANS during the pressurization process of the T45S

mutant. (B) bis-ANS binding to MS2 bacteriophage and mutants
under conditions of increasing urea concentrations. Structural changes
were also analyzed by a fluorescent probe (bis-ANS) emission, at a
final concentration of 2 l
M
(d) MS2 bacteriophage and the ts mutants
(m)T45Sand(j) D11N. The excitation wavelength was 360 nm and
the emission wavelength range was 400–600 nm. The data shown are
representative of three experiments.
142 S. M. B. Lima et al. (Eur. J. Biochem. 271) Ó FEBS 2003
(corresponding to the b-sheet) is smaller. An appealing
interpretation for these findings is that the lower stability of
the coat protein shell is counterbalanced by a more
structured RNA, which results in a similar infectivity.
The increased stability of the D11N mutant presents a
puzzle. Like the others, this mutant was isolated for its ts
growth phenotype, implying that it possesses decreased
thermal stability. Why then is the virus apparently more
stable? The X-ray structure of MS2 shows that Asp11
participates in both intra- and interchain interactions. It
appears to form a salt bridge to Lys113 on the alpha-helix of
the adjacent subunit of the dimer, and may thus help to
stabilize the dimer interface (residues in red and blue colors
in Fig. 1B). But Asp11 makes another interaction. It resides
in b-strand A, just before the turn that connects it to bB.
Here it seems to position itself to the H-bond, through a
carbonyl oxygen of its side-chain, to the main chain amide
of Gly13 within the turn, thus possibly exerting a stabilizing
influence on the turn. The D11N substitution might be
expected to have a destabilizing effect on both interactions,

making it difficult to understand its stabilizing effect on the
virus particle. Thus, the stabilization produced by the D11N
substitution is probably associated with a change in the
overall conformation propagated from the local replace-
ment. One can envisage that the salt-bridge in the WT
capsid protein locks the helix in a conformation leading to
cavities in the interior of the hydrophobic core. King and
colleagues [51,52] characterized a large collection of ts
mutants of the phage P22 tailspike protein. Many of these
are so-called temperature-sensitive for folding or tsf muta-
tions. They have the property of preventing folding at the
non-permissive temperature without altering appreciably
the stability of the native state, once formed. A dispropor-
tionate fraction of these mutations affect residues at or near
b-turns, suggesting that the proper formation of such turns
represents a crucial step in the folding process. Substitutions
in positions near b-turns may destabilize important folding
intermediates, rendering them aggregation-prone. D11N is
apparently in this category.
All the mutants were subjected to a test of their stability
to assemble into virus-like particles, which is an activity
sensitive to folding defects. As virus-like particles have a
characteristic electrophoretic behavior, this provides a
simple means of assessing a mutant for acquisition of native
protein structure (D. S. Peabody, unpublished data).
Therefore, the changes in the stability observed here are
not the result of defects in the assembly pathway.
In conclusion, our studies show the high sensitivity of
single amino-acid substitutions in the coat protein of small
RNA viruses. Lower stability was correlated with a local

increase in the cavity. Interaction with RNA is also sensitive,
and its perturbation (in the case of T45S replacement) also
leads to lower stability. On the other hand, capsid proteins
cannot be highly packed otherwise they would lose the
flexibility needed for virus assembly. In this context, the
intricate interactions between capsid protein and RNA are
Fig. 8. Cavity increase occuring in the M88V mutation. Using the program
VMD
and a probe radius of 1.4 A
˚
, we analyzed the pdb coordinates of
wild-type (WT) bacteriophage and the M88V mutant for the existence of internal cavities in this region. The figure shows the region on coat protein
in the asymmetric unit of the capsid around residue 88. The methionine residue is shown in yellow and the valine is represented in green. A
significant cavity was identified in the WT phage structure (A and B) in the neighboring region of the Met88 residue. After the substitution of this
residue with valine (C and D), the cavity volume increased to 43 A
˚
3
, reflecting a reduction in the surface area of the residue (% 59 A
˚
2
)andinthe
interactions occurring there.
Ó FEBS 2003 Cavities and stability in MS2 capsid mutants (Eur. J. Biochem. 271) 143
crucial for assembling the whole particle. It is also interest-
ing that the mutants had a ts phenotype, which means a
high sensitivity to both high and low temperatures. The
presence of cavities that confer the decreased stability to
pressure is also usually related to both decreased stability to
low and high temperatures [15,17,18].
Acknowledgements

We gratefully acknowledge Alan Witer Sousa da Silva from Labo-
rato
´
rio de Fı
´
sica Biolo
´
gica at Instituto de Biofı
´
sica Carlos Chagas
Filho/UFRJ for help with the
VMD
program, Emerson Gonc¸ alves for
competent technical assistance, and Cristiane Dinis Ano Bom and
Professors Fa
´
bio Almeida and Ana Paula Valente from CNRMN/
UFRJ for helpful comments and suggestions. This work was
supported, in part, by an International Grant from the Howard
Hughes Medical Institute to J.L.S. and by grants from Programa de
Nu´ cleos de Exceleˆ ncia (PRONEX), Conselho Nacional de Desenvolvi-
mento Cientı
´
fico e Tecnolo
´
gico (CNPq), Fundac¸ a
˜
odeAmparoa
Pesquisa do Estado do Rio de Janeiro (FAPERJ), Fundac¸ a
˜

o
Universita
´
ria Jose
´
Bonifa
´
cio(FUJB)ofBraziltoJ.L.S.andA.C.O.,
and by a grant from the National Institutes of Health (NIH) to D.S.P.
References
1. Johnson, J.E. (1996) Functional implications of protein–protein
interactions in icosahedral viruses. Proc. Natl Acad. Sci. USA 93,
27–33.
2. Tuma, R., Tsuruta, H., Benevides, J.M., Prevelige, P.E. Jr &
Thomas, G.J. Jr (2001) Characterization of subunit structural
changes accompanying assembly of the bacteriophage P22 pro-
capsid. Biochemistry 40, 665–674.
3. Turner, B.G. & Summers, M.F. (1999) Structural biology of HIV.
J. Mol. Biol. 285, 1–32.
4. Skehel, J.J. & Wiley, D.C. (2000) Receptor binding and membrane
fusion in virus entry: the influenza hemagglutinin. Annu. Rev.
Biochem. 69, 531–569.
5. Oliveira, A.C., Gomes, A.M.O., Almeida, F.C., Mohana-Borges,
R.,Valente,A.P.,Reddy,V.S.,Johnson,J.E.&Silva,J.L.(2000)
Virus maturation targets the protein capsid to concerted dis-
assembly and unfolding. J. Biol. Chem. 275, 16037–16043.
6. Oliveira, A.C., Ishimaru, D., Gonc¸ alves, R.B., Smith, T.J.,
Mason, P., Sa
´
-Carvalho, D. & Silva, J.L. (1999) Low temperature

and pressure stability of picornaviruses: implications for virus
uncoating. Biophys. J. 76, 1270–1279.
7. Silva, J.L., Oliveira, A.C., Gomes, A.M.O., Lima, L.M., Mohana-
Borges, R., Pacheco, A.B.F. & Foguel, D. (2002) Pressure induces
folding intermediates that are crucial for protein–DNA recogni-
tion and virus assembly. Biochim. Biophys. Acta 1595, 250–265.
8. Carey, J., Cameron, V., de Haseth, P.L. & Uhlenbeck, O.C. (1983)
Sequence-specific interaction of R17 coat protein with its ribo-
nucleic acid binding site. Biochemistry 22, 2601–2610.
9. Romaniuk, P.J., Lowary, P., Wu, H.N., Stormo, G. & Uhlenbeck,
O.C. (1987) RNA binding sites of R17 coat protein. Biochemistry
26, 1563–1568.
10. Johansson, H.E., Dertinger, D., LeCuyer, K.A., Behlen, L.S.,
Greef, C.H. & Uhlenbeck, O.C. (1998) A thermodynamic analysis
of the sequence-specific binding of RNA by bacteriophage MS2
coat protein. Proc. Natl Acad. Sci. USA 95, 9244–9249.
11. Ling,C.M.,Hung,P.P.&Overby,L.R.(1969)Specificityinself-
assembly of bacteriophages Q beta and MS2. Biochemistry 8,
4464–4469.
12. Witherell, G.W., Gott, J.M. & Uhlenbeck, O.C. (1991) Specific
interaction between RNA phage coat proteins and RNA. Prog.
Nucleic Acids Res. Mol. Biol. 40, 185–220.
13. Valegard, K., Liljas, L., Fridborg, K. & Unge, T. (1990) The three-
dimensional structure of the icosahedral bacterial virus MS2.
Nature 345, 36–41.
14. Golmohammadi, R., Valegard, K., Fridborg, K. & Liljas, L.
(1993) The refined structure of Bacteriophage MS2 at 2.8 A
˚
resolution. J. Mol. Biol. 234, 620–639.
15. Silva, J.L. & Weber, G. (1993) Pressure stability of proteins. Annu.

Rev. Phys. Chem. 44, 89–113.
16. Weber, G. (1992) Protein Interactions. Chapman & Hall, New
York.
17. Silva, J.L., Foguel, D. & Royer, C.A. (2001) Pressure provides new
insights into protein folding, dynamics and structure. Trends
Biochem. Sci. 26, 612–618.
18. Balny, C., Masson, P. & Heremans, K. (2002) High pressure
effects on biological macromolecules: from structural changes
to alteration of cellular processes. Biochim. Biophys. Acta 1595,
3–10.
19. Silva, J.L. & Weber, G. (1988) Pressure-induced dissociation of
BromeMosaicVirus.J. Mol. Biol. 199, 149–161.
20. Silva, J.L., Luan, P., Glaser, M., Voss, E. & Weber, G. (1992)
Effects of hydrostatic pressure on a membrane-enveloped virus:
high immunogenicity of the pressure-inactivated virus. J. Virol. 66,
2111–2117.
21. Da Poian, A.T., Oliveira, A.C., Gaspar, L.P., Silva, J.L. & Weber,
G. (1993) Reversible pressure dissociation of R17 bacteriophage:
the physical individuality of virus particles. J. Mol. Biol. 231,
999–1008.
22. Prevelige, P., King, J. & Silva, J.L. (1994) Pressure denaturation of
bacteriophage P22 coat protein and its entropic stabilization in the
icosahedral shells. Biophys. J. 66, 1631–1641.
23. Leimkuhler, M., Goldbeck, A., Lechner, M.D. & Witz, J. (2000)
Conformational changes preceding decapsidation of bromegrass
mosaic virus under hydrostatic pressure: a small-angle neutron
scattering study. J. Mol. Biol. 296, 1295–1305.
24. Frye, K.J. & Royer, C.A. (1998) Probing the contribution of
internal cavities to the volume change of protein unfolding under
pressure. Protein Sci. 7, 2217–2222.

25. de Sousa, P.C. Jr, Tuma, R., Prevelige, P.E. Jr, Silva, J.L. &
Foguel, D. (1999) Cavity defects in the procapsid of bacteriophage
P22 and the mechanism of capsid maturation. J. Mol. Biol. 287,
527–538.
26. Bradford, M.M. (1976) A rapid and sensitive method for the
quantification of microgram quantities of protein utilizing the
principle of protein–dye binding. Anal. Biochem. 72, 248–254.
27. Paladini, A.A. & Weber, G. (1981) Pressure-induced reversible
dissociation of enolase. Biochemistry 20, 2587–2593.
28. Cadwell, R.C. & Joyce, G.F. (1992) Randomization of genes by
PCR mutagenesis. PCR Methods Applications 2, 28–33.
29. Shaklee, P.N. (1990) Negative-strand RN A replication by Qb and
MS2 positive strand replicases. Virology 178, 340–343.
30. Davanloo, P., Rosenberg, A.H., Dunn, J.J. & Studier, F.W. (1984)
Cloning and expression of the gene for bacteriophage T7 RNA
polymerase. Proc. Natl Acad. Sci. USA 81, 2035–2039.
31. Peabody, D.S. (1993) The RNA binding site of bacteriophage
MS2 coat protein. EMBO J. 12, 595–600.
32. Peabody, D.S. (1990) Translational repression by bacteriophage
MS2 coat protein expressed from a plasmid. J. Biol. Chem. 265,
5684–5689.
33. Stonehouse, N.C., Valegard, K., Golmohammadi, R., van den
Worm, S., Walton, C., Stockley, P.G. & Liljas, L. (1996) Crystal
structures of MS2 capsids with mutations in the subunit FG loop.
J. Mol. Biol. 256, 330–339.
34. Da Poian, A.T., Oliveira, A.C. & Silva, J.L. (1995) Cold dena-
turation of an icosahedral virus: the role of entropy in virus
assembly. Biochemistry 34, 2672–2677.
144 S. M. B. Lima et al. (Eur. J. Biochem. 271) Ó FEBS 2003
35. Gaspar, L.P., Johnson, J.E., Silva, J.L. & Da Poian, A.T. (1997)

Different partially folded states of the capsid protein of cowpea
severe mosaic virus in the disassembly pathway. J. Mol. Biol. 273,
456–466.
36. Rosen, C.G. & Weber, G. (1969) Dimer formation from 1-amino-
8-naphthalenesulfonate catalyzed by bovine serum albumin. A
new fluorescent molecule with exceptional binding properties.
Biochemistry 8, 3915–3920.
37. Horowitz, P., Prasad, V. & Luduena, R.F. (1984) Bis (1,8-anili-
nonaphthalenesulfonate). A novel and potent inhibitor of micro-
tubule assembly. J. Biol. Chem. 259, 14647–14650.
38. Silva, J.L., Silveira, C.F., Correia Junior, A. & Pontes, L. (1992)
Dissociation of a native dimer to a molten globule monomer:
effects of pressure and dilution on the association equilibrium of
arc repressor. J. Mol. Biol. 223, 545–555.
39. Lim, F. & Peabody, D.S. (1994) Mutations that increase the affi-
nity of a translational repressor for RNA. Nucleic Acids Res. 22,
3748–3752.
40. Lim, F., Spingola, M. & Peabody, D.S. (1994) Altering the RNA
binding specificity of a translational repressor. J. Biol. Chem. 269,
9006–9010.
41. Peabody, D.S. & Chakerian, A. (1999) Asymmetric contributions
to RNA binding by the Thr45 residues of the MS2 coat protein
dimer. J. Biol. Chem. 274, 25403–25410.
42. Dertinger, D., Dale, T. & Uhlenbeck, O.C. (2001) Modifying
the specificity of an RNA backbone contact. J. Mol. Biol. 314,
649–654.
43. Consonni, R., Santomo, L., Fusi, P., Tortora, P. & Zetta, L.
(1999) A single-point mutation in the extreme heat- and pressure-
resistant Sso7d protein from Sulfolobus solfataricus leads to a
major rearrangement of the hydrophobic core. Biochemistry 38,

12709–12717.
44. Ericsson, A.E., Baase, W.A., Zhang, X.J., Heinz, D.W., Blaber,
M., Baldwin, E.P. & Matthews, B.W. (1992) Response of protein
structure to cavity-creating mutations and its relation to the
hydrophobic effect. Science 255, 178–183.
45. Lee, C., Park, S.H., Lee, M.Y. & Yu, M.H. (2000) Regulation of
protein function by native metastability. Proc.NatlAcad.Sci.
USA 97, 7727–7731.
46. Humphrey, W., Dalke, A. & Schulten, K. (1996) VMD: visual
molecular dynamics. J. Mol. Graph. 14, 33–38. (.
uiuc.edu/Research/vmd/)
47. Powell, A.J. & Peabody, D.S. (2001) Asymmetric interactions in
the adenosine-binding pockets of the MS2 coat protein dimer.
BMC Mol Biol. 2,6.
48. Lima, L.M.T., Foguel, D. & Silva, J.L. (2000) DNA tightens the
dimeric DNA-binding domain of human papillomavirus E2 pro-
tein without changes in volume. Proc. Natl Acad. Sci. USA 97,
14289–14294.
49. Inoue, K., Yamada, H., Akasaka, K., Herrmann, C., Kremer, W.,
Maurer, T., Doker, R. & Kalbitzer, H.R. (2000) Pressure-induced
local unfolding of the Ras binding domain of RalGDS. Nat.
Struct. Biol. 7, 547–550.
50. Suarez, M.C., Lehrer, S.S. & Silva, J.L. (2001) Local hetero-
geneity in the pressure denaturation of the coiled-coil tropomyosin
because of subdomain folding units. Biochemistry 40, 1300–
1307.
51. Galisteo, M.L., Gordon, C.L. & King, J. (1995) Stability of wild-
type and temperature-sensitive protein subunits of the phage P22
capsid. J. Biol. Chem. 270, 16595–16601.
52. Teschke, C.M. & King, J. (1995) In vitro folding of phage P22 coat

protein with amino acid substitutions that confer in vivo tem-
perature sensitivity. Biochemistry 34, 6815–6826.
Ó FEBS 2003 Cavities and stability in MS2 capsid mutants (Eur. J. Biochem. 271) 145