Báo cáo khoa học: Monomeric molten globule intermediate involved in the equilibrium unfolding of tetrameric duck d2-crystallin pdf

Monomeric molten globule intermediate involved in the equilibrium
unfolding of tetrameric duck d
2
-crystallin
Hwei-Jen Lee
1
, Shang-Way Lu
1
and Gu-Gang Chang
2
1
Department of Biochemistry, National Defense Medical Center, Taipei, Taiwan;
2
Faculty of Life Sciences, National Yang-Ming
University, Taipei, Taiwan
Duck d
2
-crystallin is a soluble tetrameric lens protein. In
the presence of guanidinium hydrochloride (GdnHCl), it
undergoes stepwise dissociation and unfolding. Gel-filtra-
tion chromatography and sedimentation velocity analysis
has demonstrated the dissociation of the tetramer protein to
a monomeric intermediate with a dissociation constant of
0.34 l
M
3
. Dimers were also detected during the dissociation
and refolding processes. The sharp enhancement of 1-anilino
naphthalene-8-sulfonic acid (ANS) fluorescence at 1
M
GdnHCl strongly suggested that the dissociated monomers

were in a molten globule state under these conditions. The
similar binding affinity (% 60 l
M
) of ANS to protein in the
presence or absence of GdnHCl suggested the potential
assembly of crystallins via hydrophobic interactions, which
might also produce off-pathway aggregates in higher protein
concentrations. The dynamic quenching constant corres-
ponding to GdnHCl concentration followed a multistate
unfolding model implying that the solvent accessibility of
tryptophans was a sensitive probe for analyzing d
2
-crystallin
unfolding.
Keywords: d-crystallin; lens protein; unfolding; dissociation;
argininosuccinate lyase.
d
2
-Crystallin, a highly concentrated yet soluble protein in
avian and reptile eye lens, acts as an important structural
protein for light refraction [1–3]. Thermodynamic stability
for crystallins is essential in maintaining lens transparency
[4]. Determining the mechanism of folding and assembly of
these proteins is important for understanding how they can
form stable transparent structures at high concentrations.
The d
2
-crystallin in lens was recruited during evolution
from argininosuccinate lyase, an enzyme involved in
arginine biosynthesis derived from the urea cycle. These

two proteins shared over 90% sequence homology and thus
have similar tertiary structures [5–7]. d
2
-Crystallin has a high
helical content and constitutes a unique liquid-like region in
the center of the duck lens [8]. The central 20-helix core
contributes to the major interactions between subunits, and
is crucial for subunit association. The active site is located at
a boundary composed of three subunits [8–11].
Identifying and characterizing of possible conformational
states in the pathways leading to folding and unfolding are
important. For d
2
-crystallin, characterization of the inter-
molecular association of the helix bundles and the partial
unfolded intermediate with exposed hydrophobic region
leading to polymerization remains to be elucidated [12].
Most of the established models of reversible unfolding
of proteins are based on experiments exploring the effect
of chemical denaturants or temperature. These models,
although providing useful information on the unfolding
mechanism, are limited to small, monomeric proteins. For
multimeric proteins, detection of partially unfolded inter-
mediate is always complicated by the dissociation step
[13,14]. Only in limited cases can dissociation and unfolding
be clearly distinguished [15–19]. In previous studies we have
demonstrated that duck d
2
-crystallin can be reversibly
dissociated and unfolded by GdnHCl [15]. The dissociation

of tetrameric d
2
-crystallin is accompanied by loss of
argininosuccinate lyase activity at around 0.9
M
GdnHCl,
which produces monomeric d
2
-crystallin as judged by gel-
filtration chromatography [15]. At higher GdnHCl concen-
trations, the monomer unfolds via a partially unfolded
intermediate before denaturation. Structural information
has revealed that tetrameric d
2
-crystallin possesses a double
dimer structure [10]. The dimeric form of d
2
-crystallin can be
observed under acidic conditions [20].
In this report, we investigate the detailed dissociation
process for d
2
-crystallin during chemical denaturation by
GdnHCl. Gel-filtration chromatography and analytical
ultracentrifugation were used to identify the dissociation
intermediate. Unfolding of the dissociated monomers was
investigated using 1-anilino-8-naphthalene sulfonate (ANS)
binding, fluorescence studies and circular dichroism (CD).
Materials and methods
Materials

Ultra-pure guanidine hydrochloride and acrylamide was
purchased from Baker (Phillipsburg, NJ, USA). Tris (base),
EDTA and NaCl were obtained from Merck AG
Correspondence to H J. Lee, Department of Biochemistry, National
Defense Medical Center, no. 161, Sec. 6, Minchuan East Road,
Neihu 114, Taipei, Taiwan.
Fax: + 88 62 87923106, Tel.: + 88 62 87910832,
E-mail:
Abbreviations: ANS, 1-anilinonaphthalene-8-sulfonic acid; GdnHCl,
guanidinium hydrochloride; CD, circular dichroism.
(Received 12 May 2003, revised 14 July 2003,
accepted 7 August 2003)
Eur. J. Biochem. 270, 3988–3995 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03787.x
(Darmstadt, Germany), and argininosuccinic acid (diso-
dium salt) from Sigma Chemical Co. 1-Anilinonaphthalene-
8-sulfonic acid (ANS) was obtained from Molecular Probe
(Eugene, OR, USA). All other chemicals were of analytical
grade and used without further purification.
Purification of duck lens d
2
-crystallin
Duck d
2
-crystallin was purified as described previously
[6,15] except that 50 m
M
Tris/HCl, 0.5 m
M
EDTA, pH 7.5
was used to equilibrate the column and d

2
-crystallin was
eluted in the same buffer containing 0.12
M
NaCl. The
pooled d
2
-crystallin showed one major band upon SDS/
PAGE analysis. The purity of the protein was further
analyzed by gel filtration chromatography using Superdex
200 HR (10/30) column equilibrated in 100 m
M
Tris/HCl
buffer, pH 7.5. The relative area of the major peak
accounted for around 85% of the total protein. Protein
concentrations were determined spectrophotometrically as
described previously [15].
Equilibrium unfolding and refolding studies
A GdnHCl stock solution (7
M
) was prepared in 50 m
M
Tris/HCl buffer (pH 7.5), and the pH of the stock
solution was readjusted to pH 7.5. Equilibrium unfolding
and refolding experiments were performed according to
the methods described previously [15]. The endogenous
argininosuccinate lyase activity of d
2
-crystallin was moni-
tored as a function of the appearance of fumarate at

240 nm. Tryptophan fluorescence was measured by the
emission spectra at an excitation wavelength of 295 nm.
Far-ultraviolet (200–250 nm) CD data were obtained in a
Jasco 810 spectropolarimeter equipped with a thermo-
statically controlled cell holder with a 10-mm path length
cell.
Fluorescence quenching measurements
Fluorescence quenching experiments were performed by
adding aliquots of stock acrylamide or KI solution into the
GdnHCl denatured proteins. The fluorescence emission
spectra with excitation wavelength at 295 nm were moni-
tored. The concentrations of quencher added were less than
0.3
M
. Sodium thiosulfate (0.1 m
M
) was added to the KI
stock solution to prevent I
–
formation. The inner filter effect
due to the absorption of acrylamide or KI at 295 nm was
corrected for by multiplying the fluorescence intensity by
10
A/2
, where A is the absorbance of the solution at 295 nm.
Fluorescence quenching data were fitted to a modified
Stern–Volmer equation [21]:
F
0
=DF ¼ 1=ðf

a
K
SV
½QÞ þ 1=f
a
where DF ¼ F
0
– F, where K
SV
is the dynamic quenching
constant and f
a
was the fractional maximum accessible
protein fluorescence.
ANS binding assay
The exposed hydrophobic surfaces of the protein were
assayed by incubating the GdnHCl-denatured protein in the
dark with ANS (50 l
M
)for4hat25°C. The fluorescence
emission spectra of the protein solution at an excitation
wavelength of 370 nm were monitored. Appropriate blank
Fig. 1. Gel-filtration profiles of the equili-
brium-unfolded duck d-crystallin in GdnHCl.
d-Crystallin 0.6 l
M
(A), or 2.4 l
M
(B), was in
equilibrium unfolded in 0–5

M
GdnHCl. The
M
r
markers (.) were (from left to right):
thyroglobulin (669 kDa), ferritin (440 kDa),
catalase (232 kDa), albumin (67 kDa), and
ovalbumin (43 kDa). The elution positions
corresponding to monomers (M), dimers (D),
tetramers (T), unfolding (U) and polymers (P)
forms are also labeled. Refolding of the
unfolded duck d-crystallin (2.4 l
M
)in
GdnHCl was examined by a 10-fold dilution
of equilibrium unfolded d-crystallins at 1, 2
and 5
M
GdnHCl to 0.1, 0.2 and 0.5
M
,
respectively, and analysis by gel filtration
chromatography (dashed lines).
Ó FEBS 2003 d-crystallin in guanidinium chloride (Eur. J. Biochem. 270) 3989
spectra of ANS in the corresponding GdnHCl solutions
were subtracted from the observed values.
Gel-filtration chromatography
Gel-filtration chromatography was performed with an
Amersham Biosciences A
¨

KTA FPLC system using a
Superdex 200 HR 10/30 column. d
2
-Crystallin (0.6 and
2.4 l
M
, 100 lL) equilibrated in buffer and 1
M
GdnHCl
was loaded onto the column pre-equilibrated with 50 m
M
Tris/HCl buffer containing the same concentration of
GdnHCl, pH 7.5. Calibrated standards were measured
under the same conditions.
The partition coefficient (K
av
) of the eluted component
was calculated by the following equation:
K
av
¼ðV
e
À V
0
Þ=ðV
t
À V
0
Þ
where, V

e
is the elution volume of the protein elution
peak and V
t
and V
0
are the total volume and the void
volume of the column, respectively.
Analytical ultracentrifugation
All analyses were performed at 20 °CusingaBeckman
Optima XL-A analytical ultracentrifuge and an An-60 Ti
rotor. For sedimentation velocity experiments, a sample
volume of 0.45 mL was used, and the radial scans were
recorded at5-min intervals at rotor speedof 50 000 r.p.m. for
2.5 h. The
SEDFIT
software was used for data analysis [22].
Sedimentation equilibrium measurements were per-
formed at two speeds, 18 000 r.p.m. and 20 000 r.p.m.
ORIGIN
software was used for data analysis. Protein
partial specific volumes in GdnHCl solution were calcu-
lated from amino acid composition [23]. The solvent
density was estimated as described [24]. For each set of
experiments, a single species model was used to estimate
the apparent weight–average molecular mass (M
W,app
)of
the protein. The data were further fitted to a T–M
association system. A molar extinction coefficient of

Fig. 2. Time-dependent unfolding of duck d
2
-crystallin analyzed by gel-
filtration chromatography. (A) Traces Ôa–cÕ correspond to d
2
-crystallin
(2.4 l
M
) incubated with 1
M
GdnHCl for 0, 10 and 60 min, respect-
ively. The M
r
markers (.) were (from left to right): thyroglobulin
(669 kDa), ferritin (440 kDa), catalase (232 kDa), albumin (67 kDa),
and ovalbumin (43 kDa). (B) Relative amount of monomers measured
from the peak height.
Fig. 3. Continuous sedimentation coefficient (A) and molecular mass (B)
distributions of duck d
2
-crystallin in GdnHCl. d
2
-Crystallin (1.3 l
M
)was
equilibrated in 0 (d), 0.6 (s), 0.8 (m), 1 (n)and2
M
(j)GdnHCl
solution.
3990 H J. Lee et al. (Eur. J. Biochem. 270) Ó FEBS 2003

1.1 · 10
5
M
)1
Æcm
)1
for d
2
-crystallin was used in calcula-
tion of dissociation constant [5]. The fit quality of the
models was examined by the residuals and by minimiza-
tion of the fit variance.
Results
Dissociation of tetrameric d
2
-crystallin in GdnHCl
A complex unfolding process is observed when tetrameric
duck d
2
-crystallin is equilibrated in GdnHCl solutions [15].
The protein appears to dissociate into monomers before
further unfolding occurs. The detailed dissociation/unfold-
ing mechanism of the protein remains to be determined.
In this study, we examined the M
r
of the protein after
equilibration in various denaturing concentrations of
GndHCl. Figure 1 shows the distribution of various species
under different GdnHCl at two different protein concen-
trations [0.6 l

M
for (A) and 2.4 l
M
for (B)]. Dissociation of
the protein at 1
M
GdnHCl and the unfolding at higher
GdnHCl concentrations are observed. The unfolded forms
easily polymerize and finally aggregated. This process can
be reversed simply by dilution.
The time course of the dissociation process of the protein
at 1
M
GdnHCl was also determined (Fig. 2). Dissociation
takes place a few minutes after addition of 1
M
GdnHCl
(Fig. 2A). More than one protein form was resolved,
including a form eluting earlier than the native tetramer and
a form of intermediate size between the tetramer and
monomer, which was ascribed to a dimeric form also
observed during refolding (Fig. 1). Dissociation of the
tetrameric form reached equilibrium after 30 min (Fig. 2B).
The peak that eluted earlier than native protein could
represent a partially unfolded tetrameric form, which is
probably a molten-globule state with exposed hydrophobic
patches as detected by ANS binding (Fig. 6A). This
partially unfolded form is easily polymerized as demonstra-
ted by chromatography with lower K
av

and was dependent
on the protein concentration. Upon dilution of the GdnHCl
concentration aggregates redissolved (Fig. 1).
Sedimentation experiments
Dissociation of d
2
-crystallin in 1
M
GdnHCl was further
examined by analytical ultracentrifugation. Native tetra-
mers apparently sedimented as a single species with a
sedimentation coefficient (velocity of sedimentation divided
by the acceleration of the force field) of 9.14 s (Fig. 3).
Incubation of the protein in 0.2
M
or 0.4
M
GdnHCl
resulted in a decrease in the s-value to 8.87 s and 8.51 s,
respectively. This result confirmed that the change in K
av
observed by gel-filtration chromatography was due to
protein unfolding. Two components appeared at 0.6
M
GdnHCl (Fig. 3). At 0.8
M
GdnHCl, two components were
observed with sedimentation coefficients of 7.8 and 3.8 s,
respectively, consistent with the species observed by gel-
filtration chromatography (Fig. 1). In 1

M
GdnHCl
solution, only one component was detected with a sedi-
mentation coefficient of 3.76 s (Fig. 3). The predominant
component detected by sedimentation coefficient distribu-
tion had a molecular mass corresponded to that of a
monomer. However, M
r
of the component with larger
s-value at 0.8
M
GdnHCl was determined to about
130 kDa, which is consistent with it being a dissociated
Fig. 4. Gel-filtration chromatography and
sedimentation equilibrium ultracentrifugation
analysis of duck d
2
-crystallin (0.6 l
M
) equili-
brated with 1
M
(A,B) and 2
M
(C,D) GdnHCl.
(A) and (C) Gel-filtration chromatography
analysis. The M
r
markers used (Ñ) were (from
left to right): thyroglobulin (669 kDa), ferritin

(440 kDa), catalase (232 kDa), albumin
(67 kDa), and ovalbumin (43 kDa). The
dashed line in (C) is the elution profile of
native d
2
-crystallin. (B,D) Sedimentation
equilibrium analysis of the same samples. s
(bottom panel) show absorbance at 280 nm.
The solid line indicates best fit to a self-asso-
ciating system. The upper panel shows the
residuals for the best fit.
Ó FEBS 2003 d-crystallin in guanidinium chloride (Eur. J. Biochem. 270) 3991
dimer (Fig. 3B). The dissociated monomers possessed
significant amount of secondary and tertiary structure as
judged by the CD and fluorescence changes (Fig. 5A).
Under the same conditions, the M
r
of the major
component was further analyzed by equilibrium sedimen-
tation. As d
2
-crystallin is apparently polydispersed in a 1
M
GdnHCl solution, a monomer–tetramer self-association
model was adopted in data analysis [25]. An equilibrium
constant (K
d
)of0.34l
M
3

was obtained which predicted that
76% of the protein existed as monomers, consistent with the
results obtained from gel-filtration chromatography ana-
lyses (Fig. 4A). Analysis of behavior at different protein
concentrations allowed the apparent molecular weight to be
estimated as 59 490 ± 210 Da by extrapolation. The value
averaged from two rotation speeds has about a 15%
discrepancy compared to the calculated molecular weight of
monomeric d
2
-crystallin.
Multistep unfolding of monomeric d
2
-crystallin
Further increases in GdnHCl concentration induced unfold-
ing of the dissociated monomers. This unfolding process
was investigated by quenching of intrinsic protein
fluorescence by KI and acrylamide. The solvent accessibility
of tryptophan residues in protein was subject to the
conformational fluctuations. KI is ionic in nature and can
selectively quench exposed tryptophan residues, while the
nonionic quencher acrylamide can nonselectively quench
both buried and exposed tryptophan residues.
The GdnHCl concentration-dependent changes of the
dynamic quenching constant were apparently a multistate
process as it did not conform to the smooth two-state
transition observed for simple protein molecules (Fig. 6).
The dynamic quenching constants for the individual
quencher interacting with d
2

-crystallin in native and 6
M
GdnHCl solution were 1.8 ± 0.1 and 9.4 ± 0.3 for
acrylamide, and 0.4 ± 0.02 and 4.1 ± 0.07 for KI,
respectively. The f
a
value for KI was only 12% implying
the low solvent accessibility of tryptophan residues of
d
2
-crystallin in the native form.
ANS binding
ANS is a sensitive probe commonly utilized to detect
conformational changes of proteins especially in the molten
globule state [26]. Binding of ANS to the hydrophobic
surface of proteins will lead to fluorescence enhancement as
well as a blue shift in its emission maximum. Native
d
2
-crystallin binds to ANS with maximum emission at
480 nm and slightly enhanced fluorescence intensity
(Fig. 5A). Denaturation of d
2
-crystallin results in complete
loss of ANS binding. The fluorescence intensity at 470 nm
increases sharply at GdnHCl concentrations exceeding
0.5
M
and reaches a maximum at around 1
M

.The
fluorescence intensity gradually decreased as GdnHCl
concentrations were further increased. The unfolding curve
using the ANS fluorescence probe suggested a multistate
process, consistent with the results obtained by protein
intrinsic fluorescence and CD. The binding affinity (K
d
)
of ANS to d
2
-crystallin in 0, 0.84
M
and 2
M
GdnHCl
solution was determined to be 53 ± 3.1, 63 ± 3.6 and
72 ± 3.5 l
M
, respectively.
Discussion
We have previously proposed a minimum model for the
dissociation/unfolding of the tetrameric duck d
2
-crystallin
in the presence of GdnHCl. The dissociation step involves
dissociation of tetramer to monomers. However, our
present results demonstrate the formation of dimers
during the dissociation step (Fig. 1). These results are in
agreement with the double dimer structure of the tetra-
meric d

2
-crystallin. The crystal structure of the protein
indicated that dimers were bound together by the
interaction of three helices, one of which is involved in
tetramer formation [10].
The close contact interface between two tightly bound
dimers of duck d
2
-crystallin structure is clearly shown by the
surface model (Fig. 6). The two subunits are structurally
complementary to each other. Area ÔaÕ of subunit A has a
perfectly structural complementarily with area ÔbÕ from
subunit B in a head-to-tail manner (high light in circle)
(Fig. 6B). The two b-sheets (label c) at subunit A and C
stack together in a face-to-face manner forming another
type of subunit association (Fig. 6C). This association is not
Fig. 5. Equilibrium unfolding of duck d-crystallin in GdnHCl. (A)
Unfolding monitored by changes in tryptophan fluorescence at
340 nm (d), CD (s), or ANS fluorescence at 470 nm (m). All
experiments were measured at 25 °Cat0.24 l
M
protein concentration.
(B) The dynamic quenching constant (K
SV
) and the fractional maxi-
mum accessible protein fluorescence (f
a
) verses GdnHCl. The intrinsic
fluorescence of duck d-crystallin (0.24 l
M

) in GdnHCl was quenched
by acrylamide (closed symbols) and KI (open symbols).
3992 H J. Lee et al. (Eur. J. Biochem. 270) Ó FEBS 2003
as close and may be weaker than that shown in Fig. 6B. The
structural and biochemical data are compatible with a T-D-
M model for the dissociation process. Dynamic exposure of
tryptophan residues in duck d
2
-crystallin revealed conform-
ational changes at different concentrations of GdnHCl
(Fig. 5B). The results implied a more complicated process
for further unfolding of the dissociated monomer.
Substantial tertiary structural changes occur in the
presence of 1
M
GdnHCl but secondary structural changes
are minimal. The drastic increase of hydrophobic patches
strongly suggests that there is a molten globule intermediate
[26]. This may be the reason why an off-pathway inter-
mediate with an exposed hydrophobic core was detected
under the present conditions (Fig. 6). However, disturbance
Fig. 6. Diagram of the surface of duck d-crystallin (1AUWÆpdb). (A) Each subunit of the tetrameric d-crystallin shows a surface with several convex
and concave areas as indicated by ÔaÕ, ÔbÕ and ÔcÕ, which participate in subunit association. Two neighboring subunits (A and B) associated in a tail-to-
head manner to form a compact and tightly bound dimer (B). The top region of the structure (C) participates in another major region of subunit
interaction (A and C). This figure was produced using
VIEWERPRO
( />Ó FEBS 2003 d-crystallin in guanidinium chloride (Eur. J. Biochem. 270) 3993
of electrostatic interactions by GdnHCl may contribute to
the promotion of hydrophobic interactions [27]. Incorrect
subunit interactions giving rise to aggregate formation

appears to be a competitive kinetic process with respect to
stable monomer [28]. More aggregates accumulate at higher
protein concentration (Fig. 3). Soluble polymerization is a
reversible process but aggregation at high protein concen-
tration is irreversible. At 5
M
GdnHCl the aggregates are
dissolved, presumably due to the stronger ionic salt effect,
which disrupts interactions between protein molecules and
the polypeptide is in an unfolded state.
Chakraborty et al. [29] have reported similar GdnHCl
concentration-dependent reversible unfolding of recombin-
ant duck d
2
-crystallin based on CD experiments, and
proposed a tetramer-dimer-unfolded monomer mechanism.
The total free energy change of the process was as high as
64.5 kcalÆmol
)1
. We have investigated this process with
multiple biophysical probes. An elaborate model has been
derived (Scheme 1) based on the following observation:
First, our results revealed that the fluorescence signal was
more sensitive probe than CD for monitoring the dissoci-
ation transition. Dissociation reflected a 50% intensity
decrease in tryptophan fluorescence and an 80% enhance-
ment in ANS fluorescence compared to only a 10%
decrease in ellipticity (Fig. 5). A CD probe for oligomeric
protein unfolding may be relatively insensitive and miss
some important information.

The shoulder at 0.6
M
GdnHCl in the ANS fluorescence
trace (Fig. 5A) and M
r
observed with sedimentation
velocity at 0.8
M
GdnHCl could represent a dimeric
intermediate (Fig. 3B). Thus, the proposed dimeric form
for human argininosuccinate lyase at % 2.5
M
GdnHCl
could be an artifact. Our analytical ultracentrifugation
results indicated that d
2
-crystallin existed as monomer at
this concentration of GdnHCl (Fig. 4). It is unlikely that
recombinant duck d
2
-crystallin has a grossly different
behavior from protein isolated from duck eyes. We believe
that the mechanism shown in Scheme 1 more accurately
describes the dissociation/unfolding process of d
2
-crystallin.
Another possible origin of the discrepancy between the
results of Sampaleanu et al. [30] may be the denaturation
time used in these protocols. We have performed time-
dependent denaturation of d

2
-crystallin experiments and
confirmed that the dimers separated upon mixing protein
with GdnHCl and their abundance decrease as incubation
time increased. Monomer was the predominant species at
equilibrium (Figs 1 and 2). Sufficient incubation time is
therefore essential for allowing the denaturation reaction to
reach equilibrium [31].
Acknowledgements
We thank Matthew D. Lloyd (University of Bath) for reading this
manuscript before publication and Yu-Chin Pon for technical
assistance. This work was supported financially by the National
Science Council, Republic of China.
References
1. Wistow, G.J. & Piatigorsky, J. (1988) Lens crystallins: the evolu-
tion and expression of proteins for a highly specialized tissue.
Annu. Rev. Biochem. 57, 479–504.
2. De Jong, W.W., Hendriks, W., Mulders, J.W.M. & Bloemendal,
H. (1989) Evolution of eye lens crystallins: the stress connection.
Trends Biochem. Sci. 14, 365–368.
3. Wistow, G.J. (1993) Lens crystallins: gene recruitment and evo-
lutionary dynamism. Trends Biochem. Sci. 18, 301–306.
4. Slingsby, C., Simpson, A., Ferszt, A., Bateman, O.A. & Nalini, V.
(1991) Molecular interactions in the eye lens. Biochem. Soc. Trans.
19, 853–858.
5. Lee, H.J., Chiou, S.H. & Chang, G.G. (1992) Biochemical char-
acterization and kinetic analysis of duck delta-crystallin with
endogenous argininosuccinate lyase activity. Biochem. J. 283,
597–603.
6. Lee, H.J., Lin, C.C., Chiou, S.H. & Chang, G.G. (1994) Char-

acterization of the multiple forms of duck lens d-crystallin with
endogenous argininosuccinate lyase activity. Arch. Biochem.
Biophys. 314, 31–38.
7. Piatigorsky, J. & Horwitz, J. (1996) Characterization and enzyme
activity of argininosuccinate lyase/delta-crystallin of the embryo-
nic duck lens. Biochim. Biophys. Acta 1295, 158–164.
8. Simpson, A., Bateman, O., Driessen, H., Lindley, P., Moss, D.,
Mylvaganam, S., Narebor, E. & Slingsby, C. (1994) The structure
of avian eye lens d-crystallin reveals a new fold for a superfamily of
oligomeric enzymes. Nat. Struct. Biol. 1, 724–734.
9. Simpson, A., Moss, D. & Slingsby, C. (1995) The avian eye lens
protein d-crystallin shows a novel packing arrangement of tetra-
mers in a supramolecular helix. Structure 3, 403–412.
10. Abu-Abed, M., Turner, M.A., Vallee, F., Simpson, A., Slingsby,
C. & Howell, P.L. (1997) Structural comparison of the enzy-
matically active and inactive forms of delta crystallin and the role
of histidine 91. Biochemistry 36, 14012–14022.
11. Vallee, F., Turner, M.A., Lindley, P.L. & Howell, P.L. (1999)
Crystal structure of an inactive duck delta II crystallin mutant
with bound argininosuccinate. Biochemistry 38, 2425–2434.
12. Radford, S.E. (2000) Protein folding: progress made and promised
ahead. Trends Biochem. Sci. 25, 611–618.
13. Clark, A.C., Sinclair, J.F. & Baldwin, T.O. (1993) Folding of
bacterial luciferase involves a non-native heterodimeric inter-
mediate in equilibrium with the native enzyme and the unfolded
subunits. J. Biol. Chem. 268, 10773–10779.
14. Gokhale, R.S., Agarwalla, S., Santi, D.V. & Balaram, P. (1996)
Covalent reinforcement of a fragile region in the dimeric enzyme
thymidylate synthase stabilizes the protein against chaotrope-
induced unfolding. Biochemistry 35, 7150–7158.

15. Lee, H.J. & Chang, G.G. (2000) Guanidinium chloride induced
reversible dissociation and denaturation of duck delta2-crystallin.
Eur. J. Biochem. 267, 3979–3985.
16. Higurashi, T., Nosaka, K., Mizobata, T., Nagai, J. & Kawata, Y.
(1999) Unfolding and refolding of Escherichia coli chaperonin
GroES is expressed by a three-state model. J. Mol. Biol. 291,
703–713.
17. Kishimoto, K., Yasuda, C. & Manning, J.M. (2000) Reversible
dissociation/association of amino acid transaminase subunits:
properties of isolated active dimers and inactive monomers.
Biochemistry 39, 381–387.
18. Hornby, J.A.T., Luo, J.K., Stevens, J.M., Wallace, L.A., Kaplan,
W., Armstrong, R.N. & Dirr, H.W. (2000) Equilibrium folding of
dimeric class l glutathione transferases involves a stable mono-
meric intermediate. Biochemistry 39, 12336–12344.
19. Park, Y.C. & Bedouelle, H. (1998) Dimeric tyrosyl-tRNA syn-
thetase from Bacillus stearothemophilus unfolds through a
monomeric intermediate. J. Biol. Chem. 273, 18052–18059.
T
4M
4I
4U
(Scheme I)
2D
Scheme 1. Reversible dissociation and unfolding mechanism of duck d
2
-
crystallin in GdnHCl. T, tetramer; D, dimer; M, molten globule
monomer; I, partially folded intermediate; U, unfolded monomer.
3994 H J. Lee et al. (Eur. J. Biochem. 270) Ó FEBS 2003

20. Chang, G.G., Lee, H.J. & Chou, R.H. (1997) pH-induced
reversible dissociation of tetrameric duck lens delta-crystallin.
Exp. Eye Res. 65, 653–659.
21. Lehrer, S.S. & Leavis, P.C. (1978) Solute quenching of protein
fluorescence. Methods Enzymol. 49, 222–236.
22. Schuck, P. (2000) Size-distribution analysis of macromolecules by
sedimentation velocity ultracentrifugation and lamm equation
modeling. Biophys. J. 78, 1606–1619.
23. Lee, J.C. & Timasheff, S.N. (1974) The calculation of partial
specific volumes of proteins in guanidine hydrochloride. Arch.
Biochem. Biophys. 165, 268–273.
24. Kawahara, K. & Tanford, C. (1966) Viscosity and density of
aqueous solutions of urea and guanidine hydrochloride. J. Biol.
Chem. 241, 3228–3232.
25. MacRorie, D.K. & Voelker, P.J. (1993) Self-Associating Systems in
the Analytical Ultracentrifuge, Beckman Instruments, California.
26. Chamberlain, A.K. & Marqusee, S. (2000) Comparison of equi-
librium and kinetic approaches for determining protein folding
mechanisms. Adv. Prot. Chem. 53, 283–327.
27. Monera, O.D., Kay, C.M. & Hodges, R.S. (1994) Protein
denaturation with guanidine hydrochloride or urea provides a
different estimate of stability depending on the contributions of
electrostatic interactions. Protein Sci. 3, 1984–1991.
28. Jaenicke, R. & Lilie, H. (2000) Folding and association of
oligomeric and multimeric proteins. Adv. Prot. Chem. 53, 329–401.
29. Chakraborty, A.R., Davidson, A. & Howell, P.L. (1999)
Mutational analysis of amino acid residues involved in arginino-
succinate lyase activity in duck dII-crystallin. Biochemistry 38,
2435–2443.
30. Sampaleanu, L.M., Davidson, A.R., Graham, C., Wistow, G.J. &

Howell, P.L. (1999) Domain exchange experiments in duck
delta-crystallins: functional and evolutionary implications. Protein
Sci. 8, 529–537.
31. Pace, C.K. (1986) Determination and analysis of urea and
guanidine hydrochloride denaturation curve. Methods Enzymol.
131, 266–280.
Ó FEBS 2003 d-crystallin in guanidinium chloride (Eur. J. Biochem. 270) 3995