We found that Bet v 4 and Phl p 7 undergo oligomerization transitions with characteristics that are markedly different from those typically found in proteins: transitions from monomers t
Trang 1by EF-hand rearrangements and domain swapping
Iris Magler1, Dorota Nu¨ss1, Michael Hauser2, Fatima Ferreira2and Hans Brandstetter1
1 Division of Structural Biology, Department of Molecular Biology, University of Salzburg, Austria
2 Division of Allergy, Department of Molecular Biology, University of Salzburg, Austria
Introduction
Allergy is a health problem that is growing at an
almost epidemic rate, with approximately 20% of the
population being affected by type I allergy worldwide
[1–6] Allergies appear in many versions, including
pol-len and food allergies, and mite dust and
environmen-tally caused allergies Pollen allergens represent the
largest subgroup, and can be classified into 29 protein
families; most of them belong to the expansin, profilin
or calcium-binding protein families [7]
Massive efforts have been directed at elucidating the characteristics and causative mechanisms underlying the action of allergens Among the biophysical proper-ties shared by allergens with the ability to breach phys-ical defense mechanisms in a susceptible host are: (a) small size, typically ranging from 5 to 30 kDa; (b) high effective concentration, implying high solubility and stability; and (c) foreignness to the affected host [8] Additionally, allergens elicit an IgE response and a
Keywords
covalently locked conformation; EF-hand
protein; protein engineering; structure;
temperature-dependent oligomerization
Correspondence
H Brandstetter, Billrothstr 11, 5020
Salzburg, Austria
Fax: +43 662 8044 7209
Tel: +43 662 8044 7270
E-mail: hans.brandstetter@sbg.ac.at
(Received 5 February 2010, revised 17
March 2010, accepted 7 April 2010)
doi:10.1111/j.1742-4658.2010.07671.x
Polcalcins such as Bet v 4 and Phl p 7 are pollen allergens that are con-structed from EF-hand motifs, which are very common and well character-ized helix–loop–helix motifs with calcium-binding functions, as elementary building blocks Being members of an exceptionally well-characterized protein superfamily, these allergens highlight the fundamental challenge in explaining what features distinguish allergens from nonallergenic proteins
We found that Bet v 4 and Phl p 7 undergo oligomerization transitions with characteristics that are markedly different from those typically found
in proteins: transitions from monomers to dimers and to distinct higher oligomers can be induced by increasing temperature; similarly, low concen-trations of destabilizing agents, e.g SDS, induce oligomerization transitions
of Bet v 4 The changes in the quaternary structure, termed molecular metamorphosis, are induced and controlled by a combination of EF-hand rearrangements and domain swapping rather than by the classical law of mass action Using an EF-hand-pairing model, we provide a two-step model that consistently explains and substantiates the observed metamor-phosis Moreover, the unusual oligomerization behavior suggests a straight-forward explanation of how allergens can accomplish the crosslinking of IgE on mast cells, a hallmark of allergens
Structured digital abstract
l MINT-7718612 : Bet v 4 (uniprotkb: Q39419 ) and Bet v 4 (uniprotkb: Q39419 ) bind ( MI:0407 )
by molecular sieving ( MI:0071 )
l MINT-7718648 : Phl p 7 (uniprotkb: O82040 ) and Phl p 7 (uniprotkb: O82040 ) bind ( MI:0407 )
by molecular sieving ( MI:0071 )
Abbreviations
GFP, green fluorescent protein; TEV, tobacco etch virus.
Trang 2clinical response, which may represent an immediate
and⁄ or late-phase response [9] Many allergens show
proteolytic activity, for which, in selected cases, a
cau-sal connection has been demonstrated [10–12];
addi-tionally, surface-exposed hydrophobic patches have
been suggested to provide allergen-typical danger
sig-nals that are recognizable to the innate immune system
[13,14]; similarly, glycosylation patterns present on
allergen surfaces are believed to be involved in
recogni-tion and endocytotic internalizarecogni-tion by innate immune
cells [15] For a recent review, see [16] The increased
biological knowledge is accompanied by an enormous
increase in the available structural database on
aller-gens, accomplished by crystallography and NMR
projects [17–32] Despite this progress, our mechanistic
understanding of the molecular principles of
allergen-icity remains unsatisfactory This is highlighted by the
fact that we are unable to predict the allergenic
behav-ior of a protein on the basis of biophysical properties,
such as its crystal structure [33,34] We lack reliable
structural motifs that could serve as hallmarks of
aller-genicity – such as a catalytic triad and an oxyanion
hole that could identify proteolytic activity
The investigations in the current study were aimed
at the identification of a biophysical hallmark that
could distinguish allergens from other proteins and
could ultimately reveal a causative mechanism acting
in a subfamily of allergens To this end, we
investi-gated the hypothesis that the ability to undergo
con-formational changes represents a distinguishing feature
of allergens The concept of molecular metamorphosis
is receiving increasing attention [35,36] We have
iden-tified and characterized this unexpected molecular
metamorphosis in the pollen allergens Bet v 4 from the
white birch and Phl p 7 from timothy grass These allergens are built from EF-hand motifs, which are exceptionally well-studied building blocks [37,38] We have identified physicochemical parameters that con-trol the oligomerization transitions, and provide a model relating the oligomerization to the ability of the allergens to crosslink already synthesized IgE antibod-ies on mast cells
Results
Bet v 4 can be expressed in a soluble, SDS-stable dimeric form
Bet v 4 and the related Phl p 7 were expressed in Escherichia coliBL21(DE3) cells Typically, SDS⁄ PAGE analysis of intact cells indicated the expression
of monomeric proteins with approximate sizes of 12.5 kDa and 11.7 kDa, as shown for Bet v 4 in
Fig 1A and Phl p 7 in Fig 1C, respectively Purifica-tion to almost homogeneity was achieved in a single step by employing immobilized metal affinity chroma-tography (Fig 1B,C)
Under standard storage conditions, both Bet v 4 and Phl p 7 were also monomeric under native condi-tions, as judged by gel filtration chromatography (Fig 3A)
Surprisingly, we observed spontaneous dimerization
of Bet v 4 with a size of 25 kDa by SDS⁄ PAGE (Fig 2A) Although we repeated the expression of dimeric Bet v 4 more than 10 times, the underlying mechanism of dimerization is partly statistical in nat-ure, because we observed dimerization in 1–2% of the expression trials only However, when dimerization
72
55
43
72 55 43
72 55 43
34
26
34 26
17 17
10
17
Fig 1 Bet v 4 and Phl p 7 protein samples appear exclusively as monomers on SDS⁄ PAGE when expressed and purified under standard conditions (A) Expression of Bet v 4 under standard conditions Lane 1: protein standard (Fermentas) Lane 2: sample before induction Lanes 3–8: Samples 4 h after induction All samples were drawn from different expression flasks (B) Bet v 4 purification by affinity chroma-tography Lane 1: protein standard Lane 2: Bet v 4 cell lysate Lane 3: flow-through Lane 4: wash step Lanes 5 and 6: eluted protein with-out impurities (C) Expression and purification of Phl p 7 Lane 1: protein standard Lane 2: sample before induction Lanes 3–6: samples from different expression flasks 4 h after induction Lane 7: unbound protein impurities (flow-through) after Ni 2+ –nitrilotriacetic acid treat-ment Lane 8: wash fraction Lanes 9 and 10: purified Phl p 7 protein.
Trang 3was observed at all, it was apparently 100% complete.
The statistical nature of the dimerization is puzzling,
and cannot be explained by obvious factors, e.g the
presence or absence of metal factors such as Ca2+ or
EDTA, as described in more detail below
To exclude the possibility of artefacts and to confirm
the identity of the protein, we cleaved the N-terminal
His6-tag by utilizing the tobacco etch virus (TEV)
protease cleavage site Figure 2B shows that, upon
removal of the N-terminal His6-tag plus linker
( 3 kDa), the migration of the protein on SDS ⁄ PAGE
corresponds to a molecular mass reduced by
approxi-mately 6 kDa, as expected As a consequence, we can
conclude that the dimer contact is not mediated by, but
is independent from, the N-terminus The identity of the
Bet v 4 protein was unambiguously confirmed by
ESI-MS The dimerization is reversible, because Bet v 4
monomers were observed by SDS⁄ PAGE after several
weeks of storage at 4C This finding, in particular,
shows that dimerization can take place at 37 C
Spontaneous in vitro dimerization of Bet v 4
When Bet v 4 was expressed as a monomer, it
remained in the monomeric state when stored at 4C
or 20C (Fig 3A) By serendipity, we identified
spon-taneous dimerization of a Bet v 4 sample that was left
on the bench in the summer for weeks, as analyzed
by gel filtration chromatography These findings
prompted us to systematically investigate possible
mechanisms that govern the unexpected and intriguing
oligomerization behavior of Bet v 4 Given the storage
at elevated temperatures over a very long time period,
we hypothesized that temperature and incubation time may affect the oligomerization behavior
Distinct oligomerization transitions in Bet v 4 and Phl p 7 can be induced by temperature changes
We systematically studied the temperature dependence
of the oligomerization state of Bet v 4 under native conditions by using gel filtration chromatography Oligo-merization was observed at 30 C, but only over time intervals of several months These long incubation times effectively excluded the option to conduct sys-tematic experiments at 30 C However, when heated
to 75 C, a mixture of monomeric and dimeric proteins appeared quite rapidly (Fig 3B) When the tempera-ture was further increased to 95C, the dimeric form
of the protein was observed exclusively (Fig 3C) Con-sequently, temperature is one key parameter that induces Bet v 4 oligomerization transitions in vitro Naturally, the question arises of whether the structure
of Bet v 4 remains intact at high temperatures; we con-firmed the structural integrity by performing overnight
CD measurements at 75C, as detailed below
Oligomerization depends on incubation time The oligomerization transitions did not occur instanta-neously, but required some incubation time To quan-tify the required time scale, we analyzed protein oligomerization after distinct incubation times
We found approximately 75% of the protein to be monomeric after 24 h at 75C and the rest of the protein to be dimeric, whereas the situation was
43 55
72 72
55 43 34
26 34
17
26
17
11 11
Fig 2 Spontaneous and complete dimerization of Bet v 4 can be observed during protein expression (A) Expression of the SDS-stable dimer of Bet v 4 in E coli BL21(DE3) cells Lane 1: mass standard Lane 2: cells before induction Lanes 3–8: samples 4 h after induction Note that the samples in the different lanes were drawn from different expression flasks and showed complete dimerization in each case (B) Cleavage of the N-terminal His 6 -tag Following the mass standard (lane 1), the His 6 -tagged Bet v 4 is shown before addition of the TEV protease (lane 2) The protein migrates at an apparent size of 25 kDa ( 2 · 12.5 kDa) Lanes 3–7: TEV-digested Bet v 4 at different time points; TEV protease is visible at 30 kDa Lane 3: TEV digest at time zero Lane 4: digest after 1 h Lane 5: digest after 6 h Lane 6: digest after 12 h Lane 7: digest after 24 h TEV protease cleavage releases the N-terminal His 6 -tag and thus shifts the size of the protein to about
19 kDa ( 2 · 9.5 kDa).
Trang 4reversed after 48 h (Fig 3D–F) After 72 h, all of the
monomeric protein was converted to higher oligomers
The process of oligomerization does not stop with
dimer formation Instead, higher oligomeric forms
were observed by size exclusion chromatography after
72 h of incubation (Fig 3F)
As mentioned earlier, we found dimerization of
Bet v 4 that was stored at room temperature for
approximately 4–10 weeks, whereas Bet v 4 stored at
4C was always found to be monomeric Significantly,
the observed oligomerization does not correspond to
an unspecific aggregation phenomenon, but reflects a
reversible transition between distinct oligomerization
states; in particular, monomer formation can be
induced by lowering the temperature to 4C within
few days Therefore, we conclude that Bet v 4
oligo-merization depends on both incubation temperature
and time in a multiplicative manner
SDS induces instantaneous monomer-to-dimer
transitions in Bet v 4 at room temperature
In contrast to monomer-to-dimer transitions in vitro,
dimerization in E coli cells can take place rapidly,
within a few hours (Fig 2A) We hypothesized that
dimerization can be efficiently catalyzed by compounds
that are presumably present in E coli in trace
amounts Therefore, we systematically screened a vari-ety of chemicals, including Ca2+and other metals, for their effect on oligomerization, both in expression con-ditions and with purified protein
Ca2+ is known to affect the 3D structure of the Bet v 4 monomer [32] Interestingly, addition of neither
10 mm Ca2+ nor EDTA had a direct effect on the dimerization behavior, as judged by SDS⁄ PAGE and gel filtration chromatography, which gave results iden-tical to those shown in Fig 3 These findings were further corroborated by CD measurements, as described below Surprisingly, we found that SDS led
to partial dimer formation in Bet v 4 at 20 and 4C The addition of 0.05% SDS led to equal amounts of the monomeric and dimeric states, as reflected by two prominent peaks at approximately 13 and 11.4 mL (Fig 4A) To a lesser extent ( 10%), a highly oligo-meric species was observed at an elution volume of
8 mL, represented by a broad peak At an SDS con-centration of 0.5%, nearly all of the protein aggregates and only approximately 10% of the protein remained
in the monomeric or dimeric state, as shown by the dashed line in Fig 4A
The bimodal oligomerization behavior of Bet v 4 contrasts with what is seen for most soluble proteins,
as confirmed by a control experiment with green fluo-rescent protein (GFP) Whereas native GFP migrates
10.0
20.0
30.0
40.0
50.0
60.0
mAU
0.0 5.0 10.0 15.0 20.0 ml
12.15 13.56
5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 mAU
0.0 5.0 10.0 15.0 20.0 25.0 30.0 ml
11.95
13.56
10.0 20.0 30.0 40.0 mAU
9.19 11.63
0.0
20.0
40.0
60.0
80.0
mAU
0.0 5.0 10.0 15.0 20.0 25.0 ml
13.75
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 mAU
11.97
13.58
5.0 10.0 15.0 20.0 25.0 30.0 35.0 mAU
0.0 5.0 10.0 15.0 20.0 ml
11.78
4 °C
A
Fig 3 Temperature and time affect the dimerization of Bet v 4 (A–C) The temperature dependence of Bet v 4 oligomerization was analyzed
by gel filtration chromatography Bet v 4 samples were incubated for 48 h at (A) 20 C, (B) 75 C, and (C) 95 C The experiments showed a monomer-to-dimer transition as a function of incubation temperature Further details are given in Experimental procedures (D–F) The time dependence of Bet v 4 oligomerization as analyzed by gel filtration chromatography Bet v 4 was incubated at 75 C and analyzed every 24 h for 3 days.
Trang 5exclusively as a monomer on gel filtration (Fig 4B,
continuous line), addition of 0.05% SDS induced
aggregation of 10% of the protein (Fig 4B, dashed
line) In summary, 0.05% SDS generates an
equilib-rium between native and aggregated GFP, resembling
the situation for Bet v 4, with the exception that
Bet v 4 can be monomeric or dimeric in its native
state
Bet v 4 can be conformationally locked in the
monomeric state
The oligomerization properties of Bet v 4 revealed
unexpected and unique features, such as its dependence
on temperature and chemicals Moreover, the
dimer-ization is apparently independent of protein
concentra-tion in the range from 0.1 to 25 mgÆmL)1 These
unique properties suggested that, in Bet v 4,
oligomeri-zation could involve not only intermolecular
recogni-tion events, governed by the law of mass acrecogni-tion, but
also intramolecular conformational rearrangements
To investigate this hypothesis, we constructed a
Bet v 4 variant containing a K25C⁄ F60C double
muta-tion On the basis of the NMR structure of monomeric
Bet v 4 [32], we devised these point mutations to form
an intramolecular disulfide bond that stabilizes the
conformation by covalently linking both EF-hand
motifs in Bet v 4 (Fig 5)
This covalent linkage is absent in the presence of
dithiothreitol If an intramolecular rearrangement does
indeed accompany the oligomerization of Bet v 4, the
oligomerization behavior of oxidized (disulfide-linked)
Bet v 4-K25C⁄ F60C should deviate markedly from
that of the wild type By contrast, reduced Bet
v 4-K25C⁄ F60C should show oligomerization behavior
identical to that of the wild type
We carried out experiments to test both the tem-perature and time dependence of the oligomerization
by incubating the disulfide-linked Bet v 4 double mutant at 20C for 7 days and at 75 C for 24 h Under both conditions, the monomer was stable over the observation period, as monitored by gel filtration (Fig 6A)
As a control experiment, we carried out similar experiments under reducing conditions using 5 mm dithiothreitol The reduced Bet v 4 double mutant was incubated at 75C for 24 h, and subsequently ana-lyzed by gel filtration The reduced protein revealed the native-like induction of higher oligomer formation (Fig 6B, continuous line), clearly contrasting with the behavior of the oxidized double mutant (Fig 6B, broken line)
0.0 20 40 60 80 mAU
0.0 2.0 4.0 6.0 8.0 10.0 12.0 ml
8.01
11.41 12.77 7.50
11.4012.74
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 mAU
5.0 6.0 7.0 8.0 9.0 10.0 11.0 ml
8.71
10.67 10.70
Fig 4 SDS induces dimerization in Bet v 4 (A) Solid line: at 0.05% SDS and 20 C, most of the Bet v 4 elutes at retention volumes corre-sponding to monomers and dimers, with only a small ( 10%) aggregated fraction eluting near the void volume Dashed line: at 0.5% SDS, most (90%) of the Bet v 4 aggregates (eluting at the void volume: 7.5 mL), and only 10% elutes at volumes corresponding to monomers and dimers (B) Solid line: Control experiment using GFP at 0.05% SDS and 20 C reveals a predominantly native monomeric form, corre-sponding to a retention volume of 10.67 mL, and a small ( 10%) aggregated fraction eluting near the void volume (retention volume of 8.71 mL) Dashed line: at 0% SDS, GFP migrates exclusively as a monomer.
C25 C60
NH 2 COOH
Fig 5 Engineering of a disulfide bridge intended to lock the mono-mer conformation of Bet v 4 The introduced K25C ⁄ F60C double mutation promotes disulfide bond formation between the two anti-parallel b-strands, and thus crosslinks the first Ca 2+ -binding EF-hand (shown in red) with the second EF-hand (blue).
Trang 6Consequently, the formation of Bet v 4 dimers and
higher molecular mass oligomers does indeed involve
an intramolecular conformational rearrangement
Temperature-induced oligomerization may be
universally conserved in polcalcins
Next, we tested whether the intriguing
temperature-dependent and time-temperature-dependent oligomerization is
spe-cific to Bet v 4 or could also be found in structurally
related proteins We selected Phl p 7 as a further
repre-sentative of Ca2+-binding EF-hand proteins, and
car-ried out oligomerization analyses analogous to those
described for Bet v 4 We found that Phl p 7 is indeed
able to undergo temperature-dependent
oligomeriza-tion: At 4C, Phl p 7 formed monomers exclusively,
whereas significant amounts of dimeric Phl p 7
accumu-lated after overnight incubation at 75C Overnight
incubation at 95C completely converted the
mono-meric form of Phl p 7 to higher oligomer forms
(Fig 7) The temperature dependence of the
oligomeri-zation state of Phl p 7 thus parallels the behavior
observed with Bet v 4 The broadness of the 95 C peak may be partly related to heat-induced denaturation
The secondary structure content is independent
of the oligomerization state and is conserved at
75C
We employed CD spectroscopy to investigate whether the secondary structure of Bet v 4 at room temperature was dependent on its oligomeric state Furthermore, as
we used heating of Bet v 4 as a tool to speed up the conformational transition, we wished to clarify whether the structure becomes disrupted at 75C Finally, we used an engineered disulfide-containing variant to test the nature of the conformer transforma-tion, which raises the question of how well this mutant resembles the wild type CD is ideally suited for pro-viding answers to these questions
To investigate the first question, we used Bet v 4 stored at 4C, corresponding to the monomeric spe-cies, and Bet v 4 that had been heated to 75C over-night, corresponding to the dimeric species The
0.0 20.0 40.0 60.0 80.0 mAU
6.0 8.0 10.0 12.0 14.0 16.0 18.0 ml
10.66 12.80
0.0 20.0 40.0 60.0 80.0 mAU
6.0 8.0 10.0 12.0 14.0 16.0 18.0 ml
10.66
12.80
Fig 6 Disulfide bond inhibits the monomer-dimer transformation (A) The gel filtration chromatogram of the disulfide-containing (oxidizing) Bet v 4 mutant; the protein was incubated at 4 C (continuous line) and at 75 C (dashed line) for 24 h The chromatogram did not change over an incubation period of up to 7 days, indicating an exclusively monomeric state (B) The gel filtration chromatogram of the reduced Bet v 4 mutant (no disulfide bond); the protein was incubated at 4 C (continous line) and at 75 C (dashed line) for 24 h The heat-treated protein eluted at a retention volume corresponding to the dimer, resembling the wild-type protein in this respect.
20
40
60
80
100
120
mAU
0.0 5.0 10.0 15.0 20.0 25.0 ml
12.62
0 100 200 300 400 500 600 mAU
10.12 12.45
0 100 200 300 400 mAU
10.08
11.50
Fig 7 Oligomerization of Phl p 7 Gel filtration chromatograms indicate the conversion from monomer to higher oligomerization states at
4 C (A), 75 C (B) and 95 C (C) over an incubation period of 24 h, qualitatively resembling the behavior of Bet v 4.
Trang 7oligomerization states were further confirmed by gel
filtration chromatography Both protein samples
yielded CD spectra that revealed well-folded a-helical
proteins (Figs 8A and S1) Therefore, the secondary
structure in the monomeric and dimeric species is
qualitatively identical
Next, we tested the a-helical content of Bet v 4 at
75C at three time points: after 15 min, after 16 h,
and after 20 h The sample was kept at 75C In all
three samples, the a-helical content was preserved, and
qualitatively coincided with that indicated by the
spec-tra measured at 20C (Fig 8B) The double minimum
structure characteristic of a-helices was less
pro-nounced in the heated samples, however Similarly, a
quantitatively reduced mean residual weight ellipticity
indicated increased flexibility in the secondary
struc-ture of the heated samples (Fig S1)
As a third experiment, we tested the
disulfide-engi-neered Bet v 4 mutant (K25C⁄ F60C) The disulfide
mutant stored at 4 C resembled the native monomer
and gave CD spectra qualitatively identical to those of
the unmodified protein, independently of whether the
disulfide bond was formed or reduced (Fig 8C, CC oxi⁄
red 4C) Interestingly, whereas the overall secondary
structure content was also conserved after heating at
75C, there appeared to be significant disorder in these
protein variants (Fig 8C) The reduction in a-helix
content is most prominent in ‘CC oxi 75C’, in which
the monomeric state is enforced by the intact disulfide
bond These findings are paralleled by the quantitative
representation of the ellipticity in Fig S1
Finally, we confirmed that Ca2+is tightly bound by
Bet v 4 and cannot be extracted by the addition of
10 mm EDTA, as demonstrated by the qualitatively unchanged CD spectrum in the presence of EDTA (Fig S2)
Discussion
Bet v 4 forms monomers, dimers, and higher oligomers
We identified several distinct oligomerization states for Bet v 4 Although these findings appear to be in con-flict with those from previous experiments, these dis-agreements may be reconciled by considering the settings of the particular experiments [32] This is of particular relevance for experiments with measurement times of days, such as NMR and ultracentrifugation, which were run at a constant temperature of 20C or
4C, respectively In fact, also in our hands, the pro-tein’s oligomerization behavior over time was tempera-ture-dependent Even when stored for several months
at 4C, Bet v 4 remained in a monomeric conforma-tion If expressed in monomeric form, Bet v 4 remained monomeric over days to weeks at 20 C However, after months, Bet v 4 adopted a dimeric con-formation at room temperature
Temperature is a universal inducer of oligomerization transitions
The consistent observations made with Phl p 7 and Bet v 4 suggest to us that temperature acts as an important order parameter controlling oligomer forma-tion in polcalcins The fact that an increase in
temper-Fig 8 CD measurements document the structural integrity of diverse Bet v 4 species Data are presented as baseline-corrected mean resi-due molar ellipticity [Q] MRW at a given wavelength (A) The spectra of monomeric Bet v 4 protein stored 4 C (continuous line) and after heat-induced dimerization (dashed line) qualitatively coincide, indicating a near-identical secondary structure content (B) Time series of CD spectra of Bet v 4 kept and analyzed at 75 C for 15 min (continuous line), 16 h (dashed line with dots), and 20 h (dashed line) The second-ary structure is mostly conserved, and does not noticeably vsecond-ary over time (C) CD spectra of Bet v 4-K25C ⁄ F60C (CC) with the disulfide bond formed (oxi) or reduced (red), each stored at either 4 C or 75 C (overnight) When stored at 4 C, the (monomeric) CC mutant adopted a native-like ellipticity spectrum, independently of the status of the disulfide bond (oxidized or reduced), indicating a native like 3D structure After heat treatment, the qualitative form of the spectrum remained conserved, albeit with a significantly reduced amplitude.
Trang 8ature induces transitions to high molecular mass
oligo-mers is surprising: with increasing temperature (T), the
entropy (S) of the protein becomes more important for
its Gibbs free energy (G) than the enthalpic
contribu-tion (E): G = E) TS Thus, dissociation of oligomers
should be favored at high temperature, because the
degrees of freedom are maximized for monomers
However, the observed behavior contrasts with these
fundamental physicochemical principles, and points to
the existence of temperature-induced intramolecular
rearrangements in Bet v 4 In other words, although
CD spectra indicate that the secondary structures of
monomers and dimers are alike (Fig 8), their detailed
conformations differ in a subtle way (Fig 10) Only
the excited conformation is able to form dimers; the
ground state conformation is monomeric Importantly,
although, for practical reasons, we performed the
experiments shown in Figs 3, 6 and 7 at
unphysiologi-cally high temperatures, these transitions do also occur
at ambient temperatures Additionally, the
tempera-ture-induced transitions may well be catalyzed by other
components present in the pollen, as discussed below
As an additional cautionary remark, we must point
out that structural integrity could be demonstrated for
temperatures up to 75C only (Fig 8); the sample at
95C may be partially unfolded
Like temperature, chemical substances induce
metamorphosis by stabilizing or destabilizing
local free energy minima
SDS is known to destabilize the quaternary and or
ternary structure of proteins [39,40] This effect is also
observed in Bet v 4 and our control protein, GFP,
leading to a broad peak near the void volume of the
gel filtration column (Fig 4A,B) Significantly,
how-ever, we found that SDS induced specific dimerization
of Bet v 4, as reflected by a sharp elution peak
(Fig 4A)
It is very likely that a number of other physiological
chemicals will affect the oligomerization of Bet v 4 In
fact, sodium chloride at 0.5 m favors the monomeric
state of Bet v 4 over the dimeric state These findings
support the notion that oligomer transformations
are relevant in the physiological environment of the
pollen
This observation can be explained by assuming a
multimodal free energy surface of Bet v 4 with several
distinct substates; in a simplified version, this surface
can be represented by two isothermal free energy
graphs (Fig 9) The ratio of the free energy minima
representing the monomeric and dimeric states changes
with temperature This property reflects the surprising
fact that dimers and higher oligomers are preferred over monomers at high temperature
The differences in free energy can be quantitatively estimated by exploiting the fact that the statistics of oligomer formation are governed by Boltzmann’s law
p¼1
Ze
DG=RT
where the probability p corresponds to the likelihood
of a dimer and is estimated to be 1%, reflecting the frequency of observation of spontaneous dimerization (see Results) Z represents the partition function, which we roughly estimate, from the number of acces-sible states, to be Z = 3 (monomer, dimer, and tetra-mer) R is the gas constant [8.3 JÆ(mol K))1], T is the absolute temperature (300 K), and DG is the change in free energy associated with the monomer-to-dimer transition Reformulation of Boltzmann’s law yields
DG¼ RT ln ðZ 0:01Þ RT ln ð3 0:01Þ 9 kJ
We admit that our estimations of both P and Z are quite inaccurate; however, as both parameters (P, Z) affect DG via logarithmic dependence, errors will trans-late to the resulting DG value only in a dampened manner
In addition to this phenomenological consideration,
we tried to develop a structural model that could pro-vide a mechanism that explains this counterintuitive behavior Such a model is developed below
High temperature Low temperature
Oligomerization State ΔG
Fig 9 Schematic free energy diagram of Bet v 4 governing the occupancy of its conformational substates The free energy depends most prominently on the temperature: at low temperature (i.e 4 C), the monomer is preferred, whereas higher oligomers are preferred at higher temperatures Additionally, the free energy depends on parameters such as ionic strength or SDS, which together result in a more complex hypersurface, as illustrated.
Trang 9Dimerization is a two-step process involving an
excited conformation of Bet v 4
The anomalous dependence of the oligomerization on
temperature and⁄ or chemical substances had already
indicated that changes in the tertiary structure of the
Bet v 4 subunits precede the monomer-to-dimer
transi-tion Proof for this hypothesis was obtained by
engi-neering a disulfide-containing variant that locks the
known 3D structure of the monomeric substate This
protein is unable to undergo dimerization or higher
oligomer formation under oxidizing conditions in
which the disulfide bridge is conserved This model
also explains the different modes of action of
tempera-ture and SDS The latter slightly destabilizes the
monomeric state and effectively lowers the separating
energy barrier, leading to a population of both
mono-mers and dimono-mers (Fig 4A) The effect of temperature
is more sophisticated: although it also helps to
over-come the separating energy barrier, an additional
mechanism is required to explain why dimers are
pre-ferred over monomers at high temperature
Our model involves a two-step process First, elevated
temperatures will induce a conformational transition
within the Bet v 4 subunit from the ground state to an
excited state, whereby the ground state represents a
closed conformation and theexcited state an open
con-formation with no EF-hand pairing Both the ground
state and excited state are monomeric (Fig 5) Two hydrogen bonds in the short antiparallel b-sheet have to
be broken in the excited state, representing an energetic barrier that matches that derived previously from Boltz-mann’s distribution ( 9 kJ) In a second step, this excited form is now able to gain enthalpy by intermolec-ular pairing of the EF-hands, and thus forms dimers (Fig 10A,B) Such an extended conformation has been observed in the crystal structure of Phl p 7 [41]
It should be noted that a straightforward extension
of the EF-hand-pairing model, shown in Fig 10C, explains the existence of higher order oligomers Here,
we assume that an initial dimer is formed by monomer pairing with one EF-hand rather than two This gener-ates a ‘sticky overhang’ dimer that provides two addi-tional EF-hand docking sites These docking sites can attract additional Bet v 4 subunits, and thus provide a mechanism to generate trimers, tetramers, and higher oligomers This model further explains why Bet v 4 proteins migrating as dimers during gel filtration may differ on SDS⁄ PAGE We propose that single EF-hand-paired dimers are less stable and dissociate to form monomers on SDS⁄ PAGE, whereas double EF-hand-paired dimers stay intact on SDS⁄ PAGE The proposed model (Fig 10) has a qualitatively unchanged secondary structure content in the mono-meric and dimono-meric states, consistent with the recorded
CD spectra (Fig 8)
Fig 10 (A, B) EF-hand pairing as a mechanism for dimerization Bet v 4 consists of two EF-hand motifs, EF1 (blue) and EF2 (red), which are connected by a flexible connecting segment (green) A C-terminal helix (a5; gray) presumably contributes to stabilization of the EF-hand pair-ing In the experimentally determined monomer structure, intramolecular EF-hand pairing occurs via strands b1 and b2, forming a central antiparallel b-sheet This structure represents the ground state conformation On the basis of the crystal structure of dimeric Phl p 7 [41],
we propose that dimerization is mediated by intermolecular EF-hand pairing via strands b1 and b2¢ and strands b2 and b1¢ For this dimeriza-tion to occur, we propose the existence of an excited state intermediate (open form) that will be increasingly common at high temperature (C) Alternatively, a singly EF-hand-paired dimer may form, as shown here, via strands b1 and b1¢; possible alternative dimers would involve strands b2 and b2¢, b1 and b2¢, or b2 and b1¢ Singly EF-hand-paired dimers will be less stable than doubly paired dimers.
Trang 10CD measurements further showed that the
second-ary structure of Bet v 4 is mostly conserved during the
overnight heat treatments (Fig 8) This lends support
to the notion that the heating protocol only accelerates
the transformation from monomer to dimer, and does
not change the reaction path of the transformation In
particular, the heat-induced dimerization does not
occur via an unfolding–folding process
Finally, the CD measurements of the
disulfide-containing variant indicate that the Bet v 4 dimer
structure will deviate in subtle details from the
pro-posed Phl p 7 dimer structure (Fig 10)
A remaining puzzle is why Bet v 4 is mostly
expressed as a monomer in E coli, but sometimes as a
dimer; moreover, if Bet v 4 is expressed as dimer, it is
exclusively dimeric It seems quite plausible that the
difference in the observed oligomerization relates to
the presence of a dimerization catalyst We propose
that a chaperone, such as a heat shock protein, could
account for the observed dimerization behavior The
expression rate of heat shock proteins varies drastically
upon subtle and difficult-to-control changes
Molecular metamorphosis provides a framework
to explain the ability of polcalcins to crosslink
IgE antibodies on mast cells
Naturally, the question arises of whether the
intrigu-ing biophysical behavior of Bet v 4 and Phl p 7
relates to their allergenic properties The molecular
metamorphosis model provides a straightforward
explanation for one allergenic key feature, namely,
the ability to crosslink IgEs on mast cells
Addition-ally, it is very plausible that the oligomerization status
of an allergen will affect its endocytosis and
endoso-mal processing Significantly, dimerization or
multi-merization has been reported for numerous allergens
[42–51] Clearly, IgE binding is necessary, but not
suf-ficient, to induce a Th2 immune response, which is
characteristic of allergy In fact, there is conclusive
evidence for selected allergens that the allergenicity,
including antibody-binding capacity, differs for
mono-mers and dimono-mers: Scho¨ll et al have shown, for the
birch pollen allergen Bet v 1, that dimers (34 kDa),
and not monomers (17 kDa), represent the allergenic
Bet v 1 species [51] The same basic mechanism has
been reported by Reese et al for the carrot allergen
Dau c 1 [45] According to the molecular
metamor-phosis hypothesis presented here, we suggest that
con-formationally locked allergens (e.g
disulfide-stabilized) will cause drastically reduced allergic
reac-tions Clearly, the validation of this hypothesis awaits
further experiments
Experimental procedures
Materials Plasmids coding for Bet v 4 and Phl p 7 (Uniprot Database accession numbers are Q39419 for Bet v 4 and O82040 for Phl p 7) were isolated from pollen, as described previously [52,53] Restriction enzymes and T4 ligase were obtained from Fermentas (St Leon-Rot, Germany) Pfu Ultra II Fusion HS DNA polymerase was obtained from Stratagene (La Jolla, CA, USA) Custom-made primers were obtained and sequence analyses were performed at Eurofins MWG Operon (Germany) E coli strain XL1 Blue (Stratagene) was used for subcloning Strain BL21(DE3) (Novagen, Madison, WI, USA) was used as host strain for protein expression For expression, LB-Lennox (Roth, Karlsruhe, Germany) was used All reagents were of the highest stan-dard available from Sigma-Aldrich (Mu¨nchen, Germany)
or AppliChem (Darmstadt, Germany)
Cloning The plasmids were cloned in the pHIS parallel II vector with an NcoI site at the 5¢-end and an EcoRI site at the 3¢-end [54] To engineer the disulfide mutant of Bet v 4,
a double mutation K25C⁄ F60C was constructed by site-directed mutagenesis using the QickChange method [55] The following primers were used: 5¢-gccaatggcgatggtTGCat-ctcAgcagcagag-3¢ [K25C forward primer, bases exchanged are underlined, silent control restriction (PstI) site is in bold]; 5¢-ctctgctgcTgagatGCAaccatcgccattggc-3¢ (K25C reverse primer, bases exchanged are underlined, control restriction site is in bold); 5¢-accgatggcgacggATGCatt-tcgttccaagag-3¢ [F60C forward primer, bases exchanged are underlined, control restriction site (NsiI) is in bold]; and 5¢-ctcttggaacgaaatGCATccgtcgccatcggt-3¢ (F60C reverse primer, bases exchanged are underlined, control restriction site is in bold) The PCR product was digested with the meth-ylation-sensitive enzyme DpnI for 1 h at 37C Products were purified (Qiagen, Hilden, Germany) and transformed into XL1-blue cells by electroporation; cells were plated on
LB agar containing ampicillin Plasmid Mini Preparation (Qiagen) was performed and the obtained plasmids were digested with the appropriate control restriction enzymes (PstI and NsiI) for 2 h at 37C to screen for plasmids with the correct mutation The correctness of the restriction-positive plasmids was finally confirmed by sequencing
Protein expression Plasmids were transformed into E coli strain BL21(DE3) via electroporation, and grown overnight in 100 mL of LB medium containing 100 lgÆmL)1 ampicillin Large-scale expression cultures (12· 600 mL) were inoculated with