The ability of the small Hsp (heat-shock protein) Lo18 from Oenococcus oeni to modulate the membrane fluidity of liposomes or to reduce the thermal aggregation of proteins was studied as a function of the pH in the range 5–9. We have determined by size-exclusion chromatography and analytical ultracentrifugation that Lo18 assembles essentially as a 16-mer at acidic pH. Its quaternary structure evolves to a mixture of lower molecular mass oligomers probably in dynamic equilibrium when the pH increases. The best Lo18 activities are observed at pH 7 when the particle distribution contains a major proportion of dodecamers. At basic pH, particles corresponding to a dimer prevail and are thought to be the building blocks leading to oligomerization of Lo18. At acidic pH, the dimers are organized in a double-ring of stacked octamers to form the 16-mer as shown by the low-resolution structure determined by electron microscopy. Experiments performed with a modified protein (A123S) shown to preferentially form dimers confirm these results. The α-crystallin domain of Methanococcus jannaschii Hsp16.5, taken as a model of the Lo18 counterpart, fits with the electron microscopy envelope of Lo18.

INTRODUCTION

Bacteria use several mechanisms including sHsp [small Hsp (heat-shock protein)] synthesis to cope with environmental stress [1]. All proteins of the sHsp family exhibit a low molecular mass ranging from 12 kDa to 43 kDa and contain an α-crystallin domain of 80–100 residues flanked by a N-terminal region of variable sequence and a short C-terminal region [2,3]. Synthesis of sHsp is up-regulated in response to various signals and more generally in response to protein aggregation. sHsp functions have been investigated both in vivo and in vitro. Their chaperone activity was demonstrated in vivo by evaluating the thermostabilization of cellular proteins in Escherichia coli cells expressing either rice cells Oshsp 16.9 (Oryza sativa Hsp16.9) or the sHsp Lo18 from Oenococcus oeni [4,5]. In some cases, E. coli expressing Oshsp16.9 or sHsp from Pyrococcus furiosus exhibited an enhanced thermotolerance [4,6]. In vitro, sHsps are known to exert a molecular chaperone activity on various protein substrates [7] by interacting with the proteins and preventing aggregation during heat shock [8]. The organization of sHsps in large oligomers is critical for their activity [2,9]. Besides, as reflected by a rapid subunit exchange under physiological conditions to form homo- or hetero-oligomers [10], their quaternary structure is highly dynamic [3,11]. The effect of temperature, pH and protein concentration has been demonstrated to be critical for oligomerization and consequently for chaperone activities [12,13].

Some sHsps were also found to have a stabilizing effect on membrane lipids [14]. For instance, SP21 from Stigmatella aurantiaca could participate to the stabilization of cellular components during stress and to the extensive reconstruction activities accompanying spore formation [15]. In yeast, Hsp12 was found to protect liposomal membrane integrity against desiccation and ethanol-induced stress [16]. The capacity to stabilize membrane has been extensively described for Hsp17 from Synechocystis PCC 6803 on liposome models with various compositions of fatty acids [14,17]. However, whereas oligomerization is a pre-requisite for in vitro chaperone activity preventing aggregation of proteins [3,18], the quaternary structure of sHsp during interaction with membrane lipids is less documented. Nevertheless, it is known that the association of Hsp17 with membranes is dependent on the oligomer dissociation [19].

O. oeni is a lactic acid bacterium mostly responsible for malolactic fermentation in wine [20]. However wine must be considered as a hostile environment for cell growth, characterized by a low pH, the presence of 10–13% ethanol, high sulfite concentration, low temperature and nutrient starvation. O. oeni has evolved adaptation strategies to cope with this harsh environment such as: (i) modifications of membrane composition by changing the proportion of saturated, unsaturated, cyclic or branched fatty acids to adjust membrane fluidity [21] or (ii) synthesis of stress proteins [22,23]. We previously investigated the heat-shock response of O. oeni [22] with a special emphasis on the role of Lo18 [24]. The transcription of the hsp18 structural gene was induced after addition of benzyl alcohol, a membrane fluidizer, and in response to this stress, the corresponding protein Lo18, was found associated to the membrane [25]. The synthesis of Lo18 is induced also during the stationary growth phase [26,27]. Lo18 was demonstrated to protect in vitro both currently used protein models from aggregation and liposomes formed with O. oeni lipids from destabilization [24,25]. We also provided evidence for a role of different amino acids of the α-crystallin domain in such in vivo activities [5].

In the present study, we have assayed the chaperone activity of the wild-type Lo18 and that of Lo18 modified in the α-crystallin domain by the A123S substitution, as a function of the pH. These activities were correlated with the dynamic behaviour of the quaternary structure of Lo18 explored at the different pH. The low-resolution structure of the wild-type protein was determined by electron microscopy at pH 5, a pH at which Lo18 is present as large oligomers that are probably inactive. Finally, we investigated the activity of Lo18 in membrane stabilization as a function of its subunit organization. Our data strongly argue for a dimer as the basic structure for oligomerization and consequently Lo18 activity.

EXPERIMENTAL

Bacterial strains and plasmids

O. oeni ATCC BAA-1163 (formerly known as IOB 8413) was grown at 30°C and at pH 5.3 in FT80 medium modified by the addition of meat extract instead of casamino acids (mFT80 medium) [28]. E. coli BL21 star (DE3) cells transformed with the plasmids pET28a-hsp18 or pET28a-A123S, named respectively E. coli Lo18 and E. coli A123S [5], were grown aerobically at 37°C in LB (Luria–Bertani) broth supplemented with 50 μg·ml−1 kanamycin.

Protein expression and purification

The E. coli strains Lo18 and A123S were used for overexpressing wild-type Lo18 and the modified form A123S, respectively [5]. Cell-free extracts were prepared from a 1 litre culture of E. coli cells grown aerobically at 37°C in LB medium supplemented with kanamycin (50 μg·ml−1). The production of Lo18 or A123S was induced by adding 50 μM IPTG (isopropyl β-D-thiogalactopyranoside) for 3 h at 37°C and shaking. All procedures were then carried out at 4°C. Cells were washed in 20 mM Tris/HCl and 250 mM NaCl (pH 8), concentrated by centrifugation (at 7500 g for 5 min) and disrupted at 1.4 kbar (140000 kPa) (Disruptor Z Plus Series Cell, Constant Systems Ltd). The suspension was centrifuged at 7500 g for 20 min at 4°C to remove unbroken cells. The supernatant was filtered on 0.2 μm-pore-size membrane and loaded on to a HIC-PHE 1 ml column (GE Healthcare) equilibrated with 20 mM Tris/HCl and 250 mM NaCl (pH 8). The protein was eluted with 20 mM Tris/HCl (pH 8) buffer and dialysed against 50 mM sodium phosphate buffer (pH 7) [25]. Purified proteins were concentrated at 3 mg·ml−1 and stored at −20°C in the same buffer.

SEC (size-exclusion chromatography)

The oligomeric state of purified Lo18 and A123S was estimated by SEC on a Superdex 200 10/300 GL column (GE Healthcare). Before each run, the column was equilibrated with 50 mM phosphate buffer at pH 5, 6, 7, 8 or 9. Proteins (300 μg) were injected and separated at a flow rate of 0.5 ml·min−1 at 10°C using an Akta Purifier device (GE Healthcare). Absorbance was recorded at 215 nm and 280 nm and 1 ml fractions were collected and analysed by SDS/PAGE (12% gel). For calibration, ovalbumin [43 kDa, RS (Stokes radius)=30.5 Å (1 Å=0.1 nm)], conalbumin [75 kDa, RS=NA (not applicable)], aldolase (158 kDa, RS=48.1 Å), ferritin (440 kDa, RS=61.0 Å) and thyroglobulin (669 kDa, RS=85.0 Å) were used. Each experiment was performed in triplicate.

AUC (analytical ultracentrifugation)

Sedimentation velocity experiments were performed using a Beckman XL-I analytical ultracentrifuge and an AN-50 TI rotor (Beckman Coulter). The experiments were carried out at 20°C in buffer containing 50 mM sodium phosphate, pH 5, 6, 7, 8 or 9. Protein concentrations were 13 μM, 34 μM and 68 μM for Lo18 and A123S corresponding to an absorbance at 280 nm of 0.2, 0.5 and 1 respectively. A volume of approximately 400 μl was loaded into 12 mm path-length two-channel cells equipped with sapphire windows. Buffer without protein was loaded in the reference channel. Sedimentation profiles were recorded overnight at 50000 rev./min every 12 min, at 275 nm and by interference. Data were analysed with the program Sedfit (available at http://www.analyticalultracentrifugation.com) using the continuous distribution of sedimentation coefficients [c(s)] analysis [29]. We used the Sednterp software (freely available at http://www.jphilo.mailway.com) to estimate the partial specific volume of the polypeptide chain (vLo18=0.7205 ml·g−1 and vA123S=0.7199 ml·g−1), the solvent density (ρ=1.004 g·ml−1) and the solvent viscosity (η=1.023 cP) at 20°C, for the c(s) analysis and calculation of the corrected s20w values. The c(s) curves were obtained considering 200 particles with sedimentation coefficients (s) between 0.1 and 20 S, with a frictional ratio, f/fmin, of 1.25 corresponding to a globular compact macromolecule. A confidence level of 0.68 was assumed.

The sedimentation coefficients, were analysed using the Svedberg equation, in terms of the molecular mass, M, considering the RS obtained from SEC: s=M.(1−vρ)/(NA.6π η RS). The ratio of RS to the minimum theoretical hydrodynamic radius of non-hydrated volume of the particle defines the frictional ratio f/fmin, related to the macromolecule hydration and shape.

Electron microscopy

Before negative staining, samples prepared for AUC were diluted at 3 μM (around 0.05 mg·ml−1) in a phosphate buffer at the desired pH. An aliquot (4 μl) of these protein samples were adsorbed on to the clean face of a carbon film on a mica sheet (carbon/mica interface) and negatively stained with 2% (w/v) uranyl acetate. Micrographs were taken under low-dose conditions with a FEI CM12 LaB6 electron microscope working at 120 kV and with a nominal magnification of 45000× using an Orius SC1000 GATAN CCD camera (1.55 Å/pixel at the specimen level). The image analysis procedure was started by selecting 800 particles in 128×128 pixel squares from 12 micrographs (sample at pH 5) by using X3D [30]. A starting model was created using one centred raw image that showed a clear circular symmetry. This single image was used to generate an initial three-dimensional starting model and to run a projection matching procedure as described in [31]. After 20 cycles of refinement because of the D8 symmetry appearance (two 8-fold symmetry rings stacked together through a 2-fold axis) of the three-dimensional reconstruction and because the reprojections of the three-dimensional structure fitted well with the raw images, we imposed full D8 symmetry to the reconstruction. A further 25 cycles of refinement was performed using 29 reprojections, and only 25% (190) of the images that had the highest correlation coefficient with the reprojection of the model were included in the reconstruction. Because of the low number of images used in the three-dimensional reconstruction, the resolution of the final model was limited to 30 Å. No contrast transfer function correction was applied. For the isosurface representation, we included the correct molecular mass of the complex (16-mer) using an average protein density of 0.84 Da/Å3. The crystal structure of the Methanococcus jannaschii Hsp16.5 monomer (PDB code 1SHS) was docked into the electron microscopy density map of Lo18 with VEDA (http://mem.ibs.fr/VEDA). The resolution of the fit was limited to 25 Å. The Hsp16.5-based model fitted into the electron microscopy density with a correlation of 81%.

Thermal aggregation of CS (citrate synthase)

An ammonium sulfate suspension of CS (Sigma) was dialysed against TE buffer [50 mM Tris/HCl and 2 mM EDTA (pH 8)] and concentrated to about 17 mg·ml−1 at 4°C. CS was then centrifuged at 14000 rev./min during 30 min at 4°C to remove precipitated protein. The exact concentration in supernatant was determined and aliquots of 30 μM CS (monomer) were prepared and stored at −20°C. CS at a concentration of 300 nM was thermally denatured in 50 mM phosphate buffer, pH 5, 6, 7, 8 or 9, at 45°C in the absence or presence of 1.2 μM, 4.8 μM or 9.6 μM Lo18 or A123S. CS aggregation was analysed by light scattering at 360 nm using a Uvicon XS70 spectrophotometer equipped with a Peltier thermosystem (Secoman, Bioserv). Each experiment was performed in triplicate. Lysozyme (1.2, 4.8 or 9.6 μM) was used as negative control.

Preparation of liposomes

O. oeni cells {50 UA (absorbance units) [1 UA=2×108 CFU (colony-forming units)·ml−1]} collected in exponential growth phase were used to extract cellular lipids according to [32]. Liposomes were obtained as described previously [25] with slight modifications. A film of extracted lipids was obtained by evaporation of chloroform using a rotavapor at 65°C (Rotary Evaporator RE100 Bibby, Odil) under nitrogen-reduced pressure followed by rehydration with pre-warmed 50 mM phosphate buffer pH 5–9 at 65°C. The liposome suspension was gently mixed and sonicated for 2×30 s (ELMA D-78224, VWR). Liposomes were stored at 4°C for a maximum of 1 week.

Fluidity measurements

Fluorescence anisotropy was measured in a cuvette filled with 250 μl of liposomes prepared as described above, containing an additional 3 μM DPH (diphenylhexatriene) probe (Molecular Probes) and in the absence or the presence of 10 μM purified Lo18 or A123S. Lysozyme (10 μM) was used as a negative control. Measurements were performed for 30 min (1 determination every 8 s). After probe insertion (10 min at 15°C), a linear gradient of temperature from 15 to 65°C (increase of 2°C per min) controlled by a Peltier system (Wavelength Electronics) was applied to the liposome suspension. Anisotropy measurements were carried out using a Fluorolog 3 spectrofluorimeter (FLUOROLOG-3, Jobin Yvon) and anisotropy values were calculated according to [33]. Excitation and emission wavelengths were 352 nm and 402 nm respectively. Each experiment was done in triplicate.

Statistical analysis

One-way ANOVA was carried out using SIGMASTAT® v. 3.0.1 software (SPSS). The Holm–Sidak test (n=3, P<0.05) was performed to find significant differences.

RESULTS

Effect of pH on the chaperone activity of Lo18 and A123S

The chaperone activity of Lo18 was checked in vitro by assaying its ability to protect the CS from thermal aggregation [24] as a function of the pH in the range 5–9. Chaperone activity of Lo18 and A123S at a concentration of 9.6 μM was reported in Figure 1 in the range of pH 5–8 only, because at pH 9 the very high CS aggregation velocity prevents accurate measurements (results not shown). When heated at 45°C, CS began to form insoluble aggregates that could be detected by recording the increase in light scattering at 360 nm (Figure 1, ●). Lo18 (Figure 1, ○) poorly protects CS from thermal aggregation at pH 5, whereas the best thermal aggregation reduction was detected at pH 6 and 7. Lo18 remains active at pH 8 even if the activity was difficult to assess accurately due to the high CS aggregation velocity observed at this pH, which may hamper the effect of Lo18. When the same experiments were performed with A123S (Figure 1, Δ), a significant chaperone activity was detected only at pH 7. The chaperone activity of Lo18 in the concentration range 1.2–9.6 μM was also assayed and found to be dependent on the protein concentration, as reported previously [24]. The same is true for A123S with the maximal activity obtained at 4.8 μM (results not shown). Finally, negative controls run in the presence of lysozyme, a protein without any known chaperone activity, gave results similar to those of the blanks without protein (results not shown).

Chaperone activity of Lo18 or A123S in the pH range 5–8

Figure 1
Chaperone activity of Lo18 or A123S in the pH range 5–8

The light scattering at 360 nm corresponding to thermal aggregation of 300 nM CS at 45°C was recorded as a function of time and pH in the absence (●) or the presence of 9.6 μM of either Lo18 (○) or A123S (Δ). Averages and error bars (S.D.) are calculated from the results of three independent measurements.

Figure 1
Chaperone activity of Lo18 or A123S in the pH range 5–8

The light scattering at 360 nm corresponding to thermal aggregation of 300 nM CS at 45°C was recorded as a function of time and pH in the absence (●) or the presence of 9.6 μM of either Lo18 (○) or A123S (Δ). Averages and error bars (S.D.) are calculated from the results of three independent measurements.

The quaternary structure of sHsp Lo18 is dependent on the pH

The quaternary structure of Lo18 was first analysed by SEC as a function of pH in the range 5–9 (Figure 2A). Lo18 eluted as a more or less symmetrical broad peak regardless of the pH value, suggesting a mixture of species. At pH 5, 6 and 7, the elution volume of 11.5 ml for the main peak allows to estimate the presence of oligomers of RS=60±2 Å, thus approximately 300 kDa, corresponding roughly to 17 or 18 subunits, considering a precise molecular mass of 16938 Da for the Lo18 monomer and a globular shape. The peak at pH 6 is the broader, suggesting a higher size heterogeneity. The peak at pH 7 shows an extra feature around 14 ml, i.e. a Stokes radius of 57±2 Å suggesting the presence of lower molecular mass species in this sample. The elution profile is dramatically different at pH 8 and 9 with peaks eluting around 16 ml corresponding to 24±4 Å. This is the expected volume for the elution of the dimeric compact form of Lo18.

SEC of Lo18 and A123S

Figure 2
SEC of Lo18 and A123S

(A) Lo18 was loaded on to a Superdex 200 10/300 GL column equilibrated in phosphate buffer at pH 5 (continuous line, high molecular mass), pH 6 (dotted line, high molecular mass), pH 7 (dashed line, high molecular mass), pH 8 (solid line, low molecular mass), pH 9 (dotted line, low molecular mass). For clarity, the chromatograms at pH 8 and 9 are represented with a different baseline. (B) Elution of A123S at pH 7. Results of the column calibration are displayed in (A).

Figure 2
SEC of Lo18 and A123S

(A) Lo18 was loaded on to a Superdex 200 10/300 GL column equilibrated in phosphate buffer at pH 5 (continuous line, high molecular mass), pH 6 (dotted line, high molecular mass), pH 7 (dashed line, high molecular mass), pH 8 (solid line, low molecular mass), pH 9 (dotted line, low molecular mass). For clarity, the chromatograms at pH 8 and 9 are represented with a different baseline. (B) Elution of A123S at pH 7. Results of the column calibration are displayed in (A).

Oligomerization of Lo18 between pH 5 and 9 was further investigated by AUC at three protein concentrations (13 μM, 34 μM and 68 μM). Examples of raw data and their fits for Lo18 at 68 μM at pH 5 and 9, and superimpositions of the distribution of sedimentation coefficients for Lo18 at 68 and 13 μM at various pH values are shown in Supplementary Figure S1 (at http://www.BiochemJ.org/bj/444/bj4440097add.htm). Figure 3 reports the main features in a peak-by-peak analysis. A main species was observed at pH 5 and 6 at around 12 S (s20,w=12.4 S). Considering a globular shape (f/fmin value of 1.25), the s-value corresponds to a 16-mer, which represented approximately 90% of the species. Combining the s-values with the Stokes values estimated by SEC gives a slightly larger number of 17–18 subunits, in an oligomer with a more extended shape (f/fmin=1.33). The measurements do not allow discriminating accurately between the two association states. These samples contained also 8–10% of the heaviest particles with s-values around 15 S, which could correspond to assemblies of 21±1 subunits considering a globular shape. A minor species of about 1.5 S (>3%) that may correspond to an elongated monomer is also present. At pH 7 and above, the sedimentation profiles analysis revealed an increasing heterogeneity. The main particles at pH 7 representing 50–75% of the population were at 10.8±0.2 S (s20,w=11.2 S). The c(s) distributions show also 15–20% of an intermediate species in the range 6–7 S and 15% of small species at around 2–2.5 S. Their presence is emphasized at pH 8 and 9, where the analysis revealed several peaks poorly resolved. Considering a mixture of discrete species in solution, the combination of the main peaks at pH 7, s=10.8 S, with a RS of 57±2 Å determined by SEC, allows us to estimate the presence of particles of ~12–15 subunits. Globular tetramers and hexamers are calculated to sediment at 4.8 and 6.3 respectively. On the other hand, the smallest species at 2.6±0.2 S (s20,w=2.9 S) is rather well defined and probably corresponds to a dimer. It contributes to ~25% of the whole species at pH 8 and 9. At pH 7–9, the population was also completed with the presence of 5–10% of oligomers larger than 20-mers. The assignment of the oligomeric species determined by AUC fits pretty well with that deduced from SEC. However their distribution does not reveal any correlation with the protein concentration, in the range used during the present study.

Distribution of the sedimentation coefficient as a function of pH for Lo18 and A123S

Figure 3
Distribution of the sedimentation coefficient as a function of pH for Lo18 and A123S

AUC experiments were performed at three protein concentrations: 13 μM (white columns), 34 μM (grey columns) and 68 μM (black columns). At each concentration, different populations of sedimentation coefficients (frames in abscissa) corresponding to different oligomeric forms were identified.

Figure 3
Distribution of the sedimentation coefficient as a function of pH for Lo18 and A123S

AUC experiments were performed at three protein concentrations: 13 μM (white columns), 34 μM (grey columns) and 68 μM (black columns). At each concentration, different populations of sedimentation coefficients (frames in abscissa) corresponding to different oligomeric forms were identified.

A123S displays a chaperone activity at pH 7, where this protein behaves as a dimer both by SEC (Figure 2B) with an elution volume around 16 ml and by AUC experiments (Figure 3) with the main peak at 2.7–3 S representing 60–70% of the particles and assumed to correspond to a dimer. An extra peak with a sedimentation coefficient of 6–6.5 S related to the presence of a putative hexamer was also detected. These results are in agreement with those obtained by in vivo cross-linking experiments where the presence of dimers and few oligomers was observed [5]. This confirms that A123S is impaired in its capacity to form large oligomers.

Electron microscopy was used to observe directly the oligomeric complexes formed by Lo18 at different pH values after negative staining (Figure 4A). The sample appeared as quite well-structured spherical oligomers with few aggregates at pH 5 and pH 6. From pH 7–9, several aggregates were observed and Lo18 appeared less structured and more heterogeneous in size, preventing image analysis. The pH 5 samples which appear to be the more homogenous ones were subjected to image analysis. The three-dimensional model we obtained is a rather low-resolution model (30 Å), but it shows the general organization of the complex: a doughnut-like structure made by two stacked rings with eight monomers per ring (Figures 4B and 4C). The dimer that is probably the building block of the complex can be identified on the side view. The two vertices of the structure are occupied by a non well-defined density. It is unlikely that this density can be assigned to extra copies of the chaperone which did not follow the imposed symmetry. The possibility that the N-terminal region flanking the α-crystallin domain is involved in such a density will be discussed. By visual inspection of the images, it is clear that there is quite a high heterogeneity in the sample: only 25% of the starting images have been used and probably other structures can co-exist with the 16-mer. It should be noticed that the size distribution can be different between electron microscopy and AUC since the protein concentration used in negative staining is approximately five times lower than the lowest one used in AUC.

Electron microscopy of Lo18

Figure 4
Electron microscopy of Lo18

(A) Negative-staining images of Lo18 at different pH values (stained with uranyl acetate). (B and C) Image analysis for the pH 5 sample. (B) Reprojections of the three-dimensional structure of Lo18 (top row) compared with the corresponding class averages used to calculate the three-dimensional reconstruction (middle row) and with some corresponding raw images (bottom row). The size of each square is 200 Å×200 Å. (C) Three-dimensional structure of Lo18 at about 30 Å resolution viewed in top view down the 8-fold axis (left-hand panel), slightly tilted (middle panel) and in side view down a 2-fold axis (right-hand panel).

Figure 4
Electron microscopy of Lo18

(A) Negative-staining images of Lo18 at different pH values (stained with uranyl acetate). (B and C) Image analysis for the pH 5 sample. (B) Reprojections of the three-dimensional structure of Lo18 (top row) compared with the corresponding class averages used to calculate the three-dimensional reconstruction (middle row) and with some corresponding raw images (bottom row). The size of each square is 200 Å×200 Å. (C) Three-dimensional structure of Lo18 at about 30 Å resolution viewed in top view down the 8-fold axis (left-hand panel), slightly tilted (middle panel) and in side view down a 2-fold axis (right-hand panel).

Relation between pH and liposome stabilization by Lo18 and A123S

Liposomes formed by lipids extracted from O. oeni cells collected during exponential growth were mixed with DPH. The variation of membrane fluidity was then measured at pH values varying from 5 to 8 by steady-state fluorescence anisotropy of DPH during temperature ramping (15–65°C), in the absence or presence of Lo18 (Figure 5A). The gradual decrease of the anisotropy values reflected the liposomes fluidizing due to temperature increase and reached 42% of the initial value at 65°C regardless of the pH in the absence of Lo18. The same results were obtained in presence of lysozyme used as a negative control (results not shown). It is important to point out that the curves displayed in Figure 5 are directly comparable in the absence of Lo18 since the kinetics of liposome fluidizing appeared to be independent of the pH. As for the chaperone activity, the best stabilization effect of membrane fluidity was obtained at pH 7. Below 31°C, there was no significant effect of the presence of Lo18 on the order of membrane lipids, but above this temperature the increase in acyl chain motion caused by the increase in temperature was significantly reduced. The membrane fluidity that represented 79% of the initial value at 31°C decreased only to 73% at 65°C. When the experiment was performed at pH 8, no significant effect of the presence of Lo18 was observed until 55°C and a very slight, but significant, effect (P<0.05) was discernable between 55°C and 65°C. The situation was different at acidic pH. Until 39°C at pH 5, the reduction of liposome fluidity was better than that obtained at pH 7 (84% compared with 79%). However, above this temperature, the fluidizing of the liposomes strongly increased and reached 57% of the initial value at 65°C, compared with 73% at pH 7. The same atypical profile was observed at pH 6, although to a lower extent. The presence of Lo18 had a significant effect on liposome stabilization between 23°C and 47°C. Above this temperature, the anisotropy values decreased and the level of fluidity obtained (43% of the initial fluidity) was not significantly different from that obtained with liposomes alone.

Evolution of liposome fluidity

Figure 5
Evolution of liposome fluidity

(A) Results are expressed as anisotropy percentage as a function of the temperature increase in the absence of protein (●, broken line) or in the presence of 10 μM Lo18 at different pH values: 5 (○), 6, (Δ), 7 (□) and 8 (◇). (B) The same experiment was reproduced in the presence of 10 μM A123S at pH 7 (●, solid line). The curves obtained without protein (●, broken line) and in the presence of Lo18 at pH 7 (□) were purposely reproduced. The initial anisotropy value (100%) was equal to 0.167±0.08.

Figure 5
Evolution of liposome fluidity

(A) Results are expressed as anisotropy percentage as a function of the temperature increase in the absence of protein (●, broken line) or in the presence of 10 μM Lo18 at different pH values: 5 (○), 6, (Δ), 7 (□) and 8 (◇). (B) The same experiment was reproduced in the presence of 10 μM A123S at pH 7 (●, solid line). The curves obtained without protein (●, broken line) and in the presence of Lo18 at pH 7 (□) were purposely reproduced. The initial anisotropy value (100%) was equal to 0.167±0.08.

The evolution of the liposomes fluidity was also assayed in the presence of A123S at pH 7 (Figure 5B). A slight protective effect was discernable until 35°C and the presence of the modified protein produced significant protection above 55°C, even though it remained modest.

DISCUSSION

As the growth of O. oeni is crucial for malolactic fermentation in wine, a further characterization of Lo18 that was demonstrated to actively participate to O. oeni adaptation to wine is of particular interest. In the present study, we provide strong evidence for a relationship between the dynamic oligomerization of Lo18 and its role in membrane stabilization and protein protection. The oligomeric status of Lo18 was explored as a function of pH in the range 5–9. The AUC data confirmed the results of SEC experiments indicating that Lo18 formed heavy oligomers between pH 5 and 7, but was no longer able to maintain this association at higher pH values. The main species representing approximately 90% of the population corresponds to 16-subunit oligomers at pH 5 and 6, and probably evolves to 12-subunit oligomers representing more than half of the species at pH 7. Increasing the pH above pH 7 leads to an increased heterogeneity of the samples with the appearance of low-molecular-mass species. Well-structured spherical oligomers appropriate for image analysis were present at pH 5. The same oligomers are also present at pH 6 when Lo18 gains a significant chaperone activity. This pH corresponds to the intracellular one when O. oeni is grown in acidic medium in the presence of L-malate [34]. As determined by electron microscopy, Lo18 appears at pH 5 as an assembly of two stacked rings of octamers, forming the 16-mer detected by AUC. Interestingly, the X-ray structure of the Hsp16.5 dimer from M. jannaschii fits pretty well (correlation coefficient >81%) with the three-dimensional electron microscopy envelope (Figure 6B). The eight monomers of each ring of Lo18 can accommodate each the α-crystallin domain of a Hsp16.5 monomer. M. jannaschii Hsp16.5 is the only prokaryotic sHsp for which an X-ray structure was determined [35]. It appears perfectly suited for such a superposition. Hsp16.5 and Lo18 share 54% of sequence similarity in their α-crystallin domains including 26% identity (Supplementary Figure S2 at http://www.BiochemJ.org/bj/444/bj4440097add.htm). On the basis of the whole sequence, they share also a very similar secondary structure prediction (Supplementary Figure S2). The ribbon representation of the Hsp16.5 dimer shows that the N-terminal 32 residues are missing (Figure 6A) in the structure. This is consistent with the model of an intrinsically disordered N-terminal arm providing diverse geometries of interaction sites with the client proteins of dodecameric PsHsp18.1 (Pisum sativum Hsp18.1) from pea [36]. Besides, the evolutionary variable N-terminal region was also demonstrated to be critical for substrate specificity [3638]. In the atomic model constructed using the synergistic application of cryo-electron microscopy and EPR [39], the N-terminal sequence of Hsp16.5 was located in the shell formed by the oligomeric structure. One can suggest a similar organization for Lo18 with the N-terminal sequence of each Lo18 protomer pointing towards the centre of the stacked rings in the doughnut-like structure, hence contributing to the two apical densities observed in electron microscopy. The doughnut-like model could represent the inactive structure of a closed molecular cage requiring major structural rearrangements to turn active. The pH-dependent oligomer plasticity of Lo18 illustrates the ability of this protein to be involved in such rearrangements, including the association/dissociation dynamics required for substrate binding in the central core of the protein. Chaperones are thought of as highly sophisticated protein machines that assist the correct folding of partner molecules by a combination of mechanisms. For instance, the chaperone activity will depend on the accessibility of the hydrophobic surface for its client protein and on a structural rearrangement around it [40]. It appeared that Lo18 with a quaternary structure corresponding to approximately 12-subunit oligomers at pH 7 is suitable for such activities. Actually, if the overall organization of the chaperone is conserved (two stacked rings), the contacts between two adjacent dimers can be quasi-equivalent when comparing hexameric (12-mer) and octameric (16-mer) rings (only the solid angle will change). At pH 5, this protein may either be less able to develop hydrophobic contact with denatured CS or is impaired in conformational adjustments. As expected, the loss of high-molecular-mass particles at basic pH matched with an alteration of chaperone activity. This most probably indicates a decrease in the subunit association constant leading to rapid association/dissociation equilibrium.

Fitting of the pseudo atomic model of a Methanococcus jannaschii Hsp16.5 subunit pair in the three-dimensional reconstruction of Lo18

Figure 6
Fitting of the pseudo atomic model of a Methanococcus jannaschii Hsp16.5 subunit pair in the three-dimensional reconstruction of Lo18

(A) Ribbon view of the Hsp16.5 dimer used in the superposition (PDB code 1SHS). The N- and C-terminus in each protomer are indicated. Note that the N-terminal 32 residues are highly disordered and not structurally defined. (B) Eight Hsp16.5 dimers were docked within the electron microscopy map of Lo18 with the help of the VEDA software (http://mem.ibs.fr/VEDA). Top view down the 8-fold axis (left-hand panel), slightly tilted (middle panel) and in side view down a 2-fold axis (right-hand panel).

Figure 6
Fitting of the pseudo atomic model of a Methanococcus jannaschii Hsp16.5 subunit pair in the three-dimensional reconstruction of Lo18

(A) Ribbon view of the Hsp16.5 dimer used in the superposition (PDB code 1SHS). The N- and C-terminus in each protomer are indicated. Note that the N-terminal 32 residues are highly disordered and not structurally defined. (B) Eight Hsp16.5 dimers were docked within the electron microscopy map of Lo18 with the help of the VEDA software (http://mem.ibs.fr/VEDA). Top view down the 8-fold axis (left-hand panel), slightly tilted (middle panel) and in side view down a 2-fold axis (right-hand panel).

It is tempting to speculate that the building block required for Lo18 oligomerization is a dimer, as suggested for the mouse protein Hsp25 [41], M. jannaschii Hsp16.5 [35], Hsp16.9 from Triticum aestivum [42] or Hsp17 from Synechocystis [43]. The structure of dimeric Hsp16.5 matches with the three-dimensional electron microscopy envelope of Lo18, but further arguments for this hypothesis are brought by AUC data that revealed also the presence of low-molecular-mass species that can be assigned to dimeric forms regardless of the pH and the significant chaperone activity of A123S at pH 7. The replacement of Ala123 by a serine residue appeared to be critical for oligomerization of Lo18 and led to the production of a protein characterized as a dimer, both by SEC and AUC experiments. Ala123 localized in a β-sheet of the α-crystallin domain is highly conserved among the proteins of the sHsp family. Its role in formation of oligomers and in chaperone activity was already evidenced in other bacterial sHsps [3,44] or in a truncation mutant of the human αB-crystallin protein [45].

The liposome stabilization activity is dependent on the ability of the protein to interact with the lipids, these interactions being mediated in part by the protein and lipid relative electronic charges. The liposomes formed with phospholipids purified from O. oeni cells [25] mimic the bacterial membrane and their phospholipidic composition probably corresponds to that encountered by Lo18 immediately after a heat shock [23]. The pKa value of the amino group of phosphatidylethanolamine, the most represented phospholipid in O. oeni membrane [46], has been estimated to be approximately 9.5 [47] and the theoretical pI of Lo18 is close to 5. This means that electronic interactions between Lo18 and O. oeni membrane should be maximal between pH 6 and 8. Actually, the importance of charge interactions between Hsp and lipids has already been pointed out. For instance, a better affinity of GroEL for membrane bilayer was noticed at acidic pH [48]. The authors suggested that the interaction between membrane phospholipids and GroEL was advantaged by hydrophobic residues maybe more exposed to the surface. A strong interaction was also described between negatively charged Hsp16.3 from Mycobacterium tuberculosis and positively charged lipids at neutral pH [49]. Besides, the dissociation of sHsp seems to be a prerequisite for the process of association with the membrane as shown for Hsp16.3 [49] or Hsp17 of Synechocystis [19]. These behaviours are in agreement with the best liposome stabilization activity of Lo18 observed at pH 7 that allows the best compromise between two crucial criteria: oligomeric status and charge distribution. Finally, dimeric A123S displayed an altered capacity to protect membrane against the deleterious effects of temperature increase compared with Lo18 at pH 7. This suggests that the inability of A123S to be involved in dynamic subunit exchange [50] may be critical for chaperone activity and for lipid interaction and protection as well. These results are in agreement with those obtained during in vivo experiments [5].

In conclusion, we have demonstrated the correlation between the dynamic oligomerization of Lo18 and its requirement for both lipid stabilization and protein protection in vitro activities. These results allow us to propose for the first time a low-resolution model for the spatial organization of Lo18 in a probably inactive conformation. The interaction between Lo18 and the lipids of the O. oeni membrane remains to be explored in more detail.

Abbreviations

     
  • AUC

    analytical ultracentrifugation

  •  
  • CS

    citrate synthase

  •  
  • DPH

    diphenylhexatriene

  •  
  • Hsp

    heat-shock protein

  •  
  • LB

    Luria–Bertani

  •  
  • Oshsp

    16.9, Oryza sativa Hsp16.9

  •  
  • Rs

    Stokes radius

  •  
  • SEC

    size-exclusion chromatography

  •  
  • sHsp

    small Hsp

AUTHOR CONTRIBUTION

Magali Maitre performed protein expression and purification, SEC, and chaperone and lipochaperone activities. She also participated in preparation of the electron microscope and AUC samples. Aurélie Rieu participated to establish the protocols for chaperone and lipochaperone assays. Daphna Fenel and Guy Schoehn recorded the electronic microscopy data, and Guy Schoen analysed the images and determined the low-resolution structure. Christine Ebel performed analysis of the AUC data. Jacques Covès, Stéphanie Weidmann and Jean Guzzo conceived the work and supervised the experiments. Magali Maitre, Jacques Covès, Jean Guzzo, Stéphanie Weidmann, Guy Schoehn and Christine Ebel contributed to writing the paper.

We thank the PSB (Partnership for Structural Biology) electron microscopy platform for access to the FEI CM12 LaB6 electron microscope. We thank also Aline Appourchaud for the AUC data acquisition and processing, and Dr Gael Goret for expert handling of the VEDA software.

FUNDING

This work was supported by the Ministère de l'Education Nationale, de la Recherche et de la Technologie and the Conseil Régional de Bourgogne [grant number 2009 B2R4 018].

References

References
1
Watson
K.
Microbial stress protein
Adv. Microb. Physiol.
1990
, vol. 
31
 (pg. 
183
-
223
)
2
Nakamoto
H.
Vigh
L.
The small heat shock proteins and their clients
Cell. Mol. Life Sci.
2007
, vol. 
64
 (pg. 
294
-
306
)
3
Narberhaus
F.
α-Crystallin-type heat shock proteins: socializing minichaperones in the context of a multichaperone network
Microbiol. Mol. Biol.
2002
, vol. 
66
 (pg. 
64
-
93
)
4
Yeh
C. H.
Chang
P. F. L.
Yeh
K. W.
Lin
W. C.
Chen
Y. M.
Lin
C. Y.
Expression of a gene encoding a 16.9-kDa heat-shock protein, Oshsp16.9, in Escherichia coli enhances thermotolerance
Proc. Natl. Acad. Sci. U.S.A.
1997
, vol. 
94
 (pg. 
10967
-
10972
)
5
Weidmann
S.
Rieu
A.
Rega
M.
Coucheney
F.
Guzzo
J.
Distinct amino acids of the Oenococcus oeni small heat shock protein Lo18 are essential for damaged protein protection and membrane stabilization
FEMS Microbiol. Lett.
2010
, vol. 
309
 (pg. 
8
-
15
)
6
Laksanalamai
P.
Maeder
D. L.
Robb
F. T.
Regulation and mechanism of action of the small heat shock protein from the hyperthermophilic archaeon Pyrococcus furiosus
J. Bacteriol.
2001
, vol. 
183
 (pg. 
5198
-
5202
)
7
de Jong
W.
Caspers
G. J.
Leunissen
J.
Genealogy of the α-crystallin-small heat-shock protein superfamily
Int. J. Biol. Macromol.
1998
, vol. 
22
 (pg. 
151
-
162
)
8
Horwitz
J.
α-Crystallin can function as a molecular chaperone
Proc. Natl. Acad. Sci. U.S.A.
1992
, vol. 
89
 (pg. 
10449
-
10453
)
9
Haslbeck
M.
Kastenmüller
A.
Buchner
J.
Weinkauf
S.
Braun
N.
Structural dynamics of archaeal small heat shock proteins
J. Mol. Biol.
2008
, vol. 
378
 (pg. 
362
-
374
)
10
Matuszewska
M.
Kuczyńska-Wiśnik
D.
Laskowska
E.
Liberek
K.
The small heat shock protein IbpA of Escherichia coli cooperates with IbpB in stabilization of thermally aggregated proteins in a disaggregation competent state
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
12292
-
12298
)
11
MacRae
T. H.
Structure and function of small heat shock/α-crystallin proteins: established concepts and emerging ideas
Cell. Mol. Life Sci.
2000
, vol. 
57
 (pg. 
899
-
913
)
12
Lentze
N.
Aquilina
J. A.
Lindbauer
M.
Robinson
C. V.
Narberhaus
F.
Temperature and concentration controlled dynamics of rhizobial small heat shock proteins
Eur. J. Biochem.
2004
, vol. 
271
 (pg. 
2494
-
2503
)
13
Chernik
I.
Panasenkoa
O.
Li
Y.
Marston
S.
Gusev
N.
pH-induced changes of the structure of small heat shock proteins with molecular mass 24/27 kDa (HspB1)
Biochim. Biophys. Acta
2004
, vol. 
324
 (pg. 
1199
-
1203
)
14
Török
Z.
Goloubinoff
P.
Horvath
I.
Tsvetkova
N.
Glatz
A.
Balogh
G.
Varvasovski
V.
Los
D.
Vierling
E.
Crower
J.
Vigh
L.
Synechocystis HSP17 is an amphitropic protein that stabilizes heat-stressed membranes and binds denatured proteins for subsequent chaperone-mediated refolding
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
3098
-
3103
)
15
Lunsdorf
H.
Schairer
H. U.
Heidelbach
M.
Localization of the stress protein SP21 in indole-induced spores, fruiting bodies, and heat-shocked cells of Stigmatella aurantiaca
J. Bacteriol.
1995
, vol. 
177
 (pg. 
7092
-
7099
)
16
Sales
K.
Brandt
W.
Rumbak
E.
Lindsey
G.
The LEA-like protein HSP 12 in Saccharomyces cerevisiae has a plasma membrane location and protects membranes against desiccation and ethanol-induced stress
Biochim. Biophys. Acta
2000
, vol. 
1643
 (pg. 
267
-
278
)
17
Horváth
I.
Glatz
A.
Varvasovszki
V.
Török
Z.
Páli
T.
Balogh
G.
Kovács
E.
Nádasdi
L.
Benkö
S.
Joó
F.
Vígh
L.
Membrane physical state controls the signaling mechanism of the heat shock response in Synechocystis PCC 6803: identification of hsp17 as a ‘fluidity gene’
Proc. Natl. Acad. Sci. U.S.A.
1998
, vol. 
95
 (pg. 
3513
-
3518
)
18
Giese
K. C.
Vierling
E.
Mutants in a small heat shock protein that affect the oligomeric state
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
32674
-
32683
)
19
Balogi
Z.
Cheregi
O.
Giese
K. C.
Juhasz
K.
Vierling
E.
Vass
I.
Vígh
L.
Horváth
I.
A mutant small heat shock protein with increased thylakoid association provides an elevated resistance against UV-B damage in Synechocystis 6803
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
22983
-
22991
)
20
Lonvaud-Funel
A.
Lactic acid bacteria in the quality improvement and depreciation of wine
Antonie Van Leeuwenhoek
1999
, vol. 
76
 (pg. 
317
-
331
)
21
Grandvalet
C.
Assad-Garcia
J.
Chu-Ky
S.
Tollot
M.
Guzzo
J.
Gresti
J.
Tourdot-Maréchal
R.
Changes in membrane lipid composition in ethanol- and acid-adapted Oenococcus oeni cells: characterization of the cfa gene by heterologous complementation
Microbiology
2008
, vol. 
154
 (pg. 
2611
-
2619
)
22
Guzzo
J.
Jobin
M. P.
Delmas
F.
Fortier
L. C.
Garmyn
D.
Tourdot-Maréchal
R.
Lee
B.
Divies
C.
Regulation of stress response in Oenococcus oeni as a function of environmental changes and growth phase
Int. J. Food Microbiol.
2000
, vol. 
55
 (pg. 
27
-
31
)
23
Jobin
M. P.
Delmas
F.
Garmyn
D.
Divies
C.
Guzzo
J.
Molecular characterization of the gene encoding an 18-kilodalton small heat shock protein associated with the membrane of Leuconostoc oenos
Appl. Environ. Microbiol.
1997
, vol. 
63
 (pg. 
609
-
614
)
24
Delmas
F.
Divies
C.
Guzzo
J.
Biochemical and physiological studies of the small heat shock protein Lo18 from the lactic acid bacterium Oenococcus oeni
J. Mol. Microb. Biotech.
2001
, vol. 
3
 (pg. 
601
-
610
)
25
Coucheney
F.
Gal
L.
Beney
L.
Lherminier
J.
Gervais
P.
Guzzo
J.
A small HSP, Lo18, interacts with the cell membrane and modulates lipid physical state under heat shock conditions in a lactic acid bacterium
Biochim. Biophys. Acta
2005
, vol. 
1720
 (pg. 
92
-
98
)
26
Guzzo
J.
Delmas
F.
Pierre
F.
Jobin
M. P.
Samyn
B.
Van Beeumen
J.
Cavin
J. F.
Divies
C.
A small heat shock protein from Leuconostoc oenos induced by multiple stresses and during stationary growth phase
Lett. Appl. Microbiol.
1997
, vol. 
24
 (pg. 
393
-
396
)
27
Guzzo
J.
Jobin
M. P.
Divies
C.
Increase of sulfite tolerance in Oenococcus oeni by means of acidic adaptation
FEMS Microbiol. Lett.
1998
, vol. 
160
 (pg. 
43
-
47
)
28
Cavin
J. F.
Prevost
H.
Lin
J.
Schmitt
P.
Divies
C.
Medium for screening Leuconostoc oenos strains defective in malolactic fermentation
Appl. Environ. Microbiol.
1989
, vol. 
55
 (pg. 
751
-
753
)
29
Schuck
P.
Size-distribution analysis of macromolecules by sedimentation velocity ultracentrifugation and lamm equation modeling
Biophys. J.
2000
, vol. 
78
 (pg. 
1606
-
1619
)
30
Conway
J. F.
Steven
A. C.
Methods for reconstructing density maps of ‘single’ particles from cryoelectron micrographs to subnanometer resolution
J. Struct. Biol.
1999
, vol. 
128
 (pg. 
106
-
118
)
31
Franzetti
B.
Schoehn
G.
Hernandez
J. F.
Jaquinod
M.
Ruigrok
R. W.
Zaccai
G.
Tetrahedral aminopeptidase: a novel large protease complex from archaea
EMBO J.
2002
, vol. 
21
 (pg. 
2132
-
2138
)
32
Bligh
E. G.
Dyer
W. J.
A rapid method of total lipid extraction and purification
Can. J. Med. Sci.
1959
, vol. 
37
 (pg. 
911
-
917
)
33
Shinitzky
M.
Barenholz
Y.
Fluidity parameters of lipid regions determined by fluorescence polarization
Biochim. Biophys. Acta
1978
, vol. 
515
 (pg. 
367
-
394
)
34
Salema
M.
Lolkema
J. S.
San Romao
M. V.
Lourero Dias
M. C.
The proton motive force generated in Leuconostoc oenos by L-malate fermentation
J. Bacteriol.
1996
, vol. 
178
 (pg. 
3127
-
3132
)
35
Kim
K. K.
Kim
R.
Kim
S. H.
Crystal structure of a small heat-shock protein
Nature
1998
, vol. 
394
 (pg. 
595
-
599
)
36
Jaya
N.
Garcia
V.
Vierling
E.
Substrate binding site flexibility of the small heat shock protein molecular chaperones
Proc. Natl. Acad. Sci. U.S.A.
2009
, vol. 
106
 (pg. 
15604
-
15609
)
37
Basha
E.
Friedrich
K. L.
Vierling
E.
The N-terminal arm of small heat shock proteins is important for both chaperone activity and substrate specificity
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
39943
-
39952
)
38
Giese
K. C.
Basha
E.
Catague
B. Y.
Vierling
E.
Evidence for an essential function of the N-terminus of a small heat shock protein in vivo, independent of in vitro chaperone activity
Proc. Natl. Acad. Sci. U.S.A.
2005
, vol. 
102
 (pg. 
18896
-
18901
)
39
Koteiche
H. A.
Chiu
S.
Majdoch
R. L.
Stewart
P. L.
Mchaourab
H. S.
Atomic models by cryo-EM and site-directed spin labeling: application to the N-terminal region of Hsp16.5
Structure
2005
, vol. 
13
 (pg. 
1165
-
1171
)
40
Giese
K. C.
Vierling
E.
Changes in oligomerization are essential for the chaperone activity of a small heat shock protein in vivo and in vitro
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
6310
-
6318
)
41
Dudich
I. V.
Zav'yalov
V. P.
Pfeil
W.
Gaestel
M.
Zav'yalova
G. A.
Denesyuk
A. I.
Korpela
T.
Dimer structure as a minimum cooperative subunit of small heat-shock proteins
Biochim. Biophys. Acta
1995
, vol. 
1253
 (pg. 
163
-
168
)
42
van Montfort
R. L. M.
Basha
E.
Friedrich
K. L.
Slingsby
C.
Vierling
E.
Crystal structure and assembly of a eukaryotic small heat shock protein
Nat. Struct. Mol. Biol.
2001
, vol. 
8
 (pg. 
1025
-
1030
)
43
Haslbeck
M.
Braun
N.
Stromer
T.
Richter
B.
Model
N.
Weinkauf
S.
Buchner
J.
Hsp42 is the general small heat shock protein in the cytosol of Saccharomyces cerevisiae
EMBO J.
2004
, vol. 
23
 (pg. 
638
-
649
)
44
Lentze
N.
Studer
S.
Naberhaus
F.
Structural and functional defects caused by point mutations in the α-crystallin domain of a bacterial α-heat shock protein
J. Mol. Biol.
2003
, vol. 
328
 (pg. 
927
-
937
)
45
Feil
I. K.
Malfois
M.
Hendle
J.
van der Zandt
H.
Svergun
D. I.
A novel quaternary structure of the dimeric α-crystallin domain with chaperone-like activity
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
12021
-
12029
)
46
Teixeira
H.
Gonçalves
M. G.
Rozès
N.
Ramos
A.
San Romäo
M. V.
Lactobacillic acid accumulation in the plasma membrane of Oenococcus oeni: a response to ethanol stress?
Microb. Ecol.
2002
, vol. 
43
 (pg. 
146
-
153
)
47
Tsui
F. C.
Ojcius
D. M.
Hubbell
W. L.
The intrinsic pKa values for phosphatidylserine and phosphatidylethanolamine in phosphatidylcholine host bilayers
Biophys. J.
1986
, vol. 
49
 (pg. 
459
-
468
)
48
Török
Z.
Horvath
I.
Goloubinoff
P.
Kova
E.
Glatz
A.
Balogh
G.
Vigh
L.
Evidence for a lipochaperonin: association of active proteinfolding GroESL oligomers with lipids can stabilize membranes under heat shock conditions
Proc. Natl. Acad. Sci. U.S.A.
1997
, vol. 
94
 (pg. 
2192
-
2197
)
49
Zhang
H.
Fu
X.
Jia
W.
Zhang
X.
Liu
C.
Chang
Z.
The association of small heat shock protein Hsp16.3 with the plasma membrane of Mycobacterium tuberculosis: dissociation of oligomers is a prerequisite
Biochem. Biophys. Res. Commun.
2005
, vol. 
330
 (pg. 
1055
-
1061
)
50
Bova
M. P.
Ding
L. L.
Horwitz
J.
Fung
B. K.
Subunit exchange of αA-crystallin
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
29511
-
29517
)

Supplementary data