Ice-induced partial unfolding and aggregation of an integral membrane protein
Iona P. Garber Cohen, Pablo R. Castello
⁎
,1, F. Luis González Flecha
⁎
Laboratorio de Biofísica Molecular, Instituto de Química y Fisicoquímica Biológicas, Universidad de Buenos Aires, CONICET, Buenos Aires, Argentina
a b s t r a c t
a r t i c l e i n f o
Article history: Received 10 March 2010
Received in revised form 10 July 2010 Accepted 28 July 2010
Available online 4 August 2010
Keywords: Protein freezing Protein stability
Freeze-induced denaturation Plasma membrane calcium pump ATPase
Inactivation mechanism
Although the deleterious effects of ice on water-soluble proteins are well established, little is known about the freeze stability of membrane proteins. Here we explore this issue through a combined kinetic and spectroscopic approach using micellar-purified plasma membrane calcium pump as a model. The ATPase activity of this protein significantly diminished after freezing using a slow-cooling procedure, with the decrease in the activity being an exponential function of the storage time at 253 K, with t½= 3.9 ± 0.6 h. On the contrary, no significant changes on
enzyme activity were detected when a fast cooling procedure was performed. Regardless of the cooling rate, successive freeze–thaw cycles produced an exponential decrease in the Ca2+-ATPase activity, with the number of
cycles at which the activity was reduced to half being 9.2 ± 0.3 (fast cooling) and 3.7 ± 0.2 (slow cooling). PAGE analysis showed that neither degradation nor formation of SDS-stable aggregates of the protein takes place during protein inactivation. Instead, the inactivation process was found to be associated with the irreversible partial unfolding of the polypeptide chain, as assessed by Trpfluorescence, far UV circular dichroism, and 1-anilino-naphtalene-8-sulfonate binding. This inactive protein undergoes, in a later stage, a further irreversible transformation leading to large aggregates.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
Proteins solutions are frequently stored frozen assuming that the stability of the frozen state will be greater than that corresponding to the unfrozen state[1]. However, it is known that freezing and thawing of protein solutions can cause substantial loss of their biological activity
[2]. To date, various hypotheses have been posited concerning the cause of freeze damage in proteins: (i) protein unfolding at low temperature (cold denaturation), (ii) the effect of low activity and the peculiar properties of the liquid water in equilibrium with ice; (iii) protein– solutes and protein–protein interactions, elicited by freeze-induced concentration of solutes; and (iv) interactions between proteins and the surface of ice[3]. It has been described that protein–ice interactions are influenced by the number and size distribution of ice crystals which determine the total ice surface area and the size of inter-granular spaces
[4]. On the other hand, the texture of ice depends, among other factors, on the cooling rate and on duration of freezing; slow cooling and prolonged annealing promote the formation of fewer and larger ice crystals[2].
Although the deleterious effects of freezing on water-soluble proteins have been well established, there are only a few studies on the freeze–thawing damage of membrane proteins (e.g.[5]). This lack of information is surprising considering that about 25–30% of the total genomes code for membrane proteins[6]. Additionally, membrane proteins are of particular interest because they constitute the target for about 70% of current drugs[7]. The understanding of membrane protein freeze–thaw denaturation and its control would have numerous positive consequences, especially in thefield of cell preservation[8,9].
The plasma membrane calcium pump constitutes a good model for testing the stability of membrane proteins against freezing. It is an integral helical membrane protein consisting of a single polypeptide chain of Mr 134000[10]. A large intracellular domain includes the catalytic and regulatory sites[11], and short external loops connect ten transmembrane segments organized in three hydrophobic clusters
[12,13]. Thermal inactivation of PMCA was characterized as an irreversible process with the inactivation rate being a function of the micellar phase composition and the storage temperature[14,15]. The half life of the purified protein at room temperature is of about 20 days. Under native conditions, PMCA reversibly dimerizes[16,17], providing additional thermal stability to PMCA structure[18].
In this work we described that PMCA enzymatic activity significantly decreased after freeze–thawing. The inactivation was found to be associated with the irreversible partial unfolding of the polypeptide chain, as assessed by Trpfluorescence, far UV circular dichroism, and 1-aniline-naphtalene-8-sulfonate binding. This inactive protein under-goes a further irreversible transformation in later stages, leading to large protein aggregates.
Abbreviations: ANS, 1-aniline-naphtalene-8-sulfonate; C12E10, poly(oxyethylene)
10-lauryl ether; CD, circular dichroism; FT, freeze–thaw; PMCA, plasma membrane calcium pump
⁎ Corresponding authors. Tel.: +54 11 49648289; fax: +54 11 4962 5457. E-mail addresses:pablo.castello@colorado.edu(P.R. Castello),
lgf@qb.ffyb.uba.ar(F.L. González Flecha).
1
Present address: Department of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, CO 80309, USA.
0005-2736/$– see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2010.07.035
Contents lists available atScienceDirect
Biochimica et Biophysica Acta
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b b a m e m2. Materials and methods 2.1. Materials
All the chemicals used in this work were of analytical grade and mostly purchased from Sigma Chemical Co. pH and free Ca2+ concentration of the solutions were measured with a Corning pH/ion meter 450, using Orion ion-selective H+and Ca2+electrodes. 2.2. Purification of the plasma membrane Ca2+pump
Calmodulin-depleted erythrocyte membranes were prepared as described by Levi et al.[15]. The membrane suspension was stored at −80 °C until PMCA purification. Ca2+pump was isolated in pure form by affinity chromatography as described[18] and reconstituted in 120 mM KCl, 10 mM MOPS, 2 mM EDTA, 2 mM CaCl2, 2 mM DTT, 800μM C12E10and 260μM soybean phospholipids (pH 7.4 at 4 °C). Fractions exhibiting the highest specific Ca2+-ATPase activity were pooled. Integrity of purified Ca2+pump was veri
fied by SDS-PAGE (single band at Mr134,000). Ca2+pump total mass was determined by densitometric analysis after SDS-PAGE using bovine seroalbumin as standard. The molar concentration of PMCA was expressed as monomeric enzyme. Prior to use, the enzyme (580 nM, specific ATPase activity 11.5μmol Pi/mg min) was kept under liquid nitrogen. 2.3. Measurement of the Ca2+-ATPase activity
Ca2+dependent ATPase activity was measured at 310 K as the initial rate of release of Pi from ATP as described previously [18,19]. The incubation medium was: 1 nM PMCA, 120 mM KCl; 30 mM Tris; 4 mM MgCl2; 1 mM EGTA; 1.1 mM CaCl2; 140μM soybean phospholipids; 800μM C12E10and 2 mM ATP (pH 7.4 and [Ca2+]free= 140μM). 2.4. Polyacrylamide gel electrophoresis
PMCA samples (1–4 μg per well) were prepared for electrophoresis as described [13], including in the sample buffer 0.8 mg SDS and 0.2μmol DTT (SDS-PAGE) or 0.1 mg C12E10, 5μmol ε-aminocaproic acid (native PAGE). SDS-PAGE was carried out according to Schägger[20]
using 10% T and 1% C. PAGE using non-denaturing blue gels was performed as described by Wittig et al[21].
2.5. Fluorescence spectroscopy
Steady statefluorescence measurements were performed at 298 K in a 3 × 3 mm quartz cuvette using a SLM-Aminco Bowman Series 2 spectrofluorometer. Both excitation and emission bandwidths were set at 4 nm. PMCA emission spectrum was registered between 305 and 400 nm after excitation at 295 nm. ANSfluorescence was registered between 420 and 550 nm following excitation at 380 nm. The spectra were corrected for background emission. Total intensity (It) was calculated as:
It=∑iIλi ð1Þ
and the wavenumber (v) corresponding to the center of spectral mass_ was determined as reported[22]:
vCM= ∑ i I vð Þ:vi iΔvi ∑ i I vð ÞΔvi i ð2Þ 2.6. Light scattering
Rayleigh light scattering was measured at a 90° scattering angle using a SLM-Aminco Bowman Series 2 spectrofluorometer.
Measure-ments were performed at 298 K in a 3 × 3 mm cuvette, setting the excitation and emission wavelengths at 690 nm and the excitation and emission bandwidths at 4 nm. The relative Rayleigh ratio was calculated as described[23].
2.7. Circular dichroism
Circular dichroism spectra of PMCA were registered at 298 K in the wavelength region of 200–250 nm using a Jasco J-810 spectropolari-meter. Data were collected in a 1 mm path length cuvette using a scan speed of 20 nm/min with a time constant of 1 s. An average of three independent measurements was used to calculate the molar residue ellipticity [θ] as:
θ
½ =ðθ⋅100⋅MrÞ
c⋅l⋅NA
ð Þ ð3Þ
where [θ] is the mean residue molar ellipticity in deg cm2dmol−1,θ is the experimental ellipticity in millidegree, Mris the molecular mass of the protein, c is protein concentration in mg/ml, l is the cuvette path length in centimeters, and NAis the number of aminoacid residues of the protein (1205 for the h4b isoform of PMCA).
2.8. Data analysis
Data presented in this work are representative of at least two independent experiments. Measurements were performed in duplicate or triplicate. Equations werefitted to the experimental data using a non-linear regression procedure based on the Gauss–Newton algorithm[24]. The dependent variable was assumed to be homoscedastic (constant variance), and the independent variable was considered to have negligible error. Parameters were expressed as the mean± standard error.
3. Results
3.1. PMCA freeze–thaw inactivation
To evaluate the stability of the PMCA against freeze–thawing two cooling procedures were used. The first one was achieved by submerging the sample in liquid nitrogen (77 K) and the other by placing it in an adiabatic chamber at 253 K. The rates of cooling at the freezing point, measured as thefirst time-derivative of temperature, were 4 Ks−1(fast cooling) and 0.05 Ks−1(slow cooling) (Fig. S1 in the Supporting Material). Frozen samples were thawed by placing them in a water bath with agitation at 298 K, with a rate of warming at the freezing point of 0.2 Ks−1.
Wefirst explored the stability of frozen PMCA by measuring the ATPase activity after various storage-time periods. As can be seen in
Fig. 1, when purified PMCA was frozen and stored at 253 K the activity decreased exponentially with the storage time:
v vo
= e−ln2
t
t1=2 ð4Þ
where v and voare the remaining and initial ATPase activities, t is the storage time and t½the time at which v diminished to half its initial value. The bestfit value of t½was 3.9 ± 0.6 h. This experiment was repeated at other storage temperatures in the range 238–263 K, and no significant changes in t½were detected (results not shown).
On the other hand, when the enzyme was frozen and stored at 77 K the activity did not change over the 30 h time period assayed. Longer times were also assayed and no significant changes in activity were observed over a time period of at least 4 years. To further characterize the inactivation process, two additional experiments were done: In thefirst one a sample which had been frozen under the
fast cooling procedure was placed at 253 K, and it was observed that the ATPase activity decreased as a function of the storage time at 253 K, paralleling the behavior of samples frozen following the slow cooling procedure. In the second experiment, a sample which had been frozen under the slow cooling procedure was placed at 77 K. In this case, the ATPase activity remained constant at a value that was just below that obtained when the enzyme was frozen following the fast cooling procedure and stored at 77 K. In both, the inactivation rate depended only on the storage temperature, while the initial activity was affected by the freezing method.
Considering that the freezing point depression is 22.3 K for an aqueous solution in which the concentration of osmotically active particles is 12 molal, we investigated the effect of PMCA storage at 253 K in the presence of 6 m NaCl. As was expected, the protein solution did not freeze, and no inactivation was detected after an overnight incubation in this medium (data not shown). This result suggests the involvement of surface interactions between PMCA and ice crystals in the protein inactivation mechanism.
To further evaluate the stability of PMCA against freeze–thawing, the purified enzyme was successively frozen and thawed with an incubation step before thawing to allow temperature equilibration. After successive freeze–thaw cycles, the remaining Ca2+-ATPase activity was measured and plotted as a function of the number of cycles (Fig. 2).
It can be observed that in both experiments the enzyme activity exponentially decreased as a function of the number of FT cycles:
v vo
= e−ln2
n
n1=2 ð5Þ
where v and voare the remaining and initial ATPase activities, n is the number of FT cycles and n½the number of cycles where the activity v diminished to half its initial value. The best-fit values of n½were 3.7± 0.2 cycles in the case of slow cooling and 9.2 ±0.3 cycles when the fast cooling procedure was used. Reducing the inter-cycle equilibration time upon the slow cooling procedure resulted in an increased n½value (Fig. S2) because of the time dependent inactivation previously described. The observed exponential decrease in the activity suggests a two-state process involving only fully active and inactive molecules.
To explore the effect of the thawing step, the same experiment was repeated placing the samples frozen following the fast cooling procedure in a water/ice bath with agitation at 277 K. The warming rate at the freezing point was in this case 0.1 Ks−1. The enzyme stability
was only slightly lower than the control (Fig. S3) suggesting that the thawing step is not critical for the damage of PMCA.
3.2. Structural changes accompanying FT inactivation of PMCA It is known that the irreversible inactivation of soluble proteins can be due to covalent modifications or non-covalent changes such as aggregation or kinetically trapped conformations[25].
To evaluate the integrity of PMCA after freeze–thaw inactivation, SDS-PAGE analysis was performed (Fig. S4A). No changes were detected in the electrophoretic pattern and mobility of PMCA following freeze–thaw inactivation, suggesting that inactivation was not associated with either major fragmentation or significant formation of SDS stable protein aggregates. Analogous information was obtained from non-denaturing gels (Fig. S4B). Two well differentiated bands were observed, in agreement with previous reports indicating that native PMCA is composed of two different oligomeric states (monomer and dimmer)
[16,17]. Neither significant aggregation, nor changes in PMCA oligomeric distribution were associated with thermal inactivation. Interestingly however, we noticed a small decrease in the total amount of protein in the inactivated samples.
Far UV CD spectra of both native and FT inactivated PMCA were registered for determining whether or not FT inactivation of PMCA is accompanied by changes in protein secondary structure (Fig. 3). A significant decrease in the ellipticity was observed when PMCA was subjected to FT cycles, irrespective of the rate of cooling. This indicates a loss of secondary structure in the protein when it is inactivated. However, no significant differences in the CD spectra were observed between the native and the freeze–thawed PMCA stored 24 h at 253 K. To further explore the structural perturbations responsible for PMCA FT inactivation, two approaches which examine the status of hydro-phobic regions were used: ANS binding, and Trpfluorescence. The fluorescent probe 1-aniline-8-naphtalenesulfonate (ANS) is character-ized by a low quantum yield when it is exposed to water, and this yield strongly increases when the probe is located in hydrophobic protein pockets ([26]and references therein). Moreover, this probe is used extensively to detect molten globule like states during protein unfolding, including freeze-induced unfolding [3,27]. Fluorescence intensity of ANS bound to PMCA was registered for the native and FT inactivated protein (Fig. 4). This experiment shows that ANS fluores-cence intensity decreased when PMCA was FT inactivated, indicating the
Fig. 1. Time course of inactivation of the plasma membrane calcium pump after freezing. Purified plasma membrane calcium pump (100 nM, Ca2+
-ATPase activity 10 ± 2μmol Pi/ mg min) was frozen following slow cooling (○, □) or fast cooling (●,■). After 40 min samples were exchanged and kept in the adiabatic compartment at 253 K (○,●) or 77 K (□,■) and, at different storage times, they were thawed, and the Ca2+
-ATPase activity was measured. Relative ATPase activity was plotted as a function of the storage time. The lines are the graphical representation of Eq. (4) with the best-fit parameter value indicated in the text.
Fig. 2. Inactivation of the plasma membrane calcium pump after freeze–thaw cycles. Purified PMCA (100 nM, Ca2+-ATPase activity 10 ± 2 nmol Pi/μg min) was frozen
following the slow (○) or the fast cooling procedures (●). After thermal equilibration (3 min at 77 K and 40 min at 253 K), samples were thawed. This sequence was repeated and after each cycle an aliquot was taken, the Ca2+
-ATPase activity was measured and represented as a function of the number of FT cycles. Lines are the graphical representation of Eq. (5) with the best-fit parameter values indicated in the text.
loss of hydrophobic regions in the protein. No significant differences were found before 4.5 h incubation at 253 K, again suggesting that the irreversible structural changes are occuring during freeze–thaw events. Trpfluorescence spectrum of native PMCA excited at 295 nm is centered at 333 nm (Fig. S5), which is characteristic of Trp in a moderate hydrophobic environment such as that of a water-membrane interface. A slight red shift in the center of spectral mass (about 2 nm) was observed upon enzyme inactivation suggesting that Trp residues increased their solvent exposure. More evidently, Trp fluorescence intensity (I) decayed exponentially toward a constant value as a function of either the number of FT cycles (n) or the storage time at 253 K, as shown inFig. 5A and B.
I = I o I∞ e−ln2 x x1=2 + I ∞ ð6Þ
where Ioand I∞are the initial and remaining intensities and x may be the number of FT cycles (n) or the storage time (t). The best-fit values of n½ were 3.7 ± 0.3 (slow cooling) and 9.5± 0.8 (fast cooling), and the best-fit value of t½was 4 ± 2. These values are similar to that found for the
inactivation process (Figs. 1 and 2). 3.3. Aggregation of freeze–thaw inactivated PMCA molecules
In the previous section we mentioned that no protein aggregates were detected by SDS-PAGE or blue native PAGE analysis of FT inactivated PMCA samples. However, we considered that this approach may not be sensitive enough to detect low amounts of aggregated protein. The scattering of light by proteins is a widely used, highly sensitive method to evaluate protein aggregation. We determined the scattered light intensity of PMCA to be a function of the storage time and the number of FT cycles. Fig. 6A shows the scattering of light by PMCA samples as a function of the storage time at 253 K. This behavior can be described by an exponential function similar to Eq 6 (with x = t). The best-fit value of t½was 11.4 ± 0.9 h. Conversely, the observed dependence of the scattering signal on the number of FT cycles appears approximately linear (Fig. 6B). Consid-ering that this corresponds to the initial portion of an exponential function similar to Eq 6 (with x = n), the n½values were estimated as ≥23 in the case of slow cooling and ≥18 under fast cooling. These values are significantly higher than those corresponding to PMCA FT inactivation and the associated changes influorescence.
To estimate the degree of protein aggregation, we measured the protein concentration of inactivated samples before and after centrifuga-tion for 20 min at 14,000g. These sedimentacentrifuga-tion condicentrifuga-tions were selected
Fig. 4. ANS binding to native and FT inactivated PMCA. The purified enzyme (blue line), the enzyme frozen in liquid nitrogen and then subjected to 10 FT cycles following the fast cooling procedure (red line), 4 FT cycles following the slow cooling procedure (orange line) and stored 4.5 hours after the slow cooling procedure (green line), were supplemented with 1.5μM ANS (protein concentration=150 nM) and the fluores-cence spectra of the probe were recorded.
Fig. 5. Trpfluorescence of native and FT inactivated PMCA. (A) Purified enzyme was frozen following the slow cooling procedures. After different storage times, samples were thawed, and the Trp emission fluorescence spectra were recorded. Total fluorescence intensity (It) was calculated from each spectrum using Eq. (1) and plotted
as a function of storage time. (B) Purified enzyme was subjected to repeated freeze– thaw cycles following fast (●) or slow (○) cooling procedures. After each cycle aliquots were taken and the Trp emission fluorescence spectra were recorded. Total fluorescence intensity (It) was calculated from each spectrum using Eq. (1) and plotted
as a function of the number of cycles. The lines represent an exponential function decreasing towards a constant value (Eq.(6)) with the bestfit parameter values indicated in the text.
Fig. 3. Far UV-CD spectra of native and FT inactivated PMCA. The CD spectra were registered for the purified enzyme (blue line), the enzyme frozen in liquid nitrogen and then subjected to 20 FT cycles following the fast cooling procedure (red line), 10 FT cycles following the slow cooling procedure (orange line) and PMCA stored 24 hours after the slow cooling procedure (green line).
because they render a supernatant with a scattering signal similar to that corresponding to the native protein. After 10 h of storage at 253 K the protein concentration decreased 1.4%. After 5 FT cycles by slow cooling the observed decrease was 6.5%, and after 8 FT cycles of fast cooling there was a decrease of 5.2%. In all cases, inactive protein constituted more than 40% of the total protein sample. Considering both, the high estimated values of n½and t½, with the small proportion of aggregated protein indicate that aggregation is a later stage in the FT inactivation process.
Moreover, pellets obtained by sedimentation were collected, separated into two identical aliquots and analyzed by PAGE. One sample was run on SDS PAGE and appears as a set of oligomeric forms (Fig. S4A). The other, evaluated by blue native PAGE, did not run in the gel indicating that the pellet is composed by large protein aggregates that are partially solubilized by SDS.
4. Discussion
4.1. Long-term storage of micellar PMCA preparations.
It is known that membrane proteins need both an aqueous and a lipidic environment to display biological activity. Purification of these proteins requires the use of detergents to extract the protein from the
membrane and, at the end of these procedures they are often reconstituted in lipid vesicles or in mixed micelles of phospholipids and detergent. To obtain functionally active membrane proteins in a purification process, selection of the appropriate detergents and phospholipids is important. PMCA is purified from detergent-solubilized erythrocyte plasma membranes followed by reconstitu-tion in phospholipid-detergent mixed micelles at a mole fracreconstitu-tion of protein: amphiphiles at about 1:2000. The biochemical properties of the enzyme in this medium are identical to its properties in a natural membrane[28], making these preparations suitable for structural and functional studies. When purified PMCA is stored in a liquid state it spontaneously and irreversibly inactivates with the inactivation rate being a function of the storage temperature with t½= 13.3 h at 310 K, and an activation energy of 222 kJ/mol[14].
In this work we have shown that purified PMCA, frozen and stored at 253 K, inactivates as a function of the storage time, leading to a completely inactivated enzyme after 24 h of incubation. Considering that any reversible step occurring in the system during the storage at 253 K was reverted during the thawing step or during the preincubation for activity measurements, our results indicate that the observed PMCA inactivation was, at least partially, irreversible. The time course of this process is well described by an exponential function, a characteristic of first order reactions involving only fully active and inactive molecules. The t½for the inactivation process was 3.9 h, a value eight orders of magnitude lower than that estimated by extrapolation at 253 K from the Arrhenius plot characterizing the thermal inactivation of PMCA at non-freezing temperatures[14]. This result indicates that inactivation after freezing is not part of the thermal inactivation process previously described.
It was reported that frozen salt solutions are non-uniform media in which the concentration of solutes can vary from one region to another, depending on the ice growth rate and the direction of the thermal gradient[29]. Moreover, Dong et al. demonstrated that the microstruc-ture of frozen protein samples is time dependent, with ice domains coalescing and the freeze-concentrated liquid regions (which form narrow channels and lagunae devoid of ice) decreasing in number but increasing in size[30]. In this context, we can assume that when PMCA samples are brought to 253 K, water will freeze out until the concentration of osmotically active particles in the liquid phase reaches about 12 molal, and PMCA will remain in this liquid concentrated phase in contact with ice crystals. Our results showed that PMCA storage at 253 K in this liquid concentrated medium without ice did not produce PMCA inactivation. Thus, surface interactions of the protein with ice crystals and/or the high pressure stress produced by ice growing, which was proposed to be enough to deform the crystal structure of hexagonal ice[31], are probably involved in the freeze-induced protein inactiva-tion mechanism.
On the other hand, when the enzyme is frozen and stored under liquid nitrogen, no changes in the enzyme activity were detected over time periods of years. At this temperature all the sample is frozen either as cubic ice crystals (ice Ic) [32] or as eutectic water salt mixtures[33], or it could form a glass phase[2,32]. Nevertheless, a water/protein weight fraction of 0.3–0.4 remains mobile and does not crystallize even at this very low temperature during long periods[34]. We also demonstrate that inactivation was only dependent on the final storage conditions, and that the freezing step only affected the early stages of the inactivation curve. This results from a structural rearrange-ment of the frozen samples when they are annealed at different sub-freeze temperatures. When samples frozen upon fast cooling were warmed from 77 K to 253 K, the freezable water fraction starts to transform into hexagonal ice at around 240 K[35], and the eutectic mixtures melt. In this scenario, the small ice crystals in contact with this aqueous phase re-crystallize generating large ice crystals[36]. Thefinal structure of the system will be similar to that reached following the slow cooling scheme to 253 K. Thus, after this initial stage, PMCA inactivates following the same kinetics in both conditions.
Fig. 6. Light scattering analysis of native and FT inactivated PMCA. (A) Purified PMCA was fractioned and aliquots were frozen following the slow cooling procedure. At selected times, they were thawed, and the relative Rayleigh ratio was calculated and plotted as a function of the storage time. (B) Another aliquots of the purified enzyme were frozen following the slow (○) or the fast (●) cooling procedures. After thermal equilibration, samples were thawed by incubation at 298 K with continuous stirring. This sequence was repeated and after each cycle an aliquot was taken and the scattering intensity was registered. The relative Rayleigh ratio calculated and plotted as a function of the number of FT cycles. Lines are the graphical representation of Eq. (6) with the parameter values indicated in the text.
4.2. Inactivation of PMCA by freeze–thaw cycles
Successive freeze–thaw cycles leads to decreases in enzyme activity, following an exponential function of the number of cycles. This behavior was observed using either the slow or fast cooling scheme for freezing, followed by the short storage time between freezing and thawing to allow thermal equilibration of the samples. It is worth noting that extrapolation to an infinite number of freeze– thaw steps leads to completely inactive protein. Again, the exponen-tial form of the inactivation curve is indicative of a two state processes involving only active and inactive molecules. The parameter n½ obtained by fitting the two-state model to the experimental inactivation data gives an accurate measure of the enzyme stability against freezing-thawing.
As was mentioned before, one FT cycle using the fast cooling scheme did not produce significant changes in enzyme activity, however, repeated FT cycles led to irreversible inactivation. This effect must be produced during the freezing or the thawing step, since in this case there are no effects of the storage time on the enzyme activity. Since only small changes in the enzyme stability were detected when the warming rate was changed in the thawing procedure, it can be proposed that freezing the samples is the main source of PMCA damage.
On the other hand, successive freeze–thaw cycles following the slow cooling procedure also inactivated PMCA, but the n½was smaller than that corresponding to FT cycles following the fast cooling procedure. In this case there is a combined effect of the liquid/solid phase transitions and the storage time between freezing and thawing. It is not easy to split both effects, because when shortening the storage time, the enzyme would not freeze. However, by using very small enzyme aliquots it is possible to reduce the time needed for thermal equilibration to reach a value at which no significant time dependent inactivation was detected. In these conditions PMCA appears to be more stable against FT cycles using the slow cooling procedure compared to the fast cooling procedure. This result is in accordance with similar ones performed with soluble protein solutions which showed that fast freezing produced more damage to proteins and gave lower recovery of activity after freezing and thawing[4,37,38], however in PMCA this effect is compensated for by the absence of time dependent inactivation.
It is worthy to mention that purified preparations of membrane protein always include surfactants, and it has been described that these substances can protect some proteins from surface inactivation during freezing[37]or possibly work adversely and enhance freeze induced damage[39]. We can not discard a protective effect of the amphiphiles against FT inactivation in membrane protein prepara-tions, in fact preliminary results indicate that lowering the total amphiphile concentration produce a significant decrease in the n½ obtained following the fast cooling freezing procedure (Fig. S6). 4.3. Freeze–thawing produces the partial unfolding of PMCA
Irreversible inactivation of soluble proteins has been extensively studied. In many cases it is due to either covalent modifications (side-chain oxidation, hydrolysis of peptide bonds, etc) or non-covalent changes leading to aggregation [25]. PAGE analysis indicates that PMCA freeze–thaw inactivation is apparently not associated with major chain modifications and thus small structural perturbations, such as kinetically trapped non native conformations[40], could be involved. To explore this, a combined approach reporting different levels of PMCA structure was used.
The circular dichroism spectrum in the far UV region gives information about the secondary structure of proteins. We have demonstrated that after inactivation by FT cycles a significant amount of secondary structure of PMCA is lost, although the inactive state of PMCA still retains considerable structure. The mean residue molar
ellipticity at 222 nm, often considered indicative of alpha helix structure, significantly decreased on those samples after inactivation by FT cycles. Conversely, no changes on the secondary structure of PMCA after incubating the enzyme for 24 h at 253 K could be detected. In all the cases the observed changes correspond to the irreversible component of the whole change, since all possible reversible alterations disappear when samples were thawed.
It was reported for freeze-labile soluble protein solutions, that the infrared spectrum before and after freezing and thawing were indistinguishable[41]. However, recent studies using infrared and Raman microscopy on frozen protein solutions allowed to sense changes in the secondary structure between the liquid and the frozen state[30,41,42]. A close inspection of frozen samples showed that the spectra collected in the interstitial space, distant from the surface of ice crystals, were very similar to spectra collected from the initial solution, whereas through ice crystals the spectrum shows an increase in bands characteristic of intermolecularβ-sheet structures[30,42]. The addition of surfactant to protein preparations prior to freezing resulted in a decrease in intermolecularβ-sheet signals in spectra of the proteins on the ice crystal surface, and much of the native state structure was recovered[42]. As was mentioned before, we can not discard a similar effect on PMCA preparations, so the actual structural change upon freezing could be much higher than those reported in this work.
To further investigate the structural perturbations responsible for PMCA freeze–thaw inactivation we analyzed the PMCA Trp fluorescence spectrum and the binding of thefluorescent probe ANS. The isoform hPMCA4b -the main isoform of the calcium pump found in human erythrocytes- possesses 12 tryptophan residues[10]. These residues are within membrane-related regions as was demonstrated by quenching experiments with a pyrene labeled phosphatidylcholine[43]. PMCA inactivation by FT cycles and long term storage at 253 K produces a decrease in Trpfluorescence intensity, and additionally a small red shift of the spectrum. Bothfindings suggest that these processes lead to protein conformations with relatively higher exposition of Trp residues to the aqueous medium. The totalfluorescence intensity exponentially decreased with the number of FT cycles, and the obtained values of n½ were similar to those found for the inactivation, suggesting that PMCA partial unfolding and FT inactivation are closely related processes. Also in this case the observed changes correspond to the irreversible changes since the reversible ones were reverted after thawing. In a similar way Strambini et al have demonstrated that the solidification of water leads to important perturbations of the globular fold in all the water soluble proteins they have examined with a broad distribution of protein conformations in the frozen medium[4]. These authors found an inverse correlation between the extent of the perturbation and the fraction of liquid water in equilibrium with ice suggesting that the perturbation may derive from the adsorption of the macromolecule to the liquid/ solid interface[3].
The fluorescent probe ANS is frequently used to monitor conformational changes in proteins, being the ANS fluorescence intensity at high probe:protein mole ratios usually taken as a measure of the size of water accessible hydrophobic regions in the protein[26]. Only few native proteins bind ANS in hydrophobic pockets, among these are membrane proteins for which ANS appears to be a reporter of the hydrophobic transmembrane domain[19,44]. The incubation of native and FT inactivated PMCA with ANS indicated that the extent of hydrophobic pockets in the protein diminished as the inactivation took place, supporting the idea that PMCA inactivation is associated to a slight exposition of hydrophobic transmembrane regions to the solvent.
In addition, blue native PAGE analysis showed that purified PMCA exists as two oligomeric states: a more abundant monomeric form and a dimer. There was not change in the distribution of oligomeric forms of PMCA after FT inactivation, suggesting that the quaternary structure is not significantly affected.
4.4. Membrane protein aggregation: a later stage of PMCA FT inactivation Protein aggregation is a common and troubling manifestation of protein instability which may be induced by a variety of physical factors, such as temperature, ionic strength, vortexing or time[25].
Although we were not able to detect protein aggregates in FT inactivated samples by denaturing or blue native gels, the significant increase in the scattering of light by the samples as a function of the incubation time at 253 K or FT cycles clearly indicates the formation of aggregates. Moreover, PAGE analysis of the pellet obtained after centrifugation PMCA samples presenting significant light scattering indicates that FT inactivation of this membrane protein leads to the formation of macromolecular aggregates which are insoluble in water but partially soluble in SDS.
Nevertheless, the t1/2and n1/2values estimated for the increasing of light scattering are significantly higher than those found for the inactivation process, and the decrease in protein concentration before centrifugation is much lower that the loss of enzymatic activity. These results indicate that PMCA aggregation is not the cause of FT inactivation, but a later stage of this process.
5. Concluding remarks
In this work we have characterized the freeze–thaw inactivation of the helical membrane protein PMCA. The exponential dependence of the decrease of the ATPase activity on the storage time and on the number of freeze–thaw cycles gives account of a two-state process involving only active and inactive molecules.
The inactivation was found to be associated with the irreversible partial unfolding of the polypeptide chain, as assessed by Trp fluorescence, far UV circular dichroism, and 1-aniline-naphtalene-8-sulfonate binding. Thesefindings are all indicative of protein denatur-ation with loss of secondary structure and exposure of hydrophobic residues to the aqueous solvent. However, it is also evident that a large portion of structure is still conserved in the inactive state. The observed changes were particularly evident when PMCA was inactivated by FT cycles, whereas after inactivation by long term storage at 253 K, only a small decrease in Trpfluorescence was evident. This inactive protein
undergoes a further irreversible transformation in later stages, leading to large protein aggregates.
Taking together these results point out to a putative inactivation mechanism as that represented in Fig. 7. When native enzyme solution is frozen the protein -as well as the other solutes- isfirst excluded from the ice crystals, thus remaining in a liquid concentrated phase at low temperature (I). The protein in this medium reaches an adsorption equilibrium on the surface of ice (II). The adsorbed molecules undergo a reversible partial unfolding driven by surface interactions or the high-pressure stress operating on the ice surface (III). This unfolded state is composed by a collection of unfolded conformations, some of these can be kinetically trapped (IV). Upon thawing, all the reversible transformations (I, II and III) revert, resulting in active molecules. Conversely, those molecules that were in the kinetically trapped partially unfolded state (IV) will appear in solution as inactive protein molecules (V). These species are able to aggregate in a slow step (VI), being the aggregates also inactive. Successive freeze–thaw cycles multiply this effect and, because the same fraction of active molecules will remain trapped in step IV during each cycle, an exponential decrease of the amount of active enzyme with the number of cycles is expected.
Acknowledgments
The authors thank Jay Wojcik for careful reading of the manuscript. This work was supported by grants from ANPCyT PICT2006 01741 and UBACyT B110.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, atdoi:10.1016/j.bbamem.2010.07.035.
References
[1] K.A. Pikal-Cleland, N. Rodriguez-Hornedo, G.L. Amidon, J.F. Carpenter, Protein denaturation during freezing and thawing in phosphate buffer systems: monomeric and tetrameric beta-galactosidase, Arch. Biochem. Biophys. 384 (2000) 398–406.
Fig. 7. Inactivation mechanism and structural changes produced in PMCA during freeze–thawing and frozen storage. N: native protein conformation, I: partially unfolded protein intermediate, A: aggregation competent intermediate, An: protein aggregates composed by n molecules of A. The subscripts indicate the phase in which the species is: (sc) solution,
[2] F. Franks, Biophysics and Biochemistry at Low Temperatures, Cambridge University Press, New York, 1985.
[3] E. Gabellieri, G.B. Strambini, Perturbation of protein tertiary structure in frozen solutions revealed by 1-anilino-8-naphthalene sulfonatefluorescence, Biophys. J. 85 (2003) 3214–3220.
[4] G.B. Strambini, E. Gabellieri, Proteins in frozen solutions: evidence of ice-induced partial unfolding, Biophys. J. 70 (1996) 971–976.
[5] R.J. Lalonde, J.R. Lepock, J. Kruuv, Site of freeze–thaw damage and cryoprotection by amino acids of the calcium ATPase of sarcoplasmic reticulum, Biochim. Biophys. Acta 1079 (1991) 128–138.
[6] E. Wallin, G.V. Heijne, Genome-wide analysis of integral membrane proteins from eubacterial, archaean, and eukaryotic organisms, Protein Sci. 7 (1998) 1029–1038. [7] M.A. Yildirim, K.-I. Goh, M.E. Cusick, A.-L. Barabasi, M. Vidal, Drug-target network,
Nat. Biotechnol. 25 (2007) 1119–1126.
[8] J. Gearhart, New Potential for Human Embryonic Stem Cells, Science 282 (1998) 1061–1062.
[9] Y. Li, J.C. Tan, L.S. Li, Comparison of three methods for cryopreservation of human embryonic stem cells, Fertil. Steril. 93 (2010) 999–1005.
[10] E.E. Strehler, P. James, R. Fischer, R. Heim, T. Vorherr, A.G. Filoteo, J.T. Penniston, E. Carafoli, Peptide sequence analysis and molecular cloning reveal two calcium pump isoforms in the human erythrocyte membrane, J. Biol. Chem. 265 (1990) 2835–2842.
[11] F. Di Leva, T. Domi, L. Fedrizzi, D. Lim, E. Carafoli, The plasma membrane Ca2+ATPase
of animal cells: structure, function and regulation, Arch. Biochem. Biophys. 476 (2008) 65–74.
[12] P.R. Castello, A.J. Caride, F.L. González Flecha, H.N. Fernández, J.P. Rossi, J.M. Delfino, Identification of transmembrane domains of the red cell calcium pump with a new photoactivatable phospholipidic probe, Biochem. Biophys. Res. Commun. 201 (1994) 194–200.
[13] P.R. Castello, F.L. González Flecha, A.J. Caride, H.N. Fernández, J.M. Delfino, J.P. Rossi, The membrane topology of the amino-terminal domain of the red cell calcium pump, Protein Sci. 6 (1997) 1708–1717.
[14] V. Levi, J.P.F.C. Rossi, M.M. Echarte, P.R. Castello, F.L. González Flecha, Thermal stability of the plasma membrane calcium pump. quantitative analysis of its dependence on lipid-protein interactions, J. Membr. Biol. 173 (2000) 215–225. [15] V. Levi, A.M. Villamil Giraldo, P.R. Castello, J.P. Rossi, F.L. González Flecha, Effects of
phosphatidylethanolamine glycation on lipid-protein interactions and membrane protein thermal stability, Biochem. J. 416 (2008) 145–152.
[16] D. Kosk-Kosicka, T. Bzdega, A. Wawrzynow, Fluorescence energy transfer studies of purified erythrocyte Ca2+-ATPase. Ca2+-regulated activation by oligomerization, J.
Biol. Chem. 264 (1989) 19495–19499.
[17] V. Levi, J.P. Rossi, P.R. Castello, F.L. González Flecha, Oligomerization of the plasma membrane calcium pump involves two regions with different thermal stability, FEBS Lett. 483 (2000) 99–103.
[18] V. Levi, J.P. Rossi, P.R. Castello, F.L. González Flecha, Structural significance of the plasma membrane calcium pump oligomerization, Biophys. J. 82 (2002) 437–446. [19] D.I. Cattoni, F.L. González Flecha, J.M. Argüello, Thermal stability of CopA, a polytopic
membrane protein from the hyperthermophile Archaeoglobus fulgidus, Arch. Biochem. Biophys. 471 (2008) 198–206.
[20] H. Schägger, Tricine-SDS-PAGE, Nat. Protoc. 1 (2006) 16–22.
[21] I. Wittig, H.P. Braun, H. Schagger, Blue native PAGE, Nat. Protoc. 1 (2006) 418–428. [22] V. Levi, F.L. González Flecha, Reversible fast-dimerization of bovine serum albumin detected byfluorescence resonance energy transfer, Biochim. Biophys. Acta 1599 (2002) 141–148.
[23] A. Paliwal, D. Asthagiri, D. Abras, A.M. Lenhoff, M.E. Paulaitis, Light-scattering studies of protein solutions: role of hydration in weak protein–protein interactions, Biophys. J. 89 (2005) 1564–1573.
[24] G.A.F. Seber, C.J. Wild, Nonlinear Regression, John Wiley and Sons, New York, 1989. [25] E.Y. Chi, S. Krishnan, T.W. Randolph, J.F. Carpenter, Physical stability of proteins in aqueous solution: mechanism and driving forces in nonnative protein aggregation, Pharm. Res. 20 (2003) 1325–1336.
[26] D.I. Cattoni, S.B. Kaufman, F.L. González Flecha, Kinetics and thermodynamics of the interaction of 1-anilino-naphthalene-8-sulfonate with proteins, Biochim. Biophys. Acta 1794 (2009) 1700–1708.
[27] E. Gabellieri, G.B. Strambini, ANSfluorescence detects widespread perturbations of protein tertiary structure in ice, Biophys. J. 90 (2006) 3239–3245.
[28] E. Carafoli, Biogenesis: plasma membrane calcium ATPase: 15 years of work on the purified enzyme, FASEB J. 8 (1994) 993–1002.
[29] G.B. Strambini, M. Gonnelli, Protein stability in ice, Biophys. J. 92 (2007) 2131–2138. [30] J. Dong, A. Hubel, J.C. Bischof, A. Aksan, Freezing-induced phase separation and spatial microheterogeneity in protein solutions, J. Phys. Chem. B 113 (2009) 10081–10087.
[31] D.B. Varshney, J.A. Elliott, L.A. Gatlin, S. Kumar, R. Suryanarayanan, E.Y. Shalaev, Synchrotron X-ray diffraction investigation of the anomalous behavior of ice during freezing of aqueous systems, J. Phys. Chem. B 113 (2009) 6177–6182.
[32] B. Luyet, On various phase transitions occurring in aqueous solutions at low temperatures, Ann. NY Acad. Sci. 85 (1960) 549–569.
[33] L.K. Rey, Thermal analysis of eutectics in freezing solutions, Ann. NY Acad. Sci. 85 (1960) 510–534.
[34] I.D. Kuntz, W. Kauzmann, Hydration of proteins and polypeptides, Adv. Prot. Chem. 28 (1974) 239–345.
[35] G. Sartor, E. Mayer, Calorimetric study of crystal growth of ice in hydrated methemoglobin and of redistribution of the water clusters formed on melting the ice, Biophys. J. 67 (1994) 1724–1732.
[36] E. Cao, Y. Chen, Z. Cui, P.R. Foster, Effect of freezing and thawing rates on denaturation of proteins in aqueous solutions, Biotechnol. Bioeng. 82 (2003) 684–690. [37] B.S. Chang, B.S. Kendrick, J.F. Carpenter, Surface-induced denaturation of proteins
during freezing and its inhibition by surfactants, J. Pharm. Sci. 85 (1996) 1325–1330. [38] S. Jiang, S.L. Nail, Effect of process conditions on recovery of protein activity after
freezing and freeze-drying, Eur. J. Pharm. Biopharm. 45 (1998) 249–257. [39] T.J. Anchordoquy, K.I. Izutsu, T.W. Randolph, J.F. Carpenter, Maintenance of
quaternary structure in the frozen state stabilizes lactate dehydrogenase during freeze-drying, Arch. Biochem. Biophys. 390 (2001) 35–41.
[40] Y. Zhou, F.W. Lau, S. Nauli, D. Yang, J.U. Bowie, Inactivation mechanism of the membrane protein diacylglycerol kinase in detergent solution, Protein Sci. 10 (2001) 378–383.
[41] J.J. Schwegman, J.F. Carpenter, S.L. Nail, Infrared microscopy for in situ measurement of protein secondary structure during freezing and freeze-drying, J. Pharm. Sci. 96 (2007) 179–195.
[42] J.J. Schwegman, J.F. Carpenter, S.L. Nail, Evidence of partial unfolding of proteins at the ice/freeze-concentrate interface by infrared microscopy, J. Pharm. Sci. 98 (2009) 3239–3246.
[43] V. Levi, J.P.F.C. Rossi, P.R. Castello, F.L. González Flecha, Quantitative analysis of membrane protein–amphiphile interactions using resonance energy transfer, Anal. Biochem. 317 (2003) 171–179.
[44] E.A. Roman, J.M. Argüello, F.L. González Flecha, Reversible unfolding of a thermophilic membrane protein in phospholipid/detergent mixed micelles, J. Mol. Biol. 397 (2010) 550–559.
