- Email: [email protected]
Thermal Stability of Horse Spleen Apoferritin and Human Recombinant H Apoferritin
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS
Vol. 325, No. 1, January 1, pp. 58–64, 1996 Article No. 0007
Thermal Stability of Horse Spleen Apoferritin and Human Recombinant H Apoferritin Simonetta Stefanini, Stefano Cavallo, Chang-Qing Wang, Paola Tataseo,* Paola Vecchini, Anna Giartosio, and Emilia Chiancone1 CNR Center of Molecular Biology, Department of Biochemical Sciences ‘‘A. Rossi Fanelli,’’ University La Sapienza, 00185 Rome; and *Department of Biology, University Tor Vergata, 00173 Rome, Italy
Received June 21, 1995, and in revised form September 14, 1995
The thermal stability of horse spleen apoferritin, a heteropolymer composed of 90% L and 10% H chains, has been studied by differential scanning calorimetry and compared with that of the human recombinant H homopolymer. The denaturation temperatures (Tm) are significantly higher for the horse spleen polymer than for the recombinant protein under all experimental conditions (e.g., at pH 7, Tm values are §93 and 777C, respectively). The thermal denaturation process displays substantial reversibility for both polymers up to a few degrees below Tm , as indicated by CD measurements in the far and near uv regions. At temperatures higher than Tm the thermograms are influenced by the exothermic contribution of aggregation and/or precipitation. The H homopolymer thermogram, which is not distorted by the exotherm, is consistent with a multistate denaturation process. Acid dissociation of apoferritin produces stable dimeric subunits. The thermal unfolding of both dimeric subunits is reversible at least up to Tm and is characterized by an inversion of stability relative to the polymers (at pH 3.5, Tm is 427C for the horse spleen and 507C for the H subunit). These results indicate that the stabilization of the polymeric structure arises mainly from interactions between dimers, in accordance with the crystallographic evidence that the dimers are the building blocks of the polymeric molecule. q 1996 Academic Press, Inc. Key Words: apoferritin; differential scanning calorimetry; thermal stability; circular dichroism; sedimentation velocity.
Ferritin, the ubiquitous iron storage protein, is characterized by an unusually high thermal stability, a 1 To whom correspondence should be addressed at CNR Center of Molecular Biology c/o Department of Biochemical Sciences ‘‘A. Rossi Fanelli,’’ University La Sapienza, Piazzale A. Moro 5, 00185 Rome, Italy. Fax: 39-6-4440062.
property which is exploited in the purification procedure of the protein. Natural apoferritin is a heteropolymer composed of L and H subunits which coassemble into a 24-mer shell with a large cavity harboring up to 4500 iron atoms. The H and L chains share the same tertiary structure which is characterized by a bundle of four antiparallel helices (A–D), an additional short helix (E), and a loop connecting helices B and C. By interdigitation of sidechains in the loop region the subunits give rise to dimeric building blocks which generate a molecular assembly with 432 symmetry and leave intersubunit channels along the three- and fourfold axes for the passage of iron and/or small molecules (1, 2). The study of H and L homopolymers obtained by means of recombinant DNA technology brought out clearly that the two chain types differ from a functional viewpoint. Thus, the H chains contain a ferroxidase center in the four-helix bundle which is missing in the L chains (2). Accordingly, the H chains have been proposed to promote iron oxidation, whereas the L chains are thought to play a role in iron accumulation (3, 4). From a structural viewpoint the L chains confer to the assembled molecule a greater stability toward chemical and physical agents than do the H chains. In fact, the L-rich apoferritin from horse spleen (90% L) denatures only under extreme conditions, namely at temperatures higher than 807C and at pH values below 2.0 or above 10.5 (5). The stability conveyed by the L chains can be ascribed in part to residues located at the intersubunit contacts along the three- and fourfold channels since H homopolymers become more stable if residues along either type of channel are substituted with L-chain residues (6). In addition, a special role has been attributed to a specific salt bridge within the four-helix bundle between Lys62 and Glu107 (6); the corresponding residues in the H chain, Glu62 and Glu107, are part of the ferroxidase center (2). This paper reports the first direct measurement by
58
0003-9861/96 $12.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
/ m4606$9213
11-30-95 16:04:27
arca
AP: Archives
CALORIMETRIC STUDY OF APOFERRITINS
59
means of differential scanning calorimetry (DSC)2 of the thermal stability for horse spleen apoferritin, the H homopolymer, and their dimeric subunits. Horse spleen apoferritin is characterized by a markedly higher denaturation temperature than the H homopolymer; in contrast the H dimeric subunits are more stable than the L subunits. The DSC data have been supplemented by far and near uv–CD measurements and sedimentation velocity experiments in order to monitor the extent of unfolding and/or dissociation of the protein. MATERIALS AND METHODS Horse spleen ferritin (90% L; HoS) was prepared as described previously (7). Human recombinant H ferritin (100% H subunit; rH) was overexpressed in Escherichia coli and purified essentially as described by Levi et al. (3). Apoferritin was usually obtained from horse spleen ferritin by reduction of iron with thioglycollic acid and chelation with 2,2*-bipyridyl. In the case of rH ferritin, iron was removed in anaerobiosis by dialysis against 50 mM MES–NaOH buffer at pH 6.0 containing 0.3% sodium dithionite and 2,2*-bipyridyl. When this latter procedure was applied to horse spleen ferritin, the resulting apoprotein was found to be identical to that obtained with thioglycollic acid. Apoferritin concentration was calculated from the absorbance at 280 nm using the extinction coefficient e 1%, 1 cm Å 9.0. Calorimetric experiments were performed with a high-resolution MicroCal MC-2 differential scanning calorimeter (MicroCal, Inc., Northampton, MA) equipped with a DA-2 digital data acquisition system. Samples were always heated at a scan rate of 607C/h. The concentration of the protein solutions was 2–3 mg/ml. All curves were corrected for the instrumental baseline obtained by filling both cells with the buffer used. Integration and deconvolution were performed using the ORIGIN software provided by Microcal, Inc., after normalizing the data for protein concentration and subtracting as baseline the straight line connecting the initial and the final temperatures of the overall transition in the case of the rH polymer, or a sigmoidal progress curve in the case of the subunits (8, 9). The values of Tm (temperature at which excess heat capacity reaches a maximum; 7C), Dhc (specific enthalpy change; cal/g), DHc (calorimetric enthalpy change; kcal/mol), and DCp (molar heat capacity change; cal/g 7C) were obtained by analysis of the resulting thermal profiles. DHc was calculated for the rH homopolymer on the basis of Mr 504 kDa and for the dimeric subunits on the basis of Mr 2 1 19.5 and 2 1 21.0 kDa for the horse spleen and recombinant protein, respectively. The van’t Hoff enthalpy change (DHvH ; kcal/mol) was calculated according to the formula 2 DHvH Å A RT 1/2 rcex,1/2/Dhcal
where cex,1/2 (Cp at Tm in our notation) is the excess heat capacity at T1/2 (Tm in our notation), i.e., the distance from the baseline to the top of the transition; Dhcal corresponds to Dhc in our notation; and A Å 4 for a nondissociating system undergoing a two-state transition (9). All values reported in calories can be converted into joules multiplying by the factor 4.184. Circular dichroism measurements were carried out using a Jasco710 A spectropolarimeter. In the experiments designed to monitor
2 Abbreviations used: HoS, horse spleen ferritin; rH, human recombinant ferritin; DSC, differential scanning calorimetry; Gu-HCl, guanidine–HCl; MES, 4-morpholineethanesulfonic acid.
/ m4606$9213
11-30-95 16:04:27
arca
FIG. 1. DSC profiles of horse spleen and rH apoferritin at neutral and acid pH in the absence and presence of 2 M guanidine–HCl. Buffers, 50 mM sodium phosphate at pH 7.0 and 40 mM glycine–HCl at pH 4.0. Apoferritin, horse spleen (—), rH (rrr). In this and the following DSC figures the curves have been arbitrarily shifted on the y axis for clarity.
the effect of high temperatures on unfolding process, a CD spectrum was first measured at 207C, the protein was then incubated in thermostatted cells at the desired temperature for 5 min, and a second CD spectrum was acquired in either 5 or 20 min; thereafter the sample was cooled back to 207C and a last CD spectrum was measured. The acquisition time varied between 5 and 20 min. In all other experiments the spectra were recorded at 207C. The molar ellipticity (deg r cm2 r dmol01) was calculated in the far uv region ([U]aa) on the basis of a mean residue molecular weight of 113 and in the near uv region ([U]Mr) on the basis of the polypeptide chain molecular weight 19.5 and 21.0 kDa for horse spleen and rH apoferritin, respectively. Sedimentation velocity experiments were carried out at a temperature of 207C in a Beckman Optima XL-A analytical ultracentrifuge. The sedimentation coefficients were reduced to s20 by standard procedures.
RESULTS
Thermal Stability of the Polymers The thermal stability of horse spleen and human recombinant apoferritins has been studied under a number of experimental conditions. In 50 mM phosphate buffer at pH 7.0 the calorimetric profile of horse spleen apoferritin shows a single peak characterized by a Tm value of about 937C (Fig. 1); at higher temperatures aggregation and precipitation take place. Heating to 897C in a first scan does not greatly affect the calorimetric pattern in a subsequent heating cycle, suggesting that the system shows substantial reversibility until aggregation occurs. In order to further investigate the extent to which the DSC data reflect reversible unfolding and/or dissociation reactions, the calorimetric experiments were supplemented with CD and sedimentation velocity experiments. Far and near uv measure-
AP: Archives
60
STEFANINI ET AL.
FIG. 2. Effect of temperature on the (A, B, C, D) far and (A*, B*, C*, D*) near uv circular dichroism spectra of horse spleen and rH apoferritins. Each graph shows the spectra acquired at 207C (—), at high temperature (rrr), and after cooling at 207C (— —). The high temperature corresponds to (A, A*) 767C, (B, B*) 817C, (C, C*) 867C, (D, D*) 727C. The samples were incubated at the desired temperature for 5 min and the CD spectra were acquired in 20 min; thereafter the samples were cooled back to 207C and the CD spectra were measured again. Protein concentration: (A, B, C, D) 0.12 mg/ml and (A*, B*, C*, D*) 3 mg/ml in 50 mM phosphate buffer at pH 7.0.
ments were carried out between 767C, which corresponds to the onset of the thermal transition, and 867C, a few degrees below the Tm value (Figs. 2A–A* and 2C–C*). Over this temperature range no precipitation occurred and the ellipticity was independent of acquisition time (from 5 to 20 min). The ellipticity decreases progressively with increase in temperature in both spectral regions and, at any given temperature, the effect is more marked in the near uv region than in the far uv. In the sample heated to 867C, for example, the molar ellipticity at 222 nm, where the secondary structure is monitored ([U]aa), decreases by 30%, whereas it decreases by 70% at 292 nm, where the environment of the single tryptophan residue in the loop region at the dimer interface is monitored (10). The decrease in ellipticity is essentially reversible in the sample heated to 767C (Fig. 2A–A*). In contrast, the recovery of the initial ellipticity is only partial in the samples heated to higher temperatures. Moreover, the extent of recovery differs in the far and near uv regions, being larger in the former (Figs. 2B–B* and 2C–C*). In order to establish whether significant changes in quaternary structure occur in horse spleen apoferritin incubated at 867C, a sample was kept at this temperature for 25 min,
/ m4606$9213
11-30-95 16:04:27
arca
brought back to 207C, and analyzed by sedimentation velocity experiments. The sedimentation velocity patterns show, in addition to the native polymeric structure, dissociated material (10%) and aggregated protein (20%) consistently with the partial loss of secondary and tertiary structure. The calorimetric profile of rH apoferritin at pH 7.0 displays a single transition characterized by a Tm value of 777C; precipitation takes place only after completion of the thermal transition, thus allowing determination of the calorimetric enthalpy change. The DHc value is much higher than DH£H (Table I) and there is no appreciable DCp . CD spectra were used to assess the extent of temperature-induced unfolding and/or dissociation in rH apoferritin. Due to the small amount of material available the measurements were performed only at 727C, a temperature 57C lower than the Tm value. As in the case of horse spleen apoferritin, incubation at high temperature induces a loss of ellipticity at 222 nm which corresponds to 35% of the initial value, while the effect is more marked in the near uv region where the peak at 294 nm is abolished. The ellipticity is partially recovered after cooling and the recovery is more marked at
AP: Archives
CALORIMETRIC STUDY OF APOFERRITINS TABLE I
Calorimetric Data of rH Apoferritina
Solvent pH 7.0 pH 7.0 2 M Gu–HCl a
Tm (7C)
Dhc (cal/g)
DHc (kcal/mol)
DHvH (kcal/mol)
77.3
4.2
2127
51.7
59.8
4.7
2369
42.1
For details see Materials and Methods.
222 than at 294 nm (Fig. 2D–D*). Sedimentation velocity experiments indicate that the polymeric structure of rH apoferritin is unaltered after incubation at 727C for 25 min. In further experiments the native structure of the two apoferritins was mildly destabilized by means of chemical agents like guanidine and/or low pH. The Tm values indicate (Fig. 1) that addition of 2 M guanidine– HCl to the pH 7.0 phosphate buffer decreases the stability of rH apoferritin (Tm Å 59.87C) much more than that of the horse spleen protein (Tm § 82.17C). The transition enthalpy of rH apoferritin was instead practically unchanged (Table I). This finding was confirmed by measurements of the near and far uv spectra. Figures 3A–A* and 3C–C*) show that the denaturant does not produce significant changes in the secondary and tertiary structure of horse spleen apoferritin, while in the rH protein the ellipticity decreases by 30–35% at 222 and 294 nm. In order to examine the behavior of horse spleen and rH apoferritins at acid pH values, the lowest pH was chosen at which both apoferritins are polymeric. This condition, which was established by means of sedimentation velocity experiments and near uv CD spectra (11, 6) corresponds to pH 4.0. Based on these results, the thermal stability of the two apoferritins was measured in 40 mM glycine–HCl buffer at pH 4.0. The Tm values correspond to 907C for the horse spleen and 677C for the rH protein (Fig. 1). In a pH 4.0 buffer containing 2 M guanidine–HCl, horse spleen apoferritin maintains its secondary and tertiary structure unaltered, based on the CD spectra in the far and near uv regions, while rH apoferritin loses about 70% of the ellipticity characteristic of the native protein in both regions (Figs. 3B–B* and 3D– D*). Under these conditions the calorimetric transition of the horse spleen protein was characterized by a Tm of 817C, while no appreciable signal was apparent in the thermograms of rH apoferritin, except a small endotherm centered around 507C. The thermal stability of horse spleen ferritin of variable iron content (200 and 2000 Fe atoms per polymer) was also studied at pH 7.0. The thermal behavior and the Tm values were always indistinguishable from
/ m4606$9213
11-30-95 16:04:27
arca
61
those of the fully iron-depleted protein (data not shown), indicating on the one hand that iron does not affect the stability of the protein and on the other that the preparations of apoferritin studied are native. Thermal Stability of the Dimeric Subunits and Their Reassociation Products To study apoferritin subunits, the polymers were incubated at pH 1.8 to achieve complete dissociation and thereafter dialyzed against 40 mM glycine–HCl buffer at pH 3.5. With this procedure both proteins yield fully folded dimeric subunits (s20 Å 3.5 S). The Tm values of the dimeric subunits are significantly decreased with respect to the polymers, and more so for the horse spleen subunits (Tm Å 427C) than for the rH ones (Tm Å 507C) (Fig. 4). Both apoferritin subunits maintain their dimeric state after heating to temperatures corresponding to the Tm value, while heating to higher temperatures leads to a slight aggregation of the samples (10–20%) as indicated by the appearance of a fast component in sedimentation velocity experiments (s20 Å 6.5 and 8.0 S for horse spleen and rH apoferritin, respectively) and to a significant loss in secondary structure, as indicated by the far uv CD spectra (Figs. 5A and 5B). In a calorimetric experiment, heating to half transition followed by rapid cooling showed more than 50% reproducibility in a subsequent heating cycle (data not shown) for both types of subunits. Calorimetric experiments have been carried out also on partially and fully reassociated horse spleen apoferritin obtained by dialyzing the dimeric subunits respectively versus 40 mM glycine–acetic acid buffer at pH 4.2 and versus 10 mM Tris–HCl at pH 8.0 containing 2 mM dithiothreitol (11). The thermal transition of partially reassociated apoferritin containing 50% polymer and 50% reassociation intermediates of 5.8 S (as determined in sedimentation velocity experiments) displays a broad endotherm over the temperature range 60– 807C (Fig. 6). The fully reassociated sample (s20 Å 16.0 S) is characterized by a calorimetric profile approaching that of the native polymer (Fig. 6). DISCUSSION
The present work provides the first direct assessment of the thermal stability of horse spleen (90% L chains) and rH apoferritin and their dimeric subunits under various buffer conditions. The DSC measurements not only confirm the higher stability of horse spleen apoferritin with respect to the rH homopolymer (6), but also demonstrate that the stabilization of the polymeric structure is governed by subunit interactions at the three- and fourfold symmetry axes since at the subunit level rH dimers are more stable than those obtained from horse spleen apoferritin.
AP: Archives
62
STEFANINI ET AL.
FIG. 3. (A, B, C, D) far and (A*, B*, C*, D*) near uv circular dichroism spectra of horse spleen and rH apoferritin as a function of pH in the absence and presence of 2 M guanidine–HCl. Buffer, (A, A*, C, C*) 50 mM phosphate at pH 7.0; (B, B*, D, D*) 40 mM glycine–HCl at pH 4.0 in the absence (—) and in the presence (— —) of 2 M guanidine–HCl. CD spectra were acquired after exposure of the samples to 2 M guanidine–HCl for 24 h. Protein concentration, (A, B, C, D) 0.25 mg/ml; (A*, B*, C*, D*) 2.5 mg/ml.
The DSC profiles of horse spleen apoferritin are characterized by significantly higher Tm values relative to rH apoferritin under all buffer conditions analyzed. Despite the exothermic influence of aggregation and/or precipitation on the DSC data of the horse spleen protein, the observed Tm values can be used as an index of the thermal stability of the protein because the system displays significant reversibility when heated at temperatures only a few degrees lower than the Tm value. Thus, when horse spleen apoferritin is heated to 867C, there is no precipitation and the ellipticity in the uv region decreases markedly, but increases again when the solutions are brought back to room temperature. In particular, at 222 nm, where the unfolding of the secondary structure is monitored, the molar ellipticity of the final solution corresponds to 80% of the initial value, whereas in the near uv region, where the tertiary structure is monitored, the extent of reversibility is less pronounced. In accordance with this finding sedimentation velocity experiments carried out in parallel show that about 30% of the protein undergoes changes in association state. It is of interest that in the experiments performed under conditions in which the native structure of horse
/ m4606$9213
11-30-95 16:04:27
arca
spleen apoferritin is destabilized without affecting the association state, e.g., at pH 4.0 or in the presence of guanidine either at neutral or acid pH, the Tm values decrease by É137C. This finding in turn confirms the unusually high stability of the horse spleen apoferritin assembly. In the case of rH apoferritin, the denaturation profile does not show exothermic aggregation during the DSC cycle and is characterized by lower Tm values (É167C at pH 7.0) than those measured for horse spleen apoferritin. As for horse spleen apoferritin, a significant recovery of the ellipticity in the far and near uv regions was observed in rH apoferritin incubated at a temperature 57C below the Tm value (Fig. 2). Thus, a quantitative analysis of the system in terms of thermodynamic parameters has been performed, an approach formally allowed only in the case of completely reversible transitions, but found permissible in systems displaying partial reversibility (12). Deconvolution of the thermal profile reveals that the denaturation endotherm is characterized by a very high DHc value. This, however, is due only to the very high molecular mass of the polymer, since the specific enthalpy change is rather small (see below). DHvH is much lower than DHc , indicating that
AP: Archives
CALORIMETRIC STUDY OF APOFERRITINS
63
FIG. 5. Thermally induced change in the (A, B) far and (A*, B*) near uv circular dicroism spectra of horse spleen and rH apoferritin subunits. Spectra measured before (—) and after (— —) a complete calorimetric cycle in DSC experiments. Protein concentration, (A, B) 0.1 mg/ml; (A*, B*) 2.5 mg/ml in 40 mM glycine–HCl buffer at pH 3.5. FIG. 4. DSC profiles of the horse spleen and rH dimeric subunits. The subunits were obtained by treatment of apoferritin at pH 1.8 followed by dialysis versus 40 mM glycine–HCl at pH 3.5.
denaturation is a multistate process (9). Therefore denaturation of the polymer is complex and gradual and many intermediates become populated between the initial native and the final denatured states of the process. The absence of a measurable DCp is an indication that few nonpolar groups become exposed to solvent during the thermal transition (13). A similar indication emerges from comparison of the specific denaturation enthalpy change of the rH homopolymer (4.2 cal/g) with that obtained for the unfolding of water-soluble proteins at 807C (§8 cal/g) (14). Both results suggest that the thermal denaturation does not lead to the complete unfolding of the molecule, in agreement with the CD data in the far uv region. The presence of substantial amounts of secondary structure at high temperatures is often observed in the case of multisubunit proteins which do not undergo dissociation upon heating (15). The rH protein is more sensitive to acid pH and/or to the presence of guanidine than horse apoferritin as indicated by the CD measurements performed at 207C (Fig. 3). The difference is apparent also in the calorimetric profiles and is particularly striking at pH 4.0 in the presence of guanidine. This condition leads only to a 127C decrease in the Tm value of horse spleen apoferritin, but to a marked destabilization of the rH polymer
/ m4606$9213
11-30-95 16:04:27
arca
FIG. 6. DSC profiles of partially and fully reassociated horse spleen apoferritin. In (A) reassociation (50% polymer and 50% reassociation intermediates) was obtained by dialyzing the subunits at pH 1.8 against 40 mM glycine–acetic acid at pH 4.2. In (B) same solution as in (A) after prolonged dialysis against 10 mM Tris–HCl buffer at pH 8.0 containing 2 mM dithiothreitol.
AP: Archives
64
STEFANINI ET AL. TABLE II
Thermodynamic Data of Horse Spleen and rH Apoferritin Dimeric Subunits Determined by Differential Scanning Calorimetrya
Protein
Solvent
Tm (7C)
Dhc (cal/g)
DHc (kcal/mol)
DC p (cal/g 7C)
HoS rH
pH 3.5 pH 3.5
42.1 50.0
2.2 2.1
83.9 89.9
0.2 0.1
a For details see Materials and Methods. HoS, horse spleen apoferritin.
as indicated by the very small endotherm, centered at about 507C, displayed by the relevant thermogram (Fig. 1). The thermograms of the dimeric subunits are characterized by one peak centered around 42.1 and 50.07C, respectively, for the horse spleen and rH subunits. Both DCp and DHc values are consistent with a complete unfolding of the polypeptides (14). This behavior is in contrast with that of the assembled molecules, which are only partially unfolded upon heating. In turn it indicates that the polymers are stabilized primarily by interactions between dimeric subunits (17) in accordance with the knowledge that dimers are the fundamental structural units of the assembled molecule (2, 7, 16). Most interestingly, the two dimeric subunits are characterized by an inversion of stability with respect to the two 24-mers, the Tm of the rH subunit being about 87C higher than that of the L subunit (Table II). The difference is significant and cannot be attributed to a different state of association of two subunits which have the same sedimentation coefficient, 3.5 S. The low Tm values of the subunits relative to the polymers (Table II), taken together with the inversion in the thermal stability, indicate that the interfaces along the three- and fourfold axes are the most important ones in determining the stability of the assembled molecule. The same conclusion was reached in guanidine–acid pH denaturation studies on L and H homopolymers and H-chain mutants with substitutions along the three- and fourfold channels (6). From the latter experiments, in which the end products were unfolded monomers, it was also inferred that a salt bridge inside the four-helix bundle of the L chain, which is absent in the H chain, is an important factor conferring stability to L-chain apoferritin. The inversion of the thermal stability of fully folded dimeric subunits is in contrast with this contention, but rather indicates that this specific salt bridge renders the tertiary structure of the L sub-
/ m4606$9213
11-30-95 16:04:27
arca
units less susceptible to the unfolding action of guanidine at acidic pH. It is of interest that the presence of the iron core does not affect the position and the shape of the thermograms as native horse spleen ferritin, partially irondepleted ferritin, and apoferritin give very similar thermograms at least up to 1057C. This observation is related to the unusual characteristics of the iron core as a ligand. The contacts it establishes with the apoferritin cavity are so loose (18) that iron does not influence the thermodynamic parameters of the protein envelope. An important practical outcome of the present study is the definition of the temperatures which should not be exceeded during the purification process of a given apoferritin to avoid its irreversible denaturation. Such temperatures correspond to 807C for horse spleen apoferritin and only to 607C for the rH homopolymer. REFERENCES 1. Harrison, P. M., Ford, G. C., Rice, R. W., Smith, J. M. A., Treffry, A., and White, J. L. (1986) in Frontiers in Bioinorganic Chemistry (Xavier, A. V., Ed.), pp. 268–277, VCH, Weinheim. 2. Lawson, D. M., Artymiuk, P. J., Harrison, P. M., Yedall, S. J., Luzzago, A., Cesareni, G., Levi, S., and Arosio, P. (1989) FEBS Lett. 254, 207–210. 3. Levi, S., Luzzago, A., Cesareni, G., Cozzi, A., Franceschinelli, F., Albertini, A., and Arosio, P. (1988) J. Biol. Chem. 263, 18086– 18092. 4. Levi, S., Franceschinelli, F., Santambrogio, P., Cesareni, G., and Arosio, P. (1989) Biochem. J. 264, 381–388. 5. Crichton, R. R., and Bryce, C. F. A. (1973) Biochem. J. 133, 289– 299. 6. Santambrogio, P., Levi, S., Arosio, P., Palagi, L., Vecchio, G., Lawson, D., Yewdall, S. J., Artymiuk, P. J., Harrison, P. M., Jappelli, R., and Cesareni, G. (1992) J. Biol. Chem. 267, 14077– 14083. 7. Stefanini, S., Chiancone, E., Arosio, P., Finazzi-Agro`, A., and Antonini, E. (1982) Biochemistry 21, 2293–2299. 8. Privalov, P. L., and Potekhin, S. A. (1986) Methods Enzymol. 131, pp. 4–51. 9. Sturtevant, J. M. (1987) Annu. Rev. Phys. Chem. 38, 463–488. 10. Stefanini, S., Chiancone, E., Antonini, E., and Finazzi-Agro`, A. (1976) FEBS Lett. 69, 90–94. 11. Stefanini, S., Vecchini, P., and Chiancone, E. (1987) Biochemistry 26, 1831–1837. 12. Edge, V., Allewell, N. M., and Sturtevant, J. M. (1985) Biochemistry 24, 5899–5906. 13. Privalov, P. L., and Makhatadze, G. I. (1990) J. Mol. Biol. 213, 385–391. 14. Privalov, P. L. (1979) Adv. Protein Chem. 33, 167–241. 15. Ginsburg, A., and Zolkiewski, M. (1991) Biochemistry 30, 9421– 9429. 16. Gerl, M., and Jaenicke, R. (1987) Biol. Chem. Hoppe–Seyler 368, 387–396. 17. Johnson, C. R., Morin, P. E., Arrowsmith, C. H., and Freire, E. (1995) Biochemistry 34, 5309–5316. 18. Massover, W. H. (1993) Micron 24, 389–437.
AP: Archives
Vol. 325, No. 1, January 1, pp. 58–64, 1996 Article No. 0007
Thermal Stability of Horse Spleen Apoferritin and Human Recombinant H Apoferritin Simonetta Stefanini, Stefano Cavallo, Chang-Qing Wang, Paola Tataseo,* Paola Vecchini, Anna Giartosio, and Emilia Chiancone1 CNR Center of Molecular Biology, Department of Biochemical Sciences ‘‘A. Rossi Fanelli,’’ University La Sapienza, 00185 Rome; and *Department of Biology, University Tor Vergata, 00173 Rome, Italy
Received June 21, 1995, and in revised form September 14, 1995
The thermal stability of horse spleen apoferritin, a heteropolymer composed of 90% L and 10% H chains, has been studied by differential scanning calorimetry and compared with that of the human recombinant H homopolymer. The denaturation temperatures (Tm) are significantly higher for the horse spleen polymer than for the recombinant protein under all experimental conditions (e.g., at pH 7, Tm values are §93 and 777C, respectively). The thermal denaturation process displays substantial reversibility for both polymers up to a few degrees below Tm , as indicated by CD measurements in the far and near uv regions. At temperatures higher than Tm the thermograms are influenced by the exothermic contribution of aggregation and/or precipitation. The H homopolymer thermogram, which is not distorted by the exotherm, is consistent with a multistate denaturation process. Acid dissociation of apoferritin produces stable dimeric subunits. The thermal unfolding of both dimeric subunits is reversible at least up to Tm and is characterized by an inversion of stability relative to the polymers (at pH 3.5, Tm is 427C for the horse spleen and 507C for the H subunit). These results indicate that the stabilization of the polymeric structure arises mainly from interactions between dimers, in accordance with the crystallographic evidence that the dimers are the building blocks of the polymeric molecule. q 1996 Academic Press, Inc. Key Words: apoferritin; differential scanning calorimetry; thermal stability; circular dichroism; sedimentation velocity.
Ferritin, the ubiquitous iron storage protein, is characterized by an unusually high thermal stability, a 1 To whom correspondence should be addressed at CNR Center of Molecular Biology c/o Department of Biochemical Sciences ‘‘A. Rossi Fanelli,’’ University La Sapienza, Piazzale A. Moro 5, 00185 Rome, Italy. Fax: 39-6-4440062.
property which is exploited in the purification procedure of the protein. Natural apoferritin is a heteropolymer composed of L and H subunits which coassemble into a 24-mer shell with a large cavity harboring up to 4500 iron atoms. The H and L chains share the same tertiary structure which is characterized by a bundle of four antiparallel helices (A–D), an additional short helix (E), and a loop connecting helices B and C. By interdigitation of sidechains in the loop region the subunits give rise to dimeric building blocks which generate a molecular assembly with 432 symmetry and leave intersubunit channels along the three- and fourfold axes for the passage of iron and/or small molecules (1, 2). The study of H and L homopolymers obtained by means of recombinant DNA technology brought out clearly that the two chain types differ from a functional viewpoint. Thus, the H chains contain a ferroxidase center in the four-helix bundle which is missing in the L chains (2). Accordingly, the H chains have been proposed to promote iron oxidation, whereas the L chains are thought to play a role in iron accumulation (3, 4). From a structural viewpoint the L chains confer to the assembled molecule a greater stability toward chemical and physical agents than do the H chains. In fact, the L-rich apoferritin from horse spleen (90% L) denatures only under extreme conditions, namely at temperatures higher than 807C and at pH values below 2.0 or above 10.5 (5). The stability conveyed by the L chains can be ascribed in part to residues located at the intersubunit contacts along the three- and fourfold channels since H homopolymers become more stable if residues along either type of channel are substituted with L-chain residues (6). In addition, a special role has been attributed to a specific salt bridge within the four-helix bundle between Lys62 and Glu107 (6); the corresponding residues in the H chain, Glu62 and Glu107, are part of the ferroxidase center (2). This paper reports the first direct measurement by
58
0003-9861/96 $12.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
/ m4606$9213
11-30-95 16:04:27
arca
AP: Archives
CALORIMETRIC STUDY OF APOFERRITINS
59
means of differential scanning calorimetry (DSC)2 of the thermal stability for horse spleen apoferritin, the H homopolymer, and their dimeric subunits. Horse spleen apoferritin is characterized by a markedly higher denaturation temperature than the H homopolymer; in contrast the H dimeric subunits are more stable than the L subunits. The DSC data have been supplemented by far and near uv–CD measurements and sedimentation velocity experiments in order to monitor the extent of unfolding and/or dissociation of the protein. MATERIALS AND METHODS Horse spleen ferritin (90% L; HoS) was prepared as described previously (7). Human recombinant H ferritin (100% H subunit; rH) was overexpressed in Escherichia coli and purified essentially as described by Levi et al. (3). Apoferritin was usually obtained from horse spleen ferritin by reduction of iron with thioglycollic acid and chelation with 2,2*-bipyridyl. In the case of rH ferritin, iron was removed in anaerobiosis by dialysis against 50 mM MES–NaOH buffer at pH 6.0 containing 0.3% sodium dithionite and 2,2*-bipyridyl. When this latter procedure was applied to horse spleen ferritin, the resulting apoprotein was found to be identical to that obtained with thioglycollic acid. Apoferritin concentration was calculated from the absorbance at 280 nm using the extinction coefficient e 1%, 1 cm Å 9.0. Calorimetric experiments were performed with a high-resolution MicroCal MC-2 differential scanning calorimeter (MicroCal, Inc., Northampton, MA) equipped with a DA-2 digital data acquisition system. Samples were always heated at a scan rate of 607C/h. The concentration of the protein solutions was 2–3 mg/ml. All curves were corrected for the instrumental baseline obtained by filling both cells with the buffer used. Integration and deconvolution were performed using the ORIGIN software provided by Microcal, Inc., after normalizing the data for protein concentration and subtracting as baseline the straight line connecting the initial and the final temperatures of the overall transition in the case of the rH polymer, or a sigmoidal progress curve in the case of the subunits (8, 9). The values of Tm (temperature at which excess heat capacity reaches a maximum; 7C), Dhc (specific enthalpy change; cal/g), DHc (calorimetric enthalpy change; kcal/mol), and DCp (molar heat capacity change; cal/g 7C) were obtained by analysis of the resulting thermal profiles. DHc was calculated for the rH homopolymer on the basis of Mr 504 kDa and for the dimeric subunits on the basis of Mr 2 1 19.5 and 2 1 21.0 kDa for the horse spleen and recombinant protein, respectively. The van’t Hoff enthalpy change (DHvH ; kcal/mol) was calculated according to the formula 2 DHvH Å A RT 1/2 rcex,1/2/Dhcal
where cex,1/2 (Cp at Tm in our notation) is the excess heat capacity at T1/2 (Tm in our notation), i.e., the distance from the baseline to the top of the transition; Dhcal corresponds to Dhc in our notation; and A Å 4 for a nondissociating system undergoing a two-state transition (9). All values reported in calories can be converted into joules multiplying by the factor 4.184. Circular dichroism measurements were carried out using a Jasco710 A spectropolarimeter. In the experiments designed to monitor
2 Abbreviations used: HoS, horse spleen ferritin; rH, human recombinant ferritin; DSC, differential scanning calorimetry; Gu-HCl, guanidine–HCl; MES, 4-morpholineethanesulfonic acid.
/ m4606$9213
11-30-95 16:04:27
arca
FIG. 1. DSC profiles of horse spleen and rH apoferritin at neutral and acid pH in the absence and presence of 2 M guanidine–HCl. Buffers, 50 mM sodium phosphate at pH 7.0 and 40 mM glycine–HCl at pH 4.0. Apoferritin, horse spleen (—), rH (rrr). In this and the following DSC figures the curves have been arbitrarily shifted on the y axis for clarity.
the effect of high temperatures on unfolding process, a CD spectrum was first measured at 207C, the protein was then incubated in thermostatted cells at the desired temperature for 5 min, and a second CD spectrum was acquired in either 5 or 20 min; thereafter the sample was cooled back to 207C and a last CD spectrum was measured. The acquisition time varied between 5 and 20 min. In all other experiments the spectra were recorded at 207C. The molar ellipticity (deg r cm2 r dmol01) was calculated in the far uv region ([U]aa) on the basis of a mean residue molecular weight of 113 and in the near uv region ([U]Mr) on the basis of the polypeptide chain molecular weight 19.5 and 21.0 kDa for horse spleen and rH apoferritin, respectively. Sedimentation velocity experiments were carried out at a temperature of 207C in a Beckman Optima XL-A analytical ultracentrifuge. The sedimentation coefficients were reduced to s20 by standard procedures.
RESULTS
Thermal Stability of the Polymers The thermal stability of horse spleen and human recombinant apoferritins has been studied under a number of experimental conditions. In 50 mM phosphate buffer at pH 7.0 the calorimetric profile of horse spleen apoferritin shows a single peak characterized by a Tm value of about 937C (Fig. 1); at higher temperatures aggregation and precipitation take place. Heating to 897C in a first scan does not greatly affect the calorimetric pattern in a subsequent heating cycle, suggesting that the system shows substantial reversibility until aggregation occurs. In order to further investigate the extent to which the DSC data reflect reversible unfolding and/or dissociation reactions, the calorimetric experiments were supplemented with CD and sedimentation velocity experiments. Far and near uv measure-
AP: Archives
60
STEFANINI ET AL.
FIG. 2. Effect of temperature on the (A, B, C, D) far and (A*, B*, C*, D*) near uv circular dichroism spectra of horse spleen and rH apoferritins. Each graph shows the spectra acquired at 207C (—), at high temperature (rrr), and after cooling at 207C (— —). The high temperature corresponds to (A, A*) 767C, (B, B*) 817C, (C, C*) 867C, (D, D*) 727C. The samples were incubated at the desired temperature for 5 min and the CD spectra were acquired in 20 min; thereafter the samples were cooled back to 207C and the CD spectra were measured again. Protein concentration: (A, B, C, D) 0.12 mg/ml and (A*, B*, C*, D*) 3 mg/ml in 50 mM phosphate buffer at pH 7.0.
ments were carried out between 767C, which corresponds to the onset of the thermal transition, and 867C, a few degrees below the Tm value (Figs. 2A–A* and 2C–C*). Over this temperature range no precipitation occurred and the ellipticity was independent of acquisition time (from 5 to 20 min). The ellipticity decreases progressively with increase in temperature in both spectral regions and, at any given temperature, the effect is more marked in the near uv region than in the far uv. In the sample heated to 867C, for example, the molar ellipticity at 222 nm, where the secondary structure is monitored ([U]aa), decreases by 30%, whereas it decreases by 70% at 292 nm, where the environment of the single tryptophan residue in the loop region at the dimer interface is monitored (10). The decrease in ellipticity is essentially reversible in the sample heated to 767C (Fig. 2A–A*). In contrast, the recovery of the initial ellipticity is only partial in the samples heated to higher temperatures. Moreover, the extent of recovery differs in the far and near uv regions, being larger in the former (Figs. 2B–B* and 2C–C*). In order to establish whether significant changes in quaternary structure occur in horse spleen apoferritin incubated at 867C, a sample was kept at this temperature for 25 min,
/ m4606$9213
11-30-95 16:04:27
arca
brought back to 207C, and analyzed by sedimentation velocity experiments. The sedimentation velocity patterns show, in addition to the native polymeric structure, dissociated material (10%) and aggregated protein (20%) consistently with the partial loss of secondary and tertiary structure. The calorimetric profile of rH apoferritin at pH 7.0 displays a single transition characterized by a Tm value of 777C; precipitation takes place only after completion of the thermal transition, thus allowing determination of the calorimetric enthalpy change. The DHc value is much higher than DH£H (Table I) and there is no appreciable DCp . CD spectra were used to assess the extent of temperature-induced unfolding and/or dissociation in rH apoferritin. Due to the small amount of material available the measurements were performed only at 727C, a temperature 57C lower than the Tm value. As in the case of horse spleen apoferritin, incubation at high temperature induces a loss of ellipticity at 222 nm which corresponds to 35% of the initial value, while the effect is more marked in the near uv region where the peak at 294 nm is abolished. The ellipticity is partially recovered after cooling and the recovery is more marked at
AP: Archives
CALORIMETRIC STUDY OF APOFERRITINS TABLE I
Calorimetric Data of rH Apoferritina
Solvent pH 7.0 pH 7.0 2 M Gu–HCl a
Tm (7C)
Dhc (cal/g)
DHc (kcal/mol)
DHvH (kcal/mol)
77.3
4.2
2127
51.7
59.8
4.7
2369
42.1
For details see Materials and Methods.
222 than at 294 nm (Fig. 2D–D*). Sedimentation velocity experiments indicate that the polymeric structure of rH apoferritin is unaltered after incubation at 727C for 25 min. In further experiments the native structure of the two apoferritins was mildly destabilized by means of chemical agents like guanidine and/or low pH. The Tm values indicate (Fig. 1) that addition of 2 M guanidine– HCl to the pH 7.0 phosphate buffer decreases the stability of rH apoferritin (Tm Å 59.87C) much more than that of the horse spleen protein (Tm § 82.17C). The transition enthalpy of rH apoferritin was instead practically unchanged (Table I). This finding was confirmed by measurements of the near and far uv spectra. Figures 3A–A* and 3C–C*) show that the denaturant does not produce significant changes in the secondary and tertiary structure of horse spleen apoferritin, while in the rH protein the ellipticity decreases by 30–35% at 222 and 294 nm. In order to examine the behavior of horse spleen and rH apoferritins at acid pH values, the lowest pH was chosen at which both apoferritins are polymeric. This condition, which was established by means of sedimentation velocity experiments and near uv CD spectra (11, 6) corresponds to pH 4.0. Based on these results, the thermal stability of the two apoferritins was measured in 40 mM glycine–HCl buffer at pH 4.0. The Tm values correspond to 907C for the horse spleen and 677C for the rH protein (Fig. 1). In a pH 4.0 buffer containing 2 M guanidine–HCl, horse spleen apoferritin maintains its secondary and tertiary structure unaltered, based on the CD spectra in the far and near uv regions, while rH apoferritin loses about 70% of the ellipticity characteristic of the native protein in both regions (Figs. 3B–B* and 3D– D*). Under these conditions the calorimetric transition of the horse spleen protein was characterized by a Tm of 817C, while no appreciable signal was apparent in the thermograms of rH apoferritin, except a small endotherm centered around 507C. The thermal stability of horse spleen ferritin of variable iron content (200 and 2000 Fe atoms per polymer) was also studied at pH 7.0. The thermal behavior and the Tm values were always indistinguishable from
/ m4606$9213
11-30-95 16:04:27
arca
61
those of the fully iron-depleted protein (data not shown), indicating on the one hand that iron does not affect the stability of the protein and on the other that the preparations of apoferritin studied are native. Thermal Stability of the Dimeric Subunits and Their Reassociation Products To study apoferritin subunits, the polymers were incubated at pH 1.8 to achieve complete dissociation and thereafter dialyzed against 40 mM glycine–HCl buffer at pH 3.5. With this procedure both proteins yield fully folded dimeric subunits (s20 Å 3.5 S). The Tm values of the dimeric subunits are significantly decreased with respect to the polymers, and more so for the horse spleen subunits (Tm Å 427C) than for the rH ones (Tm Å 507C) (Fig. 4). Both apoferritin subunits maintain their dimeric state after heating to temperatures corresponding to the Tm value, while heating to higher temperatures leads to a slight aggregation of the samples (10–20%) as indicated by the appearance of a fast component in sedimentation velocity experiments (s20 Å 6.5 and 8.0 S for horse spleen and rH apoferritin, respectively) and to a significant loss in secondary structure, as indicated by the far uv CD spectra (Figs. 5A and 5B). In a calorimetric experiment, heating to half transition followed by rapid cooling showed more than 50% reproducibility in a subsequent heating cycle (data not shown) for both types of subunits. Calorimetric experiments have been carried out also on partially and fully reassociated horse spleen apoferritin obtained by dialyzing the dimeric subunits respectively versus 40 mM glycine–acetic acid buffer at pH 4.2 and versus 10 mM Tris–HCl at pH 8.0 containing 2 mM dithiothreitol (11). The thermal transition of partially reassociated apoferritin containing 50% polymer and 50% reassociation intermediates of 5.8 S (as determined in sedimentation velocity experiments) displays a broad endotherm over the temperature range 60– 807C (Fig. 6). The fully reassociated sample (s20 Å 16.0 S) is characterized by a calorimetric profile approaching that of the native polymer (Fig. 6). DISCUSSION
The present work provides the first direct assessment of the thermal stability of horse spleen (90% L chains) and rH apoferritin and their dimeric subunits under various buffer conditions. The DSC measurements not only confirm the higher stability of horse spleen apoferritin with respect to the rH homopolymer (6), but also demonstrate that the stabilization of the polymeric structure is governed by subunit interactions at the three- and fourfold symmetry axes since at the subunit level rH dimers are more stable than those obtained from horse spleen apoferritin.
AP: Archives
62
STEFANINI ET AL.
FIG. 3. (A, B, C, D) far and (A*, B*, C*, D*) near uv circular dichroism spectra of horse spleen and rH apoferritin as a function of pH in the absence and presence of 2 M guanidine–HCl. Buffer, (A, A*, C, C*) 50 mM phosphate at pH 7.0; (B, B*, D, D*) 40 mM glycine–HCl at pH 4.0 in the absence (—) and in the presence (— —) of 2 M guanidine–HCl. CD spectra were acquired after exposure of the samples to 2 M guanidine–HCl for 24 h. Protein concentration, (A, B, C, D) 0.25 mg/ml; (A*, B*, C*, D*) 2.5 mg/ml.
The DSC profiles of horse spleen apoferritin are characterized by significantly higher Tm values relative to rH apoferritin under all buffer conditions analyzed. Despite the exothermic influence of aggregation and/or precipitation on the DSC data of the horse spleen protein, the observed Tm values can be used as an index of the thermal stability of the protein because the system displays significant reversibility when heated at temperatures only a few degrees lower than the Tm value. Thus, when horse spleen apoferritin is heated to 867C, there is no precipitation and the ellipticity in the uv region decreases markedly, but increases again when the solutions are brought back to room temperature. In particular, at 222 nm, where the unfolding of the secondary structure is monitored, the molar ellipticity of the final solution corresponds to 80% of the initial value, whereas in the near uv region, where the tertiary structure is monitored, the extent of reversibility is less pronounced. In accordance with this finding sedimentation velocity experiments carried out in parallel show that about 30% of the protein undergoes changes in association state. It is of interest that in the experiments performed under conditions in which the native structure of horse
/ m4606$9213
11-30-95 16:04:27
arca
spleen apoferritin is destabilized without affecting the association state, e.g., at pH 4.0 or in the presence of guanidine either at neutral or acid pH, the Tm values decrease by É137C. This finding in turn confirms the unusually high stability of the horse spleen apoferritin assembly. In the case of rH apoferritin, the denaturation profile does not show exothermic aggregation during the DSC cycle and is characterized by lower Tm values (É167C at pH 7.0) than those measured for horse spleen apoferritin. As for horse spleen apoferritin, a significant recovery of the ellipticity in the far and near uv regions was observed in rH apoferritin incubated at a temperature 57C below the Tm value (Fig. 2). Thus, a quantitative analysis of the system in terms of thermodynamic parameters has been performed, an approach formally allowed only in the case of completely reversible transitions, but found permissible in systems displaying partial reversibility (12). Deconvolution of the thermal profile reveals that the denaturation endotherm is characterized by a very high DHc value. This, however, is due only to the very high molecular mass of the polymer, since the specific enthalpy change is rather small (see below). DHvH is much lower than DHc , indicating that
AP: Archives
CALORIMETRIC STUDY OF APOFERRITINS
63
FIG. 5. Thermally induced change in the (A, B) far and (A*, B*) near uv circular dicroism spectra of horse spleen and rH apoferritin subunits. Spectra measured before (—) and after (— —) a complete calorimetric cycle in DSC experiments. Protein concentration, (A, B) 0.1 mg/ml; (A*, B*) 2.5 mg/ml in 40 mM glycine–HCl buffer at pH 3.5. FIG. 4. DSC profiles of the horse spleen and rH dimeric subunits. The subunits were obtained by treatment of apoferritin at pH 1.8 followed by dialysis versus 40 mM glycine–HCl at pH 3.5.
denaturation is a multistate process (9). Therefore denaturation of the polymer is complex and gradual and many intermediates become populated between the initial native and the final denatured states of the process. The absence of a measurable DCp is an indication that few nonpolar groups become exposed to solvent during the thermal transition (13). A similar indication emerges from comparison of the specific denaturation enthalpy change of the rH homopolymer (4.2 cal/g) with that obtained for the unfolding of water-soluble proteins at 807C (§8 cal/g) (14). Both results suggest that the thermal denaturation does not lead to the complete unfolding of the molecule, in agreement with the CD data in the far uv region. The presence of substantial amounts of secondary structure at high temperatures is often observed in the case of multisubunit proteins which do not undergo dissociation upon heating (15). The rH protein is more sensitive to acid pH and/or to the presence of guanidine than horse apoferritin as indicated by the CD measurements performed at 207C (Fig. 3). The difference is apparent also in the calorimetric profiles and is particularly striking at pH 4.0 in the presence of guanidine. This condition leads only to a 127C decrease in the Tm value of horse spleen apoferritin, but to a marked destabilization of the rH polymer
/ m4606$9213
11-30-95 16:04:27
arca
FIG. 6. DSC profiles of partially and fully reassociated horse spleen apoferritin. In (A) reassociation (50% polymer and 50% reassociation intermediates) was obtained by dialyzing the subunits at pH 1.8 against 40 mM glycine–acetic acid at pH 4.2. In (B) same solution as in (A) after prolonged dialysis against 10 mM Tris–HCl buffer at pH 8.0 containing 2 mM dithiothreitol.
AP: Archives
64
STEFANINI ET AL. TABLE II
Thermodynamic Data of Horse Spleen and rH Apoferritin Dimeric Subunits Determined by Differential Scanning Calorimetrya
Protein
Solvent
Tm (7C)
Dhc (cal/g)
DHc (kcal/mol)
DC p (cal/g 7C)
HoS rH
pH 3.5 pH 3.5
42.1 50.0
2.2 2.1
83.9 89.9
0.2 0.1
a For details see Materials and Methods. HoS, horse spleen apoferritin.
as indicated by the very small endotherm, centered at about 507C, displayed by the relevant thermogram (Fig. 1). The thermograms of the dimeric subunits are characterized by one peak centered around 42.1 and 50.07C, respectively, for the horse spleen and rH subunits. Both DCp and DHc values are consistent with a complete unfolding of the polypeptides (14). This behavior is in contrast with that of the assembled molecules, which are only partially unfolded upon heating. In turn it indicates that the polymers are stabilized primarily by interactions between dimeric subunits (17) in accordance with the knowledge that dimers are the fundamental structural units of the assembled molecule (2, 7, 16). Most interestingly, the two dimeric subunits are characterized by an inversion of stability with respect to the two 24-mers, the Tm of the rH subunit being about 87C higher than that of the L subunit (Table II). The difference is significant and cannot be attributed to a different state of association of two subunits which have the same sedimentation coefficient, 3.5 S. The low Tm values of the subunits relative to the polymers (Table II), taken together with the inversion in the thermal stability, indicate that the interfaces along the three- and fourfold axes are the most important ones in determining the stability of the assembled molecule. The same conclusion was reached in guanidine–acid pH denaturation studies on L and H homopolymers and H-chain mutants with substitutions along the three- and fourfold channels (6). From the latter experiments, in which the end products were unfolded monomers, it was also inferred that a salt bridge inside the four-helix bundle of the L chain, which is absent in the H chain, is an important factor conferring stability to L-chain apoferritin. The inversion of the thermal stability of fully folded dimeric subunits is in contrast with this contention, but rather indicates that this specific salt bridge renders the tertiary structure of the L sub-
/ m4606$9213
11-30-95 16:04:27
arca
units less susceptible to the unfolding action of guanidine at acidic pH. It is of interest that the presence of the iron core does not affect the position and the shape of the thermograms as native horse spleen ferritin, partially irondepleted ferritin, and apoferritin give very similar thermograms at least up to 1057C. This observation is related to the unusual characteristics of the iron core as a ligand. The contacts it establishes with the apoferritin cavity are so loose (18) that iron does not influence the thermodynamic parameters of the protein envelope. An important practical outcome of the present study is the definition of the temperatures which should not be exceeded during the purification process of a given apoferritin to avoid its irreversible denaturation. Such temperatures correspond to 807C for horse spleen apoferritin and only to 607C for the rH homopolymer. REFERENCES 1. Harrison, P. M., Ford, G. C., Rice, R. W., Smith, J. M. A., Treffry, A., and White, J. L. (1986) in Frontiers in Bioinorganic Chemistry (Xavier, A. V., Ed.), pp. 268–277, VCH, Weinheim. 2. Lawson, D. M., Artymiuk, P. J., Harrison, P. M., Yedall, S. J., Luzzago, A., Cesareni, G., Levi, S., and Arosio, P. (1989) FEBS Lett. 254, 207–210. 3. Levi, S., Luzzago, A., Cesareni, G., Cozzi, A., Franceschinelli, F., Albertini, A., and Arosio, P. (1988) J. Biol. Chem. 263, 18086– 18092. 4. Levi, S., Franceschinelli, F., Santambrogio, P., Cesareni, G., and Arosio, P. (1989) Biochem. J. 264, 381–388. 5. Crichton, R. R., and Bryce, C. F. A. (1973) Biochem. J. 133, 289– 299. 6. Santambrogio, P., Levi, S., Arosio, P., Palagi, L., Vecchio, G., Lawson, D., Yewdall, S. J., Artymiuk, P. J., Harrison, P. M., Jappelli, R., and Cesareni, G. (1992) J. Biol. Chem. 267, 14077– 14083. 7. Stefanini, S., Chiancone, E., Arosio, P., Finazzi-Agro`, A., and Antonini, E. (1982) Biochemistry 21, 2293–2299. 8. Privalov, P. L., and Potekhin, S. A. (1986) Methods Enzymol. 131, pp. 4–51. 9. Sturtevant, J. M. (1987) Annu. Rev. Phys. Chem. 38, 463–488. 10. Stefanini, S., Chiancone, E., Antonini, E., and Finazzi-Agro`, A. (1976) FEBS Lett. 69, 90–94. 11. Stefanini, S., Vecchini, P., and Chiancone, E. (1987) Biochemistry 26, 1831–1837. 12. Edge, V., Allewell, N. M., and Sturtevant, J. M. (1985) Biochemistry 24, 5899–5906. 13. Privalov, P. L., and Makhatadze, G. I. (1990) J. Mol. Biol. 213, 385–391. 14. Privalov, P. L. (1979) Adv. Protein Chem. 33, 167–241. 15. Ginsburg, A., and Zolkiewski, M. (1991) Biochemistry 30, 9421– 9429. 16. Gerl, M., and Jaenicke, R. (1987) Biol. Chem. Hoppe–Seyler 368, 387–396. 17. Johnson, C. R., Morin, P. E., Arrowsmith, C. H., and Freire, E. (1995) Biochemistry 34, 5309–5316. 18. Massover, W. H. (1993) Micron 24, 389–437.
AP: Archives