- Email: [email protected]
Pressure studies on protein folding, misfolding, protein-DNA interactions and amyloidogenesis
Trends in High Pressure Bioscience and Biotechnology R. Hayashi (editor) 9 2002 Elsevier Science B.V. All rights reserved.
79
Pressure Studies on Protein Folding, Misfolding, Protein-DNA Interactions and Amyloidogenesis. D. Ishimaru, L. M. T. R. Lima, A. Ferrfio-Gonzales, P. A. Quesado, L. M. Maiolino, J. L. Silva and D. Foguel. Centro National de Ressonfincia Magnetica Nuclear, Departamento de Bioquimica Medica, lnstituto de Ci~ncias Biom6dicas, Universidade Federal do Rio de Janeiro. ABSTRACT Hydrostatic pressure is a useful tool for dissecting macromolecular interactions at the molecular level. Nonpolar interactions are determining factors in protein folding, protein aggregation and protein-nucleic acid recognition. Because nonpolar interactions are entropic and compressible, they are more sensitive to pressure and low temperatures. We have studied problems of macromolecular recognition using hydrostatic pressure as the primary tool and employing several spectroscopic techniques, especially fluorescence, circular dichroism and high-resolution nuclear magnetic resonance. High pressure has the unique property of stabilizing partially folded states of a protein which degree of dissimilarity from the native state may range from drifted conformations to moltenglobule states. The competition between correct folding and misfolding, which in many proteins leads to formation of insoluble aggregates is an important problem in the biotechnology industry and in human diseases such as amyloidosis, Alzheimer's, prion and tumor diseases. Because of its ability to sequester folding intermediates, pressure has been used to direct the folding in one direction or the other and to explore intermediates, which are at the junction of the routes for folding and aggregation.
1. INTRODUCTION Proteins play the major functions in cells either as isolated molecules or forming macromolecular complexes held together by noncovalent interactions. A unique feature of the "oiologicar world, as compared to the 'inorganic' world, is the large number of interactions among its components that involve relatively weak energies (less than 10 kcal/mol). The energetic coupling among these interactions is likely the basis of the high specificity of macromolecular recognition. Life depends on macromolecular recognition, especially at the level of protein folding and protein-nucleic acid (NA) interactions. The mechanism by which unstructured proteins spontaneously fold to their native functional form is still not completely understood (1). To this end, it is necessary to isolate and describe the intermediate structures that occur during folding. Here, it is reviewed that in several cases these intermediates have been trapped under pressure and their dynamics and structure have been characterized by fluorescence, light scattering, NMR and hydrodynamic methods. Pressure affects the equilibrium between denatured or dissociated and native forms in the direction of the form that occupies a smaller volume (2-5). The structural region of the protein that is most sensitive to pressure is the hydrophobic core, especially when cavities are present (5, 6). It is described below the unique characteristic of pressure to isolate molten globules or to favor the state that appears to have a segment of the protein in a partially folded state ("partial" molten globule). The isolation of folding intermediates is crucial to the understanding of protein misfolding and protein aggregation. In the last decade, several diseases, such as Alzheimer's disease and other amyloidogenic diseases, spongiform encephalopathies (caused by prions), inherited emphysema, cystic fibrosis and likely many cancers are caused by protein misfolding (7). Biotechnology companies also face many problems with proteins aggregating into inclusion bodies when they are expressed in
80
bacteria. Figure 1 shows how functional proteins and macromolecular assemblies may divert from the folding pathway to a misfolding and aggregation dead end.
1
2
g;
3
/
"-...
4 Figure 1. Schematic representation of the folding and assembly of protein and multimolecular complexes. (1) Unfolded protein; (2) Folding intermediate; (3) Tertiary structure; (4) Protein misfoldmg - Aggregate; (5) Protein-DNA interaction; (6) Virus; (7) Protein-ligand interaction. To control misfolding and aggregation, we have to understand the energetics, the dynamics and the mechanisms of correct assembly. As described below, pressure has been useful not only to investigate the correct pathways (leading to 3, 5, 6 and 7 products - Figure 1), but also to the incorrect, misfolded form (product 4).
1.1. Protein folding and protein-protein interactions: The volume cavity model for pressureinduced dissociation and unfolding. The mechanisms by which one dimension information (linear sequence of amino acid residues) is transferred to four dimensions (3D-structure and dynamics) remains one of the major challenges in Biology. Although most of the thermodynamic studies have been done with conventional perturbing agents, such as high temperature, urea, guanidine, etc, they offer limited information, especially because they cause drastic changes in protein structure, not likely to occur at physiological conditions. Folding of a protein, or association of one protein with another, is accompanied by an increase in volume because of the additive effects of the formation of solvent-excluding cavities and the release of bound solvent. Water is released as nonpolar amino acid residues are buried, as well as when salt linkages are formed (2-5). The Gibbs free energy (and the equilibrium constant) for an interprotein or intraprotein interaction will depend on the standard volume change (AV) of the reaction: AG(p)=AG(o)+p AV ( 1) In (ttpn/(1 - ap))=p(AV/RT) + In (Kdo/nnC(n-l)) (2) where AG~) and AG(o) are the free energies for association/folding at pressure p and at atmospheric pressure respectively; Kdo is the equilibrium constant for dissociation or denaturation at atmospheric pressure; AV is the volume change, c~ is the extent of reaction at pressure p, and n is the number of dissociating subunits. The packing among the different components play a major role in stabilizing proteins. Because of the huge number of atoms involved, packing defects cannot be avoided, which lead to formation of cavities (8). Cavities are also related to the metastability of some proteins and macromolecular assemblages (9-12). Recent studies show that hydrostatic pressure is a useful tool to investigate packing and cavities (6, 10-12).
81 Coiled-coil proteins are excellent models to evaluate the balance between packing and hydrophobic interactions. Studies with the large two-stranded coiled-coil protein tropomyosin show that, under pressure, this protein is substantially denatured (13), but that enough residual coiled-coil region remains to maintain a 'denatured dimer'. On the other hand, the short coiled-coil dimer, formed by the 31 amino acid polypeptide chains of the leucine zipper GCN4-pl, dissociated completely under pressure to unfolded monomers, with the decrease in volume reflecting the rupture of the hydrophobic interactions (14). The volume change of unfolding for the leucine zipper dimer corresponds to a decrease in volume of 141 A 3 per molecule of dimer; whereas AV for a tetrameric mutant corresponds to a decrease in volume of 315 A 3 per tetramer. The larger volume change for the tetramer reflects a cavity in the middle of each leucine and isoleucine layer as seen by X-ray diffraction data (15). Royer and coworkers have obtained strong evidence (6) for the hypothesis that the internal cavities in the protein's structure contribute to the magnitude of the observed volume change. Experimental and theoretical approaches indicate that the underlying mechanism of pressure unfolding is the penetration of water into the protein matrix (16, 17). 1.2. Protein folding intermediates isolated under pressure It has been generally accepted that to unveil protein folding it is necessary to isolate and describe the intermediate structures. Some of these intermediates have been trapped under pressure and their dynamics and structure have been characterized by different methods. Several model proteins have been used for studies of protein folding and dimerization, including the E2 DNA-binding domain (E2DBDD) from human papillomavirus (18, 19), LexA repressor (20) and Arc repressor (16, 21-23). Arc repressor dimer reversibly dissociates into partially folded subunits under pressure (21) and the structure of the pressure-denatured monomer was partially determined by two-dimensional ~H NMR (24, 25). The pressure-induced population of partially folded intermediates has also been found for lysozyme (26), ribonuclease A (27), apomyoglobin (28), E2-DBD (19) the Ras binding domain of RalGEF (29) and DHFR (30) as determined by high-resolution NMR. The partially folded conformation of pressure-denatured proteins binds a substantial amount of water (16) and pressure denaturation does not proceed when water is withdrawn. This latter result has been corroborated by a theoretical approach that postulates the infiltration of water into the protein matrix ( 17, 3 I). Small monomeric proteins are the best models for studying protein folding but they are often too stable for the pressures normally attainable in standard equipment. Recently, we made use of the lower stability of a non-covalent complex formed by two complementary fragments of the chymotrypsin inhibitor-2 to study the folding equilibrium (32) and kinetics (33). The pressuredenaturation of the complex was completely independent of protein concentration, indicating the formation of a denatured heterodimer. The structural characteristics of the pressure denatured state of the CI-2 complex can be explained by a molten globule conformation, similar to those obtained for E2-DBD (18, 19) and Arc repressor (21-24). Further kinetic studies under pressure confirmed a denatured CI-2 complex with residual persistent structure (33). 2. PROTEIN FOLDING AND PROTEIN-DNA INTERACTIONS:
Genes are turned on and off by the binding of proteins to specific DNA sequences (34). In several cases, the formation of a specific protein-DNA complex is followed by an increase in nonpolar interactions and release of solvent, which explain the entropy-driven character of the specific binding, while solvent is not displaced in nonspecific complexes (23). On the other hand, eukaryotic transcription factors seems to be less sensitive to hydration effects and cooperative interactions among the proteins (homo or hetero interactions) play a dominant role in the binding to the specific DNA sites (19).
82
2.1. The role of specific DNA in tightening protein-protein interactions in Arc and LexA repressor:
DNA recognition by Arc repressor is tightly coupled to the competence of the given sequence of base pairs in stabilizing the native dimeric state of Arc repressor (22). The Arc repressor-operator DNA complex was cold-denatured at sub-zero temperatures under pressure, indicating that the formation of the specific complex is followed by an increase in nonpolar interactions and release of solvent, which would explain its entropy-driven character, while solvent would not be displaced in nonspecific complexes (23). With LexA repressor, we found that the protein is a dimer at nanomolar concentrations and that the dissociation constant lies in the picomolar range (20). Whereas nonspecific DNA has no stabilizing effects, specific DNA induces tightening of the dimer and a 750-fold decrease in the Ka. Accordingly, the LexA dimer only loses its ability to recognize a specific DNA sequence by RecA-induced autoproteolysis. 2.2. E2 DNA-Binding Domain (E2c) Protein of Human Papillomavirus 16.
E2 proteins from papillomaviruses are involved in viral DNA replication and regulation of transcription. The papillomavirus E2 protein is comprised of an N-terminal transactivation domain (E2n) separated by a flexible hinge region from the C-terminal DNA binding and dimerization domain (E2c). This dimeric domain, from a high-risk human papillomavirus (HPV-16), can be pressuredissociated to a monomeric state that presents substantial residual structure, through a highly reversible transition (18). Further unfolding of this high-pressure state can be achieved by low concentrations of urea. NMR studies have confirmed that the pressure-dissociated monomer of E2c is highly folded (19). The dissociation of E2c-DNA complexes (with specific and non-specific sequences) cannot be obtained using pressure alone (19). Taking into account that pressure and urea studies on the stability of E2c provide equivalent thermodynamic parameters, we obtained pressure isotherms in different urea concentration for both specific and non-specific DNAs. A plot of AGo versus pressure (Figure 2.A) is linear for both the E2c complexes of E2c dimer. Both specific and non-specific DNA sequences promoted a large stabilization of E2c dimer when compared to the protein alone, with a higher change in free energy for the specific DNA sequence (AAGo = 6.7 kcaVmol) than for the nonspecific DNA (AAGo = 4.5 kcal/mol). The difference between the effects produced by specific and nonspecific DNA is likely the basis for the sequence discrimination.
'8t.16
A
9 9
~ o
14
X
12
I
m!,
' IIII
4"t
_~
0
!
9
I
J
|
.... I
1
,
I
=------
0
_ m, IIII
0
+! II
II
'!
500 1000 1500 2000 2500 3000
Pr~n'e (bar) Figure 2. A) Plot of AGo versus pressure for (A) 1.0 ~tM E2c alone, ( I ) 250 nM [E2c:polyAT] complex (18 mer) and (O) 250 nM [E2c:E2DBS] (18 mer); B) Diagram of the monomer-dimer equilibrium of E2c bound to DNA. Volume change upon protein-protein and protein-DNA interaction correspond to hatched symbols.
83 The free energy change increases dramatically on DNA binding with no effect in the volume change, which indicates that no extra surface area is involved when E2c dissociates on the DNA. These results indicate that E2c dimers dissociate into monomers that are still bound to DNA (Figure 2.B). The retention of a large amount of tertiary structure by the dissociated monomer as evidenced by fluorescence (18) and NMR (19) data corroborate the hypothesis of a functional monomeric unit of E2c. These studies open new avenues to the development of drugs targeted to the monomer that will make it possible to prevent or treat human papillomavirus refection.
3. INCORRECT FOLDING, OFF-PATHWAY AGGREGATION AND AMYLOIDOGENESIS The competition between correct folding and misfolding, which in many proteins leads to formation of insoluble aggregates, is an important problem in the biotechnology industry and in human diseases. The off-pathway aggregation of proteins often occurs in vivo with heterologous proteins that are over-expressed in Escherwhm coil, resulting in the formation of inclusion bodies or amorphous aggregates within the cell. Pressure may affect the two pathways (aggregation/misfolding and correct folding) as recently demonstrated in the tailspike protein of bacteriophage P22 (35), rhodanese (36), myoglobin (37) and in the amyloidogenic protein transthyretin (38).
3.1. P22 Tailspike High pressure (2.4 kbar) treatment dissociated tailspike aggregates and resulted in formation of monomers and native folded trimers, without changes in buffer, dilution of reagents, or modification of reaction conditions (35). Our results also indicate that aggregation intermediates have similar specific chain recognition events as those that are involved in proper folding events, and suggest that an increased understanding of this specificity will lead improved folding methodologies. In the practical side, the use of high pressure appears to be an efficient tool for combating aggregation in a variety of research and industrial settings (35, 39).
3.5 lh at 3.5 kbar
=~~F~
C~
-
3.0
C, /
/
72h at 35 ibar
12~n
2.5
c,
.
o
o L.~ 2.0 :~ c 1.5
4
1.0
0
50
100
150
Time (min)
200
I
I
I
I
0
1000
2000
3000
Pressure (bar)
Figure 3. Left panel: Aggregation of TI'R after a cycle of compression-decompression as followed by light scattering (LS) increase. (O) At pH 5.6, 3.5 ~tM TFR were compressed at 3.5 kbar during 1 h or 72 h (A). Then pressure was release and the LS was followed and normalized by the initial value (LS0). Right panel: Bis-ANS binding as a function ofpH. (1) pH 7.5; (A) pH 5.6 and (O) pH 5. The hollowed symbols represent the decompression pathways. I T r r l = l~tM and [bis-ANS] = 10 ~tM.
3.2. Transthyretin (TTR) Transthyretin (TIR) is a tetrameric human plasma protein responsible for familial amyloid polyneuropathies (FAP) and senile systemic amyloidosis (SSA) (40). In vitro studies have shown that
84
at acidic pH T I R dissociates into partially folded monomers and refolding from this denatured state lead to aggregation (41). The main hypothesis is that aggregation occurs after dissociation into unfolded or partially folded monomers. Hydrostatic pressure produced reversible denaturation of TFR to a state that maintained residual structure and still associated as tetramers (Figure 3) (38). Among other properties, partially folded TI'R induced by pressure binds bis-ANS, similar to the aciddenatured protein that undergoes aggregation into amyloid fibrils (Figure 3). Under extreme high pressure, there was no aggregation, but at intermediate pressures, after compression and decompression, the protein aggregated into amyloid fibrils (38). These data show that pressure can be used as a controlled tool to explore intermediates, which are at the junction of the routes for folding and misfolding/aggregation.
3.3. Tumor Suppressor Protein P53 The tumor suppressor protein p53 is a 393 amino acid residue-transcriptional factor with an important role at the cell cycle control, specially after cellular stresses such as genotoxic damage, cytokines, hypoxia and alterations of ribonucleotide pools (42). Most human cancers (=50%) result from mutations in the p53 protein, mainly at its DNA binding domain (p53/DBD), affecting the DNA binding ability and/or the protein stability. Our results demonstrate that p53/DBD undergoes denaturation when submitted to hydrostatic pressures at 37~ leading to a misfolded and aggregated state. However, when we investigated the pressure effects at 4~ the protein denaturation was displaced to higher pressures and aggregation was non-cooperative with denaturation (Figure 4). Under pressure and subzero temperatures, we observed the formation of a partially folded state of the tumor suppressor protein with no aggregation. Taken together, these results suggest that, by carefully tuning high pressure and temperature, we might achieve stable intermediates of the protein folding, providing unprecedented targets for the development of antagonists capable of blocking protein misfolding and aggregation, potential drugs against tumoral diseases. The pressure denaturation curves have a large hysteresis, a characteristic of a metastable state.
~Lso O
o
.~4o
m 400 -80O
20
A------~
-12oo 0 <]
0
1
2
Pressure (kbar)
3
0
1 Pressure
2 3 (kl:mr)
Figure 4. Pressure-induced denaturation of the tumor suppressor protein p53 up to 3.1 kbar at 37~ (squares) or at 4~ (circles). Closed symbols correspond to the compression cycle and open symbols represent values after return to atmospheric pressure. (A) Center of spectral mass based on the tryptophan emission fluorescence.. (B) Light scattering data. Figure 5 shows a scheme that can be applied to tmnsthyretin and to p53/DBD. In both cases, pressure populates a partially folded, pre-aggregated state. Depending on the temperature and applied pressure, the decompression may lead to a metastable state or to an aggregate. The folding "halffunnel" represented in Figure 5 shows clearly how pressure can allow us to fully explore intermediates and determine their structures, potential targets against tumor and amyloidogenic diseases.
85
G (kcal/moi)
m
Figure 5. Free-energy diagram for dissociation, denaturation and aggregation of proteins. The native oligomer and the monomer are represented in circles, the altered, monomer, oligomer and aggregates in squares, the monomer with residual structure in triangle and unfolded protein as a line. Distance along vertical axis indicates differences in Gibbs free energy among different protein states.
4. CONCLUSIONS AND PERSPECTIVES: The increasing use of structural methods, such as NMR, to characterize the pressure-denatured state has placed pressure as one of the more promising techniques to unravel the missing links in the "Protein Folding Problem". We have learned that nonpolar interactions are determining factors in protein folding, protein-nucleic acid recognition and protein misfolding/aggregation. Pressure has permitted characterization of the high degree of plasticity of proteins. Recent detailed studies on the kinetics of pressure unfolding and folding (43) together with several theoretical approaches have increased our understanding of the pressure effects. Furthermore, the isolation of folding intermediates is crucial to the understanding of protein misfolding and protein aggregation, which have been recently tackled by pressure. As outlined in the free-energy diagram for protein folding and misfoding/aggregation (Figure 5), the population of an amyloidogenic intermediate without proceeding to aggregation is a unique property of pressure, which opens the prospect to characterize the structure of the amyloidogenic form.
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.
S.E. Radford, Trends Biochem. Sci., 25 (2000) 611. J.L. Silva and G. Weber, Annu. Rev. Phys. Chem., 44 (1993) 89. J. Jonas and A. Jonas, Annu. Rev. Biophys. Biomol. Struct., 23 (1994) 287. V.V. Mozhaev, K. Heremans, J. Frank, P. Masson and C. Balny, Proteins, 24 (1996) 81. J.L. Silva, D. Foguel, A.T. Da Poian and P.E. Prevelige, Curr. Opin. Struct. Biol., 6 (1996) 166. K.J. Frye and C.A. Royer, Protein Sci., 10 (1998) 2217. E.H. Koo, P.T. Lansbury Jr. and J.W. Kelly, Proc. Natl. Acad. Sei. USA, 96 (1999) 9989. A.E. Eriksson, W.A. Baase, X.-J. Zhang, D.W. Heinz, M. Blaber, E.P. Baldwin and B.W. Matthews, Science, 255 (1992) 178. 9. C. Lee, S.H. Park, M.Y. Lee and M.H. Yu, Proc. Natl. Acad. Sci. USA, 97 (2000) 7727. 10. D. Foguel, C.M. Teschke, P.E. Prevelige and J.L. Silva, Biochemistry, 34 (1995) 120. 11. P.C. De Souza, R. Tuma, P.E. Prevelige. J.L. Silva and D. Foguel, J. Mol. Biol., 287 (1999) 527.
86 12. L.P. Gaspar, A.F. Terezan, A.S. Pinheiro, D. Foguel, M.A. Rebello, J.L. Silva, J. Biol. Chem., 276 (2001) 7415. 13. M.C. Suarez, S.S. Lehrer and J.L. Silva, Biochemistry, 40 (2001) 1300. 14. J. L. Silva and J. R. Cortmes. Submitted (2001). 15. E.K. O'Shea, J.D. Klemm, P.S. Kim and T. Alber, Science, 254 (1991) 539. 16. A.C. Oliveira, L.P. Gaspar, A.T. Da Poian and J.L. Silva, J. Mol. Biol., 240 (1994) 184. 17. G. Hummer, S. Garde, A.E. Garcia, M.E. Paulaitis and L.R. Pratt, Proc. Natl. Acad. Sci. USA, 95 (1998) 1552. 18. D. Foguel, J.L. Silva and G. Prat-Gay, J. Biol. Chem., 273 (1998) 9050. 19. L.M.T.R Lima, D. Foguel and J.L. Silva, Proc. Natl. Acad. Sci. USA, 97 (2000) 14289. 20. R. Mohana-Borges, A.B. Pacheco, F.J. Sousa, D. Foguel, D.F. Almeida and J.L. Silva, J. Biol. Chem., 275 (2000) 4708. 21. J.L. Silva, C.F. Silveira, A. Correia and L. Pontes, J. Mol. Biol., 223 (1992) 45. 22. J.L. Silva and C.F. Silveira, Protein Science, 2 (1993) 945. 23. D. Foguel and J.L. Silva, Proc. Natl. Acad. Sci. USA, 91 (1994) 8244. 24. X. Peng, J. Jonas and J.L. Silva, Proc. Nail. Acad. Sci. USA, 90 (1993) 1776. 25. X.D. Peng, J. Jonas J and J.L. Silva, Biochemist~, 33 (1994) 8323. 26. D.P. Nash and J. Jonas, Biochemistry, 36 (1997) 14375. 27. J. Zhang, X. Peng, A. Jonas and J. Jonas, Biochemistry, 34 (1995) 8631. 28. S. E. Bondos, S. Sligar and J. Jonas, Biochim. Biophys. Acta, 1480 (2000) 353. 29. K. lnoue, H. Yamada, K. Akasaka, C. Herrmann, W. Kremer, T. Maurer, R Doker and H.R. Kalbitzer, Nat. Strucr Biol., 7 (2000) 547. 30. R. Kitahara, S. Sareth, H. Yamada, E. Ohmae, K. Gekko and K. Akasaka, Biochemistry, 39 (2000) 12789. 31. N. Hillson, J.N. Onuchic and A.E. Garcia, Proc. Natl. Acad. Sci. USA, 96 (1999) 14848. 32. R. Mohana-Borges, J.L. Silva and G. de Prat-Gay, J. Biol. Chem., 274 (1999) 7732. 33. R. Mohana-Borges, J.L. Silva, J. Ruiz-Sanz and G. de Prat-Gay, Proc. Natl. Acad. Sci. USA, 96 (1999) 7888. 34. C.O. Pabo and R.T. Sauer, Annu. Ver. Biochem., 61 (1992) 1053. 35. D. Foguel, C.R. Robinson, P.C. de Sousa Jr, J.L. Silva and A.S. Robinson, Biotechnol. Bioeng., 63 (1999) 552. 36. B.M. Gorovits and P.M. Horowitz, Biochemistry, 37 (1998)6132. 37. L. Smeller, P. Rubens and K. Heremans, Biochemistry, 38 (1999) 3816. 38. A.D. Ferrao-Gonzales, S.O. Souto, J.L. Silva and D. Foguel, Proc. Natl. Acad. Sci. USA, 97 (2000) 6445. 39. R.J. St John, J.F. Carpenter and T.W. Randolph, Proc. Natl. Acad. Sci. USA, 96 (1999) 13029. 40. J.W. Kelly, Curr. Opin. Struct. Biol., 8 (1998) 101. 41. Z. Lai, W. Colon and J.W. Kelly, Biochemistry, 35 (1996) 6470. 42. P.A. Hall and D.P. Lane, Curr. Biol., 7 (1997) R144. 43. C. A. Royer, in High Pressure Molecular Science, J. Jonas and R. Winter (eds.), Kluwer Academic Publishers, The Netherlands, (1999) p473-496.
79
Pressure Studies on Protein Folding, Misfolding, Protein-DNA Interactions and Amyloidogenesis. D. Ishimaru, L. M. T. R. Lima, A. Ferrfio-Gonzales, P. A. Quesado, L. M. Maiolino, J. L. Silva and D. Foguel. Centro National de Ressonfincia Magnetica Nuclear, Departamento de Bioquimica Medica, lnstituto de Ci~ncias Biom6dicas, Universidade Federal do Rio de Janeiro. ABSTRACT Hydrostatic pressure is a useful tool for dissecting macromolecular interactions at the molecular level. Nonpolar interactions are determining factors in protein folding, protein aggregation and protein-nucleic acid recognition. Because nonpolar interactions are entropic and compressible, they are more sensitive to pressure and low temperatures. We have studied problems of macromolecular recognition using hydrostatic pressure as the primary tool and employing several spectroscopic techniques, especially fluorescence, circular dichroism and high-resolution nuclear magnetic resonance. High pressure has the unique property of stabilizing partially folded states of a protein which degree of dissimilarity from the native state may range from drifted conformations to moltenglobule states. The competition between correct folding and misfolding, which in many proteins leads to formation of insoluble aggregates is an important problem in the biotechnology industry and in human diseases such as amyloidosis, Alzheimer's, prion and tumor diseases. Because of its ability to sequester folding intermediates, pressure has been used to direct the folding in one direction or the other and to explore intermediates, which are at the junction of the routes for folding and aggregation.
1. INTRODUCTION Proteins play the major functions in cells either as isolated molecules or forming macromolecular complexes held together by noncovalent interactions. A unique feature of the "oiologicar world, as compared to the 'inorganic' world, is the large number of interactions among its components that involve relatively weak energies (less than 10 kcal/mol). The energetic coupling among these interactions is likely the basis of the high specificity of macromolecular recognition. Life depends on macromolecular recognition, especially at the level of protein folding and protein-nucleic acid (NA) interactions. The mechanism by which unstructured proteins spontaneously fold to their native functional form is still not completely understood (1). To this end, it is necessary to isolate and describe the intermediate structures that occur during folding. Here, it is reviewed that in several cases these intermediates have been trapped under pressure and their dynamics and structure have been characterized by fluorescence, light scattering, NMR and hydrodynamic methods. Pressure affects the equilibrium between denatured or dissociated and native forms in the direction of the form that occupies a smaller volume (2-5). The structural region of the protein that is most sensitive to pressure is the hydrophobic core, especially when cavities are present (5, 6). It is described below the unique characteristic of pressure to isolate molten globules or to favor the state that appears to have a segment of the protein in a partially folded state ("partial" molten globule). The isolation of folding intermediates is crucial to the understanding of protein misfolding and protein aggregation. In the last decade, several diseases, such as Alzheimer's disease and other amyloidogenic diseases, spongiform encephalopathies (caused by prions), inherited emphysema, cystic fibrosis and likely many cancers are caused by protein misfolding (7). Biotechnology companies also face many problems with proteins aggregating into inclusion bodies when they are expressed in
80
bacteria. Figure 1 shows how functional proteins and macromolecular assemblies may divert from the folding pathway to a misfolding and aggregation dead end.
1
2
g;
3
/
"-...
4 Figure 1. Schematic representation of the folding and assembly of protein and multimolecular complexes. (1) Unfolded protein; (2) Folding intermediate; (3) Tertiary structure; (4) Protein misfoldmg - Aggregate; (5) Protein-DNA interaction; (6) Virus; (7) Protein-ligand interaction. To control misfolding and aggregation, we have to understand the energetics, the dynamics and the mechanisms of correct assembly. As described below, pressure has been useful not only to investigate the correct pathways (leading to 3, 5, 6 and 7 products - Figure 1), but also to the incorrect, misfolded form (product 4).
1.1. Protein folding and protein-protein interactions: The volume cavity model for pressureinduced dissociation and unfolding. The mechanisms by which one dimension information (linear sequence of amino acid residues) is transferred to four dimensions (3D-structure and dynamics) remains one of the major challenges in Biology. Although most of the thermodynamic studies have been done with conventional perturbing agents, such as high temperature, urea, guanidine, etc, they offer limited information, especially because they cause drastic changes in protein structure, not likely to occur at physiological conditions. Folding of a protein, or association of one protein with another, is accompanied by an increase in volume because of the additive effects of the formation of solvent-excluding cavities and the release of bound solvent. Water is released as nonpolar amino acid residues are buried, as well as when salt linkages are formed (2-5). The Gibbs free energy (and the equilibrium constant) for an interprotein or intraprotein interaction will depend on the standard volume change (AV) of the reaction: AG(p)=AG(o)+p AV ( 1) In (ttpn/(1 - ap))=p(AV/RT) + In (Kdo/nnC(n-l)) (2) where AG~) and AG(o) are the free energies for association/folding at pressure p and at atmospheric pressure respectively; Kdo is the equilibrium constant for dissociation or denaturation at atmospheric pressure; AV is the volume change, c~ is the extent of reaction at pressure p, and n is the number of dissociating subunits. The packing among the different components play a major role in stabilizing proteins. Because of the huge number of atoms involved, packing defects cannot be avoided, which lead to formation of cavities (8). Cavities are also related to the metastability of some proteins and macromolecular assemblages (9-12). Recent studies show that hydrostatic pressure is a useful tool to investigate packing and cavities (6, 10-12).
81 Coiled-coil proteins are excellent models to evaluate the balance between packing and hydrophobic interactions. Studies with the large two-stranded coiled-coil protein tropomyosin show that, under pressure, this protein is substantially denatured (13), but that enough residual coiled-coil region remains to maintain a 'denatured dimer'. On the other hand, the short coiled-coil dimer, formed by the 31 amino acid polypeptide chains of the leucine zipper GCN4-pl, dissociated completely under pressure to unfolded monomers, with the decrease in volume reflecting the rupture of the hydrophobic interactions (14). The volume change of unfolding for the leucine zipper dimer corresponds to a decrease in volume of 141 A 3 per molecule of dimer; whereas AV for a tetrameric mutant corresponds to a decrease in volume of 315 A 3 per tetramer. The larger volume change for the tetramer reflects a cavity in the middle of each leucine and isoleucine layer as seen by X-ray diffraction data (15). Royer and coworkers have obtained strong evidence (6) for the hypothesis that the internal cavities in the protein's structure contribute to the magnitude of the observed volume change. Experimental and theoretical approaches indicate that the underlying mechanism of pressure unfolding is the penetration of water into the protein matrix (16, 17). 1.2. Protein folding intermediates isolated under pressure It has been generally accepted that to unveil protein folding it is necessary to isolate and describe the intermediate structures. Some of these intermediates have been trapped under pressure and their dynamics and structure have been characterized by different methods. Several model proteins have been used for studies of protein folding and dimerization, including the E2 DNA-binding domain (E2DBDD) from human papillomavirus (18, 19), LexA repressor (20) and Arc repressor (16, 21-23). Arc repressor dimer reversibly dissociates into partially folded subunits under pressure (21) and the structure of the pressure-denatured monomer was partially determined by two-dimensional ~H NMR (24, 25). The pressure-induced population of partially folded intermediates has also been found for lysozyme (26), ribonuclease A (27), apomyoglobin (28), E2-DBD (19) the Ras binding domain of RalGEF (29) and DHFR (30) as determined by high-resolution NMR. The partially folded conformation of pressure-denatured proteins binds a substantial amount of water (16) and pressure denaturation does not proceed when water is withdrawn. This latter result has been corroborated by a theoretical approach that postulates the infiltration of water into the protein matrix ( 17, 3 I). Small monomeric proteins are the best models for studying protein folding but they are often too stable for the pressures normally attainable in standard equipment. Recently, we made use of the lower stability of a non-covalent complex formed by two complementary fragments of the chymotrypsin inhibitor-2 to study the folding equilibrium (32) and kinetics (33). The pressuredenaturation of the complex was completely independent of protein concentration, indicating the formation of a denatured heterodimer. The structural characteristics of the pressure denatured state of the CI-2 complex can be explained by a molten globule conformation, similar to those obtained for E2-DBD (18, 19) and Arc repressor (21-24). Further kinetic studies under pressure confirmed a denatured CI-2 complex with residual persistent structure (33). 2. PROTEIN FOLDING AND PROTEIN-DNA INTERACTIONS:
Genes are turned on and off by the binding of proteins to specific DNA sequences (34). In several cases, the formation of a specific protein-DNA complex is followed by an increase in nonpolar interactions and release of solvent, which explain the entropy-driven character of the specific binding, while solvent is not displaced in nonspecific complexes (23). On the other hand, eukaryotic transcription factors seems to be less sensitive to hydration effects and cooperative interactions among the proteins (homo or hetero interactions) play a dominant role in the binding to the specific DNA sites (19).
82
2.1. The role of specific DNA in tightening protein-protein interactions in Arc and LexA repressor:
DNA recognition by Arc repressor is tightly coupled to the competence of the given sequence of base pairs in stabilizing the native dimeric state of Arc repressor (22). The Arc repressor-operator DNA complex was cold-denatured at sub-zero temperatures under pressure, indicating that the formation of the specific complex is followed by an increase in nonpolar interactions and release of solvent, which would explain its entropy-driven character, while solvent would not be displaced in nonspecific complexes (23). With LexA repressor, we found that the protein is a dimer at nanomolar concentrations and that the dissociation constant lies in the picomolar range (20). Whereas nonspecific DNA has no stabilizing effects, specific DNA induces tightening of the dimer and a 750-fold decrease in the Ka. Accordingly, the LexA dimer only loses its ability to recognize a specific DNA sequence by RecA-induced autoproteolysis. 2.2. E2 DNA-Binding Domain (E2c) Protein of Human Papillomavirus 16.
E2 proteins from papillomaviruses are involved in viral DNA replication and regulation of transcription. The papillomavirus E2 protein is comprised of an N-terminal transactivation domain (E2n) separated by a flexible hinge region from the C-terminal DNA binding and dimerization domain (E2c). This dimeric domain, from a high-risk human papillomavirus (HPV-16), can be pressuredissociated to a monomeric state that presents substantial residual structure, through a highly reversible transition (18). Further unfolding of this high-pressure state can be achieved by low concentrations of urea. NMR studies have confirmed that the pressure-dissociated monomer of E2c is highly folded (19). The dissociation of E2c-DNA complexes (with specific and non-specific sequences) cannot be obtained using pressure alone (19). Taking into account that pressure and urea studies on the stability of E2c provide equivalent thermodynamic parameters, we obtained pressure isotherms in different urea concentration for both specific and non-specific DNAs. A plot of AGo versus pressure (Figure 2.A) is linear for both the E2c complexes of E2c dimer. Both specific and non-specific DNA sequences promoted a large stabilization of E2c dimer when compared to the protein alone, with a higher change in free energy for the specific DNA sequence (AAGo = 6.7 kcaVmol) than for the nonspecific DNA (AAGo = 4.5 kcal/mol). The difference between the effects produced by specific and nonspecific DNA is likely the basis for the sequence discrimination.
'8t.16
A
9 9
~ o
14
X
12
I
m!,
' IIII
4"t
_~
0
!
9
I
J
|
.... I
1
,
I
=------
0
_ m, IIII
0
+! II
II
'!
500 1000 1500 2000 2500 3000
Pr~n'e (bar) Figure 2. A) Plot of AGo versus pressure for (A) 1.0 ~tM E2c alone, ( I ) 250 nM [E2c:polyAT] complex (18 mer) and (O) 250 nM [E2c:E2DBS] (18 mer); B) Diagram of the monomer-dimer equilibrium of E2c bound to DNA. Volume change upon protein-protein and protein-DNA interaction correspond to hatched symbols.
83 The free energy change increases dramatically on DNA binding with no effect in the volume change, which indicates that no extra surface area is involved when E2c dissociates on the DNA. These results indicate that E2c dimers dissociate into monomers that are still bound to DNA (Figure 2.B). The retention of a large amount of tertiary structure by the dissociated monomer as evidenced by fluorescence (18) and NMR (19) data corroborate the hypothesis of a functional monomeric unit of E2c. These studies open new avenues to the development of drugs targeted to the monomer that will make it possible to prevent or treat human papillomavirus refection.
3. INCORRECT FOLDING, OFF-PATHWAY AGGREGATION AND AMYLOIDOGENESIS The competition between correct folding and misfolding, which in many proteins leads to formation of insoluble aggregates, is an important problem in the biotechnology industry and in human diseases. The off-pathway aggregation of proteins often occurs in vivo with heterologous proteins that are over-expressed in Escherwhm coil, resulting in the formation of inclusion bodies or amorphous aggregates within the cell. Pressure may affect the two pathways (aggregation/misfolding and correct folding) as recently demonstrated in the tailspike protein of bacteriophage P22 (35), rhodanese (36), myoglobin (37) and in the amyloidogenic protein transthyretin (38).
3.1. P22 Tailspike High pressure (2.4 kbar) treatment dissociated tailspike aggregates and resulted in formation of monomers and native folded trimers, without changes in buffer, dilution of reagents, or modification of reaction conditions (35). Our results also indicate that aggregation intermediates have similar specific chain recognition events as those that are involved in proper folding events, and suggest that an increased understanding of this specificity will lead improved folding methodologies. In the practical side, the use of high pressure appears to be an efficient tool for combating aggregation in a variety of research and industrial settings (35, 39).
3.5 lh at 3.5 kbar
=~~F~
C~
-
3.0
C, /
/
72h at 35 ibar
12~n
2.5
c,
.
o
o L.~ 2.0 :~ c 1.5
4
1.0
0
50
100
150
Time (min)
200
I
I
I
I
0
1000
2000
3000
Pressure (bar)
Figure 3. Left panel: Aggregation of TI'R after a cycle of compression-decompression as followed by light scattering (LS) increase. (O) At pH 5.6, 3.5 ~tM TFR were compressed at 3.5 kbar during 1 h or 72 h (A). Then pressure was release and the LS was followed and normalized by the initial value (LS0). Right panel: Bis-ANS binding as a function ofpH. (1) pH 7.5; (A) pH 5.6 and (O) pH 5. The hollowed symbols represent the decompression pathways. I T r r l = l~tM and [bis-ANS] = 10 ~tM.
3.2. Transthyretin (TTR) Transthyretin (TIR) is a tetrameric human plasma protein responsible for familial amyloid polyneuropathies (FAP) and senile systemic amyloidosis (SSA) (40). In vitro studies have shown that
84
at acidic pH T I R dissociates into partially folded monomers and refolding from this denatured state lead to aggregation (41). The main hypothesis is that aggregation occurs after dissociation into unfolded or partially folded monomers. Hydrostatic pressure produced reversible denaturation of TFR to a state that maintained residual structure and still associated as tetramers (Figure 3) (38). Among other properties, partially folded TI'R induced by pressure binds bis-ANS, similar to the aciddenatured protein that undergoes aggregation into amyloid fibrils (Figure 3). Under extreme high pressure, there was no aggregation, but at intermediate pressures, after compression and decompression, the protein aggregated into amyloid fibrils (38). These data show that pressure can be used as a controlled tool to explore intermediates, which are at the junction of the routes for folding and misfolding/aggregation.
3.3. Tumor Suppressor Protein P53 The tumor suppressor protein p53 is a 393 amino acid residue-transcriptional factor with an important role at the cell cycle control, specially after cellular stresses such as genotoxic damage, cytokines, hypoxia and alterations of ribonucleotide pools (42). Most human cancers (=50%) result from mutations in the p53 protein, mainly at its DNA binding domain (p53/DBD), affecting the DNA binding ability and/or the protein stability. Our results demonstrate that p53/DBD undergoes denaturation when submitted to hydrostatic pressures at 37~ leading to a misfolded and aggregated state. However, when we investigated the pressure effects at 4~ the protein denaturation was displaced to higher pressures and aggregation was non-cooperative with denaturation (Figure 4). Under pressure and subzero temperatures, we observed the formation of a partially folded state of the tumor suppressor protein with no aggregation. Taken together, these results suggest that, by carefully tuning high pressure and temperature, we might achieve stable intermediates of the protein folding, providing unprecedented targets for the development of antagonists capable of blocking protein misfolding and aggregation, potential drugs against tumoral diseases. The pressure denaturation curves have a large hysteresis, a characteristic of a metastable state.
~Lso O
o
.~4o
m 400 -80O
20
A------~
-12oo 0 <]
0
1
2
Pressure (kbar)
3
0
1 Pressure
2 3 (kl:mr)
Figure 4. Pressure-induced denaturation of the tumor suppressor protein p53 up to 3.1 kbar at 37~ (squares) or at 4~ (circles). Closed symbols correspond to the compression cycle and open symbols represent values after return to atmospheric pressure. (A) Center of spectral mass based on the tryptophan emission fluorescence.. (B) Light scattering data. Figure 5 shows a scheme that can be applied to tmnsthyretin and to p53/DBD. In both cases, pressure populates a partially folded, pre-aggregated state. Depending on the temperature and applied pressure, the decompression may lead to a metastable state or to an aggregate. The folding "halffunnel" represented in Figure 5 shows clearly how pressure can allow us to fully explore intermediates and determine their structures, potential targets against tumor and amyloidogenic diseases.
85
G (kcal/moi)
m
Figure 5. Free-energy diagram for dissociation, denaturation and aggregation of proteins. The native oligomer and the monomer are represented in circles, the altered, monomer, oligomer and aggregates in squares, the monomer with residual structure in triangle and unfolded protein as a line. Distance along vertical axis indicates differences in Gibbs free energy among different protein states.
4. CONCLUSIONS AND PERSPECTIVES: The increasing use of structural methods, such as NMR, to characterize the pressure-denatured state has placed pressure as one of the more promising techniques to unravel the missing links in the "Protein Folding Problem". We have learned that nonpolar interactions are determining factors in protein folding, protein-nucleic acid recognition and protein misfolding/aggregation. Pressure has permitted characterization of the high degree of plasticity of proteins. Recent detailed studies on the kinetics of pressure unfolding and folding (43) together with several theoretical approaches have increased our understanding of the pressure effects. Furthermore, the isolation of folding intermediates is crucial to the understanding of protein misfolding and protein aggregation, which have been recently tackled by pressure. As outlined in the free-energy diagram for protein folding and misfoding/aggregation (Figure 5), the population of an amyloidogenic intermediate without proceeding to aggregation is a unique property of pressure, which opens the prospect to characterize the structure of the amyloidogenic form.
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.
S.E. Radford, Trends Biochem. Sci., 25 (2000) 611. J.L. Silva and G. Weber, Annu. Rev. Phys. Chem., 44 (1993) 89. J. Jonas and A. Jonas, Annu. Rev. Biophys. Biomol. Struct., 23 (1994) 287. V.V. Mozhaev, K. Heremans, J. Frank, P. Masson and C. Balny, Proteins, 24 (1996) 81. J.L. Silva, D. Foguel, A.T. Da Poian and P.E. Prevelige, Curr. Opin. Struct. Biol., 6 (1996) 166. K.J. Frye and C.A. Royer, Protein Sci., 10 (1998) 2217. E.H. Koo, P.T. Lansbury Jr. and J.W. Kelly, Proc. Natl. Acad. Sei. USA, 96 (1999) 9989. A.E. Eriksson, W.A. Baase, X.-J. Zhang, D.W. Heinz, M. Blaber, E.P. Baldwin and B.W. Matthews, Science, 255 (1992) 178. 9. C. Lee, S.H. Park, M.Y. Lee and M.H. Yu, Proc. Natl. Acad. Sci. USA, 97 (2000) 7727. 10. D. Foguel, C.M. Teschke, P.E. Prevelige and J.L. Silva, Biochemistry, 34 (1995) 120. 11. P.C. De Souza, R. Tuma, P.E. Prevelige. J.L. Silva and D. Foguel, J. Mol. Biol., 287 (1999) 527.
86 12. L.P. Gaspar, A.F. Terezan, A.S. Pinheiro, D. Foguel, M.A. Rebello, J.L. Silva, J. Biol. Chem., 276 (2001) 7415. 13. M.C. Suarez, S.S. Lehrer and J.L. Silva, Biochemistry, 40 (2001) 1300. 14. J. L. Silva and J. R. Cortmes. Submitted (2001). 15. E.K. O'Shea, J.D. Klemm, P.S. Kim and T. Alber, Science, 254 (1991) 539. 16. A.C. Oliveira, L.P. Gaspar, A.T. Da Poian and J.L. Silva, J. Mol. Biol., 240 (1994) 184. 17. G. Hummer, S. Garde, A.E. Garcia, M.E. Paulaitis and L.R. Pratt, Proc. Natl. Acad. Sci. USA, 95 (1998) 1552. 18. D. Foguel, J.L. Silva and G. Prat-Gay, J. Biol. Chem., 273 (1998) 9050. 19. L.M.T.R Lima, D. Foguel and J.L. Silva, Proc. Natl. Acad. Sci. USA, 97 (2000) 14289. 20. R. Mohana-Borges, A.B. Pacheco, F.J. Sousa, D. Foguel, D.F. Almeida and J.L. Silva, J. Biol. Chem., 275 (2000) 4708. 21. J.L. Silva, C.F. Silveira, A. Correia and L. Pontes, J. Mol. Biol., 223 (1992) 45. 22. J.L. Silva and C.F. Silveira, Protein Science, 2 (1993) 945. 23. D. Foguel and J.L. Silva, Proc. Natl. Acad. Sci. USA, 91 (1994) 8244. 24. X. Peng, J. Jonas and J.L. Silva, Proc. Nail. Acad. Sci. USA, 90 (1993) 1776. 25. X.D. Peng, J. Jonas J and J.L. Silva, Biochemist~, 33 (1994) 8323. 26. D.P. Nash and J. Jonas, Biochemistry, 36 (1997) 14375. 27. J. Zhang, X. Peng, A. Jonas and J. Jonas, Biochemistry, 34 (1995) 8631. 28. S. E. Bondos, S. Sligar and J. Jonas, Biochim. Biophys. Acta, 1480 (2000) 353. 29. K. lnoue, H. Yamada, K. Akasaka, C. Herrmann, W. Kremer, T. Maurer, R Doker and H.R. Kalbitzer, Nat. Strucr Biol., 7 (2000) 547. 30. R. Kitahara, S. Sareth, H. Yamada, E. Ohmae, K. Gekko and K. Akasaka, Biochemistry, 39 (2000) 12789. 31. N. Hillson, J.N. Onuchic and A.E. Garcia, Proc. Natl. Acad. Sci. USA, 96 (1999) 14848. 32. R. Mohana-Borges, J.L. Silva and G. de Prat-Gay, J. Biol. Chem., 274 (1999) 7732. 33. R. Mohana-Borges, J.L. Silva, J. Ruiz-Sanz and G. de Prat-Gay, Proc. Natl. Acad. Sci. USA, 96 (1999) 7888. 34. C.O. Pabo and R.T. Sauer, Annu. Ver. Biochem., 61 (1992) 1053. 35. D. Foguel, C.R. Robinson, P.C. de Sousa Jr, J.L. Silva and A.S. Robinson, Biotechnol. Bioeng., 63 (1999) 552. 36. B.M. Gorovits and P.M. Horowitz, Biochemistry, 37 (1998)6132. 37. L. Smeller, P. Rubens and K. Heremans, Biochemistry, 38 (1999) 3816. 38. A.D. Ferrao-Gonzales, S.O. Souto, J.L. Silva and D. Foguel, Proc. Natl. Acad. Sci. USA, 97 (2000) 6445. 39. R.J. St John, J.F. Carpenter and T.W. Randolph, Proc. Natl. Acad. Sci. USA, 96 (1999) 13029. 40. J.W. Kelly, Curr. Opin. Struct. Biol., 8 (1998) 101. 41. Z. Lai, W. Colon and J.W. Kelly, Biochemistry, 35 (1996) 6470. 42. P.A. Hall and D.P. Lane, Curr. Biol., 7 (1997) R144. 43. C. A. Royer, in High Pressure Molecular Science, J. Jonas and R. Winter (eds.), Kluwer Academic Publishers, The Netherlands, (1999) p473-496.