- Email: [email protected]
Formation of a Molten-Globule-like State of Myoglobin in Aqueous Hexafluoroisopropanol
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
233, 687–691 (1997)
RC976524
Formation of a Molten-Globule-like State of Myoglobin in Aqueous Hexafluoroisopropanol John R. Cort and Niels H. Andersen1 Department of Chemistry, University of Washington, Seattle,Washington 98195
Received March 25, 1997
The effects of aqueous hexafluoroisopropanol (HFIP) media on the structure of myoglobin are reported. Circular dichroism (CD) spectra of this a-helical protein in as little as 4% (v/v) HFIP indicate that native-like amounts of secondary structure remain while rigid tertiary structure is lost. However, thermal studies suggest some residual cooperativity of unfolding in this state. At much higher HFIP concentrations, the helicity exceeds the native value and the protein behaves as a series of independent helices which do not interact with each other. We did not observe cold denaturation of myoglobin, even though this phenomenon has been observed for molten globule states of myoglobin, as well as for monomeric amphipathic a-helices when moderate quantities of HFIP are present. The pH dependence of trifluoroethanol-induced disruption of tertiary structure revealed that the degree of disruption increases as the enthalpic advantage of the folded state is diminished at low pH. q 1997 Academic Press
Aqueous fluoroalcohol (FA) solutions are known to promote helix formation in medium-sized peptides of modest helical propensity (1, 2). FA solutions are also known to affect protein structure. For example, aqueous 2,2,2-trifluoroethanol (TFE) solutions have been found to disrupt quaternary structure in the dimeric protein Troponin C (3), to increase helicity of many bsheet-containing proteins (4), and to produce a moltenglobule state of lysozyme (5). In some cases, non-fluorinated alcohols produce similar effects in proteins (6). TFE is by far the most common FA used in polypeptide studies and was used for the protein studies cited above. However, a second FA, 1,1,1,3,3,3-hexafluoroisopropanol (HFIP), has also been used with peptides and proteins, albeit less frequently and occasionally 1 To whom correspondence should be addressed. Fax: (206) 685– 8665. E-mail: [email protected]. Abbreviations: CD, circular dichroism; FA, fluoroalcohol; HFIP, hexafluoroisopropanol; TFE, trifluoroethanol; UV, ultraviolet.
with results not observed in aqueous TFE. Aqueous HFIP has been used to induce helicity in medium-sized peptides in much the same way as TFE (7, 8). It has also been used both alone and with water as a cosolvent to induce a-helix and b-sheet formation in amino acid homo- and co-polymers (9, 10, 11). Aqueous HFIP has been used as a sovent for proteins as well, inducing conformational changes in histones (12) and, remarkably, unfolding a 60-residue snake toxin with four disulfide bonds (13). Recent work in this research group has led to the discovery of some novel behavior of medium-sized peptides in 4 to 10 volume-% HFIP: formation of b-sheet structure in a 37-residue amyloid-forming peptide which becomes helical on further addition of HFIP (14), and ‘cold denaturation’ of a variety of monomeric peptide helices (15). In the latter report, we postulated that cold-induced unfolding of peptide helices in aqueous HFIP was due to a large positive DCp reflecting solvation of non-polar side chains. If non-polar side chains in the molten-globule state of a protein are accesible to the bulk solvent, and if hydrophobic forces are primarily responsible for the stabilization of this state, then one might expect cold denaturation of helical proteins in this medium as well. We chose myoglobin to investigate this question. Myoglobin is an a-helical protein which displays cold denaturation at pH 3.6-3.9 (16) when it is on the verge of acid denaturation. Several molten-globule states of apomyoglobin have been reported: notably, upon addition of millimolar quantities of trichloroacetate at pH 2, unfolded apomyoglobin forms a molten-globule state that displays cold denaturation (17) and has a radius of gyration similar to that of the 100ms intermediate in the apomyglobin refolding path (18). Partially folded and highly helical (greater than native helicity) states of apomyoglobin have been observed at 10 and 40% TFE, respectively; the highly helical state was also observed for the holoprotein (4). MATERIALS AND METHODS Horse skeletal muscle myoglobin was purchased from Sigma (St. Louis, MO) and used without further purification. HFIP (DuPont) 0006-291X/97 $25.00
687
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
Vol. 233, No. 3, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. (A) Far UV CD spectral traces for the native state of myoglobin at pH 4 (10mM acetate), upon acidification of to pH 3.2 and upon addition HFIP (to 4 vol-%) to both of these solutions. (B) Near UV spectra for the same four solutions. The curves are scaled to [u]222 Å 023,7007 for the native state (17); see Methods. However, our value for the near UV band around 270 nm is higher: [u]270 – 4 Å / 200 (versus /120 for myoglobin and /30 for apomyoglobin (17)). The published [u]222 for apomyoglobin is 019,000 (17). Our value for the 415 nm band of myoglobin in these conditions is [u]415 Å /320.
and TFE (Sigma) were distilled prior to use. CD spectra were acquired on a Jasco J-720 which had been calibrated as previously described (19) and are reported in units of residue-molar ellipticity ([u], deg-cm2 / dmol-residue). The quantity of myoglobin present in aqueous stock solutions was determined by CD and is based on the [u]222 value at pH 6.0, 023,270 (17) [other [u]222 values have been reported for different conditions: ca. 022,000 at pH 3.72 (16), and 024,900 at pH 6 (4)]. A metal-jacketed cell holder connected to a thermostated circulator containing a glycol/water mixture was used to maintain constant temperatures. 1 mm cells were used for far UV measurements and a 1 cm cell was used for measurements in the near UV.
RESULTS AND DISCUSSION To visualize what could be small effects, we focused on an acid pH range in which DHU is less positive. The far and near UV CD spectra which we recorded for myoglobin at pH 4 and upon addition of HFIP to either the pH 4 solution or with prior acid-denaturation (acidification to pH 3.2) are compared to those of the aciddenatured state in Fig. 1. Upon addition of HFIP (to 4 vol-%), the heme signal at 415nm and the aromatic sidechain band at circa 274 nm disappeared by the time the CD was recorded. This suggests that tertiary structure is thoroughly and rapidly disrupted by HFIP, and the heme binding site no longer exists or has greatly diminished affinity for heme. In contrast, the [u]222 value is similar to that of apomyoglobin, which retains tertiary structure. Myoglobin retains circa 85% of its helical secondary structure in 4 vol-% HFIP. Among the hallmarks of the molten-globule state are the retention of a large portion of the native secondary structure with some compactness but without ‘rigid’ tertiary structure (20); the partially denatured state of myoglobin present at 4 vol-% HFIP appears to fit this description. Upon further addition of HFIP, 0[u]222 actually increases. The formation of additional helical structure at 25 vol-% HFIP likely reflects the general helix-stabilizing properties of this FA.
To assess whether these states retain some cooperatively-formed structure, we examined the thermal dependence of [u]222 at several levels of HFIP content (Fig. 2). Monomeric linear peptide helices in aqueous FA media typically display a nearly linear dependence of [u]222 upon temperature (21) reflecting both increasing endfraying and more extensive helix unfolding due to the positive configurational DSU term. Rather similar helix ‘melting’ behavior is observed in aqueous media: e.g. [u]222 Å 020200 / 320rT(7C) for a RNAse C-peptide analog (unpublished data, this laboratory). A gradient of /1037/7C has been reported for longer helices (22); also see data reported by Munoz and Serrano (23). Linear dependence of [u]222 upon temperature has also been observed for hen egg-white lysozyme in 50 vol-% TFE, [u]222 Å 027500 / 235rT (5). At 25% HFIP, the myoglobin ‘melting transition’ was linear over the 0 0 70 7C temperature range examined, [u]222 Å 030200 / 162rT, (R2 Å 0.99). This suggests a fully unfolded structure with no compactness, a series of independent peptide helices linked by random coil strands. At 2 - 8 vol-% HFIP, some degree of cooperativity is apparent in the melting curves, decreasing as the level of HFIP increases. Even at 8% HFIP the temperature gradient of CD helicity below the melting point, 020000 / 62rT, resembles that of the native state (023900 / 32rT, T Å 0 0 407C at pH 4) rather than that expected for independent helices. We view this as further evidence of a molten globule state at low HFIP concentrations, where some compactness is present and the protein resists unfolding. Nonetheless, cold denaturation of myoglobin was not observed at any HFIP concentration. At 8% HFIP the CD spectra for the thermal unfolding experiment display a blue-shifted isodichroic point [200 nm vs. 204 nm for both 4% and 25% HFIP] (Fig. 3). At this time we do not have a rationale for this observation. However, we note that our lab and others have noted atypical behavior of peptides between 4 - 10 vol-
688
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
Vol. 233, No. 3, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Plots of [u]222 vs. temperature for myoglobin in pH 4 buffer, and with FA added. (A) HFIP; (B) TFE.
% HFIP (14, 15, 24, 25). The thermal unfolding isodichroic point remains at 204 nm for all proportions of TFE in water.2 Although the CD spectra of myoglobin at 4 vol-% HFIP (Fig. 1) are essentially identical (both lack indicators of tertiary structure in the near UV) whether the sample is prepared from the native state or the aciddenatured state; the resulting solutions have different properties. For example, the solution formed from the acid denatured state tended to become cloudy and aggregate after storage for a day at room temperature, while the ‘molten-globule state’ solutions which had never been taken below pH 4 remained clear for weeks. 2 Some published physical data also suggest that HFIP/water is a distinctly different solvent system from TFE/water and is probably highly structured, in particular at 4–10 vol-%. Denda et al. (26) report large negative excess molar enthalpies of mixing for HFIP and water in this composition range, while those for TFE/water were positive at compositions with an equivalent helix-promoting capability. Kuprin et al. (27) note that aqueous HFIP displays enhanced low angle X-ray scattering which falls off at the composition where the helix-promoting effect levels. Finally, Murto et al. (28) noted that HFIP/water displays a sudden large negative partial HFIP molar volume difference (v 0 v0) at xHFIP Å 0.008 (about 4.5 vol-%). TFE/ water mixtures displayed a much smaller (v 0 v0) anomaly at a comparable xFA level (29) which, for TFE, is well below the level where helix induction begins.
The solution produced by adding HFIP to the pH 3.2 solution probably contains a larger equilibrium population of the random coil state due to the higher acidity. Unfortunately the interpretation of pH electrode measurements in aqueous HFIP [pKa Å 9.3] (30) has not been clarified. TFE is considerably less acidic [pKa Å 12.4] (31) and its effects can be examined over a wider pH range. Shiraki et al. (4) report identical [u]222 values (026110{1507) for equine holo- and apomyoglobin in 40% aqueous TFE at pH 2 and 6, versus 249007 and 196007, respectively, for these two forms of the protein in 0% TFE at pH 6. Near UV CD spectra of myoglobin in 10% TFE at pH 6.5, 5.0, 4.0 and 3.2 appear in Fig. 4. At pH 6.5, partial loss of the 274nm feature (which is due to both the heme unit and aryl side-chain chromophores) is more advanced than loss of the 415nm heme feature, which occurs incompletely at this pH. No additional decrease in the heme signal occurs after about 15 minutes. In 10% TFE, loss of structured aromatic chains also appears to be incomplete even at pH 4; some residual positive ellipticity is retained at circa 260nm. Based on this comparison, HFIP is a much more effective disruptor of the tertiary structure of myoglobin. We expect that this will be a general feature. Thermal unfolding experiments at 12 and 40%
FIG. 3. CD spectra (wavelength, nm vs. residue molar ellipticity, [u]) of myoglobin at different temperatures in pH 4 buffer with (left to right) 4, 8, and 25% HFIP present, showing the blue shifted thermal isodichroic at 8% HFIP. Temperatures (K) at which the spectra in each thermal study were recorded are: for 4% HFIP— 275, 297, 327, 355; for 8% HFIP— 275, 297, 333, 355; for 25% HFIP— 275, 297, 341. 689
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
Vol. 233, No. 3, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Near UV CD spectra of myoglobin solutions containing 10% TFE at different pH values, and at pH 4 in the absence of TFE.
TFE (Fig. 2, panel B) yield curves similar to those seen for 4% and 25% HFIP, respectively, though the slopes in the high temperature, linear portion of the curves are slightly steeper for 40% TFE ([u]222 Å 028760 / 191rT). This probably reflects the decreased helix-stabilizing effect of TFE versus HFIP. CONCLUSIONS The addition of HFIP to as little as 4 vol-% disrupts the tertiary structure of aqueous myoglobin with complete loss of the aromatic side chain CD signals in the near UV as well as the bands in the visible and near UV due to the heme unit. Comparisons with TFE indicate that comparable behavior requires more than 10 vol-% of TFE but the disruption or fluxionalization of tertiary structure probably occurs in much the same way.3 At higher levels of added fluoroalcohol, myoglobin behaves as a linked set of independent helices. The myoglobin state that results upon addition of limited amounts of HFIP to the native form displays properties suggesting molten globule character, including some cooperativity in melting. The degree of helicity (essentially equal to that of native apomyoglobin) and dimished thermal fraying of this state would not be expected in the absence of some stabilization by tertiary interactions. Analysis of inherent helix propensities along the protein sequence - using i) a modified Lifson-Roig helix/coil transition model (33), ii) molecular dynamics trajectories (34) and iii) NMR spectros-
copy of peptidic fragments of the sequence (35) - all indicate that only two or three of the helices would be well populated in this medium if each were behaving independently. Furthermore, solvent exposed, partially formed helices would be expected to display unfolding upon cooling in this medium (15). The complete absence of cold denaturation at all HFIP concentrations examined is remarkable. The trichloroacetate-induced molten globule state of apomyoglobin displays dramatic cold denaturation at pH 2 (17) and isolated amphipathic helices display this behavior in 4 - 10 vol-% HFIP. We suggest that this indicates a state that is still compact in which the lipophilic sides of the helices have little bulk solvent exposure. Quite possibly the interior of the globule is accessible to HFIP but not to water. While cold-induced unfolding was not observed, the thermal isodichroic observed in 8 vol-% HFIP does indicate that this solvent composition has unique properties for protein solvation as well as peptides. Studies of the effects of HFIP upon other proteins and upon helix bundles are in progress in order to provide further definition of the physicochemical basis for both the helix-inducing and tertiary structure disrupting properties of fluoroalcohols. REFERENCES 1. Nelson, J. W., and Kallenbach, N. R. (1986) Proteins 1, 211–217. 2. Cammers-Goodwin, A., Allen, T. J., Oslick, S. L., McClure, K. F., Lee, J. H., and Kemp, D. S. (1996) J. Am . Chem. Soc. 118, 3082– 3092. 3. Slupsky, C. M., Kay, C. M., Reinach, F. C., Smillie, L. B., and Sykes, B. D. (1995) Biochemistry 34, 7365–7375. 4. Shiraki, K., Nishikawa, K., and Goto, Y. (1995) J. Mol. Biol. 245, 180–194. 5. Buck, M., Radford, S. E., and Dobson, C. M. (1993) Biochemistry 32, 669–678. 6. Bychkova, V. E., Dujsekina, A. E., Klenin, S. I., Tiktopulo, E. I., Uversky, V. N., and Ptitsyn, O. B. (1996) Biochemistry 35, 6058– 6063. 7. Mayer, R., Lancelot, G., and H’elene, C. (1983) FEBS Lett. 153, 339–344. 8. Andersen, N. H., Harris, S. M., Lee, V. G., Liu, E. C.-K., Moreland, S., and Hunt, J. T. (1995) Bioorg. and Med. Chem. 3, 113–124. 9. Parrish, J. R., and Blout, E. R. (1971) Biopolymers 10, 1491– 1512. 10. Parrish, J. R., and Blout, E. R. (1972) Bioploymers 11, 1001– 1020.
3
A comparative study of bovine b-lactoglobulin in aqueous HFIP and TFE appeared (32) as this paper was being submitted. HFIP effects a native (b-sheet) to helical state transition with DGU Å 0 at circa 6 vol-% (0.60 M, approximately one third the concentration required for a comparable effect by TFE). The authors note that the enhanced activity of HFIP cannot be explained by the additive effects of the constituent parts (CF3 surface area, etc.) and suggest that the ‘‘cooperative formation of micelle-like clusters of HFIP is important.’’ Evidence for other unusual media effects at this HFIP / water composition are given in footnote 2.
11. Chen, F., Lepore, G., and Goodman, M. (1974) Macromolecules 7, 779–783. 12. Beaudette, N. V., Okabayashi, H., and Fasman, G. D. (1982) Biochemistry 21, 1765–1772. 13. Galat, A., Degelaen, J. P., Yang, C. C., and Blout, E. R. (1981) Biochemistry 20, 7415–7423. 14. Cort, J., Liu, Z., Lee, G., Harris, S. M., Prickett, K. S., Gaeta, L. S. L., and Andersen, N. H. (1994) Biochem Biophys. Res. Commun. 204, 1088–1095.
690
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
Vol. 233, No. 3, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
15. Andersen, N. H., Cort, J. R., Liu, Z., Sjoberg, S. J., and Tong, H. (1996) J. Am. Chem. Soc. 118, 10309–10310. 16. Privalov, P. L., Griko, Yu. V., and Venyaminov, S. Yu. (1986) J. Mol. Biol. 190, 487–498. 17. Nishii, I., Kataoka, M., Tokunaga, F., and Goto, Y. (1994) Biochemistry 33, 4903–4909. 18. Eliezer, D., Jennings, P. A., Wright, P. E., Doniach, S., Hodgson, K. O., Tsuruta, H. (1995) Science 270, 487–488. 19. Andersen, N. H., and Palmer, R. B. (1994) BioOrg. Med. Chem. Lett. 4, 817–822. 20. Ptitsyn, O. B. (1995) Adv. Protein Chem. 47, 83–229. 21. Merutka, G., and Stellwagen, E. (1990) Biochemistry 29, 894– 898. 22. Scholtz, J. M., Qian, H., York, E. J., Stewart, J. M., and Baldwin, R. L. (1991) Biopolymers 31, 1463–1470. 23. Munoz, V., and Serrano, L. (1995) J. Mol. Biol. 245, 297–308. 24. Wei, J., and Fasman, G. D. (1995) Biochemistry 34, 6408–6415. 25. Mchaourab, H. S., Hyde, J. S., and Felix, J. B. (1993) Biochemistry 32, 11895–11902.
26. Denda, M., Touhara, H., and Nakanishi, K. (1987) J. Chem. Thermodynamics 19, 539–542. 27. Kuprin, S., Graslund, A., Ehrenberg, A., and Koch, M. H. J. (1995) Biochem. Biophys. Res. Commun. 217, 1151–1156. 28. Murto, J., Kivinen, A., Kivimaa, S., and Laakso, R. (1967) Suomen Kemistilehti B 40, 250–257. 29. Murto, J., and Heino, E.-L. (1966) Suomen Kemistilehti B 39, 263–266. 30. Dyatkin, B. L., Mochalina, E. P., and Knunyants, I. L. (1965) Tetrahedron 21, 2991–2995. 31. Ballinger, P., and Long, F. A. (1959) J. Am . Chem. Soc. 81, 1050–1053. 32. Hirota, N., Mizuno, K., and Goto, Y. (1997) Protein Sci. 6, 416– 421. 33. Andersen, N. H., and Tong, H. (1997) Protein Sci., in press. 34. Hirst, J. D., and Brooks, C. L. III (1995) Biochemistry 34, 7614– 7621. 35. Waltho, J. P., Feher, V. A, Merutka, G., Dyson, H. J., and Wright, P. E. (1993) Biochemistry 32, 6337–6347.
691
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
233, 687–691 (1997)
RC976524
Formation of a Molten-Globule-like State of Myoglobin in Aqueous Hexafluoroisopropanol John R. Cort and Niels H. Andersen1 Department of Chemistry, University of Washington, Seattle,Washington 98195
Received March 25, 1997
The effects of aqueous hexafluoroisopropanol (HFIP) media on the structure of myoglobin are reported. Circular dichroism (CD) spectra of this a-helical protein in as little as 4% (v/v) HFIP indicate that native-like amounts of secondary structure remain while rigid tertiary structure is lost. However, thermal studies suggest some residual cooperativity of unfolding in this state. At much higher HFIP concentrations, the helicity exceeds the native value and the protein behaves as a series of independent helices which do not interact with each other. We did not observe cold denaturation of myoglobin, even though this phenomenon has been observed for molten globule states of myoglobin, as well as for monomeric amphipathic a-helices when moderate quantities of HFIP are present. The pH dependence of trifluoroethanol-induced disruption of tertiary structure revealed that the degree of disruption increases as the enthalpic advantage of the folded state is diminished at low pH. q 1997 Academic Press
Aqueous fluoroalcohol (FA) solutions are known to promote helix formation in medium-sized peptides of modest helical propensity (1, 2). FA solutions are also known to affect protein structure. For example, aqueous 2,2,2-trifluoroethanol (TFE) solutions have been found to disrupt quaternary structure in the dimeric protein Troponin C (3), to increase helicity of many bsheet-containing proteins (4), and to produce a moltenglobule state of lysozyme (5). In some cases, non-fluorinated alcohols produce similar effects in proteins (6). TFE is by far the most common FA used in polypeptide studies and was used for the protein studies cited above. However, a second FA, 1,1,1,3,3,3-hexafluoroisopropanol (HFIP), has also been used with peptides and proteins, albeit less frequently and occasionally 1 To whom correspondence should be addressed. Fax: (206) 685– 8665. E-mail: [email protected]. Abbreviations: CD, circular dichroism; FA, fluoroalcohol; HFIP, hexafluoroisopropanol; TFE, trifluoroethanol; UV, ultraviolet.
with results not observed in aqueous TFE. Aqueous HFIP has been used to induce helicity in medium-sized peptides in much the same way as TFE (7, 8). It has also been used both alone and with water as a cosolvent to induce a-helix and b-sheet formation in amino acid homo- and co-polymers (9, 10, 11). Aqueous HFIP has been used as a sovent for proteins as well, inducing conformational changes in histones (12) and, remarkably, unfolding a 60-residue snake toxin with four disulfide bonds (13). Recent work in this research group has led to the discovery of some novel behavior of medium-sized peptides in 4 to 10 volume-% HFIP: formation of b-sheet structure in a 37-residue amyloid-forming peptide which becomes helical on further addition of HFIP (14), and ‘cold denaturation’ of a variety of monomeric peptide helices (15). In the latter report, we postulated that cold-induced unfolding of peptide helices in aqueous HFIP was due to a large positive DCp reflecting solvation of non-polar side chains. If non-polar side chains in the molten-globule state of a protein are accesible to the bulk solvent, and if hydrophobic forces are primarily responsible for the stabilization of this state, then one might expect cold denaturation of helical proteins in this medium as well. We chose myoglobin to investigate this question. Myoglobin is an a-helical protein which displays cold denaturation at pH 3.6-3.9 (16) when it is on the verge of acid denaturation. Several molten-globule states of apomyoglobin have been reported: notably, upon addition of millimolar quantities of trichloroacetate at pH 2, unfolded apomyoglobin forms a molten-globule state that displays cold denaturation (17) and has a radius of gyration similar to that of the 100ms intermediate in the apomyglobin refolding path (18). Partially folded and highly helical (greater than native helicity) states of apomyoglobin have been observed at 10 and 40% TFE, respectively; the highly helical state was also observed for the holoprotein (4). MATERIALS AND METHODS Horse skeletal muscle myoglobin was purchased from Sigma (St. Louis, MO) and used without further purification. HFIP (DuPont) 0006-291X/97 $25.00
687
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
Vol. 233, No. 3, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. (A) Far UV CD spectral traces for the native state of myoglobin at pH 4 (10mM acetate), upon acidification of to pH 3.2 and upon addition HFIP (to 4 vol-%) to both of these solutions. (B) Near UV spectra for the same four solutions. The curves are scaled to [u]222 Å 023,7007 for the native state (17); see Methods. However, our value for the near UV band around 270 nm is higher: [u]270 – 4 Å / 200 (versus /120 for myoglobin and /30 for apomyoglobin (17)). The published [u]222 for apomyoglobin is 019,000 (17). Our value for the 415 nm band of myoglobin in these conditions is [u]415 Å /320.
and TFE (Sigma) were distilled prior to use. CD spectra were acquired on a Jasco J-720 which had been calibrated as previously described (19) and are reported in units of residue-molar ellipticity ([u], deg-cm2 / dmol-residue). The quantity of myoglobin present in aqueous stock solutions was determined by CD and is based on the [u]222 value at pH 6.0, 023,270 (17) [other [u]222 values have been reported for different conditions: ca. 022,000 at pH 3.72 (16), and 024,900 at pH 6 (4)]. A metal-jacketed cell holder connected to a thermostated circulator containing a glycol/water mixture was used to maintain constant temperatures. 1 mm cells were used for far UV measurements and a 1 cm cell was used for measurements in the near UV.
RESULTS AND DISCUSSION To visualize what could be small effects, we focused on an acid pH range in which DHU is less positive. The far and near UV CD spectra which we recorded for myoglobin at pH 4 and upon addition of HFIP to either the pH 4 solution or with prior acid-denaturation (acidification to pH 3.2) are compared to those of the aciddenatured state in Fig. 1. Upon addition of HFIP (to 4 vol-%), the heme signal at 415nm and the aromatic sidechain band at circa 274 nm disappeared by the time the CD was recorded. This suggests that tertiary structure is thoroughly and rapidly disrupted by HFIP, and the heme binding site no longer exists or has greatly diminished affinity for heme. In contrast, the [u]222 value is similar to that of apomyoglobin, which retains tertiary structure. Myoglobin retains circa 85% of its helical secondary structure in 4 vol-% HFIP. Among the hallmarks of the molten-globule state are the retention of a large portion of the native secondary structure with some compactness but without ‘rigid’ tertiary structure (20); the partially denatured state of myoglobin present at 4 vol-% HFIP appears to fit this description. Upon further addition of HFIP, 0[u]222 actually increases. The formation of additional helical structure at 25 vol-% HFIP likely reflects the general helix-stabilizing properties of this FA.
To assess whether these states retain some cooperatively-formed structure, we examined the thermal dependence of [u]222 at several levels of HFIP content (Fig. 2). Monomeric linear peptide helices in aqueous FA media typically display a nearly linear dependence of [u]222 upon temperature (21) reflecting both increasing endfraying and more extensive helix unfolding due to the positive configurational DSU term. Rather similar helix ‘melting’ behavior is observed in aqueous media: e.g. [u]222 Å 020200 / 320rT(7C) for a RNAse C-peptide analog (unpublished data, this laboratory). A gradient of /1037/7C has been reported for longer helices (22); also see data reported by Munoz and Serrano (23). Linear dependence of [u]222 upon temperature has also been observed for hen egg-white lysozyme in 50 vol-% TFE, [u]222 Å 027500 / 235rT (5). At 25% HFIP, the myoglobin ‘melting transition’ was linear over the 0 0 70 7C temperature range examined, [u]222 Å 030200 / 162rT, (R2 Å 0.99). This suggests a fully unfolded structure with no compactness, a series of independent peptide helices linked by random coil strands. At 2 - 8 vol-% HFIP, some degree of cooperativity is apparent in the melting curves, decreasing as the level of HFIP increases. Even at 8% HFIP the temperature gradient of CD helicity below the melting point, 020000 / 62rT, resembles that of the native state (023900 / 32rT, T Å 0 0 407C at pH 4) rather than that expected for independent helices. We view this as further evidence of a molten globule state at low HFIP concentrations, where some compactness is present and the protein resists unfolding. Nonetheless, cold denaturation of myoglobin was not observed at any HFIP concentration. At 8% HFIP the CD spectra for the thermal unfolding experiment display a blue-shifted isodichroic point [200 nm vs. 204 nm for both 4% and 25% HFIP] (Fig. 3). At this time we do not have a rationale for this observation. However, we note that our lab and others have noted atypical behavior of peptides between 4 - 10 vol-
688
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
Vol. 233, No. 3, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Plots of [u]222 vs. temperature for myoglobin in pH 4 buffer, and with FA added. (A) HFIP; (B) TFE.
% HFIP (14, 15, 24, 25). The thermal unfolding isodichroic point remains at 204 nm for all proportions of TFE in water.2 Although the CD spectra of myoglobin at 4 vol-% HFIP (Fig. 1) are essentially identical (both lack indicators of tertiary structure in the near UV) whether the sample is prepared from the native state or the aciddenatured state; the resulting solutions have different properties. For example, the solution formed from the acid denatured state tended to become cloudy and aggregate after storage for a day at room temperature, while the ‘molten-globule state’ solutions which had never been taken below pH 4 remained clear for weeks. 2 Some published physical data also suggest that HFIP/water is a distinctly different solvent system from TFE/water and is probably highly structured, in particular at 4–10 vol-%. Denda et al. (26) report large negative excess molar enthalpies of mixing for HFIP and water in this composition range, while those for TFE/water were positive at compositions with an equivalent helix-promoting capability. Kuprin et al. (27) note that aqueous HFIP displays enhanced low angle X-ray scattering which falls off at the composition where the helix-promoting effect levels. Finally, Murto et al. (28) noted that HFIP/water displays a sudden large negative partial HFIP molar volume difference (v 0 v0) at xHFIP Å 0.008 (about 4.5 vol-%). TFE/ water mixtures displayed a much smaller (v 0 v0) anomaly at a comparable xFA level (29) which, for TFE, is well below the level where helix induction begins.
The solution produced by adding HFIP to the pH 3.2 solution probably contains a larger equilibrium population of the random coil state due to the higher acidity. Unfortunately the interpretation of pH electrode measurements in aqueous HFIP [pKa Å 9.3] (30) has not been clarified. TFE is considerably less acidic [pKa Å 12.4] (31) and its effects can be examined over a wider pH range. Shiraki et al. (4) report identical [u]222 values (026110{1507) for equine holo- and apomyoglobin in 40% aqueous TFE at pH 2 and 6, versus 249007 and 196007, respectively, for these two forms of the protein in 0% TFE at pH 6. Near UV CD spectra of myoglobin in 10% TFE at pH 6.5, 5.0, 4.0 and 3.2 appear in Fig. 4. At pH 6.5, partial loss of the 274nm feature (which is due to both the heme unit and aryl side-chain chromophores) is more advanced than loss of the 415nm heme feature, which occurs incompletely at this pH. No additional decrease in the heme signal occurs after about 15 minutes. In 10% TFE, loss of structured aromatic chains also appears to be incomplete even at pH 4; some residual positive ellipticity is retained at circa 260nm. Based on this comparison, HFIP is a much more effective disruptor of the tertiary structure of myoglobin. We expect that this will be a general feature. Thermal unfolding experiments at 12 and 40%
FIG. 3. CD spectra (wavelength, nm vs. residue molar ellipticity, [u]) of myoglobin at different temperatures in pH 4 buffer with (left to right) 4, 8, and 25% HFIP present, showing the blue shifted thermal isodichroic at 8% HFIP. Temperatures (K) at which the spectra in each thermal study were recorded are: for 4% HFIP— 275, 297, 327, 355; for 8% HFIP— 275, 297, 333, 355; for 25% HFIP— 275, 297, 341. 689
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
Vol. 233, No. 3, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Near UV CD spectra of myoglobin solutions containing 10% TFE at different pH values, and at pH 4 in the absence of TFE.
TFE (Fig. 2, panel B) yield curves similar to those seen for 4% and 25% HFIP, respectively, though the slopes in the high temperature, linear portion of the curves are slightly steeper for 40% TFE ([u]222 Å 028760 / 191rT). This probably reflects the decreased helix-stabilizing effect of TFE versus HFIP. CONCLUSIONS The addition of HFIP to as little as 4 vol-% disrupts the tertiary structure of aqueous myoglobin with complete loss of the aromatic side chain CD signals in the near UV as well as the bands in the visible and near UV due to the heme unit. Comparisons with TFE indicate that comparable behavior requires more than 10 vol-% of TFE but the disruption or fluxionalization of tertiary structure probably occurs in much the same way.3 At higher levels of added fluoroalcohol, myoglobin behaves as a linked set of independent helices. The myoglobin state that results upon addition of limited amounts of HFIP to the native form displays properties suggesting molten globule character, including some cooperativity in melting. The degree of helicity (essentially equal to that of native apomyoglobin) and dimished thermal fraying of this state would not be expected in the absence of some stabilization by tertiary interactions. Analysis of inherent helix propensities along the protein sequence - using i) a modified Lifson-Roig helix/coil transition model (33), ii) molecular dynamics trajectories (34) and iii) NMR spectros-
copy of peptidic fragments of the sequence (35) - all indicate that only two or three of the helices would be well populated in this medium if each were behaving independently. Furthermore, solvent exposed, partially formed helices would be expected to display unfolding upon cooling in this medium (15). The complete absence of cold denaturation at all HFIP concentrations examined is remarkable. The trichloroacetate-induced molten globule state of apomyoglobin displays dramatic cold denaturation at pH 2 (17) and isolated amphipathic helices display this behavior in 4 - 10 vol-% HFIP. We suggest that this indicates a state that is still compact in which the lipophilic sides of the helices have little bulk solvent exposure. Quite possibly the interior of the globule is accessible to HFIP but not to water. While cold-induced unfolding was not observed, the thermal isodichroic observed in 8 vol-% HFIP does indicate that this solvent composition has unique properties for protein solvation as well as peptides. Studies of the effects of HFIP upon other proteins and upon helix bundles are in progress in order to provide further definition of the physicochemical basis for both the helix-inducing and tertiary structure disrupting properties of fluoroalcohols. REFERENCES 1. Nelson, J. W., and Kallenbach, N. R. (1986) Proteins 1, 211–217. 2. Cammers-Goodwin, A., Allen, T. J., Oslick, S. L., McClure, K. F., Lee, J. H., and Kemp, D. S. (1996) J. Am . Chem. Soc. 118, 3082– 3092. 3. Slupsky, C. M., Kay, C. M., Reinach, F. C., Smillie, L. B., and Sykes, B. D. (1995) Biochemistry 34, 7365–7375. 4. Shiraki, K., Nishikawa, K., and Goto, Y. (1995) J. Mol. Biol. 245, 180–194. 5. Buck, M., Radford, S. E., and Dobson, C. M. (1993) Biochemistry 32, 669–678. 6. Bychkova, V. E., Dujsekina, A. E., Klenin, S. I., Tiktopulo, E. I., Uversky, V. N., and Ptitsyn, O. B. (1996) Biochemistry 35, 6058– 6063. 7. Mayer, R., Lancelot, G., and H’elene, C. (1983) FEBS Lett. 153, 339–344. 8. Andersen, N. H., Harris, S. M., Lee, V. G., Liu, E. C.-K., Moreland, S., and Hunt, J. T. (1995) Bioorg. and Med. Chem. 3, 113–124. 9. Parrish, J. R., and Blout, E. R. (1971) Biopolymers 10, 1491– 1512. 10. Parrish, J. R., and Blout, E. R. (1972) Bioploymers 11, 1001– 1020.
3
A comparative study of bovine b-lactoglobulin in aqueous HFIP and TFE appeared (32) as this paper was being submitted. HFIP effects a native (b-sheet) to helical state transition with DGU Å 0 at circa 6 vol-% (0.60 M, approximately one third the concentration required for a comparable effect by TFE). The authors note that the enhanced activity of HFIP cannot be explained by the additive effects of the constituent parts (CF3 surface area, etc.) and suggest that the ‘‘cooperative formation of micelle-like clusters of HFIP is important.’’ Evidence for other unusual media effects at this HFIP / water composition are given in footnote 2.
11. Chen, F., Lepore, G., and Goodman, M. (1974) Macromolecules 7, 779–783. 12. Beaudette, N. V., Okabayashi, H., and Fasman, G. D. (1982) Biochemistry 21, 1765–1772. 13. Galat, A., Degelaen, J. P., Yang, C. C., and Blout, E. R. (1981) Biochemistry 20, 7415–7423. 14. Cort, J., Liu, Z., Lee, G., Harris, S. M., Prickett, K. S., Gaeta, L. S. L., and Andersen, N. H. (1994) Biochem Biophys. Res. Commun. 204, 1088–1095.
690
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC
Vol. 233, No. 3, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
15. Andersen, N. H., Cort, J. R., Liu, Z., Sjoberg, S. J., and Tong, H. (1996) J. Am. Chem. Soc. 118, 10309–10310. 16. Privalov, P. L., Griko, Yu. V., and Venyaminov, S. Yu. (1986) J. Mol. Biol. 190, 487–498. 17. Nishii, I., Kataoka, M., Tokunaga, F., and Goto, Y. (1994) Biochemistry 33, 4903–4909. 18. Eliezer, D., Jennings, P. A., Wright, P. E., Doniach, S., Hodgson, K. O., Tsuruta, H. (1995) Science 270, 487–488. 19. Andersen, N. H., and Palmer, R. B. (1994) BioOrg. Med. Chem. Lett. 4, 817–822. 20. Ptitsyn, O. B. (1995) Adv. Protein Chem. 47, 83–229. 21. Merutka, G., and Stellwagen, E. (1990) Biochemistry 29, 894– 898. 22. Scholtz, J. M., Qian, H., York, E. J., Stewart, J. M., and Baldwin, R. L. (1991) Biopolymers 31, 1463–1470. 23. Munoz, V., and Serrano, L. (1995) J. Mol. Biol. 245, 297–308. 24. Wei, J., and Fasman, G. D. (1995) Biochemistry 34, 6408–6415. 25. Mchaourab, H. S., Hyde, J. S., and Felix, J. B. (1993) Biochemistry 32, 11895–11902.
26. Denda, M., Touhara, H., and Nakanishi, K. (1987) J. Chem. Thermodynamics 19, 539–542. 27. Kuprin, S., Graslund, A., Ehrenberg, A., and Koch, M. H. J. (1995) Biochem. Biophys. Res. Commun. 217, 1151–1156. 28. Murto, J., Kivinen, A., Kivimaa, S., and Laakso, R. (1967) Suomen Kemistilehti B 40, 250–257. 29. Murto, J., and Heino, E.-L. (1966) Suomen Kemistilehti B 39, 263–266. 30. Dyatkin, B. L., Mochalina, E. P., and Knunyants, I. L. (1965) Tetrahedron 21, 2991–2995. 31. Ballinger, P., and Long, F. A. (1959) J. Am . Chem. Soc. 81, 1050–1053. 32. Hirota, N., Mizuno, K., and Goto, Y. (1997) Protein Sci. 6, 416– 421. 33. Andersen, N. H., and Tong, H. (1997) Protein Sci., in press. 34. Hirst, J. D., and Brooks, C. L. III (1995) Biochemistry 34, 7614– 7621. 35. Waltho, J. P., Feher, V. A, Merutka, G., Dyson, H. J., and Wright, P. E. (1993) Biochemistry 32, 6337–6347.
691
AID
BBRC 6524
/
6929$$$501
04-10-97 13:15:25
bbrcg
AP: BBRC