HX -- Hydrogen Exchange

Folding and Design Lab

Members

John Osterhout

Research

alpha-t-alpha
Peptide Dissection
HX

Group Alumni

Youcef Fezoui
Tanya Knubovets
Marcela Oslin
Diane Schaak
Wujing Xian

Protein Folding Protein Structure Discussion Group

Talk Schedule

Hydrogen Exchange

Protein Folding and Design Laboratory

H(C) <=> H(O) -> D(O) <=> D(C)

Hydrogen Exchange

Hydrogen exchange is an excellent way to look at the local stability of proteins. The rates of amide proton exchange for individual protons can be related to equilibrium constants for opening of individual hydrogen bonds. Knowing the equilibrium constants, one can calculate the free energy for the conformational transition which allows exchange to occur. See the discussion in Fezoui et al., 1999 Biochemistry 38, 2796-2804 (below) and (importantly) references therein. When certain protons are only exposed in the completely unfolded form then the equilibrium constants and deltaGs correspond to the global unfolding reaction. These protons are usually the slowest exchanging protons in the molecule.

A powerful variation of amide proton exchange is the quenched exchange experiment. In this experiment amide proton exchange is allowed to occur under one set of conditions, the exchange is then stopped (quenched) by a change in solvent conditions (primarily pH), and the degree of exchange is monitored by two-dimensional NMR at one's leisure. For instance, protein folding intermediates which form on the millisecond time scale can be studied in this way by stopped flow/quench.

We have been involved in measuring stability or conformational changes in a variety of types of protein structures or environments.

These include:

alpha-t-alpha, a peptide model for a folding intermediate
Kinetic folding intermediates of Villin 14T
Bovine Pancreatic Trypsin Inhibitor (BPTI) in the solid
Lysozyme in 99% glycerol

The Villin 14T, BPTI and Lysozyme projects were performed as collaborations as noted below.

ata -- HX From a Folding Intermediate Model System

Dissection of the de Novo Designed Peptide alpha-t-alpha: Stability and Properties of the Intact Molecule and Its Constituent Helices.

Fezoui, Y., Braswell, E. H., Xian, W. and Osterhout, J. J. (1999) Biochemistry 38, 2796-2804

Abstract

ata is a de novo designed 38-residue peptide (Fezoui et al, Protein Science 4: 286-295, 1995) which adopts a helical hairpin conformation in solution (Fezoui et al, (1994) Proc. Natl. Acad. Sci. USA 91, 3675-3679 and (1997) Protein Science 6: 1869-1877). Since ata was developed as a model system for protein folding at the stage where secondary structures interact and become mutually stabilizing, it is of interest to investigate the increase in stability that occurs with helix association. ata was dissected into its component helices and the relative stabilities of the individual helices and the parent molecule were assessed. The deltaG of unfolding of ata measured by guanidinium hydrochloride denaturation was determined to be 3.4 kcal/mole. The equilibrium constant for folding of ata was estimated from the deltaG as 338 and from hydrogen exchange measurements as 259. The stability of the helices in intact ata over the individual helices increased by a factor of at least 37 based on amide proton exchange measurements. Sedimentation equilibrium studies showed very little association of the peptides to form either homo or heterodimers suggesting that helix association is stabilized by the high effective concentration of the helices caused by the presence of the connecting turn. The effects of salt and pH on the helicity of the component peptides are largely reflected in the intact molecule implying that short range interactions still make important contributions to the conformation of the intact molecule even though significant stabilization is caused by helix association.

Commentary

Hydrogen exchange turned out to be the best way to compare the stability of the individual helices and the intact molecule. (Hydrogen exchange works better than CD because ata is too stable for CD to provide a good estimate of the equilibrium constant.) (See also the Peptide Dissection Page.) The basic conclusion from this study is that the peptide representing the individual helices are not very stable, Keq for folding = 6-7, while the Keq for the folding of the intact molecule is about 260 by hydrogen exchange.

Recently, we produced partially labeled ata by solid phase peptide synthesis using 15N labeled Ala, Gly and Leu (manuscript in preparation). This project has given us a good idea of nature of local exchange in the peptide. Watch this space!

Villin 14T -- HX From a Folding Intermediate

Folding Kinetics of Villin 14T, a Protein Domain with a Central beta-Sheet and Two Hydrophobic Cores.

Choe, S. E., Matsudaira, P. T., Osterhout, J., Wagner, G., and Shakhnovich, E. I., (1998) Biochemistry 37, 14508-14518

Abstract

The thermodynamics and kinetics of folding are characterized for villin 14T, a 126-residue protein domain. Equilibrium fluorescence measurements reveal that villin 14T unfolds and refolds reversibly. The folding kinetics was monitored using stopped-flow with fluorescence and quenched-flow with NMR and mass spectrometry. Unfolding occurs in a single-exponential phase in the stopped-flow experiments, and about 75% of the total amplitude is recovered in the fast phase of refolding. the remaining 25% of the amplitude probably represents trapping in cis-trans proline isomerization pathways. At 25 degrees C, the stability estimate obtained by extrapolation from the transition region of the stopped-flow chevron matches the stability value from equilibrium urea titrations (delta G = 9.7 kcal/mol, m value = 2.2 kcal /mol M). At low final urea concentrations, however, the refolding kinetics deviates from the two-state model, indicating the formation of an intermediate. Under these conditions, quenched-flow followed by NMR and mass spectrometry show no detectable hydrogen-bonded intermediate in the fast refolding phase. In contrast, agreement is observed between the equilibrium and kinetic estimates of stability at 37 degrees C (deltaG = 6.0 kcal/mol, m value = 1.6 kcal/mol M), at all observed urea concentration, demonstrating apparent two-state folding at this temperature. This result shows that a two-state folding model, previously applied to small domains with single, central hydrophobic cores, can also describe the folding of a larger domain with multiple core structures.

Commentary

For this commentary I am going to focus on the contribution of the hydrogen exchange experiments although they played a relatively small part in this paper. The chevron plot of the folding and unfolding phases of villin 14T in urea at 25 degrees C shows a deviation in linearity at the low concentrations of urea. This behavior is commonly interpreted to mean that there are intermediates populated at the low concentrations of urea. Quenched flow experiments were performed to attempt to detect stably hydrogen bonded structure in these intermediates. Thirty one amides exchange slowly enough from the native state to be studied by the quenched flow technique. Of these 31, 30 show a refolding time constant of approximately 60 ms which corresponds to the 60 ms time constant determined from stopped-flow fluorescence experiments. (The odd proton, that of valine 31, was thought to be aberrant because of local structure formation). Protection of the amide protons from exchange under these conditions correlates with overall structure formation and there is no evidence for a stably hydrogen bonded intermediate. Quenched flow experiments were also monitored by mass spectrometry with similar results: the major kinetic phase has a time constant similar to the fluorescence detected phase and there is no evidence for intermediate formation.

One can put a limit on the protection factor for exchange that can be detected in a quenched flow experiment. In this case the minimum protection factor detectable is 33 which is lower than the maximum protection factor in ata, our peptide model system (described above). The conclusion from the quenched flow experiments is that the equilibrium constant for hydrogen bonding in the intermediate must be less that about 33 which is quite low

From comparison of the m value for the fluorescence phase and the m value calculated for the intermediate the approximate solvent accessible surface of the intermediate can be calculated. This calculation suggests that the intermediate buries about 65% of the total surface area buried during refolding. A picture of the intermediate emerges: the amount of hydrophobic surface area buried and the change in tryptophan fluorescence suggest a hydrophobic collapse. The low protection factor for hydrogen exchange from the intermediate says that, while secondary structure may be formed, it is not very stable.

BPTI-Hydrogen Exchange from the Solid

Protein Structure in the Lyophilized State: A Hydrogen Isotope Exchange/NMR Study with Bovine Pancreatic Trypsin Inhibitor.

Desai, U. R., Osterhout, J. J. and Klibanov, A. M. (1994). J. Am. Chem. Soc. 116, 9420-9422

Abstract

The structure of a stable model protein, bovine pancreatic trypsin inhibitor (BPTI), in the lyophilized form has been investigated using the hydrogen isotope exchange/high-resolution NMR methodology. Six amide protons of BPTI that are buried in the protein interior and strongly hydrogen-bonded in aqueous solution are found to exchange with water vapors within hours in the lyophilized state: in aqueous solution, most of these protons do not exchange appreciably even after a week under otherwise identical conditions. When BPTI is lyophilized in the presence of the lyoprotectant sorbitiol, no significant hydrogen isotope exchange of these protons in the solid state is detected. On the basis of these and other observations it is concluded that the structure of BPTI is partially (and reversibly) denatured on lyophilization. This conclusion, if true for other proteins, may explain the drastically reduced enzymatic activity in nonaqueous media compared to that in water.

Commentary

Six protons which are stable to exchange in aqueous BPTI exchanged readily from lyophilized BPTI. The straightforward conclusion from this observation is that BPTI is unfolded in lyophilized powder. When BPTI is colyophilized with sorbitol no exchange is observed in the same period from the lyophilized powder. This observation suggests that sorbitol allows BPTI to remain folded and hence protect its amides from exchange. A second possibility is that the orbital interferes with the ability of the water vapor to penetrate the lyophilized powder and so accounts for the lack of exchange. The "other observations" mentioned in the abstract concern primarily an interesting set of controls. The controls involved using N-acetylglycine-N'-methylamide, AGMA, to measure the exchange from essentially unprotected amides. Exchange experiments with AGMA in the presence and absence of sorbitol showed that sorbitol actually increased the exchange slightly . The lack of exchange from BPTI in the presence of sorbitol therefor cannot be attributed to interference from the lyoprotectant.

Lysozyme-Hydrogen Exchange in 99% Glycerol

Structure, thermostability, and conformational flexibility of hen egg-white lysozyme dissolved in glycerol.

Knubovets, T., Osterhout, J. J., Connolly, P. J. and Klibanov, A. M. (1999). Proc. Natl. Acad. Sci. USA 96, 1262-1267

Abstract

Hen egg-white lysozyme dissolved in glycerol containing 1% water was studied by using CD and amide proton exchange monitored by two-dimensional 1H NMR. The far- and near-UV CD spectra of the protein showed that the secondary and tertiary structures of lysozyme in glycerol were similar to those in water. Thermal melting of lysozyme in glycerol followed by CD spectral changes indicated unfolding of the tertiary structure with a Tm of 76.0 +/- 0.2 degrees C and no appreciable loss of secondary structure up to 85 degrees C. this is in contrast to the coincident denaturation of both tertiary and secondary structures with Tm values of 74.8 +/- 0.4 degrees C and 74.3 +/- 0.7 degrees C respectively, under analogous conditions in water. Quenched amide proton exchange experiments revealed a greater structural protection of amide protons in glycerol than in water for a majority of the slowly exchanging protons. The results point ot a highly ordered, native-like structure of lysozyme in glycerol, with the stability exceeding that in water.

Commentary

Here again, amide proton exchange was measured by a variation of the quenched technique. In this case lysozyme was allowed to exchange in 99% glycerol/1% water before being diluted into low pH aqueous solution (to halt the exchange) and worked up into an NMR sample. For controls the exchange from poly-D,L-alanine was measured spectraphotometrically in both water and glycerol under the same conditions as the exchange experiments from lysozyme. The ratio of exchange of PDLA and lysozyme in water approximates the protection factor (uncorrected for sequence effects). The ratios of the protection factors in the two situations is the stability factor or the increase in stability due to the change in solvent conditions (an assumption is that the effects of sequence in the unfolded state are approximately the same and cancel from the calculation). The stability factors of the various amides are color coded onto the lysozyme backbone below. Most of the amide protons we measured had stability factors greater than one (the stability in glycerol was equal to or greater than that in water). The exceptions (red spheres in the figure below) were mostly near the ends of secondary structures.

An interesting aspect of this work concerns the melting curves of lysozyme in glycerol and water measured by CD. The near- and far-UV CD curves of lysozyme in water and glycerol are similar suggesting that approximately equal amounts of secondary and tertiary structure. In water the unfolding curves of the near- and far-UV CD signals are coincident as they are in many globular proteins. However, in glycerol the near-UV CD signal melts at approximately the same temperature as lysozyme in water while the far-UV CD signal hardly changes up to 85 degrees C. This indicates that in glycerol at high temperature lysozyme adopts a conformation which is rich in secondary structure and lacking in detailed tertiary structure (the conformational restrictions on the aromatic side chains are lost.). This is a situation reminiscent of the molten globule.

The melting experiments suggest how lysozyme can melt in glycerol at the same temperature (by near-UV CD) as in water and yet be up to 40-fold more stable in places as implied by stability factors. The suggestion is that there is a switch in unfolding mechanism between the two situations, that in glycerol, when lysozyme unfolds first to an intermediate which contains significant secondary structure rather to an unfolded form. The intermediate would have to have a degree of amide protection at least equal to the change in stability factor which at this point is maximally 40-fold. Since folding intermediates, molten globules, and our own peptide model system have protection factors in the range of 200-1000 it does not seem unreasonable that a molten globule-like intermediate could provide sufficient stability to account for the changes in stability factors which we observe.

Stereo view (cross-eyed) of a ribbon diagram of lysozyme with stability factors of the amides colored by class: Sf < 1 (red spheres), 1<Sf<10 ( green spheres), Sf > 10 (blue spheres) and Sf which could not be reliably estimated from our data (gray spheres). Note: this figure is similar to Fig. 4 from the paper but includes an additional class of stability factors (Sf >10).

Folding and Design Lab Members John Osterhout Research alpha-t-alpha Peptide Dissection HX Group Alumni Youcef Fezoui Tanya Knubovets Marcela Oslin Diane Schaak Wujing Xian Protein Folding Protein Structure Discussion Group Talk Schedule	Hydrogen Exchange Protein Folding and Design Laboratory H(C) <=> H(O) -> D(O) <=> D(C)
	Hydrogen Exchange Hydrogen exchange is an excellent way to look at the local stability of proteins. The rates of amide proton exchange for individual protons can be related to equilibrium constants for opening of individual hydrogen bonds. Knowing the equilibrium constants, one can calculate the free energy for the conformational transition which allows exchange to occur. See the discussion in Fezoui et al., 1999 Biochemistry 38, 2796-2804 (below) and (importantly) references therein. When certain protons are only exposed in the completely unfolded form then the equilibrium constants and deltaGs correspond to the global unfolding reaction. These protons are usually the slowest exchanging protons in the molecule. A powerful variation of amide proton exchange is the quenched exchange experiment. In this experiment amide proton exchange is allowed to occur under one set of conditions, the exchange is then stopped (quenched) by a change in solvent conditions (primarily pH), and the degree of exchange is monitored by two-dimensional NMR at one's leisure. For instance, protein folding intermediates which form on the millisecond time scale can be studied in this way by stopped flow/quench. We have been involved in measuring stability or conformational changes in a variety of types of protein structures or environments. These include: alpha-t-alpha, a peptide model for a folding intermediate Kinetic folding intermediates of Villin 14T Bovine Pancreatic Trypsin Inhibitor (BPTI) in the solid Lysozyme in 99% glycerol The Villin 14T, BPTI and Lysozyme projects were performed as collaborations as noted below.
	ata -- HX From a Folding Intermediate Model System Dissection of the de Novo Designed Peptide alpha-t-alpha: Stability and Properties of the Intact Molecule and Its Constituent Helices. Fezoui, Y., Braswell, E. H., Xian, W. and Osterhout, J. J. (1999) Biochemistry 38, 2796-2804 Abstract ata is a de novo designed 38-residue peptide (Fezoui et al, Protein Science 4: 286-295, 1995) which adopts a helical hairpin conformation in solution (Fezoui et al, (1994) Proc. Natl. Acad. Sci. USA 91, 3675-3679 and (1997) Protein Science 6: 1869-1877). Since ata was developed as a model system for protein folding at the stage where secondary structures interact and become mutually stabilizing, it is of interest to investigate the increase in stability that occurs with helix association. ata was dissected into its component helices and the relative stabilities of the individual helices and the parent molecule were assessed. The deltaG of unfolding of ata measured by guanidinium hydrochloride denaturation was determined to be 3.4 kcal/mole. The equilibrium constant for folding of ata was estimated from the deltaG as 338 and from hydrogen exchange measurements as 259. The stability of the helices in intact ata over the individual helices increased by a factor of at least 37 based on amide proton exchange measurements. Sedimentation equilibrium studies showed very little association of the peptides to form either homo or heterodimers suggesting that helix association is stabilized by the high effective concentration of the helices caused by the presence of the connecting turn. The effects of salt and pH on the helicity of the component peptides are largely reflected in the intact molecule implying that short range interactions still make important contributions to the conformation of the intact molecule even though significant stabilization is caused by helix association. Commentary Hydrogen exchange turned out to be the best way to compare the stability of the individual helices and the intact molecule. (Hydrogen exchange works better than CD because ata is too stable for CD to provide a good estimate of the equilibrium constant.) (See also the Peptide Dissection Page.) The basic conclusion from this study is that the peptide representing the individual helices are not very stable, Keq for folding = 6-7, while the Keq for the folding of the intact molecule is about 260 by hydrogen exchange. Recently, we produced partially labeled ata by solid phase peptide synthesis using 15N labeled Ala, Gly and Leu (manuscript in preparation). This project has given us a good idea of nature of local exchange in the peptide. Watch this space!
	Villin 14T -- HX From a Folding Intermediate Folding Kinetics of Villin 14T, a Protein Domain with a Central beta-Sheet and Two Hydrophobic Cores. Choe, S. E., Matsudaira, P. T., Osterhout, J., Wagner, G., and Shakhnovich, E. I., (1998) Biochemistry 37, 14508-14518 Abstract The thermodynamics and kinetics of folding are characterized for villin 14T, a 126-residue protein domain. Equilibrium fluorescence measurements reveal that villin 14T unfolds and refolds reversibly. The folding kinetics was monitored using stopped-flow with fluorescence and quenched-flow with NMR and mass spectrometry. Unfolding occurs in a single-exponential phase in the stopped-flow experiments, and about 75% of the total amplitude is recovered in the fast phase of refolding. the remaining 25% of the amplitude probably represents trapping in cis-trans proline isomerization pathways. At 25 degrees C, the stability estimate obtained by extrapolation from the transition region of the stopped-flow chevron matches the stability value from equilibrium urea titrations (delta G = 9.7 kcal/mol, m value = 2.2 kcal /mol M). At low final urea concentrations, however, the refolding kinetics deviates from the two-state model, indicating the formation of an intermediate. Under these conditions, quenched-flow followed by NMR and mass spectrometry show no detectable hydrogen-bonded intermediate in the fast refolding phase. In contrast, agreement is observed between the equilibrium and kinetic estimates of stability at 37 degrees C (deltaG = 6.0 kcal/mol, m value = 1.6 kcal/mol M), at all observed urea concentration, demonstrating apparent two-state folding at this temperature. This result shows that a two-state folding model, previously applied to small domains with single, central hydrophobic cores, can also describe the folding of a larger domain with multiple core structures. Commentary For this commentary I am going to focus on the contribution of the hydrogen exchange experiments although they played a relatively small part in this paper. The chevron plot of the folding and unfolding phases of villin 14T in urea at 25 degrees C shows a deviation in linearity at the low concentrations of urea. This behavior is commonly interpreted to mean that there are intermediates populated at the low concentrations of urea. Quenched flow experiments were performed to attempt to detect stably hydrogen bonded structure in these intermediates. Thirty one amides exchange slowly enough from the native state to be studied by the quenched flow technique. Of these 31, 30 show a refolding time constant of approximately 60 ms which corresponds to the 60 ms time constant determined from stopped-flow fluorescence experiments. (The odd proton, that of valine 31, was thought to be aberrant because of local structure formation). Protection of the amide protons from exchange under these conditions correlates with overall structure formation and there is no evidence for a stably hydrogen bonded intermediate. Quenched flow experiments were also monitored by mass spectrometry with similar results: the major kinetic phase has a time constant similar to the fluorescence detected phase and there is no evidence for intermediate formation. One can put a limit on the protection factor for exchange that can be detected in a quenched flow experiment. In this case the minimum protection factor detectable is 33 which is lower than the maximum protection factor in ata, our peptide model system (described above). The conclusion from the quenched flow experiments is that the equilibrium constant for hydrogen bonding in the intermediate must be less that about 33 which is quite low From comparison of the m value for the fluorescence phase and the m value calculated for the intermediate the approximate solvent accessible surface of the intermediate can be calculated. This calculation suggests that the intermediate buries about 65% of the total surface area buried during refolding. A picture of the intermediate emerges: the amount of hydrophobic surface area buried and the change in tryptophan fluorescence suggest a hydrophobic collapse. The low protection factor for hydrogen exchange from the intermediate says that, while secondary structure may be formed, it is not very stable.
	BPTI-Hydrogen Exchange from the Solid Protein Structure in the Lyophilized State: A Hydrogen Isotope Exchange/NMR Study with Bovine Pancreatic Trypsin Inhibitor. Desai, U. R., Osterhout, J. J. and Klibanov, A. M. (1994). J. Am. Chem. Soc. 116, 9420-9422 Abstract The structure of a stable model protein, bovine pancreatic trypsin inhibitor (BPTI), in the lyophilized form has been investigated using the hydrogen isotope exchange/high-resolution NMR methodology. Six amide protons of BPTI that are buried in the protein interior and strongly hydrogen-bonded in aqueous solution are found to exchange with water vapors within hours in the lyophilized state: in aqueous solution, most of these protons do not exchange appreciably even after a week under otherwise identical conditions. When BPTI is lyophilized in the presence of the lyoprotectant sorbitiol, no significant hydrogen isotope exchange of these protons in the solid state is detected. On the basis of these and other observations it is concluded that the structure of BPTI is partially (and reversibly) denatured on lyophilization. This conclusion, if true for other proteins, may explain the drastically reduced enzymatic activity in nonaqueous media compared to that in water. Commentary Six protons which are stable to exchange in aqueous BPTI exchanged readily from lyophilized BPTI. The straightforward conclusion from this observation is that BPTI is unfolded in lyophilized powder. When BPTI is colyophilized with sorbitol no exchange is observed in the same period from the lyophilized powder. This observation suggests that sorbitol allows BPTI to remain folded and hence protect its amides from exchange. A second possibility is that the orbital interferes with the ability of the water vapor to penetrate the lyophilized powder and so accounts for the lack of exchange. The "other observations" mentioned in the abstract concern primarily an interesting set of controls. The controls involved using N-acetylglycine-N'-methylamide, AGMA, to measure the exchange from essentially unprotected amides. Exchange experiments with AGMA in the presence and absence of sorbitol showed that sorbitol actually increased the exchange slightly . The lack of exchange from BPTI in the presence of sorbitol therefor cannot be attributed to interference from the lyoprotectant.
	Lysozyme-Hydrogen Exchange in 99% Glycerol Structure, thermostability, and conformational flexibility of hen egg-white lysozyme dissolved in glycerol. Knubovets, T., Osterhout, J. J., Connolly, P. J. and Klibanov, A. M. (1999). Proc. Natl. Acad. Sci. USA 96, 1262-1267 Abstract Hen egg-white lysozyme dissolved in glycerol containing 1% water was studied by using CD and amide proton exchange monitored by two-dimensional 1H NMR. The far- and near-UV CD spectra of the protein showed that the secondary and tertiary structures of lysozyme in glycerol were similar to those in water. Thermal melting of lysozyme in glycerol followed by CD spectral changes indicated unfolding of the tertiary structure with a Tm of 76.0 +/- 0.2 degrees C and no appreciable loss of secondary structure up to 85 degrees C. this is in contrast to the coincident denaturation of both tertiary and secondary structures with Tm values of 74.8 +/- 0.4 degrees C and 74.3 +/- 0.7 degrees C respectively, under analogous conditions in water. Quenched amide proton exchange experiments revealed a greater structural protection of amide protons in glycerol than in water for a majority of the slowly exchanging protons. The results point ot a highly ordered, native-like structure of lysozyme in glycerol, with the stability exceeding that in water. Commentary Here again, amide proton exchange was measured by a variation of the quenched technique. In this case lysozyme was allowed to exchange in 99% glycerol/1% water before being diluted into low pH aqueous solution (to halt the exchange) and worked up into an NMR sample. For controls the exchange from poly-D,L-alanine was measured spectraphotometrically in both water and glycerol under the same conditions as the exchange experiments from lysozyme. The ratio of exchange of PDLA and lysozyme in water approximates the protection factor (uncorrected for sequence effects). The ratios of the protection factors in the two situations is the stability factor or the increase in stability due to the change in solvent conditions (an assumption is that the effects of sequence in the unfolded state are approximately the same and cancel from the calculation). The stability factors of the various amides are color coded onto the lysozyme backbone below. Most of the amide protons we measured had stability factors greater than one (the stability in glycerol was equal to or greater than that in water). The exceptions (red spheres in the figure below) were mostly near the ends of secondary structures. An interesting aspect of this work concerns the melting curves of lysozyme in glycerol and water measured by CD. The near- and far-UV CD curves of lysozyme in water and glycerol are similar suggesting that approximately equal amounts of secondary and tertiary structure. In water the unfolding curves of the near- and far-UV CD signals are coincident as they are in many globular proteins. However, in glycerol the near-UV CD signal melts at approximately the same temperature as lysozyme in water while the far-UV CD signal hardly changes up to 85 degrees C. This indicates that in glycerol at high temperature lysozyme adopts a conformation which is rich in secondary structure and lacking in detailed tertiary structure (the conformational restrictions on the aromatic side chains are lost.). This is a situation reminiscent of the molten globule. The melting experiments suggest how lysozyme can melt in glycerol at the same temperature (by near-UV CD) as in water and yet be up to 40-fold more stable in places as implied by stability factors. The suggestion is that there is a switch in unfolding mechanism between the two situations, that in glycerol, when lysozyme unfolds first to an intermediate which contains significant secondary structure rather to an unfolded form. The intermediate would have to have a degree of amide protection at least equal to the change in stability factor which at this point is maximally 40-fold. Since folding intermediates, molten globules, and our own peptide model system have protection factors in the range of 200-1000 it does not seem unreasonable that a molten globule-like intermediate could provide sufficient stability to account for the changes in stability factors which we observe.

	Stereo view (cross-eyed) of a ribbon diagram of lysozyme with stability factors of the amides colored by class: Sf < 1 (red spheres), 1<Sf<10 ( green spheres), Sf > 10 (blue spheres) and Sf which could not be reliably estimated from our data (gray spheres). Note: this figure is similar to Fig. 4 from the paper but includes an additional class of stability factors (Sf >10).