327
Biochimica et Biophysica Acta, 5 8 0 ( 1 9 7 9 ) 3 2 7 - - 3 3 8 © E l s e v i e r / N o r t h - H o l l a n d Biomedical Press
BBA 38286
STABILITY OF PHAGE T4 LYSOZOMES II. UNFOLDING WITH GUANIDINIUM CHLORIDE
M.L. E L W E L L * a n d J . A . S C H E L L M A N
Institute of Molecular Biology, University of Oregon, Eugene, OR 97403 (U.S.A.) (Received March 3rd, 1979)
Key words: Lysozyme; Guanidine denaturation; Mutant protein; Stability; Temperature sensitivity; (Phage T4)
Summary The denaturation by guanidinium chloride of three phage lysozymes (wild type and two mutants) was investigated. The study of solvent denaturation permitted the investigation of the relative stabilities of the proteins at neutral pH, in contrast to thermal denaturation studies reported earlier which could only be performed in acid pH. The results were interpreted assuming t h a t the free energy of solution of proteins is a linear function of denaturant concentration. Using standard t h e r m o d y n a m i c formulas this permits the calculation of the stabilities of the three proteins in the absence of guanidinium chloride. The single point m u t a t i o n Trp 138 -* Tyr leads to relatively large changes in stability and the interaction of the protein with guanidinium chloride. The changes associated with the subsequent double mutation, Trp 126 -~ Tyr, Trp 158 -* Tyr, are much smaller indicating a relatively smooth adjustment of the protein structure to the changed side chains. Models of the structural effects of point mutations are discussed. It is found t h a t the m u t a t i o n at position 138 does n o t fit a model in which the effect of a substitution is to introduce an energetic strain in the structure. It does fit a model in which there is a partial unravelling of the structure as a result of the mutation. However, there are no changes in the backbone circular dichroism spectra associated with the mutation. The two observations are not necessarily in conflict. Further physical studies are required for the resolution of the problem. * Present address: Department of Chemistry, University of Texas at Dallas, Richardson, TX 75080, U.S.A. Abbreviations: Phage T4 lysozyme, T4 lysozyme; Phage eR1-75 T4 lysozyme, eR1-75 T4 lysozyme; Phage eRCIOeRleRD1 T4 lysozyme, eRRR T4 lysozyme.
328
Introduction
This paper explores the effect of point mutations on the unfolding of phage T4 lysozyme in solutions of guanidinium chloride. The preparation and properties of the enzyme have been described in an earlier publication [1] and we have recently reported the effect of the same point mutations on its thermal stability and t h e r m o d y n a m i c properties [2]. Briefly, the wild-type enzyme has three t r y p t o p h a n residues at positions 126, 138 and 158. Tryptophans 126 and 158 are non-essential for enzyme activity. Tryptophan 138 is essential for activity but 50% of the activity is retained when tyrosine is substituted for t r y p t o p h a n at this position. The two mutants to be discussed in this paper are the amber revertant eR1-75, which has tryptophan 138 replaced by tyrosine and m u t a n t eRRR which is a triple revertant with all three tryptophans replaced tyrosines. The previous paper demonstrated that the substitution of tyrosine for t r y p t o p h a n at position 138 had a pronounced effect on the thermal stability of this enzyme, but that the additional effect of tyrosine substitutions at positions 128 and 158 was small by comparison. The present paper extends these comparative studies to solvent denaturation, specifically denaturation in guanidinium chloride. The use of guanidinium chloride makes it possible to work at neutral pH and higher protein concentration without the solubility problems encountered in the thermal denaturation, thereby supplying important information on the stability of the protein at physiological pH values. If the free energy of interaction of the protein and guanidinium chloride is known, it is possible to obtain the free energy of stabilization of the protein in the absence of solvent denaturant by extrapolation [3]. This is the key piece of information that is required for a discussion of stability. The extrapolation is made provisionally in this paper on the basis of a linear free energy law. Evidence for the applicability of this law will be adduced in Discussion. Materials and Methods The materials, methods and instrumentation employed were identical to those used previously [1,2]. The viral strains and methods of purification and isolation of T4 lysozyme have been previously described [1,2]. The fluorescence emission was monitored as a function of Gdn HC1 concentration using either the Hitachi-Perkin-Elmer MPF-2A spectrofluorimeter or the Schoeffel RRS 1000 spectrofluorimeter. Fluorescence measurements were made at protein concentrations of 0.07 mg/ml. Aromatic CD measurements were made at protein concentrations of 1 mg/ml. Far ultraviolet CD measurements were made at 0.02 mg/ml protein. The stock Grin HC1 solutions were 5--6 M Gdn HC1 in H20 and were diluted to the desired Gdn HC1 concentration. For neutral pH experiments, the stock Gdn HC1 was diluted with 0.2 M sodium phosphate buffer, pH 6, 10 -3 M mercaptoethanol. For acid pH experiments, the stock Gdn HC1 was diluted with water and the pH adjusted with HC1. Gdn HCI concentration was determined by measuring the index of refraction of diluted Gdn HC1 samples containing no T4 lysozyme [4]. Baselines for the calculation of equilibrium constants were extrapolated horizontally as in Fig. 3.
329
Solvent denaturation Our experimental results indicate that the unfolding of the wild-type enzyme as well as the eR1-75 and e R R R m u t a n t lysozymes may be interpreted with a two state model. Evidence for this is the fact that all solvent and thermal unfolding curves are monophasic and follow a course which is independent of the physical technique used to follow them. (See Results and Ref. 1.) For a two state model N~D the change in standard chemical potential which accompanies the unfolding of a protein may be written as A/2 = / ~ / 0 + A/Aint
(1)
where A~ ° is the change in standard chemical potential in the absence of guanidinium chloride and A~ is the change in standard chemical potential in its presence. A~in t represents the difference in free energy of interaction of the D and N states with the added guanidinium chloride. To parallel another discussion, [3] we write it as A~/in t = R T A b 2
where R T A b 2 is the change in excess free energy which acompanies the unfolding i.e. the free energy of interaction with guanidinium chloride, and subscript 2 indicates the protein c o m p o n e n t of the system. Introducing equilibrium constants via - - R T in K = A~ in K = In K ° - - Ab2
(2)
where K ° = [D]/[N] in the absence of guanidinium chloride and K = [D]/[N] in its presence. Using C3 to symbolize guanidinium chloride concentration we define Cm (the analog of Tin) as the concentration of denaturant at which [D] = [N]. At Cm, K = 1 so we have in K ° = Ab2
(C3 = Cm)
(3)
or
A~ ° = - - R T A b 2
(4)
where b2 is to be evaluated at Cm. A more detailed t h e r m o d y n a m i c analysis [3] shows also that {O l n g ~
= AF32
(5)
where a is the activity of the guanidinium chloride and AF is the change in the t h e r m o d y n a m i c binding parameter F which occurs when the molecule is unfolded. If the guanidinium binding is stoichiometric, AF = Ar where Ar is the increase in bound ligand upon unfolding based on multiple equilibrium theory. More generally, F represents the excess guanidinium chloride in the solvation domain of the macromolecule even if this excess is not describable in terms of denumerable complexes.
330
A special case of interest is when Ab: is a linear funct i on of d e n a t u r a n t concen tr atio n Ab2 = Ab°~C3
(6L)
where Ab°3 is the virial coefficient for the interation of c o m p o n e n t s 2 and 3. The label L on the formula indicates the restriction of a linear free energy model. In this case Eqns. 3 and 4 may be rewritten as In K ° = Ab°~Cm
(3L)
A~ 0 = --RTAb°3Cm
(4L)
Experimentally one evaluates K as a function of C3. Cm is the c o n c e n t r a t i o n at which K = 1; Ab°3 is found from the slope of In K vs. C3 via the relation o In K _ Ab:3 ~C3
(7L)
The binding is determined by plotting in K vs. in a (Eqn. 5). Activity data for guanidinium chloride are available from two investigations [5,6]. The values of Cm and Ab°3 obtained above permit the evaluation of the stabilization free energy of the protein in the absence of the d e n a t u r a n t via Eqn. 4L. If the linear free energy relation is n o t valid then Ab2 may be written as a p o w e r series m b 2 = m b ° 3 C 3 4- A c 203 3 C 3 2 -Jr- A d 2 03 3 3 C 33 + higher terms
(8)
where it is assumed that the protein solution is sufficiently dilute that proteinprotein interactions are u n i m p o r t a n t and that the guanidinium chloride concen tr atio n is sufficiently high that the Debye-Hiickle term in C3 is swamped o u t [7]. In this case --ABE must be evaluated by a careful analysis of K as a f u n c t i o n o f C3 not only near Cm but over as wide a range as possible using the relations In K
=
in K ° - -
Ab°3C
3 --
A c 2 03 3 C
23 - -
A d 2 303 3 C
33 - -
higher terms
(9)
or O In K . 3C3
Ab03 . .
2 A c .0 3 3 C 3
0 3Ad2333 C32 -- higher terms
(10)
Clearly the free energy of stabilization in the absence of d e n a t u r a n t can be calculated when all of the contributing viral coefficients are known. It is assumed that the ionic strength is sufficiently high from salt c o m p o n e n t s in the system that the electrolyte effect of guanidinium chloride plays a small role at low guanidinium concentration. The linear model is a special case of Eqns. 8--10 and should n o t be used w i t h o u t confirming evidence t ha t In K is linear in C3 over a wide range o f concentrations. This t y p e of data is missing from this r e p o r t partially because the m e t h o d of analysis was not envisaged at the time the experiments were p e r f o r m e d and partially because phage l y s o z y m e is in scarce supply and oft en precipitates at m o d e r a t e concentrations. Consequently most of our w ork was d o n e at low c o n c e n t r a t i o n where accurate data far from K = I are hard to obtain.
331 Nevertheless we shall provisionally make use of the linear free energy model in interpreting the data on the mutants at neutral pH. This is temporarily justified by the findings of Green and Pace [8] and Aune and Tanford [6] that a linear relation is followed over the full range of accessible concentration for several proteins. The former investigation included urea as well as guanidinium chloride as a denaturant. However, in a personal communication Dr. Pace has indicated that he does not believe that the interaction parameter Ab: is linear in guanidinium chloride for the denaturation of myoglobin. In addition Pfeil and Privalov [9] have found that the enthalpy of interaction of guanidinium chloride with hen egg white lysozyme is essentially linear for both the native and denatured forms. While this is only the enthalpic part of RTAb2 it adds credibility to linear t h e r m o d y n a m i c relationships. Another cross-check can be obtained by comparing the stabilization free energies obtained by extrapolating thermal data to ambient temperature (25°C or 37.5°C) with those obtained by the linear extrapolation using Eqns. 7L and 4L. This serves as an independent test. Unfortunately most of our guanidine studies were performed at neutral pH where the thermal transition cannot be studied. However, amongst our data and subsequent investigations of R. Hawkes (unpublished data), two different companion sets of sata (AH and Tm from thermal studies, Ab°3 and Cm from guanidinium studies at the same pH) were found. These were consistent with the lineal model. R. Hawkes is at present performing experiments which will test the dependence of the free energy on guanidinium chloride concentration in a more systematic fashion. Results
Since studies were made on three enzymes using four experimental parameters in two pH ranges, we shall present representative experimental results rather than giving full details. The four experimental parameters were {1) circular dichroism at 223 nm which is a measure of backbone conformation; (2) circular dichroism in the aromatic region 270--300 nm, which is sensitive to aromatic side-chain conformation; (3) fluorescence emission intensity for excitation at 280 nm, and (4) the wavelength of the fluorescence maximum. The latter two measurements are dominated by signals from t r y p t o p h a n residues except of course for e R R R in which all t r y p t o p h a n residues have been replaced. For studies of thermal denaturation we were forced to work in the acid region of the pH spectrum because of the insolubility of denatured T4 lysozyme in neutral aqueous solution. In the presence of guanidinium salts, however, it proved possible to perform unfolding studies at neutrality because the denatured form is solubilized by the denaturing agent. Consequently, studies were undertaken both at neutral pH, because this is the condition where the enzyme is active, and at acid pH, for the purpose of comparison with the results of the thermal denaturation studies [2].
Wild-type T4 lysozyme Data on this protein have been reported earlier [1] and are partially presented in Fig. 1 for purposes of comparison with the m u t a n t lysozymes.
332
? 0 x
-I0 i
,E,
340 E ,
Biochimica et Biophysica Acta, 5 8 0 ( 1 9 7 9 ) 3 2 7 - - 3 3 8 © E l s e v i e r / N o r t h - H o l l a n d Biomedical Press
BBA 38286
STABILITY OF PHAGE T4 LYSOZOMES II. UNFOLDING WITH GUANIDINIUM CHLORIDE
M.L. E L W E L L * a n d J . A . S C H E L L M A N
Institute of Molecular Biology, University of Oregon, Eugene, OR 97403 (U.S.A.) (Received March 3rd, 1979)
Key words: Lysozyme; Guanidine denaturation; Mutant protein; Stability; Temperature sensitivity; (Phage T4)
Summary The denaturation by guanidinium chloride of three phage lysozymes (wild type and two mutants) was investigated. The study of solvent denaturation permitted the investigation of the relative stabilities of the proteins at neutral pH, in contrast to thermal denaturation studies reported earlier which could only be performed in acid pH. The results were interpreted assuming t h a t the free energy of solution of proteins is a linear function of denaturant concentration. Using standard t h e r m o d y n a m i c formulas this permits the calculation of the stabilities of the three proteins in the absence of guanidinium chloride. The single point m u t a t i o n Trp 138 -* Tyr leads to relatively large changes in stability and the interaction of the protein with guanidinium chloride. The changes associated with the subsequent double mutation, Trp 126 -~ Tyr, Trp 158 -* Tyr, are much smaller indicating a relatively smooth adjustment of the protein structure to the changed side chains. Models of the structural effects of point mutations are discussed. It is found t h a t the m u t a t i o n at position 138 does n o t fit a model in which the effect of a substitution is to introduce an energetic strain in the structure. It does fit a model in which there is a partial unravelling of the structure as a result of the mutation. However, there are no changes in the backbone circular dichroism spectra associated with the mutation. The two observations are not necessarily in conflict. Further physical studies are required for the resolution of the problem. * Present address: Department of Chemistry, University of Texas at Dallas, Richardson, TX 75080, U.S.A. Abbreviations: Phage T4 lysozyme, T4 lysozyme; Phage eR1-75 T4 lysozyme, eR1-75 T4 lysozyme; Phage eRCIOeRleRD1 T4 lysozyme, eRRR T4 lysozyme.
328
Introduction
This paper explores the effect of point mutations on the unfolding of phage T4 lysozyme in solutions of guanidinium chloride. The preparation and properties of the enzyme have been described in an earlier publication [1] and we have recently reported the effect of the same point mutations on its thermal stability and t h e r m o d y n a m i c properties [2]. Briefly, the wild-type enzyme has three t r y p t o p h a n residues at positions 126, 138 and 158. Tryptophans 126 and 158 are non-essential for enzyme activity. Tryptophan 138 is essential for activity but 50% of the activity is retained when tyrosine is substituted for t r y p t o p h a n at this position. The two mutants to be discussed in this paper are the amber revertant eR1-75, which has tryptophan 138 replaced by tyrosine and m u t a n t eRRR which is a triple revertant with all three tryptophans replaced tyrosines. The previous paper demonstrated that the substitution of tyrosine for t r y p t o p h a n at position 138 had a pronounced effect on the thermal stability of this enzyme, but that the additional effect of tyrosine substitutions at positions 128 and 158 was small by comparison. The present paper extends these comparative studies to solvent denaturation, specifically denaturation in guanidinium chloride. The use of guanidinium chloride makes it possible to work at neutral pH and higher protein concentration without the solubility problems encountered in the thermal denaturation, thereby supplying important information on the stability of the protein at physiological pH values. If the free energy of interaction of the protein and guanidinium chloride is known, it is possible to obtain the free energy of stabilization of the protein in the absence of solvent denaturant by extrapolation [3]. This is the key piece of information that is required for a discussion of stability. The extrapolation is made provisionally in this paper on the basis of a linear free energy law. Evidence for the applicability of this law will be adduced in Discussion. Materials and Methods The materials, methods and instrumentation employed were identical to those used previously [1,2]. The viral strains and methods of purification and isolation of T4 lysozyme have been previously described [1,2]. The fluorescence emission was monitored as a function of Gdn HC1 concentration using either the Hitachi-Perkin-Elmer MPF-2A spectrofluorimeter or the Schoeffel RRS 1000 spectrofluorimeter. Fluorescence measurements were made at protein concentrations of 0.07 mg/ml. Aromatic CD measurements were made at protein concentrations of 1 mg/ml. Far ultraviolet CD measurements were made at 0.02 mg/ml protein. The stock Grin HC1 solutions were 5--6 M Gdn HC1 in H20 and were diluted to the desired Gdn HC1 concentration. For neutral pH experiments, the stock Gdn HC1 was diluted with 0.2 M sodium phosphate buffer, pH 6, 10 -3 M mercaptoethanol. For acid pH experiments, the stock Gdn HC1 was diluted with water and the pH adjusted with HC1. Gdn HCI concentration was determined by measuring the index of refraction of diluted Gdn HC1 samples containing no T4 lysozyme [4]. Baselines for the calculation of equilibrium constants were extrapolated horizontally as in Fig. 3.
329
Solvent denaturation Our experimental results indicate that the unfolding of the wild-type enzyme as well as the eR1-75 and e R R R m u t a n t lysozymes may be interpreted with a two state model. Evidence for this is the fact that all solvent and thermal unfolding curves are monophasic and follow a course which is independent of the physical technique used to follow them. (See Results and Ref. 1.) For a two state model N~D the change in standard chemical potential which accompanies the unfolding of a protein may be written as A/2 = / ~ / 0 + A/Aint
(1)
where A~ ° is the change in standard chemical potential in the absence of guanidinium chloride and A~ is the change in standard chemical potential in its presence. A~in t represents the difference in free energy of interaction of the D and N states with the added guanidinium chloride. To parallel another discussion, [3] we write it as A~/in t = R T A b 2
where R T A b 2 is the change in excess free energy which acompanies the unfolding i.e. the free energy of interaction with guanidinium chloride, and subscript 2 indicates the protein c o m p o n e n t of the system. Introducing equilibrium constants via - - R T in K = A~ in K = In K ° - - Ab2
(2)
where K ° = [D]/[N] in the absence of guanidinium chloride and K = [D]/[N] in its presence. Using C3 to symbolize guanidinium chloride concentration we define Cm (the analog of Tin) as the concentration of denaturant at which [D] = [N]. At Cm, K = 1 so we have in K ° = Ab2
(C3 = Cm)
(3)
or
A~ ° = - - R T A b 2
(4)
where b2 is to be evaluated at Cm. A more detailed t h e r m o d y n a m i c analysis [3] shows also that {O l n g ~
= AF32
(5)
where a is the activity of the guanidinium chloride and AF is the change in the t h e r m o d y n a m i c binding parameter F which occurs when the molecule is unfolded. If the guanidinium binding is stoichiometric, AF = Ar where Ar is the increase in bound ligand upon unfolding based on multiple equilibrium theory. More generally, F represents the excess guanidinium chloride in the solvation domain of the macromolecule even if this excess is not describable in terms of denumerable complexes.
330
A special case of interest is when Ab: is a linear funct i on of d e n a t u r a n t concen tr atio n Ab2 = Ab°~C3
(6L)
where Ab°3 is the virial coefficient for the interation of c o m p o n e n t s 2 and 3. The label L on the formula indicates the restriction of a linear free energy model. In this case Eqns. 3 and 4 may be rewritten as In K ° = Ab°~Cm
(3L)
A~ 0 = --RTAb°3Cm
(4L)
Experimentally one evaluates K as a function of C3. Cm is the c o n c e n t r a t i o n at which K = 1; Ab°3 is found from the slope of In K vs. C3 via the relation o In K _ Ab:3 ~C3
(7L)
The binding is determined by plotting in K vs. in a (Eqn. 5). Activity data for guanidinium chloride are available from two investigations [5,6]. The values of Cm and Ab°3 obtained above permit the evaluation of the stabilization free energy of the protein in the absence of the d e n a t u r a n t via Eqn. 4L. If the linear free energy relation is n o t valid then Ab2 may be written as a p o w e r series m b 2 = m b ° 3 C 3 4- A c 203 3 C 3 2 -Jr- A d 2 03 3 3 C 33 + higher terms
(8)
where it is assumed that the protein solution is sufficiently dilute that proteinprotein interactions are u n i m p o r t a n t and that the guanidinium chloride concen tr atio n is sufficiently high that the Debye-Hiickle term in C3 is swamped o u t [7]. In this case --ABE must be evaluated by a careful analysis of K as a f u n c t i o n o f C3 not only near Cm but over as wide a range as possible using the relations In K
=
in K ° - -
Ab°3C
3 --
A c 2 03 3 C
23 - -
A d 2 303 3 C
33 - -
higher terms
(9)
or O In K . 3C3
Ab03 . .
2 A c .0 3 3 C 3
0 3Ad2333 C32 -- higher terms
(10)
Clearly the free energy of stabilization in the absence of d e n a t u r a n t can be calculated when all of the contributing viral coefficients are known. It is assumed that the ionic strength is sufficiently high from salt c o m p o n e n t s in the system that the electrolyte effect of guanidinium chloride plays a small role at low guanidinium concentration. The linear model is a special case of Eqns. 8--10 and should n o t be used w i t h o u t confirming evidence t ha t In K is linear in C3 over a wide range o f concentrations. This t y p e of data is missing from this r e p o r t partially because the m e t h o d of analysis was not envisaged at the time the experiments were p e r f o r m e d and partially because phage l y s o z y m e is in scarce supply and oft en precipitates at m o d e r a t e concentrations. Consequently most of our w ork was d o n e at low c o n c e n t r a t i o n where accurate data far from K = I are hard to obtain.
331 Nevertheless we shall provisionally make use of the linear free energy model in interpreting the data on the mutants at neutral pH. This is temporarily justified by the findings of Green and Pace [8] and Aune and Tanford [6] that a linear relation is followed over the full range of accessible concentration for several proteins. The former investigation included urea as well as guanidinium chloride as a denaturant. However, in a personal communication Dr. Pace has indicated that he does not believe that the interaction parameter Ab: is linear in guanidinium chloride for the denaturation of myoglobin. In addition Pfeil and Privalov [9] have found that the enthalpy of interaction of guanidinium chloride with hen egg white lysozyme is essentially linear for both the native and denatured forms. While this is only the enthalpic part of RTAb2 it adds credibility to linear t h e r m o d y n a m i c relationships. Another cross-check can be obtained by comparing the stabilization free energies obtained by extrapolating thermal data to ambient temperature (25°C or 37.5°C) with those obtained by the linear extrapolation using Eqns. 7L and 4L. This serves as an independent test. Unfortunately most of our guanidine studies were performed at neutral pH where the thermal transition cannot be studied. However, amongst our data and subsequent investigations of R. Hawkes (unpublished data), two different companion sets of sata (AH and Tm from thermal studies, Ab°3 and Cm from guanidinium studies at the same pH) were found. These were consistent with the lineal model. R. Hawkes is at present performing experiments which will test the dependence of the free energy on guanidinium chloride concentration in a more systematic fashion. Results
Since studies were made on three enzymes using four experimental parameters in two pH ranges, we shall present representative experimental results rather than giving full details. The four experimental parameters were {1) circular dichroism at 223 nm which is a measure of backbone conformation; (2) circular dichroism in the aromatic region 270--300 nm, which is sensitive to aromatic side-chain conformation; (3) fluorescence emission intensity for excitation at 280 nm, and (4) the wavelength of the fluorescence maximum. The latter two measurements are dominated by signals from t r y p t o p h a n residues except of course for e R R R in which all t r y p t o p h a n residues have been replaced. For studies of thermal denaturation we were forced to work in the acid region of the pH spectrum because of the insolubility of denatured T4 lysozyme in neutral aqueous solution. In the presence of guanidinium salts, however, it proved possible to perform unfolding studies at neutrality because the denatured form is solubilized by the denaturing agent. Consequently, studies were undertaken both at neutral pH, because this is the condition where the enzyme is active, and at acid pH, for the purpose of comparison with the results of the thermal denaturation studies [2].
Wild-type T4 lysozyme Data on this protein have been reported earlier [1] and are partially presented in Fig. 1 for purposes of comparison with the m u t a n t lysozymes.
332
? 0 x
-I0 i
,E,
340 E ,
