Bux'htmica et Biophystca Acta, I [178 ( 1991 ) 336-338 'r lt)ql Elsevier Science Publishers B V. 016%4838/91/$03.50 ADONIS 016748389100233M
336
Proton pathways in lysozyme S. Bone h:~muw ~;,[3h)lcc ular and Btomt~lct uhlr E/e( tr,mlcs. Unit erstt'~ of Wales Bangor. Ban,;or. G~tvm'dd ¢U. K
(Received t3 November 1990)
Key gords: Protonic conduction: Hydrogen bonding. Lysozyme; Protein.bound ~ater
Protonic conduction studies are reported for lysozyme as a function of the number of bound water molecules. Lysozyme samples employing proton-injecting palladium black electrodes exhibited conductivities up to eight orders of magnitude greater than those retained between control (copper) electrodes. The results indicate that water involved in multiple hydrogen bond contact with the enzyme together with hydrogen bonded segments of the enzyme structure provide a h)drogen bond network which is capable of supporting considerable protonic conduction.
Introduction X-ray crystallography has indicated the existcn¢c of considerable hydrogen bonding in the structure of lysozyme. In addition to the hydrogen bonding associated with helical and ~B main chain conformations, there are at least 33 hydrogen bond interactions between polar side chains, most of which involve clusters of acceptor and donor groups situated on the surface of the lysozym¢ molecule [1]. Crystallography has also gone some way towards elucidating the positions and structure of water molecules hydrogen bonded to the enzyme. Some 33 to 35 water molecules have been located in positions consistent with their making multiple hydrogen bonds with the lysozyme molecule whilst 95 to 105 waters appear to be engaged in single hydrogen bonding to the enzyme [2]. Despite the existence of detailed structural information, very little is known about how the extensive hydrogen bonding affects the physico-chemical properties of the enzyme, particularly with regard to proton transiocation, and the implications this might have in enzyme activity. It is likely, for example, that the proton translocational motions involved in the catalytic process at the active site take place via hydrogen bonding associated with polar side chains and enzyme-bound water molecules rather than through contact with the bulk aqueous solvent. It has been shown that lysozyme is enzymatically active at hydration levels well below true solution conditions [3]
Corrcspondcncc: S. Bone. Institute of Molecular and Biomolccular Electronics, Uni,~ersity of Wales Bangor. Dean Street, Bangor. Gwyncdd LL57 IUT. U K .
and many enzymes are also active in the hydrated state suspended in organic media [4,5]. In these cases, in the absence of bulk water, proton motions must be supported by the hydrated enzyme. The work presented here describes the ability of hydrated lysozyme to support these protonic motions which are crucial to enzyme activity. Materials and Methods Hen egg-white lysozyme (Sigma, Grade 1) was extensively dialysed against ultra-pure water before use. This procedure removed all ions other than those bound as counterions to the ionised side groups of the enzyme. Hydration isotherms were obtained for dialysed lysozyme using a Sartorius vacuum mierobalanee. For electrical measurements, compressed polycrystaltine samples of density ll00 K g . m -3, cross-sectional area 1.3 × 10 -a m 2 and nominal thickness 2 x 10 -3 m, were retained between spring-loaded electrodes in an environment in which the relative humidity could be controlled to within 0.5~. In order to assess the role played by water-based hydrogen-bonds in proton conduction through the enzyme, it was necessary to vary the amount of enzyme-bound water and this required the u ~ of polycrystalline rather than single crystal enzyme samples. In each series of measurements, control samples having polished copper electrodes were simultaneously investigated with sampies retained between 'proton injecting' electrodes composed of electrochemically deposited palladium black on a copper substrate, following the method used by Freund and Wengeler [6] in their proton conduction studies of solid ionic hydroxides. On attaining equilib-
337 rium with the desired ~alue of water vapour prc,%urc, the samples were subjected to an applied electric field of 1 × 10: V m - t , and the resulting steady state conduction current monitored using a Keithley electrometer. Dry hydrogen gas at atmospheric pressure ~as then admitted into the chamber containing the samples and the changes in conduction current obse~'ed.
~
~
~~ i ~ + ~ +
+ . . . . . . . . . . . . . . .
/
/ -7-
[
+e-
,.,~
~...g! "
Results On admitting hydrogen to the test chamber the current exhibited by the samples retained between the palladium electrodes was observed to increase by a factor of up to 10". The injection of hydrogen gas did not significantly alter the conduction current observed in the samples with the copper electrodes. An example of typically observed responses is shown in Fig. 1 for samples containing 35 water molecules per lysozyme molecule. The hydration dependence of the conductivity is shown in Fig. 2 for both copper and palladium black electrodes. The copper electrodes exhibit an insignificant proton injecting capability and therefore the observed conductivity in this case arises from mobile protons and ions that are intrinsic to the bulk or surface structure of the enzyme samples. Recent conduction and solid-state electrolysis measurements on iysozyme containing varying ionic concentrations have indicated that, for samples employing non-proton injecting (copper) electrodes, the dominant charge carrier in the low hydration region ( < 35 water molecules/lysoz~Tne) is protons while at higher water contents the conduction is primarily due to ions [7]. For the dialysed lysozyme studied here. therefore, it is likely that the sample conductivity obser,'ed with copper electrodes at water contents abo~'e 35 water
..+J
-6-
+. 5
.~ ,
.•
40-' _J
_+,!
/
.....
Pig. I. Timc-cx:mr~ of lh¢ e l ~ cl ol a h)~+ogen pulse on the c~,md~ctivities o f I ~ z ' ~ m ¢ +,ample~ (l~drali~m of 35 ~ a t e r m o l ¢ c u l e ~ / l y ~ - m e ) r e t a i n e d bet~'een c o o e r e l e c l r ~ e s (C) anal ptoto~n-injecli n s p a l l a d i u m black e l e c t r o d e s ( × )+
g
-~" ~
t c~
o
+
F+g 2 Variation of the conducti-,tt~ of ly-oz-~me ~, a lunction ,at the n u m b e r of bound ~,atcr motecule~, at =.'9/~ K. (:-1 copper clcctrt~dc~: ! ~ ~ pr+~ton.in~ec,,ing p a l l a d i u m black elcctrcMes
m o l e c u l e s / i y s o ~ m e is due to migration of counterion species bound to ioni~d ~i~c chains which were not removed b~ dialysis. The effects of the proton injecting electrodes and hydration in increasing the observed conductivity is clearly evident in Fig. 2 and gives support to the concept that i y m ~ m e is capable of supporting a significant protonic conduclivit3+ It can be wen that at relati~rely low hydrations ( = 35 water molecules w r '+ysoz~,me) the conductivity with proton-inlecting efcc+ trodes is some ~,+ighl orders of magnitude greater than that of the control+ The observed protonic conductivity can be ~+cn to r i ~ rapidly for hydration values up to approx+ 35 water molecules per ty~z3'me after which, at higher hydrations, it remains relativeb insensitive to increasing hydration. This is c l o ~ to the number (.~ water molecules per lysozyme) found, in previous dielectric studies [8,ql on the binding of water to hen egg-white lysozyme, to be irrotationally bound to the enz~Tne structure. T h e ~ waters were considered bound by two or more hydrogen bonds+ an interpretation consistent with the X-ray crystallographic studies of Blake et ai, [2] who detected 33 and 35 water molecules making two or more hydrogen bond contacts to tortoi~ egg-white and human l~.'soz3me, respectively. The lredration-dependence of the protonie conduction shown in Fig. 2 therefore indicates that bound water molecules form an essential part of the h~drogen btmd network on the enzyme surface. In p:~rticular, it appears that multiply hydrogen bonded water may provide links in the proton
338 path,~a~s bctwccn hydrogen-bonded side chain scgSingly hydrogen bonded water molecules, that is those bound a: water contents in excess of approx. 35 water molecules per lysozyme, do no~ appear to influence the level of protonic conduction. This is in marked contrast to the conduction of ions which, as shown in Fig. 2 (copper electrodes), increases exponentially with increasing water content. In this case the hydration-induced increase in conductivity is probably due to the ability of loosely bound water to e[ectrostaticaUy screen the ionic charges from polar groups in the enzyme, thereby reducing the energies of interaction.
bonds has also been demonstrated [13] with a number of hydrogen bonded water molecules being shared by two or more hydrogen bond circles. The existence of such non-standard hydrogen bonding in disordered systems may provide the basis for a protonic conduction mechanism different from that envisaged in hydrogen bonded crystal lattices. It is possible that flip-flop and circular h~,drogen bonding exists on the surface of large, hydrophilic molecules such as enzymes where proton acceptance and donation at the active site are often crucial steps in enzyme activity.
Acknowledgements
Discussion
I would like to thank Dr H. Morgan and Professor R. Pethig for helpful discussions.
The accepted theory of proton conduction in hydrogen bonded crystals such as ice as reviewed by Onsager [10] involves two alternating processes. In the so-called h o p / t u r n mechanism, an excess proton hops to an adjacent site and this is followed by the turning of the donor group to assume its original orientation. Protons can be repeatedly transported along the hydrogen bond chain through the alternation of these two processes. If the h o p / t u r n mechanism were solely responsible for proton translocation in lysozyme, this would exclude the involvement of those hydrogen bonds associated only with the peptide groups since the reorientation of such groups would involve prohibitively large activation energies [11]. In structures of a more disordered nature such as proteins, hydrogen bonding not constrained by the symmetry requirements of the crystal lattice and therefore more free to arrange is likely. For instance in ¢t-cyclodextrin a far less rigid hydrogen bond structure occurs with networks containing a proportion of socalled 'flip-flop' hydrogen bonds which exhibit cooperalive and concerted rearrangements of hydrogen atoms (domino effect) [12]. in addition, the existence of extcndcd networks o,~ water-based circular hydrogen
References 1 Imoto. T,, Johns~m. L.N., North. A.C.T.. Phillip,~ D.C. and Rupicy, JA. (1972) in The Enzymes (Boyer, P.D,, ed,), Vol. VII, pp. 665-868, Academic Press, New York. 2 Blake. C.C.F., Pulford, W.C.A, and Arty®niuk, P.J. (1983)J. Mot. Biol. 167. 6'~~-723. 3 Rupley, J.A., C,alton, E. and Careri. G. (1983) Trends Biochem. Sci. 8(1), 18-22. 4 Zaks. A. and Kibanov, A.M. (i988) J. Biol. Chem. 263 (17). 8017-8021 5 Hailing, P.J. (1587) Biotech. Adv_ 5, 47-84. 6 Freund, F. and Wcngeler H. (1980) Ber. Bunscnges. Phys. Chem. 84, 866-873. 7 Morgan, H, and Pethig, R, 1986) J, Chem, So¢, Faraday Trans. ! 82, 143-156, 8 Bone, S. and Pethi6, R. (1982) J. Mot Biol_ 157, 571-575 9 Bone, S= and Pethig, R. (19851J. Mo|. Biol. 181,323-326. 10 Onsager. L. (1973) in Physics and Chemistry of Ice (Whalle~, E., Jones. S.J. and Gold. L.W., eds.I, pp. 7-12, Royal Society. OtIo~,a. 11 Onsager. L (1970) in Physical Principles of Biological Membranes (Snell. F,. Wolken, J , Iver~m~ G. and l_am. J , eds), pp. 137 141+ Gordon and Breach. Ne~ Y o r k 12 Sacngcr. W., Borzoi. Ch,, Hingcrty, B. and Brow,n, G.M. ~1982) Nature 296. 581-583, 13 $aengcl, W (1970~ Nature 279. 343-341-.
336
Proton pathways in lysozyme S. Bone h:~muw ~;,[3h)lcc ular and Btomt~lct uhlr E/e( tr,mlcs. Unit erstt'~ of Wales Bangor. Ban,;or. G~tvm'dd ¢U. K
(Received t3 November 1990)
Key gords: Protonic conduction: Hydrogen bonding. Lysozyme; Protein.bound ~ater
Protonic conduction studies are reported for lysozyme as a function of the number of bound water molecules. Lysozyme samples employing proton-injecting palladium black electrodes exhibited conductivities up to eight orders of magnitude greater than those retained between control (copper) electrodes. The results indicate that water involved in multiple hydrogen bond contact with the enzyme together with hydrogen bonded segments of the enzyme structure provide a h)drogen bond network which is capable of supporting considerable protonic conduction.
Introduction X-ray crystallography has indicated the existcn¢c of considerable hydrogen bonding in the structure of lysozyme. In addition to the hydrogen bonding associated with helical and ~B main chain conformations, there are at least 33 hydrogen bond interactions between polar side chains, most of which involve clusters of acceptor and donor groups situated on the surface of the lysozym¢ molecule [1]. Crystallography has also gone some way towards elucidating the positions and structure of water molecules hydrogen bonded to the enzyme. Some 33 to 35 water molecules have been located in positions consistent with their making multiple hydrogen bonds with the lysozyme molecule whilst 95 to 105 waters appear to be engaged in single hydrogen bonding to the enzyme [2]. Despite the existence of detailed structural information, very little is known about how the extensive hydrogen bonding affects the physico-chemical properties of the enzyme, particularly with regard to proton transiocation, and the implications this might have in enzyme activity. It is likely, for example, that the proton translocational motions involved in the catalytic process at the active site take place via hydrogen bonding associated with polar side chains and enzyme-bound water molecules rather than through contact with the bulk aqueous solvent. It has been shown that lysozyme is enzymatically active at hydration levels well below true solution conditions [3]
Corrcspondcncc: S. Bone. Institute of Molecular and Biomolccular Electronics, Uni,~ersity of Wales Bangor. Dean Street, Bangor. Gwyncdd LL57 IUT. U K .
and many enzymes are also active in the hydrated state suspended in organic media [4,5]. In these cases, in the absence of bulk water, proton motions must be supported by the hydrated enzyme. The work presented here describes the ability of hydrated lysozyme to support these protonic motions which are crucial to enzyme activity. Materials and Methods Hen egg-white lysozyme (Sigma, Grade 1) was extensively dialysed against ultra-pure water before use. This procedure removed all ions other than those bound as counterions to the ionised side groups of the enzyme. Hydration isotherms were obtained for dialysed lysozyme using a Sartorius vacuum mierobalanee. For electrical measurements, compressed polycrystaltine samples of density ll00 K g . m -3, cross-sectional area 1.3 × 10 -a m 2 and nominal thickness 2 x 10 -3 m, were retained between spring-loaded electrodes in an environment in which the relative humidity could be controlled to within 0.5~. In order to assess the role played by water-based hydrogen-bonds in proton conduction through the enzyme, it was necessary to vary the amount of enzyme-bound water and this required the u ~ of polycrystalline rather than single crystal enzyme samples. In each series of measurements, control samples having polished copper electrodes were simultaneously investigated with sampies retained between 'proton injecting' electrodes composed of electrochemically deposited palladium black on a copper substrate, following the method used by Freund and Wengeler [6] in their proton conduction studies of solid ionic hydroxides. On attaining equilib-
337 rium with the desired ~alue of water vapour prc,%urc, the samples were subjected to an applied electric field of 1 × 10: V m - t , and the resulting steady state conduction current monitored using a Keithley electrometer. Dry hydrogen gas at atmospheric pressure ~as then admitted into the chamber containing the samples and the changes in conduction current obse~'ed.
~
~
~~ i ~ + ~ +
+ . . . . . . . . . . . . . . .
/
/ -7-
[
+e-
,.,~
~...g! "
Results On admitting hydrogen to the test chamber the current exhibited by the samples retained between the palladium electrodes was observed to increase by a factor of up to 10". The injection of hydrogen gas did not significantly alter the conduction current observed in the samples with the copper electrodes. An example of typically observed responses is shown in Fig. 1 for samples containing 35 water molecules per lysozyme molecule. The hydration dependence of the conductivity is shown in Fig. 2 for both copper and palladium black electrodes. The copper electrodes exhibit an insignificant proton injecting capability and therefore the observed conductivity in this case arises from mobile protons and ions that are intrinsic to the bulk or surface structure of the enzyme samples. Recent conduction and solid-state electrolysis measurements on iysozyme containing varying ionic concentrations have indicated that, for samples employing non-proton injecting (copper) electrodes, the dominant charge carrier in the low hydration region ( < 35 water molecules/lysoz~Tne) is protons while at higher water contents the conduction is primarily due to ions [7]. For the dialysed lysozyme studied here. therefore, it is likely that the sample conductivity obser,'ed with copper electrodes at water contents abo~'e 35 water
..+J
-6-
+. 5
.~ ,
.•
40-' _J
_+,!
/
.....
Pig. I. Timc-cx:mr~ of lh¢ e l ~ cl ol a h)~+ogen pulse on the c~,md~ctivities o f I ~ z ' ~ m ¢ +,ample~ (l~drali~m of 35 ~ a t e r m o l ¢ c u l e ~ / l y ~ - m e ) r e t a i n e d bet~'een c o o e r e l e c l r ~ e s (C) anal ptoto~n-injecli n s p a l l a d i u m black e l e c t r o d e s ( × )+
g
-~" ~
t c~
o
+
F+g 2 Variation of the conducti-,tt~ of ly-oz-~me ~, a lunction ,at the n u m b e r of bound ~,atcr motecule~, at =.'9/~ K. (:-1 copper clcctrt~dc~: ! ~ ~ pr+~ton.in~ec,,ing p a l l a d i u m black elcctrcMes
m o l e c u l e s / i y s o ~ m e is due to migration of counterion species bound to ioni~d ~i~c chains which were not removed b~ dialysis. The effects of the proton injecting electrodes and hydration in increasing the observed conductivity is clearly evident in Fig. 2 and gives support to the concept that i y m ~ m e is capable of supporting a significant protonic conduclivit3+ It can be wen that at relati~rely low hydrations ( = 35 water molecules w r '+ysoz~,me) the conductivity with proton-inlecting efcc+ trodes is some ~,+ighl orders of magnitude greater than that of the control+ The observed protonic conductivity can be ~+cn to r i ~ rapidly for hydration values up to approx+ 35 water molecules per ty~z3'me after which, at higher hydrations, it remains relativeb insensitive to increasing hydration. This is c l o ~ to the number (.~ water molecules per lysozyme) found, in previous dielectric studies [8,ql on the binding of water to hen egg-white lysozyme, to be irrotationally bound to the enz~Tne structure. T h e ~ waters were considered bound by two or more hydrogen bonds+ an interpretation consistent with the X-ray crystallographic studies of Blake et ai, [2] who detected 33 and 35 water molecules making two or more hydrogen bond contacts to tortoi~ egg-white and human l~.'soz3me, respectively. The lredration-dependence of the protonie conduction shown in Fig. 2 therefore indicates that bound water molecules form an essential part of the h~drogen btmd network on the enzyme surface. In p:~rticular, it appears that multiply hydrogen bonded water may provide links in the proton
338 path,~a~s bctwccn hydrogen-bonded side chain scgSingly hydrogen bonded water molecules, that is those bound a: water contents in excess of approx. 35 water molecules per lysozyme, do no~ appear to influence the level of protonic conduction. This is in marked contrast to the conduction of ions which, as shown in Fig. 2 (copper electrodes), increases exponentially with increasing water content. In this case the hydration-induced increase in conductivity is probably due to the ability of loosely bound water to e[ectrostaticaUy screen the ionic charges from polar groups in the enzyme, thereby reducing the energies of interaction.
bonds has also been demonstrated [13] with a number of hydrogen bonded water molecules being shared by two or more hydrogen bond circles. The existence of such non-standard hydrogen bonding in disordered systems may provide the basis for a protonic conduction mechanism different from that envisaged in hydrogen bonded crystal lattices. It is possible that flip-flop and circular h~,drogen bonding exists on the surface of large, hydrophilic molecules such as enzymes where proton acceptance and donation at the active site are often crucial steps in enzyme activity.
Acknowledgements
Discussion
I would like to thank Dr H. Morgan and Professor R. Pethig for helpful discussions.
The accepted theory of proton conduction in hydrogen bonded crystals such as ice as reviewed by Onsager [10] involves two alternating processes. In the so-called h o p / t u r n mechanism, an excess proton hops to an adjacent site and this is followed by the turning of the donor group to assume its original orientation. Protons can be repeatedly transported along the hydrogen bond chain through the alternation of these two processes. If the h o p / t u r n mechanism were solely responsible for proton translocation in lysozyme, this would exclude the involvement of those hydrogen bonds associated only with the peptide groups since the reorientation of such groups would involve prohibitively large activation energies [11]. In structures of a more disordered nature such as proteins, hydrogen bonding not constrained by the symmetry requirements of the crystal lattice and therefore more free to arrange is likely. For instance in ¢t-cyclodextrin a far less rigid hydrogen bond structure occurs with networks containing a proportion of socalled 'flip-flop' hydrogen bonds which exhibit cooperalive and concerted rearrangements of hydrogen atoms (domino effect) [12]. in addition, the existence of extcndcd networks o,~ water-based circular hydrogen
References 1 Imoto. T,, Johns~m. L.N., North. A.C.T.. Phillip,~ D.C. and Rupicy, JA. (1972) in The Enzymes (Boyer, P.D,, ed,), Vol. VII, pp. 665-868, Academic Press, New York. 2 Blake. C.C.F., Pulford, W.C.A, and Arty®niuk, P.J. (1983)J. Mot. Biol. 167. 6'~~-723. 3 Rupley, J.A., C,alton, E. and Careri. G. (1983) Trends Biochem. Sci. 8(1), 18-22. 4 Zaks. A. and Kibanov, A.M. (i988) J. Biol. Chem. 263 (17). 8017-8021 5 Hailing, P.J. (1587) Biotech. Adv_ 5, 47-84. 6 Freund, F. and Wcngeler H. (1980) Ber. Bunscnges. Phys. Chem. 84, 866-873. 7 Morgan, H, and Pethig, R, 1986) J, Chem, So¢, Faraday Trans. ! 82, 143-156, 8 Bone, S. and Pethi6, R. (1982) J. Mot Biol_ 157, 571-575 9 Bone, S= and Pethig, R. (19851J. Mo|. Biol. 181,323-326. 10 Onsager. L. (1973) in Physics and Chemistry of Ice (Whalle~, E., Jones. S.J. and Gold. L.W., eds.I, pp. 7-12, Royal Society. OtIo~,a. 11 Onsager. L (1970) in Physical Principles of Biological Membranes (Snell. F,. Wolken, J , Iver~m~ G. and l_am. J , eds), pp. 137 141+ Gordon and Breach. Ne~ Y o r k 12 Sacngcr. W., Borzoi. Ch,, Hingcrty, B. and Brow,n, G.M. ~1982) Nature 296. 581-583, 13 $aengcl, W (1970~ Nature 279. 343-341-.
