THE JOURNAL OF BIOLOGICAL CHEMIWRY 0 1990 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 265, No. 20, Issue of July 15, pp. 11’/&3-11795,199O Printed in U.S.A. The Role of Va16’(E1 1) in Ligand Binding to Sperm Whale Myoglobin SITE-DIRECTED MUTAGENESIS OF A SYNTHETIC GENE* (Received for publication, February 13,199O) Karen D. Egeberg, Barry A. Springer& and Stephen G. Sligarg From the Departments of Biochemistry and Chemistv, University of Illinois, Urbana, Illinois 61801 Theodore E. Carverli, Ronald J. Rohlfsnll, and John S. Olson** From the Department of Biochemistry and Cell Biology, Rice University, Houston, Texas 77251 Site-directed mutants of sperm whale myoglobin were prepared to probe the functional role of the highly conserved distal pocket valine residue, Val’*(El 1). This amino acid was replaced with Ala, Ile, and Phe to examine the effects of the side chain volume at position 68 on ligand binding. Three double mutants were also constructed in which the distal Hisa4(E’7) was replaced with Gly and Vale8 was replaced with Ala, Ile, and Phe to determine the effects of size at position 68 in the absence of the distal histidine. Association and disso- ciation rate constants for 02, CO, and alkyl isocyanide binding were measured by stopped-flow rapid mixing, conventional flash, and laser photolysis techniques at pH 7, 20 “C. The association rate constants for the binding of all eight ligands to the single mutants de- creased in the order Alaee > Valss (native) > IlesS myoglobin, indicating that the 68(E 11) residue is part of the overall kinetic barrier. A similar pattern was observed for the association constants of the double mutants: Glya4/Alase > Glys4/Vales > Glye4/Ilees. Thus, increasing size of the E 11 side chain inhibits the rate of ligand binding even in the absence of histidine at position 64. Substitution of Ala for Vales had little effect on 02 affinity but did increase the affinities for CO and isocyanide binding. The affinities for all of the ligands were decreased for the Ileas mutant. The ligand binding affinities for the Glye4/Alaas, Glye4/Valae, and Glye4/Ilees myoglobins displayed an analogous trend to that of the single mutants, indicating that the equilib- rium interactions between the position 64 and 68 side chains and the bound ligand are roughly additive. Both the association rate constants and dissociation rate constants for O2 and isocyanide binding were de- creased for the Phess mutant myoglobin. These kinetic parameters result in little change in O2 affinity and an increase in &cyanide affinity, relative to the native * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC. Section 1734 solely to indicate this fact. $ Present address: Dept. of Chemistry, University of California, Berkeley, CA 94720. 5 Supported by United States Public Health Service Grants GM- 33775 and GM-31756. To whom reprint requests should be addressed. ll Recipients of graduate fellowships from the National Institutes of Health, Training Grant GM-07933, from the National Institute of Medical Science. (1 Current address: Dept. of Physiological Chemistry, Ohio State University, Columbus, OH 43210. ** Supported by United States Public Health Service Grant GM- 35649, Grant C-612 from the Robert A. Welch Foundation, and Grant 4073 from the Advanced Technology Program of the Texas Higher Education Coordinating Board. protein. Thus, the large benzyl side chain of phenylal- anine at position 68 inhibits the rate of ligand move- ment up to and away from the iron atom but not the final bound state. The role of the distal histidine residue, His(E7),’ in mam- malian myoglobins and hemoglobins has been studied exten- sively and is thought to be the key residue for regulating oxygen affinity (Olson et al., 1988; Springer et al., 1989; Mathews et al; 1989, Rohlfs et al., 1990, and references therein). The distal valine, Val(Ell), is also of interest be- cause it contacts both the bound ligand and His(E7), restrict- ing the size of the binding site in these proteins (Fig. 1). Perutz (1970, 1989) Phillips (1980), Shaanan (1983), and others have suggested that the Val(E1l) residue in both myoglobin and hemoglobin may orient bound oxygen toward the t-amino nitrogen of His(E7) for more efficient hydrogen bonding and that this close proximity to the bound ligand inhibits CO binding due to the preferred linear FeCO geom- etry. Kuriyan et al. (1986) have shown that in sperm whale myoglobin bound CO occupies at least two different, bent orientations. The imidazole ring of His64(E7) is moved slightly away from the bound CO by rotation about the C,-CB bond, but no change was observed for the position of the isopropyl side chain of Vala when compared to that in either oxy- or deoxymyoglobin. Even larger movements of His64 were ob- served in the structure of ethyl isocyanide sperm whale myo- globin reported by Johnson et al. (1989), but again, the posi- tion of Va16s was unchanged. The lack of movement of Va16s does not necessarily imply the absence of steric interactions with bound CO or ethyl isocyanide, but rather that the side chain of His? is more mobile. In both the CO and ethyl isocyanide complexes, the Fe-ligand bonds are severely dis- torted from linearity suggesting significant hindrance by both Hisa and Va16s. Examination of the distal pocket in sperm whale myoglobin reveals no obvious route for ligands to enter or leave due to the tightly packed globin structure. On the basis of theoretical dynamics calculations, Case and Karplus (1979) suggested that a major pathway for ligand entry into the distal pocket 1 The alphanumeric codes (e.g. E7 and Ell) refer to the position of the residue within the helices and loops of the myoglobin folding pattern (Dickerson and Geis, 1983). In the case of native sperm whale myoglobin, E7 and El1 correspond to positions 64 and 68, respec- tively, in the amino acid sequence. The amino acids at the distal His(E7) and Val(E1l) positions in the site-directed mutants are referred to as 64 and 68 for comparison with the native protein even though the E. coli myoglobins have an additional Met at the NH, terminus. 11788 by guest on February 12, 2018 http://www.jbc.org/ Downloaded from
Transcript
THE JOURNAL OF BIOLOGICAL CHEMIWRY 0 1990 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 265, No. 20, Issue of July 15, pp. 11’/&3-11795,199O Printed in U.S.A.
The Role of Va16’(E1 1) in Ligand Binding to Sperm Whale Myoglobin SITE-DIRECTED MUTAGENESIS OF A SYNTHETIC GENE*
(Received for publication, February 13,199O)
Karen D. Egeberg, Barry A. Springer& and Stephen G. Sligarg From the Departments of Biochemistry and Chemistv, University of Illinois, Urbana, Illinois 61801
Theodore E. Carverli, Ronald J. Rohlfsnll, and John S. Olson** From the Department of Biochemistry and Cell Biology, Rice University, Houston, Texas 77251
Site-directed mutants of sperm whale myoglobin were prepared to probe the functional role of the highly conserved distal pocket valine residue, Val’*(El 1). This amino acid was replaced with Ala, Ile, and Phe to examine the effects of the side chain volume at position 68 on ligand binding. Three double mutants were also constructed in which the distal Hisa4(E’7) was replaced with Gly and Vale8 was replaced with Ala, Ile, and Phe to determine the effects of size at position 68 in the absence of the distal histidine. Association and disso- ciation rate constants for 02, CO, and alkyl isocyanide binding were measured by stopped-flow rapid mixing, conventional flash, and laser photolysis techniques at pH 7, 20 “C. The association rate constants for the binding of all eight ligands to the single mutants de- creased in the order Alaee > Valss (native) > IlesS myoglobin, indicating that the 68(E 11) residue is part of the overall kinetic barrier. A similar pattern was observed for the association constants of the double mutants: Glya4/Alase > Glys4/Vales > Glye4/Ilees. Thus, increasing size of the E 11 side chain inhibits the rate of ligand binding even in the absence of histidine at position 64. Substitution of Ala for Vales had little effect on 02 affinity but did increase the affinities for CO and isocyanide binding. The affinities for all of the ligands were decreased for the Ileas mutant. The ligand binding affinities for the Glye4/Alaas, Glye4/Valae, and Glye4/Ilees myoglobins displayed an analogous trend to that of the single mutants, indicating that the equilib- rium interactions between the position 64 and 68 side chains and the bound ligand are roughly additive.
Both the association rate constants and dissociation rate constants for O2 and isocyanide binding were de- creased for the Phess mutant myoglobin. These kinetic parameters result in little change in O2 affinity and an increase in &cyanide affinity, relative to the native
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC. Section 1734 solely to indicate this fact.
$ Present address: Dept. of Chemistry, University of California, Berkeley, CA 94720.
5 Supported by United States Public Health Service Grants GM- 33775 and GM-31756. To whom reprint requests should be addressed.
ll Recipients of graduate fellowships from the National Institutes of Health, Training Grant GM-07933, from the National Institute of Medical Science.
(1 Current address: Dept. of Physiological Chemistry, Ohio State University, Columbus, OH 43210.
** Supported by United States Public Health Service Grant GM- 35649, Grant C-612 from the Robert A. Welch Foundation, and Grant 4073 from the Advanced Technology Program of the Texas Higher Education Coordinating Board.
protein. Thus, the large benzyl side chain of phenylal- anine at position 68 inhibits the rate of ligand move- ment up to and away from the iron atom but not the final bound state.
The role of the distal histidine residue, His(E7),’ in mam- malian myoglobins and hemoglobins has been studied exten- sively and is thought to be the key residue for regulating oxygen affinity (Olson et al., 1988; Springer et al., 1989; Mathews et al; 1989, Rohlfs et al., 1990, and references therein). The distal valine, Val(Ell), is also of interest be- cause it contacts both the bound ligand and His(E7), restrict- ing the size of the binding site in these proteins (Fig. 1). Perutz (1970, 1989) Phillips (1980), Shaanan (1983), and others have suggested that the Val(E1l) residue in both myoglobin and hemoglobin may orient bound oxygen toward the t-amino nitrogen of His(E7) for more efficient hydrogen bonding and that this close proximity to the bound ligand inhibits CO binding due to the preferred linear FeCO geom- etry. Kuriyan et al. (1986) have shown that in sperm whale myoglobin bound CO occupies at least two different, bent orientations. The imidazole ring of His64(E7) is moved slightly away from the bound CO by rotation about the C,-CB bond, but no change was observed for the position of the isopropyl side chain of Vala when compared to that in either oxy- or deoxymyoglobin. Even larger movements of His64 were ob- served in the structure of ethyl isocyanide sperm whale myo- globin reported by Johnson et al. (1989), but again, the posi- tion of Va16s was unchanged. The lack of movement of Va16s does not necessarily imply the absence of steric interactions with bound CO or ethyl isocyanide, but rather that the side chain of His? is more mobile. In both the CO and ethyl isocyanide complexes, the Fe-ligand bonds are severely dis- torted from linearity suggesting significant hindrance by both Hisa and Va16s.
Examination of the distal pocket in sperm whale myoglobin reveals no obvious route for ligands to enter or leave due to the tightly packed globin structure. On the basis of theoretical dynamics calculations, Case and Karplus (1979) suggested that a major pathway for ligand entry into the distal pocket
1 The alphanumeric codes (e.g. E7 and Ell) refer to the position of the residue within the helices and loops of the myoglobin folding pattern (Dickerson and Geis, 1983). In the case of native sperm whale myoglobin, E7 and El1 correspond to positions 64 and 68, respec- tively, in the amino acid sequence. The amino acids at the distal His(E7) and Val(E1l) positions in the site-directed mutants are referred to as 64 and 68 for comparison with the native protein even though the E. coli myoglobins have an additional Met at the NH, terminus.
HpaI (Springer and Sligar, 1987). For the single mutants at position 68, mutagenesis was performed by ligating two different sets of mixed oligonucleotides into the BglII-HpaI digested vector. In one set of mixed oligonucleotides, the normal codon for Val@ (GTC) was changed to ATC or TTC generating Ile or Phe at this position. A second set of mixed oligonucleotides was used to generate Ala (GCC) at the same position. For the double mutants at positions 64 and 68, two mixed sets of oligonucleotides were generated to include the Valm mutants, as stated above, in concert with the change of the His” (CAT) codon to Gly(GGT). In all constructions, the HpaI site was deleted for screening purposes. DNA manipulations were made as described by Maniatis et al. (1982). Mutant sequences were confirmed by dideoxy DNA sequencing as described by Hattori and Sakaki (1986) using Bethesda Research Laboratories sequencing reagents. All enzymes used in producing mutant constructions were purchased from Bethesda Research Laboratories except polynucleotide kinase which was purchased from New England Biolabs. Oligonucleotides were synthesized by the University of Illinois (Urbana-Champaign) Biotechnology Center and purified by reverse-phase high performance liquid chromatography as described by Springer and Sligar (1987). All mutant constructions were expressed in Escherichia coli strain TB-1 (ma, A(lac-pro), strA, thi, ~8OdlacZAMIS,r-,m’; T. 0. Baldwin, Texas A & M, College Station, TX) and the resultant myoglobins purified as described by Springer and Sligar (1987) and Springer et
FIG. 1. The distal pocket of the O2 complex of sperm whale myoglobin. The coordinates were taken from the structure deter- mined by Phillips (1981). The view is from a position near the back of the distal pocket looking out toward the solvent.
of myoglobin was between Va16’ and His&. This idea is sup- ported by the more recent molecular dynamics calculations of Kottalam and Case (1988) and x-ray crystallographic studies of imidazole, phenylhydrazine, and ethyl isocyanide bound to myoglobin in which these bulky ligands force the His64 side chain out of the distal pocket opening a path to solvent (Bolognesi et al., 1982; Ringe et al., 1984; Johnson et al., 1989). Ligand binding studies of site-directed mutants of sperm whale myoglobin and human hemoglobin also indicate that the distal histidine is the dominant barrier to ligand binding (Nagai et al., 1987; Olson et al., 1988; Springer et al., 1989; Mathews et al., 1989; Rohlfs et al., 1990). Although direct movement of the myoglobin Va16’ side chain may not be required for ligand entry, its conserved presence may reflect an interplay between orienting bound ligands for interaction with His64 and allowing a channel for ligand migration to the heme iron.
By using the synthetic sperm whale myoglobin gene de- scribed by Springer and Sligar (1987), three mutant proteins were constructed with Ala, Ile, and Phe substituted for Va16’ in order to examine the functional role of the distal valine in myoglobin and to allow comparisons with the El1 mutants of human hemoglobin prepared by Nagai and co-workers (Math- ews et al., 1989). These mutants, along with native myoglobin, provide a series of proteins containing apolar residues of increasing size at the El1 position: Ala68, Va16’, Ile6’, and Phe@ corresponding to side chain volumes of 25, 75, 102, and 137 A3, respectively (Creighton, 1983). The effects of these substitutions on the functional properties of myoglobin were assessed by measuring rate and equilibrium constants for 02, CO, methyl, ethyl, n-propyl, n-butyl, isopropyl, and tert-butyl isocyanide binding to each of the proteins. The alkyl isocya- nides were used to probe the size of both the distal pocket and the ligand pathway. Finally, comparisons of the roles of His’j4and Valm were made by measuring the ligand-binding properties of double mutants in which His’j4 was replaced with Gly, and Val@ was mutated to Ala, Ile, or Phe within the same protein. The results for these proteins, Gly64/Ala68, Gly64/Ilea, and Gly64/Phe68, allowed a direct measure of the relative contributions of the E7 and El1 residues to the overall kinetic and equilibrium barriers.
MATERIALS AND METHODS
Preparation of Mutants-Mutagenesis of the synthetic sperm whale myoglobin gene (pMb413) was performed by using cassettes inserted at the construction unique restriction enzyme sites BglII and
al. (1989). Rohlfs et al.-(1990) have shown that wild-type myodobin expressed in E. coli exhibits kinetic and equilibrium ligand-binding parameters identical to those of native sperm whale myoglobin ob- tained from Sigma. Phillips et al. (1990) have shown that the three- dimensional structure of synthetic wild-type metmyoglobin is iden- tical to that of native protein except in the region of the NHs-terminal methionine.
Kinetic Measurements-Association rate constants, k’, for CO and alkyl isocyanide binding to the single mutants, Ala=, Ilea, and Phe@’ and native (Va16’) protein were measured by both rapid mixing and conventional flash photolysis techniques as described previously by Rohlfs et al. (1990). The reaction conditions were 5-10 pM myoglobin in 0.1 M potassium phosphate, pH 7.0, 20 “C. Association rate con- stants for O2 binding were measured bv laser photolvsis using the pulsed dye system (Phase-R Model 21OdB) desciibed Gy Rohlfset al. (1990). The 0, comDlexes of Alam. IleG’. and Phea mvoglobin were
” Y
stable and handled in the same way as native myoglobin; however, measurements of 0, association and dissociation with the double mutants containing Glys4 were complicated by high rates of autooxi- dation (see Springer et al., 1989, for a description of how replacement of Hiss4 increases the autooxidation rate). Dissociation rate constants, k, were determined from the analysis of replacement reactions carried out in the stopped-flow rapid mixing apparatus. In these experiments Mb02 or Mb-isocyanide complexes were mixed with high concentra- tions of CO, and MbCO was mixed with NO. The final values of k’ and k for 02, CO, and alkyl isocyanide binding to the single mutants (Table I) were obtained by simultaneously fitting sets of observed association and replacement rates as described by Rohlfs et al. (1990). Equilibrium association constants, K, were calculated from the ratio of the rate constants: K = k’lk.
RESULTS
Association Rate Constants for VuP Mutants-The ob- served time courses for ligand binding to the Ala@, Ile6’, and Phe6’ single mutants were monophasic and fit well to single exponential expressions. The resultant pseudo first-order rate constants depended linearly on ligand concentration, and the kinetic results measured in rapid mixing experiments were the same as those obtained from flash photolysis experiments.
The observed association rate constants (k’) for the Valm mutants are given in Table I, and displayed in Fig. 2A as a function of the size of the amino acid at position 68 (Ell). For all ligands, there is a monotonic decrease in k’ with increasing size of the El1 amino acid for the series Ala@’ to Va16’ (native) to Ilem myoglobin, which suggests that this residue limits the size of the opening for ligand approach to the iron atom. Steric constraints are imposed by the native valine residue as evidenced by the increased association rate constants for 02, CO, and isocyanide binding to Alaa myoglo- bin (Fig. 2A). Even greater hindrance is observed in the Ilem
TABLE I Rate and equilibrium constants for ligand binding
to sperm whale myoglobin containing single substitutions at position 68 (EIl) at pH 7.0,20 “C
The values for native sperm whale myoglobin (Val=) were taken from Springer et al. (1989) and Rohlfs et al. (1990). These authors showed that the rate constants for wild-type Vale’ myoglobin are identical to those for the native protein. The relative errors for the native rate and equilibrium constants were +I6 and f20%, respec- tively. These errors are assumed to apply to the mutant parameters (i.e. +20%). The parameter values were rounded off to two significant figures after K was calculated. The isocyanide abbreviations are: MNC, methyl; ENC, ethyl; nPNC, n-propyl; nBNC, n-butyl; iPNC, isopropyl; tBNC, tert-butyl isocyanide.
Ligand El 1 residue (position 68)
k’ k K
02
co
MNC
ENC
nPNC
nBNC
iPNC
tBNC
Ala 22 18 Val (native) 14 12 Ile 3.2 14 Phe 1.2 2.5
X IF8 M-’
1.2 1.2 0.22 0.48
Ala 1.2 0.021 56 Val (native) 0.51 0.019 27 Ile 0.05 0.024 2.1 Phe 0.25 0.018 14
Ala 0.38 0.76 0.50 Val (native) 0.12 4.3 0.028 Ile 0.050 21 0.0024 Phe 0.013 0.030 0.43
Ala 0.18 0.070 2.6 Val (native) 0.069 0.30 0.23 Ile 0.047 3.4 0.014 Phe 0.0061 0.0035 1.7
Ala 0.11 0.022 5.0 Val (native) 0.042 0.39 0.11 Ile 0.016 1.5 0.011 Phe 0.0040 0.0019 2.1
Ala 0.14 0.011 13 Val (native) 0.030 0.69 0.044 Ile 0.018 1.5 0.012 Phe 0.0058 0.0077 0.75
Ala 0.014 0.0098 1.4 Val (native) 0.012 0.53 0.023 Ile 0.0046 3.0 0.0016 Phe 0.0016 0.0091 0.18
mutant which exhibited 5- to IO-fold decreased association rate constants for CO and O2 binding relative to the native myoglobin. Large decreases in the bimolecular rates of O2 and isocyanide binding were also observed for Phe6’ myoglobin; however, the association rate constant for CO binding to the Phe@ mutant was 5-fold greater than that for Iless myoglobin and only P-fold lower than that observed for the native (Vala) protein. This apparent dichotomy indicates that k’co is gov- erned by steps that are physically distinct from those limiting the bimolecular rates of O2 and isocyanide binding and that the steps for CO binding are less hindered by the benzyl side chain of Phe6’.
The dependence of k’ on ligand size is shown in Fig. 3A. The observed patterns are roughly parallel for all four proteins and show overall decreases in k’ with increasing chain length and Lu-substitution of the ligand side chain. Increasing the
n 01
III co q M
q E
0 nP
cl nn
n 02
t&i co
q M
q E
0 nP
0 nB
-2’ ’ I Ala68 ‘le68 Phc68
Mutant
FIG. 2. A, dependence of the association rate constant (k’) on the size of the Ell(68) amino acid in sperm whale myoglobin. Logarithms of the ratio of k’ for a given mutant to that for native Vales myoglohin were calculated for 02, CO, methyl (M), ethyl (E), n-propyl (nP), and n-butyl (nZ3) isocyanide binding to Alaa, He@, and Phea myoglobin. The legend in the figure reading from top (4) to bottom (nB) refers to the bars going from left (solid bar) to right (open bar) for each protein. B, dependence of the association equilibrium constant (K) on the size of the Ell(68) amino acid in sperm whale myoglobin. The legend and bar symbols are the same as those in A. In this case, each bar represents the logarithm of the ratio of K for the mutant to that of native Vale myoglohin.
size of the amino acid at position 68 exerted roughly the same effect on the kinetic barrier for each of the isocyanides. These results contrast markedly with those for unhindered chelated protoheme in soap suspensions (open circles, Fig. 3A). For this model compound, the logarithms of the association rate con- stants for isocyanide binding increase linearly with increasing surface area of the alkyl side chain due to preferential parti- tioning of the large ligands into the soap micelle (i.e. a hydrophobic effect; Olson et ~1. 1983). Substitution of Val- with Ala does not produce a completely accessible active site since k’ decreases with increasing ligand size. This is in contrast to our previous study in which substitution of His64 with Gly resulted in bimolecular rate constants for 02, CO, and isocyanide binding which were similar to those of un- hindered protoheme in soap micelles (Rohlfs et al., 1990). Thus, the distal histidine appears to cause greater inhibition of active site accessibility than the distal valine.
Dissociation Rate Constants and Equilibrium Constants fOF
VaP Mutants-The rate constant for CO dissociation was invariant with substitution at position 68 (Table I). As a result, the equilibrium association constant for CO binding is governed exclusively by the association rate constant, de- creases monotonically for the Ala6’ to Va16* to Ilea series, and then increases for Phe6’ myoglobin (Table I, Fig. 2B). Simi- larly, the rate constants for oxygen dissociation were essen- tially the same for Alam, Va16’ (native), and Ile@’ myoglobin (Table I). The slight increases in both k’o, and ko, observed for Ala6’ myoglobin resulted in no change in 02 affinity, whereas the 6-fold decrease in Ko, for Ile6’ myoglobin is due entirely to a decrease in the association rate constant (Table I). For Phe68 myoglobin, both k’o, and ko, decreased 12- and
FIG. 3. Dependence of the association rate (k’) and equilib- rium (K) constants on ligand size and stereochemistry. A, log(k’) versus ligand length (n series) and a-substitution (isopropyl and tert-butyl); B, log(K) versus ligand length and substitution. Ligand abbreviations are given in Table I. Symbols: open circles, unhindered chelated protoheme in soap micelles (taken from Mims et al., 1983); closed triangles, Ala@ myoglobin (Mb); solid circles, Vale8 (native) Mb; closed squares, Ilees Mb, open triangles, Phe@ Mb.
&fold, respectively, with the net result being only a 2.5-fold decrease in Ko, relative to that for native myoglobin. The values of all the equilibrium association constants for Phea myoglobin were consistently greater than those of the Ile@ protein, regardless of ligand size or chemistry, even though the benzyl side chain of phenylalanine is -30% larger in volume than the set-butyl side chain of isoleucine. This result suggests that rotation about the /? carbon of Phea accommo- dates bound ligands with considerably less steric hindrance than that which occurs for the P-substituted side chain of IleGS.
The isocyanide dissociation rate constants increase dra- matically for the series Ala68 to Va168 (native) to Ile68 myoglo- bin. Ala‘j8 myoglobin exhibited 5- to 50-fold decreased k values relative to native myoglobin, whereas those for the IleB8 mu- tant were 2- to 5-fold greater. These results indicate that Val% and Ilea sterically hinder the bound ligand and destabilize the iron-isocyanide bond. Substitution of Ala for Valm par- tially relieves this hindrance resulting in decreased dissocia- tion rate constants. This conclusion is supported by the equilibrium association constants which decrease -loo-fold when comparing Ala- to Ile- myoglobin. Thus, Vala in native myoglobin appears to limit access to the heme iron atom and also to hinder sterically the final bound state of the isocya- nides.
The most unexpected results of this study were the marked decreases in the isocyanide dissociation rate constants ob- served for Phe@ myoglobin compared to those for native myoglobin (Table I). The value of k for methyl isocyanide decreased 140-fold, whereas lz’ decreased only lo-fold result- ing in a 14-fold increase in the equilibrium association con- stant for methyl isocyanide binding to Phe‘j8 myoglobin. Sim- ilar results were observed for all of the isocyanides (Table I, Fig. 2B). Thus, the benzyl side chain of phenylalanine appears to adopt a conformation which stabilizes the bound states of
the isocyanides relative to those of native myoglobin, and the extent of this stabilization is roughly the same as that ob- served for the AlaM mutant (Fig. 2B). However, unlike Ala’j8 myoglobin, the kinetic barriers for ligand association and dissociation in the PheM protein are increased dramatically (Table I, Fig. 2A).
Relationship between the Equilibrium Association Constant and Ligand Size-As can be observed in Fig. 3B, K depends on both the length and extent of a-substitution of the iso- cyanide ligand and the size of the amino acid at position 68. A linear increase in log K is observed for the binding of the n-series, methyl through n-butyl isocyanide to Ala@ myoglo- bin. This increase is presumably due to favorable hydrophobic effects for the larger ligands by analogy with the results observed for Gly64 myoglobin (Rohlfs et al., 1990) and model heme in soap micelles (Fig. 3B). The 20-fold difference ob- served between the equilibrium association constants for the model heme and the AlaM mutant is due to the presence of His64 which sterically hinders all isocyanides to roughly the same extent (-2.0 kcal/mol; Rohlfs et al., 1990). As described above, the equilibrium constants for methyl and ethyl iso- cyanide binding to the Phem protein are very similar to those for AlaGS myoglobin. The benzyl side chain inhibits the bind- ing of larger ligands, particularly n-butyl and tert-butyl iso- cyanide, but in all cases, the K values for native (Val@) myoglobin are lower than those for the Phe68 mutant. The presence of either Val or Ile at position 68 restricts the size of the isocyanide which can exhibit a favorable hydrophobic effect. For example, the equilibrium association constants for n-propyl and n-butyl isocyanide binding to native and IleGs myoglobin are less than or equal to those observed for ethyl isocyanide. These latter results provide functional evidence that ValB and Ile@ restrict the size of the final equilibrium binding site, even for the smaller ligands, whereas Phe@ does not.
Rate and Equilibrium Constants for the Double Mutants, Gly64/Alu68, Gly64/Ile68, and Gly@/PheGB-Ligand binding to the double mutants was measured by stopped-flow rapid mix- ing and flash photolysis techniques. Typical results for CO binding to Glya/AlaM myoglobin are shown in Fig. 4. In the flash experiment (Fig. 4A), all of the rebinding occurred rapidly in a single phase and was complete in less than 3 ms. In contrast, only a small, slow phase was observed in the stopped-flow experiment at high CO concentrations, and most of the expected absorbance change occurred within the dead time of the mixing apparatus (-3 ms). At lower CO concen- trations (Fig. 4B), part of the rapid binding phase could be observed in the stopped-flow experiments and corresponded in rate to that observed in the flash experiments under com- parable conditions. The rates of CO binding to Gly64/Ala68 measured by photolysis depended linearly on ligand concen- tration over a lo-fold range (from 10 to 100 PM), whereas the second phase observed in the mixing experiments was 20 s-’ regardless of CO concentration. The slow phase measured in the stopped-flow apparatus suggests that the distal pocket had partially collapsed in some of the deoxymyoglobin mole- cules, whereas in the photolysis experiment, ligand rebinding was complete before this conformational transition could begin. Similar kinetic behavior was observed for Gly64/Phe68 myoglobin and to a lesser extent for Gly64/Ile68 myoglobin. A definitive explanation for the heterogenous time courses ob- served in the mixing experiments requires further study, and as a consequence, we limited our study of the double mutants to photolysis reactions with Or, CO, and methyl isocyanide.
The association rate constants presented in Table II were obtained from flash photolysis experiments and the dissocia-
FIG. 4. Time courses for CO binding to the double mutant Glye4/Alaes myoglobin. A, reaction of 3 PM Gly6’/Alaa myoglobin with 24 pM CO (after mixing) monitored at 424 nm. The Flow time course was observed in the stopped-flow spectrometer using a 2-cm path length cell, and the absorbance changes were divided by 2. The dead time of the mixing apparatus was 3 ms, and the dashed line simply connected the first observed data point to the change observed at zero time in the flash experiment. The effluent from the stopped- flow experiment was collected into a l-cm fluorescent cuvette, and then flashed with an intense, 0.5-ms light pulse. The resultant Flash time course was fitted to a single exponential function and the observed rate was 1380 s-‘. The observed rate for the Flow time course was 20 s-l. B, reaction of 3 pM Gly64/Ala68 myoglobin with 9.6 PM CO. The Flow time course exhibited two phases. The rate of the fast phase (-70% of the total absorbance change after dead time correc- tion) was 470 s-l and the rate of the slow phase was 22 s-l. Again, the dashed line is not theoretical and simply connects the first observed point with the expected total change at zero time.
TABLE II
Rate and equilibrium constants for 0% CO, and methyl isocyanide binding to position 68 (El 1) mutants of sperm whale myoglobin
containing Gly’YE7) Errors and abbreviations are given in Table I. The parameters for
native (Hi@“/Val@‘) myoglobin and the Gly64/Va168 single mutant were taken from Snrineer et al. (1989) and Rohlfs et al. (1990).
tion rate constants from ligand replacement reactions, both of which exhibited simple, monophasic kinetic behavior. The rate and equilibrium parameters for Gly64/Ala68 myoglobin
provide compelling evidence that the His64 and Va16* residues in native myoglobin are responsible for most of the resistance to ligand binding. The association rate constants for CO and methyl isocyanide binding to this protein are 100- and ZOOO- fold greater, respectively, than those for native sperm whale myoglobin and lo- and ZO-fold greater, respectively, than those for the single mutant, Gly6*/Va1@ myoglobin (Table II). Large increases in I( were also observed for CO and methyl isocyanide binding to the GlyU/Ala6’ double mutant, confirm- ing the view that this myoglobin contains a completely un- hindered distal pocket.
The association rate constant for O2 binding to Gly&/Ala@ myoglobin (1 x 10’ M-’ s-‘) is the same as that for the GUYED/ Vala protein and probably represents the diffusion controlled limit (see “Discussion”). Surprisingly, the O2 dissociation constant for Gly64/Ala68 myoglobin is lo-fold less than that of Gly64/Va168, and the equilibrium association constant for this double mutant is the same as that of native myoglobin (Table II). This suggests that the open distal pocket in Gly‘j4/ Ala6’ myoglobin is more polar due to the increased presence of water molecules and/or Arg45 which may help stabilize the polar Fe. O2 complex. A similar explanation was proposed for the lo-fold increase in Ko, for Gly‘j4 myoglobin when compared to those for the Phe64, Leu”, and VaY mutants (Springer et al., 1989; Rohlfs et al., 1990). However, the open structure also facilitates rapid autooxidation which makes kinetic meas- urements with the oxygenated double mutants very difficult (Springer et al., 1989).
The association rate constants for the binding of all three ligands to the double mutants decreased when the size of the El1 iesidue was increased from Ala- to ValGs to Ile@. These results demonstrate that Va16’ and Ile6’ inhibit the rate of ligand binding even without interference by the distal histi- dine. In contrast to the results for the single mutants at position 68 (Table I), the Gly64/Phe68 protein exhibited greater values of k’ for O2 and methyl isocyanide binding than those observed for Gly64/Ile6s myoglobin. Thus, the inhibitory ef- fects of Phe6’ on the rates of ligand association are greatly reduced when His64 is replaced with Gly. Large decreases in Kc0 and KMNC were also observed for the series Gly6’/Ala6’ to Gly64/Va168 to Gly64/Ile68. Similar, but smaller effects were observed for O2 binding. As with the single mutants at position 68, the equilibrium association constants for methyl isocyan- ide binding increase for the Gly64/Va168 to Gly64/Phe68 muta- tion and approach those for the Gly64/Ala68 protein (Table II). Thus, steric hindrance of equilibrium binding by Val@? and Ile= and stabilization of the bound ligand by the Ala6’ and Phe6’ side chains occur even in the absence of the His64 side chain.
DISCUSSION
Relative Zmportance of HiP4 and VaP in Determining Equi- librium Association Constants-Mutagenesis at positions 64 (E7) and 68 (Ell) allowed estimation of the free energy contributions of the naturally occurring, distal amino acids to the overall affinities for OS, CO, and isocyanide binding. These contributions are shown in Table III and were computed as the incremental change in binding free energy (AG = -RTlnK) when the position 64 residue was changed from His to Gly and when the position 68 residue was increased in size from Ala to Val to Ile.
Neither Hisa nor Val- in native sperm whale myoglobin appear to hinder bound oxygen sterically, presumably because the iron.Oz complex is inherently bent and readily accom- modated in the distal pocket (Fig. 1; Phillips, 1980). Substi- tution of Ala for Va16’ had no effect on the oxygen equilibrium
TABLE III Contributions of Hi.?‘(E7) and Va16s(Ell) to the &and binding free
energies of sperm whale myoglubin at 20 “CpH 7.0 The 6(AG) values were calculated as RTln[K.,,,,,/Kc~,ti],
RZ’ln[R,.,i,./R,,~,~], and RTln[R,at,vc/Kiiees]. The equilibrium associa- tion constants were taken from Table I, Springer et al. (1989), and Rohlfs et al. (1990).
Ligand His6’ --f Gly Val- + Ala Val’ + Ile
VA@ kcal/mol
% +1.4 -0.1 +o.f3 -1.0 -0.4 +1.4
MNC -2.3 -1.7 +1.4
TABLE IV Ratio of CO and OS association equilibrium constants for position 68
(Eli) mutants at 20 “C pH 7.0 The M values were calculated from the equilibrium constants listed
association constant, whereas the change from His64 to Gly produced a lo-fold decrease in K. The latter, unfavorable effect is due to the loss of a hydrogen bond between the t-amino nitrogen of His’j4 and the second bound oxygen atom (Phillips and Schoenborn, 1983; Olson et al., 1988; Springer et al., 1989; Rohlfs et al., 1990). Bound CO and methyl isocyanide are hindered by both His64 and VaY as evidenced by the negative 6(AG) values upon replacement of these resi- dues with Gly and Ala, respectively. The largest effects were observed for the mutation of His@ to Gly which stabilizes CO and methyl isocyanide binding by -0.6 kcal/mol more than that observed for the replacement of Va16* with Ala. The favorable effects of these mutations were even greater for the larger isocyanides shown in Table I.
The results in Table III show that Hisa plays the more dominant role in determining ligand equilibrium association constants. Bound oxygen is stabilized by this residue, and CO binding is hindered to a greater extent by His64 than by Va16s. As a result, the distal histidine is the residue primarily re- sponsible for the ability of native myoglobin to discriminate against CO in favor of O2 binding (Springer et al., 1989). As shown in Table IV, the ratio Kco/Ko, (M value) increases 3- fold when comparing Ala- to Vala (native) myoglobin. How- ever, replacement of His64 (native) with Gly increases the M value to a much greater extent, 20- to ‘IO-fold regardless of the amino acid at position 68 (Table IV).
The x-ray structures of 02, CO, and ethyl isocyanide myo- globin reveal that the Val@ residue is in close contact with all bound ligands (Fig. 1; Phillips, 1980; Kuriyan et al., 1986; Johnson et al., 1989). The 6(AG) values in Table III for the Ile6’ mutant provide independent, quantitative evidence of these van der Waals contacts: increasing the size of the 68 residue by one 6 methyl group hinders the binding of all ligands, including 02, by -+l.O kcal/mol.
Relative Contributions to Kinetic Barriers-The effects of reducing the size of the residues at positions 64 and 68 on the association rate constants are compared in Fig. 5A. For the
FIG, 5. Comparisons of the effects of position 64 and 68 mutations on the association rate constants for ligand binding to sperm whale myoglobin. Logarithms of the ratio of k’ for a given mutant to that for native (Hisa/Val@) myoglobin were calcu- lated for 02, CO, and methyl (M) isocyanide binding to the Gly& and Ala@ single mutants and the Gly”/AlaGs double mutant (A) and the Glyti and Ile- single mutants and the Gly64/Ile68 double mutant (B). The rate constants were taken from Tables I and II.
single mutants, the largest increases in k’ were observed for GlyG4 myoglobin; replacement of Vala with Ala caused only 1.5- to 3.0-fold increases. Interestingly, the rate enhancements for CO and methyl isocyanide binding appear to be roughly additive: substitution of Gly for His64 increased k’co and kIMNc lo- and 80-fold, respectively, and further IO-fold increases were observed when Ala replaced ValGS in the same protein (Gly”/Alam myoglobin in Fig. 5A; Table II). The latter result supports the idea of a kinetic pathway involving both the position 64 and 68 residues since reduction in the size of the 68 residue affects the rate of ligand binding even in the absence of a side chain at the distal histidine position. The lack of any further increases in k’o, for Gly64/Ala68 myoglobin indicates that the diffusion limit is reached with removal of the imidazole side chain in the single mutant containing Gly at position 64 (association rate constants for 02, NO, and isocyanide binding to unhindered model heme compounds are never greater than about 2 X lo8 M-l s-‘; Collman et al., 1983). In contrast, the effects of the single substitutions in the Gly@/ Ilem double mutant are not additive (Fig. 5B). Hindrance of the rate of O2 and CO binding by Ile68 dominates even in the absence of the His& side chain, due presumably to the 6 methyl group of Ile6’ occupying a position directly over, or very near to, the iron atom (Fig. 1). Nagai and co-workers (1987) have shown that this conformation occurs in the deoxy structures of hemoglobins containing either LY Ile-(E11)2 or p Ile-(Eli). As a result, CO and O2 are still inhibited in their approach to the heme iron atom in Ilem myoglobin even when His64 is replaced with Gly.
Ph@’ Myoglobin-Distinction between Kinetic and Equilib- rium Effects-As shown in Table I, the major effect of sub- stituting Phe for Val= is a substantial increase in the kinetic
* J. Tame, S. Fermi, and K. Nagai, unpublished x-ray studies quoted by Mathews et al. (1989).
FIG. 6. Effects of E7 and El 1 mu- 2 tations on the association rate con- stants of sperm whale myoglobin and the subunits of R-state human hemoglobin at 20 “C, pH 7. Loga- i 1 rithms of the ratio of k’ for a given g mutant and to that of native myoglobin 2 were calculated for: A, sperm whale myo- L globin; B, R-state a subunits; C, R-state 4 ’ fl subunits. Gly-E7 represents the B His(E7) to Gly mutation; Ala-Eli, the Val(El1) to Ala mutation; and Ile-Eli,
j,
the Val(E11) to Ile mutation. The data X -1 2
for hemoglobin subunits were taken from Mathews et al. (1989) and that for the His(E7) to Gly mutant of myoglobin from Snrinser et al. (1989) and Rohlfs et
barrier for O2 and isocyanide association and dissociation. Little change is observed for the O2 equilibrium association constant, and K for isocyanide binding actually increases -lO- fold although the side chain of Phe is roughly twice as large as that of Val. In contrast, the association and dissociation rate constants for CO binding to the Phe6’ mutant are essen- tially the same as those for native myoglobin (Table I, Fig. 2). Frauenfelder et al. (Doster et al., 1983) and Gibson et al. (1986) have shown that the CO association rate constant is determined by the product of the equilibrium constant for ligand migration into the distal pocket and the rate of bond formation with the iron atom. Gibson et al. (1986) have also shown that the rate of CO dissociation from most heme proteins is determined solely by the rate of bond disruption and is little affected by the rates of ligand migration away from the iron atom and out of the protein matrix. Thus, a conformation of Phe6’ which inhibits ligand entry into the distal pocket, but exerts little influence on the bound state, would not be expected to affect k’cc and kce. In contrast, the overall association and dissociation rate constants for O2 and isocyanide binding should decrease for such a conformation since these parameters are governed by the rates of ligand movement through the protein.
of geminate recombination from within the distal pocket (see Gibson et al., 1986).
Preliminary x-ray crystallographic data for Phe- metmy- oglobin show that the benzyl group is pointed toward the back of the heme pocket with the first phenyl carbon atom occu- pying a position analogous to that of the yl methyl carbon atom of Va16* in native myoglobin (toward the viewer in Fig. 1).3 In difference Fourier maps, a large negative peak was found where the Va16’ y2 methyl group was located in native metmyoglobin. The absence of a position 68 carbon atom near the iron probably accounts for the high affinities observed for isocyanide binding to the Phe6* protein. Similar relief of steric hindrance between the y2 methyl group and bound isocyanides occurs in the Alaa mutant which also shows a lo-fold increase in K for isocyanide binding compared to native myoglobin (Table I). Unfortunately, a clear explanation as to why the rates of O2 and isocyanide association and dissociation de- crease markedly for Phe@ myoglobin is not evident from these initial crystallographic results. Further refinement of the met- myoglobin structure and determinations of the high resolution structures of the deoxygenated and liganded forms of ferrous Phe@ myoglobin will be necessary for a complete interpreta- tion of the unusual kinetic properties of this mutant. Nano- second and picosecond laser photolysis experiments are also being carried out in order to measure more directly the rates of ligand migration into and out of the protein and the rates
Comparisons with Human Hemoglobin-Comparisons of the effects of active site mutations on the association rate constants for 02, CO, and methyl isocyanide binding to myo- globin and hemoglobin are shown in Fig. 6. The ligand binding site in R-state /3 subunits is less hindered than that in either a subunits or myoglobin and is easily accessible to diatomic gases as judged by the large association rate constants of the native /3 subunit (i.e. k’ O2 (/3) = 10 X lo7 M-’ s-‘, k’ O2 (a) =
2.8 X lo7 M-’ s-l, and k’ O2 (Mb) = 1.4 X lo7 M-’ s-l) and by the lack of effect of the His(E7) to Gly and Val(E11) to Ala substitutions (Fig. 6C; Mathews et al., 1989). These mutations also have little effect on the equilibrium constants for ligand binding to p subunits, and thus, bound O2 in this subunit must be stabilized by some mechanism other than hydrogen bonding to the E7 side chain (Mathews et aZ., 1989). Curiously, the R-state /I subunits are markedly affected by substitution of Ile for Val(Ell), whereas the a subunits are not and myoglobin shows intermediate effects. In both R-state (Y sub- units and myoglobin, bound O2 is stabilized by hydrogen bonding (-1 kcal/mol) as predicted by the close proximity of the t-amino nitrogen of His(E7) to the O(2) oxygen atom (2.6 A in Mh and 2.8 A in a subunits versus 3.5 A in /l subunits; Shaanan, 1983), and this orientation of the bound ligand appears to be fixed by the position of the isopropyl side chain of Val(E1l) (Fig. 1). As a result, the favorable hydrogen bonding interaction is achieved at the expense of increased steric crowding in the distal pocket of these proteins, limiting access to the ligand-binding site. This restriction is mani- fested as lower association rate constants for all ligands and substantial increases in these kinetic parameters when smaller residues are substituted for His(E7) and Val(E1l) in R-state ar subunits and sperm whale myoglobin (Fig. 6, A and B).
REFERENCES
Bolognesi, M., Cannillo, E., Ascenzi, P., Giacometti, G. M., Merli, A., and Brunori, M. (1982) J. Mol. Bid. 158,305-315
Case, D., and Karplus, M. (1979) J. Mol. Biol. 132, 343-368 Collman. J. P., Brauman, J. I., Iverson, B. L., Sessler, J. L., Morris,
R. M.; and Gibson, Q.‘H. (i983) J. Am. Chem. Sot. 106, 3052- 3064
Creighton, T. E. (1983) Proteins: Structures and Molecular Properties, p. 242, W. H. Freeman and Co., New York
Dickerson, R. E., and Geis, I. (1983) Hemoglobin: Structure, Function, Evolution and Pathology, Benjamin/Cummings, Menlo Park, CA
Doster. W.. Beece. D.. Bowne, S. F., DiIorio, E. E., Eisenstein, L., Frauenfelder, H.; Reinische, L., Shyamsunder, E., Winterhalter, K. H.. and Yue. K. T. (1982) Biochemistrv 21.4831-4839
3 G. N. Phillips, R. M. Arduini, and K. A. Johnson, unpublished Gibson, Q. H.,‘Olson, J. S., McKinnie, R: E., and Rohlfs, R. J. (1986) results. J. Biol. Chem 261,10228-10239
Hattori, M., and Sakaki, Y. (1986) And. Biochem. 152, 232-238 Johnson, K. A., Olson, J. S., and Phillips, G. N. (1989) J. Mol. Biol.
207,459-463 Kottalam, J., and Case, D. A. (1988) J. Am. Chem. Sot. 110, 7690-
7697 Kuriyan, J., Wilz, S., Karplus, M., and Petsko, G. A. (1986) J. Mol.
Biol. 192,133-154 Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Motefutar
Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
Mathews, A. J., Rohlfs, R. J., Olson, J. S., Tame, J., Renaud, J.-P., and Nagai, K. (1989) J. Bid. Chem. 264,16573-16583
Mims, M. P., Porras, A. G., Olson, J. S., Noble, R. W., and Peterson, J. A. (1983) J. Biol. C/rem. 258,14219-14232
Nagai, K., Luisi, B., Shih, K., Miyazaki, G., Imai, K., Poyart, C., De Young, A., Kwiatkowskv, L., Noble, R. W., Lin. S.-H., and Yu, N. T. (1987) Nature 329, &3-$60
Olson. J. S.. McKinnie. R. E.. Mims. M. P.. and White. D. K. (1983) J. Am. hem. Sot. lb5,1522-1527 ’
Olson, J. S., Mathews, A. J., Rohlfs, R. J., Springer, B. A., Egeberg,
K. D., Sligar, S. G., Tame, J., Renaud, J.-P., and Nagai, K. (1988) Nature 336, 265-266
Peru& M. F. (1970) Nature 228, 726-739 Per&z, M. F. (1989) Trends Biochem. Sci. 14, 42-44 Phillips, S. E. V. (1986) J. Mol. Biol. 143,531-554 Phillips, S. E. V. (1981) The X-ray Structure of Deoxy-Mb (pH 8.5)
at 1.4-A Resolution, Brookhaven Protein Data Bank Phillips, S. E. V., and Schoenborn, B. P. (1981) Nature 292,81-82 Phillips, G. N., Arduini, R. M., Springer, B. A., and Sligar, S. G.
(1990) Proteins, in press Ringe, D., Petsko, G. A., Kerr, D. E., and Ortiz de Montellano, P. R.
(1984) Biochemistry 23,2-4 Rohlfs, R. J., Mathews, A. J., Carver, T. E., Olson, J. S., Springer, B.
A., Egeberg, K. D., and Sligar, S. G. (1990) J. Biol. Chem. 266, 3168-3176
Shaanan, B. (1983) J. Mol. Biol. 171,31-59 Springer, B. A., and Sligar, S. G. (1987) Proc. Natl. Acad. Sci. U. S.
A. 84,8961-8965 Springer, B. A., Egeberg, K. D., Sligar, S. G., Rohlfs, R. J., Mathews,
A. J., and Olson, J. S. (1989) J. Bid. Chem. 264, 3057-3060
