His73, Often Methylated, Is an Important Structural Determinant for Actin

His73, has been proposed to regulate the release of Pi from the interior of actin following polymerization-dependent hydrolysis of bound ATP. Although it is a 3-methylhistidine in the vast majority of actins, His73 is unmethylated in S. cerevisiae actin. We mutated His73 in yeast actin to Arg, Lys, Ala, Gln, and Glu and detected no altered phenotypes associated with the mutationsin vivo. However, they significantly affect actin functionin vitro. Substitution of the more basic residues resulted in enhanced thermal stability, decreased rate of nucleotide exchange, and decreased susceptibility to controlled proteolysis relative to wild-type actin. The opposite effects are observed with the neutral and anionic substitutions. All mutations reduced the rate of polymerization. Molecular dynamics simulations predict a new conformation for the His73 imidazole in the absence of a methyl group. It also predicts that Arg73 tightens and stabilizes the actin and that Glu73 causes a rearrangement of the bottom of actin's interdomain cleft leading possibly to our observed destabilization of actin. Considering the exterior location of His73, this work indicates a surprisingly important role for the residue as a major structural determinant of actin and provides a clue to the impact caused by methylation of His73.

His 73 , has been proposed to regulate the release of P i from the interior of actin following polymerization-dependent hydrolysis of bound ATP. Although it is a 3-methylhistidine in the vast majority of actins, His 73 is unmethylated in S. cerevisiae actin. We mutated His 73 in yeast actin to Arg, Lys, Ala, Gln, and Glu and detected no altered phenotypes associated with the mutations in vivo. However, they significantly affect actin function in vitro. Substitution of the more basic residues resulted in enhanced thermal stability, decreased rate of nucleotide exchange, and decreased susceptibility to controlled proteolysis relative to wild-type actin. The opposite effects are observed with the neutral and anionic substitutions. All mutations reduced the rate of polymerization. Molecular dynamics simulations predict a new conformation for the His 73 imidazole in the absence of a methyl group. It also predicts that Arg 73 tightens and stabilizes the actin and that Glu 73 causes a rearrangement of the bottom of actin's interdomain cleft leading possibly to our observed destabilization of actin. Considering the exterior location of His 73 , this work indicates a surprisingly important role for the residue as a major structural determinant of actin and provides a clue to the impact caused by methylation of His 73 .
His 73 is conserved in all actins studied to date, and in the vast majority of actins, this residue is post-translationally methylated to produce N--3-methylhistidine. The only exception to this rule had been the actin from Naegleria gruberii (1). However, we recently demonstrated that the actins from the yeasts Saccharomyces cerevisiae and Candida albicans were also nonmethylated (2). This residue lies on the outside of the protein in a cleft between subdomains 2 and 4 as part of a section of elongated polypeptide that extends downward from the top of subdomain 2 (Fig. 1). Based on the structure of skeletal muscle ␣-actin and nonmuscle ␤-actin which have methylhistidines, His 73 is adjacent to a glutamic acid at position 72 and lies across the cleft from Asp 184 in subdomain 4 (3)(4)(5). The conservation of a positively charged residue at this position suggests it might be necessary to offset the high neg-ative charge density in the vicinity which otherwise might disrupt actin structure. Previously, we demonstrated that mutation of His 73 in muscle actin would not prevent the incorporation of radiolabeled mutant monomers into wild-type actin filaments in vitro (6). However, in this case, it was entirely possible that the constraints imposed upon the mutant monomer by the WT 1 monomers in the filament masked any effects that the mutation would have caused in a pure mutant filament.
Based on molecular dynamics simulations, Wriggers and Schulten (7) have proposed that the imidazole moiety of this histidine residue blocks the opening of a channel by which water can gain access into the interior of the actin molecule where the bound nucleotide resides (7). This imidazole is over 6.5 Å away from the ␥-phosphate of the bound ATP and is thus too far away to directly interact with the phosphate. However, these investigators subsequently proposed that following scission of the ␤␥-phosphate bond during polymerization, movement of the P i away from the ADP toward the outside of the channel would allow it to interact with a positively charged imidazole thereby stabilizing the phosphate's negative charge. The interaction between the two moieties, promoted by protonation of the imidazole, would retard the release of the phosphate from the actin (8). Since the state of the nucleotide in the filament determines filament stability (9 -11), this phosphate retention hypothesis, if correct, suggests that His 73 might be an important factor in controlling cytoskeletal dynamics.
To investigate the effect of His 73 on actin function, we have mutated this residue in S. cerevisiae actin. Substitution with Arg and Lys maintains a positive charge at this position, substitution with Ala neutralizes the charge. Replacement with Glu introduces a negative charge in place of the original positive charge, and finally substitution of Gln gives a neutral substitution isosteric with Glu. We have isolated yeast cells expressing each of these as the only actin in the cell and have assessed the effects of these mutations in vivo. Prior to assessing the mutations' effects on ATP hydrolysis and P i release in terms of the phosphate-release model, it was necessary to determine the effects of the mutations on actin monomer integrity and polymerizability. We have thus isolated the actins and characterized the effects of the mutations on monomer stability, tertiary structure, nucleotide exchange, and the ability to form filaments at different temperatures. The results of these studies are reported here.

MATERIALS AND METHODS
The site-directed mutagenesis kit was purchased from Stratagene Corp. Oligodeoxynucleotides used for site-directed mutagenesis were synthesized in the DNA Core Facility at the University of Iowa. The Sequenase Version 2.0 DNA sequencing kit was purchased from Sigma. DNase I was obtained from Worthington. Affi-Gel 10 was obtained from Bio-Rad. Etheno-ATP, ␣-chymotrypsin, trypsin, and subtilisin were purchased from Sigma.
Oligonucleotide-directed Mutagenesis-Site-directed mutagenesis was used to construct a mutant actin sequence carried in a centromeric vector pRS314 (12) marked with the TRP1 gene. The oligodeoxynucleotide 5Ј-GG-TTACCCAATTGAAC(G/A)A(C/G)A(T)GGTATTGTCACCAACTGG-3Ј was used to generate the mutant actins in which the codon for His(CAC) was mutated to that for Ala(GCT), Gln(CAA), Glu(GAA), Lys(AAA), and Arg(CGT). The underlined sequences are the mutated codons. The DNA was sequenced in each case to verify the desired mutation.
Generation of Cells Producing Mutant Actins-pRS314 plasmids containing the mutant coding sequences were introduced into a trp1,ura3-52 haploid cell in which the chromosomal ACT1 gene had been disrupted by replacement of the coding sequence with the LEU2 gene. Wild-type actin was expressed in these recipient cells from another centromeric plasmid containing the URA3 gene. Following transformation with the mutant plasmid and selection on tryptophan-deficient medium, surviving cells were subjected to plasmid shuffling to eliminate the plasmid carrying the WT actin gene. The mutant plasmid was rescued from surviving trp ϩ ,ura Ϫ cells and sequenced to ensure that the mutation was still intact. Viable cells were readily obtained for all mutants.
Purification of Yeast Wild-type and His 73 Mutant Actins-Wild-type and His 73 mutant actins were purified using a combination of DNase I affinity chromatography, DEAE cellulose chromatography, and subsequent polymerization-depolymerization cycling as described previously (13). Calcium/G-actin was stored in calcium/G-buffer (10 mM Tris-HCl, pH 7.5, 0.2 mM CaCl 2 , 0.2 mM ATP, and 0.5 mM dithiothreitol). Calcium/ actin was converted to the magnesium form by treatment with 0.2 mM EGTA in the presence of 0.1 mM MgCl 2 at 25°C for 10 min as described by Chen and Rubenstein (14) and Pollard (15). Magnesium/G-actin was used immediately after the conversion from calcium/G-actin. Actins were stored at 4°C following purification and used within 4 days.
Actin Polymerization-Actin polymerization was assessed by the increase in light scattering that occurred following the addition of 2 mM MgCl 2 and 50 mM KCl to a G-actin solution in the thermostatted cuvette chamber of a Spex Fluorolog 3 fluorescence spectrometer. Excitation and emission wavelengths were set at 360 nm. For all experiments, wild-type actin was run as a control, and experiments were repeated with at least three different batches of actins. To assess cold sensitivity, actin was polymerized at 25°C until a plateau was reached. The temperature was then lowered to 4°C over a 15-30-min period of time and subsequently maintained at this temperature during which the change in light scattering was monitored.
Thermal Denaturation-The apparent melting temperatures of G-actins were determined by circular dichroism according to Chen and Rubenstein (14). 1.4 M G-actin was heated at a constant rate of 1°C/min over a range from 20 to 80°C with constant stirring of the samples over the entire range tested. Changes in the ellipticity of actin samples were monitored at 222 nm in an AVIV 62 DS spectropolarimeter. Data were fitted to a two-state model with a single transition between a native and a denatured form of the protein and the T m value was defined as the temperature when 50% of the G-actin was in the denatured form.
Nucleotide Exchange-Unbound ATP was removed from a 5 M Gactin solution by gel filtration, and the actin was incubated with 0.3 mM etheno-ATP at 4°C for 2 h. Excess nucleotide was removed by gel filtration. The subsequent actin solution was mixed with 0.1 mM ATP and the decay in fluorescence that accompanied release of the etheno nucleotide from the actin was monitored as a function of time with excitation and emission wavelengths at 340 and 410 nm, respectively. For rapid reactions, a stopped-flow fluorescence spectrometer was employed. Nucleotide exchange rates were derived by fitting the data to a single exponential expression.
Electron Microscopy-Actin (5 M) polymerized at 25°C was applied to carbon-coated Formvar grids and visualized following negaitve staining with 1.5% (w/v) uranyl acetate using a Hitachi 7000 electron microscope (University of Iowa Electron Microscope Facility).
Limited Proteolysis of G-actin-Alterations in the tertiary structure of G-actins were monitored by the susceptibility of the actins (9 M) to limited digestion by the following three enzymes at the indicated actin/ protease ratios (w/w): trypsin (80:1), subtilisin (400:1), and chymotrypsin (16:1) essentially as described previously (16). Incubations were carried out at room temperature for the desired time. The digestion with trypsin was stopped with 10 g/ml trypsin inhibitor, and the digestion with the other two proteases was stopped with 1 mM phenylmethylsulfonyl fluoride. The samples were then separated by electrophoresis on a 12.5% SDS-polyacrylamide gel and visualized by staining with Coomassie Blue.
Molecular Modeling Studies-Molecular dynamics simulations of actin were carried out using the program X-PLOR (17) with the CHARMM-22 all-hydrogen force field (18). The structure of ATP-actin taken as the initial model was from the actin-gelsolin segment-1 complex (5), prepared with magnesium as divalent cation and solvated with 1,159 explicit water molecules as described in Ref. 8. The originally methylated MeHis 73 (5) was replaced by unmethylated histidine in three different protonation states (proton on the ␦1 nitrogen or on the ⑀2 nitrogen or on both) for the control simulations, and by charged arginine and glutamate for the H73R and H73E mutants, respectively. The systems were subsequently refined by simulated annealing protocols. Simulated annealing is a conformational search technique that allows biomolecules to pass barriers between local minima in the complex energy landscape of proteins, faster than would be realized by simulations at room temperature. The method, thereby, accelerates the relaxation of the system from any less favorable interactions created by the mutation of His 73 . Simulated annealing protocols were successfully employed in earlier computational studies of kinesin (19) and apolipoprotein A-1 (20). Here we chose protocols of 170-ps length with a maximum temperature of 500 K.

RESULTS
Mutagenesis of His 73 -We wished to address the question of how His 73 affects the stability, tertiary structure, nucleotide exchange, and polymerization of actin. For this purpose, we used site-directed mutagenesis to substitute His 73 with Arg, Lys, Ala, Gln, and Glu and replaced the WT actin gene with the mutant actin genes in S. cerevisiae as described under "Materials and Methods." Viable haploid yeast cells producing only the desired mutant actins were readily obtained for each of the substitutions, and the presence of the desired mutation was confirmed by rescuing the plasmid from yeast and resequencing the DNA.
Effect of the His 73 Mutant Actins on Yeast in Vivo-To determine any defects caused by His 73 mutations in vivo, we first determined the growth rates of cells expressing each of the mutant actins. In liquid YPD medium at 10, 30, and 37°C no appreciable difference was found between the generation times of each of the mutant cells and wild-type cells at the temperature tested (data not shown). Phase-contrast microscopy revealed no overt difference in either the size or morphology of the mutant cells compared with their wild-type counterpart. Calcofluor staining (21) showed that all the mutant cells displayed well formed chitin rings at the sites of previous budding Shown is a polypeptide backbone three-dimensional structure of skeletal muscle actin with specific residues depicted using a space-filling mode. The figure was generated using the program Rasmol. events and the axial budding polarity typical of yeast (data not shown). To visualize the actin cytoskeleton directly, we stained wild-type and mutant cells with rhodamine/phalloidin (22). Again, there appeared to be no distinct difference in the phalloidin/staining patterns exhibited by mutant and WT cells in terms of the presence and distribution of cables and cortical patches.
Effects of His 73 Mutations on the Kinetics of Actin Polymerization at 25°C-In order to interpret results obtained from phosphate release assays, we first had to assess the effects of these mutations on actin monomer structure and polymerizability. Fig. 2 shows the polymerization of the calcium form of each of the mutant and WT actins induced by the addition of salt to the appropriate G-actin solutions as described under "Materials and Methods." At 25°C, each of the mutations had a noticeable effect on the rate of polymerization. The cationic substitutions, Lys and Arg, produced the mildest retardation in rates followed by Ala and Gln. The Glu mutation appeared to cause the most extensive retardation in polymerization.
Despite the slower polymerization of the mutant actins compared with WT actin, the extent of polymerization for all the actins except for H73E was essentially the same, suggesting that the critical concentrations of the mutant and wild-type actins are roughly identical. H73E might either have a higher critical concentration than wild-type or is less stable since the extent of polymerization of this mutant varies on a preparation to preparation basis between about 60 and 100%. Results with the magnesium form of the mutants showed the same relative order of effects, although in each case, the magnesium form of these actins polymerized faster than the corresponding calcium form. This has long been known to occur with WT actin. Finally, negatively stained polymerized samples of the Arg, Ala, Gln, and Glu mutations yielded filaments under the electron microscope with a morphology similar to that observed with WT actin.
To determine the effects of the His 73 mutations on the conformation of subdomain 2 of G-actin, we carried out limited proteolysis of WT and His 73 mutant actins as described under "Materials and Methods." Fig. 3 shows that although controlled digestion of G-WT and G-His 73 mutant actins by trypsin, subtilisin, and ␣-chymotrypsin produced the same digestion patterns, the extent of digestion of H73K actin was less than that of its WT counterpart. In contrast, under the same conditions, the proteolysis of H73E actin appeared to be more extensive than that of its WT counterpart. These results indicate that the digestion sites of these proteases in subdomain 2 of H73K mutant actin are more protected and those of H73E mutant actin are more accessible than those of WT actin. H73R showed an even more protected pattern than H73K (data not shown). Thus, cationic substitutions appeared to confer greater resistance to proteolysis than that observed with WT actin and substitution of Glu for His 73 greatly increased susceptibility. Substitution of the neutral residue Ala produced a susceptibility roughly equivalent to that of WT actin. These results were observed with both the magnesium and calcium forms of actin although in all cases, the magnesium/G form is less susceptible than the calcium/G form, consistent with previous results. stopped-flow fluorescence spectrometer (Fig. 4B). Exchange data for the mutants were fit to a first-order decay plot, and the t1 ⁄2 values were calculated. Our results, shown in Table I, reveal that there is a hierarchy of effects of the different mutations on nucleotide exchange that correlates with the differences seen in protease susceptibility. The Arg substitution produces the longest delay compared with WT actin (almost ϫ20). Lys slows the exchange but to a much lesser extent than does Arg. Ala is slightly faster than WT as is Gln. Glu greatly accelerates the rate of exchange by about 14 times compared with WT actin. Thus, mutations which seem to produce a tighter, more compact monomeric structure also result in a retarded rate of nucleotide exchange while those which appear to open the structure accelerate the rate of exchange.
Thermal Stability of His 73 Mutant Actins-Our results so far have suggested that the basic residues result in a tighter, more compact monomer while the neutral and acid residues lead to a looser structure with greater conformational flexibility. To assess this situation further, we carried out circular dichroism experiments to determine the overall thermal stability of the His 73 mutant actins. The molar ellipticity at 222 nm, which reflects the content of ␣-helical structure in G-actin, was mon-itored as a function of temperature. The results are shown in Fig. 5. Based on the total change in dichroism, the degree of denaturation at each temperature was calculated and the data replotted as the fraction native versus temperature. The relative 50% denaturation value is defined as the apparent melting temperature (T m ) for the protein, and this value for each of the mutants is shown in Table II. The altered T m value of all the mutants suggest that the nature of the substitution has a direct effect on the thermal stability of the actin monomer, even though residue 73 is on the exterior of the protein. Strongly cationic substitutions increase the stability of the protein over that of the WT protein whereas neutral substitutions cause slightly decreased stability. Substitution of glutamate greatly

values of ATP exchange of His 73 mutants
The exchange data for actins in the calcium form were fitted to a first-order decay plot and the t 1/2 values calculated. Polymerization of Some His 73 Mutant Actins Is Cold-sensitive-One major interstrand force stabilizing the actin filament might be a "plug pocket" arrangement in which a hydrophobic plug at the tip of a loop between subdomain 3 and 4 interacts with a hydrophobic pocket formed by the interface of two subunits on the opposing strand of the helix (27,28). Our previous work has shown that mutations in the hydrophobic plug produce effects consistent with a weakening of this proposed plugpocket interaction and lead to disassembly of filaments at 4°C (12,29). Although residue 73 is not directly involved in the proposed interaction, it is on a strand of the polypeptide directly attached to the part of subdomain 2 hypothesized to be part of the "pocket." If alteration of the conformation around His 73 led to an altered conformation in these hypothetical pocket residues, the His 73 mutations might produce a "cold sensitivity" like that observed with the hydrophobic plug mutations.
Wild-type and mutant actins were polymerized at 25°C. The temperature was then lowered to 4°C, and the light scattering was monitored over time. Wild-type actin filaments were stable at 4°C, as well as were the H73R and Lys filaments (data not shown). In contrast, the three other mutations H73A, Gln, and Glu resulted in actins that depolymerized at 4°C (Fig. 6A). Glu always depolymerized completely while the Ala and Gln mutants showed about 80 -90% depolymerization depending on the experiment. To determine the reversibility of this cold sensitivity, the temperature was returned to 25°C in each of the cases, and polymerization was restored to almost the original level (Fig. 6B). Thus, cold-induced depolymerization did not result in irreversible changes in actin structure.
Molecular Modeling of WT, H73R, and H73E Actins-We next performed molecular modeling simulations on WT, H73R, and H73E mutant actins to determine if the absence of a methyl group on His 73 of yeast actin caused a realignment of residues at the bottom of actin's interdomain cleft. Also, we wished to better understand the molecular basis for the changes in actin behavior caused by these two His 73 mutations.
The methyl group on the ⑀2 nitrogen of MeHis 73 , present in the crystal structure of muscle actin, was removed to simulate the WT state of yeast actin. The nonionized form of the imidazole ring can exist as two tautomers, with the proton on either the ␦1 or ⑀2 nitrogens (30). The hydrogen is usually found on the ⑀2 nitrogen, but the pK a of the ␦1 atom is only about 0.6 pH unit lower than that of ⑀2. Thus, the position of the proton depends on the conditions of the local environment in the protein, and both forms may exist. Furthermore, histidine is readily protonated with a pK a near 7, in which case the positive charge is shared by the two nitrogens by resonance. We have simulated all three charge states (proton on either atom or on both) starting with the crystal structure of the imidazole for a realistic modeling of the WT yeast actin. Fig. 7, A-D, compares the three charge states of His 73 to the position of MeHis 73 in the crystal structure (5). In the cases where the ␦1 nitrogen is protonated, the conformation of His 73 is close to that of MeHis 73 in the crystal structure regardless of the protonation state of ⑀2. The imidazole heavy atoms exhibit a root mean square deviation of only 0.7 and 0.9 Å from corresponding atoms in MeHis 73 , for protonated and unprotonated ⑀2, respectively. The relative stability of these two cases is due to a ␦1-proton-stabilizing hydrogen bond with the backbone oxygen of Gly 158 (hydrogen acceptor distances: 1.8 and 1.9 Å, donor hydrogen acceptor angles: 25 and 8°, for protonated and unprotonated ⑀2, respectively). In contrast, if ⑀2 alone is proto-   6. Panel A, cold sensitivity assay of His 73 mutant actin polymerization. H73A (curve 1), Gln (curve 2), and Glu (curve 3) mutant actins (4 M) were polymerized at 25°C. The temperature was then lowered to 4°C, and the light scattering was monitored over time. Panel B, following cold-induced depolymerization, the temperature was returned to 25°C, and the increased in light scattering due to repolymerization was monitored. The curves are numbered as described in panel A. nated, the stabilizing hydrogen bond no longer exists, and the side chain flips about its long axis so that the proton on ⑀2 is oriented toward the carboxylate of nearby Asp 184 across the cleft separating the two domains of the protein. In this case, the imidazole exhibits a large root mean square deviation from Me-His 73 (1.8 Å), and a salt bridge is formed with Asp 184 (hydrogen acceptor distance: 1.9 Å, donor hydrogen acceptor angle: 11°).
The conformation of residue 73 in the two simulated mutants, Arg 73 and Glu 73 , are depicted in Fig. 7, E and F, respectively. The positively charged side chain of Arg 73 coordinates both carboxylate oxygens of Asp 184 (hydrogen acceptor distances: 2.4 and 3.1 Å). On the other hand, Glu 73 does not form a salt bridge with Asp 184 but with Arg 183 (hydrogen acceptor distance: 1.7 Å, donor hydrogen acceptor angle: 16°). Residues 73, 183, and 184 are exposed to adjacent actin monomers in the WT filament (28), and their rearrangement in the Glu 73 mutant frees the negatively charged Asp 184 while neutralizing the positive charge on Arg 183 . DISCUSSION Our results strongly suggest a role for this residue that is much more complex than simply controlling the rate of P i release from actin as had been originally proposed. His 73 appears to be a major determinant of stability and conformational flexibility of the actin monomer despite the fact that it is located on the protein exterior where it can interact with solvent molecules.
First, our simulations of actin with His 73 , as occurs in yeast, instead of the more common MeHis 73 ,predicted the possibility of a second conformational state for this residue and provide a potential clue as to the function of the methyl group found on His 73 in the vast majority of actins. This new state, expected to be pH and environment dependent, is stabilized by a salt bridge of the ⑀2 proton with Asp 184 . In contrast, the MeHis 73like state involves a weaker hydrogen bond between the ␦1 proton and the Gly 158 oxygen. The salt bridge with Asp 184 would significantly stabilize the protonation of the ⑀2 and even that of the ␦1 nitrogen. In T4 lysozyme, a similar salt bridge between residues Asp 70 and His 31 altered the pK a values of these two residues from the normal values of 3.5 and 7 to 0.5 and 9, respectively (31) contributing 3-5 kcal/mol to the free energy of folding of the protein. Since we started with the crystal conformation that is stabilized by the weaker hydrogen bond, a transition to the salt bridge with Asp 184 was not observed in this work for doubly protonated His 73 . However, a tunneling through the singly ⑀2 protonated state would lower the barrier for the transition from the MeHis 73 -like conformation to the salt bridge conformation also for doubly protonated His 73 . This scenario is likely because ⑀2 protonation, that is, the state leading to salt bridge formation, is favored over ␦1 protonation by the pK a difference of the two nitrogens.
We have demonstrated that substitution at this residue with amino acids ranging from more basic to neutral and finally to an acidic residue produces a gradient of effects on actin structure and function. The magnitude and direction of these effects are consistent with the hypothesis that the imidazole of His 73 may interact electrostatically with nearby acidic residues, in particular Asp 184 , on the opposite side of the interdomain cleft in actin to determine the range of movement of the two domains of actin relative to one another. Substitution of His 73 with strong cationic residues Arg and Lys seemed to result in a monomer with a tighter conformation as evidenced by a decreased nucleotide exchange rate, decreased rate of proteolysis of subdomain 2, and increased thermal stabilization. A plausible hypothesis is that the stronger positive charge allowed a more stable interaction with Asp 184 leading to a less flexible more protected interior of the monomer. The marked difference in effects of the Arg versus the Lys substitution might be due to the higher pK a of Arg coupled with the lower flexibility of the protein's side chain allow a stronger interaction to occur with Asp 184 . The opposite effects were observed with H73E. These results suggest that charge repulsion with one of the acidic residues and subsequent rearrangement of amino acids at the base of the cleft results in a more open and flexible conformation responsible for the greatly enhanced rate of nucleotide exchange, greater protease susceptibility, and lower thermal stability.
The results of molecular modeling studies with the Arg and Glu mutants were consistent with the idea of an altered interaction involving Asp 184 . Substitution of Arg produced a structure with a much stronger level of interaction predicted between the guanidinium group and Asp 184 in the second state than would occur with His 73 . This would lead to the hyperstabilization of the protein we observed experimentally. Substitution of Glu at residue 73 resulted in formation of a predicted salt-bridge between Glu 73 and Arg 183 with the concomitant breaking of a number of hydrogen bonds due to the resulting distortion of the tertiary structure of the protein. Interactions between residue 73 and 183 or 184 are likely to affect the stability of the actin fold since they link actin subdomains across the nucleotide binding cleft (3)(4)(5). In this context, the salt bridges formed by the Arg 73 and Glu 73 mutants are remarkable. The stronger bond formed by Arg 73 would hyperstabilize contacts around the nucleotide leading to the effects we observed experimentally. Conversely, the Glu 73 -Arg 183 salt bridge would lead to a shift of the electrostatic charge on the surface of actin and alter the binding face and binding affinity to the adjacent monomer in the filament (28). Such a movement might easily alter the ATP-binding site leading to the enhanced nucleotide exchange kinetics, thermal instability, and the altered protease susceptibility we observed.
The observed polymerization kinetic effects with little if any change in critical concentration are consistent with our proposal of an electrostatic clamp near the base of the interdomain cleft. G-actin must undergo a conformational change to the F-monomer conformation before it can become incorporated into a growing filament (32). An increased interaction of residue 73 with Asp 184 might result in the retardation of the rate at which this change occurs leading to slower polymerization. Elimination of the positive charge might impart enough increased conformational flexibility to make it easier for the protein to sample additional conformational states thereby retarding attainment of the F-monomer conformation. The rearrangement at the base of the cleft predicted with the Glu mutation might alter the rate at which the protein enters the F-monomer state resulting in even slower polymerization. However, the numerous inter-monomer contacts that develop during filament assembly might compensate for these altered monomer conformations with little net effect on critical concentration.
The cold-sensitive polymerization of the Ala, Glu, and Gln mutations but not the Arg and Lys species is intriguing. According to the hydrophobic plug-pocket proposal for filament stabilization, residues 63 and 64 along with residues 40 -45 would constitute part of the hydrophobic surface that comprises the pocket in this interaction. Hydrophobic interactions are cold-sensitive, and we had previously demonstrated that mutations in plug residues resulted in cold-sensitive polymerization behavior. Neutralization of the positive charge at residue 73 produced an increased protease susceptibility in this region and moderate cold-sensitive behavior which was accentuated by substitution of an anionic residue. Altering the geometry of residues at the bottom of the interdomain cleft in actin likely results in a change of the surface at the top of subdomain 2. Such an alteration would cause a distortion in the pocket proposed by Holmes et al. (33), thereby decreasing the energy of the proposed hydrophobic interaction resulting in cold sensitivity. The smaller effect observed with the His 73 mutations as opposed to that seen with the plug mutants may simply reflect the change in total hydrophobic surface that a single mutation in the plug makes relative to one in the larger pocket.
A second possible explanation is that His 73 mutations alter the conformation of the plug in the same monomer by conformational changes propagated through the nucleotide bridge or interdomain cleft. However, the relatively large distance of the His 73 imidazole from the ATP ␥-phosphate coupled with the cold sensitivity of the neutral mutations despite their relatively small effect on monomer structure argues against this second possibility. In either case, the lack of a clear defective phenotype in vivo most probably reflects the fact that the conformational alterations produced by the mutations are at least in part cancelled by monomer-monomer interactions and the association of actin with different actin-binding proteins in the cell.
We originally wished to test the hypothesis that His 73 regulates the rate of P i release from F-actin following actin-catalyzed hydrolysis of ATP. However, from this study, His 73 clearly plays an important role as a structural determinant of the actin monomer with concomitant effects on actin function. Therefore, one will have to be cautious in the interpretation of assays that may indicate a difference in the rate of P i release from WT versus the His 73 mutant species.