Using proteolytic susceptibility as a probe, we have identified four regions of the actin polypeptide chain where structural rearrangements, dependent on the nature of the tightly bound metal ion and/or nucleotide, take place. Replacement of the tightly bound Ca2+ by Mg2+ in ATP-actin strongly affected the regions around Arg26 and Lys68, as judged from nearly complete inhibition of tryptic cleavages of the polypeptide chain at these residues. It also significantly diminished the rates of splitting by trypsin of the peptide bonds involving carbonyl groups of Arg372 and of Lys373 in the C-terminal segment. Conversion of ATP-actin to ADP-actin (with Mg2+ as the tightly bound cation) abolished the protective effect of Mg2+ on specific tryptic cleavage and, in contrast, largely inhibited proteolysis at specific sites for subtilisin and for a novel protease from Escherichia coli A2 strain within a surface loop of residues 39-51. We also examined the effect of proteolytic cleavage or chemical modification at certain sites on the kinetics of proteolysis at other sites of the molecule. These experiments demonstrated structural relationships between loop 39 -51 and regions involving Lys61 and Lys68. It is suggested that the conformational transitions reflected in the observed changes in proteolytic susceptibility may underlie the known influence of the nature of the tightly bound cation and nucleotide on the kinetics of actin polymerization and stability of the polymer.The ability of monomeric actin to polymerize into helical filaments is a fundamental property of this protein. It is well established that polymerization is strongly influenced by the kind of divalent cation and nucleotide occupying the single high-affinity site for each of these ligands. Actin containing bound Mg2+ (Mg-actin) has a lower critical concentration and polymerizes faster than actin with bound Ca2+ (Ca-actin). Similarly, actin prepared by procedures commonly in use, which contains ATP at the high-affinity site for nucleotide (ATP-actin), has a lower critical concentration and polymerizes faster than ADP-actin. These differences themselves, as well as a number of other observations, are indicative of nucleotide-dependent and metal-ion-dependent alterations in the monomer conformation (for reviews, see [l, 21). Little information, however, is available on the location of these changes in the actin polypeptide chain. Most extensively examined were the alterations in the environment of Cys374 monitored as changes in the fluorescence of actin labelled at this residue with N-iodoacetyl-K-(5-sulfo-l-naphthyl)-

Actin can be specifically cleaved between residues 42 and 43 with a novel protease from Escherichia coli A2 strain (ECP) [Khaitlina, S. Y., Collins, J. H., Kuznetsova, I. M., Pershina, V. P., Synakevich, I. G., Turoverov, K. K. & Usmanova, A. M. (1991) FEBS Lett. 279, 49-51]. The resulting C-terminal and N-terminal fragments remained associated to one another in the presence of either Ca" or Mg". The protease-treated actin was, however, neither able to spontaneously assemble into filaments nor to copolymerize with intact actin unless its tightly bound Caz+ was replaced with Mg". Substitution of Mg2+ for the bound Ca2+ was also necessary to partially restore the ability of the protease-treated actin to inhibit the DNase I activity.The critical concentration for KC1-induced polymerization of ECP-treated ATP-Mg-G-actin, determined by measuring the fluorescence of pyrenyl label, was approximately threefold higher than that for actin cleaved between residues 47 and 48 using subtilisin, and 36-fold higher than the critical concentration for polymerization of intact actin under the same conditions. Morphologically, the filaments of ECP-treated actin were indistinguishable from those of intact actin. Comparison of the fluorescence spectra of pyrenyl-labelled actins and chemical cross-linking with N, Nf-l,2-phenylenebismaleimide have, however, revealed structural differences between the filaments assembled from ECP-treated actin and those of intact as well as subtilisin-treated actin. Moreover, the filaments of ECP-treated actin were easily disrupted by centrifugal forces or shearing stress unless they were stabilized by phalloidin. The results are consistent with the direct participation of the region around residues 42 and 43 in the monomer/monomer interactions as predicted from the atomic model of F-actin [Holmes, K. C., Popp, D., Gebhard, W. & Kabsch, W. (1990) Nature 347, [44][45][46][47][48][49] and suggest that the interactions involving this region are of primary importance for stabilization of the actin filament. The mechanism of the regulation of actin polymerization by the tightly bound divalent cation is also discussed.The atomic model of F-actin presented by Holmes et al.[ 11 has resolved the earlier controversies about the monomer orientation in the F-actin filament and, for the first time, provided precise information on the intermonomer contact sites. This model is, however, based on the monomer structure in G-actidDNase I complex in which certain regions of the actin molecule could be distorted by the interaction with DNase I (21. Moreover, it is well documented that the monomer undergoes conformational changes at various stages of the polymerization process [3], although so far very little is known of the exact nature of these changes.According to the model of Holmes et al., the N-terminal segment (residues 40-45) of a surface loop 39-51 on the top of subdomain 2 of the monomer [2] is involved in extenCorrespondence to H.

Effects of proteolytic modifications of the DNase-I-binding loop (residues 39-51) in subdomain 2 of actin on F-actin dynamics were investigated by measuring the rates of the polymer subunit exchange with the monomer pool at steady state and of ATP hydrolysis associated with it, and by determination of relative rate constants for monomer addition to and dissociation from the polymer ends. Cleavage of actin between Gly-42 and Val-43 by protease ECP32 resulted in enhancement of the turnover rate of polymer subunits by an order of magnitude or more, in contrast to less than a threefold increase produced by subtilisin cleavage between Met-47 and Gly-48. Probing the structure of the modified actins by limited digestion with trypsin revealed a correlation between the increased F-actin dynamics and a change in the conformation of subdomain 2, indicating a more open state of the filament subunits relative to intact F-actin. The cleavage with trypsin and steady-state ATPase were cooperatively inhibited by phalloidin, with half-maximal effects at phalloidin to actin molar ratio of 1:8 and full inhibition at a 1:1 ratio. The results support F-actin models in which only the N-terminal segment of loop 39-51 is involved in monomer-monomer contacts, and suggest a possibility of regulation of actin dynamics in the cell through allosteric effects on this segment of the actin polypeptide chain.

Truncated derivatives of actin devoid of either the last two (actin-2C) or three residues (actin-3C) were used to study the role of the C-terminal segment in the polymerization of actin. The monomer critical concentration and polymerization rate increased in the order: intact actin < actin-2C < actin-3C. Conversely, the rate of hydrolysis of actin-bound ATP during spontaneous polymerization of Mg-actin decreased in the same order, so that, for actin-3C, the ATP hydrolysis significantly lagged behind the polymer growth. Probing the conformation of the nucleotide site in the monomer form by measuring the rates of the bound nucleotide exchange revealed a similar change upon removal of either the two or three residues from the C-terminus. The C-terminal truncation also resulted in a slight decrease in the rate of subtilisin cleavage of monomeric actin within the DNAse-I binding loop, whereas in F-actin subunits the susceptibility of this and of another site within this loop, specifically cleaved by a proteinase from Escherichia coli A2 strain, gradually increased upon sequential removal of the two and of the third residue from the C-terminus. From these and other observations made in this work it has been concluded that perturbation of the C-terminal structure in monomeric actin is transmitted to the cleft, where nucleotide and bivalent cation are bound, and to the DNAse-I binding loop on the top of subdomain 2. Further changes at these sites, observed on the polymer level, seem to result from elimination of the intersubunit contact between the C-terminal residues and the DNAse-I binding loop. It is suggested that formation of this contact plays an essential role in regulating the hydrolysis of actin-bound ATP associated with the polymerization process.

Homogeneous preparations of actin devoid of the three C-terminal residues were obtained by digestion of G-actin with trypsin after blocking proteolysis at other sites by substitution of Mg2+ for the tightly bound Ca2+. Removal of the C-terminal residues resulted in the following: an enhancement of the Mg(2+)-induced hydrolysis of ATP in low-ionic-strength solutions of actin; an increase in the critical concentration for polymerization; a decrease in the initial rate of polymerization; and an enhancement of the steady-state exchange of subunits in the polymer. Electron microscopy indicated an increased fragility of the filaments assembled from truncated actin. The results suggest that removal of the C-terminal residues increases the rate constants for monomer dissociation from the polymer ends and from the oligomeric species.