Influence of Tightly Bound Mg2+ and Ca2+, Nucleotides, and

Biochemistry 1998, 37, 14529-14538

14529

Influence of Tightly Bound Mg2+ and Ca2+, Nucleotides, and Phalloidin on the Microsecond Torsional Flexibility of F-Actin† Conrad A. Rebello and Richard D. Ludescher* Department of Food Science, Rutgers, The State UniVersity, New Brunswick, New Jersey 08901-8520 ReceiVed May 26, 1998; ReVised Manuscript ReceiVed July 28, 1998

ABSTRACT: To better understand the relationship between structure and molecular dynamics in F-actin, we have monitored the torsional flexibility of actin filaments as a function of the type of tightly bound divalent cation (Ca2+ or Mg2+) or nucleotide (ATP or ADP), the level of inorganic phosphate and analogues, KCl concentration, and the level of phalloidin. Torsional flexibility on the microsecond time scale was monitored by measuring the steady-state phosphorescence emission anisotropy (rFA) of the triplet probe erythrosin-5-iodoacetamide covalently bound to Cys-374 of skeletal muscle actin; extrapolations to an infinite actin concentration corrected the measured anisotropy values for the influence of variable amounts of rotationally mobile G-actin in solution. The type of tightly bound divalent cation modulated the torsional flexibility of F-actin polymerized in the presence of ATP; filaments with Mg2+ bound (rFA ) 0.066) at the active site cleft were more flexible than those with Ca2+ bound (rFA ) 0.083). Filaments prepared from G-actin in the presence of MgADP were more flexible (rFA ) 0.051) than those polymerized with MgATP; the addition of exogenous inorganic phosphate or beryllium trifluoride to ADP filaments, however, decreased the filament flexibility (increased the anisotropy) to that seen in the presence of MgATP. While variations in KCl concentration from 0 to 150 mM did not modulate the torsional flexibility of the filament, the binding of phalloidin decreased the torsional flexibility of all filaments regardless of the type of cation or nucleotide bound at the active site. These results emphasize the dynamic malleability of the actin filament, the role of the cation-nucleotide complex in modulating the torsional flexibility, and suggest that the structural differences that have previously been seen in electron micrographs of actin filaments manifest themselves as differences in torsional flexibility of the filament.

F-Actin, an asymmetric polymer found in all eukaryotic cells, is an essential structural and contractile element involved in cellular functions as diverse as cell locomotion, cell division, maintenance of specific cellular shape, and muscle contraction (1). The structure of actin is known at atomic resolution from X-ray diffraction of complexes of skeletal R-actin with DNAase I (2) and with segment 1 of gelsolin (3) and of β-actin with profilin (4). Actin is a 42 kDa globular protein folded into distinct domains separated by a deep cleft. The actin filament (diameter = 90 Å) can be described as a two-start, right-handed helix. A model of the filament at atomic resolution was generated by fitting the actin-DNAase I structure to low-resolution X-ray fiber diffraction patterns from oriented F-actin gels (5, 6). This model now forms the basis for molecular interpretations of F-actin structure and function (7-10) and appears to be consistent with other data on filament structure (5). Actin contains a single high-affinity (Kd ≈ 10-8 M for both Ca2+ and Mg2+) (11) divalent cation binding site which † This work was supported by a grant to R.D.L. from the Muscular Dystrophy Association and by a graduate assistantship awarded to C.A.R. from the New Jersey Agricultural Experiment Station. This is publication number D-10567-2-98 of the New Jersey Agricultural Experiment Station. * To whom correspondence should be addressed. Telephone: (732) 932-9611, ext 231. Fax: (732) 932-6776. E-mail: ludescher@ aesop.rutgers.edu.

has been located at the bottom of the deep cleft between the large and small domains. The nucleotide is also bound in this cleft, and interactions between the cation and the nucleotide phosphates indicate that cations bound to the highaffinity site are actually bound to the nucleotide. Under cellular conditions, the cation is Mg2+, but most standard actin preparations isolate Ca2+-actin (11). The nature of the divalent cation bound to the nucleotide at this site influences the large-scale structure and dynamics of F-actin. Several studies have demonstrated that the structure of the actin filament can be altered by varying either the type of tightly bound divalent cation (Ca2+ or Mg2+) (9, 10, 12-14) or the nucleotide (ADP or ATP) (15) that is present in monomeric G-actin prior to polymerization. On the basis of these structural differences, differences in the bending flexibility of actin filaments have been proposed. However, when the bending flexibility of actin filaments which differed in the type of tightly bound divalent cation was measured directly (16) or inferred from persistence length measurements (17), these predictions were not fulfilled. Similarly, when the bending flexibility of actin filaments was determined as a function of the bound nucleotide (ATP or ADP) the data have either affirmed (18) or refuted (10, 17, 19) the predictions of the structural studies. Such apparently contradictory results suggest that our understanding of the manner in which the structural changes in F-actin manifest themselves as flexibility changes is incomplete.

S0006-2960(98)01240-9 CCC: $15.00 © 1998 American Chemical Society Published on Web 09/25/1998

14530 Biochemistry, Vol. 37, No. 41, 1998 Most studies of F-actin dynamics have focused in on the flexural (bending) flexibility of the filament, since this dynamic property is believed to be important for actin’s ability to carry out its diverse functions in both muscle and nonmuscle cells (17, 20). The torsional flexibility of the filament, however, has not been adequately addressed despite important functional implications for the formation of bonds that hold bundles of actin filaments together in the microvilli, for the formation of actin bundles in limulus sperm, and for any as yet undefined role in actomyosin function (16, 21). In addition, Tsuda et al. (22) have demonstrated that the actin filament is more sensitive to torsional twisting than it is to bending. Since the studies summarized above have indicated that the observed structural differences do not show up as bending flexibility differences, we hypothesize that these structural differences may manifest themselves as differences in torsional flexibility. We have thus set out to monitor the torsional flexibility of F-actin as a function of the identity of the tightly bound divalent cation (Ca2+ or Mg2+) or the specific nucleotide (ADP or ATP) present in G-actin prior to polymerization, the solution ionic strength, and the presence of the filamentstabilizing ligand phalloidin. The torsional flexibility of F-actin was monitored by measuring the steady-state phosphorescence emission anisotropy of the phosphorescent probe erythrosin-5-iodoacteamide covalently attached to Cys-374 of G-actin. The phosphorescence emission anisotropy is sensitive to rotational motions taking place in the actin filament on the micro- to millisecond time scale. Previous studies (23, 24) have established that the microsecond rotational motions that take place in F-actin are dominated by torsional twisting motions. Our measurements indicate that the torsional dynamics of actin filaments are remarkably sensitive to the chemical structure of the divalent cationnucleotide complex bound at the active site cleft. MATERIALS AND METHODS Protein Preparation and Labeling. Actin was isolated from acetone powder prepared from the leg and back muscles of New Zealand white rabbits of either sex as per established protocols (25). G-Actin was isolated from acetone powder by extraction into G-buffer (GB) [1 mM EPPS,1 0.2 mM CaCl2, 1 mM NaN3, and 0.2 mM ATP (pH 8.5)] containing 0.5 mM DTT as described by Thomas et al. (26); such a procedure generates G-actin with Ca2+ at the high-affinity site (G-Ca2+-actin). Actin was stored as filaments with continuous dialysis against filament buffer (FB) [10 mM MOPS, 100 mM KCl, 2 mM MgCl2, 0.2 mM CaCl2, 1 mM NaN3, and 0.2 mM ATP (pH 7.0)] at 0 °C; the dialysis buffer was changed weekly. The actin concentration was determined by the absorbance at 290 nm using an extinction coefficient of 0.63 M-1 cm-1 for actin. The protein was assayed for purity using SDS-PAGE. Actin was labeled with the phosphorescent probe erythrosin-5-iodoacetamide (Molecular Probes, Inc., Eugene, OR) as per previously reported procedures (27) and was stored as F-actin by dialysis against FB. 1 Abbreviations: Ap A, P1,P5-di(adenosine 5′)-pentaphosphate; DTT, 5 dithiothreitol; EGTA, ethyleneglycol bis(β-aminoethyl ether)-N,N,N′,N′tetraacetic acid; EPPS, N-(2-hydroxyethyl)piperazine-N′-3-propanesulfonic acid; FB, filament buffer; MOPS, 3-(N-morpholino)propanesulfonic acid.

Rebello and Ludescher F-Ca2+- or F-Mg2+-actin was prepared by polymerizing G-actin with FB as per established protocols (27) and was dialyzed against this buffer. An actin filament containing predominantly Ca2+ at the high-affinity divalent cation binding site, F-Ca2+-actin, was prepared by polymerizing G-Ca2+ with a modified filament buffer containing only calcium as the divalent cation (Ca2+-FB) [10 mM MOPS, 100 mM KCl, 2.2 mM CaCl2, 1 mM NaN3, and 0.2 mM ATP (pH 7.0)] and was dialyzed against this buffer. G-Actin containing Mg2+ at the high-affinity divalent cation binding site (G-Mg2+-actin) was prepared according to the protocol of Orlova and Egelman (10) by incubating G-Ca2+ with 0.4 mM EGTA and 0.2 mM MgCl2 for 10-15 min at 4 °C. G-Mg2+-actin was then polymerized to F-Mg2+-actin (actin filaments containing predominantly Mg2+ at the high-affinity divalent cation binding site) in a modified filament buffer (Mg2+-FB) [10 mM MOPS, 100 mM KCl, 0.2 mM EGTA, 2.2 mM MgCl2, 1 mM NaN3, and 0.2 mM ATP (pH 7.0)]. G-Actin containing Mg2+ as the tightly bound divalent cation and ADP as the nucleotide (G-Mg2+-ADP-actin) was prepared from G-Mg2+-ATP-actin by incubating a 2 mg/mL sample with 30 units/mL hexokinase (Sigma Chemical Co., St. Louis, MO) and 2.5 mM glucose for 2 h on ice as per the protocol of Isambert et al. (17). F-Mg2+-ADP-actin was prepared by polymerizing G-Mg2+-ADP-actin with a Mg2+ADP buffer [10 mM MOPS, 100 mM KCl, 0.2 mM EGTA, 2.2 mM MgCl2, 1 mM NaN3, 0.2 mM ADP, and 5 µM Ap5A (pH 7.0)]. F-Mg2+-ADP-Pi-actin was prepared by polymerizing G-Mg2+-ADP-actin in Mg2+-ADP buffer containing 30 mM Pi. ADP-actin containing BeF3-, F-Mg2+-ADP-BeF3-actin, was prepared by polymerizing G-Mg2+-ADP-actin with Mg2+-ADP buffer containing 100 µM BeSO4 and 10 mM NaF as per the procedure of Isambert et al. (17). These proteins were stored by dialysis at 0 °C against the appropriate buffer. Spectroscopic Measurements. Steady-state fluorescence measurements were taken on a SPEX model F1T11i spectrofluorometer (SPEX Industries, Metuchen, NJ) equipped with a 450 W high-pressure xenon lamp, single-grating excitation and emission monochromators, dual emission monochromators in a T-format, Glan-Thomson crystal polarizers on excitation and emission, and a circulating water bath for controlling the sample temperature. The instrument is under control of a microcomputer running DM3000F software (SPEX Industries). Phosphorescence measurements were taken with this instrument equipped with a pulsed, lowpressure Xe flash lamp and a model 1934C phosphorimeter attachment (SPEX Industries). Samples for spectroscopic measurements are prepared by mixing labeled and unlabeled actin to adjust the total concentration of erythrosin to 1 µM and the total concentration of actin (labeled and unlabeled) to 0.5 mg/mL. All luminescence measurements were taken under anaerobic conditions using the enzymatic deoxygenation protocol of Horie and Vanderkooi (28) and at a temperature of 20 °C. Steady-state fluorescence emission anisotropy spectra were collected for both intrinsic tryptophan and the erythrosin label; tryptophan emission anisotropy over the range of 340350 nm was measured using 295 nm excitation, while erythrosin emission anisotropy over the range of 560-570 nm was measured using 500 nm excitation. Steady-state phosphorescence emission anisotropy measurements were

Microsecond Torsional Flexibility of F-Actin

Biochemistry, Vol. 37, No. 41, 1998 14531

taken at 684 nm with excitation at 534 nm using a time delay of 0.07 ms and integrating the emitted photons over a total time window of 1.5 ms; four polarized intensities corresponding to vertical and horizontal excitation and emission (Ivv, Ivh, Ihv, and Ihh, where the subscripts refer to the excitation and emission polarizations, respectively) were collected and used to calculate the anisotropy r with the equation r ) (R - 1)/(R + 2), where R ) (Ivv/Ivh)(Ihh/Ihv). Phosphorescence intensity decay measurements were taken with excitation at 534 nm and emission at 688 nm, using an initial time delay of 0.07 ms and a total decay window of 1.5 ms, and collecting intensity every 0.01 ms. Additional details are reported elsewhere (27, 29). Total intensity decays were analyzed for single-exponential lifetimes using the nonlinear least-squares fitting program NFIT (Island Products, Galveston, TX); the goodness of fit was evaluated by examination of the residuals (data minus fit) and the fit χ2. Singleexponential functions gave good fits with residuals randomly distributed around zero and χ2 values in the range of 1-1.4. Determination of the “True Anisotropy” of F-Actin. When there exist in solution two chemical species such as G-actin and F-actin which differ in anisotropy, the measured anisotropy of the resultant solution is the intensity-weighted sum of the anisotropies of the two species (27). If we designate rM as the measured steady-state anisotropy of the actin solution and rGA and rFA as the anisotropies of G-actin and F-actin, respectively, which contribute fractional phosphorescence intensities of fGA and fFA to the solution, respectively (such that fGA + fFA ) 1), then

rM ) fGArGA + fFArFA

(1)

Since we know from experiment that rGA ) 0 for steadystate phosphorescence (27), then

rM ) fFArFA ) (1 - fGA)rFA

(2)

The term fGA is a function of the concentrations of G-actin ([G]) and F-actin ([F]) and a term IFA/IGA which represents the relative phosphorescence emission intensity of F-actin versus G-actin.

fGA ) [G]/[[G] + [F](IFA/IGA)] Since the critical concentration of actin (that is, [G] in the above expression) is known to be in the range of about 0.010.02 mg/mL (30), at the concentrations of actin used in our experiments (0.2-0.5 mg/mL), fGA can be approximated by the following equation:

fGA ≈ [G]/[A](IFA/IGA)

(3)

where [A] is the total concentration of actin. IFA/IGA ) 2.81, which is estimated from the ratio of mean lifetimes of G-actin (0.099 ms) and F-actin (0.278 ms) (27). If the critical concentration was 0.05 mg/mL and the total actin concentration 0.5 mg/mL, the approximation of eq 3 would introduce a 6% error. Substituting eq 3 into eq 2, we obtain the following expression relating the measured anisotropy to the total actin concentration [A]

rM ) rFA - ([G]/[A]IFA/IGA)rFA Since the concentration of G-actin in solution is the critical concentration, cc, then

rM ) rFA - (cc/IFA/IGA)rFA/[A]

(4)

Thus, a plot of rM versus 1/[A] provides a measure of the true anisotropy of the actin filament, rFA, and the critical concentration, cc. RESULTS Influence of the Tightly Bound DiValent Cation on F-Actin Dynamics. Previous studies indicate that the steady-state phosphorescence anisotropy of erythrosin attached to Cys374 provides a reliable measure of the average rotational dynamics of F-actin (27, 29, 31). Time-resolved studies (23, 24) have demonstrated that the optical anisotropy on the triplet time scale reflects torsional twisting motions about the long axis of the actin filament. Such an interpretation is also supported by fluorescence microscopic studies of the reptation of labeled F-actin in solutions of unlabeled actin (32) which show that at actin concentrations of g0.5 mg/ mL the actin filament is embedded within a gel-like meshwork of entangled filaments that effectively constrains any end-over-end motions of the filament. We have thus used the steady-state phosphorescence anisotropy as an indicator of the average torsional motions of the filament; a theoretical treatment (33) indicates that it also provides a measure of the torsional rigidity of the filament. To determine the effect of the tightly bound divalent cation on the torsional flexibility of F-actin, F-Mg2+-actin and F-Ca2+-actin filaments were prepared by polymerizing G-Mg2+-actin or G-Ca2+-actin with a solution containing either MgCl2 and KCl or CaCl2 and KCl, respectively. The torsional flexibilities of these filaments at 20 °C were compared to those of filaments prepared from G-Ca2+-actin monomers, but polymerized with CaCl2, and MgCl2 and KCl; we refer to this sample as F-(Ca2+/Mg2+)-actin. Additional filament samples were prepared in the absence of KCl by polymerizing G-Ca2+-actin and G-Mg2+-actin in buffers containing only the respective divalent cation but no KCl. F-Ca2+-actin and F-Mg2+-actin exhibit differences in critical concentration regardless of the type or concentration of polymerizing salt used (11, 34); since the phosphorescence anisotropy of G-actin is 0.0 (27), variable amounts of G-actin within a filament solution will introduce artifacts by lowering the measured anisotropy. The anisotropy was thus determined as a function of actin concentration so the true anisotropy of these filaments could be determined (see Materials and Methods, especially eq 4). The steady-state phosphorescence emission anisotropies at 20 °C of the various actin samples determined as a function of actin concentration are shown in Figure 1. The y-intercepts of the regression lines of the anisotropy versus 1/[actin] curves provide the true anisotropy of F-actin (rFA), while the slope provides an estimate of the critical concentration of the actin solution (cc) (this procedure provides a novel technique for determining the critical concentration of an actin solution). Estimates of the critical concentrations of G-actin in each of these solutions (cc) and of the true anisotropy of each of these types of actin filaments (rFA) are summarized in Table

14532 Biochemistry, Vol. 37, No. 41, 1998

Rebello and Ludescher

Table 1: Influence of Divalent Cation, KCl, and Phalloidin on the Steady-State Phosphorescence Emission Anisotropy of Erythrosin-Labeled F-Actin F-actin F-Ca2+-actin F-Mg2+-actin F-(Ca2+/Mg2+)actin F-Ca2+-actin F-Mg2+-actin Ph-F-Ca2+-actin Ph-F-Mg2+-actin Ph-F-(Ca2+/Mg2+)actin

phosphorescence emission anisotropy (rFA)

phosphorescence lifetime (ms)

critical concentration (µM)

Ca2+-FB Mg2+-FB FB

0.083 ( 0.002a 0.066 ( 0.002a 0.084 ( 0.002a

0.306 ( 0.005b 0.263 ( 0.005b 0.281 ( 0.008b

1.6 ( 0.39e _d 0.62 ( 0.81e

Ca2+-FB (no KCl) Mg2+-FB (no KCl) Ca2+-FB Mg2+-FB FB

0.083 ( 0.002a

0.334 ( 0.008b

1.8 ( 0.14e

0.066 ( 0.002a

0.305 ( 0.007b

0.39 ( 0.34e

0.098 ( 0.005 0.080 ( 0.005 0.102 ( 0.006

0.293 ( 0.004c 0.271 ( 0.005c 0.276 ( 0.004c

_d _d _d

buffer

a From intercept on the y-axis of the regression line in Figure 1. b Mean ( standard deviation over the actin concentrations in Figure 1. c Mean ( standard deviation of at least three replicates. d Unable to determine the critical concentration due to zero slope. e Mean ( standard deviation based on propagation of errors calculated from the slopes of lines in Figures 1 and 2.

FIGURE 1: Measured steady-state phosphorescence emission anisotropy (rM) of various F-actin solutions plotted as a function of the reciprocal of the actin concentration (see Materials and Methods, eq 4, and the text for details): (0) F-(Ca2+/Mg2+)-actin polymerized in FB, (O and b) F-Ca2+-actin polymerized in Ca2+-FB and Ca2+FB (no KCl), respectively, and (4 and 2) F-Mg2+-actin polymerized in Mg2+-FB and Mg2+-FB (no KCl), respectively.

1. The critical concentration of F-Ca2+-actin polymerized with CaCl2 and KCl was 1.6 µM, and that of F-Ca2+-actin polymerized with only CaCl2 was 1.8 µM; both values were considerably larger than the critical concentration of F-Mg2+actin polymerized with only MgCl2 (0.4 µM) or that of F-(Ca2+/Mg2+)-actin polymerized with CaCl2, MgCl2, and KCl (0.6 µM). (The critical concentration of F-Mg2+-actin polymerized with Mg2+ and KCl could not be determined from this analysis; however, the absence of a concentration effect in the presence of Mg indicates that the concentration of free monomer is probably