HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Razzaq, A. Articles by Sparrow, J. C. Search for Related Content PubMed PubMed Citation Articles by Razzaq, A. Articles by Sparrow, J. C. Social Bookmarking                      What's this?

J Biol Chem, Vol. 274, Issue 40, 28321-28328, October 1, 1999

Actin Residue Glu⁹³ Is Identified as an Amino Acid Affecting Myosin Binding^*

Azam Razzaq^§, Stephan Schmitz^¶, Claudia Veigel, Justin E. Molloy, Michael A. Geeves^¶**, and John C. Sparrow

From the  Department of Biology, University of York, P.O. Box 373, York YO10 5YW, United Kingdom and ^¶ Max-Planck-Institut für Molekulare Physiologie, Rheinlanddamm 201, D-44139 Dortmund, Germany

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Many mutants have been described that affect the function of the actin encoded by the Drosophila melanogaster indirect flight muscle-specific actin gene, Act88F. We describe the development of procedures for purification of this actin from the other isoforms expressed in the fly as well as in vitro motility, single molecule force/displacement measurements, and stop-flow solution kinetic studies of the wild-type actin and that of the E93K mutation of the Act88F gene. We show that this mutation affects in vitro motility of F-actin, in both the presence and absence of methylcellulose, and the ability of the ACT88F actin to bind the S1 fragment of rabbit skeletal myosin. However, optical tweezer measurements of the actomyosin working stroke and the force transmitted from the rabbit heavy meromyosin to and through F-actin are unchanged by the mutation. These results support the proposal (Holmes, K. C. (1995) Biophys J. 68, (suppl.) 2-7) that actin residue Glu⁹³ is part of the secondary myosin binding site and suggest that myosin binding occurs first at the primary myosin binding site and then at the secondary site.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Actin and myosin are ubiquitous in eukaryotic cells, where they form one of the major motor protein systems. The atomic structure determinations of the actin monomer (2) and the S1 fragment (motor domain) of the chicken skeletal muscle myosin (3) were major landmarks in our understanding of how this motor system works. The "docking" of the atomic structures of the motor domain to that of F-actin (4, 5) showed that myosin binds two neighboring actin monomers within the F-actin helix and identified the amino acids potentially involved in the actomyosin interface. The two contact sites are referred to as the primary and secondary binding sites (4, 5).

The limited resolution of the S1-decorated actin structure leaves considerable doubt as to which amino acid residues are involved in the actomyosin binding sites, particularly the secondary one. In addition, the S1-decorated F-actin is in a rigor conformation, and it is not known to what extent the monomer crystal structures are relevant for the complex. In fact, adjustments were made to the S1 structure to fit the EM densities (4, 5). Little is known of the actomyosin structure during other steps of the cross-bridge cycle. One way to probe the interactions between residues in both proteins is to make amino acid substitutions in both proteins and determine the functional consequences.

We and others are taking a genetic approach to understand the details of the actomyosin interactions by making and studying mutations within actin genes from Drosophila (6-8), Dictyostelium (9), and yeast (10-13). This involves using predictions from the actomyosin structures to generate mutants in residues believed to be within the actomyosin interface and, using a combination of assays, to ask in which ways the mutants affect the cross-bridge cycle.

The Act88F actin gene of Drosophila is expressed specifically in the indirect flight muscles (IFMs)¹ (14) and is the only actin expressed in these muscles (15). This has allowed the only genetic study of a muscle actin to date. Many Act88F mutants have been recovered either by selecting for flightless mutants (16-18) or by in vitro mutagenesis and germ line transformation (7, 8, 19, 20). However, because flies are small and multicellular and express six actin genes (21, 22), recovery of pure ACT88F actin has been a challenge that has limited the biochemical and biophysical techniques that can usefully be applied to study mutant actins (6). Since the IFMs account for approximately one-fifth of the body volume of the adult fly and the ACT88F actin is the major isoform, we believed that it should be possible to isolate pure ACT88F actin.

To study actomyosin interactions we have 1) developed a minipurification of ACT88F actin from dissected IFMs that yields sufficient quantities (1-10 µg) for in vitro motility, force/displacement measurements, and some biochemical assays; 2) developed a large scale purification (0.1-1.0 mg) of the ACT88F isoform from a whole fly actin preparation for more detailed stop-flow kinetic studies; and 3) used these approaches with the E93K mutation to investigate whether residue Glu⁹³ is involved in the actomyosin interface and to what extent it affects this binding. Our particular interest in choosing to study this mutant was to examine the proposal for the secondary myosin binding site on actin, for which there is currently little experimental support, and in which residue Glu⁹³ has been identified as having a role.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Fly Strains-- An ry⁵⁰⁶ homozygous strain (ry⁵⁰⁶ is an allele of the rosy eye-color gene) was used to prepare wild-type (WT) Drosophila actins. Whole fly actin preparations lacking the IFM-specific isoform, ACT88F, were made from ry⁵⁰⁶ KM88 e^s flies (KM88 is an Act88F gene null mutation; Ref. 20). The E93K mutant actin was isolated from the ry⁵⁰⁶ KM88 e^s P[Act88F^E93K] transgenic strain, which has a single copy of the E93K mutant Act88F gene inserted into the KM88 strain by P-element-mediated transformation (8). Flies were grown on standard Drosophila media. Young adults were collected and stored at 20 °C prior to processing.

Miniactin Extraction from Dissected IFMs-- Approximately 5 µg of pure ACT88F actin (co-purified with arthrin, an actin-ubiquitin conjugate; see Ref. 15) was isolated from the dissected IFMs (see Ref. 23) of a minimum of 10 flies. Dissected dorso-longitudinal muscles in 0.5 ml of York modified glycerol (20 mM KPO₄, 50% (v/v) glycerol, 0.5% (v/v) Triton X-100, 2 mM MgCl₂, 1 mM dithiothreitol, 1 mM NaN₃, pH 7.0) were spun at 13,000 rpm for 2 min in a benchtop centrifuge. The pellet was extracted in 500 µl of high salt buffer (20 mM KPO₄, 800 mM KCl, pH 7.0) at room temperature for 2 min. After centrifuging, the pellet was washed twice with deionized water (1 ml and then 0.5 ml) to remove salt and dehydrated using 50 and then 100% ice-cold acetone, followed by air-drying at room temperature for >15 min.

Actin was extracted from the acetone powder with two consecutive 30-µl aliquots of ACEX (2 mM Tris-Cl, 0.2 mM CaCl₂, 0.2 mM ATP, 1 mM dithiothreitol, pH 8.0) for 45 min each. Extracts were combined, and the actin was polymerized by the addition of 6 µl of 10× polymerization buffer (50 mM Tris-Cl, 500 mM KCl, 20 mM MgCl_2, 10 mM ATP, pH 8.0) for 2 h at room temperature. Solid KCl was added to a final concentration of 800 mM and incubated for 30 min at 37 °C to release tropomyosin and troponin contaminants from the F-actin. After spinning at 100,000 rpm for 10 min at 4 °C, the F-actin pellets were resuspended in 25 µl of ACEX and labeled with rhodamine-phalloidin (24).

Maxiactin Preparation-- The actin purification scheme used for the miniactin preparation from IFMs was scaled up to purify 100-µg to 1-mg quantities of Drosophila ACT88F from thousands of flies. A myofibrillar preparation was made from 50 g of flies (50,000 flies) according to the method of Saide et al. (25). All centrifugations at this stage were performed at 11,500 rpm for 20 min at 4 °C in a Sorvall RC5B centrifuge using a GSA rotor. After a myosin extraction with 200 ml of 20 mM KPO₄, 800 mM KCl, pH 7.0, on ice for 10 min, followed by centrifugation, the pellet was quickly washed twice with 600 ml of deionized water and then dehydrated using acetone (see above) but finally air dried overnight at room temperature.

Actin extraction from the acetone powder followed essentially the methods of Bullard et al. (26) and Pardee and Spudich (27) using two successive volumes of ACEX (see above) for 30 min each at 4 °C. After the first extract was spun at 20,000 rpm for 15 min at 4 °C, the pellet was re-extracted for 30 min with another aliquot of ACEX, and a one-tenth volume of 10× polymerization buffer (see above) was added to the first extract to initiate polymerization. The second extract was treated similarly and combined with the first, and polymerization continued for a further 60 min at 4 °C. The contaminating thin filament proteins were removed by adding KCl to 800 mM, incubation at 37 °C, and centrifugation (see above). The F-actin was resuspended in ACEX, homogenized, and dialyzed overnight in 3× 1-liter changes of ACEX. Optical density at A₂₉₀-A₃₁₀, with an extinction coefficient of 0.63 cm²/mg, was used to estimate G-actin molarity. Actin purity was assessed by one-dimensional SDS-PAGE.

The ACT88F isoform was purified from the isoform mixture in whole fly actin extracts by anion exchange chromatography. G-actin preparations were clarified by spinning at 90,000 rpm for 20 min at 4 °C in a Beckman TLA100.3 rotor and then filtered through a 0.20-µm Millipore filter before loading onto a 1-ml anion exchange column (Amersham Pharmacia Biotech Mono-Q HR5/5). The column was washed with 20 mM MOPS, pH 6.5, and elution was performed with a 28-min, segmented gradient of 0-500 mM NaCl in 20 mM MOPS pH 6.5 (see Fig. 2). Individual peak fractions were analyzed by one- and also two-dimensional gel electrophoresis (28) using a Bio-Rad Mini-Protean II gel electrophoresis system.

Rabbit Skeletal Muscle Heavy Meromyosin (HMM) Preparation-- Rabbit skeletal muscle myosin and HMM were prepared using a modification of the method of Margossian and Lowey (29). The HMM concentration was estimated spectrophotometrically with A₂₈₀ = 0.60 cm²/mg. The stock was drop-frozen in liquid nitrogen and stored at 80 °C.

In Vitro Motility Assays-- In vitro motility assays were performed according to standard methods (24, 30, 31, 32) with nitrocellulose-coated flow cells. All assays used the same batch of rabbit HMM and were performed over a 3-day period. Assay buffers (25 mM imidazole-HCl, 4 mM MgCl₂, 1 mM EGTA, 2 mM ATP, pH 7.4) supplemented with KCl (0-50 mM) were degassed before use.

In vitro motility assays were performed at 23 °C (± 0.5 °C) and recorded onto video tape. Retrac^TM software was used to both grab video images every 0.5 s from at least two separate assay slides and track filaments. Only filaments that moved smoothly and continuously for >3 s were tracked for velocity determinations.

Single Molecule Mechanical Experiments-- The single actomyosin cross-bridge mechanical experiments were carried out using an optical tweezers transducer in the "three-bead" configuration (33-35). Two independently trapped latex beads, attached to the ends of an actin filament, held the filament close to a glass bead coated with a low surface concentration of rabbit skeletal HMM molecules, allowing interactions between single myosin heads and the actin filament. Beam steering of the traps and calibrations were performed as described previously (34-36). Actomyosin interactions were measured at a single trap stiffness of 0.02-0.04 pN nm¹.

The size of the working stroke of rabbit HMM interacting with the different actins was determined by analyzing many displacement events (for details, see Refs. 34 and 35). Histograms of the displacement events could be fitted by a Gaussian distribution whose midpoint shifted from zero, reflected the size of the working stroke, and whose width was determined by the trap stiffness (34).

The cross-bridge stiffness measurements were made by applying a sinusoidal forcing function (100 Hz) to one bead (Fig. 4a) (35). The applied force was calculated from the movement of the driven bead, and the induced extension of the attached cross-bridge was derived from the bead movement of the bead at the other end of the actin filament.

Stop-flow Kinetics: Proteins and Reagents-- Rabbit skeletal F-actin (RSA), pyrene-labeled rabbit F-actin (pyr-RSA), and rabbit myosin subfragment 1 (S1) were prepared, and their molar concentrations were determined as described by Kurzawa and Geeves (37). The absorbance of the pyrene label at 280 nm was calculated by multiplying the absorbance at 344 nm by 1.059 (38).

Phalloidin was used to prevent F-actin from depolymerizing at the low actin concentrations used. Phalloidin-RSA and phalloidin-pyr-RSA stock solutions were made by incubating a 10 µM concentration of the respective F-actins with equimolar concentrations of phalloidin in 20 mM MOPS, 5 mM MgCl₂, 10 mM KCl, pH 7.0 (low salt buffer) overnight at 4 °C.

The fly ACT88F F-actin stock solutions were prepared as follows. Peak B fractions from the anion exchange chromatography runs were collected in 10× polymerization buffer. Excess phalloidin was added to the pooled fractions, and actins were polymerized overnight at 4 °C. F-actin was pelleted at 100,000 rpm for 20 min at 4 °C in a Beckman TLA 100.4 rotor. The pellet was washed and resuspended in the low salt buffer to a concentration of about 10 µM.

Stop-flow Assays-- Stop-flow measurements were performed at 20 °C as described by Kurzawa and Geeves (37) using a Hi-Tech Scientific SF-61MX stop-flow spectrophotometer in fluorescence mode. Transients shown are the average of 5-10 consecutive shots of the stop-flow machine, and the concentrations after mixing are quoted, since these are the relevant ones for kinetic analysis of the exponential curves. In the secondary plot, the concentrations present before mixing in the stop-flow are plotted, since these are the relevant concentrations to assay the fraction of actin bound to S1 at equilibrium. The experimental buffer was 20 mM MOPS, 5 mM MgCl₂, 10-100 mM KCl, pH 7.0.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Minipreparation of Actin-- The minipreparation of the ACT88F isoform depends upon IFM dissection, since the Act88F actin gene is the only actin gene expressed in the IFMs (14, 15, 22). The preparation allows rapid purification of small amounts (~5 µg) of pure ACT88F actin (Fig. 1), sufficient for in vitro motility/force measurements of Act88F mutants. However, these preparations contain arthrin, a stable post-translationally ubiquitinated form of ACT88F (15). No differences in polymerization, activation of myosin ATPase, or myosin binding have been found between actin and arthrin from either Lethocerus (39) or Drosophila.²

View larger version (32K): [in this window] [in a new window]   Fig. 1.   One-dimensional SDS-PAGE separation of the final protein obtained from the miniactin preparation to show that the material (lane 3) consists almost exclusively of ACT88F actin (ACT), mass 42,000, and arthrin (AR). Arthrin has a mass of 51,000 but an Mr of 55,000 due to the shape of the molecule, which involves an isopeptide bond between ubiquitin and an unidentified actin lysine -amino group. Lane 1 shows the major myofibrillar proteins from "skinned" dorso-longitudinal muscles. M, myosin heavy chain, mass 210,000; TnH, troponin-H33 and -H34, mass 55,000, but Mr of 78,000 and 71,000 respectively (see Ref. 44); TnT, troponin-T; GST-2, glutathione sulfuryl transferase-2 (see Ref. 52); TnI/Tm, troponin-I and tropomyosin; RLC, myosin regulatory light chain. Protein band identification was by matrix-assisted laser desorption ionization-time of flight spectroscopy (53). Lane 2, actin extract after polymerization.

ACT88F Purification from Whole Flies-- The calculated isoelectric points of the isoforms from the six Drosophila actin genes fall into three charge groups: 1) ACT57B at 5.23; 2) ACT42A, ACT79B, ACT87E, and ACT88F at 5.30; and 3) ACT5C at 5.37. In the sequences used for these calculations, the first two N-terminal amino acids, methionine and cysteine, were removed, since it is a common feature of class II actins (40), but no other post-translational changes were considered. Separation of the actin isoforms from WT flies on two-dimensional gels produces four actin spots (see Fig. 2a). The largest and most basic isoform, arthrin, is a stable conjugate of ACT88F actin and ubiquitin (15). The three main actin spots are labeled I-III in order of increasing basicity (41) and might be expected to correspond to the three charge groups above. When the Act88F gene is transcribed and translated in vitro under conditions allowing complete N-terminal processing, including acetylation, ACT88F runs, as predicted, at spot II (42). However, the most basic actin, labeled III, on two-dimensional gels of the whole fly actin extract is the mature ACT88F isoform (43), which we refer to as ACT88F-III. This difference relates to the finding that most of the ACT88F undergoes an unknown in vivo post-translational modification under the control of the mod gene (43) to gain a more basic charge. The basic charge of ACT88F predicts that as a major adult fly actin it is separable from the other isoforms.

View larger version (43K): [in this window] [in a new window]   Fig. 2.   a, two-dimensional gel PAGE separation of proteins from a WT actin preparation. The actins separate into three spots, two major ones, labeled III and II, and a minor, more acidic one, I. Arthrin, ubiquitinated ACT88F, is the more basic spot with reduced mobility due to its increased molecular mass (~51,000 Da). Panels b-d are anion exchange separations using a Pharmacia Mono-Q HR 5/5 column on a fast protein liquid chromatography system. The actin isoform mix was applied at 1 mg/ml. Elution was with a segmented salt gradient of 0-500 mM NaCl in 20 mM MOPS, pH 6.5 (buffer A), with a 1 ml/min flow rate. The gradient of 0-150 mM NaCl (5 min), 150-300 mM NaCl (18 min), 300-500 mM NaCl (5 min) was created by pumping buffer A with buffer B (buffer A plus 500 mM NaCl). Fractions were 250 µl. b, separation of WT Drosophila actins into four separate peaks; c, separation of the KM88 strain actins; d, separation of actins from the E93K strain.

The whole fly actin preparations were relatively pure, each containing a major actin band (by one-dimensional SDS-PAGE, not shown) and a band corresponding to arthrin. The major band includes the different isoforms found in adult flies, including nonmuscle actins. Anion exchange chromatography resolved the actins into four separate peaks, labeled A-D according to their elution order from the column (see Fig. 2b). Peak A is the first and most basic actin fraction; peak D is the last and most acidic. The four discrete peaks match the two-dimensional separation of the mixed actins into arthrin and three actin spots (III-I). The identity of peak A as pure arthrin was confirmed by its mass in one-dimensional PAGE (data not shown). To prove that the actin in peaks B and C represented one or more isoforms of a particular pI, two-dimensional gel electrophoresis of different combinations of peak fractions was performed (data not shown). The two nonoverlapping spots found upon mixing of peak B and C samples confirmed that each peak contained one or more actin isoforms of a particular charge, coincident with the expectation that these represent actin spots III and II, respectively.

Since ACT88F is the major adult actin and runs as spot III (14, 43), this is likely to be the major component of peak B. To confirm this anion exchange separations of actin extracts from two Act88F mutants were performed. First, the separation was used on actin preparations from KM88 strain flies that lack any ACT88F because KM88 has a nonsense (stop) mutation in codon 79 of its Act88F gene (20). Anion exchange chromatography of the KM88 actin extract generated three actin peaks (Fig. 2c). These were depicted as Bi, C and D since they eluted at positions corresponding to peaks B, C and D in a WT chromatogram (Fig. 2b). As expected no peak eluted at the position corresponding to the arthrin peak A. Whereas the relative magnitudes of peaks C and D in the KM88 profile were similar to the corresponding peaks in a WT chromatogram, peak Bi is very much smaller than WT peak B. It can therefore be argued that peak B constitutes ACT88F-III in a WT profile. Peak Bi must be another isoform, and the calculated isoelectric points (see above) suggest that it is most likely ACT5C, a cytoplasmic isoform. Since peak Bi in a WT profile will be obscured by co-elution with the ACT88F-III isoform, the WT peak B fractions are only greater than 90% pure ACT88F-III.

Second, due to the amino acid substitution, ACT88F-E93K is more basic than the other isoforms, and peaks A and B from the ACT88F-E93K strain should elute earlier than the corresponding WT peaks. As expected, four major peaks eluted (Fig. 2d), with peaks A and B eluting earlier than the corresponding WT peaks, confirming that peak B is the ACT88F-III actin; the positions and magnitudes of peaks C and D were unaltered. The premature peak B had a similar amplitude to its WT counterpart. Two additional minor actin peaks were distinguishable between peaks B and C, labeled Bi and Bii (Fig. 2d). The earlier elution of peak B in the E93K chromatogram uncovered the same small Bi actin peak seen in the KM88 chromatogram (Fig. 2c). Peak Bii is assumed to be the E93K-ACT88F-II fraction (ACT88F actin unchanged by the mod gene product; Ref. 43), which would also elute earlier than its WT counterpart due to increased basicity. From calculated isoelectric points, peak C (spot II on a two-dimensional gel) is probably a mixture of ACT79B, ACT88F-II, ACT87E, and ACT42A. Peak D is probably the ACT57A isoform (spot I).

Approximately 30% of extracted fly actin is the ACT88F-III isoform. A single column run produces 250 µg of ACT88F-III isoform in a 1-2-ml volume. However, significant amounts are lost due to actin polymerization on the column. By pooling successive runs, 1-mg quantities of WT and E93K actins were attainable. Nanoelectrospray mass spectroscopy of protein eluting in peak B shows only a single species with a mass of 41,477 ± 4 Da. This is 44 Da less than the calculated mass of fully processed ACT88F and is accounted for by the absence of the N-terminal acetyl group normally found in higher eukaryotic class 2 actins. This has been confirmed by N-terminal sequencing.³

In Vitro Motility-- Under standard motility assay conditions (25 mM imidazole-HCl, 25 mM KCl, 4 mM MgCl₂, 1 mM EGTA, 2 mM ATP, pH 7.4, 23 °C) WT actin filaments moved at a velocity of 3.7 ± 0.24 µm/s, whereas E93K filaments showed no signs of interacting with the substrate surface. To promote binding and movement of E93K filaments the in vitro motility assays were performed at lower salt concentrations, where both WT and E93K filaments displayed smooth, stable sliding.

The ionic strength dependence of WT and E93K in vitro motility was investigated (Fig. 3a). As ionic strength was raised, the velocities of WT and E93K filaments increased. At higher salt concentrations a large proportion of bound filaments were released from the surface, but those that remained attached displayed fast, slightly unstable movement. At the critical "wash-off" point, all of the filaments had released from the surface. We define the wash-off point as the lowest salt concentration at which actin filaments exhibit no interaction with the substrate surface. The wash-off points for WT and E93K were estimated as 50 and 25 mM KCl, respectively. On average, E93K actin filaments moved 40% slower than WT over the range of salt concentrations tested (Fig. 3a). Due to the instability of movement at ionic strengths approaching the wash-off points, filament velocities could only be accurately measured at KCl concentrations 5 mM below these points.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   In vitro motility of WT () and E93K () actins at increasing salt (KCl) concentrations in the absence of methylcellulose (a) and in the presence of 0.7% methylcellulose (b).

The salt dependence of WT and E93K in vitro motility was also investigated in the presence of 0.7% methylcellulose (Fig. 3b). As ionic strengths approached wash-off concentrations seen in assays without methylcellulose (Fig. 3a), actin filaments continued to display smooth, stable sliding at high velocities until a "stalling point" was reached. At KCl concentrations just before this point, filaments exhibited periods of intermittent sliding, which became less frequent as the salt concentration was raised further until movement ceased and filaments stalled. In 0.7% methylcellulose and as ionic strength increased (Fig. 3b), the average E93K velocity was 32% slower than that of WT. At ionic strengths above those required for wash-off in the absence of methylcellulose (Fig. 3a) E93K filaments showed more than a 32% reduction in velocity compared with WT.

Single Molecule Mechanical Experiments-- The effect of the E93K mutation on the cross-bridge cycle was investigated in further detail in single molecule mechanical experiments using an optical tweezers transducer. The working stroke and the stiffness of single cross-bridges formed between rabbit HMM and ACT88F actins, WT and E93K, were measured. Records (Fig. 4a) show movement of one trapped bead in parallel to the filament axis versus time. Periods of reduced Brownian motion indicate cross-bridge attachment events that add additional stiffness to the system, thereby reducing Brownian motion of the trapped beads (34, 35). The apparent amplitude of individual displacement events was calculated relative to the mean bead position in the absence of cross-bridge attachment. The size of the myosin working stroke was determined from displacement histograms (Fig. 4, b and c). The working strokes measured were as follows: RSA, 5.6 nm (S.E. = 0.5 nm, n = 595); WT actin, 6.3 nm (S.E. = 1 nm, n = 218); and E93K actin, 6.8 nm (S.E. = 0.5 nm, n = 751). These values have been corrected to account for series compliance in the actin-to-bead connections. There are no significant differences in the size of the working stroke for skeletal HMM interactions between any of these actins.

View larger version (36K): [in this window] [in a new window]   Fig. 4.   Single molecule mechanical experiments. a, displacement records of single HMM molecules interacting with Drosophila E93K mutant actin (3 µM ATP). Records show position data in parallel to the filament axis (x axis) versus time at a low trap stiffness (ktrap ~0.02 pN nm1, 1-kHz sampling rate). Intervals of reduced noise indicate cross-bridge attachments. b and c, amplitude distribution of attachment events of rabbit HMM interacting with WT actin and E93K mutant actin. The amplitude of individual attachment events was calculated relative to the mean bead position in the absence of cross-bridge attachment; for details see "Experimental Procedures." d and e, stiffness measurements on rabbit HMM cross-bridges interacting with WT actin and E93K mutant actin. The stiffness was measured by applying a sinusoidal forcing function to the actin-attached cross-bridge by moving one of the laser traps (for details, see "Experimental Procedures"). The average stiffness for each attachment event was calculated from the average applied force and the average induced cross-bridge extension during the attachment event. The stiffness data are plotted as histograms: WT actin (n = 68) and E93K mutant actin (n = 59).

The stiffness of attachments between rabbit HMM and each of the actins was measured to estimate the forces produced by the actomyosin interactions. A sinusoidal forcing function was used (see Ref. 35), and the measurements were performed at low ATP (3 µM). At a time resolution of ~10 ms (100-Hz sinusoidal forcing function), we could not resolve any changes in stiffness during individual attachment events and therefore determined an average stiffness value. The stiffness data are plotted as histograms in Fig. 4, d-e. The distributions were skewed. The higher values might be explained by variation in myosin binding to the nitrocellulose substrate or attachment of multiple cross-bridges. An average stiffness value for a single attached cross-bridge was estimated from the data by fitting a single Gaussian distribution to the first peak of the distribution. There was no significant difference in cross-bridge stiffness between WT and E93K actins. The mean stiffness with WT actin was 0.67 pN nm¹ (S.D. = 0.34 pN nm¹, n = 68), and with E93K actin it was 0.59 pN nm¹ (S.D. = 0.24 pN nm¹, n = 59).

Measurement of K_d for S1 Binding of Wild-type and E93K Actin-- The affinity of the actins for S1 was examined using the stop-flow method developed by Kurzawa and Geeves (37) to use 1-µg quantities of proteins. In stop-flow titration measurements, the amplitude of the fluorescence change observed upon adding ATP to a preincubated mixture of pyr-RSA and S1 was used to estimate the fraction of pyr-RSA bound to S1. The dependence of the observed amplitudes on S1 concentration were fitted to a quadratic equation to give the K_d for pyr-RSA binding to S1 (K_d*). K_d* values were measured at 10, 25, 50, 75, and 100 mM KCl, and the average of 5-7 measurements (5-7 different batches of pyr-RSA) at each salt concentration was within a factor of 2 of the K_d* values reported by Kurzawa and Geeves (37).

In competitive titrations, the same reaction and stock solutions of pyr-RSA and S1 were used to estimate the fractions of pyr-RSA bound to S1 in the presence of increasing concentrations of unlabeled actin (WT or E93K). The unlabeled actins were preincubated with pyr-RSA and S1 before the addition of ATP in the stop-flow. Fig. 5 shows the transient for a control reaction in which 6 µM ATP was used to dissociate 15 nM pyr-RSA and 45 nM S1 at 50 mM KCl (conditions where 80% of the pyr-RSA is complexed with S1). In the presence of 100 nM WT and 200 nM E93K actins, the rate of the dissociation reactions was unaffected, but the amplitude was reduced by 66 and 24%, respectively. The dependence of the observed amplitude on the unlabeled actin concentration (Fig. 6) was fitted to a competitive binding model to give the K_d for unlabeled actin binding to S1 with values of 12 and 144 nM for WT and E93K actins, respectively, compared with 14 nM for pyr-RSA; i.e. the affinity of S1 at 50 mM KCl for pyr-RSA and WT is very similar and about 10-fold weaker for E93K. K_d values were measured at 10, 25, 50, 75, and 100 mM KCl, and the average of 5-7 measurements for RSA and WT (5-7 different batches) and of two or three measurements for E93K (two or three different batches) at each salt concentration are presented in Fig. 7. In agreement with previous results for 100 mM KCl (37), RSA bound up to a factor of 1.7 tighter than pyr-RSA and showed a similar salt dependence. RSA and WT have similar K_d values at all measured salt concentrations, while E93K shows a significantly lower affinity for S1.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Influence of the presence of unlabeled actin on the ATP-induced dissociation of pyr-RSA.S1. Shown are transients observed for the dissociation by 6 µM ATP of 15 nM pyr-RSA and 45 nM S1 (control) and preincubated with 100 nM WT and 200 nM E93K actins, respectively. The best fit single exponential is superimposed on the data, and the parameters for the control, WT, and E93K are kobs = 34.5, 34.1, and 33.0 s1 and amplitude = 0.41, 0.14, and 0.31 V respectively. Reaction conditions are as described under "Experimental Procedures" except that the salt concentration was 50 mM KCl.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Competitive stop-flow titration of ATP-induced dissociation of pyr-RSA and S1 with WT and E93K actins (protocol as for Fig. 5). Plots of the amplitude of the fluorescence change against the concentration of WT () and E93K () are shown. The best fit is superimposed on the data, and the fitted Kd values for the WT and E93K are 12 and 144 nM, respectively, compared with 14 nM for pyr-RSA.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Debye-Hückel plot to show the dependence of Kd for actin binding to S1 on ionic strength. , RSA; slope = 5.2 M0.5, intercept = 9.4. , WT; slope = 6.2 M0.5, intercept = 9.7. , E93K; slope = 11.0 M0.5, intercept = 9.7.

These assays require relatively small amounts of material. Since the affinity of pyr-RSA for S1 is known, relative affinity can be measured in competition with pyr-RSA (under conditions where the K_d of the labeled actin is very small, e.g. at an ionic strength of 10 mM KCl) with as little as 5 µg of an unlabeled novel actin.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The E93K mutation was recovered as a flightless mutant (44) whose myofibrils lack sarcomeres (45). The aberrant myofibrillar structure limited skinned fiber mechanics experiments, but the fibers go into rigor producing stiffnesses comparable with WT fibers. This and the observation of rigor cross-bridges in electron micrographs (45) indicates that E93K actin can bind myosin, but not whether there are quantitative changes in this binding.

E93K actin binds to rabbit HMM in an in vitro motility flow cell in the absence of ATP, confirming the rigor interactions. However, under standard salt conditions (25 mM KCl) E93K actin did not bind the rabbit HMM substrate. At reduced KCl concentrations, E93K actin moved but at lower velocities (~60%) compared with WT actin. As the salt concentration was increased, E93K actin eventually washed off but at a significantly lower concentration than that which released WT actin. Similar results were seen when the viscosity-enhancing methylcellulose was present. This agent reduces the rate of diffusion of the filaments away from the surface (46). Significant differences in WT and E93K velocities were measured in the presence and absence of methylcellulose over a wide range of salt concentrations, indicating that E93K affects a step independent of the presence of methylcellulose. However, at higher temperatures (30 °C), the difference between E93K and WT velocities is much reduced (47).

Single molecule, optical trap measurements of the working stroke and stiffness of rabbit HMM interacting with Drosophila WT or E93K mutant actin showed no significant differences. This is not surprising. Since the displacement is thought to be produced by conformational changes in the myosin, a changed displacement would require premature detachment of the mutant actin before the working stroke was complete. We observed no differences in the forces transmitted from the myosin through the actomyosin interface and filaments made of E93K, WT, and RSA actins. However, the interface might be stiffer than other components of the actomyosin complex. Attachment lifetimes for the three actins at low ATP concentrations were the same (data not shown). Since attachment lifetimes depend on ATP concentration, this experiment measures the rate of ATP binding to the myosin. If the actin mutation, for example, affects actin affinity for HMM, this might not be apparent in the optical trapping experiments, since it would affect mainly the detached lifetime, a difficult parameter to obtain, since many factors contribute to it.

The K_d values of rabbit S1 binding to WT and RSA actin at different salt concentrations show that both actins have similar affinities for rabbit myosin S1. The fitted slopes from the Debye-Hückel plot are 5.2 M^0.5 and 6.2 M^0.5 respectively (Fig. 7). These are not significantly different and suggest that the charge shielding effects of salt are essentially the same. Given the number of amino acid differences between ACT88F actin and RSA in the proposed myosin binding site, especially the reduction from four to three N-terminal acidic residues in ACT88F actin, a difference in affinity would not have been surprising. These results suggest that such a difference does not affect the affinity and that rabbit S1 is a suitable substrate for assessing the actomyosin binding of Drosophila mutant actins.

Comparison of the K_d values of WT and E93K actins binding to rabbit S1 shows a significant reduction in affinity for the mutant actin. This and the greater salt dependence of the E93K-S1 interaction explain the increased binding sensitivity of the E93K F-actin filaments in the in vitro motility experiments. The fitted slopes to the Debye-Hückel plot are 6.2 M^0.5 and 11.0 M^0.5, respectively (Fig. 7), suggesting that the net overall charge in the binding of WT to S1 changes in E93K. This is a large change for a single amino acid substitution from glutamate to lysine. It is likely that residue Glu⁹³ is part of a charge cluster in the secondary myosin binding site and that a lysine at this position affects the ability of this charge cluster to fully participate in myosin binding. The Debye-Hückel plots (Fig. 7) all have similar intercepts at zero salt (RSA intercept = 9.4, WT intercept = 9.7, and E93K intercept = 9.7), suggesting that at zero salt the affinity of the three actins for myosin would be very similar at <1 nM.

Rayment et al. (4) proposed a secondary myosin contact site consisting of actin residues 40 and 42, the C terminus of the 79-92 -helix, and residues in the loop from residue 92 to 100. This secondary myosin binding site is very poorly resolved but is thought likely to be predominantly ionic. The effects of the E93K mutation on in vitro motility and S1 binding support the contention that this residue is part of a charge cluster in the secondary binding site. Glu⁹³ is a highly conserved residue in 141 actin sequences (48), being invariant except for a conservative substitution to aspartate in a Leishmania major actin.

There are no amino acid differences between rabbit skeletal and Drosophila ACT88F in the vicinity of residue Glu⁹³. The side-chain of Glu⁹³ projects from the actin surface in the rabbit actin monomer structure (2). Using the molecular graphics program Syble, we modeled the Glu⁹³ to Lys substitution in the actin monomer structure. The results (Fig. 8) show that the hydrogen bonds formed in the region around Glu⁹³ are unchanged and there are no obvious clashes or restrictions on residue movement apparent in the modeled Lys⁹³ representation. It is therefore unlikely that E93K causes local conformational changes in the actin structure.

View larger version (35K): [in this window] [in a new window]   Fig. 8.   The molecular environment around the actin residue Glu93. a, a wire frame representation of an actin monomer (atomic coordinates from Ref. 2) is viewed from the front right at about 45° above the horizontal x axis. A sphere of radius 8 Å centered at about residue 93 is highlighted above the junction between subdomains I and II. b and c, wire frame representations of the 16-Å spheres around residues Glu93 and Lys93, respectively. There are no residue substitutions between rabbit and Drosophila actin in the vicinity of residue Glu93. The molecular graphics program Syble, version 6.5 (Tripos Inc., St. Louis, MO) was used to model the Glu to Lys substitution in the rabbit actin structure. Residue 93 in each sphere is shaded. Hydrogen bonds between donor and acceptor atoms are shown as dashed lines. The -carbon atoms of residue 93 and its nearest neighbors are indicated by dots. The hydrogen bonding in the Glu93 and Lys93 spheres is identical. Glu93 is a residue whose side chain projects from the actin surface. No obvious clashes or restrictions on residue movement are apparent in the modeled Lys93 representation.

The increased in vitro velocity of RSA, WT, and E93K actin filaments at increasing KCl concentrations is a common observation. It is presumably related to reduced actomyosin affinity with increased ionic strength. Homsher et al. (49) proposed that reduced filament movement at low, suboptimal salt concentrations might be due to increased drag from weak cross-bridges, but they could not confirm this. At salt concentrations above an optimum for velocity, filaments first show reduced velocities and then wash off or, in the presence of methylcellulose, stall. This suggests that as actomyosin affinity is reduced, productive interactions are eventually reduced to a point where no movement is possible, and the filaments diffuse away from the surface. The viscosity-enhancing effect of methylcellulose maintains filament movement at much higher salt concentrations by maintaining them in close proximity to the surface myosin, thereby increasing the effective actin concentration and the probability of productive actomyosin interactions.

Methylcellulose has a similar effect on WT and E93K actin filaments but does not reduce the relative differences between these two actins as regards velocity or the salt concentration at which stalling occurs. All of these results are consistent with a reduced myosin binding of both actins with increased salt but with an increased salt sensitivity for the E93K actin. These results are clearly different from those reported in Refs. 11 and 12 for charge mutants in the primary myosin binding site of yeast actin. In this case, mutants reducing charge of the N-terminal charge cluster or causing substitution of aspartate residues 24 and 25 by alanine did not exhibit in vitro motility unless methylcellulose was present, where they then displayed WT velocities. These differences suggest that the E93K mutation and these others (11, 12) affect different stages of actomyosin interactions. The work of Miller et al. (11, 12) is consistent with their charged change mutations primarily affecting initial weak binding. The different behavior of E93K actin could indicate that an actin and myosin contact is first established through the primary binding site before contact is made with the secondary binding site (on the neighboring monomer). This supports Holmes' proposal (1) that initial actin and myosin contacts probably occur through the charged residues of the N terminus. These, together with the subsequent formation of stereospecific hydrophobic interactions, are followed by a major conformational change that may involve additional structural changes at the actomyosin interface. These changes may correspond to the three events observed in the docking of myosin S1 to actin in solution kinetic measurements (50), where the first step is the formation of collision or encounter complex (strong ionic strength dependence), the first isomerization has low ionic strength dependence but is perturbed by temperature and organic solvents, and the second isomerization is a major conformational change involving both ionic and hydrophobic contacts.

What is the role of Glu⁹³ in this docking process? Our data show that E93K actin increases K_d and reduces in vitro motility compared with WT actin but has no effect on force transmission. The stop-flow and in vitro motility data with their altered salt dependence indicate that Glu⁹³ is involved in electrostatic actomyosin interactions, i.e. formation of the collision complex or step 2. In terms of the above docking process, K_d is defined by k₁/(1 + K₂)K₀k₊₁. In preliminary studies, we found that the E93K mutation changes the apparent rate constant of actin binding (K₀k₊₁) by about a factor of 2. The much larger changes observed in K_d therefore require a large change in the apparent rate constant of actin dissociation from the complex (k₁/(1 + K₂). K₀ and K₂ are expected to involve ionic interactions, and therefore the charge change in E93K is most likely to affect K₂; i.e. the effects of E93K on K_d can be explained by a small reduction in k_on (primarily K₀) and an increased k_off (K₂). K₂ is the equilibrium constant for the isomerization of the relatively weakly bound A-M state to the rigor-like A·M state. A reduction in K₂ need not change the step size or stiffness of the cross-bridge, but it is consistent with a reduction in the number or frequency of productive transitions, which would reduce in vitro motility.

Rayment et al. (4) proposed that charged residues 99 and 100 were included in the secondary myosin site. Removal of charge in double mutants by converting residues Glu⁹⁹/Glu¹⁰⁰ in either yeast actin to alanines (11, 13, 51) or in Dictyostelium actin to histidines (9) reduced in vitro motility. However, since motility was restored to WT velocities in the presence of methylcellulose, these residues may primarily affect the weak binding of myosin to actin in a similar manner to the N-terminal charged residues in the actin atomic structure. The resolution of the fit of the actin and myosin atomic structures to the EM reconstructions of S1-decorated F-actin (4, 5) is such that it is difficult to assign the involvement of residues 99 and 100 to either the primary or secondary sites with any certainty. In addition, the secondary contact is largely through a myosin loop (residues 567-578), which is unresolved in the chicken S1 atomic structure (3) and for the Dictyostelium myosin used in the latter model (5) is probably too short to reach this part of the proposed secondary binding site.

The mutant approach (9, 11-13) has already proved valuable in determining the detail of the myosin binding site. Miller et al. (13) discuss the contributions that actin mutants can make to characterizing the roles of different parts of the actomyosin interface in weak and strong binding, arguing that their results point to a clear distinction between weak and strong binding sites. Our results from E93K suggest that further distinctions are possible and that residue Glu⁹³ may be involved in the transition from the weak to rigor-like binding. Clearly, since different mutants give different responses, by applying a variety of techniques to mutant actins, we expect to learn in much more detail about the roles of specific residues during the actomyosin interaction and the sequential recruitment of different parts of myosin binding sites during the binding of actin to myosin.

	ACKNOWLEDGEMENTS

We thank Sam Clark, Ann Lawn, and Nancy Adamek for excellent technical assistance and Kenji Mizuguchi for help with the molecular graphics.

	FOOTNOTES

^* This work was supported by BBSRC, British Heart Foundation, Wellcome Trust, DFG-SFB Grant No. 394, Royal Society, and European Union Contract CEC CHRX-CT94-0606.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Recipient of a University Studentship (University of York), partially supported by the Max Planck Society.

^§ Present address: Dept. of Genetics, University of Cambridge, Downing St., Cambridge, CB2 3EH, UK.

^** Present address: Dept. of Biosciences, University of Kent, Canterbury, Kent, CT2 7NJ, UK.

To whom all correspondence should be addressed: Dept. of Biology, University of York, P.O. Box 373, York YO10 5YW, UK. Tel.: 44-1904-432826; Fax: 44-1904-432860; E-mail: jcs1@york.ac.uk.

² S. Schmitz, A. Razzaq, M. A. Geeves, and J. C. Sparrow, unpublished observations.

³ S. Schmitz, J. Clayton, U. Nongthomba, H. Prinz, C. Veigel, M. Geeves, and J. Sparrow, submitted for publication.

	ABBREVIATIONS

The abbreviations used are: IFM, indirect flight muscle; HMM, heavy meromyosin; MOPS, 4-morpholinopropanesulfonic acid; PAGE, polyacrylamide gel electrophoresis; pyr-RSA, pyrene-labeled rabbit F-actin; RSA, rabbit skeletal actin; S1, myosin subfragment 1; WT, wild-type; pN, piconewtons.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES