ARCHIVES
OF BIOCHEMISTRY
Vol. 293, No. 1, February
AND BIOPHYSICS
14, pp. llO-116,1992
Removing the Two C-Terminal Residues of Actin Affects the Filament Structure’ S&an I. O’Donoghue,’ Masao Miki, and Cristobal G. dos Remedios Muscle Research Unit, Department of Anatomy, The University of Sydney, NS W 2006, Australia
Received June 28,1991, and in revised form October 29, 1991
We define conditions under which the two C-terminal residues of actin, Cys-374 and Phe-375, can be selectively removed by proteolysis with trypsin. This modification had little effect on the secondary structure of actin detected by Fourier-transform infrared spectroscopy. However, removing these residues caused small but significant decreases in the critical concentration of actin, in its ability to activate myosin ATPase, and in its interaction with tropomyosin and troponin. Removing residues 374-375 caused dramatic changes in the actin filament as seen by electron microscopy. The filaments had a much greater and more irregular curvature and were intertwined into disordered multifilament bundles. Removing 374-375 also significantly lowered the flow viscosity of fllamentous-actin solutions. These data suggest an increase in the flexibility and fragility of the filament, supporting the idea that the C-terminus forms one of the o 1992 ACTmajor intermonomer contacts in the filament. demic Press, Inc.
Actin is a highly conserved protein believed to be present in large amounts in nearly all eukaryotic cells. It has many roles in the cell, but its major function is the interaction with myosin which results in movement of subcellular organelles and whole organisms. At low ionic strength, actin exists in a globular form, called G-actin,3 i This work was supported in part by grants from the National Health and Medical Research Council of Australia. * To whom correspondence should be addressed at Department of Biochemistry, The University of Sydney, NSW 2006, Australia. 3 Abbreviations used: actin-,, actin residues l-373; ADP-F-actin, Factin assembled from ADP-containing monomers; ATP-F-actin, F-actin assembled from ATP-containing monomers; F-actin, filamentous actin; FTIR, Fourier-transform infrared spectroscopy; G-actin, globular actin; IAEDANS, N-iodoacetyl-N’-(l-sulfo-5-naphthyl)ethylene~~ine; IAF, 1,5-iodoacetamidofluorescein; IAF-actin, actin labeled at Cys-374 with IAF; pyrene, N-(1-pyrenyl)iodoacetamide; pyrene-actin, actin labeled at Cys-374 with pyrene; RSS, residual sum of squares; Sl, myosin subfragment 1; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TLCK, N-p-tosyl-L-lysine chloromethyl ketone; Pipes, 1,4-piperasinediethanesulfonic acid; EGTA, ethylene glycol bis(&aminoethyl ether) N,N’-tetraacetic acid.
which has a single nucleotide binding site usually occupied by ATP. Upon addition of salts, G-actin self-assembles into a two-start helical polymer called filamentous actin (F-actin). Soon after the polymer is formed, the bound ATP is hydrolyzed, resulting in stiffening and strengthening of the filament (1). This suggests that the energy released by hydrolysis is used either to form or to strengthen intermonomer bonds; however, it is not yet clear which bonds are affected. The penultimate residue in the primary structure of actin, Cys-374, is the most reactive cysteinyl in most actin isoforms. Consequently, this residue has been used extensively for spectroscopic probe studies (e.g. (2-6)) and cross-linking (e.g. (7,8)). These studies revealed that the C-terminal region of actin is very close to another monomer in the filament (6-S). Based on the amino acid sequence similarity between the C-terminal region of actin and proteins which cap or sever the actin filament, Tellam et al. (9) predicted that the region of homology, residues 330 to 375, forms an important interfilament contact site which is mimicked by the filament-severing proteins. This agrees remarkably well with the atomic-resolution actin models published recently (10,ll). In the X-ray structure of G-actin, this region corresponds to a small, well-defined region located on the bottom of subdomain 1, following the nomenclature of Kabsch et aZ. (10). The proposed filament model based on this structure (11) places most of this region in direct contact with a neighboring monomer. Labeling of other actin cysteinyls has been hampered by the high reactivity of Cys-374. Although Cys-374 can be blocked with sulfydryl-specific probes, these may also partially block the other cysteinyls. Thus we were interested in removing this residue by proteolysis to facilitate the labeling of other cysteinyls. Previous investigators have used proteolysis to modify this region of actin. Trypsin normally digests actin at several sites (61-62, 68-69, and 373-374), but the 373374 site is the more reactive (12). Suck et al. (13) found that mild trypsin digestion of the G-actin:DNase I complex removed only residues 373-375. This mild trypsin treatment improved the quality of the crystals obtained 0003-9861/92
110
$3.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction
in any form reserved.
REMOVAL
OF ACTIN
C-TERMINAL
RESIDUES
(13) and thus was used in deriving the recent atomicresolution structure (10). Carboxypeptidase A has been used to remove Phe-375 and then Cys-374 from actin (14,15). Removing Phe-375 alone has little effect on actin or on its interaction with myosin. However, the conditions for removing this cysteine residue have not previously been defined, and the effect of this proteolytic cleavage on actin has not been characterized. We found that Cys-374 residues were removed from all monomers after 18 h digestion with 1 mol carboxypeptidase A per 10 mol of G-actin (S. I. O’Donoghue, unpublished data). However, we decided to use trypsin rather than carboxypeptidase A for these experiments since it is cheaper and easy to obtain in the quantities needed. To date, there has been no systematic investigation of the effect of removing residues 374 and 375 on the structure and functions of actin. To be a useful technique, it is important that this modification does not greatly alter the conformation of actin or its ability to polymerize and interact with myosin. If there is an effect on any of these properties, this would have important implications for accepting the X-ray crystal structure of actin. Our aim in this paper is to define conditions for removing these residues using trypsin and to characterize the effects of this modification on the functions of actin. We found that removing these residues caused only subtle changes in the secondary structure of actin and its interaction with other proteins. However, we found major changes in the filament shape as seen by electron microscopy and in the viscosity of F-actin. These changes suggest that the normally rigid actin filament is more flexible when these residues are removed. MATERIALS
AND METHODS
Reagents. Bovine trypsin (type XI) and TLCK (1-chloro-3-tosylamido-7-amino-Z-heptanone hydrochloride (N-p-tosyl-L-lysinechloromethyl ketone)) were purchased from Sigma Chemical Co. Carboxypeptidases A and B were obtained from Boehringer-Mannheim Biochemicals. Pyrene (N-(l-pyrenylhodoacetamide), IAEDANS (N-iodoacetyl-iV’-(l-sulfo-5-naphthyl)ethylenediamine), and IAF (1,5-iodoacetamidofluorescein) were purchased from Molecular Probes, Inc. Actin was prepared from rabbit skeletal muscle according to the method of Spudich and Watt (16). ADP-G-actin was prepared as described by Pollard (17) using hexokinase from Boehringer-Mannheim. Pyrene-actin (labeled at Cys-374 with pyrene) was prepared according to the method of Kouyama and Mihashi (3). The molar ratio of attached pyrene to actin was 0.95. IAF-actin (labeled at Cys-374 with IAF) was prepared according to the method of Wang and Taylor (18). The molar ratio of attached IAF to actin was usually 0.6. Myosin was prepared according to the method of Tonomura et al. (19). Myosin subfragment 1 (Sl) was prepared by chymotryptic digestion of rabbit skeletal myosin according to the method of Weeds and Pope (20). Sl was either freeze-dried in the presence of 0.1 M sucrose and stored at -2O’C or kept in 50% glycerol at -20°C. Tropomyosin, troponin, and regulated F-actin were prepared as described previously (21). IAEDANS-labeled tropomyosin (1.4 mol IAEDANS per mole of tropomyosin) was prepared as described previously (22, 23). The concentration of unmodified G-actin was determined using an extinction coefficient of 26.6 X lo3 M-’ cm-i at 290 nm (24). The Sl concentration was determined using an extinction coefficient of 82.5 X
AFFECTS
FILAMENT
STRUCTURE
111
lo3 Me1 cm-’ at 280 nm (20). Optical absorbence measurements were made with a Philips PU 8800 spectrophotometer. Protein concentrations were also determined by the Bradford method using the Pierce Coomassie blue G-250 protein assay reagent. The molecular masses of Sl and actin were taken to be 110 kDa (25) and 42.3 kDa (26), respectively. Trypsin digestion. F-actin (approximately 50 PM) was digested with trypsin (1 mol of trypsin to 5 mol of actin) in F-buffer (2 mM MgClz, 0.2 mM ATP, 3 mM NaN,, 20 mM KCl, and 20 mM Tris-HCl, pH 7.6) for 15-25 min at room temperature (about 22°C). The reaction was stopped by adding a IOO-fold molar excess of TLCK over trypsin. The actin was then taken through a polymerization/depolymerization purification cycle. This involved pelleting F-actin in a centrifuge at 100,OOOg (g = 9.8 m s-i) for 90 min at 20°C resuspending in F-buffer, centrifuging again, resuspending the pellet in G-buffer (0.2 mM CaCl,, 0.2 mM ATP, 3 mM NaN,, and 2 mM Tris-HCl, pH 8.0), dialyzing exhaustively against G-buffer, and, finally, clarifying the solution of insoluble impurities by centrifuging (100,OOOgfor 90 min at 4’C). The total amount of protein obtained from this digestion was about 60% of the starting amount or less, depending on the activity of the trypsin. The above method was designed to produce a “clipped” actin missing residues 374-375. For each clipped actin preparation, two methods were used to check the extent of 374-375 removal. The first was to add a lofold molar excess of IAF to the clipped actin (in G-buffer). After 3 h the mixture was dialyzed overnight versus F-buffer to remove excess label and then centrifuged (100,OOOgfor 90 min at 20°C). Samples from the pellet were used to detect the presence of IAF bound to actin with 12% w/v sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) according to the procedure of Laemmli (27). The second method used was to add an equimolar amount of Cu+* to the clipped actin and measure the change in optical absorption at 347 nm. While actin does not normally absorb light of this wavelength, upon binding copper an absorbence peak appears centered at 347 nm with an extinction coefficient of 3.0 X lo3 M-’ cm-’ (28). Copper binding is lost upon removal of Phe-375 (29); hence, the extent of 375 removal can be determined from the loss of 347-nm absorbance. Carboxypeptidase B digestion. The clipped actin produced from the trypsin digest was further digested with carboxypeptidase B for 18 h at 4”C, polymerized by the addition of 10X F-buffer, and pelleted by centrifuging as described above. The supernatant was cleared of protein by adding pechoric acid (to 5% w/v) and analyzed for the presence of amino acids by high-performance liquid chromatography using a precolumn derivatization method described by Haynes et al. (30). Characterization of the actin fragment. Fourier-transform infrared spectroscopy (FTIR) was used to estimate the secondary structure features of actin as described by Susi and Byler (31). Spectra were obtained for clipped and unmodified actin in G-buffer using a DIGILABS FTS20/80 spectrophotometer equipped with a HgCdTe detector. Fourier transforms and analysis of these spectra were done using DIGILAB software. The critical concentration for clipped actin was measured by copolymerizing with pyrene-actin as described by Pollard (32). In this method, the large increase in pyrene-actin fluorescence upon polymerization is used to measure the extent of polymer formation. Samples were placed in l-cm quartz cuvettes (stirred and temperature regulated at 15’C) and fluorescence intensities were measured using an SLM 8000 spectrofluorimeter operating in ratio mode. The superprecipitation of actin and myosin was studied using lo-ml samples of F-actin (0.02 mg/ml) and myosin (0.2 mg/ml) in 25 mM KCl, 1 mM MgCl,, and 20 mM Pipes (pH 7.05). These were thoroughly mixed in a Teflon-glass homogenizer and transferred to a 2 X 2-cm glass cuvette (continuously stirred and temperature regulated at 23’C). A total of 10 ~1 of 10 mM ATP was added to start superprecipitation. The extent of superprecipitation was determined from the change in light scattering at 650 nm. The rate was determined by the time required to reach half of the maximum change in scattering. The maximal change varied by about 10% between samples. ATPase measurements were carried out with a Radiometer pH-stat apparatus (ABU 12; TTT 2; SBR 3; PHA 943) at pH 8.0 using 2 mM
112
O’DONOGHUE,
MIKI,
KOH as a titrant. The conditions were 0.05mg/ml Sl, 30 mM KCl, 2.5 mM ATP, and 1 mM Tris-HCl (pH 8.0) at 23°C. The concentration of actin was varied. The reaction was started by adding Sl to the 2-ml samples. The ATPase activation of regulated actin was measured by preparing a mixture of Sl, actin, tropomyosin, and troponin at 0.05 mg/ml, 0.25 mg/ml, 0.086 mg/ml, and 0.087 mg/ml, respectively. The measurements were done as above but in the presence of Ca+s (0.1 mM CaCl,) or in the absence of free Ca+’ (with 2 mM EGTA). The association of F-actin (1.7 mg/ml) with IAEDANS-tropomyosin (0.2 mg/ml) in F-buffer was examined by cosedimentation at - 180000g for 30 min in a Beckman airfuge. The concentration of labeled tropomyosin in the supernatent (= [Tm]r, the concentration of unbound tropomyosin) was determined by measuring the fluorescence intensity at 480 nm with the SLM fluorimeter. The excitation wavelength was 360 nm. Since the total concentrations of actin and tropomyosin ([A], and [Tm],) are known, the association constant (K.) can be determined from [Tm]r using the formula:
1Tmlf = (lTm1, - [Al, - K’ + v(K;’
+ [A], - [Tm],)2 + 4[Tm]JK,)/2.
This was derived by substituting the equations [Tmlf = [Tm], - [A.Tm] and [AIf = [A], - [A.Tm] into the definition of K,,, [A.Tm] K’ = [Tmlf[Alf
DOS REMEDIOS
a
1
b
2
1
2
FIG. 1. SDS-PAGE of trypsin-digested actin treated with IAF. The gel in Fig. la shows trypsin-digested actin (lane 1) and undigested actin (lane 2) stained with Coomassie blue to show total protein. Both samples had been treated with IAF as described under Materials and Methods. Figure lb shows the same gel illuminated with a uv lamp prior to staining. The fluorescent band corresponding to undigested actin is due to IAF bound to Cys-374. Clearly, trypsin-digested actin has not been labeled by IAF, indicating that the limited proteolytic digestion resulted in essentially complete removal of Cys-374.
’
where [A.Tm] is the concentration of the actin/tropomyosin complex, and [Tmlf and [AIf are the concentrations of unbound tropomyosin and actin. A residual sum of squares (RSS) function was defined which measured the differences between the measured and the calculated [Tmlf values. The value of K. which best fitted the RSS function was found by nonlinear minimization. F-actin was diluted to 0.05 mg/ml and quickly Electron nicroscopy. transferred onto carbon-coated parlodion-filmed grids. After blotting off excess sample, the grids were washed with 0.1 M ammonium acetate and stained with 1% uranyl acetate for 30 s. The samples were viewed in a Philips EM 301 electron microscope operated at an accelerating voltage of 80 kV. Viscosities were measured using an Ostwald Viscosity mmsurements. capillary viscometer (Cannon Corp.) with a sample size of 0.5 ml and a flow time for water of 25 s at 25°C. The specific viscosity (s,) was calculated from
%p =
AND
(
flow time of actin sample _ 1 flow time for water 1’
RESULTS Trypsin Digestion The trypsin digestion procedure and subsequent polymerization/depolymerization cycle outlined under Material and Methods resulted in a “fragment” which ran as a single band on SDS-polyacrylamide gels with the same apparent molecular weight as unmodified actin (see Fig. 1). This indicated that most residues were still intact; hence we called this “clipped” actin. This clipped actin had lost all ability to bind IAF as determined by fluorescence SDS-PAGE (see Fig. l), indicating that most of the Cys-374 had been removed. To quantitate the extent of this removal, the change in optical absorbence at 347 nm was measured upon adding copper. Only a small
change in absorbence was detected for all clipped-actin preparations (less than 5% of the absorbence change for unmodified actin), indicating that over 95% of the actin in these preparations was missing its C-terminal residue (Phe-375). The Edman method was used in an attempt to degrade the trypsin-treated fragment with phenyl isothiocyanate. The fragment would not degrade, indicating that the Nterminus was blocked as in unmodified actin. This suggests that the first residue of the fragment is the acetylated Asp-l found in untreated actin. Carboxypeptidase
B Digestion
The clipped actin was further digested using carboxypeptidase B to identify the residues at the C-terminus. Using HPLC to detect amino acids released by this digestion, only two were detected; lysine (0.89 mol per mole of actin) and arginine (0.84 mol per mole of actin). Since residues 372 and 373 in the actin sequence are arginine and lysine, respectively, this result confirms that the trypsin digestion removed only residues 374 and 375. Thus, on the basis of the known trypsin cleavage sites on actin (12) and the known amino acid sequence of actin, we conclude that the limited trypsin digestion removed only residues 374-375. The clipped actin produced by the trypsin digest, residues l-373, is hereafter denoted actin-z to indicate removal of the last two residues. Characterization
of the l-373 Actin Fragment
By measuring the optical absorption of G-actin at 290 nm and comparing it to the concentration calculated by the Coomassie blue assay, we determined that G-ac-
REMOVAL
OF ACTIN
C-TERMINAL
RESIDUES
AFFECTS
FILAMENT
113
STRUCTURE
Using the coprecipitation method outlined under Materials and Methods, the affinity constant for IAEDANStropomyosin binding was determined to be K, = 99 X lo3 M-’ for F-actin and K,, = 196 X lo3 M-’ for native Factin (see Fig. 3). Electron Microscopy
l/lactinl
(JIM’ )
plot of Sl ATPase activity versus actin FIG. 2. Double-reciprocal concentration. The maximal rate of Sl ATP hydrolysis (i.e., the inverse of the ordinate intercept) is 2.8 pmol Pi min-’ mg-’ in the presence of actin (0) and 3.5 pmol Pi min-’ mg-’ in the presence of unmodified actin (m). The rate for Sl alone was 0.04 pmol Pi min-’ mg-‘. Both graphs intersect the abscissa at the same point which indicates that removing 374-375 did not affect the affinity of actin binding to Sl.
tin-z had the same extinction coefficient as unmodified G-actin. The secondary structure predicted from FTIR for Gactin-z was 31% a-helix, 47% &sheet, and 22% random coil. This compares with 32% a-helix, 43% P-sheet, and 25% random coil predicted for control G-actin. These differences (~4%) are probably not significant since the accuracy of the FTIR technique in predicting secondary structure is from 0 to 5% (31). We therefore conclude that the secondary structure of actin was not greatly altered by removing residues 374 and 375. The critical concentrations of unmodified actin and actin2 in F-buffer were determined by copolymerization with 4% pyrene-actin at six different actin concentrations (between 0 and 5 PM). The critical concentration for unmodified actin was 0.7 f 0.2 PM, consistent with the value reported by others (32). The critical concentration for actin was 0.27 +_0.04 PM (these values and standard errors were calculated from a linear regression of six points). We found no significant difference in the extent of superprecipitation of myosin with F-actin and with F-actin-, as determined from the 650-nm absorbence. The time for half-maximal change in absorbence varied from 1 to 3 s for both samples. Figure 2 shows a double-reciprocal plot of Sl ATPase activity versus concentration of actin. Removing residues 374 and 375 did not affect the affinity of actin binding to Sl, but reduced the maximal activation of the Sl ATPase from 3.5 to 2.8 pm01 Pi min-’ mg-‘. These ATPase measurements were repeated with regulated actin and actin (i.e., complexed with tropomyosin and troponin). For both actin samples, activation of Sl ATPase was not inhibited in the presence of Ca+‘. However, in the absence of Caf2, the activation by F-actin-2 was inhibited by about 50% while the activation by Factin was inhibited by approximately 80%.
At high magnification, unmodified actin filaments (Fig. 4c) are long and straight with a rather low curvature and have a uniform diameter. The filaments are mostly separate from each other, although sometimes several filaments (usually just two) run parallel to form straight multifilament cables. In contrast, the filaments of actin at the same magnification (Fig. 4d) have a kinked appearance, and a greater and more irregular curvature. These filaments appear to have a less uniform diameter, although this is difficult to quantitate since they are usually twisted around each other. The electron micrographs we obtained for F-actin-, look very similar to those published by Janmey et al. (1) for filaments assembled from ADP-containing monomers (hereafter denoted ADP-Factin, depending on the nucleotide bound to the monomer). Normally, F-actin is assembled from monomers containing ATP (ATP-F-actin). At lower magnification, the difference between filaments of unmodified actin (Fig. 4a) and actin (Fig. 4b) is dramatic. The filaments of actin are mostly entwined together into multifilament bundles. Figure 4d shows one of these bundles (going from the bottom left to the top right corner) at higher magnification. The arrangement of filaments in these bundles is very irregular. Viscosity Measurements
Figure 5 shows the specific viscosity as a function of actin concentration for ATP-F-actin, ADP-F-actin, and
Fluorescence intensity (arbitary units) 0.4.. 0.2.. 1
1
2
3
4
5
6
FIG. 3. Binding of IAEDANS-tropomyosin to F-a&n-, (0) and native F-actin (m). These data show the loss of fluorescence intensity in the supernatent due to coprecipitation with increasing molar ratios of Factin. These data were used to find the K,, for the association of tropomyosin with actin as described under Materials and Methods. The curves are calculated from the best-fitting values of K,--99 X lo3 M-’ for Factin-* (RSS = 0.27%) and 196 X lo3 M-’ for native F-actin (RSS = 0.572).
114
O’DONOGHUE,
MIKI,
AND
DOS REMEDIOS
REMOVAL
OF ACTIN
C-TERMINAL
RESIDUES
F-actin-, solutions. At high concentrations (>50 PM), the viscosity of ADP-F-actin and F-actinpa was about half that of ATP-F-actin. This difference in viscosity could easily be seen by slowly inverting the solution and watching it flow. The data in Fig. 5 are for a single preparation of actinep, but the same reduction in viscosity was measured in all F-actinpz preparations.
AFFECTS
FILAMENT
115
STRUCTURE
7
l8 T
Specific viscosity
1o
DISCUSSION
8
The limited trypsin digestion resulted in an actin fragment consisting of residues l-373. The evidence for this conclusion is that the trypsin-treated actin lost its ability to bind IAF, indicating that Cys-374 had been removed, but had the same extinction coefficient and apparent molecular weight as unmodified actin. Furthermore, the Nterminus was blocked as in unmodified actin and digestion with carboxypeptidase B released lysine and arginine corresponding to residues 373 and 372 in the actin sequence. This conclusion contrasts with the finding of Suck et al. (13), who found that trypsin removed Lys-373 as well as residues 374-375. This disparity is not surprising since their digestion conditions were different to ours; they digested the G-actin:DNase I complex, while the digestions described here were done with F-actin. It is quite plausible that Lys-373 is protected from cleavage in F-actin, but more exposed in G-actin. Removing the last two residues did not greatly alter the secondary structure of actin or its ability to polymerize. This suggests that the tertiary structure was mostly unaffected by this modification. We found a significant difference in the critical concentration required for polymerization to occur. Other modifications of this region also lower the critical concentration (33). A small but significant difference was also found in the maximal rate of Sl ATPase activation. This is consistent with previous reports that the C-terminal region is involved in the interaction of actin and myosin (3, 34) but is not critical for this interaction (14). The regulation of the actomyosin interaction by tropomyosin and troponin appeared to be slightly impaired by this modification. However, this may be partially explained by later experiments which showed that the affinity of F-actinez binding to tropomyosin was lower than that for native F-actin. This reduced binding affinity may be due to steric blocking of F-actinMp by the formation of the filament bundles seen in Figs. 4b and 4d. The variable curvature of F-actinqa filaments, as seen under the electron microscope at high magnification, indicates that these filaments were more flexible than unmodified actin filaments. Janmey et al. published very
6
0
20
40
60
80
100
[actin] (KM) FIG. 6. Flow viscosity measurements. The specific viscosity is plotted as a function of actin concentration for ATP-F-actin (m), ADP-F-actin (X), and F-actin-, (0).
similar electron micrographs of actin polymerized from monomers containing ADP rather than ATP (1). They also concluded that the difference was due to increased filament flexibility. Normal actin filaments can form interfilament crosslinks by self-association (35) which result in the parallel cables as seen in the electron microscope (Fig. 4~). The basis of this contact is probably electrostatic attraction. Removing residues 374-375 appears to increase the extent of filament cross-linking. It is possible that this is caused by an extra, specific, filament-filament interaction, perhaps involving the three basic residues 371-373 which may be more free to interact with acidic residues on adjacent filaments. A more likely explanation is that the electrostatic interaction, which results in parallel cables of unmodified actin, is complemented by the mechanical intertwining of the filaments which results from the increased flexibility of F-actin-,. This may explain the emergence of the disordered filament bundles seen at low magnification in the electron microscope. The decrease in flow viscosity may be partially accounted for by the bundling of F-actinez. However, the viscosity decrease may also be explained by filament breakage. Janmey et al. (1) found that mechanical breakdown of ADP-F-actin occurs even when very low strain (less than 1%) is applied to the solution. In contrast, ATPF-actin will not break down until a relatively large strain is applied (over 10%). We could not determine the strain
of F-actin-, and unmodified F-actin. At low magnification, it is apparent that F-actin-, filaments (b) aggregate into FIG. 4. Ultrastructure bundles, while unmodified F-actin filaments (a) form a loose network. At higher magnification it is clear that the F-actin+ filaments (d) are more entwined about one another and more variable in curvature than F-actin filaments (c). All samples were prepared under identical conditions. Scale bars are 0.50 pm.
116
O’DONOGHUE,
MIKI,
amplitude exerted by the viscometer on the solution; however, it was likely to be greater than 1%. Hence, the decrease in viscosity of ADP-F-actin measured in this paper is probably due to filament breakage. Since F-actin-z has a similar viscosity decrease, actin-z filaments may also be more fragile than ATP-F-actin; this agrees with the results obtained by Stournaras et al. (36), who found that chemical modification of Cys-374 decreased the mechanical stability of F-actin. However, this can not be established without further, more sophisticated, rheological measurements as done by Janmey et al. (1). While we cannot be certain of the cause of the apparent increase in flexibility and fragility of F-actin-P, it seems likely to result from weakening of intermonomer bonds. This idea is supported by the filament model proposed by Holmes et al. (ll), in which one of the major intermonomer contacts is formed by residue 374 (and residues 166169) with residues 41-45 on the next monomer. Other support comes from a variety of physicochemical techniques: Elzinga and Phelan (7) and Sutoh (8) showed that Cys-374 can be cross-linked to Lys-191 on an adjacent monomer. They concluded that the distance between these two residues must be 0.8-1.4 nm. The NMR probe studies of Barden et al. (6) show that Cys-374 experiences a large decrease in mobility following polymerization, consistent with it being close to an intermonomer binding site. These studies suggest that removing residues 374375 may disrupt an important intermonomer contact, causing a weakening of the filament which may account for the increased filament flexibility which we observed. Since the rheological and ultrastructural effects of removing residues 374 and 375 are so similar to those observed for ADP-F-actin, we may speculate that removing residues 374 and 375 somehow promotes hydrolysis of the ATP bound to G-actin so that filaments are then assembled from ADP monomers which would thus be more flexible. Alternatively, Cys-374 may be involved in one of the bonds which are strengthened by ATP hydrolysis after polymerization. However, these explanations seem unlikely since both X-ray crystallography (10) and fluorescence resonance-energy transfer measurements (2, 4, 5) place Cys-374 a long way (about 3 nm) from the nucleotide site of actin. In summary, our data show that removing residues 374 and 375 did not greatly alter most properties of actin but increased the flexibility of the actin filaments and promoted filament bundling. This supports the idea that the C-terminus is directly involved in actin-actin contact in the filament. ACKNOWLEDGMENTS We gratefully acknowledge the help of Drs. Ernest and Dallas MacFarlane from the University of Sydney, who made the FTIR measurements, and Dr. Andrew Gooley from Macquarie University, who did the amino acid analysis.
AND
DOS REMEDIOS
REFERENCES 1. Janmey, P. A., Hvidt, S., Oster, G. F., Lamb, J., Stossel, T. P., and Hartwig,
J. H. (1990) Nature 347,95-99.
2. Miki, M., and Mihashi, K. (1977) Biochim. Biophys. Acta 633,163172. 3. Kouyama, T., and Mihashi, K. (1981) Eur. J. B&hem. 114, 3338. 4. Cheung, H. C., and Liu, B. M. (1984) J. Muscle Res. Cell Motil. 5, 65-80. 5. Miki, M., and Wahl, P. (1984) B&him. Biophys. Acta 786, 188196. 6. Barden, J. A., Phillips, L., Cornell, B. A., and dos Remedios, C. G. (1989) Biochemistry
28,5895-5901.
7. Elzinga, M., and Phelan, J. J. (1984) Proc. Natl. Acad. Sci. USA 81,6599-6602. 8. Sutoh, K. (1984) Biochemistry 23,1942-1946. 9. Tellam, R. L., Morton, D. J., and Clarke, F. M. (1989) Trends Biochem. Sci. 14,130-133. 10. Kabsch, W., Mannherz, H. G., Suck, D., Pai, E. F., and Holmes, K. C. (1990) Nature 347,37-44. 11. Holmes, K. C., Popp, D., Gebhard, W., and Kabsch, W. (1990) Nature
347,44-49. 12. Jacobson, USA 73, 13. Suck, D., Sci. USA 14. Field, V.
G. R., and Rosenbusch, J. P. (1976) Proc. N&l. Acad. Sci. 2742-2746. Kabsch, W., and Mannherz, H. G. (1981) Proc. N&l. Acad. 78,4319-4323. L. S., and Bowen, W. J. (1968) Arch. B&hem. Biophys.
127,59-64. Malm, B. (1984) FEBS L.&t. 173,399-402. Spudich, J. A., and Watt, S. (1971) J. Bid. Chem. 216,4666-4871. Pollard, T. D. (1984) J. Cell Biol. 99, 769-777. Wang, Y., and Taylor, D. L. (1980) J. Hi&o&m. Cytochem. 28, 1198-1206. 19. Tonomura, Y., Appel, P., and Morales, M. F. (1966) Biochemistry 15. 16. 17. 18.
5,515-521. 20. Weeds, A. G., and Pope, B. (1977) J. Mol. Biol. 111, 129-157. 21. Miki, M. (1987) Eur. J. Biochem. 164, 229-235. 22. Lamkin, M., Tao, T., and Lehrer, S. S. (1983) Biochemistry 22, 3053-3058. 23. Miki, M., and Hozumi, T. (1991) Biochemistry 30,5625-5630. 24. Houk, T. W., and Ue, K. (1974) And. Biochem. 62,66-74. 25. Margossian, S. S., Stafford, W. F., and Lowey, S. (1981) Biochemistry 8,2151-2155. 26. Collins, J. H., and Elzinga, M. (1975) J. Biol. Chem. 250, 59155920.
27. Laemmli, U. K. (1970) Nature 227,680-685. 28. Lehrer, S., Nagy, B., and Gergely, J. (1972) Arch. Biochem. Btiphys. 150,164-174. 29. Drabikowski, W., Lehrer, S., Nagy, B., and Gergely, J. (1977) Arch. Biochem. Biophys. 181,359-361. 30. Haynes, P. A., Sheumack, D., Kibbly, J., and Redmond, J. W. (1991) J. Chromatogr. 640, 177-187. 31. Susi, H., and Byler, D. M. (1986) in Methods in Enzymology (Hirs, 32. 33. 34. 35. 36.
C. H. W., and Timasheff, S. N., Eds.), Vol. 130, pp. 290-327, Academic Press, San Diego. Pollard, T. D. (1984) J. Muscle Res. Cell Motil. 4, 253-262. Tait, J. F., and Frieden, C. (1982) Biochemistry 24,6046-6053. Brauer, M., and Sykes, B. D. (1986) Biochemistry 25,2187-2191. Griffith, L. M., and Pollard, T. D. (1982) J. Biol. Chem. 267,91359142. Stournaras, C., Drewes, G., Blackhohn, H., Merkler, I., and Faulstich, H. (1990) Biochim. Biophys. Acta 1037,66-91.
OF BIOCHEMISTRY
Vol. 293, No. 1, February
AND BIOPHYSICS
14, pp. llO-116,1992
Removing the Two C-Terminal Residues of Actin Affects the Filament Structure’ S&an I. O’Donoghue,’ Masao Miki, and Cristobal G. dos Remedios Muscle Research Unit, Department of Anatomy, The University of Sydney, NS W 2006, Australia
Received June 28,1991, and in revised form October 29, 1991
We define conditions under which the two C-terminal residues of actin, Cys-374 and Phe-375, can be selectively removed by proteolysis with trypsin. This modification had little effect on the secondary structure of actin detected by Fourier-transform infrared spectroscopy. However, removing these residues caused small but significant decreases in the critical concentration of actin, in its ability to activate myosin ATPase, and in its interaction with tropomyosin and troponin. Removing residues 374-375 caused dramatic changes in the actin filament as seen by electron microscopy. The filaments had a much greater and more irregular curvature and were intertwined into disordered multifilament bundles. Removing 374-375 also significantly lowered the flow viscosity of fllamentous-actin solutions. These data suggest an increase in the flexibility and fragility of the filament, supporting the idea that the C-terminus forms one of the o 1992 ACTmajor intermonomer contacts in the filament. demic Press, Inc.
Actin is a highly conserved protein believed to be present in large amounts in nearly all eukaryotic cells. It has many roles in the cell, but its major function is the interaction with myosin which results in movement of subcellular organelles and whole organisms. At low ionic strength, actin exists in a globular form, called G-actin,3 i This work was supported in part by grants from the National Health and Medical Research Council of Australia. * To whom correspondence should be addressed at Department of Biochemistry, The University of Sydney, NSW 2006, Australia. 3 Abbreviations used: actin-,, actin residues l-373; ADP-F-actin, Factin assembled from ADP-containing monomers; ATP-F-actin, F-actin assembled from ATP-containing monomers; F-actin, filamentous actin; FTIR, Fourier-transform infrared spectroscopy; G-actin, globular actin; IAEDANS, N-iodoacetyl-N’-(l-sulfo-5-naphthyl)ethylene~~ine; IAF, 1,5-iodoacetamidofluorescein; IAF-actin, actin labeled at Cys-374 with IAF; pyrene, N-(1-pyrenyl)iodoacetamide; pyrene-actin, actin labeled at Cys-374 with pyrene; RSS, residual sum of squares; Sl, myosin subfragment 1; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TLCK, N-p-tosyl-L-lysine chloromethyl ketone; Pipes, 1,4-piperasinediethanesulfonic acid; EGTA, ethylene glycol bis(&aminoethyl ether) N,N’-tetraacetic acid.
which has a single nucleotide binding site usually occupied by ATP. Upon addition of salts, G-actin self-assembles into a two-start helical polymer called filamentous actin (F-actin). Soon after the polymer is formed, the bound ATP is hydrolyzed, resulting in stiffening and strengthening of the filament (1). This suggests that the energy released by hydrolysis is used either to form or to strengthen intermonomer bonds; however, it is not yet clear which bonds are affected. The penultimate residue in the primary structure of actin, Cys-374, is the most reactive cysteinyl in most actin isoforms. Consequently, this residue has been used extensively for spectroscopic probe studies (e.g. (2-6)) and cross-linking (e.g. (7,8)). These studies revealed that the C-terminal region of actin is very close to another monomer in the filament (6-S). Based on the amino acid sequence similarity between the C-terminal region of actin and proteins which cap or sever the actin filament, Tellam et al. (9) predicted that the region of homology, residues 330 to 375, forms an important interfilament contact site which is mimicked by the filament-severing proteins. This agrees remarkably well with the atomic-resolution actin models published recently (10,ll). In the X-ray structure of G-actin, this region corresponds to a small, well-defined region located on the bottom of subdomain 1, following the nomenclature of Kabsch et aZ. (10). The proposed filament model based on this structure (11) places most of this region in direct contact with a neighboring monomer. Labeling of other actin cysteinyls has been hampered by the high reactivity of Cys-374. Although Cys-374 can be blocked with sulfydryl-specific probes, these may also partially block the other cysteinyls. Thus we were interested in removing this residue by proteolysis to facilitate the labeling of other cysteinyls. Previous investigators have used proteolysis to modify this region of actin. Trypsin normally digests actin at several sites (61-62, 68-69, and 373-374), but the 373374 site is the more reactive (12). Suck et al. (13) found that mild trypsin digestion of the G-actin:DNase I complex removed only residues 373-375. This mild trypsin treatment improved the quality of the crystals obtained 0003-9861/92
110
$3.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction
in any form reserved.
REMOVAL
OF ACTIN
C-TERMINAL
RESIDUES
(13) and thus was used in deriving the recent atomicresolution structure (10). Carboxypeptidase A has been used to remove Phe-375 and then Cys-374 from actin (14,15). Removing Phe-375 alone has little effect on actin or on its interaction with myosin. However, the conditions for removing this cysteine residue have not previously been defined, and the effect of this proteolytic cleavage on actin has not been characterized. We found that Cys-374 residues were removed from all monomers after 18 h digestion with 1 mol carboxypeptidase A per 10 mol of G-actin (S. I. O’Donoghue, unpublished data). However, we decided to use trypsin rather than carboxypeptidase A for these experiments since it is cheaper and easy to obtain in the quantities needed. To date, there has been no systematic investigation of the effect of removing residues 374 and 375 on the structure and functions of actin. To be a useful technique, it is important that this modification does not greatly alter the conformation of actin or its ability to polymerize and interact with myosin. If there is an effect on any of these properties, this would have important implications for accepting the X-ray crystal structure of actin. Our aim in this paper is to define conditions for removing these residues using trypsin and to characterize the effects of this modification on the functions of actin. We found that removing these residues caused only subtle changes in the secondary structure of actin and its interaction with other proteins. However, we found major changes in the filament shape as seen by electron microscopy and in the viscosity of F-actin. These changes suggest that the normally rigid actin filament is more flexible when these residues are removed. MATERIALS
AND METHODS
Reagents. Bovine trypsin (type XI) and TLCK (1-chloro-3-tosylamido-7-amino-Z-heptanone hydrochloride (N-p-tosyl-L-lysinechloromethyl ketone)) were purchased from Sigma Chemical Co. Carboxypeptidases A and B were obtained from Boehringer-Mannheim Biochemicals. Pyrene (N-(l-pyrenylhodoacetamide), IAEDANS (N-iodoacetyl-iV’-(l-sulfo-5-naphthyl)ethylenediamine), and IAF (1,5-iodoacetamidofluorescein) were purchased from Molecular Probes, Inc. Actin was prepared from rabbit skeletal muscle according to the method of Spudich and Watt (16). ADP-G-actin was prepared as described by Pollard (17) using hexokinase from Boehringer-Mannheim. Pyrene-actin (labeled at Cys-374 with pyrene) was prepared according to the method of Kouyama and Mihashi (3). The molar ratio of attached pyrene to actin was 0.95. IAF-actin (labeled at Cys-374 with IAF) was prepared according to the method of Wang and Taylor (18). The molar ratio of attached IAF to actin was usually 0.6. Myosin was prepared according to the method of Tonomura et al. (19). Myosin subfragment 1 (Sl) was prepared by chymotryptic digestion of rabbit skeletal myosin according to the method of Weeds and Pope (20). Sl was either freeze-dried in the presence of 0.1 M sucrose and stored at -2O’C or kept in 50% glycerol at -20°C. Tropomyosin, troponin, and regulated F-actin were prepared as described previously (21). IAEDANS-labeled tropomyosin (1.4 mol IAEDANS per mole of tropomyosin) was prepared as described previously (22, 23). The concentration of unmodified G-actin was determined using an extinction coefficient of 26.6 X lo3 M-’ cm-i at 290 nm (24). The Sl concentration was determined using an extinction coefficient of 82.5 X
AFFECTS
FILAMENT
STRUCTURE
111
lo3 Me1 cm-’ at 280 nm (20). Optical absorbence measurements were made with a Philips PU 8800 spectrophotometer. Protein concentrations were also determined by the Bradford method using the Pierce Coomassie blue G-250 protein assay reagent. The molecular masses of Sl and actin were taken to be 110 kDa (25) and 42.3 kDa (26), respectively. Trypsin digestion. F-actin (approximately 50 PM) was digested with trypsin (1 mol of trypsin to 5 mol of actin) in F-buffer (2 mM MgClz, 0.2 mM ATP, 3 mM NaN,, 20 mM KCl, and 20 mM Tris-HCl, pH 7.6) for 15-25 min at room temperature (about 22°C). The reaction was stopped by adding a IOO-fold molar excess of TLCK over trypsin. The actin was then taken through a polymerization/depolymerization purification cycle. This involved pelleting F-actin in a centrifuge at 100,OOOg (g = 9.8 m s-i) for 90 min at 20°C resuspending in F-buffer, centrifuging again, resuspending the pellet in G-buffer (0.2 mM CaCl,, 0.2 mM ATP, 3 mM NaN,, and 2 mM Tris-HCl, pH 8.0), dialyzing exhaustively against G-buffer, and, finally, clarifying the solution of insoluble impurities by centrifuging (100,OOOgfor 90 min at 4’C). The total amount of protein obtained from this digestion was about 60% of the starting amount or less, depending on the activity of the trypsin. The above method was designed to produce a “clipped” actin missing residues 374-375. For each clipped actin preparation, two methods were used to check the extent of 374-375 removal. The first was to add a lofold molar excess of IAF to the clipped actin (in G-buffer). After 3 h the mixture was dialyzed overnight versus F-buffer to remove excess label and then centrifuged (100,OOOgfor 90 min at 20°C). Samples from the pellet were used to detect the presence of IAF bound to actin with 12% w/v sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) according to the procedure of Laemmli (27). The second method used was to add an equimolar amount of Cu+* to the clipped actin and measure the change in optical absorption at 347 nm. While actin does not normally absorb light of this wavelength, upon binding copper an absorbence peak appears centered at 347 nm with an extinction coefficient of 3.0 X lo3 M-’ cm-’ (28). Copper binding is lost upon removal of Phe-375 (29); hence, the extent of 375 removal can be determined from the loss of 347-nm absorbance. Carboxypeptidase B digestion. The clipped actin produced from the trypsin digest was further digested with carboxypeptidase B for 18 h at 4”C, polymerized by the addition of 10X F-buffer, and pelleted by centrifuging as described above. The supernatant was cleared of protein by adding pechoric acid (to 5% w/v) and analyzed for the presence of amino acids by high-performance liquid chromatography using a precolumn derivatization method described by Haynes et al. (30). Characterization of the actin fragment. Fourier-transform infrared spectroscopy (FTIR) was used to estimate the secondary structure features of actin as described by Susi and Byler (31). Spectra were obtained for clipped and unmodified actin in G-buffer using a DIGILABS FTS20/80 spectrophotometer equipped with a HgCdTe detector. Fourier transforms and analysis of these spectra were done using DIGILAB software. The critical concentration for clipped actin was measured by copolymerizing with pyrene-actin as described by Pollard (32). In this method, the large increase in pyrene-actin fluorescence upon polymerization is used to measure the extent of polymer formation. Samples were placed in l-cm quartz cuvettes (stirred and temperature regulated at 15’C) and fluorescence intensities were measured using an SLM 8000 spectrofluorimeter operating in ratio mode. The superprecipitation of actin and myosin was studied using lo-ml samples of F-actin (0.02 mg/ml) and myosin (0.2 mg/ml) in 25 mM KCl, 1 mM MgCl,, and 20 mM Pipes (pH 7.05). These were thoroughly mixed in a Teflon-glass homogenizer and transferred to a 2 X 2-cm glass cuvette (continuously stirred and temperature regulated at 23’C). A total of 10 ~1 of 10 mM ATP was added to start superprecipitation. The extent of superprecipitation was determined from the change in light scattering at 650 nm. The rate was determined by the time required to reach half of the maximum change in scattering. The maximal change varied by about 10% between samples. ATPase measurements were carried out with a Radiometer pH-stat apparatus (ABU 12; TTT 2; SBR 3; PHA 943) at pH 8.0 using 2 mM
112
O’DONOGHUE,
MIKI,
KOH as a titrant. The conditions were 0.05mg/ml Sl, 30 mM KCl, 2.5 mM ATP, and 1 mM Tris-HCl (pH 8.0) at 23°C. The concentration of actin was varied. The reaction was started by adding Sl to the 2-ml samples. The ATPase activation of regulated actin was measured by preparing a mixture of Sl, actin, tropomyosin, and troponin at 0.05 mg/ml, 0.25 mg/ml, 0.086 mg/ml, and 0.087 mg/ml, respectively. The measurements were done as above but in the presence of Ca+s (0.1 mM CaCl,) or in the absence of free Ca+’ (with 2 mM EGTA). The association of F-actin (1.7 mg/ml) with IAEDANS-tropomyosin (0.2 mg/ml) in F-buffer was examined by cosedimentation at - 180000g for 30 min in a Beckman airfuge. The concentration of labeled tropomyosin in the supernatent (= [Tm]r, the concentration of unbound tropomyosin) was determined by measuring the fluorescence intensity at 480 nm with the SLM fluorimeter. The excitation wavelength was 360 nm. Since the total concentrations of actin and tropomyosin ([A], and [Tm],) are known, the association constant (K.) can be determined from [Tm]r using the formula:
1Tmlf = (lTm1, - [Al, - K’ + v(K;’
+ [A], - [Tm],)2 + 4[Tm]JK,)/2.
This was derived by substituting the equations [Tmlf = [Tm], - [A.Tm] and [AIf = [A], - [A.Tm] into the definition of K,,, [A.Tm] K’ = [Tmlf[Alf
DOS REMEDIOS
a
1
b
2
1
2
FIG. 1. SDS-PAGE of trypsin-digested actin treated with IAF. The gel in Fig. la shows trypsin-digested actin (lane 1) and undigested actin (lane 2) stained with Coomassie blue to show total protein. Both samples had been treated with IAF as described under Materials and Methods. Figure lb shows the same gel illuminated with a uv lamp prior to staining. The fluorescent band corresponding to undigested actin is due to IAF bound to Cys-374. Clearly, trypsin-digested actin has not been labeled by IAF, indicating that the limited proteolytic digestion resulted in essentially complete removal of Cys-374.
’
where [A.Tm] is the concentration of the actin/tropomyosin complex, and [Tmlf and [AIf are the concentrations of unbound tropomyosin and actin. A residual sum of squares (RSS) function was defined which measured the differences between the measured and the calculated [Tmlf values. The value of K. which best fitted the RSS function was found by nonlinear minimization. F-actin was diluted to 0.05 mg/ml and quickly Electron nicroscopy. transferred onto carbon-coated parlodion-filmed grids. After blotting off excess sample, the grids were washed with 0.1 M ammonium acetate and stained with 1% uranyl acetate for 30 s. The samples were viewed in a Philips EM 301 electron microscope operated at an accelerating voltage of 80 kV. Viscosities were measured using an Ostwald Viscosity mmsurements. capillary viscometer (Cannon Corp.) with a sample size of 0.5 ml and a flow time for water of 25 s at 25°C. The specific viscosity (s,) was calculated from
%p =
AND
(
flow time of actin sample _ 1 flow time for water 1’
RESULTS Trypsin Digestion The trypsin digestion procedure and subsequent polymerization/depolymerization cycle outlined under Material and Methods resulted in a “fragment” which ran as a single band on SDS-polyacrylamide gels with the same apparent molecular weight as unmodified actin (see Fig. 1). This indicated that most residues were still intact; hence we called this “clipped” actin. This clipped actin had lost all ability to bind IAF as determined by fluorescence SDS-PAGE (see Fig. l), indicating that most of the Cys-374 had been removed. To quantitate the extent of this removal, the change in optical absorbence at 347 nm was measured upon adding copper. Only a small
change in absorbence was detected for all clipped-actin preparations (less than 5% of the absorbence change for unmodified actin), indicating that over 95% of the actin in these preparations was missing its C-terminal residue (Phe-375). The Edman method was used in an attempt to degrade the trypsin-treated fragment with phenyl isothiocyanate. The fragment would not degrade, indicating that the Nterminus was blocked as in unmodified actin. This suggests that the first residue of the fragment is the acetylated Asp-l found in untreated actin. Carboxypeptidase
B Digestion
The clipped actin was further digested using carboxypeptidase B to identify the residues at the C-terminus. Using HPLC to detect amino acids released by this digestion, only two were detected; lysine (0.89 mol per mole of actin) and arginine (0.84 mol per mole of actin). Since residues 372 and 373 in the actin sequence are arginine and lysine, respectively, this result confirms that the trypsin digestion removed only residues 374 and 375. Thus, on the basis of the known trypsin cleavage sites on actin (12) and the known amino acid sequence of actin, we conclude that the limited trypsin digestion removed only residues 374-375. The clipped actin produced by the trypsin digest, residues l-373, is hereafter denoted actin-z to indicate removal of the last two residues. Characterization
of the l-373 Actin Fragment
By measuring the optical absorption of G-actin at 290 nm and comparing it to the concentration calculated by the Coomassie blue assay, we determined that G-ac-
REMOVAL
OF ACTIN
C-TERMINAL
RESIDUES
AFFECTS
FILAMENT
113
STRUCTURE
Using the coprecipitation method outlined under Materials and Methods, the affinity constant for IAEDANStropomyosin binding was determined to be K, = 99 X lo3 M-’ for F-actin and K,, = 196 X lo3 M-’ for native Factin (see Fig. 3). Electron Microscopy
l/lactinl
(JIM’ )
plot of Sl ATPase activity versus actin FIG. 2. Double-reciprocal concentration. The maximal rate of Sl ATP hydrolysis (i.e., the inverse of the ordinate intercept) is 2.8 pmol Pi min-’ mg-’ in the presence of actin (0) and 3.5 pmol Pi min-’ mg-’ in the presence of unmodified actin (m). The rate for Sl alone was 0.04 pmol Pi min-’ mg-‘. Both graphs intersect the abscissa at the same point which indicates that removing 374-375 did not affect the affinity of actin binding to Sl.
tin-z had the same extinction coefficient as unmodified G-actin. The secondary structure predicted from FTIR for Gactin-z was 31% a-helix, 47% &sheet, and 22% random coil. This compares with 32% a-helix, 43% P-sheet, and 25% random coil predicted for control G-actin. These differences (~4%) are probably not significant since the accuracy of the FTIR technique in predicting secondary structure is from 0 to 5% (31). We therefore conclude that the secondary structure of actin was not greatly altered by removing residues 374 and 375. The critical concentrations of unmodified actin and actin2 in F-buffer were determined by copolymerization with 4% pyrene-actin at six different actin concentrations (between 0 and 5 PM). The critical concentration for unmodified actin was 0.7 f 0.2 PM, consistent with the value reported by others (32). The critical concentration for actin was 0.27 +_0.04 PM (these values and standard errors were calculated from a linear regression of six points). We found no significant difference in the extent of superprecipitation of myosin with F-actin and with F-actin-, as determined from the 650-nm absorbence. The time for half-maximal change in absorbence varied from 1 to 3 s for both samples. Figure 2 shows a double-reciprocal plot of Sl ATPase activity versus concentration of actin. Removing residues 374 and 375 did not affect the affinity of actin binding to Sl, but reduced the maximal activation of the Sl ATPase from 3.5 to 2.8 pm01 Pi min-’ mg-‘. These ATPase measurements were repeated with regulated actin and actin (i.e., complexed with tropomyosin and troponin). For both actin samples, activation of Sl ATPase was not inhibited in the presence of Ca+‘. However, in the absence of Caf2, the activation by F-actin-2 was inhibited by about 50% while the activation by Factin was inhibited by approximately 80%.
At high magnification, unmodified actin filaments (Fig. 4c) are long and straight with a rather low curvature and have a uniform diameter. The filaments are mostly separate from each other, although sometimes several filaments (usually just two) run parallel to form straight multifilament cables. In contrast, the filaments of actin at the same magnification (Fig. 4d) have a kinked appearance, and a greater and more irregular curvature. These filaments appear to have a less uniform diameter, although this is difficult to quantitate since they are usually twisted around each other. The electron micrographs we obtained for F-actin-, look very similar to those published by Janmey et al. (1) for filaments assembled from ADP-containing monomers (hereafter denoted ADP-Factin, depending on the nucleotide bound to the monomer). Normally, F-actin is assembled from monomers containing ATP (ATP-F-actin). At lower magnification, the difference between filaments of unmodified actin (Fig. 4a) and actin (Fig. 4b) is dramatic. The filaments of actin are mostly entwined together into multifilament bundles. Figure 4d shows one of these bundles (going from the bottom left to the top right corner) at higher magnification. The arrangement of filaments in these bundles is very irregular. Viscosity Measurements
Figure 5 shows the specific viscosity as a function of actin concentration for ATP-F-actin, ADP-F-actin, and
Fluorescence intensity (arbitary units) 0.4.. 0.2.. 1
1
2
3
4
5
6
FIG. 3. Binding of IAEDANS-tropomyosin to F-a&n-, (0) and native F-actin (m). These data show the loss of fluorescence intensity in the supernatent due to coprecipitation with increasing molar ratios of Factin. These data were used to find the K,, for the association of tropomyosin with actin as described under Materials and Methods. The curves are calculated from the best-fitting values of K,--99 X lo3 M-’ for Factin-* (RSS = 0.27%) and 196 X lo3 M-’ for native F-actin (RSS = 0.572).
114
O’DONOGHUE,
MIKI,
AND
DOS REMEDIOS
REMOVAL
OF ACTIN
C-TERMINAL
RESIDUES
F-actin-, solutions. At high concentrations (>50 PM), the viscosity of ADP-F-actin and F-actinpa was about half that of ATP-F-actin. This difference in viscosity could easily be seen by slowly inverting the solution and watching it flow. The data in Fig. 5 are for a single preparation of actinep, but the same reduction in viscosity was measured in all F-actinpz preparations.
AFFECTS
FILAMENT
115
STRUCTURE
7
l8 T
Specific viscosity
1o
DISCUSSION
8
The limited trypsin digestion resulted in an actin fragment consisting of residues l-373. The evidence for this conclusion is that the trypsin-treated actin lost its ability to bind IAF, indicating that Cys-374 had been removed, but had the same extinction coefficient and apparent molecular weight as unmodified actin. Furthermore, the Nterminus was blocked as in unmodified actin and digestion with carboxypeptidase B released lysine and arginine corresponding to residues 373 and 372 in the actin sequence. This conclusion contrasts with the finding of Suck et al. (13), who found that trypsin removed Lys-373 as well as residues 374-375. This disparity is not surprising since their digestion conditions were different to ours; they digested the G-actin:DNase I complex, while the digestions described here were done with F-actin. It is quite plausible that Lys-373 is protected from cleavage in F-actin, but more exposed in G-actin. Removing the last two residues did not greatly alter the secondary structure of actin or its ability to polymerize. This suggests that the tertiary structure was mostly unaffected by this modification. We found a significant difference in the critical concentration required for polymerization to occur. Other modifications of this region also lower the critical concentration (33). A small but significant difference was also found in the maximal rate of Sl ATPase activation. This is consistent with previous reports that the C-terminal region is involved in the interaction of actin and myosin (3, 34) but is not critical for this interaction (14). The regulation of the actomyosin interaction by tropomyosin and troponin appeared to be slightly impaired by this modification. However, this may be partially explained by later experiments which showed that the affinity of F-actinez binding to tropomyosin was lower than that for native F-actin. This reduced binding affinity may be due to steric blocking of F-actinMp by the formation of the filament bundles seen in Figs. 4b and 4d. The variable curvature of F-actinqa filaments, as seen under the electron microscope at high magnification, indicates that these filaments were more flexible than unmodified actin filaments. Janmey et al. published very
6
0
20
40
60
80
100
[actin] (KM) FIG. 6. Flow viscosity measurements. The specific viscosity is plotted as a function of actin concentration for ATP-F-actin (m), ADP-F-actin (X), and F-actin-, (0).
similar electron micrographs of actin polymerized from monomers containing ADP rather than ATP (1). They also concluded that the difference was due to increased filament flexibility. Normal actin filaments can form interfilament crosslinks by self-association (35) which result in the parallel cables as seen in the electron microscope (Fig. 4~). The basis of this contact is probably electrostatic attraction. Removing residues 374-375 appears to increase the extent of filament cross-linking. It is possible that this is caused by an extra, specific, filament-filament interaction, perhaps involving the three basic residues 371-373 which may be more free to interact with acidic residues on adjacent filaments. A more likely explanation is that the electrostatic interaction, which results in parallel cables of unmodified actin, is complemented by the mechanical intertwining of the filaments which results from the increased flexibility of F-actin-,. This may explain the emergence of the disordered filament bundles seen at low magnification in the electron microscope. The decrease in flow viscosity may be partially accounted for by the bundling of F-actinez. However, the viscosity decrease may also be explained by filament breakage. Janmey et al. (1) found that mechanical breakdown of ADP-F-actin occurs even when very low strain (less than 1%) is applied to the solution. In contrast, ATPF-actin will not break down until a relatively large strain is applied (over 10%). We could not determine the strain
of F-actin-, and unmodified F-actin. At low magnification, it is apparent that F-actin-, filaments (b) aggregate into FIG. 4. Ultrastructure bundles, while unmodified F-actin filaments (a) form a loose network. At higher magnification it is clear that the F-actin+ filaments (d) are more entwined about one another and more variable in curvature than F-actin filaments (c). All samples were prepared under identical conditions. Scale bars are 0.50 pm.
116
O’DONOGHUE,
MIKI,
amplitude exerted by the viscometer on the solution; however, it was likely to be greater than 1%. Hence, the decrease in viscosity of ADP-F-actin measured in this paper is probably due to filament breakage. Since F-actin-z has a similar viscosity decrease, actin-z filaments may also be more fragile than ATP-F-actin; this agrees with the results obtained by Stournaras et al. (36), who found that chemical modification of Cys-374 decreased the mechanical stability of F-actin. However, this can not be established without further, more sophisticated, rheological measurements as done by Janmey et al. (1). While we cannot be certain of the cause of the apparent increase in flexibility and fragility of F-actin-P, it seems likely to result from weakening of intermonomer bonds. This idea is supported by the filament model proposed by Holmes et al. (ll), in which one of the major intermonomer contacts is formed by residue 374 (and residues 166169) with residues 41-45 on the next monomer. Other support comes from a variety of physicochemical techniques: Elzinga and Phelan (7) and Sutoh (8) showed that Cys-374 can be cross-linked to Lys-191 on an adjacent monomer. They concluded that the distance between these two residues must be 0.8-1.4 nm. The NMR probe studies of Barden et al. (6) show that Cys-374 experiences a large decrease in mobility following polymerization, consistent with it being close to an intermonomer binding site. These studies suggest that removing residues 374375 may disrupt an important intermonomer contact, causing a weakening of the filament which may account for the increased filament flexibility which we observed. Since the rheological and ultrastructural effects of removing residues 374 and 375 are so similar to those observed for ADP-F-actin, we may speculate that removing residues 374 and 375 somehow promotes hydrolysis of the ATP bound to G-actin so that filaments are then assembled from ADP monomers which would thus be more flexible. Alternatively, Cys-374 may be involved in one of the bonds which are strengthened by ATP hydrolysis after polymerization. However, these explanations seem unlikely since both X-ray crystallography (10) and fluorescence resonance-energy transfer measurements (2, 4, 5) place Cys-374 a long way (about 3 nm) from the nucleotide site of actin. In summary, our data show that removing residues 374 and 375 did not greatly alter most properties of actin but increased the flexibility of the actin filaments and promoted filament bundling. This supports the idea that the C-terminus is directly involved in actin-actin contact in the filament. ACKNOWLEDGMENTS We gratefully acknowledge the help of Drs. Ernest and Dallas MacFarlane from the University of Sydney, who made the FTIR measurements, and Dr. Andrew Gooley from Macquarie University, who did the amino acid analysis.
AND
DOS REMEDIOS
REFERENCES 1. Janmey, P. A., Hvidt, S., Oster, G. F., Lamb, J., Stossel, T. P., and Hartwig,
J. H. (1990) Nature 347,95-99.
2. Miki, M., and Mihashi, K. (1977) Biochim. Biophys. Acta 633,163172. 3. Kouyama, T., and Mihashi, K. (1981) Eur. J. B&hem. 114, 3338. 4. Cheung, H. C., and Liu, B. M. (1984) J. Muscle Res. Cell Motil. 5, 65-80. 5. Miki, M., and Wahl, P. (1984) B&him. Biophys. Acta 786, 188196. 6. Barden, J. A., Phillips, L., Cornell, B. A., and dos Remedios, C. G. (1989) Biochemistry
28,5895-5901.
7. Elzinga, M., and Phelan, J. J. (1984) Proc. Natl. Acad. Sci. USA 81,6599-6602. 8. Sutoh, K. (1984) Biochemistry 23,1942-1946. 9. Tellam, R. L., Morton, D. J., and Clarke, F. M. (1989) Trends Biochem. Sci. 14,130-133. 10. Kabsch, W., Mannherz, H. G., Suck, D., Pai, E. F., and Holmes, K. C. (1990) Nature 347,37-44. 11. Holmes, K. C., Popp, D., Gebhard, W., and Kabsch, W. (1990) Nature
347,44-49. 12. Jacobson, USA 73, 13. Suck, D., Sci. USA 14. Field, V.
G. R., and Rosenbusch, J. P. (1976) Proc. N&l. Acad. Sci. 2742-2746. Kabsch, W., and Mannherz, H. G. (1981) Proc. N&l. Acad. 78,4319-4323. L. S., and Bowen, W. J. (1968) Arch. B&hem. Biophys.
127,59-64. Malm, B. (1984) FEBS L.&t. 173,399-402. Spudich, J. A., and Watt, S. (1971) J. Bid. Chem. 216,4666-4871. Pollard, T. D. (1984) J. Cell Biol. 99, 769-777. Wang, Y., and Taylor, D. L. (1980) J. Hi&o&m. Cytochem. 28, 1198-1206. 19. Tonomura, Y., Appel, P., and Morales, M. F. (1966) Biochemistry 15. 16. 17. 18.
5,515-521. 20. Weeds, A. G., and Pope, B. (1977) J. Mol. Biol. 111, 129-157. 21. Miki, M. (1987) Eur. J. Biochem. 164, 229-235. 22. Lamkin, M., Tao, T., and Lehrer, S. S. (1983) Biochemistry 22, 3053-3058. 23. Miki, M., and Hozumi, T. (1991) Biochemistry 30,5625-5630. 24. Houk, T. W., and Ue, K. (1974) And. Biochem. 62,66-74. 25. Margossian, S. S., Stafford, W. F., and Lowey, S. (1981) Biochemistry 8,2151-2155. 26. Collins, J. H., and Elzinga, M. (1975) J. Biol. Chem. 250, 59155920.
27. Laemmli, U. K. (1970) Nature 227,680-685. 28. Lehrer, S., Nagy, B., and Gergely, J. (1972) Arch. Biochem. Btiphys. 150,164-174. 29. Drabikowski, W., Lehrer, S., Nagy, B., and Gergely, J. (1977) Arch. Biochem. Biophys. 181,359-361. 30. Haynes, P. A., Sheumack, D., Kibbly, J., and Redmond, J. W. (1991) J. Chromatogr. 640, 177-187. 31. Susi, H., and Byler, D. M. (1986) in Methods in Enzymology (Hirs, 32. 33. 34. 35. 36.
C. H. W., and Timasheff, S. N., Eds.), Vol. 130, pp. 290-327, Academic Press, San Diego. Pollard, T. D. (1984) J. Muscle Res. Cell Motil. 4, 253-262. Tait, J. F., and Frieden, C. (1982) Biochemistry 24,6046-6053. Brauer, M., and Sykes, B. D. (1986) Biochemistry 25,2187-2191. Griffith, L. M., and Pollard, T. D. (1982) J. Biol. Chem. 267,91359142. Stournaras, C., Drewes, G., Blackhohn, H., Merkler, I., and Faulstich, H. (1990) Biochim. Biophys. Acta 1037,66-91.
